AMPHIBIAN EGGS AND THE MONTANE ENVIRONMENT

Transcription

1 AMPHIBIAN EGGS AND THE MONTANE ENVIRONMENT OLIVE B. GOIN AND COLEMAN ]. GOIN 1 Department of Biology, University of Florida, Gainesville Received December 17, 1961 While the question of the origin of the amniote egg is not susceptible to direct proof, it is at least possible, by a consideration of various indirect lines of evidence, to arrive at a conclusion as to the most probable cause of this major evolutionary step. The traditional viewpoint has been that the animals that developed the amniote egg were already fully terrestrial and that the selective advantage of the cleidoic egg lay in the release of the adults from the necessity of returning periodically to water to breed. Romer (1957) points out that the earliest reptiles were still semiaquatic as were their amphibious ancestors. His argument that the terrestrial egg evolved before the terrestrial adult seems valid. Romer suggests that the land egg arose under conditions of periodic drought through the selective pressure resulting from desiccation of the eggs and larvae in ponds subject to drying. Tihen (1960) accepts Romer's basic thesis that the terrestrial egg preceded the terrestrial adult, but points out the inherent improbability that semiaquatic animals would have survived for long in regions where there was not a sufficiently dependable water supply to allow for aquatic development of eggs and larvae. Tihen further stresses that the modern amphibians that have terrestrial reproduction live in humid environments and that the basic amphibian adaptation to an intermittent water supply seems to be an acceleration of the developmental period. Tihen believes that reptiles evolved in humid, tropical or subtropical swamps, and that the selective force leading to the development of the terrestrial egg was heavy predation on the vulnerable eggs and early larvae. This suggestion has also been made by Lutz 1 Work supported by a grant G-13325, from the National Science Foundation. EVOLUTION 16: September, (1948). The egg came ashore, not to avoid desiccation, but to avoid being eaten. Consideration of the ecologic distribution of the modern amphibians that lay terrestrial eggs suggests still a third possibility as to the conditions under which the reptiles evolved from the amphibians. This is that the event took place in a mountainous region. In such a region, there are few broad, open expanses of quiet water. The most usual aquatic habitat is a cool, rapidly flowing stream in which sperm, eggs, and larvae are all exposed to the danger of being swept away by the current. Reproduction in such an environment requires modifications, both behavioral and structural, in animals that presumably evolved in quiet waters. Many such adaptations are known for modern amphibians that breed in rapidly flowing water. The eggs are adhesive or are laid under stones; the larvae are streamlined, or flattened for crawling over rocks, or have special suctorial organs. It seems evident, though, that one very common adaptation to a montane habitat is the development of terrestrial breeding habits. It should be stressed that in the following discussion, montane habitat refers more to ruggedness of relief than to altitude. Where mountains rise abruptly from the sea coast, as they do especially on many islands, typically montane aquatic habitats may occur at rather low elevations. There are three prerequisites for the development of the cleidoic egg. One is internal fertilization, since the sperm must enter the egg before the shell is deposited around it. Another is increase in amount of yolk, since the embryo must have a sufficient supply of nutriment to allow it to develop to the adult body form without going through a free-swimming, feeding, larval stage. And the third is simply that

2 AMPHIBIAN EGGS 365 the egg be laid, and development take place, on land. Although the first two steps may have preceded the third, we shall here emphasize the latter, since this was the major adaptive shift that marked the evolution of the reptiles. TERRESTRIAL BREEDING HABITS Modern amphibians show many intermediate stages in the development of true terrestrialism. Sometimes the eggs are simply deposited in a burrow in the bank or on a leaf above a stream and the larvae on hatching fall into the water below. Sometimes the eggs or larvae are transported to the water by one or another of the parents. Sometimes the larvae metamorphose within the nesting site without ever moving to open water. And sometimes the larval stage is omitted entirely and the young hatch, or are born, with the adult body form. In the following survey by families of amphibians with terrestrial breeding habits, we have not attempted to include all forms in which terrestrialism is only partially developed. We have stressed those amphibians that lack entirely the free-swimming, aquatic larval stage. In order to avoid bias, we first compiled a list of such amphibians and then went back and added such data as we could find on the distribution of these forms. The list is not complete and could probably be expanded considerably, particularly if more information on breeding habits were available for some of the tropical forms. Family Caecilidae.