We conclude that, although many therapies have been advocated, no regimen has proved to be consistently effective in the treatment of idiopathic male infertility. Couples in which the husband is identified as having idiopathic infertility should be advised of the inconsistent and often low conception rates obtained with medical therapy. This should be weighed against the possibility of greater success with in vitro fertilization and the likelihood of success with artificial insemination by donor. Matson and colleagues performed in vitro fertilization on 75 couples in which the husband was oligospermic. When the husband was moderately (5.1 to 11.9 million motile sperm per milliliter) or severely (less than or equal to 5 million motile sperm per milliliter) oligospermic, fertilization rates were 56 and 30 per cent, respectively. This is in comparison to a fertilization rate of 72 per cent in normospermic couples. Following embryo transfer, pregnancy rates were similar in all groups. In vitro fertilization, although expensive and often not covered by insurance policies, may yield results in 1 month. Pharmacologic treatment of the male, which is less expensive, requires several months before improvement might be expected. The decision as to which course to recommend should be made after careful consultation with the couple. If empiric therapy is decided upon, the choice of an agent is somewhat arbitrary. Reasonable initial choices for the oligospermic patient are tamoxifen (or clomiphene citrate) or HCG (HCG may also be used in the patient with idiopathic asthenospermia). Testosterone rebound, with its risk of permanent azoospermia, is not an acceptable initial therapy. Similarly, the results of studies of testolactone, GnRH, pentoxifylline, and kallikrein either demonstrate low pregnancy rates or are too preliminary to recommend at this time. Regardless of the choice of therapy, it should be administered for at least 3 months to include the length of one spermatogenic cycle. The performance of randomized, double-blind, placebo-controlled, cross-over studies of present and future treatments will allow more definite conclusions to be drawn.