Platelet counts were performed in 43 Cavalier King Charles Spaniels (CKCS, a breed predisposed to macrothrombocytopenia) and in 10 control dogs using 3 automated systems and 3 manual methods (erythrocyte-lysing agents + counting chamber or evaluation of blood smear). Good correlations were found between platelet counts using all methods (all P < 0.0001; R2 = 0.71-0.85). Best correlations were found between the manual methods. Significantly larger platelets were found in CKCS with platelet count < or = 100,000/microl when compared with control dogs and CKCS with platelet count > 100,000/microl (both P < 0.0001). All platelet counts--except when made with the 2 counting chamber methods--were underestimated at platelet counts < or = 100,000/microl.