There are things about one's life and especially about one's children that cannot be known in advance and to which it would be foolish to assume a right outcome or a wrong one. How, for example, could you possibly know whether you and your family would be better off having had a boy child and then a girl, or a girl first, or two girls or two boys? What would your standard for comparison be? Which child would you have given up in order to have had the imaginary right family? You might have thought you wanted a girl, and then gotten one not at all like the girl you expected—not gentle, not Barbie-doll loving, utterly unlike your childhood self. You might have thought that having one child of each sex was perfect, and then found that family harmony eluded you nonetheless. Certainly, the more precise your expectations of your children and the more convinced you are that those expectations can be met (indeed, that you deserve to have them met), the more disappointed you are likely to be. Without some humility in the face of the unknown and unknowable—without some fundamental willingness to accept and to love whatever child you get—parenthood would be a very different, and a lesser, thing.

One might think that all this would be obvious to anyone who was ever part of a family, or read a novel about one. But in an era when people can shop, online, for sperm and egg donors of specified height and hair texture and IQ—in an era, in other words, of consumer-driven, privatized eugenics—it's easy to forget about what we cannot control or predict. In this environment radical reproductive technologies tend to slide into use when someone wants them and can pay. There's always at least one fertility doctor out there who will do anything, no matter how risky or ethically dubious. Religious absolutists fume; free-market purists and biotech groupies celebrate; and the rest of us scratch our heads, troubled by a vague sense that something's wrong here.

For several years now clients of an infertility clinic called The Genetics and IVF Institute, in Fairfax, Virginia, have had access to a sperm-sorting technique that allows them to choose—with about 90 percent reliability for girls and 73 for boys—the sex of their child. Some people use this service to avoid passing on serious diseases that can be inherited only by boys. But most use it for what might be called social or aesthetic reasons: they think one child of each sex makes for the ideal family; they have particular goals for their children, which they imagine are best fulfilled by one sex or the other. For these customers the Genetics and IVF Institute has a principle: it's called "family balancing," and it means that institute doctors will do the procedure only for people who already have one or more children of one sex and want a child of the opposite sex. This is supposed to be an answer to the charge of gender discrimination, the charge that by offering this service they are catering to people who value one sex over the other—people like the parents in India and China who have substantially upset the balance of the sexes in some regions by aborting (categorically undesirable) female fetuses. But because most of the clients using the sperm-sorting method presumably did not use it the first time around (and therefore had to take whatever they got), there's no guarantee that they do not value one sex over the other.

Still, the problem with this method, as far as some people are concerned, is not that it is being marketed at all but that it is not quite accurate enough. A more sophisticated, fail-safe method of sex selection uses pre-implantation genetic diagnosis (PGD)—a procedure by which embryos created outside the body can be tested for certain genetic disorders before they are transferred to the uterus. In this case, sex is treated as a disease—something to be screened for, like Huntington's chorea or cystic fibrosis. Clients can pick a boy or girl embryo and discard the rest. The Genetics and IVF Institute is already offering PGD for family balancing, but quietly; neither their Web site nor their literature mentions it.

This use of PGD had a brief airing in the news last September, when an enterprising fertility doctor named Norbert Gleicher sought the blessing of the American Society of Reproductive Medicine to start offering this service. John Robertson, the acting chair of the ASRM ethics committee, gave the go-ahead, but a small flap ensued when it transpired that Robertson had acted on his own. This proved slightly embarrassing for the ASRM, which had in 1999 issued a statement discouraging the use of PGD for sex selection on the grounds that it undercut the "acceptance of offspring for themselves and not their nonessential characteristics" and that fertility treatment posed "health risks and burdens" and should not be undertaken whimsically. But the truth is that the ASRM ethical pronouncements lacked teeth, and there is no other regulatory body equipped to stop sex selection—or almost any other gene-shopping innovation.

The real trouble with sex selection goes beyond sex discrimination. The real trouble is that it allows us, for the first time, to use a medical procedure to select or reject a child on the basis of a characteristic that has nothing to do with life and death, that is not in any sense of the word pathological, that cannot possibly be construed as sparing a child any pain and suffering. It might sound harmless enough, maybe even kind of cute—this impulse to pick and choose, pink or blue. But if we allow people to select a child's sex, then there really is no barrier to picking embryos—or, ultimately, genetically programming children—based on any whim, any faddish notion of what constitutes superior stock. This time the old and overused accusation would actually be true: we would be playing God—and none too well, in all likelihood, given how little we really know about what makes individual human beings the way they are. A world in which people (wealthy people, anyway) can custom-design human beings unhampered by law or social sanction is not a dystopian sci-fi fantasy any longer but a realistic scenario. It is not a world most of us would want to live in.