Human chorionic gonadotropin was administered intramuscularly in an open study of 32 women whose last three pregnancies had ended in spontaneous abortion. An initial dosage of 10 000 iu on earliest diagnosis of pregnancy was followed up by 5000 IU twice weekly up to wk 12, then once weekly up to wk 16. Only 2 of the 32 patients aborted (6.2%). Using the same regime a further 20 habitual aborters were randomly allocated double-blind to either hCG or placebo. All 10 women on active therapy continued (100%) compared with only 3 on placebo (30%) (P less than 0.01). These results suggest that hCG should be considered as treatment in women with habitual abortion where other specific causes have been ruled out and a potentially treatable hormonal imbalance surmised.