3 
equivalent separation of the drops in another place, and that, therefore, 
whatever increase in the quantity of rain at the lower level may be so 
occasioned in a particular place at a particular time, must be exactly 
counterbalanced by a decrease at other places, or at the same place in the 
long run. No explanation of this kind can account for a phenomenon 
which is constant both in time and place. 
Again, it has been urged that there are certain mechanical causes tend- 
ing to deprive a rain-guage of its proper share of rain— especially eddies of 
wind formed by the gauge itself, and loss by splashing — and that these 
causes, being operative in proportion to the exposure, will affect the upper 
gauge more than the lower. But in a properly constructed gauge the loss 
by splashing must be inappreciable, and with regard to the assumed eddies 
of wind, I believe their effect would be the opposite of that which has been 
supposed — that they would, in fact, tend to concentrate and condense the 
rain in the funnel of the gauge, and so to increase the quantity, as we see to 
happen in eddies of snow and of dust. But in any case the effect due to 
such a cause must be infinitesimal in amount, and quite inadequate as an 
explanation of the large differences observed. 
The greater number of meteorologists have always held that the excess 
of rain collected in the lower of two gauges is due to an actual forma- 
tion of rain in the lowest strata of the atmosphere. Indeed this conclusion 
has been to most minds irresistible, although the mode in which the 
supplementary formation of rain takes place has been a standing puzzle. 
We can comprehend pretty well how rain is formed in the clouds. We 
cannot indeed say certainly what determines the first beginning of a rain- 
drop, nor is it likely that we shall be able, until we know more accurately 
than at present the intimate constitution of a cloud. But we may be 
certain that a drop once formed of sufficient size and weight to fall through 
the cloud, will gather volume as it descends by incorporating with itself 
not only the minute particles of the cloud, but also all the smaller drops 
which it may chance to overtake in its course, for the larger the drop the 
faster it will fall. In this way the large drops that fall in thunderstorms 
are to be explained. The clouds are then of great vertical depth and each 
drop that falls represents the amalgamation of many smaller drops. But 
we cannot in this way account for an increase in the total quantity of rain 
at the lower level. The cloud being the source whence the rain is derived, 
the total quantity falling through a given horizontal area should be the 
same (excluding considerations of evaporation and condensation) at all 
levels between the cloud and the earth. 
