(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Quarterly journal of experimental physiology and cognate medical sciences"

QUARTERLY 
JOURNAL OF 
EXPERIMENTAL 
PHYSIOLOGY 



EDITORS 

E. A. SCHAFER, EDINBURGH 

F. GOTCH, OXFORD 

W. D. HALLIBURTON, LONDON 
C. H. SHERRINGTON, LIVERPOOL 
E. H. STARLING, LONDON 
A. D. WALLER, LONDON 



VOLUME I 



LONDON: CHARLES GRIFFIN AND COMPANY, LIMITED 

EXETER STREET, STRAND 

1908 



Ql- 



o 



y^ "p 



W 




928615 



CONTENTS OF VOL. 



Florencb Buchanan. On the Time taken in Transmission of Refiex 
Impulses in the Spinal Cord of the Frog. (From the University 
Museum, Oxford. ) With 1 3 figures in the text .... 1 

C. S. Sherrington. Some Comparisons between Reflex Inhibition and 
Reflex Excitation. (From the Physiology Laboratory, University of 
Liverpool) With \0 Jigures in the text . . .67 

John Tait. The Freezing of Frog's Nerve, with special reference to its 
Fatigability. (From the Physiology Department, University of 
Edinburgh.) With b figures in the text .79 

R. A. Wilson and W. Cramer. On Protagon : Its Chemical Composition 
and Physical Constants, its Behaviour towards Alcohol, and its 
Individuality. (From the Physiology Department, University of 
Edinburgh) ... ... 97 

J. A. GuNN. The " Fly-catching " Reflex in the Frog. (From the 

Pharmacology Department, University of Edinburgh) . . .111 

F. H. A. Marshall and W. A. Jolly. On the Results of Heteroplastic 
Ovarian Transplantation as compared with those produced by Trans- 
plantation in the same Individual. (From the Physiology Department, 
University of Edinburgh.) With 1 figure in the text 115 

P. T. Herring. The Histological Appearances of the Mammalian Pituitary 
Body. (From the Physiology Department, University of Edinburgh.) • 
With \& figures in the text 121 

P T. Herring. The Development of the Mammalian Pituitary and its 
Morphological Signiricance. (From the Physiology Department, 
University of Edinburgh.) With \\ figures in the text 161 

P. T. Herring. The Physiological Action of Extracts of the Pituitary Body 
and Saccus Vasculosus of certain Fishes. Preliminary Note. (From 
the Physiology Department, University of Edinburgh) 187 



W. Cramer. Note on the Action of Pituitary Extracts upon the Enucleated 

Frog's Eye. (From the Physiology Department, University of Edinburgh) 189 

John Tait and Jas. A. Gunn. The Action of Yohimbine on Medullated 
Werve, with special reference to Fatigability. (From the Physiology 
Department, University of Edinburgh.) With 5 fiyures in the text 191 

W. Page May and C. E Walkbr. Note on the Multiplication and 

Migration of Nucleoli in Nerve Oells of Mammals. With 2 plates 203 

Florence Buchanan. The Electrical Response of Muscle to Voluntary, 
Reflex, and Artificial Stimulation. (From the University Museum, 
Oxford.) With 10 figures in the text 211 

W. EiNTHOVEN (in collaboration with A. Flohil and P. J. T. A. Eattabrd). 
On Vagus Currents examined with the String Galvanometer. (From 
the Physiological Laboratory of the University of Leyden.) With 1 figure 
in the text . . . . 243 

John Tait. A Simple Method of observing the Agglutination of the 
Blood Corpuscles in Gamraarus. (From the Physiology Department, 
University of Edinburgh) ...... 247 

A. D. Waller. Ou the Time taken in Transmission of Reflex Impulses in 

the Spinal Cord of the Frog. With i figures in the text . . 251 

Sutherland Simpson and Francis H. A. Marshall. On the Effect of 
stimulating the Nervi Erigentes in Castrated Animals. (From the 
Physiology Department, University of Edinburgh) . . . 257 

P. T. Herring. A Contribution to the Comparative Physiology of the 
Pituitary Body. (From the Physiology Department, University of 
Edinburgli.) With 1 plate and % figures in the text . .261 

P. T. Herring. The Effects of Thyroidectomy upon the Mammalian 
Pituitary. Preliminary Note. (From the Physiology Department, 
University of Edinburgh.) With 2 plate.i . . . .281 

C. E Walker and Alice L. Embleton. Observations on the Nucleoli in 
the Cells of Hydra Fusca. (From the Laboratory of Cytology, 
University of Liverpool.) With 1 plate .... 287 

S. Kajiura, Imperial Japanese Navy. Is Choline present in .the Cerebro- 
spinal Fluid of Epileptics? (From the Physiological Laboratory, King's 
College, London.) With 1 plate . . . .291 

Otto Rosenheim and M. Christine Tebb. Ou so-called " Protagon." 
(From the Physiological Laboratory, King's College, London.) With 
1 plate ......' .297 



T. Addis. The Coagulation Time of the Blood in Man. (From the 
Physiology Department, University of Edinburgh.) With 5 figures in 
the text and 2 plates ....... 305 

W. Emerson Lek. The Action of Tobacco Smoke, with special reference to 
Arterial Pressure and Degeneration. (From the Pharmacological 
Laboratory, Cambridge.) With li figures in the text and 1 plate . 335 

LocKB, F. S. Contributions to Physiological Technique. With 8 figures 

in the text ........ 359 

W. EiNTHOVBN and W. A. Jolly. The Form and Magnitude of the 
Electrical Response of the Eye to Stimulation by Light at Various 
Intensities. (From the Physiological Laboratory of the University of 
Ley den.) With 2b figures in the text . . . . .373 



ALPHABETICAL TABLE OF CONTENTS 



AuDis, T. Coagulation Time of the Blood in Man 
Buchanan, Flokknce. Electrical Response of Muscle 

Transmission Time of Reflexes in Frog .... 

Cramer, W. Action of Pituitary Extracts upon Enucbated Frog's Eye 

EiNTHOvKN, W. (with A. Flohil and P. J. T. A. Battaerd). Vagus Currents 
cxaniihed with the String Galvanometer 

EiNTHOvEN, W., and W. A. Joluy. Electrical Response of the Eye 
Stimulation by Light ..... 

GuNN, J. A. " Fly-catching " Reflex in the Frog 

Herring, P. T. Comparative Physiology of the Pituitary Body 

Development of the Mammalian Pituitary 

Effects of Thyroidectomy upon the Mammalian Pituitary 

Histological Appearances of the Mamm.alian Pituitary Body 

Physiological Action of Extracts of Pituitary Body in Fishes 

Kajiuea, S. Is Choline present in the Cerebro-spinal Fluid of Epileptics ? 

Lbe, W Emerson. Action of Tobacco Smoke 

Locke, F S. Contributions to Physiological Technique 

Marshall, F. H. A., and W, A. Jolly Heteroplastic Ovarian Ti-ansplanta 
tion ....... 

May. W. Page, and C E. Walker. Nucleoli in Nerve Cells of Mammals 

Rosenheim, Otto, and M. Christine Tebb. On so-called " Protagon " 

Sherrington, C. S. Reflex Inhibition and Reflex Excitation 

Simpson, Sutherland, and Francis H. A Marshall. Nervi Erigentes in 
Castrated Animals ...... 

Tait, John. Agglutination of the Blood Corpuscles in Gammarus . 

Freezing of Frog's Nerve ..... 

and Jas. A. Gunn. Action of Yohimbine on Medullated Nerve 

Walker, (X E., and Alice L. Emblbton. Nucleoli in the Cells of Hydra 
Fusca ....... 

Waller, A. D. Transmission Time of Reflexes in Frog 

Wilson, R. A., and W. Cramer. On Protagon 



QUARTERLY JOURNAL 
OF EXPERIMENTAL PHYSIOLOGY 

ON THE TIME TAKEN IN TRANSMISSION OF REFLEX IM- 
PULSES IN THE SPINAL CORD OF THE FROG. By 
Florence Buchanan. (From the University Museum, Oxford.) 

CONTENTS. 

PAGE 

I. Introduction — 

General method, and preliminary experiments on excised nerve-muscle 

l^reparatious ........ 1 

II. The Same-Limb Reflex Time in the Normal Cord — 

Methods of preparation and of recording, and of recognising the reflex 

effect in the records. Interpretation of the records ... 5 

Results, estimation of time to be deducted for transmission in nerve. 
Inquiry as to the efifects of alteration in strength of stimulus, tempera- 
ture, fatigue . . . . . . . . 10 

III. The Same-Limb Reflex Time in the Strychnine Cord — 

The influence of the drug on the reflex efi'ect and on the reflex time. 
Inquiry as to eftects of alteration of strength of stimulus, temperature, 
fatigue on synapse time . . . . . . . 18 

IV. The Same-Limb Reflex Time in the Phenol Cord ... 29 

V. The Crossed Reflex Time in the Normal, the Strychnine, and the 

Phenol Cord — 
The difliculty in obtaining the true reflex efi^ect in a normal cord. The 

extra delay in the stryclinine cord and inquiry as to the influence upon 

it of strength of stimulus (proximate and ultimate) and of temperature 30 

Significance of results, difl^erentiation of synapses and the relative actions 

of drugs and of fatigue upon them ..... 55 

Discussion as to the significance for normal cords of results obtained ^nth 

strychnine and phenol cords ...... 58 

VI. Concluding Remarks ....... 63 

VII. Summary ......... 64 

VIII. Explanation of Figures ....... 66 

I. Introduction. 

The estimate which seems to be generally accepted ^ of the reduced reflex 
time in the lower part of the frog's spinal cord in a reflex contraction 
of the simplest kind is that which was formed by Wundt in 1876,- after 
making experiments in which the mechanical responses of a gastrocnemius 

» See article on "Nerve Cell" in Schiifer's Text Book of Physiology, 1900, 
vol. ii., p. 609. 

■•^ W\indt, Unters. z. Mechanik d. Nerven u. Nervencentren, Abth. II. : " Ueber d. Re- 
flexvorgang u. d. Wesen der centralen Innervation," Stuttgart, 1876. 

VOL. I. — JAN. 1908. 1 



2 Buchanan 

were recorded, when, on the one hand, its sciatic nerve, and, on the other, 
one or two of the posterior roots of this same nerve, were stimulated by 
a single break induction shock. 

The question of the duration of this reflex time, and of the larger one 
dependent on it, of the time taken to pass a synapse or neurone-junction,^ 
is one of sucli importance, not only to the physiologist but to the psy- 
chologist,^ that one ought to be very sure that the answer to it has been 
supplied by methods which would be likely to give it correctly. 

Wundt's records were made on a pendulum myograph, the two curves 
being on the same abscissa, and the difference of time taken, after the 
moment of stimulation, for the muscle to begin to lift the lever in the 
two cases, was measured. The difficulty in determining latencies with 
exactitude by such a method is well known, but the chief objection to 
which Wundt's experiments seem to me to be open is that he intention- 
ally chose, for comparison with the reflex contraction, a contraction to 
stimulation of motor nerve of equal amount, and therefore a submaximal 
one, his reason for so doing being, that he found variations of direct latency 
with varying strength of stimulus to be more marked than the differences 
between the two kinds of latency.^ With the method he used of record- 
ing the mechanical response, the increase of latency observed when the 
stimulus to the motor nerve was submaximal instead of maximal was 
probably * due to a smaller number of fibres being excited ; for those 
excited, having the resistance of the whole of the rest of the muscle to 
overcome, would fail to move the lever as soon as when all, or a greater 
number of fibres are in action. It does not necessarily follow, and it seems 
to me improbable (see p. 27), that the smaller contraction evoked when the 
sensory root is stimulated owes its smallness to the same cause, so long 
as no stronger contraction can be obtained by strengthening the stimulus 
to the root. A further objection to Wundt's method is that, as he himself 
had shown in an earlier work,^ the latency may vary a good deal with the 
particular piece of one and the same nerve stimulated, and in his later 
experiments the stimulus was applied to spots of different excitability^ 
(loc. cit., p. 45). 

Wundt's experiments showed that the so-called "total latency," i.e. the 
time which elapses between the excitation of the posterior root and 
the beginning of the contraction of the whole muscle, varied in different 
preparations'" between 0*025 and 0*050 second. From this time he de- 
ducted what he found in each case to be the time which elapsed before 
the whole muscle began to contract, when the motor nerve was excited 

1 Schafer(loc. cit.). 

2 See, e.g., M'Dougall, Brain, xcvi., pp. 588-9, 1901. 

3 Wundt, loc. cit., p. 16. 

* Judging from some experiments made in the Oxford Physiological Laboratory, of 
which an account is given in the Journal of Physiology, vol. xxviii., 1902, p. 412. 

5 Wundt, Unters., etc., Abth. I. : Erlangen, 1871, pp. 192, 193. 

^ He says, " with the strength of the stimuhis," but none of the experiments of which 
he has published the details seem to me to bear out this statement. 



Transmission-time of Reflexes in Spinal Cord of Frog 3 

by a submaximal stimulus of such strength as to give a curve of the same 
height as the reflex contraction, and he thus obtained the result that the 
delay in the spinal cord, together with the time taken in traversing part 
of the dorsal roots, the ventral roots, and a small part of the sciatic nerve ^ 
(a time which, he says, was too small to be measurable) varied between 
0"008 and 0015 second. It is Wundt's lowest estimate of this same- 
limb reflex time to which hitherto most importance has been attached. 
As this seemed to be hardly justifiable, I have applied another method 
which I regard as less open to objection (one which I had been using for 
other purposes for several years in conjunction with Sir John Burdon- 
Sanderson, by whom it was first introduced for measuring time relations 
in physiological processes), to the purpose of determining the time-value 
in question, and others in more complex reflexes. 

As indicator I have used, not the mechanical, but the electrical response 
of a muscle, recording on a photographic plate, moving at a known and 
equable rate, the movement imparted to the meniscus of the capillary 
electrometer the moment an electrical change occurs at a spot connected 
with this instrument. The fact that the electrical response in the organ, 
as well as its manifestation in the recording instrument, occurs without 
any delay as soon as the recording spot is reached, and that it is not 
necessary for the whole muscle to be implicated before a record can be 
obtained, obviates what seems to me to be the principal objection to 
Wundt's experiments. The electrical response has the further advantage 
over the mechanical response, for the measurement of brief time intervals 
between different events occurring in a muscle, that the eflfect not only 
begins without delay, but, when in existence, outlasts the stimulus (which, 
either directly or indirectly, produces it) by a so much shorter time. A 
second effect, therefore, occurring in the muscle only a few thousandths 
of a second after a first, would have quite a distinct manifestation when 
recorded by such an instrument as the capillary electrometer, whereas in 
the record of the contraction of the muscle two such effects would be 
merged into one. This being so, there should be no difficulty in recording 
on the same photographic plate, and measuring the time interval between, 
the two electrical effects produced at one and the same spot of the muscle 
in response to simultaneous excitation of eflferent and afferent nerve 
respectively ; nor is there any such difficulty, provided that the cord is 
sufficiently sensitive for an effectual response to be obtained from it at all 
when the stimulus applied to the aflferent nerve is single and instan- 
taneous, as for the purpose in hand it must be. 

To eliminate any effect which the physiological or the physical condition 

' The frogs he used were large, being sometimes as much as 21 cm. long, so that 3 cm. 
more of nerve may have been traversed in the case of the reflex response than was traversed 
when the (ujjper part of the) sciatic nerve was stimulated. If one may infer, since it is 
not otherwise stated, that he used the same species of frog and under the same conditions ot 
temperature as were used for the experiments in the first part of his treatise, the species 
was R. viridis, and the temperature between 15° and 17^° C 



4 Buchanan 

o£ a particular piece of nerve excited might have on the time of arrival of 
the response in the muscle, and to ensure that the two stimuli were applied 
.■simultaneously, the simplest method to adopt seemed to be that of exciting 
a mixed nerve at one and the same spot, by one and the same stimulus, 
■•so as to obtain in succession in the muscle, first the direct and then the 
ireflex eftect. 

To be able to rely upon this method it was essential to know whether 
or not there is any difference in the rate of propagation of an impulse, 
-either through nerve, end-organ, or muscle, according to the strength of 
the stiuuilus producing it; for the strength of the stimulus immediately 
jproducing the reflex eflect could hardly be so strong in a normal cord as 
.that which has to be artificially applied to the mixed nerve in order to 
produce a reflex effect at all. With this object I made some preliminary 
experiments on excised nerve-muscle preparations, and found that, with 
a big resistance in the secondary circuit (as was employed in all the 
experiments referred to in this paper, unless the contrary is expressly 
stated), there is no difference — at least none that could be measured on 
plates travelling at the rate of about 85 cm. a second — in the time taken 
hy an impulse just strong enough to produce an effect at all, and that 
taken by one strong enough to produce a " maximal " effect, when each 
traverses in turn the same portion of a particular nerve and muscle — 
provided, however, that the preparation was a sensitive one. In less 
sensitive preparations a slight difference was occasionally manifested, but 
one hardly amounting in any of my preparations to as much as a 
thousandth of a second.^ As no electrical response to an excitation pro- 
duced reflexly by the application of a single break induction shock to 
afferent nerve can be obtained at all, except in very sensitive pre- 
parations, there seemed therefore to be no need to use for comparison 
with such response one produced by the application to the efferent nerve 
of a stimulus of smaller strength, as Wundt had considered necessary. 

In the same and other experiments made with excised yjreparations, I 
found, however, somewhat to my surprise, seeing that the exciting current 
was of such brief duration, that the transmission time is appreciably and 
very definitely affected by the direction of the induction current applied 
to the nerve when this is at all strong, i.e. when it is nearly strong enough, 
or just strong enough, to be felt on the tongue. A strength of excitation 
so great as this was seldom used in my experiments. When used, the 
extra delay, which sometimes even amounted to nearly two-thousandths 
of a second, and always occurred in passing the spot to which the ex- 
citing needle connected (indirectly) with the zinc of the battery was 

1 The difference, such as it is, when present, is probably one in end-organ delay, 
since Engelmann (A. f. d. ges. Physiol., Ixvi., p. 574, 1897) has shown with the curarised 
sartorins, excited by maximal and submaximal stimuli, that there is no difference in the 
rate of propagation of the mechanical response in muscle, and Gotch (J. Physiol., xxviii., 
p. 402) has shown that there is none in the rate of propagation of the electrical response 
in nerve. 



Transmission-time of Reflexes in Spinal Cord of Frog 5 

applied, svas so easily detected and measured in the direct response of the 
muscle, that allowance for it could be made in estimating the time which 
elapsed in the cord. 

II. The Same-Limb Reflex Time in the Normal Cord. 

It is essential for measuring the transmission-time in the cord that the 
.stimulus should be instantaneous and single. To overcome the well-known 
difficulty in producing a reflex eflfect by a stimulus of this kind without 
the aid of drugs, I have had recourse, in part, to a method recommended by 
Biedermann,^ which consists in keeping the decerebrated frog at a low 
temperature (2°-6° C.) for from one to Ave days before making the experi- 
ment. (Animals caught in warm weather were kept in the cold for some 
days or weeks before being decerebrated.) 

The species used was Rana temporaria. For the most part the 
specimens were small, the body length being about 6 cm. They were 
decerebrated by section through the optic thalami in the way recommended 
by Goltz,'^ the great advantage of which is that it is easily accomplished, 
with little shock, and little, if any, loss of blood or disturbance of circulation. 

After ligaturing the iliac artery on one side, the sciatic nerve was freed, 
so that a pair of needle electrodes could be placed on it without coming in 
contact with any other part of the preparation. The corresponding 
gastrocnemius muscle was then prepared, and, the whole preparation being 
then placed in a moist chamber, its tendon end and a spot on its dorsal 
surface were connected by non-polarisable electrodes with the Hg and 
H2SO4 respectively of a capillary electrometer, which was clamped on to 
the stage of a horizontally-placed projecting microscope. The vertical slit 
upon which the column of mercury was projected, and the cylindrical lens 
about 10 cm. behind it, were fixed in the front wall of a long dark-box. 
Inside this box a trolley for carrying the photographic plate was arranged 
to run at equable rates, special care being taken that in so doing it should 
produce no vibrations. The distance of the plate was such as, with the 
objective used, to magnify the image about 300 times. As the trolley 
passed the slit, it broke a platinum contact which had been completing a 
circuit containing a single Daniell cell and the primary, core-less, coil of 
a Kronecker inductorium. The induction current that it thus produced 
in the secondary circuit was used for exciting the sciatic nerve, the flne 
(steel) needles which served as exciting electrodes being placed, 2-3 mm. 
apart, on the nerve. There was always, unless the contrary is expressly 
stated, a large resistance in the secondary circuit.^ The length of the nerve, 
the position of each needle with regard to it, the length of the nuiscle, and 
the distances of the two tied-on, leading-ott" electrodes from its ends, were 
all carefully measured and noted. While the plate was pas.sing behind 

» Biedermanii, A. f. d. ges. Pliysiol., l.\.\x., 1900, y. 408. 

-' Goltz, Beitnige zur Leliie der Nervenceutreii des Frosches, Berlin, 1809. 

■■' I used one of 60,000 oluns. 



Q Buchanan 

the slit, a wheel-like cardboard disc (called by Dr Garten of Leipzig, who 
first introduced it for this particular purpose, an " episkotister ") was 
rotating in front of it at such a speed that the light was obscured by a 
spoke between 700 and 850 times a second. The exact rate at which the 
spokes were passing the slit at the time it was taken could always be 
determined with accuracy on the developed photograph, either by means 
of the record on the plate of an electromagnetic signal set into vibration 
by a 100 fork, or by the vibrations of a rod in connection with the break- 
key inside the dark-box, the primary function of which was to mark on 
the plate the moment at which the nerve was excited. The value of 
O'l second was determined (with the aid of the 100 fork record) in terms 
of the vibrations of this rod as they appear on the photographs on several 
occasions, and was found to be quite constant (so long, of course, as no 
alteration was made in the length of the vibrating rod). Whenever the 
key for any reason had to be interfered with, a fresh estimation of this 
value was again made. The number of spokes obscuring the light in 
0"1 second, as determined by the lines on the plate, was counted, and the 
value of the interval between two such lines, in time, was calculated for 
every photograph. 

Two methods of distinguishing a reflex response in a record suggested 
themselves, and, accordingly, when the preparation was ready, the experi- 
ment was continued in one of two ways. In the one case, the strength 
of induction shock to the nerve required for a " maximal " contraction 
of the muscle was first determined. Then, the preliminary experiments 
having shown that the maximal electrical effect is produced by the same 
strength of stimulus as the maximal mechanical efl'ect, the electrical re- 
sponse to a stimulus of this strength, or to one but little stronger than 
it, was recorded. By means of the graduated scale with which the induc- 
torium was provided, the secondary coil was next so far pushed up that 
the induction current might be three or four times as strong, and another 
record was taken. Responses to just maximal and to supra-maximal 
stimuli were then recorded alternately three or four times, sometimes also 
those to still stronger stimuli, in which case one record at least was taken 
with the direction of the current reversed. 

Two consecutive records obtained in one experiment of this kind are 
reproduced in fig. 1. (A) represents the response to a stimulus just 
over the "maximal," (B) that to one three times as strong. The direct 
response of the muscle to stimulation of the efterent part of the nerve 
began under the proximal electrode, in each case four-thousandths of a 
second (4(t) after the excitation of the nerve. In the second record there 
is seen in addition a much smaller effect, the position of which shows that 
the muscle was again electrically active under the proximal electrode 2S<t 
later than when it became so the first time, 27cr therefore after the 
excitation. In the particular experiment to which these records refer it 
was only in the two first responses to the supra-maximal stimulus that this 



Transinis.sioii-time of Reflexes in Spinal Cord of Frog 7 

second effect occurred, l)ut in each it occurred at the same time after the 
hrst. It did not show itself in any of the responses to the just maximal 
stimulus. That the second effect, which (in records taken without the aid 
of a drug to increase the excitability of the cord) was seldom larger than in 
this experiment, is the reflex effect is perhaps more convincingly showna by 
the records obtained when the second method of experiment was employed. 
This consisted in first recording the response when the nerve was 
excited by a stimulus five or ten times as strong as would probably be 
required to produce a maximal response to excitation of the motor nerve. 
Then, after obtaining not more than four records of this response, one of 
which was taken with the direction of the induction current reversed, the 




Fig. 1. —First and second electrical responses of the gas- 
trocnemius of a normal preparation, obtained when 
the intact sciatic nerve of the same side was excited : 

A, by a stimulus which was little more than just maximal 
(1000 units). [Time lines 760 per second.] B, by a stimulus 
of three times the strength. | Time lines 730 per second.] 



sciatic nerve was divided between the exciting electrodes and the cord, and 
the response obtained when the peripheral end was excited by the same 
supra-maximal stimulus applied to the same spot on it was recorded two or 
three times. The small second effect which was sometimes there in all the 
records taken with the nerve intact was never seen after its severance from 
the central nervous system. The disadvantage of this method is that one 
cannot alternate the two things which have to be compared as one can 
with the other method ; and since the reflex efl'ect disappears long before 
the direct effect, seldom being obtainable in a normal preparation in 
response to more than a very few excitations of the mixed nerve, it is 
important in using it to excite as few times as will suflice to make 
absolutely sure that the effect, if present, is by no possibility accidental, 
and to ascertain the time of its occurrence, before dividing the nerve. 



8 Buchanan 

Controlled, however, by experiments made according to the first method, 
there is no room for doubt as to what in the records represents the reflex 
response of the muscle. Moreover, it ,gives evidence as to the reflex 
response which is not obtainable by the first method alone. For a supra- 
maximal stimulus is apt to produce in the record of the electrical response 
of the gastrocnemius not one, but two, effects which are not there when the 
stimulus is just maximal. The second such effect, instead of being much 
smaller than the direct one and occurring some two-hundredths of a second 
after it, is usually about equal to it in amount (as indicated by the steepness 
of the rise in the curve), and occurs some six- to ten-thousandths of a second 
after it. Two records of responses to supra-maximal stimuli in which this 




Fic. 2. — Fourth and seventh electrical responses of the 
gastrocnemius of a normal preparation (Exp. 14) 
obtained : 

' A, when the Intact sciatic nerve of the same side was exciteil 

by a supra-maximal stimulus. I'J'inie lines 770 per second.] 
B, when the peripheral end of the same nerve, after divid- 
ing it, was excited at the same place by a stimulus of the 
same strength. [Time lines 780 per second.] 

effect is seen are reproduced in fig. 2. The upper curve represents the 
fourth of four very similar responses obtained from a gastrocnemius when 
its intact sciatic nerve was excited ; the lower one the third of three very 
similar responses obtained from the same muscle excited at the same place 
and in the same way, but after the division of the nerve above the place of 
excitation. In both, and indeed in all the records obtained with this 
muscle (except the very first, which was that of a response to a weaker 
stimulus), the second effect, resembling the first direct effect, is seen. It 
indicates, therefore, that, whether the nerve was in physiological connection 
with the cord or not, the contact between the muscle and the proximal 
electrode became again negative to the distal, and by about the same 
amount as when the impulse first reached that spot along the motor nerve, 
and, moreover, that it did so after an interval of 6a: In all the records 



Transmission-time of Reflexes in Spinal Cord of Frocr 9 

taken before the severance (including the very iirst), but in none taken 
after it, there is seen a much smaller, and again double, after-effect. This 
indicates that when the nerve was intact, the muscle, as sampled by the 
two spots connected with the electrometer, began to undergo the same sort 
of disturbance in its electrical equilibrium as was the case previously m 
response to the excitation of the motor nerve, but one smaller in amount, 
and that this occurred 16 to 19 -So- after the first. In the particular 
response to which the photograph here reproduced refers, it began 1 Do- 
later. Its complete absence in the records taken after the severance of the 
nerve from the cord makes the central origin of the stimulus which 
immediately provoked it almost a certainty. 

What the effect, which I propose to call the second peripheral effect, 
signifies, does not here directly concern us. I believe it to be in some way 
dependent on the arrangement of the fibres in the gastrocnemius muscle, 
and perhaps on the special spot to which the proximal electrode happened 
to have been applied, since, with this muscle, which I had seldom used 
before for electrical purposes, I have now come across it frequently in 
response to strong stimuli, whereas I do not recall ever having seen it in 
the thousands of records I have taken of the electrical response of sartorius 
muscles to single break induction shocks, nor has it been present in any of 
the few excised nerved preparations of sartorius muscles of these cooled 
frogs which I have tried. I do not, however, wish to express a definite 
opinion on this point, as I have not really investigated the matter, and the 
occurrence of a succession of electrical changes resembling one another in 
response to a single instantaneous stimulus in the sartorius muscles of frogs 
suffering from drought^ suggests another interpretation. Whatever may 
give rise to it is quite independent of the central nervous system, and 
that is all we need know for our present purpose. The strength of the 
stimulus required to produce it is sometimes less and sometimes more than 
that required to produce a reflex eflect. 

I have • thought it superfluous for the chief object I had in view, to 
deduce from the capillary electrometer records the curves indicating the 
actual difterences of potential prevailing between the two spots of the 
muscle led off from during each response. To anyone accustomed to 
interpreting such records it is easy to find, without such deduction, the 
place which indicates the coming into existence of a difterence of potential 
of one particular sign, and thus to ascertain the time at which it occurs ; 
also, if need be, to see where, and consequently when, its maximum is 
reached, when it ceases to exist, or is reversed. To those who are not fully 
conversant with the reading of such records, it may be of some assistance 
to state here briefly that every rise on the photographic curve indicates a 
movement of the meniscus which inscribes it towards the orifice of the 
capillary — an adostial movement — and that such a movement, with the 
1 See Durig, Arch. f. d. ges. Physiol., xcvii., p. 457, 1903. 



XO Buchanan 

particular arrangement adopted of connecting the muscle with the electro- 
meter, always denotes that the contact between the proximal electrode and 
the muscle was during that time (galvanometrically) negative to the other 
contact. When the mercury is moving most quickly, when the rise in the 
record is steepest, this negativity is greatest. When the mercury is moving 
most slowly (as indicated, for instance, by a summit on the curve), this nega- 
tivity is least, and the difference of potential is either just about to cease, or 
is being reversed. A descending curve, according to its steepness, i.e. an 
abostial movement of the mercury, according to its quickness, denotes 
either that the distal contact is now negative to the proximal, or that there 
is still no difference of potential between them and that the meniscus is 
returning in its own time to its original position. It means the former before 
it means the latter when the contacts are made with two parts of the muscle, 
each of which in turn becomes electrically, then mechanically, active. 

If the electrical disturbance produced by the excitation of the motor part of 
the nerve is over, under both electrodes as it usually is, before that produced 
by the excitation of the sensory part has begun to manifest itself under 
the proximal electrode, the beginning of this second electrical event is 
luarked on the curve either by a fresh rise (as in figs. 1, B. and 2, A) or by 
a checking of the course of the descending curve. 

Having learned to identify the reflex effect in the records, there is no 
difficulty in ascertaining the time which elapsed between the moment at 
which the primary direct response reached the first recording spot on the 
muscle, and that at which the reflex effect reached the same spot ; i. e. the 
time taken for the impulse to travel a known length of nerve to the cord, and, 
after passing through the cord, to travel the same length of nerve back again. 

The shortest time interval between the arrivals of the two effects at 
the proximal electrode in those preparations which were not under the 
influence of any drug was 14o- ; but it was only in one response (the first) 
in one preparation [Exp. 15] that I obtained a value so low,^ it being about 
1<T longer in the three other responses recorded with the same preparation. 
In a record obtained in one other experiment [Exp. 31], the interval 
appeared to be equally short, but this was an experiment in which the 
comparison of the records taken with the exciting current in opposite 
directions showed that there was a block at the kathode which it took the 
impulse Icr to overcome in going straight to the muscle, so that probably 

1 The measurements on which the time vahies in all my experiments depend were 
made for each record not only by myseh' but independently by another person who knew 
nothing about their !-ignificance or the conditions under which the experiments were made. 
(I have been enabled to obtain this, and other, valuable assistance in the measurement of 
records, by the kindness of Dr Osier.) On each occasion in which our final results 
differed by more than 0"5o-, and especially if there were any relative difference in the 
values we each obtained in any one experiment, the whole process was gone over again by the 
one or the other of us. The number of thousandths of a second in the time values given 
in the text or in the tables may therefore be relied upon, but the fractions of a thousandth, 
when given, make no profession of being absolutely accurate. 



Transuiissioii-tiine of Reflexes in Spinal Cord of Frog 



11 



the real difference of time in this one response, the only one taken with 
the current descending, was nearer ISo- than 14(r. In the other responses 
obtained from this preparation (by means of ascending break induction 
shocks to the nerve), the interval between the two arrivals appeared to be 
about IGor in two and about 19cr in the third. It was probably slightly, 
but not so much as lo-, shorter (see p. 14). 

In most of the preparations the interval was between lOcr and 21o-, but 
in three it was 23o-, and in three others it was longer still. Five, out of 
twenty-three, of the preparations gave records which showed no sign of 
any reflex effect ; and in the records obtained with two others, the reflex 
effect, though represented in some, was too small for the place at which 
the curve began to alter its course to be determined with accuracy. 
Each cord which yielded more than one measurable reflex effect gave it in 
about the same time in all the responses. In seven preparations from 
which records of the reflex response were obtained more than once, the 
recurrence of the same interval (even under different conditions as to strength 
of stimulus) was very striking. This may be judged of by the measurements 
obtained from the records in one experiment (the longest of the seven), the 
conditions obtaining, and the results obtained in which are given here in 
tabular form. 



Exp. 58. Dec. 17, 1906. Room temp. 12° C. 



Induction current 
to nerve. 


Length, in millimetres, of 


Time, in thousandths of a second (o-), 


Strength.! 


Direction 
(descending 
or ascend- 
ing) 


muscle 
from en- 
trance of 
nerve to p.2 


nerve 

from Cu 
electrode 
to muscle. 


nerve 
from Cu 
electrode 
to cord." 


taken by 

impulse to 

reach p 

directly 

(measured). 

4-4 

4-4 
4-4 
4 4 

t\ 

4-4 
4-4 


interval 
between 
arrivals of 
direct and 
■ if reflex 
effects at p 
(measured). 


to be de- 
ducted 
for trans- 
mission 
in nerve =' 
(assumed). 


Probable 

delay in 

cord. 

i 

17-2 

17-2 1 

17-4 

17-4 

17-4 

17-4 

171 

17-1 


10,000 

12,000 
10,000 

3,000 
10,000 

3,000 
10,000 
10,000 


. d 

d 

d 
d 
d 

d 

t 


I 


14 

14 
14 
14 
14 
14 
14 
12 


40 

40 
40 
40 
40 
40 
40 
42 


19-8 

19-8 

20 

20 

20 

20 

19-5 

19-8 


2-(i 

2-6 
2-6 
2-6 
2-6 
2-6 
2-6 
2-7 



' The strength of the current in all the tables is given in units, read off from the 
graduated scale of the Kronecker indnctorium, 1 Daniell being in primary circuit, 60,000 
ohms in secondary. With such an arrangement the break inductioti current could not usually 
be felt on the tongue with strengths under 12,000. 

2 p stands for the proximal leading-oil' electrode. 

^ The length of nerve from tlie i)rincipal exciting electrode to the cord wius not in 
every case directly measured, althougli the length of the sciatic nerve itself in the thigh 
was always measured. I ascertained that the ratio of sciatic plexus length to nerve length 
in a certain number of pre])arations was as 1 to OTB, and have in the majority of crises 
calculated the length from this. A few millimetres of nerve more or le.ss would make so 



12 



Buchanan 



[The first record was taken with the right gastrocnemius to excitation of 
rio-ht sciatic. The rest were taken with the left gastrocnemius to excitation 
of left sciatic. No reflex response was obtained when the strength of the 
current was 1000 units.] 

In three other preparations the variation in the interval was slightly 
greater, but hardly exceeded Itr. There were, however, three experiments 
(in two of which four reflex responses, in the third nine, were recorded) in 
which the shortest and the longest interval on the diflerent occasions varied 
by as much as 4cr. What the conditions were in the longest of these three 
experiments w411 be seen from the following table. It will be referred to 
again later (p. 17). 



Exp. 37. Nov. 6, 1906. Room temp. 11° C. 



Induction current 
to nerve. 



Length, in millimetres, of 



Time, in thousandths of a second (o-), 



Strength. 



5,000 
10,000 
10,000 
r),000 
3,000 
10,000 
10,000 
10,000 
10,000 



muscle nerve nerve 

ni,.^n*ir.n ^rom cH- | fpom Cu from Cv 

uireciion. ^j.^^^,^ ^f electrode electrode 

nerve to 7;. to muscle, to cord. 



interval 

taken by between 

impulse to arrivals of 

reach p direct and 

directly of reflex 

(measured), effects at ^j 

1 (measured). 



to be de- i 

fi^'fr^t , Probable 

for trans- 
mission in I 
nerve I 
(assumed). ' 



d 


17 


d 


17 


d 


17 


d 


I7 


d 


17 


a 


17 


d 


17 


a 


17 



18-5 

17-8 

18-4 

18-9 

20-2 

22 

20 

21 

22 



2-1 
2-1 
2-1 
2-1 
2-1 
2-1 
1-9 
2-1 
1-9 



delay in 
cord. 



16-4 
15-7 
16-3 
16-8 
18-1 
19-9 
18-1 
19-1 
20-1 



The length of nerve traversed from the place of excitation, through the 
plexus, to the cord and back, was in most of my preparations about 60 mm. 
Only occasionally, when large animals were used, was it 70 or even 80 mm. 
Many different observers have measured the rate of propagation of an 
impulse along fresh frog's nerve and found it to be about 30 metres per 
second. It is also known that there is no delay in the dorsal ganglia.^ To 
ascertain the time which elapsed in the cord itself during each response we 
must therefore deduct as a rule about 2o-, but occasionally as much as 2-6cr, 

little difference to the time to be deducted for transmission in nerve as compared with the 
time taken in the central nervous system, that this method seemed to me to be exact 
enough for the pre.sent purpose. The relative length of nerve traversed in different 
responses of the same preparation has, however, been in so far taken into account that 
usually more has been deducted for transmission when the current was ascending than 
when it was descending (see also p. 14). The sense in which I have used the words 
" ascending " and " descending " will be obvious. It is not strictly correct in reference to 
the muscle when the sensory fibres are being considered, and not the motor only. 
1 Moore and Reynolds, J. Physiol., xxiii., 1898, Suppl. 



Traiisniission-tiiue of Reflexes in Spinal Cord of Frot{ 



18 



from the measured time interval between the arrival of the direct and that 
of the reflex eflect at the flrst recording spot of the muscle. 

Only when the current applied to the nerve to excite it was so strong, 
or the part of the nerve to which it was applied of such a nature, that a 
temporar}^ obstruction was produced under the kathode when the current 
was broken, must a longer time be deducted for transmission in nerve when 
the kathode was between the anode and the cord, and a shorter one (from 
the measured interval between the two arrival-times in the particular 
record) when the kathode was between the anode and the muscle. 
Although there is no difiiculty in detecting a delay-producing moment 
of this kind, and in measuring the time taken to overcome the obstruction 
at the kathode by the impulse going straight to the muscle, there is a 
slight uncertainty about the exact amount to be deducted for delay caused 
by such obstruction to the impulse starting in the opposite direction. The 
reason for this uncertainty may be best explained by introducing here the 
measurements of records taken alternately with (relatively) strong ascend- 
ing and descending induction currents (in a strychnine preparation) with 
a view to elucidating the matter. 

Exp. 48. Nov. 27, 1906. Room temp. 16° C. One minim 0-01 per cent, 
liq. strych. injected one hour before preparing nerve and muscle. 



Induction current 
to nerve. 


Length, in millimetres, of 


Time, in thousandths of a second (or), 


Strength. 


Direction. 


muscle 1 nerve 
from en- i from Cu 
trance of electrode 
nerve top. to muscle. 


nerve 
from Cu 
electrode 
to cord. 


taken by 

impulse to 

reach j) 

directly 

(measured). 


interval 
between 
arrivals of 
direct and 
of reflex 
effects at p 
(measured). 


to be de- 

(assumed). , 

i 


10,000 
14,000 
14,000 
14,000 
14,000 
14,(100 
14,000 
14,000 
10,000 


d 

d 
a 
d 
a 
d 
a 
d 
d 


11 16 
11 16 
U 12 
11 16 
11 12 
11 16 
11 ' 12 
11 16 

11 16 

1 


37 
37 
41 
37 
41 
37 
41 
37 
37 


5 
5 

3-7 

5 

3-7 

5 

3-7 

i) 

5 

! 


14 

13-7 
ir)-6 
13-7 
15-6 
131 

iri 

12-.-) 
12--) 


{ "-? •- 

{:?■? '^' i 
1 +^''* 11 J 

\-±r, "■' 



Allowing Olo- for the extra 4 mm. of nerve traversed when the current 



14 Buchanan 



was 



descending, it took the impulse V'la- to overcome the obstruction under 
the Zn electrode on its way to the milscle. Hence, although the reflexly- 
produced impulse actually arrived at the muscle ISTa- later, the difference 
between the two times of arrival would have been 14>-9ar, had it not been 
for the delay to the direct impulse, unless, as is not at all improbable, the 
obstruction was still there (in an attenuated form, i.e. causing a shorter 
delay) when the reflexly-produced impulse reached the same spot later, in 
which case it would have been somewhat shorter. That it did so persist is 
one alternative suggested by the preciseness with which the measured time 
taken for the reflex effect to reach the muscles recurred alternately with 
the ascending and descending current in the first four responses. Instead 
of reaching the muscle l-2(r later, when the current was ascending, as one 
might have expected, it reached it only Q-Qcr later than when it was 
descending (in IQSa- in the one case, IS'To- in the other). Allowing Olo- 
for the extra 4 mm. of nerve which w^ere traversed by it in the former case, 
and granting that the obstruction persisted, the whole time taken to reach 
the muscle was therefore not more than O'So- longer when the impulse met 
with the block at the start (current ascending) than when it met with it 
first on its return from the cord some 2 -So- later (current descending). 
Assuming, as one can hardly help doing, that the delay at the outset at one 
spot when the current is going in one direction is the same as it is at 
another spot 4 mm. away, when the current is going in the opposite direction, 
it looks either as though the reflex effect were delayed O'To- on its way to the 
muscle when the current was descending, and not delayed at all on its way to 
the muscle when the current was ascending, in which case the cord delay 
would be 14*9cr — 0-7(T— 2*5cr = ll'7cr; or as though the impulse produced 
by the ascending break induction shock was accelerated when it had to 
pass a second time, but in the reverse direction, the spot at which it had 
before been delayed, in which case the cord delay would be 14-9(t— 2-5o- = 
12-4(7, and the supposed accelerating factor would be such as to make the 
impulse traverse the spot O'lcr more quickly than it would otherwise have 
done. Not knowing which of these alternatives, or what other alternative, 
best expresses what actually occurred, I have in such experiments (showing 
kathodic obstruction) given a value to the cord delay intermediate between 
the two which it seems to me that it might have (see above). The error so 
introduced is hardly greater than another error which could not be avoided, 
that, namely, which arises from the assumption made that the rate of trans- 
mission along nerve where there is no block, is the same for all nerves and 
at all temperatures, i.e. all at which my experiments have been made 
(11° C.-18° C). But neither solving the doubt nor removing either error 
would alter the absolute value of the cord delay by more than a fraction 
of a, and would not alter the relative value in successive responses in any 
one preparation at all ; at any rate not in such of them as were taken with 
the induction current in one and the same direction. 



Transiiiission-time of Reflexes in Spinal Cord of Frog 15 

On deducting from the measured time the time which was, according to 
what has just been said, assumed to have been spent in traversing, in each 
response in each experiment, the kno\\Ti length of nerve, we find that the 
time spent in the cord itself in the same-limb reflex, the time taken, as it 
seems to me, for the impulse to pass from the branched ultimate endings of 
the several afferent fibres concerned, each across a synapse, to their respective 
motor cells, may be, though it only once was so in my experiments, as short 
as 12(7 in the normal frog's spinal cord, but that it is more frequently some- 
thing between 14cr and 21cr. If one maj^ use an analogy, which is possibly 
something more than an analogy, this is the time taken for the endings of 
the aflerent fibre or fibres, when sufficiently charged, to discharge themselves 
across the gap to the oppositely charged, or uncharged motor cell. 

With regard to the influence of strength of stimulus on such time of 
discharge, if such it be, it is not easy to make a definite statement, mainly 
on account of the difficulty in getting the reflex effect (without the use of 
drugs) sufficiently often in one preparation.^ In two experiments only, the 
details of which have already been given in tabular form, did the reflex effects, 
in response to different strengths of stimulus, give a definite result. Exp. 58 
shows that stimuli three or four times as strong as one which cannot be far 
from just producing the reffex effect, ma}^ produce it in almost exactly 
the same time. In Exp. 37 the reflex times were more varied, but did not 
consistentl}^ vary inverselj^ with the strength of the induction current. It 
is, however, possible and even probable that there is for each preparation 
some strength of stimulus, not quite weak enough to be wholly ineffectual, 
with which the reflex time is longer, although the fact that it is not until the 
experiment is over and the photographs developed that one knows exactly 
what has happened creates a difficulty in deciding this point experimentally. 

The strength of the reflex effect, as indicated by the steepness of the 
rise in the muscle record, is almost always less, and very considerably less, 
than that produced directly by the stimulation of the motor part of the 
nerve. Only with one undrugged preparation have I obtained records 
which indicate that the effect of the central stimulus was not much less 
than that of the artificial stimulus to the motor nerve. In this experiment 
all the four records of the response, when the intact nerve was stimu- 
lated, showed this. In the first response (to an artiHcial stimulus of 5000 
units) the reflex effect was strongest; its strength ma}^ be judged by 
comparing the steepness of the rise in the curves representing the reflex 
and the direct response in fig. 3. In the second record it was nearly but 
not quite as strong ; in the third and fourth it was decidedly weaker (but 
about the same for each); the artificial stimulus used to obtain the last 
three records was twice as strong as that used to obtain the first. 

^ This is a ditiiculty wliiili I liavi- now overcome by using wholly tloxor mviscles (see 
footnote, p. 30) to indicu'to cord delay, and by keeping the preparation at a low tempera- 
ture. The results fully confirm those obtained from the two exiieriments referred to in the 
text. [November IQOf.] 



Ig Buchanan 

It should here be noted that the strength of the direct effect was 
appreciably altered in not a few of my experiments by the direction of 
the induction current the break of which was used to excite the motor 
nerve, even when the strength of the current was supra-maximal. The 
steepness was in such cases more frequently less with the induction current 
descending; but if several responses were recorded with the current in 
either of the two directions alone, the direct effect was apt to be smaller 
in a record afterwards taken with the direction reversed. The experiment 
just referred to affords an instance of this : the first three responses were 
to the break of ascending currents, the fourth to that of a descending 
current. While in the second and third the direct effect was about as 
strong as in the first, in the fourth it was very little stronger than the 
reflex eftect on the same occasion. The direct effect obtained with the 
descending current did not necessarily, when less in amount, begin to manifest 
itself later, altliough sometimes, as in this instance, it did so. Such a 




Fig. 3. — First electrical response of the gastrocnemius 
of a normal preparation obtained when the intact 
sciatic nerve of the same side was excited by a supra- 
maximal stimulus. [Time lines 730 per second.] 

difference in the strength of the effect (when it occurs) seems to me to 
indicate some temporary impairment of the particular spot of nerve under 
the anode, preventing perhaps the participation of the whole number of 
fibres (in some cases the condition being brought about by this having 
previously been a kathode), for in some of the preliminary experiments in 
which the nerve was excited at two different parts of its length in turn, 
the difference of the effect according to the difference of direction of 
current might be marked when the one part and not when the other was 
used. I have, however, made no serious attempt to understand the pheno- 
menon, because it seemed to me to have little, if any, bearing on the matter 
which is now concerning us. Moreover, the direct effect may get weaker 
in the course of an experiment even without reversing the direction of the 
current (see fig. 12 (E), p. 56). In none of my experiments (with the 
possible exception of No. 49 ; see note to it on p. 42 ) was the strength of 
the reflex effect altered by altering the direction of the induction current 
applied to (the afferent part of) the nerve either in the normal or in the 
strychnised cord. 

The probable cord delay in the one normal preparation from which 



Transmission-time of Reflexes in Spinal Cord of Frog 17 

comparatively strong reflex effects were obtained was not shorter than in 
other normal preparations, nor in this preparation was it shorter the 
stronger the effect. It was respectively, taking the four responses in order, 
after allowing for transmission time in nerve (including that taken to 
overcome the block under the kathode, which was in this preparation lo-) : 
14-3cr, 13-9cr, 16-2o-, 13o-. The strength of the central stimulus, as estimated 
by the strength of the effect it produces, is not therefore a function of the 
time taken by the impulse to affect the motor cell. 

There is the same difficulty in making experiments with regard to the 
influence of temperature on the time occupied in the normal cord, as with 
strength of stimulus. On every occasion when, after an experiment has 
begun, the gastrocnemius being used as indicator, I have either cooled or 
warmed the back of a normal preparation, the reflex effect, if present 
before, was no longer to be seen in records taken after such treatment. 
In a few of the experiments I tried to make use of Biedermann's 
experience that a reflex contraction could be obtained with greater 
certainty in a cooled spinal frog if, after making the preparation, it was 
given a long rest with a bag of ice applied to the spine. In eight 
preparations, among which was the one which gave the lowest time value 
for the cord delay, there had been for about half an hour, on the back of 
the frog, a bag (an india-rubber finger-stall) containing ice. This was not 
the case in the other experiments with normal cords, and it was in three of 
these that the measured time interval between the two arrivals was longer 
than 21cr, i.e. that the probable cord delay was over 19o-. The reflex time 
in the experiments in which ice was used was not therefore universally 
longer than in the experiments in which it was not. I do not of course 
think that this proves that cold has no influence on the delay in the cord, 
but I have npt yet been able to make an experiment that would either 
prove it or disprove it satisfactorily.^ In the strychnised cord, which is 
capable of giving a reflex effect a large number of times in succession, the 
delay is most certainly increased by cold; but whether we should be 
justified in applying, without reserve, to a normal preparation what we 
know to be true of a drugged one is not, in my opinion, to be answered 
straightway in the affirmative. 

The results obtained from the records taken in Exp. 37 (see p. 12) 
make it appear that fatigue may have some influence on the time of the 
cord delay. While in the first four responses this did not exceed (when 
reduced) l7o-, in all of the last five it was over 18o-, and became finally as 
long as 20(7. None of the other preparations gave evidence of this, but 
with the one exception of Exp. 58 they all gave (or were allowed to 
give) too few records of reflex effects to have been able to show it. This is 

1 I have now [Nov. 1907] made two experiments, using the biceps fenioris as indicator, 
which show that the probable cord delay may become some lOo- longer in a normal cord by 
a reduction of the temperature of the whole preparation from 12° C. to 6° C, and may agjun 
become shorter when the temperature is raised once more to 12° C, though not regaining 
its original value. 

VOL. I. — JAN. 1908. 2 



18 Buchanan 

one of the many matters witli regard to which further experiments are 
needed.^ 

III. The Same-Limb Reflex-Time in a Cord the Excitability of 

WHICH HAS BEEN RAISED BY STRYCHNINE. 

Most of tlie frogs used for these experiments were treated in the same 
way as those used for experiments without drugs : that is to say, they were, 
after being decerebrated, kept in the cold for one or more days before 
making an experiment. A preparation then made in the same way as 
has already been described for " normal " animals, but the injection into 
the dorsal lymph sac of a very minute dose (0-005 — 002 mgr.) of strych- 
nine some hours, or even days, before, or that of a somewhat stronger 
dose (0-03 — 0-06 mgr.) immediately before,- never failed to produce the 
reflex electrical response in the gastrocnemius when the mixed nerve was 
stimulated by a single break induction shock. This response, instead of 
being usually weaker than that to direct excitation of the efferent part of 
the nerve, was, as a rule, of nearly equal strength with it, and sometimes 
even stronger (rise on curve steeper). It was, however, not always its 
strongest at the beginning. It sometimes lasted no longer than the direct 
response, but very soon after the administration it began to lengthen, be- 
coming from two to four times as long as the direct response. It was not 
until the influence of the drug had become so great that it showed itself 
in the movements of, or in the attitude assumed by, the brainless animals, 
that the reflex electrical response began to assume the serial character so 
often described, and capillary electrometer records of which I have 
published elsewhere.^ In none of the photographic records reproduced in 
the present paper does more than the first period of a response to a stimulus 
of central origin appear, even when more were present. The undulations 
in the contour which may be seen in most of the records reproduced in 
figs. 4 to 12 are such as I have shown elsewhere* to be of purely muscular 
origin, although the photographic records which really prove it have not 
yet been published. When the effect of the strychnine begins to wear off 
the number of periods of central (proximate and, I think, also ultimate ; 
see p. 29, footnote) origin (recurring five to ten times a second) is reduced, 
until finally there is again only one such period. 

Most of the experiments made for the purpose which is now concerning 
us were made on preparations in the early stages, when the action of 
the drug was incipient, or in the late stages, when its action was vestigial. 

' They have now been made [Nov. 1907]. 

2 I cannot state the dose in milligrammes per body weight, because I did not weigh 
each frog. It seemed to me that little would be gained by doing so, seeing that the 
effectiveness of the drug varies so much, and in a way that has not yet been sufficiently 
studied, with the temperature of the frog and other conditions. 

2 Buchanan, Journ. Physiol., xxvii., 1901, Plates VIII. and IX. 

■* Burdon-Sanderson and IBuchanan, Journ. Phvsiol., xxviii., 1902; Proc. Physiol. 
Sec, p. xxix. Garten (Abh. k. Sachsischen Ges. Wiss.,' xxvi., p. 333, 1901) has also given 
experimental evidence of the fact. 



Transmission-time of Reflexes in Spinal Cord of Frog 1!> 

In most, therefore, the response with the nerve intact, whether electrical 
or mechanical, was not much longer than when, after dividing it, its 
peripheral end was stimulated. A few, however, were made when the pre- 
paration was in the attitude characteristic of strychnine poisoning, or 
only just beginning to recover from it. 

The conditions obtaining, and the results yielded in the several responses 
in nineteen tj'pical experiments, have been tabulated in the same way as 
were those for normal cords. Certain of the conditions, and the results 
obtained in thirteen of them, will be found on the left-hand side (first six 
columns) of the tables beginning on p. 35. In almost every case in which 
the animal was only lightly drugged, the mechanical, as well as the 
electrical response was recorded, in order that its strength and duration 
might be compared with those of the twitch (einfache Zuckung) which was 
recorded at the end of the experiment. 

It was only in three experiments that the cord delay was through- 
out of longer duration than appears to be characteristic of good normal 
cords under similar conditions. The majority of the experiments do not 
therefore confirm the conclusion which Wundt came to from his obser- 
vations on the mechanical reflex response of the strychnised animal, 
that, namely, the efi'ect of strychnine on the cord is to lengthen the delay 
The three preparations in which the cord delay, after deduction for 
transmission time in nerve in the same way as before, was found to be 
throughout 24cr or more, were the only ones in which not only were the 
arms flexed, but the legs were rigidly extended when the preparation was 
made. The reflex electrical responses in all three were serial. Verworn^ 
has shown that in acute stages of strychnine poisoning not only the central 
nervous system but the heart also is affected by the drug. I have there- 
fore examined the heart at the end of these and other experiments. In 
one of the three [Exp. 50], made on a specimen which had been in the 
attitude characteristic of strychnine poisoning for about two hours, and 
in which the buccal respiratory movements had already ceased, the heart 
was hardly beating at all, the blood was very blue, and tlie preparation 
seemed to be nearly asphyxiated ; both the direct and the reflex electrical 
eflfects were very feeble (rise of curve ver}^ gradual) ; the probable cord delay 
was between 25cr and 30o- in four of the responses ; it was as much as o9a 
in the first one, and 41o- in the second. Of the other two preparations, the 
one had been in the attitude characteristic of strychnine poisoning for half 
an hour only, while the other had been in it for several hours. In both, 
the buccal respiratory movements were laboured. The heart was beating 
feebly. The reflex effect was not as strong as the direct effect : and in 
most, but not all, of tlie responses, it did not become maximal until some 
(J-10(T after its commencement ; sometimes, in the second of the two experi- 
ments (in which the stimuli to the nerve were kept close to the threshold 
value in strength) [Exp. 55], not imtil 20o- after. Fig. 10, A (see p. 52) 
' Verworn, Archiv f. (Anat. u.) Physiol., 1900, p. 385. 



20 Buchanan 

represents the third of the live responses to excitation of the intact nerve 
of the same side, recorded with the first preparation [Exp. 52]. It was the 
only one of the five in which the reflex effect attained its maximum early. 
Fig. 11, A and C (see p. 54) represent the first and eighth, respectively, of 
the sixteen responses to excitation of the nerve of the same side, recorded 
with the second preparation, and both show, as indeed did all but two 
out of the sixteen, the late development of the maximal reflex effect. It 
will be seen (pp. 43 and 45) that in the one of these two preparations 
the probable cord delay varied between about 26o- and 30o- ; in the other 
between 24cr and 33o-. 

In the only two other experiments made upon animals with their arms 
flexed, recovery had so far progressed that the legs were no longer stiff 
and extended, buccal respiration appeared to be normal, and the heart was 
beating very fairly. In neither of these were the reflex electrical responses 
serial, and the cord delay was not longer than is frequently the case in 
normal frogs. Fig. 7, A (see p. 50) represents the first of six responses 
from one of these preparations [Exp. 40]. In this and in the second 
response, but not in the others, the direct effect was double ; the reflex 
eftect in this response and in some of the others was less strong ; it did not, 
in this one, reach its maximum at once, although it did so in those in which 
it was stronger. Fig. 12, A, C, and E (p. 56), represent the first, second, 
and seventh of the eight responses recorded (by one of the two gastro- 
cnemius muscles used) with the other preparation in which the arms were 
flexed [Exp. 56 R], when the sciatic nerve of the same side was excited. 
The reflex effects, as will be seen, were in the first two very nearly if not 
quite as strong as the direct eftects ; they attained their maximal strength 
at once and in all the responses. In four of the later responses they 
were even stronger than the direct effects owing to the latter having 
become weaker. One of these is shown in tig. 12, E. 

In experiments made upon animals in which the drug was either just 
beginning to take effect or had nearly lost its effect, the heart was, as a 
rule, beating quite normally. Measurements of the records taken in these 
remaining experiments have led me to the conclusion that if strychnine 
has any direct influence at all on the length of the delay in the cord 
in the case of the same-limb reflex, it diminishes rather than increases it. 
Thus, whereas without strychnine only one preparation (out of eighteen) 
gave a value so low as 14o- for the fime taken by the impulse to travel 
from the particular spot on the nerve stimulated, to, and through, the cord 
and back again, as many as (six out of nineteen) strychnine preparations 
[Nos. 4, 8, 14, 16, 42, 48] gave as low a value as this, and four of these 
each gave it more than once or gave a still lower value (shorter by from 
1(7 to 2cr) in one or more of the responses. 

In three experiments (one of which only [Exp. 14] is included in the six 
just referred to) records were taken first of the electrical responses of the 
one gastrocnemius when its nerve was stimulated, before the administration 



ransmission 



■time of Reflexes in Spinal Cord of Frog 



21 



of strychnine ; then of those of the second gastrocnemius when the second 
sciatic nerve was stimulated after the administration of strychnine in 
sufficient quantity, and for a long enough time to make the contraction in 
response to a single break induction shock to the mixed nerve something 
stronger than a twitch. Exp. 14, which has already been referred to (p. 8), 
and of which one of the records obtained when the cord was normal has 
been reproduced in fig. 2, A, is perhaps the most valuable of these, for the 
reason that the cord was only excited four times, and gave an excellent 
response each time, before the administration of the drug, and that as 
many as nine equally good reflex responses were recorded with the other 
muscle after its administration. In seven of these it occurred in almost 
exactly the same time after the direct response, and in all seven the interval 
was from la to l|o- less than in any of the responses obtained from the 
same cord without strychnine. In the other two responses it was shorter 
still, but in these two (both of which, as it happened, were to stronger 
stimuli after weaker ones) the eflect, when it began, instead of being strong, 
was very weak, and it only became maximal, in the one case (fourth 
response) 2ar, and in the other (seventh response) 4(t, later. 

The following table gives the data for both parts of the experiment : — 



Exp. 14. Oct. 1, 1906. Room temp. 18° C. Records taken first with 
the right gastrocnemius to excitation of the right sciatic. Cord normal. 



Induction current 
to nerve. 



Strength. 



Length, in millimetres, of 



Time, in thousandths of a second (<r). 



muscle 
from en- 
trance of 
nerve to i). 



nerve 
from Cw 
electrode 
to muscle. 



taken by 



interval 
between 



to be de- 
ducted 



electrode 
to cord. 



reach p direct and 
directly of reflex 

^(measured). effecUa^t^p (,3,^,,). 



in nerve 



delay in 
cord. 



2-6 ' 19-5 2 17-5 

2-6 15-6 2 13-6 

2-6 18 1-8 16-2 

2-6 19 2 ' 17 



5,000 d 7 14 

10,000 \ d 7 14 

10,000 \ a 7 17 

10,000 \ d 7 14 30 

Records then taken with the left gastrocnemius to excitation of the left 
sciatic, prepared a quarter of an • hour after the injection of ^ minim 
0-1 per cent. liq. strych. into the dorsal lymph sac. 



10,000 


d 


8 


9 


35 


2-6 


14-2 


2-3 


11-9 


5,000 


d 


8 


9 


35 


2-1 


14-4 


2-3 


121 


5,000 


d 


8 


9 


35 


2-4 


141 


2-3 


11-8 


10,000 


d 


8 


9 


35 


2-2 


13-4 


2-3 


111 


5,000 


d 


8 


9 


35 


2-2 


14-1 


2-3 


11-8 


5,000 


d 


8 


9 


35 


21 


14-5 


2-3 


12-2 


10,000 


d 


8 


9 


35 


2 


11-4 


2-3 


91 


5,000 


d 


8 


9 


35 


2-2 


141 


2-3 


11-8 


10,000 


a 


8 


6 


38 


2 


14 


2-4 


IIG 



22 Buchanan 

It will be noticed that the probable pord delay was not only longer, but 
more variable before the administration of the drug. The record of the 
second response of the second muscle is given in fig. 4, for comparison with 
the fourth of the first muscle reproduced in fig. 2, A. The comparison 
brings out very forcibly the first effect of strychnine on the electrical reflex 
response of muscle, namely — the response (to a weaker ultimate stimulus) 
has become as strong (curve as steep) as that to the direct excitation of the 
motor nerve ; it has also become somewhat longer than it was when 
produced by the same cord when normal. I believe it to be also character- 
istic that it occurs somewhat earlier. It will be seen that the direct effect 
was again double. 

With regard to the two other experiments of this kind : they were both 
made with preparations, in the normal cord of which the delay was long. 
In the one [Exp. 18] the records taken with the second muscle showed that 




Fig. 4. — Second electrical response of the other 
gastrocnemius of the same preparation as was 
used for taking the records to which tig. 2 
refers, obtained when the intact sciatic nerve 
belonging to it was excited (Exp. 14 L). After 
the first muscle had been used, ^ minim 0-1 per 

cent, liquor strychnise had been injected into the i 

dorsal lymph sac. [Time lines 795 per second.] 

the probable cord delay, which did not vary much, and was only just under 
24o- in the five reflex responses which were obtained with it before 
administering the drug, was less constant in those taken with the 
second muscle after so doing. It was in two responses longer than 
before ; in the next, of about the same duration ; and in the last, shorter 
than before. In none of the responses was the reflex effect any stronger, or 
longer, than before. In the other experiment [No. 45], after five responses 
of the one muscle to excitation of its nerve, three of which only showed a 
reflex effect, had been recorded with the cord normal, two records were 
taken of the responses of the same muscle excited in the same way a 
quarter of an hour after the administration of strychnine, and before 
preparing the second muscle and its nerve. These both gave a reflex 
response no stronger than before, and still much weaker than the direct 
effect, but in a very much shorter time (probable cord delay in both of 
them 12o- instead of the 23o- which it had been in all three responses when 



Transmission-time of" Reflexes in Spinal Cord of Frog 28 

the cord was normal). The records then taken with the muscle of the 
opposite side when its nerve was stimulated (the first and the fourth of 
which are reproduced in fig. 8, A and C, p. 51) show that the central 
stimulus was then, a quarter of an hour later, producing as big an effect 
in the muscle as the peripheral one, and also that the interval between the 
arrivals of the two effects was shorter in all the responses, except the very 
last, than it had been in the records taken before the administration of the 
drug, the probable cord delay being 20cr or 19cr instead of the 28o-. 

Three other experiments [Nos. 16, 42, and 48 (the part of it referred to 
on p. 13)], all made very soon after injecting the strychnine and when its 
action was incipient, further lend some support to the view that strychnine, 
so long as it is affecting the spinal cord alone, tends to somewhat shorten 
the delay in it in the same-limb reflex, in so far as they all show that the 
reflex time gets shorter as the time the drug has had to take effect becomes 
longer. 

In four experiments [Nos. 8 (both sides in turn), 88, 46, and 54], all 
made very soon after the administration of an extremely weak dose, one 
would not have known from the records that any strychnine had been 
administered at all : that is to say, the reflex effect was no stronger nor 
longer than usually obtained from a normal cord. The probable cord 
delay varied about 14cr in the first two, about 16o- in the third, and about 
21a- in the fourth. There was one record obtained with the second muscle 
in Exp. 8, in which the reflex response was slightly but decidedly stronger 
than in the rest, and in this one it occurred 2o- sooner than in any of the 
others. Moreover, the delay was lo- longer in the two responses which 
showed a reflex effect with the first muscle in Exp. 8, than it was in an}' of 
the four taken with the second. 

Taking all these facts into consideration, I think the conclusion is 
warranted that strychnine, when it is producing no other effect than one 
on the excitability of the cord, does not greatly alter the length of time 
which elapses there in such a simple reflex as the one with which we have 
hitherto been dealing, but that it does tend to reduce this time slightly, i.e. 
by one- or two-thousandths of a second. Only when the heart also has 
been affected by the drug is the time lengthened, owing to insufficiency in 
the supply of oxygen to the cord or to some other consequence of defective 
circulation, the impulses arising along the afferent nerve fibres taking then 
a much longer time to be transmitted to their respective motor cells or else 
the motor cells taking a longer time to react to their several impulses, some 
perhaps taking longer than others. 

It was particularly when using weak stimuli to excite the dorsal roots, 
stinuili too weak to have produced any ett'ect at all had the cord been 
normal, that Wundt found the reflex mechanical latency in strychnine 
preparations so very much prolonged, it becoming sometimes as much as 
six or even ten times as long as in normal preparations. The variation of 
latency with strength of stimulus appears indeed from his curves to become 



24 Buchanan 

very marked under the influence of ^ strychnine. It is therefore at first 
sight somewhat surprising to find that the measurements of my records, 
examples of which are given in the protocols on pp. 35 to 48, give no more 
evidence of cord delay varying with strength of stimulus than do 
those obtained with normal cords. Thus, while in six experiments [16, 40, 
first part of 42 (p. 28), 46 (which, however, was one of those in which the 
records give so little evidence of the presence of strychnine), 52, and 53] 
made with the streng-th of the induction current kept constant throughout, 
the cord delays varied in successive responses by as much as from 2(t to 4(7, 
they varied hardly at all in five other experiments [14 L, end part of 42, 
47, 49 R, and 56 R] made with the strength of the stimulus varying a good 
deal, even though, as in the three last of these, the strengths used to obtain 
some of the responses cannot have been much greater than just sufficed to 
produce a reflex effect at all. It is true that in 56 L the one response 
recorded with the weakest stimulus had the longest cord delay (l-2o- 
longer than the rest), and it is also true that in Exp. 14 L (see p. 21) a strong 
stimulus did on one occasion evoke a response in a time that was 2(t shorter 
than it was on any other occasion ; but these are isolated records, and in 
Exp. 14 L a stimulus equally strong was applied on three other occasions 
without the delay being shorter than it was with weaker stimuli. 

My records, therefore, furnish no evidence of an inverse relationship 
between cord delay and strength of stimulus, even in preparations in which 
the effect of strychnine on the cord is well marked, so long as the 
strength has a certain value which is not far above the threshold 
value. The strength may be doubled, trebled, or even increased to four 
or five times its value without altering the cord delay (see footnote 
to p. 54). But although there can be no doubt as to what is the case with 
stimuli above a certain strength in relative value, we find ourselves on far 
less certain ground when we try to discover experimentally what happens 
when the strength of the stimulus is almost at its threshold 
value. For here we are met with difficulties as great as those met 
with in making the same investigation in the normal cord, although they 
are of another kind. While strychnine cords have undoubtedly so far the 
advantage over normal ones, for the investigation of the influence, if any, 
exerted by this and other physical conditions, that any one of them will 
give as a rule a comparatively large number of responses, they have this 
disadvantage for the investigation of the particular question as to the 
influence of strength of stimulus, that the threshold value, i.e. the strength 
of the stimulus just sufficient to produce an effect, is apt to change during 
the experiment. This was the case most strikingly in Exp. 55, in which 
stimuli at first too weak to produce any effect at all became capable of 
doing so in the course of the experiment. The cord delay, as the table 
shows, varied throughout a good deal in the diflferent responses, but did 
not do so in any definite regular relation to the strength of stimulus 
evoking them, near as this must have been at times to the threshold value. 



Transmission-time of Reflexes in Spinal Cord of Frog 25 

In spite of this fact the records do, however, it seems to me, afford evidence 
that there is a region near the threshold where strength of stimulus and 
synapse delay vary inversely, and that they at the same time suggest 
an explanation for any discrepancy there may appear to be between such 
observations as those of Wundt (which have been confirmed in mammals 
b\^ Sherrington, H. Franck, and others) and mine. I have already 
mentioned (p. 19) that in so many of the responses in this preparation (as 
well as in certain others) the reflex electrical eflect does not become maximal 
until some time after its first appearance ; and if we grant, as we shall 
presently find reason for doing, that only a single set of synapses is con- 
cerned in the reflex we are here considering, this fact, to my mind, can only 
mean that there was great want of co-ordinate action amongst the motor 
cells, due to variations obtaining either in the times taken to pass individual 
synapses, or in the promptitude with which individual motor cells were 
reacting. Whichever of these two factors determining what I call synapse 
delay it may be, the few motor cells which first come into play would 
control only a certain number of the fibres under the recording spots, not a 
sufiicient number for the whole central stimulus to produce an electrical 
eflect there even equal in amount to the small one which may generally be 
obtained from a normal cord (to a stronger peripheral stimulus), and not a 
suflicient number, I think we may add, to overcome the resistance of the 
rest of the muscle, and cause the whole to contract. Only when the other 
synapses too have been crossed and the other motor cells have reacted, are 
all the fibres excited together (the prolongation of the discharge so charac- 
teristic of the action of strychnine having brought it about that the cells 
belonging to the few synapses which were passed first were still discharging 
at the time the others began to be aflected later), and then only does the 
electrical response attain its maximum ; and only when a sufficient number 
of muscle fibres are brought into play to overcome the resistance of the rest 
of the muscle would the mechanical response begin. 

Since the strength of the stimulus was the same for all, this great 
variability in the delay at the individual synapses of the same set can only 
have been due, it seems to me, to great variability in the thresholds, i.e. in 
the resistances, of the separate synapses. For a few of them the stimulus 
possesses that strength which by analogy with what is known for muscle 
and nerve we may call " maximal " ; for the greater number it possesses 
submaximal values. We know that in this preparation the thresholds for 
those synapses or cells which were most readily aflected were changing 
after the experiment had begun. It is no great assumption to make that 
others at the end of the experiment had the same threshold which those 
had had at the beginning, and that these others at the beginning had 
thresholds lower than those had then. In seeking for the cause of this 
variability we are struck by the fact that the phenomenon in our records 
which indicates it (the long delay after the response has begun before it 



26 Buchanan 

produces its maximal electrical effect) is only at all commonly present in the 
responses of those preparations in which not only the cord but the circula- 
tion was affected by the drug, and in which (perhaps, to some extent, 
conse(|uently ^) all the muscles were in spasm, and in which cord delay (to 
all strengths of stimulus) was longer than in normal, and still more so than 
in lightly strychnised, preparations. What strychnine acting on the cord 
alone effects, appears to be counteracted in certain respects by what strych- 
nine acting on the circulation, and hence indirectly on the cord, effects. We 
have seen that this is the case as far as cord delay is concerned in the 
reflex we have so far been considering. It will be still more striking in 
the reflex we shall next have to consider, and for this second more complex 
reflex we shall see this conflict between the direct and indirect effect of 
strychnine on the cord showing itself perhaps in another way also. The 
cord of the decerebrate frog used in Exp. 55 had been for some hours under 
the predominating influence of the indirect action of strychnine. By the 
time the preparation was made, that effect showed signs of wearing off, and 
it would appear that the more specific action of the drug upon the cord — 
the lowering of the threshold at any rate — was beginning to re-assert itself, 
but it had only as yet done so for some of the synapses concerned in the 
response of this particular muscle, whatever it may have done for others, or 
possibly the same, synapses concerned in the responses of other (e.g. anta- 
gonistic) muscles.- The fact that the temperature of the room was rising 
while this experiment was being made probably contributed to bringing 
about variability of threshold, since there can be no doubt that strychnine 
is much more effective at higher temperatures than at lower. 

A record such as the one reproduced in fig. 11, C, seems tome, therefore, 
to show that synapses with higher resistances, i.e. those which required a 
stronger afferent stimulus to be forced at all, took longer (up to as much as 
20o- longer) to be passed by impulses hardly above their threshold values in 
strength than did those with lower resistances for which an afferent 
stimulus of the same strength would have a higher value ; and that thus 
variations produced in the thresholds of the structures excited have given 
us information that we could not succeed in obtaining by altering the 
strength of the exciting stimulus. The question, however, as to whether we 
may apply this information to a more normal, even though strychnised, 
animal, remains. It may be that it is only in an animal with circulation 

1 I would here refer to Bethe's observations, in which •' Erstickungskrampfe " (p. 248), 
" welche sich von Strychnmtetanus nicht unterscheiden lassen " (p. 247), are obtained in frogs 
under conditions which he regards as signifying increased consumption, and therefore 
requirement, of oxygen (Eosenthal's Festschrift, 1906). 

2 It may well be that this want of co-ordinate action of the synapses concerned in the 
contraction of one muscle, which seems so often to prevail when strychnine spasms are at 
their height (I have other examples of the phenomenon indicating it in records taken 
years ago), is correlated with the opening of paths to other muscles which are normally 
closed by the high resistance of the synapses or the ]jarts of the synapses (if we regard as a 
whole synapse, the boundary, or gap, between all the afferent fibre-terminations, and the 
one motor cell they adjoin) concerned in their contraction, and normally obtaining when 
the resistance in the first synapses, or parts of synapses, was low. 



Transmission-time of Reflexes in Spinal Cord of Frog 27 

impaired that synapse time is eventuall}' prolonged V)y insufficient strength 
of stimulus, or it may be that the impairment has only emphasised some- 
thing which does actually hold true for a normal, or more normal, animal, 
but on a smaller scale, the difficulty in discovering it being in such greater 
restriction of the threshold region. 

I would suggest therefore that in all responses in an}^ of my prepara- 
tions the records of which showed that the maximal effect was not developed 
until late, i.e. that the total central stimulus was not exerting its full force 
at the beginning, there was want of accord in the time taken to pass the 
several synapses, the greatest effect not being produced until all that can 
be passed have been passed, and all the muscle fibres belonging to them 
affected; and that this want of accord is brought about not by the 
direct action of strychnine on the cord, but by its indirect action. 

The fact that, as a rule, when strychnine is affecting the cord only, the 
effect at the recording spots after the administration of the drug is as 
great, and becomes so in as short a time, as in the case of the direct re- 
sponse, and that it is sometimes even greater than it (some of the responses 
in Exps. 45, 47, 49 L, and 56 R), together with the fact that no increase 
in the strength of the stimulus to the normal cord will make the reflex 
effect larger than it otherwise is, seems to me, on the other hand, to indicate 
that the smallness of the reflex response which is usual in a normal pre- 
paration is not due to a smaller number of fibres being excited than by 
the direct stimulus, but rather to the effect produced in each being smaller. 
In this connection it should be mentioned that the reflex effect, as it is seen 
in records taken with the normal cord, takes very nearl3^ if not exactly, 
the same time to become maximal as the direct effect. 

The influence of altering the temperature of the back of the pre- 
paration, and consequently of the cord, is well marked after the adminis- 
tration of strychnine. This is shown by the experiment of which the 
data are given on the following page. 



[T.\HLE. 



28 



Buchana 



Exp. 42. Nov. 15, 1906. Room temp. 15° C. One minim 0-01 per cent. liq. 
strych. injected one hour before making the preparation. 



Induction current 
to nerve. 



Strength. 



10,000 
10,000 
5,000 
10,000 
10,000 
10,000 



Length, in millimetres, of 



muscle 
from en- 
trance of 
nerve to^. 



nerve 
from Cu 
electrode 
to muscle. 



nerve 
from Cm 
electrode 
to cord. 



Time, in thousandths of a second (a), 



taken by 

impulse to 

reach p 

directly 

(measured). 



4-2 

4-2 

4 

4-2 

4-2 



interval 
between 
arrivals of 
direct and 
of reflex 
effects at p 
(measured). 



13-2 
14-4 

no reflex 
14-4 
13-2 
12-1 



to be de- 
ducted 
for trans- 
mission 
in nerve 
(assumed). 



2-3 
2-3 

effect 
2-3 
2-3 
2-3 



Probable 

delay in 

cord. 



10-9 
12-1 

12-1 

10-9 

9-8 



Bag of ice then put on the back of preparation for five minutes. 



10,000 I d 10 14 I 35 j 4-2 23 2-3 217 

10,000 d 10 14 35 4-2 24 2-3 22-7 

10,000 d 10 14 35 I 4-2 20-7 2-3 18-4 

5,000 d 10 14 I 35 I 4-2 22-2 2-3 19-9 

10,000 d 10 i 14 35 , 4-2 21 2-3 18-7 

Ice-bag removed. A second minute dose of strychnine given at tlie same 
time (in the hope of securing responses to the excitation of the other 
sciatic also), which fact may (though I do not think it does) detract 
from the value of the results obtained from the records taken five 
minutes later. 



10,000 


d 1 


10,000 


d 


5,000 


d 


4,000 


d 


5,000 


d 


10,000 


, 



35 


4 


13-1 


2-3 


35 


4 


14-3 


2-3 


35 


4 


14-3 


2-3 


35 


4 


14-4 


2-3 


35 


4 


13-4 


2-3 


35 


4 


12 


2-3 



10-8 
12 
12 
12-1 
11-1 
9-7 



It is also shown in Exp. 48, of which the data, when the whole prepara- 
tion was at room temperature, have already appeared in tabular form on 
p. 13, and of which most of the data, after the back had been cooled, 
appear on the left-hand side in the table p. 41. 

A number of experiments I made some years ago for another purpose, 
and in which the reflex eflfect was produced by excitation of the skin, not 
of nerve, show the same sort of thing. To these I shall have to refer in 
another connection on another occasion. What is very striking in the two 
recent experiments is that in the one case [Exp. 42] the cord delay is 
almost exactly doubled by cold, in the other [Exp. 48] almost exactly trebled. 
But whether any significance should be attached to these facts the evidence 
is not as yet great enough to decide. The further question suggests itself 
as to whether the influence of cold is a direct one on the cells (or the 



Transmission-time of Reflexes in Spinal Cord of Frog 29 

synapses) of the cord itself, or whether it is some indirect one, as, for 
instance, on the oxygen supply to the cells. 

The fact that strychnine acts on cord delay in the opposite direction 
to that in which fatigue would act, creates a difficulty in studying the 
influence of this factor in strychnine preparations. Although Exps. 40 and 
45, and the comparison of the two opposite same-limb reflex times in 
Exp. 56, show that the times got longer as the experiment went on, both 
Exps. 45 and 55 show that it was not shortened by rest ; so that I do not 
think we should be justified in straightway attributing the increase of 
length in the three experiments just mentioned to fatigue in the sense in 
which this word is commonly employed. We shall see that this seems to 
be otherwise in the case of the more complex I'eflex which we are going 
to consider immediately. 

A way of investigating the eflect of fatigue on cord delay, or rather, 
as I should prefer to call it, on synapse delay, which seems to me likely 
to be fruitful, is that of recording long serial responses obtained from well- 
strychnised preparations, and measuring the time intervals between suc- 
cessive periods. For whatever is the ultimate source of the impulses which 
give rise to the second and following periods,^ they must each in turn cross 
the synapse forced in the first instance by the impulse transmitted through 
the afferent nerve. It is well known that the periods may get longer towards 
the end of such a response, if this is a long one ; but what determines the 
duration of the response requisite for them to do so has not yet been 
suflSciently studied. 

IV. The Same-Limb Reflex Time ix a Cord the Excitability of 

WHICH HAS BEEN RAISED BY PhEXOL. 

I have only as yet made two experiments with this drug.- As, however, 
both show that the cord delay in the reflex we are studying was of the 
same order as in the normal cord or in the strychnine cord, I think they 
are worth referring to. Ten minims of a O'l per cent, solution, containing 
therefore about 0-6 mgr., were injected subcutaneously two or three hours 
before making the experiment. In the one preparation a reflex effect was 

' I still hold that this is to be sought in the central organ itself, and not in the peri- 
pheral organs, as Baglioni (loc. cit., and Z. f. allg. Physiol., ii. p. 556, 1903) believes. 
Besides the evidence against Baglioni's view brought forward by Sir J. Burdon-Sanderson 
and myself (loc. cit. and Physiol. Centrall)l., 1902), there is the further objection that the 
time between two successive periods is often not long enough (it may be as short as O'Oo 
second) for an impulse not only to have got to the muscle and back to the cord, then back 
again to the muscle, but for it to have made the muscle contract to such extent the first 
time as to excite the sensory organs in it (or its tendon) and start the mipulse back to the 
cord. Moreover, by Baglioni's method of treatment of the cord it is difficult to believe 
that impairment of circulation would not have made cord delays longer than usual. 1 
reserve, however, the discussion of the question for a later paper. As the text shows, it is 
to something in the cord, other than the motor cell, that I would attribute the ]ieriodicity. 

- Since this jiaper was written I have made several, but as they all confirm what is 
said here (and on p]>. 31 and 61), and as this paper is already too long, I must leave what 
they have further to tell for a future communication. 



30 Buchanan 

seen from the records to have arrived at the tirst recording spot of the 
muscle 20-2o-, 20-5cr, 20-7o-, 20-lo-, IDl^, and 24-4o- after the direct effect, in the 
successive responses to stimuli applied to the sciatic nerve of the same side, 
of respective strengths 10,000, 5000, 3000, 5000, 3000, 10,000. A record of 
one of these (the fourth) is reproduced in fig. 13, A (see p. 62). Above 
(fig. 13, B) may be seen the record of the response of the same muscle 
obtained later, when the peripheral end of the sciatic nerve was excited 
after severing it from the cord. All the records taken with the nerve 
intact very closely resembled one another. 

In the other preparation the reflex responses occurred 15cr, 15"6cr, 15o-, and 
25cr after the direct response, the strengths of the stimuli being respectively 
10,000, 3000, 10,000, 10,000 units. The direct and reflex responses very 
closely resembled one another, and were about equal in strength ; both were 
weak in the last response, strong in the others. 

V. The Crossed-Reflex Time. 

I have not yet succeeded in getting from a normal cord a true crossed- 
reflex effect in response to a single induction shock, i.e. a reflex effect in 
one gastrocnemius^ when the sciatic nerve of the opposite side was 
stimulated, although I have tried to get it often in well-cooled decerebrate 
frogs. Even from the one normal cord, which gave strong same-side reflex 
responses (fig. 3), no crossed-reflex effect, either mechanical or electrical, 
could be evoked by such stimulus. I have, however, not infrequently 
obtained, when very strong induction shocks have been used to excite the 
sciatic nerve (so strong as to be easily felt on the tongue), responses, 
mechanical and electrical, in the gastrocnemius of the opposite side, which 
T might have mistaken for reflex responses, had I not already acquired 
information about these when the nerve of the same side is excited. Had 
I only been able to record the mechanical response and to observe the 
electrical one, it would have been very difficult to know whether they were 
or were not reflex. The developed photographs, however, showed that the 
record of the electrical response in these cases was almost precisely identical 
with the direct efl'ect obtained in response to stimulation of the same-side 
sciatic, and that it occurred but 1 to I'So- later than this direct response, 
and consequently long before any reflex response to excitation of the nerve 

' Nor in the seiuitendinosus, nor the biceps femoris (see, however, footnote to p. 59), 
both of which muscles I have now used a good deal for my experiments (in the summer 
and autumn of 1907), Professor Sherrington having pointed out to me that muscles 
which are wholly flexor would be more likely to respond to reflex excitation than the 
gastrocnemius. The records of electrical responses taken with these two muscles afford 
confirmatory evidence of all the conclusions come to from experiments made with the 
gastrocnemius mentioned in this paper. As one would expect, they show that the 
response is purely reflex (and not preceded by any direct effect) when the sciatic nerve of 
the same side as the recording thigh muscle is excited near the knee, although it some- 
times happens that the conditions become, or can be made, favourable for the appearance 
in the electrical response of the counterpart of what in a mechanical response is known 
as " paradoxical contraction," and this then precedes the true reflex effect. 



Transmission-time of Reflexes in Spinal Cord of Frog :i\ 

of the same side. In more than one excitable preparation, such a response 
was obtained by merely removing the large resistance in the secondary 
circuit of the current used to excite the nerve of the opposite side, or by 
reducing it to one of not more than 10,000 ohms, if the secondary coil 
were right up or nearly so. The difference of time in such response, 
according as the nerve of the same side or that of the opposite side was 
stimulated, is about that which would be taken for an impulse to traverse 
between 30 to 45 extra mm. of nerve, and the only suggestion I can make 
as to what happened when such abnormally strong exciting currents were 
used is that these escaped to some nerve fibres in connection with the 
motor cells, or nerve, of the opposite side. The same effect can be produced 
not only by using abnormally strong induction currents, but also by making 
something in or about the ventral part of the cord, according to Baglioni ^ 
the motor cells, abnormally excitable, by the injection of phenol into the 
circulation. In the two experiments I have as yet made with such phenol 
preparations, records taken of responses to stimuli not quite but very 
nearly strong enough to be felt on the tongue, applied to the sciatic nerve 
of the opposite side, all showed, and in both preparations, besides the true 
crossed-reflex effect (which I shall have to refer to immediately), an effect 
occurring but Icr later than the direct effect as obtained by excitation of the 
same-side sciatic. This had every appearance of being a direct effect itself, 
and I cannot conceive that it was anything else, in spite of the fact thai it 
was produced by stimulation of the nerve of the opposite side. When the 
exciting current was weakened, the true crossed-reflex effect appeared as 
before, and was equally strong, but it appeared alone. I have not, as yet, 
further investigated the phenomenon, because it has seemed to me that all 
the evidence obtained by the use of stinuili more nearly approaching, 
though still exceeding in strength, the stimuli which must occur in nature, 
goes to show that no physiological interest would be furthered by so doing. 
That some morphological or even pathological interest would be served 
thereby seems to me, on the other hand, to be not at all improbable, but it 
is not with such interests that we are, for the moment, concerned. 

I think there can be no doubt that what Rosenthal- described as a 
crossed-reflex contraction to be obtained only with very strong exciting 
currents, and occurring at the same time as, or sometimes even before, what 
he considers to be the same-side reflex (whether or not it was so), was 
nothing but such a direct eflect due to the abnormal strength of the current. 
There is nothing in Wundt's treatise to show that he ever obtained a 
crossed-reflex contraction without the use of strychnine. The consideration 
of the conclusion that he came to that the crossed-reflex time was only Aa 
longer than that of the same-side reflex (" uncrossed," as we may now. if we 

' Baglioni, A. f. (Anat. u.), Pliysiol., 1900, Supplement. 

■^ Rosenthal, Abb. Berliner Akad., 1873, p. 104. The short mechanical latencies 
observed by Francois Franck in what he considers to be the crossed reflex in the guinea- 
pig, when very strong electrical stimuli were used, would fall, I believe, into the s;uue 
category. 



32 Buchanan 

like, call it), may therefore be postponed until the results obtained from my 
own experiments on strychnine frogs have been stated. To these we may 
now turn. 

Even in preparations in which the excitability of the cord has been 
raised by strychnine sufficiently to make certain of the presence of the 
same-limb reflex effect in the records of the electrical responses, it is by no 
means certain that a reflex effect will be also obtainable when the nerve 
of the opposite side is stimulated by a single break induction shock. There 
is, however, every likelihood of obtaining a response to such a stimulus 
when the strychnine has been allowed a certain time (varying with tempera- 
ture, season, dose, etc.,) to act, and there is hardly Siny doubt about obtaining 
it in a preparation which has been injected several hours before and is 
either still showing in its attitude all the external symptoms of strychnine 
poisoning, or has begun to recover. 

The crossed-reflex response has been recorded in sixteen preparations, 
many of which have already been referred to in connection with the same- 
side reflex. In all, the scia tic nerve of the opposite side was prepared, usually 
before beginning the experiment, in exactly the same way as its fellow, 
and a second pair of needle electrodes was applied to it. A Pohl reverser 
without crossed wires in the secondary circuit made it easy to excite the 
two nerves alternately. 

The measurements obtained from the records of the several responses, 
and the probable cord delay estimated therefrom in each, are given for 
thirteen of the experiments on the right-hand side in the tables beginning 
on p. 35, which are so arranged that the time, measured and estimated, 
in any crossed-reflex response may be readily compared with that, measured 
and estimated in the same way, in the same-side reflex response, recorded 
immediately before or after it, and given on the left-hand side. 

In order to estimate the cord delay when the nerve of the opposite side 
was stimulated, I have subtracted from the whole measured time, firstly, the 
time known, from the measurement of the same-limb reflex record, to be 
taken by an impulse to reach the recording spot on the muscle directly (i.e. 
through the motor part of the nerve) from tlie near anode, and, secondly, 
the time which would have been taken to traverse the measured length of 
nerve from the far anode to the near one, on the assumption that it 
travels at the rate of 30 metres per second. The first of these time values 
would be the same as when the same-limb reflex was recorded, provided that 
there had been no obstruction at the kathode to be overcome at the time this 
was taken, in which case allowance would have to be made for the fact. The 
second of these time values would also be very nearly the same in the two 
kinds of responses, the distance up one nerve and down the other as far as 
the near anode being, if the needles were on corresponding parts, just about 
the same as the distance from the near anode to the cord and back, and 
the time taken to traverse it only being different if very strong ascending 
currents were being used. In the last column of the table, I have given 



Transmission-time of Reflexes in Spinal Cord of Froc( 83 

the extra delay in the case of the crossed reflex as it appears to have been 
in each response. It is estimated by the subtraction of the cord delay which 
is shown to have probably occurred in the case of the immediately preceding 
or immediately following same-limb reflex response, from the probable 
cord delay in the case of the crossed reflex. I have indicated, by bracket- 
ing the lines representing them together, which same-limb reflex response 
I have employed for the purpose in each case. 

Two of the experiments in which also crossed-reflex responses were 
obtained, but which do not appear in the tables, may be first referred to. 
The one [Exp. 38] was one of the four already mentioned (p. 23) as 
showing no sign of the effect of strychnine in the records, electrical or 
mechanical, of the same-side reflex responses. It was one in which an 
extremely minute dose (0"012 mgr.) had been injected only a quarter of an 
hour before records began to be taken. After two had been taken with 
the sciatic nerve of the same side as the recording muscle excited, which 
subsequently showed on measurement that the probable cord delay had 
been in the two responses respectively 13cr and 14"ocr, a third record was 
taken with the opposite sciatic nerve excited. A reflex response extremely 
weak, but lasting about twice as long as when obtained by the excitation of 
the other nerve, was recorded ; but measurement showed that it did not 
occur until 83cr after excitation, the cord delay being therefore 634o-(83cr — 
3"5cr — 1'6(T— 14-5o-) longer than in the simpler reflex. The cro.ssed-reflex 
response could not be obtained a second time, although three more records, 
taken when the first nerve was excited, showed that the same-side reflex 
response was st.ill being obtained unaltered, and with an unaltered probable 
cord delay, this being respectively in the three : 14"5cr, 14o-, and 13'8o-. 
From none of the other three preparations, in which, though strychnine 
had been given, there was no evidence of the fact in the records of the same- 
side reflex response, could a crossed-reflex response be obtained at all. 

The second experiment, which may be described at once [Exp. 33], was 
one made on a small frog into which five times as much strychnine had 
been injected as in the one just mentioned, after the muscle and nerves hlld 
been prepared and the response without strychnine recorded. Almost 
immediately after the injection one good record was obtained of the 
same-side reflex and one of the crossed-reflex response. These are re- 
produced in fig. 5. The upper curve (A) represents the response when 
the nerve of the same side was excited. It shows first the response to 
direct excitation of the motor nerve, which was in this pi-eparation again 
double, then that to reflex excitation (24-7(r later). The lower curve (B) is 
the response of the same muscle when the sciatic of the opposite side was 
excited. The impulse took rather more than twice as long to reach the 
muscle (54'5o-), but when it began to aflect it, it aftected it much in the 
same way. There was an unavoidable delay of 10 to 15 minutes before the 
experiment could be continued. It was then seen that both reflex electrical 
responses had become serial. Their records showed that both had become 

VOL. 1. — JAN. 1908. 3 



34 



Buchanan 



very feeble, but that the crossed-reflex effect manifested itself 20o- earlier 
than before. The probable cord delay in the case of the same-side reflex 
was still 227 (T (24-7(7— 2 0o-); in the case of the crossed reflex, it had been 
at first 47'9or (54"5o- — 4"8cr— l"8cr), and was now about 28o-, the extra delay 
in its case having been at first 25'2cr, then, after the dose (which was a 
good deal strono-ei- than was usualh^ employed for a frog of such small size) 




Fig. 5. —Electrical responses of the gastrocnemius of a small frog 
which, after being prepared and immediately before records 
began to be taken, had been injected with 1 minim O'l per cent. 
liquor strychniae (Exp. 33). 

A, first response obtained when the intact sciatic nerve of the same side 
was excited. [Time lines 710 per second.] B, First response obtained 
when the sciatic nerve of the opposite side was excited. [Time lines 
700 per second.] 



had so taken efi^ect as to produce one of the best known symptoms of the 
action of strychnine, becoming about Qcr. 

Before discussing the change which took place in the course of this 
particular experiment, it will be well to consider what happened in other 
preparations which were made either a longer time after the injection of a 
very minute dose of the drug, or still early, but mth a more moderate dose 
than in the second of the two experiments just referred to, and which 
yielded crossed-reflex responses more frequently than only once or twice. 
It is to thirteen of these that the following tables refer. 



Transmission-time of Reflexes in Spinal Cord of Frog 



35 



Exp. 4. (L. Gastroc). Sept. 17, 1906. Room temp. 16' C. Large frog in- 
jected with I minim 0*1 per cent. liq. strych. two hours before ; spasms 
now general when any part of skin touched. The right gastrocnemius 
had already given a few weak same-side reflex responses, which were 
recorded, but no crossed response could be obtained from it, nor could 
this be obtained at first with the second muscle. 



Induction current 


Time, iu thousandths of a second, in 


Extra 
delay 
in the 
case 
of the 
crossed 
reflex. 


to nerve. 


Same-limb reflex. 


Crossed reflex. 


Strength. 


Direc- 
tion. 


Measured 
time taken 
by im- 
pulse to 
reach p 
directly. 


Pleasured 
time in- 
terval 
between 
arrivals of 
direct and 
reflex 
effects. 


Assumed 
time to be 
deducted 
for trans- 
mission 
in nerve. 


Pro- 
bable 
cord 
delay. 


Measured 
time taken 
by impulse 
to reach p. 


Time to be 
deducted 
for trans- 
mission 
from Cu 
electrode 
on opposite 

side to p 
(me.-jsured). 


Time to be 
deducted 
for trans- 
mission 
in nerve 

from one Cv 

electrode to 
the other 

(assumed). 


Pro- 
bable 

cord 
delay. 


r 12,000 
\ 10,000 
( 10,000 
/ 10,000 
1 10,000 


a 
a 
d 
a 
d 


3-7 


12-5 
12 -'5 


2-7 
2-7 


9-8 
9-8 


30 

25 

27 


3-7 
3-7 

3-7 

... 


2-6 

2-6 

2-6 


23-7 

18-7 

26-7 


13-9 
8-9 

10-9 



All the responses, direct and reflex, were weak ; the reflex ones were 
prolonged, but not serial. 



Exp. 16. (R. Gastroc). Oct. 4, 1906. Room temp. 16° C. One minim 01 
per cent. liq. strych. injected after preparation of the nerve and muscle 
and after -taking three records with the cord normal (none of which 
showed a reflex eftect) of responses to a fairly strong stimuli. Nor 
did two records show it, taken respectively five and ten minutes after 
the injection. Five minutes later still it was present. 



Induction current 
to nerve. 






Time, in 


thousandths of a second, in 


Extra 
delay 
in the 
case 
of the 
crossed 
reflex. 


Same-limb reflex. 


Crossed reflex. 


Strength. 


Direc- 
tion. 


Measiu-ed 
time taken 
by im- 
pulse to 
reach p 
directly. 


Measured 
time in- 
terval 
between 
arrivals of 
direct and 
reflex 
effects. 


Assumed 
deduction 
for trans- 
mission 
of nerve. 


Pro- 
bable 
cord 
delay. 


Measured 
time taken 
l)y impulse 
to reach p. 


Time to be Time to be 
deducted deducted 
for trans- for trans- 
mission mission 
from Cu in nerve 
electrode from one Cu 

on opposite electrode to 
side to p i the other 

(measured). | (assumed). 


Pro. 
bable 
cord 
delay. 


10,000 


d 


3-7 


17-5 


1-9 


15-6 






... 






10,000 


d 


3-7 


18-6 


1-9 


16-7 


... 










) 10,000 
/ 10,000 


d 


3-7 


17-5 


1-9 


15-6 












d 










36-3 


3-7 


1-8 


30-8 


15-2 


I 10,000 
t 10,000 


d 


3-7 


15 


i-9 


13-1 












d 








... 


34 


3-7 


1\S 


28-5 


15-4 


10,000 


d 


3-7 


13-7 


1-9 


11-8 













All the electrical effects were weak. The reflex effect in the last three re- 
sponses was two to three times as long as the direct effect. In the first 
two it was of about the same duration. The contractions were all feeble. 



36 



Buchanan 



Exp. 18. (L. Gastroc). Oct. 8, 1906. Room temp. 16° C. Responses of the 
right gastrocnemius to excitation of the right sciatic had first been 
recorded with the cord normal. They showed a reflex effect when the 
stimuli was 5000 units in strength ; none when it was 1000 units. 
After preparing the left gastrocnemius and recording two responses to 
excitation of the left sciatic, neither of which (though the strengths 
were 5000 and 10,000 units) showed a reflex effect, 1 minim 0"1 per 
cent. liq. strych. was subcutaneously injected. Records taken I hour 
later. 











Time, in thousandths of a second, in 








Induction current 














Extra 
delay 


to nerve. 


Same-limb reflex. 


Crossed reflex. 








Measured 
time in- 
terval 
between 
arrivals of 
direct and 
refle.x 
effects. 






Time to be 


Time to be 




Strength. 


Direc- 
tion. 


Measured 
time taken 
by im- 
pulse to 
reach p 
directly. 


Assumed 
deduction 
for trans- 
mission 
of nerve. 


Pro- 
bable 
cord 
delay. 


deducted 
Measured | '''^^^^■ 
time taken i i?''f ^.^ 
^y;-P^- lleZol 
to reach p. ^^ opposite 
side to p 


deducted 
for trans- 
mission 
in nerve 
from one Cu 
electrode to 
the other 


Pro- 
bable 
cord 
delay. 


in the 
case 
of the 
crossed 
reflex. 

i 












(measured). 


(assumed). 






5,000 


d 


3-7 


28 


2 


26 










1,000 


d 


3-7 


no 


reflex efF 


ect 










10,000 


d 


3-7 


28 


2 


26 






... 


... 1 


5,000 


d 










no effect 








r 10,000 


d 










43 


3-7 


1-8 


37-5 


15-5 


\ 10,000 


d 


3-7 


24 


2 


22 












('10,000 


d 










41 


3-7 


i-8 


35-5 


15 


5,000 


d 


3-7 


22-5 


2 


26-5 












1 10,000 


d 




... 






38 


3. 


1-8 


32-5 


12 1 



All the effects (direct and reflex) were weak. 



Transmission-time of Reflexes in Spinal Cord of Frog 



37 



Exp. 40. (R. Gastroc). Nov. 13, 1906. Room temp. 12" C. One minim 
0*04 per cent. liq. stryeh. injected the day before making the experiment. 
Body still somewhat stiff; limbs no longer so. Muscles responding 
reflexly to slightest touch of skin. 





















Induction current 
to nerve. 
















F.xtra 
delay 
in the 
case 
of the 


Same-limb reflex. 


Crossed reflex. 






Measured 
timetalien 


Measured 
time in- 
terval 


Assumed 
time to be 


Pro- 


Pleasured 


Time to be 
deducted 
for trans- 


Time to be 
deducted 
for trans- 
mission 
in nerve 
from one Cu 
electrode to 
the other 


Pro- 


Strength. 


Direc- 
tion. 


by im- 
pulse to 
reach p 
directly. 


between 

arrivals of 

direct and 

reflex 

effects. 


deducted 
for trans- 
mission 
in nerve. 


bable 
cord 
delay. 


time taken 
l)y impulse 
to reach p. 


from Cu 

electrode 

on opposite 

side to p 


bable 
cord 
delay. 


crossed 
reflex. ; 


) 5,000 












(measured). 


(assumed). 






d 


5 


22 


2-3 


19-7 










) 5,000 


d 


... 




... 




48 


5 


2-4 


40-6 


•2()-9 


( 5,000 


d 


5 


20-5 


2-3 


18-2 








... 




} 5,000 


d 










46 


5 


2-4 


38-6 


20-4 


( 5,000 


d 


5 


20-5 


2-3 


18-2 












j 5,000 


d 


... 








55 


5 


2-4 


47-6 


21-9 


1 5,000 


d 


5 


28 


2-3 


25-7 








... 




( 5,000 


d 










51 


5 


2-4 


43-6 


18-1 


^5,000 


d 


5 


27-8 


2-3 


25-5 












( 5,000 


d 






... 




44 


5 


2-4 


36-6 


111 



Reflex eflfects less strong than the direct effect, except in the 2nd and 3rd 
same-side and in the 2nd crossed-reflex response. For records of first 
two responses see fig. 7, p. 50. The direct eflect was only double the 
first two times it was recorded. 



38 



Buchanan 



Exp. 42. (R. Gastroc). Nov. 15, 1906. Room temp. 15" C. One minim 
0-01 per cent. liq. strych. injected just before making the preparation. 
The same-side reflex had been ah'eady recorded several times with the 
cord at different temperatures (see p. 28). No crossed-reflex could be 
obtained until a second dose (of 2 minims 0-01 per cent. liq. strych.) 
had been injected, and then at first it could not be obtained with any 
streno-th under 14,000 units. 



Induction current 


to nerve. 


Strength. 


Direc- 
tion. 


/ 14,000 
1 10,000 


d 


d 


t 14,000 
i 10,000 


d 


d 


5,000 
12,000 


d 


d 


/ 4,000 
\ 11,000 


d 


d 


/ 5,000 
\ 10,000 


d 


d 


10,000 


d 


10,000 


d 


10,000 


d 



Time, in thousandths of a second, in 



Sarae-limb reflex. 



time taken 
by im- 
pulse to 
reach p 
directly. 



Measured 
lime in- 
terval 
between 
arrivals of 
direct and 
reflex 
efifects. 


Assumed 
time to be 
deducted 
for trans- 
mission 
in nerve. 


131 


2'3 


14-3 


2-3 


14-3 


2-3 


14-4 


2'3 


13-4 


2-'3 


12 


2'3 


12 


2"3 



bable i time taken 
cord by impulse 
delay. ! to reach p. 



Time to be 
deducted 
for trans- 
mission 
from Cu 
electrode 

on opposite 
side to p 

(measured). 



Time to be 
deducted 
for trans- 
mission 
in nerve 

from one Cm 

electrode to 
the other 

(assumed). 



10-8 

12 
12 

12-1 i 

iTij 

9-7J 
9-7 1 



321 

29'8 

28-8 
36" 

28 

28 



2-3 

2-3 



4 2-3 

4 i 2-3 

.. I ... 
4 2-3 

4 2-3 



Pro- 
bable 

cord 
delay. 



25-8 
23-5t 

22"5t 
23-7t 
21-'7+ 
21-7 



Extra 
delay 
in the 
case 
of the 



reflex. 

15 
11-5 

10-5 

"11 -6 

( l("»-6 } 

12 ' 



The crossed-reflex effects were in the first two responses very weak ; they then 
became stronger, though not always attaining their maximal strength 
until late.^ Their duration was somewhat shorter than that of the 
uncrossed-reflex effects, but, with the exception of the first two, was 
longer than the direct effect. 

1 The t when it occurs in the tables signifies that the moment at which the effect 
became maximal was deferred. 



Transmission-time of Reflexes in Spinal Cord of Frog 



39 



Exp. 45. (L. Gastroc). Nov. 20, 1906. Room temp. 12 C. Five minims 
0'005 per cent. liq. strych. injected -| hour before. Records had been 
taken of the electrical responses of the right gastrocnemius both before 
and after the injection. No cross-reflex response had been obtainable 
with that muscle. 



1 








Time, in thousandths of a second, in 








Induction cuirent 
to nerve. 


















Extra 
delay 


Same-limb reflex. 




Crossed reflex. 


















Time to be 








Measured 
time taken 


time in- 
terval 


Assumed 
time to be 


Pro- 


Measured 


deducted 
for trans- 


deducted 
for trans- 


Pro- 


in the 
case 
of the 


Strength. 


Direc- 
tion. 


by im- 
pulse to 


ijetween 
arrivals of 


deducted 
for trans- 


bable 
cord 


time taken 
l)y impulse 


from Cm 

electrode 

on opposite 


in nerve 

from one Cu 

electrode to 

the other 


bable 
cord 


crossed 
reflex. 






reach p 
directly. 


direct and 
reflex 


mission 
in nerve. 


delay. 


to reach p. 


delay. 


















(measured). 


(assumed). 






j 12,000 
} 12,000 


d 










39-3 


5 


2-5 


31-8 


11-8 


d 


5 


22 -.5 


2 '5 


20 


... 






... 




/ 10,000 


d 










39-3 


5 


2-5 


31-8 


12-9 


1 10,000 


d 


5-2 


21-4 


2-5 


18-9 










... 


) 10,000 
1 10,000 


d 










39-3 


5 


2-5 


3i-8 


12-3 


d 


5-'5 


22 


2-5 


19-5 












r 10,000 

{ 10,000 
1 10,000 


d 










36-9 


5 


2-5 


29-4 


10-5 


d 


5-'5 


2i-4 


2-5 


18-9 


... 




... 






d 










369 


5 


2-5 


29-4 


l6-5 










Ten mi 


lutes 


later :— 










10,000 


d 


5-3 


27-5 


2-5 


25 












10,000 


d 




... 


... 


_^'_ 


no 


effect 


... 




... 



The same-side reflex effects were of about the same strength and hardlj^ 
longer than the direct effect. The cross-reflex effects were weak 
throughout. For records of the 1st and 2nd responses see fig. !S 
(B and A); of the 7th and 8th, fig. 8 (D and C), p. 51. 



40 



Buchanan 



Exp. 47. (R. Gastroc). Nov. 23, 1906. Room temp. 16° C. ; moist chamber 
kept at 14° C. Two minims O'Ol per cent. liq. strych. injected I hour 
before preparing the muscle and nerve. No crossed effect could be 
obtained at first with a stimulus under 14,000 units in strength. 



- 








Time, in thousandths of a second, in 




1 




Induction current 
to nerve. 


















Same-limb reflex. 




Crossed reflex. 




Extra 
















Time to be 


Time to be 


1 


delay 






Measured 


Measured 
time in- 


Assumed 






deducted 

fr>r trnnB- 


deducted 


1 


in the 
case 






time taken 


terval 


time to be 


Pro- 


Measured 


mission 
from Cu 
electrode 


mission 


Pro- 


of the 


Strength. 


Direc- 


by im- 


between 


deducted 


bable 


time taken 




bable 


crossed 


tion. 


pulse to 


arrivals of 


for trans- 


cord 


by Impulse 


from one Cu 


cord 


reflex. 






reach p 


direct and 


mission 


delay. 


to reach p. 


on opposite 


electrode to 


delay. 








directly. 


reflex 
effects. 


in nerve. 






side to p 
(measured). 


the other 
(assumed). 






/ 10,000 


d 


4-2 


19-8 


2-5 


17-3 












\ 14,000 


d 








35-1 


4-'2 


2-3 


28'6 


ir-3 


r 10,000 


d 


4-2 


19-8 


2-'5 


17-3 










\ 14,000 


d 








35 


4'2 


2-3 


28-5 


ll'-4 


j 10,000 


d 


4-8 


18-3 


j +0-6 


16-4 










/ 10,000 


d 






\ -2-5 


33-2 


4-'2 


2-'3 


26"7 


10'3 


/ 5,000 


d 


4-2 


19'5 


2-5 


17 








... 


{ 9,000 


d 








31-2 


4-2 


2'3 


24"7 


7-7 1 


/ 5,000 


d 


4-5 


19-5 


2-5 


17 








... 


\ 8,000 


d 








29-0 


4"2 


2-3 


23 


6 


/ 4,000 


d 


4'6 


20 


2-5 


17-5 ... 






... 




\ 7,000 


d 




... 




30-2 


4'6 


2-3 


23-3t 


'5-8 


f 3,000 


d 


4-6 


19-5 


2'5 


17 










■\ 6,000 

1 


d 






... 


31-5 


4-6 


2-3 


24-6 


'7-6 



The records show that the same-side reflex response was, to begin with, 
somewhat weaker, but of about the same duration as the direct response; 
but that it soon became as strong, and four times, then six or seven 
times, as long. The direct response was reduced in strength by the 
weakening of the stimulus to 4000 and 3000, while the reflex response 
was not ; so that this became finally not only longer, but stronger than 
the direct response. 

The crossed-reflex effect in the 1st response was very feeble. In the two 
next there was apparently a second stimulus of central origin, affecting 
the muscle 47(t and 37(t respectively after the 1st. There was this 
again in the last response but one (the 6th), whereas the only same- 
side reflex responses which showed it were the 4th and 6th. In the 
3rd, 4th, 5th, and 6th responses the crossed-reflex effect was quite as 
strong as that of 'the same-side reflex, although in the 6th it did not 
attain its maximum until some llcr after it had begun to be effectual. 
Its duration was shortest, being about twice that of a direct effect, in 
the 4th. 

For records of the 5th and 10th responses (of the whole series) see fig. 9. 
p. 52. 



Transmission-time of Reflexes in Spinal Cord of Frog 



41 



Exp. 48. (R. Gastroc). Nov. 27, 1906. Room temp. 16° C. (see p. 13). 
No reflex responses could be obtained to excitation of the nerve of the 
opposite side, even with strength of current 14,000, while the responses 
to excitation of the nerve of the same side of which the time measure- 
ments are given on p. 13, were showing quite strong reflex effects. 
It was even then very small, but two records of it were taken. 



Induction current 
to nerve. 



Time, in thousandths of a second, in 



Same limb reflex. 



Strength, ^ire^" 



14,000 
10,000 
14,000 
14,000 





Measured 


Measured 


time in. 


time taken 


terval 


by im- 


between 


pulse to 


arrivals of 


reach p 


direct and 


directly. 


reflex 




effects. 


5 


12-5 


5 


12-5 i 


.... 





Assumed 
time to be 
deducted 
for trans- 



Pro- 
bable 
cord 
delay. 



in nerve. 



1-7 
1-7 



10-8 
10-8 



Crossed reflex. 



