Synapsids include mammals and all extinct amniotes more closely related to mammals than to reptiles. Synapsids are the dominant large terrestrial animals worldwide, and they have also invaded the oceans (whales, pinnipeds) and the air (bats). The oldest known synapsid is an ophiacodontid from the Middle Pennsylvanian (320 million years ago) of Joggins, Nova Scotia. By the Lower Permian, therapsids (the group that includes mammals and most of their Upper Permian and more recent relatives) had appeared (Laurin and Reisz, 1990, 1996). However, the oldest known mammal only dates back to the Jurassic (Rowe, 1988). The fossil record of synapsids is one of the most extensive of any groups of vertebrates. This fossil record has been used to illustrate the concept of evolution (Hopson, 1987) and to test macroevolutionary patterns (Kemp, 1985). The largest gap in this fossil record is between the Permo-Carboniferous synapsids and therapsids.
The fossil record provides conclusive evidence that synapsids are the first amniotes to diversify. Synapsids quickly became the most diverse, widespread and most common amniotes in the Late Carboniferous, and they maintained this predominant position throughout the Paleozoic. Only during the Early Mesozoic are the synapsids eclipsed by the evolutionary radiation of reptiles (Benton, 1983; Charig, 1984). Within the Late Carboniferous and Early Permian, two different herbivorous and several faunivorous synapsid groups can be recognized.
The fossil record suggests that during the Carboniferous and Early Permian, synapsids and other amniotes were restricted to the paleoequatorial and subequatorial regions. During the Late Permian, the distribution of synapsids, and of some of the other amniotes becomes cosmopolitan. However, the evidence for this pattern rests on rather weak negative evidence (i. e., no Permo-Carboniferous synapsids have been found outside paleoequatorial regions, but other areas have not been prospected intensively).
Early synapsids had a sprawling posture and a small brain, like most early tetrapods. The parasagittal gait characteristic of most mammals (Fig. 1) appeared gradually because some therapsids were apparently capable of sprawling and parasagittal gait, and this character may have appeared in the hind limb before the fore limb. Early synapsids were moderately large (body length between 50 cm and 3 m) and most were carnivorous or insectivorous, but caseids and Edaphosaurus were herbivorous. Mammals include the largest tetrapods that ever lived (the blue whale is larger than any dinosaur), as well as very small species.
Most mammals are viviparous, but the platypus (a monotreme) is oviparous, and most, if not all, non-mammalian synapsids were probably oviparous. All mammals have mammary glands, but the presence of these structures cannot be determined in extinct taxa. The oldest preserved synapsid hair appears to belong to a Middle Jurassic docodontan mammaliaform (a stem-synapsid) from China (Ji et al., 2006). However, hair is infrequently fossilized and the next oldest record is from a Paleocene multituberculate (Meng and Wyss, 1997). Therefore, its real date of appearance is unknown. The position of this group has been debated, but they are probably mammals (Rowe, 1988), so the presence of hair in multituberculates was predictible.
Figure 1. An extant synapsid (a black-tailed deer). Like most of the
other 3500 extant synapsid species, it is endothermic and has fur, but these
features were certainly absent in all Paleozoic synapsids. Photograph by Roger Laurin.
The earliest synapsids were probably ectothermic (their body temperature depended on the environment), as shown by the presence of a large "sail" on the back of edaphosaurids and some sphenacodontids. This sail was composed of greatly elongated neural spines that was grooved anteriorly and posteriorly to accommodate blood vessels, and skin probably connected all the neural spines together. This sail may have allowed these early synapsids to raise their body temperature faster than similarly-sized tetrapods lacking a sail. It was also thought that the sail was used to radiate excess heat into the environment, but recent studies suggest otherwise (Haack, 1986). The presence of the sail suggests that early synapsids lacked hair. Hair usually does not fossilize, but its presence in most monotremes, marsupials, and placentals suggests that it was present in their last common ancestor. Endothermy is also difficult to determine in extinct taxa, but it probably appeared slightly before the first mammal was born. The presence of epidermal horny scales cannot be determined in most fossil taxa, but bony ventral scales were primitively present, as shown by their presence in most reasonably complete fossils of early synapsids.
All known early synapsids had a lateral temporal fenestra (a hole in the lateral surface of the skull behind the orbit) that is still present in a modified form in mammals. Primitively, the lateral temporal fenestra was bordered by only three bones (Fig. 2).
Figure 2. Temporal fenestration primitively found in synapsids.
Abbreviations: J, jugal; Po, postorbital; Sq, squamosal. Scale equals 1 cm.
The occiput was broad and it faced posterodorsally. In other early amniotes, the endochondral bones of the braincase formed a narrower surface, and it was more or less vertical.
The pubis had a long, ridged anterodorsal edge.
Evolution and Systematics
Discussion of Phylogenetic Relationships
Phylogeny of synapsids, after Reisz (1986) and Laurin (1993). Terminology after Kemp (1982), Reisz (1986), and Reisz et al. (1992).
Early synapsid phylogeny is relatively uncontroversial. Recent studies by Reisz (1980, 1986), Brinkman and Eberth (1983), Gauthier et al. (1988), and Laurin (1993) have resolved the relationships between the largest groups of early synapsids. More recent studies by Reisz et al. (1992) and Berman et al. (1995) have confirmed these findings.
Reisz (1980) produced the first phylogeny of synapsids using phylogenetic systematics. He argued that most carnivorous synapsids (ophiacodontids, varanopseids, sphenacodontines, and therapsids formed a clade that excluded the herbivorous edaphosaurids and caseids.
Brinkman and Eberth (1983) argued that varanopseids were closely related to caseids and that edaphosaurids were closely related to sphenacodontids:
Reisz (1986) agreed with Brinkman and Eberth (1983) that edaphosaurids were closely related to sphenacodontids, but he argued that varanopseids were more closely related to ophiacodontids than to caseids:
This new position of varanopseids has not been directly questioned by subsequent studies. The phylogeny of Gauthier et al. (1988) resembled that of Brinkman and Eberth (1983), but they did not test the monophyly of the varanopseid-caseid clade.
Gauthier et al. (1989), Hopson (1991), and Reisz et al. (1992) suggested that Haptodus, previously believed to be a sphencodontid, did not belong to this group:
Laurin (1993) studied the relationships of several problematic taxa previously called haptodontines. He suggested that haptodontines were a paraphyletic group:
See the page on Evolution of early synapsids for synapomorphies justifying the latest phylogenies and for anatomical drawings.