Bryan Gee
LightNews LightNews
banner
koskinonodon.bsky.social
Bryan Gee
@koskinonodon.bsky.social
1.4K followers 390 following 190 posts
- Research data librarian @ UT Austin
- 🐶 + 🐱 dad
- Amphibian paleontologist ☠️🐸
- BA in geology, PhD in ecology/evol bio

he/him; all opinions my own

🔗: https://bryangee.weebly.com/
Posts Replies Media Videos
koskinonodon.bsky.social
And for included taxa, more than a third (31 of 88) underreported the largest published value by 10% or more. Mehmood et al. sometimes report higher values than TEMNOS - sometimes this is b/c they extrapolated from a reconstruction (I don't do this yet), but some of these values are clearly errors.
Comparative graph showing the maximum skull length recorded in TEMNOS vs. the value reported by Mehmood et al. for all species where the difference is at least 10%. 31 have a higher value in TEMNOS; 7 have a higher value in Mehmood et al. Graph showing percent difference between TEMNOS and Mehmood et al. for species where the values differ by at least 10%
August 5, 2025 at 2:52 PM
koskinonodon.bsky.social
Despite claiming to have comprehensively sampled temnospondyls, this study failed to include published data from nearly as many species that met inclusion criteria (graph) as species that were included. Max. skull length of every missed species that was published by 2024 is shown below.
Plot showing maximum recorded midline skull length of every temnospondyl species that met inclusion criteria for Mehmood et al.'s study but that were not included
August 5, 2025 at 2:52 PM
koskinonodon.bsky.social
Phylogenetic analysis and qual comparisons suggest Aspidosaurus nests higher in the dissorophid tree than was previously thought on the basis of the (probably) chimeric A. binasser, putting it much closer to the well-known Cacops. Still a lot in flux re: dissorophid phylogeny though (15/n)
Phylogenetic results showing Aspidosaurus chiton clusters with dissorophids like Cacops
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
Broili's description of the skull is very brief - no sutures or much to go on - but superficially, the Smithsonian specimen looks about as similar to the lost holotype as is possible from 120yo figures and can be readily differentiated from all other dissorophids with skulls (11/n)
Screenshot of illustrations from Broili (1904) of his new species, Aspidosaurus chiton. The figures include a partial skull in dorsal view (incomplete posteriorly) and various views of postcrania, mainly shingle-shaped osteoderms and associated vertebrae
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
This new specimen comprises a nearly complete skull and substantial amounts of postcranial material, including a block of several articulated osteoderms and vertebrae. When I looked at the osteoderms, I immediately recognized them as identical to what Broili had figured for Aspidosaurus chiton (9/n)
Photographs of the skull of the new specimen in right and left lateral views. Parts of the skull are missing or damaged. Photographs of the osteoderms associated with the Smithsonian specimen in various views. The osteoderms are dorsally ornamented with circular pits of equal size and are about as wide as they are long. They slope downwards to the side, forming an inverted-V shape if viewed from the front/back. Each osteoderm slightly overlaps onto the back of the one in front of it. In a broken cross-section, we can see a second, hidden osteoderm that caps the tip of the spine and that is much smaller, barely larger than the tip. It is thus obscured entirely by the ornamented osteoderm when articulated
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
A number of new taxa were named exclusively on the basis of a single osteoderm, typically not a solid basis for a new taxon, and the genus has an extreme temporal range of over 25myr as a result. Various examples of some holotypes of other species are shown below (6/n)
Photographs of various holotypes of species assigned to Aspidosaurus, depicting the marked variation in osteoderm morphology. For example, some osteoderms are very wide and dorsally curved, while others are narrow and have a distinct apex or ridge along the midline
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
In 1901, a German paleontologist named Ferdinand Broili visited the U.S. to collect Permian fossils from the classic Texas redbeds with Charles Sternberg. Several years later in 1904, he published a lengthy monograph in which he named a new dissorophid, Aspidosaurus chiton (2/n)
