If you scan "The Compact Thescelosaurus" with the data sorted by publication year, dubious names are quite common as late as the 1960s, although note that's only a third or so of the list. (Also, people were still taking random theropod fossils and making them species of Megalosaurus into the 1970s, because it worked so well the first couple dozen times. Megalosaurus chubutensis, anyone? Readers, if ever you are confronted with a theropod tooth or vertebra and it occurs to you to describe it as a new species of Megalosaurus, find something else to do until the urge passes.) The stretch ending with 1876, just before the Morrison dinosaurs show up in bulk, is especially dire. Out of 96 entries, I have 79 marked as dubious or potentially dubious, and I don't think I've been unusually kind or unkind with the metaphorical red pen. That means that for the first five decades or so of formal study of non-avian dinosaurs, a little more than four out of every five names are really difficult to apply to anything but that original specimen. The problem is obvious in hindsight: they didn't know all that much about dinosaurs (in particular, they seriously overestimated the taxonomic utility of dinosaur teeth) and they didn't have the best material, so what they were doing amounted to throwing things at the wall to see what stuck, and all they got was a mess. (For a change of pace, once decent material started showing up, we got a different kind of mess, which was "every decent specimen becomes a new species". You can't win.)
The end result was a pile of names based on fossils of marginal quality. The question becomes how to handle them, a tug of war between priority (oldest name takes the cake), stability (try not to shake up accepted usage), and utility (the more complete and well-preserved the type specimen, the more useful it is). Once we throw in the special sauce that is human nature, we get the delightful pastime of arguing validity and synonymy. Of course, validity and synonymy are subjective and liable to change (except of course for objective synonymy, where multiple species are named for the same fossil), although some cases are more subject to change than others. (Again, you can't win.) For the record, nearly 200 years of data have shown that naming a dinosaur from teeth, finger or toe bones, ribs, chevrons, or isolated scutes is demanding trouble. Isolated vertebral centra are also up there, but if you can get the neural arch as well, that helps.
If you are faced with one of these old names, there are a number of ways you can go. Each ancient name has different circumstances. Just remember that you cannot satisfy everyone, but people will certainly be less satisfied if you declare synonymy or dubiousness without going into details, act in such a way that is suspiciously convenient for a species you named, or both.
It's important to distinguish between validity and synonymy. If a species is identified as dubious, the implication is that the specimen it's based on is inadequate to allow you to assign more material to the species, or to distinguish it from other species. If you can show that supposedly dubious Species X is the same as Species Y, Species X is not dubious. It's clearly not indeterminate at the species level if you can convincingly match it to another species. Note that this does not hold for "we only know of one [tyrannosaurid, hadrosaurid, etc.] in this formation, so all of the remains must belong to one species". It only takes one find to prove you wrong, and when dealing with Mesozoic rocks, to think that we can know everything about the fauna of a particular formation is hard to swallow. Furthermore, it does not account for neighboring areas that were unfortunate enough not to leave a rock record. The odds may be pretty darn good, but I'm not going to base an argument to resurrect a name like Thespesius occidentalis or Manospondylus gigas on the grounds that there is only one known species of hadrosaur or tyrannosaur in the latest Cretaceous of western North America.
Sometimes an old species is pretty easy to ignore, which is the usual thing done with any species that is not the type species of a genus. The type species of Megalosaurus is Megalosaurus bucklandii, but of course lots of other fossils were named as species of Megalosaurus later. Because dinosaur paleontologists focus on genera, those secondary species of Megalosaurus named from random scraps have basically evaporated as the years have gone by. It would be great if these and other secondary species were all restudied, but it's unlikely we would end up with a lot of hidden gems, barring, say, some advance in the study of microscopic internal tooth structures that could allow us to distinguish teeth of different related species.
Let's say it's not a secondary species, but the type species of a genus. Here's where the three-sided tug of war really kicks in, and often situations end up with an official opinion from the International Commission on Zoological Nomenclature (ICZN). Let's start with a comparatively simple situation: you've got a venerable genus and species, used widely, but based on poor material, and there is either no competing name or the competitor has been named recently. An example of this is Coelophysis bauri, which had become attached to the Ghost Ranch theropod even though its real type material wasn't all that great. In the interest of stability, the ICZN permitted one of the Ghost Ranch skeletons to become the neotype, or new holotype. (That's the short version.) Anchisaurus polyzelus underwent something similar, although I am not all certain that the original holotype is necessarily indeterminate; it's just really convenient to have a skull. Allosaurus fragilis is currently in this process (ICZN Case 3506). A. fragilis itself is a case where usage has beaten out an older genus and species based on even cruddier material, the immortal Antrodemus valens, which is likely the same thing but only known from a partial caudal centrum.
In other cases, paleontologists have decided to preserve the genus but not the type species, and have petitioned the ICZN to make a better known species the type species. Oftentimes this new type species has served as the unofficial "face" of that genus for years. This has happened to Cetiosaurus, Diplodocus, Iguanodon, and Stegosaurus in recent years. The action wasn't unusually controversial for most of these, but the Iguanodon revision had the odd effect of making the type species of this most English of dinosaurs the Belgian Iguanodon bernissartensis. Iguanodon fell victim to being a very popular wastebasket for Cretaceous iguanodont fossils, and when the reckoning came, I. bernissartensis had been the scientific "face" beyond any of the original fossils. I do think that the "Mantel-piece" would have been appropriate, though, but there was really no way to salvage historical "Iguanodon" short of accepting that it was around for 20+ million years and most of its "species" looked suspiciously like distinct genera.
A different fate awaits the questionable old species when it has passed into obscurity and another name has become established. Tylosteus ornatus was suppressed for Pachycephalosaurus wyomingensis, although it could be argued that if Tylosteus ornatus was really based on indeterminate material, such a move was unnecessary. Better safe than sorry, though. Manospondylus gigas would almost certainly get the same treatment versus Tyrannosaurus rex if someone decided to get formal about it. (People love to toss around "nomen oblitum", but it's really hard to do if you're trying to do it properly.) Thespesius occidentalis is another candidate, being based on some scraps that anatomically could belong to any hadrosaur, but by time and place probably belong to Edmontosaurus annectens. Again, I'd argue that in the absence of anything convincing in the bones themselves, the time/place argument isn't sufficient.
It's reasonably safe to let an old name be if its time and place were shared by multiple similar animals. Deinodon horridus is based on tyrannosaurid teeth from the Judith River Group, and was a rough contemporary of Gorgosaurus libratus and species of Daspletosaurus. It probably represents one of them, but with no compelling reason to choose, it just sits there, waiting for enough people to remember that Deinodontidae predates Tyrannosauridae.
This is the boat that Troodon formosus now finds itself in. Troodon itself had become a wastebasket for North American troodontid fossils from the last 10–15 million years of the Cretaceous, and now that we can distinguish multiple troodontid species with similar teeth in that sample, Troodon formosus itself should be set aside for now. Barring new discoveries on how to separate troodontids based on their teeth, there is one thing that could be done for it, though: someone could transfer the name to another, better specimen, ideally from Judith River rocks if one is found in the near future, before the name drifts into memory. You could technically do the same thing for Deinodon horridus or some of its friends (Trachodon mirabilis, Palaeoscincus costatus, Ceratops montanus, etc.), but with most of them out of the scientific and public eye for decades, there's been very little interest in such an undertaking.
|Joseph Leidy's Judith River Class of 1856, taken from here.|