“hehe. a
twee

ino

A dtr oe
Serres < fey
wage moe Nee oe
. tor atew owes
ey ot -
Se

ate ae
Se eating
Si pareseem

eon ee
Cat oayrs rata

ee
as

Fre ae et SF

oe eT

sa et tt

1 ee vw ae

eS ee eo ee oe
: eS a
7
hr

ee _—— ee

ca |

,
"aye
i mi

WN) Fy 8
lie
ear ”
Weal

my 5% ‘
| ots

rhe gf

7

THE

AMERICAN JOURNAL

OF

PHYSIOLOGY

VOLUME XXXVIII
|
sil
all”

BALTIMORE, MD.

ve

CONTENTS

No. 1. Juny.1, 1915

Tue Respiratory Drap Space. By Yandell Henderson, F. P. Chilling-
PDT A GTS na UP OUULTBO INE & (25. t ce aaa «choke Bthel so 4% ew Nateeieate Med obs
THE VARIATIONS IN THE ErrectTive Drap Space In BrwatHiIne. By J. 8S.

CONTRIBUTIONS TO THE PHYSIOLOGY OF THE STOMACH. XXIV. THE Tonus
AND HUNGER CONTRACTIONS OF THE STOMACH OF THE NEW Born. By
EAE RAT: /SO7h ONL His GON SOUT tia FEM Ass. Se Se ahe Use skid 2 weiss Aajesses eS
Factors AFFECTING THE CoAGuLATION Timp oF Bioop. VII. Tue INFiv-
ENCE OF CERTAIN ANESTHETICS. By Walter L. Mendenhall............
A CALORIMETRIC CALIBRATION OF THE KrocH BicycLp ERGoMETER. By F.
Gen LACAN TE GAGE I eM OM OG TICES Ae SRA Bee a ERIS OSU SE ae nanbe Smite
Tue Errect or ADRENALIN ON THE Heart-Rate. By Walter J. Meek and
EE STOEL A iy tt RE. 8 Set eye Cases a AIA, Dra eae SES oa nb weak erapends
Tue INFLUENCE OF THE OIL OF CHENOPODIUM ON THE CIRCULATION AND
Respiration. By William Salant and A. E. Livingston...............
INFLUENCE OF DIIODOTYROSINE AND JODOTHYRINE ON THE SECRETION OF
CEREBROSPINAL Fiuip. By Charles H. Frazier and Max Minor Peet...
THE RESPONSE or THE VASODILATOR MECHANISM TO WEAK, INTERMEDIATE,
AND STRONG SENSORY StimuuaTion. By E. G. Martin and W. L.
CTECLENIAEL Us ee MTA MRR ne a loeb Os coin ley BiIOIA 2 cesamd oa inlaraie, tes
SPINAL ANAESTHESIA IN THE Cat. By G. G. Smith and W. T. Porter......
Tur ConpucTION WITHIN THE SPINAL CoRD OF THE AFFERENT IMPULSES
PrRopDUCING PAIN AND THE VASOMOTOR REFLEXES. By S.W. Ransom and
LS URES ER Se SS igs OEE ee aan ee nr ere ae
RHYTHMICAL CONTRACTION OF THE SKELETAL MuscLe TisSuE OBSERVED IN
fissum Cortores. By Margaret Reed Lewis. .....2.5.026.s00ca0 needs

No. 2. Auveust 1, 1915

THE PRESERVATION OF THE LIFE OF THE FROG’S EGG AND THE INITIATION OF
DEVELOPMENT BY INCREASE IN PERMEABILITY. By J. F. McClendon...
Tue AcTIon or ANESTHETICS IN PREVENTING INCREASE OF CELL PERME-
AUVELUEY aeMEN 1) vel coe ACO LETE LOIRE ©. 4/1 4. 053. Sy x:8tsl tin Vee Bh oot Mee hase eA o>
New HyproGEN ELECTRODES AND Rapip METHODS OF DETERMINING HypDRo-
GEN Ion ConcentrRATions. By J. F. McClendon.....................
A Direct READING POTENTIOMETER FOR MrAsuRING HyproGen Ion Con-
ERMA TIONS: seid ol. MA GOLENDON 0.08 6 eiess Ve cdee Weesascte dee se iees
Acipiry CURVES IN THE STOMACHS AND DUODENUMS OF ADULTS AND INFANTS,
PLOTTED WITH THE AiD OF IMPROVED MrtTHopDs oF MEASURING HYDROGEN
low Gancentuavtion.. By J. F. MéClendoti. sit cicscscncceccevecsacs

ili

PAGE

29

153

NG CONTENTS

. PAGE
Tue PERFUSION OF THE MAMMALIAN MEDULLA: THE EFFECT OF CALCIUM AND

oF PoTASSIUM ON THE RESPIRATORY AND Carpiac Centers. By D. R.

PLoOW Er. : «oes. Ree edie 6 een 5 Ee ay Ee EE ss ee 200
An INTERPRETATION OF THE MEMBRANE MANOMETER CURVES AS AFFECTED
BY VARIATIONS IN Buoop Pressure. By J. D. Pilcher............-... 209

SruDIES ON THE HypROoGEN ION CONCENTRATION IN BLOOD UNDER VARIOUS
ABNORMAL ConpiTions. By M. L. Menten, M.D., and G. W. Crile, M.D. 225

THE ORIGIN OF ANTITHROMBIN. By George P. Denny, M.D. and George R.
Minot MDs. oe spine oti leas» 2 oe et a OA eA oy oe ae 233

CoNTRIBUTIONS TO THE PHYSIOLOGY OF THE STOMACH. XXV. A Notre on
THE CHEMISTRY OF NoRMAL HuMAN Gastric Juice. By A. J. Carlson,

Assisted in part of the experiments by H. Hager and M. P. Rogers....... 248
Tue ConTENT OF SUGAR IN THE BLoopD oF CATS UNDER THE INFLUENCE OF
Cocaine. By Edward Waldo Emerson Shear............-.-...++++++> 269

CoNTRIBUTIONS TO THE PHYSIOLOGY OF THE StoMacH. XXVI. THE RELA-
TION BETWEEN THE DIGESTION CONTRACTIONS OF THE FILLED, AND THE
HunGER CONTRACTIONS OF THE ‘‘Empty’’ Sromacn. By F. T. Rogers
and LJ oHatdtetd os cx Soe ho ee Ce ee 274

A ConrTRIBUTION TO THE PuysioLoGcy or Lactation. By W. L. Gaines..... 285
No. 3. SEPTEMBER 1, 1915
An ANALYSIS OF CERTAIN PHOTIC REACTIONS, WITH REFERENCE TO THE
WeEBER-FECHNER Law. I. THE REACTIONS OF THE BLOWFLY LARVA TO
OproseD Brams or Ligut. By Bradley M. Patten..................... 313
Tue INFLUENCE OF THE EXTRACT OF THE POSTERIOR LOBE OF THE HYPOPHYSIS
UPON THE SECRETION OF SALIVA. By G.O. Solem and P. A. Lommen..... 339
THE INFLUENCE OF THE VAGUS NERVE ON THE GASEOUS METABOLISM OF THE
Kipnry. By Roy G. Pearce and Edward\P.Carter.2>--.-- - =.» eee 350
CHANGES IN IoDINE CoNTENT OF THE THYROID GLAND FOLLOWING CHANGES
IN THE BLoop Ftow THrRouGH THE GLAND. By C. F. Watts........... 356
PHOTOELECTRIC CURRENTS IN THE EYE OF THE FisH. By Edward C. Day.... 369
THE CoMPARATIVE RATE, AT WHICH FLUORESCENT AND NON-FLUORESCENT
BACTERIA ARE KILLED By ExposuRE TO ULTRA-VIOLET. By W. E.
Burge and A. J. Neill. 00. ae tb. <. oda be eee a 399
Tue Errects oF CHANGE IN AURICULAR TONE AND AMPLITUDE OF AURICULAR
SYSTOLE ON VENTRICULAR Output. By Robert Gessel................-. 404

No. 4. Ocroper 1, 1915

Srup1es IN EXPERIMENTAL GLycosuRIA. IX. THr LEVEL OF THE BLOOD-
Sucar iN THE Doc Unprr Laporatory Conpirions. By J. J. R.
Macleod and RG. Pearce... ...:< occcuste ees oe eee 415

Strupies IN ExPERIMENTAL GLycosuRIA. X. THE SUGAR RETAINING POWER
OF THE LIVER IN RELATIONSHIP TO THE AMOUNT OF GLYCOGEN ALREADY

PRESENT IN THE OrGAN. By'J. J. R. Macleod and R. G. Pearce........ 425
Tue DirreRENTIAL Errects oF ADRENIN ON SPLANCHNIC.AND PERIPHERAL

AnTERIns. By Frank A. Hartman. : os. 80s. iis daues poe eee 438
THe Osmotic PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE IN

RELATION TO THOSE oF Livinc Cetis. By Edward B. Meigs........... 456
BRD IK os seelee o's «vo a's e si0ie eb o.ae eo dy bites cd ood ne maine 491

THE AMERICAN
JOURNAL OF PHYSIOLOGY

VOL. 38 JULY 1, 1915 No. 1

THE RESPIRATORY DEAD. SPACE

YANDELL HENDERSON, F. P. CHILLINGWORTH anp J. L. WHITNEY
From the Physiological Laboratory of the Yale Medical School

Received for publication April 13, 1915.

A correct estimation of the volume of the respiratory dead space is
of fundamental importance in connection with many of the prob-
lems of the regulation of breathing and related topics. The interpreta-
tion of the data accumulated by many investigators depends upon the
decision of the question whether the dead space is a fixed or a variable
quantity. If the latter is the case, what are the nature, cause, and
mode of control of such variations?

Although earlier investigators recognized the dead space as a factor
in breathing,! Loewy,” seems to have been the first to report a measure-
ment of it. By means of a plaster cast of the cavity of the mouth,
pharynx, trachea, bronchi and bronchioles of a cadaver he found a vol-
ume of 144 cc. He noted, however, that in the living subject even in
an expiration much smaller than the volume of the dead space there
are always to be found considerable amounts of COs, and that as expira-
tion progresses the percentage of CO. in successive portions increases
gradually, not abruptly: facts which later investigators have not always
kept in mind.

Basing their opinion upon these observations, Zuntz and his col-
laborators in their extensive investigations have assumed the dead
space to be an unvarying volume—the same no matter whether the sub-
ject were at rest, or breathing deeply from physical exertion, or on a
mountain. All of their calculations of the alveolar ventilation and of
the composition of the alveolar air involve this assumption. The
method employed by them was based on the proportion: the composi-

1

2 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

tion of the alveolar gases (i.e., the excess of COs over, or the deficiency

‘of oxygen under, the inspired air) is to the composition of the (total,
mixed) expired air as the tidal volume is to the tidal volume minus
the dead space. It is noteworthy that by this method the constancy
of composition of the alveolar air was not discovered. As Haldane
and his co-workers have demonstrated that such a constancy normally
exists, the natural inference is that the dead space is of widely varying
volume.

This inference was followed up by Douglas and Haldane* who found
when the subject was at rest a volume of 160 cc., and a progressive
increase with physical exertion up to more than 600 cc. Their method
was based (as all indirect determinations must be) upon the same pro-
portion as that used by Zuntz, Loewy and others, but with this differ-
ence that Douglas and Haldane, instead of assuming the dead space
and calculating the alveolar gases, have determined the composition_
of the alveolar air_by direct analysis, and calculated the dead space.
The other terms in the proportion are the (mean) tidal volume as
determined from the total volume of air expired into a Douglas bag
in a certain number of breaths, and the composition of this expired air.

Even before this work of Douglas and Haldane, Siebeck,* had found
that during the hyperpnoea of physical exercise the dead space is
increased. He used a method involving only a single breath—instead
of the average volume of a series of expirations—in which 500 ce. or
1000 cc. of hydrogen were inspired, and then expired into a gasometer.
The percentage of hydrogen in the alveolar air (from a sample at the
end of expiration), the percentage in the mixed expired air, and the
volume of the expiration afforded three terms in the proportion from
which the fourth, the dead space, was calculated. He concluded that
the dead space is larger with the lungs inflated, and suggested that this
might be due to participation of the bronchi in the respiratory move-
ments. Thus far, as we believe, Siebeck was correct. He concluded
also, however, that the depth of breathing does not influence the dead
space, and that it is even decreased during the deep breathing caused
by inhaling CO,. On these points we think that Biebeck was misled
by not using sufficiently deep inspirations.

The latest considerable contribution to the problem of the dead space
has been made by Krogh and Lindhard.® They have returned emphati-
cally to the doctrine of a dead space of unvarying volume. They first
subjected the method of Siebeck to a decidedly damaging criticism,
and then adopted it themselves and founded quite precise conclusions

a inAd

ao” oe

ary

THE RESPIRATORY DEAD SPACE 3

upon it. They have recently published a number of papers contain-
ing a large amount of data on yarious respiratory problems. In the
interpretation of their data they have used a fixed dead space, and, as
they themselves have remarked, if the dead space is a variable they
will have to recalculate all of these experiments, and reconsider their
arguments. ; :

As we shall show that the dead space is a variable, it is of interest to
note how Krogh and Lindhard were caught in a trap set (quite unin-
tentionally) by Douglas and Haldane. The latter suggested that the
enlarged dead space, which they found during hyperpnoea, serves to
facilitate the flow of air, and is due to an active broncho-dilatation
under the influence of nervous or chemical stimuli. This highly
teleological suggestion has not, so far as we can discover, led Douglas
and Haldane into any error of fact, but it has led Krogh and Lindhard
to confuse the question of an augmented dead space with the quite
distinct problem of an active broncho-dilatation.

Accordingly, Krogh and Lindhard, in testing the volume of the
dead space at_rest and after exercise, seem to have arranged intention-
ally that the subject should take no deeper inspiration under one con-
dition than under the other. The two measurements were equal;
and they concluded therefore (correctly) that there is no evidence of

~ active, teleological, broncho-dilatation during exercise. They did not,

however, make adequate tests of the volume of the dead space in breaths
as deep as occur during hyperpnoea. If they had, they would doubt-
less have found, as we do, a dead space varying up to as great a maxi-_
mum as Douglas and Haldane claim.’

The true explanation of these variations is in the main, we believe, that
during extra deep breathing the bronchi and bronchioles are dilated passively
with the rest of the lungs. Thus a deep inspiration involves an increased
dead space no matter whether the subject is at rest or exercising. We
recognize, of course, that broncho-dilatation and constriction are func-
tions under sympathetic or autonomic control, that they are influenced
by drugs, and are subject to pathological disturbances.’ Our results
demonstrate merely that the ordinary variations of the dead space as
between quiet and deep breathing are not for the most part of this
active character.

Axial flow of gases through the tubes. As several of the recent investi-
gators in this field have evidently left out of account the peculiarities of
the movements of liquids and gases in tubes, we may begin the report
of our data with a few simple experiments on this topic. They illus-

4 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

trate why it is that with a dead space of 150 cc. the first 150 cc. expired
(or even the first 50 cc.) are not free from a considerable admixture of
pulmonary air. They afford the reason why a dog during heat poly-
pnoea may have a tidal volume considerably less than the volume of
the dead space. They suggest that, during rapid shallow breathing
_in man, what we may call the physiological dead space is a much
smaller volume than the anatomical dead space of Loewy’s plaster
cast. There may easily be a gaseous exchange sufficient to support.
life even when the tidal volume is considerably less than the dead space.

If one takes a glass tube of, say, a meter length and one or two centi-
meters bore, and blows tobacco smoke into one end, he sees that the
smoke does not move along the tube in a cylindrical column, filling the
tube from side to side, but in the form of a very thin spike. If the tube
is held vertically, so that gravitational effects are avoided, the spike
follows the axis of the tube, and the tip of the spike begins to issue from
the upper end before the tube as a whole is more than a quarter or a
third filled with smoke. The quicker the puff, the thinner and sharper
the spike, and the more smoke must be blown through before all the
clear air is washed out.

If, at the moment that the tip of the spike reaches the upper end
of the tube, the puff is stopped by applying the tongue to the lower
end, the spike breaks instantly everywhere; and the tube is seen to be
filled from side to side with a mixture of smoke and air, thin at the
upper end, a thicker mixture at the lower, and all gradations between.

If now an inspiration is made, a thin spike of clear air projects itself
down along the axis and into the mouth of the operator along with some
of the thicker mixture in the lower end of the tube.

An even more striking demonstration is obtained with a glass bulb
having inlet and outlet tubes on opposite sides. When smoke is blown
in through one of these tubes, the column at first shoots across the cen-
tre of the bulb and out through the other tube with little contamina-
tion of the clear air surrounding the stream. If it is stopped suddenly,
a complete mixing of smoke and clear air occurs almost instantaneously,
and thereafter a very large volume of air must be drawn through the
bulb before the last trace of smoke is washed out.

For instrumental purposes these frictional effects are easily neutralized
by filling the large glass tube with some loose fibrous material such
as glass wool, jute, disks of wire gauze, ete. The friction being then
the same at every point in the cross section, the column of smoke pushes
the air ahead of it like the plunger of a pump. This holds true even

THE RESPIRATORY DEAD SPACE 5)

when the fibrous material is so loosely packed as to afford no noticeable
resistance to an expiration.

We have employed a tube of this sort to determine the composition
of successive fractions of an expiration. At intervals along the tube
were inserted side tubes to which were connected small pipettes of a
type described in a previous paper from this laboratory.* The appara-
tus is shown in figure 3, and examples of the results obtained in figure
4. From them it is clear that in a quick expiration through the mouth
the “spike”’ of alveolar air begins to issue between the lips with the first
25 to 50 cc. of air expelled. They show also that for an approximately
complete washing out of the dead spate, so that the last portion shall
consist of practically undiluted alveolar air, an expiration (especially

Fi 9. / Fig. 2

Fig. 1. (a) Shows a “‘spike’’ of smoke moving through a glass tube. (6)
Shows the condition when the current is suddenly stopped and mixing instan-
taneously occurs. (c) Shows clear air drawn in.

Fig. 2. Shows how a column of smoke crosses a bulb with little mixing or
sweeping out of the air within it.

if rapid) of at least 400 cc. or three times the volume of the dead space
is required. This applies to ordinary quiet breathing. During hyper-
pnoea the larger dead space necessitates a much greater expiration be-
fore an undiluted sample of alveolar air can be obtained.
We have used and compared five methods with minor variations for
determining the dead space:
First, or CO: method. The subject began with the deepest possible
expiration with the mouth open, and followed this with an inspiration
- sufficient in volume to fill the dead space with entirely fresh air. In
some cases this breath was held for seven to ten seconds; in others the
next act followed immediately. This consisted in the subject making
the deepest possible expiration through a rubber tube into a gasometer

6 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

or rubber bag. The subject then kept his tongue against the end
of the tube until it was closed with a spring clip. A sample of alveolar
air was drawn from the tube as in the method of Haldane and Priestley,
and its CO, percentage (a,;) was determined with a Haldane apparatus.
(The little apparatus graduated up to 10 per cent for CO, only is most
convenient for this purpose.) A sample of the mixed air (mm) in the
gasometer was also analyzed. The volume of the (second) expiration

Fig. 3. Apparatus for analyzing successive portions of an expiration. The
volume expired is measured by means of the gasometer. The tube (volume
400 ce. and length 100 em.) is filled with loosely shredded glass wool or jute. The
pipettes hanging from it are filled with ,~,; normal baryta. Immediately after
an expiration the clip at the end of the tube is closed. By opening the clips on
the pipettes, exactly one-half (15 ec.) of the baryta is then allowed to run out
of each, thus drawing in a sample of air from the tube. The clips are reclosed,
and the remaining half of the baryta is thoroughly shaken with this air, and is
then drained into a small flask, and tightly stoppered. After complete sedi-
mentation of the BaCO;, a sample (5 cc.) of the supernatant baryta is titrated
with ;y normal acid; and, with corrections for the barometric pressure and tem-
perature when the air sample was taken, its CO, percentage is calculated.

(£) was read from the gasometer, or by means of a gas meter con-
nected with the rubber bag. The dead space (d) of the apparatus,
i.e., the volume of the tube, was determined once for all. It should
be as small as convenient. By using a readily collapsible tube, or by
filling the tube with water before each experiment, it can be eliminated ~
entirely. Although more exact these arrangements are also more
trouble and not necessary.

THE RESPIRATORY DEAD SPACE 7

Per cent COz

7i9. 4

Fig. 4. Showing the percentages of CO, in successive portions of quick
expirations in four experiments with the apparatus of figure 3.

The (virtual) dead space (D,) for COs. was calculated from these
data by the formula:
my

D, = E ray E d
ay
It is not necessary that the alveolar samples should have a normal
CO, content, as it is not the absolute amount but the relation of this
factor to the CO, in the mixed air which counts. For instance, in two

experiments (on Y. H.) the data and results were:

E Ai Mi d Di
3810 3.75 2.90 195 670
3850 6.20 4.70 195 665

The characteristics of the results obtainable with this method are
shown by figure 5.

These data and their significance may be summarized as follows:
(1) The dead space as determined by this method (without a pause)
varies with the extent to which the lungs are inflated. Below the in-
flation of quiet breathing it may be as small as 60 ce. or less. At the
normal level volumes of 150 or 160 cc. are obtained in close agreement
with other observers. When the lungs are inflated to an extent cor-
responding to that of hyperpnoea, the volumes are much larger (up
to 600 or 700 cc. or more). But they are the same no matter whether
the subject is at rest or exercising. This indicates that the increased
dead space observed by Douglas and Haldane under the latter condition
is chiefly a mere mechanical stretching, not an active broncho-dilatation.

(2) When inspiration is as far as possible of a costal character, the

8 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

values obtained even by fairly great inflation (3100 to 3800 cc.) are only
about half as large as when it is of the ordinary diaphragmatic charac- *
ter. This indicates that the expansion of the dead space between or-
dinary and deep breathing depends to a great extent upon the lengthen-
ing of the bronchi and bronchioles with the downward movement of
the diaphragm.

CC Virtval dead

tr
space for C02

Extent of Chestinflation
as sd ie by pt hn

final Geepest expiral Ton

Fig. 5

Fig. 5. A diagram of results obtained by our first method. Ordinates ex-
press the (virtual) dead space, and abscissae the extent to which the lungs were
expanded above the residuum at deepest expiration. In the subject (Y. H.)
during quiet breathing with a tidal volume of 500 cc. inspiration reaches a level
of about 1600 cc. as indicated by the arrow at the left, and an active hyperpnoea
raises the level to between 2900 and 3630 as indicated by the arrows at the right.

The values for the dead space when the final expiration was made immediately
after the inspiration are indicated by the little circles (0)., and upper line.

The values obtained when the inspiration was held for seven to ten seconds
are expressed by the little squares (0) and the middle line.

The influence of making the inspiration and pause so far as possible with the
costal mechanism without contraction of the diaphragm is shown by the incom-
plete squares ‘C) and the lower line.

All of the foregoing observations were made while the subject was at rest,
sitting on a stool in a room of comfortable temperature. The solid circles and
squares (@ @) on the contrary express observations by identical procedures,
but taken while the subject was hyperpnoeic from vigorous exercise on a station-
ary bicycle. As they correspond closely with the observations during rest, they
show that the increased dead space during exercise is not due to active broncho-
dilatation, but to passive distension incident to deep inspirations.

ae

THE RESPIRATORY DEAD SPACE 9

(3) When the inspired air is held for seven to ten seconds before the
final expiration is made, the values obtained for the virtual dead space
are at all levels of chest inflation considerably smaller (less by a third
to one-half) than when no pause is made. This is probably due to dif-
fusion of CO, in appreciable amounts through the mucosa of the mouth,
trachea, bronchi, etc. (A study of the diffusion of CO. and oxygen
in the mouth is now under way in this laboratory, has afforded con-
firmation of this view, and will be reported in a later paper.)

Second, or oxygen method. The procedures were the same as in our
first method except that the gas samples were analyzed for oxygen and
the percentages found (m,. and a2) were subtracted from the oxygen
percentage of the inspired air. Thus the dead space for oxygen (D2)
was found by the formula |

_ 720-93 — ms cv
20.93 — ay

It is noteworthy that when both this and the CO, method were car-
ried through on the same breath the (virtual) dead space for COz was al-
ways smaller than that for oxygen. The difference between them was
greater (owing to the fact that D,; was reduced much more than D,)
when the breath was held for several seconds than was the case when
no pause was made. For example we found (on J. L. W.):

DT

“

E Di Dz Di: Dz
With pause......... 720 33 78 42:100
With pause......... 1450 66 192 34:100
Wath pause: ..&..... 3430 569 1020 55:100
Without pause...... (800) 130 | 154 84:100
Without pause...... (1450) 156 164 95:100
Without pause...... (3000) 563 809 69: 100

The last three experiments here instanced were really made with
our fourth and fifth methods which give the average dead space for a
series of breaths. But they suffice to show that the difference between
D, and Dz is much less without, than it is with, a pause. From this
fact as from the evidence of the first method above discussed, it is clear
that there ts a very considerable diffusion of COz from the walls of the res-
piratory passages. The oxygen exchange on the other hand between
the air in these tubes and the blood in their walls is comparatively
small. It is probable that the great part of the blood affected by this

10 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

diffusion is not in the pulmonary, but in the systemic circulation and
flows from the bronchial veins to the right heart.*

_ This bronchial CO, diffusion affords an explanation of the fact that
the respiratory quotient calculated from alveolar analyses is always
lower than that found from the total expired air. Indeed this explana-
tion was pointed out to one of us some years ago by Dr. J. 8. Haldane,'®
and it was his forecast which led us to look for a smaller (virtual) dead
space for CO, than for oxygen.

Third, or hydrogen method. This is the method of Siebeck as em-
ployed by Krogh and Lindhard. In our use of it we made a pause of
some six to eight seconds before the final expiration. We also varied
the depth of the inspiration, the point which, as already explained,
Krogh and Lindhard overlooked. In several experiments we deter-
mined on a single breath the dead space, both for oxygen and for hy-
drogen. The formula for the method, in which H is the percentage of
hydrogen in the gas inspired, m3 and a; are the percentages of hydro-
gen in the mixed and alveolar samples, and the other terms, as in pre-
vious formulae, is
Dig”

d

D,=E-E

Our results with this method are given in the fourth column of the
table of comparative results.

The dead space for hydrogen is, in general, of approximately the
same size as that for oxygen. Like that both for oxygen and for COs,
it varies with the extent to which the lungs are inflated. These obser-
vations do not, however, invalidate the demonstration of Krogh and
Lindhard that with breaths of the same size the dead space is the
same during exercise as during rest.

Fourth and fifth, or Douglas bag methods. The fourth method which
we used was identical with that of Douglas and Haldane. The total
expired air for a certain length of time was caught in a Douglas bag
and measured. This volume was divided by the number of breaths
made in the period to find the mean volume of expiration (e). A
sample was analyzed for CO., and a separate determination of the
subject’s alveolar air was made. The formula was the same as that
for our first method, except that the mean tidal air (e) replaces (2)

*Since this paper was written, correspondence with Dr. Haldane has changed
our opinion on this point. The greater part of this blood must go into the pul-
monary Veins and to the left heart, See note at the end of this paper.

THE RESPIRATORY DEAD SPACE 11

the deepest possible expiration. The extent of the lung dilatation (2)
when needed, was determined separately with a small graduated
spirometer.

The fifth method was the same as the fourth except that oxygen
analyses also were made on the expired and alveolar air. The formula
was the same as for our second method with (e) substituted for (2).

With these methods the dead space for oxygen was always larger
than that for CO.. But as there was no pause between inspiration
and expiration the differences were not nearly so great as between the
results of the first and second methods with a pause.

The effects of exercise (fast walking) were found, in agreement with
the observations of Douglas and Haldane, to include a considerable
increase in the dead space. But this was not to any considerable ex-
tent assignable to active broncho-dilatation, for it was of practically
the same amount as that obtained by the other methods with equal
chest inflation while the subjects were at rest. Furthermore, with
the bag methods when the subject sat perfectly still and voluntarily
made deep, but slow breaths, the dead space worked out to a volume
as great as, or greater than, that during the hyperpnoea of exercise.
For example, the dead space for CO, and for oxygen (Ds and Ds) were
found (on J. L. W.) to be:

TIDAL AIR Ds Ds

At rest, shallow breathing with

constricted che3t................ 180 130 154
At rest, natural breathing. ........ 403 189 198
Fast walking and natural hyper-

TOTDVEIEE. tocol Bee Se SiR oa eee ee 1373 407 650
At rest, deep slow breathing....... 1384 563 809
At rest, deep slow breathing....... 2116 917 1237

Comparative results. In the table of comparative results are shown
data for a single subject by all five methods. The first column indi-
cates the extent to which the lungs were dilated above deepest expira-
tion. In this subject when seated and breathing quietly the tidal
air amounted to 400 or 450 ec. In ordinary inspiration his lungs were
dilated to 1400 or 1500 cc. above deepest expiration. The “vital
capacity’? was 3800 or 3900 ce. The data of the fourth and fifth meth-
dds are inserted at places in the table corresponding to the extent to
which the lungs were expanded at inspiration.

12 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

Table of comparative results of determinations of the volume of the dead space of a
single subject (J. L. W.) by five methods (D, to Ds) arranged according to the

extent to which the lungs were dilated (EF)

E | Di | Dz
720 33 78
725

860 58
1450 66 192
1480
1500 191 296
1650 82 245 °
1710 86 186
1860 370
2430
2540 156
2710
2770 290
3220 1116
3400
3430 569 1020
3450 847
3460 289 620
3550 406
3580
3620 796 1212

cc.

138

102

267
139

234

1105
653

725

928

Ds | Ds
130 154
154 163
132 154
189 198
407 650
563 809
917 1237

Below level of normal
breathing, i.e., with con-
tracted chest.

At level of normal breath-
ing.

At level of hyperpnoea of
vigorous exercise.

The table clearly demonstrates by all methods the passive dilatation
of the dead space with chest expansion.
CO, from the walls of the dead space by the facts that the figures for
D, are somewhat smaller, and those for D, (with a pause) are very
much smaller, than those obtained by the hydrogen (D;) and the two
oxygen methods (Ds and Ds).

There is also to be noted a contrast between the first three methods
(on single breaths) and the fourth and fifth (with the Douglas bag)
in that, while the latter increase fairly uniformly, the former are quite

It shows the diffusion of

THE RESPIRATORY DEAD SPACE 13

irregular. These discrepancies are not due to analytical errors which
can scarcely exceed 10 per cent (except possibly in the third, or hydro-
gen method, where they may reach 20 per cent), nor to variations in
the final measured expiration which should not err more than 100
or 200 cc. in totals of from 1000 to 4000 cc. They are due to the
fact that, while the fourth and fifth methods give the mean dead space
for several minutes, the first three methods determine it at a single
instant. This suggests that the dead space is continually undergoing
active variations which are recorded by our first three methods and
averaged by the last two.

Rhythmic variations in the dead space. When one makes a series of
determinations at regular intervals by any one of our first three methods
on a single subject under conditions and at a chest expansion as nearly
uniform as possible, it seldom happens that any two successive results
agree to within even the extreme errors of the method. This fact which
we have verified repeatedly puzzled us even after many months had
been spent in unravelling the influence of chest expansion on the dead
space. The suggestion was made to us by Dr. A. L. Prince of this
laboratory that there might be variations of tonus in the non-striated
muscle fibers of the bronchi similar to those in other organs contain-
ing such tissue. A survey of the literature shows that Einthoven"
many years ago observed slight rhythmic variations of bronchial tonus
in dogs, although more recent investigators seem not to have noticed
them.

Accordingly we carried through by our first method two series of obser-
vations (on J. L. W.) in which the dead space was determined every
three minutes for more than an hour.* The results are shown in figures
6and 7. In figure 6 are to be seen rhythmic variations of a periodicity
of about 7.5 minutes, large waves amounting to thirty or forty per
cent of the whole volume of the dead space alternating with waves
only half as large. In figure 7 the results obtained on the same sub-
ject the next day exhibit equally marked variations, but of slower and
less clearly marked periodicity.

We may fairly conclude from these experiments, supported by a
mass of observations too voluminous for detailed publication, that
the respiratory dead space, like other cavities having non-striated muscle
' fibers in their walls, is subject to considerable active variations of a more
or less rhythmic character.

*We are indebted to Dr. Prince for assistance in carrying out the necessary
rapid succession of analyses.

14 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

Reflex influences of close and fresh air on bronchial tonus. One of us
(Y. H.) is very susceptible to the ill effects of a close and crowded room.
‘The nasal mucosa becomes congested almost to the occlusion of the
nostrils, and asthmatic sensations also develop. On going out of doors,
or feeling a cool breeze on the face, there occur a reflex constriction
of the nasal blood vessels, and an ease of breathing, after a few deep
breaths, suggestive of an active reflex change from constriction to dila-
tation of the bronchi. The stimulus seems to be the cool sensation
from the face. The effect is too rapid to depend on body temperature.
Sensations of ‘‘stuffiness” and obstructed breathing have been noticed
also in a crowded railroad car when there was no perceptible perspira-

s
Virtual dead space

ie BERS...
cl De

Time Smtr.
frtervals

Fig. 6. A series of dead space determinations (averaging 500 cc.) at inter-
vals of three minutes for over an hour made by our first method on J. L.W. The
chest expansion was as nearly uniform as possible: averaging 3300 cc. for E.
Active variations in the volume of the dead space are here seen to occur rhythmi-
cally, large waves alternating with small at intervals of about seven and one-half
minutes.

tion, no impression of being too warm, and occasionally even when it.
was rather too cold for comfort. We have noticed also that in a Turkish
bath, persons who are unaccustomed to it, or who do not perspire readily,
may experience obstructed breathing, nausea and faintness, while
the practiced bather breathes quite easily and feels exhilarated. The
same holds true, however, in a steam room (Russian bath) with an
atmosphere at 45° C., and saturated with moisture. Under the latter -
conditions body temperature must rise practically equally whether one
perspires or not. The obstructed breathing, if it is a broncho-eon-
striction, together with the nausea and faintness, appear to us, there-

THE RESPIRATORY DEAD SPACE 15

fore, to be rather due to reflexes from the skin than to the central
effects of hyperthermia.”

Early in these investigations there were carried out on Mr. Edmund
Andrews, then a student in this laboratory, a series of observations which
indicated that in a “close” room with a distinct feeling of stuffiness the
dead space is abnormally small; that standing in an agreeably cool breeze
from an open window for even a few seconds induces an enlargement;
that a distinct chill on the contrary induces in one coming from a close
room, not an increase, but on the contrary a further decrease in the volume
of the dead space—in one case the subject “caught cold;” and that a warm
room with no sensation of stuffiness but free perspiration tends to cause
enlargement of the dead space. The methods employed for these ob-
servations were crude (the apparatus shown in figure 3 was used) and

C.C Virteal dead

| JS SSeS ae
moe ee ea
_ aS ae ae

Nee Ap =

Be |-\
ie

20
; mimesis tie ticle Pl fh ris Is
120 ny, rrille Js

7g a
Fig. 7. A series of observations similar to those in figure 6 on the same sub-

ject the next day. The rhythm is much less regular. The mean (330 cc.) of these
determinations is less than before because E was smaller, averaging 2650 cc.

we had not then puzzled out the relation of the volume of the dead
space to chest inflation. The observations are, however, worthy of
record as a starting point for future investigations.

Using our second method and expirations as nearly equal as pos-
sible (3260 and 3140 cc.) we found the dead space (in J. L. W.) in a
close room to be 645 cc. and a few minutes later out of doors—air cool
and dry, and sky clear—a volume of 816. Owing,- however, to our
demonstration of the type of variations discussed in the preceding sec-
tion, it will require many additional observations to settle the point
as to whether this really indicates broncho-dilatation.

We have, however, carried out several experiments on the influence
upon the dead space of steam, hot air, and cold shower baths. The

16 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

results are rather discordant, although they suggest that when one begins
taking such baths the hot rooms cause broncho-constriction, and the
‘cold shower a dilatation, but that after a little practice reactions of
the opposite type occur. Our initial experiments on the effects of the
hot room (65° C. with dry bulb thermometer and 37° with wet bulb),
steam room (44° C. with both wet and dry bulb thermometers), and
cold shower (16°) of a Turkish bath gave the following results for the
oxygen dead space (D.) at nearly uniform chest inflation (£):

SUBJECT Waves Je awe
FE; or amount of lung inflation................ 3150-3320 ce. 3340-36850 ce.
D2 in cool well ventilated room.............. 230 1346
Dz after fifteen minutes in steam room....... 104 218
Dz: after fifteen minutes more in hot room....
Dprattericola:showersean. sede ce oe 634
D; after dressing cool room:.3.2....2.-. a. <*. 338 1124

A few days later, the following results were obtained.

SUBJECT ease Jeune
Ei or amount of lung inflation................] 3650-4100 ec. 3550-38800 ce.
Dei cooldressingroum see. 300 505
D2 after ten minutes in steam room.......... 385 777
Dz after ten minutes more in hot room....... 752 1288
alsen, Cold Shower tos Fa hae oi ae oe 234 557

Dz after dressing in cool room............... 360 1270

One experiment on this topic was made also with the Douglas bag
methods. It showed that (in J. L. W.) after ten minutes in the steam
room and similarly in dry heat the (virtual) dead spaces for CO, and
for oxygen were 200 cc. and 396 ec. respectively, as compared with
189 cc. and 198 cc. when the subject was sitting in a room of
ordinary temperature. But the volume of air breathed per minute,
and the tidal volume were about twice as great in the hot rooms as in
the cool, so that the chest inflation (which we failed to measure) was
probably enough to balance to an unknown extent the apparent broncho-
dilatation. This incomplete experiment, however, yielded one new
observation: it indicated that in great heat the exhalation of COs from the
walls of the dead space is enormously increased. This may well be due to
an active hyperaemia. ‘This exhalation is shown in the fact that although

THE RESPIRATORY DEAD SPACE 17

the subject was breathing twenty times a minute the virtual dead
space for CO, (D4) was only half of that for oxygen (D;). In agree-
ment with the observations of Haldane," the alveolar CO. was consider-
ably lowered. The respiratory quotient for the mixed expired air was
0.928, while the quotient calculated from the alveolar air was only
0.717: an extraordinary difference.

The dead space in asthma. Some observations on asthmatic sub-
jects (by F. P. C. at the Tulane Medical School in New Orleans) show
dead spaces (D;) smaller than in normal subjects although not to the
extent expected.“ It is very probable, however, that their chests
were dilated much above the normal level and that they failed to make
expirations of normal depth in the tests. When these measurements
were made, we were still unaware of the influence of chest expansion
on the dead space. Assuming their chests to have been expanded to
the extent of 1000 or 1500 cc. above the normal their dead spaces were
only about one-third of that of normal subjects at such chest expan-
sions. Jt appears probable that in asthmatics the abnormally expanded
condition in which the chest is held affords a passive stretching of the
bronchi and bronchioles which partially compensates for the active broncho-
constriction.

CONCLUSIONS

From a consideration of the axial flow of gases through tubes, and
from determinations of the CO, content of successive fractions of
the expired air it is found that in man some alveolar air (i.e., CO. in
increasing amounts), begins to issue from the nose and mouth even in
the first 50 cc. of an ordinary expiration. A tidal volume even much
smaller than the volume of the dead space may thus afford u very
considerable gaseous exchange, as in animals during heat polypnoea.
With the ordinary expansion of the chest during quiet breathing an
expiration of at least 400 cc. is necessary to effect an even approximately
complete washing out of a dead space of 150 cc. During hyperpnoea
a much larger expiration is necessary before a sample of pure alveolar
air is obtainable.

Five methods of determining the dead space on man have been em-
ployed with generally concordant results and some significant differ-
ences. The results show that the dead space expands and contracts
passively with the movements of the thoracic walls and lungs. At
the level of ordinary breathing the dead space is about 150 ce. (as
practically all previous observers have found). At shallower levels,
it is, however, considerably less; with deeper breathing it is much

18 Y. HENDERSON, F. P. CHILLINGWORTH AND J. L. WHITNEY

more (400 to 600 cc.); and with the deepest breaths the (virtual) dead
space may exceed a liter in volume. It is much more affected by dia-
phragmatic than by costal movements.

The dead space is found to be of practically the same volume during
rest and exercise, providing that the determinations are made at equal
extents of chest inflation. This fact indicates that the enlargement
of the dead space during hyperpnoea is essentially a passive stretch-
ing and_not an active broncho-dilatation. 3

The dead space for oxygen is always larger than that for CO:. This
is shown to be due to the diffusion of CO, in considerable amounts from
the walls of the mouth, trachea, bronchi, ete.

The dead space, even at a uniform extent of chest inflation, is con-
tinually undergoing active variations in volume. At times these varia-
tions exhibit a distinct rhythm of a period of several minutes. They
may amount to as much as 30 per cent of the mean volume of the dead
space in quiet breathing.

Some facts are reported which suggest that in a “close and stuffy”
room the bronchi, etc., are constricted, and that a distinct chill con-
stricts them further, while pleasantly fresh cool air, on the contrary,
induces broncho-dilatation. Experiments in a Turkish bath have not,
however, afforded concordant results, except that they have shown
that during profuse perspiration and cutaneous hyperaemia there
is a greatly augmented diffusion of CO, from the walls of the dead space.
This doubtless indicates a hyperaemia of the respiratory passages.
As a partof the CO, given off in the passages is inspired into the alveoli
before being expired, and as the alveolar respiratory quotient may be
only a little above 0.7 when the quotient of the mixed expired air is
considerably above 0.9, it appears that under such conditions as much
as one-half of the total CO, exhaled by the subject may come from the
dead space.

In asthmatics, the chronic dilatation of the chest stretches the pul-
monary passages passively, and thus tends to compensate to some
extent for their active contraction.

Note: After our investigations were completed and this paper was
ready for publication, one of us (Y. H.) wrote to Dr. J. S. Haldane of
the results obtained. By return mail, Dr. Haldane replied that he had
been at work on the same topic and had obtained practically identical
results as regards the passive expansion of the dead space.

A few days later he sent to us the manuscript of his paper with
the suggestion that it should either be combined with ours or pub-

THE RESPIRATORY DEAD SPACE 19

lished simultaneously. As publication of the two papers uncombined
appears to us to be the most effective method of carrying conviction
to the minds of others, and as each paper has some special aspects,
Dr. Haldane’s paper (which he has curtailed and modified after read-
ing our paper) appears elsewhere in this number of this Journal. It is
certainly a rare event that investigators, working entirely independently
on opposite sides of the Atlantic, reach so nearly the same conclusions
on a topic which has been in a state of confusion as long as has that of
the respiratory dead space. The only important point of difference,
between Dr. Haldane and ourselves was as to whether the greater part of
the blood, which has lost CQOz directly to the dead space, passes to the
right heart, as we supposed, or to the left heart, as Dr. Haldane sug-
gests. In the latter case the large amounts of COz2 given off to the
respiratory passages during hyperpnoea, and especially during heat
hyperpnoea and bronchial hyperaemia (see our experiments in the
Turkish bath) may prove to be of great importance. After reading
the papers of Miller® to which Dr. Haldane refers there is no doubt in
our minds that Dr. Haldane is correct in considering that the passive
stretching of the dead space occurs principally in the atria, and that
the greater part of the blood from the dead space flows in to the pul-
monary veins and left heart.

1 For literature cf. Siebeck: Skandinavisches Archiv fiir Physiologie, 1911,
KV. Pe OL. :

2 Loewy: Pfliiger’s Archiv, 1891, lviii, p. 416.

3 Douglas and Haldane: Journal of Physiology, 1912, xlv, p. 235.

4 Siebeck: Loc. cit.

5 Krogh and Lindhard: Journal of Physiology, 1913, xlviii, p. 30.

6 Cf. Campbell, Douglas and Hobson: Journal of Physiology, 1914, xlviii, p.
303.

7 Similar observations have been made but misinterpreted by Carter.: Jour-
nal of Experimental Medicine, 1914, xx, p. 81.

8 For literature see Jackson, D. E.: Journal of Pharmacology and Experi-
mental Therapeutics, 1914, v, p. 479.

° Henderson and Russell: American Journal of Physiology, 1911, xxix, p. 441.

10 Pike’s Peak Expedition: Phil. Trans. B. cciii, p. 231.

41 Kinthoven: W. Pfliiger’s Archiv. f. d. gesammte Physiologie, 1892, li, p. 415.

1 Cf. Henderson: The unknown factors in the ill effects of bad ventilation.
Transactions of Fifteenth International Congress on Hygiene and Demography,
1913, ii, p. 622.

13 Haldane: Journal of Hygiene, 1905, v. p. 494.

‘4 Hoover, C. F. and Taylor, L. have reported similar observations but have
interpreted them differently: Archives of Internal Medicine, 1915, xv, p. 1.

15 Miller, W. S.: Journal of Morphology, 1893, viii, p. 165; and Anatomomische
Anzieger, 1906, xxviii, p. 433.

THE VARIATIONS IN THE EFFECTIVE DEAD
SPACE IN BREATHING

J. S. HALDANE
From Cherwell Laboratory, Oxford

Received for publication April 13, 1915

In the foregoing paper by Messrs. Yandell Henderson, Chillingworth
and Whitney, clear evidence is brought forward that the increase in
the respiratory dead space during hyperpnoea, as noted by Douglas
and myself during muscular work, and by Campbell, Douglas and
Hobson for hyperpnoea caused by COz,! is due, not to active dilatation
of the bronchi, as we believed, but to mechanical stretching of the
lungs. I had meanwhile reached the same general conclusions, and it
is unnecessary for me to repeat what my American colleagues have so
admirably expressed; but as my methods differed in certain respects
from theirs, and in one or two points I have been led to a different
interpretation of the data, it seems worth while to put my experiments
on record along with theirs.

Douglas and I obtained our results for the dead space with the
existing natural breathing, deep during hyperpnoea, and compara-
tively shallow during rest. I had been led to suspect that the appar-
ent divergence between our results and those of Krogh and Lindhard
depended on the depth of breathing, and in order to determine the
influence of varying the depth per se, without any hyperpnoea, I made
use of the fact, discovered by Priestley and myself, that the frequency
of breathing may be varied within wide limits, without altering the
alveolar CO, percentage, provided that the depth of breathing is
allowed to regulate itself naturally, with no forcing or holding back.
Thus by varying the frequency one can greatly vary the depth, without
altering the mean alveolar CO. percentage, and without true hyper-
pnoea being present. I have verified this on myself within wider
limits than in our original experiments, in groups of experiments,
the experiments in each group succeeding one another at as short in-
tervals as possible, but the different groups being on different days.

1 Journal of Physiology, xlv, p. 235, 1912, and xlviii, p. 303, 1914.
20

VARIATIONS IN THE DEAD SPACE IN BREATHING 21
It was easy enough to reduce the frequency to three breaths a minute
without any discomfort, provided the inspirations and expirations
were sufficiently slow and regular. The results obtained were as
follows:
Alveolar CO: percentage

F La teeth lated geal END OF INSPIRATION END OF EXPIRATION MEAN
J 30 5.66 5.70 5.67

} 4 5.24 6.09 5.66

24 5.48 5.49 5.48

6 5.40 5.73 5.56

36 5.63 5.73 5.68

4 5.11 6.34 5.72

3 5.19 6.24 5.71

| 60 6.17 6.16 6.16

The constancy of the alveolar CO: percentages with frequencies of
from three to thirty-six breaths is very striking. The cause of the
paradoxical rise in alveolar CO: percentage when the frequency was
increased to sixty will be discussed later.

The effective dead space was now determined by our method with
varying frequencies of breathing, and consequent variations of depth,
the expired air being collected over a period of three minutes, and the
rate of breathing having been accurately adjusted by a clock for at
least two minutes before the collection of samples was begun. Inspira-
tion and expiration were timed to be of about equal duration and
with no pause between them. For reasons which will appear below,
the results, which are given in the following table, are stated somewhat
fully.

This table shows clearly that in spite of the absence of hyperpnoea
the effective dead space increases enormously with increased depth of
breathing, the increase in dead space bearing a rough proportion to
the increase in depth. This fact explains the apparent divergence
between our own results and those of Krogh and Lindhard, since
the latter observers made their determinations with a constant and
relatively small depth of breathing.

The data for oxygen bring out a further point. It will be noticed
that the differences between the oxygen percentages at the end of
inspiration and end of expiration are much greater than the differences
in the CO, percentages: also that the respiratory quotient as calculated

HALDANE

s.

J.

22

661
9é1
1&Z
&6G

G6E
619
06

3G WOIy

SST
It
T9I
TZT
~PGG
GLG
LOV
€89

ZOO woz

peyeENoyVH | pexe[NoyBH INGILOnNod

‘00 ‘ToqIdHLAOW

JO LVHL SANIW Govds

davad GAMOadAat

AUOL
-Vuldsawt
UVIOAATY

00° ST
G9 &T
GL &l
v8 &T

sé Pl
20’ FI
OL &T

uo

uoly
-Bi1dxo
jo puny

GTI
88° ST
WW El
G6 &T
ts vi

86°F
99 TI

uoly
-ei1dsut

jo pul

ADVINGAONAd 6M UVIOAATYV

|

18°¢

metro 2 oO
OF ae
19 D> DOS

1 Se

15 19 19 19 19 19 19.29
oa >

aoa (=)
SS
C © 19 219 19 19 19 O
oD
us
1 21D 1D 19 AD 19 19

AN oO
~ Sis

uory uory
ure -Bidxe -eirdsut

jo pul JO pul

ADVINGOUAd GOO UVIOMATY

028'0 | $4 8T 68°
082°0 | ST ZT GG
888'0 | 86°9T 8o°&
288°0° | T0°2L |= 6S_6

“(118° Q> | Lo LT Ig'&

¢06'0 | 2h 91 ivana
G18°0 | 16 91 9c F
Shs'0 | 20°91 66 F

qgueryonb
A10yw %*O %OO
-11d50 YY,

eS

uly AAuIdxXa

L&E 09
OIF xa
€h9 L Lt
O¢9 O'LI
€89 ost
SIF i)
SEG P
F866 €

‘00 NI da | ALON
-Lvuoiys| uad
olg iv |SNolLvu

SNOIL -IdSau
-Vuldxa a0
40 Hidad | ANNAN

NVIWN -aud

VARIATIONS IN THE DEAD SPACE IN BREATHING 23

from the alveolar samples is lower than the true quotient as calculated
from the composition of the expired air. To the latter point and its
probable explanation attention had already been called by Douglas,
Henderson, Schneider and myself in our account of the Pike’s Peak
Expedition.? It follows that, as shown in the table, the effective dead
space calculated from the oxygen percentages is greater than that
calculated from the CO. percentages, mny results in this respect being
entirely confirmatory of those given in the previous paper.

It will also be seen that in the three experiments with normal breath-
ing, and a depth of breathing of about 650 cc., the dead spaces found
differed considerably, thus also confirming the conclusions of the
previous paper.

It seemed probable to Douglas and myself that the increase in the
dead space during hyperpnoea is due to general relaxation of the bron-
chial muscular coat, so that air can pass more easily. As, however,
the increase occurs without any hyperpnoea, this view becomes un-
tenable: the more so as Professor Dixon informs me that he had mean-
while found in direct’ experiments on animals that no broncho-dilata-
tion occurs on administering air containing CQO..* It is thus necessary
to seek for another explanation.

Considering that the walls of the bronchi are very thick selaivaly
to the walls of the freely distensible air spaces of the surrounding lung-
tissue, it does not seem probable that any considerable dilatation of
the bronchi can be brought about by mere mechanical distension of
the lungs in deep breathing. The position at which the increase
of dead space occurs must therefore, I think, be sought beyond the
terminal bronchioles. The manner in which a terminal bronchiole in
_the mammalian lung breaks up was carefully worked out by recon-
struction and other methods by W. 8. Miller, and is clearly described
and figured in his paper. Miller’s work seems to furnish the key
to the interpretation of the increased dead space. Each terminal
bronchus (see fig. 7 and 8 of Miller’s paper), ends in several openings or
“vestibules,” each of which leads into an air-cavity or “atrium,” lined

Primi. Trans.; B, cciii, p. 221.

’ The method he used was that employed in the experiments of himself and
Ransom on broncho-dilator nerves (Journal of Physiology, xlv, p. 413, 1912).
Einthoven observed broncho-constriction under the influence of CQ: with the
vagi intact, and no effect after vagus section. Einthoven, W: Pfliiger’s Archiv
f. d. gesammte Physiologie, 1892, li, pp. 411 and 423.

4W.S. Miller, Journal of Morphology, viii, 1893, p. 165.

24 J. S. HALDANE

by alveoli. From each atrium several openings lead onwards into “air-
sacs,’ which are main cavities of which the walls are constituted by
alveoli or air-cells. By far the greater number of the lung alveoli be-
long to the air-sac system, but a very appreciable number belong to
the atria, and the latter act partly as air-passages to theair-sacs, and
partly perform the same respiratory functions as the air-sacs themselves.
The walls of the atria have the same general structure as those of the
air-sacs, and must be just as free to expand when air enters the lungs.

It is evident that the atria must have a far greater supply of fresh
air than the groups of air-sacs beyond them, since all the fresh air sup-
plied to the air-sacs passes through the atria, and at the end of an in-
spiration they will be left full of relatively pure air. They will there-
fore contribute to the “effective” or ‘‘virtual’’ dead space due to the
bronchi and upper respiratory passages; and as they will expand freely
with a deep inspiration, and be washed out more thoroughly, the dead
space will increase with a deep inspiration.

If the lungs as a whole are over-ventilated by temporary forced
breathing, the respiratory quotient, as calculated from the composition
of the expired air, is extremely high, since over-ventilation extracts
much extra CO, from the blood, but cannot impart appreciably more
oxygen to it. As, however, the atria are, as it were, constantly over-
ventilated, the part of the expired air coming from them will have a
high respiratory quotient. The air from the air-sacs must therefore
have a lower quotient, so that the mixed expired air, coming from atria
and air-sacs, has a quotient representing that for the body as a whole.
As, moreover, the pulmonary blood passes partly through the atria,
though mainly through the air-sacs, the partial pressure of CO, in the
mixed arterial blood will be slightly lower than that of the blood from
the air-sacs, though higher than that of the blood from the atria. This
conception explains the fact, noted above, that with very shallow and
frequent breathing the excess of CO. and deficiency of oxygen in the
air-Sac air increases. j

It is doubtless true that the respiratory exchange of the bronchi and
upper air-passages must make some contribution to the total exchange
represented in the expired air; but considering the thickness of the
bronchial epithelium and the very small massof the whole mucous mem-
brane lining the bronchi, ete., this contribution must be very small:
whereas the extra CO, represented by the difference in respiratory quo-
tient between alveolar and expired air represents about 12 per cent of
the total CO, given off by the body. Moreover it is only during ex-

VARIATIONS IN THE DEAD SPACE IN BREATHING 25

piration that the bronchial mucous membrane can contribute towards
raising the respiratory quotient above that of the alveolar air, since
any CO, coming off during inspiration is carried down to the alveoli.
It seems, therefore, that the respiratory exchange of the bronchial mu-
cous membrane contributes hardly anything to the difference in res-
piratory quotient between alveolar and mixed expired air. In this con-
clusion I was confirmed by finding that the respiratory quotient of the
first part of the expired air is not strikingly different from that of later
parts. The first part would be expected to have a very high respira-
tory quotient if the respiratory exchange of the bronchi were responsible
for the respiratory quotient of the expired air being so much higher
~ than that of the alveolar air.

Priestley and I found that when, during normal breathing, air is
expelled sharply from the lungs, the partial pressure of CO, in the ex-
pired air is constant after a certain amount of air has been expelled, and
we inferred that the air of constant CO, pressure is alveolar air. It now
appears that the air in question is alveolar air from the “air-sacs”’ of
Miller’s nomenclature, and that the air from the alveoli of his “atria’’
is of a different and more variable composition. I have made a few
experiments in order to test more definitely than in our original experi-
ments the depth of expiration needed in order to obtain air of constant
composition. A rubber bag, of which the capacity when inflated could
be varied at will by the adjustment of a large wooden clamp, was at-
tached to the far end of the piece of tubing used for obtaining samples
of alveolar air. As the inflation of this bag stopped the expiration,
samples of the air leaving the mouth after any desired depth of expira-
tion could be obtained. With ordinary resting breathing (about 18
breaths per minute in my case) the following results were obtained in
successive trials on the same day.

pzorit oF exernaniow | PERCENTAGE OF CO*IN pron oF expmamiox | PERCENTAGE Of COs

1350 5.51 650 Deal °
190 3.03 650 5.32

335 4.37 650 5.18

510 5.17 1350 5.57

510 4.91 950 5.49

1350 5.39 950 | 5.53

1350 Nea yi 1350 5.44

650 4.95 1350 5.63

650 5.23

26 J. S. HALDANE

If we average this series the results are:

DEPTH OF EXPIRATION COs ECE ee areaEn aaa
190 3.03
335 4.37
510 5.04
650 5.19
950 5.51
1350 5.48

In ordinary determinations of the alveolar CO, percentage about
1350 ce. was, so far as I could judge, about the depth of expiration usu-
ally employed in my own case. By somewhat forcing the expiration
about 1750 ec. could, however, be expired without causing undue delay.
A further series (on a different day) was therefore made, to see if any
change could be detected in the deepest portions of the expired air.
The results were that in six successive determinations the mean per-
centage of CO. was 5.39 with an expiration of 900 cc., and 5.36 with an
expiration of 1750 cc. Hence the deeper part of the expiration “con-
tained no more CO, than the middle part.

A few experiments were also made in order to see what depth of ex-
piration is needed in order to reach a constant CO: percentage in a
sample taken at the end of inspiration when the breathing was deep.
The mean results were as follows, with breathing at a frequency of four
per minute, and a depth of about 2000 ec. at 12°.

COz PERCENTAGE IN AIR ISSU-
DEPTH OF EXPIRATION ING FROM THE MOUTH IN A
IN cc. AT 12° SHARP EXPIRATION AT THE
END OF INSPIRATION

First 460 4.74
series 910 Lye bil
2550 5.34

Second ees 5.10
series Selo

2550

It appears from these results that a depth of expiration of about
1500 cc. would be needed to obtain a sample of undiluted alveolar air

VARIATIONS IN THE DEAD SPACE IN BREATHING ai

at the end of an inspiration of 2000 cc. With still deeper breathing the
depth of expiration needed would doubtless be greater.

If some of the current descriptions of the manner in which the ter-
minal bronchioles are connected with the alveoli were correct it would
be hard to offer any explanation of why the composition of the expired
air becomes constant after a certain depth of expiration: for according
to these descriptions the further away an alveolus is from the terminal
bronchus the less fresh air will it receive. Miller’s investigations have
made it possible to explain the actual facts, including the increase of
the virtual dead space with deep breathing.

With regard to methods used for determining the dead space, it seems
worthy of remark that the ‘effective’ or “‘virtual” dead space is a
physiological, and not an anatomical conception. The magnitude of
this space depends on the physiological efficiency of the respiratory sur-
faces in relation to the supply of venous blood and fresh air, It there-
fore seems wrong in principle to use the method of hydrogen inhalation
for the purpose of estimating the effective dead space. It is also evi-
dent that the varying magnitude of the effective dead space with dif-
ferent depths and types of breathing, and the differences of the dead
spaces calculated for oxygen and COs, make the calculation of the com-
position of the alveolar air from that of the expired air a very uncertain
matter. In this connection I may perhaps put on record that my own
experiments have confirmed the observation of Messrs. Henderson,
Chillingworth and Whitney that a pause at the end of inspiration
greatly reduces the effective dead space, as would be expected. In-
crease of depth of breathing must tend to increase the effective dead
space, since the atria are more expanded, but increase of frequency.
must have the same effect, since the air remains for a r a shorter time in
the atria. In very anallow and frequent breathing these two factors
appear to counterbalance one another, as seen in the results obtained
with a frequency of 60 per minute. see

SUMMARY ees
ae

1. phe effective dead space increases enor Pah ith ier eased depth
causes which seis also affect fie dead spat oe

2. This increase is apparently due to Aa
“atria’’ into which the terminal bronchioles

28 J. S. HALDANE

3 The dead space for oxygen is greater than for COs.
4, Estimates of the composition of alveolar air from the composition

are fallacious.

CONTRIBUTIONS TO THE PHYSIOLOGY OF THE
STOMACH

XXIV. Tue Tonus anp HUNGER CONTRACTIONS OF THE STOMACH
OF THE NEwW-BorN

A. J. CARLSON anp H. GINSBURG

From the Hull Physiological Laboratory of the University of Chicago, and the Pres-
byterian Hospital

Received for publication April 16, 1915

The gastric hunger mechanism is probably inherited. At any rate,
the frequency and duration of the periods of gastric hunger con-
tractions are related to the feeding habits of the individuals or the
species only in so far as the feeding time and the food quantity are
factors in the time required for emptying of the stomach, and hence
for the appearance of the hunger contractions.!. On the other hand,
the hunger mechanism determines to a certain extent the feeding
habit. Animals and children probably eat as soon as the stomach is
nearly empty, if food is at hand, and the greater frequency of the
gastric hunger periods in the young, as shown by Patterson for the dog,”
is probably related to the more continuous feeding on the part of the
young animal.

We have now made observations on a number of new-born infants,
and on two pups, born 8-10 days before term, with results showing
- that the empty stomach at birth and in the prematurely born ex-
hibits the typical periods of tonus and hunger contractions of the
adult, the only difference between infant and adult being the greater
frequency and relatively greater vigor of these periods in the young.
In the case of the two pups, and in some of the infants, the observa-
tions were made before their first nursing. It is thus clear that in the
normal mammal the gastric hunger mechanism is completed, physiolog-
ically, and is probably active some time before birth.

! Carlson: This Journal, 1914, xxxiv, p. 169.
2 Patterson: This Journal, xxxiili, p. 423.
3’ Carlson: This Journal, 1912, xxxi, pp. 151, 175.

29

A. J. CARLSON AND H. GINSBURG

0

3

‘IOJIUOUBUT ULIOJOLO] YO

“T9}OUOUBUL UNOJOIOTYD *seynurUL CF : OIL],
PY} SHON “Bursanu 10978 sanoy ¢@ ‘VUBJUI plo-Aep-G B Jo suoly

‘SOINUIU ge ou, ‘porsed
eT FV “Sursunu 4sag a10jaq “JuByuI po INOY-6 B UI SUOTIOBIYU09 1asuny d1148e8 jo potsod xv

iasuny Surpeoosd ey} JO pus ayy uses

‘polied ay} spua YoryM snueyey e}yo/dwo0ur
OBIFUOD LEBUNY o11}sed jo polsed & Sutmoys

SULlIvIT

iu

‘% “BIY

oq ABUL Sure} jo pua

SUIMOYS Sulevry, *] ‘SI

CONTRACTIONS OF STOMACH OF NEW-BORN 31

The recording of the gastric hunger contractions of the new-born hu-
man infant offers no great difficulties. We used delicate rubber balloons
of 15 cc. capacity, attached to a flexible rubber catheter of 2 mm. diam-
eter. Most of the infants swallowed this apparatus without difficulty
and went to sleep in our arms during the observation periods. The
results were always most satisfactory with the infants asleep, as that
eliminated all nervous inhibitory factors, and the disturbances from
body movements and from irregularities in respiration. Practically
nothing can be done with the balloon method if the infant is at all

Fig. 3. Tracings of contractions of the empty stomach of a pup born 8-10
days before term. No food given before securing tracings. Time: 10 minutes.
Chloroform manometer.

restless. All of our observations were made on healthy and vigorous
infants.

The two premature pups were very small, and the balloon introduced
into the stomach via the oesophagus had a capacity of only 4 ce.

RESULTS

1. Human infants. Periods of gastric tonus and hunger contrac-
tions are in evidence shortly after birth and before any food had en-
tered the stomach. These gastric hunger periods exhibit all the pe-
culiarities of the gastric hunger contractions of the adult, except that

32 A.’ J. CARLSON AND H. GINSBURG

the periods of motor quiescence of the stomach between the hunger peri-
ods are on the whole much shorter (10-15 minutes). When the gas-
tric hunger contractions become very vigorous the sleeping infant
may show some restlessness, and even wake up and ery. If the infant
is awake the very vigorous hunger contractions frequently induce
crying and restlessness. Two tracings showing typical hunger periods
in a nine-hour old infant before first nursing, and in a nine-day-old
infant three hours after nursing are reproduced in figures 1 and 2.
The reader’s attention is called to the fact that in both of these infants
the gastric hunger periods end in incomplete tetanus, an index of youth
and vigorous stomach.

2. Prematurely born pups. The observations were made _ before
any food was given to them. The empty stomach of these very small
pups exhibited a continuous motor activity of a character shown in
figure 3. These contractions are not identical with the digestion
peristalsis, because the latter contractions in the dog occur at 15-18
seconds intervals. , The contractions shown in figure 3 last for 30 to
60 seconds or longer, and at times seem to be periods of gastric tetanus.
This is the type of motor activity one might expect to observe with the
slightly inflated balloon in the cardiac end and the empty stomach
in very great tonus.

We are under obligations to Dr. N. 8. Heaney of the Presbyterian
Hospital for facilities in part of this work. We also wish to thank Mr.
I. Tumpowsky and the Misses Jacobson, Rautsche, Clapp, Windmiller,
and Jones for their willing assistance.

FACTORS AFFECTING THE COAGULATION TIME
OF BLOOD

VII. Tue INFLUENCE OF CERTAIN ANESTHETICS

WALTER L. MENDENHALL
From the Laboratory of Physiology in the Harvard Medical School

Received for publication April 1, 1915

In preceding papers of this series evidence was presented that in-
jections of adrenalin result in a hastening of the coagulation time
of blood.t. Other evidence showed that stimulation of the splanchnic
nerves produces a like effect upon coagulation time.2 Several investi-
gators have shown that artificial stimulation of the splanchnic nerves
leads to a discharge of adrenalin into the blood.* Also it has been
proved that certain emotional reactions such as fear and rage occurring
in the normal life of an animal induce a discharge of adrenalin. This
latter effect has been proved to be due to the passage of impulses along
the splanchnic nerves.* Elliot has shown that the adrenalin content
of the suprarenal glands is reduced by administration of various anes-
thetics;> and that this effect with ether and chloroform is due to stimu-
lation of the suprarenal glands through the splanchnic nerves. A
similar result has been shown by another investigator, Oliva.6 The
experiments of the latter showed that chloroform discharges the adrenal
glands more completely than does ether. It was also shown in the
same experiments that the adrenalin content is more quickly regained
after ether anesthesia than after chloroform anesthesia. The question

‘Cannon and Gray: This Journal, 1914, xxxiv, 232.

* Cannon and Mendenhall: This Journal, xxxiv, p. 243.

3 See Dreyer: This Journal, 1898-99, ii, p. 219; Tscheboksaroff: Archiv fur die
gesammte Physiologie, 1910, cxxxvii, p. 103; Asher: Zentralblatt fur Physiologie,
1910, xxiv, p. 927; Kahn: Archiv fur die gesammte Physiologie, 1911, cxl, p. 240;
Meltzer and Joseph: This Journal, 1912, xxix, p. xxxiv, 34; Elliott: Journal of
Physiology, 1912, xliv, p. 400; Cannon and Lyman: Loe. cit., p. 377.

*Cannon and Mendenhall: This Journal, xxxiv, p. 255.

5 Elliott: Loc. cit., p. 388.

§ Oliva: Lyon Chirurg., 1914, ii, p. 11.

33

34 WALTER L. MENDENHALL

of the effect of anesthetics upon coagulation time has long been of prime
importance to both surgeons and obstetricians. Their chief concern,
however, has been in the after-effects as agents productive of post-
operative or post-partum hemorrhages. Chloroform seems to be the
one most often recognized as causing a change in the coagulation proc-
ess. Whipple and Hurwitz’ recently have shown that several hours
after administration of large doses of chloroform to dogs the coagulation
time is unchanged; they call attention, however, to the weak con-
sistency of the clots. They ascribe the cause of post-operative hemor-
rhages following administration of chloroform as a failure of the- clot
to hold firmly rather than a retardation of clotting processes. That
the liver is concerned in the coagulation of blood has been shown 2
“many observers.

The foregoing evidence led to the question of the effects of ether and
chloroform upon blood coagulation during the administration of the
drugs. Inasmuch as experiments recorded in previous papers of this
series were concerned with immediate factors affecting coagulation
time it was thought logical to study the immediate effects of ether and
chloroform upon the coagulation process. These drugs, furthermore,
have been shown to exert action upon organs that are intimately in-
volved in blood coagulation, i.e., the liver and adrenals. It was hoped,
if changes occurred during anesthesia by these drugs, that such changes
might be of value in studying their after-effects or in explaining after-
effects of this form of anesthesia upon coagulation time, and also that
they might throw some light upon the complex of organs involved in
the coagulation mechanism.

The method of drawing blood and recording the coagulation time was
the same as described in a previous paper.* Decerebrate animals
(cats) were used throughout this investigation. Two reasons led to
the adoption of this type of animal; first the animals of the whole series
were placed under practically uniform conditions, and second, the
animal was free from the anesthetic whose action it was desired to
study. It was necessary in the beginning of each experiment to induce
anesthesia for a short time in order to perform decerebration. Ether
was used therefore in the beginning of the experiment. Care was
taken to produce not too profound anesthesia and to remove the cere-
brum as quickly as possible after beginning the administration of the
ether. The usual routine was as follows. Simultaneously with secur-

7 Whipple and Hurwitz: Journal of Experimental Medicine, 1911, xiii, p. 136.

8 Cannon and Mendenhall: This Journal, xxxiv, p. 225.

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 39

ing the animal on the board the ether was applied with a cone, and the
neck was prepared by clipping the hairs. By this time anesthesia
was deep enough to permit operative procedures. The animal was then
tracheotomized, a tracheal cannula inserted, and both ecarotids tied.
Then it was turned over and decerebration performed according to
the method described recently by Forbes and Sherrington.’ The
total time elapsing from the application of ether to its removal never
exceeded fifteen minutes, usually it was from ten to twelve minutes.
After decerebration the femoral artery was prepared according to direc-
tions given in a preceding paper of this series.!° The temperature of
the animal was maintained when necessary by an electric heating pad.
A thermometer was inserted into the rectum. The ether or chloroform
was given by means of the bottle used in ordinary laboratory operations.
It consisted of a small bottle of about 75 cc. capacity. It was stoppered
by a rubber cork through which passed two right angle glass tubes,
each 1 cm. in diameter. One of these tubes conducted air to the
surface of the anesthetic; the other conducted the ether-air mixture
to the animal by means of a short rubber tube connected to the tracheal
eannula. This rubber tube had an oblique cut in the wall so that by
shifting the bottle more air could be mixed with the ether if the animal
showed signs of asphyxia. The corneal reflex was used to determine
if anesthesia was present; also vibrissae, ear and tail reflexes were used.
After all operative procedures were finished the animal was left undis-
turbed for forty-five minutes or an hour. This was done in order that
the animal might be free from ether when observations were to be made,
and also because of the discovery recorded in a previous paper that
operative procedures may shorten the coagulation time. It was felt
that the time mentioned above sufficed to free the animal from the
preliminary small dose of ether and also any hastening factor that may
have been aroused by operations. All experiments began with observa-
tions taken at intervals of ten minutes for forty minutes or an hour to
determine the normal coagulation time of the animal, then anesthesia
was induced by the means described above and observations con-
tinued every ten minutes for an hour.

A total of sixty-three successful experiments were e performed. Pre-
liminary to the investigation of chloroform a number of experiments
were made with chloral hydrate. It was thought that this drug would
give some valuable data which would be indicative of the action of the

®Forbes and Sherrington: This Journal, 1914, xxxv, p. 367.
10Cannon and Mendenhall: Loc. cit., p. 227.

36 WALTER L. MENDENHALL

whole series of chlorine containing anesthetics. Moreover it might
reduce the number of animals which would be necessary for the study of
chloroform. Thus the fatalities resulting from the powerful toxicity
of the chloroform would be reduced. In actual practice, however, the
fatalities due to chloroform were surprisingly small. Inasmuch as
chloral hydrate is frequently used for its anesthetic effect, the study of
its influence upon coagulation is not without value.

TABLE 1
Effect of chloral hydrate

= DOSE IN NORMAL PER CENT PER CENT
EXE SEMEL De SEX MGMS. COAGULATION INCREASE DECREASE
HG: SDS PER KILO TIME COAG. TIME | COAG. TIME
lips apae ine, cee 120 Female 90 2.4 50.0
Re aR eee 2 Male 65 2.8 60.0
chal ce ne 4 Male 65 2.8 142.0
Fela 3 ee en ae 4 Female 65 3.6 30.5
ales tics EER PORE 2 Female 65 4.0 5.0
SSE Sette diss 3 Male 70 6.4 10.9
SOs sete each: 4 Male 70 6.5 1525
SIO LVS 2 eee mare 4 Female 100 6.8 0.0 0.0
ae ane ot 3 Female 70 6.8 20.8
heat eee ye 3 Male 100 6.9 8.7
Ghee Aenea 3 Male 100 7.0 7.0
ARI Oe Po. 1 Female 70 Wat 20E7;
A SA ms Mtr. dee 3 Male 80 7.9 623
Silla ee ete lata So: 2 Female 100 8.0 1337
7 Cs ae lle a 3 Female 70 8.6 — 6.9
AS RT eee ti 0 Male 70 9.3 Use
Gore IR: Mite hee 0 Female 80 9.3 1128
Average...... 6.2 OT oll 10.3

Effect of chloral hydrate. A total of twenty-three experiments were
performed with chloral. Table 1 shows the results of seventeen of
these experiments arranged in ascending order according to length
of the normal coagulation time. The increase or decrease of coagula-
tion time is represented in per cent of the normal. The doses were
given intravenously. Injection was made slowly. It usually took
three to five minutes to introduce the drug. The dose varied from 65
to 100 mgm. per kilogram. Six experiments were performed with
large toxic doses (150-165 mgm. per kilogram). These are not in-
cluded in the table because they are of no interest except from a toxi-

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION oT

cological standpoint. Three of them showed an increase, one no change,
and two a decrease in coagulation time. A glance at the table reveals
the curious fact that the effect which chloral hydrate has upon coagula-
tion bears a distinct relation to the coagulation time before chloral
was administered. It is noted that if the normal coagulation time of
the blood was 6.8 minutes or less, chloral prolonged the coagulation;
whereas, if the normal coagulation time was 6.9 minutes or more, then
chloral decreased the coagulation time. That this effect is not due
to size of dosage is revealed by the table; furthermore it is unlikely
that it may depend upon the sex of the animal or the length of time
it had been in stock. Thus Experiment 42 was a female in stock one-
half day and received a dose of 70 mgm. per kilogram. Its normal
coagulation time was 6.8 minutes. Chloral increased the coagulation

q

oo eile is 2025 S085. 40 45 75095) 60 165 75 80-85 90
Fig. 1. Effect of chloral hydrate.

time 20.5 per cent. Experiment 41 was a female in stock one day;
it received the same dose per kilogram; its normal coagulation time was
- 7.7 minutes. Chloral in this instance decreased the coagulation time
20.7 per cent. This shows clearly that sex and size of dosage are not
the determining factors in the effect of chloral hydrate upon coagula-
tion time. A point of further interest in the above two experiments
was in the weights of the animals. It was a mere coincidence that
their weights were exactly the same 2.6 kgm., and therefore each re-
ceived the same size of dose of chloral. Figure 1 is a composite curve
based upon the results obtained in Table 1. The straight line at the
beginning of the curve represents the average normal coagulation time.
It is extended as a line of dashes through the length of the curve. The
ordinates represent minutes of time for coagulation to occur. The

38 WALTER L. MENDENHALL

abscissae represent intervals of five minutes from the time when the
drug was given. The general averages of these experiments showed
a normal coagulation time of 6.2 minutes. After chloral was admin-
istered the average coagulation time increased 4.8 per cent. Figure 2
is a composite curve constructed in the same manner as the one above.
In this curve only those experiments were used whose normal coagula-
tion time was 6.8 minutes or less. It is noted that only once did the
coagulation fall below the normal, and then only 0.1 minute. The
curve shows the striking effect that chloral hydrate has upon a short
coagulation time. The average increase in coagulation time as shown
by this curve amounted to 28.2 per cent. Figure 3 is a curve which
is composed of those experiments whose normal coagulation was de-

cj

05 pK 2 0 9 0 1 50 55 WU DM GB BO Bs
Fig. 2. Effect of chloral hydrate (short normal).

creased in these observations. The average decrease amounts to 7.5
per cent. The decrease does not correspond to the increase. In the
first paper of this series the per cent of average error due to the method
is stated as 6 per cent. Therefore, the result when the normal coagula-
tion time was long might be regarded as nil. The fact that chloral hy-
drate exerts its retarding effect more when a short normal is present
suggests the idea that it acts antagonistically to hastening factors pres-
ent in the blood or that it may depress activity of organs which produce
or activate hastening factors. The action of chloral hydrate upon
the liver is too well known to need description here. There seems to
be no reference available in regard to its effect upon the adrenals. If
chloral hydrate caused the production of factors that retarded coagula-

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 39

tion one should expect it to exert its retarding effect even though the
normal coagulation was long. Only one clear instance of this is shown
and then a large toxic dose of chloral hydrate was used. Thus in
Experiment 36 the average coagulation time for a half hour preceding
the injection of chloral hydrate was 8.8 minutes; for forty minutes fol-
lowing the injection the average was 9.6 minutes, an increase of 9.0
per cent. This result is not far from the average per cent of error.
Experiment 32 was another (normal 8.0 minutes) in which a toxic
dose of the drug was used. This showed an increase of 16 per cent
over the normal. The evidence in this experiment was not clear, in-
asmuch as the animal became extremely irritable after decerebration
and because of twitching made difficult the drawing of blood. The

q

ee 50535 S55. 45. 30. SO i tS 0. 8h Ae
Fig. 3. Effect of chloral hydrate (long normal).

general disturbance present in this animal made the normal coagula-
tion time doubtful. The animal became quiet after the chloral was
injected. Experiment 33 was another instance in which a long normal
(10.2 minutes) was increased by a toxic dose of chloral hydrate. Here
again the evidence is not clear, because the animal had just come
into the laboratory and in observations lasting an hour before chloral
was given the coagulation time fluctuated from 6.5 minutes to 14.5
minutes. After chloral was given the respiratory center was paralyzed
and artificial respiration became necessary. The increase amounted
to 13.7 per cent. Contrary to this experiment is one (Exp. 37), in which
a toxic dose was given to an animal whose coagulation time was short
(6.0 min.). Here there was practically no change whatever. This
was a male animal, in stock two days. It fought furiously while
being placed upon the board and continued thus until anesthesia be-

40 WALTER L. MENDENHALL

came effective. Experiment 8 is an example of a short coagulation re-
maining unchanged, but the dose was not toxic. This animal, a fe-
male, had been in stock four days. The weight of evidence obtained
in these experiments shows that chloral hydrate, if it affects the coagula-
tion at all tends to prolong it. The prolongation is greatest when the
normal coagulation is short. The evidence does not warrant a con-
clusion that retarding factors are produced.

Effect of chloroform. Fifteen experiments were performed with
chloroform. Table 2 shows the results obtained. The amount of
chloroform used varied somewhat, the average being 10 cc. Anesthesia,
as noted by reflexes, was usually complete in from three to four minutes.
With two exceptions, Experiments 58 and 50, chloroform behaved

8

05 AK 0D 0 HS 05 0 § OO EG CO ee
Fig. 4. Effect of chloroform.

similarly to chloral hydrate—thus if the normal coagulation time was
short chloroform prolonged it, whereas, if it was long a decrease re-
sulted. In these experiments as in those with chloral hydrate a com-
posite curve shows little effect of the drug upon coagulation time other
than to make it irregular. Figure 4 is such a curve constructed upon
the basis of the observations in Table 2. The average normal coagula-
tion time in this set of experiments with chloroform was 6.7 minutes.
The average coagulation time during administration of chloroform was
the same, 6.7 minutes; hence the change in per cent was 0.0. This,
however, was a mere coincidence, since leaving out any one experiment
would alter the figures slightly one way or the other. <A curious fact
is noted in the point where chloroform action changes from an increase
to a decrease of coagulation time. It occurs at 7.5 minutes. With
chloral hydrate it was 6.8 minutes, a difference of less than a minute.

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 41

TABLE 2
Effect of chloroform

EXPERIMENT DAYS IN sex | coacttariox INCREASE DECREASE
[3.5 SURE Peles ree eee 2 Male 4.7 10.6
CO ae 1 Male 5.3 9.4
S135 Renter eae 3 Female al 3.6
0) os nee 1 Female 5.8 ee,
igs sh eae 2 Male 6.0 25.0
ho i i! Female 6.2 8.0
7 2 Male 6.4 Sul
Tk ote 3 Male 6.6 16.6
OTL oer 1 Male 6.6 18.1
5). a 3 Male 7 sil 2S
Sn 215s eee 2 Female eo BAG
Pog bees Ee 1 Female 8.0 18.7
(Cll ood ee 3 Female 8.5 8.2
(TO 3 eee eee 4 Male 9.0 5.5
Uh 3 Male 9.2 8.6

Average...... Gus 10.6 (het

Figure 5 is a curve constructed from observations made in those experi-
ments in which the normal coagulation time was 7.5 minutes or less.

8

~I

[os

n

G5 10 a0 2s $0; 35 40 45-50 55 00 65 1 75
Fig. 5. Effect of chloroform (short normal).
The average normal coagulation time in these experiments was 6.1

minutes. After the chloroform was given there was an increase of 4.9
percent. Figure 6 is a curve in which the normal coagulation time was

42 WALTER L. MENDENHALL

over 7.5 minutes. The average normal was 8.6 minutes. After the
chloroform was given it decreased 11.6 per cent. The increase in coagu-
lation time when the normal was short was not as pronounced as in
-the chloral hydrate experiments. In figure 5, however, there were
included two experiments in which there was a decrease in coagulation
time. If the average decrease is calculated in all experiments that
show a decrease it is found to be 7.7 per cent; whereas, if the average
increase is estimated, it is found to be 10.6 per cent. Moreover the
average short coagulation normal with chloral hydrate was 4.6 minutes,
while with chloroform it was 6.1 minutes. Chloroform is known to
affect two organs that are important in coagulation processes, 1.e.,

10

—O0OS HE DBHSH HHH HW & DF bea
Fig. 6. Effect of chloroform (long normal).

the liver and the adrenals. It would be natural to suppose that its
effect at any one time upon coagulation would depend upon various
interrelations among many factors. If the adrenals were discharged
completely it might still exert an effect upon coagulation by disturb-
ance of liver function; or if the adrenals were highly charged, the re-
sulting outpour might or might not be effective because of an impair-
ment of liver function. The following protocol is typical of cases of
chloroform anesthesia in which coagulation was prolonged.

Protocol Experiment 54.

February 10, 1915. Female cat, in stock 1 day, weight 2.5 kgm.
1.05 Placed on board and anesthetic (ether) begun.
1.15 Tracheal cannula placed and carotids tied.

1.20
1.40
2.45
2.55
3.05
3.15
3.25

3.35
3.45
3.55
4.05
4.15
4.25
4.35

4.36
4.45
4.55

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 43

Decerebration completed, ether removed.

All operative procedures completed.

Coagulation time
Coagulation time
Coagulation time
Coagulation time
Coagulation time

Chloroform administered

Coagulation time
Coagulation time
Coagulation time
Coagulation time
Coagulation time
Coagulation time

Chloroform removed

Coagulation time
Coagulation time

6.0 minutes.
6.5 minutes.
5.5 minutes.
6.5 minutes.
6.5 minutes.

Average 6.2

6.0 minutes.
7.0 minutes.
6.0 minutes.
5.0 minutes.
8.0 minutes.
8.5 minutes.

Average 6.7

8.5 minutes.
6.0 minutes.

Temperature variation in animal 1° C.
Water bath constant

Figure 7 gives a summary of this experiment.

25° C.

In this experiment

the increase amounts to 8.0 per cent, but the curve is typical of the
increase shown by chloroform, i.e., the increase is more noticeable

about an hour after the beginning of the anesthetic.

Previous to the

increase there may be evidence of a disturbance of balance between
opposing forces with a final predominance of the factors that retard
coagulation or perhaps a decrease in effectiveness of hastening factors.
After the chloroform is removed the hastening factor again appears.
The following protocol is of interest because of the opposite effect that
is shown.

Protocol Experiment 61.

March 1, 1915. Female cat, in stock 3 days, weight 3.2 kgm.
Animal placed on board and etherization begun.
Tracheal cannula placed and carotids tied.

1.10
1.20
1.25
2.10
3.15
3.25
3.35

Decerebration completed, ether removed.

All operative procedures completed.

Coagulation time
Coagulation time
Coagulation time

8.5 minutes.
8.5 minutes.
8.5 minutes.

Average 8.5

44 WALTER L. MENDENHALL

3.36 Chloroform administered

3.45 Coagulation time 10.0 minutes.
4.00 Coagulation time 8.0 minutes.
4.10 Coagulation time 7.5 minutes.
4.20 Coagulation time 5.5 minutes.
Average 7.8

4.21 Chloroform removed

5.40 Coagulation time 10.0 minutes.
Temperature variation in animal 2.5° C.
Temperature water bath constant 25° C.

The experiment is summarized in figure 8. In this experiment the
average normal was 8.5 minutes. Forty-five minutes following the
introduction of chloroform the coagulation time was shortened 35.2
per cent. One hour and forty minutes after chloroform was removed
the coagulation time had increased over 50 per cent. The average
coagulation time throughout the whole forty-five minutes of chloroform
anesthesia showed a decrease of only 8.2 per cent. The evidence
obtained in these experiments with chloroform shows a marked resem-
blance to that obtained with chloral hydrate. If the coagulation time
is affected at all it is usually retarded, except when the normai
coagulation time is long, when it may be decreased. The evidence
does not indicate that retarding factors are produced.

Effect of ether. A total of 21 experiments were performed with
ether. Thirteen were made with adrenals intact and eight with adrenals
removed. Table 3 shows the results obtained in the experiments in
which the adrenals were intact. In no instance did ether increase the
coagulation time. An average of 50 cc. of ether was used in each
experiment. The percentage decrease in coagulation time varied
from 0-24 per cent. Figure 9 is a composite curve based upon the
observations obtained in Table 3. The normal coagulation time was
7.2 minutes. Ether decreased it 15.2 per cent. There was no dis-
tinct relation between normal coagulation time and the effect of ether.
The longest normal was 9.4 minutes; this was decreased 11.7 per cent;
the shortest normal was 5.6 minutes; it was decreased 24.4 per cent.
Experiments 16, 17, and 22 showed practically no change, but these
may readily be explained. In Experiment 16 the animal had been in
stock only 12 hours. During the first 35 minutes of ether anesthesia
the coagulation time increased, reaching 37.0 per cent above the normal;
twenty-five minutes later the coagulation time was 19.3 per cent shorter
than normal. It continued with some variation for over an hour.

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 45

|
70 20 30 40 50 60 70 80 90 700 10 720 73

Fig. 7. A summary of Experiment 54. Chloroform was given at 1, removed at 2. Dotted lines
resent averages of coagulation time before and after chloroform was given.

5
Oh: 4/0 75 10 AP OG 0, 5 30 P 60 65 70 HO 15

Fig. 8. Summary of Experiment 61. Chloroform given at 1, removed at 2.
Dotted lines represent averages before and after chloroform was given.

46 WALTER L. MENDENHALL

This caused the final average to appear the same as the normal, when
the real effect was a marked diminution. Experiment 17 showed a
similar initial rise followed by a drop in the coagulation time. The
animal had been in stock 20 hours. In Experiment 22 the animal
died suddenly one hour after the ether was started. The coagulation
time was irregular throughout the experiment. The following protocol
is typical of the ether effect.

TABLE 3

Effect of ether (adrenals intact)

NORMAL PER CENT PER CENT
EXPERIMENT NO. DAYS IN STOCK SEX COAGULATION INCREASE DECREASE
TIME COAG. TIME COAG. TIME
19 3 Male S26 24.4
14 2 Female 6.1 9.8
16 3 Male 6.2
17 3 Male 6.3 ay
21 1 Female 6.5 7.6
12 150 Female 6.8 13ee
13 150 Female 6.8 17.6
11 150 Female fal 23.9
18 2 Female 7.6 21.0
15 2 Female 8.5 15-3
22 2 Female 8.8 1 Ee |
20 6 Male 9.1 8.7
10 ? Male 9.4 WAG
Average lee, 0.0 15.2

Protocol Experiment 16.

November 11, 1914. Animal in stock 2 days, female, weight 2.4 kgm.
1.50 Animal placed on board, etherization begun.

1.55 Tracheal cannula placed and carotids tied.

2.00 Decerebration completed, ether removed.

2.15 Operations completed.

3.00 Coagulation time 7.5 minutes.
3.10 Coagulation time 8.0 minutes.
3.20 Coagulation time 10.0 minutes.
3.35 Coagulation time 8.5 minutes.
3.45 Coagulation time 8.5 minutes.
3.55 Coagulation time 8.5 minutes.

Average 8.5

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION AT

4.00 Ether

4.05 Coagulation time 9.5 minutes.
4.15 Coagulation time 6.5 minutes.
4.25 Coagulation time 7.0 minutes.
4.35 Coagulation time 8.5 minutes.
4.45 Coagulation time 6.5 minutes.
4.55 Coagulation time 6.5 minutes.
5.05 Coagulation time 7.0 minutes.
5.15 Coagulation time 6.5 minutes.

Average 7.2

Animal’s temperature showed variation of 1.0° C.
Temperature of water bath constant 25° C.

TABLE 4

Effect of ether (adrenals removed)

EXPERIMENT DAYS 1N SPER L NE eoraekas acd ie
N AY! N Z ~ - = INCREASE DECRE E
NO. STOCK SESS hee COAGULATION Bonar udavon
? TIME TIME
2D. on eee 4 Male 5.9 1325
PL ee a 3 Female 6.9 2.8
2) 3 SO 0 Male We, Boh
io = : 2 Male fee Se2
i SD eae 5 Male Himes
a) “6 5 Female 7.9 12
23). 7 ee 3 Male oy 6.8
2.0: oe 0 Male 9.3 hes) S
Average...... 725 es

Figure 10 shows a summary of the experiment.

The consistent action of ether suggested the idea that only one factor
- of coagulation, the hastening factor, was affected. Since Elliott has
shown that ether discharges the adrenal gland it was thought desir-
able to remove the adrenals and see if adrenalin was the factor which
was affected. Table 4 shows the results of 8 experiments with adrenals
removed. The results are interesting in that one shows an increase of
13.5 per cent. This was the only experiment with ether that showed
an increased coagulation time. The remaining experiments all showed
a decrease, but the decrease was so slight that it was practically nil.
All might easily fall within the range of per cent of error. Figure 11
is a composite curve based upon observations recorded in Table 4.
The effect of removal of adrenals is clearly shown by comparing this

48 WALTER L. MENDENHALL

20 30 40 50 60 80

Tig. 9. Effect of ether (adrenals intact).

0 10 20 30 40 50 60 70 80 90 100 10 720

Fig. 10. Summary of Experiment 15. Ether given at 1. Dotted lines represent averages befo
after ether was given.

) /0 20 30 40 50 60 70 80

Fig. 11. Effect of ether (adrenals removed).

“ INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 49

curve with figure 9 where the adrenals are intact. The average decrease
per cent with adrenals out was 1.3 per cent. The evidence shows
therefore that ether decreases the coagulation time and that the effect
is exerted through the adrenal glands.

Since the experiments with chloral hydrate and chloroform show so
much irregularity it is conceivable that more than one organ is involved
by them in the coagulation changes. That both drugs affect the liver
is well known and that chloroform discharges the adrenal glands has
been described by two investigators. The action of chloral hydrate
upon coagulation time points to involvement of a hastening and a
retarding process. Possibly the adrenals are discharged by chloral
hydrate, just as they are by chloroform. It was not shown in these
experiments that chloral hydrate or chloroform produces or stimulates
the production of a factor retarding coagulation. The experiments
in which increase in coagulation time was shown might well be explained
in another way. For example, in Experiment 37 mentioned above,
the normal coagulation was 6.0 minutes. A toxic dose of chloral hy-
drate produced practically no change in the coagulation time. This
seems to add additional evidence to support the suggestion made in a
previous paper that a hastening process may involve the action of a
factor upon the liver (or intestine).!!. It may readily be conceived
here that chloral hydrate discharged the adrenal gland, and at the
same time rendered the liver incapable of being acted upon by the
adrenal secretion. As a result no hastening of coagulation occurred.
The animal in this experiment was a male in stock two days and had
fought vigorously. He might easily have largely discharged his ad-
renal glands so that no more was discharged by chloral hydrate. This
would have left chloral hydrate free to stimulate at least temporarily
the production of a retarding factor if it had such power. The evidence,
however, is negative. Thus the coagulation time before the introduc-
tion of chloral hydrate was as follows: 6.0, 5.5, 5.5; afterward it was
6.0, 5.5, 6.5, 5.5, 6.0. The actions of chloroform, which is known to
discharge adrenalin (a hastening factor), may well be explained in terms
of the factor of adrenalin and its action upon a susceptible or non-
susceptible liver. Chloroform showed usually a retarding influence
upon coagulation. This may have been due to its action upon the
liver. Thus it may stop the action of adrenalin upon the liver and
in that way remove the factor that was keeping the coagulation time

1 Cannon and Mendenhall: Loc. cit., p. 250.

50 WALTER L. MENDENHALL

low. In figure 7 above it is seen that the prolongation did not occur
until several minutes after the chloroform was given. Ten minutes
after the chloroform was removed the coagulation process was still
prolonged, but in the following ten minutes it returned to normal.
This may be explained by the liver again becoming susceptible to the
action of adrenalin. The decrease in coagulation shown by chloro-
form may be explained in terms of the adrenalin factor also. Thus in
figure 8 above, the animal had been in stock three days, had become
quiet, and naturally the hastening factor, adrenalin, was not being
discharged. Its coagulation time was consequently long, 8.5 minutes.
It has been shown that a darge amount of adrenalin may produce first
a retardation and then an acceleration of coagulation.’ The first

TABLE 5

Summary of data

roraxd. | oo XOMa oe | Se

RSet SIs TIME TIME TIME
1 LO a See eae 6.2 Chloral hydrate 4.8
OR Set oF: 4.6 Chloral hydrate 28 .2
Site arek te: 8.0 Chloral hydrate 128
GS Se nae 6.7 Chloroform
16 en ee 6.1 Chloroform 4.9
AS Es cee 8.6 Chloroform 11.6
Secor tee wee Ether (adrenals

intact) 1522
Sisk iseee, ac 3 if Js Ether (adrenals

removed) 1.3

application of chloroform may have discharged an amount of adrenalin
which acted upon the liver and producedthe retarding factors before
the liver became insusceptible to the action of adrenalin. As the
chloroform was administered it continued to discharge adrenalin and
when it was removed and the liver again became susceptible to its
action it acted the same as a large dose of adrenalin at any time. It
prolonged the coagulation time.

The evidence with ether points clearly toward involvement of one
factor of coagulation, adrenalin. It might be regarded as a bloodless
method of injecting adrenalin intravenously. Thus in Experiment 16
above the preliminary rise is typical of a large dose of adrenalin and

1 Cannon and Gray: Loc. cit., p. 238.

INFLUENCE OF ANESTHETICS ON BLOOD COAGULATION 51

the subsequent fall is typical. Experiment 17 is a similar effect.
This preliminary rise frequently caused the final average to appear
smaller than it actually was. That adrenalin is the factor involved
is shown in the experiments in whichthe adrenals were removed. That
most of them still showed a decrease may easily be due to discharge
from accessory chromaffin tissues. Some control experiments in which
nothing was given showed that the coagulation time decreased slightly
as the experiment proceeded. No explanation is offered for the one
instance in which ether prolonged the coagulation time when the
adrenals were excluded. Table 5 shows briefly the results obtained in
all the experiments.

SUMMARY.

The observations in these experiments seem to warrant the following
conclusions.

1. Coagulation time is little altered by chloral hydrate unless it is
normally short, then it is prolonged.

2. Coagulation time is affected by chloroform as by chloral hydrate,
i.e., if the process is affected at all it more usually is prolonged rather
than hastened.

3. The effects of chloral hydrate and chloroform are probably the
result of disturbance and consequent interaction between two or more
organs which are important in the coagulation process; probably liver
(intestine?) and adrenal glands.

4. The evidence is not sufficient to prove that a retarding agent is
produced.

5. Coagulation processes are hastened by ether anesthesia.

6. The effect of ether is exerted wholly through its action upon the
adrenals. |

A CALORIMETRIC CALIBRATION OF THE KROGH
BICYCLE ERGOMETER

F. G. BENEDICT ann L. E. EMMES

Contribution from the Nutrition Laboratory of the Carnegie Institution of Wash-
ington, Boston, Mass.

Received for publication April 19, 1915.

A quantitative study of the relation between muscular work and
total metabolism requires the use of some form of apparatus which
will transform the muscular work into measurable units. In a con-
siderable part of the earlier research on this problem with man, the
arms have been employed to rotate ergometers of various forms. Much
valuable research has been carried out by Zuntz and his associates
with a piece of gymnasium apparatus designated by them as the Gaert-
ner ergostat, with the brake ergometer of Zuntz,! and with the dynamom-
eter of Fick.2. Johansson, Tigerstedt, and their associates have also
made studies with Johansson’s extraordinarily ingenious ergostat.®

In recent years the use of some form of bicycle has been recognized
as affording the most practical method for securing large amounts of
external muscular work and at the same time obtaining a relative fixa-
tion of the body of the subject, so as to permit the measurement of the
gaseous metabolism by some type of breathing appliance. One of the
earliest uses of the bicycle form of ergometer was that made by Atwater
and Benedict* in which a small pulley attached to the armature shaft
of a generator pressed against the rear wheel of a stationary bicycle.
When the subject was pedaling, the friction between the wheel and
the pulley developed an electric current in the armature which could
be measured. Subsequently a special type of bicycle ergometer was
devised on the electric brake principle, which was used in a series of
investigations with the respiration calorimeter at Wesleyan University,

' Zuntz: Archiv fiir Physiologie, 1899, p. 372.

* Fick: Archiv fiir die gesammte Physiologie, 1891, 1, p. 189.

§ Johansson: Skandinavisches Archiv fiir Physiologie, 1901, xi, p. 273.

‘Atwater and Benedict: U. 8. Department of Agriculture, Office of Experi-
ment Stations, Bulletin 109, 1902, p. 20.

52

CALORIMETRIC CALIBRATION OF BICYCLE ERGOMETER 53

Middletown, Connecticut, in studying the effective work produced by
several bicycle riders. The apparatus was briefly described in a publi-
cation by Benedict and Carpenter;? more recently a detailed descrip-
tion of the apparatus, together with a series of interesting calorimetric
calibrations made with it, was published by Benedict and Cady.*

Bowen’ published in 1903 a description of an ergometer in which the
rear wheel of a bicycle was replaced by a grindstone, weighing about
50 kilograms, the apparatus being also supplied with a brake. Later
some researches made with this ergometer were published.®

As a means for securing a considerable amount of muscular work,
Haldane fitted up a tricycle, weighting the wheels heavily to convert
them into fly wheels, and using a brake consisting of a strap passed
around the brake-wheel. The records were made on a spring balance.?®

Zuntz exhibited an apparatus at the Dresden Hygiene Congress in
1911 in which a brake was applied to a bicycle ergometer fashioned
much after the design of the Zuntz brake ergometer. To our knowledge
no description of this apparatus has thus far been published. Personal
inspection of it by one of us showed that it is very well constructed
and should give most satisfactory results.

Amar also used a bicycle, the rear wheel being supplied with a special
weight attached with a broad steel band. He adapted a brake to the
instrument and connected it with either a dial dynamometer or a
balance.!®

Boussaguet likewise used a bicycle, fitted with a Prony brake, in
studying the work of miners."

