Amniotes include most of theland-dwelling vertebrates alive today, namely, mammals, turtles, Sphenodon,lizards, crocodylians and birds. It is a diverse clade with over 20000 living species. Amniotes include nearly all ofthe large plant- and flesh-eating vertebrates on land today, and they liveall over the planet in virtually every habitat. They also sport disparateshapes - chameleons, bats, walruses, Homo sapiens, soft-shelled turtles,ostriches and snakes are but a few examples - and they include some of thesmallest (sphaerodactyline geckoes) and largest (mysticete whales)vertebrates (Figs. 1 and 2). Although fundamentally land dwellers, several cladessuch as ichthyosaurs, plesiosaurs, pinnipeds and cetaceans have returned tothe sea. A few forms are gliders - the Flying Dragon lizards, HoneyCreepers, and Flying Squirrels - and powered aerial flight has originatedthree separate times, first in pterosaurs, then in birds, and later stillin bats.
Figure 1. Two extant amniotes (diapsids). The rattlesnake can detect its prey atnight using an infrared-sensitive organ that allows it to detect the warmbody of small mammals. It then kills its prey with it poisonous fangs.Parrots eat nuts and fruits. Pictures copyright © 1996 Michel Laurin.
An extensive fossil record documents the origin and early evolution ofAmniota, and that record has played a key role in understandingphylogenetic relationships among the living amniotes (Gauthier et al.,1988b). The oldest amniotes currently known date from the MiddlePennsylvanian locality known as Joggins, in Nova Scotia (Carroll, 1964).The relationships of these fossils indicate that amniotes first divergedinto two lines, one line (Synapsida) that culminated in living mammals, andanother line (Sauropsida) that embraces all the living reptiles (includingbirds). One Joggins fossil, the "protorothyridid" Hylonomus, appears to bea very early member of the line leading to Sauria (Crown-clade diapsids),the clade encompassing all living diapsids. This suggests that the more inclusive clade of which turtles (Testudines) are part (Anapsida) in most morphological phylogenies had diverged as well, even though its current record extends back only to the Lower Permian (Laurin & Reisz, 1995).
An older amniote (from the Lower Carboniferous) was reported (Smithson, 1989). However, more recent studies suggested that it was only a close relative of amniotes (Smithson et al., 1994), and the latest study even suggested that it was more likely to be a stem-tetrapod or an early amphibian than a relative of amniotes (Laurin & Reisz, 1999).
Figure 2. More extant amniotes. The insectivorous elephant shrew (a synapsid) resembles vaguely the earliest placental mammals. The omnivorous terrestrial box turtle Terrapine (an extant anapsid) eats mushrooms, fruits, insects, and worms. Like all turtles, it is not affected by senescence but eventually succumbs to a disease, an accident, or a predator. Pictures copyright © 1996 John Merck.
Many amniote synapomorphies arewidely interpreted as adaptations to the rigors of life on land. Indeed,Amniota owes its name to what may be its most distinctive attribute, alarge "amniotic" egg. While most of us are most familiar with the hard-shelled eggs found in birds, Stewart (1997) showed that the first amniotic eggs probably had a flexible outer membrane, and that a mineralized (but still flexible) outer membrane is a synapomorphy of reptiles. The heavily mineralized, hard shell is a synapomorphy of archosaurs (crocodiles and birds), and it also appeared at least three times in turtles, and a few times in squamates. This probably explains why the oldest known amniotic egg (Coyne, 1999) only dates from the Lower Triassic (220 My), whereas the oldest amniote dates from the Upper Carboniferous (310 My); the eggs of most (if not all) Paleozoic amniotes must have had a flexible, poorly mineralized or unmineralized outer membrane, and thus had a low fossilization potential (Laurin, Reisz & Girondot, 2000).
