680 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
which only some outlines are visible.” The farther up you go in the 
order of consequences, the farther away you get from real observa- 
tions and the more does certainty, or rather probability, diminish. A 
system is true not in proportion to the logic of its reasoning, but to 
the sum of positive facts introduced into it. 
It is to these philosophical conceptions that may be attributed to a 
certain extent Berthelot’s opposition to constitutional formule. A 
slave to facts, he would not admit these systematized signs to which 
some went so far as to attach an objective reality. ‘“ The symbols of 
chemistry present in this respect some strange allurements by the 
algebraic ease of their combinations and by the tendencies of the 
human mind. They naturally lead to the substitution, in the place of 
a direct conception of things, never absolutely determined, the more 
simple and apparently more comprehensive view of their representa- 
tive signs. It would be a strange misconception of the philosophy of 
the natural and experimental sciences to attribute to such mere ma- 
chinery for working a fundamental importance. In fact, in the study 
of the sciences, all depends on the discovery of general facts and of 
the laws that bind them one to another.” Berthelot saw in these for- 
mulze only a chemical language, and it meant no more to him than that 
the facts could be translated into one or another language. 
I may be permitted to recall that on reaching the college labora- 
tory at the close of a lecture, when Berthelot had explained his ideas 
on notations and chemical formule, I respectfully suggested to him 
that it would be more logical for him to use a language adopted by 
the majority of chemists. It was following this conversation that I 
presented to the Academy of Sciences the first work from the labora- 
tory of Berthelot with atomic formule. Some time after, in a work 
performed in collaboration with my teacher, on the chlorine deriva- 
tives Berthelot definitely gave up notation in equivalents for atomic 
notation (1890). 
Berthelot, at least at the time that I knew hin, attached only a 
secondary importance to theories. This, moreover, is a trait com- 
mon to nearly all learned men who have pursued a long scientific 
career. They have seen so many systems rise and fall that they 
arrive in the end at skepticism. I presented to him one day a short 
paper containing some theoretical ideas to which I attached some 
degree of importance and I carried it to him proud of my theoretical 
explanation of the facts observed. Glancing rapidly over my paper, 
Berthelot seized a pencil and quickly crossed out all that part on 
which I expected to be complimented. I was still a beginner, and 
notwithstanding all the admiration that Berthelot commanded from 
those about him, I confess that I consoled myself for my disappoint- 
ment by considering the act as that of a scholar grown too old. This 
little incident springs to my mind whenever I come across an old 
