RELATIVES OF THE HERB ROBERT 185 
necessary conditions, it would only be a matter of 
time. 
It would also be desirable, in view of the occasional 
self-pollination of the type, to discover whether, as 
seems possible or even likely, it is resorted to more 
regularly in the shingle variety, owing to the com- 
parative scarcity of insects. Lastly, there was a rather 
curious difference in two other points that came under 
my notice. Firstly, the interval between the pollin- 
ation and the maturity of the seeds of my 1917 plants 
was only seventeen days instead of twenty-one; and 
secondly, the cotyledons from some of the seeds that 
I sowed in 1917 appeared above ground in eleven days, 
that is to say, ten days earlier than for the type. 
These seeds were sown in a large pot of soil covered 
over with small stones, and I watered them in with 
a syringe as being the best substitute for driving 
rain : it occurred to me while doing so that the absence 
of threads made it easier for them to be washed into 
the shingle on the seashore by a less violent downpour. 
Now, it is easy enough, and probably a great deal too 
easy, to make shrewd guesses, but guessing is not a 
very profitable way of studying Nature, and although 
one can understand that the long threads or an 
abundance of glandular hairs might not be of much 
use where not only insects but also birds and other 
animals are comparatively scarce, one must not 
conclude that the plant suppresses them for this or 
any other reason: it is just as likely that on the 
hot dry shingle it could not produce them if it wanted 
to. But, however that may be, it is good to note the 
differences even if we must leave their explanation 
to more experienced observers or perchance to a future 
generation. 