More recently Martin’ has described a simple and convenient
form of bicycle ergometer, maintaining that the instrument has an
error of less than 1.0 per cent. A special cast iron wheel is used in place

® Benedict and Carpenter: U. 8. Department of Agriculture, Office of Experi-
ment Stations, Bulletin 208, 1909, p. 11.

6 Benedict and Cady: Carnegie Institution of Washington Publication No.
167, 1912; Cady and Benedict: Physikalische Zeitschrift, 1912, xiii, p. 920.

7 Bowen: Contributions to Medical Research dedicated to Victor Clarence
Vaughan, June, 1903, p. 462.

8’ Higley and Bowen: This Journal, 1904, xii, p. 311.

Haldane: Journal of Physiology, 1905, xxxii, p. 225.

10 Amar: Le rendement de la machine humaine, Paris, 1910, p. 30; Journal de
Physiologie et de Pathologie générale, xiv, p. 298, 1912; Le moteur humain, Paris,
1914, p. 390.

1 Boussaguet: Recherches expérimentales sur les conditions physiologiques
du travail des mineurs. Thése, Paris, 1912.

12 Martin: Journal of Physiology, 1914, xlviii, p. xv.

54 F. G. BENEDICT AND L. E. EMMES

of the rear wheel of the bicycle; a fabric band is wrapped around this
wheel and connected with two spring balances which indicate the
amount of tension. As yet we have seen no publication of researches
carried out with this ergometer.

Finally Krogh of Copenhagen has modified and greatly improved
the electric brake bicycle ergometer. He has most ingeniously applied
the electric brake principle and has also been able to weigh the
work done with the apparatus. Krogh’s instrument, which was
first exhibited at the International Congress in Groningen, appeared
to be so well adapted for quantitative studies of the relation between
muscular work and metabolism that it was secured for the Nutrition
Laboratory.

With all forms of electric brake apparatus there is always an element
of uncertainty as to the variations in the mechanical friction with
load. Furthermore, there is always present the possibility of the dis-
tortion of the magnetic field as a result of the rotation of the copper disk.
With the original form of ergometer it was observed that there was a
most singular calibration curve and that the amount of work done
was not directly proportional to the number of revolutions, since
the heat per revolution varied considerably with different speeds.
The wholly unexpected calibration curves found by Benedict and
Cady" have made all who attempt to work with this type of apparatus
somewhat uncertain as to the exact values to be obtained at different
speeds. Theoretically, the electric brake ergometer of Krogh elimi-
nates in large part all of the errors incidental to the earlier form of
the bicycle ergometer; nevertheless it seemed necessary to calibrate
the new instrument carefully, preferably in heat units.

When an attempt was made to calibrate the first electric brake
ergometer, it was suggested by Professor Atwater that the instrument
be placed inside of the respiration calorimeter and rotated by means
of a shaft and motor outside of the calorimeter chamber. The amount
of heat generated per revolution could then be directly determined
by the calorimeter. This method of calibration proved most satis-
factory and a series of tests made with the calorimeter at Middletown
and a few years later with the chair calorimeter in the Nutrition Labora-
tory showed a remarkably close agreement.

Since ultimately in metabolism studies a knowledge of the calories
resulting from muscular work is particularly desired, the measurement

'S Krogh: Skandinavisches Archiv fiir Physiologie, 1913, xxx, p. 375.

‘4 Benedict and Cady: loc. cit.

CALORIMETRIC CALIBRATION OF BICYCLE ERGOMETER Do

in heat units of the work done has distinct advantages. Furthermore
such a method of calibration takes into account all errors due to fric-
tion which cannot be estimated by ordinary methods. The friction
tests with the first form of ergometer showed that in all probability,
with a well constructed bicycle, using sprocket wheels and roller bear-
ing chains, the friction was practically negligible. Nevertheless,
the uncertainty as to the exact values obtained with the ergometer
is well pointed out by Krogh in discussing the sources of error in the
apparatus. After conference with Dr. Krogh it was decided that the
instrument would probably prove of such general application for
muscular work experiments that its calibration in one of the respira-
tion calorimeters of the Nutrition Laboratory would be highly desir-
able. Accordingly we mounted the apparatus in a large calorimeter
and have made a series of tests with it which are described in this paper.

The Krogh ergometer consists of an ordinary bicycle from which
the front wheel is removed. The rear wheel is replaced by a copper
disk having a lead ring around the periphery, this weighted wheel act-
ing more or less as a fly wheel. The four magnets, instead of being
mounted permanently as in the original bicycle ergometer, are mounted
on a metal frame which is attached with ball bearings to a prolonga-
tion of the rear axle of the bicycle. A side view of the apparatus, as
partially dismantled and mounted in the chair calorimeter, is shown
in figure 1. In order to save space in the chamber, the front handle
bars and fork have been removed. Another view of the apparatus,
as seen from above, is shown in figure 2. For further details regard-
ing the apparatus, reference must be made to the detailed and dia-
grammatic figures given by Krogh in his description.

To simplify calculations the exact distance between the centre of
the axle to the point of suspension of the pan upon which the weights
are placed has been made equal to 0.3182 meters. By means of a
screw counterpoise the weight of the arm and pan can be accurately
compensated and the whole system brought into perfect equilibrium.
If, then, a weight is placed upon the pan and a current passed through
the fields, when the disk is rotated the pan and weight will be elevated
and suspended free in the air. With a considerable strength of field
current there will be a drag upon the magnets which will ultimately
raise a weight as large as 6 kilograms and hold it in equilibrium. Under
these conditions the work per revolution will be represented by the
weight on the pan multiplied by the circumference of a circle of which
the hub of the wheel is the centre and the point at which the pan is

56 F. G. BENEDICT AND L. E. EMMES

suspended is on the circumference. The ergometer is so constructed
that the circumference of this circle is exactly 2 meters. Thus, for
every revolution of the wheel with a weight of 1 kilogram on the pan,
there will be 2 kilogrammeters of work performed.

Krogh first depended upon hand regulation for adjusting the cur-

Fig. 1. Krogh bicycle ergometer as mounted for calibration in the chair
calorimeter. (Side view.)

rents through the magnets to hold the balance system in equilibrium
and while he subsequently applied an automatic arrangement, hand
regulation was used in all of our tests. To prevent gross movements
of the system due to the irregularity of pedaling by a man, Krogh
has also added a damping device which consists of a heavy sheet of

~I

CALORIMETRIC CALIBRATION OF BICYCLE ERGOMETER 5

i
:

)
HY BD

Fig. 2. Krogh bicycle ergometer as mounted for calibration in the chair
calorimeter. (View from above.)

lead attached to the magnet frame. This lead paddle dips in a trough
containing a thick syrup solution. While this damping device is of
advantage in the practical use of the apparatus, in calibrations of the
ergometer in which the apparatus is driven by an electric motor, irregu-
larities of movement do not exist and hence it was unnecessary to
employ the damping device in the calibrations.

58 F. G. BENEDICT AND L. E. EMMES

The Krogh apparatus permits very considerable variations in the
amount of work performed. Thus Krogh has shown that as large an
amount of work as 2900 kilogrammeters per minute may readily
be done upon the apparatus. It seemed desirable in our observations,
therefore, to calibrate over a considerable range and we were at once
confronted by the fact that while the chair calorimeter, which was
used for these calibrations, is normally not used for experiments in which
the total heat production—that of a sitting man—is much, if any, over
125 calories per hour, here we had to consider the possible production
of 6 large calories per minute, exclusive of the heat developed in the
fields of the magnets by the magnetizing current and the heat due to
a small electric fan, which is used in securing temperature equilibrium
inside the chamber. A more extended system of heat absorbers was
therefore required for these tests and accordingly the ordinary heat
absorbing system, which is shown in the original description of the
chair calorimeter,!® was supplemented by several lengths of heat ab-
sorbing coil, making a total length of 8.7 meters. The disposition
of these extra coils is clearly shown in figure 1.

Under these changed conditions, it was necessary to test the respira-
tion chamber calorimetrically by developing 6 calories per minute
electrically inside the chamber. We therefore checked all of the
calibrations of the ergometer by making electrical check tests in which
approximately the same amount of heat was developed as was actually
developed in the ergometer calibration. The electrical check tests
were usually three to five hours in length and the heat measured varied
from 150 to 350 calories per hour. The results showed that as a eal-
orimeter the apparatus had a very high degree of accuracy.

In mounting the apparatus it was of course necessary to have the
driving mechanism outside of the chamber and for this purpose a shaft
was carried through the copper wall'® of the calorimeter and attached
directly to the pedal bearing of the ergometer. By means of a system
of reduction gears, which varied somewhat with the speeds employed,
this shaft was connected with a 2 kilowatts electric motor outside of
the calorimeter. The connection between the motor, gearing, and
the ergometer is shown in figure 2. A revolution counter attached to

‘6 Benedict and Carpenter: Carnegie Institution of Washington Publication
No. 123, 1909. ; ,

‘To prevent interchange of heat along the shaft the temperature of the
calorimeter laboratory was always kept the same as that of the interior of the
calorimeter.

CALORIMETRIC CALIBRATION OF BICYCLE ERGOMETER 59

the reduction gear shaft recorded the number of revolutions of the
pedal per minute. The weights on the swinging pan of the ergometer
could be varied as desired. The arm of the pan was projected some-
what by the pointer which was held in equilibrium by variable resist-
ance outside the chamber in parallel with the magnets.

To equalize temperature distribution inside the chamber, an electric
fan has been found advantageous in practically all of our work with
the calorimeters, and was retained in the calibration tests. The heat
brought away by the cooling water current of the calorimeter included
the heat developed as a result of the mechanical work, plus the electric
energy required to magnetize the fields and the heat developed by the
fan. In our final computations these two latter factors are deducted.
Throughout the entire series of calibration tests, frequent records were

-

TABLE 1
Results of calibration test of Krogh bicycle ergometer, December 10, 1914

CORREC-
CcOR- TION FOR |HEAT DUE aie TOTAL
WEIGHT] RECTED TOTAL |CHANGEIN/ TOCUR- | )oongp | NUMBER | HEAT PER
TIME OF TEMPERA-| HEAT TEMPERA-| RENTS IN SRG OF REVO- | REVOLU-
WATER | TURE DIF-|MEASURED| TURE OF | MAGNETS |... cURpp| LUTIONS TION
FERENCE CALORIME-| AND FAN ‘ OF DISK
TER
p.m. kilos. SOn cals. cals. cals. cals. cals.
12.26-1.26 | 80.67} 4.482 357 .5 —1.0 85.8 270.7 9627 0.02812
1.26-2.26 | 80.65} 4.310 347 .6 —0.4 90.7 256.5 9024 0.02842
0 1

aalixsw) gant 250. 8874 | 0.02818

2.26-3.26 | 80.62} 4.242 | 342.

made of the voltage and current on the fan and likewise that required
to magnetize the fields of the ergometer. With a rate of revolution
so constant as that obtained with an electric motor, the variations in
the magnetizing current would be very slight and readings every four
- minutes could be relied upon to give an average value throughout the
experiment.

The results of a specimen experiment are given in Table 1. In this
experiment the load on the pan was6 kilograms. There were 47.0revolu-_
tions per minute of the pedals, which corresponded to 152.9 revolutions
per minute of the disk or rear wheel, as the ratio of the pedals to the
rear wheel is 1 to 3.25. This ratio was obtained by the number of
teeth on the sprockets, the large sprocket having 26 teeth and the
smaller sprocket 8 teeth. In the first column of the table is given the
time for each period; in the second the weight of water passing through
the cooling system; in the third the temperature difference corrected |

60 F. G. BENEDICT AND L. E. EMMES
for pressure on the thermometer bulbs; and in the fourth the total
calories measured. Slight capacity corrections are shown in column
five, the heat required for magnetizing the fields and for the fan is given
in column six, and in column seven is given the corrected measurement,
i.e., the total calories measured less the capacity correction and the
heat due to the magnetizing of the fields and to the fan. The total
number of revolutions of the disk or rear wheel is given in the next
to the last column and finally the calories per revolution are computed.
It will be seen that in the three 1-hour periods the heat measured
per revolution was 0.02812, 0.02842, and 0.02818 large calorie respec-

TABLE 2
Results of calibration tests of the Krogh bicycle ergometer, made with the chair
calorimeter

ane EVOUNS HEAT PER REVOLUTION OF DISK }

DATE "or tzsr | EOAP | “sew or | MIN. OF: Ratio of
PEDALS DISK Found Theory found to

theory

1914 hrs. kilos. cals. cals. per cent

Dec 10... 3 6 47.0 152.9 | 0.02824 | 0.02813 100.4
Weewilant vases 4 6 43 .4 141.2 | 0.02828 | 0.02813 | 100.5
Dec lOt ce 3 6 58 .2 189.2 | 0.02814 | 0.02813 100.0
Deeesliipaeee. 3 3 124.6 405.1 | 0.01414 | 0.01406 | 100.6
WEG Silex tevscnts 3 3 84.3 274.1 | 0.01406 | 0.01406 | 100.0

1915

Janseelier 3 Be 144.5 469.7 | 0.00946 | 0.00938 | 100.9
aM eles sey: 3 4 82.7 268.8 | 0.01880 | 0.01875 | 100.3
Janse St eyes 4 § 70.1 227.9 | 0.02351 | 0.02344} 100.3
NETO ott Arete c 4 1 144.3 469.1 | 0.00479 | 0.00469 102.1
Aenne We eyeies Se 4 1 105.3 342.2 | 0.00471 | 0.00469 | 100.4

tively, with an average of 0.02824 large calorie.

Since each revolution

of the disk was equivalent to the load (6 kilograms) multiplied by the
circumference of the circle (2 meters), we have here 12 kilogrammeters
as the amount of work done. Using 426.6 kilogrammeters” as the
equivalent of 1 large calorie, we find that 12 kilogrammeters correspond
to 0.02813 large calorie. Thus in this particular experiment we find
that the ergometer produced 100.4 per cent of the theoretical energy
computed from the weight on the pan and the revolutions of the disk.

The results of all of the ergometer calibrations have been summarized
in Table 2. The tests included calibrations with weights on the pan

‘7 Armsby: Principles of animal nutrition, New York, 1906, p. 233.

CALORIMETRIC CALIBRATION OF BICYCLE ERGOMETER 61

ranging from 1 to 6 kilograms and with revolutions of the pedals per
minute varying from 43.4 to 144.5 and of the disk or rear wheel from 141.2
to 469.7. In other words, the tests include great varieties of speed
and weight, extending quite outside the range of ordinary physiological
experimentation. It will be seen that the percentage of theory found
with this ergometer is striking, the average of the 10 tests being 100.55
per cent with variations from this value exceeding 1 per cent in only
one instance. Since the rate of speed and the load were so greatly
varied, it is reasonable to conclude that the friction plays a wholly
insignificant role with this instrument and may accordingly be entirely
neglected in any computations of the work done. With the Krogh
ergometer, it is therefore justifiable to use without corrections the

formula
W = 2pR

in which W is the work in kilogrammeters, p is the weight in kilograms
and R the number of revolutions of the rear wheel.

SUMMARY.

A Krogh electric brake bicycle ergometer was placed inside of a
bed calorimeter and rotated from the outside by means of an electric
motor. The heat developed was measured by the calorimeter and
has been compared with the computation of the work done in sus-
taining various loads on a suspended balance pan. These experi-
ments showed that friction and other extraneous factors may be entirely
neglected in using the Krogh bicycle ergometer and that the results
obtained by calibration were within 0.5 per cent of theory. The experi-
ments included tests at different rates of speed and with different

weights on the balance pan.

THE EFFECT OF ADRENALIN ON THE HEART-RATE

WALTER J. MEEK anv J. A. E. EYSTER
From the Physiological Laboratory of the University of Wisconsin

Received for publication April 22, 1915

In a previous article! we have attempted to determine some of the
mechanisms by which the immediate increase in pulse-rate at the be-
ginning of a period of exercise is brought about. Our conclusions were
that the increase in rate is produced principally by a decrease in vagal
tone. Although acceleration may take place through the accelerators
the evidence seemed to indicate that they act chiefly in maintaining
the resting pulse-rate and that their accelerating action is a factor of
safety superimposed on the more labile vagus mechanism. Incidental.
to this work we were able to make some observations on the action of
the adrenals.

Animals with the accelerators removed and the vagi cut did not show
a noticeable increase in heart-rate unless there were symptoms of as-
phyxia. Soon after the vagotomy when the signs of respiratory dis-
tress were most marked, exercise produced a great increase in pulse-
rate. In the succeeding days as the signs of asphyxia decreased the
exercise acceleration grew very much less. That the marked increase
of rate had been due to the asphyxial secretion of the adrenals we
proved by ligating these organs. Since the adrenalin seemed to be a
factor only when exercise was pushed to an extreme we concluded that
in the normal animal with the vagi intact the secretion of adrenalin
could hardly be considered as one of the mechanisms which immedi-
ately increase the heart-rate on exercise. We still believe this holds
for moderate exercise when uncomplicated by emotional excitement.

It is to be noted that our argument does not rest on the fact that
adrenalin as usually given experimentally produces cardiac inhibition if
the vagi are intact, but that the conclusions are based on our direct
experiments. In our discussion we did overlook the fact however that
as exercise increased in amount and the vagus tone grew less and less

1 Gasser and Meek: This Journal, 1914, xxiv, p. 48.

62

EFFECT OF ADRENALIN ON HEART-RATE 63

there must of course come a time when any secretion of adrenalin might
have an accelerating effect on the heart-rate.

Recently Hoskins and Lovellette? have objected to any conclusions
that adrenalin does not always cause acceleration in exercise on both
theoretical and experimental grounds. They point out in the first
place that the doses of adrenalin experimentally shown to give inhib-
itory effects have been far in excess of the physiological secretions
determined by their previous work. The rate of injection too has been
much more rapid than the normal adrenal discharge. There is no gain-
saying these important and just criticisms which, however, we feel do
not apply to the experimental part of our work.

In the second place Hoskins and Lovellette present a table of 28 ex-
periments on dogs in 18 of which there is increased heart-rate following
the injection of adrenalin in physiological amounts. An examination
of their data shows that the initial pulse-rate was in all cases very high.
The lowest rate recorded is 144 per minute and the highest 258. The
average is 188 per minute. This indicates a low vagal tone which was to
be expected from the ether anaesthesia. It is by no means surprising
that under these conditions 18 of the dogs showed an increase in heart-
rate with small doses of adrenalin. That the heart itself is exquisitely
sensitive to adrenalin we have had occasion to observe in work on the
isolated heart. As soon then as the vagus mechanism becomes suf-
ficiently depressed and the secretion of adrenalin becomes sufficiently
large stimulation of the heart would of course result. The animals of
Hoskins and Lovellette represent therefore to our mind a picture of
extreme exercise, the decrease in vagal tone in their case being due to
the ether anaesthesia. Limited to such conditions their conclusions are
eminently correct. Adrenalin in a time of extreme muscular exertion
when the decrease in vagal tone has done all the accelerating it can,
- is then able to show its direct stimulating effect on the heart muscle.

Although our previous work seemed to show that in dogs the adre-
nals were not concerned in any large degree in the increased pulse-rate
which resulted from two minutes running, it has seemed worth while
to study directly the effect of adrenalin when injected in small amounts
into the unanaesthetized animal.

To accomplish this the dogs were laid on the floor and held until
they had become perfectly quiet. The injection was made into an
ear vein by a hypodermic needle which was connected with a burette,

2 Hoskins and Lovellette: Journal of the American Medical Association, 1914,
Ixili, p. 316.

64 WALTER J. MEEK AND J. A. E. EYSTER

the fluid being under air pressure in the burette. By means of this
pressure and the proper opening of the burette cock, the injection could
be made absolutely regular and at any desired speed. Usually the
needle could be inserted without the dog even moving. In case the
animal flinched at the prick of the needle, it was again quieted before
the experiment proceeded. The right front and left hind leg of the dog
were connected to the string galvanometer and electrocardiographic
records of the heart-beat were taken at one-half intervals immediately
before, during and from two to four minutes after the injection ceased.
Each record was at least one respiratory cycle long, usually several be-
ing included, in order to avoid respiratory variations obscuring the
count.

Adrenalin hydrochlorid and crystalline adrenalin were used. The
amount injected varied from 2-38 ec. and the strength from 1: 50,000
to 1:500,000. The duration of the injections varied from 1 minute 20
seconds to 3 minutes 55 seconds. These quantities were chosen so
that the injection might simulate in some degree at least the physio-
logical secretion.

The results of our experiments are presented in Table 1. It will be
seen that in 25 injections on nine different dogs, in all but two the heart
was slower at the end of the injection than at the beginning. Even
these two cases are really not exceptions for in experiment 5° the rate
was as low as 55 per minute during the second minute of the injection.
By the end of the injection it had however risen to its former level.
In experiment 9° also the rate at the end of the first minute had dropped
to 102 but by the end of the injection it had risen to 132 per minute.
In all cases then a slow injection of adrenalin in doses approximating
its physiological secretion has given us in the intact resting animal a
slowing of the pulse.

According to the generally accepted belief that this slowing takes
place as a reflex through the vagus, we should expect less effect in those
cases with a depressed vagus mechanism, i.e., those hearts with a high
rate due to loss of vagal tone. The data bear this out in a few instances
rather definitely. Variations in dose and time of injection make only
a few comparisons possible, but in experiment 3 it may be noted that
a rate of 150 was reduced only to 141 while under almost identical
conditions a rate of 129 fell to 108.

The pulse rates for the successive half minutes show that the effect
of adrenalin in these doses comes on in from one-half to one minute
after the injection starts. With the higher dilutions particularly the

EFFECT OF ADRENALIN ON HEART-RATE 65

pulse-rate may begin to increase again before the injection is ended.
For this reason the table does not always give the lowest rate recorded.
With the stronger concentrations the heart did not begin to resume its
normal until the injection was finished.

TABLE 1

Showing the influence of adrenalin in small doses on the pulse-rate of nine un-
anaesthetized dogs. In experiments 7 and 8 crystalline adrenalin was used; in all °
others adrenalin hydrochlorid. The exponents represent successive injections in the
same experiment

: PULSE RATE
EXPERIMENT NO. specs rials aa cc. INJECTED STRENGTH poop ideae Se o

il 88 305 1: 50,000 be 64
a 140 3.0 1: 50,000 PHB De 65
3} 93 3.0 1: 50,000 Ze ADe 60
2 69 3.0 1: 50,000 PES Ve 45
33 100 3.0 1: 100,000 PAE AN Ved 90
3 129 Das 1: 100,000 3 108
3° 150 3.0 1: 100,000 3 05" 141
41 124 3.0 1: 100,000 1/0307 90
42 138 3.0 1: 50,000 WesOe 70
43 122 3.0 1: 50,000 Sie 90
5} Uz 3.0 1: 50,000 ee: 68
52 87 3.0 1: 500,000 Nee20% 60
53 67 220 1: 100,000 BEAD 67
6! 99 3.0 1: 100,000 ashe 66

2 90 3.0 1: 500,000 1’ 40” 75
68 86 2.0 1: 500,000 3 78
7} 96 Pas 1: 50,000 of 66
7? 90 20 1: 200,000 150" 48
81 108 2.0 1: 50,000 12052 75

2 99 2.0 1: 50,000 1 ate 72
83 117 2.0 1: 200,000 PIG 96
84 120 2.0 1: 200,000 1’ 45” 84
g} 117 2.0 1: 50,000 1b? Bsc! 66

2 123 2.0 1: 50,000 DOSE 96
93 123 220 1: 200,000 DATO™ 132

The pulse-rate was followed in all cases for from one to four minutes
after the injection. In nearly all experiments the heart had reached its
former rate or exceeded it by the end of the third or fourth half-minute.
In 12 of the 25 separate injections the rate after the injection had been
finished rose somewhat above its previous level.

66 WALTER J. MEEK AND J. A. E. EYSTER

SUMMARY

Intravenous injection of physiological amounts of adrenalin into in-
tact unanaesthetized dogs with good vagal tone invariably causes a
decrease in the heart-rate.

The action of the adrenalin is doubtless twofold; it accelerates the
heart by direct stimulation and inhibits it reflexly through the vagus.
In our experiments the net result of this balanced mechanism was al-
ways a decrease in pulse-rate.

For this reason a secretion of adrenalin could hardly play a part in
the immediate cardiac acceleration which follows moderate exercise.

In exercise which has been pushed to an extreme degree, however,
the accelerating action of adrenalin might well become predominant,
for in this condition we have a great decrease in vagal tone. In regard
to the heart-rate then adrenalin seems to act just as it does in so many
other instances; that is its stimulating effect becomes apparent at a time
of great physiological need.

THE INFLUENCE OF THE OIL OF CHENOPODIUM ON THE
CIRCULATION AND RESPIRATION

WILLIAM SALANT anp A. E. LIVINGSTON

Pharmacological Laboratory, Bureau of Chemistru, U. S. Department of
Agriculture

Received for publication April 26, 1915

The frequent occurrence of circulatory and respiratory disturbances
observed in the course of studies on the toxicity of the oil of chenopo-
dium (1), (2) suggested the present investigation. The experiments
were carried out on different animals, several anesthetics being used
for the operative procedures. “Morphine-ether anesthesia or chloretone
in alcohol was employed in experiments on dogs, ether alone was given
to cats, while rabbits were anesthetized with ether or urethane, the
latter being given by mouth through a stomach tube.

The oil of chenopodium was introduced in the form of 1 or 2 per
cent emulsion, which was made up by adding the oil to 0.9 per cent
salt solution, usually but not always containing 5 per cent of acacia, 1
per cent cocoanut oil, and a few drops of sodium carbonate solution.
The mixture was then vigorously agitated in a shaking machine until
a very fine emulsion resulted. This was injected into the femoral vein
from a burette. An emulsion containing all the ingredients in the same
proportion, but without chenopodium, was used as a control, and was
_ Injected into the femoral vein on the opposite side. It may be stated
in advance that blood pressure, after the administration of the control
solution, was only slightly increased or not affected at all. No change
in respiration was observed at any time. The injections were made
from 50 cc. graduated burettes which were held in a Liebig condenser
containing water that was maintained at body temperature or a little
below by allowing it to circulate through a glass jar, the contents of
which were kept warm by means of an electric bulb immersed in the
liquid. Blood pressure was recorded by a mercury manometer con-
nected with the carotid artery in the usual fashion, a saturated sodium
sulphate solution being used as an anti-coagulating fluid. Respiration
was recorded by two receiving tambours, one applied to the thorax

67

638 WILLIAM SALANT AND A. E. LIVINGSTON

and the other to the abdomen, each of these being connected with a
recording tambour by means of rubber tubing. In some experiments
the respiration was recorded by a tambour connected with the trachea.

The volume of the kidney was recorded by means of the Roy oncom-
eter in the earlier experiments but air transmission alone was sub-
stituted later, the kidney being placed into a covered aluminum box
provided with an opening for the renal pedicle. By means of lanolin, the
oncometer was made
air-tight. One end of
a brass tube was fitted
into an opening in the
top of the oncometer,
the other end being
connected by means of
rubber tubing, with a
recording tambour.
Variations in volume
of the kidney could
thus be readily ob-
served.

NAKA AM ri a 4ehh
WAAAY is hi A

EXPERIMENTS ON DOGS

pace 2% Ol Chemaprdinny. Effect on the circu-
| ist ps MOA He lation
eediamny cerbortl peer so

(Le Fone ie 053 A fall of blood pres-
sure, varying in extent
and duration was usu-
ally observed after the
Fig. 1. Dog 218. First injection of 0.02 ce. ra evoneae injceue
chenopodium per kilo. Fall of blood pressure, of the oil of chenopo-
apnoea, and decreased amplitude and rate of re- dium. In experiments
spiration. in which morphine-
ether was employed,

the initial dose of 0.02 cc. per kilo caused a decrease in blood pressure
of 10 to 15 mm. Hg or 7.5 to 10 per cent (fg. 1). When a second in-
jection was made, the effects of the same amounts were more pro-
nounced. The fall of blood pressure increased from 7.5 to 12 per cent
in One experiment, in another from 10 to 24 per cent, in a third from
9 to 14 per cent, there being little or no difference in the speed of in-

|
i}

nti

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 69

jection. In one experiment, however, the effect was more marked, and
was also reversed, the initial injection of the same dose hav ing produced
a fall of blood pressure amounting to 35 per cent, while a second injec-
tion produced a fall of blood pressure of 16 per cent.

lil lil te
ayy)

AVA AAAAL ELTA nh ii HAAR
Lartcie (Qapeicriticn!

/4CC 2% Oik of Chamopedinend <“ SH acacia!
iw 2.9% Wa C6 +7 Po Cocwamith Dib anc | trope vp

[Ok eet corbensl/ her jooce A forerel ie ad boy Lough
gr encod dd 4G

Fig. 2. (a) Dog 221. Chloretone—alcohol anesthesia. Marked fall of blood
pressure al diminished volume of kidney after the injection of 0.02 ce. cheno-
podium per kilo. Respiration only slightly depressed. (b) Also shows recovery
15 minutes later.

When anesthesia was produced by means of chloretone and alcohol.
the initial injection of 0.02 ce. oil of chenopodium per kilo decreased
blood pressure 40 to 50 per cent (fig. 2). A second dose, given after
an interval of 30 minutes, was followed by the same effect in one ex-
periment, but in two others blood pressure fell considerably less, the

70 WILLIAM SALANT AND A. E. LIVINGSTON

decrease being 20 instead of 50 per cent in one experiment, and about
30 per cent in another in which the initial injection caused a fall of
blood pressure of 65 per cent.

Lh ws gee’ Ofiety

asce 2% OL Chausfee desir
an Emulsion of acaccay
org Aan” :
a EASA he 2 a ES Mee re

Aiwes Suter hee ace J
MareburrcisrLsvicveercrverscuseOcse veer ir beriststonrueeevecicsesasuriceteese

Curstl Pest sure USUTELeVSTeSELSEMTLINUSSNSITELETESUITEETEVULIVNVETULTUCTOeTeCUEIVUEE NO THtSEESEIT Tener CL cccrebee CLE

Fig. 3. Dog 210. 0.06 ce. chenopodium per kilo. Showing marked fall of
blood pressure and apnoea.

A number of experiments with larger doses have also been carried
out. The initial injection of 0.05 and 0.06 ce. oil of chenopodium per
kilo produced in dogs, under morphine-ether anesthesia, a fall of blood
pressure from 160 mm. to 80 mm., or 50 per cent in one experiment
(fig. 3). In another it fell from 155 to 120 mm., or 21 per cent, although

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 71

the rate of injection in both was the same, being 1 ce. in about four
seconds. In a third subject, likewise under morphine-ether anesthesia,
in which the speed of injection was 1 cc. in 16 seconds, 0.05 ce. oil of
chenopodium per kilo produced a fall of blood pressure from 115 to
75 mm., or 34 per cent. It is evident, therefore, that the rate of in-
jection does not modify the action of oil of chenopodium on blood
pressure. Similar results were obtained with smaller doses in several
experiments. The rate of injection is an important factor, however, in
determining the character of the fall of blood pressure, which was
abrupt when carried out rapidly, but gradual when the speed was
reduced.

The effect of repeated injections varied in our experiments, a second
dose of the same size, as already noted, producing in some cases an
increased fall of blood pressure, In others, on the contrary, this was
much less than that caused by the previous initial injection. Subse-
quent injections decreased the reaction to chenopodium until the re-
sponse of the circulation completely disappeared (fig. 4) even when
larger amounts were introduced. Although the quantity required to
produce this effect varied in different individuals, a progressive de-
crease in the intensity of the reaction was observed until blood pres-
sure was no longer affected by chenopodium. This condition was usu-
ally attained when the total amount received exceeded 0.2 cc. per kilo,
although in two experiments the depression of the circulation was only
slightly affected after doses of 0.105 and 0.12 cc. per kilo had been in-
troduced. The action of repeated injections is illustrated in the fol-
lowing experiments:

In one experiment a dose of 0.085 ce. per kilo produced a maximum
fall of blood pressure of 55 per cent. The decrease in the next injec-
tion of the same amount was 35 per cent, but after the following in-
jection the fall of blood pressure was only 20 per cent.

In another experiment (fig. 3) the initial injection of 21 cc., or 0.06
ec. per kilo, produced a fall of blood pressure of 50 per cent, and only
7 per cent at the seventh injection when the same dose was given, the
total amount previously injected being approximately 0.3 cc. In the
same experiment 0.04 cc. per kilo, given in a third injection was fol-
lowed by a fall of blood pressure of 48 per cent. Such a dose given in
the eighth injection had no effect on blood pressure, the rate of injec-
tion being practically the same.

In a third experiment 0.02 cc. per kilo, introduced in 15 seconds at
the fifth injection, caused a fall of blood pressure, amounting to 20 per

7(P4 WILLIAM SALANT AND A. E. LIVINGSTON

cent. Double this dose per kilo at the tenth and eleventh injections,
after a total of 0.236 cc. had been given, failed to produce any change
in blood pressure.

rR. &

40 68 24 OL LOhrtebinrr
SEES vial Ne reas Fe"
FSH Oren otk and ° Fhe
10 Faerie Carte nake/ he, soot
Ake Srmveral cee o¢ ag Ch

~~

Ly
i AAS Be Saew

Fig. 4. Dog 219. Prolonged apnoea; 33 minutes after injection of chenopo-
dium, then dyspnoea. Blood pressure not affected by the injection of 0.08 ce.
chenopodium per kilo. Total amount previously injected 0.24 ec. per kilo.

In a fourth experiment 0.04 ec. per kilo, injected in 35 seconds, low-
ered blood pressure 20 per cent at the sixth injection. A dose of 0.085
ec. per kilo, given at the ninth injection, produced a rise of blood pres-
sure of 11 per cent (which might be expected if the same volume of the
control solution were injected). Such a dose injected 34 minutes pre-

viously lowered blood pressure 25 per cent.

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 73

Similar results were obtained in several other cases. The condition
of the blood pressure at the time the failure to respond to oil of cheno-
podium occurs may appear to be an important factor in the determina-
tion of this result, for blood pressure at the time of the injection was
considerably lower than at the beginning of the experiment. Analysis
shows, however, that this is not the case. In two experiments in which
blood pressure fell from 130 to 85 mm. Hg. approximately the same
amount of chenopodium produced a fall of blood pressure, in one case
35 per cent, in the other 23.5 per cent. In three subjects in which
blood pressure fell from 155-160 mm. Hg. at the beginning of the ex-
periment to about 100 mm. Hg, the reaction to chenopodium was
fairly good in two, a fall of blood pressure of 35 per cent being observed
in one case after injecting one gram, or 0.085 cc. per kilo. In another
0.06 ec. per kilo produced a fall of blood pressure of 30 per cent. Ina
third 0.04 lowered blood pressure 10 per cent.

It is evident, therefore, that the reaction to chenopodium may per-
sist though blood pressure had fallen one-third. The experiments car-
ried out under alcohol chloretone anesthesia afford additional evidence
that the reaction of the circulation to oil of chenopodium is not de-
termined by the height of the blood pressure, for it will be recalled that
small quantities caused a very marked effect when the maximum blood
pressure was only 100 mm. Hg, while in many instances in which the
height of the blood pressure was 40 to 50 mm. Hg, or even less, the ad-
ministration of oil of chenopodium produced depression of the circula-
tion to a very considerable degree. Decreased reaction or complete
failure of the circulation to react to drugs after repeated injections
have been reported by several investigators. Sollmann and Pilcher (3)
observed in dogs that large doses of caffeine lowered blood pressure to
50 or 60 mm. Hg, but additional injections failed to cause a further
decrease. This is attributed by them to the low blood pressure pre-
ceding the injection, since a dose of caffeine which was without effect
when blood pressure was under 45 mm. Hg produced a considerable
change if injected during periods of high blood pressure. Blackford
and Sattford (4) found that the initial dose of an extract of goitre low-
ers blood pressure in dogs but subsequent injections produced little or
no effect. Roth (5) reported similar observations recently with Pitui-
trin in different animals.

Changes in the volume of the kidney indicate that the fall of blood
pressure is of cardiac origin since the oncometric records have almost
always varied directly with the changes in blood pressure, being almost

74 WILLIAM SALANT AND A. E. LIVINGSTON

parallel after some injections (fig. 2). In some experiments in which
blood pressure no longer reacted to oil of chenopodium, its introduction
into the blood stream, especially when large amounts were injected, was
followed by an appreciable increase in volume of the kidney, although
smaller quantities failed to produce similar changes. That cardiac
depression was produced by chenopodium was also indicated by a signifi-
cant reduction in the
rate of heart action,
which may be seen on
examination of figures
6 and 7. A marked de-
crease in cardiac vagus
irritability was ob-
tained, even after mod-
erate doses were ad-
ministered. When the
vagi were divided after
the large amounts of oil
of chenopodium were
injected, blood pressure
rose a few millimeters or
remained unchanged.

Effect on respiration

The intravenous in-
jection of the oil of
chenopodium was like-
wise followed by depres-
sion of respiration.

Fig. 5. Dog 209. 0.021 cc. chenopodium per kilo. This varied with the
Apnoea, with decreased respiratory amplitude and size of the dose and
fall of blood pressure. sometimes also with the

speed with which it was
introduced into the blood stream. Small doses, 0.02 to 0.028 ce. oil per
kilo of body weight, produced with few exceptions, a decrease in the
amplitude, the rate of respiration being affected much less frequently
when such amounts were introduced directly into the circulation. In
some instances arrest of respiration for a brief period (Dog 209, fig. 5),

with prompt recovery was observed. The activity of larger doses was
much more marked. The injection of 0.04 cc. oil of chenopodium per

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM

a |
or

kilo produced slowing of respiration more frequently than with small
doses, the amplitude of thoracic respiration being decreased consider-
ably at the same time. The effect on abdominal respiration was less
pronounced (fig. 6, Dog 210). Very striking results were noticed in

7

hhore mal Jiro

YUU UL

Fig. 6. Dog 210. Apnoea. Respiratory amplitude decreased. Marked fall
of blood pressure after the injection of 0.04 ec. chenopodium per kilo.

some experiments (fig. 7), when the doses were still more increased.
After the injection of 0.06 cc. oil of chenopodium the rate was very
frequently diminished; the amplitude of thoracic respiration also showed
considerable reduction, abdominal respiration, however, being much less

76 WILLIAM SALANT AND A. E. LIVINGSTON

affected in this case. The initial injection of such a dose, produced
apnoea which lasted nearly 20 seconds. Respiration improved at the
end of this time but the rate was only 50 per cent of that preceding the

| \
i Atdomn wal Tick

2ée
an

Shine footiume

cut

Ci
¢ Ln a

Bark

—~

-J~

Fig. 7. Dog 210. Showing prolonged apnoea and diminished frequency of
respiration after the intravenous injection of 0.085 ec. chenopodium per kilo.
Also marked fall of blood pressure and slowing of heart action.

injection. Recovery took place in two minutes. A second injection
of 0.085 ce. per kilo (Dog 210, fig. 7), given 16 minutes later caused
apnoea which lasted about 50 seconds. In another experiment the initial
injection of the same dose was followed by a decrease of frequency of

~I

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 7

about 50 per cent. The amplitude of thoracic respiration decreased
about 25 to 30 per cent. Abdominal respiration, on the other hand,
became more prominent in this case.

The action of repeated injections indicated a tendency to cumula-
tive effect. Even after complete recovery from previous treatment
was obtained, another dose of the same size was followed by a more
pronounced depression of respiration which was obtained with small,
as well as with large doses.

Observations were also made on the effect of the speed of injection.
When 0.02 cc. per kilo were introduced at the rate of 15 to 30 ce. per
minute the action was practically the same, but the introduction of
0.05 cc. to 0.06 cc. per kilo, given at the rate of 7.5 ce. to 18.0 cc. per
minute, showed differences that were very striking. The slower in-
jections of small and of large amounts produced a slight decrease of
amplitude of thoracic respiration which was preceded by slight stimu-
lation, but rapid injections were followed in the case of small doses by
a marked reduction in rate and decrease in amplitude of thoracic respi-
ration lasting more than 13 minutes, while large doses were followed
by apnoea lasting about a half minute and disturbed rhythm with
diminished frequency of respiration from which it recovered in about
23 minutes.

It will be observed, therefore, that the behavior of the respiratory
mechanism towards chenopodium presents important differences from
that of the circulation. While the latter frequently showed a tendency
towards lessened response after one or two injections of moderate
amounts have been made, as already pointed out, the reaction to che-
nopodium finally disappearing altogether even after large doses have
been introduced, respiration, on the contrary, became increasingly
slower until it was arrested altogether after the administration of
amounts that were much less effective in the earlier stages of the ex-
periment. Blood pressure and cardiac action, which have nearly al-
ways been fairly good at the time oil of chenopodium ceased to be
effective for the circulation, continued with little or no change for a
considerable period of time, while respiration manifested signs of in-
creased disturbance. In some experiments the frequency of the heart
beat was still well within the normal rate at this time, the blood pres-
sure being 50 mm. Hg and in some experiments reaching a height of 80
to 90 mm. Hg for about 2} minutes after respiration had ceased (Dog
219, fig. 4).

Prolonged periods of apnoea were also observed in the final stages

78 WILLIAM SALANT AND A. E. LIVINGSTON

of most of our experiments, usually a short time after the last dose of
oil of chenopodium. In one experiment apnoea set in 23 minutes after
the final dose and lasted 3} minutes. The heart stopped at the end

Fig. 8. Dog 220. Apnoea (upper curve). Cardiae paralysis (lower curve)
and dyspnoea. Middle curve shows kidney volume. Time, 5 seconds (lowest

line X

of this period, apnoea being succeeded by well marked dyspnoea which
lasted-2 minutes. In another experiment (fig. 8) 6 minutes after the
final dose, which was 0.08 ec. per kilo in this case, several periods of
apnoea varying between 20 to 80 seconds were observed, blood pressure

during this time was 55 mm. Hg. As blood pressure fell apnoea was

~]
(den)
we)

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM

at once followed by dyspnoea and death. This sequence of events,
prolonged apnoea, cardiac paralysis, and dyspnoea was observed in
most experiments on dogs.

INANANARAAA

y v

Fig. 9. Cat 287. Intravenous injection of 0.04 cc. chenopodium per kilo.
Moderate fall of blood pressure with marked depression of respiration.

EXPERIMENTS ON CATS

The action of the oil of chenopodium when given intravenously was
somewhat different in cats. Its effect on respiration was more marked
than in dogs. The initial injection of 0.02 to 0.04 cc. per kilo (fig.
9), produced a fall of blood pressure of 6 to 8 per cent, while respi-
ration was depressed to a marked degree, the amplitude, as well as the

80 WILLIAM SALANT AND A. E. LIVINGSTON

rate having been decreased very considerably. A second injection pro-
duced a greater effect. Blood pressure fell nearly 10 per cent after
smaller doses and 25 per cent after larger doses. Thoracic respiration
ceased immediately after injection; abdominal respiration continued,
but became much slower. The fall of blood pressure after the initial
injection of a dose of 0.06 cc. per kilo was 32 per cent in one case. In
another experiment in which 0.085 cc. per kilo was given blood pressure
fell 45 per cent.

The increasing effects of oil of chenopodium on the circulation and
respiration observed after repeated doses (figs. 10 and 12) failed to
manifest themselves, however, in some cases. As shown in Cat No.
320, respiration was not affected either by the sixth or seventh injec-
tion, while the blood pressure fell 10 mm. or 11 per cent as a result of
‘the seventh injection.

Cat 287. Weight, 1600 grams. Ether anesthesia

First injection at 2.07 p.m. of 3.2 cc. (0.04 ce. per kilo), 2 per cent emulsion of
oil of chenopodium. Injected in 30 seconds. Blood pressure fell from 125 mm. to
115 mm. Hg, or 8 per cent. Recovered in 5 minutes. Abdominal respiration de-
creased in amplitude fully one-third, while thoracic respiration also became very
superficial. The rate decreased from 20 to 6 per minute. Considerable improve-
ment took place during the next 5 minutes, but recovery was incomplete at this
time.

Second injection at 2.12p.m. 3cc. of 2 per cent emulsion of oil of chenopodium
were injected in 60 seconds. Blood pressure fell from 120 to 90 mm. Hg, or 25 per
cent, and remained at this level for about 2} minutes. The rate of cardiac ac-
tion was also appreciably decreased. Recovered 7 minutes after injection. Tho-
racic respiration stopped immediately after injection and began to recover slowly
two minutes later. Abdominal respiration became slower after injection but
improvement began within one and a half minutes.

Third injection at 2.19 p.m. Blood pressure 120mm. Hg. Three cc. of emul-
sion of oil of chenopodium were injected in 40 seconds. Blood pressure equals
95 mm. Hg. Frequency of respiration decreased 50 per cent, thoracic respira-
tion disappeared. Abdominal respiration markedly decreased. Note cumula-
tive effect on respiration.

Cat 300, female. Weight, 3080 grams. Well fed

December 4, 1914. Ether anesthesia.

1.45 p.m. Received 18 ec. 5 per cent cocoanut oil emulsion in acacia in 2 min-
utes. Slight rise of blood pressure.

1.58 p.m. Central end of left vagus stimulated for 5 seconds, the right vagus
being intact. Distance of P. from 8S. coil 8 em. Marked fall of blood pressure
and slowing of the heart. Respiration stimulated.

ye Soyn
I0}jB SoyNuUIUI EZ “ ‘ Omeol oun : ‘O[Ty Jad 99 90'0 = “OT “37
: UBD JO UOTZDOTUT 9Y4 104Jv U no11d pue Oo UO YDOJJO BSUIMOYQ “OOF

thi TenanyyTen Teareneneresyerre
oer 4

ei Saas

HOODOO FG oe

ee ed VA Go aw

momo a 19 hy 226

Oey Bin oe

monger 9 Se 79 Ke a2

mem pepo 3. ~% 7] Fee 20 t/

OIL OF CHENOPODIUM

OF

ia Vv Ore (9

por omen.

Via Pia ede

Zz
iS
&
i)
a
ml
oO
Ll
Oo
io)
4
©
7)
<<
_—
=
om
<
a8}
Ay

gry Poy Sil
f " \ W ' wy Saint

\ i \ | \| \|
I) \y \| | WW WW WA \ | eae | \\ |
NUN A A \\ Wy yt | i oh | | \

Pee |

82 WILLIAM SALANT AND A. E. LIVINGSTON

2.07 p.m. Central end of left vagus stimulated. Distance P. from S. coil 16
em. Slight fall of blood pressure. Respiration slowed.

2.08 p.m. 9 ce. 2 per cent oil of chenopodium given in 5 per cent emulsion
of acacia. Blood pressure fell from 185 to 125 mm., or 32 per cent. Respira-
tion stopped for about 20 seconds, then returned, but was much slower than
before. Amplitude of thoracic respiration was much more depressed than ab-
dominal.

2.12 p.m. Blood pressure recovered. Central end of vagus stimulated 5
seconds. Distance of P. from S. coil8 em. No effect on blood pressure but res-
piration distinctly stimulated.

2.16 p.m. 12 cc. 2 per cent emulsion oil of chenopodium given in 5 per cent
acacia and cocoanut oil. Blood pressure fell from 190 to 110 mm., or about 42
percent. Respiration stopped for about 20 seconds then returned but was much
slower. Thoracic respiration very feeble, abdominal respiration slower than
before injection but depth not markedly less than before. While blood pres-
sure remained low, 18 cc. of an emulsion of 5 per cent cocoanut oil in 5 per
cent acacia were injected at the rate of 12 cc. per minute. Blood pressure
rose gradually from 105 to 140 mm. in 40 seconds but there was no noticea-
ble improvement in respiration. Respiration, on the contrary, became much
slower. Almost immediately after injection of cocoanut oil emulsion, blood
pressure fell about 40 per cent; respiration became irregular, then slow.
Another injection of 12 cc. 5 per cent cocoanut oil emulsion was made in 70 sec-
onds; blood pressure steadily rose from 60 mm. to 150 mm. Hg in half a minute.
Respiration became deeper and more frequent toward the end of the injection,
but this lasted half a minute and stopped altogether.

2.39 p.m., or 16 minutes later, the injection of 9 cc. of 2 per cent oil of cheno-
podium in 20 seconds caused a sudden drop of blood pressure from 170 to 70 mm.,
respiration becoming slow.

2.41 p.m. No improvement in blood pressure. Respiration became slower
than before. Eighteen cc. cocoanut oil were injected; blood pressure became
still slower, respiration stimulated at first, then stopped. Animal dead.

Cat 286. Weight, 2740 grams

1.43 p.m. 12 cc. (0.087 cc. per kilo) emulsion of 2 per cent oil of chenopodium
injected in 70 seconds. Blood pressure fell from145 to 80 mm. Hg, or 45 per cent,
by the time the injection was completed. Two minutes later no change in blood
pressure. Respiration much slower, thoracic respiration became very super-
ficial, abdominal respiration deeper, but much slower than before injection.
Ten minutes later blood pressure was 130 mm. Hg. Respirationwas much im-
proved but did not return to normal.

Second injection was made of 12 cc. 2 per cent emulsion of oil of chenopodium.
Blood pressure fell from 130 to 60 mm. Hg by the end of the injection, 100 seconds.
Effect on respiration was very marked; abdominal respiration less depressed than
thoracic. Forty-five seconds after injection respiration stopped, blood pressure

was but a few mm. Hg, heart continued to beat 40 seconds after respiration
stopped. ;

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 83

‘Cat 320. Weight, 3.3 kilos

3.15 p.m. Injected 7 cc. 1 per cent oil of chenopodium in 50 seconds. Blood
pressure before injection was 160 mm. after injection 150 mm. Respiration be-
came much slower, amplitude was decreased.

3.19 p.m. Blood pressure was 160mm. Hg. Injected 7 cc. oil of chenopodium
in 53 seconds. Blood pressure was 145. Blood pressure began to rise at once;
at 3.23 p.m. the blood pressure was 160 mm. Hg. Injected 15 cc. 1 per cent oil of
chenopodium in 5 minutes.

3.28 p.m. Blood pressure was 140 mm. Hg. Recovered in 3 minutes.

3.31 p.m. Blood pressure was 160 mm. Hg.

3.33 p.m. Injected 13cc. 1 per cent oilof chenopodium in 1} minutes. Blood
pressure before injection was 155mm. Hg. At the end of the injection blood pres-
sure was 125 mm. No change in 3 minutes, then rose gradually. Respiration
decreased in force and frequency.

3.49 p.m. Blood pressure was 145mm. Hg. 13 ec. 1 per cent oil of chenopo-
dium injected in 72 seconds. Blood pressure before injection was 145mm. Hg at
the end of the injection 100 mm. Hg, which continued without change 6 minutes
before recovery began. Respiration was much slower and weaker. Thoracic
respiration just perceptible; abdominal respiration regular and fairly strong,
but weaker than before injection.

4.15 p.m. Blood pressure 110mm. Hg. Injected 13 cc. 1 per cent oil of cheno-
podium in 3}; minutes. Blood pressure 80 mm. Hg. Respiration remained un-
changed.

4.28 p.m. Blood pressure 90 mm. Hg, injected 13 cc. 1 per cent oil of cheno-
podium in 1} minutes. Blood pressure 80mm. Hg. No change in respiration.

4.34p.m. Blood pressure 70mm. Hg. Injected 30 cc. 1 per cent oil of cheno-
podium in 3} minutes, At the end of injection blood pressure began to decline
rapidly and reached zero in about one minute. Respiration stopped before heart
action was arrested. A few respiratory movements noticed-after heart stopped.

EXPERIMENTS ON RABBITS.

The circulatory changes produced by chenopodium as observed in
rabbits, under ether or urethane anesthesia, indicate that even a small
dose may produce considerable depression. Blood pressure fell 16 to
20 per cent after initial doses of 0.022 to 0.027 cc. oil of chenopodium
per kilo were injected intravenously (fig. 13). The rate of injection
was about 7 cc. per minute, in some experiments, while the time occu-
pied in other experiments was about 5 seconds for the introduction
of the entire amount. Te fall of blood pressure was usually prompt,
but it began to rise almost immediately, often within 30 seconds from
the time the lowest level was attained. Recovery was accomplished
in some experiments in 3 to 5 minutes, in others the process occupied
more than double this time. Larger doses may be more active though
this was not always the case. In two experiments with 0.04 and 0.08
ec. oil of chenopodium per kilo the effect of the initial injection did not

LIVINGSTON

E.

A.

AND

SALANT

Ss

WILLIAM

84

‘apngiyjdwme A10jvaldser Jo osvaloop 9}v19poul
puv ainsseaid poo[q jo [[BJ peyavur smoys o[Ty 10d wntpodousyo
‘09 SO'O JO UoIQooluUT SNOUSARIZUT OUT, “ZZZT MAQVY CPL Sly

GLa

y IRAE om gore pyre 10-7 mys
fe Oy gs EE

Dose penn gunn! gre a2 001 ar agousyrry marine? % o/

porvees b= rn famed ming a do 70 L/ w@PL

-said poo[q Jo [vy Zurmoyg

‘oly ted umtpodousyo
‘09 ZOO 109Je UoryeIIdser JO UOTYR[NUIIys puR soins

“SOLT FIGQqQBY

"eL SI]

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 85

differ from that of smaller doses. But this is in all probability excep-
tional as the results were different when the experiments were repeated
in other rabbits. The initial injection of 0.08 cc. of oil of chenopodium
per kilo (fig. 14) produced a fall of blood pressure amounting to 45 mm.
Hg or 39 per cent in one experiment. The relative effect of subsequent
injections did not differ appreciably from that produced by the initial
dose. This was observed after experiments with small as well as with
larger doses, and may be readily seen from the following abbreviated
protocols.

Rabbit 1772. Gray female. Weight 3.19 kilos.
March 8, 1915. Received 6 grams urethane by mouth through stomach tube.

EMULSION BLOOD PRESSURE | FALL OF Bees
OF 1 PER PRESSURE
D DURATION OF
Cen SRENS INJECTION
OPODIUM Betore After . é

INJECTED | injection |. injection

1.08 | 25.0 115 7 45 39 | 3 min. 20 sec.
1.41 25.0 90 58 32 35 4min. 5 sec.
9.19» © "oe: 25:0 65 35 30 46 | 2min. 40 sec.
2.54 20.0 | 2 min. 50 sec.

Norte: Blood pressure gradually fell and rabbit died at 3.01 p.m.

Rabbit 1763. Gray male. Weight 2.3 kilos.
March 3, 10.30. Received’ by mouth through stomach tube, 2.5 grams
urethane.

95 75 20 21 | 20 sec.

12.01 4.6
12.15 5.0 105 80 25 24 | 36sec.
12.30 5.0 102 85 17 16 43 sec.
12.40 5.0 100 85 15 15 90 sec.
12.46 | 10.0 100 70 30 30 70 sec.

Rabbit 1718. Weight 2.2 kilos. Ether anesthesia.
6 cc. control solution injected—no effect.

I
EMULSION - : FALL OF BLOOD °
melt ceaes BLOOD PRESSURE aeeauain | dt ete
CENT CHEN- a= Pee
OPODIUM Before After H Pesteank NLRs WEES
INJECTED | injection | injection | ™™: 7S Se
ce.
d5o 6.0 118 95 23 20 55 sec.
UE: 6.0 110 84 26 23 .6
2.19 3.0 125 96 29 23-++ 1 min.
2.30 5.0 106 94 12 11+ 1 min.
2.48 6.0 85 65 20 23-++ 75 sec.

86 WILLIAM SALANT AND A. E. LIVINGSTON

A progressive increase in the fall of blood pressure was observed,
however, in some experiments after chenopodium. A second injec-
tion of the same amount as the first was twice as effective, while a third

injection was approximately three times as active as the initial dose,

the amounts introduced each time being the same.

The absolute and

relative fall of blood pressure are shown in the following table:

1696 1660 1702 | 1661
0.08 cc. 0.02 cc. 0.027 cc. 0.04 ce.
1 20mm. or 23) 25 mm. or 18) 15 mm: vor 16) (222) mms ore
per cent per cent per cent per cent
2 32 mm. or 46 | 45 mm. or 37 | 30 mm. or 36 | 28 mm. or 28
per cent per cent per cent per cent
3 50 mm. or 50 | 40 mm. or 50} 35 mm. or 35
per cent per cent per cent
Remarks | Died after re- | Died after re- | Died after re- | Died after re-

ceiving 0.16
per kilo.

ceiving 0.083
per kilo.

ceiving 0.08
ec. per kilo.

ceiving 0.24
ec. per kilo.

It may be remarked that the fatal dose in the experiments presented
in this table was considerably smaller than in those rabbits in which
no cumulative effect on the circulation could be demonstrated. The
acute fatal dose of chenopodium when administered intravenously var-
ied in our experiments between 0.3 and 0.35 cc. per kilo and in one in-
stance it was 0.48 cc. per kilo. The results obtained in Rabbit 1661
in which the effect of repeated injections were much less marked than
in the other three rabbits given in the table suggest that the resistance
to chenopodium might be a factor in determining cumulation. It
may be of interest in this connection to call attention to the condition
of the heart in Rabbit 1660. Post mortem examination indicated the
presence of serious damage to the myocardium, brought about possibly
by repeated bleeding extending over a period of several weeks which
were made by thrusting a hypodermic needle into the organ. Although
this case does not afford sufficient evidence for any positive statement
regarding the relation of the condition of the heart to circulatory dis-
turbance caused by chenopodium, it is nevertheless suggestive and
may perhaps account for similar effects observed in other experiments.

As shown in the experiments cited below, cumulation was likewise
indicated when larger doses of chenopodium were injected after con-
siderable quantities had previously been introduced into the circulation.
Perhaps also in these cases the increased action was due to damage of
the heart caused by the previous treatment with chenopodium.

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 87

Rabbit 1709. Weight, 2275 grams

No previous injection.
35 cc. 1 per cent chenopodium injected in 8 minutes.
Blood pressure fell from 100 to 80 mm. Hg, or 20 per cent.

Rabbit 1681. White male. Weight, 2250 grams.

Total quantity previously injected 0.25 ce.

Seven minutes later 22.5 cc. 1 per cent chenopodium injected in 15 minutes.

Blood pressure fell from 88 mm. before injection to 40 mm. or 54 per cent
when injection was discontinued.

Rabbit 1718. Weight, 2.2 kilos.

Total quantity of chenopodium injected 0.4 cc, or 0.18 ce. per kilo; 5 minutes
after last injection 22.5 cc. 1 per cent emulsion of oil of chenopodium injected in
83 minutes.

Blood pressure fell from 85 mm. to 40 mm. Hg, or 52 per cent.

The action of chenopodium on respiration was in several respects
different from its effects on the circulation. The initial injection of
small and moderate doses, the duration of which varied from 5 to
60 seconds, did not produce any respiratory changes in most experi-
ments. In some eases slight, in others transitory but well marked
stimulation was observed, even when the initial dose was large (fig. 13).
A dose of 0.08 ec. per kilo produced in one rabbit (1772, fig. 14) a mod-
erate decrease of amplitude, while in another it caused a slight depres-
sion only. The rate of injection in this case was 4.5 cc. 1 per cent
emulsion of oil of chenopodium per minute. Subsequent injections of
large doses were much more active, but the case was different when
smaller amounts were administered. A second or third injection of a
small dose which was not previously effective sometimes produced
respiratory depression but this was not always the case. In one experi-
ment a number of injections were made without affecting respiration;
only after the fifth injection when 0.05 cc. per kilo at one dose was
given the total amount previously received being 0.1 ec. per kilo, was
any effect noticed in this experiment, thoracic respiration being marked-
ly depressed but abdominal respiration remained unchanged. Similar
results were obtained in another experiment (1681). In this case two
doses, each of 0.022 cc. per kilo, produced a slight increase of ampli-
tude. In the third the amplitude of thoracic respiration decreased 33
per cent after a dose of 0.066 cc. per kilo. No change in abdominal
respiration was observed. Another injection of 0.1 ec. per kilo begun
7 minutes later and carried out in 15 minutes, produced a still greater

88 WILLIAM SALANT AND A. E. LIVINGSTON

effect on thoracic respiration, the amplitude being decreased 65 per
cent, but abdominal respiration remained unchanged. In neither test
was the rate affected.

Although the amount required to depress respiration varied a good
deal in different individuals, it may be safely stated that when a suf-
ficient amount was introduced into the circulation, respiration became
decidedly weaker, amplitude, as well as the rate being decreased. It
may be pointed out, however, that the frequency of respiration was
but moderately depressed. This was also observed in rabbits in which
stimulation occurred at first. Although respiration seemed to be less
affected by chenopodium than the circulation, arrest of respiration usu-
ally took place either at the same time or before heart action ceased.
In several experiments in which the heart was exposed it was found
beating feebly and continued some time in this condition—15 minutes
to one hour.

DISCUSSION.

A survey of the results obtained in the foregoing experiments shows
that the effect of oil of chenopodium on different animals is essentially
the same notwithstanding the occasional departure from the general
type of action and the marked quantitative differences produced in the
same animals under changed conditions of experimentation. This was
illustrated by the characteristic action on the circulation and respira-
tion and also by the resistance to chenopodium as indicated by the
amounts tolerated, which varied in rabbits between 0.15 ce. and 0.48
ce per kilo, being in most cases between 0.3 cc. and 0.35 ce. While the
[largest amounts in dogs were 0.47 cc., the smallest was 0.3 per ec. per
kilo. Cats suecumbed when the total quantities reached 0.2 to 0.36 ce.
per kilo. The average is, therefore, approximately the same in all the
animals examined. Intravenous injections were usually followed by
circulatory and respiratory disturbances. The initial dose produced a
fall of blood pressure in dogs, cats, and rabbits but the action when
this was repeated was greater in the dog and rabbit than in the cat.
The effect on dogs was markedly enhanced, the fall of blood pressure
being several times greater when chloretone anesthesia was substituted
for morphine and ether. The action in different animals varied still
more when the injections were repeated. Thus the failure of the cir-
culation to react after a large amount of chenopodium has been intro-
duced in divided dose, which was frequently observed in dogs, was
seldom seen in the cat and was absent in rabbits.

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 89

That chenopodium is a respiratory depressant was abundantly shown
in our experiments but the behavior of different animals in this re-
spect likewise presented interesting differences. Respiratory depres-
sion after the same amounts in proportion to body weight was more
marked in cats than in dogs. In rabbits, it will be recalled, small
doses may produce stimulation at first, depression having been noticed
only after the dose was repeated several times. Besides, apnoea was
seldom observed in rabbits even after large doses. Bruning’s (6) ex-
periments on the behavior of the red blood cells toward chenopodium
are of interest in this connection since he found that hemolysis and
methemoglobin may be produced in vitro. It is conceivable that dis-
turbance of respiration may be caused by the hemolytic action and the
formation of methemoglobin. We have, therefore, carried out a num-
ber of observations with the oil of chenopodium to determine whether
methemoglobin and hemolysis may be produced in vivo. Doses of
0.02 cc. to 0.24 cc. per kilo were injected intravenously. A spectro-
scopic examination of the blood obtained from animals thus treated
failed to indicate the presence of methemoglobin. Hemolysis was
also absent even after large amounts were administered. This was
studied in cats after the introduction of two grams per kilo of the oil
into the stomach or small intestine, and in rabbits which received 0.1 to
0.16 ce. per kilo intravenously. A moderate degree of hemolytic ac-
tion was present in dogs’ blood taken after oil of chenopodium was in-
jected intravenously and kept at low temperature for 24 hours, but this
was about the same in degree which may be observed when blood is
obtained from anesthetized dogs that have not received oil of cheno-
podium. Depression of the circulation may also be thought of as a
factor in determining the respiratory changes observed, but the results
obtained in a large proportion of our experiments do not support this
view. As shown in figure 2, blood pressure fell about 40 per cent but
respiration was only slightly depressed. On the other hand, respira-
tory depression was observed when the circulation no longer reacted
to chenopodium. Moreover, in experiments on rabbits a fall of blood
pressure and a coincident stimulation of respiration often occurred.

Paralysis or depression of the vagus may also be thought of in this
connection as a possible cause, but some of our observations with che-
nopodium were made on dogs and cats in which both vagi were cut.
As the characteristic effect on respiration was present in both these
animals it would seem that the action was central. It is evident that
no satisfactory explanation of the respiratory effect can be offered at

90 WILLIAM SALANT AND A. E. LIVINGSTON

present. The work of Loevenhart and Grove (7) is, however, sugges-
tive. After the intravenous injection of the salts of iodoso and iodoxy
benzoic acid, the effects on respiration which these investigators ob-
served were strikingly similar to those of chenopodium. They believed
that the apnoea was due to the liberation of oxygen in the blood by
these substances. As shown by Nelson (8) and later by Wallach (9),
ascaridole, which is the active principle in the oil of chenopodium, is a
peroxide. Although it does not oxidize guaiac, it is quite possible that
its action in vivo may be similar to that of the salts of benzoic acid.
The depression of respiration and apnoea may be due, therefore, to the
oxygen liberated by ascaridole. This explanation does not seem to ap-
ply to the action of oil of chenopodium in the rabbit, but may, indeed,
contradict it. That this is not the case, however, appears when the
relative amounts of carbon dioxide and oxygen in the blood of different
animals, which formed the subjects of our experiments, are taken into
consideration. According to Heinz (10), the carbon dioxide, as com-
pared with the oxygen, in the blood of the rabbit, is relatively greater
than in the dog or cat. Since carbon dioxide stimulates respiration,
according to Henderson (11) and others, its action would be antagonis-
tic to that of oil of chenopodium, thus tending to neutralize the effect
produced by the liberation of oxygen. The relatively greater amount
of carbon dioxide in the blood of the rabbit would, therefore, prevent
the onset of respiratory depression and apnoea. This may also explain
the difference in the final stages of intoxication. The simultaneous
failure of the circulation and respiration frequently observed in cheno-
podium poisoning in the rabbit is probably due to direct action on the
heart, since respiration is impaired but not abolished. Cardiac asphyxia
(Henderson, |. c.), such as would occur in the dog or cat on account of
prolonged apnoea is not probable in the rabbit, there being enough oxy-
gen to sustain heart action while the respiration goes on, though very
much diminished. Finally the anesthesia produced by chenopodium
may also be considered as a factor in the production of apnoea and respir-
atory depression. According to Henderson the respiratory center is
much less sensitive to carbon dioxide in profound anesthesia, but the
evidence obtained does not favor this explanation as in profound anes-
thesia, such as was produced by chloretone and alcohol, the adminis-
tration of chenopodium, according to this explanation, should have
produced apnoea much more readily than when the same was given in
morphine-ether anesthesia.

PHARMACOLOGICAL ACTION OF OIL OF CHENOPODIUM 91

SUMMARY

1. The intravenous injection of doses of 0.02 to 0.085 cc. of cheno-
podium per kilo produced a fall of blood pressure in dogs, cats and
rabbits. Recovery was observed.

2. The effect was greater in dogs than in rabbits or cats.

3. A second injection of the same dose produced a greater effect,
but when this injection was repeated until the total amount reached
about 0.2 ec. per kilo, no response of the circulation could be observed.
This was especially the case in dogs, but to a much smaller extent in
cats. This phenomenon was absent in rabbits.

4. Fall of blood pressure was of cardiac origin, as the volume of the
kidney decreased with the fall of blood pressure.

5. Frequency of heart action was diminished after oil of chenopodium.

6. A very marked decrease of vagus irritability was observed after
oil of chenopodium.

7. Respiratory depression such as decreased amplitude and rate, with
apnoea, was also caused by chenopodium, but the effect with small
doses was less constant than on the circulation.

8. Cats react more readily than dogs.

9. Small doses may stimulate respiration in rabbits. Apnoea was
very seldom observed in the rabbit, even after large doses.

10. No methemoglobin or hemolysis was observed even after the in-
travenous injection of 0.02 to 0.024 cc. per kilo, or the introduction of
two grams per kilo into the stomach or small intestine of the cat.

11. Liberation of oxygen in the body by ascaridole is suggested as a
possible cause of respiratory depression and apnoea.

12. Action of chenopodium on respiration is independent of its effect
on the circulation.

13. Reduction of sensitiveness of respiratory center to carbon di-
oxide is not the cause of action of chenopodium on respiration.

14. Amounts of chenopodium tolerated by intravenous injection var-
ied in the same animals. The average is approximately 0.03 to 0.35 ce.
per kilo in dog, cat and rabbit.

15. The less depressant action of chenopodium on respiration in the
rabbit is attributed to relatively larger amounts of carbon dioxide in
the blood.

WILLIAM SALANT AND A. E. LIVINGSTON

BIBLIOGRAPHY

SaLant: Journal of Pharmacology and Experimental Therapeutics, vol. 2,
p. 391, 1911.

SALANT AND Newtson: American Journal of Physiology, vol. 36, p. 440,
1915.

SOLLMANN AND PitcHEeR: Journal of Pharmacology and Experimental
Therapeutics, vol. 3, p. 19, 1911.

BLACKFORD AND SATTFORD: Medical Record, vol. 84, p. 378, 1913.

Roru: Bulletin Hygienic Laboratory No. 100. p. 5.

BruninG: Zeitschrift fiir Experimentelle Pathologie und Therapie, vol. 3,
p. 564, 1906.

LOEVENHART AND Grove: Journal of Pharmacology and Experimental
Therapeutics, vol. 3, p. 101, 1911.

Newtson: Journal of the American Chemical Society, vol. 33, p. 1404, 1911;
vol. 135, p. 84, 1913.

WauutacH: Annalen der Chemie, vol. 393, p. 59, 1912.

Hertnz: Handbuch der Experimentellen Pathologie und Pharmakologie,
vol. 2, part 1, p. 384, 1906.

HENDERSON: American Journal of Physiology, vol. 25, p. 310, 1910.

INFLUENCE OF DIIODOTYROSINE AND IODOTHYRINE ON
THE SECRETION OF CEREBROSPINAL FLUID

CHARLES H. FRAZIER anp MAX MINOR PEET

From the Department of Surgical Research, University of Pennsylvania

Received for publication April 26, 1915

In a previous communication we have shown that thyroid extract
slows the rate of secretion of cerebrospinal fluid. The present work
was undertaken to discover, if possible, what constituent of the thyroid
gland possesses the power of inhibiting the cerebrospinal fluid secre-
tion. In this paper we present the results of experiments with iodothy-
rine, a commercial derivative of the thyroid gland, and of diiodotyro-
sine, a synthetic substance closely related to the iodine complex of the
thyroid gland. The diiodotyrosine was prepared and kindly furnished
for these experiments by Dr. Treat B. Johnson of Yale University.

The anesthesia, rate of secretion of cerebrospinal fluid, and the blood
pressure records were obtained with the technique described in a pre-
vious paper (1).

Diiodotyrosine. The injection of solutions of this substance into the
femoral vein causes no change in blood pressure or in respirations, so
that all changes in the outflow of cerebrospinal fluid which result from
the injection may be considered as due to the effect of the substance on
the secretion of the choroid plexus.

The results of four experiments with diiodotyrosine are given in the
table on next page.

Experiment I. The injection of 0.1 gm. of diiodotyrosine intra-
venously causes a marked decrease in the flow of cerebrospinal fluid.
This is pronounced even in the first half hour after injection, the flow
dropping from the normal 3.78 cc. to 1.28 cc., a decrease in rate of se-
cretion of 0.0833 cc. per minute. The flow of cerebrospinal fluid gradu-
ally grows less to the end of the experiment, two and one-half hours
after the injection, although slightly more fluid escapes in the last period
than in the two preceding it. The average flow for the five 30-minute
periods after injection is only 0.8 ce.,a decrease from the normal flow of
2.98 cc. The normal rate of secretion is 0.126 cc. per minute, while

93

94 CHARLES H. FRAZIER AND MAX MINOR PEET

the average rate for the five periods is 0.026 cc. per minute, a decrease
in rate of secretion of cerebrospinal fluid after the injection of 0.1 gm.
of diiodotyrosine of 0.1 ec. per minute.

Experiment II. The injection of 0.2 gm. of diiodotyrosine in this
instance does not show as pronounced an effect on the amount of secre-
tion in the first half hour as was noted with 0.1 gm. in Experiment I.
The flow is lessened, however, and steadily decreases after the second
period. The flow for the first half hour is 1.28 ec. against 1.59 ee. for
the normal period. This is a decrease in rate of secretion from 0.053
ec. per minute to 0.042 cc. per minute. The flow for the second period

Diiodotyrosine

FLOW IN CC. OF C. 3. F. FOR 30-MINUTE PERIODS

PERIODS en
Experiment I. |Experiment I1.|Experiment III.|Experiment LY.
0.1 gm. injec. | 0.2 gm. injec. | 0.2 gm. injec. | 0.1 gm. injec.

Before injection............ 3.78 1.59 2.82 1.62
After myection 1). 4ee- ar 1.28 1.28 1.63 1.28
Di: P53 FORT 0.99 1.39 0.66 1.29
BE aye bets fete 2 0.56 1.23 0.77 1.23
Ae it mane dae 0.51 0.88 0.78 0.88
Djs tek 0.67 0.47 0.28 0.47
oie ser setae as 2 0.11 0.11
Vion i eaeentee 0.26 0.26
Average flow for 30 minutes. 0.80 0.80 0.82 0.78
Decrease from normal....... 2.98 0.79 2.00 0.84

is larger than that for the first, but decreases below the first in the third
period.

The flow then rapidly lessens till in the sixth period only 0.11 ce. is
secreted an average rate of 0.003 cc. per minute. The flow in the
seventh period is somewhat more rapid than in the preceding 30 min-
utes, but still has a rate of only 0.008 cc. per minute. The average
flow for 30 minutes for the seven periods (three and one-half hours) is
0.8 cc. a decrease of 0.79 ce. from the normal flow. This is a decrease
in secretory rate of 0.027 cc. per minute. .

Experiment IIT. The injection of 0.2 gm. of diiodotyrosine in an-
other dog gives more pronounced results than in Experiment II. In
this experiment the reduction in flow is pronounced in the first half
hour with a further marked drop in the second period, while in the
third and fourth the flow is practically uniform. The decrease in

SECRETION OF CEREBROSPINAL FLUID 95

flow for the first period is 1.19 cc., a reduction of 0.04 ec. per minute
in rate of outflow. The second period shows a decrease from the first
period of 0.97 cc. or a reduction in rate of outflow of 0.032 cc. per min-
ute. The third and fourth periods have a slightly greater flow than
the second, 0.77 cc. and 0.78 cc. respectively against 0.66 cc. for the
second. The fifth period, however, is much slower, only 0.28 cc. escap-
ing in the 30 minutes. The average 30-minute flow for the five periods
is 0.82 cc., a decrease from the normal of 2 cc. or a reduction in rate of
0.067 cc. per minute.

Experiment IV. The injection of 0.1 gm. of diiodotyrosine in this
experiment gives less striking results than in Experiment I. In fact
the results are very similar to those of Experiment II in which twice
the amount was injected. The decrease in flow is noticeable in the
first half-hour and remains practically the same for the first hour and
a half after injection. It then rapidly decreases for one and a half
hours, with a slight increase in the last half-hour. The flow in the first
period decreases 0.34 cc. from the normal, or a decrease in rate of se-
cretion of 0.011 cc. per minute. The flow is then practically unchanged
until the fourth period when 0.88 cc. escapes in 30 minutes. The fifth
and sixth periods show reductions to 0.47 cc. and 0.11 cc. respectively.
The seventh or final period shows a slight increase to 0.26 cc. The
average flow for 30 minutes for the seven periods is 0.78 cc., a decrease
from the normal amount of 0.84 cc. or a decrease in rate of secretion
of 0.028 cc. per minute.

It is evident from the above experiments that diiodotyrosine in solu-
tion when injected intravenously has an inhibitory action on the se-
cretion of the choroid plexus very similar to that of saline extracts of
the thyroid gland. The results on different dogs are not uniform, but
seem to depend somewhat upon the normal rate of cerebrospinal fluid
secretion.

Iodothyrine. The intravenous injections of iodothyrine solution pro-
duce no changes in blood pressure or respiration. The results of four
injections are given in the table on next page.

Experiment I. The injection of 0.3 gm. iodothyrine gives a decrease
of 0.32 cc. cerebrospinal fluid for the first half hour, a decrease in rate
of 0.11 cc. per minute. The flow shows a steady reduction which in
the fifth and sixth periods has decreased to 0.4 cc., a decrease from
normal of 0.95 cc. The average 30-minute flow for the six periods is
0.66 cc. a drop from normal of 0.69 ce. or a decrease of 0.023 ec. per
minute in rate of secretion.

96 CHARLES H. FRAZIER AND MAX MINOR PEET

Experiment II. The injection of 0.05 gm. iodothyrine has very little
influence on the flow of cerebrospinal fluid.

The flow in the half hour before injection is 1.41 ec. and in the first
period after injection is 1.10 cc., a reduction of 0.31 cc. or a decrease
in rate of 0.11 ec. per minute. However, in the second period the flow
returns to normal. In the succeeding two periods it again decreases,
the average 30-minute flow for the four periods after injection is 1.15
ec., a decrease of only 0.26 ce. or in rate of only 0.009 ce. per minute.
This is almost within the range of the normal decrease.

Experiment III. In this experiment 0.5 gm. iodothyrine gives very
little change in the first half hour, the decrease only amounting to 0.09

Todothyrine

FLOW IN CC. OF C. 8. F. FOR 30-MINUTE PERIODS

PERIODS MbpP itn = MitinL= ais i:
Experiment I. |Experiment II].|/Experiment 1II. Experiment IV.

0.3 gm. injec. | 0.05 gm. injec. | 0.5 gm. injec. | 0.3 gm. injec.

Before injection............ 1.35 1.41 1.74 2.97
After injection 1............ 1.03 LO 1.65 1.41

Ds ys CUBE Yon: 0.92 1.40 ea 7s 0.57

BREN Red 0.73 1323 0.63 0.11

7 ines el A 0.53 0.90

Bee cued Alene 0.40

6 0.40
Average flow for 30 minutes 0.66 1.15 1.15 0.69
Decrease from normal... ... 0.69 0.26 0.59 2.28

* 1.5 gm. injected.

ce., a second injection (1.5 gm.) is then given and the next period shows
a more marked reduction. This response reduced the flow 0.57 ce.
below the normal. The third period shows a further reduction from
normal of 1.11 ec. Owing to an unavoidable accident the experiment
was terminated at this point. The average 30-minute flow for the
periods after injection is 1.15 ec., a decrease of 0.59 ec. or a drop from
the normal rate of 0.058 cc. per minute to 0.038 ee. per minute.
Experiment IV. The injection of 0.3 gm. iodothyrine in another dog
gives a more marked decrease than in any of the three previous experi-
ments. The reduction in the first half hour after injection is pro-
nounced, slightly less than half the normal amount of fluid flowing
from the cannula. The actual decrease in this period is 1.56 ec. The
flow in the third period is slightly larger than in the second, while in

SECRETION OF CEREBROSPINAL FLUID 97

the third it is again greatly reduced. The average flow for a 30-minute
period after injection is 0.69 cc., a drop below the normal of 2.28 cc. or
a decrease in secretion rate from 0.099 cc. per minute to 0.023 cc. per
minute, a reduction of 0.076 ce. per minute.

The experiments with iodothyrine tend to show that small amounts,
such as 0.05 gm. have very little influence on the rate of secretion of
cerebrospinal fluid. Larger amounts result in a slower rate, but with
the exception of Experiment IV, the decrease is not as marked as that
produced by diiodotyrosine or saline extracts of thyroid gland.

It is evident from the experiments, both with diiodotyrosine and io-
dothyrine, that different dogs react very differently to the same amount
of substance. As a general rule the most marked responses are in those
animals which have normally a very rapid secretion. Sex and age
gave apparently no influence, while the weight of the animal seems to
make very little difference.

Control experiments have shown that the decrease in flow noted in
the above experiments is due to the action on the choroid plexus of the
substance injected. A decrease does occur normally, but this is so
slight as to be negligible in the two to four hours which are considered
in these experiments.

CONCLUSIONS

1. The intravenous injection of diiodotyrosine has an inhibitory in-
fluence on the rate of secretion of the choroid plexus. This decrease in
rate usually appears in the first half hour after injection, but is not
so marked as that obtained with saline extracts of fresh thyroid gland.

2. Iodothyrine in solution, injected intravenously, has little influence
on the rate of cerebrospinal fluid secretion when given in small amounts
(0.05 gm.). In amounts of 0.3 gm. and 0.5 gm. there is some inhibi-
tion of the rate of choroid plexus secretion, but not as marked as that
produced by diiodotyrosine or saline extracts of fresh thyroid.

THE RESPONSE OF THE VASODILATOR MECHANISM TO
WEAK, INTERMEDIATE, AND STRONG
SENSORY STIMULATION

E. G. MARTIN anp W. L. MENDENHALL
From the Laboratory of Physiology in the Harvard Medical School

Received for publication April 28, 1915

In the study of vasomotor reflexes the desirability has long been reec-
ognized of supplementing by means of plethysmographic studies of in-
dividual organs the information gained from observations of blood-
pressure changes. Recently Tschalussow! has described a method of
employing the nasal cavity as a plethysmograph, and has found it to
be of exceptional delicacy.

In certain studies from this laboratory the characteristic depressor
effect of weak sensory stimulation has been emphasized,? and it has
seemed to us worth while to supplement these studies with observa-
tions of the effects of weak sensory stimuli on the volume of the nasal
cavity.

METHOD

Our experiments were performed on cats. In some cases urethane
anesthesia was employed; in others rapid decerebration under ether
was performed. The method of preparing the nasal cavity was the
modification of Tschalussow’s original one described by one of us.®
Simultaneous blood-pressure records were made from a femoral artery.
For sensory stimulation electrodes (Sherrington type) were placed cen-
trally on branches of the brachial and sciatic trunks, and in some ex-
periments on the vagus or saphenous. Induction shocks of known in-
tensity (Z units) were used for excitation. The rate varied between 5
and 15 per second.

1 Tschalussow: Archiv fiir die gesammte Physiologie, cli, 19138, 524.
2 Martin and Lacey: This Journal, xxxiii, 1914, 212; Martin and Stiles: ibid
xxxiv, 1914, 106.
3 Mendenhall: This Journal, xxxvi, 1914, 58.
4 Martin: The measurement of induction shocks, New York, 1912, 73.
9S

RESPONSE OF VASODILATOR MECHANISM TO SENSORY STIMULI 99

The nasal cavity is peculiarly adapted for an analysis of depressor re-
flexes, since it receives both vasodilator and vasoconstrictor fibres,® and
can respond, therefore, to reflex influences acting either through the
vasoconstrictor or the vasodilator central mechanism.

With our method of recording nasal volume-change, as in the experi-
ments of Tschalussow (loc. cit), an up-stroke of the lever signifies
vasodilation, and a down stroke vasoconstriction. Changes in nasal
volume following stimulation of sensory nerves can be interpreted as
signifying particular vasomotor effects only when recorded in connec-
tion with simultaneous records of general blood pressure, for only thus
can passive changes be distinguished from active ones. Even when so
accompanied, the distinction cannot always be made with certainty.
To be sure, active nasal vasoconstriction occurring simultaneously with
general vasoconstriction is readily distinguished, since here we have a
down-stroke of the nasal recorder with upward movement of the blood-
pressure tracing; likewise active nasal vasodilation in the presence of
general vasodilation presents no difficulty since in this case there will
be an upward stroke of the nasal recorder with downward movement
of the blood-pressure tracing. The difficulty arises when we attempt
to determine whether local active vasomotor changes may occur in the
presence of general vasomotor changes of opposite sign. To deter-
mine whether the nasal vessels are actively dilating or are merely being
gorged passively by rising general blood pressure requires the most
careful analysis of the records.

For purposes of comparison we have found it desirable, as did Tscha-
lussow,® to bring about known passive changes in nasal circulation.
Means to this end are compression of the abdominal aorta, which brings
about passive engorgement of the vessels of the head; or compression
of the thorax, or peripheral stimulation of the cut vagus to cause the
blood supply to the head to be lessened. In our experience the nasal
cavity always responds immediately to these procedures. Tschalussow
found (loc. cit., 52), that the nasal vessels react to passive engorgement
by energetic local contraction. This we have seen in a few, but by no
means in all, of our experiments. As an index of passive engorgement,
therefore, we do not consider this after contraction wholly reliable.

Observations. In general the nasal volume changes brought about by
stimulation of sensory nerves correspond with expectation. That is

5 Langley: Philosophical Transactions, London, clxxxiii, 85; Dastre and Mo-

rat: Recherches sur le systeme nerveux vaso-moteur, Paris, 1884, 115 et seq.
6 Tschalussow: Loc. cit., 527. ,

100 E. G. MARTIN AND W. L. MENDENHALL

to say, weak stimuli, such as cause drop in general blood pressure, cause
simultaneously nasal vasodilation, and strong stimuli, which cause a
rise of pressure, produce nasal vasoconstriction. The nasal mucosa,
therefore shares actively in the reactions by which both pressure-fall
and pressure-rise come about.

22 Z 60 Z 86.5 Z 127 Z

Fig. 1. The transition from a depressor to a pressor response with increasing

strength of stimulation. The time signal is at the line of zero pressure. Time

in 30 second intervals. Stimulation strengths in Z units. Peroneal nerve
stimulated.

In the presence of a mechanism of this sort, in which weak stimula-
tion produces one sort of response and strong stimulation of precisely
the same nerve trunks a diametrically opposite response, a degree of
interest naturally centres about the point of reversal, for here, at the

RESPONSE OF VASODILATOR MECHANISM TO SENSORY STIMULI 101

point where one sort of influence relinquishes control of the mechanism
to another sort, we may hope to gain some insight into the nature and
mode of action of the influences themselves. The great sensitiveness
of the nasal plethysmograph lends itself to this pos-
sibility.

So far as we are aware the character of the blood-
pressure curve under sensory stimulation near the
reversal point has not hitherto been reported in
detail, although it has been observed in this labora-
tory from the beginning of our quantitative study
of vaso-motor reflexes, and has undoubtedly been
observed incidentally by other investigators. A typi-
cal example of the transition from a depressor to a
pressor reflex is presented in figure 1. As the strength
of sensory stimulation is increased above that which
brings about pure pressure-drop a tendency develops
toward the replacement, during the period of stimu-
lation, of the initial pressure-drop by a pressure-rise.
This tendency becomes more and more marked, un-
til, in most cases, with the application of sufficiently
strong stimulation, the initial depressor phase is
completely suppressed, leaving the rise of pressure
as the only visible response. In the particular ex-
periment illustrated in the figure the region of re-
versal occurred with somewhat weaker stimulations
than we have ordinarily found adequate. The aver- Ree
age of all our experiments places the “‘reversal thres- 60 Z
holds” in the neighborhood of 200-250 Z units as Fete! Active
compared with 60-90 in this experiment. be Sal eacadilation

The responses of the nasal plethysmograph to persisting through-
sensory stimulation in the region of “reversal” are out avasomotorre-
instructive. During the initial pressure drop there Seas ie
is active dilation of the nasal mucosa. As the de- fed treat te
pressor phase is succeeded by the pressor there is peyoneal nerve
often a corresponding change to active constriction stimulated.
of the nasal vessels. This latter reaction is not,
however, invariable. We have sometimes seen the active nasal dila-

tion persist throughout the pressor phase, particularly when the latter
' was not very pronounced. Such a case is illustrated in figure 2. In this
same experiment stronger stimulation brought about a typical pressor
reaction with the usual active nasal constriction.

102 E. G. MARTIN AND W. L. MENDENHALL

A further suggestive observation is that in a fair percentage of cases
of ordinary pressor stimulation active constriction of the nasal mucosa
does not begin at the instant o pressure-rise but several seconds
thereafter. The latent period for the nasal mucosa, where a well
marked one oécurs, averages about 15 seconds. Obviously the dis-
charge of the vasoconstrictor centre which brings about the pressure-
rise does not, in these cases, involve immediately the nasal mucosa.
A simple explanation would be that in this region, which is also under
the control of the vasodilator mechanism, active vasoconstriction is
held in abeyanee for a time by a simultaneous vasodilator discharge.

From all these observations we may draw the inference that depres-
sor influences and pressor influences can be aroused simultaneously by
suitable stimulation of individual sensory nerve trunks, and that the
resultant effect on the peripheral vasomotor mechanism depends on the
balance that is established between the opposing influences. To put
the point in other words, when strong sensory stimulation brings about
a rise in blood pressure, it may be that depressor influences are not
suppressed, but merely overpowered.

In any attempt to analyze the interaction of simultaneous pressor
and depressor influences we have to bear in mind the potentially i wo-
fold nature of each. _Pressor influences may act positively on the vaso-
constrictor centre or negatively on the vasodilator, either separately,
or, as Bayliss maintains,’ on both together. Likewise depressor influ-
ences may operate either through lowering constrictor tone or heighten-
ing dilator tone or both. A method of separating these factors used
by Bayliss (loc. cit.), and also by Fofanow and Tschalussow,* consists
in cutting either the constrictor or the dilator efferent paths to the or-
gan upon which plethysmographic studies are being made. By this
means observed vasomotor changes can be assigned definitely to one or
the other mechanism. We have applied this method in our experi-
ments with the nasal plethysmograph in order to determine whether
the depressor influences manifested during weak sensory stimulation
involve activity of the vasodilator mechanism. We find that marked
active nasal dilation is induced by weak stimulation when both cervi-
cal sympathetic nerves are severed. This procedure cuts the nasal
mucosa off from its connection with the vasoconstrictor centre while
leaving it in communication with the dilator mechanism through the

7 Bayliss: Proceedings of the Royal Society of London, 1908, Ixxx, B, 339.

8 Fofanow and Tschalussow: Archiv fiir die gesammte Physiologie, cli, 1913,
548.

RESPONSE OF VASODILATOR MECHANISM TO SENSORY STIMULI 103

Vidian nerves.‘ The observed dilation is the result, therefore, of posi-
tive discharges through the vasodilator system.

By the same procedure we have attempted to show whether the de-
pressor influences which we have pictured above as present but inef-
fective during strong sensory stimulation involve posit ve action of the
vasodilator mechanism. We have repeated the observations of Fofa-
now and Tschalussow on the effects on the nasal mucosa of stimulating
the vago-depressor trunk!® and confirm their finding that strong stimu-
lation of this trunk with both cervical sympathetics cut brings about
active reflex vasodilation, indicating a positive action upon the vaso-
dilator apparatus.

The attempt to demonstrate by this method activity of the vaso-
dilator mechanism during ordinary pressor stimulation is rendered dif-
ficult by the circumstance that when the cervical sympathetic nerves
are cut the nasal mucosa may respond passively to the general rise of
blood pressure, giving, as already pointed out. a tracing that is difficult
to distinguish from active vasodilation.

There are two features, however, in which active dilation may differ
from passive. The first is in time relationships, the second in degree
of response.

A passive nasal volume-change must necessarily synchronize accu-
rately with the blood-pressure rise or fall which produces it. The de-
gree of response, moreover, will show some relation to the amount of
general pressure-change, unless, indeed, the local reaction described by
Tschalussow occurs (loc. cit., 52), in which case the response is immedi-
ately identified thereby as a passive one. In an experiment in which
both cervical sympathetics are cut, a nasal dilation during pressor
stimulation, which definitely outlasts the general pressure rise; or a
series of dilations of equal extent in connection with pressor responses
of different amounts, indicate strongly, if they do not prove conclus-
ively, that the dilator mechanism is active during strong stimulation of
sensory nerves. Both these appearances we have encountered in the
course of our work, and they have been sufficiently clean cut to satisfy
us that they were not accidental. For example, in one experiment,
strong sensory stimulation (900 Z units) applied for 30 seconds, caused
a reflex rise of pressure which fatigued quickly; so that the pressure
had returned to the original level 5 seconds before the period of stimu-
lation was over. Well marked nasal dilation accompanied the response

° Tschalussow: Loc. cit., 539.
10 Fofanow and Tschalussow: Loe. cit., 554.

104 E. G. MARTIN AND W. L. MENDENHALL

and continued for four seconds after stimulation ceased, outlasting the
period of heightened pressure a total of 9 seconds. In another experi-
ment, one with an exceptionally low pressor threshold, sensory stimu-
lation of 35 Z units caused a rise of pressure of 14 per cent, and stimu-
lation on the same nerve of 127 Z units caused a rise of pressure of 29
per cent. The nasal dilation was slightly more pronounced in the first
case than in the second. The mechanical limit of the nasal mucosa
was not approached either time, for compression of the abdominal
aorta, within 9 minutes of these observations, produced a passive
nasal dilation nearly double the greater of them.

DISCUSSION

On the basis of the observations reported above, and those formerly
reported from this laboratory we would suggest the following as a
possible picture of the responses of the vasomotor mechanism to sen-
sory stimulations of increasing strength. Weak sensory stimuli excite
the vasodilator apparatus to activity. The result is a drop in general
blood pressure, with active dilation in such organs as are provided with
vasodilator nerves. Our data yield no positive mformation as to
whether weak stimuli affect the vasoconstrictor centre; but if this cen-
tre is affected appreciably it is inhibited. As stronger and stronger
sensory stimuli are used there begins to develop, concurrently with the
active vasodilation, excitation of the vasoconstrictor centre. The lat-
ter is more potent than the concomitant dilation although its latency
appears to be ordinarily somewhat longer, for the typical response to
intermediate stimulation is an initial active dilation followed by, and
overpowered by, a subsequent active constriction. With still stronger
stimuli the active constriction develops with less delay, so that, save in
exceptional cases, no indication is afforded of the vasodilator dis-
charges which are occurring simultaneously with the predominant vaso-
constrictor activity. When the sensory nerve stimulated is the vago-
depressor trunk, the situation is the same, except that with this nerve
strong stimulation inhibits the vasoconstrictor centre, reinforcing, in-
stead of opposing, the concurrent dilator discharges.

This suggestion of concurrent excitation of both mechanisms is di-
rectly contrary to the view of Bayliss (loc. cit.), according to which
vasoconstrictor excitation is accompanied by inhibition of vasodilator
tone. Bayliss states that most of his attempts to demonstrate inhibi-
tion of dilator tone gave negative results." He reports, however, two

11 Bayliss: Loc. cit., 349.

RESPONSE OF VASODILATOR MECHANISM TO SENSORY STIMULI 105

experiments which he interprets as supporting his contention (loc. cit.,
350). In the first of these, during a pressor reflex with rising general
blood pressure, active vasoconstriction occurred in the ear of a rabbit
both of whose cervical sympathetics had been cut. Since virtually all
vasoconstrictor fibres were presumably severed Bayliss considers that
this active vasoconstriction was brought about through inhibition of
vasodilator tone. This
interpretation involves
the rather difficult me-
chanical conception
that the arterioles are
so forcibly held open
through their dilator in-
nervation that when
this innervation sub-
sides they will constrict
forcibly against a rising
blood pressure in the
complete absence of
constrictor excitation.
The alternative possi-
bility suggests itself
that in this particular
animal an aberrant
vasoconstrictor inner-
vation to the ears was
present. As supporting
this alternative we wish
to cite an experiment of
our own. On February

)

1300 Z 1300 Z

a b
eniois we were mak- hee oe Active nasal vasoconstriction ae
; 7 pressor stimulation. a, cervical sympathetics in-
ing observations on a tact; b, cervical sympathetics cut. Peroneal nerve
cat under urethane - stimulated.

anesthesia. After elicit-

ing typical pressor responses from stimulation of the left peroneal and
left radial nerves, with active nasal vasoconstrictions with each re-
sponse, we cut both cervical sympathetics and repeated the sensory
stimulations. Contrary to our experience in numerous other experi-
ments, and to our great surprise, each pressor reflex was accompanied,
precisely as before, by active nasal vasoconstriction instead of by the ex-

106 E. G. MARTIN AND W. L. MENDENHALL

pected dilation. We performed ten sensory stimulations, three on the
left radial, four on the left peroneal, and three on the left popliteal with
results as stated. Figure 3 gives two tracings from this experiment.
In both of these the stimulation was applied to the peroneal; the
strength of stimulus was 1300 Z. Stimulation a shows the response
before the cervical sympathetics were cut; b, the response after they
were cut. The interval between stimulations a and b was 12 minutes.
The cervical sympathetics were cut 5 minutes after stimulation a. The
animal was vagotomized, and we searched both carotid sheaths care-
fully for minute nerve strands that might represent portions of the
cervical sympathetics separate from the main trunks. Comparison of
the nasal tracings in a and b of the figure indicates strongly that if the
first represents vasoconstrictor excitation the second does also. We
consider aberrant vasoconstrictor innervation a much more probable
explanation of this and of Bayliss’ phenomenon than depression of
dilator tone.

Bayliss’ second experiment is one in which diminution of volume of
the hind leg occurred in an animal, presumably with abdominal sympa-
thetics cut, from stimulation of the median nerve, with simultaneous
excitation of the depressor and the peripheral end of the cervical sym-
pathetic. The conditions of stimulation in this experiment are such as
to make the interpretation of the results difficult. An investigation
now in progress in this laboratory on the effects of simultaneous stimu-
lation of the depressor and other sensory nerves reveals a complexity
of interaction requiring careful quantitative comparisons for analysis.
The results of this investigation will be reported in due course. In
this connection we wish to say only that we should hesitate to admit
the possibility of inhibition of vasodilator tone during adequate stimu-
lation of the depressor nerve, unless all other possibilities, such as aber-
rant vasoconstrictor innervation, or passive redistribution of blood,
were more rigorously excluded than they appear to have been in this
experiment.

SUMMARY

Depressor (weak) stimulation of sensory nerves produces active di-
lation of the vessels of the nasal mucosa.

Pressor (strong) sensory stimulation produces active constriction in
the nasal mucosa.

Sensory stimuli of intermediate intensity produce typically an initial
fall in general blood pressure followed by a rise. The nasal mucosa

RESPONSE OF VASODILATOR MECHANISM TO SENSORY STIMULI 107

shows active dilation during the initial depressor period. Sometimes
the dilation persists through the second or pressor phase; at other times
there is active nasal constriction during the pressor phase.

In many instances the active constriction of the nasal mucosa under
pressor stimulation shows a longer latency than does the general blood
pressure rise.

Both these observations suggest an active dilator influence which is
able for a time to oppose successfully the constrictor influence occurring
simultaneously.

When both cervical sympathetic nerves are cut weak sensory stimu-
lation produces active dilation of the vessels of the nasal mucosa, show-
ing that the vasodilator mechanism is excited by weak stimuli.

We confirm the observation of Fofanow and Tschalussow that with
strong stimulation of the vago-depressor trunk, when the cervical sym-
pathetics are cut, the vasodilator mechanism is excited.

Observations are reported which indicate that the nasal dilation seen
when strong (pressor) stimulation is applied to ordinary sensory nerves,
with both cervical sympathetics cut, is, in part, at least, the result of
excitation of the vasodilator mechanism.

On the basis of these experiments the suggestion is made that the
vasodilator apparatus is ordinarily excited whenever sensory nerve
trunks are adequately stimulated. Strong stimuli, by exciting concur-
rently the vasoconstrictor mechanism, may overpower the dilation in
regions where both innervations obtain.

Although this suggestion is based altogether upon the results of arti-
ficial stimulation of sensory nerve trunks and contains no implication
that receptor stimulation may not be selective as regards the elicitation
of vasodilator or vasoconstrictor responses independently, it may not
be amiss to point out that an important feature of the physiological sig-
nificance of the high blood pressure induced by vasoconstriction is in
the diversion of much of the blood of the body into those organs which
appear to have a vasodilator innervation but no constrictor, namely,
the skeletal muscles. This distribution of the blood would be favored
by the mechanism here postulated.

SPINAL ANAESTHESIA IN THE CAT

G. G. SMITH anv W. T. PORTER
From the Laboratory of Comparative Physiology in the Harvard Medical School

The dangers and discomforts of ether anaesthesia have long been
recognized. Inevitable nausea, possible pneumonia, difficult res-
piration, renal injuries, make ether anaesthesia impossible in certain
cases, dangerous in others, and distressing in all. To meet these dif-
ficulties, spinal anaesthesia was devised for use in fields where general
anaesthesia is superfluous, such as operations on the perineum. As /
a rule beneficent, spinal anaesthesia is nevertheless by exception open |
to a grave and sudden danger. In the course of surgical procedures
otherwise completely successful, the vasomotor apparatus may sud-_
denly give way. The fall in blood pressure is immediate, sometimes |
profound, always disquieting. Nor can the surgeon predict in what )
patient it may appear.

