J. AITKEN ON THE NUMBER OF DUST PARTICLES IN THE ATMOSPHERE. 17 



checks any shock being communicated from the pump to the air in the receiver. In the 

 Plate it will be noticed that there is an enlargement P at the top of the exit tube C. 

 The interior of P is packed with cotton wool, and a cover is put over all to keep off 

 the showers of fine rain. This arrangement has been found to work perfectly, and it has 

 in a simple manner removed one of our principal difficulties, keeping the air free from 

 spontaneous condensation with even higher expansions than results from one stroke of 

 the pump. 



It may be asked, Why was the use of such high degrees of expansion persisted in, 

 when it was accompanied by so many difficulties ? One reason was the very existence of 

 these difficulties themselves. They had to be satisfactorily explained before any con- 

 fidence could be placed in the method of working. If we had confined our attention to 

 lower expansions, we could not, for instance, have been certain that the finest particles 

 were thrown down and all the dust counted. But another reason was that the high 

 degree of expansion makes the drops much larger ; they therefore settle more quickly, 

 and are more easily counted. 



Passing on to another necessary precaution. In making these tests, as has been 

 already explained, it is necessary from time to time to turn the receiver upside down, and 

 move it about in such a manner as to cause the water in it to come into contact 

 with all parts of the interior. When using the 12-inch filter it was noticed that, after 

 doing this, rainy condensation frequently took place on making the first expansion. It 

 was thought this time that the failure might be due to some straining of the joints from 

 the twisting of the tubes when moving the receiver, or that it might be due to the 

 saturation being more perfect after the walls were newly wetted, and the air consequently 

 more highly supersaturated on expansion, and therefore in a more favourable condition 

 for producing condensation. This class of failures was, however, traced to the manner 

 of wetting the inside of the flask, the particles being produced by the splashing of the 

 water. If the wetting was roughly done, and there was much splashing, a considerable 

 number of particles were manufactured, and a shower always took place on expansion. 

 But if the water was quietly moved round inside the receiver, no condensation appeared. 



On another occasion copious showers made their appearance even when using every 

 precaution, and the showers persisted, so that it looked as if the apparatus had broken 

 down somewhere. At last the source of these particles was traced to a drop of water, 

 which had got into the inlet pipe, and wet the lips of the stopcock. The air rushing 

 over this wet surface, on its way to the receiver, had torn up the water into fine spray, 

 which supplied active centres of condensation. A similar disturbance was also produced 

 when the air was allowed to rush too quickly out of the gasometer into the receiver. The 

 rush of air over the wet lip of the pipe, where it enters the gasometer, manufactured 

 particles enough to give a good shower. 



There are some other precautions which require attention in working the apparatus 

 but, as there are some developments of the apparatus at present under consideration, in 

 which it is hoped these will be avoided, reference need not be made to tbem here. 



VOL. XXXV. PART 1. C 



