130 TRANSPIRATION AND ASCENT OF SAP en. 



each case, was determined by its reaction with ferric 

 chloride applied in a piece of bibulous paper to the upper 

 surface of the cylinder. 



The mean of the entire series gives 6*9 cm. per hour as 

 the velocity at unit head. As all the known errors, such 

 as the introduction of bubbles, clogging and injury of the 

 tracheidal tubes, tend to reduce the result, it is probable 

 that the velocity in the intact tree would be at least 7 to 

 8 cm. per hour under the same pressure. The occasional 

 high results obtained indicate a still higher figure as the 

 probable velocity. 



Table 15. 



In Fig. 20 I have plotted these results. The ordinates 

 represent the lengths traversed in one hour, while the 

 abscissae indicate the pressures, considering a head equal 

 to the length of the branch as unity. 



