International Society for the Study

and Conservation of Amphibians

(International Society of Batrachology)

SEAT

Laboratoire des Reptiles et Amphibiens, Muséum national d'Histoire naturelle, 25 rue Cuvier, 75005

Paris, France. — Tel.: (33).(1).40.79.34.87. — Fax: (33).(1).40.79.34.88. — E-mail: ohler@mnhn.fr.

BOARD FOR 1996

President: Raymond F. LAURENT (Tucumän, Argentina).

General Secretary: Annemarie OHLER (Paris, France).

Treasurer: Dominique PAYEN (Paris, France).

Deputy Secretaries: Britta GRILLITSCH (Wien, Austria); W. Ronald HEYER (Washington, U.S.A.);

Stephen RICHARDS (Townsville, Australia).

Deputy Treasurers: Jean-Louis DENIAUD (Paris, France); Mark L. WyGopa (Lake Charles, U.S.A.).

Councillors: Janalee P. CALDWELL (Norman, U.S.A.); Marissa FABREZI (Salta, Argentina); Jon LOMAN

(Lund, Sweden).

TARIFF FOR 1996

Individuals Institutions

Subscription to Alytes alone 280 FF / 56 $ (regular) 560 FF /1125

140 FF / 28 $ (student)

Subscription to Alytes + ISSCA + Circalytes 300 FF / 60 $ (regular) 600 FF / 120$

150 FF / 30 $ (student)

Back issues of Alytes: single issue 70 FF/14$ 140 FF /28$

Back issues of Alytes: one complete volume (4 issues) 225 FF / 45 $ 450 FF / 90 $

Back issues of Alytes: complete set volumes 1 to 13 2340 FF / 468 $ 4680 FF / 936$

Five-years (1996-2000) individual subscription to Alytes: 1120 FF / 224 $.

Five-years (1996-2000) individual subscription to Alytes + ISSCA + Circalytes: 1500 FF / 300 $.

Life individual subscription to Alytes from 1996 on: 5600 FF / 1120 $.

Life individual subscription to Alytes + ISSCA + Circalytes from 1996 on: 7000 FF / 1400 $.

Patron individual subscription to Alytes from 1996 on: 11200 FF / 2240 $ or more.

Patron individual subscription to Alytes + ISSCA + Circalytes from 1996 on: 14000 FF / 2800 $ or

more.

Important notice: from 1996 on, any new life or patron individual subscriber to Alytes will be offered a

free complete collection of back issues of Alpes from the first issue (February 1982) until the start of

her/his subscription.

Circalytes is the internal information bulletin of ISSCA. Back issues of this bulletin are also available:

prices can be provided upon request by our Secretariat.

Inclusive Section or Group affiliation to ISSCA: 250 FF / 50 $.

Individual subscription to the ISSCA Board Circular Letters: 200 FF / 40 $.

MODES OF PAYMENT

In French Francs, by cheques drawn on a French bank payable to “ISSCA”, sent to our Secretariat

(address above).

In French Francs, by direct postal transfer to our postal account: “ISSCA”, Nr. 1-398-91 L, Paris;

if you use this mode of payment, add 15 FF to your payment for postal charges at our end.

In US. Dollars, by cheques payable to “ISSCA”, sent to Mark L. WyGopa, Department of

Biological and Environmental Sciences, P.O. Box 92000, MeNeese State University, Lake Charles,

Louisiana 70609-2000, U.S.A.

Source : MNHN, Paris

AITES

INTERNATIONAL JOURNAL OF BATRACHOLOGY

March 1996 Volume 13, N° 4

Alytes, 1996, 13 (4): 109-139. 109

Comparative morphology of phytotelmonous

and pond-dwelling larvae

of four neotropical treefrog species

(Anura, Hylidae, Osteocephalus oophagus,

Osteocephalus taurinus,

Phrynohvyas resinifictrix, Phrynohyas venulosa)

Luis C. ScHiesaRi *, Britta GRILLITSCH ** & Claus VOGL **

* Departamento de Zoologia, Instituto de Biociéncias, Universidade de Säo Paulo, CP 11294,

05422-970, Säo Paulo, SP, Brasil

Icschies@usp.br

** Institute of Laboratory Animal Science, University of Veterinary Medicine of Vienna,

Linke Bahngasse 11, 1030, Vienna, Austria

brittagr@ping.at

We dedicate this paper to the memory of Rudolf WyTek (f August 25, 1995, Vienna)

External and buccopharyngeal morphology of phytotelmonous and pond-

dwelling larvae of four neotropical hylid species were analyzed with respect to

differential diagnosis and ecomorphology. Larvae typically live in bromeliads

(Osteocephalus oophagus), tree holes (Phrynohyas resinifictrix), ponds in

rainforest (Osteocephalus taurinus), and ponds in open areas (Phrynohyas

venulosa). Discriminant analysis of morphometric external characters revealed

only slight differences between the phytotelmonous species, but two well

separated subgroups among the pond-dwelling species. AIl species showed

average body proportions, macrophagous habits, and pulmonary as well as

branchial respiration. Tadpoles living in phytotelms were characterized by

reduction of peribuccal and buccopharyngeal structures, and differed notably

in jaw sheath morphology and lung development among species. Tadpoles

developing in ponds were characterized by high numbers of tooth rows and low

balance values. The genera Osteocephalus and Phrynohyas were distinguished

best by development of secondary upper tooth rows, position of the eyes, and

gross body proportions. Bibliothèque Centrale Muséum

(LI

3 3001 0011

110 ALYTES 13 (4)

INTRODUCTION

In this study we describe the tadpoles of Osteocephalus oophagus Jungfer & Schiesari,

1995, Osteocephalus taurinus Steindachner, 1862, Phrynohyas resinifictrix (Goeldi, 1907),

and Phrynohyas venulosa (Laurenti, 1768). Larval gross body characters and external

buccal features are examined through series comprising almost the entire larval period.

Internal buccopharyngeal surface features and detailed morphometry are considered in

advanced tadpole stages only.

Tadpoles of the four species have already been described by a number of authors

(Table I). However, with the exception of P. venulosa, those descriptions were based on

small samples and provided little information on ontogenetic and intraspecific variation.

Information on buccopharyngeal larval morphology was available only for P. resinifictrix.

The neotropical hylid genera Osteocephalus Steindachner, 1862 (seven species;

JUNGFER & SCHIESARI, 1995) and Phrynohyas Fitzinger, 1843 (five species; FROST, 1985) are

well defined by their adult morphology (TRUEB, 1970; TRUEB & DUELLMAN, 1971). Within

both genera, larval habitats vary significantly among species and include phytotelms,

ponds, and streams (Table VIII). The larval habitats for the four species in this study are

principally of two types: phytotelms and ponds. Larvae of O. oophagus typically develop

in rainforest bromeliads at the ground or off-ground up to 2 m high; those of P.

resinifictrix dwell in spacious tree holes in the canopy up to 35 m high. In contrast, larvae

of O. taurinus typically develop in rainforest ponds and those of P. venulosa in ponds of

open areas.

To better understand larval morphology of the species analyzed in depth in this study,

we finally include comparative data from the literature on other Osteocephalus and

Phrynohyas species (Table VIIL, fig. 7) and refer to phenotype-ecotype interdependencies

recognized for anuran larvae (WAssERSUG, 1980; LANNOO et al., 1987; WASSERSUG &

HEYER, 1988; ALTIG & JOHNSTON, 1989).

MATERIAL AND METHODS

Collection data, as well as the number and range of developmental stages of the

tadpoles examined, are summarized in Table Il. Tadpoles were preserved in 4%

formaldehyde solution. Most of the series were preserved in the field shortly after

collection. Some series of P. resinifictrix and O. taurinus Were reared in the laboratory

from spawn or tadpoles collected in the field. Specific assignment of the tadpoles collected

by L. C. SCHIESARI was confirmed by specimens raised in the laboratory either from

identified spawn or until neometamorphosed stages (exceeding stage 46, GOsNER, 1960),

and was confirmed by congruency with the museum tadpole series indicated in Table II.

P. venulosa larvae were identified by J. P. CALDWELL.

Determination of larval developmental stages follows GosnER (1960). Morphometric

parameters (Table III) are defined as in GRiLLITsCH et al. (1993). Within stages, the

Source : MNHN, Paris

Table I. - Literature survey on larval morphology of Osteocephalus oophagus, Osteocephalus taurinus, Phrynohyas resinifictrix and Phrynohyas venulosa.

Information present (+) or not present (-).

Features

Species | Stages Localities General Oral Buccopharyngeal References

appearance disk cavity Mor-

O. oophagus | 37 | Manaus, Amazonas, Brasil - + - ; = - - HERO (1990) !

19-40 | Manaus, Amazonas, Brasil + + + - - - + | JUNGFER & ScHiEsari (1995)|

O. taurinus ? | 40-41 | Manaus, Amazonas, Brasil - + - + - - - HERO (1990)

P. resinifictrix - + - + - - - - LANNOO et al. (1987)

39 | Manaus, Amazonas, Brasil - + - + - - - HERO (1990)

28-38 Panguana, Perû + + + + + + + | GriLLrrsox (1992)

P. venulosa | 18-41 | Bejuco, Panamä, Panamä + + + + : - + | ZweFeL (1964)

17-46 | Encinal, Veracruz, México | + + + + - : + | PyBuRN (1967) à

38 | Encinal, Veracruz, México | + + + + - - - DUELLMAN (1970)

38 Leticia, Colombia + - + # - - DUELLMAN (1978)

Rio Negro, Edo. Miranda, + + + + - - - RADA DE MARTINEZ (1990)

Venezuela

1. Tadpoles described under the name Osteocephalus sp.

2. Déscription of O. saurinus tadpoles in DUELLMAN & LESCURE (1973) must be referred 10 Hyla geographica (CALDWELL, 1989);

is also applies to the ©. raurinus tadpole described

in DUELLMAN (1978). Original description of Hyla elkejungingerae (HENLE, 1981) includes larval morphology. This species has been considered a possible synonym of O.

taurinus (FROST, 1985), but was redefined by HENLE (1992) as Osteocephalus elkejungingerae and is considered as a distinct species in the present study.

3. Tadpoles described under the name Phrynohyas spilomma..

4. Reexamination of the series of ZWEIFEL (1964) and PYBURN (1967), and comparison with a further series from Tepic, Nayarit (México), support conspecificity of the tadpoles of

ZWEIFEL and PYBURN.

190À ® HOSLITINS) ‘VSHIHOS

Source : MNHN, Paris

112 ALYTES 13 (4)

Table II. - Material investigated: Museu Nacional do Rio de Janeiro, Brasil (MNRIJ);, Museu de

Zoologia da Universidade de Säo Paulo, Brasil (MZUSP); L. C. Scmsart field numbers

of material deposited in the MZUSP (LCS).

Localities: Reserva Florestal Adolfo Ducke, Manaus, Amazonas, Brasil (A);

Reservas INPA No. 3402, 1501, 1401, Instituto Nacional de Pesquisa do Amazonas -

World Wildlife Fund, Amazonas, Brasil (B); Juruä, Rio Xingü, Par, Brasil (C); Boa

Vista, Roraima, Brasil (D).

Collectors: see notes.

Morphometry: number of specimens and range of stages of the specimens

investigated for gross body morphometry (Table III), detailed morphometry (Table 111)

and tooth row counts (Table IV).

Gross Detailed Tooth row

Specimen | Localities | Dates of morphometry morphometry counts

©. oophagus A‘

LCS 345 26.01.93 7 27-37 7 37-39 8 27-37

LCS 361"? 09.03.93 20 31-40 1 37 17 31-40

©. taurinus

MNRI 7971 !? A 26.10.85 6 35-39 à 37-39 6 35-39

MZUSP 66336 B° 01.-05.93 3 36-39 2 39 3 36-39

LCS 343 ? A‘ 27.01.93 33 25-26 - - 34 25-26

LCS 364 ? A‘ 12.03.93 17 27-28 - - 16 27-28

P. resinifictrix AS

LCS 342 ? Tree 1 |21.01.-11.03.93] 6 30-36 - - 5 30-36

LCS 355 Tree 1 18.02.93 9 25-26 s : 4 25-26

LCS 362 Tree 1 10.03.93 2 25 - 2 25

LCS 372-375! | Treel |13.02-12.06.93| 30 | 25-40 3 37 19 | 2641

LCS 347 Tree 2 28.01.93 8 26-35 - - 7 26-35

LCS 356 Tree 2 18.02.93 6 27-31 - - 4 27-30

LCS 376377? | Tree2 |29.01.-17.04.91| 17 25-39 3 37-39 16 | 25-39

LCS 378 ! Tree 3 29.01.91 14 | 25-37 1 37 12 26-37

LCS 366 Tree 4 18.03.93 2 25 - : 2 25

LCS 379 Tree 5 30.01.91 5 25 - : 4 25

LCS 357 ? Tree 6_|24.02-09.03.93| 8 27-37 2 37 8 27-37

P. venulosa ce

MZUSP 64359 15.01.87 5 30-34 . ë 5 20-34

SP 64360 ! 1987 2 35-37 Û 37 2 35-37

27.01.87 4 37-40 3 3738

05.02.87 1 39 2 39 3 :

14.02.87 2 25 - - 3 25

D’ 23.06.91 2 36-37 1 37 2 36-37

C. SCHIESAR

HERO,

Coll. €. GASCON.

Coll. J. P. CALDWEIL.

;

Source : MNHN, Paris

Table III. - Comparison of morphometric parameters describing the body proportions of the tadpoles of Osreocephalus oophagus, Osteocephalus

taurinus, Phrynohyas resinifictrix and Phrynohyas venulosa, for developmental stages pooled within two different ranges.

Parameters (distances as defined in GRILLITSCH et al., 1993): maximum diameter of eye (ED); maximum height of tail

(HT); maximum height of lower tail fin (LF); internarial distance (NN); naro-pupillar distance (NP); maximum width of oral disk

(OD), interpupillar distance (PP); rostro-narial distance (RN); distance tip of snout - opening of spiracle (SS); distance tip of snout -

insertion of upper tail fin (SU); distance snout - vent, snout-vent length (SV); maximum height of upper tail fin (UF); distance vent -

opening of spiracle (VS); distance vent - tip of tail, tail length (VT).

For each parameter, the following information is provided: mean value + standard deviation; range in parenthesis; number

of specimens in brackets; coefficient of regression of ratios with total length / Spearman correlation coefficient of ratios with stage in

waved brackets, * P < 0.05, ** P < 0.01.

LE mere

Parameters Stages pooled O. oophagus ©. taurinus P. resinifictrix P. venulosa

Vent - ail tip / VT/SV] 26-40 [1200.13 (0.82-1.47) (271/1.36 + 0.08 (1.22-1.52) (38]|1.47+0.11(1.25-1.72) (87]| 1.38 + 0.08 (1.26-1.50) [14]

snout - vent {0.0000/-0.25} {0.0017/0.01} -0.0030/0.12} 40.0061/0.41}

36-40 [1.17 + 0.14 (0.82-1.41) 112]] 1.38 + 0.09 (1.22-1.49) 1 7]/1.47 + 0.12(1.25-1.70) [18]] 1.41 # 0.08 (1.26-1.50) [8]

Vent - ail üip / VT/HT| 26-40 |2.42+0.26(1.99.3.07) [231] 3.26 + 0.29 (2.77-4.23) [30] 2.53 + 0.26 (2.06-3.05) [471] 2.50 + 0.18 (2.14-2.75) [14]

tail height 40.0216/0.05} {0.0075/-0.01} {0.0070/0.16} 4-0.0111/-0.32}

36-40 [2.47 + 0.31 (2.04-3.07) [10)] 3.60 + 0.34 (3.20-4.23) [51/2.55 + 0.12(2.34-2.83) [12]] 2.44 + 0.18 (2.14-2.69) [8]

Tail height / HT/UF 26-40 3.15+ 0.15 (2.90-3.41) [24] | 3.44 + 0.28 (3.05-4.21) [30]| 3.10 + 0.29 (2.71-3.71) [50]| 2.84 + 0.11 (2.63-3.03) [16]

upper fin height 40.0025/0.26} 40.0074/-0.31 +} 40.0329 **/ 0.15} {0.0023/-0.02}

36-40 [3.19 + 0.14 (2.94-3.41) [10]] 3.63 + 0.39 (3.06-4.21) [51/3.07 + 0.25 (2.74-3.56) [13)] 2.82 + 0.10 (2.63-2.94) 19]

Upper finheigt/ [UF/LF| 26-40 |1.070.07(0.92-1.19) [24]| 1.03 + 0.09 (0.79-1.16) [30]|0.99 4 0.09 (0.79-1.16) [50] 0.98 + 0.06 (0.86-1.10) [16]

lower fin height {0.0085 **/0.48 **} {0.0039 +/0.18} 10.0023/0.35 **} 4-0.0033/-0.16}

36-40 [1.10 + 0.06 (0.99-1.19) 110] 1.11 + 0.07 (0.99-1.22) [51/1.02 + 0.08 (0.90-1.16) [13]] 0.97 + 0.05 (0.86-1.04) (91

Snout - vent / SV/SUT 26-40 |1.66#0.15(1.32-2.15) [271] 1.51 #0.08 (1.37-1.71) (31]| 1.65 0.20 (1.36-2.19) (50]| 2.07+ 0.20 (1.81-2.39) [15]

snout - upper fin -0.0063/-0.17} {0.0000/0.11} 40.0062/0.22} {0.0079/0.22}

36-40 [1.65 + 0.19(1.32-2.15) 112)] 1.50 + 0.09 (1.44-1.66) 141/1.72 + 0.18 (1.42-1.93) [13]] 2.16 + 0.22 (1.84-2.39) [8]

Snout-spiracle/ | SS/VS| 36-40 | 1.20 + 0.15 (0.98-1.49) (8) | 1.23 + 0.17 (0.94-1.46) [51] 1.18 + 0.15(0.94-1.41) [91] 1.35 + 0.16(1.12-1.64) (7)

vent - spiracle

Interpupillar / PP/NN 36-40 1.51 + 0.05 (1.43-1.59) [ 8]| 1.42 + 0.06 (1.36-1.51) [5] | 1.67 + 0.07 (1.51-1.75) [9]| 1.51 + 0.04 (1.45-1.58) [7]

internarial.

Rostro-narial / RN/NP] 36-40 |1.03 + 0.10(0.85-1.20) [81| 0.92 + 0.14(0.73-1.15) [51] 0.78 + 0.15(0.53-1.13) [91] 0.86 + 0.15 (0.63-1.04) [7]

naro-pupillar

Interpupillar / PP/OD| 36-40 [1.61 + 0.06(1.55-1.74) [8]] 1.51 + 0.09(1.36-1.62) [51] 2.24 + 0.12(2.09-2.43) [91] 1.87 + 0.06 (1.81-1.96) [6]

oral disk width

Internarial / NN/OD] 36-40 |1.07 + 0.04(1.03-1.15) [81] 1.06 + 0.07 (0.99-1.16) [5] 1.34 + 0.08 (1.22-1.45) [9]| 1.25 + 0.03 (1.21-1.28) [6]

oral disk width

Interpupillar / PP/ED 36 - 40 3.24 + 0.26 (2.92-3.63) [8] | 3.15 + 0.22 (2.88-3.49) [5] | 4.17 + 0.42 (3.54-4.85) [9]| 3.92 + 0.07 (3.78-3.99) [7]

eye diameter

190À # HOSLITIINO “RIVSHIHOS

€IT

Source : MNHN, Paris

114 ALYTES 13 (4)

Table IV. - Comparison of collective median formulae (DuBois, 1995) of labial tooth rows (median

values, range in parentheses, number of specimens in brackets, presence and position of

gaps not indicated) of Osteocephalus oophagus, Osteocephalus taurinus, Phrynohyas

resinifictrix and Phrynohyas venulosa, in relation to the stage of larval development.

Median collective tooth row formulae

O. oophagus ©. taurinus P. resinifictrix P. venulosa

213(G4) (27| 2/3G4 (15 213

2/4(34) [10]| 2/3:4(84) [16]

215 51 | 2/4G4 tt

2/5(56 [11] 214 [51

e 21464 [3]

- 2/4(35) (6)

21484) [6]

2/5

214 ]

2/44) 13]

216 214

215 (4-6) 214

217:6 (6-7) 2/4

2/6(56) 214(84)

214

2/4

collective labial tooth row formula (Dugois, 1995) is described by median value and range

of tooth row counts in the anterior and posterior labium, respectively. Balance values of

tooth row counts (i.e., number of rows in the upper labium minus number of rows in the

lower labium) are according to ALTIG & JOHNSTON (1989), who categorized tooth row

formulae as balanced (equal number of rows on upper and lower labia), negatively

imbalanced (more rows on lower labium), or positively imbalanced (more rows on upper

labium). Terminology of jaw sheath morphology follows, e.g., KAUNG & KoLLRoS (1976)

and Fox (1984). Terminology of buccopharyngeal structures is in accordance with

WassERsUG (1976, 1980) and WassERSUG & HEYER (1988). Typological assignment of

breeding habitats (Table VIII) corresponds to DUELLMAN & TRUEB (1985), LANNOO et al.

(1987), DUELLMAN (1988), and ALTIG & JOHNSTON (1989).

The tadpoles analyzed comprise a wide range of developmental stages (Table II). For

all specimens, developmental stage, total length (fig. 1), gross body proportions (Table

HT), and tooth row counts (Table IV) were determined. Investigations on further

morphometric parameters (Tables III and V) and on buccopharyngeal structures were

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL

115

Table V. - Comparison of (A) external and (B) internal features of Osteocephalus oophagus, Osteocephalus

taurinus, Phrynohyas resinifictrix and Phrynohyas venulosa in advanced larval developmental

stages (38 + 2, numbers of specimens according to those in Tables 11 and IV).

A. External features.

P. venulosa

Parameters O. oophagus | O.taurinus | P. resinifictrix

Mean total length (mm) 29.7 32.7 37.0 418

Maximum total lengeh (mm, present study) 36.2 (stage 40) | 38.3 (stage 39) | 47.0 (stage 39) | 39.9 (stage 39)

Maximum total length (mm, from the literature) ‘| 26.0 (stage 37) | 31.7(stage 40) | 38.7(stage 39) | 49.4 (stage 41)

Labial tooth row formula (present study)? 213 2/6(4-7) 2/4(34) 3 (3-4) / 5 (4-5)

Labial tooth row formula (from the literature) M 213 2/6(4-7) 2/4(3-4) 3(3-4)/ 5 (4-7)

Total number of labial papillae ? 98:100 (88-114) | 160(146-162) | 120(90-144) | 168:182 (150-210)

Number of labial papillae per 100 jam * 1.52 2 œ 1.5

Height of exposed part of apical cone cells (um) * 10-12 21-27 35-47 27-32

Number of apical cone cells per 100 jm * 10 67 67 67

Number of labial keratodonts per 100 um d 10-14 10-13 69 9-12

Number of keratodont serrations * 8-10 10-12 12-14 14-16

B. Internal features.

Parameters O. oophagus | O.raurinus | P.resinifictrix | P. venulosa

Total number of postnarial papillae 3-4 34 34 3-4

Number of lateral ridge papillae per side 1 1 1 1

Number of buccal roof arena papillae per side * (30-45) 40-50 (25-35) 50-60

Diameter of secretory pits in dorsal velar

glandular zone (am) 15-20 15-20 5-15 10-25

Angle between longitudinal axis of nares and

transversal body axis 40°-45° 40°-45° 10°-20°

Number of praelingual palps per side 1 1 1 1

‘Total number of lingual papillae 2 2 2 2

Number of marginal buccal floor

arena papillae per side 36 11-16 6-10 10-15

Number of central buccal floor

(3-15) 13-15 CE) 12-14

Diameter of secretory pits in ventral velar

glandular zone (um) 20-30 15-30 5-10 15-40

Folding pattern of filter rows : 2°13° 3° 47) 2° (3°) 4°

Branching pattern of gills * 32 3: 2 x]

Length of gill tufts (am) * 350-400 400-500 300-350 450-650

Extension of lungs * A 2 4 34

ording 10 Table 1.

ranges in parentheses.

position on oral disk.

don on upper jaw sheath.

. Indistincly developed papillac in parentheses.

Secondary, tertiary, quaternary, features found

In median position on second cératobranchial.

m position on continuous upper 100 row.

Mean maximum caudal extension of inflated lungs in the abdominal cavity.

some cases in parentheses, found exceptionally in brackets.

Source : MNHN, Paris

116 ALYTES 13 (4)

restricted to specimens of stages 38 + 2. We excluded material below developmental stage

26 from subsequent statistical analysis because of small sample sizes.

For examination of buccopharyngeal features, scanning electron microscopy (SEM,

two specimens per species) and stereo light microscopy (methylene blue staining, three

specimens per species) were used. Data on external features were based on stereo

microscopic examination. Oral disk structures were confirmed by SEM (two specimens per

species). Preparation for SEM examination (Jeol JSM-35 CF) followed a standard

procedure (ethanol dehydration, critical-point-drying, gold sputter surface-coating).

