for Twenty Years in succession on the same Land. 319 
manures. For, when used alone, the increase so reckoned as due 
to the nitrogen only, includes that due to the associated mineral 
constituents of the rape-cake ; but when used with mineral 
manures, the increase due to the mineral constituents directly 
applied is deducted. On this point it may be mentioned that, 
it the increase were, in all four experiments with rape-cake, 
calculated over the unmanured produce, the result would appear, 
both actually and relatively, more favourable where mineral 
manures were also used, than the figures in the Table show. 
The comparison between the ammonia-salts and the rape-cake 
is, of course, so far as the nitrogen is concerned, the fairest where 
the mineral conditions were the most equally favourable with 
both manures ; that is where superphosphate was used. The less 
favourable result with the rape-cake under these conditions is, 
doubtless, due to its nitrogen becoming less rapidly available 
than that of the ammonia-salts. Still, upon the whole, it would 
appear that not very much more nitrogen is required in rape- 
cake than in ammonia-salts to yield a given amount of immediate 
increase ; and an advantage of the rape-cake is, not only that it 
itself supplies mineral constituents, so that with it less super- 
phosphate, if any, will be required, but that its nitrogen will 
probably be less liable to loss by drainage than that of ammonia- 
salts or nitrate of soda. On the other hand, a given amount of 
nitrogen costs more in rape-cake than in either sulphate of am- 
monia or nitrate of soda. 
The last illustrations relate to the results obtained by farmyard 
manure. As in the case of the rape-cake, the quantity of nitrogen 
applied can only be approximately estimated. In the calculations 
it has been assumed that the dung contained 0'64 per cent, of 
nitrogen = 0 77 per cent, of ammonia. This result is arrived at 
by calculations founded on the average composition of the matters 
supposed to enter into the dung. It agrees almost precisely with 
determinations recently made in dung from the farmyard at 
Rothamsted ; but it is rather less than has been found here in 
good box dung. It is almost exactly the mean of the results of 
Boussingault and Voelcker for fresh dung. But it is considerably 
higher than results recently published by Profess o>r Anderson. 
As has been stated, the produce on the farmyard-manure plot 
has increased considerably in recent years ; and accordingly 
the Table shows much less nitrogen = ammonia required to 
yield a bushel of increase in the later than in the earlier years. 
There has indeed been a great accumulation, the effects of which 
have been only very gradually developed. Taking the average 
of the 20 years, however, it has required 8 lbs. of ammonia, or 
its equivalent of nitrogen, in dung, to yield one bushel increase 
of barley, and its straw ; in other words, nearly four times as 
