Download Analysis of Golgi Tendon Organ Responses to Muscle Contraction and Electrical Stimulation and more Summaries Biophysics in PDF only on Docsity!
J. Physiol. (1981), 318, pp. 429-443 429 With 7 text-figures Printed in Great (^) Britain
RESPONSES OF ISOLATED GOLGI TENDON ORGANS OF THE CAT TO
MUSCLE CONTRACTION AND ELECTRICAL STIMULATION
BY YASUSHI FUKAMI
From the Department of Physiology and Biophysics, Washington University School of
Medicine, 660 South Euclid Avenue, St. Louis, Missouri 63110, U.S.A.
(Received 17 November 1980)
SUMMARY
1. Responses of Golgi tendon organs isolated from cat tail muscles to contraction
of muscle fibres inserting directly into the receptor (GTO-muscle fibres) as well as to
pulses of electrical current applied extracellularly through the sensory axon were
studied.
2. Analysis of the responses to GTO-muscle fibre contraction indicated that the
active force developed by each muscle fibre constituted an equally potent input to
the receptor in^ proportion to^ the^ developed force.
- The sensitivity of a Golgi tendon organ remained almost constant with changes
in muscle length up to a length where maximum active tension was developed (lo).
Beyond lo, the sensitivity tended to decrease.
4. The absolute force threshold (passive + active) at lo for initiating an impulse in
the afferent nerve was estimated for five preparations to be 45-14 mg. It was also
demonstrated that the contraction of a single GTO-muscle fibre may initiate impulse
discharge from the receptor.
5. A constant depolarizing current applied extracellularly to a GTO through its
axon initiated a train of impulses, probably originating from a site near or within
the receptor capsule. Analysis of responses to constant currents of various intensities
suggested that a single impulse initiation site was involved.
6. During combined stimulation, responses of a tendon organ to GTO-muscle fibre
contraction simply added to the response initiated by a constant current pulse,
suggesting that the impulse initiation sites activated by each mode of stimulation
were identical, or situated very close to each other in the nerve terminal.
INTRODUCTION
The Golgi tendon organ or^ GTO, one^ of the^ major receptors involved^ in^ the central
control of muscle contraction, is a tension receptor that responds to contractions of
certain number of motor units whose muscle fibres lie directly in series with the
receptor (Houk & Henneman, 1967; Houk & Simon, 1967; Reinking, Stephen &
Stuart, 1975; Jami^ &^ Petit, 1976). Based^ on^ the^ analysis of^ GTO^ responses to
contraction of these motor units, the threshold force and sensitivity of GTOs have
been calculated (Houk, Singer & Henneman, 1971; Binder, Kroin, Moor & Stuart,
430 Y.^ FUKAMI
1977). Based on the number of motor units acting^ on^ a^ single^ GTO in cat soleus^ and
the number of muscle fibres^ in^ series with^ each receptor, Houk & Henneman^ (1967)
have suggested that a motor unit contributes only one or two muscle fibres^ that
actually insert into^ a^ tendon^ organ^ (also^ Reinking et al. 1975; Jami^ &^ Petit,^ 1976).
More recently, Gregory & Proske (1979) examined the responses of GTOs to^ combined
stimulation of^ different^ motor units^ acting on single receptors and demonstrated^ that
the relation between firing^ rate^ and tension for^ combined stimulation^ may^ differ
markedly, depending on the combination^ of motor^ units^ stimulated.^ Based^ on^ this
and other findings they^ suggested^ that^ the response^ seen^ in^ the^ parent axon^ reflects
the interaction of impulses originating^ from^ multiple^ sites^ along^ the^ sensory^ ending.
Since the above^ studies^ on^ GTOs^ have used^ whole^ muscle^ preparations^ in^ situ, these
approaches are obviously indirect. A^ more^ direct^ approach^ is^ to^ employ^ isolated
preparations where various experimental parameters^ may^ easily^ be^ controlled.
