progress, wliicli merely involves tlie separation of the endo-
clirome of the parent frustules into two portions, but does not
iiiohide such a differentiation of these portions as renders them
capable of the coiijiigative act ; the endochrome capable of conjugating
with these segregated portions must be sought for in
other frustules ; hence the process in these genera involves the
presence of two parent frustules, and results in the production
of two sporangia.
Ill the second mode, met with in Ilimantidium, the progress
of separation is arrested at a still earlier stage ; no differentiation
has taken place, and conjugation intervening, necessitates the
union of the entire contents of two parent frustules to form a
single sporangium.
In the third mode, the progress of the separation of the endochrome
ill the parent frustule must be considered as so far
advanced that complete differentiation has taken place. In
every respect but the formation of new valves, self-division has
been completed ; the incomplete frustules are therefore prepared
for conjugation, which, intervening at this stage, leads the
observer to believe that but one frustule has been concerned in
the production of the single sporangium. This we see in Melosira
and the other genera mentioned under this class.
And lastly, self-division occurring during the progress of conjugation,
the endochrome becomes segregated in the very act of
intermingling, and a single frustule, whose contents have been
already differentiated, gives rise to two sporangia, as in Ach-
nanthes and Bhabdonema.
Nor is the self-dividing disposition in all cases permanently
arrested by the complete formation of the sporangium. Having
assumed the form of the parent frustules, with a great increase
in size (the enlargement in dimensions being in some cases due
to the accumulation of the contents of the two conjugating frustules,
and in others, to a rapid assimilation of nutritive material
from the surrounding medium), the sporangial frustule immediately
submits to self-division, and-by the repetition of this
act developes a series of frustules equal in size to the original
product of the conjugating process. This is notably the case in
the filamentous species, as may be easily seen in Melosira,
Plate XLIX. 3 3 9 /', and Plate LI. 331./. sp. ; Orthosira, Plate E.
337. HI. IV. &V.; and i n P l a t e I). 280. IV. How
far this self-division may be carried in the sporangial frustules is
at present unknown ; it is probably of short duration, as we
rarely meet with any considerable number of frustules characterized
by the enlarged size of the sporangial form. In most
cases an arrest of growth, and consequently of self-division,
seems immediately to follow the complete formation of the
sporangia, and the reproductive body assumes the quiescent
character which belongs to the seed of the higher plant, its
vital function remaining dormant until circumstances favour its
further development, and the production of the young frustules,
of which it is the destined parent.
The mode of this development is imperfectly understood, the
dispersion of the sporangial frustules by the dissolution of the
investing mucus usually removing the reproductive bodies from
the sphere of ready observation ; cases, however, have fallen
under my notice, which seem to indicate that the further process
of reproduction consists in the resolution of the contents of
the sporangium into a “ brood ” of Diatoms having the same
form and specific characters as the original frustules which
originated the sporangia. In the gathering of Cocconema Cistula
made in April 1852, which contained numerous instances
of the conjugating process, I observed the frequent occurrence
of cysts enclosing minute bodies, variable in theii- number
and size, and many of which had the outline and markings
of the surrounding forms, and were obviously young frustules
of the Cocconema : these cysts and their contents are figmed
in Plate C. 221. III. IV. & V. It would appear from these
figures, that the production of the young frustules is preceded
by the separation and throwing-off of the siliceous valves of the
sporangium, and the constriction or enlargement of its primordial
utricle, according to the number of young frustules originating
in its protoplasmic contents. In this gathering, forms
of every size intermediate between the minutest frustule in the
cyst and the ordinary frustules engaged in the conjugating
process (Plate C. 231. I.) were easily to .be detected, and the
conclusion was inevitable, that the cysts and their contents were