-Although details are unknown for many species, apparently reproduction in the Caecilidae is primarily terrestrial. They have internal fertilization, and may either lay eggs or bear living young. As far as is known, the eggs are always relatively large-yolked and are laid on land. The young hatch, or are born, either as fully metamorphosed individuals or as larvae that must make their way to water to complete their development. Of the American forms, Dunn (1942) says that the center of the group range is northwestern South America, which is also the area of greatest abundance of genera and species. This is the region of the northern Andes. "In North America the only temperate areas inhabited are montane cloud forest; it is probable that the animals occur in savanna only in galeria forests along rivers" (Dunn, 1942). No summary of distribution is available for African forms, but scattered locality records suggest a primarily montane habitat. Of the nine forms reported by Loveridge (1957) from East Africa, two are from lowland localities and seven from mountains. Stolomorphis kirki of Nyasaland is reported to be "restricted to the leaf mould of the floor of the mountain forest" (Mitchell in Rose, 1950). Ichthyophis of Asia lays eggs on land but the young hatch as larvae and must finish their development in water. It inhabits hilly parts of southern India and Ceylon, the eastern Himalayas, Malaya, the East Indies, and the Philippines (Boulenger, 1912). In Malaya it is found up to 1,200 m, and in the Philippines it reaches elevations of 900 m (Inger, 1954). Family Salamandridae.-The terrestrial Black Salamander, Salamandra atra, bears from two to four fully metamorphosed young. This salamander is found only in the Alps, from 800 to 3,000 m (Angel, 1947). Family Plethodontidae.-These are the most terrestrial of the salamanders. A number of the genera (e.g., Plethodon, Ensatina, Aneides) lay eggs on land and have no aquatic larval stage. Dunn (l926a) says: "The family is then a group derived from primitive Salamandrids... specialized beyond them in adaptation to mountain brook life... Appalachia is the center of the group range, it is a mountain region..." Dunn's map of the distribution of the family shows well its primarily montane distribution. Family Leiopelmidae.-These primitive frogs live near the mountain tops of New Zealand. The eggs are laid on land (though they are apparently also able to develop in water) and the young hatch out

3 366 OLIVE B. GOIN AND COLEMAN J. GOIN as metamorphosed frogs, though still possessing tails (Stephenson, 1955).? Family Pelobatidae.-A. P. Butler twice found in the Larut Hills of Malaya clusters of about a dozen large frog eggs under damp moss on tree trunks. These eggs contained tadpoles with well-developed hind limbs. Butler believed the eggs to be those of M egaphrys longipes, which he had found to be the commonest frog in the hills above 900 m, and which lacks an aquatic larval stage (Boulenger, 1912). While the identity of these eggs has not been confirmed, they do indicate the presence in the mountains of Malaya of a frog with terrestrial breeding habits. Family Ranidae.-The little African Chirping Frog, Arthroleptella lightfoati, lays relatively large eggs on damp moss. The young, which hatch after the hind legs have developed, wriggle about on the moss for a few days before metamorphosing, but are apparently helpless when dropped into water (Rose, 1950). Rose says: "He appears to be confined to mountainous districts, where as a rule the only available water is a rushing torrent which the tadpole of such a tiny frog could not possibly stem. In this probably lies the explanation of the unusual breeding habits..." Another South African ranid, Anhydrophryne rattrayi, lays about twenty eggs in a hole in the ground. The young metamorphose within the chamber and the tadpole will drown if placed in water. This species is found in the Amatola Mountains of Cape Colony (Rose, 1950). Several species of the African genus Arthroleptis have been reported to be terrestrial