I have argued that because human sexual reproduction inevitably involves the creation and destruction of embryos, it is a problematic activity for those who believe that the embryo is “one of us.” Or, if it is not a problematic activity, then neither is the creation and destruction of embryos for a purpose of comparable moral seriousness—the development of lifesaving therapy, for example. I assume that, whereas it is possible for the very first act of unprotected intercourse to result in a live human baby, and hence not in a given case cause any embryo loss, this is a rare event and that, statistically, for every live birth from three to five embryos must be created only to die. For dramatic effect, I assume that five is a reasonable figure, giving each embryo a 20% chance of survival; however, nothing in the argument depends on any specific figures being correct.Søren Holm has been an important source of ideas for this paper. Julian Savulescu has also been very generous in commenting in detail on this paper, and I have taken account of the critique that I have received.