In my presidential address to the Society for the Study of Evolution last night, I started off, as I had to, by defining what I meant by speciation (my talk was, after all, about speciation). I said that I was using the “biological species concept” (BSC), which conceptualizes species as populations separated by reproductive isolating barriers: genetically based features, like mate discrimination and hybrid sterility, that prevent gene flow between different groups. And when I put up a slide of the BSC, I said it was the “right” species definition.
That was meant to be both funny and provocative: I know that different groups of biologists have different species concepts, and in some cases they should—you can’t use the BSC in bacteria, for example, which don’t really have sexual reproduction. But in my book Speciation with Allen Orr we note that there’s no one “right” species concept: concepts are more or less useful depending on the biological problem you’re addressing.
Since my problem is addressing the origin of distinct groups that coexist in one place without much interbreeding, the BSC is the only one that makes sense. Indeed, regardless of the species concept proposed by biologists—and there are many—when biologists go about studying the process of speciation in sexually reproducing groups, all of them, nearly without exception, adhere to the BSC. That is, when they study speciation they invariably study the origin of barriers to gene flow. So if the origin and existence of discrete groups of sexual taxa nature is the problem—why there are blackbirds and robins and crows and sparrows, all living distinctly and discretely in one place—then the BSC is the go-to concept.
I used the word “right,” of course, to make people laugh, to provoke them (after all, the meeting includes a lot of systematists, many of whom abjure the BSC), and because it’s well known that I’m a staunch advocate of using the BSC for the “species problem” described above and have been a vociferous critic of other species concepts.
So when Carl Zimmer made this (ugh) “tweet” on the evol#11 Twitter feed, he was sort of on the mark:
I wasn’t purely joking but I was partly joking. And Carl, my good man, if you want to know what I meant, all you have to do is ask.
34 thoughts on “The “right” species concept”
When you’re right, you’re right.
And when you’re useful, you’re useful.
I’m a little over half way through “Speciation”. I’ve learned a lot.
“Tweet”, “blog”…Jerry, ya gots ta get with the lingo — all the cool kids are sayin’ it.
What’s needed is a combined platform.
From which one could “bleet”.
It seems that the BSC is ‘right’ because it cuts through the arbitrariness of how much sequence or morphological divergence warrants distinct species status. Are these two sets of populations now on separate evolutionary trajectories, or not (regardless of their current level of divergence)? Of course, as Darwin noted, there will always be no shortage of difficult applications (beyond the problem with bacteria, paleospecies, etc., because one will encounter cases of incipient speciation.
I believe you’re conflating the BSC and the evolutionary species concept.
Regarding the BSC–how do we avoid arbitrariness by adopting the BSC? How much reproductive isolation is enough?
Regarding the ESC–evolutionary trajectories, so far as I can tell, are fundamentally non-empirical. Your statement that whether sets of populations are on separate evolutionary trajectories is independent of their current level of divergence indicates the problem nicely. What, then, does indicate that separate evolutionary trajectories exist? I haven’t seen a coherent answer to this in the ESC literature.
I probably was unclear, but I was referring to separate evolutionary trajectories simply in the context of current levels of gene flow. Zero gene flow means a separate trajectory regardless of current differences. At the other extreme, abundant gene flow indicates conspecific populations. Your point about how much reproductive isolation is enough is what I was trying to get at by referring to incipient speciation. In such cases, you might see a set of populations where A interbreed freely with B, B does with C, but A and C don’t seem to recognize each other pre-zygotically. So the BSC does not eliminate messiness but, by focusing on reproductive isolation, it gets at whether populations tend to be evolving separately or in concert.
“Zero gene flow means a separate trajectory regardless of current differences.”
But how do we infer an absence of gene flow? By an absence of differences, of course!
It seems like you’re trying to separate a property inferred from data and the data used to infer that property. I’m not sure what this accomplishes.
“Your point about how much reproductive isolation is enough is what I was trying to get at by referring to incipient speciation. In such cases, you might see a set of populations where A interbreed freely with B, B does with C, but A and C don’t seem to recognize each other pre-zygotically.”
However, we don’t even need to invoke ring species to get messiness under the BSC. Suppose we have two sets of populations, A and B. We infer gene flow between the two to be low but greater than 0. Where’s the cutoff?
That the BSC focuses on reproductive isolation, IMO, does nothing to reduce messiness. If anything, it increases the problem by focusing on a property that must be inferred indirectly, rather than focusing directly on observed patterns of data.
No, we can infer lack of gene flow in several ways, including complete sterility or sexual isolation between the species (this can also be measured in the lab) or by looking at the genes themselves and seeing if there is any introgression, which is something we’ve done in our two island species. And the “messiness” of the BSC is wauy overrated; for most sexually reproducing species, and that includes plants, there is no evidence of hybridization whatsoever between close relatives. In birds, 10% of species have been observed to form hybrids, but many of those hybrids are either intrinsically or behaviorally sterile, so cannot be conduits for gene flow.
Many of the questions raised in this thread can be answered by reading chapter 1 of Speciation, by Coyne and Orr, where we address incomplete reproductive isolation, messiness, and other questions. You don’t have to buy the book; just take it out of the library.
Consider it done, kind sir!
Crap, I screwed up that first couple of sentences. It should read: “But how do we infer an absence of gene flow? By the presence of fixed differences, of course!” Or something along those lines.
