Using the criteria of generalized anaphylaxis, the SchuhzDale test, and mast cell degranulation occurring on contact with antigen in vitro, a study has been made of the ability of antisera, prepared in various species, to sensitize guinea pig tissues passively. Whereas small amounts of guinea pig and rabbit antibodies and rather larger amounts of monkey, dog and human antibodies were effective, antibodies from goat, horse, rat, fowl and a human auto-immune thyroid antiserum were ineffective. These results are in general agreement with those in the literature. Using antibodies labelled with 131I, no gross differences were found between those which sensitized and those which did not, either in respect of uptake on guinea pig mesentery in vitro, or retention in the tissue after uptake in vivo. By means of antigen labelled with 131I it was also shown that the antibody adsorbed on guinea pig mesentery could combine equally well with antigen, whether the antibody came from rabbit (sensitizing) or rat (non-sensitizing). These experiments also indicated that the amount of adsorbed antibody necessary for sensitization is very small. Reversed passive anaphylaxis was studied with a variety of different antigen-antibody systems, using the criteria listed above. It occurred when, but only when, the antigen was a γ-globulin, and when both antigen and antibody globulins came from species whose antibodies were able to sensitize guinea pigs for direct passive anaphylaxis. It is concluded that the explanation of the differences between antibodies or antigens which sensitize guinea pigs passively, and those which do not, is to be found in the nature of the antigen-antibody combination as well as in the manner of ‘fixation’ on the tissues. The possible importance of ability to activate guinea pig complement is discussed. [ABSTRACT FROM AUTHOR]