2. In tissue imprints made with the spleens of antigen-stimulated animals, no morphological distinction was evident between areas showing non-specific fluorescence and specific fluorescence.

3. Elimination of non-specific fluorescence was not achieved by any one, or any combination of the following: (a) conjugating only gamma globulins with fluorescein isothiocyanate; (b) removal of dialyzable fluorescent products on sephadex, followed by concentration through the use of pressure dialysis; (c) use of crystalline preparations of fluorescein isothiocyanate.

4. Individual preparations of fluorescent antibodies were separated by gradient elution chromatography on diethylaminoethyl (DEAE) cellulose into fractions possessing different numbers of fluorescein radicals per molecule of globulin.

5. The coupling ratio of 50 mg fluorescein isothiocyanate (FITC) per gm of protein, as commonly advocated, can not be recommended for the precise localization of antibody globulin in tissues owing to the capacity of the coupled products to give non-specific fluorescent staining. When crystalline preparations of FITC are used instead of the amorphous product at 50 mg/gm protein, far too high non-specific fluorescence results.

6. A fraction with bright specific fluorescence and no or negligible nonspecific fluorescence was obtained from each fluorescent antibody that was prepared by using 6 to 8 mg of crystalline fluorescein isothiocyanate per gm of globulin and was then subjected to DEAE-cellulose chromatography and gradient elution to eliminate the most highly coupled molecules.