Measurements were determined using a digital display length-measuring unit (Wild MMS

235) attached to a stereo microscope (Wild M8).

For statistical analyses, SPSS for Windows (Version N° 5.0.2.) was used. Gross body

proportions were described by basic descriptive statistics. Since material examined

included a wide range of developmental stages, we tested for covariation of ratios (Table

ID) with total length (univariate regression for each species) and correlation of ratios with

stage (Spearman rank correlation), as well as for isometry of growth. Gross body measures

were log-transformed and regressed against total length ({n y = In a + b In x; GouLp,

1966). Where the regression coefficient (b) was significantly different from 1.0, the

isometry hypothesis (H,) was rejected. We also estimated the regression of the ratios

describing the gross body proportions on total length.

With discriminant analysis, we estimated the optimum linear combination of variables

for differential diagnosis and the probability of misclassification of individuals. We used

the ratios of gross body proportions for these analyses and included the number of upper

and lower tooth rows. The latter two are meristic characters, and number of upper tooth

rows did not show any variation in three of the species studied. Therefore, some within

group variance-covariance matrices were singular. Discriminant analysis is fairly robust

against this type of violation of assumptions, and, as an exploratory tool, provided concise

results.

To estimate the relative influences of phylogenetic relationship (genus) and contem-

porary larval habitat (ecotype), univariate as well as multivariable analyses of covariance

(ANCOVA, MANCOVA) were performed (stages 38 + 2). The dependent variables were

logarithms of absolute gross body measures and of numbers of upper and lower tooth

rows. À full factorial mixed model was used with ecotype (phytotelm versus pond) and

genus (Osteocephalus versus Phrynohyas) as factors and total length as covariate (JOHNSON

& WICHERN, 1988; MORRISON, 1990).

For the material examined, each ecotype was represented in each genus, hence, all

cells of the model were occupied. However, the number of individuals in each cell was

unbalanced, and the individual sums of squares did not add up to the total sum of squares

in the ANCOVA. Since our design is “not extremely unbalanced” (SHAW, 1987), results

are not substantively compromised, and the application of other estimators will compare

to the ANOVA estimator (SWALLOW & MONAHAN, 1984; SHAW, 1987).

Source : MNHN, Paris

TOTAL LENGTH

26 28 30 32 34 36 38 40 42

STAGE

Fig. . — Size-stage graphs (scatter plots, regression lines) of developmental stages (GOSNER, 1960) and total lengths (mm): Osteocephalus

oophagus (A, squares); Osteocephalus taurinus (B, multiplication signs); Phrpnohyas resinifictrix (C, circles); Phrynohyas venulosa (D,

addition signs).

190À ® HOSLITIINO) ‘THVSHIHOS

LIT

Source : MNHN, Paris

118 ALYTES 13 (4)

RESULTS

For complete and concise description, presentation of results is based on the following

conventions:

() Descriptions apply to all four species where no interspecific differences are

indicated.

(2) Relative descriptive terms without quantification or reference of comparison refer

to external morphology of “‘generalized pond-type hylid larvae”’ (DUELLMAN, 1970) and to

the terminology for the description of buccopharyngeal morphology as used by, e.g.,

WaASsERSUG (1980), VIERTEL (1982), INGER (1985), and WASsERSUG & HEYER (1988).

(3) Precise values of the ratios describing body proportions and corresponding

abbreviations are presented in Table III, metric and meristic values of oral disk and

buccopharyngeal features in Table V.

For stages 26 through 40, almost all ratios describing gross body proportions covaried

with neither total length nor stage (Table II), and therefore were pooled, in addition to

the présentation for tadpoles in advanced developmental stages (38 + 2). In test for

isometry, the regression coefficients did not differ significantly from 1.0, i.e., tadpoles”

growth was isometric from stage 26 on in all gross body measures, except for tail height

in ©. oophagus and P. resinifictrix. In these species, the tail height grew proportionately

less than total length.

Through all stages examined, mean total length increased from O. oophagus, through

O. taurinus and P. venulosa, to P. resinifictrix (fig. 1).

GENERAL DESCRIPTION

Body slightly depressed; in dorsal view, elongate elliptical in P. venulosa, elongate

ovoid in O. taurinus, round ovoid to piriform in ©. oophagus, elongate elliptical to piriform

in P. resinifictrix. In P. resinifictrix, O. oophagus and O. taurinus, body width correlated

with amount of ingested eggs. Those eggs were visible through the abdominal wall, notably

in O. oophagus, whose integument was comparatively pale, thin and transparent.

Spiracular tube sinistral, directed posterodorsally, tightly attached to the body,

opening slightly below midline of body, at about one-half to two-thirds of distance from

tip of snout to opening of vent tube (SS/SV). Vent tube of moderate size, opening medially

to subdextrally at edge of ventral fin. Tail length equal to three-halves the snout-vent

length; tail proportionately shortest in O. oophagus and longest in P. resinifictrix (VT/SV).

Dorsal tail fin extending moderately onto body, inserting one-half to two-thirds the

snout-vent length distant from tip of snout; insertion of dorsal tail fin most anterior in P.

venulosa (SV/SU). Tail length twice to four times the tail height; tail height three to four

times the height of dorsal tail fin; ©. taurinus with least maximum tail height in relation

to tail length and shallowest upper tail fin compared to the maximum height of tail

(VT/HT, HT/UF). Fin edges arched; compared to the margins of caudal musculature,

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 119

more convex in P. venulosa than in P. resinifictrix, fairly parallel in Osteocephalus. Upper

and lower fins almost equally high (UF/LF), gradually tapering. Caudal musculature

moderate, nearly reaching the obtusely pointed tip of tail.

In dorsal view, snout nearly truncate in O. oophagus, bluntly rounded in the other

species. In lateral profile, snout acutely rounded in O. oophagus, rounded in P. resinifictrix

and P. venulosa, bluntly rounded in O. taurinus. Oral disk subterminal; position correlated

with shape of snout (Table V.A); most anterior and, if expanded, partly visible in dorsal

view in ©. oophagus, and increasingly more posterior from P. resinifictrix trough P.

venulosa to ©. taurinus. Eyes moderately large, situated dorsolaterally; widely spaced,

directed laterally, visible in ventral view in Phrynohyas, slightly less separated, directed

dorsolaterally, not visible in ventral view in Osteocephalus. Nostrils rimmed, directed

anterolaterally, about midway between pupillae and tip of snout in Osteocephalus, slightly

more anterior in Phrynohyas (RN/NP); internarial distance about two-thirds the

interpupillar distance (PP/NN).

ORAL DISK

Oral disk (fig. 2) medium-sized, moderately expanded laterally, slightly trilobate

ventrally. Labial papillae average sized; extension of dorsomedian gap about one-fifth of

continuous upper tooth row in ©. oophagus, P. resinifictrix and P. venulosa, one-third of

continuous upper tooth row in O. taurinus. In the fully expanded oral disk, marginal labial

papillae arranged in a single or alternating double row; submarginal papillae scattered in

the lateral oral disk portions, most frequently ventrolaterally, often in lateral continuation

of the outermost lower tooth rows. Labial papillae more numerous in P. venulosa and O.

taurinus than in P. resinifictrix and O. oophagus (Table V.A); denticulate papillae in the

lateral portions of the oral disk and in lateral extension of secondary tooth rows frequently

present in P. venulosa, exceptionally present in P. resinifictrix.

In early larval stages, two upper and three lower labial tooth rows present (Table IV,

fig. 7). Labial tooth row formula 2/3 retained throughout entire larval period in ©.

oophagus, and occasionally in advanced stages in P. resinifictrix; in the other species,

additional lower tooth rows appearing with increasing developmental stages (Table IV).

One or two additional upper tooth rows developed only in P. venulosa (Tables IV and

V.A). Maximum total number of tooth rows 5 in ©. oophagus, 7 in P. resinifictrix, 9 in

©. taurinus and P. venulosa.

Morphology of ontogenetically basic upper two and lower three tooth rows quite

homogeneous in all species. Outer upper primary row continuous, inner one shortly

interrupted medially, with its median ends often covered by the outer tooth row, both

coextending far towards the lateral corners of the oral disk. Ridges of lower three primary

rows frequently indented medially when not fully expanded; innermost always continuous

in P. resinifictrix, occasionally with a very narrow median interruption in the other species;

outer two always continuous; all three typically of broad and almost equal lateral

extension, outermost often shorter than the inner ones in O. oophagus.

Tooth rows in excess of the basic 2/3 pattern added centrifugally in the upper

(P. venulosa) as well as in the lower labium; typically, with broad median interruption in

Source : MNHN, Paris

120 ALYTES 13 (4)

Fig. 2. — Drawings of oral disk (top), floor (middle) and roof (bottom) of buccopharyngeal cavity

after SEM micrographs: À, Osteocephalus oophagus (193/3.87); B, Osteocephalus taurinus

(2.33/4.94); C, Phrynohyas resinifictrix (1.84/5.90); D, Phrynohyas venulosa (2.33/4.62). Maxi-

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 121

mum horizontal (graph) diameter (in mm) of oral disk / of buccopharyngeal cavity in

parentheses. Measures according to SEM micrographs; SEM preparation shrinkage factor about

0.70.

Source : MNHN, Paris

122 ALYTES 13 (4)

the upper labium (P. venulosa), continuous in the lower labium; shorter and more

frequently broken the more distally positioned.

Labial keratodonts in a single series on each ridge, cone-shaped with spatulate apical

portions bearing acute marginal denticles; size and density of keratodonts, and number of

apical indentations slightly varying among species and genera (Table V.A).

Jaw sheaths (fig. 2) broadly dark pigmented, robust, wide, reaching far towards

lateral corners of oral disk; front surfaces of average curvature; median part of occlusive

margins slightly convex and rectilinear; jaw sheaths moderately narrower and more

delicate in ©. oophagus than in the other species. Edges of upper and lower jaw sheaths

smooth in O. oophagus, finely serrated in the other species. In SEM examination, exposed

parts of apical cone cells incisiviform in ©. oophagus, caniniform in the other species (fig.

3); shape subrectangular with nearly straight distal edges and tight lateral attachment in

O. oophagus, acutely pointed, cone-shaped in ©. taurinus, lanceolate in Phrynohyas;

longest and most acutely tapered in P. resinifictrix (Table V.A).

BUCCOPHARYNGEAL CAVITY

Buccopharyngeal surface features quite homogeneous in all four species. Variation

mainly restricted to number and size of the papilla-derived structures, being longer and

more numerous in ©. taurinus and P. venulosa than in ©. oophagus and P. resinifictrix,

With O. oophagus showing the simplest and P. venulosa the most differentiated pattern

(Table V.B, fig. 2).

Buccopharyngeal roof. — Prenarial arena broad, centrally with stout tuberous

pustulations, scattered or fused to a median knob or ridge, of variable arrangement even

within a species. Internal nares elongate, obliquely oriented; relative length of internal

nares in Phrynohyas about two-thirds that of Osteocephalus; angle between longitudinal

axis of internal nares and transversal body axis smaller in P. venulosa than in the other

species (Table V.B). Anterior narial wall lined by tiny, laterally slightly more elongate

papillae; posterior narial wall valve smooth-edged, slightly lobate; narial valve projections

faint or slightly lobate. Postnarial papillae arranged in an anteriorly convex arch except

for some scattered minor pustulations, well separated from each other in Osteocephalus,

more basely fused in Phrynohyas. One lateral ridge papilla per side, broad-based, palp-like,

bearing rather stout or conical pustulations. Median ridge average sized, flap-like;

triangular, more slender, elongate, small-based, distant from the lateral ridge papillae, with

a pointedly lobed margin in ©. oophagus and P. resinifictrix, semicircular, more stout,

broad-based, laterally extended towards the lateral ridge papillae, with the margin tightly

bordered by a row of small pustulations in ©. taurinus and P. venulosa (Table V.B).

Papillae in the spacious buccal roof arena comparatively small in all species. Papillae

bordering the arena more distinct; in the lateral corners of the arena slightly elongate in

©. taurinus and P. venulosa, almost absent in ©. oophagus and P. resinifictrix. Lateral roof

papillae scarce but minor pustulations. Dorsal velum continuous across midline with the

medial edge bare of papillation. Glandular zone distinct in all species. Width of glandular

zone and diameter of secretory pits less in P. resinifictrix than in the other species (Table

VB, fig. 2).

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 123

Fig. 3. — SEM micrographs of edge of median upper jaw sheath (scale line equals 10 pm): A,

Osteocephalus oophagus; B, Osteocephalus taurinus; €, Phrynohyas resinifictrix; D, Phrynohyas

venulosa.

Source : MNHN, Paris

124 ALYTES 13 (4)

Buccopharyngeal floor. — Prelingual arena scattered with more or less papilliform,

stout, laterally and posteriorly more frequent pustulations; almost bare in ©. oophagus. One

average sized, broadly based prelingual palp per side; somewhat larger with the papilliform

marginal lobations more elongate, finger-like, and less numerous in O. oophagus and P. resi-

nifictrix than in the other species. One pair of slim cylindrical lingual papillae. Buccal floor

arena well defined, center almost bare; papillae bordering the arena moderately enlarged,

conical, simple; lateroposteriorly most distinct and frequently fused basely.

Buccal pockets large; orientation almost transversal in ©. oophagus, oblique in the

other species. Prepocket papillae scarce stout pustulations. Ventral velum distinct with

evident spicular support; three spiculae per side. One marginal velar projection per filter

cavity; least developed in O. oophagus. Free edge of ventral velum lined by a distinct

glandular zone; secretory pits largest in P. venulosa, slightly smaller in Osteocephalus, less

prominent in ©. oophagus than in O. taurinus, considerably smaller in P. resinifictrix

(Table V.B). Branchial food traps with distinct secretory ridges.

Median notch broad, leaving glottis fully exposed; glottal lips broad, elevated;

exposure of glottis less distinct in O. oophagus than in the other species (fig. 2). Lungs well

developed in all species (Table V.B, fig. 5) already in early ectotrophic stages; almost

extending to caudal curvature of abdominal cavity except for O. oophagus (Table V.B).

Esophageal funnel spacious.

Depth of branchial baskets and complexity of the filter rows — i.e., degree of

branching (Table V.B), height, depth, and density of filter rows (fig. 4) — decreasing from

P. venulosa through O. taurinus to P. resinifictrix and O. oophagus. Internal gills least

developed in P. resinifictrix, next least in O. oophagus, most differentiated in P. venulosa

(Table V.B, fig. 4).

DISCRIMINANT ANALYSIS

Discriminant analysis resulted in three, linearly independent, discriminant functions

(Table VI). The first discriminant function was dominated by the number of upper tooth

rows. The second discriminant function was determined by the ratio of length of tail to

height of tail (VT/HT), but was also influenced by the ratio of length of tail to snout-vent

length (VT/SV) and the number of tooth rows in the lower labium. The latter two

characters also dominated the third and least significant discriminant function. The ratios

of height of tail to height of the upper fin (HT/UF), height of the upper fin to height of

the lower fin (UF/LF), and of snout-vent length to the distance between tip of snout and

insertion of the upper fin (SV/SU) poorly discriminated the species (Table VI.A).

Misclassification of specimens into species was remarkably rare and restricted to the

phytotelmonous species (Table VI.B). The group centroid of P. venulosa was clearly

separated on discriminant function one. On discriminant function two, group centroids

were highly positive in ©. taurinus, while they were negative in ©. oophagus and P.

resinifictrix. On the last discriminant function, group centroids of ©. oophagus and P.

resinifictrix were distinctly separate. In summary, the phytotelmonous species clustered

closely, while the two pond types formed two more separated clusters (Table VI.C, fig. 6).

Source : MNHN, Paris

Fig. 4. — SEM micrographs of internal gills and filter of median part of second ceratobranchial (scale line equals 100 am): À,

Osteocephalus oophagus; B, Osteocephalus taurinus; C, Phrynohyas resinifictrix; D, Phrynohyas venulosa.

190AÀ # HOSLITINO) “RIVSHIHOS

STI

Source : MNHN, Paris

126 ALYTES 13 (4)

Fig 5. — SEM micrograph of floor of the buccopharyngeal cavity and lung sacs of Phrynohyas

venulosa (scale line equals 1000 um).

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL

127

Table VI. - Discriminant analysis of variables describing gross body proportions (abbreviations

according to Table III), including number of tooth rows in the upper and lower

labium, for Osteocephalus oophagus, Osteocephalus taurinus, Phrynohyas resinifictrix

and Phrynohyas venulosa (stages 38 + 2).

A. Coefficients of standardized canonical discriminant functions.

B. Classification of individuals into species (correctly classified individuals on the main diagonal).

Parameters Function 1 Function 2 Function 3

Eigenvalues 13.838 3.470 1.576

Vent - tail tip / snout - vent VT/SV 0.001 -0.602 1.018

Vent - tail tip / tail height VT/HT 0.103 1.131 -0.265

Tail height / upper fin height HT/UF -0.080 -0.209 0.189

Upper fin height / lower fin height | UF/LF -0.079 -0.203 -0.103

Snout - vent / snout - upper fin SVISU 0.080 0.021 -0.012

Number of upper tooth rows UTR 0.978 -0.010 -0.057

Number of lower tooth rows LTR 0.075 0.527 0.496

Actual numbers

Assigned numbers of cases

Species

of cases O. oophagus | O.taurinus | P. resinifictrix | P. venulosa

©. oophagus 20 19 0 1 0

©. taurinus 29 0 29 0 0

P. resinifictrix 43 [l 0 4 0

P. venulosa

C. Canonical discriminant functions at group centroids.

Species Function 1 Function 2 Function 3

O. oophagus -1.569 -1.243 -2.320

O. taurinus -1.094 2.886 0.036

P. resinifictrix -1.347 -1.319 1.077

P. venulosa 10.090 -0.194 -0.076

Source : MNHN, Paris

(+) € SALATV

65 0 5 10 15 20 25 er 4 a co pt à

DF1 DF3

Fig. 6. — Scatter plot of scores of specimens and group centroids (arrows) on discriminant functions (DF) 1 and 2, and 3 and 2:

Osteocephalus oophagus (A, squares); Osteocephalus taurinus (B, multiplication signs); Phrynohyas resinifictrix (C, circles); Phrynohyas

venulosa (D, addition signs).

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 129

Table VII. - Analysis of covariance of variables describing gross body proportions (abbreviations

according to Table III), including number of tooth rows in the upper and lower

labium, for Osteocephalus oophagus, Osteocephalus taurinus, Phrynohyas resinifictrix

and Phrynohyas venulosa (stages 38 + 2).

ANCOVA: sums of squares (* P < 0.05, ** P < 0.01; individual sums of

squares do not sum to total sum, because occupation of cells is unbalanced).

MANCOVA: coefficients of standardized canonical discriminant functions.

ANCOVA MANCOVA

Parameters

Fée Genus |Ecotype Les nee Residuall Total | Genus |Ecoiypel US

ISnout - vent SV |0.16**| 0.00 | 0.02**| 0.02**| 0.30**! 0.08 | 0.38 | 0.121 | 0.378 |-0.387

Tail height HT | 0.07**| 0.33**| 0.13**| 0.16**] 1.16**| 0.15 | 1.30 |-0.522 | 0.535 |-0.444

ISnout - upper fin SU | 0.18**| 0.25**| 0.08* | 0.07* | 0.48**| 0.40 | 0.89 | 0.121 |-0.001 | 0110

Number of upper tooth rows|UTR| 0.00 | 0.28**| 0.48**| 0.41**! 1.27**] 0.07 | 1.34 |-0.737 |-0.678 |-0.661

Number of lower tooth rows| LTR| 0.00 | 0.01 | 1.52**| 0.44**| 2.26**| 0.28 | 2.26 |-0.009 |-0.718 | 0.609

ANALYSIS OF COVARIANCE

With ANCOVA, influence of total length, ecotype, genus, and the interaction between

ecotype and genus, were significant for gross body measures with the exception of genus

on snout-vent length. The tooth row characters were not influenced by total length;

influence of ecotype and the interaction between ecotype and genus were significant for

both upper and lower tooth rows, whereas the influence of genus was only significant for

the number of upper tooth rows (Table VII).

MANCOVA produced three, linearly dependent, discriminant functions (Table VII).

Both ecotypes and genera were best differentiated by number of upper tooth rows but also

by height of tail. Number of lower tooth rows only differentiated between ecotypes. For

the effect of genus, tooth row characters and height of tail enter with identical signs,

whereas, for the effect of ecotype, these variables enter with opposite signs. Interactions

were influenced mainly by the tooth row characters with opposite signs, but also by height

of tail and snout vent length, both in the same direction with the number of upper tooth

rows. Craniad extension of upper fin discriminated only weakly between any factor or

interaction (Table VII).

The observed standardized MANCOVA discriminant functions associated with

genus, ecotype, and their interaction, condensed the morphometric results: compared with

the pooled pond-dwelling species, the pooled phytotelmonous species were characterized

by fewer upper and lower tooth rows, while, relative to their total length, they had higher

tails. For the pond-dwelling species, higher number of lower tooth rows was found in both

genera while higher number of upper tooth rows was found only in P. venulosa.

Source : MNHN, Paris

130 ALYTES 13 (4)

Discrimination in height of tail between ecotypes and genera was influenced mainly by ©.

taurinus, which had the relatively lowest tail, while the tail characteristics of the other

species were comparatively uniform.

DISCUSSION

Tadpoles of both the pond-dwelling and the phytotelmonous species studied are

characterized by overall average body dimensions and, thus, resemble typical hylid pond

tadpoles (DUELLMAN, 1970). P. resinifictrix and P. venulosa larvae are longer, have

proportionately longer tails, higher upper tail fins, more lateral eyes, and more anterior

nares than ©. oophagus and O. taurinus larvae. As far as known from the literature (see

caption of fig. 7), these intergeneric differences in height of tail and position of eyes and

nares also apply to the other species of the two genera.

Collectively, buccopharyngeal features of the four species studied represent a

relatively uniform type, which we consider omnivorous to macrophagous and capable of

both branchial and pulmonary respiration. Nevertheless, number and size of buccopharyn-

geal papillae, complexity of branchial filter system, development of velar secretory tissues,

differentiation of gills, along with number of tooth rows and of labial papillae, correspond

to the principal larval habitat types as usual among anuran larvae: relative structural

“simplification” characterizes the phytotelmonous larvae, wheras “elaboration” charac-

terizes the pond-dwelling ones (WASSERSUG, 1980; ALTIG & JOHNSTON, 1989).

Less differentiated external and internal buccopharyngeal features along with the

more anterior position of the oral disk in the phytotelmonous species are explained by

their predominantly macrophagous feeding habits. Among the phytotelm-dwelling species

analyzed, however, most internal buccopharyngeal features are less differentiated in O.

oophagus than in P. resinifictrix. This divergence may be explained by different degrees of

“specialization” (TRUEB, 1973) to macrophagous nutrition: ©. cophagus is obligatorily

macrophagous, feeding on conspecific fertilized eggs and tadpoles (HôbL, 1993), while P.

resinifictrix is omnivorous (GRILLITSCH, 1992; SCHIESARI, 1993), predominantly macro-

phagous (ScHiesari, 1993), feeding mainly on conspecific fertilized eggs, but also on

detritus (SCHIESARI & GoRDO, 1993). Among the pond-dwelling species, buccopharyngeal

surface features are more differentiated in P. venulosa, indicating a greater reliance on

microphagous feeding than in O. taurinus. O. taurinus larvae have been observed in the

field to be voracious egg-eaters (SCHIESARI, personal observation), which matches the

morphological indication of macrophagy. Osteocephalus elkejungingerae, whose tadpoles

are highly cannibalistic when laboratory bred (HENLE et al., 1983), represents a further

species within the genus with macrophagous larvae.

Lungs are spacious in the four species analyzed, although more expanded in

Phrynohyas than in Osteocephalus. Comparing among phytotelmonous larvae, glottis and

lungs are large in P. resinifictrix as in the tree hole-dwelling Philautus sp. and Theloderma

stellatum (WaAssERSUG et al., 1981), but are medium sized in O. oophagus as in the

bromeliad-dwelling Osteopilus brunneus (LANNOO et al., 1987). Among the two types of

Source : MNHN, Paris

ERA

©. buckleyi

©. elkejungingerae

©. langsdorffii

©. oophagus

©. taurinus

©. verruciger

P. mesophaea

P. resinifictrix

P. venulosa

Fig. 7. — Variation of larval tooth row formulae among Osteocephalus and Phrynohyas species.

Schematic drawings; median interruptions and relative lengths of lower tooth rows not

considered.

References: Osteocephalus buckleyi (HERO, 1990); Osteocephalus elkejungingerae (HEN-

LE, 1981); Osteocephalus langsdorffii (DUELLMAN, 1974); Osteocephalus oophagus (present study

and as in Table 1); Osteocephalus taurinus (present study and as in Table 1); Osteocephalus

verruciger (TRUEB & DUELLMAN, 1970); Phrynohyas coriacea (SCHiESARI & MOREIRA, in press);

Phrynohyas mesophaea (LUTZ, 1973; SCHIESARI, personal observation); Phrynohyas resinifictrix

(present study and as in Table I); Phrynohyas venulosa (present study and as in Table 1).

