geodiversitas fait suite,

avec la même tomaison, au Bulletin du

Muséum national d’Histoire naturelle,

4® série, section C, Sciences de la Terre.

Rédacteur en chef :

Hervé Lelièvre

Conseil éditorial :

Annemarie Ohler (Nomenclature)

Assistante de rédaction :

Florence Kerdoncuff

Corrections-relecture :

Hélène Bertini

Sandrine Hoffart-Muller

Juliette Thomas

Comité scientifique :

Jean Broutin (UPMC, Paris)

Bruno David (CNRS, Dijon)

Jean-François Deconinck (USTL, Lille)

Patrick De Wever (MNHN, Paris)

Jean Marcoux (U. Denis Diderot, Paris)

Christian Ravenne (IFP, Paris)

Abonnements pour l’année 1998 (prix HT)

Annual subscription rates 1998 (excluding VAT)

Abonnement général / General subscription : 1800 FF

zoosystema : 800 FF

adansonia : 500 FF

geodiversitas : 800 FF

geodiversitas peut être obtenu par voie d'échange.

Pour toutes informations s’adresser à :

geodiversitas may be obtained on an exchange

basis. For further information please write to :

Service des périodiques et des échanges de la

Bibliothèque centrale du Muséum national

d’Histoire naturelle

38 rue Geoffroy Saint-Hilaire

75005 Paris

Tél. : (33) 01 40 79 36 41

Fax : (33) 01 40 79 36 56

geodiversitas

Editions scientifiques du Muséum, Paris

Photocopies :

Les Publications Scientifiques du Muséum adhèrent au Centre

Français d’Exploitation du Droit de Copie (CFC), 20 rue des

Grands Augustins, 75006 Paris. Le CFC est membre de

Y International Fédération of Reproduction Rights Organisations

(IFRRO). Aux États-Unis d’Amérique, contacter le Copyright

Clearance Center, 27 Congress Street, Salem, Massachusetts 01970.

d’Histoire naturelle, Paris, 1998

Photocopies:

The Scientific Publications of the Muséum adhéré to the Centre

Français d’Exploitation du Droit de Copie (CFC), 20 rue des Grands

Au^stins, 75006 Paris. The CFC is a member of International

Fédération of Reproduction Rights Organisations (IFRRO).

In USA, contact the Copyright Clearance Center, 27 Congress Street,

Salem, Massachusetts 01970.

Archaeocyathids from the Atdabanian (Lower

Cambrian) of the Altay-Sayan Foldbelt, Russia

Dena V. OSADCHAYA & Dmitriy V. KOTEL’NIKOV

All-Russian Geological Institute, Sredniy prospekt 74,

Saint Petersburg 199026 (Russia)

Osadchaya D. V. & Kotel’nikov D. V. 1998. — Archaeocyathids from the Atdabanian (Lower

Cambrian) of the Altay-Sayan Foldbelt. Russia. Geodiversiîas 20 (1) : 5-18.

KEYWORDS

Archaeocyathids,

Atdabanian,

Altay-Sayan,

corrélations.

ABSTRACT

The stratigraphie distribution of Irregular archacocyaths (archaeocyathids)

species in the Atdabanian Stage of the Altay-Sayan Foldbelt is established.

One assemblage characterizes the Bazaikha Superhorizon and two assem¬

blages are observed in the Kameshki Horizon. These assemblages are consis¬

tent independently of the faciès type and traceable throughout Batencvskiy

Ridge (carbonate faciès), Tuva (volcanic-carbonate faciès) and Eastern Sayan

(siliciclastic-carbonatc faciès). This observation confirms the existence of a

single Altay-Sayan basin by the middle Atdabanian dme. Three new species

(Loculicyathus voznâscnsküy Mikhnocyathus irregularis, Sakhacyathus

karpinskit) are described and the most typical species are revised and redescri-

bed.

MOTS CI^

Archeocyathida,

Atdabanien.

Altai-Saïan,

corrélations.

RÉSUMÉ

Archéocyathes de VAtdabanien (Cambrien inferieur) du massif plissé d'Altaï-

Satan, Russie. La répartition stratigtaphique des archéocyarhes irréguliers

(archéocyathides) du massif plissé d’Alraï-Saïan esc établie pour l’Atdabanien.

Un assemblage caractérise le Superhorizon de Bazaikh et deux assemblages

sont observés dans l'Horizon de Kameshki. Ces assemblage.s concordent,

indépendamment du type de faciès, cf peuvent sc suivre à travers la crête de

Batenev (faciès carbonate), la région de Tuva (fiiciès volcano-carbonaté) et le

Sai'an oriental (faciès silicoclascique et carbonate). Cette observation confirme

l'existence d’un seul bassin dans rAltaï-Saïan au milieu de J’Aidabanien.

Trois nouvelles espèces sont décrites (Loculicyathus voznesenskiiy

Mikhnocyathus irregularis, Sakhacyathus karpinskit) ; les autres espèces sont

révisées et redécrites.

GEODIVERSIIAS • 1998 • 20(1)

Osadchaya D. V. & Kotel’nikov D. V.

INTRODUCTION

The Alcay-Sayan Foldbelt is a mountaiiieous

région of Siberia (SW of thc Siberian Platform).

It iiicludes Kuznerskiy Alatau, Batenevskiy

Ridge, Altay Mountains, Shoriya Mountains,

Eastern and Western Sayan and Tuva. Thick and

continuüus Lowcr Cambrian strara of rhe

Altay-Sayan FoJdbeli are, in addition to those of

the Siberian Platform, a kcy sequence tor ihe

Lower Cambrian stage and zone subdivisions

that arc acceptcd in Russia; Botomian and

Toyonidn archaeocyuthan zones are establishcd

here. Howcver, ihese zones were eiuirely based

on regular archaeocyaths whose géographie dis¬

tribution is‘ limiied and, as a result, iheir signifi-

cance Jor corrélations.

Irregular archaeocyaths (archaeocyathids) of the

Altay-Sayan Foldbelt were poorly known. IVjo

many species were commonly and wrongly attri-

buted to the gênera ^Dictyocyathus" and

Loculuyathus'. Cravesiock (19S4) iniriated nevv

researches into the morphology of irregular

archaeocyaths and this was cxpnnded in che

major hook ol Debrenne 6 l Zhuravlev (1992).

The work of the larter autbors develops a nevv

systematic o( irregular archaeocyaths based on

morphology, onlogcny and homologous variabi-

lity; they provide a complété diagnosis, revised

systematic affiiiities and an outlinc of stratigra¬

phie distribution of ail généra. There, we révise

the Atdabanian irregular archacocyath -assem¬

blages ol the Altay-Sayan Foldbelt in the framc-

work of this new systematics based on abundanr

material that we collected front Batenevskiy

Ridge, Eastern Sayan and luva, mainly from the

stratotypc.s of zones of the Bazaikha

Superhorizon and Kameshki Horizon. These

local subdivisions of the Altay-Sayan Foldbelt

correspond to the Atdabanian Stage.

GEOLOGICAL SE I FING

BATENtvsKrv RnXtF. (Sukhie Solontsy

MASS iU-CARnoNAi i- Sec tion Tyet

Calcimicrobial and calcimicrobial-archaeo-

cyathan rcefal carbonates dominate in the

Batenevskiy Ridge. These are massive gray and

light gray limestones with unevenly spaced fos-

sils. Irregular archaeocyaths arc studied from rhe

Solontsy Biohermal Massif in Sukhie Solontsy

Valley, Tolcheya Village area (Fig. 1). The

Solonrsy Biohermal Massif spreads latirudinally

for 8 km as a discontinuons belt of Power

Cambrian recfal limestones from 1000 to 1200 m

in thickness. The Atdabanian .strata are 600 m in

îhicknes.s.

rbc Nochorolcyaihus marivnkn Zone is at the

hase of thc Bazaikha Horizon in Batenevskiy

Ridge. The most complète assemblage of this

zone is derived from the référencé section of

Krutoy Log. Addiiional data arc obtaincd from

scction.s of 740.6 m. 786 m and 803.5 m alti-

Tudes. ] he Nochouncyûihtis manhnkii Zone is

rypified by gênera Qimhrotyaihellusy Neoloctili-

cyarbtiis DictyocyatbtiSs Diclyosycon. lalmlacya-

ibellus and L^irtyofatnis. l'hc most comtnon

species are Gimbnnymhellns commnfiis (Fonin),

C. niuiliseptu'^ (Voronin), C tifbrrcHlnti4s

(Vokigdln), C. kundnius (Zhuravieva), C. mlnu-

tui (Vologdiri), Neolacnllc'^ailnis primus Voronin,

Loculicydtbus rncmbranivoitites Vologdin,

Dictytfcyatbus conférons Fonin, Arrhaeopharetra

mnrginatii (Fonin), Divtyosycon radintus

(Zhuravieva). Diityofavus sp., D. kpidits (Fonin),

D. ühiîisîts Gravxstock, Tahxdacynthellîts sp. and

l\ bidjaeusis Missarzhcvskiy. Species of

Qimbwtyathelliis arc cspccially abondant at this

Icvcl.

Froin the ba^t of the followtng Gordinikyathus

howeUi Zone, the former archaeocyathan assem¬

blage steadily déclinés. Single repre.sentativês of

the earlier assemldage only exist at thc top of the

Cordonhytithui howelli Zone. In the lowcr part of

thts zone Cambnxcyatbelhi% sp., C communis,

C. sinülisepmSs C. fuberçnlatns. C kuridatusy

Nadùculicyathîis prh^un. Lvculicyafhui niêntbrani-

i^Uiies^ Ptminacydthiii ? sp., Mikhdocyinhtis ivre-

gnlarh Kotcl'nikov n.sp., Dittyocyathus confertiiSy

AnhiU'ophdrêira xHurglnata. Ùiccyosycon sp.,

Dit'tyofuvm Icpidm^ Tabulacyathellus bldjncnsisy

and Kechikûcyathus ? sp. are présent. Curnbro-

cyatbdhi^ miburghimis (Vologdin) appears in the

upper part of the zone and Loculicydthus meni-

hramvesntei hecomes the most ahundant species

there (Fig. 2).

The same species is the only common archaeo-

GEODIVERSITAS • 1998 • 20(1)

Altay-Sayan archaeocyachids

Fig. 1. — Silualion of sections (Altay-Sayan Fotdbelt. Russia). Localitles 5100-5105' Solontsy Biohermal Massif, Sukhie Solontsy

Valley, Krutoy Log section. Batenevskiy Ridge; locality 369; Bazaikha River section. Batenevskiy Ridge (Eastern Sayan); locali-

ties 30, 542: Bayan Kol River section (Tuva); locality 9075: Vadi Bala section; locality 218: Terektyg Khem River section; localities

207, 208: Tapsa river Basin, H’chir River section (Central Tuva). Scale: 1/10 000 000.

cyathid of the Kameshkl Horizon in the Krutoy

Log and other sections of the Solontsy Biohermal

Massif Rare Cambrocyathellus^ Neoloculityathus

and Divtyat'yathiis occur at the base of rhe huri'

zon.

Tuva Volcanic-cakbonaiu Section Tveiî

Bayan Kol River Section

The Bayan Kol River is a Icft tribufary of the

Enise)'^ RiA'cr in Central lljva. The référence sec¬

tion occurs in the lowcr course ot the Bayan Ko)

River (Fig. 1). The lower subformarit)n of the

Bayan Kol Pormatiou which thickness is I lOÛ m

is of carly Ardabanian âge. On the left bank of

rhe Bayan Kol River, the stibformarion inchides

sandsrone-gravelstone-conglomerare unit with

restricted red and Hghr gray reefal limestones.

Reelal limestranes consîsr of kalipira, kalipttate

bioherms, biohermal and biostromal beds.

The Nochnrokyathus mariinskii Zone of the

Bazailcha Superhorizon is restricted to the basal

red reefs and is characterized here by

CiinibrncyathelliLi .sp., C. tuben ulatîiSj. C kunda-

tus, Sakhacyathus karpinskii Osadehaya et

Kotefnikov n.sp., Loculicyathus voznesenskii

Osadehaya et Kotefnikov n.sp.i Archacopharetra

mar^imttih Dictyofavus lepidus, and Dictytuyathus

confertus. Modulât furms of Carnbfocyathellus^

Arvhaeopharetra and Dictyofitinis are commun.

Cambwcyathellîis L the niost diverse genus and

Archae&pharetnt dominâtes volumentrically.

Archaeocyathan assemblage ot the GordouK

(yaihus howclll Zone is présent in the upper gra)*

reefs of the lower suhformation. OnJy rare repré¬

sentatives of the earlicr assemblage. Gamhru-

cyatlfellus tuheîxulatus, C. kundatus, Sakhaiyathus

karpinskii^ Ncoloculicyathus ex gr. prinius, and

Archaeophureira marginata are found. The

Kameshkl Horizon assemblage is not included in

rhe présent stiidy.

Tapsa River Sections

The sections of the Ttpsa River basin occur along

rhe V^îdi Bala Valley and Bolshoy Ifchir Creek in

the Cherhi Village area of the laapsa and Kaa-

Kliem rivers interfluve. Central Tuva. The Ifchir

Formation which rs iip to 2000 m in rhkkness,

represents the Atdabanian Stage here. The lower

GEODIVERSITAS • 1998 • 20(1}

Osadchaya D. V. & Kotel’nikov D. V.

part of formation (200-700 m) œntains lUac tuff-

conglomerates, niffgravelates, tuflltsand tuOiand-

stones with scarce limestone lenses. The upper

carbonate part of formation (700-1300 m)

includes dark gray platy dolomitic Ümestones with

marker beds of oolitic lime.scone, above which a

reefal limestone with archaeocyaths occurs.

Vadi BaJa Section. Archaeocyathan assemblage

of the Nocbürvicyathm rnariinskn Zone, Bazaikha

Superhorizon, is restricted to a massive reefal

limestone of the Vadi Bala Biohermal Massif and

includes Cambrocyaihdlus simÜiseptîis, C. tuber-

culatm^ C. NealoadicyathuS sp., Dictyo-

cyathiis sp., Archaeophareîra marginata, Dictyo-

sycon radïatus and Dictyofaims lepidui.

Bol shoy Il’clur Creek Section. Archaeocyaths

of the Gordonicyathus howelli Zone, Bazaikha

Superhorizon, arc found in vvhitc and red massi¬

ve reelal limestone overlaying the basal tufteon-

glomcrate, They are represented by

CambrocyuthvlUis sp., C, minutus, C. kundatus,

Loculicyathiis membranvvestites, L. voznesenskii-,

Mikhnotyatbm ? sp., Archaeopharetra marginata^

and Tuhulacyatbidbis sp., amoug which Loculi'

cyUthus mvmbratiivestites:, L. voznesenskii and

Mikbnocyaxhus ? sp. appear ac this levcl.

To the tüp of ihe Bolshoy Jl chir Crcck Section,

along the creek upstrearn, archaeocyaths of the

Kameshki Horizon arc collected front limestone

débris. These archaeocvath.s are, Cnmbrocyathel-

Fig. 2. — Distrlbiition of Atdabanian Irregular archaeocyath species in the Bazaikha Superhorizon of the Solontsy Blohermal Massif,

Batenevskiy Ridge. Square. Nochoroicyathus mariinskii Zone-, circle, Gordonicyathus howelli Zone.

GEODIVERSITAS • 1998 • 20(1)

Altay-Sayan archaeocyathids

lus sp., C. minutus, Loculicyathtis mevxbranives-

tites, L tolli Vblogdin, Mikhnocyathus trregularis

aiid Dictyocyathîis sp. A sîmilar assemblage occurs

along che Tëjcktyg Khem Creek of the Tapsa and

Kaa KJiem rivers intcrfluve. This assemblage Is

distinctive in its abundance of branching

Cambrocyuthellus minutus and Mikhnocyathus

irregularis.

Archaeocyathid assemblages of Tu va. Th us.

the following irregular archaeocyath assemb!age.s

of tlie Atdabanian âge are established in Tuva:

Fourteen species ot cight gcneta, Cambrocya-

tbellus, Neoloculuyathus^ Lociilicyathus, Sakha-

cyathusy Dictyocyathusy Dictyofavus, Dictyosycon

and Archaeopharetra characterizc the Nochoroi-

cyathus maninskïi Zone of the Bazaikha Super-

STAGE

Horizon

Zones

Archaeocyalhs

Cambrocyaihellus communis

C. tuberculus

C. similiseptus

Archaeopharetra marginata

Dictyocyaihus confertus

Neolocuhcyathus primas

Tabulacyaihelius bidzhaensis

Dictyofa^rus iepidus

Dictyosycor^ radiaius

Cambrocyathelfus Hundatus

Ardrossacvathus karpinskü

Loculicyathus vosnesenskfi

Dictyocyathus smoljaninoi/ae

Loculicyathus membranivestites

Cambrocyaihellus mmutus

Dictyofavus obtusus

Neolocuticyathus chabaHovi

Dictyosycon sp.

Tabulacyaihelius sp.

Kechikacyathus n.sp.

Cambrocyaihellus neiburgianus

Mikhnocyathus Irregularis

Dictyocyathus sp.

Neoloculicyathus sp.

Loculicyathus tolli

Cambrocyathellus sp.

Paranacyathus sp.

ATDABANIAN

Fig. 3. — Distribution of Atdabanian irregular archaeocyath species in the Altay-Sayan Fold Belt.

GEODIVËRSITAS • 1998 • 20(1)

Osadchaya D. V. & Kotel’nikov D. V.

horizon. Hiis generic ând spécifie association is

almost iclcntical (o thaï tjf the Solontsy

Biohermal Massii from rhe Ravenevskiy Ridge.

Species ol Camhrocyatht'llus cspccially diverse,

and Anhûcophciyetrtt ynurgiiHUtt dominâtes in

abondance,

Nine speciês are known from the Gordonicyatlms

howelli Zone of the Bazaikha Superhorizon in

Tuva: Cambrocyathellu^ sp., C mwutus^ C. fubet'

culatiLu C kufidatuu Neolocidicynthüs ex gr. prt-

muSy Laculicyarhns fnrmhranwsntc's^ L. vozne'

senskii, Anhaeopharetm mnrgiytata and Tabula-

cyathelliis sp. LoculiLyathus and Tabulacyaihellus

généraappeai at this level for the hrst time.

Only six species hâve beon found in the

Kameshki Horizon: CawbyocYalhellu'i sp.,

C. miiiutiiSy Lcn'ulicyâtbm uîeUibrauivesr/tcs,

L. tolli^ Mikhfwcyathus irregularis and D/ctyo-

cyathus sp. l'he archaeocyachan species and gene¬

ric diversiry decreased ai this level in Tuva> as

wcll as in the Batenevsldy Ridge. In .some sec¬

tions of’l'uva, Carnbrocyathellus minutus and

Mikhmeyathus trirgitiam becomc very common.

Easteun Sayan Silic:iclastic-Carbonai'e

Type Sectkin'

Sections of the Bazaikha River

The Bazaikha River ents che Camhrian strata in

the Northwest of Eastem 5ayan on the junction ol-

the Altay-Sayan FoldbcU and the Siberian

Platform, in the Toigashiiio Ridge. The

Torgashino Massif which occurs in the Mana

Depre.s.sion has been propo.sed to be a large reefai

massif (ZadorovJmaya 1983). The massif formed

during tbe wholc Early Cambrian and ihe begin-

ning of Middic Cambrian. Atdabanian strara

cropped oui on ihc Southern slope of tbe

Torgashino Ridge, in the Kaltar River mouih area.

The tdorhoroityathus mariinskii Zone ol rhe

Bazaikha Superhori/.on is restricied to ihe basal

Bazaikha Mentber whicli consists of red limcy

sandsrones, graveliies ;ind limesroncs. Six généra

of archacocyafbids are found In ihe member

which include Cambrotyaibellusv Neoloi ulicya-

thusy Loctilkyathus. Dictyofavusy Üictyosycon and

Sakhacyathm. The assemblage etmsisrs of rvs^elve

species which are Caminvcyathellus .sp,, C' tuber-

culatuSy C ueîhîtrguinusy C. ex gr. sirniliseptus\

NeolocuLicyathus sp., N. primusy Sakhacyathus sp.,

I.otulicyathus membranivestiteSy Diityofavus lepi-

dusy D. (d/tususj Dictyosycofi sp. and D. ex gr.

radiatm. There is strlking similarity of this

assemblage to tbe coevaJ unes dLscussed above.

The only différence is an absence of species of

Dktyùcyatbus and TabuLtcyaîbeUiis.

The Gonlunicyathus hoivelli Zone assemblage is

re.siricced to tlie upper part of ihe siliciclastic-

carbonatc Bazaikha Member and ihc lower part

of massive lighr gray reefai Umestones of che

Torgashino Formation. Fhc samc irregular

archaeocyath association occurs here, namely

CambrocyatheUtis sp., C, similiseptus. Q ex gr.

kiwdatusy Neolôculicyinhm sp.» N. prlmusy N. ex

gr. cbabakovi Rouyusbkot^ Loâdüyafbus manbm-

nimtiteiy Parafunyarhus ? stp^yAtchaeophamm sp.,

Dicîyotycôn sp., D. ex gr, radiatus and D, ex gr.

gravis. Sakhaiyatbusy Dictyocyatbus and Mikho-

cyathus appcar in the cransitional beds to the

Kameshki Horizon. Peculiarities of rhis assem¬

blage consist in a diversity of Nealoculicynthus

and Dictyosycon généra and almost complété

absence ol Dictyofavus which is abundant in the

imdcriying strata.

Cc)Nt:i-üsit)N.s

Ihc possibiliiy to use archacocyathld assem¬

blages (instoad of ajadcy'athid ones) for corréla¬

tion of Early Cambrian strata in rhe Aray^Sayan

Folcibclt is clemonstrared bere for the fitst time.

A constaiicy in the spécifie and generic composi¬

tion of Atdabanian archaeocyathid assentblages

through rhe Altay-Sayan Foldbelt is observed in

spite of a fades différence betw'een carbonates of

ihe Baiencvskiy Ridge, volcano-carbonaces of

fuva and .s'ilicidascic-carbonates of Eascern

Sayan.

7'be Nochorokyathus mariinskii Zone assemblage

déclines stcadily through che Gordonteyathus

howelli Zone strata uniil an almost complété

disappcarancc ac the Bazaikha Superhorizon-

Karneshki Elorizon boundary. At the Bazaikha

Superhorizon-Ramesbki Elorizon boundary, a

notable shift in an archaeocj'ath composition is

indicated. Loetdkyathus dominâtes from rhe base

ofthe Karneshki Horizon.

An icientity of arcliacocyaihid asscjnblagcs in ihe

Batenevskiy Pidge, Tuva and Eastern Sayan sug-

GEODIVERSITAS • 1998 • 20(1)

Altay-Sayan archaeocyathids

gests the existence of single basin during the

Atdabanian epoch throughout the whole territory.

SYSTEMATIC PALEONTOLOGY

New and revised species arc dcscribed in this sec¬

tion. Authors fbilow ihe s)^stcmadcs and termi-

nology of irregular archaeocyaths (archaco-

cyathids) developcd by Dcbrennc & Zhuravlev

(1992). The figured type marcrial is houscd in

the Mustfum national d’Histoire naturelle

(MNHN), Parisj France; other specimens are in

the Central Scientific-Research Geological

Exploring Muséum (CNIGRm), Saint Peters-

burg, Russia. Orhcr abbreviations include the

Palaeontological Insdtute, Russian Academy of

Sciences, Moscow, Russta (PIN) and the South

Au.stralian Muséum, Adélaïde, South Australia

(SAM).

Ail dimensions are in millimétrés.

Phylum PORIFERA Grant, 1872

Class ARCHAEOCYATHA

Bornemann, 1884

Order ARCHAEOCYATHIDA

Okulicch, 1935

Suborder LOCULICYATHINA

Zhuravleva, 1954

Superfaniily LOlULICYATOOIDUA

Zhuravleva, 1954

Family LOCULICYATIUDAK Zhuravleva, 1954

Genus Loadicyathus Vologdin, 1931

Loculicyathus voznesenskîi

Osadchaya et KotePnikov, n.sp.

(Fig. 4)

Holotype. - MNFIN M810055 (Fig. 4); Central

Tuva, Bayan Kol River, localiry 542-1; Lower

Cambrian, Atdabanian Stage, Baziukha Superhorizon,

Nochoroicyathm niariimk 'n Zone.

ETYXtOLOGY. — The spccies is named after the geolo-

gisc V. D. Vüzncscnskiy.

Other MATFRIAt-* — Three well-preserved specimens

in transverse sections (CNIGRm).

Disi'RIRU riÜN. — Ba/.aikha Superhorizon,

Atdabanian .Stage; Ontral Tuva.

DlAGNOStS. — Spccies wilh funncl-shaped outer wall

pores whicli plerce ihc wall in two rows per inter.sepr,

inner wall .simple bcaring spines.

Description

Cup diameter i.s up to 11.5. Outer wall simple,

0.15-0.2 thick, bcaring two rows of funnel-

shaped pores per intersept (pore diameter =

0.2-0.3, lime! width = 0.2). Intervallum (width =

2.5) contains pseudosepta which are pterced by

fivc to seven porc rows (pore diameter = 0.2-0.3,

pseudoseptum thickness = 0.1). Radial coef¬

ficient = 3.7, înterseptal coefficient = 1:5-1:6.

Inner wall simple {thickness = 0.2-0.25)i with

one pore row per intersept and spines (pore dia¬

meter = 0.4). Vesicles in the intervallum and pel-

lis on both walls are common.

Discussion

This new species diffèrs frorn Loculicyathus mem-

branivestjtes Vologdin by less numerous pore

rows per intersept of the outer wall (two againsc

four), by funnel-shaped outer wall pores and by

inner wall spines.

Genus Cambrocyathelhis I960

Cambrocyathellus simîliseptus

(Voronin, 1979)

(Fig. 5)

Robustocyathus similiseptus Voronin, 1979: 99, pl. IX,

fig- 1-3.

Cambrocyathellus similiseptus (Voronin) - Debrenne

& A. Zhuravlev 1992: 122.

Holotype. — PIN 2404-11; Voronin, 1979. pl. IX,

fig. 1 : Western Mongolia, Khasagt Khairkhan Ridge;

Lower Cambrian, Atdabanian Stage.

Other MAIERIAI- — Tweniy wcll-preserved speci¬

mens (CNIGRm).

Distribution. — Bazaikha Superhorizon,

Atdabanian Stage; Kuznetskiy Alatau, Batenevskiy

Ridge, Tuva, Easteni Sayan, Mongolia,

Df^scription

Cup diameter is up to 5-6. Outer wall simple of

GEODIVERSITAS • 1998 • 20(1)

Osadchaya D. V. & Kotel’nikov D. V.

Fig. 4. — Locuficyâthus ooinesenskii Osadchaya et Koternikov,

n.sp., hololype MNHN M01OO5. Iransverse section,

Nochorolcyathus maninskü Zone. Bazatkha Superhorizon.

Atdabanlan Stage, locality 542-1. Bayan Kol River section,

Central Tuva. Russia. Scale bar. 1 mm.

Cambrocyathellus-xypt:, 0.1 thick, bearing a single

row of rounded pores per inrersepr (pore diame-

ter = 0.1). Intervallum (widrh = 1-1.5) is filled

with pseudosepta which are pierced by four to

five pore fows (pore diameter = 0.1, pseudo-

septum thickness = 0.1). Radial coeflFicient = 7-5,

intersepr coefficient = 1:4. Inner wall simple,

with one pore row per intersepr (pore diameter =

0.15, wall rhickncs.s = 0.05). Vesicles are présent

in the intervallum.

Discussion

Cambrocyathellus similiseptiis differs from

C tuberciilatiis (Vologdin) by thicker pseudo¬

septa (0.1 against 0.03) and by thinner inner

wall (0.05 against 0.1).

Gcnxxs MikhnocyathîisMzslov 1957

Mikhnocyathus irregularis

KoternikoV) n.sp.

(Fig. 6)

Holotype. — MNHN M81006; Centra! Tuva,

Tapsa River basin, Terektyg Khcm River, locality 218;

Lower Cambrian, Aidabanian Stage, Kameshki

Horizon, Nalivkinicyathiis cyroflexus Zone.

Fig. 5. — Cambrocyatfiellus simtiiseptus (Voronin), specimen

CNIGRm 12917/4, transverse section. Nochorolcyathus mahms-

kii Zone. Bazaikha Superhorizon. Aidabanian Stage, locality

9075-2, Vadi Bala section. Central Tuva, Russia. Scale bar:

1 mm.

EtyMCÏLOGY. — ïrreguLtris (Lat.) = irregular.

OtHER MATERIAL. — Fifty well-preserved specimens

(CNIGRm).

Distribution. — Gordonicyathui hotvelU Zone,

Bazaikh-a S\iy^cxï\onio\\-Niilivkînicyathus cyroflexus

Zone, Kameshki Horizon, Atdubanian Stage;

Baicncvskiy Ridge, Tuva.

DjAGNOSIS. — Bowl-like CLips; outer wall with ntime-

rous pore rows per intersepr, highly porous pseudo¬

septa and abundant plate tabulae.

Description

Cup diameter above 30. Outer wall simple of

CambrocyatheUus-XYŸCy pierced by four to five

rows of rounded pores per intçrsept (pore diame¬

ter = 0 07, lintel width =0.1, wall thickness =

0,1). Intervallum of 1.5 în width, containing

pseudosepta pierced by seven to ten pore rows

(pseudo.septum thickness = 0.05, pore diameter

0.05), together with plate tabulae. Interseptal

coefficient = 1:2-1:2.5. Tabulae are dcveloped at

different levels ïn adjacent intersepts, a single

tabula may be ijontînuous tor two to rhrec inter¬

septs. Tabulae are tonvex, bearing five lo six pore

rows per intersepr (pore diameter = 0.05-0.07,

GEODIVERSITAS • 1998 • 20 (1)

Altay-Sayan archaeocyachids

Fig. 6. — Mikhnocyathus irregularis Kotel'nikov, n.sp.; A, holotype MNHN M81006, fragment of transverse section; B, paratype

MNHN M81007. fragment of transverse section. Nallvkinicyathus cyroflexus Zone. Kameshki Horizon, Atdabanian Stage: loca-

lity 218. Terektyg Khem River section. Central Tuva, Russia. Scale bar: 1 mm.

tabula thickness -= 0.05). Inner wall simple, with

three to four pore rows per intersept (wall thick¬

ness = 0.07, pore diamerer = 0.05). Vesicles are

abundanc in thc intervallum.

Discussion

This new species diflers from Aîikhnocyathus

zolaensis Maslov by more numerous pore rows

per intersept of the outer wall (four to fivc

against three) and of pseudosepta (seven to ten

against five to six), as well as by more abundant

tabulae, few being présent since juvénile stages

(cup diameter - 3.5'9,5 with interseptal coeffi¬

cient = l:l-l:1.5 only).

Superfamily SakHACYATHOIDFA

DebrenneefA. Zhuravlev, 1990

Family SakHACYATHIDAE

DebrenneefA. Zhuravlev, 1990

Genus Sakhacyathm

Debrenne et h, Zhuravlev, 1990

Sakhacyathus karpinskii

Osadehaya et Kotel’nikov, n.sp.

(Fig. 7)

Holotype. — MNHN M81008; Central Tuva,

Bayan Ko! River, locaÜty .30-1; Lower Cambrian,

Atdabanian Stage, Bazaiklia Superhorizon.

Etymology. — The species is named after the

Acadcmician A. P Karpinskiy.

OtheR MATERIAL. — Five well-preserved specimens

in transverse and oblique sections, MNHN M81007,

paratopes in CNlGRm.

DiSTRinuTION. — Bazaikha Superhorizon,

Atdabanian Stage; Central Tuva.

DtAGNOSlS. — Species with two outer wall pore rows

per intersept.

Description

Modidar sheet-like and solitary skeletons with a

transverse foldlng. Outer wall pustular bearing

two rows of pores per intersept (wall thickness =

0.15-0.20, pore diameter = 0.15-0.2, lintel

width = 0.08-0.1). Intervallum (1-1.5 in width)

concains non porous to sparsely porous pseudo¬

septa (pore diameter = 0.1-0.12, thickness =

0.07-0.1). Interseptal coefficient = 1:3. Inner

wall simple, 0.1-0.2 ihick, with two to three pore

rows per intersept (pore diameter * 0.1-0.2).

Vesicles are présent in the intervallum and a

GEODIVERSITAS • 1998 • 20(1)

Osadchaya D. V. & Kotel nikov D. V.

Fig. 7. — Sakhacyathus karpinskii Osadchaya et Kotel’nlkov, n.sp., holotype MNHN M81008, oblique transverse section of a modu-

lar form. Nochoroicyathus mahinskiiZone, Bazaikha Superhorizon. Atdabanian Stage: locality 30*1. Bayan Ko! River section, Central

Tuva, Russia. Scale bar: 1 mm.

secondary diickening is devclopcd in the lower

part of the cup.

Discussion

This new species differs from Sakhacyathus sub-

artus (Zhuravieva) by more numerous pore rows

per intersept of the outer wall (two against one).

Suborder ARCHAEOCYATHINA

Okulitch, 1935

Superfamily DicitocyaTHOIDEA

Taylor, 1910

Family Dictyocyathidak

Taylor, 1910

Genus Dictyocyathus Bornemann, 1891

Dictyocyathus conferttis Fonin, 1982

in WoTomn et al. 1982

(Fig. 8)

Dictyocyathus confertus Fonin - Voronin et al. 1982:

94, pl. XXIII, fig. 7.

Hoiotypi:. — FIN 3302/360: Vomnin et al. 1982,

pl. XXIII, fig. 7; We.stcrn Mongolia, Khusagt

Khaiikhan Ridgc, Salaany Go! River, locality

N-226/290: Lower Canibrian, Atdabanian 8tage,

Alataucyathus jacoH'hf^-l^tchi-, fabulacyathelhts bidz-

huensis-^ Pretiosoc^athui iuhtUis beds.

Othbr MATERIAL. — Ten well-preserved specimens

(CNIGRm).

Distribution. — Bazaikha Superhorizon-Kameshki

Horizon, Atdabanian Stage; Batenevskiy Ridge, Tuva,

Eastern Sayan, Mongolia.

GEODIVERSITAS - 1998 • 20(1)

Altay-Sayan archaeocyathids

Fig. 8. — Dictyocyathus confertus Fonin. specimen CNIGRm

12917/5, oblique transverse section, Noctioroicyathus mariinskii

Zone, Bazaikha Superhorizon. Atdabanian Stage: locality

5101-3. Sukhie Solontsy Valley, Batenevskiy Ridge, Russia.

Scale bar: 1 mm.

Description

Cup diameter is up ro 7-3- Outer wall basic

simple, 0.7 iliick., bearing a single pore row per

intersept (porc diameccr = 0.12-0.15, lintel

widrh = 0.5). Inccrvallum of 1.5 in width is filled

with a dictyonal nei^vork. Tacniae are pierced by

nine to ten vertical pure ix>ws (porc diameter =

0.15-0.25, tliitkiiess ol intcrvalarelemeiii.s includ-

ing taeniae and .synapticulae = 0.03). Synaptl-

culae arc rcgularly airanged in six co ciglit rows

per intervalluivi width. Radial cûcfficiciu =■

12-13, interscptal coeffitient = 1:10. limer wall

simple, wiili oiie pure row per incersepi (porc

diameter = 0.1).

Discussion

Dictyocyathus confertus differs from other species

of Dictyocyathus by thinner skeictal éléments and

by higher radial coefficient.

Superfamily ArcKAKOCVATHOIDEA

Hinde, 1889

Family ARCHAEOCYA THinAK Hinde, 1889

Genus Archaeopharetra

R. Bedford et W. R. Bedford, 1936

Archaeopharetra inarginata

(Fonin, 1982) in Voronm et al. 1982

(Fig. 9)

Satanviyathus niarginatus Fonin - Vomnin et al. 1982:

96, pj.XXlVJlgs 1-3.

Saianycyatfms ordinatus Fonin - Voronin et al. 1982:

97, plXXV,f!gsl-3.

Salan\>c\dthus dhertus Fonin — Voronin et al. 1982;

98, pl.XXV,fig.4.

Flindcrsicyathus ^aàlis Fonin — Voronin et al. 1982:

109, pl.XXXUl.Fgs 1-5.

Archaeopharetra marginata (Fonin) — Dcbrcnnc

A. Zhuravlcv 1992: 121.

Holoitpi - — I>IN 3302/368; Voronin étal. 1982,

pl. XXIV. (‘ig. 1 î, Western Mongolia, Khasagt

Khairkhan Ridge, .Salaany Gol River, locality

N-224/ÎS0; Lower Cnmbrian, Atdabanian Stage,

Alataucyatlms jaroschevitschi-, Tahulacyathelliis hiaz-

haemh-., Pntwsocyuilms suhtilis bed.s.

O'lUHR MATI-RIAI.. — Tcn wcll-prcserved .spccimens

CNRIRni.

DlSIRlBUTKW. — Bazaikha Superhorizon-Kanieshki

Horizon, Atdabanian Stage; Kuznctskiy Alatau,

Batenevskiy Ridge, Central Tuva, Eastern Sayan,

Mongolia.

Description

Cup diameter is up to 18. Outer wall centripetal,

0.1 thick. Two to rhrec pore rows cross an inter-

valar ccll whicb is formed by pseudotaenial lin-

tels and synapticulae (pore diameter = 0.05-0.1).

InCervjIlum of 1.5-2 in width bearing occasional

segmented rabniae and porous pscudolaeniae

picreed by Hve to six vertical pore rv>ws (porc dia¬

meter = 0,25-0.35). Synapticulae are irregularly

spaced in rwo to ihmc éléments per intcrvallum

width. Radial coefficient = 12-14, interscptal

coefficient 1:4-1:5. Inncr wall simple, with one

pore row per intersept and spines (thickness =

0.05, porc diameter = 0.1-0.2).

Discussion

Archaeopharetra marginata differs from A. itma-

GEODIVERSITAS • 1998 • 20(1)

Osadchaya D. V. &c Kotel’nikov D. V.

Fig. 9. — Archaeopharetra marginata (Fonin). A. specimen CNlGRm 12917/8, oblique longitudinal section, locality 9075-1. Vadi Bala

section, Central Tuva; B. specimen CNlGRm 12917/9, transverse section, locality 5101-8, Sukhie Solontsy Valley, Balenevskiy

RIdge; Nochoroicyathus maninskiiZona, Bazalkha Superhorizon, Atdabanian Stage: Russia. Scale bars: 1 mm.

tiensis (Belyaeva) by hlghcr radial coefRcicnr Ûlcl^^ocyathus lepidm^i^x\\n-Votom\\etaL 1982:93,

(twelve to fourteen against six to seven) and by pl^ XXllI, figs 5-6. . v ,

Il J. c ^ il r Chouherttcyathus lepiaus (Fonin) - Debrenne & A.

smaller diameter or oufer wall pores; rrom zimravlev 1992- 123

A. yarbili (Rodionova) by thicker outer wall (0.1 Diayofauus sp. '- Debrenne & A. Zhuravlev 1992,

against 0.04) and by less numerous synapticulae. pl. )Ô(V111. fîg. 2.

Suborder DICTYOFAVINA Debrenne, 1991

Superfamily DtcrTYOFAVOIDEA

Debrenne et A. Zliuravlev, 1992

Family DïClTOFAVlDAH

Debrenne et K. Zhuravlev, 1992

Genus Gravestock, 1984

Dictyofavus lepidus (Fonin, 1982)

inVoTonm et ai 1982

(Fig. 10)

Holotype. — PIN 3.302/358; Voronin et al 1982,

pl. XXIIR fia. 5; Western Mongolîa, Kha.eagt

Khairkhan RjJgc, Salaany Gol River, locality F-13;

Lower Cambrian, Atdabanian Stage, Alataucyathus

jaroschevitschi-, Tabidacyathellus biclzhaensis~y

Pretiosovyathus suhtilis bcd.s.

OtHER M/\TER1AL. — Twenty well-preserved spéci¬

mens (CNlGRm).

Distribution. — Bazaikha Superhorizon-Kameshki

Horizon, Atdabanian Stage; Batenevskiy Ridge,

Central Tuva, Eastern Sayan, Mongolia.

GEODIVERSITAS • 1998 • 20(1)

Altay-Sayan archaeocyathids

Description

Modular massive or solitary skeletons. Diameter

of solitary cups is 5-7- Ourer and inner walls

rudimentary, wirh cell openings 0.4-0.7 in size.

Intervallum is filled wich calicles hexagonal in a

cross section. Calicles are front 0.5-0.6 ro 2.0-2.5

wide, thickne.ss of a calicle wall is 0. l. Calicles

bear one, rarely rwo pore rows per facet. Pore

diameter 0.2-0.5. One of che osculi is commonly

larger than che orherSi Vesicles are abundam

through the whole cup.

Discussion

Dictyofaviis lepidus differs from other species of

Dictyofaviis by rhicker intervalar éléments and by

différenciation of osculi.

Dictyofaviis obtustis GxdiWcstocV.y 1984

(Fig. 11)

Dictyofavus obtusus Gravestock, 1984: 98,

fig. 50C-F. - Zhuravlev & Gravestock 1994: 46,

fig. 6B (synonymy).

Hoi.OTNTE. — SAM P21666 is not figured. A para-

type SAM P2I668 which is figured by Gravestock,

1984, fig. 50D-E, Ls chosen here as the Icctotype;

South Austr.alia, Wilkawillina Gorge, section I; Lower

Cambrian, Atdabaiiian Stage. Warrwotacynthus wilka-

willmensis Zone.

Other MATERIM.. -- Fifrcen wcll-prescrved spéci¬

mens (CNIGRm).

Distribution. — Bazaikha .Superhorizon,

Atdabanian Stage; Batenevskiy Kidge, C.entral Tuva,

Easrern Sayan and Warrwotacyathus wilkawillinensis-

Spirillicyathtis tenuis Zones, Atdabanian Stage,

ÎTinders Ranges and Yorke Penin.sula, -South Ausrralia.

Description

Modular massive skeletons commonly consisting

of rwo to three individuals. Diameter of a single

individuai is up co 6, 2-3.5 apart from each

other. Outer and inner walls are rudimentary,

with cell openings 0.2-0.3 in size. Intervallum is

filled with calicles hexagonal in a cross section.

Calicles are up to 0,3-0.4 wide, thickness of a

calicle wall ls 0.05. Calicles bear one pore row

per facet (pore diameter = 0.01-0.04). In places,

pellis is présent on the outer wall.

Fig. 10. — Dictyofavus fepidus {Fonin), A, specimen CNIGRm

12917/10, tfans-vefse section, tocsIIJy 369, Bazaikha River sêt>

lion. Easlern Sayan; B. specImen 12917/11. fragment of IrarTS^

verse section. Central Tuva; C. specimeri CNIGRm 12917/12.

fragment of transverse section; locallly 5101-7. Suktrie Solontsy

Valley, Baienevskiy Ridge: Nochoroicyathus mahinskii Zone,

Bazaikha Superhorizon, Atdabanian Stage; Russia. Scale bars:

2 mm.

GEODIVERSITAS • 1998 • 20(1)

Osadchaya D. V. &c Kotel’nikov D. V.

Fig. 11, — Diclyofavus oblusus Gravestock. A. specimen

CNIGRm 12917/6. transverse section, locality 1808. Bazaikha

River section, Easiern Sayan; B specimen CNIGRm 12917/7,

oblique longitudinal section, locality 5100-9. Sukhie Solontsy

Valley. Batenevskiy Ridge; Nochoroicyathus mariinskii Zone,

Bazaikha Superhorizon. Atdabanian Stage; Russia. Scale bars:

1 mm.

Discussion

Dictyofavîis obtusiis differs from D. lepidns

(Fonin) by smaller calicles and by thiniier calicu-

lar Wall.

Acknowledgements

We are gratcFul to Françoise Dcbrenne (Muséum

national d'Hisroirc naturelle, Paris) for her great

hclp wirh translation and publication of rhis

paper and to André}' Zhuraviev (Palaeontological

Institure, Moscow) and Rachel Wood (Earth

Sciences, Univensity of C'ambridge) wbose com-

ments as referees significanlly improve our paper.

This paper is a contribution to ibe IGCP

Project 366 “Ecrdogical Aspeers of Cambrian

Radiation".

REFERENCES

Dcbrcnnc F. Zhuniviev A. Yu. 1992. — Irrcgular

Arcliaeocyarlis: Morpliology, Clnrogcny, Sysce-

matics, Bîostratigr.iphy, Pulaeoecol(.>gy. i't-ihiers de

CNRS Editions, Patîs, 2!2 p.

C.ravcstocl< D. 1. 1984. — Archaeoq'atha from lower

parts of che Lower Cambrian carbonate séquence in

Soiitb Ausfralia. Assartiitiafi of Austmlasian

PaUfontologists Mémoire 2: 1-1 39.

V^oronin Yu. I. 1979. — Ayaisiisiacidy SSSR

(Ajacicyatids of the U.S.S.R.). Trudy Paleontolo-

^icheshogo lustituta Akademii miuk SSSR 176: 1-148

[in Russian].

Vuronin Yu. L, V^oroiiuva 1 . G., Grigor'eva N. V.,

Dro/dova N A., Zlicgallo F.* A., Z.huravlev A. Yu.,

Ragt>/ina A. I.., Rozanov A. Yu., Savutina 1. A.,

Sysoev V. A. & I onin V. L). 1982. — Granirsa

dokembriy.» i kcmbriv.i v gc^'^inkIina^^ykb oblas-

tyakh (oporn>q' tazre/ Salany-Col. MNR) [The

Precambrian/CJambriaii buundary in rlie géosyncli¬

nal areas (Salany-Cîol rLlêieiicc section, MRP)].

Tnid\< Sovmeunoy Sovetsko-MongoPskoy poleontologi-

cbeskoy i'kspeditPti 18: 1-152 |in Riissianf

/ado)x»-tbnaya N. M. 198.3. — l orgashinsldy rifos^yy

kompleks (nizhniv kembriy, Vosiochnyy Sayan).

[lorgashino rectal cornplt'x (l ower Cambrian,

l’.astcrn Sayan)], />/ Betekhrinn L>. A. & Zburavleva

J. I . (cds). Sreda I /bizn'v geologiche.skom prosh-

lom (Palcobiogeografiya I paleockologiya)

[Environment and lilc in the geological past

(Palaeobiogeography and palaeoecolugy)]. Trudy

hntituto geohgii I geofiziki Sihinkogn oïdeleniya

Akfuiemit Hiiuk SSSR, N.mka, Novosibirsk 569:

138-151.

Zhur.iviev A. Yu. & Gravestock F). I. 1994. —

Archaeoeyacha from Yorke Peninsula, South

Ausrralia and aichaeucyathan Early Cambrian

zonation. Ait'heringü 18 (1): 1-54.

Submitted for publication on 7 April 1997;

accepted on 9 September 1997.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes ferox, a borhyaenoid (Metatheria,

Mammalia) from the early Palaeocene of Bolivia.

Phylogenetic and palaeobiologic implications

Christian de MUIZON

LIRA 12 du CNRS, Laboratoire de Paléontologie, Muséum national d'Histoire naturelle,

8 rue de Buffon, F-75231 Paris cedex 05 (France)

Muizon C. de 1998. — Mayulestes ferox. a borhyaenoid (Metatheria, Mammalia) from the

early Palaeocene of Bolivia. Phylogenetic and palaeobiologic implications. Geodiversitas

20(1) : 19-142.

ABSTRACT

Mayulestes ferox is a borhyaenoid marsupial from the early Palaeocene of

Tiupampa (Bolivia). The holocype and only known spccimen is a partial ske-

leton which is described and discussed below. Mayulestes ferox is a member of

the Family Mayulestidae- a taxon which also includes the species Allqokirus

australis from the same locality and age> but which is only known by a few

isolated molars. Mayulestes ànà Alleiokirus are the nvo oldesr known borhyae-

noids. Mayulestes diffcrs from Allqokirus in the morphology and proportions

of its molars. A major feature of the molars of both généra is the réduction of

the entüconid which is regarded here as a synapomorphy of the Mayulesti-

dae. Mayulestes lias the plesiomorphic marsupial dental formula (15/i4;

Cl/cl; P3/p3; M4/m4) and its molar morphology approaches the plesio¬

morphic marsupial cheek tooth pattern. Mayulestes ferox does not hâve a

tympanic process of the alisphenoid, a structure whose presence is gcnerally

regarded as a marsupial synapomorphy. Comparison wirh oiher borhyaenoid

taxa indicates that the lack of tympanic process of the alisphenoid is in fact a

plesiomorphic character State for rhe .superfamily, and it is su^ested that this

feature appeared scveral rimes during marsupial évolution. The ear région of

Mayulestes beats a conspicuous médial process of chc squamosal and there is a

shallow cavity (the roof of the alisphenoid sinus) between the foramen ovale

and the glenoid cavity. excavated within the squamosal anterîoriy, the perio-

rie po.steriorly, and rhe alisphenoid berween. The contribution of the squa-

mosal to the roof of the alisphenoid sinus is regarded as the key

synapomorphy of the bothyaenoids. Other borhyaenoid synapomorphies

are: the loss of the prootic canal, ihc réduction and ihe loss of ihe antéro¬

latéral process of the mxxilla, and the probable loss of cpipubic boncs. The

postcranial skelcton of Mayulestes is represented by rwcnty complété or par¬

tial vertebrae> a few fibs and raost major limb bones. A comparison with

living didelphids. Pucadelpbys, orher borhyaenoids, and sevcral arboreal (or

probably arboreal) mammals such as sciurids» tupaiids» procyonids. multi-

tuberculatcs morganucodontids, triconodontids, and Henkelotherium rcveals

that many features of rhe postcranial skeletpn gf Mayulestes are indicative of

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

KEYWORDS

Marsupialia,

Borhyaenoidea,

Palaeocene,

Bolivia,

phylogeny,

functional anatomy.

arboreality. Thc^e traits arc: probable prclicnsiliry ot'rhc tail; postcrodorsally

extended posrerodorsnl angle ofthe scapula; antcriorly and distally projecred

acromion; low tubciclcs of the hunicnis; circulât shape of rbe head of the

humcriis; large si/c of the epicondyloid ridge and distomcdially protruding

médial épicondyle ot the humérus; deep flexor fossa on rhe media! side of

the olecranon of the ulna; morpholog)' of the MeV; gieai mobiÜty ofvhc hip

attested by the shaJlownc.ss of the acctabuliim and the srrong development of

the fémoral irochaïucrs; sigmoid shape of ihe tibia and morphology of its

distal aniculation; shape and orientation of the ccial facct ofthe calcanéum;

large size of the peroneal process; transversely compresscd tuber calcanei.

SeveraJ other features (size of the neural spine and transverse process of the

lumbar vertebrae; morphology of the zygapophyscs of the last thoracics and

lumbar vertebrae; long, arueriorly benr olecranon of the ulna; éversion of the

iliac wing; relative deprh of the fémoral rrochlea; flattened distal epiphysis of

the tibia; grcat length of the tuber calcanei) indicate rhat Mayulesies was a

relativcly agile, scansorial animal capable ofbounding. Mayulesies i.s regarded

as a partially arborcal predaceous mammal capable ofbounding and of some

relatively fast but short runs. Mayulestes was certainly fairly agile and could

bave had an ecological niche close co that of wcasels or martens, although

more arboreal tlun rhe former. Several arborcal features oi Mayulestes arc aiso

found in PucadelphySy a didciphid marsupial from ihe same locality.

Consequenrly, rhis genus is also regarded as partially arboreal, although to a

lesser extern than Mayulestes. The fact that the rwo oldest skeletons of

American marsupials dénote arboreal habits reinforces the hypothesis that

arboreality is probably a symplesiomorphy within marsupials.

RÉSUMÉ

Mayulestes ferox, un Borhyaenoidea (Metatheria^ Mammalia) du Paléocène

inférieur de Bolivie. Implications phylogénétiques et paléobiologiques. Mayulestes

ferox est un marsupial Borhyaenoidea du Paltfocènc inférieur de Tiu|)ampa

(Bolivie). L’holotype et unique spécimen connu est un squelette partiel com¬

prenant le crâne complet, la mandibule incomplète et la plupart des os des

membres, lesquels .sont décrits et discutés ci-dessous. Alayulcstes ferox est un

représentant de la famille des Mayulestidae, un taxon qui inclut également

l’espccc Allqokmis australis, provenant de la même localité et du même âge,

mais connue uniquement par quelques molaires isolées. Mayulestes et

Allqokmis .sont les deux plus anciens Borhyaenoidea connus. Mayulestes dif¬

fère ésÀllqokirus par la morphologie et les proportions de ses molaires. Un

caractère important des molaires des deux genres est la réduction de renioco-

nide qui est considérée ici comme une synapomorphic des Mayulestidae.

Mayulestes possède la formule dentaire plésiomorphe pour les marsupiaux

(I5/i4 ; C/c ; P3/p3 ; M4/m4) et la morphologie de ses molaires esc proche

du patron plésiomorphe des dents jugales de marsupiaux. Mayulestes ferox ne

présente pas de processus tympanique de Palisphénoïde, une structure dont

la présence est généralement considérée comme une synapomorphie de mar¬

supiaux. Des comparaisons avec les autres taxons de borhyénoïdcs indiquent

que l’absence de processus tympanique de lalisphénoïde est en fait une plé-

siomorphie pour la superfamille, et il est émis l'hypachèsc que cc caractère est

apparu plusieurs fois au cours de l'évolution des marsupiaux. Le squamosal

de Mayulestes présente un processus médial bien développé. Dans la région

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

MOTS CLÉS

Marsupial ia,

Borhyaenoidca,

Paléocène,

Bolivie,

phylogénie,

anatomie fonctionnelle.

auditive, entre le foramen ovale et la cavité glénoïde du squamosal, on obser¬

ve une cavité peu profonde (le toit du sinus alisphénoïde), creusée dans le

processus médial du squamosal antérieurement, dans le périotique postérieu¬

rement et dans ralisphénoïde entre les deux. La participation du squamosal à

la constitution du toit du sinus alisphénoïde est considérée ici comme la

principale .synapomorphie des Borhyaenoidca. Les autres synapomorphics de

la superfamille sont la perte du canal prootique, la réduction et la perte du

processus antérolatéral du maxillaire et la perte probable des os epipubiens.

Le squelette post-crânien de Mayulestes esc connu par une vingtaine de ver¬

tèbres complètes ou partielles, quelques côtes, et par la plupart des princi¬

paux os des membres. Une comparaison avec les didelphidés actuels,

PucaMphys, les autres Borhyaenoidca et plusieurs mammifères arboricoles

(ou supposés arboricoles) tels que les sciuridés, les tupaidés, les procyonidés,

certains multituberculés, morganucodontes, triconodontes et Henkelo^

therium révèle que beaucoup de caractères du squelette post-crânien de

Mayulestes indiquent un mode de vie arboricole (queue probablement pré¬

hensile ; angle postéro-dorsal de la scapula étiré po.s'téro-dorsaIcmcnt ; acro-

mion projeté antérieurement et distalement ; tubercules de Thumérus

relativement bas ; forme circulaire de la tête de l'humérus ; grande taille de la

crête épiconylienne et projection disto-mcdiale de Tépicondyle ; profonde

fosse des fléchisseurs sur la face médiale de l'olécrâne de Tulna ; morphologie

du MeV ; grande mobilité de la hanche attestée par la faible profondeur de

Tacétabulum et le développement des trochanters fémoraux ; forme sigmoïde

du tibia et morphologie de son articulation distale ; forme et orientation de

la facette ectale du calcanéum ; grande taille du processus péronéen et tuber

calcanei comprimé transversalement). Plusieurs autres caractères (taille de

l’épine neurale et des apophyses transverses des vertèbres lombaires ; mor¬

phologie des zygapophyses des dernières vertèbres dorsales et des lombaires ;

olécrane de l’ulna, long et recourbé antérieurement ; éversion de l’aile de

l’ilium ; profondeur relative de la trochléc fémorale ; épiphyse distale du tibia

aplatie antéro-postérieurement et grande longueur du tuber calcanei) indi¬

quent que Mayulestes était un animal relativement agile, capable d’adopter

une démarche rapide et de bondir. Mayulestes est interprété comme un pré¬

dateur partiellement arboricole capable de bonds et de course rapide mais de

courte durée. Mayulestes était certainement assez agile et a pu avoir une niche

écologique voisine de celle des martres et des belettes actuelles. Plusieurs

caractères arboricoles de Mayulestes sont aussi présents chez Pucadelpljys^ un

marsupial didciphidé provenant du même gisement. En conséquence, cene

forme est également interprétée comme étant panicllement arboricole, bien

qu’à un degré moins poussé que chez Mayulestes. Le fait que les deux sque¬

lettes les plus anciens de marsupiaux américains possèdent des caractères liés

à l’arboricolie renforce l’hypothèse selon laquelle ce mode de vie est proba¬

blement une symplésiomorphie chez les marsupiaux.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

INTRODUCTION

Borhyacnoids are highly carnivorous South

American marsupials which are known from the

early Palaeocene to ihe late Pliocène. The remains

are generally isolated teech and jaws but,

although uncommon, .some complété or partial

skeletons are known. The majoi collection is that

described by Sinclair (1906) which consists of

several partial skulls and skeletons obtained from

the Santa Cruz beds (midclle Miocene) of Para-

gonia. The specimens are plated in four taxa:

Borhyaena tuberatUy ProthyLteynus patagonicm,

Cladosîctis patay^ornms, and Sipalotyou graciUs,

Several other skulls ol Proîhylacynus and

Cladosictis from the Santa Cru?- beds are known

and hâve heen figured (although nor described)

by Marshall (1979a, 19S1). In Columbia, the

“Monkey unit" (late Miocene) of the Honda

Group from the upper Magdalena Basin has yiel-

ded a complété .skeleton ol Lycopsis longirostris

(Marshall 1977a). In Cacamarca Province

(Departamento de Belen). the Monthermosean

beds of the Corral Quemado Pormation hâve

produccd îwo partial skeletons uf Thylacosfrtilus

atrox (Riggs 1934). No skeletons are kfuwn from

the Palaeogene and the only skulls are those from

the Deseadan (late Üligücene-early Miocène)

beds of Salla*Luriba}' (Bolivia) referred lo

Sallacyo}i hoffitetteri and PamborhyutfU't holiviuna

(Pettei & IdoffstetTer 1983) and an undescribed

basicraoium retèrred co Notagak uader study by

the author. Borhyaenoids from the Early Tertiary

are rate and no complété skulls and/ot partial

skelerons bave been found in ihc Palaeogene.

This State of things drastically enhances the

importance of the discovety of a partial skeleton

of borhyaenoid {Mtjyulcstcs frrox) from the early

Palaeocene of the Santa Luda Pormanon at

Tiupampa (Muizon 1994), which is one ot the

two oldesi known borhyacnoids and probably

one of che most complète spécimens known in

the wholc superfamily. The other oldest borhyae-

noid is Allqok 'mis australis from the samc locality

and âge as Mayuksm^ but knosvn by a tdv teerh

only. There are more than 4ü Ma between

Mayukstes ferox and the Santa Cruz borhyacnoids

skeletons, more than berween rhe Santacruzian

and the last borhyacnoids (approx. 15 Ma).

In a more général context, skull and skeletons of

fossil marsupials in the Earlv Tertiary are extre-

mcly rare and rhe only other marsupial skeleton

known from rhe Palaeocene is that of

Pucadelphys iuidinus from the sa me locality

(Marshall & Muizon 1995; Marshall &

Sigogneau-Russell 1995). Auatherium resbetovi

from the Late Cretaceous of Mongolia

(Trofimov Szalay 1994; Szulay tk IToRmov

1996) is rhe only Mesozoic mennherian known

hy a lairly complète skeleton. I hLs form is regar-

ded hene as a crue mclathcriun, although its very

spccialiscd ccerh may Icad sonie aurhors to ques¬

tion its membership in Metatheria. These lacts,

rhcrelore. raise the importance uf Mayukstes ai

lire Icvcl of Mctatheria and, cynsidering the scar-

clty of early Palaeocene and Lace Cretaceous

mammals, the discovety of the skeleton of

Mayukstes ferox represents a major event in the

knowledge ol early nvammals.

rhe description and interprétation of the skull

and rhe postcranial skeleton of Mayukstes ferox

are presented below. As is generally the càSc, it is

rhe skull that provides most of the information

un the phylogony of borhyacnoids and marsu-

pials while the postcranial skeleton is more Tofor-

mative on the locomotion and habitat ol

Aïayukstes, although the latter also beats some

intercsnng data on mammal évolution.

Abbreviatlons

AiMNH

DGM

FMNH

MHNC

MNHN

MIT

YPFB Pal

YPMPU

USNM

American Muséum of Nacural Flistory,

New York, US.\;

Divisâo de Geologia e Mineralogia do

Depart.imcrtto Naclonal de Produçâo

Minerai, Rio de Janeiro, Brazil;

Ficld Muséum of Natural History,

Chicago. USA;

Museo de Flistoria Natural de

Cochahamba, Cochabamba, Bolivia;

Muséum national d'Histoirc n,uurelle,

Paris, ITancc:

Museo de La Plaça. L.i Plata, Argeniina;

Paleonrolog)' collections of Yacimientos

Petroiifcros fiscales de Bolivia, Santa

Cruz, Bolivia;

Yale Peabody Muséum, New Haven (ex-

collcctions of the Princeton University),

USA;

United States National Muséum of

Natural Hisrory, Washington D. C.,

USA.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

SYSTEMATIC PALHONTOLOGY

Légion F’RIBOSPHENIDA McKenna, 1975

Infraclass ME TAPHEILEA Huxley, 1880

Suborder DIDELPHIMORPHIA (Gill, 1872)

Infraorder SPARASSODONTA

Ameghino, 1894

Superfamily B0RHYA1:N0U‘)Fj\ Ameghino, 1894

DiaijNom-s, — Camivuroiii niarsupuds wiih a middle

car epitympanic sinus locarcd anrerolateral to the pro'

montoriiim and formed by the altsphenoid anteriorly

(and sonictimcs vcntrally), dic pciro^al posteriori)',

and ihc squamosal laicrally. The Rorhyaenoidea art

chc oiily knttwn nietatlnrlans svherc the squamosal

participâtes iii the eomposition of the alisphen<*id

sinus and thb (vauire ix'prcseiUs the key synapunior-

phv ol tho Borhyaenoidca- ritrtherniofc, ihc superfa¬

mily is ;dso diugnosed by the réduction and ihe loss of

the anierolaivral protess of the niaN-illu which fonu

the huerai Wall of the fossy for rhe lower canine on the

skull, rhe loss of the prooric canal (a character siate

found in some other lineages of marsupials), and, pos-

sibly, the loss of ihe epipubic bones (only probable in

four gênera, MaytlLatcs. Clacloiicth, Prothylucynus and

Lycopiis, a Icaiure unknown in oihcr taxa).

Nota. — Borhyaenoids also présent the h)'percarni-

vorous dental l'unctii.tnjl complex rtlîiTcd lo the posi-

vallum-prevallid .vhear (reilucrion and loss ol nieta-

conid and entoconid. increase in si/e of paraconid,

réduction ofialonid. réduction of protocone, paraco-

ne and Mvlar shclf. increase iti si/.e ot metaeonc and

inctastylar crest). HovvevcT, as shown by Archer

('1982), Fox (19'>5) and Mui/on & l.angtvlVadrc

(1997), ihis functional complex Isa highly liomoplas-

tic feature which appears indepcndently in scvcral

group of mamnials and wilmin thèse gruups.

I hcrefore, if It is truc that ihis synapomotohy dia¬

gnoses the Borhyaenoidca. it bas a low phyiogenetic

value and cerrainly cannot represent a key synapo-

morphy of the Borhyaenoidca.

Family Mayulhstidae Muizon, 1994

DlAGNOSls. — Borhyaenoidca diagno.sed by an

important réduction ol the entoconid and che resul-

ting lingual opening of the rolonid basin.

Typk cil'.NUS. — Mayulestes Muizon, 1994.

