142 BOTANICAL GAZETTE [February 



the suberin. If suberin is for any reason more abundant in the 

 diseased bark, an error would thus be introduced which might 

 invaUdate any comparisons based on the results obtained. Sub- 

 ject to this possible correction the results shown in table XXVII 

 confirm those obtained in the microchemical analysis; that is, 

 they show a progressive decrease in tannin as the bark is more and 

 more affected by the disease. Healthy bark was found to contain 

 on the average 4.06 per cent of tannin, diseased 2.99, and dead 

 o .97. If sample i healthy, which gave a high figure, and sample 7 

 dead, which gave a low figure, be eliminated, the averages become 

 healthy 3.51, diseased 2.99, dead 1.33. The figures for samples 

 3, 6, and 9, all from the same tree, are healthy 3 .38, diseased 2 .93, 

 dead i .51. There is undoubtedly a difference between bark from 

 a sound limb and seemingly healthy bark from a limb that is badly 

 cankered. The latter is usually shghtly browned throughout 

 when first cut ofT and rapidly becomes reddish brown on exposure 

 to air. Really healthy bark under such conditions shows only a 

 slight browning. 



Whatever the results with apple bark may mean, they are not 

 in agreement with the statement made by Kerr (see Cook and 

 Wilson, 15, p. 26, footnote) that because of the greater stability 

 of tannin and the disappearance of other constituents "all decayed 

 wood and bark give higher tannin contents, no matter what causes 

 the decay." If confirmed by further analyses they would indicate 

 a different relation between host and parasite with reference to 

 tannin in the case of blister canker than obtained in any of the 

 cases studied by Kerr. Leaching of tannin may account for 

 the low percentage found in dead apple bark, as suggested by the 

 chemist of the Chestnut Tree Blight Commission (15, p. 6) for old 

 cankers of chestnut blight, but can hardly be responsible for the 

 condition found in diseased bark. 



Summary 



I. Measurements with the simpHfied Bunzell apparatus show 

 that apple bark attacked by Nummularia discreta causes about 

 twice as much oxidation of pyrogallol, pyrocatechin, guaiacol, and 

 benzidine as does healthy bark. 



