8 THE ROYAL SOCIETY OF CANADA 



Experiments have been performed also with Kata Thermometers 

 directly exposed first in the free air and then in vacuo under circum- 

 stances of controlled convection and radiation. In this way it has 

 been possible to consider separately the effects of natural convection, 

 independent air currents, and radiation. 



The corrections necessary on account of gustiness and radiation 

 for the formula which is applicable under the conditions of wind 

 tunnel experiments have been found to have a value of the same order 

 as that of the discrepancies obtained in comparisons with the ane- 

 mometers. 



The formulation and checking of these corrections over a large 

 range of conditions is at present in progress. 



McGill University. 