Select digitized screenshots of the holotype of Aspidosaurus chiton, including the partial skull and articulated series of osteoderms. Illustrations from Broili (1904)
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
New paper out today in @zoojlinnsoc.bsky.social for #FossilFriday AKA here's-Bryan's-one-paleo-paper-this-year-to-remind-people-I-still-do-paleo-sometimes: a revision of the #Permian #dissorophid #temnospondyl Aspidosaurus chiton based on a new specimen: doi.org/10.1093/zool... (1/n)
Photograph and illustration of the skull of the new neotype of Aspidosaurus chiton in dorsal view. The skull is an orange-brown color, longer than wide, and largely undistorted.
July 11, 2025 at 3:48 PM
koskinonodon.bsky.social
And discusses the need for (re)curation, lots of case studies about current/previous systemic metadata issues. Dryad publication years in DataCite all out of sorts right now. Also, I think I discovered a huge screw-up where Figshare assigned the wrong affiliation to thousands of datasets! (5/5)
Line graph comparing three date fields, two from DataCite and one from Dryad, for Dryad datasets. At some point, almost all 2024 publications got reclassified as 2025 in DataCite, does not occur for other repositories but occurs in Dryad for both UT Austin and entire corpus of data.
July 3, 2025 at 1:35 PM
koskinonodon.bsky.social
Also discusses some of the metadata assessments we're doing (i.e. what are we doing with the data once we have them). Examples including identifying mislabeled deposits (e.g. datasets with code+data or just code; licensing) (4/5)
Horizontal bar chart showing distribution of UT Austin datasets based on whether they have a software file in them. Most datasets do not or don't crosswalk that metadata, but about one in six of the datasets that do crosswalk it have at least one software file. Horizontal bar chart comparing licensing treatment among UT Austin datasets. CC0 is the most common but this is mainly due to Dryad only using CC0 and our institutional repository setting that as the default. CC0 is third if those two are removed. Rare software licenses again indicate some mislabeled deposits.
July 3, 2025 at 1:35 PM
koskinonodon.bsky.social
The preprint is mainly designed to explain the gory details of the process, although it does show data for UT Austin specifically, like how many different ways people write "UT Austin" and whether an affiliated researcher is first/last author or not (proxy for how involved they were) (3/5)
Horizontal bar chart comparing permutations of UT Austin (n=17) that were identified through the DataCite API. The official university name (The University of Texas at Austin) is the second most common, being outranked by the permutation without a leading "The". Unpredicted variations of punctuation and casing were also found. Horizontal bar chart showing distribution of UT Austin datasets by whether an affiliated researcher is listed as first, last, first and last, or neither in dataset metadata. Single author datasets are actually the most common. Ones without an affiliated first or last author are more common than those with both, only first, or only last, individually
July 3, 2025 at 1:35 PM
koskinonodon.bsky.social
As I've previously mentioned, I've gotten creative with this and can get datasets with no DOIs that do have affiliations (like NCBI) and datasets with DOIs but no affiliations (like automated Figshare deposits mediated through journal partners)
Workflow diagram showing how Figshare deposits with no affiliation metadata can be linked to a given institution through a complicated backdoor using scholarly articles and metadata linkages. The process involves a sweeping capture of Figshare deposits and a targeted capture of affiliated articles from select publishers and then cross-referencing them.
July 3, 2025 at 1:35 PM
koskinonodon.bsky.social
I see our cats have similar taste...
Photograph of a houseplant that has some mysterious bite marks on the ends of leaves
May 1, 2025 at 7:48 PM
koskinonodon.bsky.social
Boardroom meeting meme where the boss asks how to get high-quality temno data. One person says the PBDB [Paleobiology Database], another says "skim papers", and a third says "Ask a temno worker", which is not well-received.
March 5, 2025 at 2:49 PM
koskinonodon.bsky.social
Astronaut gun meme where one astronaut says "wait the PBDB [Paleobiology Database] isn't reliable for temnos?" The one with the gun says "Never has been."
March 5, 2025 at 2:44 PM
koskinonodon.bsky.social
Feeling spicy today

