162 
Saunders. — The Leaf -skin Theory of the Stem: 
recognizing that these categories are not sharply divided, for it is evident 
that inheritance of some characters does not take place in equal measure 
in different types. The hairy character affords a case in point. The 
conditions which lead to general hairiness in the vegetative shoot of such 
a form as Matthiola incana prevail throughout this region and throughout 
the plant’s life, and with such uniformity that we can detect no difference 
between individuals. We have here an example of strict or constant 
inheritance. In other instances, such as some of those described above, 
in which the hairiness is localized, the constitution of all the individuals 
of the species may be alike in regard to potentiality for hair production, 
so that if the appropriate conditions are maintained the hairy character is 
always exhibited in an expected manner. But it may frequently happen 
that this is not the case. We have then what we may describe as 
conditional inheritance. The character may not be manifested, but if it is, 
it will appear in due order in relation to the general development, as is well 
seen e. g. in Asclepias linifolia , where the physiological condition existing 
in a lateral axis may be sufficiently different from that in the main axis for 
a conspicuous hair line to be formed in the one case while it is absent in 
the other. But in another type a more general state of plasticity may 
exist, so that a further development of hairs may occur without a definite 
time or place relation. The character then ranks among those which are 
fluctuating or non-inherited. With fuller knowledge classification in this 
manner may become unnecessary, but in the meantime the use of some 
terminology such as that here adopted, which indicates these distinctions, is 
convenient. 
From this point of view anthocyanin production in the vegetative 
shoot may be classed with hair formation. Localization of the colouring 
matter in a particular tissue or layer can be observed in many species. 
Where there is inheritance of the ‘ constant ’ type we should expect this, 
though an overflow beyond these limits might be anticipated under 
favourable conditions. In Lobelia fulgens , where the inheritance appears 
to be ‘ constant ’, the pigment is confined to the epidermis and perhaps 
a cell here and there in the first hypodermal layer. Although present in 
such quantity as to produce a very deep coloration it is absent from the 
unicellular hairs (see foot-note to p. 154). On the other hand, in Lopezia 
coronata , where also the anthocyanin occurs in the epidermis, the hairs are 
coloured as well. In other cases, as in the uniformly coloured stem of 
a red-stemmed Dracaena , the pigment is present in the first hypodermal 
layer and is absent from the epidermis. Though in a particular species the 
potentiality for anthocyanin production may be constantly exhibited in 
a particular cell layer, an alteration in physiological equilibrium arising 
from internal causes may lead to the appearance of the pigment elsewhere. 
Thus in the hypocotyl of Antirrhinum Orontium a slight amount of 
