If you look at the answers that people have given, you will see that all three possible choices (1 species, 2 species, and 3 species) have received some votes. The most popular answer is that there are 2 species, with populations A and B being put together as a single species, and population C being given status as a separate species. The people who have chosen this option focused on the obvious differences in fertility for the crosses involving population C. The person who voted for three species did so based on the high likelihood that all three populations are on separate evolutionary tracks. The people who voted for a single species did so based on the fact that, despite the male sterility, population C is still interfertile with populations A and B. Several people also asked for more information. I'll try to satisfy some of those requests in this post.
All three arguments are good ones, and I can't say that any of them is flat out wrong. I can, and will, tell you which of those positions matches the current scientific classification for the group, but that does not mean that the other two are necessarily wrong. This group definitely falls into a grey area. The three populations are somewhere in the process of differentiating from each other, and drawing the line becomes a bit tricky in these cases.
Populations A and B are currently considered to be in the same species - Drosophila grimshawi. Population A is the Molokai population; population B is found on Maui. There is also a population of these flies on Lanai. Population C is considered to be a separate species, D. pullipes, and is endemic to the Big Island.
The data that I presented on the fertility of male hybrid offspring was taken from Alan Ohta's 1977 Dissertation. A published version of the material can be found in:
Ohta, Alan T. 1980. Coadaptive Gene Complexes in Incipient Species of Hawaiian Drosophila. The American Naturalist. v. 115(1), pp.121-131.Here's where it gets fun, though: I only gave you part of the situation with this group in the example, and the part that I did talk about is actually the more clear-cut part.
In addition to the populations that I talked about in the earlier post, there are also populations on Lanai, Oahu, and Kauai. The Lanai population acts much like the Molokai and Kauai populations. This isn't surprising, since the channels connecting Maui, Molokai, and Lanai are narrow and shallow - so shallow, in fact, that the three islands were connected several times during the ice age. As a result, it is common for evolutionary biologists to treat those islands as a single entity, often referred to as either the Maui Complex of islands, or "Maui Nui" ("Big Maui"). The Oahu and Kauai populations do not fit in quite as well with either the Maui Nui populations or with the Big Island population.
I'm not going to present numbers this time the way I did last time (see the reference cited above if you're interested), but I am going to try to provide more of a broad overview. Some of this was already covered in the earlier post, of course, but I'm going to include it to try to keep everything in context.
Pretty much all of the extant Hawaiian Drosophila are restricted to the native forests above 1000'. This is probably related in part to human-caused habitat changes, but in the case of the picture-wings (including the species discussed here) the restriction is more climate-related. These species prefer cooler temperatures than other Drosophila species, and tend to do poorly at temperatures over about 68 F (20 C).
The Big Island flies (D. pullipes) can be distinguished from the others only on the basis of some minor color changes in the legs and sides of the thorax. All of the other populations are physically indistinguishable from each other.
This is the most obvious phenotypic difference separating populations. The Oahu and Kauai populations and D. pullipes will only deposit eggs in the presence of rotting bark from plants of the genus Wickstroemia. The Maui Nui populations will ovideposit virtually anywhere. Larvae have been reared from 12 plant families, including 2 that are entirely invasive to the islands (the figure of 14 that I gave in the comments was incorrect). To answer a question asked earlier, the Maui Nui populations have been known to use Wickstroemia, but I think it's unclear whether that is a retained trait or whether they're simply indiscriminate. Looking at the study (Montgomery, 1975) that determined this, it appears that their larvae has been reared out of pretty much any plant found in the same climate/vegetation zone as the flies. It should also be noted that specialized ovipositional behavior is the norm in the picture-wings. Only four other species are known to use more than three families, and only one is known to use more than grimshawi.
All of the populations can be crossed successfully in the lab. As I previously reported, the Maui Nui populations are fully interfertile with each other, but most of the crosses involving D. pullipes produce sterile male hybrids and decreased numbers of fertile female hybrids. (For the exact numbers, see Ohta, 1980). The Kauai and Oahu populations are interfertile with each other, and produce infertile males when crossed with D. pullipes.
Here's where it really starts to get interesting, though: when crossed with the Maui Nui populations, the Oahu and Kauai populations produce fertile hybrids, but the fertility of the F2s (the offspring of hybrid x hybrid matings) is decreased, as are most of the backcrosses (hybrid x parental species). This is particularly true for crosses involving a hybrid and the popuation used for the male parent of the hybrid, but much less so when the backcross is with the female parental population. Again, the numbers can be found in Ohta, 1980, but here are the high points: the smallest drop in the F2 fertility was from 90% to 50%. The smallest drop for backcrosses into the male parental population was from 93% to 63%. This indicates that while the differentiation between these populations isn't as great as that involving D. pullipes, it's still substantial.
In his dissertation, Ohta examined about a dozen allozyme loci (those are enzymes where more than one version [allele] of the enzyme is present in the group you are looking at). He computed Nei's genetic distances for the different populations.
Here's a quick summary:
Within the Maui Nui populations, the largest distance separates the Lanai and Molokai populations (0.246), and the smallest separates Molokai and Maui (0.05).
The Hawaii population (D. pullipes) is most similar to the Molokai population (D=0.245) and most distant from the Kauai population (D=0.419).
The genetic distance between Kauai and Oahu, despite their behavioral and morphological similarity, is larger than any of the distances involving D. pullipes (0.375).
The greatest genetic distances are those separating the Oahu/Kauai populations from the Maui Nui populations - the range is from 0.576 (Kauai x Lanai) to 0.760 (Kauai x Maui).
The new questions:
So, now that we've got a bigger view of the picture, how should the group as a whole be handled? In particular, what should (or, for that matter, can) we say about evolution in this group, and how should we classify the relationship between the Oahu, Kauai, and Maui Nui populations? I'll post on this again in another couple of days - let's see what the comments bring this time.
Montgomery, Steven L. 1975. Comparative Breeding Site Ecology and the Adaptive Radiation of Picture-Winged Drosophila (Diptera:Drosophilidae) in Hawaii. Proceedings of the Hawaiian Entomological Society. V. 22(1), pp.65-103
Ohta, Alan T. 1980. Coadaptive Gene Complexes in Incipient Species of Hawaiian Drosophila. The American Naturalist. v. 115(1), pp.121-131.