n6 - pnas · 11 cajal, s. r., "l'anatomie fine de la moelle 6pinilre," atlas der...

592 PHYSIOLOGY: LORENTE DE N6 AND CONDOURIS PROC. N. A. S.

zero and ceiling is an expression in the inverse sense of the rate at which the meta-bolic processes are operating.

It is of interest that the dendritic membrane, given adequate oxygen, normalcarbon dioxide and maintained at normal body temperature, should be set at alevel of polarization so far from that at which maximal response is possible. Onemust suppose that dendritic conduction with decrement toward the tips is thenormal occurrence and important to the normal working of the motoneurons.It would occur, for instance, not only in artificial "antidromic" action, but alsoin association with monosynaptic reflex transmission, for the presynaptic endingsconcerned are located on the cell body and the thick proximal dendrites,'1' 12 andthe impulse generated there would travel toward the dendrite tips as well as alongthe axon to the periphery.

1 Renshaw, B., "Effects of presynaptic volleys on spread of impulses over the soma of themotoneuron," J. Neurophysiol., 5, 235-243 (1942).

2 Lorente de N6, R., "Action potential of the motoneurons of the hypoglossus nucleus," J.Cell. Comp. Physiol., 29, 207-288 (1947).

3 Lloyd, D. P. C., "The interaction of antidromic and orthodromic volleys in a segmentalspinal motor nucleus," J. Neurophysiol., 6, 143-152 (1943).

4 Lloyd, D. P. C., "Electrical signs of impulse conduction in spinal motoneurons," J. Gen.Physiol., 35, 255-288 (1951).

6 Lloyd, D. P. C., "Note on the lateral dendrites of plantar motoneurons," these PROCEEDINGS,45, 586-588 (1959).

6 Chang, H. T., "Cortical neurons with particular reference to the apical dendrites," ColdSpring Harb. Symp. Quant. Biol., 17, 189-202 (1952).

7 Lloyd, D. P. C., "Influence of asphyxia upon the responses of spinal motoneurons," J. Gen.Physiol., 36, 673-702 (1953).

8 Lloyd, D. P. C., "Mediation of descending long spinal reflex activity," J. Neurophysiol., 5,435-458 (1942).

9 Lloyd, D. P. C., and A. K. McIntyre, "Analysis of forelimb-hindlimb reflex activity inacutely decapitate cats," J. Neurophysiol., 11, 455-470 (1948).

10 Lorente de No, R., A study of nerve physiology, Studies from The Rockefeller Institute, 131,132, 1947.

11 Cajal, S. R., "L'anatomie fine de la moelle 6pinilre," Atlas der pathologischen Histologie desNervensystems (Berlin: Hirschwald, 1895).

12 Sprague, J. M., "The distribution of dorsal root fibers on motor cells in the lumbosacralspinal cord of the cat, and the site of excitatory and inhibitory terminals in monosynaptic path-ways," Proc. Roy. Soc., B149, 534-566 (1958).

DECREMENTAL CONDUCTION IN PERIPHERAL NERVE.INTEGRATION OF STIMULI IN THE NEURON

BY R. LORENTE DE N6 AND G. A. CONDOURIS*THE ROCKEFELLER INSTITUTE, NEW YORK CITY

Communicated February 23, 1959

The purpose of this communication is to reinstate an important piece of knowl-edge, the doctrine of decremental conduction in peripheral nerve, which existedin the classical literature, but which was lost some 35 years ago, when, mistakenly.physiologists of prestige deemed it erroneous.

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

VOL. 45, 1959 PHYSIOLOGY: LORENTE DE N6 AND CONDOURIS 593

Historical Background.-The concept of decremental conduction was enunciatedin 1881 by Szpilman and Luchsingerl to explain the results of this experiment. Asegment of the nerve of a nerve-muscle preparation is enclosed in a narcotizingchamber. The nerve is stimulated alternatively through electrodes outside thechamber, near the central end of the nerve, and through electrodes inside thechamber. Contraction of the muscle serves as indicator of the effectiveness ofstimulation.t As the effect of the anesthetic (ether or alcohol vapor) develops,inside stimulation soon fails to cause a muscle contraction while the effect of out-side stimulation remains apparently unchanged. Nevertheless, if the stimulatingcurrent is increased, inside stimulation again becomes effective. As the anesthesiaadvances a stage is reached during which outside stimulation, however strong, failsto cause a muscle contraction, while strong inside stimulation still is effective.Finally, also strong inside stimulation becomes ineffective. During the recoveryfrom anesthesia inside stimulation becomes effective first and some time lateralso outside stimulation becomes able to elicit a muscle contraction.That at the beginning of the anesthesia inside stimulation fails, unless the stimu-

lating current is increased, indicates that the anesthetic raises the threshold ofstimulation of the nerve without preventing conduction of impulses. Szpilmanand Luchsinger did not follow this change in detail, but a few years later Werigo2showed, in a systematic investigation, the accuracy of which cannot possibly beincreased by the use of present-day techniques, that as the anesthesia progresses thestimulation threshold rises continuously. The increase in threshold, however,cannot explain the fact that after outside stimulation has failed, strong inside stimu-lation still results in the initiation of impulses, which cause the muscle to contract.To explain this fact, Szpilman and Luchsinger developed the concept of decrementalconduction with an argument so lucid that it will be reproduced here with minorchanges in terminology. Let a, b, c, d, ... be successive segments of a nerve.At a the stimulating current initiates an impulse of magnitude 'a, given by thisequation

'a = 8AS)S (1)

'a measures the magnitude of the action current (one may take the height of theaction potential as a measurement of that magnitude) which the impulse at acauses to flow; ,u(S) is a function of the stimulating current. The impulse at ainitiates at b an impulse of magnitude Ib = alIa: The impulse at b initiates at can impulse of magnitude I, = a2Ib, etc., so that during propagation the nerve im-pulse has the following successive magnitudes

Ib = ajla= AualSIC = a2Ib = lUala2S (2)Id = a3I, = Malaa2a3S

In equations (2) the a's are functions of the magnitudes I and of the properties ofthe nerve. It is clear that if

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

594 PHYSIOLOGY: LORENTE DE NO AND CONDOURIS PROC. N. A. S.

a, = a2 = a3 = ... = 1 (3)

the impulse will be conducted with constant magnitude.If, however,

al, a2, a3, ... > 1 (4)

the impulse will increase in magnitude during propagation. (Incremental con-duction occurs, for example, whenever the impulse passes from a zone of decrementinto untreated nerve; see below, Fig. 7.)

If, on the other hand,

al, a2, a3,. . < (5)

the impulse will decrease in magnitude during propagation, i.e., it will be conductedwith decrement; eventually it will be extinguished. The length of the segment ofnerve through which an impulse may propagate itself depends on two factors, themagnitude that it has at the cathode of the stimulating current, and the magnitudeof the a's. The smaller are the a's, the more rapid is the decrement and con-sequently the shorter the segment through which the impulse can propagateitself.Equations (1), (2), and (5) apply to nerve, which has been submitted to the action

of anesthetics. Outside stimulation fails, while inside stimulation still is effective,when the values of the a's are such that the nerve impulse can propagate itselfthrough half the length but not through the entire length of the narcotizing cham-ber; inside stimulation fails when the values of the a's are such that the impulsecan no longer propagate itself through half the length of the chamber. Duringthe recovery the values of the a's increase progressively so that the impulse againbecomes able to propagate itself first through half the length and then throughoutthe entire length of the chamber.The remarkable fact should be emphasized that according to the doctrine of

decremental conduction anesthesia does not block conduction because it abolishesthe ability of the nerve fibers to produce impulses; anesthesia blocks conductionbecause it creates the conditions for decremental conduction and the nerve impulseis extinguished during propagation through the treated segment. On the otherhand, it should be emphasized that in Szpilman and Luchsinger's formulation nodefinite statement was included regarding the value of gA(equation 1) in normal nerve.It was not until some 30 years later that the crucial nature of this problem could beunderstood.The doctrine of decremental conduction survived a 30 year long period of discus-

sion and elaboration by a large number of physiologists. During this periodweighted objections were raised against it by Werigo,2 but the objections werepositively removed by the work of Dendrinos3 in Hering's laboratory and by thework of a number of authors (Frbhlich, Boruttau, Ishikawa, etc.) in Verworn'slaboratory (cf. literature in Verworn4). The vast majority of those experimentswere done with the classical techniques of comparing the effects of stimulation out-side and inside the narcotizing chamber or determining the length of the segmentsthrough which at various depths of anesthesia the impulses can propagate them-selves, but in 1912 Adrian5 described a new type of experiment (see below, Fig.

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


7,1 , II), so elegant and so convincing, that by itself it was sufficient to establish thedoctrine of decremental conduction upon sound foundation.

