204 Prof. L. Hill, Messrs. H. M. Vernon, and D. Hargood-Ash. 



It will be noticed that the rate of cooling, both of the dry " kata " and the 

 wet " kata," begins to take an abnormal course at the same point, viz., an air 

 velocity of 1 metre per second. There can be little doubt that the abnor- 

 malities are due to convection currents. Currents of warm air, heated by the 

 cooling " kata," tend to rise vertically from the sides of the bulb, and this air 

 considerably impedes the rate of cooling of the " kata " in still air. It can 

 have little or no effect in the presence of horizontal air currents of fair 

 velocity, as it will be swept aside, but air currents of low velocity are not 

 sufficiently powerful to effect this completely, and in consequence the rate of 

 cooling is retarded. 



Determination of Swirl. 



It has already been mentioned that in the whirling arm experiments 

 account must be taken of the swirl produced ; that is to say, the true velocity 

 of the " kata " relative to the air is less than the actual velocity, because the 

 air is carried round with the revolving arm to a certain extent. A method 

 similar to that described by L. V. King* was adopted to determine the amount 

 of swirl. In the case of the Oxford experiments, a " kata " was fixed at one 

 end of a horizontal arm and it was revolved at the rate of 1, 2, or 3 metres 

 per second, and its bulb passed close to the bulb of another " kata," which was 

 fixed in a stationary position. Cooling observations of the stationary " kata " 

 were made while the moving one revolved at room temperature. When the 

 centres of the bulbs were 2 cm. apart there was only 2 mm. space between 

 the bulbs. 



From the series of data obtained we were able to calculate what the H/# 

 values would have been had it been physically possible for the bulb of the 

 stationary " kata " to occupy the place through which the bulb of the moving 

 " kata " passed in its revolutions. From them it has been possible to calculate 

 the velocity of swirl of air against the stationary " kata " bulb by using an 

 approximate formula already determined from observations taken without 

 allowance being made for swirl. These swirl velocities vary from 6 - 5 to 

 7"96 per cent, of the velocity of the revolving " kata," but they are all subject 

 to a certain deduction. Now the velocity of swirl found by King when using 

 a wire was 6 per cent. ; therefore, assuming that this value holds likewise 

 with the "kata" — the assumption being warranted by the above experi- 

 mental results — then it follows that the values must be reduced by 6 per cent. 

 This correction has been applied to all the Oxford results. 



Similar experiments at Hampstead gave a value for the swirl of about 

 9 per cent., which was deducted from the calculated velocities before 

 recording. 



* L. V. King, ' Phil. Trans.,' A, vol. 214, p. 373 (1914). 



