ed constant, then an increase of galvanometer
deflection would accurately indicate a heightened or depressed E.M.
response, due to greater or less excitability of tissue caused by the
reagent. But, by the introduction of the chemical reagent, the
resistance of the tissue may undergo change, and owing to this cause,
modification of response as read by the galvanometer may be produced
without any E.M. variation. The observed variation of response may thus
be partly owing to some unknown change of resistance, as well as to that
of the E.M. variation in response to stimulus.

We may however discriminate as to how much of the observed change is due
to variation of resistance by comparing the deflections produced in the
galvanometer by the action of a definite small E.M.F. before and after
the introduction of the reagent. If the deflections be the same in both
cases, we know that the resistance has not varied. If there have been
any change, the variation of deflection will show the amount, and we can
make allowance accordingly.

I have however adopted another method, by which all necessity of
correction is obviated, and the galvanometric deflections simply give
E.M. variations, unaffected by any change in the resistance of the
tissue. This is done by interposing a very large and constant resistance
in the external circuit and thereby making other resistances negligible.
An example will make this point clear. Taking a carrot as the vegetable
tissue, I found its resistance plus the resistance of the
non-polarisable electrode equal to 20,000 ohms. The introduction of a
chemical reagent reduced it to 19,000 ohms. The resistance of the
galvanometer is equal to 1,000 ohms. The high external resistance was
1,000,000 ohms. The variation of resistance produced in the circuit
would therefore be 1,000 in (1,000,000+19,000+1,000) or one part in
1,020. Therefore the variation of galvanometric deflection due to change
of resistance would be less than one part in a thousand (cf. fig. 49).

#The advantage of the block method.#--In these investigations I have used
the block method, instead of that of negative variation, and I may here
draw attention to the advantages which it offers. In the method of
negative variation, one contact being injured, the chemical reagents act
on injured and uninjured unequally, and it is conceivable that by this
unequal action the resting difference of potential may be altered. But
the intensity of response in