I thought a short essay on the treatment of holotypes would be worthwhile. In zoological sciences, many of us are involved in describing and naming new species - in order to maintain taxonomic stability, a type specimen must be designated in a new publication naming a new species. In plain english, the type specimen - also known as a holotype - is the specimen demonstrating the physical evidence for which a new species is named upon. Often when a new species is discovered, researchers will leap to name it - so often, we're stuck with a species named upon a single specimen. It happens frequently in paleontology; finding good fossils is (often) more difficult and more dependent upon sheer luck than going out and finding a second specimen of a newly discovered modern species. As a result, much of the vertebrate fossil record is composed of "singletons" - fossil species represented by single specimens. This problematic nature of the vertebrate fossil record is a favored talking point of young earth creationists, and has also unfortunately contributed to the bizarre opinion amongst some radical cladists that ancestor-descendant relationships are impossible to determine in the fossil record. Radical ideas by folks such as Nature editor Henry Gee come to mind, who rejects all non-cladistic methods in paleontology as being unscientific - an easy position to take if used to the fragmentary vertebrate record of singletons, but difficult to maintain in the face of the enormous and densely sampled record of fossil invertebrates (mollusks and foraminifera come to mind, in which non-cladistic methods can identify speciation events and bouts of anagenesis in vertical successions of fossil assemblages).
The typically fragmentary nature of the vertebrate fossil record, and dealing primarily with singletons, can lead one astray. Subdisciplines divorced from extant relatives and flooded with researchers who do not really look at modern species to get an idea of morphological variation - dinosaur paleontology comes to mind - may be dominated by an overabundance of taxonomic "splitters" (disclaimer: I have many friends who are dinosaur paleontologists and have a rather sober idea of morphological variation). For the uninitiated, splitters are researchers who tend to name more species than is probably likely; the opposite are "lumpers", who tend to lump species together into a more reasonable framework of fewer names. Many in biology prefer to use a third but unnamed category for the middle ground - I do not, as I consider the "lumper" category the middle ground, and "extreme lumpers" to be on the other end. One marine mammal related example of an extreme lumper is work by Caretto (1970) who lumped all known fossil balaenopterids into a single subspecies, including fossils now known to likely represent several genera. Most "lumpers" I know tend to use morphological variation in related extant species as a tool for interpreting species-level identifications/naming in the fossil record, which sounds reasonable ("eminently sensible" as my adviser would say) - and so I consider anyone reasonably taking a middle ground approach to fossil identification/naming a lumper (versus an extreme lumper).
Many smaller subdisciplines dominated by just a handful of researchers working more or less "unopposed" (i.e. without professional rivals to rigorously dispute work) may be plagued by splitting until some young turk who's spent more time playing with modern skulls comes along and rewrites everything - this has happened time and again in mammalian paleontology. I won't use any examples close to home as many of those researchers involved are still extant, so this discussion is largely going to consist of hypotheticals.
If you spend too much time looking at a relatively small group of fossils - and don't bother looking at modern specimens of extant relatives to take a stab at what sort of range of morphological variation you ought to expect - it's easy to fall into the trap of "typology". I don't necessarily mean the reliance upon type specimens (although I'll eventually get back to that), and the definition of typology is generally given as "a classification according to general type, especially in archaeology, psychology, or the social sciences." In paleontology, typology is a style of thinking in which every fossil species conforms to a different morphological/anatomical "type" (NOT to be confused with the similarly vaguely defined "kind" from young earth creationism). In other words, it should be relatively easy to identify two fossils (or, alternatively, living individuals) of the same species because according to typological thinking, they should look close to identical. For over a century, this was the founding principle of paleontology in practice: if you find something that looks a little different, it's a new species. Done. That's partially why folks like Cope, Marsh, and Leidy published so many damn papers.
The problem is, and this is a concept well known to most paleontologists - biological entities don't give a shit what you think about them. Domesticated animals and humans are some of the more extreme examples - all breeds of domesticated dogs belong to Canis familiaris and can healthily interbreed, but if the entire species was extinct and all we had were skeletons (no ancient DNA), paleontologists would readily name the skulls of a doberman and a chihuahua as different species. With humans, it's a similar case: obviously we know all extant humans can interbreed, but there is a fair amount of morphological diversity, and the human mind is hard wired to recognize familiar faces and as a result exaggerate differences between different populations of humans. Someone who is familiar with other primates would readily note that all humans share much more in common with one another relative to gorillas and chimpanzees, but someone who only looked at human faces and skulls would tend to focus more on the differences rather than the similarities.