INCLUDI-.D GENERA. — Mayulestes Muizon, 1994;

AllqokirusWài'sh^W cr Muizon, 1988.

Genus Mayulestes Muizon, 1994

Type SPECII-S. — Mayulestes ferox Muizon, 1994.

DiagnoSIS. — The samc as the only species referred

to Mayulestes.

Maytdestes ferox Muizon, 1994

Diagnosis. — Mayulestes ferox difîers from Allqokims

australîs in the snialler L/W ratio ot its upper molars>

in dic groaier depth of ihe cctnflexiis. in liic larger sty-

lar cusp D wiih .i low lingual cicsr which riins

towaids ihe lingual extremiry of che metacrisra, in rhe

meracrista Avhick does not overhang rhe base of the

Crown po.steriorlvi in ihesiraigln (>osterior edgt of ihe

upper molar.s, in the more robust and longer proto¬

cone, and in its narrower and more slender lower

molars.

Ntvi A. — Since die only oihcr spccies of the family

(Àütjokh'us australis) is known by upper and lower

rnolurs only, ihe diagnosis ol Mayulestes ferox refers lo

rnolar morphology i»nly.

Hoioiype. - MH NC 1249, a partial skcleton wirh:

almost complété skull; right hemim.tndible lacldng

the vertical ramus. p2 and p3; lefi hemimandible lac-

king the vertical ramus, portion ol ihc horizontal

ramus anterior to the posterior root of p2, and the tri

gonid of in.5; atlas laeking the ventral arch; axis; cen-

iruni of the ?fhird tcrvioil vcrtchra; ^fifrh cervical

vertebra laeking rhe neural arch; centra of rss'o aiue-

rior thoracic verrebrae; one complété anterior thoracic

venebra; rwo lasr thoracic vertebrae: five first lumbar

verrebrae wiih L4 laeking most ol the neural arch;

cwü anterior caudal vertebrae, rwo posienot caudal

vertebrae; rwo anterior ribs; rwo médian ribs; nghi

Ncapula, almost complète; borh humeri; complète lelt

uina; proximal exrremicy of the righr uina; lc(t radius;

right unciform; lelt Mclll and V; left innominatc lac*

king the jnterovcJitral part ol dit pubb; part ol the

right innnminarc- (acerabtihim and posterior portion

ot the ilium): lefr fémur laeking pan of the diaphysis

and rhe nicdial distal corid)dc; nght fémur laeking

pan of the grcater irochaïucr and pari of ihc diaphy-

sis; right tibia; distal extremity ol the righr fibula;

right calcanéum; distal extremity of rhe left talca-

neuni; Icft Mtlll and IV.

H\Tt>D|GM. — T)'pc spccinicn only.

LgcAUTY, horizon AKI) Atil'. — d he specimen was

collccrtd on the site 1 (“ihe quarry"), at rhe locaJity of

Tiupampa, situated about 95 km sourheasi of

Cocliabamba (Mizquc Province, Department of

Cochabamba, Bolivia). Site 1 (sec Gayet et al. 1992;

Muizon & Marshall 1992: Marshall et al. 1995 for

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

locality map) Tiupampa has yieldcd an abundant

vertebratc fauna of rhe Sanra Lacia Formation.

The vertebratc fauna includcs; Oitcichthycs (Dipnoi,

Telesotei), Amphibia (Anura, Gymnophiona),

Reptilia fChelonia, Sauamata (Lacertilia and

Ophidia), Crocodilia], and Mammalia.

An updaitfd lîst of rhc mammal iauna is given bclow.

Supralamilial cFassificaiion of marMupials fo!Iow.s

Szalay (1994) and Sparas.sodonta are fcgarded as an

infraorder of rhe suborder Didelphimorphia. follo-

wing Muizon H tiL fl997). The ïuher Didelphi¬

morphia are incluaed in rhc new infraorder

Didelphodonta;

Class MAMMALIA

Infra-class MEd’A'l’HERIA

Order DIDELFHIDA

Suborder ARCH IM ETATHERIA

Family PEHADriCTIDAE

Pemeiectes cf. atatrinum

Suborder SUDAMERIDELPHIA

Family CaROLOAMECHINIIDAE

Rohtrthoffstcttma nationalgcographica

Suborder DfOELPHIMORPHlA

Infraorder DIDELPHODON'I A new

Family Pl'CADFLPHYDAE new

Pucadelphys andinus

? Andifwaelpim codmhambensis

Family ? DlUKlEHin.^E

încadelphyi anHijutis

Mizanedilphys pilpinensis

Tiufordia floresi

Family jASKHAnaPfiYDAE

Jaskhadelphys mitîutus

Infraorder SPARASSODONTA

Family MAytJl.ESrr>;\E

Mayukut% ferox

Aliqukinu itusmdh

Order GONDWANADELPHIA

Suborder MfCROBIOTHERlA

Family MiCROBÏOTI lERIlDAE

Khasin cordillierenm

Infra-class EUTFiERlA

Order LEPTICTIDA

Family PaiAKORYCTIDAE ?

Cf. CimoUsies

Family indet.

Gen. and sp. Indet.

Order PANTODONTA

Familv AlcIDEDORBIGNYIDAE

Alciaedorhigi'iyd inopinata

Order CONDYLARTHRA

Family MloriAENTD.AE

Molinodus suarezi

Thiclatnns minutia

Tiuvlaenus n.sp. 1

Tiuclaenus n.sp. 2

Pucanodus gagnieri

Family MlOCLAENiaÆ or DiDOLODONTIDAE

Mioclaenidae or Didolodontidae n.g., n.sp.

Andinodus boliviensis

Family ? Perjptychidae

aff. ? Mhnatuta

Order CON DYLARTH RA incertae sedis

Family KOI.I.EANIIDAE

KoUpania tïuparnpina

Order NOTOUNGULAIA

Family Hr.NRICOSBORNiiDAE or

OlDhihi IHOMAMID.AE

The âge of rhe fauna is now commonly accepeed to be

Palaeocenc (Lîayer et nL 1992; Muizon 1992;

Muizon 6c 1. Brito 1993; Marshall et al. 1995)

contrary co the Lute Creraceous âge Initially ossigned

to the Tiupampa mammals (Marshall et al. 1983;

Muizon et al. 1983; Marshall et al. 1985; Marshall &

Muizon 1988).

d hc position of the Tiupampa (site 1) mammal launa

within the Palacocene has been distusscd by Van

Valcu (1988), Muizon 6c I. Briio (1993), Bonaparte

et al. (1993), and Marshall et al. (1993) and assigned

an early Palacocene âge. According to rhese aurhors,

the Tiupampian (early Palacoccnc) local mammal

fauna (type locality; site 1 at Tiupampa) is older than

the Peligran local mammal fauna (type locality: Punta

Pcligro, Paiagonia, Argcntina) (Muizon & 1. Brito

1993; Bonaparte et al. 1993: 36, 37, Bonaparte &

Morales 1997). The relative age of ihe faunas is based

on the comparison ol rhe cvolutionary stage of ihe

condylarths they hâve yielded. Although thisapproach

i.s certainly quc.siionable, it is obvious chat the

tlircc condylardis of the Punta Pcligro fauna axe more

derivcd than ihc relativcly generalised forms of

Tiupampa (six gênera), Furtherinore, a new ungulate

from Puntu Pcligro (Bonaparte & Morales 1997)

shows a c|earlv prelophodont motpholoey which

recalls, although more primitive, Notonycîiops from

the Rio Loro Formation of norrhern Argentina (Soria

1989). The Punta Pcligro fauna has been found in the

Banco Negro Infcrior and its age i.s early Palacocene.

Ir probably spans from 63.2 to 61.8 Ma. as stated by

Pascual & Ortiz-Jaureguizar (1992: 564), who provid-

ed a feview ol rhc age of the Banco Negio Infcrior.

The Tiupampa local fauna is here regarded as older

than the Puma Pcligro fauna and, rhereforc. if the

dare of rhc base of me Banco Negro infcrior is cor¬

rect, ihe \ iupampa fauna i.s older than 63.2 Ma.

Therefore, acœrding to Haq ôd Van Eyslnga (1994),

it would bf pre-Torrejonian in age and contempora-

ncou.s with tne Pucrcan. The Punta Pcligro fauna is

probably contemporaneous with the Torrejonîan of

North America. I therefore agréé with the gross équi¬

valence, proposed by Bonaparte et ai (1993)> of the

Tiupampian with the Puercan and the Peligran with

the Torrejonian.

Recently, Marshall et al. (1997) and Semperc et al.

(1997) hâve correlated the Tiupampa beds to the

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

magnetic anomal)' chron 26r whîch would corrclate

the Tiupampian land rnammal agc with the early

'I iffanian North American land maninial âge {ue,

carly late Palaeocene). However, ihc panrodonts and

condylarths (groupe ihat can be compared in both

.^ubconiinertt.';) of rhe 'riffiinian faunas of

NortJi America are much more derived and more

diverse chan tliQse of d iupampa and clearly seeni to

represent a younger fauna. Furrhermore, since rhe

eniire sériés of ihe Santa Liicia Formation ar

Tiupampa corresponds to a (possible?) single reversed

chron, the corrélation of the Tiupampa beds with rhe

orher Maa.siiichtian-Palaeocene serie.s nf Bolivia (La

Palca, sce Marshall et al. 1997) has been based on

lithological hicies and scdimenrological sequences.

Marshall et ai (1997J include ihc tnatninal hearing

beds ai Tiupampa in the niiddlc scquencc of the

Santa Lucia Formation, wbith ihey hâve conelared

with the lasr sequence of rhe underJying El Molino

Formation to the chron 26r. Idowevcr, the authors

hâve not presented ihc sedimcniological argumcnls

which jIIow the suçgcstcd corrélation and the .sedi-

mentoiogy and ieposirional environment at

riupampa are ncither descrîbcd aor diseussed.

Fürtherniore, some basic considérations seriously wea-

ken the corrélation proposed by Marshall et al, (1997)

and Sempere et al (1997): (1) the Tiupampa sériés is

located on the edge- of the El Molino-Santa l.ucta

Basin and theretore the sédimentation may he atypical

and may bave very local characteristics; this observa¬

tion obviousiv makes corrélations based on sedimen-

lological data very hazardous; (2) the Tiujiampa beds

wcrc deposited on an alluvial plain, an environinenr

which is generally charactertzed by an irrcgular rate of

sédimentation; (3) scveral palcostds with root casts

have bcen inentioned ai Tiupampa (Muiz.on ci ai

1983) which cûuld correspond to important iiuerrup-

rions in the sédimentation- For example, the single

reversed iincrval of Tiupampa could very well corres¬

pond to chron 27r and 26r wiih lack of sédimenta¬

tion, or crusion of sedirncnis correspondmg to

chron 27n. Since no radioineiric dates art available,

so far, ar Tiupampa, this dîHiculty cannot bc re.solved;

(4) alrhough considerably improved dufing the lasT

Fifteen years, rhe knowledge oi the Lare Cretaccous

and Palaeocene of Bolivie is sril! very incomplète, a

stacement ro which must be added general considéra¬

tions concerning the s'ariabiliry of the rates of sédi¬

mentation and ine great ditficulty in derennining the

importance of scdimcnurion gaps.

Furthermore, at Vila Vila. at about 5 km east of

riupampa, iii the same ouicrop, a Maasrrichrian ftsh

launa (Gayet ci al. 1992: 400) indicaies rhe présence

ot Maastrichtian beds ol rhe Hl Molino Formation

(probably belonning to die lower seqtience of the for¬

mation). fhose Deds have a faciès ol mainly thick red

sandstones cortesponding to a more detrital sédimen¬

tation, uncommon for the El Molino Formation and

probably indicative of the local border of the basin.

which could cxplain the atypical faciès found at Vila

Vila and Tiupampa.

In spire ot Marshall et al. (1997)*s daim of having the

only accepfabic approath u» ilic rclarive chronology,

rhe corrélation they suggested c>t rhe Tiupampa beds

to the early Tiffanian is still weakened by numeroas

unknowns and approximations, l'heir âge assignment

is certainly not convincing (as far as the rnammal

fauna is concerned) and, in spite of iKe appearances of

a rigorous quantitative analysis, it is not more objecti'

vc than a raunistic approach, which. as mentioned.

above. présents the Uiiccriaintic.s that cvctybody

knows. Marshall et ai (1997) may bc correct

(although tt would be excrcinely surpnsing) in assign-

ing a Tilbtnian âge to the Tiupampa beds, however

they failed to demonsiraie it.

Therefore, on ilve basis of huuiisik comparisons, rhe

âge corrélation of the Tiupampa rnammal fauna tn

rhe i*uercan is srill accepted here, following Van

V'^alcn (1988), Bonaparte et ai (1993), Muizon &

I. Brito (1993), Marshall et al. (1995) and

Bonaparte 6i Morales (1997).

DHSCIUPTION

.Serial désignation for teeth follows Luckett

(1993) contra Archer (1978), Hershkovitz

(1982), Marshall & Muizon (1988) and Muizon

(1992), i.e. prcmolars arc PI, P2, P3; dcciduous

tooth is dP3: permanent molars arc Ml, M2,

M3, M4: terminology for molar structure fol'

lows Marshall & Muizon (1995); usage of

Metaiheria follows Szalay (1994) and S/ulay &

IVofimov (1996); usage of 7’heria follows Kielan-

jaworowska et ai (1987). and Tribosphenida toi-

Itjw.s McKenna (1975); supragencric ranks ol

MeialherJa lollow Marshall et ai (1990). AJl

measuremenrs are in millimétrés (mm).

Abbreviaüons of teeth are as follows: c. lower

canine; C. upper canine; i, lower incîsor; I.

upper inciser; m, lower mular; M, upper molar;

p, lower premolar; P, upper prcmolar.

Anaromical clemenrs are described in che follo¬

wing orderi upper dentition, lower dentition,

skuil, dentary, venebrae, forelimb and hindlimb,-

When necessary to the understanding of the

text, the description will include some compari-

sons, mainly with the other borhyaenoids and/or

with didelphids. In the following description of

the dentition of Mayulestes feroxy comparisons

will be made essentially with two borhyaenoids

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

(Allqokiriis australis from thc same locality and

Patene simfmni froni the early late Palaeocene of

Brazil) and with Ddtatheridiurn praetritubercu-

lare, a rribosphenidan from the Lare Cretaceous

of Mongolia.

Skuli

The dental formula of Mayidestes ferox is the

complété primitive marsupial formula: I5/i4;

C/c; P3/p3; M4/m4 (Owen 1868, but see also

Luckett 1993).

Upper dentition

Upper incisors. (Fig. 1) The number of upper

incisors is nor reduced to four as in other

borhyaenoids. } lowever, it is noteworthy that the

number of incisors is unknuwn in Allqokims,

Patene and NemoleTtts, thrcc other Palaeocene

borhyacnoids. The incisor tooth row is parabolic

and differs from the stratglit (or sUghtly concave

posteriorly) row observed in Miocene borhyae-

noids (Cladvsiitis, Sipalocyon^ Prothylacynus,

Acrocyon). In the hulotype lA' Mayidestes, thc left

Il is not preserved and the clowns of the Icft 12

and right 13 are broken. Il is thc largesr incisor.

In diameter of the crown II > 12 < 13 > M > 15-

The crown of Tl is conical, slightiy recurved pos-

teriorly and approximately iwicc as high as wide

at the base. The ape^ of the crown of 12 is bro-

ken; this tooth is the second smallest upper incL

sor in diameter, The section of the crown is

subcircular, slightiy flaitened labiolingually. The

anterior base of ihc crown t.s jîrotrudlng ante-

riorly. ‘l'he 13 has a much lowcr crown than the

II; it is triangular in labial view and clcarly flat-

tened labiolingually. l’hc crown of 13 is protru-

ding anteriorly and posteriorly slightiy below thc

juncrion with the root. l'hc mcsiodistal diameter

of the crown i.s only slightiy smaller than its

heighr. The 14 lias a much smaller crown than

thc 13. It is triangular in labial view, les maxi¬

mum mcsiodistal diameter is also locaccd below

the crovvn-root contact but, contrax)'^ to 13, it is

slightiy larger than rhe height of the crown. The

15 is the smallest of the upptr incisons. The

Crown is conical and not constricted ar the

contact with thc root. 'Lite labial edges of thc

alveoH of 11-4 are ac thc same Icvel while that of

15 is situated a little tnore dorsally at the anterior

edge ol a large premaxillomaxillary fossa for the

tip of the lower canine. Tliere is no true dias-

lema, althougli 11-4 are not as close to each

other as 14 and 15. A distinct diasrema separates

15 from C.

Upper canine. (Fig. I) l’hc tooth is large, poin-

icd and slcndcr. It is complété on the right side

only where k has been partially expelled from its

alveolus during fossili'/ation Lhe right C there-

fbre appears ro be longer than it probably was on

the üving animal. The crown is Icss than half of

the length of thc observable portion of thc tooth.

It beats an important anterior wcar facct lor thc

lowcr canine, which continues dor.sally on thc

root. In cross section, the tooth is ovoid, being

Fig. 1. — Mayulestes ferox, holotype (MHNC 1249). Anterior part of the palate with incisors canines and anterior premolars. Scale

bar: 1 cm.

GEODIVERSITAS • 1998 * 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

longer than wide (1: 4.9 mm; w; 1.9 mm). The

tooth is recurved posrcriorly but ro a lesser

cxtent than in PiicadelphySy as the tip of rhe

Crown is anterior to rhe level of the posterior

edge of rhe alvcolus. It approaches the condition

observed in Sipalocyon^ although il is approxima-

tely 40% smaller.

Upper prcraolars- (Fig. 2) The three iipper pre-

niolars of Mayulestes are double-rooted. The size

of the teeth increases from PI lo P3. Pi is consi-

derably smaller than the other premolars. The

différence in size between PI and P2 is more

important than in other borbyaenoids and

Piicadelphys but approaches the condition obser¬

ved in Deltatheridium, Pl bas a litiy single-

cusped asymmetrical crown. In labial view, its

posterior edge is strongly oblique (in relation to

the alveolar border) and rccfilincat whilc the

anterior edge is subvertical and convex anterior-

ly. At the base of the postctolabial edge of the

crown is a miiuue cuspule. Pl is .set slighdy obli-

quely in the maxilla, as observed in Piicadelphys

(Marshall èc Muizon 1995) and Andinodelphys

(Mtiizon étal, 1997). However, the obliquity of

the left Pl has probabiy been incrcased by the

deformation duç to lossilization. Fhe crown of

the P2 is triangular and longct than high in

labial view. In occliisal view, it is flattencd trans-

versally. It présents a small cuspule at its anterior

base and a tonspicuous cusp at its posterior base.

The apex of ihe crown lies veiitrally below the

posterior edge of the anterior root. The cristae

running from the apex to the basal cusps arc

straight and rhe posterior is longer than the ante¬

rior. The P3 is the liiglicst of ihe chcck-tecth. It

is a littic longer than P2 and much hlgher. ‘Fhc

différence between the proportions ol P2 and P3

is more important in Mayulestes than in

Pttvadelphys and the other borliyaenoids but

re.scmbles ihe condition observed in Delta-

theridiiON. Phe crown of the P3 is clcarly higher

than long, as in Pucadclphys^ Patene and

Sipalotyon, The anterior edge of the crown is a

smootit crista, convex anteriorly and possessing a

small heel ar its base. The posterior edge is a wcll

marked cristHv concave posteriorly with a

well-defined cusp at its base, 'fhe apex of the

crown lies ventrally to the anterior edge of the

posterior root and, in latéral view the crown is

slightiy recurved posteriorly. On P2 and P3 there

is a small posrerolabial cingulum which disap-

pears anteriorly.

Upper molars. (Fig. 2) In the only specimen

knovvn, the rmdars are slightiy worn but the den¬

tal morphology is clearly recognisable. In lengih.

Ml < M2 > M3 > M4 and tn widrh Ml < M2 <

M3 > M4 (Table 1 ). The molars of Mayulestes

are relarively shorter anteroposreriorly than in

Allqahrus and Patene, On M1-3 the prorotonc is

relatively large for a borhyaenoid but narrow

anrcroposteriorly .ind sligluly inflated basallyi

especially on its posterior sidc. In Allqoleims^ a

borhyaenoid from the same localiry, rhe proro-

cône is even shorter anteropostcriorly antl not

iifflaied basally; ihc protocone of Mayulestes is

shorter than in Pateiie, l he prolocri.stae arc

straight: the preprotocrista is slightiy shorter

than the postprotocrista and, as a consequctice,

the apex of the protocone is located in an ante¬

rior ptysition on the cusp. VUe trigon is wtll basi-

ncd. The conules are stnall but wcll marked; they

are distinctiy V-shaped and vvell separared from

the paracone and rnetacone as observed in

Atlqokînis. Because of the sligfii wear of tlie teeth

it is not possible tu deterrnine with certainry the.

relative size ol tbc conules. A distinct paracingu-

liim unités rhe prepnraconulc crista to the siylar

cusp A. In face, rhe paracingulum is continuons

with the preparaconule crista as in Allqokirus and

Patene. The para- and rnetacone are wcll-devclo-

ped, well separated and constitiice the central

cusps of the molars. The rnetacone is slightiy

compressed transvcrsally, this being particularly

obvions on M3 as in Allqokirus. It is distinctiy

larger in height and volume than the paracone.

In occlusal view, the paracone is approximately

30% shorter and narrower than the rnetacone.

On the four molars the paracone is more worn

than the rnetacone and on M3 the paracone is

relatively worn whilc the rnetacone is almo.si

unworn. In spitc of this> the diftcrcnce in height

of the rwo cusps is so important thar ic is clcar

char the paracone was lower than the rnetacone.

l'hc centrocrista is straight in occlusal view. In

labial vic-w, the centrocrista is dceply V-shapcd

and présents a notch at the junction point of the

premetacrista and postparacrista. From this

point, the highest of the centrocrista, a deep

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

groove runs labially dividing rhe srylar shelf

transversally but noi reaching the labial cingu-

lum. The same condiiion is observed in

Allqokinis to a lesser extent, in Paierie. The

preparacrista links the paracone to rhe stylar

cusp B. It is traiisverse, al musc perpendicular to

the centrocrista. On Ml'2, ir is sUghtly concave

posteriorly. On M3-4, it is slightiy worn but it

seems thar it was almost straighr. Tt is distinctiy

concave vcntrally in anterior view. Its length

increases irom Ml ta M4 where ir is almost

three times longer rhan on Ml. The postmcta-

crista is well-devcloped and rclativcly long but

fairly low. It is approximately twice as long as the

preparacrista on Ml-3. It is almost straighr in

occlusal view and slightiy concave in postcrior

view. On M1, it forms an angle of approximately

45° with the anreropo.sierior axis of the skull:

this angle increases on M2 and it reache.s almost

80° on M3. Posterior to the preparacrista and

anterior to the postmetacrisca, the stylar shelf

beats two elongated fossae directed postcrolabial-

ly and anterolabially, respectively. Similar struc¬

tures are observed in Allqokirus and Paierie. The

stylar shelf is similar to that oi Allqokirus h\xt it is

larger and deeper than in Paierie. Jt differs Irom

the condition in other borhyaenoids, where the

srylar shelf is either very reduced or, generally,

toially absent. Its width increases from Ml to

M3. The stylar cusps are relatively well-develo-

ped with the exception of stylar cusp C, which is

absent. Stylar cusp A is medium-si/ed on Ml-2

and distinctiy lowcr than stylar cusp B. The two

cusps arc connate but easily discernible. On M3,

stylar cusps A and B hâve the same height and

are connate but not totally luscd. On M4 they

are fused, 'l’hc stylar cusps A and B of Paierie are

smaller than in Mayulestes but their relative size

and relationship are similar The stylar cusps A

and B of the M3 of Allqokirus (the only molar

known) are similar to those of Mayulestes. As

mentioned above there is no stylar cusp C in

Mayulestes as m Allqokirus and Paierie. In rhese

ihree gcncraj there is a distuict labial cmgulum

in the Stylar cusp C position. The stylar cusp D

is wcll'developed in Mayulestes, contrary to

Allqokirus where it is very small and Paierie

where it is absent. In Mayulestes. a small crista

runs lingiially from the cusp w'irhin the basin of

the Stylar shelf, but does not reach the postmeta-

Fig. 2. — Mayulestes ferox, holotype (MHNC 1249). Left upper molars and premolars: A, occlusal: B, labial views. Scale bar; 5 mm.

GEODIVERSITAS • 1998 • 20(1)

MayuUstes, a borhyaenoid from the Palaeocene of Bolivia

crista or the metacone. This crista is absent in

Allqokirus and Patene. The stylar cusp E is small

as in most marsupials and in facr constitutes the

labial extremiry of the postmetacrista. On Ml-3

ühc eccoflcxus is vcry dccp and its depch incrcases

from Ml to M3. It is deeper than in Allqoktrus

and much deeper than in Patene and Sallacyon.

An ectoflcxus is absent or extrcmely reduced in

the other borhyaenoids.

The M4 is very short anteroposteriorly (much

shorter than in Patene) but still recains a meta¬

cone as in Patene and Sallacyon (il is tiny in this

genus), contrary to the other borhyaenoids. The

trigon is still wel! basined and the stylar shelf is

reduced but présent. The stylar cusps A and B

are totally fused in a very large “parastyle”; there

is just a hint of stylar cusp D and the stylar

cusp E has totaJly disappeared.

Power dentition

Lower incisors, (Fig. 3) In the description of the

lower incisors of Mayulestes the médiat inciser

will be regarded as the i2 and the latéral as the i5

following Mershkovicz (1982). The four incisors

are preserved on the right dentary only. Their

relative volume in occlusal view is as follows:

i2 < i3 > i4 > i5. The incisors hâve a single eus-

ped Crown. The i2 is slighrJy compressed mesio-

distally and i3-5 are compressed labiolingually,

i4 and i5 being clearly spatulate. The incisor row

is a regular arch but the root of i3 is staggered

and buttresscd labially a.s ob.verved in most didch

phidst several borhyaenoids (Borhyaena,

Siaplocyon, Cladosictis^ jy^ylacinus), somc dnsyu-

rids and peramelids (Hershkovitz 1982). A stag¬

gered i3 is also présent in Pucadelphys ;ind

Andinodelphys (Muizon et ai 1997)-

Lower canine. (Fig. 3) The lower canine is large

ahhough conspicuously .smaller and shorter than

chc upper canine. It is longer than widc and

compressed cransversally (this is particulaiiy clear

on the root). The large axis of the tooth is sec ai

an oblique angle in relation to the cheek tooth

sériés. It is genily recurved posceriorly and its

posterior edge beats a distinct wear surface from

the upper canine.

Lower premolars. (Fig. 3) The pl is only pre¬

served on the right hemimandible; the posterior

heel of the p2 and complété p3 are preserved on

the left hemimandible. The three premolars are

double rooted. The pl is very small, single eus-

ped and criangular. It is set slighdy obliquely in

the dentary, as in the Borhyaenidae (ahhough

this feature is much more emphasised in the

représentatives of this family). In labial view the

posterodorsal edge of the crown is three rimes

longer than the anrerodorsal. Therefore, the

tooth is strongly asymmetrical and its apex is dis-

placed anreriorly and occurs above rhe middie of

the anterior roor. l'he anterior edge of the crown

overhangs the anterior border of rhe anterior

root. The p2 is known by its po.sterior heel only,

which is similar in s'izt with that of p3- The p3

has a high crown. It is highet than long, as in

Patene and Sipalocyon but contrary to Cladosicth.

As in Patenty the apex of the tooth is located

above rhe anterior root, contrary to Cladmktn

and Sipalocyon where it is located above the

middie of the tooth. The anterior edge of the

main cusp is convex and the posterior edge

concave; the former is approximately 40% shor-

cer than the latter, Because of the shape of its

edges, the main cusp of the tooth is slightiy

recurved po.sceriorly. In occlusal view, the labial

side ol rhe main cusp is .strongly convex. while

the lingual side is reJativcly fiat. The anterior

edge of the tooth has no basal cusp but the ante¬

rior base of tbe crown protrudes anteriorly and is

elevated. The concave area below the protruding

base of the crown receives the posterior heel of

the p2. Such a condition is aiso observed in

Patene and in most didelphids. It is absent in

other borhyaenoids where die check tooth row is

less compressed anteroposteriorly. There is no

diastema between the canine and pl and bet-

ween the premolars.

Lower molars. (Figs 3, 4) In length and height

ml < m2 < m3 < m4 and in widvh of the trigo-

nid m 1 < m2 < m3 > m4. However, the increase

in size is moderate, contrary to the condition in

most other borhyaenoids., The lower molar

structure is similar to that of Allqokirus but che

tooth proporrion.s are more slender and the m3

of Mayulestes (the only lower molar known in

Allqokiras is an m3) is smaller and narrower than

that of Allqokirus, Besides the size, the main dif¬

férence between anterior and posterior molars is

the shape of the trigonid in occlusal view, which

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

becomes wider and less opened lingually froni

ml to m4. On thc four molars the trigonid is

much higher and wider than rhe talonid. The

protoconid is very large and is much higher than

the other cusps of the trigonid. It has a triangu-

lar section and is swollen labially while the lin¬

gual side is flat. This morphology is commonly

observcd amoiig borhyaenoids but is bcttcr mar-

ked in Mayulestes^ Allqokirus^ Patène and

Nemolestes than in ihc odier borhyaenoids. The

paraconid is similar in proportion to rhat of

Patene, but is smallcr than in mosi other

borhyaenoids. It has a triangular section in

occlusal view. On che anterolingual edge of the

Fig. 3. — Mayulestes ferox, holotype (MHNC 1249). Right hemimandible; A. labial; B, occlusal; C, lingual views. Left hemimandible:

D, lingual; E, occlusal views. Scale bar: 1 cm.

GEODIVERSiTAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

paraconid is a large, salient prcparacrisrid,

apprcssed againsc thc lingual sidc of the hypoco-

nulid of the prcccding rooth. In fact rhe latter

interlocks with thc anterior edge of the folJowing

Fiq. 4. — Mayulestes ferox, holotype (MHNC 1249). Right lower

molars: A, labial; B. lingual; C, occlusal views. Scale bar; 5 mm.

rooth. 7’he anterolingual sidc of dic paraconid is

concave and receives the preceding hypoconuUd

which is limited lingually by dic paracriscid and

labially by rhe lingual edgc of the prccingulid.

This featurc is likely lo be plesiomorphic sincc ir

is obscrved in ail the othcr borhyacnoids

(although somehow altercd in hums with vcry

robust Jenixtions)» in didclphids, in several

Cretaccous Theria (Kielantherruni, Asioryctesy

Prokennalestesy Kokopcllia^ Eodelphis^ Didd-

phodofh Alphadofty Glmbius) and în the eupanro-

there Pertimm.

Whcn unworn [i.e. on m4), che paraconid and

the mctaconid arc subequal in hcight. The meta-

conid is slighdy less robust and shorrenn lingual

view than ihc paraconid. It has a subcircular sec¬

tion in occlusal view and is rclatively large for a

borhyaenoid. Ir resembJes in sh.c rhosc of

Allqokirus and Paiene (although, in rhe lacccc

genus, it is somerimes slighdy smaller than the

paraconid), but ir i,s Urger than in ail rht other

borhyacnoids which stül retain a metaconid

{Nernolestes^ Pharsophorus^ Plestofelis), In these

gênera thc metaconid is always much smaller

than the paraconid. The paracristid and proto-

cristid are sharp and show a conspicuous carnas-

sial notch ai thc limir of the cusps (respectively

paraconid and protoconid; protoconid and meta¬

conid). The protücristid is almost tninsver.se

while tlîc paracristid is sirongly oblique in rela¬

tion lo the rooth row (the obliquity is more pro-

nounced in the anterior than in rhe po.sccrior

molars). Al thc antvrolabial base ot thc trigonid is

a strong prccingulid which runs from the labial

angle of thc paraconid to the base of the protoco-

nid above the middle of the anterior root.

The talonid is narrower than in Allqokirus.

Patene-, Sipalocyon and Clndosictis. The basin of"

the talonid is moderately devcloped and opeirs

lingually hetween the eiitoconid and the metaco¬

nid. The talonid is characterized by thc large size

of its hypoconid and ihc great réduction of iis

entoconid. The hypoconid is ihc largest cusp of

the talonid. It is much larger than the entoconid

and more voluminous than the hypoconulid,

Although ît is slightly worn on ihc spécimen des-

cribed, on ml-3 it was probably as high as the

hypoconulid and slightly lower on m4. The

entoconid is very small and much less volumi-

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 5. — Mayulestes ferox, holotype {MHNC 1249). Skull: A, dorsal; B, ventral; C, latéral views. Scale bar: 1 cm.

32 GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

nous and lowcr rhan the hypoconid. In this res¬

pect, it resembles the condition observed in

Allqokirus but dilfcrs (rom Paierie where the

entoconid (alihough Icss voluminous than the

hypoconid) is less reduced than in Mayidestes

and has approximately the samc height (some-

times it is higher) as the hypoconid. A well-

devcloped crLstid obliqua connccts the anterolin-

gual angle of the hypoconid lo the po.sterolin-

gual angle of the protoconid slightiy labial to the

notch of the protocristid. The hypoconulid is

moderately large on m I -2 where it is only slight¬

iy higher than the entoconid. On m3-4 it is

voluminous and high. On rhe four molars it is

very salienc posteriorly» contrary to the ochcr

borhyacnoid.s. It Is located in a lingual position

but not connate to the entoconid. This condi¬

tion is probably due to the grcat réduction of the

entoconid, sincc this feature is very clear on the

other borhyaeiioids. On ml-3 a large posteingu-

lar shelf extends ventrolabially from rhe tip of

the hypoconulid across the posterolabial surface

of the hypoconid. The posteingulid is very redu¬

ced on m4 and does not reach the tip of the

hypoconulid.

Bony skull

General features. (Figs 5, 6) The only skull

known of Mayidestes is dorsoventrally crushed,

but it is probable that it lirrle aftected the ventral

and dorsal views of the skull The major damage

to the skull is in the basicranium where the

basioccipital, the basisphenoid and the alisphe-

noid hâve been pushed down into the braincase

(the skull ha-s fossilised with the palate facing

dorsally). The petrosals, the glenoid fossae and

the occipital condylcs hâve resisted the pres.surc

berter and remained salienr. The skull is small

and relaiively short anieroposieriorly when com-

pared to gracile forms such as Cladosicîis and

Sipalocyon. Ir is not as stout as that in Borhyaena

and approaches the proportions observed in

Pnnhylacynus. Alchough the ro.strum seems to be

shorter than in orher borhyaenoids. Table I

(p. 86) shows that rhe ratio of length of the ros-

rrum (from the tip of ihc prcmaxillae to ilie

anterior foramen of rhe infraorhital canal) to

total length of the skull is higher in Mayulestes

than in any Santa Cruz borhyaenoids. The ros-

trum of Mayulestes does not hâve the strong

postcanine constriction observed in the

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Santa Cruz borh)'^cnoids and, in rhis respect, il

clearly resembles the morphology observed in

didelphids. Tht* temporal fossae are long and

represent more than half of the skull length. The

maximum width of the skull is located at the

posteriof e.xtremity of the temporal fossae. slight-

ly antcrior to ihc glenoid tossae. l ht* zygomiuic

arches are regularly curved from iheir antetior

root 10 the glenoid cavity> whcrc ihcy form an

almosi right angle beforc coiitaccmg chc brain-

case. This condition is similat to that observed

in Borhyaena, and PrathyLtcymis^ Imt

differs from that in Sipalocyon where the zygo-

matic arch is' regularly convex frc>m its anterior

to its posterior tout and where the maximum

width of the skull is more anterior. The interor¬

bital constriction is weak and bas a width close

to that üi the temporal fossa. In ail other

borhyacnoids U is clearly natrower. The brainca-

se is large and not clearly separated from the

interorbital bridge. Irs latéral edges are almost

straighl, regularly divergent posteriorly and not

convex as in other horhyaenoid skulls,

A small but well marked sagittal crest is présent.

It begins on the posterior edge nf the fronrals^

runs posteriorly on the parierais and rises at its

contact wirh rhe lambdoid cresu It is low and

bas the same élévation ail along ihe pariétal. The

posterior part of ihe braincase shows a well-

developed dorsoposteriorly directed lambdoid

crest. Both the sagittal and lambdoid crests are

more redueed than in the other borhyacnoids

and the lambdoid crest dues not overhang the

occipital condyles posteriorly.

In ventral view, thé palate is wide and the dental

rows arc siraight to convex latcrally, contrnry to

other horhyaenoid where they arc gencrally

concavo-convcx. The glenoid fossa is located

posteriorly, latéral to ihc anterior cxtrcmic)'^ of

the promontorium of the pars petrosa and to rhe

foramen Ovale. The lattcr is located at the posiero-

ventral angle ot the alisphenoid. Ivs anterior bor¬

der is formed by the alisphenoid and its posterior

edge is formed by the petrosal.

Nasal. (Figs 5A, 6A) ‘The bones extend antcrior-

ly beyond rhe anteriormost point of the nasal-

premaxilla suture^ The anterior halves of the

nasals are eiongace and narrow. They become

progressively narrower posteriorly and are narro-

west at rhe level of a rransverse line joining the

anterior foramina ol the infraorbiral canal.

Posterior ro rhis line, rhey strongly and rapidiy

widen and reach rheir largest width ai the triple

junction wirh the Irrjncal and the lacrimal where

rhey form a small postctolaterally directed horn.

The fronial-nasal suture is markcdly W-shaped,

with the buse of the W facing posteriorly. The

nasals of Mayulcsies ferox hâve a clcar contact

with ihc lacnmal, a plesiomorphic featuie obser¬

ved in ail oihcr borhyacnoids.

Premaxilla. (Figs 5, 6) The anterior edge (rhe

alveolar border) of tlic prcmaxillae is gently

arched when viewed dorsally or ventrally and the

bones proirudc anicriorly conirary to what is

observed in other borhyacnoids, wlicrc the ante¬

rior edge of the premaxillac is almost straight

and transversal. As a conséquence, the incisor

tooth rowv in venrnil view, is a vvidc open para-

bola while ir is gencrally straight in other

borhyacnoids. The ascending pmcc.ss of the pre-

mxxllla is long and siender and insères berween

the nasal and the maxilla postciodorsally. The

length of rhe preniaxilla-nasal suture is approxi-

mately 23% of the total length of ihe nasal. Fhe

width of ibe ascending process is rpughly

constant from its anreroventral extremity to the

anterior point of the nasal-prcmaxllla suture,

rhis pan of the process, approximately hnlfof it.

forms the latéral edge of tlic cxtcrnal narcs. The

posterior half ot the process articulating with rhe

nasal is very sharp and progressively tapers poste¬

riorly.

At the anterior point of chc nasal-prcmaxilla

suture the anterodorsal border of the ascending

process shows a distinct anierodorsal protrusion

which corresponds to the point of junction bev-

ween the plane of the external nares and ihe dor¬

sal surface of the rostriim. The two planes form

an angle of upproxiinutely 140°. Latcrally, the

maxilla-preniuxilla suture run.s anterovenrrally

and rcachcs the alveolar border slightly beyond

rhe upper canine and the hottom of the fossa for

the lowcr canine. In this fossa the maxilla-

prcmaxilla suture runs posteromedially and

reaches the lingual alveolar border at the antero-

mcdial angle of the canine. At this point the

suture takes an almost anteroposterior direction.

On the lacerai border of the canine fossa, the

GEODIVERSITAS • 1998 • 20(1)

Mayulestes-, a borhyaenoid from the Palaeocene of Bolivia

premaxilla is iii CDiitaci with the antérolatéral

process of the maxilla which torms the postcrior

third of the latéral wall of the canine fossa. Its

two anterior thirds :uc foimcd by the prem;ixilla

itsclf. In other borhyacnids, the maxilla forms no

part of the latéral border ol the lossa for the

lower canine while In the didclphids and in

Pucadelphys (carly Palaeocene of Bolivia) the

maxilla forms the entire latéral wall of the fossa.

The horhyacnoids arc charac.terl 7 .cd by the loss of

the maxillary antérolatéral process which borders

the canine tossa laterally. The ctmdition of

Mayulestes represents an incipicnc deveiopmem

of this borhyaenoid feature, since the antérolaté¬

ral process of the maxilla is greaily reduced when

compared to the didelphoids but has nor tocally

disappeared as is observed in lhe other borhyae-

noids. The fossa lor the iiiferior canine is located

anterior to the upper canine. les anterior part is

in the premaxilla just behind and dorsal to the

15. l’hc battt)m of this fossa is in the small por¬

tion of the maxilla. antérolatéral to the upper

canine and its latéral rim is formed by the m;ixilla.

Its dorsal extremity is situated more ventrally

than in Cladosktis^ Sipalocyony AcrocyoUy and

Borbyaena. The incisive fbramina are elongated»

narrow and slightiy concave laterally. Their ante¬

rior extremity is ai the levcl of the posterior end

of the 13 and their posterior end rcacJics the ante¬

rior border of the canine. The nacdial edges of the

incisive foramma arc made of the ventromedial

processes of the premaxillae, whjch seem to

diverge slightiy posteriorly* It is also poswsible thaï

the divergence of the processes is tho resuit of the

deformation of the skull due to fossilization.

Maxilla. (Figs 5, b) The dorsal process of the

maxilla is a low rounded bladc, articulating ante-

riorly with the premaxilla, dorsoincdially with

the nasal, postcriorly with the lacrima) and late¬

rally with the jugal. It has no contact with the

frontal. ITc posterior edge of the dorsal process

has a transversc suture that is roughly convex

anteriorly. It articulâtes laterally with the jugal at

the level of rhe anterior part of M2 and passes on

the ventral side of the skull. The jugal-maxilla

sutures are almost parallel to the alvcolar border,

slightiy converging anienorly. The anterior ope-

ning of the infraorbîtal canal opens dorsal to P3.

It is of moderate to small size and probably

slightiy elongated dorsoventr.ally. This observa¬

tion is difTicult becausc of the crushing of the

skull. riie posterior opening is flattened by

deformation but it is likcly that it wa.s elongated

transvcrsally. It opens dorsal to M2. It.s dorsal

rim is lormed by the lacrimal and its ventral rim

by the maxilla. It is not possible to déterminé il,

as in Pitcadelphys. ihc palatine also participâtes to

the médial rim of tfie loramen. In Récent rneta-

therians, the infraorbital canal transmits the

infraorbital branch of the V2, a branch of tbc

infraorbital attery and a small vein (Archer

1976). The flüor of the orbit is formed by the

maxilla. Ils suture with the palatine, medially, is

not observable.

On che ventral side of tbc palare, ihcrc is a large

postpalatine foramen with an orientation similar

to that observed in Didelplm. It faces antcromc-

dially with a small ventral component. Becausc

ut this condition, in ventral view, the postpalatine

foramen appears as an elongated slit with an

anterolaieral-po.steromedial orientation. This is

clear on the lelt .side of che specimen where the

foramen is intact. The postpalatine foramen of

Mayulestes is proportionally much larger than in

other borhyaenoids. Us antérolatéral edge is for-

med by chc maxilla and its dorsal, ventral and

posteromedial edges are formed by che palatine.

The palatine-maxilla suture is difTicult to déter¬

miné but il probably lormed a regular parabola

with an apex at the levcl of Ml and with

branches reaching ihe anterior border of the

postpalatine foramen. Altbough the palace has

been defornicd it Ls possible to observe that it

was foirly deep and hollowed. The palatal por¬

tion of the maxilla bas no vacuiries,

The anterior extremity ol the maxillae on the

palate .shows two small antérolatéral processes

which fbrm the posterolarenil rim of the incisive

foramitia. On its anieromedial edge each maxilla

shows a tiny anteromedial process wedged bet-

weeti the ventromedial process of the premaxilla

and the vomer.

Palatine. (Figs 5, 6) On the posterior extremity

of the palate, the ventral edge of the choanae is

relacively wide. It is approximatcly half of the

distance berween the protocones of the M4

while it is close to one thîrd of the distance in

Borbyaena, Prothylacynus^ Cladosictis and

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Sipalocyon. The edge is thickened but does noi

form A True postpalatine torus as observed in

most didelphoids (including Putadelphys, ihe

oldest membcr of rhe superfamîly knowa from

skulls and skeletons), where it is salienc ventrally

and larerally. In chis respect il resemblcs ihe

condition observed in oiher borhyaenoids and

difFers from that in ihc didelphoids. In antero-

ventral view of rhe palate ihe ventral rim of the

choanae is markedly concave ventrally A.r the

interpalatine suture, a médian spine is fbrmed by

the nvo médial spincs of each bone. A lacerai

spine is locared on tlie posterior edge of the pose-

palatine foramen and onented ventrally. Ir arti¬

culâtes wirh the anterior extremiry of the

pterygoid wing and is closely appressed lo it ft

also articulâtes with the maxilla anteriorly. The

latéral spine and its articulation with the ptery'

goid are located ventrally to rhe media! spine.

Although the palatine bones aie broken, it is

possible to observe that they had no palatal

vacuities a condition found in ail the other

borhyaenoid.s.

The orbital portion of the palatine is badly

damaged becaase of the crushing of the skull,

and, therefore, its rclationships with the lacrimal

and the maxilla arc very difficult to observe.

However, on the right side of rhe specimen, a

large sphenopalatinc foramen, opening poste-

riorly, is located ventrally and slighcly po.stei'ior

to the latéral exlremity of the frontal-lacrimal

suture. Il is well anterior to and dorsal to chc

postpalatine foramen. The sphenopalatine fora¬

men transmits the sphenopalatine artery and

nerve into the nasal caviry.

Posteromedially, the palatines liave rwo large

sphenoidal processes which roof the choanae and

the palatine gutter. 'I hc processes underlap the

presphenoid and ihe basisphenoid and hâve a

.small contact with the anterior processes of the

alisphenoids. This condition is probably similar

CO thaï observed in the didelfïhids.

Pterygoid. (figs 5C, 6C) Boih pterygoids are

preserved on rhe specimen described bere. The

prerygoid of Mayulestes i.s a large (for a liorhyae-

noid), friangular blade, located latéral lo the roof

of ihe choanae. it is high and has a wcll-dcvcloped

hanuilus on its posteroventral extrcinity. It arti¬

culâtes mosrly with rbe posreromedial extensions

of ihe palatines which form the roof ol the choa¬

nae and underlaps the presphenoid and the

basisphenoid posieriorly. The pterygoid contacts

the alisphenoid posteriorly and, anteriorly, has a

very small conracr with rhe maxilla postérolatéral

to rhe postpalatine foramen. The pterygoid of

Mayulesttfs ferox i.s niuch larger that of the didel-

phids. ïr is ver)*^ reduced m other borhyaenoids

where it loses the haniular processes.

The choanal gutier of is very similar

to that uf dldelphids: it is short anteroposceriorly

and is bordered mostly by the pterygoids. In the

other borhyaenoids the pterygoids are reduced

and the choanal gutter is elongated. Its walls are

formed by rhe sphenoidal processes of rhe pala¬

tines and the pterygoid processes of tlie aJisphe-

noids. They are thick and their latéral side is

buttressed by a salient ridgc.

Lacrimal. (higs 5, 6) The lacrimal forms the ante¬

rior border of chc orbit. It has a large iriangular

portion on the dorsal surface of the skull. Its laté¬

ral portion is a small horn that articulâtes on the

mediaJ side of the anterior extremiry of the jugal.

The lacrimal articulâtes with the maxilla anterior¬

ly, witli the fn^nral posreromcdially, wirh rhe nasal

FrG. 6. — Mayuiesîes ferox. Reconstruction of the skull and mandible (latéral view): A, dorsal: B. ventral; C. latéral views (the ascen-

ding ramus of the dentary. lacking in the holotype, has been reconstructed from the mandibie ot a recently discovered specimen pro¬

bably referable to Mayuiestes). Abbrevialions: AS, alisphenoid; ahs, alisphenoid hypothympanic sinus; BO, basioccipital;

BS, basisphenoid; corpd, coronoid process of dentary; cpd, condyloid process of dentary; OEN(ar). ascending ramus of dentary;

OEN(hr), hori 2 onlal ramus ol dentary doc. dorsal occipital condyio; carn. oxiornal auditory moatiMi; cf, entocarotid foramen;

EO, exooc^pitîil; er, epityrnpanic tecess: Hc. fusàa (or lüWe> caniné. Im. forefoen magoiirn. fo. foramen ovale FR. frontal; tr, foramen

rotundum; gf. glonoid fossa: hapt, hamular procoso of thu ptüryqoid: if, mclsivo foramen. Hc. iiuraorbiial cane). JU. jugal. LA. iacri-

mal; le. lambdoidal crest (■ nuchal crest), It lacnmal toramon. most, ma&sotenc fo«a; ment momal foramen; mp, mastoid process:

mpf, médial paiadne foranien. mp$, merliai process of ttin squa<no.sat. MX maxilla. NA nasal. RA. pariptal; pgf, po^tglenoid fora¬

men; pgp. postglenoid process of Ihe squaniosai; PL. palatine. PMX, pft«rnaxilla, P(pm). pais mestoidèa of pMiromasfoirt (* rrasloid

s.str.y, P(pp). pars potrosa ol potromastoid (- potrosal s^sir.): PT. pterygoid: ppl posipalaUno foramen, ppl. poslpalat-ne torus;

prgp, preglenu'd prQceâ& ot juQât ps, pWaiür^<ii psmf. posloripr ot rrasr.otenc to.ss.i; pip, posttympanic process;

SC, sagittal crest; sica, sulcus for the internai carotid artery; SO, supraoccipital; spf, supraorbital process of frontal: SQ. squamosal.

Scale bar: 1 cm.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

rhe towl lengrii ot rlie bone) and, in lareral view>

it is less oblique rban in the other borhyaenoids.

The anieriormost extremir)' of rhe jugal is latéral

to the anterinr edge of M2. In vcnrral view, the

jugal-maxilla suture Is equidisranr along all irs

lengrh from rhe alveolar border and the sutures

on both side ol the specinien are slightly diver¬

gent posteriorly. Mcdially on its anterodorsal

side, the jugal has a small articulation with the

lacrimal. In latéral view, rhe bonc is gently sig-

nioid. fts anierior portion is concave dorsally

and forms the ventral border of rhe orhit. The

transition trom the orhit to the temporal fossa is

hardly marked by rhe inflexion ol the bone. The

posterior portion of rhe bone îs convex dorsally

and is approxiniatcly 60% ot tlic total Icngth of

the bonc. Av a levcl sHgbtly anterior ro the post-

orbital constriction, the jugal articulâtes dorsally

with the zygoniatic process ol the squamosal. Its

posterior extremity cuntributes ro rite antérolaté¬

ral edge ol the glenoid iossa, Frorn the posierit)r

end of the jugal-maxilla articulation to chc ante¬

rior end of chc jugal-squamosal articulation, ihc

jugal has a constant widrh. h tapers rapidiy

towards its anierioi extremity and gradually

towards its posterior extremity- On chc latéral

side of the bonc, a signioid rim and a shallow

sulcus for rhe insertion of the niassctcr muscle

run from the dorsal région of the posterior extre¬

mity ro rhe ventral région of the jugal, at the

level of the posterior extremity of rhe

jugal-maxilla suture.

Frontal. (Figs 6) In dorsal view, rhe fronrals

form the inrerorbiial bridge located hetween the

slightly longer rostrum and braincase. l*he fron¬

çais represent ap proxi ma tel y 28% of the total

length ol the skulh Anreriorly they have a dis¬

tinct W-shaped siitnre with the nasals and, on

their antérolatéral edge, a small suttire with the

lacrimals. Fhe posterior suture with the parietals

also has rhe shape of a W widely opened posie-

riorly. Latéral ro the frontal-parieial suture, the

dorsal pari of the Irontal-alisphenoid suture is

subverrical; the venrral portion of the suture is

not observable because of ihe crushing of the

specimen. Very weak and rounded supraorbical

processes (they resembic humps more than pro¬

cesses) are located dorsally in the middle of the

latéral edge of the bone and a reduced interorbi-

ral constriction is locyted slightly anterior to the

frontal-parietal suture. Supiaorbital processes

and interorbital constriction are mucli less deve-

loped than in rhe other borhyaenoids and the

constricrion is in a more posterior position than

in the other borhyaenoids. In dorsal view, the

Irontals are niuch less opened anieriorly (ie. the

latéral edge.s ol the l)one.s are less diverging ante-

riorly) and much narrower posteriorly than in

the other borhyaenoids. Thcre is apparcntly no

supraorbical foramen as in other borhyaenoids

and in didelphids, alrhough a liny foramen

helow the righr .supraorbital “hump*' could

represent an actual supraorbical foramen.

On tbe lareral side of tbe skull, rhe relationships

of rhe frontal with the palatine, the orbito-

sphenoid and rhe lacrimal arc not observable

bccause of the crushing of the skull.

Pariétal and intcrparietal. (Figs 5, 6) The two

boues arc rightiy fused and no suture can be dis-

tiilguished hetween tbem. l’he bones are relati-

vcly fiat wliich dénotés small cérébral

hémisphères. In the middle of cach pariétal, and

mcdial to ihe anterior edge of the posterior root

of rhe zygomaric arLh, is a small prominence

(corresponding to rhe cérébral hemispheres) and,

located behind, a small dépréssion just posterior

to the anterior extremity of the bone. Behind the

prominence is another depressed area which cor¬

responds to rhe intcrparietal région anterior ro

the lambdoici cresi Except for the relarively

weak (for a borhyaenoid) sagittal and lambdoid

cresrs, the mnscular artjchment for the temporal

muscle is not clearly évident as Is observed in

Pucaddphys byseveral irregular scars. Each parié¬

tal has a V-sbaped suture wirh the fronral ante-

liorly (rhe V being wide open posteriori) ), an

oblique concave suture with the alisphenoid (/>.

ilie two parietal-ali.splienoid sutures diverge pos-

icriorly in dorsal view) and a concave, anteropos-

icrioly oriented suture with the .squamosal. The

triple point beween the pariétal, the alisplicnoid

and the squamosal is located latéral to the hemis-

phcrical promincnces ot rhe pariétal, appro.xtma-

tely in the middle of the latéral edge of the

pariétal. Several small grooves or pits roughiy ali-

gned transversally are observed in the middle of

die cranial vauh, on both sides oi chc sagittal

crest. They do not scem to correspond to muscu-

GEODIVERSITAS • 1998 • 20(1)

Mayulestes-, a borhyaenoid from che Palaeocene of Bolivia

lar attachments but probably represent scars duc

to injuries resulring hoin the attack ot a predator

on ihe head of die holotype of Mayulestes firox,

cacching it from above and ro the side.

Orbîtosphenoid. (Figs 5, 6) This boue is preser-

ved but ic is badly crushcd. U can be ob.served in

chc bottom ot the large optic-orbital foramen.

The latéral wall of rhis loramen is fonncd by the

alisphenoid and the media] wall is formcd by the

orbirosphenoid. The foramen is locatcd dorsola-

teral ro rhe base ot the hamular process of die

pierygoid. In rhe spccimen describcd hcrc, a por¬

tion of the alisphenoid arrificially ovcrlaps the

anteromedial edge of rhe opric-orbitai foramen.

Approximately 5 mm beyond rhe postérolatéral

rim of the optic-orbital foramen, at the suture

between the orbirosphenoid and the frontal is

rhe erhmoidal foramen, which is locaied dorsola-

terally to tlie anterior third of the pterygoid. The

opnc-orbiial foramen and the ethmoid foramen

arc located in a slightiy more posterior position

than in che living Diclelphis. The optic-orbitaJ

foramen transmits the crantai nerves II, III, IV,

VI, VI, the ophtalmie artery and a vein which

drains the eye to the cavernous sinus. The eth-

raoid foramen iransmirs a branch of the internai

carotid from ihc orbit into the olfactory région

of rhe cranial c.wity (Archer 1976).

Alisphenoid. (Figs 5C, 6C> 7,. 8) d his hone

form.s the anterovenrrolateral région of the brain-

case. It bas a large articulation wiih the frontal

anteriorly and laterally; with the pariétal dorso-

laterally and with the squamosal posierolaterally.

The alisphenoid articulâtes with the orhitosphe-

noid mcdially, rhe palatine and the pierygoid

anteriorly and rhe basisphenoid and the pcrioüc

posceriorly. Anieromedially, rhe alisphenoid

forms the posterior rim of rhe optic-orbiral fora¬

men. Postero-lateral to ir is a large foramen

rorundum (which transmits che maxillary branch

of the trigcminal nerve, V2) which open.s ante¬

riorly. Latéral to the foramen rotimdum is a wide

groove bordered laterally by a sliarp crest, obli-

qucly oriented (Le. in anteromedial-posterolateral

direction). This structure represents the ventral-

mosi extension of the origin of rhe temporalis

muscle, which was apparently welLdeveloped in

Mayulestes. This structure is also observed in

Didelphis, although hcre it is much less pro-

nounced. It is absent in Borhyaena^ Cladosictis

and Sipalocyan^ but it is very welLdeveloped in

Prothyldcynus where sirong fossae, grooves and

crests arc présent in that région of che skull.

Posteromedial to rhe foramen rotundum, the

alisphenoid borders the ba.sisphenoid laterally

and, îU ihe postérolatéral extremîty of this bonc,

it contributes to rhe formation of the dorsal wall

of the etnocarorid foramen (which transmits the

internai carotid artery and a small vein from the

inferior pecrosal sinus). Latéral to rhe cntocarotid

foramen, the alisphenoid forms rhe anterior edge

of the foramen ovale and the foramen lacerum

medium. Because of the deformation of the skull

it is not certain whether the fonimen ovale and

rhe foramen laccrum medium were separated

from each other by a bony wall or were

confluent, although the first iriterpretarion is

more probable. Postcrolarerally, the alisphenoid

contributes to the formation of the alisphenoid

hypotympanic sinus (see Muizon !99i) whert: it

is wcdged between the période posrerornedially

and the médial process of the squamosal antero-

lareraJly. The alisphenoid does not possess any

tympanic process, not even incipienr. No trans¬

verse canal is observed latéral to rhe suture wiih

the basioccipital. Laterally, the ali.sphenoid

contributes ro the formation ot chc anteromedial

angle ol the glenoid fossa.

Basisphenoid. (Figs 5C. 6 C, 7ï 8 ) Ir is a large

trapezoid bone narrow anrenorly and wide pos-

teriorly. It is anteriorly undcrlapped by the sphe-

noitl processcs ot ihc pterygoids which hide its

suture with the presphenoid. The bonc is borde¬

red by rwo latéral cresrs which are higher in their

anienor portion. Along thèse crests, the basis¬

phenoid is tigluly articulated with the alLsphc-

noid but ihe suture is clearly visible on borh

sides of the specirnen. On rhe postérolatéral

angle, a reasonably dcvcloped eniocarorid fora¬

men opens wherein rhe veniromcdial edge is for-

med by the basisphenoid- The bone is Hat on

most ol its .surface, excepr for the latéral crests

and a srnall médian ridge that occurs in Jts' anre-

rior région and which îapers rapidly posceriorly.

’l'he suture with the basioccipital is siraight,

trans-verse and joins rhe anterior extremit}' of the

periotics.

Squamosal. (Figs 5-8) Dorsally, the squamosal

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

articulâtes wirh the posterior latéral half of thc

pariétal in a convex suture. Anterolaterally, it has

a subvertical suture with the alisphenoid which

turns laterally and passe.s on the ventral sidc of

the skull, where it eues obllquely the anterome-

dial angle of the gtenoid fossa, It then runs

medially, almosi reuching thc latéral border of

the foramen ovale. In this area the squamosal

possesses à médial indentation Icalled here the

media! process of die .squamosal (Muizon 1994)]

which is also présent in thc other borhyacnoids,

and in Andino-delphys Piuadelphys but absent in

other marsupials (see discussion below). The

posterior edge of this process is excavated by the

roof of the alisphenoid hyporympanic sinus and

its ventral sidc bears a conspicuous rldge which

almost joins the médial edge of the glenoid fossa

to the latéral border of the foramen ovale (the

anterior crest of the alisphenoid hypotympanic

sinus; Fig. 7). T he médial process of the squa¬

mosal articulâtes wirh the alisphenoid and, at the

point where the suture leaves the alisphenoid

hypotympanic sinus posteriorly, the squamosal

articulâtes with thc pcriotic. The suture runs

along thc latéral border of the epityinpanic rccess

and the fossa incudis. Further posteriorly, the

squamosal has an irregular contact with the laté¬

ral surface of the pars mastoidea of the période.

The petiotic'Squamosal suture is visible in dorsal

view of thc skull where a small portion of the

pars mascoidea is not covered by the squamosal

and is part of the lambdoid crest.

The glenoid fossa is deep and elongated transver-

sely. In this respect it tesembles that of

Borhyaena, more than that of any other borhyae-

noid. The axes of rhe glenoid fossae are not

exactly parallel a.s in Borhyaena and Prothylacynm

but are slighily oblique in relation lo the antero-

posterior axis of ihe skull. Its anieromedial angle

is formed by a small contribution ol the alisphe¬

noid as observed in mosc didelphid.s and perame-

lids, contrary lo the other borhyaenoids. The

postglenoid process is very wide and high. It has

a rounded outline and is approximately symme-

trical in posterior view [ue. its greatest ventral

expansion JS located at the middie of the glenoid

fossa). The preglcnoid proce.ss of the jugal is

much smaller and is located at the antérolatéral

angle of the glenoid fossa. The postglenoid fora¬

men is located on the posteromedial edge of the

postglenoid process. The vcssels it transmits

continue rheir course in a groove along thc ven-

tromedial edge of the postglenoid process. The

groove disappears approximately in the middie

of the ventral edge ol the process. The postgle-

noid foramen is exclusively lôrmed by die squa¬

mosal. while intcrnally the mcdial wall of the

canal is formed by the laicral side of thc pcriotic

and the mcdial vvall is formed by thc squamosal.

rhe postglenoid foratnen transmits the spheno-

panetaJ eniissary vein (which externally becomes

the postglenoid vein) from the prootic sinus

(Wihle 1990) and the postglenoid artery (Archer

1976; Wible 1990).

The subsquamosal foramen is preserved oa the

Icft side of thc specimen but it is redueed to a

small slit (probably slighrly flartened by the

deformation of the skull) posterodorsal to the

postglenoid foramen, jusr dorsal to the extcrnal

acoustic mearus. ïr differs from what is observed

in ail orher borhyaenoids where rhe subsquamo¬

sal foramen is aiways much larger, but clearly

resembles that of the stagodontid Eodelplm cittleri

from thc Late Crccaccous of Canada. The sub-

squamosal foramen opens into ihe postglenoid

foramen and rransmirs an artery from the post¬

glenoid foramen onto rhe pariétal area of ihe cra-

nium, which supplies rhe tempotaJis muscle, and

a vein from thc sphenoparietal emissary vein,

which exits through rhe postglenoid foramen

(Archer 1976; Wible 1990).

rhere arc no obvious postzygomaric foramina,

contraty to the condition observed in

Pucadelphys. Howevcfj rhe small foramen located

in the groove ol che postglenoid foramen could

represent the postzygomatic foramen which

rnigrated medially from its position dorsomcdial

to thc apex of che glenoid process. The condition

of Alayukstcs is found in the other borhyaenoids

and in Diclclphh. The postzygomaric foramen

transmits a vein from the posterior root of the

zygoniaiic arch lo the sphenoparietal emissary

vein (Archer 1976). In some other Borhyae-

noidea (ProthylacytitdS, Cladosictis^ Sipalocyon and

Sallacyon) another foramen is présent on the

latéral face of che zygomauc process of the squa¬

mosal, jusr above the glenoid fossa. In didel-

phids, when présent, this foramen is connected

GEODIVERSITAS • 1998 • 20 (1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 7. — Mayulestes ferox, holotype (MHNC 1249). Basicranium. Scale bar: 1 cm.

to the postzygomatic foramen. The same is pro-

babiy true in borhyacnoids although the second

foramen is locatcd turther anteriorly chan in

didelphids and thercfore more distant from the

postzygomatic foramen.

