532 Piercy . — The Structure and Mode of Life of a- 
The fact that successive points on the curves are fairly consistent suggests 
that, in the worst case, errors are not likely to exceed those given in the 
above examples. In view of the present lack of information on the subject, 
no attempt has been made to smooth out the curves. 
To return to the desiccation experiments themselves, both afforded 
the same general result. During the first few weeks of desiccation, the 
proportion of living cells containing granules increased till it reached a 
maximum of nearly ioo per cent., when it began to decrease, although this 
decrease was less marked in the second than in the first experiment. 
The proportion of dead cells increased rapidly at first, as though certain of 
the living cells had been surprised by the advent of drought. Then it 
remained roughly constant for a period which coincided approximately with 
the main production of granules, a fact which seems to point to this as the 
means by which the cells were coping with the desiccation. Finally, when 
the production of granules ceased, the proportion of dead increased. 
There were, however, minor differences in the behaviour of the Alga in 
Experiments I and II. Thus, in the first experiment, an actual decrease 
in the proportion of granular cells occurred during the first week, so that it 
became reduced to 15 per cent. ; the subsequent increase to the maximum 
was very rapid, being completed within the three following weeks ; the final 
decrease was considerable. The granules in individual cells never became 
very numerous and were mostly small. The period of low mortality was 
brief, and the death-rate, both before and after, high ; at the end of nearly 
two months not a single cell remained alive. In the second experiment 
there appeared no preliminary decrease of granular cells, and the increase 
to the maximum proportion was only reached after six to nine weeks. 1 
During the latter part of the period of increase the granules were large and 
very abundant in the individual cells ; big, refractive masses were also frequent. 
Only a small decrease in the proportion of granular cells followed, and in 
the succeeding few weeks there was no change. Eventually, during the 
fifteenth to twentieth weeks, the proportion of granular cells increased to 
another maximum, diminishing during the final stage of the experiment. 
The greater abundance and slower accumulation of granules in the second 
experiment are probably to be attributed to the increased temperature and 
sunshine and the more gradual desiccation. The mortality during the first 
five weeks of desiccation was slight and occurred, on the whole, at a slow 
rate during the remaining period of the experiment. 
The accumulation of large granules in individual cells, together with 
the big increase in the proportion of living cells with granules, occurring 
when mortality is low, places beyond doubt the production of granules 
1 The actual readings give two maxima, one after six weeks and another after nine weeks. (See 
Fig* 5> a 2-) 
