Updated June 23, 2017 with the removal of Phlegethontia after taxon additions attracted that taxon to the Aïstopoda, where it traditionally nests.
I asked for the challenge.
Dr. David Marjanović (DM) responded. He thought the traditional collosteids should nest together, as they do in Marjanović and Laurin 2017. By contrast, in the large reptile tree (LRT, 1012 taxa) only two nest together. Dr. David Marjanović also did not like Colosteus and kin nesting between Osteolepis and Panderichthys. Rather, Marjanović and Laurin 2017 reported, “Colosteidae is consistently found in a position one node more rootward than Baphetoidea and one node more crownward than Crassigyrinus.” In the Marjanović and Laurin study, relatives of Baphetes include Spathicephalus, Eucritta and Megalocephalus and Crassigyrinus nests between the collosteids and Tulerpeton. The LRT does not support this topology.
Marjanović and Laurin report,
“Deltaherpeton is one of the oldest known colosteids.” Bolt and Lombard 2010 report, “Deltaherpeton is unique among colosteids in having an internasal and single midline postparietal. An additional midline pair of cf. ‘interfrontonasals’ may be present. Synapomorphies which unite Deltaherpeton, Colosteus, Greererpeton, and Pholidogaster as Colosteidae are:
- premaxilla with fang pair;
- dentary with notch for receipt of premaxillary fang;
- mandible with single elongate exomeckelian fenestra;
- pre-narial infraorbital lateral line terminating at ventral margin of premaxilla just anterior to external naris; and
- post-narial infraorbital lateral line terminating at the ventral margin of the maxilla just posterior to the external naris.
Let’s test to see
if this list is just a Larry Martin list of a few traits that are overwhelmed by other synapomorphies in the LRT. And at the same time, let’s see if these few traits have a wider, but overlooked, distribution and to see if they are valid for every included taxon.
Premaxilla with (lateral) fang pair
is indeed present in the four named taxa, if only barely in Deltaherpeton. Overlooked, perhaps, the lateral premaxillary tooth is also the largest in Acanthostega, Ventastega, Pederpes, Sclerocephalus, Ichthyostega among taxa related to traditional colosteids. More on premaxillary fangs below.
The dentary notch
Unfortunaely I see this trait only on Greererpeton and Pholidogaster. In Colosteus (Figs. 1, 6) and Delatherpeton (Figs 3, 5) it is not apparent.
Collosteid traits 3-5 (above)
include the elongate exomeckelian fenestra and the two lateral lines (Fig. 3) are difficult to see or not see in some taxa. Note in the labeled image of Deltaherpeton by Bolt and Lombard 2010 (Fig. 3). Even they were unable to draw the naris and its surrounding lateral lines (white glow), but provided a diagram (Fig. 3) on another figure.
I have not looked for lateral line/naris patterns
in other taxa, but Bolt and Lombard note the lateral line is continuous and straight below the naris along a lateral rostral plate in Eusthenopteron, Panderichthys (Fig. 4) and Ichthyostega. The lateral rostral plate is below the naris in Eusthenopteron. In Panderichthys the lateral line does not cross the lateral rostral plate, if that is what it is, because it is illustrated by Vorobyeva and Schultze (1991) with teeth, so it may just be a broken portion of the maxilla and the lateral rostral plate is no longer present. The naris of Ichthyostega is at the jawline leaving little room for a lateral rostral plate on the exterior surface. Would have been better for Bolt and Lombard to provide both the data, for verification, and the diagram, because now doubts arise.
Two taxa separate Greererpeton and Pholidogaster in the LRT:
Panderichthys and Tiktaalik. Both lack the lateral premaxillary fang. Notably and despite their antiquity, both are derived and distinct from related taxa in having a very flat skull with orbits close to the midline. All marginal teeth are relatively tiny, which is also distinct from related taxa. Apparently when the skull flattened in these two the lateral premaxillary fangs shrank. Perhaps we should look for them in undiscovered basal taxa, probably originating n the early Late Devonian and lasting who knows how long.
In the large reptile tree
(subset Fig. 3) flat Greererpeton nests with other flat taxa like Spathocephalus, Trimerorhachis and Gerrothorax and Ossinodus. Deltaherpeton nests with other tall- and narrow-skulled taxa, like Crassigyrinus and Ventastega. Pholidogaster and Colosteus nest with other round skull taxa. Despite their readily apparent differences no other taxa in the LRT share as many traits with clade members. And these three are the few representatives of a radiation covering about 60 million years. Just think what the mammals did in 60 million years.
The taphonomic crushing of the Deltaherpeton skull suggests it was wider than it was. Finding the palate by excavating from the other side of the matrix would provide precise data.
Did temnospondyls return the water?
Or never leave it? Did tetrapods develop fingers and toes more than once? Did basal tetrapods develop the ability to raise their bellies off the substrate more than once? The LRT provides provisional answers to these questions (Fig. 7). Convergence is apparent here. The LRT collosteids are separated from start to finish by about 60 million years, so changes can be expected.
Rather than pulling a Larry Martin, I did not list the few or many traits shared by Collosteus relatives in the LRT. Those can be gleaned from the matrix and most certainly will find convergences elsewhere on the cladogram. Remember, its not just one or a dozen traits that nest taxa as a clade, but the suite of traits that can really only be recovered by software like PAUP.
As to Dr. Marjanović’s challenge:
The traditional list of collosteids certainly does fall into a much narrow spectrum of sizes (Fig. 1), as opposed to the LRT list of Collosteus relatives (Fig. 6). And I did reexamine several issues and red flags. Some scores were revised. Deltaherpeton shifted two nodes and I think I understand it much better now. The list of classic collosteid traits is not found in all members and some traits extend to other clades. Finally, the phylogenetic distance between classic collosteids is not far from each other in the LRT, and in both studies both collosteid clades nest toward the base of the Tetrapoda. The details will work themselves out with further study on both sides. All interpretations are provisional, especially in basal tetrapods given their lateral lines that sometimes look like sutures and both camouflaged by a maze of skull texture.
As a suggestion for the future:
if colleagues would colorize their skull photos, paying attention to broken pieces and parts that just barely peek out from overlying material, that would go a long way toward improving the present system of either just showing the specimen or creating a freehand outline of the specimen, or just labeling bones with abbreviations and arrows without noting sutures.
Marjanović D and Laurin M 2017. Reevaluation of the largest published morphological data matrix for phylogenetic analysis of Paleozoic limbed vertebrates. PeerJPrePrints (not peer-reviewed).
Vorobyeva EI and Schultze H-P 1991. Description and systematics of panderichthyid fishes with comments on their relationship to tetrapods, in Schultze and Trueb (eds.), Origins of the Higher Groups of Tetrapods Comstock, pp 68-109.