Well...the title is a little creepy, but sums up the melange in this report, Study finds that a woman's chances of having twins can be modified by diet. But there is more than diet, researchers have long known that genetics plays a role in twinning rates, it is heritable in that some proportion of the population variation correlates with genotypic variation. And we also know that the rise of fertility medicine has resulted in a boom of multiple births in the modern world. Twinning then neatly encapsulates the manifold aspects of many phenomena of interest which exhibit a biological angle. The researcher found that vegan women had 1/5 the twinning rates as non-vegan women. I don't have access to the original paper in The Journal of Reproductive Medicine, so I don't know how well variables were accounted for, but the magnitude of the difference seems great enough that given a decent sample size there is something there. Importantly, he also found that "The concentration of IGF in the blood is about 13 percent lower in vegan women than in women who consume dairy." IGF is a growth factor, and the article notes that the basal concentration of IGF is greatest in Africans, least in Asians and middling in Europeans, and this is exactly the rank order of fraternal twinning across the populations. So, not only do you have a suggestive correlation, but you have a suspect in this case for why that correlation might be cropping up. Also, northern Europeans have higher dizygotic twinning rates than southern Europeans. This is interesting because of course northern Europeans have a more dairy-rich diet, so the difference here might be accounted for by diet as opposed to genetics. One should also note that East Asians tend to be lactose intolerant, and traditionally do not consume dairy as adults, but sub-Saharan Africans also tend to be lactose intolerant, so it seems to be in this case that endogenous basal IGF levels may be implicated in this case (one could check the twinning rates of the Masai, who are lactose tolerant, vs. their Bantu neighbors who usually are not). A final note, there is a locus, igf2, which is a major illustration of genomic imprinting in humans (it obviously has a role in the generation of IGF). Paternal copies tend to be expressed, and induce a larger offspring, and this is opposed by the action of another locus, igf2r. I don't know enough about molecular genetics, but there might be some epigentic process going on here where milk consumption is modulating these loci through some pathway.