1916] GRAVES—CHEMOTROPISM IN RHIZOPUS 349 
It is evident from the results shown in table I that for small 
numbers, from roo to 200, the experimental error may vary from 
o to 16 per cent; but by taking a large number, such as 774, the 
error is reduced to about 10 per cent. Also, apparently owing to 
the personal equation of the observer, the error is always minus.® 
TABLE I 
RESULTS OF TEST COUNTS TO DETERMINE ACCURACY OF METHOD 
Class 
Preparation Total gh oon ge 
A B Cc D 
ie ae 50 34 58 38 180 pal 
ee 57 34 57 43 Igt wd 
oe ee ae 33 34 52 37 156 == 36 
Rie es. 24 28 39 36 127 —16 
Bee ee ay 18 32 23 04 20 
748 
(Average error for 748 hyphae=—11 per cent) 
Be eG 3: 29 31 35 35 oo 3 
2 Nes 54 oe 73 55 233 it 
Bee 41 29 40 20 139 — 8 
ee rere 25 28 36 33 $22 me © 
DOr is caa ss 33 30 48 39 sag a. 
774 
(Average error for 774 hyphae= —1o per cent)* 
* The same average error of 10~11 per cent was obtained ~— a recount was made after a few days’ 
interval, so that the method can be trusted to give uniform results. 
It is also noteworthy that the large error, up to 16 per cent, 
occurs only when the hyphae are not markedly turned in any one 
direction. In the results given later, where a marked turning 
occurs, the error would be much smaller. 
FuLton’s METHOD.—In this connection, the method employed 
by Futron (5) may be considered. It is stated as follows: 
Hyphae within a radius of one opening diameter from the margin of each 
opening were considered, the hyphae within such an area were classed in the 
counts as those turning toward the openings, those turning away from the 
It is possible that the extra number in the C class may be due to some factor 
hitherto overlooked, which causes a slight repulsion from the holes on both sides of 
the plate, even when the conditions seem uniform in the two media. 
