ON CERTAIN PECULIARITIES OF THE KNEE-JERK 
IN SLEEP IN A CASE OF TERMINAL DEMENTIA. 
WILLIAM NOYES, M. D., 
Assistant Physician and Pathologist, McLean Asylum . 
From the Laboratory of the McLean Asylum, Somerville, 
Mass. 
Reprinted from the American Journal of Psychology, Vol. IV, No. J, 
April, 1892. ON CERTAIN PECULIARITIES OF THE KNEE-JERK IN 
SLEEP IN A CASE OF TERMINAL DEMENTIA. 
From the Laboratory of the McLean Asylum, 
Somerville, Mass. 
By William Noyes, M. D., Assistant Physician and Pathologist. 
In the more precise investigations of the physiological con- 
ditions modifying the knee-jerk all investigators have found 
that the mental condition of the normal subject entered largely 
as a disturbing factor, the different emotional states and the 
varying conditions of the nervous system invalidating the 
result to a certain extent, or at least rendering necessary an 
enormous number of observations before the " normal" knee- 
jerk for any individual could be obtained. Emotional states 
being almost completely absent in cases of advanced dementia 
it was thought that an investigation of the knee-jerk in this 
state might yield some results on the conditions modifying 
it. An excellent subject was found in an elderly man, ad- 
mitted to the asylum in 1841, and with the exception of the 
seven years preceding 1850 having spent this whole period at 
the McLean Asylum. 
In 1843 it is stated that his 11 mental faculties [are] mostly 
gone; never makes an inquiry ; spends his time in wandering 
about the yard; slovenly in his dress;" and but little or 
nothing could be added concerning his mental condition at 
the present time. His dementia is complete, he having 
absolutely no knowledge of his surroundings, not even know- 
ing his name, is unable to answer questions relevantly, his 
talk being utterly incoherent. He is good natured and docile 
and has never made the slightest objection to the experiments, 
which almost always mean an extra nap for him. In fact he 
1 2 
NOYES: 
will go to sleep in almost any position in which he is placed, 
and remains perfectly quiet and contented. Several demented 
men were tried, but with the exception of this one they all 
proved resistive or restless, and were given up one after 
another. The patient finally selected was a peculiarly good 
one for a prolonged series of tests on the condition of his 
nervous system as he was subject to periods of depression 
and exhiliration, such as not infrequently occur in chronic 
dements, and a record was made of all his bodily functions, 
pulse, temperature, weight, blood pressure so far as this 
could be measured, mental condition, and finally knee-jerk 
in the hope of finding some change in his bodily condition 
coincident with the change in his mental state. Unfortunately, 
from the point of view of gaining any information on this 
particular point, after passing through one cycle of exhilira- 
tion and depression he settled down into a state of compar- 
atively even mental and physical life, and as the obser- 
vations taken during this cycle would need further con- 
firmation no mention will here be made of them. The 
original problem, also, the theoretically lesser variability of 
the knee-jerk in a demented person, gradually changed into 
another as will appear subsequently, and the present paper 
will be confined to a consideration of two important points 
that came out in the course of the investigation. 
Apparatus and Method of Experimentation. 
After a trial of different hammers that of Lombard1 was 
finally adopted, that of Prof. Bowditch2, which was kindly 
loaned for an experimental trial not proving strong enough 
for this patient in whom a rather heavy blow was necessary. 
The hammer was of cast iron, the handle being an iron rod 
20 cm. long, on the end of which a similar piece of iron rod 
4 cm. long was fastened horizontally, serving as an axis and 
supported in an iron frame work. An index on the handle 
moved over a graduated scale, so that by raising the hammer 
any given number of degrees the same strength of blow was 
given. Throughout the experiments the hammer always fell 
45°, except in a few instances where the strength of the blow 
was changed for a short time for some special purpose. The 
hammer was supported by horizontal rods and clamps to a 
piece of gas-pipe, serving as an upright support fastened to a 
firm wooden base that could be securely clamped to the table. 
The arrangement combined firmness and strength, while at 
the same time the hammer could be quickly raised or lowered, 
1 American Journal of Psychology, 1887, Vol. I., p. 12. 
- Bowditch and Warren Journal of Physiology, 1890, XI, p. 27. PECULIARITIES OF THE KNEE-JERK. 
3 
or moved in any direction horizontally, and thus easily ad- 
justed to hang at zero of the scale, when the patient's leg was 
put in position. When pulled back into position to strike 
the blow the hammer was held by a large electro-magnet, the 
support of which was also adjustable. The hammer was 
pulled back by a cord and released by a simple circuit- 
breaking key. The blows were always given at 5 seconds 
intervals, except that the interval was occasionally changed 
for special purposes, but in the reports of the experiments 5 
second intervals are always to be understood unless other- 
wise distinctly stated. The patient reclined on his right side 
on a mattress placed on a table, with a firm support extend- 
ing the length of his back. The left knee was supported on 
a wooden arm extending from a wooden upright that was 
firmly clamped to the table, and the left foot was supported 
in a stirrup hanging from the ceiling allowing free movement 
of the leg. Attached to the stirrup at the point on which the 
heel rested was a steel rod 3 feet long, passing backwards and 
supported on pulleys; on its further end was fastened a 
thread passing over pulleys and attached to a short vertical 
steel rod, moving up and down through two brass supports, 
and suspended from above by a light spring. A cork placed 
on the rod carried on its side the writing point, a piece of 
light stiff celluloid, which pressed lightly against the drum 
of a Bal tzar Kymograph. When the blow was struck the 
foot moved forward pulling the steel rod with it, and this by 
its thread attachment pulled down the writing point, leaving 
the record of the full length of the kick on the smoked sur- 
face of the paper. There was of course always a slight back- 
ward kick which appears in the records above the horizontal 
line made by the drum revolving under the writing point 
when this was at rest. 
