10 Scientific Proceedimis, Roijal Duhlin Society. 



soon as possible after their formation, consist, to a large extent, of the slowly- 

 moving type. They are probably, as a whole, not very effective for purposes of 

 recombination. The e.xtent to which their activity affects our values of charge is 

 difficult to estimate. It should not vary much for different pressures of the 

 sprayer, but its effect is possibly less at higher pressures, owing to the more rapid 

 removal of the air. On the whole, we conclude that the electrical recombination 

 should not affect the shape of our curves seriously beyond a possible apparent 

 tilting up of the electrical values at higher pressures (smaller drops). 



We have next to consider the possibility that coalescence takes place to some 

 extent among the drops. This would result in our underestimating the extent of 

 the pulverization undergone by the water. As is well known from the experi- 

 ments of Lord Eayleigh, drops of pure water colliding will rebound if uncharged. 

 A moderate degree of charging will promote coalescence, while a higher degree of 

 charging will prevent it. It might be said, then, that all our measurements are 

 affected by coalescence of the drops ; that the effect is less with the purer water, 

 owing to the higher charges developed ; that the purer water being sprayed into 

 smaller drops (fig. 3) merely means that in that case there has been less coalescence 

 between drops. We are not in the possession of any quantitative data by which 

 this view could be tested, but we are unwilling to believe that coalescence takes 

 place between the drops to any serious extent, for a variety of reasons : (1) Except 

 in the immediate vicinity of the nozzle of the sprayer, there is nothing to "cause 

 collision between drops. For most of their time, between production and capture, 

 they are moving steadily through air that is practically at rest ; and even when 

 they are being driven forward by the air-current from the sprayer there is no 

 appearance of whirling or violent motion which would throw them against one 

 another. (2) When the vertical distance traversed by the drops before capture 

 was varied from 20 to 80 cms. in steps of 10 cms., there was no appreciable change 

 in the size of the drops examined. This experiment shows that the drops are 

 certainly not coalescing in the later stages of their existence. (3) In fig. 3 the 

 first part of the graph (pulverization against pressure) is common to the three 

 samples. But this is just the stage at which the first part of the Eayleigh eff'ect 

 (moderate charges promoting coalescence) should show up distinctly. The charges 

 on the three samples here differ most widely ; for example, on the more impure 

 sa.mple the charge at 5 cm. pressure is practically zero. We should therefore be 

 able to trace the effect of charges increasing from zero and promoting increasing 

 degrees of coalescence, the effect developing more rapidly in the purer samples. 

 But we have no evidence of any action of this kind. (4) If the separation of the 

 later part of the graph (fig. 3) into three parts be taken to indicate the eff'ect of 

 the second part of the Eayleigh phenomenon (increasing charge preventing 

 coalescence), it is difficult to see why the transition should occur at a certain value 

 of the spraying pressure and size of drop, the charges at that stage being in the 

 ratio roughly 1:3:8, 



We believe therefore that recombination of drops, if it occurs at all, does not 

 affect our measurements seriously. The tendency of the three curves in fig. 4 to 

 join into one, which would seem to be very approximately a straight line passing 

 through the true origin, we interpret as indicating that the charging of the water 

 is a genuine surface effect. The charge produced is proportional to the area of 

 new surface for any sample of water if the water is broken up to a sufficient 

 extent. The full electrical separation is inhibited by impurities present if the 

 degree of pulverization is small. There is apparently a critical size of drop at 

 which the inhibiting effect of impurities becomes ineffective. The effect of the 



