Christian O. Buchacher

Institut für Zoologie der Universität Wien, Althanstraße 14, A-1090 Vienna, Austria

Abstract. A population of the Small Surinam toad, Pipa arrabali, was studied at an isolated pool site in undisturbed primary forest, 70 km north of Manaus, Brazil. During the eleven months of examination 371 individuals were toe-clipped. The toads preferred more persistent pools where they preyed upon tadpoles of other anurans. If a home pool dried up or a source of food became available in one of the other pools the toads migrated over land for short distances. Of a total of 84 recaptures, 28 (33%) were found in a pool other than the original. P. arrabali bred the entire year, but females with larvae in the dermal pockets on their backs were particularly frequent at the end of the rainy season (April to June). No free swimming larval stage occurs. Depending on the size of the female, between 6 and 16 eggs develop directly into small toads. Snout-vent length at emergence ranges from 10 to 14 mm.

Introduction

Female Neotropical pipids bear their offspring in small pouches on their back from which tadpoles or fully developed toads emerge. Trueb and Cannatella (1986) recognised seven species of Pipa: Pipa arrabali, P. aspera, P. carvalhoi, P. myersi, P. parva, P. pipa and P. snethlageae. Rabb and Snedigar (1960) and Rabb and Rabb (1960, 1963b) described the mating and egg-laying behaviour of Pipa pipa and Weygoldt (1976) that of P. carvalhoi. Schütte and Ehrl (1987) gave additional information on keeping and breeding of P. pipa. Papers on other Pipa spp. deal with distribution and taxonomy (Müller, 1914; Ruthven and Gaige, 1923; Müller, 1924; Noble, 1927; Dunn, 1948; Gines, 1958; Izecksohn, 1976; Trueb, 1984; Trueb and Cannatella, 1986; Cannatella and Trueb, 1988). Few field data are available for American pipids. Only Carvalho (1939) compared the habitats of P. carvalhoi and P. pipa.

Pipa arrabali was described by Izecksohn (1976). It is known from Guyana, Surinam, eastern Venezuela and northern Brazil and was first recorded for Central Amazonia in the Instituto Nacional de Pesquisas da Amazônia - World Wildlife Fund (INPA-WWF) reserves, located 70 km north of Manaus, Brazil (Zimmerman and Rodrigues, 1990). No study, to my knowledge, has ever focused on biology or ecology of this species. Trueb and Cannatella (1986) only discussed its taxonomic status because of the resemblance to P. aspera. They also report that the eggs develop directly into small toads, but they give no further information.

In this paper I describe phenological pattern of breeding and habitat use of P. arrabali. Special emphasis was given on brood care. Habitat use and brood care are compared to other Pipa-species.

Materials and Methods

Investigations were carried out in the INPA-WWF reserve "Gavião," 70 km north of Manaus, Brazil, and were part of the Biological Dynamics of Forest Fragments Project (Lovejoy et al., 1984; Lovejoy and Bierregaard, 1990). Forest, soil and geology of the study area were described in Lovejoy and Bierregaard (1990).

Pipa arrabali occurred at a site consisting of about 30 pools of varying size and degree of permanency. In November 1987 thirteen pools that in the experience of a local field assistant were the most persistent were selected for investigation (fig. 1). One pool (no. 18) was permanent, three others kept water for almost nine months (no. 19, 26, 35). Nine pools retained water only for short periods and dried up several times during the rainy season (fig. 2). The pools had no connection to streams and were exclusively supplied by rainwater. They ranged from a half to five square meters and had a maximal depth of one meter. They contained no water plants, but there was a substantial layer (~ 0.5 m) of rotting leaves. The bottom was sandy and the water slightly muddy. The water temperature ranged from 25.3°C (rainy period) to 26.2°C (dry period), and the average annual water temperature was 25.6°C.

Rainfall was seasonal, with a pronounced dry season. From December 1987 to June 1988 I measured 2314 mm of precipitation, in the drier months November 1987, July and August 1988, 193 mm, representing 7.5% of the rainfall of the study period were measured. Long term average data indicate that the precipitation during the study period was relatively high, but the rainfall was untypically distributed (fig. 2).

