Examined the relationship between the structure of caregiver personal networks, support, and perceived distress. It is argued that network structure should have little direct effect on distress but should have indirect effects via support. Data from 83 family caregivers to chronically mentally ill family members living at home in Summit County, Ohio, are used to test these assertions. The results provide mixed support for the general hypothesis. Most of the effects of network proportion of kin are indirect via support. Network size shows a similar pattern. The effects of network density on distress are mostly direct.