How Many Species 3: Finally, some answers.

Feb 15 2008 Published by under Biology, Science

Last Thursday, I presented some data about three populations of an insect and asked you to try and figure out how many species scientists think these populations should be grouped into. On Monday, I added data from two more populations, and asked the same thing - try and figure out how many species are present. Now, I'm going to try and answer the question myself, and tell you what other scientists have said about these insects.

A quick review is probably in order before I get to the "answers":

The five populations are arranged in a line, with each separated from the next by a minimum of 5 km. The two populations at the NW end of the line share a reproductive characteristic with the population at the SE end, while the two populations between them do something different. The SE end population looks very slightly different from the other four populations. In lab experiments, none of the populations proved to be completely reproductively isolated from the others. However, the SE most population is largely reproductively isolated from all the others, and there's at least a slight reproductive barrier separating the two NW-most populations from the next two in the line.

Personally, I don't think you could make a good case for there being four species present here (at least from the information I presented), but I do think that reasonable arguments can be made for these populations being classed as 5 species, 3 species, 2 species, or 1 species. Below the fold, I'll give you the official answer. If you haven't read the first two posts in this series yet, you might want to look at them before you look at the "official" answer.

The insects in question are members of the Hawaiian Drosophila. They're related to the little flies that generations of students have looked at in lab classes, but only distantly. The Hawaiian lineages split off from the rest of the genus quite a while back - well before any of the current high islands were formed - and have been evolving on their own since then. (If you're interested in learning more about the relationships within Drosophila, RPM did a nice post discussing evolution in this genus a few weeks ago.)

During the long course of their evolution, the Hawaiian Drosophila have split into several distinct lineages. The flies I've been talking about belong to the most well-known of these lineages: the picture-winged flies. These flies get their names because their wings have very distinctive patterns of pigmentation on them. They're also a good bit larger than most of the flies in the genus. Drosophila melanogaster, which is undoubtedly the most famous of the genus, is slightly larger than a pinhead. Most of the picture wings are about the size of a housefly.


The larger of the two flies is Drosophila differens (it's not from one of the five populations I've been talking about, but it's very similar in size). The smaller of the two is D. melanogaster, the well-known experimental animal.

Geographically, each of the five populations I've been talking about is found on a different island. Population A is found on Molokai, Population B on Maui, Population C on the Big Island, Population E on Kauai, and Population D on Oahu. Ten years ago, these populations were considered to be two species; currently, they're considered to be three.

The Big Island population has been grouped as a separate species (Drosophila pullipes) for quite a while now, based on the slight difference in appearance that separates it from the rest of the species. The remaining populations were all classified as a single species (Drosophila grimshawi) until 1999, when they were split into two species on the basis of the difference in reproductive behavior. Currently, the populations on Maui and Molokai (as well as a small population on Lanai that's indistinguishable from the Maui population) make up D. grimshawi. The populations on Kauai and Oahu are now classified as D. craddockae.

Personally, I think that the current classification is as good as any, but I really wouldn't have argued if they were grouped as a single species, two species, or five. Good reasons can be given for any of those choices. There isn't absolute reproductive isolation, so if you're using that as the standard you can clearly call this a single species. The reproductive barrier separating D. pullipes from the others is nearly complete, and it looks a bit different. That was good enough for scientists for years, and it's still reasonable. D. craddockae acts differently from D. grimshawi, and there's more of a barrier to reproduction between the two species than there is within the populations that belong to each of them. That's the basis for the current classification, and I wouldn't argue against that one, either. Finally, each of the populations lives on a separate island, there's virtually no possibility of migration between islands, and you can see hints of the formation of reproductive barriers between the populations that are currently assigned to a single species. If someone wanted to claim five species on that basis, I wouldn't argue with that, either.

It's not that I'm being wishy-washy here, or that I don't think that species are real, or that they matter. Species evolve. Isolated populations do not become new species overnight. They diverge from each other over periods of time.

In many ways, the process of becoming a new species is like the process of becoming an adult. It's often very clear when populations belong to the same species, and when they are definitely different species. It's also often clear when someone is a child, and when someone is an adult. But in both cases, there's a large grey area during the "becoming" stage. The five populations that we are talking about are adolescents, or possibly young adults. Whether you call them populations or species is going to depend on where you choose to draw an arbitrary line.

