674 HENRY H. DONALDSON 
For the Chicago frogs we had no October observation, but in 
this case we do have one and it is seen that it occupies the position 
which we should expect, and thus supplements and extends the 
Chicago data. 
When first computing the values of C for the Minnesota frogs, 
taken in October, it seemed important to use specimens having 
the same body weight as those used in the preceding series, so the 
first entry for October made in tables 4 and 5 is for the six speci- 
mens having an average body weight of 48.5 gms. This entry is 
designated 3. Later the value of C was determined for the re- 
TABLE 5 
Data on frogs from Minnesota, 1998 
NUMBER OF | 
ae BIRO 
SPECIMENS VALUE OF C 
SERIES a 
Probable 
DATE 
M. | #F Mean error of the Range 
meant 
1908 
ile 2 | 2 | March 26 28.1 +0.30 (27 .0-29 ..4) 
Di 0 > | June 10 Sil i +().65 (28 .8-34.8) 
gS 58 A ee ed 3 3 | October 19 PASS Pe +().66 (24.5-32.6) 
cea ene 3 9 October 19 28.3 +().36 (25 .1-32. 
* Series 3 (six cases) has an average body weight of 48.5 gms. 
+ Series 3’ (twelve cases) has an average body weight of 64.9 gms. 
t See comments on pp. 671-672. 
maining twelve specimens having an average body weight of 64.9 
gms. This latter record is entered in the tables as 3’. The value 
of C is the same in both series. This gives me the opportunity 
to correct a statement previously made, (Donaldson 710, p. 14) 
to the effect that the value of Cis in a measure influenced by the 
absolute size of the frog. This conclusion I now think erroneous. 
It may be added that in the paper just cited the argument is 
not altered by the introduction of the data for the groups of frogs 
there excluded from comparison on account of their body weight. 
The most striking difference between the observations on the 
Minnesota frogs and those on the Chicago frogs is the high value 
