In 1982, 43 ragweed-sensitive patients receiving maintenance injections of full doses of ragweed allergoid were selected for a study of the immunologic and clinical efficacy of booster injections only four times a year. These patients had participated for 2 to 7 years as part of a trial of mixes of up to four allergoids to common pollens in the mid-Atlantic area tailored to each patient's skin test sensitivity. They were divided into a group (21 patients) to receive injections every 3 months and a group (22 patients) to receive injections about every 6 weeks (eight injections per year). Patients were rerandomized after 1 year so that half of each original group switched to the alternate treatment, and this program was continued until after the ragweed season of 1985. Thirty-four patients were still under study the last year. Doses, per injection, were 100 allergoid units (1 allergoid unit equals 100 PNU) of each allergen in the mixture. Symptom scores during the 8 weeks of each of the four ragweed seasons were not significantly higher in the 3-month treated group. IgG antibody levels to Amb a I (antigen E) were followed until early 1984 and were not significantly different in the two groups, even though the 6-week treated patients received a two times higher cumulative dose per year. Rates of local and systemic reactions (percent of injections eliciting reactions) were not different in the groups, which means that the 3-month treated group had about half as many reactions by virtue of taking half as many injections.(ABSTRACT TRUNCATED AT 250 WORDS)