328 Dr. Frederick A. Dixey on 
legitimately employed to designate the adoption of any 
new features borrowed from another species. In so far as 
A copies the appearance of B, it may properly he said to 
“mimic” the latter, whether the object be to suggest 
the presence of a disagreeable flavour which it does not 
really possess, or merely to convey the impression 
that it and its model are alike in all respects. Miil- 
lerian assimilation may be quite as deceptive as Batesian, 
in the sense of leading to confusion between species 
essentially distinct ; and in the case of a “ weak”? species 
being associated with a “strong” one, the departure of 
a form from the typical aspect of its congeners by the 
development of strictly imitative features may be as well 
marked in the one kind of mimicry as in the other. But 
although either of the terms “ Miillerian mimicry” or 
“convergence ”’ would appear to express quite adequately 
the general idea of the mimetic relation between inedible 
Species, a separate term is wanted to designate the 
peculiar give-and-take changes which we have seen are 
theoretically possible to a greater or less extent in every 
case of Millerian association, and which in fact do actually 
occur in several. It is to supply this want of a term that 
I have proposed the expression “ reciprocal mimicry,’’* 
which is meant to convey, besides the general idea of 
convergence, the special information that in the cases to 
which the term is applied, the convergence is brought 
about not by the simple imitation of one form by another, 
but by the interchange of features between forms, and 
their consequent simultaneous approach to an intermediate 
position. 
The foregoing remarks will, I think, have made it 
sufficiently clear, (1) that reciprocal mimicry can only take 
place in Miillerian associations, not in Batesian; and that 
it is therefore, as I have elsewhere said ‘‘ good evidence 
of the distastefulness of all the forms between which it 
can be shown to occur;” (2) that although a mimic 
which is of relatively plentiful occurrence must be Miil- 
lerian, it does not follow that a mimic which is scarce 
must necessarily be Batesian. An inedible mimic may 
be either rare or common ; an edible mimic must be rare. 
Judging by these principles, we must conclude that the 
association of Pieris locusta f$ with Heliconius cydno is 
“ Trans. Ent. Soc. Lond, 1894, p, 298. 
