Early hypotheses (such as
lizards) or
rhynchocephalians (such as the
tuatara, pictured) When first named by Merriam in 1904, Thalattosauria was only known by the species
Thalattosaurus alexandrae. Based primarily on the overall skull shape, it was hypothesized to have been close to the reptile order
Rhynchocephalia, which includes
Sphenodon (the living
tuatara). Nevertheless,
Thalattosaurus was recognized as distinct enough to be given its own order, and was tentatively grouped along with Rhynchocephalia in the group
Diaptosauria, a collection of various "primitive" reptiles now known to be
polyphyletic. Within Diaptosauria, thalattosaurs were also considered very closely related to
choristoderes and "Proganosauria" (
parareptiles). Comparisons were also made with Parasuchia (
phytosaurs), Lacertilia (
lizards), and
Proterosuchus, but dismissed as incompatible with proposed evolutionary schemes. Further discussion by Merriam (1905) considered a relationship with
ichthyosaurs due to their similar ecology, but questioned why their skull and vertebral anatomy would diverge so widely if they had a close common ancestor. He proposed that potential similarities were best explained as convergent evolution. The possibility that thalattosaurs diverged from reptiles close to lizards (such as
Paliguana) was described in more detail, with thalattosaurs serving as a short-lived early attempt for near-lizards to return to the sea, an evolutionary process later repeated more successfully when
mosasaurs evolved from true lizards. Nevertheless, Merriam found no clear evidence that any previously known reptile group was directly ancestral to thalattosaurs or vice versa. They were probably descended from land-dwelling Permian reptiles, and not closely related to other marine reptile groups which first evolved in the Triassic.
Modern classification and external relationships , An analysis by
Müller (2004) has even considered thalattosaurs to belong just outside of Sauria. Unusually, thalattosaurs have an affinity to shift near
ichthyosaurs (in the group
Ichthyosauromorpha) when certain basal saurians or near-saurians are excluded from the data set. Some analyses derived from Müller (2004) group thalattosaurs in a "marine superclade" with ichthyosauromorphs and sauropterygians, and sometimes with turtles, archosauromorphs, or lepidosauromorphs as well. For example, Simões
et al (2022) classify thalattosaurs as the
sister group of the sauropterygians, with their clade being sister to the ichthyosauromorphs, and all three being basal
archosauromorphs. However,
cladograms generated by these analyses change in unpredictable ways through alterations to their methodology (such as including or excluding aquatic adaptations or switching between
parsimony and
bayesian inference), leading some to have concerns over the validity of the "marine superclade". While thalattosaurs are almost certainly diapsids, the large degree of uncertainty surrounding their
outgroup relations has led most modern paleontologists to classify them as Diapsida
incertae sedis.
Internal relationships One of the first phylogenetic analyses specifically focusing on thalattosaurs was part of
Nicholls (1999)'s reevaluation of
Thalattosaurus and
Nectosaurus. She used a restricted definition of Thalattosauria which referred to a
clade including all reptiles more closely related to
Nectosaurus and
Hescheleria than to
Endennasaurus or
Askeptosaurus. The more inclusive group including
Askeptosaurus,
Endennasaurus, and traditional thalattosaurs was given the name
Thalattosauriformes.). Meanwhile, the clade containing reptiles closer to askeptosaurids is termed
Askeptosauroidea''
However, uncertainty over Endennasaurus'
s thalattosaurian ancestry led to it being excluded from these analyses. After Müller et al
. (2005) re-affirmed that Endennasaurus
was closely related to Askeptosaurus
, all thalattosaurs known at the time were finally combined into phylogenetic analyses. Studies by Rieppel, Liu, Cheng, Wu, and others continued to identify new Chinese taxa such as Miodentosaurus and various species of Anshunsaurus
and Xinpusaurus
, though homoplasy in these new taxa has led to little resolution in the structure of the two major branches of Thalattosauria. In an attempt to remedy this problem, new phylogenetic analyses were developed by Liu et al
. (2013) during the description of Concavispina'', The internal relationships of thalattosaurs is still considered tentative and inconclusive, although the fundamental structure of the group (a monophyletic Thalattosauria clade split into askeptosauroids and thalattosauroids) is very stable. Some paleontologists have attempted to divide thalattosaurs into families. One family,
Askeptosauridae, is typically considered to include
Askeptosaurus and
Anshunsaurus, with a few studies also placing
Miodentosaurus or
Endennasaurus within it. Another family, Thalattosauridae, was originally used to group
Thalattosaurus and
Nectosaurus, was later redefined to exclude
Nectosaurus, and later still encompassed practically all thalattosauroids. Many thalattosaur-focused paleontologists avoid using family names due to their inconsistent usage and questionable validity. The following
cladogram represents the results of a thalattosaur ingroup
phylogenetic analysis by Druckenmiller
et al. (2020). }} ==List of genera==