Having dealt with the basics of taxonomy, we can now begin to delve deeper into the field, and there are some common issues, terms and quirks that should be brought struggling into the light when possible. In time I’ll try and deal with priority, the binomial system as a whole, higher order taxonomy, and those beastly nomen dubia, numen nuda and the rest. However here I want to deal with one of the most vexing and controversial is the near age-old issue of lumping and splitting that comes up again and again*, and causes all kinds of problems and, it should be noted, feelings often run deep.
*and again
Naming taxa, and especially species, can often be somewhat subjective. In fossils particularly it can be hard to account for individual variation (obviously not all individuals look exactly alike), possible sexual dimorphism (differences between males and females), and even ontogentics (changes in shape as a result of growth and maturity) when working with a morphological species concept. Combine this with early practices that tended to favour naming any old bits of bones as a new species and you can understand why many people think that some species have been named unnecessarily, and that these should be brought together. This is the practice of lumping – quite literally lumping several species (or other ranks of taxa, though it is obviously most common and the species and generic levels) together under a single name because there is not enough variation to justify the extra names and divisions. The opposite of this is of course splitting – pulling apart one genus or species to make two or more. But of course there is more to it than just this.
The practice of splitting and lumping (by of course ‘splitters’ and ‘lumpers’) refers not just to revisions of old taxa, but taxonomic philosophies as a whole – to name more, or less species. As I mentioned in my taxonomy post, species are arrived at more or less through consensus, (typically) if enough researchers agree it’s distinct, then it’s distinct. The obvious problem here is where there is no consensus, which can be pretty common since often there are very few researchers working on a given clade – if there are only two or three and they fundamentally disagree on which species are valid, it’s very hard to resolve to everyone’s satisfaction and difficult for outsiders not familiar with the fine details to understand.
Certainly, each approach (if take to even a mild extreme as an example) can have very serious consequences. If you want to look at diversification rates of lizards for example and the guy who works on extant lizards is a bit of a lumper and the guy working on recently extinct lizards is a bit of a splitter you might find a huge decrease in the number of species and their diversity rates, swap the researchers and you might find the opposite. Within groups of course this does not normally happen – taxonomists are familiar with each others work and approach and one can make allowances for this, but when looking across all lizards or dinosaurs or birds this gets much harder. You will not be aware of what is going on and the results of analyses of diversification, extinction, rates of evolutionary changes, and similar patterns can all be affected. This is just as much a problem in some fields of biology as well as palaeontology and macroevolutionary studies – look at conservation work.
If a species with a large overall population but several at risk is suddenly split into several what happens? A population of 10 000 animals can suddenly become one of 8 000, one of 1 500 and one of just 500. Now two of these new species are under great threat and require protection both legal and physical, or they could be gone in months or years, and in zoos and breeding programs people will be struggling to restore genetic lines that have been separated and interbred between three species. This costs time and money. This is of course fine, assuming the decision was correct and justified – if not huge amounts of resources are going to unnecessary projects and with something as resource intensive as conservation this is a serious issue.
Typically these problems are not manifest since taxonomists are rarely in such disagreements over ranks and identities (though there are some extreme practitioners for both philosophies). However, they do have real world casualties and effects (look at what happened to the tuatara) and should not be taken lightly. There is certainly a general tendency towards splitting right now with the advent of molecular and DNA analyses, barely a week goes by without some previously well defined species being split (most often with the well studied primates). Often these were well recognised as separate populations with unusual traits or just well isolated from the main groups but additional data from biochemical studies allow them to be erected as separate species. I am not in a position to judge if this is excessive splitting or perfectly normal taxonomic revisions just going on at an accelerated rate with new techniques, but as with all taxonomy it does need to be checked and revised and this does not appear to be happening much at the moment.
Dave Hone's Archosaur Musings 
As I’ve probably made clear in the past, I have a lumper mentality when it comes to, at the very least, non-avian dinosaurs. My own pale-artist rule of thumb is this: If two skeletons are so similar that the outside appearance of the animal would be the same or extremely similar*, it is a species-level distinction. If the two would look very different, it’s a genus-level distinction. We’re assuming that species, under this scheme, would be separated based on large geographical or chronological distances.
*Barring any unknown soft-tissue or coloration differences.
Certainly, there are still problems with this. Sexual dimorphism isn’t really taken into account (although sometimes it’s obviously present, like in Protoceratops), and neither is individual variation or pathology.
Or actual skeletal variation Zach. That might be your rule of thumb, but it would be impractical in the extreme when dealing with a skeleton. Take Aereosteon for example, externally it could look identical to another species (or even a member of another family quite possibly) but the derived pneumatic system would mark it out as very different. Most palaeo taxonomic characters *are* internal ones – pleurocoels, hypantra articulations, shape of prezygaphophsys, vertebral laminae, the palate and braincase, the shape of the cnemial crest or fourth trochanter, ungual shapes, number of vertebrae, the atlas – axis articulation and more would not necessarily (and in most cases would not) show any external morphological changes, yet you could probably find 50 or even 100 characters to separate them out. In the opposite case,m you could have two animals that are completely identical, but one had an extra horn and thus an obvious external distinction that would now make it a new genus. It just doesn’t work. You have to evaluate all the evidence and make a decision based on this, not some arbitrary distinction that only works when you have most of a skeleton available.
Ah, good point. I didn’t think about internal aspects of the skeleton. Has anybody done a comparison of living, related animals (like, say, crocodilians) and a well-known group of dinosaurs (tyrannosaurs, etc.) to see whether minor skeletal variations in extinct taxa is or is not mirrored in living taxa, and what that could mean for basing extinct taxonomies on minor skeletal variations is sound?
Another way to say the same thing: Let’s say that two populations of nuthatch can be distinguished based on the shape of the tibia–maybe one population has a prominent crest on the tibia, and the other population does not. However, both populations can breed and produce viable offspring.
Now, we’d call those the same species. What does that kind of intraspecies variation between clearly related fossil taxa, like Allosaurus and Saurophalanx? I suppose this kind of question gets us back to square one, though, and brings us back to how to define a species.
Yes, its basically all interleaved at it depends what methods you use, what critera you have and what the ‘consensus’ is in the clade being studied. We have to use a morphological species concept in the fossil record (rightly or wrongly) and our species are based on that (though with most vertebrates, we effectivly use genus as the most basic unit really rather than species per se). Bird workers tend to use songs, plumage and distribution as these are easy to study despite the fact that there are plenty of molecular and anatomical differences (thoiugh less of the latter than you might expect). I’ve been doing a lot of reading on thsi recently, and yes, it all just comes down to common sense and practical application – life is a continuum, species are artifical, but we have to make the two meet somehwere or we’ll never get anyhting done.