The more persistent pools were breeding sites of other anurans. Tadpoles of ten species occurred in the pools the entire year (Hylidae: Osteocephalus taurinus, Phyllomedusa bicolor, Ph. tomopterna, Ph. tarsius; Dendrobatidae: Colostethus marchesianus, Epipedobates femoralis; Leptodactylidae: Leptodactylus knudseni, L. rhodomystax; Microhylidae: Chiasmocleis shudikarensis, Ch. sp.). In addition, dip netting yielded fish of the genus Rivulus and the sub-family Callichthyinae, and water insects, mainly larvae of dragon-flies (Aeshnidae and Libellulidae).

Field surveys were carried out from November 1987 to September 1988. The selected pools were sampled weekly by two people with dip nets if they contained water. Dip netting turned up leaves and mud and made it difficult to locate the toads. Therefore, beginning with March 1988, in addition to dip netting, a trap was laid out once per week in each of the 13 pools. The traps were 30 x 30 cm square baskets made of fly-screen (synthetic). A funnel-shaped entrance was sewed into one side, with an outside opening of 50 mm diameter and inside opening of 20 mm. The basket traps were lined with wet leaves from the pools, baited with raw liver and placed in shallow water after nightfall. They were collected in the morning of the next day. I cannot exclude that some toads escaped during the night.

The toads were toe-clipped using a number code, and the snout-vent length (SVL) was measured. Pipa arrabali is capable of regenerating digits (pers. obs.). Juvenile toads (individuals with a SVL of less than 30 mm were assumed to be juvenile because the smallest female found with larvae in her back measured 32 mm) were able to regenerate fore and hind digits within a few weeks. Therefore juveniles were not toe-clipped. Full-grown toads regenerated totally clipped fore digits within a few weeks. Hind digits regenerated within three months if less than one phalange was clipped in order not to damage the inter-digital webbing. Hind digits apparently do not regenerate if one entire phalange is clipped at the joint, a method applied only during the last four months of the study (One Pipa arrabali held in captivity for more than two years clearly shows its marking.). Thus, until May 1988 the identification of recaptures was only successful during the first three months after toe-clipping. Individual identification of toads was only possible for a few weeks after toe-clipping, then the fore digits were regenerated.

I could detect no sexual dimorphism, unless a female bore larvae in the dermal pockets of her dorsum.

Results

Behavioural observations

During the day I observed individuals only when they swam to the surface to gulp air. They then immediately returned to the bottom of the pool where they hid in the leaf-layer. At night the animals swam slowly above the bottom of the pools with outstretched arms groping their way through the leaves. P. arrabali mainly fed on tadpoles which it discovered swimming around. Snapping at food was followed by a grasping movement. With a kick of their hind legs the toads jerked backwards. A similar movement was observed by Weygoldt (1976) on Pipa carvalhoi.

When I set out a baited trap on the edge of a pool, the toads immediately became excited and began to swim towards the bait. The animals repeatedly swung their fore legs in rapid movements from the sides to their mouths, as it was described for Xenopus (Russel, 1954) and Pipa carvalhoi (Weygoldt, 1976).

When a toad encountered a piece of liver it snapped at it and devoured it entirely using its fore limbs. The hind feet were employed in pulling and tearing at larger bits of liver. Debris was removed from the mouth, again using the hands. Motion of a prey-sized object did not appear to be a necessary visual stimulus for P. arrabali, as it is known for other pipids (Hymenochirus boettgeri, Rabb and Rabb, 1963a). During transport, or after frequent handling, the toads disgorged larger pieces of food.

Captures, Recaptures and Migrations

Between November 1987 and September 1988 I toe-clipped 371 toads. 156 (42%) were caught in the single permanent pool, 143 (38%) in the three semi-permanent pools (no. 19, 26 and 35) and 68 (18%) in the remaining nine ephemeral pools (table 1). The SVL of the majority (77%) of toe-clipped toads ranged from 30 and 44 mm. The largest individual was a female measuring 57 mm.

84 individuals, this is 23% of all toe-clipped toads were recaptured. 40 recaptures (48%) were made in the single permanent pool, of which 36 had also been toe-clipped there, representing 23% of all captures made in the permanent pool. Of the remaining 44 recaptures, 28 (63.6%) were toads found in a pool other than the original (table 1, fig. 1).

The toads generally migrated to adjacent pools, only in two cases I measured a distance of about 40 meters (fig. 1). The toads not only migrated to other pools when their home pool had dried up (for example, from nine toads recaptured in pool no. 35 in August 1988 six originated from pool no. 33 which had dried up one week before; fig. 1), they also frequently moved between water containing pools. I observed individuals after heavy rainfalls in flooded pools which lacked tadpoles and which I knew to hold water for only a few days, and in 20x20 cm puddles which contained only wet leaves and mud.

