366 Gates. 
number of cases in various plants and animals. Only by breeding 
can it be determined whether a character will regress or remain true. 
There is no other criterion. DE VRIES formerly made the distinction 
that (non-heritable) fluctuations are only plus or minus variations, 
while mutations are, on the contrary, in all directions. But it is now 
evident that this distinction will not hold, and that whether variations 
are germinal or non-germinal, can only be decided by observing their 
offspring. Variations which can be most easily interpreted as due to 
some purely quantitative change, such as the present case of O. rubri- 
calyx, breed true, just as do more complicated mutants in which there 
has been an alteration in many characters. In how far the time-worn 
distinction between quantitative and qualitative changes is a valid 
one when analyzed, it is at present impossible to say. 
It might be considered, on account of the definite localization of 
the increased color pattern of O. rubricalyx in the hypanthium and 
the under surface of the rosette leaves, that there was here a morpho- 
logical as well as a physiological change. Yet on analysis I think the 
character change will be found to be fundamentally and purely a 
physiological one. The localization of the color pattern will be found 
to be concerned with the physiology of ontogeny. Thus it appears 
that every part of the plant has an increased capacity for anthocyan 
production. The particularly conspicuous accumulation of anthocyan 
in such regions as the hypanthium and the ventral surface of the 
rosette leaves may be explained as due either to a greater abundance, 
in these regions, of the materials for anthocyan production, or to 
its accumulation and storage in these organs after being produced 
elsewhere in the plant. The latter explanation might apply to its 
occurence in quantity on the ventral surface of the rosette petioles, 
because we know that this region is largely shielded from light although 
light is necessary for the production of anthocyan in the plant. 
I have considered certain features of anthocyan production in a 
previous paper already mentioned (GATES, 1910). The chemistry of 
the subject is not yet sufficiently far advanced to decide between 
the various more or less divergent views which have been proposed 
by E. OVERTON (1899), PALLADIN (1908), COMBES (1909) and Miss 
WHELDALE (1909, 1910), to mention only a few of the workers in 
this field. But certain conclusions of general interest may already 
be drawn from the behavior of the mutant O. rubricalyx. In the 
first place, the obvious interpretation, that O. rubricalyx is due to a 
germinal transformation which is fundamentally of a positive nature, 
