1909. | BLOOD-PARASITES OF FRESHWATER FISHES. 15. 
mass. After about five hours I slipped off this coverslip also and 
made a preparation of the blood, but it only shows trypanosomes 
very badly preserved and stained, and apparently quite degenerate 
in structure. In the control drop of pure blood the trypanosomes 
were all dead or moribund after four hours, and I did not proceed 
further with it. If any conclusion can be drawn from these ex- 
periments, it is that the trypanosomes live, and remain normal, 
longer when tiie blood iv vitro is mixed with water than when it 
is mixed with salt-citrate solution or left pure; a result which I 
certainly did not expect. 
In preparations the trypanosomes of the perch show a con- 
tinuous gradation of sizes from the smallest to the largest (figs. 8- 
14); the larger forms being, however, by far the commonest. 
The best means of classifying them is by the free flagellum, which 
in the large stout forms is very short (figs. 13, 14), but in the 
medium-sized (fig. 12) and small forms (fig. 8) is much longer. 
The cytoplasm stains a very deep blue with the Giemsa stain, so 
deeply in fact that it is very difficult to obtain satisfactory pre- 
parations of the stout forms; they appear often as bluish opaque 
masses in which the intensity of the stain obscures all details of 
structure. In the same preparations, on the other hand, more 
slender forms may be found perfectly stained. The cytoplasm in 
the largest forms usually appears blotchy, with lighter and darker 
parts, often with tiny vacuole-like spaces, not very sharply limited. 
Any of the forms may contain red-staining granules to a greater 
or less extent, sometimes very numerous, sometimes absent alto- 
gether. The granules in question are of fair size and more or 
less irregular in form. In one specimen I saw them frequently 
in pairs, and sometimes rod-shaped, suggesting division (fig. 104) ; 
and the idea occurred to me that they might perhaps be intrusive 
organisms of the nature of Bacteria. In never-dried preparations. 
staloed with iron-hematoxylin the cytoplasm appears more or legs 
evenly and coarsely granular, according to the degree of extraction 
of the stain (figs. 96-99). ‘The above-mentioned granules are not 
brought out by this stain. I see, therefore, no reason for re- 
garding them as chromidia. 
The nucleus appears, in the smears stained by Giemsa’s method 
(figs. 8-14), as a red patch, often obscured and difficult to make 
out clearly in the large stout forms. It varies in size with the 
dimensions of the trypanosomes, and is much larger in the large 
forms. With the Romanowsky stain the details of nuclear 
structure appear to vary greatly; sometimes a distinct sharply 
limited karyosome, stained a deeper red than the rest of the 
nucleus, can be made out (fig. 13), sometimes not. It would 
appear as if the method of drying had the frequent result of 
distorting or breaking up the fluid or plastic karyosome, thus 
producing different appearances in different cases. On the other 
hand, in never-dried smears stained with Heidenhain’s iron- 
hematoxylin, the structure of the nucleus appears quite uniform 
in all its principal features (figs. 97-102), in all cases, and can be 