Source : MNHN, Paris

132 ALYTES 13 (4)

phytotelms, tree holes offer the more anaerobic aquatic environment: dissolved oxygen

was 0.2 mg/l (surface water 26°C, 25 1 water volume) in tree hole water dwelled by P.

resinifictrix larvae, but 2.6 mg/l (26°C, 10-15 ml water volume) in the water of bromeliad

leaf axils inhabited by ©. oophagus larvae (SCHIESARI, personal observation). Correspon-

dingly, development of lungs indicates greater importance of pulmonary respiration in tree

hole-dwelling larvae than in the bromeliad-dwelling ones, whereas gill development

indicates that the contrary applies to branchial respiration. Comparing within a genus,

development of internal gills and lungs indicate greater reliance on branchial respiration

but also on pulmonary respiration (especially in O. taurinus) in the pond-dwelling species.

“Internal oral structures of anuran larvae can be used to make reasonably sound

predictions about the feeding and respiratory ecology of anuran larvae” (WASSERSUG,

1980). Respiratory structures in the species studied indicate comparatively low average

levels of dissolved oxygen also in the larval pond habitats. However, since “lungs appear

to be advantageous to aquatic organisms even in normoxic water in that they allow

buccopharyngeal surfaces to be dedicated fully to feeding rather than respiration”

(WASsERSUG & MUrPHY, 1987), extensive development of lungs in the pond-dwelling

species studied might further correlate with their less macrophagous, more omnivorous

nutrition and, at last, with correspondingly higher motility and metabolic rate in these

species which, in contrast to the phytotelmonous ones, develop without “parental” food

supply in a, typically, less confined habitat.

For the four species studied, shape of labial keratodonts represents a type very

common in Ranoidea tadpoles (e.g., HÉRON-ROYER & VAN BAMBECKE, 1889; GOSNER,

1959; INGER, 1985; ALTIG & JOHNSTON, 1989). Gross jaw sheath morphology shows no

considerable peculiarities in the pond-dwelling larvae but is remarkable in the phytotel-

monous ones (e.g., DUELLMAN, 1970): edges are smooth in ©. oophagus, whereas P.

resinifictrix shows elongate, acutely pointed serration. Smooth edged jaw sheaths are rare

in anuran larvae. Among phytotelmonous tadpoles, the upper jaw sheath is smooth and

the lower finely serrated in two oophagous species, the bromeliad-dwelling Hyla zeteki

(DUELLMAN, 1970: 326, first paragraph) and the tree hole-dwelling Philautus sp.

(WassEersUG et al., 1981); furthermore, in some egg-eating Jamaican hylids, jaw sheaths are

not denticulate (NOBLE, 1929). However, some stream-dwelling tadpoles (Hyla mixe, Hyla

mixomaculata) also bear smooth-edged jaw sheaths (DUELLMAN, 1970), and, in contrast,

fine uniform jaw sheath serration is frequently reported for phytotelmonous, oophagous

larvae (e.g., Osteopilus brunneus, LANNOO et al., 1987; Theloderma stellatum, WASSERSUG

et al., 1981). Thus, various jaw sheath patterns are apparently suitable for oophagous

feeding. Smooth-edged jaw sheaths are likely to have different functional correlates in

rheophilous and oophagous tadpoles. In rheophilous larvae, smooth jaw sheaths may be

most effective for grazing on constrained epilithic substrates; in obligatorily macrophagous

oophagous larvae, such as O. oophagus, serration simply may have become unnecessary or

even disadvantageous for ingesting eggs as a whole.

In the ontogenetic sequence of tooth row appearance, the labial tooth row formula

2/3 is primary in both genera (fig. 7). Additional, secondary upper tooth rows develop in

Phrynohyas (except in P. resinifictrix), where they are added distally. In contrast, they are

absent in Osteocephalus, with the exception of O. elkejungingerae, where they are proximal

and poorly formed. Hence, the two genera are well distinguished by different derived types

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 133

of ontogenetic tooth row increase (types A to E in ALTIG & JOHNSTON, 1989): in both

genera, tooth rows are added centrifugally in the lower labium, but addition of tooth rows

in the upper labium is absent (type not considered in ALTIG & JOHNSTON, 1989) or

centripetal (type B) in Osteocephalus and centrifugal (type C) in Phrynohyas. The basic 2/3

tooth row pattern (type A) persists only in the phytotelmonous O. oophagus and

occasionally in P. resinifictrix.

Total numbers of tooth rows of 5 or more than 5, as in ©. oophagus and in P. resinific-

trix respectively, compare to the highest known for phytotelmonous hylids (LANNOO et al.,

1987). Likewise, among non-phytotelmonous hylids and anuran larvae in general (ALTIG &

JOHNSTON, 1986, 1989), total number of tooth rows is comparatively high in all Osteocepha-

lus and Phrynohyas species (fig. 7): in the pond- and stream-dwelling Osteocephalus species,

maximum total numbers of tooth rows vary from 7 to 10, which is typical for lotic but not

rheophilous hylid tadpoles. Surprisingly, the highest tooth row counts of up to 10 or 11 are

present in the evidently lentic Phrynohyas coriacea, P. mesophaea and P. venulosa, as in the

pond-dwelling tadpoles of the hylid Trachycephalus jordani (MCcDiarMiD & ALTIG, 1990).

These tooth row counts exceed the upper limit of the range of variation known for other

lentic hylids, and, furthermore (compared to the data in ALTIG & JOHNSTON, 1986), are within

the upper third of the range of variation in lotic hylid tadpoles.

Upper and lower tooth row counts are negatively imbalanced in the two genera

studied. Balance values of — 1 and —2 as shown in O. oophagus and P. resinifictrix are

moderate among phytotelmonous tadpoles (LANNOO et al., 1987), and represent the most

frequent type among hylids as well as among anuran larvae in general (ALTIG & JOHNSTON,

1986, 1989). In all other Osteocephalus and Phrynohyas species, number of tooth rows in

the lower labium exceeds that in the upper labium notably (fig. 7): balance values of —3

to —4 are the most common in the two genera (fig. 7) and in Osteocephalus even reach

— 5 (0. taurinus) and —6 (0. buckleyi). However, in anurans in general, balance values of

—3to —6 are rare and are most common to lotic larvae (usually neotropical Hyla species),

though —3 may be also found in phytotelmonous larvae (e.g., Hyla bromeliacea;

DUELLMAN, 1970).

Among hylids, the combination of a wide dorsomedian interruption of the peribuccal

papillary margin, as typical for “generalized pond-type” hylid tadpoles (DUELLMAN, 1970),

with a number of tooth rows exceeding the typical pond-type 2/3 pattern, is rare and

characterizes both Osteocephalus and Phrynohyas. À wide dorsomedian papillary gap and

increased number of lower tooth rows, as is typical in Osteocephalus, is known, e.g., in the

bromeliad Hyla dendroscarta and the pond-dwelling Hyla rufitela larvae (DUELLMAN,

1970). A wide dorsomedian papillary gap and increased numbers of both upper and lower

tooth rows, as is typical in Phrynohyas, is only known in the pond-dwelling tadpoles of

Hyla geographica (BOKERMAN, 1963; HERO, 1990; RADA DE MARTINEZ, 1990) and

Trachycephalus jordani (McDiarmiD & ALTIG, 1990). Let us mention here that TRUEB

(1970) suggested comparatively close phylogenetic relationship for the genera Osteocepha-

lus, Phrynohyas and Trachycephalus. This proposal based on geographical and adult

morphological evidence is supported by larval oral disk morphology and is not

contradicted by the other larval features examined in this study.

Source : MNHN, Paris

134 ALYTES 13 (4)

SUMMARY AND CONCLUSIONS

Although breeding sites of the species studied in depth are assigned to two principal

types (phytotelms and ponds), with discriminant analysis, the external morphological

characters analyzed cluster the four species into three distinct groups (fig. 6). The first is

the phytotelmonous group with only slight differences between the bromeliad species (0.

oophagus) and the tree hole habitating species (P. resinifictrix), suggesting comparatively

little ecological diversity among these species. The other two morphotypological groups

are both pond forms (O. taurinus and P. venulosa), suggesting greater ecological diversity

in that habitat.

Among phytotelmonous hylids (as reviewed in LANNOO et al., 1987), external larval

morphology assigns both ©. oophagus and P. resinifictix to a relatively “generalized”

(TRusB, 1973) larval type in that they show “typical pond tadpole”’ (LANNOO et al., 1987)

body proportions, and oral disk features similar to those phytotelmonous species which

feed mainly on detritus. For both species analyzed, dietary information and buccopharyn-

geal morphology indicate predominating oophagous, carnivorous macrophagy, reduced

microphagy, and branchial as well as pulmonary respiration with evidently greater reliance

on generalized diet in P. resinifictrix. For P. resinifictrix, LANNOO et al. (1987) therefore

stated that “they appear restricted to larger aquatic bodies, which are more likely to occur

in tree holes than in leaf axils”. P. resinfictrix, in fact, is exclusively known to breed in

spacious tree holes (Table VIII). Comparatively high degree of morphological congruency

of P. resinifictrix and O. oophagus corresponds to their collectively relatively low degree

of “specialization” (TRUEB, 1973), which, for O. oophagus, may be explained by its

remarkable flexibility in breeding habitat selection: although typically breeding in

bromeliads, this species has also been reported to breed in other, considerably diverse

water-filled plant structures (Table VIID).

For the non-phytotelmonous larvae of the two genera, data from the literature on

external larval morphology of other species greatly match the intergeneric differences

observed in the species studied in depth in this study. Collectively, comparatively high tail

fins, lateral eyes, and balanced tooth row formulae, as in Phrynohyas, are typical for

nektonic lentic tadpoles, while the opposite, as shown by Osteocephalus, is typical for

benthic, commonly moderately lotic larvae. Data compiled from the literature on breeding

habitats of non-phytotelmonous congeners (Table VIII) apparently parallel the above

morphotypological grouping: all Phrynohyas species regularly breed in ponds, whereas

Osteocephalus larval habitats comprise lentic, and facultatively as well as permanently lotic

habitats. Most of the Osteocephalus species, in fact, breed in a stream habitat.

However, total number of tooth rows in the lentic Phrynohyas species matches or

exceeds that of the most lotic Osteocephalus species, and, thus, corresponds to habitat

inversely than usual among anuran larvae. Furthermore, compared to typical pond-type

larvae, tooth row counts are unusually high and balance values are unusually low at least

in the non-phytotelmonous species of both genera, more like in lotic rather than in lentic

tadpoles. Our observations might be paralleled in other groups of neotropical hylids:

WASsERSUG (1980) found a mosaic of stream and pond related features in the

Source : MNHN, Paris

Table VIII. - Literature survey on breeding habitats of the Osteocephalus and Phrynohyas species with information on larval external morphology available

fig. 7). Regions réfer to the sites of observation and do not necessarily cover the species” entire range of distribution.

Species Principal habitat types Habitats Regions

Stream | Pond_[Phytoælm|

O. buckleyi + - = [Smallsreams Central and upper HERO (1990); HÔDL (I

Amazon basin RODRIGUEZ & DUELLMAN (1994)

(Brasil, Perd)

©. ekejungingeræe| + E = [Small shallow, slowlÿ to moderately fast | Forest Lower eastern Andean | HENLE (1981); HENLE Cal. (1983);

flowing waters slopes (Per) HENLE (1992)

O. langsdorfi - F = [Temporary ponds Foresrborder | Atlantic forest (Brasil) | DUELLMAN (1974). FROST (1985);

SCHIESARI (pers. obs.)

O. oophagus = = + |Epiphyuc or ground bromelads, palm leaf | Primary and Central Amazon basin | JUNGFER & SCHIESARI (1999);

axils, palm bracts lying in the ground, tre | secondary forest | (Brasil) JUNGFER & WEYGOLDT (1995)

holes up to about 35 m high, plastic basins on

the ground

g E + | Leaf axils in arboreal plants, bromeliads, palm | Forest Cenwal Amazon basin | HERO (1990)

leaf axils, equivalent arboreal plant structures (Brasil) HÔDL (1990, 1993) ‘

O: species - - + |Bromeliads Primary forest | Upper Amazon basin | RODRIGUEZ & DUFLLMAN (1994)

(Perë)

O. taurinus + hi . Large and small, temporary and permanent Primary and Central and upper BOKERMANN (1964); HERO (1990);

ponds, streamside ponds and isolated forest secondary forest | Amazon basin (Brasil, | HÔDL (1990, 1993); ZIMMERMAN &

ponds, large and small streams, in forest and Peré) RODRIGUES (1990); GASCON (1991,

forest-edge sites 1993); RODRIGUEZ & DUELLMAN (1994)

O. verruciger # . pe Quiet pool in a stream ‘Humid montane Lower eastern Andean | TRUEB & DUELLMAN (1970, 1971)

forest slopes (Ecuador)

F. coriacen = F = [Temporary ponds Primary forest | Central and Upper HÔDL (1990, 1993);

Amazon basin RODRIGUEZ & DUELLMAN (1994);

(Brasil, Perd) SHIESARI & MOREIRA (in press)

F.mesophaea = F = |Temporary ponds = “Atlantic forest (Brasil) _| FROST (1985); _SAZIMA (1974)

P. resinificrix = - + [Spacious cavities High in large trees Primary forest | Amazon basin References in ZIMMERMAN & HODL

(Brasil, Perd) (1983); HERO (1990, 1991); GRILLITSCH

(1992); HÔDL (1993); SCHIESARI (1993);

RODRIGUEZ & DUELLMAN (1994)

P. venulosa + = | Shallow temporary ponds Open, nonforested | Neotropical lowlands | References in

‘but also primary ZIMMERMAN & HÔDL (1983);

and secondary RODRIGUEZ & DUELLMAN (1994)

forest sites

1. Described under the name Osteocephalus sp. (W. HÔDL, personal communication).

190À ®@ HOSLITINO) VSHIHOS

S€T

Source : MNHN, Paris

136 ALYTES 13 (4)

bromeliad-dwelling Hyla dendroscarta tadpoles. Within species, closely related to H.

dendroscarta, he recognized a variety of breeding habitats, which, as in the genera

Osteocephalus and Phrynohyas, comprise ponds, streams, and phytotelms.

In summary, Osteocephalus and Phrynohyas larval habitats are notably diverse. But

morphologically, larvae among the two genera are less diverse in that they share similar

oral disk and buccopharyngeal features as well as overall average body proportions, high

number of tooth rows, low balance values, omnivorous to macrophagous diets, and

branchial as well as pulmonary respiration. Particularly, number of tooth rows and lung

development do not correspond to habitat in the usual anuran larval fashion.

Two, not mutually exclusive, biological explanations for the generally high number of

tooth rows and the well developed lungs in the extant non-phytotelmonous species

examined are possible:

(1) Adaptation to contemporary environment. — Their contemporary larval envi-

ronment is collectively characterized by comparatively high temperatures and correspon-

dingly low oxygenation. If aerial respiration is the expected major factor in the

development of lungs (as reviewed in WASsERSUG & SEIBERT, 1975 and WASSERSUG &

MurPHyY, 1987), low levels of dissolved oxygen favor early and extensive development of

lungs in lentic but also in lotic tadpoles (NOBLE, 1929). If adhesion to substrate is the

expected major action among the suggested functions of labial teeth (as reviewed in ALTIG

& JoHNsTON, 1989), increased number of tooth rows, i.e., increased adhesive efficiency of

the oral disk, might be an adaptation to compensate the hydrodynamic disadvantage

(NoBe, 1929; WaAssERSUG, 1980) of well developed lungs especially in lotic environments.

This explanation is more likely to apply to the lotic benthic Osteocephalus species than to

the lentic nektonic Phrynohyas species. For both genera, increase in adhesive capacity of

oral disk might also be a not yet considered correlate to macrophagous nutrition.

(2) Persistent influence of ancestral patterns. — Some features, such as high number

of tooth rows in the lentic species, might represent an ancestral lotic pattern and might

have persisted relatively unchanged. If breeding in phytotelms is “derived” (TRUEB, 1973)

in the genera studied, other features, such as pulmonary respiration and omnivorous,

predominantly macrophagous nutrition in non-phytotelmonous larvae, might further be

exaptations (GouLp & VRBA, 1982) to life in lowly oxygenated and ‘“confined” (e.g.,

LANNOO et al., 1987) phytotelmonous habitats.

RESUMEN

La morfologia externa y bucofaringea de larvas habitantes de fitotelmata y charcos

de cuatro especies de hilidos neotropicales fueron analizadas con relacion a diagnosis

diferencial y ecomorfologia. Los häbitats larvales tipicamente comprenden bromélias

(Osteocephalus oophagus), hoquedades en ärboles (Phrynohyas resinifictrix), charcos en

äreas de selva (Osteocephalus taurinus) y charcos en âreas abiertas (Phrynohyas venulosa).

Un anälisis discriminante de caracteres externos morfométricos revelé dos subgrupos

ligeramente diferentes dentro del grupo de habitantes de fitotelmata pero dos subgrupos

bien separados dentro del grupo de los habitantes de charcos. Todas las especies

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 137

mostraron proporciones corporales del tipo de un renacuajo generalista, häbitos

macréfagos y respiraciôn tanto branquial como pulmonar. Los renacuajos que viven en

fitotelmata fueron caracterizados por la reducciôn de las estructuras peribucales y

bucofaringeas, y difirieron marcadamente en la morfologia del pico corneo y en el

desarrello de los pulmones. Los renacuajos que se desarrollan en charcos fueron

caracterizados por un alto nûmero de filas de denticulos y por valores de balance bajos.

Los géneros Phrynohyas y Osteocephalus se distinguieron mejor por el desarrollo de filas

de denticulos superiores secundarias, por la posiciôn de los ojos y por las proporciones

corporales groseras.

ACKNOWLEDGEMENTS

We are indebted to the University of Vienna for funding this research (Scientific Exchange

Program between the University of Säo Paulo and the Faculty of Natural Sciences, University of

Vienna) as well as to the support of the University of Veterinary Medicine of Vienna. W. HôDL

initiated and promoted our cooperation. M. GoRDo helped in the collection of P. resinifictrix larvae,

T. LoserT in SEM techniques, K. REPP in preparation of drawings, and R. WYTEK in statistical

analyses. G. SkUK translated the Spanish resumen. H. GRiLLITSCH, A. C. MARQUES and H. L.

NEMESCHKAL provided discussion and valuable advice. L. C. SCHIESARI expresses his gratitude to Eva

and Angelika RIEDER for hospitality in Vienna. We graciously acknowledge R. WASSERSUG for review

and constructive advice.

LITERATURE CITED

ALTIG, R. & JOHNSTON, G. F., 1986. — Major characteristics of free-living anuran tadpoles. Smithson.

herpet. Inf. Serv., 67: 1-15.

me 1989. — Guilds of anuran larvae: relationships among developmental modes, morphologies, and

habitats. Herper. Monogr., 3: 81-109.

BOKERMAN, W. C. A., 1963. — Girinos de anfibios brasileiros. 1 (Amphibia - Salientia). An. Acad.

brasil. Ciéne., 35: 465-474.

CaALDWELL, J. P., 1989. — Structure and behavior of Hyla geographica tadpole schools, with

comments on classification of group behavior in tadpoles. Copeia, 1989: 938-950.

Dumois, A., 1995. — Keratodont formulae in anuran tadpoles: proposals for a standardization. J.

zool. Syst. evol. Res., 33: I-XV.

DuELLMAN, W. E., 1970. — The hylid frogs of Middle America. Monogr. Mus. nat. Hist. Univ.

Kansas, 1 (1): i-xi + 1-427; 1 (2): 429-753, pl. 1-72.

Le 1974. — À reassessment of the taxonomic status of some neotropical hylid frogs. Occas. Pap.

Mus. nat. Hist. Univ. Kansas, 27: 1-27.

ee 1978. — The biology of an equatorial herpetofauna in Amazonian Ecuador. Univ. Kansas Mus.

nat. Hist. mise. Publ., 65: 1-352, pl. 1-4.

ee 1988. — Patterns of species diversity in anuran amphibians in the American tropics. Ann.

Missouri bot. Gard., 75: 79-104.

DuELLMAN, W. E. & LESCURE, J., 1973. — Life history and ecology of the hylid frog Osteocephalus

taurinus, with observations on larval behavior. Occas. Pap. Mus. nat. Hist. Univ. Kansas, 13:

1-12.

DuELLMAN, W. E. & TRUEB, L., 1985. — Biology of amphibians. New York, McGraw-Hill: i-xix +

1-670.

Source : MNHN, Paris

138 ALYTES 13 (4)

Fox, H., 1984. — Amphibian morphogenesis. Clifton N.J., Humana Press: i-xv + 1-301.

Frosr, D. R. (ed.), 1985. — Amphibian species of the world. Lawrence, Allen Press & Assoc. Syst.

Coll: [i-iv] + i-v + 1-732.

Gascon, C., 1991. — Population- and community-level analyses of species occurrences of central

Amazonian rainforest tadpoles. Ecology, 72: 1731-1746.

GascoN, K. L., 1993. — Breeding-habitat use by five Amazonian frogs at forest edge. Biodivers.

Conserv., 2: 438-444.

Gosner, K. L., 1959, — Systematic variations in tadpole teeth with notes on food. Herpetologica, 15:

203-210.

en 1960. — A simplified table for staging anuran embryos and larvae, with notes on identification.

Herpetologica, 16: 183-190.

Goup, S. J., 1966. — Allometry and size in ontogeny and phylogeny. Biol. Rev., 41: 587-640.

GouLp, S. J. & VrBA, E.S., 1982. — Exaptation - a missing term in the science of form. Paleobiology,

8: 4-15.

GrizciTsCH, B., 1992. — Notes on the tadpole of Phrynohyas resinifictrix (Goeldi, 1907).

Buccopharyngeal and external morphology of a tree hole dwelling larva (Anura: Hylidae).

Herpetozoa, 5: 51-66.

GRILLITSCH, B., GRILLITSCH, H., DUBOIS, A. & SPLECHTNA, H., 1993. — The tadpoles of the brown

frogs Rana [graeca] graeca and Rana [graeca] italica (Amphibia, Anura). Alytes, 11: 117-139.

HENLE, K,, 1981. — Hyla elkejungingerae, ein neuer Hylide aus dem peruanischen Regenwald

(Amphibia: Salientia: Hylidae). Amphibia-Reptilia, 2: 123-132.

æe 1992. — Zur Amphibienfauna Perus nebst Beschreibung eines neuen Eleutherodactylus

(Leptodactylidae). Bonn. zool. Beitr., 43: 79-129.

HENLE, K., EHRL, A. & PILGRAM, C., 1983. — Zum Biotop und zur Aufzucht des peruanischen

Laubfrosches Hyla elkejungingerae Henle 1981. Herpetofauna, 24: 8-9.

HERO, J.-M., 1990. — An illustrated key to the tadpoles occurring in the central Amazon rainforest,

Manaus, Amazonas, Brasil. Amazoniana, 11: 201-262.

HÉRON-ROYER & VAN BAMBEKE, C., 1889. — Le vestibule de la bouche chez les têtards des batraciens

anoures d'Europe; sa structure, ses caractères chez les diverses espèces. Arch. Biol., 9: 185-309,

pl. 12-24.

HôoL, W., 1990. — Reproductive diversity in Amazonian lowland frogs. In: W. HANKE (ed.), Biology

and physiology of amphibians, Fortschr. Zool., 38: 41-60.

ES 1993. — Amazonien aus der Froschperspektive. Zur Biologie der Frôsche und Krôten des

Amazonastieflandes. Kataloge OÙ. Landesmus., (N. F.), 61: 499-545.

INGER, R. F., 1985. — Tadpoles of the forested regions of Borneo. Fieldiana: Zool., (n.s.), 26: i-v +

1-89.

JOHNSON, R. A. & WICHERN, D. W., 1988. — Applied multivariate statistical analysis. London,

Prentice-Hall: i-xii + 1-640.

JUNGFER, K.-H. & SCHIESARI, L. C., 1995. — Description of a new central Amazonian and Guianan

treefrog, genus Osteocephalus (Anura, Hylidae), with oophagous tadpoles. Alytes, 13: 1-13.

JUNGFER, K.-H. & WEYGOLDT, P., 1995. — Reproductive biology of the central Amazonian frog

Osteocephalus oophagus. Abstracts 8th O.G.M. Societas Europaea Herpetologica, Bonn, 1995:

69.

KAUNG, H. L. C. & KoLLRoOS, J. J., 1976. — Cell turnover in the beak of Rana pipiens. Anat. Rec.

188: 361-370.

LANNOO, M. J., TOWNSEND, D. S. & WASSERSUG, R. J., 1987. — Larval life in the leaves: arboreal

tadpole types, with special attention to the morphology, ecology, and behavior of the

oophagous Osteopilus brunneus (Hylidae) larva. Fieldiana, Zool., (n. s.), 38: 1-31.

