Lead (Pb(2+)) affects neuronal and endocrine systems that influence social interactions. By providing potential hiding locations, spatial heterogeneity may affect Pb(2+)-induced behavioral outcomes. Therefore, a test chamber was designed into which a refuge could be inserted. The refuge allowed test subjects to escape from the mirror image that stimulated agonistic interactions. Behaviors with a mirror were compared with baseline activity patterns without a mirror. Adult (12-month old) male and female zebrafish, exposed to Pb(2+) (0-10 μM) as embryos (2-24 hours post fertilization), were tested individually for 5 min in each chamber design within 2 h of feeding. Behaviors were evaluated for % time in mirror zone, distance traveled (=activity level), and attacks on the mirror image. When there was no refuge, significant concentration-dependent increases occurred in male % time in mirror zone, activity level, and number of attacks. Increases in these variables were less pronounced in females. When there was a refuge, there were significant differences for males only in activity level and attacks at the higher developmental exposure concentrations; % time in mirror zone followed a similar pattern and level as without refuge. Females displayed Pb(2+)-induced behavioral changes only for attacks on mirror. Since the presence of refuges that is, environmental enrichment, reduced Pb(2+)-induced agonistic behavior in both sexes, experimental spatial design can be considered an important factor when interpreting behavioral outcomes.