Time to be I Time to be 

deducted deducted 

Measured for.tra»s- for trans- p^ 
time taken J°'t^>°» P"*^'°° bable 

to reach p. ^,j opposite , electrode to ; ^^^^^ 

side top I the other 

(measured), (assumed). 



35 
31-5 



3-7 



2-6 

2-6 



29-7 
25-2 



Extra 
delay 
in the 
case 
of the 
crossed 
reflex. 



18-9 
14-4 



Bag of ice then put on the back of the preparation for \ hour. 



14,000 


d \ 


10,000 


d 


10,000 


d 


8,000 


d 


10,000 


d ' 


10,000 


d 


10,000 


d 


10,000 


d 


10,000 


a 


14,000 


a 


14,000 


a 



5-5 



6 
4-5 



4-5 



33-3 


34'5 


33-3 



37-5 
36 



1-7 

i-V 

1-5 

i-3 

4 





72-8 


3-7 


2-6 


31-6 










72-8 


3'7 


2-6 


32-'8 






... 




73-4 


4 


2-6 


3l'8t 


74-1 


4 


2-6 


34-7 


... 


... 


... 


33-5 






... 




79-4 


4 


2-6 


32 


... 


... 


... 



66-5t 34 
66-5 33-7 



66-8 I 
67-5 I 

72-8t 



35 
32'8 



39-3 

or I 
40-8 



The two kinds of reflexes closely resembled one another in their effects 
after the cord had been cooled. None of the effects were strong, and 
in the 1st and 5th crossed-reflex responses and in the 3rd uncrossed 
one, the maximal strenofth was not attained at once. 



42 



Buchanan 



Exp. 49. (R. Gastroc). Nov. 30, 1906. Room temp. 15° C. One and a half 
minims O'Ol per cent. liq. strych. injected the day before. Recovery 
almost complete. 











Time, in thousandths of a second, in 








Induction current 


















to nerve. 




Same-limb reflex. 




Crossed reflex. 




Extra 








Measured 
time in- 
terval 






Time to be 


Time to be 




delay 






Measured 
time taken 


Assumed 
time to be 


Pro- 


Measured 


deducted 
for trans- 
mission 


deducted 
for trans- 
mission 


Pro- 


in the 
case 
of the 


Strength. 


Direc- 


by im- 


between 


deducted 


bable 


time taken 


from Cm 
electrode 




bable 


crossed 


tion. 


pulse to 


arrivals of 


for trans- 


cord 


l)y impulse 


from one Cu 


cord 


reflex. 






reacli p 
directly. 


direct and 
reflex 
eflfects. 


mission 
in nerve. 


delay. 


to reach p. 


on opposite 
side to p 


electrode to 
the other 


delay. 








5-4 








(measured). 


(assumed). 






/ 7,000 


a 


19 4 


2 


17-4 










1 


1 t 7,000 


a 






... 




36-3 


'5 


'2 


29'3 


li'9 1 


( 5,000 
\ 5,000 


a 


5-1 


19-4 


2 


17-4 












a 










35 


5 


'2 


28 


l6-6 


j 5,000 
1 \ 5,000 


d 


6 


19'l 


2 


17-1 












d 










33-3 


5 


"2 


26-3 


"9-2 


I ( 5,000 


d 


5 


19-5 


'2 


17-5 










... 


s \ 5,000 


d 










33-2 


5 


"2 


26"2 


8-7 


,(3,000 


d 


5" 


19-5 


"2 


17-5 


... 










1 2,000 


d 


no effect, 


either di 


rect or re 


flex 












/ 5,000 


d 








... 


32-9 


5 


2 


25'9 


8-7 


\ 3,000 


d 


5 


19-2 


"2 


17-2 












/ 5,000 


a 


5 


18-3 


2 


16-3 












15,000 


a 






... 




33 


5 


2 


26 


9'7 



None of the reflex effects were quite as strong as the direct. The crossed- 
reflex effect was, however, throughout, strong ; the uncrossed became 
weaker the 5th and 6th times it was recorded. All the reflex effects 
had a duration which was about double that of the direct effect. 



Transmission-time of Reflexes in Spinal Cord of Frog 



43 



Exp. 52. (R. Gastroc). Dec. 4, 1906. Room temp. 16° C. Two minims 
0-02 per cent. liq. strych. injected three hours before. Arms had been 
flexed and legs extended for | hour. Heart still beating, though 
slowly. 

















1 


Induction current 
to nerve. 
















Extra 
delay 
in the 
case 
of the 


Same-limb reflex. 


Crossed reflex. 






Measured 
time taken 


Measured 
time in- 
terval 


Assumed 
time to be 


Pro- 


Measured 


Time tu be 
deducted 
for trans- 


Time to be 
deducted 
for trans- 
mission 
in nerve 
fiom one Cu 
electrode to 
the other 


Pro- 


Strength. 


Direc- 
tion. 


by im- 
pulse to 
reach p 
directly. 


between 

arrivals of 

direct and 

reflex 

effects. 


deducted 
for trans- 
mission 
in nerve. 


bable 
delay. 


time taken 
by impulse 
to reach i). 


from Cu 

electrode 

on opposite 

side to p 


bable 
cord 
delay. 


crossed 
reflex. 














(measured). 


(assumed). 






/ 5,000 
\ 5,000 


a 


3-6 


32-3 


1-8 


30-5f 




... 








a 










55-5 


3-6 


1-9 


50 


19-5 


/ 5,000 


a 


3-6 


30-5 


i-8 


28-7+ 








... 


... 


{ 5,000 


a 








... 


53-8 


3-6 


1-9 


48-3 


19-6 


/ 5,000 


a 


3-6 


27-9 


i-'s 


26-1 




... 








\ 5,000 


a 










51-5 


3-6 


1-9 


46 


19-9 


) 5,000 


a 


3-6 


29-2 


i-8 


27-4+ 




... 


... 


... 




I 5,000 


a 






... 


... 


5i 


3-6 


1-9 


45-5 


18-1+ 


/ 5,000 


d 










52-9 


3-6 


1-8 


47-5 


20-2 


\ 5,000 


d 


3-6 


29 


it 


27 -St 






... 







All the reflex electrical responses serial. Maximum strength of crossed- 
reflex effect sometimes greater than the same-side effect. ]\Iaximum 
not so great in either as in the direct effect. For records of the 5th 
and 6th responses see flg. 10 (A and B) on p. 52. 



44 



Buchanan 



Exp. 53. Large Frog (R. Gastroc). Dec. 6, 1906. Room temp. 14° C. 
Two minims 0*02 per cent. liq. strych. injected the day before. Recovery 
ahnost complete. 











Time, in 


thousandths of a second, in 








Induction current 
to nerve. 


















Extra 
delay 


Same-limb reflex. 


Crossed reflex. 








TH^ 1 








Time to be 


Time to be 








Measured 


Measured 
time in- 


Assumed 






deducted 
for trans- 
mission 


deducted 
for trans- 




in the 
case 






time taken 


terval 


time to be 


Pro- 


Measured 


mission 


Pro- 


of the 


Strength. 


Direc- 


by im- 


between 


deducted 


bable 


time taken 


from Cm 


bable 


crossed 


tion. 


pulse to 


arrivals of 


for trans- 


cord 


by impulse 


from one Cu 


cord 


reflex. 






reach p 
directly. 


direct and 
reflex 
effects. 


mission 
in nerve. 


delay. 


to reach p. 


on opposite 

side to p. 

(measured). 


electrode to 
the other 
(assumed). 


delay. 




/ 5,000 


d 


4-8 


20-4 


2-3 


18-1 












\ 5,000 


d 










43'3 


i's 


2"2 


36"3 


18'2 


( 10,000 


d 


... 








45-2 


4-8 


2-2 


38-2 


19-8 


I 10,000 


d 


4-8 


20-7 


2'3 


18-4 












( 10,000 


d 






... 




43 


4-8 


2-2 


36 


n'a 


( 10,000 

10,000 

( 10,000 


d 










39 


4-8 


2-2 


32 


15-3 


d 


4-8 


19 


2'3 


16-7 












d 










43 


4-8 


2-2 


36 


19'-3 


j 10,000 
I 10,000 


d 










40-6 


4-8 


2-2 


33-6 


17-4 


d 


4-8 


18'5 


2-3 


10'2 












{ 10,000 
} 10,000 


d 










40'4 


4-'8 


2-2 


33-4 


17'3 


d 


i'-'s 


18-4 


2-3 


16-1 












) 10,000 
} 10,000 


d 








... 


35-7 


4-8 


2-2 


28-7 


13-6 


d 


4-9 


17'4 


2-3 


15-1 




... 









(The big resistance had been removed from the secondary circuit when the 
last two responses were obtained.) 

The same-side reflex effects very nearly as strong as, and hardly longer than, 
the direct eflects, which were double throughout. Crossed-reflex effects 
extremely weak throughout, and weak out of all comparison with the 
uncrossed-reflex eflects. For records of the last two responses (before 
section of the nerve), i.e. the 13th and 14th, see fig. 6 (B and A) on p. 49. 



Transmission-time of Reflexes in Spinal Cord of Frot 



45 



Exp. 55. (R. Gastroc). Dec. 11, 1906. Room temp, rising from 10° C. to 
14° C. One minim 002 per cent. liq. strych. injected four days before, 
and 2 more minims the day before making the experiment. When 
this was begun, tlie preparation was in the attitude characteristic of 
strychnine poisoning. Heart beating but feebly at the end of the 
experiment. No response of any kind, direct or reflex, could be 
obtained to begin with, with stimuli weaker than 2000 units. 



Induction current 
to nerve. 




Time, in thousandths of a second, in 



nutes I rest now 
I 5 



Same-limb reflex. 




Measured 






time in- 


Assumed 




terval 


time to be 


Pro- 


between 


deducted 


bable 


arrivals of 


for trans- 


cord 


direct and 


mission 


delay. 


reflex 


in nerve. 




effects. 


2 


26t 


28 


30 


2" 


28t 


ect, direc 


t or refl 


ex 


26 


2 


24 


29 


"2 


27 1 


given 






31 


2 


29t 


30 


2 


28t 



Crossed reflex. 



Measured 
time taken 
by impulse 
to reach p. 



Time to be 
deducted 
for trans- 
mission 
from Cm 
electrode 
on opposite 

side to p 
(measured). 



lime to be 
deducted 
for trans- 
mission 
in nerve 

from one Cu 

electrode to 
the other 

(assumed). 



Pro- 
bable 

cord 
delay. 



48 
50 

52 
51 

57t 
6it+ 



Extra 
delay 
in the 
case 
of the 
crossed 
reflex. 



) 28 or 
I 25 
24 



33 



Reflex responses could now be obtained with induction currents of 1000 units, 
but not with any that were weaker. Ten minutes' rest was given before 
continuing the experiment. 



/ 1,000 


a 






... 


... 


61 t 


\ 1,000 


a 


5 


35 


2 


33 




j 1,000 


a 


... 


... 






68 


\ 1,000 


a 


5 


34 


2 


32tt 




Five mi 


nutes' 


rest now 


given 




... 




■ 1,000 


a 










59 


1,000 


a 


5 


33 


2 


sit 


... 


' 1,000* 


a 










69 


1,000* 


a 


5 


28 


2 


26t 




1,000 


a 










no 


\ 5,000 
/ 5,000 


a 










58 1 


a 


4-8 


31 


2 


29t 


1 


1 5,000* 
1 5,000* 


a 










56 I 


a 


5-2 


29 


2 


27t 




5,000 


a 


5 


28 


2 


26+ 


... 



effect 
5 



2 


54 


21 


2 


61++ 


29 


2 


... 
52 


21" 


2 


t52++ 


<31 ■ 
/ or 36) 


2 


51+ 


22 ^ 


'2 


49+ 


22 



The *, when it appears in the first column, denotes that the big resistance 
had been removed from the secondary circuit when the particular 
response was evoked. 



46 



Buchanan 
Exp. 55. — continued. 



Time, in thousandths of a second, in 




Induction current 




Extra 

delay 


to nerve. 


Same-limb refle.x. 


Crossed reflex. 






1 




Time to be 


Time to be 








Measured 
Measured time in- 
time talcen terval 


Assumed ! 
time to be Pro- 


Measured 


deducted 
for trans- 
mission 
from Cti 
electrode 

on opposite 
side to p 

(measured). 


deducted 
for trans- 


Pro- 


in the 
case 
of the 


Strength. 


Direc- 
tion. 


by im- between 
pulse to arrivals of 


deducted bable 
for trans- cord 


time taken 
by impulse 


in nerve 
from one Cu 
electrode to 

the other 
(assumed). 


bable 
cord 


crossed 
reflex. 






reach p direct and 
directly. reflex 
effects. 


mission delay, 
in nerve. 


to reach p. 


delay. 




Reflex responses could now be obtained to excitation of the sciatic of the opposite 


side with induction currents of 500 units ; soon afterwards it could be also 


obtained when that of the same side was so excited. Ten minutes' rest was 


given before continuing the experiment. 


( 500 a 










56 


5 


2 


49 


18 1 


\ 500 


a 


5 


33 


2 


Bit 


... 


... 






... t 


{ 500 


a 










69 


5 


2 


62+t 


31 


) 500* 


a 










56 


5 


2 


49t 


22 


\ 500* 


a 


5 


29 


2 


27t 












) 5,000 
\ 5,000 


a 








... 


55 


5 


2" 


48t 


24 


a 


5 


26 


2 


24t 








... 





l''he reflex eftects were serial throughout, the separate large rises of the 
mercury being distinct enough from one another to be counted by the 
eye in each response. There were sometimes as many as fifteen periods, 
the number being usually greater in the crossed reflex than in the 
uncrossed. As the first period was far from being over on a plate 
which took 0"15 second to pass the slit, they were probably recurring 
at a rate of about five a second. The records of the responses giving the 
cord delays marked f showed that the reflex eflfect was not as strong 
at the beginning as it subsequently became. All the records so marked 
had the character seen in the reproduction of the first of these (fig. 11, A, 
p. 54). The records of the responses giving the cord delays marked 
f f showed that the moment at which the eflfect became strongest was 
yet longer deferred, and that the strength of the eflTect was very 
small indeed at the beginning. The records so marked had the 
character shown in fig. 11, C and D. The crossed-reflex response was 
more readily produced than the same-side reflex by the weaker 
stimulus, i.e. one of 1000 units, and, later, one of 500 units was capable 
of producing a reflex eftect when applied to the left sciatic, before it 
was capable of producing it when applied to the right sciatic. The 
number of periods in the electrical reflex responses was no fewer with 
the weaker stimulus. There was the usual two minutes' rest given 
between recording successive responses, except when otherwise stated. 



Transmission-time of Reflexes in Spinal Cord of Frog 



47 



Exp. 56. Dec. 13, 1906. Room temp. 1 1° C.-13' C. Two and a half minims 
0-02 per cent. liq. strych. injected two days before. Arms still flexed, 
but legs no longer rigidly extended. 

Right Gastrocnemius. (Exp. 56 R.) 











Time, in thousandths of a second, in 








Induction current 
to nerve. 


















Extra 
delay 
in the 
case 
of the 




Same-limb reflex. 






Crossed reflex. 








Measured 
time taken 


Measured 
time in- 
terval 


Assumed 
time to be 


Pro- 


Measured 


Time to be 
deducted 
for trans- 


Time to be 
deducted 
for trans- 


Pro- 


Strength. 


Direc- 
tion. 


by im- 
pulse to 


between 
arrivals of 


deducted 
for trans- 


bable 
cord 


time taken 
by impulse 


mission 
from Cu 
electrode 
on opposite 
side to p 


mission 

in nerve 

from one Cu 

electrode to 

the other 


bable 
cord 


crossed 
reflex. 






reach p 
directly. 


direct and 
reflex 
effects. 


mission 
in nerve. 


delay. 


to reach p. 


delay. 
















(measured). 


(assumed). 






/ 3,000 


a 


5-3 


19-3 


2-2 


17-1 












\ 3,000 


a 










38'5 


4-'8 


2-'2 


31-5 


14-4 


) 3,000 


d 


4-8 


26'l 


2-1 


18 




... 








} 3,000 


d 


... 








38-5 


4-8 


2-1 


31-6 


13-6 


j 3,000 


d 


4-8 


26-1 


2'l 


18 


... 










/ 3,000 


d 










39-2 


4-'8 


2-1 


32-3 


14-3 


/ 2,000 


d 


4-8 


26'l 


2-1 


is" 












\ 2,000 


d 










38"8 


4-8 


2-1 


31-9 


13-9 


/ 1,000 


6 


4-8 


20-5 


2-1 


18-4 


... 










1 1,000 


d 










38-7 


4-'8 


21 


31-8 


13-4 


/ 3,000 


d 


4-8 


19-7 


21 


17-6 












\ 3,000 


d 










38-5 


4-8 


2-1 


3i'6 


14 


/ 9,000 


d 


5-2 


19-3 


2-1 


17-2 






... 






\ 9,000 


d 










37-3 


4-8 


21 


30-4 


13'2 




Inte 


rval, du 


•ing whi 


cli mecl 


lanict 


il respon 


5es were 


recorded. 







3,000 


d 


5-3 


21 


2-1 


18-9 












3,000 


d 










391 


4-8 


2-1 


32-2 


13-3 


500 


d 










66-1 


4-8 


21 


59-2 


(?)40-3 , 



Both reflex effects lasting about four times as long as direct effect in all, 
and at first about ecpial to it in strength. For records of the first four 
and of the 13th responses, see fig. 12, A, B, C, D, E, on p. 56. 



48 



Buchanan 



Left Gastrocnemius. (Exp. 56 L.) 





Time, iu thousandths of a second, in 


Extra 
delay 
in the 
case 
of the 
crossed 
reflex. 


to nerve. 


Same-limb reflex. 


Crossed reflex. 


Strength. 


Direc- 
tion. 


Measured 
time taken 
by im- 
pulse to 
reach p 
directly. 


Measured 
time in- 
terval 
between 
arrivals of 
direct and 
reflex 
effects. 


•Assumed 
time to be 
deducted 
for trans- 
mission 
in nerve. 


Pro- 
bable 
cord 
delay. 


Measured 
time taken 
by impulse 
to reach p. 


Time to be 
deducted 
for trans- 
mission 
in nerve 

from one Cij 

electrode to 
the other 

(assumed). 


Time to be 
deducted 
for trans- 
mission 
from Cm 
electrode 
on opposite 

side to p 
(measured). 


Pro- 
bable 
cord 
delay. 


j 3,000 
/ 3,000 
/ 2,000 
{ 2,000 
J 3,000 
\ 3,000 
/ 1,000 
\ 1,000 
/ 3,000 
\ 3,000 


d 
d 
d 

d 
d 
d 
d 
d 
d 
d 


5-3 
5'3 
5-3 
5-3 
5-3 


23 

23 

... 
23 

24'2 

23-4 


2-1 
2-'l 
2-1 
2-'l 

■ 2-'l 


20-9 

... 
20-9 

20-9 

22'-l 

2i'3 


4i'5 
42'-5 
43'5 

47-8 
43 


5-3 
5-3 

5-*3 
5-3 
5'3 


1-9 

i'-b 

1-9 

r-9 


... 
34-3 

35"3 

36-3 

40-6 
35'8 


13-4 
14-4 
15'4 
18-5 
14-5 



Both the direct and reflex efiects were less strong (curves less steep) in 
these records than in those taken with the other muscle, but they were 
about equal to one another. Reflex effects somewhat longer than when 
the muscle of the other side was recording. Contractions also not quite 
as strong, but of somewhat longer duration. 

The left sciatic nerve in this preparation was more excitable than the right, 
for both the crossed reflex and the uncrossed reflex could be obtained 
with it in response to a stimulus of 500 units' strength. The uncrossed 
reflex with this strength was recorded mechanically, but not electrically ; 
it was quite as strong as with a current of 1000 or more units. Xo 
reflex response of either kind could be obtained when the right sciatic 
was excited by a stimulus of 500 units' strength. Unfortunately, I did 
not, after cutting the left sciatic, try whetlier the muscle still responded 
when so weak a stimulus as one of 500 units was applied to the 
peripheral end — so that, although I did ascertain the contrary in the 
case of the right sciatic, it cannot be stated with certainty whether it 
was the nerve or cord of which the excitability varied. 



In four of these experiments [Nos. 4, 16, 42, 48] the crossed-reflex effect 
was unobtainable when the experiment was begun, even when, as in the 
two last, fairly strong reflex effects were being obtained in response to 
excitation of the sciatic nerve of the same side as the recording muscle. 
It was not until after respectively nine, three, ten, and nine same-limb 
reflex responses had been recorded (in all four cases at least a quarter of an 
hour later) that any response at all could be obtained when a single break 
induction current, even one strong enough to be felt on the tongue, was 
applied to the sciatic nerve of the opposite side. In Exps. 42 and 48, it 
was not until after a second very minute dose had been administered, and 
even then not for some ten minutes later, that a crossed-reflex eflect could 



Transmission-time of Reflexes in Spinal Cord of Frog 49 

be obtained. In all four experiments the crossed-reflex effect never became 
strong, though in the first two the same-side reflex effect was no stronger. 
In all these preparations, in which the first appearance, as it were, of a 
crossed-reflex effect in response to a single instantaneous stimulus was 
observed, the extra cord delay in the case of the crossed reflex was very 
nearly the same as, sometimes somewhat longer than, the whole cord delay 
in the case of the corresponding same-side reflex. In two out of the four 
[Exps. 4, 42] it was longer the first time the effect appeared than it was 
in any of the subsequent responses, and this notwithstanding the fact that 
in all of them stronger stimuli were used (and had been obliged to be used) 
to get the effect at all at the beginning. 

In one other experiment [No. 53], made on a preparation which had, so 
far as external appearance went, completel}'^ recovered from the effects of a 




Fig. 6. — Electrical responses of the gastrocnemius of a 
large frog which had been injected the day before 
with 2 minims 0*02 per cent, liquor strychniae, and 
had apparently quite recovered (Exp. 53). 

A, sixth response obtained when the intact sciatic nerve of the 
same side was ex ited. [Time lines 810 per second.] 
B, eighth resiimse ol.iiinLMl when the sciatic nerve of 
tlie opposite side was excited. (Time lines 815 per 
second.] 

dose of strychnine so small (in proportion to the size of the frog) that these 
had never been marked, the extra reflex delay in the one case was about 
equal to the whole delay in the other. Although it could be obtained the 
first time the sciatic of the oppo.site side was excited, the crossed-reflex 
effect was extremely feeble, and in this case it remained so throughout the 
experiment. It was strongest the eighth and last time it was recorded; 
but fig. 6, B, shows how weak it was even then. The same-side reflex 
responses in this experiment, on the other hand, the record of the sixth of 
which (taken immediately after that of the eighth crossed reflex) is repro- 
duced, had been throughout about as strong as the direct effect (which was 
again double in all the records taken of it). There was a corresponding 
diflference in the strength of the mechanical responses. 

The cord delay in the crossed reflex was about double what it was in 
the simpler reflex in two other preparations [Nos. 40 and 55], both of which 
were well strychnised, and in both of which the reflex response could be 
produced as readily, i.e. at the start, and to a stimulus of the same strength 

VOL. I. — JAN. 1908. 4 



50 



Buchanan 



as the simpler one. Records of the first response to excitation of each 
nerve in turn in the two preparations are reproduced in figs. 7 and 11 
respectively. The muscle used for recording in Exp. 40 was again one 
of those in which the direct effect to excitation of the motor part of the 
nerve was double ; the same-limb reflex time was longer in the last two 
responses than in the others, and the whole crossed-reflex time also became 
longer in the third and fourth responses, so that the extra reflex time at 
first remained unchanged ; it finally in the last response became shorter. 
Exp. 55 will be referred to again (p. 53). 

In all the other experiments the extra reflex delay was shorter through- 
out than the same-side reflex time in the same preparation, although it 
was as a rule no shorter than this time frequently is. In three of these 
[Nos. 18, 45, and 47] the action of the drug was incipient. In Exp. 18 the 




Fk;. 7. — Electrical responses of the gastrocnemius of a frog which had been 
injected the day before with 1 minim 0'04 per cent, liquor strychniae. The 
jireparation was abnormally excitable, but the limbs were no longer at all 
stitt'(Exp. 40). 

A, first response obtained when the intact sciatic nerve of the same side was excited. The 
direct effect was double. [Time lines 745 per second.] B, first response obtained wlien 
the sciatic nerve of tlie opposite side was excited. [Time lines 745 per second.] 

same-side reflex delay was exceptionally long ; the effects were weak, and 
resembled one another in both kinds of reflexes. In Exp. 45 the crossed- 
reflex effect was throughout weak, and at first a good deal weaker than the 
same-side reflex effect ; but as this became weaker each time it was recorded, 
they were in the end not unlike one another in strength. Fig. 8 allows of 
the comparison of the two kinds of reflex responses at the beginning and the 
end of the experiment. In Exp. 47 the crossed-reflex effect was very weak 
the first time it was recorded, but became in subsequent responses of nearly 
the same strength as the same-side reflex eftect ; it varied, however, some- 
what in strength as well as in duration in the different responses. Records 
of the third same-side and of the fifth crossed-reflex responses are repro- 
duced in fig. 9. This was the only preparation, besides No. 33, referred 
to on p. 33, which gave an extra reflex delay decidedly shorter than the 
same-side reflex time, as we have now learnt to know it, in strychnine 
preparations. To begin with, it was about llo", but became in the fifth and 



Transmission-time of Reflexes in Spinal Cord of Frog 



51 



sixth responses as short as 6cr. The records of these two responses show — 
but it is much more evident in that of the sixth than in that of the fifth 
(the one reproduced) — that the central stimulus was not exerting its full 
strength at the start. Three of the crossed-reflex responses, and two of 
those of the same side, show, as those of no other preparation in so early 
a stage of strychnine poisoning do, that a second stimulus of central 
origin was affecting the muscle some 50-40(t later than the first began to 
aflfect it. This indicates that the drug was taking effect rapidly. In two 
of the responses in another preparation [Exp. 49] the extra delay became 
as short as da- ', but this is hardly so short as to be called exceptional as 





Fig. 8. — Electrical nsponses of the second gastro- 
cnemius of a frog which, after records had 
been taken with the other gastrocnemius, 
before the injection of strychnine, was then 
injected with 5 minims 0"005 per cent, 
liquor strychnise (Exp. 45). 

A and C, first and fourth responses obtained when the 
intact sciatic nerve of the same side whs excited. B 
and D, first and fifth responses obtained when the 
sciatic nerve of the opposite side was excited. 
[Time lines 840 per second in all four.] 

compared with the whole delay in the same-side reflex, which was once in 
one strychnine preparation (Exp. 14, see p. 21) equally short. 

The extra cord delay in all three preparations [Nos. 50, 52, 55], which 
were in the attitude characteristic of strychnine poisoning when the records 
were taken, and in which the heart as well as the cord was aflected by the 
drug, was long, although it was only in Exp. 55 that it was as long as the 
whole same-side reflex time. With all these preparations the crossed-reflex 
effect was stronger at the beginning in some of the responses than the 
same-side reflex eflect. This may be seen, for instance, by comparison of 
the records of the two kinds of response in Exp. 52, reproduced in fig. 10. 
In Exp. 50, which is not introduced into the table, the extra delays in 
responses recorded alternately with the same-side responses referred to on 
p. 19, were, in order, 85cr, 20o-, 19cr, 14(r. 



52 



Buchanan 



The extra cord delay in the case of the crossed-reflex response may, as 
will have been noticed, vary a good deal in preparations in which the 
cord delay .in the case of the same-side reflex response is keeping fairly 



7^7CN/vVS,'VVVV%A/^s^V,^/v~'^^^'~~ 


H-i 


s 


rf 


L 









Fig. 9. — Electrical responses of the gastrocnemius 
of a frog which had been injected ^ hour 
before being prepared with 2 minims 01 
])er cent, liquor strychniae (Exj). 47). 

A, third response obtained when the intact sciatic nerve 
of tlie same side was excited. [Time lines 820 per 
second.] (A time mariner giving 400 d. v. per second 
was being used this day, and its record is shown at 
the top of the pliotograph.) B, fifth response when 
the sciatic nerve of the opposite side was excited. 
[Time lines 820 per second.] 

constant. This is especially well illustrated by Exps. 47 and 49. On the 
other hand, the extra cord delay may remain very fairly constant in pre- 
parations in which the same-side reflex delay is getting either longer [Exp. 
40] or shorter [Exps. 16, 42, and 52]. 



I 




Fig. 10. — Electrical responses of the gastrocnemius of a small frog 
which had been injected 2J hours before being prepared with 
2 minims 0'02 per cent, liquor strychniae, and which has 
been in the attitude characteristic of strychnine poisoning for 
half an hour (Exp. 52). 

A, third response obtained when the intact sciatic nerve of the same side 
was excited. [Time lines 830 per second.] B, third response obtained 
when the sciatic nerve of the opposite side was excited. [Time lines 
830 per second.] 

The extra delay does not vary with the strength of the effect 
produced in the muscle, i.e. with that which represents the strength 
of the physiological stimulus coming from the cord. Thus, both in Exp. 



Transmission-time of Reflexes in Spinal Cord of Frog 53 

45 (fig. 8) and Exp. 53 (fig. 6) the crossed-reflex eflect is very weak 
throughout the experiment; yet in the one [Exp. 53] the extra delay is 
approximately equal to the whole same-side reflex delay, whereas in the 
other [Exp. 45] it is a good deal shorter. Again, in Exp. 55, throughout 
which there was considerable variation in the length of the delay before 
the muscle began to respond, and in both kinds of reflexes, the ultimate 
strength of the effect remained about the same (see fig. 11). On the other 
hand, when the crossed-reflex effect is variable in strength during an ex- 
periment [Exp. 47], the extra delay, although it also may vary, does not 
vary with the strength of the effect. 

Nor does the extra delay in the case of the crossed-reflex response vary, 
any more than the same-side reflex time does, inversely with the strength of 
the artificial stimulus evoking it: that is to say, it does not do so when 
this is once fully above the threshold value, whatever it may do when it is 
very close to it. Thus, in Exp. 47, in which the stimulus was being gradually 
weakened, the extra delay was getting shorter, not longer ; and in Exp. 56 R, 
the extra delay remained very fairly constant, notwithstanding a good deal 
of variation in the strength of the stimulus, only in the last response, one 
to an extremely weak stimulus, apparently becoming considerably longer. 
(As, however, no. response could be obtained with this muscle when the 
sciatic of its own side was excited by an equally weak stimulus, we do not 
really know how much should be deducted to find the extra delay in the 
case of the crossed reflex ; moreover, the effect in this last response was 
so weak that, for a reason which will appear immediately, I think its 
significance for giving information about cord delay at all is doubtful.) 
The same preparation possibly gave direct indication of lengthening of 
extra cord delay with the weakening of the artificial stimulus when the left 
gastrocnemius was recording ; for, as with the whole delay in the same-limb 
reflex, only to a more marked degree, the extra delay was in this case longest 
in the one response recorded with weakest stimulus, but here at any rate 
this must have been very near its threshold value (see note to experiment). 

To obtain more information regarding the influence of stimuli very close 
to threshold strength we must again, as with the same-limb reflex, have 
recourse to observing what happens in preparations in which there is 
reason to believe that the threshold resistance of the individual synapses 
is variable, when these are attacked by the impulses, presumably of equal 
strength, produced by a single stimulus. So far as Exp. 55 is concerned, 
all that I have said on p. 25 with regard to the same-limb reflex delay 
would apply to the crossed-reflex delay, and it is perhaps significant that 
a long extra delay in attaining the maximal effect was more frequently 
observed when the stimulus was applied to the nerve of the crossed side 
than when it was to that of the same side. Thus there were three other 
records closely resembling fig. 11, D, and none other resembling fig. 11, C. 
The evasiveness of the threshold under the influence of strychnine is well 
seen in most of the preparations in which the drug was only beginning to 



54 



Buchanan 



take effect, and in which the (^rossed-retlex response was coming into 
existence under our eyes, as it were. Thus, in Exps. 42, 47, and 48, the hrst 
time it could be obtained it was only to a much stronger stimulus than 
such as afterwards sufficed to produce it. In two of these [Nos. 42 and 48] 
the extra cord delay was longer in the first response recorded than it was 
subsequently, the stimulus, although stronger, being nearer to the threshold 
value than weaker ones were later. This seems to me to be the best piece 
of evidence we have at present that strength of stimulus, when near its 




Fig. 11. — Electrical responses of the gastrocnemius of a frog which had been 
injected the day before with 2 minims 0-02 percent, liquor strychnise, and 
which was in the attitude characteristic of strychnine poisoning when it was 
prepared (Exp. 55). 

A ami C, first and eighth responses respectively obtained when the intact sciatic nerve 
of the same side was excited. [Time lines 840 and 770 per cent, respectively.] B and 
D, first and tenth responses respectively obtained when the sciatic nerve of the 
opposite side was excited. (Time lines 830 and 820 per second respectively.] 

threshold value, is likely, in the normal animal, to affect synapse time, since 

the cords in these two experiments were in less abnormal condition than 

was that used in Exp. 55.^ 

The effect of altering the temperature of the cord on the extra cord 

^ Since this paper was first prepared for the press I have made a good many more 
experiments in which the strengtli of the stimulus was altered. The results amply confirm 
the conclusion already come to, namely, that, increasing the strength of the stimulus when 
once its value is above a certain amount does not shorten the time taken to pass indi- 
vidual synapses, although it may increase the number crossed abreast. They will be dealt 
with in a future communication, together with the results obtained in more co'mplex reflexes 
in which cord delay, and possibly the number of synapses crossed in series, appears to be 
less constant in any one preparation than it is in the two which are here considered. 



Transmission-time of Reflexes in Spinal Cord of Frog 55 

delay can only be studied in one of the experiments [No. 48], since it was 
only in this one (of the two which had the temperature altered) that a 
crossed-reflex could be recorded, before, as well as after, the cooling. As 
may be seen from the table, not only was the same-side reflex delay 
lengthened by cooling in this experiment, but the difference between the 
two cord delays was likewise lengthened, showing that the temperature 
of the cord affects the extra cord delay in the same way as it affects the 
same-limb reflex time.^ 

We have, therefore, in the case of the crossed reflex, a loss of time 
additional to that obtaining in the same-limb reflex, which is of about the 
same order, and which varies, or does not vary (as the case may be), in the 
same way with external physical conditions ; but which, when these are 
constant, may vary independently of this same-limb reflex cord delay. 

These facts suggest very strongly that there is interposed in the 
conductive path, in the case of the crossed reflex, some element 
of the same nature as the one which is alone interposed in the 
path of the same-limb reflex. In other words, one can hardly 
refrain from inferring that, whereas in the same-limb reflex 
a single synapse has, in the case of each fibre, to be passed; in 
the crossed-limb reflex two such synapses have to be passed in 
turn by every individual fibre concerned. 

The one set of synapses in the case of the crossed reflex would be the 
same, or would belong to the same set, as those which are alone concerned 
in the simpler reflex. They may, therefore, be called the primary synapses. 
The principal cells concerned in them would be presumably the motor cells 
of the ventral cornua. The second set involved in the case of the crossed 
reflex may be called secondary synapses. They would lie on the opposite 
side of the cord, and in view of the most interesting experiments of 
Baglioni with regard to the action of strychnine on the cord, the locating 
of them somewhere in the dorsal part of the cord suggests itself. 

Whether the primary synapses involved are individually the same, and 
necessarily so, in the two kinds of reflexes, is another question to which 
my records suggest an answer. Certain experiments show in the records 
a very striking difference in the muscle response in the two kinds of reflexes. 
The difference in Exp. 53 (in which all the six same-side responses very 
closely resembled the one of them reproduced in fig. 6, A, and all the eight 
crossed-reflex responses the one of them reproduced in fig. 6, B) was chiefly 
one in strength, and might either mean that each primary synapse cell was 
giving a weaker discharge when the impulses arrived by way of the secondary 
synapses, or that a smaller number of them were brought into play. But 
in other experiments, e.g. Nos. 55 and 5G R (in both of which especially 
weak exciting stimuli were being employed), the two kinds of reflex 
responses were equally strong, and yet exhibited characteristic diflerences. 
To illustrate this, I have had reproduced in fig. 12 the records of four 
> I have now made several experiments wliicli confirm tliis [Nov. 1907]. 



56 



Buchanan 



successive responses of the same spots of the right gastrocnemius, obtained 
by the alternate stimulation of the right and the left sciatic in Exp. 56. 
No one can fail to see (and it is not only in these four records that the 
phenomenon is to be seen) that the undulations on the curves which 
represent the crossed-reflex response have a character of their own, different 




Fig. 12. — Electrical responses of the gastrocnemius of a frog 
which had been injected with 2^ minims 0*02 per cent, 
liquor strychniae two days before, and which had been in 
the attitude characteristic of strychnine poisoning for some 
time, but in which the limbs were no longer more than 
somewhat stiff (Exp. 56 R). 

A, C, and E, first, second, and seventh responses respectively obtained 
when the intact sciatic nerve of the same side was excited. [Time 
lines 835, 830, and 827 per second respectively.] B and D, first and 
second responses obtained respectively when the sciatic nerve 
of the opposite side was excited. [Time lines 830 per second in 
both cases.] 



from that of those on the curves representing the same-limb reflex response. 
I have elsewhere shown (loc. cit.), and I have still better evidence than any 
that has yet been published (namely, the records themselves) of the fact 
that these undulations represent something for which the muscle alone is 
responsible. The difference between the records, therefore, suggests that not 
all the same muscle fibres, and consequently not all the same efferent nerve 



Transmission-time of Reflexes in Spinal Cord of Frog 57 

fibres, and not all the same primary synapse cells were in play in the two 
kinds of reflex response. In Exp. 55 all the records of the same-limb reflex 
responses referred to with a f in the table had at the beginning the character 
exhibited in the first of them (the one reproduced in fig. 11, A), while the 
crossed-reflex responses more usually began in the way shown in fig. 11, B. 

At first sight the suggestion that some at any rate of the primary 
synapses concerned may be different in the two kinds of reflexes may seem 
to be incompatible with a suggestion I made before (p. 27) with regard to 
the same-limb reflex response in the normal cord. It is, however, quite con- 
ceivable that all the primary synapses and their cells, controlling the 
particular muscle, may be brought into play when either sciatic alone is 
stimulated by a strong induction shock, and yet that some come into 
play more readily when the afferent fibres of the one nerve are excited, 
others when those of the other are excited, and that such differentiation 
takes place, though it may not always be manifest, when the exciting 
currents are weak (even when strong enough to excite all the afferent fibres 
of each nerve), or when certain synapses have failed to become, or to 
remain, sufficiently responsive. The records obtained in Exp. 55, showino-, 
as they do (see fig. 11), that neither reflex effect was ever so strong as 
the direct, would lend support to such a view. The records obtained in 
Exp. 56 might, however, at first sight seem to make it untenable, in so far 
as they show that the reflex effects were, to begin with (fig. 12), about as 
strong as the direct effect. But such a fact does not necessarily always 
imply that all the muscle fibres have been excited by the stimulus coming 
from the cord when this has had its excitability increased by strychnine. 
The increased discharge which this drug enables the motor cells to give may 
quite well so increase the strength of the eflect in the muscle fibres as 
to obscure a small reduction in the number of fibres which are in plaj'-, 
as compared with the number brought into play in the direct response. 
That this is sometimes actually the case is shown, I think, by records 
obtained in a few of my experiments (in 56 R amongst others) of same- 
limb reflexes in which the direct effect was less strong than the reflex 
effect. Occasionally it was so when the experiment began, as in the one 
[Exp. 45] to which fig. 8, A, refers ; the curve derived from this record 
would certainly show that the difference of potential when it first came 
into existence in each response was somewhat greater in the case of the 
reflex than in that of the direct response. Usually, as was the case in 
Exp. 56 R, the reflex effect only became the stronger by the direct effect 
becoming weaker (compare, in fig. 12, E with A or C). Since the one 
muscle effect remained strong, the weakening of the other is likely to 
have been due to some difference in the part producing the excitation, 
i.e. in this instance in the motor part of the nerve, and the difference 
seems to me to be most probably one in the number of fibres excited, 
some of them having become in some way temporarily or even per- 



58 Buchanan 

manently impaired.^ Thus, at the end, in such an experiment as 56 R, 
the whole reflex discharge was producing a stronger electrical efltect in 
the muscle than the artificial stimulus to the nerve could produce, whether 
or not more fibres were in play, and in neither case (probably) all being 
in play. As Exp. 45 went on, the reflex eflect as well as the direct 
efiect got weaker, as may be seen by comparing curves C and A in 
fio". 8, the eflfect of the strychnine, if I am right in my interpretation, not 
being great enough in this preparation to overcompensate for a reduction 
in the number of efferent nerve fibres, and consequently of muscle fibres, 
which were in play.'^ 

The comparison of the crossed-reflex responses, among themselves even, 
in Exp. 55 suggests a further differentiation of synapses belonging to one 
and the same set ; only it appears to me more probable that the set differ- 
entiated in this case was that of the secondary synapses, and that the 
agent which differentiated them was fatigue. A glance at the table on 
p. 45 shows how much the extra delay in the case of the crossed reflex 
was reduced by an extra few minutes' rest each time this was allowed. 
The examination of the records brings out certain peculiarities in the four 
responses referred to by the ff, one of which is reproduced (fig. 11, D), 
which are not presented by the other crossed-reflex responses. The fact 
that in all of these the total cord delay was almost the same, and 10-13a- 
longer than it was in most of the other crossed responses, suggests a further 
inference, the discussion of which, I will, for the present, postpone. 

Results such as these, and others obtained in the few experiments which 
have hitherto been made with stimuli not far from the threshold value in 
strength, seem to me to indicate that it will be more fruitful for the further 
investigation of processes occurring within the central nervous system to 
continue to use weak stimuli rather than strong ones. At the same time 
it must not be forgotten that the experiment in which stimuli nearest to 
the threshold value were used [Exp. 55] was one of those in which the 
circulation was not quite normal, and that some of the peculiarities in the 
responses may be due to the effect which defective circulation has been 
shown to have in lengthening cord delay, this being probably greater (see 
p. 26) on some of the synapses than on others. 

We have seen that one of the first actions of strychnine on the cord 
delay, in the case of the same-limb reflex (on the time taken, if we may 
so consider it, to pass the primary synapse), is to somewhat shorten it. 

1 In those cases in which the direct effect was weakened by a change in the direction 
of the current (see p. 16), the reflex effect again usually retained its full strength. 

2 Another interpretation of the regularly recurring difference occasionally met with 
between the two kinds of reflex eft'ects in the muscle would have suggested itself if the 
fibres of the gastrocnemius were supplied, as are those of the sartorius and certain other 
muscles of the frog, by more than one efferent nerve fibre ; but Sandemann (A. f. 
(Anat. u.) Physiol., 1885, p. 246) has shown that this is not the case, or rather he has 
shown that each filare of the frog's gastrocnemius is provided with but one motor nerve 
end-organ, which makes it improbable that it should be the case. 



Transmission-time of Reflexes in Spinal Cord of Frog 59 

This action appears to be far stronger on the factor which determines 
the extra delay in the case of the crossed reflex, i.e. on the time taken 
to pass the secondary synapse. This time may apparently be ver}- 
considerably shortened by it. In two preparations, as has been seen, in 
both of which the drug was taking eflect very quickly, it was reduced in 
the course of the experiment to bu and 6cr respectively. According to 
Wundt, it may sometimes be reduced to as little as 4(t. Are we to infer 
from this, and from the fact that the only time a crossed-reflex response 
was obtained in a preparation which was hardly influenced at all by the 
drug [Exp. 38], the delay was as much as 63cr longer than in the cor- 
responding same-limb reflex response, that the time taken in passing the 
secondary synapse is in the normal cord a good deal longer than it usually 
is in the strychnised cord ? I think not, although, if ever I succeed in 
obtaining and recording the crossed reflex in response to a single in- 
stantaneous stimulus in a normal preparation, I should not be at all 
surprised to And that the delay in the manifestation of the effect in the 
muscle was still longer.^ With regard to its long duration in normal or 
very nearly normal preparations, I would suggest that the same sort of 
explanation may apply as applies to the postponement of the moment at 
which the eflect becomes maximal in the case of some of the responses to 
just efficient stimuli in fully strychnised preparations. If we compare, for 
example, the maximal strength of the reflex effect attained in the responses 
to which fig. 11, C and D, refer, with what it was when it started in these 
responses, and then compare the maximal strength of the reflex effect in a 
normal cord as it usually is, and as it was in Exp. 38 (steepness of rise about 
the same as in the record reproduced in fig. 1, B), with something less than 
it, in the same proportion (or in the proportion which the crossed-reflex eff'ect 
sometimes bears to the uncrossed in strychnine preparations, e.g. fig. 6), 
it is evident that this something, if it existed, would be imperceptible with 
the recording instrument and arrangements which were being used. 
Whatever is the explanation of the postponement of the moment at which 
the eflect reaches its maximum in the one case, seems to me likelj^ to be 
the explanation of the late appearance of any effect at all in the other case. 
The postponement in the strychnine preparations, since it may appear 
in the response of the uncrossed as well as in those of the crossed reflex, is 
likely to be due, in part and sometimes entirely, to some factor in the set 
of primary synapses preventing the appearance of the full eflect at once. 
I have already suggested (p. 27) that this factor is the want of accord in 
the moment of dischai'ge of the separate cells, brought about by impair- 

1 Since this paper was sent in I have at last succeeded in obtaining in two much 
cooled but undrugged preparations a trui- crossed-reHex response in the biceps femoris. 
Although the records of it. were extremely minute, the moment at which it began could 
be determined. In one preparation the effect appeared after an interval which was twice 
as long as in the same-limb reflex ; in the other, after an interval which was, both relatively 
and absolutely, considerably longer. The cord delay in the same-limb reflex was very long 
(about 50(r) in each preparation. [Nov. 1907.] 



()0 Buchanan 

rnent of circulation, and variatiorts in the threshold resistances of individual 
synapses, some of them being forced and not others, or others not until 
later. 

The late appearance of a crossed-reflex effect, if it appears at all, in the 
normal preparation, seems to me, on the other hand, to be much more 
probably due to some factor in the set of secondary synapses preventing its 
earlier appearance, since, while the same-limb reflex response is com- 
paratively easy to evoke in the normal cord (it was obtained in about 70 
per cent, of my cooled preparations), the crossed-reflex response is rarely, if 
ever, evokable from it by a single instantaneous stimulus to the aflerent 
nerve fibres, and even in the strychnine cord the crossed-reflex response is 
so often not to be evoked when that of the same side is showing in its record 
the eflects of the drug. It is clear that in such case something else has to 
happen before the extra excitability thus shown to have been brought about 
in the primary synapses or their cells, can be taken advantage of by the 
impulses which may come to them from the opposite side of the cord. If 
these arrive at all at the primary synapses in the normal, or in the lightly 
strychnised cord, they behave like impulses produced by artificial stimuli 
too weak to be effectual, applied to the same-side aflerent nerve fibres. It is 
only when the collective discharge from the cells of the secondary synapses 
has reached, and passed, its threshold value with regard to the primary 
synapses and their cells, that it is able to produce an eflfect in the muscle 
supplied by these. Yet we have seen that strychnine may, when it has 
acted for a longer time (insufiicient, however, to make the electrical effect 
serial or the mechanical one a spasm), have a considerable influence on that 
factor in the secondary synapse which determines the delay in it, and it is 
most unlikely that what it influences so strongly, it influences less soon 
than what it influences less strongly, namely, the delay-regulating factor 
in the primary synapse. Just for this reason, and for the one that other 
factors also, such as fatigue, seem to have greater influence on secondary 
than on primary synapse time, I would suggest that it is the individual 
variability of the time taken in passing the secondary synapses (the 
variability in their threshold resistances perhaps) in the normal cord which, 
by preventing the discharges from being synchronous, prevents them from 
affecting the primary synapse cells and hence the muscle. When, either by 
making the excitability of the secondary synapse cells more alike, or by 
making the resistances of the secondary'' synapses more equal, the separate 
discharges become not only stronger, but more synchronous, they are able to 
effect what even discharges of the same strength could not do, one by 
one, with broken step (if one may use the expression). To make effectively 
to work together things the temporary capacities of which vary, would 
take a longer time for the drug to accomplish than when it has to 
deal with things the capacities (Fahigkeiten) of which vary but little, 
as in the case of the primary synapses. The fact that the first time 
a crossed-reflex eflect appears after failures to obtain it, in any particular 



Transmission-time of Reflexes in Spinal Cord of Frog 61 

preparation, it is almost always very weak, even when it subsequently 
becomes strong, lends support to this view. If it is correct, it implies 
that the probable cord delay (or extra cord delay, as the case may be) 
given for each response in my tables represents the time taken to pass 
the primary (or secondary) synapse by a sufficient number of individual 
impulses at the same moment, to produce an effect on the muscle. They 
therefore give the shortest time taken by a sufficient number of them 
to pass simultaneously. When the effect is its strongest, as it usually 
is, almost as soon as it begins, the probable cord delay given is the time 
taken by the large majority, if not all, of the impulses arriving through their 
several afferent channels to pass the synapse in that response. If any have 
passed it more quickly, they were too few in number to produce an effect in 
the muscle. Those that passed their synapses more slowly would only 
produce an additional effect if a sufficient number of them do it in the same 
time, and this apparently was only sometimes the case. 

That strychnine actually does make the time taken to pass the 
individual secondary synapses get shorter, and does not only serve to 
make them work more synchronously, seems to me to be attested by the 
measurements of records in several individual experiments. If we may 
express it in terms of what I have just ventured to set forth, these show 
that the time taken by the greater number, the modal time, to borrow Pro- 
fessor Karl Pearson's word,^gets shorter, not longer, as the drug becomes 
more effective, and it would be most unlikely that those synapses which 
can be passed in shortest time would be the last to be affected by the drug. 
I hold, therefore, that the probable extra cord delays obtained from 
records taken with strychnine preparations just so much affected 
by the drug as to give a decided and definite crossed-ref lex response 
at all, are likely to be those which would be obtained from each of 
the cords when normal, could we from such a cord evoke a measurable 
response ; and that the extra cord delays obtained from records taken at, 
or very soon after, what we may call the critical moment in the action 
of the drug, represent each the modal time (possibly also the average 
time of the whole set) taken by the secondary synapses in any particular 
normal cord. To know this modal time, the time taken to pass the greater 
number of synapses, and to know the variations from the mode in any 
particular set of synapses, seems to me far more important than to know 
the shortest time in which a synapse may be passed. 

Another reason for the conclusion that the results obtained with 
strychnine cords may be applied to normal cords is that, if the excitability 
of the cord is raised by other drugs, the delay in traversing the cord, when 
the nerve of the opposite side to the recording muscle is stimulated, is 
longer than the delay when the nerve of the same side is stimulated, by 
an amount of the same order as when strychnine was employed. This was 
so in the two experiments I have so far made with phenol preparations.- 
1 Phil. Trans., 186, Scries A, 1895, p. 345. - See footnote to p. 29. 



62 



Buchanan 



A record of one of the crossed-reflex responses taken with the first of these 
is reproduced in fig. 13, C The cord delay was 12-5cr longer than it was in 
the same-limb reflex response recorded immediately afterwards (with the 
same strength of the artificial stimulus), and reproduced in fig. 13, A. It 
was of about the same duration in two other responses, one to a stimulus 
somewhat stronger, and the other to one a good deal weaker than was 
employed when the record reproduced was taken. But there is something 



M^A^^^W■MVAv 



AA'^Af' 




Fig. 13. — Electrical responses of the gastrocnemius 
of a large frog which had been injected 
2^ hours before being prepared with 10 minims 
0*1 per cent, phenol. 

A, fourth response obtained when the intact sciatic 
nerve of the same side was excited. [Time lines 
750 per second.] B, response to excitation of the 
peripheral end of the same nerve after It had been 
divided. Record taken at the end of the experi- 
ment. [Time lines 835 per second.] C, third re- 
sponse obtained when the sciatic nerve of the 
opposite side was excited. [Time lines 750 per 
second.] 



peculiar about the extra delay when the crossed sciatic is excited by 
strong stimuli (10,000 units) in phenol frogs (besides that which has 
already been mentioned on p. 31), the discussion of which I must defer. 
My experiments should, and I hope will, when they are carried further, 
throw light upon the results obtained from the most ingenious and in- 
teresting experiments of Baglioni made for the purpose of ascertain- 
ing the action of strychnine and phenol on nerve cells in Vertebrates 
and Molluscs. I do not think, however, that they will confirm his 



Transmission-time of Reflexes in Spinal Cord of Frog 63 

conclusions that one kind of cell behaves in a diflferent way from another 
to the same drug, the one having its excitability increased greatly, the 
other remaining wholly unaflected. They rather suggest, so far as they 
have gone at present, that strychnine can increase the charge made by both 
the sets of cells ; but that in the one case and not in the other it has first 
to bring them into array, and that the difference lies not so much in the 
cells themselves as in the way in which collateral or other fibres are 
distributed to the one or the other. With regard to these matters, it seems 
to me that the electrical responses of muscle, when recorded photographi- 
cally by such an instrument as the capillary electrometer, are capable of 
giving information about the physiological stimuli which evoke them 
which can never be obtained from records of the mechanical response 
alone. 

It will give me pleasure to supply any physiologist with copies of 
photographs relating to any particular experiment in which he may be 
interested. A great deal of the value of some of the experiments lies 
in the resemblances (or differences, as the case may be) between the 
records of successive responses, and it is only samples of them, and those 
only of comparatively few experiments, that could be reproduced within 
the limits of this paper. 



All the experiments described in this communication were made with 
apparatus which belonged to Sir John Burdon-Sanderson, and on which 
he had spent a great deal of labour, time, and money in trying to make as 
complete as possible. I had already been using it in its different stages for 
some years, and any work that has yet been done or will be done with 
it may in a sense be regarded as a continuation of his work. That 
this should be continued, and that what was already done in his lifetime 
should be made use of to the fullest extent possible for suggesting new 
lines of research and for the acquirement of new knowledge, is, I feel sure, 
the tribute to his memory which he himself would most desire. 

The recording part of the apparatus was made under the direction of 
Mr H. S. Souttar, by whom it was to a considerable extent designed. 
He has also supplied me with the very quick capillary electrometers with 
which all the records have been obtained, and I am greatly indebted to him 
for so doing. I have also to thank Professor Sherrington for valuable 
criticism and for the interest he has taken in the work. 

The working expenses have been defrayed out of a grant from tlu' 
Government Grant Committee of the Royal Society. 



64 Buchanan 



SUMMARY. 



By means of photographic records taken with the capillary electrometer, 
it has been shown that : — 

1. The delay in the normal cord of the frog, in the same-limb reflex, as 
determined by the interval of time between the two electrical responses of a 
particular spot of a muscle (the gastrocnemius) when the efferent and afferent 
fibres of the mixed nerve supplying it (the sciatic) were simultaneously 
excited by a single break induction shock, varies in different preparations 
between 0012 and 0022 second. It is very rarely as short as 0"012 second, 
and only occasionally longer than 0-022 second. These numbers refer to 
cord delay alone, time for transmission along the known length of nerve 
having been deducted on the assumption that an impulse travels at the rate 
of 30 metres a second along fresh frog's nerve. 

2. When the cord and the cord alone has been acted upon by a very 
weak dose of strychnine, this delay is somewhat diminished. It then 
varies in different preparations between 0'009 and 0'020 second at room 
temperatures, though it is rarely as short as 0'009 second. It is seldom longer 
than 0'020 second, except when the circulation as well as the cord has been 
affected by the drug. In such cases the cord delay may become as long as 
0"033 second, and the central stimulus may fail to exert its full strength 
when it first begins to affect the muscle. 

3. Alteration in the strength of the artificial stimulus applied to the 
nerve does not alter the delay in either the normal or the strychnised cord. 
This statement is certainly true for stimuli above a certain strength (vary- 
ing with the sensitiveness of the preparation), but probably does not apply 
to stimuli the strength of which is only just above the threshold value. The 
difficulties in the way of determining the threshold value for any one 
preparation have so far prevented direct satisfactory investigation of the 
effects of just adequate stimuli.^ 

4. Cooling the cord by applying ice to the back of the preparation greatly 
increases the delay in the strychnised cord. Cold seems also to increase 
to some extent the delay in the normal cord. 

5. Repeated stimulation (fatigue) may lengthen the delay in the 
normal cord. 

6. The direct action of strychnine on the cord consists principally, so far 
as the same-limb reflex is concerned, in making the discharge stronger, 
and eventually longer, in response to a weaker stimulus. So far as the 
crossed reflex is concerned, it also brings about something which before 
prevented an effectual response from being obtained in the muscle to a 
stimulus so brief as a single induction shock, applied to the sciatic nerve of 
the opposite limb. 

^ A little direct evidence has, however, been obtained since this paper was written that 
cord delay, even in the same-limb reflex, is longer when the strength of the stimulus is only 
j ust adequate to produce a reflex response in the muscle. 



Transmission-time of Reflexes in Spinal Cord of Frocr 65 

7. In strychnine preparations, from which an effectual response can only 
just be obtained when the nerve of the opposite limb is stimulated, the 
cord delay is always roughly about double what it is when the nerve of the 
same limb is stimulated. 

8. Whereas the delay in the case of the same-limb reflex is only some- 
what shortened by the action of strychnine on the cord, the extra delay 
in the case of the crossed reflex may be very considerably shortened by the 
continued action of the drug, so long as its action is confined to the cord. 
It may be reduced to half, or even to a fifth of what it was, in the course 
of an experiment in which the drug was taking effect rapidly. It seems, 
however, never to be reduced to less than 0-004 second, and seldom becomes 
so short as this. It may vary in any one preparation independently of the 
whole delay in the simpler reflex, but if the circulation has been affected by 
the drug it also is long. 

9. The extra delay in the crossed reflex is no more affected by the 
strength of the stimulus applied to the nerve than is the whole delay in 
the same-limb reflex. It is affected, and in the same way, by changes in 
the temperature of the cord. 

10. The cord delays in the uncrossed and in the crossed reflex are of the 
same order when the excitability of the cord has been raised by phenol 
instead of by strychnine, and bear the same sort of relation to one another, 
provided that in the case of the crossed reflex the strength of the artificial 
stimulus is not very greatly in excess of (not more than five times) the 
strength just necessary to produce it. 

It has been inferred : — 

11. That in the same-limb reflex there is normall}' a single sj^napse 
interposed in the conductive path of each individual fibre concerned, and 
that the time taken to pass it in the normal animal probably lies between 
0-010 and 0020 second. 

12. That in the crossed reflex investigated there are normally two 
sjmapses interposed in the conductive path of each individual fibre con- 
cerned, and that the time taken to pass the additional (" secondary " ) one is 
about the same as that taken to pass the primary one in a normal animal, but 
that it is subject to greater variation in each individual cord, and that there 
is, determining it, something which is more susceptible to such agents as 
drugs and fatigue. 

An abstract of this communication, under a somewhat different title, 
appeared in the Proceedings of the Royal Society, B., vol. Ixxix., 1907, p. 503. 



VOL. I. — JAN. 1908. 



66 Transmission-time of Reflexes in Spinal Cord of Frog 



EXPLANATION OF FIGURES. 

On each positive used, an exactly vertical line was ruled from the spot in the 
record of the signal which represents the moment of excitation to the zero line of 
the curve. Subsequently (for the sake of economising space) the signal-record itself 
was cut off in all the positives, with the exception of those used for figs. 1, 2, 4, 9 A, 
and 13 A and B. The photographs have not been "touched" in any other way, 
except that in some the shaded part has been darkened for purposes of photographic 
reproduction. They are to be read from right to left, and the ruled vertical line 
is consequently to be seen on the right-hand side of each. In the photographs 
themselves the vertical lines mark the time, and their number per second is stated 
below for each. The horizontal lines, when present, mark in millimetres the height 
to which the projected image of the meniscus rose. 



(>! 



SOME COMPARISONS BETWEEN REFLEX INHIBITION AND 
REFLEX EXCITATION. By C. S. Sherrington. (From the 
Physiology Laboratory, University of Liverpool.) 

I. Grading of Intensity of Reflex. 

Opinion regarding relation between strength of the stimulus exciting a 
reflex action and the intensity of the resulting reflex is undergoing change. 
It was thought that something like the " all or nothing " rule observable 
for the relation between stimulus and response of the vertebrate myo- 
cardium held good for spinal reflex arcs. The statement still often is that 
within but a narrow range does variation of intensity of external stimulus 
aflect the intensity of the spinal response. Internal condition of the reflex 
arc does certainly enormously influence the intensity of the arc's reaction. 
But recently instances have been forthcoming to show that grading of 
reflex eftect follows closely the grading of the external stimulus/ and in 
some cases through a wide range of intensity of stimulus.^ 

There has to be remembered, in dealing with this question, the property 
of temporal summation so transcendently displayed by reflex arcs.^ A 
series of weak stimuli may by summation become more potent than a 
stimulus of much greater physical intensity, but single or of relatively few 
or infrequent repetitions. Mere duration of the stimulus comes, therefore, 
to be equivalent to intensity. A simple waj^ of eliminating this source of 
confusion is to employ as external stimulus an agent of variable intensity, 
but of duration practically infinitely brief ; its period then becomes 
negligible. A single induction shock may be regarded as furnishing such 
a stimulus. The single induction shock has, however, figured very rarely 
as the stimulus for evoking a reflex reaction. There has existed a belief 
that a single induction shock must, in order to excite a reflex reaction, be em- 
ployed in very high strength. Indeed, authorities have questioned whether a 
single induction shock can excite a reflex at all. Some reflexes are, it is 
true, extremely difficult to evoke by a single induction shock ; thus, the 
" scratch-reflex " of the spinal dog I was on no occasion able to elicit by a 

1 Merzbaclier, L., Pfliiger's Arch., Ixxxi., 1900. Langelaan, J. W., Aivh. f. Physiol., 
Suppl. Bd., 1903. Sherrington, C. S., Proc. Physiol. Soc, .March 1904; Joiirn. of 
Physiol., xxxi., p. xvii. ; Journ. of Physiol., xxxiv., p. i. ; Intograt. Action of the 
Nervous System, p. 70, 190(i. Pari, G. A., Zeitschr. f. allgeni. Physiol., iv., 1904 ; Arch, 
italiennes de biolog., 42 ; Atti. Instit. Veneto, 65, 1906. Baglioni, S., Analyse d. 
Reflex-funktion., 1907. 

2 Sherrington, C. S., op. cit. ^ Stirling, W,, Ludwig's Arbeiten, 1874, p. 245. 



€8 Sherrington 

sino-le shock. But in various other mammalian reflexes this is not the case, 
and the " flexion-reflex " of the limb is elici table by a single induction shock, 
either make or break, and of such slight intensity as to be imperceptible to 
the tongue.^ 

With the single induction shock as stimulus, therefore, and with the 
" flexion-reflex " as reaction, observations can be made on the relation 
between intensity of stimulus and intensity of reflex response. For 
obtaining comparable break shocks in the following observations, the 
opening of the primary circuit has been operated by a pendulum. To 
obviate changes in resistance, a box of 100,000 ohms has been employed in 
the secondary circuit. The "flexion-reflex" can be readily excited by a 
break shock, whether applied to an afferent nerve of the limb or to some 
point of skin in the " receptive field " of the reflex. For the following 
observations it seemed preferable to apply the stimulus directly to the 
afferent nerve rather than to the skin. It is true that, as Baglioni^ has 
pointed out for the frog, application to the nerve is probably not so 
favourable as application to the skin for the obtaining of the full amount 
of grading of the reflex. On the other hand, however, by applying the 
external stimulus to the aflferent nerve direct, the unknown factors entering 
into the reaction are somewhat reduced, a desideratum where so many 
variables are perforce included. The afferent nerve to which the stimulus 
was applied remained the same for all the experiments; it was the 
musculo-cutaneous branch of the peroneal taken about 4 centimetres below 
the knee. The electrodes were silver pins placed on each side of the nerve 
5 mm. distant along its length. The direction of the current in all the 
observations was the same. 

The " intensity " of the reflex reaction has several forms of expression.^ 
The reflex reaction as it increases in intensity tends to involve an increasing 
number of muscles. It also involves with greater intensity the several 
muscles individually."* The greater reflex movement of the limb which dis- 
tinguishes a stronger reflex effect from a weaker is probably usually due to 
both these factors.^ In the following observations the increase of intensity 
of reflex reaction has been examined as it occurs in the individual muscle. 
The question of irradiation of the reflex discharge over a wider or narrower 
field of musculature has not here been entered on. In regard to any such 
grading of intensity of action as might be found in the individual muscle, 
the inquiry had in view its comparison in the excitatory and inhibitory 
sides of the reflex respectively. The " flexion-reflex " ^ is a reflex of 
simultaneous double sign (± reflex); flexors of hip, knee, and ankle, and 

1 Sherrington, C. S., Proc. Eoy. Sue, vol. Ixxvi. B, p. 270, 1905. 

2 Baglioni, S., ibid. "^ Sherrington, C. S., Integrat. Action etc. 

■* Sherrington, C. S., Journ. of Physiol., loc. cit. ; Langendorff, 0., Nagel's Hand- 
buch d. PhysioL, iv. i., footnote, p. 240, 1904. 

s Sherrington, C. S., Integrat. Action etc. 

" Sherrington, C. S., Proc. Roy Soc, loc. cit. ; Integrat. Action etc., p. 83 ; Ergebn. 
d. Physiol, 1905, p. 834. 



Comparisons between Reflex Inliibition and Reflex Excitation 



69 



abductors and internal rotators of hip express the excitatory ( + ) side of 
the reflex, by contracting ; extensors of hip, knee, and ankle, and adductors 
and external rotators of hip express the inhibitory ( — ) side of the reflex 
by relaxing. To serve as samples of these opposed groups reacting under 
the reciprocal reflex innervation, two muscles were chosen, which act on 
one and the same joint, but oppositely. These were semitendinosus as 
flexor of knee, vasto-crureus as extensor of knee. The reflex preparation 
was made as in previously reported observations,^ deep chloroform narcosis 
being employed until after the destruction of the brain. 

The reflexes obtained from the isolated semitendinosus show clear 
grading of intensity of contraction following grading of intensity of the 
break shock, figs. 1 and 2.- The grading, even under experimental 
conditions, is sufficiently delicate and covers a sufficient range of gradation 
of stimulus to indicate that under natural conditions strength of stimulus 
must influence minutely and widely the extent and force of the actual 




Fig. 1. 



movement in the flexion-reflex of the limb. This result is welcome, because 
it conforms with what would be expected from the standpoint of teleology. 
In the examples figured (figs. 1 and 2), the sequence of change of intensity 
of the external stimulus has been from weaker to stronger. The grading 
of the reflex response has, however, been equally obvious when the direction 
of the sequence has been reversed. A feature noticeable in the myograms 
is that not only does the amplitude of the contraction increase with 
increased strength of stimulus, but also the duration of the contraction 
increases. There is marked persistence of contraction in the stronger 
reflexes, recalling the " after-discharge " ^ of strong reflexes excited by longer 
stimuli. An advantage in using tlie single induction shock as a stimulus 
is that fatigue tends to occur hardly at all. The interval between the 
successive reflexes in the examples shown was one minute, but even with 
much shorter intervals no evidence of fatigue was obvious. Care has to be 

^ Proc. Roy. Soc, loc cit. 

2 All the figures read from left lo right ; in all the time record i.s marked in fifths of 
seconds. 

3 Sherrington, C. S., Journ. of Physiol., loc. cit. ; Integral. Action etc. 



70 



Sherrington 



taken, however, lest the interval be so short as to allow facilitation 
(Bahnung) of a reflex by its immediate predecessor. This I have seen 




occur even across an interval of 15 seconds (%. 3). Bahnung is 
particularly likely to occur when a number of reflexes are excited in fairlv 



Comparisons between Reflex Inhibition and Reflex Excitation 71 

quick succession from a preparation that has previously lain quiescent for 
some length of time. The staircase phenomenon seen in skeletal and 
cardiac muscle may then make its appearance in the reflexes (fig. 3).^ 

Turning to the inhibitory side of the reflex, the reactions exhibit there 
a grading not less delicate and extensive than those of the excitatory side. 
With increase in the strength of the stimulus (break shock), the amplitude 
of the reflex relaxation increases (fig. 4). Also the speed of the relaxation 
becomes greater (fig. 4). Increase in speed of progress of the relaxation 
is particularly evident when, instead of a single momentary stimulus, the 
stimulus used is faradic. Fig. 5, A B C, illustrates this. For each of the 
three reflexes in the figure the stimulus consisted of a series of double 
shocks obtained by an iuterruptor vibrating 40 per second in the primary 
circuit. In the reflex of fig. 5, A, the secondary coil stood at 40 units on 
the Kronecker scale ; in reflex B at 100 units ; in reflex C at 500 units. The 




reflex relaxation is not only greater in B than in A, but its progress is more 
rapid ; and in C than in B. This greater speed of progress of the relaxa- 
tion under stronger stimuli finds a close counterpart in the greater speed 
of augmentation of contraction in the excitatory reflex under stronger 
stimuli.^ It is a point of likeness between the inhibitory and the excita- 
tory sides of this "flexion-reflex." The relaxation in reflex C (fig. 7) is 
followed, after a latency longer than that of the initial reflex inhibition, 
by the contraction termed the "rebound-contraction,"^ ascribable to "suc- 
cessive spinal induction." * That this occurs in C and not in B or A is in 
accord with the description of the phenomenon given elsewhere,^ and with 
the explanation there oflered. 

When the inhibitory relaxation is weak it is, in my experience, not 
rarely accompanied by tremor. This is seen in fig. 5, A and B, ; also in the 

1 Cf. Stirling, W. (in the frog), op. cit. 

2 Sherrington, C. S., Proc. Roy. Soc, Ixxvi. B, pp. 272, 274. 

3 Sherrington, C. S., Proc. Rov. Soc, vol. Ixxvi. B, p. 160, 1905 ; ibid., Ixxvii. B, 
p. 478, 1906 ; ibid, Ixxix. B, p. 347, 1907 ; ibid., Ixxx. Integrat. Action etc., ]k 206. 

* Ibid. 5 Ibid. 



72 



Sherri no-ton 



weaker reflexes in fig 4. The tremor is sometimes more marked than 
in those examples, as in fi(,^ 6. In the reflex of fig. 6 the stimulus was 




Fig. 4. 

faradic and of little more than threshold value ; the inhibitory relaxation 
is correspondingly weak, and when it occurs the muscle exhibits tremor, 
and this outlasts considerably the period of application of the inhibitory 



J 



Comparisons between Reflex Inhibition and Keflex Excitation 73 




74 



Sherrington 



stimulus. Weak reflex relaxations come thus to have a hesitant character, 
very obvious even to simple inspection. This hesitant character constitutes 
a point of likeness between weak inhibitory and weak excitatory reflexes ; 
examples of the latter have been furnished elsewhere.^ 

The myogram of the inhibitory reflex differs from that of an excitatory 
reflex in the former's showing little or no recovery by the muscle (under 
the conditions of experiment) of the length it had prior to initiation of the 
reflex. Subsequent to the inhibition-reflex the muscle continues to remain 
of the new length it assumed under the relaxation due to the reflex- (fig. 4). 
After an excitatory reflex the muscle fairly quickly resumes the length it 
had before its reflex contraction. 

The amplitude of the reflex relaxation caused by a single induction 
shock, although showing regular increase, within certain limits, with in- 




FlG. 



crease in the strength of the momentary stimulus, does not in my experi- 
ence reach, even under the strongest of such momentary stimuli, the extent 
it attains when, instead of a single induction shock, a short series is delivered, 
as in faradic stimulation (fig. 7, A and B). The extent of the reflex relaxa- 
tion which a faradic stimulus, even far below maximal, can produce exceeds 
the relaxation which a maximal single induction shock produces. This 
constitutes a further point of resemblance between reflex inhibition and 
reflex excitation. The most ample and powerful of the reflex contractions 
elicited by the single break shocks (figs. 1 and 2) fall in amplitude and 
strength much below reflex contractions easily elicitable from the same 
preparation by faradic stimuli, even brief and quite submaximal. Fig. 7 
exhibits in A the reflex relaxation of vasto-crureus evoked by a single 
break shock of 12,000 units of the Kronecker scale. Fig. 7, B, is the reflex 

Sherrington, C. S., Proc. Eoy. Soc, vol. Ixxvii. B, p. 494; Integral. Action etc., 



p. 70. 



Sherrington, C. S., Proc. Roy. Soc, vol. Ixxvi. B, p. 273. 



Comparisons between Reflex Inhibition and Reflex Excitation 



75 




relaxation from a short series of double shocks at 6000 units delivered 
at 30 per second. The reaction B was observed one minute after the 
reaction A, no change 
having been made in 
the position of the 
electrodes, or in the 
reflex preparation 
during the interval. 
From the difference 
between the two re- 
flexes it is evident 
that summation is in 
marked degree a 
physiological pro- 
perty of the reflex 
arc in its inhibitory 
as well as in its excitatory reactions. 

II. Some Features of the Reflex 
Summation. 

It was shown above that in inhibitory re- 
flexes excited by single induction shocks, the 
muscle, after elongating, tends, under the con- 
ditions of experiment, little or not at all to 
return to the length it had prior to the reflex 
which relaxed it. The relaxations, therefore, 
evoked by the individual shocks of a faradic 
series easily sum. Fig. 8 exhibits the relaxa- 
tion caused by a make shock and break shock, 
the latter ensuing somewhat less than one 
second after the former. The elongation due 
to the break reflex practically adds itself to 
that due to the make reflex. This result is 
not like that which happens under similar 
circumstances in the excitatory side of the 
reflex. There a make shock and a break 
shock following at a second's interval give as 
result two short reflex contractions, the second 
not superposed on the height of the flrst, but 
.starting practically from the same base line as Fig. 7. 

its precursor. The reflex contractions are not 

superposed unless the time interval between the two single stimuli be less 
than some 150o-.^ With this the inhibitory reflexes stand in apparent 

^ Sherrington, C. S., Integral. Action etc., p. 44. 




76 



Sherriiipfton 



contrast. Superposition of the mechanical results of single stimuli evoking 
the inhibitory reflex occurs readily when the time interval between the 
stimuli extends to 4000cr-5000(T ; in some records the interval is very much 
longer still. 

Stimuli which, taken singly, produce no perceptible or scarcely percept- 
ible relaxations, produce on repetition relaxations of large extent by 
summation. P'ig. 9 exemplifies this. On the signal line each descent 
marks a break of the primary circuit. Fig. 9, A, gives the effect of sixteen 
feeble break shocks, each of 20 units, on the Kronecker scale, delivered 
in the course of 4-4 seconds. The total relaxation is considerable. Fig. 9, B, 
shows the degree of correspondence between the rate of succession of the 
single weak stimuli and the incidence of the separate reflex relaxations. 
The relaxation produced by a single occurrence 
of the weak stimulus (break shock of 15 units 
of the Kronecker scale) is seen at the top left- 
hand corner. Then follow twenty-two similar 
break shocks of slow and irregular repetition. 
The effect of each of these is evident on the 
graphic record. Next ensues an interval of 
nearly one second without stimulus, and no 
further relaxation occurs during that time. Then 
follow twenty-four more of the break shocks at 
more irregular intervals, and these again sum 
as to the mechanical effect of their reflexes on 
the muscle. The later members of the series of 
stimuli succeed each other rather more rapidly 
than do the earlier, and their combined effect 
is seen to be greater. The greater effect of the 
summation when the individual stimuli follow 
each other more rapidly is seen better in fig. 10. 
Here sixty feeble break shocks (intensity 10 units of the Kronecker scale) 
were delivered. The first twenty-five of these were delivered at the 
average rate of 4"5 per second, and produced relatively little relaxation ; 
the last thirty-five shocks of the series were delivered at an average 
rate of sixteen per second, and they produce much more relaxation — 
more than four times as much. The more quickly delivered shocks 
were, owing to the quicker rotation of the interruption key, given bj^ 
more sudden opening of the primary circuit, and were individually there- 
fore somewhat more intense ; and this applies also to the reflex of fig. 9. 
But the much greater reflex eflfect of the more frequent than of the less 
frequent series suggests that with the former their central reactions, 
and not merely their mechanical elongations of the muscle, summed. 
With summations such as that shown in fig. 8, an interpretation which 
obviously can be put upon the result is, that the eflfect of each stimulus 
may be divisible into two successive parts. The flrst part appears to be 




Fig. 8. 



Comparisons between Reflex Inliibition and Reflex Excitation 77 

a depression of the motoneurone's discharging activity ; the second appears 
as the supervention of a state in which, although there is no further 
depression, there is no restitution, or only a very slow one, of the moto- 




neurone's previous discharge. In the tirst part or period of the etlect 
central inhibition is active ; but nothing appears to indicate that it is active 
in the second. In such summations as that of fig. 8, the summation can be 
explained without supposing tliat tlie process of central inhibition due to 



78 Comparisons between Reflex Inhibition and Reflex Excitation 

the second stimulus overlaps in time with that occasioned by the first 
stimulus. In such summations as that, for instance, of fig. 10, it seems 
probable that the greater reflex effect of the more rapidly following stimuli 




II 



i| 



Fig, 10. 



is due to summation and overlapping in time of the periods of inhibitions 
occasioned centrally by the individual stimuli. 

It is evident that the reflex arc is delicately responsive to the time 
relations of its external stimuli in the inhibitory aspect of its function, 
as well as in the excitatory. 



1^ 



THE FREEZING OF FROG'S NERVE, WITH SPECIAL REFERENCE 
TO ITS FATIGABILITY. By John Tait. (From the Laboratory 

of Physiology, Edinburgh University.) 

During the course of an investigation into the influence of low tempera- 
ture on the conductivity of frog's nerve (1), (2), I often had experiments 
interrupted by the freezing of the nerve. I used the gastrocnemius nerve- 
muscle preparation, taking the muscle response as an index of the state of 
the nerve, of which the middle portion was cooled, while the extreme 
central and peripheral ei ds, as well as the muscle, were more or less pro- 
tected by insulation from cold radiation. When freezing occurred the 
muscle would suddenly begin to twitch, or then pass at once into a condition 
of prolonged tetanic contraction, followed after a minute or so by irregular 
convulsive movements. These continued for two or three minutes, the 
contractions becoming feebler and feebler until they finally ceased. The 
nerve then refused to conduct. In other cases again, one or two little 
twitches of the muscle, followed by absence of conductivity, was all that 
indicated the onset of freezing. In every case after such irregular muscular 
movements the nerve was found on examination to be hard and stifi", and 
had a white, opaque appearance as if covered with hoar-frost. 

At first it seemed that the frozen nerve was killed, for no recovery 
occurred after thawing. Further observation showed that the death of the 
nerve in such instances was due to mechanical injury during the frozen 
condition, for, out of curiosity, I had been in the habit of touching the 
frozen nerve with a pencil to test its rigidity. On ceasing to interfere with 
it in this way I found that subsequent thawing restored the conductivity. 
Since then I discovered that Boruttau had previously made similar 
observations (3). 

The conductivity of the frozen nerve does not, as a rule, come back 
immediately on thawing. Some time usually elapses before conductivity is 
restored. In cases of doubt as to the actual vitality of the nerve, I always 
found that after a rest of an hour or two in Ringer's solution the con- 
ductivity did return. We may consequently take it that a frozen nerve, 
while very susceptible to mechanical injury, is not necessarily killed by 
freezing. 

Such a result is not at all surprising when we consider that poikilo- 
thermic animals in general may be frozen hard without sacrifice of life. 
The experiments of Raoul Pictet (4), to whom most of our know- 



80 Tait 

ledo-e of this subject is due, sliowed that fishes which had been cooled down 
in a block of ice to —15° C still remained alive after careful warming, 
although others similarly treated could be pounded like ice into powder. 
After a temperature of — 20° C the fishes were found to be dead. Frogs, 
again, withstood a temperature of — 28° C. 

The exact temperature at which freezing of frog's nerve occurs I have 
been unable to determine. Boycott (5), who also records some experiments 
on freezing of nerve, was inclined to place the freezing-point somewhere 
about —7° C. With my apparatus, which probably gave a more exact 
determination than his, I found that freezing occurs at a temperature 
somewhat higher than this, viz. between —3° and —5-5° C. 

To study more exactly the changes produced in nerve by means of 
freezing, it seemed better to freeze a minute length than to freeze a longish 
portion. By means of a piece of glass-tubing with an hour-glass constric- 
tion in the middle, on to which the nerve suspended between two pairs of 
electrodes could be lowered, it was possible to freeze a very short length of 
the nerve by running cold alcohol through the tube, and to subsequently 
thaw it again by running in warmer fluid. The outside diameter of the 
constricted part of the tube was 2 mm. The alcohol was cooled by means 
of an ice and salt cooling mixture, consequently the temperature of the 
frozen part of the nerve might vary from the freezing-point right down to 
within a few degrees of — 22° C. The vessels containing the supply of cold 
fluid and of warm fluid respectively were connected with the glass tube by 
means of a three-way stop-cock. A small screw clamp connected to a piece 
of rubber-tubing at the further end made it possible to regulate the rate of 
flow and thus to roughly graduate the temperature. The two pairs of 
stimulating electrodes on which the nerve was suspended were connected 
by means of a Pohl commutator, from which the cross wires were removed, 
to a standard Kronecker induction coil, in the primary circuit of which 
was placed a small accumulator charged to 4J volts. By means of these 
electrodes the nerve could be stimulated either centrally or peripherally 
to the cooled portion. 

I. Changes during Freezing. 

It is difficult to determine the exact cause of the muscular twitchings 
that accompany freezing of the nerve. Prolonged observation showed that 
by the time the muscle twitchings began the whole of the small portion of 
nerve lying over the cold tube was already opaque in appearance. This 
might indicate that the nerve is frozen right through before twitchings 
occur ; but, on the other hand, the white appearance may be due to deposi- 
tion of hoar-frost from the atmosphere before the salt solution inside the 
nerve has reached its freezing-point. The generally accepted idea is that 
the twitchings and subsequent loss of conductivity in the nerve are due to 
mechanical compression owing to the expansion of the water-substance in 
the nerve during freezing, and it has been shown (6) that mechanical 



Freezing of Nerve, with Special Reference to Fatigability 81 

compression applied from the exterior to a normal nerve may temporarily 
abolish conductivity. On physical grounds it is improbable that the out- 
side coating of ice nips the nerve. A hollow cylinder of water undergoing 
freezing would expand and not contract. 

Observations made by botanists on the freezing of plant tissues (7) show 
that crystals of ice become deposited first of all between the cells, and only 
at a later period, when the temperature is still further lowered, inside the 
actual cells. Although there are no observations published, so far as I am 
aware, on the histological appearances inside frozen nerve, we cannot imagine 
that the solution of salts inside the nerve becomes suddenly solid throughout. 
Once freezing starts the water is probably extracted from the surrounding 
solution by degrees, thus raising the concentration of the salts in the non- 
frozen portions. To any definite fixed temperature below the freezing- 
point there would probably correspond a definite concentration of the 
solution in the still fluid portions, while only at a very low temperature 
would the nerve be solid throughout. The twitchings would, from this 
point of view, correspond rather to the twitchings that occur in a muscle 
when its nerve is dried. 

At one time it seemed as if some light might be thrown on the freezing 
process by a study of the muscle twitchings which ensue upon freezing of 
the nerve, for at first I imagined I could detect a correspondence in the 
behaviour during freezing of two preparations from the same frog. 
Further investigation, however, showed that such a correspondence, even 
when all precautions are taken to make the conditions similar, is much less 
constant than I at first thought. 

Another method of investigation which occurred to me was to subject 
some frogs before dissection to continuous evaporation. Durig (8) found 
that frogs may lose in two to three days, by evaporation, from 20 to 30 pei- 
cent, of their weight : in one case he succeeded in diminishing the weight 
of a frog by 39 per cent, without killing it. In the process of drying a 
great concentration occurs in the body fluids, the concentration being least 
marked in the brain and heart of the animals. On the assumption that 
the nerves might participate in this drying-up process, I took a number 
of frogs and, after weighing them, placed them in an open wire cage outside 
the laboratory window. The outside temperature was but a few degrees 
above zero, and yet in eighteen to forty hours they had lost from 12 '8 per 
cent, to 23 per cent, of their body weight. Preparations were examined in 
various conditions of dryness, care being taken not to moisten the nerve 
with saline solution before freezing. Except that one case, viz. the one 
which was least dried, showed a more markedly tetanic and longer-lasting 
response than usual, the muscle twitchings were not different' in character 
from those in ordinary cases. Apparently little help can be got from this 
method of investigating the question. 

The character of the muscle response during freezing of the nei've 
depends to a large extent on the method of freezing. When one freezes 

VOL. I. — JAN. 1908. 6 



82 Tait 

a minute length of nerve, using 'the thin glass tube previously described, 
the muscular response tends to be convulsive and not tetanic, while the 
period of time during which the twitchings continue is not long. When, 
on the other hand, 2 or 3 cm. of the nerve is frozen in a cold chamber, the 
cooling being effected by radiation, the muscle tends at first to go into a 
prolonged tetanus, while the succeeding conviilsive movements may last for 
many minutes. Whether this difference is due to more rapid freezing in 
the case where the cooling is by conduction, or whether it depends simply 
on the shorter length of nerve frozen, I have not as yet determined. 

On looking over my charts, I find that freezing of thick nerves from 
large-sized frogs is, as a rule, accompanied by more twitching of the muscle 
than is the case when preparations from smaller frogs are used. Whether 
this is a general rule or merely accidental I cannot say. 

One thing, however, became apparent during the investigation of the 
muscular twitchings from freezing of the nerve, viz. that under certain 
conditions freezing may occur without any twitching of the attached 
muscle. Previous experiments on the cooling of nerve (9) had shown that 
conductivity may disappear at temperatures lying entirely above the 
freezing-point, while in other cases again no such disappearance of the 
conductivity occurs even though a considerable length of the nerve be 
cooled right down to any temperature short of the freezing-point. It was 
a natural idea that, in those cases in which freezing was unaccompanied by 
muscular twitching, the temperature of such disappearance of conductivity 
was high ; for then, before the molecular disturbance due to freezing occurred, 
the nerve for some considerable distance around might be quite incapable 
of conducting as the result simply of cooling. One or two experiments 
sufficed to show that this is the correct explanation of the phenomenon. 

Nerves in which conductivity disappears at a high tempera- 
ture may undergo freezing without any excitation being trans- 
mitted from the site of the freezing to outlying parts. Other 
nerves, in which the temperature of disappearance of con- 
ductivity lies low, are thrown into a condition of excitation 
throughout their whole extent when freezing occurs. 

Seeing that absence of conductivity in a cooled nerve may arise in two 
ways, (1) in certain cases as the result of cooling to a temperature still 
above the freezing-point, (2) as the result of freezing, it is well to clearly 
distinguish the two conditions. Botanists use the term "cold rigor" to 
indicate the condition in which function ceases in a plant as the result 
of cooling when the temperature still remains above the freezing-point. 
Without committing ourselves to any theory as to the process by which 
conductivity becomes abolished in a nerve which has not been cooled to 
the freezing-point, we shall, for convenience, use the term " cold rigor " to 
denote the condition in question. 

One phenomenon I got to look on as indicative of the near approach of 
freezing in a cooled nerve. If one gradually lowers the temperature, 



Freezing of Nerve, with Special Reference to Fatigability 83 

testing the conductivity at every stage, a sudden marked improvement in 
conduction occurs just before freezing sets in. Before we discuss this 
question, however, a word or two is necessary as to the method adopted to 
test the conductivity. 

The work of Wedensky (9) and of Frohlich (10) on anassthetised nerve 
has shown the value of rapid rhythmical stimulation at varying intensities 
for detecting slight changes in conductivity at a time when, by ordinarily 
employed methods, such changes are not discoverable. By the method of 
applying single maximal shocks to the nerve, depreciation of conductivity 
is indicated only by a falling off in the height of the muscle twitch follow- 
ing stimulation of the nerve. Apart from this change, conductivity seems 
to remain unaltered almost up to the point at which it suddenly disappears 
for good. By the method of rhythmical stimulation, on the other hand, as 
Frohlich showed, one is .able not only to determine roughly the degree to 
which the amplitude of the excitation is cut down, but also to show changes 
in the refractory period. 

With progressive anaesthesia the refractory period of the nerve becomes 
longer and longer ; consequently, when one stimulates at such a rate that 
the interval between the individual excitations is less than the refractory 
period, or about equal to the refractory period, the muscle response becomes 
abnormal. It may be simply an initial twitch of the same height as that 
evoked by one maximal excitation ; it may be a tetanus which attains its 
maximal height at the very start, to fall off very rapidly thereafter, and 
finally to drop to the base line again, thus showing that the conductivity 
of the nerve has temporarily ceased. Indeed, Frohlich proved that this 
latter form of tetanus was an expression of fatigue on the part of the 
nerve, and a tetanus of this special form he names a " fatigue tetanus." 

Such effects are more readily obtained with strong than with weak 
stimulation, for the duration of the refractory period seems to vary with 
the intensity of the corresponding stimulation. As previously mentioned, 
they are to be detected at a time when the conductivity, as tested by the 
method of isolated maximal break shocks, is apparently present in un- 
diminished degree. Consequently, in testing the conductivity of cooled or 
of frozen nerve, I availed myself of this method. The rate of rhythmical 
stimulation used was such as could be obtained by means of the springs 
supplied with the Kronecker coil, and varied with occasion from 30 to 256 
stimulations per second. 

Fig. 1 is a record of an experiment in which a portion of nerve was 
gradually cooled from 0^ C. to — 3"5° C, the state of the conductivity being 
meantime tested every few seconds by short rhythmical stimulation of 
constant intensity (30 Kronecker units) and rate (144 per second). In 
this case the cooling was not effected by means of the thin glass tube, but 
about 7 mm. of the nerve was cooled in a cold chamber. The first six 
stimulations produce a kind of tetanic response of the muscle, each succes- 
sive response, apart from the preliminary uprise, being less marked than the 



84 



Tait 



preceding. These are tjT)ical " fatigue tetani." As cooling proceeds there 
comes a time when only single maximal twitches are got ; these gradually 




decline in height until finally, when the conductivity is just on the point of 
going away, the muscle response takes on the form of a complete tetanus 



Freezing of Nerve, with Special Reference to Fatigability 85 

The next stimulation also evokes tetanus, but of a less height. Immediately 
thereafter the muscle goes into freezing tetanus. When next tested, con- 
ductivity is seen to be abolished. This is a case in which, by an accident, 
the temperature of cold rigor corresponding to the definite length of nerve 
cooled and the freezing-point of the nerve coincide, and the immediate 
improvement in conductivity just before freezing sets in is well shown. 
Similar tracings, none of them, however, so strikingly illustrative as this 
one, might be repeated indefinitely. 

Once or twice it was noted that mere stimulation of the nerve, when 
the temperature was very low, sufficed to send the muscle at once into a 
tetanic contraction as if over-cooling had occurred and the molecular com- 
motion due to stimulation had suddenly induced freezing : but such cases 
are rare. 

During the process of freezing the conductivity is not necessarily at 
once abolished. After the muscle has begun to twitch convulsively, 
rhythmical stimulation of the central end of the nerve may produce a 
smooth and elevated tetanus which stands up above the irregular twitches 
immediately preceding it, and corresponds exactly in duration to the 
period of stimulation on the cessation of which the irregular twitching 
begins again. If the freezing be interrupted during this stage, the conduc- 
tivity is not necessarily abolished. Further, even after the muscle twitch- 
ings due to the occurrence of freezing have entirely ceased, the property of 
conductivity may still be retained by the frozen portion. The muscle 
may react by single twitches to isolated shocks, and to rapid rhythmical 
stimulation applied centrally ; but these twitches quickly fall off in height 
and the frozen portion soon ceases to conduct. Hitherto I have not ob- 
served tetanus of the muscle to follow upon rapid rhythmical stimulation 
under these conditions. 

To sum up: — When freezing takes place the nerve first becomes 
white externally. About the same time an improvement in con- 
ductivity suddenly occurs: excitatory processes are not so much 
cut down in amplitude on passing through the cold area and 
the refractory period becomes shorter. 

Immediately thereafter, in consequence of the molecular dis- 
turbance due to freezing, the whole nerve is, as a rule, thrown 
into excitation. In cases, however, where the temperature of 
cold rigor is normally high, the disturbance arising at the site of 
freezing is not propagated outwards because of the existence 
of cold rigor in immediately adjoining parts of the nerve. 

When the nerve is in this state of more or less tumultuous 
agitation due to the occurrence of local freezing, it may retain 
for some time the property of conducting rhythmical stimuli 
of external origin throughout its whole length, including the 
actual site of freezing. Conductivity, finall}', becomes abolished 
if the nerve be sufficiently cooled, though often enough it does 



86 Tait 



ive 



not disappear until some little time after the convuls 
twitchings of the muscle accompanying freezing of the nerve 
have ceased. These facts indicate that the absence of con- 
ductivity induced by freezing does not come on abruptly, but 
gradually and progressively. 

During the frozen condition the nerve, while very susceptible 
to mechanical injury, is not necessarily dead. 

II. Changes ox Subsequent Thawing. 

The process of thawing of the nerve, unlike the process of freezing, is 
not accompanied b}^ any muscular twitchings. In the case of a nerve 
which has been frozen until all convulsive twitchings of the muscle have 
stopped, the only external sign of subsequent thawing is the disappearance 
of the hoar-frost round the frozen part of the nerve. If the thawing take 
place while the muscle is still contracting vigorously as a result of the 
freezing process, the muscle contractions cease almost abruptly, and the 
lever, after one or two diminished excursions, drops back to the base line, 
to remain there till external stimulation of the nerve be applied. The 
melting of the ice in the nerve, and the presumable re-dilution of its fluid 
contents, consequently sets up no marked excitation of the structure. 

Nor is the vitality of the nerve in any way prejudiced by the rapidity 
with which thawing occurs. Pictet (4) found that, when the living animal 
is frozen, the thawing process must be gradual if life is to continue. 
Although I have thawed nerves at very different rates, in no case have I 
seen the slightest ultimate harm ensue from rapid thawing. 

As was said before, the conductivity of the nerve does not necessarily 
return at once on thawing. The time taken for recovery varies consider- 
ably in different experiments. According to the time taken and the mode 
of recovery, we might distinguish three different types of cases : — A. In 
some cases recovery is rapid, what is apparently full conduction being 
restored within the space of a minute after thawing. B. In other cases 
(the majority in my experiments) recovery is delayed for a period varying 
from a few minutes to an hour or even longer. C. In a few cases a partial 
recovery of conduction occurs after a minute or two, this is succeeded again 
by absence of conductivity, and finally, after a more or less prolonged period, 
full recovery occurs. 

As to the conditions under which these various eventualities occur, I 
have no information to offer. In any one experiment there are a number 
of variable factors, e.g. the original state of the nerve itself as regards 
nutrition, thickness, temperature at which cold rigor occurs, etc. ; the length 
of nerve frozen ; the rate at which freezing occurs ; the ultimate temperature 
to which the frozen nerve is lowered; the rate of thawing and of sub- 
sequent rise of temperature ; the degree to which the temperature is 
ultimately raised; and finally, the extent to which the nerve has been 



Freezing of Nerve, with Special Reference to Fatigability 87 

stimulated before freezing, during the frozen condition and after thawing. 
For all we know, any one of these factors may influence the result, and 
although I have carried out experiments to determine the possible influence of 
various of them, I have as yet reached no positive conclusion. 

It must not be assumed that in every case recovery of the nerv^e occurs 
after freezing. In a certain small number of cases in my experiments the 
nerve did not recover even after a prolonged rest. In each of these cases, 
however, there was reason to believe that the nerve before b^ing frozen 
was not in a normally healthy condition, so that we may look on death of 
the nerve from freezing as quite an unusual phenomenon. 

No matter whether the return of conductivity after thawing be rapid 
or delayed, it is always possible to detect stages in the re-establishment of 
full conduction. As recovery occurs the nerve first of all regains the power 
of transmitting strong excitations ; only at a later period is it able to conduct 
weak excitations, and in a case where recovery takes place slowly the 
gradual improvement of the nerve in the transmission of successively 
weaker and weaker excitations is easily followed. Such observations may 
be made by using single-break shocks of different intensity to excite the 
nerve. 

When one uses, on the other hand, series of rhythmical stimuli of 
varying rate so as to examine the duration of the refractory period, one 
finds that at first single initial twitches are got even with low rates of 
stimulation, at a later stage tetanus; only after some time does tetanus 
occur on rapid stimulation, but eventually even with the most rapid rates 
at one's command the muscle is always thrown into tetanus. This points 
to a gradual and progressive shortening of the refractory period throughout 
the stage of recovery. 

As regards the duration of the refractory period corresponding to stimuli 
of different intensities, the result seems to fall out in two different ways 
according to circumstances. As a rule it is found that the refractory 
period corresponding to a strong excitation is shorter than that correspond- 
ing to a weak one. This is indicated by the fact that rhythmical stimulation 
of high intensity tends to produce full tetanus of the muscle at a time when 
stimulation at the same rate but of lower intensity causes only an initial 
twitch. In a minority of cases an opposite effect is got, viz. full tetani 
occur on weak stimulation, while initial twitches follow on strong 
stimulation, pointing to the fact that in these cases the refractory period 
is shorter for weak stimulations than for strong. The conditions determin- 
ing which of these two effects should follow I have been unable to discover. 

The fact that, under the special conditions of freezing followed by 
thawing, the refractory period of nerve may apparently be shorter thf 
more intense the preceding excitation, is worthy of special attention. 
Frohlich, who first systematically investigated changes in the refractory 
period of nerve by means of the method of rapid rhythmical stimulation (10), 
was inclined to conclude, from the coexistence in ana3sthetised ner^•e of long 



88 Tait 

refractory period with strong excitations, that such coexistence is a general 
phenomenon. As I have found exceptions to this rule, not only in nerve that 
has been frozen and thawed again, but also in nerve which is recovering 
from simple cold rigor, we are not at liberty to generalise and say that a 
strong excitation in nerve is always accompanied by a longer refractory 
period than a weak excitation. The exceptional behaviour in this regard 
of nerve which has been frozen and then thawed would indicate that the 
internal state of such nerve is not quite comparable with that of 
anaesthetised nerve. 

If at any stage during the recovery of a nerve which has been frozen 
and warmed again to room temperature, the nerve be once more frozen at 
the same spot as before, freezing occurs this time without any twitching of 
the muscle. The reason of this is clear when, instead of actually freezing 
the recovering nerve, we simply lower its temperature a few degrees below 
that of the room.^ Cold rigor then occurs at a temperature many degrees 
above zero, although the nerve, previous to being frozen, ina,y have retained 
its conductivity at all temperatures down to the freezing-point. Evidently, 
therefore, a nerve which has been frozen, and subsequently thawed, is, for a 
time at least, very susceptible to the influence of low temperature. 

This condition of the nerve, whereby it loses its conductivity so readily 
with depression of temperature after having been once frozen and warmed 
again to room temperature, is not limited to the first stages of recovery 
after freezing. Long after the conductivity, as tested by rhythmical 
stimulation at the fastest rate possible with the Kronecker apparatus, may 
have to all appearance completely returned, the nerve still tends to go into 
cold rigor at moderately high temperatures, though as time goes on this 
tendency progressively diminishes. By testing the nerve in this way we 
see that changes are produced by freezing which persist for a long time 
after thawing. 

In showing such susceptibility to the depressing action of low tempera- 
ture, thawed nerve does not stand alone. A similar condition is found in 
nerve which, without being frozen, has been cooled to some temperature 
sufficiently low to abolish its conductivity, and has then, during the stage 
of recovery which ensues upon warming, been fatigued by rhythmical 
stimulation.^ Furthermore, under certain nutritive conditions nerve may 
normally, i.e. without any special experimental treatment, show a similar 
extreme susceptibility to the action of slightly lowered temperature. The 
temperature relations in regard to the conductivity of nerve under different 
conditions have as yet been very imperfectly worked out. 

As was previously mentioned, the time for recovery of conductivity after 

1 By arranging that the freezing apparatus should fit on to the apparatus which I 
used for simple cooling of nerve, it was possible, after freezing and then thawing a minute 
portion of the nerve, to subsequently cool this portion to any desired temperature without 
any shifting of electrodes. 

2 For details as to the method of treating the nerve, see paper previously referred to 
on Fatigue of Medullated Nerve by the Method of Cooling (2). 



Freezing of Nerve, with Special Reference to Fatigability 89 

freezing may be long. If the nerve be frozen twice at the same spot, the 
time for recovery of conductivity after the second freezing is still further 
prolonged, though full recovery of conduction, as far as it can be tested 
with the most rapid rate of stimulation possible with the Kronecker 
apparatus, does ultimately occur. Hitherto I have carried out no experi- 
ments to test the effect of repeated freezings and thawings carried out on 
the same nerve. Botanists have shown that plants subjected to repeated 
freezings show gradual diminution in vitality with each freezing and 
ultimately die. Arguing from the result of two freezings on nerve, it is at 
least possible that freezing never occurs in excised nerve without some 
slight irreparable alteration in its structure. 

To sum up at the present stage : — 

Nerve which has been frozen and thawed again is not, for 
some time at least, in normal condition. After thawing, slow 
changes begin to occur which ultimately lead to the more or 
less complete restoration of its function. The time for such 
restoration may vary greatly in different cases. As recovery 
proceeds the nerve first of all becomes capable of transmitting 
strong excitations; only at a later period is it able to transmit 
weak excitations, and this improvement in functioning power 
takes place gradually. Meantime, the refractory period becomes 
progressively shorter and shorter. At any given stage the 
refractory period corresponding to strong excitations may 
apparently be shorter than that corresponding to weak; in 
other cases again the reverse holds true, and the refractory 
period for strong excitations is longer than that for weak. 

After a frozen nerve has been warmed to room temperature 
and its conductivity is beginning to return, it may again cease 
to conduct if simply lowered a few degrees in temperature. 
By testing the nerve in this way for disappearance of con- 
ductivity at different temperatures, it is found that a tendency 
to cold rigor at relatively high temperatures persists for a con- 
siderable time after freezing, though, as time goes on, this passes 
off. In consequence of this exaggerated tendency to cold rigor, 
a second freezing of the nerve following closely upon a first is 
attended by no twitching of the attached muscle. 

On numerous occasions now I have observed that some time after 
freezing, and especially after two acts of freezing, the muscle, which had 
been lying quite still, would suddenly begin to move convulsively, or pass 
into an irregular tetanus like a Hitter's tetanus. The result was in no way 
due to drying either of the muscle or of the nerve, for care was taken 
throughout to keep the preparation moistened with Ringer's solution ; 
nor did any electrical stiunilation act on the preparation. That the source 
of the disturbance lay in the previously frozen portion of nerve is indicated 
by the fact that, on one occasion, when the muscle was thus in active 



90 Tait 

contraction, the tetanus ceased at once when the nerve was snipped across 
with scissors just peripheral to the site of freezing. It would consequently 
seem that, under certain conditions, after freezing the nerve is in a state of 
highly unstable equilibrium and may easily pass into a state of violent 
commotion. On one occasion at least, after such a violent disturbance had 
occurred in the preparation, the conductivity of the portion of nerve affected 
by freezing was found to remain permanently absent although the muscle 
still responded to peripheral stimulation of the nerve. 

One further phenomenon requires to be mentioned. When the nerve, 
after being thawed, has recovered to such an extent that tetani of the 
muscle follow on rapid stimulation, these tetani often tend to take at 
first a peculiar form. This is illustrated in fig. 2. (Although the figure 
illustrates more than the point immediately in question, a complete ex- 
periment has been shown, partly because it happens to demonstrate in 
small compass a number of points already described, partly because the 
conditions under which this peculiar form of tetanus is got are thereby 
rendered more clear.) 

The tracing marked 1 shows the condition of the conductivity at the 
commencement of the experiment, the whole preparation being at room 
temperature. The muscle tetani corresponding to rhythmical stimulation 
(rate 144 per second) are smooth and elevated : in other words they are 
normal tetani. Tracing 2 shows a freezing tetanus followed by ultimate 
loss of conductivity. The nerve was then warmed by turning on fluid at 
room temperature at the spot marked " hot." Tracing 3, which illustrates 
the point immediately under discussion, shows a progressive and rapid 
return of conductivity, rhythmical stimulation at the same original rate 
(144 per second) being used throughout to test the condition of the nerve. 
The first three muscle responses are practically single twitches, the nerve 
being stimulated at intensities 30, 100, and 300, respectively. The next 
three are examples of the peculiar form of tetanus in question. In each of 
these the muscle first gives a single twitch and immediately thereafter 
relaxes. Then it begins to contract again, not in the rapidly summated 
fashion of the tetani in tracing 1, but more deliberately and slowly. With 
each successive series of stimulations the relaxation of the muscle after the 
preliminary twitch becomes less marked ; this change is due, as other 
experiments have shown, to the rapid recovery going on in the nerve. 
Further, the preliminary twitches all show a certain regularity in height, 
thus indicating that the corresponding nerve processes are somewhat 
similar and of approximately the same magnitude. 

A little later the tetani following rhythmical stimulation are free from 
irregularities (tracing 4), though the tetani are not so high as in tracing 1. 
The cold fluid was now made to run slowly through so as to cool the 
nerve gradually, and as a result the muscle responses become single initial 
twitches of approximately the same height as the preliminary twitches in 
tracing 3. These finally become abolished, and the nerve freezes again 



Freezing of Nerve, with Special Reference to Fatigability 



91 



with only one minute twitch to indicate what has occurred. This time, 
after warming, the conductivity did not return for at least half an hour. 




tc 


O 


= g'2 


_c 


a 

i 


■13 

1 


|_2^ 


■3 


-2 


13 


— ' *^ "3 






2 


III 

5 3 '^ 






rt 










■^ 5 m 




a 


•^ 


>.2 


^ 


^ 






2 










o 




'"o §"5 




>> 


a> 


a, -a . 


S 


3 


H 




2 


o 




- g.8 


•5 


1 

3 


2 




X 

2 


r3 




Z-, S I^ i2 






-^ 


^-^ 2 2 


■ o 


3 




|H teg 


■If 


O 

1 




m 


c8 g 

¥ 


'% 


o 


■3 i- ® -^ 


S O 


>> 




^ S3 


'> 


3 


%>=4 






s 


mi 

\, <o a B 








« -■ o .3 




1 


'o 


^ +3 a. 


+3 S 
^3 


■5 


3 




P 


"o 
.2 




ilil 




a 
o 


ll 


ilil 




o 




.St.-—. 


.2 


s 


mil 




ill 




• 2 *^ 

111 






i:^ 35 33 


cS 


"S ?= 


.S ^ ^3.2 


III 


-^ 


c« =■ 


^ o o t2 -w 




O -g 






E 


P 


X 33 


33 33 



B 

33 CO ! 

1.1 



though two hours later the uiu.selc again responded by complete tetani to 
rhythmical central^stimulation. 



92 Tait 

Throughout this experiment the drum was kept uniformly revolving, 
and the figure shows a record of eight minutes' duration. 

As to the internal processes in the nerve corresponding to the kind of 
tetanus seen in tracing 3, all we can say is that the nerve readily transmits 
at least the first excitation of the series, but this preliminary effort is 
followed by temporary disablement which, as the stimulation proceeds, is 
gradually'' recovered from. That the preliminary twitch in the tetanus 
represents only one excitation is probable from its correspondence in height 
with the twitches seen in tracing 5, which, from other experiments, we 
know to correspond to one excitation alone. But why the first excitation 
of the series is followed by a temporary fall in conductivity while later 
excitations are not, is hard to say. 

III. Nerve Fatigue after Freezing. 

In a previous paper (2) I have shown that, after being in cold rigor, 
nerve may be readily fatigued during the stage of recovery of conduction 
due to subsequent warming, and that the fatigue then induced may be 
more or less long-lasting. The method adopted to produde such fatigue 
was sometimes by means of alternate central and peripheral rhythmical 
stimulation, sometimes by means of central stimulation alone. As some- 
what similar results have been obtained after freezing of the nerve, I shall 
now describe them. 

(a) Short-lived Fatigue. 

On p. 87 it was pointed out that, during the recovery stage after 
freezing, the refractory period corresponding to different intensities of 
stimulation may sometimes be shortest when the intensity is greatest, 
while in other cases again the reverse is the case. It would follow from 
this that in certain cases during recovery the refractory period may be 
almost the same for all intensities of stimulation. The example shown in 
fig. 3 is apparently a case in point. 

In the experiment of which this is a record, a minute portion of the 
nerve was frozen and the conductivity found to remain absent for some 
considerable time after thawing. The nerve was taken off the electrodes 
and put back into Ringer's solution for an hour and a half. At the end of 
that time it was found to have almost entirely regained its conductivity, 
the tetani on central stimulation being practically as high as those obtained 
on peripheral stimulation. The previously frozen part of the nerve was 
now gradually cooled to —1° C, at which temperature cold rigor came on. 
On gradually warming the nerve again conductivity returned, at first 
imperfectly and then more and more completely. At a temperature of 
-l-3"5° C. isolated make-and-break shocks at intensity 200 are seen to 
produce single twitches, while rhythmical stimulation at a rate of 144 per 
second produces " fatigue tetani " both at 200 and at 500 units of intensity. 



Freezing|of Nerve, with Special Reference to Fatigability 



93 



On repeating the rhythmical stimulation, fatigue tetani of much the sam* 
form in each case are got at intensities 300, 400, 500, and 100. 




This special form of tetanus is in no way due to fatigue of the muscle, 
for control experiments have again and again shown that peripheral stnnu- 



94 Tait 

lation of the nerve under such conditions causes full and complete tetanus 
of the muscle. Furthermore, had peripheral stimulation been used here, 
the conductivity of the aiFected portion of nerve would thereafter have 
been abolished and the experiment temporarily stopped. The fatigue is 
in the nerve and in it alone. 

From the improvement in the muscle response after each short rest 
from stimulation it is evident that the fatigue in each case is short-lived. 
This corresponds to the fatigue already demonstrated by Frohlich (10) in 
anaesthetised nerve. Although in the present instance the nerve had been 
in cold rigor subsequent to the actual freezing process, still similar curves 
are often obtained during the stage of recovery that follows directly upon 
freezing and thawing. 

(b) More or Less Lasting Fatigue. 

The tracings seen in the next two figures (figs. 4 and 5) are more interest- 
ing. They are again from a preparation of which the nerve was frozen, 
the process of freezing having been accompanied in the case by more or less 
prolonged tetanus and much subsequent twitching of the muscle. On 
thawing conductivity remained absent for at least twenty minutes. The 
nerve was put away to rest in Ringer's solution, and its conductivity 
examined two and a half hours later at room temperature. 

Rhythmical stimulation (rate 144 per second) of the proximal end of the 
nerve first produced the effect seen in fig. 4, tracing 4, viz. " fatigue tetani," 
Almost immediately afterwards the muscle responses were of the form seen 
in tracing 5, after which peripheral stimulation of the nerve was carried 
out — tracing 6. The smooth and elevated tetani which follow upon 
peripheral stimulation show that the muscle is in normal condition and not 
fatigued. After a rest of about five minutes, stimulation was now applied 
to the proximal end of the nerve, with the result seen in tracing 7. It can be 
seen that with each succeeding series of excitations the corresponding muscle 
responses, which are of the " fatigue tetanus " form throughout, become less 
and less vigorous, and tend eventually to die away. 

Another rest of about five minutes was given, and series of rhythmical 
stimulations at a constant intensity of 500 units once more applied to the 
central end of the nerve, with an exactly similar result to the previous one 
(see fig. 5, tracing 1), only that this time the muscle responses died away 
completely in the end. A rest of four minutes was given, and another two 
tetani inscribed with central stimulation. Seeing that the same process of 
gradual extinction of the muscle response was evidently going to be 
repeated, the nerve was then stimulated peripherally so as to make 
absolutely sure that no fatigue of the muscle had occurred ; the result of 
this peripheral stimulation, however, is that the nerve now refuses to 
conduct excitations applied at the central end. As explained in the above- 
mentioned paper (2), the result of strong peripheral stimulation, when the 
nerve is already in a fatigued condition due to rhythmical central stimula- 



^ 



Freezing of Nerve, with Special Reference to Fatigability 95 

tion, is temporarily to abolish the conductivity. The excitations which start 




from the peripheral side run backwards along tlie ner^■e, and probably tluis 
disable the fatigued portion. 