After decremental conduction had been securely established, i.e., after it had be-come clear that the magnitude of the nerve impulse could be measured by the lengthof the zone of decrement through which the impulse can propagate itself, it becamepossible to attempt to determine for normal nerve the relationship of the magnitudeof the nerve impulse to that of the stimulating current. Since, in experiments donewith the narcotizing chamber technique (Fig. 2, below), after outside stimulationhas failed, to increase the outside stimulus apparently has no significant effect,Lodholz6 and Adrian7 concluded that in untreated nerve the impulse has a fixedmagnitude, i.e., that the all-or-nothing law is valid for the individual nerve fibers.Also toward the end of the period of elaboration proof was obtained by Adrian8 andLucas9 of an important fact, namely that when the conditions necessary for decre-mental conduction have been created, the decrement is intensified during the rela-tively refractory period. Among other reasons, the fact is important because itexplains Wedensky's inhibition phenomenon.10

In its final form the doctrine of decremental conduction included two proposi-tions (cf. Verworn,4 Lucas11).

(a) All-or-nothing law. In untreated nerve the magnitude of the nerve im-pulse is independent from the magnitude of the stimulus: the nerve fibers eitherdo not produce an impulse or produce one of maximal magnitude. Consequently,in untreated nerve the impulse is conducted with constant magnitude, at uniformspeed.In terms of Szpilman and Luchsinger's formulation the all-or-nothing law would

have had this meaning. In untreated nerve, 1A (equation 1) is a constant and themagnitude of the impulse at the cathode is

Ia = ASt (6)

where St is the threshold stimulus, i.e., the magnitude of that applied current whichis just sufficient to initiate an impulse. Therefore to increase the stimulating cur-rent may shorten the latency, but it cannot increase the magnitude of the nerveimpulse.

(b) Decremental conduction. In nerve depressed by any one of a large varietyof agents the all-or-nothing law is not valid. The magnitude of the impulse at thecathode decreases when the stimulus which produces a maximal impulse is reduced,and in any one segment of the nerve the magnitude of the nerve impulse is suchthat it can initiate in the next segment only an impulse of smaller magnitude.Consequently, the magnitude of the impulse and the speed of conduction decreaseprogressively during propagation; eventually the impulse is extinguished.

In terms of Szpilman and Luchsinger's formulation proposition b is defined byequations (1), (2), and (5).The importance of the doctrine of decremental conduction for the understanding of

the physiology of the central nervous system was fully appreciated by the classicalauthors. The all-or-nothing law could be valid for peripheral nerve, which ap-parently has no other function than to conduct impulses throughout its entirelength; but there should be in the central nervous system zones where decremental

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


conduction occurs. In those zones processes of summation and of inhibition wouldtake place, much in the manner in which they take place in a segment of peripheralnerve, which has been caused to conduct with decrement (cf. Wedensky,'0 Adrianand Lucas'2). To implement this thought ingenious arguments were developedby Lucas" (cf. also Forbes'3).The doctrine of decremental conduction did not survive the objections that were

raised against it by Kato,"4 15 Davis, Forbes, Brunswick, and Hopkins,'6 and Katoand Teruuchi.'7 It seems unbelievable, but it is true, that although the objectionswere soon proved to be unjustified by a number of authors (Ishikawa and co-workers,'8 Frbhlich,'9 Baruch,20 Wiersma,2' Reswjakoff,2 Woronzow,23 etc.),the doctrine of decremental conduction disappeared from the literature and thereremained, to mold the thinking of neurophysiologists, only the all-or-nothing lawin the generalized, inflexible form given to it by Kato: The magnitude of the nerveimpulse depends upon the state of the nerve fibers, but in any case the nerve fibersproduce either a maximal impulse or none at all. If the impulse is small it will beconducted at a reduced speed, but it will be conducted, without decrement, through-out the entire length of the nerve.

It is indeed difficult to explain why the doctrine of decremental conduction wasabandoned, in spite of the existing, gigantic body of evidence that supported it.Two factors may have played a role. Before it had been sufficiently elaborated, anew technique, which became available at that time (the study of responses ofsingle fibers),+ seemed to indicate that the all-or-nothing law is valid under anycondition (cf. Tasaki24). On the other hand the celebrated researches of Adrianand co-workers25 had established the existence of a theretofore unsuspected mecha-nism of graded excitation, the regulation of the number and of the frequency ofimpulses conducted by nerve fibers.Soon after the all-or-nothing law had been accepted as generally valid for pe-

ripheral nerve it became clear that if the law also should apply to the bodies anddendrites of the neurons, the integrative role of the neuron could not be understood.Let this situation be made clear. Toward the end of the past century analysisof the fine structure of the nervous system led Cajal to the establishment of the nowgenerally accepted neuron theory.26 An essential concept of the theory is this:The neuron integrates the stimuli that it receives at the numerous synapses thatnerve fibers of varied origins establish on its body and dendrites and initiates in itsaxon a wave of excitation which is delivered to other neurons with which the axonestablishes synaptic junctions. To develop this concept it was not necessary tounderstand in detail the nature of nervous conduction; consideration of the planof structure of the nervous system and the assumption of a flow of "nervous energy"in the neurons from the synaptic junctions toward the axon were sufficient.

Later, however, after nerve physiologists had proved the discontinuous nature ofthe nerve impulse, to explain how the neurons can play their integrative role becamea serious problem. If the body and dendrites of the neurons should conduct all-or-nothing, i.e., maximal, nerve impulses and if the impulses could be initiated byactivation of a small group of closely neighboring synapses anywhere in the neuron,then summation of synaptic stimuli delivered to distant synapses, for example tosynapses in different dendrites or to synapses in two different segments of a longdendrite, could never take place.

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


The problem of summation of impulses delivered at distant synapses is par-ticularly apparent in the cerebral cortex, where as a rule the number of synapsesestablished on the bodies of the neurons is very small in relation to the number ofsynapses established on long dendritic processes,26-28 and where there are types ofneurons which could never be activated by afferent fibers unless impulses deliveredat thin branches of the main dendritic shaft could summate to initiate a self-propagating nerve impulse. Thus, when the attempt was made to ascertain, onthe basis of detailed anatomic knowledge, how nerve impulses could travel throughthe cerebral cortex, the conclusion had to be reached that the all-or-nothing lawcould not be valid for the bodies and dendrites of the neurons; instead, the existenceof graded responses capable of spreading without leaving a refractory period, andcapable of summating, had to be postulated.27 Consideration of the same problemhas recently led Bishop29 to a similar conclusion. In the recent literature gradedresponses of the type of subliminal postsynaptic potentials (Eccles30' 31) have beenbelieved to represent important components of cortical potentials (Bishop,32Bremer,33 Grundfest34). In addition to subliminal synaptic potentials, and toother subliminal processes discussed by Bremer,35 to explain the integrative role ofthe neuron we may consider now the kind of graded responses postulated in theclassical doctrine of conduction with decrement: nerve impulses of variable magni-tude capable of propagating themselves through variable, but limited, lengths.

Reinstatement of the Classical Doctrine.-The existence of decremental conductionhas been demonstrated in an extensive series of experiments done with spinalroots of bullfrogs and with sciatic nerves of frogs and bullfrogs, either with intactsheath or desheathed.§ Two classes of agents have been used to cause the nervesto conduct with decrement. (1) Agents which produce a progressive depolariza-tion of the nerve fibers, such as ethyl alcohol, an excess of potassium ions or anexcess of hydrogen ions, and (2) agents which do not alter the resting membranepotential, such as sodium-free or sodium-deficient solutions, procaine, cocaine,xylocaine, ethyl-urethane, and phenyl-ethyl-urethane. With oscillographic tech-nique the classical experiments with stimulation outside and inside the narcotizingchamber can be repeated in such a manner that the causes of error suspected byWerigo2 and by Kato14 15 can be positively eliminated. It has been found that theobservations on decremental conduction made by the classical authors, and inparticular the results of the refined experiments done by Adrian5' 7 and Lodholz,6were perfectly correct (Figs. 1 and 2). On the other hand, with oscillographictechnique it is possible to observe directly, during propagation of the nerve impulses,the progressive decrement of both the height of the action potential and the speed ofconduction (Figs. 3 to 6).Whether the depressant agent reduces the value of the membrane potential or

not, two changes in the properties of the nerve fibers always take place beforedecremental conduction becomes demonstrable. The threshold of stimulationincreases and the height of the maximal action potential at the cathode decreases.In all probability decremental conduction develops very early because conductionwith decremental speed is observed very soon, but since segments of bullfrog nervewith uniform diameter longer than 30 mm are not available the decrement in theheight of the action potential cannot be demonstrated with certainty until thedecrement is rapid enough for nerve impulses to be extinguished during propagation

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

598 PHYSIOLOGY: LORENTE DE N5 AND CONDOURIS PRoc. N. A. S.

through a 30 mm long segment of nerve. With advancing time those two changesbecome progressively greater and the rate of decrement during conduction increases,so that the nerve impulses are extinguished during propagation through progres-sively shorter segments of nerve. In each nerve fiber propagation ceases when theaction current of the decrementing impulse fails to reach the threshold of stimula-tion of a certain segment of the fiber; in this segment, therefore, the action currentcreates only a subliminal electrotonic potential. Decremental conduction ceasesentirely when the maximal impulse at the cathode is so small that its action currentcannot initiate an impulse in the next segment of the nerve. The action potentialof the unconducted impulse spreads only electrotonically, i.e., passively, much inthe same manner as does the electrotonic potential created by the applied current(Fig. 4). Finally, the impulses initiated at the cathode become so small that theiraction potentials are undistinguishable from the electrotonic potential created bythe applied current.

