No. 378.] ARTIFICIALLY HATCHED FISH. 409 
If we now tabulate the dorsal fin-rays of an equal number 
of flatfish from another locality, it is evident that if the fishes 
in both localities are alike, the curves will coincide. If they 
are different, eyen slightly so, the lack of coincidence will 
indicate the difference. The curve drawn at 8, Fig. 1, is 
based on the enumeration of the dorsal fin-rays of 100 flatfish 
taken at Waquoit, Mass., from a small bay only eight miles 
east from Woods Holl. Compared with curve A, the Waquoit 
curve lies further to the left, has a longer base, and a less 
altitude. The Waquoit collection thus contains several fish, 
the number of dorsal fin-rays of which are less in number than 
those of fish taken at Woods Holl. The Wadquoit fish are 
more variable, the amplitude at Woods Holl being from 62 to 
72 (eleven points), while the amplitude at Waquoit is from 60 
to 71 (twelve points). The depressed curve of distribution in 
the second curve is an indication of greater variability and 
general indifference to the “ideal mean.” The arithmetical 
mean, represented by the vertical line, is 65.2, the Waquoit 
fishes averaging about one dorsal fin-ray less than the Woods 
Holl specimens. 
Curve C represents the distribution of 100 flatfish from 
Bristol, R. I., from a body of water located about fifty miles 
west of Woods Holl. Compared with curve 4A, the Bristol 
curve lies further to the left and has a broader base, though its 
culminating point is very definitely indicated. The arithmetical 
mean is 64.9. 
It is thus seen that there isa measurable difference between 
collections of fish from different localities, even though the fish 
individually present no perceptible difference. 
There is correlated with the increase or decrease in the 
number of dorsal fin-rays, an increase and decrease in the num- 
ber of anal fin-rays, as shown in Fig. 2. The Woods Holl 
specimens average a large number of dorsal and also a large. 
number of anal fin-rays, 66.1 dorsals and 49.7 anals. The 
Waquoit specimens average a less number of dorsal fin-rays 
(65.2), and they also have a less number of anal fin-rays (48.6). 
The Bristol specimens average only 64.9 dorsal fin-rays and 
48.7 anal fin-rays. The individuals also partake of this corre- 