96 Freezing of Nerve, with Special Reference to Fatigability 

A rest of ten minutes was now given, and the nerve once more subjected 
to regular series of centrally applied rhythmical stimulations. The intensity 
was much lower than that previously employed, being 30 units instead of 
500, and 3^et the same effect as before was repeated, the nerve ultimately 
ceasing to conduct. A rest of four minutes suffices to restore the con- 
ductivity in some measure — tracing 4. Peripheral stimulation this time, 
however, does not absolutely abolish conductivity, possibly because the 
intensity of stimulation is low. The nerve, however, soon ceases to conduct 
when repeatedly stimulated. 

Now, in this experiment we can throughout watch the gradual fatigue 
of the nerve as it is subjected to successive series of stimulations. With 
each more or less prolonged rest it recovers to a certain extent, to break 
down whenever regular series of stimulations are again applied. 

Nerve which has been frozen and thawed again is readily 
fatigued by means of rapid rhythmical stimulation. As a rule 
the fatigue is short-lasting, and on cessation of the stimulation 
recovery occurs with great rapidity. In some cases, however, 
recovery may be delayed for a considerable period. 



REFERENCES. 



(1) Tait, "The Influence of Low Temperatures on Nerve," Proc. Physiol. Soc, 
Jour, of Physiol., vol. xxxiv., 1906. 

(2) Tait, "Fatigue of Medullated Nerve by the Method of Cooling," Zeitschr. 
f. allgem. Physiol., vol. viii., 1908. 

(3) BoRUTTAU, "Beitrage zur allgemeinen Nerven- und Muskelphysiologie," 
Arch. f. d. ges. Physiol., vol. Ixv., p. 11, 1897. 

(4) PiCTET, Kaoul, "Das Leben und die niederen Temperaturen," in Revue 
scientifique, 52, 1893. 

(5) Boycott, " On the Influence of Temperature on the Conductivity of Nerve," 
Jour, of Physiol., vol. xxvii., p. 488, 1901-2. 

(6) Howell, Text-Book of Physiology, 1906, p. 109. 

(7) Pfeffee, "Physiology of Plants," trans, by A. J. Ewart, vol. ii., 1903, p. 239. 

(8) DuRiG, " Wassergehalt und Organfunktion," Arch. f. d. ges. Physiol., vol. 
Ixxxv., p. 401, vol. Ixxxvii., p. 42, vol. xcii., p. 293. 

(9) Wedexsky, "Die fundanientalen Eigenschaften des Nerven unter Einwerk- 
ung einiger Gifte," Arch. f. d. ges. Physiol., vol. Ixxxii., p. 134, 1900; "Die 
Erregung, Hemmung und Narkose," Arch. f. d. ges. Physiol., vol. c, p. 1, 1903. 

(10) Frohlich, "Die Erniiidung des markhaltigen Nerven," Zeitschr. f. allgem 
Physiol., vol. iii., p. 468, 1903-4. 



The expenses of this research were defrayed partly by a grant from the Carnegie 
Trustees and partly by a grant from the Crichton Research Scholarship Fund of 
Edinburgh University. 



ON PROTAGON: ITS CHEMICAL COMPOSITION AND PHYSICAL 
CONSTANTS, ITS BEHAVIOUR TOWARDS ALCOHOL, AND 
ITS INDIVIDUALITY. By R. A. Wilson and W. Cramer. (From 
the Physiological Laboratory, University of Edinburgh.) 



Protagon is a Product of a Definite Composition. 

Within recent years the existence of protagon as a definite compound 
has been categorically denied by Gies and his collaborators, whose views 
have been endorsed by Rosenheim and Tebb, and positive assertions have 
been made concerning the part played by protagon in the history of 
neurochemistry, assertions which it is difiicult to test by the exact methods 
of experimental investigation. Whether these authors are right or not. 
wliether protagon is a mixture or not, there is one outstanding fact that 
must be clearly recognised : namely, that from brain a substance of a 
constant chemical composition can be extracted which has received the 
name protagon. 

Gam gee, in his investigations on this substance, prepared a number of 
samples, which on analysis gave figures agreeing inter se, and which re- 
tained this composition after recrystallisation. From his analyses he gives 
the following figures for the composition of protagon : 



C. 
16-39 



H. 

10-69 



N. 
2-39 



1-06 per cent. 

Gam gee's results were confirmed by a number of observers, who by 
identical and by different methods extracted from nervous tissues substances 
having the same composition as protagon, as will be seen from the following 
table : 





C. 


H. 


N. 


P. 


S. 




per cent. 
66-39 


per cent. 


per cent. 


per cent. 


per cent, 

1 


Gamgee 


10-69 


2-39 


1-06 




Baumstark . 


66-48 


11-12 


2-35 


1-02 




Ruppel .... 


66-29 


10-75 


2-32 


1-13 


0-096 


Lesem and Gies . 


66-11 


10-90 


2-02 


1-23 


0-77 


Cramer .... 


66-37 


10-82 


2-29 


1-04 


0-71 



We have not here considered observations in which substances were 
prepared from brain, but only partially analysed, as was done, for instance, 



VOL. I. — JAN. 



98 



Wilson and Cramer 



by Gulewitch, Posner and Gies, Lochhead and Cramer. As a rule, 
only the phosphorus content was determined in these cases, which is, indeed, 
a fairly good indicator of the nature of the preparation. It must, however, 
not be forgotten that preparations have been obtained from brain by Kosse 
and Freytag, by Noll and by Cramer, which, although having the same 
phosphorus percentage as protagon, differ in their nitrogen or carbon 
contents. 

By a method which is described at the end of this paper, we have ex- 
tracted from brain, by means of boiling absolute alcohol, a substance giving 
the following analytical results : ^ 





C. 
per cent. 


H. 

per cent. 


per cent. 


P. 

per cent. 


S. 
per cent. 


Sample C, 4th recryst. . 
Sample A, 4th recryst. . 


66-64 
66-57 
66-40 


10-92 
10-98 
11-01 


2-41 
2-39 
2-33 


0-92 
0-93 
0-97 
0-99 


6-73 



Another sample was prepared by a slight 
method (see p. 104), and gave on analysis : 



lodification of Gam gee's 



Sample D, 3rd recryst. 



C. 

66-40 



H. 

10-71 



N. 

2-55 



P. 

1-02 



S. 
0-68 per cent. 



It is this constancy of the chemical composition which has induced many 
physiological chemists to consider protagon as a definite chemical compound ; 
and, whether this view is right or not, we must insist that the name 
protagon cannot be applied to substances of a different chemical composition. 
As a rule, if a chemist tries to prepare a known substance by a given 
method and fails to obtain the same substance as previous workers, he is 
inclined to suspect his technique. Some workers on protagon, however, 
use that name for any substance that has been prepared according to a 
known method, irrespective of its chemical composition. If a substance of 
a composition different from protagon is obtained, this is taken as evidence 
for " the variability and indefiniteness of the protagon mixture." A recent 
paper on this subject by Rosenheim and Tebb illustrates very instruc- 
tively the confusion which has arisen with regard to the use of the name 
protagon. 

In order to bring forward evidence to support Thudi chum's view that 
protagon is identical with cerebrote, a substance isolated from ox brains by 
means of alcohol extraction by Couerbe in 1837, Rosenheim and Tebb 
have repeated Couerbe's method and have analysed the substance obtained 
in this way. The analytical figures are given here : 

1 The nitrogen was determined by Dumas' method, the phosphorus by Neumann's 
method, the sulphur by Carius' method. 



On Protagon 



99 





C. 
per cent. 


H. 

per cent. 


N. ! P. 
per cent, j per cent. 


Couerbe's cerebrote ; crude product 
Substance prepared by Rosenheim 

and Tebb ; crude product 
Thrice recrystallised .... 


67-82 


11-10 


3-40 2-33 

4-34 1-09 
2-08 0-65 

1 



Although the composition of Couerbe's cerebrote differs markedly from 
Rosenheim and T ebb's substances, they call their substances Couerbe's 
cerebrote, because they have obtained these substances by using Couerbe's 
method. A comparison of these figures with the analyses of protagon given 
above shows that not one of these " cerebrotes " has the same composition 
as protagon. Nevertheless Rosenheim and Tebb conclude that protagon 
and cerebrote — it is not said which of the three cerebrotes — are the same 
substance under two different names, because the same solvent has been 
used for their preparation and because there is some superficial resemblance. 

It seems almost unnecessary to demand that investigations on protagon 
intended to demonstrate the composite nature of this substance should be 
made on material the identity of which with protagon is beyond doubt. 
But a critical survey of these observations will show that even here the 
same tendency to identify protagon by its method of preparation, against 
the evidence of the analytical results, is apparent, although not in such an 
exaggerated form as in the case just discussed. 

The substance which Worner and Thierfelder called protagon, and 
from which they isolated cerebron, was prepared according to Gam gee's 
method. The analytical results of the carbon, hydrogen, and nitrogen 
determinations show variations from 62-37 per cent. C. to 68-97 per cent. C, 
and from 2-39 per cent. N. to 3-39 per cent. N. The phosphorus was not 
determined. 

Posner and Gies have obtained substances by means of Gam gee's 
method with a phosphorus percentage varying from 1-73 to 0-89. All 
their products they call indiscriminately "protagon." This is the more 
remarkable since Gies in his first investigations on this subject had worked 
with substances which had the same chemical composition as protagon, 
and since Posner and Gies themselves had obtained substances the 
phosphorus percentage of which was identical with that of protagon. 
Although these observers had the personal experience that a substance 
of the definite composition of protagon can be obtained, tliey do not 
hesitate to apply this name also to a substance of a difterent composition. 

Rosenheim and Tebb even go so far as to apply the name protagon 
to substances which show an entirely opposite behaviour towards a certain 
solvent. On page 6 these authors state that acetone is a suitable solvent 
for protagon. On the next page they give the results of subjecting a sample 
of protagon twice recrystallised to fractional crystallisation from acetone. 



100 Wilson and Cramer 

The portion remaining insoluble is called "insoluble protagon," and is 
purified by recrystallisation from alcohol and then called " protagon thrice 
recrystallised." This protagon is therefore insoluble in boiling acetone. 
On page 11a method for the preparation of protagon is described, which 
consists in extracting brain with boiling acetone. The substances crystallis- 
ing out on cooling from the various extractions are alleged to be " typical 
protagon." We wish to point out that even if protagon is a mixture, it 
cannot be at the same time soluble and insoluble in one and the same 
solvent. We doubt whether the substances obtained by the method of 
acetone extraction are protagon. The evidence given by Rosenheim and 
Tebb on this point is neither very clear nor very conclusive. The nitrogen 
and phosphorus contents of the crude products only are determined. 
Although they differ from each other and partly from protagon, it is not 
said which product is considered to be protagon. No attempt is made to 
show that a substance of a constant chemical composition identical with 
protagon is obtained after repeated recrystallisation. Besides, in our ex- 
perience, protagon is not easily soluble in acetone. When the experiments 
described in a previous paper by Lochead and Cramer were carried out, 
acetone was tried as a means of extracting protagon from brain and found not 
to be a suitable solvent. The samples of protagon which we have prepared 
are also not readily soluble in acetone. We believe that we have found an 
explanation of the apparent contradiction between these observations on the 
solubility of protagon in acetone, and shall refer to it later. 

Here it may be sufficient to point out the inconsistency in applying the 
name protagon both to a substance soluble in acetone and to a substance 
insoluble in this solvent. 

It is evident that many observations, on the basis of which the existence 
of protagon as a definite chemical compound has been denied, were made 
on material called protagon, which, however, was not identical with 
protagon, but at the best represented a crude product containing protagon, 
together with other substances. 

This looseness of designation is, as we have seen, due to the fact that, 
contrary to the ordinary rules of chemical investigations, the chemical 
composition is not considered to be a criterion of protagon but its method 
of preparation. In this way the impression has gained ground that 
protagon is such a variable substance that it is a mere accident if a sample 
is obtained of the same composition as Gamgee's protagon. We believe 
that we have found an explanation of these failures to obtain protagon, and 
shall show later that protagon can be prepared without difficulty, if certain 
conditions are observed during its preparation and recrystallisation. 

Protagon is a Product of Definite Physical Constants. 
We have spoken of the chemical composition as the only means of 
identifying protagon, because its physical characters, even its crystalline 
form, are neither very conclusive evidence of its identity nor do they allow 



On Protagon 



101 



of a comparison of various samples by a quantitative method. We have 
therefore endeavoured to determine some physical constants of protagon, 
namely, the specific rotatory power and the refractive index of its solution. 
Some difficulty was experienced in finding a solvent which would dissolve 
a sufficient amount of protagon without the use of higher temperatures. 
We found that protagon was comparatively soluble in pyridine, even at 
room temperature. By making our determinations at 30°, a 3 per cent, 
solution could be used. The polarimeter was a Laurent apparatus. The 
determinations were made in a 10 cm. tube at 30"" with sodium light. The 
refractometer was a Pulfrich instrument. A high density prism was em- 
ployed. The pyridine used was Kahlbaum's pyridine, and had at 30'" the 
lefractive index lo062. The results are given in the following table : 



Sample. 


Sjjecific rotation Refractive index 
r "130 of 3 per cent, solu- 
L"Jd= tionat30°. 


A 
B 
C 
D 


+ 6-66 1-5034 
+ 6-90 1-.5033 
+ 6-61 1-5034 
+ 7-01 1-5032 



The results for the specific rotatory power agree well with each other, 
especially if it is considered that the actual reading taken was very small. 
They stand in striking contrast to the observations of Rosenheim and 
Tebb, which were published shortly after our first preliminary communica- 
tion on this subject had appeared. According to these authors, the specific 
rotatory power of protagon varies between 2-7" and 7-5°, and thus demon- 
strates again the composite nature of protagon. As Rosenheim and Tebb 
include, under the term protagon, the substance prepared according to 
Couerbe's method for cerebrote, substances insoluble in acetone and 
substances soluble in acetone, and as most of these substances differ in 
their chemical composition both inter se and from protagon, their physical 
properties could hardlj^ be expected to show any agreement. 

From our observations we conclude that different methods of extraction 
isolate from brain a substance, protagon, of a definite and constant chemical 
composition, which will retain this composition after repeated recrj'stallisa- 
tion and possesses definite and constant physical properties. These facts are 
quite independent of the question to be discussed later, as to whether this 
substance, protagon, is a definite chemical compound or not. 

The Action of Warm Alcohol upon Protagon. 

We have already seen that the plain postulate, that investigations on 
the nature of protagon should be made on protagon and not on some crude 
product po.ssibly containing protagon, has not alwaj's been fulfilled. These 
observations need therefore not be taken into consideration. Nor can the 



102 Wilson and Cramer 

failure of some workers to isolate protagon from brain be considered as 
evidence against its existence, seeing that other workers have obtained this 
substance without difficulty. If we exclude all these observations, there are 
still a number of experiments which show that, by means of a certain 
process, which has been called a "process of fractional crystallisation," 
protagon is split up into substances varying widely in their phosphorus 
percentage and in their solubility in alcohol and ether. This was demon- 
strated conclusively by Gies and his collaborators. In a former paper by 
one of us it was suggested that the protagon of Lesem and Gies may have 
been contaminated with pseudocerebrin. But as the analytical figures of 
their preparations are identical with those of Gam gee's protagon, their 
material must be considered as representing typical protagon, and Posner 
and Gies are right in contending that, if their preparations were con- 
taminated with pseudocerebrin, the same may be said of Gam gee's and 
Cramer's protagon. In order to remove any possible objection, Posner 
and Gies have recrystallised protagon ten times until the phosphorus 
percentage of the crystalline product and of the mother liquid was almost 
the same and identical with that of protagon. Even from this preparation 
different substances could be isolated when the so-called process of fractional 
crystallisation was applied. 

This process consists in the treatment of protagon with a quantity of 
warm alcohol, insufficient to dissolve it, over periods lasting many hours 
(20-24 hours). After separating the soluble from the insoluble part, the 
solution is allowed to cool slowly, and the substances crystallising out at 
different temperatures are collected separately and then show the differences 
mentioned above. 

Of the correctness of these facts there can be no doubt ; it is only in 
their interpretation that we differ from Gies and his collaborators. In 
order to interpret these facts as proving conclusively the composite nature 
of protagon, it is of course essential that the prolonged treatment with 
warm alcohol does not effect any change in the protagon — in other words, 
that the process is really one of recrystallisation and not one of decomposition. 
This last possibility has, indeed, been considered by Lesem and Gies, who, 
is speaking of their results of fractional crystallisation, say : " They show 
that protagon is either a mixture of bodies or else a substance decomposing 
quite readily under the conditions of such experiments." Rosenheim and 
Tebb simply dismiss the second possibility by saying that the process of 
fractional crystallisation evidently cannot effect any serious chemical 
decomposition. A priori there is no reason why these results should not 
prove with equal force that the prolonged treatment with warm alcohol has 
induced a decomposition. Before the treatment protagon is, as we have 
seen, a substance of a constant composition retaining this composition after 
simple recrystallisation, which involves only a short contact with warm 
or boiling alcohol, the mother liquor and the crystalline product having an 
almost identical phosphorus percentage : in the course of its preparation it 



On Protagon 



103 



has been subjected (in the precipitated stage) to a very thorough washing 
with cold ether. After the treatment a substance remains less soluble in 
warm alcohol than protagon, and almost phosphorus free ; the substances 
crystallising out vary considerably in their phosphorus percentage from 
that of their respective mother liquids, the difference sometimes exceed- 
ing 1 per cent., and yield to ether a considerable quantity of a substance 
containing less phosphorus than protagon. If protagon is a mixture of all 
these substances, it must follow that in the process of simple recrystallisation 
these substances dissolve and crystallise out in constant proportions, as this 
process always leads to the same product — protagon. 

The assumption that protagon is a mixture of these substances, so 
different in many respects, is therefore not without difficulty ; and if the 
chemical individuality of protagon had not been under suspicion, these 
results would have been interpreted as proving the instability of protagon 
towards warm alcohol. The statements of previous workers on this 
question, whether protagon is decomposed or not by warm or boiling 
alcohol, are very contradictory. The reason for this is probably that 
any change that might take place would not be very obvious, and could 
only be demonstrated by an elaborate experiment. Indeed, if one had 
wished to study the action of warm alcohol on pi'otagon, one could only 
have done so by experiments similar to those of Gies and his collaborators. 

The determination of the physical constants and their constant value 
has made it possible for us to obtain direct and conclusive evidence on this 
question. The two samples of protagon, A and B, the physical constants of 
which had been determined, were treated with 80 per cent, alcohol at 44° 
for 22 hours, in the manner prescribed for the process of fractional crystal- 
lisation. The vessels were then cooled in ice and the alcohol evaporated in 
vacuo, no filtration having taken place. The remaining residue was dried 
in vacuo over concentrated sulphuric acid and its specific rotatory power 
and the refractory quotient were determined as before. As nothing had been 
removed by filtration the proportion of the elements must have remained 
the same as in protagon, so that on analysis the figures obtained would be 
identical with those for the composition of protagon. As a control a sample 
was dissolved in boiling alcohol, the solution was kept boiling for Ih 
minutes, and the vessel cooled in ice. After evaporating the alcohol in 
vacuo without previous filtration, the residue was dried and its physical con- 
stants determined as before. The results are given in the following table : 



Sample. 


Treatment. 


Rotation 

r T^" 


Refractive index 

of 3 per cent. 

solution. 


A 

B . 

C . . 


85 per cent, alcohol at 44° for 22 hours 

)) 55 55 

Boiling alcohol for 1^ minutes 


13-43 

13-08 
6-69 


1-5041 
1-5038 
1-5034 



104 Wilson and Cramer 

The results show that a short contact with boiling alcohol is without 
effect on protagon, while prolonged treatment with warm alcohol produces a 
(•hano-e. The product obtained after this treatment is not identical with 
protagon, as the physical constants show, although its analytical figures 
would agree with protagon. The results obtained by the so-called process 
of fractional crystallisation, therefore, do not give any conclusive indication 
as to whether protagon is a mixture or not, the method being an unsuitable 
one, but have to be interpreted as proving the relative instability of 
protagon towards warm alcohol. This conclusion again is not in any way 
dependent upon either the composite or the uniform nature of protagon : 
the products which are found after the treatment are, apart from 
unchanged protagon, decomposition products of either the definite compound 
protagon or of the substances constituting the mixture protagon. 

The statement made in a former paper by one of us that " protagon is 
not decomposed by warm ether or boiling alcohol," must therefore be 
corrected.^ In the case of boiling alcohol it is true only if the solvent 
is prevented from acting on protagon for some time. We have not 
investigated the effect of ether on protagon, but the instability of this 
substance towards warm alcohol makes it probable that other solvents are 
not without disintegrating effect. This is perhaps the explanation of the 
contradiction above referred to, that, although protagon is not readily 
soluble in acetone, a method of acetone extraction has been proposed for the 
preparation of protagon. As in this method the brain material is boiled 
with acetone for four hours, it seems doubtful, in view of our experiments, 
whether decomposition has not taken place, and whether the substances 
extracted are not decomposition products of protagon, possibly together 
with some unchanged protagon. 

The relative instability of protagon towards warm alcohol is a fact 

which throws light on some obscure and controversial points of the protagon 

problem. It is clear that a prolonged treatment with warm alcohol, both in 

the preparation and in the recrystallisation of protagon, must be avoided ; 

otherwise decomposition products are formed which are extracted together 

with protagon. This is probably the explanation of the lack of uniformity 

in the composition of products obtained by some workers, especially when 

Gamgee's method has been used. It is usually stated that Gamgee's 

method was followed in every detail. We have been unable to find in 

Gamgee's communications any definite information about the time of 

recrystallisation, and most other workers who have made use of this method 

do not give any detailed statement with regard to this point. The sample 

of protagon which we prepared by Gamgee's method, slightly modified, and 

which was identical in its chemical and physical properties with protagon, 

1 While freely acknowledging this error, I wish to point out that the statement in 
the sarne paper, that " choline is the only Imse formed " on hydrolysis by baryta water, has 
been misunderstood by Kosenheim and Tebb. If read in connection with the paper, this 
statement appears as the result of experiments intended to decide whether choline or 
neurine or both bases are formed. [W. C.] 



On Protagon 105 

was extracted for four hours with alcohol and three times recrystallised, 
the solution being effected in each case by warm alcohol in three-quarters 
of an hour. The products which Posner and Gies obtained by this method, 
and which had such an abnormally high phosphorus constant, were prepared 
by extractions lasting for twenty-four hours. In purifying the crude 
product by recrystallisation it was treated with alcohol for periods varying 
from five to fifteen hours. 

In order to test the validity of our conclusion that it is the prolonged 
treatment with warm alcohol which is responsible for the failure to obtain 
protagon, we have applied Gamgee's method in the same way as before, only 
extending the extraction to twenty-four hours, and the treatment with 
warm alcohol for the solution of the crude product in the recrystallisation 
process to twelve hours. The specific rotatory power of the substance 
which had been prepared in this way was [ajn — 13"12, refr. index 1"5039. 
This substance was, therefore, not protagon, whatever the results of the 
chemical analysis would have been. 

We can also understand the apparently paradoxical fact that a protagon 
sample, obtained after repeated crystallisation and identical in every respect 
with protagon, may appear after a further recrystallisation to be con- 
taminated with pseudocerebrin or other substances. Formerly pseudocerebrin 
was held to be extracted from brain, together with protagon, from which it 
had to be separated. We know now that this substance (which is identical 
with cerebrin) is formed from protagon by the hydrolysing action of warm 
alcohol. If, therefore, in the process of recrystallisation the alcohol has been 
allowed to be in contact with protagon for a longer time than in the 
previous recrystallisation — and as the destructive action of warm alcohol 
has not been suspected previously, such a condition may have easily 
occurred — decomposition products will be formed, and after this recrystal- 
lisation the protagon may really be less pure than it was before. 

Method for the Preparation of Protagon. 

In order to shorten as much as possible the time during which the hot 
or warm solvent is in contact with the material in the preparation of 
protagon, we have adopted the following procedure, which we have found 
to be a most convenient method for the prepai'ation of protagon. 

The brain mass is made into a pulp and treated repeatedly with 96 per 
cent, alcohol in a wide-mouthed bottle. The extraction is accelerated by the 
use of a shaking machine, the material being kept afterwards in an ice chest. 
After three to four extractions ether is added instead of alcohol, and the 
treatment continued until lecitliin and cholesterin are completely extracted. 
After removing the ether by filtration and drying the remaining mass by 
exposing it to the air at room temperature, a brown mass remains which 
can be made easily into a fine powder. To this powder any solvent can be 
applied directly. We have prepared protagon from it according to 



106 Wilson and Cramer 

Gamgee's method by means of warm alcohol (see p. 104) and by extraction 
with boiling absolute alcohol. In the latter case the boiling solvent is 
poured on the powder and the mixture kept boiling for one to two minutes 
in a water bath, moving the mixture all the time. The alcoholic solution is 
filtered through a hot-water funnel ; the filtrate is allowed to drop into a 
vessel cooled in ice. The same process of extraction is repeated twice. 
The crude crystalline product is washed with ether and dried in vacuo. 

Recrystallisation is effected by pouring boiling absolute alcohol on the 
sample of protagon. The solution is kept boiling for one minute and then 
filtered as before. 

This method offers many advantages. Water and all the substances 
soluble in cold alcohol and cold ether are removed before the extraction 
begins, so that the bulk of the material is greatly reduced and less of the 
hot solvent is necessary. In this way even a large quantity of material, 
fifteen to twenty ox brains, can be worked up easily, while with other 
methods the bulk of the material and the volume of alcohol are so 
great that the manipulations cannot be carried out neatly and require 
a longer time. 



Is Protagon a Definite Compound or a Mixture of Phosphatids 

AND CeREBROSIDES '* 

In the preceding pages we have considered protagon simply as a 
substance prepared from brain, and having constant and definite chemical 
and physical properties. The question whether it is a definite chemical 
compound or a mixture we have left open, so that the facts which we have 
observed and the conclusions which we have drawn remain independent 
of this controversial subject. 

Conclusive proof of the chemical individuality of protagon can onl}' 
be brought by synthesis. Evidence to the contrary could be obtained 
by isolating the substances of which the mixture protagon is con- 
stituent and to reconstitute protagon from them. Rosenheim and 
Tebb hope to be able soon "to reconstitute a pure protagon" with a 
phosphorus percentage varying from 0'9 to 1"26, by making a mechanical 
mixture in certain proportions of substances nearly phosphorus free and 
substances containing about 3 per cent, phosphorus. We have no doubt that 
it is possible to obtain in this way a mixture of the same phosphorus per- 
centage as protagon. But if that were to prove that protagon is a mixture, 
one might also prove that fat is a mixture of glycerine and fatty acids, 
because a mixture of these substances in certain proportions would have 
the same carbon percentage. If Rosenheim and Tebb wish to prove that 
this mixture of alcohol-soluble and alcohol-insoluble substances is identical 
with protagon, they will have to show that this mixture retains its com- 
position after repeated recrystallisations and that it has the same specific 
rotatory power as protagon. 



On Protagon 107 

Until this question can be decided conclusively, either one way or the 
other, an objective interpretation of the known facts must be sufficient. 

We have emphasised already the fact that protagon has a definite 
chemical composition and retains this composition after repeated recrystal- 
lisation. This substance has been obtained by various observers and by 
various methods. As we have pointed out in a former communication, this 
fact is evidence in favour of the view that protagon is a definite chemical 
compound. To this evidence we add further the fact of the constancy of 
its physical properties. The crystalline form of protagon we have never 
considered to be of much value in recognising the nature of protagon, as it 
is well known that mixtures of complex organic compounds frequently 
crystallise out together in a definite crj^stalline form. The weight of the 
analytical evidence has been admitted even by those who hold different 
views on the nature of protagon. Lesem and Gies discuss their analytical 
results of four samples of protagon as follows : " Much to our surprise, these 
results accord as well as many analytical series given for what are un- 
doubtedly individual substances. Our data in this connection, considered by 
themselves, would seem to harmonise with the older view of the integrity 
of protagon." 

Against this \aew the results of the so-called process of fractional 
crystallisation have been adduced as demonstrating that protagon is a 
mixture of substances differing widely in their solubility, differing widely in 
their chemical composition and constitution, and in their physical constants. 
These results appear in a new light, since we have been able to prove that 
they are due to a factor which has not been recognised before, namely, the 
instability of protagon towards warm alcohol. This fact changes the 
process of fractional recrystallisation into one of partial decomposition, 
and invalidates the conclusions which have been drawn from these 
experiments. 

This property of protagon is responsible for the fact that, bj- following 
known methods for the preparation of protagon, substances have been 
extracted from brain which differ from protagon in their chemical com- 
position. By calling all the substances protagon which were prepared 
according to an acknowledged method, these results have been used as 
additional evidence for the variability of the protagon mixture. In the 
only case in which the account of the technique employed was detailed 
enough to repeat the process, it was possible to show that the failure to 
obtain protagon was due to the prolonged treatment with warm alcohol. 

Not only is it incorrect to interpret these results as evidence against 
the existence of protagon, but we must protest against such reasoning, 
which threatens to deprive the protagon problem of its very basis. What- 
ever protagon is, the name protagon has been given to a substance of a 
definite chemical composition, having, as we have seen, definite physical 
constants. Like every other chemical substance, protagon is identified by 
these properties and not by its method of preparation. Many of the 



108 Wilson and Cramer 

substances which figure in recent papers as protagon have no claim to this 
name. 

There is therefore no evidence in favour of the view ofThudichum and 
his followers that protagon is a mixture of substances differing from each 
other in almost every respect. On the contrary, we must conclude that 
the substances found after the prolonged treatment with warm alcohol are, 
besides unchanged protagon, decomposition products of protagon. One of 
these decomposition products, the phosphorus free phrenosin (cerebron, 
pseudocerebrin), has indeed been found to be identical with cerebrin, which 
is obtained by hydrolysing protagon by baryta water. The phosphorised 
moiety of the protagon molecule, however, does not behave in the same way 
towards warm alcohol and towards baryta. This last reagent carries the 
hydrolysis to the ultimate constituents; of which choline, sphingosine, 
glycerophosphoric acid and fatty acids have been isolated. The action of 
alcohol does not go so far. This is quite clear from the investigations of 
Thudichum, who isolated numerous substances from brain by methods 
involving prolonged extraction with boiling alcohol and distillation of the 
alcoholic solutions regardless of any decomposition. These substances w^ere 
hydrolysed further by means of baryta water. The treatment with warm 
alcohol, instead of being a means of separating the substances which are 
present in the mixture protagon, would appear therefore to be a useful 
method for the study of the more complex groups which enter into the 
composition of the compound protagon. 

Seen in this light, the observations of Gies and of Rosenheim and 
Tebb, so far from being opposed to the view that protagon is a definite 
compound, are a valuable contribution to the study of the constitution of 
this substance. 

The question has also to be considered whether the substances which 
have been isolated from brain, and which at the same time have been 
isolated from protagon hydrolysed by alcohol, such as, for instance, phrenosin 
(pseudocerebrin, cerebrin, cerebron), exist preformed in the brain or are 
formed only in the process of extraction. 

Although our results support the view that protagon is a definite 
compound, we do not exclude the possibility, which has been considered 
already in a former paper, that several protagons exist just as several 
lecithins exist. The existence of such substances, differing perhaps only 
in the nature of the fatty acid radicle which they contain, would be no 
more evidence against the existence of protagon than the existence of 
several lecithins is considered to disprove the existence of a definite 
compound lecithin. Such protagons would be distinguished only by slight 
differences in the carbon and hydrogen contents, and would resemble each 
other very closely in every respect, so that they could not be separated 
easily. It is therefore possible that protagon is a mixture of such sub- 
stances, although there is at present no evidence for it. Such a view, which 
is quite compatible with the idea that protagon is a definite compound, is 



On Protagon 109 

fundamentally different from the view that protagon is a variable and 
indefinite mixture of cerebrosides and phosphatids. 

COXCLUSIOXS. 

1. Protagon is a substance of a definite chemical composition, retaining 
this composition after repeated recrystallisation. 

2. Protagon is a substance with definite and constant physical properties. 
The specific rotatory power and tlie refractive index of several samples of 
protagon have been determined. 

3. Protagon is identified by its chemical composition and by its physical 
constants. Many substances to which the name protagon has been given 
on account of their method of preparation, do not conform to these con- 
ditions, and therefore have no claim to the name protagon. Couerbe's 
cerebrote is not identical with protagon, but probably a mixture of 
substances of which protagon is one. 

4. Protagon is decomposed by a prolonged treatment with warm alcohol. 
The so-called process of fractional crystallisation is therefore in reality a 
process of partial decomposition. The conclusions which have been drawn 
on the assumption that it is a process of recrystallisation are not valid. 
There is, consequently, no evidence for the view that protagon is a mixture 
of cerebrosides and phosphatids. 

5. The constancy of the physical and chemical properties of protagon 
support the view that protagon is a definite compound. The substances 
isolated from protagon after prolonged treatment with warm alcohol, and 
formerly held to exist as such in the mixture protagon, must now be con- 
sidered to be the constituents of the protagon molecule. They are the 
intermediate decomposition products of protagon. 

6. Details of a method for the preparation of protagon are given, by 
means of which a prolonged contact with warm or boiling alcohol can 
be avoided as much as possible. 

Addendum by W. Cramer. 

The paper by Lochhead and Cramer has called forth a polemical 
paper by Gies (Journal Biolog. Chemistry, iii. 4, p. 339) which is mainly a 
restatement of the views of Posner and Gies and does not adduce any 
new facts. Gies believes that our results support his view, and that we 
obtained " different mixtures by extracting brain with different solvents." 
He applies to our results a different standai'd from that which led him to 
state of the protagon samples of Lesem and Gies that the analytical 
results "accord as well as many analj^tical series for what are undoubtedly 
individual substances." The phosphorus content of the purified protagon 
samples of Lesem and Gies varies from 0'89 per cent, to 126 per cent., 
that of Lochhead and Cramer's purified products from 0-96 per cent, to 
1"07 per cent. Even if our non-purified products are included, the 



no On Protagon 

difference in the phosphorus percentage is only 0*03 per cent, more than in 
the samples of Lesem and Gies, the variation in these being 0"94 per cent, 
to 1'34 per cent. 

Gies further states (in his recent paper on Paranucleo - Protagon, 
American Journal of Physiology, xx. 2, p. 379), referring to our results : 
" When the phosphorus contents of their protagon products were lowered by 
recrystallisation to the percentage amount that appeared to them to be 
about right, they arbitrarily discontinued in each case the recrystallisation 
process, in spite of the fact that repetition of it promised to decrease further 
the proportionate contents of phosphorus." I have looked in vain through 
our paper for any statement which could warrant this wanton suggestion. 



PAPERS REFERRED TO. 

Baumstark, Zeitschrift fur physiologische Chemie, vol. ix., 1885, p. 145. 

Cramer, Journal of Physiology, vol. xxxi., 1904, p. 31. 

Gamgee, Text-Book of Physiological Chemistry, London, 1880, p. 441. 

Gamgee and Blankenhorn, Zeitschrift flir physiologische Chemie, vol. iii., 
1879, p. 260. 

GuLEWiTSCH, quoted from Noll. 

KosSEL and Freytag, Zeitschrift fiir physiologische Chemie, vol. xvii., 1893, 
p. 431. 

Lesem and Gies, American Journal of Physiology, vol. viii., 1902, p. 183. 

Liebreich, Liebig's Annalen der Chemie und Pharmazie, vol. cxxxiv., 1865, 
p. 29. 

Lochhead and Cramer, Biochemical Journal, vol. ii., 1907, p. 350. 

Noll, Zeitschrift fiir physiologische Chemie, vol. xxvii., 1899, p. 370. 

PosNBR and Gies, Journal of Biological Chemistry, vol. i., 1905, p. 59. 

Rosenheim and Tebb, Journal of Physiology, vol. xxxvi., 1907, p. 1. 

RUPPBL, Zeitschrift fiir Biologie, vol. xxxi., 1895, p. 86. 

Thudichum, Die chemische Konstitution des Gehirns, Tubingen, 1902. 

Wilson and Cramer, Seventh International Congress of Pliysiology, Heidelberg, 
1907. 

WoRNER and Thierfelder, Zeitschrift fiir physiologische Chemie, vol. xxx., 
1900, p. 542. 



The expenses of this research have been defrayed by a grant from the Moray 
Fund of the University of Edinburgh. 



THE " FLY-CATCHING REFLEX " IN THE FROG. By J. A. Gunn. 
(From the Pharmacology Department, University of Edinburgh.) 

(Received for puhlicatioti l\th February 1908.) 

One of the most conspicuous symptoms which result in the frog from the 
administration of toxic doses of yohimbine is the appearance of a fly- 
catching reflex. This symptom is interesting not only on account of the 
invariability of its occurrence, but also because it seems to illustrate the 
close resemblance which may obtain between the effects of toxic action 
and operative lesion on the central nervous system. 

The following experiment will serve to show the general effects produced 
by a large sublethal dose of yohimbine in frogs, and also the relation which 
the symptom under consideration bears to other symptoms. 

A healthy male frog (R. temporaria) weighing 29 grammes was used. 

At 3.7 p.m. the throat respirations were 30 in 10 seconds, the heart-beats 
8 in 10 seconds. 

At 3.20, 0"7 cc. of a solution of yohimbine lactate (0"01 gm. in 5 cc.) was 
injected into the dorsal lymph sac. This was equal to 0*048 gm. per kilo, 
the minimum lethal dose being 0*05 gm. per kilo. 

At 3.35 the normal respirations had entirely ceased, and were replaced 
by infrequent gulping movements. The head was lowered and the limbs 
not fully drawn up. The frog made no spontaneous movements, but when 
pinched jumped well, and when laid on his back recovered his usual posture 
quickly. The conjunctival and limb reflexes were acute. The lower eyelid 
covered half the eye. At 3.55 if pinched he did not jump, but moved 
forward on his abdomen by kicking. When laid on his back he recovered 
once, but when placed on his back a second time immediately after he was 
unable to do so. 

At 4.15 when laid on his back he kicked vigorously, but coukl not 
turn over. 

At 4.53 (one hour thirty-tliree minutes after injection) when laid on his 
back he made a few feeble movements and then lay still with legs ex- 
tended. When a foot was now pinched the leg was drawn up quickly. 
Movements of any nature brought on fatigue very soon. When the 
animal's hand or nose was touched he snapped in tlic direction touched, 
extending his tongue as a frog does when catching a fly. 

At 7.0 he jumped feebly when pinched. Tliough in jumping he could 
raise himself off the ground, his movements were badly co-ordinated ; for 
example, one hind limb would get flexed behind the other. 

VOL. I. — APRIL 1908. 8 



112 ' Gunn 

At 9.0 when pinched he went into a convulsion, in which the hind Kmbs 
were rigidly extended with slight opisthotonus, followed by emprosthotonus, 
with the body flexed and the arms flexed under it. The fly-catching reflex 
was elicited as before. 

At 7.20 a.m. next morning he still had when pinched convulsions 
similar to that described. The eyes were fully open. If a bright object 
was held near his eyes, he snapped at it, extending his tongue. He could 
not turn over when placed on his back. 

During all this day he remained in nearly the same condition, with, 
however, some increase of co-ordinating power, for by the end of the day he 
could turn over, though still with some difllculty, when laid on his back. 

On the second day after injection he sat with his limbs fully drawn 
up, made frequent spontaneous movements, could jump well, and turned 
over rapidly when laid on his back. When his hand or nose was touched 
he no longer snapped, and when a bright object was held near his eyes he 
moved away as a normal frog does. He appeared to have practically com- 
pletely recovered, except that his respirations were only ten per ten seconds. 
Two days later, however, the rate of the respirations was the same as 
before injection. 

When the earlier symptoms are analysed with a view to their explana- 
tion, it is apparent that the cessation of voluntary movements, the inco- 
ordination of movement when such movement is elicited by stimulation, 
and the loss of the power of jumping are symptoms which so much resemble 
those which result from operative destruction of the cerebrum, mid-brain 
and cerebellum as to lead one to infer that yohimbine early in its action 
abolishes the functions of these parts. 

That the medulla oblongata, too, is involved is shown by the cessation 
of respiration, and by the inability of the animal to recover its normal 
posture when laid on its back. These eflfects are not the result of a peri- 
pheral paralysis, for, as I have ascertained from control experiments, the 
motor nerves and the voluntary muscles at this time react normally to 
electrical stimulation. 

Since pinching the foot elicits withdrawal of the leg as promptly as 
before injection, yohimbine does not impair the functions of the spinal cord 
as it does those of the higher parts of the central nervous system. 

Is there anything in the nature of the action of yohimbine on the central 
nervous system to explain the appearance of the " fly-catching reflex " ? 

In every one of a large number of experiments with doses of yohimbine 
lactate ranging from 0*22 gm. to 0-08 gm. per kilo subcutaneously, I have 
found this reflex elicitable at some time during the experiment. With the 
largest of these doses (0-08 gm. per kilo), which killed the frog in one 
and a half hours, it was obtained only once, namely, about an hour after 
injection, though it was tested for at regular intervals of a few minutes. 
With the smallest of these doses this reflex could be elicited for over 
twenty-four hours. 



The "Fly-catching Reflex" in the Frog 113 

The frog poisoned by yohimbine snaps and generally extends his tongue 
if his hand or nose be touched ; sometimes if a bright object be brought 
near his eyes. If the hand of a normal frog be touched the arm is drawn 
away ; if the nose be touched the head is depressed and the eyes closed ; 
while if a bright object be approached close to its eyes, the animal merely 
moves away or closes its eyes if it react at all. But Schrader (1) has 
shown that the snapping for food is a reflex from slight stimulation, and 
that a frog deprived of its cerebrum will catch flies under suitable circum- 
stances, and given a long enough time for recovery from operation. He 
has also shown that if the brain is destroyed down to the fore part of the 
medulla oblongata there is developed a somewhat difterent snap reflex. In 
this case the frog snaps if its nose or hand be touched lightly, the head 
being directed as far as possible towards the place of stimulation. 

The remarkable resemblance between these reflexes and those occurring 
after injection of yohimbine led me to conclude that yohimbine, by a para- 
Ij^sing action on the upper part of the central nervous system, imitates the 
operative lesions and results of such lesions as Schrader describes. 

The evidence which I have cited above shows that yohimbine does 
paralyse the functions of certain parts of the supraspinal portion of the 
central nervous system. Though the appearance of this fly-catching reflex 
has not to my knowledge been described hitherto as the result of the action 
of a toxic agent on the frog, still, as a consequence of such action, symptoms 
simulating the effects of operative lesions are well known to occur. 

However, in the case of poisoning of an intact animal, when touching 
the muzzle or hand or bringing a bright object near the eyes evokes a fly- 
catching or snapping reaction, there is a possibility that this reaction is 
a voluntary one, abnormal it is true, but rendered elicitable, for example, 
by a depression of some normal inhibitory influence of the cortex cerebri. 
In order to decide this point I destroyed the cerebra of two frogs and 
administered yohimbine to one of them immediately after operation. The 
snap reflex occurred in a few hours in the poisoned frog, and in it alone. 
Since it occurs when voluntary impulses are cut off", it is a true reflex. 

There is another possible explanation of the production of this reflex 
bj^ yohimbine, still in keeping with the known method of action of drugs. 
There would seem to be in this condition a hypersensitiveness of the centre 
involved, or, what may be the same thing, a condition in which varied and 
divergent afferent stimuli find a path of least resistance in reflexion 
through the centre for the apprehension of food. The action of yohimbine 
may therefore consist in " facilitation of the discharge of force already 
latently present, and the rendering of the liberating forces more ettective 
tending to thwart inhibition " (2), and this action may be exerted especially 
on this centre in the bulb. Certain facts go to show that yohimbine may 
so act on the nervous system. The fact that small doses of j'ohimbine, by 
an action on the respiratory centre, induce an increase in the rate and 
amplitude of the respirations (3), may be explained by the same kind of 



114 ' The "Fly-catching Reflex" in the Frog 

action. Also, later than the fly-catching reflex, there come on spinal con- 
vulsions. These may be due to the supposed action on the bulb — an action 
similar in nature to that of strychnine — spreading to the spinal cord. 

A third circumstance may possibly have some bearing on the ease with 
which this snap reflex is elicited in a frog poisoned by yohimbine. "A 
point of general interest in the physiology of the great alimentary nerve 
centre in the bulb is the high degree to which it employs inhibition. Each 
subdivision of it is depressible by inhibitory fibres from some afierent nerve 
trunk, e.g. respiration by fibres in the superior laryngeal, deglutition by 
fibres in the superior and partly in the inferior laryngeal nerves " (4). I have 
noticed in the frog, poisoned either by yohimbine or by other substances 
which paralyse the respiration, that cessation of the normal respiratory 
movements is followed for a short time by gulping movements which appear 
to be movements rather of deglutition than of respiration. The swallowing 
movements are induced especially by slightly disturbing the frog. Is it 
possible, as a converse of the stimulation of the laryngeal nerves, that 
paralysis of the respiration removes some normal inhibitory effect, and so 
allows a more ready elicitation of the movements of deglutition, and also 
perhaps of the more complicated fly-catching reflex ? 

Whether yohimbine produces this reflex by a paralysing action on the 
upper part of the central nervous system (as by operation), or by an action 
on the medulla oblongata facilitating the elicitation of a latent reflex, or 
by both actions combined, the phenomenon is interesting as illustrating in 
a particular manner the close resemblance in the effects produced by 
operative lesion and toxic action on the nervous system of the frog, and 
on the other hand, as showing the selective action of a toxic agent on the 
nervous system, since there may occur in yohimbine poisoning a paralysis 
of the respiratory centre coincident with an exaggerated activity of the 
closely related centre for the apprehension of food. 



i 



REFEEENCES. 



(1) ScHRADER, Arch. f. d. ges. Physiol., Bonn, 1887, Bd. xh. 

(2) Sherrington, Schiifer's Textbook of Physiology, vol ii., 1900, p. 837. 

(3) GuNN, Arcliives de Pharniacodynamie, 1908. 

(4) Sherrington, loc cit., p. 887. 



ON THE RESULTS OF HETEROPLASTIC OVARIAN TRANS- 
PLANTATION AS COMPARED WITH THOSE PRODUCED 
BY TRANSPLANTATION IN THE SAME INDIVIDUAL. 
By F. H. A. Marshall and W. A. Jolly.^ (From the Physiology 
Department, University of Edinburgh.) 

(Received fur jmblication l^dh January 1908.) 

In a paper published in the Transactions of the Royal Society of Edinburgh 
we recorded the results of a series of experiments on rats in which the 
ovaries were removed from the normal position and transplanted beneath 
the skin or on to the peritoneum. In the present paper an account is 
given of certain further experiments, the results of which, on the whole, 
confirm and extend our previous conclusions. One of these experiments 
was upon a monkey, but the remainder were upon rats, the ovaries being 
transplanted on to the peritoneum or on to the tissue of the kidney. 

In the case of the latter operation the technique adopted was as follows : — 
The ovaries were removed from the normal position. An incision was 
made on the external margin of one of the kidneys, either in the same or in 
another rat. One or both of the ovaries were then placed inside the 
incision so that they were in direct contact with the highly vascular cut 
surface of the kidney. The incision in the kidney with its contained ovary 
was then sewn up with a catgut stitch and the peritoneal cavity closed. 

The following is an account of the separate experiments : — 

(1) The ovaries were removed from the normal position and grafted 
together on to the peritoneum of the same rat (homoplastic transplantation) 
in the manner described in our previous paper. Fourteen and a half 
months afterwards the rat was killed, when the grafted ovaries were found 
in position. Microscopic sections showed that the ovarian tissue was 
normal, several corpora lutea being present. The uterus was also normal. 

(2) The ovaries were removed from the normal position and trans- 
planted on to the peritoneum of the same individual. At the time of the 
operation the uterus appeared somewhat distended. After thirteen months 
the rat was killed, when ovarian tissue was found in the position of the 
graft. Microscopic examination showed, however, that it had undergone 
partial degeneration. The uterus now appeared normal, both superficiallj^ 
and liistologicall}^ 

(3) The ovaries were removed from a rat, and one of them was trans- 
planted into the right kidne}^ of another rat belonging to a different litter. 

1 Carnegie Fellow. 



^l(j Marshall and Jolly 

When the last-mentioned animal was killed two and a half months sub- 
sequently, it was found that the graft, as far as was observed, had been 
absorbed, only a certain amount of scar tissue being left in its place. The 
rat's own ovaries had not been removed. 

(4) This experiment was identical with the preceding one, the result 
being merely a persistence of scar tissue. 

(5) The ovaries were removed from a young rat and grafted together 
into the right kidney of another rat belonging to a different litter. The 
latter animal's own ovaries were not removed. Two and a half months 
later it was killed, when it was found that the grafts had been absorbed, 
their position being occupied by connective tissue. 

(6) This experiment was similar, but the grafted ovary was found to be 
entirely degenerated after a little more than one month. 

(7) In this experiment the ovaries of a young rat were grafted into 
the right kidnej^ of a male. On the latter being killed it was found that 
the graft was degenerated and palpably in process of absorption. 

(8) The ovaries were removed from a young rat and grafted into 
the right kidney of another female belonging to the same litter. About one 
and a half months afterwards this rat was killed, when it was found that 
the graft had taken perfectly, the ovaries containing normal follicles in 
various stages of development and at least two recently formed corpora 
lutea. Sections showed also that the ovarian tissue was in almost complete 
continuity with the kidney tissue in which it was embedded (see figure). 
The animal's own ovaries had not been removed. 

(9) This experiment was similar to the last, the ovaries being grafted 
into the right kidney in a whole sister. After three months the rat with 
the transplanted ovaries was killed, the organs being found in situ in a 
state of partial preservation. 

(10) An ovary from one female rat was grafted into the right kidney 
of another belonging to the same litter. The rat with the graft was killed 
three months afterwards. Microscopic sections through the kidney revealed 
ovarian tissue and follicles with ova in the periphery, but the central 
part was less well preserved. The animal's own ovaries had not been 
removed. 

(11) The ovaries were removed from a white rat. About two months 
afterwards the ovaries were removed from a piebald rat, and one of them 
was grafted into the left kidney of the white rat previously castrated. 
After another six months (or eight months after the first operation) the 
white rat was killed, when it was found that the transplantation had been 
perfectly effected. Normal ovarian tissue in abundance, and containing 
numerous ova, was observed in sections through the kidney. The uterus 
showed little or no indication of degeneration. It was evident, therefore, 
that whatever degeneration this organ may have undergone during the first 
two months after the removal of the animal's original ovaries was arrested 
by the successful ovarian graft, the uterus being restored (or almost restored) 



Oil the Results of Heteroplastic Ovarian Transplantation 117 






C.I- 



..^^' 



Section through part of kidney containing transplanted ovary. (Experiment 8.) 
, artery ; c.l., corpus luteum ; ijl., glomerulus of kidney ; </./., Graafian follicle ; o., ovum : ov. St., ovarian stroma ; 
r.t., renal tubules ; z.g.t., zone of granulation tissue between the kidney tissue on the right and the ovarian tissue 
on the left of the figure. 



118 Marshall and Jolly 

to the normal condition. It is hardly possible, however, that anything more 
than a slight uterine atrophy could have occurred in so short a period as two 
months, but our previous observations have shown that the degenerative 
process may be far advanced after six months' castration. The white rat 
into which the ovary was grafted did not belong to the same litter as the 
castrated piebald rat, and so far as known was not a relative of it.^ 

In another experiment the ovaries were removed from a monkey and 
grafted on to the peritoneum of another monkey (heteroplastic transplanta- 
tion). At the same time the ovaries of the latter were removed from the 
normal position and also grafted on to the peritoneum (homoplastic trans- 
plantation). About two months later the monkey with the grafted ovaries 
was killed, when it was found that the heteroplastic ovaries had been 
absorbed, while the homoplastic ovaries were still in position but had 
undergone a certain amount of fibrous degeneration. 



Conclusions. 

As a result of these experiments, taken in conjunction with those 
described in our former paper, the following conclusions may be 
drawn : — 

(1) Greater success attends transplantation of the ovaries into the kidney 
than on to the peritoneum, probably on account of the greater vascularity 
of the kidney. 

(2) Homoplastic transplantation of ovaries is very considerably easier 
to perform successfully than heteroplastic transplantation. This fact can 
scarcely be ascribed to differences in the technique of the two operations, 
since this was identical in each experiment, the two animals being operated 
upon simultaneously in the case of the heteroplastic transplantations. 

(3) Heteroplastic transplantation of ovaries is apparently easier to per- 
form successfully when the two animals employed in the experiment are 
near relatives of each other. In our experiments there were few exceptions 
to this rule. 

(4) The presence of an animal's own ovaries does not seem to exert any 
inhibitory influence on the successful attachment and growth of additional 
ovaries obtained from another individual. 

(5) The presence of a successfully grafted ovary in an abnormal position 
in the body, whether obtained from the same or from another individual, 
is sufficient to arrest the degenerative changes which habitually take 
place in the uterus after the complete extirpation of the ovaries, as other 
experiments have shown. It may be concluded, therefore, that the 
ovarian influence on the uterus is chemical rather than nervous in 
nature. 

1 It is possible that the two rats employed in this experiment might have been sisters 
belonging to different litters, since they were obtained from the same breeder. 



On the Results of Heteroplastic Ovarian Transplantation 119 



Literature. 

An account of the literature of ovarian grafting down to the beginning 
of 1907 is given in our previous paper. Since its publication a few further 
cases have been placed on record. Guthrie in a preliminary note has de- 
scribed certain experiments on heteroplastic transplantation of ovaries in 
fowls. The ovaries are stated to have become successfully attached, and in 
some instances to have afterwards given rise to ova which were fei-tilised in 
the ordinary way. Moreover, the chickens so produced are supposed to have 
inherited some of their characteristics from their " foster mothers " — that is 
to say, from the hens into which the ovaries were grafted. 

Some experiments on ovarian transplantation are reported in the 
" Mlinchener medizinische Wochenschrift," but very few details are given. 
Foges records a case of ovarian grafting into the spleen of a hare, which 
appears to have been partially successful; and Bucura, in commenting on 
Foges' results, states that he successfully transplanted ovaries and testes 
from guinea-pigs into a castrated female rabbit. 

Pankow reports nine cases of ovarian transplantation in the human 
subject, seven of them being homoplasts and two heteroplasts. In the 
former the menstrual periods are said to have started again at intervals of 
from three to six months after the respective operations, but there was no 
evidence that the heteroplastic transplantations were successful. 

The following case recorded by Kronig was omitted from our previous 
account of the literature: — In a woman suffering from osteomalacia the 
ovaries were removed from the normal position and transplanted on to the 
peritoneum. The result was beneficial, but with the return of menstruation, 
which occurred about two months afterwards, the symptoms of the disease 
reasserted themselves. The inference is, therefore, that the grafts were 
successful, but there was no direct evidence of the fact. 

The expenses of tliis investigation were defrayed by grants from the 
Moray Research Fund for the University of Edinburgh, and the Carnegie 
Trust for the Universities of Scotland. 



REFERENCES. 



Edges, " Ovarientransplaiitiition in die Milz," Wieu. med. Gosell., Miinehenev 
med. Wochenschr., May 28, 1907. 

Guthrie, "Successful Ovarian Ti-ansplantation iu Fowls," International 
Congress of Physiology, Heidelberg, 1907. Abstract in Zentr. f. Phys., vol. xxi., 
1907. 



120 On the Results of Heteroplastic Ovarian Transplantation 

Kronig, Discussion on Everke's paper, "Die Osteonialalde in Westfalen," 
Naturforsclierversammlung zur Stuttgart, Zentr. f. Gynak., No. 44, 1906. 

Marshall and Jolly, " Results of Removal aud Transplantation of Ovaries, 
Trans Roy. kSoc. Edin., vol. xlv., 1907. . x- • 

Pankow, "Ueber Reimplantation der Ovarien bemi Menschen, Verem J^rei- 
burger Aerzte, Mlinchener nied. Wochensclir., February 26, 1907. 



THE HISTOLOGICAL APPEARANCES OF THE MAMMALIAN 
PITUITARY BODY. By P. T. Herring. (From the Physiology 
Department, University of Edinburgli.) 

{Received for jmhlication llth February 1908.) 

Introduction. 

The structure and significance of the pituitary body have long been objects 
of much speculation. Erroneous conceptions of its structure are responsible 
for some of the many theories which have been advanced with regard to 
its functions. The pituitary, indeed, derives its name from the old idea 
that it was a gland which discharges a secretion — pituita — into the 
nostrils. 

Rathke (32) discovered the double origin of the pituitary, and on 
developmental grounds classed it among glands. Other observers looked 
upon it as part of the brain. Luschka (23) called it a "nerve-gland" 
in which the two parts are separated from one another by pia mater. 
Ecker (8), on the other hand, held the view that both portions of the 
pituitary combine to form a unit of the nature of a " blood-vessel gland." 

Burdach (4), Luschka (23), and Virchow (46) regarded the posterior 
lobe as the anterior terminal end of the cerebro-spinal canal, a "filum 
terminale anterius," resembling in structure the filum terminale of the 
spinal cord. Virchow also compared the anterior lobe to the thyroid 
gland, and described in it vesicles containing colloid material which show 
a striking resemblance to the follicles of the thyroid. Rogowitsch (34), 
H. Stieda (43), Schonemann (39), and others have attached great 
importance to this resemblance, and ascribe similar functions to the two 
glands. Removal of the thyroid is, according to their observations, 
followed by a compensatory hypertrophy of certain parts of the glandular 
lobe of the pituitary. 

In bS8() Marie drew attention to a relationship between changes in 
the pituitary and the disease acromegaly or gigantism. Clinical and 
pathological experiences have led to the theory which as.signs to the 
pituitary the role of regulating the normal development of the body, more 
especially of the extremities and bones. The nature of the change that the 
pituitary undergoes in acromegaly is uncertain, and before any light can 
be thrown upon its pathology it is necessary that the significance of the 
various histological elements that constitute the normal pituitary should 



122 Herring 

be understood. Moreover, it appears that acromegaly may occur without 
any apparent change in the pituitary, and that tumours of tlie pituitar}^ 
are not always attended by acromegaly. A feature as constant as acro- 
megaly in affections of the pituitary is the occurrence of polyuria with 
or without sugar in the urine (Hansemann (16), Sternberg (41)). 

Oliver and Schafer (28) in 1895 described the presence of a substance 
in saline extracts of the pituitary, which, when injected intravenously, 
produces a rise of blood-pressure. Howell (18) showed that this sub- 
stance is only present in the posterior lobe. Magnus and Schafer (24) 
in 1901 noticed that intravenous injection of saline extract of the 
posterior lobe is followed by a marked increase of urine flow. Schafer 
and Herring (37) confirmed this observation, and showed the striking- 
parallelism which exists between the suprarenal capsules and the pituitary 
in development, structure, and functions. In each there are two parts, one 
of which, a highly vascular epithelium, yields no active extract, while the 
other, of neuro-ectodermic origin, gives an extract which has a remarkable 
physiological effect upon the heart and arteries. The view was conjectured 
that in the epithelial part of each organ the material which is to furnish 
the active agent of the secretion passes through certain stages of formation, 
and that its production is merely completed in the neuro-ectodermic part, 
in which part alone the full activity of the secretion is acquired. That 
the posterior lobe of the pituitary should furnish an active secretion is 
difficult to reconcile with the usual views held on its structure. The older 
anatomists, W. Miiller (27), Schwalbe (40), and Toldt (45), looked upon 
it as a mass of connective tissue cells and flbres which during development 
have destroyed all trace of the original nerve tissue. Berkley (2), on the 
other hand, describes in it a complex arrangement of nerve cells and nerve 
fibres, besides neuroglia and ependyma cells. Kolliker (19) takes up an 
intermediate position, and believes that there are no true nerve cells, but 
neuroglia and ependyma, a view similar to the one held by Virchow. 
Peremeschko (30) first recognised that the posterior lobe has an epithelial 
investment. Osborne and Swale Vincent (29) state that extracts of 
the central part of the posterior lobe are more active than extracts of 
the margin of the lobe, and believe that the epithelial investment would 
be found to be inactive if it could be properly isolated. 

The pituitary body is found in all vertebrates, and, although differing 
widely in structure and in the arrangement of its component parts, possesses 
many features common to all. In fishes, the posterior lobe has a complex 
vascular structure of a glandular nature, which was called the "saccus 
vasculosus" by Gottsche (12). L. Stieda (44) proved that the saccus 
vasculosus communicates with the brain cavity, and Rabl-Riickhard (31) 
named it an infundibular gland. Their researches have been confirmed 
by Kupffer (21). The function of the saccus vasculosus is unknown, but 
its secretion, if it is a secretory gland, apparently mixes with the fluid 
contents of the ventricles of the brain. According to Kupffer, the 



Histological Appearances of the Mammalian Pituitary Body 123 

posterior lobe of the mammalian pituitary in its early development 
retains for a time a glandular structure. In the adult mammal the 
epithehal investment of the posterior lobe is regarded by Kolliker as 
the representative of an infundibular gland. B. Haller (14) states that 
in mammals — as a type of which he takes the mouse — and in all other 
classes of vertebrates the anterior lobe of the pituitary and epithelial 
investment of the posterior lobe form a gland, the tubules of which open 
by a small median and ventral mouth into the space between the pia and 
dura mater. Haller believes that the pituitary in all vertebrates secretes 
directly into the subdural space. Edinger (9) denies that this is true 
of the human pituitary, Salzer (36) could find no opening in the pituitary 
of the rat and mouse, and Sterzi (42) found none in the pituitary of 
Petromyzon. 

There are other views on the structure and functions of the pituitaiy 
body. Boeke (3) and Gemelli (11) describe appearances in the posterior 
lobe of fishes which they regard as indicative of sense organs. Cyon (6) 
looks upon it as an organ which regulates the amount of blood passing 
to the brain. Guerrini (13) and others believe that the pituitary 
produces a secretion which has a vague antitoxic action. 

Our knowledge of the structure of the pituitary body is, therefore, 
far from exact, and is inadequate to account for the physiological effects 
which follow intravenous injection of extracts, especially of the posterior 
lobe. Even the important question as to whether the glandular portion 
secretes directly into the subdural space is still unsettled. The work, the 
results of which are given in this paper, was begun with the intention of 
investigating the physiological histology of the posterior lobe, but the two 
portions of the pituitary were found to be so closely associated that no 
part would be complete without careful consideration of the other. The 
development and comparative anatomy of the pituitary body have been 
examined, but are only touched upon in this paper where reference to them 
throws light upon the particular point considered. 



Material axd Methods Employed. 

The cat furnishes some of the best material for the study of the pituitarj^ 
bodj^ for in this animal the posterior lobe retains throughout life its 
original cavity in free communication with the third ventricle of the brain. 
The structure of the posterior lobe in the cat is thus rendered simpler 
because the arrangement of the cells which line the cavity persists in the 
adult in much the same manner as obtains in the developing organ. The 
parts which are derived from the buccal mucous membrane form an almost 
complete investment for the nervous portion, and the original lumen of 
the epithelial pouch also persists throughout life in the form of a well- 
marked cleft. The so-calle*d colloid cysts are also prominent features in 
the pituitary of the cat. 



124 Herring 

The pituitary of the monkey more closely resembles that of man, and 
is a t3'pe in which greater fusion of the original elements from which 
it is developed has taken place. The posterior lobe is solid throughout. 
Its investment by the epithelial portion is not so complete as it is in the 
cat, and only a small cleft remains as the representative of the original 
buccal pouch. 

The pituitary of the dog offers in some respects a type which is inter- 
mediate between that of the cat and that of the monkey. The posterior 
lobe is solid, but the cavity of the third ventricle of the brain is continued 
downwards and backwards towards the neck of the posterior lobe. The 
epithelial investment is very complete, and the cleft in it well developed 
as in the cat. The colloid cysts are more numerous than in the pituitary 
of the monkey, and their arrangement and structure present features 
which distinguish them from those of the cat's pituitary. The morphology 
of the pituitary bodies of the cat, dog, and monkey will be described briefly, 
and the structure of the various parts more minutely detailed in the cat. 

For the investigation of the finer structure of the pituitary body 
Flemming's fixative gives the best results; a 10 per cent, solution of 
formol and saturated corrosive sublimate have also been employed. 
Sections have been cut serially in a vertical antero-posterior plane ; these 
show the relations of the various parts of the pituitary to one another 
better than do sections cut in other directions. Most of the material has 
been cut in parafiin, but the freezing microtome has also been used, and 
the Golgi preparations cut by hand. 

The structure of the anterior lobe is shown to the best advantage by 
staining with eosin and methylene blue, or by the employment of some 
of the many methods devised for the staining of blood films. Many 
preparations w^ere made by Cajal's silver reduction method, which is 
especially valuable for showing the fibrils of the neuroglia, and the 
ependyma cells of the posterior lobe. Cox's modification of Golgi's 
method was also adopted for the investigation of the nervous elements. 
Fresh tissues have been teased out and examined in salt solution and 
in osmic acid, and chromic acid fixed preparations have been cut by the 
freezing microtome. The blood-vessels were also injected from the common 
carotids with carmine gelatine, and the vascular supply of the pituitary 
body studied in thick sections. 

A word must be said about the removal of the pituitary body for 
purposes of examination. In order to investigate the question raised by 
B. Halle r as to the presence of an opening on the median ventral aspect 
connecting the epithelial cleft with the subdural space bj' means of a 
lymph space, it is almost essential to remove the sella turcica and part of 
the brain from below, to decalcify the bone and cut sections of the pituitary 
in situ. This can be more readily done in the young animal. For most 
purposes it is sufficient in the adult animal to dissect the bone piecemeal 
from the dura mater, which forms an envelope to the pituitary, thickened 



Histological Appearances of the Mammalian Pituitary Body 125 

at certain points, especially behind. Great care must be taken not to 
rupture the thin layer of epithelium which in the cat is continued back- 
wards from the anterior lobe, to be reflected at the place where the 
blood-vessels enter the posterior lobe to form a closely fitting investment 
over the ventral aspect of the latter. Removal of the pituitary from the 
cranial cavity by raising the brain and dissecting from above is almost 
invariably followed by rupture of the neck of the posterior lobe. The 
dura mater should always be- preserved intact without being pulled upon, 
and the best way to do this is to dissect off the bone from below, disturbing 




^4^lg;Xj^ 




Fiu. 1. — Mesial sagittal section through pituitary body and sella turcica of 
new-born kitten. (Semi-diagrammatic.) 
a, optic chiasma ; b, tonRue-like process of pars intermedia ; c, third ventricle ; ff, anterior lobe proper ; 
e, epithelial cleft ; /, central cavity of posterior lobe ; </, nervous substance of posterior lobe ; h, posterior 
reflection of epithelium. 

the base of the brain as little as possible. A portion of the brain can 
then be cut out with the pituitary attached, and the piece trimmed after 
hardenino-. 



Morphology. 

The relations of the anterior and posterior lobes of the pituitary to 
one another, and to their immediate surroundings, can be most readih- 
appreciated by reference to the comparatively simple pituitary of the 
new-born kitten. Fig. 1 is a diagram of a mesial sagittal section throuo-h 
the pituitary and sella turcica of a new-born kitten. The infundibulum 
cerebri is a continuation of the brain backwards and slightl}' down- 
wards, and consists of a comparatively thin wall oi brain substance 



126 Herring 

enclosing a cavity which is a continuation of the third ventricle. The 
infundibulum has a funnel-shaped origin from the base of the brain, 
narrowing as it passes backwards to a tubular neck, then expanding to 
form a hollow club-shaped body which makes up the larger portion of 
the posterior lobe. The central cavity also enlarges behind the neck 
of the infundibulum. 

The anterior lobe, composed of epithelial cells, lies below, and its thickest 
portion is in front of the infundibulum. It extends for some distance 
anteriorly, forming a tongue-shaped projection which reaches to the under 
surface of the tuber cinereum. The anterior lobe also spreads further 
laterally, and enfolds the sides of the infundibulum, the neck of which 
is encircled completely, so that, as in the figure, a portion of the anterior 
lobe appears above it. In some kittens the wrapping of the epithelium 
round the posterior lobe is more complete, and the only part of the lobe 
which is never covered by epithelium is a small part behind where the 
blood-vessels make their entrance. A narrow and somewhat S-shaped 
space lies inside the epithelium close to and following in its outline the 
under surface of the nervous portion of the posterior lobe, but separated 
from it by several layers of epithelium. The space or cleft is, as Kolliker 
(20) pointed out, the remnant of the cavity of the pouch of buccal 
epithelium from which the anterior lobe is derived. The layer of epi- 
thelium which lies between the cleft and the part of the posterior lobe 
developed from the brain is comparatively thin, and very closely applied 
to the nervous substance, thus forming an investment to it which is more 
or less complete according to the degree in which the anterior lobe has 
grown round the posterior. The cleft extends laterally, and in some cases 
almost surrounds the body of the posterior lobe. The posterior lobe as 
separated from the anterior by the cleft is therefore a composite body 
derived from the brain and from buccal epithelium, and it is to this 
structure of elements derived from two sources that the name of posterior 
lobe is usually applied, although strictly speaking the epithelial investment 
belongs developmentally to the anterior lobe. The cleft is sometimes 
more complicated, and branches of it may run into the substance of the 
anterior lobe. Serial sections show no opening below such as has been 
described by B. Haller (14), nor does there appear to be an opening at 
any point ; the cleft is a closed cavity in the kitten, but very great care 
has to be taken in the removal and preparation of the pituitary to prevent 
rupture of the thin layer of epithelium which is continued backwards 
from the anterior lobe. 

The greater portion of the anterior lobe is a solid structure made up of 
columns of cells and wide blood-channels. Granules are present in many 
of the cells of this part, but are not so marked a feature of the anterior 
lobe in the new-born kitten as they are in the adult cat. Colloid cysts are 
not found in the pituitary of the new-born kitten. 

The relation of the pituitary body to the sella turcica is shown in fig. 1. 



Histological Appearances of the Mammalian Pituitary Body 127 

The gland lies on the body of the sphenoid bone, and is separated from it 
by the dura mater, which is thickened in front and behind, and contains 
blood-vessels, and what appears to be a lymph space. This space, which 
was described by B. Haller, is not a marked feature in the kitten, but is 
more pronounced in the foetus of the ox, where, at an earlier stage in 
development, it penetrates for some distance into the body of the sphenoid 
bone. Ossification of the bone at this point is delayed by the gradual 
disappearance of the epithelial stalk which connects the anterior lobe of 
the pituitary with the buccal epithelium. No trace of this connection is 




Fio. 2. — Mesial sagittal section through the pituitary body ot" 
an adult cat. (Semi-diagrammatic.) 



a, optic chiasma ; b, tongue-like process of pars intermedia ; 

, epithelial cleft ; /, cavity of posterior lobe ; /', cavity of neck of posterior lobe : 



third ventricle ; d, anterior lobe proper ; 
fj, nervous substance 
of posterior lobe ;' i,' epithelial investment of posterior lobe. 
The dark shaiiing indicates the distribution of the characteristic cells of the anterior lobe ; the lighter shading 
shows the position of epithelium belonging to the pars intermedia. 



found in the new-born kitten, and there is no evidence of any opening of 
the glandular tubules or cleft of the pituitary into the lymph space either 
in the foetal ox or kitten. 

The pituitary body of the adult cat is ver}^ similar in structure to that 
of the kitten, but presents several important modifications. In mesial 
sagittal section the posterior lobe appears larger than the anterior; the 
latter is, however, the larger, and extends further laterally, embracing the 
posterior lobe. The central cavity of the posterior lobe persists, and a 
tapering process of it runs upwards and backwards towards the place of 
entry of the blood-vessels into the infundibulum (fig. 2). This process is 

VOL. I. — APRIL 1908. 9 



128 



Herring 



always present, and frequently runs up to the epithelial investment. The 
neck of the infundibulum is narrow and its lumen small. 

The anterior lobe and epithelial investment of the nervous portion of the 
posterior lobe are separated by the cleft, which, as a rule, persists in its 
entirety. Occasionally the cleft is closed up to a large extent, especially in 
its posterior part, and the two layers of epithelium are more or less fused, 
but a space always remains between the main part of the anterior lobe and 
the epithelial investment of the neck of the infundibulum. The main mass 
of the anterior lobe in front of the cleft contains cells holding granules 
which stain deeply with eosin. These granular cells are not present in the 
epithelial investment of the posterior lobe, nor are they found in the tongue- 
like process of the anterior lobe which runs forwards towards the optic 









Fig. 3. — Mesial sagittal section through pituitary body of an adult cat. 
(Photograph.) Compare with tig 2. 



chiasma. The epithelial investment of the posterior lobe is well ma.rked 
and thickened, especially round the neck of the infundibulum, where there 
are many layers of cells. The cells are frequently arranged in groups 
round a central lumen which contains a colloid material. These colloid 
vesicles are for the most part small and scattered at intervals ; they are 
especially well developed round the neck of the infundibulum and in the 
tongue-like process, where they show a resemblance to the vesicles of the 
thyroid gland. They are rarely found among the eosinophil cells of the 
anterior lobe, and appear to be placed never far distant from the nervous 
portion of the pituitary. 

Considerable variations occur in different cats in the relative size and 
arrangement of the parts described. This is especially the case with the 
eosinophil cells of the anterior lobe, which are sometimes continued far 
backwards over the posterior lobe, but separated always from it by the 



Histological Appearances of the Mammalian Pituitary Body 129 

cleft. At other times they end abruptly, and there is nothing but a thin 
layer of connective tissue with occasional epithelial cells in it extending 
backwards from the posterior margin of the anterior lobe to the reflection 
on to the posterior aspect of the infundibulum. The thinning and partial 
disappearance of epithelium in this situation in the adult cat may possibly 
allow a communication between the cleft and the subdural space, but there 
is no direct opening to be seen, and where the epithelium persists, as it 
often does, serial sections show that the cleft is completely closed by it. 
The readiness with which rupture may take place here is easily appreciated 




Flo. 4.— Mesial sagittal section through the j>ituitary bo<ly of 

an adult dog. (Senii-diagraminatic.) 

rt, optic cliiasnia; b, tongiie-like process of pars iiitermeiiia ; c, third ventricle; d, anterior lobe proper ; 

d\ part of anterior lobe appearing above; e, epithelial cleft; <t, nervous subsUnce of posterior lobe : 

i, epithelial investment of posterior lobe. , .^. , ^, .^, ,. 

The dark shading indicates the anterior lo))e proper ; the lighter shading sIk.ws the position of the epithelium 

of the pars intermedia. 

from the appearance in tig. 8, which is a photograph of an actual specimen 
of the cat's pituitary. 

The pituitary body of the dog (tig. 4) presents furtluT ditierences in the 
structure and arrangement of its parts. The body of the posterior lobe is 
solid, but a cavity occurs in its neck which opens by a comparatively wide 
mouth into the tliird ventricle of the brain. 

The attachments of the pituitary body to the base of the brain are very 
similar in the cat and dog. A thin lamina of brain sub-stance runs forwards 
from the neck of the infundibulum for some distance to merge with the 
tuber cinereum. This lamina is closely invested below by the tongue- 
shaped process of epithelium which runs forward from the anterior lobe. 



130 Herring- 

The posterior lamina, after leaving the neck of the infundibuluni, is sharply 
bent back upon itself, and appears in sagittal section as a long thin strip of 
brain substance tapering as it passes backwards until it joins the septum 
between the corpora niamillaria. The bend in the lamina encloses 
epithelium which is continuous with the epithelial investment of the 
posterior lobe, and which has in the cat a distinctly tubular character in 
this situation. In coronal section the opening of the neck into the third 
ventricle is not so wade, and the lateral laminae are shorter. The neck is 
really funnel-shaped, but compressed from side to side. It is completely 
invested by epithelium. 

In the dog the anterior lobe almost completely embraces the posterior, 
but the main mass of the lobe containing eosinophil cells lies below it and 
at its sides. Prolongations of the anterior lobe pass over the dorsal aspect 
of the posterior lobe to unite with one another. The epithelium is reflected 
at the neck, and at the postero-superior extremity of the posterior lobe, to 
form an investment which covers the nervous portion of the lobe. The 
reflected portion of the epithelium is separated from the outer covering by 
the cleft, which is extremely well developed in the dog's pituitary. Finger- 
like processes of the epithelium which invests the posterior lobe frequently 
project into the cleft, and sometimes join with the outer layers of epithelium, 
forming strands across it. The cleft is a closed cavity in the dog, but the 
epithelium bounding it is very thin at the posterior reflection, and conse- 
quently liable to rupture there in course of preparation. 

The investment of the posterior lobe is thick, and portions of it pass 
deeply into the nervous substance. It contains no eosinophil cells, but 
numerous colloid vesicles. The vesicles are larger in the dog than in the 
cat, and occur in groups which are for the most part situated in the deeper 
layers of the epithelium, and are not infrequently found in the adjacent 
nervous substance. 

The pituitary body of the monkey (fig. 5) presents a very different type. 
The posterior lobe is solid and the cavity of the third ventricle is not even 
prolonged into its neck. The attachment of the pituitary to the brain is 
by a narrow solid stalk of nervous substance, which is surrounded by a 
thin layer of epithelium continuous with the anterior lobe. 

The anterior lobe lies in front of the posterior, and is partly separated 
from it by the cleft. The main mass of the anterior lobe lies in front of 
the cleft, and is made up of columns of cells, many of which stain deeply 
wath eosin, and of blood -channels. The epithelial investment of the 
posterior lobe is moderately thick behind the cleft, and contains no 
eosinophil cells. It is continued round the posterior lobe and completely 
invests the neck, spreading on to the adjacent parts of the brain, but is 
usually deficient tow^ards the middle line on the posterior aspect of the 
nervous lobe. The main blood-vessels enter and leave the posterior lobe 
in this situation, and in all animals examined the epithelial investment 
stops more or less short of this place. The epithelium frequently dips 



Histological Appearances of the Mammalian Pituitary Body 131 

into the nervous substance, and strands of it may pass quite deeply into 
it, and even into the brain tissue in the neighbourhood of the floor of the 
third ventricle. The cleft is here again a closed cavity, and is not nearly 
so well developed as in the pituitaries of the cat and dog. In some cases 
very little of it remains, but the epithelium which lies between it and the 
nervous substance of the posterior lobe is always distinct from that of the 
main mass of the anterior lobe and contains no eosinophil cells. The same 




Fio. 5.— Mesial sagittal section through the ])ituitary body of au adult 
monkey. (Semi-diagrammatic. ) 
«, optic chiasma; 6, tongue-like process of pars intermedia; c, tliird ventricle; rf, anterior lobe proper; 
e, epithelial cleft: g, nervous substance of posterior lobe; i", epithelial investment of posterior lobe; 
k, epithelium of pars intermedia extending over and intu adjacent brain substance. 
The dark shading indicates the anterior lobe proper; the lighter shading shows the position of the epithelium 
of the pars intermedia. 

is true of the epithelium which invests the neck and sides of the posterior 
lobe. In this respect the epithelial investment of the nervous portion 
resembles that of the pituitaries of the cat and dog, and like them, too. 
may contain colloid-holding vesicles, but they are comparatively scarce in 
the monkey, and are not always present. 

In the pituitary body of the monkey there is, tiien, a very complete 
fusion of the tissues derived from the buccal mucous membrane and from 



132 Herring 

the brain. The cleft is rudimentary and may be almost completely closed, 
though great differences in this respect occur in difterent individuals. The 
pituitary of the monkey closely resembles that of man. The reactions to 
certain staining reagents, such as hsematoxylin and eosin, differentiate it 
into three parts : an anterior glandular, which constitutes the bulk of the 
epithelial lobe and which contains eosinophil cells ; a posterior lobe of 
nervous origin; and an intermediate portion (Edinger (9)), which is 
composed of epithelial cells closely investing the nervous portion. The 
intermediate portion, although derived from the same source as the main 
anterior lobe, differs from it in adult mammals in that it contains no 
eosinophil cells, and may exhibit the presence of vesicles resembling the 
colloid vesicles of the thyroid gland. The pituitary bodies of the ox, pig, 
and rabbit, also belong to the third type. Traces of a central cavity are 
sometimes found in the neck of the posterior lobe, but in general the 
pituitary bodies of these animals conform to the type illustrated in fig. 5. 



Structure of the Anterior Lobe. 

The anterior or glandular lobe as seen from below when attached to 
the brain is the more prominent portion of the pituitary body, surrounding, 
as it does, a large part of the posterior lobe ; it also makes up the greater 
bulk of the organ. For the sake of description, however, it is convenient 
to consider as the anterior lobe only that portion of it which has already 
been distinguished from the " pars intermedia." The part thus designated 
as the anterior lobe is separated by the cleft and by the pars intermedia 
from the nervous portion of the pituitary. It is continuous with the 
epithelium of the pars intermedia above in the region of the neck of the 
infundibulum, and behind with the thin layer of epithelium which passes 
backwards to be reflected on to the body of the posterior lobe. It is a 
solid structure made up of columns of cells separated from one another by 
large and numerous blood-vessels and a small amount of connective tissue. 
The distinguishing histological feature of this lobe is the presence in it of 
two main kinds of cells, one of which has a marked affinity for certain 
staining reagents. 

The occurrence of two kinds of cells in the anterior lobe of the pitui- 
tary was recognised by Hannover (15) in 1844, but little attention 
was bestowed upon them until the researches of Flesch (10) and 
Dostoiewsky (7), appearing independently of one another in 1884, 
definitely established their existence. Both Flesch and Dostoiewsky 
described one kind of cell possessing a large, round, or polyhedral bod}' 
full of big granules, which retain a deep red colour when treated with 
eosin and hsematoxylin, and differentiated in alcohol. These cells are 
called "chromophil" cells by Flesch. Lothringer (22) states that they 
are probably identical with the " Mutterzellen " of Luschka. The other 



Histoloo-ical Appearances of the Mainuuilian Pituitary Body 13.S 

kind of cell is small, contains a large nucleus and little protoplasm, which 
is decolourised by the same method of procedure. This variety is the 
" chromophobe " cell. The two kinds of cells occur together in strings or 
clumps, sometimes the one preponderating, sometimes the other. The 
clumps are surrounded by a basement membrane, and vary in size according 
to the number and character of the enclosed cells. The distribution of the 
two kinds varies in different animals. Dostoiewskj^ says that in man and 
the ox the clear cells are chiefly grouped together in the central part of 
the gland, while in small animals, rat, cat, and rabbit, they are more 
scattered throughout the lobe. The gland is extremely vascular, and the 
blood-vessels are of the nature of wide channels. Rogowitsch (34) calls 
the "chromophobe" cell of Flesch the " Hauptzelle," recognises the 
" chromophil " as a distinct cell, and states that a third variety exists in 
the form of nucleated masses of embryonic tissue. H. S tie da (43) comes 
to a similar conclusion and describes as "Kernhaufen" masses of em- 
bryonic tissue full of closely packed nuclei, having little protoplasm which 
behaves like that of the " Hauptzellen " to stains, and no cell borders. 
Schonemann (39) goes still further, and believes that most of the so- 
called " Hauptzellen " have no real borders, and that they are to be 
regarded as " kernreiches Protoplasma." According to Rogowitsch, 
Stieda, and Schonemann, the changes in the pituitary which follow 
removal of the thyroids are confined to the cells of the anterior lobe. 
Their results have a certain general agreement, but difler considerably in 
detail. Rogowitsch tinds colloid in the "chromophil" cells, and states 
that it passes directly from them into the blood-vessels, both of wliich 
observations are strongly combated by Stieda. The latter believes that 
thyroidectomy is followed by increase in size of the " Hauptzellen," and 
that no formation of colloid takes place. Rogowitsch describes In-per- 
trophy of the " Kernhaufen " with vacuolisation and colloid formation. 
Schonemann is of the opinion that "chromophil" cells are not a 
prominent feature of the healthy pituitary, that their development after 
thyroidectomy is a degenerative process, and, further, that they undero-o 
colloid change, accompanied by proliferation of connective tissue and 
blood-vessels. 

Saint-Remy (3.5) in 1892, after careful examination of the pituitarv 
bodies of many vertebrates, came to the conclusion that there is only one 
kind of cell in the anterior lobe, and that the varieties previousl}^ described 
are merely the expressions of different functional stages of the same cell. 
The " chromophil " cell is really a " Hauptzelle," or principal cell, in the 
protoplasm of which deeply staining granules have accunnilated. The 
granules are probably transformed into some product of secretion and 
eliminated from the cell, which then becomes a smaller body recognisable 
as a principal cell. All stages between these extreme forms may be 
recognised in the normal gland. 

Claus and Van der Stricht (5) came to similar conclusions. Benda (1) 



134 



Herri 



ring 



has more recently confirmed Saint-Remy's views. He distinguishes three 
main forms showing transitional stages. The small, poorly granular cell 
is the young form, while the large, deeply-staining granular cell marks 
the acme of functional development. A third variety is the large cell 
devoid of granules, which he regards as a cell the function of which is 
temporarily or permanently interrupted. Benda pointed out that there is 
no evidence of any of the cells being the products of degenerative changes 
as supposed by Schonemann, and further that the granules bear no 






.*??: 



^ 



0' 



■^%^'0^f(! 



>ni'- 



n^. 



^K.'--- 



^->iim^hi :,:^tf:m. 






"'-iS ■-* ■' 






m^ ■'' 



Fig. 6. — Mesial sagittal section through part of the pituitary body of an adult cat. 

a, anterior lobe showing different forms of epithelial cells : the blood channels are collapsed and 
their position indicated only by endothelial cells; 6, epithelial cleft separating anterior from 
posterior lobe; c, epithelial iuvestmeut of posterior lobe — " Epithelsaum " of Lothringer; 
d, nervous substance of posterior lobe. 

relation to the formation of colloid material. He believes that the granules 
break down into a secretion which passes directly into the blood-vessels by 
diffusion through their thin walls. 

My own observations are to a large extent confirmatory of the views 
expressed by Benda. In the anterior lobe of the cat's pituitary there 
exist three main varieties of cells : a small polygonal cell with large nucleus 
and little protoplasm, containing few or no granules ; a larger cell with 
similar nucleus and protoplasm, which may be clear, but frequently shows 
a diffuse arrangement of fine granules ; and, lastly, cells which are full of 
deeply-staining material. The latter kind of cell is sometimes smaller than 
the diffusely granular form (cf. fig. 7), but this is not always the case, and 



Histological Appearances of tlie Mammalian Pituitary Body 135 

the deeply-staining cell may be quite as large. In fig. 7 it looks as though 
there were two kinds of cells : a large, clear, and diffusely granular kind, and 
a smaller, deeply-staining cell. The picture presented by these cells varies 
according to the method and depth of staining, and differs in different parts 
of the lobe in the same section. The fixative employed has also a great 
influence on the staining reactions, and with formalin or corrosive sublimate 
fixation there appear to be only two kinds of cells : the granular and clear. 
Occasionally a cell is seen which is diffusely granular in most of its body, 
but contains around its nucleus protoplasm of the deeply-staining variety. 
It is extremely diflicult to decide whether these appearances indicate 






m jT-^-^frT^i f»'-'' 













•^ 



^^1 



':^^^w:S 



Fi( 



7. riinioi^'rapli (if part of a;;'ii i . - :. ;■ " t cat, x 500. 

Fi.xed in Flemining's solution, staiiird with hot ah?oholi(j eosiu, ditteventiated 
in alkaline alcohol, and counterstaiiied with picro-fuchsine. 

Shows clear cells and granular cells. The latter appear to be of two varieties — one kind 
consists of large cells whose protoplasm takes on a diffuse stain ; the other of smaller 
cells which are full of deeply-staininj; granules. The blood-channels are collapsed 
and their position is indicated by the dark lines. 

distinct forms of cells, or whether they are merely expressions of different 
functional stages of one and the same kind of cell. 

The different cell forms are distributed fairly uniformh" throughout the 
lobe ; the clear cells sometimes predominate in certain localities, but are not 
constantly distributed in these positions. Where anterior lobe blends with 
the epithelium of the pars intermedia, a gradual transition is sometimes 
seen between the two kinds ; at other times the dividing line is sharply 
marked (cf. fig. 8). The cells typical of the anterior lobe may spread right 
round the neck of the posterior lobe, or backwards over the body of tlie 
posterior lobe to the posterior reflection of epithelium. The general 
appearances of the relation of the cells of the anterior lobe to the cleft and 



136 Herring 

opitlielium of tlie pars intermedia is shown in %. 6, which is a drawing 
from a sagittal section through part of both lobes of the pituitary of the 
cat. In this specimen and in the one from which fig. 7 is taken, the blood- 
channels are collapsed and their position is indicated by endothelial cells. 
The cells are arranged in solid columns, between which run thin-walled 
blood-channels. The columns show no central lumen, nor is there any 
colloid met with either in the cells or between them. Where colloid is 
present it lies among the clearer cells of the pars intermedia, and never in 
relation to the characteristic granular cells of the anterior lobe. 

The cells of the anterior lobe and of the pars intermedia are derived 
from the same origin and become differentiated during fcetal life. In the 
kitten the cells of the anterior lobe do not show such marked differences in 
size and the possession of granules as they do in the adult. The clear cell 
and the granular cell are recognisable, and there are transitional forms. 
While it is almost impossible as yet to settle the exact nature of these cells, 
I am inclined to believe that all the varieties represent varying stages of 
functional activity of one and the same kind of cell, and that the deeply- 
staining material is the product of the cell destined to be poured as an 
internal secretion into the blood-vessels. 

The blood-vessels of the anterior lobe are extremely numerous and wide. 
When injected with carmine gelatine from the carotids they are seen to 
form wide channels resembling to some extent the sinusoids of the liver. 
The endothelial cells are closely applied to the epithelial cells without 
intervening connective tissue cells. In this respect also they resemble the 
sinusoids of Minot. There is, however, no evidence of any intracellular 
canalisation of the epithelial cells, such as is found in the liver (17). A fine 
reticulum of connective tissue is present in most places, resembling the 
" Gitterfasern " of the liver lobules. Whether lymphatics exist or not is 
doubtful ; the sinusoidal character of the blood-vessels and the closely fitting 
endothelial cells render their presence unlikely in many parts of the anterior 
lobe. In certain situations near the cleft and pars intermedia true capillaries 
and connective tissue are found, and lymphatic vessels appear to exist in 
these situations. 

The anterior lobe in tlie cat is usually separated from the cleft by a 
single layer of flattened cells, which are larger than endothelial cells, and 
are continuous at the anterior and posterior ends of the cleft with the cells 
of the epithelial reflection (fig. 6). 

In the dog the anterior lobe is permeated by extraordinarily large, thin- 
walled blood sinuses. Lothringer (22) compared the structure of the 
anterior lobe in this animal to cavernous tissue. In the monkey, too, the 
blood-vessels are in the form of wide, thin- walled sinuses running more or 
less parallel to one another in an antero-posterior direction. 

The changes in structure of the anterior lobe which have been alleged 
to follow thja-oidectomy in the rabbit require further investigation. The 
normal variation in structure and arrangement of the cells varies within 



Histological Appearances of the Mammalian Pituitary Body 18/ 

wide limits. Different methods of fixation and staining give very diverse 
pictures. The most useful method for showing the finer structure of the 
pituitary body as a whole is Flemming's fixative followed by Muir's 
eosin and methylene blue .stain. Some of the clear cells of the anterior lobe 
occasionally seem wanting in outline, but careful staining shows that they 
are not " Kernhaufen." Cajal's silver reduction method leaves no doubt 
that they are cells, and their outlines are readily seen when this method is 
employed. 

The anterior lobe of the pituitary is evidently an important glandular 
body, and probably furnishes a secretion which passes directly into the blood- 
vessels ; in this sense it is a blood-vessel gland, as was surmised by Ecker. 
Its function is unknown ; extracts of it, when injected into the blood-vessels, 
have no immediate physiological action beyond that common to most 
glandular extracts. It is possible that this part of the pituitary has some- 
thing to do with the regulation of the growth of the body, but in the mean- 
time there is not evidence enough to form a basis for any definite statement. 



Structure of the Ixtermedi.\te Part of the Pituitary. 

The intermediate part of the pituitary body has its origin in common 
with the anterior lobe. It arises from the epithelial pouch which grows 
inwards from the buccal mucous membrane, being a development of that 
portion of its wall which is closely applied to the nervous portion of the 
pituitary. It is separated from the anterior lobe by the cleft throughout 
a large part of its extent in the cat, but is continuous with it, in front 
round the neck of the infundibulum, and behind at the posterior reflection. 
The connection between it and the nervous portion is very intimate. 
The portion which surrounds the neck of the infundibulum shows a 
structure differing somewhat from the part which covers the body of 
the posterior lobe. In the cat the epithelium surrounding the neck of 
the infundibulum is distinctly tubular, but the lumen is not a continuous 
one. The cells are arranged round a central lumen, which frequently 
contains a colloid material. The tubules are continued forwards in the 
tongue-like process already mentioned. Between them are numerous 
large blood-vessels ; this portion of the gland is very vascular. Fig. <S 
shows the structure of the tongue-like process of the pituitar}^ of an 
adult cat, and its line of separation from the granular cells of the 
anterior lobe. 

The tubules do not appear to open into the subdural space, and are 
probably columns of cells in which lumina only appear at intervals where 
the colloid material accumulates between the cells. In the region of the 
anterior part of the cleft the tubules .sometimes appear to open into the 
latter, but their lumina are frequentl}^ interrupted. Colloid material has 
been noted by many observers in the cleft, and may enter it in this 



138 Herrino; 

manner, but the occurrence of colloid in the cleft is, in my experience, 
rare, and where found histologically has been in large, thin-walled cysts 
belonging to the epithelium of the neck of the posterior lobe.^ 

Connective tissue and lymphatics are found between the cell columns, 
and the whole structure is closely united to the under surface of the brain, 
into which blood capillaries freely penetrate. 

The epithelial cells do not show the regular arrangement which is 
so characteristic of the thyroid vesicles ; the walls are irregular and may 
be composed of one or several layers of cells. The cells are small and 
clear ; fine granules may be present m their protoplasm. The colloid 



fi^' 



^A 



f:f/^m. 



>^ 



'^^^^'^ 



Fk;. 8. — ^Mesial sagittal section through tongue-like ]ii'i((~- if (lar,^ mtemiedia 
and adjacent part of brain and of anterior lobe juoiier >.l an adult cat. 
(Photograph x 100.) 

o, third ventricle; 6, portion of anterior lamina of neck of posterior lobe; c, tongue-lilte 
process of pars intermedia, consisting of epithelial cells in form of solid columns and 
tubules : many of the latter contain colloid ; d, granular cells of anterior lobe proper. 

material varies in amount in different animals ; it does not stain deeply, 
and has a somewhat different appearance from the colloid met with in 
the thyroid. Whether it is the same kind of material or not is open to 
(juestion. Sclinitzler and Ewald (38) state that thyreo-iodine occurs 
in the pituitary. 

The junction of the intermediate portion with the anterior lobe is 

1 A pituitary body from an apparently healthy female cat, that had had a litter of 
kittens a short time previously, showed a large mass of colloid substance surrounding the 
neck of the infundibulum and occupying a large part of the cleft. The material did not 
lie free in the cleft, but was surrounded by a single layer of flattened epithelium. The 
cyst was nearly as large as the posterior lobe, and originated from the eijithelium sur- 
rounding the neck of the infundibulum. The substance in the cyst was not of a 
homogeneous nature, but consisted of irregular solid masses lying among a clearer material. 



Histological Appearances of the Maimualiaii Pituitary Body 139 

usually well defined by the occurrence of coarsely granular cells in the 
latter. The cells of this part are very like the clear cells of the anterior 
lobe, and are closely packed together in solid columns ; occasionally they 
spread downwards a little over the front of the anterior lobe (see fig. 2). 
Behind they are continuous with the epithelial covering of the posterior 
lobe, the cells of which they closely resemble. 

The epithelial covering of the posterior lobe was described by 
Peremeschko (30), who gave it the name of " Markschicht." He found 
it to vary in thickness in different situations, and to be firmly attached 
to the nervous substance. He also noted the presence of colloid vesicles 
in the " Markschicht " of the dog's pituitary, and that the cells are unlike 
those of the " Korkschicht " or anterior lobe. Lothringer (22) gave 
it the name of " Epithelsaum," a term which is used by Retzius (33) and 
later observers. Peremeschko was the first to point out that the cleft 
lies in the part of the pituitary which is composed of epithelium, and 
that it does not separate epithelial from tissue of nervous origin, as had 
previously been thought. In addition to the ordinary epithelial cells 
which are found in the "Epithelsaum," Lothringer described marginal 
cells which lie between the others and reach the free border, or are 
arched back upon themselves. Similar cells have been figured by 
Retzius (33) in Golgi preparations, and compared by him to neuroglia 
cells. According to Retzius they are, for the most part, small and 
thread-like, and reach through the whole border. Other cells do not 
pass right through, but are branched. The cell nuclei are often placed 
near the outer end, while the inner end widens to a three-cornered foot, 
which is placed against the nervous tissue of the posterior lobe. The 
structure of the epithelial border is shown in fig. 9, which is taken from 
a section of kitten's pituitary prepared by Cajal's silver method. Long, 
thin nucleated cells of a spindle shape are numerous, and take a vertical 
course through the epithelium. They appear to be of ectodermic origin, 
and act as supporting cells. Similar cells are found in the adult, but are 
better seen in the young animal. A section through the epithelial 
covering of the adult pituitary gives appearances shown in fig. 6. The 
cells are arranged in several layers over the greater part of the body 
of the posterior lobe, but are much thicker in some places than in others. 
In the cat a great accumulation of epithelial cells is found in the region 
of the lower part of the neck of the infundibulum, forming a thick mass 
between the cleft and nervous substance. Colloid material may be present 
in rounded spaces between adjacent cells in any part of the covering, but 
the vesicles are largest and most numerous in the thicker parts round the 
neck. In some cats there is a great development of the epithelium at the 
sides of the posterior extremity of the nervous lobe. In this situation there 
are distinct tubules which open into the cleft. The cells forming the walls 
of the tubules are clear and devoid of granules, and there is an absence 
of colloid material. This tubular arrangement is not always present. 



140 Herring 

and appears to be the result of proliferation of the epithelium at the 
end of the cleft, a lumen being retained in the outgrowths continuous 
with the original cavity. 

In the dog large vesicles are frequently seen in the epithelium ; they 
are larger than those of the cat's pituitary, and often occur in groups. 
They are for the most part deeply situated and often separated from 
the nervous substance by a single layer of flattened cells, and even these 
may be deficient, so that the colloid material is partly bordered by the 
nervous substance. In the monkey the vesicles are less numerous, and 
occur for the most part in the epithelium at the ends of the cleft. 

The cells which make up the greater part of the epithelial border are 
well-defined polygonal cells resembling somewhat the clear cells of the 




Fig. 9. — Vertical section througli e])itlielial investment and nervous substance of posterior lobe 
of new-born kitten. Prepared by Cajal's method. 

The epithelial investment contains numerous spinille-shaped cells, the marginal cells of Retzius ; the nervous 
substance is composed of neuroglia cells and granular-looking matrix. Numerous blood capillaries are 
found in the nervous substance, but do not penetrate into the epithelium. 

anterior lobe, but they stain rather more deeply and their protoplasm is 
more granular. Occasionally deeply-staining cells are met with, but they 
never show the eosinophil granules of the cells of the anterior lobe. The 
colloid material appears to be a product of the epithelial cells, and 
accumulates between adjacent cells, giving rise to the appearances of 
vesicles. The spaces thus formed may communicate, but are for the 
most part closed. 

The epithelial investment of the posterior lobe presents another notable 
feature which distinguishes it from the remainder of the epithelial part of 
the pituitary^ — it rarely contains blood-vessels. In the pituitary of the 
cat no vessels occur in the thinner portions, and where the epithelium is 
much thickened a few capillaries only enter it from the adjacent nervous 
substance. Large blood-vessels, chiefly veins, and capillaries are very 



Histological Appearances of the Maiinnalian Pituitary Body 



141 



numerous at the junction of the epithelium with the nervous portion (cf. 
fig. 9). In the dog, capillaries are rather more numerous. In the monkey 
they are absent. A very striking difference is thus afforded between the 
highly vascular epithelium of the anterior lobe and the non-vascular 
epithelium which covers the posterior lobe. This peculiarity indicates a 
difference between the two parts, if not of function, at any rate of the mode 
of absorption of the products of the epithelial cells. Another significant 
feature relates to the position of the colloid material. This occurs for the 
most part in extra-vascular situations, and never far distant from the 
nervous portion of the pituitary. If it is a product of secretion utilised by 
the animal, it must pass inwards to the nervous tissue and be carried away 
by the numerous blood-vessels situated immediately under the epithelium, 
or pass into lymph spaces in the nervous portion. This point will be dis- 
cussed later in connection with the structure of the nervous portion. In 
the other place, where colloid material is common in the cat, viz. in the 
tongue-like process (fig. 8), the epithelium containing it is surrounded by 
connective tissue, blood-vessels and lymphatics, and even here a close rela- 
tion exists between the epithelial cells and adjacent brain substance. The 
structure of the intermediate part of the pituitary body shows such marked 
differences from the structure of the anterior lobe that it is quite probable 
it has a different function. 



Structure of the Nervous Part of the Pituitary. 

On no part of the pituitaiy body have there been so many and different 
opinions expressed with regard to structure as on the portion derived from 
the brain. Virchow (46) regarded it as a "filum terminale anterius," 
consisting of ependyma cells, networks cf white fibres, and finely 
granular masses in which cells appear. He could find no nerve cells in it. 
Peremeschko (30) described a feltwork of connective tissue fibres and 
spindle-shaped cells, with a few ganglion cells among them. The latter, 
he says, lie mostly two or three together surrounded by "connective tissue, 
and differ from ordinary ganglion cells in that their protoplasm is scantier 
and their nuclei flattened. W. Miiller (27) and Mihalkovics (25), who 
studied the development of the pituitary, agreed that with the ingrowth of 
blood-vessels into the nervous portion, a proliferation of connective tissue 
cells accompanying them destroys the original brain tissue, converting the 
infundibulum into a connective tissue appendage of the brain. W. Miiller 
(27) compared the arrangement of the connective tissue cells to the appear- 
ances shown by a spindle-celled sarcoma, a simile which has been frequently 
employed since by authors of text-books of anatomy. Loth ringer (22) 
described the bulk of the organ as made up of bundles of fibres, with nuclei 
resembling those of plain muscle fibres, crossing one another at sharp angles, 
and holding in the meshwork thus formed round, angular, or polygonal 
cells. He believed it to be chiefly neuroglia tissue, and poor in nerve cells. 



142 Herring 

Berkley (2) examined the nervous lobe of the pituitary of the dog, em- 
ploying Golgi and other methods. He gives a description and diagram of 
the posterior lobe of the dog's pituitary, and divides the organ into three 
parts, all of which are microscopically dissimilar in appearance. " There is 
first an outer lamina of slightly irregular ependymal cells, three or four 
deep, arranged after the manner of the cuticular epithelium, separated into 
divisions by thin processes extending from the fibres of the surrounding 
capsule and terminating at a definite line where a separation from the 
more internally lying elements occurs ; then follows a more internal zone 
of varying depth, containing epithelial cells of a secretory type, which in 
places along the posterior and inferior border of the lobe, as well as 
occasionally in the more central regions, are arranged into distinct closed 
acini lined with a low variety of cylindrical epithelial cell, and often hold 
in their lumen collections of a colloid substance." Berkley in this passage 
clearly describes as ependymal cells the long thin cells of the epithelial 
investment, the nature of which has already been discussed. Berkley 
further states : " The secretory region gradually merges into a central region 
of small, rounded, and polygonal cells separated by extensive connective 
tissue partitions, carrying blood-vessels, and scattered widely among these 
cells are others of larger dimensions. Some of the latter are of spindle 
form, others of pear or rounded shape, and still others of very irregular 
form with their borders ill-defined ; all having a finely granular appearance, 
with here and there larger granules scattered among them, that are tinged 
by osmic acid a blackish colour. In the region of the neck of the infundi- 
bulum the epithelial elements, except the outer ependymal row, become 
segregated into groups of mainly oval and pear-shaped cells separated by 
a fine stroma, with small nuclei here and there in it, and are now easily 
recognised as nerve cells. A certain number of the spindle cells are very 
long as well as broad, and probably correspond to those described by 
Krause as spindle cells of uncertain function, but they are undoubtedly 
nerve cells." Berkley gives a correct description of the structure of the 
lobe, but with his statement that the epithelial cells become nerve cells, 
I am quite unable to agree. He finds several varieties of neuroglia cells, 
chiefly of the moss and spider type. No less than six varieties of nerve 
cells, including large, medium, and small pyramidal cells, are described by 
him as occurring in the posterior lobe of the pituitary. The axis cylinders 
have a general tendency to pass upwards and forwards. Berkley con- 
cludes that " the pituitary gland retains in the dog, as one of the highest 
orders of vertebrates, its double role of secretory and nervous functions, 
intact ; the former perhaps modified, the latter, the original special sense 
organ, probably lying quiescent, not atrophied, and only changed in so far 
as to admit of a slightly different arrangement of its constituent elements." 
Berkley is quoted at length, because it is on the result of his work that 
the belief in the presence of true nerve cells in the pituitary is mainly based. 
Retzius finds no true nerve cells, and no medullated fibres, and inclines to 



Histological Appearances of the Mammalian Pituitary Body 143 

the opinion that, although nerve fibres may be present, the bulk of the 
posterior lobe consists of neuroglia cells and fibres, and ependyma cells. 
Kolliker (19) takes up the same position, and holds that the apparent 
nerve cells are really neuroglia and ependyma ; many of the fibres of the 
latter run for a considerable distance in a longitudinal direction, and form 
thick bundles. Kolliker agrees with Retzius, that, in man and the 
higher mammals, there are no true nerve elements in the infundibular lobe, 
and that the occurrence of glandular structures in it betokens the formation 
of an infundibular gland in the sense of Kupffer. 

The posterior lobe of the pituitary body, when fixed in formalin, cor- 
rosive sublimate, or other fixing agent in common use, and stained in thin 
sections by hsematoxylin and eosin, presents in its interior a structure which 
has a general resemblance to connective tissue. There are numerous 
flattened, spindle-shaped, and branching cells, and a matrix large in amount, 
finely granular, and holding numerous ill-defined fibres resembling white 
fibrous tissue. Blood-vessels are fairly numerous, especially small arteries 
and capillaries, and the tissue of the nervous lobe frequently appears to be 
arranged in whorls around them, a layer of condensed matrix lying next 
the vessels, and outside that a lighter zone anastomosing with similar 
layers around adjacent vessels. This disposition of the matrix is very well 
brought out by the iron-alum-heematoxylin method of staining. A some- 
what similar arrangement is seen in the neck of the infundibulum, where 
there are two very distinct layers surrounding the central cavity, an inner 
layer in which run longitudinally placed fibres, and an outer layer, which 
is finely granular and apparently traversed by much finer fibres, having for 
the most part a vertical direction. 

The cells in the posterior lobe do not resemble true nerve cells, and 
when the sections are stained by Nissl's method nothing of the nature of 
Nissl bodies is found in them. In sections prepared by Cox's modification 
of Golgi's method, the nervous substance is seen to be composed of neu- 
roglia cells and fibres (fig. 10). 

The neuroglia cells are numerous, and their branches interlace, forming a 
dense network of fibres througliout the body of the infundibulum. Many 
of the branches end in relation to the blood-vessels, which they frequently 
bend round ; others run to the periphery of the lobe, but do not penetrate 
into the epithelial investment. In addition to the neuroglia cells and fibres, 
there are cells lining the central cavity, continuous through the neck of the 
infundibulum with the cells that line the cavity of the third ventricle. 
They are undoubtedly ependyma cells, and are extremely well developed in 
the pituitary. The fibres from the cells take various directions, according 
to the position of the cell bodies, but most of them assume a longitudinal 
direction, and pass into the neck of the infundibulum. In this .situation 
they at first run in the internal layer, then turn into the outer layer, and. 
breaking up into extremely fine fibrils, take a course at right angles to theii- 
previous direction, and end at the margin of the neck in proximity to the 

VOL. I. — APRIL 1908. 10 



144 



Herring 



epithelial columns surrounding it. The arrangement of these fibres is not 
easily made out in Golgi preparations, but can be more readily followed in 
thinner sections prepared by Cajal's reduced silver method. Cajal's method 
shows that the posterior lobe is pervaded by a dense network of fibres which 
agree in their manner of disposition with the arrangement revealed by the 
Golgimethod,but are more uniformly stained than by the latter. Theependyma 
fibres are large and thick in the adult cat, especially near their origin from 




/ V 



Fir;. 10.— Section of part of nervous substance of posterior lobe of the pituitary body 
of an adult cat, showing neuroglia cells and fibres. Prepared by Cox's modification 
of Golgi's method. 

the cell. They appear to begin abruptly in the cell protoplasm, often by 
several rootlets, which soon fuse to form a single process or remain separate. 
In the body of the posterior lobe the ependyma fibres take various courses. 
Some run outwards and are lost in the neuroglial network. At the posterior 
end of the lobe where the central cavity approaches the epithelial covering, 
fibres may be traced outwards to end immediately under the latter. Most 
of the ependyma fibres, and especially those of the anterior part of the lobe 
and the neck, take a longitudinal direction, as though passing upwards and 
forwards into the brain. The two layers already mentioned as occurring in 
the neck of the infundibulum are particularly well defined by Cajal's 
method. Fig. 12 is a drawing of a longitudinal section througli part of the 



Histological Appearances of the Mammalian Pituitary Body 145 




Fk;. 11.— Diagram showing arrangement of the ei)endyma cells lining the central 

cavity of the posterior lohe of an adult cat. 

The ependynia cells send their processes forwards into the neck of the posterior lobe. Drawn 

from a section prepared by Oajal's method. 



'■J-" •>: yr ^■'''\ '^'yr^^:^^'^S^'\^'' -■'^'■ 






.«.jv.si>>s?i;_ d 




^i./r^ i:- '. 



i;- .0^^-!;^ — a 



>^;}^;?;. ^'^ "^- 'Vh -■■" * '■ '■ '' " '" ' 



Fig. 12.— Mesial sagittal section through neck of the posterior lobe 
of a kitten. Drawing from a section prepared by Cajal's method. 

a, epithelial cells of pars intermedia surrounding neck; 6, outer layer of 
nervous substance of neck ; c, inner layer of ependynia fibres ; d, ependyina 
cells lining neck ; c, central cavity of neck. 



146 Herrinoj 

neck of tlie posterior lobe of a kitten in which the two layers are shown, 
surrounded by the epithelial columns of the intermediate portion of the 
pituitary. 

The distinction between the two layers begins at the junction between 
body and neck of the posterior lobe, and is continued throughout the 
laminae which connect it to the brain. In mesial sagittal section the ap- 
pearances of a Cajal preparation are very striking. The finely granular, clear 
outer layer extends anteriorly to the tuber cinereum, increasing in size as it 
passes forward. The inner layer of longitudinally running fibres diminishes 
in thickness, but is quite well marked where it fuses with fibres and cells 
of the tuber cinereum. In the posterior lamina the two layers are eijually 
distinct, but the longitudinal fibres are fewer in number, and both layers 




Fig. 13. — Mesial sagitt:il section through anterior lamina of nervous 
substance in floor of third ventricle of adult cat. Drawing from 
a section prepared by Cox's modification of Golgi's method. 

a, ependyma cells ; 6, inner layer of horizontal fibres ; c, outer layer of vertical 
fibres from ependyma cells; d, position of cells of pars intermedia of tongue-like 
process. Cf. flg. 8. 

become very thin towards their junction with the brain substance in the 
region of the corpora mamillaria. The cells which line the cavity and 
neck of the infundibulum vary somewhat in appearance, according to the 
methods of fixation and staining employed. In the kitten, as revealed by 
Cajal's method, there may be more than one layer of rounded cells from 
which, as in fig. 12, fibres arise. None of the cells are ciliated in the cat. 
In the adult there is one layer of cells, but others often lie among the fibres, 
and appear to give origin to some of the latter. The anterior lamina in 
front of the neck of the infundibulum is lined by cells, which, when 
prepared by Golgi's method, have the appearance shown in fig. 13. Thej^ 
are well-developed ependyma cells, and their branching processes pass 
outwards, giving rise to the vertical striation which in Cajal and other 
preparations is characteristic of the outer layer of the neck of the infuijdi- 
bulum. In Golgi preparations these cells appear to give off" processes which 



Histological Appeai-ances of the Maninialian Pituitary Body 147 

nm with the other longitudinal fibres in the inner layer. Similar appear- 
ances were described by Berkley, but it is almost impossible to decide 
whether longitudinal hbres arise from these cells or not ; they may be fibres 
horn cells further back, which cross the cell processes in this situation. 

