Feder (1992) suggests that in amphibians, "fluxes seemingly permit only limited independence from the immediate environment. Accordingly, the internal milieu of amphibians may be far less fixed than that of many other vertebrates." The first amniote, we suggest, accomplished the transition from amphibian anamniote to reptilitan amniote mainly by altering its style of water balance. Restriction of skin permeability permitted greater control over the state of body hydration. Early amphibians remained tied to water for feeding, reproduction, and most of their locomotion, using land probably primarily as a refuge from predators or as an escape route from one drying or degraded aquatic habitat to another, still essentially aquatic, home. Reptiles completed the transition to land by truly exploiting the land's resources as a permanent habitat.
Minimally, the proto-amniote that came to fully exploit the land must have greatly reduced cutaneous evaporative water loss by decreasing skin permeability. However, reducing permeability of the skin also reduced cutaneous reuptake of water, reduced cutaneous respiration, altered acid-base regulation, greatly reduced the utility of ammonia as a waste product, and reduced the cooling effect of evaporation from the body surface, changing heat budget parameters. The simple reduction of water permeability alters nearly all of this animal's systems. Therefore, amniotes have the potential, and probably the necessity, for more refined regulation of blood plasma volume, osmolarity, pH, and body temperature control systems than anamniote tetrapods; in short, the transition to land provides the potential, and possibly the necessity, for increased metabolic homeostasis. To accomplish this, several or all of the following changes were necessary.
The permeability of amphibian skin to water is under some hormonal control. Arginine vasotocin (AVT) increases permeability for osmotic uptake of water, but does not affect resistance to evaporation (Shoemaker et al., 1992). This is because the main resistance to evaporation is the boundary layer above the skin, not the permeability itself. In addition, AVT acts as an antidiuretic (Uchiyama and Pang, 1985). During the transition to full terrestriality, AVT secretion may have increased to reduce urinary loss of water, and skin permeability must have been de-coupled from this hormone. Scales are not necessary for reduction of evaporative water loss, as a comparison of scaled and scaleless reptiles has shown (Bennett and Licht, 1975). Heavily keratinized epidermal layers are not particularly necessary in the initial stages, as the "waterproof' frog Phyllomedusa has only one layer of stratum corneum, coated in wax and lipids, and yet has greatly reduced permeability to water (Shoemaker et al., 1992). Making the skin permanently and irreversibly impermeable to water solved the problem of water loss but intensified the problem of water gain.
An animal eating succulent plants ingests more dietary water than a carnivore does, although plants may be more difficult to digest. Green plants have approximately 1.5 times as much nitrogen as meat or insects per unit of energy and an order of magnitude more water content (Shoemaker and Nagy, 1977). Plants also have higher osmolytes than animal food (Shoemaker and Nagy, 1977). Fossil evidence indicates that the early amniotes were insectivorous (Carroll, 1988), but that herbivory developed early in the amniote lineage (Hotton et al, this volume). Insects have only slightly more water per kilocalorie than meat but have more than double the nitrogen, much of which would have to be excreted as wastes. Omnivory would have been ideal for the earliest amniotes in order to maximize benefits and minimize problems of water balance. Increased body temperature would have facilitated increased use of plant matter in the diet.
A further means of adding dietary water would be the use of drinking. This behavioral change is another potential adaptation for life on land, and could have been accomplished initially by the use of tongue flicking, with minimal anatomical or physiological adjustments.
Excreted Nitrogenous Wastes without Losing Excessive Water
The cessation of urine excretion while on land is the extreme response by extant amphibians, necessary to conserve water because of their high cutaneous evaporative water loss . The excretory system of an amniote must excrete nitrogenous wastes terrestrially, in a relatively nontoxic form, using relatively little water, and in addition, it is likely that the excretory system would be required to take an increased role in acid-base regulation. If the animal ate a diet of plants, or a diet of insects or meat and drank water, it could probably meet its dietary needs and have sufficient water to excrete nitrogenous wastes as urea, so it may be that the earliest amniotes did not require uric acid for excretion. Sphenodon uses urea, not uric acid (SchmidtNielsen and Schmidt, 1973), indicating that this may have been the ancestral condition. Turtle eggs use urea as a nitrogenous excretory product, as well (Packard et al., 1984). The proto-amniote may have continued to use a urinary bladder for storage, and later, reabsorption of water, as do desert tortoises (Nagy and Medica, 1986).
Reproduction on Land without Returning to Fresh Water
Internal fertilization is necessary for the cleidoic egg because of the necessity of a protective shell, secreted before the egg is laid (Packard and Seymour chapter, this volume). Although internal fertilization is necessary for terrestrial reproduction, an intromittent organ is not, as in terrestrial salamanders, Sphenodon, and birds. We believe that the presence and structure of extraembryonic membranes are so similar for all amniotes that the amniotic egg probably evolved only once, early on. In its earliest incarnation, it probably had a membranous shell that allowed entry of water and oxygen. A large yolk provided energy. It would probably have been laid in relatively small clutches in a moist environment, perhaps buried in sand or leaf litter. The waste products could have been uric acid, urea, and a little ammonia, although uric acid my not have been necessary in early amniotes. Uricotelism in adults may have preceded evolution of the cleidoic egg (Needham, 1931), because increasing terrestriality of the adult rather than the egg would probably confer an advantage earlier in the evolution of terrestriality. The evolution of the extra-embryonic membranes, in particular the allantois, a ventral out-pocketing of the gut, may have been more an adaptation for sequestering the nitrogenous waste products than a water-conserving measure. On the other hand, the embryo may have tolerated relatively high levels of urea, and the increased osmolarity would have decreased the water potential of the egg and promoted water transport from the soil, as in some turtle eggs (Miller and Packard, 1992).
