410 Marion L. Durbin 
of the curve or region of rapid decline represents a sudden decrease 
in the rate, whereby it is reduced to nearly half its maximum in 
three to five days. The fourth region is one of gradual decline 
representing the time of slow additions to length which terminates, 
in all probability, in the suspension of further regeneration of 
length. The third and fourth regions of the curve are not dis- 
tinctly separate from each other; this is especially evident in the 
fourth experiment in which the transition between the two 
regions is more gradual than in the others because measurements 
were made more frequently. 
It is also evident that tadpoles of this species do not regenerate 
as much tail as has been removed. ‘There is, however, no reason 
to believe that this fact influences the curve of regeneration. 
In Experiment I a series maintained for fifty-one days had ceased 
to show increase in length of regenerated tissue when it had 
replaced 62 per cent of the 15 mm. removed. In Experiment II 
the operated series had nearly completed the process of regenera- 
tion on the thirtieth day after the operation when 56 per cent of 
the amount removed had been replaced. This incomplete regen- 
eration was also observed by Ellis. The irregularities in the rate 
curves may each be associated with some irregularity in the 
external conditions to which the series was subjected. Experi- 
ments I and III afford examples of an irregularity in food supply. 
The tadpoles of Experiment I were starved for four days, during 
which time the regenerating part of the tail shrank .5 mm. The 
irregularity in the food supply during Experiment III is of a 
different nature. The series was fed upon Oedogonium until 
the rate of regeneration was near the end of its interval of rapid 
decrease, then Spirogyra was given for two days. At the end of 
this period it was found that the rate, instead of having continued 
to decrease as it usually does at this time had increased slightly. 
During the following three days no food was given to the series 
and the rate of regeneration decreased to almost zero. In Experi- 
ment II (Table III and Fig. 2) the period between May 11 and ~ 
May 15 included a cold Sunday, during which the laboratory 
was not heated. The rate for this period was much less than 
that for the following period. The irregularities occurring late 
