440 ELECTRO-PHYSIOLOGY CHAP. 



has been previously heated gives quite a different reaction ; the 

 negative variation in this case quickly reaches its maximum, and 

 only diminishes slightly if the excitation of the nerve continues. 

 The magnet only returns to rest when the circuit is broken. 

 The effect is therefore similar to that in striated vertebrate 

 muscle. We have stated above that the positive variation of 

 the fatigued muscle tends at first to a permanent increase of 

 the demarcation current, while later on, with repeated excitation, 

 it compensates itself each time slowly, but completely. A 

 negative after-variation, such as is often observed on strongly 

 curarised preparations, is regularly absent in fatigued muscle. 



The next question connected with the appearance of the 

 positive variation on indirect tetanisation of the adductor muscle is 

 whether there is here a discontinuous alteration of state in 

 the muscle -substance, as in the negative variation which 

 otherwise accompanies excitation. We have unfortunately not 

 been able to decide this point from frog's nerve -muscle pre- 

 parations, since, under even the most favourable conditions, it 

 seems impossible to elicit secondary twitch, or secondary tetanus, 

 from the adductor muscle. Investigation of these effects by 

 means of the capillary electrometer, which we have not yet been 

 able to carry out, might here also give the desired solution. In 

 the light of other observations on the same preparation to be 

 discussed below the experimental results point to an explana- 

 tion, which is directly connected with Gaskell's theory, as above 

 quoted. Remembering that the adductor muscle of the crab's 

 claw is supplied, demonstrably, by two functionally distinct nerves, 

 inhibitory and excitatory (assimilatory, dissimilatory), which on 

 excitation produce opposite changes of state in the muscle-sub- 

 stance (expressed by antagonistic changes of form on the one 

 hand, and by contrary electromotive action on the other), it 

 must be assumed that the galvanic effect observed with a moder- 

 ate intensity of the artificial stimulus to the nerve is, as a rule, 

 the result of co-operation between two contrary and simul- 

 taneously excited processes, the alternating ratio of which depends 

 on the one hand on the strength of excitation, on the other on 

 the temporary condition of the muscle. 



The strongest argument in favour of a masked diphasic 

 response even in such galvanic effects as, with strong excitation 

 of the normal adductor muscle, yield experimentally negative 



