The Scientific American headline asks, "Can Babies Be Made to Order?" It is typical of SciAm, as well as of many members of the research community, that the more important question was overlooked: "Should Babies Be Made to Order?"

The article is an interview with Maria Lalioti, an expert in pre-implantation genetic diagnosis, or PGD. In PGD, embryos are created in vitro and then screened for certain genetic traits, which may include traits incompatible with life or merely traits that increase risk for certain diseases.

The embryos that fail the screening are labelled "undesirable." I'm not making that up: Lalioti says the parents decide the fate of "undesirable embryos" whose genes are imperfect. They may be experimented on, or they may be killed. Lalioti's euphemisms, respectively, are "donated to science" and "discarded."

Let us be crystal clear about what PGD does and does not do. PGD does not save lives — it takes them. It kills young, tiny humans before a couple invests significant affection and financial resources on them. It kills them before they are cute enough to inspire sympathy.

PGD is a ghastly "solution" to preventing diseases with a genetic component. It's the solution of Adolf Hitler, who also took undesirable people and discarded them or donated them to science. At least his victims could look their killers in the eye.

The real kicker occurs near the end of the interview, in which Lalioti self-righteously announces that her clinic does not permit sex selection of interviews "for ethical reasons." She derisively adds that some clinics "are making a lot of money" doing just this.

As for the question "Can Babies Be Made to Order?" The answer, according to Lalioti, is probably not. We don't know enough about the genes that influence cosmetic traits like hair and eye color. After all, researching those genes does nothing to teach us how to save lives. PGD does not save lives, either; it takes them.

Correction: When first published, this post incorrectly stated that Dr. Lalioti's clinic "does permit sex selection." This was a typo; it was supposed to read "does not permit sex selection." The error has been corrected and I apologize for any misunderstanding.