sition; but this, too,
in a sufficiently large group, would necessarily have a statistical
constancy. As a matter of observation, the resemblance between
persons of the same "genus" (in the sense of "generic," as already
explained) is sufficiently great to admit of making good pictorial
composites out of even small groups, as has been abundantly shown.

Composite pictures, are, however, much more than averages; they are
rather the equivalents of those large statistical tables whose totals,
divided by the number of cases, and entered in the bottom line, are
the averages. They are real generalisations, because they include
the whole of the material under consideration. The blur of their
outlines, which is never great in truly generic composites, except
in unimportant details, measures the tendency of individuals to
deviate from the central type. My argument is, that the generic
images that arise before the mind's eye, and the general impressions
which are faint and faulty editions of them, are the analogues of
these composite pictures which we have the advantage of examining at
leisure, and whose peculiarities and character we can investigate,
and from which we may draw conclusions that shall throw much light
on the nature of certain mental processes which are too mobile and
evanescent to be directly dealt with.

III. COMPOSITE PORTRAITURE.

[_Read before the Photographic Society, 24th June, 1881_.]

I propose to draw attention to-night to the results of recent
experiments and considerable improvements in a process of which I
published the principles three years ago, and which I have
subsequently exhibited more than once.

I have shown that, if we have the portraits of two or more different
persons, taken in the same aspect and under the same conditions of
light and shade, and that if we put them into different optical
lanterns converging on the same screen and carefully adjust
them--first, so as to bring them to the same scale, and, secondly,
so as to superpose them as accurately as the conditions admit--then
the different faces will blend surprisingly well into a single
countenance. If they are not very dissimilar, the blended result
will always have a curious air of individuality, and will be
unexpectedly well defined; it will exactly resemble none of its
components, but it will have a sort of family likeness to all of them,
and it will be an ideal and an averaged portrait. I have also shown
that the image o