Si A. FRANKLIN SHULL 
Of the second factor, little can be said except in a descriptive 
way. In Hydatina it progressively changes so that the propor- 
tion of male-producers decreases with the age of the partheno- 
genetic line: Whether this change is due to continued breeding 
under uniform conditions, or to some other cause, is not known. 
Little more can be said of the variable element, whether separate 
from or only a featureof the progressive one. Fluctuations in the 
‘sexuality’ of daphnians occur, such’ that periods of few sexual 
forms may alternate with periods in which sexual individuals 
are numerous. Woltereck (11) attributes the form of the cycle 
to antagonistic substances, now the one, now the other gaining the 
ascendancy in a rhythmical manner. I have found in Hydatina 
just such fluctuations, which I have not been able to trace to 
any external agent. Nevertheless, it appears that the extent of 
the fluctuations is not independent of external conditions. Thus, 
in my earlier starvation experiments (Shull, ’10, fig. 1), both the 
starved and the well-fed lines show simultaneous fluctuations in 
the same direction, but in every case the wave is more marked 
in the well-fed line than in the starved. Even if the external con- 
ditions (chemical substances in the water, for example) are not the 
cause of this fluctuation, they do modify its amplitude. 
Regarding the first element of the internal nature of Hydatina, 
the genotypic constitution (zygotic constitution of Punnett, ’06), 
we fortunately have more evidence. The crossing experiments 
described in my former paper (Shull, ’11 a), together with the 
results of inbreeding described in this article, enable us at least 
to eliminate certain possible views regarding the internal cause of 
the form of the life cycle. 
The proportion of male-producers can not be dependent on the 
simple quantity of some substance present. For it is difficult to 
see why, in some crosses, the F, line should be intermediate in the 
proportion of male-producers between its parent lines, while in 
other crosses the proportion in F,; should exceed not only that of 
either parent line alone, but of both parent lines combined (op. 
cit., Experiments 36 and 35). 
Among Mendelian explanations, it can not be assumed that the 
life cycle in a given line is dependent on a single gene or a pair of 
