Hone and Benton 2007 revisited – part 2

Yesterday was part one of a deep dive
into Hone and Benton 2007 (= HB 007). the paper that attempted to un-discover the origin of pterosaurs in Peters 2000. HB 007 was written by professor and paleo textbook author, MJ Benton, and his student, DWE Hone, now a PhD. This was their first paper together.

Here is part 2,
let’s start with a quote from HB 007: “For this paper the datasets of both Bennett (1996) and Peters (2000) were re-analysed to confirm the results they produced. Figures are not included, as although each analysis performedhere produced a different result to those published, differences in topologies were subtle.”

Back then, datasets were printed on paper. Now datasets are available as electronic files online, which is much better = more accurate and easier to transfer for everyone. When I say ‘printed,’ that means long lists of taxa, long lists of characters and their possible scores, followed by page-filling columns and rows of printed scores. The LRT would have been impractically enormous to publish on paper back then.

When HB 007 reported ‘figures were not included’, that also means they did not examine any pertinent specimens. None. Very odd to do this. With his status, professor Benton gave himself and his student Hone license to skip this step. Pterosaur experts like Kevin Padian and SC Bennett told me I had to visit and examine specimens first hand, as standard practice, in order to be ‘taken seriously,’ so I did so.

The lineage of pterosaurs recovered from the large reptile tree. Huehuecuetzpalli. Cosesaurus. Longisquama. MPUM 6009. ” data-image-caption=”

Figure x. The lineage of pterosaurs recovered from the large reptile tree. Huehuecuetzpalli. Cosesaurus. Longisquama. MPUM 6009.

” data-large-file=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg?w=460″ src=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg?w=460″ alt=”The lineage of pterosaurs recovered from the large reptile tree. Huehuecuetzpalli. Cosesaurus. Longisquama. MPUM 6009.” class=”wp-image-6903″ srcset=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg?w=460 460w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg?w=67 67w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg?w=135 135w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2012/08/huehuecuetzpalli-pterosaur.jpg 588w” sizes=”(max-width: 460px) 100vw, 460px” />

“Bennett’s (1996) main analysis was one of 13 ingroup taxa and a total of 126 characters. The shapes of the trees were only subtly different, with Scleromochlus and the Pterosauria appearing successively below Lagosuchus, rather than as sister taxa in this position (Bennett 1996: fig. 2).”

Note here that HB 007 reported results: but not all of the results. Bennett nested pterosaurs with Scleromochlus between Ornithosuchidae and Lagosuchus + Dinosauria in analysis one. In analysis two (sans Scleromochlus) Bennett nested pterosaurs between two heavy-set, croc-like and likely aquatic clades, Proterosuchidae and Erythrosuchidae. He was brave and forthright to announce those unlikely results. Bennett was just looking in the wrong place, following dogma and tradition.

None of these taxa had a short metatarsal 5 with robust elongated p5.1 phalanx, as seen in tanystropheids and Triassic pterosaurs. None had uropatagia, as seen in Cosesaurus, Sharovipteryx and pterosaurs. The list goes on (see Peters 2000, ignored by HB 007).

See what taxon exclusion gets you? No wonder pterosaurs were so mysterious until Peters 2000 and 2007 added pertinent, overlooked and omitted taxa.
HB 007 fought against the process and the results.

BTW – Scleromochlus was recovered alongside ‘the Pterosauria’,‘ a suprageneric taxon in Bennett 1996. It is much better to get down to the genus level in phylogenetic analysis in order to report which in-group taxon was most similar to the outgroup taxon. In that way the basalmost pterosaur from which all other pterosaurs evolved could be pinpointed, subject to change, of course.

“Peters (2000) presented four datasets that had been previously published and then modified by him. The three analyses from which Peters constructed edited data matrices (Evans 1988; Bennett 1996; Jalil 1997) were briefly reanalysed. However, Peters (2000) did not state which outgroup taxa were used and he did not state what settings were used in PAUP∗ or how the trees were compiled (where consensus trees were shown). He did not perform bootstrap analyses.”

Note here that HB 007 reported NO results published in Peters 2000, distinct from their treatment of Bennett 1996. When HB 007 reported, ‘modified by him’, that sounds bad, but actually means I simply added previously omitted taxa and scored them according to characters previously published by the authors of the four datasets. Also see below.

Note: Among the competing analyses, remember Bennett 1996 was one of the datasets Peters 2000 added taxa to.

