later weighed to determine the health and stress status of each rat at time of death.

Plasma ACTH and CORT concentrations from the rats were later analyzed with

radioimmunoassay (RIA). ACTH RIAs were run with kits from Alpco Diagnostics

(Salem, NH) and CORT RIAs were run with kits from Diagnostic Products Corporation

(Los Angeles, CA).

 Statistical Analyses

 Potential between-groups differences in number of defeats and latency to first

defeat were analyzed using two one-way analyses of variance (ANOVAs) to compare the

five groups of intruders (i.e. three repeatedly stressed, and two just acutely stressed) in

Experiment la during their first exposure to social defeat stress. A 3x6 (group x session)

repeated-measures ANOVA (for Experiment la) and a 2xl2 (group x session) repeated-

measures ANOVA (for Experiment lb) were used to examine potential differences in

number of defeats between groups and across experimental sessions for the repeatedly

stressed groups. A 3x6 (group x session) repeated-measures ANOVA (for Experiment

la) and a 2xl2 (group x session) repeated-measures ANOVA (for Experiment lb) were

used to examine potential differences in latency to first defeat between groups and across

experimental sessions for the repeatedly stressed groups.

 Potential between-groups differences in freezing responses were analyzed using a

one-way ANOVA to compare freezing behavior between the five groups in Experiment

la during their first exposure to social defeat stress. A 3x6 (group x session) repeated-

measures ANOVA (for Experiment la) and a 2xl2 (group x session) repeated-measures

ANOVA (for Experiment lb) were used to examine potential freezing behavior

differences between groups and across experimental sessions for the repeatedly stressed

groups.