After intensively treating 18 agoraphobic patients with 13 1/2 hours of exposure in vivo, we examined the effects on both patients and their spouses over a period of six months, using questionnaire measures of symptoms and marital adjustment. Those patients whose marriages were rated as unsatisfactory before treatment improved less during treatment, and were significantly more likely to relapse during follow-up, than those patients with satisfactory marriages. The marriages of nine patients appeared to be adversely influenced by their symptomatic improvement, and two distinct types of marital interaction were observed in relation to this. In one pattern, the patients' symptoms appeared to strengthen aspects of largely affectionless "compulsory" marriages; in the other, the patients' symptoms appeared to protect their spouses from recognizing or examining aspects of their own personal and interpersonal problems.