In fall of 1974, Richard McCormick proposed a novel interpretation of meaning of valid proxy consent in case of children. Addressed as it was to a perplexing problem in medical ethics, legitimacy of nontherapeutic experimentation on children, McCormick's article, Proxy Consent in Experimental Situation,' has had a wide influence. It is my contention, however, that McCormick's ethical analysis is incorrect, that it suffers from a moralistic quality, and that it is bound to weaken protection of children in this age of research medicine. Indeed, it may have already done so. One can surely describe McCormick's position as novel. By insisting on a voluntary consent of human subject for all experimentation, Nuremberg Code seemed to rule out altogether nontherapeutic experimentation on children or incompetent. Other medical codes adopted since then have allowed parents and guardians to enter children and incompetent persons into nontherapeutic research, but this right was ascribed simply by stipulation. No attempt was made to argue that validity of proxy consent could be grounded in subject's presumable consent. That is what McCormick does, with far-reaching consequences. McCormick begins his argument with a review of medical codes and of professional ethicians. Not surprisingly, he finds profoundly diverging views in this literature. Nonetheless, the ethician who has discussed this problem at greatest length, writes McCormick, is Princeton's Paul Ramsey; and it is against my that McCormick develops several of his arguments. Ramsey, he writes, denies validity of proxy consent in nonbeneficial (to child) experiments simply and without qualification (p. 8). As evidence he submits following passage from my book The Patient as Person: To attempt to consent for a child to be made an experimental subject is to treat a child as not a child. It is to treat him as if he were an adult person who has consented to become a joint adventurer in common cause of medical research. If grounds for this are alleged to be presumptive or implied consent of child, that must simply be characterized as a violent and a false presumption.2 This, in fact, remains my view of matter; I have read nothing that persuades me that this, my original position, is incorrect. Nonetheless, in view of necessity sometimes claimed for nontherapeutic research with uncomprehending subjects, several years ago I did explore, by way of a thought experiment, an alternative position. If today we mean to give such weight to research imperative, I argued, then we should not seek to give a principled justification of what we are doing with children. It is better to leave research imperative in incorrigible conflict with principle that protects individual human person from being used for research purposes without either his expressed or correctly construed consent. Some sorts of human experimentation should, in this alternative, be acknowledged to be borderline situations in which moral agents are under necessity of doing wrong for sake of public good. Either way they do wrong. It is immoral not to do research. It is also immoral to use children who cannot