I0I4- RiDDELSDELL. — The British Forms of Hdosdadium. 97 
tlieir true character. It is quite common for the signs of 
rooting to extend to the end of the branch, though ap- 
parently sometimes with omissions in the middle. But 
sometimes these side stems, even if prostrate on bare ground, 
show no evidence of the rooting chaiacter. Again, on 
different flowering branches of one and the same plant 
the rooting character is strongly present, and nearly 
absent." " At one time, I thought the rooting character 
a means of diagnosing the soil rather than the variety 
of H. nodiflorum ; but that opinion now appears to me 
at variance with actual facts. It seems impossible at 
present to explain the variations of this feature. 
Other characters, e.g., length of peduncle, shape of leaflet, 
are as untrustworthy as that of rooting, for purposes of 
critical diagnosis. The form taken by the species is very 
largely determined by the surrounding vegetation, and by 
the amount of water present. It is a common thing to find 
the same series, or even the same plant, changing suddenly 
from the small- or medium- leafed form, with its compact 
foliage and low growth, as it runs among the open grass by 
a tiny rill, to the large ditch-form, with its tufts of great 
upright leaves, and large coarse stems and branches ; the 
change being due only to the shade of a copse or bramble." 
Again, where large plants of vulgar e have been broken 
off short during the summer, if the w^ater has (as in 19 11 
and 19 13) receded, and left the plants on mud, the later 
growth from the same roots takes the form of a small, com- 
pact plant, rooting at the joints, which would have to be 
placed very close to pscudo-rcpens, H. C. Watson. The 
circumstances have changed, at least in one vital respect, 
and the plant has changed with them. 
From this is derived, with some hesitation, a third 
conclusion, which needs, especially in some directions, 
to be submitted to the test of cultivation. " Some of the 
supposed varieties at any rate are simply transient forms 
due to special circumstances " : though I think there is 
at least one exception. " My behef is that almost 
all the named varieties of H. nodiflorum are simply states 
due to difference of soil, water supply, and cover. They 