Having read Jerry’s book and having spoken to him for my reporting on speciation, I felt pretty sure I got the joke. And by “Nope,” I meant exactly what Jerry says here–he was not *purely* joking. But thanks for the expansion. I will direct people to it for their own enlightenment–with a tweet (gasp).
I once heard that species is the only meaningful or quantifiable classification level that can be measured genetically. Does this have anything to do with BSC? Which btw is the only concept of species I”m familiar with, and therefore the only one that makes any sense to me! lul
Re: “All you have to do is ask.”
I’ve always wondered why some people are afraid to ask a direct question directly of the source.
To me, it’s the only way to gain clarity.
But I think a lot of people in general, not necessarily limited to science types, seek to avoid direct confrontation. Lest they be seen as being “dickish”.
Agreed. If I have already clarified an email I sent and get another request, or if I don’t understand someone else’s, I pick up the phone and talk to them. Two minutes on the phone will create a better understanding than 10 more emails.
I view the BSC as simply a statement that, in sexual lineages, morphological and genetic distinctions between groups of individuals are accompanied by rarity or absence of interbreeding between those groups. I don’t think there’s any basic disagreement on that point (although there are practical complications that tend to get swept under the rug), but it’s not clear to me what it accomplishes as a definition of “species”.
I observe some pattern of data, and the BSC tells me that sex is a key process in the origin of that pattern. Well, OK, but as a practicing taxonomist, how does this help me? I need to use that pattern to infer species boundaries. Knowing that it results from some set of reproductive relationships is not particularly helpful for this goal.
I’d argue that a definition for “species” that is inapplicable to most of the lifeforms on the planet isn’t terribly good, or at least isn’t terribly general. Are we supposed to toss out the notion of “species” for non-sexually reproducing organisms? If they have “species”, can we even say that they are at all conceptually similar to the “species” that the BSC delineates in sexual lineages?
“Species” seems like porn — people can’t give a coherent definition of it, but know it when they see it.
I agree entirely that the BSC lacks generality. Further, I don’t consider it particularly useful from a practical point of view within the realm in which it -is- applicable.
IMO, the BSC looks like a good idea for those who are: 1) studying sexual lineages; 2) not dealing directly with the inference of species boundaries (i.e., non-taxonomists). In other contexts, it has limitations that decrease or eliminate its applicability and usability.
It would appear that Carl knew full well.
Did you mean @CarnyEvolution instead?
I considered two forms which might have been ecophenotypes of a single species or two separate species, fish or fishes. At the time they were not know to occur in syntopic association, although sympatric. I was able to reproduce both of them in the laboratory, but they would have nothing to do with each other under the same conditions. I was thinking they were different species based on a single difference in fin ray counts, and felt much better about describing one of them as new after my failed crossing experiments. They are sister species according to our later DNA analysis, but have quite different karyotypes, and different breeding behavior in the wild. We later found them syntopic in a disturbed area and saw no evidence of hybridization over several generations. BSC worked pretty good in this instance.
I have often wondered why the canines called Great Danes & Chihuahuas are not considered different species, as the interbreeding barrier is naturally “insurmountable”.
If all the sizes in between, the beagles, airdales and plain mutts, were to die off they almost certainly would be. As it is, variation (incl. fertility) is too close to continuous between the extremes to separate them. You could cross the dane with a shepherd, and the offspring of that cross with a beagle, and those pups with a dachshund and …
There are similar messy cases in nature.
I have long thought that extinction is one of the most important forces in speciation — sometimes the linking populations just get killed off.
More problematically, you can also cross a shepherd with either a coyote or a wolf. Or you can cross a coyote with a wolf; a dog with a jackal; a wolf with a jackal.
What this means under the BSC depends on whether we make inferences from ability to interbreed alone (in which case they should all be the same species) or focus on patterns of morphological and genetic distinction in natural populations regardless of ability to interbreed (in which case they should be different species). In the latter case, we would reach the same conclusion with other functional species concepts. The former interpretation is the BSC’s unique contribution to the problem.
Ah yes, how stupid of me!
Such as those lizards of (I think) the west coast of the USA.
And, if I remember correctly some northern sea-birds.
Oh, and I cannot refrain from observing:
This has been successfully acheived in Scandinavia many times going back before recorded history.
(Only not with canids.)
Great Danes and Chihuahuas, if they have the right friends, can share genes, given enough time. They are two ends, so to speak, of a cline.
Maybe organisms which do not do biparental sexual reproduction are more or less honorary species just to avoid coining another term. On the other hand, self fertilizing hermaphrodites, all female species, and species which only reproduce vegetatively all fit the reproductive isolation from other forms criteria
I understand, and thank thee.
So, to paraphrase the marvellous Ben Goldacre:
I think I’m finding it’s a bit more complicated than than I first thought!
” I know that different groups of biologists have different species concepts, and in some cases they should—you can’t use the BSC in bacteria, for example, which don’t really have sexual reproduction.”
Ha. Bacteria figured out this genetic mixing bid’ness quite awhile ago. Instead of all those interesting, complicated methods of exchanging/ melding genetic code that “higher” creatures use, they just crap it out for all to enjoy and suck it in to their heart’s content (so to speak).
In other words the problem with the “species” concept is that we are trying to push a discrete classification onto variables that are intrinsically continuous, and sometimes there is no clear breakpoint that we can use to define the classification.
I think that terms such as “intrinsically continuous” are an approximation of atomic reality.
At the very least level of granularity, codons are discrete units. The genetic continuum stops there.
The lowest level of classification is inherently discrete.