I50 



NA TURE 



[June 14, 1900 



condensation, whilst many receive no charge of vapour." Instead 

 of being an addition to our difficulties, does not this rather sug- 

 gest a method by which, even if the air entering the base of a 

 cloud be very impure, it may become freed from its dust ? For it 

 follows that even in such air a comparatively small number of 

 drops will be formed in each c.c. when the saturation level is 

 reached. What becomes of the nuclei which do not there form 

 active centres of condensation ? If the presence of a few slightly 

 more efficient nuclei has prevented them from coming into play, 

 the same number of actual drops will be at least equally effective 

 in this respect. Will the dust particles then remain free until 

 they are carried up beyond the reach of the drops, and there 

 become active centres of condensation as Mr. Aitken suggests ? 

 It seems to me that, after a considerable vertical thickness of 

 cloud has accumulated, this is highly improbable ; such a cloud 

 is likely to act as a very efficient air filter. For if even very 

 impure air be kept in a small vessel with wetted walls the dust 

 particles are removed in a comparatively short time — the shorter 

 the smaller the vessel — by coming in contact with the walls. 

 Dust particles in air travelling through a cloud must be very 

 favourably situated for removal by contact with the drops. They 

 are thus not likely to survive aS free nuclei long enough to be 

 able to come into play at the upper surface of the cloud, unless 

 the time taken to traverse the cloud has been comparatively 

 short. A cloud, due to an ascending air current containing near 

 its lower surface as many dust particles (7700 per c.c. ) as that 

 encountered by Mr. Aitken on one occasion on the Rigi Kulm, 

 even if it receive a continuous supply of equally or more impure 

 air from below, may thus have no dust particles left in its upper 

 portions beyond what are contained in the drops ; while the 

 number of drops per c.c. may amount to only a small fraction of 

 the number of dust particles originally present, the size of each 

 being correspondingly greater. 



Mr. Aitken refers to the possible re-evaporation of drops due 

 to the tendency of the larger ones to grow at the expense of the 

 smaller. But all drops which have survived the great tendency 

 to evaporate which accompanies the initial stages of their growth 

 will surely continue to grow so long as the rate of expansion 

 remains the same, or even if it be much reduced. The effect of 

 the size of the drops on the vapour pressure necessary to cause 

 water to condense on them is in fact relatively unimportant 

 except in the case of very small drops ; if we apply Lord 

 Kelvin's formula to the case of drops even as small as io~* 

 cm. in radius we find that the vapour pressure exceeds by 

 only about one part in a thousand that over a flat surface of 

 water ; the evaporation from the drop of one part in 30,000 of its 

 mass would cool it sufficiently to counterbalance this difference. 



With respect to the power of sunshine to manufacture nuclei 

 in air containing various gaseous impurities specified by Mr. 

 Aitken, it may be observed that there is no evidence of such an 

 effect of sunlight in normal atmospheric air, and that all the 

 substances mentioned by Mr. Aitken (ammonia, nitric acid, 

 &c. ) being very soluble in water would be dissolved out of the 

 air in passing through a cloud of water drops. It is true that 

 sunshine does appear to produce in pure air nuclei (which how- 

 ever require a fourfold supersaturation to make water condense 

 on them), and that strong ultra-violet light produces large nuclei 

 like dust particles {Phil. Trans. 192, p. 403) ; but these effects 

 have not, so far as I can see, any immediate bearing on the 

 subject of the possibility of supersaturation in the atmosphere. 



I do not know of any evidence to show whether the small 

 drops in clouds tend to coalesce to form larger ones or not. 

 Such coalescence would tend to hasten the process of separation 

 of dust-free air from the cloud, by increasing the downward 

 velocity of the drops relatively to the air ; but it is unnecessary 

 to assume its occurrence. 



We have now seen reason for believing that the drops in the 

 upper portion of a cloud produced in ascending air are likely, 

 before the air around them has lost any very large proportion of 

 its vapour, to have grown large enough to lag behind the ascend- 

 ing air at quite an appreciable rate ; and that the air between 

 them is likely to be dust-free. Under these conditions a dust- 

 free layer will be formed above the cloud, and will continually 

 increase in vertical thickness. This layer will be saturated with 

 moisture at its lower edge, above this it will be supersaturated ; 

 the amount of supersaturation being greatest near its upper 

 limit, and depending on the vertical distance through which the 

 air has risen since escaping from the cloud. Now to produce in 

 air initially saturated the supersaturation (approximately four- 

 fold) necessary to cause water to condense on negative ions, it is 



NO. 1598, VOL. 62] 



