540 Lloyd Williams. — Slit dies in the Dictyotaceae. 
number of cases makes it certain that the great majority mature very soon 
after the others. In some cases, particularly when the conditions are 
unfavourable, they are arrested shortly before maturity, degenerate in situ, 
and very soon fall off. In very exceptional cases, more especially when 
very late in starting, they may remain on the plant till the succeeding 
crop. It is evident that in the first case there must be considerable 
acceleration as compared with the crop-optimum. This is probably due 
to the greater vigour of the part. Now, if younger sori can be started 
in the manner just described, at the ‘wrong period/ so to speak, why could 
similar ones not be interpolated among the older sori on the other parts 
of the thallus ? The probable answer is that the formative reproductive 
material or energy is diverted into the adjacent sori, which being already 
started form centres of reproductive metabolism for the neighbouring cells. 
When a plant gets old it frequently happens that the distal sori, instead 
of being later, actually liberate before the others. The reason for this 
would probably be found in the senility of the older parts and the 
greater metabolic activity of the distal portions, which have not as yet 
produced many crops. 
io. In most cases the crop-band, if accurately plotted, would show 
a greater width at the initiation end than at the other ; hence there must 
be acceleration of the later among the ordinary sori. (Here the exceptional 
distal sori discussed in paragraph 9 are excluded.) This is clearly shown in 
Diagram I, where the difference amounts to as many as four tides. 
In Diagrams III and IV the optima of the two sets of crops above 
discussed are compared. The interval between the highest spring tide and 
the discharge of the gametes in the two cases is well shown, and the agree- 
ment between the crops in the respective groups is most striking. The 
diagrams also show the variation in length of the optima, but it does not 
indicate the corresponding variation in the length of the tidal periods. 
In the above discussion the crops of October and early November are 
left out of consideration. In view of the regularity of the August and 
September crops those of the later months disclose a very remarkable 
retardation, which progressively increases until initiation coincides with the 
rising spring tides, and liberation is effected at, or just before, the lowest 
neaps, thus almost reversing their usual tidal relations. It would only be 
natural to expect a retardation towards the close of the season when so many 
unfavourable conditions supervene ; the remarkable thing, however, is that 
the optimum is not lengthened by a single tide. This is well shown in Table 
III, where the percentages in column 5 are 76 and 74. Here, then, although 
there is great retardation, the periodicity is as clearly marked as ever. 
In the second paper of this series I suggested that it would be 
interesting to find out whether the periodicity, which up to that time 