Lurz, B., 1973. — Brazilian species of Hyla. Austin & London, University of Texas Press: i-xix +

1-265, pl. 1-7.

MCcDraRMDD, R. & ALTIG, 1990. — Description of a bufonid and two hylid tadpoles from western

Ecuador. Alytes, 8: 51-60.

MoRRIsON, D. F., 1990. — Multivariate statistical methods. New York, McGraw-Hill: i-xvii + 1-493.

None, G. K., 1929. — The adaptive modifications of the arboreal tadpoles of Hoplophryne and the

torrent tadpoles of Staurois. Bull. amer. Mus. nat. Hist., 58: 291-334.

Source : MNHN, Paris

SCHIESARI, GRILLITSCH & VOGL 139

Pysur, W. F., 1967. — Breeding and larval development of the hylid frog Phrynohyas spilomma in

southern Veracruz, Mexico. Herpetologica, 23: 184-194.

RADA DE MARTINEZ, D., 1990. — Contribucin al conocimiento de las larvas de anfibios de

Venezuela. Mem. Soc. Cienc. nat. La Salle, 50: 391-403.

RoDRIGUEZ, L. O. & DUELLMAN, W. E., 1994. — Guide to the frogs of the Iquitos region, Amazonian

Peru. Univ. Kansas nat. Hist. Mus. Special Publ., 22: 1-80.

SazIMA, L., 1974. — An albino hylid frog, Phrynohyas mesophaea (Hensel). J. Herpet., 8: 264-265.

ScHirsari, L. C., 1993. — Estudo da morfologia da larva de Phrynohyas resinifictrix (Anura,

Hylidae). 3° Congr. Lat.-Amer. Herpet., 1993: 144.

Scmesart, L. C. & Gorpo, M., 1993. — Aspectos da histéria natural do Canauarü, Phrynohyas

resinifictrix (Anura, Hylidae). 3° Congr. Lat.-Amer. Herpet., 1993: 143.

ScHiesari, L. & MoREIRA, G., in press. — The tadpole of Phrynohyas coriacea (Peters, 1867) (Anura:

Hylidae), with comments on the species’ reproduction. J. Herpet., in press.

SHAW, R. G., 1987. — Maximum-likelihood approaches applied to quantitative genetics of natural

populations. Evolution, 41: 812-826.

SwaLLOW, W. H. & MONAHAN, J. F., 1984. — Monte Carlo comparison of ANOVA, MIVQUE,

REML, ML estimators of variance components. Technometrics, 26: 47-57.

TRUEB, L., 1970. — Evolutionary relationships of casque-headed treefrogs with co-ossified skulls

(family Hylidae). Univ. Kansas Publ. Mus. nat. Hist., 18: 547-716.

pr 1973. — Bones, frogs, and evolution. In: J. L. ViaL (ed.), Evolutionary biology of the anurans,

Columbia, Univ. Missouri Press: 65-132.

Trus8, L. & DUELLMAN, W. E., 1970. — The systematic status and life history of Hyla verrucigera

Werner. Copeia, 1970: 601-610.

= 1971. — A synopsis of neotropical hylid frogs, genus Osteocephalus. Occas. Pap. Mus. nat. Hist.

Univ. Kansas, 1: 1-47.

VIERTEL, B., 1982. — The oral cavities of central European anuran larvae (Amphibia). Morphology,

ontogenesis and generic diagnosis. Amphibia-Reptilia, 4: 321-360.

WASsERSUG, R., 1976. — Oral morphology of anuran larvae: terminology and general description.

Occas. Pap. Mus. nat. Hist. Univ. Kansas, 48: 1-23.

- 1980. — Internal oral features of larvae from eight anuran families: functional, systematic,

evolutionary and ecological considerations. Univ. Kansas Mus. nat. Hist. mise. Publ., 68: 1-146.

WASSERSUG, R. J., FROGNER, K. J. & INGER, R. F., 1981. — Adaptations to life in tree holes by

rhacophorid tadpoles from Thailand. J. Herpet., 15: 41-52.

WASsERSUG, R. & HEYER, W. R., 1988. — A survey of internal oral features of leptodactyloid larvae

(Amphibia: Anura). Smithson. Contrib. Zool., 457: i-iv + 1-99.

WASsERSUG, R. J. & MUrPHY, A. M., 1987. — Aerial respiration facilitates growth in suspension-

feeding anuran larvae (Yenopus laevis). Exper. Biol., 46: 141-147.

WaASsERSUG, R. J. & SEIBERT, E. A., 1975. — Behavioral responses of amphibian larvae to variation

in dissolved oxygen. Copeia, 1975: 86-103.

ZIMMERMAN, B. L. & HôDz, W., 1983. — Distinction of Phrynohyas resinifictrix (Goeldi, 1907) from

Phrynohyas venulosa (Laurenti, 1768) based on acoustical and behavioural parameters

(Amphibia, Anura, Hylidae). Zool. Anz., 1983: 341-352.

ZIMMERMAN, B. L. & RODRIGUES, M. T., 1990. — Frogs, snakes, and lizards of the INPA-WWF

reserves near Manaus, Brazil. Jn: A. H. GENTRY (ed.), Four neotropical rainforests, New Haven

& London, Yale University Press: 426-454.

ZweireL, R. G., 1964. — Life history of Phrynohyas venulosa (Salientia: Hylidae) in Panamä. Copeia,

1964: 201-208.

Corresponding editor: Alain DuBoIs.

Source : MNHN, Paris

Alytes, 1996, 13 (4): 140. Announcements

Prize for best student paper in Alytes

ISSCA is glad to announce the creation of a yearly prize for the best paper published

by a student in A/ytes. The winner of this prize will be offered a complete collection of

Alytes from volume 1 (1982) to the volume containing the prizewinning paper.

To be eligible for this prize, a paper should be authored by a student alone or by

several students, but should not be a multi-author paper by student(s) and non-student(s).

Studentship should be attested by an administrative document or by a letter from the

student’s professor or supervisor. Candidacy to this prize should be indicated by the

student when returning the corrected proofs of her/his accepted paper to the chief-editor

of Alytes. The choice of the winning paper will be made by the ISSCA Board among

papers published by students in the preceding year’s volume of Alytes. The first prize will

be awarded at the beginning of 1997.

Free collections of Alvtes for life subscribers

From 1996 on, any new life or patron individual subscriber to A/ytes will be offered

a free complete collection of back issues of Alytes from volume 1 (1982) to the start of

her/his subscription.

Source : MNHN, Paris

Alytes, 1996, 13 (4): 141-166. 141

Systematics, morphometrics

and biogeography of the genus Aubria

(Ranidae, Pyxicephalinae)

Annemarie OHLER

Laboratoire des Reptiles et Amphibiens, Muséum national d'Histoire naturelle,

25 rue Cuvier, 75005 Paris, France

New data are given on geographic distribution and morphometric variation

with reference to geographic origin and sex in Aubria subsigillata (Duméril,

1856) and Aubria masako Ohler & Kazadi, 1990. The recently described

Aubria occidentalis Perret, 1995 is here considered as a synonym of the

former. Different statistical approaches are used to describe the taxa and the

observed variations. À hypothesis of femoral gland function is formulated in

the genera Aubria and Pyxicephalus. À phylogeny of the group Aubria-

Pyxicephalus (Pyxicephalinae) is discussed.

INTRODUCTION

Naturalists conducting systematics and biogeographical surveys often encounter two

patterns of distribution. A species can have a distribution-limited area, with little

morphological variation, or a very wide distribution, within which one frequently observes

significant variation that may lead to more than one interpretation. Often discrimination

of distinct units is not discrete, and taxonomic decision difficult if collection localities are

distant from each other. Even when we are conscious of variation problems in biology and

use appropriate sampling methods, old samples generally include too few specimens for

statistical comparison.

A recent study of frogs of the ranid genus Aubria (OHLER & KAZADI, 1990) led us

to distinguish a new species (Aubria masako) from the central region of Zaire, North of

the Congo. This species is clearly distinct from the type-specimen of Rana subsigillata from

Gabon (DUMÉRIL, 1856). Specimens from 17 other localities of western Africa were

referred to the species Aubria subsigillata by use of morphometrical (internarial distance,

snout-nostril distance, diameter of tympanum) and morphological (presence of femoral

glands, presence of mid-dorsal stripe) data. However, it was not possible to assign three

frogs in the Paris Museum collection from the Sangha region, because they showed

characters which distinguished them from our samples. The collection of further data was

needed to resolve the question of their relationships.

Another problem presented by the literature was the presence of femoral glands in

some of the frogs of populations without mid-dorsal stripes and their absence in others (DE

Source : MNHN, Paris

142 ALYTES 13 (4)

WITTE, 1930; PARKER, 1936; PERRET, 1966). PERRET (1966) had indicated that variation

exists in this character among the frogs of Cameroon, and recently (1995) he discussed

variation of this character in relation to sex. Finally, according to AMIET (personal

communication), in Cameroon there are two different mating calls in Aubria, character-

izing two kinds of frogs that are also distinct in distribution and ethology.

Recently, I collected important additional information on frogs of the genus Aubria.

This gives a more detailed knowledge of the distribution of these frogs, contributes to the

biogeography of Africa, and to the study of morphometrical variation in the taxa involved.

Discussion of homology of femoral glands in this genus might give some interesting clues

for use of these glands as characters in phylogenetic analysis.

In a recent paper, published after this work was completed and submitted, PERRET

(1995) described a new species, Aubria occidentalis, from West Africa (type-locality in

Ivory Coast). Its synonymy with Aubria subsigillata will be discussed under this species.

MATERIAL AND METHODS

ABBREVIATIONS

The names of the collections where the studied specimens are deposited are

abbreviated as follows: BMNH, British Museum (Natural History), London; CAS,

California Academy of Sciences, San Francisco; K, KAazapi Mpetemba collection;

KMMA, Koeninglijk Museum voor Midden-Afrika, Tervuren; MHNG, Muséum d’His-

toire naturelle de Genève; MNHN, Muséum national d'Histoire naturelle, Paris; MRHN,

Institut Royal des Sciences naturelles de Belgique, Brussels; NMW, Naturhistorisches

Museum, Wien; ZMB, Zoologisches Museum, Berlin, ZMH, Zoologisches Museum,

Hamburg.

For abbreviations of measurements, see Table I.

SPECIMENS STUDIED

For this study I measured 117 adult and juvenile frogs from 41 different collecting

sites. Of these, samples from 31 sites contain adult specimens (a total of 65 individuals),

including 21 localities with one specimen, 5 with two, 4 with three to eight and one locality

with 15 adult specimens. As my data were collected over a five-year period, in the

beginning the list of measurements was not exhaustive and consequently some individuals

had to be excluded from statistical analysis.

COMPOSITION OF SAMPLES

The unit used for morphometrical analysis (sample) is a group of individuals of about

the same age, of the same sex and of the same locality. For this purpose three age-groups

Source : MNHN, Paris

OHLER 143

Table I. - List of abbreviations for measurements.

Abbreviations Descriptions of measurements

Eye length

Distance from front of eye to nostril

Distance from maximum incurvation of web between fourth and

fifth toe to tip of fourth toe

Femoral length (from sagittal axis of body to knee)

Fore limb length (from elbow to base of outer palmar tubercle)

Foot length (from base of inner metatarsal tubercle to tip of toe)

Fourth toe length (from base of proximal subarticular tubercle)

Greatest diameter of femoral gland

Distance of femoral gland to sagittal axis of body

Hand length (from base of outer palmar tubercle to tip of finger)

Head length (from back of mandible to tip of snout)

Head width

Distance between backs of eyes

Distance between fronts of eyes

Length of inner metatarsal tubercle

Internarial distance

Inner toe length from distal edge of inner metatarsal tubercle

Distance from back of mandible to back of eye

Distance from back of mandible to front of eye

Distance from back of mandible to nostril

Distance from distal edge of metatarsal tubercle to maximum

incurvation of web between fourth and fifth toe

Distance from distal edge of metatarsal tubercle to maximum

incurvation of web between third and fourth toe

Distance from nostril to tip of snout

Snout-vent length

Third finger length (from base of proximal subarticular tubercle)

Distance from maximum incurvation of web between third and

fourth toe to tip of fourth toe

Tibia length

Greatest tympanum diameter

Distance from tympanum to back of eye

Source : MNHN, Paris

144 ALYTES 13 (4)

were recognized: larvae, with a tail; juvenile frogs, from metamorphosis to sexual maturity;

and adult frogs, including females with ripe ovaries and oviducts and males with well

developed testes (DuBois, 1976). Secondary sexual characters are absent in adult Aubria,

so it is difficult to determine adult males based on external and internal observation. But

morphometrical analysis showed an important allometric difference between subadult and

adult males that allows staging of the specimens. Subadult males, when analysed using the

Laurent’s technique (described below), cluster in a group of their own, independent of

species membership, as do females (results not shown).

The best way to identify interpopulational variation is to treat the different samples

separately. Only some samples have enough specimens for such comparisons. As sample

sizes are usually too small for statistical analysis, I also grouped the frogs on a

morphological basis, and compared the homogeneity of the two species.

In order to optimize analysis and to find an intermediate way between analysis of

single population samples and species comparisons, I grouped specimens from populations

of homogeneous biogeographic areas (WÜSsTER & THORPE, 1992). The areas were defined

following the studies of ScHioTz (1967) and HAMILTON (1988) who have discussed possible

biogeographic zones. The following areas were used (fig. 1): (1) Sassandra Valley, Ivory

Coast; (2) Ghana; (3) Kovié, Togo; (4) Nigeria, West of Niger; (5) Nigeria, between Niger

and Cross Rivers; (6) lowlands of Guinean Coast south of Cross River; (7) Sangha region

at the intersect of Cameroon, Gabon and Zaire; (8) western part of Zaire; (9) eastern part

of Zaire (this latter division is possibly due to a discontinuity in collecting).

MEASUREMENTS AND TRANSFORMATIONS

Measurements were taken with slide-calipers and binocular microscope. The precision

varies from 0.1 % to 1 % of the measurement; for distances smaller than 3 mm, precision

is only 1 % to 3 %.

Snout-vent length is given in millimetres. The other measurements are expressed as

thousandths (%o) of snout-vent length:

RX = (X/SVL) x 1000.

The position of the femoral gland on the thigh is expressed as the ratio of “femur”

length (FL) to distance of gland from sagittal axis: FLGLDT = (FL / GLDT) x 1000.

Webbing is given as the ratio (TFTFOL) of webbing from the tip of the fourth toe to the

incurvation between the third and fourth toe (TFTF) divided by the length of the foot

(FOL), or as the ratio (MTFFOL) of webbing from the distal border of the metatarsal

tubercle to the incurvation between the third and fourth toe (MTFF) divided by the length

of the foot (FOL).

The logarithm of the measurement was used in discriminant analysis:

X = InX.

Source : MNHN, Paris

WATHO

Fig. 1. — Distribution of Aubria subsigillata and Aubria masako in relation to vegetation types of tropical Africa. Black circle, Aubria subsigillata;

white star, type-locality of Rana subsigillata A. Duméril, 1856; white circle, type-locality of Phrynopsis ventrimaculata Nieden, 1908; black

triangle, type-locality of Aubria occidentalis Perret, 1995; black square, Aubria masako; black star, type-locality of Aubria masako Ohler &

Kazadi, 1990. Forest types: À, Sudanian woodland; B, lowland rain forest; C, mangrove; D, Afromontane vegetation; E, swamp forest; F,

mosaic of swamp forest and wetter lowland Guineo-Congolian rain forest; G, wetter lowland rain forest; H, drier Guineo-Congolian rain

forest; I, mosaic of G and H. Biogeographical areas: 1, Ivory Coast; 2, Ghana; 3, Togo; 4, western Nigeria; 5, eastern Nigeria; 6, coastal area

of Cameroon and equatorial Guinea; 7, Sangha region; 8, western Zaire basin; 9, eastern Zaire basin.

Source : MNHN, Paris

146 ALYTES 13 (4)

To calculate Laurent’s distance, SVL was used as a size factor and all measurements

were corrected as follows:

IX = [InX — InSVL].

All these corrections were used to standardize for isometric size changes.

STATISTICS

Mean, standard deviation, minimum and maximum were calculated for all variables

of all groups on a personal computer using the SPSS program (Norusis, 1992). To

estimate homogeneity, Haldane’s coefficient of variation was caculated (DELAUGERRE &

Dugois, 1985).

Non-parametric statistics permit comparison of samples of relatively small sizes

(Dusois, 1984; DELAUGERRE & DuBois, 1985) and are therefore appropriate to use on

samples from old zoological collections. Standard non-parametric statistics (ZAR, 1984)

were applied to compare different populations. The Kruskal-Wallis test was used on all 35

measurements to measure group homogeneity. Those measurements which showed

significant differences were subsequently treated by Tukey type B tests (on ranks of

variables). The Mann-Whitney U test was used for comparison between any two groups.

LAURENT (1955, 1981) developed a method that uses distance measurements to

compute phenograms. Cluster analysis using Laurent’s procedure was made for adult

males including the 35 measurements (for female frogs the sample size was too small to

perform an analysis) on SPSS using Manhattan distance dissimilarities matrix and Ward’s

method (Norusis, 1992).

To investigate variables that distinguish the two morphologically recognized species,

I performed a discriminant analysis. In the first analysis, all log-transformed measure-

ments were included. In a further step, five variables with good discrimination

characteristics (small in-group variation indicated by a relatively small Wilk’s lambda)

were retained to assign cases to their group based on a few measurements.

DESCRIPTIVE METHODS

The webbing formula of Myers & DUELLMAN (1982) was used. Tadpoles were staged

following Gosner (1960). Keratodont formulae (DuBois, 1995) are given according to

ANNANDALE’S (1912) system.

RESULTS

The genus Aubria is widely distributed in western and central Africa (fig. 1). Its

distribution includes forested areas of the following countries: Guinea, Ivory Coast,

Ghana, Togo, Nigeria, Cameroon, Congo, Republic of Central Africa, Zaire, Equatorial

Guinea, Gabon. In life these are very colourful frogs with a typical yellow and violet

Source : MNHN, Paris

OHLER 147

ventral pattern (PERRET, 1966: Cameroon; PERRET, 1995: West Africa; OHLER & KAZADI,

1989: Zaire; LAURENT, personal communication, letter of 22 June 1994: Zaire). The frogs

of the different regions have a superficial resemblance — they were referred to one species

from the description of Rana subsigillata in 1856 to 1990 — but detailed morphological

studies show extensive interpopulational variation.

External morphology clearly separates the frogs into two groups: one with

midfemoral glands and the other with close-to-knee femoral glands (see also PERRET,

1995). In a previous study (OHLER & KAZADI, 1990), because of an insufficient number

of specimens, we did not realize the presence of near-to-knee femoral glands in Aubria

masako. The character of gland position is correlated with dorsal colour and ontogenetical

variation of ventral pattern.

The morphometrical homogeneity of morphologically determined groups was tested

using the Kruskal-Wallis test. The test showed no significant differences at the P = 0.05

level for samples of adult males and females of Aubria subsigillata for the following

measurements and ratios: SVL, RHW, RHL, RTL, RIN, RTYD, RTFL, LFGLDT,

MTFFOL. Male and female samples of Aubria masako were tested for the same

measurements. Only the samples of males showed significant differences in the ratio RIN

G2 = 114, DF = 5; P = 0.04 *).

COMPARISON OF SPECIES

Univariate comparisons

Interspecies comparisons were performed separately for males and females. The

observed differences were the same as found in our previous study (OHLER & KAZADI,

1990), but treating the sexes separately underlines the differences observed. The

morphometrical differences concern the head, the femoral gland position and the webbing.

Comparison of Aubria subsigillata and Aubria masako by means of Mann-Whitney U

tests shows that there are statistically significant differences in both sexes in the ratios to

SVL of several measurements (Table II). Aubria subsigillata is larger and has a longer tibia

than Aubria masako. The webbing of its foot is more extensive. Differences exist between

males and females in interspecific differentiation. In female Aubria, the Mann-Whitney U

tests show no significant differences for several ratios concerning head (RHW, RHL,

RTYD) and webbing (TFTFOL) that show significant differences in males.

Discriminant analysis

Using only five morphometric variables, adult specimens of Aubria can be attributed

to one of the two species recognized (Table III). This permits also to classify the holotype

of Aubria occidentalis in the Aubria subsigillata group. The variables are position and size

of femoral gland, tibia length, metatarsal tubercle, distance between eyes and size of

tympanum. They describe various aspects of the body form. Other variables were excluded

because of missing values or of great variability of the measurements.

Source : MNHN, Paris

148

ALYTES 13 (4)

Table II. - Species differences between Aubria subsigillata and Aubria masako, all

localities pooled, males and females treated separately. N, number; SD,

standard deviation; V, coefficient of variation, U, Mann-Whitney U; P,

probability; SUB, À. subsigillata; MAS, À. masako; for other abbreviations,

see Table I and text.

—— @ —

Measurement | _ Species N | Mean | SD v u P

Males

SVL SUB 17 79.1 4.1 5:31 16.0 |0.000***

MAS 12 71.4 4.5 6.43

RHW SUB 17 349 11.63| 3.37 40.0 |0.006**

MAS 12 363 16.3 | 4.58

RHL SUB 17 412 11.9 | 2.93 50.0 |0.021*

MAS 12 428 20.7 | 4.94

RTL SUB 17 388 14.1 3.68 30.0 |0.001**

MAS 12 368 16.0 | 4.26

RIN SUB 17 66 3.8 5.84 8.0 |0.000***

MAS il 60 1.7 | 2.89

RTYD SUB 17 73 5.9 | 8.20 33.0 |0.002**

MAS 12 83 8.2 |10.09

LFGLDT SUB 13 370 45.7 |12.59 0.0 |0.001***

MAS 6 516 21.3 | 4.30

RMTTF SUB 15 242 15.3 | 6.42 21.0 |0.006**

MAS 9 222 14.7 | 6.80

TFTFOL SUB 17 445 38.8 8.85 42.0 |0.008**

MAS 12 480 27.0 | 5.75

Females

SVL SUB 17 86.0! 4.1 4.84 28.0 |0.004**

MAS 10 78.4 6.0 7.84

RHW SUB 17 348 9.3 | 2.71 77.0 |0.688 ns

MAS 10 344 21.0 | 6.26

RHL SUB 17 403 13.9 | 3.50 68.0 |0.393 ns

MAS 10 399 16.1 4.14

RTL SUB 17 385 17.4 4.59 6.0 |0.000***

MAS 8 352 18.1 5.30

RIN SUB 16 66 4.3 | 6.62 13.0 |0.001***

MAS 9 58 3.6 | 6.38

RTYD SUB 17 74 ST. 7.81 47.0 |0.056 ns

MAS 10 79 7.0 | 9.08

LFGLDT SUB 13 334 55.8 |17.03 0.0 |0.000***

MAS 8 487 53.4 |11.31

RMTTF SUB 15 230 18.8 8.31 3.0 |0.003**

MAS 5 196 9.72| 5.21

TFTFOL SUB 16 450 37.5 8.46 61.0 |0.317ns

MAS 10 456 28.0 | 6.29

Source : MNHN, Paris

OHLER 149

Table HI. Use of discriminant function to distinguish Aubria subsigillata and Aubria

masako (adult males and females together) based on 5 morphometric

parameters (see "Material and methods").

A. Statistical significance

Canonical Chi-square Degrees of

correlation freedom

Eigenvalue

3.8584 0.8912 68.76 5

B. Standardized canonical discriminant function coefficients.

Morphometric character Function 1

Tibia length 0.69064

Tympanum diameter - 0.34850

Distance of femoral gland - 0.45144

Diameter of femoral gland - 0.38166

Distance between posterior part of eyes 0.60099

C. Classification success

Predicted

Actual group em

Aubria subsigillata | Aubria masako

Aubria subsigillata 30 (100%) 0

Aubria masako 0 15 (100%)

Aubria occidentalis holotype (ungrouped) Î 0

Source : MNHN, Paris

150 ALYTES 13 (4)

Laurent's morphometric distance

Measurements of males from all localities were included in LAURENT’s (1955, 1981)

procedure, which separated the two species very clearly (fig. 2). One group includes the

type-specimens of R. subsigillata and À. occidentalis and corresponds to the morphologi-

cally determined 4. subsigillata. The second group includes specimens from Masako,

Boteke and Sangha and the type series of 4. masako.

COMPARISON OF POPULATIONS

The inter-group differences that are significant at the P < 0.05 level obtained by the

Tukey test are shown in fig. 3 for males and in fig. 4 for females.

In males, the differences between OTUSs 2, 3, 4, 6 and OTUSs 7, 8, 9 appear clearly.

In females, where the number of specimens is very small even when the populations are

grouped, the relations between the OTUSs appear slightly different. The population 3 of 4.

subsigillata from Togo seems to be morphologically closest to 4. masako populations 7,

8, 9. Comparison of the specimens from population 7 (Sangha region) with populations

from western Zaire (8) and eastern Zaire (9) shows that they should be included in 4.

masako.