Disappearance of the Knee-Jerk in Sleep and the Effect of 
Auditory Stimuli. 
The disappearance of the knee-jerk in sleep has been 
previously observed.1 Its complete disappearance was early 
noticed in our patient and this is therefore no departure 
from the normal. As would naturally be expected the 
patient went to sleep much more easily than a person with 
an active mind, and it was thus possible to observe the 
phenomena a large number of times. 
The behaviour of the knee-jerk when sensory stimuli are 
received by the patient during this period of sleep appears, 
'Bowditch and Warren, loc. cit. p. 59. 
Lombard, loc. cit. p. 53. 4 
NOYES: 
however, to be a distinct departure from the normal, and 
suggests that the weakened state of his brain permits the 
sensory stimulus to exert its effect in reinforcing the knee- 
jerk for a much longer period than in a person with a sound 
brain. 
That the knee-jerk is increased when the patient clinches 
his hand or makes any violent movement coincident with the 
blow was first shown by Jendrassik1. "He also thought 
that stimulation of the sensory nerves had a similar influence 
on the tendon reflex, but considered his experiments on this 
point incomplete and that such an influence was more difficult 
of determination." (Quoted from Bowditch and Warren.) 
Mitchell and Lewis2 made a study of the conditions under 
which the knee-jerk is increased and diminished, and found 
that volitional acts directed to other parts of the body, pain- 
ful stimulation of the nerves of the skin either by pinching or 
by the application of heat, cold or electricity, caused a rein- 
forcement, as did also a burning magnesium wire exposed to 
the eyes. 
Lombard found that sensory irritations, voluntary move- 
ments and strong emotions when synchronous with the blow 
increased the knee-jerk. The investigation of Bowditch and 
Warren had for its object a study of the exact relations in 
time between the knee-jerk and the reinforcing act, and was 
suggested by the statement of Mitchell and Lewis " that the 
muscular action or circuit closing, must precede the tap, in 
order to reinforce it, by a period which is, as yet, undeter- 
mined." The conclusion of Bowditch and Warren with 
regard to auditory stimuli was that "the effect of a sudden 
auditory stimulus on the extent of the knee-jerk was, in the 
three subjects of experiment, almost wholly positive, though 
great individual differences were observed. The maximum 
effect was produced when the interval between the sound and 
the blow was 0.2"-0.3"." 
As the results here to be recorded have to do with the 
knee-jerk in sleep the experience of previous investigators on 
this point is of interest. 
Bowditch and Warren3 found that " the monotonous char- 
acter of the experiment was often found to produce a decided 
tendency to sleep in the individual experimented on. To 
counteract this tendency and to insure a certain degree of 
attention to the phenomena, the subject of the experiment 
was required to declare after each knee-jerk whether or not a 
'Jendrassik: Beitrage zur Lehre von den Sehnen reflexen. Deutsches 
Arch. f. Klin. Med. 1883, XXXIII, 177. 
2 Medical News, Feb. 13 and 20, 1886. 
3Loc. cit. p. 58. PECULIARITIES OF THE KNEE-JERK. 
5 
sensory-stimulus (i. e. sound, flash, etc.,) had been perceived, 
or in other words, whether the knee-jerk was normal or rein- 
forced. In spite of this precaution the tendency to sleep was 
sometimes quite irresistible, and in eight or nine cases the 
experiment was continued after the subject had yielded to it 
and was sleeping soundly. It was then found that the knee- 
jerks, both normal and reinforced grew gradually smaller, 
and when sleep was profound, disappeared entirely, the blow 
upon the knee being absolutely without effect. This result 
is not what might have been expected from our knowledge of 
the effect of sleep on the ordinary cutaneous reflexes, e. g., 
that produced by tickling the sole of the foot. Whether this 
can be regarded as an argument against the reflex character 
of the knee-jerk, or whether we have here an essential differ- 
ence between deep and superficial reflexes, are questions to 
be decided by future investigations." 
Prof. Lombard has kindly gone over with me his own 
curves made in 1887 in New York, generally on healthy 
medical students, and many of these show an absolute dis- 
appearance of the knee-jerk in sleep, and they also show a 
sudden rise from 0 from some accidental stimulus such as the 
entrance of a person into the laboratory, but this effect lasted 
over but a few kicks. 
In our demented patient when the knee-jerk has entirely 
disappeared in sleep an auditory stimulus causes an increase 
in the length of the kick, and this increase is visible over a 
long series of jerks. This latter phenomenon is apparently a 
departure from the normal and a peculiarity of this subject. 
This in well shown in Fig. 1. The patient was asleep 
from the beginning of the experiment and soundly asleep with 
total disappearance of the knee-jerk during the minute and 
three quarters preceding the time when two light taps were 
given on the table with the wooden handle of the needle that 
was being used by the observer to make records on the drum. 
The knee-jerk rose from 0 to 9.5 mm.; in 5 seconds rose to 
39.5 mm.; in the next 5 seconds to 59 mm.; then fell to 38 
mm.; through the 29 succeeding kicks gradually fell with 
slight fluctuations to 6.5 mm.; remained at practically this 
length for 3 kicks ; rose to 15 mm., and then to 28 mm.; fell 
to 16 mm. and 14 mm.; then came 5 kicks averaging 6.6 mm.; 
then came a rise without apparent cause to 44 mm.; then 
followed a fall to 0 after fluctuations extending over 27 
kicks. Thus one auditory stimulus had an apparent effect 
extending over 72 kicks occupying 6 minutes. During this 
whole time the subject was sleeping soundly, being apparently 
as sound asleep as when the auditory stimulus was first given. 