Single tendon organs with^ or^ without^ muscle fibres^ inserting^ into them^ can^ be
isolated from cat tail muscles, and will survive^ well in vitro^ for^ many^ hours.
Information obtained from^ isolated^ single GTOs^ without^ muscle fibres^ attached^ has
already been published (Fukami & Wilkinson, 1977). This^ report^ deals with^ analysis
of responses of^ isolated^ single GTOs^ to^ GTO-muscle^ fibre contraction^ as^ well^ as^ to
constant current pulses applied extracellularly^ through the sensory^ axon.^ Interaction
between the^ responses to^ these^ two^ modes of^ stimulation^ was^ also^ examined. The
results provide information on^ threshold^ force for^ initiating^ an^ impulse^ from^ the
receptor, the sensitivity as a function of^ muscle^ length,^ and^ the^ site^ of^ impulse
initiation in the sensory terminal. Evidence^ is^ also^ presented^ that single^ GTO-muscle
fibre contractions can initiate impulse discharge from the^ receptor.
METHOD Preparation Cats were anaesthetized by i.P.^ injection of^ pentobarbitone^ (30^ mg/kg).^ Immediately^ following transaction of the^ skinned^ tail^ from the^ rest^ of the^ body,^ the^ dorsolateral^ tendon^ was^ retracted laterally to^ expose muscles^ to^ the^ bathing^ solution^ (composition^ in^ mM:^ NaCl^ 145, KCl 1-5,^ CaCl 1-8, MgSo4 1 0, KH2PO4^ 1-0,^ HEPES buffer^ 5-0, glucose^ 5-0,^ pH^ adjusted^ to^ 7-4^ at^ room temperature). Under^ a^ dissecting^ microscope^ single^ tendon^ organs^ at^ the^ musculotendinous^ junction can be identified. A part of the muscle containing a^ Golgi tendon^ organ^ and its nerve was^ dissected out (^) together with a piece of the tail bone from which^ the^ muscle^ originates;^ the^ preparation^ was transferred to another small chamber where^ further dissection and^ experiments^ were^ performed. Using fine^ needles and scissors,^ a^ single^ GTO^ together^ with its^ nerve^ was^ first^ isolated^ from^ the surrounding connective tissue and tendon. Isolation of muscle^ fibres^ inserting^ directly^ to^ the receptor (GTO-muscle^ fibres)^ then^ proceeded^ from the tendon^ side^ toward^ their^ origin.^ The^ bathing solution was changed several times during the^ course^ of this dissection. After^ isolation,^ the^ piece of tail bone was firmly fixed by magnet-based pins to^ the bottom of the^ chamber, the end of^ the tendon tied to a strain gauge (sensitivity 0-5^ 1V/mg,^ compliance^20 4m/g,^ resonant^ frequency 800 Hz, drift+ 1-5 ,uV/hr) mounted^ on^ a^ micromanipulator,^ and the^ nerve^ lifted into oil^ for recording impulse activity using a^ pair of Pt^ wire^ electrodes. Another^ pair^ of Pt wire^ electrodes was placed on GTO-muscle fibres^ for^ stimulation. Stimulation and^ recording To stimulate GTO-muscle^ fibres,^1 msec^ electrical^ pulses^ of^ various^ intensities^ were^ applied directly to^ the^ muscle^ fibres^ through^ a^ pair^ of^ stimulating^ wire electrodes^ which,^ in order to^ reduce stimulus artifacts in^ nerve^ records,^ were^ placed^ away^ from^ the receptor,^ usually^ near^ the muscle origin. The^ technique^ employed^ for^ electrical^ stimulation of the^ sensory^ ending^ was^ similar^ to^ that
Y. FUKAMI
Tension (mg) 0-12rC
0-08 1-
004 -
