COXJUGATIOX IX PARAMECIUM 
293 
Within a very short period after conjugation it is usually pos- 
sible to induce a new conjugation b}' again adding hay. It is 
probable however that time must be given for loss of the plump- 
ness and well-fed condition due to the previous feeding. Appar- 
ently three steps are required: (1) a thin, ill-fed condition, with 
httle multipHcation, followed by (1) abundant nutrition, causing 
plumpness, and rapid multiplication, and this followed by (3) 
a falling off in the nutrition, plumpness and multiplication. It 
is at the very beginning o^ this third period, when its effects are as 
yet hardly noticeable, that conjugation occurs. The interval 
between two successive conjugations appears to depend merely 
on the length of time required for inducing these stages in succes- 
sion. As we have seen, intervals were observed in the case of k, 
varying from five days to two weeks; to a month or more. 
CONDITIONS FAVORABLE TO CONJUGATION IN OTHER RACES 
In the other races which I kept in the laboratory, epidemics 
of conjugation usually occurred under conditions similar to those 
observed for the strain k, but in most other races the epidemics 
occurred much more rarely, by no means taking place regularly 
whenever the conditions described were supplied. The differ- 
ence between the strain A- and other strains in this respect is well 
brought out in certain experiments that were planned for deter- 
mining what external conditions would induce conjugation. 
Thus, in the experiments with kb just described, T carried out at 
the same time a parallel series of experiments, identical in every 
detail, with individuals of the race D. This race had not con- 
jugated for a long period, while kb had conjugated within a month. 
As we saw above, in kb conjugation occurred with absolute reg- 
ularity as soon as the required fall in nutritive conditions was pro- 
duced. Yet under the same treatment there was no conjugation 
in D at an}- time. 
Again on June 9, 1909, 1 placed in five drops of a fresh rich stand- 
ard infusion two individuals each of ten different strains, desig- 
nated 82a, 21a, 21b, 16a, 16d. kb, c, i, L2, D. The first five strains 
had been isolated from a wild conjugating culture about a month 
