SIMPLEST NEURAL ARCS. 41 



investigations, the Blix-Sandstrom kymograph has one rather serious 

 defect. It is never noiseless. In our measurements of the knee-jerk 

 the noise itself was probably negligible, but the correlated vibration 

 tended to transmit itself through the table to the axis of the recording 

 levers. When this occurred an irregular base-line was produced, which 

 more or less obscured the moment of muscle contraction. These 

 vibrations of the lever axis may be largely eliminated by suitable 

 independent supports. Before the cure was found, however, these 

 vibrations ruined a number of early knee-jerk records. 



A final difficulty which appeared as the experiments progressed was 

 the fact that the knee-jerk of a few subjects was highly refractory. 

 In all our subjects a knee-jerk was elicitable, but in some only by 

 reinforcements, by extra heavy hammers, or by considerably increased 

 velocity of the hammers. Under these exceptional circumstances, the 

 knee-jerk measurements were omitted, since intense stimulation tended 

 to produce not merely mechanical disturbances of the muscle, but also 

 unpleasant mental correlates, and an involuntary tendency to stiffen 

 the leg-muscles for the blows. Any one of these factors would operate 

 to make interpretation of the records questionable. 



EXPERIMENTAL PROCEDURE. 



The subject was seated comfortably in a slightly reclining chair at 

 the edge of the main apparatus table (position I, fig. 1). The experi- 

 menter moved the chair so that the subject's left leg fitted comfortably 

 into the double V supports (fig. 4) ; and the whole was oriented with 

 respect to the apparatus table so that the middle point of the quad- 

 riceps of the left leg was directly beneath the recording lever. Before 

 the first records of a day were taken, the height of the pendulum- 

 hammer system was controlled and accurately adjusted, so that the 

 blow was delivered on the middle of the patellar tendon. 



The recording-lever was then adjusted to its proper position. The 

 muscle end of the lever was placed in position and secured by an elastic 

 band which was passed around the thigh and fastened at proper tension. 

 The recording end of the lever was adjusted so that it was perpendicular 

 to the axis of the drum and tangential to its surface. 



The kymograph was set in motion and allowed four revolutions to 

 gain regular speed. (Measurements showed that our instrument gains 

 regular speed in three revolutions, when run at the rate of 100 mm. per 

 second.) The time-marker was set in operation. The subject was 

 instructed to relax completely, but to say "Ha" each time the knee was 

 struck. This was done in an effort to control both the attention and 

 the respiration. At each revolution of the kymograph, offsets from the 

 shaft broke the circuit of the electromagnets which controlled the 

 hammers, at a definite point of each revolution. This regulated the 

 interval between the stimuli and determined the position of the records 

 on the smoked paper. To insure regularity of the first stimulation, 



