I14 THE AMERICAN NATURALIST. [Vor. XXXIV. 
of the square of pattern A. This gives us rows which answer 
to the description of Müller quoted above, but this pattern evi- 
dently was not the one ob- 
served by him. This pattern 
was first figured by Beer (98) 
in Blennius, in which the 
large hollow square formed is 
filled by a great number of 
rods. The actual pattern 
found in the eye of Blennius 
was of interest (Fig. 12), 
inasmuch as double series of 
twin cones, one with axes 
horizontal and one with axes 
vertical, were, in several 
places, interpolated without 
disturbing the pattern already 
present. 
Pattern G is formed by 
adding a single cone in the 
center of the hollow square 
of pattern 7. This pattern 
was described by Ryder (95) 
in Salmo. A pattern but a 
little less regular was found 
in some larval Coregonus 
(Figs. 13-17).} 
asmuch as these specimens of Coregonus were of interest in several 
respects, a brief note may be added to the above. The material examined consisted 
of recently hatched larvae preserved in 70 per cent alcohol. The material came 
from the hatching station of the U. S. Fish Cóimniséloti. The heads were 
imbedded in paraffine, sectioned and stained in haemalum-eosine or in Biondi- 
Ehrlich's ‘ren ls mixture. Both methods gave excellent results. The twins 
scopo with the single cones in such a way that each is surrounded by four 
ividing plane points toward the space occupied by the single cone. 
The DESS is not mathematically accurate in the eye. Many of the twin cones 
are twisted, as it were, on their axes, and occasionally an extra single cone is to 
be found besides the single one in the unoccupied space between four twin cones, 
i.e. at the angles of the square. Each component of a twin cone consists of a 
slender outer segment, a much thicker body, and a more or less vacuolated base 
