26 THE ROLE OF WATER IN THE EVOLUTION OF THE TERRESTRIAL VERTEBRATES BY J. GRAY, M.A., King's College, Cambridge. (From the Zoological Laboratory, Cambridge.) (Received igth January 1928.) (With Three Text-figures.) As pointed out by Watson (1925) the main problems of evolution centre round the possibility of deriving one morphological type from another without postulating or necessitating any functional discontinuity of the organs involved. We have, therefore, to assume that the changes in bodily structure attending evolution of the earliest terrestrial vertebrates (from purely aquatic organisms) occurred in such a way as to maintain the organism, at all times, in a state of physiological efficiency. This particular phase of evolution must have been influenced if not largely controlled by the fact that, whereas water is present in abundance in the environment of a fish, it is not so present in the case of an animal living in air. Since the tissues of all living vertebrates, irrespective of environment, contain nearly 8. per cent, of their weight of water (Table I) it seems almost certain that the terrestrial forms were evolved without seriously reducing the percentage of water in their tissues. Table I. Vertebrate type Dogfish Trout Goldfish Frog Newt Salamander Lizard Sparrow Fowl Approx. % water in embryo, or in non-bony adult tissues 81 84 75 8 79 78 72 78 83 If a fish, frog, newt, or salamander is removed from water and exposed to air, the loss of water which occurs is both rapid and extensive; this suggests that the acquisition of a skin which is relatively impermeable to water occurred after and not before the origin of the earliest terrestrial types. Fig. 1 illustrates these facts in respect to the newt, Triton cristatns. It is seen that a live newt loses water by
The Evolution of Terrestrial Vertebrates 27 evaporation from the skin as rapidly as one that is dead, and on being replaced in an aquatic medium water is very rapidly reabsorbed by the same route. An animal of this type may be said to be in dynamic equilibrium with the water in the surrounding environment. A typical reptile {Lacerta viridis), on the other hand, only very slowly loses water when exposed to air (Fig. 2) and this character persists long after the animal is killed; the organism is almost, in fact, in static equilibrium with the water of its environment. This adaptation to conservation of water is undoubtedly located in the dermis, since if the skin'of the lizard be removed the tissues lose water as rapidly as do those of a newt or fish (Fig. 3). ALIVE IN WATER 5 TRITON CRISTATUS 5 6 7 8 9 O 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 Time in hours Figs, i and 3. Illustrating that when a newt is exposed to air it very rapidly loses water through its skin, whereas the loss from a lizard is very much less. Just as a lizard does not lose water through its skin, so it is unable to absorb water rapidly by this organ. We find, in fact, that reptiles are the earliest type of vertebrate which drinks water through the mouth and absorbs it by the alimentary canal. Typical amphibia on the other hand do not drink; they imbibe water over the whole surface of their bodies. We may, therefore, conclude that vertebrate terrestrial life remote from water became possible as soon as the adult organism acquired a skin which was relatively impermeable to water. Complete adaptation to terrestrial life includes the faculty of laying eggs on land and of supplying the developing embryo with an adequate supply of water. In a previous paper (Gray, 1926) it has been shown that the eggs of a fish {Salmo fario) absorb a significant amount of water from their environment, and there are good reasons for supposing that a similar phenomenon occurs in nearly every other aquatic type of egg. Table II shows that in a variety of cases 5 percent, of the weight
J. GRAY of vertebrate eggs is composed of solid materials; two-thirds of this material is converted into embryonic tissues whose water content is roughly 8 per cent. It is clear that a significant amount of water in the full time embryo must be derived from the external medium. Table II. Type Scyllium canicula Salmofario Ranafusca R. temporaria (Maja squinado) {Sepia officinalis) Tropidonotus Gallus gallus Moorhen % solid material in the yolky phase of the egg Si'4 4i 41 A 42-6 43-6 47-3 567 49 ; 5i 6 S 5 T3 8 4 o 3 M CO O s 2 o 1 \ I7ARQ with sk'm 1 2 3 4 5 Time in Hours Fig. 3. Illustrating that the power, possessed by a lizard, of preventing loss of water from its tissues is due to the properties of its skin. It is necessary, therefore, to determine the means whereby the embryo of a reptile or bird is supplied with an adequate supply of water, and secondly to consider how such a type of egg could have been evolved from purely aquatic types without any discontinuous change in function on the part of any essential organ. There can be no doubt that the embryos of oviparous amniotes derive their water supply from the so-called albumen deposited round the egg by the walls of