The layers of the neck of the infundibulum have been described by 
several authors. Mihalkovics (25) noted them in the developing pituitary, 
but believed that the inner layer is composed of connective tissue fibres, 
arising from cells which enter the posterior lobe with the blood-vessels. 
Lothringer (22) described them, and stated that the layers are prolonga- 
tions of the tissue of the tuber cinereum into the posterior lobe, neuroglia 
fibres increasing and nervous elements diminishing from before backwards. 
According to Lothringer, the longitudinal fibres have their origin in the 
tuber cinereum. This view has been adopted by most subsequent authors 
except Berkley, who, on the other hand, regards them as nerve fibres 
arising from cells in the posterior lobe, and passing upwards and forwards 
to the brain. Cajal (quoted from Kolliker's " Gewebelehre," Bd. ii., 
S. 604, 1896) found in two-days-old mice nerve fibres originating in the 
tuber cinereum, and passing downwards into the body of the posterior lobe. 
In the lobe itself he described one of the thickest plexuses of nerve fibres 
known, and cells of a doubtful nature. 

The fibres are not medullated nerve fibres, and do not stain as such with 
osmic acid in fresh teased preparations. They are much thicker than 
ordinary non-medullated fibres, and are devoid of nuclei along their course. 
Many of them resemble white fibrous tissue, but they are stained by Cajal's 
method, although not so deeply as the nerve fibres in the adjacent brain 
substance. The posterior lobe of the pituitary body of the cat, owing to 
the persistence throughout life of its original cavity, is a particularly good 
subject for the elucidation of the question of the origin of these longitudi- 
nally running fibres. In this animal the fibres clearly arise from cells lining 
the cavity of the body and neck, and pass forwards and upwards towards 
the brain. They are the fibres of ependyma cells. After a course of varying 
length, they break up into fine fibrils which enter the outer layer and ter- 
minate on its external surface. The origin of the fibres can be traced in the 
developing organ, and their elongation and oblique course explained. The 
cell bodies, from which the fibres proceed, move downwards and backwards 
with the growth of the infundibulum, but the outer ends of the fibres remain 
attached to the junction with the epithelial part, and do not participate in 
the movement. New fibres arise during growth, some of them apparently 
taking origin from cells which lie deeper and not lining the cavity. In the 
front portion of the anterior lamina most of the ependyma fibres run verti- 
cally, and have only a very short course (fig. 13). There are numerous cells 
in the tuber cinereum, and fibres pass into or out of this body ; it is difficult 
to tell where their cells of origin lie. Some fibres may arise in the tuber 
cinereum, but they resemble the ependyma fibres, and probably belong to 
ependyma or neuroglia cells. 



148 Herring 

Similar bundles of fibres are found in the solid neck of the posterior 
lobe of the monkey's pituitary. Here, too, the cells of origin lie below, 
and the fibres run upwards; they must be regarded as ependyma cells 
which have become enclosed in the body of the organ. 

In the body of the posterior lobe the ependyma and neuroglia fibres 
make up a very thick network. The presence of true nerve cells in the 
pituitary is very doubtful. In preparations made by Cox's modification 
of Golgi's method the cells stained resemble neuroglia cells and are all of 
the same type. That this is not an accident is shown by the uniformity 
of the results in different specimens, and by the fact that when the sections 
include brain substance the true nerve cells in the latter are well and 
characteristically shown. The neuroglia cells, like those of adjacent parts 
of the brain, are spider cells with very numerous processes. In Cajal 
preparations the fibres run through the cell, the body of which is often 
difficult to distinguish. 

The nerve supply of the glandular portion of the pituitary body is stated 
by Berkley to be derived from the sympathetic system. Branches enter 
the gland with the blood-vessels, and end among the epithelial cells. Cajal's 
method shows fine fibres entering amongst the cells of the epithelial invest- 
ment, especially in its thicker portions near the neck of the infundibulum. 
The fibres come from the nervous portion, where they are closely associated 
with the blood-vessels. I have not been able to trace their origin, but for 
developmental reasons am inclined to believe that they accompany the 
blood-vessels into the gland, and are, as Berkley states, derived from the 
sympathetic system. Neither by Golgi's nor Cajal's method have I been 
able to find nerve fibres in the anterior lobe, but I have not specially 
investigated this point. 

The nervous tissue of the posterior lobe of the pituitary appears, then, 
to have the structure assigned to it by Virchow in 1857, and to be made 
up of neuroglia and ependyma cells and fibres. There are, however, other 
very important elements present. The epithelial covering is in contact 
with the nervous portion from an early stage of development, and grows 
around it. But the relation becomes still more intimate, for the epithelial 
cells invade the nervous portion. This ingrowth may take place at any 
part of the lobe, but, in the cat's pituitary, is most marked in the region 
of the neck of the infundibulum, and at the posterior reflection of 
epithelium. In the latter situation the epithelial cells frequently ac- 
company the blood-vessels for some distance into the lobe. Strands of 
cells retaining their connection with the epithelial investment are often 
seen passing into the nervous substance ; at other times cell islets of a 
similar nature are found at variable distances from the epithelial covering. 
The islets consist of well-defined epithelial cells, and in preparations fixed 
in Flemming's solution are vevy distinct. They react to stains in a 
manner which identifies them with the cells of the epithelial covering. In 
the adult cat they are often very numerous, especially in the neck of the 



Histological Appearances of the Mammalian Pituitary Body 141) 

infundibulum ; but they may occur at any part, and are not infrequent in 
the middle of the posterior lobe, lying partly in the cavity. The islets vary 
in size from a few cells to twenty or more, and are usually compact, but 
in the neck of the infundibulum have a looser structure, and individual 
epithelial cells occur among the fibres of the nervous portion. These are 
probably the cells referred to by Berkley, but they are not nerve cells; 
they are derived from the epithelial investment. Occasionally a pituitary 
body is met with in which the posterior lobe is quite transformed by the 
ingrowth of epithelium. One of the most remarkable features about the 
epithelial investment of the pituitary is the property it possesses from a 
ver}^ early stage of development of spreading over and around the structures 
with which it comes into contact. It also tends to inv^ade them, and may 
even spread for some distance into the base of the brain in some animals. 
In the monkey (cf. tig. 5) the cells often penetrate towards the cavity of 
the third ventricle behind the neck of the infundibulum. The same is true 
in the case of other animals ; and it is not confined to mammals — a similar 
invasion is common in birds. The extent of the ingrowth shows great 
variations in different individuals of the same species. Careful and good 
fixation of the tissue is necessarj^ to show it ; for, if any shrinkage takes 
place, the epithelial cells are more difficult to distinguish from the other 
cells of the posterior lobe, and are easily overlooked. When the islets are 
large and deeply placed they are readily seen. 

In well-fixed preparations, and especially after fixation by Flemming's 
solution, the posterior lobe is found to contain small hyaline bodies, highly 
refractive when unstained. These bodies lie scattered throughout the 
nervous substance of the lobe, and stain indifferently with eosin or 
methylene blue. They are not so prominent in formalin fixed preparations, 
but can be seen. They are not found in the pituitary of the new-born 
kitten, but in adult pituitaries are invariably present, and in all animals 
examined. They occur in the bod}^ of the lobe and in the neck, extending 
for some distance upwards towards the brain, but not into the brain itself. 
Their appearance suggests that of cells which have undergone hyaline 
degeneration ; no nucleus is present, the outline is often irregular, and no 
structure can be made out in them. Sometimes, however, they have a 
distinctl}' granular appearance, and are not unlike the granules of the cells 
of the anterior lobe, but do not stain so deeply. The significance of these 
bodies is difficult to determine. They seem to be of the nature of a 
secretion, and are not unlike diluted colloid material. In some situations 
the substance lies in what look like lymph spaces lined with endothelium. 
The question naturally arises as to whether this material represents the 
physiologically active principle of tlie posterior lobe ; its situation certainly 
agrees with the position in which that is found. The substance often lies 
between the i^pendyma cells near the central cavity, and may possibly be 
a secretory ])r()duct of these cells, in which case the nervous part of the 
pituitary might be regarded as a glandular structure, but it is not alwaj'-s 



150 



Herrinj 



confined to the nervous part, and occurs occasionally among the cells of 
the epithelial investment, especially where that is thickened just below the 
neck of the infundibulum. This fact points strongly to its being a product 
of the cells of the epithelial investment. The material is most abundant 
in the neck and around the central cavity and in the neighbourhood of 
the epithelial islets. 




> t 


d 


e 


c 




^ 






*^' 








V -:. &' 


'■ :V:n.-- 




1;,:;^ 


> : 













'K 





m 






Fig. 14.— Drawing of part of the posterior lobe of the pituitary body of an adult 
cat. From specimen fixed in Flemming's solution and stained with eosin and 
methylene blue. 

a, cells of epithelial investment ; b, granular body ; c, e, islets of epithelial cells similar in character to 
those of the epithelial investment ; d, colloid or hyaline body ; e, ependyma and neuroslia fibres. 

The drawing is from part of the section in the neighbourhood of the neck of the posterior lobe. 
Granular and colloid bodies occur in close proximity to the epithelial cell islets. 

The accompanying figure (fig. 14) is a drawing of the appearances seen 
in a portion of the posterior lobe of an adult cat near the neck of the 
infundibulum. Several epithelial islets and scattered cells lie among the 
neuroglial and ependyma fibres. In close relation to the cells are seen 
masses of a hyaline or granular character. The .substance often stains very 



Histolot^ical Appearances of the Mammalian Pituitary Body 151 

like the so-called colloid material, and may be of that nature ; but is unlike 
in many respects the colloid of the thyroid gland. In the posterior lobe 
of the dog's pituitary somewhat similar appearances present themselves, 
but in this case the epithelial islets frequently consist of a number of cells 
which group themselves round a central cavit}^ containing colloid material. 
Isolated cysts, the walls of which are composed of a single layer of cells, 
are not uncommon, and were noted by Loth ringer. In the dog, too, the 
larger cysts of the epithelial investment are not always complete, and the 
contained material may abut against the neuroglial tissue. The colloid 
substance, when completely or partly enclosed by epithelial cells, varies 
considerably in its staining properties in different parts of the same 
pituitary. As a rule it takes on little depth of colour with stains, and 
is unlike the colloid of the thyroid in this respect ; it has a hyaline rather 
than a colloid appearance. Occasionally, however, the substance is denser 
and takes on a deeper stain, and in the dog sometimes looks as though it 
were a swollen cell with disintegrating nucleus. The hj'aline material of 
the nervous portion of the posterior lobe also varies in appearance, and 
in its staining properties ; most of it might be of the same nature as the 
so-called colloid of the epithelial investment and of the intermediate part 
generall}^, but in a diluted form. Some of this material must be the 
product of the epithelial cells, for it occurs in places where no other kind 
of cell is present. The universal occurrence of this material in the nervous 
substance, often at considerable distances from epithelial cells, is difficult 
to explain, unless we can suppose it to be carried from them by lymphatic 
vessels. The substance often does lie in distinct spaces lined by w^hat 
appears to be endothelial cells. The general tendency of the direction of 
the material seems to be towards the neck of the infundibulum in the cat, 
and it increases in amount towards this situation. Large masses are some- 
times seen lying among the ependyma cells, and similar material ma}^ be 
present in large amount in the central cavity, connnunications between the 
two being evident in places. 

In fig. 15 a typical portion of the neck of the infundibulum of a cats 
pituitary is seen. The central cavity is lined by ependyma cells, outside 
whicll are cells with large nuclei and little proto])lasm. Occupying the 
central cavity is a mass of hyaline material, and masses of a similar 
substance lie beneath the ependyma cells and between the ependyma 
fibres. In other places in the posterior lobe the material is distinctly 
present in lymph channels accompanying the blood-vessels. Evidence 
strongly points to the probabilit}' that the material is on its way to the 
central cavity, and so into the ventricles of the brain. In this sense the 
posterior lobe of the pituitary is a gland which pours its secretion into the 
third ventricle of the brain. It is possible that the ependyma and neuroglia 
cells have also a secretory function, but improbable that they secrete tlie 
material described. They may, however, have some influence upon it. 
The most likely supposition is that the ependyma and neuroglia cells 



152 Herring 

rorni a scaffolding for the posterior lobe of the pituitary, upon which is 
built up a covering of epithelial cells.^ The relations of the two structures 
Ijeconie very intimate ; their blood supply is derived from arteries which 
enter the nervous substance posteriorly ; the veins begin immediately below 
the epithelium, and return through the nervous substance. Interchange 
of material between blood-vessels and cells must be through the medium 
of lympli, seeing that most of the epithelial investment is extra- vascular. 
Cells from the investing layer grow into the nervous framework, giving 
rise to epithelial columns and cell islets. Secretion goes on either by an 
emptjang of material from the cells into the lymph, or possibly by a 
breaking down and destruction of the wliole cell. The latter indeed is 






''^•1 .^ -.T> *^^ '^t>'^ '>**♦ '-^ 



O' ^ ^ ^* 



;••;• -ssr- 



Fig. 15. — Section of part of neck of ])osterior lobe of the pituitary body of an adult cat. 
From specimen fixed in Fleraming's solution and stained with eosin and methjdene blue. 
Shows colloid materia,l lying among ependyma fibres and in the central cavity of the neck. 

a, ependyma cells lining central cavity; b, cnlloid material in central cavity; c, ependyma fibres; 
d, colloid material; e, granular body lying among ependyma filires. 

the more probable fate of isolated epithelial cells, and seems to occur at 
times in the epithelial investment itself. The material known as colloid 
substance, which has been supposed by many to be identical with the 
colloid of the thyroid gland, occurs in comparatively large amounts in the 
pituitary of the cat, an animal which cannot long survive thyroid extirpa- 
tion. Further chemical and experimental research is necessary to prove 

' In the posterior lobe of a Belgian hare, a ganglion with large ganglion cells and 
medullated fibres was found. Tiie ganglion occupied a large portion of one side of the 
posterior part of the lobe, and the fibres appeared as though entering it obliquely from the 
side and not through the neck of the lobe. Red bone marrow was also present, and the 
ganglion may have been part of one of the Gasserian ganglia, which, in the rabbit, are very 
clo.se to the pituitary body. A large part of the nervous portion of the lobe was destroyed 
by it. The intermediate portion was large in amount, and its cells thickly massed around 
the neck. The animal was a healthy adult. 



Histological Appearances of the Mammalian Pituitary Body 153 

whether the colloid substance of the pituitary is identical in composition, 
or in its physiological action, with the colloid of the thyroid gland. The 
physiological action is apparently quite different, but this may be due to 
the presence in the lobe of other substances. Any secretion formed in the 
pars intermedia of the pituitary must pass into the adjacent substance of 
nervous origin either by blood-vessels or lymphatics. There is a possibU- 
exception to this rule in the tongue-like process of the cat's pituitary, but 
even here blood-vessels from it pass into the adjacent anterior lamina 
connecting the neck of the infundibulum with the tuber cinereum, and tla- 
tubules are surrounded by connective tissue containing l^n^nphatics, the 
course of which is unknown. They may possibly accompany the blood- 
vessels. The epithelial cells do not resemble in staining properties the 
cells of the medulla of the suprarenal capsule ; they have no affinity for 
chromic acid. Nor do they resemble them in their relations to blood- 
vessels, the absence of which is so characteristic of most of the pars 
intermedia. 

A question of importance arose in the earl}" investigations of the 
structure of the pituitary. Peremeschko (30) described the cleft of the 
epithelial part as being continuous with the central cavity of the neck of 
the infundibulum, and so with the third ventricle of the brain. If this 
were the case it would furnish a proof of Kupffer's view (21) that the 
epithelial portion represents a " palseostoma " or old mouth of an ancestral 
form of vertebrate, and there would be in the mammalian embryo a 
communication between the neural canal and buccal cavit}^ Subsequent 
observers have denied the accuracy of Peremeschko's observations. In 
the pituitaries of the pig and man, in which Peremeschko described the 
continuation, there is obviouslj^^ no such thing, for the body of the infundi- 
bulum is solid in both cases. Nor is there any indication of it in the 
embryo of either pig or man, so far as I have been able to see. It is far 
more likely to occur in the pituitary of the cat, in which the central cavitj* 
of the infundibulum is well developed and prolonged far backwards. In 
the adult cat the epithelium, as already stated, frequentl}' invades the 
posterior end of the cavity, so that epithelial cells may even form part of 
its lining. This peculiarity affords some support for Kupffer's view. In 
the adult cat I have never been able to find a direct coinmunication 
between the cleft and the cavity of the infundibulum, although many 
specimens have been examined with this object. In one embryo kitten, 
however, I have found such a connnunication. Cleft and cavit}' in this 
specimen are undoubtedly in direct continuity at the postero-superior 
angle of the posterior lobe. In other kittens, at a comparativeh' late stage 
of embryonic life, a direct continuity' is occasionally seen between tubular 
epithelium at the end of the cleft, and the ependyma cells lining the 
central cavity of the infundibulum. In the kitten this coming together 
of the two portions occurs some time after the epithelial duct between 
buccal mucous membrane and epithelial portion of the pituitary has 



154 



H( 



disappeared, so that there is never a direct continuity between the neural 
canal and the exterior. The observations, nevertheless, give support ta 
Kupft'er's views on the morphological significance of the pituitary. 



Vascular Supply of the Pituitary Body. 

The arrangement of the blood-vessels in the pituitary body has already 
been described along with the structure of its several parts, but a general 




Fi<;. 16. — Mesial sagittal section of pituit<iiy body of adult cat ; blood-vessels 
injected witli carmine gelatine. (Photograph.) 

a, ojitic cliiasnia ; 6, tongue-like process of pars intermedia; c. third ventricle; d, anterior lobe; e, pars 
intermedia lying above neck of posterior lobe; /, posterior lobe ; <;, central artery entering posterior 
lobe at its postero-superior angle ; h, large vein lying between nervous substance and epithelial invest- 
ment of posterior lobe. 

survey of the vascular distribution in the cat s pituitary may be given.. 
The most noticeable feature of the injected organ is the difference in vascu- 
larity of the two lobes. The anterior lobe is filled with wide channels, making 
it one of the most vascular structures of the body ; the posterior lobe, on 
the other hand, resembles in the number and arrangement of its vessels the 
adjacent white matter of the brain. A large blood sinus is found on either 
side of the pituitary body below. The anterior lobe is supplied by arteries,, 
apparently from the internal carotid, which enter it at the sides of the 
infundibulum above, and break up immediately into large, thin-walled' 



Histological Appearances of tlie Mammalian Pituitary Body 155 

vessels. The posterior lobe is supplied by a median artery which enters it 
at its postero-superior extremity ; branches run forwards near the central 
cavity, and break up into capillaries. The veins of the posterior lobe are 
situated immediately beneath the epithelial investment, and converge 
towards the place of entrance of the artery ; leaving the lobe in this situation, 
they turn outwards to join the large lateral sinus on either side. Some of 
the veins from the anterior lobe appear to take a similar course, passing 
through the epithelial investment of the neck of the infundibulum to run 
in the nervous portion ; others leave the anterior lobe at its neck and pass 
outwards into the lateral sinus. The presence of so many large arteries 
and veins in the immediate vicinity of the pituitary body would render the 
operation of removing it an extremely difficult one, even were the organ so 
situated as to be convenient of access. Mi not (26) states that the blood- 
vessels of the anterior lobe of the pituitary will probably be found to have 
a sinusoidal development. This appears to be partly true at any rate of 
the anterior lobe of the pig's pituitary. 



COXCLUSIOXS. 

We may conclude from the histological appearances of the mammalian 
pituitary body that it is an organ of physiological importance It may be 
divided into two parts, which show structural differences probably indicative 
of distinct functions. 

The anterior lobe, consisting of large granular cells and numerous blood- 
vessels, is a gland producing an internal secretion which is poured directly 
into the blood. It is a blood-vascular gland, the function of which is 
undetermined, but which may exercise an influence on growth. The care- 
ful examination of the pituitary body in cases of acromegaly ma}' tiirow 
some light upon this question ; at present any statement as to its probable 
functions must be purely speculative. 

The posterior lobe is made up of two structures. Of these, the part 
developed from the brain and consisting of neuroglia and ependyma cells 
and fibres acts as a framework. It is more or less surrounded and invaded 
by epithelium, which probably furnishes its active part. There is histologi- 
cal evidence of a secretion produced by the epithelial cells, which apparently 
passes into lymph- vessels, and is destined to enter the ventricles of the brain 
The posterior lobe of the mammalian pituitar}^ is a brain gland, not b\ 
virtue of tissue of brain origin, but by the growth into it of epithelial cells 
of ectodermic origin. Extracts have the property of producing marked 
effects on cardiac and plain muscle fibres comparable in some respects to 
the action of the medulla of the suprarenal capsule. They have also a 
selective action upon the kidney, causing dilatation of the renal blood- 
vessels and diuresis. Disturbances of the posterior lobe of the pituitary 
are probabl}'' responsible for the occurrence of the diabetic conditions 



15() Herring 

which liave been so frequently recorded at some time or other in the 
history of cases of acromegaly and of affections and lesions associated 
with the base of tlie skull. 

Summary. 

Three types of mammalian pituitary body are recognised. In one, e.g., 
the cat, the posterior lobe is hollow and its cavity is in free communication 
with the third ventricle of the brain, while the epithelium of the anterior 
lobe affords an almost complete investment for the posterior lobe ; in the 
second type, e.g., the dog, the body of the posterior lobe is solid, but the 
neck is hollow, and communicates with the third ventricle : the posterior 
lobe is here again almost completely surrounded with epithelium ; in the 
third type, e.g., man, monkey, ox, pig, and rabbit, the body and neck of the 
posterior lobe are solid, although traces of a cavity are occasionally found 
in the neck ; in this type the epithelium does not invest the posterior lobe 
so completely, but is aggregated around the neck and spreads over and into 
the adjacent surface of the brain. 

The epithelial portion of the pituitary body is differentiated into two 
distinct parts : an anterior lobe proper, consisting of solid columns of cells, 
between which run wide and thin- walled blood-channels ; and an inter- 
mediate portion, which lies between the anterior lobe and the nervous tissue 
of the pituitary, forming a closely-fitting investment of the latter. 

The anterior lobe contains cells which are clear or hold in their proto- 
plasm varying amounts of deeply-staining granules. They are probably 
different functional stages of one and the same kind of cell, and the granules 
give rise to a secretion which is absorbed by the blood-vessels. 

The intermediate portion consists of finely granular cells arranged in 
layers of varying thickness closely applied to the body and neck of the 
posterior lobe and to the under surface of adjacent parts of the brain. The 
part of it which is separated from the anterior lobe by the cleft is almost 
devoid of blood-vessels. In the cat the portion lying in front of the anterior 
lobe has a tubular appearance and is very vascular. Colloid material occurs 
between the cells of the pars intermedia, and in most situations appears to 
pass into the adjacent nervous substance, to be absorbed by blood-vessels 
or lymphatics. 

The nervous portion of the pituitary body is made up of neuroglia cells 
and fibres. Ependyma cells line the central cavity in the cat and send long 
fibres forwards and upwards towards the brain, most of which terminate in 
the outer part of the neck. There are no true nerve cells and the nerves 
supplying the pituitary probably reach it through sympathetic fibres 
accompanying the blood-vessels (Berkley). The nervous portion is invaded 
to a large extent by the epithelial cells of the pars intermedia. Columns of 
epithelial cells grow into it, especially in the region of the neck, and 
islets of these cells are frequently found throughout the posterior lobe ; 



Histological Appearances of the Mammalian Pituitary Body 157 

in the pituitary of the cat epithelial cells may even grow into its central 
cavity. 

A substance histologically resembling the colloid of the thyroid gland, 
but probably of a different nature, occurs in large quantities in the nervous 
portion of the posterior lobe. It appears to be a product of the epithelial 
cells, and, in the cat at any rate, to be carried by lymphatics into the 
central cavity, and so into the third ventricle of the brain. In this respect 
the posterior lobe of the pituitary is an infundibular gland. Whether this 
substance is modified by its passage through the nervous substance or not 
is unsettled. Its distribution corresponds with the site of the tissue, the 
extracts of which have active physiological results when injected into the 
blood. 

The anterior lobe of the pituitary is extremely vascular and its circula- 
tion sinusoidal. The posterior lobe is supplied for the most part by a 
central artery which enters it at its postero-superior angle and runs forward 
giving off branches; the veins begin immediately below the epithelial 
investment and run backwards in this situation, to emerge near the entry 
of the artery. The veins of both lobes enter large blood sinuses lying close 
to the sides of the pituitary body. 

Histological evidence is against the statement of Be la Hall er that the 
anterior lobe is a tubular gland which pours its secretion directly into the 
subdural space. 



I am indebted to Mr Richard Muir for the care with which he has 
executed the drawings and photographs accompanying this paper. 

The expenses of the research have been defrayed by a grant from the 
Moray Fund for the prosecution of research in the University of Edinburgh. 



LITERATURE REFERRED TO IN THE TEXT. 

(1) Benda, "Beitnige zur nornialen und pathologisclien Histologie der menscli- 
lichen Hypophysis cerebri," Berlin, klin. Wochenschr., Bd. xxxvii., S. 1205, 1900. 

(2) Berkley, "The Finer Anatomy of the Infundilmlar Region of the Cerebrum, 
including the Pituitary Gland," Brain, vol. xvii., p. 515, 1894, 

(3) BoEKE, "Die Bedeutung des Infundibulums in der Entwickelung der 
Knochenfische," Anat. Anz., Bd. xx., S. 17-20, 1902. 

(4) BuKDACH, Vom Baue und Leben des Gehirns, Bd. ii., S. 108-9, and 
Bd. iii., S. 469, Leipzig, 1819-26. 

(5) Glaus and Van der Stkiciit, " Contribution a IV'tiide anatomique et 
clinique de I'acromegalie," Annale-s et Bulletin de la Sue. dc Mt'd. de Inuid, p. 71, 
1893. Quoted from Benda (1). 

(6) Cyon, " Beitriige zur Physiologic der Schilddrvise und des Herzens," 
Pfluger's Archiv, Bd. Ixx., S. 213-16, 1898. 



158 Herring 

(7) DoSTOiEWSKY, " XJeber den Bau der Vorderlappen des Hirnanhanges," 
Arch. f. mikr. Aiiat., Bd. xxvi., S. 592, 1886. First account in Militararztl. Journ., 
Oct. 1884 (Russian). 

(8) EcKEB, Icones Physiologies, Tafel 6, Fig. 9, Leipzig, 1851-9. 

(9) Edinger, Vorlesungen liber den Bau der nervosen Centralorgane, 6 Aufl., 
S. 140 and 293-4, Leipzig, 1900. 

(10) Flbscii, Tageblatt der 57. Versammlung deutscher Katurforscher und 
Aerzte zu Magdeburg, S. 195, 1884. Quoted from Dostoievsky (7) and 
Lothringer (22). 

(11) GeiMELLI, " Sur la structure de la region iiifundibulaire des poissons," 
Journ. de I'anat. et de la physiol., annee xlii., No. 1, 1906. 

(12) GoTTSCHE, " Vergleichende Anatomic des Gehirns der Gratenfische," 
Miiller's Archiv, Jahrgang 1835, S. 437. 

(13) GuEERixi, " Sulla funzione della ipofisi," Lo Sperimentale, anno Iviii., 
p. 872, 1904. 

(14) Haller, B. " Untersuchungen iiber die Hypophyse und die Infundilmlar- 
organe," Morpholog. Jahrbuch, Bd. xxv., S. 101, 1896. 

(15) Hannover, Recherches microscopiques sur le systeme nerveux, \). 26, 1844. 

(16) Hansemann, " Ueber Akromegalie," Berlin, klin. ^Yochenschr., S. 420, 
1897. 

(17) Herring and Simpson, "On the Relation of the Liver Cells to the 
Blood-vessels and Lymphatics," Proc. Roy. Soc, B., vol. Ixxviii., 1906. 

(18) HowKLL, "The Physiological Effects of Extracts of the Hypophysis Cerebri 
and Infundibular Body," Amer. Journ. of Exper. Med., vol. iii., p. 245, 1898. 

(19) KoLLiKER, Gewebelehre des Menschen, 6 Aufl., Bd. ii., S. 603, 1896. 

(20) KoLLiKER, Entwickelungsgeschichte des Menschen und der hoheren Thiere, 
2 Aufl., S. 531, 1879. 

(21) KuPFFER, "Die Deutuug des Hirnanhanges," Sitzber. der Gesellschaft 
f. Morphol. u. Physiol, in Miinchen, S. 59, Juli 1894. 

(22) Lothringer, " Untersucluingen an der Hypophyse einiger Saugethiere 
und des Menschen," Arch. f. mikr. Anat., Bd. xxviii., S. 257, 1886. 

(23) LuscHKA, Der Hirnanhang und die Steissdriise des ]\lenschen, Berlin, 1860. 

(24) Magnus and Suhafer, "The Action of Pituitary Extracts upon the 
Kidney," Proc. Physiol. Soc, July 20, 1901. 

(25) MiHALKOVics, " Wirbelsaite und Hirnanhang," Arch. f. mikr. Anat., Bd. xi., 
S. 389, 1875. 

(26) Minot, " Genetic Interpretations in tlie Domain of Anatomy," Amer. 
Journ. of Anat., vol. iv.. No. 2, p. 262, 1905. 

(27) MiJLLER, W., "Ueber den Bau der Chorda dorsalis und des Processus 
iufundibuli cerebri," Jenaische Zeitschr. f. Naturwissenschaft, Bd. vi., S. 354, 1871. 

(28) Oliver and Sohafer, " On the Physiological Action of Extracts of 
Pituitary Body and certain other Glandular Organs," Journ. of Physiol., vol. xviii., 
p. 277, 1895. 

(29) Osborne and Vincent, " A Contribution to the Study of the Pituitary 
Body," Brit. Med. Journ., vol. i., p. 502, 1900. 

(30) Peremeschko, "Ueber den Bau (l<s Hirnanhanges," Virchow's Archiv, 
Bd. xxxviii., S. 329, 1867. 



Histological Appearances of the Mammalian Pituitary Body 159 

(31) Rabl-Ruckhard, "Das Grosshirn der Knochenfische und seine Anhangs- 
gebilde," Arch. f. Anat. u. Physiol., anatom. Ahth., Jahrgang 1883, S. 317. 

(32) Rathkb, " Ueber die Entstehung der Glandnla pituitaria," Muller's Archiv, 
S. 482, Jahrgang 1838. 

(33) Rbtzius, Biologische Untersuchungen, Bd. vi., Taf. 12, S. 21, 1894. 

(34) RoGowiTSCH, "Die Veranderungen der Hypophysf; iiach Entfernung der 
Schilddriise," Ziegler's Beitrage zur patholog. Anat., Bd. iv., S. 453, 1889. 

(35) Saint-Rbmy, " Contribution a I'histologie de I'hypophyse," Archives de 
Biologie, torn, xii., p. 425, 1892. 

(36) Salzer, " Zur Entwickehmg der Hypophyse bei Saugern," Arch. f. mikr. 
Anat., Bd. h., S. 64, 1898. 

(37) ScHAFER and Herring, " The Action of Pituitary Extracts upon tlie 
Kidney," Phil. Trans., B., vol. cxcix., p. 29, 1906. 

(38) Schnitzler and Ewald, " Ueber das Vorkomraen des Thyreo-jodins ini 
menschlichen Kbrper," Wiener klin. Wochenschr., S. 657, 1896. 

(39) Schonemann, " Hypophysis und Thyreoidea," Virchow's Archiv, Bd. cxxix., 
S. 310, 1892. 

(40) ScHWALBE, "Lehrbuch der Neurologie," Hoflfmann's Lehrbuch der Anatomic, 
Bd. ii., S. 476, 1881. 

(41) Sternberg, "Die Akromegalie," Specielle Pathologic und Therapie, von 
Nothnagel, Bd. vii., Th. 2, 1897. 

(42) Sterzi, "Intorno alia struttura dell' ipofisi nei vertebrati," Atti Accad. 
Sc. Veneto-Trentino-Istriana, CI. sc. nat., lis. e. mat., vol. i., ]). 72, 1904. Abstract 
in Schwalbe's .lahresber., Bd. xi., Th. 3, S. 678, 1907. 

(43) Stieda, Hermann, " Ueber das Verhalten der Hypophyse des Kaninchens 
nach Entfernung der Schilddriise," Ziegler's Beitrage zur patholog. Anat., Bd. vii., 
S. 537, 1890. 

(44) Stieda, Ludwig, " Studien iiber das centrale Xervensystem der Knochen- 
fische," Zeitschr. f. wissensch. Zoologie, l!d. xviii., S. 44, 1868. 

(45) ToLDT, Lehrbuch der Gewebelehre, 2 Aufi., S. 290, 1884. 

(46) ViRCHow, Untersuchungen iiber die Entwickehmg des Schiidelgrundes, 
S. 91-4, Berhn, 1857. 



VOL. I. — APRIL 1908. 11 



THE DEVELOPMENT OF THE MAMMALIAN PITUITARY AND ITS 
MORPHOLOGICAL SIGNIFICANCE. By P. T. Herring. (From 
the Physiology Department, University of Edinburgh.) 

{Received for imhlicilion Wth February 1908.) 

Introduction. 

The development of the pituitary body has been a favourite subject of 
research by embryologists. Its position in the embryo, forming as it were 
a meeting-point for the anterior end of the neural canal, buccal invagination, 
archenteron, and notochord, gives to the pituitary an importance, the signifi- 
cance of which has been the object of much speculation. Some authorities 
have looked upon its relations to these structures as more or less accidental ; 
others have attached great weight to them. Kupf f er, indeed, regarded the 
pituitary body as an important key to the phylogeny of the vertebrate head. 
The morphological significance of the pituitary is also of interest from a 
physiological point of view, and some of the theories which have been 
advanced regarding it will be briefly discussed in this paper. 

Nearly all the work that has been done on the development of the 
pituitary body has been concerned with its mode of origin and ^^^th the 
early stages of its growth. The later stages, although probabl}^ of greater 
physiological importance, have been comparatively neglected. The difier- 
entiation of the epithelium of the anterior lobe, the relations of epithelium 
to the nervous tissue of the posterior lobe, and the extraordinary difierences 
in the vascularity of its several parts are all features which need investiga- 
tion. Its development in mammals has been followed chiefly in animals in 
which the posterior lobe of the pituitary becomes a solid structure at a 
comparatively early stage. In the cat, this lobe remains hollow throughout 
development, and presents peculiarities of morphological interest which are 
not found in the pituitaries of other animals. The structure of the posterior 
lobe in the cat is also of a simpler character, and the nature and arrange- 
ment of the cells found in it can be interpreted more readily than in the 
case of those animals which possess a solid lobe. For these reasons the 
pituitary body of the cat forms the basis of the description in this paper. 
The embryos of man, ox, and pig, which furnish a ditt'erent type of pituitar}^ 
have also been examined, and some of the more important features presented 
by them receive attention. 



162 



Historical. 



The pituitary body was at one time tliought to be wholly derived from 
the brain, but Rathke (26) in 1838 described the invagination of mucous 
membrane which is now known as Rathke's pouch. Rathke rightly 
assigned to this pouch the origin of the epithelial portion of the pituitary, 
but was mistaken in believing it to be derived from the entoderm of the 
fore-o-ut. His view was not at once accepted. Reichert (28) failed to find 
the invagination, and put forward the theory that the epithelial portion of 
the pituitary is a structure of mesodermic origin derived from the anterior 
end of the notochord. His (14) lent additional support to Re i chert's view, 
but made no special investigations of the subject himself. Both Rathke 
(27) and Reichert (29) subsequently changed their opinions, the latter 
believing the anterior lobe to arise from a proliferation of the cells of the 
pia mater. Dursy (8) sought to unite the original view of Rathke with 
that of Reichert and His, and described the origin of the epithelium of 
the pituitary from the fore-gut, and the origin of its vascular stroma from 
the tissue of the head of the notochord. W. Miiller (23) demonstrated 
that the anterior lobe of the pituitary is derived from Rathke's pouch, 
but fell into the same error as Rathke and Dursy in believing it 
to be of entodermic origin. The later researches of Gotte (12) and 
Balfour (3) showed that the pouch described by Rathke is derived, not 
from the fore-gut, but from the epithelium of the buccal cavity imme- 
diately in front of the oral plate. The pouch is now recognised as an 
ectodermic structure. 

The posterior lobe of the pituitary was at first believed to represent the 
anterior extremity of the brain (v. Baer (2)). Gotte (12) showed that in 
amphibians this is not the case, the infundibulum being a later formation. 
The researches of Mihalkovics (21), van Wijhe (35), Kupffer (20), and 
others have demonstrated that the infundibulum cannot be regarded as the 
representative of the anterior end of the brain axis ; it is an outgrowth 
of the "Zwischenhirn" or thalamencephalon. 

The proximity of the anterior end of the notochord to the developing 
pituitary body led to the belief, not only that the notochord enters into the 
structure of the pituitary, but that it also exercises a mechanical influence 
upon the formation of the infundibulum. Both His and Dursy considered 
that a close union between the notochord and the wall of the cerebral 
vesicle is the dominating factor in the development of the infundibulum, 
but were not agreed as to the exact manner in which this is brought about. 
W. M tiller believed that the head of the notochord anchors a portion of the 
wall of the brain, and that with the growth of the surrounding tissues the 
rest of the brain is carried forwards, leaving a diverticulum of its wall, the 
infundibulum, attached to the notochord. The attachment is subsequently 
dissolved b}^ a proliferation of connective tissue cells. Mihalkovics (21) 
and others showed that the head of the notochord does not come into 



The Development of the MammaHan Pituitary Body 163 

immediate relationship with the brain, and cannot therefore act upon it in 
this manner. 

The most complete account of the early development of the pituitary 
body is that given by Mihalkovics (21), who investigated the subject in 
rabbit and chick embryos. Mihalkovics found that the anterior lobe is 
developed from Rathke's pouch, which, in mammals as in amphibians, is 
of ectodermic origin. The beginning of the pouch or hypophysial angle 
lies in front of the oral plate, where the epidermis bends round the base of 
the brain to the nasal mucosa. In the rabbit, Mihalkovics states that the 
end of the notochord is in contact with the epidermis at the back of 
Rathke's pouch. When the oral plate ruptures, its upper stump, con- 
taining in its upper part the head of the notochord, bends forward and 
narrows the mouth of the epithelial pouch, leading to the formation of a 
definite sac — the hypophysial sac. The wall of the sac presses upon the 
base of the anterior brain vesicle, giving rise at its upper extremity to a 
fold in the wall of the brain which becomes the primitive infundibulum. 
Mihalkovics denied that the end of the notochord is ever united to the wall 
of the fore-brain ; it does not enter into the formation of the infundibulum 
at all, but has some influence upon the hypoph3\sial sac, by preventing this 
from extending backwards. The primitive infundibular process comprises 
the surrounding tissue of the tuber cinereum as well as the origin of 
the infundibulum, and the true infundibulum is formed at a later stage 
by its own growth from a portion of the primitive infundibular process. 
Mihalkovics made a careful investigation of the relations of the notochord, 
and found that its head touches the lower part of the posterior wall of the 
hypophysial sac in rabbits, but is placed at a higher level in birds ; it 
exercises no traction upon the sac in either, and, beyond presenting a 
barrier to the backward growth of the sac, takes no part in the formation 
of the pituitary body. 

The main conclusions of Mihalkovics' researches have been confirmed 
by Kolliker (16), Kraushaar (18), Minot (22), Kupf f er (19), Salzer (33), 
and others. Kupf f er described an additional origin of part of the anterior 
lobe of the pituitary from the entoderm of the fore-gut. According to 
Kupffer, the pituitary body of amphibians is built up from three separate 
sources : part of the epithelial lobe is derived from Rathke's pouch, and 
part from the anterior end of the fore-gut, while the infundibulum is of 
brain origin. In mammals, e.g. the sheep, the hj^ophysial pouch appears 
behind the " Riechplakode." Behind this and ventral to it is a swelling, 
the " Haftscheibe," which is an important larval organ in Lepidosteus. 
Then comes the double-layered oral plate (" Rachenhaut "), and behind 
this an outgrowth of entoderm directed dorsally and forwards, known as 
Seessel's pouch. The third portion of the pituitary, the cerebral, appears 
later, after the disappearance of the oral plate and median "Riechplakode." 
In the next stage the growth of entoderm increases, but is cut oft' from 
Seessel's pouch ; no cavity is to be found in it, and the end of the notochord 



164 Herring 

remains in contact with it. The infundibuknii now begins to grow. In 
the older embryos, e.g. 11 -mm. sheep, the entodermic part degenerates and 
appears as a string-Hke appendage of the notochord; it eventually dis- 
appears, and does not enter into the formation of the adult mammalian 
pituitary. 

Kupf f er came to the conclusion that the intimate relationship between 
infundibulum, mouth, and intestine is not an accidental one, but denotes an 
ancestral communication between the brain tube and the anterior part of 
the intestinal canal. A structure resembling in many respects the early 
stao-es of development of the vertebrate pituitary is found in Ascidians, and 
is known as the subneural gland. Jul in (15) in 1881 pointed out that 
this gland is probably homologous with the hypophysis of higher vertebrates, 
and since then it has been frequently spoken of as the Ascidian hypophysis. 
Kupf f er believes that the direct ancestors of vertebrates showed the same 
relations as are seen to-day in the tailed Ascidian larva. In a scheme of the 
ancestral vertebrate he describes the mouth (" Palseostoma ") opening dorsally 
in front of the brain. The brain tube is in communication with the anterior 
part of the intestine by a canal running through the base of the anterior 
brain vesicle. This canal has developed upon it a subcerebral gland. 
Ventral to the palfeostoma is the " Haf torgan " on the anterior pole of the 
body. In the course of development the new vertebrate mouth (Neostoma) 
is formed, in agreement with Dohrn's hypothesis, from a pair of gill- clefts 
below the " Haf torgan." The part of the intestine between the old and 
the new mouth, or preoral intestine, is reduced, but persists to a certain 
extent in some vertebrates. The palseostoma is lined by epidermis, and its 
representative in vertebrates is Rathke's pouch ; it also forms the outer part 
of the nasal duct (" Nasenrachengang") of Myxine, and the entire nasal 
duct of Petromyzon. The remains of the canalis neurentericus anterior, 
with its appertaining glands, are to be seen in the infundibular process and 
saccus vasculosus. In mammals, the only representative of the preoral 
intestine is the transitory appearance of the solid mass of cells formed from 
Seessel's pouch, but in amphibians it persists as part of the anterior lobe of 
the pituitary. 

Kupf f er's views on the morphological significance of the pituitary body 
have not met with general acceptance. Willey (36) states that the present 
relation of the hypophysis to the infundibulum in the craniates, however 
intimate it may be in some cases, is, nevertheless, incidental and secondary. 
Willey believes that the hypophysis arose in connection with a functional 
neuropore. B. Haller (13) criticises Kupff er's results and differs from him 
in many important particulars. He believes the nasal duct of Cyclostomata 
to be a secondary structure and not related to the origin of the hypophysis. 
He also states that the anterior lobe of the pituitary of mammals and 
other vertebrates is a tubular gland which pours a secretion into the 
subdural space. The latter statement has not been confirmed by subse- 
quent observers. Gaskell (9) quotes Hall er's results in support of the 



The Development of the Mammalian Pituitary Body 165 

theory that the glandular hypophysis was originally the coxal gland of 
Arthi'opoda. 

Kupffer's description of the threefold origin of the pituitary body has 
received support from observations by J. Nusbaum (24) and Saint- 
Remy (31). Nusbaum found that in dog embryos of 9 mm. Seessel's 
pouch is well developed, and its anterior extremity abuts against the 
posterior wall of Rathke's pouch. In 80 per cent, of older embryos examined 
it gives rise to a column of cells which unites with the epithelium of 
Rathke's pouch, and thus enters into the formation of the anterior lobe. 
In the remaining embryos no such appearance is seen, and the anterior 
lobe is entirely ectodermic in origin. • Traces of a lumen were noticed by 
Nusbaum in the column of cells growing from the fore-gut, but not a 
definite communication between the interior of the buccal invagination and 
the fore-gut. The connection is not preserved for long, and the entodermic 
cells disappear, with the exception of a few which join the posterior 
wall of Rathke's pouch. What further part these cells play — if any — in 
the formation of the anterior lobe of the pituitary Nusbaum did not 
determine. 

Saint- Remy (31) described a budding of Seessel's pouch in the embryo 
chick towards the seventieth hour of incubation. The bud acquires a fine 
lumen, and, reaching Rathke's pouch, afibrds a direct communication between 
the interior of the latter and the fore-gut. The connection lasts a little, 
then disappears, the cells of Seessel's pouch never actually uniting with 
those of Rathke's pouch. Saint-Re my agrees with Kupffer that the 
entodermic origin is rudimentary in birds and mannnals, and does not enter 
into the formation of the adult pituitary body. It is, however, of morpho- 
logical importance, and betokens the existence in lower forms of vertebrates 
of a communication between the intestine and the buccal invagination. 

Dohrn (7) looked upon the pituitary as the remains of a preoral gill- 
cleft. Salvi (32) has brought forward evidence in support of this view, 
and states that in reptiles part of the pituitary is developed from the walls 
of the premandibular cavities, which he believes to be the representatives 
of gill-clefts. Valenti (34) describes the origin of the anterior lobe in 
amphibians from an invagination of the fore-gut arising some distance 
behind Seessel's pouch. The invagination, he considers, has not the signifi- 
cance attributed by Kupffer to Seessel's pouch, but is rather to be regarded 
as the representative of a gill-cleft. Valenti therefore supports Dohrn 's 
theory. Dohrn 's view was based chiefly upon the assumption of a bilateral 
origin of the anterior lobe of the pituitary. Dohrn himself described a 
bilateral origin in Hippocampus, and Gaupp (10) found something similar 
in reptiles. Gaupp, however, described a median origin in addition to 
lateral ones, and believes all to be formed from the buccal cavity. 

Yet another interpretation of the significance of the pituitary has been 
put forward by Beard (4), who believes the anterior lobe to be homologous 
with the permanent mouth of Annelids. 



166 



Herrinj 



Author's Observations. 

My own results are confirmatory of those of Mihalkovics and Kupf f er. 
In a 4-mm. cat embryo, the youngest I have had the opportunity of ex- 




FiG. 1. — Mesial sagittal section through head of 4-mm. kitten. (Diagram.) 

a, hypophysial angle formed by buccal mucous membrane ; 6, depression in wall of cerebral vesicle where 
the infundibulum is formed; c, blood-channel; d, anteiiur end of fore-gut or Seessel's pouch; e, head 
of notochord ; /, upper stump of ruptured oral plate. 

amining, the appearance is that indicated in fig. 1. The oral plate (f) 
between buccal invagination and fore-gut has just ruptured. Immediately 
in front of the oral plate is the hypophysial angle described by 
Mihalkovics. The anterior limb of the angle is composed of buccal 






The Development of the Mammalian Pituitary Body 



167 



epithelium, which in this situation is closely adherent to the wall of the 
anterior cerebral vesicle. The posterior limb of the angle, also composed 
of buccal epithelium, leaves the wall of the brain and bends downward to 
form the anterior layer of the upper stump of the oral plate. At this stage 
there is no invagination of the wall of the cerebral vesicle to form the 
infundibulum, but its site is indicated by a definite depression. The 
anterior end of the notochord doBs not touch the posterior limb of the 



d b c e 




Fig. 2.— Mesial sagittal section tlirougli part of head of a Gmiu. kitten. 

a, buccal invagination or Rathke's pouch : b, beginning of invagination of wall 
of cerebral vesicle to form the infundibular process; c, blood - channel : 
d, clump of cells derived from anterior end of fore-gut ; c, head of notochord. 

hypophysial angle, but is separated from it by a large blood-channel (c). 
Behind the oral plate is a small dorsal invagination of the wall of the 
fore-gut, which is the only indication of anything resembling Seessel's 
pouch. Its wall is not thickened, and there is no evidence of any ento- 
dermic origin for the pituitary in this specimen. 

In a 6-mm. cat embryo (fig. 2), the remains of the oral plate have 
disappeared. The hypophysial angle has become a definite sac (a), Rathke's 
pouch. This change appears to have been brought about by a bending 
forwards of the upper stump of the oral plate and a proliferation of the 



168 Herring 

cells in its wall. The pouch is widening out behind the neck, and the latter 
is found to be constricted when the sections next in series to it are examined. 
The anterior wall of Rathke's pouch is closely applied to the wall of the 
cerebral vesicle, and at the dorsal extremity of the pouch an invagination 
of the wall of the cerebral vesicle is forming the primitive infundibulum. 

The head of the notochord bears no immediate relation to Rathke's 
pouch, and is separated from it by a clump of cells which is continuous with 
the epithelium of the fore-gut, and appears to be formed by a proliferation 
of the cells of the latter. A large blood-vessel (c) is also seen in this 
specimen, lying between Rathke's pouch and the head of the notochord. 

At this stage it is difficult to determine where ectoderm ends and 
entoderm begins ; the upper stump of the oral plate has disappeared as such, 
and its representative is uncertain. There is no indication of a pouch in 
the fore-gut, but the clump of cells appears to be derived from the wall of 
the latter. Minot (22) makes the fold of epithelium at the posterior 
margin of Rathke's pouch homologous with the upper lip of Petromyzon. 
If this is the case, the fore-gut must begin behind this fold. The close 
relation between the head of the notochord and the cell clump makes it 
likely that the latter is derived from the fore-gut, for, in the 4-mm. embryo, 
the head of the notochord is some distance behind the oral plate, and the 
epithelium opposite it is that of the fore-gut. The clump of cells is the 
only structure which resembles the proliferation of entoderm described by 
Kupf f er. It is not found in any of tlie older embryos that I have examined, 
but the amount of suitable material at disposal for this purpose has been 
limited. Rathke's pouch is the only part that enters into the formation of 
the anterior lobe of the pituitary ; it is single and median in origin, and 
there is no indication in the embryos of the cat and the pig of any other 
" Anlage " for the anterior lobe. I have not found any communication 
between the epithelium of Rathke's pouch and that derived from the fore- 
gut, as described by Nusbaum in the dog, but cannot say that this does 
not occur. In the specimens I have examined there is nothing to indicate 
in the slightest degree that Rathke's pouch is reinforced by epithelium 
from the fore-gut. The epithelial proliferation of the latter disappears as 
stated by Kupff er, and takes no part in the formation of the pituitary. 

One of the most important characteristics of the developing pituitary is 
the close union maintained between buccal and cerebral portions from the 
earliest stage. Minot (22) emphasised its importance in mechanically 
keeping the two parts together, and thus explaining their intimate relations. 
Salzer (33) also noted it, and states that he could find no connective tissue 
between the infundibular process and hypophysial sac. With these obser- 
vations I thoroughly agree. The buccal epithelium in the anterior part of 
the hypophysial angle is intimately connected with the epithelium of the 
cerebral vesicle, without the interposition of connective tissue. In the 
further growth of the embr^'o this close union is preserved, but in other 
parts connective tissue develops and separates the buccal epithelium from 



The Development of the Mammalian Pituitary Body 



169 



the wall of the cerebral vesicle. No doubt this process is contributory to 
the formation of Rathke's pouch and infundibulum, but it is probably of 
morphological significance as well, and betokens the existence in an ancestral 
vertebrate of a communication between buccal cavity and neural canal. 

A later stage of development is shown in fig. 3, which is taken from a pig 
embryo of 12 mm. The buccal mucous membrane is now widely separated 




Fig. 3.— jMesial sagittal section through part of head of a 12-mm. pig embryo. 

a, Rathke's pouch ; b, beginning of infundibular process ; c, blood-channel ; d, remnant of Seessel's 

pouoh (V) ; e, notochord. 

from the wall of the cerebral vesicle, except at that part where the 
anterior wall of Rathke's pouch closely adheres to it. The infundibulum 
is only beginning to form, and the cells lining the cerebral vesicle at this 
point have proliferated and elongated, and look more like ependjana 
cells. There is no indication of any proliferation of cells of the fore-gut. 
Rathke's pouch is median in situation, its neck is constricted, and serial 
sections show that there is no lateral origin of the pituitary. The note- 
chord persists, but has no immediate relation to Rathke's poucli. It takes 



170 



Herring 



no par-t niechanicall}^ or otherwise in the formation of the pituitary. Its 
situation, nevertheless, is not without significance ; its arrest behind the 
anterior end of the brain tube allows the latter to communicate with the buccal 
epithelium and possibly with the fore-gut in some animals. Had the noto- 
chord grown further forward, a median origin for the anterior lobe of the 




Fig. 4. — Mesial sagittal sectiuii through part of head of an 18-mm. kitten. 

a, hypophysial sac now closed below ; 6, infundibular process ; g, neck of sac connected with nasal 
mucous membrane ; h, cartilage of sphenoid bone. 



pituitary would have been impossible. The median origin of the pituitary, 
or rather the ancestral condition which this implies, may indeed explain 
why the head of the notochord has been arrested in this situation. 

In an 18-mm. cat embryo considerable changes have taken place (fig. 4). 
Rathke's pouch has become a closed sac, but its wall is still connected by 
a stalk of epithelium with what is now becoming nasal mucous membrane. 
The narrowing of the neck of the sac, its closure and ultimate disappear- 



The Development of tlie IVIanimalian Pituitary Body 171 

ance are due to tlie strong growth of connective tissue around it and the 
development of the sphenoid bone. It is unnecessary to attribute the 
change to the pressure exerted on the neck of the sac by the carotid arteries 
and their growing adventitia, as did W. Mliller. The above explanation 
of Mihalkovics is probably the right one. The hypophysial sac is now 
of considerable size and extends laterally, its anterior wall being still in 
close connection with the wall of the cerebral vesicle. A well-marked 
invagination of the wall of the latter constitutes the infundibular process ; 
its anterior wall is also closely invested by the epithelium of the wall of 
the hypophysial sac. The infundibular process becomes the nervous 
portion of the posterior lobe of the pituitary, while the part of the wall of 
the hypophysial sac adhering to it constitutes the epithelial covering of 
the posterior lobe or " Epithelsaum " of Lothringer. Epithelium and 
nervous tissue have been in close contact with one another from their first 
appearance. The cavity of Rathke's pouch persists throughout life as the 
epithelial cleft which partly separates the anterior from the posterior lobe. 
In its lateral extension the sac is beginning to envelop the sides of the 
infundibular process. Its walls are composed of cylindrical cells which 
closely resemble those of the infundibular process. They are thickened in 
the anterior part of the sac in the region of its neck, where there is a 
distinct fold ; the thickening in this situation is the beginning of the 
anterior lobe proper of the pituitary. Its cells are not as yet differentiated 
from the cells of the remainder of the sac. 

During subsequent development the pituitary body is removed further 
and further from the nasal mucosa by the growth of the sphenoid bone. 
Ossification in the latter is delayed for some time by the persistence of a 
cord of epithelial cells connecting the anterior lobe of the pituitary with the 
nasal mucous membrane. This connection is still found in cat embryos of 
from 35 to 40 mm., but is then imperfect and soon after disappears, allowing 
the opening in the bone to close up. Differentiation between anterior lobe 
proper and the pars intermedia now begins to take place. The anterior 
lobe is formed by a proliferation of the cells of the lower part of the 
anterior wall of the sac just above its neck. Solid columns of cells are 
formed in this situation, and invade the cavity of the sac so as gradually to 
fill it, leaving only a narrow space or cleft between them and the epithelium 
covering the posterior lobe. The anterior lobe also grows forward and 
laterally. The neck of the sac retains a tubular character for some time, 
and becomes somewhat convoluted. One of these convolutions (fig. 5, Z) 
applies itself to the under surface of the brain and gives rise to the tongue- 
shaped process which extends forwards from the anterior lobe towards the 
optic chiasma. 

The structures which enter into the formation of the pituitary are 
closely related to large blood-vessels from their earliest appearance. In 
figs. 1 and 2 a large blood-channel is seen lying immediately behind 
Rathke's pouch, in front of the notochord. Dursy, indeed, as already 



172 



Herrinof 



stated, sought to derive the origin of the blood-vessels of the pituitary from 
the tissue of the head of the notochord. The exact manner in which the 
blood-vessels of the anterior lobe of the cat's pituitary are formed is some- 
what difficult to make out, but in the pig embryo their origin is partly 
sinusoidal. The cell columns of the anterior lobe grow into large blood- 




Fm. 



-Mesial sagittal section through part of head of 35-mm. kitten. 



remains of cavity of flathke's pouch now recognisable as the epithelial cleft of the pituitary ; 
b, central cavity of infundibular process ; g, remnant of epithelial duct connecting hypo- 
physial sac with the nasal mucous membrane ; ;', third ventricle of brain ; k, part of 
epithelial duct which becomes the tongue - like process of pars intermedia ; I, cells of 
anterior wall of hypophysial sac (pais intermedia); in, anterior lobe proper; n, vascular 
knob covered with ependynia cells projecting into cavity of the infundibular process. 



sinuses, pushing the endothelial lining before them. In a pig embryo of 
60 mm. the appearance of the anterior lobe is very like that of the develop- 
ing liver. It is not, however, entirely sinusoidal, and some ingrowth of 
blood-vessels with accompanying connective tissue takes place, the latter 
being always small in quantity. This method of development of the blood- 
vessels of the anterior lobe was first pointed out by Gaupp (10) in reptiles. 



The Development of the Mammalian Pituitary Body 173 

Gaupp found large blood-spaces of a venous nature, and states that the 
epithelium grows into them, passing through their walls. Mi not also 
inferred from the structure of the anterior lobe of the pituitary that the 
development of its blood-vessels is partly sinusoidal. While the anterior lobe 
and tongue-like process of epithelium are extremely vascular, that part of 
the wall of the original sac which is applied to the brain remains devoid of 
blood-vessels. Its cells proliferate and spread round the nervous substance 
of the posterior lobe, forming a covering of epithelium of varying thickness. 
The nervous portion of the posterior lobe has meanwhile grown in length 
and expanded to form a definite body, which, in the cat, retains a large 
central cavity. The neck is constricted, but remains hollow. In a 35-mm. 
cat embryo (fig. 5) the epithelium lining the central cavity is composed of 
ependyma cells with thin processes, the nervous tissue is small in amount, 
contains few cells, and is chiefly made up of the processes of the lining 
cells. At the postero-superior angle of the lobe there is frequently seen a 
knob-shaped body (n) of large and deeply staining ependyma cells, behind 
which are blood-vessels. This vascular knob appears to be growing into 
the central cavity, and marks the entry of blood-vessels into the posterior 
lobe. The thickening of epithelium does not persist, but disappears ; the 
blood-vessels, however, grov^^ into the lobe in this situation. The appear- 
ance is not a constant one, but when it occurs the deeper staining of the 
ependyma cells and the vascularity of the tissue behind them are striking 
features. The blood-vessels of the posterior lobe of the pituitary are, in 
the cat, almost entirely derived from an ingi-owth in this situation ; true 
capillaries are formed in the lobe and are accompanied by a small amount 
of connective tissue. The latter is nev^er present in large quantities, and 
the posterior lobe of the pituitary does not become a connective tissue 
appendage of the brain, as stated by W. Miiller and many others; there 
is remarkably little connective tissue in the posterior lobe. The appear- 
ance which W. Miiller likened to a spindle-celled sarcoma is very marked 
in the older pituitary. It is, however, not due to the presence of connective 
tissue fibres and cells, for when the pituitary is prepared b}^ Cajal's silver 
method it is found that the appearances described by W. Miiller are caused 
by the presence of large numbers of ependyma- and neuroglia-cells and 
fibres, chiefly the former. In the developing pituitary there is never any 
sign of true nerve cells. The ependyma cells lining its central cavity are 
at first like those lining the third ventricle of the brain. Their fibres 
run vertically and end at the outer surface of the lobe. As the posterior 
lobe elongates the peripheral ends of the fibres remain attached, the cells 
become more numerous, and are moved fui'ther and further from the points 
of attachment of their fibres. In this way the ependyma fibres of the neck 
of the posterior lobe acquire an oblique direction, and finally run almost 
longitudinally, their cells of origin being situated much further back than 
the outer ends of their fibres. In the posterior lobe of the kitten, especiallv 
in the region of the neck, the ependyma fibres become very numerous, and 



174 



Herring 



their arrangement, as shown by Cajal's method, is very complex. In the 
adult cat the individual fibres are much thicker, and the cells fewer in 
proportion to the size of the lobe. Neither by Cajal's nor Golgi's method 
have I been able to find true nerve fibres entering the posterior lobe of the 
cat's pituitary through the neck. The ependyraa fibres take on a lighter 
stain by Cajal's method than the nerve fibres in the brain. Fibres can be 
seen in the fully developed pituitary which are thinner and stain more 




Fig. 6. — Mesial sagittal section through developing pituitary body of a human foetus 
(fifth month). Drawing from a photograph. 

a, optic chiasnia ; b, tongue-like process of epithelium ; c, third ventricle : rf, anterior lobe ; e, neck of 
posterior lobe ; /, epithelium surrounding neck ; g, epithelial cleft ; h, posterior lobe. 

deeply. Some of them enter the epithelium round the neck of the posterior 
lobe and ramify there. They are probably true nerve fibres which have 
entered with the blood-vessels, and, as Berkley (5) states, derived from 
the sympathetic. 

In the embryos of man, ox, and pig the posterior lobe is solid from an 
early stage, and is relatively smaller than the anterior lobe. There is the 
same close connection between the epithelium and the nervous tissue. In 
fig. 6 a mesial sagittal section through the pituitary body of a human 



The Development of the MammaHan Pitiiitaiy Body 



175 



embryo (5th month) is illustrated. The anterior lobe is a compact structure 
of columns of epithelial cells devoid of lumina. Many of its cells are 
granular in character, and are beginning to differ from the clearer cells, 
which are in closer relation with the nervous tissue. There is a well- 
marked cleft in the epithelium, which, in this specimen, is carried right 
round the neck of the posterior lobe. The cells of the pars intermedia, or 
that portion of epithelium derived from the anterior wall of Rathke's 
pouch, which is closelj'- adherent to the wall of the cerebral vesicle, are 




Fig. 7. — Sagittal section through same pituitary as shown in fig. 6, but further to one side. 
Drawing from a photograph. 

6, tongue-like process of epithelium spreading forward ; b', epithelial cells spreading backwards over surface of the 
brain ; c, third ventricle ; d, anterior lobe ; g, epithelial cleft ; /(, posterior lobe ; p, l>Tnph space. 

widely spread over the surface of the neck and body of the posterior lobe : 
they also tend to break up the neck of the lobe by passing into its 
substance along with blood-vessels, and extend forwards in a thin layer 
for some distance over the tuber cinereum. Fig. 7 shows the appearances 
presented by the same pituitary in a section further from the middle line. 
The anterior lobe is rather larger here than it is in the mesial plane. The 
posterior lobe is small, and its neck, which is a very thin one, is not seen. 
The cleft is still seen, but the epithelium covering the posterior lobe is 
small in amount and is confined to that part of it which borders the cleft. 
In the human embrj'o at this stage the intermediate part of the pituitary 
VOL. I. — APRIL 1908. 12 



176 Herring 

covering the neck and part of the posterior lobe is relatively smaller than 
it is in the cat embryo. But the intermediate part is not really reduced ; 
it has changed its position, and in the human embryo is found to extend 
further over the surface of the brain. A thin layer (fig. 7, h) is prolonged 
forwards and backwards (6') over the brain substance adjacent to the neck. 
The cells are arranged in columns, which may be a single cell thick, no 
lumen being found in them. Blood-vessels accompany this layer, and 
pass freely inwards into the brain substance, often carrying with them cells 
of the pars intermedia for a short distance along their course. The 
intimate relation of the cells of the pars intermedia to pars nervosa, and 
their differences in .structure from the cells of the anterior lobe proper, 
appear to indicate that they are physiologically as well as anatomically 
connected with the brain. In the cat they are aggregated around the neck 
and body of the posterior lobe, which are hollow and in connnunication 
with the third ventricle. In animals which have a solid posterior lobe 
they are disposed more in relation to the brain substance adjacent to the 
neck, and in the monkey may spread inwards almost to the floor of the 
third ventricle. 

In experiments which have been made on the physiological action of 
extracts of the posterior lobe, the material has usually been taken from the 
pituitary of the ox, on account of its size. An illustration (fig. 8) is given 
of the developing pituitary of the ox. In this animal the posterior lobe is 
a thin, solid, elongated structure. The epithelium of the intermediate part 
spreads widely over its anterior surface, as seen in the figure, but embraces it 
laterally as well, and passes for considerable distances in the form of columns 
of cells into the substance of the body of the lobe. Its epithelium is 
therefore closely bound up with the nervous substance of the lobe during 
development, and forms an important element in its composition. The 
question of the derivation of the active physiological principle of extracts 
of the posterior lobe has been discussed in a previous paper, and reasons 
have been given for regarding it as in great part derived from the 
epithelium of the pars intermedia. It is of interest, therefore, to find that 
in the development of the pituitary of the ox epithelial cells pass freely 
into the substance of the posterior lobe. 

The disposition of the cells of the pars intermedia is such as to bring 
them into close relation with the neural canal in the region of the third 
ventricle. This of course follows from the history of the mode of 
development of the pituitary, but the spreading of epithelium over the 
surface and into the brain itself seems to indicate some further connection 
between the two. In fishes, e.g. the cod, the epithelium of the anterior 
lobe appears to have the same intimate connection with the nervous 
tissue of the posterior lobe which obtains in mammals and birds. The 
posterior lobe is hollow and has connected with it a large saccus 
vasculosus lined with folds of columnar epithelium. The saccus vasculosus 
is said by all who have worked at its development to be derived from the 



The Development of the Maininalian Pituitary Body 



177 



brain, and its secretion, if it is a secretory structure, passes into the third 
ventricle through the infundibuluin. For this reason Rabl-Ruckhard 
(25) called it an infundibular gland. No such structure is found in 
mammals, but Retzius (30) has described in them a slight swelling in the 
form of a clover leaf appearing on the surface of the brain behind the 
infundibulum, which he called the eminentia saccularis, and believed 
to be the homologue of the saccus vasculosus of fishes. Retzius 
described its external appearance only. In section, it is found to be a 




Fig. 8. — ^lesial sagittal section through developing pituitary body of ox. Drawing 
from a i)hotograph. 

a, toiifiue-like process of epithelium spreadiiiK forwards; b, third ventricle; c, anterior lobe; d, epithelial cleft ; 
(', epithelium of pars intermedia ; /, posterior lobe ; g, large lymph space extending into body of ossifying 
sphenoid bone. 

thinning of brain substance in the tioor of the third \ entricle in front of 
the corpora mainillaria. It is doubtful if this eminence is really homo- 
logous with the saccus vasculosus of fishes ; it is not in the position one 
would expect, and should be sought rather in the postero-superior angle of 
the posterior lobe of the pituitary, where the blood-vessels enter, and where 
there is fre(|uently in the cat, during development, the vascular knob 
already mentioned. 

Although the mammalian brain has no saccus vasculosus, the posterior 
lobe of the pituitary possesses an investment of epithelium which difiers 



178 Herring 

from that of the anterior lobe, and it is because of the occurrence of colloid 
vesicles in this situation that the mammalian posterior lobe has been 
termed by K oil ike r (17) an infundibular gland. Whether the epithelium 
of the pars intermedia of the mammalian pituitary has a similar function 
to that of the saccus vasculosus of fishes is a difficult question to answer. 
Extracts of the saccus vasculosus of the cod do not appear to have the same 
physiological action as extracts of the posterior lobe of the mannnalian 
pituitary when injected into the blood ; but ni}' investigations into the 
comparative phj^siolog}" of the vertebrate pituitary are not sufficiently 
advanced to make any conclusive statement on this point. The presence of 
epithelial cells of the pars intermedia in the interior of the cavity of the 
posterior lobe of the cat's pituitary renders it probable that they furnish 
some material which passes in the direction of the brain. There must be 
some significance in the fact that in all craniate vertebrates the cerebro- 
spinal canal has in close proximitj' to its anterior end, and intimatelj' bound 
up with it, a glandular organ connected with the mouth. In the Ascidian 
larva a subneural gland or hypophj'-sis cerebri occupies the same position. 
Andriezen (1) described in the Ammocoete and larval Amphioxus a sub- 
neural gland, a duct lined with ciliated epithelium affording a communica- 
tion between the buccal and the neural cavities, and a group of nerve cells 
around and at the back of the upper opening where the duct widens into 
the ventricular cavity ; an arrangement, in fact, which is veiy similar to 
that found in the larval Ascidian. Andriezen believed that the buccal 
ventricular duct serves as an inlet for oxygenated water to the .spinal cord, 
the nerve cells acting as a sensory mechanism to test the quality of the 
water admitted. In higher animals the water vascular duct disappeai's 
and the posterior lobe gradually loses its nerve substance. The anterior 
lobe or subneural gland alone remains functional, but its secretion is carried 
to the brain by lymphatics and blood-vessels. Sensory structures have 
been described in the infundibular region by Boeke (6) and by Gemelli 
(11) in fishes. There are, as already stated, no appearances indicative of 
such in the posterior lobe of mammals. Andriezen's view is to some 
extent similar to the one expressed by Kupf fer, but his anatomical data 
do not agree with the description of the larval Amphioxus as worked out 
by Willey and other authorities. 

If the anterior lobe of the pituitary body is to be regarded as the 
remnant of an old mouth into the neural canal, it is possible that such a 
connection will occasionally show itself in the course of development. In 
one cat embryo I have met with a communication between the epithelial 
cleft and the central ca%'ity of the posterior lobe. The opening between the 
two is at the postero-superior extremity of the posterior lobe, and it has 
been rendered possible by the spreading of Rathke's sac further backwards 
than usual. The cleft or remnant of the original lumen of the buccal 
invagination is in open continuity through the infundibulum with the 
cavity of the third ventricle. The opening is a median one, and 



The Development of the Mammalian Pituitary Body 179 

consists of a narrow canal lined by ependyma cells passing backwards from 
the wider lumen of the cavity of the body of the posterior lobe to meet 
with the fold of buccal mucous membrane and pass through it into its 
cleft. The ependyma cells cease where the canal meets epithelium, and the 













Fig. 9.-3resial sagittal section tlirougli postero-superior angle of the 
posterior lobe of the pituitary „f a kitten near full time. Drawing 
from a j)hotograpli. * 

"' ^n? nnL?rii'j.^"?' ''^w' ' ''' r,"*''" '""^'^y °f posterior lobe ; c, nervous substance 
of posterior lobe ; d, caiml from cavity of posterior lobe opening into cleft 

lining of its actual opening into the cleft is formed by the epithelium of 
the buccal invagination. The opening takes place at a comparatively late 
stage in development, after the buccal invagination has become separated 
from its origin by the growth of the sphenoid bone. I have examined many 



180 



Herrini; 



cat embiyos, but have not met witli another specimen showin^^ actual con- 
tinuity of both cavities, altlioucjh it is not unusual to find the appearances 
shown in figs. 10 and 11, where a somewhat similar condition exists; a 
canal runs backwards from the central cavity of the posterior lobe to meet 
the epithelium formed from the buccal invao;ination. In some kittens 

he c <1 f 



%. 



n, 



J 



•'V,V 












Fig. 10.— Mesial sagittal section through postero-supeiior angle of the 
posterior lobe of the pituitary of another kitten near full time, 
drawing from a photograph. 

h, central cavity of posterior lobe ; c, substance of posterior lobe ; d, canal from 
cavity of posterior lobe running backwards into e, epithelial cells of tlie pars 
intermedia; /, cells of pars intermedia investing posterior lobe below. 

there is an indication of a tu])ular character of tlie epithelium of the pars 
intermedia in this situation, and the central cavit}" of the posterior lobe 
appears to run into it. In older animals epithelial cells often invade the 
tissue of the posterior lobe in the position in which the canal is indicated 
in the figures, and may come to lie in the central cavity. In one adult 



The Development of tlie Mammalian Pituitaiy Body 



LSI 



cat the nervous tissue of tlie posterior lobe is almost entirely destroyed by 
an overgrowth of epithelial cells of the pars intermedia. These observa- 
tions and the occurrence of epithelial cells in the brain substance of the 
floor of the third ventricle in the adult monkey seem to point to some 
physiological connection between epithelial cells and cerebro-spinal canal. 

h e _d f 



: .i^f]i^,:^^ 






*•:: 



* • '..0 •' V. t .'• •••»•!• •.•%.•• / 



Fig. 11. — j\lesial sagittal section through {jostero-supeiior angle of the 

posterior lobe of the pituitary of a third kitten near full time. 

Drawing from a photograph. 
b, central cavity of posterior lol)e ; d, canal leading from this cavity into c, epithelial 

cells of pars intermedia ; /, cells of pars intermedia jind cleft lying below the 

posterior lobe. 

Since the intimate relations that exist between the two from the earliest 
stages of development are not only maintained but emphasised in the adult 
mammalian pituitary, it is unlikely that they are accidental, and Kupf f er's 
hypothesis as to the signiflcance of the pituitar}' body appears to ha%e 
many facts that support it. The nature of the connection between epithe- 
lium and cerebro-spinal canal from a physiological point of view awaits an 



182 Herring 

explanation. It is possible that the epithelial cells secrete some substance 
which is necessary for the brain. Andriezen's view that the secretion of 
the cells of the anterior lobe of the pituitary is carried by lymphatics and 
blood-vessels to the brain is unlikely, owing to the vascular arrangements in 
the lobe ; but it is probable that something of the kind occurs with the secre- 
tion of the cells of the pars intermedia. B. Haller regards the anterior lobe 
as a tubular gland which provides a secretion for the membranes of the 
brain and spinal cord. It is remarkable that there should be so large a 
cleft in the pituitaries of the dog and cat, unless it has some function. An 
external opening of the cleft is frequently seen, but it may quite well be 
an artiticial one, and in carefully prepared specimens I have been unable 
to tind it. The anterior lobe is not a tubular gland, and the only cells that 
can pour a secretion into the cleft are those of the pars intermedia. On 
the other hand, the cleft is not always well developed even in the cat's 
pituitary, and may be almost entirely closed by fusion of the anterior lobe 
with the cells of the epithelial covering of the lobe. In the pituitary of 
the monkey there may be little remnant of the cleft, and certainly no 
opening from it into the subdural space. It is rare to find any histological 
evidence of a secretion into the cleft, and where colloid has been present it 
has been enclosed in a thin- walled cyst and not lying free in the cleft. In 
the rabbit's pituitary it is not uncommon to find the cleft filled with red 
bone-marrow and fat cells. Lymph spaces in the dura mater below the 
pituitary body are frequently seen, but there is no evidence that they are 
specially connected with it : they probably belong to a system of lymphatics 
present at the base of the brain. 

Conclusions and Summary, 

Development of the pituitary body begins very early in embryonic life. 
In mammals the epithelial portion is derived entirely from the ectodermic 
wall of the buccal invagination known as Rathke's pouch. Its origin is 
single and mesial. The epithelium is ditierentiated at an early stage into 
two parts, which show difierences in arrangement, structure, and vascularity. 
One of these, which has been termed the pars intermedia, is closely adherent 
to the wall of the cerebral vesicle from its earliest appearance, and remains 
attached to it throughout life. It forms a layer of cells of varying thick- 
ness over body and neck of the posterior lobe and adjacent parts of the 
brain, and tends to arrange itself in positions where it can approach as near 
as possible to the cerebro-spinal canal. The cells of the pars intermedia 
are further characterised by the absence of deeph^ staining granules from 
their protoplasm, by their tendency to form a colloid substance in the adult 
organ, and by their compai-atively poor supply of blood-vessels. Its relation 
to the nervous part of the pituitary and to the adjacent wall of the brain 
tends to become even more intimate as development proceeds, by the 
ingrowth of its cells into these structures. 



The Development of the Mammalian Pituitary Body 183 

The other portion of buccal epithelium (ijives rise to the anterior lobe 
proper. The lower portion of Rathke's pouch, which is not adherent to 
the brain, forms a solid mass of cells which grow into surrounding blood- 
channels and into the cavity of the pouch itself. Its cells become tilled with 
deeply staining granules and form columns without any lumen, separated 
from one another by blood-channels of a sinusoidal character. The original 
cavity of Rathke's pouch persists as a narrow cleft separating the anterior 
lobe proper fi-om the epithelial investment of the posterior lobe. The cleft 
remains a closed cavity, which varies in extent in different species and in 
different individuals of the same species. In the cat embryo there is 
evidence of some proliferation of cells of the anterior end of the fore-gut ; 
these soon disappear, and do not enter into the formation of the adult 
pituitary. 

The infundibulum is an invagination of part of the wall of the thalam- 
encephalon which is adherent to the anterior and upper wall of Rathke's 
pouch. It therefore possesses an epithelial covering derived from the latter. 
The infundibular process grows backwards, and, in the cat, retains its 
central cavity. It is lined by ependyma cells which during development 
become elongated, so that ependyma fibres run obliquely in its neck. The 
body of the lobe consists of ependyma and neuroglia cells and fibres ; no true 
nerve cells are present in it, and there is very little connective tissue. The 
posterior lobe of the pituitary is, from the first, a composite structure of 
epithelium of the pars intermedia and of neuroglia and ependj^ma, and the 
relations between the two tissues become more and more intimate. Its 
vascular supply is derived from a different source from that of the anterior 
lobe ; blood-vessels grow into it at its posterior-superior angle and form true 
capillaries in the lobe. 

The intimate nature of the connection between the wall of Rathke's 
pouch and the cerebral vesicle, and the maintenance of a close relationship 
between the cells of the pars intermedia and the cerebro-spinal canal, render 
it probable that the pituitary body of mammalia is to be regarded as the 
representative of an old mouth opening into the canal of the central 
nervous system. Such an arrangement exists in its simplest fOrm in the 
Ascidian larva. A connection between Rathke's pouch or original mouth- 
cavity and the interior of the infundibulum is sometimes seen in the 
developing cat, and in the adult cat it is not uncommon to find epithelial 
cells, derived from the buccal cavity, lying inside the posterior lobe in com- 
munication with the third ventricle of the brain. The relations between 
epithelium and nervous tissue are not accidental in the maunnalian pitui- 
tary. The latter may have arisen, as Willey stated, from a functional 
neuropore, but is more likely to have been produced in the manner indicated 
by Kupffer. There is less probability of Dohrn's view being a correct 
solution of the problem. The question is one of great interest, and is by no 
means settled. The anterior lobe proper is a gland whose secretion must 
enter the blood directly, and so pass into the general circulation. The pars 



184 Herring 

interiaoflia, on the other liand, appears to .secrete into the brain tissue, and 
must be regarded as a brain gland. Tlie nature of these secretions, and the 
question as to whether that of the pars intermedia is modified by its passage 
through brain substance, await further investigation. 

I have to express my indebtedness to Mr Richard Muir for the care 
with which he has executed the accompanying illustrations. The expenses 
of the research have been defrayed by a grant from the Earl of Moray 
fund for the prosecution of research in the University of Edinburgh. 



LITERATURE REEERRED TO IN THE TEXT. 

(1) Andriezbn, "The Morphology, Origin, and Evolution of Eunction of the 
Pituitary Body, and its Relation to the Central Nervous System," Brit. Med. Journ., 
vol. i., p. 54, 189-1. 

(2) V. Baer, Ueber Entwickelungsgeschichte der Thiere, Th. i., S. 104 und 
130, 1828. 

(3) Balfour, "On the Development of the Elasmobranch Eishes," Quart. 
Journ. Micr. Science, vol. xiv., p. 362, 1874. 

(4) Beard, J.. "The Old Mouth and. the New," Anat. Anz., Bd. iii., S. 15, 1888. 

(5) Berkley, " The Finer Anatomy of the Infundibular Region of the Cerebrum, 
including the Pituitary Gland," Brain, vol. xvii., p. 517, 1894. 

(6) Boeke, "Die Bedeutung des Infundibulums in der Entwickelung der 
Knochenfische," Anat. Anz., Bd. xx., S. 17, 1902. 

(7) DoHRN, " Studien zur Urgeschichte des Wirbelktirpers," Mittheilungen der 
Zoolog. Station zu Neapel, iii., S. 252. 

(8) DuRSY, Zur Entwickelungsgeschichte des Kopfes, S. 76, Tu])ingen, 1869. 

(9) Gaskell, "On the Origin of Vertebrates," Journ. of Anat. and Physiol., 
vol. xxxiv., p. 534, 1900. 

(10) Gaupp, E., "Ueber die Anlage der Hypophyse bei Saurien," Arcli. f. mikr. 
Anat., Bd. xlii., S. 576, 1893. 

(11) Gemelli, " Sur la structure de la region infundibulaire des poissons," Journ. 
de I'Anat., Annee xlii. 

(12) GoTTB, Entwickelungsgeschichte der Unke, S. 288, 317, Leipzig. 

(13) Haller, B., " Untersuchungen liber die Hypophyse und die Infundibular- 
organe," Morpholog. Jahrbuch, Bd. xxv., S. 31, 1896. 

(14) His, Untersuchungen iiber die erste Anlage des Wirbelthierleibs, S. 134, 
Leipzig, 1868. 

(15) JuLiN, "Rechei'ches surl'organisation desAscidies simples. — Sur I'hypopliyse 
et quelques organes qui s'y rattachent," Arch, de Biologie, ii., p. 211, 1881. 

(16) KoLLiKER, "Entwickelungsgeschichte des Menschen und der lioheren 
Thiere," 2te Aufl., S. 527-531, Leipzig, 1879. 

(17) KuLLiKER, Gewebelehre des Menschen, Bd. ii., S. 605, 1896. 

(18) IvRAUSHAAR, " Entwickelung der Hypophysis und P]piphysis l)ei 
Nagethieren," Zeitschr. f. wiss. Zoolog., Bd. xli., S. 79, 1885. 



The Development of tlie Mammalian Pituitary Bcjdy 185 

(19) KuPFFEU, "Die Deutung des Hirnaiilianges," Sitzber. tier Gesellschaft f. 
Morpholog. u. Physiolog. in Miinchen, iS. 59, Juli 1894. 

(20) KuPFFEK, Studien zur vergleichende Entwickelungsgeschichte des Kopfes 
des Kranioten, Heft i., 1893. 

(21) MiHALKOvics, " Wirbelsaite und Hirnanhang," Arch. f. raikr. Anat., 
Bd. xi., S. 389, 1875. 

(22) MiNOT, Human Embryology, pp. 571-575, New York, 1892. 

(23) MiJLLER, W., "Ueber Entwickelung und Bau der Hypophysis und des 
Processus Infundibuli Cerebri," Jenaische Zeitschr. f. Medicin, Bd. vi., S. 354, 1871. 

(24) NusBAUM, J., " Einige neue Thatsachen zur Entwickelungsgeschichte des 
Hypophysis cerebri bei Saugethieren," Anat. Anz., Bd. xii., S. 161-167, 1896. 

(25) Rabl-Ruckhard, " Das Grosshirn der Knochenfische und seine Anhangs- 
gebilde," Arch. f. Anat. u. Physiol., Anat. Abth., Jahrgang 1883, S. 317. 

(26) Rathke, " Ueber die Entstehung der Glandula pituitaria,'' Arch. f. Anat. 
Physiol, u. wiss. Med., Bd. v., S. 482, 1838. 

(27) Rathke, Entwickelungsgeschichte der Wirbelthiere, S. 100, Leipzig, 1861. 

(28) Reichert, Das Entwickelungsleben im Wirbelthierreich, S. 179, 
Berlin, 1840. 

(29) Reichert, Der Bau des menschlichen Gehirns, Th. ii., S. 18, Leipzig, 1861. 

(30) Retzius, " Ueber ein dem Saccus vasculosus entsprechendes Gebilde am 
Gehirn des Menschen iind anderer Siiugethiere," Biolog. Untersuchungen, neue 
Folge, Bd. vii., S. 1-5, 1895. 

(31) Saint-Rkmy, " Sur la signification morphologique de la poche pharyngienne 
de Seessel," C. R. de la Soc. de Biologic, p. 423, Paris, 1895. 

(32) Salvi, "Sopra la regione ipofisaria e la cavita premandibolari di alcuni 
Saurii," con fig., Studi Sassaresi, An. i., Sez. ii., F. 2, S. 17. Quoted from Schwalbe's 
Jahresber., Bd. vii., Th. 3, S. 463, 1902. 

(33) Salzer, "Zur Entwickelung der Hypophyse bei Siiugern," Arch. f. mikr. 
Anat., Bd. Ii., S. 55, 1898. 

(34) Valenti, "Sullo sviluppo dell' ipofisi," Anat. Anz, Bd. x., S. 538, 1895. 
" Sulla origine e sul significato dell' ipofisi," Atti Accad. Med. Chir. Perugia, vol. vii., 
fasc. 4, 1894. 

(35) Van Wijhe, ''Ueber den vorderen Neuroporus imd die phylogenetische 
Function des Canalis neurentericus der AVirbelthiere," Zoolog. Anz., Rd. vii., 
8. 683, 1884. 

(36) Willey, Amphioxus and the Ancestry of Vertebrates, p. 285, New 
York, 1894. 



I 

I 



THE PHYSIOLOGICAL ACTION OF EXTRACTS OF THE PITUI- 
TARY BODY AND SACCUS VASCULOSUS OF CERTAIN 
FISHES. Preliminary Note. By P. T. Herring. (From the 
Physiology Department, Universit}^ of Edinburgh.) 

{Received for publication \Oth March 1908.) 

In certain fishes, elasmobranchs and teleosts, the infundibular reo-ion of 
the brain is distinguished among other things by the presence of an 
extremely vascular gland — the saccus vasculosus. In elasmobranchs, e.g. 
the skate (Raja batis), the saccus vasculosus is large and paired, and its lobes 
open by a common median passage into the infundibulum, and so into the 
third ventricle of the brain. In teleosts, e.g. the cod (Gadus morrhua), the 
saccus vasculosus is single and situated in the middle line between the lobi 
inferiores ; it opens into the infundibulum immediately behind the posterior 
lobe of the pituitary. In both the skate and the cod the saccus vasculosus 
consists of a complicated sac lined by a single layer of columnar epithelium 
which is separated from numerous large and thin-walled blood-vessels by 
a thin basement membrane. The wall is thrown into frequent folds, 
especiallj^ in the cod, thereby reducing the size of the cavit}', but increasing 
the surface area of its interior. 

Extracts of the saccus vasculosus made by boiling it in Ringer's fluid 
have no marked physiological action when injected into the blood-vessels 
of a cat. Whether taken from the skate or the cod they do not produce 
a rise of blood pressure, but rather a slight fall ; kidney volume is a little 
increased, but there is no effect upon the secretion of urine. The results 
are practically those of an injection of Ringer's fluid. The saccus 
vasculosus of fishes does not yield the active principles which are 
characteristic of the posterior lobe of the mammalian pituitary-. Gentes,^ 
on anatomical grounds, believes that the saccus vasculosus is a ventral 
choroid plexus. 

The pituitary body of the skate, and, according to Gentes, of elasmo- 
branchs generally, has no posterior lobe. Neither does it possess the 
granular cells of the anterior lobe of higher vertebrates. It is, neverthe- 
less, a large body, and presents the features of an internally secreting gland. 
Extracts produce in the cat a slight fall of blood pressure, a dilatation of 

^ Gentes, " Reclierches snr rhyj)opliyst' et le sao vasculaire des vertebres," Soc. scientif. 
d'Arcachoii, Station liiologujuc, Travaux (ies labnratoires, p. 2(58, fasc. 1. Bordeanx, 
1907. 



188 Physiological Action of Extracts of the Pituitary Body 

the kidney, and some increase in urine How. The blood pressure raising 
substance is apparently absent, or is in such small amount that its action 
is entirel}^ masked by the depressor effects of the glandular extract. 
Extracts of the lobi inferiores and adjacent brain substance produce the 
well-marked fall of blood pressure and diminution of kidney volume which 
are characteristic of extracts of the central nervous system. 

In teleosts the pituitary body consists of an anterior lobe proper 
characterised by the presence in it of deeply-staining granular cells, an 
intermediate part of smaller clear cells, and a nervous portion. The latter 
is surrounded and invaded by the cells of the pars intermedia. Extracts 
of this portion of the pituitary body, pars nervosa and pars intermedia, 
produce in the cat the typical effects of extracts of the posterior lobe of 
mammals, viz. rise of blood pressure, dilatation of the kidney, and increase 
of urine. 



NOTE ON THE ACTION OF PITUITARY EXTRACTS UPON THE 
ENUCLEATED FROG'S ■ EYE. By W. Cramer. (From the 
Physiology Department, Edinburgh University.) 

{Received for -publication bth March 1908.) 

Extracts of the posterior lobe of the pituitary body of the ox produce a 
distinct dilatation of the pupil of the enucleated frog's eye. 

By using strong extracts made from the desiccated posterior lobe the 
action on the pupil becomes apparent within an hour or two. The following 
experiment, in which a solution made from 0*4 g. desiccated pituitary in 
3 c.c. of Ringer's solution was used, may be taken as an example. The 
size of the pupil, which in complete contraction is slit-like, was measured 
by determining the length of the short diameter by means of a pair of 
compasses. 



Time. 


Left eye 4- pituitary 
extract. 


Eight eye+Kinger's 
solution. 


Size of pupil. 


Size of pupil. 


12.30 p.m. (before the exjjeri- 
ment) 

2 „ ... 

3 „ ... 

4 „ ... 


1 mm. 

3 mm. 

4 mm. 
4-5 mm. 


1 mm. 

1 mm. 

1 mm, 

0"75 mm. 



The action of a solution of adrenalin (hemisin) 1 : 10000 is more rapid but 
not so lasting, as will be seen from the following simultaneous experiment. 



Time. 


Left eye + adrenalin 
1 : 10000. 


Right eye + Einger's 
solution. 


Size of pupil. 


Size of pupil. 


12.30 p.m. (before tlie experi- 
ment) 

2 , 

3 ]] '. . '. 

4 „ ... 


2-5 mm. 

4 nun. 

4 mm. 

3-5 mm. 


3 mm. 

2-5 mm. 
2 mm. 
2 mm. 



190 Action of Pituitary Extracts upon the Enucleated Frog's Eye 

If more dilute pituitary extracts are used the effect only becomes visible 
after twelve or more hours. After that time the pupil of a frog's eye 
immersed in Ringer's solution is, as a rule, completely contracted. In the 
following experiment 3 c.c. of a 2 per cent, pituitary extract were used. 



Time. 


Right eye and pituitary 
extract. 


Left eye and Ringer'.s 
sohition. 


Size of pupil. 


Size of pupil. 


Before the e.xperiment . 
After 16 hours 


2-5 mm. 

3 mm. 


2-5 mm. 
0"75 mm. 



A frog's eye placed in Ringer's solution together with the fresh posterior 
lobe of a cat's pituitary body shows a distinct dilatation after sixteen 
hours. If the anterior lobe is used instead of the posterior lobe the pupil 
is completely contracted after sixteen hours. 

The experiments of Schafer and Herring^ have shown that the effects 
of pituitary extracts on the kidney and on the circulation are independent 
of each other, and probably due to different substances. The question arises 
which substance is concerned in the action of the pituitary extracts on the 
pupil. Our observations tend to show that the substance acting on the 
pupil is not identical with the substance which stimulates renal activity. 
For some of the desiccated preparations after having been kept for several 
months proved inert when tested against a frog's eye ; their intravenous 
injection, however, still produced a marked effect on the flow of urine, 
while at the sam.e time their action on the blood-vessels was very weak. 
But any preparation which produced a dilatation of the pupil gave, 
on intravenous injection, a typical effect both on the kidney and on the 
circulation. 

' Schiifer and Herring, Pliil. Trans., Series B, vol. cxcix., 1906, p. 1-29. 



THE ACTION OF YOHIMBINE ON MEDULLATED NERVE, WITH 
SPECIAL REFERENCE TO FATIGABILITY. By John Tait 
and Jas. A. Gunn. (From the Physiology Department, University 
of Edinburgh.) 

{Received for publication 21.s< February 1908.) 

Experiments carried out in recent years by Gotch and Burch (1), 
Boycott (2), Boruttau (8), and F. W. Frohlich (4), have shown that when 
medullated nerve is thrown into activity by an external stimulus a certain 
short period of time must elapse before it can function again. This period 
of inexcitability, or refractory period, is normally very short — not more 
than -002 second for the sciatic nerve of the frog — but can be much pro- 
longed by subjecting the nerve to special conditions. Thus low temperature 
(1) (2), anaesthesia (3) (4), and asphyxia (4), all greatly prolong the refractory 
period. 

The readiest method of demonstrating the existence of this refractory 
period is to excite the nerve of a nerve-muscle preparation by two successive 
maximal stimuli separated by a very short interval of time. By the 
response of the attached muscle it is then possible to tell whether the nerve 
has conducted two excitations or only one. If the muscle response is a 
summated one it is evident that both excitations have been transmitted ; if 
summation does not occur, then the second stimulus must have been in some 
way ineffective. Absence of summation occurs only when the interval of 
time between the two maximal stimuli is sufficiently short. In such cases 
it has been shown by means of the capillar}- electrometer (1) that tlie block 
to the second excitation is seated in tlie nerve. 

The length of the refractory period would seem to be dependent on the 
intensity of the preceding excitation. Generally speaking, it has been found 
that the stronger the stimulus applied to the nerve the longer does the 
nerve take to recover its functional capacity. By combining powerful 
stimulation with anaesthesia or asphyxia or cooling of the nerve, one would 
therefore expect to get a maximal refractory period, and indeed Frohlich 
was able to prolong it to '1 second (4). 

Besides the method of applying two successive stimuli it is obvious that 
a series of rapid recurring stimuli might be applied to the nerve ; and if a 
sufficiently high rate of excitation could be attained, one would expect that 
at least some of the excitations would be ineffective. 

VOL. 1. — Al'RIL 1908. 13 



192 Tait and Gunn 

Experiments with a rapid rate of stimulation have been carried out 
on numerous occasions. Neglecting in the meantime those in which the 
enormously high rates afforded by the discharge of Leyden jars, etc., have 
been used, where the interval between the individual stimuli is of a 
different order of magnitude from the refractory period, we shall mention 
experiments where the rate of stimulation has been over 400 a second and 
yet not greater than tens of thousands. 

Bernstein (5), using a rate of 500 a second, obtained only a single 
initial contraction of the muscle. Roth (6), with a rate of 1000-5000, 
obtained tetanus. Langdon and Schenck (7), with a rate of 1800-2000 
per second, also got tetanus. Kronecker (8), using a special device whereby 
he claimed to attain a rate of 20,000 per second, found tetanus, which on 
repetition became an initial twitch and subsequently failed. From the want 
of uniformity in these results it was for a time difficult to draw any general 
conclusion. 

Of late, however, thanks to the work of Wedensky (9) and of 
Frohlich (4), it has become possible to reach a definite generalisation. 
Wedensky, who happened to combine the method of rapid stimulation 
with anaesthesia of the nerve, found that strong excitations at a rate of about 
100 per second applied to the proximal end of a nerve whose middle portion 
is deeply ansesthetised, produces simply a single twitch of the muscle. 
Frohlich pointed out that this twitch is of the same height as the twitch 
evoked by one single maximal excitation, and that the result occurs only 
when the successive stimuli are separated by an interval less than the 
corresponding refractory period of the angesthetised nerve. When, on the 
other hand, the stimuli succeed each other at an interval greater than the 
refractory period, then tetanus of the muscle occurs. 

The fact that one initial maximal twitch follows upon repeated stimula- 
tion of the nerve shows that only the first excitation of the series has taken 
effect ; this excitation prevents the second from being effective, the second 
prevents the third, and so on : consequently when once the excitations are 
applied at a sufficiently rapid rate the nerve refuses to conduct any more 
than the first excitatory process. 

Such an effect, though at first sight suggestive of fatigue of nerve, is not 
necessarily fatigue. It is conceivable that a conducting mechanism built on 
simple physical principles might give the same result. Nevertheless, by a 
closely analogous method Frohlich succeeded in showing that the nerve 
does actually become fatigued when subjected to rapid stimulation. When 
during anaesthesia of the nerve he selected a rate of stimulation which just 
about coincided with the definite refractory period corresponding to the inten- 
sity of stimulation used and to the given degree of anaesthesia, he obtained, 
not a single twitch, but a tetanus of peculiar form. This tetanus, instead of 
gradually climbing, after the normal fashion of a muscle tetanus, attained 
its maximum almost immediately, and then rapidly fell off in height until 
in the space of a second or so the muscle ceased to contract altogether. 



The Action of Yohimbine on Medullated Nerve 193 

Interruption of the rhythmical stimulation but for a second sufficed to 
restore the conductivity of the nerve to its previous condition. 

Now, on the assumption that the degree of anaesthesia does not increase 
during the short period of observation, this peculiar form of tetanus points 
to a progressive lengthening of the refractory period due to the con- 
tinuous activity of the nerve, and any change in the direction of 
depression of function which is solely due to activity is fatigue. That the 
gradual prolongation of the refractory period is not due to a temporary 
and coincident deepening of the anaesthesia is sufficiently clear from the 
consistent regularity with wliicli the effect occurs even when the anaesthesia 
is passing off. 

Frohlicli's work, while establisliing the fact that nerve can be fatigued 
— a fact of fundamental importance in regard to our views as to the nature 
of the nerve impulse— serves at the same time to emphasise the high powers 
of restitution possessed by the structure. Even when anaesthetised almost 
to the point of complete absence of conductivity, the nerve still required to 
be stimulated uninterruptedly in order to maintain the fatigued condition. 
Interruption of the rhythmical stimulation for a fraction of a second left 
time for an apparently complete recovery. , 

The anesthetic agents which were found by Wedensky and Frolich 
to prolong the refractory period of nerve include most of the common 
anaesthetics known to medicine (ether, chloroform, cocaine, phenol, etc.). In 
spite of the chemical differences between these substances, the kind of 
change produced in nerve by means of all of them seems to be virtually the 
same. Further, this change corresponds identically with that caused both 
by asphyxia (4) and by cooling (9), so that one can recognise a common 
element in the action of all of these things. 

The present paper deals with the changes produced in nerve by means 
of yohimbine. This substance, which is an alkaloid derived from the bark 
of the Yohimbehe tree (10), was shown by Magnani in 1902 to be a local 
anaesthetic (11). We have investigated its action on the sciatic nerve of 
the frog, availing ourselves of the method of Wedensky— i.e. rapid 
rhythmical stimulation, in order to show changes in conductivity. The 
response of the attached gastrocnemius muscle was used as an index of 
the condition of the nerve. 

The investigation has shown : — 

(1) That fatigue changes in nerve may be demonstrated more readily 
by the application of this substance than by any method known to us. 

(2) Yohimbine seems to differ somewhat in action on nerve from other 
anaesthetics. 

For our experiments we used a 2 per cent, solution of yoliimbine 
lactate in Ringer's fluid, which we applied to the middle portion of the 
dissected nerve. In order to keep the solution in contact with this part of 
the nerve, strips of blotting-paper (usually about 3 centimetres long) 
moistened with the solution were laid under and over the middle portion. 



194 



Tait and Gunn 



Thus the proximal and distal ends of the nerve were left unaffected by the 
solution, and to each. of these parts a pair of electrodes was applied. To 
ensure that the solution should not run along the uncovered parts of the 
nerve and affect either the proximal or distal ends, the middle portion was 
kept at a slightly lower level than the two ends. The whole nerve-muscle 
preparation was kept in a moist chamber. The electrodes were connected 
by means of a Pohl commutator from which the cross wires were removed 
with a standard Kronecker coil (original pattern), in the primary circuit 
of which was an accumulator charged to 4^ volts. 

The result of soaking the nerve for a number of hours (the time varied 
from two to three hours in our experiments) is to abolish conductivity in 




30 30 3» 




I 



l'i(i. 1 (reduced to two-thirds). —Yohimbine lactate, 2 per cent., applied for 1 hour to 3 cm. of nerve. Proximal 
stimulation, rate 128 per sec. Intensity, 30 Kronecker units. Rate of drum, 1 mm. per sec. Tracing (2) 
taken 40 sees, after tracing (1). 

Tracing (1) shows (i.) irregularity of tetanic responses ; (ii.) diminution in height of successive tetanic responses. 
I'racing (2) shows (i.) diminution in extent of successive tetanic responses with diminishing intervals between stimulations;, 
(ii.) subsequent improved responses with increased intervals of rest between stimulations. 

the part affected by the yohimbine solution. As is the case when other 
anaesthetics are applied to nerve, the abolition of conductivity does not 
occur abruptly but comes on gradually, so that long before the nerve has 
actually lost the power of conduction, changes can be detected which 
indicate a depression of function. 

Thus when the proximal end of the nerve is stimulated at some fixed 
rate lying between 100 and 200 excitations per second, the tetanic 
responses of the muscle begin to undergo a change ; instead of being 
smooth-topped, they become irregular in form, and the muscle, instead of 
remaining in continuous contraction, ultimately twitches more or less 
spasmodically (see fig. 1). On the other hand, the muscle response to distal 
stimulation is a smooth and regular tetanus, showing that the irregularity 



The Action of Yohimbine on Medullated Nerve 195 

of the muscle tetanus in the former case is not due to fatigue of the muscle 
or of the nerve ends in the muscle. In this respect Yohimbine resembles 
in its action other anaesthetics. 

At a somewhat later stage the abnormality in the muscle response is 
clearly seen to be of a definite type. To any given series of continuously 
applied rhythmical excitations the immediate response of the muscle is a 
summated tetanus which quickly begins to decline in height, and finally 
becomes feeble and irregular, or ceases altogether. Thus the general form 
of the tetanus approaches that of the " fatigue tetanus " described by 
Frohlich. (See fig. 1, tracing (1), and fig. 2.) 

In the case of other angesthetic agents applied to nerve it was found 
by Wedensky that the muscle response is largely dependent on the 
intensity of stimulation used. Wedensky showed (9) that at any given 
stage of anaesthesia, provided the rate of stimulation is kept constant, there 
is one definite intensity of stimulation (optimum of intensity) which 
produces a maximal height of tetanus ; intensities either above or below 
this optimum cause a less height of tetanic response. In other words, when 
the nerve is anaesthetised, say by ether or cocaine, weak rhythmical 
stimulation produces a tetanus of submaximal height, stimulation at some 
moderate intensity causes maximal height of tetanus, while strong 
stimulation produces again submaximal tetanus. It is found, too, that 
tetani of the form which Frohlich calls "fatigue tetani" are more readily 
obtained with strong stimulation. Furthermore, with deep anaesthesia 
and strong rhythmical stimulation, the muscle response, as already 
mentioned, is a single twitch of the same height as the twitch evoked by 
one single maximal excitation (4). The same is the case when nerve is 
asphyxiated (4). A similar effect has been shown by one of us (Tait) to 
occur when nerve is cooled. All these facts indicate that under these 
conditions the refractory period of the nerve corresponding to strong 
stimulation is longer than that corresponding to weak. 

Yohimbinised nerve does not conform in this regard to nerve subjected 
to these other influences. At almost all stages of yohimbine anaesthesia in 
which alterations of the muscle response to rhythmical stimulation can be 
detected, this response takes the form of a "fatigue tetanus" — i.e. the last 
part of the tetanus is at least markedly lower than the first. Furthermore, 
the highest and best sustained muscle response is not produced by 
stimulation at moderate intensity, but in every case strong stimulation is 
more effective than any moderate stimulation as regards both the height 
and duration of the corresponding tetanus. (See fig. 2.) Thus it is 
evident that the refractor}^ period of yohimbinised nerve does not increase 
with the intensity of the stimulation. If anything, the contrary would 
seem to be the case. 

An examination of the tracings in figs. 1 and 2 shows that when series 
of rhythmical stimulations are applied in closely succeeding sets or groups 
to the proximal end of yohimbinised nerve, the successive muscular 



196 



Tait and Gunii 



responses tend to fall off in heiglit with repetition of the successive series 
of stimulations. Thus in %. 1, tracing (2), it is very clear that the 
responses to the first six series of stimulations become progressively lower 
and lower. This eflfect may be due either to rapidly deepening anaesthesia 
of the nerve or to fatigue. That it is due to some fatigue condition 
and not to progressive and rapid aneesthesia, is indicated by the fact that 




Fig. 2 (reduced to about one-half).— Length of nerve anaesthetised, 3 cm. Duration of application 
of yohimbine, 40 minutes. Rate of stimulation, 144 per sec. Rate of drum, 1'5 mm. per 
sec!^ Six series of responses are shown, corresponding to intensities varying from 30 to 500 
Kronecker units. 
Note (i.) the tetani are all of the "fatigue " form ; (li.) the responses corresponding to strong stimulation are more 
marked than those corresponding to weak ; (iii.) in any given series with constant intensity of stimulation 
the height of the responses varies as the duration of the period of rest between stimulations. 

if longer intervals of rest are allowed between the successive sets of 
excitations, the effect does not occur. Further, the effect is most marked 
when the intervals of rest between successive series of stimulations are 
made progressively less and less, as is the case in the first six responses of 
fig. 1, tracing (2), or in the middle series of responses in fig. 3. 



The Action of Yoliimbine on Medullated Nerve 



197 




198 Tait and Guiin 

On the other hand, when the successive muscle responses have become 
less and less marked as a result of a steady diminution in the intervals of 
time between the sets of stimulations, they gradually increase in height 
again when the intervals of rest between sets of stimulations are made 
longer and longer. (See the last three responses in fig. 1, tracing (2), or the 
last five responses in fig. 3.) Provided the same short interval of time 
elapses in each case between the successive sets of stimulations, the second, 
third and fourth, etc., responses of the muscle may be all of about the same 
magnitude, whereas the first response of any series— beginning after an 
adequate interval of rest— is more marked. (See generally the tracings in 
fio-s. 2 and 3.) At any given stage of anaesthesia, therefore, the efficiency 
of the muscle response induced by stimulation of the proximal end of the 
nerve is directly proportional to the interval of time during which the pre- 
paration has rested from activity. This is clearly a fatigue phenomenon. 

During the later stages of anaesthesia with yohimbine the muscle 
responses to rapid rhythmical stimulation tend to resemble simple muscle 
twitches rather than tetani (see fig. 4), and this is the case whether strong 
or weak stimulation is used. These seeming simple twitches are, however, 
in reality summated muscle responses. This is readily seen when one 
compares the height of the muscle contractions evoked on the one hand by 
rhythmical stimulation, and on the other by single maximal break shocks 
applied to the proximal end of the anaesthetised nerve (care being taken in 
each case to examine the preparation after an adequate interval of rest). 
In every instance the effect of rhythmical stimulation is to produce a much 
higher muscle response than that produced by a single maximal excitation. 
In this respect the action of yohimbine is once again different from that of 
other anaesthetics, for in the later stages of anaesthesia with, say, ether or 
cocaine, rapid stimulation, especially when strong, produces indeed a muscle 
twitch, but this twitch is of the same height as the response to one single 
maximal excitation of the nerve. 

If we come now to the interpretation of this phenomenon we must 
conclude that when a series of excitatory processes are made to travel in 
rapid succession from a normal portion of nerve into a portion deeply 
anaesthetised with yohimbine, probably the first few excitatory processes 
succeed in traversing the anaesthetised part, but the passage of these unfits 
the affected portion of nerve for the immediate transmission of further 
excitatory processes. Only after an adequate interval of rest is the nerve 
able to function again, and an examination of the tracings in fig. 4 will show 
that this interval must be spread over many seconds to restore the conduct- 
ing mechanism to exactly the same degree of functional capacity as before. 
From the fact that even w^ith a relatively rapid rate of stimulation 
(between 100 and 200 per second) the deeply anaesthetised nerve is at the 
start able to transmit more than the first excitatory process, while after 
the passage of a few excitations it temporarily ceases to function, we infer 
that the refractory period of yohimbinised nerve is dependent, not so much 



The Action of Yohimbine on Medullated Nerve 199 

on mere degree of anaesthesia by itself as on the extent to which the 
anaesthetised nerve is within any given short period of time thrown into 
activity. It is activity during the anaesthesia rather than the anaesthesia 
itself which causes the prolongation of the refractory period. 

The progressive impairment of function of the nerve with activity is 
equally well shown if during this stage of deep anaesthesia the nerve is 
stimulated at a slow rate (4 per second), with break shocks of a strength 
that is just maximal. (See tig. 4, last two tracings.) Then the individual 
twitches of the muscle consistently decline in height with each repetition 
of the excitation and finally, after a certain small number of responses have 
occurred, die away entirely. A rest of a considerable number of seconds 
(not more than thirty) suffices to restore the nerve to its previous condition, 
when the same process can be repeated again by rhythmical stimulation at 
the same slow rate. On the other hand, if the interval of rest be not 
sufficientlj^ long (say only two to five seconds), the process of recover}^ is 
not so complete, and the next set of muscle responses are fewer in number 
and of less height. Meanwhile, if the nerve is stimulated at a part distal 
to the anaesthetised portion, the muscle responds by a continuous series of 
maximal twitches. (See fig. 5.) 

Such experiments demonstrate in striking fashion not only the existence 
of fatigue in yohimbinised nerve, but also the gradual nature of the 
recovery from fatigue. In every case after a period of continuous activity 
the nerve becomes exhausted and requires a rest of a considerable number 
of seconds before it has regained its previous state of functional efficiency. 
Nevertheless, by stimulating the nerve after a shorter period of rest it can 
be shown that the recovery process, though incomplete, has still gone on to 
a certain extent. Further, the fact that in every case recovery does occur 
after fatigue indicates that nerve is characterised not so much by non- 
fatigability as by the possession of an extremely efficient mechanism for 
repair after fatigue. 

How far the refractory period of nerve may be prolonged when the 
nerve is under the influence of yohimbine anaesthesia we have not deter- 
mined exactly. Much depends on the signification in which the term 
" refractory period " is used. If we take the term to mean that period of 
time which elapses before one maximal stimulus following upon another of 
equal intensity can be fully effective, then the refractor}^ period has been 
prolonged to at least '25 second. If, however, we extend the definition to 
include the period of time necessary for complete recovery of the nerve 
after the application of a series of stimuli applied in rapid succession, then 
the refractor}^ period has been prolonged to a number of seconds (more 
than five). 

In the later stages of yohimbine anaesthesia when the proximal end of 
the nerve is stimulated by isolated maximal shocks at long intervals, the 
corresponding muscle responses are definitely lower in height than those pro- 
duced by similar stimulation of the distal end. This points to a diminution 



200 



Tait and Gunn 




= ■2 



■§«'-f 



;■- .o^ tcs 



m g 


S = =«^ >> 


■^ 3 


^o -Of g = 


°2 




"^ "o 


il 'III! 


:;^i^ 


■£i ::«sl 




^1 ll^-s 




•Slsil'^fe 


.^c-^ 


■2S«55Sg 


~ ii 


oe-^So-?:! 


•:| 


fi-iii 


o 


II • lel 



The Action of Yohimbine on Medullated Nerve 201 

in anipHtude of the excitatory processes as they traverse the yohimbinised 
area. When the conductivity by deepening of the anaesthesia is just 
about to disappear, these responses become minimal. In this respect the 
anaesthetic action of yoliimbine corresponds with the action of coohng, of 
aspliyxia, and of anaesthesia produced by means of the more commonly 
used anaesthetics. 

Before concluding, we ought to say that the fatigue effects which we 
have ascribed to the action of yohimbine can be due only to a change in 
the nerve itself. In a normal nerve-muscle preparation, strong stimulation, 
especially of the distal end of the nerve, may so alter the nerve-endings in 
the muscle that subsequent stimulation of the nerve may produce effects 
which might be erroneously ascribed to fatigue of the nerve. In all the 
experiments carried out by us which show fatigue phenomena this fallacy 
is excluded by the fact (1) that the intensity of the stimulation applied to 
the nerve was at no time more than just maximal ; (2) that the fatigue 
phenomena in each case appeared before distal stimulation was used. 



