Speciation Fail: Single Bona Fide Example of Animal Speciation is Later Retracted
Three populations of polychaete worms were collected from the coastline around the Long Beach, California area. One population ("Lab") went through various bottlenecks, followed periods of population growth. The other two populations (P1 and P2) were collected directly from the field and crossed with the Lab population. While P1 and P2 could produce viable offspring when crossed with one-another, crosses between Lab and P1, and Lab and P2, could not. It appears that Lab is reproductively isolated from P1 and P2.
In this one instance, the FAQ cited a paper (Weinberg et al., 1992) which actually claimed to have witnessed the creation of complete reproductive isolation in the lab. But as the notable evolutionary biologist Theodosius Dobzhansky (1972) reminds us, "Reproductive isolation evidently can arise with little or no morphological differentiation."1 That seems to be the case here as the paper reports no morphological change between the populations.
But there's a very important epilogue to this story. Four years later, in 1996, the lead author of the original study co-published a follow-up study which essentially retracted and refuted claims of speciation in the lab. The follow-up paper states:
A critical assumption in Weinberg's experiment is that the P1 and P2 populations are, in fact, representatives of the natural population from which the Lab population hypothetically had diverged and speciated in the laboratory. We have tested this hypothesis by assaying 18 electrophoretic gene loci in the Lab, P1 and P2 populations and in an Atlantic population of a different species, used as a reference control. If the Lab population had speciated from P1 or P2, we would expect that randomly selected electrophoretic markers should be largely similar between the Lab and P1 or P2 populations. However, no common alleles between Lab and P1 or P2 are found in 13 (725) loci, and at two more loci the alleles fixed in Lab are at low frequencies in P1 and P2. The genetic distances between Lab and P1 or P2, are 1.75 ± 0.51 and 1.76 ± 0.52, larger than between most pairs of congeneric species in many sorts of organisms; and roughly similar to the distance between P1 or P2 and the reference population from the Atlantic (D=1.36 ±0.40). The Lab population is genetically depauperate, most likely as a consequence of the founder event, but this reduced variability contributes only trivially (about 1%) to the genetic differentiation between the populations. We conclude that the Lab population was already a species different from P1 and P2 at the time when it was originally sampled in 1964.2In other words, what happened was the investigators had originally sampled a naturally occurring independent species of polychaete worms, and used those individuals to establish the "Lab" population, and then later mistakenly concluded that an entirely new species had formed in the lab.
Thus, this is not an example of speciation in the laboratory, but the original investigation had simply sampled two naturally occurring separate species. Unfortunately the Speciation FAQ has not been updated to accommodate these findings, reported 15 years ago in 1996.
The initial Weinberg et al. (1992) paper which originally reported this alleged example of speciation stated: "the entire process of speciation has rarely been observed."3 This example did not remedy that problem.
For additional details, please see the full response to the TalkOrigins Speciation FAQ.
[1.] Theodosius Dobzhansky, "Species of Drosophila," Science, Vol. 177 (4050):664-669 (August 25, 1972).
[2.] Francisco Rodriquez-Trelles, James R. Weinberg, and Francisco J. Ayala, "Presumptive Rapid Speciation After a Founder Event in a Laboratory Population of Nereis: Allozyme Electrophoretic Evidence Does Not Support the Hypothesis," Evolution, Vol. 50 (1996): 457-461 (emphasis added).
[3.] James R. Weinberg, Victoria R. Starczak, Daniele Jörg, "Evidence for Rapid Speciation Following a Founder Event in the Laboratory," Evolution, Vol. 46(4):1214-1220 (August, 1992).