220 
A similar temperature change, brought on gradually (three hours) and 
continued for three hours, only produced about a 40 per cent. fatality 
among the composites. 
Changes in moisture, either in the soil or in the air, found a ready 
response in the behavior of the composite seedlings. A diminution of 
moisture, which would not even produce wilting in the seedlings of other 
families, would not infrequently prove fatal to those of the composites. 
The effect of moisture increase was not so readily seen, although there 
exists under such conditions a tendency on the part of the stem of the 
seedling to rot near the ground, a tendency apparently shared by Abutilon, 
and which seems about the only check put upon the increase of this 
latter form. 
As might be expected direct sunlight works against the seedlings of 
composites, as, indeed, against all others, but with a peculiarly fatal effect 
in these sensitive forms. It has been found necessary in all cases to pro- 
tect them from the direct sunlight for several days, usually until after the 
development of the first foliage leaves. Such extreme sensitiveness was 
rare, if at all present in the other families studied. 
Freru.— Cotyledon leaves of nearly related forms closely resemble each other, a 
resemblance often carried on in the earlier true leaves. 
The cotyledon leaves of Arctium, Cnicus lanceolatus and Cnicus mu- 
ticus are almost exactly alike. The only dissimilarity observable being 
that in Arctium the green is a trifle darker than in Cnicus. The re- 
semblance is so exact that I discarded the first germination experiments 
with Cnicus, supposing that through inadvertence Arctium achenes had 
been planted. With the appearance of the first true leaves Arctium is 
plainly marked off from Cnicus, but the two species of Cnicus are not to 
be separated until the appearance of at least the third foliage leaf. By 
this statement I mean that I think that at this point I can detect the be- 
ginnings of the leaf characters of the forms. The same conditions are 
found. in the Lactucas. Scariola cannot be separated from Canadensis 
in any of the seedling stages so far as my observations go, indeed, though 
I carried the seedlings through the development of the seventh foliage 
leaf I found no marked indication of specific foliar differences. I think, 
without multiplying instances, that in very many cases supposed rela- 
tionships between the species of a large genus, and certainly between 
many genera, might find corroborative evidence in a study of the early 
foliage of the seedlings. 
