+08 NATURAL SCIENCE [December 
that this hypothesis would gain in probability if it could be 
proved that the protoplasm was withdrawn from the chambers 
when the two halves of the frustule parted from one another 
during cell-division, a circumstance which he regards as by no 
means improbable. Pfitzer does not consider the hypothesis 
summarised above very probable, but believes that the chambers 
with their very thin outer walls would probably relieve the osmotic 
pressure; Lauterborn, however, seems justified in his contention 
that the two functions in no way exclude one another. 
Structure of the Frustule in Surrirella calcarata.—This 
diatom is one of the largest and most beautiful fresh-water forms. 
Examples measuring 0°3 mm. in length and 0°2 mm. in lateral 
breadth were not uncommon, so that, under favourable conditions, 
they could be readily perceived with the naked eye. 
Seen in transverse section (Fig. 4) the outline is roughly 
rectangular, with two sides (‘Gurtelseiten’) flattened, and the 
other two (‘Schalenseiten’) more or less hollowed out on each 
side of a median ridge or keel. At each corner is an outstanding 
process (ala, ‘ Fliigel’). These processes vary in appearance with 
the plane of section (see description of Plate) and three aspects are 
shown in the figure, combining features exhibited by different 
sections. The drawing also illustrates the complicated character 
of the chromatophores, and their relation to the central bridge-like 
mass of protoplasm containing the nucleus. When viewed in its 
entirety, and with one of the lateral surfaces (flattened sides in 
section) turned towards the observer, the diatom presents the 
appearance of a more or less broad wedge, the median portion 
occupied by the lobed chromatophores (superficial focus) and 
flanked on each side by the alae which are now seen to consist 
of a number of parallel transverse canals connected at their 
extremities by a longitudinal canal, and separated from each other 
by tolerably wide interspaces. In a surface view (‘Schalenseite ’) 
the contour of the frustule is almost lanceolate with a median 
longitudinal ridge produced at either end into a spur-like process, 
the anterior one being the largest. 
To return to the alae, Figure 5 shows a portion of one of these 
wing-like processes viewed from the surface. Along the edge, for 
its whole extent, passes a rather narrow canal (‘ Fliigelrandkanal’ or 
‘ Fliigellangskanal’) a connection being established between this and 
the interior of the cell by numerous short transverse canals. 
Internally, these latter commence as tolerably wide tubes of rounded 
section; towards the longitudinal canal (/c.), however, they steadily 
diminish becoming more elliptical in section. The transverse canals 
are separated from one another by U-shaped intermediate pieces, 
formed of apposed portions of the cell-wall, whilst the spots where 