breeders. A. stenodactylus, which has a wide altitudinal distribution in East Africa, lays its eggs in burrows away from bodies of water and, hence, presumably lacks an aquatic tadpole (Loveridge, 1957). A. wageri of southern Africa lays large eggs under leaves on the ground and has no free-swimming tadpole (Rose, 1950). We have no data on the altitudinal distribution of this species. An interesting group of ranid frogs is found in the Solomon Islands. Three of the genera (Batrachylodes, Ceratobatrachus, and Palmatorappia) occur nowhere else. Discodeles is found also in the Bismarcks and Admiralties. Cornufer ranges from Burma and the Philippines to the Fiji Islands and Platymantis (Cornufer, d. Inger, 1954) from Borneo and the Philippines to Fiji. Of these frogs, Brown (1952) says: "The close relationship between Ceratobatrachus, Cornufer, Discodeles, and Platymantis is shown... also in certain reproductive modifications. The eggs of species within these genera, as far as I have been able to observe (fourteen of twenty-nine recognized species) are relatively large and unpigmented. This suggests that the larvae of all the species may complete their development within the egg capsule as is, indeed, known for Cornufer guentheri and Discodeles opisthodon. Maturing, ovarian eggs of Palmatorappia are unpigmented. "... eggs of B [atrachylodes] trossulus are relatively large and unpigmented and a modification of developmental habits similar to that of the genera Cornufer and Discodeles is also suggested for this genus." The Solomon Islands are mountainous, the larger ones reaching elevations of 3,000 m. They have a humid, tropical climate with heavy rainfall and are largely forested, although extensive grassy areas are found on some of the larger islands. Streams are short, often large, and generally rapid, and freshwater lakes are few. Geological evidence suggests that formerly the islands were even more steep and rugged. Brown (1952) suggests that the reproductive specializations of these ranids "could have enabled these frogs to maintain themselves near their present center in the Solomon Islands at some time in the past, under conditions when permanent or semipermanent bodies of still or slow moving water may have been less readily available as breeding sites." Of the species' of Cornu fer (+Platymantis) in the Philippines, Inger (1954) says: "... what little is known indicates that all

4 AMPHIBIAN EGGS 367 species of Cornufer deposit large, non-pigmented eggs... probably out of water. The few larvae that are known do not pass through a free-swimming stage; the character of the eggs suggest that none do." The Philippine Islands are mountainous, and Inger's ecological notes suggest a montane habitat for five of the seven species listed. Family Rhacophoridae.-Rhacophorus microtympanum of Ceylon lays about twenty large eggs on land. The developing embryo resembles that of Eleutherodactylus and there is no tadpole stage. These frogs are found mainly in the mountains, between altitudes of about 1,200 to 1,800 m (Kirtisinghe, 1957). Family Microhylidae.-A number of microhylids show a strong tendency toward terrestrial reproduction. All members of the subfamilies Asterophryninae and Sphenophryninae for which data on breeding habits are available lay their eggs on land and omit the tadpole stage. These subfamilies are centered in the Papuan region, being most common on New Guinea, though Sphenophryninae ranges from the Philippines to northern Australia. They are primarily montane in distribution (Parker, 1934). On New Guinea Sphenophryne brevicrus was found "between 10,000 and 12,000 feet above timber line in shallow recesses under tussocks. One, together with 14 eggs, each about 5 mm. in diameter, was taken in moss under a tussock at 11,000 feet. A woodeny croaking call, presumably produced by this species, was heard as high as 13,000 feet" (Darlington in Loveridge, 1948). Inger writes of the Philippine Oreophryne annulata: "The eggs of annulata are laid in small clusters. Three such clutches, found in moss covering tree trunks on Mount McKinley by the Philippine Expedition, contain three, eight, and nine eggs, respectively. The eggs vary from 3 to 4 mm. in diameter; including the capsules, they measure about 6 mm. The larvae do not go through a free-swimming tadpole stage, but instead undergo metamorphosis within the egg capsule. The larvae of one Mount McKinley clutch were reared through to hatching in the field." Inger's notes indicate that he considers this to be primarily a montane species. Parker (1934) believes that the members of the subfamily Brevicipitinae are likewise terrestrial breeders. This is a small