d honourable men working at
the same subject have arrived at conclusions different from mine. Most
freely granted; but let me here recur to the remarks already made in
speaking of the experiments of Spallanzani, to the effect that the
failure of others to confirm his results by no means upsets their
evidence. To fix the ideas, let us suppose that my colleague comes to
the laboratory of the Royal Institution, repeats there my experiments,
and obtains confirmatory results; and that he then goes to University
or King's College where, operating with the same infusions, he obtains
contradictory results. Will he be disposed to conclude that the
selfsame substance is barren in Albemarle Street and fruitful in Gower
Street or the Strand? His Alpine experience has already made known to
him the literally infinite differences existing between different
samples of air as regards their capacity for putrefactive infection.
And, possessing this knowledge, will he not substitute for the
adventurous conclusion that an organic infusion is barren at one place
and spontaneously generative at another, the more rational and obvious
one that the atmospheres of the two localities which have had access
to the infusion are infective in different degrees?

As regards workmanship, moreover, he will not fail to bear in mind,
that fruitfulness may be due to errors of manipulation, while
barrenness involves the presumption of correct experiment. It is only
the careful worker that can secure the latter, while it is open to
every novice to obtain the former. Barrenness is the result at which
the conscientious experimenter, whatever his theoretic convictions may
be, ought to aim, omitting no pains to secure it, and resorting only
when there is no escape from it to the conclusion that the life
observed comes from no source which correct experiment could
neutralise or avoid.

Let us again take a definite case. Supposing my colleague to operate
with the same apparent care on 100 infusions--or rather on 100 samples
of the same infusion--and that 50 of them prove fruitful and 50
barren. Are we to say that the evidence for and against heterogeny is
equally balanced? There are some who would not only say this, but who
would treasure up the 50 fruitful flasks as 'positive' results, and
lower the evidential value of the 50 barren flasks by labelling them
'negative' results. This, as shown by Dr. William Roberts, is an
exact inversion of the true order