Foam nest in Scinax rizibilis ( Amphibia : Anura : Hylidae )

During the intervals of February 1993 to January 1994 and November 1994 to February 1995, in the southern São Paulo state, we studied the breeding activity of Scinax rizibilis (Bokermann, 1964), the only known hylid species with oviposition in foam nests. The foam nests were constructed by female jumps, during the oviposition. The clutches contained 850-1250 eggs, which were almost black, except for the small clear vegetative pole. The construction of foam nest in S. rizibilis is unique among the other species with this characteristic. The complexity of a foam nest is intermediate, and egg development was faster when eggs were surrounded by foam. It is possible to recognize a progression from less developed structures, represented by the bubble nests of some microhylid frogs, to more complex examples, such as the foam nests of Leptodactylidae or Leiuperidae.


MATERIAL AND METHODS
Scinax rizibilis was observed in a temporary pond at the Fazendinha São Luiz (24°21'S, 48°44'W, 800 m altitude) in the municipality of Ribeirão Branco, southern São Paulo state, Brazil, from February 1993to January 1994, and from November 1994to February 1995.We visited the pond either fortnightly or monthly, and monitored it for 2-6 nights during each visit, totaling 148 hours in 46 visits.The pond area was approximately 1,950 m2 and the distribution of vegetation was regular with predominance of Juncaceae.The pond was bordered by typical Atlantic Forest flora.
Nocturnal observations were conducted with a 6 V spotlight covered with sheets of thin red plastic to reduce the stress on the animals (ROBERTSON 1990).Focal-animal, all occurrences, and sequence samples were used for behavioral records (LEHNER 1996).
Pairs found in amplexus were collected manually.We measured the snout-vent length (SVL) of individuals to the nearest 0.1 mm with a caliper ruler and weighted them with a Pesola ® balance to the nearest 0.05 g.The clutches obtained were preserved in 5% formalin.Ten eggs of each clutch were measured under a stereomicroscopic using a micrometric ocular.
The pairs in amplexus (n = 5) were put into separate aquariums (25 x 8 x 20 cm) with water at a depth of 3 cm.The subjects were filmed with a video camera (two pairs) and pho- In order to test the foam influence on egg development, we monitored four egg masses over 24 hours, two with foam and two without foam, maintained in a plastic recipient (10 x 10 x 10 cm) with water at a depth of 8 cm.We extracted some eggs from the foam nest with a fine-mesh dipnet.Developmental stages following GOSNER (1960) were determined using a stereomicroscopic.For statistical analyses, we used Pearson correlation coefficient (ZAR 1996) with the significance level of 0.05.
Foam nest construction (n = 2) lasted 38 or 40 minutes, and was performed by females during the amplexus (Fig. 4).The beginning of the oviposition is marked by a circular swimming by the females.The male pushed its two feet very close, forming a channel between its cloaca and the female cloaca.After that, the pigmented eggs appeared.Since the mucus secreted by the reproductive traits of the females is transparent, we were not able to observe the exact moment of its release.
This sequence of oviposition was repeated several times.During the entire process the female performed alternate movements of the legs to join mucus and eggs.After egg expulsion, the female jumped up in the water and the impact of its body against the water allowed the retention of air bubbles in the mucus, forming the foam nest.The end of oviposition is marked by a characteristic posture signalization of the female; the female arched its back inwardly, the head was elevated (at 45° to it body), and legs and arms were distended.Then, the male slipped laterally off the female body and the process of oviposition ceased (Fig. 4F).The egg development was faster when eggs were surrounded by foam (Fig. 5).

DISCUSSION
The clutch size is positively correlated with SVL and female mass, as observed in other Neotropical anuran species (MARTINS 1988, BASTOS & HADDAD 1996), according to the model  HADDAD et al. (1990), in his description of the foam nest of S. rizibilis (as Hyla cf.rizibilis), suggested the females have an active part in nest construction; herein we confirm it.The characteristic posture of the female for signalizing the end of oviposition was observed in the leiuperid Physalaemus cuvieri (J.P. Pombal Jr pers.obs.), and may be common in anurans.
The function of foam nests is controversial.Some authors have suggested that it might: 1) reduce exposure to aquatic predators (HEYER 1969); 2) protect the eggs from desiccation (HÖDL  & GETTINGER 1985); and 4) supply oxygen for eggs and embryos (SEYMOUR 1999, SEYMOR & LOVERIDGE 1994), accelerating development (HADDAD & HÖDL 1997).HADDAD et al. (1990) suggested that foam nests in Scinax rizibilis evolved mainly as a protection against insolation and desiccation of eggs and embryos.However, the data obtained in this study corroborates HADDAD & HÖDL (1997), because eggs without foam (little oxygen) developed slower than eggs with foam (more oxygen) (Fig. 5).Development acceleration and protection are not mutually exclusive functions.Therefore, we cannot say thatthe foam nests of S. rizibilis only provide either one of these two benefits.For example, eggs of species that breed in temporary ponds, such as S. rizibilis, hatch quickly into larvae, decreasing the risk of desiccation and predation by conspecifics (HÖDL 1992).It is possible to recognize a sequence from a less elaborate floating device, represented by the bubble nest, to an elaborate structure represented by the foam nest (Tab.I).The foam   & DAVIES 1979).The more complex foam nests are those built by Hyperolidae, Leiuperidae, Leptodactylidae, and Rhacophoridae whose eggs are surrounded by many small air bubbles, being deposited on the water surface (HEYER 1969, HÖDL 1992) or on leafs/litter (JENNIONS et al. 1992, KADADEVARU & KANAMADI 2000).
As stated by HEYER (1969), the pre-adaptations to foam nest construction are widespread among anurans because many species are able to secrete mucus during oviposition.HADDAD et al. (1990) manually beat the mucus of Scinax hiemalis and obtained a foam nest.The fact that foam nests are known from seven anuran families and that the construction procedures differ among them indicates that the evolution of the foam nest may have originated independently among these families.
Figures 1-3.Relationship between the: (1) snout-vent length and number of eggs of females of S. rizibilis; (2) mass and number of eggs of females of S. rizibilis; (3) diameter and number of eggs of S. rizibilis.

Figure 4 .
Figure 4. Stages of foam nest construction in S. rizibilis.Drawing based in slides.

Figure 5 .
Figure 5. Developmental stages of tadpoles of S. rizibilis with and without foam.

Table I .
(HADDAD & HÖDL 1997)nuran foam nest of selected species.producedbyS.rizibilis is intermediate between the two extremes(HADDAD & HÖDL 1997).The less complex foam nests are built by Limnodynastidae, Microhylidae, and Myobatrachidae, which deposit eggs on the water surface surrounded by few large air bubbles which are produced by male and female(HADDAD & HÖDL 1997), or by females that paddle the forelimbs to start a flux of water (TYLER nest