Becausc the pars mastoLdea of the periotic has

been displaccd from irs original contact wirh the

squamosal, it is diflficult to observe the posetem-

poral foramen. Howeven the pars mastoidea of

the right pcrioiic shows a clcar small notch on its

vencrolatcral angle which could represent the

postcemporal foramen. If this inteiprcration is

correct, the haramep would be in a much lower

position than in Diddphis and Pucadclphys. The

posctympanic foramen carries the arteria diploc-

tica magna and the vena dipiocrit^ magna which

pass through a canal bordered by the periotic

medially and the squamosal laterally fWiblc

1990).

Periotic. (Figs 7, 8) This bonc includes two

components: the pars mastoidea which largely

contributes to the formation ol the latéral part of

the occipital vîcw and houscs the snbarcuate

fossa on the cercbellar sidc; and the pars petrosa

which houscs the innci car, semicircular canals

and cochlea in the promontorium, and, on the

cerebellar surface, the internai auditory meatus.

The large pars mastoidea ol Mayulestes forms the

latéral area of tl^e occipital shield. It dilTers from

thac of ail the ochcr borhyacnoids wherc it is

torally internai and does not outerop on the pos-

terior face of the skiill In Mayidestes^ it is wed'

ged berween the exoctipital ventromedially, the

supraoccipital dorsomedially and the squamosal

anteriorly. A small mastoid foramen is présent on

the dorsomedial région of the occipital side of

the pars mastoidea, just latéral to its suture with

the supraoccipital. In posterior vie^v, the pars

mastoidea has a reniform shape with a concave

médial edge and a latéral edge srraight in its

middie and more or les.s convex ai its extremities.

The Surface ol the pars mastoidea is mostly Hat,

but concave in the area of the mastoid foramen.

The anteroventral edge of the pars mastoidea is

formed by the caudal tympanic process of the per¬

iotic (sensuWibït^ 1990). Its ventrolateral extremi-

ty bears a small and rounded mastoid process but

is celatively salient venirally and posteriorly,

Dorsolateral to the mastoid process is a small rim

which is bordered laterally by the squamosal, In

the-middie of that ridge is a foramen ol rcasonablc

size which has not been tbund in other Recent or

living majsLipials. Its fiinction has nor been cliici

dated. Anrerodorsally to die ma.stoid process is a

small saddle-shaped groove lor the passage of the

facial nerve: the stylomastoid notch.

The ventral or tympanic sidc of the periotic is

lormed by the large teardrop-shaped promunto-

rium which represents the pars cochlearis, and

by the portion of the periotic posterior and laté¬

ral to rhe promontorium, the pars canalicularis.

The posterior région of the promontorium is the

thickesi and bears a small vuberclc just ventral to

the cochlear window. A simtlar tubercle is also

observed in Sipalocyon. This morphology' has

been noted by Atelier (1976: 291) who observed

on rhe right promontorium of AMNH 9254

{Sipalocyon g^radlu) a iransverse welling poste-

rior to a small depresMon that he regards as

homologous to the tympanic wing of the petro-

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

sal in other mar.supit:-d.rniv<jres. In Maytdestes the

tubercle is exactly in rhc samc position as the

rostral cyinpanic proccss ol the pcciofic [= tym-

panic wing of the petrosal t)t Archer (1976)1

observed in living didelphids. Froni thaï tubercle

a shallow depression runs anteromedially, turns

anteriorly and bccomcs a deep groove at the apex

of the promontorium. This sulcus marks the

route oi passage oi the ititernal carotid artery,

which enters the skull via the entocarotid fora¬

men located antcromedial to the foramen ovale,

and is fonned by the alisphenoid dorsolaterally

and the basisphenoid ventromedially. The passage

of the infernal carotid is ihus situated medially

(sensu Presley 1979). l.accra! to the sulciii; for the

internai carotid artery is a large fossa on die laté¬

ral side of the promontorium for the insertion of

the stapedial muscle. T’his fossa is séparated from

the sulcus for the internai carotid by a high>

bony wall. l’he rnedial sidc of the promonto¬

rium has a sigmoid outline: it is deeply concave

in its posterior région (mcdial to the tubercle

mentioned above) and largely convex in its ante-

rior part forming a blade-like expansion which

articulâtes wirh the basioccipital. The media!

border of the promonroriiim is the ventral bor¬

der of a deep media! sulcus for the passage ot the

inferior petros.d sinus. This vessel is a vein which

connects the cavernous sinus (cncasing the pitui-

tary gland and che opric chiasma) with the jugu-

lar vein just bcforc it emerges from rhc inferior

petrosal foramen (Maclntyre 1972: 291). The

sulcus for the jugulât vein was misidenrified as a

sulcus for a brandi of the internai carotid artery

by Patterson (1965), Clemcns (1966) and

Marshall (1977a, 1978). The concave edge of

the posteromedia) border of rhc promontorium

corresponds to the dorsolareral rim of the infe¬

rior petrosal foramen.

Two openings exist on the posterior région of rhc

promontorium: che fenescra cochleae and the

fenestra vesribuli. The fenesrra cochleae is situa¬

ted on the postérolatéral surface of the promon¬

torium, jusr médial ro die srylomascoid notch. It

has an ovoid shape with tes ventral rim more

convex thati the dorsal as observed in

Pucadelphys and othci* borhyaenoids. The fossula

fenestra cochleae, which is well marked in

Pucadelphys and in the Late Cretaceous petrosal

type A, B and C of Wthle (1990), is absent in

Mayulestes as in the other borhyacnoids.

PüSteromcdiodorsal to che fenescra cochleae is a

funnel-shapcd pit opening posrerolaterally,

which houses the opening of rhe aqueductus

cochleae which opens in rhe posterior lacerate

foramen, fhc anccrolatcral rim of rhe posterior

lacerate foramen is formed by rhc periotic and its

pusterumedial rim formed by the exoccipiral. On

thé unierolateral sidc of the promontorium is che

fenestra vestihuli which is hardiy visible (as ic is

reduced to a .simple slit) bccause of the dorsoven-

irai squeezing of che spccimen. 1rs posterior end

is located just media! ro rhe bony rrahecula that

makes the junction berween the pars rnastoidca

and the pans petro.^a: ihc cri.sta parocica. On che

right periotic it is jîossiblc to observe that the

fenestra vestibuli is located within a shallow

depression, the fossula fenestra vestibuli.

Laterally, the promontorium is bordered by a

deep, l.-sliaped groove, which passes berween the

promontorium and the cpicympanic rccess ante¬

riorly and bcivvccn the promonrorium and the

caudal tympanic proccss of rhc periotic, poste-

riorly. The posceronicdial extremity of cite groove

Ls formed by the anrerolareral wall of vhe po.ste-

tioc lacerate foramen. The cavity located latéral

ro this walh posierior to che posterior border of

ihe promontorium and anterior to rhe caudal

tympanic process of the periotic, is the mastoid

epitympanic sinus. Antérolatéral to this sinus

and latéral to rhe fenescra vestihiili is a shallow

cupule which receives the origin of che stapedius

muscle; the fossa stapedius. The stylomasroid

notch forrns ihe angle of che L and ventral ro it

and lacerai to the stapedius fossa is the crista

parocica, which séparâtes che fossa stapedius

medially and the fossa incudis laterally.

Anterior to the fossa incudis is the epitympanic

recess. This structure has been dcfincd by Van

der Klaauw (1931: 73) and Archer {1976a: 226)

and a clcar and simple définition has been given

by Wible (1990: 188): it is “[...j the extension of

the middle ear cavity chat lies dorsal to the i:ym-

panic membrane and coniains the mallear-

incudal articulation”. In Didclphis the epiîympa-

nie recess is a smoll elongatcd fossa located dorso-

medial to ihe dorsal rim of che cxtcrnal auditory

meatus, ventral to the prootic canal and latéral

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from rhe Palaeocene of Bolivia

Fig. 8. — Mayulestes ferox, left ear région: A, holotype (MHNC 1249); B, reconstruction. Abbreviations: ahs, alisphenoid hypotym-

panic sinus: ar. antsrior ridge ol me alispher^o'td tiypoiympanic sinus; AS. alisphenoid; BO. basioccipiral; BS. basisphenoid;

et, condyloid toramen: cfpp, caudal tympanlc pror.ess ot pars mastoidea ot petromastoid: eam, external auditory mealus; ef, entoca-

rotid foramen. EO. exoçciprtal; epa. enfoglenoid process of lhe alisphenoid; er, epitympanic recess; fc, fenesîra cochleae; fl, fossa

incudis; fo. foramen ovale; fr, toramen rotundum, fs, facialsulcus: fslm, lossa for slapedlal muscle: fv. fenestra vestibuli; gf. glenoid

fossa; icn, inlercondylar notch: ipf. intehor petrosal foramen; JU, fugal; mes. mastord opitympanic sinus; mp. mastoid process, mps,

médial process ol lhe squamosal; oc. occipital condyle; pgf. poslglenoid foramen; pif posierior lacerate foramen; P{pm), pars mas*

toidea of periofic; pgp. postglenoid process ot squamosal; po. promontorium of pars petrosa of petromastoid;

prgp, preglenoid process of lhe jugûl; PT, plerygoid; pte, petrosal crest; plp, posttympàfiic process; sW, secondary facial foramen;

sica, sulcus for internai carotld artery; SQ, squamosal; ttf, tensor tympani fossa. Scale bars: 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

and parallel ro a line joining the secondary facial

foramen and the fenestra vestibuli. k is limitcd

laterally by the dorsal rim of the external audi-

tory meatus of the squamosaJ and anteriorly by

an elevated oblique (antcrolatcral-posteromedial)

ridge called the pctrosal crest. In living didel-

phids, the petrosal crest joins the triple point

periotic'squamosaJ-alisphcnoid (laterally) and a

small crest [roughly antcroposteriorly oriented

(precisely, anteromedial-postcrolatcral)] flooring

the mcdial opening of the prootic canal (medial-

ly). The petrosal crest also delineates the postéro¬

latéral border of the alisphenoid sinus which is

excavated in the période in ics posterior portion.

The resc of the aJisphenoid hyporympanic (in

didelphids the sinus is actually, ar least in part,

hypotympanic) sinus {Le. most of it) is excavated

in the alisphenoid. The posterior région of the

alisphenoid in the sinus underlaps a stnall por¬

tion of the période. The posterior porrion of the

alisphenoid hypotympanic sinus, the anterior

portion of the epicympanic recess and the ridge

that separares chem* form, in ihe période, a small

ventrally direcced ttiedra chat Wible (1990,

fig. 4) termed the latéral wall of the epitympanic

recess. Consequently^ the fossa anterior to the

petrosal crest, in Wible's illustrations (1990,

figs 2F, 4C), do nor represent part of the epitym¬

panic recess but the periotic portion of the alis¬

phenoid hypotympanic sinus. As it will be

shown below, the sO'Called alisphenoid hypo¬

tympanic sinus is nor always excavated mainly in

the alisphenoid and is not always hypotympanic

[a contradiction also noted by Archer (1976a:

127)]. However, since rhe morphology of the

middle ear sinus of Aîayulestes îs regarded here as

homologous to that ol the other horhyaenoids

and although the term hypotympanic is inappro-

priate in the case of ALayulesies» it will bc used

here (as clsewhere, Muizon 1994) in order to

avoid confusion.

In Aîayulestes, the epitympanic recess is an elon-

gated fossa Ümited posteriorly by the dorsal por¬

tion of the crista parotica and anteriorly by rhe

petrosal crest. The small pit at îts postérolatéral

extremity', latéral to the dorsal part of the crista

parotica, is the fossa incudis (or fossa crus breve

incudis) where the ligament of the small process

of the incus is attached. The epitympanic recess

and the fossa incudis are bordered laterally by

the dorsal rim of the external audirory meatus of

the squamosal. However, beciuse of the crushing

of the specimen described here, they are opened

laterally since their latéral wall is displaced late¬

rally. Following Wible (1989, fig. 4A, C),

Muizon (1994, fig. 2a) misidentified the epitym¬

panic recess of Mayulestes. What I identified as

the epitympanic recess is in fact rhe posterior

part of the alisphenoid hypotympanic sinus and

the fossa incudis (/>? Muizon 1994) is the epity-

mapnic rece.ss. In its anterior portion, the epi¬

tympanic recess is bordered by two small crests.

The latéral oiic was probably in contact with the

squamosal or close to it and is médial to the ven-

traJ opening of the sulcus for the prootic ^inus.

The médial crest forms rhe floor of a small

trough located just latéral to the secondary facial

foramen and dorsoiareral to the petrosal crest.

This trough, în didelphids, bouses the médial

opening of the prootic canal. In Aîayulestes the

prootic canal is absent. On the latéral sidc of the

righr periotic,. the sulcus for the prootic sinus is

cicarly visible because of the opening of the

squamosal-periotic .suture due to the crushing of

the skull. Peering ihrough ihe sinus, it is not

possible to observe any latéral foramen of the

prootic canal. Médial to the mcdial extremity of

the petrosal crest is ihc .secondary facial foramen

which opetis posteriorly and ihrough which is

passing the facial nerve.

Beyond the petrosal crest is a large (but .small for

a hothyacnoid) alisphenoid hypotympanic sinus,

k is made of three different boncs, the periotic,

the alisphenoid and the squamosal. lis posterior

half is excavated in the periotic. Jn Aîayulestes it

is approximarely half of the sinus whilc in

Didelphis that région of the periotic is totally

covered by the alisphenoid ventrally. The aritero-

latcral half ol ihc sinus is formed by a small strip

of ihe alisphenoid, posteromedially, and by the

posterior wall of the médial process of rhe squa¬

mosal, anterolaterally. The alisphenoid-squamosal

suture has a distinct sigmoid morphology. The

anterior border of rhe alisphenoid sinus is fbr-

med by the médial proces.s of the squamosal

which bears a conspicuou.s ridge (che anterior

ridge of the alisphenoid hypotympanic sinus)

which almost joins the médial border of the gle-

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

noid fossa to rhe latéral border of the foramen

ovale.

Anreriorly, between the fossa for the tensor tym¬

pan! muscle and its lacerai border, the periotic

forms rhe posterior border of the foramen ovale.

This condition is primitive for marsupials

(Muizon étal. 1997) and will be discussed below.

Ectotympanic. No ectotympanîc has bccn

found with the type 5pecimen of Aîayulestes. lii

other borhyaenoids, tlic ectotympanic articulâtes

with the squamosal bcnveen the médial cdgcs of

die postglenoid process and postglenoid foramen

and the mcdial process of the squamosal. The

alisphenoid also participâtes in the articulation

and small contact may exist with the posttympa-

nic process of the squamosal posteroventralJy. A

shallow groove and small ridges can be observed

in this area and évidences ihc articulation with

the ectotympanic. In Mayulestes this région of

the skull is damaged on both sides of the spéci¬

men. However. on rhe righr side of the skull, it is

possible to observe the contact between the

squamosal, médial lo the postglenoid foramen,

and the periotic, latéral to the latéral excrcmity of

rhe petrosal cresr and the periotic part of rhe

hypotympanic sinus. A fairly good reconstruc¬

tion of this région of rhe skull is- possible, and

apparentLy, rhete was no groove or ridge for rhe

articulation of the ectotympanic, although very

subtie undulatiuns of the squamosal iii this area

are observable with incident light. Thereforcj il

seems that the ectotympanic of Mayidlestes was

not tighcly imbricated with the squamosal as in

other borhyaenoids. This bone was probably

maincained m its position by ligaments only, but

perhaps apptessed againsr the squamosal in a

position that could represent an incipient deve¬

lopment of rhe condition obsen^ed in the other

borhyaenoids. This condition seems ro be inter¬

médiare bemeen that of DifJelphis - where the

ectotympanic is attached ro the alisphenoid by

ligaments only, and has no truc articulation with

the skull -- and that of other borhyaenoids.

Basioccîpital. (Figs 5Cj 6C, 7) Because of the

dorsoventral crushing of the skull, rhe bone has

been pushed dorsally within the braincase and

occupies a position dorsal to the periotics whe-

rein the inferior pevrosaJ sinus is exposed médial-

ly. The basioccipital is a short, broad, trapézoïdal

bone which contrasts with that of the other

borhyaenoids where it is gcncrally longer and

narrower. In this respect, it is more similar to

that observed in Didelphis. The basioccipital has

a broad, transversc suture with the basisphenoid

anteriorly. Posterolaterally, at its contact with die

période, the basioccipital is inflatcd, a morpho-

logy which indicates the passage of the inferior

petrosal sinus. Its articulation with the exoccipL-

tal is formed by two oblique line.s which join the

posteromedial angle of rhe promontoria ro the

central région of the ventral rim of the foramen

magnum which is formed by the basioccipital.

1 be bone has a small médian keel which starts at

the basisphenoid-basioccipital suture and which

affects rhe antetior two thirds of the bone. On

each side of this keeJ are two deep fossae for the

origin of the recrus capitis ventralis muscle.

Ëxoccipîtal. (Fig. 9) This bone contacts rhe

periotic medially and forms rhe posterodorsal

Wall of the inferior petrosal foramen, and the

mediai and posterîor rîni of die foramen lacerum

posterior (which transmics the nerves IX, X, XL,

and probably a small branch of the sigmoid

sinus, to die internai jugulai vein). Anteriorly it

fias a small contact with the pars petrosa of the

periotic by the septum between the inferior perro-

s'al and posterior lacerate foramiaa and laterally

ic contacts die pans mastoidea of the période.

On ihe ventral side of the exoccipital, anterior to

ilie ventral portion of the occipital condyle, are

iwo condyloid foramina which probably trans-

niicicd branches of the nerve XII (Jollie 1962)

but may also carry branches of the sigmoid sinus

to rhe internai jugulât vein (Archer 1976).

In posterior view, because of the crushing of the

skull, the foramen magnum is flattened dorso-

ventrally and the occipital condyles are Tolded"

and bnoken. Ir is however possible to note that

the exoccipiials had a large, convex suture with

rhe pars mastoidea of the periotic {Le. the latéral

edge of the exoccipical is convex). The suture

with die supraoccipital is slighdy convex (/.f. the

dorsal edge of the exoccipital is convex) and runs

from the dorsomedial région of the mastoid pro¬

cess to the dorsal rim of the foramen magnum.

The paroccipital process, if présent, was very

small. The occipital condyles are very large but

much less salient than in the other borhyaenoids.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 9. — Mayulestes ferox. skull in occipital view. A, hololype. {MHNC 1249) . B. reconstruction. Abbreviations: EO. exoccipital;

fm, foramen magnum; icn, intercondylar notch; JU, jugal; le. lambdoldal crest; mf. mastoid foramen; mp, mastoid process: doc, dor¬

sal occipital condyle; P(pm). pars masloidea of the periotic; pgf, posiglenold foramen; pgp, postglenoid process of the squamosal;

prgp, preglenold process of the jugal; SO. supraoccipital; SQ, squamosal. Scale bars. 1 cm.

In this respect thc)^ are more simîlar to those of

Didelphis, On the dorsal rirn of the foramen

magnum, the exoccipitals are broadly separated

as observêd in Pttcadelphys, concrary to the

condition of Didelphis, The condition in the

other borhyaenoids was difficult to evaluate as in

ail the specimens available during this study the

exoccipitals were tighily fused to the supraoccipi¬

tal.

Supraoccipital. (Fig. 9) This bone occupies the

dorsal central area of the occipital shield and

forms an important portion of chc dorsal rim of

the foramen magnum. The bone articulâtes ven-

trolaterally with the exoccipital, laterally with the

pars mastoidca of the periotic and dorsally with

the postparietal. Apparently, the suture with the

postparietal i,s located on rhe lambdoldal crest

itself. The surface of rhe supraoccipital is rclati-

vely smooth- However, chc posccriorly protru-

ding lambdoidal crest forms a very deep fossa for

the atcachmcnt of the nuchal muscles. Several

small foramina are présent, mainly ut the base of

the lambdoid crest.

Deiitciry (F'ig. 3)

l'he deiuûry of Mayulest^es is known by jts hori¬

zontal ramils and fragments of the condyle and

angular process. The proportions of the horizon¬

tal ramus compare favourably with Cladosictis. It

is relatively high compared to didclphids and

Sîpalocyon and is more slcndcr than in

Prothyldçynxis and Borhyaena. The anrerior part

of the ramus (below the incisors, c^anine and pl)

has a relatively srraight ventral border wiiich

makes an angle of approximately 45" with the

axis of the tooth row. Below p2 and p3 chc ven¬

tral border of the ramus is slightiy concave and

genriy convex below rhe tnolar.s. Thi-s morpho-

logy' is similar to that ob.servcd in CUdosictis cen-

tralis and Sipalocyort gracilis although less

empivasised in rhese species rhan in Mayulestes

ferox. Two large mental foramina occur below

the anterior roor of p2 and the middle of ml. A

much smaller foramen is locared below rhe pos-

terior roor of p2. McdiaÜy, the articulât surface

of the symphysis' is similar in shape ro rhat of

Sipalocyon and Cladosictis but slightiy shorter

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

anteroposteriorly. Its posterior extrcmity is situa-

ted below the posterior root o! p2 while in the

tormer tvjo généra ir is below the anterior root of

p3 (it is somctimcs below p2 in Sip/ilocypn). The

articular surface is less riigose rhan in Sîpalocyo7i

and Cladoüctis. As in diose généra, rhe ardcular

surface is roughly parallel to the plane of the

horizontal tamus below the molars while in

didelphids it is al an angle of approximately 20®.

The mylohyoid groove is well niarked and ends

anteriorly below mL

POSTCRANIAl. SKFXETON

Atlas (Fig. 10)

The atlas lacks the ventral arch {the intercen-

trum of rhe atlas) which was not fused to the rest

of the vertebra and was lost prior to fbssilizarion.

The dorsal atch is long anteroposteriorly in its

dorsomedial part (5.5 mm) and short ventrolatc-

rally, above tlie transverse process (3.2 mm). In

chis respect, the dorsal arch of the atlas of

Mayulestes rcscmblcs more chat of Borhyaena

rhan those of Cladoslctis, Sipalocyon and

ProrhyldcyuHs. This condition is also présent in

Pucadelphys (Marshall & Sigogneau-Russell

1995) but less marked. The dorsal arch of the

atlas of Mayîdèstès has srrongly convex aiitêrior

and posterior edges and the anterior is dearly

recurved vcntrally in its médian région. Scrongly

convex anterior and posterior borders of the dor¬

sal arch arc also found in Borhyaenn while it is

either straight or concave in Prothylacynus,

Cladosicthi Sipatacyony Dldelphts and Thylacinm,

In Borhyaenay howêver, rhe anterior border of rhe

arch is not recurved ventrally. The dorsal arch of

Mayulestes is wide transversally and the neural

canal is almost twice as widc than high as obser-

ved ir\ Didelphis. It differs in this respect from

rhe other borhyaenoitU wbere the dorsaJ arch is

shorter transversally and wherc the neural canal

is only slightly wider than high. This condition

is unknown in Pucadelphys. On the latéral extre-

inities of the anterior and posterior edges of riic

dorsal arch chcrc are deep grooves (the posterior

grooves are deeper than the anterior grooves) for

the passage ot rhe First cervical nerve and the ver¬

tébral artery anteriorly, and a ramification of the

latter posteriorly. The condition of Mayulestes is

different from that of ail the other borhyaenoids

for w'Iiidi the atlas is known. In Borhyaena the

anterior groove is présent bm the posterior is lâc-

king. In ProthylaiynuSy Cladosictis and Sipalocy'on,

tlie anterior sulcus is closed in an intervertébral

foramen (or atlanul foramen) and the posterior

is absent. The absence of the posterior sulcus in

the Santa Cruz borhyaenoids h probably a

conséquence of lhe présence of a smaJl trûnsverse

foramen, a structure absent in Mayulestes. An

intervcrtcbral foramen is also found in Didelphis

and l'hytacinus but is absent in Marmosûy

Perameles zuà Monodelphis. 'l ’he condition obser-

ved in Aîdytdestes is similar to chat observed in

Pucadelphysy a Palaeocene didelphoid which has

neither intervertébral nor transverse foramina

(Marshall Sigogneau-Russell 1995).

The transvetse proces.ses or wings of the atlas are

partially broken in the specimen described here,

but it is clear that rhey were snialler than those

of the other borhyaenoids. In Mayulestes the pro¬

cesses are strongly constricted at their bases

bccause of the depth of the artcrial grooves. On

the ventral border of the right transverse process

(which is better preserved than the Icft one), aç

its base, is an anteropusterior groove which wâs

transmitting the vertébral arcery posteriorly to

the axLs.

lire anterior articulât facct.s with the occipital

condylcs are more opened anteriorly {i.c. they

arc facing more anteriorly than medially) than in

the other borhyaenoids whcre they are faeJng

more medially than anteriorly. 'Fhis condition is

also found in Didelphis and Thylacinm. The pos-

terior articulai* facets with the axis are simpiy

reniform and similar to those of Didelphis. They

are rnuch less concave than the occipital facets.

Axis{F\^, 11 )

The most spectacular feature of rhe axis is the

size of its extremely long, triangular-shaped spi-

nous process* If is fairly similar to chose of

BorhyaenUj Prnthylatynus and Cladosictis but

clcarly differs from. the highiy specialLsed process

of Didelphis {Didelphis has cervical vertebrae 2 to

5 with transversely thickened spmaiis processes

which tend to synostose with eacb other). It is

much longer anteroposteriorly than in

Pucadelphys and any other didelphid. The dorsal

edge of the spinous process of the axis of

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 10. — Mayulestes ferox, holotype (MHNC 1249). Atlas: A. anterior: B. dorsal; C, posterior views, Scale bar: 5 mm

Mayulesres is very convex anteriorly and almost

straight in its posterior rwo thirds. The postero-

ventral edge is straight and oblique and forms an

angle close ro 35° with rhe dorsal edge. The

anteroventral edge is regularly concave and

passes lo the anterior border of the pcdicles of

the neural arch which contact the body of the

axis in a point more posterior chan in ocher

borhyaenoids. Anteroposteriorlys tlic pcdicles arc

proportionally shorter and the whok neural arch

is located in a niorc posterior position than in

Borhyaena. ProîbyLitjma and Cbulosiciis'. in dor¬

sal view die visible portion of ebe anterior part of

the body Imainly the portion conci;pouding to

the centrum of the arias) is larger, and in ventral

view^ the visible portion of the neural arch is aiso

larger. The posti) gapophysial facets are more

widely separated than in Borhyaena mà the neu¬

ral canal is proportionally müch larger. Ventrally

to the postzygapophyses. the cransverse processes

are divided into a large dorsal portion and a

small ventral ridge. As a consequenccj the traris-

verse foramen is opened laterally and the dorsal

transversc proces.ses overhang an elongated fossa

(transverse sulcus) for the passage of the vertébral

artery. As in Pzicadelphys, thcrc is no transversc

foramen, contrary to the condition of Borhyaena-,

Prothylacynusy CladosictiSi Didelphis and

Thylacmm,

The atlantal and axial componenr of rhe axis are

coossified and their suture Ls observable on the

ventral surface of the vertehra as an clevated

transverse ridge. The centrum of rhe vertebra

(composed of the centrum of rhe atlas, anterior-

ly, ftised to the centrum of the axis, posreriorly)

is wide and relatively shorter than m the other

borhyaenoids. Anteriorly, it bears a small odon-

toid process berween the rwo anterior articular

facets for rhe arias, ail rhree similar in propor¬

tions to tliosf of Borhyaena. C)n Uie vcnrral face

of the centrum are two deep fossac for the

attachment of the longus colli muscle, an impor¬

tant depressor of rhe head. The fossae occupy

the total width of the body pusteriorly and

slightly narrow anteriorly, whcrc the body is

wider becausc of the antcrioi articular tacefs {the

body of the atlas), ’l'hc apiccs ol rhe fossae are

roundcd and ainiosc reach the base of the odon-

toid process. The fossae for the longus colli are

more developed than in Borhyaena^ whcrc they

are friaiigular and narrow anteriorly^ and in

ProthyldcynuSy whcrc they do nor reach the hase

of the odontoid process, l'hc longüs colli fossae

are separated by a sharp médian cresr concave in

latéral view and forniing a small posterior

tuberclc projecting ventrally. The médian crest

and rubercle of Maytdestês are less developed and

projêct less ventrally than in other borhyaenoids,

which could indicate a slightly weaker muscula¬

ture. d’he fossae are bordered antcrolatcraliy by

t\vo crests for the attachment of the longus capi-

tis muscle, another depressor of the head. Since

therc is no transversc foramen in Mayulestesy

these crests do not extend laterally on the ventral

edge of the transversc process venrrally (therefore

ventrally ro rhe vertébral arterv), as it observed

in Borhyaena^ Prothylacynus, Cladosictisy and

ThyladniÀS. In Aîayulestes the origin of the longus

capitis was probably resrricted co the antérolaté¬

ral région of rhe ventral face of rhe centrum,

In posterior view, rhe centrum is much lower

than in BorhyaenUy ProthylacymiSy Cladosictis and

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

A B ;

Fig. 11. — Mayulestes ferox, hoiotype (MHNC 1249). Axis:

A, latéral; B, dorsal views. Scale bar. 5 mm.

Thylacinushm shows a condition simibr to that

oi Didelphis. The articulai surface for the third

cervical vertebra faces posterodorsally.

Other cervical vertehnie

Iwo oihcr partial cervical vertebrae are known

(the Pthird and the ?fifth). The ?third cervical

vertebra is only known by its centrum, wbJch is

very low and short like that of the axis. The

?fifth cervical vertebra is betccr prescrved but is

missing most of the neural arch. The centrum is

proponionally shorter ihan in the Santa Cru/,

borhyaenoids. lis ventral surface has a weakly

devcloped relief, being excavaied ventrally by

two shallow fossac for the longus colli muscles. Jt

clcarly differs from ihc cervical vcrrebrac of

Prothylacynus and Borhynena which beat a very

strong ventral crest wiih a large posrerior

rubercle (exrremely thick in YPM PU 15120).

Dorsally, the centrum présents two deep excava¬

tions probably related to the passage of the ven¬

tral spinal artery and its ramifications to the

vertébral artery, At the ba.se of the rranst'erse pro-

cess is a relaiivcly large transver.se foramen for

the passage of the vertébral artery. The pedicle of

the neural arch is short.

Thoracic vertebrae ( Fig. 12)

Five thoracic vertebrae are known, three anterior

and two posterior. Of the three anterior verte-

brac only one is complété. It is referable to Tl or

T2 since the prezygapophyses are facing dorso-

medially and the post/.ygapophyscs arc facing

ventrally (tangential). This condition is observed

on Tl in Caluromys, Didelpbis and Phalatiger,

but on T2 in Monorlelpim and Metachinis. Since

the firsr three généra are arboreal and the lasc

two are terrestrial and, given rhe interprétation

given hclüw of rhe mode ot life of Mayuleates (at

least pariially arboreal). it is more likely ibat tliis

vertebra represents the Tl. les neural arch is very

widc and ibc prezygapophyses are greatly .sépara-

ted. \ he articulât facets lace dorsoriicdially and

not dorsally as in the more posterior vertebrae.

They are locared in a latéral position, on the

anteromedial edge ol the transverse process and

not on the anterior edge of rhe neural arch. The

spinous process i.s long, narrow and less iaclined

posccriorly than in Borhyaena and P)^thylac)>mis.

Its anterior edge is straight as in Proîhylavynus

and Borbyaena, connary to the condition in

Cladosh'tis where it is concave anieriorly and

where rhe spinous process is recurved anterodor-

sally. d'he maior characreriscic of char vertebra is

the shortness of the pedicle of its neural arch,

which is shorter chan half of the Icngth of ihe

centrum, wdiile ii is longer in Borhynena,

ProthylacyuuSi Cladosicth and in didelphids. 'T'he

centrum is proportionaJIy shorter and lowei rhan

in Borhyaena, Prothylacynus, and Cladoùcîis

(although ro a lesser extent in this genus). The

ventral side of the centrum bcars a well marked

cresi which .séparâtes the insertions of the long!

dorsi musclç-s as in Cbulosictis, The condition of

Mnyidestes Irom that of Prothylacynus Anà

Borhyaena which show no crest on the ventral

side of the centrum of the anterior dorsal vcnc-

brae. Two other anterior thoracic vertebrae are

known by their centra only. Thcy show the same

characteristics as the vertebra described above,

The last tw-o thoracic vertebrae were found asso-

ciated with the first five lumbar vertebrae.

Contrary to what is observed on the ccrvic.!! and

first thoracic vertebrae, rhe last thoracic is rclari-

vely much longet than in Prothylaeynus ^t\d

Cladoskfis, On the neural arch, ihe prezygapo¬

physes are oriented Icss dorsally than in

Prothybuymis Md rhe articulàr surface of the pre-

zygapophyseÿ occupies ail its médial surface,

whilc in Pi'ùThylacymis they are overhung by a

strong metapophysis. In latéral view the postzy-

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 12. — Mayulestes ferox, hololype (MHNC 1249). Anterior

thoraclc verlebra (T?1). A, latéral: B, dorsal views: T?12;

C. latéral; D. dorsal views: T'> 13-L1: E. latéral, F. dorsal views.

Scale bar: 5 mm.

gapophyses are longer chan high, contrary to

those of Prothylacynus which arc higher rhan

long. The arucrior edge of rhe neural sptne is

concave aiitcrodorsally, ïrs posterior edge Lv

straight and the .spine occupics the posterior two

thirds of the neural arch; rhe neural spine is che-

refore slightly inciined posterioiiy and rhe last

rwo thoracic verrebrae of Mayulesies are anrerior

to rhe anticlinal verrebra. This condition differs

from that of Prothylacynus rhe neural .spine

of the la.si two thoracic vertebrae is oriented

anteriorly (the vertebrae are posterior to the anti¬

clinal vcricbra) and has a straight anterior edge

and a concave po.sterior edge. In faci, in

Prothylacynus^ the posterior part of the neural

arch (/.£'. the neural spine and the postzygapo-

physes) seems to hâve been pulled anteriorly

when compared to the condition observed in

Mnyulestes, l he anapophyscs arc proportionally

longer than in Prothylacynus and bear, on their

latéral edge, a marked ridge for the insertion of

the longi.ssimu.s dorsi and sacrocaudalis dorsalis

muscles. This lidge Is wcH-devcloped on rhe ele-

venth thoracic verrebra of Prothylacynus

(YPM PU l')7()0). The centrtim is relatively

lowcr dorsovcntrally chan in Prothylacynus its

ventral surface if less rounded.

Lumbür vertebrae 12. 13)

Five lumbar vertebrae are known and, if one

assumes that Mayulestes WaA six lumbar vertebrae

as Pucadelphys, Üving didelpbid.s and Cladosictis^

ir is rhe last lumbar which is missing since the

five vertebrae were tound assuciared. Like the

thoracic vertebrae, ihey are relatively longer than

in Prothylacynus and Cladosictis. The mosc

remarkable fearurc of the lumbar vertebrae of

Mayulestes i.s the shapc and orientation of the

neural spine. On the first rhrcc vertebrae* they

are small, low, long (antcroposteriorly)» oriented

posrcriorly and ilie>^ occupy rhe posterior rwo

chird of rhe neural arch (/.e. their anterior edge is

concave anrerodorsally). The spinal process is

not preserved in the fourih lumbar. ’î’hc fifth

lumbar has a .short (anreroposteriorly) but very

high neural spine which is .slightly oriented ante-

rîorly and occupics the total Icngth of the neural

arch, "l'his vertebra is very similar to lhat of

I^tcadelphys. Mayukstes and Pucadclphys difter in

the morphology and orientation of ihe neural

spine from the morphology observed in the

other didelpbids, whcre the spine is generally

low and oriented posteriorly or, if anteriorly, to a

very sliglit extent. In the sb; lumbar vertebrae of

CLtdüdctis (YPM PU 15170), the neural spines

are high, short (anreroposteriorly) and scrongly

inciined anteriorly, and their posterior edge is

concave posterodorsally. The same condition is

found in the fourih, fifth and sixth lumbar of

Prothylacynus, although in this genus the neural

spine is strnnger, longer (anreroposteriorly) and

less inciined anteriorly than in Cladosictis, On

the holocype of Mayulestes, che transverse pro-

ces5C5 are intact on the fifth lumbar only They

arc more slcnder than those of Prothylacynus and

Cladosictis (to a Icsser extent in thU genus) but

resemble those of Pucadelphys. They are propor-

GEODIVERSITAS • 1998 • 20(1)

Mnyulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 13. — Mayuiestes ferox, holotype (MHNC 1249). Lumbar

vertebrae; L2: A. latéral; B, dorsal views; L3: C, latéral; D, dor¬

sal views; L4. E. latéral: F, dorsal views: L5: G. latéral; H, dorsal

views. Scale bar 5 mm.

tionally longer (transversally) than ihose of

Prothylacynus (which are almost complété on rhe

fifth and sixth lumbar of’^TM PU 13700) and

more recurved venrrally. Their bases occupy

approximarely one third of the length of the cen-

triim while they occupy half of its length in

Prothylacynus and Cladosictis. The /ygapophy.ses

of rhe fitih lumbar vcrtebra are more evertcd

than in Prothylacynus and (orm a conspicuous

X-shape Figure in dorsal view as obscrvcd in

Cladosictis. As in Prothylacynus and Cladosictis,

the anapophyse.s reduce from the flrst fo the

third lumbar vertebra; sincc they are small but

présent in rhe third and absent on rhe fifth they

were either very reduccd or absent on the foiirth

lumbar. The centra of the lumbar vertebrae are

relatively slighrly lower than those of

Prothylacynus but match in this respect the

condition obscrvcd in Cladosictis.

Sacral vertebrae

No sacrai vertebrae arc preserved in the holonpe.

Caudal vertebrae l4)

Four vvell preserved caudal vertebrae arc known,

two aiuerior and tvvo posterior. The anterior

caudal vertebrae are very similar to those of

Cladosictis, Borhyaena and Prothylacynus. The

slight diflercnces obscrvcd could be rdated lo

their position in the taîl (which cannoi be dcfi-

ned wiih précision — pos-sibly Cl and C3) rather

than CO the actual morphology of rhe taxon.

When compnrcd to Caluromys, the morphology

of the TWO posterior caudal vertebrae corres¬

ponds to C7 and C8. When compared to

Pucadelphys, the vertebrae correspond ro C8 and

C9 and when compared lo Cladosictis and

ProihyltTcynus, tlicv scem lo correspond lo Cl 1

and Cl2. They are proportionally more slcnder

and longer than in (Jladosictis and Prothylacynus.

The ventral side of the centrum of borh verte¬

brae diifers from that ol Didetphh, Caluromys

and, to a lesser extern, Metachirus, which bear.s a

ventral sulcus lor the ventral médian coccygeal

arrery. On the edges oF this sulcus attaches the

sacroccygeus venrralis medialis muscle, a flexor

of chc tail [tbi.s sulcus doe.s not rcccivc rhe

abductor muscles as stated by Marshall &

SigogneaU'Russell (1993), a rcrm which is inap-

propriate in the case of a flexion of the tail]. In

Mayulestesy the sulcus is présent but very wcak; it

is observable in rhe anterior and posterior third

of each of che rwo posterior caudal vertebrae

while, in che médian région of che centrum, ic is

reduced to a fiat strip. A similar condition rs

observed on C5 to C8 of Pucadelpf/ys andinus

(.specimen YPFB Pal 6106) and Mnnodelphis.

ThLs feaUire .shows some iiidividual variation

sincc the sulcus is more pronouneed m auother

specimen of Pucadelphys (YPFB Pal 6110) whose

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 14. — Mayulestes ferox, hololype (MHNC 1249). Caudal vertebrae. C?1: A, latéral; B, dorsal views; C?3: C, latéral; D. dorsal

views; C?8; E, dorsal view; C?9: F, dorsal view. Scale bar: 5 mm.

C7 and C8 hâve been further prepared. No sul-

cus is observed in Marmosa, The anterior and

posterior tran.sversc processes of thc posccrior

caudal of Mtiyiileites arc similar in shape and

relative size to chose of Caluromys and Puca-

delphys They are slightly larger and more

blade-Iike rhan chose of Didclphis, Metachirus,

Monodelphis and Murmosn whcrc the proce.sscs

are more knob-like. l*he morphology of thc pos¬

terior caudal vertebrae of Mayidesles dénotés an

important strengrh of thc tail musculature that

could bc related to prcliensility (see below for

discussion).

Ribs (Fig. 15)

Four ribs of the holotype of Maylulestes are

known. Two are probably the first or second

right and left ribs. They hâve a smaller tubercu-

lum and a capitulum wider and flatter at its base

than in Borhyaena. One of the other two ribs

(R8?) is a relatively anterior rib since its curvatu-

re is well pronounccd. Il has à relatively small

tuberculum and a long capitulum. It i.s rcgularly

curved and docs not show an angulation bet-

ween the tuberculum + capitulum région and

the test of the bone, as ît îs observed in

Borhyaena and CladosictU. The other rib (RIO or

11) is posterior and has a weaker curvature. Both

ribs were flatter rhan in Borhyaena and

Cladosictis.

Forelimb

Scapula. (Figs 16, 17) In the following descrip¬

tion, thc spine of thc scapula will be oriented

verricaliy.

The general shape of thc scapula is triangular

and significantly differs from that of the other

Fig. 15. — Mayulestes ferox, holotype (MHNC 1249). Ribs:

A, right RI or 2 in anterior view; B. left RI or 2 In anterior view;

C, left R?8 in posterior view; D, right RIO or 11 in posterior view.

Scale bar: 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

borhyaenoids, Metachïrm^ Phtlander and

Didelphis which is roughly oval-shaped or qua-

drangular; ic shows important .sirnilarities witli

chose of Caluromys and Monodelphis. In

Maytdcstes, the antcrior edgc of tbc scapula (Le.

of the supraspinatus fossa) présents a marked

angulation (almost a righr angle) when it is regu-

larly convex in oiher borhyaenoids. The ante-

riorly protruding supraspinatus fossa of

Mayulestes is very similar to chose of Monodelphis

and Caluromys (although a lîttle more rounded

in this genus). The posierior border of rhe sca¬

pula (;.e. of the infraspinatus fossa) is slighriy

concave as in Caltp'otnys and is oriented postero-

dorsally, whcreas it is convex in the other

borhyaenoids and in Didelphis^ and straight in

Pucadelphys.

The supraspinatus fossa is much wider antero-

po.stcriorly in its middle région than the Lnfraspi-

natus fossa but is slightly shorrer proximodiscally,

Its surface is a little superior to that of the infras¬

pinatus fossa as observed in Caluromys. Both fos-

Fig. 16. — Mayulestes ferox, holotype (MHNC 1249). Right scapula: A, latéral, B, médial, C, proximal views. Scale bar: A, B, 1 cm;

C, 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 17. — Mayulestes ferox. holotype (MHNC1249) Right scapula: A. latéral; B. médial; C. proximal views. Abbreviations: ac, acro-

mion; cp, coracoid process; gc. glenoid cavity; hp, hamatus process; ff, infraspinatus fossa; n. necK; sf. supraspinatus fossa;

shp, suprahamatus process; sp, spine; sscf, subscapuiar fossa; si, supraglenoid tubercie Scale bar: 1 cm.

sae are \cry narrow in their proximal pan and

less cxtended antemposteriorly ihan in ihe oihcr

borhyaenoid& and Didclphis. TKe supravSpinacus

fossa is much narrowcr in its proximal portion

than in rhc Santa Cruz borhyaenoids where irs

anteroproximal border strongly protrudes anre-

riorly, The morphology o! ihe scaptda o[ che

Santa Cru/, borhyaenoids is ai.so présent to a Ics-

ser extent in Didelpbis whilc Pucndelphy^

resemble^s Mayulestes in rhis respect. The supras-

pinatus fossa is relarivcly shallow while the

infraspinatus fossa is very tieep, with its posterior

border strongly defleefed laierally, a condition

common in the other borhyaenoids whose sca¬

pula is known.

The infraspinatus fossa is narrow and wîdens

moderateiy toward its distal end contrary to the

condition oi Pi’otbyLtcynus and Borhyaena. When

the spine is oriented verrically, the highest point

of the scapula is located almost at the postero-

dorsal angle of the infraspinatus fossa as in

Caluromys, whiJe in Cludosictis^ Didelpbis,

Metachiras zwà Marmosa it is located ar the dor¬

sal extremis' of the spine or at the same levcl as

the angle. The posterodor.sal angle of the scapula

receives the origin of the tercs major posteriorly

and the insertion ol the rhomboldeus amcriorly.

As in Cabirtfmys the spine is more elevaied rhan

in the other didelphids. It is sÜghtly deflected

posteriorly and its latéral edge is convex in ante-

rior view. The acromion is large and forms a tri-

angular plate which hears a long anterior process,

the hamatus process, and a short posterior pro-

cess^ the suprahamatus process (the paracro-

mion), located heliind rhe anterior proce.ss. Since

it is very ch in and fragile, ît Js generally lost in

fossils; it is broken in ail the borhyaenoids speci-

men.s dcscribcd by Sinclair (I906). In anterior

view the acromion of Mayulestes is flar and the

apex of the hamatus process is slightly benr

medially. It is similar to that of Caluromys but it

differ.s from that of Didelphis where the hamatus

process is shorter and less individualised. In

Piicadetphys^ ihe acromion aiso has a long hama¬

tus process but the wholc structure is smailcr

than in Mayulestes. In mcdial view, the ventral

extremity of the acromion is visible helow the

glenoid fossa as in Pucadelphys falthough rhis is

noi shown on figure 36 of Marshall &

Sigogneau-Russell (1995), it is clear in ihc fur-

iher prepared right scapula of YPFB F*al 6106

(Fig. 18)], Manmna and Caluromys (ro a lesser

extenr), contrary ro Didelphis where the acro¬

mion does nor extend venrrally helow ihe glc-

noid fossa. In latéral view, che hamatus process

overhangs anieriorly ihe supraglenoid tuberosity

and, in proximal view, the anterior apex of the

hamatus process is slightly anterior to the cora-

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 18. — Pucadelphys andinus (YPFB Pal 6106). Right sca-

pula: A. latéral; B, medial; C, proximal views. Scale bar: 5 mm

coid process. This condition is similar to that

observed in Cahmfmp wh\\^ in thc othcr didcl-

phids the iinterior apex of the hamatus proccs.s is

posterior ro the antcrinc cdgc of the supraglenoid

ruberosity. The condition in Pucadelphys is close

to that observed in Mayulestes but thc acromion

is slightly Icss anterior than in Mayulestes and

apparently relativeJy slighdy larger (b'ig. 18). The

latéral side of the acromion bears part of* the ori-

gin of the deltoideus muscle and, on ihe anterior

border of the acromion, h inserted the atlantoa-

cromialis muscle (= omotransversariiis muscle

pro parte).

The neck is the portion of the scapula located

above the glenoid fbssa at the level of the supra-

coracoid incisure. also called the scapular notch

(which marks thc ventral extrcmicy of the supras-

pinatus fossa). le is long, narrow and well mar-

ked as in Calurotnys Pucttdelphys, contrary to

what is observed in the otlier borhyaenoids and

Didelphis.

The head is small and bears an anteroposteriorly

elongafed glenoid cavity. Tr is more flattened

transversally tlian in Cladosii’tù, Borhyamay

Prothylacymis and Thylaiinus, but less than in

Didelphis. The coracoid process is a small apo-

physis uf bone strongly rccurvcd medially, wherc

the coracobrachialis muscle originates. *rhç cora¬

coid process of Mayulestes is longer, more slender

and more rccurvcd meclially than in Borbyaeruu

ProtbylacynitSy Cladosictisy Pucadelphys and

Didelphis. As in Cladosictis, it is not wcll separa-

ted from the supraglenoid tubercle, contrary to

what is observed in Thylacinus and, to a Icsscr

extent, in Didelphis. As in Caluromys-, Metachirus

and Marnima^ the coracoid f.micess extends ven-

trally fiirther chân thc supraglenoid tubercle,

contrary to the condition observed in Pucadcl-

phys and Didelphis |Marshall 6l Sigogneau-

Russell (l'995i fïg. 36) hâve confused rhe

coracoid process and the supraglenoid tubercle

or tuber scapulae]. The supraglenoid tubercle

bears thc origin of ihe biceps brachii muscle, a

powcrful flexor of thc dbow. On the medial side

of the scapula, just dorsal to the posteromedial

extrcmicy of thc glenoid cavicy, is a shallow fussa

for the origin of the caput longurn of the iricep.s

brachii muscle. ‘Fhat muscle atcachment is relâü-

vely weak when compared to Didelphis oi Thyla-

cirnis. However, the weakness of this attachmenr

scems to hc cornmon in borhyaenoids sincc it is

aLso litrie developed in Borhyaeruu Prothylacynus

and Cladosictis.

On t he medial side of thc scapula, thc subscapu-

laris fossa reflects the relief of the lacerai side of

die boue. U is strongly convex in its posterior

région corresponding to the infraspinatus fossa

while rhe relief is moderate in the anterior

région. In the dog {Evans ôc Christensen 1979),

thc subscapularis fo.ssa receives the insertion of

thc serratus vcntralis muscle in its pmximal rhird

and the origin of thc subscapularis muscle in its

two distal thirds. The subscapulari.s is inserted

on the lesser tubercle of thc humérus and the

serratus vcntralis originates on the transverse

processes of thc last fivc cervical vertebrae. In

Didelphis^ thc latter arises from the cransverse

proce.sses of the last four or fivc cervical vertebrae

(Jenkins & Wcijs 1979).

Humeras. (Figs 19, 20) Both humeri of the

holocype of Mayulestes are known. They show a

significani difFerence in their length and propor¬

tions of their proximal epiphysis, alrhough no

pathological deformation can be observed. This

variation is probably the conséquence of a slight

post-mortem deformation of rhe bones sincc thc

right humérus lias been slightly compressed

anteroposteriorly in its proximal half, squeezing

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fiq 19 _ Mayulestes ferox, holotype (MHNC 1249). Right humérus: A, anterior; B, posterior; C, latéral; D, médial; E, proximal

views. Scale bar: A-D, 1 cm; E, 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from che Palaeocene of Bolivia

Fig. 20. — MayuiBStes ferox, holotype (MHNC 1249), Right humérus: A, antenor; B, latéral; C. posteriori D, médial views.

Abbreviations bicipital groove; c, capitulum; dpc, deltopectoral crest; dt. deltoid tuberosity; ec, epîcondyloid crest; ef, entepicon-

dylar foramen; gt. greater luberde; h, headi le. latéral épicondyle; II. lesser tubercte; me, medial epicondyle; of. olecranon fossa; rf,

radial fossa; t, Irochlea; tl. Iricipital line; ttm, tuberosity for the teres major. Scale bar: 1 cm.

the epiphysis and the bicipital groove. The left

humérus, which seems to hâve suffercd very little

(if any) deformation, will be referred to in the

following description and comparison. The bone

will be described vcrtically.

In proximal view, the proximal epiphysis is rela-

tively shorter anteroposteriorly than in the gene-

ra uscd for comparison in this study (othcr

borhyaenoids, didelphid.s and Thylacimis). The

condylc tor articulation with the scapula (the

head) is sÜghtly wider than long in Mayulestes as

in Pucadelphys (Figs 21, 22)j while in

Prothylacynus it is as long as wide. In Caluromysy

Chironectes-, Didelphîs^ Monodelphis and

Murmosa it is slightiy longer than wide, and in

Metachims il is rnuch longer than wide. In latéral

view, the condyle has a posteroproximal orienta¬

tion. However, the proximal component of rhis

orientation is more important than in

Prothylacynus^ Cladosictis and DidelphrSi whose

condyle seems to be more “recurved'' posteriorly.

This is especiaJly clear in the shapc of the poste-

rior side of the shafe jiist below che condylc,

which is strongly recurved and directed discally

in Prothylacynusy Didelphis and ThylacinuSy while

it is oriented posteroproximally in Mayulestes.

This condition is aiso clearly observable in poste-

rior view of the humérus where che visible por¬

tion of the condyle is much more reduced in

Mayulestes and Caluromys than in the other

borhyaenoids and didelphids. In fact, the orien¬

tation of the condyle of Mayulestes resembles

more thar of Caluromys than any other genus.

The greater tubercle is slightiy lower than the

condjde as in Didelphis and Caluromys^ while in

Prothylacynus it is slightiy hîgher and in

Tbylacînus it is much hîgber. In proximal view

the greater tubercle is elongated, more than

nvice longer titan wide and obliquely oriented,

In rhese respects, it is similar to those of

Pivthylaiynus, Caluromys and DidelphiSy bue dif-

fers from that of Thylacinus which is very thick

and massive. The greater tubercle of Mayulestes

is, however, rnore oblique and les.s salient ante-

riorly than those ol Prothylacynus and Thylacinus

where it greatly extends anteriorly, well beyond

the le.sser tubercle. The lesser tubercle is rounded

and separated from the condyle by a sulcus. It is

relaiively salient anteriorly and laterally and

almost rcaches the level of the greater tubercle

anteriorly. It is fairly similar to those of

PucadelphySy Didelphis and Caluromys but clearly

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

differs from chat ot Prothylacynus vvhich is elon-

gatcd, obliqudy orientcd. appresscd agaiast thc

condylc, and which remains tar bchind the grea-

ter trochanter anteriorly. These conditions of

Prothylacyimis arc observcd in rhe dog, a terrcs'

trial cursnrial carnivuran.

The bicipital gfoovc is wcll marked and faces

anteromcdially. It e\icnds on thc antcrior side of

thc shaft of thc bone where il is deeper than in

ProthyUcynus. CLtlimrnys xwA Didelfhh. Latéral ly,

below the greater tubercle, on the proximal 40%

of the Icngth of the shafi, ihcrc is a sharp and

elcvated tricipital Une. It is more devclopcd and

more salient laterally than in Prothylacynm and

Didelphis, where il is very wcak, bui rescmbles

that of Pucadelphys is \cry prominenr.

On thc anterior sidc of the proximal two thirds

of the shaft, rhe dcitopectoral crest ru ns distally

from ihe anreromcdial angle of the greater

tubercle on approximaiely 60% oi tlie Icngth of

the bonc. ‘ITic dcltopectoral crest is rclativcly

narrow, salient and slightiy concave laterally, It

resembles chose of Diddplm and Calurowys but

differs from thaï of Pucadetphys wblch is shorter,

and from thar of ProthyUuynus which is much

thicker and scraigho the latter featutes are eyen

more pronouticed in Thylnànm. l’he deltopecto-

ral surface, limited by thc deltt)pectoral crest

anteriorly and the tricipital line posteriorly, is

relative!)' wide and concave and faces more ante¬

riorly than laterally. In thèse respects it resembles

that of Pucadclphys and living didclphids but dif¬

fers froni those of Prathyheynus and Thylaiinus

which arc narrow, fiat or convex and face more

laterally than anteriorly. On thc antcrior border

of thc tricipital Une, on its proximal half, is a

very salient crest oriented laterally and which

receives the origin of the caput lateralis of the tri¬

ceps brachii. This' tricipital crest approaches the

condition observed in Pucndelphys and

MonodelphiSy alrJiough it is slightiy Icss salient in

thèse gênera. It differs from those of

Prothylntynus^ Didelphh and Thylm inus which

arc much weaker.

On the médial side of ihe proximal ihird of thc

diapbysis, below che Icsser aihcrclc, is a telacively

weak tuberosicy for rhe teres major muscle as in

Didelphh, and the fossa posterodistal to rhe lesser

tubercle is sinall and shaJlow. Thaï negion of the

humérus of Mayulestes strongly difters from that

of Prothylacynus where che tuberosiiy for the

teres major is very large and salient, and forms a

sharp crest which runs proximally rowards the

lesser tubercle and forms a deep fossa located

distally to rhe postérolatéral angle of the lesser

rubercle and rhe posteromeclial side of the

condyle. Laterally, ihe fossa is limited hy a wide

and rounded ridge which runs distally (rom the

posterior edge of thc condyle alorig the proximal

half of thc poNlerior sidc of thc dtaphysis. On thc

proximal extremity of the fossa and rhe ridge is

loc'ated the origin of thc actcsstiry head of the

triceps whitli was very strong in Prothylacynus.

Sinçe in Aitiyulestei thc relief ol ihc attachments

oi ihis muscle was interniediatc between those of

Didelphis and Calîtromys on che one hand and

Prothylacynus on the other hand, k is likely that

the muscle was stronger than in ihe former but

much wcaker than in the latter.

The distal extremity of the humérus is markedly

flattened anteroposteriorly and plate-likc as in ail

the burhyaeiioids and didclphids. The cpicondy-

loid crest is very large and extends along about

30% of thc diaph)^si5: its proximal extremity

reaches a point, on the shaft, latéral to rhe distal

extremity of the dcltopectoral crest as in

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Prothylacymis. 'fliis point of thc shaft is located

more proximally chan in Prothylacyntis. In

Clarivsicth^ the distal extrcmiry of ihe deltopecto-

nil crest is in a Far more distal position than in

Pwthylacyyius ■â.wil Mayult'Sîes reaches the dis¬

tal quarter of rite bone; in ihis genus, thc proxi¬

mal extrcmity of the cpicondyloid crcst is located

more proximally than thc distal extrcmity of the

deltopecioral cresi, In Mayulestes thc proximal

extrcmity of the cpicondyloid crest has a rcgular

concavoconvex contact wich the diaphysis as in

Claclosictis, Metachlrus and Marniostr and is not

rccurvcd anteriorly; it differs from the condition

of Prothylacyntis, Pucaddphys-, Didelphis and

Caluromys whose cpicondyloid crest is limited

proximally by a marked notch and recurved

anteriorly in that région.

On the médial sidc of the humérus, the entepi-

condyle is very long and strongly projccted dis-

tomcdially, more than in Protbylacynus,

Cladosictis much more than in Didelphis ^rxà

Caluromys. It resembles the condition observed

in Pucadelphys. It almost reaches the level of the

médial crest of the trochlea disfally (Fig, 22) as

in PucadclphySy contrary ro the condition obser¬

ved in Protbylacynus and most Uving didclphids.

lt,s apex, whcrc thc flcxors of thc carpus and

digit-s originale, is rounded and subcircular as in

Prothylacytim and Pucadelphys and not clongated

and oval-shaped (in prnximodiscal view) as in

Didelphis and Caluromys. l’hc ridge ol bone that

uTiiis the distal extremiiy ol thc deliopectoral

crest with the apex of thc ctucpicondylc and

passes above the entepicondylar foramen is more

salienr anteriorly than in Didelphis, Caluromys^

Prothylacynus and CladosictLi, which contribute.s

to give a more nvisted aspect to that part of the

humérus of Mayulestes than in the généra cited

above. Mayulestes resembles Pucadelphys in this

respect. On thc anterior side of the bone, just

proximal to thc capiculum, is a radial fossa dee-

per rhan in Prothylacynus, Cladosictis, Didelphis

and Caluromys.

The entepicondylar foramen is large and oval-

shaped. It is proportionally similar in size and

shape to chose of Proihylacynus, Cladosictis,

Didelphis and Caluromys. The groove Ibr the pas¬

sage of the médian arcery (the internai articular

sulcus of Osgood 1921), located between the

Fig. 22. — Pucadelphys andinus (YPFB Pal 6106). Lefl humé¬

rus: A. anterior; B, posterior; C, médial; D, latéral, E, proximal

views. Scale bar: 5 mm.

médial border ol the trochlea and the latéral

edge of the épicondyle, is deeper than in

Didelphis, Caluromys and Cladosictis but

resembles that of Prothylacynus. On the posterior

sidc of the bone, the olecranon fossa is much

shallower than in Prothylacynus, Cladosictis and

Didelphis but a litrle deeper rhan in Caluromys

wherc it is very shallow. In posterior view, the

trochlea has relacively well marked crests as

observed in Didelphis, but not as sharp and

salient as in Prothylacynus. As in Didelphis, the

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 23. — Mayulestes ferox, hoiotype (MHNC 1249). Left uina; A, anterior; B, latéral; C, médial views. Scale bar: 1 cm.

trochlea of Mayulestes is deeper than in

Prothylacynus; ic grcady differs from the trochlea

of Caluromys whicii is wider, deeper and has

lower crests dian in Mayulestes. The capituium is

more salieni and convex than in Prothylacynus

and Cladosiçîisy less than in Caluromys and

PucadelphySy but resembles that of Didelphis. In

Prothylacynus the capituium is less elongated

transversely, longer proximodistaDy and, as a

whole, more roundcd. The distal articular surfa¬

ce of Pucadelphys differs from that of Mayulestes

in its slightly narrower trochlea with a lower

medial crest and its wider capituium,

Ulna. (Figs 23, 24) The bone is notably short, as

are those of Borhyaena, Prothylacynus and

Cladosictis, and differs fronr] the longer ulnae of

Thylacinus-, Didelphis and Caluromys. The proxi¬

mal half of the bonc is markecUy recurved ante-

riorly as in Pucadelphys (Fig. 25). Ir strongly

differs from those of Prothylacynus and

Cladosicîis^ which hâve a globally straight ulna

with a poscerior border concave in its middle

third, and from those of Borhyaena and Thyla-

cinus, which are bent posteriorly. It resembles the

condition observed in Caluromys but differs from

that of Didelphis the proximal extremity is

only slightly bent anteriorly.

In anterior view, the olecranon and the articular

area are markedly deflected medially as it is

observed, to a lesser extern, in Borhyaena, Puca¬

delphys and Didelphis but contrary to Prothy¬

lacynus, Cladosictis, Thylacinus mA Caluromys. In

latéral view, the posterior and proximal edges of

the olecranon of Mayulestes form an angle of

more than 120°. Ic resembles that of Caluromys

wherc the angle is smaller but clcarly greater

than 90°. It differs from those of Borhyaena,

Prothylacynus, Cladosictis, Thylacinus, Didelphis,

Chironectes^ Metachirus and Caluromys which

almost form a right angle, sometimes a little less

than 90“ as in Borhyaena, or a little more, as in

Didelphis. The condition of Marmosa and

Piuradelphys is close to that of Mayulestes. In ante¬

rior view, the olecranon is long as observed in

Borhyaena, Prothylacynus^ Cladosictis and

Thylaetnus. It is slightly longer than in Didelphis,

Caluromys and Chironectes. The medial side of

the olecranon and proximal half of the shaft bear

GEODIVERSITAS • 1998 • 20(1)

Mayulestesy a borhyaenoid from the Palaeocene of Bolivia

Fig. 24. — Mayuiestes terox, holotype (MHNC 1249). Left uina: A, médial; B, anterior; C, latéral views. Abbreviations: bo, beak of

the olecranon; brf. brachlalisfossa; cop. coronold process; ccl, crest for attachment of the posterior and transverse part of the ulnar

collateral ligament; fap, fossa for the abduclor pollicis longus and extensor indicius proplus: Wl. fossa for the flexor digitorum profun-

dus; ioc, interosseous crest; o. olecranon; prr. pronator ridge: rdn. radial notch; sp, stytoid process; spc, supmator crest; spf, supi-

nator fossa: stecu, sulcus for the tendon of the extensor carpii ulnarls; trr>. trochlear notch. Scale bar: 1 cm.

a very deep fossa for the flexor carpii ulnaris and

flexor digitorum profundus. That fossa is much

deeper than in any of the three borhyaenoid

compared here to Mayulestes (although Prothy-

lacynus has a deeper fossa than Cladosictu and

Borhyaena). The fossa of Mayulestes is also deeper

than in Dldelphis but approaches the condition

observed in Calutomys and Mamiosa. The flexor

fossa is limited posteriorly and proximally by an

extremely strong crest for the atrachrnent of the

posterior and transverse parts of the ulnar colla¬

teral ligament.

In médial view, the trochlear notch (= greater

sigmoid caviry = articulât surface for the humé¬

rus) is relatively wide open (in médial view) and

shallow as in Caluromys (although to a greater

extent than in this latter genus). It differs from

Borhyaena^ Prolhylacyaus, Thylacinus, Cladosictis^

DidelphiSy Chironectes and MetcLchinis, where the

trochlear notch is very deep, concave and less

open (although this is less marked in Cladosictis),

but strongly resembles the condition observed in

Caluromys. In anterior view, the beak of the ole¬

cranon (the proximal crest of the trochlear

notch) is less salient than in the other borhyae-

noids but more than in Caluromys\ it approaches

the condition of Didelphis. Its width is greater in

Mayulestes than in didelphids but resembles in

this respect the condition observed in other

borhyaenoids. The coronoid process, the médial

extension of the distomedial portion of the tro¬

chlear notch, is large and oriented mediodistally.