0' -^ I^ I^ I^ I^ I 40 80 120 Time (msec) Fig. 1. Impulse discharge from an isolated GTO-muscle preparation in (^) response to various levels of tetanic contraction of GTO-muscle fibres. A, three examples of single trace records where upper traces represent impulses and lower traces tension. Each impulse is marked by a dot. Other spikes are stimulus artifacts. The intensity of muscle stimulation at 32 Hz increased from top to bottom. Time calibration, 0 5 sec. Tension calibration, 100 mg. B, the relation of plateau tension vs. discharge rate. Crosses were obtained at a just taut muscle length (- 10 mm, Al =^ 0), open circles at 3 mm longer length (Al =^3 mm), and filled circles at 5 mm longer muscle length (Al =^5 mm) at which a sustained spontaneous discharge occurred. Correlation coefficient by rank (^) (rs), 1-0 for crosses, 0-975 (P [two-tail t test] < 0-001) for open circles, and^ 0-886 (P < 0-001) for filled circles. A regression line calculated for each (^) group of data within (^90) mg or less tension is also (^) shown. The (^) slope of these lines (sensitivity) is 240 i.p.s./g for the continuous line (correlation coefficient, r = (^) 0974), 250 i.p.s./g for the dashed line (r = (^0) 987), and 147 i.p.s./g for the dotted line (r =^ 0982). C, plot of coefficient of variation (C.V.) against mean intervals estimated by measuring 5-12 interimpulse intervals in single trace records at various plateau tension. Al =^3 mm.^ rs =^0 90, P^ <^ 0-001.
A
I I tj1 (^) _~~~~
B
50
0
1 I I
1 60 msec
Iiiiii
nmv-~~~~~
RESPI'ONSES OF^ GOLGi^ TENDON^ ORGAN^433
Gregory &^ Proske^ (1979)^ from the^ experiments^ in GTOs in^ medial^ gastrocnemius^ of
cat by various combinations oftetanic stimulation of single motor units^ each ofwhich,
when stimulated singly, evoked a response from a GTO. This discrepancy is possibly
due to the active force developed by muscle fibres of single motor units not in series
but in parallel with the GTO, thus decreasing the sensitivity measured. Beyond 10
the sensitivity tended to decrease. For example,^ the^ sensitivity of a GTO beyond^10
decreased progressively from 281P6 to 160 i.p.s./g at 1-3 10 and 100 i.p.s./g at 1-5 l0.
This decrease may be ascribed to partial Na-channel inactivation at the impulse
initiation site by prolonged depolarization of the sensory terminals and/or saturation
of the receptor potential generating mechanism. Or, it may be mechanical in origin;
if the compliance of a GTO decreases progressively beyond a certain level of tension,
this would result in progressively less nerve responses to a given change in tension
because of less extension of the nerve ending for a given increment of^ tension.
mg
200 L 100 o (^) S (i.p.s./g)
100 Nu- , (^100)
0 I aftI 0 I 0 1 2 3 4 5 6 7 l (mm) Fig. 2.^ Sensitivity and threshold^ active^ force^ for^ initiation of^ an^ impulse as a^ function^ of muscle (^) length. 0, x, maximum tetanic tension and passive tension, respectively (left
ordinate, upper figures). 0, sensitivity, 8, in^ i.p.s./g (right ordinate). *, threshold^ active
tension (left ordinate, lower^ figures). Abscissa, muscle^ length (Al) relative^ to^ the^ just taut
length (= 10 mm).