DETAILED ACCOUNT OF SPECIES

Aubria Boulenger, 1917

Aubria Boulenger, 1917: 988. — Type species by monotypy: Rana subsigillata Duméril, 1856.

The types of Aubria subsigillata and of Aubria masako have been described in detail

(DumériL, 1861: 224, pl. XVIII fig. 1; OHLER & KAZADI, 1990). Here I will give a

description of the variation and of sexual dimorphism within populations, descriptions of

tadpoles and young frogs, and diagnoses for the two species.

Aubria subsigillata (A. Duméril, 1856)

Rana subsigillata A. Duméril, 1856: 560. — Holotype: MNHN 1566. — Type-locality: Gabon.

[Rana (Aubria) subsigillatay: BOULENGER, 1917: 988.

Aubria subsigillata: LAURENT, 1953: 27.

Phrynopsis ventrimaculata Nieden, 1908: 499. — Holotype: ZMB 20134. — Type-locality: Longj,

Cameroon. — Synonymy fide OHLER & KAZADI, 1990.

Aubria occidentalis Perret, 1995: 258. — Holotype: MHNG 2129.17. — Type-locality: Banco

forest reserve, Ivory Coast. — New synonymy.

Specimens studied. — GABON. MNHN 1566, holotype of Rana subsigillata A. Duméril,

1856; Biligone River: MNHN 1974.1130; 50 km SW of Lambaréné: MNHN 1901.564. —

EQUATORIAL GUINEA. Benito River: ZMH A.03133. — CAMEROON. Yabassi District (4°N,

10°E): BMNH 1938.6.10.9; SW-Province, “Korup”: BMNH 1982.746; Yaoundé Road, 4

Source : MNHN, Paris

Rescaled distance cluster combine

5 10 15 20 25

SK1

MCI

Fig. 2. — Phenogram of 21 adult males belonging to the species Aubria subsigillata and Aubria

masako: Laurent’s distances based on 35 measurements were included in distance calculations,

using Ward’s method. S, A. subsigillata: K, Kovié, Togo: 1, MNHN 1989.2054; 2, MNHN

1993.1462; 3, MNHN 1989.2056; 4, MNHN 1993.1966; 5, MNHN 1993.1469; 6, MNHN

1989.2050; 7, MNHN 1989.2053; I, Ibadan Swamp, Nigeria: 1, BMNH 1964.237; L, Lambarene,

Gabon: 1, MNHN 1901.564; O, Obuasi, Ghana: 1, BMHN 1917.4.13.13; G, Gabon: 1, MNHN

1566, holotype of R. subsigillata (black cercle); P, Port Harcourt, Nigeria: 1, BMNH

1956.1.10.84; N, Banco forest, Ivory Coast: 1, MHNG 2129.17, holotype of 4. occidentalis (black

square). M, 4. masako: K, Kisangani, Zaire: 1, MNHN 1989.2775, holotype of À. masako (black

triangle); 2, MNHN 1989.3305, paratype; B, Boteke, Zaire: 1, KMMA 85.30.B.368; 2, KMMA

85.30.B.362; S, Sangha, Congo: 1, MNHN 1993.2831; 2, MNHN 1989.2830; M, Mabali, Zaire:

1, CAS 145297; C, Coquilhatville, Zaire: 1, CAS 113967.

Source : MNHN, Paris

152

ALYTES 13 (4)

23 4 6 7 8 9 23 4 6 7 8 9 23 4 6 7 8 9

2 *x[*]S 2 *x]H 2 !

3 #fll Vs [ #]Ls *[ [118

4 * te 4 * E

6 AE 6 6 *

RE 7 JEEIEIE

8 + [#8 8 8

ORE *k ORE [ EFIEIE

HW MN IN

2.3 4 6 7 8 » 2 3 4 6 7 8 9

2 [js * US: M L

s[* NUS x| [+ T, F

411] D: E 5

6 6 o D

7 %, * AE T

ORE 8

9, OREIE cs

ITL TFTFOL

Fig. 3. — Significant differences (P < 0.05) of measurements among males of Aubria, grouped

populations of biogeographically homogeneous areas, using the Mann-Whitney U test. 2, Ghana;

3, Togo: 4, western Nigeria; 6, coastal area of Cameroon and equatorial Guinea; 7, Sangha

region; 8, western Zaire basin; 9, eastern Zaire basin. For key to measurement abbreviations see

Table I.

12346789 1234 6789 123 4 6 7 8 9

Ho Jet

1 *

LE 4 | M 9 ;

3 x OME GE 3

4 4 4

6 6 «| [|

4. GORE 7 * L)

Û AE Û # OAAAEE

, , , LL

TL EN IBE

123 4 6 7 8 9 12346789

ds Fi ’

METEO

3 AE s[x JE

4 n =

‘ 0E L

7 DE 7 ARE

8 GE û

Û GE , ARE

LFGLOT MTFFOL

Fig. 4. — Significant differences (P_< 0.05) of measurements among females of Aubria, grouped

populations of biogeographically homogeneous areas, using Mann-Whitney U test. 1, Ivory

abbreviations see Table I.

2, Ghana; 3, Togo; 4, western Nigeria; 6, coastal area of Cameroon and equatorial

, Sangha region; 8, western Zaire basin; 9, eastern Zaire basin. For key to measurement

Source : MNHN, Paris

OHLER 153

mi E of Douala: CAS 103804. — NiceriA. NMW 2552; Calabar, Edge of Great Kwa

River: BMNH 1980.1275; Calabar, 20 km north on MCC Road: BMNH 1980.1276;

Calabar, 15 km north on MCC Road: BMNH 1980.1277; Ibadan: BMNH 1969.3000;

Ibadan Swamp: BMNH 1964.23ÿ; Ijebu Ode: BMNH 1969.2999; Lagos: NMW 2555; Port

Harcourt: BMNH 1956.1.10.84. — GHANA. Eastern Region, Tafo Cocoa Research

Institute: CAS 146050-51, CAS 144214-215; Eastern Region, Kadé, Agricultural Research

Station of University of Ghana: CAS 103783, CAS 103818, CAS 103836, CAS 104016,

CAS 104051, CAS 125534; South Ashantee, Obuasi: BMNH 1917.4.13.12-13; Western

Region, “Wasaw Akropong” (north of Tarkwa): CAS 97967. — Too. Kovié: MNHN

1989.2047-2056 and 1993.1463-1472; Mission Tové: MNHN 1989.4094. — GuinEA. Diéké:

MNHN 1920.147. — Ivory CoasT. Daloi-Lobo: MNHN 1993.929-930 (formerly A.929

and A.930); Sassandra: MNHN 1993.2832-2834; Soubré: MNHN 1989.4095; Banco forest

reserve: MHNG 2129.17, holotype of Aubria occidentalis Perret, 1995.

Diagnostic characters. — Aubria subsigillata has femoral glands in a position midway from

knee to vent in a ventral position on the thigh (fig. 5). The femoral gland is rounded. Feet

are more webbed (I 1 — 2 11 1 — 2 III 1 — 2.5 IV 2.5 — 1 V) than in Aubria masako.

Coloration of dorsum is uniform and structure of skin is smooth. Ventral mottled pattern

tends to disappear in adult frogs on throat first.

Synonymy. — Direct comparison of the holotype of Aubria occidentalis Perret, 1995 with

the holotype of Rana subsigillata A. Duméril, 1856 (see DuMéRi, 1861, pl. XVILL, fig. 1)

and morphometric analysis (fig. 2) clearly indicates synonymy of the two names. This

study includes also material from Ghana, included in the list of localities of paratypes of

A. occidentalis by PERRET (1995). Another piece of evidence is given by the distribution of

Aubria occidentalis, which coincides exactly with the distribution of Aubria subsigillata.

PERRET (1995) did not examine material of both species from Gabon (see OHLER &

KaAZaDI, 1990: 34, fig. 6, for the “mid-thigh gland” form of Gabon). In particular, he did

not study the type specimen of Rana subsigillata. Morphological analysis does not show

major differentiation within frogs of the genus Aubria from western Africa. Morphomet-

rical comparison of larger samples of Aubria subsigillata from Gabon with western

populations and studies using other techniques (such as bioacoustics), might lead later to

division of À. subsigillata. The name Aubria occidentalis Perret, 1995 would then be

available for the western taxon.

Sexual dimorphism. — In the genus Aubria, no secondary sexual characters are known.

Males lack vocal sacs, nuptial pads on fingers, and spinosities on different parts of the

body. Thus sexes cannot be distinguished externally.

A sample (from Kovié, Togo) large enough to allow statistic comparison between

males and females was studied. The most striking morphometrical difference between

males and females is their size (Table IV). Females are significantly larger than males. Of

the ratios to SVL of the 34 other measurements taken, only two show significant

differences (Table IV). The eye is relatively longer in males than in females and the femoral

gland is nearer to the base of the thigh (sagittal axis) in females than in males.

Source : MNHN, Paris

154 ALYTES 13 (4)

Fig. 5. — Position of femoral glands in females of Aubria subsigillata (above, MNHN 1993.1468,

Kovié, Togo) and of Aubria masako (below, MNHN 1993.2830, Ivindo, Gabon). Ventral view

of right thigh. Scale: 10 mm.

Morphological comparison shows that femoral glands are more developed in females

than in males. These composite glands are structures that are prominent and more brightly

colored and yellowish in females (fig. 6). In males the presence and position of the glands

can be recognized primarily by colour.

Source : MNHN, Paris

OHLER 155

Table IV. - Morphometrical differences between males and females of a population of

Aubria subsigillata from Kovié, Togo. SD, standard deviation, MIN,

minimum; MAX, maximum; U, Mann-Whitney U; P, probability.

Measurement Mean SD MIN MAX U test

g'n=8 77.83 4.70 67.2 83.4 U=0

SVL

$ n=7 86.44 2.60 83.6 89.7 P<0.001 ***

g'n=8 108.73 6.09 100.7 121.0 U=7.0

$ n=7 100.74 4.94 93.7 104.8 0.01 <P<0.05 *

dg'n=8 159.70 13.64 140.1 183.0 U=7.0

$ n=7 130.55 22.33 95.3 164.6 0.01<P<0.05 *

Table V. - Measurements (in mm) on tadpoles of Aubria masako as compared to

Aubria subsigillata from ScHigTz (1964). SUB, Aubria subsigillata; MAS,

Aubria masako.

MAS (stage 37) MAS (stage 40)

Total length 41.3 46.4

Snout-vent 18.8 19.0

Eye-eye 22 3.24 3.44

Nare-nare 2.0 1.55 1.75

Tail height 7.5 7.78 8.80

Keratodont formula 2:4+4/3 2:3+3/3 2:4+4/3

Source : MNHN, Paris

156 ALYTES 13 (4)

Fig. 6. — Sexual differentiation of femoral glands in Aubria subsigillata from Kovié, Togo. Above

gland of left thigh of female (MNHN 1993.1468) and below gland of left thigh of male (MNHN

1993.1466). Scale: 5 mm.

Larval characters. — A larva of this species was described by SCHIaTZ (1963) who gave

drawings of a general view and of mouthparts. Larval keratodont rows formula: 2:4+4/3

(Table V). Body length: 16 mm; tail length: 23 mm.

Distribution. — The species occurs in western and central Africa to Gabon in the coastal

area. The westernmost area of Aubria subsigillata is the valley of Sassandra river in Ivory

Coast and Diéké in southern Guinea, near the Liberian border (fig. 1). Lambarene

(Gabon, 0°41'S) is the most southern collecting point known.

Aubria masako Ohler & Kazadi, 1990

Aubria masako Ohler & Kazadi, 1990: 29. — Holotype: MNHN 1989.2775. — Type-locality: Masako

forest, near Batibongena village, 15 km from centre of Kisangani on the ancient road to Buta,

Zaire.

Specimens studied. — ZAIRE. Masako: MNHN 1989.2775, holotype; MNHN 1989.3305-

3311, K 1049, K 1266, K 1324-1326, K 1463-1464, K 2505, K 2510-2511, K 2520, K 2528,

K 2626, K 3500, K 3933; Boteke swamp: KMMA 85.21.B.123-141, KMMA 85.30.B.365-

367; near Coquilhatville: CAS 113967-113968; Equateur Province, Bikoro Territory,

Lotende Swamps near Mabali: CAS 145297; Sankuru Province, Lodja Territory,

Omaniundu: CAS 145276; Yangambi: KMNH 15478. — GaBon. Ivindo: MNHN

Source : MNHN, Paris

OHLER 157

1993.2830-2831. — ConGo. Sangha: MNHN 1945.13, 1994.1665-1666. — REPUBLIC OF

CENTRAL AFRICA. Zimba: MNHN 1993.4451. — CAMEROON. ZMH A.03134; Batouri

District (4°N 14°25'E): BMNH 1934.12.1.2; Batouri District (3°75°N 13°75'E): BMNH

1937.1.1.1; Bitye: NMW 2554 (3 specimens), NMW 2557; Ya River (Dja): NMW 2553;

Efa Yong (Efangono): NMW 2556.

Diagnostic characters. — The femoral glands are not in mid-femoral position, but closer

to the knee (fig. 5) and in a more posterior position than in Auwbria subsigillata. The

femoral glands are more elongated and feet are less webbed (1 1% — 2 II 1 — 2% III 2

— 31V2% — 1 V)thanin Aubria subsigillata. The dorsal pattern is clear with darker

spots on slight warts. A mid-dorsal line may be present. In adults the ventral mottled

pattern disappears on vent region, and remains more visible on throat.

Chresonymy. — Specimens from Ivindo forest (Gabon) listed under the name Aubria

subsigillata by PERRET (1995) are morphologically similar to specimens from Zaire and

should bear the name Aubria masako.

Adult morphology. — Before this study, this species was known only from the type-locality

and neighbouring regions (OHLER & KAZADI, 1990). Morphological variation in this

group is very important. In fact three specimens of the Sangha region of Cameroon, that

could not be assigned clearly to one of the two species in 1990 due to fixation problems,

are here tentatively placed in this group on the basis of femoral gland position, as well as

other specimens from other localities, although some of them show major differences with

the specimens from the type-locality. In the population from western Zaire (Boteke), that

is different in many morphometrical characters (fig. 4), 4 specimens out of 20 (20 %) have

a mid-dorsal stripe, a frequency lower than that observed in the type-locality (i.e. 65.2 %,

N = 23) (OHLER & KaAZaDI, 1990).

Size at metamorphosis. — 1 was able to study a sample from Boteke (Zaire) composed of

specimens of all three age-groups. This series includes numerous froglets in metamorpho-

sis, exhibiting the specific characters already present, and two tadpoles with complete

mouthparts.

The size just before metamorphosis (Gosner stage 45) varies from 18.3 to 20.0 mm

(mean SVL = 19.37; N = 13; SD = 0.46).

After metamorphosis (mean SVL = 22.6 mm; extremes 22.1-23.2 mm; N = 3; SD =

0.55), the froglets already have femoral glands in the characteristic position. These three

specimens are without mid-dorsal line and ventral pattern is complete (dark with whitish

spots).

Larval characters. — Two tadpoles with complete mouthparts were studied (Table V; fig.

7). The younger tadpole (Gosner stage 37) is lighter in colour than the older tadpole

(Gosner stage 41) and metamorphosing tadpoles. The vent and lateral parts of the body

both already have the typical coloration of Aubria: dark with white spots. In general

appearance it resembles the tadpole figured by ScHierz (1963). Measurements are very

similar with the exception of internarial distance that seems to be smaller in masako

Source : MNHN, Paris

ALYTES 13 (4)

158

) from Boteko, Zaire (dorsal and lateral

Fig. 7. — Tadpole of Aubria masako (KMMA 85.21.B.141

view, Gosner stage 37). Scale: 5 mm.

Source : MNHN, Paris

OHLER 159

tadpole (51 % of total length in the single tadpole of subsigillata and 38% in both

tadpoles of masako) just as it is in adults (Table II). AIl characters, including the

keratodont rows, must be studied on larger samples of tadpoles of both species.

DISCUSSION

SPECIES DIFFERENTIATION IN THE GENUS AUBRIA

Geographic morphological variation seems to be less important in Aubria subsigillata

than in Aubria masako. In À. subsigillata morphology seems to be very similar over the

wide range of distribution from Ivory Coast to Gabon despite important collection gaps.

No major morphometrical differences were detected between samples of different

geographic origins.

It seems doubtful that gene flow alone can explain this homogeneity. Some

populations must have been separated for an extended period of time, however no major

variations in morphology can be observed that would indicate heterogeneity in the

underlying genotype. Such observation lends support to the hypothesis of the unity of the

genotype (MAyR, 1975) due to molecular, genetic and evolutionary constraints.

In contrast, À. masako seems to be a more variable group with probably several

subgroups, at least one in western central Africa (Gabon, Congo, western Zaire), and

another one in eastern and central Zaire. These two subgroups might be interpreted as

subspecies, but more detailed data on distribution and morphological variation of Aubria

masako should clarify the situation.

PHYLOGENETIC CONSIDERATIONS

At present there is no general accepted classification of ranid frogs (DUELLMAN, 1975;

FRosT, 1985; LAURENT, 1986; FEI et al., 1990; Dugois, 1992; BLOMMERS-SCHLÔSSER, 1993;

EMERSON & BERRIGAN, 1993), as the classification of Ranoidea is in a state of revolution.

It did not change for almost a century. BOULENGER (1920) proposed to subdivide the genus

Rana in several subgenera. This was followed in the African region by subsequent

recognition of several genera. But the same did not occur in the other areas, particulary

Asia, where members of this group show the most important radiation. Works on

phylogeny and systematics of Asian ranids were published in Chinese (Lu & Hu, 1961; Fer

et al., 1990) and ignored in Western countries. The work of CLARKE (1981) was a first step

to analyse relationships of ranid subgroups and led DuBois to propose his 1987b

classification. Dugois (1992) included many new data on morphology of ranid frogs. As

the main problem remains the classification of the genus Rana, which is clearly

polyphyletic, subdivision of this genus in provisional subgenera allows formulation of

phylogenetic hypotheses more easily. Further work is needed to resolve the higher

categories (tribes becoming subfamilies, subfamilies becoming families, etc.). This does not

affect nomenclatural stability, because generic allocation of many species in Rana is

Source : MNHN, Paris

160 ALYTES 13 (4)

maintained. Nevertheless keeping of Dicroglossinae in the genus Rana is no longer

possible, as members of the Dicroglossinae comprise a well identified monophyletic group

(OuLer & DuBois, 1989; EMERSON & BERRIGAN, 1993). For these reasons, I here follow

Dusois’ (1992) classification. According to this classification, the genera Aubria and

Pyxicephalus, which have long been known to be closely related (PROCTER, 1919),

constitute together the subfamily Pyxicephalinae.

Femoral glands are present in phylogenetically distinct anuran groups. They can be

found in Pelobatidae and in several families of Ranoïidea. In the Mantellidae, Phrynoba-

trachidae and the ranid subfamily Ranixalinae, the glands are present in male specimens

only or are much more conspicuous in males than in females. This suggests that these

glands are not homologous to those present in the Pyxicephalinae, where the reverse is

observed (see below). In fact such a puzzling distribution of this structure among anurans

suggests a high level of homoplasy. For Pyxicephalinae the presence of femoral glands can

be interpreted as an apomorphic character supporting the monophyly of the group.

Three important characters allow distinction of the two species of Aubria (see also

PERRET, 1995): the position of the femoral gland, the mid-dorsal line and the ventral

pattern. Only the position of the femoral glands is constant in all post-metamorphic

specimens of a given species. In some old alcohol preserved specimens that were dried and

had their colour changed to blackish, the femoral glands are hard to recognize.

The position of the glands is either mid-femoral or close-to-knee. No ontogenetic

changes were observed. In specimens (males and females) of the genus Pyxicephalus,

femoral glands are found in the close-to-knee position (FLGLDT = 534; SD = 84.5; N

= 23) asin 4. masako. No variation in gland position has been found in this genus (SUEUR

& OHLER, in preparation). This position of femoral glands is therefore interpreted here as

the plesiomorphic character state. The mid-thigh position of femoral glands present in 4.

subsigillata is here interpreted as the apomorphic character state.

The mid-dorsal line is a feature observed in two of the 40 populations studied; both

have femoral glands in close-to-knee position. In other frog species, the presence-absence

of a mid-dorsal line has been shown to be determined by a single gene (MoRIWAKI, 1953;

Dugois, 1979; BERGER & SMIELOWSKI, 1982). This coloration pattern can be found in

Pyxicephalus and in the Dicroglossinae, the possible sister-group of the Pyxicephalinae

(CLARKE, 1981; Dumois, 1987b, 1992). In the populations of 4. subsigillata, a mid-dorsal

line was never observed. The permanent absence of the mid-dorsal line, here viewed as a

secondary loss of an allele from the gene pool of the species, is the apomorphic character

state. The possible presence of mid-dorsal line is the plesiomorphic state of this character.

The third character is the ventral colour pattern. Three different states can be distin-

guished and form an ordinated transformation (0 — 1 — 2). The absence of ventral color-

ation pattern can be found in Dicroglossinae and in Pyxicephalus and is the plesiomorphic

state (0). The intermediate state is the presence of a ventral pattern of whitish rounded

patches on a dark ground in young ontogenetic stages (older larvae and juvenile frogs) and

the gradually disappearing in adult stage (1). This is observed in Aubria masako. In the third

state, the colour pattern remains distinct in adults (2). I interpret this as a partial paedomor-

phism sensu DuBois (19874): a somatic feature (here, ventral colour pattern) shows juvenile

character in an adult phenotype. This state is present in Aubria subsigillata.

Source : MNHN, Paris

OHLER 161

The subfamily Dicroglossinae, represented by the genera Hoplobatrachus (species

occipitalis and tigerinus) and Conraua, is defined by CLARKE (1981) by one character: 7,1

P — anteriorly reduced preorbital process of pars fascialis of the maxilla.

The Pyxicephalinae (Pyxicephalus and Aubria) are a monophyletic group defined by

CLARKE (1981) by 5 apomorphic characters: 2,1 — presence of cranial exostosis; 4,1 P —

presence of occipital canal; 6,1 — a zygomatic ramus of the squamosal much longer than

the otic ramus and articulating with the postorbital process of the pars fascialis of the

maxilla; 14,3 — strong overlap of the anterior border of parasphenoïd ala by the medial

ramus of the pterygoid articulating along at least 1/2 the anterior width of the

parasphenoid ala; 19,1 — sternal style a long bony element tapering markedly anteriorly

to posteriorly.

The genus Pyxicephalus is defined by three apomorphic characters: 9,1 — pterygoid

process of maxilla well developed, directed postero-medially, overlapping anterior ramus

of pterygoid, with which it forms a suture; 16,1 P — base of the omosternum slightly

forked, the greatest space between the arms being less than half the width of one arm; 22,4

— terminal phalanges of fingers and toes reduced, almost cone-like.

Superposition of the new characters defined above on the cladogram proposed by

CLARKE (1981) (fig. 8) corroborates the phylogeny of CLARKE (1981), and one apomorphic

character of Aubria has been defined.

BIOGEOGRAPHY

Aubria is generally considered as a genus limited to high forests (LAMOTTE, 1966;

ScHieTz, 1967). Comparing our locality data with vegetation maps (WHITE, 1983) seems

to agree with these statements. Nevertheless, one must be very careful with this kind of

data, because geographical maps may not reflect the detailed local situation. In many

regions, forest is very sparse and fragmented due to human activities. When precise

collection data are available for Aubria, they mention forest localities or forest border

areas. Distribution of the two species is in fact closely related to rain forest distribution

in Africa.

African tropical forest is not a homogeneous ecosystem. Several forest types can be

recognized (WHITE, 1983). Several biogeographic refuges or core areas can be recognized

from distribution data of vertebrates (HAMILTON, 1988). Two major refugia areas can be

found on the western and eastern border of the Zaire basin. Two biologically somewhat

more impoverished areas can be found in western Africa in Sierra Leone and Liberia, and

in eastern Ivory coast and western Ghana. The Zaire basin seems to be an area of disjunct

distribution for many taxa. The distribution gap corresponds to the area of the swamp

forest (WHITE, 1983). Aubria seems to be absent from this swamp forest; this could be due

to the presence in this forest of an open canopy which resembles secondary forest, where

Aubria seems to be absent (SCHISTZ, 1967).

However, AMIET (1989) included Aubria among the species occurring in modified

forest habitats. He found choruses in secondary habitats, in forest swamp areas or even

Source : MNHN, Paris

162 ALYTES 13 (4)

Pyxicephalinae Dicroglossinae

Aubria Pyxicephalus

subsigillata masako

Fig. 8. — Interrelationships of Pyxicephalinae as proposed by CLARKE (1981). The tree was rooted

with Dicroglossinae (genus Hoplobatrachus) as the outgroup. Nodes 1, 2 (in part), 5 and 6 have

been defined by CLARKE. Node 2 is corroborated by one addititional synapomorphy discussed in

the present study (femoral glands present in both sexes close to knee position). Nodes 3 and 4

are each supported by an apomorphic state of the transformation series of the ventral pattern

coloration (node 3: ventral colour pattern in juvenile stages; node 4: ventral colour pattern also

in adult stage). Node 4 is further supported by an additional apomorphy (permanent absence of

mid-dorsal line).

in open swWamp areas (AMIET, personal communication). Further exploration of central

African forest should clarify the ecological preferences of the two species of Aubria.