Splitting versus lumping continues to be one of the biggest challenges in paleontology - is the new species you've named real, or is it just a morphological extreme? Or is it the opposite gender in a species that is sexually dimorphic? A recent paper just proposed that the Flores hominin - the "Hobbit", Homo floresiensis - is actually just a Homo sapiens with Down's Syndrome. As more and more paleontologists leave dark dusty fossil collections and dabble in neontology, variation is becoming more and more understood and embraced. In the 19th century, vertebrate zoology and paleontology was figuratively the Wild West (and in the case of Marsh and Cope, quite literally). New species which would later turn out to be junior synonyms were published by the bucketload; reading through Victorian era zoological papers often reeks of taxonomic anarchy, with the proliferation of all sorts of species and genus names no longer in use. In many cases, naturalists genuinely were unaware of some papers and discoveries - some species of modern whales and dolphins were independently named and "discovered" literally dozens of times (from different strandings on different continents) during the late 19th century - that's just one of the many problems facing naturalists in the pre-internet world. Some of the synonymy lists for modern dolphins are astoundingly long.
For those unfamiliar with taxonomy but have continued to read this anyway, one of the most important rules set out by the International Code of Zoological Nomenclature (ICZN) is the rule of priority: the species name published first has taxonomic priority. If you name a second species that turns out to be based on an individual from an already-named species, the earlier published name is used. This also applies to genus names - and can spark "nomenclatural wars" like the Aetogate controversy a few years back. If two researchers investigating the same fossil species, and 1) it is an already described species placed in the wrong genus and a new genus name is needed or 2) both parties have found separate fossils of the same, as yet unnamed species - whoever publishes the new name first "wins", so to speak - all other considerations (ethical or otherwise) are irrelevant according to the ICZN. This can both cause extreme tensions in an otherwise small and close knit discipline, or result in the proliferation of synonymous names published in disappointingly brief and poorly figured papers (sometimes called name grabbing, and is the paleontological equivalent of "saving a seat" for yourself).
Now that the discussion of zoological nomenclature has come full circle, it's time to address the title. When in the business of "splitting" and naming new species - particularly when dealing with a record dominated by singletons - it's sort of a no brainer to focus on differences between holotype specimens. There are other types of types - there are paratypes, if you want to clearly designate some key specimens other than a holotype, there are syntypes or cotypes, where a type series is designated but no holotype is selected. There's also lectotypes, where someone selects a single syntype as a retroactive holotype-like designation, and there's also the concept of a neotype - if a type specimen is lost, a neotype conforming enough to the anatomy of the original may be designated afterwards.
But what happens if you have a fossil record that's... better? Admittedly, many parts of the vertebrate fossil record are still awful - but many subsets of the record are pretty damn excellent. What happens when you find a new fossil you think is the same species, but better preserved than the holotype? Paleontologists call these referred specimens - any specimen possessing some evidence that it represents the same species can be considered a referred specimen (e.g. you refer the specimen to a particular taxon). Referred specimens are of immense importance because they can provide 1) information on the anatomy of parts not preserved in the holotype and provide 2) information on anatomical variation for parts that overlap with the holotype. We already know that modern species display an impressive amount of variation, and paleontology is science after all, so in theory we should be using as much evidence as possible to improve our understanding of extinct species in the fossil record.
Some researchers, however, prefer to focus only on holotypes. You can easily lead yourself into a trap where you assume that the holotype specimen either 1) faithfully represents the morphological "type" of a species or 2) faithfully represents the morphological midpoint of a particular species (and is thus not an outlier in terms of its morphology). Both of these are actually pretty terrible assumptions to make, because in the absence of additional specimens, there's no objective way to defend (or even evaluate) either assumption. I've been told to "stick to holotypes" before, and have heard some Researcher A casually dismiss the careful work of Researcher B who referred additional specimens to a species originally named by Researcher A - and the reasoning was that they were dealing with non-types! I've even had some curators and collections managers give me puzzled looks when I asked to see some obscure referred specimen that often gets overlooked by visiting researchers in favor of the (often more incomplete) holotype.