In another series of experiments the threshold active tension^ for^ initiation^ of^ an
impulse was^ measured^ during^ the^ rising^ phase^ of^ a^ twitch^ contraction^ at^ the time when an impulse was initiated. In^ the^ GTO-muscle preparation shown^ in^ Fig. 2 the threshold (^) (U) remained (^) relatively constant over the muscle lengths examined. The
absolute threshold (Binder et^ al.^ 1977) measured^ at^1 for^ five^ preparations
(active + passive tension in^ mg)^ was^ 1-5^ +^ 3,^ 2-8 +^ 10,^ 11-3^ +^ 3,^ 20 +^ 2 and^ 8X0^ +^ 6. In order to examine whether the^ active^ tension^ developed by a^ single GTO-muscle
A
RESPONSES OF GOLGI TENDON ORGAN
D
- -0^ ^--R''-XS1|-'a'1ld'-|e--aam--
B
A_ ..4' I
F
-j
Fig. 3. Impulse discharge from a GTO in response to intracellular stimulation of (^) a single GTO-muscle fibre. Records A, B and C represent respectively (^) responses of a (^) GTO-muscle preparation to intracellular stimulation (^) of three different muscle fibres. In A, a muscle action potential shown in the bottom trace (^) (stimulus artifacts seen as downward and upward deflexion preceding the action potential) caused a twitch contraction (middle trace) and an impulse discharge from the organ (top trace). In (^) B, the first (^) stimulation of a second fibre failed to initiate any response whereas the second stimulus produced a (^) twitch (bottom trace) and an impulse (top trace). In C, the first twitch contraction initiated an impulse but the second one failed to do so. In B and C intracellular records were omitted. Resting potential -65 mV for all three fibres. Records D and E represent records taken from two out of three impaled fibres of another GTO-muscle preparation. In D, the intracellular (^) electrode following a muscle action potential got out of the cell (bottom trace) causing muscle^ contracture (middle trace) and a train of impulse discharge from the organ. Note that the same stimulating current when applied extracellularly caused no response at all. Inset record in E shows the muscle action potential at a faster time base (time calibration, 10 msec). Resting potential -80 mV (^) for both fibres. In record F, in response to suprathreshold stimulation of a muscle fibre, two muscle (^) action potentials (middle trace) and two nerve impulses from the organ (top trace) were (^) initiated for each trial without any noticeable tension development (bottom trace). Four (^) single trace records were superimposed. Resting potential -70 mV. Time calibration (^) in C (250 msec) applies to B, (^) C, D and E. (^) Tension calibration in C, 5 mg for A, B, C and E and 10 mg for D. Volt- age calibration in D for intracellular recordings, 100 mV for A and E and 50 mV for D. Time calibration (^) in A and F, 10 msec. Vertical bar in F, 50 mV for intracellular recording and 5 mg for tension.
E
M_._.A- (^) --- (^) 0-1111-1-MO
I
435
I
Y. FUKAMI
A
I L
8 ....^.. ..-. ......... ....- 4 -
0
L 0 20
D
40 Time (sec)
I
60
(^008) [-
d 0041
L 101
I (^) I I I I I l0 120 140 160 Time (msec) Fig. 4. Impulse discharge from an isolated Golgi tendon organ in response to a constant current applied through the innervating axon. As shown in the upper record in A impulse discharge continued^ throughout the^ entire^ period (- 80 see) of^ current^ application (lower trace in A). Time and current calibration in A, 30 see and 10-7 A, respectively. In B the mean discharge rate determined by counting the number of impulses during consecutive 1-5 see periods is^ plotted on^ the^ same^ time^ scale^ as^ in A.^ C, mean interimpulse intervals of ten consecutive (^) impulse intervals measured (^) up to (^60) see from the start of current application. D, scatter plot of coefficient of variation (^) (C.V.) of ten consecutive (^) interimpulse intervals against mean impulse intervals (^) during the same (^) period as in C. Correlation coefficient by rank, (^) r. =^ 0-64, P (two-tail t (^) test) < 0-01.
8
160 C
EE 140k E (^120) -
100 _ *
0
been reported for (^) many other sensory receptors including frog muscle spindles (Buller,
Nicholls & Strom, 1953; Buller, 1965; Brokensha^ &^ Westbury, 1974), snake muscle
spindles (Fukami, 1970) and optic nerve fibres of Limulus (Ratliff, Hartline & Lange,
1968). To assess the variability of interimpulse intervals, the coefficient of variation
for each data point in Fig. 5B was plotted against the mean interval (Fig. 5D). There
were no obvious abrupt changes in the relation between these two parameters, the
coefficient of variation increasing smoothly with the increase in mean intervals
(rs =^ 0.81,^ P^ <^0 01).^ As^ discussed^ by^ Brokensha &^ Westbury^ (1974),^ ifthe relationship
between the^ variability and the^ mean rate^ of^ discharge (or interimpulse intervals)
were different for each impulse generator, then the lack of any abrupt change in this
relation would indicate a single origin of the impulse train.
