918 THE AMERICAN NATURALIST. (VoL XXXII. 
Professor Cockerell has raised the question as to whether the 
colony, which at first showed a strong tendency to split-band 
varieties, may not be reverting to the European type, in which 
the split-band varieties are much less common. From informa- 
tion gathered from data furnished by Major Morrison, he esti- 
mates 100 split-band shells in a series of 2200, or about 4.8%. 
In Mrs. Brooke’s list “ D,” the split-band forms are about 4%. 
In my own series the proportion is in “A” 4.4%, in “B” 4.1%, 
in“ C” 3.4%. A large proportion of the variation in the Lex- 
ington colony is along this line. Of varieties previously enu- 
merated, 52 out of 108, or 48%, are split-bands. Of the 277 
new varieties listed in this paper, 112, or 40%, show a split 
band. It is rare to find two split-band shells alike; indeed, 
among the 134 split-band shells in my three series, no less than 
112 different varieties are comprised. It is true that there 
seems to be a slight diminution in the proportion of split-band 
shells in the later series, but the diminution is small, and I do 
not know that Major Morrison’s series was intended to be 
made up impartially. In any case the difference is less than 
the local difference between series “A” and “C.”’ In mylists 
varieties with extra bands (x bands) are not counted as split- 
band. I do not know whether this was the case in Major 
Morrison’s list. 
One further table shows very strikingly the divergence of 
series “A and Z trom “G.” 
TABLE IV.— RELATIVE FREQUENCY OF FUSED BANDS. 
(By percentage.) 
’ B. Cc. D. 
All bands fused into one. . . 0.4 0.6 3-9 o. 
Five banded shells with two or more bands w 18.3 17.1 30.6 _ 179 
All varieties with two or more bands fused. . : 23-3 24.7 43-2 22.5 
From these facts there are certain conclusions which may 
be drawn. While the variation is very “ scattering,” there is a 
predominance of tendency along certain lines. The most fre- 
quent variation is the fusion of bands 4 and 5. Bands 12 
and 123 have considerable tendency to fuse, but band 3 rarely 
fuses with band 4 till all the other bands are fused. 
