530 Pier cy . — The Structure and Mode of Life of a 
doors under a large bell-jar. The dish containing the Hormidium rested on 
an inverted seed-pan having several large holes in its base. To allow the 
passage of air through the apparatus, the seed-pan was supported on two 
bricks, and a glass tube was led into the top of the bell-jar through a rubber 
cork. The tube was bent into two right angles, the free end being long and 
directed downwards. Thus, when the rubber cork had been thoroughly 
covered with vaseline and the bell-jar cemented to the seed-pan, the entry 
of rain was prevented. The bell-jar was weighted down to resist the wind. 
The deposit of dew which might have occurred under these circumstances 
was negligible, since the temperature throughout the experiment was low 
and the diurnal range small (see Table I). 
Fig. 5. Curves to show effect of desiccation. Abscissae indicate time under drought in weeks , 
ordinates indicate percetitages. Curves A 1} a 2 , show variation in the percentage of granular living 
cells to total living cells during Experiments I and II respectively. Curves B l5 b 2 , show variation in 
the percentage of dead cells to total cells during Experiments I and II respectively (cf. pp. 526, 527). 
In Experiment II, begun early in March and continued during the 
following six months, the Hormidium was simply placed in a deep glass bowl 
and protected from rain by a somewhat shallower covering dish, supported 
slightly above the first to allow the entry of air. The average temperature 
during the first two months was a little higher than in the first experiment, 
and increased subsequently ; also, there was a greater diurnal range of tem- 
perature, especially during the last weeks of the experiment (see Table II). 
These conditions of temperature tended to cause a greater deposit of 
dew than occurred during Experiment I, as did also the prevalence of much 
