Peterson and Bossio (1991) have proposed that pessimistic people, by default, will have predominantly pessimistic friends. We tested that notion by examining the correlation between the scores of close friends on both separate and direct measures of optimism and pessimism (via the O/P instrument) and a measure of explanatory style (EASQ), which purports to tap optimistic-pessimistic tendencies as a bipolar dimension. In two studies, only pairs of male friends yielded significant correlations, and only on the direct optimism measure (r= .65, both studies). Another significant correlation (r=-.37) was found in Study 2 between sensation seeking scores of mixed-sex friends. No between-friend correlations were found in Study 2 for loneliness and assertiveness. Finally, both pessimism, as directly measured, and loneliness were correlated with typical friendship duration (r= -.24 and -.32 respectively), while none of the personality measures were correlated with number of reported friends.The possession of social networks is predictive of physical health, psychological health, and mortality (for reviews see Cobb,