The Zoque people of southern Mexico pray to the rain gods every year by releasing a leaf-bound paste made of lime and the ground-up root of the barbasco plant in to cave waters. This mixture is a natural fish toxin.
Researchers from Oklahoma State Univ. and Texas A and M found out that fish populations exposed to this poison have evolved a resistance to it over centuries.
From the Science Daily press release:
However, a team of researchers led by Dr. Michael Tobler, an evolutionary ecologist at Oklahoma State University, and Dr. Gil Rosenthal, a biology professor at Texas A&M, has discovered that some of these fish have managed not only to develop a resistance to the plant's powerful toxin, but also to pass on their tolerant genes to their offspring, enabling them to survive in the face of otherwise certain death for their non-evolved brethren.
The use of the phrase ....these fish have managed not only to develop a resistance.. may be read by many who are not familiar how evolution works to mean that some individual fish over their lifetimes develop a physiological resistance to the poison and then pass on that trait to their offspring.
That is inheritance of acquired characteristics and that is not how evolution by natural selection works. Rather some fish within a population will by chance happen to have genes that confer some resistance to the poison. These individuals will leave behind more descendants than individuals who don't possess that version of the gene. Over time the poison tolerant gene will become more common in the population.
The press release does better later in the article:
Mollies able to tolerate the poisonous conditions survived and passed those traits to their offspring, resigning those that perished to their fate of serving as a ceremonial feast for the Zoque.
The two came out meaning different things to me in terms of how evolution works... the second paragraph sounds more accurate.
Link to paper: An indigenous religious ritual selects for resistance to a toxicant in a livebearing fish. Biology Letters, 2010; DOI: 10.1098/rsbl.2010.0663