The Evolution of Terrestrial Vertebrates 29 the oviduct. This fact is illustrated by the figures in Table III which are largely self-explanatory. In the sample of eggs used, the average egg weighed 58 gm., and contained 38-7 gm. of water when newly laid. Of this water, 8*8 gm. were in the volk and 29*9 gm. in the albumen. At the end of incubation the embryo contains 27'4gm. of water according to Murray's (1925) figures, so that at least two-thirds of this is derived from the albumen. Table III. Day of incubation Embryo (Murray) Amount of water (in gm.) in Yolk Albumen and amniotic space Water lost in evaporation at 23 % humidity (Calc. from Murray) Water gained by combustion of yolk 6 8 1 12 14 16 18 2 o-4 I-I 2-5 4-6 79 I2'O 18-1 27-4 8-5 8-45 8-4 8-2 7-8 6-9 5'3 2-3 i-o 29-9 27-2 25-4 23-2-4 16-9 12-6 9-2 2-2 2*4 3 'i 5-6 6-7 7-8 8-8 9-8 O'OI -5 O-I2-27 -5 o-8o 1*2 2-O That two-thirds of the water in the embryo is provided by the albumen is shown by the fact that as incubation proceeds the amount of water in the albumen falls steadily whilst the percentage of water in the yolk does not materially decrease after the 1th day. It will be noticed that 7 per cent, of the water in the albumen is required for the formation of the embryo leaving 3 per cent, available for evaporation. I have calculated the loss due to this cause at 23 per cent, humidity, since this appears to be a reasonable figure for the conditions of incubation of Gallus gallus in its natural habitat. As far as is known no reptilian or avian egg exists without an albuminous phase, and we may assume that its function is not radically different from that performed in the case of the chick. It is, therefore, fairly clear that the amniota have solved the problem of providing their embryos with an adequate supply of water by enveloping the egg with an aqueous mass of protein secreted by the walls of the oviduct. Now it is hardly conceivable that the albuminous layer of the amniote egg arose de novo as an adaptation to terrestrial life, for this would involve a sudden change in the structure of the oviduct. By the principle of physiological continuity it is much more reasonable to regard this essential layer as the equivalent of such homologous structures as are found in the anamniota. Among fishes, tertiary egg membranes rich in water are found in the Dipnoi, and they appear to be present in all amphibia. These membranes apparently protect the egg against destruction by predatory animals although there may be subsidiary functions associated with the incubation of the embryo. In most amphibia, where the tertiary envelope is of a mucoid nature, the full water content of the envelope is not attained until after the egg has been deposited in water, but interesting and suggestive modifications are found in the eggs of those
3 J. GRAY amphibia which deposit their eggs on land. In Phyllomedusa and in Rhacophorus the protective function of the mucoid envelope is to a large extent replaced by other devices, and it is difficult to resist the conclusion that the envelopes are themselves largely devoted to the provision of water to the embryo. In Phyllomedusa hypochondrialis the eggs are deposited in the folds of a leaf. The mucilaginous egg capsules rapidly liquefy after oviposition and provide a fluid medium in which the eggs develop. Agar (199) observed that a certain percentage of capsules contain no eggs, and this suggests that the function of these membranes, if any, is to augment the amount of water available for the larvae. The essential point is that the whole of the water necessary for development is provided by the oviductal walls of the mother. Similarly, the eggs of Rhacophorus schlegelii are laid in a subterranean burrow. Having formed the burrow, the female secretes into it mucilage which, by means of her feet, is rapidly worked into a froth. Into this froth the eggs are laid, and as development proceeds the froth is gradually liquefied. Here again the water for