1904] MWEWCOMBE & RHODES: CHEMOTROPISM OF ROOTS 25 
started in the plant house in a full culture solution, except for 
the absence of potassium nitrate. After the plants had produced 
secondary roots, fresh culture fluid of the same composition was 
exchanged for the old, and there was inserted in each jar a glass 
dish of solution of potassium nitrate, whose density was approxi- 
mately the same as that of the other liquid, as determined by 
the means already described. These glass dishes of potassium. 
nitrate, holding 30°, were closed by parchment paper, or by 
hardened filter paper, and the dishes were set with the membrane 
parallel with the main roots and 2 to 3™™ distant from them. 
The preparations were continued in the plant house for six 
weeks, or long after all stored food was exhausted, both fluids 
being renewed every ten days. The roots were shielded from 
the light to avoid heliotropic curving. The species used were 
Raphanus sativus L., Fagopyrum esculentum Moench., Lupinus albus 
L., and Pisum sativum L. There was, however,no chemotropism 
shown. 
To make sure that the failure of potassium nitrate solution to 
act as a stimulus was not due to the relatively small quantity 
employed, the foregoing experiment was repeated in all its 
details, except that the two solutions employed were nearly equal 
in volume. The preparation was continued for three weeks, and 
the solutions were renewed in the middle of the period. The 
time of year was summer, and the growth was good. 
It would seem as though this preparation would bring curves, 
if the seedlings employed were chemotropic. Here wasa row of 
roots parallel with and within 2 or 3™™ of a parchment membrane 
g™ in diameter having different solutions on its opposite sides. 
It was as though the jar had been divided into halves by a parch- 
ment septum, and the roots suspended on one side close to the 
septum. One can hardly think that in nature conditions can be 
more favorable for the manifestation of chemotropism. And 
this preparation did show a certain response. Nine seedlings 
of Lupinus albus and forty-four of Raphanus sativus were used. 
The former showed no curves nor other response ; but the Rapha- 
nus showed a considerable number of primary and secondary 
roots bending toward the potassium nitrate, and a much greater 
