In later experiments, ** f and ** f were prepared by conjugating 8 mg of FITC per gram of globulin. These conjugates ** f had much less nonspecific staining than the previous conjugate (with 50 mg FITC per gram of globulin) while the specific staining was similar in both cases. Nonspecific staining could be satisfactorily eliminated by passing these conjugates through a DEAE-cellulose column as described for ** f. The best staining procedure with this purified ** f consisted of staining with the conjugate for 30 minutes and washing in PBS for 15 minutes. The specificity of staining with ** f was established as follows: ** f specifically stained tumor sections but not sections of healthy sweet clover stems or of crown gall tumor tissue from sweet clover. Sections of tumors incited by WTV were not similarly stained with conjugated normal serum or conjugated antiserum to potato yellow-dwarf virus.

After passing ** f through DEAE-cellulose, the titer of antibodies to WTV in the specific fraction was 1: 4 of the titer before such passage (precipitin ring tests by R. F. Whitcomb); but mere dilution of the conjugate to 1: 4 did not satisfactorily remove nonspecific staining. This indicates that increase in specificity of ** f after passing it through DEAE-cellulose was not merely due to dilution.

Specific staining by DEAE-cellulose treated ** f and ** f, although clearly distinguishable under the microscope from either nonspecific staining or autofluorescence of cells, was not satisfactorily photographed to show such differences in spite of many attempts with black and white and color photography. This was chiefly because of the bluish white autofluorescence from the cells. The autofluorescence from the walls of the xylem cells was particularly brilliant.