What exactly are type specimens for? What is their purpose? According to the ICZN, type specimens are nothing more than a physical record of a species to which a zoological name may be permanently attached. In fact, the use of many historical names through to the modern day demonstrates that holotypes need not even exist - the sperm whale and narwhal do not have type specimens. Extreme cases like this are "grandfathered" in, so to speak, and this practice is no longer acceptable (nor even recommended). Type specimens can even be a plaster cast of a fossil specimen that does not even exist any more, or a single feather plucked from a newly discovered and named but extraordinarily rare "living type" specimen still flying around in the jungles of Borneo. The holotype specimen of the archaeocete Zygorhiza kochii is a relatively useless and miserable shitty braincase without any redeeming features - nobody uses that specimen, and mostly everybody in our field examines USNM 11962, the "de facto reference specimen" which Remington Kellogg monographed in the 1930's.
If the utility of type specimens has no real scientific basis other than mere housekeeping, then there's no real reason we should exclude non-type specimens. I call this weird practice "holotype worship", and I think it moreso has to relate with stamp collecting rather than actual science. Holotype worship - and typological thinking, I believe, are responsible for resistance to many proposed instances of synonymy - the most famous ones that come to mind are the hypotheses that the horned dinosaur Torosaurus is simply an old adult Triceratops, advanced by my close friend John Scannella at Museum of the Rockies, that Nanotyrannus is a juvenile Tyrannosaurus (advanced by Jack Horner and others), and that the dome-headed dinosaurs Dracorex and Stygimoloch are juvenile Pachycephalosaurus, advanced by Mark Goodwin and Jack Horner. Each of these hypotheses incorporates information past holotypes and "de facto reference specimens", uses a larger body of data, and from additional sources - and as a result, arguably explains the fossil record in a superior manner to hypotheses based on fewer data. As an aside, many young folks I've met who are rabidly against John's Torosaurus-Triceratops synonymy hypothesis come from the fanboy part of the spectrum without a strong educational background in science (other than "dinosaurs are cool, ankylosaurus is my favorite"). Other cases of synonymy more relevant to the subject of this blog are the early pinnipedimorphs Pteronarctos goedertae (1989), Pteronarctos piersoni (1990), and Pacificotaria hadromma - which Annalisa Berta argued in 1994 were a single species, a conclusion I tend to follow.
George Gaylord Simpson challenged the focus on holotypes back in 1940, and even then it was apparent that holotype worship was an antiquated form of research from the prior century. Simpson proposed the concept of the Hypodigm - the complete body of all known material assignable to the species under consideration - in other words, the type and all referred specimens. Many researchers have embraced this 75-year old concept, but in some subdisciplines within paleontology there has been surprising resistance. Many older paleontologists are reluctant to let young turks synonymize some of the species they originally named in the process of cementing a long career. Another pet peeve of mine - and this is particularly relevant to large reviews of various fossil groups - but I've read many papers that flat out only discussed the type species of a particular genus, and ignored those described as additional species within formerly monotypic genera (e.g. a genus with only one species). All fossil species are important! Don't ignore some species just because they were named later (but ignore them if they are synonyms, sure), or worse - don't ignore them because they were named by somebody you don't necessarily like.
Every month I see new papers bringing paleontology into the mid-20th century (and occasionally, late 20th and even 21st centuries) in terms of methods and scope, which gives me hope. Marine mammal paleontology is advancing faster than ever before, largely due to an influx of young researchers attacking new problems with a set of new ideas and tools. Although nearly a century old now, Simpson's idea of the Hypodigm is still as relevant as ever yet hasn't caught on as much as it should. Do your part to help stop "holotype worship" - and use as many specimens as are available!
Berta, A. 1994. New Specimens of the Pinnipediform Pteronarctos from the Miocene of Oregon. Smithsonian Contributions to Paleobiology 78:1-30.
Caretto, P. G. 1970. La balenottera delle sabbie plioceniche di Valmontasca (Vigliano d’Asti). Boll. Soc.
Paleontol. 9: 3–75.
Simpson, G.G. 1940. Types in modern taxonomy. American Journal of Science. 238: 413-431.