and Gramineae. It is based upon the fact that the histological
differentiation of the epidermis of their root is that generally
characteristic of Monocotyledons, whilst they have two cotyledons--the
old view of the epiblast as a second cotyledon in Gramineae being
adopted. But the presence of a second cotyledon in grasses is
extremely doubtful, and though there may be ground for reconsidering
the position of Nymphaeaceae, their association with the grasses as
a distinct class is not warranted by a comparative examination of the
members of the two orders. Ovular characters determine the grouping in
the Dicotyledons, van Tieghem supporting the view that the integument,
the outer if there be two, is the lamina of a leaf of which the
funicle is the petiole, whilst the nucellus is an outgrowth of
this leaf, and the inner integument, if present, an indusium. The
Insemineae include forms in which the nucellus is not developed, and
therefore there can be no seed. The plants included are, however,
mainly well-established parasites, and the absence of nucellus is only
one of those characters of reduction to which parasites are liable.
Even if we admit van Tieghem's interpretation of the integuments to
be correct, the diagnostic mark of his unitegminous and bitegminous
groups is simply that of the absence or presence of an indusium, not a
character of great value elsewhere, and, as we know, the number of the
ovular coats is inconstant within the same family. At the same time
the groups based upon the integuments are of much the same extent as
the Polypetalae and Gamopetalae of other systems. We do not yet
know the significance of this correlation, which, however, is not an
invariable one, between number of integuments and union of petals.

Within the last few years Prof. John Coulter and Dr. C.J. Chamberlain
of Chicago University have given a valuable general account of the
morphology of Angiosperms as far as concerns the flower, and the