1885.] On the transpiration-stream, in cut branches. 333 



column of air a few millimeters in length has entered the tube b, 

 the vessel c is replaced ; by this means a bubble of air is enclosed 

 in b ; the bubble travels up the tube and serves to indicate the 

 rapidity of the current of water in the tube b. The time which 

 the bubble takes to traverse a measured length of tube is read by 

 means of a stop-watch. Then by taking the reciprocals of these 

 readings a series of figures proportional to the amounts of water 

 absorbed by the branch in a given time are obtained. Thus if 

 10" is the reading, the reciprocal being 0"1, the rate of absorption 

 is put down as 100 ; five seconds being entered as 200 ; 20" as 50, 

 and so on. The actual quantities of water corresponding to these 

 figures vary according to the size of the tube used. A rate of 100 

 in our experiments means that some quantity between 4 and 8 

 gramms of water are absorbed per hour. 



Details. It will be seen that at each reading a small bubble 

 enters the potometer, these bubbles collect at e, and can be got rid 

 of at intervals by taking out the cork e and filling up the vacant 

 space with water. In some rare cases bubbles have been found to 

 collect in the limb a underneath the cut end of the branch, this is 

 of course a serious error, and to avoid the possibility of its oc- 

 currence, a different form of instrument was occasionally used, in 

 which any bubbles entering from the end of the branch could be 

 collected with the bubbles used as indices of the rate of absorption. 

 Practically however it was found that the form given in fig. 1 did 

 not lead to errors from this cause. 



There is naturally some resistance to the drawing of the bubble 

 of air into the end of b (fig. 1) ; thus it follows that the bubble 

 enters with a jerk, and does not settle down to a steady pace until 

 it has travelled some way up the tube. It is therefore necessary 

 that the mark on the tube forming the lower limit of the fixed 

 distance to be traversed by the travelling air-bubble should be at 

 some little distance from the free end of the tube; in our experi- 

 ments the length of be was usually about 10 c. m. It was found 

 convenient to employ the upper end (f) of the thermometer tube 

 as the upper limit of the measured space. 



In using the apparatus it is of importance that the length 

 of the bubble employed as an index should be constant, since 

 it was found that by using long and short bubbles alternately the 

 readings could be made to vary ; the longer bubbles giving slightly 

 quicker rates than the shorter ones. This point was carefully 

 attended to, each bubble being regulated to a measured length as 

 it entered. With these precautions fairly uniform readings can be 

 taken. The following series is not specially selected, but shows the 

 kind of uniformity easily attained to — and greater uniformity was 

 often reached : 



23—2 