subfamily of frogs found mainly in the mountains of eastern and southern Africa. The first recorded clutch of eggs of the South African Breuiceps gibosus contained about 30 large eggs and was deposted in a burrow "well up on the mountain slopes" (Rose, 1950). H oplophryne of the subfamily Melanobatrachinae lays its large eggs in internodes of bamboo and banana axils, where there is little or no water. The young hatch as late larvae and metamorphose within the breeding site. H oplophryne inhabits the Uluguru, Usambara, and Magrotto mountains of Tanganyika (Parker, 1934). Plethodontohyla tuberata (subfamily Cophylinae) has been reported to undergo direct development in a burrow on land (Guibe, 1952). Guibe reports that this Madagascan species is found on the volcanic massif of Ankaratra. Family Bufonidae.-Alone among the anurans, N ectophrynoides bears living young which metamorphose before birth (Noble, 1931). The two species of this genus are found in the mountains of Tanganyika (Loveridge, 1957). Some other genera of bufonids show a trend toward terrestrialism. Thus Inger (1958) states: "Judging by the little we know of its breeding habits, Pelophryne probably is not aquatic at any time after metamorphosis. Its eggs are very large, non-pigmented, very few in number, and laid in the smallest rain-filled depressions in the forest floor. The tadpoles do not feed, subsisting entirely on the yolk supply of the large egg (Inger, in manuscript). Associated with this habit, the oral disk is much reduced and lacks functional labial teeth and beaks. The basic modification of

5 368 OLIVE B. GOIN AND COLEMAN J. GOIN Pelophryne is adaptation towards increasing terrestriality." Elsewhere Inger (1954) says: "Most of the species of Pelophryne come from highaltitude forests..." Family Hylidae.-The center of distribution of the hylids seems to be northwestern South America. Many of them show transitional steps toward the development of terrestrialism. Eggs may be laid on leaves above water or carried on the back of the female. When they are carried by the female, they may hatch either as tadpoles or as metamorphosed young. Metamorphosis precedes hatching in Cerathyla, which is distributed from Peru to Panama and has been reported from an altitude of 2,000 m in the Central Andes (Dunn, 1944). Crytobatrachus, with similar breeding habits, is found in the northern Andes. In Colombia it has been taken at an elevation of 2,500 m (Dunn, 1944). Family Leptodactylidae.-This is another family with a strong tendency toward terrestrialism. The eggs of Zachaenus parvulus are deposited in an earth basin away from water. Metamorphosis takes place within the nest (Lutz, 1944). These frogs are found under leaves in mountain forests in southeastern Brazil (Cochran, 1955). All of the very many species of Eleutherodactylus and its allies apparently lay their eggs on land and lack a tadpole stage. These frogs are widespread in tropical America and the West Indies and are especially numerous in the northern Andes of western South America. Dunn (1944) says that the center of geographic dispersion of Eleutherodactylus extends from Central America to the humid woodlands of the Pacific coast of Colombia and Ecuador. The genus has a wide altitudinal distribution, some species being found in lowlands and others high in the mountains. Lynn (1940) lists twelve species of Eleutherodactylus on the island of Jamaica. Of these, two recently introduced forms, ricordi and martinicensis are found in the lowlands. Two, luteolus (= gossei) and cundalli have wide altitudinal distributions, being found from the lowlands to 1,500 m. The other eight are all montane species. Among the Australian leptodactylids, the two species of Kyarranus lay their eggs in nests in sphagnum or moist earth. Since these nests are located far from water, Moore (1961) believes that there is no aquatic stage. Kyarranus inhabits the mountains of New South Wales and Queensland, from 800 to 1,500 m (Moore, 1961). Darwin's frog, Rhinoderma darwini, is found in the Andes of Chile. The eggs are laid on land, and when nearly ready to hatch they are picked up by the male and passed down into his vocal pouch. Here they hatch and the young metamorphose (Noble, 1931). Sminthillus limbatus of Cuba lays a single large egg on land and the young hatches as a fully formed frog (Dunn, 1926b). Sminthillus sometimes occurs in lowland areas, but seems to be most generally distributed on the mountain ranges of eastern Oriente. Dunn found it near Soledad in thickets and hillsides especially in regions of limestone outcrops. INTERNAL FERTILIZATION Internal fertilization is another prerequisite for