1. A two per cent, solution of yohimbine lactate has been 
applied to the middle portion of the sciatic nerve of a frog's 
gastrocnemius preparation, and alterations in the conductivity 
of the nerve observed by means of the muscular response to 
rhythmical stimulation applied proximally to the alkaloid- 
affected portion. The rate of stimulation varied between 144 
and 4 per second. 

2. In its action on nerve, yohimbine resembles in many 
respects the already known action of other anaesthetics, of low 
temperature, and of asphyxia. It ultimately abolishes conduc- 
tivity. The process of abolition of the conductivity is gradual, 
and is characterised by a progressive diminution in the ampli- 
tude of excitatory processes which traverse the affected part of 
the nerve, and by a prolongation of the refractory period of 
the nerve. By means of it, too, fatigue changes may be shown 
to occur in the nerve. 

3. On the other hand, the action of yohimbine differs in 
important respects from that of asphyxia of low temperature, 
and of anaesthesia with ordinary agents. The tetanic responses 
of the muscle corresponding to rapid rythmical stimulation of 
the proximal end of a yohimbinised nerve are always of one type, 
and resemble the "fatigue tetani " described by F. W. Frohlich. 
In stages of deep anaesthesia it is not easy to demonstrate the 
occurrence of initial non-summated maximal twitches as a result 
of rapid rhythmical stimulation of the nerve. The duration of 
the refractory period does not seem to vary directly with the 
strength of the stimulus applied to the nerve, and is more 



202 The Action of Yohimbine on Medullated Nerve 

clearly dependent on the amount of previous activity than is 
the case when other agents are used to depress the function 
of nerve. 

4. The anaesthetic action of yohimbine lactate as applied in 
solution to the outside of a dissected nerve is characterised by 
great evenness and regularity. Partly for this reason, and 
partly because of the unusual prolongation of the refractory 
period due to yohimbine, it has been shown that nerve is a very 
convenient tissue on which to study the process of fatigue and 
recovery from fatigue. 

5. In spite of the ready fatigability of 3'ohimbinised nerve, 
complete restoration of function seems in every case to follow 
the katabolic changes due to activity. 

6. The refractory period of nerve has been prolonged to 
•25 second; while fatigue changes lasting for more than five 
seconds have been demonstrated. 

The expenses of this research were defrayed by a grant from the 
Carnegie Trust. 



KEFERENCES. 



(1) GoTCH and Burgh, Proc. Physiol. Soc, Jour, of Physiol, vol. xxiii. ; 
vol. xxiv. p. 10, 1899. 

(2) Boycott, Jour, of Physiol., vol. xxiv. p. 144, 1899. 

(3) BoRUTTAU, Arch. f. d. ges. Physiol,, Bd. Ixxxiv. p. 402, 1901. 

(4) F. W. Fruhlich, Zeit. f. allg. Physiol, Bd. iii. p. 468, 1903-4. 

(5) Bernstein, Untersuchungen liber den Erregungsvorgang im Xerven- und 
Muskel-systeme, 1871, p. 100. 

(6) Roth, Arch. f. d. ges. Physiol, Bd. xlii. p. 91, 1888. 

(7) Langdon and Sghexck, Cincin. Lancet-Clinic, 1896. 

(8) Kronegker, Arch. f. Anat., Physiol u. wiss. Med., p. 1, 1878. 

(9) Wedensky, Arch. f. d. ge.s. Physiol, Bd. Ixxxii. p. 134, 1900; ibid., Bd. c. 
p. 1, 1903. 

(10) Spiegel, Chemiker Zeitung, 1896, No. 20; Ber. d. Deutsch. chem. Ges., 
p. 169, 1903. 

(11) Magnani, Annali di Ottalmologia, 1903. 



NOTE ON THE MULTIPLICATION AND MIGRATION OF 
NUCLEOLI IN NERVE CELLS OF MAMMALS. By W. Page 
May and C. E. Walker. (With Two Plates.) 

(Beceived for publication 29th February 1908.) 

Methods. 

The tissues mainly employed in the observations here described were the 
Oasserian and cerebro-spinal ganglia of rats, rabbits, cats, monkeys, and 
chimpanzees. Other nerve cells were also similarly examined, notably 
those of the red nucleus, cerebral cortex, and other large motor and sensory 
cells throughout the central nervous system, and similar phenomena to 
those described in the present communication were also found to obtain 
in them. Further details bearing on this subject we hope to bring forward 
later. The animals were, with one or two exceptions, young adults. 

Absolutely fresh material was fixed in Flemming's fluid (strong formula), 
or a modification of Zenker's fluid (G. Arnold's). It was dehydrated by 
increasing the percentage of alcohol by 10 per cent, at each stage, and 
the greatest care was taken to prevent any possibility of maceration by 
shortening the time between removing the material from the fixative and 
getting it into 70 per cent, alcohol. Imbedding was carried out at a 
temperature of 45° Cent., and this process did not occupy more than an 
liour and a half ; thus any undue exposure to heat was avoided. The 
manipulation used in mounting and staining was according to the strictest 
eytological methods. Various staining methods were employed and all 
gave similar results, though it was found that particular methods rendered 
certain points in the observations clearer than others. The main processes 
adopted were : — 

A. Basic f uchsin, followed by methylene blue and Unna's orange tannin. 

B. Safl'ranin, followed by methylene blue and Unna's orange tannin 
(Breinl method). 

C. Thionin counterstained wnth Bordeaux red, etc. 

In the present communication the term " nucleolus " is used in its 
strictest sense, i.e. it is restricted to the structure which has sometimes 
been called the "true nucleolus," whilst such bodies as the so-called 
" chromatin nucleoli " contained in the nucleus are definitely excluded.^ 

' Cf. Wilson, "The Cell in Development and Inheritance," p 34; Macmillan, 
London and New York, 1904. Walker, "The Es.sentials of Cytology," pp. 12 and 13; 
Constable, London, 1907. 



204 Page May and Walker 

The nucleolus here dealt with is generall}- spherical in shape, occasionally 
oval. It is bounded by a definite membrane, and the contents are usually 
homogeneous or finely granular in structure. As compared with the 
chromatin present with it in the nucleus, its staining reaction is acid. 
Small aggregations of chromatin are almost invariably found lying upon 
the outer surface of the membrane of the nucleolus. These small collections 
of chromatin are often continuous with the chromatin granules in the linin 
of the nucleus, and are so closely applied to the outer surface of the membrane 
tliat it is only in specimens specially stained for the purpose that their 
actual position can be ascertained. Without this examination it is also 
difficult to be sure as to the staining reaction of the nucleolar material 
proper, as this is masked by the chromatin lying upon the surface.^ 

Several nucleoli are frequently found in the nuclei of nerve cells, indeed 
the present observations indicate that more than one is usual. Frequently 
there are four or five, or more,^ which may vary greatly in size. 

In many cases a small excrescence may be seen at one point at the 
margin of the nucleolus, and perhaps this is more usual w^here only one 
large nucleolus is present (tigs. 1 and 2). Generally, if not always, one or 
more of the chromatin aggregates already mentioned as being observed 
upon the nucleolar membrane are found to be present upon the outer 
surface of these excrescences, in sections specially stained for the purpose 
(figs. 1 and 2). Other cells are found where the excrescence has apparentl}^ 
increased in size and travelled away from the nucleolus, being still attached 
to it by a process of the membrane. This process seems to persist until 
what was originally the excrescence has grown to a considerable size (figs. 
3 to 5). Cells are also frequently to be found which exhibit traces of 
what was evidently the connection between the nucleolus and the excres- 
cence, and where the excrescence has attained a size which approaches or 
equals that of the nucleolus (figs. 6 and 7). A careful examination of the 
appearance of the nucleolus and of the excrescence up to the time when, 
according to the present interpretation, they are separated from each other, 
shows that both are exactly similar in structure, and that there may be 
several similar nucleoli undergoing similar processes in the same nucleus. 
It is, in fact, easy to find, in the same slide, every stage between the single 
nucleolus with one or more small excrescences, and two distinct nucleoli, 
which are almost exactly similar to each other. From this it is equally 
easy to pass on to cells with four and five nucleoli, or even more (fig. 8). 
The structure of these bodies is so definite that there is no possibility of 
mistaking them, in a properly preserved specimen, for any other nuclear 
constituent, such as a mass of chromatin. 

Very rarely a nucleolus may be seen dividing by a process analogous 
to amitosis, or to the division of a drop of viscous fluid into two (fig. 9).^ 

The many observations with regard to the migration of the nucleoli 

^ Method C. - We have counted as many as nine in one nucleus. 

^ "We have onlv observed this twice in inanv hundreds of cells. 



Multiplication and Migration of Nucleoli in Nerve Cells of Mammals 205 

into the cyptoplasm of the cell ^ seemed to indicate the destiny of these 
bodies, which are apparently continuously produced in the nerve cells of 
the animals here investigated, and a particular study of this phenomenon 
was therefore made. It has been claimed that the alleged migration of the 
nucleoli is due to some mechanical force such as the dragging or pressing 
of the edge of the microtome knife in the process of cutting sections, or to 
the action of gravity.- 

In the present observations nucleoli partly extruded through the nuclear 
membrane and in the cytoplasm clear of the nucleus were frequently found. 
In the great majority of these cases the apparent extrusion of the nucleolus 
was undoubtedly due to the action of the knife. Such nucleoli were always 
in an exactly similar position in relation to the nucleus and cytoplasm in 
every section in a series. Thus, if one were found half-way through or 
outside the nuclear membrane to the lower right-hand side of the centre 
of the nucleus, the other nucleoli would be displaced in exactly the same 
direction in other cells, not only in the same section, but in the other 
sections of the series. Furthermore, the nucleolus thus displaced carried 
with it a considerable amount of liniu with its contained chromatin gran- 
ules, and a large empty space in the nucleus with the broken ends of the 
strands of linin round it could be found in all these cases (fig. 10). Also 
the nuclear membrane was always definitely ruptured, and there was never 
any sign of its being reconstructed, which one would expect to find, and, 
as will be seen, actually was found, when the nucleoli were extruded 
normally ; for nuclei, even when they contain several nucleoli, are usually 
found to possess a membrane without any large ruptures, and fresh 
nucleoli are apparently constantly being produced. An interesting fact 
is, however, indicated by this occurrence. The nucleoli must be of a highly 
dense and resistent structure as compared with the rest of the cell, as the 
edge of the knife, when it happens to catch one of them, does not cut 
through it as it does through the other cellular constituents, but carries it 
bodily along for a considerable distance. 

In a, comparatively speaking, few cells, however, a true migration of 
the nucleolus was observed, and the phenomenon is here described in 
some detail, as it appears to differ in some respects from what has been 
stated to occur by other authors : some points seem to have escaped notice 
altogether. 

The staining reaction of the nucleoli when inside the nucleus was very 
marked with certain combinations used in the present investigation. Thus 
with Method A the nucleolus stains blue or violet, while with Method B 
it stains brilliant scarlet. 

The way in which the passage of the nucleolus from the nucleus occurs 
is apparently as follows : — A nucleolus lies for some time against the 

^ Montgomery, Rhode, Hatai, and others. 

^ Herrick, " Movi'inents of the Nncleolns through the Action of Gravity," Anatoni. 
Anz., Bd. X. 95. 



206 Page May and Walker 

nuclear membrane. The nuclear membrane is often seen to protrude con- 
siderably at this point (fig. 11), and then the nucleolus passes through into 
the cytoplasm (figs. 12 and 13). Sometimes this protrusion is very large 
compared with the size of the nucleolus, and then it seems to be depressed in 
the middle in a form not unlike the crater of an extinct volcano, the nucleolus 
lying at the bottom of the crater. When seen under the microscope, this 
formation gives the appearance of two protrusions, one on either side of 
the nucleolus. This is due to the fact that the crater formation is seen in 
actual or in optical section. The nuclear membrane is re-formed very quickly, 
and is, as far as can be ascertained, always re-formed long before the 
nucleolus leaves the outer surface of the nuclear membrane. That the 
nucleoli remain contiguous to the nuclear membrane for some time is 
rendered highly probable, if not actually certain, by the fact that in every 
slide examined a very large number were seen in this position. That the 
passage of the nucleolus is brief, is rendered probable by the fact that 
nucleoli in the act of passing through are very rarely to be found in 
comparison with those in any other position either inside or outside the 
nucleus. That the nuclear membrane is re-formed before the nucleolus 
leaves its outer surface is rendered almost certain by the fact that nucleoli 
are comparatively frequently found lying adjacent to or upon it, but no 
breach in the neighbourhood was ever observed. In the latter case, the 
surface of the nucleolus that is touching the nuclear membrane being con- 
cave, a comparatively large area of membrane is covered by it (figs. 14 
and 15). Whether the extruded nucleolus always remains thus attached 
to the membrane, however, appears doubtful, as it has been found in this 
position comparatively seldom. More often it is found adjacent to the 
nucleus, but not compressed upon its membrane (fig. 16). No breach in 
the nuclear membrane has been found that could legitimately be connected 
with the passage of a nucleolus at any other time than during the actual 
process of passing through. 

The nucleoli that migrate appear to be usually, if not always, among 
the largest found in the nuclei of the nerve cells. After they have passed 
into the cytoplasm they increase in size, often to a considerable extent, and 
the contents seem generally to become definitely granular. 

One of the most remarkable facts in connection with this migration of 
the nucleoli is that as they pass into the cytoplasm their staining 
reaction alters. Thus with Method A the nucleoli inside the nucleus 
are blue or violet. Those passing through are purple or red. Those definitely 
outside are bright red or pink, and those which have travelled away from 
the nuclear membrane are pink or red. With Method B the nucleoli inside 
the nucleus are brilliant scarlet. Those passing through are reddish orange, 
and those which have passed through a pale orange or yellow. Nucleoli 
that have been artificially forced out of the nucleus by the knife of the 
microtome stain exactly as do those that are found in the nuclei. This 
suggests strongly that some important chemical or physical change 



Multiplication and Migration of Nucleoli in Nerve Cells of Manunals 207 

takes place in the nucleolus when it passes into the cytoplasm. 
It also seems to ofter a simple and obvious means of judging at once 
whether a nucleolus has been extruded naturally or otherwise. 

The pseudopodial processes observed in the nuclei of the nerve ganglion 
cells in adult and young animals seem frequently to be intimately connected 
with the phenomenon of the migration of the nucleoli. The protrusion of 
the nuclear membrane described above seems to per.sist in many cases long 
after the nucleolus has left the neighbourhood of the nucleus. This is 
particularly noticeable in cases where the nuclear membrane has been 
depressed in the middle, as already described. In optical section this gives 
the appearance of two protrusions with the concavity between them directed 
towards the extruded nucleolus. As to whether these protrusions are 
always the protrusions formed in connection with the passage of the 
nucleolus which have persisted in cases where the nucleolus does not 
remain attached to the nuclear membrane, or whether they represent 
maybe a separate phenomenon, the present observations do not appear to 
give any definite suggestion. In any case, the occurrence of the extruded 
nucleoli and the protrusions in certain definitely relative positions seem 
to be too frequent to be due to a mere coincidence (figs. 17-19). 

Whatever may be the case with regard to the nerve ganglion cells in 
embryos, it has been found absolutely impossible to demonstrate centrosomes 
or astral rays in the material used in the present observations.^ 

After its extrusion from the nucleus, the nucleolus travels towards the 
periphery of the cytoplasm of the cell (figs. 17, 18, and 19). When it 
reaches the periphery of the cell it sometimes passes bodily through the 
surface membrane and is set free among the surrounding cells (fig. 20). 

This is, however, the sequence of events in the case of some onl}' among 
the nucleoli. In other cases the nucleolus ma}^ be seen Ij'ing on the inside 
of the surface membrane of the nerve cell in the immediate neighbourhood 
of a leucocyte or in the case of cerebro-spinal ganglia of the nucleus of a 
capsular cell. Here the substance of the nucleolus seems to pass piece- 
meal through several small openings, and to be absorbed into the cytoplasm 
of the neighbouring cell. The absorbed material seems to lie close to the 
nucleus of the cell that has absorbed it (figs. 21 and 22). 

Those nucleoli that pass out of the nerve cells without being disintegrated, 
seem sometimes to be taken bodily into the cytoplasm of a capsular cell or 
of a leucocyte, where they are probabl}- disintegrated (figs. 23 and 24). 

In studying cerebro-spinal ganglia, while it is quite easj- in very many 
cases to say definitely whether a particular cell is a leucocyte or a capsular 

^ Tlie term " centrosonie " appears lo liave been used somewhat loosely upon some 
occasions. The sense in wliich it is used here is that generally accei)ted by oy tologists, i. e. 
minute structures, oval or bean-shaped, generally two in number, sometimes surroimded 
by an archoplasm or attraction sphere, which is usually contiguous to the nucleus. We 
think it possi1)k^ tliat the extruded nucleolus lying upt)n the nuclear membrane, as shown 
in figs. 14 and 15, may very ]n'obably have been mistaken for the archoplasm, particularly 
as the nucleolar contents are granular in appearance at this stage. 

VOL. I. — APRIL 1908. 14 



208 Page May and Walker 

cell, other cells intermediate in character between these two very different 
types were found during the present investigations. The nucleoli seemed 
to be taken into the cytoplasm of the capsular cells, of the intermediate 
forms, and of the leucocytes indifferently. 

The present observations do not seem in accordance with the interpreta- 
tion of Flemming or of Oscar and Richard Hertwig, as regards the 
function of the nucleolus. These observers held that the nucleoli supply 
nutriment which contributes to the formation of the chromosomes during 
the process of mitosis, and this interpretation is perhaps strengthened by 
the fact that the nucleolus, when present, seems then to disintegrate. 
As, however, the phenomenon of mitosis has not hitherto been observed 
among the nerve cells of the adult mammal, and apparently does not occur, 
the function of the nucleoli here dealt with must be something entirely 
independent of the formation of chromosomes as they appear during 
mitosis. That the nucleoli are constantly being increased in number, even 
in adult life, by the process described in this paper, proves that they 
subserve some important function, but whether this is in the nature of an 
excretion, secretion or some other form of cell phenomenon, there does not 
as yet appear to be any direct evidence. 



CONX'LUSIOXS. 

The conclusions arrived at from the observations here described are — 

1. That the nucleoli of the nerve cells described multiply continually, 
generally by a process of budding, more rarely by an equal division in bulk 
of a pre-existing nucleolus. 

2. That the nucleoli pass out of the nucleus, and in the process their 
-staining reaction changes. 

3. After passing out of the nucleus the nucleoli become granular in 
appearance, and increase in size. The increase in size would appear to be 
due to the lessened density of the contents, rather than to an increase in 
substance. 

4. The nucleoli sometimes pass bodily out of the nerve cell (peri-karyon), 
and are taken into the cytoplasm of leucocytes or capsular cells. At other 
times the substance of the nucleoli seems to pass in small portions from the 
cytoplasm of the nerve cell into that of a capsular cell or a leucocyte. 



Qiirn-tt'dij -hin-wd fif Erpcrln^cntfil Phifsiolnrjii, Vul. I.. April ignS.] [Plate I. 



'- ^ 


\ 




Foq / 


n^.3. 


Fccj.S. 


>:^^ 






0^ 






rcy.^. 


Fo^ S. 


Fcg. 6. 


^4 


rlPp^-' 


• 






% 


r.y.^. 


r.y.p 


^^yy 




e ». 

W 


w 


C.EWalkci-d.L 


''f " 


Fc^ /'Z 



W. Pack .Mav ami C. K. Walkki:, 



qi,avin-hj .Toimidl nf K.,',,erhnri,lid Physlnh,,,,!, V„l. /., April 1'.>0S.\ 



[Platk II. 





^<'^'\ 




1 ■^^'^ 




Fo^.f:^ 


/iy /^ 


fCg./S. 




• ' 


/Wi 


'-\^)' ' 




'^"■® 


Fi^Jc. 


F^.IJ. 


^^y. /^'^ 




.■ ■ ■ .' ■-, ^ 


m . -.. 


o 


i? ^ 


. '^^^iS' ■' 


& 






F,yf 


/iy^<? 








m 




\^^ ig' 




%? 


#/''■ 




,,y^3. 


Foj.?^- 



W. Pa(;k .May ainl C. K. Walki 



Multiplication and Migration of Nucleoli in Nerve Cells of Mammals 209 



DESCRIPTION OF PLATES. 

Fig. 1. Cell from Gasserian ganglion of a rabbit, .showing an early .stage in the 
Ijudding of the nucleolus, and some chromatin aggregates upon the membrane of 
the nucleolus. 

Fig. 2. Ditto. 

Fig. 3, A cell from the same ganglion, in which a bud has travelled some 
distance from the nucleolus, but is still attached to it. 

Fig. 4. Ditto. 

Fig. 5. Ditto. 

Fig. 6. A cell from the same ganglion, in which the attacliraent of the bud to 
the nucleolus is broken. 

Fig. 7. A cell showing the same phenomenon, with several buds separated oft' 
from the nucleolus. 

Fig. 8. A cell from the same ganglion, showing several nucleoli. 

Fig. 9. A rare mode of nucleolar division. 

Fig. 10. A cell from the same ganglion, showing a nucleolus that has been 
pushed out of the nucleus by the action of the microtome knife. There is an empty 
space in the nucleus, and the nuclear membrane shows no signs of being regenerated. 
Here the staining reaction of the nucleolus is the same as in the case of nucleoli 
contained in the nucleus. 

Fig. 11. A cell from a spinal ganglion of a cat, showing the nucleolus pressing 
out the nuclear membrane. 

Fig. 12. A cell from the same ganglion, showing the nucleolus passing through 
the nuclear membrane. Here the staining reaction of the nucleolus is changing. 

Fig. 13. A cell from the same ganglion, wdiich shows the appearance, in optical 
section, of the crater formation of the protrusions of the nuclear membrane described 
in the text. The nuclear membrane has re-formed behind the nucleolus. 

Fig. 14. A cell from the same ganglion, in which the extruded nucleolus is 
adherent to the nuclear membrane. The nucleolus has increased considerably in 
size, and has become definitely granular. Here its staining reaction is quite diff'erent 
from what it is in the case of the nucleoli still contained in the nucleus. 

Fig. 15. Ditto. 

Fig. 16. An extruded nucleolus still adjacent to the nuclear membrane 
(spinal ganglion of cat). 

Figs. 17, 18, and 19. Nucleoli passing towards the periphery of the cell after 
being extruded from the nucleus (spinal ganglion of cat). The pseudopodial 
processes of the nucleus are interpreted as being the remnants of the protrusions 
produced by the passage of the nucleolus. 

Fig. 20. A nucleolus passing bodily out of a nerve cell (spinal ganglion of cat). 

Figs. 21 and 22. The contents of nucleoli being taken from the nerve cells into 
the cytoplasm of adjacent cells (spinal ganglion of cat). 

Fig. 23. A nucleolus derived from a nerve cell in the cylojilasm of a capsular cell. 

Fig. 24. A nucleolus derived from a nerve cell surrounded by leucocytes. 



I 



THE ELECTRICAL RESPONSE OF MUSCLE TO VOLUNTARY, 
REFLEX, AND ARTIFICIAL STIMULATION. By Florence 
Buchanan. (From the University Museum, Oxford.) 

{Received for publication May 29, 1908.) 



CONTENTS. 

I. Introduction : The question at issue ..... 

II. The Two Kinds of Rhythm exhibited in Capillary Electrometer 

Records of the Reflex Response of Frog's Muscle in 

Strychnine Spasm : 
Experiments in which the temperature of the recording muscle alone 

was varied .....-• 

Experiments in which the temperature of the spinal cord alone was 

varied ......•• 

Conclusions to be drawn from the.se experiments as to the nature 

of the stimulus immediately provoking the response 

III. The Rhythm exhibited in Capillary Electrometer Records of the 

Response of Muscle to Voluntary Effort in Man : 
Methods of obtaining, and of testing the significance of, the effect 
Attempts to alter the effect by varying . . . • • 

(a) The contacts ......•• 

(b) The recording muscle ...... 

(c) The voluntary effort .....•• 

IV. Contrast between Capillary Electrometer Records of the Volun- 

tary Response in Man and that of the same Muscle to 
Artificial Excitation of its Nerve by a Series of In- 
duction Shocks of a Frequency of about 50 per second . 

V. Conclusions to be drawn from the foregoing Experiments and 

further Experiments which favour these Conclusions : 
The special advantages possessed l)y a quick capillary electrometer for 
the investigation of relative differences of potential of small but 
unknown amount, recurring with an unknown fre(i^uency 
Evidence that the rhythm observed in voluntarily contracting muscle 
has its origin in the muscle ..... 

Futility of the objections to such a view .... 

The fact that the frequency is not altered by the strength of the 
stimulus . . . . . . • • 

Differences of rhythm obtaining in the responses of different 
muscles in man, have their counterpart in those obtaining in 
different muscles of the frog, subjected to cerUun kinds of 
artificial, as well as to reflex, stimulation. Suggestion as to 
the significance of these dirterences 

VI. The Reflex Electrical Response in Max 
VII. Summary .....-•• 

VIII. Note on the Effect on the Electrocardiogram of Man of the 
Contraction of Voluntary Muscles . 
IX. Note on the Misuse of the Word "Rhythm" . 

AddenduxM ......•• 

References ..... • • 

VOL. L, NO. 3. — 1908. 15 



PAGE 

212 



214 
216 
220 



221 
222 
223 
226 
226 



228 



231 



234 

234 



23a 



237 
238 
240 

240 
241 
241 
242 



212 Buchanan 

Thixking that the action currents of human muscle in voluntarj^ contrac- 
tion might throw light upon the nature of the normal stimulus to skeletal 
muscle, I had several times made the attempt to record such action currents 
with the capillary electrometer, but had not succeeded in doing so until 
August of last year. As the records did not seem to me to giv^e the 
required information, I was meaning to content myself with merely 
drawing attention to the fact that they can be obtained, when there 
appeared a paper by Dr Piper of Kiel (1) in which were reproduced a 
number of records of human muscle in voluntary contraction, taken with 
the string-galvanometer, which seemed to him to give the solution to the 
problem. 

Electrophj^siology is so much beset with the difficulty of culling, from 
the instrumental effects observed, the underlying physiological phenomena, 
and of excluding everj^thing for which the recording instrument alone may 
be responsible, that the observation of the same phenomenon with different 
recording instruinents is always likely to be of value. It is the more so 
when, as apparently' in this instance, the records do not show the same 
thing, and when the conclusions drawn from them by two observers are 
diametrically opposed. 

Dr Piper's kindness in informing me of results contained in his two 
later papers (2), (3), before their publication, and his readiness to let me 
examine several of his still unpublished records there referred to, make 
me feel sure that he is as anxious as I am myself to come to a true under- 
standing of the physiological significance of what we are both recording, 
and that he will prefer a criticism of his views to undisputed acquiescence 
in them. I trust, therefore, that he will not take amiss any criticism which 
appears in the following pages. 

He has recorded, with different patterns of Einthoven's string-galvano- 
meter, the electrical responses of several different muscles in man to their 
normal stimulus. He finds in each record one particularly prominent 
rhythm (see note on p. 241) which he regards as constant for each muscle 
or group of muscles. This rhythm, he thinks, indicates the rate at which 
successive stimuli are sent to the muscle from the central nervous system. 
He concludes, for instance, that the flexor muscles in the lower arm, which 
are the muscles which both he and I have investigated most, are normally 
supplied by impulses arriving with a frequency of 47 to 50 per second (1), 
(2). He bases his conclusion — to a large extent — on the fact that the 
galvanometer record obtained from the same muscle, when the median nerve 
is artificially stimulated by induction shocks of about this frequency, has a 
similar character. As will presently appear (pp. 228 and 241), the records 
given by my electrometer of the responses of the same muscle under the same 
two conditions exhibit very little resemblance to one another ; but even were 
it much stronger than it is, and in the records of both instruments, the fact 
that an effect can be imitated in one pai-ticular way affords, of course, 
no evidence whatever that there is no other wav in whicli the effect may 



The Electrical Response of Muscle 213 

have been produced. It has, as a matter of fact, been shown both by 
Garten (6) and by myself (4) that a rhythm of a frequency of from 50 
to 100 per second may be observed in the electiical response of excised 
frog's muscle when either it or its motor nerve is subjected to a continuous 
stimulus (whether actually so, or consisting of instantaneous stimuli recurring 
in such rapid succession that the whole may be so regarded) or to a variety 
of other stimuli of short duration which have notliing discontinuous in 
their nature. The frequency of this rhythm varies (as we have each of us 
shown) with the condition of the muscle much more than with the 
particular nature of the exciting stimulus. The central stimulus which 
provokes a normal contraction in any animal might just as well partake 
of the nature of any of these non-discontinuous stimuli as of that of the 
particular discontinuous one which Piper found to produce an effect on 
his galvanometer of a similar character. I had, indeed, laid stress on the 
fact [(4), p. 149] that the central stimulus intervening in the reflex 
response of the muscle of a frog in certain stages of strychnine poisoning is 
not to be regarded as of the nature of a series of in.stantaneous stimuli, 
although its effect on the muscle can be imitated (as in photo. 35, on pi. vii.) 
by that of a series of such stimuli, recurring with a frequency of, say, 50 
per second, interrupted at intervals. The view then expressed was after- 
wards put to the test and conlirmed. The experiments which confirm it 
deserve more than the passing mention they received in the following year. 
(5), and since the results obtained have so distinct a bearing on tlie subject 
to be chiefly discussed in the present paper, namely, on the significance to 
be attached to the rhythm observed in the electrical response of a muscle in 
voluntary contraction, I propose to begin by giving some account of them 
and to reproduce a few typical records. 

II. The Two Kinds of Rhyth.m exhibited in the Reflex 
Hlei TKicAL Responses of Fro(;'s Muscle in Strychnine Spasm. 

For the .sake of clearness I shall henceforward designate J:hose undula- 
tions which occur in the records with a frequency of from 40 to 100 per 
second as wavelets, the curves recurring with a fretpency of from 8 to 14 
per second (on which the wavelets may be superimposed) as waves. 

The independence of waves and wavelets is shown by the fact that 
either may be present without the other. I ha\e already reproduced 
records of frog's muscle in strychnine spasm showing (a) wavelets exclu- 
sively or almo.st exclusively [(4), pi. vii., ph. 37: })!. ix.. ph. 45, 46, 47); 
(b) waves exclusively [(4), pi. viii.. ph. 31), 40, 41 : })1. ix.. ph. 50] : and (c) 
the two side by side [(4), pi. vii.. ph. 3S : pi. ix.. ph. 4S, 4i>]. The records I 
have now to reproduce are to bear witness to what was stated in 1902, namely, 
that what I now call wavelets depend for their frecjuency upon something 
in the muscle, that the waves depend for theirs upon something in the 
spinal cord. 



214 Buchanan 

Experiments in wliicli the temperature of the recording muscle 
was varied, while that of the rest of the frog remained 
constant. 

For these experiments the gastrocnemius was chosen as recording 
muscle, and was prepared so that it remained in connection with the rest 
of the preparation by its sciatic nerve only. The nerve ran through an 
aperture in the wall dividing two moist chambers, in the one of which was 
the muscle, in the other the decerebrate frog. The temperature of the 
muscle chamber was varied by placing over it now a tray of iced water, 
now one of warm water. 

In the experiment to which tig. 1 relates the iliac artery had been 
ligatured on the left side and 5 minims of 1 per cent, curare injected into 
the dorsal lymph sac : then, when the curare had taken effect, 1 minim 
()"1 per cent, strychnine acetate had been injected and half an hour later 
the left gastrocnemius prepared and arranged in the way just described. 
Its tendon end was connected in the usual way with the mercury of the 
electrometer, a spot on its dorsal surface with the acid. The muscle was 
excited each time by a single break induction shock applied to the skin of 
the back of the frog. The temperature of the moist chamber in which the 
body of the frog was lying remained at 12° C. throughout the experiment. 
Twelve records were tirst taken with the muscle also at 12° C. None of 
these exhibit more than a single wave, on the rise of which are wavelets 
which in nine of the records have a frequenc}' of 100 ± 3 per second, in 
two of them (the fifth and the sixth) one of about 90 per second, and in 
one (the first) one of 110 per second. The twelfth record is reproduced in 
fig. 1, A. 

The muscle chamber was then cooled, and when its temperature was 
10" C. the record reproduced in fig. 1, B, was taken. The f requeue}" of the 
wavelets is 81 per second. A second record, taken after the muscle had 
been cooled to 9" C, showed a wavelet frequenc}' of 70 per second. 

The preparation was now left for an hour and a half, the muscle 
chamber remaining covered b}' the ice-tray, and its temperature being 7' C. 
at the end of tlie time. The only muscle which contracted when the skin 
was stimulated was, as before, tlie recording gastrocnemius, which had been 
protected from the influence of the curare. The contraction was now no 
longer twitch-like as it had been in the morning, but ^as, each time, a 
long and steady spasm. All the electrical responses were now serial, waves 
as well as wavelets being seen in the records when the plate was travelling 
at a sufficiently slow rate to show them. Five records of the reflex re- 
sponse were taken with the muscle at 7° C, of which the last is reproduced 
in fig. 1, C. The muscle chamber was then allowed to return to the tem- 
perature of the room and the record reproduced in fig. 1, D, was taken. 
It was then wanned beyond the temperature of the room and five records 
were taken with the temperature of the muscle rising to 16 C. The last. 



Tlif Electrical Response of Muscle 



215 



four records were taken while the muscle was being cooled again to 9° C. 
The table on the following page gives the period of the first wave in each 
record, and of a second when a second complete one was on the plate, in 
thousandths of a second (cr). It also gives the frequency of the wavelets 
on the first wave in as many of the photographs as they were distinct 
enough to be counted. 




oi sec. 




i 



Fig. 1. — KeHex electrical ii's(ioiisi'.s ol' tiog's gastroc ii eiui us (strvcluiiiie). Si)iual cord at 
1-i' C. in all. The muscle at I'i' C. in A. at 10' C. in B, at 7° C. in C, at 12' C. in D. 
Recording surface describing an arc. R;ite of movement indicated by 500 fork tracing 
above. It was considerably slower in C and D than it was in A and B. s is tlie signal 
key, the break of which prouncfd the cxcitntion. 

The results obtained from each experiment of this kind were so uni- 
form that only six of them, not all on curarised frogs, were made. In all 
the wavelet freciuency changed when the temperature of the muscle was 
changed, while that of the waves remained unaltered, i.e. the wave-length 
varied no more than it usually does when a succession of responses are 
recorded all with the muscle (as well as the cord) at the same temperature, 
and it did not vary in any definite direction with the temperature of the 
muscle. In only one, as it happens, out of this set of experiments were 



216 



Buchanan 



there no wa\-e.s in the records, these being of the type whicli in my previous 
paper [(4), p. 146] I called " type i," and presenting a long series of wavelets 
only (lasting for some 0-3 second) the frequency of which again varied 
directly with the temperature of the muscle. This series of experiments 
shows, therefore, that the frequency of the wavelet rhythm is dependent 

(Jastrocnemius of strychnine frog, D 82. Apr. 7, 1902. (Stimulation by 
break induction shock to skin of the back, throughout.) Temperature 
of spinal cord constant (12" C). 



Temperature 
of muscle. 


Total reflex-time. 


Duration of waves. 


Frequency of wavelets on 
1st wave. 


7°C. 


52^ 


1. 
172rr 


2 

198,7 


54 per sec. 


7°C. 


r^-2<T 


184- 


210,7 


41 


7°C. 


5-2a 


183<r 




40 


7°C. 


5bT 


ISOrr 




? 


7°C. 


52t 


180,7 


195,7 


41 


i2°a 


5U 


174,7 




90 „ 


13° C. 


59rr 


110 2ud 


wave 


96 


14° C. 


.-)0<r 


188,7 




100 


15° C. 


47^ 


183,r 


209,7 


96 „ 


16° C. 


46<r 


184,7 


217^ 


106 


16° C. 


47n- 


184<r 




110 


11° C. 


47tr 


180,7 




? 


10° C. 


M)a- 


175,r 


210,7 


70 „ 


9°C. 


50,7 


174,7 


186,7 


9 


8i°C. 


47<T 


187'^ 




? 



on the muscle, but does not, of course, show that it is independent altogether 
of the cord. This however is, I think, sliown by the second series of 
experiments. 

Experiments in which the temperature of the spinal cord was 
varied, and that of the recording muscle kept constant. 

A few of these were made in 1902, and se\eral more have been made 
within the last year. The temperature of the cord was altered hy running 
water at different temperatures through a glass tube passing under or over 
the back of the frog, and so shaped that it went no nearer to the recording 
muscle. The muscles used to record in these experiments were : the 
gastrocnemius, the sartorius, the biceps and the triceps femoris, and the 
semitendinosus. I shall have to refer later (p. 237) to differences of wavelet 
rhythm characterising these different muscles. Here I will only give the 
details of one typical experiment and reproduce two typical records. In 
the experiment to which fig. 2 refers the triceps was the recording muscle, 
the distal leading-otf electrode was on the tendon end and connected with 
the mercury of the electrometer, the proximal one was about one centi- 
metre away from it, and so on that part of the muscle in which there is no 



The Electrical Response of Muscle 217 

external evidence of its threefold origin. The temperature of the moist 
chamber in which the whole preparation lay remained throughout the 
experiment at 12° C. Through the glass tube lying over the back of the 
frog iced- water at 4° C. had been running for nearly half an hour before 
the experiment was begun. After five responses had been recorded, water 
at 22° C. was passed through the tube for two minutes, and three more 
responses were recorded. Then water at 5°C. was run through for three 
minutes and three more responses were recorded. Finallj', water at 24' C. 
was once more run tlirough for three minutes and three more records wei-e 
taken. The responses in each case were to the excitation of three different 
afferent nerves, to the brachial and sciatic of the opposite side, and to the 



0-1 sec. 



I > <> W I W |I W ^^^^ ^ ^^^ 




Fk;. 2. — ReHex electrical responses of frog's triceps f em or is (strychnine). Muscle at l-l' C. 
in both. Water running through glass tube over spinal cord at 22° C. in A, at 5" C. in B. 
Recording surface moving horizontally, rate indicated by 100 fork tracing above the signal s. 

sciatic of the same side. Enlargements of the first part of two photogra])hs, 
taken when the opposite sciatic was stimulated by a break induction shock, 
are reproduced in fig. 2. When A was taken the temperature of the cord 
was about 22° C, and it liad been subsequently cooled to about 5° C. when 
B was taken. The table on p. 218 shows the frequency of the wavelets 
on the first wave in all the responses in which it could be determined, and 
the time duration of each of the first four waves. 

In the records of this as in those of most experiments in which the 
electrical responses of frog's muscle in strychnine spasm were recorded, 
whether the cord were cooled or not, there is a good deal of variation in 
the frequency of the wavelets, not only in successive resix)nses, but also at 
difierent times in one and the same response. When the record is .serial 
the wavelet frecjuency shows that the rh3'thm usually remains the same for 



218 



Buchanan 



the time indicated by the ascent of any one wave, so that in cord-cooling 
experiments one can compare the frequency in corresponding waves in the 
successive responses as in the experiment to which the table refers. When 
there is little or no indication of waves in a record, as in the one repro- 
duced in fig. 3, one frequency may give place to another abruptly, and in 
cord-cooling experiments it is more difficult to select the corresponding 
parts. But in the records obtained in several of such experiments made on 
frogs in which the action of strychnine was not at its height and the 
electrical reflex response non-serial, I have not been able to discover any 
definite relation between the frequency of the rhythm and the temperature 

L. triceps of strychnine frog, F 160. Feb. 24, 1908. (Stimulation by 
break induction shock to the L. sciatic, to the R. sciatic, or to the R. 
brachial nerve.) Temperature of muscle constant (12° C). 



Temperature 
















of water 


Nerve 


Total r6tlcv- 










Fre(iuency of 


passing over 


stirvrn- 


time. 




Duration 


of waves. 




wavelets on 1st 


the back of 


lated 












wave. 


the frog. 






















1 


2. 


3. 


4. 




4''C. 


E. br. 


100<r 


180(7 


leOcr 


175cr 


198cr 


88 per sec. 




R. sci. 


87<r 


190<r 


160<r 


ISOo- 


215^ 


90 ' „ 




L. sci. 


50<r 


21.50- 


1650- 


18.50- 


230,r 


100 „ 




R. sci. 


87(r 


183<r 


172<T 


183^^ 


210(r 


92 




R. br. 


95<r 


180<r 


155o' 


170.T 


230(r 


97 


22° C. 


R. br. 


72<r 


115(r 


lOOo- 


1150- 


120ff 


94 ,, 




R. sci. 


62<r 


115<r 


1050- 


1-20<T 


\30<T 


100 „ 




L. sci. 


40<r 


125^ 


lOOcr 


UOcr 


llaer 


90 „ 


5°C. 


R. br. 


lOOtr 


ISOt 


130^ 


155o- 


170a- 


95 




R. sci. 


750- 


1650- 


148(r 


170^ 


no 5th wave 


100 




L. sci. 


48-^ 


1820- 


135,T 


175^ 


1720- 


100 „ 


24° C. 


R. br. 


87.T 


115<r 


I05o- 


120<r 


IOOt 


100? „ 




R. sci. 


60rr 


llTtr 


100.T 


113(r 


165^ 


94 „ 




L. sci. 


40T 


124<r 


106.r 


115a 


no 5th wa\ fc 


100 „ 



of the cord, when this was alternately raised and lowered. The wavelets 
frequently are more marked when the cord is cool, i.e. the ascent and 
descent of each wavelet is steeper (as is often the case also in the serial re- 
sponses), but there is no constant lowering of fre<iuency in any part of the 
record with cold. 

The waves, on the other hand, in all experiments I have made in which 
the responses have been serial, vary in frequency in a very definite way 
with the temperature of the cord. Thus in the experiment already cjuoted 
we find that in the first five records, taken when the cord was at a tem- 
perature of about 4° C, the frequency of the waves (estimated from the 
four which alone appeared on each plate) was from 5"6 to 5 per second. In 
the next three, in which the cord was nearly at 22° C, the frequency (again 
estimating from the first four waves onl}^ although there were more on the 



The Electrical Response of Muscle 



219 




220 Buchanan 

plate) was from 8-8 to 80 per second. In the next three, with the cord 
cooled again, it was from (iG to 6'0 per second ; and in the last three, with 
the cord warmed again, it was either 90 or S'O per second, and would have 
been 9"0 in all three if the estimation had been made from the first three 
waves only in the second response. 

It would appear, therefore, from these two sets of experiments, that in 
the electrical responses of frog's muscle to a stinuilus which comes, in the 
last instance, from the central nervous system, two kinds of rhythm may 
manifest themselves in capillary electrometer records. The one of them, 
that of the wavelets, which is very rarely absent in the responses of muscle 
in strychnine spasm, gives us no positive information as to the nature of 
the stimulus which immediately provokes the response ; but, from the con- 
ditions under which the frequency may be modified and from those which 
fail to modify it, we learn that the existence of a rhythm of this kind in 
the response is no indication of the existence of a similar rhythm in the 
stimulus, and since a single wavelet has a duration similar to that of a 
response to an instantaneous stimulus, we may infer that the central 
stimulus is not of the nature of a series of such stimuli. 

The other kind of rhythm, that of the waves, which is much more 
frequently absent from the records, and the absence of which seems to 
be determined b}^ the extent to which the animal is affected by the 
drug at the time the records are taken (I say advisedly " the animal " 
and not " the central nervous system "), does apparently, when present, 
tell us the rate at which stimuli were coming from the cord at the time. 
The form of each wave also tells us that the effect of each stimulus 
was reduced, sometimes even to zero, before the effect of the next began, 
and suggests (but this is by analogy only at present) that each stimulus 
had a duration corresponding only to that part of the wave in which 
wavelets are present. Moreover, the different forms of wave which 
occur in different experiments suggest that this stimulus (" Zeitreiz " of v. 
Kries) sometimes rises to its maximum strength quickly and sometimes 
slowly. My records, in which waves are altogether absent, suggest, as did 
V. Kries 's observations (7) and (8), that a single central stimulus may have 
quite a long duration — not infrequently one of a third of a second — some- 
times even lasting for a whole second [see (4), pi. ix., ph. 46]. It is, 
therefore, by a study of what occurs in this 3 to 14 per second rhythm, 
of what occasions its absence or presence, and of the conditions which 
modify it, that we can alone hope, it seems to me, to come to any satis- 
factory conclusions as to the natui-e of the normal stimulus, and to ascertain 
whether or not it is rh^^thmical. For this purpose we must use animals 
in which it is possible to modif}^ conditions for any one part independently 
of other parts ; and, since temperature is the easiest condition to alter, it 
ma}' be that polikothermal animals will give us most information. But the 
experiments will have to be repeated with warm-blooded animals before we 
are prepared, when we meet with a rhythm in the effect of so complicated a 



The Electrical Response of Muscle 221 

thing as the electrical response which accompanies the normal contraction 
of skin-covered muscles in man, to say an}i;hing as to what it indicates. 
Only if such experiuients gave positive results — e.g. if it could be demon- 
strated that the rhj^thni obtaining in the electrical reflex response of the 
gastrocnemius of the rabbit, which, according to Piper [(1), p. 382], has a 
frequency of 50 per second like that of the lower-arm flexors of man, can 
l)e deflnitely reduced in frequency by cooling any part of the central 
nervous s^'stem while the temperature of the muscle remained constant — only 
then should I feel inclined to admit Dr Piper's conclusion that the rhythm 
lie has observed in his records is also that of the normal stimulus. Until 
1 have any evidence of this sort, I cannot help thinking that the rhythmical 
effect which both he and I have recorded in human muscles contracting to 
the normal stimulus, partakes of the nature of the wavelets seen in m}- frog 
records rather than of that of the waves. But before giving my reasons 
for so thinking, I must give some account of my own experiments on such 
muscles, in which the electrical response was recorded with the same 
instrument or with one having the same properties as the one used for the 
frog experiments I have just described. 

III. The Rhythm exhibited ix Capillary Electkometek Records 

OF THE ReSPOXSE OF MuSCLE TO VoLUXTARY EFFORT IX MaX. 

I have taken records of the response of the lower-arm flexors with 
about twenty diflerent people, with some of them on more than one occasion. 
I am indebted to about a dozen Oxford undergraduates and to a few other 
friends (four of them women, and two of them children) for their kindness 
in acting as subjects for these experiments. 

Methods. — As leading-ofl' electrodes I have, as a rule, used sponges 
soaked in .salt .solution, and tied up with muslin to prevent possible altera- 
tion of area of contact when the muscle contracted. They were applied to 
the skin, the one over the particular part of muscle required to give its 
record, the other to that over some other part of muscle or other tissue. 
Outside each sponge was a zinc rod with binding screw. I have also used, 
in certain experiments, non-polarisable electrodes similar to those used by 
Piper, or others made by interpolating a zinc-sulphate-containing sponge 
between the sponge which touched the skin and the zinc. There is not 
the same necessity for using non-polari.sable electrodes with the capillary 
electrometer as there is with the galvanometer, and therefore, when I found 
that it made no diflerence to the result (the record) which of these three 
kinds of electrode I emploj-ed, when trying them all in turn on the same 
person, I felt justified in continuing to use the simplest and most con- 
venient kind which I had used in the flrst instance. The electrodes were 
each held in position by elastic bands, the positions being, with each person 
to begin with, the same as tho.se cho.sen by Piper. The arm rested 
supine on an insulating support, and in the hand was a dynamometer in 



222 Buchanan 

the form of a compressible steel ellipse with scale. That electrode which 
was over the part of the muscle felt to become most tense when the 
dynamometer was squeezed was, to begin with, always connected with the 
acid of the electrometer. The image of the meniscus was projected through 
a slit on to a photographic plate moving at a known and equable rate 
inside a dark box. There being no eye-piece to the microscope, the image 
was inverted. The distance of the plate from the image was such as to 
magnify it with the objective used about three hundred times in all cases. 
The late of movement of the plate was shown by the vibrations of a spring 
in unison with, and driven by, a 100 tuning fork. The trolley carrying 
the photographic plate broke a key when it began to pass the slit which 
left a record on the plate of the moment at which it was broken. The 
subject was, as a rule, told to contract the muscle that was being investigated 
— to clench his tist or his jaws, for instance — only when he heard this key 
break, so that while he was reacting to sound the meniscus might inscribe 
its resting position on the plate. 

The effect observed. Evidence tluit it is not to be attributed 
to inequalities of pressure of the skin on the electrodes. — Fig. 4 
is typical of the effect produced upon the electrometer by the flexores 
digitorum. After recording it for the first time, I made the following 
experiment for the purpose of ascertaining whether or not what I had re- 
corded could be due to variations in the pressure of the skin against the 
sponge, for such variations there must undoubtedly be when the muscle 
suddenly becomes tense. A piece of dry bandage was tied loosely round 
the arm of the subject, electrode and all, at the place where the tension of 
the muscle becomes greatest when the fist is clenched. A second person, 
holding the ends of the bandage, pulled them as hard as he could, so as to 
tightly press the zinc rod and intervening sponge against the arm of the 
subject while the plate was passing the slit, the subject himself being passive. 
The plate was passed through under such conditions alternately with other 
plates which were passed through while the subject was actively- clenching 
his fist, beginning to do so at the sound of the signal. While all 
photographs taken with the muscle contracting at the given signal showed 
an effect of the kind reproduced in fig. 4, all those taken with the muscle 
passive, but the pressure on the skin increased by the second person, 
showed a meniscus tracing as smooth throughout as that seen during the 
reaction time in the alternate photographs. This experiment makes the 
presumption strong that the effect observed when the muscle contracts 
actively is really a muscular effect. 

What strikes us most in all the records is the great irregularity of the 
undulations which indicate electrical variations, irregularity both in duration 
and in the steepness which in capillary electrometer records indicates 
amount, or strength, of variation. The irregularity is of the same kind 
and the variations have the same sort of frequency as have the wavelets 
in records obtained with many frog's muscles in strychnine spasm 



The Elect i-ical Response of Muscle 223 

(see tiu-. 8, p. 219) or in frog's muscle made to give a persistent contraction 
by the application of a descending constant current to the motor nerve, by- 
breaking an ascending constant current running through the nerve, or by the 
stimulation of either the nerve or the muscle itself by a series of instan- 
taneous stimuli following one another in such quick succession that they 
are unable each individually to produce a distinct effect. So great is the 
irregularity of the excursions in the records obtained with these arm 
muscles in most people, that it is very difficult to say what their frequency 
is. One frequency ma}- prevail for a few hundredths of a second, then 
another, then the first back again, or yet another. In most of the records 
a frequency of 100 to 120 per second manifests itself somewhere, also one of 
about 60 to 80 per second. It was only very rarely that one as low as 50 per 
second also appeared, only, in fact, in four of all the people with whom I 
have so far taken records ; in one of whom, however, it even at times became 
as low as 40 per second (see table on p. 227). The effect, although it varies 
greatly in different people, does not, so far as my present experience goes, 
seem to be altered in any definite way either hy sex or hy age. The jags 
on the curve are often — and in the records taken with some people more 
than in those taken with others — thrown into groups which recur with a 
frequency of from 14 to 30 per second, but the curve outline is so 
irregular that I can lay but little stress upon the presence of these groupings. 
It may, however, be worth mentioning that groupings of the same sort of 
frequenc}^ occur in records taken with the masseters, where, as we shall 
presently see, the excursions which fall into groups have a much greater 
and more constant frequency than thej' have in the arm muscles. 

In the hope of getting more information as to what it was tluit my 
records represented, I have, with one or more of the subjects, attempted to 
alter the effect by \arying certain conditions. 

(a) The contacts. — The connections with the electrometer were 
usually such that the part of the flexors which became most tense when 
the fist was clenched was connected with the acid. The first excursion 
which the meniscus made from its resting position was then alwaj's 
adostial (upwards in the records), not only when the second electrode 
was distal to it, as it usuall}- was, but also when it was on the elbow or on 
the shoulder. Thus the spot chosen as being mechanicall}- the most active 
one always becomes galvanometrically negative before any other spot, 
whether or not these also become negative afterwards. When the connec- 
tions with the electrometer were reversed the first movement was alwa3\s 
abostial. This is, of course, only what was to be expected on the assump- 
tion that what we are recording is essential!}- a muscular phenomenon, and 
the fact therefore helps to support the assumption. The records give 
however, as a rule, little, if any, evidence of the way in which the con- 
nections had been made except at the start, so that unless this appears on 
the plate I doubt whetiier anyone could tell with confidence from the 
records which way the connections had been made. 



224 Buchanan 

This means that it is difficult to find any place on the arm which, 
to judge from the records, remains wholly electrically inactive when 
the fist is clenched. I tried to do this by varying the position of 
the second leading-off electrode, leaving the one applied to the 
most active part of the flexors and connected with the acid of the 
electrometer, in the same position. Records were always first taken 
with the second electrode on the skin over the tendons of the flexors 
near the wrist (position i.). It was then sometimes moved so as to 
be also on the contracting muscle at a place about 10 cm. distal to 
the other electrode, i.e. presumably on a spot of nearly the same 
activity (position ii.); or it was placed (iii.) either on the back of the 
elbow or on the shoulder, i.e. on a spot likely to remain inactive when 
the flexors contracted. So long as the two electrodes made contact 
with equal areas of the skin, the records taken with the second one 
in positions i. and iii. gave as little evidence that the effect represented 
related to what was happening under the first electrode only, as did those 
taken with it in position ii. The amplitudes of the excursions were, how- 
ever, less when the second electrode was in position ii. than when it was in 
position i. It was in position i., in the records reproduced in figs 4, 5, 
and 6. There is, so far as I can see, no way of assuring oneself of the 
neutrality of any particular spot in the arm of a living person while it is 
performing any voluntary action. It would be easier to assure oneself of 
the neutrality of some quite other part of the body so far as mechanical 
action is concerned, but no other part of the body save the limb to 
which the fixed electrode is attached will serve the purpose when it is 
the electrical action which is concerned, because the changes due to the 
heart's electrical action are apt to appear whenever the two electrodes 
are far apart on the body and add further complication to the record 
(see note on p. 241). 

By reducing the size of the area of contact made with the 
second electrode I have, how^ever, obtained records which, to anyone 
accustomed to the interpretation of them, would suggest that the spot under 
one electrode was recording alone, or to a much larger extent than the 
other. Fig. 5 shows the records of the voluntary response given by 
the flexors when, on tlie one hand (A), tlie two electrodes (in position i.) 
were each 3 to 4 cm. in diameter ; and when, on the other (B), the one of them 
(that connected with the mercury) was reduced so that it made contact 
only with an area of skin Ii cm. in diameter. The rounded summits of 
the individual excursions in B contrast strongly with the pointed summits 
in A. The capillary electrometer shows the same sort of difference between 
the action currents (Einzelschwankung) of a frog's sartorius when, on the 
one hand, the tendon end (connected with the mercury) is devitalised, and 
when, on the other, it is sound. 

All the records furnished by the flexors of the same individual as those 
reproduced in fig. 5 were far more regular than were those supplied by any 



o 
'^ if) 



li 



J. 



Tlie Electrical Response of :Muscle 

< CO 



r 



_S o 



3 2 

— o 



226 ' Buchanan 

other individual I luive so far tried, and this was true on each of three 
separate occasions. Siuiihir records have since been obtained with one other 
person (see Addendum). All the experiments in which the position of the 
leading-ott" electrodes was altered were made upon him also. When most 
regular the frequency was slower than it is in most people, and each 
individual etiect was stronger. The strength of the mechanical effect 
was not in any way remarkable (see p. 227). 

When both electrodes were small the ascents as well as the descents of 
the separate excursions were less steep, i.e. the effects, as Piper has shown, 
are weaker. I attribute the difficulty I had experienced until lately in 
recording the electrical response of voluntary muscle in man to the fact 
that I had previousl}' tried to obtain it by leading off from surfaces too 
small for such negativity as might prevail under them to be able to affect 
the electrometer. 

(b) The recording muscle. — Besides the muscles in the lower arm, I 
have used those in the hand and those in the jaw. In the hand muscles 
a response frequency of 100 to 140 per second appears more often than 
any slower one. Records taken with the masseters show a series of 
very fine teeth the frequency of which is more uniform than it usually 
is in records obtained from the arm muscles. It is always high, 
frequencies of 170 to 180 or even 200 per second being met with. 
I shall have to refer again to the difference of frequency obtaining in 
different muscles (p. 237). 

With one subject an attempt was made to alter the temperature of 
the fibres, and records were taken when the recording arm was now at 
room temperature (13^ C), now in an incubator kept at 35° C. Although 
the records were good they exhibit the usual irregularity, so that I 
should hesitate to draw any conclusions from them. I hope to repeat the 
experiment with the subject who gave the fig. 5 records. 

(c) The voluntary effort. — With each person records were taken 
when he was grasping as hard as he could, and the dynamometer scale 
was read. With some people records were also taken when they were 
purposely not exerting their full strength, but such as to make the scale 
read about half or two-thirds of their maximum. The difference in the 
strength of the effect, i.e. in the steepness of the ascents and descents of the 
undulations on the curve, was very marked in any two such records taken 
with the same person, just as the difference of amplitude of the swings of the 
fibre of the string-galvanometer was marked in Piper's records. The whole 
response was, however, of the same type, i.e. the outline of the curve was 
equally irregular in the two cases with most people; and it was just as 
regular when the contraction was weaker as when it was strong in the flexors 
of the individual with whom the fig. 5 records were taken, although the 
jags were sometimes so small that they were difficult to count. To ascertain 
whether or not the frequency of the rhythm varies with the strength of 
the effect, the records obtained with this individual are the most useful. 



The Electrical Response of Muscle 227 

The following table shows the frequencies which obtained, and the 
relative length of time for which each obtained, in eleven successive records, 
seven of which were taken when he was squeezing the dynamometer as 
hard as he could, the four others when he was purposely doing less : — 

Dynamo- 
meter Frequencies per second obtaining for the lengths of time indicated hy the lengths of line, 
reading. 

.57 67 86 58 115 

15 ; 1 1 ! 1 > 



13 - - 


50 

1 


? 


64 

1 1 


? 






55 


1 


115 


90 ' 


55 


1 


62 


83 




60 


86 5S 


86 

1 1 


55 


19 




1 






42 




45 54 


44 


80 










1 




1 




1 





83 60 78 62 115 62 74 60 66 

21 1 1 1 i ^ 1 1 , 

59 115 70 60 115 87 74 50 



14 



81 55 100 57 88 72 84 74 

' 1 1 1 1 1 1 



60 74 100 



18 -^- 



42 100 77 ? 66 'I 60 80 

1 j 1 1 ' 1 1 



? 65 74 80 52 75 60 

20 < , 1 i 1 1 1 



53 76 54 80 



20 < 

0-1 .-^ec. 



-I 1- 



As his reaction time was extraordinarily long (3 to 5 tenths of a second), he 
was given the signal to clench before the trolley had quite reached the break 
key when all but three of these particular records were taken. The etiect 
however, must have onl}- just begun when the plate began to pass the 
slit. In two of the records which show the start the frequency is certainly 
lower than it ever seems to be subsequently, i.e. during the length of time 
the plate took to pass (a little over a second). I have noticed such a slow 
beginning in records taken with a few other people, but it is not universal, 
nor did the attempt to give a slow instead of a sharp contraction to the 
maximum make it appear when it was not otherwise so. In the present 
instance, as in most others, the attempt was made to bring the pointer of 
the dynamometer up to whatever was intended to be its maximum reading 
as quickly as possible after the signal had been given. I can tind no 
evidence in the measured records of any definite relation between strength 
VOL. I., NO. 3. — 1908. 10 



228 Buchanan 

of mechanical effect and frequency, so that I am inclined to agree with 
Piper that the amount of the voluntary effort does not affect the response 
fre(|uency, even though I find this to be so very far from constant. The 
two records reproduced in fig. 5 were the third and the last, in both of 
which the squeeze was maximal. 

When we come to compare the mechanical with the electrical effect of 
a voluntary effort in different people, we find that the stronger mechani- 
cal effect is by no means always accompanied by the stronger, or even by 
the more regular, electrical effect. Thus the maximal reading of the 
dynamometer produced by the person just referred to as giving the strongest 
and most regular electrical response was 21, and on most occasions when 
he was trjang to do his utmost the reading fell short of this and was only 
18, 19, or 20. Almost immediately afterwards the experiment was repeated 
(with exactly the same recording arrangements) with the flexores digitorum 
of another subject who brought the reading up to 83 while the electrical 
response was being recorded. The records showed that the separate 
excursions of the meniscus were less strong (ascents less steep) and also 
less regular, although more regular than in a good many people. It would, 
I think, be premature to do more than call attention to this fact at the 
present moment. It is illustrated by a comparison of figs. 4 and 6 A with 
fig. 5 A, the scale reading 25 with each of the two people whose flexors gave 
figs. 4 and 6 A. So far as the strength of the electrical efl'ect is concerned, 
since we know that in the same person this varies with the strength of 
the mechanical efl'ect, we must look to something outside the muscle and 
the voluntary eflbrt to account for the differences in different people. 
Variation in conductivity of tissue (fat and skin) interposed between muscle 
and electrodes is what first suggests itself, but I have as yet made no 
attempt to measure this. 

IV. Contrast between the Capillary Electrometer Record of 
THE Voluntary Response and that of the same Muscle to 
Artificial Excitation of its Nerve by a Series of Induction 
Shocks of a Frequency of about 50 per second. 

Below the record (tig. 6 A) given by the flexor muscle in response to the 
will there is reproduced (flg. 6 B) the record (on a plate moving at the same 
rate) given by the very same muscle immediately before (when the leading- 
ofl' electrodes and the surface to which they were applied were precisely 
the same) in response to a series of induction shocks applied to the skin 
over the median nerve where it comes nearest to the surface just under 
the lower end of the biceps. The artificial stimulus was supplied by 
the vibrator of a Kronecker inductorium with secondary coil right up 
and no core in the primary coil. It produced a strong sensation. 
Contact was made and then broken 44 times a second. Whereas the 
subject was exerting himself in order so to compress the steel ellipse that 



The Electrical Response of INIuscle 

< CO 



ri: 



229 



U 






280 Buchanan 

the pointer niio-ht o-iv,' a hio-h reading when Photo. 6 A was taken, the 
artificial stimulus used when Photo. 6 B was taken only caused his 
fingers to bend over so as to touch the dynamometer, but hardlj- to com- 
press it at all, and not enough to make the pointer move. The electrical 
response began about two-hundredths of a second after a short-circuit to 
the induction shocks had been broken by the break-key inside the dark 
box, and appears as a series of separate meniscus excursions succeeding 
one another at regular intervals of time (corresponding to that between 
each two breaks), not all precisely alike but each being the expression of 
a short and strong electrical variation beginning with relative negativity 
at the proximal contact and ending with relative negativit}* at the distal 
contact. To begin with each contact seems to have become in turn negative 
to the other by the same amount, but after about half a second the relative 
negativity of the distal contact (made with the skin over the tendons of 
the muscle) seems to have become greater than had been the relative 
negativity of the other contact (made with the skin over the most actively 
contracting part of the mu.scle) immediately before. The same indication 
(greater steepness of descent than of ascent) of the relative negativity 
becoming greater at the contact made with the skin over the tendons, 
appears in many of the records obtained when the will supplies the stimulus, 
but only when, as here, the distal electrode made as large a surface of 
contact as the proximal (e.g. fig. 5 A). It must be remembered that 
negativity of a contact with an active spot of muscle to one with a spot 
known to be inactive is often followed by positivity to it (Journ. of 
Physiol., xxiii., p. 335), so that when two possibly active spots are record- 
ing, the relative negativity of the last one to become active may be the sum 
of the negativity of its own contact and of the positivity of the proximal 
contact. I have never noticed greater steepness of descent in the records 
(greater velocity of the abostial than of the adostial movement of the 
meniscus) when both contacts have been made with skin directly over the 
muscle itself, i.e. presumably with nearly equall}' active parts of the 
muscle. 

The steepness of each ascent and descent in the artificial tetanus record 
presents a contrast to that observed in any of the undulations which 
can be recognised as distinct in the record obtained when the very same 
muscle (or group of muscles) was contracting in response to the will, and 
producing a far greater mechanical effect. The meniscus always came to 
rest between each two excursions ; and it did so also, although for a shorter 
time, when the core was in the primary coil, as is shown by two records taken 
under such conditions. When the exciting electrodes were not near enough 
to the median nerve to provoke contraction of the muscle (i.e. movement 
of fingers or of wrist or of both), the photographs showed that the meniscus 
remained at its resting position throughout, even when the shocks were 
distinctly felt down the arm as well as up it (see Addendum, p. 241). 



The Electrical Response of Muscle 231 

V. Conclusions to be drawn from the Foregoing Exi^eriments. 
Further Experiments which favour them. 

The conclusions that I have come to from my own experiments, com- 
paring the records obtained with human muscles with those before referred 
to as obtained under simpler experimental conditions in the frog, and 
having regard in each case to the difference of character between the 
electrical responses to artificial discontinuous stimulation, by means of 
induction shocks at the rate of about 50 per second, and to a central 
stimulus, are : — 

(i.) There is no certain indication that the central stimulus is dis- 
continuous at all (unless interrupted by other stimuli), but if it be so, the 
separate stimuli are likely to be each what v. Kries has called a Zeitreiz. 
Whatever its nature it is certainly not a series of instantaneous stimuli 
recurring with a frequency of 50 per second. 

(ii.) The rhythm which appears in electrometer records of human 
skin-covered muscle in voluntary contraction is as purely muscular in 
origin as is that proved by experiment to be so in the bared muscle 
of the frog thrown reflexly, or in a variety of other ways, into per- 
sistent contraction ; its frequency may vary in different muscles, and 
in different kinds of fibre in the same muscle, between 50 and 200 per 
second. 

Piper's records, which lead him to conclusions so diametrically opposed 
to these, differ from mine chiefly in exhibiting one pai'ticularly dominant 
rhythm, which he thinks is constant in frequency for each muscle, so that 
when a muscle is artificially excited by stimuli of the frecjuency denoted by 
this rhythm, he finds great resemblance where I find great contrast 
when a comparison is made with the voluntary response. 

The first thing which, of course, suggests itself in trying to account for 
the discrepancy is the difference of recording instrument, or rather of 
the particular pattern of each instrument, that we have been respectively 
employing. In order to judge of the suitabilit}^ of a particular instrument 
for recording electrical changes of unknown, but of possibly high, frequency, 
and of unknown, but of probably varying, strength, we must in the first 
place assure ourselves of the quickness with which the recording part of the 
instrument can come back to its position of rest after it has been disturbed, 
and ascertain that this is independent of the amount by which it has been 
disturbed. To show to what extent the particular capillary electrometer I 
have been using is suitable for the purpose, I therefore reproduce here not a 
comparison curve (Aichungscur ve) but what seems to me to be more to the 
point, the photographic record obtained when a series of break and make 
induction shocks of a frequency of about 50 per second are allowed to 
escape into a dead muscle, which is connected with the two terminals of the 
electrometer. In order to be able to measure the intervals between the 
successive ett'ects, the photograph (fig. 7) was taken on a much more rapidly 



232 



Buchanan 



moving plate than are any of the others introduced into this paper. It 
shows that, when the vibrator of the primary coil was breaking and making 
a contact 54 times a second, each induction shock was producing a separate 
excursion of the meniscus, the one in the one direction the other in the 
other. The break effect was strong (ascent, considering the quick rate of 
movement of the plate, steep), the make effect was a good deal weaker, but 
each had the same duration, namely, I'ocr. Between each break and make 
effect, and again between make and break effect, there was an interval 
lasting respectively about 6 and 5 times as long as either of the excursions 
themselves, i.e. about 9a- and Her, during which the meniscus resumed its 
resting position. I conclude, therefore, that the instrument I have been 
usins: would be capable of registering as distinct each of a succession of 
effects produced immediatel}' b}^ instantaneous electrical changes whatever 
their strength, and wliether all of equal strength or not, provided their 




Fifi. 7. — Diagram showing the effect produced upon the cajiillary electrometer 
when four make alternating with five break induction shocks were 
allowed to escape into it. Quick rate of plate, as shown by 100 fork tracing 
above. 



frequency was below 700 per second. When the effects are not produced 
immediately by such changes, the highest frequency with which they 
can appear as distinct depends, of course, also upon the duration of each 
intermediate electrical change. When the electrical \ariation of the flexor 
muscle was interposed between the induction shocks and their effects on 
the electrometer, we may gather from the photographic record reproduced 
in fig. 6 B that the duration of the muscle effect was becoming shorter 
and shorter to successive stimuli. To begin with, it outlasted the stimulus 
producing it by nearly 0"02 second ; before two-tenths of a second were 
over it was outlasting it by only about O'Ol second, and later by an even 
somewhat shorter time still ; but whatever it was, the time taken by the 
instrument remained but a small fraction of it. 

One would like to know, for each of the three string-galvanometers 
which were used by Piper, the time taken by the " string " (quartz fibre or 
platinum-tungsten wire) to come to rest, when disturbed by the immediate 
action upon it of a single induction shock. If we may judge by the escapes 
which are represented in the records reproduced in his second paper [(2), 



The Electi'ical Response of Muscle 233 

pi. i., figs. 1, 2, 3], preceding the response of the flexors to a single break 
induction shock applied to the median nerve, it would appear that the tibre 
then in use took about, or nearly, three times longer to regain its resting 
position (i.e. 4 to 5cr) than did the meniscus of my electrometer, and, more- 
over, that the greater the excursion caused by the shock the longer was the 
time it took to regain it. 

A record which Dr Piper was kind enough to send me at my request, 
of a series of very weak induction shocks of a 50 per second frequency, 
led into the fibre of the galvanometer he was using in his first set of ex- 
periments, showed that the fibre hardly halted (if it did so at all) at its 
resting position while passing it in either direction. The only record 
reproduced of the response of the flexors to artificial excitation of the 
median nerve by induction shocks of this frequency [(1) pi. i., fig. 3] was 
taken with the same fibre, and shows that it then never halted and that the 
swings were all precisely identical. 

Knowing from Einthoven's work (9) how beautifully the string- 
galvanometer can be adapted to record accuratel}- any kind of small 
electrical changes, I do not for a moment doubt that a fibre of (i.) such 
normal sensitiveness, (ii. ) of such length, (iii.) of such resistance in proportion 
to that of the electrodes used and of the skin and the tissues lying between 
the electrodes and the muscle, (iv.) of such tension that its excursion is 
shorter than that, not only of some, but of all the variations it is likel}^ to 
have to record, and yet (v.) so damped that it has no periodicity of its own 
— that such a fibre could be chosen and placed in such a magnetic field that 
it would reproduce the electrical variations which occur in human muscle in 
voluntary contraction better than the capillary electrometer can ever hope 
to do. Yet just because it has to be adapted to what it has to record, it 
seems to me that the string-galvanometer can never take the lead in giving 
information about electrical changes the strength of which and the frequency 
of which are unknown. A fibre of such kind and so arranged (as Piper 
says his was) as to reproduce accurately the electrocardiogram of man, is 
by no means therefore fitted to reproduce accurately electrical varia- 
tions, seldom of much greater amount and some of them of smaller amount, 
recurring with a much greater frequency. It seems to me that it is not 
until a quick capillary electrometer has shown the nature of the varia- 
tions to be registered that the time has come for the string-galvanometer 
to help us to understand better than we can from its own records what 
these mean. That it can then do it better follows from the fact that it 
records directly the relative diflerence of potential at each moment, 
whei-eas capillary electrometer records have always first to be inter- 
preted in order to show this ; and although it is easy enough for anyone 
accustomed to dealing with them to roughly interpret them at a glance, 
to do so accurately involves a great deal of time and labour. 

Until I know, therefore, that the special fibres used by Piper were so 
chosen and so arranged that they had no periodicity of their own, and that 



234 * Buchanan 

an excursion produced by a single induction shock was of as short a 
duration as it was with my capillary electrometer, and that it was equally 
short whatever the amount of the excursion, I cannot accept any evidence 
given by his records which is not in accordance with that given by mine. 
I cannot, that is to say, grant that for the lower-arm flexors the rhythm 
has a constant frequency of 47 to 50 per second, as he says it has in 
his lirst two papers, or even one varying only between 47 and 58 per 
second [(3), p, 510]. With regard to the masseters it may be noticed that 
he himself came to different conclusions as to their response frequency when 
he was using different instruments. With the first [(1), p. 332] he found it 
to be very irregular and varying between 60 and 80 per second ; with the 
second [(2), p. 412] he found it to be 60 to 64 per second ; with the third 
[(3), p. 509] 88 to 100 per second. Judging from the unpublished masseter 
records which he has been kind enough to let me see, taken, I believe, with 
the third instrument, I should myself have said that the response was 
characterised by greater regularity than was obtained with any other muscle, 
and that, when most regular, frequencies of 96, 100, 115 or 120 per second 
prevailed, although elsewhere it was sometimes 70 to 80 per second. My 
own records also show a more regular response with this muscle than is, as a 
rule, obtained with the flexores digitorum, although the frequency in the one 
person whose masseter response I have i-ecorded was as high as from 170 to 
200 per second. Some of Piper's other records, referred to, but not pub- 
lished, in his third paper, taken with several different muscles for a few 
seconds at a time, would also seem to me to show, as my own records do, 
that now one response frequency obtains, now another. With the gastro- 
cnemius, for instance, I should have said that the records he sent me showed 
that frequencies of 73, 87, and 100 per second prevailed rather than the 44 
which he estimates it at ; with the extensor pollicis brevis 63 per second 
rather than 47 to 50 ; with the quadriceps femoris I would agree with 
him that one of the prevailing frequencies is about 44 per second, but I 
find others in the same, or in other records, of 52, 66, and 76 per second. I 
am, of course, counting something different from what he is, but the fact 
that two people may come to such different estimates with regard to the 
frequency from the same records shows the need for something more 
striking than he has yet given us, in the way of evidence that the muscles 
supplied from any one special centre in the cord or in the medulla oblongata 
all exhibit the same response frequency [(3), p. 516]. 

To my second conclusion (p. 231), that the rhythm observed originates 
in the muscle rather than in the central nervous system. Piper raises 
two objections [(1), p. 328], namely: — 

(i.) That if the rhythm were not due to the central nervous system, we 
should expect to see in the records indications of another, and slower, 
rhythm as well, and that we do not. I do not see the necessity for expect- 
ing it. It by no means always appears, as I have shown (both in 1901 and 
in the present paper) in reflex responses of frog's muscle, even when in 



The Electrical Response of Muscle 235 

strychnine spasm. Granting for the moment that a series of Zeitreize are 
coming from the cord, their frequency would only manifest itself in the 
records if the interval between one Zeitreiz and the next was so long that 
the effect of the first was over before that of the next began. It very 
frequently becomes so, or can be made to become so, in the strychnine frog, 
and then the rhythm of from 3 to 14 per second, already described, mani- 
fests itself in the records, with the quicker rhythm of 40 to 120 per second 
superimposed upon it. On the other hand, I am not at all sure that it is 
true that we have no evidence of a slower rhythm as well as of one vary- 
ing between 40 and 120 per second, in the lower-arm flexors of man 
contracting at will. Many of my records show, as I have said (p. 223), a 
grouping of the undulations on the curve, the groups occurring with a 
frequency of from 14 to 30 per second. The beginning of such a group is 
always a particularly steep ascent, which means a particularly strong 
variation, such as I have shown (1901) to be frequently present at the 
beginning of each wave in the strychnine spasm record of the frog. 

(ii.) Piper's second objection to my view is that, " keeping to physio- 
logical conditions of experiment " (by which, I suppose, he means avoiding 
the use of drugs such as strychnine) a muscle never responds with 
rhythmical oscillations to a single stimulus. This is true perhaps with 
most muscles, if the single stimulus is instantaneous. But, as I have 
recently shown [(10), fig. 2], even to a single instantaneous stimulus the 
gastrocnemius of the frog may give more than one oscillation in the record 
of the electrical response ; and it and other muscles certainly do so when 
excited by break ascending galvanic currents [(4), pi. vii., ph. 33, 34]. 
Moreover, with a stimulus of somewhat more appreciable duration, such as 
that given by Garten's guillotine [(6), p. 340], oscillations seem always to 
occur. So far as I am aware, no records have been taken with a quick 
capillary electrometer of the response of muscle excited by v. Kries's 
Federrheonom (7), and the oscillations in cjuestion are far too <|uick to 
be observed by the eye. 

The view which I am now upholding, that the rhythm of the electrical 
response accompanying voluntary contraction is mainly of peripheral 
origin, is that of Wedensky [(11), p. 260]. I was unable to accept it in 
15)01 [(4), p. 152], partly because I had then no records of the voluntary 
response to compare with my records obtained with frog muscles, and 
partly because I was not always able to follow, as I then said (p. 139), 
his descriptions of the sounds he heard with the telephone. 

The fact that the frequency is not altered by the strength of 
the central stimulus provoking it — a fact insisted upon by Piper and 
not contradicted by anything in my records — seems to me, as far as it goes, 
to furnish evidence in favour of this view, or rather against the view that 
the response f)-e(]uency represents the innervation frecjuency. I have already 
shown [(4), p. 135] that when artificial stimuli of a frequencj^ capable of 
impressing itself upon the muscle, are applied to the nerve of the frog's 



236 



Buch 



sartorius, the muscle responds by a lower frecjuency if the stimuli are 
weak [(4), pi. vi., ph. 80 and 81]. Experiments made soon afterwards (in 
1902), but of which no account has yet been published, showed what I was 
then not expecting, namely, that when stimuli of a frequency so great as to 
be incapable of impressing itself upon the muscle were applied to the nerve, 
however nuich their strength was reduced, the response frequency was not 
reduced ; that is to say, although the response frequency differs in different 
records, as it always does when it bears no relation to the exciting fre- 
quency, it does not vary with the strength of the stimulus. Fig. (S A 
is a record of the response of a sartorius muscle the nerve of which 
was excited by the vibrations of a 500 fork acting upon a telephone 
transmitter, with a whole Daniell in the primary circuit of the telephone 
transmitter. The undulations are seen to recur Avith a frequenc}^ of about 




Fig. 8. — Electrical response of frog's sartorius to excitation of the nerve by currents of high 
freijuencj^ (to a tuning fork giving 500 d.v. per second, sounded in front of a telephone trans- 
mitter). Recording surface describing an arc. Rate of movement indicated by a 500 fork. 
A, a whole Daniell in primary circuit. B, only a third of a Daniell in primary circuit. 

88 per second. Fig. 8 B is the record obtained when only one-third of 
the Daniell was in the primary circuit. The muscle gave a much smaller 
contraction, and all four times that the electrical response was recorded 
with the strength of the stimulus so much reduced, it began, as in the 
record reproduced, a much longer time after the currents began to act 
on the nerve, i.e. after the break of a short circuit by the signal s. 
When the response began, however, the undulations had always a fre- 
quency as great as in any of the records taken with a stronger stimulus. 
In all four it was about 100 per second, whereas in the eleven records 
which were taken with the stimulus stronger the frequency varied between 
65 and 100 per second. A stimulus therefore, which may be regarded as 
a continuous one may produce the same rhythm in muscle whatever its 
strength, whereas a series of stimuli of a frequency low enough for the 
muscle to follow, as one of 50 per second would be, and is, may fail to 
be all of them effectual when their strenofth is reduced. 



The Electrical Response of Muscle 237 

Granting that the rhythm in voluntary contraction is, or is mainly, of 
peripheral origin, how are we to account for the differences in fre- 
(piency which prevail in my records of the responses of one and the same 
muscle, and for those whicli both Piper and I find to exist in different 
muscles '. To anyone who has taken records of the electrical response of 
tlie same and of different muscles in a large number of strychnine frogs, 
there is nothing surprising in the amount of variation which occurs in the 
frequency of anything which may be compared to the wavelets. As I have 
said elsewhere [(4), p. 148], the frequency of these in the frog's sartorius 
in strychnine spasm maj^ vary in different preparations from 40 to 100 
pjr second. Although I did not actually state for the strychnine reflex 
i-esponses that the frequency is generallj^ greater in the gastrocnemius than 
in the sartorius, the records reproduced [(4), pi. vii. ph. 36 (sartorius) and 
38 (gastrocnemius)] showed it, and the}^ were, and are, typical. I did, as a 
matter of fact, draw attention to the difference of behaviour of the two 
muscles in this respect when the response was to a continuous stimulus, 
or rather to one of very high frequency [(4), p. 139] ; or when it was to 
the break of an ascending current through the nerve [(4), p. 142, and pi. vii. 
ph. 33 (sartorius) and 34 (gastrocnemius)]. 

In the reflex responses of strychnine a wavelet f requeue}" of about 100 
per second is most often met with in records taken with the frog's gastro- 
cnemius, and the frequency is much more constant (whether there are waves 
or not) than it is with any of the other muscles I have used. In records 
taken with the sartorius a wavelet frequency of 50 to 60 per second more 
usually presents itself, either by itself or side by side with a quicker rhythm. 
Recent experiments have shown that with the triceps femoris as with the 
gastrocnemius a quick rhythm is more usual, while with the biceps femoris 
and the semitendinosus two or even three different frequencies are apt to 
prevail in turn in one and the same response, so that the}* behave more like 
the sartorius. 

With regard to the signiflcance of these differences of rhythm in the 
different muscles of the frog, I would suggest that the}- are due to the 
differences which Bonhoffer (12) has shown to exist in the kind of fibre 
composing the different muscles, and in the proportion of one kind to 
another in a single muscle. In view of the fact that it is the thick, '' quick " 
ffbres (of Grlltzner), those which are poor in sarcoplasm, which are present 
almost exclusively in the gastrocnemius and the triceps, it would be these 
ffbres which exhibit a response frequency of about 100 per second to any 
kind of stimulus which is not of such nature as to impress a rhythm of 
its own upon them. In the sartorius and biceps there are as many, or 
even more, thin, "slow" ffbres, rich in sarcoplasm, present. Bonhoffer 
says nothing about the semitendinosus. My experiments suggest that in 
it also the two kinds of fibres would be found. 

I have not been able to find any description of the histological structure 
of the masseters of man compared with that of the flexores digitorum 



238 ' Buchanan 

(or other arm muscles), but my records of the electrical voluntary response, 
taken with the two muscles — those with the one giving principally a 
rhythm of over 170 per second, those with the other giving one tliat varies 
between 40 and 120 per second — suggest that histological investigation 
would show that the tibres composing the masseters are much more like 
one another than are those of the flexores digitorum, and that they are all 
of the " quick," thick variety. I do not see why it should be necessary to 
assume, as Piper does [(1) p. 331], that waves must travel with the same 
speed along all the fibres of the flexors. My recently obtained records lead 
me to expect to find fibres of different kinds in the limb muscles of man 
just as much as they are known to occur in those of the frog. By the same 
token I can now no longer doubt that the rhythmical electrical effects so 
often observed in frog's muscle responding to continuous stimulation, are 
normal, a point upon which I hesitated to express an opinion in 1901 
[(4), p. 138]. The occasional absence of rhythmical effects with the three 
kinds of non-discontinuous stimulation I then employed, and its universal 
absence in veratrinised muscle, may be accounted for in a way which I hope 
to discuss elsewhere. 

VI. The Reflex Electrical Response in Max. 
When the median nerve is stimulated by a single strong break induc- 
tion shock while the flexor muscles in the lower arm are connected with 
the capillary electrometer in the usual way (p. 221), the meniscus begins to 
inscribe a curve on the plate about ocr later. It indicates that the proximal 
contact and then the distal became each in turn negative to the other, and 
that at the end of about a hundredth of a second, although there was slight 
persistent relative distal-contact negativity, the effect of the direct stimulus 
was nearly over ; and that then, some Ave or six hundredths of a second 
later, the proximal contact and then the distal each became negative in turn, 
three or four times in succession, but by a very much smaller amount than 
before. In view of the records obtained when the intact mixed nerve 
supplying the gastrocnemius of the frog is stimulated by a break induction 
shock, which I have recently published [(10), ffgs. 1, 2, 3], I cannot help 
regarding the second effect as a reflex effect, i. e. as the effect produced by 
the artiflcial stimulus on afferent flbres, and only on the muscle after the 
intervention of the central nervous system. Fig. 9 shows one such response. 
The second effect, lasting some three or four hundredths of a second, only 
made its appearance when there was a core in the primary coil, but it then 
appeared (in the one person so far who has been always successful in apply- 
ing the small exciting electrodes used, to the only place on his arm which 
will serve to produce the direct response, and therefore the same person 
who supplied the two records reproduced in ffg. 6) with the secondary coil 
at 7000, there being as usual only one Daniell in the primary circuit. The 
six records in which the second effect is seen, some of them taken on one 
day and some on another, all show that it occurred at almost precisely the 



The Electrical Response of Muscl( 



289 



same niouient after the stimulation of the median nerve, i.e. after 72 to locr. 
1 f I am right in m}' interpretation of it, and if we allow from 20 to SOa- 
for the time taken by the impulse to travel the distance of about a metre 
along the nerve from the place of stimulation to the neck and back again, 
and if we deduct this and the Scr taken b}" the impulse to reach the first 
recording spot of the muscle along the motor fibres, from the 75cr, we should 
arrive at the conclusion that 40 to 80o- had been spent in crossing the 
synapse, or the synapses, in the part, or parts, of the central nervous 
system concerned in such a simple reflex. 

If we examine the records taken bj^ Piper of the same thing [(2), pi. i., 
figs. 1, 2, 3], we see also, some 860- after the first effect is over, and thus 
some llo(T after the excitation, a very small second effect followed a little 




Fig. 9. 



•DiiTi't. and lefiex electrical resj)Oiise of the f I e x oi'e s d i git i> rii 
break induction shock applied to the median nerve. 



later by others of the same kind. He has just drawn attention to its 
presence on p. 399. I would suggest that it also is a reflex effect. 

The contrast between the direct and the reflex effect in fig. 9 seems to 
me to be only an exaggeration of the contrast between the direct response 
to artificial stimulation recorded in fig. 6 B and the voluntary response 
recorded in fig. 6 A. The one effect is strong and single, stronger than any 
of the effects (produced without the core in the primarj' coil) in fig. 6 B ; the 
other effect is very weak, weaker than the voluntary response, but pro- 
longed. This difference between the electrical effects pi-oduced when, on the 
one hand, the motor nerve is excited by an instantaneous stimulus, and, on 
the other, by a stinmlus coming (proximateh-) from the centre, together with 
the fact brought out so plainly- in my experiments (pp. 228 and 241) that 
the mechanical effect produced by the central stimulus may be so very 
much greater than that of a series of artificial direct stimuli, sugofests that 



240 Buchanan 

the mechanical effect depends for its strength not so much on the strength 
of the stinnikis as on its duration. An artificial instantaneous stimulus or 
each of a series of such produces a greater electrical variation than is, in 
the same muscle of the same person, produced by a central stimulus, but it 
lasts so much shorter a time that it is unable to efiect what a more pro- 
lono-ed, even though weaker, stimulus may do. In this connection I was 
interested to learn recently from Dr Waller that he had long ago recorded 
the reflex mechanical efiect of the extensors of the leg in man to a single 
instantaneous stinuilu.s to the nerve, and that he had found it to be much 
stronger in proportion to the direct effect than my records show to be the 
case for the reflex electrical effect. 

VII. Summary. 

In the electrical response of human muscles to their normal stimulus a 
rhythm presents itself which for the f lexores digitorum has no constant 
frequency but may vary during one and the same response between 50 and 
about 120 per second, but which for the masseters is more uniform and 
of a higher frequency (170 to 200 per second). 

This rhythm is to be regarded as of peripheral origin, because it is of 
the same order and subject to the same sort of variation as is a rhythm 
which experiments on frogs have proved to be of such origin. 

As the corresponding rhythm may be produced in the response of frogs' 
muscle to various kinds of artificial stimuli which are not discontinuous, 
the rhythm obtaining in the electrical response of human muscle to the 
will has as yet given us no information as to whether the natural stimulus 
is rhythmical. Such information can only be gained by further experiments 
upon animals (not only upon cold-blooded, but on warm-blooded ones), in 
which it is possible to simplify and control the conditions to a greater 
extent than is possible in man. 

VIII. Note on the Effect ox the Electrocardiogram of the 
Contraction of the Voluntary Muscles in Man. 

If, while the electrocardiogram is being inscribed by the meniscus of the 
capillary electrometer, the subject having his two hands in two basins of 
salt solution connected with the two terminals, the one fist is clenched, a 
number of line teeth appear on the curve which before was smooth, both 
on that part of it which corresponds to the systole and on that which 
corresponds to the diastole. It was the observation of this effect which 
suggested the experiments on arm muscles described in this paper. The 
teeth do not appear when muscles of other parts of the body alone are 
put into action. They have a frequency of from 100 to 140 per second, 
which I take therefore to be that of the electrical response of the muscles of 
the hand. But there is another effect produced by clenching one fist on 
the electrocardiogram which is equally striking, namely, the sudden and 



'Die Electrical Response of Muscle 2-il 

immediate shortenini^ of the time between each two periods of activity of 
the heart. Even so slight an action as this, wliich does not affect the 
respiration, may considerably reduce the pause between the next two 
systoles, and within the next five seconds it may become half what it was 
just before when the whole body was at rest. The quickening is not 
maintained for long, so that in counting the pulse for one minute the 
increased frequency is less striking. There is, of course, the same sudden 
temporary quickening when other' muscles are put into action or when a 
deep breath is drawn. 

IX. Note on the JMisuse of the Word "Rhythm." 
I tried in my former paper (4) to avoid the use of this word for a suc- 
cession of changes characterised by not recurring at regular intervals. I 
still think it a misnomer to so apply it, but the word is so convenient, 
and has been adopted by other people, e.g. Garten (6) and Piper (1), 
regardless of any exception which might be taken to it, that I have 
now myself applied it and its derivatives to phenomena which recur with 
approximate, although not absolute, regularity. 

The working expenses involved in making all the experiments referred to 
in this paper, with the exception of those made in 1902, have been defrayed 
out of a grant from the Go\'ernment Grant Conunittee of the Roj'al Society. 

I have to thank Professor Bourne and Profes.sor Osier for providino- 
me with the accommodation necessary for cai-rying on this and other 
experimental work. 

Addexdu^i. 
Since the above paper was written, another medical undergraduate has 
enabled me to record the electrical response of his lower-arm flexors to a 
series of artificial stimuli applied to the median nerve. It contrasts with 
the response to the will in the same way as that to which fig. (> B refers 
contrasts with that recorded in fig. 6 A, but the records are of exceptional 
interest because the voluntary electrical efi'ect was exceptionally- strong 
and regular, and the involuntary electrical ettect accordingly a great deal 
stronger and more regular, although accompanied as before by a very much 
smaller mechanical efi^ect. Records of the voluntary response show that 
the electrical effects recurred at difierent times with frequencies of 100, 77, 
.56, and 52 per second. In fig. 10 A, part of a record is reproduced which 
exhibits chiefly a frequency of 56 per second. Fig. 10 B is the record of 
the very next response of the same muscle (with the contacts in nowise 
altered), not to the will, but to a series of 35 double induction shocks 
succeeding one another at the rate of 50 per second — the secondary coil 
being right up, but no core in the primary. The dynamometer was still in 
the .same position in the hand, but while the voluntary eflbrt producing the 
('fleet recorded in A made the pointer move to 87, the strong artificial 
excitation producing the efl'ect recorded in B made the pointer move only 



242 



Buchanan 



to 2 on the scale. Fig. 10 B gives the same evidence as fig. 6 B, that the 
first few shocks produce ettects which outlast them by a longer time than 
the later shocks, and that between the successive effects, when these are 
each short, there is an interval during which there is little or no difference 
of potential between the leading-ofi' contacts. Owing, however, to the 




Fig. 10. 

strong positive after-efi'ect (relative distal negativity) at the end of each 
individual response, the absence of a fresh change at the proximal contact 
is denoted not by the meniscus being at rest as it was in the intervals 
when fig. 6 B was recorded, but by its striving to return to its zero position 
in the way which it eventual]}' does when all stimulation has ceased. 



REFEREXCES. 



(1) Piper, Arcli. f. d. ges. Physiol., vol. cxix., pp. 301-338, 1907. 

(2) Piper, Zeitschr. f. Biol, vol. 1., pp. 393-420, 1 908. 

(3) Piper, Z. f. Biol, pp. 504-517, 1908. 

(4) Buchanan, Journ. Physiol., vol. xxvii., pp. 95-160, 1901. 

(5) Burdon-Sandersox and Buchanan, Proc. Physiol. Soc, July 1902 ; Journ. 
Physiol., vol. xxviii. 

(6) Garten, Abb. d. .siicbs. Ges. f. "Wissenscb., matb.-pbys. Klasse, vol. xxvi., 
pp. 331-414, 1901. 

(7) v. Kkies, Arcb. f. (Anat. u.) Physiol., 1884, pp. 337-371. 

(8) V. Kries, op. cit., 1895, pp. 130-141. 

(9) EiNTHOVEN, Annal. d. Physik (4), vol. xxi., pp. 483-700, 190G. 

(10) Buchanan, this Journal, vol. L, pp. 1-66, 1908. 

(11) Wkdensky, Arch, de Physiol, norm, et path. (5), vol. iii., 1891. 

(12) BoNHOFFER, Arch. f. d. ges. Physiol., vol. xlvii., pp. 125-146, 1890. 



ON VAGUS CURRENTS EXAMINED WITH THE STRING GAL- 
VANOMETER. By W. EiXTHOVEN (in collaboration with A. Flohil 
and P. J. T. A. Battaerd). (From the Physiological Laboratory of 
Leyden. ) 

{Received for publication 6tli July 1908.) 

The demarcation current which may be derived from the peripheral end of 
a divided vagus nerve shows a negative variation whenever the lungs are 
expanded. This phenomenon was lirst demonstrated by Lewandowsky^ 
with the aid of a Deprez-d'Arsonval galvanometer; later Alcock and 
Seemann- studied it more in detail with the capillary electrometer. 
In a research made by means of the string galvanometer we have been 
able completely to confirm the results of the above-mentioned authors, and 
also to obtain evidence of the presence of another source of stimuli 




Figure showing the electrical changes in the vagus nen'e which accom- 
pany the respiratory and heart movements. 

V, electrovagogram ; ;;, respiration record (ascent of curve, inspiration; descent, 
expiration) ; c, pulse record. 

situated in the heart — the vagus shows rliythmic action currents which 
synchronise with the heart-beats. 

A vagus nerve in the neck of an anaesthetised dog is isolated for a con- 
siderable distance and divided at a spot high up in the neck. The current 
is led ofi" from the peripheral cut end of the nerve by means of a pair of 
non-polarisable electrodes, one of which is brought in contact with the cross- 
section, the other with the surface, the space between the electrodes being 

' Max Lewandowsky, " Ueber Schwankungen des Vagusstromea bei Voluniander- 
ungen der Lunge," Pfliiser's Arch. f. d. ges. Physiol., Bd. Ixxiii., S. 288, 1898. 

- N. H. Alcock aim John Seeiuann, " Ueber die neg-alive Scliwaukung in den Lungen- 
fasern des Vagus," PHiiger's Arch. f. d. ges. Physiol., Bd. cviii., S. 426, 1905. 

VOL. I., XO. .-3. — 1908. 17 



244 ' Einthoven 

about 1"5 cm. The demarcation current of the nerve is compensated in 
the usual way, and the electrodes are connected with the galvanometer in 
such a manner that a decrease of the demarcation current, or an action 
current, causes the image of the string to move in an upward direction. 
In the accompanying figure, 1 mm. of the abscissa corresponds to 2 sec. ; 
1 mm. of the ordinates, to 2'7 microvolts. The upper curve (v) represents 
the action cui-rents of the vagus nerve, and may be called the " electro- 
vagogram " ; the middle one (p) reproduces the respiratory movements of 
the animal in such a way that every inspiration corresponds to an ascent 
and every expiration to a descent of the curve. This pneumogram is 
obtained by causing the dog to breathe into a large bottle and transmitting 
the oscillations of the pressure of the air in the bottle to a suitably placed 
recording Marey's tambour. The lower curve (c) indicates the sphygmogram 
of the crural artery. 

It can be seen that the electrovagogram exhibits undulations having a 
double rhythm, viz. : (1) synchronous with the respiratory movements, (2) 
synchronous with the heart-beats of the animal. That these undulations 
are caused by the real action currents of the nerve and not by current 
escape from other organs, can be proved in different ways. 

In the first place we may rely on the method employed, which does 
not appear subject to this source of error. We have further repeatedly 
submitted it to the usual proofs of control. Thus, if the nerve is ligatured 
peripherally to the electrodes or is killed by means of a drop of ammonia, 
the rliythmic oscillating image of the string is brought completely to rest. 

In the dog, the respiratory oscillations of the right vagus are larger 
than those of the left one ; while, on the other hand, the heart-beat oscilla- 
tions of the left vagus surpass those of the same nerve on the right side. 
In the rabbit the vagus nerve shows exclusively respiratory oscillations, the 
depressor nerve solely heart-beat oscillations. 

The excitatory state of the vagus endings in the lungs is determined by 
the volume of these organs and not by the intrapulmonary pressure. If 
during the cessation of the natural respiratory movements of the animal 
tlie lungs are insufflated artificially, the electrovagogram shows an undula- 
tion in the same direction as if the animal is making a natural inspiration. 
Nevertheless the insufflation is accompanied by an increase, the natural 
inspiration by a decrease of the intrapulmonary pressure. 

The experiments which we have performed on the effects of insufflating 
and deflating the lungs have given evidence of the presence of two kinds 
of pulmonary vagus fibres : those having expiratory and those having 
inspiratory efiects. The latter act more weakly, are sooner tired, and are 
killed more easily by injuries than the former, so that the two kinds of 
fibres can be separated from each other as to their action. Especially is it 
easy to isolate the effect of the expiratory fibres. The theory of Hering^ 

^ E. Hering, "Die Selbststeuerung der Athiuung dureh den Nervus Vagus," Wiener 
Sitzungsberichte, 2 Abt., Bd. Ivii., S. 672, 1868. 



Vagus Currents Examined with the String Galvanometer 245 

and Breuer ^ on the self-regulation of tlie respiratory movements obtains 
in this way fresh support. 

The connection of the respiratory and the heart-beat undulations in the 
electrovagogi-am also throws light on the association which virtually exists 
between the heart's action and the respiratory movements. This associa- 
tion finds inter alia its expression in the constant ratio between the 
number of respiratory movements and the number of heart-beats per 
minute. It is well known that this ratio usually remains unaltered with 
variations in pulse frequency. Since we now learn that the beating heart 
sends rhythmic stimulations along the vagus to its centre, we may 
reasonably ask whether these stimulations may not reach the respiratory 
centre and there cause with each variation of the pulse frequency a 
proportional variation of the frequency of the respiratory movements. 
In this way an automatic regulation of the respiratory movements would 
be promoted by the heart's action. 

Yet another automatic regulation may be pointed out, viz. that of the 
heart's action by itself. We must in all probability look to this regulation 
as furnishing the cause of the vagal tonus, this tonus being thus maintained 
automatically by the action of the heart itself.- 

• J. Breuer, ibid., Bd. Iviii., S. 909, 1868. 

- A more detailed description of our experiment.-^ will appear in Pfliiger'.s Archiv f. d. 
ges. Physiol. 



I 



1 

- i 



A SIMPLE METHOD OF OBSERVING THE AGGLUTINATION OF 
THE BLOOD CORPUSCLES IN GAMMARUS. By John Tait. 
(From the Laboratory of Pliysiology, Edinburgh University.) 

{Received for 'publication 12th July 1908.) 

If the antennae of Gammarus marinus are observed with a low power of 
the microscope the circulation of the blood is readily seen. Gammarus 
has two pairs of antennae of approximately equal length. The terminal 
part in each consists of a many -jointed filament which gradually tapers 
towards the free end ; it is in this filament that the circulation is best 
seen. The anterior antenna is better suited for observation than the 
posterior, which is covered with brushes of setae that obscure the view 
of the main stem. 

In carrying out observations, the animal, after being washed in two or 
three changes of filtered sea- water, is laid on a glass slide and a drop of 
sea-water, free from any suspended particles, so placed as to wet the glass 
all round the antennae. Trouble may be experienced on account of the 
tendency of the creatures to jerk about and move out of the field, but if a 
sufficient supply of the animals is at hand some will always be found to lie 
quiet for a minute or two at a time. If it is necessary to keep the animals 
absolutely still for a prolonged period, a simple plan is to use specimens 
which have become asphyxiated by being innnersed along with many 
others in an insufficient supply of water. If a dozen are kept in a small 
dish of water overnight some will be found in the morning to be apparently 
dead, having ceased to swim or to execute breathing movements. Many of 
such apparently dead animals are alive, their heart continuing to beat, and 
the circulation can now be watched with ease, for the animals continue 
long in this narcotic condition. 

The blood flows in a very definite course in the anteunary filament. A 
main artery leads down the filament ; this forms communications by means 
of cross capillaries (one of which occurs opposite each joint) with a re- 
turning vein, the system resembling a ladder in which the two main 
vessels form the supporting poles while the connecting capillaries 
correspond to the rungs. The existence of these vessels or channels is 
inferred from the very definite path traversed by the blood corpuscles. It 
is only exceptionally that the outline of the vessels themselves can be seen. 

From this arrangement it follows that the blood-flow is more rapid and 



248 Tait 

abundant in the proximal part of the filament than in the distal part, where 
the corpuscles can be seen to travel slowly and in single file. In the vein, 
and still more so in the artery, the movement of the corpuscles is not 
uniform but jerky, each jerk corresponding to a heart-beat. This can 
readily be verified by observations on a young specimen (preferably cooled 
to slow the heart), in which the pulsating dorsal heart is easily visible. 
In the capillaries again and in the longitudinal vessels at the top of the 
antennae the corpuscles are carried slowly and uniformly along, sometimes 
sticking at one spot and then being knocked away again by the impact of 
succeeding corpuscles. Sometimes in a narrow vessel a row of corpuscles 
ma}^ be piled up each one in contact with its neighbour : every time a 
new corpuscle strikes the proximal end of the column one becomes detached 
from the distal end and floats away. From the fact that the capillary flow 
is not intermittent one might infer that the venous pulse is due to direct 
lateral transmission of pressure from the arterial vessel, the whole system 
being enclosed within a relatively rigid tube formed by the calcareous 
rings of the filament. 

Except where the corpuscles accidentally come in contact they float at 
a wide distance from each other in a relatively large volume of plasma. 
They vary both in size and in form. Generally speaking, they are flattened 
and somewhat irregular in shape, though showing no processes that might 
be called pseudopodia. They roll along as they travel, now presenting 
their narrower edge to view and again being seen on the flat. When 
removed from the animal and examined with a high power the largest of 
them are seen full of highly refractile granules which stain deeply with 
eosin. Other smaller ones contain either basophil granules or clear proto- 
plasm. In none is the nucleus lobed as is the case in some of the white 
corpuscles in vertebrates. In all, four or five different varieties of corpuscles 
can be made out by staining. 

If now, during the time that the circulation in an antenna is being 
observed, one amputates the terminal portion with a sharp knife, the rate of 
flow in the artery is greatly increased, while the flow in the vein may cease 
altogether. The blood meantime pours out at the amputated end, the cor- 
puscles being washed to a distance in the water surrounding the antennae. 
If the animal is breathing these corpuscles get washed away in the stream 
of water continually flowing past the antennae owing to the breathing 
movements. If the animal is in a narcotic condition and not breathing 
they settle down on the slide all around the amputated end. Occasionally 
the plasma is brown in colour, in which case it can be seen pouring out in 
the surrounding sea-water like smoke from a funnel. 

Some of the corpuscles, however, from the start begin to adhere to the 
amputated end. At first they stick round the edge of the divided filament, 
but soon increase in number and form a clump or mass all around the end. 
The formation of this clump does not immediately stop the flow of blood, for 
the central part is still kept tunnelled by the rush of outflowing corpuscles. 



Agglutination of the Blood Corpuscles in Ganiniarus 249 

which causes vibrations and upheavals of portions of the clump. The ap- 
pearance at this stage calls to mind that of an active volcano belching forth 
stones and cinders, some of which get piled up round the crater, while every 
now and then the whole mass is shaken and torn hy the eruption of 
material from the centre. No portion of the adhering mass of cells gets 
torn away, however. When once the cells stick in the mass thej' remain 
firmly attached. 

This flow from the central tunnel soon ceases, and the clot becomes 
entirely closed in by the adhesion of the cells all around. The mass which 
they form is at first porous and still allows of the escape of plasma, though 
the corpuscles which are carried down the filament to the spot are held 
back and become piled in a column inside the portion of the filament near 
the wound. The piling up may extend back over one or two segments, but 
not as a rule very far, for a way of escape for the corpuscles is found in one 
of the cross capillaries, and now the venous circulation begins again as if 
nothing had happened. 

This whole process takes from a minute to two or three or more minutes 
according to the thickness of the portion amputated. A wound in the very 
end of the filament is naturally closed up sooner than one nearer the proximal 
end. Once the internal plugging of the vessel has occurred, the terminal 
clot may be removed without any further risk of bleeding. Often enough 
animals may be found in which for some reason or other a whole antennary 
filament has become filled with a mass of corpuscles. Amputation of such 
a filament produces no bleeding. 

The cells which form the terminal clump do not long retain their 
rounded shape. During the time that the blood is still escaping they 
undergo changes in shape, becoming at first wrinkled and still later losing 
their outline and fusing with neighbouring cells. This fusion may have 
taken place in the cells which lie nearest the cut surface of the filament 
even -at the time when other corpuscles are still escaping from the end. 
When the bleeding stops it is not long before the whole clump becomes 
felted into one irregular mass in which no outline of any cell is visible. 
At the same time the mass is seen to have become somewhat shrunken. 
Whether it is one special form of cell among the white blood corpuscles 
that contributes to the formation of the clot, or whether all forms get caught, 
I have not yet determined. 

This method of observing the agglutination of the cells in bleeding is 
simple and retjuires little manipulative skill : it might well be adopted for 
class demonstration. The animals are readily procured, and the process may 
])e repeated two or three times on the antennas of the same specimen. 



ON THE TIME TAKEN IN TRANSMISSION OF REFLEX IM- 
PULSES IN THE SPINAL CORD OF THE FROG. By 
A. D. Waller. 

(Received for p^Mication I4th July 1908.) 

The observations of Miss Buchanan on the time taken in transmission of 
reflex impulses in the spinal cord of the frog have reminded me of some 
experiments I made on this subject in 1884, the conclusions from which are 
in some respects similar to those drawn by Miss Buchanan. In other 
particulars the facts respectively observed by Miss Buchanan and by 
myself supplement each other, obtained as they have been by different 
methods, and altogether independently. 

Miss Buchanan used the capillary electrometer, so that the reaction in 
only one limb at a time could be recorded ; I used a double myograph, so 
that the reactions in two limbs to the same stimulus were simultaneously 
recorded. 

The two methods are indeed complementary of each other, each atibrding 
information of its own not afforded by the other. I must say, however, 
that as regards one principal item — the comparison of delays of \'arious 
kinds of reflexes — it appears to me that the electrical method followed by 
Miss Buchanan is distinctly inferior to the mechanical method of which I 
made use. By the former plan the comparison has to be established between 
successive individual effects, and individual differences of time are not out 
of the question, whereas by the latter it is easy to obtain for comparison the 
simultaneous effects of two different kinds of reflex contraction. 

The two points in Miss Buchanan's results that have most aroused my 
interest in connection with my own results are : 

1. The prolongation of reflex time in consequence of the action of 
strychnine on the spinal cord. 

2. Tlie fact that the cord delay is roughly about twice as great for a 
crossed as for an uncrossed reflex, and the inference therefrom that there 
are normally two spinal synapses interposed in a crossed path and a single 
synapse in any uncrossed path. 

In my own experiments the points that seemed to me to be most note- 
wortliy were : 

1. The excessive prolongation of reflex times after strychnine injection. 

2. The consequent dissociation that can be distinguished of the com- 
ponent parts of which crossed and uncrossed reflexes are made up. 



252 ' Waller 

The inferences which I then drew and still draw differ from those of 
Miss Buchanan: naturall}', at that time I did not think or speak in terms 
of " synapses." Speaking in terms of the afferent and efferent aspects of 
spinal cells, I inferred from the facts under my observation that the delay 
of reflex action, which increases as the toxic effects of strychnine deepen, 
is the index of a gradually increasing block of transmission in the spinal 
cord, and that this block of transmission occurs at first exclusively and 
later chiefly at the junction of the afferent nerve with the cord. 

I have reviewed this conclusion by a re-examination of the facts upon 
which it rested, and in the light of Miss Buchanan's experiments, in order 
to see whether it could be expressed in terms of synapses, and whether the 
conclusion so expressed was in harmony with Miss Buchanan's conclusion. 

But, firstly, as regards the facts : — I found, as Miss Buchanan did, that 
the normal time of a simple reflex was approximately yoo^^^ sec, and that 
of a crossed reflex was approximately j-§„ths. (My actual figures were 
O'OOS and 0"016.) But, unlike Miss Buchanan, I found that this proportion 
of 1 to 2 was not maintained in the prolonged times of simple and crossed 
reflexes observed in strychninised frogs. The times (fig. 3) were, e.g., simple 
0'038, crossed 0-046, difference 0008 ; i.e. the difference of xoo^^^ ^^^- attribut- 
able to transverse transmission, or, let us say, to the interposition of a second 
synapse in the reflex arc, remained unaffected in the presence of a prolonga- 
tion of time as regards the first synapse of approximately yfo^hs over and 
above the normal delay of yooth. 

The case of longitudinal transmission lends itself to a similar argument. 
Fig. 4 of an experiment in which muscular contractions of the arm and of 
the leg of a strychninised frog were recorded, stimulation (by a single break 
induction shock) being applied to the skin of the leg, gives the times : simple 
= 0-060, transmitted = 0-608, difference = 0-008 ; and this normal difference of 
approximately yjo^'h sec, attributable to longitudinal transmission, or, let 
us say, to the interposition of a second synapse in the reflex arc, remains 
unaffected in the presence of a prolongation of time as regards the first 
synapse of approximately y^^jths over and above the normal delay 
of j-^^tli- 

I have expressed values in hundredths of a second for the sake of 
clearness, and for the sake of further clearness I will set out in detail in 
hundredths, and in terms of the synapse conception, what I consider to be 
the components of the total delays exhibited in fig. 4. 





Simple contraction. 


Transmitted contractio 


1st hundredth 

2nd „ ^ 
3rd 


Muscular delay and nerve 
transmission. 


Muscular delay and nerv 
transmission. 


4th „ 
5th „ 
6tli „ J 


First synapse 


First synapse- 


7th „ 




Second synapse. 



Traiismi.ssion of Reflex Impulses in Spinal Cord of the Frog 253 

I must confess, however, that this analysis in terms of synapse delay 
does not give me much satisfaction, for it would require us to say that the 
spinal delay caused by strychnine occurs in the first and not in the second 
synapse. And I still prefer, in the absence of further guidance, to limit 
myself to the conclusion I drew in 1885,^ to the effect that the prolongation 
of reflex time by strychnine is principally caused by a block or functional 
obstruction at the junction of aflerent fibre with spinal cell. 

I have said " principally " where it might have appeared permissible to 
say " exclusively." On attentive consideration of the analysis of flg. 4, the 
whole added delay appears to belong to the first synapse, or — as I prefer 
to say — to the aflerent side of the centre, and there is no appreciable re- 
tardation attributable to the second synapse; i.e. the excitatorj- process 




Fig. 1. — Norma 



(1) Direct muscular contraction. The lost time is 0-008 sec. 

(2) Muscular contraction caused l)y a single break shock to 
upper part of tlu- spinal cord. The lost time is 0-016 sec. 

(3) Muscular contraction (reflex) caused liy a single break 
shock applied t« the .skin of the arm of the game side. The 
lost time is uOlCi sec. 

passes from cell to cell within the cord with normal rapidity. But in 
extreme cases of delay where, as I have observed, the reflex time maj' reach 
^th sec, the ordinary j j\o^h sec. occupied by transmission in the cord by 
one cell to another (or by one synapse to another) is considerably exceeded. 
Thus in the case of fig. 80 of my 1885 paper the values are as follows : — 

Simple reflex, 0-200. 
Crossed reflex, 0-228. 

