On “ Fascialion.” 
57 
which, along with the consideration of phylogeny, would have 
otherwise solved their problems. For instance (and here I at 
once begin to treat of certain cases of what I regard as normal 
cohesion), the andrcecium of Cyclanthera , one of the Cucurbitaceas, 
consists of a single large organ arising as a central column in the 
flower; the older botanist, wholly ignorant of the proper method 
of procedure for solving such a structure as this, naively described 
it as an axial organ ! But comparison with other members of the 
same order 1 is sufficient to shew that it almost certainly consists of 
a fusion of several (probably 5) stamens which are, of course, foliar 
organs. As no sign of such fusion is seen during the development 
of the organ, it must take place congenitally. What do we mean 
by this expression ? We mean that the ancestors of the plant 
must have possessed five free, distinct stamens which in the course 
of time became fused into a single structure shewing no trace of 
its real composite character, so that to-day the organ arises at 
birth as a single, undivided structure. Yet we may imagine fusion 
occurring within the parent-tissues in this way : that at each of five 
centres cell-divisions took place which, before they culminated in 
the production of organs above the surface, spread on either side 
and towards the centre, so as to form a single large mass of 
dividing tissue which eventually became the seat of formation of 
the columnar androecium, but even this mass of tissue may have 
been formed simultaneously, so as to leave absolutely no trace in 
the ontogeny of the five original stamens. Another quite similar 
case is that of the ovary of the Primulaceae, arising as a homo¬ 
geneous ring-wall with no trace of distinct primordia ; and yet, from 
the comparative method of research, we know perfectly well that it 
is composed of five carpels. 
In most such cases, however, distinct rudiments first make 
their appearance, as in the gamopetalous corolla of Ericaceae, 
indicating the presence of five petals; but congenital fusion of 
these petals early sets in, and the lower part of the corolla 
developes as a tube or cup; yet the polypetalous condition must he 
regarded as primitive. 
It is well-known that the ovule and seed are amongst those 
organs where very considerable adaptations in structure have taken 
place in the course of time; amongst these, as especially well seen 
in Gymnosperms, are fusions between the envelopes, and between 
these and the nucellus enclosed by them. To cite an example: it 
1 There are other genera which exhibit a partial fusion of the 
stamens. 
