FATIGUE IN NERVES. 151 



gave nearly as prolonged contraction as it did before the long stimula- 

 tion. 



Lastly, the absence of effect was not due to loss of irritability in 

 consequence of degeneration. In the frog the dorsal roots, as also other 

 nerves, very slowly lose their irritability. In one case I cut the roots 

 and left them for 10 days. At the end of this time they still caused 

 the normal amount of movement. 



It is thus seen that the seat of fatigue is in the root itself, and also 

 for the production of this fatigue there must be excitation of the cells 

 in the cord. 



3. Stimulation of roots, one severed and the other unsevered 

 from the spinal ganglion. 



In order to see if the dorsal root would activate the cells in the cord 

 longer if it were still connected with the nerve cells in the spinal 

 ganglion, the following three series of experiments were made. 



1. In the first set a pair of electrodes was passed under the uncut root, 

 and the cut root also laid on the same electrodes so that the two sides 

 were stimulated at the same time. At first fatigue is obtained equally 

 rapidly on both sides, and at first recovery is as rapid on both sides. 

 After a time the cut root begins to lag behind, both in its power of 

 recovery and in its effectiveness in keeping the movement prolonged, 

 and finally the cut root is ineffective while the unsevered root is quite 

 good, and in some cases nearly as effective as at first. As the meta- 

 bolism in frogs is slow, and this slowness is no doubt increased by the 

 low temperature at which these frogs are kept, it is some hours after 

 the experiment has commenced before any difference is noted in the 

 two roots. 



Although I tried always to regulate the current so that it gave only 

 the direct reflex, yet one could not be certain that a movement might 

 not be partly due to the impulse from the other root, and as the uncut 

 root is very liable to be detached from the cord by a sudden movement 

 of the frog, this plan was abandoned, but not before four experiments 

 had been completed and several others nearly so. 



2. In the second set of experiments the plan was adopted of 

 stimulating the cut root and the corresponding cut nerve in the pelvic 

 plexus. By means of a commutator the two sides were stimulated 

 alternately and approximately equally for equal periods. In order to 

 equalize the sides the two roots (anterior and posterior) of the one side 