Two species of Aubria are present in the central African forest area. From geographic

distribution, Aubria subsigillata appears to be a form distributed in lowland rain forest

near the west and central African coast. Aubria masako is found in the basin of the Congo

river and in the Cameroonian plateaus. In Cameroon and Gabon, where the two species

occur, the distinction of a coastal form (subsigillata) and a “continental” form (masako)

is clear (PERRET, 1995, as occidentalis and subsigillata respectively).

The character analysis indicates an eastern Zaire basin origin for Aubria. Starting

from the Zaire basin, Aubria has colonized the entire area of tropical forest of central and

western Africa. Splitting up of the forest in historical dry periods might be a factor of the

evolution of two species. The observed intraspecific variation might result from ecological

isolation. Further investigations using call analysis and adequate samples for morphomet-

rical studies might discriminate further taxa.

Source : MNHN, Paris

OHLER 163

FUNCTION OF FEMORAL GLANDS

The femoral glands, either in mid-femoral or close-to-knee position, are always

present, in males, females and juveniles examined (smallest specimen in metamorphosis,

KMMA 85.21.B.125: SVL = 18.3 mm). Their observation is sometimes difficult due to

fixation problems. Often a small incision of the skin in the presumed gland region can

resolve the doubt: if a gland is present, the skin shows a basal thickening due to the large

size of glandular cells.

The difference in position of the femoral glands is distinctive but not sufficiently

significant to question gland homology. The correlation of a large set of characters

between the two species and to their outgroup Pyxicephalus confirms this homology.

Nothing is known of the biological signification of this gland. In a large sample of

Aubria subsigillata (23 specimens from Kovié, Togo), the development of the gland

(prominence) seems to depend on the sex of the frogs. Adult females have clearly more

prominent glands than males whose glands can be recognized only by their different colour

(fig. 6). This is confirmed by PERRET’s (1995) observation on samples from Ivory Coast

(under the name 4. occidentalis). Nevertheless in other specimens it is still difficult to

distinguish male and female specimens on external characters only. In further investiga-

tions attention should be paid to seasonal variation of femoral gland size, development

and colour in Aubria.

In Mantellidae and Phrynobatrachidae, femoral gland aggregates are visible or more

developed in males. As femoral glands, and other ventrally positioned glands, of males are

in contact with females during amplexus, a stimulatory function of femoral gland

secretions is assumed (DUELLMAN & TRUEB, 1985: 58).

In Aubria, schooling of tadpoles has been observed (ScHioTz, 1963). A similar

behaviour is known in Pyxicephalus adspersus, another species of the subfamily

Pyxicephalinae (POYNTON, 1964: 95). Tadpoles are “very gregarious”’ and tend to swarm

around the male of Pyxicephalus remaining in the water. Even juveniles have been known

to form swarms (POYNTON, 1964). It would be interesting to understand the nature of these

aggregations.

Tadpoles of various species recognize their siblings and distinguish tadpoles of

different clutches. This behaviour has been particularly studied in the American toad Bufo

americanus (WALDMAN, 1985). Factors that allow sibling recognition seem to be fixed or

modified in ontogenetic development. A factor permitting sibling recognition might be

transferred between the sibling tadpoles, or it might be transferred from the mother to her

offspring. Differential recognition of maternal half-siblings by separately reared tadpoles

suggests contribution of a factor of the maternal parent (product of oviduct for example)

(WaLDMAN, 1981). The secretion of female femoral glands might be such a primary factor

for Aubria and Pyxicephalus offsprings. It would provide a basis for learning the sibling

smell chemical characteristics. This gland secretion might be deposited by the mother on

the eggs while laying them and gland function might be determined by the ovulatory

hormone system. This hypothesis is more congruent with having active glands in females

than the traditional hypothesis of stimulatory function during amplexus. Another kind of

Source : MNHN, Paris

164 ALYTES 13 (4)

“sexual character reversal” in anurans is known in Limnodynastes peroni (Myobatrachi-

dae, Limnodynastinae), where females bear lateral fringes on the fingers (DUELLMAN &

TRUEB, 1985: 56-57): in this species, females use their hand in paddling movements for

stirring water and spawn into a foam nest. The use of a secondary sexual character in

offspring care would be an interesting parallelism. Experimental investigations could be

carried out on Aubria and Pyxicephalus to test the hypothesis of chemical transmission of

family recognition.

ACKNOWLEDGEMENTS

1 would like to thank the following curators and collection managers for the loan of and the

possibility to study the specimens from their collections: Barry T. CLARKE (BMNH), Jens V. VINDUM

(CAS), Danny MeiRTE (KMMA), Jean MarIAUX (MHNG), L. WALSCHAERTS (MRHN), Franz

TIEDEMANN and Heinz GRiLLITSCH (NMW), Rainer GÜNTHER (ZMB), Hans-Wilhelm KOEPCKE

(ZMH). For their helpful comments on the different drafts of this paper 1 thank Roger BOUR, W.

Ronald HEYER and Hussam ZAHER, and especially Alain Dumois. Jérôme SUEUR provided

Pyxicephalus measurements used in this study. Jean-Louis AMIET, Barry T. CLARKE, Raymond F.

LAURENT and Arne SCHIOTZ read the initial manuscript and gave me interesting remarks concerning

ecology and biogeography of African Amphibians.

LITERATURE CITED

AMET, J.-L., 1989. — Quelques aspects de la biologie des amphibiens anoures du Cameroun. Année

biol., 28: 73-136.

ANNANDALE, N., 1912. — Description of the tadpole of Rana pleskii, with notes on allied forms. Rec.

ind. Mus., 2: 345-346.

BERGER, L. & SMIELOWSKI, J., 1982. — Inheritance of vertebral stripe in Rana ridibunda Pall.

(Amphibia, Ranidae). Amphibia-Reptilia, 3: 145-151.

BLOMMERS-SCHLÔSSER, R. M. A., 1993. — Systematic relationship of the Mantellinae Laurent 1946

(Anura Ranoïdea). Ethol. Ecol. Evol., 5: 199-218.

BOULENGER, G. A., 1917. — Sur la conformation des phalangettes chez certaines grenouilles

d'Afrique. C. r. Acad. Sci., 165: 987-990.

ee 1920. — A monograph of the South Asian, Papuan, Melanesian, and Australian frogs of the

genus Rana. Rec. indian Mus., 20: 1-126.

CLARKE, B. T., 1981. — Comparative osteology and evolutionary relationships in the African Raninae

(Anura Ranidae). Monit. zool. ital., (N.S.), 15 (suppl.): 285-331.

DeLauGERRE, M. & Dunois, A, 1985. — La variation géographique et la variabilité intrapopula-

tionelle chez Phyllodactylus europaeus (Reptilia, Sauria, Gekkonidae). Bull. Mus. natn. Hist. nat.,

(4), 7 (A) (3): 709-736.

Dunois, A., 1976. — Les grenouilles du sous-genre Paa du Népal (famille Ranidae, genre Rana). Cah.

nép. Doc., 6: i-vi + 1-275.

1979. — Anomalies and mutations in natural populations of the Rana “esculenta” complex

(Amphibia, Anura). Mit. zool. Mus. Berlin, 55: 59-87, pl. I.

- 1984. — Sample-size constraints in the use of the nonparametric Mann-Whitney U test for the

comparison of two independent samples: consequences in anuran amphibians systematics.

Alytes, 3: 20-24.

1987a. — Neoteny and associated terms. Alytes, 4: 122-130.

1987b. — Miscellanea taxinomica batrachologica (1). Alytes, 5: 7-95.

Source : MNHN, Paris

OHLER 165

1992. — Notes sur la classification des Ranidae (Amphibiens Anoures). Bull. Soc. linn. Lyon, 61:

305-352.

-- 1995. — Keratodont formulae in anuran tadpoles: proposals for a standardization. J. zool. Syst.

Evol. Res., 33: I-XV.

DuELLMAN, W. E., 1975. — On the classification of frogs. Occ. Pap. Mus. nat. Hist. Univ. Kansas,

42: 1-14.

DuELLMAN, W. E. & TRUE, L. 1985. — Biology of amphibians. New York, McGraw-Hill, “1986”:

ixvii + 1-670.

DumékiL, A., 1856. — Note sur les reptiles du Gabon. Rev. Mag. Zool., (2), 8: 369-377 + 417-424

+ 460-470 + 553-562.

-— 1861. — Reptiles et poissons de l'Afrique occidentale. Etude précédée de considérations générales

sur leur distribution géographique. Arch. Mus., 10: 137-268, pl. XIII-XXIIL.

EMERSON, S. B. & BERRIGAN, D., 1993. — Systematics of Southeast Asian ranids: multiple origin of

voicelessness in the subgenus Limnonectes (Fitzinger). Herpetologica, 49: 22-31.

Fe, L., YE, C. & HUANG, Y., 1990. — Key to Chinese Amphibia. [In Chinese.] Chongging, Editions

of Sciences and Techniques: [i-iv] + 1-2 + 1-364.

FRosT, D. R. (ed.), 1985. — Amphibian species of the world. Lawrence, Allen Press & Assoc. Syst.

2 [iv] + iv + 1-732.

Gosner, K. L., 1960. — A simplified table for staging anuran embryos and larvae with notes on

identification. Herpetologica, 16: 183-190.

HAMILTON, A. C., 1988. — Guenon evolution and forest history. /n: A. GAUTIER-HION, F. BOURLIÈRE,

J.-P. GAUTIER & J. KINGDON (eds.), À primate radiation: evolutionary biology of the African

guenons, Cambridge, Cambridge University Press: 13-34.

LAMOTTE, M., 1966. — Types de répartition géographique de quelques batraciens dans l'Ouest

africain. Bull. I.F.A.N., 28 (A) (3): 1140-1148.

LAURENT, R. F., 1953. — Reptiles et batraciens récemment parvenus au Musée royal du Congo belge.

Bull. Cercle zool. congolais, 21: 21-29.

pr 1955. — Une méthode pour la recherche des meilleurs caractères taxonomiques fournis par les

proportions. Ann. Soc. r. zool. Belg., 84: 271-282.

_— 1981. — Phénogrammes d’anoures basés sur la morphométrie. Monit. zool. ital., (N.S.), 15

(suppl): 1-22.

-— 1986. — Sous-classe des Lissamphibiens (Lissamphibia). Systématique. /n: P.-P. GRASSE & M.

DeLsoL (eds.), Traité de zoologie, 14, Batraciens, Fasc. 1-B, Paris, Masson: 594-797.

Liv, C.-C. & Hu, S.-C., 1961. — The tailless amphibians of China. [In Chinese.] Shangh:

1-364, pl. I-VI + I-XXVIII.

Mar, E., 1975. — The unity of the genotype. Biol. Zbl., 94: 371-388.

MoriWaki, T., 1953. — The inheritance of the dorso-median stripe in Rana limnocharis Wiegmann.

J. Sci. Hiroshima Univ., (B), div. I (Zool.), 14: 159-164.

Myers, C. W. & DUELLMAN, W. E., 1982. — A new species of Hyla from Cerro Colorado, and other

tree frog records and geographical notes from Western Panama. 4m. Mus. Novitates, 2152: 1-32.

NIDEN, F., 1908. — Die Amphibienfauna von Kamerun. Mit. zool. Mus. Berl., 3: 489-518.

Norusis, M. J., 1992. — SPSS for Windows. Base system user's guide. Release 5.0. Chicago, SPSS

Inc: i-xvi + 1-672.

OHLER, A. & Dumois, A., 1989. — Démonstration de l’origine indépendante des ventouses digitales

dans deux lignées phylogénétiques de Ranidae (Amphibiens, Anoures). C. r. Acad. Sci., 309 (3):

419-422.

Our, A. & KAZADI, M., 1990. — Description d’une nouvelle espèce du genre Aubria Boulenger,

1917 (Amphibiens, Anoures) et redescription du type d’Aubria subsigillata (A. Duméril, 1856).

Alytes, 8: 25-40.

PARKER, H. W. 1936. — The amphibians of the Mamfe Division, Cameroons. Proc. zool. Soc. London,

1936: 135-136.

PERRET, J.-L. 1966. — Les amphibiens du Cameroun. Zoo!. Jb. (Syst.), 93: 289-464.

_— 1995. — Revision of the genus Aubria Boulenger 1917 (Amphibia Ranidae) with the description

of a new species. Trop. Zool., (1994), 7: 255-269.

POYNTON, J. C., 1964. — The Amphibia of Southern Africa: a faunal study. Ann. Natal Mus., 17:

1-334.

ixvi +

Source : MNHN, Paris

166 ALYTES 13 (4)

PROCTER, J. B., 1919. — On the skull and affinities of Rana subsigillata A. Dum. Proc. zool. Soc.

London, 1919: 21-27.

ScHiorz, A., 1963. — Amphibians of Nigeria. Vidensk. Medd. fra dansk naturh. Foren., 125: 1-92, 4

pl.

= 1967. — The treefrogs of Western Africa. Spolia zool. Mus. Haun., 25: 1-346.

‘WALDMAN, B., 1981. — Sibling recognition in toad tadpoles: the role of experience. Z. Tierpsychol.,

56: 341-358.

= 1985. — Sibling recognition in toad tadpoles: are kinship labels transferred among individuals?

Z. Tierpsychol., 68: 41-57.

Wie, F., 1983. — The vegetation of Africa. A descriptive memoir to accompany the Unesco /

AETFAT / UNSO vegetation map of Africa. Natural Resources Research, 20: 1-356 + 4 maps.

Wirre, G. F. DE, 1930. — Liste des batraciens du Congo Belge (collection du Musée du Congo Belge

à Tervuren). Première partie. Rev. Zool. Bot. afr., 19: 232-274.

Wüsrer, W. & THORPE, R. S., 1992. — Asiatic cobras: population systematics of the Naja naja

species complex (Serpentes: Elapidae) in India and Central Asia. Herpetologica, 48: 69-85.

ZAR, J.H., 1984. — Biostatistical analysis. Englewood Cliffs, New Jersey, Prentice-Hall International,

Inc: i-xiv + 1-718.

Corresponding editor: W. Ronald HEYER.

Source : MNHN, Paris

Alytes, 1996, 13 (4): 167-178. 167

Site fidelity and homing ability in Hvla labialis

(Anura, Hylidae)

Horst LÜDDECKE

Departamento de Ciencias Biolôgicas,

Universidad de los Andes, Santafé de Bogotä, A. A. 4976, Colombia

A Colombian highland population of Hyla labialis was studied for 73

consecutive weeks by applying a unique toe clip combination to each of 304

adult male and 181 adult female individuals captured and released in the field.

The analysis of a recapture record of 74% for males and 43% for females

revealed that frogs released at their capture site were reencountered almost

exclusively there (98% males, 89% females), hence demonstriing * site

fidelity. The majority of frogs translocated to a release site at a distance of up

to 200 metres returned to their original capture site (58% males, 53%

females), hence demonstrating homing ability. The frogs’ spatial behaviour

varied in relation to the breeding cycle of the population. After translocation

between two successive breeding seasons, most males as well as females

homed. When translocated during the breeding season, relatively more males

homed and relatively more females remained at the foreign release site.

Almost 30 % of all translocated frogs remained at the foreign release site and

about 15 % were found at a third site. The short-term reaction to translocation

was stronger in males (53 % homed) than in females (31 % homed), but as time

passed the discrepancies leveled off and a year later no statistically significant

differences remained between male and female spatial behaviour.

INTRODUCTION

Short or long lasting site fidelity and homing ability have been reported for several

species of anuran amphibians belonging to different families (review in SINsCH, 1992b).

Site fidelity may be related to stationary behaviour of territorial or brood-caring

individuals (MCVEY et al., 1981; STEWART & RAND, 1992; VAN WINGAARDEN & VAN

GooL, 1994), but in many cases seasonal migratory long distance movements are involved,

where each individual exhibits its site fidelity and homing ability through active round

trips between two or more places (breeding site, estivation site, hibernation site), carried

out in a repetitive manner according to the reproductive cycle of the population (HEUSSER,

1968; GLANDT, 1986; SINscH, 1990).

Site fidelity and homing ability have also been demonstrated experimentally by

displacing individual frogs from their capture site to a different place at a certain distance,

where they were released and either followed in order to see where they were heading

(OLDHAM, 1966; TRacY & DoLr, 1969; SiNsCH, 1988) or marked with the expectation

Source : MNHN, Paris

168 ALYTES 13 (4)

of recapture near their original capture site (SINSCH, 1990; RiTkE et al. 1991; PaPt,

1992).

Most of these studies on site fidelity and homing ability have been conducted in the

temperate zones of Europe (HEUSSER, 1969; RyYsER, 1989; SINsCH, 1992a) and North

America (OLDHAM, 1966; DoLE, 1972). However, these faculties have also been revealed

for some neotropical frog species (DIXoN & STATON, 1976; MCVEY et al., 1981; CRUMP,

1986; SINsCH, 1988; STEWART & RAND, 1992; VAN WINGAARDEN & VAN GooL, 1994).

The purpose of the present study was to examine site fidelity and homing ability in

the neotropical frog species Hyla labialis. In this paper I report the results of an experiment

designed to survey the spatial behaviour of this high mountain frog and to analyse it in

relation to gender and reproductive activity of the population.

MATERIAL AND METHODS

STUDY AREA

The field study was carried out in the Parque Nacional Natural Chingaza nature

reserve at 3500 metres in the eastern chain of the Andes mountains near the Colombian

capital Santafé de Bogotä. This pâramo is open grassland interspersed with spongy moss

cushions and rosette plants (VUILLEUMIER & MONASTERIO, 1986). The ground is bog-like

in many parts. Higher vegetation is sparse and patchy, consisting of small-leafed shrubs

and the characteristic tall frailejones (several species of the genus Espeletia, family

Compositae). The climate at this high altitude is cold and humid. There is a strong daily

but only a small yearly temperature cycle. Daytime air temperatures may rise to 22°C

when it is sunny, but mostly stay below 12°C, due to the usually heavy cloud cover and

frequent rains. Nightly frosts occur occasionally, particularly in the dry season (December

to March). The precipitation pattern in Chingaza is unimodal, peaking in June with about

260 mm of a total yearly rainfall of 1900 mm (SARMIENTO, 1986).

Hyla labialis was studied in a valley that stretches and slopes gently in a north-south

direction for about 700 metres, where six groups of shallow ponds 100-200 metres apart

(identified by the letters C, I, V, B, P and L) offered suitable breeding sites for this species

(fig. 1). H. labialis was the only frog species that used these ponds for egg deposition. The

only other frog species present in the study area, Colostethus subpunctatus, Eleuthero-

dactylus bogotensis, and E. elegans, deposit terrestrial eggs.

BREEDING BIOLOGY OF HYL4 LABIALIS

These frogs (male mean SVL 51 mm, female mean SVL 63 mm) usually breed in

permanent ponds. Amplexus occurs in the water, where calling males are approached by

receptive females. Whereas males spend several months per year in or near the breeding

pond, females enter the water mainly for oviposition, which usually takes one night.

Source : MNHN, Paris

LÜDDECKE 169

about 100 metres

edge of -

western eu sé GE

mountain Le

sbpes mountain

slopes

Fig. 1. — Sketch of the mountain valley where the spatial behaviour of Hyla labialis was studied.

Control frogs were captured and released in the central breeding area C. Experimental frogs

from five breeding areas (I, V, B, P and L) were translocated to area C and released there.

Arrows indicate the homeward directions and rough distances.

Females spend most of their life on land, up to hundreds of metres away from the water.

Two main breeding seasons per year last from about one to three months each: one peaks in

February-March, the other in September-October. Occasional breeding activity may occur

any time, even during the driest months of the year, when there is little water in the ponds.

TESTING FOR SITE FIDELITY AND HOMING ABILITY

From June 1991 until August 1992, the six breeding areas were searched during

daylight hours for adult H. labialis about once a week by slowly walking around the ponds

in order to spot frogs visually. AIl encountered frogs were captured and censused. Many

of these animals were already individually marked by toe-clipping from previous studies

(LÜDDECKE, unpublished). All newly caught frogs were marked individually in the same

manner. Most animals were released on the same day of capture, but some were kept in

the laboratory for a week prior to their release. The search for marked frogs was continued

until December 1992, but after August 1992 all encountered animals were released at their

capture site.

Source : MNHN, Paris

170 ALYTES 13 (4)

To test for site fidelity, 187 frogs found in the central breeding area C were released

at their capture site. To determine their homing ability, 298 individuals captured in the five

breeding areas surrounding area C were translocated to area C and released there.

Recapture of these translocated frogs at their original capture site would indicate homing

behaviour.

STATISTICAL ANALYSES

Due to disparate amounts of data for males and females, and lack of normal

distribution of data, many comparisons were made after converting actual counts into

percentage numbers, or applying chi-square tests or non-parametric statistics (NEAVE &

WORTHINGTON, 1988).

RESULTS

RECAPTURES

A total of 485 individual adult Hyla labialis were marked and released. Of these, 306

individuals were recaptured, yielding an overall recapture rate of 63.1 %. Recapture rates

differed significantly between males and females (7? = 45.6, P < 0.01). In both sexes,

recapture rates were slightly higher for translocated frogs than for home-released frogs

(Table I). Recapture success was highest during peaks of breeding activity when frogs were

abundant at the ponds. Time intervals between capture and first recapture ranged from

one week to 73 weeks. The mean time interval between release and first recapture was

shorter for males (18 weeks) than for females (23 weeks). During 26 weeks after initial

release, the time span between two reproductive peaks, relatively more males (167 of 227,

74%) than females (46 of 79, 58 %) were recaptured.

SITE FIDELITY

Of 86 individually marked males originally captured and released in the central

breeding area C, 85 (98.8 %) were recaptured in the same area. The recapture rate for

females (25 of 28, 89.3 %) at the original capture site was not significantly different from

that of males (y? = 3.22, P > 0.05). Four frogs left the home area and each moved to

a different site.

HOMING ABILITY

Because there were no significant differences in homing between frogs from the five

breeding areas surrounding C (x? = 5.15, P > 0.05), I pooled the data from these areas

Source : MNHN, Paris

LÜDDECKE 171

Table I. - Recapture record of individual adult male and female Hyla labialis handled in site fidelity and

homing ability experiments in the Péramo de Chingaza.

Treatment

Released at capture site 86 of 119 28 of 68

Released after translocation to another site | 141 of 185 51 of 113

Table IL. - Long-term gpacial behaviour of 141 adult male and 51 adult female Hyla labialis after

translocation from their original capture site to a common central release site.

Returned to original capture site

Remained at release site

Moved to third site

Table II. - Comparison of long-term spatial behaviour during and between breeding seasons of adult male

female Hyla labialis afier capture and release in brocding area C and afier translocation to

C from all other breeding areas combined.

A. Breeding area C.

Site specific 46 98 17 89 (le 38 97 8 89

Moved to another site 1 2 2 11 1 # 1 il

B. All other breeding areas combined.

During breeding seasons Between breeding seasons

Miles Females Males | Females

N % N % N x N %

Returned to capture site 35 55 16 52 47 61 ul 55

Remained at release site 21 33 7 22 21 27 7 n:

Moved to third site 8 12 8 26 Le 12 2 10

Source : MNHN, Paris

172 ALYTES 13 (4)

(Table I). The majority of the frogs recaptured after translocation had returned to their

original capture site. Some individuals homed within a single week, others were not seen

again until more than a year later. Many translocated frogs were recaptured at the release

site and others were found at a third site. The spatial behaviour of translocated frogs was

significantly different from random (Goodness-of-fit test, y? = 53.7 , P < 0.001). Both

sexes showed the same general tendency and, in spite of some disparities, there was no

significant difference between male and female spatial behaviour (7? = 1.76, P > 0.05).

SPATIAL BEHAVIOUR IN RELATION TO REPRODUCTIVE CYCLE

During the study period there were three reproductive peaks: October 1991, February

1992, and October 1992. The approximate length of each breeding season was determined

by the presence of amplectant pairs and recently deposited egg clusters of H. labialis in the

breeding ponds. Over the entire study period, males and females from the central breeding

area C were equally site-specific, regardless of being released during or between breeding

seasons (Table IT).

To estimate the spatial behaviour of translocated frogs in relation to breeding activity,

as a first approach I used first-recapture data covering the entire study period in order to

look for long-term patterns (Table III). There were only slight and insignificant differences

( = 0.41, P > 0.05) in male behaviour: a few less homed and a few more remained after

translocation during a breeding season compared to translocation between breeding

seasons. Females homed about equally well anytime and performed similarly to homing

males. Considerably fewer females remained at the foreign release site and instead moved

to a third site when translocated during the breeding season compared to their behaviour

after translocation between breeding seasons. Due to this shift, there was a highly

significant behavioural difference between females translocated during or between

breeding seasons (7? = 17.4, P < 0.01).