The points of interest in this phenomenon are as follows: (1) The
extent and the time relations of the fall in blood pressure and the
effect of this fall on the efficiency of the central nervous system. ‘(2)
The region paralyzed. (3) The structures affected, whether the vaso-
motor center, the roots of the spinal nerves, the afferent or the efferent
paths in the body of the spinal cord. (4) The extent to which the
drug may pass along the cord from the point of injection, as modified
by the per cent of the drug in the solution used, the bulk of this solu-
tion, the force of gravity, and the possible fixation of the drug by the
tissues which it bathes. (5) The duration of the phenomena. (6)
The influence of adrenalin. (7) Remedial measures, directed to rais-
ing the fallen blood pressure. -

Obviously, these factors: cannot be studied with complete satis-
faction in man, in whom the condition of experimentation cannot
be varied at will. We present, therefore, a systematic investigation
of spinal anaesthesia in animals.!

1The only experimental study of blood pressure in spinal anaesthesia so far
as we are aware, is that of Gray and Parsons (Quarterly Journal of Medicine,
1911, v, p. 339), who concluded that the slight fall of blood pressure they ob-
tained in all cases was due to relaxation of the muscles of the abdomen and the
lower limbs and that the greater fall obtained in some cases was due to paralysis
of the intercostal muscles and the consequent diminution in the pumping power

of the chest.
108

SPINAL ANAESTHESIA IN THE CAT 109

METHOD

Fifty cats were used. In a number of these, two or more intra-
spinal injections were made, so that, in all, 72 experiments were done.
In 18 cases, ether alone was used. In all the animals, the preliminary
operations were done under ether anaesthesia. In 32 cases, in which
muscular reactions would have been a vital source of error, ether was
followed by curare. Enough dilute curare solution to paralyze the skele-
tal muscles was slowly injected through the femoral vein. The carotid
blood pressure was recorded by a membrane manometer. Graduation
scales for this manometer are shown in figures 1, 2 and 3. The con-
dition of the vasomotor system and of the sensory afferent tracts was
determined by measuring the changes in blood pressure on stimula-
tion of the central end of the brachial and sciatic nerves and on stim-
ulation of the dorsal columns of the cord. The induction currents
employed were just perceptible to the tongue. In order to make sure
that the drug entered the subdural space, the injection was made under
the guidance of the eye. Laminectomy, therefore, was always done at
the level of the injection. To determine the spread of the drug by
direct stimulation of the cord, laminectomy was often done at other
levels as well. A 2 ec. all-glass Luer syringe with 24 gauge needle was
used. In the cat, the space between the cord and the dura is so shallow
that, except in the lower lumbar region, the needle cannot be inserted
perpendicularly to the long axis of the cord without impaling the cord
itself. The direction of the needle, whether pointing cephalad or
caudad, was found to be a factor of some-influence on determining the
level to which the drug diffused.

In most of the experiments undertaken to ascertain the effect of grav-
ity, the foot of the board was raised.

_In 35 injections, tablets “C’’ containing 0.05 g. novocaine and
0.000083 g. adrenalin were used; in 18 injections, tablets ‘“D”’ contain-
ing 0.2 g. novocaine and 0.06 g. sodium chloride; in 17 injections, a
fluid preparation put up in ampoules, each of which contained 1.3
ec. of 5 per cent tropacocaine and 0.00017 g. suprarenin chloride;
and in 2 injections, a fluid preparation in ampoules containing 1 cc.
of 5 per cent tropacocaine in 0.6 per cent sodium chloride solution.
Taking as a test the paralysis of the tissue directly bathed by the
drug when injected, the novocaine with adrenalin gave 28 per cent
unsatisfactory results, ranging from partial to complete failure; novo-
‘caine and salt, 38 per cent; tropacocaine and adrenalin, 23 per cent;
only two injections were made with tropacocaine and salt and both

Fig.1. The original size. Injection of 0.01 g. novocaine and adrenalin in dilute
solution (1 cc.) at Lumbar

VII causes paralysis of dorsal columns extending
to Dorsal XI, and perhaps above. Brachial rise reduced from 65 to 33 per cent.

Lower curve—left to right U pper curve—left to right

1. Sciatic stimulation 12.10 p.m. 1. Brachial stimulation 12.35 p.m.
2. Brachial stimulation 12.14 2. Lumbar VII stimulation 12.37
8 Lumbar VII stimula- 12.15 2 Dorsal XI stimulation 12.39
ulation
( RI hye t. Sciatic stimulation 3.25
| (12.22 Injec- 5. Lumbar VIl stimulation 3.27
1. Record of blood pres- | tion of drug) 6. Dorsal XI Stimulation 3.29
sure VE2o
12.28
| 12.3
5. Sciatic stimulation 12.32

Seale: 50, 70, 90, 110, 130, 150, mm. Hg.
Experiment 45, curarized cat. February 26, 1915.

110

SPINAL ANAESTHESIA IN THE CAT (es

were successful. The percentages of failure were considerably higher
than in the clinical use of the same drugs, a difference probably due to
the conditions obtaining in experimental work on so small an animal,
and to the use of inadequate doses in certain experiments.

Following is a typical protocol.

Experiment February 26, 1915. A lightly etherized cat was tracheotomized
and cannulas placed in the left carotid artery and the left femoral vein. The
left brachial and sciatic nerves were tied and cut distal to the ligature. Lami-
nectomy was done at Lumbar VII and Dorsal XI. The carotid cannula was
connected with a membrane manometer.

11.50 a.m. 1.2 cc. 0.5 per cent curare solution in 15 ee. normal saline solution
injected into femoral vein.? Artificial respiration.

12.10 p.m. Sciatic nerve stimulated with induction currents. The blood
pressure rose from 105 mm. to 170 mm. Hg. (see fig. 1). Stimulation of brachial
nerve and of dorsal columns at Lumbar VII gave slightly greater increase.

12.17. Blood pressure recorded.

12.22. 1.0 ce. of 1.0 per cent novocaine and adrenalin C (made with distilled
water) was injected very slowly into the dural sac at Lumbar VII. The needle
was inserted perpendicularly to the long axis of the cord.

12.25. Blood pressure has fallen from 100 mm. to 55 mm.

12.28. Blood pressure 65 mm.

12.31. Blood pressure 71 mm.

12.32. Stimulation of sciatic nerve. The reflex change in blood pressure
has disappeared.

12.35. Brachial stimulation causes a rise from 75 to 100 mm. (33 per cent)
instead of the rise from 100 mm. to 165 mm. (65 per cent) shown before novocaine.

12.37. No reflex rise on stimulating cord at Lumbar VII

12.39. No reflex rise on stimulating cord at Dorsal XI.

3.25. Stimulation of sciatic, brachial, Lumbar VII, and Dorsal XI cause
normal reflex increase in blood pressure.

In all the experiments, especial care was taken to avoid errors from
ether, curare, and artificial respiration.’

CHANGES IN BLOOD PRESSURE

Extent of fall. In Table 1 are recorded 20 experiments in which novo-
caine or tropacocaine was injected in the lumbar region in strength
sufficient to block all afferent impulses set up by stimulation of the
sciatic nerve. In two cases, the blood pressure fell more than 40
per cent. A fall of 40 per cent is, however, not necessarily alarming.

2 This small dose of curare is excreted after artificial respiration has con-
tinued some time and the curare must then be renewed.
3 For precautions, see W. T. Porter: this Journal, 1910, xxvii, pp. 281, 282.

SITE OF
INJECTION

SS eS | ee

VII

Cervical
TT
III
1B
Til
III

Ve

III
Ill
III
III
Ill
III
‘EI

TABLE 1

BLOOD
PRESSURE

from to

138-105
J00- 90
150-100
130-100
140-130

100

130-140

80-

70

120-100
170-140
130-120

90
100-
120-

80
95

160-100

100-
100-—
150-
120-

55
70
80
90

190-170 .

120-
130-
115-
120-
100—
100-
120-

100- §

‘110-
120-
130-
115-

80-
100-
120-
120-
110-

80-

110- 4

125— 7

100-
90-

95- ¢

80-

120- §

80-

80
80

55

110-160

100-

60

110-100
120-110

100-—
130-

112

90
70

~

ABSOLUTE
FALL

PERCENTILE
FALL

GaSexSSRzeVs

53

SPINAL ANAESTHESIA IN THE CAT iis

The criterion is not the absolute or percentile fall of blood pressure per
se, but whether there remains sufficient blood pressure to carry on,
for a time at least, the work of nerve cells in the brain and cord. In
ome of the two cases just cited, the pressure fell from 150 mm. to 80
mm. Hg, in the other it fell from 100 mm. to 55mm. The danger line
may probably be placed at 60 mm. In only one instance out of twenty,
therefore, was there a serious fall in consequence of a lumbar injection.

In the dorsal region, 19 injections were made, in 9 of which the
blood pressure fell to 60 or less.

In the cervical region, there were 13 injections and in 5 the blood
pressure fell to 60 or below.

In the cat the residual blood pressure, after the extirpation of the
spinal cord‘ is from 28 to 31 mm. In our present experiments, the
lowest blood pressure after lumbar injection was 55 (one instance);
after dorsal injection, the pressure fell in two cats to 30 mm. and in
three cats to 40 mm.; after cervical injection, the pressure fell once
to 40 mm. In several instances, therefore, the vasomotor apparatus
was absolutely paralyzed and the function lost as completely as if the
spinal cord had been extirpated. These instances were, in each case,
the result of injection in the dorsal or cervical regions.

Duration of low blood pressure. The injury to nerve cells caused by
low blood pressure and the consequent impaired nutrition, depends
on two variables; namely, the extent to which the pressure falls and
the duration of the low pressure. Recovery from the slight falls in
blood pressure usually took place rapidly. After the more severe
falls, partial but sufficient recovery took place in from 30 to 90 minutes.
The duration of low blood pressure appeared to depend more upon the
amount of drug injected than upon the site of injection.

Our experiments give evidence that in the majority of instances the
vasomotor system was not seriously impaired by the pressures noted
in Table I. After 70 to 180 minutes the reflex change in blood pres-
sure on stimulation of the sciatic and brachial nerves returned almost,
if not quite, to normal. Thus in Experiment 48, sciatic stimulation
caused the biood pressure to rise from 90 to 145 mm., 61 per cent.
After injection at Dorsal VI, the pressure fell from 110 to 45 mm.
Seventy minutes later, the blood pressure was 70 mm. and on sciatic
stimulation it rose to 110 mm., 57 per cent.

* Porter and Storey: this Journal, 1907, xviii, p. 196.

114 G. G. SMITH AND W. T. PORTER

THE REGION PARALYZED

It is obvious that drugs like novocaine, which are given to interrupt
the afferent conducting paths in the spinal cord, may also interrupt
the efferent paths. The object in view is to suspend sensations of
pain without at the same time suspending some function essential
to the well-being of the patient. Such injuries must depend on the
importance of the several regions paralyzed. Reflection upon the anat-
omy of the spinal cord will show that the vasomotor and the respira-
tory functions are especially to be considered.

The vasomotor function. The reader is reminded that the master
cells controlling the tonus of the arteries and thus the weight of the
blood pressure are situated in the bulb. Their axis cylinder processes
descend the cord in the antero-lateral tracts, bend into the gray
matter, and end there in contact with the spinal vasomotor cells. The
axis cylinder processes of the spinal vasomotor cells leave the cord in
the anterior roots of spinal nerves from the first Dorsal to about the
first Sacral. The vascular areas served by these fibers are most of them
without importance in the present investigation. Thus in the eat,
the sciatic nerve, bearing vasomotor fibers for the hind limb, may be
severed without causing any considerable fall in blood pressure. In
fact, only the splanchnic nerves, given off from the upper dorsal region,
innervate a vascular area large enough to be a dangerous factor in spinal
anaesthesia. This area, comprising the abdominal viscera, is, how-
ever, so extens.ve as to make splanchnic paralysis very serious, in that
the dilatation of the arteries controlled by the sp'anchnic vasomotor
fibers may cause so much blood to enter the corresponding veins that
not enough is left in the bulb and cord to support properly respiration
and other vital functions. The rabbit, for example, may be bled to
death into its own portal system, by section of the splanchnic nerves.
In studying the regions affected by spinal anaesthesia, special. atten-
tion should, therefore, be paid to that containing the splanchnic fibers.
Spinal anaesthesia can be successful, only when the afferent paths are
paralyzed without at the same time paralyzing enough splanchnic
cells or splanchnic root fibers to lower the blood pressure to a degree
that may threaten the continued activity of the centers situated in
the cervical cord and the bulb.

The importance of the splanchnic area becomes clear when atten-
tion is directed to the average percentile fall in blood pressure after
injections in the several regions of the cord (Table 1). The averages
are: lumbar, 19 per cent; dorsal, 43 per cent; and cervical 27 per cent.

SPINAL ANAESTHESIA IN THE CAT 115

‘The regional averages just presented, show that the fall in blood
pressure, following dorsal injection, is due to paralysis in the splanch-
nic region and not to paralysis of the bulbar vasomotor center. For,
if the fall were due to interference with the bulbar center or with the
vasomotor fibers connecting it with the spinal vasomotor cells, the use
of novocaine in the cervical region, nearer the bulbar center, should

!

TABLE -2 _

Maximum fall in blood pressure in relation to dosage, region injected, and ~*

direction of injection
Fatt or 30 Per Cent or Less Fatt or More THAN 30 PeR Cent
EXPERI- PER |EXPERI- PER
PER NEEDLE PER NEEDLE

MENT | BULK “| LEVEL 5 CENT] MENT | BULK LEVEL = CENT

a0. CENT TOWARDS | Fr| No. CENT TOWARDS | Fir,
ce. ce
17 0.2 a C. HI Head 5 46 0.2 5 C. II Head 46
18 0.3 4 C> TEL Head 18 16 0.4 2 C. Ill Tail 33
2 0:4 1. 235° )D: II Tail 73
9 0.1 1 D. IV Head 10 8 0.1 1 D. IV Tail 37
35 0.4 2.5 DD: ES Spine 27 5 0.3 2 1D IV Tail 37
12 0.5 2 L. I Tail 24 7 0.1 2 D. IV Tail 60
1l 0.4 2 ty SEE Tail 24 6 0.2 2 D. IV Tail 60
10 0.2} 2 L. III | Head 25 43 0:2°F .&: IDSs. VI. Spine 54
49 0.2) 5 tr VE Spine 25 3 0.5 | 2 Dee. EX Head 38
47 0.5 2 L. VIl Head 30 2 0.5 1 D. XIII Head 33
injected
forcibly.

42 0.2 5 Le. VEL Spine 16 25 0.2 2.5|L VI 33
50 0.2 | 10 L. VII | Spine 10 44 PAGS ies) | Ee VE: Spine 37
45 £0) DOG. VEL Spine 45
48 0.8 2.5 |L VII Spine 47

In all the above experiments a mixture of novocaine and adrenalin C. was used.
““Spine”’ means that the needle was pointed at right angles to the long axis of the cord.

SUMMARY
Injection away from Injection in splanch-
splanchnic area nic area
11 Experiments 14 Experiments
LATER CEG OS = 0.0093 g. = 0.0082 g.
PAREESP ESTE Kee se ee ck eh ce alale — i (nsces = 0.42 ce.
PAMOIAE SC IGT GEIG (550.5 2s aco .soees wie = 3.6 per cent = 2.3 per cent
Average blood pressure fall......... = 19.5 per cent = 45.2 per cent

cause as great a fall as its use in the dorsal region—at any rate, the
fall should not be less.

The conclusion that the vasomotor paralyses of spinal anaesthesia
are to be sought in the splanchnic area rather than in the bulbar vaso-
motor center is further supported by the observations on the fall of
blood pressure as affected by the direction in which the injection is
made.

116 G. G. SMITH AND W. T. PORTER

Fig. 2. The original size. Injection of 0.01 g. novocaine and adrenalin (0.2
cc.) at Dorsal VI causes fall in blood pressure from 110 mm. to 45 mm. in three
minutes with abolition of vasomotor reflex from sciatic and brachial.

Lower curve—left to right Upper curve—left to right
1. Sciatic stimulation 8.55 p.m. 1. Brachial stimulation 9.15 p.m.
2. Brachial stimulation 8.58 2. Dorsal VI stimulation 9.17
3. Dorsal VI stimulation 9.01 3. Brachialstimulation 9.42
( 9.02 4, Brachial stimulation 10.12
| (9.03 Injection 5. Sciatic stimulation 10.14
4. Record of blood pres- } of drug) €. Dorsal VIstimulation 10.15
sure 9.06
' 9.09
| 9.12
5. Sciatic stimulation 9.13

Scale: 30, 50, 70, 90, 110 ,130, 150, mm. Hg.

>

Experiment 43, curarized cat. February 19, 1915.

SPINAL ANAESTHESIA IN THE CAT 117

Fig. 3. The original size. Injection of 0.01 g. novocaine and adrenalin (0.2
ce.) at Lumbar VII causes fall of blood pressure from 120 mm. to 95 mm. in three
minutes. Paralysis of sciatic incomplete eleven minutes after injection. Eleva-
tion of foot of board at angle of 30°, eighteen minutes after injection, is followed
by complete paralysis of sciatic nerve.

Lower curve—left to right Upper curve—left to right
1. Sciatic stimulation 9.11 p.m. 1. Sciatic stimulation 9.48 p.m.
2. Brachial stimulation 9.13 2. Lumbar VII stimulation 9.49
3. Lumbar VII stimu- 9.15 3. Lumbar I stimulation 9.51
lation
( 9.17 4. Brachial stimulation 9.54
| (9.22 Injection 5. Sciatic Stimulation 10.22
4. Record of blood pres- | of drug)
sure ) 9.25
| 9.28
| 9.31
5. Sciatic stimulation 9.33

Scale: 70, 90, 110, 130, 150, 170, 190 mm. Hg.
Experiment 42, curarized cat. February 16, 1915.

118 G. G. SMITH AND W. T. PORTER

Table 2 deals with the fall of blood pressure in relation to dosage,
region injected, and the direction of the injection. On the left side
of this table are placed the cases in which the fall of blood pressure
was 30 per cent or less; on the right side are those in which the fall
was more than 30 per cent of the initial pressure. In every case in
which the fall was 30 per cent or less (except Experiment 35), the in-
jection was so made that the drug was driven away from the area in-
cluded between Dorsal I and IX (figs. 2 and 3). Contrast Experiment
8 with Experiment 9; in both, 0.1 ce. of 1 per cent novocaine and
adrenalin was injected at Dorsal IV. In Experiment 8, in which the
drug was driven towards the tail, the fall was 37 per cent; in Experiment
9, in which the solution was driven towards the head, the fall was
10 per cent. Again, in Experiment 16, 0.4 ec. of 2 per cent solution was
injected at Cervical III caudad; the fall was 33 per cent; in Experiment
4, 0.4 ce. of 2.5 per cent solution, injected caudad from Dorsal II, much
nearer the fall-producing area, was followed by a fall of 73 per cent.
The conclusion again appears justified that with moderate but ade-
quate doses, the fall in blood pressure in spinal anaesthesia is caused
by paralyses in the splanchnic area.

The clinical use of spinal anaesthesia is limited to the injection
of the drug in the lumbar region. As the drug diffuses towards the
head, the first part of the vasomotor mechanism affected by it will be
the roots in the thoracic area. It seems justifiable to assume that ih
clinical, as well as in experimental spinal anaesthesia, the fall of blood
pressure is caused by paralysis of the splanchnic area.

Paralysis of respiration. Out of a total of 18 experiments in which
no curare was used, the injection was made in the cervical or upper
thoracic region ten times. In four of these ten injections, the drug
was driven towards the tail from a point below the phrenic nerve
and there was no paralysis of respiration. In the other six, the drug
was injected towards the fifth cervical level; in four of these cases,
respiration was paralyzed. In the other two cases, the dosage was
very small (0.1 ec. of 1 per cent, 0.1 of 2 per cent novocaine and
adrenalin. )

In closing this discussion of the regions affected by spinal anaesthesia,
it is important to answer the very practical question, How often will
surgical anaesthesia of the lumbar and sacral region be complicated
by a serious fall in blood pressure or by an interruption of the breath-
ing? In our experiments, there was but one case out of twenty lumbar
injections in which the fall in blood pressure (to 55 mm.) might have

Fig. 4. The original size. Injection of 0.01 g. novocaine and adrenalin (0.2
e.c.) at Cervical II cephalad causes fall of blood pressure which is gradual rather
than abrupt, due probably to slower action of drug on cord itself than on thoracic
roots. Dorsal columns blocked, but vasomotor mechanism below D I is un-

affected.

Lower curve—left to rizht Upper curve—left to right
1. Sciatic stimulation 9.40 p.m. 1. Brachial stimulation 10.05 p.m.
2. Brachial stimulation 9.438 2. Cervical II stimulation 10.07
3. Cervical II stimula- 9.46 3. Dorsal II lateral surface
tion stimulation 10.12
9.48 4. Sciatic stimulation 11.40
| (9.49 Injection 5. Brachial stimulation 11.43
4. Record of blood pres- 9 ae gene
ia 9.55
9.58
10.01
5. Sciatic stimulation 10.03

Experiment 46, curarized cat. March 4, 1915.

119

120 G. ¢. SMITH AND W. T. PORTER

been serious, and in the eight injections in which no curare was used
there was no paralysis of respiration.

THE STRUCTURES AFFECTED

The usual site of paralysis being in the splanchnic area, we should now
enquire whether the drug affects the anterior nerve roots or the paths
bringing vasoconstrictor impulses from the bulb.’

Since paralyses of respiration are so infrequent in spinal anaesthesia,
we have not attempted to differentiate paralysis of the phrenic root
fibers from that of the bulbo-phrenic respiratory path.

It may at once be stated that a strength of the drug sufficient to
paralyze the afferent sensory paths (so that stimulation of the central
end of the sciatic nerve produces no reflex) will also paralyze the effer-
ent vasomotor fibers (fig. 4). This is illustrated by Experiment 23,:in
whidh 0.5 ce. of 2.5 per cent tropacocaine and adrenalin were applied
to all surfaces of the cord at Cervical II. The dura at that level was
laid open. The blood pressure fell from 80 to 55 mm.; stimulation of
the sciatic and brachial nerves and the anterior surfaces of the cord
at Cervical III produced no response. Stimulation of the anterior
surface of the cord at Dorsal II, however, was followed by an excellent
rise in blood presgure, thus proving the integrity of the vasomotor mech-
anism below the paralyzed portion.

It is possible, on the other hand, to secure paralysis of the nerve
roots without disturbing the conductivity of the vasomotor paths in
the substance of the cord, as in Experiment 15. In this cat, 0.2 ec. of
4 per cent novocaine and salt solution ‘“‘D’’ was injected at Cervical
III. The stimulation of the sciatic caused the blood pressure to rise
from 80 mm. to 140 mm., but brachial stimulation caused no rise. The
brachial roots in this experiment were paralyzed but the afferent paths
conveying sciatic impulses remained unaffected.

There is some evidence to show that different functions may be af-
fected differently. For example, Experiments 2, 3, and 8 showed that
the motor paths are paralyzed more easily than the sensory paths

Experiment 2. 0.5 ec. of 1 per cent novocaine and adrenalin was injected at
Dorsal XIII. Stimulation of left sciatic nerve followed by rise in blood pres-

5 We have at present no satisfactory method of isolating effects limited to the
splanchnic cells, if indeed the cells are ever paralyzed independently of the nerve
paths.

SPINAL ANAESTHESIA IN THE CAT 121

sure throughout experiment, whereas right hind leg was completely paralyzed
for 55 minutes.®

Experiment 8 shows that the vasomotor reflex may persist although
spontaneous motion of the extremities is lost.

Experiment 8. In an etherized cat, 0.1 cc. of 1 per cent novocaine and adrena-
lin was injected at Dorsal IV at 12.04p.m. At 12.31, 12.36 and 12.44, left sciatic
stimulation was followed by rise in blood pressure from 105 to 120, 100 to 115,
and 110 to 120. The right hind leg remained paralyzed for 65 minutes.

THE DIFFUSION OF THE DRUG ALONG THE CORD

In studying the diffusion of the drug along the spinal cord,’ it seemed
well to fix a reasonable interval between the moment of injection
and the testing of the resultant paralysis. This period was set at fif-
teen minutes, in which time the drug seemed to have exerted its maxi-
mal effect. Care was taken not to manipulate the spine after the
injection, lest the fluid injected should be pumped or driven to a more
distant level.

Per cent of drug. In the following experiments, the same quantity
of solution was injected, but the solution contained different amounts
of anaesthetic (a constant mixture of novocaine and adrenalin “‘C’’).
Paralysis of the dorsal column to direct stimulation was the test’ em-
ployed to fix the limits to which the drug had spread.

Experiment 26. 0.225 cc. of 2.5 per cent (0.0056 g.) at Lumbar VII diffused
6 vertebrae.

Experiment 42. 0.2 cc. of 5 per cent (0.01 g.) at Lumbar VII did not diffuse
5 vertebrae.

Experiment 49. 0.2 ec. of 5 per cent (0.01 g.) at Lumbar VI did not diffuse
2 vertebrae.

Experiment 50. 0.2 ce. of 10 per cent (0.02 g.) at Lumbar VII did not diffuse
4 vertebrae.

The average diffusion here was less than four vertebrae.

6S. Baglioni (Centralblatt fiir Physiologie, 1910, xxiii, pp. 869-873), has shown
that after the subdural injection of stovain, sensations disappear in this order:
pain, cold, heat, pressure; they return in reverse order. This seems to show a
varying degree of resistance to the effect of drugs.

7 Some writers in this field emphasize movements of the spinal fluid due to the
effects of respiration upon the emptying and filling of the cerebro-spinal venous
system. That this possible factor in the diffusion of the drug is done away with
when the dorsal sac is opened to atmospheric pressure, an operation found essen-
tial in our experiments, we are not prepared to deny. Other, and more important,
factors affecting distribution within the dural sac can be studied by our method
and indeed the problem is simplified by the removal of confusing influences.

122 Cc. Gi. SMITH AND W. T. PORTER

Varying bulk of fituid. In the following experiments the amount
of fluid injected ek varied, while the percentage of the drug (novo-
caine and adrenali/n “‘C’’ remained the same. The paralysis of the
dorsal columns w/as again the test of diffusion.

Experiment 25. | 0.225 ce. of 2.5 per cent (0.0056 g.) at Lumbar VII diffused
6 vertebrae.

Experiment 2,

4. 0.25 cc. of 2.5 per cent (0.0062 g.) at Lumbar VI diffused 6

vertebrae. |
Experiment #. 0.4 ce. of 2.5 per cent (0.01 g.) at Dorsal II diffused 9 vertebrae.
Experimenté 35. 0.4 cc. of 2.5 per cent (0.01 g.) at Dorsal IX diffused 11
vertebrae.

Experimy ent 48. 0.8 cc. of 2.5 per cent (0.02 g.) at Lumbar VII diffused 8
vertebraue

The average diffusion was eight vertebrae.

Additional information is afforded by certain experiments in which
the drug was injected at approximately the same levels (Lumbar VI
or VII) but in which dilute and concentrated solutions are examined
with regard to their effect upon blood pressure. The greater the fall,
the further the drug progressed toward the splanchnic area.

Divute solutions

Experiment 48. 0.8 ec. of 2.5 per cent (0.02 g.) caused blood pressure to fall
46 per cent.

Experiment 44. 1.0 cc. of 1.5 per cent (0.015 g.) caused blood pressure to fall
37 per cent. ;

Experiment 45. 1.0 ce. of 1.0 per cent (0.01 g.) caused blood pressure to fall
45 per cent.

Experiment 47. 0.5 ec. of 2.0 per cent (0.01 g.) caused blood pressure to fall
39.5 per cent.

The average fall was 41.9 per cent. The average dose was 0.014 g.

Concentrated solutions

Experiment 50. 0.2 cc. of 10 per cent caused blood pressure to fall 11 per cent.
Experiment 41. 0.3 cc. of 5 per cent caused blood pressure to fall 16 per cent.
Experiment 49. 0.2 ec. of 5 per cent caused blood pressure to fall 25 per cent.
Experiment 42. 0.2 cc. of 5 per cent caused blood pressure to fall 17 per cent.
Experiment 40. 0.2 cc. of 5 per cent caused blood pressure to fall 20 per cent.

The average fall was 17.8 per cent. The average dose was 0.013 g.
The analysis of the observations on diffusion does not show any
very definite laws, probably because the number of experiments is

SPINAL ANAESTHESIA IN THE CAT 123

necessarily limited. On the whole, the bulk seemed a factor of greater
importance than the strength of the solution. Dilute solutions seemed
to spread further than concentrated solutions. But in some cases,
a dose of small bulk and containing a small amount of the drug pro-
duced a more widespread effect than a dose larger both in bulk and in
percentage of drug injected in a manner as nearly similar as possible.

Effect of gravity. We are aware of the possibility of error in attempt-
ing to determine the effect of gravity upon the diffusion of a drug in-
jected into a dural sae which is exposed at the highest point to atmos-
pheric pressure, whereas normally the cord is protected by its bony
envelope. To avoid this source of confusion as far as possible, we
tilted the animal and allowed the blood pressure and spinal fluid to
become settled after the change of position, before injecting the drug.
After the injection, the board was left tilted for fifteen minutes, then
returned to level and the dorsal columns were stimulated to determine
the extent of the diffusion.

The following experiments compare three animals in the horizontal
position with four in which the head was tilted down at a varying angle.

Horizontal

Experiment 42. 0.2 ec. 5 per cent novocaine and adrenalin C at Lumbar VII,
did not diffuse 5 vertebrae.

Experiment 49. 0.2 cc. 5 per cent novocaine and adrenalin C at Lumbar VI,
did not diffuse 2 vertebrae.

Experiment 29. 0.2 ec. 5 per cent novocaine D at Lumbar VII, diffused 3
vertebrae.

Tilted, head down

Experiment 31. 0.2 ec. 5 per cent novocaine D at Lumbar VII, diffused 6
vertebrae.

Experiment 32. 0.2 ce. 5 per cent tropacocaine and adrenalin at Lumbar VII,
diffused 7 vertebrae.

Experiment 39. 0.2 ec. 5 per cent novocaine D in 5 per cent glucose at Lumbar
VII, diffused 8 vertebrae.

Experiment 40. 0.2 cc. 5 per cent novocaine C in 5 per cent glucose at Lumbar
VII, diffused 8 vertebrae.

It appears that tilting the animal board at an angle of 40, head down,
increases the diffusion of novocaine and salt solution, and that the
diffusion is increased to a slight degree when the drug is carried in a
5 per cent glucose solution.

Fixation of the drug. The complete fixation of the drug in some
loose chemical combination with the tissues of the cord would be greatly

124 G. G. SMITH AND W. T. PORTER

to the advantage of the surgeon. If such a bond existed, the action
of the anaesthetic would soon be localized. If the paralysis had ex-
tended far enough to affect seriously the blood pressure, the patient
could then be tilted head down, thus keeping by the force of gravity a
supply of blood in the brain. If, however, it can be demonstrated that
the drug is not entirely fixed, it would follow that tilting the patient
might cause the unfixed remainder of the anaesthetic to flow towards
the head, invading more of the splanchnic region, and even reaching
the phrenic cells, and finally the spinal bulb.
We present three observations upon fixation:

Experiment 34. One ce. of 1.0 per cent tropacocaine and salt solution was in-
jected at Lumbar VII. Twenty minutes later, stimulation of Lumbar I pro-
duced a fair rise in blood pressure. Five minutes after that, the dura at Lumbar
III was opened. Evidently the manipulation drove the drug upwards, for after
that Lumbar I no longer reacted.

Experiment 38. 0.4 ec. of 5.0 per cent novocaine and salt solution D was in-
jected at Cervical III. Blood pressure fell from 100 to 90 mm., but returned to
100 mm. in 13 minutes. Sixteen minutes after the injection, the dura was opened
at Dorsal X and as the spinal fluid flowed down the cord the blood pressure fell
from 100 to 80.

Experiment 40. 0.2 ce. of 5.0 per cent novocaine and adrenalin C was in-
jected at Lumbar VII. Paralysis of the sciatic did not occur. Eighteen minutes
after the injection, the board was tilted and immediately afterwards complete
paralysis of the sciatic was found to have taken place.

From these three experiments we may conclude that after 25, 16, and
18 minutes respectively, enough drug free from fixation was present
to paralyze other nerve fibers.

THE DURATION OF THE PHENOMENA

The duration of paralysis of the vasomotor reflexes was studied
in relation to the absolute amount of drug injected, and also in relation
to the percentage of the drug in solution. In many cases, a low dos-
age or one of weak percentage (1 to 2 per cent) secured as long a paraly-
sis as did stronger or larger doses. The minimum dose could not be
determined with any finality, for in one experiment 0.1 cc. of 1 per
cent solution would be fairly effective, while in another a much larger
dose would not give the desired effect.

In Experiments 26, 27, 29, and 30, the spinal fluid was drained off
15 minutes after the injection of the solution, but this did not shorten
the duration of the paralysis.

i

SPINAL ANAESTHESIA IN THE CAT 125

In Experiments 5, 6, 8, 20, 21, 22, 24, and as the effects of the drug
began to wear off, the stimulation of the sciatic nerve was followed
by a fall of blood pressure instead of a rise. In cases in which this
phenomenon occurred, the normal reflex returned before the blood pres-
sure rose to its original level.

THE INFLUENCE OF ADRENALIN

In order to learn whether adrenalin was a factor in the phenomena
following spinal anaesthesia, we twice injected adrenalin chloride alone.

Experiment 37. 0.5 cc. of 1-10,000 adrenalin chloride was injected cephalad
from Dorsal XI. No change in blood pressure followed, but sciatic reflex was
temporarily diminished, perhaps because the solution was cool.

Experiment 38. 0.5 cc. 1-10,000 adrenalin chloride, warmed, was injected
caudad from Cervical III. The blood pressure fell from 120 to 110 in five min-
utes. The reflexes were not affected.

A comparison of the action of novocaine and adrenalin C with that
of novocaine and salt solution D is given in Table 3. The injection
was caudad in Experiment 15 and cephalad in all the others.

TABLE 3

Comparison of action of novocaine + salt (““D’’) and novocaine
+ adrenalin (‘‘C’’)

BA et 2. &
A os ° Aa DURATION OF - ° °
z On | ahs PARALYSIS a 2 £
sre DOSE LEVEL | & 4 2 a a e ———| 2 < é Zn
ae at: ee at and ee
: On B AD sae reflexes < ap 2 &
per cent
15-D ie ec. of 4% (caudad)........ Gat Tiialign2 20 25 26 No Yes
17-C 0.2 ee. of 4% (cephalad)...... Co FER 3 5 5 0 No No
29-D WP MIGE TOLD Voss. occ0 co decle ts cows s L. VII 0 0 0 25 Yes
. 42-C aren OF Va. ae a cits cc L. VII 3 16 60+ 60+ | Yes Yes
30-D | 1 CE n/n L. VII | Rise 0 27 Yes Yes
49-C (V2) fo ON N Oe Re ee Le Vi 9 25 23 28+ | Yes Yes
36-D Ree CGH OO G20. sta onc ce ee CG. TE Rise 35 Yes ?
46-C TT TE a) TY A ee ae GC: Tf 12 46 110+ 110 Yes fy
38-D Sp ECAORD DUG Se he: Co In 05 9 4 20 No Yes
16-C ATCC OR 2 gs3 005 Slag lone oS Cc. Lil 3 33 16+ 0 No Yes
39-D a cee. of 5% glucose.......... L. VII 0 0 0 105+ | Yes Yes
40-C 0.2 ce. of 5% glucose.......... L. Vil 3 20 60 60 Yes Yes

Table 3, showing six pairs of experiments, exhibits a markedly greater
fall of blood pressure in the cases in which adrenalin was used. This
occurs in five of the six pairs. In the remaining pair, the fall of blood
pressure with novocaine and salt is not great (20 per cent). We do

126 G: G. SMITH AND W. T. PORTER

not attempt to explain this fact. It may be that the pressure of salt

in solution D is a factor.

It is also noteworthy that although in three experiments the novo-

caine D was more effective as regards duration of paralysis of the re-
flexes, the average duration of the paralysis after use of D was 40 min-
utes, whereas, of the four experiments in which C produced paralysis
at all, the average duration was 64 minutes. The longest action of
D was secured when the solution was made up in 5 per cent glucose.
C failed twice to produce paralysis; D never failed. On the whole,
the honors seem to be divided fairly evenly.

MEASURES TO RAISE THE FALLEN BLOOD PRESSURE

In five experiments an effort was made to raise the lowered blood
pressure by the intravenous injection of salt solution, adrenalin chloride
or pituitrin. This was done both when the blood pressure was lowered
by section of the cord, and by spinal anaesthesia. To be of value,
such experiments should be done only after section of the cord as other-
wise the natural return of blood pressure as the spinal drug wears off
will influence the results.

Saline solution, in the two experiments in which it was injected in-
travenously, did not materially affect the blood pressure.

The effect of pituitrin and adrenalin were tried with the cord cut

across at Cervical III. Blood pressure stood at 50 in Experiment

36, at 55 in Experiment 38. In Experiment 36, 0.5 cc. pituitrin in 5
ec. H,O was given. The blood pressure rose in one minute from 50
to 100, and four minutes later had fallen again to 60.

In Experiment 38, 0.5 ec., 1-10,000 adrenalin chloride in 5 ec. NaCl
was given. The blood pressure rose from 55 to 160 at once, but in
four minutes after the injection was back at 50.

CONCLUSIONS

1. In our experiments on spinal anaesthesia, there was in twenty
animals but one case in which a moderate but adequate injection
in the lumbar region caused a fall in blood pressure that might have
been serious (to 55 mm.); and in the eight cases in which no curare
was used there was no paralysis of respiration after lumbar injection.

2. Even after marked falls in blood pressure partial but sufficient
recovery took place in from 30 to 90 minutes. The duration of low
blood pressure appeared to depend more upon the amount of drug
injected than upon the site of the injection.

SPINAL ANAESTHESIA IN THE CAT 127

3. The fall in blood pressure seen after lumbar and dorsal injection
is due to paralysis in the splanchnic region, In our numerous obser-
vations, it was not due to paralysis of the bulbar vasomotor center.

4. A strength of the drug sufficient to paralyze the afferent sensory
paths in the cord (so that stimulation of the central end of the sciatic
nerve produces no reflex) will also paralyze the efferent vasomotor
fibers.

5. The nerve roots may in some cases be paralyzed without dis-
turbing the conductivity of the vasomotor paths in the substance of
the cord.

6. There is some evidence that different functions may be affected
differently; thus in three experiments the motor paths were paralyzed
more easily than the sensory paths.

7. Regarding the diffusion of the drug, the bulk seemed on the
whole a factor of greater importance than the strength of the solution.
Dilute solutions usually but not always spread further than concentrated
solutions.

8. Gravity is a factor of some importance; tilting the animal at an
angle of 40°, head downward, increased the diffusion of the drug.

9. Fixation of the drug is only partial. In three experiments, after
25, 16, and 18 minutes respectively, enough remained free to paralyze
other nerve fibers.

10. In seven experiments, as the effect of the drug began to wear
off, the stimulation of the sciatic nerve caused a fall of blood pressure
instead of the usual rise. In these cases, the normal reflex rise returned
before the blood pressure attained its original level.

11. A greater fall of blood pressure occurred in the cases in which
adrenalin was used in connection with tropacocaine or novocaine.

12. Measures taken to raise the fallen blood pressure were of little
value. It was easy to restore the blood pressure to normal but the
normal level could be maintained but a few minutes.

THE CONDUCTION WITHIN THE SPINAL CORD OF
THE AFFERENT IMPULSES PRODUCING PAIN
AND THE VASOMOTOR REFLEXES

S. W. RANSON anp C. L. VON HESS

From the Anatomical and Pharmacological Laboratories of the Northwestern Uni-
versity Medical School

Received for publication May 7, 1915

One of the problems which has been puzzling investigators for many
years is that of the varieties of cutaneous sensation. How are im-
pressions of pain, touch, heat and cold differentiated, and how are the
underlying afferent impulses propagated along peripheral nerves and
the spinal cord. So far as the peripheral nerves are concerned the
problem has remained very obscure, and yet in the last decade advances
have been made in our knowledge of the physiology and histology of
afferent nerve fibers which promise to throw some light on the question.

On the physiological side Head (1) has shown that cutaneous sen-
sations fall into two groups. In general it may be said that pain
and temperature correspond to his “protopathic’’ group and light

touch to his ‘‘epicritic.’”” He has shown that there must be two kinds

of cutaneous afferent nerve fibers corresponding to his two types of
sensation. Fibers of one kind mediate protopathic sensation; fibers
of the other sort mediate epicritic sensation. These two kinds of
fibers differ as to anatomical distribution and rate of regeneration.

On the histological side Ranson (2) has shown that there are two
kinds of afferent nerve fibers, the medullated and non-medullated.
Although the non-medullated afferent nerve fibers are very numerous,
in some nerves more than twice as numerous as the medullated, they
remained unknown, until in 1911 a differential axon stain was developed
and applied to the peripheral nerves. A remarkable parallel exists
between Head’s account of the protopathie fibers and the facts which
have already been ascertained in regard to the non-medullated fibers.
Space does not permit us to make a detailed comparison. We will
restrict ourselves to a comparison of their course in the spinal cord.
Head’s protopathic fibers after entering the cord turn at once into the

128

st

BOA as ev

intermediate layer.

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 129

gray matter. The epicritic fibers ascend for longer or shorter distances
in the posterior funiculus before they enter the gray matter. This
difference in the arrangement of the protopathic and epicritic fibers .
in the cord corresponds exactly to the difference in the course of the
medullated and non-medullated dorsal root fibers. While most of
the medullated fibers enter the posterior funiculus, all the non-medul-
lated fibers turn at once lateral-ward into the tract of Lissauer in the
apex of the posterior gray column (fig. 1.) Their course up or down
the cord in Lissauer’s tract is very short, the length of one or two
cord segments. There is good reason to believe that they terminate
in the substantia gelatinosa Rolandi (fig. 1), and that this is the nucleus
of reception of these non-medullated afferent fibers.

On the basis of many such
points of similarity between
the protopathic and the non-
medullated afferent fibers one
of us formulated the theory
(Ranson 1914) that the non-
medullated afferent fibers con-
veyed pain and temperature
sensation. It was suggested

1 that the tract of Lissauer, the

Fig. 1. Diagrammatic section of the pe ostan tae labinns eae
spinal cord of the cat at the level of the and the bundle of non-medul-
first lumbar segment. m.f., medullated lated fibers in the ventrally
dorsal root fibers; n.f., non-medullated lying intermediate layer (fig.
dorsal root fibers; L.t., Lissauer’s tract, 1) constituted a mechanism
s.g.-R., substantia gelatinosa Rolandi; 7.1.,

for the reception and conduc-
tion of pain and temperature
sensations. It was further suggested that this assumption ‘“‘does not
exclude the possibility that we are dealing here with a center for
vasomotor and pilomotor control, as suggested by Sano. In fact
these autonomic functions are of necessity closely correlated with
the afferent impulses which find their conscious expression in the form
of sensations of pain, heat and cold. It is thus possible that the
apparatus in question has a double function, serving as a central
autonomic apparatus and for the reception and conduction of pain
and temperature sensations.”

It has been pretty well established that pain is transmitted up the
cord in the anterior part of the lateral funiculus (Van Gehuchten (3),

130 S. W. RANSON AND C. L. VON HESS

Petrén (4), Piltz (5), Bertholet (6), May (7) and Rothman (8) ). Ziehen
(9) however stated that pain is conducted upward in the apex of the
posterior horn; and we were inclined, because of our studies on the
tract of Lissauer, to believe that he was correct. At least it seemed
probable that the apex of the posterior horn was an important part
of the mechanism for the reception, and conduction of the ‘‘nocicep-
tive” (Sherrington (10)) afferent impulses which are represented in
consciousness by the sensation of pain.

With such possibilities in mind we determined to make a careful
study of the conduction within the cord of the nociceptive afferent
impulses. Such a study divides itself into two parts, dealing first,
with the conduction of these impulses to the cortical centers for con-
scious pain, and second, with the conduction of these impulses to the
centers for the autonomic reflexes, notably the vasomotor center. It
will be clear that the paths in the cord for the conduction of these
impulses to the cerebral cortex on the one hand and to the autonomic
centers on the other need not necessarily be the same. The study of
paths in the spinal cord for the nociceptive afferent impulses as deter-
mined by the ordinary tests for pain and as determined by the vaso-
motor reflexes forms two mutually supplementary lines of investi-
gation which have been carried out by us on the same series of animals.
The results of both lines of investigation can, therefore, be directly
compared.

TECHNIQUE

Cats were used throughout this series of experiments. Various
spinal cord lesions were produced, and after recovery each cat was
subjected to two kinds of tests to determine the effect of the lesion,
first, on the conduction of pain; and second, on the character of the
vasomotor reflexes. To avoid possible variations according to age
only full grown cats were used.

The operations were made at the level of the first lumbar segment,
which lies under the spine of the second lumbar vertebra. A very
high grade of asepsis was maintained. In addition to the usual pro-
cedures to secure asepsis, three special precautions were taken. All
skin was excluded from the operative field by sterilized oiled muslin
placed under the laparotomy-sheet and slit to correspond to the skin
incision. The cut edges of the skin and the margins of the slit in the
muslin were held together by skin clips in such a way that no skin could
be seen through the opening in the laparotomy-sheet. The operator’s

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 131

fingers were kept out of the wound, and gauze, needle and sutures were
handled only with instruments. In closing the wound a layer of super-
ficial fascia was drawn over the line of closure in the deep fascia and
sutured to the deep fascia half an inch beyond this line. This layer
of superficial fascia becomes adherent to the deep fascia in a few hours’
and forms a perfect protection for the deeper part of the wound. In
a few cases the cutaneous incision was torn open by the cats, but in
none of the twenty-five did infection develop beneath this layer of im-
bricated superficial fascia.

The dura was exposed and opened in the usual way. Dorsal and
lateral hemisections and sections of the posterior funiculis were made
with a small sharp knife. The apex of the posterior horn was de-
stroyed on bothsides of the cord in six cats by dissecting with a specially
designed probe between the dorsal and lateral funiculi. All lesions
were made in the first lumbar segment, except in four of the earlier
experiments, in which by mistake the lesion was in the second and one
in which it was in the third lumbar segment.

Attempts to produce ventral hemisections were unsuccessful. Each
of the three trials resulted in serious injury to the lumbrosacral cord,
probably due to a disturbance of its blood supply. From all the other
operations the animals recovered promptly and moved about nor-
mally after periods varying from two days to two weeks, except in the
ease of cat A3 in which there was permanently some disturbance of
motion and some atrophy of the hind limbs. The cats gained in weight
‘and were in excellent condition when their vasomotor reflexes were
tested.

The pain sense was tested in the unanaesthetised animals by prick-
ing with a needle and by pinching the skin. These crude methods of
inducing pain were regularly supplemented by the use of the cutaneous
needle electrodes, used by Martin, Porter and Nice (11) in determining
the “sensory threshold for faradic stimulation in man.” Faradic
stimulation of uniform strength through these electrodes gives much
more uniform and dependable pain reactions than can be obtained by
pinching or pricking the skin. The cord lesion was at the level of the
first or second lumbar segments and hence above the level of origin
of all the nerves going to the hind limb. All tests of the hind limbs
were made over the area of distribution of the sciatic nerve which
takes origin below the level of the fifth lumbar segment. A sharp
ery and generalized struggling movements were taken as criteria of
pain.

132 S. W. RANSON AND C. L. VON HESS

After complete recovery of the animal from the lesion and after the
conduction of pain in the cord had been tested the vasomotor reflexes
were studied. Under ether anaesthesia a tracheotomy was performed
and the ether bottle attached. Connections were made as rapidly
as possible for carotid blood pressure and respiratory tracings. Both
sciatic nerves were exposed, ligated and cut distally to the ligature.
The central end was laid bare for some distance above the ligature
and could be handled by the attached thread. On the left side three
brachial nerves, the median, ulnar and internal cutaneous were ex-
posed, ligated together, and cut distally to the ligature and there-
after treated as a single nerve. Care was taken not to stretch the
nerves at any time. When not being stimulated, they were kept
covered with other tissues to prevent drying.

To eliminate passive dilatation of blood vessels in the areas supplied
by the divided nerves the limbs were constricted with heavy cord placed
proximal to the elbow and knee joints. Fluctuations in blood pres-
sure from pressure on the abdomen by flexion of the limbs during stim-
ulation of a nerve were prevented by securely tying the legs to the
animal board.

The stage of ether anaesthesia is of the greatest importance as the
vasomotor reflexes are affected by the slightest overdose, (Porter (12)).
This seems to be especially true of the depressor reflex which is diffi-
cult to obtain under deep anaesthesia. The animal was kept relaxed, -
with regular respiration, with pupils half contracted, and with a brisk
corneal reflex. Any tendency toward cyanosis was prevented because -
asphyxia, being a powerful vaso-constrictor stimulus, would obscure
the results of sciatic stimulation (Sollman and Pilcher (13)).

Faradic stimulation of the central ends of the cut nerves was used
to elicit the reflexes. Standard platinum electrodes were applied at
least half an inch from the cut ends of the nerves, held suspended by
the threads with which they had been ligated. The electrodes were
moved slowly along the nerve during stimulation. The source of the
current was a Stoelting inductorium No. 7090 through the primary
of which passed a constant half ampere current. The primary current
was obtained by shunting the instrument circuit across part of the
resistance of a two ampere one hundred ten volt direct current ‘‘indi-
vidual unit’? system (von Hess (14) ) attached to an ordinary lamp
socket. By actual observation at repeated intervals the current
through the primary coil was found to be very constant. The rate of
interruption of the primary current was made fairly slow, about 25 per

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 133

second, in order to avoid possible overlapping of the currents induced
in the secondary coil (Erlanger 1914). A knife switch in the primary
circuit insures uniform contact. It is believed that with this system
the strength of the induced current for each position of the secondary
coil does not vary greatly from day to day. In this way stimulation
of the same strength can be applied to several nerves in the same cat
and to the nerves of different cats and the results compared.

Curare was used in some of the experiments. It was injected in-
travenously very slowly, until the stimulation of the peripheral end
of a cut nerve gave no response.

After the reflex vasomotor and respiratory responses had been re-
corded the animal was killed, the cord exposed and the level of the
lesion accurately determined. A stretch of cord about seven milli-
meters long containing the lesion was removed and prepared by the
pyridine-silver method (Ranson (2) ) and cut into serial sections. The
serial sections were then studied to determine the extent of the lesion.

LITERATURE ON THE VASOMOTOR REFLEXES

There is a vast number of articles on the conduction of pain in the
spinal cord. Good reviews of this literature have been given by Bertho-
let (6), May (7) and Karplus and Kreidl (15).

Considerable work has been done on the vasomotor reflexes, but
relatively little ts known concerning the anatomical location of the
reflex arcs involved. It is generally stated that there exists in the
brain stem a center for the regulation of blood pressure. It is sup-
posed that this is primarily a vasoconstrictor center and is associated
with the pressor reflexes. Little or nothing is known concerning the
location of a center or centers for the depressor reflexes. In addition
to the “‘principal’”’ vasomotor center in the bulb “‘secondary”’ segmental
centers are located in the thoracic and upper lumbar portions of the
spinal cord, i.e., in those segments of the cord with which the white
rami are associated.

Concerning the paths in the cord along which afferent impulses as-
cend to the vasomotor centers little is known. Dittmar (16) ob-
tained the usual pressor effects from sciatic stimulation after section
of the posterior columns and gray matter of the spinal cord. Sherring-
ton (17) obtained results which led him to conclude that the afferent
path to the vasomotor center lay in the anterior part of the lateral
funiculus and states that Miescher had previously obtained similar

134 S. W. RANSON AND C. L. VON HESS

results. Bikeles (18) also reports some experiments which led him to
the same conclusion. None of these papers give evidence of a thorough-
going investigation of the question and the results are open to numerous
objections such as the testing of the vasomotor reflexes before the
animal has recovered from the effects of the operation, the location of
the lesion too high (9th Th. Seg.) and the failure to distinguish between
pressor and depressor reflexes.

Aside from these experiments little has been done to determine the
effects of cord lesions on the vasomotor reflexes. Stursberg (19)
showed that complete destruction of the cord at the level of the 7th
thoracic segment involved the fibers which produced a coérdination
of the vasoconstrictors of the arm and leg.

The observations of Sherrington (10), Pike (20) and Porter and
Muhlberg (’21) on “spinal shock” have an important bearing on our
work. After transection of the spinal cord at the 8th cervical segment
or during complete cerebral anaemia the blood pressure falls but soon
returns to normal. The vasomotor reflexes are at first abolished
but return in a short time. According to Sherrington the pressor
reflex becomes very good after a few weeks and must now be purely
spinal.

It has recently been shown that stimulation of the central end of
the sciatic produces depressor or pressor reflexes according to the
strength of the stimulus. (Porter (12), Sollman and Pilcher (22),
Martin and Lacey (22).) Weak stimulation gives depressor reflexes,
strong stimulation gives pressor reflexes. Martin and Lacey argue
that the depressor reaction from sciatic stimulation is the more
normal, since the pressor reactions are produced only by excessive
stimulation.

VASOMOTOR REACTIONS IN NORMAL CATS

In order to determine the character of the normal vasomotor reac-
tions we obtained tracings from fourteen normal adult cats. In all
of these, excepting two, stimulation of the sciatic or brachial nerves
with strong currents (s.c. 4 to 6=secondary coil at 4 to 6) gave an
increase in blood pressure, Table I. This pressor response varied
considerably in extent. The cat giving the least response showed
a rise in blood pressure of 11 mm. The greatest reaction obtained
from these normal cats was a rise of 48 mm.

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 135

Stimuli in the depressor range (s.c. 13 to 18) were given to nine
of these normal cats. They all showed a drop in blood pressure vary-
ing in extent from 4 to 22 mm. Hg. With medium strengths of cur-
rent the results varied greatly, some cats showing a rise others a fall
in blood pressure. These results corroborate those of Porter (12),
Sollman and Pilcher (22) and Martin and Lacey (23) that stimulation
of the central ends of afferent spinal nerves gives depressor reflexes
with weak and pressor reflexes with strongstimulation. In the theoreti-
cal discussion of our results we will attempt to give an explanation of

TABLE I.
Normal cats.

CHANGES * IN BLOOD PRESSURE IN MM. HG. FOR INDICATED

Ee cesar POSITIONS OF SECONDARY COIL
PRESSURE
17-18 15-16 13-14 11-12 9-10 7-8 5-6 4-5
WEE or ios cles c's 142 +13
1 Gee 129 sed
1 ee eee 139 +23
1 Se 140 +11
) ee 121 +18 | +14 +48
le 122 — 8} -—11 | — 5} —22) —14} —12] —-11
(bo 2 144 15 +45 +38
lS 2) aes 130 — 9; —12} —18} —18} —14] —11 | +24] +10
[DS A ae 154 —22 —28 | —26 +24 | +24
oo 124 —6;|-4 —10 +14/+4+ 8
Lk Dea, ee 171 — 8 —18 | —22 —16
1007 2 144 —22 —20 +24 | +16
lh 8 36> 3 ae 153 —3|-— 6 +20
ls, DS 178 —10 | —20 +38

*A rise in blood pressure is indicated by +, a fall by —.

this change in direction of the reflexes with increasing strengths of
stimulation.

In our series of normal cats the sciatic nerve was used almost exclu-
sively. Occasionally, however, tests were madeon the brachial nerves—
median, ulnarand internal cutaneous beingstimulated alone or together.
So far as our results show there was no difference in the reflexes obtained
from sciatic and brachial nerves. This agrees with the observations
of Porter and Richardson (24), who found by a long series of experiments
with stimuli of given intensity, applied to the sciatic and brachial
nerves in cats and other animals, that the result did not depend in any

136 S. W. RANSON AND C. L. VON HESS

way on which afferent spinal nerve was stimulated. These experiments
were all with strong stimuli in the pressor range. Weak stimulation
also gives a depressor response which is independent of the individual
nerve stimulated, (Martin and Lacey (23)).

LATERAL HEMISECTIONS

The vasomotor reflexes were tested on five cats which had suffered
a right lateral hemisection of the spinal cord 17 to 63 days before the
tracings were taken. All five had recovered perfectly and had full use
of both hind legs.

Post mortem examination showed the lesion located in the first lum-
bar segment in three of these cats and in the second lumbar segment in
two. In one (cat O7) the lesion when
studied microscopically proved to be a
perfect hemisection. In all the others
part or all of the right anterior funicu-
lus escaped injury (fig. 2). In two the
lesion extended slightly beyond the mid
line posteriorly into the left posterior
funiculus (fig-2). Since the conduction
of pain and the reflex changes in blood
pressure were the same in cat O 7 in
which the hemisection was a perfect

Fig. 2. Diagram of the second one as in the others in which the lesion
lumbar segment of cat O 8. The was incomplete ventrally we conclude
shaded eee shows the part involved that these variations in the lesion have
a tue right lateral hemisection. OLA EE omen

As far as could be determined by pricking the skin and by the use

of cutaneous needle electrodes pain was conducted up the cord equally °

well from all four legs. We had hoped that by the use of the cutaneous
electrodes we could locate accurately the pain threshold and determine
if there were any difference on the two sides of the body below the lesion.
We found that the threshold on the normal limb varied considerably
and these normal variations were greater than any difference between
the two hind limbs or between the front and the hind limbs in cats
with lateral hemisection of the cord. We do not deny that in the
cat some difference may exist in the conduction of pain from the two
sides after lateral hemisection but such difference as may exist is too
small to be readily recognized. In the same way we found it impossible

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 137

TABLE II

Depressor reflexes after lateral hemisection

NN

‘ POSITION E DROP IN BLOOD
OF ile (ile PRESSURE IN MM. HG. : :
CAT SECONDARY NERVE AVERAGE
corn Individual tests

Brachial oe als: 20

16 Right sciatic | 24 22 23

‘06 Left sciatic Hae. er 5 6
I. B. P. 166 Brachial | 20-019 31. 27 24
14 Right sciatic 2119 25 26 23

Left sciatic ey aie era ia 9

Brachial 21 19 20

16) Right sciatic alee 20) 21

0-7 Left sciatic —-4-5 0 5 —]

I. B. P. 166 Brachial 20 18 31 28 24
. 14 Right sciatic | 21 18 27 26 28

Left sciatic Bye 9 7

Brachial 20 19 13 17

12 Right sciatic UCN ksh 7 15

0-8 Left sciatic | 9-3 9 9 6

I. B. P. 142 Brachial 16 16
8 Right sciatic 23 23

Left sciatic 5 5

Brachial 1 15 15 16 15

18 Right sciatic US Eee abe a0) 11

ee Left sciatic i y-8 29 8

I. B. P. 166 Brachial 14 23 22 20
16 Right sciatic 20; 10 12 °8 13

Left sciatic iMcyemalS} 1) as 11

AS Brachial ype ; iy

17 Right sciatic 15 15

eae. 176 Left sciatic 5 5

to demonstrate any loss in sensibility to pain in the hind limbs as com-
pared to the front limbs.

It is clear that the conduction of pain in the cat’s cord takes place
bilaterally. This agrees with the results of other observers on animals
(Bertholet (6) and May (7)). But it is usually stated that, while

138 S. W. RANSON AND C. L. VON HESS

TABLE III

Pressor reflexes after lateral hemisection

POSITION RISE IN BLOOD
CAT Pe ee NERVE Sr a tae Ae
‘ ene Individual tests
0-6 Brachial 5 0 0 2
I B.P. 193 5 Right sciatic 0 0 0
‘oe dl Left sciatic 6 6 14 |* 9g
0-7 Brachial 29X 32X 31X
I. B. P. 166 5 Right sciatic 0 0 0 Oc gay
rte se Left sciatic 9x 10%: J6% 12x
Brachial 12 12
4 Right sciatic 0 0
0-8 Left sciatic 8 8
gee an Brachial bx 0 16 | 10
2 Right sciatic 0 0
Left sciatic 10 10
Brachial 28X ~25x 27X
6 Right sciatic 8X 8X
re Left sciatic 11 1l
ee as Brachial 26x 32x 29x
4 Right sciatic 5 0 -—2 1
Left sciatic 4 0 0 |
que Brachial 22X 31X 27X
I B. P. 176 5 Right sciatic 0 0
i, ee Left sciatic 9 9x9 9

xX indicates that the rise was followed by a considerable fall in blood
pressure.

conduction of pain is bilateral in animals, it is more contralateral than
homolateral. It is usually said that after a hemisection in animals
there is partial analgesia of both hind limbs and that this is most notice-
able on the side opposite the lesion. Karplus and Kreidl (15) have
recently shown that after section of both halves of the cord in the
cat at different levels, pain is still felt in the hind limbs. This would
indicate that in the cat short spinal paths take an especially important
part in pain conduction. We were unable to demonstrate any hypal-
gesia in the hind limbs of our cats with laterally hemisected cords.
Despite these negative results as to the conduction of pain, these cats

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 139

showed characteristic departures from normal in their vasomotor
reactions.

As will be seen by a study of Table II the depressor reactions obtained
from the brachial and right sciatic nerves were normal and approximate-
ly equal. They were of the same average extent as the drops obtained
in normal cats and showed no greater variation than these. In strik-
ing contrast is the reaction obtained from the left sciatic which was as
a rule not more than a third as great as that obtained from the brachial
and right sciatic nerves (fig. 3.). Each nerve was stimulated several
times and the results of the individual tests were very consistent as

Fig. 3. Blood pressure tracings from cat O 8, with right lateral hemisection
of the cord. Base line raised 57 mm. a, drop of 18 mm. Hg. on stimulation of
the left brachial nerves with weak faradic current s.c. 14.; b, drop of 19 mm.
Hg. on stimulation of the right sciatic nerve with weak faradic current s.c. 14,
c, negligible drop on stimulating the left sciatic nerve with weak faradic current
Ree.

will be seen by the table. Each of the five cats gave the same results.
It is, therefore, clear that after right lateral hemisection of the cat’s
spinal cord at the level of the first lumbar segment the depressor re-
actions obtained from the brachial and right sciatic nerves are normal
while those obtained from the left sciatic nerve are greatly reduced.
This would indicate that the conduction in the spinal cord of the affer-
ent impulses producing the depressor reflexes is chiefly contralateral
but to some extent also homolateral.

The changes in the pressor reflexes were not so clear. In every case
the pressor reaction obtained from strong sciatic stimulation, is below
the average pressor response obtained by stimulating the sciatic in
normal cats with the same strength of current, Table III. With one

140 S. W. RANSON AND C. L. VON HESS

exception the pressor response from either sciatic was considerably
less than from the brachial. Since the pressor reflexes from both
sciatic nerves are decreased after lateral hemisection it would seem
that the a‘erent impulses bringing about this rise in blood pressure
must pass up the cord bilaterally.

It will be noted further that the rise from right sciatic stimulation
is less than that from stimulation of the left sciatic. This may be
due to the impulses passing up the cord somewhat better homolaterally
than contralaterally. Or it may be due to the fact that the antagonis-
tic depressor reflex is almost eliminated from the left side, while on the
right side the depressor reflex is normal and tends to overpower the
weakened pressor. It seems probable, therefore, that the afferent
impulses producing a rise in blood pressure are conducted bilaterally
in the cord, and either equally well on both sides or somewhat better
homolaterally. |

Since the best depressor reactions were obtained from the sciatic
on the side of the lesion and the best pressor reactions from the sciatic
on the side opposite the lesion, it is clear that the afferent paths in the
- cord involved in these two reflexes are not the same.

POSTERIOR HEMISECTION

A posterior hemisection was performed on six cats at the level of the
first or second lumbar segments. The autopsy, performed 6 to 84
days later, showed that in two (A 1
and A 2) the lesion was in the first,
in three (O 2, O 3 and A 3) in the
second, and in one (O 4) in the
third lumbar segment. Microscopical
examinations showed that in each
case the lesion involved all of the
posterior funiculus, the posterior part
of the lateral funiculus and all of the

gray matter except the anterior horns
a (fig. 4). In eat O 2 the entire gray

Fig. 4. Diagram of the second substance was destroyed at the level
lumbar segment of cat A 3. The of the lesion.
shaded area shows the part involved
in the posterior hemisection.

So far as could be determined by
careful tests these. cats felt pain
equally well in all four extremities. These lesions in the posterior
half of the cord had had no appreciable effect on the conduction of

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 141

pain impulses from the hind limbs to the cortex. This is in keeping
with the results of most other recorded experiments which place the
pain path in the anterior part of the lateral funiculus. The interest
in these experiments lies in the fact that they show, contrary to the
assumption on which this investigation was started, that the apex of
the posterior horn is at least not the chiefh of pat pain toward the
cerebral cortex. It is also of interest to note that in cat O 2 the gray
matter was completely destroyed at the level of the lesion, showing that
pain is not transmitted upward in the gray matter of the cord.

The vasomotor reactions on these cats were tested from 6 to 84 days
after the operation. It was found that in each animal stimulation

TABLE IV
Reflex changes in blood pressure after posterior hemisection
CHANGES IN BLOOD PRESSURE IN MM. HG.

INITIAL BLOOD BoeTOy Oe

CAT = SECONDARY
eae Bou Brachial Right sciatie Left sciatic
U2 so Ae 138 4.5 +18? —-2 | — 2°
ao ae 152 4.5 +31? + 3? + 3?
(1) Gee 148 4.5 +29 +15 +15
NS hoe eee 148 6.0 ai — —13?
1 ee 160 8.0 + 16? : + 4
ne 160 0.0 +20 — 3 — 2
Pe eyar are eS 3 22 104 6.0 +14 — § — 6
Beit Tee <i. 104 4.0 +12 Oe eae

+ indicates arise.
— indicates a fall.
Small figures indicate the number of individual tests which have been averaged.

of the brachial nerves with strong currents (s.c. 0 to 8) gave normal
pressor responses varying in height from 12 to 31 mm. Hg. (Table IV).
With the same stimulation little or no response was obtained from either
sciatic (fig. 5). The slight reaction obtained was more often a drop
than arise. In one cat a rise of 15 mm. Hg. was obtained from sciatic
stimulation, but in this animal we had failed to tie the legs securely
to the animal board and it is possible that flexion of the thigh on the
abdomen may have caused this rise by increasing intra-abdominal
pressure. Leaving this result to be explained, the other five cats showed
a negligible rise of 3 to 4 mm. Hg. ora drop of 2 to 8 mm. Hg. on sciatic
stimulation with currents that gave strong pressor reactions from the
brachial nerves. It is thus clear that the nociceptive impulses which

142 S. W. RANSON AND C. L. VON HESS

produce a reflex rise in blood pressure travel up the posterior half of
the spinal cord along paths which are not the same as those taken by
the nociceptive impulses toward the cortical centers for conscious pain.

No systematic attempt was made in these cats to develop the de-
pressor reflex with weak stimulation. Such tests, as were made, indi-
cate that the depressor mechanism was either normal or somewhat

i,

My
My 2
iy :
Mh h
Mi ri 3
Hn ernie i

atin
i WN Aan aye
Why AN i PWetetitvani AYR TL ALA AL

D.

Fig. 5. Blood pressure tracings from cat O 3 with posterior hemisection of
the cord. Base line raised 57mm. Upper tracing, rise of 42 mm. Hg. on stimu-
lation of the left ulnar with strong faradic current s.c. 43. Lower tracing, slight
drop on stimulation of the right sciatic with strong faradic current s.c. 4}.

reduced in activity. No exaggerated depressor reactions were ob-
tained in this series.

This question now suggests itself: What part of the posterior half
of the cord is involved in this pressor vasomotor reflex? The following
paragraphs will show that it is not the posterior funiculus.

SECTION OF THE POSTERIOR FUNICULUS

At the level of the second lumbar segment the medial three-fourths
of the posterior funiculus contains all of the dorsal root fibers of the
lower lumbar and sacral roots which have reached this height in the

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 143

cord. The lateral one-fourth at this level contains only fibers from
the second and third lumbar nerves. Hence by cutting the medial
three-fourths of the posterior funiculus in the second lumbar segment,
as was done in cat O 9, it is possible to cut all the fibers in this funiculus
associated with the sciatic nerve and yet run no risk of injuring the
apex of the posterior horn. This was attempted in two cats. In
eat O 10, the apex of the posterior horn was damaged on one side and
the vasomotor reactions showed evi-
dences of this. But in cat O 9 a satis-
factory lesion was obtained, although
in this cat in addition to a section of
the medial three-fourths of the pos-
terior funiculus considerable damage
was done to the gray matter around
the central canal (fig. 6).
As was to be expected from the
results on posterior hemisection both
6, of these cats showed no hypalgesia
Fig. 6. Diagram of the second in either hind limb. The vasomotor
lumbar segment of catO9. Shaded reactions in cat O 9 were entirely
area shows the part involved in the normal (Table V). Since in this cat
section of the medial three-fourths the impulses producing both the pres-
of the posterior funiculus. :
sorand the depressor reactions traveled
up the cord in a normal manner it is clear that the posterior funiculus
takes no part in the conduction of the nociceptive impulses to the
vasomotor centers.

TABLE V

_ Reflex changes in blood pressure after section of the posterior funiculus—Cat 0-9

GanrTO REFLEX CHANGE IN BLOOD PRESSURE IN MM. HG.
INITIAL BLOOD Beane
PRESSURE ein
Brachial Right sciatic Left sciatic
1 Oe 16 —14 —10 —27
6 +31 +28 +20

— indicates a fall.
+ indicates a rise.

144 S. W. RANSON AND C. L. VON HESS

BILATERAL LESIONS IN THE APICES OF THE POSTERIOR HORNS

In the apex of the posterior horn is situated the tract of Lissauer
and the substantia gelatinosa Rolandi. These structures are associated
with the dorsal roots and belong on the afferent side of the nervous
system. But their special function has remained unknown.

Six cats were operated on with the idea of destroying the apex of
the posterior horn on both sides at the level of the first lumbar vertebra.
The autopsies showed the lesion in
each case located in this segment,
but microscopical examination showed
that the lesion was fairly complete in
only two cats A 6 and A 8 (fig. 7).
In each of the other cats the destruc-
tion of the apex of the posterior horn
was incomplete on one or both sides.
In cat A 11 the tract of Lissauer was
destroyed on both sides and a little

if of the substantia gelatinosa on one

Fig. 7. Diagramof the first lum- side. In cats A 7, A 10, and A 12
bar segment of the cord of cat A 8. the tract of Lissauer was entirely
Shaded area shows the part involved destroyed on one side and more or less
in the bilateral lesions of the apices .
of-thetwadienson Nana intact on the other. In these three

cats the substantia gelatinosa suffered
but little injury on either side. In cats A 6, A 8, and A 11 the
bilateral destruction of the apex of the posterior horn was most com-
plete and they gave the most characteristic vasomotor reactions.

As was to be expected from the results of posterior hemisection none
of these six cats showed any hypalgesia of the hind limbs. These
animals tested from 11 to 23 days after the operation showed a very
remarkable variation from the normal vasomotor reaction, Table VI.
With weak stimulation (s.c. 16 or 17) normal depressor reactions were
obtained from both sciatics and the brachial nerves. With increas-
ing strengths of current the reflex drop became greater and greater.
With normal cats the greatest drop was obtained with weak or moderate
stimulation (s.c. 17 to 12) and as the current was increased beyond
this optimum depressor stimulus the drop became smaller and soon
gave place to arise. The greatest drop obtained from a normal cat was
28 mm. Hg. In these cats with lesions in the apices of the posterior
horn the drop continued to increase as the strength of the stimulus was

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 145

TABLE VI
Reflex changes in blood pressure after bilateral lesions in the apices of the posterior
horns
CHANGES IN BLOOD PRESSURE IN MM. HG.
INITIAL POSITION OF
CAT BLOOD SECONDARY
SSS Ure CON Brachial Right sciatic Left sciatic
Siac sce 158 17 — 5 —l1 —4
—12 —23 =—10
—46 —47 —35
4 —29 —40 —26
ep eds. 2-2. 2 136 16 — 9 — 8 0
«4 +12 +10 Pal
—44 —32 —30
Se 5 140 17 —11 —13 —12
4 —44 —52 —42
/\= | RA 136 17 —20 = & —10
9 —21 — 26 —15
8 — 4 — 10
*6 {+28 +16 +25
\-—38 —38 —31
7 138 17 —15 —18 —14
12 —20 —43 —35
9 —33 —44 —31
6 —60 —58 —36
a 130 17 —10 —22 — 4
9 —20 — 28 —27
8 =i —27 —21
6 +43 +28 +26

+ indicates a rise.
— indicates a fall.
* A-7 and A-10 show rise followed by marked drop.

increased until with strong stimuli (s.c. 4 to 9) which would normally
produce strong pressor responses the fall in blood pressure became two
or three times as great as any that was obtained from normal cats
(figs. 8,9, 10). From the right sciatic a drop of 40 mm. Hg. was ob-
tained with cat A 6, a drop of 52 mm. Hg. with cat A 8, and one of 58
mm. Hg. with cat A 11. It should be noted that these reactions are
two or three times as great as can be obtained by stimulating the same
nerves in normal cats and were obtained by strength of stimulation

146 S. W. RANSON AND C. L. VON HESS

Fig. 8. Blood pressure tracing from cat A 11, with bilateral lesions in the
apices of the posterior horns. Base line raised 26 mm. Drop of 58 mm. Hg.
on stimulation of the right sciatic with strong faradic current s.c. 6.

Fig. 9. Blood pressure tracing from cat A 8 with bilateral lesions in the
apices of the posterior horns. Base line raised 33 mm. Drop of 52 mm. Hg.
on stimulation of the right sciatic nerve with strong faradic current s.c. 4.

Fig. 10. Blood pressure tracing from cat A 7 with bilateral lesions, in the
apices of the posterior horns. Base line raised 42 mm. Slight rise followed by
drop of 44 mm. Hg. on stimulation of the left brachial nerves with strong

faradic current s.c.4.

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 147

which would normally produce the greatest pressor reaction. No
pressor responses could be obtained from these three cats. It will be
noted that in all these cats the brachial nerves gave the same response
as the sciatics. This is difficult to explain but an attempt at an ex-
planation will be made later.

The cats in which the tract of Lissauer was not completely destroyed
on both sides (A 7, A 10, A 12) do not give clear cut results (Table VI) ;
but each shows a tendency for the depressor reaction to predominate.
In cats A 7 and A 10 strong stimulation (s.c. 4 to 6) gave a moderate
rise in blood pressure followed by a much greater fall. The character
of such a reaction is shown in figure 10. The drop carried the blood
pressure two or three times as far below the original level as the rise
carried it above that level.

The abruptness of the drop in cats A 7, A 10, and A 11 may indicate
that vagus inhibition of the heart plays an important part in this re-
action. Yet this cannot be the chief cause of the fall in blood pressure
since in cats A 6 and A 8 section of both vagi did not affect the reac-
tion. In fact the tracing from cat A 8 (fig. 9) was taken after both
vagi had been divided.

In cat A 12 the lesions in the tracts of Lissauer were not complete
bilaterally. Well marked pressor responses were obtained from sciatic
stimulation but a greater one was obtained from the brachial. The
results would be easily explained on the assumption that the pressor
‘mpulses pass up in the tract of Lissauer and that the partial lesions of
these tracts in cat A 12 had partially shut off these impulses arising
in the sciatics from reaching the vasomotor centers. In fact the re-
sults from cat A 12, though unlike those from the other five of the
series are more easily explained than the others.

_ The question arises, why were not pressor responses obtained from
the brachial nerves in the other cats of this series? Since the brachial
nerves are connected with the cord above the level of the lesion and
the lesion was below the segments from which most if not all of the
vasomotor fibers leave the cord, it is difficult to understand why the
vasomotor responses from the brachial nerves should have been altered
at all. One possible explanation would be that the cutting off of
the impulses normally ascending in the apex of the posterior horn from
all of the nerves below the first lumbar so lowered the tone of the
series of short relays by which the impulses are transmitted that the
pathway may become less conductive than normal even to impulses
from the brachial nerves. But this explanation is not altogether

148 S. W. RANSON AND C.'L. VON HESS

satisfactory and the most we can say at present is that lesions of the
apices of the posterior horns in upper lumbar segments of the cord
cause a complete reversal of the vasomotor responses from strong
stimulation of the spinal nerves. Other experiments along this line
are In progress.

EFFECT OF CORD LESIONS ON REFLEX CHANGES IN RESPIRATION

In the six cats with bilateral lesions of the apices of the posterior
horns, stimulation of the sciatic caused the same increase in rate and
depth of respiration as in normal cats. Since the pressor responses
were eliminated in five of these cats it would seem that the afferent
paths to the respiratory centers are not the same as those to the pressor
vasomotor centers. Section of the medial three-fourths of the posterior
funiculus also had no effect on reflex changes in respiration. We con-
clude that the path to the respiratory center is neither in the posterior
funiculus nor in the apex of the posterior horn.

In the six cases of posterior hemisection the reflex changes in respira- °

tion from sciatic stimulation were greatly reduced though not entirely
eliminated. This indicates that the impulses leading to reflex changes
in respiration pass up in part at least in the posterior half of the lateral
funiculus.

RELATION OF PAIN TO THE VASOMOTOR REFLEXES

All of the cats whose vasomotor reflexes were tested felt pain equally
well in all four extremities. So far as could be determined none of them
showed any hypalgesia. The complete elimination of the pressor
reflex in many of these cats shows that within the cord the pressor
path and the path for conscious pain are not the same. On the other
hand the depressor reflex was not eliminated in any of these cats. This
path was shown to be chiefly crossed and is probably located in the
lateral funiculus. Although we could not demonstrate that the pain
path was chiefly crossed in cats this has been shown to be the case
in most other mammals andin man. It may be that the more accurate
blood pressure tracings showed a difference in the two sides which the
more crude pain tests failed to show. It is therefore possible that
the afferent impulses which are felt as pain and those which produce
the depressor vasomotor reactions travel the same paths in the spinal
cord. .

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 149

THEORETICAL CONSIDERATIONS

1. Vasomotor reflex arcs. It is difficult at first to see how the alter-
ations in the vasomotor reflexes which we have described can be ac-
counted for. It is possible however to formulate a conception of the
vasomotor reflexes which would explain all our observations. This
we present in a purely tentative way as a working hypothesis.

It has been shown by Pike (20) and Sherrington (10) that after
section of the cervical cord and during cerebral anaemia the vascular
tone and pressor reflexes, which are at first lost, quickly return. This
in our mind throws considerable doubt on the importance of the “‘chief”’
vasoconstrictor or pressor center in the medulla and in our theory we
ignore the brain entirely, although nothing in our theory would be
inconsistent with the afferent paths passing by way of the medulla. It
is well known that the vasomotor impulses ultimately leave the cord
by way of the thoracic and upper lumbar spinal nerves. We shall
consider the paths by which impulses from the sciatic ascend to the
vasomotor centers without committing ourselves as to the location of
these centers.

We assume that in the vasomotor reflexes there are two separate
paths from the sciatic nerve to the efferent vasomotor neurones; and
for brevity we will speak of a pressor path and a depressor path accord-
ing to the vascular responses produced by the impulses which they
carry. That these are separate paths seems to be true, since after
lesions of the spinal cord one type of reflex may be decreased or de-
stroyed while the other is normal or augmented.

The pressor path, according to this hypothesis, is in the apex of the
posterior horn and is either equally bilateral or chiefly homolateral.
This would account for the loss of the pressor reflex from sciatic stim-
ulation after posterior hemisection and after bilateral injury to the
apices of the posterior horns.

The depressor path is not involved in section of the posterior funic-
ulus nor in bilateral injury to the apices of the posterior horn. It
must therefore be ventrolateral to these structures probably in the
lateral funiculus. Some damage seems to have been done in this path to
posterior hemisection but as to this our results are not clear. The
results obtained on cats with lateral hemisections showed that the
depressor path was chiefly crossed.

The impulses traveling along the two paths are mutually antagonistic
and in normal animals vasomotor reactions represent a balance between

150 S. W. RANSON AND C. L. VON HESS

them. The depressor path has a low threshold; and by it weak stimuli
are able to reach the efferent vasomotor neurones. The pressor path
has a high threshold since it is composed of the short fibers and many
synapses of the apex of the posterior horn. With weak stimuli the
impulses pass up the depressor path only, resulting in a drop in blood
pressure. With stimuli of medium intensity they pass up both paths;
stimulation and inhibition balance each other with little or no change
in blood pressure. With strong stimuli they still pass up both paths
but the strong pressor overcomes the depressor impulses, resulting
in vasoconstriction and increased blood pressure.

After bilateral lesion of the apices of the posterior horn the pressor
path is so impaired that with strong stimulation the inhibitory impulses
along the depressor path predominate and produce marked depression.
This would account for the marked depression obtained in this series
from sciatic stimulation. But similar depressions were obtained from
the brachial nerves in this set of experiments, although no cord lesion
was interposed between the origin of these nerves and the vasomotor
efferent neurones in the thoracic cord. In order to explain this result
it is necessary to assume that the normally high resistance of the pressor
path has been increased above the lesion by the cutting off of the im-
pulses normally ascending in this path from the lower half of the body.

2. The intraspinal conduction of nociceptive afferent impulses. It is
generally admitted that painful afferent impulses are associated with
uncontrollable nervous discharge through the autonomic system, one
expression of which is the vasomotor reflex (Sherrington in Schaffer’s
Text of Physiology). Obviously only such part of these impulses as
ascend to the cortex are responsible for the production of pain. Hence
‘nociceptive’ is a better general term than ‘‘painful’’ and includes
the afferent impulses producing the pain reflexes as well as those pro-
ducing conscious pain. The apex of the posterior horn is involved in
the conduction of the nociceptive impulses producing reflex vasocon-
striction but not in the upward conduction of those which are felt as
pain. It seems probable to us that the tract of Lissauer and the sub-
stantia gelatinosa which are well developed in all vertebrates form
the primitive mechanism for the reception and intersegmental con-
duction of nociceptive afferent impulses. In lower vertebrates there is
no spinothalamic path for the conduction of painful afferent impulses
toward the cerebrum, but these lower forms must have some mechanism
for the reception and intersegmental conduction of nociceptive im-
pulses. The tract of Lissauer and the substantia gelatinosa are present

.

SPINAL SENSORY PATHS FOR PAIN AND VASOMOTOR STIMULI 151

in well developed form in these lower vertebrates and their structure
is such that they are especially adapted for intersegmental conduction.
Even in the mammal they retain the function of conduction of some
nociceptive impulses as is evidenced by their relation to the pressor reflex.
With the development of the cerebral cortex in the higher vertebrates
this mechanism in the apex of the posterior horn would still retain,
according to this theory, its function of reception of all nociceptive
impulses. Connected with it there would be developed a new path-
way to the cerebral cortex, the spinothalamic, which would carry the
pain impulses from the primitive nociceptive mechanism to the cortical
centers for conscious pain..

CONCLUSIONS

1. Lateral hemisection in the upper part of the lumbar cord results
in a great reduction of the depressor reaction obtained from stimu-
lation of the sciatic on the side opposite the lesion. The depressor
reactions from the sciatic on the side of the lesion and from the brachial
nerves are normal. There is considerable reduction in the pressor re-
actions from both sciatics, the greater reduction being in the pressor
reactions from the sciatic on the side of the lesion.

2. Posterior hemisection in the upper part of the lumbar cord almost
entirely obliterates the pressor reflex from sciatic stimulation but has
less effect on the depressor reflex.

3. Section of the posterior funiculus in the lumbar region is without
influence on the vasomotor reflexes.

4. Bilateral lesions in the apices of the posterior horns obliterates the
pressor reflex. The depressor reflex from weak stimulation is normal
but becomes greatly exaggerated when strong stimuli are used. This

-amounts to a complete reversal of the normal reaction. In this group

of experiments the same results were obtained from brachial and sciatic
stimulation.

5. A comparison of the effects of the same lesions on the conduction
of pain and on the reflex changes in respiration and blood pressure shows
that the afferent spinal path involved in the pressor reflex cannot be
the same as the path for conscious pain nor the same as the afferent
path to the respiratory center. On the other hand it is possible that
the afferent spinal path involved in the depressor reflex is the same
as that involved in reflex changes in respiration, and in the conduction
of pain.

152 S. W. RANSON AND C. L. VON HESS

6. None of the lesions (lateral hemisection, posterior hemisection,
section of the posterior funiculus, destruction of the apices of the pos-
terior horn) had any noticeable effect on the conduction of pain in these
cats.

7. It seems probable that there are two separate afferent spinal paths

involved in the vasomotor reflexes: the pressor path, equally bilateral -

or chiefly homolateral in the apices of the posterior horn; and the
depressor path chiefly crossed and located in the lateral funiculus.

It is a pleasure to acknowledge our indebtedness to Professors Mc-
Guigan and Hoskins for many helpful suggestions which we have re-
ceived from them during the progress of this investigation.

BIBLIOGRAPHY

(1) Heap AND SHERREN: Brain, 1905, xxvii, 117; HEAD AND THompson: Brain,
1906, xxix, 537.

(2) Ranson: Amer. Journ. Anat. 1911 xii, 67; Journ. Comp. Neur. 1912,
xxii, 159: Journ. Comp. Neur., 1913, xxiii, 259; Am. Journ. Anat., 1914,
Aly jo, Ole

(3) VAN GEHUCHTEN, 1901, cited after Bertholet.

(4) Perrin, 1902, cited after Bertholet.

(5) Prutz, 1904, cited after Bertholet.

(6) BrertHouet: Le Névraxe, 1906, vii, 285.

(7) May: Brain, 1906, xxix, 742.

(8) RorumMan, 1906, cited after May.

(9) ZimHEN: Nervensystem—Handbuch der Anatomie des Menschen. Karl
von Bardeleben, Jena, 1899.

(10) SHERRINGTON: The integrative action of the nervous system. New York,
1906.

(11) Martin, Porter anp Nicer: Psych. Review, 1913, xx, 196.

(12) Porter: This Journal, 1910, xxvii, 276.

(13) Som~tMANN AND PitcHeER: This Journal, 1911, xxix, 100.

(14) von Hess: Science, 1914, xl, 566.

(15) Karpius AND Kremu: Pfliiger’s Archiv, 1914, elviii, 275.

(16) Dirrmar, 1873, cited after Hill; Schifer’s Physiology, ii, p. 137.

(17) SHERRINGTON: Brain, 1886, ix, 342.

(18) BrkevEs: Zentralbl. f. Physiol., 1899, xii, 346.

(19) SturssperG: Deutsch. Arch. f. klin. Med., 1911, civ, 262.

(20) Pike: This Journal, 1908, xxi, 359; 1909, xxiv, 124; 1912, xxx, 436.

(21) Porrrer AND Muuuserc: This Journal, 1900, iv, 334.

(22) SOLLMANN AND PitcHeER: This Journal. 1910 xxvi, 233.

(23) Martin AND Lacry: This Journal, 1914, xxxiii, 212.

(24) PorTER AND RicHARDSON: This Journal, 1908, xxiii, 131.

RHYTHMICAL CONTRACTION OF THE SKELETAL MUSCLE
TISSUE OBSERVED IN TISSUE CULTURES

MARGARET REED LEWIS

Carnegie Institute at Washington
Received for publication May 22, 1915

The following observations upon the skeletal muscles of chick em-
bryos in tissue cultures show that isolated fibers and myoblasts possess
the property of automatic, rhythmical contractility, in the medium
employed.

Among a number of preparations of myotomes of the tadpole ex-
planted into frog lymph, Harrison (3) noticed, in a few instances when
the myotome was thin, the differentiation within the myotome of
the primitive myoblast into a cross striated fiber. He states that
such fibers do not contract unless a part of the neural canal remains
attached to the explanted myotome. This result of Harrison may be
due to the fact that lymph was used as a medium in consequence of
which the necessary salts to stimulate the muscle fiber may not have
been present, or it may have been due to the fact that Harrison was
dealing only with the myoblast within the thick explanted myotome and
not with myoblasts and muscle fibers, which had grown out into the
surrounding medium. The following observations show that in the
case of the chick skeletal muscle, it may contract rhythmically when
no nerve tissue is present. It is interesting to note that Harrison ob-
‘served the primitive myoblast differentiate into the striated muscle
fiber, since Champy (2) finds that the striated muscle fiber of the
rabbit, when explanted into plasma, dedifferentiates into an indifferent
mass of cells, through which the scattered remains of the myofibrillae
can be observed. Champy does not state whether the muscle tissue
contracted rhythmically at any time during this process.

Burrows (1), described the rhythmical contraction of an isolated heart
muscle cell at the periphery of the growth of a tissue culture in a plasma
medium. Burrows states also that a group of heart muscle cells,
which had divided and differentiated from cells of the explanted piece,
but which had become entirely separated from the other growth and

153

154 MARGARET REED LEWIS

also from the explanted piece, continued to beat rhythmically, although
with a changed rhythm from that of the old piece. The explanted
piece of embryo chick heart may contain ganglion cells from which
nerve fibers may grow, but such nerve fibers are so easily distinguished
among or over the cells of the new growth, that it is possible to deter-
mine at a glance whether the preparation contains nerve fibers or not.
In the case of the isolated heart muscle cell, which Burrows observed
in rhythmical contraction, it is needless to state, there was no possi-
bility of nervous influence. These observations of Burrows determine
that it is possible for a heart muscle cell to contract rhythmically when
entirely free from the nervous system and subject only to the stimulus
of the surrounding plasma medium.

Rhythmical contraction of muscle tissue independent of nerve in-
fluence has been demonstrated by means of numerous physiological
experiments among which some of the well known observations are
those of Howell, Loeb, Stiles and Lingle.

Howell (4) found that

a strip of vena cava from the heart of the terrapin may be kept in continuous
rhythmic pulsation during forty-eight hours or more when immersed in baths
containing only the inorganic salts, sodium chloride, potassium chloride and
calcium chloride. Under normal conditions the stimulus that leads to a heart
contraction is dependent upon the presence of calcium compounds in the liquids
of the heart, but for rhythmic contractions and relaxations a certain propor-
tion of potassium compounds is also necessary. The sodium chloride seems to
be essential only in preserving the osmotic relations between the tissue and the
surrounding liquid.

Loeb (9), observed that the skeletal muscle of a frog contracted
rhythmically when placed in a 0.7 per cent solution of sodium chloride.
Loeb concluded from his observations that it is only the antagonistic
action of the calcium and magnesium salts of the blood, which prevent
the skeletal muscles from contracting rhythmically.

The solutions of Na-salts produce rhythmical contractions only if the muscle
cells contain Ca-ions in sufficient number. As soon as there is a lack of Ca-ions
in the tissue the Na-ions are no longer able to cause rhythmical contractions.
On the other hand, if we add Ca-salts in sufficient quantity to the NaCl solu-
tion, it will no longer cause rhythmical contraction in a fresh muscle of the frog.
It therefore looks as if the presence of a certain quantity of Na-ions caused con-
tractions, but if the quantity of the Na-ions becomes too great in proportion
to the Ca-ions, the muscle loses its irritability. On the other band if there are
too many Ca-ions present the rhythmical contractions become also impossible.—
Loeb.

RHYTHMICAL CONTRACTION OF SKELETAL MUSCLE Me

OV

Stiles (10) showed that rhythmic contraction can be set up in the
smooth muscle of the oesophagus of the frog. His conclusions cor-
respond more with Howell’s than with those of Loeb, i.e., he finds that
the calcium salts have a direct action in the stimulation of the smooth
muscle cells to contraction.

While a strip of smooth muscle in a 0.7 per cent solution of sodium chloride
may contract irregularly for about an hour, in a Ringer’s solution it will con-
tract rhythmically for many hours. When the strip comes to rest in Ringer’s
solution, it can be again stimulated by successive quantities of calcium until
the calcium chloride reaches 0.1 per cent when the calcium becomes depressive
if not toxic. The calcium and potassium are both necessary to maintain rhythmic
contraction for any length of time, but the sodium is the necessary salt, although
the muscle frequently fails to contract at all in the sodium solution until the
calcium is added.

Lingle (7) (8), described the rhythmical contractions of the isolated
ventricle of a tortoise heart which does not beat in the blood when
placed in pure sodium chloride solution. If the ventricle remains
in a sodium chloride solution it ceases to beat. The pure solution
of sodium chloride acts as a poison. A small amount of calcium how-
ever acts antagonistically to the injurious action of the sodium chlor-
ide. Calcium cannot start the contraction of the ventricle, but it is
necessary to maintain the rhythmical contraction started by the
sodium chloride.

The above together with other later work shows that it is possible
for any muscle (heart, smooth or skeletal) to contract rhythmically,
provided the necessary salts to stimulate the muscle to contraction
are present in the tissue itself and in the surrounding medium.

In the following observations tissue cultures were made from pieces
of the legs of 4, 5, 6, 7, 8, 9 and 10 day chick embryos, explanted into
‘Locke’s solution +0.5 per cent dextrose +10 per cent boullion accord-
ing to the method of Lewis and Lewis (5) (6).!

The growth of cells is rapid and luxuriant, anywhere from one to
fifty muscle fibers may have grown out radially from the explanted
piece at the end of 24 hours (fig. 1). Also numerous isolated fibers

1 Most of the failures to obtain growth in tissue cultures in Locke’s solution
by other observers is probably due to some error in their technic, as I have
found it possible to obtain as Jarge a growth in Locke’s solution as in plasma.
Dr. C. C. Macklin has shown that certain attempts to obtain growth have proven
unsuccessful because evaporation of the medium has caused too great a concen-
tration of the medium (unpublished).

156 MARGARET REED LEWIS

and many primitive myoblasts are scattered throughout the new
growth. There is an abundant growth of connective tissue and primi-
tive mesenchyme cells. These cells are larger than the myoblasts
and contain a larger nucleus. They extend further away from the
explanted piece than most of the muscle cells. The connective tissue
cells divide rapidly by mitosis and in some cultures the greater num-
ber of the connective tissue cells are in some phase of mitosis. Amitosis
has never been observed in these cultures but mitotic figures are to
be seen on all sides, frequently as many as one mitotic figure to every
four or five resting cells. An attempt has been made to count the
mitotic figures present in the growth at a given time and frequently
more than 100 such figures have been present in one preparation at
the same time.

Numerous myoblasts are always present in the growth which develops
from an explanted piece of embryo chick leg of any of the above ages
(4, 5, 6, 7, 8, 9 and 10 days) and the greater number of the fibers are
less differentiated than is normal for a chick embryo of the age from
which the explantation was made. Certain cultures have been kept
in a healthy condition (many mitotic figures) for 10 and 12 days by
frequently washing with the culture medium and in these cultures
the muscle fibers redifferentiate somewhat, but a typical striated
muscle fiber has never been observed in any growth, not even in that
from a more adult chick embryo.

Nerve fibers do not grow out from the explanted piece of chick em-
bryo leg and a nerve fiber has never been observed in any of the tissue
cultures used for this study of skeletal muscles.

Mitotic figures are not present in the multinucleated muscle fibers
(fig. 2) but the isolated myoblasts which contain a single nucleus fre-
quently divide by mitosis. This division results in two typical
myoblasts and not in an indifferent tissue such as that described by
Champy.

The muscle fiber has the appearance of a gigantic cell with from 2 or
3 to as many as 60 or more nuclei in a spread out protoplasmic end of
the muscle fiber (fig. 2). In a few instances such a protoplasmic end
has split up into numerous myoblasts with each a single nucleus. These
myoblasts migrate away and divide by mitosis.

A description of the growth and differentiation of the skeletal mus-
cle in tissue culture will appear in a separate paper.

Fig. 2

Fig. 1. Muscle fibers and connective tissue of a 72-hour growth from an 8-
day chick embryo wing in Locke’s solution + 0.5 per cent dextrose + trace of
yolk. Zenker fixation. Mallory stain. Oc. 4. Obj. 16.

Fig. 2. Protoplasmic end of a single muscle fiber, which contains many
nuclei and primitive myofibrillae. Same culture as figure 1. Oc. 4. Obj. 4.

157

158 MARGARET REED LEWIS

Fig. 3. Isolated muscle fiber, which contracted 120 times a minute for 8 hours.
A 56-hour growth from a 9-day chick embryo leg in Locke’s solution + 0.5 per
cent dextrose + 10 per cent chicken bouillon. Zenker fixation. Mallory stain.
Oe. 4). (Obj. 4.

Fig. 4. Two living myoblasts, which contracted slowly (cell 1 once in 15
seconds, cell 2 once in 23 seconds) for one-half hour. Five-day growth from a
6-day chick embryo leg in Locke’s solution + 0.5 per cent dextrose + 10 per cent
chicken bouillon. Oc. 4. Obj. 4.

RHYTHMICAL CONTRACTION OF SKELETAL MUSCLE 159

RHYTHMICAL CONTRACTION

Among the numerous muscle fibers present in the growth in a tissue
culture, one may occasionally contract rhythmically. The time
interval of the rhythmical contraction has been different for each
muscle fiber as well as for each myoblast that has so far been observed
to contract. Some fibers contract very rapidly, i.e., as often as 120
times a minute; others contract once in 2 or 3 seconds; others not of-
tener than once in 15 or 20 seconds; and some contract very infrequently,
i.e., only as often as once in from 1 to 5 or 10 minutes.

Figure 3 is a photograph of an isolated multinuclear muscle fiber,
which was not only free from the explanted piece, but also free from
other muscle fibers. This fiber was first observed in rhythmical con-
traction when the culture was 48 hours old and it continued in rhythmi-
cal contraction (120 times a minute) throughout the day, at the end
of which time it was fixed and stained. The stain used was Mal-
lory’s connective tissue stain and the primitive myofibrillae can be
seen to extend from one end of the fiber to the other. The growth
was from the leg of a 9-day chick embryo and was 56 hours old when
fixed. The muscle fibers of the explanted piece were without doubt
striated at the time of explantation and therefore it seems reasonable
to suppose that this fiber must be either the result of dedifferentiation
of the striated muscle fiber (Champy) or of the differentiation of a
myoblast.

Figure 4 is a drawing of two living myoblasts during the resting
period of a rhythimeal contraction which continued for over half
an hour after it was discovered in a 5 days’ growth from an explanted
piece of a 6-day chick embryo. The time interval of contraction was
different for the two cells. Cell 1 contracted every 15 seconds and
cell 2 every 23 seconds.

Some well developed muscle fibers with a few cross striated myo-
fibrillae have been observed to contract, but more frequently the
property of rhythmical contraction is exhibited by less well developed
muscle fibers such as figures 2 and 3 or by the isolated myoblast itself.

The ability of the muscle fiber to contract is not determined by the
age of the culture for rhythmical contraction has been observed in cul-
tures from 24 hours to 5 days old. Also the property of contractility
is never general for any one culture, i.e., while one or two myoblasts

160 MARGARET REED LEWIS

or muscle fibers of a given culture may contract, the other 50 or more
fibers appear inactive or else the time interval of contraction is so
slow that it is not observed.

EXPERIMENTAL WORK

An attempt has been made to stimulate the muscle fibers to con-
tract by a change in their environment such as has been shown to
stimulate rhythmical contractions by Howell, Loeb, Stiles and others.

The cover slip on the underside of which was the growth to be experi-
mented upon was removed from the vaseline ring and a drop of a differ-
ent solution at 39° C. was placed upon the growth. The cover slip
was then replaced on a hollow ground slide and the growth was studied
under the microscope in a warm chamber with the following results.

Sodium chloride. A drop of 0.9 per cent solution of NaCl in distilled
water placed upon the growth in a few instances resulted in the im-
mediate contraction of one or more of the muscle fibers. This contrac-
tion was repeated a few times and then ceased and did not occur again.
In no case did the solution of NaCl stimulate a rhythmical contraction
of all the muscle fibers or cells of the new growth. The muscle cell
or fiber, which had previously been contracting, continued to con-
tract for a short time and then ceased. It is possible that the NaCl
penetrated the extremely thin growth very rapidly and had a toxic
action upon the cells. If the NaCl solution was replaced by one which
contained CaCl, as well as NaCl no change occurred. When the
erowth was again bathed in the medium, which had been used for the
explantation (Locke’s sol. + 0.5 per cent dex. + 10 per cent bouillon)
the growth sometimes recovered and mitosis continued, but more
frequently the growth became degenerate.