Decremental conduction is independent from the direction of travel of the nerveimpulses. Impulses initiated near the central margin of a treated segment of nerveare extinguished during propagation peripheralward before reaching the peripheralmargin, and impulses initiated near the peripheral margin are extinguished duringpropagation centralward before reaching the central margin.The very fact that decremental conduction occurs leaves no doubt that the magni-

tude of the nerve impulse increases with that of the stimulus and that large impulsesare propagated through longer distances than small ones. With multifibered nerves,however, it is not possible to establish an accurate relationship between the heightof the action potential of the individual nerve fibers and the magnitude of the stimu-lating current, even though it is reasonable to assume that the height of the maximalaction potential at the cathode is proportional to the height of the individual fiberspikes. For this reason it is important to mention that in recent work with singlefibers spectacular failures of the fibers to obey the all-or-nothing law have beenobserved (Altamirano, Coates and Grundfest,36 Mueller36a). When the magnitudeof the current which is capable of initiating a maximal action potential is reducedthe height of the action potential decreases.The reinstatement of the doctrine of decremental conduction is a matter of

importance for nerve physiology. It has appeared that the magnitude of the nerveimpulse may be made to vary from the normal maximal magnitude down to zeroby two different changes in the ionic composition of the external medium of thenerve fibers: a decrease in the amount of sodium ions or an increase in the amount ofpotassium ions. And the same result may be obtained, while the ionic composi-tion of the external medium is maintained normal, by means of agents as varied asanoxia, alcohol and ether, which depolarize the nerve fibers, or anesthetics, whichdo not alter the value of the membrane potential. Consequently it is impossibleto accept any hypothesis, such as the Hodgkin-Huxley-Katz hypothesis (cf.Hodgkin37) in which the magnitude of the nerve impulse is supposed to be directlydetermined by the concentration of sodium and potassium ions outside the nervefibers.

Experimental Evidence.-A few crucial experiments on decremental conductionwill be now briefly described.The experiments illustrated by Figures 1 and 2 are a repetition, with the advan-

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


tages offered by modern oscillographic technique, of the classical experiments withstimulation inside and outside the narcotizing chamber. The purlose of the ex-periments is to answer two objections raised by Werigo2 and by KWto4 15 againstthe classical explanation. The first objection was this. According to Werigo andto Kato, after outside stimulation has become ineffective, inside stimulation remainseffective, not because impulses are initiated in the treated segment, but because theapplied current spreads and initiates impulses in the untreated segment of the nervebeyond the peripheral wall of the narcotizing chamber.The experimental arrangement is indicated in the diagram at the bottom of Figure

2. The nerve is mounted inside a rectangular vessel divided into three chambers bytwo walls, 1.5 mm thick, provided with vaseline seals. The lateral chambers arefilled with Ringer's solution and the central chamber with the test solution. Thesolutions are removed during the brief periods of time needed for oscillographicanalysis. Between the periods of oscillographic analysis the solutions are renewed,and therefore effectively stirred, very frequently; indeed, they have been routinelyrenewed at 30 second intervals. Thereby the concentration of the test substance ismaintained permanently constant throughout the entire length of the narcotizingchamber, and permanently zero in the lateral chambers. A pair of (outside) stimu-lating electrodes is placed in one of the lateral chambers (Fig. 2, o), with the cathodeat some distance from the wall, and a pair of recording electrodes is placed in theother lateral chamber, the first recording electrode being in contact with the walland the second electrode at some 15-20 mm from it, in contact with an uninjuredpoint of the nerve. In the central chamber there is a pair of movable electrodes(m). After stimulation through the outside electrodes, o, has become ineffectiverectangular pulses of current are delivered to the treated segment through electrodesm, first with the cathode in contact with the wall of the chamber and then with thecathode at progressively increasing distances from the wall. As the cathode ismoved away from the wall the magnitude of the stimulating current is increased.Whether the stimulating electrodes o are placed on the central segment of the nerveor in the peripheral segment does not modify the experimental results.The argument underlying the experiment is this. As is at present well known,

when a rectangular pulse of current is applied to a segment of nerve of uniformproperties impulses are initiated when the catelectrotonic potential reaches thresh-old value. Since the catelectrotonus has its highest value at the cathode andrapidly decrements in height with distance from it, the impulses are always initiatedat the cathode itself. If the current is just above threshold the impulses are initiatedafter a long utilization time, up to 1 msec or even more, but if the magnitude of thecurrent is increased the utilization time decreases rapidly, because the catelectro-tonic potential reaches threshold value at a much earlier time. With currents twoor three times larger than that which initiates impulses in all the A fibers the utiliza-tion time is negligible.When the current is applied near the margin of a segment of nerve that has been

treated with a blocking agent the situation may be different, because, since theagent causes a rise in the stimulation threshold of the treated segment, the catelec-trotonic potential may reach threshold value in the untreated segment at somedistance from the cathode before reaching threshold value at the cathode itself.The problem therefore is to differentiate between impulses that may have been

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


initiated in the untreated segment byelectrotonic spread of the applied cur-rent and impulses that have been ini-tiated at the cathode in the treatedsegment. The problem has proved tohave a simple solution.

In the experiment illustrated byFigure 1, a 30-mm long segment ofnerve was treated with 1.5 mM phenyl -

U. urethane dissolved in Ringer's solutioni.e., with an agent, which like anyother nondepolarizing anesthetic, in-creases the spread of the electrotonusbecause it increases the transverse resistance of the nerve membrane. The

W -AI observations were begun 15 minutes_s_ after stimulation through the outside

electrodes had become ineffective.

Two strengths of current were usedwhen the cathode of the movable elec-trodes (m) was in contact with the wall

mmd_||-of the chamber, i.e., approximately

1.5 mm from the first recording elec-trode. The weaker current createdonly a subliminal electrotonic potential

--_| (Fig. 1, 1), but the stronger current wasable to initiate a small action potentialafter a long utilization time (Fig. 1, 2).Since the assumption can only re-inforce the argument, let it be assumedthat the action potential in Figure 1, 2belonged to impulses initiated just out-side the central chamber. Then, theelectrotonic potential in Figure 1, 2

FIG. 1.-Decremental conduction in a bullfrog measures the minimal height that thesciatic nerve, with intact sheath, treated with catelectrotonus must reach at the first1.5 mM phenylurethane added to Ringer's solu- recording electrode when impulses aretion. Arrangement of electrodes as indicated inthe diagram at the bottom of Figure 2. The initiated just outside the central cham-magnitudes of the stimulating currents used for ber. (If the assumption were erro-records 2 to 12 are given as multiples of the cur- .

..

trent used for record 2. The distances in mm. neous, i.e., if the impulses to which thegiven on the records are the distances from the action potential in Fig. 1, 2 belongedcathode of the movable electrodes (m) to thewall of the central chamber. Stimuli of any had been initiated inside the treatedmagnitude applied through the outside electrodes segment, the argument that follows(o) had become ineffective 15 minutes before theobservations of decremental conduction through would be reinforced, because in thatthe treated segment were begun. After outside case the electrotonic potential instimulation had become ineffective the nerve wasleft suspended in moist air. Figure 1, 2 would be the measure of a

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

VOL. 45, 1959 PHYSIOLOGY: LORENTE DE NO AND CONDOURIS 601

stimulus which was still subliminal for points outside the narcotizing chamber.)On the basis of this information it can concluded, with absolute certainty, that

all the other action potentials that appear in Figure 1 belonged to impulses whichwere initiated within the treated segment, since they are superposed upon electro-tonic potentials which are smaller than the electrotonic potential in record 2(record 4) or even smaller than the subliminal electrotonic potential in record 1(records 6 to 12). Particularly convincing are records 6 and 8 which were obtainedwith the cathode at 6 and 8 mm from the wall. They present large action potentialshaving shorter latencies than the action potentials in records 2 or 4. Therefore,the impulses, which had to be conducted through 6 or 8 mm of treated nerve, musthave been initiated at the cathode after very short utilization times.An important feature of decremental conduction is illustrated by Figure 1.