Five kicks of the value of 0 then followed, and there then 6 
NOYES: 
came on another series of increased kicks with no apparent 
cause, lasting over 16 kicks, with another total disappearance. 
There then followed 28 kicks with almost total disappearance 
of the knee-jerk and then two taps (B) brought out the series 
of reinforced kicks again, with a peculiar series of groups of 
kicks extending to near the end of the experiment, with occa- 
sional total disappearances. Toward the end (C) two taps 
produced almost no effect, causing a rise only to 10 mm. from 
a preceding 7, then followed 18, 10.5, 4 and 0. After three 
of 0 value another lengthened group came on with no ap- 
parent cause, and at this point the subject was awakened by 
being spoken to. It was noted that this day he was decidedly 
more dull than on the day before, that he slept from the 
beginning to the time he was awakened, and that all efforts 
to rouse him at this point amounted to but little as he would 
simply mutter a little and then drop off to sleep again. There 
were no accidental sensory reinforcements that were sufficient- 
ly noticeable to be brought to the consciousness of the ob- 
server who was on the watch to note them. The peculiar set 
of 11 groups " of kicks occurring during the course of this 
experiment will be discussed later on. 
This prolonged effect of a sensory stimulus is seen again in 
Fig. 2, where two clicks were given on a telegraph sounder at 
A, after the knee-jerk had entirely disappeared. Here the kick 
rose from 0 to 18 mm., and then to 42 and 44, falling through 
17 kicks to 3 mm., not reaching 0 again for 17 more kicks. 
There were three kicks of the value of 0, a short rise over 4. 
kicks, a total disappearance during 2 more, the needle not 
even making a dot on the line, and then there came another 
of the " groups," B, with no apparent exciting stimulus, 
lasting over 10 kicks before these entirely disappeared again. 
There were slight rises above the 0 line, and at C a noise on 
the floor below caused another series of reinforced jerks, and 
the subject awoke. Practically this same effect of stimuli is 
seen in Fig. 3, where at A and B walking on the floor below 
caused the rises there shown ; the much lesser effect of the 
second stimulus is seen in the fewer number of kicks that 
were caused by it. Blows were delivered at the regular 5 
second intervals although without the effect of moving the 
needle from the straight line, until at C two taps with the 
needle-handle and a distant locomotive whistle a little 
later caused the prolonged series of increased kicks that 
followed. Here again it must be remembered that to all 
outward appearances the patient was sleeping as soundly 
as when the sensory stimuli were received. Fig. 4 shows 
the same effect once more, the knee-jerk having entirely 
disappeared during the period represented by the straight PECULIARITIES OF THE KNEE-JERK. 
7 
line preceding A, blows being struck at the regular 5 
seconds intervals, while at A, a passing barrow caused a 
slight rise and at B the noise of a passing cart caused a much 
greater rise, with a secondary rise a little later, with another 
rise before the knee-jerk entirely reached 0 again, and then 
after three kicks of the value of 0 there was still another rise, 
when there was again total disappearance. 
In Fig. 5, the effect is shown of giving the rapping-stimulus 
seven times in succession. The largest effect was after the 
second time where the reinforcement continued over 13 
kicks before the zero point was reached. After the third 
stimulus there were 8 reinforced jerks before the kicks sank 
again, not to 0, but to 4 mm. Stimulus no. 4 called out only 
two reinforced kicks, no. 5 called out 2, no. 6, 4, and no. 7, 
6. At no. 8 the patient was awakened. 
These would seem to show that the subject had gradually 
become accustomed to the stimulus and thus its effect was 
lessened. If it is argued that the patient was more wakeful 
at 6 and 7 as shown by the fact that the kicks did not return 
to 0, this would mean also that stimuli did not have so great 
an effect during this more wakeful condition as during the 
condition of deeper sleep. This diminution of the effect of a 
reinforcing stimulus is shown in Fig. 6 where 2 clicks were 
given on a telegraph sounder at (1), causing a marked effect 
lasting through 6 kicks, and then disappeared, to be followed 
by a secondary rise during 7 kicks, and then a total disap- 
pearance. A repetition of the clicks however at (2) was 
followed by no response for the first kick but the second and 
third kicks rose respectively to 1 and 2 mm., which ordin- 
arily would be accounted for by the jarring of the apparatus, but 
in this particular instance there had been no movement even 
from the jarring, so it seems fair to attribute these two move- 
ments of the writing needle, slight as they are, to the knee- 
jerk. A third repetition of the stimulus (3) called out a 
knee-jerk of 3 mm., followed by a total disappearance during 
5 kicks, then a kick of 8 mm. with no apparent preceding 
stimulus, a disappearance of three kicks, and then a repetition 
of the clicks, this time three clicks instead of two, produced a 
kick of 8 mm. with another total disappearance. A repetition of 
the three clicks (5) caused no response for the first kick, but 
the second rose to 26 mm., the third to 43, and from this 
time on there was no disappearance but a kick followed 
every blow of the hammer. 