Interaction between responses to muscle contraction and to electrical stimulation
The above results raise^ the^ question of whether^ impulses elicited by active muscle
force and those initiated by electrical pulses do originate from the same site or from
different sites along the sensory ending. In^ the former^ situation where the same site
is involved, depending upon the distance between the impulse initiation site and the
sensory transduction site in the terminal, two possibilities may be considered. If the
two sites were situated close to each other, then the responses during combined
stimulation might be less than simple addition of responses to each mode of
stimulation because of the increased conductance during the receptor potential which
would short circuit the applied current. In contrast, if the two sites were situated
far apart, then the response to combined stimulation^ would be a simple summation
of responses to each mode of stimulation. In the latter situation where more than
one source is involved, one may expect to see evidence for interaction of impulses
originating from two independent sources such as resetting (Matthews, 1931 Fukami,
1980), increase in^ both^ mean^ frequency and^ regularity of^ discharge (Eagles &^ Purple,
1974; Fukami, 1980), and/or irregular discharge due to impulse collision along
terminal branches or probabilistic mixing of impulses (Clifford & Sudbury, 1972;
Fukami, 1980). The above possibilities were^ tested by independent analysis of the
response to each mode of stimulation, and analysis of the response to combined
stimulation. It was found that the discharge rate in response to GTO-muscle fibre
contraction simply added to the response to electrical stimulation. Example records
of this experiment are shown in Fig. 6 where the top record A shows the response
to a constant current, record B the response to a tetanic contraction of GTO-muscle
fibres, and the bottom^ record^ (A + B) the^ response during combined stimulation.
During combined stimulation the rate of discharge increased gradually, reaching a
rather regular firing rate which appears higher than that of the response shown either
in A or B. This finding suggests the absence of both resetting and impulse mixing
during this period. For further analysis, instantaneous frequency of discharge in
response to^ each mode^ of stimulation^ as^ well^ as to^ combined stimulation^ were^ plotted
as a function of time measured from the onset of current application. Two sets of
examples are shown in Fig. 7 where each graph represents superimposed plots of
measurements made on two to four single trace records of the response to electrical
stimulation, the response to tetanic contraction of GTO-muscle fibres and the
response to combined stimulation. As seen in this Figure, the response to tetanic
438 Y. FtTKAMI
RESPONSES OF GOLGI TENDON ORGAN 439
contraction during combined stimulation, was simply superimposed on top of the
response to electrical stimulation. To^ examine^ this^ point further, the^ average rate
of discharge was determined during the period of 800-1200 msec following the onset
of current application when^ the^ response to^ tetanic contraction^ appears relatively
stable. The value thus obtained for the response to combined stimulation was
compared with the value for the^ response to^ tetanic contraction alone^ plus that for
A
A (^) +8B
Li LLLL
4*4 - - - I. -- -. ... .............
Fig. 6. Sample records illustrating the response to applied current (A), the response to tetanic stimulation at 25/sec of GTO-muscle fibres (B) and the response to combined stimulation (A + B) of an isolated GTO-muscle preparation. Upper trace^ in^ each^ record represents axonal discharge. Time calibration, 0-5 sec.^ Vertical calibration^ bar, 9 x^ 10-8^ A for current and 50 mg for tension. Each^ impulse in B^ is^ marked^ by a^ dot.^ Other^ spikes in B and smaller spikes in^ the^ upper trace^ in^ A^ + B^ are^ stimulus^ artifacts.
the response to current stimulation. These two values for each set of graphs in Fig.