Calcium chloride. A drop of .025 per cent CaCl, resulted in the
degeneration of the muscle fibers and other cells. A drop of Locke’s
solution had no effect, i.e., the muscle fiber continued to contract
but the inactive fibers were not stimulated to contract. A drop of
Locke’s solution, which contained an increased amount of CaCle either
had no effect or resulted in the degeneration of the growth.

Dilution of the medium. <A slight dilution of the medium used for
the explantation had no definite action upon the contractility of the
muscle tissue. A dilution of 10 per cent or more frequently stimulated
certain of the muscle fibers and cells to contract, although it also
resulted in the degeneration of many of the muscle fibers. The degen-
eration of the muscle fibers in this case was quite different from that

ES ee

RHYTHMICAL CONTRACTION OF SKELETAL MUSCLE 161

which usually takes place due to other causes. In this case the entire
muscle fiber flowed from one extremity to the other and left no trace
of its previous elongated structure. The mass of coarsely granular
protoplasm which resulted from the rounding up of the muscle fiber
later disintegrated.

Although the method used to stimulate the muscle tissue in these
tissue culture growths to rhythmical contraction have so far not been
successful, due probably to the fact that the method used was not
sufficiently delicate, nevertheless the fact remains that by some so far
unknown change, either in the tissue itself or in the surrounding medium,
the myoblast and also the muscle fiber of the skeletal muscle of the chick
embryo possess the property of rhythmical contractility when entirely
free from nerve influences and subject only to the stimulus, which
emanates from its environment.

It is interesting to note that the muscle cells inherit this property
from cell to cell even when several generations of cells have taken
place in a simple Locke’s solution. The cytoplasm of the myoblast
is quite different from that of the surrounding connective tissue cells.
It is less thinly spread out and has a different refraction from the other
cells.

CONCLUSION

The skeletal muscle tissue which grows out from an explantation of a
piece of embryo chick leg into Locke’s solution +0.5 per cent dextrose
+10 per cent chicken bouillon, may contract rhythmically in the
absence of nervous tissue.

LITERATURE LIST

(1) Burrows: Miinch. med. Wochensch., 1912, xxvii, 1473.

(2) Cuampy: Arch. de Zool. Exper., 1913, liii, 42.

(3) Harrison: Jour. of Exp. Zool., 1910, ix, 787.

(4) Howet.: This Journal, 1898, ii, 47.

(5) Lewis anp Lewis: Johns Hopkins Hosp. Bull., 1911, xxii, 126.
(6) Lewis anp Lewis: Amer. Jour. of Anat., 1915, xvii, 339.

(7) Linete: This Journal, 1900, iv, 265.

(8) Linete: This Journal, 1902, viii, 75.

(9) Lorn: Festschrift fiir Professor Fick, Wiirzburg, 1899.

(10) Srizes: This Journal, 1901, v, 338.

THE AMERICAN
JOURNAL OF PHYSIOLOGY

VOL. 38 AUGUST 1, 1915 No. 2

THE PRESERVATION OF THE LIFE OF THE FROG’S EGG
AND THE INITIATION OF DEVELOPMENT, BY
INCREASE IN PERMEABILITY

J. F. McCLENDON
Contribution from the Physiological Laboratory of the University of Minnesota}

Received for publication April 29, 1915

If the frog’s egg is removed from the body of the female and placed
in Water it soon dies and disintegrates. This is true whether the egg
is removed from the ovary, the body cavity or the oviduct. Even
though the egg were laid in the normal fashion, it would die if not
stimulated by a spermatozoon or in some other manner. Bataillon?*
was able to cause unfertilized frog’s eggs to segment by pricking them
with a needle. But I (1) found that a much more convenient and ef-
ficient method of causing them to segment is to stimulate them with
the ordinary 110-volt alternating current from carbon or platinum
electrodes placed about two inches apart in the water containing the
eggs, for about one second. Bataillon thinks the introduction of a for-
eign protein, such as is used to induce anaphylaxis, is necessary for the
development of these eggs. I found the jelly in which the eggs are
imbedded to be permeable to hemoglobin, and it is possible to think
of protein diffusing in without the needle prick. But eggs removed so
carefully that there is no possibility of foreign protein gaining access to
them, segment when stimulated electrically. It is certain that no for-
eign protein is necessary for the beginning segmentation of this egg,

1 The nephelometer used in this work was bought with a grant from the Re-
search Fund of the Graduate School.
2 Bataillon: Comptes Rendus de la Société de Biologie, 1911, Ixx, 562.

163

164 J. F. McCLENDON

though it may be necessary for its later development. In the leopard
frog, Rana pipiens, this segmentation usually stops before the first cy-
toplasmic division is completed, although nuclear divisions may con-
tinue for several days.

The initiation of cleavage in the frog’s egg by the simple prick of a
needle, which I have observed in American frogs, might be interpreted
as increased permeability. Since I have shown that-electrical stimu-
lation of muscle causes increased permeability (2), electrical stimulation
of these eggs might be supposed to have the same effect. In order to
determine this point, I made some preliminary experiments last year,
but owing to difficulty in obtaining material, could offer but meagre
data to prove the point (3). This year the leopard frogs bred in the
laboratory, and furnished abundant material.

A large number of females of the same size, and that were clasped by
males, were placed in a tank with two inches of water. As soon as a
female deposited her eggs, these were carefully removed to distilled
water. Two females were opened and their eggs placed in separate
dishes of distilled water, one as a control and the other for electrical
stimulation. |

In removing the eggs, the female was washed free from sperm, the
head cut off, most of the viscera removed and the blood washed
out. The distended oviducts were carefully removed and the blood
washed out of the vessels in their walls. A hole was made in each
oviduct and it was dragged across the floor of a dry glass dish, leaving
a trail of adhesive eggs. When distilled water was added the eggs re-
mained adhering to the glass.

In order to guard against slight differences in the masses of eggs,
the two oviducts of one female were emptied in separate dishes and
one used as an unfertilized control and the other stimulated electrically.
The batch of fertilized eggs of another female was cut in two equal
parts, and one part placed in a third dish. .

The eggs were washed in repeated changes of distilled water for an
hour. During this time the gelatinous coverings of the eggs partially
swelled. The sperm traverses the jelly and reaches the inseminated
eggs in about one hour, and the eggs rotate until their black poles are
turned upward. It is now time to stimulate the unfertilized eggs in
one dish. Two clean platinum electrodes, about one inch apart are
passed along opposite sides of each row of eggs, until all of the eggs in
the dish have been stimulated. These eggs now begin to rotate, and
some of them are completely rotated in one minute. If the rows of

INFLUENCE OF PERMEABILITY ON LIFE OF FROG’S EGGS 165

eggs are slender so that the jelly is uniformly swollen, all of the eggs ro-
tate at about the same time. None of the eggs in the unfertilized con-
trol rotate, no matter how much the Jelly swells. This is due to the
fact that the jelly sticks tight to the egg, whereas when they are stimu-
lated, a fluid is secreted between the egg and the jelly.

The water is now poured or syphoned off of all the eggs and a meas-
ured quantity added to each dish. Ordinary distilled water was boiled
and redistilled in an automatic quartz apparatus (4) and used in the
experiments. At the end of certain intervals of time, the water in
each dish is stirred, and 25 ec. removed for analysis of chlorides with
the nephelometer. At the end of seven hours all of the water is re-
moved from all of the eggs. An aliquot part of it is evaporated in
quartz for ash. Water is added to all of the eggs and another series
of determinations made.

The following table gives a sufficient number of the results to indi-
cate their character.

CC. OF A x}; NORMAL CHLORIDE SOL. GRAMS ASH
NO. OF OVIDUCTS TO ce. HO INH as ;
EACH DISH = SSS SS SESS

0-3.5 hrs. 0-7 hrs. 7-30 hrs. 0-7 hrs.
Paateriiized 2.2.0. ....- 500 0.70 1.20 0.38 0.005
TSG cele 500 1.40 2.24 122, 0.008
Loni rr 500 iL Oe 2.20 1.20 0.007
2 unierbilized......2...:.- 1000 1.40 2.40 1.42 0.015
Panimalated. oo... 02... 1000 1.90 3.60 2.40 0.024
mulerinlineds. os... 2s. 1000 1.80 4.20 2.00 0.023
APIMIETGUIZed. 222 ye.s.. 2000 2.80 4.80 3.00 0.031
cuaiinn | GCC 2000 4.00 7.60 4.80 0.053
80 7.20 4.20 0.045

PETIT ZOO Ss 2 =< Sign. cios s << 2000 3.

The first column shows the number of oviducts used in supplying
each dish with eggs, and the second gives the number of ec. H.O added.
In the first 3.5 hours as much chloride diffused out of the eggs as in 23
hours, beginning with the eighth hour. The table gives the total quan-
tity of Cl in the H,0 as ealeulated from analysis of a 25 cc. sample,
and the result may be converted into grams by multiplying by 0.035.

In obtaining the ash, three-fourths of the H,O was evaporated down
in quartz beakers. During the evaporation the foam was skimmed off
and transferred to a platinum crucible. When the volume of the
water was reduced to a few cc. it was transferred to the crucible. The

166 J. F. McCLENDON

ashing was done at as low a temperature as possible by admitting air
into the crucible through a small metal tube while it was heated. Some
chlorides may have been lost, but the results are comparative. The
ash was found to contain about 60 per cent of insoluble salts and 40
per cent of soluble salts. The ash contained Na, Li, K, Ca, Mg, Cl,
SO, and COs, but no phosphates were detected. The metals were de-
tected in the flame spectrum and the spark spectrum (fulgurator).
Mg was also detected by means of the displacement of the bands of
alkanin. Na was also determined quantitatively as sodium uranyl ace-
tate and as silicofluoride but the quantities were too small to be accur-
ate. SO, was precipitated with BaOH, and CO, driven out of the dry
ash with acid.

The diffusion of salts from the electrically stimulated eggs is about
the same as from the fertilized eggs, but nearly double that from the
unfertilized eggs. This difference was found to continue for 50 hours,
when the determinations were discontinued as the unfertilized control
had died.

In making prolonged experiments it is important to change the water
frequently in order to keep down the bacteria, which might bind the
salts or interfere with the nephelometer reading. In previous years, I
have charred and extracted the samples before nephelometric deter-
minations, but think it a bad practice and only necessary in the pres-
ence of bacteria or protein. The traces of dissolved glyco-protein from
the jelly did not cause a measurable error as a great excess of AgNO;
(5 drops of a saturated solution) was added to the 25 ec. sample, and
the reading was taken in one minute, whereas several hours are re-
quired for perceptible reduction of silver by these solutions in the in-
tensity of light to which they were exposed. I found it advantageous
to use open tubes closed at one end with black rubber stoppers in the
nephelometer, in order to avoid the reflection from the bottom of test
tubes, but this was not absolutely necessary with the Richards neph-
elometer, because a part of the large field shows no reflections, and test
tubes can be more easily cleaned.

The supposed relation between increased permeability and cleavage
has been fully discussed in a previous paper (5) in which I showed that
an increase in permeability of the sea urchin’s egg was produced by any
treatment which causes cleavage.

The fertilized or stimulated frog’s egg may live a long time, whereas
the unfertilized frog’s egg soon dies when placed in water. The ques-
tion arises whether the increased permeability saves the life of the egg.

— ~~ #o Stns

INFLUENCE OF PERMEABILITY ON LIFE OF FROG’S EGGS 167

In order to decide this we must first consider the effect of water on the
egg.

The frog’s egg swells when placed in water, and I found that the un-
fertilized egg swells more than the fertilized or stimulated egg, until
the death of the former terminates the physiological comparison. In
eggs of the same female, the mean diameter of 23 unfertilized eggs 30
minutes after being placed in water was 1.52 mm. whereas the mean
diameter of 23 that had been stimulated by the electric current and
left in water 30 minutes was 1.47 mm. After cleavage begins the esti-
mation of the volume of the eggs becomes more difficult, and on the
formation of spaces between the cells of the fertilized egg the apparent
volume is much greater than the real volume of the protoplasm. But
I think that one effect of fertilization is clearly the retardation of the
swelling of the egg. One need only suppose that the death of the un-
fertilized egg is caused by rapid swelling, in order to explain the life-
saving action of fertilization or stimulation. Bachman found the
unfertilized salamander egg bursts in two to seven hours if placed in
water but not if in salt solution. It is not so easy to determine the ex-
act moment of death of the frog’s egg.

The retardation of the swelling of the fertilized or stimulated egg is
evidently due to the loss of osmotic substances, such as the salts which
I have given in the above table. At the moment of increase in permea-
bility, a concentrated solution of these substances passes from the egg
and pushes out an adhering membrane which we may call the fertili-
zation membrane. Bialaszewicz’ claims that a measurable decrease in
volume of the egg takes place at this time. We may take this de-
crease in volume of the egg to equal the volume of the secreted fluid,
but after this perivitelline fluid has once been secreted it increases by
absorption of water through the fertilization membrane. I found this
membrane to be easily permeable to salts since it cannot be easily
plasmolyzed (shrunken) by immersion in hypertonic salt solutions.
Hence it is probable that the perivitelline fluid contains other less dif-
fusible substances. An analysis of the perivitelline fluid of the Am-
blystoma egg showed “that it contained salts and organic substances
but only a trace of protein. The perivitelline fluid of the egg of the
giant salamander contained 0.16 per cent of dissolved substances.
Backman and Runnstrom‘ suppose these substances to be chiefly se-
creted by the suckers that develop on the head of the frog embryo,

3 Bialaszewicz: Bull. Acad. Sc. Cracow, Math.-Nat., Oct. 1908.
Backman and Runnstrom: Pfliiger’s Archiv, 1912, exliv, 313.

168 J. F. McCLENDON

but this is hardly in harmony with the fact that the salamander embryos
have no suckers.

The function of the perivitelline fluid seems to be to shia room for
rotation of the egg and extension of the embryo, by pushing out the
fertilization membrane, which is chiefly effected by the osmotic pres-
sure of the dissolved substances that it contains. The loss of salts
must be more or less independent of this, since salts are lost continu-
ously during development, and salts would not be very effective in
pushing out the membrane since it is permeable to them.

This loss of salts explains the results of Backman and Runnstrom?®
who found that the osmotic pressure of the frog’s egg is reduced when
it is fertilized and placed in water. They attempt to explain it, how-
ever, by erroneously assuming that fertilization causes a coagulation
of the proteins and that the coagulation adsorbs the salts. They seem
to consider this egg as a diphasic system in which the watery phase
forms the main bulk of the egg. On the contrary, I found the cyto-
plasm to be a 4-phase system in which the watery phase is a very
small fraction of the total volume (6). The salt would have to be lost
from this watery phase only, in order to reduce the osmotic pressure
of the egg, and the more permeable the plasma membrane is to them
the less osmotic pressure they can exert while in the egg. I found the
watery phase to be 16 per cent of the whole egg and to contain 82 per
cent water (7). These figures are a little too high owing to slight ad-
mixture of other phases, but after removal of these and of the dissolved
proteins, the water-soluble substances in this layer formed only 1.7
per cent of the weight of the egg.

The chlorides diffusing out of the egg in seven hours would make a
solution in the watery phase more concentrated than 7} normal and
the total salts diffusing out in this time would form nearly 1 per cent
of the watery phase. Since the osmotic pressure of the ovarian egg
corresponds to a 0.166 normal solution of NaCl and is probably lowered
in the passage of the egg through the oviduct, the increased permeability
and diffusion seem to be sufficient to account for the failure of the egg
to die in the very hypotonic water in which it develops.

By an analysis of the ovarian eggs, I found that the chlorides were
not sufficient to raise the osmotic pressure of the watery phase to equal
that of frog’s blood or a 0.166 normal NaCl solution, and that they
were not bound during coagulation of the protein (furthermore there is

5 Backman and Runnstrom: Biochem. Zeitsch., 1909, 390.

|

INFLUENCE OF PERMEABILITY ON LIFE OF FROG’S EGGS 169

no evidence that fertilization causes coagulation), About 25 ec. (26.6
g.) of ripe ovarian eggs of the leopard frog were freed from lipoids,
dried, powdered and boiled in 200 ce. of distilled water, slightly acidu-
lated with acetic acid in order to coagulate the proteins, and filtered.
The ash of the filtrate required 1.55 cc. of ;'5 normal AgNO; to pre-
cipitate the chlorides, whereas the ash of the coagulum required only
0.2 cc., which might have been due to that part of the filtrate adhering
to the coagulum. Since this egg contains 53 per cent water, about
seven times as much water was used to dissolve out the chlorides as is
contained in the egg, but since the egg is permeable to salts, the whole
pond is available to wash out the salts when the egg is laid in natural
surroundings.

The supposition that the osmotic pressure of the frog’s egg falls
while it is in the oviduct is supported by the fact that such a change
happens to the bird’s egg. Atkins® found that the osmotic pressure
and chloride content of the egg white is much less than that of the bird’s
blood. The egg white corresponds roughly to the jelly of the frog’s
egg, both containing some glyco-protein and being secreted by the ovi-
duct. A coat of jelly of low osmotic pressure would cause a lowering
of the osmotic pressure of the egg, provided the protoplasm is at all
permeable to salts or water. Bialaszewicz’ found that the ovarial egg
of the hen has an osmotic pressure nearly equal to that of the hen’s
blood, but that the osmotic pressure of the yolk as it descends the ovi-
duct, steadily decreases. The yolk extracts the water from the white
so that the latter becomes denser.

Backman and Runnstrém state that the osmotic pressure of the fer-
tilized frog’s egg is as low as that of the pond water (A = 0.045) but
they admit that some of the jelly was adhering to the eggs that were
used in the cryoscopic determination. Notwithstanding the permea-
bility of the frog’s egg it continually swells after the slight shrinkage
due to the exit of perivitelline fluid, as I have repeatedly observed.
Bialaszewiecz made careful measurements of this swelling. Backman’
observed the same process in the eggs of the toad and the salamander
(Triton). Therefore we must conclude that the osmotic pressure of
the frog’s egg does not fall to the extent that Backman and Runnstrém
suppose.

6 Atkins: Biochem. Jour., 1909, iv, 480.
7 Bialaszewicz: Roux’s Archiv, 1912, 489.
8’ Backman: Pfliiger’s Archiv, 1912, cxlviii, 141.

170 J. F. McCLENDON

During the development of the frog’s egg and tadpole the osmotie
pressure rises to become that of the adult frog. This is probably
brought about by the solution of the reserve proteins (8) and the for-
mation of osmotic substances, rather than the absorption of salts from
without, since the same process takes place in the hen’s egg which is
cut off from osmotic exchange with the environment. Since the defin-
itive osmotic pressure is maintained fairly constant, Backman and
Sundberg? suppose that the skin is relatively impermeable to water,
from without toward the interior, and that the elastic contraction of
the cell walls resist the entrance of water produced by the osmotic
pressure. On the contrary the skin of the frog continually absorbs
water at a rapid rate (9), and it is only the Activity of the kidneys that
preserves the frog from dissolution. Any effect of the elasticity of
cell walls or the skin, in preventing swelling, must be very slight, since
the osmotic pressure is so great (8.9 atmospheres) and the frog will
burst if the action of the kidneys is eliminated. The pronephros de-
velops before the egg hatches and the kidney function probably devel-
ops before any danger of flooding the tissues with water occurs.

Although swelling may cause the death of the unfertilized eggs, a
normal degree of swelling seems necessary for development. The egg
contains 46 per cent of solid (dry) materials and the viscocity of the
protoplasm is very great. If the viscocity is not reduced by the ab-
sorption of water, the segmentation of the cells of the white pole is
prevented and the downgrowth of the cells of the black pole retarded.
Also by the action of centrifugal force, the solid materials may be
packed in the white hemisphere and its cleavage prevented (10). This
hard white hemisphere exists in less degree in eggs whose swelling is
prevented by immersion in solutions of too high osmotic pressure. For
a number of years I have repeated experiments in which the osmotic
pressure of the water was raised by the addition of salts. One-tenth
molecular solutions of NaCl, and isotonic solutions of chlorides and ni-
trates of Na, K and Li have the same effect. If the eggs are placed in
these solutions immediately after they are fertilized, and allowed to
remain 24 hours, the segmentation of the white pole of some of the eggs
is prevented. If they remain 48 hours in the solutions a large number
are so affected, and if they remain indefinitely none of them develop.
The addition of Ca or Mg does not prevent this effect, so it is probably
due to osmotic pressure.

® Sundberg: Pfliiger’s Archiv, 1912, exlvi, 225.

— i 4uve »

ee Ew ——————— ll
= ee, ky

INFLUENCE OF PERMEABILITY ON LIFE OF FROG’S EGGS 1 Ai |

These salts are not entirely without toxic action, for this is mani-
fested in the swelling of serous cavities. The swelling of the pericar-
dium is greater in lithium solutions than in isotonic solutions of Na or
K. In more concentrated solutions, the individual cells separate from
one another, 1 condition which Roux called framboisea of the embryo.

A marked contrast exists in the effects of pure salt solutions on the
eggs of the frog and those of many fish. These fish eggs are normally
impermeable to salts and water and the osmotic effects are never ob-
served. The toxic effects of salts are, however, more manifest (11).
These toxic effects, resulting in the swelling of serous cavities and
other abnormalities, are associated with increase in permeability (12).
Though a slight increase in permeability does no damage, a greater
increase causes abnormalities and a still greater increase causes death.
It is probable that the swelling of the pericardium of the frog’s egg is
produced by increase in permeability in the same way as the swelling
of the pericardium of the fish egg. The difference between the two is
that, whereas the fish egg is normally impermeable to salts and to water,
the frog’s egg is permeable to water and to a less degree to salts. Fer-
tilization increases this permeability to the optimum degree, and pure
salt solutions increase it too much.

SUMMARY

Fertilization or electric stimulation increases the permeability of the
frog’s egg so that Na, K, Li, Mg, Ca, Cl, SO, and CO; diffuse out at a
faster rate. Probably the swelling of the pericardium, due to the ac-
tion of pure salt solutions, is caused by too great increase in permea-
bility. Notwithstanding the outward diffusion of salts, the egg ab-
sorbs’ water, and the addition of osmotic substances to the medium,
preventing the absorption of water, prevents the segmentation of the
white hemisphere, and gastrulation.

The abnormally rapid swelling of the unfertilized egg probably is the
cause of its death, and thus the increase in permeability on fertilization
lowers the osmotic pressure and saves its life.

The low osmotic pressure of the egg observed by Backman and Runn-
str6m was probably caused by three processes: (1) The lowering of
the osmotic pressure as the egg descends the oviduct; (2) the increased
permeability and loss of salts; (3) the admixture of jelly of low osmotic
pressure with the eggs in taking the freezing point.

172 J. F. McCLENDON

REFERENCES

(1) McCLenpon: Science, 1911, xxxiii, 629; and Am. Jour. Physiol., 1912, xxix,
298.

(2) McCienpon: Am. Jour. Physiol., 1912, xxix, 302.

(3) McCuLenpon: Science, 1914, xl, 70.

(4) McCuenpon: Int. Zeit. f. physik.-chem. Biol., 194, i, 31.

(5) McCienpon: Am. Jour. Physiol., 1910, xxvii, 240.1 ~

(6) McCienvon: Archiv f. Zellforschung, 1910, 386, fig. 1, a.

(7) McCienpon: Roux’s Archiv, 1909, xxvii, 247.

(8) McCienpon: Am. Jour. Physiol., 1909, xxv, 198.

(9) McCienpon: Internat. Zeitsch. f. Physik.-chem. Biol., 1914, i, 169.
(10) McCienpon: Roux’s Archiv, 1909, xxvii, 253.
(11) McCienpon: Am. Jour. Physiol., 1912, xxix, 289.
(12) McCienpon: Int. Zeit. f. Phys.-chem. Biol., 1914, i, 28.

THE ACTION OF ANESTHETICS IN PREVENTING
INCREASE OF CELL PERMEABILITY!

J. F. McCLENDON
From the Physiological Laboratory of the University of Minnesota
Received for publication May 5, 1915

Héber and Lillie have attempted to associate anesthesia with per-
meability. In a lecture at Woods Hole in 1911, I attempted to ex-
press the relation as follows: “If stimulation consists in increase in
permeability, we should expect anesthetics to prevent this change.’’

Most of the attempts at correlating anesthesia and permeability
have been indirect. Thus Héber* found that anesthetics prevent
the action of salts in producing a demarcation current in muscle. Lil-
lie‘ observed that anesthetics prevent the action of salt solutions in
causing artificial parthenogenesis and cytolysis of Echinoderm eggs.
He also observed that anesthetics prevent the outward diffusion of
cell pigment produced by salts, and thinks this a direct proof that
anesthetics prevent increase in permeability. Arrhenius and Bubo-
novic® observed that anesthetics retard the hemolytic action of hypo-
tonic solutions. H6éber attributes a similar observation to Traube,
but I have not as yet found the reference.

It is clear, therefore, that anesthetics in the anesthetizing concentra-
tion, tend to prevent the outward diffusion of pigment from pigmented
cells. It seemed to me of interest to determine whether non-pigmented
substances are affected in this way, especially the salts contained in
the cells. In 1911, I had observed that the Fundulus egg is impermeable
to salts, but found that Mg came out of the eggs when placed in a slightly
toxic solution of NaCl.®

1 This research assisted by a grant from the Research Fund of the Graduate
School.

2 McClendon: Biol. Bull., 1912, xxii, 139.

3 Héber: Pfliiger’s Arch., 1907, exx, 492, 501, 508.

4Lillie: This Journal, 1912, xxix, 373, 1912, xxx, 1; 1913, xxxi, 255; Journ.
Exper. Zool., 1914, xvi, 591.

5 Arrhenius and Bubonovic: Meddel. K. Ventensk. Akad. Nobel institut, 1913,
li, no. 32.

6 McClendon: This Journal, 1912, xxix, 296.

173

174 J. F. McCLENDON

It was later found that various toxic solutions increase the permea-
bility of the egg to salts.7. The addition of anesthetics to these salt
solutions tends to prevent this increase in permeability.’

In making these last experiments, Fundulus could not be obtained
and the much more delicate eggs of the pike were used. It was found
that practically no salts came out of these eggs when they were placed
in distilled water. Since the egg contains salts, and it does not swell
or burst when placed in distilled water, it must be impermeable to water
as well as to salts. When placed in slightly toxic solutions of nitrates
or anesthetics the chlorides diffuse out of the egg very rapidly. But
when a toxic salt solution (=45 molecular NaNO3) contains an anesthetic
in the proper concentration for anesthesia, the chlorides diffuse out of
the egg at a slower rate than in the pure salt solution. Thus anesthetics
have two effects on these eggs, at a certain concentration they produce
enesthesia and at a greater concentration they are toxic. A 3 per cent
solution of alcohol or a 0.5 per cent solution of ether produces anesthesia
(retards development) whereas a 6 per cent solution of alcohol, or a
2 per cent solution of ether is slightly toxic. Twice as much chloride
diffuses out of eggs placed in pure NaNO; solution as in the same solu-
tion containing 3 per cent alcohol. One and a half times as much
chloride diffuses out of the egg in the pure NaNO, solution as in the
same solution containing 0.5 per cent ether.

These experiments were extended this year, and the chlorides de-
termined more accurately with a Richards’ nephelometer.

MATERIAL

The eggs of the pike (Esox) were used. These eggs are very delicate,
and are especially susceptible to changes in temperature, lack of oxy-
gen and to the presence of various substances in the medium. The
eggs were shipped from the breeding grounds in northern Minnesota
to the State Fish Hatchery in Saint Paul by being spread out in
thin layers on stretched muslin in refrigerator boxes. From this
point they were brought to the laboratory in a refrigerator box, and
those in good condition selected one by one and placed in water that
was distilled, boiled and later redistilled in an automatic quartz still.’
No other water was used in the experiments.

7 McClendon: Internat. Zeitschr. f. Physik.-chem. Biol., 1914, i, 28.
® McClendon: Sci., 1904, xxxviii, 280.

or

EFFECT OF ANESTHETICS ON CELL PERMEABILITY 17

METHODS

The eggs are washed in the redistilled water until the salts practi-
cally cease coming out of them, and measured in a series of graduated
cylinders of the same bore. That the chlorides have all been washed
out of the transparent egg shell follows from the following facts: The
last wash waters contain so little chlorides that they cannot be detected
with the nephelometer unless the washings are boiled down to a small
volume. Tap water usually contains 100 to 10,000 times as much
chlorides. The eggs are made permeable to chlorides by ;'5 molec-
ular NaNO; solution, and while they remain in this, twice as much
chloride diffuses out of the eggs in six hours as in three hours. If these
chlorides were held in the thin shell we would expect more to come out
in the first three hours than in the second three hours. Since the
shell is at all times easily permeable to salts, the source of the chlorides
is not the fluid between the egg and the shell, and the small amount of
shell substance is probably not the source of such large amounts of
chlorides as were obtained from the eggs. I found that when the
Fundulus egg is made permeable to chlorides the other salts diffuse
out also, and when the frog’s egg is made permeable to Cl that SO,,COs,
Na, K, Li, Mg and Ca diffused out also. This is probably true of the
pike egg also, and enough salts could probably be obtained by dif-
fusion to equal the entire weight of the egg shell, which is protein in
composition. The most reasonable hypothesis is that the plasma
membrane is practically impermeable to salts (and to water) and
that it is made permeable by toxic solutions. The permeable egg may
live and develop for several days, showing that the entire store of
salts is not necessary for its immediate needs.

EXPERIMENTS

The same volume of eggs was placed in each of a series of Stender
dishes, with 30 cc. of the solution to be tried. At the end of six hours,
25 ce. of this solution was removed with a pipette having a trap to
prevent contamination with saliva. The sample was transferred to
the nephelometer tube, and five drops of a saturated solution of AgNO;
added and mixed. The two matched test tubes used in the nephelom-
eter were filled up to the zero point of the scale by the 25 cc. + 5
drops.

In case any eggs died during an experiment, this experiment was
thrown out. Since the material was available in great abundance,

176 J. F. McCLENDON

a number of experiments were performed, of which the following are
typical:

(1) 5 ce. of eggs in each dish. Chlorides expressed in arbitrary units. Ethyl]
alcohol expressed in per cent by volume.

Solution;<:..-25 75 NaNO; |+1%alcohol| +2% ale. | +3% ale. +4% ale.
C@hiorides)>.:.-. 100 60 50 45 60

(2) 15 ee. of eggs in each dish

Seton | * NaNO; ®. NaNO;+2% ale. | 3 NaNO;+3% ale.
Chlorides..... | 100 50 40

The duration of the experiment, six hours, was selected because this
was the maximum duration that could be considered safe. Even
then, some of the experiments had to be thrown out because one or
two eggs died before they were completed. The NaNO; has to be pure
in order to have the right degree of toxicity. The ethyl alcohol was
redistilled over metallic sodium. Although the dishes were covered
in all experiments, the experiments with ether were discontinued after
it was found that ether has the same effect as alcohol, owing to the fact
that some of the ether diffused into the air above the eggs and the
concentration of that remaining in the solution was no longer known.
It is probable that all anesthetics have the same action.

A word of caution to any one who wishes to make similar experiments
on pike eggs: Some of my early experiments were irregular in results.
I believe the reason for this is that the egg is made more permeable by
increase in temperature,’ and that different eggs are not affected by
the same temperature. I did not determine the maximum temperature
at which it is safe to work, but a temperature of about 8° is safe and is
easily maintained in a refrigerator. Fundulus eggs may he used at
room temperature in Woods Hole, and are preferable in every way.
It is clear that 2 to 3 per cent by volume of ethyl alcohol partially
inhibits the permeability-increasing action of a pure NaNOs solu-
tion. It is theoretically possible to entirely prevent this increase in
permeability if the toxicity of the NaNO ; solution is low enough.
The difficulty in demonstrating this lies in the fact that the toxicity
of the sodium salt does not depend on its absolute concentration, but

*Osterhout: Biochem. Zeitschr., 1914, Ixvii, 272.

EFFECT OF ANESTHETICS ON CELL PERMEABILITY 177

on the ratio of sodium to calcium. With a very mildly toxic solution
of NaNO; the Ca diffusing out of the egg at the first increase of per-
meability lowers the toxicity and hence the permeability to such an
extent that no permeability increase can be measured by means of the
nephelometer, and electric conductivity experiments with their large
and numerous sources Of error would have to be substituted.

- That 2-3 per cent ethyl alcohol is really the anesthetic concentration,
follows from the fact that it retards the development of these eggs.
The same concentration may not be correct for every species or every
tissue. In general, it seems that nerve tissue requires a less concen-
tration of an anesthetic for anesthesia than other tissues. But it is
hardly justifiable to assume that this affect of anesthetics on per-
meability is peculiar to egg tissue. Pike embryos were found to behave
the same as eggs up to the time of the development of kidney function,
when the excretion of salts interferes with the method used to measure
permeability. It is probable that anesthetics retard the increase in
permeability of any cell by any “stimulus.”

DISCUSSION

In 1910 I observed that chloroform, when added to the sea water,
reduces the electric conductivity of sea urchin eggs. This experiment
was not repeated, owing to lack of time and the large number of sources
of error that must be guarded against in order to be sure that decreased
conductivity indicates decreased permeability. Osterhout,’® by find-
ing a tougher material, was able to show that anesthetics decrease
the permeability, at least of certain plant cells. Joel'! found that
anesthetics decrease the permeability of erythrocytes. The question
arises whether anesthetics prevent increase in permeability by decreas-
ing permeability. If a cell is absolutely impermeable its permeability
cannot be decreased. The Fundulus egg is so nearly impermeable to
salts that it would be extremely difficult to measure a decrease in
permeability. The pike egg may be more permeable to salts but it
would be no easier to measure a decrease in permeability, owing to the
delicate nature of the egg and danger of non-uniformity of material.
It seems probable, however, that the anesthetic and toxic substances
act on the same constituent of the cell surface or plasma membrane.
If the cell surface is composed of a mosaic of different constituents,

10 Osterhout: Sci., 1913, 111.
U Joel: Pfliiger’s Arch., 1914.

178 J. F. MCCLENDON

and one constituent is permeable, the cell is permeable. But if one
constituent is impermeable the whole cell is not necessarily impermeable,
since diffusion can take place through the other constituent. If the
NaNO; makes the cell permeable by acting on a protein it is difficult
to see how the anesthetic could antagonize this effect by acting on a
lipoid.

Stimulation and anesthesia seem to be antagonistic states. I have
shown that the permeability of striated muscle is increased on stimu-
lation” and that the permeability of the eggs of the sea urchin and the
frog increases when they pass from the state of repose into that of
activity. The question arises whether this is true of other cells,
such as those of glands. In order to decide this we must discuss the
psycho-galvanic reflex.

If a constant electric current is sent through the body from non-
polarizable electrodes and the person is given a nervous shock, as by
sticking a pin in him unexpectedly, the strength of the current is mo-
mentarily increased. Leva't showed that the degree of this change
is in direct ratio with the number of sweat glands per unit area of the
skin, and therefore concluded that the sweat glands produce this
phenomenon. Gildemeister®’ was able, by improvements in the
electric conductivity method, to show that this is due to the increased
permeability, presumably of these glands. Apparently the gland
cells are stimulated by the sympathetic nerves and their permeability
is increased.

It seems, therefore, that the increased permeability on stimulation
is a general phenomenon, and we might expect the action of anesthetics
in preventing this increase to be general, also.

Warburg" has shown that anesthetics retard the oxidation of oxalic
acid by blood charcoal to about_the same degree as they decrease the
respiration of nucleated erythrocytes. I have repeated and confirmed
these experiments. Tashiro and Adams! have shown that the nerve
fiber gives out less CO» when it is anesthetized. Although it was shown
by Warburg! that the respiration of sea urchin eggs is only slightly

1 McClendon: This Journal, 1912, xxix, 302.

13 McClendon: This Journal, 1910, xxvii, 240, and ibid, in press.

14 Leva: Miinch. Med. Wochenschr., 1913, 2386.

15 Gildemeister: Miinch. Med. Wochenschr., 1913, 2289; see also Schwartz:
Zentlbl. f. Physiol., 1913, xxvii, 734.

16 Warburg: Pfliiger’s arch., 1914, elv, 147.

17 Tashiro and Adams: Int. Zeitschr. f. Physik.-chem. Biol., 1914, i, 451.

18 Warburg: Zeitschr. f. Physiol. Chem., 1910, Ixvi, 306.

a

EFFECT OF ANESTHETICS ON CELL PERMEABILITY 179

reduced during anesthesia, it seems generally true that anesthetics may
antagonize oxidations by cells, oxidases and some inorganic katalyzers.
The question whether permeability has any relation to the oxidative
processes cannot be finally settled until we know more about the mech-
anism of the latter. At present we can, at most, make the generaliza-
tion that in the presence of Oz cell respiration varies more or less with
cell permeability so long!’ as the cellis alive. The only observation that
I know of that might appear to extend this rule to a dead cell is that of
Warburg.*®° He found that the oxidation in the young erythrocytes
of the goose is increased by freezing and thawing provided they are
closely packed. Freezing and thawing causes hemolysis and increases
permeability, and the hemolyzed cells might be considered dead.

SUMMARY

Anesthetics in the concentration that retards development (2-3 per
cent alcohol dr 0.5 per cent ether) tends to inhibit the permeability-
increasing action of a ;5 molecular solution of NaNO; on the eggs
and embryos of the pike (Esox).

19 McClendon and Mitchell: Journ. Biol. Chem., 1912, x, 459.
20 Warburg: Zeitschr. f. Physiol. Chem., 1911, 419.

NEW HYDROGEN ELECTRODES AND RAPID METHODS
OF DETERMINING HYDROGEN ION
CONCENTRATIONS

J. F. McCLENDON
From the Physiological Laboratory of the University of Minnesota

Received for publication May 8, 1915

The technique of hydrogen electrodes as applied to pure chemistry
is in a high state of perfection! but the impression is sometimes given
that these electrodes require a skilled physical chemist to use them
correctly. It is true that some solutions give trouble, for example,
a pure KCl solution. Starting with Merck’s highest purity KCl,
I recrystallized it five times in fused silica dishes and dissolved it in
conductivity water, but could not obtain a neutral reading with the
hydrogen electrode until I poured it boiling hot into the electrode and
passed a rapid stream of hydrogen through it while cooling and while
making the reading. This difficulty is never experienced with solu-
tions containing considerable amounts of carbonates, phosphates or
proteins, and hence biological fluids are less difficult than some inor-
ganic solutions. In fact, it is not necessary with biological fluids to
change the hydrogen in the electrode during the reading.

The time required to determine the reaction of a fluid depends chiefly

on the time required to saturate the electrode with hydrogen. Since °

gold absorbs comparatively little hydrogen, it becomes quickly satu-
rated. I found No. 36 gold wire very satisfactory from this stand-
point but the electrodes made of it could not be so conveniently cleaned
by heating, as platinum electrodes. Drucker made electrodes of films
of iridium burned on Jena glass, but the same objection holds true of
them. By reducing the thickness of the platinum, the saturation time
may be reduced. I found that platinum foil 0.02 mm. in thickness,
coated with platinum black, requires less than two minutes for satura-
tion, provided it is separated from the hydrogen by only a film of the
solution. Since the electrodes which Michaelis designed for rapid

' Ostwald-Luther: Physiko-chemischer Messungen, 3d ed. 1910, Leipzig.
180

Ne eek on eee el

DETERMINATION OF HYDROGEN ION CONCENTRATION 181

work require thirty minutes for saturation? a considerable saving of
time is thus accomplished by simply using narrow strips of thin foil
instead of the wire that he used.

The chief difficulty in determining the H* concentration of biologi-
cal fluids arises from the fact that they contain dissolved gases. The
dilution of the hydrogen with other gases causes an error in the direc-
tion of greater acidity. The loss of CO, from the solution increases
its alkalinity, hence the passage of COs. and O» from the solution into
the hydrogen causes two errors which tend to oppose one another, and
hence the reading might happen to be correct. Héber, Hasselbalch,
Michaelis and others have guarded against the error due to escape
of CO, from the solution. The method of Michaelis depends on the
use of a very small volume of Hy, in ratio to the volume of the solu-
tion, and is best adapted to rapid work.

The chief difficulty with other gases is experienced in determinations
on arterial blood. The oxyhemoglobin gives out so much QO, into the
H, as to cause a greater error than arises from the escape of
CO,. Milroy* centrifuged the blood and then poured it
through the air into the electrode. In order to obviate
errors arising from this procedure, as well as other errors,
and to shorten the time required for a determination, I de-
signed the following electrode:

This electrode consists essentially of a U-tube with one
end constricted, figure 1. Next to the constriction, a strip
of platinum foil, 0.02 mm. in thickness or thinner, is fused
through the glass so that the plane of that portion of the
foil which protrudes into the interior of the tube passes
through the axis of the tube. The free end of the foil, A, is bent up
so that when a loop of wire is passed over the constricted end of the
tube electric contact will be made with the foik A short piece of
rubber tube, B, is attached to the constricted end of the glass tube,
and closed with a small Langenbeck clip or pinch cock at C. The rub-
ber tube is as short and of as small bore as practicable, and its free
end is connected with a hypodermic needle. The large end of a hy-
podermic needle is filed down so that it may be inserted into the rub-
ber tube. The platinum is coated with platinum black and cleaned
in the usual manner.

Fig. 1

2 Michaelis: Die Wasserstoffionenkonzentration, 1914, Berlin.
3 Milroy: Quart. Journ. Exper. Physiol., 1914, 141

182 \ J. F. McCLENDON

Before filling the electrode, the needle is dried thoroughly be means
of a suction pump and filled with oil. It may then be boiled for sterili-
zation. The rubber tube is filled with a concentrated solution of hiru-
din in water or Ringer. The needle is inserted into the artery or vein
and the Langénbeck clip removed. The U-tube is held in such a posi-
tion that only the first few drops of blood come in contact with air,
as this first blood covers and protects the rest of the blood. When
the U-tube is filled, the Langenbeck clip is put on the rubber tube and
the needle is removed.

The U-tube is now placed in the shield of a centrifuge, counter-
balanced, and centrifuged a few minutes. It is then removed and a
pipette inserted into the free end of the rubber tube. By sucking
on the pipette and at the same time partially opening the Langenbeck
clip; the blood corpuscles that remain in the rubber tube are removed.
A tube from which pure hydrogen is flowing is instantly put in place
of the pipette, without admitting any air. By cautiously opening the
Langenbeck clip, hydrogen is admitted until the platinum foil is sur-
rounded by the gas. A film of plasma adheres to the platinum and
sides of the glass tube and establishes electrical connection with the
blood below. After waiting two minutes for the platinum to be satu-
rated with He, the U-tube is shaken so as to bring a fresh portion of the
plasma in contact with the platinum, and the reading is immediately
taken. Care should be taken not to shake so hard that any of the
corpuscles rise high enough to liberate any oxygen into the hydrogen.
In order that no time be lost in connecting this electrode with the
calomel electrode, a ball of cotton cord soaked in a saturated solution
of KCl is kept at hand and a piece cut off previously, with which to
make the connection. The inclosure of this conducting cord in a tube
is unnecessary.

Since the reading can be made in a few seconds with a proper potenti-
ometer, the CO, does not have time to diffuse out of the film of plasma
covering the platinum and cause an error. Since the erythrocytes
have all been precipitated away from the surface layers, and the ex-
posure is but little more than two minutes, the contamination of the
hydrogen with other gases is very slight, and the error due to this
unmeasurable.

This electrode has been indispensable in my work on blood, but it
can be used with any biological fluid. In studying stomach or duo-
denal contents, the rubber tube may be connected to the smallest size
stomach or duodenal tube, having a strainer (bucket) on the end that

DETERMINATION OF HYDROGEN ION CONCENTRATION 183

is swallowed. The filling of the U-tube may be assisted by aspiration
through a rubber tube attached to its free end, but the suction should
not be sufficient to cause bubbles to appear in the fluid.

In case the readings are made at room temperature, it will be neces-
sary to insert the U-tube in mercury in order to cool it with sufficient
rapidity. If blood or any other fluid that requires to be centrifuged
is used, the centrifuge shield may be filled with water so that most of
the cooling takes place without loss of time.

If extreme accuracy is not required and the fluids to be investigated,
contain no oxyhemoglobin, the electrode designed by Michaelis may be

used, provided thin platinum foil is substituted, and a
bulb is blown on the open end of the glass tube to pre- u ay 8
vent spilling while introducing the He. A strip of foil ;

0.05 mm. thick may be drawn through a small rubber
stopper by means of a needle and thread, instead of
sealing it in a glass stopper. This foil may be cleaned F
by heating, provided the stopper is first wet with dis-
tilled water, whereas a glass stopper is liable to crack.
In measuring the acidity of the gastric contents, it was
found possible to lower an electrode into the stomach.
The apparatus designed for work on the stomach con-
tents consists chiefly of a rubber tube 60 cm. long and 3
mm. bore, and two No. 40 silk covered copper wires, that
were coated with rubber cement and dried several times
(fig. 2). One wire, M, extends through the rubber tube,
JJ, and the other, N, passes down outside of it until by
entering the hole, EH, it connects with a platinum wire
that is fused into the lower end of a short piece of glass
tube that is inserted into the rubber tube. The lower end
of the glass tube and copper-platinum junction is covered
with sealing wax, A. A drop of pure mercury is dropped into the
lower end of the glass tube so as to connect with the platinum wire
at the level of B. Above the mercury a little calomel washed with
concentrated KCl solution, C, is placed, and the rest of the glass tube
packed with moist KCl crystals, D, and the hole, E, stuffed with cot-
ton soaked in KCl solution. This forms a calomel electrode, and is
separated off from the remainder of the tube by a short piece of
glass rod, F. Above F several holes are cut in the rubber tube at
the level of G, and from this point a fine platinized platinum wire
extends through the lumen of the tube and is held in place by fusion

184 J. F. McCLENDON

to a bump on the inside of a short piece of glass tube at the level
of I. This platinum wire then connects with the wire M and the
junction is coated with rubber. The rubber tube is connected at K
with a tube, L, leading from a hydrogen generator, and a slow stream
of Hy passes down the rubber tube and out at G, thus converting the
platinum wire from H to F into a hydrogen electrode. Whereas
the results are not quite as accurate as those obtained with larger
electrodes, I think them sufficiently accurate for stomach contents,
where there are such great individual variations. It is necessary to
have a source of hydrogen of sufficient pressure to prevent the stomach
contents from rising in the tube higher than H. Fresh crystals of KCl
must be put in D before the appa-
ratus is used, and the end of the
tube from A to G may be immersed
in saturated KCl solution so as to
keep it moist until it is swallowed.
A correction is made for hydrogen
pressure (subtract 0.17 mv. for 1
cm. increase in pressure).

The time necessary to calculate
the H+ concentration from the
potentiometer reading may be
saved either by using a conversion
table or making a potentiometer
that reads off directly the H+ con-
centration. It is most convenient
to express results in the form of the
hydrogen ion exponent, PH. Thus
0.001 normal is 10-*, or PH=8.
Figure 3 is a conversion table to be used at 23°, provided the hydrogen
electrode is connected with a calomel electrode of the saturated type
(containing KCl crystals). The codrdinates are to be followed only
to the edge of the diagonal band. Starting with the potentiometer
reading in millivolts on the ordinate, the corresponding PH on the
abscissa may be read.

The temperature coefficient is small and is practically zero for an
H* concentration of 0.0001 normal (PH=4). That is to say a 0.0001
normal solution of hydrogen ions will give a reading of 481 milli-
volts whether the temperature is 19°, 25° or 37°. If 23° is taken as
the standard, at 19° there is a variation of +3 millivolts for PH =0 and

~

DETERMINATION OF HYDROGEN ION CONCENTRATION 185

—5 mv. for PH=10. At 37° the variation is —11 mv. for PH=0 and
+16 mv. for PH=10.

Since the PH of gastric juice is not far on the acid side of 4, that
of urine is about 4 and that of the other biological fluids (except those
containing much bile) is not far on the alkaline side of 4, a slight
variation of the temperature from 23° does not make a serious error
in using this conversion table or a potentiometer reading the PH
directly.

None of the potentiometers on the market can be easily adapted to
read the PH directly, but I made one that will serve all such purposes.
It consists of 2200 em. of No. 30 hardened German silver wire stretched
over cross section paper. By means of a Weston cell, and 2 movable
contacts, the current of a lead storage cell passing through a variable
length of the wire is so adjusted that each centimeter has a fall of
potential of one millivolt. By means of a conversion table the PH is
marked on the cross section paper. We thus have a scale reading the
PH instead of the millivolts. See also the description of a direct read-
ing potentiometer in this journal.

SUMMARY

The method of H+ concentration as used by Michaelis is modified
so as to reduce the time necessary for a determination from forty min-
utes to about two or three minutes.

An electrode which eliminates errors due to O2 (from oxyhemoglobin)
and COs, is described. An electrode that may be lowered into the
stomach is also described.

A potentiometer reading the Ht concentration directly, instead of
millivolts, is described.

A DIRECT READING POTENTIOMETER FOR MEASURING
HYDROGEN ION CONCENTRATIONS

J. F. McCLENDON
From the Physiological Laboratory of the University of Minnesota

Received for publication May 14, 1915

One disadvantage of the use of the hydrogen electrode in determining
the hydrogen ion concentration of biological fluids is the time re-
quired to calculate the results, and the liability to error by one not ex-
perienced in the use of mathematics. In order to avoid part of the
calculation, some workers stop the calculation as soon as the logarithm
is obtained. These logarithms cannot be used in compiling statistical
data in the same way as decimal fractions, and hence the calculation
that is saved at one time may have to be done at another. The direct
reading potentiometer does the entire calculation for you in zero time.

The difficulty in making a direct reading potentiometer for hydro-
gen ion concentrations lies in the numerous temperature coefficients
that would have to be considered in making an extremely accurate
and universal apparatus. But Ihave shown in a recent paper! the
temperature coefficient of the whole apparatus is negligible in modern
heated rooms with wall thermoregulators. During the summer the
temperature variation would be serious only for determinations on
blood, where the differences observed are comparatively minute. The
error due to temperature is certainly not greater than the same error
in reading a burette. The instrument is correct at 23°.

The potentiometer is made to be used in connection with a calomel
electrode filled with crystals of KCl, because this avoids the danger of
error due to the change in concentration of the KCl solution by dif-
fusion or evaporation. A Weston cell is used. If it does not give the
theoretical voltage, it should be standardized and its voltage marked
on it.

The potentiometer may be made in two sizes, a large size that is
easily made accurate and a small size with evident advantages.

1 McClendon: This Journal, 1915, xxxviii, 180.

1S6

DIRECT READING POTENTIOMETER 187

The large form is essentially a resistance wire 23,000 mm. long, with
movable contacts. Suppose the Weston cell is correct, that is to say,
the voltage marked on it is 1.0185 at 23° the temperature of standardi-
zation of the potentiometer, if the current from it is passed through
10,185 mm. of the wire each centimeter will have a fallof potential of 1
millivolt. A storage cell current is sent through and balanced against
the Weston cell current by changing the length of the wire. Since the
lead storage cell gives a current from about 1.85 to 2.23 volts, the length
of the wire should be capable of being changed a corresponding amount.
By means of the half dial, C, and the sliding contact, D, figure 1, the

eS
-

g PTEREST eM CELE ORT CALO MEL ELECTRODE
S
eS ere » ;:° ~ %® xy aa
+ Cass wo Oo Ot nw os 4 v4 a
a 5 =
aoe = = 2.8 Py eB = 3
= : : ; : : =
5 22: 103
ft

TOSTORAGE CELL —

Fig. 1

length of the wire can be changed from 18,000 to 23,000 mm. corre-
sponding to 1.8-2.3 volts. The slide wire, D, is 1 meter and hence out
of proportion in the figure. In the figure, the length of wire between
two buttons is marked by a number parallel to the wire and the total
length of wire up to a button is marked perpendicular to the wire.
The measurement of the wire for the half dial and Slide wire, men-
tioned, need not be accurate, since the results are not affected by these
measurements.

In order to determine the E. M. F. of the hydrogen + calomel elec-
trodes, they are connected with the contacts on the dial, A, and the

188 J. F. MCCLENDON

slide wire, B. Since with a normal solution of hydrogen ions in the
hydrogen electrode, the apparatus has a voltage of 0.2468, therefore
2468 mm. of wire are included between the dial and the slide wire.
Since the moving of the decimal point in the hydrogen ion concentra-
tion one place, causes a change in voltage of 0.0587, there are 587 mm.
of wire in the slide wire, B, and the same length between each pair of but-
tons on the dial, A. There are 14 buttons, and hence the length of
the wire up to the slide wire is (13x587)+2468 = 10,099, and hence
86 mm. on the slide wire is the place for leading off to the Weston cell,
or a total of 10,185 mm. <A temperature scale may be made for the
Weston cell. It is 10,186 at 20°, 10,184, at 25°, and 10, 181 at 30°.

WO. AP CAL.E.
SS ae

WESTOW
CELL

ba
oy

4
Ss, 6 5 ¥ 3 “9 ad
<j + 9 °
“ aye H. concentration 23
Ld ~ ’
x aes ELECTRODE ?
B <2 Bis .
. x10 35%, : ‘.@ x10! Hy : a
men So rer Cl: as
: : : = EL. SAT. wiTH KOl, 24
{ as aa =e xio 2 CALOMEL EL.SA ; ¥

The scale under the slide wire, B, for reading the hydrogen ion con-
centration, is logarithmic, and expressed in millimeters is as follows:
27+30+34+39+46+57+73+4+ 103+ 178 = 587.

There is a binding post for the negative and one for the positive
pole of the storage cell, one to the capillary electrometer, one for the
Weston cell and one for the calomel electrode, and a double throw
switch to connect the slide, B, either with the calomel electrode or the
Weston cell.

A diagram of the face of the potentiometer is shown in figure 2.
After the apparatus is set up in the usual manner, the contact of the
dial, A, and the slide, B, and the double-throw switch are moved to the

DIRECT READING POTENTIOMETER 189

points marked Weston cell, and the half dial, C, and slide, D, moved
until a balance is obtained. Then the switch is thrown over and the
dial, A, and slide, B, moved until a balance is obtained. The hydrogen
ion concentration is then read off directly, for instance 0.4 on the
slide wire and x10~ on the dial, or 0.4x10-?=0.004. Since it saves
paper to write 0.3x10-", for instance, than the equivalent decimal
fraction, the reading is in the most convenient form for general work,
without any calculation whatever. The range is from 10~° to 10-%,
but this may be increased on the acid side as high as it. is possible
for the acidity of any solution to rise, in case it is desired to use the
method on strong solutions of mineral acids.

37 Z 23 B e
250
~~ ~los, sees
TS eo Ale!
eon Face,
* S pe
= oy"
FS >
S ey
ier : :
°
S od A ° : G Le
ao y es
~~ e
28 ° .
aS ° y Ne
oe ot) e y
~) en og’
250 wot 2213
+ 213 D
Fig. 3

The lengths of wire in millimeters for the small instrument are shown
in figure 3. The length of the slide wire, B, is 250 mm. This is the
same length as the logarithmic scale on small-sized slide rules, and
logarithmic codrdinate paper of this length may be obtained of Keuffel
and Esser or other dealers. Therefore it is not worth while to calculate
the divisions on the scale. Some samples of logarithmic paper are a
little short, especially in a very dry room, but they may be moistened
until they expand to the required length and then glued tight, especially
at the ends, under the slide wire. By comparing figure 2, the orienta-
tion and labeling of the scale may be easily done. The mark on the
slide wire scale for the Weston cell is 37 mm. from the left end. The

190 J. F. McCLENDON

temperature correction for this point is so small that it may be judged
with the eye. In general this mark should be shoved approximately
1 mm. to the right for 5° fall and 1 mm. to the left for 5° rise in tempera-
ture. At 30° it should be about 1.5 mm. to the left.

Instead of flexible wires to connect the sliding contact, B, with the
double throw switch, ‘and the sliding contact, D, with the binding
posts, it is better to use brass rods with movable sleeves bearing the
contact points, as in figure 3. :

There are 250 mm. of wire between each two buttons on the dial,
A, and 1051 mm. between this and the slide wire, B. If this wire is
measured accurately the instrument will not need calibration for bio-
logical purposes, because the non-uniformity of well made wire, taken
fresh off of the original spool, will be small, and notwithstanding slight
errors in the wire the instrument will show differences accurately.

There are 9 buttons on the half dial, C, with 213 mm. of wire be-
tween each pair, and 215 mm. on the slide wire, D.

The more resistant the wire the better the storage cell will hold up.
Manganin (86 Cu, 12 Mn, 2 Ni) invented by Weston, or Constantan
(60 Cu, 40 Ni) of from 32 to 36 gauge, are the best materials for the
wire, but German silver will do, especially if it is bare and strung in
a ventilated box under the dials to avoid heating effects. Some of
the wire should be saved to renew the slide wires when necessary.
The connecting wires marked by heavy lines in figure 1, should be
heavy copper (resistance free). The leads to the potentiometer do
not have to be heavy, as only enough current passes through them
to operate the electrometer.

I do not have a contact key on the potentiometer, because this
key when released, must short circuit the electrometer, and a telegraph
key connected directly with the electrometer, is used.

The small instrument is about 12x7 inches.

SUMMARY

The hydrogen ion concentration is read on the scale, B, figure 2,
and the position of the decimal point read on the dial, A.

ACIDITY CURVES IN THE STOMACHS AND DUODENUMS
OF ADULTS AND INFANTS, PLOTTED WITH THE AID
OF IMPROVED METHODS OF MEASURING HYDROGEN
ION CONCENTRATION

J. F. McCLENDON
From the Physiological Laboratory of the University of Minnesota

Received for publication May 14, 1915

The hydrogen ion concentration’ of the gastric juice and stomach con-
tents after test meals is very well known.' According to Michaelis
and Davidsohn? hyperacidity means a hydrogen ion concentration
of from 0.011 to 0.088; average acidity means 0.028—0.0015 and hy-
poacidity means 0.00041—0.0000001.

Owing to the number of factors involved in the physiology of the
stomach, as elucidated by the extensive work of Cannon and of Carl-
son, it occurred to the writer that the giving of a scant test meal, poor
in protein, and the complete evacuation of the stomach all atonce,
might not reveal all that could be learned of the acidity of the stomach
under normal conditions. Consequently an attempt was made to
trace the progress of digestion after a normal meal, by lowering a hy-
drogen electrode into the stomach, and also by removing a few cubic
centimeters every half hour by means of a very small tube with a
strainer (bucket) on the end that is swallowed. The new departures
in the technique are given in two other papers.*

It was found possible to keep the tube in the stomach all day except
during the ingestion of the meals. After a little practice it was not
very difficult to keep the strainer well in the interior of the main part
of the stomach. The acidity at the cardia fluctuated more or less
owing to the intermittent entrance of saliva. The pylorus became
acid more rapidly than the rest of the stomach. But it is not true that
' the pylorus is always more acid than the fundus. My observations

1 Frankel: Zeitschr. f. Exper. Path. u. Therap., 1905, i, 431.
* Davidsohn: Zeitschr. f. Exper. Path. u. Therap., 1910, 398.
’ McClendon: This Journal, 1915, xxxviii, 180 and 186.

191

192 J. F. McCLENDON

indicate that the stomach contents begin to mix immediately after the
ingestion of the food, and that the mixing is very thorough in about
two to three hours.

_ The meals that were eaten were governed by the appetite, except
in one case when a scant meal was purposely eaten for comparison.
The real efficiency of the stomach to take care of these meals may be
determined by finding the time required for the acidity of the main
bulk of the food to rise to that degree most favorable to peptic digestion.
This may be determined with little discomfort and little or no loss of
nourishment.

The usual method of pumping out the entire stomach contents
after an abnormal meal and titrating the acids and the salts of weak
bases, is not calculated to shed very much light on the digestion. Such
a procedure implies that acidity is some indestructible property of
certain kinds of matter. The same procedure that is used to esti-
mate matter, such as total nitrogén, cannot be used for studying acidity.
In the first place new hydrogen ions are formed during the titration
and the original number cannot be estimated in this way. The acidity
may be determined more accurately by tasting than by titration.
In the second place, the same number of hydrogen ions in the stomach
may favor or retard digestion according to how they are distributed.
Peptic digestion is most rapid in a hydrogen ion concentration of about
0.03 normal! and decreases when the acidity increases or decreases. The
total stomach contents after being pumped out and mixed may be at
the optimum acidity, whereas in the stomach, part may have been too
acid and part not acid enough. It is much better to determine the
hydrogen ion concentration of a sample taken from the large part of
the stomach, away from the wall, because the main bulk of the food is
in this location. Whereas digestion may commence immediately
in the pylorus and next to the wall, the quantity of food so affected
appears to be small, and general digestion is delayed until there has
been a thorough mixing of the food with the gastric juice.

THE REACTION OF THE ADULT STOMACH AND DUODENUM

Figures 1 and 2 show a number of curves, representing the rise in
acidity of the adult stomach during digestion. On the ordinates are
the hydrogen ion concentrations and on the abscissae are the hours

‘Sorensen: Biochem. Zeitschr., 1909, xxi, 131; Michaelis and Davidsohn:
loc. cit.; Michaelis and Mendelssohn: Biochem. Zeitschr., 1914, Ixiv,

ACIDITY CURVES OF STOMACH AND DUODENUM 193

after finishing the meals. Curves 1-4 are taken from the same individual
to show the effect of meals of varying bulk, and hence varying protein
content. Curve 1 was taken after a light meal of milk, curves 2 and
3 after average meals and curve 4 after a heavy dinner. Curve 5,
from another individual, was taken after a meal that was not only light
but also very poor in protein (chiefly carbohydrates). It is clearly
seen that the heavier the meal and the more protein it contains, the
more slowly the acidity rises. This is due to the fact that the protein
neutralizes the acid and the acidity of the stomach cannot rise very
high until all of the protein is changed into acid albuminate.

The curves in figure 2 are all taken from different individuals and in
general steeper than those in figure 1. This is due to personal character-
istics and not to the pro-
tein content. Whether Lee
this personal difference is 4
due to difference in rate
of acid secretion by the
stomach, or difference in __,,,,
the quantity of saliva
swallowed, is not clear.
Curve 12, showing a very
slow rise after a medium
meal, was taken from a 70! ; ; ; oi8

person having a very sen- ;
Sat. and notices Fig. 1. Curves 1-4, male, 34 years; curve 1,
a wee luncheon, 500 cc. milk. Curve 2, breakfast, 1

bly large quantities of orange, 250 cc. milk, 200 cc. water, 2 eggs, 2 slices
saliva were swallowed. of buttered toast. Curve 3, luncheon, roast
In the average individual, lamb, potato, bread and butter, lettuce, ice
the tube produces no Cream. Curve 4, dinner, soup, beefsteak, potato,
ag : bread and butter, tomato, 250 cc. milk, 200 ce.
stimulation of the mucous : ite

z water, slice of pie. Curve 5, male, 32 years,
membranes. Curve 11, breakfast, bacon, cereal, bread and butter.
taken after a light meal,
rises steeply, whereas curve 6 taken from another individual, after a light
meal, chiefly carbohydrate, does not rise so steeply.

It is well known or generally assumed that the height to which the
acidity rises is a personal characteristic. This is well illustrated in
curves 1—4 (all from the same individual) all of which reach the same
maximum. Curves 7 and 8 are from two brothers and show the same
maximum. This suggests that the acidity of the stomach is influenced

by heredity.

-00000!

194 J. F. McCLENDON

After reaching the characteristic maximum, the acidity of the stomach
remains constant for considerable time, at least until the greater part
of the food has left the stomach. After this it is difficult to take a
sample from the interior, that is, away from the wall, and fluctuations
that are sometimes observed may be due to this fact. After all of
the food has left the stomach the acidity may fall, due to cessation or
decrease in the secretion of acid and also to the swallowing of saliva

+
t
mH is is
SS
Ol
6
ae 7B 4
7
.00] a

000! 3 as
}
{]/

-00001

-00000} /

900000!

' hour 2 3 y

Fig. 2. Curve 6, male, 40 years, breakfast,
toast and buttermilk. Curve 7, male, 24
years, dinner, ham, eggs, potato, pie.
Curve 8, male, 19 years, dinner, veal roast,
rhubarb, soup, buttermilk. Curve 9, male,
18 years, luncheon, 4 roast beef sandwiches,
chocolate creams. Curve 10, male, 22 years,
dinner, roast beef, 2 eggs, bread and butter,
strawberries. Curve 11, male, 25 years,
luncheon, ham, potato, bread and butter,
coffee, sauce. Curve 12, male, 25 years,
ham, potato, bread and butter, sweets,
milk, rhubarb. Curve 13 (unfinished), fe-
male, 19 years, luncheon, vegetables, bread
and butter. Curve 14, compiled from 27
samples of stomach contents of infants of
the first month.

that is about neutral in reac-
tion and contains protein capa-
ble of neutralizing acid.

The hydrogen ion concentra-
tion of a number of duodenal
samples taken with the duode-
nal tube were all very close to
2x10-°. This may be generally
true of all normal individuals,
but it is not always true in dis-
ease. A sample which I de-
termined for Dr. Schneider
was very far from this value,
but I will not give it here be-
cause he may wish to work up
the subject from the patho-
logical side. :

THE CALIBRATION ©,’ INDI-
CATORS ~

During the experiments de-
scribed, I took occasion to
calibrate a number of indica-
tor solutions and _ papers.
Michaelisand Davidsohn’ give
a table of indicator solutions
that are to be used after the

Ewald breakfast. My results, after normal meals correspond closely with
theirs except in case of methyl violet (and tropaeolin 00). Since my re-
sults correspond closely with those of Sorensen on non-protein solutions,
this difference is not due to the higher protein content of my samples,

5 Davidsohn: Zeitschr. Exp. Path. u. Therap., 1910, viii, 398.

ACIDITY CURVES OF STOMACH AND DUODENUM 195

and some of my samples had as low a protein content as those of Michae-
lis. The difference is probably due to the fact that Michaelis’ samples
were highly colored with tea. The yellow tea and blue indicator
gave a green at 0.01 whereas in my experiments green never appeared,
and in those of Sorensen it appeared at 1 normal and yellow appeared
at 2 normal. I confirmed Sorensen’s results on non-protein solu-
tions and used different brands of methyl violet, Merck’s 3B and Grub-
ler’s 5B, on gastric contents, but never obtained green at 0.01 when the
contents were originally colorless. A similar difference was observed
in the case of tropaeolin 00. In my tests it became orange at 0.01
and in those of Michaelis at 0.0033.

This trouble with the color of the sample increases as we go into the
duodenum or test the urine. For this reason I calibrated a number
of indicator papers to see if such a thing were possible. It is generally
considered that indicator papers are less sensitive than the solutions.
One reason for this is that the paper used contains substances that
preserve the reaction (alkaline earths?). Bausch and Lomb congo
paper begins to turn at 0.001 whereas some that I made begins to turn
at 0.00005. I made a number of papers by soaking Schleicher and
Schull’s ash-free paper no. 589 in the indicator solutions and drying,
and found them but little less sensitive than the solutions. The end
of the paper is dipped into the sample and held there a little while.
As the sample rises in the paper any coloring matter is adsorbed very
quickly and prevented from rising further. The acid or alkali rises
highe ->d gives the characteristic tinge to the indicator. Pure water
rh higher, so that only the middle of the wet region should be
observed. Proteins, and some other interfering substances also, are
partially held back from ascending in the paper.

The original solutions and the color of the papers were as follows:
- methyl violet 0.2 per cent (violet), tropaeolin 00 2 per cent (deep yellow),
dimethylamidoazobenzole 0.5 per cent (light yellow), Congo red,
0.5 per cent (red), methyl orange 2 per cent (yellow), alizarine sodium
sulphonate 0.5 per cent (salmon pink), p-nitro phenol 2 per cent (light
yellow), litmus 2 per cent (violet), neutral red 0.5 per cent (red).

The following table gives the observed changes in adult gastric
contents: (—)=no change, B=blue, G=green, O=orange, R=red,
V=violet, W=white and Y=yellow, (+) =a slight change:

196 J. F. McCLENDON

Normal H?..... .|0.05) 0.01 | 0.005! 0.001} 0.0001 | 0.00001! 0.000001) 0.000005
Met. violet...... B B — —_ = os ES =

_ iropaolinw, see: R O — _ = = =e a
2 met.a.a.ben....| R R O _ =: a at =
Convowee eee B B B V V = = <
Met. orange..... Ray) JR R R O _ _ --
Alizarine....... Ye MY YC "Ys ¥: oa = a
LBL e Aco Perea os a a R R R a = nit
P-nitro-ph....... wi WwW W W W + = a
Neut. red.......| — | — — - _ = a O

These indicator papers were tried on infants’ stomach contents with
the same result.

The indicator papers used on duodenal contents were: neutral red
0.5 per cent (red), rosolic acid 0.5 per cent (orange), cyanin 0.5 per cent
(blue) phenol-phthalein 0.5 per cent (colorless). All of the normal duo-
denal contents investigated (0.00000005—0.00000001) affected the papers
in the same way. Neutral red was changed to orange, rosolic acid
was reddened, cyanin and phenol-phthalein were unchanged.

THE REACTION OF THE INFANT STOMACH AND DUODENAL CONTENTS

The work on the infant was done at the request of Prof. J. P. Sedg-
wick, head of pediatrics, University Hospital. The determinations of
acidity were done by me with the hydrogen electrode, except for a
few that were done by Dr. Rood Taylor under my direction. The
samples were taken by Dr. Taylor, and the radiographs taken by Dr.
F. S. Bissell, to both of whom my sincere thanks are due. Dr. Sedg-
wick gave me some references to the literature.

It was observed by Huenekens® that the acidity of the infant’s
stomach depends on the diet, and Hess,’ that the titratable acidity of the
stomach of the new born before it has taken any food, is high. I was
able to obtain several samples of the latter and the acidity was 0.005,

_but the acidity of the empty infant’s stomach is in general high, as
will be shown later. Since the protein of the milk binds the acid, we
should expect the acidity to be higher the less food is in the stomach,
and such was found to be the case. In other words, the regulatory
mechanism of gastric acidity in the infant is very imperfect.

Although 27 samples of the gastric contents of infants of the first
month were investigated, I had no control over the manner in which

6 Huenekens: Zeitschr. f. Kinderheilk., 1914, xi, 297.
7 Hess: Amer. Journ. Diseases of Children, 1913, vi, 264.

» tee

ACIDITY CURVES OF STOMACH AND DUODENUM 197

the samples were taken, and an actual curve of the rise in acidity can-
not be drawn. The entire stomach contents were pumped out and
mixed in each sample. Since no relation between age of the infant and
acidity was found, the various determinations may be used to construct
a theoretical curve. Individual variations cannot be investigated,
owing to the few determinations made on one infant. If we consider
all samples taken one hour after nursing, for instance, it would be in-
correct to use the mean acidity in computing the one-hour point in
the curve, because a single stomach that emptied itself in one hour
would have more influence than several stomachs that remained full for
one hour. But if the variation curve of these determinations is made,
the ‘‘mode”’ of the curve may be taken instead of the ‘‘mean.’’ Most
of the determinations are very near the ‘‘mode,”’ and to use the “‘mode’’
has the same effect as throwing out the extreme variations. The results
are as follows:

Time after nursing
in hrs. 0.25 0.5 1.00000 ins Uber is 2 Pls 3 4
H* 0.000006 0.000005 0.000006 0.00006 0.00005 0.00012 0.00005 0.001 0.01

The acidity rises to 0.00012, two hours after nursing, and then drops
to 0.00005, only to rise again very rapidly until the stomach is empty
or until the next meal. This fall is due to the fact that there were
not enough samples between two and three hours to obtain a good
variation curve from which to determine the ‘‘mode.” If the acidity
curve is smoothed, we obtain the one shown in figure 2, curve 14.

Since woman’s milk is neutral* and the earliest samples taken
were of a much greater hydrogen ion concentration, it is probable that
the stomach contains some gastric juice before the milk enters it. The
acidity of this juice is high, as shown by curve 14. At the end of four
hours there is practically no milk or even curds in the stomach, and
at this time the acidity equals that of the adult stomach.

From a quarter to one hour after nursing the acidity remains prac-
tically stationary. During this time the protein is being transformed
into acid albuminate and all of the acid secreted is used in the process.
It was during this period and the hour following it that Davidsohn,?
and others took their samples and observed the very low acidity.
Schackwitz,'° who did not take samples in the same way as Davidsohn,
observed a greater acidity in a few cases.

8 Davidsohn: Zeitschr. f. Kinderheilk, 1913, ix, 11.

® Davidsohn: Zeitschr. f. Kinderheilk, 1911, ii, 420.
10 Schackwitz: Monatschr. f. Kinderheilk., 1903, xiii, 73.

198 J. F. MCCLENDON

It seems probable that the milk does not begin to leave the stomach
until the end of one hour. From this time on, it is gradually dimin-
_ ished in quantity in the stomach and the acid that it secreted has less
protein to neutralize it, and this protein is already partially saturated
with acid, so that the acidity of the stomach continually rises. When
the acidity has risen high enough for rapid peptic digestion, the quan-
tity of milk left in the stomach is so small and passes out so quickly
that it is safe to say that protein digestion practically does not occur
in young infants’ stomachs after a milk diet.

As a control on the collecting of the samples, the enzymes were
determined. Pepsin was determined by the edestin method, trypsin
by the casein method and lipase by the splitting of tributyrin as deter-
mined by the surface tension measured with Traube’s stalagmometer
(Michaelis). The stomach contained pepsin and gastric lipase (and
milk lipase) but no trypsin.

Twenty-three samples of the duodenal contents of the same babies,
taken by means of a catheter, were examined. Some of these were
not bile stained but contained trypsin. By means of X-rays and de-
terminations of bile and enzymes and the length of the tube swallowed,
the origin of the samples was determined. There is no relation be-
tween the time after nursing and the acidity. I exclude three samples
that were not bile stained, taken about four hours after nursing in the

attempt to get duodenal contents, and showing an acidity of about 0.01.
Dr. Taylor considered these as stomach contents, and I have included.

them in making the curve of stomach acidity.

The variation curve of duodenal acidity has a ‘‘mode”’ of about 0.0008,
amaximum of 0.004 and a minimum of 0.00000015. When the stomach
is very acid the acidity of the duodenum is lower than that of the
stomach, but when the stomach is weakly acid the acidity of the duo-
denum is higher than that of the stomach. Very probably the acidity
of the pylorus is always high, and causes the high acidity of some of
the duodenal samples. The less acid samples have a greater admixture
of bile. The bile is not sufficient to make the duodenal contents
alkaline.

Unless the succus entericus is alkaline and abundant, the entire in-
testine must be acid. The question arises, what is the nature of the
infant’s digestion? The enzymes were determined in the same way
was those of the stomach. I did not find it possible to distinguish
between gastric and pancreatic lipase, but presumably gastric, pan-
ereatic and milk lipase were present. Pepsin was present. ‘Trypsin

ACIDITY CURVES OF STOMACH AND DUODENUM 199

was probably always present, although in three samples outof fifteen
it seemed doubtful.

It seems certain that a high acidity is necessary for very rapid pep-
tic digestion, but it is probable that after the pepsinogen is once
activated by acid, slow digestion may take place at a very low acidity.
Sorensen" observed peptic digestion in an acidity of abut 0.00008.
According to Michaelis and Davidsohn® tryptic digestion ceases when
the acidity is raised to 0.00001. We see therefore that the acidity of
the infant’s stomach is high enough to activate the pepsinogen, and
that the reaction of the duodenum is as favorable for peptic digestion
as it is for tryptic. Probably both forms of proteolysis proceed si-
multaneously in the infant’s intestine. According to Rona and Arn-
heim" the reaction curve for erepsen is about the same as that of trypsin.

Very little is known of these enzymes, except the processes which
they hasten and the conditions under which they act. Since it is
certain that proteolysis occurs in the alimentary canal of the infant,
the next thing to be determined is the reaction of the remainder of
the intestine.

SUMMARY

The acidity of the adult stomach rises rapidly during the first 1.5
to 3 hours after a meal, after which it remains stationary until the
food has nearly all left the stomach.

The rapidity of rise is less the heavier the meal and the more protein
it contains, but it seems also to depend on the efficiency of the individual
stomach.

The height to which the acidity rises is a personal characteristic.

The adult duodenal contents are slightly alkaline. The hydrogen
ion concentration is about 0.00000002.

The acidity of the infant’s stomach (of the first month) rises slowly
during the time from fifteen minutes to one bour after nursing, after
which it rises rapidly until the stomach is empty. Four hours after
nursing it may be as acid as the adult stomach. The acidity is sufficient
to activate the pepsinogen, but there is so little milk left in the stomach
after the acidity has risen half way that the peptic digestion in the
stomach seems unimportant.

The infant’s duodenum is more acid than the average acidity of the
stomach. Pepsin is always present and peptic digestion must take place.

11 Sorensen: Biochem. Zeitschr. 1909, xxi, 294.

12 Davidsohn: Biochem. Zeitschr., 1911, xxxvi, p. 280.
1% Rona and Arnheim: Biochem. Zeitschr., 1913, lvii, 84.

THE PERFUSION OF THE MAMMALIAN MEDULLA: THE
EFFECT OF CALCIUM AND OF POTASSIUM ON THE
RESPIRATORY AND CARDIAC CENTERS

D. R. HOOKER
From the Physiological Laboratory of the Johns Hopkins University

Received for publication June 10, 1915

Using a perfusion apparatus the principle of which has already been
described (1) the author has sought to perfuse the medullary centers in
the dog. The method is now sufficiently advanced to report upon it
and in this paper it is proposed to give the results obtained from a study
of the effects of calcium and potassium upon the cardiac and respiratory
centers.

An investigation of the effects of these salts upon the respiratory
activity in the frog (2) while in part apparently conflicting showed
clearly that, as compared with a balanced solution, predominance of
calcium over potassium caused excitation and predominance of potas-
sium over calcium caused depression. The same effects are now found
in the case of the respiratory activity in the dog. It is found further
that the salts in question have a definite action upon the tonicity of
the cardiac centers, and it is interesting to note that while the resultant
effect of potassium is to cause a slowing of both respiration and heart
rate the one effect is apparently due to inhibition of one medullary
center (respiratory), while the other is due to the excitation of another
center (cardio-inhibitory).

The most recent work on the perfusion of the mammalian medulla is
that of Herlitzka (3) and of Winterstein (4). Herlitzka fed defibrinated
blood to the heart in dogs in which the circulation was limited to the
head region. Under these conditions he noted the cardiac and respira-
tory movements and observed that the repeated head circulation of
defibrinated blood, even when oxygenated was inadequate to preserve
function for any length of time. The substitution of fresh defibrinated
blood, however, temporarily improved conditions. Winterstein per-
fused new-born rabbits with saline solution at room temperature and
observed that the addition of carbon dioxide and other acids produced

200

a

PERFUSION OF MAMMALIAN MEDULLA 201

respiratory discharge in centers previously quiescent and concluded
that the hydrogen-ion concentration of the perfusate governed the
function of the respiratory center. Both of these authors review the
earlier work in this field. Hirschfelder and Brown (4) have reported
on certain pharmacological reactions in the central nervous system
perfused with defibrinated blood but their results are not yet accessible
to the writer.

Method. Such success as has been obtained by the present method
is probably due in
large part to the use
of a new device for
aerating the blood
perfusate. The ap-
paratus is shown in
figure 1. It con-
sists of an inverted
bell-jar covered with
a brass plate which
clamps air-tight
over a rubber wash-
er. Openings in this
plate permit the in-
troduction of a
thermometer, a tube
for the gas mixture
and a tube for the
entering venous
blood; a third tube
leads outward
through a trap for
the escape of gas. Fig. 1. Apparatus for aerating blood. Description
In the center of the in text.
plate a hollow axle
(a) carries a hub (6) which extends above and below the plate; below,
it supports a flat rubber dise (c) of a diameter slightly less than that
of the bell-jar, and above, a pulley (d) rotation of which causes the
rubber disc to rotate. As the blood falls upon the rotating disc it is
thrown against the side of the bell-jar and runs -down in a thin film
exposed to the contained air and collects in a suitable reservoir below.
(The figure shows only the rubber stopper adapted to close the union

202 D. R. HOOKER

between the chamber and the reservoir.) The hollow axle is closed
by the entrance of insulated wires which lead to a small electric bulb
by means of which the temperature is regulated. The bell-jar now in
use is about 15 by 20 cm.

In the work here reported the perfusate consisted of washed cor-
puscles suspended in Ringer’s solution. There was in consequence
little tendency to froth. In numerous preliminary experiments, how-
ever, to test the maintenance of function in the medullary centers and
in experiments directed to other ends defibrinated and hirudinized
bloods have been used successfully. The apparatus permits the ex-
posure of circulating blood to any chosen gas mixture and obtains ade-
quate aeration without the frothing so likely to occur when protein
solutions are agitated. So far as tried the size of the chamber is ade-
quate to experimental needs. It is obvious that a larger one could be
substituted if the volume of blood flow demanded it.

Technical procedure. The animal is prepared for perfusion under
chloretone anaesthesia. The thorax is freely opened under artificial
respiration after ligation of the internal mammary vessels. The right
subclavian artery and vein are ligated. Loose ligatures are laid under
the superior vena cava and left subclavian artery. The innominate
is ligated close to the aortic arch and an inflow cannula is inserted distal
to this ligature. An outflow cannula is placed in the right external
jugular vein. The preparation is now connected with the perfusion
apparatus and the ligatures about the left subclavian artery and superior
vena cava are permanently tied. The head circulation is thus isolated
except for anastomotic branches along the cord which doubtless unite
the spinal branches of the intercostal with the spinal branches of the
vertebral arteries. The arterial supply to the brain by the latter paths
is insufficient to maintain life. A more serious difficulty lies in the
venous anastomoses along the cord by which path there is a continuous
seepage of blood from the head circulation into the body. The tech-
nical difficulties of occluding this path are so great that it has seemed
best to ignore it at least until experimental results are likely to be
invalidated by the entrance of chemical substances into the systemic
circulation.

After the perfusion is established 300-400 cc. of blood are allowed to
waste from the left external jugular vein before the latter is connected
to return the perfusate to the apparatus. The isolated circulation as
thus accomplished includes the neck parts and the whole head as sup-

PERFUSION OF MAMMALIAN MEDULLA 203
plied by the branches of the carotid and the external carotid. The
perfusate enters by way of both carotids and the right vertebral artery.

During an experiment artificial respiration is maintained to preserve
the cardio-vascular system as a recorder of the medullary effects on
this system. The heart-rate has been recorded with a Hiirthle manom-
eter connected with the femoral artery and the respiration by record-
ing the movements of the epigastrium.

With the procedure as stated there is a possibility that spinal and
vertebral vascular anastamoses might connect the systemic with the
isolated circulation and that the former
might in part contribute to sustain medul-
lary life. Opposing this possibility are
these facts: (a) medullary life ceases at
once after ligation of both carotids and
both vertebrals or upon stopping the per-
fusion in an experiment. It is interesting
to note, however, that when the medulla
has been isolated and perfused for some
time the respiratory center will continue to
discharge for a little while after the perfu-
sion has been stopped just as the isolated

Fig. 2.

ruary 10. Tracing of the re-

Experiment of Feb-

mammalian heart will continue to beat
under like conditions. (b) Medullary
activity (respiratory movements) contin-
ues on the isolated circulation long after
the heart has stopped beating. In some
experiments lasting an hour or more the
heart was not beating at any time.

Figure 2 is reproduced to indicate the
reliability of the method employed. It
shows the respiratory movements in a dog
two hours and fifteen minutes after per-
fusion was begun.

diluted with about an equal volume of Ringer’s solution.

spiratory movements from a
dog in which the head was

perfused with defibrinated
blood. Record 2 hours, 15
minutes after perfusion was
begun. Heart ceased beating
30 minutes and eyelid reflex
disappeared 2 hours after per-
fusion was begun. Perfusion
pressure 60 mm. Hg. Up-
stroke indicates inspiration.
Time in seconds.

The perfusate consisted of defibrinated dog’s blood

The per-

fusion pressure at the time the record was taken was 60 mm. Hg.
Respiratory movements continued regular for ten minutes longer or

for a total period of two hours and twenty-five minutes.

In this ex-

periment the heart stopped beating in the first half hour, a fact which
supports the assertion that the isolation of the medulla is complete.

204 D. R. HOOKER

It is further of interest to note that the eyelid reflex disappeared thirty
minutes before the respiratory center ceased to discharge; in the major-
ity of experiments, however, the eyelid reflex persisted longest. It
should be stated that this is the longest period the respiratory center
has been kept alive. Because the heart had stopped beating no ex-
perimental procedures were instituted; this undoubtedly contributed
to the period of activity of the preparation because it has been repeat-
edly observed that changing the experimental conditions in prepara-
tions with such a narrow margin of safety markedly reduces the period
of survival. Nevertheless a number of experiments were continued
successfully for two hours or more. The temperature of the blood in
the venous outflow cannula was about 32° C.

Experimental. In the present experiments dogs’ corpuscles sus-
pended in salt solution made up the perfusate. An anaesthetized
animal was bled, transfused with Ringer’s solution and bled again.
The blood was defibrinated and the corpuscles were separated in a
centrifuge. The corpuscles were then washed twice in and finally
suspended in the following solutions:

per cent per cent per cent
NBC aenisuks ere 0.90 JI Od Lee ioe es 0:90 NaGl!, 22. See 0.90
Ca Ole tics chime tts 0.03 CaCl 25>. Aa 06 KG ty ieee 0.06
1 OSS ape ee 0.03

The solutions, tested before the corpuscles were added, were all
slightly alkaline. The Ringer’s mixture had a value of P, = 8.4 +;
the potassium solution P, = 8.2; while the calcium solution fell be-
tween the others. This variation in H ion concentration, if it prevailed
after the corpuscles were added, was in all likelihood too small to affect
the results. Furthermore, it will be noted, the potassium solution
which depresses respiratory activity tends to be more acid than the
calcium solution which acts as a stimulant.

In order to obtain the requisite volume of perfusate the corpuscles
were suspended in a volume of salt solution about twice as great as that
of the serum. While this dilution was not determined accurately in
the several experiments care was observed to have the dilution the
same in the suspensions used in any one experiment. The perfusate as
thus prepared was kept on ice over night, in some cases over two nights,
before being used. It was hoped that the washing would yield a final
solution free of calcium or potassium when the absence of one or the
other of these salts was desired so that the results might be correlated

PERFUSION OF MAMMALIAN MEDULLA 205

with those obtained from the frog referred to in the introduction of this
paper. This was not possible, however, as the supernatant fluid was
never free of either calcium or potassium as shown by analysis. The
experiments therefore bear only upon the effect of a preponderating
amount of caleium and of potassium. The perfusate was aerated in
all the experiments with pure oxygen.

Figure 3 shows the stimulating effect of a predominance of calcium.
Between the marks the perfusate low in potassium and high in calcium
content was substituted for the balanced solution. The respiratory
center, previously quiescent, is caused to discharge and the heart-rate
is markedly increased. This record, in common with others, shows a
latent period of considerable length. The construction of the perfusion
system is such that the change of perfusion solution is made some dis-
tance from the organ
under investigation
which no doubt ac-
counts in large part ry on Fore
for the delayed effect bt TI DD eee ll
sincesometimemust _iae WN
elapse before the new
solution reaches the
medulla. After the
return to the control

Fig. 3. Experiment of February 9. To show the
: effect of a preponderance of calcium on the heart-rate
perfusate the respi- and respiratory movements. Downstroke indicates
ratory movements inspiration. Time in seconds.

quiet down and the
heart-rate, somewhat more slowly, returns to the former rate.

Figure 4 shows the depressant effect of a predominance of potassium.
Between the marks the perfusate low in calcium and high in potassium
content was substituted for the balanced solution. A transitory in-
crease in respiratory activity is noted which is followed by a period of
complete depression. This momentary increase of respiratory activity
was observed in several but not in all the records. The analagous con-
dition of momentary depression when changing to the calcium-strong
perfusate was also occasionally seen. This effect is not unlike that
sometimes exhibited by the isolated heart when the feeding solutions
are changed. Along with the depression of respiratory activity there
is a marked slowing of the heart-rate. This slowing is taken to be a
vagus effect caused by stimulation of the cardio-inhibitory center since
it largely disappears after vagus section. The conclusion therefore

206 D. R. HOOKER

seems inevitable that potassium inhibits one center in the medulla while
stimulating another. The data from this series of experiments, in
which the cardiac nerves were intact, are presented in Table I.

An analysis of the response of the cardio-inhibitory and cardio-
accelerator centers independently of one another was attempted after
section of the vagus or accelerator nerves. These results are collected
in Table IT.

In the single experiment in which the accelerator nerves were cut
potassium increase produced a retardation of the heart-rate comparable
to that observed in the animals with cardiac nerves intact. The in-

|
|
|
\ |

WAU I i IMI UM

vu Hy wilt i

VRPORDAGOGOREL

Fig. 4. Experiment of February 2. To show the effect of a preponderance
of potassium on the heart-rate and respiratory movements. Downstroke indi-
cates inspiration. Time in seconds.

crease in heart-rate following calcium increase was relatively insignifi-
cant. The result is, however, decisive enough to justify the belief that
-alecium exerts some Ba ae effect on the cardio-inhibitory center.
Five experiments were directed to show that calcium increase stimu-
lates the cardio-accelerator center directly. The difficulty in this pro-
cedure is to obtain a slow heart-rate after vagus section so that accele-
rator effects will be apparent (6). In four of the experiments a slowing
of the rate was attempted by reducing the pulmonary ventilation (7);
the result in one of these was negative, in two it was insignificant and
in the fourth the increase amounted to twenty beats per minute. In
the last of these there was a very definite increase in the amplitude of

PERFUSION OF MAMMALIAN MEDULLA 207

beat. In the fifth of this series of experiments the heart-rate was
slowed by pilocarpine in the systemic circulation. By this procedure
(experiment of June 4) a very definite increase in both rate and ampli-
tude of beat was obtained by calcium increase.

TABLE I

Respiratory and heart-rates per minute before and during medullary perfusion
of solutions containing a preponderance of calcium or of potassium. Cardiac nerves
intact. The records were counted for 30 seconds.

RESPIRATION HEART RATE RESPIRATION HEART RATE

Before Ca |During Ca} Before Ca |During Ca|| Before K | During K| Before K | During K

Feb. 2.... 3 0 72 30

18 0 104 30
mag. |- 0 52 38 122
0 12 34 124 |
Mar.5.... 18 0
29 0
15 0
ManG...|.> 6 12 70 82 6 0 54 28
TABLE II

Heart-rate per minute before and during medullary perfusion of solutions con-
taining a preponderance of calcium or of potassium after accelerator section and
after vagus section. The records were counted for 30 seconds.

HEART RATE HEART RATE
ap ps =) hes | PROCEDURE
Before Ca | During Ca| Betore K | During K
0 a 142 150 150 50 After accelerator section
iS ee 124 134 116 102 After vagus section
May 20 oo... 102 110 After vagus section
May 22... 2. 144 164 After vagus section
wane 4.27... 11 20 After vagus section
9 24 After vagus section

It seems therefore probable that both cardiac centers are responsive
to each of the salts investigated. The effect of calcium is however
much more pronounced on the accelerator center while the effect of
potassium is much more pronounced on the inhibitory center. The

208 D. R. HOOKER

last described results in conjunction with several negative experiments
emphasize further that the marked effect of the salts of calcium and
potassium on the heart-rate seen in the experiments in which the cardiac
nerves were intact is not due to a direct cardiac effect produced by the
entrance of the perfused solution into the systemic circulation otherwise
the heart would be influenced as much in the one set of experiments as
in the other.
CONCLUSIONS

1. The medulla of the dog may be successfully perfused and the car-
diac and respiratory centers will continue to function for a period of
two hours or more on a saline solution in which dog’s red blood corpus-
cles are suspended.

2. An apparatus is described which appears adequate for the aeration
of blood used in perfusion experiments.

3. The effect of calcium predominating over potassium in a solution
containing both of these salts is to stimulate the respiratory center and
to increase the heart-rate. Conversely potassium predominating over
calcium inhibits the respiratory center and slows the heart-rate.

4. When the accelerator nerves are cut potassium increase causes
pronounced cardiac slowing while calcium increase causes slight cardiac
acceleration. When the vagi are cut calcium increase causes definite
cardiac acceleration and augmentation while potassium increase causes
slight cardiac slowing.

5. The results obtained cannot be ascribed to incomplete isolation
of the medulla.

BIBLIOGRAPHY

(1) Hooxesr: This Journal, 1910, xxvii, 24.
(2) Hooxer: Journ. Pharm. and Exper. Therap., 1913, iv, 443.
(3) HerurrzKa: Pfliiger’s Arch., 1911, exxxviii, 185.
(4) WuInTERSTEIN: Pfliiger’s Arch., 1911, cxxxviii, 167.
(5) HirscHFELDER AND Brown: Reported at 1914 meeting, American Society
for Pharmacology and Experimental Therapeutics.
Hooker: This Journal, 1907, xix, 417.
(7) Henperson: This Journal, 1908, xxi, 126.

Oe ee eee

AN INTERPRETATION OF THE MEMBRANE MANOMETER
CURVES AS AFFECTED BY VARIATIONS IN
BLOOD PRESSURE

J. D. PILCHER

From the Pharmacological Laboratories, Schools of Medicine, Western Reserve Uni-
versity, Cleveland, and the University of Nebraska, Omaha

Received for publication June 7, 1915

Introduction. While there is a large literature on the subject of the
tracings formed by the membrane blood pressure manometer, this does
not bear directly on the changes in the membrane manometer curve as
related to the different mechanisms by which the blood pressure changes
were caused. During the past few years a large number of tracings
have collected in the laboratories! in which the membrane manometer
of the Harvard type has been used. As these tracings contain a large
number of curves showing various types of circulatory phenomena,
such as vasoconstriction and dilation, hemorrhage, etc., they have been
studied with the idea of determining if there were not a fairly definite
relationship of this kind; to learn, for example, if such phenomena as
vasoconstriction, cardiac stimulation or depression, etc., are not repre-
sented by characteristic peculiarities in the manometer curves. It
seemed probable that such an interpretation would disclose type curves
that would give an indication of the cause of a change in blood pressure,
which might be of especial value when other, possibly more exact, data
were lacking. This we think has been accomplished.

No reference will be made to the form of the individual pulse record
for the paper will be limited largely to a presentation of the modifica-
tions in the systolic and diastolic pressures caused by various physio-
logical and pharmacological actions on the circulation. The pulse
pressure will also be discussed.

Methods. The tracings were made with a Harvard membrane ma-
nometer attached by a Y tube to the carotid artery; to the other end of
the tube a mercury manometer was attached, damped to give the mean
blood pressure only and to prevent any changes in the membrane ma-
nometer that might arise from oscillations of the mercury. The dog

1 With but few exceptions in the Pharmacology Laboratory at Cleveland.
209

210 J. D. PILCHER

was the usual experimental animal but occasionally the cat was em-
ployed. As the work extended over a period of several years, a number
_ of rubber membranes of different thicknesses were used. This fact
precludes an exact study of the relation between the amount or extent
of the changes in the systolic and diastolic pressures and the degree of
the rise or fall in blood pressure. Another fact that must be borne in
mind is that the higher the blood pressure, the greater the tension on
the membrane, so that a given impulse would probably cause a lesser
change at a high level of pressure than at a low level. Neither of these

factors, however, would alter the type

ea) hoe ; rf ] 1 j 1
Prem = of the curve, which is the question

Aes Movi under discussion.
Pe a) Experiments showing fling of the

lever are disregarded except when this

followed the experimental procedure;
rmmet 14 note will be made of such results.
ae Se Method of presenting the results:

Big. ds pe eee eee eee The results will be presented accord-
themembrane manometer. Upper 18 to the form of the curve displayed
level, the systolic pressure; lower by the tracing rather than by the dif-
level, the diastolic pressure. Type ferent physiological or pharmacologi-
A—formed by general vasocon- ca] procedures. This classification
See Cn (aie Pig ceed ee SS eomn to be somewhat simpler as most
rate. Type B—formed by general ; :
vasodilation and by decreased 0 the curves fall into definite group
heart rate. Type C—formed by types. In all curves the upper limit
cardiac depression and by de- represents the systolic and the lower
creased volume of blood. Type limit the diastolic pressures.
ee RN he ue ts Gs The following type curves include
and by increased volume of blood. res

the great majority of the curves:

Type A—In this type both systolic and diastolic pressures are ele-
vated, but the diastolic much more than the systolic pressure, so that
the amplitude of the pulse excursion is lessened (fig. la).

Type B—Both systolic and diastolic pressures are lowered, but the
diastolic more than the systolic pressure, so that the amplitude of excur-
sion is increased (fig. 1b).

Type C—Both pressures are lowered, but the systolic more than the
diastolic pressure, so that the amplitude is lessened (fig. 1c).

Type D—The systolic pressure is raised somewhat. while the diastolic
remains the same or is slightly lowered or raised; in any case the ampli-
tude is usually somewhat increased (fig. 1d).

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 211

To discuss the individual curves: Type A curve is formed from two
distinct physiological phenomena, viz., vasoconstriction and increased

heart rate.
Vasoconstriction.

This was produced by several procedures, by

faradic stimulation of sensory nerves (sciatic), by asphyxia and by large

doses of nicotine and aconite, all of
which stimulate the vasomoter center;
also by the administration of epineph-
rin, pituitary extract, ergotoxin and ergot
whose primary action is a direct stimula-
tion of the arterioles although cardiac
stimulation plays a part in the action.
A great majority of the curves from
these procedure fall under type A, that
is as the blood pressure rises both systolic
and diastolic pressures are raised but the
elevation of the diastolic pressure is dis-
proportionately great, so that the ampli-
tude of the pulse excursion (pulse pres-
sure) is lessened.

The purest example of this class of
vasoconstriction is that of faradic stimu-
lation of the sciatic nerve (fig. 3 A), for
the cardiac action of the drugs employed
complicates their action somewhat, al-
though as their predominant action is on
the vessels the results are quite similar to
that of pure vasoconstriction. The dias-
tolic pressure is the determining feature
of this type of curve, with the systolic
pressure playing a secondary part. This
point is well illustrated in experiment 33,
figure 2. Four separate stimulations of
the sciatic nerve caused a moderate rise
in blood pressure each time; in each
case the diastolic pressure was raised
considerably while the systolic pressure
remained the same or was very slightly
elevated. The experiments were made
with normal blood pressure and when the

Fig. 2.
stage
type of curve. Experiment 33,
dog; the upper curve sfrom the
membranemanometer; themid-
dle curve the mean blood-pres-
sure from the damped mercury
manometer; the signal line is
the zero blood-pressure. 4—
Sciatic stimulation; note the
moderate rise in pressure, with
the diastolic level raised con-
siderably but the systolic very
slightly.

To illustrate the early
in the vasoconstriction

7 WW J. D. PILCHER

pressure had been lowered by phenol or curare. A further illustra-
tion of the greater importance of the diastolic over the systolic pres-
sure in this type of curve (vasoconstriction) is that as the blood pres-
sure rises on stimulation of the sciatic nerve, the earliest effect to be
observed in the curve is the prompt elevation in the diastolic level,
usually before any change is noticeable in the systolic level.

Huerthle! found that, during the rise in blood pressure from
sciatic stimulation in the rabbit, the pulse pressure was diminished. He,
however, makes no reference to the importance of the diastolic pressure
under such conditions.