Attempts to find a similarly marked reinforcement from 
auditory stimuli while the patient was awake did not meet 
with as good results as while he was asleep. Fig. 7 shows 
a portion of a curve taken from the middle of a tracing while 8 
NOYES: 
the subject was fully awake. At the places iudicated by dots 
two clicks were given on the telegraph sounder and an 
apparent reinforcement appears at times, and again an appar- 
ent inhibition. This disagreement may be due to the fact that 
the interval of time by which the sound preceded the blow was 
not measured, so that the blows may have been struck at the re- 
inforcing interval at one time, and at the inhibiting interval at 
another. The irregularities in the kicks, during the period that 
the reinforcing signals were being given, are no greater than the 
irregularities in the kicks preceding and following the rein- 
forcing signals. The sounds both inside and outside the labor- 
atory, that had so much effect during sleep, never appeared to 
have a corresponding effect when the subject was awake. To 
have settled this point definitely would have necessitated a 
repetition of the elaborate experiments of Bowditch and 
Warren. Nor was the interval by which the sound preceded 
the blow in sleep measured; it probably usually varied 
between one and two seconds. As the peculiar prolongation 
of the reinforcement made its appearance apparently irrespect- 
ive of the interval by which the sound preceded, particular 
attention was not given to this point. It would be interest- 
ing to determine if an interval could be found at which the 
stimulus would inhibit such a prolonged series of kicks as 
shown in some of the curves. On several occasions the first kick 
after an auditory stimulus did not rise as high as the second, 
as shown in Fig. 3, and this raises the question whether the 
inhibiting interval may not have accidentally been struck here, 
but the effect of the inhibition passing off during the succeed- 
ing 5 seconds, the stimulus exerted its full force and the kick 
rose to the maximum. Should this be so, it would seem to 
point to the necessity, in measuring the interval by which 
the blow must be preceded by the sensory stimulus to pro- 
duce inhibition or reinforcement, of following the first blow 
by several more at comparatively short intervals. With 
regard to sensory stimuli that reached this patient during the 
time he was awake it can only be stated that these appeared 
to have a very slight and triflina: effect compared with those 
that reached him during sleep. 
It should be added that previous observers of the disap- 
pearance of the knee-jerk in sleep and of its rise in this con- 
dition in response to external stimuli have not delivered the 
blows at the same intervals as in these experiments, conse- 
quently it is possible that some of the effects of the stimuli 
may not have been recorded in their tracings. Attempts were 
made to settle this point by experiments on normal indi- 
viduals, but the knee-jerk did not entirely disappear in sleep 
in the trials that were made. In Fig. 12 there is shown a PECULIARITIES OF THE KNEE-JERK. 
9 
portion of one of these tracings from one of the medical house- 
pupils. The experiment was begun at 10.10 p. m. and con- 
tinued until 11.15 p. m. The portion shown is from about 
the middle of the tracing, after the subject had become 
thoroughly drowsy, and a grdual diminuation of the length of 
the kicks is seen as the experiment progressed. When it be- 
came probable that the knee-jerk would not completely disap- 
pear the customary auditory stimuli, two taps with the needle- 
handle, were given at (1), (2), (3) and (4). Although there 
was a response in each case the effect of the stimuli extended 
over a much shorter period than in the demented patient. 
Beginning with the tenth kick preceding the point at which 
the stimulus was first given at (1), the length of the kicks in 
millimetres is given below. 
5.5 
2 
2.5 
10 
7-(4) 
5 
3.5 
9.5 
7.5 
14 
5.5 
4 
2 
3 
5 
5.5 
14 
2 
3.5 
8 
3.5 
9 
2 
3 
2 
10 
5 
5 
3 
7 
4 
10 
7 
7-(3) 
6 
4.5 
6 
15 
4.5 
10 
4 
3 
4 
3 
7 
1.5-(1) 
5 
15 
15 
9 
8 
2.5 
4 
3 
6 
13 
15-(2) 
5 
4 
8 
4 
3 
12.5 
2 
6 
11 
In each of the four cases the two taps were given im 
mediately preceding the kick designated by the figures (1), 
(2), (3) and (4). 
It will be seen that the effect of the first stimulus can be 
traced over the three following kicks, and then the knee-jerk 
fell to 4 mm. The second stimulus caused a rise from 2.5 mm. 
to 15 mm., but at the next kick the knee-jerk fell to 3 mm. 
The third stimulus caused a rise from 3 mm. to 7 mm., and 
the next kick fell to 4.5 mm. As the effect of the stimuli was 
evidently diminishing, the fourth stimulus (4) was made much 
louder than the preceding ones, and the knee-jerk rose from 
2 mm. to 7 mm., and then to 14 mm.; the succeeding three 
kick were 5 mm., 8 mm., and 2 mm. 
In no instance did any such prolonged effect from the 
stimuli occur as is shown in the tracings from the demented 
patient. Fig. 12 also shows the "groups" of kicks that 
could not be identified with any external stimulus; these 
groups are well shown before the stimuli were given by the 
two taps. During the time that these groups appeared the 
subject was in the same condition of half-sleep as when the 
groups came out best in the demented patient. 10 
NOYES: 
It was possible on several occasions to find well marked 
evidences of this rhythm while the subjects were fully awake. 
Fig. 11 is a tracing from the demented patient while awake, 
and the wave like movement of the tops of the kicks is very 
evident. 
As has already been stated there were suggestions of this 
rhythm (?) all through the tracings, from the first of the 
experiments, even during the waking state; but in none of the 
other tracings during the waking condition is it as well shown 
as in Fig. 11. This same suggestion of a rhythm is also 
shown in Fig. 13, which is of a tracing taken from a 
case of well advanced general paralysis. This patient was 
in the quiet and apathetic stage of the disease, much demented, 
with ataxic gait, and slow, stammering speech. He was 
awake during the whole experiment. The same wave-like 
appearance of the ends of the knee-jerks is well marked. 