When the cathode of the movable electrodes (m) was displaced from 4 to 6 mm fromthe wall, that current which a 4 mm had initiated an action potential capable ofreaching the first recording electrode (Fig. 1, 4) failed to do so at 6 mm from thewall (Fig. 1, 5). The explanation of this phenomenon lies at hand. The relativelysmall current that was used to obtain records 4 and 5 could initiate only smallimpulses, which could propagate themselves through a 4 but not through a 6 mmlong segment of treated nerve. When the applied current was increased it becameable to initiate larger impulses which readily propagated themselves to reach theuntreated segment of the nerve (Fig. 1, 6). A similar phenomenon was observedwhen the cathode was displaced from 6 to 8 mm from the wall (Fig. 1, 7) and thecurrent was strengthened (Fig. 1, 8).

All the records of the series 8 to 12 (Fig. 1) were obtained with the same stimulat-ing current. Consequently, the progressive decrease in height of the action po-tential gives a fair measure of the intensity of the decrement during conduction.As already mentioned, the observations presented in Figure 1 were begun 15

minutes after stimulation through the outside electrodes (o) had failed. Duringthis time the anesthesia had become somewhat deeper and no fiber could conductimpulses through a segment longer than 16-17 mm. If the observations had beenmade immediately after outside stimulation had failed, conduction with decrementthrough a 26-28 mm segment of nerve could have been demonstrated, as has indeedbeen repeatedly demonstrated in this laboratory. On the other hand, if the con-centration of the anesthetic is increased the intensity of the decrement grows veryrapidly, so that within a few minutes decremental conduction through a segment ofnerve more than 6-8 mm long can no longer take place. If phenylurethane is usedat the concentration 5 mM the anesthesia becomes so deep that no decrementalconduction at all can take place. To be sure, with the technique used for the ex-periments of Figures 3 to 6 it is found that small action potentials can still beelicited at the cathode, but those action potentials cannot propagate themselves,they only spread electrotonically.Even more convincing are the results obtained with the use of depolarizing agents.

In the experiment illustrated by Figure 2, a 30mm long segment of nerve was treatedwith 0.106 N potassium chloride (added to Ringer's solution). The observationswere begun 10 minutes after stimulation through the outside electrodes (o) had be-come ineffective. At each position of the movable electrodes two magnitudes ofcurrent were used; the weaker one initiated only a small action potential, which was

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

602 PHYSIOLOGY: LORENT'E DE NO AND CONDOURIS PROC. N. A. S.

S-

I~~~~~~~~~~ ~ 0_~~~~~~~~~~~~

FIG.2.-ecreentl coducion n abullfrog sciatic nerve, with intact sheath, treated withalareecesofpotssim ins(0.106 N KCl added to Ringer's solution). The voltages used toiniiatthrsposesinthesecndand fourth columns of records are given on the records. Thedistnceinmm. ive ontherecords of the first and third columns are the distances from thecatodeofhe ovbleeletroes(in) to the wall of the chamber. Stimuli of any magnitude de-livredthrughtheoutideelectrodes (o) had become ineffective 10 minutes before the observa-tiosoderemnta coducionthrough the treated segment were begun. After outside stimula-

in records 8, 10 and 12 have a fi elevation. (Lorente de N6 and Catlin, unpublished.)

kept approximately constant throughout the experiment, while the stronger currentinitiated a nearly maximal a spike at 0 and 4 mm from the wall and a maximal aresponse at all other points. It hardly needs be mentioned that in order to keepthe small response nearly constant the magnitude of the weaker current had to beprogressively increased so that it could initiate impulses large enough to be prop-

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


agated through a zone of decrement of -progressively increasing length. Themagnitudes of the small stimuli were not recorded. Those of the larger stimuliare given on the records.

Since at the time when the observations were made the repolarizing flow ofdemarcation current must have kept the stimulation threshold at the wall still nearnormal, there is no reason to believe that the impulses to which the action potentialsin records 1 and 2 (Fig. 2) belonged were initiated by spread of the stimulating cur-rent beyond the wall of the central chamber. For the sake of argument, however,let it be assumed that in the case of records 1 and 2 (Fig. 2) the impulses wereinitiated just outside the central chamber. (The assumption serves only to rein-force the argument that follows.) On this assumption record 1 gives the heightthat the electrotonic potential must have at the first recording electrode when anear threshold a response is initiated just outside the central chamber after a longutilization time, and record 2, the height that the electrotonic potential must havewhen a nearly maximal a response is initiated just outside the central chamber,after a brief utilization time.

It will be noted in Figure 2 that, with the cathode at 4 and 8 mm from the wall,the small responses (records 3 and 5) appeared superposed upon electrotonic po-tentials which were much smaller than that in record 1, and that the near maximala response in record 4 and the maximal a response in record 6 appear superposedupon electrotonic potentials which are not only smaller than the potential in record2, they are even smaller than the supposedly threshold potential in record 1.Consequently, there cannot be the slightest doubt that the responses in records3 to 6 were initiated within the treated segment of the nerve. As to the rest of theresponses reproduced in Figure 2 hardly a comment is necessary. The actionpotentials appear superposed upon electrotonic potentials which are either ofnegligible magnitude or entirely undetectable. The impulses, therefore, wereinitiated inside the chamber; in the case of the large currents the impulses must havebeen initiated after negligible utilization times, so that the latency of the responsesis a rather accurate measure of the time of conduction through the treated segment.An interesting feature of decremental conduction appears in Figure 2. The

large action potentials initiated at 0, 4, and 8 mm from the wall include only spikesof a fibers because the currents used did not reach the stimulation threshold of:fibers. The currents used for the rest of the large responses were strong enough toinitiate impulses in ,3 fibers. Nevertheless fB elevations appear only in the responsesinitiated at 12, 17, and 20 mm from the wall. This result indicates that the decre-ment in conduction was more rapid in f3 than in a fibers.Owing to the peculiar manner in which a large excess of potassium ions acts upon

the nerve fibers, after the test solution had been removed from the central chamber,whereby the nerve was left suspended in moist air, the intensity of the decrement inconduction increased progressively with advancing time. Consequently, the lengthof the segment through which impulses in a fibers could propagate themselves toreach the untreated segment of the nerve decreased progressively, from 27.5 mm(Fig. 2) to 20, 15, 10, and 6 mm. Finally, one hour after the observations presentedin Figure 2 had been made, conduction with decrement ceased entirely. Then, itwas found that with the cathode at less than 4 mm from the wall of the chamber largeapplied currents could still initiate impulses beyond the margin of the chamber, but

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


with the cathode at more than 4 mm from the margin no impulse could be initiated,not even by currents 5 times greater than that which with the untreated nerve hadbeen sufficient to initiate impulses in all the A fibers. Such a result had to be ex-pected. In depolarized nerve the spread of the electrotonus is greatly reduced andconsequently with the cathode at more than 4 mm from the wall of the chamber theelectrotonic potential cannot reach threshold height at sufficient rate beyond themargin of the chamber.11

It is now possible to give a satisfactory explanation of the second objection raisedby Werigo2 and Kato14 15 against the classical doctrine. When using a stronganesthetic, Werigo found that the length of the segment of nerve enclosed in thenarcotizing chamber had an influence upon the time needed to establish the conduc-tion block only when the narcotized segment was less than a certain limit length(4-5 mm). Werigo thought that the limit length measured the length of thesegment of nerve through which the action currents of the blocked impulse canspread to initiate impulses in untreated nerve, beyond the narcotizing chamber.Kato found that the limit length was different for different blocking agents andexplained it by assuming irregularities in the diffusion of the anesthetic. 14, 15, 17The results obtained with the techniques used for the experiments of Figures 1

and 2 and for the experiments of Figures 3 to 6 have proved that depending uponthe concentration, and upon the time of action of the blocking agent, conductionwith decrement can be demonstrated, with any blocking agent, through segments ofnerve ranging from 0 to 20 mm with frog sciatic nerve, and ranging from 0 to 30 mmwith bullfrog sciatic nerve or bullfrog spinal roots. (Longer segments of nervewith uniform diameter are not available.)The meaning of the limit length is therefore this. It is a length slightly greater

than that of the segment through which with the particular blocking agent and theparticular concentration used conduction with decrement is still possible. Clearly,if the limit length has been reached, to increase the length of the segment of nervein the narcotizing chamber can have no influence upon the time necessary for theestablishment of the conduction block. If Kato and co-workers had used otherconcentrations they would have observed other limit lengths, shorter or longer.