It differs from that of the Santa Cruz borhyae¬

noids which is stronger, longei and protrudes

mediaUy. Among didelphids. it resembles that of

Caluromys in this respect and differs from those

of Pucadelphys, Dldelphis and Metachirus. The

radial notch (= lesser sigmoid caviry = articulât

surface for the radius) is located distal and latéral

to the trochlear notch (articulât surface for the

humérus) and faces anterolaterally. The line of

contact between the humerai and radial surfaces

has an anteromedial-posterolateral orientation. It

is a low ridge anteromedially and it is fiat and

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

continuous posterulatcrally. In other words, the

radial and humerai facers hâve an angular

contact antcromcdially and a Hat contact poste-

rolaterally. In the oiher borhyacnoids and in

ThylacuniSy the radial notch i.s aiways separated

from the rrochlear notch by a sharp crest anccro-

medially and by a sinall ridge postérolatérally

{Le. the angle between the rwo articulât surfaces

is more pronounced). In Prothylacynus, Borhy-

aenaznà Thylacinus xht angle between the ante-

romedial contact of buih articulât surfaces (in

anterior view) is cqual lo or smaller ihan 90*^; it

is more than 120^ in Muyulesies. In the

Didelphidac ihc angle varies around 90" but can

be close to 120" in some specimens of Didelphu

or more in Caluwmy^. The radial notch is shal-

low and does not excavate the aiiicTolareral face

of the uina as in tlic Santa Cru/, borhyaenoids

and ThyLuirms. It is roughly ctiangular and not

divided into two portions by a proximal

inflexion of its distal border as obsciwed in the

other borhyacnoids and Tbvlaeînih. In Alayu-

lestes and Ciilurornyu this articulât surface is

almosr parallel to the diaphysis, but, conrrary to

what is observed in Mayulestes^ the médial part

of the radial notch in CAilurQYnysx's not in contact

with the trochlear notch antcromcdially.

Distal ro the coronoid process on the antcromc-

dial side of the iiina iç a well-developcd latcrally

oriented fossa, for rhe insertion of the brachialis

and biceps muscles, h differs Irorn tliat of didcl-

phids whicb is slightK' larger and faces antcrinrly.

On the anterior face of the diaphysis, latcrally, is

a well marked supinator crest wliich runs distally

from the latéral edge of the radial articulation.

Between this crest and the latéral border of the

brachialis fossa is an elongated depressed area

facing anterolarerally. Thi.s fossa and the supina¬

tor crest probably represent the origin of the

supinatot (= supinator brevis). This muscle is

absent in Didelphis (Coues 1872) and probably

also in the other living didelphids since none of

them has a crue supinator crest and fossa. In

didelphids, the supination fijnetion is performed

by the brachioradialis. In CLidosictis, the supina¬

tor crest is présent but smaller than in AJaytilestes

and this structure is even smaller in Prothylacynus

and Thylachius. It is absent in Borhyaetia.

On the antérolatéral edge of the shaft is a robust

Fig. 25. — Pucadelphys andinus, holotype (YPFB Pal 6105).

Left uina; A, anterior; B, latéral: C, médial views. Scale bar:

5 mm.

interosseou-s crest which Ls in continuity with the

supinator crest. It is more salicnt than in

Didelphis, Aleiachiriis and Borhyaena where it is

fairly rounded. It is clearly not as sharp as in

Caluromys but approaches the condition of

P}'Othyla{y}nn. On the latéral side of tlie shaft is

an elongated dépréssion, relativcly deep in the

middle third of the shaft but becoming shallowcr

m its distal extremity. This area corresponds to

the origin of the abductor pollicîs lortgus and

probably to the extensor indicus propriu.s fthe

excensor pollicis longus is absent in Didelphis

(Coues 1872) and probably in didelphids in

général]. The latéral fossa of the uina of

Mayulestes is deeper ihan in mosc of the living

didelphids, excepr Ciî/urywyj where rhe fossa is

more pronounced. The condition of Mayidestes

is siinilar to those of Prothylacynus and

Cladosiciis but clearly differs from those of

Borltyaemi and Thylacinus where the fossa is very

rediiced.

On ihe médial side of the distal quarrer of the

diaphysis i.s a wcll-dcvcloped crest for rhe origin

of the pronatoi quadracus. It is sharper than in

the Santa Cruz borhyaenoids and approaches the

condition observed in Pucadcdphys and some

living didelphids (Didelphis, Monodelphih- On

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 26. — Mayulestes ferox, holotype (MHNC 1249). Left

radius: A. anterion B, poslerior C, latéral, D, médial: E. proxi¬

mal: F, distal views. Scale bar: A-D. 1 cm; E-F, 5 mm.

the anrcrolatcral sidc ot chc di.sral cxtrcmiiy of

thc diaphysis is a shaJIow groove which extends

on the epiphysis and probahly received the ten¬

don of the extcnsor carpii ulnaris. This structure

Fig. 27 — Mayulestes ferox. holotype (MHNC 1249). Lett

radius: A, anfenor. 8, postenpr views. Abbreviations: bt. bicipital

tuberosity: ptapi, passage of the tendon for the abductor polltcfs

tongus; sp styloiü process' stecr, sulcus for the tendon ot the

exteosor carpii radialis' siedc sulcus for lhe tendon of the

extensof digitorum communis; uf. ulnar facet, Scale bar; 1 cm.

is absent in Cladosictis^ Borhyaena and

Thylacimis. It is well-developed in Prothylacynm,

Didelphis and Caluromys.

The styloid procesf: o( the ulna of Mayulestes is

smail. regiilarly conical and flattened anteropos-

teriorly. [t diflers from those ot ihe other

borhyacnoids which are more rounded and bcar

a clcai anicrorncdial notch.

Radius. (Figs 26, 27) The bone is short as the

ulna and markedly recurved postciiorJy like

those ot Proihylacyuus^ C/adosiciis, l'hylaciuus,

Didelphis Caluromys, but diflers from that of

Borhyaena sn\s\cU is rclatively scraight. As preser-

ved, the liumeral facet is .sirongly oval-sluiped

(ahnost twîce longer than- widc) and resernbles

those of Cladositiis^ Borhyaena and Ihylacinus

but differs from the fiunieral lacets of

Prothylacynus, Didelphis and Caluromys which

tend 10 bc less elôngatcd transvcrsully. However,

it is likely that the anterior border of the articu-

lar facet has been slighily eroded (and perhaps

anteroposteriorly compressed) during fossilisa¬

tion. rherefore, it is probable that thc humerai

articular facct of thc radius of Mayulestes was

slightly less transverse than what is aciually

observed on the oniy known radius of

Mayulestes, possibly approaching the condition

GEODIVERSITAS • 1998 • 20<1)

Muizon C. de

of the living didciphids. The uln.tr facer is vcry

short proxiniodistally, It is shorter than in

Prothylacyuus^ Clachsictis and Borhyacna and

much shorter than ihose of Didclphis and

Caluromys. The bicipital tubemsity is weak; it is

smaller and locatcd furthcr distally on the shaft

than in Borhyuena and Cladosii'tii but rescinbles

the condition of Prothylacyuui. On the latéral

edge of the distal hall of the shalt is a conspi-

cuous ridgc Jimitcd by shallow grooves anteriorly

and posrerioriy and whcrc the radiaJ part of the

origin of the abductor pollicis longus probably

insertcd. This ridgc is better marlccd than in

Borhyaeiut, Ctadosicùs, Thyladtius and Didelphu

but considerably weaker than in Prothylacynm

and Caluromys. In the latter, it extends as a vcry

thin and salicnt blade from chc distal border of

the bicipital tuberosity (a little more distal in

Prothylacynus) to the latéral side of the distal

excremity of the shaft.

The distal epiphysis beats a well-developed sty-

loid apophysis which is bercer individualiscd

than in Borhyacna and Cladosktis and larger than

in Diddphb and Caluromys. It rotigfily resembles

that of Prothylacynus. On ihe ancerior side of the

distal epiphysis. lateioJ co the styloid process, are

nvo shallow grooves for the passage of the ten¬

don of the extensor carpi radiales medially and

extensor digitorum communis laccrally. The

médial side of the sryloid process ïs a fiat area

that transmirted the tendon for the abductor

pollicis longus. Those structures are better indi-

vidualised than in the other borhyacnoids but a

little less than in Didelphis; their development is

similar co chat observed in Caluromys. In distal

view of rhe bone, the distal epiphysis is flattened

anteroposteriorly, the scapholunar articulation is

oval-shaped, very concave» and the styloid pro¬

cess is lociued anieromedially. It differs from the

condition of rhe other borhyacnoids, Thylaànus

and Didclphis, where the scapholunar articula¬

tion is shallower and where the styloid process is

generally thicker and located medially. It

resembles the condition observed in Caluromys

although in this gcrius the styloid piocess is loca-

red medially and the articulât surlàce is shallower

and wider. 'l'he distal epiphysis ol the radias of

Mayulestes is not thickened medially and, on its

latéral side, the ulnar facet is hardly discernible.

Fig. 28. — Mayulestes ferox. holoîype (MHNC 1249) Unciform:

A. dorsal; B, proximal views. Abbreviations: eu. articular facet

witb the cuneiform; lu, articular facet with the lunar; mcIV, articu*

lar facet with the mcIV; mcV. articula? facet with the mcV. Scale

bar: 2 mm.

The latéral side of the distal extremity of rhe

.shaft i.s not thickened and not excavaced to receive

ulna.

Carpus. (Fig. 28) l’he only bonc known of the

carpiis of A'Iayulcstes is the right unciform. The

anterior side of the bonc is roughiy rriangular

with a developed medioproxlmal process which

articulâtes with the lunar as in Didelphh. The

uncilorm of Borhyacna (YPM PU 15701)

(Sinclair 1906, pl. 54, fîg. 4) is partially broken.

However, it appears to hâve hecn more quadrate

in anterior view and the lunar process must hâve

heen more massive than in Mayulam. The arti¬

cular facet for the cuneiform Is very sigmokl and

face.s proximolaierally as in Didclphis while it

face5 more proximally in Borhyacna. The arricu-

lar facet for the MclV and V is triangular as in

Didclphis and Borhyacna.

Metacarpus. (Fig. 29) The lefr McIII of the

holor}^pe is preserved. Il is incomplète and lacks

the distal epiphysis. On die proximal epiphysis,

the ;uücular facet for the McIV is more concave

than in Clndosicns and Didclphis. Contrary to

the condition observed in Didclphis^ ihe part of

the epiphysis which beats this facet protrudes

latcrally.

The leff McV of the holorype is known. It is

more siender and longer than those of Cladosktis

(YPM PU 13046), Sipaheyon (YPM PU 15154)

and Borhyacna (YPM PU 15701). The bone is

flattened dorsoplaniarly, a (caiure which is not

found in the Santa Cruz borhynenoid.s and

didelphids. The epiphyses are flattened in rhe

same plane, i.e. their main axes are parallel. As a

resuit, when the proximal epiphysis articulâtes

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 29. — Mayulestes ferox, holotype {MHNC 1249). McV:

A. dorsal: B, médial views: MclM: C, dorsal view. Scale bar:

5 mm.

with the McIV, che flexion axis of the distal cpi-

physis of the McV rends ro be closct to a perpcn-

dicular than to a parallel position in relation to

the dorsopalmar plane of the manus. rherefore,

the plane of articulation of digit V tends to he

pcrpendicular to chat of digit 111. In Didelphis,

the proximal epiphysis is flattened tran.sver.sely

and roughly pcrpendicular to the dorsopalmar

plane. Howeven ihe plane of the di.scal epiphysLs

is not parallel that of the proximal (as in

Mayulestes) and the articulation plane of digit V

makes an angle of approximatcly MO"" with that

of digit III. The medial surface of the proximal

epiphysis (which articulâtes wiih the McIV) is

very salicnt in Mayulestes^ and alrnost forms a

small condylc rhus indicaring a g<iod mobilfty of

the finger (espccially abduction and adduction),

This articulation i.s Ic.ss convex in DUlelphts. On

the distal epiphysis, in Mayulestes^ rhe condyie is

smallcr but more convex than in Didelphis and

there is a well marked articular fossa. whÜc it is

almosr absent in Didelphis. This condition

dénotés a grearer mobility of the fingers in

Mayulestes. Furthermore, as in Didelphis but to a

grcater excent, the distal epiphysis is recurved

medially. In Sipalocyon and Cladosicih the McV

are straight and in Borhyaena the distal epiphysis

is recurved laterally.

Hindlimh

Pelvis. (Figs 30, 31) The major characteristic of

the pelvis consists in the shape and proportions

of its ilium. The ilium is approximately 57% of

the total length of ihe bone as in Cladosiais (in

YPM PU 15170). Il is siniilar to the condition

fbund in Thy/acinus (AMNH 35244) where the

ilium is approximatcly 54% of ihc total length of

the bone but signiflcantly differs from Didelphis

and Caluromys where the proportions are 66%

and 67% respcctively. The body of the ilium

(the portion of the bone between the wing ante-

riorly and the acciabulum posieriorly) i.s longer

and more .slender ihan in Cladosictisj

Prothylacynus and Thyladnus but ir is shorter

than in Caluromys. It is fairly simiJar lo ihar of

Didelphis and Pucadelphys. The wing is longer

than in Pucadelphys but shorter ihan in

Didelphis, Marmosa and Caluromys. It approxi-

mates the relative size and proportions observed

in Prothylacy7iusa.n<\ Cladosictis.

In latéral view, ihe wing of the ilium has a

roughly rectangular outline. Its dorsal edge is

slightly convex and its ventral edge is wcakly

concave. In this respect it approaches the condi¬

tion observed in Cladosictis and Thylacimis.

However, in these généra, as well as in

Prothylaiynus^ the profile of the dorsal edge of

the ilium is less salient dorsally, and gives che

wing a more triangular morphology.

The po.sterodorsal iliac spme is well marked in

Mayulestes h\xx. it is much less developed than in

Cladosictis and in Ihothylacyniis, where ii is extre-

mcly salient and forms the anterior edge of a

very deep greater sciatic notch. The anccrodorsal

iliac spine is virtually absent in Mayulestes and

that angle of the ilium is rounded and certainly

does not deserve the namc of spine. In

Prothylacynus, Cl-adosictis and Thylacinus, chc

anterodorsal angle of the ilium is evcii more

rounded than. in Mayulestes d,x\A gives the ilium a

ventrally deflêcted shape. This is emphasised by

rhe morpholog)^ of the ventral edge of rhe ilium,

more concave in Prothylacyrius. Cladosictis and

Thylaciniis than in Mayulestes. Furthermore, ihc

anrerovenrral iliac spine Ls much more salient in

Prothylacynus, Cladosictis and Thylacimis than in

Mayulestes. On the medial side of the iliac spine

are inserted the quadratus lumborum (also on

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 30. — Mayulestes ferox, holotype (MHNC 1249). Right innomlnate: A, latéral; B, dorsal; C, ventral views. Scale bar: 1 cm.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 31. — Mayulestes ferox, holo-

type (MHNC 1249) Right Innomlna-

te in latéral view Abbreviations. ac,

acetabulum; adis, anlerodorsal iliac

spine; af. acetabular fossa; avis

anteroventral iliac spine. bü. body

of tne ilium; gl, gluieal fossa:

gsn. greater sclatic notch; ie. iliopu

bic eminence; H. iliac fossa; IL.

ilium; IS, ischium; isp, ischiatic

spine; It, ischiatic tuberosity Is

lunate surface, of obturator fora¬

men; pdis, posterodorsai iliac

spine. ssrt, smaller sciatic notch;

til, tuberosity for the rectus femoris;

wil, wing of the ilium. Scale bar:

1 cm.

the ventroniedial border of iKe ilium), a power-

ful flexor of lhe vertébral column and part of the

erector spinac (also on the dorsomedial face of

the ilium), the major excensor of the vertébral

column. In Aiayu/mes, a small iliac tuberosity or

posterovenrral iliac spine is présent while it is

absent in Prothylacynus^ Cladosictis and

Thylacinus. In Didelphis and CaluromySr the

ilium is longer and narrower than in Mayîdestes

and the iliac spines are relatively smoorh, like in

hdayulestes. In Pucadelphyi, the ilium bas an

intermédiare morphology bciween tliose of

Mayukstes and the Santa Cray borhyaenoids: as

in rhe former, there is a clear posterovenrral iliac

spine but, as in the latter, the postcrodorsal iliac

spine U more salieiu ihan in Mayulatvs, the grea¬

ter sciatic notch is well marked, and the very

roundcd anterodorsal angle and the concave ven¬

tral edge oi the ilium give the bone a ventrally

deflectcd morphology.

The lacerai side of the wing of the ilium of

Mayulestes bears a low and roundcd ridge which

arises from the body of the ilium, runs anteriorly

and vanishes in the anterior third of the wing.

That ridge is absent in Prothylacynus and

Cladosictis and very poorly marked in Thylacinus,

!n Caluromys, the ridge is more pronounced chan

in Mayîdestes and rcaches the anterior extremity

of the ilium; it is exrremely salienc in Didelphis

and divides the latéral .side of the bone into rwo

wcll defined fossac. The superior fossa mainly

receives the origin of the gluteus medjus (gluteaJ

fossa) and the lower that of the iliacus (iliac

fossa). In Mayulestes^ the two fossae are approxi-

mately of the same size as in living didelphids,

while in Pucadelphys the gluteal fossa is clcarly

larger than the iliac fossa. Jn dorsal view, lhe wing

of the ilium is thin as m Pucadelphys and conspi-

cuously everced as in Cladosictis^ Prothy-lacynus

and Pucadelphys^ contraty to Didelphis, Marrnosa

and Calurornys where the bouc is almost straight.

On the médial side of the ilium rhe articulation

with the sacrum is vciy littlc marked but it scems

thar only one sacral vertebra was articulating wirh

lhe btïne as in Cladodcth and Thylacinus ratlicr

than two as in Didelphis and Caluromys. As in

Pucadelphys and CladosiciiSy the wing of the ilium

is more evcited than in the living didelphids. The

important eversion of rhe ilium provides a larger

insertion area for the erector spinac.

On the latéral side of the body of the ilium, jtist

anterior to the acetabulum, the tuberosity for the

rectus femoris is small. It is, however, slighriy lar¬

ger than ihosc of Caluromys and Didelphis but

differs from chose of Prothylacynus and Cladosictis

which arc large and very salient. In Thylacinus,

rhe tuberosity is not as dcvelopcd as in ihese

généra but ic is larger than in Mayulestes. It is also

relatively large in Pucadelphys. On the ventral side

of the pelvis of Mayîdestes, ar a point which cor¬

responds îo îhe junction of the ilium and rhe

pubis, Ls a small tuberosity, the iliopubic eminen-

ce where is atrached the tendon of the psoas

minor, a flexor of the lumbar part of the vertébral

column. The iliopubic emmence of Mayulestes is

weaker than in Cladosictis and Prothylacynus but

more developed than in Didelphis and Caluromys

where it is sometimes totally absent.

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Fig. 32. — Mayulestes ferox, holotype (MHNC 1249). Left fémur: A, anterior; B, posterior; C, latéral; D, médial; E, proximal. Right

fémur: F, anterior; G, posterior; H, distal views. Scale bar: A-D, F, G, 1 cm; E, H, 5 mm.

GEODIVERSITAS • 1998 • 20(1)

MayuUstes, a borhyaenoid from the Palaeocene of Bolivia

The acetabulum is shallower and more open

than in Cladosictis, ProthyUcynus and Didelphis.

It similar co thac of Caluromys and indicates a

greater mobility at the articulation. The lunate

surface is the arricular surface wich the femur; it

is composed of a ventral and a dorsal lobe sepa-

rated by the acetabular fossa. In Mayulestes the

dorsal lobe of the lunate surface is smaller and

narrower than the ventral onc. The same condi¬

tion exisrs in Cabiromys, Didelphis and Pucadel-

phys while in Cladosictis and Prothylacynus both

lobes are approximately the same size and in

ThyUcinm the dorsal lobe is longer and wider

than the ventral one- The anterior border of the

acetabulum is salicne laterally and thickened; the

dorsal border, in dorsal view, is well excavated

{i.e. concave laterally), The condition of

Mayulestes is similar to that of Caluromys^

Phalangevy Petaums and Pucadelphysy but differs

from those of Cladosiais, Prothylacynus^ Meta-

chirus^sxA Didelphis.

The ishium is longer than in the Didelphidae

and similar in size to chat of Cladosictis. On the

dorsal edge of the bone, posterior to the acetabu¬

lum, is a small tubercle, the ischiatic spine. In

Mayulestes ir is sLightly larger than in CaturomySf

Didelphis and PucadelphySy but slighcly less deve-

loped than in Cladosictis^ Pi’othylacynus and

Thylacinus, Posterior co the ischiatic spine is the

smaller sciacic notch which, consequently, is less

pronounced in Mayulestes than in the Santa Cruz

borhyaenoids. In its posterior portion, the

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

ischium of Mayulestes is longer and narrower

than in Caluromys and Didelphis but resembles

in these respects the condition obscrved in

Cladosictis. The posierodorsal angle of the

ischium is chc ivschiatic tuberosity where originatc

the biceps fernoris and the semiiendlnosus, boih

extensors of the thigh. The ischiatic cuberosity is

weak in Mayulestes and Calaromys. le is stronger

in Didelphis and Cladosictis and much stronger

in Thylachius.

The posterodorsal angle of the ischium is not

modified in a true iscliiatic tuberosity and is not

slightly recurved vcrurolatcrally as ir is observed

in didelphids. In Cladosictis and Prothylacynus

they are more developed rhan in Mayidestes and,

in Thylacwus, ir is stronger than in the former

généra. The ischiatic tuberosity of didelphids is

reduced but a lictlc stronger in Didelphis,

Metachirus and Monodelphis than in the other

Fig. 33. — ferox. holotype (MHNC 1249). Left

fémur: A. anterior; B, postenor. Abbreviations; fc, tovea capitls;

ft. fémoral trochlea; gt, greater irochanler; h head; if. intercon*

dylar fos&a: le, latéral condyle: le, latéral epicoodyle; It. lesser

trochanter; Itc, latéral troohlear crest; me. médiat condyle:

me, médial épicondyle; mtc, médial trochlear crest; n, neck;

tf. trochantehc fossa; tt. third trochanter. Scale bar: 1 cm.

living généra of the family. In Caluromys and

Piicadelphys h is extremely fiat as in Mayulestes

but more salicnt.

In ventrolaterai view, the dorsal borders of the

ilium and the ischium of Mayulestes make an

angle of approximately 155"^. In didelphids, pha¬

lange rid s, Cladosictis^ Prothylacynus and

Thylacwus the dorsal borders of the ilium and

the ischium are roughiy parallcl and aligned.

Femur, (Figs 32, 33) Neithei of the femora of

the holotype of Mayulestes is complété. The left

fémur is the best preserved but lacks a .small por¬

tion of the neck, ihe apex oi the lesser trochan¬

ter, part of the diaphy.sis and the mcdial distal

condyle; furthermorc, the distal part oJ the shaft

shows a slighc post-mortem deformation. The

right fémur lacks the apex of the greater trochan¬

ter, the distal third of the diaphysis and a .srnall

posterior portion of ihe médial distal condyle.

Both bones, however, allow an accuratc recons¬

truction of the fcmiir oi Alnyulestes.

The lemur of Mayulestes Ls relatively short when

compared co thosc of Cladosictis and Prothylacy¬

nus but rcscmblcs that of Borhyaeua which is

motc massive. It is shorter than that of Calu¬

romys but its proportions are close to thaï of

Didelphis. In latéral view, the proximal epiphysis

is clearly bent anteriorly as in Cladosictis,

Pticadelphys (Fig. 34) and Caluromys, coatrary to

Borhyaena where it is almost straighi. In

Didelphis and Thylachius, the curvuiure is pré¬

sent but much less pronounced than in

Mayulestes. That feature is uiso very clear in

proximal view. The proximal condyle (or head)

is slightly elongated in proximomedial view

(Le. cornpressed anteropo.steriorly) and the arti¬

culât surfece extends on tire neck laterally (there

is a small variation beeween both femora since

the condyle of che right fémur is slightly more

cornpressed anieroposteriorly), This œndition is

relatively similar to that of Calurotnys, Didelphis

and the otiier borhyacnoids. Kowever, che proxi¬

mal condyle of Mayulestes is less globular than in

these forms and more resembles in this respect

that of l'hylaiinus. The fovea capitis for the

atcachment of the ligament of the head of the

fémur is locaicd on the posteromedial side of the

head, close to the border of the articular surface.

It differs from what is observed in Didelphis,

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Calîtromys and Cladosktisy where ir is somewhar

more central on the condyle; ir is doser ro the

position observed in Borhyacna and Thyladnus,

The fovea capitis is larger rhan in didelphids and

other borhyaenoids, suggesting a much stronger

atcachmcnc of the ligartienl in Mayulestes. The

neck is short as in CMluromys and difTers from

that of Clcidosiclh and, to a Icsscr cxtcnc,

Didelphk which arc longer. In this respect, it

somehow rcsembles those of Borhyuena and

Prothylacynus. fhe head and rhe neck are orien-

ted less proximally and the distal side of the neck

is more concave ihan in CLxdosicth axmX Didelphis

but they rcscmblc the condition observed in

Caluromys, The grcater irochantci is for the

insertion of the thrcc glutei, rvvo of which hâve

their origîn un rhe latéral side of the ilium. The

gluteus médius is inserted on the posteroproxi-

mal angle, and rhe gluteus profundus on the

anteroproxima) angle. 1 hc grcater trochaiuer of

Mayulestes is higher tlian the condyle, a condi¬

tion aiso found to a lesser extent in Borhyaena

and Thyladnns. In Cdadvsicth, Prothylacynus,

Didelphis and Caluromys the gfeater trochanter is

always lower than the coiulyle. In latéral view,

the grcater trochanter hus an angular apex with

salieni insertion areâs for the giuiei médius and

profundus. On the antêrior side of the greater

trochanter is an clongatcd shallow fo-ssa, for the

vasri lateralis and intermediits, which runs distal-

ly along approximately one quarter of the Icngth

of ihe bmie. That fo-ssa is absent in Cladosictis,

Pro tbyla cy n us, Thylaein us. Calu ro mys a n d

Didelphis, in Borhyaena the origin of the vavsri

lateralis and intermedias is a slightly depressed

area locared on the anterior side of the greater

tmehanrer. The lesser trochanrer is a large médial

triangular blade on rhe apex of which is inserted

the tendon of the iliopsoas (iliacus + psoas

major). It is clearly more developed than in

Prothylacynus, Cladosicîis, Borhyaena and

Thytadnus. It rcsembles those of Didelphis and

Caluromys, although larger and thinner. Distal to

the grearer trochanrer, on the latéral side of rhe

bone, is rhe third trochanrer. In fact, in Mayu-

lestes, il is more a latéral expansion of rhe bone

rhan a trochanter, which is in conrinuiiy^ with

the antérolatéral crest of tlie greater trochanter.

The third trochanrer receives the insertion of the

gluteus superficialis, an important abductor and

inverrer of the hip. In Mayulestes, rhe area of the

ihird trochanter Is markcdly expanded laterally

while it is siraight in Cladosictis, Prothylacynus,

F'G. 34. — Pucadelphys andinus {YPFB Pal 6106). Right fémur:

A, anterior; B. posterlor; C, latéral; D, médial; E, proximal;

F, distal views. Scale bar: 5 mm.

GEODIVERSITAS • 1993 • 20(1)

Muizon C. de

Fig. 35. — Mayulestes ferox, holotype (MHNC 1249). Right tibia: A, anterior; B, posterior; C, latéral; D, medlal; E, proximal; F, distal

views. Scale bar: A-D, 1 cm; E, F, 5 mm

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Borhyaena and Thylacinus. In Didelphis, postero-

medial to the insertion of the gliueus superficia-

lis and lacerodistal to the apex of the lesser

trochanter is a stnall rubercle for the insertion of

the quadratus femoris* an extensor of the hip

and an evcrtor or the thigh. No such tubcrcle

exists on the holorype of Mayulestesf whde ir îs

strongly developed in Q-adosictis, ProthylacynuSy

Borhyaena and Thylacinus. In Cladostctis it is

especialiy salient and connected to the lesser tro¬

chanter proximomedially through an oblique

ridge. On the posterior side of the gceater tro¬

chanter, the trochanteric tossa rcceives the obtu-

ratorii internus and externus and the gemelli

muscles. The fossa is deep but not very elonga-

ted proximodistally. Its distal extremity is slighdy

more proximal than the level of the apex of the

lesser trochanter. This condition differs from

ihose of Chdosictisj Protfjylncynus, Borhyaena and

Thylacinus, whcre the ibssa is much longer and

reaches the distal extremity of the lesser trochan¬

ter blade distally. The trochanteric fossa of

Mayulestes is much more redueed than in these

généra and rescmbles those of Didelphis and

Caluromys. In latéral view of the bone the shaft is

straight as in didelphids and the Santa Cru 2

borhyaenoid.s. It differs from Thylacinus whose

fémur is beni posteriorly, a cursonal featurc.

The anterior face of the distal extremity of the

fémur beats a large rrochlea (the patella was pav

bably absent in Mayulestes) lor the passage of the

tendons of the vasti and recrus fcmori.s muscles,

It is deep wlien compared to didelphids and has

an important proximal extension on the antenor

side ol the shaft; ihe extension of the trochlca is

more developed on the latéral side than on the

médial. The rrochlea is .shifted laterally, approa-

ching the condirion of Eozostradon (jenkins ÔC

Farrington T^76). The. latéral and médial bor-

ders of the rrochlea form sharp crests, the latéral

one being more elevated than the médial one.

The rrochlea of Mayulestes is more developed

than in the other borhyaenoids: it is wider, has

more pronounced relief and expand? more di.s-

tally on the shaft, It is, however, doser to that of

Cladostctis than to those of Borhyaena and

Protbylacynus, althoiigh ît is wifler. It differs from

that of Didelphis which has a less pronounced

relief, and is notably different from that of

Caluromys, whicb is distoproximally very short

and very fiat. The distal extremity of chc shaft,

proximal to the trochlea, beats a shallcrw sulcus

for rhe passage of the tendon of the vasti and

rectus femotis. A similar condition is observed in

Thylacinus. In the other borhyaenoids and in

Didelphis there is no groove but a fiat surface

wliich aiso corresponds to the passage of the ten¬

don. That région of the femur in Caluromys is

slightly convex.

In distal view, the latéral condyle is conspicuous-

ly larger and wider than the médial. This condi¬

tion is close to that of Didelphis-, Caluromys and

Phalanger wherc the médial condyle is narrower

than the latéral (alihough more pronounced in

these généra), but differs from chat observed in

Cladosictis, Borhyaena and Prothylacynus, whcie

the lacerai condyle is narrower chan or subcqual

CO the media) one, Since the médial condyle of

Mayulestes is noi compleiely preserved, it is not

possible lo observe the relative anteroposccrior

extension of borh condyle.s. However, ii is pro¬

bable chat ihe médial condyle was ac leasi as long

as the latéral one or even slightly longer. In lace¬

rai view it is noteworchy char the latéral condyle

is much less globular than in Cladostctis, Prothy-

tacynusy Didelphis and Caluromys. The ariicular

surface of the lacerai fémoral condyle of Mayu¬

lestes is not recurved po.steroproximally as in

these gçnçra and rhe condyle is protruding pos¬

teriorly, a condition that must hâve redueed the

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

amplitude ol thc movcrnent posteriorly (rhe

flexion of ihe knce). Tfiere is a Ijrge and deep

po.stcondviar los.sa fcalled popliteal surface by

Evans &C Christensen (1979) in tbe dog], proba-

bly a conséquence of tlic littic rccui ved posterior

border of tbe condyles. In anterior or posterior

view, tbe médial condyle is slightiy lower {i.e.

more distal) tban tbe latéral oiie as in Prothy-

lacynus, Cladoiiciis, Borhyaena, l'bylucinus,

Didelphis and Caluromys. On rhe latéral sidc ol

tbe latéral condyle is the fessa fer the origin ol

tbe poplitcus muscle (whieb logically should be

called popliteus fessa). It is small and shallow as

in Didvlphis and Cdluromys but differs from

Cladosictiu ProihylacynuSy Borhyaena and

ThyLîciuns where it i*» well marked. Tbe épicon¬

dyles are not signiflcantly diflerenr frt>m tbose of

the généra cunsidered here. In distal view, rhe

distal epiphysis of the fémur is less flattened

anteroposteriorly rhan in Didelphis, Caluromys,

Monodelphh and Philander but approaches the

condition of Mtruchirus. It is however clearly

wider than long and approaches tbe conditioii

observed in the Santa Crux borhyaenoids; it is

proportionally slighily longer anteroposteriorly

than in Pruthylaçymis and Borhyaena but slighily

shorter than in Cladosktis.

Patella. No patella bas bcen fdund associated to

the holorype and it is suggested rhat it was not

ossified in Alayulestcs as it generally occurs in

didelphids.

Tibia. (Figs 35, 36) i'hc bonc has the typical

didelphid sigmoid shape, a condition that is

found in ail the représentatives of the family. In

Mayulestes it is tnore pronounced rhan in

Didelphis and Caluromys', Ic.ss rnarked rhan in

Chironeetes but simÜar t<j whac is observed in

Marmosa and Lutreolina. In Prothylacynus and

Thylncintis rhe tibia is .srraight and, in Cladosictis,

its distal exircmity is sügbtly bent mcdialJy. Jn

Mayuksies, in posterior view tbe proximal third

of the sbaft is concave laterally and rhe cwo distal

rhirds arc convex. The same condition is found

in Didelphis, Marmosa and Caluromys-, in

Chironeetes, Metachirus and l.utreolina the

inflexion point is located al the middle of the

shaft. The tibia of Thylacinus is aiso slightly

concavoconvex with the inflexion point at the

middle of the shaft. In latéral and distal view, the

F»g, 36. — Mayulestes ferox. holotype (MHNC1249). Right tibia;

A, anterior; 6, posterior views. Abbreviations; aica, anterior

intercondyloid area, t1, tibular facei; ic, intercondyloid eminence;

le, latéral condyle; me. medial condyle. mm media» malleolus;

tb. tibial tuberosity; te, tibial crest. Scaie bar: 1 cm.

shaft of the tibia of Mayulestes is rcgularly bent

posteriorly as in didelphids, contrary to the

condition uf rhe Santa Cruz borhyaenoids.

In proximal view, rhe proximal epiphysis is short

anteroposteriorly as in Caluromys and Didelphis

(to a Icsser extent in this geniis) and differs from

rhe condition observed in Prothylacynus and

Cladosictis, whose proximal epiphysis of the tibia

is markedly rriangular. The tibial cuberosit}' is

small and very lirtie salienr anreriorly, fr is smal-

ler rhan in Caluromys and Didelphis and much

smailcr rhan in Prothylacynus, Cladosictis and

Thylacinus, whene it.s development i.s re.spon.sible

fer the triangular shapc of the proximal epiphy¬

sis. In latéral view tlie tibia! tuberosity is trunca-

red and forms a rcgular slope on the antero-

proximal angle of the tibia. This condition is

found in ail didelphids and boibyaenoid but dif¬

fers Irom rhat observed in Thylacinus whcrc the

ruberosiiy is very salicne and forms a right angle

with the shaft. Between the tuberosity and the

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 37. — Pucadelphys andinus (YPFB Pal 6106). Right tibia:

A, anterior; B. posteriori C. latéral: D, médial; E, proximal;

F, distal views. Scale bar: A-D. 5 mm; E. F, 2.5 mm.

condyles îs a surface ot rugose bone called the

anterior intcrcondyloid area. This area is very

short antcropo-steriorly as in Caluromys and

Didelphis and differ from the very long area

observed in Prothylncy-rms, Cladosiclis and

Thylacinus, Tlie latéral tibial condylc is subtrian-

gular. almo.st Hat, the médial condylc is reni-

forni, cxcavated and is a ürtle more distal rhan

the latéral. The two condyles are aineroposte*

riorly shorter than in Caluromys, Didelphis

Thylacinus and probahly Cladosicth (the two spé¬

cimens I cüiild observe during this study had a

poorly preserved proximal exiremity). In

Prothylatyrius the condyles arc even shorter than

in Aiaytdcstes the médial condylc is subcircu-

lai in shapc. The fibulur faccr of Mayutntn is

elongated transversely and has a postérolatéral

orientation as in ail didelphids and borhyaenoids

obser\Td during this study. The intercondyloid

emincncc is wcll-developcd and rountled. Je is

larger than in Prothylacyims but smaller rhan in

the didelphids and Thyladnus. The posterior

inrercondyloid area is almosr absent and corres¬

ponds to rhe posterior slope of che intcrcondy¬

loid erninence. Thcre is a small posterior inter¬

condyloid area in ProthyLicytcus, Cladosîctts and

Thylacinus.

On rhe anterior edge of rhe proximal extremity

of rhe shafr is an unatural ovoid caviiy. le i.s diffi-

cult to derermine if this structure i.s pre-niortem

{i.e. paihological or rraumatological) or post-

moriem (i.e. caphonomical), although the

smoothness of ics edge.s would indicaie .some

bonc growrh and would favour the pathological-

trau matological hypothesis.

The tibial crest is smoorhly convex and does not

form U roLinded keel as in Caturomys^ Didelphisy

Pwthyhicynus and CUdosictis, The keel is much

more salient and much narrower in Thylacinus.

The latéral tibial fossa is small bur well marked.

It is smaller than in Prothylacynus'Ans\ Thyhiânus

but deeper rhan and not as Hat as in Didelphis

and Caluromys. The médial ribial fossa is

well-developed as in Caluromys and difFers from

the condition observed in ProthylacynuSy

CLidosicm. Thylacinus and Didelphis^ where ic is

more a Hat area than a fo.ssa. On the posterior

sicle of the proximal extremity of rhe shaft is a

deep fossa located distal to the popliteal notch.

As it stands in Mayukstfs, it is probable that it

has been emphasised by a post-mortem deforma¬

tion. However, ît scems to hâve been relatively

deeper than in rhe didelphids, a condition simi-

lar to thac observed in Prothylacyniis and

Cladosictis. J he distal two thirds of the shaft are

transversely compreSsed a$ in Prothylacynus,

Cladosictis and Thylacinus. This condition is also

found in Caluromys and Didelphis^ where it is

more pronounced. Ün the postérolatéral side of

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

the shaft is a marked crest ruaning distally from

the posterior ribial lossa toward thc distal extre-

miîy until thc distal third of thc bone. This crest

is probably for the intcrosseous membrane which

unités the tibia and fibuJa as in aJJ didelphids.

This structure is strongcr than in Prothylacynus

but it is much weaker than in Cladosictis whcre it

almosr reaches the distal extremity of the shaft.

The distal extremity bears ihe articular surface

for the astragalus which is divided into a media!

and a latéral faeet* In Mayulesteshoih feccts form

a marked angle a little larger than 90°. It

resembles the condition of Prothylacynus and

Cladosictis^ where the angle is close co 90” (or

slightiy sraaller) and difFcrs from the condition

observed in thc living didelphids where the angle

is much more open. The malleolus is large and

high and occuplcs thc en cire Icngth of thc ante-

remédiai .side of die distal epiphysts, concrary ta

the didelphicl condition where the malleolus is

anteroposreriorly shorter, The condition of this

feature in Cladosictis is intermedlate between

those of Mayulestes and didelphids. l*he malleo¬

lus is flattened rransvcrscly and its major axis has

a posteromedial-antcrolateral orientation and

forms an angle of approximately 48* wiih rhe

transverse axis of thc tibial cçmdyles. tn the

Santa Cruy. borhyacnoids and in Thylacinus, the

plane o( die malleolus is ai 90" with the trans-

versc axis of thc femorotibial articulation (the

functional interprétation and significance of this

feature are discussed below). The médial faeet of

the astfagaloribi.al articulation is large as in the

other borhyuenoidsi unlike in didelphids, where

it is short, anteroposteriorly and proximodistally.

It is strongly convex and faces poscerolateraliy.

The latéral faeet of the articulation is reniform,

slightiy concave and oriented slightiy obliquely

in relation to the boundart'^ (in latéral view) bet¬

ween the epiphysJs and the shaft (Fig. 35C).

This condition is intermédiare between that ot

Prothylacynus and Thylacinusy wherê it is parallel,

and that of didelphids, where the faeet aiways

makes an angle of 30 to 45* with the limit bet¬

ween the epiphysis and die shaft. As a consé¬

quence, the latéral astragalotibial faeet of

didelphids is helical and screwsaround che mal¬

leolus. The astragalotibial articulation of

Cladosictis is also slightiy oblique as in Mayulestes

but its astragalotibial articulation is noT helical as

in thc didelphids. The latéral faeet îs short ante¬

roposteriorly a.s in ProtMacynus and ThyLzcinus

and ditfers from those of Cladosictis and

Didelphis which are longer. In Pucadelphys the

condition ot the distal articulation ot die tibia is

similar co that of Mayukstes (Fig. 37). The mal¬

leolus torms an angle of approximately 69" with

che axis of the femorotibial articulation and die

lacerai and médial facets of the astragalotibial arti¬

culation arc approximately at right angle

The articular faeet for the fibula is located on the

latéral edge of the distal epiphysis. In Mayulestes,

it is shorter proximodistally than in Prothy-

lacynus and Cladosictis and occupies the whole

postemlareral edge ofthe distal epiphysis. It dif-

fefs from that of Thylacinus where it is rescricted

ro the antérolatéral angle of die epipliysis'. As in

Prothylacynus and CLidosictls, the lareral edge of

rhe discal epiphysis is more salicne than in

Thylacinus and m didelphids (where ir is almost

in continuity with the latéral side of rhe shaft).

Fibula. (Fig. 38) Only the distal extremity of the

right fibula is known. T'he distal epiphysis has a

subtriangular shape in distal view. Contrary to

what is observed in didelphids, ir is ver)' salient

medially and it is probable that the distal extre-

mity of the shafts of the bones were broadly

separaied. The articular surface for the astragalus

is concavoconvex; il is not possible to observe

whether it was aniculating also with the calca-

Fig. 38. — Mayulestes ferox, holotype (MHNC 1249). Right fibu¬

la; A, anterior; B, posterior; C, distal vtews. Scafe bar: A, 6,

1 cm; C, 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the PaJaeocene of Bolivia

neum or not. In the didelphids, rhere is no fibu-

localcanear arriculatîon, except in Didelphis,

Metachirus and Chironecîes (Szalay 1994; 196,

340). On the latéral side of the epiphysis is a

well marked ga>m'e for rhe passage of the tendon

of the extensor digitorum lateralis and peroneus

brevis and longus rauscles. A similar condition is

found in ihe living didelphids. In Prothylatymis

and CladosictiSy the passage of the tendon is a flat

area and there ’ts no groove.

Tarsus. (Figs 39, 40) The calcanéum Is the only

bone of the carsus that is known in Mayidestes.

The right calcanéum is complété; only ihe distal

extremity of rhe left one is preserved. Compa-

risons of rhat bone will be made with rhose ot

Sîpalocyon (YPM PU 15154), Cladosictis (YPM

PU 15046) and with six large (for the fauna) cal-

canea trom the early laie Palacoccne of Itaborai

(Bra/.il) which are most likely rcferable to

borhyaenoids although not necessarily to the

same taxon. Because of their size, the four smal-

ler specimens (DGM 1. 175-M, 176-M» 178-M,

179-M) belong to the IMG ([cabotai Metathe-

rian Group) ^ ol Szalay (1994) and could fit

the genus Patène while the two larger ones

(DGM 1. 180-M, 184-M) belong to the

IMG XIII of Szalay (1994) and could fit

Nemolestes, lo the following description, for

practical reasons they will be referred to the

‘‘Itaborai borhyaenoid calcânea” and no référencé

is made to spécifie gênera.

The ruber calcanei of Mayuknes is relatively lon-

Fig. 39. — Mayulestes ferox, holotype (MHNC 1249). Right calcanéum: A, anterior; B, posterior; C, latéral; D, medlal; E, distal views.

Scale bar: 5 mm. '

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

ger and more slcndcr than in Cladosictis,

Sîpabcyon and the Itaboraian calcanca. The ectal

facet (the latéral articular fecet for the astragaliis)

is smaJI, short proximodisfally and very narrow.

Latéral to the ectal lacet rhere was either no cal-

caneofibular (CaFi) lacer or a very narrow

strip-like facct. A calcancofibular lacet is présent

in Cladosicth and Sipalocyon and in ail the

Itaborai borhyacnoid calcanca menüoned above.

In onc spccimen Hgured by Szalay (1994,

fîg. 6.27 A-C), the calcaneofibular facct is very

similar to tliat whicli coiild exist in MayuLe^tes in

its length, narrowness and orientation, while in

the other speciinens die facet is generaJIy wider

and/or ohliquely ofiented. In the Itaborat

borhyacnoid calcanca, the artictilar contact with

the fibula is generally smaller rban in Cladosictis

and Sipdlocyon (Szalay 1994. fig. 6.27, 6.31),

although it is larger (as large as the ectal facet) in

DGM J.I60-M, a specimen which “falls within

the range expecied lor Patcdc^ (Szalay 1994:

177). On ihe latéral side of the protubérance

which beats the cctal and fibular lacets is a well

marked insertion area for the calcancofîbular

ligament. Medtally, the dorsal side of the susten-

raculum rali beats the su.srenracular facct (the

medial articulation for the ascragaliis). It is smal-

1er and more gracile rhan in Cladosicns and

Sipalocyon and, as in chese généra, its articulai*

facet occiipies the entire length of its dorsal side

and reaches distally the dorsolateral edge of rhe

cuboid facct, conrrary to ihe condition observed

in didelphids. The sustentacular facet is slightly

oblique reUtively to rhe axis of the tuber calcanei

as observed in the Itaboraf calcanca and

Sipttlocyou. concrary to the condition of

Chidosktiu wherc it is parallci to çhc cuber. The

orienrarion of the plane of the sustentacular facet

is mainly medial with a srnall dorsal component.

In Slpuloc^>(tn its orientation is almost torally dor¬

sal and in didelphids it is intermediate between

Mayukstes and Sipalocyon. The peroncal process

of ihe calcanéum is expanded distolaterally and

bears a deep groove for the passage til rhe per-

oncus longLis. In Mayulestes the peroneal process

is larger and the groove is deeper than in

Cladosictis and Sipalocyorr, in thèse respects, they

are similar to those of the Itaborai calcanca.

Furthermore, the latéral wall ol the sulcas is very

thick and longer than the medial in Mayulestes as

in the Itaborai calcanca, contrary to (ILrdosictis

and Sipaloiyon. The large and laterally expanded

peroncal process ol Muyîdestes Icaves a large space

for the passage ot the tendon of the peroneus

brevis on its anterior (ace, and lor rhe abductor

digiti quinti on its posrerior face (Godinot ôc

Prasad 1994; Prasad & (iodinot 1994). The pos-

terior face of the calcanéum of Mayulesies possess

a small ridge which rtins disndly fioin the posce-

romedial edge of rhe tuber calcanei and reaches

the small distal planrar tubercle on ihe postcrior

border of the c;dcaneücuboid articulation. The

posrerior side of the peroneal process is widc and

deeply concave and char of the susientaculum

cali is relatively fiat. In Sipalotyon and Cladasictis,

the ridge is much more developed than in

Fig. 40. — Mayulestes ferox, holoiype (MHNC 1249). Righi catcaneum: A. anterior; B. posterior, C. latéral; D. medial. E. distal views.

Abbreviations: at. anterior piantar tubercte; cf. cuboid lacet; cff. calcaneofibular facet; dpt. distal plantar tubercle; ef, ectal facet; icfl,

insertion for calcaneofibular ligament: pp. peroneal process; pr, posterior ridge; spl, sulcus for the passage of lhe peroneus longus;

St, sustentaculum tali; stf, sustentacular tacet; te, tuber calcanei. Scaie bar: 5 mm.

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Mayulesm, thc posterior side of the sustcntacu-

lum is convex and the poiterior sidc t)f ihe per-

oneal process is narrow and only slightiy

concave. When compared ro char of Mayulestes^

the posterior sidc of the calcaitcum of Sipalocyon

and Cladoskin is intlatcd po.sicriorly,. probably in

ordcr to strcJigrhen che bone. In the habtjrai cal-

canca, ihc condition is intermediacc bctween

thar of Mayulestes and tho.sc of Cladosktis and

Sipalocyon since diey hâve a slightiy chickcr pos¬

terior ridge. As a coJi.sequencc of the inorpholo-

gy of the posterior sidc of the calcanéum, in

distal view, tlie arricular facet tor the cuboid has

a .straighr border in Mayulestes, while ii is convex

püsteriorly in Cladosuils and Sipalocyon. 1 he

condition of the Iraborai calcanca resembles

more chat ot Mayulestes ihan thac of the

Santacru/ian généra. As in che Itaborai calcanea.

the cuboid facer is deeper in Mayulestes than in

Sipalocyon and Cladoskth.

The calcanéum of Mayulestes from thar of

Didelphisyn\\\c\\ Ls more spccialiscd. In Didelphh

the cuber calcanei is more robust and short, the

octal facet is very narrow, the suscenracular facet

does Ilot reach che cuboid facet disially, and the

sustcntaculum tali is notched mcdially. The sus-

tentacular and ccral facers are very close to onc

anorher in Didelphh, while they are well separa-

ted in Mayulestes, rhe peroneal process is small,

does not reach the latéral border of the cuboid

facet disially, and ihc groove for rhe pcroneal

muscle has losc its incdial wall. Purthermore, the

calcanéum of Mayulestes dlKers from thar of

most didelpliids [a secondarily acxjuircd CaFi

facet is présent in Didclphis (ahhough small in

this gemts), Metchirus^ and Clnronectes (Szalay,

1994)1 which hâve lost the calcaneofibular

(CaFi) facet, a derived condition aiso found in

peradectids (Szalay 1994). Contrary to che scatc-

menc by Marshall ÔC Sigugneau-Russell (1995),

the calcanéum of Pttcadelphys has a clcar calca¬

neofibular facet distinclly observable on the fur-

ther prepared left tais-us of YPfb Pal 6106

(Fig. 4l).

One of the mosc characteristic modifications of

the calcanéum of didclphids is the presence, pos-

teriorly, of a deep notch on che posterior border

of che cuboid facet for the articulation of a large

proximal styloid process of the cuboid, around

which che calcanéum probably has some rotation

ability (CaCup facet of Szalay 1994). Both

cuboid (Cacud and CaCup) facets arc separated

by a well marked .scmicirculai tidge. This fearu-

rc, which is a dldelphid synapomorphy (Szalay

1994), is absent in Mayulestes. Contrary ro rhe

statemenr by Marshall Sigogneau-Russell

(1995: l48), rhe calcanéum of Pticndelphys Aots

noi show, even incipiently, any trace of the very

typicil proximal indenracion of the cuboid facet,

as it is observed in che livmg didelphids and in

cwo fossil didelphids from the Palaeocene of

FiC. 41. — Pucadelphys andinus (YPFB Pal 6106). Right calca¬

néum: A, anterior; B, posterior; C, latéral; D, médial; E, distal; F,

proximal views. Scale bar: 2.5 mm.

GÊODIVERSITAS • 1998 • 20(1)

Muizon C. de

Itaboraf (BraziJ) (Szalay 1994, fig. 6.23). In rhis

respect, che calcanéum ol PucadAphys Is simiJar

to rhat of Mayulestes and to most Norch

American Laie Crctaceous and Tertiary marsu¬

pial calcanea figured by Szalay (1994;

Chapter 6). rhe calcanéum of Pticadelphys does

not possess a CaCup lacet. In fact, the concavity

of the plantai aspect ot rhe cuboid facet (what

Marshall Sigogncau-Russell questionably

regard as an incipient development of che didel-

phid condition) is more pronounced in

Mayulestes than in PucadclplTys (Y?FB Pal 6105

and 6106). The specimen YPFB Pal 6110

(Marshall & Sigogneau-Russell 1995, fig. 50) is

chat of a juvénile and the po.srerior border of the

cuboid facet Is probably slighdy worn (this has

been poinied oui lo me by D. Sigogneau-

Russell). The conséquence is chat the proximal

extension of the cuboid facet of that specimen

appears more developcd than it actually was, and

more than in the other specimens (YPFB

Pal 6105 and 6106). Whatever ihe orienration of

the calcancocuboid facet of PucadtHphys is, it is

clear that it is simple (/.c*. il does not hâve a

CaCup (àcet) and does not présent an incipient

development of rhe calcaneocuboid didclphid

synapomorphy Therefore, rhe calcanéum of

Pucadelphys does not hâve nvo of the major cal-

canear synapomorphie.s (ound in didelphids

(présence of CaCup and loss of CaFi facets).

Metatarsius. (Fig. 42) Iwo metapodials are refer-

red to left Mtllï and MrIV. However, the déter¬

mination of ihe Mtlll is uncertain since part of

the proximal extremin' is lacking. The relative

length and proportions of che Mtlîl are similar

ro chose of Sipalocyon and Ctadosictis. It is clearly

shorter than in Protijylacynus, In anterior view,

the articulât lacet for the ectocunoiform is more

convex and more bent anteriorly than in

Sipalocyon allowing probably better flexion

movemenrs of the fooc. On the distal extremity,

the condyle is less globular than in Sipalocyon

and there is a deep articulât fossa on rhe anterior

side of the bone, proximal to the artîcular

condyle. Thi.s latter condiciun aiso dénotés better

articulation and wider movemeni of the digits.

The MtIV is relatively shorter rhan in Sipalocyon

and does not présent rhe weak latéral curvarure

observed in this genus. les proximal articulât

Fig. 42. — Mayulestes ferox, holotype (MHNC 1249). A. Left

Mtlll in dorsal view; B. Left MttV in dorsal view. Scale bar:

5 mm.

facet with the cuboid is more inclined anteriorly

and lacerally than in Cladosictis and would also

indicates more agiÜty in the movements of the

foot. The distal articulation is similar to chat of

theMtI]l.

Mtlll (13 mm) is slightiy longer than MtIV

(12.2 mm). They are both slightiy longer than

onc third ol rhe length of r.hc tibia.

DISCUSSION

Cranial characters

Tecth

Incîsors. The number of incisors of Mayulestes

(I5/i4) is obviousiy plesiomorphic when compa-

red to that of the other borhyaenoids (I4/i3).

Befbre the disetwer}^ of Mayulestes, the réduction

of the incisors number to l4/i3 was regarded as

diagnostic of rhe Borhyaenoidea (Marsh^tll &

Kiclan-Jawmrowska 1992), Among marsiipials,

an incisor formula ol I5/i4, which is alwa)Ti pré¬

sent in didelphids, is regarded as plesiomorphic.

However, Winge (1941) has noted that the II

was not occluding with any lowcr incisor and

suggested that the il had been lost in marsupials.

Fhereforc, the firsr lower incisor is the i2.

Hershkovitz (1982: 186) agréés with rhis inter¬

prétation which is supported by slrong embryo-

logical evidence (Woodward 1893; Berkovitz

GEODIVERSITAS • 1998 • 20(1)

Mayulestesy a borhyaenoid from the Palaeocene of Bolivia

Fig. 43. — Dorsomedial view of the dissected anterlor région of a mandible of Pucadelphys andinus (YPFB Pal 6473) showing the

staggered i3. Scale bar: 2.5 mm.

197S). So far, no marsupials hâve been found

with a complété incisor formula and the plesio-

morphic incisor formula for marsupials is chat of

Mayulestes: 15/14.

Mayiilestes has a staggered 13 {i.e. second lower

incisor). Hershkovir?, (1982) has shown rhai the

second lower incisor (i3) of most polyprocodont

marsupials has a root which is shitted (staggered)

posteriorly and dorsally. Conscqucntly, the ante-

rior alvcolar border of the i3 on the dentary is

thickencd or buttressed. According to

Mershkovitz, that feature is présent in didel-

phids, borhyacnoids, scvcral dasyurids,

Thylacinus, sevcral peramelids and an Early

Cretaceous portion of Icft mandible (FMNH

PM583, Thcria invertae sedis) interpreted (by

Hcrshküvitz) as a metarherian. Contrary to the

statemenr by Marshall &L Muizon (1995: 68),

the i3 of Pucadelphys is actually staggered, as

shown by speamens YI'FB Pal 6473 and 6474

(Fig. 43). The staggered condition could not bc

observed in Pediomys becau.se of ihc .siate of pré¬

servation of ihc specimetis. The staggered i3 is

absent in che Microbiotheriidae and is not obser¬

vable in other South American familles where

hyperspécialisation of the muzzlc hides thaï

condition (Caenolescidae, Groeberiidae. Argyro-

lagidae Peramelidae, and Myrmecobiidae).

Among Lace Cretaceous taxa, the i3 Is apparencly

not .staggered in Alphadoriy Kokopellia^ Eadelphis

browni (AMNH 14149) and Didelphodon

(USNM 2136): (Cifdli Muizon 1997, 1998).

The staggered i‘3 has been regarded as a synapo-

morphy of marsupials (Hershkovitz 1982), but

Muizon et al. (1997) tentativcly regard this cha-

racter as a synapomorphy of the Southern radia*

tion of marsupials (South America and

Australid). The microbjotherüd condition musc

hence be regarded as a reversai.

T he upper incisor niw of Mayulestes is deeply

archcd posteriorly as in didelphids. In the other

borhyaenoids, where that part of the skiill h

known, the upper incisor row is almost straight

and rransverse (in Pharsophorus, Sipaloijoth

Cladasictisy Borhyaena^ Prothylacyaus, Acrocyon),

A straight upper incisor row is a derived condi¬

tion wichin rhe Borhyaenoidea and rhe arched

upper incisor row of Mayukstes is a plesiomor-

phic condition for the superfamily.

PremoLirs. The firsi upper and lower premolars

of Mayulestes arc sÜghtly obliquely set in the

maxilla and dentary as ît îs observed to a much

grcater extent in the Borhyacninac and chc

Prohorhyaenidae. The Prorhylacininae have a pl

strongly oblique ih the dentary but Pl i.s appa-

tcnrly not oblique. Marshall et .ai (1990) have

regarded the oblique implantation of pl as asyna-

pomorphy of the Borhyacnidac, However, the

presence of this feature in the Prohorhyaenidae

{Paraborhyaena) introduces a contradiction in the

cladogram of Marshall et al. ( 1990, fig. 2), as does

chc presence of an oblique PI in the Borhyaeninae

and the Prohorhyaenidae. In face, node 24

(Borhyaenidae) of Marshall et al (1990, fig. 2) is

relatively weakiy supported since che other syna-

pornorphy they use to diagnose the Family is “ani¬

mais of medium to large size”. The presence of

slightly oblique upper and lower first premolar in

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Mayulestes would suggcst that this is the plesio-

morphic condirion vvidiin the Borhyaenoidea and

that it could rcprcsent a synapomorphy of the

superfimiily The loss of the obliqiiity in some taxa

would rherefore hâve to he regarded as an apo-

morphic character State nelated to a k-ngtlieiiing of

the tooth row (Ckdosicth, Sipalocyon). However, it

is notcworthy rhar the ohliqiiity of the Urst lower

premolar in the Borhyacnidae and Proborhyae-

nidae is much more pronoimced than in

Mayulestes and probabJy aiso rcpresents an apo-

morphk trend relatcd ro the acquisition of a very

short and stout rostrum.

Molars. The molars of Mayulestes show a slight

increase in size from Ml to M3 and front ml to

m4. In face, the last- lower molar of Mayulestes is

subequal in height and volume to m3 and only

very slightly longer than m3 (lengrh m3 = 3.70,

length m4 = 3.74). ’l'hc character statc *'trend for

molars to increase rapidly Irorn ml to m4 and

from Ml to M3” ha^ bcen regarded by

Marshall Kielan-Jaworowska ( 1992) as a syna¬

pomorphy of the Borhyacnoidca. Mowever. it is

noteworihy that an increase in size from ml to

m4 and from Ml to M3 is aIso found in the

Stagodontidac and in the Dasyuroidea

(Thyladmis, Sarcophilus). In the Creodonta the

last lower molar is frequchtiy the largest of the

tooth row {HyaeUôdon^ Pterodouy Dissopsatis,

Quercvlheriun}. Cyiiohyaenodon)- This is aiso rrue

in several felids and hyaenids. In faci, the ten-

denc) CO increase the size of the Ust lower molar

and penuliimatc upper molar is a highiy homo-

plastic character State rclated to hypercarnivo-

rous diet. Whatever the function is (shearing,

crushing or botli), the greaiest force is located at

the posteriormost end of the tooth row (rhe cio-

sest possible to the rotation axis of the condyle),

which is probnbly relared ro the rendency of

varions groups i>f mammals to increase the size

of the posteriormost lower tooth. l'hereforc, rhe

phylogenetic value of the character stare ‘Vapid

increaxe in .size from ml to m4 and from Ml ro

M3” is qiiestionahie since it is a highiy homo-

plastic feattire. Furrhcrmorc, a last lower molar

slightly larger than, or subequal in size to. rhe

preceding tooth is aiso présent in the Eatly

Crecaceous eutherian PwkennalesteSy and in some

species of Cimolestes. Therefore, it is probable

chat rhe condition of Mayulestes represents the

plesiomorphic character stare. The same is pro-

bably true for Eodelphis and Pariadens, wiiere the

rn4 is only very slightly larger than the ni3.

The molar morphology of Mayulestes is very

similar to that of Allqokirus from the same locali-

cy (Fig. 44). The holotype of Allqokhus is an

upper molar referred to an M2 or M3 by

Marshall & Muizon (1988). The proportions of

the holot)'pc of Allqokirus (1, = 3.25 mm, W =

3.8 mm, L/W = 0.85) clearly differ from chose

measured on the M2 (L = 3.05. W = 4.34,

L/W = 0.7) and M3 (L = 2.8, W = 4.64, L/W -

0.6) ol the holorype of Mayulestes, Furthermore,

the M2-M3 of Mayulestes difter from the holotype

of Allqokirus in having a much larger scylar

cusp D, a low cre.st descending from the stylar

tusp D loward ihe lingu.il excremity of the mera-

crLsta, a mecaensta which does not strongly over-

haiig rhe base of the crown (as it is observed in

Allqokirus), a straight posecrior edge of the tooth

Fig. 44. — Occlusal views of M3 of: A. Mayulestes: B, AUqo-

kirus. Occlusal views of m3 of: C. Mayulestes: D, Allqokirus.

Lingual views of m3 of: E. Mayulestes: F, Allqokirus. Scale bar =

2 mm.

GEODIVERSITAS • 1998 • 20(1)

MayulesteSy a borhyaenoid from the Palaeocene of Boiivia

(vvhereas il (orms an angle nf approxlmately 150'^

in Allqoktrus)^ a deeper ectoflexus and a more

robust and longer protocone. One m3 bas been

referred to rhe holorype of Allqokirus hy

Marshall & Muizon (1988). The m3 of

Mayulestes diflers from thaï of Allciokims in being

narrower and more slcnder. The lower molars of

Mayulestes resembic ihose ol Allqokirus in mosi

of rhcir srructurc and both arc pcciiliax in having

a rcduccd cnloconid and a lulonid basin opcncd

antcromedially. The Icvcl ol morphological si mi'

laricy existing ber^vccn Mayulestes and AlUjokirus

is similar to lhat existing belween Cladoslctis and

Sipalocyoyt^ two diiîerent genera fouird associated

in die samc localitics Irom die Colhuehuapian

and Saniacruzian beds of PaUgonia.

