Population density is one of the most important factors influencing immune function. Social stress induced by higher density may account for the immunosuppression according to the endocrine hypothesis. To test this hypothesis, male striped hamsters (Cricetulus barabensis) were classified into the One/Cage (n = 9), Two/Cage (n = 6), and Three/Cage (n = 9) groups, and the treatment lasted for 45 days. The titers of immunoglobin (Ig)G15 and IgM 10 were lower in the Two/Cage group compared to the One/Cage group, indicating that higher housing density suppressed humoral immunity. However, the masses of thymus and spleen, phytohaemagglutinin (PHA) responses at 6 h, 12 h, 24 h, 48 h, and 72 h after PHA injection, the titers of IgG 5, IgG10, IgM5, and IgM15 were all not affected by housing density. Blood glucose level was higher in the One/Cage group than the other two groups, leptin titers did not differ among the three groups, whereas corticosterone concentration was higher in the Two/Cage and Three/Cage groups than in the One/Cage group. Moreover, negative correlation was observed between corticosterone concentration and the titers of IgG5, IgG10, IgG15, and IgM10. These results suggested that humoral immunity was reduced by higher stress levels induced by higher housing density, which supported the endocrine hypothesis. White blood cell (WBC) count was higher in the Two/Cage group than in the One/Cage group, and intermediate granulocytes (MID) were higher in the Two/Cage group than in the One/Cage and Three/Cage groups, indicating the fight and injury might have arisen in the higher housing density.