Experimental data under type A curve. Faradization of the sciatic nerves:
Forty-five uncomplicated experiments were made on 17 dogs; 35 curves were of
type A and 9 of type D (increase in amplitude by a disproportionate elevation of
the systolic level). The explanation of these exceptional cases is not clear; how-
ever the respiratory improvement and, secondarily, the cardiac improvement by
the relief of a partial asphyxia may account for the increased systolic excursion.
This does not include a large number of curves (19) made under abnormal con-
ditions but all of which followed type A; they will be mentioned later. The exper-
iments were made at all levels of blood pressure from 50 to 160 mm. and in all
experiments the pressure rose considerably (from 10 to 50 mm.). While in gen-
eral it may be stated that the greater the rise in blood pressure, the greater was
the elevation of the diastolic level, yet the change in amplitude (pulse pressure)
seemed to be fairly independent of the rise in pressure, not only in different but
in the same experiments. Probably carefully conducted experiments with this
end in view would show a closer relationship between the rise in pressure and the
variation in the pulse pressure, but it must be borne in mind, as was mentioned
in the introduction, that the higher the pressure the greater the tension on the
manometer membrane, and this would tend to lessen the systolic excursion
while not affecting the diastolic relaxation; this is a factor that would be difficult
toestimate. It seems, then, that the form of the curve is the determining feature
of this type of curve and not the extent of the change in the amplitude of the
excursion.

Sciatic stimulation during low blood pressure from other procedures. The curve
did not differ from that of the normal pressure. This type of curve wasmet
twelve times while the pressure was low from curare, four times from phenol and
three times from nitrite. With both curare and phenol there were experiments
in which sciatic stimulation caused a rise in mean blood pressure with elevation
of the diastolic pressure but without change in the systolic pressure as measured
by the membrane manometer.

Asphyxia. Slight asphyxia was induced three times in one animal; each time
as the blood pressure rose somewhat the manometer curve assumed the form of
type A curve.

Aconite and Nicotine. With the great stimulation of the vasometer center
from these drugs, type A curve was met once from aconite and twice from nico-
tine. In two other experiments with nicotine the pulse excursions were in-

1 Quoted from Erlanger and Hooker (1).

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 213

creased somewhat by the greater elevation of the systolic level. As nicotine
stimulates the heart as well as the vasomotor apparatus, this may explain the
increase in systolic excursion. In a single instance nicotine paralyzed the vaso-
motor center; the blood pressure curve assumed type B, typical of vasomotor
depression which is to be described later.

A B

Fig. 3. Type A Curve—To illustrate vasoconstriction and increased heart
rate. Upper curve is from the membrane manometer; middle curve is the
mean blood-pressure from the damped Hg manometer. A—Sciatic stimulation;
experiment 28. B—Injection of pituitary extract immediately after curve A.
C—Injection of epinephrin; experiment 34. D—Section of the vagi; experiment
79. A, B and C illustrate the rise in pressure from vasoconstriction; D from
section of the vagi.

Epinephrin. The curve of epinephrin resembles that of sciatic stimulation
very closely in the majority of experiments (7 of 10). (Fig. 3C) As the blood
pressure rises the diastolic pressure usually is elevated before there is any change
in the systolic pressure, and the maximum decrease in the excursion occurs, as a
rule, during the rise in pressure; at the maximum level the amplitude of excur-
sion becomes larger but still remains below the normal. Tbisincrease in the

214 J. D. PILCHER

excursion at the height of the curve may be due to the cardiac stimulation, for
three times the amplitude was above the normal at this point but early in the
rise in pressure there was a brief period in which the amplitude was lessened
solely by the elevation of the diastolic pressure; these experiments are the three
exceptional ones mentioned above so that in reality all the epinephrin curves
follow type A. In these exceptional experiments, however, sciatic stimulation
also caused an increased pulse pressure. When epinephrin is given in acute car-
diac failure, of course the rise in pressure is accompanied by a greater increase in
systolic than in diastolic pressure.

Pituitary extract. Type A curve was formed in five of seven experiments; in
the other two both pressures were equally elevated; possibly the cardiac element
was more marked in these cases. The curve at times resembles that of sciatic
stimulation very closely as is shown in figure 3B, in which sciatic stimulation

immediately preceded the injection of the pituitary extract. There may be .

considerable decrease in the excursion solely by the elevation of the diastolic
level, with the systolic level remaining the same: in experiment 227, on the injec-
tion of pituitary extract the blood pressure rose from 75 to 130 mm. and the
excursion was lessened from 24 to 10 mm. solely by the elevation of the diastolic
pressure; however, too much stress should not be placed on but one experiment
for there may have been a certain amount of fling of the lever that was not noted
during the experiment and this would have given a greater original systolic excur-
sion than really existed; the point that is to be emphasized is that the diastolic
pressure is the determining feature of the pituitary curve.

Ergotoxin and ergot. Each drug formed type A curve in but a single experi-
ment. ‘

Type A curve from increased heart rate. The general form of the type
is maintained in this group but as a rule the blood pressure rose more
gradually than in the members of the previous group (especially sciatic
stimulation and epinephrin) so that the changes were not so pronounced,
although in a few instances there was little difference between the two
classes (fig. 3D). The heart rate was increased by section of the vagi
or by the administration of atropin to paralyze the vagal endings.

Vagus section. In nine instances this resulted in an increased heart rate and
a rise in blood pressure (average 35mm.). All manometer curves assumed type
A with moderate decrease in the amplitude of excursion, Jargely by the elevation
of the diastolic pressure; occasionally the systolic level was but very slightly
changed, if at all.

Atropin. When atropin caused an increased heart rate and a rise in blood
pressure, type A form of curve was found in three of four experiments; in the
fourth case type D was formed.

Erlanger and Hooker (1) have pointed out that acceleration of the
heart, such as occurs after section of the vagi, as a rule diminished the
pulse pressure, a statement that these results confirm.

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 215

Type B curve. In this type the amplitude of the excursions is in-
creased by a great lowering of the diastolic pressure; the systolic pressure
is also usually lowered but exceptionally may remain unaffected. The
type is met during the fall in blood pressure from two distinct phenom-
ena, vasodilation and cardiac slowing, conditions that are just the re-
verse of those that produced the preceding type of curve.

eG) 9: 3) 7 ae
d

Cc

Fig. 4. Type B Curve—To illustrate the effect of the fall in blood-pressure
from cardiac slowing and vasodilation. Upper curve is that of membrane
manometer; the middle curve is the mean blood-pressure from a damped Hg
manometer; the signal line is the zero blood-pressure. A—Cardiac slowing from
cevadin; experiment 49. B—Vasodilation from nitroglycerin; experiment 90.
C—Vasodilation from amyl nitrite; experiment 55.

The vasodilator group. The vasodilation was induced most typically
and in the purest form by the administration of nitrites—amyl and
sodium nitrites and nitroglycerin (fig. 4B, C); also by chloroform,
phenol, atropin (in large doses only), and by the intravenous adminis-
tration of curare which probably causes peripheral vasodilation. The
typical dilator action of chloroform and phenol, however, is modified

216 J. D. PILCHER

by their cardiac action. In this group the heart rate was practically
unaffected.

The Nitrites. In all the experiments (19) with the nitrites as the blood pres-
sure fell the diastolic level was greatly lowered, while the systolic level, though
usually also lowered considerably, was frequently unaffected or but very slightly
lowered; this caused a great increase in the amplitude of the excursion of the
pulse-pressure. Too much emphasis should not be placed on the occasionally
unchanged systolic level as the sudden emptying of the heart into the dilated
vessels would undoubtedly tend to cause fling of the manometer lever. It is
rather striking that there seemed to be a tendency for the systolic level to re-
cover to the normal somewhat earlier than the diastolic level, but as many of the
experiments were interrupted by other procedures, the evidence for this point is
not complete.

There was no material difference in the curves of the three members of the
group, amyl and sodium nitrites and nitroglycerin. There were twelve experi-
ments with nitroglycerin (0.5 to 1.0 mgm. per kilogram), five with amyl nitrite
and two with sodium nitrite (5and10 mgm). In each case there was a pronounced
fall in blood pressure, varying from 25 to 65 mm; the experiments were made at
all levels of pressure from 70 to 205 mm, without showing any striking differences
in the manometer curves. The amplitude was more than doubled in several
curves.

Chloroform. This has a twofold action, depressing both the heart and vaso-
motor systems. When administered not too rapidly by inhalation the vasomotor
action predominates; when given by vein in suitable dosage direct cardiac de-
pression can be obtained. The intravenous results will be discussed under the
cardiac depressants.

Chloroform was given by inhalation seven times to six ‘wale and in all of
them type B curve was formed as the blood pressure fell. In asingle experiment,
in which the administration was pushed rather rapidly, the systolic pressure fell
about as much as the diastolic, thus approaching the cardiac depression type of
curve; the blood pressure also fell more rapidly in this experiment. The average
fall in pressure was 50 mm., extremes of 20 and 90 mm.

Phenol. As the blood pressure fell from the intravenous administration of
phenol in six of seven experiments type B curve was formed, thus indicating that
the cause of the fall in pressure was vascular, at least in part, and for this there
is direct evidence from the perfusion method elsewhere described (2). Direct
cardiac depression is the other factor in the fall in pressure; there is, however,
but a single instance in which there is evidence for this statement: in one experi-
ment the systolic level fell more than the diastolic, thus decreasing the amplitude
of the excursion in a manner that indicates cardiac depression (to be discussed
later). In all experiments the blood pressure fell sharply, an average of 45 mm.,
extremes of 25 and 65 mm.

Atropin. When atropin caused a fall in blood pressure the curve assumed type
B, indicating that the fall in pressure was due to vasomotor depression as the
heart rate was not materially affected under the conditions of the experiment.

There were six experiments in which atropin in doses of from 0.1 to 1.0 mgm.
per kilogram, and single experiments with 5 and 10 mgm., caused a mean fall in

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 217

blood pressure of 15mm. (extremes 5 and 72mm.). All but one tracing followed
type B. The amplitude was increased a variable amount, usually moderately,
but in one experiment from 5 to 35 mm., part of which was probably due to fling
of the lever. The heart rate was somewhat increased twice and decreased once;
in the other experiments there was practically no change in rate as the vagi were
divided or previous doses of atropin had been given.

Curare. During the fall of blood-pressure from the intravenous administra-
tion of curare there is always a great lowering of diastolic pressure, and usually,
but not always, the systolic pressure falls also; in either case there is a marked
increase in the amplitude of the cardiac excursions. The lowering of the dia-
stolic level precedes the fall in systolic pressure, and the former falls abruptly
while the latter falls gradually, indicating that the lowering of the diastolic level
is the essential cause of the fall in blood pressure. As this curve is of the same
type as the nitrite curve it gives further evidence that peripheral vasomotor de-
pression is the’ cause of the curare fall; peripheral because the vasomotor center
is not depressed but rather stimulated (2).

Curare was given a total of 23 times to 23 dogs. The average fall in pressure
was 40 mm. (extremes 12 and 95 mm.). In several experiments the systolic
pressure was either raised slightly (five times) or remained unchanged (twice);
with but one exception the blood pressure fall was considerably less than in the
other experiments, so that when the curare action is marked the systolic level
also falls. Fling of the lever may also account for the maintenance of the sys-
tolic pressure in such experiments. When the minimum blood pressure was
reached, usually the systolic pressure had fallen considerably so that the pulse
pressure was lessened, although there were numerous exceptions as noted.

Hales,” and later Bernard,” pointed out that a fal] in blood pressure increases
the amplitude of the pulse. Our observations confirm this statement as regards
the fall in pressure from general vasodilation and from decreased heart rate; we
wish to emphasize that the increase in the pulse pressure is brought about largely
by the disproportionate lowering of the diastolic pressure. The lesser change in
the systolic level is nicely put by Marey:? ‘‘If we assume that the more or less
rapid and more or less abundant entrance of blood into the arterial system result
from the excess of ventricular pressure over the pressure in the arterial system,
then it is evident that a fall of arterial pressure would be exactly similar to an
increase in the force of the heart beat, and there is, therefore, nothing surprising
in this augmentation of the force of the pulse in all conditions that lower the blood
pressure.’’

The effect on the pulse curve of lowered blood pressure from hemor-
rhage will be discussed later.

Type B from decreased heart rate. The curve is similar to that of the
vasodilator group except that usually the systolic level is not lessened
to as great an extent. The curve is simply that of the “vagus pulse”
characterized by a prolonged diastole, during which the pressure falls
greatly, while the systolic pressure is but moderately lowered or may

? Quoted from Erlanger and Hooker (1).

218 J. D. PILCHER

even remain unaffected; in any case the pulse pressure is greatly
increased.

The heart rate was decreased by faradic stimulation of the peripheral
end of the divided cervical vagus, by the intravenous administration of
veratrum viride and its alkaloid cevadin, and in a single experiment by
aconite.

Veratrum viride, as the tincture, was given in doses of from 1 to 10 mgm. per
kilo of body weight; cevadin in 0.05 mgm. per kilo doses. In all there were 13
experiments on 11 animals and the two drugs gave similar results. With the ad-
ministration of the drug the heart rate was markedly decreased so that the blood
pressure fell from 30 to 70 mm. in each experiment. All experiments gave type
B form of curve (fig. 4A).

Vagus stimulation gave results similar to veratrum in three cases. In other
experiments the heart was completely inhibited so that the systolic pressure fell
to zero.

Type curves C and D. In the preceding curves the characteristic
feature has been the changes in the disatolic pressure; in the curves to
be described next (C and D) the characteristic feature is the systolic
variation in pressure although the diastolic pressure may also be con-
siderably changed. In type C both levels are lowered, but the systolic
much more than the diastolic level so that the pulse pressure is lessened.
This type is met typically following sudden arterial hemorrhage and
depends upon a decreased quantity of blood (fig. 5 B); and from cardiac
depression. In type D the systolic pressure is somewhat elevated while
the diastole remains unaffected or is slightly raised or lowered. The
curve follows an increase in the volume of blood, either from the injee-
tion of defibrinated blood or normal saline solution into the bled animal,
or from the injection of saline solution into the normal animal (fig.
5 A, C). Type D is also met from cardiac stimulation, although the
experiments showing pure cardiac stimulation are somewhat limited in
number. As the two curves follow one another experimentally it is
simpler to discuss them together.

Hemorrhage. The animals were bled rapidly from the femoral artery
in successive small hemorrhages until a very low point in blood pressure
was reached; then either the defibrinated blood or normal saline solu-
tion, or each in succession, was injected into the femoral vein. At the
onset of the hemorrhage the systolic level usually fell promptly, and the
diastolic level also fell but to a much lesser extent, so that the amplitude
of the pulse excursions was lessened. This effect was met with small
hemorrhages of but 5 cc. per kilo and increased with succeeding hemor-

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 219

rhages, the pulse pressure becoming smaller and smaller as the total
amount of blood withdrawn increased. The individual experiments
showed considerable differences in the extent of the change in the max-
imum and minimum pressures because of the variation in the thickness
of the membrane, so that ‘t is not feasible to make quantitative com-

Fig. 5. Types C and D Curves—To illustrate the effect of saline infusion
and hemorrhage; experiment 18, dog; the upper curve is from the membrane
manometer; the middle curve the mean pressure from the damped Hg manometer.
Curve A—Saline infusion before hemorrhage. Curve B—Hemorrhage, two por-
tions of 10 cc. each. Curve C—Saline infusion after hemorrhage. Curves A and
B have been interrupted.

parisons. Between the successive hemorrhages there was _ usually
very little recovery in the systolic pressure but as a rule the diastolic
pressure recovered somewhat.

Reinjection of blood and saline solutions. When the blood pressure
had fallen to the minimum point compatible with safety, the defibrinated
blood or saline solution were injected. As the pressure rose the ma-

220 . -» J. D. PILCHER

nometer curves showed changes practically the reverse of those formed
from the withdrawal of the blood, i.e., while both the systolic and dias-
tolic pressures were raised, the systolic was raised more than the dias-
tolic level so that the pulse excursions were increased. If the injections
were sufficiently large, the curve returned to the normal.

The injection of saline solution into the normal animal. This resulted
in a constant, though variable, elevation ‘of the systolic pressure; the
diastolic pressure was variable, slightly raised, unchanged or lowered a
little. The amplitude of the excursions were increased by the predomi-
nance of the systolic rise. The slight lowering of the diastolic pressure
is of little import as it may have been caused by fling of the lever.

Daia on hemorrhage and saline infusion. There were 16 experiments on 12
dogs. The blood was withdrawn in 5 to 10 ce. portions until the minimum point
in pressure was reached, about 30 to 40 mm. In 12 experiments the systolic
pressure was lowered disproportionately so that type C curve was formed; in
one experiment the maximum and minimum pressures were equally lowered and
in three the curve was irregular but the diastolic pressure recovered before the
systolic which is the usual occurrence. When the defibrinated blood (5 experi-
ments) and saline solution (2 experiments) were injected there was a prompt
elevation of the systolic level while the diastolic level was but.slightly affected
as mentioned above until rather large amounts were injected when conditions
returned toward the normal.

Saline infusion into the normal animal. Ten experiments were made on six
dogs, injecting from 5 to 35 cc. per kilo. All experiments gave an increased am-
plitude of excursion after type D curve. The diastolic pressure was slightly
raised four times and slightly lowered or unaffected each three times. In each
experiment the blood pressure rose moderately, average 13 mm., extremes of 4
and 20 mm.

Type C curve from cardiac depression. Direct cardiac depression was
produced by the intravenous injection of chloroform or by saline solu-
tion saturated with chloroform. The drug, of course, depresses the
vasomotor system as well, but by giving small doses by vein the primary
action seems to be direct cardiac depression as the action is of such brief
duration and recovery prompt and complete; in one instance, however,
there was evidence of marked temporary depression of the vasomotor
center without altering the cardiac depression form of curve. On the
injection of 0.1 to 0.2 cc. of chloroform the blood pressure fell imme-
diately and considerably, but returned to the normal usually within
two to four minutes. During this period the cardiac excursions were
greatly lessened, largely by a great lowering of the systolic level although
the diastolic level was also lowered somewhat (fig. 6). Usually, also,

OKC Oe

INTERPRETATION

OF MEMBRANE MANOMETER TRACINGS 221

the diastolic pressure recovered before the systolic, which indicates

weakening of the heart.
except in three instances when there was consider-
able slowing without altering the type of the curve.
There were ten such experiments on six animals.
The blood pressure fell an average of 45 mm., ex-
‘tremes of 20 and 67 mm.

Cardiac depression was also observed in ten ex-
periments from asphyxia when the vasomotor center
was paralyzed, thus preventing the usual rise in
pressure from central vasomotor stimulation. In
each instance as the pressure fell there was a grad-
ual lowering of the systolic pressure before the
diastolic pressure fell as well, thus following type C
curve.

Experiments showing cardiac stimulation are sug-
gestive only, for they are not free from the element
of vasomotor action. Thus strophanthus gives the
curve of vasoconstriction somewhat more frequently
than what may be considered the type of cardiac
stimulation, that is with the systolic level raised
more than the diastolic level, thus increasing the
excursion. With caffein the number of experiments
with good membrane manometer curves is limited
and these show a variable response: In two in-
stances stimulation was indicated by a rise in blood
pressure, the amplitude of excursion was either in-
creased by a greater elevation of the systolic level
or it remained unchanged with both pressures some-
what raised, both*curves indicating cardiac stimula-
tion. In other experiments, with fall in pressure,
the systolic level was twice lowered more than the
diastolic level, indicating cardiac depression; in still
another instance vasodilation was indicated. In
the acute fall in pressure from the intravenous in-
jection the curve was also irregular, indicating
either cardiac or vasomotor depression or both,

The heart rate was not materially affected

Fig. 6. To illus-
trate cardiac de- -
pression. Experi-

ment 91, dog; the
upper curve is from
the membrane
manometer; the
middle curve the
mean pressure from
the damped Hg
manometer; signal
line is the zero pres-
sure. At the signal
chloroform in
line solution was
given intrave-
nously.

Sa-

The irregularity of the caffein curves is readily appreciated when it is
recalled that the drug primarily stimulates the heart and depresses the
vasomotor apparatus and later depresses the heart as well.

222 J. D. PILCHER

Discussion. After the experimental results had been classified as in
the text, the explanation of the formation of the type curves was sought
on a priori grounds and are here presented. The main factor in the
formation of the curves is, of course, the physiological processes, modi-
fied, it may be to a certain extent, by the mechanical influences of the
rubber membrane of the manometer. As the various types of phenom-
ena were classified in the text as if they occurred in what may be called
the pure state, that is, unmodified by other circulatory influences, the
same plan will be followed in this discussion. It must be borne in mind,
however, that the various factors that determine the circulation inter-
act with, and modify, one another, so that probably a strictly “pure”
type does not exist.

The vasoconstriction type. This is characterized by the dispropor-
tionate elevation of the diastolic level. The diastolic pressure is deter-
mined by the rate of blood-flow through the arterioles to the venous
side of the circulation, therefore the greater the resistance offered by
the arterioles during the vasoconstriction, the less will be the passage of
blood through them and consequently the higher the diastolic pressure
will rise. The systolic pressure is determined by the rate and force of
the discharge of blood into the aorta; the greater the resistance in the
aorta the smaller the quantity of blood that can be forced into it with a
given strength of heart beat. The higher the diastolic pressure rises the
nearer it approaches the systolic intracardiac pressure, and this means
that with each systole a lesser quantity of blood will be discharged into
the aorta because of the increasing resistance to the discharge, and the
added increment of pressure will be less and less as the diastolic pressure
rises. Therefore, as was stated in the text, it is the diastolic pres-
sure that determines the vasoconstriction type of curve, with the
systolic pressure playing a secondary part. The mechanical factor
that might influence the form of the curve is that the higher the blood
pressure the greater the resistance offered by the rubber dam of the
manometer. This would tend to lessen the systolic excursion while not
affecting the diastolic relaxation. However, as this type of curve is the
same at all levels of blood pressure and with different manometer mem-
branes, the mechanical factor is evidently of little importance in deter-
mining the type of curve.

The vasodilation type, characterized by excessive depression of the dias-
tolic pressure is lowered by the rapid escape of blood through the dilated
arterioles into the venous side of the circulation. The systolic pressure
tends to approach the normal as the expulsive force of the heart remains

INTERPRETATION OF MEMBRANE MANOMETER TRACINGS 223

practically unchanged and because of the lessened resistance in the
aorta to the systolic discharge. Thus the determining feature of the
vasoconstriction curve is the diastolic pressure.

Cardiac depression. The systolic level is lowered by the lessened
force of the eardiac contraction. The diastolic pressure is primarily
unaffected as the resistance to the blood-flow from the arterial to the
venous side is primarily unaffected.
~ Cardiac stimulation. The systolic level is raised because of the in-
creased force of the systolic contraction. The diastolic pressure is not
materially affected.

Change in the volume of blood. - Decrease in the volume of blood would
result in a lessened discharge, hence a Jowered systolic pressure. In-
crease in the volume would result in a greater systolic output and a
higher systolic pressure. In either case the diastolic pressure would be
a secondary feature. This agrees with the experimental findings. The
change in the systolic pressure was greater in the case of hemorrhage as
the compensatory mechanism would not be as efficacious as when the
volume of the blood is increased by the addition of saline solution.
When the infusion followed hemorrhage the increase in the systolic
level was correspondingly great.

To recapitulate in another way. The diastolic pressure is determined
mainly by the outflow through the arteries during the diastolic pause
and it is, therefore, affected mainly by the calibre of the vessels and by
the rate of the heart; so that changes which affect mainly the diastolic
pressure are either vasomotor or heart rate. The systolic pressure is
determined mainly by the pressure with which the heart forces blood
into the vessels and is, therefore, determined by the force of the cardiac
contraction and by the volume of blood at the disposal of the heart; so
that changes in the systolic pressure implies either variations in the
force of the cardiac contractions or in the volume of the blood at the
disposal of the heart.

Conclusions. The membrane manometer gives evidence of the cause
of a variation of blood pressure by the form of the curve assumed.
These curves are classified in the text.

Vasoconstriction is indicated by the elevation of both diastolic and
systolic pressures, but the former much more than the latter, so that
the amplitude of the excursion (pulse pressure) is greatly lessened.

Vasodilation is indicated by a lowering of both pressures, with the
diastolic pressure lowered much more than the systolic, so that the
amplitude is greatly increased. The diastolic pressure is the determin-
- ing feature in both vasoconstriction and dilation.

224 J. D. PILCHER

When the heart rate is increased sufficiently to raise the blood pres-
sure, the curve resembles vasoconstriction; while a lowered pressure
from decreased rate is similar to that of vasodilation.

Following hemorrhage both systolic and diastolic pressures are low-
ered, but the systolic disproportionately, so that the amplitude of the
excursions are lessened. The infusion of blood or saline solution in-
creases the excursion largely by elevating the systolic level. In both
hemorrhage and infusion the systolic pressure largely determines the
form of the curve.

I am glad to express my indebtedness to Dr. Sollmann for many help-
ful suggestions, especially in the discussion of the results.

LITERATURE

(1) ErRuanGeR AND Hooker: Johns Hopkins Hospital Reports, 1904, xii, 155.
(2) SoLuMAN AND PiucHER: This Journal, 1910, xxvi, 233.

EEE Ee

STUDIES ON THE HYDROGEN-ION CONCENTRATION IN
BLOOD UNDER VARIOUS ABNORMAL CONDITIONS

M. L. MENTEN, M.D. anp G. W. CRILE, M.D.

From the Cushing Laboratory for Experimental Medicine, Western Reserve
University, Cleveland, Ohio

Received for publication June 16, 1915

INTRODUCTION

Since the application by Hoéber (1) in 1900 of the Nernst hydrogen
concentration chain to the measurement of the reaction of the blood, our
knowledge of that subject has been greatly extended. Many modifi-
cations of the original method have been adopted by various investi-
gators, but practically all the results published reveal the fact that the
reaction of normal blood lies within comparatively narrow limits, vary-
ing according to the earlier observers between 0,3.10~7 and 0,7.107~7
(2) and according to the later and more accurate studies of Hassel-
balech (3) on Mammalia at body temperature between 3,5.10~* while
Michaelis (4) records an average value of 2,75.10~° at a temperature
of from 18°C to 20°C. Both these later observers have noted that
venous blood is always slightly more acid than arterial blood, due,
probably, to the greater carbon dioxide tension in the former.

Even the investigations of blood in pathological conditions, where
acidosis occurs, have failed to show any considerable deviation from the
narrow limits already quoted, and only in diabetes, (5) (6) where deep
coma has been reached, does there occur any demonstrable increase in
the hydrogen-ion concentration of the blood, although oxybutyric and
diacetic acids are known to be produced in not inconsiderable amounts
in that disease. The only remaining conditions in which an increased
acidity of the blood has been observed, where the technique is free from
serious criticisms is that of narcosis, reported by Michaelis (7).

The investigations outlined in the following pages deal with changes in
the hydrogen-ion concentration, noted in animals in various conditions.

225

226 M. L. MENTEN AND G. W. CRILE

METHODS

The method employed for estimating the acidity of the blood was that
of measuring the hydrogen-ion concentration by means of the gas chain;
and since the technique used was essentially that already published in
detail by Michaelis (8) no further description need be given here. It
may be mentioned however that all measurements were made on animals
at laboratory temperature varying from 18°C to 24°C., and for these
variations corrections were made. The quantities of blood used vary-
ing from 2.5 ec. to 3.5 ce., were diluted to 6.5 cc. with 0.85 per cent
sodium chloride containing hirudin in solution. When possible the
blood was allowed to drop directly into the electrode from the ear of the
animal employed; in other cases a cannula wasinsertedin the vessel and
the blood passed from this into the electrode. In every experiment
duplicate readings were made. —

Instead of expressing our results as the hydrogen concentration
C,, = 10" as has been done in the introduction, we have followed the
Sérensen procedure in using throughout the hydrogen-ion exponent
pH = p.

RESULTS AND DISCUSSION

In some of our earlier estimations of the hydrogen-ion concentrations
of the blood it was observed that in anaesthetized animals the acidity of
the blood was considerably higher than that reported for normal ani-
mals. A systematic study was then made of the blood of animals anaes-
thetized with chloroform, ether and nitrous oxid. For these experi-
ments dogs and rabbits were chosen because of the comparative ease in
obtaining blood from the ear blood vessels of these two animals. As
previously stated the blood was bled from an incision in the blood vessel
of the ear directly into the electrode; the same animal was then anaes-
thetized and blood again taken from the same site in a similar manner
so that in every case the blood in anaesthesia could be compared with
the normal blood of the same animal.

In a large number of experiments the values obtained for normal
blood show a wide variation, ranging in the dog from pH = 7.64
to pH = 7.32. In the rabbit the variations are even more marked, the
minimal and maximal figures obtained being pH = 7.18 and pH = 7.70,
a phenomenon which will be discussed later.

Under the influence of the three above mentioned anaesthetics, when
anaesthesia reaches the stage in which the reflexes are completely

HYDROGEN-ION CONCENTRATION IN BLOOD 227

abolished, the reaction of the blood measured at 20°C. may fall to
pH = 7.00. Although our experimental studies of the blood reaction in
anaesthesia number over twenty-five, no protocols are recorded here
because the degree of acidity produced is, as far as could be ascertained,
dependent solely upon the amount of the drug inspired, if corrections
are made for the different weights of the various animals. Correspond-
ing more or less closely to the degree of anaesthesia therefore, values
extending from pH = 7.00 to pH = 7.60 may be obtained. The deeper
the anaesthesia the more marked the acidity, and vice versa; also pro-
longed duration of the anaesthesia is apparently not a factor in increas-
ing the hydrogen-ion content of the blood. Further, the increase in
acidity commences as soon as the anaesthetic enters the circulation,
and a diminution takes place immediately on the cessation of the ad-
ministration of the drug, until 45 minutes subsequently the blood again
resumes its normal hydrogen-ion concentration.

Narcosis produced by morphia, even when this drug is given in such
large doses as 75 cc. of 1 per cent morphia sulphate in two hours, fails to
reveal any demonstrable changes in hydrogen-ion concentration.

As it is interesting on account of the prolongation of the duration of
the induced acidity, mention is made of one experiment with ethyl
aleohol, where a dog whose normal blood reaction was pH = 7.47 was
given by stomach tube a mixture of 60 ec. of 95 per cent alcohol in 40 ce.
of water. Forty-five minutes subsequently the blood acidity was
pH = 7.32 and 6 hours later, although the animal had apparently
gained complete consciousness, the reaction remained at the high value
of pH = 7.28.

Finally, in two cases of surgical operations in the senior author’s
clinic, during nitrous oxide anaesthesia, measurements showed the blood
reaction to be respectively pH = 7.20 and pH = 7.22.

Not only in anaesthesia but also in other abnormal states the blood
reaction was observed to become more acid. The most striking of
these was fright, a phenomenon remarkably shown in the rabbit. Ordi-
narily if blood is taken from the ear of a rabbit, so that the animal is
disturbed as little as possible, the blood shows a fairly high alkalinity.
Under the influence of fright this is altered in an astonishingly short
time and to a most profound extent. Another noteworthy feature is
the rapidity with which the blood regains its normal reaction. The
following protocols will serve to illustrate these facts:

I. Normal blood reaction of rabbit was pH = 7.67. Animal was
comfortably placed in a box covered with a wire screen and frightened

228 M. L. MENTEN AND G. W. CRILE

for ten minutes by a barking dog. Blood immediately taken from the
ear blood vessels showed an acidity of pH = 7.17. The rabbit was
then allowed to rest quietly 45 minutes and again frightened, directly
following which the animal was bled, and the blood reaction was found
to be pH = 7.34. After being allowed to rest quietly for 30 minutes,
blood examined in like manner gave a value of pH = 7.60.

II. Rabbit whose normal blood was pH = 7.60, after being frightened
for 5 minutes in the manner above described, showed a hydrogen-ion
concentration of pH = 6.98. Fifteen minutes subsequently the blood
again gave anormal reaction. No signs of distress other than shallower
and more frequent respirations seemed to accompany the increased acid-
ity and the animal apparently suffered little inconvenience during or
subsequent to the fright and 15 minutes afterward appeared to be quite
normal again. It is probable however that such a process frequently
repeated throughout life must leave some traces. It is indeed, most
probable that the varying figures obtained for the normal blood of the
rabbit noted earlier in this paper, and which have been previously re-
ported by Hasselbalch and Lundsgaard (9) have their explanation in the
above phenomenon, the repetition of which during the life of the rabbit
may account for the varying pathological conditions of the blood vascu-
lar apparatus of that animal, encountered by so many investigators and
which, therefore, make it such an undesirable experimental animal.
Similar observations regarding such marked effects of fright have been
very rarely observed by us in dogs.

Since the stimulus of fright can produce such a prompt and intense
response it is to be surmised that in other emotional disturbances, sim-
ilar changes in acidity would be encountered. Experiments to test
this supposition were made for anger in cats, and it was found that the
expression of that emotion is also accompanied by profound changes in
the blood of the animal. Unfortunately the cat is not as desirable for
such experiments as the rabbit. To obtain blood from the cat without
general anaesthesia is much more difficult than from the rabbit, largely
because in the former the blood vessels must be exposed under local
anaesthesia. Further if the operation is not skillfully and quickly done
so as to minimize the pain and eliminate as far as possible the excite-
ment of the animal, the blood from the beginning assumes the high acid
value indicative of emotional disturbance. The following abbreviated
protocols serve to illustrate the character and results of the experiments.

I. Normal blood withdrawn by syringe from the femoral artery, ex-
posed under cocaine, had a value of pH = 7.43. The cat was angered

HYDROGEN-ION CONCENTRATION IN BLOOD 229

by a barking dog and in blood then withdrawn from the femoral artery
acidity was pH = 7.20. The animal was kept tied down and at the
end of an hour the blood reaction was pH = 7.37.

II. Cat was tied down and femoral artery exposed under cocaine.
Normal blood withdrawn with syringe had a hydrogen-ion concentra-
tion of pH = 7.20. Cat was then frightened for 15 minutes and blood
acidity had risen to pH = 7.02. It may be noted, that in these experi-
ments on cats the normal reaction of the blood acidity is invariably
high, due, doubtless, to the effect of fright and some unavoidable distress
of the animal during the experiments.

The possibility that the cause of the increased acidity in the above
reported experimental data was an accelerated oxidation with the
accumulation of its accompanying products in the circulation led to
the estimation of the blood reaction during and immediately follow-
ing insomnia, where it was thought the prolonged metabolic processes
taking place might give rise to a similar condition of the blood. But
tests on a series of six rabbits, housed in a warm place, abundantly sup-
plied with food and water but kept awake for 100 consecutive hours,
were quite negative on this point.

Evidence of a similar sort is also obiained by the following experi-
ments. In pithed animals immediately following the severing of the
spinal cord there is a rapid rise of hydrogen-ion concentration, due un-
doubtedly to the decreased lung aération and the consequent faulty
elimination of carbon dioxide, for on the establishment of artificial
respiration the normal reaction of the blood is promptly restored.
When in such a pithed animal where a normal hydrogen-ion concentra-
tion was being uniformly maintained by artificial respiration, violent
contractions of the skeletal muscles were produced by electrical stimu-
lation of the cut end of the cord in communication with the musculature
of the thorax, abdomen and extremities, no evidence of an increased
acidity of the blood was found except when the animal was dying.

One other condition in which the blood showed an increased acidity
was that of shock. Only two experiments were performed but as the
results in both cases were very definite, they are now outlined.

I. Blood from dog’s ear had a hydrogen-ion concentration of
pH = 7.61. The animal was anaesthetized with ether, the abdomen
opened and intestines manipulated for 12 hours at the end of which
time the blood pressure was exceedingly low. Thirty minutes subse-
quent to this manipulation and 40 minutes after the ether was discon-
tinued, the acidity of the blood was pH = 7.11; and 60 minutes follow-

230 M. L. MENTEN AND G. W. CRILE

ing intestinal manipulation and 70 minutes after the cessation of the
administration of the ether, the acidity had risen to pH = 6.98.

II. Reaction of dog’s blood at the beginning of the experiment was
pH = 7.48. Shock evidenced by a marked fall in blood pressure was
produced by a manipulation of the intestines. One hour after ether
was discontinued, blood from the vena cava had a value of pH = 7.08.
Blood pressure at this time was so low that sufficient blood for a
reading could not be withdrawn from the femoral artery with a syringe.
That the increased acidity was not due to imperfect elimination of ether
was indicated by the return to consciousness of both animals in from
25 to 30 minutes after the anaesthetic was discontinued.

In addition to the conditions above reported, where increased acidity
of the blood obtains, is appended lastly a very interesting observa-
tion where a diminution of the hydrogen-ion concentration was noted;
namely, in the blood flowing from the adrenal gland. In order to ob-
tain blood which would contain, as far as possible, the maximal con-
centration of adrenal secretion, it was found necessary to clamp the
adrenal vein just before it joins the vena cava, on the one side, and be-
fore it passes over the adrenal gland on the other side. The portion of
the vein, between the clamps therefore, lay directly over the gland and
contained only blood from that organ. When this blood was with-
drawn with a syringe and measured the value of the hydrogen-ion con-
centration was always from pH = 0.10 to pH = 0.12 lower than the
blood taken from the immediately adjoining vena cava or from that
part of the adrenal vein lying distal to clamp. While this difference is
not large it isa very constant feature of adrenal blood. The assumption
that this increased alkalinity is due to adrenalin is supported by the
fact that the addition of adrenalin to blood serum, whose hydrogen-ion
concentration is known, lowers its acidity and the diminution corre-
sponds to the weight of adrenalin added when this is below the amount
necessary for saturation. Thus if adding a definite amount of this sub-
stance, which as has been shown by Aldrich (10) and others to be a very
strong base, causes a certain increase in the hydroxyl ion concentration,
then when twice or four times that amount is added the increase in
alkalinity is multiplied by two and four respectively. Since at no time
was it possible, with the apparatus used, to obtain blood from the vena
cava, in the immediate vicinity of the opening of the adrenal vein which
was appreciably more alkaline than blood from any other part of the
vena cava, the adrenalin apparently causes a measurable modification
of the blood in a very limited area. Moreover, that the influence of the

HYDROGEN-ION CONCENTRATION IN BLOOD aa

adrenalin on the circulatory fluids is local in character is further evi-
denced by the fact that the removal of the gland caused no change
whatever in the reaction of the blood until the animal was moribund.
Although blood from the pancreas, liver and thyroid, as well as from the
internal and external jugular veins was compared with blood from the
vena cava of the same animal, no evidence of any variation in the hydro-
gen-ion concentration could be obtained. Further the removal of these
organs singly or in combination caused no change in the reaction of the
blood.

The data presented in this paper proves conclusively that a very
marked increase in the hydrogen-ion concentration may occur under cer-
tain abnormal conditions, and that the existence of this high acidity of
the blood is not,incompatible with life. As to the cause of the phenom-
enon and regarding its significance, the authors feel the present data to
be too limited to warrant much speculation. Concerning the first point,
however, one or two suggestions may not be amiss. The immediate
and intense response of the blood to emotional stimuli by a marked rise
in the hydrogen-ion concentration with the corresponding alteration of
the character of respiration, and the rapid disappearance of these on
the removal of the exciting cause, indicate that under these cireum-
stances the carbon dioxide is a major factor; it is obvious however that
all increased acidity of the blood cannot be ascribed to this source.
This explanation does not suffice for shock and anaesthesia, since in the
former the carbon dioxide in the blood is markedly diminished as has
been shown by Henderson (11), and in the latter the carbon dioxide in
the blood is increased according to the researches of Buckmaster and
Gardner (12). Indeed in such a composite fluid as blood, containing
so many complex chemical compounds, it is quite conceivable as has
been suggested by Robertson (13) that the amphoteric character of
certain proteins must be of extreme importance in regulating the acidity
of the blood.

SUMMARY

1. The hydrogen-ion concentration of the blood during certain emo-
tional disturbances, such as fright in rabbits and dogs, and- anger in
cats, is markedly increased and at a temperature of 20°C., frequently
reaches an acidity corresponding to pH = 7.00. This is probably due
to increased carbon dioxide tension.

2. In anaesthesia caused by ether, chloroform and nitrous oxid.
the hydrogen-ion concentration may be increased to the same ex-

232 M. L. MENTEN AND G. W. CRILE

tent. The change in acidity begins when the inspired anaesthetic com-
mences to react with the blood, and depends, approximately, on the
degree of the anaesthesia. The restoration of the normal reaction of
the blood is completed in 45 minutes after the administration of the
drug is discontinued.

3. In two cases of shock the acidity of the blood was very much
increased.

4. The blood flowing from the adrenal gland is always more alkaline
than venous blood elsewhere in the body. This increased alkalinity is
local, not extending to any appreciable extent beyond the immediate
vicinity of the adrenal vein, and is due to the dissolved adrenalin which
it contains.

REFERENCES

(1) H6éseEr, r: Pfliiger’s Arch., 1900, 81, 522 and 1903, 99, 572.

(2) Micuartis, L.: Die Wasserstoffionen Konzentration, Berlin, 1914.

(3) HasseL.patcH: Biochem. Zeitschr., 1911, 30, 317 and HassELBALCH AND
LUNDSGAARD: Biochem. Zeitschr., 1912, 38, 77.

(4) Micuartis, L.: Loe. cit.

(5) Benepixt: Pfliiger’s Arch., 1906, 115, 106.

(6) MicHartis, L.: Loe. cit.

(7) MicHaE.is, L.: Loe. cit.

(8) ABDERHALDEN: Handbuch der Arbeitsmethoden, v, 11.

(9) HASSELBALCH AND LUNDSGAARD: Skand. Arch. f. Physiol., 1912, 27, 13.

(10) Aupricu: Journ. Amer. Chem. Soce., 1905, 27, 1074. Complete bibliog-
raphy is given in Barger’s Monograph ‘‘The simpler natural bases.’’

(11) Henprrson, Y.: Amer. Journ. Physiol., 1908, 21, 126.

(12) BuckMASTER AND GARDNER: Journ. Physiol., 1910, 41, 246.

(13) Rosertson, G. B.: Journ. of Biol. Chem., 1908, v, 155; 1909, vi, 313;
1910, vii, 351.

THE ORIGIN OF ANTITHROMBIN

GEORGE P. DENNY, M.D., Boston, GEORGE R. MINOT, M.D., Boston
From the Physiological Laboratory of the Johns Hopkins Medical School

Received for publication June 17, 1915

INTRODUCTION

It is a well-known fact that Witte’s peptone when injected intra-
venously into dogs produces in their blood an anti-coagulating sub-
stance which causes blood withdrawn from such animals to remain
fluid for long periods. The plasma from peptonized animals has the
power to delay or inhibit the coagulation of normal blood in vitro.

Many workers have studied peptonized animals with a view of find-
ing the origin of the anti-coagulating substance, among them being
Delezenne (1), Nolf (2). Doyon (3), and Popielski (4). All agree that
the liver is the most important site; in fact, with the exception of Popiel-
ski they feel it to be the only site. The last named author thinks the
intestine and extremities play a large part.

Almost all experiments on the origin of antithrombin have been
carried out by means of peptone injections and it has been found that
in animals with the liver cut out of the circulation peptone is without
effect on coagulation. The liver then seems essential to the production
of the anti-coagulating substance which we shall call antithrombin
according to Howell’s theory (5).

Just in what way peptone acts on the liver is a disputed point but all
workers agree that Witte’s peptone itself has no action. It is some
substance formed in the blood which stimulates the liver to produce
antithrombin. Delezenne believed the destruction of leucocytes to
play the whole part. Nolf’s work does not substantiate this. Both
Nolf and Popielski believe the liver endothelium produces antithrombin
but this point is not as yet adequately proven. No attempt will be
made in this paper to discuss the various theories of the method of pep-
tone action but the above brief summary serves to indicate that the
liver is the organ most concerned.

233

234 GEORGE P. DENNY AND GEORGE R. MINOT

Up to the present time almost all determinations of antithrombin
have been made by determining the effect of the peptone plasma on
whole blood. This of course is a very rough method and serves only
to show a great increase in antithrombin.

According to Howell’s theory of coagulation antithrombin is a normal
constituent of the blood. By its antagonizing action to prothrombin
it prevents the latter from being converted to thrombin and tends to
neutralize the action of any thrombin that may be formed. In other
words it is the normal excess of antithrombin which prevents intra-
vascular clotting. If antithrombin is normally present in the blood,
and if, as previous workers have shown, it is formed in the liver, it
would seem likely that in a normal animal blood taken from the
hepatic vein would show a greater amount of antithrombin than blood
from other parts of the body.

Doyon (6) has reviewed the numerous articles on the coagulation
time and fibrinogen content of blood so taken and finds the results so
greatly at variance that no conclusions are possible. He concluded
from work of his own that such blood may coagulate more quickly than
carotid blood. Nolf found that whole blood introduced into a living
washed liver clotted normally. |

We have taken up the study of blood after its normal passage through
various organs and after stasis in the same organs.

TAPPING OFF OF ORGANS

In order to determine how the antithrombin content of the blood may
vary in different parts of the body a series of six experiments were done,
the results of which are tabulated below.

Method. A dog was etherized, after previous administration of
morphia, and a tracheal cannula inserted for continued etherization and
artificial respiration. Dissecting down on the jugular vein a specimen
of blood was drawn from it as a control, ten to twenty minutes before
a head thorax circulation was established. In all but Experiment IT
ligatures were then placed in the following order: around the inferior
vena cava in the thorax close to the diaphragm, around the inferior
cava directly below the liver, and around the aorta directly below the
diaphragm. In Experiment II the order of the ligatures at the liver
was reversed, first below and second above.

It took about ten minutes to place these ligatures after the vessels
were exposed. The ligatures were allowed to remain until after the
experiment was completed.

ORIGIN OF ANTITHROMBIN B35

Specimens of blood were drawn as rapidly as possible after the stasis
was produced, usually in the following order: first, from the portal
vein, second, from the inferior vena cava below the lowest ligature, and
third, from the hepatic vein. In Experiments III and V the order was,
the inferior cava, the hepatic vein and the portal vein. It was about
ten minutes from the time the last ligature was tied until the last of the
above samples of blood was obtained. The antithrombin determina-
tions of these specimens are recorded in the table as “before stasis.”’
To be sure, a few minutes of stasis of the blood had occurred before
these samples were obtained, but they represent blood before prolonged
stasis had taken place. After the blood below the diaphragm had re-
mained stagnant and a head thorax circulation had existed for varying
lengths of time, 30-70 minutes, as recorded in each instance in the table,
other specimens of blood were drawn from the same regions for com-
parison.

Samples of blood were drawn from the splenic and renal veins in three
experiments before the ligatures creating a head thorax circulation were
tied; then the arteries and veins of these organs were clamped and after
an interval of stasis blood was again drawn to be compared with other
specimens.

The specimens of blood to be examined were drawn with a Luer
syringe previously rinsed with salt solution. A-definite volume of this
blood was put into a tube with a constant amount of 1 per cent sodium
oxalate in 0.9 per cent sodium chloride solution and well mixed, centri-
fuged for 20 minutes, and the plasma pipetted off, heated to 60°C. to
precipitate the fibrinogen and destroy prothrombin, filtered and exam-
ined for antithrombin by the method described by Howell (7) and dis-
cussed by Minot and Denny (8).

In brief the test used was as follows: 1 drop of the antithrombin
solution (after heating to 60°C. and filtering) from each specimen was
mixed in a tube with the varying amounts of a solution of thrombin.
After these two substances had been in contact for a given interval, a
constant amount of a fibrinogen solution was added to each tube so that
the clots formed in a period of time suitable to distinguish differences.
The time taken for clotting was upon the first appearance of a clot, that
which clotted first having the least antithrombin and that last the most.
At least four suitable series of determinations on the different speci-
mens were made. To compare the specimens of any single experiment,
the time that their antithrombin allowed thrombin and fibrinogen to
clot was sufficient.

236 GEORGE P. DENNY AND GEORGE R. MINOT

But in order to compare those of different experiments time could
not be satisfactorily used, owing to differences in the reagents, so that
antithrombin is expressed as the antithrombin factor (8). The average
time of clotting of a series of suitable determinations of a specimen was
divided by a similar figure obtained for a control. The control in these
experiments was the specimen from the jugular vein before any stasis
was produced and its factor was taken as unity.

By referring to the table it will be noticed that in the six different
experiments the antithrombin in the venous blood of the jugular or
superior vena cava before and after a head thorax circulation remained
about the same.

The two observations of splenic blood before stasis showed less anti-
thrombin than the jugular blood. The three specimens of splenic vein
blood after stasis all gave a plasma in which some definite hemolysis
had taken place. It has been shown that large amounts of fatty acids
occur in the spleen and this may be the reason that the blood standing
in the spleen is hemolyzed. Except for one specimen of renal plasma,
none of the other specimens showed any hemolysis. It was felt that the
hemolysis in this renal specimen might have been due to the fact that
the syringe contained water and was not washed out with salt solution
before the blood was drawn.

Antithrombin, being ‘‘neutralized’”’ by thromboplastin, cannot be.
determined satisfactorily in a plasma in which hemolysis is apparent, so
that the figues for antithrombin after stasis of blood in the spleen are
probably lower than they should have been owing to the presence of
thromboplastin liberated from the hemolyzed corpuscles. The two
observations of antithrombin in renal vein blood before stasis were sim-
ilar to the control with a slight tendency to rise after stasis.

Portal blood before prolonged stasis was studied in five experiments;
in four it contained less antithrombin than the control, in one a greater
amount. In the three observations after an interval of stasis from
32-60 minutes there was a slight rise.

Inferior vena cava blood (taken from below the lowest ligature) in the
four experiments in which it was studied showed less antithrombin than
the control, with a definite rise after an interval of 32 to 48 minutes of
stasis in three instances, and essentially no change in the fourth.

Hepatic vein blood before prolonged stasis showed an antithrombin
content essentially like the control blood. Just before the specimen
to be tested after stasis of 30 to 70 minutes was obtained, varying
amounts of blood were drawn off from the vein and discarded so as to

ORIGIN OF ANTITHROMBIN 237

obtain blood that had been in actual contact with the liver cells. Owing
to the short length of the hepatic vein it was easy to draw blood from
the cava that had stood between the clamps above and below the
liver instead of obtaining hepatic vein blood. To obviate this diffi-
culty in the last three experiments just before the hepatic blood after
prolonged stasis was drawn, a clamp was placed across the cava above
the liver, just below the point where the hepatic vein enters it.

After prolonged stasis of blood in the liver there was in the last four
experiments a very marked rise in the antithrombin, a much greater
amount than in blood from other sources. In the first experiment
there was a slight rise and in the second a fall. Why there was not a
greater rise in Experiment 1 and a fall in Experiment 2 is not clear, but
it is possible that the blood obtained had not been stagnating deep in
the liver tissue.

Not only does the antithrombin show a distinct increase after stasis
of blood in the liver, but the clot formed by the thrombin and fibrinogen
in the presence of the antithrombin from this source was of a weak
character such as occurs when one has increased antithrombin.

The portal blood and the blood from the inferior cava and spleen,
obtained before stasis, had less antithrombin in most instances than
that from the jugular. The renal blood on the contrary had about an
equal amount, as did the hepatic.. Blood taken after stasis in any or-
gan showed in most instances a slight rise in antithrombin. This was
particularly evident in the inferior cava below the lowest ligature, but
in no case was found the great rise observed after stasis in the liver.

In Experiment VI the time of clotting of the recalcified oxalated
plasma was determined. Precautions were taken to find the optimum
amount of calcium for this clotting and the time obtained designated
as the prothrombin time. It constitutes a test of the relative efficacy
of the prothrombin present. To a series of tubes each with 5 drops of
plasma from each specimen 2, 3, 4 and 5 drops of 0.5 per cent CaCl, solu-
tion were added and the time of clotting noted. The tubes with 3
drops of calcium were the first to clot in all instances, i.e., had the
- optimum amount of calcium. The shortened time occurring with
plasma from the splenic blood after stasis in this organ, is undoubtedly
due to the hemolysis and the consequent liberation of thromboplastin
as we have frequently noticed on other occasions.

Prothrombin times are as follows:

Jugular vein before stasis 5 minutes.

Jugular vein after stasis 5 minutes.

238 GEORGE P. DENNY AND GEORGE R. MINOT

Splenic vein after stasis 3 minutes.

Renal vein before stasis 5 minutes.

Renal vein after stasis 5 minutes.

Portal vein before stasis 5 minutes.

Portal vein after stasis 5 minutes.

Hepatic vein before stasis 5 minutes.

Hepatic vein after stasis 35 minutes.

A striking delay in the prothrombin time occurred with the hepatic
blood after stasis. This delay is probably wholly due to the increased
antithrombin content, although possibly there was also a decreased
amount of prothrombin.

Examination of arterial and venous blood taken from the heart has
not shown any difference in either antithrombin or prothrombin content.

PERFUSION OF LIVER

Perfusion of the liver with mixtures of blood and peptone, plasma
and peptone, etc., has been shown by Delezenne (1), Nolf (2) and
others to produce antithrombin. Both of these authors have perfused
the liver with defibrinated blood alone and Delezenne came to the con-
clusion that such blood acquired a very slight anticoagulating power.
Nolf also obtained negative or doubtful results.

Doyon (9), by attaching the carotid artery of one dog to the portal
vein of another which had been bled to death found that on emer-
gence from the inferior cava the blood was slow to coagulate and re-
tarded the clotting of normal blood. Positive results were obtained
only with young, fasting dogs. With the exception of the above work
we have been unable to find in the literature any successful demon-
stration of a formation of antithrombin by perfusion of the liver with a
normal circulating medium without drugs added to the perfusing ma-
terial.

We have perfused the liver with defibrinated blood in the following
manner:

Method. A large dog was bled to death while under ether and the
blood defibrinated. Part of this was diluted 1 to 2 with normal salt
solution and used for irrigation and the rest was used to perfuse the
liver of a second dog.

Dog II was given morphia and ether, the abdomen opened and the
portal vein exposed. A large cannula was placed in the portal vein
and the inferior cava quickly clamped above its tributaries, close to the
liver. The hepatic artery was also clamped at this time.

08 OL Og'e % JA Jusuredxy
xe /99 | 00°E | 00°T 200 ROGes ee OnOh |e Ohen | cm ens Roe /0& | 08°0 | 08°0 | ,02 GT} OT : ;
0€ WP | GOS
03 10¥. || OL T A juowtedxy
él OF | 81% 188 | 901 | 89°0 | ,S& | 00T | 89°0 | 0G OT] OT
08 SP | 20% AI Juourrsed xT
06 8h | S8T | 20 T | OF | OFT | 82°0 06 | GT OT} $6 | 08°0 | S8°0 | ,S& (Oe |} 1! ;
86 L& | O'T TI] Jueursodxg
st S$ | OST} LOT} .& | €2T | 22°0 660 08 OT|OT >
ST 10S | 80] 00T| OF | TZ0| 20} ,c& | 42°0}) 990 109), 90 10% II Jucewpodxiy a
| ee Ss ee ee ee a ee ee eS EF ee ae
oy | 380 J Juoeurnsod xo
0Z =| ,0€ jeer |20T eT & L0°T if $90 8 £6°0 | OT Y
99
RinGalce lee. |e ee lreeslsen) oe le onl os le ye) Spl ae
aca] ¢ |g2/82| € |sh/52| 2 | s2/S2| 2 | 82/22] 9 | s2| SEARS sels oe
on be e teas oO ek oe oct & Sats oo rl ast oe 2 ee ot SPS sea sos
toap = oe ae = +5 ae Ee ae ae 5 e+e | a oe 4E 1] a8 |atPrela Bilisienics
$2) co ° a ° an ° is) is ie} po + ° © ct lo)
e,g-| » |BS/EG| 2 | BSIES| & | BS ES] ® | 28/88) = 128) 8s SE") S8ises
Boas 2. Sle@ol Q1*e¢lao] Si *erlas| & Se inet | ews |e ooo) ace Zo Slee nee
4 (i es eae 2 a | § Be feraea (ss Ele e erate Geile Eo oat) Se sellin oe ele eS
ay is) ®
sce e | Be} F| & | Bl F)e | Bl F) | Bl Fle | BF) Pe) eel FF
AUOALVOIT L84 VAVO VNGA
NIGA OliVdaH -MOT MOTAG VAVO NIGA IvLuod NIGA ‘IVNAU NIDA OINAIdS wolMmados XO UV

VNGA HOIMAANI -Op0f 'IVWNUGLXa

240 GEORGE P. DENNY AND GEORGE R. MINOT

While one operator was doing this the other had opened the chest
and introduced a cannula into the inferior cava just above the liver and
washing with diluted blood was immediately started through the portal
vein. As soon as the washing fluid had been run through perfusion
was started with defibrinated blood. The time during which the liver
had no circulation varied from 4 to 7 minutes.

The apparatus used for perfusing was in most cases very simple. It
consisted of a large tank filled with water. In this was placed the
reservoir for the perfusing and washing fluids from which tubes lead to
the inflow cannula in the portal vein. The flow was by gravity. A
thermometer was introduced at a point just before the entrance into
the liver and the perfusing fluid was kept at about 38°C. by regulation
of the temperature in the large water bath.

The outflow was caught in a container kept at about 38° by means of
a water bath. The outflow was whipped, when defibrinated blood
was used as a perfusate, since it was found that the first specimens com-
ing through after washing showed some tendency to clot. Defibrina-
tion was continued only during the first part of the experiment and was
necessary probably because the liver had not been washed entirely free
from blood. In some experiments oxygen was bubbled into the outflow
container and in one an artificial lung devised by Dr. Hooker was used.
Oxygenation seemed to make little or no difference since those experi-
ments in which no attempt was made to oxygenate gave similar results.
The outflow was poured back by hand into the reservoir above and
thus allowed to circulate many times.

Perfusion was carried on under a pressure of 90 to 120 mm. of mer-
cury. Although there was a certain amount of loss in the amount of
the perfusing liquid the latter was not concentrated enough to account
for any significant increase in antithrombin content. The method,
however, has the objection that the system was not closed.

Specimens were taken of the perfusate before perfusion was started
and at intervals during perfusion. These were taken in 5 ce. lots and
mixed with 0.7 ce. of 1 per cent oxalate in normal saline. After centri-
fuging, antithrombin tests were done on the serum by the method of
Howell previously described.

Four experiments were done.

ExperimentI. Perfusion with Ringer’s solution.
Dog—Morphia and ether. Cannulas placed as described. Liver washed with
3 liters of Ringer’s solution until outflow was clear.

ORIGIN OF ANTITHROMBIN 241

Perfusion then started with about 200 cc. of Ringer’s solution. This was
clear at the start but quickly became bloody. Whole system cleaned out and
started again with fresh Ringer’s. Perfused in first case for about one hour and
in the second for one and a half hours. The latter became bloody and a third
perfusion was done with fresh Ringer’s after washing. After one hour and forty-
five minutes perfusion was again bloody. The specimens taken during the last
perfusion showed a heavy precipitate on heating to 60° which was not fibrinogen
and was evidently protein from the dying liver. None of the seven specimens
taken during the perfusion showed any more antithrombin than might be ex-
plained by the blood which had been picked up during the repeated passage
through the liver.

Experiment II. Perfusion with defibrinated blood.

Preparation as described.

Liver washed out with 100 cc. defibrinated blood plus 200 ce. normal saline.
Perfusion started with 570 cc. defibrinated blood after a control specimen had
been taken. This is labeled I in the table. Specimens taken 30, 40, 50, 58 and
64 minutes after the start of perfusion are labeled II, III, IV, V, and VI.

All specimens centrifuged at high speed for 20 minutes, serum pipetted off,
heated to 60° and filtered. On heating none showed any precipitate.

One drop of the heated and filtered serum (antithrombin) was added to the
thrombin in varying amounts and after an interval of 15 minutes 10 drops of
fibrinogen were added.

Antithrombin determination

THROMBIN I Ir Ill Iv v ae SPECIMEN NUMBER

4 gtt. 18’ 30’ Bik 41’ 46’ 48’ | Figures represent time in
minutes which it took for
5 gett. 12’ PPK 22’ 2! 38’ 40’ clot to form after addition
of fibrinogen.

| The two series correspond well and show a very marked gradual in-
crease in antithrombic power.

Experiment III. Perfusion with defibrinated blood.

Large dog bled to death and cannulas inserted into the portal vein and inferior
eava of another dog in the usual way. Liver washed with 100 ce. defibrinated
blood plus 200 ee. normal salt solution.

Perfusion started after control specimen had been taken and 20 minutes later
Specimen II was obtained.

Just after this was taken a rather large amount of water was spilled accident-
ally into the outflow jar. This of course produced great hemolysis as well as
dilution in the circulating blood and the effect in lowering antithrombin will be
seen in the table.

Specimens taken 65, 95 and 108 minutes after start of perfusion are labeled
III, IV and V in the table.

On centrifuging, those specimens taken after the accident were much more
hemolyzed than those before it. Antithrombin tests were run in the usual way.

242 GEORGE P. DENNY AND GEORGE R. MINOT

Antithrombin determination

THROMBIN I Il III IV nf . SPECIMEN NUMBER
5 gtt. PAN ie i igs 14’ 16’ | Figures represent time in minutes
which it took a clot to form after

6 gtt. 10’ iY ue 10’ 12% addition of fibrinogen.

No. III represents the first specimen taken after spilling water into
the blood. Nos. IV and V show a gradual increase in antithrombin
content although the antithrombin never reached the concentration
that it had at the beginning.

Experiment IV. Perfusion with defibrinated blood.

This experiment was carried on in the usual way except that the outflow from
the liver was connected with an artificial lung devised by Dr. Hooker. This en-
sured thorough oxygenation of the blood. The liver was washed with mixture
of 300 cc. normal salt solution and 150 ec. defibrinated blood. Perfusion started
after taking control, Specimen I. Specimens IJ, III, IV and V taken 30, 59, 70
and 83 minutes respectively after the start of perfusion. After Specimen V was
taken 1.5 gm. peptone in solution was added to the perfusing material (about 400
ce.). Specimens taken 15, 24 and 34 minutes after addition of peptone are labeled
VI, VII, and VIII in the table.

On centrifuging, all specimens showed hemolysis as usual, the degree increas-
ing somewhat as the experiment went on.

Antithrombin determination

Q Z
ZA 2p
of a)
THROMBIN I II Ill IV Vv & a VI VII Vill S 2
aS Bz
5 ett. 10’ 13’ 13’ 19’ | No No clot in
clot 30’
in 30’
6 gtt. 8’ 8’ 8’ 9’ Very Poor clot in
poor 21’ 26’ 5
clot
15/

From this table it will be seen that antithrombin was not acquired by
the perfusing fluid for about 1 hour, Specimen IV being the first to
show any change. Specimen V shows a well marked increase. The
specimens taken after addition of peptone show a still further increase
in antithrombin content but not to the extent which one would expect

ORIGIN OF ANTITHROMBIN 243

in.a fresh liver perfused with a blood and peptone mixture. Probably
much of the antithrombin had been taken.up from the liver in this case
before the peptone was added to the perfusate.

These experiments show that the frequent passage of defibrinated
blood through the liver results in a gradual increase is the antithrombin
content of the perfusate.

Specimens of defibrinated blood were obtained from perfusions of
the spleen, head and hind leg done by other workers in this laboratory,
and although in some of these cases the perfusions were carried on much
longer than in our experiments with the liver, we were unable to demon-
strate any marked increase in antithrombin; in fact, usually the
amount was decreased. Three perfusions of the head showed in each
instance a very slight increase of antithrombin.

_ The decrease in antithrombin was probably due to hemolysis and
deserves a word in connection with our experiments.

The centrifuged serum from defibrinated blood always shows a certain
amount of hemolysis. In perfusing for any length of time the amount
of .this hemolysis increases distinctly up to a certain point. Since
hemolysis liberates an active thromboplastic substance some of the
antithrombin in the perfused blood must have been neutralized and
therefore the increase in antithrombin which we actually found after
perfusion of the liver probably did not represent the total increase
obtained. If we could have done away with this factor of hemolysis our
results would probably have been more striking.

ATTEMPTS AT STIMULATION OF LIVER

| Having shown that stasis in the liver causes an increase in anti-
thrombin and that perfusion with defibrinated blood likewise causes
an increase we next attempted to stimulate the liver in various ways.
Doyon (10) has attempted this also and claims success as a result of
intravenous injection of bile and bile salts. He found that bile salts
injected intravenously could cause an incoagulable blood, and that doses
which were ineffective in the general circulation were capable of delay-
ing coagulation when injected into the mesenteric veins. He did not
show that the long coagulation time was due to an increased anti-
thrombin.
The substances which we have used were all injected into the mesen-
teric veins, and all the dogs used were starved 24 hours or longer.
These experiments will not be given in detail because they were en-
tirely negative.

244 GEORGE P. DENNY AND GEORGE R. MINOT

Experiment 1. Effect of bile salt. Dog, weight 7 kg.

9.5 cc. of 10 per cent solution of sodium glycocholate was injected into the
mesenteric vein. In 10 minutes the coagulation time had gone from 40 minutes
(normal) to one hour and twenty minutes and remained in this vicinity for three
hours.

All specimens taken after injection showed a very marked hemolysis and the
antithrombin was uniformly reduced as compared with the control specimen.
Prothrombin not changed to any marked degree. We felt that larger doses might
cause greater delay in coagulation time but that the increasing hemolysis would
make it impassible to say if it were due to an increased antithrombin.

Experiment 2. Effect of bile by mouth.

A dog, weight 8 kg., having had control specimen of blood taken, was given 16
gm. of dried ox bile in 200 cc. water by stomach tube. Allowed to come out of
ether. Vomited a little (40-50 ec.).

Two hours and fifteen minutes after giving bile another specimen of blood was
taken. 17 gm. of ox bile in 150 ce. water then given by stomach tube. Specimens
taken one and two hours later showed no hemolysis, no change in coagulation
time and no change in prothrombin and antithrombin content.

Experiment 3. Effect of secretin.

Dog, weight 7 kg. Morphia and ether.

10 ce. of a secretin solution prepared according to Bayliss and Starling (11)
injected into the mesenteric vein. Specimens taken 10 minutes after, showed
coagulation time of 56 minutes against the control time of 28 minutes—20 min-
utes after injection coagulation time was 29 minutes. 9 cc. more injected and
specimens taken 16 and 24 minutes later, showed a coagulation time of 41 and 55
minutes respectively.

Large amounts of secretin introduced into the mesenteric veins with a bu-
rette failed to further lengthen the coagulation time.

No pancreatic fistula having been made, it is not certain that the secretin
used was active. There was a slight lengthening of the coagulation time after
the first two injections but on the whole the experiment must be considered nega-
tive. There was no change in the antithrombin content of the numerous speci-
mens taken after the various injections.

Experiment 4. Effect of nerve stimulation.

Dog, weight 8 kg. Morphia and ether. Control specimens taken for coagu-
lation time and antithrombin. Abdomen opened and coeliac plexus exposed.
Electrode passed under the ganglion and stimulation applied with a rather strong
current from an induction coil. Duration of stimulation 20 seconds with an
interval between of 40 seconds. The procedure was continued for one hour,
specimens being taken at 15, 35 and 50 minutes from the start of stimulation.

These specimens showed no change in coagulation time or antithrombin con-
tent. The hepatic branches of the plexus were next isolated and stimulated with
the result that no change was obtained either in coagulation time or antithrombin.

Experiment 5. Effect of injection of thrombin.

Dog, weight 8 kg. Morphia and ether.

Injection of 0.4 gm. of pure thrombin in which there was some salt (this solu-
tion was very active in clotting fibrinogen) into the mesenteric vein. In a spec-
imen taken 12 minutes after injection there was lengthening of coagulation time

ORIGIN OF ANTITHROMBIN 245

from 30 minutes (normal) to 43 minutes and an increase in antithrombin from 10
minutes to 17 minutes.

Davis (12) found an increase in antithrombin by injecting thrombin
in very large amounts into the general circulation. Here we see a
small amount into the mesenteric vein giving a slight but definite effect.

PHOSPHORUS POISONING
Antithrombin as Influenced by Phosphorus Poisoning in Dogs

If the liver is the site of antithrombin formation it would seem pro-
able that a liver destroying substance such as phosphorus would lower
the antithrombin content of the blood. It has been shown by many
Jacoby (1900); Loeb (1903); Morawitz (1905); Doyon (1906); Nolf
(1908) ; and Whipple and Hurwitz (13) that the fibrinogen of the blood
in phosphorus poisoning is greatly diminished or almost absent.

We have studied the blood of five dogs with varying degrees of phos-
phorus poisoning, all of which showed extensive liver destruction at
autopsy. The specimens of blood were kindly given us by Drs. Good-
pasture and Marshall, to whom we wish to express our thanks.

Antithrombin determination

PHOSPHORUS NORMAL ANTITHROMBIN
BOG DOG FACTORS
Det S00 ee 9’ 14’ 0.64
2 I Oa 8’ Ae 0.53
“7 0 (DUS 1 25/ 0.44
1) Oe a ae 16’ 0.12
0.15

NSS le er 2! 13’

Figures represent time of formation of a clot in the usual antithrom-
bin test. It will be remembered that the normal antithrombin factor is
approximately 1.

Dog IV was the most severely poisoned and shows the greatest dim-
nution in antithrombin; in fact, almost a complete absence of any anti-
coagulating power. All specimens showed by rough tests a diminution
in fibrinogen.

These results tempt speculation as to what part of the liver is con-
cerned in the formation of antithrombin. It will be remembered that
Nolf and Popielski are inclined to think that the endothelium in the
liver is responsible.

246 GEORGE P. DENNY AND GEORGE R. MINOT

While it cannot be stated that phosphorus does not injure the endo-
thelium of the liver there can be no doubt of its destructive action upon
the liver cells and we suggest from our results that antithrombin may

very likely be a true liver cell product. However, further work on this’

subject, with careful histological study, must be done in order definitely
to demonstrate this point.

Numerous clinical cases of liver disease which we have studied have
shown no constant variation in antithrombin content of the blood.
A few with low fibrinogen have shown a low antithrombin. Some
cases having a normal antithrombin showed very extensive destruc-
tion of liver substance at autopsy but it is a fairly well recognized fact
that only small amounts of active liver tissue are necessary to carry on
the functions of the liver.

CONCLUSIONS

1. Antithrombin is formed in the liver.

2. Venous blood taken from the liver, spleen, kidneys and intestines
shows no appreciable difference from jugular vein blood in antithrombin
content.

3. When stasis is produced in these organs the liver blood alone shows
a definite increase in antithrombin.

4. Perfusion of the liver with defibrinated blood causes an increase
of antithrombin in the perfusate.

5. Perfusion of the head, spleen and hind leg showsno marked increase
and generally a decrease in antithrombin content.

6. Attempts to stimulate the liver to production of antithrombin by
means of bile, bile salts, secretin and electrical stimulation were nega-
tive. Injection of small amounts of thrombin into the portal circula-
tion caused a slight rise in antithrombin.

7. Dogs with phosphorus poisoning and liver destruction show a very
marked decrease in antithrombin content of the blood as well as dimin-
ished fibrinogen.

We wish to express our gratitude to Professor Howell for invaluable
advice and suggestions and to thank Dr. D. R. Hooker for material
from his experiments and for help with our perfusion work.

(1) DELEzENNE: Compt. Rend. d. Soc. Biol., 1898, v, 354.
(2) Nour: Arch. Internat. d. Physiol., 1910, ix, 407.
(3) Doyon and GAuvTIER: Compt. Rend. d. Soc. Biol., 1910, Ixviii, 450.

ORIGIN OF ANTITHROMBIN 247

(4) Popretskxi: Zeitschr. f. Immunitatsfurschung, 1913-0, 1 Teil, xviii, 542.
(5) Howe: Text Book of Physiology, 5th ed., 1914.

(6) Doron: Journ. Physiol. et Path. gen., 1906, viii, 1003.

(7) Howe tt: Arch. Int. Med., 1914, xiii, 76.

(8) Minor anp Denny: Unpublished work.

(9) Doron: Comp. Rend. d. Soc. Biol., 1910, Ixviii, 752.

(10) Doron: Jour. Physiol. et Path. gen., 1910, xii, 197.

(11) Bayess anp Staruine: Journ. Physiol., 1902, xxviii, 425.

(12) Davis: Am. Journ. Physiol., 1911, xix, 160.

(13) WurtpreLe anp Hurwitz: Journ. Exp. Med., 1911, xiii, 136.

CONTRIBUTIONS TO THE PHYSIOLOGY OF THE STOMACH

XXV. A Nore on THE CHEMISTRY OF NorMAL HuMAN GASTRIC
JUICE

A. J. CARLSON
ASSISTED IN PART OF THE EXPERIMENTS BY H. HAGEeR anp M. P. Roggrs

From the Hull Physiological Laboratory of the University of Chicago
Received for publication June 29, 1915

The gastric juice employed in this work was obtained from Mr. V.,
our gastric fistula case with complete cicatricial stenosis of the oesoph-
agus of 20 years’ standing. The gastric juice is therefore at no time
mixed with saliva. The appetite juice is thus secreted during the
mastication of the food and some data relative to the secretion of
appetite and hunger gastric juice in Mr. V., as a type of normal adults,
have already been reported (1). A few samples of appetite gastric
juice was secured from a second gastric fistula case with partial cica-
tricial stenosis of the oesophagus. Samples of appetite and of hunger
gastric juice were also secured from one of the authors (A. J. C.) by a
method that will be referred to later. A few controls are also noted
on dog’s gastric juice, that is, the secretion from the accessory stomach
pouch prepared according to Pawlow. Most of the present work refers
to appetite gastric juice. A few determinations were made on samples
of the hunger gastric juice. The reader is referred to our previous
report for an account of the continuous or hunger gastric secretion.

I. The solids of the gastric juice. 100 ce. of six different lots of appetite
gastric juice of Mr. V. were evaporated on water bath and then dried
to constant weight in an oven at 112°C. The incineration was made
at dull red heat so as not to drive off the potassium chloride, which ~
according to Rosemann is more abundant in gastric Juice than the
sodium chloride (2). The results are given in Table I. On three of
these lots of appetite juice the specific gravity, the freezing point, and
the total acidity (phenolthalein) was also determined. The one deter-
mination on hunger gastric juice represented many separate collec-
tions of juice, as no juice was included in this lot if it was secreted by

248

1

The

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 249

the empty stomach at a greater rate than 3-4 cc. per hour. For the
purposes of comparison the analyses of two lots of the juice or fluid
found in the stomach free from food one hour after putting 200 ce. of
water into the stomach are also given in Table I. It will be seen from
Table I that the total solids of the appetite juice vary from 0.48. gr.
to 0. 58 gr. per 100 cc., of which 0.34 gr. to 0.47 gr. is organic, and 0.11
er. to 0.14 gr. inorganic material. The hydrochloric acid is, of course,
expelled in the process of evaporation and drying of the gastric juice
residue.

The hunger gastric juice of Mr. V. is distinctly higher than the appe-
tite juice in total, and in organic solids. It appears also to be higher
than the average appetite juice in inorganic solids.

The gastric juice or fluid in the stomach free from food one hour
after ingestion of 200 cc. of water is distinctly more dilute than the

TABLE I

composition of pure human gastric juice. The solids of appetite gastric juice, hunger

juice, and content of ‘“‘empty’’ stomach per 100 cc. of juice.

HUNGER |

: ; CONTENTS OF
APPETITE GASTRIC JUICE cases Prva pele
I II Til IV V VI I | II
yecific gravity .. 1007 1009 1006 1008 1004 1006
A —0.550 —0.580 —0.530 —0.31 —0.45

ity (total)...| 0.45 0.48 0.42 0.34 0.22 0.33
ptal Bolids....... 0.557 er. 0.606 gr. 0.480 gr. | 0.571 gr. |0.580 0.544 0.667 gr. 0.3€0 gr. 0.450 gr.
‘ganic » Ee 0.420 gr. 0.466 gr. 0.341 0.452 gr. |0.470 0.430 0.521 gr.| 0.260 gr. 0.350 gr.
organic Besa aes 0.137 gr. 0.140 gr. 0.139 0.119 0.110 gr.} 0.114 gr. | 0.146 gr. 0.100 gr. 0.100 gr.

appetite juice, although it may approach the concentration of the
latter in cases where the rate of the continuous secretion is considerable
(Lot IT).

The above figures on Mr. V.’s appetite gastric juice are slightly
higher than those given by Sommerfeld (3) for tHe gastric juice from a
ten-year-old girl, namely, 0.40 gr. to 0.47 gr. Sixty years ago Schmidt
(4) reported on the gastric juice of a human gastric fistula case, the
oesophagus being partly patent. He found total solids 0.58 per cent,
of which 0.32 per cent was organic, and 0.26 per cent inorganic. But
Schmidt did not work with pure gastric juice This is evident from
his method of obtaining the juice, as well as from the fact that the
acidity of the juice was only 0.20 per cent, which is less half that of
normal human gastric juice. The figure of 0.26 per cent for inorganic
solids is probably also too high Albu (5) reports one experiment 02

250 A. J. CARLSON

a patient with hypersecretion finding the percentage of solids only
0.24 gr. practically all of which (0.23 gr.) was inorganic salts. Albu
also reports one determination on normal human gastric juice (pure
appetite juice) in which the inorganic solids were 0.18 per cent; the
organic solids are not given.

Our results on Mr. V. agree closely with most of those reported for
the gastric juice of dogs, as the following figures will show:

Total
Observer Solids Organic Inorganic
per cent per cent per cent
Pawlow and Schoumow-Simanowksi (6)............. 0.47
dS Gop aKa ya) (0) Bie (VA) RS prontens ono oro athens tom sin as Oto lac 0.48
SChHOUMOW- SLM ANO WES See ees olen eee eee 0.528 0.393 0.135
FVOSETMAININ (2) aos scares cpa es ariae rots. 2 ee ree eae 0.427 0.294 0.1382
Nenckavandssieberk(G) passe een... a ee ee 0.306

The total concentration of organic and inorganic substances is there-
fore about the same in the normal gastric juice (appetite juice) of man
and dog. The very low solids (in one case only 0.16 per cent) reported
by Nencki and Sieber are clearly exceptional and must be due to some
unusual condition of their animal.

II. The specific gravity of the gastric juice. The specific gravity of
Mr. V.’s appetite gastric juice varies between 1006 and 1009 with an
average of 1007. This is the average of twenty tests on an equal
number of gastric juice samples. It will thus be seen that 1009 is an
exceptional concentration.

The specific gravity of the hunger juice is certainly higher than that
of the appetite juice, since the total solids are higher, but no actual
measurements were made on the hunger juice. The specific gravity
of the fluid or juice found in the empty stomach thirty to sixty min-
utes after ingestion of water varies from 1003 to 1006, with an average
of 1004.

The specific gravity of the appetite gastric juice of the dog, as re-
ported by Schoumow-Simanowksi, Konowaloff, Friedenthal (10), and
Rosemann varies from 1002 to 1007, with an average of 1004. This
low average figure is probably due to the fact that in most of the
experiments the gastric juice was collected for several hours after only
a few minutes sham feeding. There is evidence that the percentage
of solids in the gastric juice is greatest during the first hours of appe-
tite or digestion secretion. The concentration of the dog’s appetite
gastric juice during the first twenty minutes of secretion will in all
probability be found identical with that of man for the same period.

vat al

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 251

III. The osmotic concentration of the gastric juice. Our data on the
osmotic concentration of the gastrie juice are brought together in
Table II. The appetite juice lowers the freezing point, —0.55° C.
to —0.62°C.; the hunger juice from —0.47°C. to —0.52°C.; the
fluid of the empty stomach —0.21° C. to —0.41° C. The juice found
in the empty stomach an hour after ingestion of water exhibits the

TABLE II

The osmic concentration of normal hwman gastric juice (Mr. V.) The total
acidity of the samples of gastric juice were determined with phenothalein as the
indicator

FLUID IN ‘‘ EMPTY ’’ STOMACH HUNGER JUICE APPETITE JUICE
Total acidity A Total acidity A Total acidity A

per cent per cent per cent

0.33 —0.38 0.40 —0.51 0.53 —0.58

0.26 —0.33 0.39 —0.52 0.52 —(0).61

0.35 —0.42 0.36 0.45 —0.57

0.26 —0.30 0.35 —0.51 0.53 —0.60

0.30 —0.39 0.32 —0.50 0.52 — O68

0.27 —0.38 0.36 —0.53 0.47 —0.59

0.18 —0.25 0.29 —0.49 0.42 —().55

0.33 —0.38 0.28 —0.45 0.47 —0.60

0.17 —0.21 0.30 —0.47 0.49 — Oren

0.21 —0.27 0.41 —0.52 0.53 —0.60

0.17 —0.25 0.36 —0.49 0.45 —0.59

0.25 —0.32 0.41 —0.53 0.50 —0:61

0.26 —0.33 0.40 —0.52 0.49 —0.60

0.27 —0.37 0.32 —0.50 0.51 —0.57

0.18 —0.28 0.25 —0.47 0.46 —0.62

—0.33 0.51 —0.60

—0.37 0.51 —0.60

0.47 —(0:57

0.48 —0.55

0.49 —0..59

—0.62

—0.62

greatest fluctuations in osmotic pressure, the appetite and the hunger
Juice being very constant. The hunger juice has, on the whole, a
lower osmotic concentration than the appetite juice. The total acidity
of the hunger juice is also uniformly lower.

The above figures for the appetite gastric juice of Mr. V. are prac-
tically identical with those reported on the pure gastric juice of other

Bae A. J. CARLSON

human fistula cases. Sommerfeld (in a ten-year-old girl) found the
freezing point to vary from —0.47° to —0.65° C.; Kaznelson (11) (25-
year-old girl) reports a variation from —0.46° to —0.54°. Umber (12)
reports two tests on the gastric Juice (pure) of a 59-year-old man with
cancer, finding a variation of —0.15° C. to —0.82° C. Assuming that
Umber’s determinations are correct, the gastric juice of this cancer
patient was clearly not normal. We question whether the normal
stomach can secrete a juice with an osmotic concentration so much
greater than the blood as the figure —0.82° C. demands. The reader
will note that the figures of Sommerfeld and Kaznelson, as well as our
own for Mr. V., indicate an osmotic pressure of the appetite gastric
juice not far below or above that of the human blood. According to
Bickel (13) the gastric juice (10-year-old child) is always hypotonic to
the blood. Lehmann (15) concludes that the osmotic pressure of nor-
mal gastric juice (gastric content) is usually less than —0.50° C. and
that a concentration above this figure indicates hyperacidity or other
pathological conditions. This view is obviously untenable.

The osmotic concentration of the dog’s appetite gastric Juice appears
to be practically identical with that of man. Sasaki (14) reports a
variation from —0.51° to —0.60°; Rosemann gives somewhat higher
figures as —0.56° C. to —0.64°C. On the other hand, Bickel (13)
reports extraordinary fluctuations in osmotic concentrations of dogs
gastric juice (Pawlow pouch) or —0.52° to —1.21°C. We question
whether the normal stomach can secrete a juice of the osmotic con-
centration —1.21° C., that is, twice that of the blood.

IV. The nitrogen of the gastric juice. 1. The total nitrogen was deter-
mined by the method of Kjeldahl on nine different lots of appetite
gastric juice-of Mr. V., with the following results:

Nitrogen per 100 cc.
Appetite gastric juice gr.

Lot I 0.074
ot UE 0.051
Lot III 0.054
Lot IV 0.057
Lot V 0.054
Lot VI 0.051
Lot VII 0.075
Lot VIII 0.057
Lot IX 0.071

YAY 2c) Oa eR OT LOREM ETE. ec AS Aas Oe 0.060

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 253

The average of all our determinations is 0.060 gr. nitrogen per
100 ce. appetite juice. The total nitrogen of the hunger juice was not
determined. If all the nitrogen is in the form of proteins, and if we
accept the figures of Nencki and Sieber, and of Pekelharing (16)
namely, that the nitrogen constitutes 14.39 per cent of the proteins
of the gastric juice, the appetite gastric juice of man would contain
on the average nearly 0.42 gr. protein per 100 ce. This is practically
all the total organic solids in the appetite juice (Table I). It is there-
fore probable that some of the nitrogen in the gastric Juice is present
in non-protein combinations, such as ammonia, amino acids, and sul-
phoeyaniec acid.

The literature does not, to our knowledge, contain any data on the
total nitrogen of pure human gastric juice. Rosemann reports nitro-
gen determinations on two lots of dog’s appetite gastric juice, finding
0.035 gr. and 0.054 gr. per 100 cc. respectively. The reader will note
that these figures are considerably lower than those on the appetite
juice of Mr. V.

2. The ammonia of the gastric juice. The ammonia of the fresh gas-
tric juice was determined by a combination of Folin’s aeration and the
Nessler colorimetric methods, using 1 to 5 ee. of the juice. It ap-
pears that the ammonia cannot be determined by the Nessler reagent
directly in pure gastric juice, as parallel tests on the same samples of
gastric juice yielded higher figures by aeration and Nessler than by
Nessler direct (using 1 cc. of the juice).

Our data, summarized in Table III, go to show that

1. Ammonia in the amounts of 2-3 mgr. per 100 ce. is a constant
constituent of pure gastric juice of man and dog.

2. The ammonia appears to be slightly more concentrated in the
continuous secretion or hunger juice than in the appetite juice.

3. The ammonia appears to be greatly increased in gastric ulcers
(dog).

Rosenheim (17) and Strauss (18) reported small amounts of am-
monia in the gastric content of man. Zunz (19), working with the
gastric content (test meals) on normal persons and on persons with
various disorders of the alimentary tract, also reports the presence of
ammonia. In the normal individuals the ammonia of the test meal
contents varied from 0.7 to 5.0 mgr. per 100 ce. In cancers of the
stomach the ammonia in the test meal content is greatly increased,
according to Zunz’s figures. Zunz concludes there is no relation be-
tween the state of the gastric digestion or the acidity of the gastric

254 A. J. CARLSON

content and the quantity of ammonia in the gastric content. Never-
theless, the test meal introduces factors (bacterial action, saliva, ete.)
not present in pure gastric juice. Sommerfeld (13), working with pure
gastric juice of a ten-year-old girl with complete stricture of the
oesophagus, states that gastric juice contains no ammonia. Nencki,
Pawlow, Zaleski (20) and Salaskin (21) reported 4 to 4.5 mgr. am-
monia per 100 ce. pure gastric juice of the dog. Rosemann (2) reports

TABLE III
: The ammonia in gastric juice
NHs In MGR. PER 100 cc. GASTRIC
; NOZCe JUICE
INDIVIDUAL MATERIAL OBSERVA-
TIONS
Maximum | Minimum | Average
Mr. V : Appetite juice 30 5.5 2.0 3.0
tate ‘ Hunger juice 10 5.6 3.0 4.0
IM ese Eee anes colar crac Appetite juice 8 3.0 Les 0
A ite jui 5 5 yf ;
Wie (O'S eienties oe 2 Peewee ee 2 eee ine aaa
Hunger juice 3 10.0 9.5 9.8
aia “foes | | Appetite juice 20 4.5 1.5 2.5
bi ie Digestion juice 24 4.5 1.5 2:5
pouches, ...=-
Dogt Pawlow
pouchand Continuous secretion 5 25 .0 10.0 18 .0

ulcer in pouch

*Mr. E. is a second gastric fistula case, a man age 24, partial cicatricial
stenosis of oesophagus from drinking lye. Gastrostomy of five months’ standing.
The man is in good general health, and shows no gastric disturbance.

+ Mr. C. noticed increased hunger pains, and at times slight epigastric pain
for a period of several months when these tests were made. There was no varia-
tion in the diet.

t This dog had an experimental ulcer in the stomach pouch produced by
intravenous injection of a strain of streptococci isolated from a gastric ulcer in
man by Dr. C. E. Rosenow. This dog was used in another line of work by my
assistant, Mr. L. L. Hardt.

the constant presence of a small amount of ammonia in the pure gastric
juice (appetite secretion) in the dog. Reisner (29) concludes that the
ammonia in gastric juice comes from the saliva.

What is the origin and significance of the gastric juice ammonia? It
is known that saliva contains traces of ammonia. We find that the
mixed saliva of man (A. J. C.) contains from 0.5 to 1.5 mgr.ammonia
per 100 cc. Salaskin found 2.5 mgr. NH; per 100 ce. in the saliva of
the dog. But in dogs with Pawlow’s stomach pouch, and in my gastric

ve ee ae

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 255

fistula case, Mr. V., no saliva can enter the stomach or the part of the
stomach yielding the juice. In the present work the saliva therefore
is not a factor.

The ammonia of the duodenal content may be a factor, as Boldyreff
has shown the frequency with which intestinal content enters the
stomach. This factor is excluded in dogs with the Pawlow stomach
pouch. In my human fistula case this factor is readily controlled by
making the ammonia determinations only on those samples of gastric
juice that are absolutely free from admixture with bile, pancreatic
juice, and succus entericus.

Rosemann points out that the gastric juice ammonia cannot be a
simple filtrate from the blood since normal blood contains only about
0.5 mgr. of ammonia per 100 ce.

The above considerations seem to limit the origin of the gastric juice
ammonia to the following factors:

1. It may be an active excretion from the blood, in which case one
would expect an increase in the gastric Juice ammonia on increasing
the blood ammonia. In normal individuals there appears to be a
decrease in the elimination of ammonia by the kidneys during gastric
secretion (Gammeltoft (28).

2. The ammonia may be formed in the secretion process itself.

3. The action of the hydrochloric acid on the proteins of the gastric
juice (conversion of alkaline or neutral proteins to acid proteins).

4. The action of the hydrochloric acid on cells of the mucosa (splitting
off of ammonia by the cells as a protective measure against the action
of strong acids).

5. In the case of gastric ulcers of infectious origin ammonia may
actually be produced continually by bacterial action in the active
focus of the ulcer.

6. In the case of traumatic or non-infectious ulcers it seems highly
probable that the action of the hydrochloric acid on the raw surfaces
of the gastric mucosa will result in the formation of ammonium chloride.

In 1898 Nencki, Pawlow and Zaleski (20) reported studies on the
ammonia concentration in the gastric mucosa and its relation to the
secretion of gastric juice. They found that, per unit of mass, there is

more ammonia in the gastric mucosa than in any other tissue of the body.

100 gr. gastric mucosa at rest contained 20 mgr. NH;. 100 gr. gastric
mucosa’after 2 hours secretion (sham feeding) contained 42 mgr. NHs3.
_ These findings were essentially confirmed by Salaskin the same year.
They seem to indicate some relation of the ammonia formation to the

256 A. J. CARLSON

secretion process itself, unless the higher ammonia content of the
secreting mucosa represent ammonium chloride in the process of ab-
sorption from the gastric juice. The other possible factors referred to
above are capable of experimenta tests and work is in progress in that
line.

3. The amino-acids of the gastric juice. The formol titration was made
on eight lots of appetite gastric juice of Mr. V., and on three lots of
appetite gastric juice of Mr. E. When deductions were made for the
ammonia nitrogen of the juice, the formol titrable nitrogen of Mr.
V.’s gastric Juice varied from 3 to 9 mgr. nitrogen per 100 cc. Mr.
E.’s appetite Juice gave 7 mgr. per 100 cc. Four lots of dog’s appetite
gastric juice (Pawlow pouch) gave only 1 to 2 mgr. of amino-acid nitro-
gen. It thus appears that normal human gastric juice contains slightly
more amino acid than ammonia nitrogen; but the greater part of the
gastric juice nitrogen is associated with the more complex proteins.

Zunz (19) working with gastric contents after test meals reports that
the amino acid nitrogen usually exceeds the ammonia nitrogen, and
that both substances are increased in cases of gastric cancer. In three
normal persons the maximum amino acid nitrogen was 10 mer. per
100 cc. of gastric content, while in several gastric cancer cases it reached
15-20 mgr. per 100 ce. of content. But these figures cannot be directly
compared with ours on pure gastric juice, because of the uncertain
factors associated with the gastric contents following a test meal.

4. The auto-digestion of the gastric juice. When fresh gastric juice is
incubated at 38° C. the following changes take place in the proteins
and the gastric mucin.

(1) All the ropy mucin and mucin flocculi are dissolved.

(2) The pink color of the biuret reaction is increased. In fact, fresh
human gastric juice gives practically a violet biuret reaction, and this
color is intensified and changed towards pink by the auto-digestion.

(3) The characteristic protein precipitation at the point of neutrali-
zation is decreased.

(4) The quantity of proteins precipitated by nitric acid and by heat
is reduced.

It is thus clear that the proteins of pure gastric juice undergoes
pepsin-hydrochloric acid digestion in the stomach itself. But some
of the gastric juice proteins are not hydrolyzed, at least not down to
the peptone stage. It has been pointed out in a previous report. (1)
that this auto-digestion of the gastric juice itself is probably a factor
in the continuous secretion of gastric juice in the way of yielding gas-
tric secretagogues.

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE

TABLE IV

257

Anaphylactic reactions of guinea pigs, using human gastric juice and human serum as
Pigs 7, 8 and 9 were sensitized with human gastric
juice that had been incubated at 38°C. for 10 days

sensitizing and toxic agents.

= SENSITIZING DOSE
1 | 0.1 ce. fresh g. juice
2 |. 0.2 cc. fresh g. juice
3| 0.5 ec. fresh g. juice
4/ 1.0 ce. fresh g. juice
5 | 1.5 ce. fresh g. juice
6] 2.0 ce. fresh g. Juice
7| 2.0 ce. digested g. juice
8 | 1.0 cc. digested g. juice
9] 0.2 ce. digested g. juice
10} 1.0 ce. fresh g. juice
|
11 | 2.0 ce. fresh g. juice
12| 5.0 ce. fresh g. juice
13 | 0.5 ce. fresh g. juice
14| 1.0 ce. fresh g. juice
15 | 5.0 cc. fresh g. juice
16 | 0.5 ce. fresh g. juice
17 | 1.0 cc. fresh g. juice
18 | 2.0 ce. fresh g. juice
19 | 10.0 ce. fresh g. juice
20 | 10.0 cc. fresh g. juice.
21 | 10.0 cc. fresh g. juice
\22 10.0 ce. fresh g. juice
23 | 10.0 ce. fresh g. juice
24 | 10.0 cc. fresh g. juice
25 | 5.0 ce. human serum
26 | 5.0 cc. human serum
27 |} 5.0 ce. human serum
28 | 5.0 cc. human serum
29 0.05 ec. human serum
30 | 0.05 ec. human serum
31} 0.10 cc. human serum
82 | 0.10 cc. human serum
33 | 0.15 ce. human serum
34 0.15 cc. human serum
35 | 0.25 ec. human serum
36 | 0.25 ec. human serum
37 | 0.5 ce. human serum
38 | 0.5 cc. human serum

a4

INTERV AL

15 days

15 days
15 days
15 days
15 days
15 days
20 days

20 days
20 days
23 days

23 days
23 days
27 days
27 days
27 days
30 days

30 days |.

30 days
18 days
18 days
18 days
18 days
18 days
18 days
18 days
18 days
18 days
18 days
27 days

27 days
days
days
days
days
days
days
days
days

NNONNNNND
es en

CS SS |

TOXIC DOSE

juice
juice
juice
juice

:. fresh g.
. fresh g.
. fresh g.
.. fresh g.
. fresh g. juice
. fresh g. juice
-c. human serum

ehh ee be bb
QQ lOM iO ae (6. 3
Col SCOECo) IES eS Ito) (O

. fresh g. juice
3 cc. fresh g. juice
0.5 fresh human serum

.O fresh g. juice
.0 fresh g. juice
.O fresh g. juice
0.5 ce. human serum
5.0 fresh g. juice
0.5 ec. human serum
5.0 fresh g. juice
.0 fresh g. juice
1.0 ee. human serum
1.0 ec. human serum
1.0 ec. human serum
1
1

or Or

Or

on

.0 ec. human serum
.0 ce. human serum
1.0 ec. human serum
10.0 fresh g. juice
10.0 fresh g. juice
10.0 fresh g. juice
10.0 fresh g. juice
1 ce. human serum

10.0 fresh g. juice
2.0 ce. human serum
10.0 fresh g. juice
2.0 ee. human serum
10.0 fresh g. juice

1.0 ec. human serum
10.0 fresh g. juice

1.0 ec. human serum
10.0 fresh g. juice

ANAPHYLACTIC
SYMPTOMS

None

Doubtful

None

Doubtful

Doubtful

None

Severe; death in
1} hrs.

None

None

Severe; death in
1 hr:

Doubtful

Slight

Doubtful

Severe; recovered

None

Severe; recovered

Doubtful

None

Severe

Severe

Severe

Severe

Severe

Severe

None

Doubtful

None

None

Severe; death in
13 hrs.

No symptoms

Severe; recovered

None

Severe; recovered

Doubtful

Severe ;death Lhr.

None

Severe;death1 hr.

None

258 A. J. CARLSON

5. The nature of the gastric juice, proteins and their relation to the
serum proteins as shown by the anaphylactic reaction.

This series of experiments on guinea pigs with human gastric juice
as a sensitizing and toxic dose for anaphylaxis were made as a pre-
liminary step for the study of pathological gastric juice. The injec-
tions were made intraperitoneally with a blunt needle. The gastric
juice was rendered slightly alkaline by titration with 7 NaOH. It is
well known that at the neutral point of the gastric juice some of the
proteins are precipitated, and that these go into solution on rendering
the juice slightly alkaline. The human serum was obtained, not from
Mr. V. who yielded the gastric juice, but from Mr. Rogers, who assisted
in this part of the work.

The results are summarized in Table IV. These data seem to show
that: (a) Normal human gastric juice contains a sensitizing substance
for human serum, but that it 7s practically devoid of toxic substances both
in relation to the proteins of the gastric juice and to serum proteins. This
is rather an unexpected situation, in view of the fact that the gastric
juice contains complex proteins in considerable amounts. (b) The
gastric juice contains some practically unchanged serum proteins cap-
able of acting as sensitizer, but incapable or in too small concentra-
tion to act toxically. This sensitizing substance is not diminished or
destroyed by auto-digestion of the gastric juice at 38° C. for ten days
(Table IV, pigs 7, 8, 9).

It is of interest in this connection to recall that according to Gay and
Adler (22) the euglobulin (of horse serum) is sensitizing but not toxic.
And Wells (23) has shown that enzyme (trypsin) hydrolysis of proteins
destroys the toxic preperties of the proteins very must faster than
the sensitizing properties, the latter persisting even after three years
continued tryptic digestion.

V. The acidity of normal gastric juice. The acidity was determined
by titration with =~; NaOH, and using dimethyl-amino-azo-benzene
and phenolthalein as indicators for the free and the total acidity re-
spectively. During the three years that Mr. V. has been under obser-
vation hundreds of determinations have been made of the acidity of
the contents of the ““empty”’ stomach, of the hunger juice or continu-
ous secretion, and of pure appetite juice. The reader will recall that
the contents of the ““empty”’ stomach is taken one hour after washing
out the stomach with 200 ce. of water. All the cases where the gastric
juice or gastric content was contaminated with bile (intestinal con-
tent) are excluded from the summaries given in Table V.

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 259

The second gastric fistula case, Mr. E., is a man 26 years of age,
healthy and vigorous. Nearly a year ago his oesophagus was cor-
roded with a solution of lye, and this led to a nearly complete cica-
tricial stenosis. The gastrostomy was made in October, 1914. I am
under obligations to Dr. Bayard Holmes for the privilege of working
with Mr. E. daily for two weeks. My observations were made in
February, 1915. At that time the oesophagus had been dilated suffi-
ciently to permit swallowing of any well masticated food and the gas-
trostomy opening was used only in the dilation processes. In this
case saliva is therefore not excluded from the contents of the empty
stomach, and possibly not from the continuous or hunger secretion,
although Mr. E. was instructed and urged not to swallow any saliva
during these experiments. The appetite juice was obtained by Mr. E.

TABLE V

The acidity of normal human gastric juice
ACIDITY

PERSON MATERIAL

NO. CF OBSER-
VATIONS
be

. 4
®
cs)
el
°
oe
r]
ee

Cont. empty stomach.| 235 | 0.10 | 0.35 | 0.18 | 0.15 | 0.40 | 0.28
Mr. V. .| ; Hunger juice..........} 180 | 0.15 | 0.35 | 0.25 | 0.20 | 0.45 | 0.34
Appetite juice.........| 285 | 0.35 | 0.44 | 0.40 | 0.40 | 0.53 | 0.48

Cont. empty stomach. 10 | 0.09 | 0.36 | 0
Mr. E. .| ; Hunger juice.......... 8 | 0.20 | 0.32 | 0.
Appetite juice......... 15 | 0.30 | 0.36 | 0

chewing palatable food, and spitting out the chewed food, care being
taken not to swallow saliva or particles of food.