The borhyaenoid rnolar morphology shows seve-

ral evolutionar)' trcrids vvhicJi represent a func-

rional complex rclaicd to postvallum-prevallid

shear, charactcrisric of hypcrcarnivorous adapra-

cittn (Muizon & Lange-Badré 1997)- On the

lower molars, thé metaconid rcdtices and disap-

pears in niosi généra; rhe paraconid is enlargcd

and crest-like, and the paracristid tends tt) rotate

coiinter-clockwise, being ahnost parallcl to the

toorh row in somc borhyacnîds. somc probo-

rhyaenids and in thylacosmilids: the talonid

rends to reduce and almost disappears in

borhyaenids, proborhyaenid.s and rhylacosmilids.

On the upper molars, the paracone and the pro-

toconc are aiways reduced and almost disappear

in some borhyaenids. proborhyaenids and thyla¬

cosmilids; rhe stylar shelf and siylar cusps B and

D arc gencrally very reduced and often totally

disappear; the postmetacrista is greaily enlatged

and tends to be aligned wirh the mctacone,

parallel tp the tooth row. These features are high-

ly adaptivc and can bc observed (some or ail) in

five other groups of mammals. Dchâihcroida,

Stagodonridac, Dasyntoidea (Thylacinus and

Sarcophilus), Creodonta and Carnivora (Muizon

1994; Muizon & Lange-Badré 1997). Therefore

the functional complex based on postvallum-pre-

vallid shear is a highly homoplasric synapomor-

phy wirh a low phylogenetic value and ic should

not bc reiained a.s a key synapomorpby of rhe

borhyaenoids.

In A'îayulestes, the molar morphology is relacivcly

unspecialised for a borhyaenoid but some of the

features relared to posn^allum-prevallid shear are

already incipicnrly developed: the paracone is

smaller in volume and height iban ihe metacone,

rhe posTmciacrista is enlargcd, the metaconid is

subequal in height to, but smaller in volume

than, the paraconid, and the paraconid is slighrly

blade-like. T'he same features are also présent in

AUqokirus. The molar morphologies of

Mayulestes and AUqokirus are very similar lo that

of Paierie. Mowevcr, the former genera differ

irom the latter in the very small size of their

entoconid and in rhe lingual opening of the ralo-

nid basin, while in Patent rhe entoconid is sube¬

qual in height and volume to the hypoconulid

and the talonid basin is not opened lingually.

The small s-ize of rhe entoconid of Allqokints has

been regarder! as a plesiomorphy within rhe

borhyaenoids by Marsh.ill 6c Kielan-Jaworowska

(1992). A small enroconid is also found in the

Deltachcroida and probably in Acgialodon\ this

cusp is absent in Kieiuniherïuni and hlypçrnyloSy

two Early Crctaceous tribosphenidans (respecti-

vely, Dashzeveg ôc KicJan-Jaworowska 1984 and

Sigogncau-Russcll 1992). Marshall & Kielan-

Jaworowska may be correct .since AUqokirus and

Aîayulestes are ilie oldesi known borhyaenoids

and sincc, in younger genera of lhe superfamily,

the entoconid, when présent, Ls always well-deve-

loped. Howcvci, various Üneages of borhyae¬

noids show a tendency to réduction of the

talonid, which is often achieved by the réduction

of the entoconid and tbe lingual opening ol the

talonid basin [Plesiofilis scblosseri (MLP 11-114),

Noto^alc rnitJs (MNHN SAL 97). Anathertum

herrente (FMNll 13521), Cbasicostylus castrai

(MLP 57-XI-2), an isolated ml (MNfIN

SAL 272) from the Deseadan beds of Salla

(Boiivia), probably referable to a small species of

Phanophorus]. Therefore, since this feature bas

appeared severùl times independenily in several

groups of borhyaenoids, the réduction of rhe

entoconid of Aîayulestes and AUqokirus may very

wcll bave also occurred in the Mayulestidae

{AUqokirus and Atayulestes) and would l)C, rhetc-

forc, a synapomorphy of the family. J'his liypo-

rhesis is tentatively retaincd here. It is

noteworthy rhat the earh'est known marsupial

genera (Kokopellui, from the j^lbian-Cenomanian

of Utah, Pariadens, from the Cenomanian of

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Utah) hâve a well-developed entoconid. A simi-

lar condirion is présent in Anatherium from the

Late Cretaceoiis of Mongol ia.

The apomorphy of the rcduced entoconid of

Allqokirm has been prcviousiy mentioned by

Szalay {1994: 321, 328). If this a-ssumption is

correct, MayiilesUs and AUqnk'irus cannot bc

regarded as probable niorphological ancestors for

Patène (as it was suggesred by Marshall &

Muizon 1988 and Muizon 1992) but could

represent a raorphological dental ancestor for a

lower molar from the late Palaeoœne of Iraborai

referred by Marshall (1978) to cf- Nemolestes

This looth (a posterior molar, m3 or m4) has a

high protoconid, a large cresilike paraconid and

a small metaconid cuspule at the posrerolingual

base o( the protoconid. The talonid is vety redii-

ced, the tilonîd basin is open lingually and the

entoconid îs cither extremcly reduccd or rotally

absent. The genus Ncmnlestes has been regarded

by Marshall (1978) and by Marshall etaL (^1990)

as rhe oldest known représentative ot the

Borhyacnidac. fl these author.s are correct, then

Allqokirm and Mayidesies would not be ancestral

the Hathlyacinidae as previously stated by

Marshall S>C Muizon (1988) and Muizon (1992)

(for Allqokirus only) but could rcpre.sent the si.s-

ter-gfoup of the Borhyaenidac. Hovvevet, this

hypothesi.s stÜI ha.s tu be tested with ihe disco-

very ofcranial reniains of Nemolcstes.

The stylar cusps and srylar shelf of the upper

molars of Mayulesses- and Allqokirus are

well-developed for borhyaenoids. The only orher

borhyaenoid genus whosc upper molars have

well-developed srylar shelf and stylar cusps is

Patene. Mayulestes and Patene rescmble cach

other in having a stylar shelf larger than in the

other borhvacnoids, conspicuous stylai cusps A

and B and no cusp in a stylar C position.

However, on the M3 of the holotype of Patene

simpsoni and on an isolatcd M3 from Itaboraf

(DGM uncaralogued specimen), a sériés of small

cuspules are ohserved in a stylar cusp C position.

Mayulestes differs from Patene in having a largci

stylar shelf, a larger stylar casp B and a conspi¬

cuous srylar cusp D (absent in Patene), A .stylar

cusp B i-5 aiso présent in Prodadasictis from ihe

Mustersan of Paiagonia. The stylar cusps and

stylar shelf of Mayulestes are extremely similar to

those of the Stagodontidae {Eoelelphis and

Didelphodon) which aIso have a large srylar .shelf

with well-developed stylar casps B and D and

which lack a cusp in a stylar cusp C position.

However, a sériés of cuspules in stylar cusp C

position, .similar to that ol Patene, is aiso obser-

ved in some spécimens of Eodelphis and

Didelphodon (Fox 1981). As in Mayulestes^ rhe

ectoflexus of Eodelphis and Didelphodon arc

deep. A well-developed stylar cusp C îs présent

on the teeth referred by Eaton (1993) to

Pnriadens kirklandt from the Ccnomanian of

Utah. This spccîes has been included within the

family Stagodontidae by CifelU &C Eaton ( 1987),

but Eaton (1993) cautiousiy added a question

mark to the familial attribution initially sugges-

ted. Eox Sc Naylor (1995) considcr that the

upper molars referred ro rhat specles by Eaton

(1993) ccrtainly do nor match rhe stagodoncid

morphology since they possess a wcU-devclopcd

stylar cusp C. The Deltatheroida lack stylar

cusp C. but a small stylar cusp D is présent in

Sulestes (Kielan-Jaworowska &c Nessov 1990) and

on the molars of the Gurlin l'sav skull. The

absence of a conspicuous cusp in a stylar cusp C

position is regarded as a plesiomorpitic character

State within mecatherians (Fox 1973; Fox 6c

Naylor 1986; Fox 1987; Marshall et al. 1990;

Eaton 1993). Furthermore, Cifeüi {1993a, b)

and Szalay (1994) have clearly expressed that the

stylar cusp C may have appeared and disappeared

•sevcral times in marsuplals évolution. Therelorc,

Mayulestes, which laclcs a stylar cusp C, piobably

rerains the plesiomorphic condition

In fact, the upper molaj patrern of Mayulestes

approaches the stem marsupial morphology,

according to Marshall et ai (1990). The folio-

wing featurcs are regarded by these authors as

representing the stem marsupial upper molar

morphology: (1) molars trausversely wide and

anieroposteriorly short; (2) cctollexus deep, cen-

trally located along the labial margin; (3) stylar

shelf wide; (4) stylar cusps B and D prominent.

with B > D; (5) stylar cusp C not developed; (6)

stylar cusp A distinct, yet much smaller rhan D;

(7) paracone and metaconc very large, subequal

in .size, set side by side and posicioncd approxi-

mately midway along transverse axis of chc

tooth; (8) centrocrista linear; (9) conules distinct

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

but not enlarged: (10) protocone tall (.spirelike)

and not expanckd, so the angle between the pro-

tocristae being acute (60“); and (11) both antc-

rior and posrerior cingula présent (shelf between

the anrerior base of stylar cusp A and the para-

cône and between the posrerior base of stylar

cusp E and the metacone respectively), This

morphology is based on that of ^'Alphadoff crc'

ber (specimens wirhout sfylar cusp C). Mayulcstes

shows most of ihe features cited by Marshall et

al. (1990), except a posterior cingulum;

Mayulestes also diverges Itom the plesiomorphic

marsupial upper molar pattern in having a para-

cone slightly smallcr than the metacone and an

ectoflexus located slightly anteriorly (the two fea-

cures are probably relateU). The condition of the

Stagodontidae is similar to fhat of Mayu/estes and

Allqokiras. Jn his cladogram of early tribosphenic

mammals, Cifelli (1993a, b) lias retained the

same fcature (présence of a posrerior cingulum)

which hc expressed diffcrcntly: “postprotocrista

of upper molars extends labially past base of

metacone (double-rank postvallum/prevallid

shearing)”. This author regards that fcature as a

synapomorphy of a monophyleric group made

of: (1) the liuthcria; (2) Kakupellia, Tygocuspis

and Falpetrus\ and (3) the Marsupialia [The

occurrence of this characrer has reccntly been

observed in Kokopellia (Cifelli &c Muizon

1997)]. However, a postprotocrista extending

labially past the metacone is absent in many

meiatherians: delratbcruidans, borbyacnoids, sta-

godontids, didelphoids, most microbiotheres,

and dasyuroids. Thercforc, the distribution of

this fcature would indicaie that il appearad seve-

ral rimes in iherian évolution. However, since ir

is ipcipienrly présent in the Palaeocene micro-

biothere {Khasia)j and absent in the Miocene

{Microbiotheriurn) and Recent {Dromiciops)

généra, it is also possible that ir has been reversed

(perhaps independencly) in some lineages

(Muizon et ai 1997). The molar morphology of

Mayulestes is, hence, probably derived in the lack

of a po.sterocingulum

To conclude, the dentition of Mayulestes shows

many plesiomorphic features not only for a

borhyaenoid but also for a rnelatherian. The

number of incisors of Mayulestes (15/i4) is the

primitive condition for marsupials and the cha-

racter State “number of incisors reduced to I4/i3”

is derived for the ocher borhyaenoids. The lack

of a posterocingulum is probably a reversai

wichin marsupials which also occurs in stago-

dontids, borhyaenoids, didclphoids, microbio-

theres, and dasyurids. The incipicntly developed

features related to prevallid/posrvallum shear

observ^ed in Mayulestes hâve been shown above to

be of low phylogejietic value sincc they appeared

independendy in ai least six groups of mammals

and probably several cimes within some of thcsc

groups (Muizon àc Lange-Badré 1997). The

same is rrue for the increase in sîze ftom ml to

m4 and Ml to M3, a feacure probably related to

carnivoroLis dict. Conscquenily, the dentition uf

Mayulesves does not cxhibic undoubted borhyae*

noid synapomorphies smee ail of them arc eiihcr

symplesiomorphies or highly harnoplascic fea-

tures. The only probable derived feature of the

molars of Mayulestes and Allejokirus is the great

réduction nf the enioconid, regarded hetc as a

synapomorphy of the Mayiilestidae. This feature

cercainly appears several times in borhyaenoids

évolution and is therefore of low phylogenetic

value. However, sincc the only éléments of

Mayulestes and Allqokirus that can be compared

are M2 or 3 and m2 or 3, and since it is the only

derived feature shancd by the molars of the rwo

gênera, ir is tentatively retained in spite of its

wcakness as a synapomorphy of rbe family. It is

clcat that crantai remains of Allqokirus are much

needed to clarily this point.

The key synapomorphies of the Borhyaenoidca

cannot bc established on the basis of dental fea¬

tures but must be searched for in cranial mor-

phûlogy.

Bony skull

Mayulestes, a weasel-sized animal, is the smallest

known borhyaenoid. Although, at first sighr, the

rostrum scems to be shorter than in the other

borhyaenoids, the measuremenis in Table 1 indi-

cate that it is even slightly longer. The rostrum

of the Hathliacynidae {Sipalocyon and

Cbdosictis) appears to be sliglitly longer mainly

because of the greater length of the jugal toorh

row and the grcacer narrowness of the palace..

Mayulestes clearly shows a shorter cheek tooth

row and a wider palate between P3 and the ante-

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Tablft 1. — Pfopoflions ot tne lengîh of lhe roslrum to me total lengîP ol iho skum îd borti/aeno^ds Le, total length ot ihe skull form

lhe tip cH the premaxiilae tq the posts'ior eHirêmfty of ihe oociptla» condyles (in YPMPU 150*36 and 15701 Ihe occipital condylas are

missing and the length of lhe sHliH is measured from lhe posîehor border of ine lamodoid cresi The error inlroduced iç. regarded

here as minor); Lr, length of the rostrum trom the anterior border of the orpil. Lctr, length of the cheektooth row from antenor border

of P1 îû posiôfior border of M4 WP3 widih of the patate beiween posierior roots of P3s: WM4. widih of the palate beween lhe M4s:

Winf, widih ol lhe rosTium al the level of tne anienor foramina of me infraorbital canal In Mayulestos WM4. WPS and Winf are

approximate because ot me dorsoventral crushing ot the skutl. Ali measurements are in millimeiers.

Lctr

WP3

WM4

Winf

Lr/Lc

Lctr/Lc

WP3/LC WM4/lr Winf/Lc

Mayulestes (MHNC P 1249)

11.5e

14.8e

0.37

0.31

0.21

0.274

Borhyaena 1 (YPM PU 15 701 )

230

78.3

0.33

0.34

0.16

0.32

0.252

Borhyaena 2 (YPM PU 15 120)

195

67.2

71.3

63.5

0.33

0.35

0.14

0.36

0.251

Prothylacynus (MACN 5931)

171

52.8

43.5

0.31

0.16

0.28

0.254

Cladosictis (YPM PU 15170)

158

57.5

15e

0.36

0.33

0.09

0.205

0.154

Cladosictis (YPM PU 15046)

142

15.5

—

0.36

0.11

0.21

0.151

Sipalocyon (AMNH 9254)

112e

39e

38.5

14.5

21.5

0.348

0.34

0.13

0.196

0.151

rior opening ot the infraorhiral canal. Further-

more, in the Santa Cru/ borhyaenoids, the grea-

ter concaviry f>t the latéral side of the inaxilla, in

the région of the anterior foramen of the infraor-

bital canah. conrribures to the narrowness of the

rostrum and to ics apparent lengili. Tn

Mayuleste^^ the rostrum is not constricted at irs

base. Therefon;, the rostrum of Mayiilestcs is rela-

tively long (for a borhyaenoid) and robusi but

the palaie is wider and shorter ïhan in ihc

Santa Cnir borhyaenoids.

The general morpholog)' of the .skull is close to

that of tlie Rorhyaenidae, From wlucli il diflfcrs,

howeven bv iis much wiiler interorbiial bridge.

The absence of a supraorbital process is prob.ibly"

a plcsiomorphic character since chis structure is

absent in sevcral I.ate Cretacéou.s eutheriaiis

from Mongolia [AsiorycîtSy Bdrwilestes^ Kenna-

lestes), and in the deltatheroidan skull from

Gurlin Tsav (Kielan-Jaworowska & Nessov,

1990; Szalay ^ Trofiniov 1996), Ir is absent in

sevcral didclphids. in caenolestoids. in severaJ

dasyiirids and in pcramelids, Among the other

borhyaenoids, a distinct supraorbital process is

also Jacking in Borhyaenii airhough, in rhis

genus. the frontal bridge is greatly widencd bet-

ween the orbirs.

As in ail ihc other borhyaenoids and in ail gene-

raliscd fossil and living marsupials, Mayulestes

retains the plcsiomorphic condition of a large

orbit confluent with the temporal fossa.

The nasals of Mayulestes aie long and postcriorly

flared. They hâve a broad contact with the lacri-

mals, a plesiomorphic coadiiion found in qmo-

donis, iritylodoniids, Vincelestcs (Bonaparte &

Rougier 1987), Morgauucodon fKcnnack et al.

1981), Siaoenvodon (Crompron 1 uo 1993),

Haldanodon {l.illegraven 6c Krusat 1991),

Deltatberidium (Kielan jaworowskii l97Sa), and

Ashitherimn ( IVolnnov 6c Szalay 1994; Szalay &

frofimov 1996). Among marsupials, a

nasal-lacrimal coniact is présent in ail the

borhyaenoids (excepi rhylacmmilus) and In

Wynyardia. Tht derived character siate is a

rnaxilla-lrontal contact which .séparâtes nasal and

lacrimal. The condition of Mayulestes is a plcsio-

morphy within mammals. *1 he nasolacrimal

contact in Wynyardia (Gregory 1920) and the

lack of contact in Thylacosmilus are regarded as a

reversais. Tn the latter, this' reversai is due to the

hyperdevclopmeiu of the maxillae rclatcd to the

development of sabre-like caniiic.s.

The lacrimal oF Mayulestes bas a large Facial

wing, a plcsiomorphic contlition loiind In cyno-

donts, IJaldanodtm (Lillegraven & Kfusat 1991),

Morgamicodtm (Kermack /?/. 1981), l^inocono-

don (Crompton & l.uo 1993), Vï}icelestes

(Bonaparte & Rougier 1987), DeltitthcYidlum

(Kielan-Jaworowska 1975a) and Puiadelphys

(Marshall & Muizon 1995). f'iinhcrmore, the

cxternal lacrimal foramen is doublcd and opens

within the orbit. This condition is plcsiomor¬

phic and ir is présent in tynodoins, MorganU'

codofiy Vmcclestes, Deltathertrlium .rnd in the

other borhyaenoids. This contlirinn Is prescrit in

most didelphids, except in Didelphis where the

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of BoÜvia

external lacrimal foramina opens dorsolaccrally

just antenor to the antcrior extremity of the

orbir. In Thylacinm the external lacrimal fora¬

men is doublet! and one foramen opens inside

the orbit while the other opens clearly outsidc of

the orbir anteroventrolaterally. The condition of

Mayulestes plesiomorphic within n'iammals.

Mayulestes has no palatal vacuiries as is observed

in the other borhyaenoids. The presence of pala¬

tal vacLiities shows an important variation within

mammals. l'hcy are absent in cynodonts,

Hnldanodon (Lillcgraven &: Krusat 1091), inor-

ganucodontids (Kcrmack et al. 1981), Shwcano-

don (Crompton ^ Luo 1993). soine multituber-

culates {Knwf)tohaatai\ Chulsanbaaîar, Limhdnp-

sallis, faeniotahis), Vhjcelestei (Bonaparte ^

Rougier 1987). Deltatberidiiim (Kielnn-

Jaworowska 1975a) and in many marstipial.s

[Pucadelphys^ CalurornySy Sparassncynnsy Dasycer-

ctiSy Dusyunnde^y some spccics ol Swinthopsis atid

AnterlfhmSy MyrmecobiuSy DactylopsUay Petaurus

and Dactylonax (Marshall ]979b)J. Thcy are

generally absent in eutherians (except leporid

lagomorphs, some rodenis, macroscelids, erina-

ccids, and Carpoleites). Palatal vacuities are also

présent in some mulrituberculatcs (hfemegt-

baatat\ Sloanhantary Bidganhaatar^ Ptilodus)^ in

the Deltatheroida !rom Gurlin Tsav (Kielan-

Jaworowska &: Nessov 1990; Szalay & Trofimov

1996), in Asiatherium (Trofimov &: Szalay 1994;

Szalay & Trofimov 1996) and in most living and

fossil marsupials (Marshall 1979b: Reig et al.

1987). Pox & Naylor (1995) recently observed

the présence ol palatal vacuities in siagodoniids

{Eodelphh m’xA DidetphoJon) and iv\ Alphndon.

The palatal vacuities arc generally regardcd as

plesiomorphic lor marsupials (Tyndale-Biscoe

1973; î'ox & Naylor 1995). flowever, Marshall

(1979b), Archer (1982) and Marshall & Muizon

(1995) stated that, since the maxillae and palati¬

ne hone in the devcloping skull of marsupials

werc generally not fenestrared [Parker (1886)

noted thaï the fenestration occurs later in onro-

geny by bonc résorption], a solid palate was pro-

bably the plesiomorphic State for mammals. The

high degree ol variability in the presence or

absence ol palatal vacuities in mammals, as well

as embryological observations, suggests that

these structures probably appeared several rimes

independcntly in axtd within each group. which

therefore significantly reduces their phylogenetic

value. Mayulestes and Pucadelphys retain the ple-

sioniorphic condition for mammals.

The prerygoid and posrpalatine région of the

skull of Mayulestes is significanily different from

that of the other borhyaenoids, The choanal

fossa of Mayulestes is relatively widc and short

anreroposteriorly. les walls are formed essentially

by che pterygoids, rhey are high and thii). The

pferygoids hâve a large, hook-like hamular pro-

ce.ss which overhang (in ventral view) the basis-

phenoid and the alisphenoid In the Descadan

and Santacruzian borhyaenoids (Mayulestes is the

only pre-Deseadan borhyaenoid whose complété

-skull is known), the choanal fossa is long and

narrow anteroposreriorly. Its walls are low, thick

and formed by rwo bony layers, on the one hand

the prerygoid medially, on the other hand the

palatine arucrolaterally and the alisphenoid pos-

ccrolatcrally. The latéral side of the \va!l shows a

thick anteroposteriur buttress. The bottom ol

rhe choanal fossa is rooted by the pterygoid. I he

hamular proce.sscs of the pterygoids arc cither

lüst or very rcduccd. As a conséquence of that

mocphology, the plane ol ihc palate pa.sses

smoorhly ro the basioccipital without rhe sirong

ditlcrcncc of Icvcl observed in Mayulestes bccau.se

ol the présence ol very saÜent hamular proces.ses.

Therian pterygoids arc fragile bony platc.s sel-

dom preserved in fossils and unknown in

Palaeogcnc or Crctaccous marsupials (except

Mayulestes). Tn rhe Üving didelphids, rhe ptery¬

goids arc very fragile small b(jny blades (often

losi during préparation) which still retain a

hamular process, although much sniallcr than in

Mayulestes. Well-developed pterygoid laminac

and hamular processes arc presenr in Barunlestes

(Kiclan-Jaworowska & Irofimov 1980) and

Astotyaes (Kielan-Jaworow.ska 1981). The condi¬

tion observed in these Lare Crctaccous euihe-

rians suggests that a pterygoid with well-

dcveloped ventral lamina and hamular process

probably represems the plesiomorphic condition

for therians. Therefore, Mayulestes retains the

plesiomorphic condition: it is more plesiomor-

phic than the other Borhyaenoidea and the

Didelphidae. As stated by Muizon (1994), the

loss of the hamular process of the pterygoid is a

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

synapomorphy of the orher borhyaenoids.

However, ir is noteworthy thar a morphology

similar to chat of the other borhyaenoids is pré¬

sent in Thylaciniu airbough, in this genus, tbe

pterygnid plate is liigher.

The jugal of MayulnteSy as in fossil and living

marsupials, reachcs chc glenoid fossa posreriorly

and torms a prcgienoid process which receives

part of the articular surface. This condition is a

plesiomorphy for therians (Marshall (Se Muizon

1995) which is al-so found in Vincclestes^ the dcl-

tatheroidan skull from Gurlin Tsav, and somc

eutherians. In spitc of ics plcsioniorphic nature

this feature was retained by Marshall 6c

Kielan-Jaworowska (1992) as a synapomorphy of

the Metaiheria (including Dcltathcroida).

The alisphcnoid ot Aiayulestes bas a large suture

with the pariétal. This character State is présent

in the Late Cretaceous eutherians Asioryctes and

Kennalesm and in the eupantothcrc. Vincckstes,

Among marsupials, il is présent in fossil and

living didelphids (including PiHadclphys)^ in

somc borhyaenoids [MayulesteSy Sallacyoriy

Sipalocyon (Archet 1976), Notogale (MNHN

SAL 271), Paraborhyaena (MNHN SAL 51)]> in

myrmecobiids and most dasyuriJs (Archer

1976). It is absent in somc borhyaenoids

{Borhyaetia-, Prothyliicynus)y rliylacinids, perame-

lids, vombatids, in some phascolarctids and some

dasyurids (Archer 1976) Mayulestes therefore

présents what is regarded here as the plesiomor-

phic State for the Thcria.

The alisphenoid of Màyidestes makes a small

contribution to the anteromedial angle of ihe

glenoid fbssa [entoglenoid process of the alisphe¬

noid of Clemens (1966: 73)]. A much larger

contribution Is also présent in ail didelphids, in

caenolesioids, in microbiotheres, in peramelids,

in dasyuroids (smaller), in some perameloids and

in several phascolarctoids (smaller). The alisphe¬

noid docN not participate in ihe glenoid fossa in

the other borhyaenoids, in stagodontids, in

Hondadelphys and in most phalangeriforms. [n

Vincelfstesy the alisphenoid dues not contact the

glenoid fossa of the squamosal and the feature Is

therefore Lrrelevant in this genus. However, it is

noteworthy thac, in ViucetesTes, a similar partici¬

pation to tlic antcromedlal angle of the glenoid

fossa is achieved by the anterior lamina of the

periotic. It is interesting u> note hcre chat the

anterior lamina of the periotic bas bcen regarded

by Presley &C Steel (1976) and Presley (1981) as

homologous with the hiade of tho alisphenoid. A

participation of the alisphenoid to rhe anterome-

dial angle ot the glenoid fossa is interprétée] here

as a plesLomorphic character State within marsu¬

pials. which disappears indcpendently in several

Jineages. Mayulestes rctains the plcsioniorphic

condition within the borhyaenoids and marsu¬

pials. In Rccenr didelphoids. the alisphenoid is

pcrloratcd by the foramen rotundum, the large

foramen ovale, the entocarotid canal, and, when

présent, rhe rransverse canal.

Thcrc is no transverse canal in Mayulestes. This

siTucrure is also absent in somc borhyaenoids

(Sipalotyon, Borhyaena and Prothylacynus).

However, Marshall (1977b; 639) noted in

Lycopsis a '*ciny foramen [...] which appears to

represenr a rudimentary rransverse canal” and a

probable transverse canal is. présent in Notogale

(MNHN SAL 271) from Salla-Luribay (Bolivia)

and in Cladosktts (YPM PU 15705). Therets no

rransverse canal in morganutodonrids, multitu-

bcrculates, Deltatheroida, in rhe Late Cretaceous

eutherians from Mongolia, in some didelphids

(Caluromys)i in some dasyurids (some species of

PUnigate)y A rransverse canal is présent in most

didelphids, most dasyurids, myrmecobiids, per¬

amelids and thylacynids. On the basis of the

important variation in its srze and morphology,

Marshall &c Muizon (1995: 71) hâve stated

{contra Archer 1976) thaï the lack of rransverse

canal was likdy to be a plesiomorphy for marsu¬

pials and ihat this structure probably appeared

several tlmes du ring marsupial évolution. The

absence of tcansverse canal in Mayulestes would

support this statement.

l'he term foramen ovale requires some discus¬

sion. In this Work it is used to designate the fora¬

men which transmics the mandibular branch of

the trigeminal nerve without considération of

the bones surrounding it {sensu Kielan-

jaworowska et al. 1986). In Vincclestes (Early

Cretaceous cupaniothere), the foramen ovale

pierces the anterior lamina of the periotic, in

Pucadelphys and Mayulestes (early Palaeocene

marsupials) it is lirniied by the alisphenoid ante-

riorly and by the periotic posteriorly, in

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Asioryctes and Kenruilestes (l.ate Cretaceous

eutherians) it only pierces tKe alLspheaoid. The

désignation toramcn pscudovalc has a variable

définition according to authors (Macintyrc

1967; Archer 1976). Mclntyic (1967) cailcd

loramen pscudovalc in eutherians the foramen

rcsulting in rhe hision ol the ttue foramen ovale

(totally enelosed by rhe alîsphenoid and rhe fora¬

men lacerum medium). For Archer (1976,

1982), the foramen pscudovalc in rnarsupiaU is

the foramen called foramen ovale in this study,

i.e. limited hy the ali.sphenoid anteriorly and by

the perioiic posteriorly and rhrough which the

V3 nerve exits the skull (Archer 1976. 1982). In

fact, the plesiomorph condition tor marsupials is

very probably that ol Pucadclphys and Aiayt4lestes

(which is aiso présent in rccently discovered

skulls of Andinodelphys from ilic carly PaJacoccnc

of Tiupampa). The derived condition lor marsu¬

pials (found. lor exarnple, in Didelphh) h rhe

formation ol a short canal (the tanal ot rhe fora¬

men ovale) in the alisphenoid which totally

encloses the V3. The postcrodorsal opening of

the canal is the primitive truc foramen ovale

found in Pucade/fihys-And Mayulestes. It is clcarly

observable on a cérébral view ol ihe basitranium

of Didelphis. Since it rcpreseais the plesiomorph

conditioa 1 suggest not to give it the name pseu-

dovale. The anieroventral opening of the canal is

a secondary fortnaiion wliich, în fact, would bei-

ler deserve rhe name pseudovale than the poste-

rodorsal opening. In order to avold confusion

with Mclnryres foramen pseudovale (fusion of

foramen ovale and foramen lacerum medium),

the name of secondary foramen ovale is more

appropriarc for the anteroventral opening of the

canal of the foramen ovale (Wroe 1997).

The conséquence of rhe formarion of a canal of

the foramen ovale is a superficial séparation of

the foramen ovale from the foramen lacerum

medium. This condition is probably indepen-

denr of the development of a lyrripanic process

of the alisphenoid since Prothylatynns has a canal

of the foramen ovale and a secondary foramen

ovale alrhûugh chis genus did not de\'elop a lym-

panic process of the alisphenoid. The lack of the

tympanic process of the alisphenoid and a fora¬

men ovale which opens hetween the alisphenoid

and the periocic are regarded here as plesiomor-

phlc conditions in Mayulestes and Pucadelphys,

This interprétation is reinforeed by rhe fact that

Hopson & Rougicr (1993: 289) scated that tlic

didelphid condition of the cmhryological deve¬

lopment of the alisphenoid and les relationships

witli the three branche.s of the trigeininal nerve

represents the plesioniorphic condition for mar-

supiais and, by extension, for ail living rherîans.

In Mayulestes rite foramen lacerum medium is pos-

sibly confluent with the foramen twale. a condi¬

tion ohseived in Barhyaena and the dasyurids

(Marshall 1977a). which, therefore, would possess

a tfuc foramen pseudovale (re?nu Mclntyre 1967).

This condition is absent in the Didelphidac and

in Pucadelphys, and if actually présent iti

A{ayule.aes is regarded hcrc as apomorphic.

Aiayulestes has no tympanic process ot the alis¬

phenoid. The absence of this structure in some

borhyacnoids {Borhyaena, Prothylacynus, Lycopsis)

and in Pucadelphys has been regarded as a deri¬

ved condition (Marshall & Kielan-Jaworowska

1992). However, Muiy.on (1994) has suggested

thaï ihc absence of ali.sphenoid huila in the

oldest known skulLs of borhvaetioid (Maytilestes)

and didelphoid (Pucadelphys and Andinodelphys)

scems TO indicate that rhe condirion in rhe three

généra is ple.siomorphic, thus indicating that a

tympanic process of rhe alisphenoid evolved

several tinies independcntly during marsupial

history. There is no tympanic process of the alis¬

phenoid in Borhyaena., Prothylaeynus^ Lycopsis^

Sailacyon (possibly) Paraborhyaena and Thyla-

cosmilus. 1rs occurrence in the généra Cladosictis,

Stpalûcyon and Motogale (MNHN S AL 271) is

regarded hcrc as a synapomorphy of the

Hathliacynidae (Muizon 1994).

Mayulestes does not liave a fosrral tympanic pro¬

cess \scmu Wihle (1990) = tympanic wing of die

petrosal part of the periotic sensu Archer

(1976a)] However, the smull cubcrcle anteroven-

rral to the fenestra cochleae is probably homolo-

gous to rhe rostral tympanic process. A condirion

simil.ar to that of Mayulestes is observed in

SîpalocyoH and Protbylacynus (nor in Cladosictis).

Wible (1990: 199) observed in Borhyaena a

“ridge rcscmblitig lliat ol petrosal Type A” (a

LaCc Cretaceous petrosal from Bug Crcck

Anthills, Montana). In an undescribed basicra-

nium (MNHN SAL 271) from the late

GEODIVERSITAS • 1998 » 20 (T)

Muizon C. de

Oligocène of Salla Lunbay (Bolivia), referred to

Notogaky a cicar ridge is présent on the médial

side of rhe promontorium. In Pamborhyaena, the

periotic hears a srrong venrrally projecring pro-

cess but, becaiise of the important modifications

of the auditory région of lliLs geniis (probably of

the family), ic is not certain that it is homologous

tü the rOsStral tympanic process of tither marsii-

pials. Pucadelfyhys^ an early Palaeoccne didel-

phoid, has a snaooth promonrorium while ail the

other members of the superfamily bave a rostral

rympanic process. A rostral tympanic process is

présent in monotretnes, mutiruberculates, most

marsupials and in sonie eutherians. The homolo¬

gies of that structure are not .simple, since, as

mentioned by Wîble (1990: 199jr processes ou

the promoncoritim rcsult from several different

ontogénies in Récent inammals. This aurhor

regards the lack of tympanic process as a ple.sio-

morphy and concludes that “rostral tympanic

processcs of the petrosal hâve cvolved îndepeit-

dently a number of limes wirhin thèse mamma-

liait taxa^ (Wible 1990: 199). Therelote. Wible’s

assertions are corroborared by the morphologies

of Mayulestes and Ptuadelphys (which hâve no

rrue rostral tympanic proce.ss) and hy the obser¬

vation of an undoubted rostral iym[>anic proces.s

in Nowgak (MNHN S AL 271)* which démons-

trates tliac thls vStruccurc appeared ai least rwice

iiidepcndcntly in marsupials. This is not surpri-

sing, since the presence of a rostral r>"mpanic pro¬

cess is probably ac least partially related to the

presence of a tympanic process of ibe alisphenoid

and/or of an alisphenoid hypocympanic sinus

and, as stated above, the alisphenoid process is a

struciurc ihai is Hkely to hâve evolved indepen-

dently several cimes wichin marsupials.

The cpitympanic recess and the alisphenoid

hypocympanic sinus are structures which require

spécial comments, The epirympanic recess is rhe

“extension of ihe niiddle ear cavicy which lies

dorsal to the tympanic membrane and contains

the mallear-incudal articulation’ (Wible 1990:

188; sce aiso Van der Klaauw 1931: 73; Archer

1976: 226). rite posterior extremity ot the cpi¬

tympanic recess is tbc (ossa incudis or fossa crus

brève incudis, a deep and narrow pii wherc the

ligament of the crus breve of the incus attaches.

In ail marsupials, except Pucadelphys and

Andinodelphys, anterior co the epirympanic recess

is a bony sinus excavated in the alisphenoid and

floored by the tympanic process of the alisphe¬

noid (absent in several borhyaenoids). It is the

alisphenoid hypotympanic sinus. In rhe didel-

phoids (wliicli arc commonly regardcd as bcaring

the basic picsiomorphic pattern for living marsu¬

pials), the epiiympanic recess and the alisphe¬

noid hypotympanic sinus are separated by the

petrosal crest. The posterior slopc of the crest is

excavated by the epirympanic recess and the

anrerior slopc by the posterior part of the alis¬

phenoid liypon mpanic .sinus. I hc roof of the

sinus is formed by the alisphenoid. The poscero-

dorsa! horder of the roof abuts against the anté¬

rolatéral boaler of the periotic at the base of the

antetior slopc of rhe pcinîsa! cre.st. l'hc resuir is

that the posterior part of rhe roof of the sinus is

formed by the periotic. The caviry of the periotic

of Diddphis ifirginiana. narned cpitympanic

recess by Wible (1990: fig. 4A) is in lact ibe pos-

cerior extremity of the alisphenoid hypolympa-

nic sinus (see above). The same is probably true

for the periorics illustrated in his figs 2H 5B and

D. The cpitympanic recess i.s locaied posterior to

the petrosal crest whith is clearly observable in

his figs 4A, 5A and C.

dhc alisjïhcnoid hypotympanic sinus of

Mayulestes is made of three componenrs: the

petrosal, the alisphenoid, and the squamosal

(Fig. 45 and see description above). The partici¬

pation of tht squamosal to the sinus has been

noted bv Archer (1976) in three other borhyae¬

noids [Sipidocyou. ProthyLuynus and Hnrhyaena).

1 hâve aIso observed it in these three généra as

well as in Cbidosictis (Mui/on 1994: 210, rontra

Archet 1976: 292), in SaHacyofu in Notogale

(MNHN SAL 271) and in ParaborhyaenarT\\c

portion of the squamosal involved partially

{Mayulestes) c»r rorally (other bijrhyaenoids) in

the formation of ihc .sinus is whal has been

named by Muizon (1994) the mcdial piocess of

the squamosal (Fig. 45). A participation of the

squamosal to the con.struction of the alisphenoid

hypotympanic sinus Ls apparently absent from ail

rhe other marsupials. Pucadelphys andlnus^ a

didclphoid Irom the early Palacoccne, aiso has a

médial process of the squamosal but, in this spe-

cies, there is no alisphenoid sinus (Fig. 46). A

GEODIVERSITAS • 1990 • 20(1)

MayuUstes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 45. — Audilory région of: A, Mayulestes-, B, Sallacyon; C, Cladosictfs (from Muizon 1994, modified). Abbreviations: aca, anterior

crest of the alisphenoid hypotympanic sinus; AL. alisphenoid; ap, alisphenoid portion of the alisphenoid hypotympanic sinus:

are, articular rjdges o! the squamosal for articulation of the ectotympantc; dwfo. dorsal walf of the foramen ovale; ef. enlocarotid

foramen; epa, entoglenoid process of lhe alisphenoid; or eprtympanic recess; fc, teneslra cochleae; fo. loramen ovale; fi, fossa

incudis; gf. glenoid fossa; K. latéral trough; mp. mastold process; mps. media) process of tlie squamosal; pap. parocciplta) process,

pgf. postglenofd loramen: po» promontorium; pp, petrosai portion of the alisphenoid hypotympanic sinus: pfc. petrosal cresl:

sica. suicus for the Internai carotid ariery: sjv. sulcus for the Internai jugular vein; sp. squamosal portion of the alisphenoid hypotym¬

panic sinus; SQ. squamosal; tpa, tympanlc process of the alishenoid. Diitted line on figure 4SC indicale the eKtension of the mediai

process of the squamosal hidden by the tvmpanic process of the alisphenoid. Not lo scale.

similar condition is présent in Andinodelphys

(Muizon et al. 1997). In Aiayulestesj a generalised

borhyaenoid, the participation of the alisphenoid

to the sinus is small whüe the periotic and the

squamosal portions are much larger (Fig. 45). In

younger borhyacnolds [Sallacyon, Notogale,

Sipalocyon and Cladosicns)^ the enlargement of

the sinus is maiidy due to the increase in size of

the alisphenoid participation. In the borhyae-

noids which do not hâve a tympanic process ot

the alisphenoid, the alisphenoid sinus is opened

ventrally and deveJops dorsally and anteriorly

within the latéral wall of the skull. It may be

small {Prothylacytius), medium-sized {Sallaiyony

Borhyaena) or very large (Paraborhyaend).

Prothylacynuss Sipalocyon. Cladosictis and

Parahorhyaena hâve a canal of rhe foramen ovale

and a seconday foramen ovale> torally surroun-

ded by the alisphenoid, conrrary to the condi¬

tion of Mayulestes. In Borhyaena. rhe condition is

similar to thac of Mayulestes since rhe foramen

ovale is probahly confluent with the foramen

lacerum medium, therefore partially bordered by

rhe periotic. In fact, it is possible to dérivé most

of lhe borhyaenoid alisphenoid morphologies

from a Mayulestes morphotype, which probably

represents the plesiomorphic condition. In

Sallacyon. rhe sinu.s is deeper than in Mayulestes

but the organisation of its components (alisphe¬

noid, squamosal, and periotic) is basically similar

lo thaï in Mayulester. the alisphenoid hypotym¬

panic sinus is excavated anterodorsally in the

mediai procc.ss of the squamosal, posieriorly in

the periotic and medially in the alisphenoid.

flowever, bccause of the inadéquate préservation

of the only known specimen, ihc absence of a

tympanic process of the alisphenoid and the

condition of the foramen ovale are uneertain. In

Borhyaena and Prothylacynus, the alisphenoid

sinus develops anccrodorsally and “pushes” the

mediai process of the squamosal within the alis¬

phenoid, the anterior part of the sinus is still

excavated in the mediai process of the squamosal

which is underlain by the alisphenoid; in

Borhyaena. the foramen ovale probably has the

same pattern as in Maytdestes [the only specimen

(YPM PU 15120) available during this study is

partially broken in this région of the skull]. In

Parahorhyaena. the sinus fiirtlier develops (ante-

rodorsomcdially) as a conical cavity which dee-

ply pénétrâtes the alisphenoid lar anterior to the

mediai process of the squamosal; the latter is

totally internai to the sinus and forms its posté¬

rolatéral wall internally; the anteroventral part of

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

the sinas is excavated in the aliyphenoid; the pas¬

sage of rhe V3 is cotally enclosed within the alis-

phenoid and there is a secondary foramen ovale.

Thylacosmilus also has a large, anterodorsally

developed alisphenoid sinus as evidenced by a

riibber endocast of the middie ear caviry of

FMNH P 14344 (Turnbuli ài Scgall 1984,

fîg. 7). Ijî Sipalocvon. Cladosictis and Notogale,

the portion of the alisphenoid anterior to the

médial protcss of rhe squamosal devclops vcn-

trally and postcriorly. underlies it and projects

posteriorly in a well-dcvclopcd tympanic process

which floors the sinus: the ventral and most of

the dorsal parts ol the sinus are cxcavaccd in the

alisphenoid; the antérolatéral part of the sinus is

cxcavatcd in the squamosal. f'he période portion

of the sinus is small when coniparcd to the alis¬

phenoid and squaniosid portions, ‘l'hcrc scenis to

be, rhcrcforc, tw'o transformation patterns to the

increase ol the alisphenoid sinus in the borhyae-

noids: ( l) an anterodorsal expan-sion which exca-

vates the alisphenoid; (2) the postcroveniral

development of a tympanic process of ihe alis¬

phenoid. Therefore, strictly spcakîng, ihe sinus

of Mayitlestes and borhyaenoids of rhe firsi pat¬

tern, which is situated above the tympanic mem¬

brane, is an alisphenoid epicympanic sinus while

that of die borhyaenoids of ilie second pauern,

located mosdy below the tympanic membrane, is

physically consistent wich rhe lerm alisphenoid

hypotynipanic sinus. Furthermorc, the term alis¬

phenoid is not very appropriare in the case of the

borhyaenoids sincc rhe sinus is never excavated

exclusively in tbi.s hone, in Mayukstes the alisphe-

noid portion of the sinus is even the smallesr ol

the threc. HcAVcvcr, as mentioned above, this term

is commonly used by authors (Van der Klaauw

1931; Aj'cher 1976; Petter & Hoffstetter 1983)

and it is retained berc in order to avoLd confusion.

Pucadelphys nvdiims (early Palaeocene), the

oldest didelphoid whose skull is' known, dues not

bave any audirory siniises, a condition which has

been regarded as plesioniorphic for marsuplals

(Marshall & Muizon 1995)- An alisphenoid

hypotynipanic sinus is pre.sent in ail the other

marsupials. Contrary to statements by

Marshall & Muizon (1995), the sinus is not

absent in some borhyaenoids; it is only small (or

reduced) in Prothylacynus and Lycopsis.

Pucadelphys and MayiUestes (as expressed above)

are respectively the most plesioniorphic marsu¬

pial and the most plesiomorphic borhyaenoid for

this character State.

Comparison of the alisphenoid hypocympanic

sinus of MayuUstes to thac of other borhyaenoids

rcvcals scveral evolutîûnary trends of die superfa-

mily: (1) incrcasc of the size of the alisphenoid

hypotynipanic sinus; (2) increase of rhe size of

the alisphenoid portion ol the sinus; (3) tenden-

cy to cover the sinus vcntmlly either by antero¬

dorsal pénétration within the alisphenoid or by

postcrovcntral development of a tympanic pro¬

cess of the alisphenoid; (4) tendency to isolate

rhe foramen ovale from the foramen laccrum

medium and m cncJo.se the course of the rnandi-

hul.ir nerve within the alisphenoid.

As noted above, the borhyaenoids, Pucadelphys

(Fig. 46), and Andinodelphys (Muizon et ai

1997) hâve a conspicuous médial process of the

squamosal, a media! prolongation of die glenoid

fossa of the squamo.sal which contacts die pério¬

de or die alisphenoid anterior to the epiiympa-

nic recess, which reaches (almosr in Mayulestes)

the latéral liorder of the foramen ovale medjally

(this fearure disappear.s secondarily when the

alisphenoid encloses the foramen ovale), and

wliich pardcipaies to ihe formation of the alis¬

phenoid hypoiynipanic sinus in die former. AJl

hoihyaetioid.s hear such a process (Mayulestes,

Notogalev Slpalocyon, Cladosictis, Borhyaena,

Prothylacynus, Lycopsis, Paraborhyaena). This fea-

tiire apparently absent in Didelphodon

(UCMP 538%; Clemens 1966) and Eodelphis

(AMNH 14169) (Matrhew 1916), I hâve no

indication on the condition in dic DcJtaihcroida

fiotn Gurlin Isav, and che holoiypc of

Asiatheriuîn is too crushed to allow its observa¬

tion. A médial process of üie squamosal is absent

in the other marsuplals, in Morganucodon

(Kermack et ai 1981), in Sinoconodon

(Crompton & Luo 1993), in mulrituberculates

(Kielan-Jaworowska et al. 1986; Miao 1988,

1993)^ in Vinedestes (Rougier ef æ/. 1992), and in

Asîoryctes frvim the I-ate Cretaceous of Mongolia

(KLclan-Jaworowska 1981). Thercforc, this

feature is probably a synapomorphy within mar¬

supials.

The presence of an alisphenoid hypotympanic

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Fig. 46. — Auditory région of Pucadelphys (photo of right side of YPFB Pal 6110) to show the médial process of the squamosal.

Scale bar: 2 mm.

sinus partially formed by the squamosal has been

regarded as a synapomorphy of the Borhyaenoi-

dea and che présence of a niedial process of che

squamosal has been regarded as a synapomorphy

of Pucadelphys + Andinodelphys t Borhyaenoidca

(Muizon étal. 1997). However, Pucadelphys

been referred to the Didelphidae bv Marshall &

Muizon (1995) on the basis ot its dental mor-

phology. If this assignment is correct, then

Pucadelphys \a the sister-group of the other didel-

phids in which the médial process has been lost.

But one could aiso question the value of the den¬

tal chaiacierA* used to reftr Pucadelphys to the

Didelphidae. [n fact, the maior fearurcs conside-

red by Marshall & Muizon are rhe V-shaped cen-

trocrisra and the metacone larger than the

paracone. These fearures hâve been seriously

questioned by Cifelli (1990a: 315; I99(Jb: 328)

as characters unique co didelphids since this

author demonstrared thar they verv probably

cvolved independently several cimes among

Norrh and South American marsupials (see aIso

Marshall et al. 1990, fig. 4). Furthermore, a.s

noted above, Pucadelphys does not hâve the

major tarsal synapomorphies of didelphids: the

loss of the calcaneofibular (CaFi) articular facet

which is clearly présent in Pucadelphys and the

presence ol a proximal calcaneocuboid (CaCup)

facet. Iv is therefore probable that Pucadelphys

andinus is- not a didelphid as stated by

Marshall & Muizon (1995) but belongs to a dis¬

tinct clade (sister-group of che borhyaenoids)

and which independendy acquired a V-shaped

centrocrista (a metacone larger chan the para-

cône is aiso found in borhyaenoids) and a

wcll-developed srylar cusp C (abs'enr in the

borhyaenoids). Attribution of Pucadelphys andi¬

nus to a new faniily uvuld explain the presence

in this species of some plestomorphic feacures,

unique among marsupials [lack of auditory

siniises, lack of t}'mpanic process of che alisphe-

noid (shared wirh Mayulestes), presence of a

small anterior lamina of che perioticl and absent

in ail didelphids. However, if Pucadelphys is

actually a didelphid (or belongs to the didelphid

sister-group), ic is aiso possible rhat the develop¬

ment of an alisphenoid sinus in the didelphkls

was achieved with exclusion of the médial pro¬

cess of the squamosal which got reduced because

of the increase in size of the alisphenoid in that

GEODIVERSITAS * 1998 • 20(1)

Muizon C. de

région oi thc skull. In chc borhyaenoids, thc

development oF the alisphenoid sinus lias inclu-

ded che nicdial process of che squaniosal.

Therefore, Piicadelphys woiild represent che ple-

siomorphlc cojidition for dideJphoids. If this

interprétation is correct then the didelphoids

(including Pticadelphys and Andinodelphys) would

represent the sister-group oi rhe borhyaenoitls

(Muizon efv't/. 1997).

Therefore, givtn the faci thaï none of the

non-dental synâpomofphies ot didelphids are

présent in PuauMphys and consideriug the weak-

ness of the dental synapomorphîes (see above) 1

formally include Pucadslphys in a new supragene-

ric taxon distinct ttoni the Didclphidae: the

tamily Pucadelphydae new. Andinodelphys is

very probably aiso a Pucadelphydae. However,

since the study of îts cranial anatomy is still in

progress, the diagnose uf the new lamÜy is re.s-

tricted hcre to that of thc spccics Pucadelphys

andiniis. The J^Icadelphydac arc regarded herc as

Didelphimorphia {sensu Marshall et al, 1990)

and probable members of rhe superfamily

Didelphoidea. liowever, ic is noteworthy that

niuch of che early histori^ of marsupials is Icnown

by tecth, and ic is clear that the discovery of

major a-anial remains of such taxa as Alphadofu

Peradectes or Pediomys would probably radically

change our poor underscanding of early marsu¬

pial évolution.

Because of the presence of a médial process of

the squamosal, whicb relates Pucadelphys to the

borhyaenoid-S, the lattcr could possihly have their

origin within unspecialised didelphoids {i.e. a

Pucadelphys-\ÙJt fdrm without dental spécialisa¬

tions) or within their primitive sister-group

(Muizon et al. 1997). This short discussion

demonstrates again how hazardous it is to

construct a phylogeny based on tecth only and

how cranial and postcranial remains ol early

marsupials arc needed to providc a safer approa-

ch to thc origin and early history of the group.

The pcriolic of Pucadelphys, didelphids, caeno-

lestoids and somc dasyurids bcars a small prootic

canal for thc transmission ol thc prootic canal

vein which unités rhe latéral head vein (which

passes in the posrerior pan of ihe facial sulcus)

and the prootic sinus, a pnmary tributary of the

latéral head vein (which runs in a sulcus on the

latéral sidc oi thc periotic and Ls bordered lateral-

ly by thc squamosal). The prootic sinus exirs the

skull through che postglenoid foramen, via the

sphenoparictaJ cmissacy vein. The médial ope-

ning of thc prootic canal is situated in a grcK)ve

of the latéral side of the facial sulcus (a relict of

rhe latéral trough of morganucodontid.s), posté¬

rolatéral to the secondury facial fciramen. The

latéral opening of thc prootic canal is in the ven¬

tral cxti'cmity of thc sulcus for rhe prootic sinus,

on the lateial sidc of the periotic. The prootic

canal of marsupials passes dorsal to rhe petrosal

cresi and anterodorsal to thc cptcympanic recess.

Aiiiong fossil marsupials, a prootic canal is pré¬

sent in hicadelphys, in Andinodelphys. in petrosal

of type.s A, B. C and D ofWible (1990). Its pre¬

sence in Didelphodon cannot be confirnted since

thc corresj^onding part of che only known perio¬

tic of this taxon is broken (Wiblc 1990). The

lack of a prootic canal in Mayuh'stes and in the

ûther borhyaenoids is a synapomorphy t>f the

superfamiiy. However, it is likcly that this (oss

occurred several rimes during marsupial évolu¬

tion (some dasyuroids, pcrajncloids, notoryc-

toids and diprotodonts do not have a prootic

canal), which thereft^re considerahly reduces its

phylogeneric value.

The ectotympanic of the holotypc and unique

specimen of Mayulestes ferox has bcen lost during

fossilisation. Archer (1976a: 293) noted in

CLtdùsictis a unique articulation of thc ectotym-

panic whose ''main body is întergrown laterally

witli thc squamosal and vcntrally wirh the tym-

panic wing of rhe alisphenoid '. As he srated,. this

condition is unique among marsupicurnivores. I

have observed Lr in Cladosictis (YPM PU 1 5170),

Payahorhyaena (MNHN SAL 51) and Noiogale

(MNHN SAL 271), wheie the ectotympanic was

preserved in situ (Fig. 47)- I have personally

removed the tympanic of YPM PU 15170

{Cladosictis patagoitirus) which was still in

contact wirh the squamosal. The articulation of

the ectotympanic wirh the st|uaniosal is characte-

rized by several interlocking rîdges and grooves.

So, even il the ectotympanic is lost during fossi¬

lisation, it is still possible to know if this peculiar

fcaturc was présent or not I have observed ridges

and grooves on the posteromedial angle of the

glenoid fossa and on the médial side of the

GEODIVERSITAS • 1998 • 20(1}

Mayulestes, a borhyaenoid from the Paiaeocene of Bolivia

Fig. 47. — Auditory région of Notogale (MNHN SAL 271) to show the ectotympanic interlocked with the squamosal. Scale bar: 5 mm.

nicdial wall of the posrglcnoicl foramen (rhe

location of ihe ectotympanic-squamosal articula¬

tion) in Prot/ry/arynus (YPM PU 15700),

Borhyaena (YPM PU 15120), Sipalocyon

(AMNH 9254) and SatUicyon (MNHN

SAL 92). It Ls apparently absent in Tlsylacostniku,

However, consideiing tbc hyperspécialisation oi:

the ear région ot that genus, it is possible rhat

this feature vvas lost in thi.s taxon. As mentioned

above, tlie case of Mayulestes is dübcult sincc it

seems ro be intermediate between that of didel-

phids and that of the other borhyacnoids.

However, the condition of Mayiilestes \% doser to

that observed in didelphids and certaitily not as

specialised as in the other borhyacnoids.

Mayulestes thercforc retains a plesiomorphic

condition wirhin the Borhyaenoidea and ir is not

certain thaï this feature was even incipicntly

developed in this gcnu.s.

"Phe periotic of Mayulestes has a large pars inas-

toidea whidi greatly contributes to Uie occiput.

This is a plesiomorphic condition which also

exists in several 1-ate Cretaceous and Paiaeocene

marsupials (Eoelelphis, Didelphodotu pctrosal

type A and B of Wible (1990), Pucadelphys),

This feature is also foUnd in diddphids, caeno-

lesioids and niost dasyuroids. Ali the other

borhyacnoids hâve a reduced pars mastoidea,

internai to the braincase and wedged between

rhe squamosal and the cxoccipital. Mayulestes

retains the tribosphenid plesiomorphic condi¬

tion. Among the other borhyacnoids, the occi¬

put is fotmed by the occipital only in CLidosictis^

Sipalocyon and Notogale. A large conlribution of

the squamosal to the occiput (in the place uf the

pars mastoidea) is observed in Borhyae?ta,

Prnthylacyrtus, Piruborhyaena (contra Mtiizon

1994) and Thylacosjnilus. In this genus, rhe squa¬

mosal participation lo rhe occiput is smaller than

in the other thice gênera but this is probably

rclatcd to ifs very spécial tympanic huila (sec

Muizon 1994 and bdow).

The présence of a small truc mastoid process and

chc absence of paroccipiral process in Mayulestes

is a plesiomorphy wirhin mammals (Marshall

Muizon 1995). In the other borhyacnoids, sincc

the rcduccd pars mastoidea is internai to chc

braincase, the (so-caI|ed) mastoid process docs

noc contain any dément of the pars mastoidea of

the periotic but is formed by the adjunedon of

ihe paroccipkal process of the cxoccipital and

the posttympanic process of chc squamosal. It is

medium-sized in Prothylacynus and Parabo-

rhyaena. It is relacively large in Borhyaena where

it is excavated anteriorly hy a paroccipital hypo-

tympanic sinus, It is welLdeveloped and projecis

anteroventrally in Cladosictis^ Sipalocyon and

Notogale. The extreme condition is présent in

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Thylacüsrnilus where the amen’or projection of

rhe posttympanic and paroccipital processes is so

hypertrophied diat rhcy contact respectively the

squamosal and the alisphenoid medially to the

postglcnoid proccss and completely floor the

tympan ic cavity. Therc is no tympanic process of

rhe alisphenoid in ThyLicosrnilus.

Conclusions on the skull

The major phylogenctic contribution of the new

borhyaenoid is a better understanding of marsu¬

pial and borhyaenoid synapomorphics. Although

tempting, I shall not considet herc the phylogc-

netic rclationships of the supetfamily as a whole

since this study will be undertaken in a work in

progre.ss which ineJudes descriptions of the basi-

crania of SalUicyon boffstetten (MNHN SAL 92)

and Notogale (MNHN SAL 271). The new data

providcd by these spccimeas allow the establish¬

ment of a phylogeny of the Boihyaenoidea

represented by major cranial rcmains. In the fol-

lowing section I shall consider the problem of

che diagnosis of the Borhyacnoîdea and the affi-

nities of Aiayulestes,

Définition of the Borhyacnoîdea. The

borhyacnoids have bcen diagnoscd scvcral times

by Marshall (1976, i977K i978> 1979a. 1981).

However, features taken inro account are ofeen

symplesiomorphies and some of rhem, regarded

as derived, are highly homoplasric and cherefore

their phylogeneçic value is redueed (emphasis on

prevallid'posrvallum shear, rapid increase in size

from Ml to M3 and Iront ml to m4. incipient

rostral tympanic proccss of the periotic).

Marshall et al. (1990) hâve included in rhe taxon

Borhyaenoidca the familles .Stagodontidae and

Hondadelphidae. l hese authors have given ano-

ther diagnosis oF the Rorhyaenoldea. In this

work, rhe Borhyaenoidea tn their tradilional

sense (Le. sensu Muizon 1994) i.s the monophyle-

tic gtüup including lhe Mayulestidae +

Hathliacynidae + Borhyacnidae + Proborhyae-

nidae + ThyJacosmilidae. Marshall et aL (1990)

have diagnoscd the monophyletic cladc made of

Hathliacynidae + Borhyacnidae ^ Proborhyae-

nidae + Thylacosmilidae with four synapomor-

phies: (l) a distinct uasal-lacrimal contact;

(2) the loss of subarcuate fossa of the période;

(3) the réduction of the talonid and protocone;

(4) rhe loss of epipubic bones. Character 1 is a

symplesiomorphy Character 2 is absent in

Mayulestes. Sallacyon, Notogale and Clcidosictis as

these généra have a well-developed subarcuate

fossa (this fearure in Mayukstes and Sipalncyon

has bcen observed by CT scanning at the

Department of Geology of the University of

Texas .it Ausdn. Rcsults of this scanning arc

under sludy by chc author, R. Cilcili and

T. Rowe). Character 3 is absent iu Mdyulestes

and Is involved in a highly homoplasric limcdo-

nal complex related to hypercarnivoroiis dict.

Character 4 may bc a synapomorphy of the

borhyacnoids. Howet^er, Il is only probable for

A4aytilestes, Clodosictis, Prothylacynus and Lycopsis

and therc is no indication that it was présent in

the other taxa of the siipcrfonriily. It is neverrhe-

less tentadvely retained herc as a borhyaenoid

synapomorphy. Thereforc. the diagnosis ot rhe

Borhyaeiiüidea given by Marshall et al. (1990) is

rcgardcd hcrc as poorly supported.

Marshall & Kielan-Jaworowska (1992: 371) have

diagnoscd ihc Borhyaenoidea with two synapo¬

morphics: “incisors redueed to 4/3; trend for

molars to increase tapidiy in size from ml to

m4i Ml lo M3’'. However* Mayxikstes has five

upper and four lower incisors and. as mentioned

above, chc increase in size of the molars is hardly

observable in Maytikstcs and is a spécialisation

related to hypcrcarmvorous diet found in several

other groups of carnivorous mammals. It is note-

worrhy that rhe increase in size of the mol.ars is

regarded by M.arshail et al. (1990) as a synapo-

tnorphy of the six following familic.s (their

Borhyaenoidca): Stagodontidac, Hondadclphi-

dae. Hathli.acynidae, Borhyacnidae, Prohorhyae-

nidae and T hylacosmilidae, while Marshall &

KieLin-Iaworowska (1992) regard this character

as a synapomorphy of the Borhyaenoidca, a

taxon which, according to these aiirhors, does

not include the Stagodontidae. Therelore, the

proposed phylogeny of Marshall 8c Kielan-

Jaworowska (1992) contradicts that of Marshall

et al. (1990). *[‘hc phylogenetic v^alue of the cha-

racter statc ‘"rapid increase in size from Ml to

M3 and from ml to m4'’ .appears ro be questio-

nable and ics wcakness is corroborated by the

contradiction pomted out above. Therefore, the

two dental characters proposed by Marshall &

GEODIVERSITAS • 199B • 20(1)

Mayulestes-, a borhyaenoid from the PaJaeocene of Bolivia

Kielan-Jaworowska (1992) arc not acceptable as

diagnostic of the Borhyacnoidea.

Jn fact, the on)y unique feature observed in ail

the skulls of borhyaenoids is che contribution of

the médial process of the sqiiamosaJ to the alis-

phenoid hyporympanic sinus. So far as known,

this feature is absent from any oiher marsupial

and is regarded here as the key-character of ihe

Borhyacnoidea. Anochcr suggested synapttmor-

phy of the borhyaenoids is the loss of che prootic

canal, a structure which cransmits the prootic

canal vein, svhich in turn links the sphenoparic-

tal cmissaiy vein to rhe latéral head vein in didel-

phids, caenolescoids and some dasyurids.

However, this character State also appears in

other lincages of marsupiaU (Wiblc 1990) and is

consequcntly of lower phylogenerLc value than

the key-synapomorphy cited above. furcher-

more, jn didelphids, the fossa for the lower

canine, antenor ro the iipper canine, is bordered

by an antérolatéral process of the maxilla lateral-

ly. In posr-Palaeocene borhyaenoids, rhis process

disappears and rhe fossa for the lower canine is

opened latérally. Jn Mayulestes^ the antérolatéral

process of the maxilla is sfiil présent but rcduccd,

announcing chcreforc the condition observed in

younger borhyaenoids. ’l*hc réduction and loss of

rhe antcrolatcral process of ihe m;L\illa is regar¬

ded here as a borhyaenoid synapomorphy. I his

feature is also présent in rhylachnis but. in this

genus, the fossa lor the lower canine is not ope-

ned lalerally {i.e. is completely excavated in the

premaxilla), contrary to the condition observed

in borhyaenoids. As mentioncd above, it is pro¬

bable that Mayitlesies lacked cpipubic bones. If

rbis hypothesis is correct, the probable lack of

epipubic bones in Cladvsictis^ Prothylacyuits and

Lycopsis (the only ihree other borhyaenoids

known by relatively complété skcleions) would

indicatc that the loss of epipubic bone is anoiher

probable borhyaenoid .synapomorphy.

