The purpose of the present study was to design a standardized model of hemorrhagic shock in dogs as a prerequisite for future studies on the pathogenesis of this syndrome. Hemorrhagic shock was induced in 86 mongrel dogs submitted to a systemic hypotension of 45 mm Hg by means of a constant pressure reservoir procedure. The following blood volumes (ml/kg) were measured during the evolution of shock: (1) initial bleeding volume, (2) secondary bleeding volume, (3) maximum bleeding volume, (4) automatic reinfusion volume, and (5) duration of the hypotensive period (min). The corresponding mortalities were submitted to a statistical analysis which yielded the following conclusions: (a) the initial, secondary, and maximum bleeding volumes do not correlate with mortality, (b) the duration of the hypotensive period is of only relative importance, and (c) the automatic reinfusion volumes shows a high degree correlation with the fatal outcome.