Analysis in this case (which I may mention in passing gives the highest 
values I ever observed) gives in place of the normal yooth sec. delaj' for 
the first cell (or synapse) the enormous value of nearly ,^foths sec, and in 

• Briti.sh Medical Journal, "Report on KxpeiiuK-nts in the Proce.ss of Fatigue 
and Recovery," 25tli July 1885. 



254 ' Waller 

place of the normal i ^gth for the second cell (or synapse) nearly y^o^hs. 
It is on account of this and similar observations that I have not said above 
" exclusively," but only " principally." 

The difference between (1) the direct and (3) the reflex contraction on 
the same side gives 0"008 as the value of the corrected reflex time. 

The coincidence of connnencement of response to (2) excitation of the 
upper part of the spinal cord and to (3) cutaneous excitation on the same 
side indicates that in both cases the mechanism of the delay is very probably 
of similar character. The longitudinal transmission delay of 0*008, like the 
corrected reflex time of 0*008, is probably occasioned by one similar nerve 
station. Lost time in nerve-fibres, at, say, 30 to 50 metres velocity per 
second, is in this connection negligible. 




Fig. 2. — After stiychnine. 
(About 0-1 milligi-amme injected.) 

(1) Direct muscular contraction, 0*010 sec. 

(2) Spinal muscular contraction, 0-020 „ 

(3) Keflex muscular contraction, 0-056 ,, 

The identical values of the lost times in the two cases (2) and (3), in 
which the cell station at the origin of motor fibre is aroused by a cord fibre (2) 
and by a skin fibre (3) respectively, so that the commencements of the two 
contractions normally coincide, are altered in consequence of the action of 
strychnine. The delay of (2) is very slightly increased ; that of (3) is very 
considerably increased. The augmentation of (2), as of (1) the direct 
muscular contraction, was so small, 0010 as compared with 0-008, as to be 
within the limits of experimental error ; but the augmentation of (3) the 
reflex time, by no le.ss than 0*040 sec, w^as clear and unmistakable : it is a 
measure of the delay suffered by the centripetal stimulus at the junction 
between afferent nerve and spinal cord. 

Miss Buchanan used much lighter doses of strychnine than I did, viz. : 
1 min. of 0-01 per 100 liq. strych. to 2 mins. of 0*02 per 100 liq. strych., 
while I used 2 mins. of 1 per 1200 liq. strych. 



ransinission 



of Reflex Impulses in Spinal Coixl of the Frog 255 



In 3Iiss Buchanan's observ^ations the prominent effects of strychnine 
on cord times were a slight diminution of time for the simple reflex 
and a considerable diminution of the additional time for the crossed 
reflex, e.g., 

Simple. Crossed. 



Normal . 
Strychnine 



0-012 to 0-022 
0-009 to 0-020 



0-010 
004 



while I got considerable augmentation of the simple reflex time with little 
or no alteration of the additional time, e.g., 




Simpl( 

0-010 
0-030 



Crossed. 

0-010 
0-010 




(1) Simple reflex contraction, 

(2) Crosse<l reflex contraction 



O-04tl 
Difference, 0-OOs 



(1) Simple reflex contraction (leg), 0060 

(2) Transmitted reflex contraction (arm), 0"06S 



There is, of course, no actual incompatibility l)etween these two sets of 
flgures taken in light and in deep strychninisatioii respectivel}-. It appears 
to nie, however, that the latter as well as the former require to be taken 
into account before we can accept Miss Buchanan's conclusion that in the 
same-side reflex a single synapse has to be passed, while in the crossed-limb 
reflex two such synapses have to be crossed in succession. Given the fact 
that in deep strychnine intoxication the times of the uncrossed and cross 
reflex movements are, e.g., 0-04 sec. and O'Oo sec, we have to assume on the 



256 ' Waller 

double synapse theory that the tirst synapse is more profoundly affected 
than the second. I still prefer, however, not to make this assumption, nor 
yet that of the double synapse, and to repeat what appears to me to be the 
most that we may legitimately infer, viz. : that under the influence of 
strychnine, transmission of a reflex impulse is blocked principally at the 
junction between aflfereut nerve flbre and spinal cord cell. 



ON THE EFFECT OF STIMULATING THE NERVI ERIGENTES 
IN CASTRATED ANIMALS. By Sutherland Simpson and 
Francis H. A. Marshall. (From the Physiological Department, 
University of Edinburgh.) 

(Keceived for publication llth Juhj iy08.) 

Eckhard^ was the first to show experimentally in the dog that the penis 
could be caused to erect by the stimulation of certain nerves arising from 
the sacral part of the spinal cord. These nerves he afterwards designated 
the nervi erigentes. In the dog they are given off from the first and 
second sacral nerves and sometimes also from the third, but their origin 
from the cord varies slightly in difTerent species of mammals. Gaskell,- 
and subsequently Morat,^ found that the nervi erigentes leave the cord 
by the anterior roots only, and these observations have been confirmed 
by other investigators. 

The erection of the penis is brought about partly through the contraction 
of the ischio-cavernosus (or erector penis) and bulbo-cavernosus muscles, 
certain of whose fibres pass over the efferent vessels of that organ, and so 
arrest the outward flow of blood. The result of this contraction is that 
whereas the blood can freely enter the dilated vascular spaces of the penis, 
its exit is retarded. This leads to a further distension of the vessels, whose 
venous outlets become still more compressed. It is clear, however, that 
whereas the constriction of the outlets assists in causing erection, it is 
incapable by itself of effecting this result, since erection cannot be induced 
by ligaturing the efferent veins.* The usual view is that the nervi erigentes 
exercise an inhibitory influence upon the tonicity of the walls of the vessels, 
and so cause them to distend, and that erection is due chiefly to the direct 
vaso-dilator action of these nerves. According to Langley and Anderson's-^ 
description, stimulation of the nervi erigentes causes inhibition not only of 
the unstriated muscles, but also of the retractor penis, when that muscle is 
present. 

' Eckhard, "Untersucliungen iiber d. Erektion d. Penis beim Hunde," Beitriige zur 
Anat. und Phys., vol. iii., Giessen, 1863. 

2 Gaskell, "On the Structure, Distribution, and Function of the Nerves which innervate 
the Visceral and Vascular Systems," Journ. of Physiol., vol. vii., 1886. 

3 Morat, " Les Nerfs Vaso-dilatjiteurs et le Loi de Majendie," Arch, de Physiol., 1890. 
* See Retterer, article "Erection" in Richet's Dictionnaire de Physiologic, vol. v., 

1902. This article cont;iins numerous references. 

^ Langley and Anderson, "The Innervation of the Pelvis and Adjoining Viscera," 
Journ. of Physiol , vol. xix., 1895. 



258 ' Simpson and Marshall 

The object of tlie present inquiry was to ascertain whether the mechan- 
ism of erection is present in animals which have been castrated prior to 
puberty. It is known that erection can take place in castrated adults, at 
any rate for a considerable time after the removal of the testes, but this 
may be due to the fact that the accessory generative organs, having once 
been developed, do not immediately atrophy. 

Our experiments were upon cats. Before proceeding to operate on 
castrated cats, we successfully carried out two experiments upon normal 
animals. The following is an account of the experiments : — 

(1) An adult male cat was anaesthetised, tracheotomised, and fastened face 
downwards. The spinal cord was exposed in the lumbar region, and the 
dura mater opened. The cord was then transected at the level of the first 
lumbar segment. The anterior roots of the sixth and seventh lumbar 
segments, and of the first and second sacral segments, were exposed on each 
side. The roots were then ligatured and divided between the ligature and 
the cord. The posterior roots of the same segments were also divided. 
The anterior root of the first sacral nerve on the left side was then stimu- 
lated by an induced current obtained from an ordinary induction coil with- 
out a Helinholtz wire in the primary circuit. The current used was slightly 
painful on applying the electrodes to one's tongue. The stimulation caused 
a gradual but distinct erection of the penis, accompanied by a partial 
ejaculation of semen. Microscopic examination of the latter revealed the 
presence of living spermatozoa. The erection gradually subsided on 
shutting off" the current, but it could again be induced on renewed 
stimulation. The same result was produced on stimulating the anterior 
root of the first sacral nerve on the right side. 

(2) This experiment was identical with the first. Erection took place 
as described above, and was almost immediately followed by ejaculation. 
On microscopic examination the ejaculated semen was found to contain an 
enormous number of living spermatozoa. 

(3) This experiment, which was upon a castrated cat, was carried out in 
the same manner, but the result was negative. The animal had been 
castrated before puberty, and when about half grown ; the extirpated 
testicles being approximately as large as peas. The weight of the cat on 
the date of castration (3rd December 1907) was 1130 grams. At the time 
(jf the stimulation experiment (3rd July 1908) the cat appeared to be fully 
grown. Stimulation of the anterior roots of the first sacral nerve on either 
side failed to produce any sign of erection 

(4) This experiment was identical with the preceding, the result being also 
identical. The cat was castrated on 3rd December 1907, when about half 
grown, its weight being 1170 grams. The stimulation experiment was on 3rd 
July, when the cat appeared to be fully grown. There was no indication of 
any erection. 

(5) This experiment was similar in every way to the last two. The cat 
was castrated on 3rd December, its weight at that date being 1060 grams. 



Effect of Stimulating the Nervi Erigentes in Castrated Animals 259 

The stimulation experiment was on 3rd July. The result was entirely 
negative. 

(6) In this experiment, which was a control, the first sacral anterior roots 
of another fully grown normal male cat were stimulated experimentally in 
identically the same way as in the preceding cases. Erection took place 
gradually as in the other two control experiments (1 and 2). Ejaculation 
also occurred, the semen being found to contain numerous spermatozoa, 
some of which were moving when examined under the microscope. 

The experiments show, therefore, that erection cannot be induced 
experimentally in animals which have been castrated prior to puberty, or, at 
any rate, that it is far more difficult to cause erection in such animals. 

It is well known that in animals castrated in early life the secondary 
sexual characters as a general rule fail to make their appearance, and that 
the correlation between the testicles and the accessory generative organs 
is a still closer one. Thus the prostate and Cowper's glands undergo 
atrophy after castration, even in cases where the operation of removal is 
performed after puberty.^ It is possible, therefore, that in animals which 
were castrated when young the muscular apparatus of the penis fails to 
develop sufficiently to admit of erection occurring, but it would seem 
unlikely that the nervous mechanism is impaired. If erection is due 
mainly to an inhibition of the vasomotors of the penis, as is ordinarily 
supposed, there would seem to be no theoretical reason why it should not 
be possible to bring about that process experimentally (or at any rate to 
produce a partial erection) in animals which have been castrated. It 
would appear, therefore, that the process of erection is very possibly a 
more complex phenomenon than is generally believed, but our experiments 
throw no further light on the mechanism of that process. 

' Griffiths, "Observations on the Formation of the Prostate Gland in Man and the 
Lower Animals," Journ of Anat. and Phys., vol. xxiv., 1890. Wallace, " Prostatic Enlarge- 
ment," London, 1907. 



VOL. L, NO. 3. — 1908. 18 



A CONTRIBUTION TO THE COMPARATIVE PHYSIOLOGY OF THE 
PITUITARY BODY. By P. T. Herring. (From the Physiology 
Department, University of Edinburgh.) (With one Plate and eight 
figures in the text.) 

(Received for publication 21st July 1908.) 

The researches of Oliver and Schafer (7), Howell (4), and others have 
demonstrated the existence in the mammalian pituitary body of active prin- 
ciples which have a specific effect upon the heart and blood-vessels when 
injected intravenously. Howell, moreover, pointed out that it is the 
posterior lobe alone which possesses this property. Magnus and Schafer 
(5), and more recently Schafer and Herring (10), showed that extracts of 
the posterior lobe have the additional characteristic of producing kidney 
dilatation and diuresis when injected. Their observations were confined to 
the pituitary body of certain mammals and of the cod, which was found to 
have a similar action. Osborne and Vincent (8) had previously shown 
that extracts of the pituitary body of the cod produce effects upon the heart 
and blood-vessels similar to those of extracts of the mammalian posterior 
lobe. The question as to the origin of the active principles found in the 
posterior lobe of the mammalian pituitary has been discussed in a previous 
paper (3), and the author has given reasons which appear to him to support 
the view that they are derived from the cells of the pars intermedia, which 
in the mammalian pituitary form so close an investment over the nervous 
tissue of the posterior lobe. It was thought that an examination of the 
structure of the pituitary body of other classes of vertebrates, combined 
with an experimental investigation of the action of extracts of the different 
parts of each, might furnish some interesting facts bearing upon the phj^si- 
ology of the pituitary body, and at the same time throw light upon the 
mode of origin of the active material. 

The difficulty of obtaining sufficient material for extracts has so far 
prevented an investigation of the pituitary bodies of reptiles, amphibians, 
and the lower orders of fishes. The present paper is confined to a descrip- 
tion of the general structure of the pituitary body, and the physiological 
action of its extracts, in birds and in bony and cartilaginous fishes. 

Methods. 
The pituitary bodies of the types examined were fixed for histological 
examination in Flemming's ffuid. Sections were cut by the paraffin method 



262 Herring 

in the sagittal plane and mounted serially. In all cases a sufficiently large 
portion of brain was included to display the immediate relationship which 
exists between the brain and the pituitary body, a precaution which is of 
importance. 

For the experimental part of the research, extracts of the various 
structures revealed by the histological investigation were prepared, the 
lobes of the pituitary being carefully separated from one another, finely 
minced, and boiled in Ringer's fluid. The animals experimented upon 
were cats, which were anaesthetised with a mixture of chloroform and 
alcohol. After a tracheal tube had been introduced, anaesthesia was continued 
by the administration of the same mixture through Brodie's apparatus, 
with artificial respiration. Injections of the extracts were made through a 
tube inserted in the external jugular vein. Blood-pressure was recorded by 
means of a cannula in the carotid artery. The left kidney was placed in a 
brass oncometer ; its movements were registered by a piston recorder. A 
tube tied into the bladder drained away the urine, which was allowed to 
fall drop by drop upon a recorder, an electrical signal marking on the paper 
the moment of the falling of each drop. 



The Pituitary Body of Birds. 
Histological Features. 

The type of bird's pituitary investigated has been that of the common 
fowl, Gallus domesticus. The pituitary body of the adult fowl bears certain 
general resemblances to that of mammals. It has two well-defined lobes — 
an anterior or glandular, and a posterior or nervous. The epithelial cleft, 
which is so prominent a feature of certain mammalian pituitaries, e.g. 
those of the dog and cat, is absent from all the specimens of fowl's pituitary 
that I have examined. The anterior lobe is a compact cylindrical body 
with its long axis in an antero-posterior direction, deeply embedded in the 
sella turcica. Large blood-vessels enter it at its lower posterior margin 
and are a prominent feature in the initial dissection. The lobe itself is 
very vascular, and contains large blood-channels running between solid 
columns of cells. The cells are for the most part small and finely granular ; 
larger cells containing granules which stain more deeply are occasionally 
met with, but do not resemble the large deeply staining cells which are so 
characteristic of the anterior lobe of the mammalian pituitary. The cells 
have a close resemblance to those of the mammalian parathyroid. The lobe 
is as a rule well defined, but in its upper part strands of cells frequently 
pass towards the neck of the posterior lobe and are continuous with narrow 
columns of cells which encircle this and spread over the adjacent brain- 
tissue. Fig. 1 (of Plate) shows the general relationship of anterior to 
posterior lobe and adjacent brain. It is taken from a median sagittal section 
of a fowl's pituitary body. 



The Comparative Physiology of the Pituitary Body 263 

The posterior lobe is smaller than the anterior, and overlaps it slightly 
behind. It is hollow, and its cavity is continuous through a narrow neck 
with the third ventricle of the brain. The lobe is occasionally much con- 
voluted, and its cavity appears at several points in the same section. Its 
wall is never very thick, and seems to consist chiefly of long ependyma 
cells, true nerve-cells being absent from it. Colloid bodies are not infre- 
quently present, and the cavity often contains much debris and occasionally 
rounded clumps of what resemble epithelial cells. In the extension of its 
cavity by recesses and the convolutions of its wall the posterior lobe sug- 
gests a glandular structure opening into the third ventricle. Like the 
posterior lobe of the mammalian pituitary, that of the fowl possesses an in- 
complete covering of epithelial cells, which are constantly found in certain 
positions. They resemble in structure and in their relationship to nervous 
tissue the cells of the pars intermedia of the mammalian pituitary, and are 
probably to be regarded as having the same significance. These cells form 
layers closely investing the nervous substance of the neck of the posterior 
lobe, and extending forwards between the anterior lamina of the neck and 
the optic chiasma (fig. 1 of Plate). The layers are few in number, and well 
supplied by blood-vessels ; in fact, the cells often appear to have extended 
along the sheaths of the blood-vessels. They spread around the neck of 
the posterior lobe and for a considerable distance backwards over the thin 
lamella forming the lower wall of the third ventricle. The body of the 
posterior lobe lies behind, directly upon the anterior lobe, but is readily 
separated from it. No epithelial cells are seen on its posterior and upper 
surface, but they are often found laterally, and extend with the blood- ve.ssels 
into the spaces between the folds of the lobe. In the fowl, therefore, the 
cells of the pars intermedia come into close contact with the nervous tissue 
of the posterior lobe, but are aggregated for the most part in the neighbour- 
hood of its neck and on the thin lamina of nervous tissue forming the floor 
of the third ventricle. It is easy to separate the anterior lobe from the 
posterior, but impossible to remove the nervous tissue of the posterior lobe 
without at the same time including epithelial cells of the pars intermedia 
and their products. 

B. Haller (2) studied tlie pituitary of Gallus doraesticus and describes 
two portions in the anterior lobe, one of which, the superior, is closely 
applied to the infundibulum. The other portion, or anterior lobe proper, 
Haller believes to be tubular, and to constitute a gland whose secretion is 
poured into the subdural space by a small mesial opening. Haller noted 
the diverticula in the posterior lobe, and states that the arrangement met 
within it gives the lobe a glandular appearance. 

Sterzi (11) examined the pituitary of several species of birds, and 
describes a division of the anterior lobe into two parts, one of which is 
made up of chromophobe cells and nearly surrounds the posterior lobe, 
while the other, more massive, consists of chromophil cells. 

Gentes (1) also describes two segments in the anterior lobe. One of 



264 ' Herring 

these he designates as the " segment juxta-nerveux," consisting of a few 
layers of cells which have little affinity for stains, and are applied to the 
posterior lobe in the middle line only. The other segment is formed by 
the anterior lobe proper, and consists of chromophil cells. Both Sterzi 
and Gentes deny the tubular character of the anterior lobe assigned to it 
by Haller. Gentes found a small cleft — the remains of the sac from 
which the epithelium of the pituitary is developed— in a young duck. 

The " segment juxta-nerveux " of Gentes and the chromophobe cells of 
Sterzi agree in most respects with the cells which are here described as 
belonging to the pars intermedia, but in all the specimens I have examined 
they have a more extensive disposition than is assigned to them by Gentes. 
The cells of the anterior lobe may be designated chromophil, but they have 
not the remarkable affinity for stains which is possessed by the larger cells 
of the anterior lobe of the mammalian pituitary. They certainly do differ 
from the cells of the pars intermedia in staining property, and occasionally 
deeply staining granular cells are found among them. 

Histological evidence points to the anterior lobe of the fowl's pituitary 
being, like the mammalian anterior lobe, a gland which secretes directly 
into the blood-vessels. The posterior lobe has an incomplete covering of 
cells which are comparable with the cells of the mammalian pars intermedia 
and are chiefly aggregated around its neck, as in some types of mammals. 
The nervous tissue of the posterior lobe, with its epithelial investment, may 
be regarded as forming a distinct organ which has probably a similar 
function to that exercised by the posterior lobe of the mammalian pituitary. 
It also resembles the mammalian posterior lobe in the occurrence within its 
nervous substance of colloid or hyaline bodies. The colloid is, however, 
confined to the nervous tissue of the lobe, and I have not seen it in or 
between the cells of the pars intermedia. No colloid is found in the 
anterior lobe proper. 



Physiological Action of Extracts of the Lobes of the 
Fowl's Pituitary. 

Anterior Lobe. 

Extracts of the anterior lobe of the pituitary body of the fowl, when 
injected intravenously, have little effect upon the blood-pressure, kidney 
volume, and urine secretion. There is no change in the force and frequency 
of the heart-beats ; the blood-pressure may show a very slight rise, as in 
fig. 2, but is not much altered. Sometimes the pressure falls slightly and 
quickly recovers, but I have not seen a marked fall after injection of 
extracts of the fowl's anterior lobe. The kidney volume increases a little, 
but not more than it does after the injection of a similar amount of 
Ringer's fluid alone. 

The secretion of urine shows no change. Fig. 1 is a typical tracing of 



The Comparative Pliysiolof^y of the Pituitary Body 



265 



the effects of the injection of 5 c.c. of an extract of anterior lobe in Ringer's 
solution into the blood-vessels of a cat. The anterior lobe of the fowl's 
pituitary does not, therefore, contain any active principles exerting an 
immediate physiological effect upon blood -pressure, kidney volume, or 
secretion of urine. In this respect it resembles the anterior lobe of the 
mammalian pituitary body. 

Posterior Lobe. 

The posterior lobe is readily separated from the anterior, and yields 
a greyish gelatinous material which dissolves to a certain extent in 
Ringer's solution. When boiled, filtered, and injected intra venouslj^ such 
extracts produce innnediate and well-marked effects. The blood-pressure 
begins to rise soon after the injection, the heart beats more rapidly, and 





Fig. 1. — Ett'ect of injection into jugular vein of a cat of 5 c.c. of an extract of anterior lobe 
of the fowl's pituitary in Ringer's fluid. 

a, blood-pressure ; k, kidney oncograph ; u, urine secretion (drops) ; t, time in 5 sec. intervals ; »■, signal. 
In this and subseguent tracings the line t represents the zero of blood-pressure. 



the large respiratory waves, when present, are abolished or very much 
diminished in size. The rise in blood-pressure occurs slowly and attains its 
maximum about two minutes after the injection. The rise is not a large one, 
but continues for some time and then gradually falls to normal. The respira- 
tory movements, in spite of the continued supply of air containing the anes- 
thetic from the air-pump, are sometimes affected. Soon after injection the 
respiration is increased or inhibited for a short time, and then resumed as 
before. The kidney volume shows a slight initial increase, followed by a 
slow and gradual expansion, which, after a 5 c.c. dose (8 glands in 40 c.c. 
Ringer), attains its maximum in about fifteen minutes, and then falls gradu- 
ally to what it was before the injection, the whole phase lasting about half 
an hour. The secretion of the urine increases with the expansion of the 
kidney, a latent period of one to two minutes usually elapsing before the 
increase begins. 

The increase of urine is ver}^ pronounced. In the example of which 
fig. 2 is a tracing, the increase is from 12 drops in five minutes before the 



266 



Herring 




g 3 

^ g 
o 



^1 

o 3 






The Comparative Physiology of the Pituitary Body 267 




268 ' Herring 

injection to 42 drops in five minutes afterwards. Where the secretion is 
very slow to begin with, the subsequent increase may be even more marked. 
The secretion is independent of the increase of blood-pressure, as was noted 
by Schafer and Herring in the case of extracts of the mammalian 
posterior lobe ; it is, however, related to the expansion of the kidney, and 
de;reases when that begins to pass off. 

A subsequent dose of the extract in the same animal, if administered 
after the kidney and urine effect have passed off, is followed by a repetition 
of the same changes, although there may be an initial fall of blood-pressure, 
followed by a slow rise (fig. 3). 

Extracts of the posterior lobe of the fowl's pituitary have, therefore, an 
effect on blood-pressure, kidney volume, and urine secretion which is very 
similar to that produced by extracts of the posterior lobe of the mammalian 
pituitary. It is impossible to determine whether the active principles in 
the posterior lobe of the bird's pituitary are products of the epithelial cells 
of the pars intermedia or are formed solely in the nervous substance. The 
large preponderance of the latter in the bird might be considered as an 
argument in favour of their nervous derivation ; but, on the other hand, if 
the cells of the pars intermedia pour their secretion into the pars nervosa 
of the lobe, it may accumulate there in larger quantities. There is evidence 
in the mammalian pituitary that the secretion is emptied into the third 
ventricle of the brain, and is furnished by the cells of the pars intermedia. 
The posterior lobe of the bird's pituitary is so constituted that a similar 
process may quite well be the normal one in it also. 



The Pituitary Body of Teleosts. 
Histological Features. 

The pituitary body of the cod — Gadus morrhua — is taken as the type. 
In this fish the pituitary is a prominent organ lying in front of and below 
the lobi inferiores. The infundibular region is complicated by the presence 
of a saccus vasculosus, which lies immediately behind the pituitary, between 
the two large lobi inferiores. The pituitary body, although forming a single 
organ, is seen to be composed of two different kinds of tissue, an anterior 
portion, reddish or white according to the amount of blood in it, and 
a posterior part, greyish in appearance. The two portions are directly 
continuous with one another, and the line of division between them can only 
be recognised by the change of colour in passing from one to the other. 
On section, the pituitary is found to be a solid organ, and to resemble in 
general structure the pituitary of mammals and birds ; it differs from these, 
however, in the arrangement of its parts. 

In a median sagittal section (fig. 2 of Plate) the general relationship of the 
different parts is readily appreciated. The pituitary is composed of three 
varieties of tissue, two of which are epithelial and the third nervous, the 



The Comparative Physiology of the Pituitary Body 269 

latter being comparative!}^ small in amount. In the anterior part of the 
pituitary a somewhat quadrilateral or wedge-shaped mass is characterised 
by the large and deeply staining cells of which it is composed. These cells 
are almost identical in appearance with the cells found in the anterior lobe 
of the mammalian pituitary, and the portion containing them is probably 
the equivalent of the true anterior lobe. The cells are arranged in columns 
with blood-channels between. There is no trace of tubules, and nothing 
to support Haller's contention that the anterior lobe of the teleostean 
pituitary is a tubular gland secreting into the subdural space. This portion 
of the pituitary of the cod corresponds with that described by Sterzi in 
other bony fishes as the chromophil segment. Gentes also noted the deeply 
staining cells met with in certain positions in the teleostean pituitary, and 
showed that they vary in situation and extent in different species. The 
chromophil portion is aggregated in the cod's pituitary in the position 
indicated in fig. 2, c, of Plate. It may be regarded as constituting the true 
anterior lobe, and its similarity to the anterior lobe of the mammalian 
pituitary suggests that it has a like function. 

The other epithelial constituent of the cod's pituitary is widely dis- 
tributed in the form of small round cells which have little affinity for 
stains. They surround and invade the nervous tissue of the pituitary, and 
resemble in this respect the cells of the pars intermedia of the mammalian 
organ. This part of the gland was described by Sterzi as the "chromo- 
phobe " portion, and there is little doubt that it corresponds with the pars 
intermedia of mammals and birds. Gentes found it in the types he 
examined, and states that it surrounds and passes between projections of 
the nervous substance of the infundibular lobe. The pars intermedia in 
the cod is divided into two main portions, which are continuous with, and 
separated from one another by, the true anterior lobe. The part which 
lies in front of the chromophil segment consists of a mass of small cells 
among which fibres of the nervous substance penetrate. The latter 
increases in amount towards the junction of the pituitary with the brain 
substance behind the optic nerves. The thin lamina of nervous tissue 
connecting the pituitary with the brain in this situation is called by 
Gentes the lamina post-optica. The main mass of the pars intermedia lies 
behind the chromophil portion and makes up the greater part of the lobe. 
On the surface of the pituitary the epithelial cells form a thick mass and 
pass deeply inwards among the fibres of the nervous portion. 

The pars nervosa of the cod's pituitary is small in amount, and appears 
to be composed of neuroglia and ependyma cells, without any true nerve 
cells. It is continuous with the brain in front by the lamina post-optica 
or anterior lamina, and at the sides by lateral laTninae. A thin layer of 
nervous tissue separates the chromophil substance from the cavity of the 
infundibulum. Behind and in the middle line the pituitary is continued 
into the wall of the saccus vasculosus. The nervous substance of the 
pituitary closely resembles in structure the pars nervosa of the mammalian 



270 ' Herring 

pituitary. It is freely invaded by cells of the pars intermedia — more so, 
indeed, than is the case in mammals. It contains, moreover, the colloid or 
liyaline bodies of mammals and birds, and like them it encloses an infundi- 
bular cavity which communicates with the ventricles of the brain. The 
pituitary of the cod furnishes another example of a brain gland similar 
in its essential structure and relationships with the brain to the pituitary 
of mannnals and birds. Pars intermedia — chromophobe portion of Sterzi — 
and pars nervosa make up a structure strictly comparable to the posterior 
lobe of mammalian and avine pituitaries. In the cod there is no epithelial 
cleft, and anterior and posterior lobes are fused together. The fusion is in 
some cases even more complete than is indicated in fig. 2 of Plate, for it not 
infrequently happens that some of the chromophil cells of the anterior lobe 
are found among the cells of the pars intermedia, and cells of the latter 
occur in the true anterior lobe. It is almost impossible, for this reason, to 
separate one portion from another exactly ; but the difference in colour of 
the two parts is, as a rule, sufficiently obvious to enable one to divide them 
for the purpose of making extracts. 

The saccus vasculosus of the cod is single and placed in the middle line. 
According to Gentes, the saccus vasculosus varies considerably in size in 
different species of teleosts, and may, indeed, be absent altogether, or present 
only in a rudimentary state. In the cod it is well developed and forms a 
wide-mouthed sac opening into the infundibulum immediately behind the 
pituitary recess. It is lined by a single layer of columnar epithelium resting 
upon a basement membrane. Numerous blood-vessels reach it in the 
middle line in the interval between the two large lobi inferiores. The 
columnar cells are large, with a nucleus in each situated near the basement 
membrane, the part of the cell next the lumen of the sac being clear. The 
epithelium is thrown into numerous folds which are suggestive of an in- 
crease of surface for secretory purposes. The arrangement and structure 
of the saccus vasculosus is such as indicates that it is a gland which secretes 
into the ventricles of the brain, Gentes believes that it is to be looked 
upon as a ventral choroid plexus, and that its function is to help in the 
formation of the cerebro-spinal fluid. It was called an infundibular gland 
by Rabl-Riickhard (9). The saccus vasculosus of the cod is attached to 
the brain behind, and its epithelium is continued for a short distance over 
the ventricular surface. The brain-tissue above it is remarkable for the 
large ependyma cells which line its internal surface. 



Physiological Action of Extracts of the Lobes of the 
Pituitary and of the Saccus Vasculosus of the Cod. 

Anterior Lobe. 

Extracts of the anterior lobe proper — chromophil portion of Sterzi and 
Gentes — have little innnediate physiological effect (fig. 4). The blood- 



The Comparative Physiology of the Pituitary Body 



271 




272 



Herrinir 



pressure may show a temporary slight fall or remain unaltered. The 
frequency and force of the heart-beat are unaffected. 

The kidney frequently shows a slight expansion, but not a continued 
one. The secretion of urine is unaltered or very slightly increased. The 
difficulty of isolating completely the proper tissue of the anterior lobe from 
the elements of the posterior make it probable that any effect obtained b}^ 
the injection of its extracts is brought about by the inclusion of a little of 
the posterior lobe. If the dissection be so made as to avoid the junction 
of the two lobes, extracts of the chromophil portion have practically no 
action. It seems, therefore, that the anterior lobe or chromophil segment 
aorees in the inactivity of its extracts as well as in its structure with the 
anterior lobes of the pituitary of mammals and birds. 



Posterior Lobe. 

The general effect of extracts of the posterior lobe is similar to that 
brought about by extracts of the mammalian and avine posterior lobes. 




Fig. 5. — Effect of injection into the jugular of a cat of 5 c.c. of an extract of the posterior lobe- 
pars intermedia and pars nervosa — of the cod's {lituitarj-. (12 glands in 40 c.c. Ringer.) 
The arrow on the kidney oncograph indicates an artificial lowering of the writing point. Notice the rise 
of blood-pressure and rapid expansion of liidney, as well as the well-marked diuresis. 

The blood-pressure is almost immediately affected. A considerable rise may 
take place, preceded sometimes by a slight fall, or the increase of blood- 
pressure may be only trivial. The same extract frequently produces 
different results when injected into different cats. In one animal the rise 
of blood-pressure may be marked, in another very slight, and that notwith- 
standing that no previous injections of any kind have been made in these 
animals. Howell pointed out that a repeated dose of pituitary extract 
does not produce the same results on blood-pressure as are brought about 
by the ffrst injection, and this observation holds true of extracts of the 
posterior lobe of the pituitary of avine and teleostean pituitaries as well. 
The immunity conferred by a first dose does not last very long, and varies 
with the amount and strength of the injection given ; but, in order to obtain 



The Comparative Physiology of the Pituitary Body 273 

the typical effect, the first injection of extract of posterior lobe can alone 
be relied upon. Subsequent doses, unless delayed for half an hour, an hour, 
or longer, according to the amount and strength of the first injection, are 
followed by a fall of blood-pressure. The same is not the case with regard 
to the effect upon kidney volume and secretion of urine. 

The force and frequenc}^ of the heart-beat are scarcely affected, but 
irregularities in the pressure-tracing due to inhibition of the heart are often 
abolished for a time. 

The kidney expands almost immediately after the injection, and this 
expansion may be rapid and very considerable, as shown in fig. 5. As was 
the case after injections of an extract of the posterior lobe of the avine 
pituitary, the expansion of the kidney may last for twenty minutes or 
longer and then gradually pass off. The amount of urine secreted begins 
to increase with the expansion of the kidney, and the increase may be, and 
usually is, very considerable. In the experiment, of which fig. 5 shows 
part of the tracing, the increase of urine was from 6 drops in five minutes 
before the injection to 31 drops in five minutes afterwards, i.e. an increase 
of five times the amount in a given time. Thei-e may be a delay of several 
minutes after the injection before any increase of urine is detected, and the 
same extract has difterent effects in this respect in dift'erent cats. 

The posterior lobe of the pituitary of the cod has, therefore, an action 
on blood-pressure, kidney volume, and urinary secretion similar to that of 
the posterior lobe of the mammalian and avine pituitaries. 

Extracts of the posterior lobe of the pituitary of other teleosts, e.g. the 
ling (Molva vulgaris) and the John Dory (Zeus faber), have been tried and 
give the same results. One may conclude, therefore, that the posterior lobe 
of the teleostean pitv^itary, corresponding as it does in structure and in the 
action of its extracts with the posterior lobe of the mammalian and avine 
pituitary, has a like function. In the case of the teleost the cells of the 
pars intermedia predominate in the posterior lobe and are inseparable from 
the pars nervosa, so that one cannot determine which produces the active 
material. It seems probable, indeed, that both are concerned ; for, wherever 
cells of pars intermedia — chromophobe cells of Sterzi — are bound up with 
pars nervosa, extracts of the resulting tissue produce the effects on blood- 
pressure, kidney volume, and urine secretion which have been associated 
with extracts of the posterior lobe of the mannnalian pituitary. 

Saccus Vasculosus. 

Extracts of the saccus vasculosus are practically inactive. There is no 
efi'ect on blood-pressure, and, although there may be some expansion of tiie 
kidney, the increase of urine, if any, is very slight (fig. (i). Almost identical 
efiects are produced by rapid injection of a similar amount of Ringer's fluid. 

It is of interest to note the observation of Gentes, that in difierent 
species of teleosts the saccus vasculosus varies greatly in size and may even 



274 



Herring 



be absent. G antes further remarks that the presence or absence of the 
saccus vasculosus brings about Jno modification of the nervous lobe. It is 
very probable that the saccus vasculosus has, as Gentes believes, a function 
similar to that of a choroid plexus. 

The Pituitaky Body of Elasmobraxchs. 

Histological Features. 

The pituitary body of the skate — Raja batis — is taken as the type. In 
the skate the pituitary is a long, club-shaped body which lies for the most 
part behind the small lobi inferiores. Its anterior extremity is thin, and 
stretches forward to the optic chiasma. Close to it is the large saccus vas- 
culosus which is bilobed. The lobes of the saccus vasculosus appear to arise 
just above the anterior part of the pituitary by a common origin with it. 




Fig. 6. — Effect of the injection into the jugular of a cat of fi c.c. of an extract of the 

saccus vasculosus of the cod. (12 glands in 20 c.c. Ringer.) 

The effect upon the kidney is no greater than that produced by rapid injection of 5 c.c. Ringer alone. 

Each lobe passes backwards and outwards, and the body of the pituitary 
lies between them. A fine prolongation of connective tissue passes from 
the under surface of the pituitary body into the cartilage of the floor of the 
cranium, binding it closely down to the latter. This is the remnant of the 
neck of Rathke's pouch, from which the pituitary is developed, and was de- 
scribed by Miclucho-Maclay (6) in the shark. In the skate all connection 
between buccal mucosa and pituitary body is lost, but a string of connective 
tissue persists in the cartilage. It is advisable for this reason, in removing 
the pituitary body, to expose it by cutting away the cartilage of the floor 
of the cranial cavity. 

On making a sagittal section through the pituitary, it is seen to extend 
for a long distance backwards from the optic chiasma, and to be quite dif- 
ferent in structure from the pituitary bodies of mammals, birds, and teleosts. 
The main body of the organ lies posteriorly, and is the part which Haller 
designates as the head of the pituitary. It at first sight appears to be com- 
posed of tubules lined by large columnar cells, but on careful examination 



The Comparative Physiology of tlie Pituitary Body 275 

the tubules are found to consist of columns of cells surrounding blood spaces. 
The lumen is a blood channel. This feature has been emphasised by 
Gentes, who states that the elasmobranch pituitary is a typical example of 
a gland whose secretion is poured directly into the blood-vessels. The 
epithelial cells surrounding the blood channels are columnar, with nuclei 
situated at their bases, the part of the cell bordering on the blood-vessel 
being clear. Outside the columnar cells and separating the vascular tubules 
from one another is a small amount of what appears to be a very vascular 
connective tissue. The vascular tubules and this connective tissue make 
up the body of the lobe. There are no deeply staining granular cells 
resembling the chromophil cells of the anterior lobe of the pituitary of 
mammals and teleosts, nor are there any cells exactly resembling those 
of the pars intermedia. 

The anterior extremity of the skate's pituitary consists of a comparatively 
thin prolongation of epithelium enclosing a cavity (tig. 3, b, of Plate) which 
is stated by Gentes to be the remains of the cavity of the original sac from 
which the pituitary develops. It is lined by columnar epithelium very 
similar in appearance to that surrounding the blood-vessels in the body 
of the lobe. The cavity appears to be completely closed, and is much 
sacculated by convolutions of its wall. Outside this sac are numerous 
blood-vessels. 

There is no differentiation of the pituitary gland of the skate into 
anterior and posterior lobes. An infundibular cavity is present which 
runs backwards and downwards to the body of the pituitary. It does 
not penetrate into the pituitary, but ends in the middle line, as shown in 
the figure. When followed laterally, however, the infundibular cavity is 
found to pass on either side into the saccus vasculosus. The nearest 
approach to anything resembling a posterior lobe is seen in the thin lamina 
of nervous tissue which bounds the infundibular cavity and passes into the 
tissue of the pituitary. Whether the line vascular tissue that lies between 
the epithelial tubules in the body of the pituitary is of nervous origin or 
not is uncertain, but it is continuous with the lamina of nerve tissue that 
forms the lower wall of the infundibular cavity. If this tissue really be- 
longs to the posterior lobe, then we have in the skate a very complete inter- 
mixture of elements derived from the brain and from buccal epithelium. 
Gentes states that the posterior lobe is completely absent in elasmobranchs. 
It is probable, however, that some representative of the processus in- 
fundibuli exists even in adult life, and that the general plan of development 
of the pituitary in elasmobranchs does not differ from that of other verte- 
brates. The nerve tissue in the anterior wall of the infundibulum must be 
regarded as a representative, in part at least, of the posterior lobe. But 
its constitution is altered by the large developnient of the saccus vas- 
culosus, and the extension of the epithelium of the saccus over the lining 
wall of the infundibulum. 

The saccus vasculosus is extivmrly well developed in the skate, and is a 

VOL. I., \(). :5. — 1908. l'> 



276 Herring 

prominent bilobed organ with a deep red colour due to the amount of blood 
contained in its vessels. Its wall is thin and convoluted, and consists of 
one or more layers of epithelial cells outside which are numerous and large 
thin-walled blood-vessels. The epithelium is continued forwards into the 
infundibulum, and in median sagittal section the common opening of the 
two sacs is seen lying above and in front of the body of the pituitary. 
The pituitary body of the skate is, then, an example of a type which is 
entirely different from those of mammals, birds, and bony fishes. There is 
no differentiation into anterior and posterior lobes, and the characteristics 
of the cell elements of these are missing. The pituitary body itself fur- 
nishes, as Gentes says, a schematic type of gland secreting into blood- 
vessels. The posterior lobe is not distinct, but is represented to some 
extent ; its infundibular surface appears to be largely devoted to the same 
purposes as the saccus vasculosus. There are no colloid bodies present in 
the thin layer of nervous substance, and no cells clearly resembling those 
of the pars intermedia. 

Physiological Action of Extracts of the Pituitary and 
OF THE Saccus Vasculosus of the Skate. 

The Pituitary Body. 

Extracts of the whole pituitary body of the skate have little effect upon 
blood-pressure, kidney volume, or secretion of urine. Strong extracts pro- 
duce a temporary fall of blood-pressure, but not a marked one (fig. 7). 
Kidney volume is slightly increased, but there is no continuous expansion, 
and the effect is merely that of the injection of Ringer's fluid. 

Kidney secretion is unaltered or very slightly increased. It is doubtful 
if this increase is a specific one : it may be solely caused by the rapid 
injection of so much fluid. The pituitary body of the skate apparently 
contains none of the active principles which are found in the pars nervosa 
and pars intermedia of mammals, birds, and teleosts. If any of these are 
present, it is only in very small amount ; there is no clear evidence of a 
histological character for the presence of these active principles, and it is 
probable that they do not exist in the elasmobranch pituitary. 

The Saccus Vasculosus. 

Extracts of the saccus vasculosus, even when very concentrated, have 
little effect. There may be, as in fig. 8, a temporary fall of blood-pressure, 
but with weaker solutions there is no change. 

Kidney volume and urinary secretion may show a slight temporar}^ 
increase, but, as was the case with extracts of the saccus vasculosus of the 
cod, it is not a specific effect, but merely the result of the rapid injection 
of so much Ringer's fluid. 

Extracts of portions of the brain adjacent to the pituitary body and 



The Compai-ative Physiology of the Pituitary Body 



277 



3 o 






O -3 



278 



Herring 



saccus vasculosus, jiiid extracts of the lobi inferiores, bring about a large 
fall of blood-pressure ; the effects are similar to those seen after injection 
of extracts of central nervous system in general. There seems to be 
nothing in the infundibular region of the brain of the skate which is com- 
parable in the action of its extract with the posterioi- lobe of the pituitary 
body of nuuninals, birds, and bony fishes. 



Conclusions and Summary. 

In mammals, birds, and bony tishes the pituitary body consists of two 
lobes, an anterior or epithelial which has the structure of a gland secreting 
into blood-vessels, and a posterior composed of nervous tissue more or less 
surrounded and invaded by epithelial cells of the pars intermedia. The 
posterior lobe may also furnish secretion into blood-vessels, but its arrange- 




FiG. 8. — Effect of the injection into the jugular of a cat of 5 c.c. of an extract of tlie 

saccus vasculosus of the skate. (5 glands in 20 c.c. llinger.) 

There is a transient fall of Vilood-pressure and slight expansion of kidney, but no diuresis. 

ment and histological features suggest a gland which pours its products 
into the infundibulum, and so into the ventricles of the brain. It may, 
therefore, be regarded, in part at least, as a special brain gland. 

Extracts of the anterior lobe have no immediate physiological action 
when injected into the blood-vessels. 

Extracts of the posterior lobe of birds and bony fishes have an action 
similar to extracts of the mammalian posterior lobe, bringing about a rise 
of blood-pressure, expansion of the kidney, and an increase in the secretion 
of urine. The tissue in which the active principles giving this result are 
found, contains, when examined histologically, bodies of a colloid nature 
such as have already been described in mammals in a previous paper. 
Whether this colloid contains the above-mentioned active principles or not, 
is undecided ; it may possibly be the expression of some other function. 
The close relationship which exists between pars nervosa and pars inter- 
media of the posterior lobe renders it probable that the active principles, 



The Comparative Phy.siology of tlie Pituitary Body 279 

and especially the colloid bodies, are furnished by the epithelial cells, but 
it is possible that the ependyrna cells have also a secretory function. 

The pituitary body of elasniobranchs differs widely in structure 
from that of the other classes considered. It is a gland which apparently 
secretes directly into the blood-vessels, but it contains none of the deeply 
staining (chromophil) cells which are characteristic of the anterior lobe of 
mammals and teleosts. Its posterior lobe is absent or merely rudimentary. 

Extracts of the pituitary body of elasmobranchs have no immediate 
physiological activity. 

The saccus vasculosus secretes its products into the ventricles of the 
brain. Its extracts are inactive, and it is probably an auxiliary to the 
choroid plexus, aiding in the production of the large amount of cerebro- 
spinal fluid which is found in fishes. 

I have to thank Mr Richard Muir for the care with which he has 
executed the accompanying illustrations. The expenses incurred have been 
assisted by a grant from the Carnegie fund for research-work. 



LITERATURE REFERRED TO IN THE TEXT. 

(1) Gentes, " Recherches sur I'hypophyse et le sac vasculaire des vertebres," 
Soc. scientif. d'Arcachon, Station biologique, Travaux des laboratoires, pp. 129-275, 
fasc. i. Bordeaux, 1907. 

(2) Hallek, B., " Untersuchungeu iiber die Hypophyse und die Infundibular- 
organe," Morpholog. Jahrhuch, Bd. xxv., S. 31, 1896. 

(3) Herring, "The Histological Appearances of the Mammalian Pituitary 
Body," Quart. Journ. of Exper. Physiol, vol. i., No. 2, p. 121, 1908. 

(4) Howell, "Tlie Physiological Effects of Extracts of the Hypophysis Cerebri 
and Infundibular Body," Amer. Journ. of Exper. Med., vol. iii., p. 245, 1898. 

(5) Magnus and Schafer, "The Action of Pituitary Extracts upon the 
Kidney," Proc. Physiol. Soc, July 20, 1901. 

(6) Miclucho-Maclay, "Beitrag zur vergleichende Anutomie des Gehirns," 
Jenaische Zeitschr. f. Naturwisscnscliaft, S. 557, 1868. 

(7) Oliver and Schafer, "On the Physiological Action of ILxtracts of the 
Pituitary Body and certain other Glandular Organs," Journ. of Physiol., vol. xviii., 
p. 277, 1895. 

(8) Osborne and Vincent, "A Contribution to the Study of the Pituitary 
Body," Brit. Med. Journ., vol. i., p. 502, 1900. 

(9) Rabl-Ruckharu, "Das Grosshirn der Knoohenfisohe und seine Anhangs- 
gebilde," Arch. f. Anat. u. Physiol., Anat. Abth., Jahrgang 1883, S. 317. 

(10) Schafer and Herring, "The Action of Pituitary Extracts upon the 
Kidney," Phil. 'IVans., B., vol. cxcix., p. 1, 1906. 

(11) Sterzi, " Intorno alia struttura dell' ipotisi nei vertel)rati,'' Atti Accad. 
Sc. Veneto-Trentino-Istriana, CI. sc. nat., fis. e. mat., vol. i., p. 72, 1904. Quoted 
from Gentes. 



280 'The Comparative Physiology of the Pituitary Body 



DESCRIPTION OF PLATE. 

Fig. 1. Median sagittal section through the pituitary body of the fowl — Gallus 
domesticus. a, optic chiasma ; h, anterior lobe of the pituitary — pars glandularis ; 
c, cells of tlie pars intermedia ; d, neck of the posterior lobe ; e, cells of the pars 
intermedia lying behind the neck ; /, third ventricle ; g, posterior lobe of the 
pituitary. 

Fig. 2. Median sagittal section through tlie pituitary body and saccus vasculosus 
of the cod — Gadus morrhua. a, optic nerve ; h, anterior part of pars intermedia — 
chromophobe cells of Sterzi; c, pars glandularis or anterior lobe — chromophil cells 
of Sterzi ; d, pars nervosa surrounded by cells of pars intermedia — posterior lobe ; e, 
infundibulum ; /, large ependyma cells ; g, saccus vasculosus ; h, space between lobi 
inferiores occupied by blood-vessels and connective tissue. 

Fig. 3. Median sagittal section through pituitary body of skate — Raja batis. 
a, optic chiasma ; h, anterior part of pituitary enclosing cavity ; c, infundi- 
bulum ; d, opening of saccus vasculosus on either side into infundibulum ; <?, body 
of pituitary. 



Quarterly JounuU vf Experimental I'hysiohKjy, Vol. I., I90S.] 
/> c d e f g 



''^si"-^ 




'■^S^ieg^t^i^* 



Fig. 3. 



P. T. FIkukin(;, " A Contriluitioii 1.. tlic Coinpaiativc riiysi.,lo^ry vf tlio riiuitarv Roc 



THE EFFECTS OF THYROIDECTOMY UPON THE MAMMALIAN 
PITUITARY. PRELIMINARY NOTE. By P. T. Herring. (From 
the Physiology Department, University of Edinburgh.) (With Two 
Plates.) 

(Received for puhlication 2bfh July 1908.) 

The occurrence in the pituitaiy body of a substance i-esembling the colloid 
of the thyroid has long been known. The removal of the thyroids is 
followed in some animals by certain well-marked symptoms ending sooner 
or later in death; in other animals it has no apparent effect. Rogowitsch 
(7) went so far as to state that the pituitary acts vicariously for the 
thyroid, and that in rabbits and other animals which can survive thyroid- 
ectomy the function of the thyroid is taken over and maintained by in- 
creased activity of the pituitary. Where the pituitary is relatively small, 
as in the dog, it fails to do this. Rogowitsch found alterations in the 
pituitaries of thyroidectomised dogs and rabbits of tlie nature of an increase 
of certain elements — " Kernhaufen " — in the glandular or anterior portion. 
He also described the formation of colloid by the chromophil cells of the 
anterior lobe, and its entry from them directly into the blood-vessels or into 
small cysts in the " Markschicht." A further change was seen in an in- 
crease of the number of vacuoles both in the chromophil cells and in the 
" Kernhaufen." H. Stieda (9) described hypertrophy of the anterior lobe 
brought about by an increase in the number of " Hauptzellen " with the 
formation of vacuoles in them. He saw no change in the chromophil cells 
and no formation of colloid. 

Various other observers, Hofmeister (4), Gley (2), Pisenti and Viola 
(()), Schonemann (8), have found changes in the pituitary body consequent 
upon removal or disease of the thyroid. Attention has been confined chiefly 
to alterations in the epithelial constituents of the pituitary body, little 
mention having been made of the condition of the nervous portion. Klebs 
(5) states that he has seen hyaline globules in the blood-vessels of the 
nervous part of the pituitary of sti'umiprivous dogs, and believes that in 
this organ is to be sought the " Ausgangspunkt " of the disturbance. 
Boyce and Beadles (1) found, in cases of myxtTedema, enlargement of the 
pituitary body with increase of colloid in the posterior part of the anterior 
lobe. Large cysts containing colloid were present in this situation which 
they termed the medullary layer. The posterior lobe was atrophied but 



282 ' Herring 

occupied in its upper portion by a ' colloid-like (? plasmatic) material which 
is not bounded by a cell membrane." 

In a previous paper (3), histological evidence has been brought forward 
to prove that the posterior lobe of the pituitary furnishes a secretion which 
passes through the nervous portion to enter the infundibulum and ventricles 
of the brain. The secretion has normally a colloid appearance, and is prob- 
ably a product of the cells of the pars intermedia. The results which 
follow thyroidectoni}" furnish additional and strong evidence in favour of 
this view. 

My observations have up to the present been confined to rabbits, cats, 
and one dog. Three rabbits, of which the thyroids had been completely re- 
moved, were placed at mj'- disposal by Professor S chafer. These animals 
showed no symptoms, and were apparently healthy when killed three months 
after thyroidectomy. The pituitary bodies were removed along with 
part of the brain, fixed in Flemming's fluid, and cut in serial sections. Un- 
fortunately, during the process of removal the posterior lobe of one animal 
was badly injured, and was not available for examination. In the other 
two there is no apparent change in the anterior lobe. Clear cells, deeply 
granular cells, and transitional forms are present, and there is no evidence 
of colloid production by them. 

The cells of the pars intermedia are distributed as usual, but are some- 
what increased in amount, and stain more deeply. Remarkable changes 
are at once apparent in the posterior lobe, and especially in that portion of 
it which lies next to the pars intermedia. Masses of a colloid nature lie 
among the cells and fibres of the pars nervosa (Plate I., fig. 1), and extend 
forwards and upwards into the neck of the lobe right up to the floor of the 
third ventricle. Some of this material is irregular in outline, and appears 
to lie in lymph spaces, but much of it is cellular. The presence of a swollen 
nucleus occupying the centre of a mass is frequently seen, and in some cases 
instead of the hyaline or colloid appearance distinct granules are evident. 
The granular cells and hyaline masses can be traced into the pars inter- 
media, from the cells of which they apparently take origin. 

In the neck of the posterior lobe the colloid is even more plentiful ; 
approaching the floor of the third ventricle it becomes more cellular in 
appeai'ance. Many of the cells in this situation are extremely vacuolated, 
others are granular or filled with an amorphous material. In the infundi- 
bular recess at the floor of the third ventricle are found numerous cells of 
varying size and masses of granular and amorphous material. The cells lie 
free in the cavity of the third ventricle (Plate I., fig. 2), having found their 
way out of the neck of the posterior lobe b}^ passing between the ependyma 
cells. After their escape into the cerebro-spinal fluid they become swollen 
and disintegrate with the production of a granular and amorphous debris. 

In addition to this increased production of colloid, there is a distinct 
budding of the ependj-ma cells which line the ventricle, and a proliferation of 
the adjacent neuroglia. The colloid bodies are not only present in the neck 



The Effects of Thyroidectomy upon the Mammalian Pituitary 283 

of the lobe, but occupy the thin anterior or post-optic lamina, with the under 
surface of which the cells of the tongue-like process of pars intermedia are 
closely associated. Changes in the ependyma and neuroglia cells are most 
noticeable in this region. 

In the cats operated upon nervous symptoms followed within twenty-four 
hours after removal of the thyroids (and parathyroids), and the animals 
were killed from four to six days after the operation. In the pituitaries of 
these animals there is comparatively little change. Colloid is still present 
in the spaces in the pars intermedia' in which it is normally found. Varia- 
tions in the amount of colloid in the pituitary of the healthy animal are 
80 great that no change can be definitely asserted, but the amount of colloid 
in the pituitaries of these recently thyroidectomised animals is not un- 
usually large. The granular bodies and hyaline material in the posterior 
lobe are more plentiful than is normally the case. Accumulations of colloid 
cells are seen in numerous places immediately beneath the ependyma cells 
both in the body and neck of the posterior lobe. They form papular pro- 
jections of the ependymal surface, and the contents are frequently seen 
escaping into the infundibular cavity. Localised proliferations of ependyma 
or neuroglia cells are also found. The changes, as in the rabbit's pituitary, 
are not confined to the body and neck of the posterior lobe, but extend 
into the anterior and posterior laminai. They correspond in extent and 
distribution with the presence of cells of the pars intermedia lying outside 
the nervous part. There is no accumulation of colloid inside the epithelial 
cleft. 

In the only dog as yet operated upon the thyroids (and parathyroids) 
were completely removed, and no symptoms occurred until five days after- 
wards, when the animal had attacks of tetany. It recovered, but tetan}^ 
again developed and extreme weakness came on. The dog was killed nine- 
teen days after the operation. The anterior lobe of the pituitary shows 
no change. There are no cystic accumulations of colloid in the pars inter- 
media, but a great formation of colloid is taking place at the junction 
between pars intermedia and pars nervosa. The nervous substance of the 
posterior lobe is granular in appearance, and contains masses of colloid 
accumulated in certain situations. It is most abundant in the neck of the 
lobe and at the lower end of the infundibular recess. In places the 
epend\nna lining is distended by colloid, some of which bursts through into 
the infundibular cavity. Here again most of the colloid is cellular, and when 
it mixes with the cerebro-spinal fiuid it takes the form of large nucleated 
cells full of granular or amorphous material. 

In the neck of the lobe appearances are seen of which fig. 3. Plate II., 
presents an illustration. Cells of the pars intermedia pass inwards invading 
the nervous substance and frequently accompanying blood-vessels. As they 
approach the infundibular cavity the cells become swollen and look like 
colloid masses. They accumulate below the ependyma and finally pass 
into the cavitv either in a cellular form or as a hvaline debri.s. 



284 . Herring 

In this dog there are also marked changes in the ependyma cells, con- 
sisting chiefly of a budding otl" of small round globules into the infundibulum 
and third ventricle. Fig. 4, Plate II., is a photograph of part of the internal 
surface of the posterior lamina, and shows the formation and liberation 
of these small globular bodies. The ependyma cells of the infundibular 
region are not ciliated, and probably have a secretory power. The com- 
bined products of the cells of the pars intermedia and of the ependyma 
cells are mixed, and form a considerable accumulation in the infundibular 
recess of the third ventricle. The nature and signiticance of this material 
is as yet undetermined ; its formation in thyroidectomised animals appears 
to be an exaggerated condition of a normal process. 

Summary. 

Thyroidectomy is followed by definite histological changes in the 
pituitary body. The anterior lobe is apparently unaffected ; it shows no 
immediate sign of increased activity as far as can be judged from the 
animals examined. Whether or not it undergoes an alteration in animals 
which survive for a long time remains to be determined. 

There is increased activity of the cells of the pars intermedia, and 
probably an increase in the number of these cells in animals which live 
for some time after the operation. In this respect the work of Rogowitsch 
and others is confirmed. 

The most striking changes are manifested in the nervous part of the 
posterior lobe and in the laminae forming the floor of the third ventricle. 
In these situations granular, hyaline, or colloid bodies become very 
numerous. They appear to be, in part at least, of a cellular nature, and to 
find their way between the ependyma cells into the infundibular recess 
and ventricles of the brain. 

There are also alterations in the ependyma and neuroglia cells in the 
same regions. The former appear to have a secretory function and give off 
small clear globular bodies into the infundibular recess and third ventricle. 
There are also localised proliferations of neuroglia. 

The significance of these changes is as yet undetermined. The colloid 
bodies appear to arise from the epithelial cells of the pars intermedia, and 
their extensive production to be an exaggeration of a normal process. Further 
work is being carried on to determine the nature of the colloid of the 
pituitary, its relations to the active principles found in the posterior lobe, 
and its influence in the production or amelioration of the symptoms, which, 
in many animals, follow removal or disease of the thyroid. 

The expenses incurred in this work are assisted by a grant from the 
Carnegie endowment for research purposes. 



The Effects of Thyroidectomy upon the Mammalian Pituitary 285 

LITERATURE REFERRED TO. 

(1) BoYCB and Beadles, "Enlargement of the Hypophysis Cerebri in Myxce- 
dema, with remarks upon the Mainmahan Pituitary," Journ, of Pathol, and Bacteriol., 
vol. i., p. 223, 1893. 

(2) Gley, -'Recherches sur la f(jnction de la glande thyroide," Arch. de. 
physiol. normale et path., p. 311, 1892. 

(3) Herring, "The Histological Appearances of the Mammalian Pituitary 
Body," Quarterly Journ. of Exper. Physiol., vol. i., No. 2, p. 151, 1908. 

(4) HoFMEiSTER, " Zur Physiologie der Schilddriise," Fortschritte der Medicin, 
S. 81, 1892. 

(5) Klebs, Allgemeine Pathologie, Th. ii., S. 433, 1889. 

(6) PiSENTi and Viola, " Beitrag zur normalen und pathologischen Histologic 
der Hypophysis und beziiglich der Verhaltnisse zwischen Hirnanhang und Schild- 
driise," Centralbl. fur die med. Wissensch., S. 450, 1890. 

(7) RoGOWiTSCH, " Die Veranderungen der Hypophyse nach Entfernung der 
Schilddriise," Ziegler's Beitrage zur patholog. Anat., Bd. iv., S. 453, 1889. 

(8) ScHONEMANN, " Hypophysis und Thyreoidea," Virchow's Archiv, Bd. cxxix., 
S. 310, 1892. 

(9) Stieda, H., " Ueher das Verhalten der Hypophyse des Kaninchens nach 
Entfernung der Schilddriise," Ziegler's Beitrage zur patholog. Anat., Bd. vii., 
S. 537, 1890. 



DESCRIPTION OF PLATES. 



Plate I. 



Fig. 1. Median sagittal section through part of posterior lobe of the pituitary of 
a rabbit three months after removal of both thyroids. (Photograph, x 300.) Granu- 
lar and colloid bodies are seen in the nervous portion of the lobe ; some of the cells 
of the pars intermedia appear on the left-hand side. 

Fig. 2. Median sagittal section through infundibular recess and neck of the 
posterior lobe of a rabbit three months after removal of both thyroids. (Photograph, 
X 200.) Granular and colloid bodies in neck of the posterior lobe. Some of these are 
definite cells which pass into the infundibular recess. Budding of ependyma cells is 
seen on the inner surface of anterior lamina on the left-hand side of the photograph. 

Plate II. 

Fig. 3. Median sagittal section through lower part of the neck of the posterior 
lobe of the pituitary of a dog nineteen days after thyroidectomy. (Photograph, 
X 160.) Cells of the pars intermedia seen below are invading the nervous tissue of 
the neck of the posterior lobe. Colloid liodies are streaming inwards and accumu- 
lating beneath the ependyma cells lining the infundibular recess. 

Fig. 4. Median sagittal section through part of the nervous substance of the 
upper part of the neck of the posterior lobe of the pituitary of the same dog. 
(Photograph, x 300.) The active budding of the ependyma cells, and the escape of 
their products into the infundibular recess of the third ventricle, are shown. 



Qninferlii Jinnnifd of ExprrivienUd Phyfiiohi(iy, Vol. /., 1908.'] 



[Plate I. 



M ^ 

'^.r 



^-fe^«^,#v, 



^2i^ 






€ 









..,». 



-.1 






'*' 



.),:« 






^^li^^^^Sl 



Fig. 1. 










r:k?S^*^*^>-^ 






r. T. IIekiiint;, "The Ellects of Thyroidectomy upon the Pituitary Body," 



jU^ 



Q'larUrhj J<:urH'il of E.rprTiT}unt'i! Phii::io!^„jy, V„l. /., 190S.'\ 



[I'lATE II. 






^ .^.r .•>- -. 1.'/' 



.,',>:;|*«»v-n»&^ 




Fjg. 3. 









ii"'' 



P. T. HEKKlNf:, "Tlie, EirectsofTliyioi.l.'otoiiiy upoti tlip Pituitary Body.' 



OBSERVATIONS ON THE NUCLEOLI IN THE CELLS OF HYDRA 
FUSCA. By C. E. Walker and Alice L. Embletox. (From the 
Laboratory of Cytology, University of Liverpool.) (With One Plate.) 

{Received for puhlvat ion, 25th July 1908.) 

Methods and Material. 

Hydra fusca was used throughout in the observations described here. 
The hydrae were fixed with Flemming's fluid (strong formula). Dehydra- 
tion, embedding in paraffin, and the various other processes, were carried 
out according to the strictest cytological methods. 

The following two processes of staining were chiefly employed, and 
only the results obtained by these are described here. 

A. Saturated solution of basic fuchsin in 80 per cent, alcohol, plus 
a few drops of liquor ammon. fort., for about two minutes. Rinse in water 
and 30 per cent, alcohol, and water again, getting rid of part of the red 
stain so as to leave the sections pink. Methylene blue (aqueous solution) 
for about a minute ; rinse in water. Unna's orange tannin until no more 
blue clouds are to be seen. Dehydrate, clear in xylol, mount in Canada 
balsam. 

B. Saturated solution of saflranin in 80 per cent, alcohol, with a few 
drops of liquor ammon. fort., for a quarter of an liour, followed by methy- 
lene bhie and Unna's orange tannin, as in first method. 

Observations. 

The nucleolus dealt with here is the " true nucleolus." It is a body 
generally spherical in shape, occasionally oval, bounded bj'^ a definite 
membrane, the contents being usually homogeneous, or finely granular in 
structure.^ 

The nucleoli in the cells of hydra, both in the ectoderm and in the 
endoderm, are very striking and definite. Generally only one nucleolus is 
to be seen in the nucleus, but in some cells we have found two, or even 
more, of diflerent sizes. While within the nucleus, the nucleoli stain dark 

' See Wilson, "Tlie Cell in Develoi^ineiit and Inheritance," p. 34; Macniillan, 
London and New York, 1904. Walker, "Tlie Essentials of Cytology," pp. 12 and 13; 
Constable, London, 1907. 



288 ' Walker and Einbletoii 

purple with method A, and bright red, sometimes with a blue area in the 
centre, with method B. They may frequently be seen to be budding in a 
manner very like the budding of the nucleoli in nerve-ganglion cells 
described elsewhere.^ 

In both endoderm and ectoderm cells the budding often takes the form 
of a small excrescence which gradually separates from the nucleolus, 
remaining joined to it, however, for a considerable period by a membranous 
process (see tigs. 1, 2, and 8). Sometimes, however, particularly in the 
cells of the endoderm, the nucleoli seem to divide more or less equally from 
the beginning, and it is impossible to describe this form of division as 
budding. Very often there is one large nucleolus to be seen and several 
small ones (fig. 2), but we also find some nuclei containing two or more 
large nucleoli nearly equal in size (fig. 4). Nucleoli may also frequently 
be seen in process of passing through the nuclear membrane. As the 
nucleolus passes through, the nuclear membrane appears to form an 
encircling lip round it, the membrane being re-formed underneath it 
very rapidly (figs. 5 and 6). Thus, when the nuclear membrane has 
re-formed behind the extruded nucleolus, the latter lies in a depression not 
unlike a crater. Almost as soon as the nucleolus has passed through the 
membrane, the dark purple colour it exhibits when stained by method A 
is lost at its periphery, leaving this pink. We thus have a spherical 
nucleolus with a dark purple area in the centre, surrounded by an area 
stained pink (figs. 1, 4, 7, 8, and 9). The purple centre appears to 
decrease rapidly, leaving the nucleolus pale pink throughout. With 
method B the blue or violet area seen in some nucleoli while within the 
nucleus is lost as the former passes into the cytoplasm. In every case the 
nucleolus becomes more and more orange in colour as it leaves the 
neighbourhood of the nucleus. In the cells of the endoderm, nucleoli may 
often be seen in process of division, after they have been extruded from the 
nucleus. It seems that after this, in the endoderm cells, the nucleoli are 
pushed towards the periphery of the cytoplasm. On approaching the 
periphery they lose the pink colour (obtained with method A) and take 
the orange stain, becoming at the same time much less defined, and appear 
eventually to disintegrate altogether. 

We have been unable hitherto to trace the destiny of the nucleoli in 
the cells of the ectoderm after their extrusion. Occasionally the nucleolus 
lies pressed upon the surface of the nuclear membrane after its extrusion 
(fig. 9). Also we have observed the contents of the nucleoli to become 
granular in a few instances. 

The depression, or crater, in the surface of the nuclear membrane 
appears to last for a considerable time, and gives the appearance of 
pseudopodial processes when seen in optical section (figs. 4, 7, 8, and 9). 

• " On the Multiplication and Migration of Nucleoli in Nerve Cells of Mammals," by 
W. Page May and C. E. Walker, this Journal, vol. i.. No. 2, 1908. 



Observations on the Nucleoli in the Cells of Hydra Fusca 289 

The cytoplasm of the endoderm cells appears to be almost structureless, 
while there is genei-ally some structure to be seen in the ectoderm cells. 

Conclusions. 

The phenomenon here recorded seems to be something quite apart from 
anything connected with cell division, whether mitotic or amitotic ; indeed, 
it is only to be observed in cells that are in the vegetative condition. 
The probability that it is intimately connected with, and dependent upon, 
metabolism taking place in the nucleus, is very strong. While it is readily 
observed in almost all the cells of the endoderm, it can be followed only to 
a very limited extent in the cells of the ectoderm. The processes connected 
with digestion are carried out by the cells of the endoderm. It would 
therefore appear probable that the kind of metabolism in the nucleus that 
is connected with digestion produces this phenomenon in a more striking- 
manner, and more frequently, than is the case with the nuclear metabolism 
taking place in the cells of the ectoderm. 

The rapid and very marked change in the staining reaction of the 
nucleolus as it passes into the cytoplasm, suggests that some important 
chemical or physical change takes place in the contents at this time. 
Similar changes have been observed in the nucleoli of nerve-ganglion 
cells.i 

It seems probable that the bodies here described as extruded nucleoli 
have been frequently described as food particles. But we would point out 
that the structures here dealt with are apparently in process of disintegration 
at the periphery of the cell, become more defined in the neighbourhood of 
the nucleus, and are found in their most definite form and in process of 
active multiplication within the nucleus itself. 

1 Page May and Walker, luc. cit. 



DESCRIPTION OF PLATE. 

(In these drawings the cytoplasm of the endoderm cells is, for the sake of con- 
venience, considerably diminished in size in proportion to the nuclei.) 

Fig. 1. Two endoderm cells showing the nucleoli budding. Some nucleoli are to 
be seen in the cytoplasm. 

Fig. 2. Two endoderm cells showing the nucleoli budding. Several nucleoli 
may be seen in the same cell. 

Fig. 3. Four ectoderm cells showing nucleoli budding. 

Fig. 4. An endoderm cell showing two large nucleoli and pseudopodial processes 
in the nucleus, as well as several nucleoli in the cytoplasm. 

Fig. 5. Three endoderm cells showing nucleoli passing out of the nucleus. 
In c a nucleolus which remains within the nucleus is seen to be as large as that 
which is passing out. 



290 Observations on the Nucleoli in the Cells of Hydra Fusca 

Fig. 6. Three ectoderm cells showing nucleoli passing into the cytoplasm. 

Fig. 7. An endoderm cell showing nucleoli lying in the cytoplasm. Some 
have a dark centre ; in others the dark centre has disappeared. The nucleus also 
exhibits pseudopodial processes. 

Fig. 8. An endoderm cell in which a nucleolus is seen pressed against the 
outside of the nuclear membrane. 

Fig. 9. Two endoderm cells, one of them containing a number of nucleoli in 
the cytoplasm. One of these nucleoli is seen to be dividing. 



L'^O 



Qiuirlerly Joi^rn(d of Experimental Physiology, Vol. /., 1D08.] 



1 

I / ^ • a b 

" ' 3 



^ 






f 









# 



y -^ 



I 



6 









r 






C. E. Wai.kku and Amck L. Kmisleton, " Ohservatious on die Xuolooli of Hydra Fusca. 



p 



Ill 



IS CHOLINE PRESENT IN THE CEREBROSPINAL FLUID OF 
EPILEPTICS ? By S. Kajiura, Imperial Japanese Navy. (From 
the Physiological Laboratory, King's College, London.) (With 
one Plate.) 

{Received for publication 30th July 1 908. ) 

In their investigations on the presence of choline in cerebro-spinal fluid 
and blood, Halliburton and Mott pointed out that it was only present 
in organic diseases of the nervous system, and that this was a fact of 
diagnostic value in distinguishing diseases where there is an actual 
breakdown of nervous material from those which are merely functional.^ 
Although these observers, and those who immediately followed them, 
employed as their principal chemical test the formation of crystals of 
the choline platino-chloride — a test which it has since been shown is not 
absolutely conclusive by itself — nevertheless the absence of choline can 
be inferred when this test gives negative results. 

On the other hand, it has been contended by Donath ^ that choline is 
generally"^ ^3«gH!d in the cerebro-spinal fluid in cases of epilepsy ; and he 
relies for the detection of choline upon the fact that the crystals of th^ 
platinum compound of choline are doubly refracting, and can thus easily 
be distinguished from platinum compounds of potassium and ammonium 
chloride, which under the ordinary microscope are liable to be mistaken for 
those of the choline salt. 

In his Oliver-Sharpey lectures,'* Professor Halliburton said: "This 
requires confirmation. Granting that the diagnosis of epilepsy in 
Donath's cases was correct, it appears necessary to remove epilepsy from 
the list of functional diseases, if the existence of choline in the fluids of 
such cases is a fact." 

The research of which this paper is the outcome was undertaken with 
the view of ascertaining whether or not Donath is correct in his state- 
ments. The difiiculty of demonstrating the presence of choline when mixed 
with potassium and ammonium salts, as always happens when one is 
dealing with physiological fluids, has been overcome by the introduction of 
the periodide test by O. Rosenheim,* and it appeared necessary to 
investigate the question by means of this characteristic and trustworthy 

* See fully Halliburton, Ergebn. d. Pliysiologie, iv., pp. 72-74, 1906. 
^ Joum. of Physiol., xxxiii., p. 211, 1905-6. 

3 British Medical Journal, May 4, 1907. 

* Joum. of Physiol., xxxiii., p. 221, 1906; xxxv., p. 445, 1907. 

VOL. I., NO. 4. — 1908. 20 



292 Kajiura 

reaction. Dr Donath had sent to Dr Rosenheim seven samples of 
cerebro-spinal fluid from cases of genuine epilepvsy, and the latter kindly 
placed them in my hands for the purpose of this research. I have to 
thank Dr Rosenheim for his care in supervising my work. I am also 
deeply indebted to Dr F. E. Batten, Dr A. Connel, and Dr J. Bernstein 
for a number of other specimens which have enabled me to extend the 
series of observations. In all cases th-e fluid was removed by lumbar 
puncture during life. 

In order to make myself acquainted with the method, and at the same 
time to test its reliability and sensitiveness, I carried out a series of tests 
by adding pure choline to an artificial cerebro-spinal fluid. The latter was 
prepared by mixing egg-albumin and dextrose in the necessary proportions 
^dth a solution of potassium and sodium chloride of a strength correspond- 
ing to that found in cerebro-spinal fluid. The choline hydrochloride 
employed was prepared from lecithin ai>d purified by repeated crystalli^- 
tion of its mercury double salt.^ Using Rosenheim's simplified iodine 
test,^ I have' been able to detect choline easily and with certainty when it 
was mixed with the artificial cerebro-spinal fluid in the proportions 1 : 1500, 
1 : 5500, and 1 : 10,500. In the last two instances, only 15 c.c. of the mixture 
were used. Dilutions greater than 1 : 10,500 I should regard as of no 
practical importance; but Rosenheim obtained positive results when the 
proportion was 1 : 20,000. 

In the examination of the material sent by Dr Donath, my general, 
plan of testing for choline was (1) to try the periodide test with the purified 
alcoholic extract; (2) to employ Donath s micro-polariscopic method, 
all his directions being carefully observed; and (3) to use Rosenheim's 
original periodide test as applied to the platinum salt. 

The results are given in Table I.^ p. 293. 

In all cases Donath's test gave a positive result; a few doubly re- 
fracting crystals could always be distinguished under the crossed nicols of 
the polarising microscope (see fig. 2). The yield of the platinirm pre- 
cipitate was in all cases extremely small, amourfbing usually to not more 
than a slight haze ; so that it was necessary to wait many hours for it to 
settle before collecting it on a filter. In no instance could the periodide 
reaction be obtained either with the alcoholic extract direct, or from the 
platinum precipitate. 

1 In his first communication on the periodide reaction, Rosenheim applied the iodine 
solution to the crystals of choline platino-chloride, and found that they dissolved and were 
replaced by crystals of choline periodide. Dr Rosenheim informs me that the mercury 
salt may be used instead of the platinum salt for this purpose. I have further examined 
the lead, zinc, and cadmiiun salts in the same way, and found that they all show the reaction. 
The cadmium salt, owing to its insolubility in alcohol, seems to be the most suitable to 
replace the platinum salt, if that should be necessary (see also F. W. Schmidt, Zeits. f. 
physiol. Chem., liii., p. 428, 1907). Still, as Rosenheim pointed out in his second paper, it 
is not really necessary to prepare a metallic salt first, for the iodine reaction can be obtained 
straight from the alcoholic extract of any fluid which contains choline. 

2 Joum. of Physiol., xxxv., p. 465, 1907. 



Is Choline present in the Cerebro-Spinal Fluid of Epileptics ? 293 

It follows from these results either that choline is present, but in 
amounts too small for the periodide test to reveal it, or that Donath's test 
is untrustworthy and choline is absent. If the first alternative is correct, 
Donath's test is much more sensitive than the periodide reaction. 

Tablk I. 



1 






Platinum compound prepared. 


Donath's 


Quantity of 
cerebro-apinal 


Periodide 

test. 


1 








fluid. 


Donath's 


Periodide 








polariscopic test. 


reaction. 


1 


33 c.c. 


_ 


+ 




2 


36-5 „ 


- 


+ 


- 


3 


24 „ 




+ 


4 


28 „ 


- 


+ 


5 


15 „ 


- 


+ 


6 


36 „ 


- 


+ 


7 


11 „ 


— 


+ 



Tlie formation of anisotropic crystals on the addition of platinum 
chloride is really not so characteristic of choline as Donath appears to 
consider. It does not exclude the possibility that other bases are present in 
cerebro-spinal fluid which form anisotropic crystals, and which do not give 
the periodide reaction for choline. Rosenheim has already pointed this 
out in one of his papers,^ and I have been able to confirm his observations. 

But, before coming to a final conclusion, it seemed advisable to examine 
a few more cases of epilepsy, and the results are given in Table II. : — 



Table II. 



Zt'^: Dise.se. 


Quantity of 

cerebro-spinal 

fluid. 


Periodide 

test. 


Platinum precipitate. 


Donath's 

test. 


Periodide 

test. 


8 Epilepsy 

9 1 

10 ; „ 
11 


5 c.c. 
3 „ 
10 „ 
2-5 „ 


1 ++ 1 


+ 
+ 
+ 

+ 


+ 



The results of the examination of these four specimens are very interest- 
ing. When sent to the laboratory, they were simply labelled epilepsy ; and 
it will be seen that in two of them choline was found by the periodide test, 
notwithstanding that in one of them (case 9) the quantity of fluid at my 

1 Journ. of Physiol., xxxv., p. 469, 1907. 



294 Kajiura 

disposal was very small. The quantity, however, was too small to enable 
me to complete the series of tests by the application of the periodide 
reaction to the platinum compound. This was also the case in No. 11. 

In view of these two positive results, a closer investigation of them 
appeared necessary, for it seemed improbable that they were cases of 
simple epilepsy ; the notes which Dr Batten kindly forwarded confirmed 
this view. 

Case 9 was a child (F. L. R.) one year old, with congenital mental 
defect, with chorea and epilepsy culminating in status epilepticus. 
Her symptoms while in the hospital appear to have been very 
obscure, and the diagnosis of meningitis and later of encephalitis 
was made. The fluid removed by lumbar puncture contained 
excess of lymphocytes. She was still in the hospital when I last 
heard of her. 
Case 10 was also a child (L. H.) ten years of age, who had had fre- 
quent fits since she was eighteen months old. It did not appear, 
however, to be a simple case of epilepsy, for there was also 
considerable mental defect, though the cause of the latter is 
uncertain. 
In case 9 the clinical evidence is quite clear that something more than 
a functional disorder was present. In case 10 the clinical evidence is not 
so clear, although it was evidently recognised as something more than 
a simple case of epilepsy. The finding of choline in both cases is there- 
fore not unexpected, and these are just the cases where the choline test 
should come to the assistance of the clinical observer, in pointing to the 
probability that an organic lesion is at the root of the malady. 

It now remains to explain how it is that Donath's method gave 
always a positive result, although choline could not be detected by the 
periodide test. 

In the first place, Donath's method always gives positive results, not 
only in cases of epilepsy, but also in normal cerebro-spinal fluid, or in the 
cerebro-spinal fluid of every disease in which I have had the opportunity of 
examining it. 

These results are shown in Table III., p. 295. 

In cases 12 and 13 the fluid was normal, and in case 16 the fluid was 
also approximately normal. 

Negative results (by the periodide test) in cases of meningitis were 
previously noted by Rosenheim;^ inflammation of the meninges does not 
necessarily cause any noteworthy change in the underlying cerebral matter. 
Yet, as the table shows, all cases gave a positive result when subjected to 
Donath's test. 

As it seemed possible, as mentioned before, that the obtaining of 
anisotropic crystals may indicate the presence of other organic bases, I 
collected the platinum precipitates of several cases (which gave a positive 
1 Journ. of Physiol., xixv., p. 467, 1907, 



Is Choline present in the Cerebro-Spinal Fluid of Epileptics ? 295 

Donath's test, and a negative periodide reaction), ignited them in a 
platinum dish, and extracted the residue with dilute hydrochloric acid. 
Donath's test was again performed with this extract. Although all the 
organic matter had been destroyed, nevertheless doubly-refracting crystals, 
although less numerous than before, were obtained. It is clear, therefore, 
that part at least of the anisotropic crystals represent inorganic substances. 



Table III. 



Number 
of case. 


DLsease. 


Amount 

of Huid. 

examined. 


Periodide 
test. 


Platinum precipitate. 


Donath's 
test. 


Periodide 
test. 


12* 
13 
14 
15 

It; 

(a) 

(b) 

\^ 

17 

18 
19 
20 


Cancer of stomach 
Nephritis 
Post-basic meningitis 

Congenita.1 liydrocephahis 

Epidemic cerebro-spinal 
meningitis 

No record given 

" 


33 c.c. 

28 „ 

10 „ 

5 ,, 

5 n 

15 „ 
5 „ 

5 » 
60 „ 
45 „ 
13 „ 

t 

10 „ 

5 » 


Ill 1 1 1 1 1 1 1 1 1 1 1 


+ 
+ 
+ 

+ 

+ 
+ 

+ 
+ 
+ 
+ 
+ 

+ 

+ 
+ 


- 



* In this case the fluid was removed twelve hours after death. 

This result led me to carry out a control experiment, using distilled 
water instead of cerebro-spinal fluid. The method was carried through in 
the usual way, and the inorganic salts employed (potavssium carbonate, 
potassium chloride, and platinum chloride) were carefully purified specimens. 
Finall3% under the polarising microscope a crop of doubly - refracting 
crystals^ indistinguishable in quantity and form from those obtained in 
cases of epilepsy, were obtained. I have repeated this several times with 
the same result. This is illustrated by the accompan3nng photographs. 
Fig. 1 shows the appearance of the doubly-refracting crystals obtained in a 
blank experiment, using 20 c.c. of distilled water instead of cerebro-spinal 
fluid. Fig. 2 shows the same from a case of genuine epilep.sy. If these are 
compared with thase published by Donath,' they will be seen to be 
practically identical. 

It is clear, therefore, that such crysttils are no proof of the presence of 
choline, but owe their origin (at least in part) to minute impurities in the 

' Loc. cit., p. 215. 



296 Is Choline present in the Cerebro-Spinal Fluid of Epileptics ? 

reagents employed. It must be remembered that the delicacy of the 
polarisation microscope is so great that impurities which cannot be 
detected by ordinary chemical methods are revealed. Whilst it may of 
course be possible to purify the salts used in the test to such a degree that 
they are free from these impurities, I have been unable up to the present 
to obtain any specimens which do not show some anisotropic crystals 
under the conditions of Donath's method. Kahlbaum's purest potassium 
chloride and potassium carbonate give the reaction easily. 

Whilst this investigation was in progress a paper appeared by 
A. Ziveri^ on the presence of choline in cerebro-spinal fluid in mental 
diseases. By applying Rosenheim's periodide test, he found choline 
absent in cases of epilepsy, only one positive result being obtained in 
twenty-six examinations. 

The conclusions arrived at by M. Kauf mann,'^ so far as they refer to 
the absence of choline in epilepsy, confirm my results ; but his further state- 
ments that the base found by him in cerebro-spinal fluid in other cases Is not 
choline do not seem to be borne out by his own experiments, and are not in 
agreeiiient with the results of previous observers and my own. 

The following are the general conclusions to be drawn from the 
investigations : — 

1. Rosenheim's periodide test for choline is both trustworthy and 
sensitive. 

2. Relying upon this test, choline is found to be absent from the 
cerebro-spinal fluid in cases of genuine epilepsy. 

3. The detection of a few fragments of anisotropic crystals by 
Donath's raicro-polariscopic test is not in itself sufficient evidence of the 
presence of choline in the cerebro-spinal fluid of epileptics ; for the same 
result may be obtained with normal cerebro-spinal fluid, or even with 
distilled water. 

' Riv. di Neuropatologia, etc., vol. i. p. 1, 1908. 

« Neurolog. Zentralbl., Na 6, March 1908, p. 280. The errors iu Kaufmann's 
experiments have been fully demonstrated by J. Donath in a paper which will be 
published in the forthcoming number of the same Centralblatt, an advance proof of 
which has reached me through the kindness of Dr Donath. 



',>itarterhj Journcd of E.rperimental Physiologii, Vol. /., 1908. 



zu' 




Fig. 1. 




Fig. 2. 



S. Kajiuua, " Choline in Cerebro-Siiimil Fluid. 



ON SO-CALLED " PROTAGOX " By Otto Rosenheim and ^I. Christine 
Tebb. (From tlie Physiological Laboratory, King's College, London.) 
(With one Plate.) 

{Received for publication 31st July 1908.) 

[n the first number of this Journal ^ Wilson and Cramer made an attempt 
to rehabilitate " protagon " as a substance of definite chemical composition 
and of constant physical properties. For this purpose the old assumption 
of Liebreich, previously abandoned by Cramer, that " protagon " is decom- 
posed by warm or boiling alcohol, is revived. The decomposition is assumed 
to be a hydrolytic one, and explains, according to Wilson and Cramer, the 
discrepancies in the analytical figures obtained for " protagon " by other 
investigators. By means of a method which is essentially a modification 
of Couerbe's (1834) original method for the preparation of what he called 
c^r^brote^ ( = " protagon "), and by limiting the time of extraction to one 
aiid a half minutes, they prepared a new standard " protagon " and maintain 
that it can be recrystallised from alcohol without decomposition, the only 
condition being to limit the action of the solvent to a short period of time. 
As an index of the decomposition they rely upon a change in the supposed 
physical constants. The specific rotation of " protagon " in pyridine is 
stated by Wilson and Cramer to be [a]^-^^ -j-G'S^, whilst " decompo.sed 
protagon" according to them shows [aj: = { + 1)m^° (1308 to 13-43). By 
a strange oversight the sign of the rotation in the latter case is not stated. 

In the course of an investigation on the behaviour of " protagon " 
solutions in polarised light, which has led us to the discovery of a new 
phenomenon (Spherorotation), we have found that any " protagon " (pro- 
vided that its constituents are present in such proportions as to make the 
amount of phosphorus about 1 per cent.) possesses an initial dextrorotation 
of [a]n = -f6-8° and a final Isevorotation of [a]™ = —13-3".^ This change 
from dextrorotation to Isevorotation, the nature of which has been 

» Journ. of Exp. Physiol., vol. i., 1908, p. 97. 

2 It .soeni.s hardly necessary, in view of the results to be discussed in the present 
paper, to enter into tlie charge of inconsistency made by Wilson and Cramer against our 
identification of "cerebrote" and " acetone- protagon '' with "protagon." It can be easily 
shown by the methods to be described that the.se two products represent the same hetero- 
geneous mixtures as, and are identical with, "protagon." With regard to the different 
solubility of "protagon" in acetone before and after extraction from brain, Wilson and 
Cramer seem to have forgotten that exactly the .same behaviour of " protagon " towards 
ether was de.scribed in Liebreich's original pa]>e,r. This is, indeed, now generally recognised 
as characteristic of the behaviour of mixtures of lijwids. 

^ Journ. of Physiol., vol. xxxvii., 1908, p. 341 and p. 348. 



298 Rosenheim and Tebb 

explained fully in the other papers just referred to, has not been mentioned 
by Wilson and Cramer. It is a property of any " protagon," whether 
the same has been subjected to the supposed decomposing influence of 
boiling alcohol or not. The value [a]i, = (+ or — )? 133° of Wilson and 
Cramer is therefore in no way characteristic for " decomposed protagon," 
and this fact alone deprives the whole decomposition theory of its 
foundation. 

During these investigations we have also obtained further chemical 
evidence, if such should be still wanted, which proves clearly that our 
previous conclusions with regard to the composite nature of " protagon " 
also apply to the new standard "protagon " of Wilson and Cramer. 

The Composition of "Protagon" is completely changed by 
Recrystallisation from Alcohol. 

The main chemical fact on which Wilson and Cramer rely for their 
view that " protagon " is of a definite composition is the possibility (which 
has never been denied) of its recrystallisation from alcohol without change 
in its composition. The one and only condition which must be fulfilled 
to attain this end consists, according to them, in limiting the exposure to 
the hot solvent to a short period of time. 

It is clear, however, that this fact, even if the above condition was the 
correct one for its achievement, would in no way prove the definite com- 
position of "protagon"; for a mixture of substances which possess 
approximately the same solubility in a given solvent, may also retain its 
relative composition under these conditions. This is well illustrated by 
the case of phytosterin, which since its discovery by Hesse ^ has also 
been assumed to be a substance of definite chemical and physical properties, 
a -product of constant composition, melting point, and optical activity being 
always obtained by recrystallisation from alcohol. Windaus and Hauth^ 
recognised, however, that Hesse's product was a mixture of two substances. 

In the recrystallisation of " protagon " there is, however, another factor 
which has been completely neglected by Wilson and Cramer. We have 
looked in vain to find in their paper any statement as to the proportion of 
" protagon " to alcohol employed for the recrystalli.sation. This proportion 
is evidently of no importance if one is dealing with a definite chemical 
substance such as cholesterin. Whatever the amount of solvent used, the 
recrystallised product will always be cholesterin. But in the case of 
a mixture of substances the amount of solvent used for recrystallisation is 
obviously of the greatest importance. 

We found that, by simply varying the proportion of " protagon " to 
alcohol, we obtained variations in the phosphorus percentage of " protagon " 
of over 50 per cent., whilst taking care at the same time to avoid the 

' Anil. d. Chem., 192, 1878, p. 175. 

' Ber. d. d. chem. Ges., 39, 1906, p. 4.378. 



On so-called " Protagon 



299 



supposed decomposing influence of alcohol by limiting the time of heating 
during recrystallisation to one minute. 

We give the following experiment in detail ; it is perfectly typical of 
several we have performed. 

" Protagon " was prepared from ox-brain, following scrupulously Wilson 
and Cramer's directions. The greatest care was taken to limit the time 
of extracti(m with hot alcohol to one and a half minutes, and the time of 
heating during recrystallisation to one minute. The extracts and solutions 
were cooled at once on ioe. An insoluble residue remained in the first and 
second recrystallisation, very little in the third and fourth. The following 
, analjrtical figures were obtained : — 





Proportion "protagon" to 


P% 


Original " protagon " 

1st recryst 

2nd ,, . . . 
3rd „ ... 
4th „ . . . 


3 extractlMiB ; 1 part brain : 
2 parts alcohol 
1:18 
1 :30 
1 :20 
1 :20 


133 

0-83 
0-54 
0-44 
039 



rt will be seen from these results that the phosphorus percentage of 
" protagon " falls during four recrystallisations from 1-33 per cent, to 039 
per cent., notwithstanding the strictest adherence to Wilson and Cramer's 
conditions. 

From the final phosphorus-poor product we obtained easily by recrystal- 
lisation from glacial acetic a<Jid (Koch), or by a slight modification of 
Thudichum's method, a substance which agrees in all its properties with 
Thudichum's phrenosin ( = Gamgee'8 " pseudocerebrin " and Thier- 
f elder's "cerebron ") and which is free from phopphofus and sulphur. 

We give below the results of a complete analysis of this substance in 
order to show the striking differences from the composition of the original 
" protagon '.' : — 

Phrenosin (our analysis) 

Protagon (Wilson and Cramer's analysis, sample D) 

The results obtained in the recrystallisation of protagon from alcohol, 
which agree with those recently published by Gies and Cohen,^ are in 
direct contradiction to those of Wilson and Cramer, who maintain that 
the phosphorus percentage of " protagon " remains constant, although they 
obtain the relatively low figure of 092 per cent, after the fourth recrystal- 
lisation. (The phosphorus percentage of their original " protagon " is not 
given.) 

' Proc. Soc. Exper. Biol, and Med., vol. v., 1908, p. 97. 



c% 


H% 


N% 


P% 


s% 


68-96 


10-32 


1-82 


none 


none 


66-40 


10-71 


2-55 


1-02 


0-68 



300 



Rosenheim and Tebb 



We were; however, able to find the correct explanation for this dis- 
crepancy, which by itself proves the indefinite composition of " protagon." 
By carrying out a tedious series of systematic recrystallisations, constantly 
controlled by analysis, we have found out the exact conditions under which 
the phosphorus percentage of " protagon " remains approximately constant. 
Wilson and Cramer seem to have arrived at this result empirically 
without, however, recognising or stating the true conditions. These depend 
in no way on the time of heating and the supposed decomposition, but 
simply on the proportion of "protagon" to alcohol employed. Wilson 
and Cramer favoured evidently a very small proportion, as is indicated 
in the paper by Lochhead and Cramer.^ We found that, by limiting the 
amount of alcohol used for recrystallisation to a minimum (" protagon " : 
alcohol = 1:5 for the first and 1 : 2 or less for the subsequent recrystallisa- 
tions), we were able to keep the phosphorus percentage at the figure held 
by Wilson and Cramer to be characteristic for "protagon." 

It is obvious that these are not the conditions which favour the intended 
purification. When using reasonable amounts of solvent for this purpose, 
as in the preceding series of recrystallisations, it will be seen that 
" protagon " undergoes a complete change of composition. 

We again quote only one typical experiment, in which another sample 
of Wilson and Cramer's "protagon" was employed. The same pre- 
cautions were taken during its preparation and recrystallisation as indi- 
cated above, the only altered factor being the proportion of " protagon " 
to alcohol. The results were as follows : — 





Proportion of 
" protagon " : alcohol. 


P% 


Original " protagon " 

1st recryst. 

2nd „ ... 


3 extractions ; 1 part brain : 

2 parts alcohol 

1 :", 

1 :2 


1-18 

0-99 
0-96 



Even here a distinct drop in the phosphorus percentage is noticeable, but 
owing to the whole of the dissolved substance being precipitated again 
by cooling on ice, no eflBcient separation of the constituents can possibly 
take place. ^ 

The above results, which agree with those of Wilson and Cramer, 

1 Biochem. Journ., ii., 1907, p. 350. See also Cramer's letter to Posner and Gies, 
publisTied by the latter in Journ. Biol. Chem., i., 1905-6, p. 79. 

^ We should like to point out that the term " recrystallisation " is hardly justified in 
this case, as the product precipitated from its hot solution by cooling on ice is amorphous. 
"Protagon" can only be obtained in crystalline form, according to Liebreich and 
Gamgee, if the temperature is allowed to fall very gradually. Under these conditions 
" protagon " would, however, be decomposed in Wilson and Cramer's sense. According 
to Gamgee and Blankenhorn (Zeitschr. f. physiol. Chem., iii., 1879, p. 277), their 
best result was obtained when the temperature fell in seventeen hours from 41-25° to 27-5°. 



On so-called " Protagon " 301 

show that, whilst it is possible to retain approximately the composition of 
" protagon " during recrystallisation, this achievement is in no way due to 
the prevention of a hypothetical decomposition. The great variations in 
the analytical figures usually obtained for " protagon " are explained in the 
simplest manner. There is, therefore, no need t6 take refuge in the far- 
fetched " decomposition " theory which, besides not being confirmed by our 
critical re-examination of Wilson and Cramer's statements as to the 
physical constants of "protagon," is also a priori most improbable in view 
of the fact that alcohol is not a hydrolytic agent. 

The Isolation of the Constituents of '"Protacon" 

BY MEANS of PYRIDINE. 

A generally recognised method of showing the uniformity of any 
substance consists in subjecting the same to fractional crystallisation or 
precipitation.^ Wilson and Cramer considered the results of our previous 
fractionations as being produced ■ by " decompasition." Although this 
criticism has now been shown to be unfounded, we thought it nevertheless 
advisable to repeat our fractionation experiments under conditions which 
make a decomposition in Wilson and Cramer's sense impossible. This 
can be easily done, as already stated in a previous communication,^ by using 
a mixture of inert solvents. 

It is remarkable that Wilson and Cramer refrain from employing 
fractional crystallisation as a test for the uniformity of their new " pro- 
tagon," especially as the solvent (pyridine) chosen by them for the deter- 
mination of the supposed physical constants of " protagon " lends itself 
admirably to this purpose. A decomposing influence at the low temperature 
in question is not to be assumed, and pyridine would not have been 
employed by them as their standard solvent if such an influence had 
been suspected. 

" Protagon " is fairly soluble in pyridine at 30" to 45° C, and a precipitate 
is formed on cooling its solution. Evidently, if " protagon " is of a uniform 
composition, the precipitate must again consist of " protagon " with 1 per 
cent, phosphorus. Although we hardly expected this from our previous 
experience with this substance, we were nevertheless surprised to find on 
analysis that the precipitate contained more than double the amount of 
phosphorus, namely, 2 5 per cent. This result led us to a systematic 
examination of the fractions into which " protagon " can evidently be 
divided by means of pyridine. 

For this purpose a 3 per cent, solution in pyridine of a recrystallised 
"protagon," prepared by Wilson and Cramer's modification of Gamgee's 
method, was employed. The solution was effected at 45° C, and the 
temperature kept at 45" for only one minute. The perfectly clear solution 

' See H. Meyer, Analyse, etc., organ. Verbindungen, p. 13. 
* Proc Physiol. Soc, xxxvii., 1908, p. 1. 



302 



Rosenheim and Tebb 



was At once cooled on ice to room temperature (15° C.) and the precipitate 
filtered either after a few minutes or in some cases after half an hour. The 
filtrate was poured into two volumes of acetone, which produced a small 
quantity of a flocculent precipitate. After the removal of the latter by 
filtration, the acetone-pyridine solution was cooled on ice, by which process 
the bulk of the dissolved product was brought down. 
The results of the analysis are given below : — 



Original " protagon "... 

Fraction 1. (insoluble in pyridine 
at +15') . . '. 

Fraction II. (precipitated by ace- 
tone at ± 0°) . 



107 


N% 


2-46 


2^1 


3-03 


0-09 


1-73 



These resultiS furnish the most striking proof for the composite nature 
of " protagon." Under the conditions of the experiment a decomposition in 
Wilson and Cramer's sense is impossible, and nevertheless "protagon" 
is divided by one fractionation into a practically phoephorus-free part 
and one with two and a half times as much phosphorus as that of the 
original " protagon." The nitrogen figures also undergo characteristic 
changes. 

(a) The phosphorus-rich constituents. — The quantity of this 
fraction averages one-third of the " protagon " employed. We possess, 
therefore, in pyridine a solvent by means of which the phosphorus-rich 
moiety of " protagon " can be easily isolated. Its phosphorus percentage 
is not appreciably raised by repeated (three times) recrystallisation from 
pyridine, a fact which would, according to Wilson and Cramer, speak 
for its definite composition. We have, however, succeeded in isolating its 
main constituent, the di-amino-phospatide, sphingomyelin, by a fractiona- 
tion method with alcohol-chloroform and acetone, which we shall com- 
municate later in detail. For the purpo.se of comparison with " protagon " 
^e give below a complete analysis of the purest preparation which we 
succeeded in obtaining so far. We have, however, reasons for believing 
that this substance, which agrees in its properties with Thudichum's 
sphingomyelin, is not yet perfectly uniform 

Sphingomyelin (our analysis) 

** Protagon " (Wil»on and Cramer's analysis, sample D) 

(b) The phosphorus-free constituents. — Fraction II. as obtained 
above may be easily rendered perfectly phosphorus-free by further re- 
crystallisation from glacial acetic acid (Koch) or by the method indicated 
previously. 



c% 


H% 


N% 


P% s% 


62-90 


11-54 


3-33 


3-46 ... 


66-40 


10-71 


2-55 


1-02 0-68 



On ao-called " Protagon " 303 



Micro-chemical Proof of the Composite Nature of "Protagon." 

We have examined the physical properties of the substances isolated 
from " protagon " somewhat more closely, and found characteristic differ- 
ences in their optical activity, melting point, solubility, etc. They possess 
further the remarkable property of crystallising from pyridine under certain 
conditions in ^uid spherocrystals,^ and we had indications that they also 
exist in a liquid-crystalline state between the solid and completely fused 
condition. During the latter observations we noticed a striking difference 
between them by the help of the polarising micrascope, which furnishes a 
further proof of the heterogeneous composition of " protagon." 

If a small quantity of the phosphorus-rich material mentioned above 
(Fraction I.) is carefully fused on a slide under a cover-glass, the clear fused 
liquid is seen to be isotropic between the crossed nicols of a polarising 
microscope. On allowing the slide to cool a shower of separate bright 
spherocrystals, showing dark crosses, appears on the black background. 
The spherocrystals grow rapidly on cooling until they touch each other, 
forming finally a complete mosaic on solidifying. The same property is 
shown still better by the specimen of sphingomyelin described above 
(see fig. 1). 

Quite different is the behaviour of the phosphorus-poor fraction (Fraction 
II.) and of pure phrenosin. An isotropic fluid is also produced on complete 
fusion, but instead of spherocrystals it will be observed that on cooling 
bright anisotropic needles shoot out on the dark background (see fig. 2). 
In ordinary white light both the needles of phrenosin and the sphero- 
crystals of sphingomyelin can only be faintly distinguished by their outlines. 

It seemed to be of interest to examine in the same way samples of 
" protagon," and this method also proved it to be a mixture. As will be 
seen from the illustration (see fig. 3), " protagon " when fused and allowed 
to cool slowly gives rise to an indefinite crystalline mosaic, in which the 
needles of phrenosin seem to predominate. The identical result is obtained 
from " artificial protagon," i.e. a mixture of the phosphorus-rich and 
phosphorus-free constituents which we described previously (loc. cit.) and 
found to be identical in chemical and physical properties with " protagon." 

Summary. — It will be remembered that the "protagon" idea was 
originally conceived by Liebreich (1865). In his opinion all the con- 
stituents of nervous tissue (known at this time as phosphorised fats, 
lecithin, cerebrin, etc.) do not exist preformed, but are derived from the 
decomposition of the one and only mother-substance, which was therefore 
called " protagon." This simple idea had no doubt a certain attraction for 
the earlier physiological chemists, especially as it was for a time at lea«t 
adopted by Hoppe-Seyler. Several additional theories, none of which 
were supported by facts or stood the test of experimental criticism, had to 

1 Joum. of Physiol., vol. xxxvii., 1908, p. 348. 



304 On so-called " Protagon " 

be made in order to keep the original theory alive. There is now no doubt 
that lecithin, to quote only one of the supposed derivatives of " protagon," 
occurs independently in brain, and besides in no way enters into the 
composition of " protagon." If we consider the diversity of the mixtures of 
lipoids occurring in other organs (as demonstrated by the recent researches 
of Erlandsen on the lipoids of the heart, of Bang and others on those of 
the blood, of Thierfelder and Stern and of Frankel on those of the 
egg), there seems to be no justification for the assumption that the lipoids 
of the brain, the most complex organ, should be derived from one uniform 
substance, " protagon." In view of the clear evidence against it, we 
think, therefore, that the time, has arrived to dismiss finally the primitive 
" protagon " idea as unfounded and contrary to modern conceptions and 
knowledge. It has for the last forty years unfortunately retarded pro- 
gress in the investigation of the chemistry of the nervous tissues, instead 
of stimulating it. We are forced by the results of our prolonged study 
of the question to repeat our previous conclusion : 

"Protagon" is a heterogeneous mixture, and the term " pro- 
tagon " has only a historical justification. 

The expenses of this research have been in part defrayed from a grant 
from the Government Grant Committee of the Royal Society. 



DESCRIPTION OF PLATE. 

Fig. 1. Crystals of sphingomyelin, x 70. Polarised light : crossed nicols. 
Fig. 2. Crystals of phrenosin. x 88. Polarised light : crossed nicols. 
Fig. 3. Crystals obtained from "protagon" after fusion, x 130. Polarised light ; 
crossed nicols. 



ijimrterly Journal of Experimental P}t,ysiolo(j (J, Vol. I., 1908. 



boH' 



/ J^ 







Fig. 2. 




Fig. 3. 
0. Rosenheim and M. C. Tebh, " On so-called Protagon. 



321 
329 



THE COAGULATION TIME OF THE BLOOD IN MAN. By T. Addis. 
(From the Physiology Laboratory, University of Edinburgh.) (With 
five figures in the text and two Plates.) 

(Received for publication 1st August 1908.) 
CONTENTS. 

PAGK 

I. A New Method of estimating the Coagulation Time of the Blood 305 

II. The Conditions which are essential for thb Accurate Estimation 

OF the Coagulation Time ...... 314 

III. A Review of other Methods ..... 

IV. The Effect of Variations of Temperature on the Coagulation 

Time •••..... 

V. The Absence of Diurnal Variation in thu Coagulation Time . 330 

VI. The Effect of the Administration of Calcium and Citric Acid by 

the Mouth on the Coagulation Time .... 331 

VII. Conclusions ......... 331 

I. A New Method of estimating the Coagulation Time. 
When a current of oil streams against the edge of a drop of blood sus- 
pended in oil, a continuous smooth flow of the corpuscles is induced, 
although the drop as a whole does not rotate. Under the microscope this 
flow will be seen to cease quite suddenly after a certain time has elapsed. 
This is due to the occurrence of coagulation in the drop. 

The method is a modification of Brodie and Russell's method (1). In- 
stead of intermittent jets of air at an unknown and variable temperature, a 
continuous stream of oil at a known and constant temperature is used. The 
end point also is entirely different. 

The whole apparatus is placed on a small table to which an upright has 
been attached (see Plates at end of article). 

A reservoir (P) of filtered mineral oil such as is used for burning in 
lamps is hung from the upright by a cord passing through a pulley so 
that its height can be varied at will. Any vessel will do for a reservoir so 
long as it contains a sufficiently large surface of oil, in order that the 
amount which runs out during an observation may not materially affect 
the pressure of the flow of oil by altering its level in the reservoir. 

The oil is conducted from the reservoir through six feet of flexible metal 
asbestos-lined quarter-inch tubing (p). The usual flexible metal tubing 
should not be used, because it is packed with rubber which soon rots under 
the influence of the oil. 



306 



Addis 



The 'last 4^ feet of the tubing are coiled spirally within a tank of 
water (V), and the lower end emerges through the side of the tank at the 
level of the stage of a microscope and terminates in a stop-cock (fig. 1). 

The next part of the apparatus is an adaptation of Bogg's modification 
of Brodie and Russell's method (2). Bogg's apparatus consists of a small 
circular metal box, the floor of which is of glass. The box is closed above 
by an inverted truncated glass cone. It is pierced on one side by a small 
metal tube which ends in a nozzle (figs. 1 and 2). 

The metal tube is screwed on to the stop-cock on the side of the tank. 




Fig. 1. — Diagram showing the stage apparatus in section. 

V, tank of water; P, flexible metal tabing ; (, metal tube acrewing into tap; e. troncated glass cone in 
positioD ; M, low-power lens of microscope ; R, box hang below the stage to catch the oil which flows 
oat of UiS coi^E^alometer. 



The box is then fitted on to it and locked by a small brass collar which fits 
into pegs driven into the tube and the side of the box, so that any rotation 
of the tube in its socket is rendered impossible. A special nozzle must be 
fitted on to the end of the tube. It is made of soft brass, and is about half 
a millimetre in diameter and 4 ram. long. It is essential that the nozzle 
should point in exactly the right direction. It should be placed so that 
when the stop-cock is turned and a jet of oil issues from it, the stream 
travels tangentially against the edge of the drop of blood which hangs 
from the end of the glass cone. This direction has to be determined ex- 
perimentally by making slight alterations in the direction of the tube with 



The Coagulation Time of the Blood in Man 307 

a pair of fine pliers, until it is found that a smooth and continuous flow is 
produced by a low pressure of oil. 

The nozzle should then be plastered round with solder to preserve it, 
for the slightest knock may alter it a little, and unless it is exactly right 
the flow of the corpuscles is rendered jerky and inconstant. 

When the stop-cock is open the box is of course filled with oil, and tbe 
jet from the nozzJe produces a current which streams across the edge of the 
drop of blood which is surrounded by oil on all sides except where it is in 
contact with the end of the cone. When everything is in position the box 
lies on the stage of a microscope, and tbe flow of the corpuscles can be ob- 
served with a low-power lens. The oil runs out of the box through a hole 
in the metal fitting of the cone and foi-ms a pool on the top in which the 
lens lies when it is in focus. 

The oil is allowed to run over the stage of the microscope and falls into 
a vessel hung beneath it, from which it is collected by a pipe which conveys 
it over the edge of the table into a receptacle on the floor. 

An arrangement is also necessary for keeping the oil which surrounds 
the blood at a constant temperature. The water in the tank is warmed 
by a small gas-jet which is regulated by S chafer's thermostat. 

This tank is round, is 7 inches high, and has a diameter of 7 inches. 
It has a dome-shaped bottom so as to raise it some distance above the gas- 
flame, and is placed on a metal ring 3 inches high which is pierced by holes 
fo the inlet of air to the flame. 

The oil in the 4^ feet of metal tubing which is immersed in the water 
soon acquires the same constant temperature. 

A finely graduated thermometer (fig. 2, T) pierces the box into which 
the cone fits, so that thotigh the shaft is outside, the bulb lies in the 
interior very close to the suspended drop of blood. In this way, the 
temperature of the oil immediately surrounding the blood is accurately 
known. By means of this arrangement the temperature of the oil in which 
the blood lies can be kept constant at any desired temperature for any 
length of time. 

The pressure of the flx3>w depends on the height of the level of the oil in 
the reservoir, and on the calibre of the nozzle. 

With the calibre of nozzle which is at present used by me, the surface 
of oil must be 10 cm. above the blood. 

The pressure may vary from the obstruction of the nozzle by dust, etc. 
To obviate this, the oil is twice filtered, before it is introduced into the 
reservoir. 

It is necessary to have a standard by which the pressure can be gauged. 
This is most simply attained by measuring the length of the jet of oil when 
the box is removed. 

When the receiving vessel, which is hung below the stage of the micro- 
scope, is 39 cm. from the end of the nozzle, the jet of oil just falls into it. 
If it does not do this, there must be something wrong. The nozzle can then 
VOL. I., NO. 1 — 1908 21 



308 



Addis 



be screWed off and the length of the jet, as it issues from the stop-cock, is 
measured in the same way : it should be 65 cm. In this way the location 
of the obstruction can be ascertained. 

This does not need to be done before every estimation, but it should be 
done every now and then, especially if the apparatus is not being constantly 
used. Of course, any considerable variation in pressure .reveals itself, by its 
different action on the blood. 

A temperature of 185° C. is a convenient one to work with. It 
is easily maintained by regulating the temperature of the water in the 
tank The correct temperature having been attained, it is advisable not 
to stop the flow of oil by turning the stop - cock between successive 




Fig. 2. 



-Diagram of theinetal box with the truncated glass 
cone removed. 



The thermometer (T) and the tube (t) are ahown piercing the side of . 
the box ; the bulb of the thermometer and the nozzle of the tube 
being inside the chamber. 



estimations, because it takes a little time for the box to be warmed up 
again. 

The parts of the apparatus which come in direct contact with the 
blood, i.e. the end of the glass cone and the instrument used for punctur- 
ing the skin, have to be freed from any possible contamination with fibrin 
ferment. 

The only way to do this with certainty is to expose them for some time 
to a temperature above 65° C. At temperatures above that point, fibrin 
ferment is destroyed. 

The puncturing instrument and the glass cone are placed in a vessel 
full of oil, the lid of which has three holes cut in it. Into one of these a 
thermometer is fitted, and through the other two the cone and lancet can be 
hung, so that they dip below the level of the oil. The vessel is placed in a 
tin of water, which is then heated until the thermometer indicates that the 



The Coagulation Time of the Blood in Man 309 

temperature of the oil is 70° C. or more. After they have been left in for 
a few minutes they are fitted into the necks of small bottles containing 
ether. The bottles are then well shaken, and the ether dissolves off all 
the oil. 