These results on my two gastric fistula cases are in agreement with
the work of Pawlow and his pupils on dogs, and the work of previous
observers on pure gastric juice of normal persons. The latter data
have recently been brought together and discussed by Boldyreff (24).

Normal human gastric juice (appetite secretion) when secreted above a
certain minimum rate shows a practically constant total acidity of nearly
0.5 per cent HCl, or the same as the gastric juice of normal dogs. The
gastric juice (appetite as well as hunger juice) secreted by the normal
stomach at a low rate shows lower than normal acidity and total chlo-
rides. The view of Pawlow based on experiments on dogs that gastric
juice is secreted at uniform and constant acidity is true for man only

260 A, J. CARLSON

in regard to the appetite, digestion, and hunger juice secreted at fairly
high rate. We must take cognizance of the equally important fact
that the normal gastric mucosa is capable of secreting a juice of sub-
maximal acidity.

Last year Rehfus and Hawk (25) reported a series of interesting
observations on the acidity of the gastric content at varying periods
after drinking water, and ingesting an Ewald test meal, which they
offer as “direct evidence of the secretion of gastric juice of a constant
acidity in the human stomach,” and as “the first direct demonstration
in the human subject in favor of Pawlow’s theory of the secretion of
a juice of a constant acid concentration.”” Rehfus and Hawk are deal-
ing with the gastric content, not with pure gastric juice. The fact
that the gastric content soon reaches and then maintains for some
time a fairly constant acidity indicates the equilibrium between the
acid secretion and the factors of neutralization, as shown by Boldyreff,
but. it does not show that the juice is poured into the stomach at a
certain constant acidity. Analysis of the Ewald meal or the continu-
ous secretion do not permit us to draw any conclusion in regard to the
acidity of the actual gastric juice. If the authors mean by the above
conclusions that the gastric juice of normal persons, that is the pure
appetite juice, or the juice secreted during the first two or three hours
of digestion exhibit a fairly constant acidity of nearly 0.5 per cent
HCl, they are undoubtedly correct; but this conclusion is not new, as
shown by Boldyreff’s review of the earlier literature, and it cannot be
based on the observations they report.

The reader will note that normal human gastric juice is equal in total
acidity to the maximum acidity reported by clinical observers for so-called
hyperacidity in man. So far as I am acquainted with the literature,
there is no evidence that the gastric glands under any pathological condi-
tions are able to or do secrete a juice of higher than normal acidity. More-
over, the presence in the stomach of gastric juice of full acid strength
leads by itself and immediately to no untoward symptoms.

The contents of the ‘‘empty”’ stomach, and the continuous or hunger
secretion (when the secretion rate is low) have uniformly a lower acidity
than the appetite juice. The total acidity of contents of the “empty”
stomach is about 0.2 per cent. The reader will note that this figure is
frequently given, especially by clinicians, as the acidity of pure gastric
juice of normal persons. The acidity of the continuous or hunger
secretion is higher, and the greater the secretion rate the higher the
acidity until it may equal that of the appetite juice. Inno instance

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 261

does the acidity of the continuous secretion exceed that of the appetite
juice.

What is the cause of the low acidity of the continuous secretion and °
contents of the empty stomach? The following factors must be taken
into account:

1. The actual acidity of the juice as secreted may increase with the
secretion rate, until the maximum acidity is reached with the high
average rate of secretion, a condition similar to that obtaining in the
case of the salivary glands where the concentration of the salts and the
organic materials increase with the rate of a salivary secretion. If
this is a factor the gastric juice secreted at a low rate should show a
lower osmotic concentration and total chlorides than the juice secreted
at high rate. The figures reported by Umber (12) for man and by
Rosemann (2) for the dog appear to support this view, the former
investigator showing particularly that the osmotic concentration (A)
of the gastric juice increases with the rate of secretion. The cryoscopic
data may, however, be misleading, as the salts produecd by the neu-
tralization of the HCl may not dissociate as freely as the acid.

2. The slower rate of secretion may give chance for the HCl to be
partly neutralized by the alkaline mucus secreted by the mucin cells
of the gastric mucosa. This is the factor emphasized by Pawlow. In
fact, Pawlow takes the position that in the normal animal gastric
juice has practically a constant acidity, irrespective of the secretion
rate, but the actual acidity of the juice in the cavity of the stomach is
purely a matter of rate of neutralization. If this is the sole factor, the
total chlorides of the gastric juice ought to show a greater constancy than
the acidity. That the hydrochloric acid of the gastric juice is in part
neutralized by the gastric mucus is obvious. But according to Boldyreff
the alkalinity of gastric mucus is only 0.05-0.10 per cent Na2CO;. That
is to say, it would require 50-100 cc. gastric mucus to reduce 100 cc.
gastric juice from the normal acidity of 0.45 per cent down to 0.25 per
cent. The importance of this factor has therefore been overestimated
by Pawlow.

3. When the gastric juice is collected from a Pawlow accessory
stomach, or from an individual with complete closure of the oesophagus,
as is the case with Mr. V., the saliva cannot be a factor in lowering
the gastric juice acidity by neutralization and dilution. When all or
most of the saliva is swallowed the acidity of the gastric juice is neces-
sarily reduced in proportion to the relative rate of salivary and gastric
secretion. This is effected by dilution rather than by neutralization,

262 A. J. CARLSON

as the titration alkalinity of saliva is low (0.08 per cent NasCO;, Neu-
meister, cited by Boldyreff).

4. According to Boldyreff the most important factor in lowering the
acidity of gastric juice from that actually secreted by the gland (0.5
per cent) to that usually found in the cavity of the stomach (0.25 per
cent) is the entrance of intestinal contents (pancreatic juice, bile, and
succus entericus) into the stomach. Boldyreff has shown that this
occurs in dogs, and probably occurs in man, when the acid in the
stomach mounts much above 0.25 per cent. I am satisfied from my
observations on Mr. VY. that Boldyreff’s view is essentially correct.
When 15-25 ce. appetite gastric juice of full normal acidity is per-
mitted to remain in the empty stomach, even a few minutes, the
pylorie sphincter dilates, and the contents of the upper end of the duo-
denum is in some way forced into the stornach. The quantity of
intestinal content entering the stomach at any one time varies from a
trace up to 10 ce., and the acidity of the gastric juice is correspondingly
reduced by dilution and neutralization. The presence of duodenal
content in the empty stomach is therefore a normal occurrence, but
the mechanism involved in this intestinal regurgitation has not been
closely worked out. We know from the work of Cannon and others
that an acid reaction on the stomach side of the pylorus tends to dilate
the pyloric sphincter. But it is difficult to see how a patent pylorus
can cause the pancreatic juice and bile to enter the stomach without
an actual antiperistalsis in the duodenum.

It is clear that it is this ‘mechanism for self regulation of the acidity
of the stomach content’? (Boldyreff) which breaks down in cases of
so-called ‘‘hyperacidity’ in man. In eases of ‘‘hypersecretion” the
quantity of juice secreted is greater than normal, and the secretion
may persist in the absence of all normal stimuli, but the neutralizing
factors suffice to reduce the acidity of the juice approximately to that
found in the normal stomach. It is purely a balance of secretion rate
and of neutralization capacity. Impairment of the neutralization fac-
tors or very excessive secretion rate of gastric*juice would tend to
render the acidity of the gastric content equal to that of pure gastric
juice; in other words, produce clinical “‘hyperacidity.”

VI. The total chlorides of the gastric juice. The total chlorides of
the gastric Juice were determined by the method of Volhard on 10
different lots of appetite juice and 11 different lots of the hunger juice
or continuous secretion. The results are summarized in Table VI.
The secretion rate and the acidity of the various lots of juice are also

—_yhe

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 263

recorded. It will be seen that the total chlorides of the appetite juice
are very constant, the minimum being 0.49 per cent and the maximum
0.56 per cent chlorine. The continuous secretion or hunger juice is
more variable in chloride content, and this variation appears to be
directly dependent on the secretion rate and on the acidity. In gen-
eral the lower the secretion rate the lower the acidity and the lower

TABLE VI

The total chlorides of normal human gastric juice

SECRETION RCETA
dg GASTRIC JUICE BANE. PHS || 1) nore CL.
HE Free Total
cc.
1 { Hunger juice 6 0.20 0.32 0.40
2 Appetite juice 560 0.42 0.48 0.56
9 { Hunger juice 4 0.18 0.25 0.25
ee Appetite juice 308 0.41 0.48 0.53
3 { Hunger juice 8 0.30 0.40 0.38
tg Appetite juice 500 0.41 0.49 0.54
4 { Hunger juice. 12 0.30 0.41 0.44
a Appetite juice 398 0.42 0.49 0.55
(| Hunger juice I 2 0.20 0.32 0.28
Gere Appetite juice 154 0.42 0.48 0.56
Hunger juice II 10 0.29 0.38 0.36
6 { Hunger juice 2 0.20 0.33 0.34
i: Appetite juice 200 0.41 0.46 0.53
7 { Hunger juice 6 0.30 0.41 0.43
ied) Appetite juice 280 0.45 0.52 0.56
8 { EMM PERMUTCES 655+. o ose saa 3 0.18 0.28 0:27
i, a Appetite juice 150 0.41 0.47 0.53
9 { Hunger juice 2 0.14 0.29 0.24
merase Appetite juice 210 0.48 0.50 0.56
10 { Hunger juice 3 0.26 0.36 0.36
Bareis iPapnetiteryuices. 4. ........ 0.0. 100 0.32 0.40 0.49
i { Hunger juice 36 0.34 0.40 0.53
eetes Appetite juice(mixedwithbile)} 228 0.32 0.37 0.56

the total chlorides. This is in agreement with the findings of Foster
and Lambert (26) on dogs.

These facts seem to point to the conclusion that the low acidity of
the gastric juice secreted at a slow rate is not due entirely to neutrali-
zation by alkaline mucus. We have apparently a secretion of gastric
juice of an acidity actually lower than that of the rapidly secreted

264 A. J. CARLSON

appetite juice. The dependence of the actual secreted acidity on the
secretion rate is not a very close one, however, as we may have very
marked fluctuation in rate without any change in chlorides. But
below a certain secretion rate (25-30 ce. per hour from the entire stom-
ach of the adult) an actual hypoacid juice is secreted.

The above figures for total chlorides in normal human gastric juice
agree closely with the findings of previous observers on the gastric
juice of dog and of man. Rosemann gives 0.54—0.64 per cent Cl for
the appetite gastric juice of the dog. The figures given by Sommer-
feld for human appetite juice varies from 0.53 to 0.59 per cent Cl.
Umber working on an old man (59 years) with partial esophageal
stenosis (malignant) reports total chlorides of the gastric juice as
varying from 0.27 to 0.60 per cent Cl. The reader will note that the
lowest figure reported by Umber is practically identical with the lowest
total chlorides found by me in the very slowly secreted hunger juice.
It is probable that the low chloride gastric juice samples of Umber
were secreted at a slow rate.

VII. The concentration of pepsin. Most of the pepsin determina-
tions were made by the method of Mett; a few tests were also made
with the ricin, and the U.S. Pharmacopceia methods. The tubes (2-23
mm. diameter) were sealed at both ends with heat and the egg albumin
coagulated by boiling for 10 minutes, and the tube left in the water
until cooled, and then set aside for two days before using.

The digestion mixture was made up as follows: 1 ce. gastric juice
+ 15 ec. § HCl, + two Mett’s tubes (13 em. length) at 37° C. for 24
hours. The + HCl was used because it approximates the free acidity
of appetite gastric juice, although we are aware of the fact that this
degree of acidity is somewhat higher than the optimum for artificially
prepared pepsin (ef. Cobb (27), Sérensen (30), Michaelis and Davidsohn
(31)). We propose to determine the optimum acidity for the pepsin in
normal gastric juice in a later work. It would seem singular if normal
gastric juice is secreted with an acidity too great for optimum peptic
digestion. It may be so, however, for under normal conditions the
actual peptic digestion must be carried out under less than normal
acidity of the gastric juice, because of dilution with saliva and the
water of the food, and the fixation of the acid by the protein of the
food.

Our results with the Mett’s test on the gastric juice of Mr. V. may
be given in the following summary:

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 265

Appetite gastric juice.

Number of tests (Lots of G. J.): 45.

Digestion in mm.: High, 83; low, 6; average, 7.
Hunger juice (continuous secretion):

Number of tests: 35.

Digestion in mm.: High, 83; low, 5; average, 6.
Contents of the ‘‘empty’’ stomach:

Number of tests: 35

Digestion in mm.: High, 4; low, 3; average, 3.

The results are stated in the length of albumin column actually
digested, because according to Cobb the law of Schiitz does not hold
for pepsin in concentrations that digest more than 4-5 mm. in 24 hours.

The variations in pepsin concentrations in these three groups of

TABLE VII

The relation of total acidity and pepsin concentration (Mett’s method) innormal
human gastric juice (Mr. V).

CONTENTS ‘‘EMPTY”’

Saat HUNGER G. JUICE APPETITE G. JUICE
SERIES pases bes ee ee eee ne
Acidity Pepsin Acidity Pepsin Acidity Pepsin
per cent mm. per cent mm. per cent mm.
27 eee eee 0.18 4 0.40 6 0.48 if
2S. . See eee 0.10 3 0.25 6 0.48 1S
339.5 5 ee 0.18 33 0.41 63 0.47 7
og Se 6 ee 0.15 4 0.38 if 0.48 TS
Li) ater eee 0.15 a 0.40 63 0.49 iz
2S. Se eee 0.22 ft 0.28 63 0.41 7
ae 0.10 3 0.35 6 0.44 62

_ gastric juice, and the absence of a close relation of pepsin concentra-
tion and acidity is further illustrated by a few typical series of tests
given in greater detail in Table VII.

The appetite gastric juice of Mr. E., our second gastric fistula case,
when tested as above in 14 experiments showed a pepsin concentration
of 5-7 mm. with an average of 6 mm., a slightly lower value than the
gastric juice of Mr. V.

We have tested as above the gastric juice (appetite, as well as diges-
tion secretion) of 25 dogs with Pawlow fistula. The dogs, normal and
in good condition, were being used in other lines of work. We were
surprised to learn that the dog’s gastric juice showed uniformly a lower
pepsin concentration than the gastric juice of man. The figures ob-

266 A. J. CARLSON

tained on dog juice was: highest, 5 mm.; lowest, 2 mm., with an average
of 34 mm. The reader will note that this is only about half the quan-
tity of digestion obtained in the human juice.

When the Mett’s tubes are placed in 16 ce. of undiluted human
gastric juice (appetite secretion) the digestion in 24 hours at 37°C.
varies from 12-16 mm., or only twice the quantity digested in the
dilution of 1 cc. juice to 15 ec. 45 HCl. This seems to indicate that
‘in normal gastric juice the pepsin is present in excess of the needs or at
least far in excess of that needed in economic digestion.

The U. 8S. Pharmacopeceia defines ‘100 per cent Pepsin as a prepa-
ration capable of digesting three thousand times its own weight of
finely divided egg white (coagulated) in three hours.’”’ The Pharma-
copeeia test is carried out as follows: 10 grams of boiled white of egg is
macerated through a No. 40 filter and placed in 40 cc. 0.3 per cent
HCl, 34 mgr. dried pepsin added, and the mixture incubated at 52° C.
for three hours, with occasional stirring When thus carried out there
is only a very small residue of undissolved egg white at the end of
three hours, but the procedure to measure the amount of this residue
does not yield very accurate results.

This test was applied to six different lots of appetite gastric juice
of Mr. V. Under above conditions, 1-14 cc. appetite gastric juice
digest 10 gr. of coagulated and finely divided egg white in three hours
practically as completely as is done by 33 mgr. “100 per cent pepsin.”
As defined by the U. 8. Pharmacopeeia, 1 ec. of human gastric juice
must therefore contain 33 mgr. pepsin, or 100 cc. of the juice, 35 mgr.
pepsin. We have:seen that the appetite gastric juice of man contains
about 400 mgr. organic material per 100 cc. That is, according to the
Pharmacopeeia definition, only about 10 per cent of the organic matter
in the human gastric juice is pepsin. Of course, we are aware of the
fact that the Pharmacopeeia definition and test is only one of conveni-
ence for standardizing commercial preparations of pepsin, and having
no direct relation to pepsin concentration in normal gastric juice.

It has been shown in a previous report that an adult normal person
if hungry secretes 600-700 cc. gastric juice on an average palatable
dinner, or about 1500 cc. gastric juice total in 24 hours. That is to
say, there is a secretion of 240-250 mgr. pepsin per dinner, capable
under proper conditions of digesting from 630 gr. to 750 gr. of protein
(coagulated and finely divided egg albumen) in three hours; and the
total pepsin secretion in 24 hours is 525 mgr., capable of digesting
134 kilo proteins (coagulated egg white) in three hours.

— — =

1 Se eee

CHEMISTRY OF NORMAL HUMAN GASTRIC JUICE 267

The practically complete digestion of 10 gr. boiled and finely divided
egg white by 1 cc. human gastric juice in three hours is a fact. The
secretion of 600-700 cc. gastric juice on an avergage dinner by the
adult human stomach is an estimate, but not far from the actual fact.
It is therefore clear that the normal human stomach secretes pepsin
far in excess of the actual needs of gastric digestion, or, more precisely,
far in excess of what can be used advantageously under ordinary con-
ditions of gastric digestion When the boiled egg white is broken up
in larger pieces, such as occurs in ordinary rapid mastication, 1 cc. of
gastric juice requires 6-10 hours for complete digestion.

This great excess of pepsin in normal gastric juice probably explains
the clinical findings of great reduction in pepsin content without any
evidence of impaired gastric digestion, provided sufficient acid is pres-
ent. It probably also explains, in part at least, the practical useless-
ness of commercial pepsin as a therapeutic measure in gastric disorders.

REFERENCES

(1) Cartson: This Journal, 1915, xxxvi, 30.

(2) Rosemann: Arch. f. d. ges. Physiol., 1907, exviii, 467.

(3) SomMEeRFELD: Arch. f. Anat. u. Physiol., 1905, Suppl., 455.

(4) C. Scumipt: Anal. de Chem. et Pharm., 1854, xlv, 46.

(5) AuBu: Zeitschr. f. exp. Pathol. u. Pharm., 1908, v, 17.

(6) PawLow anD ScHoumow-Simmanowsk!: Arch. f. Anat. u. Physiol., 1895, 58.
(7) KnowatorrFr: Maly’s Jahrb., 1893, xxiii, 289.

(8) Schoumow-Simanowsk!: Arch. f. exp. Pathol. u. Pharm., 1894, xxx, 342.
(9) NeNcKI AND Sreser: Zeitschr. f. physiol. Chem., 1901, xxxii, 302.

(10) FriepeNnTHAL: Arch. f. Physiol., 1900, 186.
(11) Kanetson: Arch. f. d. ges. Physiol., 1907, exviii, 327.

(12) Umper: Berl. kl. Woch., 1905, xlii, 56.

(13) Bickex: Berl. kl. Woch., 1905, xlii, p. 60.

- (14) Sasakr: Ibid., p. 1386.

(15) Lenumann: Zeitschr. f. exp. Pathol. u. Pharm., 1906, iii, 359.

(16) PEKELHARING: Zeitschr. f. physiol. Chem., 1902, xxxv, 8.

(17) RosenHEmM: Zeitschr. f. kl. Med., 1892, xiii, 817.

(18) Strauss: Berl. kl. Woch., 1893, 398.

(19) Zunz: Intern. Beitr. z. Pathol. u. Therap, d. Ernarungsstérungen, 1910,

re eye

(20) NeNck1, PawLow anp Za.eski: Arch. f. exp. Pathol. u. Pharm., 1898,
_ XXxvVil,. 26.

(21) Sauaskin: Zeitschr. f. physiol. Chem., 1898, xxv, 448; 1902, xxv, 246.

(22) Gay anp ApteR: Journ. Med. Research, 1908, xviii, 433.

(23) Wetts: Jour. Inf. Dis., 1908, v, 449.

(24) BotpyrerrF: Quart. Journ. Exp. Physiol., 1914, viii, 1.

(25) Renrus anp Hawr: Jour. Am. Med. Assoc., 1914, lxiii, 2088.

268 A. J. CARLSON

(26) Foster anp LamBert: Journ. Exp. Med., 1908, x, 820.

(27) Coss: This Journal, 1905, xiii, 448.

(28) GammuttortT: Zeitschr. f. physiol. Chem., 1911, Ixxv, 51.

(29) Reisner: Zeitschr. f. kl. Med., 1903, xlvili, 110.

(30) SérENSEN: Bioch. Zeitschr., 1909, xxil, 131.

(31) Micuaniis AND DaviIpson: Zeitschr. f. exp. Pathol. and Therap., 1910,
Vili, 2.

THE CONTENT OF SUGAR IN THE BLOOD OF CATS UNDER
THE INFLUENCE OF COCAINE

EDWARD WALDO EMERSON SCHEAR
From the Department of Physiology of Columbia University

Received for publication June 29, 1915

It has become apparent that the chemical composition of the blood
is very rapidly changed by the conditions to which animals are sub-
jected both preliminary to and during the operation of drawing sam-
ples of blood. Especially is this true with reference to the carbohydrates.

Thus Pavy (1) shows not only that it is necessary that the animals
be kept free from excitement while the sample is being drawn, but also
that when the animal is killed and all the bloodisused, “. . . it
is necessary to proceed with the utmost expedition in order that
no change may take place in the contents of the circulation from the
post mortem production of sugar in the liver.” These results have
been substantiated by many investigators since the publication of
Pavy’s work and are treated in considerable detail in a recent paper
by Scott (2), in which the results of extensive researches are given,
with numerous references to the literature both on glycaemia and
glycosuria. Still more recently Shaffer (3) shows that the sugar con-
tent of the blood of the dog when ‘‘free from excitement or pain is
surprisingly low and constant.’ Scott also discusses the effects of
- chloroform and ether, and records a few results from the use of cocaine.

It is the purpose of this paper to put the effects of cocaine to a more
thorough analysis by means of a larger series of experiments. The
chemical technique used in the experiments here recorded is the same
as that described by Scott in the paper above referred to, and need not
be repeated. The work was performed under the immediate direction
of Dr. Scott, and it is a pleasure to acknowledge my indebtedness
to him.

DESCRIPTION OF EXPERIMENTS

In these experiments rather large doses of cocaine hydrochloride
dissolved in M/8 sodium chloride solution were given. It is obvious
269

270 EDWARD WALDO EMERSON SCHEAR

that if results to which any reliability can be attached are to be ex-
pected, certain standard conditions must be determined upon which
the effect of the drug itself can be superimposed. The conditions
adopted in this work are the same as those reported by Scott in the
paper above referred to, pages 278 to 293. The chief essentials are
that the animals were kept in captivity for from ten to twenty days
under excellent conditions and were fed exclusively on meat, the natu-
ral diet of the cat.

It is well known that certain other processes of metabolism are
affected by cocaine injections; therefore it was necessary to proceed
with extreme caution in the carrying out of the experiments, as well
as in their interpretation. For example, Araki (4) found that lactic
acid was eliminated in the urine of frogs and rabbits after cocaine
injections; also one rabbit out of four secreted sugar with the urine.
Underhill and Black (5) report that while small doses of cocaine, 10
mgm. per kilo of body weight, cause no appreciable influence upon the
course of nitrogenous metabolism or the utilization of protein and
fat, larger doses, 15 to 20 mgm. impair, first, fat utilization and, sub-
sequently, that of protein, while a marked decline in body weight is
manifest. It has also been shown by Scott and others that the char-
acter of the diet upon which animals are fed will affect the sugar con-
centration. Besides this it is quite evident from the work of a number
of investigators that the glycosuria which has been found in many cases,
as well as the hyperglycaemia which is known to exist even before any
sugar is excreted, is frequently due to the emotional condition of the
animal entirely apart from the diet or the anesthetic.

For these and other reasons only such animals as conformed to the
standard conditions were used in these experiments. Throughout the
work great care was taken to avoid everything that would in any way
tend to excite the animals before drawing the blood for analysis. It
was of course impossible to do this with absolute success in every case,
inasmuch as some of the cats which it was found necessary to use
were unaccustomed to being handled, even though not particularly
wild. Others were somewhat wild. When, however, anything occurred
that might indicate other causes than that of the drug for hypergly-
caemia strict records of such conditions were made. The animals were
killed by’ sudden decapitation, and the blood was obtained from the
severed vessels of the neck.

In table 1 are given the results from five animals which were almost
entirely quiet throughout the handling preliminary to the injection

INFLUENCE OF COCAINE ON SUGAR CONTENT OF BLOOD 271

and during the operation itself. It will be noticed that the amounts of
sugar in the blood here recorded are very close together, the highest
variation from the mean being only 15 per cent. In fact the variation
is but little greater than that found by Scott in normal cats, or than
is to be expected from unavoidable variations in the internal rela-
tionships with their necessary readjustments, as is pointed out by
Cannon (6).

TABLE 1
: TIME 5
NO.OF | sex | “ow | 2OP¥ | toe. | FROM | stoop | “mw | Peon sone
cael DIET 5 INJ. SaaS DRAWN] BLOOD | MEAN
[2 gm. min. gm. gm.
1 F 18 | 2.80 | 0.070} 6 | 62.83/0.0736} +11 | Fairly quiet
2 F 18 | 3.30 | 0.070) 6 | 79.85/0.0720/ + 9 / Quiet
3 F 12 | 3.00 | 0.050} 10 | 77.38/0.0679} + 2 | Ideal
4 F 14 | 2.20 | 0.050} 10 | 50.11/0.0619} — 7 | Practically quiet
5 F 17 | 1.90 | 0.050} 8 | 49.90/0.0546} —15 | Slightly excited be-
fore, but quiet
during injection
Mean 0.0664

The results from the four excited animals reported in table 2 would
scarcely be expected to vary greatly from those of table 1, since the
excitement, while evident, was not at all excessive. The mean is
somewhat higher than the mean of table 1, which result is in accord
with the results obtamed by various investigators already referred to.

TABLE 2
TIME oO sd
DAYS 2 AMT. AMT. |%SUG.| VAR.
eRe SEX ON ean coc. ania BLOOD IN FROM REMARKS
; NJ. que N AD
DIET INJ DEATH | DRAWN| BLOOD | MEAN
Kee gm. min. gm. gm.

6 F 18 | 2.6 | 0.045) 11 | 81.45/0.0815| + 9 | Excited from first
Cried during injec-
3 tion

7 F 12 | 2.3 | 0.050) 5 | 58.63/0.0643} —14 | Cried during injeo-
tion

2.8 | 0.050} 5 | 80.00/0.0832} +11 | Restive

2.15 | 0.050) 6 | 64.65/0.0695| — 7 | Restless; slightly
, excited

Mean 0.0746

272 EDWARD WALDO EMERSON SCHEAR

This conclusion is strengthened by the interesting examples of the
two individuals recorded in table 3. These animals showed excep-
tional reaction to the drug. In the first there was no particular ex-
citement manifest before the injection, but as soon as the operation
was over marked excitement was noticed.

Protocol of Experiment 10

Time of injection 10.05. Quantity, 27 mgm. per kilo of body weight.

10.08. Cries, due to drug; had not been excited.

10.09. Evidence of nausea; deep respiration; vomiting.

10.10. Pupil reflex slow.

10.12. Hind legs weak; respiration rapid; cries; salivation.

10.15. Spasms; after about 10 seconds the first spasm diminished slightly,
and the animal was killed just as the second spasm set in; blood seemed cyanotic.

TABLE 3
TIME
: DAYS _ | AMT AMT. | %SUG
pg SEX ON tite coc. ea taty BLOOD IN REMARKS
DIET INJ pEaTH | DRAWN] BLOOD
k. gm. min. gm. gm.

10 F 11 2.6 | 0.070) 10 | 54.52/0.1393) Quiet at start; drug show-
ed marked influence;
nausea and spasms.

0.035) 27 | 82.92/0.2215| Cat fought during injec-
tion

11 F 11

i)
or

In the other experiment, No. 11, the cat was difficult to handle and
fought from the first. The injection was not entirely successful, only
about one-half of the syringeful of cocaine being introduced. Since
the cat was rather large the injection was thought hardly sufficient
to yield adequate results, hence the animal was rejected and returned
to the cage. In a short time, however, the characteristic symptoms
were manifest: intense salivation, dilated pupils,ete. The cat was killed
twenty-seven minutes after the injection. It will be noticed that the
result in No. 10 was higher than in any other except No. 11, yet the
quantity of cocaine used in the former was approximately the same
as that in other cases; in the latter case the quantity used was only
about half as great, yet the carbohydrate recovered is extremely high.
The reason why the result in No. 11 is higher than in No. 10 is possi-
bly twofold. In the first place the excitement manifested by the cat
was very much greater, and, secondly, the time between the injection
and the killing wat longer. This, however, merely substantiates the
results of a number of investigators previously reported.

INFLUENCE OF COCAINE ON SUGAR CONTENT OF BLOOD 273

CONCLUSIONS

1. If we compare the amount of Sugar in the blood of cocainized
cats in table 1, mean 0.0664, with Scott’s standard of 0.069 for nor-
mal animals, it is seen that the mean of the series here presented
is slightly lower than the standard. This might be taken to indicate
that the direct influence of the cocaine tends to diminish the concen-
tration of the carbohydrate in the blood, and is thus in harmony with
the short series reported by Scott. This difference, however, is so
slight that for most purposes it may be neglected.

2. While an increased concentration of sugar was noted in the blood
of several cocainized animals it seems obvious that this was due to
excitement. It seems a matter of indifference whether this excitement
is due to the action of the drug or to other causes. In this connec-
tion it may be noted that while the cocaine may truly cause hyper-
glycaemia it does so only in an indirect manner.

3. While cocaine may be used for local anesthesia, in drawing blood
for sugar analyses its use involves an element of doubt that would
better be obviated if possible.

LITERATURE

(1) Pavy, F. W.: Journ. Physiol., 1899, xxiv, 479.

(2) Scorr, E. L.: This Journal, 1914, xxxiv, 271.

(3) SHarrer, P. A.: Journ. Biol. Chem., 1914, xix, 297.

(4) Araxi, T.: Zeitschr. f. Physiol. Chem., 1891, lv, 546.

5) UNDERHILL, F. P. and Buacx, C. L.: Journ. Biol. Chem., 1912, xi, 235.

(6) Cannon, W. B., SHont, A. T. and Wricut, W.S.: This Journal, 1911, xxix,
280.

CONTRIBUTIONS TO THE PHYSIOLOGY OF THE STOMACH

XXVI. THE RELATION BETWEEN THE DIGESTION CONTRACTIONS OF
THE FILLED, AND THE HUNGER CONTRACTIONS
OF THE ‘Empty’? STOMACH

F. T. ROGERS anp L. L. J. HARDT
From the Hull Physiological Laboratory of the University of Chicago

Received for publication June 29, 1915

Boldyreff in 1905 (1) first reported that there occurred contractions
of the empty stomach of the dog during starvation. Cannon and
Washburne (2) reported that the same occurred in man. Carlson (3)
by more delicate methods, analyzed the various types of motor activi-
ties of the empty stomach during prolonged hunger. His work showed
that the empty stomach exhibits three types of activity.

1. Rhythmical contractions occurring about three times per minute
designated the “twenty second rhythm.”

2. Very vigorous contractions occurring periodically. The duration
of each of these contractions is about thirty seconds.

3. Tonus changes of the stomach musculature.

In his first paper Carlson states that probably the twenty second
rhythm is the peristaltic activity of the antrum pylori. The thirty
second rhythm or hunger contractions were thought to be contractions
of the fundus. His later work on dogs with Pawlow pouches showed
that the twenty second rhythm is not confined to the pylorus, but is
also exhibited by the fundus (4). The general trend of all of Carlson’s
experiments is toward the conclusion that the first period of hunger
contractions apparent after a meal is an exaggeration of the motor
activities of the stomach occurring during gastric digestion.

In experiments on the rabbit (5) undertaken at Dr. Carlson’s sugges-
tion it was found difficult to differentiate the hunger contractions of
the stomach from the digestion peristalsis. No other conclusion seemed
possible than that the hunger contractions were augmented peristaltic
contractions.

274

CONTRACTIONS OF EMPTY AND FULL STOMACH 275

In conference with Dr. Carlson the following experiments were out-
lined to determine whether or not the same conclusion applied to man
and dog:

We used the two standard methods of investigating the movements
of the stomach: the rubber balloon and the X-ray methods. To study
the genesis of the hunger contractions the rubber balloon, which was
connected by small size rubber tubing with the recording manometer,
was swallowed shortly after a meal and continuous graphic records of
the intragrastric pressure variations were made until after the onset
of a typical hunger period.

Soon after eating an average meal the subject of the experiment
swallowed the balloon and put himself into a comfortable position either
sitting in a chair or lying on a cot. The results were similar, irrespec-
tive of the position; the best results were obtained while the subject
was asleep for then inhibiting psychic influences were removed. Dur-
ing a long continued series of tracings we tried to keep the balloon in
the upper part of the stomach by pulling it up at intervals and swallow-
ing it again. We realize it is possible for it to be pushed to the pyloric
end of the stomach, but by the X-ray observation in the dog we found
that the balloon introduced after feeding remained in the cardiac end
of the stomach for a considerable time.

In order to actually see the hunger movements we coated a balloon
with a bismuth paste and observed its movements in the stomach with
the X-ray.

The only practicable way we found of preparing these balloons con-
sisted in painting the outer wall of one balloon with a bismuth paste
prepared by mixing bismuth subnitrate with vaseline. This balloon
was then inclosed within another of the same size; thus the two balloons
were separated by a thin wall of bismuth paste. These balloons had
to be prepared anew for each experiment, as the peristaltic waves tend
to sweep all the bismuth to the lower end of the balloon and the vase-
line soon causes the rubber to lose its elasticity. We were able to
make graphic tracings with simultaneous direct fluoroscopic observation.

Our records show that the fundus is quiescent immediately after
eating a meal. The pressure upon the balloon is maintained at a
steady level. If a light meal is taken the tonus variations may be
demonstrated immediately after eating. At first they are so slight
as to seem insignificant but they increase in vigor and are usually visible
thirty minutes after the meal. In one experiment an unusually large
_ meal was eaten and the tonus waves were distinctly in evidence twenty

276 F. T. ROGERS AND L. L. J. HARDT

minutes later (fig. 1). They increase in intensity and may but do not
always, become more rapid. Each wave is of one to three minutes’
duration. When the stomach is nearly empty (as determined by the
stomach tube or induced vomiting) they become conspicuous and at
this stage of digestion there usually appear superposed upon them,
stronger contractions which increase in vigor and are felt by the sub-
ject as hunger pains (fig. 2). Although it is by no means always the

Ll ly
Wf Ih
mle :

if

|

Nal

Fig. 1. Tonus rhythm of fundus of stomach (man) 20 minutes after a hearty
dinner. Time in seconds.

Fig. 2. Tonus rhythm of stomach (man) three hours after dinner (beef-
steak, spaghetti, bread, butter, apples and cream).

case, it is significant that the first contractions felt may occur when
the stomach still contains traces of food.

Carlson (6) in his first paper describing the hunger contraction noted
that the onset of a hunger period was marked by the appearance of a
slow tonus rhythm which gradually increased in vigor and culminated
in the hunger contractions. As stated above this tonus rhythm is

CONTRACTIONS OF EMPTY AND FULL STOMACH 277

present not only as an immediate precursor of the hunger period but
also throughout the course of normal gastric digestion.

Many workers have employed the rubber balloon in recording the
stomach movements but none of the published tracings so far as we
are aware, indicate the presence during normal digestion of a slow
continuous rhythm as herein described.

Moritz (7) reported contractions of the fundic end of the stomach
occurring two and one-half to three and one-half times per minute.
According to Dietlan (8) the time required for the peristaltic contrac-
tions of the pyloric end of the stomach is twenty to twenty-four sec-
onds. We, therefore, think as suggested by Cannon and Carlson that
the tracings of Moritz record the intragastric variations due to the
pyloric peristalsis.

Sick (9) using the same method reported that there occur tonus
variations of the fundic end of the stomach which give rise to the
peristaltic waves. These occur at the rate of two to four per minute
in the full or empty stomach. The tracings of Sick show three kinds
of pressure variations: Respiratory, cardiac, and “‘ Magen tonus schwan-
gungen,”’ the latter averaging twenty seconds each. The duration of
these “‘stomach tonus variations’ coincide with the time intervals
required for the pyloric peristalsis.

Is the hunger contraction simply an augmented peristaltic contrac-
tion or a contraction of the fundus as a whole? To answer this we
studied the stomach movements in the dogs and in ourselves by the
“balloon X-ray’? method. We were able to make direct observations
of the movements of the balloon in the stomach and at the same time
note the character of the graphic record.

The upper part of the balloon was held in the cardiac end of the
stomach. A young and vigorous dog was employed for the experi-
ments. The dog was starved for intervals of thirty-six to forty-eight
hours and in order not to have the hunger contractions completely
inhibited by the excitement attendant upon the X-ray examination,
frequent repetitions were made until the dog became accustomed to
the necessary manipulations. The vigor of the stomach contractions
thus made visible to the naked eye is surprising. The weaker type
of hunger contractions begin as a constriction in the cardiac end of the
stomach and pass down toward the pyloric end as a rapid peristaltic
wave. In’the very vigorous contraction the wave spreads over the
stomach so rapidly that it is difficult to decide whether there is a con-
traction of the fundus as a whole or a very rapid peristalsis (fig. 3).

278 F. T. ROGERS AND L. L. J. HARDT

Such a condition may well be compared to the peristaltic rush of the
intestine as described by Meltzer and Auer (10).

When the X-ray observation of the hunger contractions were made
in ourselves, with subject in reclining position a series of photographs
were taken of the balloon at times when the record showed the stomach
to be comparatively at rest and at varying intervals during a period

Al]
q A! }
‘ned

Fig. 3. Hunger contractions of dog’s stomach after 30 hours’ starvation.
A, Outline of stomach (as seen by X-ray and bismuth balloon), between hunger
contractions with stomach relaxed. B, Outline of stomach at the height of a

hunger contraction.

of hunger contraction (figs. 4, 5, 6, 7). We made sure the balloon was
just below the cardia by pulling the distended balloon upward as
far as it would come. By careful observation we were able to tell
from the movements of the balloon when a hunger pain was being felt
by the subject. In every case the graphie record and the subject

CONTRACTIONS OF EMPTY AND FULL STOMACH 279

confirmed our own observations. When a hunger pang was felt, a
series of constrictions passed rapidly over the balloon. Beginning at
the cardiac end, they swept rapidly toward the pyloric end increasing
in strength as they proceeded. It was readily seen that the hunger
contractions are powerful peristaltic contractions, which, arising at or
near the cardiae sphincter, swept downward over the entire stomach.
During a typical hunger period the stomach exhibits movements which
resemble very closely the movements which have been described by
clinicians in patients after a bismuth meal as hyperperistalsis, but
described by Cole as normal peristalsis of a stomach that contains small
quantities of food. Observations of the hunger contractions of the
dog were made after thirty-six hours’ deprivation of food, on ourselves
after fifteen hours. Prolonged starvation in the case of the dog was
- necessary to overcome the inhibitory influence of the excitement at-
tendant upon the X-ray examination. Whether or not the stomach
of man after a longer period of starvation would show a condition ap-
proaching that described for the dog, we are not in a position to state.

The current teaching with reference to the part played by the fundus
during digestion is that it is a reservoir, exerting a tonic grasp upon its
contents. The kinematographic figures of the stomach published by
Kastle, Ridder, and Rosenthal show that the fundus is not quiescent
during digestion. Cole (12) has shown that when food is in contact
with the cardiae end of the stomach, contractions begin in the fundus
and frequently are as deep in this region as in the pylorus. Forsell
(13) describes the fundus changes as follows:

In the standing position, the fundus shows no peristalsis. In the reclining
position there occur typical persitalses of the fornix. The fornix empties itself
principally by a concentric contraction of the whole wall. At the conclusion of
gastric digestion there occur circular contractions of the wall. One may then
observe stages in which the fornix assumes a more oval form: then a circular
contraction presses the food into the korpus: then a stage of rest in which the
fornix takes a more spherical shape. The peristaltic period is begun by a trans-
verse division of the fornix which presses the fornix into a more cylindrical shape
and forces the contents into the korpus. This is repeated at intervals until the
fornix is empty.

These circular contractions of the stomach may be the cause of the
rhythmical pressure variations which we found and which we here
speak of as tonus rhythm. That there is a certain degree of independ-
ence between the slow tonus rhythm and the hunger contractions is
shown by the behavior toward inhibiting agents. Carlson (14) showed

Fig. 4. X-ray photograph (Dr. Rowntree), of the bismuth balloon in the
stomach when the graphic record shows the stomach to be quiescent.

Fig. 5. X-ray photograph (Dr. Rowntree), of the balloon in the stomach at
the height of a moderately strong hunger contraction.

280

Fig. 6. X-ray photograph (Dr. Rowntree), of the balloon in the stomach at
the height of a very strong hunger contraction.

Fig.7. X-ray photograph (Dr. Rowntree) of the balloon in the stomach near
the end of a hunger contraction.

281

282 F, T. ROGERS AND L. L. J. HARDT
that numerous substances and circumstances inhibit reflexly the hunger
contractions. This we found true in ourselves, but this tonus rhythm
does not show the same sensitiveness toward the same inhibiting influ-
ences. Thus tasting sugar or acid, swallowing water or 0.36 per cent
hydrochloric acid solution in small quantities (10—25 ec.) do not inhibit
this rhythm or only very slightly (fig. 8). This is in contrast to the
immediate effect of the same substances on the hunger contractions.

VT
i ih Kilt

un"

¢

\

vi ih ‘a fy an Nyt Ni HH

ua

ul ii

i nH Ai i (hi
wt iN Mi HAN nisin i thi i ry W MN
| a i thi Ne HM ie i

}

Fig. 8. Tonus rhythm of stomach (man) during gastric digestion is not
inhibited to the same degree by substances which completely inhibit the hunger
contractions. A, at x 25 cc. of 0.36 per cent HCl was swallowed. B, at x S0ece.
of cold water swallowed. C, Tonus rhythm just before a hunger period; atz
50 ec. of cold water was drunk.

Large quantities of liquids do inhibit the tonus rhythm probably as
a result of the distension which a sudden increase in volume of gastric¢
contents would necessitate. A quantity of liquid which does not
inhibit the rhythm in a filled stomach may do so in an empty stomach.
Excitement, interest, worry or any strong psychic influences immedi-
ately inhibit the rhythm. The twenty-second rhythm is not so easily
inhibited as either the hunger contractions or the slower tonus rhythm.

CONTRACTIONS OF EMPTY AND FULL STOMACH 283

- Our findings interpreted in the light of the contributions of Cannon,
Carlson and Forsell show that the normal stomach exhibits the follow-
ing types of muscular activities:

1. A tonic grasp of the upper stomach musculature upon the food.
This tonic condition exhibits slow rhythmical variations.

- 2. Peristaltic contractions of the antrum pylori.

3. Peristaltic contractions of the entire stomach (the hunger contrac-
tions of Boldyreff, Cannon and Washburne, and the thirty-second
rhythm of Carlson).

4. In the young dog the vigorous hunger contraction begins with a
constriction at the cardiac end and is followed by a contraction of the
entire fundus. (If this type of contraction. be peristaltic the waves
are too rapid to be detected by our methods.) :

During normal digestion the peristaltic waves sweep over the lower
part of the stomach. The fundus exhibits a rhythm of the same rate
as was shown by the work of Carlson and Orr. In the meantime there
are slow rhythmical tonus variations of the upper part of the stomach.
As the stomach empties itself the peristaltic waves arise from. points
higher and higher toward the cardiac end of the stomach run over the
entire stomach, culminating in a more or less tetanic. contractions of
the antrum. The hunger contractions are therefore strong peristaltic
contractions in a stomach whose condition compared with its normally
filled condition is hypertonic. Evidently these two phases of stomach
contractions are most intimately related. The tonus may become so
great as to run into tetanus or it may maintain a level on which the
separate peristalses are superposed. These two different activities are
not necessarily similarly affected by the same inhibiting influences.
Both may be inhibited by introducing food or liquid into the mouth
or stomach or by psychic influences but not to the same degree.

CONCLUSION

1. Both in man and in the dog, the fundic end of the stomach, dur-
ing normal digestion, exhibits a slow tonus rhythm of one to three
minutes’ duration. During the emptying of the stomach and the onset
of the hunger period, this tonus rhythm becomes more vigorous.

2. In the dog after thirty-six hours’ starvation the stomach is mark-
edly hypertonic; a vigorous hunger contraction is a contraction of the
fundus as a whole.

3. In man after fifteen hours’ starvation the hunger contractions

284 F. T. ROGERS AND L. L. J. HARDT

are vigorous peristaltic contractions in groups of 2-4, superposed upon
the slow tonus rhythm. They begin at the cardiac end of the stomach
and sweep over the whole stomach.

4. A given inhibiting agent does not necessarily have the same effect
upon the different types of gastric contractions; thus the pyloric peri-
stalsis is the most resistant to reflex inhibiting influences; fundic peri-
stalsis is least resistant, while the tonus rhythm occupies a position
between the two.

We wish to express our thanks to Dr. Rowntree of the Presbyterian
Hospital under whose supervision the X-ray work on ourselves was
carried out, and to Dr. A. J. Carlson for his help and criticism.

BIBLIOGRAPHY

(1) Boupyrerr: Arch. d. Sci. Biol., 1905, xi, 1.

(2) CANNON AND WASHBURNE: This Journal, 1912, xxix, 441.

(3) Carztson: This Journal, 1914, xxxiil, 95.

(4) Carson, OrR AND McGratu: This Journal, 1914, xxxiii, 119.

(5) Rogers: This Journal, 1915, xxxvi, 183.

(6) Caruson:. This Journal, 1912, xxxi, 151.

(7) Moritz: Zeit. f. Biol., 1895, xxxii, 356.

(8) Diertan: Ergebn. d. Physiol., 1913, 87.

(9) Sick: Deutsch. Arch. f. Klin. Med., 1906, Ixxxviii, 190.

(10) Merurzer anp AugR: This Journal, 1907, xx, 259.

(11) Karstite, RrepER AND RosentTHAL: Miinch. Med. Wochenschr., 1909,
No. 6.

(12) Coxe: Arch. of Roentgen Ray, 1911, 242.

(13) Forsetu: Miinch. Med. Wochenschr., 1912, No. 29.

(14) Caruson: This Journal, 1913, xxxi, 212.

A CONTRIBUTION TO THE PHYSIOLOGY OF LACTATION

W. L. GAINES

From the Hull Physiological Laboratory, University of Chicago

Received for publication June 29, 1915

FOREWORD

The development and functioning of the mammary gland in the
female mammal is closely associated with, and dependent upon the

reproductive function.!

The normal course of activity of the gland is

somewhat as follows.

Embryonic—

Puberty

|

Pregnancy

|

Delivery

|

Pregnancy—

Delivery—

The first sign of the mammary organs, the milk line,
appears at an early stage in embryonic life, and
growth is continued until the gland reaches a con-
siderable development at

At birth a milky fluid may be expressed from the
gland in either sex, indicating a secretory activity,
but this activity soon ceases and the gland remains
quiescent until

At puberty under the influence of the ovaries growth
is resumed and carried to a greater or less develop-
ment, varying with the individual and the occurrence
of

Pregnancy induces a great hypertrophy of the gland,
and it reaches a high state of development, accom-
panied by the accumulation of colostrum, by the
time of

At this time milk secretion proceeds actively and if
the milk is removed at short intervals, by nursing or
artificially, secretion continues for a period of days,
months, or years. The rate of secretion, however,
after a time gradually decreases to zero. This
decrease is favored by a succeeding

Further growth of the gland ensues but not as
marked as in the preceding gestation, and at
milk secretion is again actively resumed.
the pregnancy-delivery cycle is repeated.

Whence,

1The mammary gland may, in turn, influence the reproductive and other
body functions, although such influence is, apparently, not so fundamental.

285

286 , W. L. GAINES

This paper presents some data and discussion based on an investi-
gation of certain phases of the above outlined course of activity of the
mammary gland carried out at the Hull Physiological Laboratory of
the University of Chicago. I wish to acknowledge especially the
assistance and suggestions of Dr. Carlson throughout the work. I am
also indebted to Dr. Mathews and Dr. Koch, among others, for various
suggestions.

EXPERIMENTAL METHODS AND RESULTS

Prenatal development. The milk lines may be distinguished as
early as the 4 mm. stage in the human embryo, according to Strahl (1)
Knoepfelmacher (2) states that in children born prematurely there is
no appearance of milk in the mammary glands. In a post mortem
examination of two kids, the mother of which died about two weeks
before term, I found mammary tissue plainly developed but no ap-
pearance of milk in the transected gland. Chemical examination
for the presence of sugar was not made and this test might have shown
the presence of milk in amount too small to be distinguished by simple
cutting of the gland. The udder of the mother, a multipara, was at
the time undergoing a rapid hypertrophy as evidenced by increase in
size. (Data on this point will be found in table 1, Goat No. 4). A
part of the increase in size was due to the accumulation of colostrum,
which exuded freely from the cut ducts of the gland after death.

The influence of pregnancy on the growth of the mammary gland. The
growth of the mammary gland during gestation is a very striking
phenomenon. This development is in preparation for its function of
milk secretion following delivery. Ribbert (3) transplanted the mam-
mary gland of a guinea pig into the skin back of the ear. Five
months later (the gestation period is about two months) living young
were born, whereupon the transplanted gland functioned promptly.
Ribbert states that hypertrophy was not sufficiently marked to make
the gland apparent beneath the skin, that the nipple did not develop,
and there was, therefore, no outlet for the secretion, but that milk was
present in the gland. Marshall (4) cites the case of the Bohemian
pygopagous twins, Rosa-Josepha, one of whom became pregnant.
The breasts of both underwent development during pregnancy
and functioned following delivery. Lane-Claypon and Starling (5)
produced a development of the mammary glands in a virgin rabbit
closely resembling that occurring during pregnancy by means of in-
jecting extracts of rabbit foetus.

PHYSIOLOGY OF LACTATION 287

On the other hand, Lombroso and Bolaffio (6) using parabiotic
methods with rats found no evidence that the pregnancy of one of a
pair of females produced any effect on the mammary glands of the
other one of the pair. Knoepfelmacher (2) injected blood serum from
a pregnant goat into a non-lactating goat. The amounts used were
55, 33, and 90 cc. taken respectively 6 days, 2 days, preceding delivery,
and 1 day following delivery. The results were negative.

I have injected defibrinated blood from a pregnant non-lactating,
multiparous goat into a non-pregnant, non-lactating, multiparous goat
with negative results with reference to increase in the volume of the
mammary glands or milk secretion. The volume of the udder was
determined by pressing a suitable vessel containing warm water up

TABLE 1

Showing the effect on the volume of the udder of a non-pregnant goat jrom the intra-
venous and intraperitoneal injection of defibrinated blood froma pregnant goat
td

GOAT NO. 4—DONOR (PREGNANT, GOAT NO. 3—RECIPIENT (NON-PREGNANT, NON-LACTATING,
NON-LACTATING, MULTIPAROUS) MULTIPAROUS)
Date see of Melee o ai Method of injection
1915 ce cc ce
1-21 935 170 90 Intraperitoneal
1-23 935 165 115 Intraperitoneal
1-25 950 175 100 Intraperitoneal
1-27 980 175 75 Intraperitoneal
1-30 1005 160 15 Intravenous
2-3 1060 170 85 Intraperitoneal
2-6 1090 170 35 Intraperitoneal

—~—

around the udder and firmly against the body wall. The amount of
water displaced (determined by the difference between the amount
left in the vessel and its capacity) was taken to represent the volume of
the glands. The method is not exact but is sufficiently accurate to
show any marked changes in volume. The data are given in table 1.
No appearance of milk resulted during the period of injections or fol-
lowing.

The figures in table 1 show that the udder of the donor was increas-
ing steadily in size but the amount of blood injected failed to affect the
size of the udder of the recipient. The latter at a previous time during
a similar stage of pregnancy as represented by the donor in the present
instance had fully as large an udder as the donor. The figures, then

288 W. L. GAINES

serve to show the enormous difference in size of the udder with the

stage of pregnancy. .
The influence of pregnancy on milk secretion. Lane-Claypon and

Starling (5) advance the theory in connection with the work referred

TABLE 2

Daily milk and fat yield oj goat No. 2, showing the effect of intravenous injections
oj defibrinated blood from a pregnant goat, No. 3; from her two days old kid;
and again from No. 3 six days after delivery. No. 2 said to have dropped her
kids about April 12. No. 3 kidded August 8, 1914

DATE MILK FAT FAT REMARKS

1914 cc. |percent| gms.
7-24 | 823 5,1 | 41.9

7-25 | 796] 6.4] 51.0
7-26 | 730] 6.4] 46.8
7-27 | 645 | 5.4} 35.0
7-28 | 720| 6.4] 46.3|f Injected intravenously 175 cc. defibrinated blood
7-29 | 463 | 69) 31.9]\ from No. 3 (10 days before term)
7-30 | 708 | 6.8 | 48.2
7-31 674 | 6.2 [ 41.7
8-1 783 | 6.8 | 53.7 |f Injected intravenously 300 cc. defibrinated blood
8-2 | 652| 7.0| 38.6|\ from No. 3 (7 days before term)
8-3 783 | 7.0 | 54.8
8-4 847 | 7.8} 66.1 Injected intravenously 175 cc. defibrinated blood
8-5 913 | 7.5 | 68.4 from No. 3 (3 days before term). Caused some
8-6 349 | 4.9 | 17.1 general depression, weather hot, flies trouble-
8-7 633 | 8.0] 50.8 some
8-8 832 6.2) | o136
8-9 (HPS || ieay || Gala
8-10} 665| 5.9 | 39.3|{ Injected intravenously 100 ce. defibrinated blood
8-11 | 538 | 6.6 | 35.4 il from male kid of No. 3. Age of kid 2 days
8-12 | 806| 6.8 | 54.7
8-13 | 873 | 6.0 | 52.4
8-14} 853] 6.5 | 55.0|f{ Injected intravenously 100 cc. defibrinated blood
8-15 | 765 | 7.0| 63.7/\ from No. 3 (6 days after delivery)
8-16 | 880| 6.3 | 55.4
8-17 | 823 | 7.3 | 59.6
6.2 | 52.3

8-18 | 848

to above that the developing foetus passes a substance into the maternal
circulation which favors the growth of the mammary gland but in-
hibits its secretory activity. The removal of this inhibitor, then, at
delivery allows the gland to function in its milk secretory capacity.
Bearing on this point I have made the following series of injections

PHYSIOLOGY OF LACTATION 289

of defibrinated blood into a lactating goat: 1, from a pregnant goat
near term; 2, from a kid of this same goat 2 days after birth; 3, from
the same goat 6 days after delivery. Data of this test are given in
table 2.

The figures indicate an inhibitory action from each of the trans-
fusions. The milk and fat data as given for each day are the sum of
two milkings made throughout at approximately 12-hour intervals.
In every case the inhibitory influence was most marked at the first
milking following the transfusion. Except in the case of the trans-
fusion made August 5 there was no apparent systemic disturbance

TABLE 3

Daily milk and fat yield of goat No. 6, showing the effect of intraperitoneal injection
of a water extract of dried, fat-freed cow’s placenta; also, defibrinated blood from
goat No.5. No.6 said to have kidded March 12,1915. No.édroppedkids April
17, 1915, and at the time of transfusion, April 24, was milking around 2000 ce.
per day

DATE MILK FAT FAT REMARKS

1915 ce. percent| gms.

418 | 475 | 6.8 | 32.3

4-19} 488] 7.3 | 35.5

420| 460| 6.4| 29.3|f Injected intraperitoneally a water extract of 5 gms.

4-21| 168) 9.4)15.8/\ of dried, fat-freed cow’s placenta

422} 2388] 8.5] 20.4

ee a a ae | Injected intraperitoneally 200 cc. defibrinated
: blood from No. 5 (fresh 7 days and giving about

4-25 | 182] 8.6 | 15.7 [.:soomtne: daily)

4-26 | 147| 7.1.) 10.5 a re

4-27 | 111} 9.4| 10.4

4-28 | 183 | 7.8 | 14.3

4-29} 154) 8.3] 12.8

in the condition of the recipient. Leaving this case out of consideration
the inhibitory action seems to be equally marked for either the blood
of the pregnant goat or that of her two-day old kid.

The transfusion made 6 days after delivery had onlya very slight
inhibitory action as seen in the yield of milk and was not at all reflected
in the yield of fat. Results in another trial with two other goats were
similar to the above except in the transfusion following delivery (see
table 3).

Effect of placental extract on milk secretion. Niklas (7) found that
an extract of placenta caused milk production in virgins and mothers.

290 W. L. GAINES

Lederer and Pribram (8) found greatly increased secretion of milk upon
intravenous injection of placental extract.

Table 3 presents the data of a test of the effect of an extract of cow’s
placenta on milk secretion in the goat. The fresh placental tissue was
ground, dried, extracted with gasoline and ether to rid of fat, and the
residue extracted with water as needed. The figures show a distinct
inhibition, quite similar in nature to that just observed from the blood
of a pregnant goat or new-born kid (see table 2).

Relation of the blood to milk secretion. The activity of the mammary
gland with reference to milk secretion varies enormously. A large
part of the time it may be at zero, but immediately following a normal
pregnancy and delivery it is at a high point. To test for the presence
of a substance in the circulating blood stream which accelerates the
secretory function of the mammary gland during the period of its
highest activity following delivery, transfusions were made from a
fresh, heavy milking goat to one giving a low yield. No increase in
secretion was obtained by this means. In table 2 no change, or a very
slight decrease was shown. In table 3 there is shown a very marked
decrease. In this latter case the donor was milking heavily at the rate
of about 2000 ce. daily, including the milk taken by two nursing kids.
The recipient although giving a relatively low yield, about 300 ce.
daily, had been fresh only about 6 weeks and would seem to have offered
a favorable subject for positive results. No general disturbance was
apparent to account for the decrease noted.

The effect of mammary gland extract on milk secretion. Gavin (9)
fed mammary gland preparations to cows without affecting the yield
or quality of the milk. The mammary gland of the same cow whose
placenta was used in the data shown in table, 3, was prepared in the
same way as given for the placenta. Results from the injection of
water extracts of this material are given in table 4.

The figures appear to show a temporary inhibition followed by an
increase to something above the previous yield. The net result of the
several injections seems to be a considerable increase, but it may be
that this should not be attributed to the injected material. Reference
to table 3 shows that goat No. 6 was, a few days before, giving more
than the yields recorded in table 4, and the apparent increase may be
nothing more than recovery following the blood transfusion of April 24.
However, the immediate inhibition, similar to that noted for the pla-
cental extract, seems to be unequivocal.

PHYSIOLOGY OF LACTATION 291

The action of pituitrin on the mammary gland. Since Ott and Scott
(10) first noted that pituitrin hada galactagogue action this material
has received considerable attention in its relation to milk secretion.
All reports of investigators on the immediate action of pituitrin confirm
the observation of Ott and Scott, viz., that its injection into the cir-
culation causes an immediate flow of milk in a lactating animal. «-Mac-
Kenzie (11) working with cats in anesthesia and observing the flow of
milk from the cut surface of the gland found pituitrin the most active
of the several animal extracts he studied. Gavin (9) working with

TABLE 4

Daily milk and fat yield of goat No. 6, showing the effect of intraperitoneal injection
of a water extract of dried, fat-freed cow’s mammary gland (pregnant). No. 6
said to have dropped kids March 12, 1916

DATE MILK FAT FAT REMARKS

1915 ce. percent} gms.

5-3 113 | 8.3 9.4
5-4 Sa Ne627 | Gd
5-5 ig. | 69
5-6 119| 7.6| 9.0|f Injected intraperitoneally water extract of 5 gms.
5-7 75 | 65) 4.9/\ cow’s mammary gland
5-8 135 | 6.2] 8.4
5-9 i435 | 6:6) 9.5
5-10] 176| 6.4 | 11.2|f Injected intraperitoneally water extract of 1 gm.
poviaie 167.|° 7.1. | 11.1 } cow’s mammary gland
o-12 | 190} 6.9} 13.1 Injected intraperitoneally water extract of 4 gms.
5-13 | 134| 7.8/10.4|\ cow’s mammary gland
5-14 | 189] 7.4 | 13.9
5-15 | 220| 7.5| 16.4|f Injected intraperitoneally water extract of 5 gms.
Donets | vit) | Lost \ cow’s mammary gland
7.3 | 14.3

o-l7 | 194

cows yielding 15-35 pounds of milk daily found no effect on quantity
or quality of milk as the average result of 3-5 days treatment with
pituitrin. In some cases, he reports, there was a distension of teat and
cistern with milk immediately following injection. Hill and Simpson
(12) report similar results. Hammond (13) has studied in some
detail the effect of pituitrin on the composition of milk secreted under
its influence. He finds that the fat content is greatly increased and that
the protein, sugar and ash content remain very constant. This refers
to the milk secured immediately after the injection of pituitrin and is
not in conflict with Gavin’s results which referred to longer periods;

292 W. L. GAINES

for the yield of milk and its fat content are later depressed. Hammond
concludes from his data that the action of pituitrin is not muscular.
Hill and Simpson (14) confirm Hammond’s results as to the composi-
tion of milk obtained after the use of pituitrin except they did not find
the depression in fat content later, as noted by Hammond.

Heaney (15) found in the human a contraction of the breast upon
injection of pituitrin followed by a return to initial volume, and at-
tributes the milk secretory action of pituitrin to its muscular effect.

Goat No A.
Milk Yiela

Ny
(AUT ASIP AIP ABP AAP ARP ABS Anat

Fig. 1. Showing fluctuation in yield of milk by a goat milked at regular 12-
hour intervals and the dependence of the flow of milk produced by the injection
of pituitrin upon this fluctuation. The heavy vertical lines represent the yield
of milk obtained upon intravenous injection of 1 cc. of pituitrin immediately
after the regular milking. It is inferred, for example, that pituitrin would
have caused a large further flow at X; while at Y little or no extra milk would
have been obtained. A = morning milking; P = evening milking.

As to the quantity of milk and fat yielded as a result of the tise of
pituitrin (Parke & Davis preparation was used throughout the work
here reported on pituitrin) based on the average of two days and not
the immediate yield, my data show no effect either way. My results,
based on three goats, confirm those of Hammond who found a decrease
in fat content of the milk for a day or two following the use of pituitrin.
I have found no marked or consistent variations in the sugar or pro-
tein content of the milk.

PHYSIOLOGY OF LACTATION 293

The flow of milk from an active mammary gland, which has not
been drained for a few hours, upon injection of pituitrin is a very strik-
ing reaction. In the goat I have never failed to secure some flow of
milk from its use even immediately after milking “dry.’”’ The amount
obtained, however, varies considerably.

Figure 1 illustrates this point. The goat was milked regularly
at 12-hour intervals but the yield is seen to fluctuate widely. The
intravenous injection of pituitrin immediately following a regular
milking produces a further flow varying in amount inversely with the
yield secured by the hand milking which preceded it. That is, in-
jection of pituitrin following a relatively high yield of milk produces a
small flow; and following a relatively low yield it produces a large flow.
A second, third or fourth dose produces only a very slight flow, 2-4 ee.
in the goat.

I have followed the yield of milk in the nursing dog very satisfactorily
by keeping the pups separated from the mother and allowing them to
nurse at 8-hour intervals. The pups were balanced on a smooth-
working balance, with shot, then allowed to nurse, and again balanced
with the standard weights. The increase in weight represents approxi-
mately the yield of milk by the mother.

Under the above conditions, after the pups have nursed fully (6
or 7 minutes), I have never succeeded in producing any further
flow of milk by the intravenous injection of pituitrin. But if the
mother be placed under ether at the regular nursing hour and the pups
then allowed to nurse the yield of milk is greatly depressed, although
the pups do their part of the act in the usual vigorous manner. Often,
in fact usually, the yield is so low as not to be detectable by the method
used (that is, it appears, well under 5 gms., considering errors in loss of
saliva, ete.). If, now, pituitrin be injected intravenously, with the
- mother still under ether, the pups immediately secure the normal yield,
or the balance of it which they had failed to get in the first nursing, and
this, in rather less time than is required in nursing under normal con-
ditions. A second dose following the first in 10 minutes, produces no
further flow.

Figure 2 shows an illustration of the effect of pituitrin under normal
conditions and also with the mother under ether, but in the latter
case some milk yielded to the pups nevertheless. Ordinarily the ether
curve would hold to the base line until the pituitrin was administered.

While pituitrin causes no further flow of milk in the dog after nurs-
ing, this does not seem to hold true for the goat. Data from two

294 W. L. GAINES

experiments are given in table 5 which show that after the kid had
nursed and secured all the milk it could, injection of pituitrin still caused
a further marked flow. The high fat content (15.8 to 19.6 per cent)
of the milk thus obtained is also remarkable.

Bearing on the nature of the mechanism involved in the flow of milk
produced by pituitrin under certain conditions, the following test was
carried out. A cannula was inserted in the teat of a goat and con-
nected with a chloroform manometer recording on the revolving drum
of a kymograph. The gland was then inflated with air through a T-

Ga i a RT
00 ES

(ta: aes SSssei
mA isl a
we cals Aca) sl). S04 [eee |
ole lila a] net [eg] NE ao
50 “ lsat a Fiow of Mitk inthe Dog :

=

tee a4 ure

Hp nject Wace Pituitrin at
~~ | /

Ad (

me FAA Pinel: Hint
Aza de 7 fiat aaa eid Pala a.

BF

a

Fig. 2. Showing failure of pituitrin to cause a flow of milk following normal
nursing in the dog; and, its restoral of the flow normal to the stimulus of nursing
when the normal flow is inhibited by anesthesia; and, the failure of a second dose
to cause any further flow. Also, showing under the stimulus of normal nursing
the latent period in the flow of milk during the first minute, followed by the very
rapid flow for the next two minutes, and then the gradual decline to zero (ef.
figs. 4-8).

connection to a pressure of 8-10 cm. chloroform. The device may be
made quite sensitive in recording any change in pressure within the
gland. . With the apparatus properly adjusted pituitrin was injected
intravenously. The effect varies with the stage of activity of the
gland. In the non-pregnant, non-lactating goat there is no apparent
effect. This is true, also, in the pregnant, non-lactating goat up to a
time close to term. <A short time before term the result, following a
latent period of 15-20 seconds, is a slight increase in pressure with a
return to the initial. In the freely lactating goat, after a latent period

PHYSIOLOGY OF LACTATION 295

of 15-20 seconds, there is an abrupt rise in pressure followed by a
gradual and nearly uniform decline to the initial pressure. The decline
occupies about 5 minutes. A series of 3 curves, taken from the same
goat before and after delivery, is given in figure 3.

A second dose of pituitrin in the freely lactating goat gives a curve
similar to the first dose, except that it is much reduced in the extent
of the rise. In the case of a goat milking from one gland and dry in
the other, each gland being separately connected as above, pituitrin

TABLE 5
Showing the effect in the goat of nursing following milking ‘“‘dry’’ by hand; and of
pituitrin following nursing. After milking by hand, kid nursed left gland and
at the same time right gland was milked by hand. Following the nursing 1 cc.

pituitrin was injected intravenously and both glands milked by hand. Goat
No. 5, kidded April 17, 19165

RIGHT GLAND LEFT GLAND BOTH GLANDS

EXPLANATION

CC: percent) gms. ce. percent| gms. cc. per cent| gms.
4-22-15
Milked ‘‘dry” by
i eae 358 | 3.5 |°12.5
Nursed by kid...... 153 } 343
Milked by hand.....| 190} 9.8 | 18.6
Injected pituitrin—
milked by hand...| 70 | 15.8 | 11.1 98 | 17.1 | 16.8 | 168 | 16.6 | 27.9
5-2-15
Milked ‘‘dry’’ by
0 Dee 315 | 4.7 | 14.8
Nursed by kid...... 175 \ 330
Milked by hand.....| 155 | 13.3 | 20.6 f
Injected pituitrin—
milked by hand...| 45] 19.1 | 8.6 40/19.6| 7.8| 85 | 19.3 | 16.4

produced the typical curve in each gland, that is, the abrupt rise and
slow decline in the active gland, and no change in the non-lactating
gland.

While the above test indicates a contraction of the ducts and alveoli
of the gland with a subsequent relaxation, it does not definitely localize
the seat of the contraction. That is, the effect might be produced by
the activity of the muscles of the large ducts. Indirect evidence on
the point is afforded by the following test.

296 W. L. GAINES

A lactating guinea pig was separated from her nurslings for two
hours to allow the accumulation of milk in the glands. The animal
was then anesthetised and a mammary gland dissected entirely away
from the body with as little injury to the gland as possible. The end
of the nipple was cut off to allow free egress to any milk. A few drops
of pituitrin were now injected into the gland substance, whereupon
there was a copious flow of milk from the cut nipple. A control in-
jection of salt solu-
tion gave no such ef-
fect.

A similar gland
preparation was cut
across with a razor,
exposing a cut sur-
face as remote as
possible from the
large ducts and ves-
sels. A drop of pitu-
itrin was placed on
the cut surface and

Fig. 3. Showing the effect of pituitrin on the pres-

sure of an air-inflated mammary gland of agoat. The l 4 few seconds
teat was connected with a cannula and tube to a Many minute white
chloroform manometer and the gland inflated to a dots appeared be-

pressure of 8-10 em. At X # ce. of pituitrin was in-
jected intravenously. The time line marks 10-second
intervals. The upper tracing was taken 37 days be-
fore delivery; the middle tracing, 17 days before de-

neath the pituitrin
and slowly swelled
to tiny milky rivu-

livery; the lower tracing, 12 days after delivery (the
daily milk yield at that time being around 1800 cc.).
The shallow waves, seen especially in the upper trac-
ing, are due to respiratory movements. The deep,
sharp irregularities are due to bleating and struggling
of the animal.

lets streaming beau-

tifully through the
clear liquid. Similar

preparations from the
cat and dog gave like

results.

The reflex involved in milking and nursing. According to Flower and
Lydekker (16) the mother whale ejects the milk from her mammary
glands into the mouth of the young without the active sucking on the
part of the young which is generally common to mammals. Sehifer (17)
states that the discharge of milk during the act of sucking or milking
is in part the result of direct mechanical pressure upon the milk reser-
voirs of the larger ducts; and partly due to a contraction of the plain
muscular tissue which accompanies these ducts, and which appears to

PHYSIOLOGY OF LACTATION 297

be set in action by mechanical stimulation of the nipple. He adds
that the flow is probably aided by the swelling of the gland due to a
reflex dilatation of its arterioles.

The curve of normal nursing in the dog given in figure 2 shows
roughly the rate at which the pups secured milk during the process
of nursing. The data on which the curve is based were obtained by
working the pups in two relays of three pups each. Group I was allowed
to nurse for 1 minute, then removed and Group II given way for 1 min-
ute, Group I being meanwhile weighed. Places were then again ex-
changed and so on until they had obtained all the milk possible.

The normal curve shows that the pups could secure only a small
amount of milk the first minute, but during the second and third
minutes there was a relatively large amount of milk available to their

Mid Nursing

Fig. 4. Showing the pressure changes produced in one gland under the influence
of kid nursing the other gland. Goat No. 5, fresh 10 days, milked 3 hours pre-
viously. Left gland connected with chloroform manometer (no inflation of gland).
Kid nursing right gland. The sharp irregularities in the curve are due to bunting
of the kid. Time line marks 10-second intervals and represents 7 cm. pressure
(ef. figs. 5, 6, 7, 8).

efforts. After the third minute the amount available declines and at
the end of 6 minutes the total supply appears to be exhausted. The
very marked changes in the amount of milk secured with the progress
of nursing are not attributable to any differences in the efforts of the
pups themselves to secure their food.

In table 5 it is shown that, after milking the mother goat as completely
as possible by hand, if a kid be allowed to nurse one gland he not only
obtains a considerable quantity of milk but there is also a practically
equal flow in the other gland which may now be readily obtained by
hand milking. The full significance of this reaction with reference to
the operation of a reflex contraction and ejection of milk from the gland
is perhaps better given by the records of the following tests.

I. Change in pressure. One gland was connected by a cannula and
tube to a chloroform manometer recording on a kymograph. A kid

298 W. L. GAINES

was then allowed to nurse the other gland.. An example of the effect
is given in figure 4. The pressure registered upon insertion of the
cannula was about 9cm. During the period when the kid was nursing
the curve shows a latent period of 60 seconds with no change in pressure
aside from the purely mechanical effect of the impatient bunting of the
kid. At the end of this latent period there appears a marked rise in
pressure. That, coincident in time with this rise in pressure, the kid
was obtaining a satisfactory supply of milk, was plainly evident from
the cessation of his bunting (shown, as well, in the smoothness of the
curve), his rapid guzzling, and the rounding out of his little paunch.
Two minutes later a second kid was allowed to nurse and a rise in pres-
sure occurred; again only after a latent period of about 1 minute.

IT. Flow of milk. A cannula was inserted in one teat and the milk
flowing through it led by a short tube to one limb of a U-tube of large
bore. The U-tube was held in a vertical position and its other limb
was provided with a float and rod recording on a kymograph after the
usual fashion of a manometer. Enough water was put in the U-tube
to allow the float to ride freely, so that the addition of any further
liquid would be properly registered. The connection of the tube from
the teat with the U-tube was open so that the liquid in both limbs was
constantly under atmospheric pressure only. The outlet for the milk
remained constantly about 10 cm. below the level of the teat. With the
apparatus adjusted a kid was allowed to nurse the free teat. Examples
of the results are given in figures 5-7.

In the experiment represented by figure 5 the right gland was con-
nected as described above and the left gland milked by hand. After
the hand milking was completed a kid was allowed to nurse the left
teat. During the first 35 seconds of the kid’s nursing the curve shows
a flow amounting to a few drops only from the right gland. At the
end of a 35-second latent period it shows a rush of milk reaching a maxi-
mum in 20 seconds and being practically complete in 80 seconds. The
form of this curve is strikingly similar to that of the curve of flow
in normal nursing of the dog shown in figure 2.

The high fat content (13.3 per cent) of the milk obtained under the
stimulus of the nursing of the kid is noteworthy and significant. The
amount of fat contained in the second portion was 50 per cent greater
than that obtained under the stimulus of the first milking operation.

In the trial shown in figure 6 preliminary hand milking was not used.
Fifteen ec. of milk flowed upon insertion of the cannula in the course
of 3 minutes. Upon nursing of the opposite gland by the kid there is

PHYSIOLOGY OF LACTATION 299

/0 sec.

tka NuUTsina- ----------~

Fig. 5. Showing flow of milk through cannula from one gland under the stimulus
of kid nursing the other gland. Goat No. 5, fresh 24 days, milked 11 hours pre-
viously. Cannula inserted in right teat and led to instrument recording the flow
of milk. 84 cc. of milk flowed immediately on insertion of cannula. Left gland
then milked as dry as possible by hand, yielding 136 cc. 45 seconds after the
commencement of hand milking a further flow started from the right gland,
yielding in 2 min. 68 ce. (results to this point not shown in the figure). Kid
then nursed left gland causing a rapid flow from the right gland, after a latent
period of 35 seconds as shown (cf. figs. 2, 4, 6, 7, 8).

Milk, Fat, Fat,

ce. per cent gms.

J Left gland, hand milked................... Wen 196 5.6 7.6

i Righieeiands cannulaflow:...¢.-...¢.0.<0a0.0: los 53 8.1
ioeti gland, kid’ nursed....u..2..0..¢0.5.. e000. 005 ye 90

\ Rienieland, cannula flowis. 2.02... 44.0.8. ose 95 oro 12.6

Se a A A ee
; SHOE CT 8009 SAN cpt aes

Occ 10 3€0.

Fig. 6. Showing same as figure5. Goat No. 5, fresh 11 days, milked 33 hours
previously. Left gland connected with flow recorder, kid nursed right gland
(no hand milking). Initial flow on insertion of cannula, 15 ec. Flow under
stimulus of kid nursing other gland, after latent period of 60 seconds, 100 ce.

Total flow 115 cc., 7.1 per cent fat (cf. figs. 2, 4, 5, 7, 8).

300 W. L. GAINES

shown a latent period of 60 seconds and then a rapid flow from the
unnursed gland. A second marked increase in flow is shown to occur
2 minutes following the first ejection. The curve shows, in compari-
son with figure 5, first,a longer latent period, and second, the occurrence
of a second ejection of milk. In some cases the curve takes the form of
4 or 5 distinct steps.

The conditions of the test shown in figure 7 are practically identical
with those of figure 6. The curve is given as showing another type of
reaction, viz., a protracted and more or less uniform flow of milk
(although a tendency to steps in the curve is shown here, also) under
the stimulus of the kid’s nursing. That the kid was also experiencing
delay in the speed with which he was served his meal is evidenced by
the length of time he continued to nurse.

Fig. 7. Showing same as figure 5. Goat No. 5, fresh 16 days, milked 4 hours
previously. Left gland connected with flow recorder. Kid nursed right gland
(no hand milking). Initial flow on insertion of cannula, 7 ee. Flow under
stimulus of kid nursing other gland, after latent period of 60 seconds, 44 ce.
(ch. figs. 254 5.65.8).

Figure 8 brings out two points: (1) The curve of milk flow with the
simple insertion of a cannula and without any other manipulation may
take a distinct step form, as shown in the left hand part of the figure.
(2) Hand milking of one gland produces a flow in the opposite gland,
similar to the effect noted for the kid’s nursing. (Other trials gave
higher and steeper curves, very like that of figure 6.) The latent period
is distinct, being 55 seconds in the present instance.

The question arises as to the nature of the factor or factors involved
that cause the further ejection of milk upon nursing of the kid following
hand milking, and which it seems to be impossible to call into play
by usual methods of milking. I have tried various methods of electri-
eal stimulation without success. I have imitated the nursing action

PHYSIOLOGY OF LACTATION 301

of the kid as actively as I could with my own mouth for several minutes
at a time without result in the way ofa further milk flow. The presence
of the kid, nosing about the udder but not allowed to take a teat into
his mouth, while using the greatest possible persuasiveness in the art of
dry-hand milking was not sufficient. Neither did it suffice to add to
the last mentioned recourse conditions of moisture and temperature
by immersing the teat in water at 38-40°C. Only by allowing the kid
to nose about the udder, while applying my own mouth to a teat in

Fig. 8. Showing flow of milk through cannula from one gland under stimulus
of hand milking of other gland. Goat No. 5, fresh 23 days, milked 12 hours
previously. Left gland connected with flow recorder. Initial flow on insertion
of cannula, 27 cc., followed by second flow of 44cc. Flow under influence of hand
milking other gland, after latent period of 55 seconds, 33 ce. Total flow 104 cc.
Kid then nursed right gland obtaining 140 cc. and causing a further flow from
left gland of 124 ec. (not shown in the figure), (cf. figs. 2, 4, 5, 6, 7).

Milk Fat Fat

ce. per cent gms.

I pee Bland whandsmilked. «4: ascineteerera eer a nee 136 3.6 4.9

Seman eCANAMLs NOW... 5-62 meee ses 2 ee see 104 3.3 3.4
Pare inser lane KT OeNULSeds... fos Pe te ls ee. ae 140 :

Perr igsi cannula flow. 2. :.:. = sees eee es ee os 5 as 124 9.3 11.5

imitation of his sucking, have 1 been able to bring about the desired
reaction without the actual nursing of the kid. The following protocol
illustrates.

302 W. L. GAINES

Goat No. §. 5/16/15, 7.15-7.45 p.m.

Milk Fat Fat

ce. per cent gms.

Milked by hand in usual manner (kid not around)...... 285 7.9 2205

Sucked teat with mouth, then milked by hand........... 00 00 00
Kid introduced and attempting to nurse but not al-

lowed to take teat in mouth, then milked by hand..... 00 00 00

Sucked teat with mouth (no milk drawn from teat), with
kid present but prevented from nursing, then milked
by: Hand 1:5: setaceee BTSs. oss eee ee eee 160 17.3 27.7

The second flow of milk induced by the treatment outlined is not quite
as great as in the data given in connection with figures 5 and 8, and
table 5, but the rise in fat content is greater, and, in this respect, is

much the same as the effect noted from the use of pituitrin given in
table 5.

wa ae
4

fe (P| | tT] Tt |e ttan] Maik Recorp-| [|
mE AAS Se fio.igo6. |_|

Wal | | Pp Evin Meee
oan JY] OY A Pp ie eel el oe
af AL aaa

. ‘| V
fj esl a
* of et
f Py ee OM

Fie. 9. Showing an illustration of the failure of a a machine, followed
by hand stripping, to secure a uniform degree of removal of the milk from the
udder in the cow. Figure from M. S. thesis of the writer, University of Illinois,
1910.

While hand milking does not remove all the milk from the udder
that it may be possible to obtain by the use of some stronger stimulus,
it usually secures a fairly uniform degree of removal. Thus, in figure

PHYSIOLOGY OF LACTATION 303

9 showing the yield of milk by a cow after a 13-hour period (morning
milking) and after an 11-hour period (evening milking) it is seen that,
with hand milking, the curve for the longer period is always above the
curve for the shorter period and the two do not cross. However, when
machine milking commences, even though it was followed by hand
stripping, the character of the curves changes and they cross and
intercross repeatedly. This means that the stimulus provided by the
machine work did not produce a uniform result. The total yield over
a period of some months, it appears, need not be decreased by this
condition.

Change in volume of the udder in milking. That the udder of the
goat and cow decreases in volume during the process of milking is a
matter of common observation. It has been stated [see Marshall
(4) p. 557] that the udder of a cow could not contain the quantity of
milk which can sometimes be obtained from it at one milking. No
quantitative data are given to back up the statement. In making
measurements of the change in volume of the udder of the goat by
removal of the milk I have found in some cases a decrease equal to the
amount of milk secured, and in other cases a decrease less than the
volume of milk drawn. An average of four determinations shows a
yield of 339 cc. of milk and a decrease in volume of udder of 323 ce.
The method of volume determination, as previously noted, was not
exact. The results of one trial were as follows:

Goat No. 5. 5/5/15, a.m.

ce. ce.
Chg EERE GUE CHR G TET Ra oe eS Ri rs A ey 1540
EULER: (Tio. eee ess Bete tees e Sera ce eae eeeeo 105
TOSS INTEC. ob hues ie eee Rie eG So 6 Gea a eee 255
SM prIntini eer ert | ROS poets: ccisichie ck ns Leen 360
WOLPERT LG PRE Dc ON ene ie ih en ee RS 1170
Decrease in volume of udder by milking.....................-. 370
Injected salt solution at 25 cm. water pressure................. 480
Wight nine Qi TG ke es RS eee RY 8) ec - 1635
HRA GUnEMAVONUIN Gs rae «tas 5 eo 6s RS ss sae lee S45 ee eee 465
eeeenimenet us 1) er Cent. fab). 5... cme eee nee eee eae 185
Milked out one hour later (2.5 per cent fat)................... 82
Pitted caren (ae tae eet. ts lo, c\cso-y 1 eS is Aric 4 Maz 2s Gawlaeee 265
Ronin recovered: (le NOUTS).. as. cnaae oss ol esses wees s- 532

The pressure of injection used was probably somewhat greater
than that at which the milk existed in the udder. It is noteworthy,
however, that in this case the decrease in volume of the udder by milk-

304 W. L. GAINES ~

ing and nursing was equal to the volume of milk removed. Also,
the internal capacity of the udder after milking and nursing was shown
to be considerably greater than the volume of milk yielded.?

Change in fat content of milk with progress of milking. Data previously
given serve to emphasize the well known fact that as milking progresses
the fat content of the milk rises greatly. The effect may be duplicated
to a certain extent by allowing milk to pass through a quantitative
filter paper. The results of such a test are given in table 6. The goat
was first turned on her back and the milk cisterns gently massaged to
destroy the results of any tendency for the milk to cream in the cisterns.
25 ec. was then drawn, designated as fore-milk; then the main portion,

TABLE 6
Showing the fat content and size of fat globules in fore-milk and stripping milk in
comparison with similar data for the main portion of the milking, mid-milk,
and two fractions of a 50 cc. portion of the mid-milk separated by allowing to
pass through a filter paper until 20 cc. had filtered

AVERAGE DIAMETER OF

SAMPLE ies FAT GLOBULES

Hore-millke 25:¢c-: eee eee Nase ne. Poel 4.8
Sin pines AO ers 755i ee ae nee pels 8.3
Mid smile "DAS eG. soci eee ee. chee 6.0
Pilea ce OU: ess... see et. Sake ae 5.3
Non-filtrate; 30) 60s, eee eee ee ee 6.6
Wile (Ont rola. sree ee ee 6.0

6.0

245 cc., designated as mid-milk; and finally the last portions, amounting
to 40 ce. and designated strippings.

A 50 ce. portion of the mid-milk was placed in a filter and 20 ce. al-
lowed to filter through, requiring about 20 minutes. The portion
remaining unfiltered was transferred with the paper to a beaker, the
paper gently washed through the milk and removed. This portion
is designated non-filtrate. As controls another 50 cc. portion of the
mid-milk was treated the same as the above except the unfiltered part

* A characteristic reaction of the goats I have worked with is for them to start
chewing the cud when milking begins and to continue during the process of milk-
ing. Upon the injection of the salt solution in this trial the goat started again to
chew her cud and continued as if she were being milked. That is, the passive
distension of the gland produced the same reflex, with reference to this particular,
as is usually produced during its active contraction during the milking act.

PHYSIOLOGY OF LACTATION 305

with the paper was added to the filtrate and the paper, after light
washing through the milk, removed. This is designated filter control.
Still another 50 ce. portion of the mid-milk was allowed to stand in a
beaker during the foregoing process and a sample drawn from the sur-
face. This is designated creaming control.

From table 6 it appears that the filter paper retards the passage of
the fat globules more than it does the rest of the milk. The fat content
and size of the fat globules in the filtrate are decreased. A certain
parallelism appears between this effect and that shown in the natural
separation of fore-milk and strippings in the removal of milk from the
mammary glands. The effect of treating cow’s milk (where the fat
globules are larger than in goat’s milk) as above, is even more marked
on the fat content. Thus, in one case a sample of cow’s milk contain-
ing 7.1 per cent fat showed in the first portion filtered 0.7 per cent fat.

DISCUSSION

lt seems highly probable that the remarkable growth of the mammary
gland during pregnancy is excited by the presence in the blood of a
specific hormone, speaking in terms of current point of view. The
evidence of such natural experiments as that of the pygopagous twins,
the transplantation experiments of Ribbert, and the foetal extract
experiments of Lane-Claypon and Starling, is readily interpreted in
sucha way. On the other hand it is clear that the hypothetical hormone
is not present in large or potent quantities in the blood of the pregnant
animal, or else that the mammary gland of the pregnant female is for
some reason more susceptible to its influence than the gland of the non-
pregnant female; for the transfusion of considerable quantities of blood
does not produce positive results. The negative results of Lombroso
and Bolaffio with rats united in parabiosis would seem to overthrow
the hormone hypothesis. But the authors state that a starved member
of such a pair suffered death as quickly as a normal starved control,
even though its parabiotic mate was given full feed. That is, there
was, apparently, no interchange of circulating food substances. From
this it might be well contended that passage of a possible hormone in the
blood of the pregnant individual of a pair to the blood of the other, non-
pregnant member, was not possible. The fact that lactation may occur
in the virgin does not argue against a hormone of pregnaacy being
normally the main factor in the peculiar increased growth activity of the
mammary gland during pregnancy. From a teleological point of view

306 W. L. GAINES

such a hormone, produced by the foetus (or foetal membranes), would
be simply an organ for its own post-natal welfare.

While a development of the mammary gland similar to that occur-
ing in pregnancy is not readily produced by the introduction of blood
from a pregnant animal into a non-pregnant female, yet such treatment
of a lactating animal immediately and temporarily depresses milk
secretion. Either inhibition of milk secretion is due to a different
hormone than the one causing growth of the gland in pregnancy, or
the gland is more sensitive to its influence in the lactating stage. The
Lane-Claypon and Starling theory that milk secretion occurs as a result
of the removal of this hormone inhibitor, with the removal of the foetus,
serves to explain lactation following normal delivery and abortion.
The preliminary preparation of the gland, however, seems to be the
essential factor for its secretory activity, since lactation ceases normally
after a time even though pregnancy does not recur, and, therefore,
without the inhibition of any foetal hormone. Further, while lactation
is initiated or increased by premature delivery it is not as large in
amount as it is following a normal delivery.’

The failure of blood transfusion from the actively lactating goat to a
feebly lactating one to accelerate the secretion of milk in the latter
indicates that lactation is not due to the presence of some particular
substance (hormone) in the blood. The fact that one gland may cease
lactation before the other also argues against such a factor. The
response of the lactating gland to pituitrin is not dependent on the
presence of anything in the nature of an activator in the blood; for its
non-lactating mate, supplied with the same blood, does not respond.
This supports, indirectly, the contention that lactation is not deter-
mined by the presence of any exciting substance in the blood stream.

The form of the curves showing the effect of pituitrin on an active,
air-inflated mammary gland shows a muscular response. The relatively
rapid rise of pressure as expressive of a contraction of the milk passages,

3 If this point of view is correct it would seem possible that the foetus, by
influencing the development and preparation of the gland preceding delivery,
might affect the yield of milk by the mother following delivery. Then, the
sire of the foetus, by means of the foetus, might influence the milk yield of the
mother. In dairy theory and practice, as far as I know, the bull to which a cow
is bred is not recognized as a factor in the yield of milk by the cow following the
birth of the calf by that mating. Probably the maximum yield of milk is fixed
by the genetic endowment of the milk secreting cells of the gland and influences
of pregnancy and delivery serve purely a qualitative purpose in initiating activity
and have no quantitative effect. The point, however, is worthy of investigation.

PHYSIOLOGY OF LACTATION 307

followed by the slow decline as expressive of a relaxation is characteristic
of the response of smooth muscle. The failure of a second dose to dupli-
cate the effect of the first is not a thing unusual to the action of drugs.
Indeed, with pituitrin itself while a first dose causes a rise of blood
pressure due to constriction of peripberal vessels, a second dose may not
cause a rise of pressure. That the pressure changes in the gland upon
injection of pituitrin are not due to changes of blood pressure seems
certain from the fact that no change is produced in the non-lactating
gland in a late stage of pregnancy, although the gland is at that time
highly vascular. That it is not due to a secretory action of the cells
is shown by the rapidity of the rise and the return to the initial pressure.
A rise of pressure through a secretory action could occur only by the
absorbtion of water and its passage into the ducts. If this occurred it
is extremely unlikely that resorbtion would commence at once and be
carried to a point that would exactly restore the initial pressure. A
muscular explanation seems the only one to fit the data. The difference
in sensitiveness of the contractile elements of the gland to pituitrin
according to the stage of activity finds an analogue in the response of
uterine muscle to adrenalin, where the gravid uterus is strongly con-
tracted and the virgin uterus practically unaffected.

The outflow of milk caused by pituitrin is well explained by a muscular
action. This seems the more likely as a correct explanation when it
is remembered that the act of nursing causes an active increase of
pressure in the gland closely resembling the effect of pituitrin. The
flow of milk produced by pituitrin when simply placed on the cut sur-
face of the excised gland containing milk shows that its effect is inde-
pendent of the circulation (except, of course, to bring it in contact with
the gland when injected), and gives, also, pretty conclusive evidence
that it causes constriction of the very small milk passages of the gland.

Hammond bases his conclusion that pituitrin stimulates the secretory
activity of the cells, and has no muscular action on the gland, on the
yield of milk obtained by hand milking during a test period of a few
hours. Rather positive evidence of a muscular action has been shown
above. It must also be clear, from the data given, that any attempt
to measure the secretory activity of the gland cells by the amount of
milk that may be obtained by hand milking in a test extending over a
few hours, is subject to a large error because of variations in the com-
pleteness of removal of the milk from the gland. Milk secretion, in
the sense of the formation of the milk constituents, is one thing; the
ejection of the milk from the gland after it is formed is quite another

308 W. L. GAINES

thing. The one is probably continuous; the other, certainly discon-
tinuous. And inasmuch as no one has shown any material increase in
the milk obtained over a period of days by the use of pituitrin any
true secretory action of the drug must be held unproven.

The data given on the flow of milk during milking and nursing show
very plainly that there is a reflex constriction of the gland (very similar
to that produced by pituitrin) involved; that the removal of the milk
from the gland is dependent upon the operation of this reflex; and that
the reflex is conditioned. The stimulus which, naturally, excites the
reflex must be found in the friction and warmth of the sucking action
of the young acting on the cutaneous sense organs of the teat; with,
possibly, a further source in the passive dilatation of the sphincter mus-
cles of the nipple by the passage of the milk.

If the mother dog be placed on her back the pups experience no
difficulty in getting the full yield of milk, although it must be secured
directly against the force of gravity. But if the mother be placed under
ether, in a normal position, the pups are unable to secure any milk.
Evidently some connection in the reflex are is broken by the anesthesia,
and the indication would seem to be that the central nervous system is
involved. The failure of vigorous and continued nursing to secure
milk under this condition indicates that the normal removal of the milk
is not passive with respect to the activity of the gland. Pituitrin
restores the flow, presumably by acting directly on the contractile
elements of the gland and producing the same sort of contraction that is
ordinarily produced by the nursing reflex.

An artificial imitation of the stimulus afforded by the nursing kid
which must have duplicated the natural stimulus very closely, failed
to excite the nursing reflex in the absence of the kid; but with the kid
present the reflex was called into effective operation by the artificial
stimulus as quickly and thoroughly as by the nursing of the kid himself.
This may be interpreted to mean that conduction in the reflex are is
dependent upon the psychic condition of the mother.

When the recently delivered cow or mare has been separated from
her young some time and her udder is distended with milk, she shows,
when brought to her calf or colt, marked symptons of maternal concern,
and accompanying this there is often a spurting of milk from the
nipples. Apparently, with the gland under this extreme tension, the
_psychic state produced by the recovery of the young induces a con-
traction of the gland and an ejection of milk. But in the case of the
goat above, with the udder previously drained of milk by hand milking,

PHYSIOLOGY OF LACTATION 309

recovery of the young, although causing plainly an altered psychic
state of the mother, was not sufficient to cause a further flow of milk.
But this psychic state in combination with the mechanical and thermal
stimuli of nursing permits a reflex more powerful than that excited by
the preceding hand milking and calls into play, as it were, a certain
reserve strength in the contractile elements of the gland which results
in a large flow of milk. In the dog this seems to be the limit of contrac-
tile power of the gland, but in the goat pituitrin causes a still further
contraction as seen in the ejection of milk. (The statement is based
on trials with one goat only.) It is of interest to note the somewhat
analogous action of adrenalin on the vascular system. The drug
produces a stronger constriction of the blood vessels than can be pro-
duced by any reflex nervous excitation; but direct stimulation of the
motor nerve to a vessel may cause an equal constriction.

It is difficult to reconcile the seeming dependence of proper nursing
or milking on a central nervous mechanism with the results of Eckhard
(18) who found no effect on yield of milk in the goat by section of the
nerves to the mammary gland; or, with those of Goltz and Ewald (19)
who found that removal of the spinal cord in the dog did not interfere
with the nursing of her pups.

The determinations of volume change in the udder with milking
here reported are hardly extensive or varied enough to be absolute,
but so far as they go the indication is that practically the entire quantity
of milk obtained at any one time is present as such in the udder at the
beginning of milking. Certainly, the capacity of the ducts and alveoli
of the udder in the goat is as great as the milk yield at one time. There
is, therefore, no a priori reason why the milk should not be regarded
as being accumulated as such in the udder during the interval between
milkings, and the milking operation regarded as a more or less effectual
method of stimulating a reflex which causes a contraction of the gland
musculature resulting in the ejection of a greater or less part of the
milk therein present, according to the effectiveness with which the
reflex is excited. It has been claimed that in the heaviest milking cows
it is impossible for the udder to contain the quantity of milk that may
be secured from it at one time, but that is a matter for determination
rather than speculation.

Histologists, as a rule, have been unable to demonstrate any con-
tractile elements in the alveoli of the mammary gland. Positive
physiological evidence, however, should outweigh negative histological
evidence. Benda (20) has demonstrated contractile elements in the

310 W. L. GAINES

basement membrane of the alveoli to his satisfaction. The muscular
mechanism of the mammary gland seems to be a very important factor
in its functioning. It is, indeed, not inconceivable that the activation
of the muscular mechanism of the gland may be an important factor
in the onset of lactation. Thus, colostrum is secreted by the cells and
fills the ducts during pregnancy long before the contractile mechanism
is. sensitive to pituitrin or to the stimulus of milking. More extensive
data than are here available are needed for definiteness, but, at least,
it is clear that the contraction of the gland under the influence of
pituitrin runs more closely parallel to the appearance of an actual
outflow of milk than does the formation of milk (colostrum) by the
secreting cells.

The high fat content of strippings milk as compared with fore milk
can not be accounted for by a creaming process in the udder for the
same phenomenon occurs in the human (Engel (21)). How far the
effect noted by simple passing of milk through a filter paper is com-
parable to the mechanism that causes the result in the mammary gland,
is not altogether clear. It is, however, suggestive that the natural
process may be partly one of physica] retardation of the fat globules
in the rapid passage of the milk through the small ducts which must
occur in milking. It is generally regarded that the fat of milk is formed
within the secreting cells and the globules then extruded through
the ruptured end of the cell. It would seem quite possible, also, that
under the strain and stress imposed upon the cells by the reflex con-
striction produced by the milking act this rupture and escape of fat
globules from their free ends may be greatly increased and thus cause
the higher fat content of the later portions of milk.

SUMMARY AND CONCLUSIONS

Transfusion of blood from a pregnant goat into a lactating one
temporarily inhibits milk secretion. Placental and mammary gland
extracts from a pregnant cow have an effect similar to the blood.
There is some indication of a subsequent accelerating action in the
case of the gland extract. Transfusion of blood from a fresh, heavy-
milking goat into a low milking one fails to accelerate milk secretion.

Pituitrin has a muscular action on the active mammary gland caus-
ing a constriction of the milk ducts and alveoli with a consequent
expression of milk. This action holds, also, on the excised gland in
the absence of any circulation. The flow of milk produced by pituitrin

PHYSIOLOGY OF LACTATION SLL

is dependent on the amount of milk present in the gland. There is no
evidence of any true secretory action. The non-lactating gland, up
to a late stage of pregnancy, is not sensitive to pituitrin.
' Nursing, milking, and the insertion of a cannula in the teat, excite a
reflex contraction of the gland musculature and expression of milk.
There is a latent period of 35 to 65 seconds. Milking is a stronger
excitant than the cannula; nursing is stronger than milking; [and the
direct action of pituitrin (in some cases) is stronger than nursing].
Removal of milk from the gland is dependent upon this reflex, and it
may be completely inhibited by anesthesia. The adequate stimulus
for the receptor in the reflex are is the thermal and mechanical effects
of nursing; but the strength of the excitation thus aroused is profoundly
modified by the psychic state of the mother—especially striking are
anesthesia which greatly weakens it, and recovery of the young after
separation which greatly strengthens it.

The internal capacity of the mammary gland of the goat is greater
than the volume of milk drawn at one time, and the udder shrinks in
volume during milking to nearly the same extent as the volume of milk
drawn. Practically all the milk drawn is present as such in the gland
at the beginning of milking, and is actively ejected by a reflex contrac-
tion of the gland musculature under the stimulus of milking.

Passing milk through a filter paper separates it into fractions of differ-
ent fat content, somewhat similar to fore-milk and strippings. A
physical filtration may account, at least in part, for the rise in fat con-
tent of the milk with progress in milking.

REFERENCES

(1) Srrawu: Verhandl. d. Anat. Gesellsch. zu Kiel, 1898, 236.