the development of the cleidoic egg. It is obviously advantageous in terrestrial reproduction, since the sperm are thereby insured a fluid medium in which to travel. It would also appear to be of immediate advantage to animals breeding in swift water, since it eliminates the danger of the sperm being swept away before the eggs are fertilized. Whether it appeared in the ancestors of the reptiles before or after they began laying eggs on land cannot be determined. It seems to have evolved independently several times among the amphibians. So far as known, all caecilians practice internal fertilization with the protrusible cloaca of the male serving as a copulatory organ. As was pointed out above, there is evidence that the caecilians are primarily montane in distribution. Two genera of frogs also have internal fertilization. Ascaphus uses an extension of the cloaca as an intromittent organ (Noble and Putnam, 1931). N ectophrynoides lacks

6 AMPHIBIAN EGGS 369 such an organ; and the sperm are transmitted to the reproductive tract of the female by cloacal apposiiton (Angel and LaMotte, 1944). Both of these frogs are montane forms. The unique method of internal fertilization by transfer of spermatophores practiced by the more advanced salamanders cannot be correlated with reproduction in swift water, since many of these salamanders breed in quiet ponds. However, at least in the eastern United States, salamanders are much more prone to breed in fast-flowing streams than are frogs, and there is nothing inherently improbable in the mechanism having evolved first as an adaptation to reproduction in a montane environment. INCREASE IN YOLK Increase in yolk is important in the development of the amniote egg. In part this is because the increase in nutriment allows development to proceed within the egg until the embryo has reached the adult body form. Perhaps more important is the mechanical effect on the early stages of development of a large supply of yolk in the ovum. Noble (1931) has pointed out the difference in gastrulation between the small-yolked egg of a frog or salamander and the large-yolked egg of a caecilian. In the former, the whole of the yolk hemisphere is covered over as the blastopore closes, while in the latter the blastopore becomes circular while the yolk hemisphere is still largely uncovered. As a result, the blastopore becomes surrounded by blastodisc while the latter still remains on the upper surface of the egg. Noble considers this an important step in the development of the amniote egg. He points out that if the developing embryo should then sink into the blastodisc until the surrounding tissue folded over it, an amnion would, in effect, be formed. Increase in size of yolk seems to be associated with reproduction in cold waters. This has been indicated for Rana (Moore, 1949), and for arctic, antarctic, and deep sea fishes (Marshall, 1953). Noble (1931) considers large egg size in salamanders an adaptation for reproduction in the cold waters of mountain brooks. Large egg size, with a concomitant reduction in the number of eggs per clutch, is also characteristic of amphibians with terrestrial breeding habits. Indeed, some authorities (e.g., Brown, 1952; Inger, 1954; Parker, 1934) feel that, in the absence of direct information on breeding habits, the presence of a small number of large, unpigmented, ovarian eggs in a female frog is suggestive of terrestrial reproduction. DISCUSSION There seems to be a marked correlation in the modern amphibians between the development of terrestrial breeding habits and a humid, montane environment. Some frogs in seasonally arid regions do lay their eggs away from water, but in places subject to flooding. Hatching is delayed until the arrival of the rainy season restores the eggs to an aquatic environment. We know of no frog in such an environment that lacks an aquatic larval stage. Speaking of Australian frogs, Main, Littlejohn, and Lee (1959) state, "No species of frog with a terrestrial larval development occurs in the desert." On the other hand, on the humid and mountainous island of Jamaica, the only frog that breeds in open water is the introduced Bufo marinus. All of the native frogs have modified breeding habits. The eggs and tadpoles of the four species of H yla develop in the small amount of water trapped in the bases of the leaves of bromeliads, and all of the Eleutherodactylus reproduce on land (Lynn, 1940). It may be, then, that the ancestors of the reptiles lived in a region in which there was orogenic activity during the Carboniferous. As the land rose, and the nature of the aquatic habitats shifted from warm, quiet pools to cool, swift-flowing streams, these animals adapted to the changed conditions by developing large-yolked eggs, and perhaps also internal fertilization. As uplift continued, and the streams became mountain torrents, the animals began laying their eggs away from the water, per-

