He has already observed that in the early middle ages, serfs could not contract legally binding marriages in Western Europe (when he was noting the point in time when that changed). And throughout the period covered by the book thus far, the institution of slavery exists in one form or another (I'm to the 1770s, where Phillips is getting into how the colonial legislatures responded to the additional freedom to allow divorce) and I know from elsewhere that slave marriages were very contingent to the extent they were recognized at all. Further, Phillips has spent a fair amount of time (and will spend more) describing separation (either through ecclesiastical or civil courts) followed by a nominally bigamous second marriage as a sort-of replacement for divorce. But when he cites divorce (_actual divorce_, not divorce replacement like the separation + bigamy thing) and attempts to put that in perspective, he puts it into perspective across the population of the jurisdiction in question.
I don't think that can possibly make sense. Even if his population calculation excluded the people we know couldn't access divorce because of being slaves (which he has not specified), I think it's safe to say that divorce was only accessed by a particular segment of the population -- the examples he gives are heavy on the aristocracy, and not the lower ranks, either. To give a sense of at what rate people accessed divorce, we'd need not an overall population figure, but a figure for that segment -- whether it be 1% or 10% or somewhere in between.
I recognize that it's probably not possible for most of this time to get any sense of more informal procedures for terminating one relationship and replacing it with another. I'm not asking that. I'm asking for a little more information on the kinds of people who did petition legislatures (or ecclesiastical courts) for relief (whether they got it or not). Maybe there were, occasionally, poor laborers -- I'd love to know! But if it really is all very rich and powerful folk, I'd rather see those numbers given against that subpopulation.