H.L. Carson's tree of the picture wing Hawaiian Drosophila is a spectacular work of evolutionary biology. Indeed, it can be considered a formal proof of species evolution. Above, we reproduce the tree in slightly altered form, with the four major species subgroups color coded and labeled. The original work is described in the references below.
Carson compared the polytene chromosomes from scores of species of Hawaiian Drosophila and discovered that certain species shared particular gene arrangements. These gene arrangements are visible in the banding patterns of each chromosome, which can be read like a "bar code":
The most common variations in banding pattern that Carson observed were "inversions". Each inversion is a 180-degree rotation of part of the chromosome, caused by two chromosome breaks. Inversions occur sporadically over time, and randomly across the genome. Once an inversion occurs, the new banding pattern acts as a permanent historical marker of the inversion event. Dozens of inversions can be mapped in each species, and each appears to be unique (see note below). Thus, inversions provide unambiguous evidence of the lineage of any given species, and how closely related it is to any other species in the group. Importantly, this inversion method is independent of the more classic lineage-tracing method in which morphological features are compared among species. It is also independent of DNA sequence comparison, a newer technique for detemining species relationships.
A fourth independent method for assessing lineage relationships is made possible by Hawaii's unique geography: the islands form on a "conveyor belt" as the Pacific plate moves slowly over a volcanic hotspot. The age of each island is known, and the newest island (Big Island) is still forming; its oldest region is under 0.5 million years old. Let's assume that one species of Drosophila landed on one of the old Hawaiian islands, and by a succession of colonization and speciation events, it gave rise to today's multitude of species. If so, the lineage should show that new chromosome inversions are accumulating on each successively formed island. If, instead, evolution were not driving speciation, then the distribution of inversions should be random with respect to Hawaiian geography, DNA sequence features, and the appearance of each fly.
The results are stunningly clear. The inversions, when plotted in tree form (independent of all other information), show a clear "flow" of species from older to newer islands (top figure). Above, the data are also plotted on the map of Hawaii; the blue arrows indicate number of species that "island-hopped", and the direction. For the most part, newer species are derived from species on older islands. Since the current islands have coexisted for ~0.5 million years, there are also cases of colonization back to older islands, and skipping of islands, but these are much less frequent as the model predicts (notice there are no colonizations back from the Big Island).
Carson's tree also matches extremely well with lineages based on the appearance of the flies (Kaneshiro et al. 1995). Our 2007 paper compares the tree with the diverse wing patterns of each species, showing a remarkable correlation. It was also recently shown that the patterns of DNA sequence change in each species closely match the inversion tree. The overall pattern of species evolution, through colonization of newly formed land, is unmistakable.
It is rare to have the opportunity to compare this many independent types of evidence for speciation. This is why Hawaii is considered a uniquely rich, natural "evolutionary laboratory".
See:
Carson HL (1970) Chromosomal tracers of evolution. Science 168:1414–1418.
Carson HL (1983) Chromosomal sequences and interisland colonizations in Hawaiian Drosophila. Genetics 103:465-482.
Carson HL (1992) Inversions in Hawaiian Drosophila. In: Krimbas CB, Powell JR, eds. Drosophila inversion polymorphism. Boca Raton, FL: CRC Press. pp 407-439.
Kaneshiro, K.Y., R.G. Gillespie, and H.L. Carson. 1995. Chromosomes and male genitalia of Hawaiian Drosophila: Tools for interpreting phylogeny and geography. pp. 57-71. IN Hawaiian Biogeography: Evolution on a Hot Spot Archipelago, W.L. Wagner and E. Funk, eds. Smithsonian Institute Press
Notes: Are inversions really unique? It is possible the same inversion could occur twice in two different lineages. This would distort the tree, causing species to be called sisters even if they are not closely related. Such duplicate inversions are nearly impossible if chromosome breaks are random: if we can resolve 200 different breakpoints on the polytene map of each chromosome, then there are 40,000 possible unique inversions per chromosome (times five major chromosomes). However, duplicates might occur if each inversion breakpoint harbors a transposon, and the inversion is caused by the transposons hopping out of the chromosome simultaneously. Any two species with those particular transposon inserts might acquire the same inversion. Alternatively, the breakpoints could be prone to illegitimate recombination because they contain repetitive sequences. In the Hawaiian flies, however, all the inversions appear to be unique events: duplicate inversions have never been demonstrated. It would be very infomative to clone the breakpoints of some of the major inversions in grimshawi now that its sequence is available.