go6 Brenchley — The Development of the Grain of Barley. 
but then, while A and B keep on rising steadily, C shows a more rapid rise 
until about six days before harvesting, when the ratio remains steady. 
Fig. 3 shows the actual amount of water contained in 1,000 grains. 
There is a steady and very rapid rise for nine to fifteen days, then a fairly 
constant level is maintained for about four periods, followed by a gradual 
fall up to the time of harvesting. Two of the plots show a somewhat inex- 
plicable rise in water content during the last period. 
Nitrogen. The percentage of nitrogen in the dry matter of the grain 
(Fig. 4) falls steadily for four to six periods, after which a rise sets in, this 
increase being most marked in A, with nitrogenous manure. The rise is 
much more gradual in both B and C. 
/o 
2 0 
175 
1-50 
D% V25 
175 !00 
150 75 
125 50 
tOO 25 
A 
B 
C 
D 
(Mean) 
N ' 
1 
'N 
- 
0 3 6 9 12 15 18 £1 24 27 30 33 35 39 42 45 48 Days 
Fig. 4 . Percentage of nitrogen in dry matter of grain. Mean curve is placed three squares 
too low for the sake of clearness. 
The actual nitrogen in 1,000 grains (Fig. 5) rises steadily at first in A 
and B. A reaches its maximum content about twelve days before cutting, 
and then probably remains constant, while B rises steadily almost to the 
end. C, on the other hand, is very different in behaviour. After running 
parallel with the other plots until about the seventh period the curve flattens 
off considerably, and after rising very slowly for a few days more, the amount 
of nitrogen present remains practically constant until harvest. 
The rise in the percentage of nitrogen in the dry matter is somewhat 
unexpected, though some indications of it had been noticed in the earlier 
experiments on wheat. The composite curve D, showing the mean pro- 
portion of nitrogen in the dry matter of the grain for the three plots, indicates 
that this rise is real. There is, of course, the possibility that such an 
increase is merely a seasonal effect, and further investigations will be 
necessary to settle the point definitely. Further corroboration of this result 