Affînities of Mayulestes ferox. fhe above dis¬

cussion of some relevant features of the skull of

Mayulestes shuws that, for most of them, it

retains the plesiomorphic character .sute for a

borhyaenoid or for marsupial. In facr, the Family

Mayulestidae {Mayulestes and AUqoktrus) repre-

sents the sisrer-group of .dl rhe other members of

the superfamily, which are diagnosed by six syna-

pomotphies: (1) the pars mastoidca of the perio-

tic, internai and not conirihuting to che occiput;

(2) the loss of the contribution of rhe alisphe-

noid to the glenoid fossa (/.r. the loss of the

entoglenoid process of the alisphenoid (Clemens

1966); (3) rhe réduction of the barnular process

and laminae of’ the pterygoid and the formation

of cwo cresLs which connect, without Icvel diffé¬

rence the posterior border of the palaie to the

basicranium;. (4) the rympanic intcrlocked wiih

the squamosal; (5) the number of incisors rcdu¬

ccd ro 4/3; (6) the dotiblc-archcd posterior edge

of the palare. The six plesiomorphic conditions

of these character States arc présent in Mayulestes.

Thcy arc: (1 ) che pars mastoidca concributing to

the occiput, (2) the présence of au entoglenoid

process of the alisphenoid, (3) che hamular pro¬

cesses of che pcert'goids wcll-developed and nor

in continuit}' wiih che basicranium, (4) rhe rym-

panic loose, attached to rhe squamosal by liga-

mencs only, (5) 3/4 incisors, (6) the single-

arched posterior edge of the palace. Mayulestes

iAV\à Allqakirus (family Mayvtlesritlae) arc diagno-

sed by the réduction of rhe enroconid and rhe

concomitant lingual opening of the talonid

basin. It has been suggested above that, becausc

of cheir lower molar morphology, the

Mayulesndae could represent the si.ster-group ol

the Borhyaenidae. However, the six cranial syna-

pomorphies of the other borhyaenoids Üsled

above demon.strate thaï the hypcithesis cannot be

accepted so far. Nevertheless, the diagnosls of the

Mayulestidae and their relation-ship.s wich the

other borhyaenoids hâve to he confirmed by the

discovery of cranial rcmains of Allqokirits and

Ncniolestcs> Aiayulestes is certainly not a

Hathliacynidae as stated bv Marshall étal. (1997),

POSTCRANIAI. CHARACTERS

Atlas

The intercentrum of the atlas of the holotype of

Mayulestes is not fused to the neural arch. The

type specinien of Mayulestes ferox is a j'oung

udult since the epiphyses of limb bones are not

completely fused and the teevh are only slightiy

worn. A similat condition is observed in

borhyaena tuhaata (YPM-PU 13120) whcrc che

intercenrrum is not fused to the neural arch and

which clearly shows incompletely ossified limb

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

bones. However, in h'othylacynus patagonicus

(YPM-PU 15700) and in CL^doslctis patagonica

(YPM-PU 15702), the intercentrum of the arias

is totally fused to the dorsal arch while the limh

bones are incompletçly ossified. In most living

didelphids the inrercentrum is completely fused

CO the dorsal arch, contrary to whaL is observed

in Pucadelpbys. The condition in Mayulestes is

plesiomorphic when comparcd ro chat of

Prothylacynusy Chtdosicth and rcsembles chat of

Pucadelphys and Borhyarna.

The absence of fully enclosed intcrv'crtcbral fora-

mina is a primitive condition in Mayulestesy

Borhyaena, Pticadelphys, Marnmsa^ Monodelphh

and Peramvles\ its présence in Prothylacynus,

Cladosictü and Sipûlocyon is a derived character

State. The absence ol a transverse foramen is pri¬

mitive in MayulesteSy Pucadelphys, Didelphis^

Monodelpfnsy Metachtrus and Asïoryctes^ while its

presence is a derived character State in ihe

Santa Cruz borhyacnoids.

Axis

The axis oi Mayulestes is clearly specialized in its

large, long and triangular spinous process, a deri¬

ved fearure of borhyaenoids, aiso found in

Thylaciniis. le îs more derived than in

Pticadelphys and other didelphids, excepr

Didelphis. A very long and iriangular spine of the

axis is also pre.sent in creodonts and carnivotans

and represents an adaptation ro hypercarnivo-

TOUS dier since chese animais kül rheir prey wiih

their jaws, whith requires grcar strength ol the

neck musculature. A large triangular spinc of the

axis is also présent in XaLzwbdalestes but, in this

Late Cretaceoiis mammal from Mongolia, che

structure of the spinc of rhe axis seems to indica-

te that the anterlor part of the neck was immobi¬

le which would be indicative of a tendency

toward salracorial habits (Kielan-Jaworowska et

ai 1979; 239). However, the axis of Mayidestes is

shorcer anreroposreriorly than that of the other

borhyaenoids and Thyla^imts^ a condition which

represents che primitive one. The lack ol a totally

enclosed rransverse loramen is a primitive fearure

also found in the Palaeocene didelphoid Puai-

delphys. In thLs respect, Mdyukstes is more primi¬

tive than the other borhyaenoid, this foramen

being always présent in the latter.

Other cervical vertebrne

The major characteristic of the cervical vertebrae

of Mnyulestes is their relative shortness when

compared to chose of the other borhyaenoids.

The relief of the ventral sidc of their centra being

le.ss pronoLinccd than in rhe Santa Cruz borhyae¬

noids, a wcaker musculature of the neck is sug-

gested. The shorter and wcaker neck Is probably

related to a Icsscr mobiliiy of ilie neck. i his

condition is clearly le.ss speciaJiscd than that of

the other borhyaenoids sîncc lengih and strength

of the neck are dassical adaptations to hyperpre-

daceous habits aiso ob.served in ctirnivorans,

creodonts and thylacynids. Mayidestes is, howe¬

ver, clearly more derived than several unspeciali-

sed (in this respect) didelphoids {Pucadelphys,

Caturornys^ Monodrlphis^ Metachirus^ Philander).

Thoracic and lumbar vertebrae

As in Pucadelphys, rhe anticlinal verrebra (the

vertebra where the spinal orientation reverses

from a posteritjr orientation in the prcanciclinal

vertebrae to an anterior orientation in the pos¬

tanticlinal t-Tertebrae) is located much more pos-

teriorly in Maynlestes than in Cladosictis and

ProthyLicynm, since it occurs on the lumbar ver-

tebrae^ becw'een 1.3 and L5 (the spinal process of

L4 is not preserved). In Cladosictis and

Prothylacyniis, the anticlinal vertebra is 1*11. The

anticlinal vertebra is Ll'in ZaLîwbdalestes^ L2 in

Pticadelphys^ between L2 and L4 (the spinc of L3

is broken) in Asùltherium, L3 in Metachirm , L5

in Mannosa and Caliiromys, L6 in Monodelphis

and Tl 1 in Perameles — a very dynamic curso-sal-

tatorial Australian marsupial (TsJovâk & Paradiso

1983). High and strongly aiuetiorly oriented

spines oi the lumbars are also observed in carni-

vorans (canids and telids) and creodonts, wliere

the spine inversion commonly Cakes place on the

last thoracics. This condit'ion is related tu a fast

running. A posterior position (i.e. lumbar) of the

inversion is observed in carly marsupials and pla-

ccntals and in most living généra of the conser¬

vative tamily Didelphidae. rhereforc, A4ayulestes

would retain the primitive condition of rhat cha-

râceer State.

The size and shape of the neural spine of the

lumbar vertebrae is also an important issue. The

elevated and anteroposteriorly short spine of the

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

L5 oï Mayulestes dearly diffcrs From thc low and

anteroposteriorly long spine observcd in rhc

living didelphids. Evcn in Metachirus, a terres-

trial didclphid with somc cursoriaJ and saltatorial

habits (Charles-Dominique, pers. comm.,

05/1996), the morphology of the spine does not

fundamenrally dJlier from that of Ca/urvmys, the

niost arborcal living didelplud. The major ditfe-

rence lies in the position oi the anticlinal verte-

bra, which is more anterior in Metachirus [in

relation ro the curso-saltarorial (lerrascaivsorial of

S’/alay 1994) habits ot this genus]. Canids (cur-

sorial) and felids (leaping cursorial) gencrally

hâve high and anteroposteriorly short neural

spines of the lumbars. Peramelids, which are

extrcmely agile cursorial and saltatorial marsu-

pials hâve very blgh. short and widely separated

neural spines o( the lasts lumbars. This rnorpho-

logy is regarded herc as related to cursorial

and/or saltatorial locomotion more rhan to fos-

sorial habits as suggested by Marshall &

vSigugneau‘Rus.sell (1995). As a matier ot fact,

nonc (with one exception) of the fbssorial mam-

mals cxamined dunng this study bears this kind

of neural process on the lumbar vertebrac: noto-

ryctids, dasypodids, talpids, bâthyergids, spala-

cids, geomyids, fossorial murids (mole-rats),

meline musrelids. le is irue thaï aardvarks hâve

high and slender processes, liowever, chey are not

oriented anteriorly. Furthermore, it is notewor-

rhy that, as stated by Novak & Paradiso (1983),

although aardvarks arc extrcmely efFicient dig-

gers, they can aiso run very fast when chased. In

rcturn» high and rclativcly slender (not always in

some carnivorans) neural proccsscs are found in

cursorial and/or saltatorial mammals: kangaroos,

canids, felids, some viverrids, murids (gerbÜs and

hopping micc), clipodids, sciurids, chinchillids

{Lagostoynus, Lagidiuyfh ChifichyUa)^ caviids

[Dolichotis), Since neirher Mayulestes nor

Pucadelphys show obvious cursorial adaptations,

it is therefore likely that they were capable of

some leaping [perhaps a son of leaping run, as

described by Jenkins (1974) in mpaüds], radier

than digging as srared (for Pucadelphys) by

Marshall Sigogneau-Russell (1995). In

Mayulestes, an elevated neural spine is known

only on L5 (it was probably présent on L4 and

L6 but it is broken on L4 and L6 is not preser-

ved). l’hc .spine of Ll-3 is rclativcly low and long

anteroposieriorly. In Cladusictis, the neural

spines ol the six lumbars are elevated, longer

anteroposteriorly and more inclined anteriorly

than in Mayulestes. In Prothylacynus^ only the last

three lumbars are known; each has a neural spine

anteroposteriorly longer (more robust) than in

Mayulestes. Becausc of the morphology of the

spine of the last thoracics and because ol the

position of the anticlinal vertebra (Tll), it i&

likely that the neural spines ot the firsr lumbar

vertebrae of Prothylacynus werc relatively similar

to chose of the poscerior lumbars. Phe morpho-

logy of the neural spine ol the lumbar vertebrae

(high, slender and widely separated) and thc

anterior position (on the last thoracics) of thc

anticlinal vertebra in Cladosictis and Protbyla-

cyuus indicates a more robust back muscul.iture

for these taxa, which tould be interpreted as

indicating some cursorial and/or probably some

büunding abÜity. Tlie more posterior position of

the anticlinal vertebra and the morphology of

the lumbar neural spine oï Mayulestes (low in thc

anterior lumbars) suggests that it was Icss specia-

liscd in these functions than Cladosictis and

Pmhylacyyius. It is obvious chat Mayulestes could

run relatively fast (as mosc living didelphids can

do), but it certainly did not hâve what is com-

monly called cursorial habits. The morphology

ol rhe neural spine oF the 1.5 would indicatc

some leaping or bounding ability in Mayulestes

although to a lesser extent than in the

Santa Cruz borhyaenoids because of the mor¬

phology of thc neural process of the anterior

lumbar and the position of the anticlinal verte¬

bra.

The long and venrrally recurved transverse pro-

cess of the last lumbars is indicative of powerfui

flexors of the vertébral column. Flexion of the

vertébral column is performed by thc quadrati

lumborum and the psoas mvijor and minor

muscles when acting jointly. Fhe size of these

muscles is compatible with that of the lumbar

epaxial museuFature (erector spinae), which is

denored by the heighf of the neural processes of

the posterior lumbars and thc eversion of the

ilia. Thcrcforc, die morphology of thc transverse

proccsscs ol rhe last lumbars is indicative of

significant mobility of the posterior part of the

GEODIVERSITAS • 1998 • 20(1)

Muizon C. de

Table 2. — Proportions of the lumbar vertebrae in some borhyaenoids and didelphids. L, length of the centrum; Wa, anterior width of

the centrum. AH measurements are In midimeters.

Wa/L

Mayulestes (MHNC 1249)

7.9

4.5

0.57

7.9

4.5

0.57

4.9

0.612

Cladosictis (YPM PU 15170)

16.4

0.82

Prothylacynus (YPM PU 15700)

Pucadelphys {YPFB Pal 6106)

3.7

0.81

4.5

2.9

0.644

4.9

2.8

0.571

Calurornys

8.3

5.2

0.626

9.4

0.532

10.7

5.3

0.495

Marmosa

2.9

0.69

3.4

1.9

0.56

3.8

1.9

0.5

Monodelphis

3.8

2.7

0.71

4.3

2.8

0.65

5.2

2.8

0.54

Metachirus

3.7

0.616

3.6

0.514

3.8

0.475

Wa/L

Mayulestes IMHHC 1249}

8.6

0.581

7.4

5.4

0.729

—

Cladosicîis (YPM PU 15170)

0.77

Prothylacynus (YPM PU 15700)

34.5

22.5

0.652

Pucadefphys (YPFB Pal 6106)

5.5

2.8

0.509

5.5

0.545

4.3

3.3

0.767

Calurornys

5.5

0.5

5.3

0.48

10.5

5.4

0.514

Marmosa

3.9

0.512

3.5

1.9

0.54

3.1

0.645

Monodelphis

5.3

2.8

0.53

5.2

2.5

0.48

5.2

2.3

0.442

Metachirus

3.9

0.487

4.2

0.525

7.3

4.2

0.575

vertébral column (last rhoracics and lumbars).

made

on L5)

. Character (4) is présent in

Jenkins {]974: 106-108) has noied four leacures, Mayulestes on 'ri2?, L1 (not observable

related ro rhe dorsovcntral mobiltry of the but probably présent), L2, L3 and L4 (littlc mar-

T1 1-T12-T1 3'Ll portion of rhe vertébral ked). The présence of rhese feacurse on the pos-

column in trcc shrews. They are: (1) the ventral terior part of the vertébral column of Alayulestes

length oi rhe centrum is shorter than the dorsal rhus indicates greac mobilicy. However,

(neural); (2) the distance between the centres of Alayulestes difFers from rhe rree shrews since the

the pre-and posr atticuJar surfaces is greater ihan mosr mobile portion seems to be locared bet-

the centrum length (measured between rlie ween T!3? and L3 or L4 as opposed to Tll and

centres of the nuclei pulposi); (3) the length of Ll in the latter (Jenkins 1974). This différence is

the zygapophyseal prearticuUr surface is compa- probably due ro rhe fact that the anticlinal verte-

table or slightiy longer than rhar of rhe postarti- bra of MayuUstes is located much more poste-

cular surface of the preceding verrebra: (4) the riorly (1.4) than in Tupaiaglis iT\Q).

anterior margin of rhe preanicular surface is The centra of the lasi thoracic and lumbar verte-

more ventral than the posierinr margin which brae are proportionally longer in Mayulestes than

gives an anteroposterior convexicy to the articu- in Cladosictis and Prothylacynus. while rhe

lar surface. In Alayulestes^ character (1) is lound contrary is observed on the cervicals. Tn this res¬

on the last thoracic (Tl3?) but is not vety well pect, Alayulestes resembles Ptuadclphys and the

marked; it is obvions on I.I-L2-1.3 and was pos- living didelphids {Calurornys^ Aloriodelphis,

sibly présent on L4. Character (2) is présent on Alarmosa, Aletachirus). Comparîson of the fourth

Tl2?, Tl3?, Ll, L2, L3 and L5 (no measurc- lumbar vertebra in Mayulestes, Santa Cruz

ment eau be made on L4). Character (3) is pre- borhyaenoids and didelphids illustrâtes chis dif-

sent from Tl2? to L4 (no measurement can be ferences of proportions well (Table 2). The rela-

100

GEODIVERSITAS • 1&98 • 20(1)

MayulesteSy a borhyaenoid from che Palaeocene of Bolivia

cive size of thc lumbar vertebrac of Alayulestes,

doser to ihat of the didelphids than ro rhat of

the Mioccne borhyaenoids, is regarded herc as a

plesiomorphic State wirhin the superfamily.

Cursorial carnivorans generally hâve elongated,

large lumbar vertebrac. The tact chat Miocène

borhyaenoids had relativdy short centra ol the

lumbar vertebrac demonstrates that they were

certainly not as hîghly specialised cursorial mam-

mais as thylacinids, canids and some felids.

However, as noted above, ihe Santa Cruz

borhyaenoids certainly had some curs'orial ability

rclaccd CO chelr hypcrpredaccous habits. In

Mayulestes. this was apparcntly absent or much

less developed than in tfie Santa Cruz horhyae-

noids.

Catiiial vertebrae

rhe morphology of the caudal vertebrae of

Mayulestes suggests some prehensile ability. A

préhensile tail is ohserved in ail living didelphids

(alchough rcducod in Lutreolîna (Novak ôc

Paradiso 1983)].

Marshall Sigogneau-Russell (1993. 118, 119)

bave suggested rhat Pucadelphys nndinHs did not

bave a prehensile tail contra Muizon (1991).

‘l'hese authors State that “in DidelphiSy the caudal

vertebrae hâve spécialisations associated with a

prehensile tail (Krause &: Jenkins 1983: 242):

e,g. [!)] tail commonly twice or more rhe length

of the precaudal vertébral column; [2)] a médian

sulcus for abducror (sic) muscles and tendons

crosses vencrally ail the vertebrac; 13)J zygapo-

physes are more vertical; [4)] cransverse ptocesses

arc broad and robusr for muscle attachments and

présent even in most distal caudals; moreover

15)] haemal apophyses, that cjiclose abductor

tendon and muscle, are large and developed

along enrire lengrh ol rail: 16)] finally, sacral spi-

nous processes are relatlvely wcll'developed,

commonly subequal to the height of the spinous

processes of posrerior lumbar vertebrae”.

Several commeuts bave to be raade on Marshall

& Sigogneau-Russell s lîst. First, leatures 2 and 3

are not cited hy Krause & Jenkîns (1983) as rela-

ted to a prelrcnsile tail. If leature 2 is indeed an

adaptation related lo a prehensile tail, why is it

absent in Marrnosa, whose tail is strongly pre¬

hensile (Novak & Paradiso 1983)? As mentioned

above, ihe ventral sulcus, well marked in

Didelphis and Calutomys, does not receive an

abductor muscle but the médian coccygeal aorta.

On the ridges which border the sulcus laterally

are inserted rhe sacrococcygci ventralis and

medialis muscles, which form a deep lurrow for

the sacrococcygeal artery. The term abductor

used by Marshall &: Sigogneau-Rus.sell (1995) is

inappropriate for the taiJ since an abduction is a

muvement ol an extremity away Irom the

médian plane; the movement they relcr to is a

Rexion of the rail* Furthermore, rhe haemal

arches (not apophyses) do not enclose muscles

but ihe sacrococcygeal artery. Marshall &

Sigogncau-Russel) (1995: 1 18) State chat ‘‘no

such spécializations except high sacral processes

and very slight ventral sulcus on caudals exist in

Philander 3Li^6 A]ctachmis \ However. both géné¬

ra also bave a taîl which is twice as long as the

presacral vertébral column and haemal arches

developed along nearly rhe entire length of thc

tail. It is true that the transverse processes of che

posceriof caudal are not as developed as in

Caluromys or Didelphis but it is also true that

they are at least as large as in thc tail oïMarmosa,

one of thc most prehen-silc among didelphids.

Furthermore, che tail of Philander is reported as

prehensile (Novak & Paradiso 1983 and

Julien-La Ferricre pers. comm.). Novak &

Paradiso (1983: 12) State that the didelphid tail

is “long, scâly, very scantily haired and prehensi¬

le'’. ConcGrmng Lutreolina, one of the most ter-

rcstrial genus, these authors State that the tail is

not as prehensile as in other didelphid.s. They

also report che observation of an individual of

Monodelphis domestica carrying a pièce of paper

by curling Iw tail downward around thc paper

Therefore, ail didelphids appear ro hâve some

dcgrcc of prehensility ol the tail. The greater abi¬

lity is found in Didelphis, Marmosa^ Caluromys

and Philander and the lesser ability is found in

Morwdelphisj LutreoUna. and Lestodelphys.

Marshall & Sigogneau-Russell (1995: 119) hâve

stated that there was no indication that the tail

of Pticadelphys prehensile. However, the pos-

terior caudal vertebrac ol Pucadelphys hear large

and robusr rransverse proccsscs and the tail is

long (estimation of 30 vertebrae). The C6 and

C7 of Pucadelphys are strikingly similar to C5

GEODIVERSITAS • 1998 • 20(1}

101

Muizon C. de

A B

Fig. 48. — Anterior caudal vertebrae of: A. Pucadelphys (C6-

C8); B, Caluromys (C5-C7). Not to scale.

and C6 ol Caluromys^ tKe most arboreal living

didelphid (Fig. 48). In che C6 of Pucadelphys

and C5 ol^ Calurornys^ thc transversc proccss is

locatcd posteriorly on thc centrutn and occupies

more than half of its length. It i.*: much longer

than in thc preceding vertebra. The C6 of

Caluromys and the C7 of Pucadelphys differ froni

the respective prcccding vertebra by the adjunc-

tion of a small anterior cransversc process, much

shorter antcroposteriorly than thc long posrerior

transversc proccss. In the posterior caudal vette-

brae of the nva générai îhe transverse processcs

are wcll-developed. as large as in Didclphis and

certainly much larger than m Marmosa or

Philauder. Bccause of the large rratisverse pro¬

cesses ot C6 to C9 ^YPFB Pal 6106) and in C16?

and Cl7? (YPFB Pal 61 10), I consider thaï

Pucadelphys had a prehensilc taÜ. It is not pos¬

sible to cs'^aluate tbe dcgiec of prehensiliry of the

tail since the tail of Pucadelphys is not complété

in the available spécimens and because there

seems ro be some inconstancy în (he corrélation

between thc anatomical feaiures regarded as rda-

ted to a prebensile tail and the actual prehensility

of the tail in living didelphids. T'hc statement by

Marshall ÔC Sigogneau-Russell (1995) that there

were no haemal arches in Pucadelphys is contra-

dicted by thc presence of a fragment of arch still

présent betvvecn C7 and C8 of YPFB Pal 6106

and of a partial arch below C4 of YPFB

Pal 6110, The lack of most haemal arches in

Putadelpfjys is regarded here as a loss duc to fossi-

lisation. d'he only feature mentioned by

Krause &: jciikins (1983) and which is absent in

Pucadelphys is the large size of the spinous pro¬

cesses of the sacral vertebrae. In Pucadelphys,

rhese processes arc broken, but iri view of their

smaller diameier it i.s likely ibai chey were not as

high as thosç of the last lumbar. However, it is

noteworrhy rhat the most important movements

of a prebensile Taîl arc the flexions of the tail and

the spinous proccss ot thc sacral vertebrae bears

attachment for extensor muscles of the taÜ.

The twu poster iox caudal vertebrae ofA/iry?//ejrer

are extrcmely similar to the C9 ot Pucadelphys

and to tbe C7 and C8 of Caluromys^ mainly in

the large size of their anterior and posterior

iransverse processes (Fig. 40). They are, however,

longer than in Pucadelphys and slightly shorter

than in Caluromys. Because of this morphologi-

cal similaricy it is probable that Mayulestes had a

prebensile tail.

Cartmill (1974. 51) has stated that most prehen-

silc-tail animais (excepr primates) “practically

never make leaps of any distance and generally

move cautiously from one support to anocher”.

As shown below, it is probable that Mayulestes

was a relarively agile animal capable of some lea-

ping run, as rree shrews, althougb certdinly slo-

wer Therefore, a contradiction would exist

betw'een the prehensility of thc tail oï Aiayulestes

and the suggested agility of thc animal. First, it is

necessary to keep in mind that only four caudal

vertebrae of Mayulestes are known, therefore, the

anatomical support of ibe preben.sile uil of

Mayulestes is .still relarively weak. Furthertnore, if

the rail of Afayulestes wa.*; indeed prehensilc, it is

possible that this lunction was liule used by the

animal and beirig lost in favour of an increasing

agility of thc locomotion. The tail could hâve

kept the characters and the ability of prehensility

although it was not (or little) used as such. A

living examplc of this condition is thc tcrrestrial

didelphid Metachirus whîch indeed has a préhen¬

sile tail but which does not use it for climbing.

rhe same can be said of the aquatic didelphid

Chironectes. Another interprétation (see below

102

GEODIVERSITAS • 1998 • 20(1)

Mayulestesy a borhyaenoid from the Palaeocene of Bolivia

Fig. 49. — Posterior caudal vertebrae of: A, Pucadelphys (Cg):

B, Mayutestes (Cg? and Cg?); C, Caiuromys (C 7 and Cg). Not to

scale.

for discussion) would be to œnsider that thc agi-

licy of the animal was mainly used on the ground

(like Metachînfs) and chat rhe arboreal locomo¬

tion was slower. In othcr respect, ît is notewor-

thy that the grear jumping agility of the

didelphid genus Afarmosa coniradicts Carrmllls

assesemenr.

The inferred présence of a prehensile rail in

Pucadelphys (probably) and Adayîdestcs (possibly)

suggests rhat thèse animais are likely to hâve had

some arboreal habits. Furvhermore, rhe générait-

sed occurrence of a prehensile tail in didelphids

(to varions extenc according to the taxa), the

most primitive famÜy ot living marsupials.

would favoLir the idea that a prehensile tail is a

plcsiomorphiç charaeter State for marsupials and

seems to reinforce rhe hypothesis (Szalay 1984,

1994) that early marsupials were primarily arbo¬

real (see below for discussion).

Forelirnb

Scapula. The anatomy of rhe scapula of

Mayulestes dénotes a more robust consritution of

che shoulder rauscularure than in other borbyae-

noids and most didelphids, but is simîlar to that

of Caiuromys. The coracoid process is large,

strongly recurved and projects more proximally

than the glenoid fossa; the acromion projects

proximally below the glenoid fossa and anrerior-

ly beyond the supraglenoid process. The ktter

receives the otigin of thc coracobrachialis muscle

whose insertion is locared on rhe posteromedial

border of che proximal half of the diaphysis of

the humérus The coracobracJiialis is an adductor

of the forelirnb and a flexor of the shoulder. On

the acromion (on thc hamatus and suprahama-

tus processes) and on rhe ventral two tbirds of

the scapular spine attaches part of the origin of

the deltoideus (acromial and spinal) muscle. The

insertion of thc deltoideus muscle is on che distal

two chirds of che delropectoral crest on the

humérus. On che ventral rhird of rhe spine and

on the anterior mai^in of rhe acromion attaches

the insertion of rhe arlantoacromialis muscle

(Jenkins ôc Weijs 1979), probahly a part of the

omotransversarius. The origin ol the atlantoacro-

mialis muscle i.s on the posterior .side of thc wing

of the arias. The deltoid muscle is an abductor of

thc shoulder and a flexor of che arm when corn-

bined with the action of rhe teres major muscle.

The acianroacromialis pulls rhe scapula anterior-

ly and makes ic rotare anriclockwise. The mor-

phology of rhe coracoid process and acromion of

the scapula of May74cstcs^ projected proximally

and anteroproximally rcspeciively, denotc.s the

grear strength of tho.se muscles and, chcrclore,

the strengrh of lhe shoulder articulation. In

arboreal rnammals, the acromion and the cora¬

coid proccs.s are gcncrally well-developed .-ind

veuirallv (and anceriorly for thc formel) projec¬

ted, often to a mucli greater extent than in

Mayul^'stes (tree shrews, possums, opossums, pri¬

mates, rree sloths, Cydopes^ kinkajou, coendou;

Fig. 50). This morphology has been relatcd lo

arboreal habits (Corruccini &: Ciochon 1976;

Ciochon &: Corruccini 1977). The proximal

élongation of thc acromion is likely to be related

to improvement of the leverage for the deltoi¬

deus (Inman et al. 1944; Larson 1993), an

abductor of the arm. Furthetmorc, a dorsoposte-

rior élongation of thc posterodorsal angle of the

scapula, as it is observed in Mayukstes and

Caiuromys., is aiso présent in most arboreal mam-

nial-s (Fig. “SO). This is cspccially obvious in pri¬

mates (Roberts 1974, Larson 1993). The

anterior projection of thc acromion and the pos¬

terodorsal élongation of the posterior angle of

GEODIVERSITAS * 1998 • 20(1)

103

Muizon C. de

Fig. 50. — RIght scapulae in latéral (top) and proximal (bottom) views; A, Mayulestes: B, Caluromys: C, Cladosictis. D, Sciurus’,

E. Tupaia: F. Propilhecus: G. Colobus; H, Hylobates. Notto scale.

the scapula can be relared to great abiliry of ante-

rior extension of the forelimb since, during this

movemeni, ai ilie etid ol ihe extension, a bettcr

protraction ot the ümb is produced by an ami-

clockwise rotation of the scapula. Such move-

ments are performed in acrobatie arboreal

activities. As demonsrrated by Lanson (1993) in

arboreal primates, a raised-arm position provokes

an important rotation of rhe scapula whÜe, in

terrestrial primates, rhe movements of the scapu¬

la are mainly anteroposterior translations and rhe

rotation is weak. In arboreal primates, scapular

rotation is brought about by the action of a mus-

cular couple (Inman et al. 1944; Lai.son 1993).

The upper unit is chc cranial trapezius and the

lower unit is the caudal trapezius and the caudal

serratLis ancerior muscles (Fig. 51). In Didelphis,

the couple of muscles acting during the anterior

extension of chc forelimb (anticlockwisc rotation

of the scapula) has an upper unit made of the

serratus venlralis thoracis and the caudal portion

of the trapezius and .i lowcr unit made of the

atlantoacromialis and the anrerior portion of the

trapezius. In Dididphis, chc serratus ventral is tho¬

racis takes origin from the First eight or nine ribs

and inserts on the caudal angle of the scapula;

the trapezius originates from the nucchal crest

and the supraspinous ligament from the occiput

to the level of rhe chirteenth thoracic vertehra.

Its anrerior portion insères on the anterior edge

of the spine of the scapula, whilc its posterior

portion inserts on the posterior edge of the distal

third of the spine. The ailanroactomialis links

the posterior side of chc wing of the adas to the

acromion and the proximal third of the spine

(Jenkins bc Wcijs 1979) Acting jointly, rhese

threc muscles rotate the scapula countcrcloclcwi-

se and exert an anterioiiy directed force on the

acromion and proximal third of the scapular

spine and a po.steriorly directed force on the pos-

terodorsal angle of the scapula.

The rwo leatures of rhe scapula discusscd above

(anterior position of the acromion and postero-

dorsal élongation of ilie posierodorsal angle) are

not as dcvelopcd on the .scapula of Mayulestes as

on the -scapul.ae of highiy arboreal primates, but

ihcy approacli the condition observed in

Caluromys, chc most arboreal didelphid (Fig. 50).

The proportions and relative sizes of the siipra-

and infraspinatus fossae are difficult to explain

104

GEODIVERSITAS • 1998 • 20(1)

Mayukstes, a borhyaenoid from che Palaeocene of Bolivia

f^iG. 51. — Scapulothoraac muscular couple involved in the

rotation of the scapuia in Mayuf&stes [based upoo Larson (1993.

fig. 2.5) and Jenkins & Weijs 0979)]. The anteroveniral unit is

made of the atlantoacromralis and the anterior portion ot the tra-

pezius; the posierodorsal unit is made of tfie serraïus ventralis

thoracis and the posierior portion of lhe trapezius.

mechanically in Mayulestes. In arboreal primates,

the infraspinatus fossa and muscle are generally

larger chan the supraspinacus (Robert.s 1974)

while the conirary is observcd in Mayulestes.

However» squirrels bave a morpholog)^ of the sca-

pilla ver)' similar ro that of Mayulestes. The pos-

tei'odorsa) angle is scrongly elongared, the

acromion is vcnirally and anceriorly projected,

the coracoid process is very long and developed

ventrally (much longer than in Mayulestes and

any didelphids), the supraspinatu.s tossa is large,

triangular and rhc infraspinatüs fossa is deep

long, scraight and narrow and its posierior edge

is almost parallcl ro the plane of tho spine.

Squirrels are well*known co be non-suspensorial

arboreal rodenrs whÜe many arboreal primates

are at leasr parrially suspensorial. The similariries

between the scapulae of Mayulestes and sciurids

are possibly relaied ro similar positional beha-

viours. 'rhfrefore. many aspects of the morpho-

logy of the scapula of Mayulestes are similar to

those observed in bighly arboreal mammai.s. Ir is

mechanically consistent, with a sfrong commit'

ment lo arborcality, pmbably to a liighcr degree

than in rnost living didelphids (except

Caluromys).

In Pucadelphys^ che morphology ol die scapula is

not completcly known. However, the best prcscr-

ved scapula of Pucadelphys andinus (Fig. 18)

shows important similarities with that of

Mayulestes: ventral and anterior development of

the acromion and the deep and narrow infraspi-

natus fossa, According to the above discussion

these features are apparently consistent with

well-develûped arboreal habits.

Humérus» ulna and radius. The proximal extre-

mity of the humérus aï Mayulestes rcsembles that

of the didelphids more than chat of che other

borhyaenoids. The head is relacively circulai* (in

proximal view)j oriented more proximally than

posteriorly and slightiy higher than ihe greater

tubercle (Table 3). In Prothylucynus^ the he;ui has

a more posrerior orientation and is lower chan

the greater tubercle, cwo features well-developed

in cursorial mammals (for instance Thylaànus).

Arboreal mammals tend ro hâve a head of the

humérus proximally oriented and a low greâter

tubercle in order to increase the mobility ot the

joint. In cursorial mammals, a grear multidirec-

tional mobility of the articulation is not essendal

since the movement of the forelimb is mostly

anteropo-srerior and importance is given to the

power of the movement. Thereiore, modifica-

dons ot che joint are focu-sed toward stabilisation

of the scapulohuincral joint; according to

Larson ôc Stern (1989), that is the most signifi-

cant rôle of the supraspinatus (which is inserred

on che greater tubercle) in rerrestrial primates,

d'he intraspinaCLis is aiso involved in that func-

don since ir is inserted on the greater tubercle

slightiy more proximally than the supraspinatus.

Table 3. — Proportion of the head of the tiumerus in varlous

marsupials. L, Length; W, width. AH measurements are in milli-

meters.

W/L

fl/layutesfes (MHNC 1249)

5-2

5.6

1.07

Prothylacynus (WM PU15700)

2.9

Ctadosicîis (YPM PU 15831)

1.77

1.35

0.76

Cladosictis{YPU PU 15556)

1.65e

1.3

0.78

Pucadelphys (YPFB Pal 6106)

3.2

Caluromys

5.8

0.96

Didoiphis

10.3

9.6

0.93

Marmosa

2.3

2.1

0.91

Monodelphis

3.6

3.4

0.94

Metachirus

7.6

5.9

0.77

Thylacinus

21.5

0.79

Tupaia

0.84

Sciurus

7.2

6.8

0.94

Vulpes

3.2

0.93

GEODIVERSITAS • 1998 • 20(1)

105

Muizon C. de

Table 4. — Relative height of the spine of IPe scapula in varlous

arboreai and lerrestrlat mammals. H, greatest heigh! of îhe

spine approximalely in its middie part; L, length of the spine at

its base trom Ils distal notch (i.e. at base of the acromion) lo the

proximal extremity. AH moasurements are in millimeters.

Mayulestes (MHNC 1249)

4.8

19.6

0.244

Pucadelphys (YPFB Pal 6105)

2.9e

14.4

0.2

Didelphis

33.2

0.18

Caluromys 1

4.2

17.6

0.23

Caluromys 2

4.3

18.5

0.23

Marmosa

2.1

9.5

0.22

Monodelphis

3.2

14.3

0.22

Metachirus 1

4.6

25.3

0.18

Metachirus 2

3.5

0.175

Philander

0.2

Sciurus

0.26

Tupaia

3.3

14.4

0.23

Potos

9.4

47.8

0.196

Herpestes

5.5

31.2

0.176

Mustela

4.3

0.12

The aurhors also conclude rhar this muscle

(supraspinatus) and, rherefore, the size of the

greater tubercle are nor related to speed or power

of the movement (Larson & Stern 1989 1992).

A small greater cubercle and a humerai head

proximally orienced generate a greater mobihey

of the .shoulder, which is required m arboreai

life. However. as mentioned by Larson &C Stern

(1989)v rhe scapulohiimeral joint also need.s sta¬

bilisation diiring acrobauc behaviour. Thç

authors conduded that “the only way for an ani¬

mal with a lower greater tubercle ro deal wirh

these heavy demands on the supraspinatus for

brachial élévation and joint stabilisation is ro

increase the ovcrall size of the supraspinatus

itselP (Larson 1993' 60), There are two ways of

increasing che size of the supraspinatus muscle,

either by increasing rhe size of rhe fossa or hy

increasing rhe élévation o( the spine. The spine

of Mayiilestes is more elevated than that of

Caluromys but lower than in sciurids and

approaclies chat of tupaiids. It is rclarively much

more robust than in Metachims, a terrestrial

didelphid. 'Labié 4 compares the relative height

of the .scapular spine in scveral marsuprals, one

sciurid, onc lupaid, one arboreai carnivore and

two terrestrial non-cursorial carnivores.

Tlic deltoid cresc of Aiayulestes is shorter than in

the Santa Cruz horhyaenoids bue matches the

leneth ob.served in the livinc didelphids

(Table 5).

The very salient tricipital crest o’t Aiayulestes

received rhe origin of a robust caput latérale of

rhe triceps hrachii muscle. The insertion of this

head of the triceps is on the olecranon of the

uina with the capiris mediale and longum. The

strength of the triceps brachü ol Aiayulestes is

also icvealcd hy the grcat length and size of the

olecranon ol: the ulna and by rhe impomnt ante-

rior curvature ol the proximal ihird of the shaft

of that bone. This morpholog)' indicates impor¬

tant tractions of the triceps on rhe olecranon. A

simitar morpbology of rhe olecranon is lound in

living didelphids, being strongly ernphasîsed in

the rnost arboreai taxa {Caluromys, Alarmosa). It

is also obvions in Puaidclphys (Figs 25, 52). In

Caluromys and Alayidestes, the tension of the tri¬

ceps brachü on the olecranon is even greater

.since its anteroproximai angle is strongly elonga-

led antcropFoxim;illy. Such a morphology of the

proximal half of the ulna exisrs, to various

extents, in many arboreai [especially arboreai

quadrupcdal or arboscansorial (Szalay 1994)]

mammals: didelphids, phalangerids, Nasua,

PoTos, sciurids and rupaiids. In primates and

xenarrhrans there is a tendency to reduce the

length of the olecranon in order to allow a grca-

ler extension of the cibow. The posterior edge of

the ulna is scraight in terrestrial mammals, even

concave in highly cursorial taxa, with the olecra¬

non posteriorly oriented (Bown et al. 1982). In

the Santa Cruz horhyaenoids the proximal thîrd

of the ulna is not bent anteriorly. In

Ptothylacynus and (lladosictis, ihe posterior bor¬

der of the ulna is concavoconvex but the bonc is

glübally srraigbt. In Borhyaeiia^ rhe posterior bor¬

der of the ulna is siraight in its proximal balf and

concave in its distal half and the olecranon is

long, robust and has a very quadrate proximal

extremity. The morphology of the ulna of

Borhyaena is indicative of a terrestrial mammal

with somê cur.sorial ability. The ulna of

Thylaiinus (a cursorial rnarsupial) is similar to

that of Borhydena, but more gracile and more

recur\'cd posteriorly.

The distal extremity of the humérus, of

106

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Table 5. — Relative length of the deltoid crest. Lh, total length

of the humérus: Lcr, length of the deltoid crest. Ail measure-

ments are in millimeters.

Lcr

Lcr/Lh

Mayulestes left (MHNC 1249)

32.3

18.4

0.57

Mayulestes r\Qt\\ (MHNC 1249)

34.6

19.2

0.555

Prothylacynus (YPM PU 15700)

161

103

0.639

Cladasictis (YPM PU 15702)

113

0.708

Pucadelphys (YPFB Pal fi106)

17.8

9.5

0.53

Didelphis

62.7

0 558

Caluromys

40.8

21.7

0.53

Marmosa

14.4

7.8

0.54

Phifander

42.8

22.8

0.53

Monodelphis

20.5

0.585

Metachirus

31.7

16.4

0.51

Thylacinus

178

129

0-724

Mayulestes bears a very strong latéral epicondy-

loid crest and a robust distomedially elongated

media) épicondyle. On the posterior side of the

epicondyloid crest attaches the distal part of the

origin ot the caput mediale ol the triceps brachii*

a powerfijJ extensor of the elbow whose insertion

is on the anterion aiitetomedial and anicrolatcral

surface ol the olecranon ot the iilna (the distal

portion of the caput médial of the triceps is the

anconeus muscle, which is apparently fused to

the médial head of the triceps in dasyurids

(Kielan-Jaworowska &: Gambaryan 1994)]. On

the anterior sidc of the latéral epicondyloid crest

and on rhe latéral épicondyle attach the origins

of the extensor muscles of carpus and digits and

the brachioradialis muscle. On the médial épi¬

condyle are the origins of rhe flexors muscles of

carpus and digits, the epitrochleoanconcus and

the pronator tercs. Thereforc, the morphology of

the distal extremiry of the humetus oîAiayxdestes

indicates the power of the flexion and extension

of the manus and a good abilîty iu the

pronation-süpination movements. The great

depth of the fossa on the médial side of the

proximal third of the ulna loi* the origins of the

flexur carpi ulnaris and flexor digitorum profun-

dus muscles as well as rhe strong médial bending

of rhe proximal rhird of the bone also contribute

to suggest ability of powerfui flexion of the digits

and maims, for instance for grasping branches.

These aspects of the morphology ofthe elbow of

Mayulestes rcscmble chose of Caluromys (ihe

most arborcal living didciphid) and Pucadelphys

(although the features mendoned above are less

pronounced in this genus; Fig. 32). They are

more pronounced in Mayulestes than in most

Fig. 52. — Latéral view of the left ulna in several marsupials; A, Mayulestes: B, Caluromys: 0, Marmosa: D, Didelphis:

E, Metachirus: F, Pucadelphys: G, Prothylacynus; H, Borhyaena. Not to scale.

GEODIVERSITAS • 1998 • 20(1)

107

Muizon C. de

other living didclphids and Santa Cruz. borhyae-

noids. Patos and Nasuu, arboreal and semiarbo-

real carnivorans rcspcctivcly, bave the same

modification of the cibow rcsponsiblc for power-

ful flexion and extension ot the manus and

digits. Sciurids aiso bave a wclbdcvclopcd lacerai

epicondyloid cresr. a distomedially projecting

médial épicondyle and a médially benr proximal

extremity ni the ulna. A long and anteriorly benr

olecranon bas been regarded as related lo arbo¬

real quadrupedalismc since, in such forms, the

forelimb is aiways in partial flexion and vhe

extensors ol ihe elbow are aiways resisring ro

flexion (Bown et ûL 1982). This feature, extre-

mely developed in a predaror like Mayjtlestes^

could also indicate bounding or leaping ability;

On the iiina ni Mayulestes, rhe trochlear notch is

more open (in médial view) and shallower rhan

in the Santa Cru7 Borhyaenoids and some didcl-

phids {LyidelphP^ Meiaehirus)^ furdicrmorc, the

trochlear and radial norches form an angle of

approximarely 120° in anrerior view. As mentio-

ned above^ this condition is found in the mosi

arboreal didelphids {CaluromySy Marmosa.

Didelphis), In a cursorial marsupial such as the

thylacine and in rhe siib-cursorial Borhyaena (see

below), the trochlear notch is less open than in

Mayulestes and rhe angle berween rhe trochlear

and radial norche.s (in ancerior view) is clearly

inferior ro 90^. In other words, the trochlear and

radial notches are well-separated by an elevated

crest and the radial notch is well-excavatcd in

lerrestrial {a priori in cursorial) forms, wliile rhe

contrary is rrue in arboreal quadrupeds. The

arboreal morpholog)' {CalunnnySi Didelphis and

Mayideslei) would indicate an elbow tbat does

not rcceivc miich articular stress and docs nor

require an important stabillty of the joint and.

therefore, would designatc rclativcly slow ani¬

mais. Howevci, this morphology of the radio-

Lilnar articulation also indicates a grear mobility

related to pronatiou-supination movements.

This is corroborated by the grear development of

rhe supinatof crest, lossa and pronator ridge. In

rhe very agile Squirrets, the trochlear notch is

similar to thdt uf Didelphis, the angulation bet-

ween both articular surfaces approaches 120°

and the radial notch is very shallow.

Furthermore, in Mayulestes^ the long and robust

olecranon indicates a powerful triceps and rhe

strong attachment crest of the ulnar collateral

ligament rnedially are indications of a fairly

stable elbow joint, whicb is in agrecment with

the postulatcd activity. lliercfore, the morpholo¬

gy ot the proximal articulation of rhe ulna of

Mayulestes scems to bc compatible wich agility.

Whatever rhat may be, ir is clear tbat the mor-

phülogy uf the humeraJ articulation of the ulna

ol Mayulestes rcsembles rhat of Caluromys, more

rhan that of any other didelphid or borhyaenuid.

The apparent weaicness of the elbow joint noted

on the aiticular surfaces vvas probably compen.sa-

red for by ligaments and muscles as in active

rupaiids and sciurids. It is also noteworrhy that

rhe relative breadth ofrhe trochlcat notch, when

coniparcd to thar of Caluromys, is a factor of sta-

bility of the elbow, as indicated by Fleagle et al.

(1925: 136) for the ulna o\ Aegyptopitheats.

Bown et al. (1982: 626) liave ILsied several fea-

tures of the ulna whicb are related to arboreal

quailrupedaiism. They are: (1) long olecranon

exrended proximrdiy îti line wirh rhe shaft;

(2) rrochlear noreb relarively shallow wich a low

coronoid process; (3) small radial notch nor dee-

ply excavared; (4) anreroposreriorly deep ulnar

shaft, posteriorly convex. According to thèse cri-

reria, the ulna of Mayulestes is rhat of a highly

arboreal quadruped.

On che radias, the proximal articulation wirh rhe

capitiilum of the humérus is transversully elonga-

Ccd and, as preserved, docs not suggest pronoun-

ced pronation-supination movements, which

contradicts rhe above srarements. The condition

nf Mayulestes is similar tu char of Borhyaena and

Cladusiitis, but difïers from Prothylacynus where

tbc radiohumcral atticulacion is more oval-sha-

ped and less transverse. Although, as noted

above, the morpholog)^ of Mayulestes is probably

rhe resuit of some post-morrem érosion .ind/or

dcfdrmarion, it is likely tbat its pronation-supi-

nation ability was slightiy less developed than

rhat of living didelphids. Ir is probable rhat rhe

movements had less amplitude, dithougfi it is

difficult ço détermine to whicb extent. In this

respect, it is noteworthy chat, in living didel¬

phids which hâve wcll-developcd pronation-

supination, rhe proximal arricularion of the

radius is never totally circulât as in arboreal pri-

108

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

mates or rree sloths but is oval, although less

transverse rhan in Mayulestes, Cladosictis and

Borhyaena. The relatively oval-shapcd proximal

articulation of the radius of Prothylacynus indi-

cates superior pronation-supination ahility wheri

compared co the other Santa Cruz borhyaenoids.

Metacarpus. As noted above, both extremiries of

McV are flattcncd in ihc same plane. Therefbre,

s'incc the articulation widi MclV (on the palmar

side of the proximal epiphysis of McV) is rough-

ly pcrpcndicular to the dorsopalmar plane of the

hand, the articulation between the McV and the

first phabmx of the diglt is aiso pcrpcndicular to

the dorsopalmar plane of rhe manus. This indi-

cates thaï digit V had a plane ul flexion whlch

tended to be ac 90‘’ with that of the other fin-

gers. This condition (which is approaching chat

of Mcl and McII in man) would indicate an

important -ability of prehensilily of the manus,

which obviuusly facilitâtes grasping and is clcarly

useful for climbing trees.

Furthermore, the medially recun'cd distal extre-

mity of rhe McV (also observed. but to a Icsscr

extent, in rhe living didelphids) renefs to shift chc

phalanges ol the fînger toward the centre of the

manus. This ct)ndition is al.so useful for gras-

ping, an action fundamental in didelphid arbo-

reality.

Hindlimb

Innominate. The ilium t»f Mayulestes is propor-

tionally slightly longer than in Cladosictis and

Prothylacynus but as shown in table 6 the diffé¬

rence is more important regarding living didcl-

phidsv w^hosc ilia are clearly long and narrow.

Such a morpholog)^ is fbund in a large number

of living raarsupials (it is absent in Thylacinus

and Peramcles), The Palaeocene didclphoid

Pucadelpbys andmus has an ilium relaiively sbor-

ter than that of the living didelphid.s and

resembles Mayulestes in this point. Therefore, che

oldest known borhyaenoid has an ilium longer

than the younger form.s and rhe tddesr known

didclphoid has an ilium shorter rhan the Recent

didelphids. Whac then is the primitive morpho-

logy of the ilium? Short or long? Marshall ^

SigogneaU'Russell (1995) regarded rhe dorsoveiv

tral expansion of the w^ing of che ilium in

Pucadelphys as a spécialisation but consider its

shortness as a primitive State. Triconodonts hâve

a narrow and elongare ilium (Jenkins &

Parrlngton 1976) although, apparencly,. noc as

long as In the living didelphids. The ilium of

Henhelotherium is longer than that Mayulestes

(Krebs 1991). The same is true in Asiaric

(Kielan-Jaworowska hi Gambaryan 1994) and

North American mulrituberculates (Krause &

Table 6. — Relative length of the ilium in various marsupials and placentals. Hll, maximum height of the ilium at mid*!ength; Lil, length

of the ilium; Lt, total length of the innominate. AH measurements are in millimeters.

Lil

Lil/Lt

Mil

Hil/Lil

Mayulestes (MHNC 1249)

25.2

0.572

6.5

0.258

Cladosictis {yPM PU 15702)

70.5

124

0.568

0.368

Cladosictis (YPM PU 15170)

60e

106

0.566

0.35

Prof/ry/acynus (YPM PU 15700)

182e

0.538

35.5

0.364

Thylacinus

104

182.4

0.57

Pucadelphys (YPFB Pal 6106)

26.4

0.568

0.33

Didelphis

50.5

75.5

0.668

0.23

Metachirus

26.2

0.65

6.4

0.24

Caluromys

30.3

0.677

6.2

0.2

Marmosû

11.8

0.69

1-6

0.13

Phalanger

47.5

75.5

0,63

10.5

0.22

Perameles

49.5

85.7

0.56

17.8

0.36

Sdurus

24.3

0.59

Nasua

87.6

0.64

Ailurus

58.3

0.61

0.43

Martes

56.5

0.56

Canis

88.5

0.64

26.2

0.47

GEODIVERSITAS • 1998 • 20(1}

109

Muizon C. de

Jenkins 1983). Bamnlestts, a Lace Crctaccous

eutherian tVom Mongolia, aiso has lodg and slen-

der ilium (Kiclan-Jaworowska 1978) althüugh

apparently not longer than in Mayulestes. A long

and narrow ilium is aIso présent in Asiatherium

(Szalay ôi Trofimov 1996). Ir sccms thcreforc

chat a long and narrow ilium could rc-present che

plesiomorphic condition of che rnarnmalian

ilium. Considering che morpholog}^ of che ilium

of Mayulcsteh longer chan char of che younger

borhyaenoids, ic is likely chat early history of thc

superfamily is characieri/ed by a shortening of

the ilium, A long and slender ilium scems to bc

fairly constant in early mamrnals, which vvould

indicate rhat it represents the plesiomorphic

condirion. Furrhermore, such a morphoing)' of

the ilium is by far che mosc common one within

marsupials. The short ilium of Pucddelphys

would hence be specialised comparcd to chat of

the living didelphids. However, skelecons of Late

Cretaceous or Palacocenc marsupiaJs are too

scarce to provude adéquate comparison and, so

far, it secms diffîcult to answct rhis question.

MayulesTCs and Puradelphys hâve conspicuously

cverted iliac wings contraiy ro rhe condition

observcd in ihe living didelphids and phalange-

rids which havc relativcly itraiglir and narrow

wings. An important extroversion of the wing of

the ilium is loimd in some cursorial (thylacine,

canids, most ungulares), (ossorial Iwombar, ban-

dicoots (aiso cursosaJratorial). aardvark, badgers,

ratel], aqiiaric (phocids. walrus, sca otter) and

arboreal mammals [koala, phalanget, small and

giant panda, arboreal iree slirews, squîrrels,

lemurs, indri, spîder monkey (the five laiicr>

capable of powcrfui leaping)]. In chc phalange-

rids, the iliac wings are slighrly evcrted ac their

apiccs. Bears and anceaters aiso hâve evcrted iliac

wings. In Cladosicüs and Prothylacynus. two

borhyaenoids from thc middie Miocène of thc

Santa Crur beds {Argentina) thc éversion of the

ilium is comparable co thaï oïMayulcsfes. On thc

médial sidc of the wing of the ilium the ertetor

spinae is insericd (consisting of thc iliocostalis

and longissimas muscles), a powcrfui extensor of

the vertébral column. The origin of rhis muscle

is on the dorsal side of rhe rhoracic and lumbar

vertebrae and ribs. An cverted ilium, therefore,

leaves more space to the posterior portion of the

erector spinae at the angle berween the vertébral

column and the pclvis and dénotés chat rhis

muscle was powerful in Mayulcstes^ as in

Cladosktis and Prothylacynus. d'hc samo can be

said for Pitcadelphys. On thc latéral side of the

iliac wing arc thc origins of the glutei médius

and superficialis, muscles inserted on rhe tip of

the greater trochanter (gluteus médius) and,

more disrally, on the latéral side of che greater

trochanter (glutens superficialis). An everred

ilium allows for more space (or the gluteus

médius and superficialis (and a greater muscular

ma.ss) than in an animal wich a non-everted

ilium as is rhe case in living didelphids. Other

muscles originating on rhe anteroventra) région

of rhe ilium are rhe sartorius. .ind ren.sor fascia

lata. These muscles are inserted on rhe parella

and an- flexors of the hip and/or exrensors nf the

leg. An everred ilium gives more power ro rhese

muscles since it increa.ses the lever arm ol their

action. 'J'he morphology of thc wing oi thc ilium

of Mayulestes thcreforc indicarcs a more impor-

tanr activity of thc vertébral column and coxo-

femoral articulation than in the living didephids.

Tite epaxial musculature ineiuioned above

coniributcs ro the propulsive sirokc of the hind

lirnb wilh vertical oi lioriztjnral (as in .'•eals and

walruses) movemenrs of the vertébral column.

Thüse movcmeius can occui in ail che ecoiypes

menrioned above. In rhe case of rhe arboreal eco-

lypcs (as noted above, Mdyuleues bas several lea-

lures of its fureliriib rhat can iiuerpreced as

arboreal), eversLou ol the iliac wing can be due to

leaping (lemurs, indri, spider monkey), ro lea¬

ping run [tree shrews (Jenkins 1974), .squirrcls],

or to a \xrtical climbing position as observcd in

koalas and vsnmerimes in cuscuses. Living didel¬

phids which generally (when iin.stre.ssed) do not

move ver)^ hisi, which walk on the branches with

the hcip of their grasping manus and pes and

which arc not jumpers, do not hâve an everted

ilium. The only living didephid which bas an

ilium slightly more everred than rhe other ones is

the rcrresirial genus Mvtachirm. This is probably

rclaied to the terrascansorial (Szalay 1994) mode

of terrestrial locomotion of the hrown foiir-eycd

opossum, which is ex^tremely agile on the ground

and whosc locomoiiuri is a sort of leajilng run

(Charles-Dominique pers. comm.). It is note-

110

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

worthy ihat the Plio-Pleistocene saltatorial mar¬

supial ArgyroLigns has an everted ilium (Simpson

1970). Therefore, the eversion of the ilium of

Maytilestes probably dénotés some Jeaping or

bounding ability.

The anterovcntral iliac spinc o’i Adayulestes is

clearly deflected ventrally as in Cladosictis and

Prothylacynus (but to a lesser extent). Among

living didelphids the only fbrm which shows an

anterovcntral iliac spine slightly deflected ancero-

ventrally is the terrascansorial Metachiriis, On

the médial side of the anteroventral région of the

iliac \\'ing is inscrted the quadratus lumborum,

an important flexor of the vertébral column.

This muscle was hcnce powerful in Mayulestes.

This condition Ls in agreement with the large

transverse processes of the lumbar vertebrae

(where rhe muscle originates) and with the infer-

red sirength of the crecior spinae (the ancagonist

muscle of the quadratus lumborum and psoas

mmor). On the anterovcntral iliac spine the obli-

quus abdomini mternus was also probably inser-

ted, as in many orher marsupials (Elftman

1929 ). This muscle is another flcxoi of the

trunk. Flexion of the vertébral column is an

important movement in cursorial and saltatorial

mammals and acts in synerg)^ with the recovery

stroke of the hindlimb when the animal is run-

ning or leaping (Elftman 1929). The fuct that

the anteroventral iliac spine of Mdyulcste& is rela-

tively more devcloped than in the arboreâl didel¬

phids ïndicates chat, although certainly not

cursorial, it was probably capable of relarively

fast running (probably for a reiarively short dis¬

tance) and/or some bounding. The greater deve¬

lopment of this spine- in Cladosictis and

Prothylacynus indicative of some cursorial abili-

ry (probably for short distances) in those

San^acru^ian forms and, very probably, boun¬

ding in order to seize their prey by surprise. Tins

morpholog)^ of the anteroventral angle of the

iliac wing is also présent in the ciirsosaltatorial

(but also partly fossorial) Perameles and in the

superfossorial Orycteropus. However, sincc it is

absent in orher fossorial mammals (wombar, fos¬

sorial rodents, armadillos, meliiie jnustelids,

giant anccatcr) k Ls probal^ly not relaied to dig-

ging. Although capable of a relatively fasc gallop

when chased (Novak & Paradiso 1983),

Oryctèropus is ccrcainly not a cursorial animal.

No interprétation is given here of tbc ventrolatc-

rally deflected anteroventral iliâc spine of

Orycteropm. In Perameles it is probably rclated to

the pronouneed leaping and tunning ability of

this animal.

It has bcen shown above that Mayulestes beats

scvcral fcacures that can bc relatcd to arborcality.

Therefore, if rhis mterpreration is correct, chc

everted ilium and the wclI-deVc|opcd anceroven-

tral iliac spine ol Mayulestes çoi\\à bc ittterprered

as indicating some running and/oi leaping abili¬

ty, perhaps approacfiing the exirernely fasi and

agile leaping run of rree shrews and squirrels but

certainly not as faSt. Marshall & Sigogneau-

Russell (1995r I 5 O) bave suggested that

Pucadelphys hdid both leaping and digging ability.

Sincc the scapula and the caudal vertebrae of

fhu'adelphys bave been .shown to fiave characters

compatible with arborcality (shapc and position

of the acromion and preherisile rail), an interpré¬

tation similar to that oi Aîayulestes could be

given for the everted ilium and the ventrally

deflected anteroventral iliac spine of Pucadelphys.

Thcy are therefore prcferentially related here to

leaping rather than digging function.

The tuberosity for the rectus femoris is slightly

larger in Mayuksles than in the living didelphids.

However, ii is much smaller thait in Cladosivth

and Prothylacyn-us. A wêll-developed tuberosity

for the rectus femoris is found in pcramclids,

sciurids, rupaiids, canids, leporids, aardvark,

giant anreater, some fossorial and some cerrestnal

rodents, lemurs, indri, Ar^irolagits. Pucadelphys

and Barunlestes. The recrus femoris is a powerful

extensor of the knee and a flexor of the hip. It is

an efficient actor during the stroke (if the exren-

sors of the hip prevent it from flexion) and

during rhe recovery phase. A powerful rectus

femoris is an important advantage to increase the

power and strength of the stroke. The need for a

powerful stroke of the hindlimb is easily unders-

tandablc in a fast runner, in a juntper or in very

agile and active small mammals (arbo- or terras¬

cansorial). Iv is also esscntial in an anijnal which

needs to be able to escape very quickly (because

of predators or any kind of social behaviour). In

Mayulestes the relatively small rectus femoris

(only slightly larger than in the living didephids)

GEODIVERSITAS • 1998 • 20(1)

111

Muizon C. de

would indic^ire a miich less active animal than a

tree shrew or a squirrel but, a little more than in

the living didelphtds. On the contrary,

Pucadelphys, with a vvell-developed tuberosity,

probably had a stronger rcctu.K fenioriü than in

the living didclphids wluch could bc rclatcd to

an important agility possibly as is observed in

squirrels or tupaiids. In Cladosictis and

Prothylacymis w\\\c\x are much larger animais, the

inferred large size of the recrus femoris tuberosiry

could be relaied to bounding alnliry, an action

used by non-cursorial predators which lie in wait

and which is generally cornbined with fast but

short run.

The ischiatic tuberosity of Mayulestes is smaller

than in the Santa Cniz borhyacnoids and most

living diddphids. It approaches rhe size of that

of Caluromys. A well-deveJoped ischiatic spine is

found in cursorial (thylacinc, canids, most ungu-

lates, Dolichotis), cursosaltacorial and saltatorial

(lagomorphs, dipodids rodents), fossorial (wom-

bat, naarsupial mole, aardvark, geomyids,

bathyergids, meline musteÜds), aquatic (sca

otter, seals, walrus) and arboreal mammals

[koala, lemuns, indri, avahi (airhough less develo-

ped in thosc forms than in chose mentioned

above)]. Btars aise hâve a large ischiatic luberosl-

ty, probably relatcd to their capacityfor bipedali-

ty (occasionally and for short distances). Among

Recem mammals the better dcveloped ischiatic

tuberosities are found in cursorial and highiy fos¬

sorial rnammals. In the othei Récent mammals

the development is gencrally not so speccacular.

However, a reasonably wcll-developed ischiatic

tuberosity is présent in some arboreal mammals

(phalangerids, DetidroLigus^ sciurids, tupaiids).

In other respects, it is notcworihy that, in the

living didelphids, the Lschiatic spine is gencrally

more developed in the terresrrial and arboreal-

terresrrial lorms than in the srrictly arboreal

lorms. Among fossil mammals, North American

(arboreal) and Asiade (terrestrial) multitubercu-

lates hâve a wcll-dcvclopcd ischiatic tuberosity

(respectively, Krausc âc Jenkins 198.3; Kielan-

Jawomwska & Gambaryan 1994); it does noc

seem to be well-developed in Hettkelotherinm

(Krebs 1991); it is weak but saüent in Gnbico-

nodon (Jenkjns Ôc Schaif 1988) and rounded in

Megazostrodon (Jenkins & Farrington 1976).

rherefore, lhe reduced size of ihe ischiatic spine

of Mayulestes indicaces that it was certainly not

cursorial nor fossorial. Its condition is not incom¬

patible with arboreal life but does not particularly

reinfiirce rhis hyporhesis.

The ischium of Mayulestes is relatively long when

compared to those of living didelphids (Table 7).

