One reason that the book was so effective was that Williams demonstrated how natural selection could influence the full course of a species’ life history. It wasn’t necessary to think of growing old as being for the good of the species, for example. Instead, Williams argued that the decline of old age could be caused by pleiotropy–in other words, the harmful side effects of genes selected for advantages they offered during youth. Just as long as the advantages of these genes outweighed the disadvantages, they would become widespread. Ironically, cancer, declining stamina, deteriorating vision, and various diseases of old age could all be the result of natural selection, says Williams: “Pleiotropy is the ultimate reason for all these things.”Pleiotropy is indeed possibly the reason why people get cancer, etc. I think of pleiotropy as this crazy wild card with which all sorts of weird and unexpected traits can evolve. A disadvantageous trait, which natural selection under normal circumstances would purge, may exist in a population (go to fixation) because an allele (i.e., a version of a gene) wholly or partly responsible for the disadvantageous trait also affects another trait that is advantageous. If the combined effect of the two (or more) traits is that the organism has higher fitness, then the disadvantageous trait will evolve. Many organisms have traits that some people propose evolved because they initially were adaptations, like big brains or feathers. And many of these may well be, but it isn't out of the question that they initially evolved via a pleiotropic effect.
Pleiotropy may aid in adaptation when epistatic interactions create a rugged fitness landscape. In a rugged fitness landscape, it will sometimes be the case that in order to adapt, a lineage/population must cross through a valley in the fitness landscape before it can reach higher ground. When several mutations are necessary to change a gene (or a set of genes) enough that the organism becomes fitter than the original, and the intermediate genotypes are of lower fitness than both beginning and end product, then pleiotropy may rescue the situation by negating (or attenuating) that negative effect, such that the lineage can manage crossing the valley (which then really isn't a valley, when the second function of the gene (or set of genes) is taken into account).
After having to explain the meaning of epistasis and fitness landscapes to some people who have zero background in biology and evolution, I have become keenly aware just how strange the above may sound when one is unfamiliar. Apologies. Questions are welcome.
I am hosting the next edition of Carnival of Evolution myself over on Carnival of Evolution on October 1st (submit your posts here). Perhaps in memory of Williams?