Dixon — Note on the Tensile Strength of Water. 65 



Fifteen minutes was occupied in raising the water of the beaker through 

 the next degree, so that the water in the capillary tube must have very 

 closely approximated to the temperature indicated by the thermometer in 

 the beaker. By proceeding in this way every effort was made to avoid 

 exaggerating the temperature at which the tube filled. The large amount 

 of water in the beaker secured that the cooling should be extremely slow 

 before the reappearance of the bubble, so that it is improbable that the 

 thermometer gave readings sensibly different from the temperature of the 

 tube. 



In spite of these precautions, differences are observed in the successive 

 experiments with the same tube. These can scarcely be due to errors of obser- 

 vation. It generally happened that when a tube was heated on successive 

 days, lower readings were obtained for the " full " temperatures in the 

 later observations. See experiments 4, 5, and 6 on tube 11. Occasionally, 

 however, the " full" temperature rises after a time, viz., experiments 1, 2, 

 and 3 on tube I. 



As a general rule, the temperature of rupture lowers with time, suggest- 

 ing that adhesion is improving. It may be noted that some very high 

 values of adhesion of water to copper were accidentally obtained in tube II., 

 in which by chance some minute shavings of copper were included. 



These experiments amply confirm Berthelot's observations of the tensile 

 strength of water containing air, and they raise the minor limit obtained by 

 him of its cohesion and its adhesion to glass from 50 atm. to over 150 atm. 

 They furtiier show that the adhesion to the walls of the conducting tracts is 

 also over this figure. It is also evident that such a tensile strength as is 

 thus ascertained for water is more tlian sufiicient to withstand the stresses 

 developed in lifting the transpiration stream in plants, and in overcomino- 

 the resistance of the conducting tracts. 