sufficient to let the volume of the air increase adiabatically to 

 I "25 times its initial value (Phil, Trans. A, vol. cxciii. p. 289) ; 

 an expansion which will result from an ascent of the air through 

 a vertical distance of 2500 metres, if we suppose the air on 

 escaping from the cloud to be at a temperature of 10° C. (at lower 

 temperatures a smaller elevation would suffice). Thus, when 

 the air in the uppermost layers of the supersaturated stratum has 

 reached a height of about 2500 metres above the level at which 

 it escaped from the cloud, a sudden change will result ; conden- 

 sation will there take place on the negative ions. The thickness 

 of the supersaturated stratum {i.e. the vertical distance which 

 the upper surface of the cloud has lagged behind the air), when 

 the condensation on the negative ions begins, may vary greatly ; 

 it may be very small if the drops are small and the ascent of the 

 air rapid ; it may amount to nearly the whole 2500 metres in 

 the case where the drops grow large enough to acquire a velocity 

 relative to the air as great as the upward velocity of the air, so 

 that the upper surface of the cloud has ceased to ascend. Above 

 any cloud in an ascending air current, however numerous and 

 small the drops, we should expect to find a supersaturated layer 

 (possibly of very small vertical thickness), provided its upper 

 surface has risen high enough for all dust particles to have either 

 come into play as condensation nuclei, or to have been removed 

 by coming in contact with drops already formed ; provided also- 

 that the heating effect of sunshine on the drops at the upper 

 surface of the cloud is not sufficient to counterbalance the cooling 

 effect of the expansion and cause them to evaporate. And if the 

 ascending current continues till a level about 2500 metres higher 

 is reached, we get condensation taking place in the dust-free 

 layer. It is difficult to avoid connecting this process with the 

 sudden appearance of "false cirrus" at the top of a cumulo- 

 nimbus cloud at the commencement of a shower. 



We must now consider what will happen to the drops con- 

 densing from the supersaturated layer. Mr. Aitken takes the 

 view that if condensation ever did take place on the ions, the 

 drops formed would fall at once as rain, and that a cloud would 

 never result. He remarks that the supersaturated air will be, 

 as it were, in an "explosive" condition, which will cause the 

 extremely rapid growth of any drop that may begin to form,, 

 thus preventing condensation on neighbouring ions. There is,, 

 however, no obvious reason for supposing the rate of increase of 

 size of a drop in supersaturated air to be of a different order 

 from that of the diminution in size of a similar drop in an un- 

 saturated atmosphere. In neither case is there anything of the 

 nature of an explosion. In the one case evaporation causes the 

 lowering of the temperature of the drop below that of the sur- 

 rounding air (to the wet-bulb temperature), the evaporation 

 being thereby retarded ; in the other case, the condensation on 

 the drop at once raises its temperature above that of the sur- 

 rounding supersaturated air, the rate of growth being mainly 

 determined by the rate at which the drop can give out to the 

 surrounding air the heat developed in it by the condensation. I 

 do not think we have the data for determining whether the 

 drops will fall at once as rain or remain in suspension till they 

 have travelled into regions where the ascending current is in- 

 sufficient to support them. In either case, if the drops fall 

 through a supersaturated layer of some thickness, they are likely 

 to reach the ground as negatively charged rain, I see, how- 

 ever, no reason to conclude that negatively charged clouds may 

 not also be produced by condensation on the negative ions. 



The foregoing considerations contain a theory of the origin of 

 rain such as I had in view when the paper, criticised byvMr, 

 Aitken, on the difference between the positive and negative ions 

 as condensation nuclei was written {Phil. Trans. A, vol. cxciii. 

 p. 289). That rain may sometimes at least have its origin in 

 supersaturated portions of the atmosphere has indeed been held 

 by V. Bezokl, Cleveland Abbe, and other meteorolc^ists. 



I do not propose to consider what is likely to happen after the 

 rain has begun to fall. It may be pointed out, however, that we 

 are likely then to have a reduction in the supply of dust particles,, 

 especially if the rain extends over a considerable area ; for the 

 inflowing air is likely to have a considerable proportion of its 

 dust particles carried down by the rain before it has penetrated 

 any great distance into the rain-washed area. In Mr. Aitken's 

 papers may be found references to the apparent dust-removing 

 power of rain. 



Mr. Aitken considers that the positive ions would not remain 

 in the atmosphere, because a slightly greater supersaturation 

 than was necessary to cause condensation in the negative ions 

 would bring them down also. It is conceivable that they may 