However, its relative length is conipatiblc with

that observed in Cladosictis and Prothylacynus. As

shown in table 7 the length of the ischium is

quite variable in the same ecotype. For instance,

fossorial mammals are considered to hâve a long

ischium. This is obviously true for ihe super-

digger Orycterofius but not for the geomyid or

hathyergid rndents. heaping mammals also hâve

gencrally long ischia (allacfaga, kangaroo). The

bcnclit of a long ischium is an increase of the

lever arm of the extensor muscles of the bip and

knee (biceps (emoris, sernimembranosus, semi-

tendinosus, quadratus femoris) and therefore a

motc powerful extension of the hind Hmb essen-

tial in digging and leaping. In tupaiids and per-

amclids the ischia arc similar in size to those of

Adayulestes and Pucadelphys (Table 7). Tupaiids

and peramelids are extrcmely active and dynamic

Table 7. — Relative length oi the ischium in various marsupials

and placentals. Lis. length of the Ischium: Lt. total length of the

innominale. AH measurements are in milllmeters.

Lis

Lis/Lt

Mayulestes (MHNC 1249)

17.7

43.5

0.4

Ciadusiciis (YPM PU 15170)

101.5

0.36

Cladosictis (VPM PU 15702)

123

0.38

Prothylacynus (VPM PU 15700)

81e

184 e

0.44

PucadelphÿsCfPfB Pal 6106)

0.41

Didefphfs

0.29

Metachirus

12.5

0.31

Monodefphis

0.32

Marmosa

17.5

0.28

Caluromys

14.5

0.32

Phalanger

0.32

Sciurus

0.36

Tupata

11.5

0.4

Ailurus

0.33

Perameles

0.4

Orycteropus

150

260

0.57

Geomys

0.3

Bathyergus

0.42

Allactaga

0.51

Argyrolagus

14.5

0.38

112

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from che Palaeocene of Bolivia

Fig. 53. — Posterior view of the proximal third of îhe left fémur in several marsupials: A, Mayulestes; B, Caluromys; C, Didelphis;

D, Pucadelphys; E. Phalanger. F. Monodelphis; G. Metachirus; H, Borhyaena; I, ProthyJacynus. Not to scale.

mammali; and the morphology of the ischia of

Mayulestes and Pucadelphys could simply bc rela-

ted to thcir agility (Perameles is also pattiaily fos-

sorial). Howcvcr. the fact that tamanduas bave

the same ratio “length of the ischium/rotal length

of the pelvis’' as squirrcls indicates that the inter¬

prétation of this feaCLirc is perhaps not as simple

as it appeïirs at flr.st.

Coxofemoral articulauon. The rclativcly open

acecabulum of the pclvjs of Mayidestes indicates a

good rnobility of the hip to a dcgrce comparable

to that obserN^d in the most arboreal didclphids

and in the Aastralian phalangcrids. As in che lat-

ter the ihickened and elevated anterior border of

the acetabulum provides a good anterior burtres-

sing System of the hip articulation. The concave

dorsal border (in dorsal view) of the acetabulum

provides a greater amplitude of abduction (and

in part of inversion) of rhe fémur which allows a

berter opening ofthe legs and gives more siabili-

ty in the case of an arboreal animal. On the

contraiy this ojndirion would be a handicap in

the case of a terresrrial cursorial animal where

the movements of ihe limbs tend to be parasagic-

tal (Kappelman 1988). li is therefore probable

that Mayulestes had a relatively more mobile arti-

culacion of the hip and greater abîlicy to have the

legs widely spread rhan Cladosictis and

Prothylacynus. The condition o{ Mayulestes was

probably approaching that of arboreal marsupials

such :is didelphids and phalangerids. It is noie-

worthy that a simÜar structure of the acetabulum

is füund in Eozostrodon, iniequeied as partially

arboreal byjenkins & Parringioit (1976).

Among living forms ir is pre.sent in rupaiids, in

most arboreal didclphids, in phalangerid.s and

petaurids. Kiclan-Jaworowska Gambaryan

(1994: 70) note rhar: “Elftman (1929) stated

chat the acetabulum is open dorsally in arbore:tl

didelphids and Pseudoi:hirus \ However, F.lfrman

(1929: 223) actually stated: ‘Mn arboreal forms,

such as Didelphys (sic) and Psetulochirus^ che cup

is more opencd allowlng greater freedom of

movemeiit of che head of ihc fémur". Elfcman

(1929: 225) aiso stated dtat "the acetabulum of

Petaurofdes is somewhat more open than chat of

Pseudochirus" and “in Petauroides the acetabulum

has a deeper notch dorsally than iit Pseudochirus

allowing e.vtreme abduction of ihc fémur diiring

gliding." In fact rhe acecabulum is never totally

open dorsally but Its dorsal border is more

concave (in dorsal view) in arboreal forms

{Caluromys, Caluromysiops) and terrestriaharbo-

rcal forms (Didelphis) than in sirictly terresrrial

forms (Metachirus).

On the proximal extremity of the fémur, the tro-

GEODIVERSITAS • 1998 • 20(1)

113

Muizon C. de

Fig. 54. — Distal view of the distal epiphysis oF the righl fémur in several marsupials: A. Caluromys; B, Didelphls; C. Philanden D.

Monodelphis: E, Metacdirus; F. Pucêdelphys: G* Phalanger H, Mayulêstes; I, Borhyaena; J, Prothylacynus: K. Thylacinus.

Abbreviatiùos: ïc. latéral condyle; me. médiat condyte. No! to scale.

chantm arc well-developed and expandcd în the

plane of the epiphysis, The size of the trochan¬

ters of Mapilestcs is similar to thac ohscrvcd in

phalangerids> slightiy greater than in the living

didelphids and much more developcd than in

the Santa Cruz borhyaenoids. Among the living

didephids, the trochanters are more devcloped in

the arboreaJ foinis {Caluromys, Didelphh^

Philander). In the cerrestrial forms {Metachiriis^

Monodelphis) their relative size varies. In

Metachiriis the greaier trochanter is relatively

high (pmbably in nelation ro ihc leaping abilirj-0

while in Monodelphis the grcater trochanter is

smaller than in other didelphids, althougli the

lesser trochanter is relatively well-developed

(Fig. 53). The development of the trochanters of

Mayulesles corroborâtes the grcat mobility of the

coxofemoxal articulation siiggestcd by the mor-

phology of the acetabulum.

The morphology of the coxofcmoral articulation

and the proximal extremity of the fémur of

Pucadelpbys are very similar to those of

Mayulestes except a .slightiy deeper acerabulum

and a slightiy lower grcater trochanter in the for¬

mer. As in Mayulesles, che morphology of the

acetabulum and proximal extremir)^ of the fémur

of Pucadelphys suggests a well-developed mobiliry

and powet of the coxofcmoral articulation and is

consistent with somc arborealit)^

Femorotibial articulation. ( hc distal epiphysis

of the fémur of Mayulestes is anteroposteriorly

longer than that of the Didelphidae and that of

the Santa Cruz borhyaenoids (T’ig. 54, Table 8).

CaluromySi the mosc arboreal didelphid, bas che

most anteroposteriorly flattened distal epiphysis

of ihe fémur while the most quadrate (in distal

view) isS found in Metachirus which is exclusively

lerrcstriul. Beivveen these two pôles are several

incermediaie types of totally or parrially arboreal

geitera. Phalangerids aiso hâve an anteroposte¬

riorly flattened distal epiphysis of ihc fémur. A

flattened distal epiphysis of rhe fémur seems

thereforc rclared ro arboreality in didelphids.

However, the distal extremity of the fémur is not

flattened in Tupaia:, Sciurus, Nasua^ Ailurus,

alihough it is so in Potos. Therefore, an arboreal

mode of life does not aiways implies the presence

oi a flaaened distal epiphysis of the fémur and

rhe proportions of rhe distal epiphysis of the

fémur of Mnyulestes (which are doser to those of

Metachiriis rh^x\ ro those of Caluromys) dn not to

argue againsr arboreal adaptation. An anreropos-

leriorly elongaied disul epiphysis of ihe fémur

indicates a greater amplitude of the movements

of fhc kncc which could bc rclaccd to somc run-

ning or leaplng ability, Highiy cursorial canids

and thylacinids and saltatorud kangaroos or Indris

bave a very irianguUr distal epiphysis of the

lemur. As a rnaiter of fact the lerre.srrial

Metacbiriis is a very active runner which shows

good leaping abilic)' and whose distal epiphysis of

the fémur is much longer anteropo.steriorly than

in the other living didelphids. Therefore, the pro¬

portions of the discal epiphy.sis of the fémur

would be more informarive on the way of loco¬

motion than on the Iiabifs (arboreal vs Icrrcstrial).

The morphology observed in Mayulestes could

pos.sibh' indicate a relatively more agile animal

than che living didelphids, capable of sonie run-

ning and/or some bounding or leaping ability.

114

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

The distoproximal élévation ol thc fémoral tro-

chlea, its dcpth and ics salicnr crcsts, when com-

pared to the living didelphids, suggest more

stress in the articulation which vvould be in

agreement with more povverftil and fasrer move-

ments. During fasr movement a deeper rrochlea

is necessarv' for a betier guiding of rhe tendon of

the vasti and reclus femoris (which in.serts on the

tibial tuberosity) and to prevenl it from disloca¬

tion. Living didclphids and diseuses which hâve

a shallow lernoral trochlea arc relatively slow

movers (although some didclphids are faster than

others, i.e. Mctachhm), Extremely fast saltatory

runners (squirrcls and tree shrews), cursorial

mammals, arbortal jumpers primates (7in'//o,

galagos, lemurs, chcirogalcids indriids) ajid salta-

torial rodents (dipodids. pedetids, some ger-

billids) gencrally have a well-marked, proximo-

distally elongated fémoral trochlea wirh relatively

clevatcd crests. As mentioned by Tirdicu (1983),

a deep aqd long trochlea with clevated crests is

lundamenral in jumping primates and allows

better rétention of the patella in its articulât

position during extension. A well-marked rro¬

chlea seems to be related to fast movements.

Slow climbers such as lofisines primates have a

wide and fiat trochlea as seen in didclphids and

phalangerids. However, thc slow nioving taman-

duas and pangolins have a very deep trochlea

with very elevated crests and ihe extremely fast

squirrels and tree slirews do not have a more

pronounced trochlea than the relatively slow

moving lesser panda. Hence. rhe significancc of

the fémoral trochlea is not clcar. Wirhin the

borhyaenoids, Miiyult'stes nesembles Cladosivtis in

having a well-marked ciochlea but not as deep as

in the fast moving jumping lemuriforms mentio¬

ned above. Borhyaemi and Prothylacynus which

have short and shallow fémoral trochleae were

certainly not a.s highiy cursorial as a rhylacine

Therefore, when c*impared with living didel-

phids, the morphology of the trochlea of

Mayulestes woiild indicate a more agile animal

probably capable of some running and/or lea-

ping> although, as mentioned above, this feature

is not as reliable as is often belîeved.

The fémoral trochlea of Mayulestes is strongly

asymmetrical and the latéral Üp is much more

elevated than the médial one. Furthermore, the

latéral side ol the trochlea extends more proxi-

mally on the anterior face of the diaphysis than

the médial skie. This is probably related to the

si/,e of the latéral œndyle which is much largcr

than thc médial one. An uuercscing comparison

can be made with the fémur of Tanumduas where

the médial condyle is the largest and where the

médial crest of rhe trochlea is much stronger and

much more elevated than rhe latéral one. Thus, a

relation «seems to exisr hemeen rhe size of one

distal fémoral condyle and the development of

thc corre.sponding crest of thc fémoral trochlea.

The élévation of the latéral lip of the fémoral

trochlea corresponds to a necessity to prevent the

joint from dislocation, more iinporiani on the

latéral side of thc joint than on the médial one.

In lyideljdns the average position of the fémur

when ihe animal Is walking is close ro horizontal,

the kiiec being sÜghtly lower than the hip

(JcrJdns 1071). In this position, a slighi dorso-

ventral (anticlockwisç for the left fémur and

clockwisc for the right one) rotation of the

fémur on ics axis will place the latéral lip of rhe

rrochlea below thc médial lip. Such a position

woLild imply rhe necessit}' for a stronger latéral

lip uf the trochlea in order to maintain thc paccl-

la (or thc vastus tendon, if there was no patella)

in the irochlear groove. A dorsoventrally rotated

Table 8. — Proportions of the dista) epiphysis of the fémur in

various marsupials. L. lenglh: W, width. AI) measurements are in

miiiimeters.

L/W

Mayulestes (MHNC 1249)

7.9

6 e

0.76

Cladosictis (YPM PU 15170)

21.5

0.79

Prothylacynus lyPM PU 15700)

0.74

Borhyaena (YPM PU 15701 )

26.7

0.74

Thylacinus

32.5

30.5

0.93

Pucadeiphys (YPfB Pal 6105)

4.8

3.3

0.68

Pucadelphys (YPFB Pal 6106 left)

4.8

3.6

0.75

Pucadeiphys (YPFB Pal 6106 right)

4.9

3 7

0.75

Pucadeiphys (YPFB Pal 6110)

4.5

3.2

0.71

Caluromys

9.5

5.5

0.58

Didelphis

13.7

9.5

0.69

Philander

10.3

0.7

Marmosa

2.5

1-9

0.76

Monodelphis

4.5

3.2

0-71

Metachirus

0.85

Phalanger

14.2

10.5

0.73

GEODIVERSITAS • 1998 • 20(1)

115

Muizon C. de

posirion of ihe fémur will hâve ihe effeci of

adducting die crus. The uiilicy of an adduLted

position of ihe crus is easily undersrandable in

an arboreal (arboscansorial) animal which would

climb grasping the branches laterally (or dorsola-

terally) as che living didelphids or tupaiids do

(Jenkins 1974, 1984). Thercfore, considcring

the position of the fémur, chc morphology of thc

fémoral trochlca oîMayulestes can be inrerpreted

as related ro arborealiry in the ca-sc of a didel-

phid-like or tupaid-like climbing srraregy (sec

below) but was probably associarcd ro fasr and

powerful movenienrs.

Another major feacure of the libiofemoral arricu-

lation of Mayulestes is the relative widrh of che

distal condyles (Fig- 54, Table 9). As in didel¬

phids and phalangcrids, thc médial condylc is

much narrowcr than che latéral one> whÜe the

contrary is obscrvcd in thc Santa Cruz borhyae-

noids. The didelphid condition varies according

to the habits of the species. The most significant

différence in the widch of the condyles is obser-

ved in Caluroinys, thc most arbojeal didelphid,

whilc in Metachîrus, a tcrrestrial genus

(Janson & Emmons 1990), thc médial condylc

is wider and the latéral condyle narrower than in

Caluromys but thc former is srill clearly narrower

than the latten 1 hereforc, it scems lhat, among

didelphids, ihe degree of arborealiry i.s invcrscly

proportional to the rclarivc widrh of chc médial

condyle of the fémur (i.e, ihc most arboreal, the

narrowest). In Mnytdesie^^ the relative size of thc

médial condyle (VCHvl/WL 0.73) is close ro that

of Mettichirm (WM/^T 0.75). It is, however,

much smaller than in the'Santa Cru/ borhyae-

noids (\VM/WL varies from 0.95 to 1.2) and it

is doser ro thc most arboreal didelphid

{Caluromys WM/WL 0.48) than to Borhyaenn

and Ctadosictis. It is noteworthy that Eozostro-

don, a genus interpreted as partially arboreal, bas

a latéral condyle of thc fémur which approxlma-

tely twice the size of thc médial condylc

(Jenkins fie Eurrington 1976). Ihc distal

condyles of the fémur arc gcjîerally subcqual in

width in the essenthilly tcrrcstrial Caenolestes and

in the cursosaltarorial Perameles. In Ptilodus,

Styginys^ ? Mvsodmct and ? pMcosmodon, arboreal

multituberculatcs Iront North America

(Krause Ôc Jenkins 1983: 221, figs 19B, 20E,

Table 9. — Relative width of the distal fémoral condyles In mar-

supials. Wl, width of the latéral condyle: Wm. width of the

médial condyle. Ail measurements are in millimeters.

WmA/VI

/Wayu/estes (MHNC 1249)

3.13

2.3

0.73

aadosictisiyPU PU 15702)

10.2

1.07

BothyaenayPU PU 15701)

12.5

1-2

PrüthylacynusyPU PU 15700)

15.2

0.95

Caluromys

3.9

1.9

0.48

Phtlander

4.2

2.3

0.54

Didelphis

6.4

3.8

0.59

Marmosa

1.28

0.67 0.52

Phafanger

7.2

4.3

0.59

Mûnodelphis

1.8

1.5

0.83

Metachirus

2.9

2.2

0.75

Pucadelphys {yPf B Pal 6106)

2.4

1.8

0.75

Pucadelphys (YPFB Pal 6105)

1.5

0.75

Pucadelphys (YPFB Pal 6110)

2.1

1.6

0.76

21 E), the latéral condyle of the distal epiphysis is

wider than the médial condylc. wliile in

Chulsanhaatar, a terrcstrial multituberculate

from the Gobi Desert (Kiclan-Jaworowska Ôc

Gambaryan 1994: fig. 17), both condyles are

subequal in width; in Kryplobaatâr, anorher ter-

resrrial mulrituberculatc from the sanie région

(Kiclan-Jaworowska fîc Gambar)'an 1994: 12),

the médial condylc is wider than rhe latéral one.

It seems, therefore, that narrowing of thc médial

distal condylc of the fémur could bc related to

somc klnd of arborealiry. The précisé l unction of

that pcculiai morphology bas not heen elucida-

ted. Flowever, ir i.s noteworthy that il is présent

in sevcral non-therian niammals \Bozo$iwdon,

Erythrotherium , Megazosfrodon, Pnlodu-U

Styginm, ? Mesodmû, ? Eucosmodon, Henkdothc’

riurn (apparently, from Krebs 1991: fig. 11,

pl. 3)). In placenta! mammals, when distal

condyles of the fémur do not bave the same

width, the narrowest is always the latéral cûndy-

Ic. Therefore, rlie presence of a médial condyle

conspicuously narrower than the médial one

could icpresent a plesiomorphy. Mayulestes

would thus retain rhe pJesiomorphîc condition

within mammals.

The médial fémoral condyle articulâtes, on che

tibia, with a reniform, elongated and concave

(almost grooved in Caluromys, Didelphis and

116

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from che Palaeocene of Bolivia

Phalanger) mcdial tibial facet which indicates

antcropostcrior movemencs and some rotational

ability of the femorotibial articulation

(Jenkins &: Parrington 1976: 424). The latéral

facet is wider, roughly quadrate and convex or

fiat and probably received the axis of rotation

(Jenkins & Parrington 1976: 42S). Cineradio-

graphic studies havc dcmonstratcd the necessity

for such a rotation durjng the représentative

phases of the walking step in DidHphh (Jenkins

1971)- It is likcly thaï rotation movcinents of che

knee are usefui too in climbing, in order to help

the inversion of the fooi while grasping the

branch on which the animal is moving (see

below). Henvever, Jenkins & Mcl.earn (1984:

216) noicd that there is no significant contribu¬

tion of feinorotibial rotation to foot reversai.

Although less pronounced than in Caluromys or

Didelphh-, the condition of Mayulestes clearly

indicates some capacity for rotation of che knec.

A simil.Tr femorotibial articulation is aiso found

in Pucadelphys, although less developed than in

mosr living didelphids.

On the proximal epiphysis of che tibia, the tibial

cuberosity is not protruding anteriorJy in

Mayulestes, didelphids, phalangerids and Phasco-

Urctos. As a conséquence, the epiphysis is flatte-

ned anceriorly and docs not hâve the crianguJar

shape obser\'ed in the Santa Cruz borhyaenoids

and in highly specialised cursorial (thyladnids,

canids, feUds> ungulates) and terrestrial saltato-

rial (kangaroos, rabbits, pedeiids) mammals

(Fig. 55, Table iO). The development of the

tibial tuberosity indicates the importance of trac¬

tion exerted on the patellar ligament or (if there

is no patella, as in didelphids) on the conjoined

rendon of the vasti and reefus femoris. The

action of these muscles is a powerfui extension of

the leg, an essential movement in fast-running

and jumping animais. On ihe contrary, rapidity

and power of tins movement is less important in

the relativcly slow-moving arborcal didelphids.

In Metachirus, a terrestrial didciphid regarded by

Szalay (1994) as subcursorial, the tibial tubero-

■sity is more developed than in the highly arbo-

real Caluromys and the proximal epiphysis of the

tibia is clearly triangular, while it is anteroposte-

liorly flattened in Caluromys and Mayulestes. In

fact, in didelphids there is a gradient of the shape

Table 10. — Comparison of the proportions of the proximal epi*

physis of the tibia in marsupials. L, anleroposterior length:

W, transverse width. Ail measurements are in millimeters.

LA/V

Mayulestes (MHNC 1249)

5.4

7-8

0.69

Cladosictis (YPM PU 15046)

19.1

19.7

0.97

Promy/acynus (YPM PU 15700)

39.5

0.91

Thylacinus

33.5

1.15

Caluromys

7.7

8.5

0.905

Didelphis

11.3

122

0.926

Metachirus

7.2

7.6

0.94

Monodelphis

3.9

0.77

Phalanger

9.2

10.3

0.89

Potos

14.8

18.5

0.8

of the proximal epiphysis of the tibia from reni-

form in Caluromys (the most arboreal didelphid)

to clearly triangular in Metachirus (a terrestrial

didelphid). In Didelphh and Philamier, both ter-

rcstiial-arborcal gênera, the morpholog)^ of the

proximal epiphysis is inteimcdiate. In Dêudro-

lagusy the arboreal kangaroo, che tibial cuberosity

is much less developed than in the terrestrial

rnacropodids. An anteroposteriorly flattened

proximal epiphysis^ of the tibia is also found in

several slow moving arboreal eutherians {Potosy

Ailurus. Tamandudy Choloepusy Bradypus). It

seems that the faster the animal is moving, the

more triangular the epiphysis will be. For instan¬

ce, the fasc-moving squirrels bave a more trian¬

gular epiphysis than Potos or Bradypus, but do

nor bave the anterior protrusion of che tibial

cuberosity observed in a cursosaltatorial rabbit.

Thereforr, the morphology of the proximal epi¬

physis of Mayulestes would indicate rclatively

slow movements, which contrasts with che rela¬

tive agility suggested by the anatomy of other

éléments of the posterania) skeleton (pelvis).

However, the Interprétation of that feature h

probably more complex .since, in che extremely

agile and fast tupaiids, the proximal epiphysis of

the tibia is relativcly short anteroposteriorly

while the distal epiphysis of the fémur is not

anteroposteriorly flattened. Furthermore, the sal-

tatorial lemuriforms (indris, galagos, lemurs)

hâve an ameroposterîorly short proximal extre-

miiy of the tibia associated wich an anteropostc-

riorly long distal epiphysis of the fémur, with a

GEODIVERSITAS • 1998 • 20(1)

117

Muizon C. de

C D

E F

Fig. 55. — Proximal view of the right tibia in several marsupials:

A, Mayulestes-, B. Dldelphis: C. Caluromys: D, Metachirus;

E, Phalanger. F, Phascolarctûs: G, Thylacinus: H. Prothy-

lacynus. Not to scale.

deeply groovcd trochlea. Therefore, the morpho-

logy of tlie proximal epiphysis of the tibia of

Mayulestes-, which Is similar to thac of the saltato-

rial lemuriforms mentioncd above, could indi-

cate sorne saliatorial ability. The combination of

an anteropoüteriorly long distal fémur with an

anteroposteriorly îïljort proximal tibia is also

found in several terrcstrial ccrCopithccids

(macaque, baboon) wliich are vcry agile and,

occasionally, good runners. l'he relative ancero'

posterior length of the distal epiphysis of the

fémur and tlie shortness of the proximal epiphy-

sis and articular faccts of the tibia would indicate

that the tibial condyles caii bave an long trajec-

tory on the fémoral condyles, which dénotés

capacity of a significant amplitude ol the move-

ments of the knee as in sahatorial Icmurilorms.

It is therefore perhaps not incompatible with an

agile and a relatively fast-moving animal. 7'hc

morphology observed on the tibia of

Prothylacynus and, to a lesser extent, Cladosictisy

with a well'developed tibial tuberosity, would

dénoté some cursoriality and bounding abiliry.

Diaphysîs of the ribîa. The weU-marked cha-

racteristic signioid rnorphology of the shaft of

the tibia (in anrerior view) of Mayulestes (proxi¬

mal chird to lialf, bowed lacerally; distal half to

rwo thirds, bowed medially) is found in ail living

didelphîds and to a lesser extern in DasyuruSy

Petauriis, Smihcbopsis. AcrobtiteSy Pseudoddms and

Dendrolagus, ail arborcal or partially arboreal

nïarsupiâl généra. A sigmoid tibia is also found

in Catnolestes and SarcophiluSy both mainly ter-

rcsirial but capable of climbing and in some

cases vcry agile (Novak & Paradiso 1983).

Pozosfrodoiiy an arboreal form according to

jenkins & Parrington (1976), also bas a sigmoid

tibia, The ribiae of rbe terrestrial Mongolian

multituberculates Chulsirnhaatiir, Krypiohantar

and Nanegtbaatar are not sigmoid, but. srraight,

slightly bent laierally and scrongly beat laterally

respectively (Kielan-Jaworowska bc Gambaryan

1994), while those of the arboreal North

American généra Pttlodus and ? Meiodma arc sig¬

moid (Krause & Jenkins 1983)- The tibia of

Hmkelatheriumy regardcd as a partLally arboreal

eupantotherc, is also sJightJy sigmoid (Krebs

1991). However, the ribiae Pacadelphys^

Phalanger and Phascolarctos^ are straight. The

tibia of cynodonts is .strongly bent laterally but

nor sigmoid. The sigmoid morphology of the

tibia is probably due to highiy vaiiable direction

of the varions tensions exerted on the bone in

the case of an arboreal form. In cursorial mam-

mais the direction of forces Is probably more

parallcl to the main axis ol the bonc, whose

straighrne.ss increascs ics mechanical résistance.

Fiirtbermore, rlic distribution of that feature

among the multiruhciculaie gênera cited above

would be indicative of a relation to some clim¬

bing ability. Tf this interprétation is correct, its

absence in ihe koalas and phalangers is probably

duc to a different n pe of arboreal locomotion in

thèse généra. It is noteworrhy that tlie tibia of

Cladosictis, Prothylacynus (probably bolli with

some cursorial abiliry. Le. tcrra-scan.sorial) and

Thylacinus (definirely cursorial) is straight. The

présence of a very sigmoid tibia in Mayulestes

ccrtatnly confirms its lack of cursorial adaptation

and probably reinforces the hypothesis of arbo-

reality. Nevertheless, biomechanical .studie.s are

118

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

needed to contimi ihc relationships ot a sigmoid

tibia to arboreality.

Astragalotibial articulation. The upper ankle

joint is importaiu in miderstanding locomotory

habits. *rhc astragalus MayuUstes is unknown,

howcver, the morphology oi the asrragalar faccts

of the tibia are highly informative. As noted

above, the flatcencd malleolus is oiientcd at an

angle of 135“ with the iransverse axis of the

tibial condyles. A conséquence f>f this morpho-

log}' is that an extension of the uppet ankle joint

will automaiically rcsult also in an inversion and

the sole of the font which will tend to face postc-

romedially. In Prothylacynus, CJadosiciis and

Thylacinus, sincc the malleolus is approximately

parallel to ihc transverse axis of the tibial

condyles, an extension of the foot will orienta te

the sole posteriorly. In didelphids, phalangerids

and petaurids, the latéral tihioastragalar faccc is

helical and an extension of rhe foot is also

accompanied by an inversion. Didelphids are

known for having very prehensile hands and feer

capable of efficient hallucial grasping (except,

probably, ihe hind foot of Chinmcctesy the aqiia-

ric didelphid). 'flte same is irue in phalangerids

and petaurids. Figure 13 of Jenkins 6: McLearn

(1984) (which is based on phorographs) well

illustrâtes how Didelphh walks on a bratïch. The

hands and fcet tend to grasp the hranch latcraJIy

or dorsolaterally which constrains the animal to

hâve irs hands and fcet facing, ar least partially,

medially. Anorher very démonstrative figure is

that of Novak & Paradiso (1983*. 24, top photo-

graph) whcre the right hindlooi of a

Calurornysiops standing on a brandi, is extended

and laces medially grasping the brandi, rhe hal-

lux being above and the other digits latéral

(Novak & Paradiso 1983; 13, 61, 64-66, 68,

70). I bave niade chc same observation on

Caluromys in captivity; in tliis genus grasping is

also often adiicwcd with digits I and H above the

branch and digits IIl, IV and V, bdow. A similar

grasping of the branch is also observed in

Tupaia, although locomotion is much fasrer rhan

in didelphids. Jenkins (1974: 98) States thaï

“when running at maximum speed on sniall

brvanches Tupaia glis supinates cach manus as

much as 90“. As the contact of both hands is

nearly synchronous, the effect is to grip the

branch beeween the righi and left manus rallier

thaii to run on ihe top of if'. The same is obser¬

ved on ihe foot whicli is also abducred and

inverced (Jenkins 1974) wheii walking on

branches less than 2 cm in diameter. Consider-

ing ihe shape of a branch, the position of the

hands and feet during didelphid and tupaid

arboreal locomotion appears to bc obvious. So

are the modifications of the upper ankle joinr

movement and articulation. Therefore, the mot-

phology of ihe distal extremity of the tibia of

Mayidestes and the rcsuliiiig inferred mtweinents

of rhe ankle are in agrccnicnt with a didelphid

type ol arboreal locomotion, although morpho-

logically different- In Pucadelphys the malleolus is

also ac an angle ot more than 90® with the crans-

verse axis of the ribial condyles. F’urthermore, in

rhis genus, the latéral tihioastragalar facct is

sUghrly helical, aldiough less rhan in living didel¬

phids. These fearurcs represent an indication of

capacity of inversion of the foot during exten¬

sion of ihe lower ankle joinr and abiliry of hind'

foot reversai as is observed in living didelphids

and many orher arboreal raarsupials and placcn-

tals (Jenkins 1974). The morphology of the

astragalotibial articulation of Pucaddphys is rhe-

rcforc indicative of arboreality. Smee ihc astraga-

lus of Mayulestes is unlcnown, It is not possible to

confirm che reversai abiliry (Jenkins & McLearn

1984) suggested by the astragalar lacets of the

tibia.

Calcanéum. The calcanéum of Mayulestes also

provides important infonnarion on rhe move-

ment of the lower ankle joint. The ecial and sus-

ccntacular faceis strongly lace medially, while in

Sipalocyon clicy face mainly dorsally and slightiy

disially (Fig. 56). In living didel|ihids and pha¬

langerids the L'cial and sustentacular lacets also

face mainly medially but lo a lesser exceni than

in Muytd.tsies\ the médial orientation is more

pronounced in Caluromys and MarnuhU than in

Dtdelpbis and Metachirm, The morphology

observed in Sipalocyon is obviousiy relared to

plantigrady. In that case, the astragalus directly

receives the weight of the body through ibc

tihioastragalar articulation with reduced shearing

forces (tangential forces). The wcight of chc ani¬

mal is transmitted to the calcanéum through rhe

ectal and sustentacular facets. Therefore, the arti-

GEODIVERSITAS • 1998 • 20(1)

119

Muizon C. de

Fig. 56. — Distal view of the nght calcanéum ot several marsu*

pials: A, Mayulestes: B. SipaIocyon:C, Caluromysr, D. Didelphis;

E, Marmosa: F. Pucadelphys: G, Metachtrus: H. Phalanger.

Abbrevialions: CaA, calcaneoastragalar facet; CaCu, calcaneo

cuboid lacet; CaCud. dorsal caicaneocuboid facet. CaCup, pos-

terior caicaneocuboid facet. CaFi, caicaneofibular facet; Su,

sustentacular facet. Not to scale.

culation musc be the most perpendicular as pos¬

sible to the direction of the force and a dorsal

orientation of the ectal facet of the calcanéum

reduces the possibilities of médial sliding of

astragalus in relation to the calcanéum. In didel-

phids, the astragalus is not posinoned above the

calcanéum but on ics médial side and the asira-

galocalcanear facets are oriented medially. As sta-

ted above, the articulât surfaces must be as close

as possible to a position perpendicular to the

direction of major forces in order to reduce the

possibilities of dislocation of the joint.

Therefore, since the upper and lower ankle joints

are grossly parallel, an increase ol the médial

orientation ol the XNtragalocalcanear articulation

will place the calcancum and the foot in an

inverted position and the sole will face medially

or ventromedially. The inversion of the foot will

tend to maintain rhe articular surface in a posi¬

tion as close as possible to a perpendicular posi¬

tion in relation ro rhe main axis of the tibia (/>.

the orientation of the main force exerted on the

articulation). Such a po.sition bas becti .shown

above to he bxsic in a didelphid-likc arboreal

locomotion, ’lliis of course docs not incan that

the most arboreal didclphid carmot walk or even

run on chc ground. whîch Cdhiromys can do per-

feedy well. It only mcans that rhese animais are

scansorial and poorly adaptecl to long or/and fast

runs. Tt is likely rhat Sipalfnyon, although cer-

tainly not as well adapted to running as a rhyla-

cine, had betrer cursorial ability rhan any

didelphid. The morphology of the lower ankle

joint ül Mayulestes wirh a medially oriented ectal

facet, is similar to thaï observed in rhe most

arboreal didelphids and is therefore compatible

wiih arboreal life. It certainly does not suggest

major tcrrestrial cursorial ability.

Table 11 illustrâtes the proportions of the tuber

calcanei of sev^eral marsu pials. Among rhe South

American taxa, Mfiyul^tes and Argytvlagiis hâve a

ruher calcanei longer rhan half of the total length

of the hone. A long cuber calcanei is obviously

an advaniage to increase rhe power of the exten¬

sion of the ankle. This movement is useful to

run (ungulatcs, canids) to dig (aardvarks, arma-

dillos, bandicoois) and to jump (kangaroos, ban-

dicootSv Argyrolagm^ jerboas, gerbils). It has been

shown above that numerous features of

Mayulestes indicate arboreaht)\ but not cursoria-

lity, Therefore, it seems more likely that its long

tuber calcanei is rclated to some leaping ability

or a leaping run as in cupaiids but probably less

agile. Digging adaptations arc improbable.

The ruher calcanei of Mayulestes is deep and nar-

row. As srated by S/alay (19^)4: .429, fig. 6.16),

this morphology indicates strong plantar flexor

musculature related to grasping and therefore

120

GEODIVERSITAS « 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Table 11. — Comparison of the proportions of the calcanéum in

various marsupials. Lt. total length; Ltu, length of the tuber. Ail

measurements are in millimeters.

Ltu

Ltu/Lt

Mayu/esïes (MHNC 1249)

9.4

5.3

0.55

S/pa/ocyon (YPM PU 15154)

0.43

Didelphts

5.7

0.47

Pucadelphys (YPFB Pal 6106)

5.9

0.5

Calurornys

8.2

3.6

0.43

Metachirus

10.7

5.3

0.5

Perameles

0.71

Argyrolagus

8.8

5.25

0.59

arboreality. A similar morphology is also found

in Pucadelphys.

‘Fhe proximodistally elongaced ecca] facec of the

calcanéum of Mayulestes and ics sustcniacular

facet, prolonged both proximally and distally,

suggests enhanced capacity of movement bet-

ween the astragalus and calcanéum (Godinor &

Prasad 1994). l’he morphology of rhe lower

ankJe joint thercforc suggests good rotational

abilicy of the toot as is observed in several livjng

arborcaJ mammals.

The large and distal peroneal process of the cal¬

canéum of Mayulestes is similar co char of

Deccanolestes hisplopi (Godinor &: Prasad 1994;

Prasad & Godinot 1994). According to these

authors, this morphology is indicative of the

rclatively large size of rhe peronel longus and

brevis and adductor digiti quinti, muscles invol-

ved in movement of eversion-inversion of the

foot. As they mentioned it indicaics “[...] fre¬

quent foot rotation movements necessary in an

arhoreal way of life” (Godinot & Prasad 1994:

80).

The susfcntaculum tali is wealccr in Mayulestes

and in most living didelphids than in Sipalotyon.

dTe large size of the sustcntaculum in Sipalocyon

is probably related to plantigrade tcrrestrial loco¬

motion. The small size of the sustcntaculum of

Mayulestes indicates that it was much iess terres-

trial than Sipaloiyon and is in agreement with the

hyporhesis of the arborcal habits.

Metatarsiis, As shown in table 12 the propor¬

tions ol the MtlII of Mayulestes are slighuly infe-

rior to one third of that of the tibia. In Didelphis,

Calurornys and Marmosa, the proportion of the

MtlII is beiween one fourrh and one fiffh the

length of the tibia, in Phalanger ir equals one

fifth and in Philander it is Icss than one fifth. In

Megazostrodon, the relative length of the médial

metararsals can be evaluated co one chird of rhe

length of the tibia. In Herikelotherium ihey are

less than one rhird of the length of the tibia and

in Ptilodus they are sitghriy longer than che ihird

of the tibia. In Pucadelphys the proportion is

intermediate betNveen one fourth and one chird,

Considering the condition of living didelphids,

it seems that the shorcer Mtlll are related ro

arhoreal forms [Calurornys, Didelphis Philander

and Marmosd) while the more tcrrestrial généra

[Metachirus and Monodelphis) hâve relatively lon¬

ger MtlII. This idea has been inipliciily expres-

sed by Marshall &: Sigogncau-Russell (1995).

However, hîghly arborcal squirrcis bave a Mtlll

approximaring one third of the length of the

tibia. Megazostrodon, llenkelosherium and

Ptilodus bave much longer MtlII than arhoreal

Recent didelphids and, neverthcless, bave been

interpreted as arhoreal respectively by Jenkins &

Farrington (1976), Krause & Jenkins (1983) and

Krebs (1991). Furrhermore, the MtTIT of

Mayulestes is proportionally longer than thaï of

Calurornys or Didelphts and it has been shown

above that this genus bears several features In its

postcran ial anatomy which are compatible with

Table 12. — Relative length of the Mtlll and the tibia In various

mammals.LMtlII. length of the Mtlll; Lti, length of the tibia. Ail

measurements are in millimiters.

LMtlII

Lti

LMtlII/Lti

Mayu/es/es (MHNC 1249)

13.3

40.5

0.328

Pucadelphys (YPFB Pal 6106)

7.1

24e

0.29

Didelphis

0.21

Calurornys

8.2

37.6

0.21

Monodelphis

6.6

26.3

0.25

Marmosa

4.4

0.21

Phalanger 1

21.5

106

0.21

Metachirus

0.27

Philander

7.6

0.185

Henkelothehum

0.28

Ptilodus

11.8

28.8

0.41

Megazostrodon

18.8

0.32

Sciurus

22.5

66.5

0.3

GEODIVERSITAS • 1998 • 20(1)

121

Muizon C. de

arborealiry. Therefore, relatively long metatarsals

appear to be nor so closely relatcd to terrestrial

habits. They do not seem to bc exclusive oi arbo-

real habits but could jast represent a plesiomor-

phic condition.

Conclusions on thc postcmmul skeletofi

The postcranial sltelccon of Mayukstes has revea-

led several features iliai indicaïc a great similarity

with the living didelphids or/and which are com¬

patible wiih arborealicy.

Charaaers relaied to arborcality are; (1) ihc pré¬

hensile lail (if actually présent in Mayulcstes)\

(2) the anterior and distal development of the

acromion; (3) thc élévation of thc spine ot thc

scapula which indicates a powcrful musculature

necessary ro the stabilicy of ihc scapiilohumeral

articulation: (4) rhe great mobility of the scapulo-

humerai articulation attested by rhe circulât

shape of ihe head, lis proximal orientation and

the relatively lovv tuberclcs; (5) rhe large sW.c of

rhe epicondyloid ridge and the distomedially

protruding medial tpicondyle which dénoté fre-

quency .md sxtength in the movements of ihe

hand and fmger.s; (6) the great Icngrh of the oie-

cranon, strongly hem anteriorly and medially

and which i.s cxcavaicd medially hy a deep fossa

for the flexor muscles of thc hand and fmgers;

(7) thc morphology of the MeV which dénotés

good grasping abilily; (8) tlie shallowness of rhe

acctabulum and the excavation of its dorsal bor¬

der which indicatc an important mobility of thc

coxofemond articulation; (9) the large si/.e of dne

trochanters of thc femur and thc medially benr

greater trochanter which aiso îndlcate a great

mobility of the hip; (10) the sigmoid morpho-

logy of the diaphysis of the tibia which indicates

the great variety of force directions exerfed on

rhe bone; (11) the morphology of the di.sial arti¬

culât surface of the tibia which indicates chat an

extension o! the frjoi was accompanied by inver¬

sion; (12) ihc narrow and proximodisially clon-

gated cctal facei of the calcanéum; (13) the

laterally oriented ecral and sustenracular facers

(Le. the lower ankie joint) which indicates chat

the proximodistal résultant forces exerted on the

joint will hc oriented perpcndicular to it (which

allows rhe best stabilicy of the joint and avoids its

dislocation) only if the foot is inverted; (14) the

proximodistal Icngth of thc sustcntacular facet

which is indicative of good rotaiional abiliry of

rhe foot; (15) the large .si/.e of the pcroneal pro-

cess which indicates the large size of ihrce

muscles involvcd in movements of eversion-

inversion; (16) the elevaced and narrow tuber

calcanei which indic.ne sirong flexors of ilie foot

and thcreftïn: good grasping abilicy.

Somc features are not strietly telated to arbo-

reality but arc compatible with this way of life

and are informative on rhe locomotion of

Aîayulestcs and indicatc a relatively agile animal

capable of boundïng. l’hcy are: (1 ) the posterior

position of the anticlinal vertebra (14) w^hich

nidicacci thaï Mnyulestes was not cursorial; (2)

thc size of the neural spine and transversc pro¬

cesses of the fîosterior lumbar vertebrae which

indicates a powerfui epaxial muscul.il urc; (3) lhe

morphology and the position of tht pre- and

postzygapohyses of thc List thoracic and lumbar

vertebrae which dénoté un important mobility of

tbe posterior vertébral coluinn; (4) thc long,

utireriorly heiit olecranon of the uina, which

indicates powcrful extensions of the cibow (for

instance for Icaping); (5) the everted îliac wing

and thc vcntrolaccrally oriented anteroventral

iliac spine which indicates a powcrful cpaxial

musculature, importance of the flexion-exten¬

sion movements of the column and mobility of

lhe coxofemolal articulation; (6) lhe relative

deptfi of the fémoral iruthlca and thc élévation

of its crcsis w'hich arc relaicd to agllity of the ani¬

mal; (7) thc flattening of the proximal epiphysis

of lhe tibia wliich is aIso found in somc saltato-

rial Icmuritorms; (8) lhe relative leiigtfi of the

tuber calcanei which indicate.s Icaping and/or

running ability,

furihermore, the élongation and the great

dimensions of the neural spine of thc axis are

indicative of a rohust nucchal musculature com¬

patible w'ith prcdaccous habits.

Uabhs /^Mayulestes ferox

The question of irboreal vs terrestrial habits bave

been dlscussed hy Jenkins âc Parringion (1976).

The aurhors stated ihat ihc behavioural factor is

essential since in arboreal groups some taxa are

terrestrial because of any kind of preferences

(feeding, physiological, ethological, etc.) even if

122

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolîvia

thcy can climb pcrfectly wcll. Howevcr, Jenkins

(1974: 91) srated rhac ‘'the evidence for locomo-

tory behaviour in boch captive and wild tree

shrcws indicates a moderatc diversicy of habitar

prefcrence. CJearly somc tree shrew specics arc

more arborcal than others; othcr spccies arc more

or less terrestrial. Perhaps the mosc significanc

fact is that ail tree shrews species can climb and,

at least occasionally, if not frequently, do so”.

riie same is true for didelphicLs, where the terres-

trial généra Metachirus and Monodelphis are

reported as good climbers (Novak & Paradiso

1983)* Hildebrand (1961: 249), who stated

(referring ro ihe didelphid gênera Metachirus,

Monodelphis, Philander, Didelphis Mamwsa) that

“the more arboreal animais differ from ihc

scmiarborcaJ and terrestrial animais in behaviour

pattern but noi in morphology. Any of them

could climb or walk well if it 'wanted' to”.

Contrary ro Hildebrand statement, clear anato-

mical différences can be related to the arboreak

semiarboreal or terrestrial habits in didelphids.

Detailed analysis of the postcranial skeleton

reveals morphological différences for almost

Fig. 57. — Reconstruction of the sKeleton of Mayufestes ferox (approx. x 0.7), The éléments of the skeleton which are missing in the

holotype are inspired by the recent arboreal généra Caluromys and Didelphis rather than the terrestrial Santa Cruz borhyaenoids. In

this reconstruction, the tail is regarded as prehensile because of the morphology of C8? and C9? and the number of caudal verte-

brae is estimated to approach 30 (as in Caluromys where it varies from 30 to 35). The manus and pes are aiso regarded as prehen¬

sile, as in living didelphids and it is hypothesized that Mayulestes could hâve used (at least sometimes) an arboreal way of

locomotion similar to that observed in the Recent arboreal didelphids.

GEODIVERSITAS • 1998 • 20(1)

123

Muizon C. de

every limb bone, somc of which can be relaied to

their babils. Furthermore, Grand (1983) bas

related rbe différence in limb proportion of

Metachirus and Momdclphis co tbcir way of loco-

motion. The most imporranr fact is certainly, as

stated by Jenkins (1974) for tree sbrews, thaï in

boch groups (tupaiids and didelphids) ail thc

species can climb well [except perhaps Meta-

chirus (Atramentovitz pers. comm.)]. This indi-

cates that for both familles arboreality is proba-

bly a plesiomorphy. Therefore, since changes in

habits is mainly due to behaviour, it is probable

that features of the postcranial skeleron of these

mammals (whechcr arboreal, semiarboreal or ler-

restrial) actually represent arboreal way of life.

Since behaviour is not observable m fossils it is

obvioas that an animal regarded as actually arbo¬

real could be a terrestrial lorm wich a skcleton

bearing arboreal features (likc Adonodetphis or

Tupaia tana). To concludc, another important

remark ol jenkins (1974: 91) is that the range of

adaptive types is probably too subtle to bc

understood in terms ot the gross categories of

arborealism and terrcstrialism. It is likely that

these two pôles are separated by a great amount

of intermediate stages reptesenting a gradient

from one condition to the other.

In the following interprétation, \\ a clear position

is taken concerning ihe way of life of Alayulestes,

it is obvious chat bchavioural parameters, which

could hâve notably modified the conclusions

exposed below. could not be taken imo account.

Given thc preceding discussions and lists ot cha-

racters, it is highiy probable chat Mayulestes was

at least partially arboreal. Its mode of locomo¬

tion could hâve approached that of tupaiids

(although certainly not as fast and agile) wea.sles.

or MetachirtiS (on thc ground). The postcranial

skeleton of Mayukites shows ability for leaping

or bounding, which is in agrecmem with preda-

ceous habits. It is likely that Mayulestes was more

agile than most living didelphids. Mayulestes was

certainly not cxclusively arboreal as is Caluromys

and it probably spent part of its time on the

ground» pcfhaps under the pressure of some ali-

mentary or crhological factors. With its short

limbs and its shon and blunt snou: (when com-

pared to didelphids), Mayié^lestes probably had an

external aspect similar to that of the living wea-

sels (Fig. 57) Although it was certainly more

arboreal (most of the vveasels climb well but are

noi considered arboreal mammals), its agility

was probably similar. Like weasels, Mayulestes

was certainly an efficient predator which could

hâve led upon the abundant fauna ol small

insectivorous marsupials (Pucadelphys, Mizqiie'

delphys^ Jncadelphys^ Tiulordia, Peradectes^ Kahsia,

Jaskhadelphys), In size, these animais are similar

TO the small rodents that make an important part

of the diet of the variou-s species of Mustela.

Weasels are known (Novak & Paradiso 1983) to

attack .sometimes animais much larger than

themselvcs {ie. adult hares). A further compari-

son of Mayulestes with Aiiisiela therefore suggests

that Urger oninivorous animais likc thc small

condylarths of T'iupampa ( Tiuelaenus, AioUnodus

or PucîDwdus) or caroloamcghiniids such as

Roberîhoffitetteria could hâve aiso repre.scnted

occasional prey for Mayi4lestes. Larger didelphids

such as AndlnodelphySj othei predaceous marsu¬

pials like Allqokirus and the large (for the fauna)

pancodont Alctdedorbignya are less probable to

hâve represented easy prey for Mayulestes.

Among the non-mammalian hiuna. the lepto-

dactylid frog Estesius is very likely to hâve repre-

sented an important part of the diet of

Mayulestes.

Evolution of the locomotion and habitat préférences

of the borhyaenoids

The othet borhyaenoids known by partial skele¬

ton arc thc four généra of thc middic Miocene

Saniu C-'iu/ beds of Patagonia. Cladosictisy

Sipalocyon, Prothylacynus and Borhyaena and the

skcleton of Lycopsis from chc laïc Mioccnc of La

Venta (Colombia, Marshall 1977a). Thcy are

much larger animais than Mayulestes and range

in size from a .small fox to a small beat. They

were terrestrial and show some cursorial adapta¬

tions, although they are nor as highiy adapted as

thylacines or canids Several features of their

postcranial skeleiOTi indicate an intermediate

morphology between Mayulestes and thylacines.

They are: (1) the anterior position of the anticli¬

nal vertebra (on the last choracics); (2) the

straightness of the ulna, whose olecranon is long

but not bent anteriorly (in Rorhyaena the dia-

physis of the ulna is even bent posteriorly as in

124

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

all cursorial mammals): (3) tlic straightness of

rlie fémur, whose proximal epiphysis is not bent

medially; (4) rhc straightness of rhc tibia; (5) the

srrong development of the tibial tuberosity;

(6) the anteropt^sterior orientation of the médial

malleolus of the tibia, which restricts the ankle

movemcnc to flexion-extension; (7) the réduc¬

tion of the hallux.

However, some features clearly Indicate thar the

Santa Cruz borhyenoids were not highiy curso¬

rial mammals. They are: (1) the development of

the epicondyloid crcst and médial épicondyle ot

the humérus, indicating a good mobility of the

hand and fingers, which is generally not found

in cursorial mammals; (2) the absence ofa poste-

rior curvature of the fémur, a fcature présent in

thylacines and canids; (3) the weak (when corn-

pared to cursorial mammals) excavation of the

fémoral trodilea, svhieh dénotés slower and wea-

ker movemencs of die knee (in othet words, less

srress in çhe femorotibial aniculâtion); (4) the

semiplantigrady attested by the morphology of

the tarsus.

Among the Santa Cru'/ borhyaenoids, however,

some différences can be noted in the anatomy of

the limbs bones and, therelure, in morphotunc-

uonal interprétation. In Profhylaiynui, the humé¬

rus has a felativcly circular head as in Maya lestes,

which indicaces a greater mobility of the shoul-

der rhan in Cladosictis, where rhc hun^cral head

is clearly elongated anteroposteriorly (Table 3)

and marches the W/l. ratio observed in

MetachirusTfiylncinus, respectively, cursosal-

tarorial and cursorial marsupials (che humérus of

Borhyaena is unknown). Ncvcrclielcss, because of

the higher greater tuberosity of its humérus, the

-shoulder of Prfnhylacynus musc hâve been less

mobile than chat of Maytikstes. In Prothylacynus

the presence of a deep fossa on rhe posterior face

of the lesser tuberosity and médial border of the

head indicates a powertui accessory head of the

triceps. This condition dénotés important capa-

cicies of extension or rétention to flexion of the

elbow. The cxtremely dcveloped entepicondylar

crest ot the humérus is related to good mobility

of che manu.s (pronation-supination and flexion-

extension). The capitulum of the humérus and

the humerai articulation of the radius are oval-

shaped (only slightly wider than long) which

indicates a fairly good capacity for pronation-

supination movements. The radius of Prothy-

lacynus bears aiong its shafe (as in Caluromys) a

wcll-developed latéral crest which receives part of

the origins of the pronator quadratus (on che

posterior sidc) and abducror pollkis longus (on

the anterior side). These muscles are generally

artached on the inrerosseous membrane and the

presence of a weU-de\'cloped médial crest of the

radius, which partially fills tlic interosseous space

indicates much stronger muscles, l'he medial

face of the olecranon of chc ulna bears a deep

medial tossa for the flexor of the manus and fin¬

gers. Therefore, the morphology of the elbow

and forearm oi Prothylacynus indicates an enhan-

ced mobility of the elbow, wrist and fingers.

In Borhyaena, the humérus is unknown, but lhe

ulna is highiy informative. The shaft ot the bonc

is slightly rccurvcd posteriorly and the apex of

che olecranon extends poscerodistally, whercas, in

Prothylacynus the ulna is globally straight but

slightly convex posteriorly. at the Icvel of the

radiohümeral articulation. In CLidosictss û\t ulna

is inrermediate ben^^en thosc ol Borhyaena and

Prothylacynus. Contrary to the condition obser¬

ved in Protbylaçyjius, the flexor fossa of

Borhyaena, on rhe medial face of the olecranon.

is shallow as in rhylacinids and canids. The bcak

ol rhe olecranon protrudes anteriorly to a greater

extent than in Ptothyalcinus and the greater sig-

moid cavity is less open than in Prothylacynus.

This condition of the elbow articulation of

Borhyaeyia indicates a greater stability ôf the joint

rhan in Prothylacynus in order to tolcrate a gréa-

ter stress (for instance when running). The

humerai articulation of ihe radius in Borhyaena

(W/L ratio - 0.59) and Cladosictis (3X^7L ratio =

0.606) is much more transverse than in

Prothylacy7ius (W/L ratio = 0.77) and the latéral

border of rhe shaff is roünded and bears no crest

as is observed in Ikothylacynus and Caluromys.

This morphology is indicative of lesser mobility

of tbe forearm, manus and fingers than in

Prothylacynus and is in agreement wirh better

cursorial abilities. The condition of che humérus

head of Cladostctis (much longer than wide)

dénotés a prédominance of parasagictal move¬

ments of the forelimb. It is therefore possible

that Borhyaena and Cladosictis were, at least par-

GEODIVERSITAS • 1998 • 20(1)

125

Muizon C. de

rially, tursorial or semicursorial. Ih'othylacynm

has a forelirtib compatible wiib agility and good

capacity of grasping (possibly sonae clinibing).

However, lhe protruding tibial tubcrosity of^

ProthyUicynus aiso indicate.v a gréai power ol thc

knee extensors compatible with some running

(or/anj bouncÜng) abilify (the tibia of Borhyaena

is unknown and that of Cladoùcth also bas a

strong tibial iiiberosity). The locomotor}' habits

and orher limh uses of ProthvUtyfius could gm.ss-

ly approach iliose of living bears or large felids

which hâve fairly mobile foreliinhs (allhough

very probably less rhan ProthyltiODius) and which

arc capable o( very tast runnitig (cspccially

felids). Prothyldcyniis could bave liad a forelimb

slightly more agile than rhar of living bears (wiih

reasonably good cliinbing ability as bears) and

could bave been a slightly bcltcr runner than

bears (bears bave an anccropostcriorly shorter

proximal epiphysis of thc tibia) b\ït certainly not

as fasc as a felid. It Ls very probable that Borby-

aena was a better runner and liad lesscr mobiÜty

of the forcarm than Prothylneytim. Borhyaend antl

Cladosîctisy secm to hâve initiatcd a cursorial

adaptive rrend whose extrême spécialisation is

observed (among marsupials) in thé rhylucine.

Protbylacynus and Borhyacnti are similar in sue

and coexisted in rhe sanie environment of the

Santa Cruz beds; they should tbcrcfbre hâve had

different ccological niches. As a marier of face,

thc morphology of chcir hmb bones certainly

indicates difîcrent habits> although thc détermi¬

nation of their mode of locomotion and other

use of tho forclimbs still romains poorly dcfincd.

The Santa Cruz borhyaenoids werc efficient pre-

dators, although to varions degrees (Sinclair

1906). Sonie feamres îndicatc rhat rhey werc

more predators rhan scuvengers. They are: (1)

the very largo and elongaied morphology ol the

neural spine ol thc axis which dénotes ihc

strength of thc ncck musculature nccessary when

seizing u prey; (2) ihc eversion of thc iliac wings

which indicates a powerfui epaxial musculature

necessary for leaping onto prey; (3) the présence

of large claws; (4) the présence of some rumiing

ability as indicated hy the morphology of the

knee.

Therefore, die large and itiedium-sizcd Miocène

borhyaenoids were terrestrial animais (with per-

haps -somc arboreality in Protbylacynus) capable

ol relatively efficient running (to a greater extent

in Borhyaena) but probably for a short distance.

Sincc they werc predators, in order to seize prey

such as the iclacively fast-runnlng notoungulates

and liiopierns, it is likely thaï they liad to hunt

lying in waîi as ambush predators, bounding on

their prey ai the appropriatc moment in a way

somewhai comparable to that ol the living lelids

(this is especlally true for Prothylaiynus). Bccausc

of ilicir relatively short limbs (Sinclair 1906) and

the faci thaï they were scmiplantigradc, k is pro¬

bable chat rhey werc not capable of such a fast

running as thc living felids and, hcncc, were pro-

bably nor as efficient.

Mayn/estes, the oldesr known borhyaenoid, was

partially arboi'eid and ils nu'jde of locomotion

was probably a moderately fast weascl- or

tupuiid-like leaping run. (.icoiogically younger

borhyaenoids from the Miocene arc terrestrial

and probably scansorial with good ability for

bounding. Borhyaena nnci (.'/adasktis were proba¬

bly better runners than Protbylacynus^ which

could hâve had some climbing ability.

Way aj lift of Fucadelphys

Marshall ik Sigogneau-Russell (1995) hâve sug-

gested a terrestrial mode of life for Pucadelphys, a

didelphoid from Tiupampa. However, diis study

lias pointed ont several tearures of Pucadelphys

which indicate that it was. at least partially, urbo-

real. 1 hc features mentioned ahove are: (l) pré¬

hensile tail {conPn Marshall Sigogneau-Riissell

1995)i (2) large acromion developed antertorly

and distally: (3) humerai head approximately as

wide as long with relatively low ruhercle deno-

ring a good mohilir}' of rhe shoiilder; (4) very

laig;e epicondyloici crest and meihal cpic<indyle

w'Iiich indicate important capacity for flexion

and extension of thc hand; (5) olecranon of the

uina strongly bent anteriorly; (6) olecranon benr

medially, wirh a deep médial fussa fur the flexors

ol rlie manus and digit; (7) salicnt crest of the

pronator quadr.ttiis on che fncdial sidc of thc dis¬

tal extremity nf thc diapbysis of the iiina;

(8) acetalnikim relatively open with an cxcavated

dorsal border (concave latcrally), which indicate

a good mobility of die coxofemoral articulation;

(9) strong development of the fémoral crochan-

126

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

ters which also îndicates good mobility of thc

hip; (10) helical astrag.dar facet on thc tibia,

which allows inversion niovemenis of rhe f'oor;

(11) large peroneal proccs.s of ilte calcanéum

which also îndicates frcciuent invcrsion-eversion

movements of the loot; (12) ele^^ated and narrow

cuber calcanei which dénotés strcnglh of the

digital flexors ol the pcs; (13) sustcntacular hicct

medially orienied and di.srally extended; (14)

capaciry of hind foot reversai as inditaied in

Marshall &c Sigogneau-Russell (1993, figs 25A,

41,45 and 48) and as corroborated by the mor-

pholügy of the a.stragalai trochlea observable

after further préparation of specimen

YPFB Pal 6106.

Considering rhc morphology ofirs lumbar vcrte-

brae and ilium, Puatdelphys was probably fairly

agile and was pmbably capable of leaping. as sta-

ced by Marshall àc Sigogncau-Russell ( 1993 ).

Digging ability suggcsred by thcse authors is pos¬

sible although Pticadclphyi docs not show aiiy

undoubted and exclusively fossorial adaptation

such as fiiigcr modifications (stoutness o( thc

mctapodials, large claws). In lact, many tion-fos-

sorial mammals can dig (somc dasyurids, many

canids, lagomorphs, many non-liypcrfossorial

rodents, somc viverrids, tcnrccids. soicnodoncids,

somc soricids, some macroscclids). Flovvcver, thc

association of arborcal adaptations to thc lack of

obvions fossorial adaptations seriousiy icduces

the credibility of truc fossorial habits in

Pucadelphyi. l'Iic spécialisations of dic pcivis of

PucadeLphys inentioncd by Marshall &

SigogneaU'Russell (1993: 149, 150) arc also, as

the authors mention, leaping adaptations.

Furrherniorc, thc authors' compare Pnciidtlphys

with Penimelvs which tlicy consider as a digging

form. If it is truc thaï Peramelvs foiagc.s ihe

ground in scarcli for food, it is also an cxtrcmdy

agile runner and Icapcr which is also responsihic

for thc morphology ot ics ilium. Vhis well-

known adaptation of Perameles (Novak iU

Paradiso 1983 ) ks never mentioned by

Marshall Sigogneau-Russdl (1995). (îiven the

discussion above, it secrm more probable that

locomotion in Pucadelphys w-as a sort of leaping

run sirnilar tn ihat of Metachirus^ although ])ro-

bably not so agile, (nfelli et al. (1988) bave

shown that, in primitive tribosphenid mammals,

"rerrcstrial spccies arc clcarly distinguished from

arboreal form on thc basis of their more heavily

piticd and deepiy striated crushing and shearing

surlaces rcspectivdy on molar crowns^. These

microwears, in cerresrrial forms is probably due

to thc dust, sand grains and eurth which is inges-

tcd with aliments on thc ground. It is llkdy that

tins condition should bc grcatly enhaneed in the

case of a fossorial specic.s. Flowever, since

Pucadelphys was probably boih arboreal and icr-

rcstrial il is probable that the observation of wcil-

marked microweur feaciircs would not be of gicat

significance. Considering thc influence ot ctho-

logical factors in arborealit)' vs terresCrialit)', it is

noi possible lo déterminé il Pucadelphys was

more ur l-ss tcrresirlal ilian arboreal. It is clear,

however, thaï it liad arboreal ability. It may bave

rcprcscntcd a stage close to the plesiomorphlc

condition.

Arborealiîy vs tenestriality

The interprétations presented herc of thc way of

lite ot Mayulestes and Pucadelphys rcquire the

combined présence of a préhensile lai) and a rela¬

tive agilit)' with reasonable running and leaping

ability. Flowever, as mentioned by Cartmill

(1974: 31). no living animal with a prehensile

lail bas an extrcmdy agile arboreal locomotion

except tIic acrobatie ceboids^ Ncvcrthdcss, the

combination of a prehensile rail with agile Ica-

ping and running is well kuown in Metachiriu, a

icrrcstiial didelplûd. In tins case, the prehensile

tail oï Aletachiriis is noi used for climbing but

probably represents a plesiomorphic feature,

héritage of tlic arboreal ancestor t)f Aletachicus.

Thcrcforc, a plausildc interprétation could be

that Maytdestes and Pucadelphys wert bodi arbo-

rcLil and tcrrestrial (which is highly probable)

and thaï their agilUy was inainly used on the

ground. Their arboreal locomotion was probably

more cautions which would be in agreement

with Cartmills statement. Two orher interpréta¬

tions in agreerneiit wiih CariïT)il!’s scatement

could bc considered: (1) Mayulestes was terres-

trial but recenily evoivcd from an arboreal ances¬

tor, thus rctaining several arboreal features on its

skeleton; the agilîty of Mayulestes would there-

lore be derived a.s is that ol iMetnehirus', (2)

Mayulestes was mainly arboreal but is derived in

GEODIVERSUAS • 1998 • 20 jl)

127

Muizon C. de

having acquired agility and losi a great part of

the prehensility of its rail; Mayulestcs woidd hâve

had an arboreal locomotion close to char of

squirrels or rree shrews. This incerprecation

implies char che plesiomorphic condition is a

relatively slow locomotion with a prehensile tail

(didelphid-like) which is far from being demons-

trared. In orher respect, ir is noteworthy thar

arboreal condirions in che carly Palaeocene do

nor hâve ro be similar to chose in the présent.

