THE INFLUENCE OF SECONDARY RESISTANCE 85 
which larger values of /3, yet, if the experiments are 
carefully performed, one can be practically certain, 
whenever the values of Z differ by more than 15 or 18 
per cent, that the larger Z means also a larger /3. With 
this degree of accuracy assured, probably the demands 
of most researches of this class are fully met, and all 
such may safely disregard both the secondary resistance 
and the cathode surfaces. 
Differing from the series of experiments quoted above 
in that they offer wider variations in both secondary 
resistance and electrode surface, and therefore greater 
likelihood of error if these factors be disregarded, is a 
series of observations on frogs' gastrocnemius muscles, 
carried out by myself. 
In the series of eighteen experiments cited in Table X 
the secondary resistances ranged from 3100 to 13,000 
and the values of A from 2600 to 13,500. Yet, in spite 
of these wide ranges in the values of the factors deter- 
mining the relation of Z to 0, this latter relation varies 
to a surprisingly moderate degree. The average ratio 
of to Z is .49. The widest departures from this are 
ratios of .32 and .64, amounting to 35 per cent and 31 
per cent respectively, while the average variation is only 
15 per cent. If the experiments of Table X represent 
fairly the variations in secondary resistance and cathode 
surface likely to be met with in experiments on frogs' 
gastrocnemii, we can safely conclude that the values of 
