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



by a micrometer head M, and was kept on that diameter by 

 the arm R passing through a guide Gr. 



The same type of pressure-gauge as above was used for 

 low velocities, but for higher velocities a much less sensitive 

 tilting gauge of a simple U type was employed. 



The pipe and its funnel were fitted to the top of a 

 gasometer consisting of three zinc cylinders A, B, and C, each 

 4 ft. 6 ins. in height and arranged as in fig. 2 B, so that C 

 could slide on rollers R^, E 2 , &c, in the annular space 

 between A and B. This space was filled with oil. The 

 diameter of C was 1 ft. 6 ins., those of A and B differing 

 from this by about an inch. 



The gasometer could be raised about 4 feet and then 

 allowed to fall, its velocity being varied by altering the 

 balancing weight W, which was attached to C over the 

 pulleys Pi and P 2 . 



Piston rods immersed in dash-pots D : and D 2 containing 

 treacle tended to render the fall uniform, and the gasometer 

 could also be accelerated or retarded on its downward path 

 by hand application to the rod H. 



When a uniform velocity had been attained, the air was 

 expelled through the experimental pipe at a uniform rate 

 which could be determined from the dimensions of the 

 gasometer and its rate of fall. 



The latter was measured by clipping to the suspending 

 wire a paper strip S which passed under a recording pen. 

 This was made to mark seconds on the paper by an electro- 

 magnetic device connected with a standard clock. 



To eliminate those parts of the fall during which the 

 velocity was not correct, a second recording pen was placed 

 side by side with the first. It was arranged that by com- 

 pleting an electric circuit, the observer could indicate with 

 this, those parts of the fall which were uniform enough to be 

 used in calculation. The procedure was then as follows : — 



The pitot was placed in a known position and the gauge 

 was clamped and set at a reading which by the process of 

 trial and error had been calculated to be right for the 

 velocity required, and for the particular position of the pitot 

 on a diameter. The gasometer having been drawn up w r as 

 set free ; when it had reached a more or less constant 

 velocity, the gauge was opened and the line kept on the 

 cross wire either by accelerating or retarding the gasometer 

 by means of the handle H. 



The experiment was then repeated with the pressure gauge 

 commutated. 



After a little practice it was generally possible to keep the 



