760 
ANALYSIS OF CONTINENTAL JOURNALS. 
humour, when examined microscopically, appears to be dis¬ 
tinguished by particular elements, would we be much more 
advanced because we had made out these elements? It would 
be necessary, at first, to prove that they had not been 
accidentally introduced; and all naturalists who have studied 
the question as to the atmospheric germs and their develop¬ 
ment in animal or vegetable infusions, know how difficult it 
is to evade the causes of error that this development imports 
into the study of the organisms proper to the physiological or 
pathological humours. Good observers have been deceived, 
and singular blunders have been committed. 
If we were not limited as to time, and if I did not consider 
the task useless and superfluous, I might refer to the work 
which has had the most popularity among those that had the 
pretention of arriving at the determination of the virulent 
elements by microscopical analysis; we might together make 
a critical examination of this work, and I could show you a fine 
list of errors of this kind; you would see what aberrations 
may occur, even with the most learned in natural history, 
when questions in experimental medicine are studied without 
exact notions as to the exigencies accompanying the scientific 
demonstrations belonging to these questions. 
Lastly, admitting that particular elements found in a 
virulent humour really and properly belong to it; for the 
simple reason that similar humours, taken from other diseases 
do not present the same elements, shall we be authorised in 
attributing to the latter the specific activity of the humour in 
which it is found ? Assuredly not—and I will not even allow 
that this attribution may be a simple probability; for, strictly 
speaking, the same probability exists, a priori , for all the other 
constituent elements of the humour, inasmuch as it cannot 
be directly demonstrated that they are strangers to the activity 
of this humour. 
Observation in such cases, no matter how perfect it may 
be, cannot yield more than in analogous cases. It orginates 
certain hypotheses in the mind as to the causes, or rather 
the conditions of phenomena. It is to experimentation that 
recourse must be had to transform to certainty the characters 
of probability of these hypotheses ; a direct demonstration —one 
that is as complete as possible, and obtained by the rigorous 
application of the logical and technical procedure of the 
experimental method—is needed. The more the subject is 
delicate and difficult, the more exigent ought we to be in 
accepting the proofs which should constitute this direct 
demonstration. And can there be anything more difficult 