7 were not significantly different (P > 0-1 for the left set, P > 0 05 for the right set
of graphs). Furthermore, the test for variance ratio (F test) examined in the similar
way indicates that^ the impulse frequency distribution during combined stimulation
is not significantly different from that of the response to each stimulation added
together during the same period (P > 0-25 for the left, P > 01 for the right set of
graphs). Taken together, these findings indicate that impulses elicited by muscle
0011- -
1 L 1 L L 1.
I
I
-[-fl-
RESPONSES OF GOLGI TENDON ORGAN
assumptions, estimated the absolute threshold force for initiation of a single impulse
from a GTO in cat soleus muscle to be as little as 4 mg. The corresponding values
obtained from isolated preparations in^ vitro^ are^ approximately 8-170 mg for^ a
sustained discharge (Fukami & Wilkinson, 1977) and 4-5-22 mg for initiation of a
single impulse. Although the^ properties of^ GTOs in^ situ^ may^ differ^ from those
described in the present work, the two sets of values obtained from in situ and in
vitro preparations are remarkably similar, both pointing toward high sensitivity of
this sense organ.
In their recent studies on GTOs in hind limb muscles of cats, Gregory & Proske
(1979) reported that the relation between firing rate and tension for combined
stimulation of^ up to^ twelve^ motor^ units^ acting^ on^ single GTOs^ may^ differ^ markedly
depending on the combination of^ motor^ units^ stimulated.^ Based^ on^ these^ and other
findings they proposed that each motor unit acts on a particular collagen bundle in
a GTO, activating the^ impulse initiation site in^ the myelinated terminal branch
innervating that collagen bundle. In other words each collagen bundle in a tendon
organ behaves relatively independently of^ the others and the^ response seen^ in^ the
parent axon reflects the interaction of impulses originating from multiple sites in the
ending. In contrast, the^ present results^ (Fig. 1) suggest that^ contraction^ of each
GTO-muscle fibre constitutes an equally potent input to the receptor, depending only upon the^ amount^ of^ active^ force^ developed; a^ tendon^ organ responds simply to^ the total force exerted upon it. This discrepancy seems to^ be due^ largely, if^ not^ entirely, to the differences in (^) experimental conditions. For (^) instance, a part of the results by Gregory & Proske (1979) could be explained, at least qualitatively, by assuming an unloading effect^ exerted^ by non^ GTO-muscle^ fibre^ contraction^ of^ single motor^ units (Houk & Henneman, 1967; Stuart, Mosher, Gerlach &^ Reinking, 1972). If^ only one or two muscle fibres of a single GTO-activating motor unit inserted directly into^ the organ (Houk et^ al. 1971), the^ larger the^ motor^ unit^ the^ more^ powerful the^ expected unloading effect. Smaller motor units might exert^ little^ such^ effect.^ Thus, when the largest motor^ unit is^ stimulated^ first, which^ would^ presumably produce a^ relatively
weak tendon organ response, the response to successive^ recruitment^ of smaller^ motor
units in order of (^) increasing potency would result in a (^) progressively steeper tension vs. response relation. On the other hand, the^ reverse^ order of^ recruitment^ would produce a^ relation^ showing a^ steep rising^ foot^ followed^ by^ a^ progressively^ less^ steep curve (Fig. 3 of Gregory & Proske, 1979). Similarly, when^ the^ firing rate^ evoked^ by each of a number of motor units of similar size differs by a relatively small^ amount, an approximately linear relation^ is^ expected between^ firing rate^ and^ tension,
irrespective of the order of motor unit recruitment.
The dynamic response of^ GTOs^ where^ myelinated terminal branches^ also^ occur (Barker, 1974) is generally much less than that of^ spindle primary endings, being comparable to^ that of the^ secondary ending. Based^ on^ the^ analysis^ of^ receptor potentials and impulse discharge from^ single GTOs^ in vitro it^ has been^ suggested that it is not (^) necessary to postulate functionally distinct multiple impulse initiation sites to account for the dynamic response (Fukami &^ Wilkinson, 1977). The present analysis of responses to tetanic contraction of GTO-muscle fibres, to electrical current pulses and^ to^ combined^ stimulation^ suggests that^ the^ impulse discharge seen in the parent axon originates from a^ single impulse initiation^ site^ in
442 Y. FUKAMI
the terminal. This conclusion,^ however,^ does^ not exclude the possibility of multiple
impulse initiation sites in myelinated branches as shown for frog muscle spindles (Ito
& Vernon, 1975); each branch, if tested individually, might initiate impulses. The
impulse initiation site examined in the present study might simply be the lowest
threshold site among others and under the present experimental conditions only this
site was activated. Or, as in the case of cutaneous type I mechanoreceptors (Horch,
Whitehorn & Burgess, 1974), the highest rate of discharge initiated from the lowest
threshold site may^ have^ suppressed^ discharge^ in^ other sites by a resetting mechanism.