The fact chat some feacurcs of Mayulestes and

Pucadelphys indicate agility combincd with a pre¬

hensile tail is perhaps an evidence which contra-

dicts Cartmils sratement. In fact, ic is probable

that the actual locomotor hiology of Mayulestes

was more cclcctic thon the four hypothe.ses pro-

posed above. The arboreal features oi Mayulestes

indicate chat ii was very probably capable of

climbing, it probably had a prehensile caif ir was

probably partially terrescrial and ir was relatively

agile. How these four characteriscics combine

and their relative importance in the biology of

Mayulestes is difficult, perhaps impossible, to

establish so far. It seems reasonable to conclude

that Mayulestes had good potcntial for arboreality,

agility and prehensility of rhc tail; Mayulestes is

not an extremely derived arboreal mammal and

cercainly was also partially terrestrial.

Thereforc» analysis of the postcranial skeleton of

Mayulestes shows that the évolution of borhyae-

nuids is characterized by a loss of arboreality.

This would indicate that this adaptation is a

marsupial symplesiomorphy wirhin the superfa-

mily. Furthermore, the almo.st universal arbo¬

reality (or climbing ability) of the didelphoids

(including Pucadelphys) also indicates that arbo¬

reality probably represents the plesiomorphic

way of life of this group. The morphofunction-

nal study of che Tiupampa marsupial skeletons

therefore seems ro corroborate the conclusions of

Szalay (1984: 254) that the stem marsupials were

primarily arboreal mammals.

GENERAL CONCLUSIONS

Mayulestes ferox is the oldest known borhyaenoid

represented by a skeleton and is the most primi¬

tive member of the superfamily. It is represented

by one of ihc two oldest known skeletons of

American marsupials. Dental morphology clearly

indicates an animal engaged in the way of hyper-

carnivorous speciali.sation, although relatively

discretely. The development of a prevallid-

posrvaltum shear (and the rclatcd transforma¬

tions of the toorh morphology) is a common

adaptation amoug meat-earing mammals and

ha.s appeared at Icast in six different groups of

mammals and very probably severai rimes within

each group. ît is therefore a highly homoplastic

apomorphy which has reduccd phylogcnetlc

value. One apomorphy oÇ Mayulestes also found

in che comemporaneous genus Allqokirus is the

réduction of the enroconîd, which is regarded

here a synapomorphy of the family Mayules-

tidae. Becausc of che presence of a reduced enco-

conid> the Mayulestidae cannot be ancestral to

the late Palaeocene genus Patene from Itaborai

which has a well-dcvcloped emoconid. On rhe

other hand, chey constitute a good potenrial

dental ancestor for cf Nemolestcs sp. from rhe

same localiry, a primitive Borhyaenidae which

has losr its emoconid and has a reduced metaco-

nid. l lowever, cranial remains from this genus

are needed to test this hypothesis.

The presence. in Mayidestes^ ol livc upper and

four lower incisors (I4/i3, in othex borhyaenoids)

reintorces the fact that I5/i4 is the primitive

marsupial incisor formula.

The skull of Mayulestes Is' highly inlormative on

marsupial basicranial structure. Comparison

with rhc other aldest-known marsupial skeleton,

Pucadelphys audinus^ rcvcals that the rympanic

process of the alisphenoid and the alLsphenoid

hypotympanic sinuses, characteristic of marsu¬

pials, are very likely to hâve appeared scvcral

times within marsupial évolution. Therefore,

they shüuld not bc comsidered as diagnostic

synapomorphics of the Mctaiheria. In ail

borhyaenoids the squamosal participâtes in the

formation of the alisphenoid sinus. 1 his charac-

ter State, which is absent from ail the other mar-

supiâls represents one of the main synapo-

morphies of the superfamily. Most of the other

crânial features of Mayulestes are therian. tribos-

phenidan* or marsupial plesiomorphies.

The postcranial skeleton of Mayulestes clearly

indicates an animal that was at least partially

128

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

arboreal. Many fcaturcs arc shared with thc

didelphoîds (including Fucadelphys) which arc

primitjvely arboreal (sccondarily terrcsrrial for

some torms). Puaidelphp is regarded here as par-

rially arboreal concrary to fornier interprétations

(Marshall & Sigogneau-Russell 1995). Among

the many arboreal features borne by Mayulestes

some arc shared with other arboreal cherian and

non-therian mammals. A mcdial distaJ condyle

of rhe fémur thar is narrower than the latéral

condylc is found in Mayulcsmi PucadelphySs the

living didciphids, phalangcrids, DeudroLagus.

Eozostrodofh Megazosreodou, Eryrhrotherium,

arboreal multitubcrculatcs (Stygimys^ Ptilodus.

? Eucosmodofh ? Mi'sodma) and Ilenkelotherium.

A sigmoid nhîa îs fbiind in Mayule^teSy in ail

living didelphids, in phalangcrids (less marked

than in didelphids), in EozosU'odan, in arboreal

multituberculatcs {Ptilodus, ? Mesodma) and in

Henkelotherium. Considering their distribution

these two features are probably plesiomorphîc-

However, it is noteworthy rhat they are présent

only in living arboreal marsupiâls (although not

in ail of them) or in fossil mammals interpreted

as arboreal. They are absent in rhe Asian miilti-

tubcrculates, regarded as terrestrial by Kielan-

Jaworowska 6c Ciainbacian (1094). Therefore,

these features arc very probably related to some

kind of arboreality. Tl these hypothèses arc cor¬

rect, rhey would indicare rhat some arboreal Tca-

lurcs are picsiomorphic for mammals. This docs

not impiy rhat arboreality is a plcsiomorphic fea-

ture for mammaLs as hypothesized by Matthew

(1904). Furthermorc, as clcarly stated by

jenkins & Farrington (1976) arborealit)' is a very

relative State since thc influence of behaviour

may greatly influence the habits of mammals.

Moreover, as stated by Jenkins & Farrington

(1976: 425, 426) the question of arboreal vs ter-

rcstrial spécialisation in diniinutive mammals is

probably invalid since ac "ground level, obstacles

rhat requîtes climbing arc common and végéta¬

tion provides a continuum of substrate possibili-

ties between the terrestrial and rhe arboreaF’. Ir is

also noteworthy rhat sevcral Late Cretaceous

mammals from Mongolia (eutherians and rnulci-

ruberculates) hâve been regarded as terrestrial

(Kielan-Jaworowska 1977> 1978, Kielan-

Jaworowska & Gambaryan 1994); undoubtedly

this is related to the very atid environment in

which they weie living. Although thc study of

rhe postcranial raotphology of Mayulestes and

Pucadelphys does not confirm Marthews (1904)

theory, it docs not contradict it and certainly

rcinforces the hypothesis of the ancestral arbo¬

reality of marsupials.

Acknowledgements

Research in the field and in Muséums (American

Muséum of Natural HLstory, New York; US

National Museurn of Natural History,

Washington D. C.; Feabody Muséum, New

Haven) was funded by the Institut Français

d'F.tudcs Andines (IFEA, Lima, Féru). The

author ts a member of thc Centre National de la

Recherche Scientifique (CNRS, Paris, France)

and part of the research was undercaken witli

funds of this insttruuon and thc Muséum natio¬

nal d'Histoirc naturelle (MNliN, Paris. France).

The Field expédition where the holotype of

MayuEstes was found ( 1992) was carried out in

collaboration with thc Asociacion Boliviana de

Palcontologla and the Fundacidn para las

Ciencias (Cochabamba, Bolivia). Spécial thanks

arc due to R. Céspedes Paz, j. Jacay Haraché and

R. Suarez Soruco for rheir collaboration and

logLstic support. The holotype of Mayulestes

ferox is property of the Museo de Historia

Natural de Cochabamb.i (Bolivia); rhe specimens

of Puctxdelpbys aadinus .ire property of rhe

Centro de Tecnologia Petrolera de Yacimicntos

Peiroli'feros Fiscales de Bolivia (YFFB, Santa

Cru/., Bolivia). |. Cuisin, M. Tranier, F. Renoulc

(MNHN, Paris), R. Emry (US NMNII,

Washington D. C.), M. Novaeek (AMNH, New

York) and M. A. Turner (Feabody Muséum,

New Haven) gave access to collection under

their care. Spécial thanks are duc to C. Tardieu

(CNRS, Paris, France) who kindly spent many

houfs with me discussing marsupial functional

anatomy. M. Atramentovitz (CNRS, Brunoy,

France), P. Charles-Dominique (MNHN,

Brunoy, France) and M.-L. Guillemain

(MNHN, Brunoy, France) provided very useful

unpublished data on Recent didelphids biology

and gave access to their Caluromys husbandry.

Review by R. L. Cifelli and F. S. Szalay allowed

GEODIVERSITAS • 1998 • 20(1)

129

Muizon C. de

considérable improvement of the manuscript.

This Work bas benefited from much fruitfui dis¬

cussions wifh R. Cifelli, L. Ginsburg, E. Jaillard,

M. McKenna, M_ Novacek, B. Senut,

D. Sigogneau-Russell, E. Sargis. F. S/alay.

E. Sargis kmdly helped in measuring rhe teeth of

Mayulestes ar rhe Analyrical Micruscopy and

Imaging Cenrer jn Antbropology (Hunier

College, CUbTV’). Phorographs are by D. Serrette

(URA 12 CNRS); drawings are by F. Pilard

(MNHN) except figure 45 which is by

M. Parrish (US NMNH, Washington D. CO¬

REFERENCES

Anieghino F. 1894. — Énurnératioti synopiiqtic des

espèces de niammifères fossiles des formaiions

éocèneS de Patagonie. Bo/e’ün de la Academia de

Ciendaa df^ CAtehdm 1259-452; Ameghino Ubras

complétas )ü; 594-863.

Archer M. 1976. — d’he basicranial région ol marsu-

picarnivoncs (Marsupinlia), rclation.ships of cainl-

vorous marsupials and uffinities uf the

insecrivorous marsupial peramelids. ZooUtgical

Jottrîiai uf the finrieau Soricty oj London 59;

217-322. ‘

— 1978. — rhe nature of the molar-premolar boun-

dary in m.irsupials and rcintcrprctacion of the

homology of marsupial clieekteeth. Meinairs of the

Queensland Muséum 18; 157-164.

— 1982. — A review of Miocene thylacinids

(Thylaanidac. Marsuplalia), die phylogenecic posi¬

tion of the d'hylacinidae and the problem of aprio-

risms in character analysis; 445-476, m Archer M.

^narsupiuL. Royal Society of New

Soudi WaJes, Syil nrv.

Barnett ( '- 11, A* Napier J. R. 1953- — Ihe form and

mobiliry of rhe fibiifa in meratberian mammals.

journalofAnatomy 53: 207-213.

Berkoviu- B. K. B. 1978. — Tooth ontogeny in

Didelphn vtrxinntnu. Australian Journal of Zoology

26; 61-68.

Bonaparte ]. F- &c Morales J. 1997. — Un primirivo

Notonychopidae (? Litopterna) del Paleoceno

Inferior de Piinra Peligro, Chubut, Argentina.

Estudkh gcolàgicos51s (5-6); 263-274.

Bonaparte J. &c Rougiei Cî, 1987. - MamlTeros del

Creticico Jnferior de Patagonia. !V Corigreso

Latinoamericano de Paleantologia^ Santa Cru/,

Bolivia 1: 343-359.

Bonaparte J. F., Van Valcn L. M. Kramartz A.

1993. — La fauna local de Punta Peligro,

Paleoceno inferior, de la Provincia dcl Chubul,

Patagonia, Argentina. Evolutwnatji ntonographs 14:

1-61.

Büvvn r. M,, Kr.iusc M., Wing S. J.< Fleagle J. G. L.,

Tiffney B. K., Sinions L. L. &C Vondra C. F.

1982. - rhe Fayum primate forcsi revisited.

Journal ofhuman évolution 11: 603-632.

Cartmill M. 1974. — Pads ainl d-aws in arboreal

locomotion: 45-83, in jenkins F. A. (cd.). Primate

Loivmotwn. Academie Pres.s. New York.

Cifelli R. !.. Î990a. — Cretaccotus mammals ot Sou¬

thern Utah. I. Marsuplals from the Kaiparowits

Formation (Judlthian). Journal oJ Vertebrate

Paleontolo^' 10 (3): 295*319.

— I99(ib. — Creiacêous mammals from Southern

Utah. ÎT. Marsupials .tnd marsupiaJ-likc mammals

from the Wahweap Formation (carly Campanian).

Journal of Vertebrate Paleontolagy 10 (3): 320-331.

— 1993a. —Early Cretaceous mamma) from North

America and the évolution of marsupial dental eha-

racters. Procecdings of the Naiionai Acudemy of

Sciences USA 90; 9413-9416.

— 1993b. — Theria of merarherian-euiherian grade

and the origin of (narsu[>ials: V5}luine 1: 205-215,

in S/.alay F. S., Novacek M. J. iSc McKenna M. C.

(eds), Mammal Pbylogeny, Springer-Verlag,

New York.

Cifelli R. L. &: Eaton J. G. 1987. — Marsupial from

uhe earliest Lite Cretaceous of Western Ü V Nature

325: 520-522.

Cifelli R. L. Muizon C. de 1997 — Denriiion and

jaw of Kuhùpel/ia juddi. a primitive marsupial or

ne-ar-mar.supial from the médial Cretaceous of

Utah. jouniaiofAlammalian F.vohiiionÀ (in press).

Cifelli R. L. &: Muizon C. de 1998 — fvlarsupial

inammal-S boni tlie Upper Cretaceous North hforn

Formation, central \SxÀ\. Journal of Pahontology 69

(in pre.ss).

Cifelli R. L., Sherburn Iv W. & Larson L. M.

1988. — Dental wear and habitat préférence in

primitive tribc»sphciiid mammal.s. Journal of

Vertebrate Pali'ontology 12 (3) Suppl.: 12A.

Cioclîon R. L. &: Corruebini R. S. 1977. — The

coracoacromial ligament and projection index in

man and other anihropoid primates, yrn^rwr// of

Anatomy 124: 62'^-632.

Clcrnen.s \V. A. 1966. - Fossil mammals ol the type

L.ance Formation, Wyoming. Part 11. Vlarsnpialia.

Vnivenity of California Publications in (teological

StmeesÙl'Ànt

Corruchini R. S. Ciochon R. !.. 1976. —

Morphornetric afHniries of the human shoulder.

American Journal ofPhysicnl Anthropolop 45: 19-38.

Coues E. 1872. — The osteology untTniyologv of

Dideiplyys virgimana. Memoirs of the Boston Socie^

ofNaturul Histor)' 2: 41 -154.

Croclict J. Y. 1980. —Les Marsupiaux du Jertiaire

d Europe, Fondation Singer-Polignac, Paris, 279 p.

Crompeon A. W. Ôc fenkins F, A. 1979. — Origin of

mammals, Mesozoit Mammals. 59-73, in

Lillegravcn j. A., Kielan-Jaworowska L, &

Clemens W. A. (eds), The First Two-Thirds of

130

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Mammafian History. Universiiy o( California PrcSvS,

Berkeley.

Crompton A. W. l.uo Z. 1993. — Rclationships oF

carly main mais, Shioconodorti Morgdntfcodon oehleri

and Dinnttheriunr. 30-44, im Sxalay F. S.,

Novacek M. J; & McKenna M. C. (eds) Mammal

Phylogeny. Mestmnc diffêrentLatiôfu Multituberai-

bues, Monotrcma. etiny Therinm and Marsupials.

Springer-Verlag, New York.

Dashzeveg D. Kielan-Jaworowska Z. 1984. —The

lower jaw of an aegialtjdonrid rnnrnmal frorn ihc

Early Cretaceous of Mongoüa. Zoologicul Journal of

the Liniiean Sonnyof London 82: 217- 227.

Dom R., Fhhcr B. L tk Maitin Cî. F. 1970. — The

venons System of the Kead and neck of the opos¬

sum (Didelphis vifxintana). Journal of Morphology

132:487-4%.

Eaton J. G. [993. — Tlierian niammais [rom the

Cenomanian fCpp^r Crticaceons) Dakota

Formation, soiuhwestern Utah. Journal of

Vertehratc Paleouiulogy 15(1): 105-124.

HiFtman H. O. 1929. — Funcrionul adaptations of

the pcivis in marsiipiaLs. Bulletin of lhe American

Muséum ofNatmal Hishny 58: 189-232.

Evans H. h. &■ Cbristcnscn G. C. 1979. — Anatomy

of the dog; V41. Saunders Philadelphia.

Flcagle J. G., Simons H. L. & Conroy G, C. 1975- —

Apc limb bone from the Oligocène ol Egypt.

Science 189: 135-137.

Fox R. C. ] 9'^5. — Molar structure ànd function in

Early Cretaceous mammal Pappothenum\ evolutio-

nary implication for Mesozoie I hcria. Canadian

Journal oftarth Sciences 12: 412-442.

— 1981. — Mammais from rhe Upper Cretaceous

Oldman Formation, Alberta. V. lioelelphis Matthew

and the evoliiiion of ihc Sragodonridae

(Marsupialiah Canadian Journal of lùirth Sciences

18 (2); 350-365.

— 1987. — Palaeuniologt' and che early évolution of

marsupials: 101-169. in Archer M. (cd.), Possums

and Opossums. Surrey Beatcy and Sons and the

Royal Society of New South Wales, Chipping

Norton. New South \X'’alcs.

Fox R. C. ôc Nuylot B. 1980. - A new species of

Didelphodon Marsh (MarsupiaJia) From rhe Upper

Cretaceous of Alberta, Canada: Palacobiology and

Phylogeny. Neut's Jcthrljuth jilf Cieologle tind

Paiaontfflo<^ie, Ahbandlungen 172: 357^360.

— 1995. — rhe relation.ships of the Sragodontidae,

Primitive Notth American F.ue Cretaceous mam-

mals: 247-2.50, w Sun A & Wang Y (eds), Sixth

symposium on Mesozoic ecosystem and hiotas. Short

paper. China Océan Press.

Gambar\'an Ik 1974. — ïioio mammais run: anatomi-

cal adaptations. Halstcd Press, Jérusalem, XI +

367 p.

Gayet M., Marsliall L. G. Sc Sempere F. 1992. —

The Meso/oic and Palaeocene vertebrates of

Bolivia and their stratigraphie context: a review.

Revista Jïcnica de Santa Cru?- 12 (3-4, die.

1991): 350-433.

Gidiey J. W. 1919- — .Significance of divergence of

tlie first digit in the primitive mammalian foot.

fournnl of the Washington Academy of Science 9

(10): 273-281.

Godinot M. ôt Prüsad G. V. R. 1994. — First

Creiaccous arborcal curhenans dcscovered in India.

Naîurwissenchaften S1: 79-81.

Goodrich E. S. 1958. -- Studies on the structure and

Development ofVerTehDUes. Dover, London, 837 p.

tîrand T. !. 19S3. — Body weighr: its rclatiuoshi() lo

chsue composition, si-^inenial distribution of mass

and nu»ior function. TII. The Didelphidae of

ITench Gin'iifM. Australian Journal of Zoology 31:

299-312.

Gregory W. K. 1910. — hhe orders of mammais.

Bulletin of the American Muséum ofNatural HtsUny

27: 1 524.

— TJ2(). - Studies in comparative myology and

osteolog)^; N IV. A review ol the évolution of lhe

lacn'^mal bone ol vertebrates wiih spécial references

to chat of mamtnals. Bulletin of the American

A/rtset//n ofNatucal Histofy 42 (2): 95-263.

Haq B. U. èc Van Fysihga F. W. B. 1994 —

Ceologiçal time lahle Fourth revised, enlarged and

updated editlom Elsevier S'cience. Amsterdam.

Hershlcovir? P. 1982. — l he scaggered marsupial

lower thiîd mci.tor (13). Géohios. Mémoire Spécial

6: 191-2(10.

Hiltlcbrand M. 1961. — Body proportions of didel-

phid (and soinc other) marsupials, with emphasis

on variahility. American Journal of Anatomy 109:

239-249.

Hopson h A- Sc Roiigier (i. W. 1993. ^ Braincase

structure in rhe olciest skull of a rherian mammal:

implications lor mamrnalian svsrcmaiic and cranial

évolution. American journal of Science 293:

26»-299.

Huxley 11. 1880. — On ilie application of ihe

laws of évolution vo the arrangement of the

Vertcbrai.i and more particularly co che Mammalla.

Proceedings of the Zoologtcal Society of London 1S80:

649-662.

Tnmam V. T., Sauiuler J. R,. F3e (... M &L Abotr L. C7

! 944. — (^bser. ations on rhe function of the shouL

dc! Jouival ofBone Joint Surgery 26: 1-30.

jaillard E.. Cappetta H., Ellenberger P., FcLst M.,

Grambasi-Fcs.surd N., Lcfranc J.-P. Ôc Sigé B.

1993. — Sedimencolog>% paleontology, biostrart-

graphy and eorelaiion of lhe l,;ue Cretaceous

Vilquethico Groufi of Southern Peru. Cretaceous

Reseatvb 14; 623-661.

Janson C. A. F.mons L. H. 1990. -- Ecologkal

structure of rhe nonflying mammal community at

Ciocha Ciashu biologital srarion, Manu National

Park. Peru; 314^338, in Gentry A. H. (ed.), For

Neotrapiail Rain Forest. Yale Lînivcrsicy Press, New

Haven.

GEODIVERSITAS • 1998 • 20(1)

131

Muizon C. de

Jenkins F. A. î971. — Limb posture and locomotion

in the Virginia opossum (D'tdelpbis mursupialis)

and in other non-cursorial mammaU. Journal of

Zoology^ London 165: 303-315,

— 1974, — Tree shrew locomotion and primate

arborcdÜsiTi: R5-I 15, in Jenkins F. A. (cd.),

Primatr f oromotion Academie Press, New York.

Jenkins I'. A. &: McLearn D. 19S4. — Mechanisms

of hind fbor reversai in climbing mammals. /oî/;7W

ofMorph(}log}> 182 (2): 197-219.

Jenkins I*. A. tk. Farrington F. R. 1976. — The posr-

craniui skeletons of tbc Tria.ssic niamrnals

Eozosirodon, Me^azosirodon and Erythrothenum.

Philosophwal îramavtiona of thr Royal Sovirty of

R 273: 387-431.

Jenkins f. A. Schaif CL R., 1988. - i'Ke Early

Cretacenns mamm.il Gohiconodon (Mammalia,

Triconodonra) from tbc Cloverly formation in

Montana. Journal af VirtehraU Pakontolo^ 8 (l):

1-24.

Jenkins F. A. & Wcijs W. A., 1979. — The functio-

nal anaromy of the shoulder in the Virginia opos¬

sum {Didelphis virginiand). journal of Zoolo§y

London 188: 379-410,

Julien-Laferricre D. 1991. — Organisuûon du peu¬

plement de inarsüpiauic en Guyane française. Revue

d'Écologie (‘ferre Viej-kb : 125-144.

Jollie M. 1962. — Chordare Morphology. Reinhold

Publisbing Corporation, New York.

Kappelman j. 1988. — Morphology and locomotor

adaptation of die bovid lemur in relation lo habi¬

tat. 198: 119-130,

Kermack K. A. Musseet F. & Rigney H. W. 1981. —

The skull of Aforganucodon. Zoologkal Journal of

the Linnedn Society of London 71: 1-158.

Kielan-Javvorow'ska Z. 1975. - Evolution of the the-

rian mammals in the Late Creraceous of Asia.

Part 1. Dcliathcridiiciae. Rcsults of the

Polisli-Mongolian P.-ikotuological Fxpediiions.

Part VI. ï^aliieontologm Polonka 33: 103-132.

— 1977. — Evolution of the therian mammals in the

Late Crctaceous ol Asia. Part II. Postcranial skele-

ton in KennaleUrs and éAiioryetts. Pulaeontologia

Polonka 37: 65-H3.

— 1978. - Evolution of the therian mammals in the

Late Crctaceous of .Asia. Part III. Postcranial skelc-

ton in ZaJambdaiestidae. Palaeonudogia Polonka

38: 3-41.

— 1981. — Evolution ofthe therian mammals in the

Late Crctaceous of Asia. Part IV. Skull structure of

Kennalestcs and Æv/oryere.v. Rcsults ol the

Polish-Mongolian Thdcoruological Expéditions.

Part IX. PaLieomologia Polonka 42: 25-78.

Kielan-Jaworowska Z. ôc Ciambaryan P. P. 1994. —

Postcranial anaromy and habits of Asian multitu-

berculatc mammals. Fossils and Strata 36: 1-92.

Kielan-Jaworowska Z. & Nessov L. A. 1990. — On

chc metatherian nature of the Deltathçroida, a .sis-

ter-group of the Marsupialia. Lethaia 23: 1-10.

Klelan-Jaworowska Z. k Trofimov B. A. 1980. —

Cranial morphology of the Crctaceous eutherian

BafuriUstes, Acta Palaeonwlogica Pohnica 25 (2):

167-185.

Rielan-jaworowska Z,, Bown 1. M. k. Lillegraven

j. A. 1979, — Eullicria: 221-258, in Lillegraven

J. A., Kiclan-J.twuri>Wwsk:i 7- 6C- Clemens W. A.

(ed.s), Afewzok AJanimaL: The First Two~TInrds of

Mammalian Mistory. Université' of California Press,

Berkeley.

Kiclan-jaworowsk.i Z.. Presley R. Poplin C.

1986. — l’hc cranial va.sciilar .system in lacniolabi-

doid niuitituberculace mammals. Philosophical

fransactioni ofthe Royal Society of London 313

(1164): 525-602,

Kielan-Jaw'orowska Z., Cromptoii A. W. &

Jenkins F. A. 1987. — The origin of egg-laying

mammals. Nature 326: 871-873.

Krause D. W. & Jenkins F. A. 1983. — The postcra¬

nial skeleton of North American mulutubertulafes.

Bulletin ofthe Muséum of Comparative Zoologyy

Harvard 150 (4): 199-246.

Krebhs B. 1991. — Das Skclcir von HenkeUaherium

guimarotae gcn. et sp. nov. (Eupantotheria,

Manunalia) aus dem Oberen Juta von Portugal.

Berli7}etgcowis^enschtiftlkhe Ahhandlintgoiy Berlin A

133, 110 p.

Larson S. C. 1993. — Jmnctional morphology of the

shoulder in primates: 45-69. in\ Gebo D. (ed.),

Posicranial adaptation tn non-human primates.

Northern Illinois Liniversity Press.

Larson S. G. & Stern J. T. 1989. — The rôle ot the

supra.spjnaïus in ihe quadrupcdal locomotion of

vervets [cereopitheeuz acihiops): implications for

inierpietation of humerai morphology. American

fournal of Physkal Anchropology 79: 369 -377.

Larson S, G. 6l .Stern J. F. 1992. — l'urther evidence

for rhe rôle of supraspjnarus in quadrupedal mon-

kcyn. American Jowiml of Physkal Anthropoluzy 87:

359-363.

Lillegraven j. A. Kxusat C. 1991. — Cranio-man-

dibular anatomy of Haldanodon euspectatus

(Docodonra, Mammalia) Irom rhe late Jurassic of

Portugal and tts implications to ihe évolution of

mammalian characters. Uninersity of Wyomlng

Cnniribiition to Geology 28: 39-138.

Luckett P, 1993. — An ontogenetic asseiîsment of

dental homologies in therian mammals: 182-204,

in Szalay F. S., Novaeek M. J. McKenna M. C.

(eds). Alnmmal Phylogeny. Mesozoic Dijferennationy

Multuiiberculatesy Early fherians arm Alarsupials.

SpringcT-Verl.tg, New York.

McKenna M. C. 1975. — Toward a phylngenetic

classification of the Mammalia: 21-46, in

Luclccti W. P. Ôl Szalay F. S', (eds)^ îHiylogeny ofthe

Primates. Plénum Press, New'York.

McManus J. J. 1970. — Behavior of captive opos¬

sum, LLidelphrs marsupialis virgirtianus. The

American Naturalist 144-169.

132

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of BoÜvia

Maclntyre G. T. 1967. — Foramen pseudovalc and

quasi mammals. Evolution 21: 834-841.

— 1972. — The trisulcate petrosal pattern of mam¬

mals; 275-303> in Dobzhansk>’ T.j Hecht M. K. &

Steere W. C. (eds), Evoludortary Biology Volume 6.

Appleton-Centurv'-Crofts, New York.

Marshall L. G. 1976. — Kvolution of the

Thylacosmilidae, extinta sabor-ïoorli mamipiais of

South America, Paleohiosiy. 1-31.

19773. — A new species of Lycopsis

(Borhy;ïenid;ie: Marsupialia) from ihe La Venta

fauna (laïc MIoccnc) of Colombia, South America.

Journal of Pahonlolo^ S ! : 633-642.

— 1977b. — Cladistic analysis of borhyaenoid,

dasyuroid, didtlphoid and tnyladnid (Marsupialia:

Mammalial aninity, Systemutic Zoology 26;

410-425.

— 1978. — Lvolurion of the Borhyaenidae, cxrincr

South Amerioan predaceous marsupiak. University

of California Pnbhcatiom in GroloCicnl Sciences 117:

1 -89.

— 1979a. — rC'view of the Prorhylacyninae, an

extinct subfamily ol South American *'dog-likc'’

marsupials. Fkldiatm Geolo^, nevv séries 3; 1-50.

— 1979b. —livoliition of metarherian and eutherian

(mammalian) characters: a review based on cladis¬

tic meihodolouy. Ztmiogical Journal of the Linnean

Society oj Lonnon 66: 369-410.

— 1981. — Review ot rhe Haihlyacyminae, an cxtinct

subfamily of South Arnerican \log-liice” marsu¬

pials. FieUluvia Cifolo^ n.s. 7: ) -120.

Marshall L. G. & Kiel ;an-Jawo!OWska Z. 1992. —

Rclulionships oi the dog-Jike marsupials» deliathe-

roidans and carly iribosphenic mammals. Lethaia

25: .561-374.

Marshall L. G. ôc Muizon C. de 1988. — ITe dawn

of the âge of mammals in South America. National

Géographie Research 4 (1): 23-55.

— 1995- — Fhe skull, in Muizon C. de (cd.),

Pucadelphvs andinus (Marsupialia, Mammalia)

from the early Palaeocene of Bolivia. Part II,

Mémoires du Muséum national d'Histoire naturelle

165 : 21-90.

Marshall L. G. &c Sigogneau-Russel! D. 1995. —

Posccranial skeleion, /// Muizon C. de (cd.),

Pucadelphys andinus (Afarsupialia, Mammalia)

from the carly Palaeocene of Bolivia, Part III,

Mémoires du Muséitm national d'Histoire naturelle

165:91-164.

Marshall L. G., Case J. A. & Woodburne M. O.

1990. — Phylogenetic relationships of the familles

of marsupials. Current Mammalogy 2\ 433-502.

Marshall L. G., Muizon C. de & Sigogneau-

Russell D. 1995. — The locaÜt)' ofTiupampa: age»

taphonomv and mammalian rauna,

Muizon C. de (ed.), Pucadelphys andinus

(Marsupialia, A4ammalia) from the early Palaeocene

of BoliviUy Part I, Mémoires du Muséum national

d'Histoire naturelle 165 : 11-20.

Marshall L. G., Muizon C. de. Gayet M.,

Lavenu A. & Sigé B. 1985. — The “Rosetta

Stonc” for mammalian évolution in South

America. National Geovraphic Research 1 (2):

274-288.

Marshall Lv G.. Sempere T. tSi Butler R. F. 1997 —

Chfonosrratigraphy of the mammal-bearing

Palaeocene ol Sriuih America. Journal of South

American FArth Sciences 10; 49-70.

Matihew W. D, 1904. —The arhoreal ancestry of

the Mammalia. Ameyican Naturalist 38; 811-818.

— 1916. — A marsupial from the Belly River

Cretaceous. Wich trirical obscr\Mtions upon the

affinitics of the Cretaccous mammals. Bulletin of

the American Muscurn of Natural History 35:

477-500.

Miao D 1988. — Skull morphology of Lambdopsalis

huila (Mammalia, Multituberculata) and its impli¬

cations (O manuïialian évolution. Contrilmtiom îo

Geolog)t^ Univeriity of Vi'yortiing, Spécial Paper 4:

M04.

— 1993- —(iranial morphology and multitubercula-

te relationships: 63-‘74, in Szalay F. S.,

Nüvacck M. J. & McKenna M. C. (eds), Mammai

Phylogeny. Volume I. Springer-Verlag, New York.

Muizon C de. 1992. La fauna de mamifcros de

Tiupampa (Paleoceno inferior, Formacion Santa

Lucla) Bolivia. Revisia técnica de Yactminevws

Pctndljeros Fiscales de Bolivia 1 2 (3-4. December

1991): 575-624.

— 1994. — A nc^v carnivorous marsupial from ihe

Palaeocene of Bolivia and the problcm of marsupial

monophyl)^ Nature 57Ù: 208-211.

Muizon C. de HsC Briio I. M. 1993. — Le bassin cal¬

caire de Sao Josc de Iraboraf (Rio de Janeiro.

Brésil): ses relations fauniques avec le .site de

Tiupampa (Cochabamba. Bolivie). Annales de

Paléontrmgie?*) (3) : 233-268.

Muizon C. de & Lange-Badré B. 1997 — Carni-

vorous dental adaptation in tribosphenid mammals

and phylogenetieal reconstruction. Lethaia 30 r

351-366.

Muizon C. de & Marshall L. G. 1992. —

Alcidedorbignya inopinata (Mammalia; Panrodonta)

from the Early Palaeocene of Bolivia; phylogenetic

and paleobiogcographic implications./owrwÆ/ of

PaUontolo^ 66 (3): 499-520.

Mui/on C. ot', Gaya M., Lavenu A.f Marshall L. G.,

Sige B. ^ VilJarrocI C. 1983. — Laïc Cretaceoas

vencbraie.s. incluJing mammals, from Tiupampa,

southcentral Bolivie. Geohios-t Lyon 16 (6) :

745-753.

Muizon C. de, Marshall L, G. 6c Sigé B. 1984. — The

mamnial launa from the El Molino Formation (Latc

Cretaceous-Maasrtichcian) at Tiupampa, southeen-

tral Bolivia. BidUiin du Muséum natiomd d'Histoire

naturelle, Paris, série 4, C 6 (4) ; 315-327.

Muizon C. de, Cifelli R. L. Cespedes Paz R.

1997 — The origin of the dog-like borhyaenoid

GEODIVERSITAS • 1998 • 20(1)

133

Muizon C. de

niarsupiaJs of South America. Nature 486-489.

Novak R. M. &L l’aracliso J. L. 1983. — Manmiah oj

the World. Volume 1; 1-644. The johns Hopkin.s

Prcîk.s, Ihililnioi'c.

OsgooJ W. H. 1921. - A monographie sriidy of rhe

Ameiicjn marsupial Qie?wlestcs. hielcl Muséum of

Natural Histoij, Zottlogy «crics 14; 1-1 S6.

Owen R. 1868. — On the anatonty nf Vertehrates.

Vol. UI, Mammaks. l.ongmam.«, Cireen and Co.

London.

Parker W. K. 1886. — On tlic «{ructuie and develop¬

ment of rhe skull in ihe M.iinmalia. Phiiosophkal

Trarisaaions of ihe Royal Society oJ London ]76:

1-27S.

Pascual R. & Ortiz |iuregui7.ar l:^. 1990. — Lvolving

cliniares and mammal faun-js in Cenozoïc South

America, foimtal of Human h.volut'um 19: 23-60.

— 1992. — El cicio raunfsiico cochabambiiio

(Palcoccno temprano): «u incidenciu. en la hîstoria

biogeogràfica de Jos maniiferos sudamericanos.

Revîsta ttkrncu de YPFB. Santa Cru/ 12 (3-4,

Die. 1991): 559-574.

Patterson B. 1965 —The auditory région of ihe

borhyaeiiid marsupial Cladosictis, Breiàora 217:

1-9.

PettcT Ci, Hoffsieiicr R. 1983. Les Marsupiaux

du Dc.séadien (Dligoccne inR^ieur) de Salla

(Bolivie). Annales de Paléontologie 69 : 175-234,

Prasad G. V. R. N. t ujdrnot M. 1994. — Euiherian

rarsal botics from rhe Lâic Cretaceous of India.

Jounufl ûfPalf-rjnSitlogy 68 (^i): 802- 902.

Presley R. 1981. — Alisphenoid équivalents in pla-

centals, niarsuptals, monotremes and f'ossils.

Nature 668-670.

— 1979. — The primitive course of the internai

caroiid arief)’^ in mammals. Acta Anatomtca 103:

238-244.

Presley R. & Steel I*. L. D. 1976. — On ilic homolo-

gy of ilie alispheiioid, Jotitna! of Anatomy 121;

441-459.

Reig Ü. A.. Kir.seb J. A. W. ôi Marshall L. G.

1987. — Systcmacic relaiionships ol the living and

Mc.sozoic opossum-like marsupials (suborder

Didelpliimorphia). with coiTimcnts on the classifi¬

cation ol tnese and of tbe Cretaceous and

Palaeoccne New World and Luropean metalhe-

rians: l-89i in Archer M. (ed,), Poisums and

Opossums. R'iyal Zt>ological Society of New South

Wales. Sydney, Australia.

Riges E. 1934. — A new marsupial saber-tooth from

me Plioeene of Argcnrina and its relationsliips ro

other South American predacious mar.supiaU.

Tramactiom of the American Phtlosophical Society,

new\crics24; 1-32.

Roberts D. LP4. — Structure and function of the

primate scapula: 171-200, in Jenkius E. A. (ed.).

Primate Locomotion. Academie Prc.s.s. New York.

Rougicr C. W., Wiblu J. R. & Hopson J. 1992. —

Reconstruction of the cranial vessels in the Early

CTetaceuus mammal Vincelesus ncuquenianus:

implications (or the évolution of ilic niammalian

cranial vascular sy.sTem, journal of Vertehrate

Paleontolo^* 12 (2): 188-216.

.Sempere E.. Butler R. L. & Marshall L. G. 1997. —

Sirangraphy and w'hronology of ihe Upper

Crctaccous-l.owcr Palcogcne .strata in Bolivia and

norhiwesi Argcntina. Genlogical Societ^^ of America,

Hullelhi )tt9 (6): 709-727.

Sigogneau* Russell D. 1992. —Hypomylos phelizoni

nov. geii. nov. sp,, une étape précoce de ( évolution

de la molaire iribosphénique ((Tétacé basal du

Maroc). Géohios 25 : 389-593'.

Simpson C. CT 1970. — The Argyrolacidac, extinct

South American marsupials. Bullettn of the

Muséum oj Gompavalive 7.oolügy I 59: 1-86.

Sinclair W. J. 1906. — Mammalia of the Saura Cruz

heds: Mar.supialia. Reports of fhc Princeton

ihiiversity Pspedittosu fo Pasagpjùa A: 3.53-460

Mijper E- b 1946. — Comparative biologic-jiiatomicai

investigations on the vertébral column and spinal

rnusciilauirc of mammals. Verhandelingen der

Koninklîjlev Nederkindsc Akadcntte tutn Wetetischap-

pen. Apfiing NatuurkumJe. Amsrcrdain 42: 1 -I 28.

Sori;i M. F. 1989. — Noiopicrna; un nuevo oïden de

mamffcros ungulados cogenos de América de! Sur.

Parte 11. Notonychop^ pou>elli gen et sp. nov.

(NdtonyLhopidac* fam. no\.) de la I ormacidn Rio

l.oro (Paleoi-eno medio), Provinci.i de lucuman,

Argentin.i. 25 (3): 259-272.

Szalay E. S. 1984. —Arboreality: is it liomologous in

mct.ithonan and eurherian m unmaix? Fiwlutionary

18 (6): 215-258.

— 1994. — Evolntûmury btstory ofthe marsupiab and

an analysis of osteologtcal characiers. Cambridge

L^niversity Press, New York, 481 p.

S/alay E. S. 1 rofimov B. A 1996. — l'he mongo-

itan Late Crct.aceoas .Asinthenum and the early phy-

Itigeny and paleobiogeography of Mctatlieria.

journal of Vertehrate Paleontology 16 (3): 4'^4-509.

Tardieu C.. 198,5. — l.’anitulaiion du genou. Analyse

morphofonctionnelle chev les primates, application

aux hominidae fossiles. Cahiers de Palèoanihropo-

%/V, éditions du CNRS, Pâtis: 1-IOS.

l'roflniov R. A. lir S.tj|jy F. S. 1994. — New

Creiaceous marsupial from Mtiitgoli.a and the early

ratllation ot Metarbcria. Proceedinçr of the National

Academyof Sciences, USA 91; 1269-12573.

'l’yndale-Biscoe CL H. 1973. ^ Life of Marsupiab.

American Elsevier Publishing Co.. New York,

254 p.

Van der Klaauw C^. J. 1913. — The auditor)' huila in

.somc fo.ssil mammals. Bnllettn of the American

Muséum ofNatural History 62: 1-352.

Van Valen L. 1988. — Palicoccne dînosaurs or

Cûctaceous iingulates in South America.

j'.volutiontiry Monogiapbs 10; 1-79.

Wible J. R. 1990. — PetrusaU of Laie Cretaceous

marsupials from North America and a cladistic

134

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

analysis of the petrosal in therian mammals.

Journal ofVertebrate Paleontology 10 (2): 183-205.

Wible J. R. & Hopson J. A. 1993. — Basicranial évi¬

dence for early mammal phylogeny: 45-62, in

Szalay F. S., Novacek M. J. & McKenna M. C.

(eds), Mammal Phylogeny. Mesozoic Dijferentiationy

Multituberculates, Monotremes, Early Therians and

Marsupials. Springer-Verlag, New York.

Winge H. 1941. — The Interrelationships of the

Mammalian Généra: 1-418. Volume I. Translaced

from Danish by E. Deichmann & G. M. Allen.

C. A. Reitzel Forlag, Copenhagen.

Woodward M, F. 1893. — Contribution to the study

of mammalian dentition. Part I. On the develop¬

ment of the teeth of the Macropodidae. Proceeding

of the Zoological Society of London 1893: 45-473.

Wroe S. 1997. — A reexamination of proposed mor-

phology-based synapomorphies for the Families of

Dasyuromorphia (Marsupialia). I. Dasyuridae.

Journal of Mammalian Evolution 4: 19-52.

Submitted for publication on 25 January 1997;

accepted on 2 October 1997.

GEODIVERSITAS • 1998 • 20(1)

135

Muizon C. de

APPENDIX

Following are the measurements of Mayulestes ferox in millimeters. e, estimated; —, measurement not

available.

Anleroposterior length (rom the tip of the right

premaxilla to the lip of the right condyle 53.4

Anteroposterior length trom the tip of the left

premaxilla to the tip of the left condyle 52.5

Bizygomatic width 33.5

Width between the lacrimal foramina

at the anterior extremity ot the orbits 17.8

Width ot the rostrum at the anterior

opening of the infraorbital canal 13.7

Width of the rostrum between the supraorbital

humps (supraorbital pnocesses) 12.8

Minimum width of the interorbital bridge 11

Length ot the right temporal fossa 11.6

Length of the left temporal fossa 11.4

Maximum width ot the nasals 12.7

Width ol trie basicramum between

the external auditory meati 20.2

Length of the right looth row (PI -M4) 16.5

Length otthe left tooth row (P1-M4) 16.6

Length ot the palale in sagittal plane 26.5

Length of the lower cheek tooth row (p3-m4) 20

Internai height of the dentary below middie

of M3 7.3

Measurements of the skul): since the crushing of the skull is

essentially dorsoventral, Ihe actual horizontal distortion is proba-

bly not very important. Therefore, the following measurements

represent a reasonable approximation of the actual dimensions

of the skull.

Height of the dentary below talonid of ml 7.3

Height of the dentary below talonid of m2 7.1

Height of the dentary below talonid of m3 6.8

Height of the dentary below posterior

talonid of m4 6.27

Length of the symphysis 9.15

Measurements of the dentary.

Upper

Right

0.6

0.5

Upper

Left

0.7

0.5

0.6

0.4

Lower

Right

0.2

0.4

0.3

0.4

0.3

0.2

Measurements of the incisors.

136

GEODIVERSITAS • 1998 • 20 (1)

MayulesteSy a borhyaenoid from the Palaeocene of Bolivia

Pir

P2r

PSI

p1r

p3l

1.12

2.2

2.6

1.3

2.2

0.78

1.4

0.6

1.2

0.87

1.5

2.5

0.9

2.6

Measurements of the premolars.

M1r M2r M3r M4r m1r m2r m3r m4l

2.95

3.12

2.96

2.39

2.86

3.06

3.36

3.72

3.66

4.08

4.55

1.83

1.11

1.21

1.43

1.2

2.48

3.27

3.72

3.68

1,24

1.84

1.87

1.98

2.86

3.09

2.89

1.95

1.74

2.2

2.09

2.34

3.29

391e

4.29

3.39

1.89

1.91

1.87

2.03

1.31

1.42

1.41

0.83

1.19

1.92

1.9

2.07

1.46

1.48

1.02

3.32

3.54

2.1

1.98

1.85

1.27

1.61

1.48

0.99

1.08

1.19

1.6

2.03

2.72

3.55

3.5

1.78

1.96

1.44

1.47

1.73

0.91

0.89

1.34

1.08

Measurements of the molars: upper (A, B) and lower (C-E) teeth

of Mayulestes ferox showing orientations for measurements:

1-11, upper molar, 1-8, lower molar. Unarrowed thick line on A

is an axis of orientation.

GEODIVERSITAS • 1998 • 20(1}

137

Muizon C. de

Maximum length of the neural arch

Maximum width of the neural canal

Maximum width between the latéral

borders of the occipital facets

Maximum width of the intercentrum

Maximum height of the left occipital facet

Maximum width of the left occipital facet

Maximum height of the left axoidian face

Maximum width of the left axoidian facet

5.5

12.5

4.2

4.3

_ Measurements of the atlas (CVI)

Maximum length of the neural spine

Maximum ventral length

Maximum height of the axis at the level

of the posteroventral border of the centrum

Maximum width between latéral borders

of the atlotdian facets

Posterior width of the centrum

Posterlor height of the centrum

Width between the postzygapophyses

Ventral length of the dens

Height of the dens

Width of the dens

9.5

8.2

4.7

2.9

1.5

_ Measurements of the axis (CV2).

CV?3

CV?5

Length of the centrum

3.7

Anterior width of the centrum

4.5

Posterior width of the centrum

4.6

4.3

Anterior height of the centrum

Posterior height of the centrum

Width between prezygapophyses

Measurements of CV?3 and CV?5.

T?1

T?12

T?13

Length of the centrum

4.2

6.2

6.4

Anterior width of the centrum

4.5

4.6

4.8

Anterior height of the centrum

2.7

2.3

Posterior width of lhe centrum

4.2

4.7

Posterior height of the centrum

Width between transverse

2.4

2.5

—

processes

11e

Width between anapophyses

Width between

“

7.1

6.7

prezygapophyses

Width between

7.3

5.7

—

postzygapophyses

4.1

Height at level of the spine

7.2

Length of the spine at apex - 3 2.9

_ Measurements of thoracic vertebrae.

138

GEODIVERSITAS • 1998 • 20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Length of the centrum

Anterior width of

7.4

8.2

7.4

the centrum

Anterior height of

5.7

4.8

5.2

5.4

the centrum

Posterior width of

—

3.4

2.8

—

3.7

the centrum

Posterior height of

4.3

4.8

5.7

5.5

the centrum

Width between trans-

3.2

2.8

3.2

3.7

verse processes

Length between

6.6e

13.7

anapophyses

Width between

prezygapophyses

Width between

6.7

7.5

—

pQStzygapophyses

Height of vertebra

5.7

5.9

—

7.8

at level of spine

Length of the spine

13.5

at apex

3.4

2.3

Measurements of lumbar vertebrae.

C?1

C?3

C?8

C?9

Length of the centrum

Anterior width of

6.7

the centrum

Anterior height of

4.3

3.5

5.4

4.8

the centrum

Posterior width of

3.5

3.3

the centrum

Posterior height of

3.5

5.4

the centrum

Width between anterior

3.3

3.6

3.7

transverse processes

Width between posterior

7.5

transverse processes

Width between

7.3

prezygapophyses

Width between

4.8

postzygapophyses

2.5

2.6

Measurement of the caudal vertebrae.

Maximum length of the infraspinatus

fossa, parallel to the spine

25.08

Maximum anteroposterior length

parallel to the spine

16.97

Length of the glenoid caviîy

5.63

Width of the glenoid cavity

3.77

Maximum anteroposterior length

of the acromion

6.14

Proximodistal length ot the acromion

9.65

Maximum height of the spine

4.00

Measurements of the scapula

GEODIVERSITAS • 1998 • 20(1)

139

Muizon C. de

Length

Transverse wîdth of the head

Anteroposterior length of the head

Length of deltoid crest

Transverse width of proximal extremity

Maximum width of the distal articular

surface in anterior view

Width of the capitulum in anterior view

Height of the capitulum in anterior view

Width of the trochlea In anterior view

Height of the trochlea in anterior view

Height of the capitulum in distal view

Height of the trochlea in distal view

Depth of the trochlea in posterior view

Angle between anterior and posterior

edge of trochlea in distal view

32.54

4.78

4.89

17.9

9.91

6.48

3.36

1.97

1.66e

1.75

3.21

2.66

0.89

56°

Measurements of the left humérus.

Length 37.74

Length from apex of olecranon to coronoid

apophysis 10.5

Length of olecranon from apex to proximal

extremity of greater sigmoid cavity

in médial view 4.72

Maximum anterior length of the olecranon 12

Proximodistai length of greater sigmoid

cavity in médial view 6.2

Width of proximal edge of the sigmoid cavity 3.53

Width of olecranon at apex 3.1

Width at the level of the coronoid apophysis 4.68

Anteroposterior length at proximal edge

of sigmoid cavity 3.7

Length of médial branch of proximal

edge of greater stgmoid cavity 2.58

Length of latéral branch of proximal

edge of greater sigmoid cavity 1.48

Angle between olecranon and shaft

in anterior view 149°

Angle between olecranon and shaft

in latéral view 166°

Measurements of the ulna.

Length

29.46

Width of proximal epiphysis

3.96

Length of proximal epiphysis

Length between proximal end and

distal border of bicipital tuberosity

5.02

Width of distal epiphysis

4.35

Length of distal epiphysis

2.74

Measurements of radius.

140

GEODIVERSITAS

1998

20(1)

Mayulestes, a borhyaenoid from the Palaeocene of Bolivia

Length

Width of diaphysis at mid-length

0.9

Anteroposterior length of diaphysis

at mid-length

1.8

Width of distal epiphysis

2.6

Anteroposterior length of distal epiphysis

1.4

Length

Length of the ilium from anterior extremity

to centre of acelabulum

Dorsoventral breadth of iliac wing

Anteroposterior diameter of acetabulum

Maximum dorsoventral diameter of

acetabulum

Length from posterior border of sciatic

spine to posterior border of ischium

42.57

25.2

7.17

6.6

6.1

11.95

Length

40.77

Proximal transverse width

10.23

Length from tip of greater trochanter

to ventral border of head

4.13

Length from tip of greater trochanter

to distal end of lesser trochanter

10.36

Mediolateral length of the head

in posterior view

6.08

Anteroposterior length of the head

in proximal view

3.84

Length from tip of greater trochanter

to distal end of trochanteric fossa

5.48

Transverse diameter at mid-length

of the shaft

3.66

Anteroposterior diameter at mid-length

of the shaft

2.76

Transverse width of distal extremity

7.54

Anteroposterior length of latéral distal

condyle

6.06

Width of left latéral distal condyle

in posterior view

4.2

Width of right latéral distal condyle

in posterior view

3.6

Width of right médial distal condyle

in posterior view

2.7

Height of right latéral condyle

in posterior view

2.35

Width of left trochlear groove

4.2

Width of right trochlear groove

3.5

Latéral height of trochlear groove

in anterior view

4.4

Médial height of trochlear groove

in anterior view

3.56

Latéral length of trochlear groove

in distal view

3.77

Médial length of trochlear groove

in distal view

3.11

Measurements of the metacarpals (McV).

Measurements of the innominate.

Measurements of the fémur.

GEODIVERSITAS • 1998 • 20(1)

141

Muizon C. de

Length

40.68

Proximal transverse width

7.5

Proximal anteroposterior length

5.2

Width of the latéral proximal condyle

3.18

Width of the médial proximal condyle

2.59

Anteroposterior length of the latéral

proximal condyle

3.17

Anteroposterior length of the médial

proximal condyle

3.55

Distal transverse width

4.1

Distal anteroposterior length

3.78

Proximodistal length of malleolus

1.84

Width of the malleolus

2.18

Width of astragalar facet from base

of malleolus to latéral edge

1.95

Angle between the flexion axis

of the knee (cagreatest width

of the proximal epiphysis)

and the plane of the malleolus

48°

Distal transverse width 4.5

Distal anteroposterior length 3.8

Length

9.5

Width of the tuber at mid-length

1.8

Height of the tuber at mid-length

3.2

Proximodistal length of ectal facet

3.3

Transverse width of ectal facet

1.8

Maximum distal width

5.5

Width of cuboid facet

3.3

Dorsoplantar length of cuboid facet

2.7

Mt?lll

MtIV

Length

13.4

12.1

Proximal width

2.2

Distal width

2.8

2.7

Measurements of the tibia.

Measurements of the fibula.

Measurements of the caicaneum.

Measurements of metatarsals.

142

GEODIVERSITAS • 1998 • 20(1)

Instructions aux auteurs

La ligne éditoriale

Elle prendra en coinpre divers aspects de la

recherche en Sciences de la Terre, en particulier

l'histoire et le comporrement des bassins sédimen-

taires, la paléobiodiversité et les palcoenvironne-

ments. Un junnéro de Gaodhfcrsitns par an pourra

erre consacre, exceptionnellement, au débat contra¬

dictoire sur un sujet d actualité, ou sur un thème

donné et sous la responsabilité d’édileur(s) invilé(s).

Les manuscrits, dont le nombre de pages irest pas

limité a priori^ devront suivre rigoureusement les

recommandations aux auteurs (voir ci-dessous) et

seront adressés à la revue

Service des Publications Scientifiques du Muséum,

Gendiversitas^

57 rue Cuvier,

F'75231 Paris Cedex 05

Tel: (33) 01 40 79 34 3«

Fax: (33) 01 40 79 38 58

e. mail : bulletin@mnlm.fr

Les chapitres de systéniaiiquc devront se conformer

aux règles du Code Internatiomtl de Nomenclature

Xoologûfue.

l our manuscrit non conforme à ces instructions

sera retourné pour mise au point. Chaque manus¬

crit est évalué par deux rapporteurs, ou plus.

Instructions atéx auteurs

Chaque manuscrit soumis (y compris les illustra¬

tions) doit cire présemé en trois exemplaires (un

original et deux copies) au format A4, avec un

double inrerligne et des marges d’au moins 3 cm ;

chaque page sera numérotée. Les illustrations origi¬

nales seront jointes au manuscrit definitif, ainsi

qu’une disquette 3.5" de format Apple Macintosh

ou compatible IBM (traitement de texte Word de

préférence), qui devra contenir également les

tableaux et éveiiUiellenicni les illustrations (Adobe

lllustraior, Photoshop ; L3eneba Canvas).

Le fonnat

Les manuscrits, écrits en français ou en anglais, doi¬

vent être structurés comme suit :

— titre si possible bref, en français et en anglais ; un

titre courant doit être proposé ;

- nom(s) et prénüin(s) de(s) auleur(s) suivis de

leur(s) udresse(.s) profcsMOiincllc(s). en indiquant si

possible le ntunéro de Fax et l’adresse électronique ;

- résumés écrits en français et en anglais (800

signes au maximum chacun), .suivis de* mots clés et

« key wt>rds » ;

- dans le texte courant. Utiliser les italiques pour

tous les noms en latin (taxons de rangs générique et

spécifique, et a /.;

- dans le texte coûtant, les références aux auteurs

seront en minuscules, ex. Dupont (2001), Dupont

(2001, 2002;, (Dupont 2001 ; Durand 2002),

Dupont (2001 : 1 ), Dupont (2001, fig. 2).

- dans le texte courant, les références aux illustra¬

tions et aux tableaux de rariiclc seront présentées

ainsi : (Fig. 1 ), (Fig. 2A, D). (Figs 3, 6), (Figs 3-5),

(Tableau 1) ;

- les remcrtienietus seront placés a la fin du texte,

avant les références bibliographiques ;

- les références bibliographiques doivent suivre les

exemples donnes ci-dessous ;

- indiquer dans la marge l’emplacement de.s illus¬

trations dans le texte définitif ;

- donner les légendes des figure.s sur une feuille

séparée.

Les illusttations

Une attention particulière sera portée à la qualité et

la pertinence de l'illustration.

Les illustrations au trait doivent être réalisées à

l’encre de Chine ou erre fournies en impression

laser, Les photographies, bien contrastées, seront

■sur fond noir ou blanc. Biles pourront être regrou¬

pées, et dans ce cas, identifiées par une lettre en

capitales (A, B, C...). Les planches photogra¬

phiques, placées dans le corps de Farticlc et non

regroupéc.s à la fin de celui-ci, doivent être traitées et

numcrotéc.s comme des figures. I.cs illusrrations

pourront être a.ssentblécs sur une largeur de colonne

(70 X 190 mm) ou sur toute la largeur de la justifî-

câiion (144 x 190 mm). La rédaction encourage la

présentation de photographies avec tout ou partie

de leur interprétation par un ou des dessins au trait.

Aucune légende, ni lettrage ne sera placé sur les ori¬

ginaux. Ils figureront sur un calque joint avec

chaque figure, la rédaction se chargeant de les pla¬

cer. Chaque figure doit comporter une échelle

métrique. Les tableaux et graphiques, à inclure dans

GEODIVERSITAS • 1998 • 20(1)

143

Instructions aux auteurs

le manascrir, doivent nécessairement pouvoir être

imprimés sur une page et rester lisibles apres réduc-

tion éventuelle. Des planches en couleur pourront

être publiées moyennani une participation finan¬

cière de ou des auteurs.

Références bibliographiques

Denison R. H. 1978. — Placodermi, in Schultze

H. P. (cd.), Hnndbook of Paleoichthyology,

Volume 2. Gustav Fischer, Stuttgart, 128 p.

Marshall C. R. 1987. — l.ungfish: phylogeny and

parsinïony, in Bemis W. F.., Burggren W. W.

& Kemp N. 1*7 (eds), The Biology and

Evolution t)f Lungfishes, lourtiat of Morphology

1:151-162.

Schultze H. P. & Aiscnauh M. 1985. — The pan-

derichthyid fish Elpistostege: a close relative to

tetrapods? Paleontnlogy 1^: 293-.309.

Schulae 11. P, 1977a. — The origin of the letra-

pod limb within rhe rhipidistian fishes:

541-544, in Hecht M. K., Goody P. C. bc

Hechi B. C. (eds), Major Patrerns in Vertebrate

Evolution. Plénum Press, New York and

London.

Épreuves et tirés à part

Les épreuves seront adressées à Fauteur ou au pre¬

mier auteur (sauf indication contraire) et devront

être rerournées corrigées sous huitaine. Les correc¬

tions, autres que celles imputables à la rédaction ou

à l’imprimeur, seront à la charge des auteurs. l.e(s)

auteur(s) recevront gracieusement vingt-cinq tirés à

part, les tirés â part supplémentaires seront à com¬

mander sur un formulaire joint aux épreuves.

Soumettre un article pour publication dans

Geodiversîtas suppose que celui-ci ou tout article

proche dans k même langue ou une autre langue,

n’ait pas été soumis dans une autre revue, meme

dans l’attente de son acceptation. Les droits de

reproduction de l'article, y compri.s des illustra¬

tions, sont réservés à la revue. La reproduction de

tout ou partie de Faniclc doit faire l’objet d’une

demande écrite préalable adressée à la rédaction.

Scope of the Journal

Geodiversîtas publishes papers which concern varied

a.spects of Earth .Sciences and particularly hi.stoiy of

sedimentary basins, palaeobiodiversicy and paleoen-

vironmenr. A complété issue of Geodiversitiu may

be devored to scveral papers on a single lopic under

the responsahility of guesr editor(s). Papers with a

systcmaric conieni should follow the International

Code of Zaologicnl Nomenclature. Manuscripts,

withoLit limitation ot the nuniber of pages, must

conform strictly with the instructions to authors,

and will be sent to rhe Editer:

Service des Publications Scientifiques du Muséum,

Geodivei'sitas,

57 rue Cuvier,

H-75231 Paris Cedex 05

Tel : (33) 01 40 79 34 38

Fax: (33) 01 40 79 38 58

e. mail : bullctin@mnhn.fr

Instructions to authors

Manuscripts, with illustrations» must be submitted

in triplicatc (onc original and iwo copies) in A4

formai, double spaccd, with margin.s' of at least

3 cm and ail pages numbered. The original figures

should be senr with the revised manusaipt, as.well

as a 3.5” diskette Apple Macintosh or IfiM-compa-

tible (Word, Word Perfcct...) format, which will

also contain tables and possibly figures (Adobe

lllustrator, Photoshop; Deneba Canvas).

Format

Papers are lo be written in simple and concise

Frcnch or English. Thcy should be organized as fol-

lows:

— a brief tille in Frcnch and English;

— a suggested running head;

— namc(s) of aiirhor(s), followed by their full pro-

fessional addrcss(cs) and, if possible, Fax number

and e-mail;

— abstracts (in English and French) not exceeding

800 signs each, wiih key words and “mots clés”;

— text with italicized words for Latin (taxa of gene-

ric and spécifie ranks, et aiy ...);

— référencés to authors in main text .should he pre-

sented, in lower case, as follows: Smith (2001),

Smith (2001, 2002), (Smith 2001), (Smith 2001;

Cary 2002), Smiih (2001: l), Smith (2001, fig. 2);

— references to illustrations .and tables should be

indicated .as follow.s: (Fig. 1), (Fig. 2A, D). (Figs 3,

6), (Figs 3-5), (Table l );

— keep acknowlcdgcmcnts short and place them at

the end of the text before référencés; please do not

144

GEODIVERSITAS • 1998 • 20(1)

Instructions aux auteurs

forget the rcvisers;

- give captions to illustrations on a separate sheet.

Illustrations

The éditorial board will pay spécial attention to the

quality and relevance of illustration.

Line drawings must be in Indian ink or high quali¬

ty laser printouts; high contrast photographs, pla-

ccd on whice or black backgrounds, are required.

These can be grouped into figures and identified

by letters A, B, C ... Plates are noi placed at the

end of the article: they will be considered as figures

and numbered as such. Arrange figures to fit one or

two columns (70 x 190 mm or 144 x 190 mm).

Associate interprétation of photograph with line

drawing. No diagram or table is to exceed one

page. Letters, numbers, etc., for cach figure, are to

be indicated on an accompanying overlay, not on

the original figure (line eut or half-tone). A scale

bar is needed for each figure.

Référencés

Denison R. H. 1978. — Placodermi, in Schultze

H. P. (ed.), Handbook of Paleoichthyologyy

Volume 2. Gustav Fischer, Stuttgart, 128 p.

Marshall C. R. 1987. — Lungfish: phylogeny and

parsimony, in Demis W. E., Burggren W. W.

& Kemp N. E. (eds), The Biology and

Evolution of Lungfishes, Journal of Morphology

1: 151-162.

Schultze H. P. & Arsenault M. 1985. — The pan-

derichthyid fish Elpistostegei a close relative to

tetrapods? Pakontology 28: 293-309.

Schultze H. P. 1977a. — The origin of the tetra-

pod limb within the rhipidistian fishes:

541-544, in Hecht M. K., Goody P. C. &

Hecht B. C. (cds), Major Patterns in Vertebrate

Evolution. Plénum Press, New York and London.

Proofs and reprints

Proofs will be sent to the fîrst author for correction

and must be returned within eight days by express

mail. Authors will receive twenty-five oflfprints free

of charge; further offprinrs can be ordered on a

form supplied with the proofs.

The submission of a manuscript to Geodiversitas

implies that the paper, or a similar one, is not being

offered for publication eisewhere. Copyright of

published paper, including the illustrations,

becomes the property of the journal. Requests to

reproduce material from Geodiversitas should be

addressed to the editor.

GEODIVERSITAS • 1998 • 20(1)

145