For example, Kato"4 (p. 118) reported that with the use of potassium cyanide(approximately 0.3 N) no influence could be detected of the length of the treatedsegment upon the time required to abolish conduction, while Wiersma21 using moremoderate concentrations of potassium cyanide readily demonstrated decrementalconduction. Another example, Kato'4 (p. 119) reported that the limit length in thecase of sodium-free sugar solutions is 15 mm, while if the stage of sodium-deficiencyis sufficiently advanced decremental conduction through the treated segment ceasesentirely.

In the experiment illustrated by Figure 3 a nondepolarizing anesthetic (procaine)was used to cause the nerve fibers to conduct decremental impulses. Two amplifierswere used to record on the same film the passage of the impulses past electrodes 2and 3. At the start of the observations both spikes 1-2 and 1-3 were of coursediphasic (Fig. 3, 1), but during the development of the action of the anesthetic twochanges took place (Fig. 3, 2): (a) The height of the first phase of spike 1-3decreased, thus indicating that the magnitude of the maximal action potential atthe cathode was being reduced, and (b) the second phase of spike 1-3 decreased much

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


1 2 3

cond 1.5 mm. trunk

FIG. 3.-Development of decremental conduction in a bullfrog ven-tral spinal root treated with moderate concentrations of procaine (4 mMand 2 mM, added to Ringer's solution). The arrangement of elec-trodes is indicated in the diagram. The whole preparation was im-mersed in the test solution, except during the brief periods of timeneeded for oscillographic analysis. The amplification used for the rec-ords obtained with the oscillograph connected to electrodes 1 and 2was slightly higher than that used for the records obtained with theoscillograph connected to electrodes 1 and 3.

more rapidly than the second phase of spike 1-2. Indeed, after the anesthetic hadacted for 24 minutes spike 1-3 was almost monophasic while spike 1-2 still had apronounced second phase (Fig. 3, 3), which indicates that a number of impulsescould propagate themselves through a 7 mm long segment of nerve, but not througha 25 mm long segment.When the concentration of the anesthetic was reduced the decrement in con-

duction was relieved and both phases of spikes 1-2 and 1-3 became greater (Fig.3, 4 to 6). If the concentration of the anesthetic had been kept low the intensity ofthe decrement in conduction to which the spikes in Figure 3, 6 correspond would haveremained practically constant for an extended period of time, but an increasein the concentration of the anesthetic rapidly resulted in an increase in the intensityof the decrement, so that again very few of those impulses which were able topropagate themselves past electrode 2 were able to reach electrode 3 (Fig. 3, 7). Todecrease the concentration of the anesthetic once more resulted in a relief of thedecrement (Fig. 3, 8, 9).

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

606 PHYSIOLOGY: LORENTE DE Nb AND CONDOURIS PROC. N. A. S.

If the anesthetic is used at a sufficiently high concentration conduction withdecrement ceases entirely and there remains only the electrotonic spread of theaction potential of the impulses initiated at the cathode. Such was the case in theexperiment illustrated by Figure 4, in which in order to analyze the situation ingreater detail 5 recording electrodes were used. At the time when the observations

---U-

5t. I 2 3 4 5+ -~ ~~1truLnk

cord- + 1.54.5 4.5 4.5 4.5mm.

FIG. 4.-1)ecremental conduction in bullfrog ventral spinal root treated with a concentrationof cocaine (10 mM added to Ringer's solution) which was sufficient, after 35 minutes of action,entirely to block propagation of the impulses, which were initiated at the cathode of the stimulat-ing current. Arrangement of electrodes is indicated in the diagram. The numbers 1-2, 1-3, etc.,indicate the pair of electrodes to which the oscillograph was connected. The whole preparationwas immersed in the test solution.The absence of conduction is demonstrated by the monophasicity of the spikes in records 1 to

4. Records 4 to 7 illustrate the electrotonic (passive) spread of the action potential and of thecatelectrotonic potential created by the rectangular pulse of current, which was applied throughthe stimulating electrodes (St.).The restoration of conduction by a long-lasting anodal current, applied through the polarizing

electrodes (Pol.) is illustrated by records 9 to 12 and 13 to 16. The anodal current was closed atthe times indicated by the upward pointing arrows, was kept flowing through the intervals betweensuccessive records (3 seconds) and was interrupted at the times indicated by the downward point-ing arrows. The zero potential level is constant for records 10 to 12 and 14 to 16. The diphas-icity of the spikes in records 11, 12 and 15, 16 proves that during anodal polarization nerve im-pulses became able to propagate themselves not only through a 6 but also through a 15 mm longsegment of the root. Records 13 to 16 were obtained at a slightly lower amplification than theother records.

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


were made the anesthetic (10 rnM cocaine at pH 7.0) had reduced the height of themaximal action potential at the cathode (as measured by the height of the crestof spike 1-5) only to about 50 per cent of normal; propagation of the impulses,however, was prevented by a several-fold increase in the stimulation threshold.That propagation had ceased is clearly shown by the fact that spikes 1-2, 1-S, 1-4and 1-5 were monophasic (Fig. 4, 1 to 4). The electrotonic spread of the un-conducted action potential is illustrated by the records obtained with electrodes1-5, 2-5, 3-5, and 4-5 (Fig. 4, 4 to 7). It will be noted that the electrotonic spreadhas a very rapid longitudinal decrement and that it takes place with very greatrapidity.At the depth of anesthesia illustrated by the records reproduced in Figure 4 it

can readily be shown that the spikes recorded in the neighborhood of the cathodeactually are action potentials of unconducted nerve impulses. If a weak long-lasting anodal current is applied to the nerve the action potential increases in heightand many impulses become able to propagate themselves through considerabledistances. For example, records 11, 12 and 15, 16 (Fig. 4) show that during anodalpolarization spikes 1-2 and 1-4 became diphasic, indicating that a number of im-pulses had been able to propagate themselves not only beyond electrode 2 but alsobeyond electrode 4.The anodal current does not relieve the decrement in conduction because it in-

creases the total value of the membrane potential. A weak short-lasting anodalcurrent does not restore conduction of impulses (Fig. 4, 10, 14) and a strong short-lasting anodal current, which rapidly establishes a large increment in the membranepotential, blocks the initiation of impulses. A weak long-lasting anodal currentrestores conduction because it creates a certain polarization potential, the slowanelectrotonus, (cf. Lorente de N638). The height of the slow electrotonus is ap-proximately equal to the vertical distance between the base lines of records 10,11 and 12 or 14, 15 and 16 of Figure 4.

If the nondepolarizing anesthetic is used at a high concentration, for example 20mM cocaine at pH 7, the maximal action potential at the cathode decreases to about20-25 per cent of normal. Anodal polarization can no longer restore conduction,but the fact that the slow anelectrotonus collapses during the spike-like deflectionleaves no doubt that the stimulating current has been able to initiate small nerveimpulses.

In the experiment illustrated by Figure 5 a spinal ventral root was treated with amoderate excess of potassium ions (18 mM). The development of decrementalconduction can be followed by examining in succession records 1 to 9 of Figure 5.In principle the sequence of changes in the action potentials was the same that wasproduced by nondepolarizing anesthetics (Figs. 3 and 4), but since the excess ofpotassium ions causes only a moderate increase in the threshold of stimulation anumber of impulses could propagate themselves through a 6 mm long segment ofnerve (note the presence of second phases in spikes 1-2 of records 8 and 9 of Fig. 5),when the height of the maximal action potential, as measured by the height of the1-3 spike, was not more than one-third of normal.Records 5 to 9 of Figure 5 were obtained immediately after the removal of the

test solution from the narcotizing chamber, whereby the root was left suspendedin moist air. The other records labeled a, b, and c were obtained after the indicated

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


151,- ~ ~ ~ ~

FIG. 5.-Development of decremental conduction in a bullfrog ventral root treated with amoderate excess of potassium ions (14 mM KCl added to Ringer's solution). Arrangement ofelectrodes as indicated in the diagram at the bottom of Fig. 3. At the start of the observations(record 1) a few fibers responded repetitively to the applied current. The amplification used forspikes 1-2 was slightly higher than the amplification used for spikes 1-3. Note in records 6 toi5c, 6 to 6c, etc., the relief of the decrement during the spontaneous recoveries which took placewhen the root was kept suspended in moist air. (Lorente de N6 and Catlin, unpublished.)

intervals of time. It will be noted that while the root was being kept in moist aira spontaneous recovery took place. The height of the action potential, as measuredby the height of the first phase of spike 1-3 increased and the intensity of the decre-

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


ment in conduction, as measured by the second phases of spikes 1-2 and 1-3, de-creased, which shows once more that the length of the segment of nerve throughwhich the nerve impulse can propagate itself is determined by the height of itsaction potential.

