18 The Effect of Climate and Weather on the Soil. 
A hot, dry summer, as we have seen, is favourable to the 
accumulation of nitrates and an improvement in the soil 
generally, and when this is followed by a dry wintei the 
fertility thus gained is preserved by the cold storage so that 
the crop starts under very favourable circumstances. When 
the hot dry summer is followed by a mild wet winter much 
less favourable results are obtained because of the serious 
losses in winter, indeed the crop may be worse off than when 
the summer has been cooler because in that case less of the 
organic reserves of the soil have been changed to nitrate. Oui 
investigations show that a hot dry summer followed by a dry 
winter is on the whole very beneficial to the soil, while a 
hot dry summer followed by a mild wet winter is likely to be 
harmful. 
Two of the Broadbalk wheat plots illustrate this point very 
clearly. Plot 2b receives farmyard manure every year and the 
plant is dependent for its food supply on the decomposition 
process brought about by the soil bacteria and is therefore 
much affected by the seasonal factors just discussed. Plot 16 
receives a complete artificial manure containing more than 
enough nitrate of soda to yield a full crop and is therefore 
much more independent of the character of the preceding 
season. Over a long period the two plots give practically the 
same yield, for example, from 1874 to 1912 the average crop 
was 34-2 bushels when dung was applied, and 33T bushels 
where artificials were used, or, taking the total produce, 
6,374 and 6,540 lb. per acre respectively. But in the 
individual years the crops were not equal. The dunged plot 
came out much the best when the preceding summer (i.e., the 
summer before sowing) had been dry and the winter dry also, 
it gave nearly 6^ bushels more grain and over half a ton more 
total produce. But when the dry summer was followed by a 
wet winter the dunged plot suffered, and lost over 2 bushels of 
grain and nearly half a ton of total produce in comparison with 
the plot receiving artificials. 
The actual yields are given in the table on the opposite page. 
Of course we cannot attribute the whole of these effects to 
the character of the preceding season because the season of 
growth obviously plays a part, but the averages for the whole 
period show that the latter effects are largely smoothed out 
over the number of years. 
It is evident from all this that when the farmer begins his 
year in October he does not start with a clean sheet so far as 
season is concerned, for his soil is much influenced by the 
character of the preceding summer. To a still greater extent 
is it affected in the spring by the character of the winter 
through which it has passed. The soil is very much the result 
