Andrew Sullivan has been trying to get an estimate on the natural miscarriage rate as background information for a planned section on abortion in his next book. Here’s the conclusion he has reached: “The scientific consensus is that, at the most conservative end of the spectrum, half of all unborn children die before they even get a chance to get implanted in the uterus (some estimates put that figure even higher at around 70 percent). Of those that successfully get implanted, the mortality rate is lower, somewhere around 35 percent.”

I think it’s a little more complicated than that. Many miscarriages, I gather, result from incomplete or defective fertilizations, and thus do not represent the deaths of embryos. From having glanced through this literature a while back, it seems pretty clear to me that Sullivan’s estimate counts these events in both his numerator as embryonic deaths and in his denominator as conceptions. The pro-life view he is examining does not compel you to do either.

I don’t think anything morally turns on this question, for reasons Ross Douthat goes through on Sullivan’s site. But to provide the kind of factual background Sullivan wants I think you’d need the answers to these questions: 1) What proportion of miscarriages are the result of failed fertilizations? 2) Leaving these events out, what’s the natural rate of early death? 3) What proportion of induced (deliberate, non-natural) abortions actually kill embryos? 4) Leaving incomplete or defective fertilizations out, what’s the ratio of natural to deliberate embryo-killings in the U.S.?