50 ROBERT STANLEY McEWEN 
1916. Finally Iam under great obligation to Dr. W. L. Severing- 
haus and Mr. E. C. Unnewehr, who, under the direction of Dr. 
Trowbridge determined for me the wave lengths and energy trans- 
mitted by my colored filters. This work was done in the Ernest 
Kempton Adams Precision Laboratory of Columbia University. 
METHOD AND APPARATUS 
A variety of methods were employed in the course of the work, 
and it therefore seems unprofitable at this point to do more than 
give a general description of the apparatus and the chief systems 
used in making the tests. 
A glass tube 26 mm. in internal diameter and 91.4 em. long 
contained the flies. This tube was graduated in inches because 
at the beginning this seemed to be about the smallest unit in 
which it was possible to observe rapidly moving flies with any 
degree of accuracy. One end of the tube was closed by a piece 
of isinglass, while the other was stopped with a cork covered 
with dead-black paper. The flies were introduced into the tube 
by means of a cardboard funnel from small vials in which they 
were kept. The tube was then placed with the open end toward 
a window in a room which was otherwise darkened. Since it 
was early discovered in accordance with the results of Carpenter 
(American Naturalist, 05) that the insects were much more 
phototropic when mechanically stimulated, the tube was held in 
the hands during the earlier trials and gently agitated. 
After a short series of experiments under these conditions it 
became evident that a more refined system must be used in order 
to get results in any way comparable or consistent. The two 
chief factors which demanded stancardization were the means of 
mechanical agitation and the source of light. The first problem 
was solved as follows. The tube was fastened in a horizontal 
position on a board by means of rubber bands and nails padded 
by thick felt. The four nails were placed in pairs, each pair 
about 22 em. from the end of the tube, in such a way that the 
latter could vibrate between the nails in a direction at right an- 
gles to its length and through a distance of about 5mm. A pen- 
dulum was now constructed from a piece of wood a meter in 