If an ectothermic animal is dry as a result of decreased skin permeability to water, then it must contend with a much more labile body temperature and therefore it is forced either to accept high variability in body temperature, or to control the internal temperature by increased behavioral regulation. The simple solution of permitting body temperature to vary with environment, which worked well in amphibians because the variation was tempered by evaporative cooling and behavioral microhabitat choices, is not as simple for the dry-skinned animal. Because of the greater temperature fluctuations, the enzymes and metabolic functioning of the animal would have to change continuously with temperature over the course of the day. The optimal solution may have been to come upon a specific, rather narrow preferred body temperature range-one that could readily be attained with behavioral means and maintained over most of the period of activity. The reptile Sphenodon shows more ancestral traits than most extant reptiles, and has a body temperature during activity that is near the minimum temperature tolerated by squamates voluntarily (Bartholomew, 1982). A higher body temperature would increase metabolism and, in turn, oxygen and food requirements. These increased metabolic needs could potentially lead to increased levels and types of activity, and more complex behavior. Thus, by reducing evaporative water loss, the proto-amniote could concommitantly increase its body temperature and its metabolic rate.
In dry-skinned animals, body size has a greater effect on body temperature than in wet-skinned animals. Even small animals may enjoy a warm body temperature if they have dry skin (Spotila et al., 1992). Early amniotes may have been relatively small (Carroll, 1988); on the other hand, larger animals could maintain a high body temperature longer during the day by inertial homeothermy, although this could increase the time necessary to warm up. Large reptiles in a thermally stable environment may need to spend little effort specifically on thermoregulation (Shine and Madsen, 1996), but they do maintain a relatively constant activity temperature. We suggest that the evolutionary trend during the transition to amniotes is toward a higher and deliberately regulated body temperature, and increased homeostasis, in the presence of increased environmental lability.
With the loss of the skin as an exchange surface, the respiratory system would be forced to increase reliance on the lungs for oxygen consumption and more effective carbon dioxide release. Enclosure of respiratory surfaces in a lung also helps to reduce water loss. The enclosure of the respiratory surfaces and the reliance on a tidal flow of the respiratory medium (Piiper and Scheid, 1975) results in intermittent rather than continuous gas exchange, which alters in particular the release of carbon dioxide. The details of the evolution of lungs have been reviewed extensively (Randall et al., 1981; Little, 1990); suffice it here to emphasize the significance of the enclosure of the respiratory surfaces on other aspects of physiology, including acid-base balance and water balance.
Now, we shall set up some hypotheses about the characteristics of the earliest amniote. It seems likely that this tetrapod had a terrestrial diet, probably herbivorous or insectivorous or omnivorous . Its skin was able to resist desiccation in air, probably by thickening and keratinization, and it may have had dermal armor or scales for protection from predators, abrasion, or UV light. Its cleidoic eggs were laid on land in a moist substrate, and had membranous shells permable to water, and true amniotic membranes. The terrestrial hatchlings were not larvae, but fully terrestrial miniature adults. Some of the earliest amniote fossils are those of the Joggins formation of Nova Scotia, in which the reptiles are found associated with tree stumps (Romer, 1966; Carroll, 1988). We speculate that these animals may have sat on tree stumps in basking behavior, similar to extant lizards. Our lines of speculation do not differentiate between particular body sizes, but larger body size could favor herbivory (Pough, 1973; Hotton et al., this volume) unless the animals could select young, tender leaves (Mautz and Nagy, 1987). Carroll (1970, 1988) favors a small size for the transition animal, while Lombard and Sumida (1992) point to a larger animal as the first amniote. Our hypothesis is that the physiological steps from anamniote to amniote may be more definitive than the morphological steps, at least as revealed by the fossil record.
Now, we will try to flesh out the animal a bit more. This putative earliest amniote may have been similar to an extant reptile. The similarities could probably include dry skin and a urinary bladder able to resorb water. This early amniote laid eggs with membranous shells that absorbed water from the environment. It may have had periods of inactivity during the year, and it may have still spent some time in an aquatic habitat, although the skin was relatively impermeable to water when compared to that of amphibians. It may have eaten plants, or insects, or possibly had an omnivorous diet. It basked and behaviorally thermoregulated in order to reach a body temperature that would enable it to be active. What reptile is the most similar to this proto-amniote picture? We suggest that among extant reptiles, turtles in general most closely fit this physiological profile.
What extant amphibian is physiologically most like the earliest amniote? It has relatively impermeable skin, terrestrial reproduction, and forages on land. It tolerates a high body temperature and excretes some of its nitrogenous wastes as uric acid and urea. It lives emerged from water, but it survives seasonal drought and high body temperatures only by becoming inactive. Arboreal "waterproof' frogs in the genera Phyllomedusa and Chiromantis (Shoemaker et al., 1987) most closely fit this model.
Was this article helpful?