It is also possible that all myelinated terminal branches in a tendon organ are not
excitable, subserving only electrotonic transmission of information from receptor
sites to the impulse initiation site where information carried by all branches would
be integrated. The short^ internodal^ length^ (05^ mm or^ less) usually^ seen^ along
myelinated terminal branches in a tendon organ is obviously unfavourable to
electrotonic transmission. This disadvantage might be overcome by frequent terminal
arborization which would increase the receptive membrane area where mechanoelectric
transduction occurs, leading to a proportionate increase in the sensitivity of this sense
organ. In^ addition,^ as^ indicated by^ computer simulation^ studies^ by^ Revenko,^ Timin
& Khodarov (1973), the impedance mismatch resulting from the increased area of
non-myelinated terminal membrane^ may be^ reduced^ because of decreased^ internodal
length proximal to the terminal and decreased diameterofthe terminal non-myelinated
segment, leading^ to^ greater^ efficacy of^ electrotonic^ transmission^ along terminal
branches.
This work was supported by grants from the National Science Foundation (BNS77-21801) and from the Muscular Dystrophy Association (Jerry Lewis Muscular Research Center, Washington University). The^ author^ is^ grateful to^ Drs C. C.^ Hunt,^ D. Purves and R. S. Wilkinson for^ reading the manuscript.
REFERENCES BARKER, D. (1967). The innervation of^ skeletal^ muscle.^ In^ Myotatic,^ Kinesthetic^ and^ Vestibular Mechanisms, ed.^ DE^ REUCK, A. V.^ S. Boston: Little Brown. BARKER, D. (1974). The^ morphology of muscle^ receptors.^ In^ Handbook^ of Sensory^ Physiology,^ ed. HUNT, C. C., vol.^ III/2. Berlin:^ Springer-Verlag. BINDER, M. D., KROIN, J. S., MOORE, G. P. & STUART, D. G. (1977). The response of Golgi tendon organs to^ single motor^ unit^ contractions. J.^ Physiol.^ 271,^ 337-349. BROKENSHA, G. & WESTBURY, D. R. (1974). Adaptation of the discharge of frog muscle spindles following a stretch. J. Physiol. 242, 383-403. BULLER, A. J. (1965). A model illustrating some aspects of muscle spindle physiology. J. Physiol. 179, 402-416. BULLER, A. J., NICHOLLS, J. G. & STROM, G. (1953). Spontaneous fluctuations of^ excitability in^ the muscle spindle of the frog. J. Physiol. 122, 409-418. CLIFFORD, P. & SUDBURY, A. (1972). The Markow property of impulse interaction^ in^ branching nerve fibres. Math. Biosci. 13, 195-203. EAGLES, J. P. & (^) PURPLE, R. L. (1974). Afferent fibres with^ multiple encoding sites. Brain^ Res.^ 77, 187-193. EDWARDS, C. (1955). Changes in the discharge from^ a^ muscle^ spindle produced by electrotonus^ in the sensory nerve. J. Physiol. 127, 636-640. Fox, J. M. (1976). Ultra-slow inactivation of^ the^ ionic^ currents^ through the membrane of myelinated nerve. Biochem. biophys. Acta^ 426, 232-244. FUKAMI, Y. (1970). Accommodation^ in^ afferent^ nerve^ terminals^ of snake muscle^ spindle. J. Neurophysiol. 33, 475-489.
