Scientists have observed that the debate lacks a clear consensus exacerbated by longstanding issues in mosasauroid phylogenetics. The construction of reliable fossil phylogenies requires extensive analysis of many evolutionarily meaningful anatomical traits (characters) of taxa. Such are encoded in a character matrix. Due to their treatment of each taxon as an operational taxonomic unit (OTU), phylogenetic analyses are sensitive to the contents of their character matrices. Incomplete or defective sampling of taxa and characters can misdirect or confound an analysis, harming the resulting topology's resolution (frequency of undesirable soft polytomies) and support (clade accuracy or confidence). This consequentially misrepresents a group's evolutionary history to varying degrees.
Two particular problems plague mosasauroid phylogenies. First, the fossil record of Cenomanian and Turonian mosasauroids is poor. Fossils of Turonian mosasaurs are particularly scant,[6] creating early ghost lineages in all three major mosasaur groups that may have lasted up to ten million years.[7] Cenomanian-Turonian aigialosaurs are better represented in diversity, but most fossils do not preserve key characters such as those located in the skull. This has resulted in particularly unstable early-diverging branches that Madzia and Cau (2017) describes as "weak spots." One solution is to reduce character incompleteness in the matrix by omitting poorly preserved aigialosaurs. For example, Palci et al. (2013) removed Haasiasaurus and Carsosaurus from their analysis due to lack of attributable skull characters, a trend followed in subsequent studies. But this comes at the cost of masking potentially critical OTUs.
The second phylogenetic problem is the character matrix itself. Virtually every 21st-century mosasauroid study recycles a 142-character matrix of North American mosasaurs and Adriatic aigialosaurs from Bell (1997)'s analysis of North American mosasaurs and Adriatic aigialosaurs. Subsequent research found that the Bell matrix is outdated; it contains several questionable characters and does not represent many traits that are significant to evolution. Taxonomic representation was also problematic as the omission of foreign taxa impedes accuracy. The matrix has since been been revised over the following decades. However, Madzia and Cau (2017) found them insufficient to adequately address the matrix's fundamental weaknesses. Zietlow et al. (2023) further identified several key traits that remain unrepresented and observed that the revised matrix is still heavily biased towards North America to the neglect of well-understood European and African species. Lively (2018) noted that now-invalid taxa are still included in analyses, misleading them with inflated OTUs.
The following cladograms illustrate the controversy of mosasaur evolution.
More information Ancestral state reconstruction by Simões et al. (2017) ...
Topology A:
Free-standing iliac crest
Ambiguous state |
Close
More information Strict consensus of maximum parsimony by Makádi et al. (2012) ...
Topology B:
Plesiopedal and plesiopelvic taxa
Plesiopedal and hydropelvic taxa
Hydropedal and hydropelvic taxa
Taxa with unknown limb structure |
Close
More information Maximum clade credibility tree by Mekarski (2017) ...
Topology C:
Mosasaur taxa |
Close
Augusta, B.G.; Zaher, H.; Polcyn, M.J.; Fiorillo, A.R.; Jacobs, L.L. (2022). "A Review of Non-Mosasaurid (Dolichosaur and Aigialosaur) Mosasaurians and Their Relationships to Snakes". In Gower, D.J.; Zaher, H. (eds.). The Origin and Early Evolutionary History of Snakes. Cambridge University Press. pp. 157–179. doi:10.1017/9781108938891. ISBN 9781108938891.
Caldwell, M.W.; Palci, A. (2007). "A new basal mosasauroid from the Cenomanian (U. Cretaceous) of Slovenia with a review of mosasauroid phylogeny and evolution". Journal of Vertebrate Paleontology. 27 (4): 863–880. doi:10.1671/0272-4634(2007)27[863:ANBMFT]2.0.CO;2.
Caldwell, M.W. (2012). "A challenge to categories: "What, if anything, is a mosasaur?"". Bulletin de la Société Géologique de France. 183 (1): 17–34. doi:10.2113/gssgfbull.183.1.7.