(2) KNOEPFELMACHER: Jahrb. f. Kinderheilk., 1902, lvi, 791.

(3) Rrsepert: Arch. f. Entwickelungsmech., 1898, vil, 688.

(4) Marsnauu: Physiology of Reproduction, London, 1910, 585.

(5) LaAnE-CLAYPON AND STARLING: Proc. R. Soc., 1906, xxvii, B, 505.
(6) LomBroso anp Bouarrio: Arch. Ital. d. Biol., 1910, liii, 447.

(7) Nuixuas: Monatsschr. f. Geburtsh. u. Ggn., 1913, xxxviul, 60 E. H.
(8) LerDERER AND Prisram: Pfliiger’s Arch., 1910, exxxiv, 531.

(9) Gavin: Quart. Jour. Exp. Physiol., 1913, vi, 13.

‘ (10) Orr anv Scort: Proc. Soc. Exp. Biol. and Med., 1910, viii, 48.
(11) Mackenzie: Quart. Journ. Exp. Physiol., 1911, iv, 305.

(12) Hi. anp Simpson: Proc. Soc. Exp. Biol. and Med., 1914, xi, 1.
(13) Hammonp: Quart. Journ. Exp. Physiol., 1913, vi, 311.

(14) Hii ann Simeson: Quart. Journ. Exp. Physiol., 1914, viii, 103.
(15) Heaney: Surg. Gyn. and Obst., 1913, xvii, 103.

312

(16)

(17)
(18)
(19)
(20)
(21)

W. L. GAINES

FLowER AND LyDEKKER: An Introduction to the Study of Mammals,
London, 1891, 228.

ScHaEFER: Text b. Physiol., Edinburg and London, 1898, i, 668.

Ecxuarp: Beitr. zu Anat. u. Physiol., 1858, i, 3.

Goutz AND Ewan: Pfliiger’s Arch., 1896, Ixili, 362.

Benpa: Dermatolog. Zeitschr., 1894, 1, 94.

EncEu: In Sommerfeld’s Handbuch der Milchkunde, Weisbaden, 1909.

THE AMERICAN
JOURNAL OF PHYSIOLOGY

VOL. 38 SEPTEMBER 1, 1915 No. 3

AN ANALYSIS OF CERTAIN PHOTIC REACTIONS, WITH
REFERENCE TO THE WEBER-FECHNER LAW

I. THe ReEAcTIONS OF THE BLOwFLY LARVA TO OpposED BEAMS OF
LIGHT

BRADLEY M. PATTEN

From the Zoélogical Laboratory of the Museum of Comparative Zoélogy at Harvard
College and the Anatomical Laboratory of the School of Medi-
cine, Western Reserve University

Received for publication July 9, 1915
I. STATEMENT OF PROBLEM

The problem dealt with in this paper is the analysis of a series of
measurements made on the motor reactions of the blowfly larva to
light.

The measurements were obtained by a method already used with
satisfactory results (8), in which light was applied as two opposed
beams the intensity of which could be easily controlled and precisely
measured. The reactions produced by two beams of light of dif-
ferent intensities, acting simultaneously on opposite sides of the same
animal, were measured in angular deflections from an original path of
locomotion. Since no deflection was produced when the opposed lights
were of equal intensity, and since a deflection appeared and increased
progressively as the intensity difference between the lights was in-
creased, the relative difference between the intensities of the opposed
lights was regarded as the value of the stimulus. Some five thou-
sand trails were measured and the results expressed graphically in the
form of curves in which the amplitude of the reaction was plotted
against the intensity of the stimulus.

313

314 BRADLEY M. PATTEN

‘The analysis of these curves was undertaken to ascertain, if possible,
the factors that determine the relation between reaction amplitude
and stimulus intensity. Because there has been considerable work
done on the so-called Weber-Fechner relationship between ‘“‘sensa-
tion value” and stimulus intensity, it seemed desirable to consider the
measurements presented here with reference to the Weber-Fechner
Law. Such a method of treatment offers the possibility of interesting
comparisons between reaction and sensation curves, and at the same
time gives a definite point of view from which to approach an analysis
of the data.

II. INTRODUCTION

The relationship between the strength of a stimulus and the re-
sponse evoked by it was first systematically studied by Weber about
the middle of the last century (11, 12). The criterion by which he
measured the effect of the stimulus was “the least detectable change
in sensation,’”’? a method necessarily limited to human physiology.
His experiments showed that the increase in stimulus necessary to
cause the least detectable change in sensation is, within a certain me-
dium range, a definite fractional increment of the acting stimulus.
Fechner (2) attempted to give this law a more quantitative and ex-
tensive application by assuming that just perceptible differences in
sensation represent actually equal amounts of sensation. Accept-
ing this assumption, we may express the relation between stimulus
and sensation, as determined by Weber’s experiments, as follows: in
order to increase the sensation by equal amounts, that is, in arithmet-
ical progression, the stimulus must be increased according to a cer-
tain factor, that is, in geometrical progression. The sensation may
therefore be regarded as a geometrical function of the stimulus. If
this relation between stimuli and sensations is represented by a curve
in which the ordinates express the sensations increasing by equal
amounts, and the abscissas the stimuli necessary to produce them, a
logarithmic curve is obtained.

There is doubtless some ground for the assumption of Fechner
that least perceptible increments of sensation are equal; but neverthe-
less it is an assumption, and one that is made in spite of conflicting
evidence. We should not, therefore, allow ourselves to be misled by
the mathematical treatment of this problem in so far as it is based on
an assumption the exactitude of which has not been proved. Waller
(9) has pointed out that the matter is further complicated by the fact

REACTION OF BLOWFLY LARVA TO LIGHT aio

that several steps intervene between the external stimulus and the
production of the sensation. That isto say, there is first the action of
the external stimulus on the end organ; then a nerve impulse in the
sensory nerve fiber; and finally the process, in the brain, which gives
rise to the sensation. It is a question, therefore, between which of
these processes the logarithmic relationship holds good.

While sensation, for the reasons indicated, is an illusive basis for
quantitative study, it is the only basis available in the higher forms.
Where reactions are profoundly modified by past experience and volun-
tary control, it is futile to measure them to determine quantitatively the
effect of the stimulus, for, instead of being expressed directly in a series
of muscular reflexes, a stimulus is then translated first into sensation,
and the reaction that follows the sensation may be largely determined
by the past experience of the individual rather than by the strength
of the stimulus. In such cases, the only thing we can measure is the
sensation, or the nerve impulse that causes it.

Hence it is hardly necessary to emphasize the desirability of obtain-
ing additional data from organisms in which the chain of events be-
tween the application of the stimulus and the resulting production, of
some measurable reaction, is not complicated by the factors of sensa-
tion and voluntary control based on past experience. If we can elimi-
nate these factors, we may regard the amplitude of the reaction as an
index to the intensity of the physiological processes initiated by stimu-
lation. Such reactions are subject to exact physical measurement,
and their amplitude may be plotted against the intensity of the stimu-
lus in the form of stimulus-reaction curves.

It is only by a careful analysis of many such curves, based on the
study of organisms having different chains of physiological processes
intervening between stimulus and reaction, that we shall be able to
ascertain at what point in the chain of events the approximately loga-
rithmic character of the curves is impressed upon them. For the present
it is convenient to speak of reactions presenting a logarithmic curve
of increase as following the Weber-Fechner Law, or better the Weber-
Fechner Curve. We should, however, be careful to avoid thinking of
all the widely divergent types of reactions commonly included in this
category as governed by the same law.

A review of the extensive work done on the measurement of sensa-
tions does not come within the scope of this paper. The amount of
experimental work on the measurement of the amplitude of responses
directly controlled by the strength of the stimulus, has been compara-

316 BRADLEY M. PATTEN

tively small. There have been many suggestions, incidental to other
work, to the effect that such and such an organism, in its response to
light or chemical stimulation, apparently conformed with Weber’s
Law; but in the majority of cases these suggestions have not been based
on sufficient data to be of any great value. When therefore measure-
ments of the light reactions of the blowfly larva, made for other pur-
poses, (8), appeared to follow, at least approximately, the Weber-Fechner
Curve, it seemed worth while to determine more carefully the exact
extent of their apparent conformity with the law, and if possible the
causes for it.

The problem in the case of the blowfly larva seemed particularly
promising: (1) because the system of measurement, worked out in con-
nection with previous experiments on this form, has yielded consistent
quantitative data; (2) because the amplitude of the reaction was directly
controlled by the intensity of the stimulus; and (3) because the stimulus
reaction curve of a simple reflex of this type would appear to be of con-
siderable value for comparison with the more familar stimulus-sensa-
tion curves of human psychology.

Ill. THE METHOD OF RECORDING AND MEASURING THE REACTIONS OF
THE BLOWFLY LARVA TO OPPOSED LIGHTS

In these experiments, the method used for stimulating the larvae and
measuring their reactions was the same as that described in det a
in an earlier paper (8,pp. 223-239). It is sufficient, here. therefore, tilo
outline the procedure very briefly.

The apparatus used was so constructed that the opposite sides of the
animal under observation could be subjected to opposed beams
of light, the actual and relative intensity of which could be varied
at will. Figure 1 shows a plan of the apparatus. It was possible to
control the actual intensity of the lights by varying the number of
Nernst glowers thrown into circuit (fig. 1, G@) and by varying their
distance (fig. 1, Gf O) from the observation stage. The relative inten-
sity of the lights was controlled by moving the observation stage from
a point midway between the mirrors (where the opposed beams would
be of equal intensity) toward one mirror or the other. By calculating
the intensities according to the law I a *. the distance through which
it was necessary to move the stage away from the center, in order to
secure any particular intensity ratio between the lights could be readily

7

REACTION OF BLOWFLY LARVA TO LIGHT 317

ascertained. With the apparatus set up for the desired intensities,
larvae were subjected to the stimulation of the opposed beams by
forcing them to crawl into the path of light at right angles to a line
connecting the mirrors.

In preliminary experiments it became apparent that there was a
wide range of individual variability in the photosensitiveness of the
larvae. It was not possible to eliminate this factor from comparative
measurements by using a single identified larva for each series of tests,
because of the
shortness of their
sensitive period,
coupled with the
rapid changes in
their degrees of
sensitiveness
with age. The
nearest approach
that could be
made to elimina- “ygjyyeme --- - - - --- ---- 4--2--|---~--~------* YN.

Fig. 1. Diagram of apparatus used to produce differ-
'Y ential bilateral light stimulation. G, five 220-volt Nernst
- glowers; M and M’, mirrors; O, center of observation
e€ stage; broken lines, central rays of beams of light from the
+ glowers reflected to O by the mirrors; d and d’, screens with
a rectangular openings; s ands’, light shields; a and b, 2 c.p.
. orienting lights with screens.

izing test used consisted-in subjecting larvae, which were
rawling under the influence of a horizontal beam of
antaneous change of 90° in the direction of the beam.
made to crawl first toward, and then away from the
hus subjected to the change of direction of the light
ht and left sides. Those larvae orienting immediately
to both changes were considered to be of “‘standard

The nature of the pairs of trails made by an animal
nd the varying abruptness with which they came into
hown in figure 2. The data presented in the follow-

318 BRADLEY M. PATTEN

ing pages were all obtained from maggots thus standardized for their
photosensitiveness, and they may therefore be considered reliable
for comparative study.

These standardized larvae were subjected to stimulation in the
following manner. Each maggot was allowed to crawl onto the obser-
vation stage immediately in front of the orienting light (fig. 1, a).
Under the influence of this light it was forced to crawl toward the
-center of the stage in a direction at right angles to the path oflight
thrown by the mirrors (fig. 1 ff’). By putting a drop of dilute methylen
blue on its posterior end, the larva was made to record its own trail
on a sheet of paper. When the animal was well onto the stage and
crawling steadily, the orienting light was turned out and the mirror

Fig. 2. Three examples of test trials which pass the stan
test. Each pair of trails was made by a different Jarva. Thou

new direction of light. The arrows mark the points at which
the light was changed.

point at
being indi-

uence of

beams were thrown on simultaneously from either side,
which the larva was subjected to the bilateral illuminati
cated on the trail. The part of the trail crawled under the
the bilaterally applied stimulus is the significant portio:
Loeb (5, p. 2), has pointed out that when “two soure
equal intensity and distance act simultaneously upon
heliotropic animal, the animal puts its median plane at ri
the line connecting the two sources of light.’”’ We should

of light of
[negatively]

REACTION OF BLOWFLY LARVA TO LIGHT 319

that a larva subjected to the action of opposed beams of equal inten-
sity, would continue crawling in a direction at right angles to a line
connecting the mirrors. That such is in fact the case was demon-
started by a large number of trials. It was, however, observed that
certain larvae tended to craw] somewhat toward either their right or
‘left even under balanced illumination. Apparently this tendency
to deflect from the ncrmal course was due to bilateral asymmetry
of photosensitiveness (8). To eliminate the distorting effect of un-
balanced sensitivity from the final measurements of the trails, each
larva was started onto the stage first from one direction and then
from the other. By thus reversing the direction of ciawling and
making the trails in “opposite-direction-pairs,” the deflection due to

Fig. 3. A set of trails made under equal, opposed lights by a symmetrically
responding larva; a, the test trails; b and c, trails under the influence of balanced
lights.

asymmetry was thrown on opposite sides of the “norm,” and would
consequently be. canceled out of the final measurements. Figure 3
is a photograph of the complete record of a larva’s responses to equal
bilateral stimulation; a shows the pair of trails made by the maggot
in the preliminary standardizing test; b, an ‘“‘opposite-direction-pair”’
of trails made under balanced illumination; and c¢, a second pair of trails
similar to those in 6, made after the larva had been rested for half
an hour. In the record sheet photographed in figure 3, c, it will be
observed that the trails do not conform exactly to the perpendicular
to the line connecting the sources of light. There is a deflection to

oY

320 BRADLEY M. PATTEN

the right in that trail made by the larva while crawling away from the
observer, and to the left in that trail made by the same larva while
crawling toward the observer. If now we measure the deflection of
the trails in degrees, by placing a protractor with its center at that
point on the trail where the head of the larva was when the lateral
lights were turned on, and its base parallel to the line connecting the’
sources of light (fig. 1, ff’) we find that the deflections to the right and
left are of equal value. By tabulating left hand deflections in plus
degrees and right hand deflections in minus degrees, the effects of
asymmetry are cancelled, and we have as a final result a measurement
identical in value with those measurements taken on larvae that have
a perfect balance of sensitiveness. As a certain amount of unbanance
of sensitiveness is far more common than perfect symmetry, this check-
ing out in all measurements of the possible effects of asymmetry is
of the utmost importance. By taking the sum of the plus and minus
deflections of a larva, and dividing it by the number of trails measured,
the average angular deflection of the larva is determined. The average
obtained by measuring four trails such as those shown in figure 3, b
and c, expresses very accurately the response of any individual larva.

The mean of the average angular deflections of a number of larvae
may be considered as expressing accurately the value of the response
which is elicited by a given set of conditions. Table 1 is an example,
taken from several similar sets of measurements that were made under
equal bilateral illumination, which serves to show the method of com-
piling the measurements. A study of the table shows that very few
trails conform exactly to the theoretical response, but the total plus
and minus deflection of each larva comes near to cancelling, and the
average of the four trails of each larva is close to zero degrees deflection.
The average response of the whole set corresponds almost exactly to
the theoretical response, the average deflection from the perpendicular
being only —0.025°. This is astonishingly accurate when one considers
that a degree on the protector of 7.6 em. radius used in measuring the
trails was only 1.5 mm.

The striking results of table 1 were confirmed by several repetitions
of the experiment. The immediate significance of these results is
two-fold. (1) The consistent closeness of the average trail to the
perpendicular, and the equal distribution of the trails on either side of
it, indicate that the method of individual measurement has eliminated
from the final results the effects due to asymmetry and placed the dif-
ferent individuals on a uniform basis for comparison. (2) The elose

REACTION OF BLOWFLY LARVA TO LIGHT 321

conformity of the aggregate response to the theoretical response,
when the lights are of equal intensity, makes a well grounded point of
departure for a series of experiments with opposed lights of unequal
intensities; for if there is no deflection toward either side following
equal bilateral stimulation, the deflection appearing under unequal
bilateral simulation may be regarded as a true expression of the physio-
logical effect of the difference between the lights. Having, therefore,
perfected a system of measuring the reactions of. the larvae to light
and established a fixed point of departure by the measurements under
equal opposed lights, we were in a position to measure the reactions to a
graded series of intensity ratios.

TABLE 1

A table giving the values of the angular deflections made by larvae under the
stimulation of equal opposed lights

NUMBERS OF THE LARVAE {of BoE ee i a ue Re aceie
DATE Ist 2nd 3rd 4th a a PERLEC-

trail trail trail trail g S enor

degrees

No. 14, 13/5/13 +10 | —16 | +15} — 7 | 28 | 25} + 2) 40.50
No. 18, 13/5/7138 0O;+6;)/+2/;]+4+ 5 0} 13 | +13 | +3.25
No. 20, 18/5/7138 +7/]— 8] +18 |} —12} 20} 25] + 5] 41.25
No. 15, 13/5/7138 +25 | —11 } —11} — 3] 22 |] 28) + 6] +1.50
No. 4, 13/5/13 +6/—6/+9]— 7) 18] 15] + 2] 40.50
No. 22, 13/5/7138 Q0|;+5!)]—-7)-1 8/| 5} — 3] —0.75
No. 21, 13/5/7138 — 5] — 3] —21 | +17 | 29! 17 |} —12 |} —3.00
No. 16, 13/5/13 +17 | —14 0} — 6} 20} 17} — 3} —0.75
No. 11, 13/5/18 — 9} — 1/413} — 6} 16} 18] — 3} —0.75
No. 9, 13/5/713 —-11/+ 5; —5|+ 3} 16 8 | — &]} —2.00
10 larvae 167°|166°; —1° | —0.025

40 trails

IV. A COMPARISON OF THE STIMULUS-REACTION GRAPHS OF THE BLOW-
FLY LARVA WITH THE WEBER-FECHNER CURVE

1. The graph expressing the values of the angular deflections produced
by a graded series of percentage differences of intensity between opposed
lights.

The problem, so far as the method of experimenting was concerned,
now resolved itself into the compiling of a series of reaction measure-
ments for comparison with various values of the acting stimulus. As

aus BRADLEY M. PATTEN

has already been pointed out, we proposed to measure, not sensations,
but the amplitudes of reflexes initiated by stimuli of various degrees
of intensity. The measurable reaction in the method of experimenting
outlined above, is the larva’s angular deflection from an original path
of locomotion. The original path of locomotion was so arranged with
reference to the lights producing the significant portion of the reaction,
that the larva came into the path of the light at right angles to the line
connecting their sources. We have already seen (Table 1, p. 321) that
when the opposed beams were of equal intensity, there was no deflec-
tion from the initial direction of locomotion. But if we subject larvae
to the action of wnequal lights, there appears a deflection toward the
weaker light, which is directly proportional to the difference of intensity
between the lights (Table 3 and figure 4). We may say, therefore, that

TABLE 2

Summary of measurements of 200 trails made to determine the value oj the angular
deflection when the opposed lights are of equal intensity

aa NUMBER OF = s N AVERAGE

Se TReHeae be psa TSRELMCHTCN SERDeCHON eae Respite =
1 40 187 165 —22 —0.55
2 40 180 176 — 4 —0.10
3 40 207 225 +18 +0.45

4 40 167 166 — 1 —0.025

5 40 246 237 — 9 —0.225
Total. 200 Average angular deflection for 200 trails —0.09

the angular deflection of the larva is made in response to a differential
bilateral stimulus, the value of which may be expressed in terms of
the relative intensity of the opposed lights, or more conveniently, in
terms of the percentage difference between the lights.!

The measurements obtained from the trails of larvae subjected to
equal bilateral stimulation were taken as the starting point of the stimu-
lus-reaction curve. Table 1 shows in detail the deflection measurements
taken from a set of 40 trails made by ten larvae. This set of data was
supplemented by four other similar sets. Without reproducing the
tables in detail, the results may be summarized as in Table 2. The
totals of Table 1 appear in Table 2 as experiment 4.

1The percentage differences throughout the experiments are computed in
terms of the stronger light.

REACTION OF BLOWFLY LARVA TO LIGHT 323

Stating the results of Table 2 in terms of the series of experiments
of which it forms the starting point,we may say that the reaction to
opposed lights of 0 per cent intensity difference, is a minus deflection
0.09°. This is so near zero degrees that it undoubtedly signifies the
absence of deflection under these conditions.

Following the experiments at equality, a series of measurements was
made at the following percentage differences between the lights: 8}
per cent, 162 per cent, 25 per cent, 334 per cent, 50 per cent, 662 per
cent, 834 per cent, and 100 per cent. As the method of measurement
and tabulation was the same throughout, a summarizing table will be
sufficient to show the results obtained.

TABLE 3

A table based on the measurement of 2,500 trails showing the progressive increase
in angular deflection with increasing differences between the lights.

Fereentage Difference in} Equal-| 83% 163% 25% 333% 50% 662% | 833% 100%
Lights. ity 0%

Number of Trails seas
ured.

Average Angular De- =) ° ° ° o ° (|| °
Hettion: \ 0.09 | —2.77°| —5.75°| —8.86°}—11.92°|—20.28°|—30.90°|—46.81°| —77 .56

These results may be expressed graphically by plotting degrees
of deflection along the axis of ordinates and percentage differences of
intensity along the axis of abscissas. The stimulus-reaction curve
thus obtained is shown in figure 4.

It is obvious that this curve is not directly comparable with the
logarithmic Weber-Fechner Curve constructed on the basis of absolute
increments in the stimulus, since the curve of figure 4 is plotted on the
basis of percentage increments in the stimulus. For purposes of com-
parison, however, a theoretical curve may be constructed of the type
which would be obtained by measuring, under the conditions of these
experiments, a reaction governed by the Weber-Fechner Law. Using
percentage differences in stimulus instead of absolute differences, and
degrees of deflection instead of sensation units, such a curve has
been plotted as a broken line in figure 4.

The construction of this theoretical curve is based on the follow-
ing simple deduction from the usual statement of the Weber-Fechner
law. If ‘‘the increase of stimulus necessary to produce a unit increase
in response is a definite fractional increment of the acting stimulus,”
then it follows that the same fractional increment in a stimulus of any

324 BRADLEY M. PATTEN

intensity will produce equal changes in response. In these experiments
(see p. 322) an increase of the intensity difference between the opposed
lights is regarded as an increase in the stimulus acting to deflect the
larva from its original path of locomotion. For a difference of 8%
per cent between the lights, there is experimentally a deflection of 2.77°
(see fig. 4 and Table 3). Dealing in round numbers let us assume, for
the plotting of the theoretical response, that a difference of 83 per cent

80

Hsu Ee EEE TR

_

iit
i

60

|

:

HU TE

50

Degrees of Deflection

20

40

30 40 50 60 70 80 90 100
Percentage Difference in Lights

Fig. 4. The solid line is a curve representing the angular deflections of the
blowfly larva under opposed lights of various intensity ratios. (See also table
3). The broken line is a theoretical curve such as would be obtained by plot-
ting, under the same conditions of measurement, a reaction governed by the
Weber-Fechner Law. In plotting this theoretical curve, a difference of 8} per
cent between the lights is assumed to produce a deflection of 3°.

Ratio of Lights.

1to ti} = 84% difference. 1 to 2 = 333% difference
lto ¢ =163% difference. 1 to ys = 413% difference.
lto 2? = 25% difference. 1 to 4} = 50% difference.

REACTION OF BLOWFLY LARVA TO LIGHT 3825

between the lights will produce a deflection of 3°. This point would
be located in precisely the same manner as the first point on the logarith-
mic Weber-Fechner curve. It is after the location of the first point
on figure 4 that the method of plotting this percentage-increment curve
begins to differ from that of plotting the absolute-increment curve.
In the former curve, each point must of necessity be figured from unity
as a starting point; in the latter, only the first point is so figured; each
succeeding point being based on the one immediately before it. The
values of the stimuli in an absolute-increment series would run: 1 + 3's
of 1 equals 1.083; 1.083 + +5 of 1.083 equals 1.173; 1.173 + 75 of 1.173
equals 1.271; etc. The values in the percentage-increment series would
run: 1 to +4 (84 per cent difference); 1 to % (162 per cent difference) ;
1 to 2 (25 per cent difference); 1 to ? (334 per cent difference); etc.;
computed each time from the same starting point (unity). Each in-
crease of 84 per cent in the difference between the lights is assumed,
in this case, to produce an increase in deflection of three degrees. The
theoretical curve plotted in this way is rectilinear (fig. 4).

The steps in this percentage-increment curve do not represent the
same sort of intervals as the least perceptible increments used by
Fechner. In experimental work of the type under consideration it is
not practical to use least perceptible changes in reaction as units of
measurement, for it involves, to too great an extent, the personal
judgment of the experimentor. It is far more accurate to measure
responses that are of readily appreciable amplitude by some standard
unit of physical measurement. If it is desirable to compare the results
thus obtained with results obtained by typical Weber-Fechner measure-
ments, it is possible to put the two types of curves on the same mathemati.
cal basis. Suppose the Weber-Fechner unit of reaction measurement to
be represented by the letter w. A response which is of considerably
greater amplitude than the least detectable change in reaction may be
represented as n times wu. If there is justification for considering least
perceptible increments as units, then we may handle n.w in any formula
in which wu could be handled. If the Weber-Fechner unit is not reliable
for quantitative work, as much experimental evidence seems to indicate,
we have, by using standard units of measurement put the whole matter
on a more sound quantitative basis. The curve of figure 4, though
constructed on a different basis from that therefore employed in work
of this nature, may be regarded as expressing the stimulus-response
relationship postulated by the Weber-Fechner Law.

Because it is based only on the value of a step at the beginning of

326 BRADLEY M. PATTEN

the experimental series, the curve shown in figure 4 is not necessarily
correct for the conditions of the entire set of measurements. It is,
nevertheless, evident that a curve expressing the Weber-Fechner
relationship plotted to percentage increase in stimulus would be rec-
tilinear. The reason for the apparent conformity of the larva’s stimu-
lus-reaction curve with the theoretical curve through the range of lower
intensity differences, is that the theoretical Weber-Fechner Curve was
constructed on the basis of units taken from the lower portion of the
experimental curve. Clearly there can be no conformity, other than
a limited fortuitous one, between the theoretical rectilinear curve of
the Weber-Fechner Law and the experimental stimulus-reaction curve
of the blowfly larva.

2. The graphs expressing the values of the angular deflextions produced
by a graded series of absolute intensities of opposed lights having their
relative intensity maintained constant.

Under the experimental conditions already described, it was possible
to make another series of measurements, which may be considered
from the point of view of the Weber-Fechner Law.

Granting for the present Fechner’s assumption that just perceptible
differences in sensation may be considered as equal differences,—
or sensation units,—we may make the deduction from the law already
referred to on page 323, namely: if the increase of stimulus necessary
to produce the least perceptible increase of response is a definite frac-
tional increment of the acting stimulus, then it follows that, within
physiological limits, a definite fractional increment of a stimulus of any
intensity would produce the same change in reaction. To put it con-
cretely, a 334 per cent addition to a stimulus of one unit should produce
the same increment in reaction as a 333 per cent addition to a stimulus
of five units, though the absolute increase in stimulus is 4 of a unit
and % of a unit in the other.

To obtain data which could be referred to this form of the Weber-
Fechner law, it was necessary to maintain the intensity ratio of the
opposed lights constant, while the lights were varied, together, through
a series of different absolute intensities. Such conditions were readily
producible with the apparatus described on page 317. The intensity
ratio of the two lights at any point between the mirrors (fig. 1, 17 and
M’) is determined by the distances, along the path of the beams, from
that point to the source of light. Therefore if the observation stage
is fixed at the point where a desired intensity ratio exists, the absolute
intensity of the common source of light (and with it the intensity of

REACTION OF BLOWFLY LARVA TO LIGHT 327

the opposed beams) may be varied at will without changing the inten-
sity ratio of the lights at the observation stage.

In these experiments the apparatus was set up for intensity ratios,
corresponding to zero, 84, 163, 25, 333 and 50 per cent differences
between the opposed lights. At each of these intensity ratios the
common source of light was varied by using from one to five Nernst
glowers. Table 4 gives the actual intensities of the opposed lights
under the above conditions, measured at the center of the observation

stage. .
TABLE 4
The intensities of the opposed beams of light measured in candle meters at the center
of the observation stage

INTENSITY OF WEAKER INTENSITY OF STRONGER
LIGHT IN CA. METERS LIGHT IN CA. METERS

PERCENTAGE
DIFFERENCE IN RATIO OF LIGHTS
LIGHTS

1 glower

2 glowers
3 glowers
4 glowers
1 glower

3 glowers
4 glowers

Equality ltol 6 .32|13 .9/24 .6|31 .5/41 . 1/6 .32/13 .9/24 6/31 .5/41 .1
83% 11 to 12 6.05/13 .4/23 .6|30.2|/39 .3/6 .61/14 .6/25 .7/32 9/42 .9

163% 5 to 6 5.77|12 .8|22 .5|28 .8/37 .5|6 .93)/15 .3/26 9/34 5/45 .0
25% 3 to 4 5 .50}12.1/21 . 4/27 4/35 .7|7 .34/16 .2/28 5/36 5/47 .6
333% 2tos 5.21/11 .5]20.2|25 .9|33 .8|7 .83/17 .1]30.5/39 .0/50.9
50% 1 to 2 4 .60}10.2/17 .9|22 9/29 .9|9 20/20. 2/35 .8]45 .9/59 .9

The reactions of the larvae to the five different absolute intensities
used at each intensity ratio of the opposed beams, were measured and
the data compiled in Table 5. The same results have also been plotted

TABLE 5
A table based on the measurements of 2700 trails, showing the average angular
deflections at five different absolute intensities and six relative differences of
intensity

degree dejrees degrees degrees dejrees degrees
AP GIOWEN. a oeJc <aces)-%'| O55 —2.32 —5.27 —9.04 —11.86 | —19.46
PGI GWENS: oor 45 83.2.) —O0-10 —3.05 —6.12 —8.55 —11.92 | —22.28
3 Glowers............| +0.45 —2.60 —5.65 —8.73 —13.15 | —20.52
4 Glowers............| —0.025 | —2.98 —6.60 —9.66 —11.76 | —19.88

BASIGWETA..¢.-)) 440s. —0.225 | —2.925 | —5.125 | —8.30 | —10.92 | —19.25

POETIC ix morn nic ascte*« —0.09 | —2.77 | —5.75 | —8.86 | —11.92 | —20.28

328 BRADLEY M. PATTEN

graphically in the form of a five-point curve to express the values of
the reactions obtained under each of the six intensity ratios. (fig. 5).

If the principle of the Weber-Fechner Law holds good, a definite per-
centage change in the value of any acting stimulus will produce the same
change in reaction (see p. 326). In these experiments, therefore, the
Weber-Fechner expectation is a horizontal rectinilear curve at each
intensity ratio. All the experimental curves, however, show marked
unevenness. Take for example the curve obtained under a 333 per
cent difference of intensity. There is a maximum deflection of 13.15°
at the three-glower intensity, and a minimum deflection of 10.92°
at the five-glower intensity. The curve, however, is of a roughly hori-

i af fal as ut oe
oe

ce os
oe

3 a z
: oo fees on

© ae a are SESE ST ere
:
© sok

0?

aneasee yypeanne— rets ott basset esse} r Css5 e
owers 3 Glowers “<2 Glowere 5 Clowese.”

Absolute Intensity of Source of Light

‘1 Glower

Fig. 5. Curves showing angular deflection made by blowfly larvae under
opposed lights maintained at a fixed intensity ratio, while the absolute intensity
was varied by using from 1 to 5 Nernst glowers at the source of light. The broken
lines represent the average position of the 5 points of each curve. For signifi-
cance of solid, horizontal lines see page 337.

zontal, rectilinear type. The interpretation of the results depends on
how much significance we attach to the variation of 2.33° between the
points of the curve most widely different in value from each other. Cases
in which a moderate intensity of light, gave a greater response than a
very high or very low intensity, have been observed by Walter for
Planaria (10) by G. P. Adams for Allelobophora (1), etc., which made
me rather expect that a similar maximum at a medium intensity might

REACTION OF BLOWFLY LARVA TO LIGHT 329

appear in the curves run at other percentage differences. This, how-
ever, was not the case, as may be seem from the graphs of the reaction
values at six different ratios of intensity (fig. 5). There is no con-
sistency in the position of the maxima; in fact, no two curves are strik-
ingly alike in any respect, save possibly the tendency to run low at
the five-glower intensity. As the maxima depart but little from the
mean, such inconsistency among the similar curves suggests that the
variations are probably merely fluctuations due to experimental error.
This is further indicated by computing the average position of the
five intensity points of a curve and comparing the amount by which
the maximum or minimum is separated from the mean, with the aver-

TABLE 6

A table showing the interrelation of percentage difference in the intensity of the
the lights, average angular deflection, average error of the trails, and departure
of the maximum from the mean of the curve.

|
7 GREATES
‘AVERAGE 7REATEST

ERROR’ — GREATEST tbs a eek hoale
RELATIVE INTENSITY OF Rae ae ep mesbon oF petegaileta ed OF THE CURVE
oH ELH EN DEFLECTION SINGLE TRAILS | CURVE FROM MEAN Yo genee
FROM MEAN OF CURVE
OF THE SET eee
degrees dejrees degrees per cent
Equality 0.09 | 1.98 | 0.53 (at 3 gl.) 97
3% difference 244 +1.58 0.45 (at 1 gl.) 28
163% difference 75 | +187 0.85 (at 4 gl.) 4]
25% difference 8.86 | +2 .93 0.80 (at 4 gl.) 27
334% difference | 11.92 +3 36 123i (Ate ele) 36
| +4.75 2.00 (at 2 gl.) 42

50%, difference | 20.28

age variability of the trails used in compiling the curve. This compari-
son shows the departures of the maximum and minimum to lie well
within the value of the ‘‘average error’ of the experimental data.
Take, for example, the case of the curve which has the greatest varia-
tion, the curve at 50 per cent difference. The deflections (see Table
5 and fig. 5) are at one glower, 19.46°; two glowers, 22.28°; three glowers,
20.52°; four glowers, 19.88°; five glowers, 19.25°. The average deflec-
tion for all five intensities is 20.28°, represented on the 50.per cent
curve of figure 5 by the (broken) “mean line.”’ The greatest depart-
ture from this mean is at the two-glower intensity. Now by referring
to Table 6 the average deviation of single trails from the mean trail
at 50 per cent difference is seen to have been +4.75°. The maximum

' 330 BRADLEY M. PATTEN

departure of this curve from the mean is but 42 per cent of the average
error of the experimental data on which the curve was based. Similar
computations have been made for the other curves and the figures
given in Table 6. The departures of the curves from a straight line
are not only well under the value of the experimental variation in
all cases, but they bear a reasonably constant relationship both to the
average error and to the total deflection at each intensity. As the
amount of response diminishes in the successively smaller intensity
differences, the average departure of the trails from the mean dimin-
ishes. And as the average error of the trails decreases, the separation
of the maximum from the mean decreases correspondingly. The last
bit of evidence is the variation in the curve at equality. This falls in
place in the decreasing series, but yet the variations from the mean
are still sufficient to indicate the fortuitous nature of the similar varia-
tions in the other curves, for if an absolutely straight line could be
obtained under any conditions, it would be at equality. There seems,
therefore, to be no room to doubt that the slight maxima of these curves
are merely chance variations, and that the significance of each curve
is a horizontal line represented by the “mean line”? computed from the
five points of a curve. The manner in which the curves of the series
grade in together without overlapping, and the confirmation of each
curve by every other curve, make the evidence practically complete.

We may therefore conclude that the curves of angular deflection
values for a graded series of absolute intensities of opposed lights having
their relative intensity maintained constant, resemble in contour
the rectilinear curves expressing the Weber-Fechner expectation. The
significance of this resemblance is a matter which is taken up at
length in the subsequent discussion of the results. It is puzzling
in view of the fact that the other series of measurements made on the
same animal under the same experimental conditions can not in any
way be regarded as following the contour of the theoretically correct
Weber-Fechner Curve.

V. DISCUSSION

Apparently it is customary to consider a given reaction to be in accord
with the Weber-Fechner Law, when the curve representing the magni-
tude of the responses produced by increasing stimulation is similar to
a certain logarithmic curve said by Fechner to express the stimulus
sensation relationship which Weber established for the sense of touch

REACTION OF BLOWFLY LARVA TO LIGHT 331

in man. Astonishingly diverse types of reactions have been held to
yield such a curve; the category runs all the way from the photic re-
actions of fungi (6) to human sensations. The heterogeneity of the
reactions said to follow this law has occasioned strangely little com-
ment. It seems hardly more justifiable to assume, without inquiring
into the mechanism of the responses, that widely different organic
reactions follow the same law because they exhibit similar curves
relating reaction amplitude to stimulus intensity, than to classify
animals on the basis of their size.

Critical consideration of the various responses said to follow Weber’s
Law will, I believe, lead to one of two conclusions: either there is some
phenomenon fundamental to all these types of reaction which impresses
its characteristics on the responses in which it occurs, whether or not
they involve a central nervous system, or we have been misled by the
lure of a mathematical formula and have sheltered under it a group
of observations only superficially similar. Which of the conclusions
is correct can be judged only in the light of further and more closely
analytical work. It is not improbable that both conclusions may
prove to a certain extent true. When the various reactions said to
follow Weber’s Law have been successfully analyzed, and when it is
known why each reaction follows the curve that it does, it may be
possible to say that a certain group of reactions follow a logarithmic
course because there is some particular phenomenon common to them
all. Quite probably the same analyses which make possible the formu-
lation of a law governing such a group of reactions will exclude from the
group certain reactions presenting a curve which is similar, but where
the similarity results from causes peculiar to the individual responses.
At present there is little justification for talking about a ‘“‘psycho-
physical Jaw.” It can be said, only, that many reactions have been
found to exhibit strikingly similar curves when increase in their ampli-
tude is plotted against increase in stimulation. When the causes for
this similarity are known, we will be in a position to formulate a law—
or to abandon an old hypothesis.

The experiments dealt with in this paper were arranged to facilitate
a consideration of the results from the point of view of the Weber-
Fechner Law. It was hoped that if these stimulus-reaction curves
did conform with the curves obtained in stimulus-sensation work, analy-
sis of the different reactions might show a dominant factor common
to both types of response. The establishing of such a factor would
go a long way toward clearing up our rather hazy ideas as to the nature

332 BRADLEY M. PATTEN

of the so-called “psychophysical relationship.” As we have seen,
however, of the two series of measurements made on the same animal
by the same experimental methods, one clearly does not accord with
the Weber-Fechner Curve, whereas the other way may be regarded as
following, within the limits of experimental error, the theoretic expec-
tation of the law. If Weber’s law held good for this reaction it would
govern the reaction from whatever point of view the measurements
were considered. Since such is not the case, we must seek some fac-
tor which is capable of producing both of the two types of curves ob-
tained experimentally. Such a controlling factor is, I believe, demon-
strable in the structural plan of the orienting mechanism of the larva.

In an earlier paper, experimental evidence was advanced which
justifies the conclusion that, in this form, orientation to opposed lights
depends upon symmetrically located sensitive areas operating bilater-
ally on the musculature (8). The evidence may be summarized as
follows: (1) When the lights acting on the opposite sides of the larva
are equal, the larva orients so that its median plane is at right angles
to the line connecting the sources of light. (2) When the opposing
lights are unequal, a deflection toward the weaker light appears. (3)
Certain larvae are asymmetrical in their response, deflecting toward
the less sensitive side when subjected to equal bilateral illumination.
(4) The blackening of one side of the sensitive region produces a de-
flection toward the side thus artificially made less sensitive (3). (5)
Asymmetry of sensitiveness may be balanced by a corresponding
inequality of the stimuli acting on opposite sides of the animal.

In view of these facts there seems to be no explanation for the res-
ponses of the blowfly larva to opposed lights, other than the assump-
tion that symmetrical sensitive areas affect the musculature of the
two sides of the animal in proportion to the stimulation received.
We do not know precisely the nature of the mechanism concerned
in the reaction, but there are certain general lines on which such a
mechanism must be based. “If the angle of orientation under opposed
beams of light i is such that the stimulation of the opposite sides is equal-
ized, the receptive mechanism must be of such a nature that varying
the axial position of the animal produces changes in the relative amount
of stimulation received on opposite sensitive areas. Otherwise there
would be no cause for the animal to assume a definite angle of orienta-
tion for each intensity difference between opposed beams of light”’ (8).

This equalization cannot be accomplished by bilaterally located sen-
sitive areas that are parallel to each other. Change in the axial posi-

ae

REACTION OF BLOWFLY LARVA TO LIGHT 333

tion of an animal having localized, bilaterally placed, parallel photo-
receptors, would result in a change of:the intensity operating upon the
sensitive surfaces of each side of the animal, but since the surfaces are
parallel their projection on a plane at right angles to the rays of light
would be the same whatever the axial position, and consequently the
amount of light falling on the two sensitive surfaces would be always
equal. If, however, we assume that the sensitive surfaces are inclined
at an angle to each other, the case is different. Changes in the animal’s
axial position with reference to the lights will then result in changes
not only of the effective illumination on both sensitive surfaces, but in
the relative intensity of the light operative on the opposite photo-
sensitive areas. It is not unjustifiable to assume as a working hypothe-
sis, therefore, that the power to balance unequal bilateral stimulation,
by a change in the position of the axis of the body, depends on the in-
clination to each other of the photo-sensitive surfaces.

It is possible to compute the value of an angle of sensitive surfaces
which would equalize the light stimulation on the opposite sides of an
animal, when the angle of orientation the animal assumes under unequal
bilateral illumination is known. This has been done in the case of the
blowfly larva, on the basis of angles of orientation experimentally ob-
tained under four different intensity ratios of opposed lights.2, The mag-
nitudes of the four angles thus constructed are in good agreement, being
82°, 83°, 83°, and 82°. The consistency of these computations, together
with the fact that some mechanism operating after the manner of sen-
sitive areas inclined to each other at an angle seems to be the only
way in which unequal bilateral stimulation can be equalized, led me
to put forward the hypothesis outlined. I believe it expressed in a
crude and much simplified form the dynamics of orientation in the
blowfly larva.®

It seemed to me possible that the structure of the reacting mechanism
might in this case, be the factor which determined the amplitudes of
the responses to various values of the stimulus. To test this supposi-

2 For a detailed discussion of the method of constructing the hypothetical
angle, and the geometrical proof of the construction, see (8) pp. 265-268.

3 This angle, I would emphasize again, probably does not represent the actual
position of the sensitive surfaces in the larva. There are too many modifying
factors intervening between the direction of the rays of light in the field and
the angle of incidence of the light on the sensitive areas to permit of basing any
conclusions directly on the magnitude of the computed angle, for the structural
peculiarities of the dioptric apparatus, or of the tissue overlying the sensitive
area would change its value to a large extent.

334 BRADLEY M. PATTEN

tion, an angle of sensitive surfaces was assumed corresponding to the
value of the angle computed in the paper referred to. Taking 80°
as the angle of inclination of the sensitive surfaces to each other, the
angle of orientation which would equalize the stimulus bilaterally was
calculated for the intensity ratios used in this seriesof experiments.
The angular deflection values thus obtained were plotted graphically
for comparison with the experimental curve (figure 6.) The two curves
correspond so exactly that they are indistinguishable at several points.

Le ae
= asaay

=

as a ake
Po en Ge escent Ge
Bge eee

Percentage Differsnces im Light

Fig. 6. A figure for the comparison of the reaction graph of the blowfly
larva, with a graph showing the angular value of the axial positions that would
equalize the light bilaterally in a photoreceptive system with sensitive surfaces
inclined at an angle of 80° to each other. Solid line, experimental curve (see
fig. 5); broken line, theoretical curve.

It will be observed, however, that the theoretical curve has not been
carried beyond the 663 per cent intensity difference. The reasons why
is not justifiable to carry the curve beyond this point are as follows.
In its normal method of locomotion the blowfly larva extends its ante-
rior end alternately to right and left, fixing its maxillal hooks in the
substratum and pulling its body up by the aid of their attachment.
(See also, 7, p. 176). The extent of the side-to-side swinging of the
head in this process is variable and very difficult to estimate accurately.

ne

REACTION OF BLOWFLY LARVA TO LIGHT 335

I should place the average swinging of the lavra when not subjected
to sudden change of illumination as approximately 30° to 40° on either
side of the mid-position. The effect of this swinging on the amount of
light intercepted by sensitive surfaces inclined toward each other at
an angle of 80° is indicated in figure 7. The line AB is drawn perpen-

dicular to the line connecting the sources of light, 1 L and 3 L. The

line OC is constructed at.an angle to AB (31°) such that the amount
of light reaching the opposite surfaces of the 80° angle, edf, constructed
about OC as a bisector shall be equal when the lights : are of a three to
one intensity ratio. The
broken lines mx and hx
are constructed at an
angle of 35° to OC. : 25 Se =
They represent the axial Sees faa: Be 22 ee

au
position of the anterior : | - ee

end of the larva at the
extreme of its swing to-
ward either side, while
the angles ghi and lmn
represent the resulting
positions of the sensitive
surfaces. In the axial
position, mx, thesensitive
surface mis shaded from
the light 3 Z and exposed
to the light 1 Z. When
the angle of orientation
(BOC) isnot great enough Fig. 7. For explanation see text, pages 334
for the side-to-side swing- BCE SD

ings to shade one sensitive area, they would not appreciably influence
orientation, because the swings being equal in extent toward either side,
the total effect of the light would be approximately equivalent to its
effect in the average position (< fde fig. 7). The particular angle of
orientation here illustrated lies at about the critical point where the
side-to-side swinging begins to alter the simple action of the lights
which holds at smaller angles of orientation. Although in this case
the extent of the are of swinging in which the effect of the shading is
felt is so small that it has little influence on the angle of orientation,
yet at the wider angles of orientation the effect of the shading be-
comes rapidly greater, causing the slope of the curve to change markedly

336 BRADLEY M. PATTEN

in the region where this second factor first becomes operative. (See
fig. 6, experimental curve.)

Calculations have been made for the upper portion of the reaction
graph on the basis of the light energy intercepted by both surfaces
in the complete are of swinging. This treatment, however, is of too
uncertain a nature to be of any quantitative value. Still it yields |
results which, in a general way, accord with the more abrupt rise in
the upper portion of the experimental curve.

We would seem, therefore, to be justified in stating that an animal
possessing the peculiarities of locomotion characteristic of the blow-
fly larva, together with bilaterally located photosensitive surfaces
inclined toward each other at an angle of 80°, which influence themus-
culature of the opposite sides in proportion to the stimulation they
receive, would, if subjected to the conditions of these experiments,
yield a type of reaction graph similar to that plotted from the measure-
ments made on the blowfly larva.

The type of curve which would be produced under the conditions
of the second set of experiments (see pp. 326-320) by a mechanism of
the nature postulated is much more readily ascertainable. It will be
recalled that the stimulus there employed was “‘a graded series of abso-
lute intensities of opposed lights having their relative intensity main-
tained constant.’”’ Suppose that under an intensity ratio of three to
one between the opposed lights, the axial position of the animal is
changed until it attains an angle of orientation such that the amount
of light falling on the opposite sensitive surfaces is equal. The position
of the photoreceptive system would be that illustrated by the angle
edf in figure 7. As already pointed out, this equalization depends on
the fact that the sensitive surface toward the light of three units inten-
sity, (3 L) presents to the rays an area of one-third that presented by
the symmetrical sensitive surface of the opposite side of the animal
to the light of one unit intensity (1 LZ). Whatever change is made in
the absolute intensities of the opposed lights, as long as this ratio is
maintained, the same axial position will equalize the stimulus bilaterally.

We should obtain, therefore, for any intensities of fixed ratio to each
other, horizantal, rectilinear reaction curves. The positions of the
graphs for the conditions of these experiments were determined theoreti-
cally when the angles of orientation which would equalize the light
at given intensity ratios were computed (see fig. 6 and p. 334). We
therefore know the contour and position of the reaction graphs which
would be produced under these conditions by the action of the mechan-

REACTION OF BLOWFLY LARVA TO LIGHT 337

ism postulated. These curves plotted for comparison with the experi-
mental curves of figure 5, fall in the position of the solid, horizontal
lines.

By computing the average variability of the experimental data and
comparing it with the values of the fluctuation of the curves, the con-
clusion was reached (p. 320) “that the significance of each curve is a
horizontal line represented by the ‘mean line’ computed from the five
points of a curve.” The close conformity of the experimental and the
theoretical curves is so striking that it needs no further emphasis.

Conclusions from such conformity as has been demonstrated above,
must be drawn with extreme caution. The fact that a mechanism
such as postulated would produce both types of curves obtained ex-
perimentally is in no wise to be regarded as proof of the existence of
precisely such a mechanism in the blowfly larva. It does, however,
present a hypothesis on which further work may profitably be under-
taken. Moreover such a hypothesis offers a logical explanation of the
contour of a group of stimulus-reaction graphs for which there is no
other explanation available.

The most interesting phase of the results as far as the Weber-Fechner
Law is concerned is the conformity with its theoretic expectation shown
by the curves of figure 5. This agreement might easily have led to
entire misinterpretation of the results, had not a second set of meas-
urements been made which was clearly unconformable with the law.
As both series of data were obtained under identical experimental con-
ditions the conclusion is unavoidable that the conformity with Weber’s
Law, where it does occur, is entirely accidental. This does not neces-
sarily signify that a following of the Weber-Fechner Law could not
be demonstrated in the photic reactions of the blowfly larva, but that
the nature of the reaction mechanism involved in the response to op-
posed lights is such that it dominates the other factors involved in the
reaction. It is entirely possible that experiments made in such a way
as to exclude the action of this mechanism for obtaining a bilateral
balance of stimulation, might yield results in accord with the Weber-
Fechner Curve.

VI. SUMMARY AND CONCLUSIONS

We have in this paper dealt with experiments devised for measur-
ing the reactions of the blowfly larva so that they might be employed
for comparison with various values of the acting stimulus. The meas-
urements were so made and collected that they could be handled from

338 BRADLEY M. PATTEN

the same point of view as the available data on the comparison of
sensation value with stimulus intensity. Analysis of the results ob-
tained, with reference to the Weber-Fechner Law, brought out the
following conclusions:

1. The graph expressing the values of the angular deflections of the
blowfly larva from an original path of locomotion at right angles to
the line connecting the sources of light, produced by a graded series of
percentage differences of intensity between opposed lights cannot be
regarded as conforming to the Weber-Fechner Curve.

2. Graphs expressing the values of the angular deflections produced
by a graded series of absolute intensities of opposed lights having
their relative intensity maintained constant, may be regarded as in
conformity with the Weber-Fechner expectation.

3. The fact that two sets of data derived from the same animals,
by identical experimental procedure, yield opposite conclusions so far
as the Weber-Fechner Law is concerned, indicates that the agreement
with the law, where it does occur, is a purely fortuitous one, and that
the reactions of the blowfly larva to opposed lights are controlled by
other factors.

4. It is possible to demonstrate geometrically that a photoreceptive
mechanism operating after the manner of bilaterally located sensitive
surfaces inclined at an angle to each other and affecting the opposite
sides of the musculature in proportion to the stimulus they receive,
would determine both types of curves which have been plotted from
experimental measurements of the photic reactions of the blowfly larva.

BIBLIOGRAPHY

(1) Apams: This Journal, 1908, ix, 26. .

(2) FecHnEeR: Elemente der Psychophysik, 2nd ed., Leipzig, 1889.

(3) Hers: Journ. Exp. Zool., 1911, x, 167.

(4) Howe: Text-book of physiology, Philadelphia, 4th ed., 1912.

(5) Logs: Studies in general physiology. Chicago, 1905.

(6) Massart: Bull. Belg. Acad. Ser. ili, 1888, t. 16, 590.

(7) Masr: Light and the behavior of organisms, New York, 1911.

(8) Parren: Jour. Exp. Zool., 1914, xvii, 218.

(9) Water: Brain, 1895, xviii, 200.

(10) Watrer: Jour. Exp. Zool., 1907, v, 35.

(11) Weser: De tactu: annotationes anatomicae et physiologicae, Leipzig, 1834.
(12) Weper: Wagner’s Handwb. der Physiol., Braunschweig, 1846, iil, p. 481.

THE INFLUENCE OF THE EXTRACT OF THE POSTERIOR
LOBE OF THE HYPOPHYSIS UPON THE SECRETION
OF SALIVA

G. O. SOLEM anv P. A. LOMMEN
From the Laboratory of Physiology, The University of South Dakota

Received for publication July 16, 1915

Of the glands of internal secretion, probably one of the least under-
stood, though one of the more important is the hypophysis, least under-
stood, because of its inaccessibility and its seeming paradox in that the
anterior lobe is the vital portion, though its extract is physiologically
inactive, while the posterior lobe, although not essential to life, fur-
nishes an extract with marked physiological properties; important,
because of the changes associated with its pathology. The action of
the extract of the posterior portion of the pituitary body on the circu-
latory system has been fairly well established, but its effect on secre-
tion is still in dispute. Schafer and Herring (1) concluded that it has
a specific action on the renal epithelium, while Houghton and Merrill
(2) maintained that the diuresis observed is due to vascular changes.
Hallion and Carrion (3) noted a vasodilation following the primary
vasoconstriction of the kidney, and also that the diuresis followed this
vasodilation. Their results were verified by King and Stoland (4)
who were unable to obtain diuresis except during vasodilation in the
kidney. Frankl-Hochwart and Frélich (5) asserted that pituitary
extract has no influence on the chorda tympani nerve nor on any of the
secretory nerves to the salivary glands. Pemberton and Sweet (6)
found that pituitary extract inhibits the secretion of pancreatic juice,
supposedly by action on the cells. Edwards (6) maintains that no
inhibition occurs except what can be accounted for by the vasocon- _
striction. In view of these results it would seem that ample oppor-
tunity was offered for further work, and at the suggestion of our in-
structor, Dr. O. O. Stoland, we decided to conduct a series of experi-
ments to determine, if possible, the effect of pituitrin upon the secretion
of saliva, and whether this effect is due to vascular changes in the
organ or to a specific action on the cells or the secretory nerves.

339

340 G. O. SOLEM AND P. A. LOMMEN

On account of its accessibility and the fact that it has definite secre-
tory and vasodilator fibers from the chorda tympani and vasoconstrictor
and secretory fibers from the sympathetic, the submaxillary gland
seemed to lend itself more readily than any other organ to our work.
It appeared to us that the results obtained with this gland could be
more readily controlled and checked. The work done by Mathews
(7) on the effect of shutting off the blood supply to the submaxillary
gland was of prime importance to us, in that his results show that the
function of the gland was not impaired by shutting off the blood supply,
but that with restoration of circulation the flow of saliva returned.
He also showed that saliva would flow for almost a minute after the
blood supply was shut off. From this we may conclude that a imere
decrease in blood supply would not be sufficient to cause an almost
immediate cessation of salivary flow.

The experiments reported in this paper were carried out on healthy,
well-nourished dogs, varying from 4 to 20 kilos in weight. The animals
had no special preparation except that they were not fed on the day of
experimentation. Ether was used as the anaesthetic throughout and
was so administered by a tracheal cannula that a relatively constant
degree of anaesthesia was attained by the time injections were to be
made. Wharton’s duct was isolated, usually on the left side, and a glass
cannula inserted. The blood pressure changes were recorded by a
mercury manometer connected with the right carotid artery. The
external jugular vein and its branches were dissected out to expose
the venous return from the gland. In most cases this consisted of a
branch from the upper and anterior portion of the gland to the exter-
nal jugular just before it passes through the capsule of the parotid
gland and a branch from the posterior portion of the gland to the
external jugular or to a branch of the latter. Variations consisted
in having only one of these, most commonly the anterior branch.
All branches of the external jugular except those coming from the
gland were ligated, and a cannula, well paraffined and oiled, was in-
serted into the external jugular vein, and the blood flow through the
gland recorded by the drop method. The flow of saliva from the
gland was recorded by the same method. All injections were made
into the femoral vein towards the heart. Continuous tracings were
taken of the blood flow and salivary flow from the gland and the gen-
eral blood pressure except in the first few experiments where the flow
of saliva and blood pressure alone were recorded. These experiments
were performed as a control against later experiments where longer

INFLUENCE OF HYPOPHYSIS ON SECRETION OF SALIVA 341
<-

exposure of the gland was necessary. We found that exposure did
not alter the results. The pituitrin made by Parke, Davis and Company
was used exclusively in our work. After the animal was ready for in-
jection, we usually found that the flow of saliva was very meager, and
in order to stimulate a flow one of two methods were resorted to: the
chorda tympani nerve was stimulated peripherally by a weak faradic
current, or the flow was augmented by the injection of small amounts
of pilocarpine. When pilocarpine was used, pituitrin was not, as a
rule, injected until the blood pressure had again returned to normal.
In order to control our results, variations in procedure were frequent,
and the results will be given in the order in which the experiments
were performed, as far as possible. The same procedure was followed
until consistent results had been obtained. One cubic centimeter of
pituitrin was injected unless otherwise stated.
_ Throughout all our work on dogs one fact, about which there could
be no reasonable doubt or question because there was no exception,
stood out prominently, namely, that the intravenous injection of
pituitary extract invariably caused a marked decrease in the flow of
saliva from the submaxillary gland. From all our experiments we found
that the average slowing of saliva due to the action of pituitrin was 82.4
per cent in 47 injections, and that of these in only 13 was there aslow-
ing of less than 75 per cent. In many cases the flow of saliva ceased
altogether for a short time. In some animals more than one injec-
tion of pituitrin was made, but we failed to find in any of these cases
that repeated injections decreased its effect at all except in one experi-
ment where a large cat was used: In this case we found that the slow-
ing of saliva was decreased after the second injection of pituitrin, and
that there was hardly any slowing after the third. In all our work,
also, an accompanying diminution of the blood flow from the gland
was noted following every injection of pituitrin. Now the question
arose, was this great decrease in salivary flow due to changes in blood
supply to the gland, associated with the vasoconstriction caused by
pituitrin, or to its effect upon the secretory nerves or gland cells of
the organ, or to a combination of both of these? In order to analyse
the causes of this great diminution in salivary secretion, we averaged
the results obtained on the injection of pituitary extract during the
stimulation of the chorda tympani nerve, during the action of pilo-
carpine, during the action of adrenalin, and during the action of chryso-
toxin. The results are recorded in Table 1.

A series of three experiments was performed where the injection

342 G. O. SOLEM AND P. A. LOMMEN

.

TABLE 1

| BEFORE THE INJECTION | AFTER THE INJECTION OF
| OF PITUITRIN 1 cc. OF PITUITRIN
TREATMENT BEFORE AND NO. OF |
DURING THE PITUITRIN ANIMALS F = a aT
ACTION be eee Blood Saliva
| No. of siqaes
} (drops per | (drops per] - 3. | Blood Saliva
: : injections
min.) min.)
Faradization of chor- |
da tympani........ 3 84 21 Go| esas 4.6
Pilocarpine stimula- |
HOM... | 13 58.6 ies, 16 16 2.5%
Adrenalin stimula- |
GON ales ote fxs ols <2'9)| 4 21.6 Hee 4 9.8 125
Action of chrysotoxin 7 48 2 5 7 15.5 0.35

of pituitrin was made during the faradization of the chorda tympani.
Here it was observed that the slowing of saliva was 80 per cent in
six injections, while the blood flow decreased only 64 per cent. We
found that in most cases the blood flow almost stopped momentarily,
but later the flow increased greatly while the salivary flow continued
decreasing (see fig. 1). This would indicate that pituitrin acts upon

Li lh a ti i tn

BP rim anata

, Par)
voto SANNA MADE imac

a NN I LAA 0 On ee ee ee |

4 ASSESS ae

Fig. 1. The effect of pituitrin upon the flow of saliva and blood from the
submaxillary gland during light faradization of the chorda tympani. BP, blood
pressure; B, drops of blood from the venous return of the submaxillary gland;
S, saliva from Wharton’s duct in drops; X, 1 ce. of pituitrin injected.

INFLUENCE OF HYPOPHYSIS ON SECRETION OF SALIVA 343

the secretory mechanism of the gland by some other method than by
vasoconstriction alone, probably by blocking or decreasing the irritabil-
ity of the secretory fibers, for, as was noted above, the salivary secre-
tion slowed even when the blood flow was fairly rapid.

We observed that during pilocarpine stimulation the slowing of the
salivary flow, following the injection of pituitrin, was equal to that
obtained during chorda stimulation, but that while the decrease in
the blood flow during pilocarpine action was 73 per cent (see fig. 2),
that during faradization of the chorda was 64 per cent. As is well
known, pilocarpine stimulates the endings of the secretory nerves,
while excitation of the chorda tympani affects both the vasodilator

B.P.

i sil RO, anne

etsy aN

AT AALLALILLLLLALAAIILLLLLLLLI 110d LLL nn rr rr 2

Fig. 2. The effect of pituitrin upon the flow of saliva and blood from the
submaxillary gland during the action of pilocarpine. The letters are same as
in figure 1.

and secretory nerves, and hence the decrease in blood flow, due to the
action of pituitrin, would not be so marked in the latter case. Some
increase in blood flow was noted on stimulation with pilocarpine, but
was evidently the result of the greater activity of the gland rather than
to any action of this drug upon the vasodilators. From this we may
assume that pilocarpine acts in the same way upon the submaxillary
gland as does faradization of the chorda, with the added advantage
that in the former case the vasodilators are probably not affected. Now
the fact that the slowing of saliva following the action of pituitrin was
the same under two different kinds of stimulation, both involving
the secretory fibers, coupled with the fact that the accompanying

344 "G O. SOLEM AND P. A. LOMMEN

decrease in venous return was not the same because the vasodilators
were probably not stimulated to the same extent in one case, would
seem to indicate that this drug inhibits in some way the secretory
fibers to the gland.

Another result obtained with pilocarpine, seemingly the oppo-
site of the normal, was the great rise in blood pressure following the
injection of this drug during the action of pituitrin. Since pilocarpine
normally causes a slowing of the heart and a fall in blood pressure, we
are at a loss for a satisfactory explanation of this apparent paradox.
The heart beat was slowed, but its amplitude was in some cases markedly
increased.

TABLE 2

The results recorded in this table show the effect of pilocarpine upon the flow of
blood and saliva before and during the action of pituitrin and the effect of adre-
nalin upon the same under like conditions.

NO. OF NO. OF BLOOD SALIVA

Hes ETE ON ANIMALS INJECTIONS pen elk pre sale
Pilocarpine:: 22h o.eca aoe a 59.5 24
PRC URIs, 4-5. aparece oe 5 df 15.4 2A
Pilocarpine during action of

PUCOHTIN. suse ees ee 5 7 20.5 3.2
Before adrenalin............. 4 4 31 2
AdrGnalink ook oo aes 4 4 49 5.5
Pituitrin before injection of

adrenaline, 32.52.0555 a.00r. 7 8 12 2.8
Adrenalin following injection

Of pitrmitrmess sees 7 8 MATE 1.5

We found further that pilocarpine was relatively inactive during the
action of pituitrin. In this series of experiments (see Table 2) an in-
jection of pilocarpine was made to stimulate a rapid flow of saliva,
followed by an injection of 1 cc. of pituitrin, and at various stages of
pituitrin action a second injection of pilocarpine was made. Table 2
shows that although the second injection of pilocarpine causes a slight
increase in blood and salivary flow during the action of pituitrin, it
is relatively inactive, because the increase can be accounted for by the
partial recovery from the action of the pituitrin. In some of our experi-
ments it was noticed that there was no increase in the flow of saliva,
even though the blood flow increased a little. This would indicate
that pilocarpine may have some action on the vasodilators, or more

INFLUENCE OF HYPOPHYSIS ON SECRETION OF SALIVA 345

likely that the activity of the gland was increased without apparent
increase in the salivary flow, possibly, by increasing the organic con-
stituents. Since we made no quantitative determination of organic
matter in the saliva, we are unable to do more than merely to state
this possible explanation. In one experiment, where two successive
injections of pituitrin were made to stop the flow of saliva, the real
effect of pituitrin was probably shown better than in any other of this
series. During this action of the hypophyseal extract it was observed
that though the flow of blood was doubled after the first injection of
pilocarpine, the flow of saliva remained the same, and when the second
injection of pilocarpine was made eight minutes later, even then the
saliva increased in less measure than the flow of blood, which is con-
trary to the normal action of pilocarpine. At this time the action of
pituitrin should have been fairly well spent, for the blood pressure had
almost returned to normal. The above observations would seem to
show that the action of pituitrin is partly, at any rate, upon the secre-
tory fibers to the gland.

In a series of four experiments we attempted to show the effect of
shutting off the blood supply on the flow of saliva normally and after
the injection of pituitrin. Normally the salivary flow ceased in less
than one minute after shutting off the blood supply, and the same
result was obtained after injection of pituitrin, the pituitrin being in-
jected half a minute before the blood supply was shut off. This pro-
cedure failed to give any definite results. The only effect noted was
that the blood flow returned slightly more slowly, and the saliva flow
did not return for one minute, while normally both returned together.
This is borne out by the work of Mathews (7) who shut off the blood
supply for long periods of time (7 to 30 minutes).

In our experiments the blood supply was shut off for one minute
In one case the pituitrin was injected while the blood supply was shut
off for 1.25 minutes and when the circulation was restored, the flow
of saliva and blood both returned immediately but later slowed due
to pituitrin in the blood. This is of interest in so far that it shows the
peripheral action of pituitrin, because the nervous mechanism of the
gland was intact. Aside from this, shutting off the blood supply fur-
nished no other evidence than that we had already obtained by other
methods, and therefore it was abandoned.

Comparing the average percentage of slowing of blood and saliva
in 33 injections of pituitrin (see Table 1), we noted that the flow of
saliva decreased 81.8 per cent, while the slowing of blood was 65 per

346 G. O. SOLEM AND P. A. LOMMEN

cent. This in itself would point to the probability that the decrease
is not due entirely to vasoconstriction but to some other factor also.
Can it be attributed to the inhibitory effect on the secretory fibers
or to the toxic effect on the gland? The latter would be ruled out by
the general lack of toxicity of pituitrin. This leaves us the hypothesis
that the inhibition of secretion is practically due to the action on the
secretory fibers to the gland, especially in view of the ineffectiveness
of pilocarpine during the action of pituitrin.

In order to obtain further evidence as to the influence of pituitrin
upon. secretion, it was concluded to try the effect of epinephrin, the
great general vasoconstrictor, which acts on the endings of the vaso-
motor nerves. However, we found that this drug invariably caused
a vasodilation in the submaxillary gland, while effecting a general
vasoconstriction. Epinephrin then presumably stimulates the vaso-
dilator endings in this gland. Epinephrin also increased the flow of
saliva; in some cases we used it to start the secretion (see Table 1).
In four injections where adrenalin was used to initiate the secretion,
the salivary flow decreased 72 per cent, while the blood flow decreased
55 per cent after the injection of pituitrin, thus practically paralleling
the results obtained during faradization of the chorda tympani (see
Table 1), a fact which would oa the hypothesis that i acts
upon the secretory fibers. ;

From Table 2, we note that’ when epinephrin was injected during
the action of pituitrin, it caused an increase in the flow of blood, but
a decrease in the flow of saliva (see fig. 3). This diminution in salivary
secretion during vasodilation in the gland must be due to the increased
amount of pituitrinized blood passing through the gland on account of
the action of epinephrin, thus augmenting the effect on the secretory
fibers. As far as we can see, the above evidence alone would go far
to prove our case.

In the hope of abolishing or decreasing the action of the vasocon-
strictors to the gland, we made injections of Jacobi’s chrysotoxin (8)
and followed this by injections of epinephrin and pituitrin. This
chrysotoxin was made by extracting powdered ergot with ether, evapo-
rating off the ether, and precipitating the oily extract with light petro-
leum. The precipitate was dissolved in physiological salt solution and
then injected. No effect was noted on blood pressure or salivary
flow, but with one preparation of chrysotoxin especially we found a
marked increase in the amount of venous return from the gland, which
may be attributed to the depression of the tene of the vasoconstrictors

INFLUENCE OF HYPOPHYSIS ON SECRETION OF SALIVA 347

to the gland. The injection of adrenalin during the action of this
ergot preparation caused only a slight primary rise in blood pressure,
followed by a fall, or in some cases there was no change at all in the
blood pressure noted; in one case we got “vasomotor reversal.’”’? The
flow of blood and saliva after adrenalin was unaffected to any marked
degree, though in some instances a slight slowing of both was observed
during the primary rise.

A further study with the method of preparation and properties of
the ergot preparations since the completion of our experiments has led
us to believe that the material used in our work was somewhat lacking
in strength and purity; its action, however, was sufficient for our in-

Fig. 3. The effect of epinephrin upon the flow of saliva and blood during
the action of pituitrin. The letters are same as in figure 1; Y, injection of 2 cc.
1-50,000 epinephrin.

tentions. Having found that chrysotoxin inhibited the action of the
vascoconstrictors to a marked degree, we were ready to try the effect
of pituitrin during the action of chrysotoxin. We found that normally
the vasoconstriction after injection of pituitrin was almost invariably
preceded by at least a slight fall. When pituitrin was injected during
the action of chrysotoxin we observed that the primary fall in blood
pressure was, as a rule, greater than normal, while the secondary rise,
though characteristic, was in most cases less marked. It was noted that
the flow of saliva diminished before the decrease in blood flow was
noticeable. In all but two cases the salivary slowing was noted while
there was an increased output of blood from the gland (fig. 4). Dur-

348 G. O. SOLEM AND P. A. LOMMEN

ing the action of this drug the vasoconstriction after pituitrin was much
less marked in most cases, the blood flowing fairly rapidly, while the
decrease in flow of saliva was as great as or greater than that obtained
in our other experiments.

The results obtained with chrysotoxin constitute our strongest
evidence that pituitrin acts by inhibiting the secretory fibers. A
number of workers have attributed the vasoconstriction caused by
pituitrin to a direct action on the muscles of the arterioles; our experi-
ments with chrysotoxin would indicate that whatever the true seat

sa MN
ea gagil id

Fig. 4. The effect of pituitrin upon the flow of saliva and blood from the
submaxillary gland during the action of chrysotoxin. The letters are same
as in figure 1. The chrysotoxin was injected about 2 minutes before the injec-
tion of pituitrin and the dose had previously been shown to be sufficient to de-
press the irritability of the vasoconstrictors.

of action may be, this substance depresses the sensibility of the mechan-
ism affected by pituitrin. Pituitrin probably acts upon the endings
of the vasomotor nerves as well as upon the involuntary muscles.

SUMMARY AND CONCLUSION

1. Pituitrin’ invariably caused a diminution in flow of blood and
saliva from the submaxillary gland, as shown from results obtained
from 30 dogs and 1 eat.

2. The decrease in flow of saliva was greater than the accompany-
ing decrease in blood flow.

INFLUENCE OF HYPOPHYSIS ON SECRETION OF SALIVA 349

3. The slowing of blood was less marked if the injection was made
during faradization of the chorda tympani than during pilocarpine-
stimulation, while the slowing of saliva was the same.

4. Pilocarpine was relatively ineffective even when injected seven
or eight minutes after pituitrin.

5. While epinephrin normally caused a vasodilation of the gland
and increase in salivary secretion, epinephrin during the action of pitui-
trin had the normal effect on the blood flow but caused a diminution
in salivary flow, probably due to the greater quantity of pituitrin com-
ing in contact with the gland.

6. When pituitrin was injected during the action of chrysotoxin,
the decrease in the flow of saliva set in before the vasoconstriction in
the gland occurred. In five out of seven cases, the flow of saliva
slowed while there was active vasodilation in the gland.

From these results we must conclude that the decrease in flow of
saliva following the injection of pituitrin is due to inhibition of the
action of the secretory nerves to the submaxillary gland, but also due
in part to the accompanying vasoconstriction, which is caused by direct
action on the muscles of the arterioles or the effect on the peripheral
endings of the vasomotor nerves, but more probably to the effect on
both. The decrease in output of blood from the gland may be also
due to the decreased activity of the gland.

We wish to express our thanks to Dr. O. O. Stoland for his valuable
suggestions and criticisms in performing this series of experiments.

BIBLIOGRAPHY

(1) Quoted from Cushing: The pituitary body and its disorders, 1910.

(2) HovucHtTon anv Merriv: Journ. Amer. Med. Assoc., 1908, 1849.

(3) Quoted from Bied1: Internal secretary organs, 1913, 324.

(4) Kine anp Srouanp: This Journal, 1913, xxxii, 412.

(5) Quoted from Bied1: Internal secretory organs, 1913, p. 326.

(6) Quoted from Wiggers: Physiology of the pituitary gland and the action of
its extracts; collected papers from the Research Laboratory, Parke,
Davis & Co., 1913, i, 79.

(7) Matuews: This Journal, 1901, iv, 16.

(8) AttEN: Commercial organic analysis, fourth edition, vii, 16.

THE INFLUENCE OF THE VAGUS NERVE ON THE
GASEOUS METABOLISM OF THE KIDNEY

ROY G. PEARCE anp EDWARD P. CARTER

From the Physiological Laboratory of Western Reserve University
Received for publication July 24, 1915

Asher and Pearce (1) have reported a series of experiments the results
of which they interpret as demonstrating the presence of renal secre-
tory fibers in the vagus nerve. In these experiments, which were
done on decerebrate animals, the urine obtained during a period of
vagus stimulation exceeded in amount that obtained during a normal
period of equal duration. The technical difficulties of the operation
made it impossible to carry out more than a few successful experiments,
so that a relatively small number of results are reported.

To secure additional data in this connect on is the object of the pres-
ent research, and for this purpose we have not depended on the rate
of urine formation, but have sought for evidence of changes in the
functional activities of the renal cel's. To do this we have taken ad-
vantage of the fact that every increase in tissue activity is accompanied
or followed by an increase in oxygen consumption. Barcroft (2)
and his co-workers have established such a relation in the case of the
kidney, their researches having shown that two types of diuretics
may be distinguished: one whose diuretic action is accompanied and
the other unaccompanied by an increase of oxygen consumption. It
may be considered that the action of the former type of diuretics (urea,
caffein and sodium sulphate), is due primarily to an actual stimulation
of the excreting cells of the kidney, while the action of the latter type
(sodium chloride, ete.) depends on an alteration in the osmotic pressure
of the blood, produced by the injection of these salts into the body.
This relationship being established, it follows that an estimation of
the oxygen absorption of the kidney will be an index of the activity
of the renal cells under varying conditions.

The diuresis produced by vagus stimulation must be a result of
increased renal activity and for two reasons: ‘viz., Ist, it is impossible
for a nerve to change directly the osmotic pressure of the blood; 2nd,

350

INFLUENCE OF VAGUS ON GASEOUS EXCHANGE OF KIDNEY 351

the blood flow through the kidney does not become altered under the
conditions of the above described experiment (3).

An increase in oxygen consumption by the kidney should accompany
the stimulation of the vagus if this nerve really contains renal secre-
tory fibers. To test this an estimation was made of the oxygen intake
of the kidney under the conditions of Asher and Pearce’s experiment;
that is to say, the percentage difference between the oxygen af the
arterial and venous bloods was compared. The reflex disturbances
arising from the manipulation of the kidney or ureters is discounted
by comparing the changes occurring in the stimulated kidney with
those obtained in the opposite kidney, which was previously denervated.
The fact that the vagus does not influence the circulation of the kid-
ney (3) makes the estimation of the volumes of the blood flow unneces-
sary in measuring the oxygen consumption, and makes it possible to
determine this by comparing the percentage amounts of oxygen in the
arterial and venous bloods before and during vagus stimulation.

METHODS

The preparation of the animal conformed in all essentials to that
followed by Asher and Pearce. Dogs were etherized and quickly
decerebrated. The vasomotor control exercised on the renal blood
vessels through the splanchnic nerves was removed on the left side by
cutting the nerve just above the adrenal gland. The kidney on the
opposite side (right) was at the same time completely removed from
the nervous system by severing the nerves surrounding the renal vessels
and searing the vessels with concentrated phenol solution. Electrodes
were placed on the vagi below the level of the heart, and the nerves cut
in the neck. The viscera were retracted and held in such a manner as
to expose the renal veins on both sides. Samples of 1 cc. each of blood
were removed from the renal veins and from the carotid artery through
fine needles into hypodermic syringes containing 7’) ec. of hirudin dis-
solved in boiled Ringer’s solution. After removing the blood, the
syringe was sealed by sucking up a small quantity of mercury and
capping the end, from which the needle had been detached, with a
closed rubber tube. The oxygen in the various samples was estimated
by the so-called differential method as described by Barcroft for deter-
mining the total oxygen in 75 cc. of unsaturated blood (4). Artificial
respiration was maintained throughout the experiments and every
precaution was taken to keep the blood pressure at a constant level.

352 ROY G. PEARCE AND EDWARD P. CARTER

In most of the experiments this was accomplished fairly well. Twenty-
five to thirty minutes elapsed between the decerebration and the taking
of the blood samples. The possible disturbing influence of anaesthesia
was thus eliminated.

The procedure followed in obtaining the samples of blood was the
same in all the experiments. The first samples of blood were removed
from the renal veins of both kidneys and from the carotid artery, and
were used to determine the normal gas exchange. Immediately after-
wards the vagi were stimulated with a current of approximately the
same strength as that used by Asher and Pearce, for alternating periods
of one minute with one minute intervals between. During the fourth
minute of stimulation the second samples of blood were removed as
above described and were used to determine the exchange occurring
during stimulation.

The oxygen percentages of the venous bloods were compared with
those of the arterial bloods of the corresponding periods, arid the num-
ber of cubic centimeters of oxygen absorbed per cubic centimeter of
blood calculated in each case. The right kidney, since its nerves
had been destroyed, could not be affected directly by stimulation
of the vagus nerve. Its oxygen consumption should therefore be con-
stant, or if any change was found in its oxygen intake, it must be
sought in some systemic condition which would effect the left kidney

in the same degree.

EXPERIMENTAL RESULTS

In the accompanying table the results of the last four experiments
are tabulated. The figures in the third column give the oxygen con-
tent of the arterial blood, and those in the fourth and fifth columns the
oxygen content of the venous blood of either kidney before and during
a period of vagus stimulation. The sixth and seventh columns give the
actual cubic centimeter of oxygen taken from 1 cc. of blood by either
kidney, and the last two columns, the percentile change in oxygen
consumption which accompanies the stimulation of the vagus.

Experiment No. 5 isthe only one of the series in which there is a greater
amount of oxygen used during a period of stimulation of the vagus
than can be attributed to systemic conditions as reflected in the secre-
tion of the right denervated kidney. In all of the remaining experi-
ments the change in oxygen consumption in the left kidney accompany-

INFLUENCE OF VAGUS ON GASEOUS EXCHANGE OF KIDNEY 353

ing vagus stimulation is less than that found in the case of the denervated
right kidney.

The operative procedure of these experiments, as in those of Asher
and Pearce, is very severe, and it is difficult to maintain the animal
in an unchanging condition in regard to the circulation, respiration
and renal function. The tables show that in spite of pains taken to
maintain the normal respiratory function, there was a marked decrease
in the oxygen content of the arterial blood during the stimulation of
the vagus. The exact cause of this change is not at present clear.
The actual oxygen consumed by each kidney, however, remained prac-
tically constant, as shown by the figures in columns six and seven of
the table.

"In the experiments reported by Barcroft and Brodie, in which the
effects of certain diuretics on the urine flow and oxygen consumption
of the kidney were investigated, an increase in urine excretion was
accompanied by a like change in oxygen used by the kidney. These
investigators were unable to estimate the actual amount of oxygen
necessary for the excretion of a fixed amount of urine, but the experi-
ments show that it is of such magnitude that the oxygen intake can be
made a test of the physiological activity of the kidney.

The failure to obtain evidence of an increased gaseous exchange
following stimulation of the vagus nerve under the conditions reputed
to produce a diuresis, as in the above series of experiments, can be in-
terpreted by assuming, either that the diuretic effect failed to appear,
or that it occurred independently of any energy change in the kidney.
It is to be regretted that the experimental procedure is such that si-
multaneous investigations of the gaseous exchange and urinary outflow
are not practicable. It is, however, reasonable to suppose that in a

series of eight experiments, some would show a change in oxygen
consumption if the vagus nerve exerts any marked secretory influence
on the kidney.

SUMMARY

The oxygen consumption of the left kidney before and during the
stimulation of the vagus below the level of the heart and after section
of the left splanchnic nerve, remains unchanged. This offers evi-
dence against the supposed existence of renal secretory fibers in the
vagus nerve. ;

354 ROY G. PEARCE AND EDWARD P. CARTER

REFERENCES

(1) AsHpr anD Prarce: Zeitschr. f. Biol. 1913, xiii, 83.

(2) Barcrorr: The respiratory function of the blood. Cambridge University
Press, 1914.

(3) Pearce, R. G.: This Journal, 1914, xxxv, 151.

(4) Barcrort: Loc. cit.. p. 302.

‘wu QT] 8S01 ounssoid
poojq ey} ‘uotR_NUIyS snsvAa
jo out} 48 poorq jo ojdwes Sut
-ye} JO oT} oY} SulINg ‘yuoUT
-140dx9 JO SuluUIseq ye “UU 08 0s00°'0 | S910°0
:eimssoid poor O8— 00=

“WIU 08 0670°0 | S8E0°0
:ainsseid poojg | ZI— 00

0gc0'0 | S8€0°0

‘quoulliodxo Jo pue 4B ‘SPT
‘UU ¢, pue SuruuIseq ye “UIUT Cs Orr0'0 | g¢c0'0
:aimssoid pool et+ Gg+

quoul

-l10dx9 ynoysno1iyy “SPT “UIUT CET G670'0 | LIF0'0
:aanssoid poojg | 2a+ L+

Aoupry Aoupry Agu ply AoUpLy
Gael Ware IPT Vay

SaUVNaAY
‘WILS SOOVA WO

aoiad ONIWAG NOW
-dWOSNOOD NGDAXO NI
GONVHO FILLNGONAd

INFLUENCE OF VAGUS ON GASEOUS EXCHANGE OF KIDNEY 355

aoold #0 *00 [
uUdd NOIdNOSNOO
N@SAXO FO “00

OOTT'O | 0660°0 | SSTT'0

ITét'O | OOTT'O | OS9T'O

OcEL'O | O&FT'O | LPST'O

Og9T'O0 | OS9T'O | SE0z2'0

([eyueULed] (peyeAsou
-xo) Aoupty|-op) oupry aide
dec WIN : :

daoold #0
‘09 [ Ud NADAXO AO “9090

uoly
-G]NUITYS
SNnSB A,

[BULO Ny
wor}
-e[NUITYS
snsvA
[BuIIO Ny
won

-B[NUIT}S
SNsvA

[BULLION
u0ly
-V[NUIT}S

Snsv A

[BULLION

1Z Avy

aqoimdd GNAWINAdxXa

a a RY

1 ATAVL

CHANGES IN IODINE CONTENT OF THE THYROID
GLAND FOLLOWING CHANGES IN THE BLOOD
FLOW THROUGH THE GLAND

C. F. WATTS
From the Hull Physiological Laboratory of the University of Chicago

Received for publication July 24, 1915

This work was undertaken at the suggestion of Dr. Carlson, primarily
to determine the influence of changes in the circulation of the thyroid
gland on the iodine content of the gland. This point does not seem to
have been controlled in the recent interesting work of Rahe, Rogers,
Fawcett and Beebe, (1) on the effect of stimulation of the thyroid
nerves on the iodine content of the gland. It seems clear that the
data reported by Rahe, et al, cannot be construed as proving the
presence of secretory nerves to the thyroids, until it is shown that the
vaso-motor changes in the thyroid induced by stimulation of the
thyroid nerves are not by themselves capable of inducing the changes
in the iodine content.

Il. LITERATURE

According to Rhinehart (2) the nerve supply to the thyroid is made
up of filaments from the cervical sympathetic, entering the gland with
the perivascular connective tissue and tunica adventitia of the thyroid
arteries. In man and the cat the filaments have been demonstrated
to arise only from the superior cervical ganglion of the sympathetic.
In the dog some observers claim that nerve filaments from the vagus
also enter the thyroids. This is denied by Rhinehart. In the thyroid
glands the nerves form elaborate perivascular nerve plexuses and end,
at least in part, in the walls of the blood vessels.

Ossakin (3) gives a brief review of the literature on the vaso-motor
and possible secretory function of the thyroid nerves. Brian stimulated
the vago-sympathetic and inferior laryngeal nerves and studied the
histological and vaso-motor changes produced in the thyroids. He
concluded that these nerves carry both secretory and vaso-motor fibers.
The vaso-motor function has been described as the sole function of the

356

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 357

thyroid nerves by Sacerdatti. V. Cyon reports experiments showing
on stimulation of the nervus thyroideus superior, a slight diminution in
the carotid blood pressure with a change in the volume of the gland.
By measuring the blood pressure in the arteriathyroidea superior V.Cyon
concluded that both vaso-constrictors and vaso-dilators are contained
in the laryngeal nerves, the constrictors predominating in the superior.

Morat and Sinakewitsch by plethysmography showed that stimu-
lation of the cervical sympathetic gives a pressor effect—a decrease
of the amount of blood in the thyroid gland. The latter investigator
showed that on extirpation of the superior cervical ganglion or paralysis
of this ganglion with nicotine no effect on the thyroid was obtained on
stimulation of the sympathetic. From this he concluded that the sym-
pathetic fibers are synapsed in the superior cervical ganglion before
reaching the thyroid gland.

Katzenstein reports marked degenerative changes in the thyroid
on cutting the thyroid nerves, and concludes that the nerves contain
true sensory fibers. These degenerative changes on cutting the nerves
are attributed solely to changes in the blood supply by Luebeke.

Wyss and Andersson failed to obtain histological change in the thy-
roid following injection of pilocarpine.

Huerthle stained the thyroid gland after stimulation of the layrn-
geal nerves and found no change. He concludes, therefore, that these
nerves are not secretory but vaso-motor.

Martini by stimulating and Horsley by cutting the thyroid nerves
report no change in the histology of the gland.

Wiener cut the vago-sympathetic and paralyzed the superior cervical
ganglion reporting no histological change in the thyreoglobulin content
of the gland. But when the inferior cervical ganglion was removed
- leaving the superior intact he reports a decrease in the weight of the
gland and decrease in the thyreoglobulin content, from which he con-
cludes that the thyroid nerves have both a trophic and secretory func-
tion.

Asher and Flack (4) working on rabbits, cats and dogs compared
excitability of the depressor nerve before and after stimulation of
the thyroid nerves, and also the effect of a small intravenous injection
of adrenalin before and after such stimulation. They conclude that
under exactly similar experimental conditions the excitability of the
depressor nerve is increased during or immediately following the stim-
ulation of the thyroid nerves. They also report that this stimulation
of the thyroid nerves sensitizes the animal to epinephrin. They as-

358 Cc. F. WATTS

cribe these effects to the action of an increased quantity of thyroid
secretion poured into the blood on stimulation of the thyroid nerves.
Stimulation of the thyroid nerves failed to produce these effects after
the extirpation of the thyroid gland. They also claim that intravenous
injections of thyroid extract act similarly to stimulation of the thyroid
nerves.

The work of Asher and Flack was repeated and apparently confirmed
by Ossakin (3) working in Asher’s laboratory.

According to the most recent work of Rahe and co-workers (1) stimu-
lation of the nerves to one thyroid lobe decreases its iodine content as
compared with that of the control lobe. They conclude that the
physiologically active substance of the thyroid is discharged into the
circulation on stimulation of the thyroid secretory nerves.

II. EXPERIMENTAL METHODS

All of the experimental work was done on dogs.

Analyses were made on the glands of ten normal animals to deter-
mine the relation of the iodine content in the two lobes. The glands
were analyzed in duplicate, for iodine content according to Hunter’s
method (5), modified according to Koch (6). Part of each lobe was
dried to constant weight for total solids.

Two series of stimulation experiments were run to determine whether
the stimulation of the nerves leading to one lobe causes a diminution
in its iodine content. The stimulation was applied to the cervical
sympathetic, isolated from the vagus sheath, and to the nerve filaments
accompanying the superior thyroid vessels.

The control and stimulated, lobes were, in a number of cases, stained
for the active secretion to determine whether the stimulation produced
histological changes parallel with the chemical changes.

The vascular changes produced by stimulation of cervical sympathetic
and the thyroid nerves, were studied both by the rate of the venous
outflow and by the plethysmographic methods.

A series of ten animals was run in which the blood supply-to one lobe
was mechanically interfered with to an extent comparable to that pro-
duced on stimulation of the nerves. The main thyroid artery was
cleaned of all connective tissue and accompanying nerve filaments.
The blood supply was momentarily, (1-2 seconds), cut off, by compres-
sing the artery for two seconds at ten second intervals.

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 359

Dogs with slightly enlarged thyroids were selected provided the two
lobes were very nearly the same size. Having the pick of a large
amount of material it was not difficult to select dogs of this type.
Glands of this type were chosen in order to have a sufficient amount
of material to run duplicate iodine determinations and to dry a por-
tion for total solids.

The details of the analytical procedure are as follows: The fresh
glands were carefully trimmed of all connective tissue and blood ves-
sels and accurately weighed; they were then hashed in a lead-pan,
re-weighed in the pan from which a portion was weighed into a care-
fully weighed glass dish to be dried for total solids, and the two other
portons weighed into nickel crucibles and covered with 6 grams of
pulvierized sodium hydrate and 2.6 grams of postassium nitrate. The
crucibles were then placed on the steam bath for a few hours, the ma-
terial going into solution in its own liquid. The crucibles were placed
in the drying oven and heat applied slowly at 110°C. running up to
140° to 150°C. A little experience made it possible to remove the
material at a stage at which it was easily pulverized. It was then
mixed with 6.4 grams of fusion mixture and covered with 10 grams of
the same, according to Hunter. From this point on the method as
modified by Dr. Koch (6) was followed exactly. Care was taken to
dry the material to constant weight in order to determine whether
the diminution in iodine on stimulation calculated on the basis of
fresh tissue is also true when calculated on the basis of dried substances.

In the stimulation experiments both lobes were left in situ until the
end of the period of stimulation in every case. We cannot consider a
lobe removed before three hours of anesthesia a control of the one
removed at the end of this period since we have no knowledge of the
nature or extent of the changes which may take place during this time.
Light ether anesthesia was used in every case. The nerves were siim-
ulated by brief periods of weak tetanizing current. The effectiveness
of the stimulus was controlled by the effect on the pupil, the current
used being 3-4 times stronger than the minimum for pupil dilatation.
Gruber (7) has shown that the threshold stimulus for the vaso-constric-
tion function of the cervical sympathetic is 6.95 B units while that for
the pupil dilation is only 2.27 B units.

Dr. R. R. Bensley has devised a new stain for the thyroid secretion.
Portions of several of the stimulated and control lobes were prepared
in Dr. Bensley’s laboratory and stained by him with a special stain for
the thyroid secretion.

360 Cc. F. WATTS

IV. RESULTS

Our results in the control series of normal dogs are summarized in
Table I. It is evident from this table that the iodine content per
gram varies widely in different individuals, but zs practically identical
in the two thyroid lobes of the same animal. The average difference
between the two lobes in the case of the fresh tissue was 0.00157 per
cent; in case of the dried tissue 0.01488 per cent.

According to Marine (8) the iodine content per gram is practically
identical in the two thyroid lobes in dogs.

It is, therefore, evident that the per cent of iodine in one lobe of the
thyroid will serve as control of the other, provided one is not dealing
with diseased glands.

In the first series of stimulation experiments the cervical sympathetic
was stimulated after isolation from the vagus sheath. The data are
given in Table II. The average loss in per cent fresh tissue, of iodine
in the stimulated lobe was 0.00794 but in Dog I, in which the per cent
of iodine was highest, the loss on stimulation was 0.01499 per cent per
gram of thyroid or approximately 12 per cent iodine. Dogs I-V were
fed about 25 grains of potassium iodide per day for three days previous
to the experiment. This was done to insure the presence of iodine in
the gland and was discontinued after coming to the conclusion that with
the method of Hunter iodine could be detected in every gland even
though we were dealing in many cases with hyperplasia. It is interest-
ing to note that in Dog I where the loss was greatest the stimulation
period was 3? hours; in Dog XII where the loss was least the stimu-
lation lasted only 13 hours. Dog XI did not give a decrease on stim-
ulation, the probable explanation being that we stimulated the lobe
with the greater per cent of iodine and the loss was not enough to make
it less but only tending to equalize the amounts. On the whole our
results show the greater decrease in the iodine content the longer the
period of stimulation of the thyroid nerves.

In the second series the thyroid nerves were stimulated near their
entrance into the gland with the superior thyroid artery. The blood
supply was not mechanically interfered with. The results of this
series are given in Table IE. It will be seen that the stimulation
caused a decrease in the iodine content in every case except Dog II.
The average iodine loss per fresh tissue, excepting Dog IT is 0.00510
per cent. The average loss calculated for dry tissue is 0.03516 per cent.

The results of these two series confirm those of Rahe, et al, but the

|
:

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 361

percentage loss is not as great as reported by them. It is apparent from
Tables II and III that there is in nearly every case a greater per cent
of total solids in the case of the stimulated lobe. Of course the per
cent of total solids varies somewhat in the two lobes of the same animal
in the normal series, (Table I), but it is obvious that we could not have
chanced in every case, to stimulate the lobe with the smaller water
content.

As stated before, the vascular changes, produced in the thyroid on
stimulation of the thyroid nerves, were studied by the rate of venous

Fig. 1. Showing vaso-constriction in the thyroid gland on stimulation of
the cervical sympathetic nerve. A, carotid (left) blood pressure. B, pehethys-
mograph record of the right thyroid lobe. C, showing stimulation of right cer-
vical sympathetic nerve. D, time 3} seconds.

ft Ser tp thi tit did pi di bE a tes ni ie

yt YENVS C8] fie. drePS

Fig. 2 Retardation and subsequent acceleration of the venous blood flow
from the thyroid vein on momentary stimulation of the cervical sympathetic
nerve. A, record of drops of venous flow. B, stimulation of cervical sympathetic
nerve. C, time 33 seconds.

outflow and by the plethysmograph. The cervical sympathetic was
ligated and stimulated centrally with the same strength of current as
in our stimulation experiments with the iodine determination. The
results with the first method are summarized in Table IV (see also fig. 2).
Even a stimulation lasting only 1—2 seconds cause a distict diminution
in the venous outflow (exp. 8, 9). In some cases the vaso-constriction
during the stimulation is followed by vaso-dilation on cessation of
stimulation (fig. 2). The plethysmograph shows marked decrease in
the volume of the gland on stimulation of the cervical sympathetic
or the thyroid nerves (fig. 1). Our experiments then confirm the work

362 Cc. F. WATTS

of Ossakin (3). The cervical sympathetic supplies vaso-constrictor
fibers to the thyroid.

A series of ten experiments were run in which the temporary and
periodic decrease in the blood flow through the thyroid gland was
produced mechanically and without any stimulation of the thyroid
nerves as previously described. The iodine and total solid determina-
tion of the thyroids in this series are given in Table V._ In every case
there is a diminution in the percent of iodine in the lobe in which the
blood flow had been periodically decreased. The decrease in iodine
is not as great as in the series of stimulation of the thyroid nerves
(Tables II and III). Control experiments showed that by one method
of periodic compression of the superior thyroid artery there was less
diminution of the blood flow through the gland than on stimulation of
the thyroid or the cervical sympathetic nerves. In 80 per cent of the
cases in this series the per cent of total solids is greater in case of the
lobe whose blood flow was periodically diminished. It is clear that
temporary vaso-constriction of the thyroid vessels, or temporary de-
crease in the blood fiow through the thyroids induced by momentary
compression of the thyroid arteries and without stimulation of the
thyroid nerves produce the same changes in the thyroid gland as does
the stimulation of the thyroid nerves, that is, a decrease in the iodine
and water content. Since the variations in the blood flow is by itself
capable of inducing these changes, and since these vascular changes
are produced in the gland on the stimulation of the thyroid nerves,
it is obvious that these effects of the nerve stimulation do not prove
the presence of true secretory nerve fibers to the thyroid glands.

Dr. Bensley’s special stain showed no difference between the con-
trol and stimulated lobes of the same gland.

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 363

TABLE I
Control series, (Dog)
Iodine content of right and left thyroid lobe of same animal

AV. DIFF. IN

LOBE | wr. oF| % oF % or i %or1 |%or1in2 oie
NO. OF ANIMAL or | FRESH] TOTAL | IN FRESH | IN DRIED ROBE, 3) O8 ee
GLAND | GLAND | sotips| TISSUE TISSUE | FRESH TIS-
pak TISSUE
iL. 2 Re ee R 38.2 | 20.5 | 0.03482 | 0.16335 | 0.00346 | 0.01686
L_ | 36.5 | 20.8 | 0.03778 | 0.18021 |
R 26.0 | 19.5 | 0.01737 | 0.08914 | 0.00153 | 0.00850
ihe. L 19.1 | 18.5 | 0.01584 | 0.08564
R .5 | 24.2 | 0.30336 | 0.13787 | 0.00212 | 0.00982
LD fo 0 So L 15.1 | 24.4 | 0.31024 | 0.12805
R 10.7 | 27.6 | 0.08886 | 0.32197 | 0.00398 | 0.06787
LD Ae i L 10.1 | 33.8 | 0.08588 | 0.25410
R 10.5 | 30.6 | 0.04172 | 0.13625 | 0.00172 | 0.02767
Va 6 ee eee L 6.6 | 26.5 | 0.04344 | 0.16392
R 19.2 | 25.0 | 0.00636 | 0.02548 | 0.00031 | 0.00078
Wilk iS i 18.9 | 27.2 | 0.00667 | 0.02470
R 23.0 | 20.4 | 0.00732 | 0.03593 | 0.00019 | 0.00352
Wiles eee L 18.6 | 22.0 | 0.00714 | 0.03341
R 9.2 | 20.4 | 0.00666 | 0.03265 | 0.00108 | 0.00605
VLG ee ee See L 10.2 | 20.4 | 0.00774 | 0.03870

Ne ee L 41.8 | 20.3 | 0.00778 | 0.03839

R_ | 49.7 | 22.0 | 0.01284 | 0.05838 | 0.00111 | 0.00744
21.2 | 0.01395 | 0.06582

Average differences: | 0.00157 | 0.01488

364 Cc. F. WATTS

TABLE II
(Dog)

Effect on iodine content of thyroid of stimulation of the cervical sympathetic nerve

(A) (a)

: 4 WT. OF| % oF % or % LOSS IN | LOSS IN

sorta | “cuawp | gent | TOM. |v reese [rw omunn | “eeeee | Zo Damp

L|L. stim. | 14.4 | 29.2 | 0.12776] 0.43715] 0.01499] 0.04640
R. cont. | 14.2 | 29.5 | 0.14265] 0.48355

IL|L. stim.| 9.4 | 23.8 | 0.03609] 0.15167] 0.00717] 0.05800
R. cont.| 8.3 | 21.0 | 0.04326] 0.20967

IIL | R. stim. | 13.2 | 25.4 | 0.06072] 0.23905] 0.01364] 0.10603
L. cont. | 14.2 | 23.0 | 0.07436] 0.34508

IV | R. stim.| 12.5 | 23.1 | 0.04034] 0.17469] 0.01179] 0.07356

L. cont. | 13.9 | 21.0 | 0.05213] 0.24825

V|R. stim.| 19.6 | 30.1 | 0.05498] 0.18267] 0.00502! 0.04037
L. cont. | 20.4 | 26.9 | 0.06000] 0.22304

VI|R.stim.| 15.7 | 22.3 | 0.01664] 0.07462] 0.00225] 0.01895
L. cont. | 16.5 | 20.2 | 0.01889] 0.09357

VII | L. stim. | 31.0 | 23.2 | 0.03175] 0.13680] 0.00612] 0.03459
R. cont.| 31.4 | 22.1 | 0.03787! 0.17139

VIII | R. stim.| 38.3 | 28.7 | 0.03746! 0.13052] 0.00461] 0.00607
L. cont. | 31.8 | 30.8 | 0.04207] 0.13659

IX | L. stim. | 10.8 | 24.3 | 0.02693] 0.11087] 0.00418] 0.04153
R. cont.| 8.1 | 23.7 | 0.03111] 0.15240

X | L. stim. | 11.9 | 24.4 | 0.01667] 0.06861] 0.00966] 0.03974
R. cont.| 11.7 | 24.3 | 0.02633] 0.10835

XI/L. stim. | 18.1 | 21.0 | 0.01452] 0.06914 0.00167] 0.00795
R. cont.| 19.2 | 21.0 | 0.01285] 0.06119

XII | R. stim.| 21.0 | 20.9 | 0.00163} 0.00500) 0.00035] 0.00166
L. cont. | 22.5 | 20.7 | 0.00138} 0.00666

Average loss in per cent 0.00794| 0.04572

(A) Average of duplicate analyses.