The amniotic egg possesses a unique set ofmembranes: amnion, chorion, and allantois (Fig. 3). The amnion surroundsthe embryo and creates a fluid-filled cavity in which the embryo develops.The chorion forms a protective membrane around the egg. The allantois isclosely applied against the chorion, where it performs gas exchange andstores metabolic wastes (and becomes the urinary bladder in the adult). Asin other vertebrates, nutrients for the developing embryo are stored in theyolk sac, which is much larger in amniotes than in vertebrates generally.Hatchling amniotes also possess an egg-tooth and horny caruncle on thesnout tip to facilitate exit from their hard-shelled eggs. The amnioticegg, together with a penis for internal fertilization, loss of afree-living larval stage in the life cycle, and the ability to bury theireggs, enabled amniotes to escape the bonds that confined their ancestors'reproductive activities to aquatic environments. It has been suggested that theoriginal function of the extra-embryonic membranes of the amniotic egg was to facilitate interactions between the mother and the embryo (Lombardi, 1994), but this hypothesis is not supported by the distribution of extended embryoretention in vertebrates, according to most proposed phylogenies (Laurin &Girondot, 1999). Some components of theamniotic egg have been variously modified within Amniota. Placentalmammals, for example, have suppressed the egg shell and yolk sac, andelaborated the amniotic membranes to enable nutrients and wastes to passdirectly between mother and embryo.
Figure 3. Development of extraembryonic membranes in an amniote egg(chick). In this early developmental stage, the yolk sac is expanding overthe yolk. The amnion and chorion are expanding over the embryo and willeventually form the amniotic chamber. The allantois is expanding towardthe chorion, with which it will form a respiratory membrane, in addition tostoring metabolic wastes of the embryo. Redrawn from Campbell (1993). Copyright © 1996 Michel Laurin.
The comparative aridity of the terrestrial environment affects all aspectsof amniote biology, and not just their reproductive systems. Thus,amniotes have a relatively impervious skin thatreduces water loss. They also possess horny nails that, among other things,enable them to use their forelimbs to dig burrows into which they canretreat during the heat of the day. The imperative to reduce water loss isequally evident in the density of renal tubules in the metanephric kidneyof amniotes, in the larger size of their water-resorbing large intestines,and in the full differentiation of the Harderian and lacrimal glands in theeye socket whose antibacterial secretions help to moisten and, along with athird eyelid (the nictitans), to further protect the eye from desiccation.The commitment of amniotes to a life on land is also revealed by anextensive system of muscle stretch receptors that enables finercoordination and greater agility during locomotion, their enlarged lungs(which are the only remaining organs of gas exchange owing to the loss ofgills), and the complete loss of the lateral line system other vertebratesuse to detect motion in water.
Many of these features are rarely preserved in fossils, but there are somenovelties in the skeleton that are no less diagnostic of amniotes. Forexample, amniotes have at least two pairs of sacral ribs, instead of justone pair. They also have an astragalus bone in the ankle, instead ofseparate tibiale, intermedium, and proximal centrale bones. Finally, theyhave paired spinal accessory (11th) and hypoglossal (12th) cranial nervesincorporated into the skull, in addition to the ten pairs of cranial nervespresent in amphibians.
Evolution and Systematics
Discussion of Phylogenetic Relationships
Phylogeny modified from Laurin & Reisz (1995) and Lee(1995); the position of Mesosauridae follows Modesto (1999). Node names follow Gauthier et al. (1988b) and Gauthier (1994). The position of turtles (Testudines) is uncertain; some authors place them approximately in the position shown above (Laurin & Reisz, 1995; Lee, 1993, 1995), while others place them among Diapsida (deBraga & Rieppel, 1996, 1997; Rieppel & Reisz, 1999; Hedges & Poling, 1999; Mannen & Li, 1999).
Generations of systematists have studied amniote phylogeny at diverse genealogical levels, and until a few years ago, its broad outlines were thought to be reasonably well understood. Indeed, recognition of the major living clades, such as mammals, turtles and birds, antedates the Theory of Descent. Relations among these taxa, and especially the connections of various fossils to them, have been contentious in post-Darwinian times. Much of that controversy can, however, be attributed to the fact that during the first two-thirds of this century, there was little thought given to what constituted evidence for phylogenetic relationships. The origins of the major extant lines of Amniota have become clearer in the post-Hennigian era. Nevertheless, the precise relations of a number of clades, most notably the turtles among extant forms and the aquatic and highly divergent ichthyosaurs and sauropterygians among extinct forms, remain contentious.
