302 
THE PLANT WORLD. 
giving the whole capsule much the aspect of a campanulate corolla. 
These capsules are clavate when closed, as is well known, and the 
separation between the valves extended about a third of the way 
down, or just to the widest part. 
Never having noticed such spreading valves in this genus 
before, my first thought was that I had perhaps discovered a 
new species. It was raining gently at the time, and had been for 
about two hours, but I did not connect this fact with the un¬ 
familiar appearance of the plant until the next morning, when I 
found the capsules all closed on some specimens I brought home. 
On experimenting then I found that they would soon open when 
immersed in water, and close again when dry. The valves seemed 
just as rigid in one condition as in the other. 
This hygroscopic movement is obviously connected with dis¬ 
semination, but just how it is an advantage to the plant to have 
its capsules open only in, wet weather is not clear. Most plants 
with dry fruits which are known to have hygroscopic movements 
act just the reverse of this, discharging their seeds in dry weather. 
The exact manner in which the seeds of KneifEa escape does not 
seem to have been determined, but these problems could doubt¬ 
less be solved by a little careful observation in the field. Most 
plants with erect capsules borneTm stiff erect stems scatter their 
seeds by means of the resilience of the stem when it is agitated by 
