34 BOTANICAL GAZETTE [ JANUARY 
time at which the curve begins. The first and third of these are 
the only comparisons that can be made from the tables. The 
second will be shown by a diagram. 
TABLE. xX. 
ROOTS SPLIT INTO HALVES AND AFTERWARD BOTH HALVES 
BRANDED ON OUTER SIDE OF TIP. 
Date Roots Hours after Both halves One straight, Both 
1897 used branding curved in one curved in straight 
Jan. 27 2 as I I 
29 4 2.5 4 
Feb. 3 3 1.5 I I I 
4 5 4 5 
5 3 3 2 
Apr. 13 8 6 8 
27 22 4 I 
: ~~ SLAG os a Re 3.8% 
It may be objected that these tables are not comparable 
because the time element is not constant. This objection is met 
by the fact that if we take table IX for the control of both VIII 
and X we have the time element varying similarly in the differ- 
ent tables. With a few exceptions the lots of the same date 
were under exactly the same conditions including the time ele- 4 
ment. Leaving out the exceptions and taking only those lots 
that were under the same conditions in all particulars, we get 4 
different numerical result, but the general showing of the tables 
is the same. The numerical change is this: 69.2 per cent. of 
the roots treated as in table IX correspond to 81.4 per cent. of 
those treated as in table X. These had both halves curved in 
toward the cut face. Also 42.4 per cent. of the roots in table 
IX correspond to 19.0 per-cent. of those in table VIII. These 
also had both halves curved in. 
Making due allowance for the variation of the time element 
a number of facts become apparent. In roots that have begun 
to curve when they are split, the half nearest to the brand, the _ 