Early phylogenetic analyses placed turtles outside of the remaining amniotes (only crown-clade names are listed to simplify the trees):
* Major living amniote clades after Gaffney (1980).
Gauthier et al. (1988a, b, and c) later placed turtles as the sister clade to Sauria (crown-diapsids), and this topology has now gained wide acceptance, at least among morphologists and paleontologists:
However, Rieppel (1994, 1995), Rieppel & deBraga (1996) and deBraga & Rieppel (1997) have suggested that turtles may be the sister clade to lepidosaurs. This requires that turtles are saurians who have lost both the upper and lower temporal fenestrae (holes in the skull associated with jaw muscles) so diagnostic of diapsid reptiles:
The three trees presented above include only extant taxa, and many phylogenetic analyses of amniotes have ignored extinct taxa. However, it is important to bear in mind that discovering the globally most parsimonious tree requires the inclusion of extinct taxa in a phylogenetic analysis (Gauthier et al., 1988b). Without fossils, the best-supported tree for amniotes inferred from morphological data is the following (although only one more step is required to switch the positions of lepidosaurs and turtles):
* Tree based on living amniotes only (after Gauthier et al., 1988b).
Recent molecular evidence for amniote relationships conflicts with paleontological and morphological evidence. Initially, some gene sequences suggested a close relationship between birds and mammals, although never with strong statistical support (e.g., Bishop & Friday, 1987; Goodman et al., 1987; Hedges et al., 1990). More recently, a study of the molecular evidence for the origin of birds (15 genes; 5280 nucleotides, 1461 amino acids) discovered strong support (100% bootstrap P value, BP) for a close relationship between birds and crocodilians (Hedges, 1994). A smaller data set of 11 transfer RNA genes (686 sites) also resulted in a bird-crocodilian grouping (Kumazawa & Nishida, 1995). A basal position for mammals was supported (99% BP) by analysis of a 3 kilobase portion of the mitochondrial genome containing the two ribosomal RNA genes (Hedges, 1994). In the same study, a Sphenodon-squamate relationship also was found, but support for that grouping and for the position of turtles was not very strong.
The most recent molecular phylogenies have generally placed turtles among archosauromorphs, and often within archosaurs (Mannen et al., 1997; Mannen & Li, 1999; Hedges & Poling, 1999). The latter placement is the least compatible with the morphological evidence, and no convincing explanation has been found so far to explain this discrepancy.
Many gene sequences of birds and mammals exist, but the relatively small number of sequences from representatives of other amniote lineages, especially tuataras (Sphenodon) and turtles, has hindered the estimation of a robust molecular phylogeny for all major groups of living amniotes. This is reflected by the low resolution of the molecular phylogeny obtained by Hedges & Poling (1999) when Sphenodon (using only sequences of genes available in sphenodon) was included:
Without Sphenodon and using the greater number of sequences available for other taxa, Hedges & Poling obtained the following fully resolved phylogeny, in which turtles are the sister-group of crocodilans:
If extinct amniotes are considered, the phylogeny is much more complex and controversial. Formerly, captorhinids were believed to be closely related to turtles (Gauthier et al., 1988b, c), but more recently, procolophonids (Reisz & Laurin, 1991; Laurin & Reisz, 1995), pareiasaurs (Lee, 1993, 1994, 1995, and 1996), and even sauropterygians (a group of Mesozoic diapsids) have been suggested to represent early relatives of turtles (Rieppel, 1994, 1995; Rieppel & deBraga, 1996; deBraga & Rieppel, 1997). The linked page Phylogeny and Classification of Amniotes provides information about the phylogenies incorporating extinct amniote taxa, and provides a detailed classification of the relevant groups. The linked page Temporal Fenestration and the Classification of Amniotes discusses how temporal fenestration has been used to classify amniotes, and how tempororal fenestration evolved.