Earlier we looked at the holotype of Pholidophorus.
Almost finless Annaichthys is definitely a pholidophorid,
but it also nests with eel-like Tarrassius at the base of the Lepisosteus (the extant long nose gar) clade. Several transitional taxa separate these two distinct taxa.
Annaichthys pontegiurinensis (Arratia 2013; MCSNB 11282a,b,c; Triassic) is known from one fossil in part and two counterparts. This taxon nests with Tarrasius. Small fins are visible on the fossil with ventral and lateral surfaces exposed. Pholidophorids are traditionally considered actinopterygii, but in the large reptile tree (LRT, 1675+ taxa) they nest near the base of a large clade of stem lobefins. That means, over deep time, tetrapods are also pholidophorids.
Previously Tarrasius entered the LRT
despite a distinctive tadpole-like morphology (straight tail and lacking pelvic fins).
Pholidophorids are traditionally considered extinct,
but in the LRT the arowana, Osteoglossum, is an extant pholidophorid. So are tetrapods.
According to Arratia 2020,
“Pholidophoriformes Berg (1937) is a poorly known assemblage of Mesozoic actinopterygian fishes whose close association with teleosts and their closest relatives makes them an important group for the understanding of the modern radiation of fishes.”
“More recently, however, the monophyly of †Pholidophoriformes and the relationships of its constituent families have been cast into doubt. The rhombic ganoid scales with peg-and-socket articulation and the elongate or fusiform body shape shared by †pholidophoriforms are now recognized to have a broad distribution within primitive actinopterygians. As a result, the nature and phylogenetic affinities of the various taxa constituting †Pholidophoriformes (collectively referred to here as ‘pholidophoriforms’) are uncertain.”
In Arratia’s 2020 memoir,
Annaichthys was considered, but Tarrasius was not. “Pholidophorus and its somewhat vague original definition (Agassiz, 1832), which has led to it becoming a taxonomic wastebasket for fish possessing rhombic ganoid scales.”
Arratia disputed Berg 1937 when reporting,
“Feature 1, scales and bones built as in Lepidosteus (= Lepisosteus), is difficult to evaluate because it could refer to many different aspects of the scale (e.g., thickness, rnamentation).Even so, there are major differences in the scales and bones of taxa identified as pholidophoriforms by Berg (1937), so it is not clear how this feature unites the group.”
Berg listed 12 traits he thought were shared by Pholidophormes.
In doing so Berg ‘pulled a Larry Martin‘. There is only one way to define any clade and that is by phylogenetic analysis as it determines a last common ancestor. Of course, this may change as taxa are added, so start with a wide gamut analysis like the LRT.
Arratia’s 2020 figure 95 cladogram
nested two dissimilar taxa, Amia (the bowfin) and Lepisosteus (the long nose gar) as outgroup taxa to a clade Teleosteomorpha, which included Pachycormus in a basal clade. The LRT includes more outgroup taxa and does not support this topology.
Arratia G 2020. Morphology, taxonomy, and phylogeny of Triassic pholidophorid fishes (Acinopterygii, Teleostei). Memoir Journal of Vertebrate Paleontology 33:sup1:1–138.
Pholidophoridae Woodward 1890