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



treated by Larmor, it is necessary, in order to show that the 

 moving system which he discusses is the system which the 

 fixed system becomes when set in motion, to restrict 

 the problem to the case in which all the electrons in the 

 system at rest are absolutely devoid of motion. Such a 

 system could not of! its own accord be in equilibrium, without 

 imposed constraints, and the introduction of such constraints 

 of non-electromagnetic origin presents a serious difficulty 

 to the mind, in view of the fact that the whole theory 

 rests on the assumption of the completeness of the 

 electromagnetic scheme. The difficulty of imagining the 

 constraints may to a certain extent be overcome, and this 

 fact, combined with the definite experimental fact to hand 

 in the Michelson and Morley experiment, provides us with 

 the confidence we need to postulate the principle of relativity 

 for uniform rectilinear motion ; but it would seem that in 

 view of the difficulties encountered in this case, the 

 postulation of the principle for non-rectilinear motion can 

 hardly be expected not to lead us into mutually contradictory 

 conclusions. 



XLII. On the Value of the Pitot Constant. By J. D. Fky, 

 Lecturer and Demonstrator in Physics, and A. M. Tyndall, 

 Af.Sc, Lecturer in Physics in the University of Bristol *. 



THE method of measuring the velocity of a stream of gas 

 by means of a Pitot tube is well known. 



A " pitot " is generally a fine tube of which one end is 

 open and directly facing the stream, and the other end 

 is connected to a pressure-gauge. A pressure is thereby set 

 up in the pitot, the value of which is a function of the 

 velocity (r) of the stream at its end. 



If P is the excess pressure over the static pressure at the 

 open end of the pitot (referred to in what follows as the 

 " pitot pressure ") and if p is the density of the gas, then 



P 



K is the Pitot Constant, and previous experiments have 

 shown that its value is not far from unity and constant over 

 a considerable range in velocity. Thus among the more 

 recent determinations of K are those by Threlfall t, who 



* Communicated by the Authors. 



t Threlfall, Proc. Inst. Mech. Eng. 1904. 



