Reproduction in Early Amniotes

See allHide authors and affiliations

Science  17 Aug 2012:
Vol. 337, Issue 6096, pp. 806-808
DOI: 10.1126/science.1224301

The conquest of dry land by vertebrate animals began with the evolution of the first four-legged, amphibious animals ∼360 million years ago (1, 2). Amniotes originated ∼50 million years later (1) and have since become the most diverse clade of land-living vertebrates, including mammals, turtles, lizards, snakes, crocodiles, and birds. Evolutionary changes in reproduction were crucial for the move from the sea via swamps to dry land. However, the reproductive structures and early life stages of amniotes fossilize poorly. Exceptional insights into early amniote reproduction are offered by recent fossil discoveries (36). The fact that these fossils come from ancient seas and lakes and not from dry land helps to explain the paradox that there is an older fossil record for live-bearing amniotes than for egg laying in amniotes.

The key evolutionary innovation that enabled amniotes to colonize habitats away from water was the cleidoic egg. Its complex structure added extraembryonic membranes (the chorion and the amnion) and a shell to the primitive vertebrate egg design with its embryo, yolk, and jelly layers (2). These membranes and eggshell enable egg laying and development on dry land. The shell and egg membranes allow gas exchange to and from the developing embryo, letting oxygen in and carbon dioxide out but retaining water. The shell may be either leathery or calcified. Phylogenetic inference shows that leathery shells evolved first; calcified eggshells evolved independently from leathery eggshells at least four times (see the figure).

The evolution of the cleidoic egg had two main effects. One was internal fertilization. The other was that eggs could no longer be laid in water, where the embryo would suffocate. This had major implications for the secondarily marine amniotes that frequently evolved from terrestrial lineages.

Soon after their origin, Amniota split into two major lineages: the mammal-line amniotes (Synapsida) and the bird-line amniotes (Reptilia). Phylogenetic inference from living animals (see the figure) suggests that amniote egg laying evolved no later than in the last common ancestor of mammals and birds, ∼310 million years ago (1). Both animal groups have a cleidoic egg, although this has been lost in mammals more derived than monotremes. It is highly unlikely that this complex egg structure evolved more than once (2). Paleontologists have inferred egg laying on dry land from skeletal indicators of full terrestriality—such as well-developed limb joints—in tetrapods close to the mammal-bird split (2).

However, fossils of the cleidoic egg from near the mammal-bird split have been hard to come by. The oldest cleidoic egg fossils postdate amniote origins by 90 million years. At sites in Argentina and South Africa, fossilized egg clutches and embryos of prosauropod dinosaurs have been found that are 220 million years and 188 million years old, respectively (7). The shells of these prosauropod eggs are very thin but sufficiently mineralized to fossilize. This suggests that calcified eggshell had to evolve before the cleidoic egg could fossilize, because a leathery eggshell and the soft contents of an egg do not preserve as a fossil.

Amniote reproductive evolution and aquatic adaptation.

This phylogenetic tree shows the evolution of amniote reproductive traits and the early appearance of live bearing in Mesosaurus, as well as the frequent co-occurrence of live bearing and an aquatic lifestyle. Extinct taxa are shown in gray. Extinct terrestrial taxa presumably laid eggs, but fossil data about their reproduction are lacking. Recent evidence of pregnant females and isolated embryos or neonates has been reported for mesosaurs, pachypleurosaurs, plesiosaurs, and choristoderes (red) (3, 6). See supplementary materials for the sources used to compile this phylogeny.

In a recent article, Piñeiro et al. (6) provide what is likely to be the earliest direct evidence to date for reproduction in amniotes. They report putative pregnant females (see fig. S1) and isolated embryos and neonates of the primitive amniote Mesosaurus tenuidens from the Early Permian of South America. Mesosaurs are early reptiles and the first amniotes that invaded the marine habitat by evolving a fully aquatic lifestyle (see the figure) (8). The current fossils (6) come from conservation lagerstättes in Brazil and Uruguay, where finely laminated dark shales and limestone beds were deposited in an epicontinental sea about 280 million years ago (8).

The fossils are adult skeletons with very small and immature skeletons inside the rib cage, as well as coiled-up, small immature skeletons found with adults or in isolation, presumably representing embryos and neonates (6). The fossils are incomplete and partially disarticulated and the sample size is small (6). However, very similar occurrences in geologically much younger amniotes strengthen the case that the Mesosaurus specimens are embryos.

The similar but better-preserved and better-studied associations of pregnant females and isolated embryos or neonates are for three species: the 235-million-year-old pachypleurosaur Neusticosaurus from the European Alps (9); another pachypleurosaur of a similar age, Keichousaurus, from China (3); and the 120-million-year-old choristodere Hyphalosaurus (4), also from China. The pachypleurosaurs are known from similar laminated sediments as Mesosaurus; they resemble it in body plan and were specialized arthropod feeders. The body plan of Hyphalosaurus (4) is strikingly similar to those of mesosaurs and pachypleurosaurs, but it lived much later.

Neither Mesosaurus nor the pachypleurosaurs and Hyphalosaurus could move on land to any extent. All share the same distinctive adaptation to a fully aquatic lifestyle, namely an increase in bone mass that affords neutral buoyancy by balancing out the tetrapod lung, allowing the animal to move effortlessly in the water column (10).

Another adaptation crucial for a fully aquatic lifestyle was live bearing, because egg laying requires some ability to go on land. Mammals had evolved live bearing long before returning to the sea, but the various clades of marine reptiles evolved live bearing independently and at different times (3, 5, 6, 11). (Sea turtles are the notable exception, illustrating the problems of a marine lifestyle combined with egg laying.) As in modern lizards, live bearing in Mesosaurus, pachypleurosaurs, and Hyphalosaurus must have evolved via embryo retention in the reproductive tract of the mother past the hatching stage; the associations of isolated embryos and adults may reflect incomplete live bearing. Fully developed live bearing is, however, evident in ichthyosaurs (5, 11) and has recently been established for plesiosaurs (5). The large size of the plesiosaur fetus suggests a reproductive strategy as in modern whales for these animals (5).

The relatively rich fossil record of pregnant amniotes (37, 1113), starting with the Middle Triassic ichthyosaurs ∼240 million years ago (11), supports the interpretation of the geologically much older Mesosaurus finds (6) as embryos inside of females instead of as last meals of cannibals. Cannibalism is well known in modern reptiles, but this is unlikely for the new Mesosaurus fossils, for reasons established by the study of other marine reptile finds with small skeletons inside large skeletons of the same species that have been interpreted as pregnant females (3, 5, 11, 12, 13).

Given that live bearing is documented so much earlier and more frequently in the fossil record than egg laying, it might be thought that the primitive mode of amniote reproduction is live bearing and that the amniote egg with its eggshell and extraembryonic membranes evolved by embryo retention (14) and not egg laying. However, several arguments can be made to counter this assumption. First, egg laying is the primitive state for both mammals and reptiles, but primitive eggs with their leathery shells are unlikely to be preserved. Second, observations on lizards show that live bearing evolves readily from egg laying (by embryo retention) but not the other way around. Finally, a key reason for the rich fossil record of live bearing is the exceptional preservation in conservation lagerstättes, which preferentially sample marine environments and secondarily aquatic amniotes, which were often live-bearing (36, 9, 11, 13) (see the figure). Even terrestrial live-bearing mammals are best known from such lagerstättes (12). Further fossils of early amniote embryos and eggs will be invaluable for further elucidating the evolution of this important clade.

Supplementary Materials

Fig. S1


References and Notes

View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article