Pollen Mother-cells of Lilium canadense . 227 
the contraction of the spirem proceeds, constantly larger aggregates of 
granules result ; and it is not improbable that the ‘ chromomeres ’ observed 
in two different cases may represent very different degrees of aggregation. 
Nor is there any reason to suppose that the smallest visible granules repre- 
sent either the ultimate units (if such there be) or, in different cases, 
homologous combinations of such units. Therefore the notion that a 
longitudinal splitting of the spirem involves a fission of ultimate hereditary 
units, while it is in consonance with observed facts, does not at present rest 
upon any sort of direct proof. 
Outward Manifestations of I dioplasmic Activity. 
In an experimental study of heredity we have to deal, at least in large 
part, with the reappearance or non-reappearance of the offspring of certain 
parental characters. The reappearance of one such character seems in 
many cases to be independent of the presence or absence in the same 
individual of any other character ; in other words, the hereditary endowment 
of an individual is at least in part analyzable into separate qualities which 
seem to act as independent units. For such a unit quality Bateson (’02) 
has proposed the name ‘ allelomorph.’ Whether the whole hereditary 
endowment of an individual is capable of analysis into unit qualities is 
as yet undetermined. Further, there is evidence that in some cases a 
correlation exists between these simple allelomorphs, so that two or more 
of them may form a sort of compound unit ; in such a case, one quality 
always appears accompanied by the correlated quality. For instance, 
Bateson and Saunders (’02) found, in their Matthiola experiments, that 
a correlation always exists between green seeds and purple or claret 
flowers ; and similar instances are common in the experience of plant 
and animal breeders. In other cases, there may perhaps be an apparently 
simple character, which, however, is really compound, and which, although 
normally transmitted as a unit, under certain conditions may split up into 
its components. Bateson and Saunders have shown that the conception of 
such ‘ hypallelomorphs,’ which may break up into simpler units as a result 
of cross-fertilization, offers at least a convenient hypothesis for the explana- 
tion of some apparently aberrant phenomena. 
These conceptions, derived from experimental study, of simple units 
(allelomorphs) and units of a higher order (correlated allelomorphs and 
hypallelomorphs), seem to find an analogy in the structure of the spirem- 
thread, containing, as it does, a series of chromomeres, which in turn 
are composed of smaller bodies. It is impossible to avoid the notion that 
this structure of the hereditary substance in some way corresponds to 
the compound nature noted in the character of an organism. More than 
this it is probably not safe to say at present. The smallest visible chromatin 
