292 
Journal of Agricultural Research 
Vol. VIII, No. 8 
run a parallel course even though both are spiral. As a rule, the streaks 
are continuous, but occasionally they are intermittent, disappearing 
from view and reappearing higher up. In all such cases examined, 
however, the streak in the wood was continuous, the span of sound tissue 
being only in the bark. These streaks may show almost every propor¬ 
tional variation in length and width. There is a strong tendency for the 
streaks that run down to be terminated by intersection with a larger 
branch, but there are frequent exceptions. 
The leaves on some part or all of the affected branch may also show 
more or less injury. The first symptoms may be apparent as early as the 
first or second day after treatment. The leaves at first lose their normal 
luster and look dull and a little lighter green. This condition may affect 
the entire leaf, but often only spots and irregular areas. In bad cases 
there is considerable wilting. Soon the spots so affected turn brown and 
become dry and crisp. It is a notable fact that the leaves so affected are 
practically always directly touched by a discolored streak or area. A 
twig touched by such a streak may have every leaf killed, while on the 
opposite side of the limb, even if it is only a half inch in diameter, a twig 
will have leaves with no visible trace of injury. 
In studying the symptoms of injury through the smooth bark in many 
different cases, one is impressed by the great uniformity of certain 
general features and the great variety of differences in minor details. 
The uniform and constant symptom is localized discoloration of bark 
and wood, beginning as an indefinite discoloration and ending after a 
period of a few weeks as a sharply defined, dark, sunken area. If the 
absorption of arsenic is slight, the discolored region does not extend 
much farther longitudinally than transversely; but if the absorption is 
considerable, as through a wound, a long and relatively narrow streak is 
produced. Beyond these general features there is little uniformity, the 
relative length and breadth of the streaks, their relative length above 
and below the point of absorption, and their branching at intersections, 
with side branches presenting an endless variety of detail. 
A feature of this work that has given us constant trouble is the 
frequent failure of duplicate and triplicate tests to check with each 
other. When two or three applications were made under con¬ 
ditions apparently identical in every respect, the resulting injury was 
not always the same in extent and character. In the majority of 
cases the results were nearly enough alike for practical purposes 
(Table VI); but all too frequently duplicate and triplicate tests 
gave widely different results, as may be seen, for example, in Tables 
IV and V. These differences were most marked in limbs that had 
been wounded just before treatment and gave much less trouble in 
those that were unwounded. As a result of this erratic behavior, most 
of the tests were repeated over and over again for confirmation; and 
the reader is warned against drawing hasty and definite conclusions 
