On the Bain and Drainage - Waters at Rotliamstcd. 253 
siderable variation. This difference is partly due to a variation 
in the quantity of ammonia present in the atmosphere. Under 
natural conditions the air will be richest in ammonia in summer- 
time, and generally when a warm wind is blowing. In towns, 
however, where much ammonia is produced by the combustion 
of coal, the winter may be the period wlien ammonia is most 
abundant in the atmosphere. At Rothamsted the rain is appa- 
rently richest in ammonia in summer-time. If we take the 
determinations made during the first 12 months at Rothamsted, 
and the subsequent two years' analyses by Way, and separately 
arrange those relating to what for our purpose we may call 
summer and winter periods, we arrive at the following average 
results : 
Table VII. — Nitrogen as Ammonia in the Eain of Summer and 
WiNTKR Periods ; Average of Three Years. 
Rainfall. 
Inches. 
Nitrogen as Ammonia. 
Per Million 
of Rain. 
Pounds per 
Acre. 
Summer (April to September) ., 
Winter (October to March) 
16-203 
12-262 
1-02 
0-85 
3-74 
2-36 
28-465 
0-95 
6-10 
The nitric acid shown by Way's analyses is also slightly 
greater in the summer period, the average of two years giving 
for the summer rain 0-122, and for the winter rain 0*109 of 
nitrogen per million. 
Another condition which has a still greater influence on the 
proportion of ammonia in rain is the amount and distribution 
of the rainfall. A heavy rainfall descending in a short time 
will always be poorer in ammonia than the rain of light showers 
distributed over a considerable period, the former rain-water 
having come in contact with a relatively smaller volume of air 
than the latter. Moreover, in a storm, or a consecutive rain- 
fall, the latter part of the rain passes through an atmosphere 
already well washed, while in light showers the atmosphere is 
more or less renewed between each rainfall. The influence of 
the quantity of the rain on the proportion of ammonia it con- 
tains will be plainly seen if we arrange the 39 monthly analyses 
of rain-water already given according to the amount of rainfall 
in each month, as is done in the following Table : 
