In this issue of theArchives(see p 52), Stevens and Merigan approach an important problem facing physicians charged with the care of immunocompromised patients. That is the question of whether or not zoster immune globulin is effective in preventing dissemination of herpes zoster when administered following the appearance of localized disease in these patients. Localized herpes zoster occurs frequently in immunocompromised patients, especially those with Hodgkin's disease, 1 and dissemination is common. 2 Although the mortality of disseminated zoster is low, there is considerable morbidity associated with this disease. 3 It is for this reason that the question regarding zoster immune globulin is an important one. It seems reasonable that the administration of zoster immune globulin might be effective in preventing dissemination of herpes zoster. Following the initial infection the virus persists as a latent infection not accessible to neutralizing antibody. When it reappears as a localized infection, it