ANNUAL RINGS IN SCALES. 485 
employed, because, after January, when the new apparatus had been 
installed this difficulty was not encountered, the fish in most cases 
appearing in quite normal condition. On account of this the results of 
the experiments between July, 1915, and January, 1916, have not re- 
ceived much attention, and although scale curves were made from each 
fish they have not been reproduced. This is not because they at all 
contradict the results of subsequent experiment, but because in most 
cases the fish had not lived long enough to exhibit much scale growth. 
Of the six fish which did survive through the whole period, only three 
were at all in good condition, the others being much emaciated (see 
Table ITI). 
From January to May the increase in length was in most cases a little 
greater than in the control tank, though the number of sclerites produced 
was practically the same. What is of great interest and importance, 
however, is the difference in the maximum and minimum breadths as 
compared with the controls. In no case was the maximum below 1-2 
and the smallest minimum was 1. Further, the difference in breadth 
between the maxima and minima is small, in only two cases exceeding 0-4. 
This is of course as it should be, as the difference in temperature during 
the period was never greater than 1° C. 
The scale curves, Fig. 6 (p. 479), are interesting in showing great 
similarity ; there is for all a sudden rise from below the 1 unit standard 
to a breadth of from 1 to 1:2. This width is maintained for a number of 
sclerites until a second rise is observed near the end of the period. 
These two rises, I think, synchronise with the increased temperature : 
first from 11-8° C., the normal for January, to 16°9°C.; and secondly, 
between the end of March and beginning of April when the temperature 
was raised 1° C. 
A further point which must not be overlooked is that at this period of 
the year, fish living in untreated water would be forming narrow sclerites, 
and their scale curve would be one showing a minimum, not a maximum. 
This I have shown to be the case with the control fish; a com- 
parison of their scale curves with those of Fig. 6, exhibits this in a 
striking way. 
Of those fish that remained alive until October, the width of the 
sclerite formed at the end of the May period is maintained constant until 
a still greater width is attained at the margin of the scale. This, I believe, 
was due to the water in the tank rising from 17-9° to 19-7° between July 
and August. 
The part of the table dealing with this period of experiment for the 
hot tank is almost a repetition of that for January to May. 
I should like to call attention at this point to the two scales No. 12 
and No. 15. The first of these shows an abnormally high width, while 