This is sufficient as regards the puncturing instrument, but it is of the 
utmost importance that the end of the cone shall be not only free from 
fibrin ferment, but that it shall also be free from any dust, or anything else 
which might impede the flow of the blood. 

In the first place, to clean the drop of coagulated blood from it, it is put 
under a strong jet of water. Then it is washed in absolute alcohol, and put 
into the hot oil. After rinsing in ether, a perfectly clean handkerchief, 
made of fine silk, is dipped in ether, and drawn once or twice gently 
across the end of the cone. After this treatment, the cone is fitted into the 
neck of a small bottle, and thus preserved from contamination with dust 
until it is required. 

The following is the method which I find best fitted to obtain at once 
a drop of blood of the right size. 

A slip-knot is placed round the finger, and the arm is swung round ten 
or twelve times. In this way the finger is filled with blood, and becomes 
bright red. When the swinging is stopped the slip-knot is tightened up, so 
that the condition of vascular engorgement is maintained. A superficial 
puncture is made, and a spherical drop of blood about 4 mm. in diameter at 
once appears. The time is then noted. 

The glass cone is gradually approached to the drop. Before it has quite 
touched it, the blood seems to leap up, and flows smoothly right up to the 
edge. It is put at once into the apparatus. The whole procedure, from the 
pricking of the finger to the fitting of the cone into its position, should 
not take more than 10 seconds. 

Jenner's vaccinostyles are well adapted for pricking the finger, but it 
is necessary to fit them with a guard of some description, to prevent too 
deep a puncture being made. 

If the wound is too deep, the drop of blood is apt to spill over, and 
a stream runs from the puncture, from which it is impossible to obtain 
the proper quantity of blood. The only way to get a constant amount 
of blood on the end of the cone is to produce a drop of the right size. 
This is a difficulty in dealing with people who have not been pricked 
before, for there is a considerable variation in the rate of flow in different 
individuals. 

That the size of drop taken up by the cone should be approximately 
constant, is of considerable importance. Variable results are obtained if 
this point is neglected. 

As soon an the cone has been introduced into the box, the reservoir must 
be raised to a little more than twice its original height. This is necessary 
because a higher initial pressure is required, partly to overcome the inertia 
of the corpuscles, but mainly, I think, to break up the blood from the 



310 Addis 

state of slight agglutination into which it passes whenever it leaves 
tlie vessels. 

That the inertia of the corpuscles is not the only factor is shown by the 
fact that the initial pressure required varies according to the time the blood 
is left exposed to the action of the air. If this time is 2 or 3 minutes, it is 
often found that a pressure of four or five times the standard pressure is 
not sufficient to start the flow. 

In different individuals, also, there seem to bo slight variations in the 
rate at which the blood agglutinates. 

If the technique as regards the taking of the blood is strictly 
observed, it will be found that a height of slightly more than twice 
the standard pressure will act uniformly. Thus, with my present ap- 
paratus, in which the standard pressure is represented by a height of 
reservoir above the blood of 10 cm., an initial height of 20 to 22 cm. 
is required. 

The outline of the drop, as the reservoir is being raised, should be 
watched. When the agglutination has been overcome, a tongue of blood is 
seen to stream out from the drop. The reservoir is, then, at once lowered ; 
but it shoukl be done slowly, so as gradually to lessen the rate at which the 
corpuscles are revolving. 

When the low power is brought into its focal position, the lens lies only 
a few millimetres above the top of the glass cone, and is immersed in the oil 
which flows out of the chamber. If the necessary conditions have been 
observed, the corpuscles are seen streaming round fairly rapidly, each one 
separate from the other. The part where the flow is most rapid is the 
point at which the oil stream impinges on the blood. This is the part 
which should be watched. After about 7 minutes have elapsed with 
out any observable change, one or two stationary streaks appear a 
little way from the edge of the drop. These rapidly increase in number 
and length, and with this there is an appreciable diminution in the rate 
of flow. 

Within a half to one minute more, the streakiness will have extended 
right up to the edge, and there will now form a laminated clot, within the 
meshes of which more and more corpuscles become entangled. Only a 
small part of the total number of corpuscles continue to flow slowly and 
interruptedly round. 

The signs of a clot associated with the cessation of flow of the great 
body of the corpuscles, is the end-point adopted. When this is attained the 
time is again taken and the estimation is complete. 

To a greater or a less extent all the conditions necessary for accurate 
results are realised in this method. There is one essential which is not 
fully carried out in any other raethod, but which is perfectly complied with 
in this, i.e. the maintenance of a constant temperature in all comparative 
observations. 

The temperature of the blood, as it comes from the capillaries is about 



I 



I 



The Coai^ulatioii Time of the Blood in Man 311 

37 (J. Ill the 10 seconds or so which are required to place it in the oil, 
the temperature falls; and it continues to fall until it reaches 185^0., the 
temperature of the oil, after which it remains constant. 

As variability of temperature is the most important source of fallacy in 
other methods, so its constancy is the chief advantage in this one. This 
is the one reason why the results are so much more constant than can be 
obtained by other methods. Another point which may be looked \ipou 
as a special advantage is that coagulation occurs under c(mditions not 
very dissiuiilar to those which obtain in an injured vessel, or one into which 
a foreign body has been introduced. . In both cases the blood is flowing, 
and thrombokinase from the tissues has been added to it. In the one case 
it is surrounded by the vascular endothelium, and in the other case by oil 
which in its neutrality as regards (!oagulati(jn is strictly comparable to the 
lining of the vessels.^ 

Coagulation may, therefore, be said to occur under conditions much 
more nearly allied to those under which blood sometimes coagulates within 
the body than when observed by any method hitherto employed. Never- 
theless the method here described is far from being a perfect one. This 
is shown by the fact that irregular variations in the time of coagulation 
still occur even with a constant temperature. These variations are not 
due to alterations in the actual coagulability of the blood, but must, I think, 
l)e attributed to experimental error. 

They appear to be due to two causes — first, slight errors in techniipu'. 
and second, a want of absolute detiniteness in the end-point. 

The errors in technique most likely to arise are those connected with 
th(i picking up of the drop of blood by the cone. 

When the margin of the drop of blood on the end of the cone is ex- 
amined under the microscope, it will be seen that it does not always come 
absolutely up to the edge of the glass surface — a thin margin is sometimes 
left. The blood Vjeing then a littli; further away, the .stream of oil will not 
artect it in quite the same wa}'. Again, the drop is then not always (juite 
circular in (jutline, and the flow of the corpuscles is liable thei-eby to be 
slightly obstructed. 

With regai-d to the errors arising from want of deflniteness in the end- 
point, it may be: ob.s(Mved that there an; three po.ssible stages whicli might 
be adopted as indicating coagulation: — 

(1) The first appearance of a streak of clot. 

(2) The stop])age of the main flow of blood and the clear appearance of 

a laminated clot ; and 

(3) The comj)lete cessation of flow. 

On an avmage 00 seconds passes between (1) and (2), and 50 seconds 
between (2) and (3). 

' In onlei' to oxjilaiii uliy tlu" blood does not loaguliHk' in the vessels, .some "have 
as.sunied that tlie vascular endothelium secretes an anti-body (anticoagnlin). Loeb (:)) has 
shown experimentally that this theory is not tenable. 



312 



Addis 



The 'first appearance of a streak of clot has been found too variable to 
use as an end-point. 

In the following columns the times given by the second stage^ and by 
the third, are compared : — 



Stage II. 




Stage III. 


Stoppage of main flow of blood 
and appearance of a lamin- 
ated clot. 




Practically complete 
stoppage. 


Min. Sec. 


Variation from 
mean. 


Min 


Sec. 


Variation from 
mean. 


7 25 


- 25 sec. 


8 





-41 sec. 


7 10 


-40 „ 


8 


4 


-37 „ 


7 55 


+ 5 „ 


9 


10 


+ 29 „ 


7 45 


- 5 » 


8 


30 


-n „ 


8 45 


+ 55 „ 


9 





+ 19 „ 


7 55 


+ 5 „ 


8 


10 


-31 „ 


7 55 


+ 5 ., 


9 


25 


+ 44 „ 


7 45 


- 5 „ 


9 


5 


+ 24 „ 


8 35 


+ 45 „ 


9 


5 


+ 24 „ 


8 5 


+ 15 „ 


8 


45 


+ 4 „ 


7 5 


-45 „ 


8 





-41 „ 


7 30 


-20 „ 


9 





+ 19 „ 



The average time taken to arrive at stage II. was 7 minutes 49 seconds, 
and to arrive at stage III. was 8 minutes 41 seconds. 

When the second stage was taken as the end-point, the average variation 
from the mean was 20 seconds ; when the third stage was adopted, it was 
27 seconds. 

The second stage lias, therefore, always been observed as the end-point. 
It is better than the third also because it is difficult to say when the 
stoppage is to be considered " practically complete." The positively complete 
cessation of flow is very variable, since a few clumps of corpu.scles .«;Ometinies 
wander slowly round for a considerable time. 

But the second stage also is not always quite definite, though it mucli 
more often is so. The main mass of the blood usually stops moving at a 
definite moment, and in the next second or two a clear laminated clot stands 
out. But sometimes this develops slowly, and in these cases it is impos- 
sible to be quite accurate, for judgment is necessary to decide when the 
clotting is distinct enough to be considered as the end-point. The amount 
of possible error is, however, strictly limited, because the fiow always stops 
completely within, at most, 90 seconds after the commencement of the 
second stage. 

Now and again the main body of the blood ceases to flow without the 
clear appearance of any clot. 

This, I think, is usually due to the agglutination of the blood not 



The Coagulation Time of the Blood in Man 31.'> 

having been properly overcome at the beginning. The time obtained in 
these cases is, as it happens, usually approximate to the mean time. 

Variations in the agglutination of the red blood corpuscles of healthy 
people, as shown by differences in the pressure of oil needed to set the 
corpuscles Mowing each one separately from the other, are slight, and have 
never given rise to serious error or difficulty. 

J found, however, when I came to apply the method to jjathological 
cases, that this agglutination of corpuscles was greatly increased in 
disease. 

In one iiisfance this was so marked that no flow could be started in the 
blood even after the pressure had been raised high enough to drive part 
of the drop off the end of the cone. I cannot at present say in what class 
of cases this condition is most marked, for I have made but few observa- 
tions on pathological conditions ; it was, however, present in a varying 
degree even in some convalescent patients. 

In one or two instances no accurate estimation of the coagulability could 
be made, because even the highest pressures failed to produce a smooth and 
even flow, the corpuscles remained sticking together in clumps, and flowed 
slowly and with jerks. 

It is interesting to note that Fleming (4) has found, in connection with 
work on the opsonic method, that in 90 per cent, of patients at St Mary's 
Hospital the red blood corpuscles are agglutinated by their own serum. 
This condition is very rare in health. Hek toen (5) says that auto-agglutinin 
is very seldom if ever demonstrable in vitro in the blood of healthy people. 
In disease the agglutination is possibly due to the action of bacterial 
haem-agglutinins, for it has been shown by Pearce and Winne (6) that 
certain bacteria produce a substance which agglutinates red blood corpuscles 
in vitro, and which when injected into animals leads to the formation of 
thrombi composed of agglutinated corpuscles. In 1873 Hueter (8) showed 
that certain thrombi in infectious diseases were due not to coagulation but 
to agglutination. Flexner (f)) confirmed this, and reproduced it experi- 
mentally by the injection of bacterial filtrates. Boxmeyer (10), Kraus 
a)id Ludwig (11), Volk and Lipschiitz (^12), Kayser (13), and Eisenberg 
(14) have also demonstrated the action of bacterial hami-agglutinins. 

In cases in which this increased U'udeiicy to agglutination has developed, 
the flow of the corpuscles is hindered and the estimation of the coagulation 
time may be impossible. 

In many pathological conditions, therefore, tlie method is inapplicable, 
though in these cases it may be of value in showing the degree of auto- 
agglutination which is present. 

Unless otherwise stated, the coagulation times given in the following 
sections were taken by this method at a temperature of 185" C. 



314 Addis 

II. The Conditions which are essential for the Accurate Estima- 
tion OF THE Coagulation Time of the Blood by any Method. 

1. The Blood must be obtained under the same Conditions 
in each Experiment. 

Pratt (15), using a modification of Brodie and Russell's method, con- 
cluded that blood from deep wounds took longer to coagulate than blood 
from superficial ones. He gives as an example a coagulation time of 7 
minutes when the cut was deep, and of 2 minutes when it was superficial. 
I have found that there is no appreciable difference in the coagulability 
of blood frpm deep and from superficial punctures. 

In eight comparative observations, the average time from deep punctures 
was 8 minutes 8 seconds, whereas, when the puncture was superficial, it 
was 8 minutes 4 seconds. 

Several observers have stated that congestion of the part from which 
the blood is obtained leads to a diminution of the time, the explanation 
being that a greater quantity of coagulation-accelerating substances from 
the tissues are added to the blood. I have not been able to confirm this. 
Very marked congestion was produced in one arm by the application of a 
Bier's bandage above the elbow. The average of twelve comparative esti- 
mations of the coagulation time of blood taken from the congested fingers 
was 8 minutes 2 seconds, while the average time of blood from the uncon- 
gested fingers was 7 minutes 51 seconds. No conclusion can be drawn from 
so slight a diflference as 11 seconds, because it is well within the limits of' 
experimental error. 

Pressure near the wound while the blood is issuing has been supposed 
to act in the same way, but again I have not been able to find that this 
makes any appreciable difference, the average time with pressure being 8 
minutes 5 seconds, and without pressure 8 minutes 4| seconds. 

The rate of flow of blood from the wound has proved to be of importance 
in so far as it affects the time during which the blood is exposed to air 
before its introduction into the apparatus. 

The other factors requisite to cause coagulation being present, air has 
a marked influence on coagulation : its temperature is variable, and dust 
particles no doubt attach themselves to the drop to a variable extent during 
its exposure to air. 

The finger is the most convenient part from which to obtain blood, 
because in it a constant size of drop and rate of outflow can be secured 
by the employment of temporary congestion. 

In continuous haemorrhage the point of time during the course of the 
bleeding at which a specimen is taken for examination is of considerable 
importance. I could not use my method to demonstrate this, as exposure 
of the blood to air leads to an increase in the agglutinability of the cor- 
puscles. M'Gowan's method (16), modified by the addition of an apparatus 



The Coagulation Time of the Blood in Man 315 

to keep the tubes at a constant temperature, showed that there was a rapid 
and progressive diminution of the coagulation time as the haemorrhage went 
on. Thus the coagulation time (at 18° C.) of the first blood to appear from 
a wound was 7 minutes 15 seconds. The blood flowing from the wound 
after 1 minute coagulated in 5 minutes 15 seconds, and after 2 minutes in 
3 minutes 40 seconds. 

It is important, therefore, to take for examination only the first blood t(j 
appear after a puncture. 

A comparison of coagulation times, taken when the blood was pro- 
tected from contact with the skin by smearing the surface with lanoline 
before the puncture was made, with the time taken in the ordinary 
way, shows that contact with the skin for the short period which elapses 
before the blood can be introduced into the apparatus has no appreciable 
effect. 

On the other hand, when the skin of the part into which the puncture 
is made is covered with recently shed blood, a very marked diminution of 
the coagulation time results. 



free from blood. 


Skin covered with a film of 
recently shed blood. 


min. sec. 

6 40 

7 50 
7 55 


min. sec. 
4 50 
4 20 
3 40 



This diminution no doubt results from the addition to the freshly issuing 
blood of preformed fibrin ferment. 

It is, therefore, necessary to make sure that there are no traces of fibrin 
ferment left on the skin, on the puncturing instrument, or on the parts of 
the apparatus with which the blood comes in contact. 

The only certain method is by destruction of the ferment by heating 
to a temperature above 65' C. This can be done for the apparatus. In 
the case of the skin thorough washing in running water, with sub-sequent 
drying by alcohol and ether is, in practice, sufficient to eliminate this fallacy. 

In connection with the method of obtaining the blood, therefore, it is 
necessary that the skin and instruments should be clean, that the rate 
of outflow should be approximately constant, and that the first blood which 
appears should be taken for examination. 

2. All Estimations must be made at the same Temperature. 

The value of any method is mainly determined by its success, or want 
of success, in maintaining a constant temperature. 

Where no attention is paid to differences of temperature the method is 
practically worthless. The influence of even the slight variations which 
occur in rooms, wards, or laboratories is so great as to make comparative 
observations valueless. 



316 



Addis 



This is in direct contradiction to the conclusions of the originators of 
some of the methods, who, while admitting the importance of large varia- 
tions, believe that slight ones have so little effect that they may, for 
practical purposes, be neglected. No one of them, however, has brought 
forward any experimental proof of this assumption. 

The extraordinaiy variations in the coagulation times obtained by the 
use of their methods are to a great extent due to variations of the 
temperature. This is illustrated in the following three charts: — - 

Figure 3 is a chart showing consecutive coagulation times, taken in 



£-3 a 






PiQ. 8. — Diagram showing the coagulation time 
(thick line) and variations of room tempera- 
ture (thin line) in an experiment upon the 
blood of a normal person taken at different 
periods of the day (Dec. 22), and estimated 
by M'Gowan's luethod. 









1 ' 


















on 




y\ 






















1 o 


-f 


\ 






















1 a 




V 




















17 


/ 




\^ 




















16 


/ 




\ 




;" 


v^ 














1 f\ 


/ 




\ 




/ 


^ 










. I 


-•— • 








> 








x/v 


•^. 




„,-^ 






1 9 






/ 


I 






/ 


















/ 


V 






f4. 
















/ 




X 




^ 




L ^ 




y 


^ 


A 






-V 








»r 






'"""*>* 




> \i 














N 












1 


g 


























7 












' 














R 



























12 1 2 3 4 5 6 7 8 9 10 11 Hours of 

the day. 

a room in which the temperature never varied so much as to make the 
room noticeably cold or hot. 

M'Gowan's method (16) was used. The originator of this method 
does not think that accurate results can be gained by the use of the 
method without the addition of some means of keeping the temperature 
constant. 

In this case, however, the tubes were simply left to acquire the 
temperature of their surroundings. 

The amount of variation in the time is very great, and the curve of 
coagulation, though complicated by experimental error due to other 
causes, is seen roughly to run in the opposite direction to the curve of 
temperature. 

In the next chart (fig. 4) the times were taken by the same method, 
modified by keeping the tubes in an apparatus I had constructed in order 



The Coagulation Time of the Blood in Man 



317 



to keep them at a constant temperature. This gave an almost constant 
temperature, and the variations are seen to be smaller. 

<i> 

•S ^ Fio. 4. — Diagram showing the coagulation time of the 

a .2 blood of a normal person taken at different periods 

•2 g of the day. M'Gowan's method, modified by 

3 's the addition of an apparatus for maintaining a con- 

§0 (3 stant temperature. The temperature wp.s approxi- 

o "" mately 20° C. throughout. 



^— — -• 

, ^ *^ ■ — 



10 11 12 



8 Hours of 
the day. 



by my 



The coagulation times in the next chart (fig. 5) were taken 
method, in which the temperature is kept absolutelj' constant. 

Here the variation is still less, and is due entirely to other experimental 
errors (see Daily Variations in the Coagulation Time, Section IV.). 



E 

O 3 

-is 



Fig. 6. — Diagram showing the coagulat.jn time of 
the blood of a normal person taken at different 
periods of the day (Addis' method). The tem- 
perature of the oil was maintained absolutely 
constant at 18-5° C. 



^r^^'^—^-^—'- 



10 



8 Hours of 



the day. 

The following is a list of coagulation times taken by my hiethod at 
temperatures between 10° C. and 20° C, which may be looked on as the 
extremes of ordinary room temperature : — 

Temperature. Coagulation time. 



•c. 


mill. sec. 


10-25 


21 30 


12-25 


16 30 


13 5 


14 32 


14-5 


12 58 


15-5 


11 46 


16-5 


10 10 


17-5 


8 27 


18-5 


7 34 


19-5 


6 2 


20-5 


5 22 



It is thus clear that variations of room temperature have a very great 
effect on the coagulation time, and it follows that methods in which the 



318 Addis 

temperature is not kept constant cannot yield comparative results which 
may be relied upon. 

3. The Contact of the Blood with the same Amount and Kind 
of Foreign Body in each Observation. 

A foreign body, as regards blood-coagulation, may be detined as any- 
thing which hastens or retards the coagulation time. 

Besides the intact endothelial lining of the vessels, there is only one 
class of substances which may be said not to act as foreign bodies, i.e. the 
oils. I do not think that the ordinary commercial mineral oils can, strictly 
speaking, be entirely excluded from classification as foreign bodies, for I 
found that the coagulation time of blood surrounded from the monient it 
issued from the wound by '" motor spirit " was less than the time taken 
when ordinary paraffin oil was used. Paraffin oil gave a time of from 70 
to 80 minutes, while with Pratt's Motor Spirit it was 50 to GO minutes. 
Nevertheless, their action as foreign bodies is so slight, in comparison 
with that of other substances, that they may be considered as having 
practically no effect at all. 

This fact may be utilised to estimate the effect of foreign bodies on the 
coagulation time. For, by placing a drop of blood in partial contact with a 
foreign body of constant nature and surrounding it elsewhere by oil, the 
complicating effect of otTier substances is excluded, and the resulting 
coagulation time gives an indication of the influence of that particular 
foreign body. 

In the following experiments coagulation was said to have occurred 
when a visible mass of fibrin was left after drawing off the fluid portion of 
the drop with filter paper. I found that, when blood was drawn under 
paraffin oil in a vessel lined with paraffin-wax, it took 70 to 80 minutes 
to coagulate. This, then, maj^ be taken as representing the coagulation 
time when the process is allowed to occur without the intervention of any 
foreign body. In this case coagulation is due simply to that amount of 
injury which, the blood receives in its passage from the wound, and to the 
thrombokinase added to it from the tissues of the wound. 

Drops of blood which were placed on clean glass slides and then 
innnersed, blood downwards, in oil, took from 20 to 26 minutes to coagulate. 
Here the only foreign body was glass. 

Other slides were coated with a smooth film of paraffin-wax and drops 
of blood placed on them. In this case the foreign body was air, and the 
coagulation time was 10 to 16 minutes. 

When the blood was exposed to both glass and air it coagulated in 
o minutes. These results are of course rough, for no attempt was made to 
keep the temperature constant ; but tliey were all done within 3 hours on the 
san»e afternoon, and they serve to indicate that different foreign bodies have 
different effects on the coagulation time. They also make it clear tiiat the 



The Coagulation Time of the Blood in Man 319 

ttiect of the wound on the tinie taken is slight in comparison with the 
fffeet of the environment of the blood after it has left the wound. Any 
method, therefore, in which the blood is not in contact with exactly the 
same amount and kind of foreign body is likely to give erroneous results. 

4. The End-point must be Clear and Definite and must always 
Indicate the same Degree of Coagulation. 

The various methods may be divided into two cla^sses, those in which 
the end-point adopted is the first appearance of fibrin, and those in whicli 
the evidence that coagulation has occurred is an indirect one deduced from 
some change in the behaviour of the blood to its surroundings 

Thus in Wright's (17), Burker's (21), Sabrazes' (22), and M'Gowan's 
(16) methods the coagulation time is the time elapsing from the issue of 
the blood from the wound to the first appearance of a thread of fibrin. 

In the second class various phenomena are taken as indicating coagula- 
tion. Thus in Hayera's (23), Milian's (24), and Brodie and Russell's (1) 
methods change in the contour of the drop of blood is the end-point. 
Biffi (25) and Kingston Fox (26) take the non-diffusibility of the blood 
when introduced into water. Vierordt's (27) method depends on the fact 
that when the blood has attained a certain degree of coagulation it no 
longer adheres to a horse-hair. The cessation of movement of the 
corpuscles in a film of blood under the action of gravity is the indication in 
Buckmaster's (28) method, and in the coagulometer which I have 
described there is a combination of the methods of both classes, for a 
fibrinous clot is seen, and the flow of the corpuscles is stopped. 

It might be thought that the recognition of a thread of fibrin was an 
absolutely certain proof that at least a certain amount of coagulation had 
occurred, but even this is not sure. Buckmaster (29) has noted that 
fibrin threads several millimetres in length may be drawn by a needle from 
a drop of blood within 10 seconds after it has issued from the wound. 
He takes this as an indication that coagulation is a gradual process which 
begins whenever the blood leaves the vessels. I have very often noticed 
the same thing when withdrawing the glass cone used in Brodie and 
Russell's method from a drop of blood which had just appeared, and yet 
the same drop will be found to have a normal coagulation time. It may be 
true that coagxilation begins at once, but I do not think that it gradually 
increases in amount up to entire coagulation, for it is a matter of common 
experience when large quantities of blood are used that it remains 
perfectly fluid for a time and then suddenly develops all the signs of 
coagulation. 

If, therefore, the fine fibrin filament which can be demonstrated so early 
indicates a certiiin amount of coagulation, it must remain very limited in 
extent until there is a sudden crisis of coagulation throughout the whole 
body of the blood. 



320 Addis 

On the other hand there is some reason to believe that the appearance 
of such a thread does not necessarilj'- imply coagulation at all, but maj'^ be 
a purely physical phenomenon. Mann, in his " Chemistry of the Proteids," 
1906, p. 382, says: "Ramsden, in a paper not yet published, states that 
fibrinogen solutions, free from fibrin ferment, can l)e made to yield mechani- 
cal surface aggi-egates " indistinguishable from typical fibrin, and that 
" fibrinogen, mechanically produced fibrin, and ferment-produced fibrin have 
the same heat-coagulation temperature, 53°-58°." 

There is, therefore, the possibility that this may be an instance of the 
mechanical production of fibrin. 

In the four methods which adopt the fibrin thread as their end -point 
the method of demonstration is by the slow withdrawal of glass from blood, 
just the circumstances under which Buck master has found this very early 
appearance of fibrin to occur. 

To adopt this as the end-point would appear, therefore, to be very 
fallacious, and to a certain extent it no doubt is so. In practice, however, 
there is very often an appearance not of a thread but of a considerable 
mass of fibrin, and there is no good reason to suppose that this is due to 
anything but fibrin-ferment coagulation. 

Nevertheless, even when a fibrin thread is taken to indicate the 
occurrence of coagulation, it is probable that it will not be very constant 
in the time of its appearance, for fibrin is primarily deposited in an amorph- 
ous and invisible condition, and it is only under the influence of mechanical 
stress that it acquires the appearance of threads or of a visible mass. These 
physical factors are impossible to control or keep constant. Possibly also 
there may be conditions of the blood which favour or retard the appearance 
of fibrin in a recognisable form. 

In the case of the indirect method of determining coagulation, various 
alterations in the behaviour of the blood to its surroundings have been 
adopted as end-points. All these are based on one change in the physical 
character of the blood, i.e. a loss of fluidity. 

When clear signs of this are seen, it is assumed that coagulation has 
occurred. This, however, is not always the case, for it is sometimes due 
simply to agglutination of the corpuscles. 

As has already been mentioned, the agglutinability of the blood is 
increased in disease, and for this reason methods which depend on an in- 
direct method of determining coagulation will probably give untrustworthy 
results in pathological conditions. 

Agglutination commences whenever the blood is drawn, and the longer 
it is left exposed to air the more marked does it become. This is well seen 
in using Brodie and Russell's method. If the blood after its introduction 
into the apparatus is not put in motion for a minute or two, agglutination 
may be so strong as to give all the signs which are considered to be conclusive 
of coagulation ; although, if movement of the corpuscles had been induced at 
once and repeated at short intervals, a coagulation time of 7 to 12 miniites 



The Coagulation Time of the Blood in Man 321 

would have been given. Agglutination is, then, a process which is entirely 
distinct from coagulation, which it may nevertheless closely simulate. 

No end-point which has yet been suggested can be considered free from 
fallacy. With the direct methods jBbrin may be inconstant in the time of 
its appearance or may possibly be sometimes due to purely physical causes, 
and with the indirect method agglutination may yield appearances usually 
considered characteristic of coagulation. 



III. Other Methods of estimating the Coagulation Time. 
Wright's Method. 

Wright (17) published his method in 1893. Since then it has undergone 
several slight changes. 

In the latest modification (1905) capillary glass tubes are calibrated by 
an ingenious method. They are filled with blood and placed in water at 
37" C. At intervals the blood is expressed from one after the other until 
fibrin is found. 

The method has probably been more widely used than any other. 
Murphy and Gould (30) compared Wright's (17) and Brodie and 
Russell's (1) methods. In 15 per cent, of estimations no result was arrived 
at with Wright's method. Ross (31) gives fourteen cases in which the 
coagulation time as shown by this method was diminished after calcium. 
He appears to think that the method is an accurate one. Coleman (32) 
preferred Brodie and Russell's (1) method. Solis-Cohen (33) made sixty- 
five observations. He says : " The results obtained were all practically 
negative, and were, moreover, unsatisfactory." He attributes this to fallacies 
in connection with the method. Douglass (34) estimated the coagulation 
time of normal, pregnant, and eclampsic women, but does not express 
any opinion as to the method. Nias (35) used it to show that strontium 
as well as calcium diminishes the coagulation time. He says : " In spite of 
criticisms which have appeared as to the sufficiency of this method, it has 
proved itself amply adequate for the purpose in hand, very consistent 
results having been obtained." Hinman and Sladen (36) think that " the 
pathological differences, and those dependent on technique, in this method, 
are of about the same relative value, which must confuse the results." 
Turner (37) made over one thousand observations on normal and epileptic 
people. His average coagulation time was 2 minutes 40 seconds, and he says 
that he often found differences of two minutes in tlie results of estimations 
of blood taken from the same individual by successive punctures. He attri- 
butes this to rapid variations in the coagulability of the blood and not to 
any deficiency in the method. 

I made forty-three estimations with the method as described in 1897, 
with the exception thiit each tube was filled from a separate puncture as is 
recommended in 1902. In nearly 50 per cent, of them only an approximate 



322 Addis 

time was arrived at, because coagulation occurred in one tube although it 
was not present in others until later. Considerable variations occurred in 
the coagulation time. These were, I think, largely due to differences of 
temperature. It is practically impossible to keep a small tin of water at 
a constant temperature simply by adding warm water at intervals in the 
manner recommended, especially since the observer's attention is fully 
occupied in filling and emptying the capillary tubes. 

The 1905 method is essentially the same, except that the tubes are kept 
at a temperature of 37° C. instead of 18"5° C. I do not think that this can 
be regarded as an improvement. 

With this apparatus Wright (17) sometimes obtained coagulation times 
as low as 30 seconds. I found that it took me nearly as long to fill the 
tube and introduce it into the water. With times so short any variation in 
the period during which the tubes are exposed to the room temperature 
must lead to considerable error. It is in connection with the end-point, 
however, that the most serious fallacy may be introduced. When the 
blood is expressed on to filter paper it is difficult to see the fibrin unless it 
is in large amount. In practice, therefore, one is apt to spread out the drop 
or to move the pipette over the filter paper while expressing the blood. It 
is here that the fallacy of the mechanical production of fibrin comes in, for 
in removing the tine tube from the blood the fibrin thread which is seen 
may not be due to coagulation at all. I think that the very low times 
recorded by Wright (17) after the administration of calcium are to be 
explained in this way. If the blood were not touched after it had been 
blown out, there would be no opportunity for this error. 



Blirker's Method. 

A drop of blood from the finger is allowed to fall into a drop of distilled 
water on a glass slide let into the lid of a box within which water at any 
desired temperature is kept. 

Fine rods of glass are passed through the mixture of blood and water 
every half minute until a thread or mass of fibrin is picked up. 

Biirker (21) states that the mixture of blood and water on the lid of 
the box very quickly arrives at the same temperature as the water beneath, 
and that variations in the temperature of the room do not materially afiect 
the temperature of the blood mixture. Even accepting this, there is still 
the difficulty of keeping the water at a constant temperature. I found that 
it was impossible to keep it from varying considerably. A thermostat 
would certainly be necessary for accurate work. Biirker (21) constructed 
a curve showing the effect of temperature on coagulation. This was 
compiled from only half-a-dozen observations, and as he later on described 
considerable daily variations in the coagulation time, it is difficult to see 
how his observations can be accepted as an accurate guide by which 
o"bseTvations conducted at different temperatures may be made comparable. 



The Coagulation Time of the Blood in Man 323 

Burker found that his results were untrustworthy unless he diluted 
the blood with water. The reason for this may be that the mechanical 
production of fibrin is prevented. Further, the end-point is not a good 
one to select, for sometimes a thread and sometimes a large mass of fibrin 
is demonstrated. In the latter case an earlier stage of coagulation has 
been missed. 

Sabraz^s' Method. 

This method depends on the fact that when blood coagulates in capillary 
glass tubes a fine thread of fibrin can be demonstrated when the tube is 
broken and the ends drawn slowly apart. The rest of the apparatus is 
intended to keep the tubes at a constant temperature. It consists of two 
superimposed glass boxes. The lower box contains warm water in winter 
and ice in summer. In the upper box the tubes are placed in close proximity 
to a thermometer. 

The end-point, as will be shown in speaking of the next method, is a 
good one, but the method is not accurate, because it is impossible to main- 
tain the temperature of the tubes constant throughout the experiment. A 
tube is taken out and broken across every half minute. To do so it is 
necessary to remove the lid, and the temperature inside at once tends to 
rise or fall according to the temperature of the air in the room. For this 
reason I found it was quite impossible to maintain anything approaching 
a constant temperature, and as a consequence the results obtained were 
very variable. Geneuil (38) has used this method, and has obtained very 
variable results in different pathological conditions. 

M'Gowan's Method. 

Glass tubes 1*5 mm. in diameter and 7 inches long are partially filled 
with blood. Portions are broken off every half minute until a thread of 
fibrin is seen. 

The author believes that for clinical work variations of temperature 
between 15° C. and 20° C. may be neglected, though he admits that to 
obtain accurate results some method of keeping the temperature constant 
would be necessary. He also suggests that in clinical work the coagulation 
time of the patient should be compared with the coagulation time of the 
observer's blood, which might be assumed to represent the normal. 

In testing this method it was at once apparent that it was useless even 
for rough clinical work unless some means were employed of keeping the 
temperature of the tubes constant. Thus when two tubes, both filled from 
the same drop of blood at the same moment, were kept at 15'5° C. and 
19° C. respectively, the coagulation times were 9 minutes 30 seconds and 5 
minutes 45 seconds. I have had an apparatus constructed which is fairly 
successful in maintaining a constant temperature, and with this addition 
I regard the method as more trustworthy than any of those hitherto used. 

In order to find whether the end-point adopted was to be depended on, 
VOL. I., NO. 4. — 1908. 22 



324 Addis 

I determined the coagulation time in tubes filled simultaneously and kept 
under the same conditions. 

Two hundred and fifty-six estimations were made, and the average 
difference in the coagulation time of each couple of comparable estimations 
was found to be 30 seconds. The slight variations which exist in the 
calibre of the tubes do not make any appreciable difference in the times, so 
that half a minute represents the average error due to the deficiency in the 
delicacy of the end-point. The average coagulation time was about 8 
minutes. This amount of error is, I think, very much less than obtains 
with the end-points adopted in other methods. 

Brodie and Russell's Method. 

The apparatus consists of a small circular chamber, floored with glass 
and roofed in above by a metal ring, into which an inverted truncated glass 
cone is fitted. A few drops of water are kept in the box, to lessen evapora- 
tion. Surrounding the sides of the air chamber is a water-jacket with inlet 
and outlet tubes for the circulation of water at a constant temperature. 

A small glass tube runs through the water-jacket and enters the air 
chamber. It is so directed that when air is blown through the draught 
impinges on the edge of the drop and causes the corpuscles to stream round. 
This is repeated at short intervals, and the movement watched under the 
microscope. Coagulation is considered to be present " as soon as a rim at 
the periphery is solid, and blowing simply indents this rim, without causing 
rotation." 

.In Bogg's(2) modification the water-jacket is discarded as unnecessary 
and a more convenient form of box is adopted. Pratt (15) did not use 
either the water-jacket or the glass cone, but simply directed a stream of 
air on to a drop hanging from a glass slide. 

Brodie and Russell (1) themselves give only a few coagulation times 
illustrating the effects of temperature. Pratt (15) found great variability 
in the coagulation time, but was unable to discover the factors which led to 
this. Murphy and Gould conclude that it is a more accurate method than 
Wright's (17). They say that in 5 per cent, of the caSes no result could 
be arrived at, because the blood would not flow at all. The times which 
they obtained were very variable. Coleman (32) obtained very consistent 
results by this method, working for the most part v/ith rabbit's blood. 
Hinman and Sladen (36), after using Vierordt's, Hayem's, Wright's 
and Milian's methods, came to the conclusion that Brodie and Russell's 
was the best " as having the greatest accuracy, combined with simplicity of 
technique." In 251 observations in health and disease, the limits were 
from 33 minutes to less than 3 minutes. 

Widely variable times were recorded even in consecutive observations on 
healthy people. 

I have made more than two hundred estimations with this method or 



Qwirterly Journal of Experimental Physiology, Vol. I., lOOd. 




T. Addis, " The Coagulation Time of the Blood in Man," Plate I. 



Quarterly Journal of Experimental Phiisiology, Vol. I., 190S. 




T. Addis, "The Coagulation Time of the Bloud in Man." Plate II. 



The Coagulation Time of the Blood in Man 325 

with Bogg's (2) modification of it. I found that with its use the coagula- 
tion time varied considerably in consecutive observations on the same 
person. Thus a time of 4 minutes might be followed by one of 12 minutes, 
and so on, without any regularity. Moreover, simultaneous observations 
with two similar coagulometers gave contradictory results. These dis- 
crepancies appear to have been due either to temperature variations or to 
the want of definiteness of the end-point, or to actual fallacies in connection 
with the end-point. 

The arrangement for the circulation round the air chamber of water 
;it a constant temperature is inadequate to keep the temperature of the air 
which surrounds the blood constant. Not only is the cone tkken off at the 
end of each observation, leaving the chamber in direct communication with 
the air of the room, but fresh air of unknown temperature is blown in 
every half minute or so when the cone is in position. The uselessness of 
the water-jacket is so apparent that, so far as I am aware, no one has used 
it except Brodie and Russell (1), who recorded the effect of running water 
at different temperatures through it. 

Another cause of error is the indefiniteness of the end-point. 
Hinman and SI ad en (36) point out that different observers have used 
three different stages as end-points. Thej" believe that a radial elastic 
movement is the most definite occurrence to fix on as indicating coagulation. 
I find, however, that it is very difficult to be sure of the exact time of occur- 
rence of this condition. Often it is found to be present quite early, but a 
blast of air stronger than usual causes it to disappear entirely. In these 
3ases it was due simply to agglutination of the corpuscles In cases where 
agglutination ia allowed to become advanced, as when the blood is not set 
in motion immediately after it is introduced into the apparatus, no amount of 
blowing is suflRcient to break it up. The blood appears to be coagulated, yet 
on removing the cone and examining it directly no coagulum can be found. 

1 believe that this end-point is due partly to tigglutination and partly 
to coagulation. If the blowing is done frequently and strongly, the amount 
of agglutination is slight or absent and the coagulation time is relatively 
long. If, on the other hand, the blowing is done at longer intervals and 
lightly, the end point is really due to agglutination, and the resulting time 
is short. 

Hayem's Method, 

Blood is received in a test-tube which is then tilted at intervals until the 
blood becomes so solid that the level no longer changes with each move- 
ment of the tube. Dastre and Floresco (39) and Brat (40) have worked 
with this method. Floresco thinks that it gives more accurate results 
than either Wright's or Vierordt's methods. Bezan(;on and Labbe 
(41 ) have used a slightly modified form of it. A great disadvantage is the 
considerable quantity of blood which is required. Moreover, Hay em 
appears to have made no attempt to exclude temperature variations. 



326 Addis 

Milian's Method. 

Drops of blood are allowed to fall on numbered glass slides. From 
time to time the slides are tilted and the contour of the drop watched. 
When the drop remains convex instead of sagging down, coagulation is 
assumed to have occurred. 

Hinman and Sladen (36) have modified this method by only including 
drops of uniform size. They recommend it as a " quick, convenient, and 
practical means of determining the coagulability of the blood," and state 
that it is thoroughly reliable. They hold, however, that " it is only marked 
differences in temperature which affect the time, which, as a rule, are not 
met with in the wards or in the laboratory." As I have already shown^ 
this is far from being true, and the great variations in the coagulation 
time as determined by this method are principally due to slight changes in 
temperature. 

Biffi's Method. 

Five loops such as are used for bacteriological work are made on a 
platinum wire which is fused into a glass rod. The rod passes through the 
cork of a jar half full of water. The wire is drawn through a drop of 
blood so that etich loop takes up a film of blood. At short intervals of 
time the films are successively pushed down into the water until one is 
reached which does not diffuse into the water but remains unaffected. This 
is taken as indicating coagulation. 

I found, in using this apparatus, that whenever the wire touches the 
water the blood in all the films becomes diluted. I was not able to find a 
way of obviating this difficulty. I have not seen any reports of work in 
which this method was used. 

Kingston Fox's Method. 

A series of calibrated capillary tubes are filled at intervals of half a 
minute with blood from the same wound. After some time a rubber nipple 
is applied and the blood expressed into water. The end-point is the time 
when the blood no longer diffuses in the water, but remains as a definite 
worm-like clot, " or until the contents have become so dense that they are 
with difficulty expiessod. This occurs, in some cases, apart from the forma- 
tion of a woiin-like clot, the mass being partially diffusible in the water." 

(Coagulation times taken at five different temperatures are given in order 
that corrections to the standard temperature of 60° F. may bt- made. No 
formula is given foi- this reduction, which " must therefore be made from 
the diagram graphically." 

Vierordt's Metliod. 

This was the first ciinlcal nujthod for determining the coagulation time. 
A graduated glass tube with a bore of I mm. is filled w^ith blood up to the 



The Coagulation Time of the Blood in Man 327 

') cm. mark. A carefully prepared white horse-hair is then introduced into 
llie tube, so that a segment of it is surrounded by blood. At intervals of time 
the hair is pulled a short distance out of the tube. As coagulation advances 
small masses of fibrin and red blood corpuscles will be found sticking to the 
hair, but later still the blood becomes so solid that none adheres to it. 
This is taken as the end-point, 

V^ierordt (27) fully recognised the importance of a uniform method of 
obtaining the blood, and the necessity of always having the same amount 
of contact with foreign bodies. He also mentions that the temperature 
should, as far as possible, be constant. - 

He made a large number of estimations, principally on himself, and he 
gave the first list of coagulation times in disease. 

His results were very variable. The limits in the case of his own blood 
were from 3^ to 17^ minutes, and he was unable to establish any regularity 
in the fluctuations. These must have been due mainly to changes of 
temperature and to want of reliability in the end-point. 

Buckmaster's Method. 

The method is based on the fact that, when a loop of wire is drawn 
through a drop of blood so that a film of blood is taken up by it, the 
corpuscle.^ can be observed by the aid of a lens to flow in response to 
changes in the position of the loop. Thus when it is held vertically they 
can be seen to fall slowly downwards until a clear circle of plasma is left 
at the upper pole, while in the horizontal position tliey spread themselves 
evenly over the film. As time goes on the flow of the corpuscles becomes 
slower and more impeded until a moment arrives at which no movement 
can be seen. The blood is then considered to have coagulated. 

The apparatus is devised with the object of keeping the film of blood at 
a constant temperature. It consists of a box, the lower part of which is 
filled with water. The film is in an inner chamber. Panes of glass are 
fitted in the sides so that it is possible to sec the blood when it is in position. 

When the wire with its film of blood has been introduced all openings 
are closed, and the film can be rotated from outside the box. The water is 
warmed from below, and a thermometer whose bulb is in the inner chamber 
records the temperature of the air surrounding the blood. 

This is an extremely original and ingenious method. No work has as 
yet, so far as I know, been done with it. Buck master (28) himself gives 
only a few times taken at slightly different temperatures. 

In attempting to make a series of observations by it at the same tem- 
perature, I found that temperature variations were apt to occur even during 
the course of an experiment, while it was altogether impracticable, without 
tedious waiting and manipulation, to get a series of times taken at a 
constant temperature. This is not remarkable when it is remembered that 
whenever the wire loop is removed for cleaning and filling with a fresh 



328 



Addis 



film, a JsTge communication exists between the inner chamber and the air 
of the room. The temperatures at which Buckmaster (28) worked varied 
between 30° and 40° C. This is so much removed from room temperature 
that there must necessarily be a rapid fall whenever the loop is removed. 
When the opening is closed again by the reposition of the wire the tempera- 
ture rises again. 

I further found it impossible to heat the water with a spirit-flame o;- a 
gas-jet so as to keep the temperature constant, and the effect of an access 
of heat to the water is not at oace apparent in the inner chamber where 
th'^ thermometer is placed, so that one is constantly at fault in attempts at 
regulation. 

By making a considerable number of observations and picking out those 
which chanced to have been taken at the same temperature, some idea oi 
the amount of variation due to other causes than temperature was obtained. 



Tempera- 


Coagulation times in minutes 


i 


! ture. 


and seconds. 




27° C. 


1 
7 30 ! 5 30 


7 15 


7 30 


9 15 


6 15 


i 29° C. 


4 15 1 5 15 


6 


5 5 






1 31° C. 


3 30 4 


3 10 


3 5 






! 35° C. 


4 15 4 45 


4 


4 10 


3 15 


4 15 


: 36° C. 


4 , 3 25 


4 10 


3 15 


3 




37° C. 


5 


5 15 


5 30 


4 30 


2 30 




1 38° C. 


5 


3 30 


4 35 


4 15 


3 40 




39° C. 


3 15 


5 


3 50 ! 4 25 

i 


... 





There are thus considerable variations even when the temperature is 
constant. 

I think that this inconstancy in my results was due to differences ol 
thickness of film. After the loop is filled with blood a shake is given to it 
in order to detach part of the blood- Unless this is done the film is so 
opaque that it is difficult to see the streaming of the corpuscles. I did not 
nay particular attention to the amount I shook off, and some of the films 
were nnich thicker than others. 

When very thick and very thin films are compared at low temperatures 
so as to give long coagulation times, there was found to be an enormous 
difference in the time. 



Tempera- 
ture. 


Thin film. 


Thick film. 


°G. 

11-5 

12-5 

13 

14 


rain. sec. 
7 30 
7 
5 30 
3 15 


min. sec. 
21 20 
17 
17 
13 45 



The Coagulation Time of the Blood in Man 329 

At higher temperatures the effect is less marked but still distinct. 



Tempera- 
ture. 


Thin film. 


Thick film. 


"C. 


min. sec. 


min. sec. 


29 


4 15 


6 


32 


3 35 


4 20 


34 


2 5 


3 30 


37 


5 15 


4 30 


38 


4 15 





If the temperature of the water were kept constant by a gas regulator 
and some means adopted for preventing the inflow of air at room tempera- 
ture when the wire is removed, the method would be a good one, for no 
doubt with practice it would be possible to shake off the extra quantity of 
blood So as to get films of approximately equal thickness. 

IV. The Effect of Low and High Temperatures on the 
Coagulation Time of the Blood. 
The low temperatures were obtained by adding ice to the water in the 
tank. The following is a list of coagulation times taken at temperatures 
from 3-26 to 51 5° 0. :— 



Temperature. 


Coagulation tirae.s. 




min. sec. 


3-25 C. 


63 20 


7-25 


32 45 


10-25 


21 30 


12-25 


16 30 


13-5 


14 32 


14-5 


12 58 


15-5 


11 46 


16-5 


10 10 


17-6 


a 27 


18-5 


7 34 


19^5 


6 2 


20-5 


5 22 


24-0 


5 


260 


3 40 


28-0 


2 55 


30-0 


2 35 


32 


2 15 


34-0 


1 40 


36-0 


1 25 


38-0 


1 30 


40-5 


1 30 


42-5 


L 40 


44-5 


2 25 


45-5 


2 55 


46-3 


3 30 


47-26 


3 10 


48-25 


3 30 


49-6 


3 30 


51-5 


5 15 


53-5 


? 



330 Addis 

When the results are plotted out so as to form a curve it can be seen 
that the effect of variation of temperature is much more marked at the 
lower temperatures. Thus from 325 to 7 25° C. there is a diminution of 
the time of over 30 minutes, while from 725 to 1225'' C. there is a difference 
of only 16^ minutes. From 12-25° C. there is at first a fall of about 2 minutes 
for every degree of temperature, but this diminishes until at about 28° C. 
it is only half a minute. From this point the periods beoome very slowly 
shorter until 36° C. is reached, when the blood coagulates in 1 minute 
25 seconds. Above 36° C. they begin to increase slowly up to 51-5° C, at 
which temperature the blood takes 5^ minute.s to coagulate. 

At temperatures below 10° C. a tendency for the red blood corpuscles 
to become agglutinated became apparent, so that twice the usual pressure 
of oil was necessary to keep the corpuscles flowing separately from 
each other. 

Above 40° C. the same phenomenon occurred, and became progressively 
more marked as the temperature rose. Above 5 15 a regular flow could 
not be induced. At 535° C. there was no evidence of any clot after 
6 minutes, but the movement was very slow and irregular. At 56° C, at 
which temperature fibrinogen would be heat-coagulated, the blood was not 
moved even by the highest pressures, but collected in whorls and clumps 
of agglutinated corpuscles. No change was apparent in their shape or size. 

V. The Absence of Diurnal Variation in the Coagulation Time. 

Vierordt (27) tried to determine whether there were any regular diurnal 
fluctuations in the coagulation, time but was unable to show directly that 
these existed, although he found differences in the averages of the observa- 
tions taken at different periods of the day. 

Btirker (21) has stated, as a result of two-hourly observations for three 
days, that the coagulation time is longest in the morning and diminishes 
until it reaches its lowest point about 2 o'clock p.m., after which it rises 
again. This statement is not. however, quite borne out by his figures. 

Coleman (32) says that the time is shortest in the morning and longest 
in the early hours of the afternoon. 

Hinman and vSladen (36) do not come to any definite conclusion, but 
are inclined to think that the time is longest in the morning. 

The fact that diurnal variations exist has never been called in question, 
simply because it was often found that different coagulation times were 
obtained at different periods of the day. This was attributed to time 
differences in the coagulability of the blood instead of deficiencies in the 
technique of the method which was used. 

When working with M'Gowan's (16) method I noticed that the more 
temperature variations were excluded the smaller was the amount of 
difierence in the coagulation times, and that fluctuations in the coagulation 
curves, which I had formerly considered to be indicative of intrinsic 



The C'o.i^julation Time of the Blood in Man 381 

changes in coaoulability, were really due to variations in tlu' surroundinof 
temperature. 

With tlie method described in Section I., in which tenijxTatuvt' xaiiation.-, 
are entirely excluded, the coagulation time was found to be lemarkably 
constant at all times of the day and night. The slight variati<jns which 
occur are irregular and are due to experimental error. 

For twenty days the coagulation time of my own l)lood and that of 
certain other persons was taken everj' hour or two, and in no one of the 
charts prepared from these observations do any variations occur which are 
outside the limit of experimental erroi-. 

The following figures are the averages of more than three hundred 
observations taken at diffVrent hours of the dav, at a tempt rature of 
ISTy" C. :— 

Time of day. A\ iia^'e coagulation time, 
mill. sec. 



8-9 a-ni. 


7 36 


9-10 ,. 


7 50 


10-11 „ 


7 24 


11-12 noon 


7 37 


12-1 p.ni. 


7 46 


1-2 .. 


7 28 


2-3 ,. 


7 49 


3-4 ., 


7 41 


4-5 „■ 


7 50 


5-6 ., 


7 54 



The assimilation of food has been described as a cause of variation in 
the coagulation time by Biirker (21), Coleman (32), and others. 

The coagulation time was estimated on ten different occasions both 
before and after breakfast. The average time before food was 7 minutes 
47 seconds, and the average after was 7 minutes 46 seconds. 

VI The Effect of the Administration of Calcium and Citric 
Acid by the Mouth on the Coagulation Time. 

The details of work done in this connection will be published elsewhere 
(42\ The general conclusion arrived at is that calcium and citric acid, when 
given by the mouth, have no influence on the coagulation time. This is 
not due to entire non-absorption, but to the fact that the change they 
produce on the calcium content of the blood is so slight that no appreciable 
effect is produced. 

VII. Conclusions. 

1. There are four conditions which must be fulfilled by any method 
which is to yield results which can be relied upon, (a) There must be a 
uniform method of obtaining the blood, (b) The temperature must be 
the same during each experiment. (c) The amount of contact with 
foreign bodies must be always the same. (d) The end-point must be 



332 Addis 

clear abd definite and such as always to indicate the same degree of 
coagulation. 

2. A method is described which conforms to these conditions. 

3. Other methods which have been employed are reviewed, and, 
where necessary, the result of work undertaken to test their accuracy 
is given. 

4. The general conclusion of this testing is that no one of these methods 
fulfils all the conditions mentioned above, and that in particular all fail 
to comply with the essential condition that in comparative observations 
exactly the same temperature must be maintained throughout. 

5. The effect of variations of temperature is described, and it is shown 
that at about the normal temperature of the body the coagulation time is 
shortest, becoming gradually longer at temperatures above 40"" C. and 
below 36^' C. 

6. The coagulation time is constant for the same individual at difFejcent 
times of the day, and even on different days. The daily variations which have 
been described do not exist, but are due to fallacies in connection with the 
methods which have been employed. 

7. The conclusion that calcium and citric acid, when administered by 
the mouth, have no effect on the coagulation time, is referred to. 

I wish to express my thanks to Professor Schafer and to Dr Cramer 
for advice and criticism during the progress of these investigations. 

The expenses entailed have been defrayed by a grant from the Moray 
Fuud of the University of P^dinburgh. 



REFERENCES. 

(1) Bbodie and Russell, Journal of Physiology, 1897, xxi. p. 403. 

(2) BoGG, Deutsches Arch. f. klin. Med., Leipzig, 1904, Ixxix., S. 539; 
International Clinics, vol. i., 18th series, p. 31. 

(3) LoEB, New York Medical News, April 1, 1 905. 

(4) Fleming, Practitioner, vol. Ixxx.^ No. 5, 1908, p. 614. 

(5) Hbktoen, Journal of Infectious Diseases, vol. iv.. No. 3, p. 297. 

(6) Pbarce and Winne, Amer. Journ. Med. Science, 1904, vol. cxxviii. p. 669. 

(7) Pearce, Journ. Med. Research, 1904 (12), p. 316; Journ. Exper. Medicine,. 
1906 (8), p. 64 ; Journ. Med. Research, 1906 (14), p. 541. 

(8) Hueter, Deutsche Zeitschr. f. Cliirurgie, 1873-1874, vol. iv., S. 105, 330. 

(9) Flexner, Univ. Penn. Med. Bull, 1902 (15), p. 324; Journ. Med. Research, 
1902 (8), p. 316; Amer. Journ. Med. Science, 1903, vol. cxxvi. p. 202; Amer. 
Journ. Med. Science, 1904, vol. cxxviii. 

(10) BoxMEYBR, Journ. Med. Research, 1903 (9), p. 146. 

(11) Kraus u. Ludwig, Wiener klin. Wochenschr., 1902, Bd. xv., S. 120, 382. 

(12) VoLK u. LipschOtz, Wiener klin. Wochenschr., 1903, Bd. xvi., S. 1894. 



The Coagulation Time of the Blood in Man 333 

(13) Kayser, Zeitschr. f. Hygiene, 1903, Bd. xlii., S. 118. 

(14) EiSENBERG, Centralb. f. Bakter., 1903, Bd. xxxiv., p. 739. 

(15) Pratt, Journ. Med. Research, 1903, vol. cxx. 

(16) M'GowAN, Brit. Med. Journ., 1907, ii. 

(17) Wright, A. E., Brit. Med. Journ., 1893, ii. 223; 1894, i. 237; 18^)4, 
ii, 57. Lancet, 1896, i. 153; 1896, ii. 807; 1897, i. 303; 1902, ii. 11; 1905, 
ii. 1104. 

(18) Wright and Knapp, Lancet, 1902, ii. 1460. 

(19) Wright and Paramore, Lancet, 1905, ii. 1096. 

(20) Wright and Ross, Lancet, 1905, ii. 1164. 

(21) BiJRKER, Arcli. f. d. ges. Physiol., Bd.cii., 1904, p. 67 ; and Bd cxviii., 1907. 

(22) SABRAzi;s, Fol. Haemat., 1904, p. 394, and 1906, p. 432. 

(23) Hayem, Du sang et de ses alterations anatomiques, Paris, 1889, p. 323. 

(24) MiLiAN, Bull, et mem. soc. med. d'hop., Paris, 1901 ; Presse M^d., 1904, 
i. 202. 

(25) BiFFi, La cronica medica, Junio 1904. 

(26) Fox, Lancet, Jan. 11, 1908. 

(27) ViERORDT, Archiv d. Heilk., 1878, xix., p. 193. 

(28) BucKMASTKR, Morphology of Normal and Pathological Blood, 1906. 

(29) BucKMASTER, Scicnce Progress, 1907. 

(30) Murphy and Gould, Boston Med. and Surg. Journ., J 904. 

(31) Ross, Lancet, 1906, i. 143. 

(32) Coleman, Bio-Chemical Journal, 1907, vol ii.. No. 4. 

(33) SoLis-CoHEN, Univ. Penn. Med. Bull., June 1907, p. 56. 

(34) Douglass, Brit. Med. Journ., 1904, p. 709. 

(35) NiAS, Lancet, 1906, ii. 436, and Jan. 1908. 

(36) HiNMAN and Sladbn, Johns Hopkins Hosp. Bull-., 1907, June-July, p. 207, 

(37) Turner, Journal of Mental Science, Jan. and Oct. 1907. 

(38) Geneuil, M^thodes pour determiner le d^but de la coagulation du sang, 
Bordeaux, 1906. 

(39) Dastre et Floresco, Arch de physiol. norm, et path., 1396, viii. 402. 

(40) Brat, Berl. klin. Wochenschr., 1902, xxxix. 1146. 

(41) Bbzan^on et Labbk, Traite d'hematologie, Paris, 1904, p. 53. 

(42) Addis, T., Quarterly Journal of Medicine, January 1, 1909. 



DESCRIPTION OF PLATES. 
Plate I. 

P, large glass bottle containing paraffin oil suspended from an upright, so that 
it can be raised or lowered ; Sc, scale on upright ; p,- flexible metal tubing leading 
the oil from the glass bottle to the spiral (sp) contained within the water vessel (V) ; 
/, metal tube co;itinuous with the spiral, and emerging from the water tank at the 
level of the microscope stage; ch, Bogg's coagulometer chamber, with thermometer 
(th) inserted into it ; r, reservoir into which the overflow of oil from the stage is 



334 The Coagulation Time of the Blood in Man 

received ; ?<;, waste tube by which the oil from r is removed ; R, Schiif er's thermostat 
(gas regulator); g, rubber tube from gas-tap; g, rubber tube from thermostat to 
gas-burner under tlie water vessel. 

Plate II. 

This gives a closer view of the stage apparatus. The lid of the Bogg's coagulo- 
meter {ch') has been taken off, and the glass cone with its truncated apex, from 
which the drop of blood hangs, is shown. 



I 



THE ACTION OF TOBACCO SMOKE, WITH SPECIAL REFERENCE 
TO ARTERIAL PRESSURE AND DEGENERATION By W. 
Emerson Lee. (From the Pharmacological Laboratory, Cambridge.) 
(With fourteen figures in the text and one Plate.) 

(Received for publication 28lh August 1908.) 



CONTENTS. 

I. Previous History 
II. The Composition of Tobaccos and Tobacco Smoke 
III. The Relative Action of the Constituents of Tobacco Smoke :— 
L On plain muscle . 
ii. On the isolated heart 
iii. On isolated vessels 
iv. On the intact animal 

(a) Central nervous system 

(b) Circulatory system 
IV. The Relative Toxicity of the Various Const 

Smoke 
V. The Effect of Smoking on Man . 
VI. The Effect of Smoking on Animals :- 

(a) Immediate effects 

(b) Remote effects 
VII. General Conclusions 



ituents II* Tobacco 



PAOK 

335 
336 



337 
338 
339 
341 
341 
342 

342 
344 

352 
356 
357 



I. Previous History. 

The following investigations were conducted witli a view to determine, 
first, the action of tobacco smoke apart from its various constituents, and 
second, whether smoking may cause aitoiial di^goncration. 

It is not necessary to refer to papers dealing with the physiological action 
of nicotine; a complete bibHography is attached to Langley's paper (1). 
The toxic effects of nicotine on man also re<juire little notice ; most of the 
important work on this subji.ct is found in Allbutt's " System of Medicine." 
The effect of tobacco smoke has received little or no attention from the 
experimental standpoint. 



II. Composition of Tobaccos and Tobacco Smoke. 

The composition of tobacco smoke, obtained by an aspirator from the 
slow combustion of 100 grams of tobacco, wjis as follows: — 

Nicotine, 1165 grams. This represented 50 per cent, of the total 
nicotine present before conibustion. 

Pyridine bases, 0146 g. Chiefly pyridine and collidine, the former 



336 Lee 

being produced during the destruction of some of the nicotine, the latter 
from the combustion of the fibres in the tobacco. 

Hydrocyanic acid, 008 g. 

Ammonia, 036 g. 

Carbon tnonoxide, 410 c.c. 

These amounts vary with many factors. Thus the length of the tube 
through which the smoke passes — by allowing the deposition of the solid 
matter and the condensation of vapour — materially affects the composition 
of the smoke J the principle of this is illustrated in the "churchwarden" 
pipe. Again, the quality of the tobacco varies within the widest limits; 
evidence will be produced on this point later. For the purposes of experi- 
ment some standard tobacco must be adopted and retained throughout. 
The tobacco which was chosen for these experiments was of two varieties : 
— 1. A sample of Virginian tobacco from the " untreated " leaf, prepared 
for smoking in cigarette form; this was kindly provided by Mr Player. 
2. A very strong variety of Manilla cigar. 

In order to determine the amount of nicotine or blood-pressure-raising 
substances present in these two tobaccos, equal weights were taken and 
macerated, each in the same quantity of normal saline solution. The 
solutions were then filtered, and the amount of nicotine or blood-pressure- 
raising substances estimated physiologically by their power of raising blood- 
pressure. For this purpose cats were pithed or anaesthetised with urethane, 
and the blood-pressure recorded by a mercurial manometer. The following 
figures show the result of these experiments : — 

One gram of cigarette tobacco and 1 gram of Manilla tobacco were 
macerated each in 100 c.c. of saline and allowed to stand for two days. 
The infusion resulting was filtered, and a fluid suitable for intravenous 
injection was obtained. 

Equal quantities of the two solutions were injected into the jugular 
vein of a pithed cat, and the relative rise of blood-pressure when the amount 
injected was not excessive was as follows : — 

Manilla tobacco, 25*6 mm. Hg. 
Cigarette „ 42-4 

A solution of the crude leaf from which this sample of cigarette 
tobacco was prepared (1 gram to 100 c.c.) produced a rise of 56 mm. Hg. 
The last result shows that during the course of preparation a considerable 
quantity of nicotine or pressor substance is destroyed. This is not surpris- 
ing, as it is well known, that fermentation reduces the amount of nicotine 
in tobacco. From these experiments it is obvious that the Manilla tobacco 
contains much less nicotine than the Virginian. It does not, however, 
necessarily follow that because one tobacco contains less nicotine than 
another, it will yield less nicotine when it is smoked. For this reason a 
second series of experiments was conducted by drawing the fumes from the 



Action of Tobacco Smoke 337 

combustion of the tobaccos through saline solution by means of a suction 
pump, 1 gram of each variety of tobacco being used with 100 c.c. of saline. 
The stroke of the pump was so arranged that the smoke was drawn through 
the saline about twelve times a minute. The procedures with the two 
tobaccos were conducted under identical conditions, and so arranged that 
the combustion of the two tobaccos was effected in equal times. The 
relative amounts of nicotine w6re determined as before, by coTnparing the 
rise in blood -pressure produced when the solutions were injected into a 
pithed cat. The results showed that when the smoke solution from the 
Manilla tobacco caused a rise of 2 mm. Hg, the smoke solution from the 
cigarette tobacco caused a rise of 1 mm. Hg. 

From these experiments the remarkable fact comes out that, whilst the 
Virginian tobacco contains a much greater percentage of nicotine than 
the Manilla, yet, after combustion, the smoke from the Manilla contains 
considerably the larger percentage. This circumstance is explained as 
follows : — During the slow combustion of a cigar, as in ordinary smoking, 
immediately behind the point of combustion is an area in which the water 
and other volatile substances in the tobacco condense ; during the act of 
smoking the greater portion of the nicotine at the seat of combustion is 
destroyed (50 per cent.), and the nicotine which finds its way into the mouth 
of the smoker is probably derived from the hot gases passing through the 
moist area and volatilising certain of the more volatile principles of the 
tobacco, of which nicotine is certainly one. So that the smaller the moist 
area behind the point of combustion, the less likely is the smoke to contain 
volatile toxic bodies. It will be immediately suggested that a thin cigai- 
or a cigarette will yield fewer of these products than a thick cigar, for the 
thin cigar or cigarette obviously permits a relatively greater evaporation to 
take place. Moreover, if a thick cigar be unrolled and made up in a thinner 
body, the percentage of nicotine destroyed during combustion is increased. 
The experience of many smokers also agrees with this hypothesis, for there 
are those who will always avoid a thick cigar because, whatever be the 
strength of the leaf from which it is made, unpleasant symptoms are 
invariably experienced. 



III. The Relative Action of the Constituents of Tobacco S\iOkE. 

It has been pointed out already that the important constituents of 
tobacco smoke are nicotine and certain pyridine bases, including especially 
pyridine itself and collidine. The following experiments were conducted 
with a view to determine the relative effects of these alkaloids. 

i. Plain Muscle. — The action on plain muscle was determined first by 
means of " ring " preparations of the frog's stomach. These were suspended 
in Ringer's solution and arranged to record on a slowly moving drum by 
means of suitably weighted levers. 



388 Leo 

A solution ol" follitline, 1 in J 000 in saline, was applied to such a piepaia- 
tion; the movements were promptl}'^ inhibited and the tonus diminished. 
Nicotine, 1 in 1000, was then applied; the muscle entered into tonic 
contraction, the waves for the time ceasinor, but it soon regained its normal 
activity. Pyridine, 1 in 1000, was administered and produced hardly any 
result; collidine was applied once more, when it again produced its typical 
effect. 

It is clear from this that nicotine and collidine aci in opposite directions : 
the former causes the muscle to increase in tonus, the latter inhibits move- 
ments and causes it to relax : pyridine, in these doses, is almost without 
action. 

ii. The Heart. — It has been pointed out that experiments made by 
dropping drugs on the frog's heart are not likely to lead to valuable results, 
since in many cases the action differs from that obtained when the drug 



Time. 




Fig. J. — Record of movements of an isolated frog's heart perfused through the hepatic vein. 
A sliow,-; the elTect of 01 per cent, nitotiiu'. B shows tlie effect of 01 i>er cent, colliiline. Time. 30 secnmls. 

passes through the circulation of the heart. The following experiments on 
the isolated frog's heart were therefore performed by perfusing Ringei's 
solution through one hepatic vein and allowing it to escape by the aorta, 
the heart-beat being recorded by the suspension method. A solution 
of I in 10,000 nicotine raised the tonus of the heart muscle and slightly 
quickened the beat. The same strength of pyridine produced no eifect, 
while tlu^ same strength of collidine produced some slowing and a fall 
in tonus. A more concentrated solution of nicotine (1 in 1000) always 
causes the heart to enter into very marked tonus. The effect of pyridine 
(I in 1000) slightly weakens the beat of the heart; it never produces 
a rise of tonus. 

From such experiments it is sixown that nicotine has much the most 
toxic effect on the heart, and that pyridine has the least. The effect of 
collidine compared with nicotine is shown in fig. 1. It will be noticed in 
these experiments, which were performed under identical conditions, that 
the nicotine (A) quickens the heart at first, but later causes some slowing • 



Action of Tobacco Smoke 339 

these effects of nicotine, we know, are due to excitation of the ganglion- 
cells. The collidine effect bears a superficial resemblance to that of nicotine, 
except that the inhibition is more pronounced in the case of collidine (B) ; 
but there is one important difference in the complete absence of rise of 
tonus. 

The action on the isolated mammal's heart was determined by perfusing 
the hearts of rabbits with Ringer's solution, by means of a modified Langeu- 
dorff apparatus. 

At A, in fig. 2, 2 ac. of a 1 per cent, solution of pyridine were injected 
by means of the lateral tube, but on its reaching the heart only a very 
small effect was noticed — there was no alteration in tonus, but there was 
evidence of some quickening of the beats. After a period of 20 minutes' 
perfusion with Ringer's solution, during which time the heart's action had 
become quite normal, 2 c.c. of 1 per cent, solution of collidine were injected 
at B, at the same rate as before,'»into the perfusing fluid. The heart was 
immediately inhibited in diastole ; but, after the drug had passed through, 
recovery grad\ially ensued until the normal ihythm w.as regained, no 
permanent depression resulted. After a further periotl of 20 minutes' rest, 
2 c.c. of a 1 per cent, solution of nicotine were injected (at C) ; the strength 
of the heart was immediately increased, the heart-beat was accelerated, and 
the tonus was gradually raised ; indeed, this gradual increase of tonus with 
nicotine is a most characteristic effect, and affords a marked distinction 
between the action of nicotine and the other constituents of tobacco smoke. 

The periods of inhibition found in the tracing, following injection of 
nicotine, are due to excitation of the intracardiac ganglion-cells ; they are 
not seen in the atropinised animal. 

I now propose to compare the action of a solution of nicotine with a 
solution through which tobacco smoke has been drawn. For this purpose 
the smoke from 1 gram of tobacco, slowly burnt, was drawn slowly through 
100 c.c. of saline solution. The isolated rabbit's heart was again used ; 
fig. 3 illustrates the two effects, at A the effect of injecting the smoke 
solution being shown, and at B the effect of nicotine. The first effect (A) 
shows initial inhibition followed by accelerati'ni of tiie heart and some 
increase in strength of the beat : there is no rise in tonus. B shows 
an almost identical result, but with the difference that the heart does 
not relax properly in diastole, so that the diastolic tone gradually rises. 
We know that the pyridine bases, especially collidine, reduce t^.mus in 
muscle, and it seems possible that while the inhibition and subsequent 
acceleration shown at A may be due to nicotine, the absence of increased 
tonus may result from the antagonistic action of the pyridine bases to 
nicotine. 

iii. Some Effects of Perfusion. — The blood-vessels of the frog were 
perfused through the right innominate artery, and the outflow from the 
veins determined by placing the frog in a funnel and allowing the drops to 
fall upon a lever that recorded on a slowly moving drum. 

VOL. I.. NO. 4. — 1908. 23 



340 



Lee 



Signal 



Signal. 




Signal. 



fiG. 2.— Isolated rabbit's heart perfused with Ringer's solution by the method of Langendorfl'. 

A shows the effect of injecting into the lateral tube 2 c.c. 1 per cent, pyridine ; B, injection Of 2 c c 

1 per cent, collidine ; C, injection of 2 c.c. 1 per cent, nicotine. Time, 1 cm. =10 seconds. 



i 



Action of Tobacco Smoke 



:341 



Ringer's solution was employed as the perfusing fluid. All three drugs 
reduce the flow through the vessels if in sufficient concentration, but this 
result is not obtained with dilute solutions such as 1 in 10,000. It is 
known that nicotine solutions of this dilution constrict blood-vessels in an 
intact animal to a very decided degree ; all that these experiments there- 
fore show is that nicotine in dilute solutions does not act directly on the 
vessel wall. 

Collidine and pyridine, in doses which one may regard as possible in 
man, have practically no action upon the blood-vessels. 



Signal 




Signal 



Fig. 3. — Isolated rabbit's heuit perfused b) La nge iid or f f ' s method. 

A shows the effect of injecting into the lateral tube 3 c.o. of the smoke solution, and U, 2 c.c. 
of 1 per cent, solution of nicotine. Time, 1 cm. = 10 seconds. 



iv. The Eflect of Intravenous Injection of the Constituents of 
Tobacco Smoke, (a) On the Spinal Cord. — Small doses of these drugs 
injected directly into the circulation of rabbits excite the spinal cord, and 
if the doses be sufficiently large, convulsions are produced. These convulsive 
movements diff"er from those produced by strychnine in that there is no 
antecedent tonic stage ; the animal passes into a condition of anaesthesia 
associated witli clonic contractions of all the muscles supplied by the spinal 
cord. This effect is much the most marked with nicotine ; it is very small 
in the ca.ses of collidine and pyridine. That the convulsions are spinal in 
origin may be shown by painting the spinal cord of a decapitated frog with 
the drug, when the muscles supplied by that part will be observed to 



342 Lee 

twitch. This has been aheady shown and commented on by Dixon (3;. 
Moreover, these flrut^s differ in their action from strychnine in that, wliilst 
this drug affects the sensory cells so that the convulsions can onl}' l)e pro- 
duced by an appropriate afierent stimulus, the drugs under C(jnsideration act 
on the motor cells. This is proved, first, because on being applied to the 
cord they produce immediate twitchings of the muscles which are supplied 
by that section of the cord ; second, because they are not reflex in origin : 
and third, because the twitchings are limited to the cells affected. That is 
to say, str3^chnine applied to one small portion of the cord produces con- 
vulsions over the whole body ; but nicotine applied in the same way causes 
twitchings only in the muscles supplied by the corresponding part of the 
affected cord. 

(b) On the Circulatory System. — Pyridine produces remarkably little 
effect ; in fig 4, A, 5 c.c. of a 1 per cent, solution were slowly injected and 
produced practically no alteration in the blood-pressure, nor in the general 
condition of the circulation. Collidine, however, in small doses cause,s 
considerable dilatation of the blood-vessels, and a corresponding fall in 
blood -pressure. This is shown in fig. 4, B, in which the upper tracing 
represents intestinal volume, and the lower the blood-pressure. It will be 
noticed that as the vessels dilate the pressure falls. In this case, however, 
there can be no doubt that some of the fall of pressure is due to cardiac 
depression. Larger doses of collidine weaken the heart, and consequently 
lower the blood-pressure to such an extent that the intestinal vessels, instead 
of filling with blood, shrink, secondarily to the fall of blood-pressure. 

The effect of nicotine is shown in fig. 4, C, for the sake of comparison. 
It also lowers the blood-pressure, as shown in the second (detached) part of 
the tracing, but only after an initial rise, I shall have occasion later, in 
dealing with the action of tobacco smoke on man, to refer again to this well- 
known rise followed by a fall in blood-pressure. 

IV. Toxicity a.s estimated by Injection into Intact Animals. 

The relative toxicity of pyridine, collidine, and nicotine was estimated 
by determining the minimal lethal doses in frogs. 

Three frogs, each within a fraction of 23 grams in weight, were 
injected with solutions containing a quarter of a grain of each drug 
respectively. The injections were made into the dorsal lymph sac. 

During the two hours after injecting pyridine the animal was very 
active, the pupils were dilated, and respirator}^ efforts were increased; there 
was no paralysis. 

The second frog, after collidine, became paralysed in two minutes, 
respiration almost ceased, the pupils were widely dilated, and reflexes were 
entirely absent. After a lapse of twenty minutes feeble respiratory 
movements became evident, and the animal gradually recovered. 

Nicotine was much the most toxic. The frog died about two minutes 



Action of Tobacco Smoke 



343 




Blood - 

pressui't 



,,Mm^^^^''^^^^^ 



Till 




Fio. 4. — Cat ; ACE mixture : wrethanc. Shows the oomparative effect of pyridine, 
collidine, and nicotine on blood -pressure. 

A, the effect of injecting into a vein 6 c.c. of 1 per cent, pyridine; B (in whicli tlie intestinal volume is also 
recorded), the effect of injecting 6 c.c. of 1 jier cent, cullldine ; C shows the effect of 1 c.c. of 1 per cent, 
nicotine. These tracings were obtained from different animals, Itnt may be taken for comparison. Time, 
30 seconds. 



344 



Lee 



after injection, from complete paralysis of the central nervous system. 
After systemic death it was found that the motor nerve endings were 
almost paralysed, but that the heart was still beating. 

After a series of similar experiments upon frogs whose weights 
approximated closely to those given in the previous experiment, it was 
decided that the following were the minimal lethal doses of the drugs 
for a 20-gram frog : — 

Pyridine . . -5 grain =039 gram. 

Collidine . . I grain =016 gram. 

Nicotine . ^^^ grain = 006 gram. 

Therefore, if the toxicity of pyridine be represented by TO, that of 
collidine will be 2"4, and that of nicotine 60. From this it will be seen 
that the toxicity of these three drugs varies in the same way as their effect 



Fig. 5, a. 



n*« 


r^ 


N 


(M- 


(N 


(N 


c4 


ci 


.N 


(Ti 


ci 


(N 


ci 


d 


CO 


CO 


eo 


CO 


130 
125 
120 
115 

110 


N 




































\ 




































k 






































^*" 




--S 




























~' 








"V- 




"^ 


^ 


=— 


— 






— 


— 



Time. 



70 70 70 68 66 



68 70 68 68 68 66 68 



68 Pulse. 



on isolated tissues. That is to say, their effects on the heart, plain muscle, 
and central nervous sy.stem all run a parallel course, nicotine in each case 
being much the most active and pyridine mucli the least. 



V. The Effect of Smoking on Man. 

A series of experiments were conducted upon men whose habit varied 
from that of the novice to that of the seasoned smoker. I have described 
below a series of experiments on the blood -pressure of men. 

The blood-pressures were taken with Martin's modification of the 
Riva Rocci instrument, the pressure band being fixed in each case to the 
bared left arm. The normal blood -pressure was always taken a number of 
times before a final figure was fixed upon. I invariably found that merely 
fixing the instrument upon the arm was sufficient to raise blood pressure 
several mm. Hg, and the final figure was never determined until the man 
was accustomed to all his surroundings and his blood-pressure was quite 
constant. The habitual smokers were required to observe abstinence from 
tobacco smoke for some six hours before being subjected to an experiment. 

Experiment 1 shows the necessity of obtaining a correct normal reading 
of the blood-pressure before determining the effect of any drug. 



Action of Tobacco Smoke 



345 



Exp, 1. A yo;ith aged 18 was placed in a chair in as comfortable a position 
as possible, and a normal chart prepared for comparison with those in 
which smoke was inhaled (see fig. 5, a). 



Time. 


B.P. 


Pulse. 


1.55 P.M. 


134 


78 


2 


128 


76 


2.5 


120 


70 


2.10 


1-20 


70 


2.15 


118 


70 


2.20 


118 


68 


2.25 


116 


66 


2.30 


114 


68 


2.35 


114 


68 


2.40 


113 


70 


2.45 


112 


68 


2.50 


110 


68 


2.55 


110 


68 


3 


110 


68 


3.5 


110 


68 


3.10 


110 


66 


3.15 

1 


100 


68 



The following experiments show blood-pressure changes and pulse-rates 
during smoking. 



Exp. 2. Youth aged 17^, an occasional smoker of cigarettes; normal blood- 
pressure, from a series of observations, found to equal 117 mm. Hg 
(systolic reading) ; pulse-rate 72. (See fig. 5, b.) 



Time. 


Blood-pressure. 


Pulse-rate. 


Observations. 


6 p.m. 


114-116 


72 


Started to smoke and inhale the stand- 
ard Manilla cigar. 


5.15 


128 




A distinct pallor of face, and sensation 
of weakness in legs. 


6.20 


128 




Eyes " sleepy " ; appears shaky. 
Feeling faint, cigar half finished. 


6.25 


128 




6.30 


78 




Intense pallor of the skin, <old sweat 
on forehead. 


5.35 




60 


Feeling very faint and weak, colicky 
painsinabdomen, cigar three-quarters 
finished. Stopped smoking. 








5.40 


114 


66 


Feels less faint. 


5.44 


108 


72 


Lip.s regaining theii colour. 


5.5() 


95 


64 


Feels better, muscles stronger but is 
still incapable of physical exertion. 


6.10 


104 


54 


lietter, but feels weak all over. 


6.25 


110 


54 


Slight indisposition. 


Next morning 


121 


72 


Normal state. 



846 



Lee 






Fig. 5, b. 



u-5 O O O 



l^Time. 



— 




^? 








^ 






























125 


/ 




\\ 




il 


/ 






p 


































/ 








1 


























y 


^ 


115 
110 
105 
100 
95 
90 
85 
80 
















1 






















^ 


y 




























































r- 


-\ 






^ 
































/ 


s 


K- 


































/ 








































/ 






































\i 








































/ 






1 




1 



















72 



76 



54 



72 Pulse. 



3 . 



Fk;. 



m o o i^ o o 



Time. 



o^iMC^ieoeoeoeoMe-s 



130 
125 
120 
115 
110 
105 



1 ^ ! 


1 M^l' 


1 

1 1 


/^A 


^^+1 


-^^ ^ ^-^ ^ 




i N / 


1 


\ / 




1 \ f 

v 

^ 


1 





84 88 80 84 90 88 88 88 92 106 104 102 



80 



Pulse. 



Fig. 



O U-, O o 



145 
140 
135 
130 



«o oo c^ 





' 1 


"l ^ 




y 


/N 


k 


















^<^ 


r ^ 


k y 


Y 




N, 


\ 
















1 . 


r 


' ! 












1 


V 










*v 


^^ 






1 












1 




L. 




- 






1 




! 






\ 




i i 









60 68 68 70 74 68 72 76 82 70 72 



66 62 62 64 Pulse. 



Action of Tobacco Srnokt 



84'i 



Exp. 3. Boy aged 1 5, uiodeiate smoker of cigarettes. Shows effect of smoking 
one and a half Manilla cigars ; occasionally inhaled. Normal blood- 
pressure and pulse-rate were found to be 120 and 84 respectively. 
(See fig. 5, c.) 



'J'iine. 


Blo()cl-i>re.s.sine. 


Pulse-rate. 


Observations. 


12 P.M. 


120 


84 


Normal .sensations. 


12.30 


122 


88 




2.50 


120 


80 


Started .smoking. 


2.55 


120 


84 




3.4 


120 


84 




316 


124 


90 




: 3.20 


130 


88 


Heavy feeling in head. 


1 3.30 


128 


88 


Feels shaky. 


3.35 


128 


88 


Slight dizziness. 


3.40 


130 


92 


Feels weak ; forehead shows beads of per- 
spiration. 


3.50 


120 


lOG 


Felt suddenly faint — nausea and misty vision. 
Smoking ctopped. 


4 


120 


104 


No change from 3.50. 


4.5 


120 


102 


Nausea— stiff feeling about the back ot neck. 
Great lassitude. 


4.30 


104 


88 


Feels slightly better. 


5 


120 


80 


Feels widl. 



Exi'. 4 Boy aged 17, smoker of cigarettes only. Smoking two Manilla 
cigars, with occasional inhalation. (See fig. 5, d.) 



Time. 


lilood-preasure. 


Pulse-rate. 


Observations. 


2.50r.M. 


134 


54 




2.55 


132 


60 




3 


132 


60 




Normal 


blood-pressure thei 


efore 132-134 


Pulse 60. 


3.5 


135 


68 


Started smoking. 


3.10 


138 


68 




3.15 


140 


70 


Pulse irregular. 


3.20 


142 


74 




3.25 


145 


68 


Pulse intormitLeiit. 


3.40 


142 


76 




3.45 


146 


82 


Pulse irregular and intermittent. 


3.55 


150 


70 


Sensations of vertigo; hands show distinct 
tremors. 


4 


145 


72 


Smoking stopped. 


4.5 


142 


78 




4.10 


136 


78 


Considerable salivation- a feeling of nausea. 


4.20 


134 


(!6 


Feels better. 


4.30 


132 


62 




4.40 


182 


62 




4.50 


133 


64 





348 



Lee 



Exp. 5. Man aged 30, moderate smoker, but not an inhaler. Smoked one 
Manilla cigar. 



Time. 


Blood-pressure. 


Pulse-rate. 


Observations. 


11.30 A.M. 


122 


52 






11.40 


122 


56 


Started smoking. 




11.45 


120 


56^ 








11.50 


123 


60 








11.55 


120 


58 




No change in sensations. No alteration 


in 


12.0 


120 


58 




colour. 




12.10 


128 


60 








12.15 


130 


58j 








12.20 


130 


60 






12.25 


130 


60 


Stopped smoking. 




12.30 


128 


58 






12.35 


126 


56 






12.40 


125 


66 






12.45 


124 


56 






12.50 


123 


66 







Exp. 6. Man aged 31J, habitual smoker, but non-inhaler. Smoked one 
Manilla cigar, inhaling all the time. 



Time. 


Blood -pressure. 


Pulse-rate. 


Observations. 


11.20 a.m. 


124 


60 




11.25 


120 


60 




11.30 


120 


60 




11.40 


120 


62 


Started smoking. 


11.45 


122 


64 




11.50 


124 


66 




11.55 


128 


66 




12 


128 


66 




12.5 


129 


69 




12.10 


128 


66 




12.15 


128 


64 




12.20 


126 


64 




12.25 


126 


66 




12.30 


128 


63 




12.35 

12.40 


126 
122 


68 
64 


Slopped smoking. 


12.45 


122 


60 




12.50 


122 


64 




12.55 
1 o'clock 


122 

122 


64 
66 


No subjective or objective changes noted 
throughout the experiment. 


1.5 


120 


64 




1.10 


120 


64 





Action of Tobacco Smoke 



349 



Exp. 7. Man aged 24, a smoker of seven years' standing. Smoked one 
cigar, not inhaling. 



Time. 


Blood-pressure. 


Pulse-rate. 


Observations. 


10.20 a.m. 


120 


72 


Average reading for 9.55-10.20. 
Started to smoke. 




10.30 


120 


72 




10.35 


120 


56 






10.40 


121 


60 






10.45 


124 


56 






10.50 


125 


56 






10.55 


127 


60 






11 


128 


64 






11.10 


130 


60 






11.15 


130 


62 






11.20 


131 


60 






11.25 


130 


62 


Stopped smoking. 




11.3^ 


128 


60 






11.35 


126 


60 


There were no subjective or 
changes noticed. 


objective 


11.40 


126 


60 






11.45 


126 


60 






11.50 


124 


60 






11.65 


122 


64 






1 o'clock 


120 


46 







Exp. 8. Man aged 29, habitual smoker, frequent inhaler of cigarettes. 
Smoked one Manilla cigar, inhaling all the time. (See fig. 5, e.) 



Time. 


Blood-preasnre. 


Pulse-rate. 


Observations. 


11.5 a.m. 


108 


68 




11.10 


108 


68 




11.15 


108 


68 


Started smokir.g. 


11.20 


110 


681 




11.25 
11.30 


108 
110 


68 1 

68 r 


No alteration in sensation. 


11.35 


108 


68) 




11.40 


108 


68 


Stopped smoking. 


11.50 


108 


68 




12 


108 


68 





One last experiment in the series was made by smoking leaves which 
are known not to contain nicotine. Dried lavender leaves were used for 
this purpose, these being sold in certain parts of this country as " boys' 
tobacco." The result of smoking such leaves is to cause a sensation of 
stinging or scalding in the throat and mouth and some slight rise in blood- 
pressure, the latter, however, not being comparable to that af tobacco. 
Fig. 5, f, shows the effect of smoking such leaves. This chart was obtained 
from the same subject as fig. 5, a and b. 



350 



Lee 



Tliese protocols are typical of the effect of tobacco smoke on man 
They may be divided into three groups : the first including those in which 
the smoker was a novice ; the second, the group of moderate smokers ; and 
the third, the group containing the " excessive smokers." 

In the case of the novice there is always an initial rise of blood-pressure 
very shortly after the inhalation has well started^ and lasting half an hour 
or perhaps even a shorter time. The height to which the blood-pressure 



Fig. 5, 






120 
115 
110 





Z^ 


d 


;;; 


^ 


^ 


^ 


d 


^ 


-■ 


a 








1 












I 










1 
§ 






















1 








,^— 




I 






»■ 


-*- 


-*" 






^ 


■^ 




""*" 



Time. 



68 68 68 



68 68 68 68 Pulse. 



rises above the normal varies, but is usually from 10 to 20 mm. Hg. This 
effect is associated with some quickening of the pulse ; for example, in 
protocol 2 the increase is from 72 to 80, and in 3 from 84 to 106. 

At first the smoker has no unpleasant symptoms, but rather a feeling 
of well-being and exhilaration. As the smoking continues, however, a 
sudden change occurs in the blood-pressure, which begins to fall rapidly, 



120 
115 

no 

105 
100 



Time. 







n 
























i 










L 






"^ 


\ 




? 




















— 








A 




i 




















V 




y^ 



























67 64 64 64 64 60 59 59 60 60 Pulse. 



SO that, as in the case shown in protocol 2, there may be a fall of 50 mm. 
Hg within five minutes. 

When the smoker, though a novice, is less affected by the inhalation, as 
in protocol 3, the fall, though vstill rapid in onset, does not so closely 
resemble a crisis as in experiment 2. This fall in blood - pressure is 
associated with all the symptoms characteristic of shock or collapse. 
The face becomes pale, the skin is covered with a clammy sweat, there is 
general weakness of all the muscles, faintness, shallow respirations, and a 
slow and feeble pulse ; sometimes nausea or vomiting may be present, and 



Action of Tobacco Smoke 351 

sometimes colicky pains are felt in the abdomen, suggesting increased 
peristalsis. 

These experiments strongly suggest that fall in blood-pressure is the 
essential factor in the production of collapse, for all the symptoms of 
collapse are such as are obtained from a sudden fall in blood-preasure. I 
believe that these simple experiments amply confirm the hypothesis of 
Crile (4), that shock and collapse are conditions resulting from a severe 
fall in blood-pressure. 

In tobacco smoke there is only one constituent (nicotine) which has the 
power of increasing blood-pressure appreciably, but there are many sub- 
stances, such as the pyridine bases, which lower the pressure. During the 
inhalation of tobacco smoke, the action of nicotine overshadows that of the 
other constituents; the nicotine stimulates nerve cells, and for a time 
exercises unchallenged its vaso-constrictor influence, with the accompanying 
rise of blood- pressure. But a stage in smoking is reached when the 
stimulation of nerve cells by the nicotine gives place under the same 
influence to their depression, with resulting vaso-dilatation and a fall in 
blood-pressure. This condition will be exaggerated by the other con- 
stituents of the smoke, such as pyridine and collidine, whicii throughout 
have been tending to lower blood-pressure. 

Tbe action, then, of tobacco smoke on man is exactly what might 
be anticipated from a knowledge of the action of nicotine : while the 
stimulation stage lasts the pressure is raised ; then, as the nerve cells 
are depressed, the blood - pressure falls. But I have obtained plenty 
of evidence of variations in the degree of idiosyncrasy for nicotine. 
Thus, in comparing experiments 2 and 3, although the subjects were 
very similar as regards their smoking habits — in fact, they were both 
almost complete novices — yet the elder was much more affectod than the 
younger, twice the quantity of smoke being required to produce a similar 
effect in 3 as in 2. 

The second group is composed of moderate smokers, and tliis gj-oup 
probably includes the majority of those who smoke regularly. Experi- 
ments 5, 6, and 7 are typical examples. In these the blood-pressure rises 
slowly, unlike those in the tirst group, where the rise is rapid ; the heifjht 
to which the pressure rises, however, is about 10 nmi. Hg, and the 
tendency is for the blood-pressure to continue rising slightl3% or at least to 
maintain the higher level, whilst the smoking lasts. With the novice, on 
the other hand, blood-pressure falls, and collapse ensues w;hilst smoking 
is actually in progress. In these temperate smokers, after smoking has 
ceased, the blood-pressure falls gradually to the normal, but shows no 
tendency to move below that level. The whole effect, then, consists of 
a small and gradual rise of blood-pressure lasting until the smoking 
ceases. The rise in pressure is usually associated with some acceleration 
of the pulse. 

Of course these experiments only show the effects of moderate smoking 



I 



352 Lee 

on the moderate smoker. If the moderate smoker smokes to excess, he 
assumes the position of a novice, and a climax would be reached in which 
the nicotine and other constituents of -tobacco smoke would accumulate 
in the blood to an extent which would paralyse the nerve cells, and 
produce the sudden fall of blood - pressure characteristic of collapse. 
Why the moderate smoker is able to withstand the action of smoking 
so much better than the novice, is a question with which I propose to 
deal elsewhere. 

The third group — the excessive smokers — an example of which is shown 
in protocol 8, is merely an exaggeration of group 2. The pressure rises 
only 2 mm. or 4 mm. Hg, and such as it is, is maintained in the same 
way as in group 2 until smoking ceases, when it returns again to the 
normal. The pulse in these cases is not affected. 



VI. The Effect of Smoking on Animals. 

(a) Immediate Effects. — The effect of inhalation of tobacco smoke on 
ansesthetised animals was determined by connecting a tracheal tube, which 
has in it a suitable lateral opening, capable of being regulated in size, with 
a lighted cigarette. The animal will then inhale the smoke with a variable 
admixture of air according to the size of the aperture in the lateral tube : 
the inhalation will be on much the same principle as when man inhales. 
The objection to this mode of procedure is the effect of the smoke on 
respiration : the animal ceases to breathe for a time, or the respiration 
becomes feeble and irregular, and this results in a variety of secondary 
circulatx)ry disturbances from the spasmodic breathing, or from partial 
asphyxia, so that it is almost impossible to say which circulatory effects are 
due directly to smoking and which are due to irritation of the respiratory 
tracts. 

To obviate this difficulty an alternative mode of administering the 
smoke was adopted. A normal animal was first killed by the destruction 
of the brain and medulla oblongata by pithing, without the ase of any 
anaesthetic, artificial respiration being started. In the course of the air 
tube was inserted a special apparatus, devised for these experiments, which 
is shown in fig. 6. The current of air was divided by a Y-tube, the one 
limb carrying the air directly to the trachea tube, the other through a 
chamber in which, by allowing a regulated quantity of air to pass through 
it, tobacco in the form of a cigar or cigarette could be burnt at any 
required rate. The two streams of air were brought together again 
by a second Y-tube, the single limb of which was connected to the 
trachea tube. 

A glance at tig. 6 shows that when a lighted cigarette or cigar is placed 
in the tube A, if the tube B is closed during the down-thrust of the pump 
all the air will pass through the tube A and through the cigarette or 
cigar, and thus the animal will receive the smoke. By regulating the 



Action of Tobacco 8inok» 



353 



354 



Lee 



amount of air passing through B, it is possible to obtain a condition in 
which the smoke is comparable in amount to that which is inhaled by 
man : this condition is gauged by the rate at which the cigarette or 
cigar burns. 

The effect of such smoking on a cat is shown in fig. 7. 

In this cat the blood-pressure rose 20 or 30 mm. Hg during the 
first five minutes of smoking, and then began to fall, in spite of the fact 
that the smoking was still going on. This condition may be regarded as 
typical, and bears a close analogy to that which occurs in man. Occasion- 
ally, when the blood-pressure began to fall, the animal showed convulsive 
movements. The ultimate effect, after smoking two or three cigarettes, was 
a considerable fall in blood-pressure. 

In one experiment, during the inhalation, a quantity of moisture from 



Inteatine. 




Fig. 8. — Cat Brain destroyed by pithing. Artificial respiration. Intestinal volume. 

Blood-pressure. 

Shows the effect of accidental inhalation of tobacco ]aic«. Time, 30 seconds. 



the cigar condensed on the glass tube, just in front of the tracheal tube. 
This during one inspiration was blown into the trachea, and produced an 
immediate rise of blood-pressure with marked constriction of the blood- 
vessels. This is shown in fig. 8, in which the upper tracing represents the 
intestinal volume, the lower the blood-pressure. It is just such sudden 
increases of blood-pressure as these which stretch and rupture the elastic 
fibres in the vessels, as described later. The experiment illustrates the 
possible dangers attendant on the u.se of foul pipes and the latter end of a 
cigar, although they may be exaggerated in this instance, since the fluid 
passed directly into the trachea and not into the mouth, as would be the 
case in man. 

The ultimate effect on blood-pressure is shown in fig. 9. It this case a 
rabbit was used, anaesthetised with urethane. The tracing A shows the 
height of the blood-pressure before the smoking commenced, while B 



Action of Tobacco Smoke 



355 



represents the height after the inhalation of three ^i^arettes, smoked during 
thirty minutes. 

The first effect of the inhalation- of tobacco smoke Ox ^. heart is shown 
in fig. 10. In this experiment the outflow of blood from, the heart was 





Fio. 9.— Rabbit. Uiethane. 

A shows normal blood-preBsiire, and B shows the fall after smoking three standard cigarettes. 
Time, 30 seconds. 

measured by the cardiometer. It will be noticed that the heart filled with 
blood, but that its systole was a little incomplete ; that is to say, it did 
.lot empty itself quite as completely as it normally does : nevertheless, the 
total output from the heart was increased. 




Cardiometer. 



P'lG. 10.— Cat, A-C-E : urethane. Artificial respiration. Cardiometer and blood-pressure. 
Shows effect of inhaling smoke from a Manilla cigar with a plentiful supjily of air. Time, 30 seconds. 



From lialf to three-quarters of an hour after smoking, the blood-pressure 
docs not rise on the administration of large doses of nicotine, and the vagi 
arc found paralysed, although adrenaline causes its normal eff'ect. This 
must mean that an excess of smoking produces paralysis of the autonomic 
nerve cells in exactly the vsame way that nicotine does. 



VOL. 1., NO. 4, — 19( 



24 



366 Lee 

(b) -Remote Effects. — Rabbits were used for this experiment, because 
they are known to be very susceptible to changes in the arterial system. 
Two were made to inhale tobacco smoke regularly for about fifteen to twenty 
minutes on alternate days throughout a period of five months. Two 
methods were used. In the one a mask was fitted over the mouth and nose 
of the animal and, with the apparatus already described, through which the 
supply of both air and smoke could be regulated, the smoke from mild 
cigarette tobacco was directed into the mask. A very dilute mixture was 
at first given, but as the animal became accustomed to the treatment the 
percentage of smoke was increased. The second method consisted in 
placing the animals in a small chamber with inlet and outlet valves ; one 
side of the chamber was of glass, to enable observations to be made, and 
through this chamber smoke in varying quantities was passed continuously. 
The latter method was chiefly employed. Whilst the animals were inhaling 
the smoke, it was noticed that salivation was present, but no other changes 
from their normal condition appeared. On being removed from the chamber, 
for the ensuing three or four minutes they appeared lethargic, although 
their reflex sensibility was somewhat heightened. At the end of five 
months each animal had received about seventy inhalations of tobacco 
smoke ; during this period they had increased in weight, as did the control 
animal, and, as far as one could see, had suffered in no way, for they took 
their food and behaved as normal rabbits, and during the later inhalations 
showed no objection to the process. 

We now know that any substance which has the power of raising blood - 
pressure suddenly and to any considerable extent, say 30 or 40 mm. Hg, 
tends to injure the aorta. This has been shown by Harvey (5) in this 
laboratory. Harvey raised the blood -pressure by compressing the ab- 
dominal aorta, thus excluding all the hypotheses and speculations which 
have been made suggesting that the " irritation " of the drug causes the 
arterio-sclerosis observed. Moreover, irritant drugs which do not raise 
blood -pressure do not cause arterial disease. 

We may, therefore, assume that if the smoking to which these animals 
were subjected was capable of raising the blood-pressure to any considerable 
degree, signs of arterial degeneration may be expected. 

Rabbit B was killed A few miliary patches were visible in the ascend- 
ing aorta, and also a small plaque in that sinus of Valsalva which does 
not give ofl" a coronary artery. The lungs showed some congestion, but 
no consolidation, and the bronchial glands were normal. In the spleen 
was an old infarct. The kidneys were normal. 

The aorta was prepared for microscopical examination ; portions of it 
were hardened in formalin and dehydrated in a series of alcohols of in- 
creasing strength up to 100 per cent.; they were cleared with chloroform, 
and embedded in paraffin. Sections were cut and stained with : — (a) 
acid hsematoxylin (Ehrlich) and aqueous eosin, or with picric acid fuchsin 
(Weigert's modification of Van Giesen); (b) acid orcein (Unna-Taenz^r 



I 



Action of Tobacco Smoke 357 

elastic tissue stain) ; and (c) silver nitrate, 5 per cent, aqueous solution 
(v. Kossa's calcium test). The sections were mounted in Canada balsam. 
Photomicrographs of the three specimens are shown in the accompanying 
Plate 1 shows extensive fibrosis of the tunica media, invading to a slight 
extent the tunica intima. It also .shows, in parts, the antecedent stage of 
■fibrasis as an inflammation, with plentiful cell-proliferation. 2 shows 
marked erosion and rupture of the elastic fibres ; some of these fibres are 
encased in calcium .salts. 3 shows considerable deposits of calcium .salts 
— which are stained black in this specimen. 

Rabbit A was killed. The animal was found to have tuberculosis of 
the lungs. The heart and aorta of this animal showed hardly any naked- 
eye change. The trachea was acutely inflamed, and with the exception of 
the tuberculous lesions, there were no other abnormalities. 

Sections from the aorta of this rabbit (A) show changes similar to 
those exhibited by section 1 of rabbit B. 

From these experiments I conclude that it is po-ssible to obtain arterio- 
.sclerotic changes by the inhalation of tobacco smoke. 

It gives me much pleasure to record my thanks to Dr W. E. Dixon for 
his advice, and to Dr W. H. Harvey for preparing the photomicrographs. 



VII. General Conclusions. 

1. Nicotine is the most important poison in tobacco smoke. 

2. Pyridine bases, in the quantities in which they are present in tobacco 
smoke, are not injurious to the .smoker. 

3. Smoking raises the blood-pressure by vaso-constriction, accelerates 
the heart and respiration, and increases intestinal movements. In excess, 
cerebral depression may occur, and, with the coexisting depression of the 
vaso-motor centre, may lower the blood-pressUre to such an extent that 
collapse may be induced. 

4. The amount of nicotine inhaled during smoking depends not so much 
on the tobacco smoked, as upon the form in which it is smoked. The 
greater the condensation area between the point of combustion and the 
entrance into the mouth, the more nicotine will be inhaled. 

5. Arterial disease may result from prolonged tobacco smoking. 



REFERENCES. 



I 



(1) Langley, Journ. Physiol, xl, p. 262. 

(2) Dixon, Journ. Physiol, vol. xxx. p. 115, 1903. 

(3) Crile, Experimental Research into Surgical Shock (1897). 



358 Action of Tobacco Smoke 

(4) Zebrowski, Centralbl. t. allgeni. Path. u. path. Anat., 1907, vol. iviiu 
p. 337. 

(5) Harvey, Recent work (not yet published) from tlie Pharmacological 
Laboratory, Cambridge. 



DESCRIPTION OF PLATE. 



The figures represent photographs taken under a moderately high power of the 
microscope of part of the wall of the aorta of a rabbit which had been caused to 
inhale tobacco smoke at intervals during five months before being killed. For further 
description see text, pp. 356, 357. 



bS8' 



Quarterly Journal of Experimental Physiology, Vol. 



A, 1908. 






W. Emerson Lee, " Action of Tobacco Smoke. " 



ir/ 



CONTRIBUTIONS TO PHYSIOLOGICAL TECHNIQUE. 

By F. S. Locke. 

(Received for publication 28th August 1908.) 

I. Signalling more than one Kind of Event with only one 
Writing-Lever. 

When events of more than one kind have to be graphically recorded 
electrically, it is usual to employ a separate electromagnetic signal for each 
kind of event. The proper arrangement of more than one electric signal in 
relation to the other recording apparatus is often a matter of irritating 
difficulty, when they do not form a permanent fixture on the kymograph 
employed (Hering-Rothe, Brodie-Palmer), and even here the limited 
adjustability of the signals in height renders the application of another 
system of signals frequently desirable. Combinations of two or more 
signals on one supporting rod suitable for use on any recording surface have 
been designed (Carl Ludwig, Langendorff, and others), but at best the 
multiplication of writing-points is not a thing to be wished for, and in many 
cases it is of distinct advantage to be able to make one writing-point 
record distinguishably in addition to time-intervals at least one other kind 
of event. I have devised two ways whereby this end is attained, one of 
which demands an electromagnetic signal of special construction, while the 
other, although necessitating more complicated electrical connections, can be 
used with any signal possessing an adjustable armature-spring and iron 
cores not exhibiting too much hysteresis. 

A Double Electromagnetic Signal with a Single Writing-Lever. 

The signal in question consists of two electrically separate electro- 
magnets, the armature of each of which, instead of being connected with a 
separate recording lever, is connected with one common recording lever by 
means of a flexure joint which serves either as point of application to the 
lever of the force of attraction of its own electromagnet, or as relatively 
fixed axis of rotation of the lever when acted on by the force of attraction 
of the other electromagnet, the movement of the writing-point attached to 
one end of the lever being in opposite directions in the two cases. The actual 
mode of construction is shown in fig. 1, and is so obvious that a detailed 
description is unnecessary. The special points worthy of mention are : — 

(]) The armature-springs, by which the armatures are connected with 
the solid brass frame of the apparatus, are of sheet phosphor-bronze, of 



360 Locke 

triangulat outline, with the armatures fastened across their apices. The 
triangular shape not only produces on flexion a rapidly increasing resistance 
to further flexion, which prevents an armature being brought in contact 
with a pole, but also makes its movement very much more dead-beat, 
owing to interference in their tendency to vibration of the longitudinal 
elements of different lengths. 

(2) The screw adjustment of the strength of the armature-springs and 
the distances of the armatures from the poles. The screws act indirectly 
on the armature-springs through two accessory springs of sheet phosphor- 
bronze, the mode of action of which is obvious in the figure. These, besides 
producing a very smooth and gradual adjustment, increase further the 
dead-beatness of the armatures in the case of downward movements. 




Fig. 1. — Double signal. | real size. The "legs" of the suspension of the writing- 
lever are a little longer than usual for the sake of greater clearness. 

(3) The armatures and pole-pieces are of special construction, the latter 
being conical in shape, and the armatures stamped out conically to fit over 
them. Magnetic attraction is thus made to take place through a longer 
range, with a strength less rapidly increasing with approximation of the 
armature than with the usual arrangement. 

(4) The suspension of the writing-lever between the two armatures 
consists of two spirals, one right- and the other left-handed, of thin German- 
silver, brass, or platinoid wire, which fit into one another with their coils 
alternating to form a tubular socket 1'2-1'4 mm. in diameter and 4 min. 
long, for the writing-lever. The ends of the two spirals make four legs 
and feet, so to speak, a pair of feet being attached to each armature by 
the same screws which attach it to its spring, and form a flexure joint 
permitting sufficient freedom of movement without the complicated con- 
struction and other disadvantages of friction-joints. 

(5) In the double-spiral socket a writing-lever about 5 cm. long of 



Contributions to Physiologic<al Technique 361 

any suitable material can be fixed. Fine straws serve admirably, but I 
prefer myself to use a strip of thin vshcet-magnalium/ 15-2 mm. wide 
sharpened and bent to form a writing-point, and stiffened by having a line 
drawn down its middle by means of a blunt metal point and straight-edge. 
Such a magnalium lever can also be stiff'ened by making its transverse 
section curved in outline by drawing it through a short piece of glass 
tubing constricted at one end in the blowpipe-flame. If the fixed end be 
suitably tapered no special aid to fixation is necessary for ordinary work, 
but if desired a tiny wood or magnalium wedge can be employed. 

The bilateral symmetry of the apparatus both in tiie transverse and 
longitudinal directions permits the writing-lever being fixed so as to point 
backwards if required on either side. This conveniently avoids the 
necessity of ever having the direction of the writing-lever against that of 
the motion of the recording surface. The necessity of the writing-lever 
projecting somewhat towards one side of the apparatus for application 
to the recording surface, is better met by a torsion to the required 
degree of the by no means fragile suspension by means of the fingers, than 
by too much flexion of the writing-lever. Any flexion it may be desired 
to give to it is better done by a sharp angle than a gradual curve. 

(6) Adjustability of the degree of pressure of the writing-point on 
the recording surface is given by means of a strong flat spring suspension, 
the tension of which is regulated by a screw. The latter is, however, 
placed longitudinally instead of, as usual, transversely. Its ht^ad remains 
therefore always conveniently accessible whatever 'the position of the 
apparatus. 

As would be anticipated from its mode of cf)nstruction, iIh; iiiol)ility of 
the signal is very considerable. It will respond easily to 100 interruptions 
per second. Beyond this I have not yet tested it. The latei>cy of the 
signal to the rupture of an electric circuit is of the same order as that of 
the Pfeil and Deprez signals (0'5-la-), and is very constant under identical 
conditions. 

Specimens of records by the double-signal are given in fig. 2. One 
electromagnet was connected up with the Brodie-Palmer clock writing 
seconds, and the other with the primary circuit of a Du Bois coil. The 
speed of the recording surface was different in each tracing. Records of 
make and break and of " tetanisation " are shown. The rate of interruption 
during the latter is well shown in the lowest tracing (52 per second). 

The double signal was shown, in a form differing only in its 
details, at the meeting of the Physiological Society on 1st March 
1902. It is made by Mr John Sinclair, Physiological Labo- 
ratory, University of London. I may add here that I have 
applied the same principle of the electromagnetic shifting of the 

' Recording levers of magnalium were employed by Dr O. Rosenheim and myself for 
heart-work in 1903. The much greater rigidity of this alloy of magnesium and aluminium, 
with its actually less den.sity, makes it much more suitable for the purpose than aluminium. 



362 Locke 

' axis of rotation of the lever recording intervals of time to the 
Jaquet Time Marker, which thus becomes an electric signal as 
well. A description of this will be published later. 

The Polygraphic Use of a Single Simple Electromagnetic Signal. 

In order to get the most polygraphically out of an ordinary electro- 
magnetic signal, it is essential that it be used to record time-intervals of 
not much less than one second, by closures of the chronograph-circuit of 
markedly unequal length to that of the openings alternating with them. 
Absolute continuity of the time-record must also be not all important, a 
condition very rarely unfulfilled in the present perfection of recording 
apparatus. Given a signal writing seconds by short makes of the battery - 



Fig. 2. — Records of time in seconds (every tenth second missed) ; of makes and breaks of a constant 
current, and of " tetanisation " by means of the interrupter of an ordinary Du Bois coil. 
The speed of the drum was increased between each tracing from above downwards. In the 
lowest tracing the interruptor was continuously in action, and two clock-contacts a second 
apart are shown. 

current through its electromagnet (e.g. by means of the Brodie-Palmer 
clock), there are three distinct ways (cf. fig. 3) in which it can be made to 
notify the occurrence of something other than the closures of the seconds- 
clock : (1) The rendering of these closures ineffective by opening the circuit 
at some other point, an Unbroken abscissa being thus produced. This will 
be called Signal 1. (2) The continuous closure of the chronograph-circuit 
by means of a key " in parallel " with the clock-contact, the abscissa being 
thus shifted in height, while seconds cease to be recorded This will 
be called Signal 2. (3) Arranging that the clock-closures, instead of 
magnetising the signal, demagnetise it, this being attained by continuous 
closure of the chronograph-circuit as in (2), while at the same time the 
clock-contact is made to act as short-circuit to the signal, the seconds-record 
being thus inverted. This will be called Signal 3. 

The sharpness of Signal 1 is obviously less than that of Signal 2 in the 



Contributions to Physiol o<^cal Technique 



363 



case of fig. 3, the possible •uncertainty of the beginning and end of the 
former approximating more or less to the time-interval of the clock.' 
Signal 1 is therefore best used for such an event as an injection. If, how- 
ever, the closure of the clock-contact be longer in duration than its opening, 
as is the case vi^ith the Leipzig clock, the case is reversed, and Signal 1 be- 
comes sharper than Signal 2. 

All three signals can be made by a single movement of the hand, by 
means of a very simple switchboard consisting essentially of two spring 
two-contact keys fixed parallel and close to one another on the same 
board/so that one or the other or both together can be depressed at will. 




'Signal 1.' 



^'Signal 2. 



' ' Signal 



Fig. 3. —For description see text. 

