Methods are described which have been used to investigate isoagglutination reactions of the blood of cats. Evidence is presented which indicates that cats possess at least three groups. The methods and results of earlier workers which indicated that grouped isoagglutinins did not occur in cats have been criticized. A genetical system for the cat groups similar to the human ABO system is shown to be untenable. Two other possible systems similar to the A 1 A 2 O, and to the Rh models of humans are suggested and briefly discussed.