Bd. V: 6) 
THE FISHES OF THE SWEDISH SOUTH POLAR EXPEDITION. 
57 
lata is probably similar, but I have also in the ventricle of this species several times 
found Polychceta. In N. coriiceps I have found mostly Crustacea, even small crabs. 
N. macroceph. marmorata appears to feed to great extent on sea-snails. Especially 
the large and middle-sized specimens contained in their stomachs the remains of 
what seemed to have been opistho-branchiate molluses of great size, often so big 
that one specimen completely filled the whole stomach. That fishes like this one 
and the foregoing, which live among the vegetation happen to swallow pieces of 
the same along with their prey which crawled on and among the algæ can be easily 
understood. I therefore regard the algæ found in the stomachs of these fishes as 
accidentally engorged. That so really is the case is proved by the fact that the 
pieces of algæ found in the posterior part of the intestine are just as little digested 
as those I found in the ventricle. 
In the ventricle of Trematomus hansoni georgianus I have not found any re- 
mains of food material that were recognizable. The specimens of Champsocephalus 
gunnari examined had also empty ventricles, but there cannot prevail any doubt 
that it preys on fish. A similar conclusion may also be drawn concerning Para- 
chœnichthys georgianus , but in the latter case it has been distinctly proved by the 
unmistakable remains of two fishes in the stomach of one specimen of this kind. 
The greatest number of pyloric caeca has been found in Trematomus hansoni 
georgianus. In that species they are 7, rather large and long. The same number 
is also found in some specimens of Notothenta coriiceps divided into two groups, 
one with 3, the other with 4, in other specimens of the same species and from the 
same locality, however, either group contains only 3, and the whole number is thus 
only 6. In N. brevipes I have counted 6 pyloric appendages, in N. sima 6 or 5, 
in N. tessellata 5. In N mizops nudifrons there are 6 such appendages, but some 
of the middle ones are so reduced in size so that it is very probable that at least 
one of them in certain instances wholly disappear. N. macrocephala marmorata 
has 5 large pyloric appendages. N. gibberifrons has 4 large and wide pyloric cæca. 
Champsocephalus gunnari is provided with only 3 pyloric appendages which, how- 
ever, are quite large, especially one of them (see fig. 4). The same number is also 
found in Parachæniclithys and Harpagifer. It may be concluded from this that 
the more specialised members of the family have been subjected to a reduction 
with regard to the number of their pyloric appendages. 
The arrangement of the intestine is practically the same in all Nototheniidæ. 
In Trematomus hansoni georgianus the intestine is comparatively long and runs 
first from the pyloric region forwards. From a place just behind the small right 
lobe of the liver it then descends along thn right body wall to the posterior end 
of the abdominal cavity from where it again curves forward to the pyloric region 
and then descends to the anus. The whole course of the intestine thus forms a 