At the beginning of the dry season (August and September 1988) I observed a concentration of Pipa in the few remaining pools that contained water (fig. 2) leading to higher capture rates and more recaptures (75% of all recaptures were made within the last four months; fig. 3). But at the beginning of the study, in November 1987, when only pool no. 18 contained water, the number of captures was low and no recaptures occurred. In the leaf-layer on the bottom of dried up pools, I could find no pipids. I also never found P. arrabali in streams.

The role of Pipa arrabali as predator and prey

During the night of March 28, 1988 I observed the larvae of Phyllomedusa bicolor dropping from their nest, a leaf coiled to form a funnel, into the permanent pool. Within a few minutes, approximately 30 pipids were swimming around excitedly, feeding on the tadpoles. Most of the tadpoles were devoured as soon as they touched the water surface. Phyllomedusa spp. only bred in the more persistent pools (table 1).

I found P. arrabali in foam nests of Leptodactylus knudseni. During the eleven months of examination nine foam nests occurred in the pool site, two of which contained three, respectively four adult P. arrabali. Both nests were laid in wet leaf litter and had no connection to open water (fig. 1). In one nest I additionally found about 30 larvae of L. knudseni, 20 mm long. The toads were still present after one night (January 21 to January 22, 1988) although all the tadpoles had disappeared.

In the stomachs of three Pipa arrabali I found tadpoles of Phyllomedusa bicolor, Ph. tomopterna, Leptodactylus knudseni, Colostethus marchesianus, and indeterminable rests of tadpoles. In another individual I found a 3 cm long and 1 cm wide piece of an earthworm. The stomach contents of one toad could not be identified and the stomach was empty in two other toads.

I observed an individual of Pipa arrabali to be eaten by a spider of the family Ctenidae and another one by a turtle (Platemys platicephala).

Brood care

I did not observe the mating and egg-laying behaviour of Pipa arrabali in the field.

I found females with larvae in the dermal pockets of their backs and freshly metamorphosed toadlets during the entire study period, but they were particularly frequent in April 1988 (fig. 3). The number and size of toadlets in the pockets on the back was correlated to the SVL of the female (table 2). The smallest female containing larvae in her back measured 32 mm SVL. The dermal pockets were covered in all stages of development and were restricted to the posterior region of the dorsum. The lids to the pouches are thick cuticular structures (pers. obs.).

I observed the emergence of toadlets in the laboratory. The female measured 55 mm SVL and weighed 18,55 g. I counted 16 pockets in the dorsal skin, three of them were already empty when the observation started (6 p.m.). At 7.10 p.m. the leg of another toadlet emerged from one of the pockets, the whole toad emerged at 7.27 p.m. The next two toadlets emerged at 7.40 and 7.54. Emergence of the first lasted 7 minutes and of the second 9 minutes. In both cases, it started with one of the hind legs. The three toadlets floated at the water surface for two minutes, then two dived and hid under their mother. One was unable to dive and died soon afterwards. By midnight five more toads emerged from the dermal pockets. The last four of the 16 emerged between midnight and 2 a.m. The toadlets measured 13 mm SVL and weighed between 0,224 and 0,229 g.

Discussion

Site fidelity and migrations

Pipa arrabali preferred the more persistent pools, where the risk of drying out was low and the food supply of tadpoles of other anurans was more continuous. However, the toads did not exhibit great fidelity even to these sites. Of 156 toe-clipped toads in the single permanent pool only 23% were recaptured. Although I could identify amputated phalanges for at least three months, recaptures always occurred within three weeks after toe-clipping. The toads apparently left the pool within three weeks after the toe-clipping to migrate to other sites. Another reason may be that the disturbance connected to the amputation of toes caused a higher migration activity (Seidel, 1987).

In the less persistent pools the majority of recaptures (63.6%) occurred in another pool than the original. The less persistent pools apparently were frequented for only a few days. The toads constantly migrated between water containing pools to search for sources of food such as Leptodactylid foam-nests. Recently flooded pools that lacked tadpoles possibly provided short-term sources of terrestrial organisms as food.