A tracing from still another patient is shown in Fig. 14. 
This man was a case of dementia, but not nearly so far ad- 
vanced as the first case. He was awake during the experi- 
ment. There is seen the same suggestion of the wave-like 
motion of the ends of the knee-jerks. 
A point not immediately connected with those already dis- 
cussed, but having a bearing on the general question, is illustra- 
ted in Fig. 15, which is a portion of a tracing from the first case 
of dementia. The subject was fully asleep as is seen from the 
total disappearance of the knee-jerk during the first part of 
the tracing shown. At R the hammer fell out of time, inter- 
rupting the regular 5 seconds interval, and as a result of this 
disturbance of the regularity of the blows the knee-jerk rose 
during the next four kicks. At R2 the rhythm of the blows was 
intentionally interrupted, with the result of causing a rise in the 
knee-jerk again, but this time less than at first, showing that 
the nervous system had become accustomed to this change of 
rhythm. At R3 the blows were delivered in as quick succes- 
sion as possible, causing a much greater increase of the knee- 
jerk than on either of the other two occasions ; after eight 
blows given at the regular 5 seconds interval the knee-jerk 
again sank to 0. 
Rhythmic Grouping. 
Besides showing the peculiar prolongation of the effect of a 
sensory stimulus Fig. 1 shows also the peculiar "groups" 
of kicks that appear in the curves with no apparent auditory 
stimulus to account for them. One is tempted to speak of 
these groups as falling into a rhythm, but they do not occur 
under circumstances justifying one absolutely in making this 
claim. Yet looking at this curve as a whole it is difficult not PECULIARITIES OF THE KNEE-JERK. 
11 
to think that there must be some rhythmic periodical activity 
of the body to produce these wave-like rises and falls. It 
will be seen that there were four of these groups before the 
auditory stimulus was given at A and the blows at the begin- 
ing of the experiment appear to have begun at the top of one 
of these crests. After the effect of the stimulus given at A 
had disappeared two of these groups came on before the 
second stimulus was given at B, and after the effect of this 
had disappeared the groups continued to appear to the end 
of the experiment. It is to be noted in this connection that 
the reinforcement at A came at about the time when a 
"group" might have been looked for, and the question 
arises did the auditory stimulus simply intensify one of these 
periodical rises in the knee jerk. At none of the places where 
these rises came on, except at A, B and C were there any 
reinforcing stimuli to account for them. At B also the 
stimulus appears to have come at about the time when a rise 
was due, while at C the stimulus seems to have come between 
two groups, and this may account for the fact that there was 
a shorter effect of the stimulus here than at A and B. 
In examining the remaining diagrams showing the prolonged 
effect of sensory stimuli to see if one of these " groups " may 
have entered as a disturbing factor the only instance where 
this could be thought to do so is in Fig. 4 where the second- 
ary rise after B may possibly be looked on as a "group," 
but the prolonged reinforcing effect of the stimulus at B is 
nevertheless sufficiently well marked. The "grouping" is 
again well shown in Fig. 8. The series is taken from the 
middle of a tracing after the patient was fully asleep. There 
was an interval of seven minutes between the first kick and the 
preceding one, and the first blows after this interval of rest 
naturally acted as a reinforcement, but the kicks soon 
diminished, and the peculiar wave-like movement of the curve 
developed. There were no sensory reinforcements from 
sounds and the subject seemed to be in the soundest sleep. 
The same phenomenon is seen again in Fig. 9, where there 
were no sensory stimuli except at 1 where two taps were given, 
and at C where a passing train of cars was heard. This shows 
also that the two taps did not invariably call out the pro- 
longed reinforcing effect that usually followed. 
It was possible on some occasions to account for the changes 
by variations in the depth of sleep, as in Fig. 10. Here at 
the points marked A there were no auditory stimuli to 
account for the rises, but in the intervals, at the points 
marked B, there were audible evidences of the soundness of 
the sleep. Usually however it was not possible to detect such 
a change, and the waves rose and fell without any noticeable 
change in the regularity of the respiration. 12 
NOYES: 
The question as to the cause of this rhythmic (?) grouping 
of the knee-jerks in sleep is a very interesting one, and con- 
siderable attention was given to it in attempting to find an 
explanation. The phenomenon was noticed very early in the 
experiments, and many of the tracings made while the 
subject was awake suggest this same tendency to a 
periodicity. The "grouping" always shows best, however, 
in a condition that might be termed half sleep, where the 
subject is to all appearances sound asleep, but that he is not 
in the profoundest sleep is shown by the fact that his knee- 
jerk is not entirely abolished. Attempts were made to find 
some connection between these "groups" and the respiratory 
rhythm, and the respiratory curve and the knee-jerk curve 
were taken simultaneously, but the results were entirely 
negative, no particular length of kick being found associated 
with a particular phase of the respiratory curve. The attempt 
was also made by taking a plethysmographic tracing from the 
arm, to find some connection between the depth of sleep and 
the variations in the knee-jerk, but these at first were equally 
unsuccessful. Later it was suggested that there might 
possibly be some connection between the groups and the 
Traube-Hering curves, and acting on this suggestion addi- 
tional plethysmographic tracings were taken. A glass 
plethysmograph, suspended from the ceiling to allow free 
movement, was placed on the patient's left arm, and connec- 
tion was made through a glass tube having rubber joints 
with a very sensitive Marey tambour, the writing-needle of 
which was placed directly over the writing-needle of the 
knee-jerk apparatus, so that the two curves were made syn- 
chronously on the revolving cylinder of the kymograph. The 
Traube-Hering curves did not always appear, and there were 
also many times when the peculiar "groupings" did not 
appear, as it was necessary for the patient to be in the con- 
dition of half-sleep already alluded to, and also that there 
should be few or no disturbing noises. The necessary condi- 
tions have been fulfilled on repeated occasions, however, 
and a series of tracings obtained where there is a good 
Traube-Hering curve and also a good series of "groups" of 
knee-jerks. A portion of one of these double curves is shown 
in Fig. 16. The pulse beats are well marked and the res- 
piratory rhythm is also well shown. The jarring of the blow of 
the hammer was sufficient to set the needle of the Marey 
tambour violently oscillating, so that the respiratory rhythm 
appears to be unduly accentuated, but this sharp upward rise 
is due to the vibration of the needle. The Traube-Hering curve is 
also well marked, the tops and bottoms of the waves being 
connected by straight lines. It will be seen that there is an PECULIARITIES OF THE KNEE-JERK. 