That might not be as satisfying an outcome for those of you who were looking for an answer as a clear-cut "this is the right answer" would have been, but that's just the way evolution works.


Kaneshiro KY, Kambysellis MP (1999) Description of a New Allopatric Sibling Species of Hawaiian Picture-Winged Drosophila. Pacific Science 53, 208-213

Montgomery, S.L., 1975 Comparative breeding site ecology and the adaptive radiation of Picture-Winged Drosophila (Diptera: Drosophilidae) in Hawai`i. Proceedings Hawaiian Ento,mological Society XXII(1) 65-103.

Ohta AT (1977) Evolutionary Relationships Among Six Populations of Hawaiian Drosophila in the grimshawi Complex of Species: A Multidisciplinary Approach, PhD Dissertation, Department of Genitics, University of Hawaii

6 responses so far

  • iRobot says:

    That ambiguity is one of the reasons so many people have problems with evolution. People dont do well with ambiguity. Throw in humans as just another animal that evolves and you can see peoples hair catching on fire.

  • Jim Thomerson says:

    As I understand it, each of the Hawaiian islands is an independent formation of volcanic origin, with ages decreasing more or less from south to north. If so, and there are picture wing flies of a single lineage on several or all islands, they must have arrived by dispersal, rather than by vicariance. Thus I see no problem with thinking that flies today occasionally disperse between islands from time to time.

  • Karl says:

    I have been struggling with the concept of "species" for some years now and have frequently tried to get an answer from various Sciblings. The example that blows up the concept (for me) is ring species. It seems that "is the same species as" (itssa) should be a transitive property. But, in ring species, A itssa B, B itssa C, ... Y itssa Z.
    But A and Z are clearly different species - by any definition. So it's not transitive. But where in the chain does it break down? There needs to be an even lower category than species (call it glorxel) based on something other than reproductive isolation. I propose that, now that we can map genes, we should define some percentage of difference in genes as being different glorxels.

  • Mike Dunford says:

    Jim -
    The combined effects of sea level changes, subsidence, and erosion have produced some significant changes in the Hawaiian Islands over time. Several of the current islands were actually connected in the past.
    There has always been a deep channel between Kauai and Oahu, so movement from or to Kauai does require dispersal. Similarly, there has always been a channel between Maui and the Big Island (although the channel was considerably narrower in the past), so migration is required there, as well.
    The channels that separate the three islands of the Maui Nui complex (Maui, Molokai, and Lanai) are quite shallow, and the three islands have been united several times in their past, including as recently as ~10,000 years ago. Oahu and Molokai were connected for a period of time as well, when the now-submerged Penguin Bank was an active island.
    In other words, there's a lot more possibility for vicariance than you might guess looking at things now.
    At the same time, remember that migrations from island to island have to be relatively common to keep the populations from drifting apart. Anything less than one migrant per generation, and they're going to keep on splitting. The populations are all currently restricted to areas on the islands that are above ~2,000', which means that a migrating fly has to make a trip of several miles in the right direction, survive the trip, find a mate, and mate.

  • Mike Dunford says:

    Karl -
    I don't really have the time to address your points right now in detail. I will try to get to it on Sunday, when I should have some time.

  • Jim Thomerson says:

    Just thinking about the allopatric speciation model, I have come to the conclusion that were I to observe speciation in process in the field, I would be unable to recognize it. If one sees two related populations, their possible fates range from: (1) extiction of one or both, probably the most likely, (2) a reuniting and merger of the two populations with little evidence that they have ever been separated, or (3)one or both populations thrive and speciate. How does one decide which of these events is most likely to occur? It seems difficult from having no more than a momentary snapshot of whatever is happening over a fairly long period of time from the past into the future.
    Do we have evidence for thinking speciation is in process in the present example. If so, where and why?
    Species as we think of them are endpoints of speciation processes. If many of these processes take long periods of time, and are uncertain of outcome, then we should see situations where it is difficult to clearly recognize species. If evolution is true, then this difficulty is to be expected, is it not?