Habitat use in comparison to other Pipa species

Pipa pipa is distributed throughout the entire drainage area of the Amazon. According to Carvalho (1939), it inhabits permanent aquatic sites such as rivers, oxbow lakes and flooded forests of large rivers (Várzea, Igapó). Both P. pipa and P. arrabali are found near Manaus. The larger Surinam toad is restricted to the flood plains, and I regularly found it in the city sewers of Manaus. P. arrabali has not been recorded for the urban area of Manaus and does not occur in the "Reserva Florestal Adolfo Ducke" which is 13 km away (Magnusson, pers. comm.). It is exclusively distributed in the Terra Firme forest, i. e. forest not subject to extensive flooding, where it inhabits small ephemeral pools. Zimmerman and Rodrigues (1990) assumed also streams and stream side pools as "probable" breeding habitats, but I never found P. arrabali in other waters than isolated upland pools.

P. pipa is adapted to large and permanent aquatic sites. It is strongly flattened and has no claws. On land it is helpless and scarcely able to move (Drewes, 1977). Under natural conditions, it lies motionless in wait for prey well disguised by its brown coloration (Rabb, 1969) and does not voluntarily leave the water (Carvalho, 1939; Weygoldt, 1976). The diet of adult P. pipa in the wild includes fish as well as crayfish (Carvalho, 1939; Rabb, 1969). Spreading occurs through the water. Here fish-like larvae would be disadvantageous, because they would have to compete with fish and could be perceived as prey by their parents (Weygoldt, 1976).

P. arrabali is adapted to temporal aquatic sites. The body is only slightly flattened and the three inner toes are clawed. It leaves the water frequently to migrate to other pools and is able to move quickly on land. It mainly feeds on tadpoles and actively hunts for prey. The advantage of brood care is even more obvious in P. arrabali, because the larvae would have to compete with the other tadpole species in the pools and risk being preyed upon by their tadpole-eating parents. Pipids with a free-swimming larval stage (f.e. Xenopus, Hymenochirus) prey upon their offspring under captivity conditions (pers. obs.).

Trueb (1984) collected Pipa myersi in lowland forest swamps in eastern Panama that are very similar to the habitat of P. arrabali. It is remarkable that, under similar conditions, P. arrabali evolved a direct development and P. myersi a development with free-swimming larvae.

P. carvalhoi inhabits small eutrophic pools, ponds and ditches in the dry steppe (Zona Agrestre and Caatinga) of the Brazilian states Pernambuco, Ceará and Paraíba (Carvalho, 1939). Like P. arrabali, it may leave the water and wander about on dry land. Meyers and Carvalho (1949) commented that P. carvalhoi was hopping about on land even in broad daylight "apparently foraging for insects," whereas P. arrabali is nocturnal. Under drier conditions P. carvalhoi may dig itself into wet mud (Weygoldt, 1976).

Comparison of the brood care of four species of Pipa

In the field, I found P. arrabali females with six to sixteen eggs embedded in the pouches on their dorsums (table 2). Greater numbers of eggs have been recorded for other Pipa-species. Drewes (1977) reported 190 embedded eggs for P. pipa and Schütte and Ehrl (1987) 192 and 193. Rabb and Snedigar (1960) found 76, while Rabb and Rabb (1960) observed only 55. On P. carvalhoi Weygoldt (1976) counted 80 to 200 eggs.

The diameter of P. arrabali eggs ranged from 3.0 to 3.2 mm. Rabb and Snedigar (1960) found the egg-diameter for P. pipa to be 6.4 to 6.6 mm, thus almost twice the size of P. arrabali eggs. For P. carvalhoi Weygoldt (1976) found an egg-diameter of only 2.4 mm, which is not surprising, since this species develops with a larval stage and in comparison to its moderate body size produces a high number of eggs. Carvalho (1939) gave an egg-diameter of 2.6 mm for P. carvalhoi. Gines (1958) reported on a P. parva female with 2 mm eggs in the pockets on her back. He kept it in an aquarium and 12 tadpoles emerged.

Whereas P. arrabali eggs are lodged only in the posterior region of the dorsum, in P. pipa after oviposition the entire dorsum from the vent to the head is swollen and full with adhering eggs (Schütte and Ehrl, 1987: fig. 3). The same applies to P. carvalhoi (Weygoldt, 1976: fig. 3F) and P. parva (Gines, 1958: fig 2).

