Tritylodontids ("three knob teeth", named after the shape of animal's teeth) were small to medium-sized, highly specialized and extremely mammal-like cynodonts. They were the last known family of the non-mammalian synapsids. One of the last cynodont lines to appear, the Tritylodontidae descended from a Cynognathus-like cynodont. The Tritylodontids were herbivorous, chewing through vegetation, such as stems, leaves, and roots.
The tritylodont's skull had a high flat crest. They retained the reptilian joint between the quadrate bone of the skull and the auricular bone of the lower jaw, but they were reduced. It is only through the retention of the vestigial reptilian jawbones that they are technically regarded to not be mammals. The back of the skull had huge zygomatic arches for the attachment of its large jaw muscles. They also had a very well-developed secondary palate. The tritylodont dentition was very different from that of other cynodonts. They did not have canines. The front pair of incisors were enlarged that were very similar to rodents of today. Traversodonts had a large gap, the diastema, that separated the incisors from the square-shaped cheek. Each of the cheek teeth in the upper jaw had three rows of cusps running along its length that had grooves in between. The lower teeth had two rows of cusps which fitted into the grooves in the upper teeth. The matching of the cusps allowed the teeth to meet in a precise bite. It would grind its food between the teeth in somewhat the same way that a modern rodent would with their food. The teeth were well suited for shredding plantsmatter.
The Tritylodonts can very much be seen as Mesozoicrodents. Tritylodonts were active animals that were probably warm blooded and burrowed like modern day rodents. For example, Oligokyphus could be compared to a weasel or mink, with a long, slim body and tail. Its legs had evolved directly beneath the body, as they have in mammals.
Because of their extremely mammal-like appearance, tritylodontids were originally placed within Mammalia. Starting with the work of British paleontologist D. M. S. Watson in 1942, a close relationship began to be favored between tritylodontids and cynodonts. Watson and other paleontologists noted that tritylodontids lacked a connection between the dentary and squamosal bones in the lower jaw that was characteristic of early mammals. Haughton and Brink (1954) were the first to classify tritylodontids within Cynodontia. Later studies identified close similarities between the teeth of tritylodontids and traversodontids, and tritylodontids were eventually thought to be descendants of traversodontids. Under this classification, which was widely accepted in the following decades, Tritylodontidae is a member of Gomphodontia, part of the larger group of distant mammal relatives called Cynognathia. The name Tritylodontoidea has also been used for the group, which traditionally includes the families Diademodontidae, Trirachodontidae, Traversodontidae, and Tritylodontidae.
More recently, tritylodontids have been reinterpreted as close relatives of mammals. Beginning with Kemp (1983), Tritylodontidae has been proposed by numerous studies as a member of Probainognathia, the same cynodont group that contains Mammalia. Gomphodontia is still used for the cynognathian group containing traversodontids and is preferred over Tritylodontoidea now that tritylodontoids are not part of it. A phylogenetic analysis performed by Liu and Olsen (2010) places Tritylodontidae very closely to Mammalia, as the sister taxon of the clade formed by Brasilodontidae and Mammalia. Ruta et al. (2013) phylogenetic analysis which is partially based on Liu and Olsen (2010) places Tritylodontidae in a more derived position than Brasilodontidae. Below is a cladogram from this analysis.
^Liu, J.; Olsen, P. (2010). "The Phylogenetic Relationships of Eucynodontia (Amniota: Synapsida)". Journal of Mammalian Evolution17 (3): 151. doi:10.1007/s10914-010-9136-8.
^Ruta, M.; Botha-Brink, J.; Mitchell, S. A.; Benton, M. J. (2013). "The radiation of cynodonts and the ground plan of mammalian morphological diversity". Proceedings of the Royal Society B: Biological Sciences280 (1769): 20131865. doi:10.1098/rspb.2013.1865.edit