In a second approach I used multiple-recapture data and set a time limit of ten weeks

between release and recapture. This 10-week interval was presumably short enough to

ensure that frogs had not yet entered their next reproductive phase. The males’ short-term

reaction was the same during and between breeding seasons. In contrast, female behaviour

differed according to season: when translocated during a breeding season, more than half

of the females remained at the foreign release site, but when translocated between breeding

seasons, more than half of the females homed (fig. 2). Again, females behaved differently

G = 8.9, P < 0.01) depending on the timing of translocation, but this time the shift

occurred between returning and remaining frogs. Few males and females moved to a third

site.

GRADUAL CHANGE IN SPATIAL BEHAVIOUR

Having found differences between the frogs’ short- and long-term reactions to

translocation, 1 examined how the spatial behaviour changed over time. I processed

first-recapture data obtained from a sample of 110 individual frogs (86 males, 24 females)

Source : MNHN, Paris

LÜDDECKE 173

TRANSLOCATION BETWEEN BREEDING SEASONS

MALE BEHAVIOUR FEMALE BEHAVIOUR

TRANSLOCATION DURING BREEDING SEASON

MALE BEHAVIOUR FEMALE BEHAVIOUR

Fig. 2. — Spatial behaviour of male and female Hyla labialis within the first ten weeks after

translocation either between or during breeding seasons, based on multiple recapture data. Black

areas: frogs returned to capture site; grey areas: frogs moved to third site; white areas: frogs

remained at release site.

captured during the same breeding season in the breeding areas surrounding C,

translocated and released at C. When calculating the proportion of frogs in each

behavioural category by accumulating data from sequential recapture-samples obtained

first at the end of that breeding season (week 14), and afterwards at ten-week intervals, a

gradual and statistically highly significant (Friedman Two-Way ANOVA, x? = 11.07, P

= 0.003) change in the proportions of frogs in each behavioural category became evident

as more and more frogs were recaptured (fig. 3).

Initially there was a large difference between the sexes: more than half of the males,

but only one third of the females had homed. In contrast, proportionately more females

Source : MNHN, Paris

174 ALYTES 13 (4)

males returned

females returned

males remained

Cumulative Individual Count (%)

females remained

males moved to third site

e.H#HO00es

females moved to third site

4 14 24 34 44 54 64 74

Weeks after translocation

Fig. 3. — Temporal change in spatial behaviour of 110 adult male and female Hyla labialis

translocated in the same breeding season.

than males remained at the foreign release site during the first weeks after translocation.

The relative numbers of males and females that moved to a third site were about equal.

Almost six months later (week 24), males had attained a distribution approaching the final

one, whereas that of females was still close to the initial one. However, when all individuals

had been recaptured, the differences between males and females had almost vanished, and

the situation resembled the long-term result for all translocated males and females (Table

Il). But the patterns of behavioural change over time differed significantly between sexes

{match test M, = 30, P < 0.05).

Source : MNHN, Paris

LÜDDECKE 175

DISCUSSION

RECAPTURE RECORD

Comparison of recapture rates reported for anuran species is difficult due to

differences in capture schedule, search intensity and marking techniques, and conspicuous-

ness, activity cycles and abundance of the species studied. With values over 70 % for males

and over 40 % for females, the recapture rates for Hyla labialis were high compared with

those reported for other anuran species (OLDHAM, 1966; Horz, 1970; RiTkE et al. 1991;

BUCHACHER, 1993). This may be related to my continuous capture-recapture programme

over a period of 73 consecutive weeks, whereas most other studies relied on one capture

period in the first year and a recapture period about one year later, both timed during

breeding activity (OLDHAM, 1966; Tracy & DoLr, 1969; DoLE, 1972; RITKE et al. 1991;

SINsCH, 19924).

SITE FIDELITY

Male as well as female H. labialis showed a high degree of site fidelity, similar to the

results obtained for H. chrysoscelis (RITKE et al., 1991), Bufo americanus (OLDHAM, 1966),

and male B. calamita (SiNsCH, 1992a). BUCHACHER (1993) obtained a lower site-fidelity rate

of 67 % for Pipa arrabali, which he ascribed to a high mobility of individuals between

adjacent ponds that were only a few metres apart. Remarkable cases of site specificity,

where individuals were recaptured only a few metres distant from the original capture site,

have been reported for male Bufo bufo spinosus in Italy (Hoïz, 1970), Leptodactylus

macrosternum during the wet season in Venezuela (DIxoN & STATON, 1976), and some

territorial dendrobatid species (MCVEY et al. 1981; VAN WIINGAARDEN & VAN GooL,

1994). This precision in site specificity was also observed in this study for some individual

males and females of H. labialis.

The absence of site fidelity in Bufo verrucosissimus has been ascribed to the instability

of suitable spawning sites in successive years (TARKHNISHVILI, 1994), and in B. calamita

females to their opportunistic choice of the spawning site in response to calling males

(SNscH, 19924). The strong site fidelity of Hyla labialis females conforms to the

environmental conditions and reproductive biology of the population studied. Although

long-term pond stability in the pâramo is undocumented, the breeding ponds in the study

area did not dry for several years, even during the strong El Niño year of 1992 (LÜDDECKE,

personal observation). Thus site-specific females seem to run almost no risk in moving to

a home pond and finding it dry. Because spawning may happen occasionally almost any

month of the year, female H. labialis also benefit from being site-specific when the scarcity

of males at the breeding ponds would offer little opportunity for long-distance phonotactic

orientation to callers.

Source : MNHN, Paris

176 ALYTES 13 (4)

HOMING TENDENCY

Translocated adult H. labialis were recaptured at the breeding ponds mostly during

the breeding season, indicating that frogs returned there for reproduction. The moderate

percentage of homing H. labialis (56.8 %) probably was not due to increased mortality

while moving across the terrain, since the recapture rates of translocated frogs were higher

than those of home-site released frogs. If frogs recaptured at a third site are regarded as

potentially homing, then about 70 % of the translocated individuals had this tendency.

This interpretation seems justified by the fact that some translocated frogs detoured to a

third site prior to returning to the home site. Another possible reason for the moderate

homing success, in spite of the strong homing tendency, is related to the high site fidelity

evidenced by the control group: a strongly site-specific individual (although it may

migrate) could be familiar with just a small fraction of the entire valley that was used as

the study area. This would mean that finding home after displacement was hampered in

individuals whose release site lay outside their migratory corridor. Nothing is known

about the orientation mechanisms used by H. labialis.

Almost 30 % of the translocated and recaptured H. labialis remained at the foreign

release site. One possible reason for this behaviour is that the release site (breeding area

C) was an appropriate place for reproduction and was therefore accepted in exchange for

the original site. Most remaining frogs were recaptured only once shortly after release and

their long-term spatial behaviour is unknown. However, some frogs first recaptured at the

release site were found at the original capture site on their second or third recapture,

indicating that they had only delayed their homeward journey.

RiTkE et al. (1991) pointed out that long intervals between release and recapture may

be due to slow recovery from a reproductive effort. The sex difference in recapture rate of

H. labialis translocated in one and recaptured in the next breeding season implies that

about half of the females skipped every other breeding season and spawned only once per

year, while most males participated in every breeding season. This would indicate that

females recovered slower than males from a reproductive event.

Since shortly after translocation during the breeding season most females were still

found at the release site (fig. 3), at first glance this would mean indifference as to where

to oviposit. But it turned out that, when recaptured, most of these females had not yet

spawned at the foreign release site. Half of the females found at a third site and most

females that had already returned to their home site when recaptured shortly after

translocation were still gravid (LÜDDECKE, unpublished data). These results are compatible

with strong site fidelity and suggest that females have the tendency to oviposit at a familiar

breeding site.

RESUMEN

Una poblaciôn de Hyla labialis de alta montaña en Colombia fue estudiada durante

73 semanas consecutivas, aplicando un marcaje ünico a cada uno de 304 machos y

Source : MNHN, Paris

LÜDDECKE 177

181 hembras adultos capturados y liberados en el campo. El anälisis de los datos de

recaptura del 74% de machos y 43 % de hembras revelé que las ranas capturadas y

liberadas en su sitio de captura fueron reencontradas casi exclusivamente alli (98 % de los

machos, 89 % de las hembras), lo que demostro su fidelidad al hogar. La mayoria de las

ranas traslocadas a un sitio de liberaciôn a una distancia de hasta 200 metros regresé a

su sitio de captura original (58 % de los machos, 53 % de las hembras), lo que demostro

su capacidad de retornar al hogar. El comportamiento espacial de las ranas variaba acorde

al ciclo reproductivo de la poblaciôn: después de la translocacin entre dos épocas

reproductivas sucesivas, la mayoria de los machos y hembras retornaban al hogar.

Después de una translocaciôn durante una época reproductiva, relativamente mâs machos

retornaban, pero relativamente mâs hembras se quedaban en el sitio de liberaciôn. Casi el

30% de las ranas translocadas se quedaba en el sitio de liberaciôn y el 15% fue

encontrada en otro sitio distinto. La reacciôn inmediata a la translocaciôn era mâs fuerte

en los machos que en las hembras (53 % y 31 %, respectivamente, regresaban al sitio de

captura), pero con el paso del tiempo las discrepancias en el comportamiento espacial

disminuyeron y un año después no quedaron diferencias significativas entre machos y

hembras.

ACKNOWLEDGEMENTS

thank Martha Lucia BOHORQUEZ ALONSO and Adolfo AMÉZQUITA TORRES for their help during

the field work. I am grateful to Ulrich SNsCH for comments on an earlier draft of the manuscript and

appreciate the advice of Janalee CALDWELL and two anonymous reviewers leading to considerable

improvements of the manuscript. The Universidad de los Andes gave financial support for this study,

and the governmental agency for natural resources INDERENA granted permission to carry it out

in the Chingaza National Park.

LITERATURE CITED

BUCHACHER, C. O., 1993. — Field studies on the small Surinam toad, Pipa arrabali, near Manaus,

Brazil. Amphibia-Reptilia, 14: 59-69.

Crump, M. L., 1986. — Homing and site fidelity in a neotropical frog, Atelopus varius (Bufonidae).

Copeia, 1986: 438-444.

Dixon, J. R. & STATON, M. A., 1976. — Some aspects of the biology of Leptodactylus macrosternum

Miranda Ribeiro (Anura: Leptodactylidae) of the Venezuelan llanos. Herpetologica, 32:

227-232.

DoLr, J. W., 1972. — Homing and orientation of the displaced toads, Bufo americanus, to their home

sites. Copeia, 1972: 151-158.

GLANDT, D., 1986. — Die saisonalen Wanderungen der mitteleuropäischen Amphibien. Bonn. zool.

Beitr., 37: 211-228.

HEUSSER, H., 1968. — Die Lebensweise der Erdkrôte (Bufo bufo L.): Wanderungen und Sommer-

quartiere. Rev. suisse Zool., 75: 928-982.

ee 1969. — Die Lebensweise der Erdkrôte (Bufo bufo L.): Das Orientierungsproblem. Rev. suisse

Zool., 76: 444-517.

Source : MNHN, Paris

178 ALYTES 13 (4)

Horz, H., 1970. — Zur Laichplatzôkologie von Bufo bufo spinosus Daudin (Amphibia, Salientia) im

tyrrhenischen Ligurien. Vierteljahrschr. Naturforsch. Gesellsch. Zürich, 115: 239-254.

McVey, M. EE. ZAHARY, R. G., PERRY, D. & MACDoUGAL, J., 1981. — Territoriality and homing

behavior in the poison dart frog (Dendrobates pumilio). Copeia, 1981: 1-8.

NEAVE, H. R. & WORTHINGTON, P. L., 1988. — Distribution-free tests. London, Unwin Hyman.

OLDHAM, R. S., 1966. — Spring movements in the American toad, Bufo americanus. Can. J. Zool.,

44: 63-100.

Part, F., 1992. — General aspects. /n: F. PAPI (ed.), Animal homing, London, Chapman & Hall: 1-18.

RirkE, M. E., BaBs, J. G. & RiTke, M. K., 1991. — Breeding-site specificity in the gray treefrog (Hyla

chrysoscelis). J. Herpe., 25: 123-125.

RYsER, J., 1989. — The breeding migration and mating system of a Swiss population of the common

frog Rana temporaria. Amphibia-Reptilia, 10: 13-22.

SARMIENTO, G., 1986. — Ecological features of climate in high tropical mountains. /n: F. VUILLEUMIER

& M. MONASTERIO (eds.), High altitude tropical biogeography, Oxford Univ. Press.

Siscn, U., 1988. — El sapo andino Bufo spinulosus: anälisis preliminar de su orientaciôn hacia sus

lugares de reproducciôn. Boletin de Lima, 57: 83-91.

- 1990. — Migration and orientation in anuran amphibians. Ethol. Ecol. Evol., 2: 65-79.

- 19924. — Sex-biassed site fidelity and orientation behaviour in reproductive natterjack toads

(Bufo calamita). Ethol. Ecol. Evol., 4: 15-32.

ee 1992b. — Amphibians. /n F. PAP1 (ed.), Animal homing, London, Chapman & Hall: 213-233.

STEWART, M. M. & RAND, A. S., 1992. — Diel variation in the use of aggressive calls by the frog

Eleutherodactylus coqui. Herpetologica, 48: 49-56.

TARKHNISHVILI, D. N., 1994. — Breeding of the toad Bufo verrucosissimus: sexual dimorphism and

shifting spawning sites. Amphibia-Reptilia, 15: 191-198.

TRACY, C. R. & DoLr, J. W., 1969. — Orientation of displaced toads, Bufo boreas, to their breeding

sites. Copeia, 1969: 693-700.

VAN WUNGAARDEN, R. & VAN GooL, S., 1994. — Site fidelity and territoriality in the dendrobatid

frog Dendrobates granuliferus. Amphibia-Reptilia, 15: 171-182.

VUILLEUMIER, F. & MONASTERIO, M. (eds.), 1986. — High altitude tropical biogeography. Oxford Univ.

Press.

Corresponding editor: Janalee P. CALDWELL.

Source : MNHN, Paris

Alytes, 1996, 13 (4): 179-192. 179

Merocrine secretion

from serous cutaneous glands

in Rana esculenta complex and Rana iberica

Giovanni DELFINO, Rossana BRiZzI & Guido MELIS

Dipartimento di Biologia animale e Genetica,

Università di Firenze, Via Romana 17, 50125 Firenze, Italy

Serous cutaneous glands of larval and juvenile specimens pertaining to

the Rana esculenta complex and Rana iberica may modulate product dis-

charge through exocvtotic release and weak compression from the myoepi-

thelial sheath. This method of secretory release is a typically merocrine

mechanism and is an unusual functional trait in these glands, which are

generally credited with holocrine bulk discharge of anti-predatory poison

products. Current trends in animal toxinology claim that the most common

active compounds in anuran poisons (i.e. biogenic amines and peptides)

played an early role in skin homeostasis during the phylogeny of this order,

before participating in chemical skin defence against predators. On the basis

of this hypothesis, we conclude that the merocrine activity described in Rana

serous cutaneous glands is an ancestral functional characteristic related to the

use of skin products in local regulative mechanisms.

INTRODUCTION

The concept of holocrine secretory activity performed by serous glands in anuran skin

was developed by FARAGGIANA (1938b), who carried out pioneering experimental studies

on the regenerative processes which follow glandular discharge (FARAGGIANA, 1938a,

1939). Her findings were later confirmed under TEM (DELFINO, 1980), and further

investigations, coupled with pharmacological approaches (BENSON & HADLEy, 1969;

HoLmes et al., 1977; HOLMES & BALLS, 1978; DELFINO et al., 1982), stressed the role of the

myoepithelial cells in the massive release of the serous product together with the secretory

cytoplasm and nuclei.

The latter method of secretory discharge is consistent with the use of the serous

secretory products as anti-predatory toxins and/or repellents (DUELLMAN & TRUEB, 1985).

Nonetheless, recent studies on the Rana esculenta complex, showing morphological

changes in the granules containing serous skin products, suggested that these substances

are also employed in skin homeostasis (BARNI et al., 1987). Furthermore, ultrastructural

evidence has proved that discrete amounts of product are released by serous glands in

juvenile Hyla arborea through exocytosis (DELFINO et. al., 1994). These findings disclose

Source : MNHN, Paris

180 ALYTES 13 (4)

new perspectives to the interpretation, both functional and evolutive, of serous skin

secretion in anurans (DALY et al., 1987). In the light of this updated approach to the

regulating function of poison secretions of anuran skin, we aimed to investigate whether

the serous cutaneous glands of two groups of European frogs, the green frogs (Italian

Rana esculenta complex) and the brown frogs (Spanish Rana iberica), can release their

cutaneous serous products through exocytosis, i. e. by merocrine mechanisms. The above

taxa were studied as they represent a genus which includes extensively studied species, both

morphologically and physiologically, commonly found in laboratories. This choice enabled

us to obtain adequate references from previous literature and at the same time allows

verification of our results.

We avoided any conditions which could elicit bulk secretion and focused on the

patterns which could refer to the modulated release of the serous product. We observed

advanced larval specimens and newly metamorphosed froglets, as these development

stages appeared to best fit our purpose. Serous glands in late tadpoles show advanced

patterns of biosynthesis but relatively undifferentiated myoepithelial cells, a condition

which may minimize bulk discharge; on the other hand, juveniles are very imperfect

terrestrial vertebrates and must maximize their cutaneous homeostatic mechanisms when

exposed to the subaerial environment.

We detected evidence of merocrine secretory processes and compared them with the

traditional holocrine mechanism reported in serous cutaneous glands of anurans.

MATERIAL AND METHODS

Larval specimens of Italian green frogs (Rana esculenta complex) and of the Iberian

brown frog Rana iberica were collected from the outskirts of Florence (Italy) and Mellid

(Santiago de Compostela, Spain), respectively. The tadpoles were reared in the laboratory

until the first specimens underwent metamorphosis. The schedule below reports the

developmental range considered, according to TAYLOR & KOLLROS (1946), as well as the

number of animals and the samples observed for each, under light (LM) and electron

microscopes (TEM).

(A) Stage XX, Rana esculenta complex: LM, 3 tadpoles, 2 samples from each; TEM,

2 tadpoles, 2 samples from each.

(B) Stage XXII, Rana iberica: LM, 3 tadpoles, 3 samples from each; TEM, 3 tadpoles,

3 samples from each.

each; TEM, 2 froglets, 2 samples from each.

(D) Stage XXV, (juvenile) Rana iberica: LM, 2 froglets, 2 samples from each; TEM,

2 froglets, 2 samples from each.

The choice of the developmental range was based on former studies which showed

that, between stages XIX and XXV, anuran skin possesses small, but already differentiated

glands which already contain large secretory accumulations undergoing post-Golgian

maturation (DELFINO, 1977; DELFINO et al., 1988, 1994). Furthermore, this developmental

Source : MNHN, Paris

DELFINO, Brizzi & MELIS 181

range also embraces the differentiation timing of the glandular myoepithelial sheath

(DELrINO et al., 1987), which plays a role in gland depletion.

Tadpoles and froglets were anaesthesized in 0.2 % aqueous chlorbutol and sacrificed.

Small bands of skin (about 4 mm?) were removed from the back and fixed in Karnovsky’s

aldehyde mixture (KarNovsky, 1965). The skin strips were then reduced into smaller

(about 2 mm?) fragments and post-fixed in 1 % OsO,. For all these procedures, a 0.1 M,

PH 7 sodium cacodylate buffer was employed, at a temperature of 4°C. The specimens

were dehydrated in a graded ethanol series, soaked in propylene oxide and infiltrated with

Epon 812 to obtain flat embeddings. The embeddings were then cut with a NOVA LKB

ultramicrotome into semithin (1-1.5 um) and ultrathin (silver-white interference colour)

sections; these sections were used for the light and electron microscope observations,

respectively. The semithin sections were stained with buffered 10 % toluidine and used to

prepare ultrastructural investigation. The ultrathin sections were collected on uncoated

copper grids and stained with a hydroalcholic saturated solution of uranyl acetate,

followed by bismuth subnitrate (0.8 mg/ml in an alkaline solution). These specimens were

finally observed under a Siemens 101 electron microscope at 80 kV.

RESULTS

Under the light microscope, serous glands at premetamorphic stages XX and XXII

(fig. la-b) exhibit the structural feature characteristic of the Anura. The secretory units,

which are syncytial in structure, are provided with a continuous contractile sheath of

myoblasts (the future myoepithelial cells, mec) as well as a cap of undifferentiated cells,

representing the regenerative matrix (intercalary tract or neck) of the gland. According to

the usual ultrastructural patterns shared by all serous glands of anurans, the poison

adenomeres of Rana tadpoles possess exiguous lumina, which are restricted to the

subintercalary level. However, in advanced larvae of the Iberian frog, these cavities are

rather enlarged (fig. la-b) when compared to other anurans studied so far. The luminal

space sinks into the secretory syncytium and resembles a multichambered compartment,

owing to the interposition of slender cytoplasmic walls, which look like thin bridges

between the secretory unit and neck (fig. la-b). Observation of serial sections revealed that

duct and gland lumina are separated only by a discontinuous cytoplasmic screen (fig. 1b).

Actually, the intercalated cells, which are arranged in an irregular doughnut structure,

partly obstruct the central opening with slender cytoplasm projections. The secretory

product may reach the duct lumen through the spaces separating these cell processes.

Secretory granules crowd around the luminal boundary and mould themselves around its

surface, so that they assume a crescent-like shape in section.

In juvenile glands, the serous syncytium is a solid structure, thus limiting the small

lumen to the neck region (fig. 1c). The secretory cytoplasm holds large amounts of poison

product, which consists of dense particles, subspherical in shape. However, larger

magnifications revealed a spongy-like substructure in some granules (fig. 1d).

Under the electron microscope, the secretory syncytium possesses a well developed

biosynthesis apparatus and nuclei scattered in a single row at the periphery. Slender rough

Source : MNHN, Paris

182 ALYTES 13 (4)

Fig. 1. — Serous gland anlagen under light microscope. Bars: 10 um.

a. — Rana iberica (XXNH). Longitudinal section tangential to the intercalary tract (it), consisting

of a few layers of undifferentiated cells. The apical portion of the secretory syncytium displays a

multichambered lumen with slender cytoplasmic partitions (arrows). Notice the serous secretory

product (s) crowded around the lumen and the thin myoepithelial cell layer (mec). m: melanophore.

b. — Rana iberica (XXH). Longitudinal section through the widest diameter of the secretory unit,

as shown by the lumen of the duct which is obvious. Only a thin, discontinuous screen (arrowhead),

formed by slender cytoplasmic processes of intercalary tract cells, separates the duct and adenomere

lumina. Coupled arrows in the duct outlet, arrow points to cytoplasmic partition between apical

chambers. s: serous secretory product.

c. — Rana esculenta complex (juvenile). Longitudinal section: the secretory unit lacks any

obvious lumen, a slender cavity (1) is detectable only in the intercalated tract-duct complex. Note two

mucous glands (mg) provided with a wide lumen. mec: myoepithelial cell layer; s: serous secretory

product.

d. — Rana esculenta complex (juvenile). Detail of the serous secretory product contained in the

syncytial cytoplasm. Arrows point to granules with a spongy-like substructure.

Source : MNHN, Paris

DELFINO, BRIZZI & MELIS 183

endoplasmic reticulum (rer) cisterns (fig. 2a-b) and Golgi stacks (dictyosomes) (fig. 2a,

2c-d) are consistent in both species observed, and occupy the peripheral cytoplasm. The

distal (trans) face of the Golgi stacks dispatches minute vesicles (fig. 2d), which merge

together to form wide, up to 4 um in diameter, secretory deposits. These large vesicles

contain a finely dispersed product (fig. 2c-e, 5a), which undergoes marked condensation

(maturation) in later stages of biosynthesis (fig. 3a). Several merging processes affect the

secretion aggregates (fig. 3b, 4), which sometimes display a spongy-like substructure, owing

to the alternation of darker and paler zones. Secretory product maturation is a gradual

process in which the granules become involved at different rates. The process produces a

remarkable polymorphism, due to the coexistence of intermediate stages of condensation.

In some instances the intermediate steps are by-passed, as revealed by condensation

patterns affecting the material freshly dispatched from the Golgi stacks (fig. 3c).

During the condensation phase, the membrane which borders the serous aggregates

becomes detached from the secretory material and leaves a transparent halo around it. In

the meantime, numerous, slender (about 25 nm in diameter), microvilli-like outgrowths of

the cytoplasm intrude this perigranular space. The microvilli are branched and hollow, and

form a network around the secretory product (fig. 3d).

From the central cytoplasm the secretory granules reach the upper level of the

syncytium, just beneath the lower cell layer of the intercalary tract (neck) of the gland. As

observed under light microscope, a proper lumen exists at this level, separated from the

neck and duct lumina only by interposition of slender cytoplasmic processes of the

intercalated tract cells. The serous deposits crowd around the lumen (fig. 4) and some

adhere to the luminal plasmalemma, so that their limiting membranes merge with it.