Whereas in P. arrabali the lids to the pouches are thick cuticular structures in P. pipa the outer egg-capsules serve as lids (Noble, 1927; Rabb and Snedigar, 1960). Most of the coverings are lost after the shedding of the female (Rabb and Snedigar, 1960), thus making it possible for the small toadlets to peer out of the pockets (Schütte and Ehrl, 1987: fig. 7). In this stage they already can be fed through the openings of the pockets with Tubifex or Daphnia (Schütte and Ehrl, 1987). I have no records that P. arrabali-females with larvae in their dorsum leave the water, but I suspect the cuticular coverings to be a protection against desiccation during migration.

Freshly emerged toadlets of Pipa arrabali measured between 10 and 14 mm (table 2). Rabb and Snedigar (1960) gave a SVL for young P. pipa at emergence of 16 to 18 mm, which is only slightly longer than freshly emerged P. arrabali, although P. pipa-eggs are almost twice as large.

Acknowledgements. I wish to thank W. Hödl and B. Zimmerman for valuable assistance throughout this project, F. Schütte, M. Roithmair and H. Tunner for reading and commenting on the manuscript and O. Pereira for assistance in the field. This study was supported by a grant from the Austrian Ministry of Science and the World Wildlife Fund - US. This is publication # 96 of the Biological Dynamics of Forest Fragments Project.

References

Cannatella, D. C. and Trueb, L. (1988): Evolution of pipoid toads: intergenetic relationships of the aquatic toad family Pipidae (Anura). Zool. Journal of the Linnean Society 94: 1-38.

Carvalho, A. L. de (1939): Notas sobre Hemipipa carvalhoi MIR. RIB. (Batrachia, Anura, Pipidae). 1^a parte. Bol. Biol. São Paulo, n. ser 3: 397.

Drewes, R. C. (1977): Surinam Toad. Pacific Discovery, Cal. Acad. of Sciences, vol. 3, No. 4: 26-29.

Dunn, E. R. (1948): American toads of the family Pipidae. Amer. Mus. Novitates No. 1384: 1-13.

Gines, H. (1958): Representantes de la familia Pipidae (Amphibia: Salientia) en Venezuela. Mem. Soc. Cien. Natur. La Salle, 18 (1): 5-18.

Izecksohn, E. (1976): Uma nova espécie de Pipa do Estado do Amazonas, Brasil Amphibia, Anura, Pipidae). Rev. Brazil. Biol., Rio de Janeiro, 36 (2): 507-510.

Lovejoy, T. E., J. M. Rankin, R. O. Bierregaard, K. S. Brown, L. Emmonds, M. E. Vandervoort (1984): Ecosystem decay of Amazon forest remnants. M. H, Nitecki (ed.) Extinctions, Univ. Chicago Press: 295-325.

Lovejoy, T. E. and R. O. Bierregaard (1990): Central Amazonian Forests and the Minimum Critical Size of Ecosystem Project. Four Neotropical Rain-forests/Alwyn H. Gentry, editor, Univ. Yale Press: 60-71.

Myers, G. S. and de Carvalho A. L. (1949): Notes on some new or little known Brazilian amphibians, with an examination of the plata salamander, Ensatina platensis. Bol. Mus. Nacional, Rio de Janeiro, No. 35: 1-24.

Müller, L. (1914): On a new species of the genus Pipa from northern Brazil. Ann. Mag. Nat. Hist. 14: 102.

Müller, L. (1924): Neue und seltene Reptilien und Batrachier der Zoologischen Sammlung des bayerischen Staates. Zoolog. Anzeiger 58, Leipzig 1924.

Noble, G. K. (1927): The value of life history data in the study of the evolution of the Amphibia. Ann. New York Acad. Sci. 30: 31-128.

Rabb, G. B. (1969): Frogs and pipid frogs. Brookfield Bandarlog 37:3.

Rabb, G. B. and M. S. Rabb (1960): On the mating and egg- laying behaviour of the Surinam Toad Pipa pipa. Copeia 1960: 271-276.

Rabb, G. B. and M. S. Rabb (1963a): On the behaviour und breeding biology of the African pipid toad Hymenochirus boettgeri. Z. Tierpsychol. 20: 214-241.

Rabb, G. B. and M. S. Rabb (1963b): Additional observations on breeding behaviour of the Surinam Toad Pipa pipa. Copeia 1963: 636-642.