7 370 OLIVE B. GOIN AND COLEMAN J. GOIN haps in crevices in the bank or under damp moss on rocks along the stream. Humidities are frequently very high in mountainous regions and eggs laid in such situations would be in little danger of desiccation. The larvae may at first still have had to return to the water to complete development, but once the egg had become terrestrial, selective pressure toward the elimination of the vulnerable larva would have been strong. This would have favored a further increase in yolk content, leading eventually to the formation of the amnion. Tihen points out that the earliest known reptile remains come from deposits that must have formed in continuously humid, tropical or subtropical swamps. This need not militate against a montane origin for the group. The absence of fossil remains of animals evolving in a mountainous region is to be expected. Probably once they achieved the amniote egg, the reptiles soon reinvaded the lowlands and began their spread over the land. SUMMARY A survey of modern amphibians having terrestrial breeding habits suggests that there is a correlation between the evolution of this reproductive pattern and a humid, montane environment in which the most usual aquatic habitat is a cool, rapidly flowing stream. Two other factors important for the development of the amniote egg, internal fertilization and increase in yolk content, would apparently also be selected for in such an environment. It is suggested that the reptiles arose in a region undergoing orogenic activity in the Carboniferous, and that the primary selective agent was the change in potential breeding sites from quiet, open bodies of warm water to cool, swift-flowing streams. LITERATURE CITED ANGEL, F Vie et moeurs des amphibiens. Payot, Paris. -- AND M. LAMOTTE Nouvelles observations sur N eciophryoides occidentalis Angel. Remarques sur Ie genre Nectophrynoides. Ann. Sci. Nat. Zool., (11) 10: BOULENBER, G. A A vertebrate fauna of the Malay Peninsula. Reptilia and Batrachia. Taylor and Francis, London. BROWN, W. C Amphibians of the Solomon Islands. Bull. Mus. Compo Zool., 107: COCHRAN, D. M Frogs of southeastern Brazil. U. S. Nat. Mus. Bull., 206: xi 423 p. DUNN, E. R. 1926a. Salamanders of the family Plethodontidae. Smith College, Northampton, Mass b. The habits 'of Sminthillus limbatus. Copeia, 155: The American caecilians. Bull. Mus. Compo Zool., 91: Los generos de anfibios y reptiles de Colombia. Primera parte: Anfibios. Caldasia, 2(10): GUIBE, J Recherches sur les batraciens de Madagascar, III. Rectification au sujet de M antipus angeli Guibe. Mem. Inst. Sci. Madagascar, ser. A, 7, fasc. I: INGER, R. F Systematics and zoogeography of Philippine Amphibia. Fieldiana: Zoology, 33: Comments on the definition of genera. EVOLUTION, 12: KmTISINGHE, P The Amphibia of Ceylon. Privately published, Ceylon. LOVERIDGE, A New Guinean reptiles and amphibians in the Museum of Comparative Zoology and United States National Museum. Bull. Mus. Compo Zool., 101: Check list of the reptiles and amphibians of East Africa (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Compo Zool., 117: LUTZ, B Biologia e taxonomia de Zachaenus parvulus. Bol. Mus. Nac., Rio de Janeiro, n.s., zool., 17: Ontogenetic evolution in frogs. EVOLUTION, 2: LYNN, W. G The herpetology of Jamaica. 1. Amphibians. Bull. Institute of Jamaica, Sci. Ser., I: MAIN, A. R., M. J. LITTLEJOHN, AND A. K. LEE Ecology of Australian frogs. Biogeography and ecology in Australia. Monogr. Biol., 8: MARSHALL, N. B Egg size in Arctic, Antarctic and deep-sea fishes. EVOLUTION, 7: MOORE, J. A Patterns of evolution in the genus Rana; in Genetics, Paleontology and Evolution. Edited by Jepsen, G. L., E. Mayr, and G. G. Simpson, Princeton. pp Frogs of eastern New South Wales. Bull. Amer. Mus. Nat. Hist., 121:

8 AMPHIBIAN EGGS 371 NOBLE, G. K Biology of the Amphibia. McGraw-Hill, New York. NOBLE, G. K., AND P. G. PUTNAM Observations on the life history of Ascaphus truei Stejneger. Copeia, 1931: PARKER, H. W A monograph of the frogs of the family Microhylidae. British Museum, London. ROMER, A. S Origin of the amniote egg. Sci. Mont., 85: Ross, W Reptiles and amphibians of southern Africa. Maskum Miller, Capetown. STEPHENSON, N. G On the development of the frog, Leiopelma hochstetteri Fitzinger. Proc. Zoo!' Soc. London, 124: TIHEN, J. A Comments on the origin of the amniote egg. EVOLUTION, 14: