508 BOTANICAL GAZETTE [JUNE 
a may be attached near the fixed end of the spring 5, thus making 
the gap longer than the distance recorded, or a condenser (such as 
is used in a telephone for instance) may be placed around the gap. 
The latter method will also tend to prevent burning of the termi- 
nals. Terminals of platinum and gold give the best results. Even 
with these precautions there will be a limit to the accuracy of a 
- S 
Ss i oe er ES Be aan 
s / Zan 
ee LT N 
Ss ee - 
i Ba : N = 
: PM. : au j 
Fic.-2.—Curve of rate of growth of the hypocotyl of a four-o’clock seedling; 
abscissas indicate time, ordinates growth per hour in microns; total time recorded, 46 
hours; total growth recorded, 1.75 cm. 
Time I 2 3 4 5 6 ” 6.4 O.) tO [aE | 12 
Aen ee 9) 20 bao.) 18 47.40 FS) WH} TE 
1s | 26 |-16 | 32 1 %4 | Is 9 5 6 5 7 6 
Checks per hour 7 6 Gj 5 wi 1 tor to 1 2o; 8 8} 
P $0,483 } 411 4 22": To 2 I 3 2 
7 am oe Pare 
single record, for the two metal surfaces will not always give electrical] 
contact at the same place. But as any experiment includes a great 
number of records, these slight errors will average up so that the 
results will be absolutely accurate. 
A record is given in fig. 2 of a four-o’clock seedling for the first 
46 hours after it appeared above the ground. The making of this 
record put the machine to an extreme test, as these seedlings are 
very delicate. It will be noted that the. curve is very even, and 
that it passes very close to the points from which it is plotted. The 
relation of the rate of growth to light is clearly shown. The 
records were made in January, when the laboratory darkened early. 
The temperature of the laboratory falls considerably at night, 
getting coldest at 4 o’clock in the morning, at which time the day 
fireman starts up the fires. It will be noted that on both nights 
the rate of growth changes sharply at this point. The first inter- 