When HB 007 reported Peters (2000) did not state which outgroup taxa were used, which settings were used in PAUP, how trees were compiled, and my paper lacked any bootstrap analyses, that’s because I was simply adding taxa to previously published analyses. That was plainly stated. Settings, compilations, etc were cited to each previous author. The present facts indicate that such nitpicking was used to hide the fact that HB 007 did not publish results from Peters 2000, as they did for Bennett 1996,

“Peters (2000) completed two analyses of Evans’ (1988) work although he only published a matrix and resultant tree for one of them. This had a total of 66 characters and 21 ingroup taxa.The outgroup taxon was not specifically identified by Peters (2000) so the most basal taxon from the figured tree (Petrolacosaurus) was selected (this was also necessary for the other analyses below). Peters (2000) recovered six trees of 186 steps using a heuristic search. However, our re-analysis produced eight trees of 191 steps, both using heuristic and branch-and-bound techniques. Although the topologies were broadly similar, consensus trees of these eight (including a strict consensus) were more resolved than the tree published (Peters, 2000: fig. 14, which was probably a strict consensus tree). A bootstrap analysis (performed as above) collapsed seven branches of the cladogram and six of the surviving branches had support of just 67% or lower.”

As stated in Peters 2000, “Evans cladistically analyzed primitive diapsids, considered the Archoauria a single taxon and did not include the Pterosauria, Langobardisaurus, Longisquama or Sharovipteryx. For this paper I duplicated Evans’s data matrix relating to early diapsids and archosauromorphs (97 characters, 31 taxa, Evans’s Nodes A-H). Other characters relating to lepidosauromorphs were examined but not included because pterosaurs do not exhibit any lepiidosauromorph synapomorphies (Bennett 1996a). Benton (1985 noted two lepidosauromoph characters: 1) single ossified sternum and 2) specialized sternal attachment for the ribs. However, the sternum in pterosaurs is a fusion of other pectoral elements (Wild 1993), including a homologous sternum that is plesiomorphic. The ‘specialized sternal attachments’ are simply ossified sternal ribs which are plesiomorphic (Bennett 1996a).”

Boy, was i wrong to believe Bennett 1996a in my freshman paper, because later studies (Peters 2007, the LRT) nested pterosaurs and tanystropheids within the Lepidosauromorpha. Back then, everyone agreed that pterosaurs were Archosauromorphs. Time heals.

Adding taxa in Peters 2007, like Huehuecuetzpalli, first described by Reynoso 1998, proved to be very important. Huehuecuetzpalli was the link, the key, the flashlight in the dark.

Add to that: Villaseñor-Amador, Suárez and Crus 2021 reported potential bipedalism in Huehuecuetzpalli. Bipedalism was a necessary first step in the origin of flapping and then the extension of the forelimbs and then their trailing extradermal membranes into wings in both birds and pterosaurs.

Be ready for change when change happens. Change can happen when taxa are added.

“Peters’ (2000) recoded analysis of Bennett’s (1996) data yielded one tree of 268 steps, but here we recovered one tree of 263 steps (using both heuristic and branch-and-bound searches) with a subtly different topology. Cosesaurus becomes the sister taxon to Longisquama, as opposed to being basal to it in the original tree (Peters, 2000: fig. 16) and Suchia also becomes the sister taxon to Parasuchia, having previously held a more basal position. Bootstrap results were high.”

Again, HB 007 refused to report results regarding pterosaur placement. Instead they nit-picked 268 vs 263 steps likely caused by Cosesaurus moving to become a sister to Longisquama. This might have happened by a typo in a score transferred from the print medium to the digital medium (see below for details). The number of recovered steps is trivial when compared to the overall results, which HB 007 side-stepped.

When HB 007 reported “recoded” that actually means taxa omitted from Bennett 1996 were added to and scored in Peters 2000. If mistakes were found, they were corrected. That’s standard practice in systematics.

“Finally, in his re-analysis of Jalil’s (1997) data, Peters (2000) recovered 120 trees with 151 steps, but here we find 140 trees at 153 with an heuristic search. Using a branch-and-bound approach, however, yielded 240 trees. The topologies were again different compared to Peters’ published tree (which is assumed to be a strict consensus as it nearly matches that of our analysis), with the pairing of Sphenodontia and Iguana nesting with the main polytomy of the other taxa and not basal to them as recovered by Peters (2000: fig. 15). Bootstrap values were low (half were less than 65% for the 6 branches retained). In all three analyses, the recovered CI and Retention Index (RI) values were comparable to those published by Peters (2000).”

Compared to Jalil’s published tree of 18 taxa, the addition of Scleromochlus, Lagosuchus, Eudimorphodon, Longisquama and Sharovipteryx resulted in loss of resolution at two nodes: the base of the Tanystropheidae and the base of the Protosauria + Proterosuchus + Choristodera + Rhynchosaurus + Trilophosaurus. This is rather similar to the one of their own making reproduced in HB 007.