Rabb, G. B. and R. Snedigar (1960): Observations on breeding und development of the Surinam Toad Pipa pipa. Copeia 1960: 40-44.

Russel, W. M. S. (1954): Experimental studies of the reproductive behaviour of Xenopus laevis I. The control mechanisms for clasping und unclasping, and the specificity of hormone action. Behaviour 7: 113-188.

Ruthven, A. G. and H. T. Gaige (1923): Description of a new species of Pipa from Venezuela. Occ. Pap. Mus. Zool. Univ. Michigan 136: 1-2.

Schütte, A. and A. Ehrl (1987): Zur Haltung und Zucht der großen südamerikanischen Wabenkröte Pipa pipa Lineanus, 1758 (Anura: Pipidae). Salamandra 23 (4): 256-268.

Seidel, B. (1987): Breeding of a Bombina variegata population in a habitat with temporary pools. Proc. Fourth ord. gen. meet. S.E.H., Nijmegen.

Trueb, L. (1984): Description of a new species of Pipa (Anura, Pipidae) from Panama. Herpetologica 40 (3): 225-234.

Trueb, L. and D. C. Cannatella (1986): Systematics, morphology and phylogeny of the genus Pipa (Anura, Pipidae). Herpetologica 42 (4): 412-449.

Weygoldt, P. (1976): Beobachtungen zur Biologie und Ethologie von Pipa (Hemipipa) carvalhoi MIR. RIB. 1937 (Anura, Pipidae). Z. Tierpsychol. 40: 80-99.

Zimmerman, B. L. and M. T. Rodrigues (1990): Frogs, Snakes, and Lizards of the INPA-WWF Reserves near Manaus, Brazil. Four Neotropical Rain-forests/Alwyn H. Gentry, editor, Univ. Yale Press: 426-454.

Table 1. Local distribution of Pipa arrabali in comparison with the reproduction of other anuran species. Pta: Phyllomedusa tarsius, Pbi: Ph. bicolor, Pto: Ph. tomopterna, Lep: Leptodactylus knudseni, tot: total number of clutches FN Leptodactylid foam nest in unnumbered pool.

Pool no.	Captures	Thereof toe-clipped	Recaptures	Thereof captured in	Clutches of other anurans
				other pools	PtaPbiPtoLeptot
18	209	156	40	4	1433424
19	37	21	7	5	----1--1
20	11	9	2	1	----------
21	3	1	--	--	----------
22	3	3	--	--	----------
23	4	3	1	--	----------
24	12	4	--	--	----------
25	2	2	--	--	----------
26	67	48	16	10	211--4
32	54	13	2	1	------11
33	39	28	--	--	1--113
34	7	5	--	--	----------
35	83	74	16	7	11--24
FN	4	4	--	--	------11
total	535	371	84	28	1856938

Table 2. Size of egg-bearing Pipa arrabali. Number and size of toadlets immediately after emergence. ^* preserved in formaldehyde with larvae in the dermal pockets on the dorsum.

No.	SVL of female (mm)	Number of embedded eggs	SVL of toadlets after emergence (mm)
1	32	6	10 - 11
2	37	6	--*
3	43	6	12
4	40	8	10 - 12
5	45	8	10 - 12
6	47	12	12 - 13
7	53	14	10 - 11
8	55	12	12 - 14
9	55	16	13

Figure 1. The 13 surveyed pools in the study site near Manaus, Brazil. Bold numbers represent the pool number, the first number in parentheses indicates the total number of captures per pool, the second, the number of toads recaptured in the same pool. Numbers associated with an arrow indicate toads recaptured in a pool other than the original. The arrows show the direction of migration based on recapture. The pools on the right side and the pools on the left are separated by a 30 m closed forest area that lacks pools.

Figure 2. Monthly distribution of rainfall in the study area in comparison with long term average precipitation (Projeto Radambrasil 1945-1975, Folha SA20 Manaus, Ministério de Minas e Energia) and permanence of the 13 surveyed pools (below). The monthly rainfall was measured in the INPA-WWF reserve "41" (distance to the study site 14 km), from November 1, 1987 to August 31, 1988. For September 1988 no data were available.

Figure 3. Temporal distribution of Pipa arrabali in the 13 selected pools of the study site, between November 1987 and September 1988. Beginning with March 1988 (arrow) traps were laid out. ^* adult females with larvae in dermal pockets on their dorsums.