|Microtus californicus californicus|
But a species is meant to be an actual thing, right? Certainly, when the word was coined by Linnaeus (in the biological sense; it has older meanings, too) in the 18th century, that was his intent. A hundred years later, Darwin elaborated on the term, realising, as earlier evolutionary theorists had, but Linnaeus hadn't, that species do slowly change into new forms over time. Today, the most common understanding, outside of scientific circles, is that two animals belong to different species if they can't interbreed to produce fertile offspring.
There are a whole host of reasons as to why this doesn't really work, not least of which is the question of how you would actually know that two animals can't (rather than won't) interbreed. You can't test all of them, even if they were willing to cooperate, which is far from assured. But, nonetheless, there is the feeling that, at some level, species must really exist; there has to be something that's a "kind" of animal, even if we can't quite put our finger on its definition.
Defining subspecies, though, is rather more of a problem.
Of course, it was obvious to Linnaeus that not all members of his "species" looked identical. Clearly, there was variation going on, but, so far as he was concerned, it wasn't really all that important in the grand scheme of things. For the remainder of the 18th century, it's probably fair to say that most naturalists agreed with him, naming something as a new species if they thought it was distinct enough to warrant it, and otherwise deeming the difference too insignificant to mention. This changed with time, as more and more specimens were collected - the westward expansion of the USA across North America has been cited as a major driver behind this.
Confusion and disagreement in the early 19th century gave way to attempts to do something about it in the decades that followed. Multiple different schemes arose for how groups smaller than species should be named, what they really were, and even how many levels of them there should be. In the 1850s, for example, Darwin referred to "varieties" which represented, to him, the smaller groups that could, in time, evolve into full species.
The current scheme was pretty much solidified by the 1880s, although it took a little longer for it become the de facto standard, and to be eventually be codified by the International Commission on Zoological Nomenclature. Under these rules, a species either has no subspecies at all, or it has at least two, one of which, the "nominate subspecies", is whichever was the first to be described, and the others being later variations on that theme. The only way a species can have just one subspecies is if all the others are extinct; they exist, and have names, but don't happen to be around any longer. Furthermore, in zoology, although not botany, there are no smaller divisions than "subspecies" that receive a formal name.
But this doesn't really get us any closer to what subspecies actually are.
In the late 1930s, Ernst Mayr developed the "biological species concept", with, at its heart, the familiar concept of a species as a population of individuals that can interbreed to produce fertile offspring. He was aware of the limitations of the idea, and did go as far as to specify that the animals had to interbreed in the wild, and not researchers had made them do so in a zoo. But even so, it's largely from Mayr that we get this idea, and its corollary, that a "subspecies" must be a distinct population that can - but usually doesn't - interbreed in this way.
Taken together with work at around the same time by Theodosius Dobzhansky, the basic idea here is that subspecies are prevented from interbreeding by something other than pure biology, and that, in most cases, that's likely to be geography. The two populations, while they absolutely could and would interbreed in the wild if only they got the chance, instead live far enough apart that they don't get the opportunity. With time, random genetic change will accumulate, and, if nothing happens to prevent it, they may eventually develop into two distinct species.
And, despite refinements to the precise definition of a species over the intervening years, the Mayr-Dobzhansky definition is still broadly where we are today. What this means is that, in most cases, subspecies must inhabit different geographical locations. After all, if they lived in the same place, there would be nothing to stop them interbreeding, and they'd merge back into one. Or, if they didn't, then they're likely already different species, or at least getting close to becoming such.
This does imply that geographic isolation - and it need only be a river that the animals can't cross - is the only way to create new species, which we know it isn't. There are at least some instances, for example, of populations feeding on different food sources within the same area so that their members rarely encounter one another, and become reproductively isolated that way. Logically, one would expect there to be a stage when such animals belong to different subspecies, but are not yet fully isolated. But they're hard to identify, so generally we stick with geography as our yardstick.
This, however, raises a couple of questions. Firstly, there's how isolated the subspecies have to be. If we look at American black bears (Ursus americanus) for example, sixteen subspecies are typically listed. Some of these live on islands, with Vancouver Island and Newfoundland (among others) both being home to their own distinctive subspecies. But many are found in regions across the mainland, with different subspecies in, for example, Louisiana/Mississippi and Florida/Alabama, and multiple ones across Alaska.
Clearly, these latter subspecies do meet up somewhere around their respective borders, and, by definition, they interbreed when they do so. This shows that it's not enough for them to be different; the change from one form to the other has to be relatively rapid across the zone where they meet. If a particular anatomic feature changes gradually, from, say, east to west, that isn't evidence of subspecies, because there's no obvious line between the populations. It's one population that changes across its area - technically called a "cline".
The second question is related: just how different do they have to be? And here, we fall entirely into subjective opinion, because there is no clear answer that everyone agrees on, even within a relatively well-defined group of species. And good luck getting mammalogists and entomologists, say, to agree on any kind of mutually meaningful definition.
These days, molecular and genetic information is used to help delineate subspecies in much the same way that it helps identify or confirm the existence of species. After all, one of the few ways that we can determine whether a population interbreeds in the wild is to to examine the genes from different individuals across that population, and see whether or not there is any evidence of mixing. In many cases, however, once we have found an important genetic difference, we often find that once we have re-examined the relevant specimens, there are subtle physical differences that we hadn't previously noticed, or paid attention to.
A recent example of this concerns the California vole (Microtus californicus), an animal that looks pretty much like every other vole you've ever seen. A total of seventeen subspecies are currently recognised across an area stretching from southern Oregon to northern Baja California, and just about reaching the Nevada border in the east. Last year, a study of populations living in the south of California identified two distinct lineages of their mitochondrial DNA.
These particular voles are not fond of arid conditions, which might make you wonder what the heck they're doing in the deserts of southern California. But the genetic study provided an answer to this, by comparing the DNA of the desert-dwelling voles with those living closer to the coast and comparing their distribution with maps of the area. By doing so, the researchers were able to show that there are two lineages because, at some point in the distant past, two different groups of California voles had entered the desert by travelling along river valleys, one from the north, and one from further south. They are now found sheltering in the few damp parts of the region, near lakes and deep valleys.
But now, a newer study has repeated this work, using physical features of the voles, instead of genetics. They found that they could correlate not only the appearance of the voles with the new genetic data, but with the five subspecies said to live in the area, all of which had been sitting around in reference books since at least 1918 without anyone thoroughly checking their validity.
I mention this because it helps explain why identifying subspecies can matter. The California vole is a reasonably widespread and common species. Where the land actually is as wet as they like to be, there can be hundreds of them per hectare, and they're clearly not going anywhere. But, in the desert, it's a different story.
Three of the desert subspecies seem to be faring reasonably well, all things considered. But one, M. c. mohavensis, is known only from two small locations on the Mojave River where the local geology permits water to gather year-round. The other, M. c. scirpensis, is found, so far as we know, only from the hot springs of Tecopa on the Amargosa River near Death Valley. (Take a look at the satellite image to see how much greenery we're talking about here).
Now, it's true that getting conservation protection for a subspecies is much less easy than for a full species; by definition, it can't be an "endangered species". But if it's not even a subspecies, you have essentially no hope at all. It has no formal name, no official identity, it's very existence as a discrete thing is likely to be in question.
The word "subspecies" may not be very well defined. But at least it gives you something you can point to.
[Photo by Jerry Kirkhart, from Wikimedia Commons.]