70 
BULLETIN OP THE UNITED STATES FISH COMMISSION. 
whicli a nucleolus was differentiated by the staining fluid. Nuclei of the same nature, 
especially near the anterior end, were also seen sparsely scattered through the spongy 
tissue. In a few cases some irregular patches of granular protoplasm were observable 
among the spongy connective tissue. No branches were certainly made out leading 
either to or from the marginal canals. The marginal canals are throughout mirch 
larger'than the aquiferous vessels. Their diameter is from to .08’“'“ in the an- 
terior and median regions of the body; towards the posterior end, however, they be- 
come somewhat smaller. 
These canals are evidently the lateral canals of Duchamp, and the 'plasmatic ves- 
sels of Kuchenmeister and Zurn. 
The aquiferous vessels in transverse sections of the anterior portion of the body ap- 
pear to be of variable number owing to the fact that branches from the peripheral ves- 
sels join them at frequent interv^als. The branches usually unite with the central 
vessels at an acute angle so that the cut ends of the branches cannot be distinguished 
from the cut ends of the main central vessels. 
Near the anterior end there are eight principal vessels lying in two groups of four 
each, central to what I have designated the marginal canals. This disposition is not 
invariable, for in many sections more than four aquiferous vessels may be seen in the 
vicinity of one of the marginal cailals. 
Cross-sections of these vessels made near the anterior end of the body were oval 
in outline with the longer diameters varying from .014 to .03 millimeters and the 
shorter diameters from .01 to .014 millimeters. In both longitudinal and transverse 
sections the appearance is that of a hollow tube with a definite wall differentiated 
from the surrounding parenchyma and fibrous tissue. In cross-sections, the walls ap- 
pear to be structureless. The fine longitudinal and circular contractile fibers seen in 
the walls of the aquiferous vessels in D. cordiceps were not observed in any of the longi- 
tudinal sections of L. catostomi. The thickness of the walls is about .0025 millimeters. 
Iteference to, the measurements which I have given and to the sketches will show that 
the aquiferous vessels are throughout much smaller than the marginal canals. The 
difference between the two is iierhaps best shown in longitudinal sections, where the 
walls of the aquiferous vessels are seen to be more or less folded, but everywhere dis- 
tinct from the surrounding tissue, and the lumen free from tissue of any sort. A lon- 
gitudinal section of a marginal canal, however, shows a slightly sinuous canal filled 
with a fine fibrous tissue, appearing irregularly striated. Both sorts of vessels are 
surrouu<led by numerous nuclei, which in the case of the canals are in part entangled 
in the circular fibers which limit the canals, while in the case of the aquiferous ves- 
sels they do not enter into the structure of the vessel walls. 
In the median and posterior parts of the body where the genitalia have already 
begun to develop the aquiferous vessels appear to be reduced to two, one at a short dis- 
tance from and central to each marginal canal. 
The aquiferous vessels in the median region of the body were about .013 millime- 
ters in the smaller and .019 millimeters in the greater diameter, outside measurement, 
and .008 millimeters by .013 millimeters inside measurement. In this region the walls 
of the tubes are quite sharply defined and in some cases even slightly separated from 
the adjoining tissue. 
I have not been able to make out ajjy communication between either the periph- 