Finally in the experiment illustrated by Figures 6a and 6b a sodium-free mediumwas used to produce decremental conduction in a desheathed bullfrog sciatic nerve.The records of the top row of Figures 6a and 6b were obtained immediately after theremoval of the test solution. At that time the decrement in conduction was sorapid that only a few impulses could propagate themselves beyond electrode 2.But while the nerve was being kept in moist air a spontaneous recovery took place.The action potential, as measured by the height of the first phase of spike 1-5increased progressively and at the same time conduction through an increasinglength of nerve was restored. The same result has been obtained in other experi-ments by immersing the nerve for a few minutes in progressively increasing, moder-ate concentrations of sodium ions (10, 12, 14, and 16 mM). In general it can besaid that when sodium-deficient nerve is restored by progressively increasing con-centrations of sodium ions, to each concentration below normal (110 mM) corre-spond a certain subnormal height of the maximal action potential at the cathode anda certain intensity of the decrement in conduction.

Application to the Central Nervous System.-Before applying the doctrine ofdecremental conduction to problems of the central nervous system an importantremark should be made. Critical consideration of the literature leads to theseconclusions, (1) the all-or-nothing law has never been proved to be strictly valid forperipheral nerve and (2) there are well-established facts which are in conflict withthe "law."That in dissected single fibers graded action potentials are readily observed,36

does not prove, of course, that the all-or-nothing law could not be valid for normal,undissected fibers. On the other hand, it is true that in intact nerve decrementalconduction develops very readily; for example, Rudolph39 studying the response ofundissected single nerve fibers found that, if the nerves have been kept in Ringer'ssolution for some time, impulses are often observed which are extinguished duringconduction between the recording electrodes. Nevertheless, the all-or-nothing lawcould be valid for nerves kept under more physiological conditions.

It is clear that if the problem has to be investigated with intact nerve kept underphysiological conditions no other method is available to investigate the validity ofthe all-or-nothing law than that underlying the experiments done by Lodholz6and Adrian,7 in which the magnitude of the nerve impulse in untreated nerve ismeasured by the length of the zone of decrement through which it may propagateitself. Moreover, the experiment must be done with the use of a blocking agentwhich does not alter the resting membrane potential, because the flow of thedemarcation current alters the properties of untreated nerve (see below, Fig. 7).The experiment does indeed prove that in normal nerve a single nerve impulsepropagates itself with constant (or nearly constant) magnitude, but it does not provethat that impulse has the maximal magnitude, which the nerve fibers are capableof producing.As a matter of fact, it is known that during rhythmic activity at low physiological

frequencies, normal nerve conducts impulses, which, measured by the height of

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

610 PHYSIOLOGY: LORENTE DE Nc AND CONDOURIS PROC. N. A. S.

U,..

FIG. 6a.-Relief of the decrement in conduction produced in a desheathed bullfrog sciaticnerve by immersion for 4 minutes in a sodium-free medium (Ringer's solution in which the so-dium chloride was replaced by diethanol-dimethyl-ammonium chloride). Arrangement of elec-trodes as indicated in the diagram in Figure 6b. The relief of the decrement was obtained byallowing the nerve to perform in moist air a spontaneous recovery from the effect of sodium-deficiency.

In each row of records the intensity of the decrement in conduction is measured by the magni-tude and the time of appearance of the second phases of the spikes. Another sign of the spon-taneous recovery was the progressive decrease in the height of the electrotonic potential producedby the applied current. (Sodium-deficiency increases the apparent resistance of the nerve mem-brane.)

their action potentials, have different magnitudes. For example, when a frog nervekept in an atmosphere containing 5 per cent CO2 is caused to conduct a train ofimpulses at the frequency 60-100 per second the action potential increases pro-gressively in height to reach a maximum after a certain number of impulses havebeen conducted. Thereafter, according to the frequency and to the temperature theheight of the action potential may remain constant or may decrease somewhat (Lo-

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


5 4 3 2

_~~~~~~~~~~

7 5.5 5 5 1 5 mm.

FIG. 6b.-Continuation of Figure 6a. The decrement in conductionat each stage during the spontaneous recovery is measured by the heightof the spikes in each successive row of records. Diphasicity of thespike indicates that impulses were able to propagate themselves pastelectrode 5. Note that in the last record (5,7) the conducting fiberswere grouped in two volleys, the second being conducted at much lowerspeed than the first.

rente de N6,318 Chapters III and XV). The significance of the fact that the nervefibers may conduct impulses of various magnitudes, may not be clear if nerve func-tion alone is considered. The fact, nevertheless, is quite important for the phys-iology of the nervous system. As has been shown by Lloyd40' 41 one of the factorsinvolved in posttetanic potentiation of monosynaptic spinal reflexes is an augmenta-tion of the nerve impulse in the afferent nerve fibers.

Since the nerve fibers can conduct impulses of various magnitudes there is no

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

612 PHYSIOLOGY: LORENTE DE NO AND CONDOURIS PROC. N. A. S.

difficulty in assuming that the same is true for the bodies and dendrites of neurons.As a matter of fact, there are reasons to believe that decremental conductionnormally takes place in the neurons. When the action potential of motoneuronswas first recorded (Lorente de No;42 43 for recent literature, cf. Coombs, Curtis,and Eccles44) it was concluded that conduction of impulses through the body anddendrites differed from conduction in axons in two important respects: the wavelength of the nerve impulse is much shorter and the speed of conduction is muchlower than in the axon. Furthermore, during repetitive stimulation the actionpotential, in particular those deflections referable to invasion of the dendrites rapidlydecrease in height, indicating that the impulses die out at progressively greaterdistances from the dendritic tips. This finding suggests that conduction withdecrement takes place and that the rate of decrement is increased during the rel-atively refractory period, as it occurs in nerve that has been caused to conduct withdecrement (Adrian,8 Lucas9). In support of this conclusion there is a remarkablephenomenon discovered by Renshaw.45 The spread into motoneurons of impulsesinitiated in the axons is increased when a volley of excitatory impulses impingesupon the motoneurons, i.e., excitatory synaptic stimuli decrease the decrement inconduction of impulses through the body and dendrites of the motoneurons. Re-cently, using internal microelectrodes Eccles, Libet, and Young46 have obtainedevidence of spike-like partial responses of motoneurons which they believe to bedue to impulses generated in dendrites, and blocked during propagation towardthe body.The doctrine of decremental conduction suggests at least two mechanisms

for the summation of excitatory impulses delivered at distant synapses. Onemechanism is this. An impulse is initiated at a point of a dendrite by simultaneousactivation of a group of closely neighboring synapses. The impulse is conductedwith decrement and is extinguished before reaching the axon, but if the rate ofdecrement is decreased by subliminal synaptic excitation along the dendrite thatimpulse may reach the cell body and enter into the axon. The other mechanismis this. Synaptic excitatory stimuli are delivered to two branches of a dendrite.In each branch an impulse is initiated which is conducted with decrement. Inisolation each impulse would be extinguished shortly after entering into the parentbranch; but if the two impulses arrive simultaneously at the junction of the branchesthey may initiate in the parent branch a larger impulse which will travel farther.This impulse may also be extinguished before reaching the cell body, but if it isreinforced by another small impulse arriving by a dendritic collateral it may travelfarther and eventually reach the cell body. A mechanism of this kind may beoperative in certain peripheral, sensory nerve endings. An impulse, capable ofpropagating itself without significant decrement, will not be initiated in the parentaxon unless several rapidly decrementing impulses have been initiated in theproper time sequence in preterminal branches of the axon.The discovery by Lloyd47 of the existence of volleys of impulses, which exert a

specific inhibitory effect upon spinal motoneurons is probably best explained byassuming the existence of inhibitory synapses, which may either reduce the effec-tiveness of excitatory stimuli or reinforce the normally existing decrement in con-duction. Nevertheless, inhibition may also be caused by unspecific processes,which result, either in a reduction of the effectiveness of summation of excitatory

