426 W. E. RITTER AND M. EF. JOHNSON 
conjecture be proved true, an exceedingly important biological 
point would have been made. 
And now as to the evidence that a periodicity corresponding 
to the future wheels does exist in the chain before its break-up. 
In discussing the results of our treatment of the data pertaining 
to the unbroken part of the chain, we said the curves, as shown in 
fig. 2, for example, ‘probably’ show a periodicity. We permitted 
ourselves to doubt to this extent, in the interest of conservatism. 
We wish now to sum up the evidence for periodicity. Its strength 
lies in the fact that it is cumulative rather than in the sufficiency 
of any one piece. 
In the first place, does not the undoubted fact of periodicity in 
the wheels themselves, and the groups that immediately precede 
them, make the presence of periodicity in the rest of the unbroken 
part of the chain probable a priort? It would seem so. In the 
second place mathematical treatment of the quantitative data 
makes it almost certain that a periodicity corresponding to 
theory actually does exist. Third and finally the probable exten- 
sion of the periods far back into the young part of the chain, leads 
us to suspect that this fact is connected with another observation 
of quite a different order, an observation, that is, which strongly . 
indicates that the periodicity is really established at least as early 
as the segmentation of the stolon itself. 
One of us has shown that in Salpa fusiformis-runcinata the very 
early segmented part of the stolon may be interrupted by an unseg- 
mented part (Johnson, 710, p. 154 and fig. 8). While such inter- 
ruptions have not been observed in Cyclosalpa affinis attention 
was called, when speaking of the first stages in the segmentation 
of the stolon, to the fact that in some cases the segmentation 
reaches to the very root of the stolon, while in others a stretch of 
unsegmented stolon exists. May not this difference indicate a 
periodicity in the segmentation corresponding to the periodicity 
in growth that we have found? 
The reader may think that the grouping, as shown in the plots 
of differences, is too variable and indefinite to warrant the con- 
clusions we have drawn. True, the groups here are not as regular 
as the wheel graphs shown at the end of the curve (fig. 2), but though 