development is all derived from the female organs. From these types it is not difficult to derive either the egg of a reptile with its solid albumen phase or the egg of a typical bird where the fluid albumen has entirely lost its power of protecting the embryo against predatory foes. It is interesting to note that, far from requiring a supply of water from external sources, the eggs of birds fail to develop unless a certain amount of water is actually lost by evaporation during incubation (Chattock, 1925). A suggestive experiment by Weldon (192) indicates that the formation of an amnion is dependent on a loss of water by evaporation 1. Since the mammals are derived from oviparous reptiles it is of interest to consider how the small eutherian egg can be derived from that of the latter group without any break in the physiological functions of the organs concerned. A conceivable fine of origin is suggested by the eggs of Monotremes. The egg of these forms has no true albumen layer and the yolky ovum lies close under the shell. As it leaves the ovary, the egg is about 2 mm. in diameter, but during its passage down the oviduct its bulk is enormously increased so that the yolky phase is about 14 mm. in diameter before the shell is deposited (Caldwell, 1887). This 3-fold increase in volume must largely be due to an absorption of water, although a certain increase in dry weight may well occur. The only significant difference between the egg of amonotreme and that of a reptile is that, in the former, the aqueous secretions of the oviductal walls are passed straight into the yolky ovum itself instead of being deposited on its surface as a separate phase. In eutherian mammals this process has gone one step further since the water contained in the mother's blood is passed, not into the ovum, but direct into the embryo. If the arguments here presented are sound, there seems good evidence to show that terrestrial vertebrates have descended from afish-likeancestor which possessed a glandular oviduct. The secretions of these oviducts were at first utilised as a 1 Weldon incubated eggs in such a way as to replace the water normally lost by evaporation without interfering with the processes of evaporation or respiration. In such eggs the amnion failed to develop normally.
The Evolution of Terrestrial Vertebrates 31 protective covering to the eggs, but eventually they made it possible for the eggs to develop OR land by providing an adequate supply of water to the embryo. It is, perhaps, permissible to note that the evidence available is entirely experimental and it would seem reasonable to hope that the application of similar methods might, if applied to the various organs of the body, provide useful evidence of evolutionary change. SUMMARY. 1. The origin of truly terrestrial vertebrates was associated with the development of a skin which is relatively impermeable to water. The skins of fishes and of most amphibia are freely permeable to water, and water is readily lost when the animals are exposed to ordinary atmospheric conditions. The skin of a lizard on the other hand is relatively impermeable to water and the animal is therefore able to withstand exposure to air without serious inconvenience. 2. Complete adaptation to terrestrial conditions involved the capability of laying and incubating eggs on land. The eggs of fish and of typical amphibia absorb, during development, considerable quantities of water from the external environment. The embryos of reptiles and of birds derive the equivalent of this water from the albumen layer secreted round the ova by the walls of the oviduct. 3. Evidence is presented to show that the reptilian type of egg is to be derived from that of a dipnoan or amphibian, where the egg is surrounded by a mucilaginous or albuminous secretion of the walls of the oviduct. 4. The mammalian egg can be derived from that of a reptile by supposing that the aqueous secretions of the oviduct are passed direct into the embryo instead of forming a separate phase round the ovum. An intermediate type is found in Monotremes where the secretions are passed into the ovum itself before the egg is laid. REFERENCES. AGAR, W. E. (199). Proc. Zool. Soc, 893. CALDWEIX, W. H. (1887). Phil. Trans. Roy. Soc. 178, B, 463. CHATTOCK, A. P. (1935). Phil. Trans. Roy. Soc. 213, B, 397. GRAY, J. (1926). Brit.Journ. Exp. Biol. 4, 215. MURRAY, H. A. (1925). Jowrn. Gen. Physiol. 9, 1. WATSON, D. M. S. (1925). Phil. Trans. Roy. Soc. 214, B, 189. WELDON, W. F. R. (192). Biometrika, 1, 365.