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


impulses (Gasser48) or in the disappearance of a background of subliminal excita-tion by internuncial bombardment (Lorente de N642). Inhibition, as was postu-lated by Lucas"' may also be an especial result of decremental conduction. Lucas'thought may now be paraphrased in this manner. An extinguished decrementalimpulse may block the spread of excitation by leaving after it a refractory period.It seems probable that some of therecent findings of Lindblom49 (Fig. AP16A) are referable to refractorinessleft by an extinguished decremental Iimpulse in branches of axons inner- Mvating touch receptors. -0 ASummary.-The best manner of

summarizing the experimental ob- D.]?servations on decremental conduc- A.P?tion in peripheral nerve is to repro-duce the diagrams given by Adrian5 IIto interpret his own experiment M(Fig. 7, I, II) and by Verwor Cinterpret Lodholz'6 experiment (Fig.7, III). The diagrams have been DYmodified to take into account the APfact that depolarizing agents wereused (ether, alcohol, an acidifiedsolution of morphine hydrochloride, III /anoxia), and that therefore the flow .of the demarcation current on the 0 1 2one hand must necessarily havecreated conditions for conductionwith decrement in the untreated A]?segments of the nerve adjacent tothe treated segments and, on the IV \other, it must have reduced the in- Mtensity of the decrement in the o 1 2neighborhoods of the margins of D.Pthe treated segment. In the dia- FIG. 7.-Diagrams intended to represent in agrams the crest height of the self- qualitative fashion the distribution of the demarca-

tion potential (D. P.) and the decrement duringpropagating action potential has conduction of the height of the action potentialbeen plotted using the nerve itself as (A. P.) when segments of the nerve have been

treated with depolarizing agents (I, II, III) or withthe axis of abscissae. The crest a nondepolarizing blocking agent (IV). o-cathodeheight of the action potential at the of the outside stimulating electrodes. 1,2-cathodes

of inside stimulating electrodes. M-muscle or re-cathode is left undetermined be- cording instrument.cause since this action potentialappears superposed upon the catelectrotonic potential created by the applied currentthe exact measurement of its crest height is fraught with unsurmountable dif-ficulties.

In the case considered in diagram I the impulse must pass through a long zone ofdecrement A before reaching the muscle (or the oscillograph). The impulse begins

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

614 PHYSIOLOGY: LORENTE DE N6.AND CONDOURIS Puoc. N. A. S.

to decrement in height before reaching the central margin of the treated segment,and within this segment it undergoes a more intense decrement so that it is ex-tinguished before reaching the peripheral margin. No muscle contraction is ob-served.

In the case of diagram II, the impulse has to pass through two short zones ofdecrement, B and C the sum of the two lengths B and C being equal to the lengthA. The impulse undergoes a decrement in segment B but it begins to increase inheight before reaching the peripheral margin of B, because the demarcation currentsexert a repolarizing action upon the B segment. After crossing through the pe-ripheral margin of B the impulse rapidly grows in size, but it cannot reach full sizebecause in the untreated segment, between B and C, the demarcation currentsexert a depolarizing action. -In segment C the impulse decrements more rapidlythan in B, because it reaches the central margin of C with a smaller magnitude;nevertheless, it reaches the peripheral margin of C with magnitude sufficient topropagate itself into the untreated segment, where eventually it reaches full magni-tude. A muscle contraction is observed.

In the case of diagram III an impulse initiated at cathode o is extinguished be-fore reaching the peripheral margin of the treated segment. A small impulseinitiated at cathode 1 by a small stimulus also is extinguished within the treatedsegment, but a larger impulse initiated at 1 by a strong stimulus or a small impulseinitiated at 2 is able to propagate itself into the untreated segment and eventuallyto reach full height.Diagram IV is diagram III as it applies to the case in which the zone of decre-

ment has been created by means of an agent which does not alter the value of theresting membrane potential (sodium-deficient solutions or nondepolarizing anes-thetics). Since no demarcation current flows, except for the negligible effect of theconcentration gradients present inside the walls of the narcotizing chamber, thetransition from decrementless to decremental conduction at the central margin ofthe treated segment is abrupt, and so is the transition from decremental to decre-mentless conduction at the peripheral margin.At the time of Adrian's5' 7 and Lodholz'6 writings the depolarizing action of the

blocking agents that they used was not known and therefore the theoretical argu-ments concerning the validity of the all-or-nothing law were based on the conceptsembodied in diagram IV (Fig. 7), while the actual experimental situations wererepresented by diagrams I to III (Fig. 7). Those situations afford conclusive proofonly of one fact, namely that decremental conduction through the treated segmentstakes place. Not even the situation illustrated by diagram IV can lead to a proofof the strict validity of the all-or-nothing law for peripheral nerve kept underphysiological conditions. All that the experiment can prove is that during con-duction along the nerve the impulse keeps a magnitude, which is constant withinthe rather wide limits allowed by the low accuracy of the measurements that canbe made with present-day techniques.Evidence has been mentioned that decremental conduction of impulses in the

bodies and dendrites of the neurons is a normally occurring phenomenon, and it hasbeen suggested that decremental conduction supplies mechanisms for the accom-plishment of the integrative role of the neuron.

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


* Present address: Seton Hall College of Medicine and Dentistry, Department of Pharma-cology, Jersey City, New Jersey.

t The diagram at the bottom of Figure 2 illustrates the classical technique, except for a detail,an oscillograph instead of a muscle was used as indicator of the effectiveness of stimula-tion through the outside electrodes (o) or inside electrodes (m).

I It should be emphasized that the technique of dissecting single nerve fibers from a nervetrunk introduces a serious artifact. The nerve fibers undergo far-reaching changes in propertiesduring the dissection. For example, Kato and Tasaki24 (pp. 76 and 99) found that 0.005 to 0.01per cent cocaine (approximately 0.15 to 0.3 mM) almost instantly suppresses the ability of thenodes to produce action potentials, while, as has been repeatedly demonstrated in this laboratory,cocaine at those concentrations, acting upon spinal roots or desheathed sciatic trunks for extendedperiods of time (60 to 120 min) does not block conduction by any fiber. It only produces a slightrise in the threshold of stimulation and a slight reduction in the speed of conduction. At a muchhigher concentration, 25 mM, cocaine blocks conduction in all the fibers of spinal roots withinless than 5 min, but the nerve fibers retain the ability to produce unconducted action potentialsat least for 100 min.

§ In view of the observations reported by Tasaki24 (p. 101) on decrementless conduction innarcotized, desheathed nerve trunks, especial emphasis has been placed in this laboratory on thestudy of decremental conduction in desheathed sciatic trunks. The demonstration of decre-mental conduction with desheathed sciatic nerves is even easier than the demonstration with spinalroots, which have no external sheath, or with nerve trunks with intact sheath. Figs. 6a and 6billustrate decremental conduction in a desheathed sciatic nerve. Similar results have been ob-tained with the use of ethylurethane or of any of the various depressant agents mentioned in thetext. In the experiments reported by Tasaki the concentration of the anesthetic was so highthat the limit length was less than 4-6 mm. With the use of weaker concentrations it is foundthat the limit length for decremental conduction of single impulses may be as long as 30 mm., andwith the use of still lower concentrations it is found that only after the decrement has been intensi-fied by a succession of refractory periods do impulses become extinguished within a 30 mm. longsegment of nerve (Wedensky inhibition phenomenon).

Il In a number of experiments reported by Kato14 16 measurements of the latency of the musclecontraction seemed to indicate that the applied current may spread so far as to initiate impulsesat points 20-30 mm. ahead of the cathode. Such spreads of current are referable to instrumentalartifacts, which were called escape of current, or, in German, Stromschleifen, by the classicalauthors. Although the natuie of the artifact was not understood at that time, it was well knownthat the artifact appears only when the stimulating shock is increased beyond a certain limit,which lies very far above the magnitudes of the shocks that were used in experiments on decre-mental conduction (cf. discussions of the problem by Lodholz6 and Frbhlich'9). The very fact,emphasized by Szpilman and Luchsingerl and by Dendrinos,3 that in the experiments done withthe classical technique outside stimulation and later inside stimulation became ineffective clearlyshows that escape of current did riot play a significant role in the classical experiments.At present, stimulation by escape of current is known to be the result of capacitative coupling

of the stimulating and recording circuits. The artifact can be minimized by reducing to lowvalues all the capacities to ground in the experimental setup. Stimulation by escape of currentcan be entirely prevented by grounding the cathode of the stimulating circuit and recording theaction potentials with a differential amplifier. The input capacities of the amplifier still cause acertain escape of current, which is responsible for the so-called shock artifacts, but the escape isfar too small ever to initiate impulses at the first recording electrode.

1 Szpilman, J., and B. Luchsinger, "Zur Beziehung von Leitungs- und Erregungsvermbgen derNervenfaser," Arch. ges. Physiol., 24,347 (1881).

2 Werigo, Br., "Zur Frage uber die Beziehung zwischen Erregbarkeit und Leitungsfahigkeit desNerven," Arch. ges. Physiol., 76, 552 (1899).

3 Dendrinos, G., "Ueber das Leitungsvermogen des motorischen Froschnerven in der Aether-narkose," Arch. ges. Physiol., 88,98(1902).

4 Verworn, M., Erregung und Lahmung. (Jena: Fischer, 1914).6 Adrian, E. D., "On the Conduction of Subnormal Disturbances in Normal Nerve," J. Ph/ysiol.,

45, 389 (1912).