13 
apparent coincidence between the two curves-that the Traube- 
Hering curve descends lowest in that part of the 4 < group ' ' 
where the kicks are longest, and at places where the Traube- 
Hering curve is highest the knee-jerks are much diminished. 
A rise in the Traube-Hering curve means of course increased 
blood pressure in the arm, and a fall in the curve corresponds 
to diminished blood pressure. On Mosso's theory that in- 
creased blood pressure in the extremities means lessened 
blood pressure in the central nervous system we should have, 
during the time that the Traube-Hering curve is at its height, 
relative anaemia of the brain and cord ; and during the time 
that the Traube-Hering curve is lowest relative hyperaemia 
of the brain and cord. There are objections to Mosso's theory, 
however, as it fails to take into account the abdominal circu- 
lation, and the possibility that a change of blood pressure in 
the extremities may mean simply an opposite change in the 
abdominal cavity and not in the central nervous system. 
Could we adopt Mosso's view it would simplify the problem 
greatly to say that if we got a rise in the Traube-Her- 
ing curve this would mean a diminished blood supply to 
the brain and cord, and a fall in the Traube-Hering 
curve would mean a corresponding increase of the blood 
supply of the central nervous system. The diminished 
knee-jerk would then naturally follow from the lessened 
functional activity of the spinal cord at the height of the 
peripheral Traube-Hering wave, while an increased knee- 
jerk from increased functional activity of the cord would 
follow at the low phase of the peripheral Traube-Hering curve. 
The occurrence of the high phase of the Traube-Hering curve 
with a diminished knee-jerk, and of the low phase with 
an increased knee-jerk has been noticed with sufficient 
frequency to give consi ierable probability to the theory 
that there may be a constant relation between the two. 
The Traube-Hering curves shown in Fig. 16 demonstrate 
that there is a rhythmic rise and fall in the blood pres- 
sure of the arm, as has been frequently proved before. 
There is thus naturally good reason to infer that with 
this rise and fall occurs throughout the whole vascular 
system, and that the vascular supply of the central nervous 
system is subject to this same periodicity. It also seems per- 
fectly fair to assume that this rhythm might not necessarily 
be the same throughout the whole vascular system of the 
body. We only need assume that the vaso motor centre in 
the medulla sets up the rhythmic contractions and dilatations 
in the vascular system that show themselves in the periphery 
in the Traub-eHering curves, but this rhythmic influence 
would not necessarily propagate itself throughout the whole 14 
NOYES: 
body within the same time, the vascular system possessing 
considerable inertia, and the amount of blood to be influenced 
being so great. Several rhythms differing in time might easily 
be present in the vascular supply of different portions of the 
body, dependent on the different rates at which the vaso-motor 
influence had propagated itself through the arterial system; 
such rhythms should show the same general characters with 
regard to rise and fall. Fig. 17 shows the result of assuming 
that this rhythm in the central nervous system differs a little 
in time from the rhythm in the peripheral circulation. In 
Fig. 17 the Traube-Hering curve shown in Fig. 16 has been 
moved back a distance representing 20 seconds in time, which 
would mean that the vaso-motor influence affected the blood 
supply of the central nervous system 20 seconds before it 
reached the arteries of the periphery, which seems a not im- 
probable supposition. Although even with this change, the 
crests of the Traube-Hering wave do not absolutely coincide 
with the points of the greatest diminution of the knee-jerk 
curve, and vice-versa, yet the coincidence is much more 
striking, and additional weight seems to be given to the infer- 
ence that there may be some connection between the two 
curves. It will be noticed that at one point of the knee-jerk 
there was a disturbing factor caused by the slamming of a 
door below, at A, sending the knee-jerk up at this point, and 
thereby apparently making the summit of this knee-jerk wave 
farther along than it otherwise would have been. 
In Fig. 18 there is a much closer coincidence in the two 
waves than in Fig. 16. Here the Traube-Hering curve makes 
a long descent between 4 and 5, with a still longer ascent to 6. 
Again it is to be noted that at 4, where the Traube-Hering wave 
descends lowest, the knee-jerk curve is also longest. 
The same general coincidence of the two curves is again 
seen in Fig. 19. This was one of the earliest tracings, and the 
writing needle of the Matey tambour, connected with the 
plethysmograph, did not make as good a record as it did later. 
At the end of the record the patient fell into his deepest sleep 
with the entire disappearance of the knee-jerk. It is interest- 
ing to note that, although the Traube-Hering curve continues 
after the knee-jerk has entirely disappeared, yet at no point 
does the Traube-Hering wave descend as low as at A, where 
there was the longest group of kicks. As in Figs. 16 and 17, 
there is the same number of waves in the knee-jerk curve as 
in the Traube-Hering curve. 
