354 Messrs. J. D. Fry and A. M. Tyndall on 



a bubble of water in benzene is used as an indicator, the line 

 of separation between the two liquids being viewed through 

 a microscope and kept coincident with the cross wire in the 

 centre of the field of view by motions of a screw. It is not 

 advisable to allow the indicator surface to be much displaced, 

 because, although it is restored to its original position by 

 tilting, slight changes in surface tension cause it to take a 

 temporary set. Hence it is necessary to insert in the gauge 

 a clamping-tap which is not in general opened until the 

 gauge has been set at the anticipated reading. The 

 particular gauge employed when working well would in- 

 dicate a pressure difference of one ten-thousandth of a 

 millimetre of water. 



The results already summarized in Curves I show that for 

 velocities 600-1400 cms. per sec, K is quite constant and 

 equal to 1*002. At lower velocities there was a one per 

 cent, variation in K which the authors never succeeded in 

 eliminating by modification of screens and so forth. There 

 was, however, always a certain very small amount of 

 pumping action at the oil joint during rotation due to 

 imperfections in the pivot A upon which the rotating system 

 turned, and it is possible that this was responsible for the 

 observed change in K. The pressure effects of such action 

 would be proportional to the first power of the velocity of 

 rotation, and hence compared with pitot readings would only 

 be appreciable at low velocities. It seems highly probable, 

 at any rate, that this change was due to experimental errors, 

 and that one is, therefore,, justified in taking 1*002 as the 

 true value of K. 



Pipe Method. — It is by this method that the recent re- 

 investigation has been made, but except in the case of very 

 small mean velocities the apparatus only differed from that 

 of 1903-05 in a few particulars. 



In this method one is met with the difficulty that the 

 mean pressure of the gas is above or below the atmospheric 

 pressure. This has been met by other experimenters in 

 various ways, the one adopted by the authors being to use 

 a differential pressure-gauge, one limb of which went to the 

 pitot and the other to a tube in the side of the pipe, and in 

 the same plane as the end of the pitot. It was assumed 

 that if this side tube ended flush with the inner surface of 

 the pipe, it would measure the static pressure of the gas in 

 this plane, except when the velocity was considerable or when 

 the pitot and side tube were too near to an end of the pipe. 



In the earlier experiments the pitot and its static pressure 

 tube were placed near the mouth of the pipe, and it was 



