48 
result in any case agrees with his anticipations, he does not 
regard it as finally confirmatory of his theory, but proceeds to 
test the truth of the theory by new deductions and new trials.”* 
It is therefore of the utmost importance that the naturalist 
should have an almost instinctive aptitude in conceiving hypo- 
theses, to be used, however, only as finger-posts directing him 
along a certain line of observation, and only to be used while 
they are useful, but to be discarded without hesitation when 
they would lead him into the quagmire of error. Hypotheses 
of this kind are only tentative, and must be regarded merely as 
the scaffolding to a more permanent erection, but must never 
be mistaken for the erection itself. 
But before we begin to build we must see that we have 
sufficient materials with which to complete the structure, so 
that it may be well-proportioned and secure. In other words, 
it is of great importance that we should collect a considerable 
number of facts before we commence theorizing; if we have 
only a few, we have no range of vision, our power of comparison 
is limited, and, consequently, the evidence in favour of any 
explanation being insufficient, the explanation or hypothesis 
erected on it will be as a cone on its apex, in very unstable 
equilibrium, easily overturned, as many such have been. 
Several naturalists have fallen into the mistake of elaborating 
theories of the universe the instant they have discovered a few 
supposed facts, which seemed new to them, instead of patiently 
gathering more, or trying to verify those previously observed. 
The more abundant the evidence the more likely is the verdict 
to be true. 
I have so far spoken only of the evidence obtained by direct 
observation ; when, however, we can by experiment repeat the 
phenomenon at will, and so verify or correct it, our confidence 
in the results we have obtained is greatly increased. But to 
speak of the absolute necessity for varied and accurate experi- 
ment in the procuring of scientific evidence would be here a 
mere waste of time; there is, however, one great result accom- 
plished by it which I would not wholly overlook, and that is, 
the deciding some of several supposed causes to be the actual 
one in the production of the observed effect. The corpuscular 
and undulatory theories of light seemed each of them for a 
time to satisfactorily account for the phenomena; but when it 
was proved by experiment that light moved more slowly 
in glass than in air, the undulatory theory which predicted this, 
was known as more likely to be true than the corpuscular, 
* J evons’ Principles of Science, vol. ii. p. 137. 
