298 MESSES. F. GOTCH AND V. HORSLEY 



the distance in question was necessarily short, the dangers of extrapolar spread were 

 increased, and the experimental results had to be controlled with great care ; this 

 danger was most formidable in the case of the anterior root, as owing to its anatomical 

 relations it could not be so conveniently separated from its contiguity with the mass 

 of the cord, and the errors thus introduced have been at present sufficiently serious 

 for us to abandon these roots for direct connection with the galvanometer, and rely on 

 those experiments performed upon the sciatic nerve with a,ll the posterior roots cut. 



(2.) The Condition of the Non-polarisdble Electrodes. 



It is scarcely necessary to observe that one obvious precaution in connection with 

 the electrodes is to ascertain, before commencing the experiment, that between the 

 electrodes themselves there is only an extremely slight and constant electrical differ- 

 ence, and noting its amount in terms of the compensator. 



There is one difficulty, however, which appears to be inseparable from a method 

 which involves a long bridge of kaolin moistened with O'G per cent, of saline, exposed 

 to the air of the room, and that is the alteration of resistance due to slow evaporation. 

 This alteration does not affect the results when observed by the electrometer, owing to 

 the very high resistance offered by the latter. It may, however, seriously affect the 

 galvanometer results. It will be obviously mitigated by employing a circuit of very 

 high resistance, and this we always took care to do, the full resistance of galvanometer 

 and electrodes amounting to little short of 50,000 ohms. 



It is, however, necessary to keep in mind the presence of such alterations, and 

 to ascertain by repeated measurement of the galvanometric effect produced by a given 

 difference in the compensation circuit, whether the resistance has so altered as to 

 vitiate the value of the results. Where experiments follow one another closely, as 

 was the case in all the instances in which the different columns of the spinal cord 

 were excited, any alteration would be obviously too slight to exercise an appreciable 

 influence upon the results, which admit, therefore, as far as this feature is concerned, of 

 strict quantitative comparison. 



It is obvious that remoistening of the electrodes must only be effected at the 

 commencement of a fresh series of observations, and must on no account take place 

 during any given series. 



(3.) The Condition of. the Animal. 



(a.) General Preliminary Condition. In all the experiments this was a very 

 important factor, which we early recognised as determining the success or failure 

 of the particular observations. If the animal were ill, or had been much exhausted 

 by the preliminary ansesthetisation, &c., it was noted that the electrical changes, 

 whether due to currents of rest or of action, in both nerve and spinal cord were 

 comparatively small in amount. This was especially the case in the Monkey. 



(b.) AncBxthesia. The degree of narcosis is an extremely important factor. It will 