#temnospondyl #temnospondyls
Anakin-Padme meme where Anakin says "we're going to do a big data study on temnos" and Padme says "you're going to consult a temno expert right?" Anakin doesn't respond and Padme repeats herself.
March 5, 2025 at 2:41 PM
koskinonodon.bsky.social
What an ominous footer on the homepage hahaha
Screenshot of the Zoobank homepage with a footer saying the site recently underwent maintenance and "should be fully functional"
March 4, 2025 at 12:16 PM
koskinonodon.bsky.social
Created an exhibit @burkemuseum.bsky.social on how 250 million-year-old amphibian fossils from Antarctica inform both our understanding of deep time environments and modern systems as part of my #NSF funded postdoc (seen by >200k visitors)

#BroaderImpacts

antarcticsun.usap.gov/science/4714/
Photograph of an exhibit display at the Burke Museum (Seattle, WA) that displays artistic reconstruction of the temnospondyl Antarctosuchus, a cast and interpretive sculpture of the temnospondyl Micropholis, and text explaining how we use deep time records (of amphibians) to understand the present.
January 29, 2025 at 2:11 PM
koskinonodon.bsky.social
Not everyone can have their own set of gifs Hans
Smithsonian paleontology curator Hans Sues questioning whether he's become too popular
January 24, 2025 at 10:24 PM
koskinonodon.bsky.social
(Here is a #metoposaurid from the LACM)
Photograph of a skull of a Late Triassic metoposaurid temnospondyl from Texas, on display at the LA County Museum of Natural History. The skull is vertically mounted with the ornamented dorsal surface facing the viewer.
January 21, 2025 at 3:53 PM
koskinonodon.bsky.social
#TemnospondylTuesday
Danny DeVito meme from It's Always Sunny in Philadelphia: "Can I offer you an egg in this trying time?", modified to swap 'temnospondyl' for 'egg'
January 21, 2025 at 3:51 PM
koskinonodon.bsky.social
What personally bugs me as a researcher is how those two harness special issues. As a reviewer, I have terrible exp with special issues - my comments are almost always entirely ignored, editors don't reject or even force revisions. Have seen unanimous reject get pub'd as is

doi.org/10.1162/qss_...
Comparative plot showing total publishing volumes among major scholarly publishers in 2016, 2019, and 2022, with proportion comprised by special issue articles colored in red. Frontiers, MDPI, and the now defunct Hindawi have significantly higher rates of special issue publication.
January 19, 2025 at 12:46 AM
koskinonodon.bsky.social
Some of us live with micromanagers unfortunately
Photo of a calico cat loafing on a MacBook preventing their servant from working
January 4, 2025 at 4:16 PM
koskinonodon.bsky.social
Elias et al. propose a novel rhinesuchid from the Middle-Late Permian of Brazil (what happened to "don't name new taxa in preprints"?) based on a partial skull. It clusters with the one Triassic rhinesuchid, Broomistega, in phylogenetic analyses. [preprint, not accepted] ssrn.com/abstract=491...
Watermarked screenshot of a figure showing a photograph of the partial skull with semi-transparent colours used to shade different cranial elements. The orbits are mostly preserved, as is most of the snout; it remains embedded in matrix on the dorsal surface.
January 3, 2025 at 3:24 PM
koskinonodon.bsky.social
Some preprints: @akufner.bsky.social et al. (CC: @calamanderso.bsky.social @paleobadger.bsky.social) provided a taphonomic analysis of a monotaxic metoposaurid bonebed from Wyoming, the 1st documentation of Buettnererpeton bakeri in the Popo Agie Fm. [preprint, not accepted] doi.org/10.1101/2024...
Photographs of skulls of Buettnererpeton bakeri from the Late Triassic of Wyoming in dorsal and ventral views, along with interpretive line drawings. The skulls are a light brown-grey colour, with the typical ornamentation and deeply incised lateral line grooves of metoposaurids. The depicted specimens are nearly complete skulls. One preserves a covering of denticulate plates that would have covered the interpterygoid vacuities in life. Comparison of the relative completeness of skeletons among three metoposaurid bonebeds: Elkins Place (type locality of Buettnererpeton bakeri in Texas), Site XIII (type locality of Dutuitosaurus ouazzoui in Morocco), and the newly described Nobby Knob site in Wyoming. The minimum number of individuals is used to calculate what percent of skeletons are complete with respect to different elements. The large, flat, ornamented elements (clavicle, interclavicle, skull) tend to be best represented, while oddly shaped elements like pelvic girdles, neural arches, and haemal arches, are not necessarily rarer in absolute count but are less-represented in terms of how many would be predicted to occur based on the minimum number of individuals (e.g., there should be dozens of neural arches in one individual compared to one interclavicle per individual).
January 3, 2025 at 3:24 PM