Rather large openings form in this way, which allow the secretory product to be released

into the lumen (fig. 4, inserts A and B), following the pattern of a merocrine process. This

mechanism appears, however, to be quite peculiar; the spaces holding the granules engaged

in the secretory release are continuous with the compartments containing other serous

aggregates, due to serial confluences (fig. 4, insert B). In this way the secretory product

may flow out toward the lumen, at a rate which depends on the number of granules

involved in reciprocal confluence.

The secretory units engaged in this exocytotic activity possess the neuro-contractile

apparatus typical of serous glands, consisting of neurites and myoepithelial cells (fig. 5a).

The thin axons are clearly recognizable as they contain parallel neurotubules (fig. 5b-d)

and electron transparent vesicles in their endings (fig. 5c), whereas in the myocytes

myofilaments occupy large portions of the cytoplasm, leaving two symmetrical zones at

the nuclear poles to accommodate scanty organelles (fig. 5a). Nevertheless, this

muscle-nerve cell machinery seems to be still immature, since myofilaments fail to fill the

myocyte cytoplasm, and neuromuscular junctions are infrequent and resemble poorly

specialized, occasional contacts (fig. 5b). In their erratic course, some neurites may also

make contact with the secretory syncytium (fig. 5d).

Serous glands in recently metamorphosed froglets display large amounts of cytoplasm

filled with dense aggregates, which arise from further condensation of the secretory

deposits described in larval adenomeres. Despite their density, secretory granules are not

homogeneous in aspect as they exhibit the spongy-like substructure (fig. 6a-b) already

Source : MNHN, Paris

184 ALYTES 13 (4)

Fig 2. — Biosynthesis apparatus in serous gland anlagen of both species. Bars: 500 nm.

a. — Rana esculenta complex (XX). Peripheral portion of the secretory syncytium, closely

contiguous to the myoepithelial layer. Note slender rer cisterns and a Golgi stack, or dictyosome (G).

mec: myoepithelial cell.

b. — Rana iberica (XXII). Parallel array of rough endoplasmic reticulum complements and

mitochondria with dense matrix.

c. — Rana esculenta complex (XX). Golgi stacks (G) frequently occur in the perinuclear

cytoplasm of the secretory syncytium. The dictyosomes fulfil the activity of the rough endoplasmic

reticulum and lead to the accumulation of a fine serous secretory material (s) inside large vesicles.

d-e. — Rana iberica (XXI). Minute vesicles (encircled in d) derive from the periphery of the

Golgi stacks (G) and merge (arrows in e) with the larger serous secretory deposits (s), contributing

material to these storage structures.

Source : MNHN, Paris

DELFINO, BRIZZI & MELIS 185

Fig. 3. — Patterns of poison maturation in both species. Bars: 500 nm.

a. — Rana esculenta complex (XX). The secretory product contained in the vesicles undergoes

a remarkable condensation, which gives rise to granules provided with a compact substructure.

b. — Rana iberica (XXII). Serous maturation, recognizable from variable density of the product,

proceeds through sequential confluence processes between secretory aggregates (arrows).

c. — Rana esculenta complex (XX). In some instances the product contained in the vesicles

dispatched by the Golgi stacks (G) is promptly condensed and the vesicle phase is by-passed.

d. — Rana iberica (XXI). Slender microvilli fill the space between the limiting membrane and

secretory product. Arrows indicate branching points which give the microvillous cluster the

appearance of a three-dimensional net.

Source : MNHN, Paris

186 ALYTES 13 (4)

Fig. 4. — Rana iberica (XXM). Bars: 1 um. Subintercalar portion of serous secretory unit. Note the

multichambered lumen (I), the characteristic cytoplasmic walls (large arrows) and secretory

aggregates of various density. Only an extremely thin screen (small arrows) separates the large

vesicle from the lumen. Inserts A and B show continuity between vesicular and luminal

compartments and between serous vesicles (B). Coupled arrows show outflow paths. it:

intercalary tract; s: serous secretory product.

described in the previous stages. In neometamorphosed froglets, the biosynthesis

apparatus typical of protein manufacturing glands decays remarkably owing to rarefaction

of rer and dictyosomes. The organelles are first reduced to the very perinuclear zone and

later disappear altogether (fig. 6a-b).

Touch stimulation of cutaneous areas, possibly painful in nature, may evoke local

responses in the myoepithelial sheath of serous glands before sacrifice. Contracted

myocytes show well-defined thickenings in their myofilament apparatus, which in turn

cause remarkable morphological changes in the shape of the nuclei. Myofilaments act as

a sphincter around the surface of the nucleus and mould it into the shape of an hourglass,

with the peripheral half contained in the myocyte and the inner one bulging towards the

secretory syncytium (fig. 6c).

The effects elicited by myocyte compression may be confirmed by patterns detectable

in the gland duct. This is a slender intraepidermal interstice that crosses the epithelial

Source : MNHN, Paris

Fig. 5. — Rana iberica (XXII). Bars: 1 pm. The neuromuscular apparatus of the serous gland anlagen

consists of myoepithelial cells (mec) and neurites (ne). nt: neurotubules; s: serous secretory

product.

a. — Note the dome-shaped, perinuclear portion of the myocpithelial cell engaged in

myofilament accumulation. The biosynthesis apparatus is obvious at the cell poles. Neurites are

contained in the interstice between secretory and contractile compartments.

b. — Detail of the previous figure. Note parallel neurotubules in the axon and non-specialized

contacts between neurite and myoepithelial cell (encircled).

c. — Detail of (a) showing a neurite ending. Small electrontransparent vesicles (arrows) are

obvious.

d. — In some instances, only a 30 nm wide gap (encircled) separates the neurite from the serous

syncytium.

Source : MNHN, Paris

188 ALYTES 13 (4)

Fig. 6 — Serous glands in juveniles. Bars: lum.

a-b. — Rana iberica (a) and Rana esculenta complex (b): the serous product consists of granules

with spongious substructure and high electrondensity; the secretory organules are restricted to the

syncytium periphery.

c. — Rana esculenta complex: in contracted myoepithelial cells the nuclei bulge toward the

secretory syncytium; arrowheads indicate a dense ring of myofilaments around the nucleus which is

shaped like an hourglass.

d. — Rana esculenta complex: gland duct reaching the body surface, lined with horny layer (hl)

and keratinocytes. The serous secretory product (s) consists of a structureless, moderately

electrondense material; coupled arrows indicate duct outlet.

e. — Enlargement of the previous micrograph, showing a detail of the duct; only cytoplasmic

projections of intercalary tract cells (ic) are interposed between duct and neck lumina. s: serous

secretory product; hl: horny layer.

Source : MNHN, Paris

DELFINO, Brizzi & MELIS 189

layers, bounded by a horny lining, continuous with the external stratum corneum of the

body surface (fig. 6d). The duct lumen contains a rather opaque, structureless material,

devoid of any membranous body (fig. 6e). The lack of membrane bounded serous material

within the lumen is consistent with merocrine secretory release.

DISCUSSION AND CONCLUSIONS

Several ultrastructural and pharmacological investigations have depicted the dis-

charge mechanism in serous cutaneous glands in anurans. The interstices between the

secretory unit and myoepithelium hold axonal endings (Bück & LERTPRAPAI, 1972;

WHITEAR, 1974) which form synaptic junctions with the myocytes and contain dense-cored

vesicles (150 nm in diameter), adrenergic in appearance (DockRAY & Hopkins, 1975;

SIÔBERG & FLOCK, 1976; DELFINO, 1979, 1991; BARBERIO et al., 1987). This sympathetic

innervation has been confirmed by nor-epinephrine stimulation which allowed pharma-

cological characterization of receptors on myoepithelial cell membranes (BENSON &

HaADLEY, 1969; HOLMEs et al., 1977; HoLMES & BaLLs, 1978; DELFINO et al., 1982). TEM

observations disclosed typical ultrastructural features in still contracted mec: myocytes

show alternating thickenings among myofilaments, whereas their nuclei are displaced

towards the secretory syncytium; the plasmalemma facing the stroma displays several folds

(DELriNo, 1980, 1991; DELFINO et al., 1987, 1990). In the Italian green frog specimens, we

observed both myofilament thickenings and nuclear displacements, although the patterns

were not so dramatic as in experimental specimens. The serous product is stored within the

syncytium cytoplasm, so that the intense compression exerted by the myoepithelial sheath,

triggered by pharmacological stimulation, causes bulk discharge of the secretory product,

syncytium cytoplasm and nuclei (DELFINO, 1980). When the secretory product consists of

dense particles, they can be collected in saline and were found to still possess their limiting

membranes (DocKRAY & Hopkins, 1975, in Xenopus laevis). This excludes that granule

release occurs through exocytosis, which involves insertion of the membrane encompassing

the secretory particle into the plasmalemma. Under pharmacological stimulation, the

activity of serous cutaneous glands of anurans may be regarded as holocrine in nature,

since it involves emission of fragments of the secretory units. This bulk emission

mechanism, which requires phasic contractions achieved through neuromuscular machin-

ery, is well consistent with the defensive role ascribed to the serous cutaneous glands of

anuran skin.

However, recent trends credit these secretory units, at least in Rana species, to

perform a regulatory role in the water balance of the skin. MizLs & PRUM (1984) assign

this function to specialized cells of the mucous and seromucous glands (the latter represent

a novel type found by these authors in Rana catesbeiana, Rana pipiens and Rana

temporaria), whereas in the Rana esculenta complex BARNI el al. (1987) stressed the role

of the secretion (venom) of large serous glands in controlling water balance. This is

suggested by changes both in granule morphology (vacuolization) and chemical compo-

sition, fitting the annual cycle of the frog. The cycle includes two periods (activity,

hibernation) with alternating changes in skin permeability. The above authors detected a

Source : MNHN, Paris

190 ALYTES 13 (4)

lower protein content during the winter phase of the annual cycle, whereas the amounts

of 5-HT — involved in controlling ionic transport and water exchange — increased

considerably.

It appears obvious that secretory mechanisms, based on holocriny, do not agree with

the above findings, which witness change in granule morphology (possibly due to

mobilisation of some component substances), rather than gland depletion. The enteramine

(S-HT) is synthesized in the hyaloplasm and later accumulates in the granules (DELFINO,

1977) by means of an active, counter gradient mechanism (DAWSON, 1970, on entero-

chromaffine cells). Conversely, biologically active proteins and 5-HT may be transferred

from their storage sites (granules) into the hyaloplasm, and then towards the dermal

environment with the contribution of the myoepithelial cells which may perform active

transport through exo-endocytotic activity (BANI & DELFINO, 1990).

In late larval specimens and juveniles of Rana esculenta complex and Rana iberica we

observed a third way of secretory release, differing both from bulk depletion and selective

mobilisation of compounds stored in the granules, which involves the merocrine discharge

of the serous product into the exiguous lumen, as well as moderate compression exerted

by myoepithelial sheath, capable of driving the product towards the surface. In some

instances, the skin poison of Rana esculenta complex has been observed passing through

the duct, and was seen to still maintain a discrete structure (FRASCHINI, 1965). It must be

stressed that the duct wall is discontinuous in structure, since it consists of funnel-like,

telescopically arranged keratinocytes (DELFINO, 1976, 1991). Thus, fluidified secretory

materials may spread through the intraepidermal interstice, an effect which is amplified

under experimental conditions and leads to the collection of secretory product beneath the

superficial horny layer (DELFINO, 1980). Following this pathway, active molecules of the

serous product may exert their effect in the intraepidermal environment. Pharmacological

studies have demonstrated that the contractile activity of myoepithelial cells of serous

glands in anuran skin fits this functional mechanism as contraction may be modulated by

adrenergic antagonist and ionic concentration (BENSON & HADLEY, 1969; HOLMES et al.,

1977; HoLmEs & BALLSs, 1978; DELFINO et al., 1982), adjusting serous discharge to

functional requirements.

This dual gland activity based on merocrine and holocrine mechanisms agrees with

the phylogenetic scenery proposed by DaALy et al. (1987). They postulate that anuran

cutaneous toxins include ancestral regulative molecules (biogenic amines and peptides)

that are widespread among living families and have been secondarily engaged in defence

strategy, according to the evolution of receptor molecules in predators. In our opinion the

exocytotic way of release is consistent with the use of cutaneous secretions in skin

homeostasis, whereas bulk discharge fits an antipredatory role.

RESUMEN

Las gländulas cutäneas serosas de ejemplares larvales y juveniles de Rana esculenta

complex y Rana iberica pueden expulsar su secreciôn por medio de un proceso merocrino.

Source : MNHN, Paris

DELFINO, BRiZZI & MELIS 191

Tal mecanismo de secreciôn contrasta con la funciôn defensiva, antipredatoria, del veneno

de los anuros, basado sobre un proceso holocrino; sin embargo, el meécanismo merocrino

es coherente con la ancestral funciôn reguladora atribuida a algunos componentes de esta

secreciôn cutänea

LITERATURE CITED

BANI, G. & DeLriNo, G., 1990. — Ultrastructure of the myoepithelial cells of the cutaneous glands

in several amphibian species. Biomed. Res., 1: 73-83.

BARBERIO, C., DELFINO, G. & MASTROMEI, G., 1987. — A low molecular weight protein with

antimicrobial activity in the cutaneous “’venom” of the yellow-bellied toad (Bombina variegata

pachypus). Toxicon, 25: 899-909.

BARNI, S., BERNOCCHI, G. & BoTTIROLI, G., 1987. — Histochemistry and morphology of the secretory

granules of skin venom glands of Rana esculenta during the active and hibernating period.

Arch. Biol., 98: 391-406.

BENSON, B. J. & HADLEY, M. E., 1969. — Jn vitro characterization of adrenergic receptors controlling

skin gland secretion in two anurans, Rana pipiens and Xenopus laevis. Comp. Biochem. Physiol.,

30: 857-864.

BôcK, P. & LERTPRAPAI, N., 1972. — Ein sarkoplasmatisches Retikulum in den myoepithelialen

Zellen der Giftdrüsen in der Haut der Gelbbauchunke (Bombina variegata variegata L.).

Cytobiologie, 6: 416-480.

Day, J. W., Myers, C. W. & WHITTAKER, N., 1987. — Further classification of skin alkaloids from

Neotropical poison frogs (Dendrobatidae), with a general survey of toxic/noxious substances

in the Amphibia. Toxicon, 25: 1023-1095.

Dawson, I., 1970. — The endocrine cells of the gastrointestinal tract. Histochem. J., 2: 527-549.

DeELriNo, G., 1976. — Structural and ultrastructural aspects of the cutaneous granular glands in

Bombina variegata (L.) (Amphibia Anura Discoglossidae). Monit. zool. ital., (n. s.), 10: 421-448.

Le 1977. — Il differenziamento delle ghiandole granulose cutanee in larve di Bombina variegata

pachypus (Bonaparte) (Anfibio, Anuro, Discoglosside). Ricerca al microscopio ottico e al

microscopio elettronico. Arch. ital. Anat. Embriol., 82: 337-363.

= 1979. — Le ghiandole granulose cutanee di Alytes cisternasii Boscà e Discoglossus pictus Otth

(Anfibi, Anuri, Discoglossidi): struttura, ultrastruttura e alcuni dati citochimici. Arch. tal.

Anat. Embriol., 84: 81-106.

ee 1980. — L'attività rigeneratrice del tratto intercalare nelle ghiandole granulose cutanee

dell'ululone Bombina variegata pachypus (Bonaparte) (Anfibio, Anuro, Discoglosside); studio

sperimentale al microscopio elettronico. Arch. ital. Anat. Embriol., 85: 283-310.

- 1991. — Ultrastructural aspects of venom secretion in anuran cutaneous glands. /n: A. T. Tu

(ed.), Handbook of natural toxins, New York, Marcel Dekker: 775-802.

DELFINO, G., AMERINI, S. & MUGELLI, A., 1982. — In vitro studies on the “venom” emission from the

skin of Bombina variegata pachypus (Bonaparte) (Amphibia Anura Discoglossidae). Cell Biol.

int. Rep., 6: 843-850.

DELFINO, G., Brizzi, R. & BORRELLI, G., 1988. — Cutaneous glands in anurans: differentiation of the

secretory syncytium in serous Anlagen. Zool. Jb. Anat., 117: 255-275.

DELrINO, G., Brizzi, R. & CALLONI, C., 1987. — Differentiation of myoepithelial cells during the

development of cutaneous serous glands in Anura. Zool. Anz., 218: 219-236.

_— 1990. — A morpho-functional characterization of the serous cutaneous glands in Bombina

orientalis (Anura: Discoglossidae). Zool. Anz., 225: 295-310.

LE 1994. — Serous cutaneous glands in the tree-frog Hyla arborea arborea (L.): origin, ontogenetic

evolution, and possible functional implications of the secretory granule substructure. Acta

Zool., Stockholm, 75: 27-36.

Source : MNHN, Paris

à 1 MARS 1996

192 ALYTES 13 (4)

DockraY, G. J. & Hopkins, C. R., 1975. — Caerulein secretion by dermal glands in Xenopus laevis.

J. Cell Biol., 64: 724-733.

DuELLMAN, W. E. & TRUE, L., 1985. — Biology of amphibians. New York, McGraw-Hill: i-xix +

1-670.

FARAGGIANA, R., 1938a. — Ricerche istologiche sulle ghiandole cutanee granulose degli Anfibi Anuri.

1. Bufo vulgaris e Bufo viridis. Arch. ital. Anat. Embriol., 39: 327-376.

- 1938b. — La struttura sinciziale e il meccanismo di secrezione delle ghiandole cutanee granulose

di Anfibi Anuri. Monit. zool. ital., 49: 105-108.

4e 1939. — Ricerche istologiche sulle ghiandole cutanee granulose degli Anfibi Anuri. II. Rana

esculenta, Rana agilis e Bombinator pachypus. Arch. ital. Anat. Embriol., 41: 390-410.

FRASCHINI, A., 1965. — Caratteristiche istochimiche e di fluorescenza delle ghiandole granulose

cutanee di Rana esculenta nel corso della metamorfosi. Boll. Zool., 32: 33-39.

Homes, C. & BALLS, M., 1978. — /n vitro studies on the control of myoepithelial cell contraction in

the granular glands of Xenopus laevis skin. Gen. comp. Endocrinol., 36: 255-263.

Hozmess, C. H., MoonDi, P. S., RAO, R. R. & BALLS, M., 1977. — Jn vitro studies on the effects on

granular gland secretion in Xenopus laevis skin of stimulation and blockade of « and

adrenoceptors of myoepithelial cells. Cell biol. int. Rep., 1: 263-270.

KarNovsky, M. J., 1965. — A formaldehyde-glutaraldehyde fixative of high osmolarity for use in

electron microscopy. J. Cell Biol., 27: 137A.

Muiss, J. W. & PRUM, B. E., 1984. — Morphology of the exocrine glands of frog skin. Am. J. Anat.,

171: 91-106.

SiôBERG, E. & FLOCK, À., 1976. — Innervation of skin glands in the frog. Cell Tissue Res., 172: 81-91.

TAYLOR, À. C. & KOLLROS, J. J., 1946. — Stages in the normal development of Rana pipiens larvae.

Anat. Rec., 94: 7-23.

WHITEAR, M., 1974. — The nerves in frog skin. J. Zool., London, 172: 503-529.

Corresponding editor: Ulrich SINsCH.

E *USÉUX Source : MNHN, Paris

PAR

AIVTES

International Journal of Batrachology

published by ISSCA

EDITORIAL BOARD FOR 1996

Chief Editor: Alain Dunois (Laboratoire des Reptiles et Amphibiens, Muséum national d'Histoire

naturelle, 25 rue Cuvier, 75005 Paris, France).

Deputy Editor: Janalee P. CALDWELL (Oklahoma Museum of Natural History, University of Oklahoma,

Norman, Oklahoma 73019, U.S.A.).

Editorial Board: Jean-Louis ALBARET (Paris, France); Ronald G. ALriG (Mississippi State University,

U.S.A.); Emilio BALLETTO (Torino, Italy); Alain COLLENOT (Paris, France); Günter GOLLMANN (Wien,

Tim HaLLIDAY (Milton Keynes, United Kingdom); W. Ronald Hever (Washington,

U.S.A.); Walter HôbL (Wien, Austria); Pierre JOLY (Lyon, France); Masafumi Marsui (Kyoto,

Japan); Jaime E. PérauR (Mérida, Venezuela); J. Dale ROBERTS (Perth, Australia); Ulrich SINSCH

(Koblenz, Germany); Marvalee H. Wake (Berkeley, U-S.A.).

Technical Editorial Team (Paris, France): Alain DuBois (texts); Roger BouR (tables); Annemarie

OHLER (figures).

Index Editors: Annemarie OuLer (Paris, France); Stephen J. RicHaRDs (Townsville, Australia).

GUIDE FOR AUTHORS

Alytes publishes original papers in English, French or Spanish, in any discipline dealing with

amphibians. Beside articles and notes reporting results of original research, consideration is given for

publication to synthetic review articles, book reviews, comments and replies, and to papers based upon

original high quality illustrations (such as color or black and white photographs), showing beautiful or rare

species, interesting behaviors, etc.

The title should be followed by the name(s) and address(es) of the author(s). The text should be

typewritten or printed double-spaced on one side of the paper. The manuscript should be organized as

follows: English abstract, introduction, material and methods, results, discussion, conclusion, French or

Spanish abstract, acknowledgements, literature cited, appendix.

Figures and tables should be mentioned in the text as follows: fig. 4 or Table IV. Figures should not

exceed 16 x 24 cm. The size of the lettering should ensure its legibility after reduction. The legends of figures

and tables should be assembled on a separate sheet. Each figure should be numbered using a pencil.

References in the text are to be written in capital letters (BOURRET, 1942; GRAF & POLLS PELAZ, 1989;

INGER et al., 1974). References in the literature cited section should be presented as follows:

BoURRET, R., 1942. — Les Batraciens de l'Indochine. Hanoï, Institut Océanographique de l'Indochine: i-x

+°1-547, pl. I-IV.

Ga, J.-D. & POLLS PeLaz, M., 1989. — Evolutionary genetics of the Rana esculenta complex. In: R. M.

DawLey & J. P. BOGART (eds.), Evolution and ecology of unisexual vertebrates, Albany, The New York

State Museum: 289-302.

INGER, R. F., VoRis, H. K. & VOris, H. H., 1974. — Genetic variation and population ecology of some

Southeast Asian frogs of the genera Bufo and Rana. Biochem. Genet. 12: 121-145.

Manuscripts should be submitted in triplicate either to Alain Dumois (address above) if dealing with

amphibian morphology, systematics, biogeography, evolution, genetics or developmental biology, or to

Janalee P. CALDWELL (address above) if dealing with amphibian population genetics, ecology, ethology or

lie history. Acceptance for publication will be decided by the editors following review by at least Lo

referees.

If possible, after acceptance, a copy of the final manuscrit on a floppy disk (3% or 5 4) should

be sent to the Chief Editor. We welcome the following formats of text processing: (1) preferably, MS Word

(1 to 6.0, DOS or Windows), WordPerfect (4.1 to 5.1, DOS or Windows) or WordStar (3.3 to 7.0); (2) less

preferably, formated DOS (ASCII) or DOS-formated MS Word for the Macintosh (on a 3 V4 high density

1.44 Mo floppy disk only).

No page charges are requested from the author(s), but the publication of color photographs is

charged. For each published paper, 25 free reprints are offered by Alyres to the author(s). Additional reprints

may be purchased.

Published with the support of AALRAM

(Association des Amis du Laboratoire des Reptiles et Amphibiens

du Muséum national d'Histoire naturelle, Paris, France).

Directeur de la Publication: Alain DuBois.

Numéro de Commission Paritaire: 64851.

Alytes, 1995, 13 (3): 81-108.

Contents

Luis C. SCHIESARI, Britta GRiLLITSCH & Claus VOGL

Comparative morphology of phytotelmonous and pond-dwelling

larvae of four neotropical treefrog species (Anura, Hylidae,

Osteocephalus oophagus, Ostocephalus taurinus, Phrynohyas

resinifictrix, Phrynohyas venulosa)

109-139

Annemarie OHLER

Systematics, morphometrics and biogeography

of the genus Aubria (Ranidae, Pyxicephalinae) ..................... 141-166

Horst LÜDDECKE

Site fidelity and homing ability in Hyla labialis

CATUPASEYIITAE) MARS RAR A nn uit ea INSERT EE 167-178

Giovanni DELFINO, Rossana Bizz & Guido MELIS

Merocrine secretion from serous cutaneous glands

in Rana esculenta complex and Rana iberica ....................... 179-192

Announcements

Prize for best student paper in Alytes ......... 140

Free collections of A/ytes for life subscribers .… 140

Alytes is printed on acid-free paper.

Alytes is indexed in Biosis, Cambridge Scientific Abstracts, Current Awareness in Biological

Sciences, Pascal, Referativny Zhurnal and The Zoological Record.

Imprimerie F. Paillart, Abbeville, France.

Dépôt légal: 1° trimestre 1996.

Source : MNHN, Paris