644 The American Naturalist. [August, E 4 
Other quarter bars are then added in front and behind this 
region. 
From the conditions represented in fig. 3, we have two P 
diverging lines of development. The one line was discussed 
in the preceding paragraph. The other line is found in speci- 
mens, Nos. 82 and 83, taken from San Marcos Spring, Texas, 
and is represented in fig. 5. We have here a splitting of the 
bars without the regular result seen in fig. 4. The lower ends 
of the whole bars have not split, in fact, they have increased . 
in width, and form a very prominent series of spots along the 
side. It will be seen that the bars radiating from the eye have 
become much more pronounced. 
'The pattern of fig. 6 can be easily derived from the preced- 
ing one by assuming that the lower half of the whole bars of 
the anterior part of the body have shifted their position back- 
ward, so that they no longer extend extirely to the mid-dorsal 
line. The 3d,4th and 5th whole bars show different degrees 1 
of shifting. The lower part of the 4th has shifted, but still * 
retains its connection with the upper part. In the 3d, the bar 
is more nearly separated, while in the 5th the separation is 
complete, and the original lower part of the bar becomes 
simply a vertically elongated spot. The bars around the eye 
are here again less developed. "The pattern of fig. 6 is the one 
occurring in Etheostoma caprodes manitou Jordan, and was 
drawn from a specimen taken from Torch Lake, Mich. Other 
specimens, taken from the same lake and from other localities, 
have the same color pattern with slight variations. Nos. 1-7, 
and 41 of Table I, are this variety. 
The line of development taken up by fig. 5 is continued in 
figs. 7 and 8, representing the specimens from Obey’s River 
and Eagle Creek in Tennessee, and from the Little South Fork 
of the Cumberland River in Kentucky. These are Nos. 56-72 
in the table. A single young specimen, No. 74, which prom- 
ised to become this form, was also taken in the North Fork of 
the Holston River, in Virginia. The two figures were drawn 
from a younger and older specimen respectively, of the same 
form. In the younger specimens, the bars have become more 
split up, and have increased in irregularity. Almost all of 