Again, HB 007 failed to report than the sole pterosaur, Eudimorphodon nested with tanystropheids + Jesairosaurus, rather than with Proterosuchus, the sole archosauriform. nor with Scleromochlus and Lagosuchus, the two archosaurs.

In summary HB 007 chose to nitpick rather than report results.
Rather than examining the specimens first-hand.
Rather than reprinting published skeletons of the taxa.
Rather than calling or emailing if confused.
The evidence shows: they were on a mission.

Two other studies
merit inclusion here…

In one of the first phylogenetic analyses using software 
Benton 1985 nested pterosaurs between the suprageneric clades Lepidosauromorpha and Rhynchosauria. This will be the last time pterosaurs nest with lepidosaurs for two decades.

Bennett 2012 reported with regard to Hone and Benton (2007, 2008), 
“Hone kindly sent me their Nexus files, and comparing them with the published data sets of Bennett (1996) and Peters (2000) revealed that the inability of Hone and Benton (2007) to replicate the analysis of Bennett (1996) was not the result of differences in the coding of Scleromochlus for Char. 43, but rather was the result of (1) four separate coding errors in retyping the data matrix and (2) leaving eight of the 9s that Bennett (1996) used to code for missing data in the matrix unchanged while using ? as the symbol for missing data and then including 9 in the Symbols statement of the Nexus file so that PAUP treated 9 as a distinct character state rather than as missing data. Similarly, their inability to replicate Peters’ (2000) analysis of a modified Bennett (1996) data set was the result of seven coding errors in retyping Peters’ published data matrix plus two 9s treated as a distinct character state.” 

See, typos happen. Typos get published in academic journals.

More to come in part 3
I hope you’re getting a taste of what it is like out there for independent researchers who add taxa and recover new interrelationships over time.

Lesson for today: Stepping on academic toes will get you a punch in the nose.
Remedy: do your best, acknowledge mistakes, report results, be brave, be kind, be skeptical, and be ready for change. Adding taxa often changes hypothetical interrelationships recovered in phylogenetic analysis.

This is science, not dogma.
That’s why college textbooks always have second, third and fourth editions.

References
Bennett SC 1996. The phylogenetic position of the Pterosauria within the Archosauromorpha. Zoolological Journal of the Linnean Society 118: 261–308.
Bennett SC 2012. The phylogenetic position of the Pterosauria within the Archosauromorpha re-examined. Historical Biology. iFirst article, 2012, 1–19.
Benton MJ 1985. Classification and phylogeny of the diapsid reptiles. Zoological Journal of the Linnean Society 84: 97–164.
Benton MJ 1999. Scleromochlus taylori and the origin of dinosaurs and pterosaurs. Philosophical Transactions of the Royal Society London, Series B 354 1423-1446.
Evans SE 1988. The early history and relationships of the Diapsida. Pp: 221–260 in: Benton MJ (ed) The phylogeny and classifaction of the tetrapods, volume 1: Amhiphians, Reptiles, Birds. The Systematics Association special volume 35A. Clarendon Press.
Hone DWE and Benton MJ 2007. An evaluation of the phylogenetic relationships of the pterosaurs to the archosauromorph reptiles. Journal of Systematic Palaeontology 5:465–469.
Jalil N-E 1997. A new prolacertiform diapsid from the Triassic of North Africa and ther interrelationships of the Prolacertiformes. Journal of Vertebrate Paleontology 17ª3):506–525. Peters D 2000b. A Redescription of Four Prolacertiform Genera and Implications for Pterosaur Phylogenesis. Rivista Italiana di Paleontologia e Stratigrafia 106 (3): 293–336.
Peters D 2007. The origin and radiation of the Pterosauria. In D. Hone ed. Flugsaurier. The Wellnhofer pterosaur meeting, 2007, Munich, Germany. p. 27. 
Peters D 2009. A reinterpretation of pteroid articulation in pterosaurs. Journal of Vertebrate Paleontology 29: 1327-1330.
Reynoso V-H 1998. Huehuecuetzpalli mixtecus gen. et sp. nov: a basal squamate (Reptilia) from the Early Cretaceous of Tepexi de Rodríguez, Central México. Philosophical Transactions of the Royal Society, London B 353:477-500.
Sanz JL and López-Martinez N 1984. The prolacertid lepidosaurian Cosesaurus aviceps Ellenberger & Villalta, a claimed ‘protoavian’ from the Middle Triassic of Spain. Géobios 17: 747-753.
Villaseñor-Amador D, Suárez NX and Crus JA 2021. Bipedalism in Mexican Albian lizard (Squamata) abd the locomotion type in other Cretaceous lizards. Journal of South American Earth Sciences 109:103299.

wiki/Huehuecuetzpalli