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020


6 Lodholz, E., "Ueber die Gultigkeit des "Alles-oder-Nichts-Gesetzes" fur die markhaltigeNervenfaser," Ztschr. f. Biol., 15, 269 (1913). "Das Dekrement der Erregungswelle in ersticken-den Nerven," Ibid., 15, 316(1913).

7Adrian, E. D., "The All-or-None Principle in Nerve," J. Physiol., 47, 460 (1914).8 Adrian, E. D., "Wedensky Inhibition in Relation to the "All-or-None" Principle in Nerve,"

J. Physiol., 46, 384 (1913).9 Lucas, K., "The Effect of Alcohol on the Excitation, Conduction and Recovery of Nerve,"

J. Physiol., 46, 470 (1913).10 Wedensky, N. E., "Die Erregung, Hemmung und Narkose," Arch. ges. Physiol., 100, 1 (1903).11 Lucas, K., The Conduction of the Nervous Impulse (London: Longmans, Green, 1917).12 Adrian, E. D., and K. Lucas, "On the Summation of Propagated Disturbances in Nerve

and Muscle," J. Physiol., 44, 68 (1912).13 Forbes, A., "The Interpretation of Spinal Reflexes in Terms of Present Knowledge of Nerve

Conduction," Physiol. Rev., 2, 361 (1922).14 Kato, G., The Theory of Decrementless Conduction in Narcotised Region of Nerve (Tokyo:

Nankodo, 1924).15 Kato, G., The Further Studies on Decrementless Conduction (Tokyo: Nankodo, 1926).16 Davis, H., A. Forbes, D. Brunswick, and A. McH. Hopkins, "Studies of the Nerve Impulse,

II. The Question of Decrement," Am. J. Physiol., 76, 448 (1926).17 Kato, G., and D. Teruuchi, "Is There "Transitional Decrement" in Narcotised Nerve?," J.

Physiol., 64, 193 (1927).18 Ishikawa, H., and Co-workers, "Verandern sich die isobolischen Substanzen in die hetero-

bolischen wenn sie narkotisiert oder gelahmt werden," Japan. J. Biophysics, 2, LXXXIX (1927).19 Frohlich, F. W., Nervenreize, in Bethe, A., Handbuch der normalen und pathologischen Physi-

ologie, (Berlin; Springer, 1929) 9, 177.20 Baruch, R., "Erregbarkeit und Leitfahigkeit des narcotisierten Nerven," Ztschr. f. Biol., 89,

48 (1930).21 Wiersma, C. A. G., "Der Einfluss der Narkose mit Cyankalium auf den Nerven. Beitrag zur

Frage der Dekrementleitung im Nerven," Proc. Konink. Akad. v. Wetensch., Amsterdam, 33, I,180 (1930).

22 Reswjakoff, N. P., "Zur Theorie der dekrementiellen Leitung im Nerven," Arch. gee. Physiol.,226, 86 (1931).

23 Woronzow, D., "Ueber die Erregungsleitung im narkotisierten Nerven," Arch. gee. Physiol.,227, 132(1931).

24 Tasaki, I., Nervous Transmission (Springfield, Ill.: Thomas, 1953).25 Adrian, E. D., The Basis of Sensation (London: Christophers, 1928).26 Cajal, S. R., Histologie du systame nerveux de l'homme et des vert~br4s (Paris: Maloine, 1909,

1911).27 Lorente de N6, R., "Studies on the Structure of the Cerebral Cortex. I. The Area Ento-

rhinalis," J. Psychol. Neurol., Lpz., 45, 381 (1934).28 Sholl, D. A., The Organization of the Cerebral Cortex (London: Methuen, 1956).29 Bishop, G. H., "Natural History of Nerve Impulse," Physiol. Rev., 36, 376 (1956).30 Eccles, J. C., The Neurophysiological Basis of Mind (Oxford: Clarendon, 1953).31 Eccles, J. C., Physiology of Nerve Cells (Baltimore; John Hopkins, 1957).32 Bishop, G. H., "The Dendrite: Receptive Pole of the Neuron," EEG and Clin. Neurophy-

siol., suppl. 10, 12 (1958).33 Bremer, F., "Cerebral and Cerebellar Potentials," Physiol. Rev., 38, 357 (1958).34 Grundfest, H., "Electrophysiology and Pharmacology of Dendrites," EEG and Clin. Neuro-

physiol., suppl. 10, 22 (1958).36 Bremer, F., Some Problems in Neurophysiology (London: London Univ., 1953).36 Altamirano, M., C. W. Coates, and H. Grundfest, "Mechanisms of Direct and Neutral Ex-

citability in Electroplaques of Electric Eel," J. Gen. Physiol., 38, 319 (1955).3 Mueller, P., "Effect of External Currents on Duration and Amplitude of Normal and Pro-

longed Action Potentials from Single Nodes of Ranvier," J. Gen. Physiol., 42, 163 (1958).37 Hodgkin, A. L., "Ionic Movements and Electrical Activity in Giant Nerve Fibers," Proc.

Roy. Soc. B, 148, 1 (1958).

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

VOL. 45, 1959 ZOOLOGY: ODUM AND SMALLEY 617

38 Lorente de N6, R., "A Study of Nerve Physiology," Studies from the Rockefeller Institute forMedical Research, 131, 132 (1947).

39 Rudolph, G., "Zur Physiologie der Impulsentstehung und Impulsfortleitung im peripherenNerven," Ann. Univ. Saraviensis, 4, 1 (1956).

40 Lloyd, D. P. C., "Post-tetanic Potentiation of Response in Monosynaptic Reflex Pathwaysof the Spinal Cord," J. Gen. Physiol., 33, 1947 (1949).

41 Lloyd, D. P. C., "Early and Late Post-tetanic Potentiation, and Post-tetanic Block in aMonosynaptic Reflex Pathway," J. Gen. Physiol., 42, 475 (1959).

42 Lorente de N6, R., "Transmission of Impulses Through Cranial Motor Nuclei," J. Neuro-physiol., 2, 402 (1939).

43 Lorente de N6, R., "Action Potentials of the Motoneurons of the Hypoglossus Nucleus"J. Cell and Comp. Physiol., 29, 207 (1947).

44 Coombs, J. S., D. R., Curtis, and J. C. Eccles, "The Generation of Impulses in Motoneurons,"J. Physiol., 139, 232 (1957).

45 Renshaw, B., "Effects of Presynaptic Volleys on Spread of Impulses Over the Soma of theMotoneurons," J. Neurophysiol., 5, 235 (1942).

46 Eccles, J. C., B. Libet, and R. R. Young, "The Behavior of Chromatolized MotoneuronsStudied by Intracellular Recording," J. Physiol., 143, 11 (1958).

47 Lloyd, D. P. C., "A Direct Central Inhibitory Action of Dromically Conducted Impulses,"J. Neurophysiol., 4, 184 (1941).

4 Gasser, H. S., "Reciprocal Innervation," in Vol. jubil. publ. en l'honneur du Prof. J. Demoor,(Libge: 1937), 212.

49 Lindblom, U. F., "Excitability and Functional Organization within a Peripheral TactileUnit," Acta Physiol. Scand., Suppl. 153 (1958).

COMPARISON OF POPULATION ENERGY FLOW OF A HERBIVOROUSAND A DEPOSIT-FEEDING INVERTEBRATE IN A SALT MARSH

ECOSYSTEM*

BY EUGENE P. ODUM AND ALFRED E. SMALLEYf

UNIVERSITY OF GEORGIA, ATHENS

Communicated by G. Evelyn Hutchinson, February 16, 1959

Very frequently the autotrophic and heterotrophic components of an ecosystemare partially separated in space in that they are stratified one above the other(vegetation-soil on land, phytoplankton-sediments in water). Also, the basicfunctions are usually partially separated in time in that there may be a considerabledelay in the heterotrophic utilization of a large portion of the net production ofautotrophic organisms. Consequently, between the first and second trophic levelsthe energy flow of the community is often divided into two broad streams resultingin two types of primary consumption: (1) direct and immediate utilization ofliving plant tissues by herbivores and plant parasites, and (2) delayed utilizationof dead tissues and stored food by other consumers. In most ecosystems manyspecies of primary consumers rely exclusively on one or the other types of foodsource while other species utilize both, or shift from one to-the other seasonally.

In addition to the problem of time-lag in utilization, one of the difficulties indeveloping a satisfactory comparative population ecology of heterotrophs is thefact that different species, and also life history stages of the same species, associatedtogether in a given area of the earth's surface differ greatly in size and rate: of

Dow

nloa

ded

by g

uest

on

Janu

ary

25, 2

020

n6 - pnas · 11 cajal, s. r., "l'anatomie fine de la moelle 6pinilre," atlas der...

Documents