One of them acts as a two-way switch, the other as a make and break 
key. A complete scheme of connections is shown in fig. 4, A, and the 
connections actually functioning in each case in fig. 4, B, C, D, and E. 
Fig 4, B, shows the actual connections when time alone is being recorded. 
Depressing the left-hand key (switch) produces Signal 1 by creating 
the connections shown in fig. 4, C. Depressing the right-liand key 
(key) produces Signal 2 by the connection shown in fig. 4, D^ De- 

1 The Brodie-Paliuer clock is usually made to drop eyery tenth second, as in the 
tracings accompanying this paper. The lack of sharpness of Signal 1 may therefore rise with 
it to two seconds, but this is of course not nece.s<ary when the time of occurrence of the 
event to be signalled can be deliberately chosen. 

-' If Signal 2 be rapidly repeated by a succession of short taps on the right-hand key, 
an additional easily di-stinguinlialile signal is produced more or less resembling on a 
slowly moving drum the upper records of " teUmisations" in fig. 2, p. 362. This variation 
can, of course, be also made use of with any ordinary signal not recording time with simply 
a make and break key in its circuit. 



364 



Locke 



pressing both keys togetlier produces Signal 3 by the connections shown 
in fig. 4, E. 

As is indicated in fig. 4, A, by dotted lines, the right-hand key 



BATTERY. 



CLOCK- 
CONTACT 

rH-i 




rrn~o 



<) (^ 



d) 6 



«).- 



c) (^ 



) '^ 



S 



I 



•h 



(J) (> 



Ji? 



Fio. 4. — For description see text. 



thanks to its spare-contact, can also if required serve as short-circuit- 
ing key to the stimulating electrodes in the secondary circuit of an 
inductoriuni, in order to signal automatically the duration of a tetani- 



Contributions to Physiological Technique 365 

sation.i Of course it is impossible to give Signals 1, 2, and 3 simultane- 
ously, but if it be necessary to signal the simultaneous occurrence of two 
events, this can be indicated by Signal 3, the separate events being 
indicated by Signals 1 and 2. 

The connections of fig. 4, A, are not the only ones possible to attain 
the end in view, and in fact are not applicable to the Brodie- 
Palmer clock on account of the permanent connection in that 
apparatus of one battery-pole with one side of the clock-contact. 
Altogether there are four modes of arranging the connections 
possible according to whether the continuous closure of the circuit 
independently of the clock in Signal 2 is effected (a) by the 
" switch " or (6) by the " key " ; and in each of these cases 
whether the clock-contact is put in permanent connection (ac, he) 
with the battery or {ad, hd) with the signal. These four different 
possible arrangements will now be referred to as ac, he, ad, and 
hd respectively. It is unnecessary to describe them in detail 
systematically ; they may be easily worked out by those inter- 
ested. 

The arrangement of connections in fig. 4, A, corresponds with ad. 
This is theoretically the most preferable of the four, because short- 
circuiting of the battery is only made use of in Signal 3. With 
the arrangement hd short-circuiting is continuous during Signal 1. 
It is therefore best avoided. With the arrangement ac short- 
circuiting of the battery tends to be, and with the arrangement 
he is certainly, produced during Signal 1, but only during the clock- 
contact closures, and is therefore practically negligible. Both of 
these arrangements can therefore be used with the Brodie-Palmer 
clock, and the way of connecting it up in the arrangement axi is 
shown specially in fig. 5. 

In order to ensure that the short-circuiting of the electromagnet of 
the chronograph effectively demagnetises it for the production of 
Signal 3, sufficient strength must be given to the armature-spring, 
and actual contact of the armature with the poles must in some 
cases be prevented, best by a screw adjustment, but extern 
poraneously by paper or other material intervening. An absence 
of inertial overthrow of the writing-lever is also desirable for 
Signal 3, which is not given in fig. 3 as well as it can be got with 
a more suitable form of signal than that actually used. 
If it be only desired to give Signal 3 — the advantage of which is, of 
course, that there is no temporary loss of time-record — there can be 

Any risk of unipolar excitation by the extra-current of the chronographic electro- 
magnet may be avoided by connecting up in parallel with the clock-contact an electrolytic 
"spark-trap" consisting of aluminium electrodes in a solution of sodium sulphate 
(ct Ostwald-Luther, Physiko-chemische Messungen, 1902, p. 397). This should, in fact, 
be in permanent use with the clock if only to prolong the life of the contact. 



366 Locke 

, usQd, instead of the switch board above described, the double switch 
frequently used as commutator on small spark-coils, consisting of 
two parallel bars playing over three contacts,^ or a Pohl's rocking 
switch with only one cross-wire. In the case of both of these also 
the spare contact can be used for the short-circuiting of stimulating 
electrodes in a secondary circuit. 

After devising the methods just described for connecting at will 
two current-paths in series or parallel (in the special case in 
question clock-contact and signal), I found that Ewald^ has 
described and figured the special arrangement be for Pohl's 
rocker with one cross- wire with a view to quite other practical 
applications than the one made here. I find too that Kronecker,^ 
in his report on chronographic methods, mentions the possibility 
of using temporary omission of the time -record as a signal 




Fig. 5. — The terminals numbered 1,2, 3, and 4 are those of the Brodie- 
Palmer Clock from right to left. (Cf. the figure showing the clock- 
connections in Proc. of the Physiol. Soc., Dec. 8, 1900, Journal of 
Physiology, xxvi., p. xii.) 

("Signal 1"). He also states that Judin^used as a signal the 
diminution of the amplitude of the excursions of the chrono- 
graph produced by the passage of a weak constant current 
through its electromagnet. I believe, although I caimot now 
recover the reference, that Kronecker has stated in a much 
earlier publication that Dew- Smith long ago employed on 
a continuous paper- kymograph a chronograph with polarised 
armature, so that currents in reverse directions produced reversed 
excursions. 

The methods of signalling described in this section were shown to 
the Physiological Society on 1st March 1902. 

' In this case the usual permanent connection of the outer two of the three contacts 
must be broken. Since the above was written, this form of switch has been applied to 
general electrophysiological work by H. G. Roaf and W. G. Smith (Proc. oi the Physiol. 
Soc, Nov. 11, 1905 ; Journal of Physiology, ixxiii., p. xiv.), wlio have made the connection 
in question breakable at will. 

2 J. R. Ewald, Arch. f. d. ges. Physiol., xlii., p. 478, 1888. 

' H. Kronecker, Arch. Ital. de Biol., xxxvi., p. 135, 1901. 

^ A. Judin. La Physiologiste Ru.sse, v., p. 67, 1898, cited after Kronecker. 



Contributions to Physiological Technique 



367 



II. A New Form of Bellows-Recorder. 

The bellows-recorder introduced by Brodie/ while capable of yielding 
the most admirable results, has in ordinary practice disadvantages well 
known to those who have worked with it, in particular the difficulty of 
making it air-tight in the first place and maintaining it so afterwards 
without stiffening its joints. The volume-recorder here described is of 





Fig. 6.— Two 



Jiie It IS seer; troiu abovi 



with the cover and recording-lever thrown back. In the upper one the 
loose open end of the condom has not been cut close off. 

very easy construction in a range of sizes suitable for a number of purposes, 
and does not tend to leak. Its air-chamber is readily renewed when 
required, consisting merely of a few centimetres of the closed end of a 
readily procurable tube of thin rubber (condom).^ 
The apparatus is shown in fig. 6. 

1 T. G. Brodie, Journal of Physiology, xxvii., p. 473, 1902. 

2 Renewal is, however, not often necessary. Made with condoms of good red rubber, 
I have had recorders in good condition for as long as two years. 



368 Locke 

A glass tube 7-8 cm. long, and of large enough bore (up to 1 cm.) is 
fixed in a rubber stopper of a diameter of 4-6 cm. flush with its broader 
upper end. It is convenient to divide an ordinary rubber stopper of suitable 
size transversely on the lathe, thus getting two of suitable diameter and 
thickness (15-20 mm.). Over the broader upper end a condom of good 
quality is drawn till only 25-5 cm. of the closed end remain clear above it. 
The condom should be of such diameter that it requires to be somewhat 
stretched to pass over the stopper, so that no creases are formed on the 
stopper while for a few millimetres it remains in light contact with the 
outer part of the upper surface of the latter. This constitutes the air- 
chamber, and it remains to make it air-tight, and furnish it with an index 
capable of recording its volume. 

A .short length (about 2o cm.) of wide (3-4 cm. diam.), rather thin-walled 
rubber tuVjing,^ is stretched over three fingers of each hand and passed 
over the end of the condom, and on to the rubber stopper, this being firmly 
held by means of the glass tube by an assistant. A little of the wide 
tubing is left above the edge of the stopper, and by virtue of its elasticity 



a 



Fig. 7. — a indicates how the rubber stopper is cut. 

and less diameter than that of the stopper lies flat upon its upper surface, 
forming an orifice through which the end of the condom suddenly pouches 
out. By varying the diameter of the wide rubber-tubing selected orifices 
of varying diameters are produced, and a corresponding variation in the 
capacity of the air-chamber obtained. The rubber-tubing around the 
stopper forms an elastic collar fixing the condom upon it. To render the 
joint ab.solutely air-tight, however, it is necessary to add a couple of turns 
of well-tied string. 

Upon the air-chamber rests a hinged cover consisting of a plate of thin 
sheet magnalium prolonged into a writing-lever stiffened by a longitudinal 
indentation. The axis of rotation of the cover is obtained in the following 
way :— 

Before the condom is applied over the rubber stopper, a transverse notch 
about 2 cm. long and 3 mm. deep is made in the latter at the edge of its 
upper surface by cutting out a slip of rubber in the way shown in profile 
in fig. 7. In the seat thus formed, but outside the condom and under the 
rubber collar, a short piece of small glass-tubing, somewhat longer than the 

1 I use for this purpose old bicycle tyre inner tubes. The short bits of tube should be 
turned inside out before being fitted, to get the seam outside. 



Contributions to Physiological Technique 369 

seat, is made to lie. The ends of this have been previously fused in the flame, 
so that they only form apertures nicely admitting a sewing-needle of No. 9 
size. Such a needle, preferably one longer than ordinary and known techni- 
cally as a " Straw," is now thrust through the rubber collar on one side, 
through the length of the glass tube and out through the rubber collar on 
the other side, taking care not to injure the condom, the projection out- 
wards of the rubber collar by the ends of the glass tube making this easy. 
Its sharp end is now cut off so that only a few millimetres project on 
either side. 

The magnalium-plate, shaped as shown in fig. 6, the hinges and w^riting- 
lever attached to it being formed by cutting and bending, is now fitted to 
the needle-axis, two beads on each side, or a bead and glass spangles as in 
fig. 6, minimising side-lash and friction. 

If preferred the plate can now be fixed to the summit of the condom by 
a touch of rubber solution. I have not myself found this to be any advan- 
tage. The merely resting plate immediately forms the condom into the 
folds natural to it, which if not disturbed remain permanently the same. 
If the condom has been fitted carefully and symmetrically, preferably with 
its summit projecting a little forward and its seam, if it have one, in 
the middle line, the folds assumed by it are surprisingly regular and 
symmetrical. 

For actual use the recorder is best fixed by means of its glass tube in a 
clamp in the somewhat oblique position shown in fig. 6. 

Many variations in the details of construction can of course be made. 
Instead of rubber, cork or suberit shives can be used to build the 
recorders on. But both of these substances are so leaky that they 
must be completely and fairly thickly covered externally with 
Prout's glue or some such substance, the edges of the condom being 
embedded in this. The cover and its hinge can be made in different 
ways of different materials. Cardboard can be used instead of 
metal, hinged by a strip of soft silk to another piece of cardboard 
slipped under the rubber collar, the rubber stopper being cut flat 
on one side to receive this. The needle-axis, if used, can be simply 
made by thrusting the needle through the thickened seam of the 
rubber collar when this is made of " inner tube " turned inside )ut, 
but this method ia by no means so lasting as the one described 
above. 

It might be thought that the use of a thin distensible rubber air- 
chamber was not permissible in a volume-recorder, but if care be taken 
that the cover moves easily on its axis and that there be no excessive 
friction of the writing-point, no rise of pressure occurs in the air-chamber 
sufficient to introduce error in ordinary volumetric work. The absence of 
any real elasticity of form in the limp air-chamber wall prevents, of course, 
any suction applied to the air-chamber acting directly upon the recording 



370 Locke 

lever, which has to follow negative changes of volume merely by virtue of 
its own weight and that of the air-chamber cover to which it is attached. 
The recorder is therefore, unlike Brodie's bellows, incapable of following 
very rapid changes in volume such as those producible by a tuning-fork 
acting on a tambour. Its rapidity of action is, however, sufficient for all 
ordinary physiological experimental work. It gives, for instance, excellent 
results as an oncograph. 

The calibration curve of the recorder resembles that of Brodie's. 
Starting from zero, i.e. with the air-chamber emptied by the weight of 
the cover and writing-lever, successive uniform increases of volume (of 
05 c.c. or 1 c.c.) produce successively greater movements of the lever for 
the first 10°-15° of its upward rotation, after which equal increases of 
volume produce sensibly et^ual movements of the writing- lever till 
distension is approached. For most purposes it may be assumed that only 
the middle portion of its range would be employed, in which no special 
reduction of the excursions obtained to correct relative values is necessary. 

The volume-recorder here described was shown to the Physio- 
logical Society on 20th January 1906. 

III. A Perfusion Stopcock. 

A stopcock was described by me four years ago,^ which among other 
possibilities permitted practically as rapid a change from one perfusion- 
fluid to another as is possible with an ordinary three-way tap, whilst 
wiving at will a bye-pass to the jfluid not being perfused, so that any ill- 
effect when turned on of its previous stagnation may be avoided. To 
accomplish this two separate three-way " keys " in one barrel were 
employed. The tap here described effects it with one key only, which is a 
considerable simplification. 

The construction, as may be seen from fig. 8, is based on the well-known 
Greiner and Friedrichs three-way tap, both the barrel and key of which 
are modified. 

To the middle of the bp-rrel at right angles to the other three tubes is 
attached a fourth tube D, to act as a bye-pass to either A or B, conveying 
the two perfusion-fluids employed. The tube C, connected with the per- 
fused organ, is functionally prolonged upwards and forwards for a quarter 
of the circumference of the barrel by a channel blown out from it and 
bulging out its front. 

Two trans.verse channels, each occupying a little more than half the 
circumference of the key, are ground out of its surface in the planes of A 
and B on opposite sides of it, so that the outer ends of the two parallel 
oblique key-borings open at the middle points of these channels, which 
form functional continuations of them. 

1 F. S. Locke, Proc. of the Physiol. Soc, March 19, 1904 ; Journal of Physiology, 
xxxi., p. xii. 



Contributions to Physiological Technique 



371 



When, therefore, the key is in the position shown in fig. 8, A is shut ofi", 
while B is connected to C. If the key be now turned clockwise through 
90°, B will still be connected with C by virtue of half of the one transverse 
channel, and the prolongation of C upwards and forwards, but A will be put 
in connection with the bye-pass D, the hitherto unemployed perfusion-fluid 
now running to waste. Another turn clockwise through 90° shuts off B 
from C, but connects A with it instead of with the bye-pass, a change fi-om 
one perfusion-fluid to the other being now effected. Another turn clock- 




Fio. 8. — For description see text. The Greiner and Friedrichs oblique 
key-borings are indicated by light dotted lines. The special additional 
passages are shown dark. The stopcock was filled with dark fluid 
before being photographed. 



wise through 90° leaves A still in connection with C, but connects B with 
D, the perfusion-fluid first employed now running to waste. Another turn 
clockwise through 90° restores the state of affairs shown in fig. 8, changing 
the perfusion-fluid to the one in use at the start. 

Midway between the second and third, and also between the third and 
fourth (first), of the positions of the key just described, A and B are com- 
pletely shut off from C and D, all perfusion being cut off, with no perfusion- 
fluid running to waste. 

Spots of different coloured enamel are fused on the ends of the key- 

VOL. I., NO. 4. — 1908. 25 



372 Contributions to Physiological Technique 

handle, and corresponding ones on the tubes A and B near the barrel, 
so that the spot on the end of the key handle pointing downwards in the 
direction of C is of the colour of the spot on the tube A, or, as the case 
may be, B, at the time in connection with it. Confusion as to fehe effect of 
movements of the key on the perfusion is thus avoided. 

The form of stopcock described was shown at the meeting of the 
Physiological Society on 20th January 1906. It is made by 
Greiner and Friedrichs, and is supplied by Messrs Baird & 
Tatlock, London. This also now applies to my earlier form of 
perfusion-stopcock (loc. cit.), which is still of special use for 
certain purposes. 



The apparatus and methods described in this paper have b^en mainly 
worked out inDr Halliburton's laboratory, King's College, London. I 
have, however, to acknowledge my indebtedness to Dr Brodie and Dr 
Waller for the facilities put at my disposal in their laboratories also. 
Grants in aid of the expenses incurred have been received from the 
Glovernment Grant Committee of the Royal Society. 



THE FORM AND MAGNITUDE OF THE ELECTRICAL RESPONSE 
OF THE EYE TO STIMULATION BY LIGHT AT VARIOUS 
INTENSITIES. By W. Einthoven and W. A. Jolly.^ (From 
the Physiological Laboratory of the University of Leyden.) (With 
twenty-five figures in the text.) 

{Received for inMicatio II 29th Sepiemher 1908.) 



CONTENTS. 



I. Historical Introduction 



11. 



III. 



VI. 



ith the Galvanomeler 



Method of Investigation- 

1. General Remarks . 

2. The Arc Light 

3. System of Lenses . 

4. Weakening the Light 

5. The Spectrum 

6. The Movable Screens 

7. The Moist Chamber and Connection of the Eye w 
Results — 

1. The Form of the Photo-Electric Reaction to Light of Moderate Intensity 

2. The Three Substances ....... 

3. Light of Different Colours ....... 

4. Rhythmical Reaction on Continuous Stimulation . . . . 

5. The Latent Period . . . . 

The Energy of the Stimulation in Absolute Measurement . 
The Relation between the Energy of the Stimulation and the 
Energy of the Reaction ...... 

Summary of Conclusions ....... 



P.\rtE 
373 

379 
379 
380 
384 
386 
387 
388 

389 
392 
405 
405 

406 
410 

414 
416 



I. Historical Introduction. 

The following paper gives an account of our study of the electrical response 
of the frog's eye to light by means of a sensitive and rapidly moving 
galvanometer. 

Before discussing our own results, it is desirable shortly to pass in 
review the principal steps by which our knowledge of the retinal currents 
in the frog has been built iip upon the foundation laid by Holmgren. 

Holmgren 2 found that when light is allowed tt) fall upon the eye of 

• The research was performed during the tenure by one of us (W. A. J.) of a Carnegie 
Research Fellowship of the University of Edinburgh; 

2 "Method at objektivera effekten af Ijusintryck po retina," Upsala Lakareforenings 
Forhandlingar, vol. i. p. 177, 1866. Physiol. Untersuch., Heidelberg, Bd. ii. j). 81, and 
Bd. iii p. 308. 



;-}74 Einthoven and Jolly 

a frog that has been kept in the dark and again when the light is removed, 
there is an increase, in the positive direction, of the current present during 
darkness. A similar result is obtained when, during continued illumination, 
the light varies rapidly in intensity. The strength of the current is, within 
certain limits, proportional to the intensity of the light. The onset and 
removal of light is attended with the same electrical changes when the 
posterior half of the eyeball alone is employed. 

In the case of the viper, rabbit, dog, and cat, a negative variation 
accompanies lighting, followed on continued illumination by a slow positive 
variation, and darkening produces a positive variation. 

Dewar and M'Kendrick,^ who rediscovered the electrical changes 
caused in the retina by light, found that " when diffuse light is allowed to 
inpinge on the eye of the frog, after it has arrived at a tolerably stable 
condition, the natural electromotive power is in the first place increased, 
then diminished ; during the continuance of light it is still slowly dimin- 
ished to a point where it remains tolerably constant ; and on the withdrawal 
of light there is a sudden increase of the electromotive power nearly up to 
its original position." 

The effect of moonlight upon the eye was found to be about equal to 
that of a candle distant a few feet. The eye is more sensitive to variations 
in light of weak intensity than to variations in light of great intensity. 

Certain of the colours of the spectrum were arranged by Dewar and 
M'Kendrick, with reference to their power of altering the electromotive 
force, in the following order — yellow, green, red, blue. 

It was found that the anterior segment of the eyeball, including cornea, 
iris, and lens, yielded a current which was not affected by light. 

The interesting fact was discovered that in the crustacean eye the 
retinal currents are reversed in direction, in accordance with the inverted 
arrangement of the sensory epithelium. In investigating the eye of the 
cat, the onset of light was found to be attended by a diminution of the 
electromotive force ; during illumination the electromotive force gradually 
rose to a point where it became steady, and on darkening a rise was well 
marked. 

Kuhne and Steiner - occupied themselves chiefly with the investiga- 
tion of the reactions of the isolated retina. They found that the electrical 
change on lighting and darkening is a complex one, the variation being first 
positive, then negative, and finally again positive. The reaction is divisible, 
according to these observers, into two parts: the first, due to the onset and 
continuance of illumination, consists of a negative variation preceded by a 
positive ; and the second, caused by disappearance of the light, consists of a 
simple positive variation. Ktihne and Steiner discuss the question 
whether the former of these parts can be divided further into a positive 
variation caused by the onset and a negative due to continuance of illumina- 

> Trans, of the Roy. Soc. Edin., vol. xxvii. p. 141, 1872-73. 
•^ Phy.siol. Untersuch., Heidelberg, Bd. iii., S. 327, 1880. 



The Electrical Response of the Eye to Stimulation by Light 375 

tion, but regard such a division as indefensible. In a later research ^ it was 
observ^ed that, in some cases, instead of a negative variation following the 
positive, the latter merely suffered diminution. This, since it occurred in 
the freshest prepatations, was regarded as the normal reaction of the 
retina. 

No current was obtained from the posterior half of the eyeball from 
which the retina had been removed, the pigment layer being left behind. 
The observers concluded from this that the pigment epithelium possesses 
no electromotive power, and no ability to give rise to electrical changes on 
stimulation by light. 

The latent period of the lighting effect M'as investigated by Fuchs^ by 
means of the rheotome. He obtained much smaller figures than have 
been observed by other workers. The method employed by Fuchs, 
however, is not applicable to the changes of long duration which constitute 
the reaction of a dark-adapted eye. 

Waller^ studied the phenomena presented by the frog's eyeball with 
the help of photographic records of the galvanometer variations. He found 
that the effects obtained in the posterior half of the eyeball or in the 
isolated retina were precisely those manifested by the injured eyeball, and 
therefore regarded the reaction of the intact eyeball as typical. Waller 
found that the deflection of the galvanometer with illumination is positive 
at the commencement and at the end, and is also positive during the 
continuance of the illumination. He regarded the final deflection as a 
subordinate feature of the main change, and described the reaction simply 
as a positive (upward) current during illumination and no such current 
during darkness. 

Some of the curves figured by Waller indicate that the positive 
deflection evoked by lighting is not simple, but of a double nature. In his 
figure 15 it is evident that the curve, after a steep ascent, is interrupted in 
a step-like manner, after which it progresses bj'' a more gradual rise and fall 
during the course of the illumination, and upon this slow deflection the 
positive effect due to darkening is superposed. The interruption of the 
rise is noted by Waller and referred to a negative restraint setting in at 
that point 

The period of latency between onset of light and positive deflection 
were found by this observer to be very long amounting in sOme cases to 
several seconds. This delay he explains as due to a period of hesitation 
dniing which two opposed currents are developed from the retina at nearly 
equal rates, and finds that a short negative swing frequently precedes the 
main positive effect. 

De Haas,* working in this laboratory, investigated in detail the strength 

' Pliysiol. Untersiich., Bd. iv., S. 64, 1881. 

2 Pflugei's Arch., BrI. 56, S. 408, 1894. . Ibid., Bd. 84, S. 425, 1901. 
■■' Pliil. Trans vol. 193, B, p. 123, 1900. 

* Liclitprikkels eii retiuastroomeii in lain qnantiUtief verband, Dissertation, Leiden, 
1903 ; and Onderzoekin<;.'n Physiol. Lah., Leiden, 2nd .ser., vol. 6. 



376 Einthoven and Jolly 

of the electrical response to light stimuli differing in intensity and duration, 
and the relation, which the response to Jight at different intensities bears 
to Fechner's law. He employed a slow-moving Deprez-d' Arson val 
galvanometer for his research. This instrument, although fairly sensitive, 
is not well adapted for following the rapid alternations of current present 
in the response of the eye. A more suitable instrument for that purpose 
is found in the capillary electrometer which was used by Gotch.^ The 
electrical reaction of the eye to light was divided by this observer into 
three portions: (1) the rise due to the sudden illumination, termed by him the 
"on-effect"; (2) the continuous change occurring during the continuance of 
illumination ; and (3) a second rise due to the sudden change from light to 
darkness, termed by him the " off-effect." In addition to these changes, 
careful examination of the records wdth a lens showed the presence in some 
cases of a small negative deflection of short duration immediately preceding 
the on-effect. The curve deduced from the electrometer records shows that 
the on-effect having reached its maximum, subsides ; this subsidence is 
checked, and a continuous effect is present during the illumination. The 
continuous effect is not necessarily steady, but in some cases gradually 
increases until it exceeds the value of the on-effect, and the off-effect 
is superposed upon it. The off-effect depends for its production upon 
previous illumination. When the illumination is of short duration — half a 
second — it does not appear ; it is just visible with slightly longer illumina- 
tion, and as the period increases the off-effect becomes more pronounced. 
The latent period of the on-eftect varies with the temperature and with the 
nature of the light. The delay was found to be shortest in the case of 
white light, longest in the ease of red light, and intennediate in duration 
with blue-violet light. It was not, however, possible under the conditions 
of the research — the light being passed through filters of coloured fluid — 
to determine the absolute intensities of the rays of diflerent wave-length 
used as stimuli. The latent period of the off-effect is shorter than that of 
the on-effect, nor does it vary with the nature of the illumination as does 
the latter. 

Having regard to the facts that the on- and off-effects are both positive, 
and that they differ in time relations, Gotch concluded that they cannot be 
regarded as merely two different aspects of one chemical change, but that 
there must be two distinct substances, one reacting: to light, the other to 
darkness. 

The results obtained by Piper ^ from the eye of the frog agree generally 
with those of Gotch. The latent periods at onset of light, according to this 
ob.server, range from 0133 sec. to 0164 sec. The lighting effect is a sudden 
increase of the electromotive force, which endures for 04 sec. to 05 sec. 
and does not exceed 1 millivolt; upoi this follows typically a slight 
diminution of the light effect. During illumination the curve remains 

' Jour, of Physiol., vol. xxix. p. 388, 1903, and vol. xxxi. p. 1, 1904. 
^ Arch. f. Anjit. u. Physiol., Suppl.-Bd., S. 133, 1905. 



The Electrical Response of the Eye to Stimulation by Light 377 

almost completely constant. The reaction on darkening has a latent 
period of about 1287 sec. and consists of a renewed increa.se of electro- 
motive force more gradual than the lighting effect and le.ss in amount. 
Immediately thereafter the electromotive force sinks at first rapidly, then 
slowly, to the original amount. Piper did not find in the case of the frog's 
eye any gradual continual increase after the first increase and diminution. 
In the course of his study of the vertebrate eye, however, he found such a 
variation. The reaction of the eye of a cat, which had not been treated 
with atropin, after a first deflection presenting the typical appearance of 
the lighting cfiect, showed a second slow positive deflection. This was not 
observed in eyes treated with atropin, and Piper attributes its presence to 
electromotive changes accompanying contraction of the iris muscles. The 
second deflection was considerably greater in amount than the initial 
positive deflection which followed the on.set of light. 

We have seen that a positive deflection during the continuance of 
illumination has been figured by several observers as an occasional, feature 
of the reaction of the frog's eye, but it is to Ishihara^ tliat w.e owe the 
recognition of this deflection as a typical constituent of the curve, inde- 
pendent of the primary positive deflection on lighting. This worker followed 
out the suggestion yielded by Waller's curves, and recognised clearly the 
double nature of the positive deflection. According to his description the 
positive deflection on lighting is at first rapid. It then continues more 
slowly until it reaches a maximum, after which it gradually diminishes. 
A close investigation shows that the ascending limb of the curve contains 
a step or notch which occurs about the same time after lighting as the 
maximum of the off-efl'ect after the moment of darkening. Ishihara 
concludes that at lighting there is a rapid positive deflection similar to that 
which occurs at darkening, although the former, owing to the slow move- 
ment of the galvanometer used by him, was less clearly visible than the 
latter, tending to be fused with tiie succeeding slower deflection which he 
names the " Helligkeitsschwankung." On one of the curves fig\n-ed by 
this obsoi'ver may be seen the positive deflection on lighting, which, after 
a slight diminution, is followed by a slow rise above the value of the fir.st, 
upon which th<i darkening reaction is superposed. 

BriJcke and Garten' have recently made valuable contributions to 
our knowledge of the retinal currents by the employment of the capillary 
electrometer and string galvanometer. By the aid of tlie latter instrument 
they have demonstrated clearly the preliminary negative variation visible 
to Gotch on examining his curves by a magnifying glass. This di'floctioti 
occurs in the majority of cases, and its latency is 0078 sec. to 00!^9 see. 
Its value is great(>st iti the freshest eyes. Tlie continuous eflect during 
illumination of the isolMted eyeball consists of a slow rise independent of. 
and two or three times greater than, the first positive variation. Its return 

' Pfliiger'.s Arch., Bil. 114, S. 5(i9, 190(5. 
-' Ibid.JBd. 120, S. 290, 1907. 



378 Einthoven and Jolly 

to the original level during lighting lasts two or three times as long as the 
ascent. The positive darkening reaction does not alter the course of the 
continuous effect. When the eye is submitted to short repeated illumina- 
tions the successive on-efFects are superposed upon the continuous effect. 
The state of adaptation of the eye has an important influence upon the 
continuous effect. When the eye is light-adapted no second rise is 
exhibited on the curve, but after the on-effect the current sinks gradually 
during illumination without, however, reaching the zero line. During a 
series of experiments on the dark-adapted eye the continuous effect is 
evoked more strongly at first and becomes progressively weaker as 
illumination is repeated. The continuous effect is exhibited not only by 
the whole eye but also by the posterior half. This is inconsistent with 
Piper's view that it is derived from the iris muscles. 

The continuous effect differs frOm that just described when the experi- 
ments are performed not upon the isolated eye but upon the eye, remaining 
in situ, of a curarised frog. Here the continuous effect does not exhibit 
subsidence during illumination but remains constant at a high level, the 
illumination being continued for half an hour. 

The latent period of the on-effect is on an average 02 sec. The very 
short latent periods obtained by Fuehs cannot be attributed to the short 
duration of the light stimuli used by him, as similar short stimuli were 
made use of by Briicke and Garten and yielded latencies of 0108 sec. to 
01 11 sec. Fatigue may be observed in the on-effect which diminishes in 
strength after repeated stimulation of the eye. The o^-effect differs from 
the continuous effect in that it is but little affected by the state of 
adaptation of the eye. The off-effect is found, as a rule, with dark- 
adapted eyes, especially after long illumination, to be stronger than the 
on-effect, while in light-adapted eyes the latter is usually as strong as or 
occasionally stronger than the off-effect. The off-effect is, as a rule, some- 
what steeper than the on-effect. 

Briicke and Garten have extended their researches over a large 
number of vertebrates and find a marked similarity of reaction in the 
different classes when the eyes are quite fresh and investigated under the 
most favourable conditions. From their results it appears probable that 
the electromotive changes caused in the eye by the stimulation of light 
are essentially of the same nature throughout the vertebrate series, and 
when allowance is made for the reversed direction of the currents, also in 
crustsuieans. 

Although the electrical response of the eye to stimulation by light has 
been studied by numerous observers, there has not, so far, been undertaken 
a systematic investigation of the electromotive changes which are caused 
by stimuli of very varying strength. Such an investigation, however, 
can, as we hope to show, contribute to our comprehension of the retinal 
processes. 



The Electrical Response of the Eye to Stimulation by Light 379 

II. Method of Investigation. 
]. General Remarks. 

We have in our work employed exclusively isolated frogs' eyes. We 
have been enabled on the one hand, by means of the string galvanometer ^ 
which, for the retinal currents, may be regarded as the most sensitive 
instrument available, tx) record and measure very weak electromotive 
forces, such as are evoked by light of extremely low intensity; on the 
other hand, we have tried by a suitable System of lenses to concentrate 
light of as great intensity as possible upon the retina of the eve under 
observation. 

2. The Arc Light. 

The weakening of any light may be continued indefinitely. On the 
other hand, the increase of the intensity of light radiating upon a given 
area is restricted within limits, theoretically as well as practically. 

We have chosen as the- source of illumination the crater of the arc light, 
which of all terrestrial sources of light possesses the greatest intrinsic 
intensity. In making this choice it has been necessary to consider, in the 
first place, whether the radiation from the crater is sufficiently constant 
for use in a series of observations an essential part of which consists in 
the use of measurable quantities of light. 

With regard to this we may observe that we have never employed the 
entire crater. If we desired to avail ourselves of the maximum of white 
light we made use of a relatively large part of the crater, while in all 
other cases only a small area of its central depression was employed. 
Thus in all our experiments the question as to tlie constancy of the crater 
light reduces itself to the question of the constancy of the light which is 
radiated per square millimetre from the crater. 

We may assume that this constancy is sufficiently guaranteed. Thus 
Violle 2 demonstrated that through variations in current intensity of 10 
to 400 amp., and of energy from 500 to 34,000 watt, the brightness of 
the crater does not change. It has even been suggested from several sides 
to take the light derived from 1 ram."^ of the crater as a light unit and 
thus raise the crater to the position of a light standard. In all probability 
the temperature of the crater is that of vaporising carbon. 

On the other hand, some difficulties arise; thus, for example, Petavel' 
says : " Even when the most favourable conditions are selected and the 
intensity of current and the length of the arc are maintained constant, 

> Cf. " Eiri neues Galvanometer," Annalen der Physik, 4 Folge, Bd. 12, S. 1069, 1903. 
" Ueber einige Anweiiduiigen dcs SaitengalvanonieterR," ibid., Bd. 1 4, S. 182, 1904. " Weitere 
Mitteilungen tiber das Saitengalvanometer," ibid., Bd. 21, S. 483, 1906. 

2 Cf. Waidncr and Burgess, Bulletin of the Bureau of Standards, Washington, 
vol. i. p. 109, 1904. 

3 Proc. Roy. Soc. London, vol. Ixv. p. 469, 1900. 



380 Einthoven and Jolly 

it is difficult to obtaiu consistent results, variations of over 5 per cent, 
being by no means unfrequent." Petavel made use of a lamp regulated 
by hand with carbons placed at right angles with one another. To what 
extent the inconstancy of his results is to be attributed to this circumstance 
we do not venture to judge. 

We have ourselves made use of a differential lamp of Siemens and 
Schuckert of 20 amperes, which burns quietly when fed either by a 
battery of accumulators at 65 volts or by a dynamo at 110 volts with 
appropriate resistances. At the same time the distance between positive 
and negative carbon is great enough to permit of our utilising the radiation 
from the centre of the crater depression. The lamp has occasionaHy 
during the course of our work shown some irregularity, but this does not 
often happen, and when it occurs the observation is repeated. The 

D, 



p 

L. 



-f 




Fig. 1. — Spectroscopic arrangement and system of lenses designed to transfer a maximum of 
spectroscopically isolated rays from the crater of the arc lamp to the pupil of the frog's eye. 

photographic records made of the movements of the string of the galvano- 
meter, where in the same way only the centre of the crater depression 
was used, give us no reason to assume that variations of 5 per cent, in the 
intrinsic brilliancy of the crater light occur with a steady lamp. 

3. System of Lenses. 

As we chiefly desired to make use of light between definite wave-lengths, 
we employed a spectroscopic arrangement. In constructing this \ve had to 
solve the question as to how a maximum of spectroscopically isolated rays 
might be obtained from the crater and transferred to the pupil of the eye 
to be investigated. 

The solution was found to be as follows: — From the slit S^ (fig. 1) a 
spectrum is formed by the aid of lens L^ and prism P in S.,. From this 
spectrum there is cut off" by the octangular diaphragm D.^ a part lying 
between definite wave-lengths. 



The Electiical Response of the Eye to Stimulation by Light 381 

Close to 1).^ there is placed a second lens L.^ which forms a sharp image 
of L^ upon a tliird lens L.,. By this last lens there is formed a diminished 
image of the spectrum wliich passes within the pupil of the eye and which 
is denoted in the figure by 83. 

Tlu} light rays which pass out from S^ and are refracted by the lenses 
and prism arrive weakened in S,. Through reHection from the refracting 
planes and absorption in tlie refracting media, a part of the light is lost, 
i^jut if we neglect this loss, then we determine as the basis of our construc- 
tion that all the light rays of a definite wave-length which are collected 
by the lens L^ from the slit must again be united in S3. 

Changes in the breadth of the slit S^ and in the horizontal dimension of 
the rectangular diaphragm D.^ biing about some variations in the nature 
of the light in use. If it is desired without changing the (juality to vary 
the intensity of the radiation, then one must especially take into account 
the height of the slit and diaphragm. We name the height of the slit and 
of the two spectra respectively -s^, ti.^ and S3. The angle of aperture of the 
cone of light falling upon lens L^ may be named <>^, that of the cone of 
light passing from this lens fOj. The analogous angles of aperture for the 
lens L3 may be named O3 and w^. 

In the space between the lenses, the prism, and the slit there is air with 
the refractive index 1. On the other hand, we leave the possibility open 
of filling the space between L3 and the eye with any medium, whose 
refractive index may be termed n. 

According to known laws we nuiy write : — 

■i3=iilH_ (1) 

Sj n sin ^0)3 

If all the rays of light which are refracted by the lens L, p;iss tlirough 
L,j, the numerical aperture of L3 must be equal to or larger than 7/ sin \u)^. 
We therefore write 

n sin 1(03 p^Xa (2) 

where N3 is tlx- numerical uperture of the lens h.^. 

In the system of Icns.-s L„ (-mnol be dispensed with. If all parts are 
exactly centred and if the requirem.'iit of formula (2) is satisfied, the light 
rays wMiich are dii-ected uj^on the point of intersection of S.^ and the optical 
axis, will all pass through the lens L.,. Here L^ does not function, but this 
lens is required to bring about for all other points of S.^ what holds good 
without the lens for the point of intersection above menti(ined. To attain 
that object the, lens L.^ must form a focussed image of the second principal 
plane of i^i upon the first principal plane of L3. 

From formuhe (1) and (2) it follows that 

*>1>1<S3 (3) 

In order to be certain that all the light which falls upon the eye can 
always pass through the pupil, we have given to S3 a fixed value which is 



382 Einthoven and Jolly 

less than the pupil diameter of any of the frogs' eyes which we have 
investigated. The question above stated as to the maximum of light 
which can be radiated from the crater upon the pupil now reduces itself 
to the question as to the most favourable values for s^, sin ^o^, and N3. 

The basis of our constiuction was, as we mentioned, that all the light 
which falls upon L^ shall enter into the pupil. If we were to leave N3 
out of account, then a maximum of light would fall upon the pupil when 
a maximum of light radiates upon Lj. For that purpose one must give 
a maximum value both to s^ and Oj ; in other words, the slit must be large, 
and at the same time the collimator lens L^ must have a large diameter 
and a short focal distance. 

But proportionally to the increase of s^ sin |0j, Ng must also be in- 
creased, and in this we soon practically reach an unsurmountable limit. 
It is among microscopic objectives that one can find lenses with the 
greatest numerical aperture. We selected from Carl Zeiss' catalogue the 
water-immersion lens D* as the most suitable lens for our purpose. The 
numerical aperture of the lens is 0-75, while the free object distance 
amounts to TS mm. The lenses with greater aperture which are 
mentioned in the catalogue have all a much smaller object distance which 
renders them unsuitable for our purpose of forming the light image in the 
pupil plane of the intact frog's eyeball. 

Formula (3) can be written in the form 

«>5i£i<N (4) 

''a — 
In our experiments Sj = 19 mm., sin JOj = 118 mm., and S3 about 033 mm. 
From this it follows that 

^jA»J^i = 0-68. 
S3 

N3 having as mentioned the value of 075, the requirements of formulae 
(3) and (4) are satisfied, and by an appropriate arrangement we can ensure 
that in fact all the light falling upon the collimator lens L^ enters through 
the pupil into the eye. 

The slit Sj is only a few millimetres distant from the crater. We 
could, of course, easily have placed a higher slit behind a larger crater, and 
much stronger collimator lenses with greater diameter are available, but 
by making use of such means we could not in any case, as we mentioned 
above, have increased the illumination of the image in the pupil. 

The intensity of illumination which we obtained by means of our con- 
struction with a lamp of 20 amp. and a simple collimator lens was already 
very great. If one imagines in place of the slit of 19 mm. in height a 
square diaphragm with side Sj = l-9 mm., and if the rays are not dispersed 
by a prism, so that in place of a spectrum of height 033 mm. in S3 a 
square white spot of light is formed of side S3 = 033 mm., then the 
illumination of this spot can be expressed in metre-candles or Lux. 



The Electrical Response of the Eye to Stimulation by Light 888 

According to Blondel ^ the light intensity of 1 mm.' of the crater surface 
amounts to 158 bougies, according to PetaveP to 147 candle-power. To 
make the two values comparable one to the other it is necessary to express 
them in the same units. We choose for this purpose the Hefner 
candle (HK). Regarding 1 bougie as equal to 1075 (HK) and 1 candle- 
power equal to 1095 (HK)^, and if the mean of the figures found by 
Blondel and Petavel be taken, a value of 165 (HK) per mm.^ crater 
surface is obtained. 

If we assume the intrinsic brilliancy of the crater light to be i(HK) per 
mm.^, then the intensity of radiation of our square spot in a direction at 
right angles to its surface equals that of a light source of 

I = Sin(HK) . . . . (5) 

The total flux of light passing out from this spot is the light flux 
which is directed upon the inner surface of a hemisphere. The centre of 
this hemisphere coincides with the centre of the diaphragm, while the plane 
upon which the hemisphere rests is the plane of the diaphragm. 

Taking Lambert's law into account we calculate the total flux of light 
as amounting to 

^ = 7rl Lumen'* .... (6) 

The flux (f>^ which forms the image in the pupil is but a part of <p, and 
indeed is 

^j=tsin2^0j Lumen . . . (7) 

where — is that part of the light which remains after the loss by absorp- 
tion in the refracting media and reflection from the refracting surfaces. 

From (f>■^^ the strength of illumination of the image in the pupil can 
easily be calculated. If the flux of light were distributed equally on an 
area of 1 M^, the illumination intensity of the surface would equal (f>^ metre- 
candles or Lux. But the flux here being concentrated on the area of the 
image =s^ square millimetres, the illumination intensity amounts to 

E=10«x^Lux . . . (8) 

To calculate the value of E numerically we remember that i = 165 HK, 
Sj=l'9 mm., S3 = 033 mm., sin |Oi = 0118. p alone remains to be deter- 
mined. Referring to Section IV., regarding the energy of stimulation in 
absolute measurement, we put p = 2,and then find for E the value 120 x 10* 
metre-candles. 

The illumination of a plane whereon the direct rays of the sun in 
zenith fall vertically through a clear atmosphere is given as 288000 

' Cf. Liebenthal, Praktische Photometrie, Braunschweig, 1907, p. 139. 

2 Loc. cit., p. 475. = Cf. Liebenthal, loc. ciL, p. 434. 

•* One Lumen is the flux of light that radiates from a puni'.Liform light source of 1 
(HK) to an area =1 of the surface of a sphere described round the light point with & 
radius =1. 



384 Einthoven and Jolly 

metre-candles.' Our image in the pupil of the frog's eye is therefore 
rather more than 400 times more strongly illuminated than this. 

As fig. 1 is only a diagram it does not show the dimensions of the 
system of lenses which we have used. We therefore give here some of 
the actual dimensions : — 

Height of slit . .1-9 mm. 



Breadth 

Distance of S^ from the anterior surface of Lj 

Diameter of Lj 

Distance of the posterior surface of Lj from 0-2 

Distance of D2 from L3 about 



11 

1650 

39-2 

825-0 

1000 



4. Weakening the Light. 

Our original plan was to weaken the light by means of smoked glasses 
in order to leave unchanged the form, magnitude, and colour of the retinal 
image, and to vary solely the intensity of the light. But it soon became 
apparent that even the best of the so-called neutral glasses, if they absorb 
a great part of the light, do not allow all colours to pass through equally. 
Before they could be used it was therefore necessary to measure the trans- 
mitting power of these glasses for each colour separately. 

As we had not an opportunity of carrying out these measurements in 
a sufficiently accurate way, we have contented ourselves with the use of 
diaphragms. In the first place the opening of lens L^ can be diminished 
by diaphragm Dj, and in the second place the opeiung of lens Lj by 
diaphragm Dg. 

The first diaphragm, which varies the illumination intensity of the image 
S3 in the pupil, leaves, it is true, the form and magnitude of this image un- 
changed, but diminishes the spot of light on the retina as it considerably 
diminishes the size of the diffusion circles which contribute not a little to 
the formation of the retinal spot. 

The .second diaphragm Dg diminishes the image S3 in the pupil, 
whereby a further diminution of the retinal image is brought about. 
Moreover, D^ intercepts the rays of light at the margin of S2. which have 
another wave-length than the rays at the centre, but the colour of the 
retinal image is thereb}' very little altered. 

By measuremejits which have been made to determine the sensitiveness 
of the human eye to very weak light, it is found that if the spot of light 
on the retina does not exceed certain limits, the illumination of the spot 
that is necessary to give rise to a sensation is inversely proportional £0 its 
area. In these circumstances the quantity of energy required for a sensa- 
tion is therefore independent of the size of the retinal area illuminated.* 

If the spot of light is not small enough the rule above mentioned does 

' Arrhenius, Lehrlnich der kosmischen Phyeik., Leipzig, p. 93, 1903. 
^ Piper and Asher ; cf. v. Kries, Zeitschr. f. Psychol, u. Physiol, d. Sinnesorg., 2 Abt., 
Bd. 41, 1907, pp. 376 and 377; also cf. Henius, Zentralbl. f. Physiol., Bd. 22, p. 229, 1908. 



The Electrical Response of the Eye to Stimulation by Light 385 

not hold good, and with larger retinal areas the amounts of energy required 
increase proportionately with the square root of the area. 

We may expect that analogous rules exist for the photo-electric reaction 
of the isolated frog's eye. In our experiments the flux of light on the retina 
was often diminished by the use of diaphragms from 1 to 10"^. When 
once it has been diminished to 10"^ or 10'* the spot of light on the retina 
has in all probability become so small that the quantity of light required 
for a photo-electric reaction has become independent of the area of the spot. 

Four groups of diaphragms were prepared. Those of the first group could 
be placed close to the collimator lens at D^, while at Dg the diaphragms 
from the three other groups served respectively for the three parts of the 
spectrum used by us. It need scarcely be mentioned that all diaphragms were 
applied centrically round the optical axis. The diameters of the diaphragms, of 
which the size was exactly controlled by aid of the microscope, were so chosen 
that the successive members of a series each weakened the light by ten times. 

The following are the dimensions of the diaphragms used: — 

Table I. 
Circular Diaphragm Di. 



Diameter. 


Flux of Light. 


39-2 mm. 


I. 


12-39 „ 


10-' „ 


3-92 „ 


10-2,, 


1-239 „ 


10-3 ^^ 


0-392 „ 


10-4 „ 


(0-124 „ ) 


10-= „ 



Table II. 
Diaphragm D.^. 



Red 


Green 


Blue 


from A = 0-670 to A = 0-590. 


from A = 0-590 to A =0-497. 


from A = 0-497 to A=0-460. 


Dimensions. 


Flux of 
light. 


Dinjensioiis. 


Flux of 
light. 


Dimensions. 


Flux of 
light. 


9-5x14 5 mm.2 


I^ 


9-5 X 27-5 mm.' 


Ig 


9-5 X 18-2 mnL^ 


lb 


(rectangular). 




(rectangular). 




(rectangular). 




Diameter 




Diameter 




Diameter 




of circular 




of circular 




of circular 




diaphragm. 




diaphragm. 




diaphragm. 




419 mm. 


10-' „ 


5-77 mm. 


10-' „ 


4-69 mm. 


10-' „ 


1-324 „ 


10-'^ „ 


1-823 „ 


io-» „ 


1-483 „ 


io-«-„ 


0-419 „ 


10-" „ 


0-577 „ 


10-3 ^^ 


0-469 „ 


10-' „ 


(0-132 „ ) 


10-^ „ 


(0-182 „ ) 


10-^ „ 


(0-148 „ ) 


io-« „ 



386 Einthoven and Jolly 

The Ikrgest of tlie diaphragms used by us at Dg has, as mentioned in 
Table II., a length of 275 mm. An image of it is formed in its full length 
by the water-immersion lens D* in the pupil of the frog's eye. The image 
so formed measures only 955 mm., while the objective is able to form 
a circular image of 1*3 mm. diameter, although this image is very much 
distorted at the margin. 

The diameter of the pupil of the eyes investigated by us has always 
been greater than 0955 mm. 

The smallest diaphragm of each group in both tables is placed within 
brackets, as it produces only on one condition the calculated weakening 
of the light. The smallest diaphragm at Dj is only used when at the same 
time the light is diminished 100 times at Dg, and the smallest diaphragms 
of the three groups at Dg are only used when at the same time the light 
is 10 times weakened at D^. 

If these conditions are not satisBed the diminution of intensity becomes 
greater than that calculated on account of the diffraction of the light. It 
can easily be shown that when the conditions are satisfied, the influence 
of the diffraction upon the image formation at all three places S^, S3, and L, 
may be neglected. 

5. The Spectrum. 

The collimator consisting of the slit S^ the lens L^ and the tube con- 
necting these two parts, is fixed upon a horizontal plank which may be 
rotated around a vertical axis. The continuation of this axis runs in 
the plane which bisects the refracting angle of the prism P, and in order 
to be able always to obtain the minimum of deviation the prism may also 
be rotated separately around this axis. 

The prism, which has walls of mirror glass, is filled with carbon 
disulphide and is 115 mm. in height. Its base is an equilateral triangle 
whose side measures 105 mm. 

The lamp is placed on the same horizontal plank as the collimator, and 
always therefore rotates together with this. 

The whole is placed upon tables which are nailed to the floor, so that 
the axis on which the lamp, the collimator, and the prism rotate holds an 
unchangeable position with regard to the system of lenses and the eye. 

There are, as already mentioned, for illuminating by the three parts of 
the spectrum, three rectangular diaphragms in use which can successively 
be placed at Dg. In order to be able to judge if the spectrum forms here 
a sharp image, and if in fact only light of the desired wave-length passes 
through the diaphragm, the pasteboard tube which lies between the prism 
and D2 for the purpose of excluding the daylight, is taken away and the 
spectrum is directly viewed. The self-regulating lamp \s replaced by an 
arc lamp with hand regulation which can easily be done without displacing 
in the slightest the collimator or other parts of the installation. Between 
the carbons of the hand-regulated lamp, a salt, either of sodium or of 



The Electrical Response of the Eye to Stimulation by Light 387 

lithium, is then placed and the light of the arc exhibits a sharp spectrum 
of lines in Dg. 

The dimensions of the rectangular diaphragms placed here, which are 
given in Table 1., are so chosen that the spectrum in its vertical dimension 
radiates exactly through the opening of the diaphragm, while each 
rectangle is so long that two previously determined easily recognisable 
spectral lines fall upon its lateral margins. 

For illumination with red the lithium line X = 0'67 and the sodium 
line. X = 059 are thrown upon the left .and right margins respectively of 
the rectangle Dgr- For illumination with green the sodium line and the 
lithium line X = 0497 are thrown upon the left and right margins of the 
rectangle D^, and for illumination with blue the lithium lines X = 0'497 
and X = 0"460 are thrown upon the left and right margins of Dab- 
When the spectrum is focussed upon Dg, in order to illuminate with 
one of these three parts, the exact position of the collimator and the prism, 
which can be easily controlled, is read from a scale. After both of these 
have been firmly screwed in their places, the hand-regulated lamp is 
replaced by the self -regulating lamp of Siemens and Schuckert. 

6. The Movable Screens. 

We must now describe the arrangement which enabled us to cause the 
light to enter the eye at the desired moment and to radiate during any 
desired time. This arrangement, which agrees in principle with that used 
by de Haas,^ is placed between the prism and Dg and consists of two parts. 
The first part is composed of a system of two black, equal-sized, vertically 
placed discs which may be rotated upon a common axis at right angles to 
their centre. This axis is parallel to the course of light rays VS^. 

The discs are pressed one against the other and cover one another 
completely, but a portion is cut out from the circumference of each, and 
by rotating one disc oyer the other, an orifice is left of any desired breadth 
through which the light can pass to the eye. 

With the aid of an electromotor the pair of discs can be rotated as 
one with exactly determinable velocity. The speed joi rotation and the 
breadth of the opening at the circumference determine tpgether the 
duration of illumination. 

In order to ensure that the light does not pass with every rotation, the 
second part of the apparatus is constructed. In this second part a small 
screen, formed like a semicircular disc, is present. By the aid of a strong 
spring the disc is rotated when a catch is withdrawn, until after half a 
rotation it is stopped by a second catch. In one position of tho screen 
the light rays may pass, in the other they are intercepted. 

The catches art; brought into action by an electromagnet which is 
ed by a battery of accumulators. The circuit wiiich carries the current 

' Loc. cit. 
VOL. I., NO. 4. — 1908. 26 



388 Einthoven and Jolly 

from tho' battery to the electromagnet is interrupted in two places. The 
first place lies near the galvanometer, where the circuit can be opened 
and closed by the observer by means of a key. The second place lies near 
the rotating discs, where the circuit is closed automatically for a short 
time after every five revolutions of the pair of discs. 

During the time when the circuit is thus automatically closed, both the 
discs and the screen are in a position to permit of the rays passing. 

In this way the observer is easily able to illuminate the eye only once 
during a definite short period. He may close at any moment the circuit 
of the electromagnet, and then merely wait until the electromagnet comes 
automatically into action. So soon as this has done its work, which is 
accompanied by a clicking sound, easily audible, the observer opens the 
circuit again. 

If it is desired to radiate the light during a longer period, the pair of 
discs are placed at rest, while the light is allowed to pass freely by the 
opening in the circumference which is made large. The action of the 
magnet is then dependent alone upon the closing of the circuit by the 
observer. 

An electrical contact arrangement is connected with the screen, by means 
of which a current is made at the exact moment at which the light passes, 
and is broken at the exact moment of interception of the light. This 
current sets in motion a signal recording these times photographically. 

The signal resembles a small string galvanometer, and consists of a 
silver strip 120 mm. long, 143 /x broad, and 6 4/* thick, stretched between 
the poles of a permanent magnet. It registers the make and break of tiie 
current with a latency which has been found by previous measurements 
not to exceed 0001 sec. 



7. The Moist Chamber and the Connection of the Eye with 
the Galvanometer. 

In a large, well-illuminated room, where the galvanometer is arranged, 
there is constructed a small dark room. This consists of a frame of wood 
covered with linen and paper. Its base is rather more than one square 
metre, and its height rather more than two metres. Within this dark room, 
resting upon a stone pillar, is the moist chamber containing the eye. The 
moist chamber, as is usual, consists for the most part of glass, and has an 
opening in front through which the tube of the microscope is inserted. 

This tube reaches the wall of the dark room, and is here closed by the 
lens Lg and the diaphragm Dg (fig. 1). Everything in front ^ of Dj, and thus 
outside the dark room, is covered with pasteboard tubes and black cloth*^ so 
that no light can enter the dark room through Dg, other than what is sent 
from the crater through the slit of the collimator. 

• Following Helraholtz, we designate the direction towards the light source, forward, 
away from the light source, backwartl. 



The Electrical Response of the Eye to Stimulation by Light 389 

The microscopical objective, Zeiss' water-immersion D* moistened with 
Ringer's solution, is placed almost in contact with the cornea of the frog's 
eye. The latter is connected with a pair of du Bois-Reymond's non- 
polarisable electrodes in such a way that one takes the potential of the 
cornea and the other of the fundus oculi (fig. 2). 

The moist chamber is so arranged that the air within can be saturated 
with water vapour, while the eye and the electrodes which are attached to 
separate glass tubes projecting freely through openings in the floor of the 
chamber, and which are nowhere in contact with its moist walls, remain 
completely insulated. 

The electrodes are, by means of insulated wires, led to the galvanometer 
in a manner similar to that which has been employed in this laboratory for 
recording electrocardiograms.^ One has in this way an opportunity of 
compensating the current of the resting retina, of measuring in a simple 
and rapid manner the resistance of the preparation, and of regulating the 
sensitiveness of the galvanometer as desired. 




Fig. 2. — Arrangement of the eye in relation to water-immersion 
lens and non-polarisable electrodes. 

We need scarcely mention that we have repeatedly controlled the 
exactitude of our adjustments. The insulation resistance lies as a rule 
between 10^° and 10^^ ohm. The crater is so large that the cone of light 
radiating through the slit extends all round beyond the margin of the 
collimator lens. All lenses are accurately centred, and we experimentally 
determined that all the light passing through the outer margins of L^ 
contributes to the formation of S3. 

In conclusion, we did not omit to ascertain that all enclosing arrange- 
ments were light tight. If Lj is covered, all diaphragms and the movable 
screens may be withdrawn without the slightest photo-electric response 
being obtained. 

III. Results 

1. The Form of the Photo-electric Reaction to Light of 
Moderate Intensity. 

If the isolated eye, which has not shortly before been exposed to strong 
light, be illuminated by rays of intermediate strength, a form of curve is 
obtained similar to that recorded by previous observer.s. 
1 Cf. "Weiteres iiber das Elektrokardiogranim," Pfliiper's Arcliiv, Bd. 122. p. 517, 1908. 



390 Einthoven and Jolly 

Fig. 3 may serve as an example. The curve shown in this figure miist^ 
like all the curves here reproduced, be read from left to right, while the 
connections of the eye with the galvanometer are made in such a way that 
a current passing from the cornea through the instrument to the posterior 
surface of the eye deflects the image of the string in an upward direction. 
An action current in this direction may be termed positive, and in the 
reverse direction negative. 

Fig. 3 gives the reaction following on a momentary illumination. In 
the rectangular system of co-ordinates 1 mm. of an abscissa represents 0'5 
sec, 1 mm. of an ordinate 100 microvolts. The illumination is by means 
of green light, which is reduced to 001 of its original intensity, and which, 
in accordance with the two tables given in the previous chapter, we may 
name 10"^ Ig. The duration of the illumination is 01 sec. 

One observes, after a latent period, a small preliminary negative deflec- 
tion A which is immediately followed by an upward movement of the 
string. After the curve has reached a somewhat acute peak B, it sinks 
first rapidly, then more gradually, but while still distant from the zero 
line it mounts again. This latter ascent begins 25 to 3 sec, after the 
beginning of the illumination, while much later the curve reaches its second 
maximum C, which lies about 1 mm. higher than the first positive peak B. 
Finally, the curve gi-adually regains the zero line. 

The potential differences which have given rise to these three summits 
amount to — 

For the preliminary negative deflection A . . — 70 microvolts. 

For the first positive summit B . . . 1050 

For the second positive summit O . . .1150 

The form of the photo-electric reaction when evoked under similar 
circumstances to fig. 3 is always essentially the same, but the absolute size 
of the deflections as well as their proportional size may differ. When a 
number of curves are compared one with another, which all agree in having 
summit B of the same height, the deflection A as well as the slow succeeding 
wave C may show very different heights. 

In the case of illuminations of short duration, the energy of the light is, 
as may be expected, the measure of the stimulation.^ A stronger light 
must shine a shorter time than a weaker in order to produce the same- 
effect. 

In fig. 4 is seen a curve which is obtained by illuminating another eye 
during 001 sec. with the full amount of Ig. In the system of co- 
ordinates 1 mm. abscissa is again 0'5 sec, but the sensitiveness of the 
galvanometer is now ten times greater, so that 1 mm. ordinate amounts to 
10 microvolts. 

Notwithstanding that the energy of the light stimulation is greater, 

' Cf. de Haas, loc. cit. 



The Electrical Response of the Eye to Stimulation by Light 391 
the amount of the potential diiferences brought about is here less than in 



}t: 



o ^ 



II 



Q . 




S I 



tTTi <; -e 



i 

fig. 3. This need not surprise us, as the curve in fig. 4 is obtained from 
another eye, arid considerable differ-ences exist in the reaction intensities o[ 
different preparations. 



392 Einthoven and Jolly 

The absolute amounts are — 

For summit A .... —15 microvolts. 
B . . . .150 

„ C . . . . 230 

In proportion to B, both A and C are greater than in fig. 3. 

If the illumination is weaker and continued for some time, then one 
observes at the moment when darkening begins after a latent period, a new 
elevation of the curve, an off-effect A^. 

In fig. 5 there is given an example. In photographing the movements 
of the string, the sensitive plate is here moving more quickly, so tjiat in the 
system of co-ordinates 1 mm. abscissa is now equal to 0-2 sec. The sen- 
sitiveness of the galvanometer is regulated in such a way that 1 mm. 
ordinate amounts to 20 microvolts. This is visible in the control curve, 
which is obtained at the end of the curve by suddenly introducing into 
the circuit a potential difference of 200 microvolts. The intensity of the 
illumination is 10~* Ig, while the duration of illumination as indicated 
by the signal amounts to 4*58 sec. 

The latent period of the preliminary deflection A is 01 sec, that of the 
off-effect Aj is 1*8 sec. 

The absolute amounts of the potential differences are — 

For A . . . . - 20 microvolts. 
„ B 384 „ 

„ C 670 „ 

The potential difference of A^ cannot easily be given. It may be 
estimated at 60 microvolts, and is measured after connecting the beginning 
and end of the curve A^ by a line running in the course of the main curve. 
We may remark here that the off-effect Aj is in general higher the longer 
the illumination has been continued. 

Further, we draw attention to the fact that the curves in figs. 3, 4, and 
5, although they may differ in details, are yet formed essentially in pre- 
cisely the same way» It may also be remarked that fig. 15, which we shall 
discuss later, shows no essential difference from figs. 3, 4, and 5. 

2. The Three Substances. 

The oomplicated structure of the curves above described and the striking 
fact that a deflection in the same direction takes place both on illumination 
and on darkening, suggest that there are in the eye two or more different 
processes occurring partly simultaneously, partly successively, whose fusion 
determines the form of the electrical reaction. 

Further investigation confirms this suggestion, and if recourse is had to 
very weak or very strong light, it seems even to be possible to bring about 
a separation of the supposed processes. The phenomena are explained in 



The Electrical Response of the Eye to Stimulation by Light 303 



mm- 



iiil 



ifHHfg 






^•■^ 



o 



!•= 



i: 



§3 



<1 o 



.4* 

5^ 



$•2 



l-i O) _ 

5 a! -C 

3 CO ^ 




£ _• 



c5 « 

II "o 

e 8 



J II 



Is 



o a 

r 

.13 

H 

I 



394 Einthoven and Jolly 

the simplest manner by the assumption that the processes are three in 
number, whether they are together dependent upon the same substance or 
each upon a separate one. For the sake of convenience we shall speak of 
three substances, and as we do not intend in the meantime to attempt to 
define them anatomically in the eye, we shall try to describe their charac- 
teristics and to mention the conditions under which their effects appear as 
pure as possible. 

The First Substance. 

The substance which we have termed the first reacts more quickly than 
the other two. On lighting it displaces the image of the string downwards, 
on darkening upwards. Its effect, which can be obtained pure only during 
a short period, is very marked in a light-adapted eye — which for the sake 
of brevity we may call a light eye^ — and the more so the stronger the 
illumination has been. 

From the nature of the case the darkening stimulation can be employed 
very strong in a light eye. and accordingly an eye which has been illumin- 
ated strongly develops on darkening a huge positive potential difference. 
The upward deflection so evoked cannot, however, be of long duration, 
because by the darkening the light eye is beginning to be changed into a 
dark eye, and therefore the effect of our first substance is no longer so clearly 
indicated. 

An example of these phenomena is reproduced in fig. 6. Here 1 mm. 
abscissa is 0-2 sec, 1 ram. ordinate is 20 microvolts. The eye has been 
illuminated for a fairly long time with almost the strongest light at our 
disposal ; that is to say, practically the full amount ^ of I„. At a a constant 
potential difference of 200 microvolts is introduced into the circuit whereby 
the image of the string is displaced 10 mm. downwards. At b the eye is 
suddenly darkened. It is seen that after a latent period which, according 
to a rough estimate, has a durati-^n of about 004 sec, a high positive 
off-effect A^ occurs, whose top is elevated 29*5 mm. above the original 
position of the string. If the indicating speed of the galvanometer left 
nothing to be desired, this height should correspond to a potential difference 
of 590 microvolts, but the galvanometer was in this case given too great 
sensitiveness to be able to follow exactly the extremely rapid current 
variation. The amount of the latter must be calculated from the form of 
the curve.2 We have not made this calculation in detail but it is easy to 
ascertain that the actual potential difference attained is considerably higher 
than the amount above mentioned. We may estimate it at about 1000 
microvolts (1 millivolt). 

' An eye which is dark-adapted may be teemed a dark eye. The terms are analogous to 
" Lichtfrosch " and " Dunkelfrosch," which are commonly used. 

^ There was placed in the path of the light rays a very weak smoked glass, whereby 
the intensity was a little lessened. 

^ For a method of calculation cf. " Weitere Mitteilungen liber das Saitengalvanometer," 
Annalen der Physik., Bd. xxi., S. 483, 1906. 



The Electrical Response of the Eye to Stimulation by Light 395 

This enormous off-efFect is thus about sixteen times higher than that 
shown in fig. 5. It is, as we have already remarked, of but short duration. 
By the darkening the light eye is beginning to be changed into a dark eye. 
The image of the string is seen to descend at first rapidly, then more 
slowly. 

Although in the light eye the conditions are less favourable for the 
lighting than for the darkening stimulus, it is nevertheless possible to 
apply the former in either of two ways. In the first place, we may 
suddenly increase the intensity of the light that is radiating on the eye, 
and secondly, we may darken the light eye for a short period, so that it has 
not yet become a dark eye., and then suddenly illuminate it. 

The second method gives better results than the first, and we possess 
numerous curves where, after a short darkening of a light eye. a strong 
light stimulus was applied. An example of this is seen at the end of fig. 6. 
The potential diflferenee of 200 microvolts which was introduced at a is 
cut out at c, and the strong white light I„ is suddenly allowed to radiate 
upon the eye at e. The reaction A of the first substance attains here as a 
downward directed deflection a value of —90 microvolts. The positive 
deflection B following thereon belongs, as we shall explain later, to the 
action of the second substance. It has nothing to do with the reaction of 
the first substance, and bcQomes smaller the more the reaction of the first 
substance appears unmixed. The positive deflection E (lighting reaction 
of the second substance) must be reckoned as beginning at the lowest point 
of the negative deflection A. The height of B must thus be measured 
from this lowest point to the peak. It amounts at e in fig. 6 to 370 
microvolts, and is therefore more than four times greater than the lighting 
reaction A of the first substance. 

But it is not difficult to increase the lighting effect of the first substance 
and at the same time to diminish that of the second substance, which is, in 
other words, to produce the lighting eff'ect of the first substance more 
purely. For that purpose one requires to darken the light eye during a 
shorter time so that it preserves better the attributes of a light eye. In 
an eye that is darkened during a very short time the lighting eflfect of 
the first substance can even surpass that of the second. The negative 
deflection A becomes then larger than the immediately following positive 
wave B. Fig. 7 may illustrate this. 

Here 1 mm. abscissa is equal to 02 sec, 1 mm. ordinate to 21 
microvolts. The eye is illuminated with white light of the full strength 
I,. The periods of lighting are, as in all our other figuree, denoted by I, 
those of darkening by d. 

If ?i2 01 fig. 7 be compai-ed with l^ of fig 6, it is clear that the reaction 
of the first substance at l^^ has become greater, that of the second substance 
has become smaller than at I.,. The reaction on lighting at /.g ^^ followed 
upon a shorter period of darkness. We may remark, however, that at ly. 
B is still twice- as great as A (see diag., fig. 8, a). 



396 



Einthoven and Jolly 



On further shortening the period of darkness the proportion becomes 
continuously changed in favour of the first substance, so that on lighting 
at l^, l^, li, l^, and l^^ the negative wav.e of the first substance is about as 
great as the positive wave of the second (see diag., fig. 8, b), and finally, 




Fig. 7.— The reaction of the first substance. Light eye. Absc. 1 mm. =02 sec. Ordin. 1 ram. = 
21 microvolts. White light. Intensity of illumination = Iw. Z, light ; rf, darkness. 

after very short periods of darkening, the positive wave of the second 
substance at l^^, l^, l^, Ig, and l^^ becomes smaller than the negative of the 
first (see diag., fig. 8, c), especially at l^, l^, and l^^ the lighting effect of the 
first substance appears almost pure. 

To understand full}- the significance of fig. 7, we must further pay 
special attention to tlie darkening deflections. As already mentioned, the 




Fig. 8. —Diagram of the reaction to strong light of a light eye which has 
previously been darkened for a short time. The preceding period of 
darkness has been longest in a, shorter in b, and shortest in c, and the 
upward deflection, due to the action of the second substance, becomes 
progressively less, while the down^eard deflection due to the first sub- 
stance becomes greater. I, light ; d, darkness. 

first substance reacts in a completely light eye on stimulation by darkness 
with the development of a huge positive potential difference, which, 
however, cannot have a long duration, since by the darkening the light eye 
begins to change into a dark eye. The descent of the image of the string 
caused thereby, a descent which is produced by the action of the second 
substance, is relatively slow. It can easily be distinguished from the very 



The Electrical Response of the Eye to Stimulation hy Light 397 

rapid descent with which the first substance reacts to a strong light 
stimulation. 

In the case of darkening of moderate duration, as for example at d^ 
and d^, there therefore appears a curve which begins with a rapid ascent, 
continues in the middle with a slow descent, and concludes with « rapid 
descent (see diag., fig. 9, a). 

The rapid ascent at the beginning and the rapid descent at the end are 
the reactions of the first substance ; the slow descent in the middle is 
produced by the action of the second substance. 

As the duration of the darkening is shortened the middle pait of the 
curve diminishes more and more until, with very short darkening (a flash 
of darkness), it entirely disappears. In these circumstances, therefore, the 
action of the second substance is totally cut out and the curve shows the 
pure reaction of the first substance. These phenomena are fully depicted 
in fig. 7. The effects of darkening at d^ and d^^ are reproduced diagram- 



J 





]- "1 T" ~T_r "ir"d ■' 



Fig. 9. — Diagram of the reaction of a light eye to darkness. The period of darkening 
is progressively shortened from a to e. The initial upward and terminal down- 
ward deflections are due to the first substance, and the intermediate slow descent 
lo the second substance. In e, where the darkening is veiy short (a flash of 
darkness), the reaction of the fir^t substance is seen pure. /, ti^ht v d, darkness. 

matically by fig. 9, b, at d^ by fig. 9, c, and the effects of the very short 
periods of darkening at d^, cLj, and dg, where the reaction of the first 
substance appears practically pure, are shown by fig. 9, e. 

The absolute amount of the darkening reaction is in d^ very great 
The deflection being 31 mm. in height, should represent a potential 
difference of 650 microvolts, if the galvanometer had been able, with the 
sensitiveness here employed, to follow the current variations exactly, but 
the actual potential difference is, for reasons explained above, much greater, 
and must be estimated at more than 1200 microvolts. 

The absolute amounts of the lighting reactions are not so great, but 
the potential differences here developed may nevertheless be termed 
considerable. At l^^ and l^ downward directed deflections occur of 17 
and 18 mm. which, if the galvanometer were rapid enough, would represent 
potential differences of 357 and 378 microvolts, but must in reality be 
estimated at more than 600 microvolts. 

In cases where the darkening is of short duration and the darkening re- 
action is greater than the lighting reaction, curves are recorded as at d^ and d^, 
where the string do*^s not return to its original position (see diag., fig. 10, a). 



398 



Einthoven and Jolly 



In ng. 10, b, on the contrary, is diagrammatically reproduced a curve 
with a long period of darkness as at d-^^, where the string descends below 
its original position. 

The Second Substance. 

The second substance reacts less quickly than the first. On lighting 
it moves the string with moderate velocity upwards, and on darkening 
slowly downwards ; thus on applying stimuli of the same kind, it develops 
potential differences which are opposed to those of the first substance. Its 
effect appears almost unmixed in a dark eye which is illuminated for a 
short time by weak light. 

If when illuminating with light of very low intensity, the darkening 
follows rapidly upon the lighting, in a similar way as in a momentary 
illumination, there is recorded a curve of simple form with a steeper 




Fig. 10. — Diagram of the reaction of a light eye to darkness. In o, where 
the darkness is of short duration and the darkening reaction is greater 
than the lighting, the curve does not return to its original position. 
In 6, where the darkening is .of long duration, the curve descends below 
its original position. I, light ; d, darkness. 



ascending part which is evoked by the lighting and a less steep descending 
part evoked by the darkening. The top of the curve lies, within certain 
limits, higher the more the energy of the illumination is increased either 
by using greater intensity or longer duration of the light. These limits 
are determined by the functioning of the other two substances, which, 
when their effects become perceptible, influence the form of the curve and 
considerably complicate it. 

The series of four photographs which are taken from one and the same 
eye and are reproduced in figs. 11, 12, 13, and 14, may serve as examples 
of illumination of this description. Here 1 mm. abscissa is equal to 0*2 
sec, and 1 mm. ordinate equals 4 microvolts. The eye is illuminated 
each time by green light, which, by making use of diaphragms, is reduced 
to • one ten millionth of its original intensity, and which we thus name 
10 '^ Ig. The duration alone of the illumination is varied. From fig. 11, 
where this duration amounts to 048 sec, it increases gradually : in fig. 12 
it is 112 sec ; in fig. 13, 19 sec. ; and in fig. 14, 3 sec 

The heights of the summits increase regularly with the increasing 



The Electrical Response of the Eye to Stimulation by Light 399 




Figs. 11, 12, 13 and 14. — The reaction of the second substance. Dark eye. 
Absc. 1 mm. =0"2 sec. Ordin. 1 mm. =4 microvolts. Green light. Intensity 
of illumination =10-^ I^. /, light ; d, darkness. / control curve precedes 
reactions 11 and 12. The durati^m of lighting increases from fig. 11 to fig. 14. 
In figs. 11 and 12 the second substance is actina; alone. In fig. 13 the action 
of the third substance is seen at C, and in fig. 14 also that of the first at A,. 



400 Einthoven and Jolly 

duration of the lighting times. In the four figures the successive amounts 
are 31-2, 66-8, 110 and 116 microvolts The first and second members of 
the series — figs. 11 and 12 — exhibit the reaction of the second substance 
practically unmixed. In the third member of the series — fig. 13 — a 
complication begins to be visible. There is formed another summit at 
C which, as we shall later explain, must be considered as the efffect of 
the action of the third substance. 

In the last member of the series — fig. 14 — the complication has con- 
siderably increased. Not only does summit C of the third substance 
appear more clearly, but the action of the first substance also becomes 
apparent; thus at A^ we perceive the darkening effect of this substance. 

The lighting effect of the first substance is the negative deflection A. 
This does not yet appear in fig. 14, which need not surprise us, since the 
first substance acts specially strongly in a light eye, and therefore from 
the nature of the case the darkening reaction must appear sooner than 
the lighting reaction. Nevertheless the lighting eflfect makes itself 
appreciable to some extent during the record of the curve ; for if we 
regard the duration of lighting and the proportional heights of the 
summits of the four figures in the series (figs. 11, 12, 13, and 14), it is 
evident that the summit height in fig. 14 is only little greater than that 
in fig. 13. The increase is only 6 on 110 microvolts, while the duration 
of lighting is increased from 1"9 to 3 sec. It is the lighting effect of 
the first substance which here hinders the development of a higher 
summit B. 

The Third Substance. 

The third substance reacts in the same direction as the second, but 
more slowly. On lighting it displaces the image of the string slowly 
upwards, and on darkening still more slowly downwards. So much slower 
is the third substance than the other two that its effect in a recorded curve 
appears, as a rule, almost entirely isolated, and can thus be easily followed. 

The summit of the wave which is evoked by the action of the third 
substance is denoted by the letter C in figs. 3, 4, and 5. In the case of 
the momentary but very strong illumination of fig. 4, this summit occurs 
16 sec. after the beginning of lighting. In the case of the less strong 
illumination of figs. 3 and 5, it occurs 20 sec. after such beginning. 

The effect of the third substance falls out under two conditions : (1) in 
a dark eye exposed tx) very faint light for a short time, and (2) in a 
completely light eye. The first condition is realised in figs. 11 and 12, 
while on the contrary, in the figs. 13 and 14 of the same series, where the 
energy of the light stimulation exceeds certain limits, the eflfect of the 
third substance appears again. 

The second condition, a completely light ej^e, is practically realised in 
fig 7. After the short lighting at l^.^ there is seen in the curve no slow 
deflection which could be the analogue of the deflection C in figs. 3, 4, and 5. 



The Electrical Response of the Eye to Stimulation by Light 401 

In the case of an eye which has been exposed for a long time to strong 
light and thereafter maintained in darkn -ss for a few minutes, so that 
it does not yet differ much from a completely light eye, we see the third 
substance acting very weakly. 

Fig. 15 may serve as an example of this. The summit C of the third 
substance is here elevated only 1 mm. or 18 microvolts above its base. It 
will clearly appear that this amount is both absolutely and relatively 
very low when the summits B, A^ and C of figs. 5 and 15 are compared 
one with another. 

Composite Curves. 

Having thus considered the effects of the three substances separately, 
it is not difficult to analyse the composite curves of figs. 3 and 4 into their 




Fig. 15. — The reaction to light of an eye which has previously been exposed for a long time to 
strong light and thereafter darkened for some minutes. The third substance is seen acting 
feebly at C Absc. 1 ram. =0*2 sec. Ordin. 1 mm. =18 microvolts. E, control curve. 
I, light ; d, dtirkness. 



component parts. The negative deflection A is evoked by the action of 
the first substance, the summit B by that of the second substance, and 
the summit C by the action of the third substance. 

In fig. 16 those three reactions are diagrammatically represented as 
three separate curves. By superposition of these curves we obtain a 
figure similar to what is shown in figs. 3 and 4. 

In fig. 5, where the lighting duration is 458 sec, there is seen in 
addition to the deflections of figs. 3 and 4 still another summit A^. This 
is produced by the darkening reaction of the first substance. As already 
mentioned, the upward movement of the string which is produced by the 
darkening reaction of this substance is only of short duration. In agree- 
ment therewith the summit A^ is superposed in such a way upon the 
slow wave C that the form of the waves may easily be recognised separately 
(see diag., fig. 17). 

What has been noted with regard to fig. 5 is also true of fig. 15. The 
two figures, which differ considerably in outward appearance, nevertheless 



402 



Einthoven and Jolly 



exhibit complete agreement in essentials. The analogous summits are 
denoted in both figures by the same letters A, B, A- and C. 

Specially remarkable are the curves obtained if the duration of lighting 
of a dark eye is systematically changed, and we wish to direct attention 
more particularly to the ofF-effect in such cases. If the duration of the 
lighting is very short and the light weak, then, as already mentioned, 
the effects of the second substance appear umnixed. The off-effect here 
consists in the descent of the curve to the zero line. 




Fig. 16. — Diagrammatic representation as three separate curves of the reactions 
to light of the three substances. A, first ; B, second ; and C, third sab- 
stance. /, light ; d, darkness. 

If the duration of the lighting is a little longer and the effects of the 
other two substances begin to become perceptible, the off-effect is determined 
by the resultant of three forces. Tlie first substance tends to displace the 
image of the string upwards. It is at first acting weakly, but its strength 
increases regularly during illumination so that it soon surmounts the 
effect of the other substances. In the case of longer lighting the off-effect 



~T^ 




FiO. 17. — Diagram of the combined action of the three substances. 
A is the lighting, and Aj the darkening reaction of the first sub- 
stancf ; B is the action of the second, and C of the third substance. 

therefore is always an upward movement which increases with the 
duration of the lighting. 

The second substance tends to depress the image of the string, acts 
first with moderate strength, but decreases gradually during lighting. As 
the second substance in particular is acting in a dark eye, the conditions 
for its functioning grow during the illumination more unfavourable. A 
strong darkening effect cannot be expected in a dark eye. 

The third substance is so slow that the darkening effects of the first 
and second take place usually at a moment when the third substance is 
still tending to displace the string upwards. The darkening effect of the 



The Electrical Response of the Eye to Stimulation by Light 403 

third substance itself, consisting in a slow descent of the string, appears 
much later and fairly isolated. 

The general result is that we can observe in a series of curves obtained 
from a dark eye, whore the light has gi-adually been lengthened in duration, 
that the darkening effect, in the first curves a negative deflection, becomes 
in the later ones a positive deflection. The latter, on further lengthening 
the duration of lighting, gradually increases in size. In the conflict 
between negative and positive deflections there is sometimes seen an 
upward movement, which is immediately preceded by a small downward 
one. 

To illustrate reactions of this description we reproduce in the first 
place fig. 18, where 1 mm. abscissa = 02 sec, and 1 mm. ordinate = 2 
microvolts. The intensity of the stimulation cannot be given exactly, as 
the weakening of the light has been brought about by the aid of a 




Fig. 18. — Conflict between the reactions of the three substances. Dark eye. Absc. 1 mm. = 0'2 
sec. Ordin. r mm, =2 microvolts. Green light. Intensity of illumination = 10-* Ig. Z, light ; 
d, darkness. Instead of the usual upward deflection on darkening, due to the first substance, 
there is seen a downward deflection Bj caused by the action of the second substance. 

coloured light screen in addition to the usual diaphragms. The light 
intensity may be estimated at 10"^ or 10"^ Ig. 

One observes that in these circumstances the negative deflection A is 
not present, while the sunnnit B is developed in the ordinary way. As a 
result of the darkening there is seen in place of the usual upward directed 
summit A^ a descent of the curve which is denoted by Bj. So deep a 
depres.sion at Bj as occurs here is not found again in our whole collection 
of photogi'aphs. 

We desire, in the second place, to draw attention to figs. 19, 20, and 21, 
which were obtained successively from the same eye. In all three figures 
1 mm. abscissa equals 02 sec, 1 mm. ordinate equals 26 microvolts. 
In each figure there is reproduced in front of the photo-electric response a 
control curve E which is obtained by introducing a potential difterence of 
200 microvolts, which remains constant for a short time. The summits A 
and B arc present in the three figures. In the first of the series, 
fig. 19, the darkening reaction at B^ is seen as a sudden descent of the 
VOL. I., NO. 4. — 1908. 27 



404 



Einthoven and Jolly 



curve after a short period of illumination. In fig. 20, where the illumination 
has la.sted somewhat longer, there is evidence of the conflict between the 
actions of substances 1 and 2 
immediately followed by a positive deflection at A^ 



At Bj is seen a negative deflection which is 




Figs. 19, 20, and 21. — Conflict between the reactions of the three substances. Dark 
eye. Absc. 1 mm. =0'2 sec. Ordin. 1 mm =26 microvolts. White light. 
Intensity of illumination = "2 Iw. /, light; d, darkness E, control curve. 
The three curves are from the same eye. The duration of lighting increases from 
fig. 19 to fig. 21. The darkening reaction is in fig. 19 a descent of the curve 
Bj ; in fig. 20 a descent Bj followed by an ascent Aj, and in fig. 21 an ascent A). 

In the last figure of the series (fig. 21), where the illumination has lasted 
longer still, the characteristics of the organ as a light eye, to which 
condition the eye m\ist attain more and more, appear in the foreground. 
Thfe darkening eff'ect at A^ here consists of a pronounced positive deflection. 



The Electrical Response of the Eye to Stimulation by Light 405 

Finally, we draw attention to the differenct- which exists between figs. 
6 and 7 on the one hand, and the series of figs. 19, 20, and 21 on the other. 
In both cases very strong light was used, but in the first-mentioned figures 
we had the organ as completely as possible a light eye. In the last- 
mentioned figures we endeavoured to keep it as completely as possible 
a dark eye. 

3. Light of Difterent Colours. 

We shall now mention some further results of our investigatrons, and 
in the first place, those obtained by illuminating with rays of different 
wave-lengths. Our expectation that we should find marked differences in 
the form of the photo-electric reaction when light of different wave-lengths 
is employed for stimulation has not been realised. When experiments are 
performed with light of a single colour there appear very different forms 
of photo-electric reactions according as we have to do with a light eye or 
with a dark eye, and according as we have illuminated the eye with 
weaker or stronger light, and during shorter or longer periods. This is 
sufficiently proved by the figures of our plates. To study the influence 
which the variation of colour exerts upon the development of these 
numerous forms would require a long and detailed investigation which we 
have not had an opportunity of carrying out. We have only been able in 
our investigations to confirm what is already known, viz. that for the 
same energy of stimulation the reaction to green rays is stronger than that 
to red and blue.^ 

4. Rhythmical Reaction on Continuous Stimulation. 

We take the opportunity of referring in a word to the possibility of a 
rhythmical reaction to constant illumination. As a rule the eye reacts to 
constant illumination with an electrical current which increases and 
decreases very gradually, but in some cases it is open to question if this 
rule holds good. In fig. 22 an example of this is reproduced. We have 
here to do with a light eye which is illuminated by strong white light 
02 I„. Here 1 mm. abscissa = 0"2 sec, 1 mm. ordinate=lo microvolts. The 
string shows, during each period of illumination l^, l^, and /g, rhythmical 
Oiscillations which fail (luring the periods of darkness d■^^ and fZg. 

After making the record shown in fig. 22, the eye was maintained for a 
(juarter of an hour in the dark and thereaftci- exposed again to light of the 
same strength, with the result given in fig. 15. The value of 1 mm. 
abscissa has reniainetl unchanged and is equal to 02 sec, while the 
value of 1 mm. ordinate is increased to 18 microvolts. At E a control 
curve is recorded, while the reaction evoked by a short exposure to strong 
light shows the usual summits A, B, A^, and C. We draw attention to the 

' Himstedt and Nagol, Bericbte der Naturforsch. Ges. zu Freiburg, Bd. xi., p. 153, 
1901. 



406 



Einthoven and Jolly 



pure line -which appears during the lighting period I, and in which every 
trace of rhj'^thmical oscillation fails.^ It is to be noted that the illumination 
employed here, where the slit of the collimator is used, is derived from the 
central part of the crater, but we cannot altogether exclude the possibility 
that the arc lamp has burnt irregularly during the recording of fig. 22. In 
that case the oscillations of potential difference in the eye during illumina- 
tion would have their natural origin in the rhythmically varying intensity 
of the light stimulation. But we did not observe an irregular condition 
of the lamp during the experiment, and must not, therefore, overlook the 
other possibility that it has been burning quietly. The oscillations in 
potential difference of fig. 22 would then have their origin in the rhythmical 
reaction of the eye itself. 

Their failure in fig. 15 may, if the latter explanation is adopted, be 




Fig. 22. — Rhythmical reaction to a continuous stimulus. Li^rht eye. Absc. 1 mm. =0"2 sec. 
Ordin. 1 mm. = 15 microvolts. White light. Intensity of illumination = 02 Iw. I, light; 
d, darkness. 

attributed to the rest of a quarter of an hour in darkness which has been 
given to the eye, during which it may be supposed to have recovered. 



5. The Latent Period. 

As we have mentioned when describing our method of investigation, 
the screen which intercepts the light rays is automatically moved by a 
strong spring set in action by means of an electromagnet. The movement 
of the hand required in closing the circuit of the electromagnet and the 
movement of the armature of the magnet itself give rise to slight jerks 
whereby the signal line and the string itself are sometimes set slightly in 
movement. One or two tenths of a second later, at the exact moment at 
which the light enters the eye or at which it is intercepted, a circuit is 
made or broken, by means of which .the signalling instrument receives its 
current. This instrument itself has, as we have "already mentioned, a 

* If we study the curve with a magnifying glass we observe very fine rhythmical 
oscillations of about 25 or 30 periods per second. These are caused by the technical 
deficiency of the recording apparatus, and have no bearing upon changes in potential 
diflference which might occur in the eye. 



The Electrical Response of the Eye to Stimulation by Light 407 

latency of only 00001 sec, so we may assume that the light stimu- 
lation begins or ceases simultaneously with the interruption of the 
signal line. 

The breadth of the image of the slit in front of the sensitive plate in 
our photographic apparatus is 005 mm., and as the majority of our 
photogi-aphs were made on a sensitive plate moving at the rate of 5 mm, 
per second, time differences of 00 1 sec. may be observed. 

So it is just possible to deduce from figs. 6, 7, 15, 19, 20, 21, and 22 that 
the latent period of the negative deflection A, as it occurs on very strong 
illumination, may be as short as 0'02 sec. or even 001 sec. 

In order to be able to measure this duration exactly, we would require 
to give greater velocity to our sensitive plate. In the meantime we must 
content ourselves with the round figures given above. 

The duration of the latent periods of the different deflections of the 
photo-electric reaction is in a high degree dependent upon the intensity of 
the stimulation.^ Thus we see that the latent period of the negative 
deflection A in fig. 5, where the light stimulation amounts to 10~* Ig, is 
increased to about 1 sec. 

In other photographs not reproduced here we have even been able to 
measure a latent period of this deflection amounting to 0-14 sec, but much 
larger periods are difficult to obtain, since, on employing weaker light, the 
negative deflection diminishes, and soon fails entirely. 

Our observations, which are in agreement with those of Brticke and 
Garten, extend over a wider range. These investigators have made exact 
measurements, but their intensity of stimulation has not been greatly varied. 
They find, as we have mentioned, for tjie shortest latent period of the 
negative deflection A 0'078 sec. and for the longest 0*099 sec. 

Fuchs ^ gives much shorter latent periods than we do, but the method 
employed by Fuchs is, as shown by Gotch,^ open to serious criticism. 

The latent period of the darkening deflection A, shows greater variations 
than that of the negative deflection A. With very strong illuminations the 
latent period of A^ is also very short, 004 sec. and less, even diminishing 
to 001 sec. With weaker illuminations it becomes longer, and since we 
can still obtain a summit A^ with very weak illuminations, there occur also 
very long latent periods of this summit. 

In fig. 5 the latent period of A^ is about 02 sec, with the intensity of 
illumination of 10"* Ig, In fig. 14 the latent period of A^ has the value 
of 08 sec, with an intensity of illumination of 10"^ Ig; while in fig. 
24, where an intensity of illumination of 10~* Ig is employed, the latent 
period attains the enormous value of 22 sec. 

The latent period of the summit B, the reaction of the second substance, 
cannot be measured when illuminating with strong light, since the beginning 
of the wave is here masked by the reaction of the first substance. Where 

1 This relation has already been mentioned by Briicke and Garten, loc. cit. 
* Loc. cit. ^ Loc. cit. 



408 Einthoven and Jolly 

a negative deflection A is present, the commencement of B cannot be 
pointed out. 

On the other hand, when employing weaker light, where the deflection 
A fails the latent period of B can easily be determined. We have found, for 



.;-::-K;"i-"""""""--ljlliini[f[!ll|; 




fi 


IIIIJI 


It- 1 tt u\[----\ 


mMMIIJllJBMfmlllwil^^^ 


\ 


1 
> 1 — 


:: :: ::^-^ 


If 



Fio. 23. — A strong reaction to a weak stimulua. Dark eye. Absc. 1 mm. =0'2 sec. 
Ordin. 1 mm, =2 microvolts. Green light. Intensity of illumination = 10-' Ig. 
I, light ; d, darkness. 

instance, with a light intensity of 10 ~^ I g, in a photograph not reproduced 
here, a value of 0*24 sec. 

With weaker illumination this amount increases considerably. With. 
10-7 i^ (figg 11 12, 13, and 14) it is on an average 06 sec. In fig. 23, 
with 10"* Ig, it is 08 sec. ; and in fig. 24, where the lighting is also 10"^ 
Ig, it even reaches 21 sec. 

Brticke and Garten^ give for the analogous values amounts which lie 




Fig. 24. — Along latent period. Dark eye. Absc. 1 mm. =0*2 sec. Ordin. 1 mm. 
= 2 microvolts. Green light. Intensity of illumination = 10~^ Ig. I, light; 
d, darkness ; p, slight jerk caused by movement of hand in closing circuit of 
electromagnet and of armature of magnet. 

between 0*108 sec. and 0244 sec. Gotch ^ mentions a minimum of 
016 sec, and a maximum of 030 sec When we consider that the dura- 
tion of the latent periods is dependent in so high a degree upon the intensity 
of the light stimulation, it may not seem too rash to suppose, in accord- 
ance with the data supplied by these investigators, that th'e intensities of 
the light stimuli employed by them ranged between 10"^ Ig and 10'^ Ig. 

' Loc. cit., pp. 312 and 315. See also Piper, loc. cit. '.Loc. cit. 



The Electrical Response of the Eye to Stimulation by Light 409 

The latent period of the wave C, the reaction of the third suKstance, is 
difficult to measure, since the commencement of the wave is always masked 
by the reaction of one or both of the other substances. 

Mention must be made of the measurements of Waller, who, in opposition 
to Gotch and Brucke and Garten, observed latent periods which are as 
great or even greater than ours. Speaking of latent periods of 3, 5, and 
even 7 seconds, Waller says: ^ " Such an interval is altogether in excess of 
any possible physiological lost time, and highly suggestive of a period of 
hesitation during which two opposed currents were developed from thf' 
retina at nearly equal rates." 

As Waller made use of a slowly acting Thomson galvanometer, he was 
entirely right in supposing that opposite forces were in question which 
compensated one another at first, while later one of them obtained the 
mastery. The forces assumed by Waller are realised in our first and 
second substances. 

Waller's explanation, however, excellent though it may be when a 
slow galvanometer is used, does not hold good for the curves recorded by us, 
for, where these show great latent periods, the actions of the various sub- 
stances are practically completely isolated, while our measuring instrument 
reacts almost instantaneously. 

From figs. 11, 12, 13, 14, 18, 23, and 24, we are, therefore, compelled to 
conclude that, in fact, on weak stimulation a considerable latency occurs in 
the action of those parts of the retina which give rise to the development 
of electricity. 

The latency found in the photo-electric reaction of the frog's eye is in 
complete agreement with the latency of light-perception in the human eye. 
As evidence for this we may quote the work of two astronomers which has 
not, so far as we know, been referred to up to the present in physiological 
literature. 

Van de Sande Bakhuyzen^ uses as lightiiig point a small opening in 
a copper plate, placed behind the flame of a petroleum lamp. Behind this 
opening there is a metal screen which can be quickly withdrawn by means 
of a strong spring. At the moment at which the margin of the screen 
passes the light point, a metal projection from the screen dips in mercury. 
In this way an electrical circuit is closed, by the aid of which a mark is 
written upon the moving strip of paper of a recording apparatus. 

The observer at a distance of 25 metres from the point of light Wews the 
appearing of the light through a telescope. At the moment at which he 
perceives it he closes a second electrical circuit whereby another mark is 
recorded on the same strip of paper. The distance between the two marks 
measures the reaction time of the observer.^ 

In order to diminish the brightness of the light point, coloured glass is 

I Loc cit., p. 143. 

* Arch N^erlandai-ses dea pc. exactes et naturelles, ser. 2, L 6, p. 727. 
For a critical review of the literature relating to measurements of reaction time see 
Nyman, Skand. Arch. f. Phy.siol., Bd. 19, S. 365, 1907. 



410 Eanthoven and Jolly 

used, and also a pair of Nicol's prisms which can be rotated upon each 
other. The brightnesses used are expressed in stars of known magnitude, 
as these are seen by a telescope of definite dimensions. We reproduce a 
table wliich gives for different star magnitudes the reaction times of the 
observer H B, in thousandths of a second. 

Observer H.B. 

m. m. n». m. m. m. m. m. m. m. m. IP. 

Star-magnitude 3-7 4-5 52 59 60 6-6 74 8-2 8-3 90 9-4 97 
Reaction time 253 253 271 285 276 288 332 375 351 450 514 605 

From these figures it is seen that for the observer H.B. the reaction time 
in the case of weak light of the star magnitude 9*7 m. is considerably greater 
than that of strong light of the star magnitude 3 7 m. The difference 
amounts to 0*605 -0253 = 0352 sec. 

If w6 assume that the light perception c)n stimulation by strong light 
takes place instantaneously, the latent period for the perception of the weak 
light here used amounts to 0352 sec. As however the perception of strong 
light must also take some time, we must conclude that the latent period for 
the perception of the weak light here used is greater than 0352 sec. 

Bakhuyzen also mentions the results which have been obtained by 
some other observers in an analogous wuy. We would draw attention in 
particular to E.B. II. where, for a star magnitude 50 m., a reaction time of 
0*279 sec. was found, and for a star magnitude 98 m. a reaction time of 
0'815 sec., and who therefore shows a latent period for the perception of 
weak light which is greater than 0-815 — 0*279 = 0536 sec. 

Particular interest attaches to the observations of Pihl,^ who has 
calculated the latency of light perception from direct observations of stars. 
For the methods employed by him we refer to his detailed communication, 
and need only mention here that Pi hi finds, for the latent period of the 
perception of weak light, amounts which exceed a full second. 

Considering what we have quoted it need not surprise us that we have 
been able to determine latent periods for the photo-electric reaction of the 
isolated frog's eye amounting to 2 sec. 

IV. The Energy of the Stimulation in Absolute Measurement. 

It may not be devoid of interest to know the amount of energy of the 
radiation employed by us in absolute measurement. Our original intention 
to determine this amount by direct bolometric observations required to be 
given up for several reasons, but nevertheless we can form a fairly good 
idea of it if we attempt to calculate it in accordance with the known laws 
of radiatioft. 

We can identify the crater of an arc light as regards its radiation 

* The stellar cluster 7 Persei micrometrically surveyed, Christiania, 1891. 



The Electrical Response of the Eye to Stimulation by Light 411 

practically with an absolutely black body. If the temperature of the body 
is known, the radiation energy of any part of the spectrum may be 
calculated with the aid of Wien-Planck's^ formula. If the energy of 
the rays whose wave-lengths lie between X and d\ is expressed hy HdX, the 
factor H according to that formula is 

In the circumstances of our investigation the error we make is 
negligible if we employ, instead of Wien-Planck'^s formula, the original 
formula of Wien, 

_« 
H = CX-5e~*^ . . . . (9) 

Here X signifies the wave-length expressed in centimetres, e the base of 
the natural logarithms, T the absolute temperature, c a constant = 146, 
and C another constant, which for the radiation from an area of 1 cm.^ has 
the value of 0-896 x IQ-^^ g gal. cm.Vsec. 

According to Lummer and Pringsheim^ the temperature of the 
crater must lie between 4200° and 3750° abs., while Waidner and 
Burgess,^ after a detailed critical and experimental investigation, think it 
most probable that the temperature of the hottest part of the positive 
carbon lies between 3900° ana 4000° abs. It is permissible, therefore, for 
us to assume that the temperature is 4000° abs. 

Further, H denotes the entire radiation which is emitted by a flat area 
of 1 cm.2 This radiation would be entirely received by an imaginary lens 
with an angle of aperture of 180°. 

We denote by Z cm.^ the area of the slit used in our experiments ; that 
is to say, the magnitude of the radiating area, while 0^ is the angle of 
aperture of our collimator lens. 

Further, we must take into account that all the light rays falling upon 
the collimator lens do not enter the frog's eye. A part of the rays, as we 
have already mentioned, is lost by reflection from the refractive surfaces 
and absorption in the refractive media of the spectroscopic apparatus. 

If we denote by - that part of the light which passes, then we find for 

Hj — the radiation actually entering the pupil — the formula 

H,=H— sin2i0i .... (10) 
P 

In our experiments Z = 0-0209 cm.^, sin i0i = 0118, while the value of 

p is taken as 2. 

> The formulae here used may be found in Kohlrausch's text-book, Lehrbuch der 
praktischen Physik, 10 Aufl., 1905. 

» Verhandl. d. Deutsch. Physikal. Gesellsch., i., p. £3. 1899 ; ibid., p. SIV 
3 Bulletin of the Bureau of Standards, Wa.shington, vol. i. p. 123. 1904. 



412 Einthoven and Jolly 

Had our absolute radiations been exactly determined by the bolometer, 
it would have been worth our while also to measure f accurately. Since, 
however, we have based our calculations of the absolute radiation on the 
ten^rature of the crater, we may content ourselves with an estimation 
of p. Further, this estimation will suffice, because jp is very small in com- 
parison with the enormous ratios of the radiating energies employed by us, 
which have ranged from 1 to more than lO^'*. 

Our estimation of p is based on a measurement which has recently been 
made by won Kries,^ who gives the analogous loss in his spectroscopic 

arrangement as =1'825. 

According to the above data the amount of the radiation entering the 
eye is calculated to be as follows : — 

Hi(A = 0-460) = 229 
Hia = 0-497) = 277 
Hi(A. = 0-590) = 379 
Hi(A = 0-670) = 418 

These results are represented -diagrummatically in fig. 25. Here the 
wave-lengths of the normal spectrum are plotted in. a system of rectangular 
co-ordinates as abscissae, and the values of H, as ordinatea. 

The radiation h of our three parts of the spectrum expressed in g. cal. 
per sec. is represented by the area of the three parts of the diagram : — 

For blue, from X = 0460 to X = 0-497 . h^ = 9-36 X 10"* g. cal. per sec. 
„ green, „ X = 0497 to X = 0-590 . hg = 30-5xl0-* 
„ red, „ X = 0-590 to X = 0-670 . hr = 31-9xl0-* 

To obtain a maximum of radiation in the eye the largest diaphragm is 
placed at Dg, viz. the diaphragm for green of 9*5 x 275 mm.* The slit of 
the collimator and the prism are removed from the spectroscopic apparatus 
and the uncoloured image of the crater thrown ditectly on the diaphragm 
Dj. The crater image, which is about 20 mm. high and 32 mm. long, has 
an elliptic form and overlaps all margins of the diaphragm. 

In order to calculate the radiation energy so obtained, we have to 
determine the part of the crater to whioh the rectangular diaphragm at 
Dg corresponds. If this is calculated from the image magnification, a 
rectangle of 1*9 mm. high and 55 mm. long is found as the piece cut out 
from the crater. It is by this rectangle, when we employ our maximum 
amount of white light, that the slit of the collimator is replaced. Its area 
is exactly five times larger than the area of the slit. 

It is useless to calculate the radiation energy of the crater light for all 
wave-lengths together. First we must bear in mind that a considerable 
part of the ultra-red and ultra-violet rays is absorbed by the glass lenses 

• Loc. dt, p. 391. 



The Electrical Response of the Eye to Stimulation by Light 413 

of the spectroscopic apparatus and the refracting media of the eye ; and 
secondly, we must remember that these rays, although they reach the 
retina, do not practically exert here a photo-electric action. 

We therefore may limit our calculation to rays of wave-lengths between 
X = 0"460 and X = 0'670. The entire active radiating energy is then easily 
calculated to be 5 times the area of diagram (fig. 25), making an amount 
of h, = 00359 g. cal. per sec. 

The photographs in our plate where the weakest radiations are employed 
are figs. 23 and 24. The strongest radiations have been employed in the 
case of figs. 6 and 7. The ratio between the weakest and the strongest 
-adiation is 10-» hg : h, ; that is, 1 : 1'18 X lO^o. 




aj,6« 



0.650 



PAio ojx 0450 qdoo 
Fig. 25. — Diagram in which the wave-lengths of the normal spectrum 
are plotted in a system of rectangular co-ordinates as abscissa, 
and the values of Hj — the radiation entering the pupil — as ordinates. 
The areas of the three parts of the diagram represent the radiation 
of the three parts of the spectrum used for illuminating in g. cal. 
per sec. 



Previous investigators have already repeatedly demonstrated that the 
eye acts as a very sensitive reagent to light. 

Waller^ mentions that radiation by moonlight during not more than 
001 sec. is able to evoke a strong action current, and that a distinct 
galvanometric deflection is caused when the eye is illuminated by a white 
card 10 feet distant from the eye which itself receives illumination from a 
standard candle 10 feet distant. 

With regard to the -sensitiveness of light perception in the human eye, 
we refer to the measurements of Langley, Grijns and Noyons and von 
kries. 

Langley -^ gives as the least energy required to cause a light perception when 
stimulating with green light 

6-7 X 10-" g. cal. = 2-8 x 10-» erg. 

» Loc. cit, p. 124. ' American Jour, of Science, p. 376, 1888. 



414 Einthoven and Jolly 

Grijns and Noyon.s' ^ minimum is 

0-95 to 26 X 10-18 g. cal. = G-4 to 11 x IQ-^o erg. 

Von Kries^ gives 

3-1 to 6-2 X 10-18 g. cal. = 1-3 to 26 x lO-^o erg. 

In fig. 23 we see the development of a fairly strong photo-electric 
reaction to a stimulus which uses 30'5 x 10-^^ g. cal. per sec., which lasts for 
484 sec, and thus possesses a total energy of 1*47 x lO-^^ g. cal. In all 
probability there may be obtained a perceptible galvanometric deflection 
with a stimulus 10 times, and perhaps even 100 times weaker, but our 
experiments have not extended further in that direction. Nevertheless, it 
is sufficiently clear that for the photo-electric reaction of an isolated frog's 
eye, there is required much more light than for the development of a 
light perception in the human eye. 

On the other hand, a comparison of the isolated frog's eye with the most 
sensitive bolometer constructed by human skill is greatly tx) the dis- 
advantage of the latter. 

V. The Relation between the Energy of the Stimulation 
AND the Energy of the Reaction. 

Generally speaking, the rule holds good that for moderate and strong 
stimuli the energy of the reaction increases or decreases much less rapidly 
than t}\e energy of the stimulus. Perhaps it is to be expected that here the 
Weber-Fechner law is true ; that is to say, that when the stimuli increase 
in geometrical progression the reactions increase in arithmetical progression. 
This should be investigated for each of the three substances separately, 
which has not so far been done. 

If very weak radiations are employed, the above-mentioned law does not 
hold good, as is sufficiently evident from the series of four photographs, figs. 
11, 12, 13, and 14. Here the energy of reaction is increasing considerably, 
whereas the increase in the energy of radiation is but small. If the 
radiations are weakened still more, the reactions decrease relatively quickly 
in energy, and it is reasonably to be expected that under these conditions 
all perceptible galvanometric deflections will soon fail. The results of the 
experiments are in complete agreement with this expectation. 

In reference to tliese problems, we may recall the work of de ilaas,^ who 
has confirmed the Weber-Fechner law within wide limits; but de Haas 
used a slowly deflecting galvanometer, so that the separation by him of the 
action of three substances w^as impossible. Presumably, his measurements 
with stronger stimuli relate principally to the action of the third substance, 
while those with weaker stimuli may have had reference to the combined 
action of the second and third substances. 

^ Engelmann's Arch. l. Physiol., p. 25, 1905. ^ Lqc (.jt. 3 Lq^. cit. 



The Electrical Response of the Eye to Stimulation by Light 415 

In conclusion, we may institute a comparison between the absolute 
amount of the energy of the stimulus and that of the response. 

We choose as an example fig. 23, where we have already determined the 
energy of stimulus as amounting to 147 x 10^^^ g. cal. The energy of 
reaction must be calculated from the form of the curve. If the galvano- 
meter is replaced by a wire of a small negligible resistance the current 

y 
passing through it is A = :^ amp. Here V is the electromotive force in 

volts developed at each moment, while R is the resistance in ohms of the 
preparation. For the present we assume that V remains unchanged, when 
the resistance of the galvanometer is diminished. The energy of the 
reaction during the time dt is expressed by YAdt, the total energy of the 

reaction by W = / YAdt or W = j dt Joule. 

In the figure 1 mm. abscissa = 0-2 sec, 1 mm. ordinate = 2 microvolts. 
The resistance of the preparation is R = 9000 ohms. 

With the aid of the above data, the amount of W, as somewhat roughly 
calculated from the form of the curve, is W = 2 x 10" '^ Joule = 48 x 1 0""" g.cal., 
and it is thus evident that in the case of fig. 23 the energy of the reaction 
is more than 30 times less than the energy of the stimulus. 

Following on this result, there are good grounds for stating as a general 
rule that the absolute energy of the photo-electric reaction is always less 
than that of the light stimulus. It is true that we have to consider the 
possibility that in a curve which is recorded under other conditions, the 
energy ratio might be altered in favour of the photo-electric reaction, but 
in our collection of photographs, taken in very varying circumstances, we 
have not found an example of this. 

The curve (fig. 23) has been chosen jilst because its energy ratio is 
specially favourable to the photo-electric reaction. 

In judging of the energy of the photo-electric reaction, we have to take 
into account that there exists short circuiting in the eye itself, and that the 
current measured by the galvanometer is presumably only a small part of 
the current passing through the eye. 

This last current is not easily determined, so we shall take, as is usually 
done, the potential difference or the current, as these are measured by an 
instrument outside the eye, to be the real photo-electric reaction. In our 
above calculations we have assumed that the electromotive force developed 
by the eye remains unchanged when the resistance of the galvanometer is 
diminished, and we have assumed the amount of this lesistance to be equal 
to zero. Both assumptions were made for the purpose of calciilating the 
possible maximum of the reaction energy for this special case. If we take 
the resistance of the galvanometer into account, we find for the reaction 
energy an amount which is 175 times less, the resistance of the galvanometer 
being 6800 ohms. 

The photo-electric reaction of the eye is, in regard to the energy ratios, 



416 The Electrical Response of the Eye to Stimulation by Light 

comparable with the electric reaction of a nerve or a muscle, not with the 
mechanical reaction of the latter, for the muscle in contracting is able to 
develop a quantity of energy whrch far exceeds that of the stimulus. 

VI, Summary of Conclusions. 

1. The photo-electric reaction of an isolated eye is specially adapted to 
the study of the effect of stimuli of very different intensities. In our ex- 
periments light stimuli have been used whose energy varies from 305 x 10~'^ 
to 3-95 X 10~^ g. cal. per sec. 

Short illumination with the strongest light has not been found to damage 
the eye, while the weakest light was capable of producing a fairly consider- 
able photo-electric reaction. 

2. The form under which the photo-electric reaction manifests itself 
under different conditions gives ground for the supposition that there occur 
in the eye three separate processes, and each of these may be dependent 
upon a separate substance. We speak of three substances for the sake of 
convenience. 

3. The first substance reacts more rapidly than the other two. On 
lighting it develops a negati\'e, on darkening a positive potential difference. 
Its action comes strongly into prominence in a light eye and appears almost 
unmixed on sudden darkening of short duration (a flash of darkness). 

The second substance reacts less rapidly than the first, and in an opposite 
sense. On lighting it develops a positive, on darkening a negative 
potential difference. Its action appears almost unmixed in a dark eye 
which is illuminated for a short time with weak light. 

The third substance reacts in the same sense as the second, but much 
more slowly. Its action fails in a completely light eye, and also in a dark 
eye which is illuminated very weakly for a short time. 

4. The latent period of the photo-electric reaction is in a high degree 
dependent upon the intensity of the stimulus. With strong stimuli it is of 
the order of 001 sec, while with very weak stimuli it may be lengthened 
to more than 2 sec. These values are in agreement with the latent periods 
of light perception in the human eye. 

5. For each of the three substances the rule holds good that with 
moderate and strong light the energy of the stimulus increases much more 
quickly than the energy of the reaction. 

6. Although the eye is far more sensitive than the most sensitive 
artificial bolometer, the energy of the reaction remains below that of the 
stimulus even in the most favourable circumstances. The photo-electric 
reaction of the eye is in this respect comparable with the action current of 
a muscle or nerve. 



O 



"■'*^ /U 1364 



QP 


Quarterly Journal of 


1 


experimental physiology 


Q3 


and cognate medical 


V. 1 


sciences 


Biological 




& Medical 




Serials 





PLEASE DO NOT REMOVE 
CARDS OR SLIPS FROM THIS POCKET 

UNIVERSITY OF TORONTO LIBRARY