Fig. 20 is from another of the early tracings, but it serves 
to show the same general characteristics that have been 
brought out by the other curves. At A a secondary wave 
appears on the long descent between 6 and 7, and slight cor- PECULIARITIES OF THE KNEE-JERK. 
15 
responding changes may be noted in the knee-jerks below. 
All the plethysmographic tracings given were taken from 
the left arm. As it would be valuable confirmatory evidence 
to obtain similar tracings from the leg, a tin plethysmograph 
in the shape of a boot was made, and in this the patient's right 
foot and leg were placed ; the rubber band that served to keep 
back the water coming just below the knee. (Into each 
plethysmograph water of 30° C. was poured to take up the 
extra air space not occupied by the leg and arm. The water 
did not quite fill the plethysmographs, a small air space being 
left at the top to allow free access of the air to the opening 
communicationg with the glass tube and Marey tambour. A 
much better tracing was made by the writing needle with the 
water in the plethysmograph than with this empty, as the 
pulsations were thus confined to the small body of air directly 
above the water, and the arc of vibration of the needle corre- 
spondingly increased). The leg plethysmograph was sus- 
pended from the ceiling in the same way as that for the arm, 
and it was still possible to have the left leg, which was still 
the upper one, in the same position as before for the knee-jerk 
experiments. Tracings were then made showing synchronous 
right leg and left arm plethysmographic curves and the left 
knee-jerk curve. It was found, however, that the jarring of the 
blow of the hammer on the left leg communicated itself 
through the bony pelvis sufficiently to affect the right leg in 
the plethysmograph, causing a serious vibration of the needle 
with each blow, interfering with the production of a good 
curve. The arm and leg plethysmographic curves were then 
taken without the knee-jerk curve with better results, and a 
general correspondence was found. The leg plethysmographic 
curve was never as satisfactory as that of the arm, for 
it was possible to place the whole arm in the plethysmo- 
graph, the strong pulsation of the radial artery being strong 
enough to give a well-marked tracing, but no similarly strong 
pulsation could be obtained from the foot and calf of the leg. 
While, therefore, the experiments with the leg plethysmo- 
graph need to be carried farther, yet so far they point to 
similar results as with the arm. 
Should the conclusions suggested by the knee jerk and 
plethysmographic curves seem to be justified, and should 
they be borne out by further research, the knee-jerk would 
thus be brought into connection with the other rhythmical and 
periodic activities of the body. The vaso motor influence that 
produces the Traube. Hering curves is necessarily constantly 
active, but its effects are usually obscured by many other con- 
ditions. It would appear probable on a priori grounds that 
the Traube-Hering curve would come out more clearly where 16 
NOYES: 
the cerebral influence was removed or inhibited, and we find 
the Traube-Hering curve coming out with marked?distinct- 
ness in our demented subject. For the same reason we should 
expect any phenomenon associated with the Traube-Hering 
curve also to come out better in such an individual than in a 
normal subject, and so we find the "groups"-if the relation 
to the Traube-Hering is a true one - coming out in this same 
patient. As the Traube-Hering curve is constantly influen- 
cing the normal respiratory rhythm, may we not also assume 
that the Traube-Hering knee-jerk curve, if it is permissible to 
call it such, is also constantly influencing the knee-jerk? This 
would explain the mysterious " rhythm" that has seemed to 
be present in many of the earliest knee-jerk curves taken in 
this patient, even when awake. We are led from this to a 
consideration of the knee-jerk of normal individuals, and if 
the inferences as to the influences affecting the knee-jerk in 
this demented man are legitimate, it is not evident why the 
same inferences do not apply to the normal subject. If this 
is the case then the original point of the investigation no 
longer has any bearing,-that is, the question as to the 
theoretically lesser variability of the knee-jerk in a demented 
person than in a healthy individual; the tops of the knee-jerks 
of a dement forming theoretically a straighter line than in a 
sound person. The "normal" knee-jerk curve,* therefore, 
could no longer be considered as theoretically a straight line, 
but as a true curve correspondingin general with the Traube- 
Hering curve. It must be admitted at once that it is extremely 
doubtful if this can ever be shown on a normal individual with 
the constantly varying emotional condition of healthy persons ; 
nor does it seem scarcely more likely that a normal individual 
will show such curves even in sleep as are seen in this patient, 
for the reason that the cerebral influences in a sound person 
would probably mask this ebb and blow ; but this is mere con- 
jecture and must be submitted to actual experiment. 
It remains to add that if the "nervous force" or "irrita- 
bility" of the spinal cord is really subject to this rhythmic 
action, the question is at once raised if the higher cerebral 
activities, especially the attention, are also subject to a simi- 
lar rise and fall, for if the rhythm (?) already described be 
really due to a vascular process of vaso-motor origin, this 
same influence must affect also the functional activity of the 
brain itself. 
Conclusions. 
The chief interest in the results brought out in this paper 
lies in the fact that the experiments were conducted on a per- 
son whose mind has been weakened by dementia of many years PECULIARITIES OF THE KNEE-JERK. 
