540 
specificity, but that it can also act as iso- and hetero-agglutinin on 
the erythrocytes of other rabbits and those of animals of widely 
different species. 
VALUE OF THE PHENOMENON AS A DIAGNOSTIC SIGN 
Before discussing this question it is necessary to emphasise the 
importance of careful observation in determining whether a certain 
blood really agglutinates or not. So far as can be gathered from 
the papers in which the existence of the phenomenon in 
trypanosomiasis has been recorded, it has been invariably decided 
from the examination of cover-slip preparations of the blood. 
Although a considerable degree ol auto-agglutination is easily 
recognised in a well-made cover-slip preparation, yet it is often 
extremely difficult, or even impossible, to decide whether the red 
cells are really agglutinated when the phenomenon is not so distinct. 
A certain amount of massing together of the erythrocytes is 
frequently evident at the edges of even the best cover-slip 
preparations of normal blood, whereas it the slide and co\er-slip 
be not perfectly clean the red cells are found to be anything but 
evenly distributed, but are grouped together into little masses and 
rouleaux, separated from one another by plasma an appearance 
closely resembling that to be observed in infected blood when the 
amount of auto-agglutination is slight. On the other hand, a slight 
degree of auto-agglutination can be easily obscured by pressure on 
the cover-slip resulting in the separation of the erythrocytes one 
from the other. 
Furthermore, it has been shown that small amounts ol 
auto-agglutinin exist constantly in the blood of many normal 
animals. In horses and donkeys auto-agglutinin is sometimes 
present in such an extent as to give rise to a more oi less 
characteristic appearance in cover-slip preparations. This is 
specially the case when the preparations are made out of doors at 
a somewhat low temperature. However, I have never observed in 
cover-slip preparations of the blood ol normal animals a con 
approaching in intensity the well-marked clumping obtaining in 
infected cases. When a high degree of auto-agglutination exists 
the corpuscles are seen to have become agglomerated into tight 
clumps, the outlines of the individual cells being indistinct, or e\en 
