in Rhodymeniales : II. Delesseriaceae. 1 79 
the corresponding pericentral cells on the opposite surface 
are examined, it is found that they also each give rise to 
a carpogonial branch which curves round the pericentral cell 
in the same direction as the carpogonial branch corresponding 
to it above. These relationships will be most readily realized 
by means of figures. Figs. 8 and 10 show the arrangement of 
the carpogonial branches as seen from the surface. Fig. 9 
gives the appearance of the pairs of carpogonial branches 
when viewed from the side. The section is supposed to be 
taken a little to one side of the median line, and the difference 
in the depth of the shading of the alternate pairs is intended 
to indicate the slight difference of level. A single thallus- 
segment thus often gives rise to from thirty to forty procarps, 
the position of all of which can be distinguished in material 
appropriately stained, and swollen in glycerine. No procarp 
ever arises elsewhere than in these situations along the 
midrib. 
It does not follow, however, that all these procarps are 
functional at the same time. They are produced in acropetal 
succession, and those whose trichogynes protrude at any time 
are a few in the apical region on each surface. Further 
back, the midrib becomes stouter by the peripheral growth of 
the sterile filaments derived from the vertical pair of peri- 
central cells. Right and left of the middle line, moreover, 
the lateral pair of pericentral cells and their derivatives give 
off cells to each surface. By this process the midrib soon 
becomes six or eight cells thick, and the procarps, which are 
at the level of the cells nearest to the axis, tend to become 
more and more immersed. A furrow may at first be detected 
on the surface on each side of the midrib, joining the points of 
exit of the trichogynes, where minute pit-like depressions for 
some time remain. It may well be that the convection of 
spermatia to the trichogynes is facilitated by the existence 
of this groove, where they might more readily lodge than on 
the even surface. 
As I have already said, I am inclined to think that some of 
the procarp-bearing segments grow out into the ordinary 