17 
duration, and that this individual shows a greater suscepti- 
bility to sensory stimuli than persons in health. The most 
reasonable explanation of this seems to be that there has been 
a weakening of the inhibitory influence normally exerted by 
the brain over the lower centres. The remarks of Mitchell 
and Lewis1 have so much bearing on this point that they may 
properly be quoted here. In discussing the cause of the 
increase of the knee-jerk from sensory stimuli, they say "It 
is very difficult to explain the fact that electricity, sensory 
impressions, and distant voluntary muscle acts increase the 
knee-jerk and the response to the muscle blow. If we conceive 
of a series of inhibitory centres extending from the meso- 
cephalon all the way down the cord, and infer that all the 
agents mentioned are capable, by more or less paralyzing these 
centres, of releasing the active reflex groups below them, we 
shall be able to comprehend that the centre thus set free may, 
by increasing tone, give to the muscle a suddenly enlarged 
capacity to respond to the tendon taps or the muscle blow. 
Nearly all the facts with which we are concerned may be 
explained by inhibition organs and the effects produced upon 
them. On the other hand, it is equally conceivable that when- 
ever a sensation reaches the cord or brain or both, an over- 
flow occurs, which shall, by increasing the excitation of the 
centres, be felt throughout the body, and reinforce any organs 
chancing to be synchronously otherwise excited from without. 
Under this view we conceive of the nervous force as not confined 
entirely to the direct paths between the centres and the muscle 
to be moved, but as overflowing so as to pass through numer- 
ous ganglia, adding a certain small increment to their effects 
when in a state of such activity as the spinal toning centres 
must be at all times. The tone centres thus stimulated send 
out a higher wave of excitability to all the muscles, and if at 
the time this reaches a muscle, that muscle is being excited 
by a tap, there is an increased response." 
Assuming such inhibitory centres as Mitchell and Lewis 
describe, these would only be kept up to their full functional 
activity by the healthy condition of the whole nervous system, 
so that when these inhibitory centres are under a weakened 
cerebral influence, as in dementia, they offer less resistance to 
what Mitchell and Lewis call the paralyzing influence of sen- 
sory stimuli. In the normal healthy individual in sleep a sen- 
sory stimulus preceding a blow on the patellar tendon pro- 
duces a rise extending at most over a few kicks. This is 
certainly the most economical method for the individual, if we 
consider that it is beneficial that the effect of the accidental 
stimuli that are continually assaulting the nervous system 
1 Loc. cit. p. 203. 18 
NOYES: 
should be prevented from spreading over a wide territory or 
through a long time. In our demented patient a stimulus 
produces an effect extending usually over a much longer time, 
even so long as three minutes. 
The experiments appear to give a graphic demonstration of 
the greater susceptibility to sounds, and all external stimuli, in 
persons with enfeebled, but not organically diseased, nervous 
systems. If this susceptibility to long continued effect 
of sensory stimuli in this patient is really due to weak- 
ened cerebral inhibitory power, it seems not improbable 
that the same effect must be produced in individuals 
with brains weakened or exhausted from any cause,-not 
necessarily from insanity. The large class of neurasthenic 
individuals naturally first occurs to one, and the question 
arises whether, when such patients ' ' feel every sound in 
their back," this is not due to the uninhibited propa- 
gation of accidental external stimuli to the lower reflex 
centres, as shown in these knee-jerk curves. If so, the thera- 
peutic corollary to this proposition, the necessity of excluding 
to the greatest possible extent all accidental stimuli from such 
patients, is graphically demonstrated in these curves. 
The conclusions with regard to the so-called rhythm have 
already been discussed. 
The results of the experiments may be summarized in the 
two following propositions : 
In a case of terminal dementia of many years duration a 
series of experiments on the knee-jerk tend to show that: 
1st. Sensory stimuli received during sleep produce a much 
greater effect and diffuse over a much longer interval than in 
healthy individuals. 
2nd. In a condition of half-sleep when the patellar tendon 
is struck by blows of uniform strength at five seconds inter- 
vals, the knee-jerks fall into groups, and synchronous 
plethysmographic tracings suggest that these groups have 
some connection with the Traube-Hering curve. 
If the truth of the second proposition can be conclu- 
sively established, several important corollaries would seem 
to follow. These are here stated as facts for the sake of 
presenting definite propositions, the truth or falsity of which 
must be submitted to further experimental investigation. 
(a) The knee-jerk curve, instead of being theoretically a 
straight line as has been heretofore assumed, is, in reality, a 
curved line, with the general characteristics of the Traube- 
Hering curve. 
(b) The spinal cord is not constantly in a condition of the 
highest potential functional activity, but its activity is repre- 
sented by a curve of rhythmic vascular contraction and dila- PECULIARITIES OF THE KNEE-JERK. 
19 
tation. During the phase of contraction of the spinal arteries, 
the spinal cord is at its least functional activity, due to a con- 
dition of relative anaemia, while during the phase of dilatation 
of the spinal arteries, the spinal cord is at its greatest func- 
tional activity, due to a condition of relative hypersemia. 
(c) The question inevitably raised by (b) is whether the 
higher activities of the brain are also subject to a rhythmic 
rise and fall synchronous with vascular dilatation and con- 
traction. 
It remains to express my obligations to several who have 
assisted me in the details of the work, and to extend my 
thanks to my colleague, Dr. D. H. Fuller, and to Drs. Abbot, 
Young, Fitz and Sawyer, medical house-pupils, for their 
assistance. To Dr. Fitz I am especially indebted for assis- 
tance and suggestions in regard to apparatus. Ttt/.I 
F(g.7/. 
FIG.XH 
A B 
C A B C 
TiaV 
Fic/Vtr 
F7gV/// 
Fig.V7 
"£'\q®E.. 
A B q 
AB 
1 ■? 3 15677 
figIX 
TiyX 
A!\qX5L 
T C 
R R2 R3 
T C 
A B A B AB 
TigXI. 
FIG.XV FIG. xxznt 
FIG ZEE 
FIG.ZK 
FIG XW 
figk