632 THE AMERICAN NATURALIST. | [Vor. XXXIX. 
on the two sides, we find that the fish in which the greater num- 
ber of gills lies on the left side (the side of the ductus cesophago- 
cutaneus), occur about twice as often as those in which the greater 
number is on the right side. But when the thoracic cavity of 
these particular fish was dissected, I found several cases where 
the number of external openings and the number of gills did not 
correspond! Of the sixty-two individuals, in which the left gill 
slits outnumbered the right, six had the last gill on the left side, 
that which should open into the ductus, entirely wanting, mak- 
ing an even number of gills on each side. Two of these six 
were in the 11-12 group, four in the 12-13 group. In regard 
to the position of the external openings only one was at the 
normal distance from the ductus, in two the distance between 
ductus and external aperture was one half the normal distance 
between these openings, in two the external aperture lay very 
close to the edge of the ductus, and in one it lay in the cephalic 
edge of the ductus. 
In other fish in which the number of gills on each side was 
uneven, where the greater number was on the left side, five had 
the last gill, that which opens into the ductus, rudimentary, vary- 
ing from one fifth to one half the size of the normal gill, and in 
one of these the opening of the ductus was one half the usual 
distance from the external aperture nearest it. 
In all of the fish examined by me that had the greater num- 
ber of gills on the right side, the gills were normal, and nor- 
mally placed. But another curious variant was found in the 
12-11 series that has a marked bearing on the question, whether 
the change in the number of gills in the different varieties is 
due to addition or suppression. This individual had twelve gills 
on the left and eleven on the right side. The twelfth gill was 
of normal size and at a normal distance from its neighbor, but 
the second gill on the left side (counting from before backwards), 
was missing, the external opening leading into a short cul de sac. 
The first gill on the same side was rudimentary. If there were 
1 It should be stated here, that in counting gill slits, 7. z., the external apertures, 
the external opening of the ductus cesophago-cutaneus is always counted as one, 
as the external aperture of the last gill on that side usually lies not on the outer 
surface, but in the ductus just inside the opening. 
