I make my own yoghurt at home using a variety of milks: pasteurized whole or low–fat cow’s milk, powdered non-fat cow’s milk, and powdered whole goat milk. (I always mix powders with the amount of water needed to reconstitute the milk to the natural concentration of milk solids.) Sometimes I use a mixture of two or three types of milk. I always use yoghurt from the previous batch as the source of bacteria (starter culture).
I have noticed that if I repeatedly use one kind of milk for several weeks or months, then, if I switch to another type of milk, or to a mixture of a new type and the one I have been using, the yoghurt is likely not to ferment and solidify much. If I then use the most solid parts of this poorly fermented batch as the starter for the next batch, and so on, I may eventually, 2 or 3 or more batches later, succeed in getting good, solid yoghurt with the new type of milk. (Sometimes I gave up after a few batches and then mixed in some store-bought yoghurt along with my own as starter for the next batch.)
I suspect that what happens in these cases is that natural selection, possibly in conjunction with back and forth mutations, results in the bacteria population becoming adapted to grow best with a certain kind of milk. For example, if they are growing for many generations in non-fat milk, they may lose the ability to deal with fat, but gain efficiency in using only non-fat milk solids.
Such shifts over time in abilities of yoghurt cultures may be confounding variables in studies on yoghurt production. For example, 2 recent studies on the effects of varying the casein protein to whey protein ratio of the milk seem to have yielded opposite results with regard to firmness or viscosity of the yoghurt. Fenelon, M.A. et al. (Irish J. of Agric. and Food Res. 39:171 (2000)—findings presented on the web as “Improving the quality of yogurt”) found that more whey protein increased firmness, whereas Amatayakul, T et al. (Food Hydrocolloids 20:314-24 (2006)) found that more whey protein decreased firmness. The difference in the results may have been due to a difference in the milks to which the bacteria were adapted when the experiments were done.