STIMULATION
TIME

3 hrs. 45 min.

Int. 10 sec.

3 hrs.
Int. 10 see.

4 hrs.
Int. 10 sec.

3 hrs.
Int. 10 see.

2 hrs.
Int. 33 sec.

3 hrs., 30min.

Int. 8 sec.

3 hrs.
Int. 10 sec.

3 hrs., 15 min.

Int. 10 sec.

3 hrs.
Int. 10 see.

3 hrs., 45 min.
Int. 10 sec.

2 hrs., 30 min.

Int. 10 sec.

1 hr., 30 min.

Int. 10 see.

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 365

TABLE IIT
(Dog)

Effect on iodine content of thyroid of stimulation of the thyroid nerves near the
entrance into the gland.

% OF % OF
‘ % or % or /0
wt. oF | % oF ee LOSS OF | LOSS OF
NO.OF | LOBE OF | pois | roraL | IODINE IODINE canna Tareas TIME OF

ANIMAL GLAND FRESH DRIED

STIMULATION
LOBE | SOLIDS! esve | TIssUE

FRESH IN DRIED
TISSUE TISSUE

I | R. stim. | 12.6 | 23.8 | 0.01299) 0.05450} 0.00152) 0.07870} 3 hrs., 45 min.
L. cont.| 8.3 | 18.4-| 0.02451] 0.13320 Int. 10 sec.

II | R. stim.} 15.1 | 24.3 | 0.01491) 0.06135) 0.00238) 0.01260) 3 hrs.
L. cont. | 14.3 | 25.7 | 0.01253) 0.04875 Int. 10 sec.

III | R.stim.| 5.6 | 27.0 | 0.03883} 0.14380} 0.00463) 0.02866] 3 hrs.
L. cont. | 6.8 | 25.2 | 0.04346) 0.17246 Int. 10 sec.
IV | L.stim.| 3.9 | 22.8 | 0.03480) 0.15263} 0.00639) 0.05229] 3 hrs., 15 min.
R. cont 3.9 | 20.1 | 0.04119] 0.20492 Int. 10 sec.

V |R. stim.| 15.4 | 22.5 | 0.00343) 0.01524) 0.00103) 0.00589) 2 hrs., 45 min.
L 21.1 | 0.00446) 0.02113 Int. 33 sec.

.8 | 0.00223) 0.01022) 0.00193) 0.01027) 3 hrs.
3 | 0.00416) 0.02049 Int. 10 sec.

Neo) le
ic)
° .
=
on
_
bo
~I

Average loss except No. II | 0.00510) 0.03516

TABLE IV

The influence of stimulation of the cervical sympathetic and thyroid nerves on the

| blood flow in the thyroid gland. In Exp. 1-3 the figures are givenincc. of blood

per minute; in Exp. 4-9 the figures are given in number of drops of blood per 20
seconds. In Exp. 8 and 9 the stimulation lasted only 2 seconds.

EXPERIMENT NO. BEFORE STIMULATION | DURING STIMULATION | AFTER STIMULATION
1 17.5 13 16
2 15.0 12 13
3 9.0 6 8
4 38 28 36
5 36 25 33
6 88 22 28
7 38 23 31
8 40 28 32 (2)
9 24 16 (1) 24 (2)

(1) 5 seconds after stimulation (2) 10 seconds after stimulation

a_i

366 Cc. F. WATTS

TABLE V

The effect on the iodine content of the thyroid gland of periodic diminution of
the blood flow through the gland, occlusion of the main thyroid artery for 1-2 seconds

every 10 seconds.

DIFF. IN DIFF. IN

R G R
no.or| tosror | 7: OF 7% OF J OF I 7o OF I % OF % OF
awiwaL| GLAND | FRESH | TOTAL | IN FRESH | IN DRIED | LOSS IN | LOSS IN
LOBE | SOLIDS}| TISSUE TISSUE FRESH DRIED

TISSUE TISSUE

TIME OF INTERF.
CIRC.

3 hrs., 15 min.

I | Right 38.5 | 24.4 | 0.00335) 0.013872) 0.00272) 0.01094
L. cont. | 37.9 | 24.2 | 0.00607) 0.02466
TI} Left 40.2 | 21.1 | 0.00166} 0.00786} 0.00126} 0.00719) 3 hrs.
R. cont. | 28.9 | 19.4 | 0.00292} 0.01505
III} Left 3.9 | 20.8 | 0.01324) 0.06365) 0.00229} 0.00694! 3 hrs., 30 min.
| R. cont.| 3.0 | 22.0 | 0.01553] 0.07059
IV| Left 14.0 | 25.1 | 0.00960} 0.03824) 0.00220| 0.00896) 3 hrs.
R. cont.| 9.5 | 25.0 | 0.01180} 0.04720
V} Right 6.5 | 21.6 | 0.00432} 0.02000} 0.00148] 0.00723] 3 hrs
L. cont. | 11.7 | 21.3 | 0.00580) 0.02723
VI\ Left 8.3 | 23.1 | 0.04281) 0.18530} 0.00532] 0.04889) 3 hrs.
R. cont | 8.0 | 21.0 | 0.04813) 0.22919
VII! Left 10.8 | 22.3 | 0.03256} 0.15964] 0.00422} 0.02243) 3 hrs.
R. cont. | 10.2 | 20.2 | 0.03678] 0.18207
VIII} Right 14.5 | 24.3 | 0.02613) 0.10753} 0.00671) 0.03061! 3 hrs., 15 min.
L. cont. | 15.8 | 23.7 | 0.03284) 0.13814
IX} Right 32.0 | 21.0 | 0.05548] 0.26419} 0.00111] 0.00528) 3 hrs.
L. cont. | 29.2 | 21.0 | 0.05659) 0.26947
X| Right 3.9 | 22.0 | 0.01280] 0.05818} 0.00290) 0.01587| 2 hrs., 40 min.
L. cont 3.4 | 21.2 | 0.01570) 0.07405
Average 0.00312) 0.01593
SUMMARY

1. The iodine percentage of the thyroid lobes in dogs
identical (Table I).

is practically

2. Stimulation of the cervical sympathetic nerve or the thyroid
nerves induces a decrease in iodine and water content of the thyroid

glands (Tables II and III).

EFFECT OF BLOOD FLOW ON IODINE CONTENT OF THYROID 367

3. The stimulation of the cervical sympathetic or the thyroid nerves
induces vaso-constriction in the thyroid gland (Table IV).

4. Decrease in the thyroid blood flow similar to that induced by stim-
ulation of the above nerves, but produced by mechanical means, causes
a decrease in iodine and water content of the thyroid gland (Table V).

5. The changes in the thyroids so far shown to be caused by the stim-
ulation of the cervical sympathetic or the thyroid nerves can all be ac-
counted for by the vascular changes.

6. No histological changes could be demonstrated following stimula-
tion of the thyroid nerves or mechanical interference with the blood
flow.

I wish to express my gratefulness to Dr. Carlson for his supervision;
to Dr. Koch and Mr. Maurer for their valuable suggestions in the an-
alytical work; and to Dr. Bensley for his help in the histological work.

ADDENDUM

These observations appear to place the thyroid gland in the same
category as the liver, that is to say, as organs very delicately balanced
in relation to the local circulation, irrespective of any secretory nerves.
Since the periodic stimulation of the thyroid nerves induce primary
vaso-constriction followed by vaso-dilatation, one cannot say at present
whether the vaso-constriction or the vaso-dilation is the active factor
in bringing about the discharge of iodine from the thyroid gland, or
at any rate upsetting the balance between iodine absorption and iodine
discharge. The thyroids have a greater blood supply than any other
organ in the body. This tremendous blood supply has in all probability
a far greater significance than the mere oxygen supply and the carbon
dioxide removal.

The fact that relatively transient vaso-motor disturbance in the thy-
roid induces such marked change in thyroid activity is probably of
significance in certain types of goitre in man.

The observations also offered another illustration of the difficulty
in deciding whether we are dealing with vaso-motor nerves or with
secretory nerves in certain organs. The adrenal glands probably
also fall in this group, although most investigators interpret the in-
creased discharge of adrenalin following stimulation of the adrenal
nerves as a true secretory nerve action.

/: ae a

368 Cc. F. WATTS

BIBLIOGRAPHY

(1) Rane, Rogers, FAWCETT AND Breese: This Journal, 1914, xxxiv, 72.
(2) Rurnenart: Am. Jour. of Anat., 1912, xiii, 91.

(3) OsSAKIN: Zeitschr. f. Biol., lxili, 444.

(4) AsHER AND FLack: Zeitschr. f. Biol., 1911, iv, 83.

(5) Hunter: Journ. of Biol. Chem., 1910, vii, 336.

(6) Koc: Journ. of Biol. Chem., 1913, xiv, 106.

(7) Gruser: This Journal, 1915, xxxvil, 266.

(8) MARINE: Personal communication.

PHOTOELECTRIC CURRENTS IN THE EYE OF THE FISH!
EDWARD C. DAY

From the Zoological Laboratory, Syracuse University

Received for publication July 27, 1915

CONTENTS
Il Linnie ICH PR aoa tek Ga Gee dp Bae 2a. 6S CR a Os ae eee ee 369
METER ORICA IST OVIGWA0 0 rls aise te ere tae eeu a ees Econ aosie Shs Sa o5's ¢ 371
PP MESLONENLS SNC ODSEIVAULONS. 0000.5. Fm es wl os oh nsads eee eac ese vas 374
ACOA aeATSOnVal gal vanometer.:oc. seat Lae soe. occ cele k oud ties 374
B. With Einthoven string galvanometer:.........-........-0..00005 376
Pei elevexposures) to: bighides 2.007 5s ae ia wo Seis od inya ged 377
MD AGk-AG apiedteyemaey. oo se eye. eels ops Stebe aieke sot hae 377
Bee P GHA ADL BYE rm aks... curs he eae oe oeuken wea 380
2. Intermittent exposures to light. Dark-adapted eye........... 382
MERINa ES OL OUSCLYALLONS oy POE sik 6 bed. BS AGRE ER. 383
MMMPIREORAION: 2500. [coer ee. slates AUR Set Seid Sanaa ee Pee ome 384
Am OT UIGI SM Ol Am CLOG Ger. <i cca eetn sire aePiemick Reine Sot aera ae 384
B. Comparison with results obtained for other animals............. 385
PeasvmiHesis OF the Oscillatory CUTWO.. <2:c25 6. oem hs sone sie scate ome 387
PAsinterpresation: of the retinogram*::<.- 0 ton 2: 2.0. ess.-)- as aeae ae a= 389
E. Similarity of the tri-substance theories ....................202-205 393
ROME INIGRTN TS :).. 5. [0 cclc sos «ods nt nn Aigtinasi dy DUM ant came spel 2 397
ER Cra IO UM OLD LA LESS sarees, = \c) pees eee beet renee Oe Gs Seasyae cate: clea cee 398

I. INTRODUCTION

It is a fact well known that electrical phenomena are intimately
associated with the normal activity of muscle and nerve. Of later
experimental acquisition however is the knowledge that the sense
organs fall within the same category; both the eye and the ear are
accompanied in their normal functioning by electric currents. A beam

1 The substance of the present paper was sent to the Arch. f. Anat. u. Physiol.,
July 15, 1914, under the title ‘‘Die photoelectrischen Vorgiinge im Auge des Fis-
ches,’’ together with the original photographic records. Since the war has in-
definitely delayed the appearance of the paper in that publication, no word hav-
ing been received regarding it after a year has elapsed from the date of send-
ing, I have deemed it only fair and proper that the results should be published in
one of the American journals.

369

370 EDWARD C. DAY

of light falling upon the eye sets up a complex change of electrical po-
tential in the retina, while sound vibrations generate electric currents
in the cochlea.

The frog and higher vertebrates have been the animals upon which
the photoelectric currents have been most successfully studied. The
phenomena which investigators have collectively brought io light by
such a study are these:

When a dark-adapted eye is connected with an instantaneously
reacting galvanometer, after placing one electrode on the cornea and
the other on the back of the eyeball, and a brief exposure is made of
the eye to light, the galvanometric deflections are, three (a negative
followed by two positive) for the ‘‘make” of light, and one (positive)
for the ‘‘break of light.””. The usage of the terms positive and nega-
tive will presently be explained. All four of these deflections are ob-
tainable for the eye of the frog, turtle, chicken, owl, dove and monkey. —
Variations occur for eyes of the cat and the dog. The studies of the
phenomenon in the eye of the fish have, for reasons to be mentioned,
been attended with meagre success.

The galvanometers at first employed were too sluggish to respond
at all to the fleeting retinal currents; and even when instruments
were found which registered the electric effects for the eye of the frog,
the same instruments and methods when applied to the fish were without
result. Whereas the extirpated eye of the frog is capable of yielding
strong retinal currents long after the operation, the extirpated eye of
the fish, on the contrary, loses that ability so rapidly that, by the time
the operator is ready to experiment, the currents are too feeble to re-
cord. In order to overcome this obstacle one must work with the
living fish. The difficulty, however, of keeping the animal’s eye slightly
above water in order to apply electrodes to it, and the rest of its body
below the surface in order that it may breathe, has been perhaps the
chief factor which has deflected investigation of the phenomenon in
the retina of the fish to a more intensive study of it in the frog and higher
vertebrates. It was with the purpose of supplying this lack of data
for a purely aquatic animal that the present research was undertaken.

The work was carried on in the Physiological Institute of Berlin
University during the winter semester of 1912-1913. To Prof. Max
Rubner, director of the Institute, I am greatly indebted for the un-
stinted privileges of research granted to me. The problem was pro-
posed to me by Professor Piper, and to him for his kind supervision and
valuable suggestions throughout the work I desire here to express my

PHOTOELECTRIC CURRENTS IN EYE OF FISH 371

thanks. I wish also to thank Dr. Arnt Kohlrausch for much kind
assistance. The opportunity for the sojourn in Berlin was afforded
by the grant of a Sheldon Fellowship from Harvard University.

There are certain matters of definition which should be made clear
at the outset. Waller (22) found difficulty in reconciling statements
of previous observers with reference to the direction of the current
at “make” and “break” of light, “because that that direction was so
frequently characterized as positive and negative with reference to
a ‘current of rest’ which was itself of variable direction.”’ Since later
investigators, in agreement with Waller, have found the current of
rest (i.e., the current occasioned simply by placing the cornea in circuit
with the back of the eyeball) taking the direction from fundus to cornea
within the eye and from cornea to back of eyeball in the outside circuit,
this direction has been regarded as positive. Since in the fish eye the
current of rest takes the same direction, the term positive will here be
used in the same sense. In the photographic curves it is always an
upward deflection.

Gotch (9) introduced the terms “On effect” and “Off effect” to desig-
nate the initial and terminal effects at make and break of light; but
since his ‘‘On effect’’ was primarily attached to the first positive deflec-
tion in the photographic curve, I wish to employ the term to cover the
negative preliminary depression and the secondary positive rise in the
curve as well, unless otherwise qualified. The term ‘Off effect” will
always refer to the simple positive deflection occurring at the break
of light. In order to refer to the mdividual features of the curve
more easily, Einthoven and Jolly (8) have been followed in denominat-
ing the negative preliminary deflection as A, the first positive elevation
as B, and the secondary on-effect as C; but instead of adopting A’
to designate the off-effect (as E and J do) which associates it upon hy-
pothetical grounds with the on-effect A, the non-partisan letter D has
been given it.

II. HISTORICAL REVIEW

In the year 1849 du Bois Reymond (7) made the discovery that the
retina of the fish reacted electromotorically. An extirpated eye of
the tench, with one electrode on the cornea and the other either on the
cross-section or on the longitudinal axis of the optic nerve, gave an
electric current. An action-current could similarly be obtained by
connecting the cross-section of the optic nerve with its long axis. It
occurred to him to test further for the effect of light striking the retina,

Si2 EDWARD C. DAY

but his experiment on a turtle with the aid of a multiplicator, an instru-
ment responsive only to the stronger currents generated during muscular
contractions, yielded no results.

It remained therefore for the Swede Holmgren (12, 13) to discover
the photoelectric reaction for which du Bois Reymond had sought
in vain. The electric currents evoked by light which he detected in
the eye of the frog and the rabbit, however, he was unable to elicit
with surety from the eye of the fish. The experiments were performed
upon extirpated eyes of pike, perch and bleak, but the results obtained
were so fleeting and irregular that, although on- and off-effects were
obtained once or twice, the question remained a matter of uncertainty.

The investigations of Dewar and M’ Kendrick (6) on the eyes of the
goldfish, rockfish and stickleback with the aid of a Thompson galvanom-
eter met with better success. From the goldfish (Cyprinus auratus)
a-positive on-effect was obtained and a negative off-effect; but whether
the former was the B or C deflection was not distinguished. Similar
though slower reactions were obtained for the rockfish (Motella vul-
garis). In the case of the stickleback, the authors remark (p. 61):
“A deflection of 400° was obtained, which was remarkably sensitive
to light, the variations occurring in very short periods of time.’”’ The
conclusion drawn from the experiment was that the differences in the
reaction-times of the various fish was to be attributed to the different
habits of the animals; or, in the authors’ words, ‘‘on comparing these
results, it will be seen that in the active alert stickleback, the variations
occurred rapidly; in the more sedate goldfish, they occurred more slowly,
while, in the case of the sluggish rockling, which hides beneath stones
near low-water mark, the variations occurred very slowly.”

Kihne and Steiner (16), in working with enucleated eyeballs of perch,
pike, barbel and eel, obtained the positive on-effect as had already
been observed by Dewar and M’ Kendrick, but instead of a negative
deflection at break of light, they got a feeble positive effect. When
cornea, iris and lens were cut away, the posterior half of the eyeball
yielded the same result, only that in this case the off-effect was decidedly
positive. From the isolated retina too, provided that it was strictly
fresh, were obtained similar though feebler results. In all cases the
current remained constant during illumination as soon as the first posi-
tive deflection had subsided. Again the negative preliminary (A deflec-
tion) escaped notice because the instrument used was too slow in reac:
tion to register it. It is also likely that the positive on-effect observed
by these investigators was the secondary C crest rather than the more

PHOTOELECTRIC CURRENTS IN EYE OF FISH 373

rapid B deflection, judging from the statement, “Die Schwankungen
der Retinastréme sind beim Kommen des Lichtes positiv sehr langsam
wachsend, gegen das Ende rasch zum Maximum gehend, wihrend
des Belichtens anhaltend.”’

From 1881 to 1907 attention was concentrated upon the phenomenon
as exhibited by the eyes of the frog and higher vertebrates, while study
of them on the fish was neglected. The researches of v. Briicke and
Garten (2), carried on with a more efficient instrument in the small
Einthoven string galvonometer, were extended to cover aquatic as
well as terrestrial vertebrates. The fish used were the pike and the
bleak. In experiments with the former, where the electrodes were
applied to the eye in situ, the on-effect obtained was described as “‘eine
positive Schwankung des normal gerichtetet Bestandstromes, die
waihrend der Belichtungsdauer an Stirke zunahm,” and the off-effect
as “eine positive Schwankung, nach der der Strom allmihlich zu seinem
Ruhewert zuriick kehrte.”’ From the enucleated eyeball, however,
the on-effect was negative while the off-effect was positive. For the
bleak only extirpated eyes were used. The current of rest obtained
was strong and ran in the outer circuit at first from the back of the
eyeball to the cornea, then, diminishing rapidly to zero, flowed in the
reverse or normal direction. These investigators were able to detect
a negative preliminary as a feature of the on-effect, the latent period
of which was 0.074 sec. in one instance, and 0.075 sec. in another.
Following this negative preliminary came a positive deflection with
a latent period measuring 0.128 sec. and 0.132 sec. in the two cases
respectively. On darkening, the eye of the bleak yielded no definite
reaction and the curve subsided gradually to zero. The combined
results from both kinds of fish exhibited therefore the full set of phenom-
ena known to occur in the frog retina, viz., the initial negative de-
pression and the two positive deflections as the on-effect, and the
positive deflection as the off-effect.

Since the work of v. Briicke and Garten no further attempt had been
made to verify or add to these facts regarding the photoelectric phenom-
enon in the eye of the fish. The substitution of the quick-reacting
string galvanometer for the older and more sluggish instruments, en-
abled Einthoven and Jolly (8) to analyse the currents in the eye of the
frog far more carefully than had been done before, and Piper (20)
_to make a thorough comparative study of the photoelectric effects in
_ the eyes of the turtle, dove, chicken, buzzard, owl, rabbit, cat, dog,
monkey and octopus as well as the frog. Since the fish was not included

374 EDWARD C. DAY

in this set of investigations, it was with the object of supplementing
the series in that direction that, first with the aid of a d’Arsonal gal-
vanometer and finally with a large Einthoven string galvanometer
‘I carried out the following experiments.

Hil. EXPERIMENTS AND OBSERVATIONS

A. With d’Arsonval galvanometer

The kinds of fish which were used in the experiments were the tench
(Tinca tinea L.) and the pike (Fsox lucius L.) These are freshwater
forms common to northern Germany and were obtainable throughout
the winter months of 1912-13 during which the investigation was car-
ried on. The pike, although a less hardy fish than the tench, yet, be-
cause of its larger eyes and more convenient shape, proved the more
satisfactory; it alone therefore served as material throughout the chief
part of the work. The fish were obtained fresh from the market every
two or three days as needed and were kept in a laboratory aquarium
supplied with running water.

The d’Arsonval galvanometer was employed in the earlier experi-
ments, but as soon as the technic of working with the live animals
was mastered, resort was made to the Einthoven string galvanometer
in conjunction with the Hut photographic register for making the final
records. .

The experiments with the d’Arsonval instrument were performed
on two tench and four pike. In order that the electrodes might be
placed upon the eye in a dry condition, the fish was wrapped first in
a wet cloth and then in a wire net to prevent any bodily movement.
Thus encaged the animal was laid on its side upon bags of sand in a
trough of water so as to immerse the snout and as much of the rest
of the body as possible, one eye being kept above the surface. While
in this position the fish could without difficulty take in water through
the mouth and irrigate the gills, and it seemed to suffer no incon-
venience during the hour or two of experimentation. The water was
aerated with compressed air. The trough was placed in a large black
box with black curtains in front. Light from a 32 C. P. electric light,
set at about a meter distant, passed through an aperture in the rear of
the box and was reflected from a mirror situated inside onto the eye
of the fish. Before being placed in position in this dark chamber the
eyeball was partially laid bare by a slight operation so that one elec-

PHOTOELECTRIC CURRENTS IN EYE OF FISH 375

trode might be brought in contact with the eyeball and the other with
the cornea. The meshes of the wire net enveloping the fish were
widened over the eye sufficiently to admit the electrodes, contact from
which being made by means of short woolen threads saturated with
physiological salt solution. The temperature of the water did not vary
more than a degree or two centigrade from that of the room (av. 17°C.).
For the first five experiments the animal was not dark-adapted; but
thereafter it was placed, after the operation on the eye had been per-
formed, in a light-proof tank of running tap-water for about three hours
previous to the experiment.

TABLE I.

Photoelectric currents in the eyes of the tench (T) and pike (P) as measured by a
d’Arsonval galvancmeter.

DEFLECTION ON GALVANOMETER
NO. HOURS
EXPERIMENT 1912 FISH DARK-
e ADAPTATION | Current of
Test

On effect Off effect

Lhe ccs ole? Abe} 0 +6. cm +0.5 em 0 cm
Tn... dle daa 0 +4. +2. 0
Se, es S'S Alpers 0 +12 +0.2 0
Ne ee MoS Deer 0 yy 3 + 1 0
cite cis: XI. 17 Ppt 0 +20 +0.5 +0.1
OL eee ele 7 12 Al 3 +25 +11 1)
Pee ok DOF peel ? —25 +0.2 0
eure Bes) a mS2 1 ops 3 —4. 8 —0.8 0
2): sae ele25 ARE a! 1e5 +15. +2. 0

(UE. See eee XI. 26 Pi 3 3 +11. + 6. 0

‘ey ae elilens | lees 3 ? + several ?

| cm.

The results of observations made with the aid of the d’Arsonval
galvanometer are given in Table I. Column three indicates the kind
of fish used (T. for tench, P. for pike) and the animal’s number;
column four the period of adaptation to dark; columns five, six and
seven give the throw in centimeters on the galvanometric scale for
the current of rest? and for the on- and off-effects of light respectively.
The plus sign signifies a current flowing in the outer circuit from cornea
to back of eyeball, and the negative sign the reverse direction. It will
be seen that in seven of the eleven cases the current of rest was posi-

2Before the beginning of each exposure this current of rest was neutralized
by means of an oppositely directed galvanic current.

376 EDWARD C. DAY

tive. The last, though unrecorded as to direction, was probably pos-
itive too, making a total of eight. Only once (Experiment 8) did the
flashing on of the light generate a current whose direction was opposed
to that of the normal current of rest, i.e., from back of eyeball to cornea
in the outside circuit. In this instance both current of rest and photo-
electric current were negative, accounted for no doubt by the feeble
condition of the animal, for it died before the end of the experiment.

The results in general showed that the eyes of both tench and pike
responded with postive currents to the onset of light and that this
effect was enhanced by dark adaptation. A negative preliminary
to the positive on-effect, could not be ascertained with this instrument.
As for an off-effect, only once or twice could a feeble positive current
be observed.

B. With Einthoven string galvanometer

The study of the photoelectric phenomena with the aid of the Eint-
hoven string galvanometer was conducted along two lines: first, the
simple effects of making and breaking the light were investigated, and
then the compound effect obtained by intermittent stimulation. In
the first part of the work both dark-adapted and light-adapted animals
were used, but for the experiments with intermittent light only dark-
adapted ones were employed. Care was taken to select from the
market fish whose physiological condition was sound. This was
especially important because the pike, the only fish used in this set of
experiments, is not a very hardy animal, and unless procured when in
prime condition yields electric currents too feeble to be photographed
satisfactorily.

The procedure was modified in some respects from that which was
followed in the first series of experiments. The operation was the same:
a small part of the orbit overshelving the eye was cut away and the
exposed muscles freed from their insertions on the eyeball so that one
electrode could be brought into contact with the equatorial region.
In some cases the operation was performed in the late afternoon and
the fish kept in the dark over night.

An arrangement for holding the fish in an immovable position was
devised as follows (fig. 1): a board of sufficient length and width to
accommodate the largest animal was provided with a large hole near
the end, while many smaller holes were scattered irregularly over its
length. The fish was wrapped in a wet cloth, the head and gills alone

PHOTOELECTRIC CURRENTS IN EYE OF FISH BY i

being left exposed, and then bound on its side to the board by means
of wires inserted in the small holes.. The under eye and gills, falling
in the open space of the large aperture, were quite free from pressure,
and the gills thus had freedom of action when running tap-water was
led into the mouth through a rubber hose. The fish had less chance
to move in this position than when enclosed in the wire net, and con-
sequently it became somewhat numb from the confinement; but it
recovered its muscular activity as soon as it was released and returned
to the aquarium. When firmly fastened to the board the fish was laid
in a large tray which had an outlet tube in the bottom to drain off the
water issuing from the gills. This tray was situated in the large dark-
box above mentioned through whose aperture at the back the light
from a single Nernst glower was reflected by a mirror onto the eye of
the fish. One of the electrodes was applied indirectly to the bulb
of the eye by means of a wool thread, while the well moistened clay
boot of the other impinged directly on the cornea.

In order to obtain a record on the photographic paper of the moment
when the light stimulus was applied, the same apparatus was used
which Piper (20) had devised for his experiments on the eyes of higher
vertebrates. It consisted of a T-tube (Fig. 1, ¢) contaming two mirrors
located in the center of the cross-arm and at right angles to each other,
so that the light entering the vertical arm was reflected, half of it through
one end of the cross-arm into the photographic register (7), and the
other half through the other end onto the eye. A photographic shutter
(s) on the end of the vertical arm made time exposures of varying
lengths possible. When the shutter was opened both the eye and the
photographic paper received a portion of the light simultaneously.
A time-marker (z), making beats at the rate of five per second, was
placed in front of the photographic register so that the shadow of its
needle together with the shadow of the string of the galvanometer
(g) fell upon the sensitized paper. Every photograph thus contained
a three-fold record: (1) that of the begininng, duration and end of stim-
ulation; (2) that of the retinal currents produced; (3) a time-scale in
fifths of seconds.

1. Single exposures to light

(a) Dark-adapted Eye. In the preliminary experiments the degree
of dark adaptation of the eye seemed to affect materially the size of
the deflections of the galvanometric needle. In this set of experiments
therefore attention was paid at first only to the phenomena exhibited

378 EDWARD C. DAY

by eyes which had been adapted to the dark for at least three hours.
The fixing of the fish to the board, the regulation of the artificial respira-
tion and the placing of the electrodes on the eye, were all done by the
light of a red electric globe. Since in the first experiments the current
of rest in normal animals was found to flow from cornea to eyeball in
the outside circuit, and the on-effect, B, to express itself by a current
flowing in the same direction, it was assumed in these experiments that
when the B deflection agreed in direction with that of the current of rest,

:
EK (2

I ATT OTT
9

EXPLANATION OF FIGURE Il

a, are light n, Nernst light
c, electric circuit from eye to galva- p, shallow tray

nometer : _ r, photographic recording apparatus
d, dark box s, photographic shutter
€, episcotister t, T-tube for dividing the light from n
f, fish board z, time-marker

g, string galvanometer

this direction was from cornea to eyeball in the outside circuit. When
the direction was in disagreement, but the on- and off-effects were
normal, it was assumed that a polarization current was masking the
current of rest. Before the usual neutralization of the current of rest
was made, the polarization current .was reversed by readjusting the
electrodes and brought to concide with it. Instances of this kind how-
ever were rare. Be it understood then, that an upward deflection in
the curves signifies a current flowing in the outside circuit from the
cornea to the back of the eyeball.

PHOTOELECTRIC CURRENTS IN EYE OF FISH 379

It may be stated at the outset that the photoelectric phenomena
in the eye of the fish show striking agreement with those exhibited by
the frog. The on-effect comprises three features: a negative pre-
liminary A, followed by a quick positive elevation B, and a slower
secondary rise C. The off-effect consists of a positive deflection D.
These phenomena will be considered first for the dark eye and then for
the light eye.

The negative preliminary A appears at some stage or other in each
of my eighteen series of records, although not in every individual
photograph of a given series. Its absence from some of the photo-
graphs will be accounted for later. Suffice it to say at this point that
it is typically present for the dark eye. It is preceded by a latent period
of 0.032 seconds after the make of light, this value being an average of
twenty-nine measurements obtained from the records of five different
animals.

On-effect. The positive on-effect B following the negative prelimi-
nary never fails to appear except in the record of an animal whose
physiological condition was at a low ebb. The latent period for B,
measured from the beginning of the light stimulus to the maximum
point of the negative preliminary depression, averages 0.075 seconds
from thirty-five measurements. The maximum point in the deflec-
tion is arrived at on an average in 0.26 seconds. Its amplitude varied
considerably in the course of the experiments owing to variations in
the resistance of the electrode connections and to differences in the
physiological state of the animals. The usual amplitude is about
equal to that obtained from a galvanic current of one-tenth millivolt.
The maximum is half again as great as this standard amplitude. In
the dark-eye this positive on-effect is characteristically large. The
ascending leg of the curve is always abrupt while the descending one is
quite gradual. In the descending phase the curve sometimes passes
below the level of the current of rest and becomes negative. The
amount by which it approaches or passes under the zero line is deter-
mined largely by the strength with which the slow-developing secondary
positive elevation C makes itself manifest.

In animals left in the dark over night and used on the following morn-
ing direct from their long period of dark adaptation, this secondary
rise following the first positive effect is quite pronounced. With each
exposure to light however, it develops at an ever retarding rate and ever
diminishing amplitude. It soon disappears altogether and the whole
curve following upon B subsides to a constant level which may be even

380 EDWARD C. DAY

below that of the current of rest. Since in most instances the C crest
developed so slowly that much photographic paper must needs have
been consumed in order to obtain its full course, only a single record
was made in which its full course was followed. It was a case in which
the fish had been dark-adapted over night and the C effect was strong.
The period from onset of light to the maximum point in the curve was
here seven seconds. The descent was slower than the ascent, and after
nearly twenty seconds of its downward course, at which point the record
was ended, the.curve had not yet returned to the zero line. Thedura-
tion of exposure in this instance was two and three-fifths seconds. The
condition necessary to the production of the secondary elevation seems
to be a long interval of dark adaptation.

Off-effect. The positive elevation in the curve, occasioned by the
shutting off of the light, has an average latent period of 0.049 seconds.
The off-effect increases in magnitude the longer the eye is previously
exposed to light. The ascending slope of the D crest is somewhat
steeper than the descending. The latter, before reaching the level
held by the curve just prior to the break of light, is often arrested in its
rapid descent to continue at a more gradual rate downwards. It
is quite possible that this second phase in the descent is but the under-
lying C elevation, upon which the off-effect had been superposed, again
cropping out, or it might be due to the still lingering descent of the B
crest.

(b) Light-adapted eye. Time did not permit an investigation of
the photoelectric phenomena in the light eye to an extensive degree.
The results obtained, however, are significant enough to be described.
In three instances a fish was brought direct from the market, the opera-
tion performed in a large porcelain sink by the light of a 50 e.p. elec-
tric light, the fish placed in position and a photographic record made.
In all three cases the first exposure revealed the fact that the off-effect
was decidedly stronger than the on-effect. In one experiment an acci-
dent to the apparatus prevented the making of a photographic record,
but this decided difference between the on- and off-effects was dis-
tinctly observed when the image of the string of the galvanometer was
projected onto a screen. In the other two experiments reliable photo-
graphic records were secured which fully confirmed this observation.
In Plate I are reproduced one set of these records.

Figures 1, 2 and 3 are curves obtained from a single animal during
one experiment, showing first a typical light condition (Fig. 1), then a
rapid transition to a condition of dark adaptation (Figs. 2 and 3),

PHOTOELECTRIC CURRENTS IN EYE OF FISH 381

in which the comparative magnitude relations of on- and off-effect
have become reversed. Examined in detail the curves exhibit these
features: Figure 1 (a record made immediately after bringing the
fish from the market and without previous dark adaptation), the ab-
sence of both A and C deflections, the presence of a diminutive B crest
and a strong D deflection; Figure 2 (a record made after ten minutes’
dark adaptation following upon Fig. 1), the feeble presence of A, with
B strongly developed and D much reduced—the slight elevation fol-
lowing B is not a C crest, as it was occasioned by the movement of the
animal’s body; Figure 3 (following Fig. 2 after six minutes more dark
adaptation), the indisputable presence of the negative preliminary
A, a further slight increase in the magnitude of B and slight decrease in
the size of D. This last might be attributed to the slightly shorter
period of illumination preceding it as compared with the length of
exposure in Fig. 1. Thus the typical condition for the light-adapted
eye is seen only in the first exposure, and it passes quickly over into that
of the dark eye with a comparatively short period of dark adaptation, -
ten to twenty minutes.

There is another phenomenon in connection with the transition from

the light to the dark condition and the reverse to which I desire to call at-
tention. If an eye of intermediate dark adaptation be subjected to light
for ten minutes, there is produced at close of exposure a gradual secondary
positive rise following the D deflection which may be designated with
the letter E. If during this secondary elevation an exposure of about
one-eighth second be made, the A, B and D fluctuations appear in the
curve without altering its general upward trend. In similar manner,
if an eye of intermediate adaptation be subjected to dark for ten min-
utes and then the brief exposure be made, a strong downward course
follows the first positive on-effect B, and upon this descending phase
the D crest may be superposed as an off-effect without altering the
general downward trend. The off-effect has also been obtained super-
posed upon the ascending phase of the C crest. Thus the D deflection
may crop out upon either ascending or descending phases of other fluc-
tuations such as the C and E crests, or upon the descending phase of
the B deflection. But it is also true for the B deflection that it too may
be superposed upon ascending and descending phases of C and E, and,
in the case of intermittent stimulation, there are often superposed the
distinct deflections of successive short-lived B-effects upon the linger-
ing descent of the first strong B deflection (slightly exhibited in Fig.
fractal, 11).

382 EDWARD C. DAY

2. Intermittent Exposures to Light. Dark-adapted Eye

For the second set of experiments it was only necessary to interrupt
the light before its entrance into the T-tube by a rotary disk of card-
board with eight radial segments cut symmetrically out of it (Fig. 1, e).
In order to intercept the light simultaneously from both mirrors situ-
ated in the T-tube, the episcotister was set at the point where the
light from the Nernst glower, after emerging from a condensor, came
to a focus close to the mouth of the T-tube, so that each segment of
the rotating disk completely and instantly intercepted the passage
of the light in turn. A photographic test in which the beams of light
emerging from each end of the cross-arm of the T-tube were reflected
side by side onto the photographic paper, exhibited exact coincidence
for the make and break of light; so that there was no uncertainty about
the eye, which received the one beam, and the photographic register,
which received the other, being struck simultaneously at every exposure.
It was of course essential for the accurate measurement of the latent
periods in the curves that this be so. The records were obtained from
dark-adapted eyes exclusively in this set of experiments.

The intermittent phenomena in the eye, corresponding to the rapid
interruptions of light by the episcotister, are shown in Figures 5, 6, 7
and 8, Plate II. In Figure 5, a record from animal No. 39, the rate
of stimulation was three flashes per second, a frequency slow enough
to allow all of the phenomena except the C variations to appear. The
first D crest is either absent or masked by the first positive on-effect.
The last D deflection is present although so closely superposed upon
the last on-effect as to be scarcely perceptible. A slight nick, how-

ever, in the ascending slope of the last fluctuation, together with the

higher altitude of its summit as compared with those preceding, is
evidence of its presence. This fact is an important aid to the inter-
pretation of the curves of higher frequency. The presence of the nega-
tive preliminary A may be inferred from the fact that the troughs
preceding the on-effects are more acute than those which succeed them.

Figure 6, the fifth exposure of a series from animal No. 37, was made
to intermittent light having a frequency of 16 flashes per second. The
intervals of dark are slightly longer than the intervals of light, the
former being 0.037 seconds and the latter 0.029 seconds on an average.
Since there are 32 crests in the curve corresponding to the 32 flashes
of light, they may represent either the make or break of light or else
an additive effect of both. An analysis of the curve will be found on

PHOTOELECTRIC CURRENTS IN EYE OF FISH 383

p. 388, in section C of division V. When the frequency was increased
from 16 to 28 stimulations per second, the eye ceased to respond with
individual currents and the general effect produced was the same
as that for continuous light, and is exhibited in Figure 7.

In order to determine whether the eye by this time might not be
fatigued, the frequency was lowered to 25 flashes per second, and it
was found that the responses of the eye, although feeble, were again
recordable. The sensitiveness of the galvanometer during this series
is shown in Figure 8, a curve produced by a constant current having
a strength of one millivolt. -

IV. SUMMARY OF OBSERVATIONS

1. The photoelectric phenomena in the eye of the fish (pike) are
quite similar to those in the frog. The current of rest flows from cornea
to back of eyeball in the outer circuit, a direction designated as positive.
The onset of light evokes in the curve a negative preliminary depression
A, a positive deflection B and a secondary positive effect C. The
withdrawal of light occasions a single positive deflection D.

2. The initial depression A appears only after a certain amount of
dark adaptation; its magnitude is relatively small, and it has a latent
period of 0.032 seconds.

3. The B deflection is feeble in a light-adapted eye, but with ten to
twenty minutes’ adaptation to darkness it becomes five to seven times
as large. Its latent period is 0.075 seconds and its maximum is reached
on an average in 0.26 seconds.

4. The C deflection appears only in a thoroughly dark-adapted eye;
its development is slow and becomes ever slower with every exposure
to light. Its magnitude may, after long adaptation to dark, exceed
that of B by two or three times. In one measured instance the summit
was reached in seven seconds.

5. The off-effect D increases in size with the length of previous illu-
mination of the eye. Its latent period is 0.049 seconds and its summit
is reached in 0.165 seconds.

6. To intermittent light the’eye yields separate responses up to a
frequency of 25 flashes per second. At 28 per second the responses
become blended together so that the curve obtained resembles that
produced by exposure to continuous light, the A and B deflections of
the first on-effect and the last off-effect alone being distinguishable.

7. A change in the direction of the current of rest may occur through

384 EDWARD C. DAY

changes in the physiolegical state of the fish forerunning death, or by
gradual change in the polarity of the electrodes.

V. DISCUSSION
A. Criticism of the Method

The factors which tended to vary throughout the experiments were
the contact of the electrodes with the eye, the tension of the string of
the galvanometer, the intensity of the hight and possibly the oxygen
supply. The temperature of the water conducted into the fish’s mouth
was that of ordinary tap water and varied from 15 to 17 degrees centi-
grade. Although in cases of extreme changes the temperature may
decidedly alter the phenomena [Gotch (9), Nikiforowsky (18)], this
fluctuation from day to day was too slight to modify the reactions in
any way. The fish, bound to the board so as not to injure it by too
strong compression, was capable of slight muscular movements, and
these together with the respiratory action of the gills, affected the
contact occasionally between the electrodes and the eye; but as muscu-
lar movements of a disturbing character were only occasioned by lack
of oxygen, when the water supply was regulated properly the fish lay
quiescent throughout the experiment. In order to obtain the best
results the tension of the galvanometer string often had to be altered
in the course of the same experiment, loosened when maximum throws
were desired and tightened when the fleeting currents set up by a high
frequency of stimulation demanded a quicker response. Since the
light from the Nernst glower was reflected into the eye by means of a
small adjustable mirror located inside the dark chamber, the intensity
varied to some extent with every readjustment.

The specific effects which an increase in light intensity has upon the
curve are summed up in the recent work of Brossa and Kohlrausch (1)
in these words:

Die Latenz wird kiirzer, der negative Vorschlag tritt erst von einer bestimm-
ten Lichtintensitiit auf und wird von da an stiirker ausgepriigt. Alle drei posi-
tiven Stromschwankungen . . . . steigen rascher und steiler an; die Ein-
trittsschwankung und die Verdunklungsschwankung erreichen friiher ihren
Gipfel. Die Senkung zwischen positiver Eintrittsschwankung und sekundiirer
Erhebung wird tiefer. Die Verdunklungsschwankung steigt in immer schirfer
werdendem Knick von der Abszisse aus an. Wohl gemerkt gehen alle diese
Formveriinderungen Hand in Hand mit einer Zunahme der Ordenatenhdhe der
Ausschlige.

PHOTOELECTRIC CURRENTS IN EYE OF FISH 385

Thus an increase of light intensity produces a shortening of the latent
periods, a steepening of the slopes of the deflections and an increase in
their altitudes.

Although these variable factors probably affected the finer character
of the curves to a certain degree as just mentioned, they did not alter
their general structure nor the direction of the deflections. In the
following discussion which deals with the structural features of both
simple and intermittent curves rather than with their magnitude re-
lations, the latter will for the most part be left out of consideration.

B. Comparison With Results Obtained for Other Animals

If we compare the curves obtained from the eye of the pike with
those obtained from the frog, we find that for both simple and inter-

TABLE II.

A comparison of the latent periods of the photoelectric reactions for the eyes of the
bleakfish, frog and pike.

< a & a
i & @ fe 2
S) S Ric ke
a a as A
oe On o6, B FROM BREAK BLENDING
= c 26 =e) See 3 OF LIGHT TO RATE FOR
eee | SNEMAL AE BE Moe Q MAXIMUM INTERMITTENT
5 3 ecle Ps OF D STIMULATION
ag aie Zo a Ss
a8 <8 me SUe 2
v. Briicke 0.074” | 0.128”
and bleak |0.075 | 0.132
Garten
Pipeee>sc.. . frog | 0.045 | 0.085 | 0.3” | 0.05’’| 0.18-0.27” | 15 per sec.
Diver. <2cs.: pike | 0.082 {0.075 | 0.26 |0.049| 0.165 28 per sec.

mittent stimulationthe phenomenaare closely similar. The correspond-
ing deflections are all like in sign. In Table II are given for comparison
the latent periods for the A, B and D deflections, together with the
time required for the B and D crests to attain the maximum points in
their courses.

The first set of values is that obtained by v. Briicke and Garten (2)
on the bleakfish (Bleie), the second set by Piper (20) on the frog, and
the last by myself on the pike.

Since the same string galvanometer had been used by Piper as was
used in the foregoing experiments, it is noteworthy to compare especially

386 EDWARD C. DAY

the last two sets of values. It should be first remarked however that
v. Briicke and Garten employed a smaller and considerably less sensi-
tive string galvanometer than Piper did. Furthermore, since they
succeeded in obtaining but two photographic records for the bleak,
measurements of which are given in the table, their results are to be
accepted with reserve.

Turning then to the second and third sets of values, it will be seen

that both the latent periods and the times required for the on- and off-.

effects to reach their maxima, are greater in the frog than in the fish;
also that the amphibian eye ceases to respond to intermittent stimuli
at a lower frequency than the fish eye does.

These facts might lead one to suppose that the photoelectric processes
in the visual organ of the pike were of a more rapid nature than those
of the frog. Considering the facts however that the values for the
fish were obtained from eyes in situ and still in communication with
the circulatory system, while those for the frog were obtained from
extirpated eyes; and that Piper employed a light of stronger intensity
(five ampere arc) than I did, the values in the table are not so closely
comparable. The interruption of the circulation in the eye through
extirpation presumably occasioned a retardation of the photoelectric
reaction, while the higher intensity probably had an accelerating effect
which tended to shorten the latent periods; but there is no way of esti-
mating how much the acceleration compensated for the retardation.
As a matter of opinion, I believe that the compensation was not enough
to make up for the depression in irritability induced by impairing the
metabolic activities of the retina.

The value 28 in the last column, the frequency of intermittent flashes
at which a blending of the individual deflections in the retinogram
occurred, was not an average but was a single value obtained in the
course of one experiment upon a normal vigorous animal. Experience
taught that much depended upon the tension of the string in the gal-
vanometer in securing the fleeting currents of the eye. Thus currents
of high frequency to which the slack fiber would not respond, could
be recorded on stretching it tighter.

Piper (20) found that dove and buzzard eyes responded to as high
as 40 flashes per second, the chicken to 35, the owl to 20, the rodent to
25, the cat and dog both to 25 and the Makakus Rhesus monkey to
17. That the fish, then, a cold-blooded animal, should possess an
eye which reacts with greater rapidity to photic stimuli than do those
of warm blooded animals like the rabbit, cat, dog or monkey, seems

PHOTOELECTRIC CURRENTS IN EYE OF FISH 387

incredible, because the fact that the rate of nerve transmission in the
cold blooded forms is slower, would lead one to expect a correspondingly
slower reaction on the part of the retina as well. Further, the fact that
. the latent periods for the on- and off-effects in the eyes of the mammals
mentioned are about twice as short as those for the fish, would like-
wise indicate that only minimal intermittent values have as yet been
obtained for the warm blooded animals, because theoretically the
animal with the shorter latent period is physiologically better capaci-
tated to react to a higher rate of intermittent stimulation than an ani-
mal with a slower reaction period. Since various drugs were employed
by Piper in his operative methods, such as ether for narcotizing, curare
for inhibiting muscular movement, and atropin for enlarging the pupil,
these may have played a réle in diminishing the sensitiveness of the
eye to light. As to the use of atropin, Gotch (9) states definitely that
it lengthens the latent periods. Other factors, too, such as the tension
of the galvanometer string, the resistance of the electrodes, the intensity
of the light, the physiological condition of the animal, etc., enter in
to make difficult a close comparison between the results obtained by
Piper on birds and mammals and those here given for the fish.

C. Synthesis of the oscillatory curve

An analysis of the deflections produced in the retinogram by inter-
mittent stimulation has been carefully made by Piper (20). He re-
gards the curve first from the standpoint of photoelectric currents set
up by a series of ‘‘flashes” of darkness, and again as compounded of
the serial effects from a sequence of flashes of light. By both methods
of approach he arrives at the conclusion that the curve consists chiefly
of negative preliminaries plus positive off-effects, i.e., of A and D de-
flections alternating in rapid succession. If the rate of stimulation is
slow, the B deflections predominate; but since with increase of frequency
the B effects diminish in amplitude while the A and D deflections retain
practically their original size, the curve comes gradually to be com-
posed of the last two kinds of deflections. His argument is based upon
three facts revealed by measurement on the intermittent curve: first,
the time elapsing from each make of light to the beginning of each de-
pression is identical with the length of the latent period for the negative
preliminary, A; second, the period elapsing from break of light to the
maximum point of the next succeeding depression is identical with the
latent period for the positive off-effect, D; third, when an intermittent

388 EDWARD C. DAY

series ends in an off-effect, the last deflection, positive, corresponds in
length of latent period to that of a deflection appearing at the end
of a continuous exposure, thus making the two identical. In the oscil-
latory curve of the frog, therefore, it is the A and D deflections which
primarily determine the oscillations when the frequency of stimulation
is relatively high.

From a similar analysis of my own curves I have arrived at a like
conclusion in regard to the synthesis of the intermittent curve for the
eye of the fish. Figure 6, Plate II, will serve to illustrate. As the
duration of each flash in this curve is about the same as that of the last
flash in Figure 5, Plate II, which, as has already been pointed out,?
occasioned a superposition of the D deflection on top of a B deflection,
one may infer that such is probably the case here too, every upward
stroke being compounded of the ascending slopes of the B and D ele-
vations. There is however another factor involved which tends to
diminish this additive effect. It is the negative preliminary A, which
exerts a depressing influence upon the positive phases of the curve in
the following way: since the sum of the interval of darkness following
a stimulation plus the latent period of the next ensuing A deflection
(i.e., 0.029 + 0.02’’) is always less than the latent period (0.09’’) needed
by the off-effect to arrive at a maximum, the A depression curtails the
full additive value of each off-effect. Therefore the summits of all
the upward deflections in the curve, as soon as the effect of the first
strong positive on-effect has subsided enough to allow them to appear,
are necessarily submaximal up to the last one; but this final one, because
of the fact that no A deflection follows to depress it, rises to a maximum.
The more the rate of stimulus is increased, the more the A deflections
overlap the B deflections and tend to diminish their altitudes.

Further, since the altitudes of B and D deflections both are depend-
ent upon the restorative periods of dark and light preceding each
respectively, according as these periods are shortened the altitudes of
the crests are diminished. It follows then that two factors, one of
interference and the other a diminution in the length of the restorative
periods after each stimulation, combine to depress the crests of the
waves to a common level as the frequency of stimulation is increased.
The rise at the end of an intermittent series is an expression of the re-
lease from the depressing effect of a subsequent negative preliminary.

3 See p. 382.

PHOTOELECTRIC CURRENTS IN EYE OF FISH 389

D. Interpretation of the retinogram

It is generally conceded that the C deflection can play no part in
producing the sensation of light, because of the slowness with which
it develops. The underlying cause for this deflection might be one
of several phenomena which take place in the eye: the chemical dis-
sociation of visual purple, the migration of pigment, the contraction
of cones or the chromolytic changes said to occur in the ganglion cells
of the retina.

As to the first phenomenon, Kiihne and Steiner (16) state that the
photoretinal currents occur in the frog after the visual purple has been
bleached out. This is confirmed by Holmgren (14). The former
authors also claim that if during dark adaptation the re-formation of
visual purple be prevented by lowering the temperature, the eyes do
not then exhibit any gain in sensitiveness, but react like light-adapted
eyes. From the general nature of the statements, however, it is im-
possible to say whether it is the B and D deflections alone which are
affected, or whether C is also.

Since visual purple and the C deflection are both characteristic for
dark-adapted eyes of many animals, two questions arise: first, whether
visual purple is present in the eyes of all animals whose retinas, when
dark-adapted, yield the C deflection upon exposure to light; and second,
whether the C deflection is obtainable from all retinas which exhibit
visual purple. An answer to these queries must await the accumula-
tion of more experimental data on the subject.

In regard to whether the migration of pigment offers a basis for the
explanation of this slow-developing fluctuation in the curve, the num-
ber of facts at hand from which to judge is meagre. The migration of
- pigment and the C deflection are both elicitible from the dark-adapted
eye of the fish, frog, dove and lobster. In the retina of the dog on
the contrary, neither phenomenon occurs, for, according to Chiarini
(4), there is no evidence of photokinetic activity in the pigment, and
upon the statement of Piper (20) the C deflection is either entirely
absent or else is very small. In the pig’s eye the migration of pigment
is very slight according to Van Genderen Stort (21); and in the case of
the cat the same thing is probably true, judging from the scant amount
of pigment present in the epithelial layer. No retinogram has ever
been made for the pig, but for the cat v. Briicke and Garten (2) and
Piper (20) have photographically recorded pronounced C deflections.
Although Hesse (11) among others has described a migration of pig-

390 EDWARD C. DAY

ment in the cephalopod eye under the influence of light, the photographic
curves of the photoelectric effects obtained by Piper (20) from cephalo-
pod eyes reveal not the slightest trace of a C deflection.

Thus the evidence, although inconclusive because the two phenomena
have not been studied side by side in the same animal and under simi-
lar conditions, makes it rather improbable that the photokinetic activity
of the pigment underlies the phenomenon of the C deflection in the re-
tinogram. A study of the photoelectric effects in the eye of an albino
rabbit where pigment is entirely lacking, would throw direct light on
the problem. For the common rabbit the C crest has already been
recorded photographically by Piper (20).

The contraction of the cones of the retina, observed for the eyes of
the fish, frog, salamander, lizard, dove and pig [Van Genderen Stort,
(21); Chiarini (8, 4); Hertel (10)] might come into consideration in
seeking an explanation for the C deflection, were it not for the fact that,
although the latter phenomenon occurs in the arthropod eye [v. Briicke
and Garten, (2)], no such contraction of retinular elements takes place
in it under the influence of light [Parker (19); Day (5)]. In palaemonetes
Parker claims that, although the “proximal pigment cells’? undergo
no change in length through exposure to light, the “distal pigment
cells” become contracted, and the “‘accessory pigment cells’’ exhibit
‘migratory activity. But if, as has previously been shown, the mechani-
cal movement of pigment granules can not account for the C deflection,
it is no more to be expected that any other kinetic activity, whether of
contraction or the shifting of cellular elements, should account for it.
It is more probably to be explained upon the basis of a chemical reac-
tion of some substance such as visual purple, or possibly in terms of
the chromolysis in the retinal ganglion cells which, according to the
observations of Chiarini (3, 4), occurs under the action of light. Until
the substance can be definitely identified which underlies the phenome-
non of the C deflection, it goes by the general designation of the “Third
Substance” in the theories of Einthoven and Jolly (8) and of Piper (20).

In regard to the other deflections in the curve it is impossible from
the experimental data at hand to conclude just what their relations to
the visual sensations are. Ishihara (15) has argued that they can
not be correlated with sensations, first, because they are of too brief
duration to represent continuous sensations, and second, because the
1) deflection which should represent a sensation of darkness is the same
in sign as the B deflection which theoretically represents the sensation
of light. But the premises upon which these conclusions are based

PHOTOELECTRIC CURRENTS IN EYE OF FISH 391

are debatable issues themselves. The fact that the B and D deflections
are alike in sign does not necessarily exclude all correlation between
them and the sensations of light and darkness respectively, if one
assumes them to be produced by the reactions of two substances which,
although reacting with like directional sign as to the electric currents
produced, possess a qualitative difference which the galvanometer is
incapable of expressing. Were the instrument, for example, to record
the B deflection as a red curve and the D deflection as a blue one, the
directional sign remaining the same, any such qualitative difference
would immediately be apparent; thus red could be associated with the
sensation of light and blue with that of darkness.

As to the first argument, the premise that sensations of light are
continuous, is invalid in the sense of an unvarying continuity. It
was however evidently with that sense in mind that the argument
was formulated. According to Ishihara, for example, the sensation
which a person experiences when an electric light is turned on is a con-
tinued, unchanging sensation of brightness which lasts until the light
is turned off. An analysis of such a sensation however discloses the
error of the premise. When one gazes at the sky one may think that
with respect to brightness the sensation thereof suffers no apparent
change; but one needs only to perform two simple experiments in order
to convince himself to the contrary.

In a room having a window which commands a free expanse of sky,
cover one eye, hold a card in front of the other so as to screen off half
the patch of sky, gaze fixedly at a speck on the pane for a moment and
then withdraw the card. The half of the sky at which you were gazing
is seen to be distinctly dimmer than the one just uncovered. Again, in a
very dimly illuminated room, when one gazes at a white spot on a black
background, the spot soon vanishes like an extinguished light. The
element of fatigue, therefore, both for the retina as a whole and for the
fovea centralis, makes quite impossible a continuous sensation of light
in the sense of an unchanging sensation. Under ordinary conditions the
eye is ceaselessly shifting on its axis so as to bring now this now that
portion of an object to fall upon the fovea for close inspection, so that
the factor of fatigue is reduced to a minimum.

The sensation of darkness is likewise assumed by Ishihara to be
continuous, and that on that account the D deflection can not be the
correlative of the sensation. But one may gaze at a black field as well
as at > sky, cutting off half of the view with a card, and find upon
xe the card that the newly exposed half is comparatively the

392 EDWARD C. DAY

darker, thus showing that the sensation of blackness steadily diminishes
in intensity as its duration increases.

The decline of a deflection after reaching its maximum altitude might
therefore be the expression of the diminution in strength of a sensation,
whether of light or of darkness. The ascending phase of a deflection
would analogously indicate the growth of a sensation to a maximum.
In the foregoing I have had B in mind rather than A as being the repre-
sentative of the sensation of light; but that the A deflection can not be
ignored will be apparent from an analysis of the intermittent curve.

When one compares the threshold value for the fusion of visual
images with that for the blending of the oscillations in the intermittent
curve, it is found that there is a certain approximation between the
two. In the former case the threshold value for the human eye ranges,
according to Piper, from 20-50 stimulations per second, at which
frequency the intermittent flashes give the sensation of continuous
light, while in the latter case the value ranges, depending on the ani-
mal, from 17—40 flashes per second, at which the individual oscillations
blend into a smooth curve. From these comparative values one may
infer that some general correlation exists between the photoelectric
phenomena and the visual sensations of light and darkness. They do
not however throw much light on the question of the significance of
the individual deflections.

If the oscillatory curve is formed synthetically as has been described,
then the sensations in the brain corresponding to the oscillations must
be due, those of light to the depressions and those of darkness to the
crests in the curve; because, from the analysis of the curve, the A de-
flection plays the chief réle in creating the depressions and the D
deflections the crests. The B deflection here seems to participate but
little and consequently has little to do with the sensation of light.
Nikiforowsky (18) has shown furthermore that the B deflection vanishes
completely when the temperature is lowered to zero degrees centigrade,
while the A and D deflections remain almost unaffected. According
to this again, the A and D deflections, being the most stable, would
represent the sensations of light and darkness respectively; because,
were the sensation of light to be attributed to the B deflection, when
the latter vanished upon lowering the temperature it would be a case
of reductio ad absurdum, of an eye capable of rendering a sensation of
darkness but incapable of rendering one of light. In the simple curve
for the normal eye, however, the B deflection is sucha prominent feature

PHOTOELECTRIC CURRENTS IN EYE OF FISH 393

that further investigation is needed to decide whether it can be left
entirely out of consideration or not.

The preceding experiments dealt, it must be remembered, with dark-
adapted animals (with but one exception), so that both simple and
oscillatory curves must yet be studied for the effects of light-adaptation
before a definite theory can be well grounded.

E. Similarity of the tri-substance theories

In looking over the literature there have come to my attention cer-
tain facts in regard to the theories proposed by Einthoven and Jolly
on the one hand and by Piper on the other which are worth drawing
attention to. The two theories are supposed to be antagonistic to
each other, but they have in fact more in common than has been sus-
pected.

When Piper evolved his theory it was with the object (p. 120) of
obviating the necessity of ascribing each fluctuation in the curve to a
separate process in the retina, as he suppposed had been done by Ein-
thoven and Jolly. I think however that a careful analysis of the two
theories will show that Piper partly misinterpreted the application of
the theory of the other two men, and that what he sought to obviate
was something which he had imputed to their theory which did not in
reality belong there.

Table III will enable us easily to compare the two substance theories.
On the left side are given the characteristics of the substances as con-
ceived by Einthoven and Jolly, and on the right as conceived by Piper.
Substances II, I and III of Piper’s are comparable to I, II and III
of the Einthoven-Jolly theory. In each cross comparison it will be
-noted that the direction of the reaction is the same for the lighting
effects and also for the darkening effects and further that the latencies
of reaction for the several substances likewise correspond in the matter
of relative duration. During illumination the curves for Piper’s sub-
stances IT and I, after having attained a maximum positive and nega-
tive level respectively, remain constant on that level up to the off-effect,
whereas, according to Einthoven and Jolly they do not remain constant,
substance I continuing in the original direction and becoming more and
more negative, and substance II changing direction after reaching a
maximum positive value and returning to zero. Except for this dif-
ference in the continuous effects during the exposure, the three sub-
stances of the one theory have identically the same characteristics as

394 EDWARD C. DAY

the three corresponding ones of the other theory. Thus Piper’s sub-
stance II is nothing more nor less than substance I of Einthoven and
Jolly, and vice versa. The third substance is identical for both.
An apparently radical difference between the two theories comes
to light when the theoretical reactions of the substances are applied

TABLE III.
The tri-substance theory of Piper compared with that of Einthoven and Jolly.

EINTHOVEN AND JOLLY PIPER
Substance I Substance II
Lighting effect........... negative...... 3 elo eee negative.
During exposure......... increasingly negative..... constant negative level.
Darkening effect......... Positive: .. .. 2... 2.36 e4) USL Eve:
Latency of reactions..... relatively short.......... relatively short.
Requirements............| light eye exposed to a

flash of darkness

Substance II Substance I
Lighting effect........... positive to a maximum...) positive to a maximum.
During exposure......... diminishingly positive....| constant positive level.
Darkening effect......... MORALIVO. oo see eee eke negative.
Latency of reactions..... MCC... <). 3.41 Ae medium.
Requirements............| dark eye and short expos-

ure to weak light.

Substance III Substance III
Lighting effect........... positive to a maximum...) positive to a maximum.
During exposure......... diminishingly positive....| diminishingly positive.
Darkening effect......... HONG; 2. os"... ee none.
Latency of reaction...... fo) 1 sre ee long.
Requirements............ dark eye and relatively

strong illumination.

to the retinogram. In the theory of Einthoven and Jolly the reactions
of the three substances apparently occur in sequence, while by the Piper
hypothesis they progress simultaneously. It would seem, then, to be
a difference between driving horses tandem and driving them abreast.
It is true there is a commencement sequence for the reactions of the
three Piper substances, but the intervals between the beginning-times

PHOTOELECTRIC CURRENTS IN EYE OF FISH 395

of each are relatively short. In the construction which Piper placed
upon the statements of Einthoven and Jolly concerning their curves
on the contrary, the reactions of substance I proceeds to its finish be-
fore the reaction of substance II begins. This seems at least to have
been the conception of the E and J. theory held by Piper when, in
referring to it (p. 120), he judged it as physiologically impossible be-
cause it demanded a separate retinal process to account for each varia-
tion in the curve. In opposition to this idea it appeared therefore more
probable to Piper that the fluctuations in the curve were interference
phenomena resulting principally from two almost simultaneous reactions
acting in opposite directions and at slightly different rates; and upon
the basis of this conception he evolved a graphic schema by which he
explains the genesis of both simple and oscillatory curves.

Although this interference conception was developed and formulated
into definite terms by Piper, its essence was nevertheless present in
the theory of Einthoven and Jolly. That these investigators had also
conceived of interference occurring between overlapping reactions will
be seen from the following quotations (in which all italics are mine).

“The lighting effect (p. 400) of the first substance . . . does
not yet appear in fig. 14, which need not surprise us, since the darken-
ing reaction must appear sooner than the lighting reaction. Never- -
theless the lighting effect makes itself appreciable to some extent during
the record of the curve; for. . . . 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.’’ Again on p.
401, “‘the summit A, is swperposed in such a way upon the slow wave C
that the form of the waves may easily be recognized separately.”” On
p. 402 the interference conception is more expressly formulated in the
statement, “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.”’

From these citations it is clear that the nucleus of an interference
hypothesis existed in the minds of Einthoven and Jolly but that it
was never developed to the extent that it was later by Piper.

In many of my records the continuous effect of illumination is not
expressed in the curve as an undeviating course on a constant level, but
by a steady and continuous descent, a phenomenon not interpretable
upon the basis of the Piper theory that the reactions of substance I

396 EDWARD C. DAY

and II remain in a state of equilibrium during the period of illumina-
tion. Either substance I must diminish in the strength of its positive
reaction or else the reaction of substance II must steadily grow in
strength negatively, in order to account for this fall in the curve which
seems to be a constant feature in retinograms of dark-adapted fish
eyes. Nor can the difference between the reactions of dark and light
eyes, a phenomenon not taken into consideration by Piper in testing
his theory, be otherwise accounted for than by the foregoing assumption.
I am therefore inclined to agree with Einthoven and Jolly that the
reactions of substances I and IT, instead of reaching a level of stability,
vary constantly, either waxing stronger or growing weaker, throughout
the period of illumination.

BIBLIOGRAPHY

(1) Brossa AND Kouurauscu: Arch. f. Anat. u. Physiol., physiol. Abt., 1913.
449.

(2) BrickE AND GarRTEN: 07. Arch. f. ges. Physiol., 1913, exx, 290.

(3) Curarini, Arch. Ital. Biol., 1904, xlii, 303.

(4) Curartni: Arch. Ital. Biol., 1906, xlv, 337.

(5) Day: Bull. Mus. Comp. Zool., 1911, lili, 305. .

(6) Dewar AND M’Kewnprick: Trans. Roy. Soc. of Edinburgh, 1874, xxvii, 141.

(7) pu Bots-Rrymonp: Untersuchungen iitber thierische Elektricitit, 1849,
Bd. 2, Abt. 1, pp. 256-257.

(8) EtInTHOVEN AND JOLLY: Quar. Jour. Exper. Physiol., 1908, I, 373.

(9) Gorcu: Jour. Physiol., 1903, xxix, 388.

(10) Herre: Arch. f: Augenheilk., 1907, lviii, 229.

(11) Hesse: Zeitschr. f. wiss. Zool., 1900, lxviil, 379.

(12) Hotmeren: Upsala Likareforenings Forhandlingar, 1866, I, 184.

(13) Hotmeren: Untersuchungen aus dem physiol. Inst. der Univ. Heidelberg.
1880, III, 278.

(14) Hotmcren: Untersuch. aus dem physiol. Inst. der Univ. Heidelberg, 1882,
II.

(15) Isurmara: Arch. f. ges. Physiol., 1906, exiv, 569.

(16) Kine anp Steiner: Untersuch. aus dem physiol. Inst. der Univ. Heidel-
berg, 1880, III.

(17) Kitune AND Steiner: Untersuch. aus dem physiol. Inst. der Univ. Heidel-
berg, 1881, IV, 64. Verhandl. d. Heidelberg. Naturhist.-Med. Vereins,
IN. Serie; 18815 i, 1.

(18) Nrxirorowsky: Zeitschr. f. Biol., 1912, lvii, 397.

(19) Parker: Bull. Mus. Comp. Zool., 1897, xxx, 275.

(20) Pirer: Arch. f. Anat. u. Physiol., physiol. Abt., 1911, 85.

(21) Van GENDEREN Srort: Arch. Neérl., 1887, xxi, 316.

(22) Water: Quar. Jour. Exper. Physiol., 1909, ii, 169.

.
PHOTOELECTRIC CURRENTS IN EYE OF FISH 397

+ DESCRIPTION OF PLATES
The figures in the two plates are tra¢ings from photographic records.
PuaTE |

Fig. 1. Experiment 40. Date, °13/III/7. 12.05 p.m. Exposure No. 2.
The fish was brought fresh from the market and photographic records were made
immediately. Exposure No. 1 yielded a curve similar to No. 2 but the latter
was selected for reproduction because of its clearer image. This record was
made after the eye had been adapted to the Nernst light for two minutes to
counteract the twe minute interval of dark adaptation necessitated by the devel-
opment of exposure No. 1. The off-effect for the light-adapted eye is here seen
to be stronger than the on-effect. B and D deflections manifest.

Fig. 2. Experiment 40. Date, ’13/III/7. 12.18 p.m. Exposure No. 3,
following exposure No. 2 (fig. 1) after 10 min. dark adaptation. The on-effect
is now stronger than the off-effect. B and D deflections present and a faint
suggestion of A.

Fig. 3. Experiment 40. Date, ’13/III/7. 12.27 p.m. Exposure No. 4,
following expos. No. 3 after 7 min. more dark adaptation. Deflection A is
now fully present, B has increased in strength while D has grown feebler.

Fig. 4. Experiment 39. Date, 713/III/6. 11.33 a.m. Exposure No. 1.
Fish dark-adapted over night. A, B, C and D deflections present. Although
the D deflection in this exposure is stronger than the B, in the other exposures
of this experiment the reverse relation is found. The larger D deflection may,
in this case, have been due to a slight condition of light adaptation at the begin-
ning. This light adaptation was not however strong enough to repress the C
deflection.

Priate II

Fig. 5. Experiment 39. Date, ’13/III/6. 12.04 p.m. Exposure No. 4.
Fish was kept dark-adapted after each of the previous exposures. The illumina-
tion, recorded by the dashes in the lowest line of the figure, was intermittent, the
rate of stimulation being about three flashes per sec. The duration of a single
flash is about 1/6 sec. A, B and D deflections are present. The last D de-
flection is superposed upon the last B deflection owing to the curtailment of the
final period of illumination.

Fig. 6. Experiment 37. Date, ’13/III/4. 4.27 p.m. Exposure No.5. Fish
was dark-adapted over night. Rate of stimulation, 16 flashes per second. Inter-
vals of light = 0.029 sec., of darkness = 0.037 sec. To the 32 flashes of light there
correspond 32 crests in the curve.

Fig. 7. Experiment 37. Date, °13/III/4. 4.47 p.m. Exposure No. 6.
Rate of stimulation, 28 flashes per sec. The individual oscillations have fused
and the effect is the same as for continuous illumination.

Fig. 8. Experiment 37. Date 713/III/4. 5.25 p.m. The three deflections
were produced by substituting a galvanic battery for the eye of the fish and send-
ing three consecutive shocks through the galvanometer having a current strength
of 0.1 millivolt. The current from the eye throughout Experiment 37 gave throws
of scarcely half this altitude.

PHOTOELECTRIC CURRENTS IN EYE OF THE FISH PLATES I anp IT

EDWARD C. DAY

.
d
refinogram b
Fig. /
time | Ve See
exposure —
ii on
4
d
“| Fig. 2
AS She es sae
d
b
*) Fig. 4
ds o, Tbs ae
a a
a
Fig. 5
dd
aa
Fig. 6
d 4
a
Fig. 7
Fig. 8

ZOOLOGICAL LABORATORY, SYRACUSE UNIV., SYRACUSE, N. Y¥.

Ono

———

THE COMPARATIVE RATE AT WHICH FLUORESCENT
AND NON-FLUORESCENT BACTERIA ARE KILLED
BY EXPOSURE TO ULTRA-VIOLET —

W. E. BURGE anp A. J. NEILL
From the Physiological Laboratory of the University of Illinois

Received for publication July 28, 1915

The source of ultra-violet used was a Cooper-Hewitt quartz mercury-
vapor burner operating at 170 volts, 3.8 amperes and 2400 candle-
power. The non-fluorescent bacteria used were B Colil communis,
B. violaceous, B. Proteus vulgaris, B. sulbtis, Sarcina aurantiaca,
Micrococcus capsulatus, and B. mucosus capsulatus. The fluorescent
bacteria were furnished by the State Water Survey of Illinois. The
Water Survey obtained the bacteria from the samples of water sent in
from different parts of the state for examination. Mr. Tanner of the
Water Survey had not gone far enough in the classification of these
bacteria at the time we obtained them from him to furnish the names
for them and for that reason they were referred to in this paper as
fluorescent bacteria numbers 1, 2, 3,4,5,6and7. He had determined,
however, that they were all different types of fluorescent bacteria and |
were non-spore formers.

The liquids containing the bacteria to be exposed to ultra-violet
were prepared in the following way. A standard loop of bacteria was
taken from a five day old culture and introduced into 25 ce. of 0.8 per
cent sterile sodium chloride solution. This liquid was filtered through
a fine grained sterile filter. Two ce. of the clear filtrate containing the
bacteria were introduced into each of eleven quartz tubes. These
tubes were placed 30 cm. from a Cooper-Hewitt quartz mercury-
vapor burner operating at 2400 candle power. At intervals of 20
seconds a tube was removed. In this manner tubes containing the
bacteria were obtained which had been exposed to the light for 0, 20, 40,
80, 100, 120, 140, 160, 180, 200 seconds respectively. The contents
of these tubes were plated with agar and incubated at 37° C. for 48
hours. At the end of this time counts were made according to the
ordinary bacteriological method.

399

400 W. E. BURGE AND A. J. NEILL

Results of these counts may be seen in Table | for eight kinds of
fluorescent bacteria and in Table 2 for seven kinds of non-fluorescent
bacteria. The numbers of colonies for the 20 and 40 second intervals
of exposure are not given, because as a rule they were too numerous to
be counted accurately. The counts for the numbers of colonies after
0 time of exposure were made possible by diluting the liquid a million
times before plating.

It may be seen in Table 1 that there were alive 31 colonies of fluores-
cent bacteria No. 1, 15 of No. 2, 3 of No. 3, 3 of No. 4, 11 of No. 5, 2

Table I. Fluorescent Baoteria
Time of Number of Number of Number of Number of Number of Mumber of Number of Number of
exposurs bacteria bacteria vacteria bacteria becteria bacteria bacteria bacillus
No.l No.k No.3 No.4 No.5 No.6 No.7 pyocyaneus
0 Seo. 72,000,000 80,000,000 28 ,000 ,000 70,000 ,000 75,000 ,000 33,000 ,000 25,000,000 79,000,000
20% www anne ens cee n nen nn = cee ene nn ewan eee ene eae en eee seme eeeee= en eeeenee = soon eaen=
400 9 nanan en nee nn nnn nn nee rere nnn - = rene aren ne enna n nen renner nnn renee en ne- teen n-----
60 750 813 55 210 460 12 361 88
80 459 623 29 79 211 10); 112 20
100 " 222 221 25 56 113 7 16 14
120 128 102 20 29 103 5 8 12
140 ” 97 88 15 18 56 4 6 8
i60 " 81 73 12 13 38 3 6 2
a¥foy 43 29 10 6 25 3 5 1
200 ” 31 15 3 3 il 2 4 B
Table II. Non-fluorescent 3acteria
Time of Number of Number of Number of umber of Number of Number of fMumber of
exposure 8,Subtilis B.Mucosus BeProteue B.Violaceus Micrococcus 8,Coli Sarcine
oapsulatus vulgaris capsulatus communie esurantiecs
0 Seo. 13,000,000 24,000,000 295,000,000 £7,000 ,000 28,000,000 268,000,000 195,000,000
S000 Maseesesed | wepeeneeee | eesenecener © eweew= nea N oeee~esseet ase bbee amet ase
POY “CD SER oReee, eee Seo oe ss WERE SSS ee eee
(oe 13 460 2,480 83 171 378 4,820
80 =" Lg 112 79 60 4 9 60
10c —=—* 7 107 71 20 2 1 10
320) 6 8 4 5 2 0 2
140 * 4 6 2 4 1 oe] 1
rT) 2 5 2 te) le) ° 1
180 " 1 ° 1 te) ° ta) 0
200 ”" 0 0 ie) 0 1°) 0° i)
of No. 6, 4 of No. 7, and 1 of B. pyocyaneus after an exposure of 200

seconds to ultra-violet while in Table 2 it may be seen that all of the
non-fluorescent bacteria were killed when exposed for a similar length
of time. This is taken to mean that the fluorescent bacteria are more
resistant to the effect of ultra-violet than non-fluorescent bacteria.

The results of these experiments are also expressed in the form of
curves in figures 1, 2, 3, 4, 5,6, and 7. The unit of the ordinate seale
is 1 em. to one hundred bacteria, that of the abscissa is 1 em. to 20

roe) ee 2 Se

Number of Bacteria Living.

20 “# oo eo 100 120 140 /60 (80 200
Pig. 1. Seconds
Fluorescent bacteria "No. 1
———— Non-fluorescent B.subtilis (two deys old)

——-— Non-fluorescent B.subtilis (two weeks ola!

Pee LL.

J
JENSSEN
eee
_SRERSGeSS

g

Nember of Bacteria Living.

=
&

40 60 60. 0 140 140 (60 (80 #00
Pig. 2 Seconds
Fluorescent bacteria No. 2

——-—- B.muicosus capsulatus

_ Tse
_
meses ET
tet ie NET
meme bok
Pea ety |
SESS
JER S6n2eae

an /eo [80 200
Seconds ©

i;

Number of Bacterra Living
i. Hex FE

.

“ eo 40 ee
Fig. 5S.
Pluorescent bacteria

(20 (#0

Ho. 5

Sarcina aurantiace

B.pyocyeneus

F928

Nember of Bacteria Living.

“ae £00

Fig. 6. Sacones

—— Fluorescent bacteria No. 6

——-—— Micrococcus capsulatus

401

#8

Pe hs
A ie
BERSSS2ees
Garth A pigs gt

Pluorescent bacteria

Number of Bacteria Living

eo

6

Number of Bacteria Living

Pane ee mane rede
ia ddl il ais Se El

b— SSS
(eo ke

a OS
Fluorescent bacterin No. 4

(lo /@o (80 200

Jecords

——-—— B.violacens

hort

No. of Bacteria Living
g

Fluorescent bacteria No. 7

———— B,¢oli communis

402 W. E. BURGE AND A. J. NEILL

seconds. Each curve marked N is plotted to a scale whose ordinate
is 10 times as great as that marked M.

It may be seen from the curves that the non-fluorescent bacteria
are killed more rapidly than the fluorescent bacteria. Of the non-
fluorescent bacteria B coli was the least resistant to the action of ultra-
violet, being killed after an exposure of 100 seconds, while B. subtilis
was the mst resistant, being killed after 180 seconds. No attempt was
made to determine how long an exposure was necessary to kill all the
bacteria in any of the cultures of the fluorescent bacteria.

The questions of most interest to us were why are bacteria killed
by ultra violet and why are the fluorescent bacteria more resistant than
the non-fluorescent. Dreyer, Hanssen! and others found that the
ordinary proteins such as serum albumin, serum globulin, egg albumin
and egg globulin, etc., are comparatively easy to coagulate by ultra-
violet. In view of this observation it seemed to us that the most
plausible explanation for the bactericidal action of ultra-violet is that
the short waves coagulate the protein or protoplasm of the bacteria and
by this means kill them. This assumption while explaining the bac-
tericdical action of the short waves does not, however, explain the
fact that fluorescent bacteria are more resistant to the action of ultra-
violet than the non-fluorescent.

Von Graefe and Helmholtz? showed that the lens possesses the prop-
erty, fluorescence, by which it converts the absorbed short waves into
longer waves and gives them off in this form. Burge,’ while carrying
out experiments in an attempt to determine the part played by ultra-
violet in the production of glass blower’s cataract, found that the lens
protein differs from most of the other proteins in that it is very difficult
to coagulate by means of ultra-violet. He also showed that the
lens protein could be modified by the salts of calcium and magnesium
and by sodium silicate, substances found to be greatly increased in
human cataractous lenses, so that the short wave-lengths of the spec-
trum could coagulate it and hence produce an opacity of the lens
or cataract. He found that these salts decrease the fluorescence of
the lens at the same time that they render the lens protein easily coagu-

‘Comptes Rendus, 1907, cxiv, 234.
* Von Graefe and Helmholtz: Ueber die Fluoreszenz der retina. Poggendorfi
Annalen der Physik und Chemie, Bd. xciv
’ Burge: This Journal, 1914, xxxvi.
Hanssen: Comptes Rendus, 1907, exiv, 234.

EFFECT OF ULTRA-VIOLET LIGHT ON BACTERIA 403

lable by ultra-violet. This last observation suggested that there might
be some relation between the fluorescence of the lens and its great re-
sistance to the coagulative effect of ultra-violet. To explain the great
resistance of the lens protein the assumption was made that the lens
by converting the short waves into longer waves and giving them off
as such disposes of the energy of the absorbed short waves, which
otherwise would have been spent in coagulating its protein. The fact
that these substances which render the lens protein more easily coagu-
lable by ultra-violet at the same time decrease the fluorescence of the
lens would seem to lend support to this hypothesis.

To explain the great resistance of fluorescent bacteria to the action
of ultra-violet an assumption is made similar to that made in explaining
the resistance of the lens protein to the action of ultra-violet, namely
that fluorescent bacteria protect themselves from the coagulative
effect of ultra-violet by converting the short wave-lengths into longer
waves and hence dispose of the energy of the absorbed short waves.
The non-fluorescent bacteria, however, are unable to dispose of the
energy of the absorbed short waves and for this reason they are killed
more easily by ultra-violet than the fluorescent bacteria.

THE EFFECTS. OF CHANGE IN AURICULAR TONE AND
AMPLITUDE OF AURICULAR SYSTOLE ON
VENTRICULAR OUTPUT

ROBERT GESELL
From the Laboratory of Physiology, Washington University Medical School

Received for publication July 24, 1915

What we know concerning the role of the auricles in the dynamics
of the heart has been largely a matter of inference: the inference being
that the auricles have a passive and an active role; acting passively
as reservoirs for the accumulation of blood which is to pass into the
ventricles, and actively as pumps, injecting blood into the ventricles
and producing a better ventricular filling. The importance of good
ventricular filling is apparent when we consider that the ventricle,
within certain limits, tends to’ expel with each systole most of the
blood it receives in the preceding diastole. The auricles regulate in-
directly by their contraction, the volume output of the ventricles.

To what extent blood pressure and ventricular output are affected
by auricular systole has not been a subject of much experimentation.
Henderson (1) using the cardioplethysmograph on the heart of the
dog, came’ to the conclusion that auricular systole had practically no
filling effect on the ventricles.

With the use of another method, I showed (2) that auricular systole
occurring at the normal time interval in the cardiac cycle, increased
the blood pressure by 55 per cent of that maintained by the filling action
of venous pressure alone. While the filling effect of auricular systole
on the ventricles was not directly recorded, it seemed a logical infer-
ence, from the experiments then performed, that the increased ventric-
ular output was due largely to an increased ventricular filling. How-
ever, at the time, it was considered probable that auricular systole
affected ventricular output in other ways. Also it was argued

How many factors are there at work in the interval of auricular systole under
consideration causing these large blood pressure changes. Whether the relative
amounts of blood forced into the ventricles, and the varied perfection of valvular
action alone can account for the changes, is very difficult to say.

404

INFLUENCE OF AURICLE ON VENTRICULAR OUTPUT 405

The physiological condition of the yentricular muscle at the moment of ven-
tricular stimulation may be of considerable importance. It is a well known fact
that a properly stretched miscle gives a far more efficient contraction to a given
stimulus than a relaxed muscle. The same probably holds true with ventricular
muscle, without in any way being at variance with the ‘all or none law.’ Straub’s
records clearly show that the tension of the ventricles is markedly increased by
auricular systole, and lasts a short but appreciable time. This period of tension
may be of considerable importance. Probably that auricular systole producing
it just at the moment the ventricular impulse reaches the ventricles is the most
efficient systole in producing the largest end result, namely, ventricular output.

It was to determine the relative importance of increased ventricular
filling, perfection of valvular action, and presystolic intraventricular
pressure, on ventricular output, and also to determine if other factors
are at work, that this research and others concerning the role of auric-
ular systole were undertaken. The experiments reported in this
paper have primarily to do with ventricular filling and valvular action.

For purposes of comparison, the method used in the previous work
will be briefly reviewed.

The heart of the dog was exposed and independence of the auricles
and ventricles established by crushing the auricular ventricular bundle
with the Erlanger heart clamp. Arterial and venous pressures, and
contractions of the auricles and ventricles were recorded. The ven-
tricles were stimulated at approximately the rate of auricular con-
traction. By such stimulation “interference waves’’ similar to inter-
ference waves of sound, were produced. That is, the auricles and
ventricles at times would beat in accord and at other times in varying
degrees of interference. Auricular systole was placed in different
phases of ventricular cycle; at one time occuring during ventricular
systole, another time early in ventricular diastole, and at another time
‘immediately before ventricular systole.

The method showed that the effects exerted by auricular systole
on increasing ventricular output vary with the place occupied by auric-
ular systole in ventricular cycle. The auricular systole placed at
about the normal time interval in the normal heart cycle was found
to be most effective. In these experiments the magnitude of auricular
systole remained constant, the varying factor was time—the time in-
terval at which auricular systole occurred in ventricular cycle.

In the present method, for a particular reason the heart of the river
terrapin was used. The auricles of that heart show two distinct types
of contraction, the rapid clonic and the slow tonic. The clonic are
superimposed upon the tonic, and, asa rule, vary in amplitude, inversely

406 ROBERT GESELL

as the height of the tonic contraction. These properties of the auri-

cles allow the studying of the effect of auricular systole on cardiac
dynamics in an entirely different way from that of the previous method
mentioned. It permits a study in which the magnitude of auricular
systole is the varying factor, while the time interval at which auricular
systole occurs in ventricular cycle is the constant factor—the direct
opposite of conditions prevailing in the experiments in Method I.
In Method I, the possibility of variation in the perfection of valvular
action, until eliminated, might be considered a possible controlling
factor. In Method II, where auricular systole precedes ventricular
systole by the same constant time interval, this possibility is reduced
to a minimum.

With changes in auricular tone, the auricular output varies greatly.
Therefore, it was thought that if ventricular output varied as a result
of auricular tone changes, these variations could be ascribed primarily
to different filling effects of the auricles on the ventricles, rather than
to variations in valvular action.

To determine whether ventricular output varied with auricular
tone, the method represented in the schema was employed.

The heart was perfused in situ with Ringer’s solution. The contrac-
tions of both auricles were recorded and the ventricular output measured
directly. The left Cuvierien duct and hepatic veins were ligated and
a cannula of maximum capacity was inserted in the right Cuvierien
duct. This cannula was connected with a Marriotte bottle which
delivered the Ringer’s solution to the base of the heart at any constant
pressure. All the arteries coming from the truncus arteriosus with
the exception of one were ligated. This artery was connected by
cannula with a differential volume recorder.

This recorder consists of a three-way connecting tube with two
long vertical arms running upward and one short one running down-
ward. The short vertical tube is supplied with a stop cock. Into one
of the long vertical tubes the volume flow to be measured is led. In
the other tube is a cork float carrying a glass writing point which
records the height of the liquid in the recorder. The stopcock is so
regulated that the mean volume flow when led into the recorder will
maintain the level of the liquid at about the center of the tube. When
the volume flow decreases, the outflow is greater than the inflow,
the level of the liquid falls to a point where its pressure will reduce
the recorder outflow to the recorder inflow and maintain the writing
point at a lower level. When the volume flow increases, the reverse

INFLUENCE OF AURICLE ON VENTRICULAR OUTPUT 407

occurs: the liquid rises until by its,increased height it produces a re-
corder outflow equal to the recorder inflow. The sharpness of the
curves indicate the suddenness of change in volume flow.

Without exception all the hearts experimented upon showed marked
oscillations of auricular tone. The frequency of the waves varied with
the animal. In some experiments in which oscillations were infrequent
additional ones were artificially produced either by mechanical irrita-
tion (tug on the auricular suspension), or by flooding with a few ccm.
of cold Ringer’s solution. The results in all cases were the same.

Numerous records were made with venous pressures varying from 10-
60 mm. of water. Tone waves occurring with all these pressures had
an effect on ven-
tricular output.
When the higher
venous pressures
were employed,
however, the tone
waves were not
nearly as marked,
and changes in ven-
tricular output ac-
companying auric-
ular tonus changes
were correspond-
ingly less marked.
With venous pres-
sure at about 25
mm. of water the Fig. 1.
tone waves were
most marked and consequently with this pressure the effects of vary-
ing amplitude of auricular systole were best studied. This pressure,
apparently, permitted heart action as normally seen in the exposed
heart of the turtle.

Figures 2 and 3 show the usual effects of changes in auricular tone
on ventricular output. The corresponding points of the curves are
marked. In the case of figure 1 the decreased ventricular output
occurring with increase of tone is large. Before the onset of a tone
wave, the output was 31 ccm. per minute. This decreased to 10 ecm.
per minute. The decrease in ventricular output occurs immediately
with the onset of the tonic contraction of the auricles, but becomes more

i
si leks MN IIAAARNNAARRAND NTN

bf Wig y aa eevee yy an
ih 4 aa ene tae Mayall
i Wily My ideal
yl Ma yl

f

Whi fu My] jalylV

RAAARNAACRAARARARR AIAN

CARRIURRIRADASAIIS

DIDI RAL DIDS

RARIARAS ANANSI SP

RADA LL LD DPDLLDRAA RA
PPA AN

rer rrtrte
An

410 ROBERT GESELL

marked at the crest of the tonic contraction. The late appearance
of the marked decrease undoubtedly is due to the masking effect of
the tonic contraction itself. The tonic contraction is fairly rapid, and
the filling effect resulting from it astonishingly effective. In some
cases where the tonic contractions are exceptionally large, the volume
output of the ventricles may be increased during the period of short-
ening of the tonic contraction, even though the amplitude of the indi-
vidual auricular systoles are progressively decreasing: the decreased
amplitude of auricular systole is largely compensated by the effective-
ness of the tonic contraction. The marked decrease of ventricular
output occurs where the tonic contraction has reached its crest and the
ventricles are filled only by venous pressure and small auricular systoles.

Figure 3 shows the decreased volume output followed by partial
recovery. The early compensatory effect of the tonic contraction for
the decreasing amplitude of auricular systole is shown again. After
the completion of the first part of the tonic contraction, which is the
most effective, the volume output decreases rapidly and begins to in-
crease again only with the relaxation of the tonic contraction and
increase in amplitude of auricular systole. The increase in volume flow
is very gradual, corresponding with the gradual increase in the amplitude
of auricular systole. The volume flow never reaches the initial volume
output for two reasons: the initial auricular amplitude does not return
and the auricle remains in higher tone, consequently having a smaller
capacity, which in turn reduces the auricular output. The objection
might be raised that the reason for the rapid decrease in ventricular
output occurring during the relaxation phase of the tonic contrac-
tion is to be found in a diminished venous pressure at the base of the
ventricle caused by blood accumulating in the easily dilated auricles.
With this objection in mind, a cannula which could deliver from 10-
30 times the ventricular output was inserted into the right Cuvierien
duct. The very gradual return of volume output occurring with the
gradual return of initial auricular amplitude in figure 2 speaks against
this objection.

In order to be in a position to ascribe changes in ventricular output
to accompanying changes in auricular tone as due purely to an au-
ricular effect, it is necessary to demonstrate that the ventricle itself
is not participating in the tone changes. Measurements of the length
of ventricular fiber with a myocardiograph were not made, because of
the danger of interfering with ventricular contraction and because
rhythmical variation in length of ventricular fiber occurring synchron-

INFLUENCE OF AURICLE ON VENTRICULAR OUTPUT 411

ously with oscillations of auricular tone would not necessarily be an
index to true tonus changes, but rather to differences in the filling of
the ventricle occurring during various stages of auricular tone oscillation.

From the experiments I performed, it seemed reasonably safe to as-
sume that true ventricular tone changes did not accompany or follow
tone changes in the auricles. The number of tone oscillations in the
right auricle were much greater than those in the left. The right and
left auricles though much more intimately connected with each other
than the auricles and ventricles, seldom showed synchronous tone
oscillations. When the oscillations were synchronous, one could not
feel sure whether the wave was conducted from one auricle to the
other or whether it occurred independently in both auricles.

Rosenzweig (3) states that tonus waves may occur synchronously
in both auricles and in auricles and ventricles. Fano (4), however, in
100 experiments found the ventricle to show tone oscillation in only
three cases.

In view of these facts, and of the fact that ventricular output in-
variably is affected by changes in auricle tone, it would seem fair to
ascribe these changes in ventricular output to auricular tone changes.

Henderson and Johnson (5) in their discussion of valve closure by
back pressure and broken jet, state that my previous results may be
adequately explained by variation in perfection of valvular action, the
valve at times closing by hinge movement from back pressure with
considerable regurgitation of blood. This explanation of my results
loses much of its force in the present experiments where auricular
systole precedes every ventricular systole by the same time interval,
thereby exerting its broken jet action on valve closure at all times at
the proper moment.

The auricular effects in these experiments apparently are not due
to variations of valvular action but to:

1. Variation of auricular diastolic volume. This is a factor of prime
importance since it permits a varying auricular output, provided the
auricles tend to empty themselves with every contraction and do not
meet with materially increased resistance when the auricular output
is large than when it is small.

2. Variation in the ability to develop tension in varying phases of
tonus oscillations. Patterson, Piper and Starling (6) place consider-
able importance on the relation between the initial length of mam-
malian ventricular fiber and tension developed in ventricular systole.
If the auricle of the turtle shows the same relation, the auricular sys-

412 ' : ROBERT GESELL

tole occurring when the auricular tone is low should be able to overcome
more tension and fill the ventricle better. (I am at present investi-
gating this question.)

1 and 2 would work in the same direction by producing a better
filling of the ventricle and in turn an increased efficiency of ventricular
fiber.

If presystolic intraventricular tension is of any importance, this
factor would vary inversely with the height of tone and work in the
same direction as the other auricular effects.

SUMMARY

The effects of variation of amplitude of auricular systole on ventric-
ular output were studied on the heart of the river terrapin.

Because of the oscillations of auricular tone, this heart is peculiarly
adapted to this research.

Perfusion of the heart with a constant venous pressure showed varia-
tions in ventricular output resulting from oscillations of auricular
tone.

Reasons are given for considering variation of valvular action as
a minor or negligible factor.

The variations in ventricular output are attributed mainly to the
direct effects of variations in ventricular filling.

The slow tonic contractions as well as the rapid clonic contractions
of the auricles exert a filling effect on the ventricle in proportion to
the amplitude of these contractions.

The decreasing effect of auricular systole of decreasing amplitude
occurring during the shortening phase of the tonic contraction is
masked by the filling effect of the tonic contraction.

The increased ventricular output resulting from auricular systole
of large amplitude is discussed and attributed to

a. Large auricular content at the disposal of auricular systole, this
tending to increase the auricular output or ventricular filling.

b. Greater length of auricular fiber, this increasing the efficiency
of the auricular contraction and therefore the volume of auricular
output, and to

c. Greater ability to increase the length and tension of the ventric-
ular fiber, thereby producing a greater efficiency of ventricular systole.

INFLUENCE OF AURICLE ON VENTRICULAR OUTPUT 4138

REFERENCES
(1) Henperson: This Journal, 1906, xvi, 325.
(2) GesELL: This Journal, 1911, xxix, 32.
(3) Rosenzweia: Arch. f. Physiol., 1903, supplement, 192.
(4) Fano: Beitr. zur Physiol. C. Ludwig gewidmet, 1887.
(5) HENDERSON AND JoHNSON: Heart, 1913, iv, 69.
(6) Patrerson, PreerR AND Starwina: Journ. of Physiol., 1914, xlviii, 465.

™ eeett ® all ALA ny on : ®
ai aie iF, rete y othe AY, =<
§ *% A sad ibe "ae? mer oa WG
; ale oral. cere
; re hs sea @ , aq
ers ’
| ove
<3 er, ro
7 ' Ce a a
acyl de eee r
¥ tine 4 fT hen
ot gi ae (TR: ea A ba ee:
aah a A os RMR :
‘ ( r | tare . i ,
wh ey af ‘ =
T Widi ay 4 ! : 7 a ae |
re Lia gr, ' AAs
' ee et ahd w °
- ¥) ; Fikaea a A ar"
rPé q
® pee. ee
. *
nh id
i.
'

&
ie
y
a
-
> » :

nh
a _
=

el

THE AMERICAN
JOURNAL OF PHYSIOLOGY

VOL. 38 OCTOBER 1, 1915 No. 4

STUDIES IN EXPERIMENTAL GLYCOSURIA

IX. Tue LevEL oF THE BLogp-SuGaR IN THE DoG UNDER
LABORATORY CONDITIONS

J. J. R. MACLEOD anp R.'‘G. PEARCE

From the Laboratory of Physiology and Biochemistry, Western Reserve Medical
School, Cleveland, Ohio

Received for publication August 21, 1915

It is a well-established fact that the anaesthetizing and preparation of
laboratory animals causes a varying and often very considerable degree
of hyperglycaemia to develop. Although there can be no doubt that
the influence of certain factors on the blood-sugar level will be masked
by the presence of this condition, yet it has so far been impossible to
obviate it by procedures which are not open to objections that are prob-
ably as serious as the presence of the excess of sugar in the blood.

Both E. L. Scott (1) and Shaffer (2) have recently shown that the
blood-sugar level in the dog and eat is really much below that usually
taken as the ‘normal’ in experiments bearing on sugar metabolism.
Seott concludes that ‘if consistent results are to be expected, the ani-
mals (cats) must be . . . . killed without pain or excitement,”’
and Shaffer, working on the dog, shows that the removal of blood from
a vein with entire avoidance of any excitement on the part of the
animal, yields blood-sugar values which are about one-half those which
have been considered by previous workers as normal. Both inves-
tigators emphasize the diabetic influence of the excitement caused by
giving a general anaesthetic, and in a later communication, Shaffer
(along with Hubbard (3)) shows that the initial hyperglycaemia may -
be avoided, or, if present, caused almost to disappear by forced respira-
tion through a tracheal cannula.

415

416 J. J. R. MACLEOD. AND R. G. PEARCE

Similar observations by other investigators are referred to fully in
the above papers.

If we are to confine our investigations on the behaviour of the sugar
level of the blood to unanaesthetized animals, however, a very limited
field will be open to us, for we shall be unable to study many of the
factors which influence the glycogenic functions of the liver. For pur-
poses of following the blood-sugar level from day to day, the precautions
advocated by the above workers must henceforth be strictly adhered
to, but for the large group of observations requiring surgical interfer-
ence, and therefore anaesthesia, an entirely different method must be
followed in obtaining normal standards with which to compare the ex-
perimental results. Shaffer and Hubbard suggest that in such cases
the initial hyperglycaemia, since it is largely due to dyspnoea, should
be avoided by forced artificial respiration. One of us (J. J. R. M.) (4)
some years ago employed this method to avoid the dyspnoeic hypergly-
caemia brought about in dogs by stimulation of the central end of the
vagus nerve. It was also used in experiments involving stimulation
of the splanchnic nerve. The forced ventilation was always found to
cause a decline in blood-sugar, but the method was abandoned as a
routine practice in subsequent work because it was concluded that the
depression in the CQ,-tension of the blood involved in using it would
introduce conditions which were at least as unphysiological as those
due to a slight excess of sugar. We have frequently reconsidered the
advisability of using forced respiration or insufflation of oxygen in our
experiments, but as work has accumulated showing, on the one hand,
the close relationship between the CO,—tension of the blood and its
reaction and, on the other, the susceptibility of the physiological activi-
ties of nerve centers, as well as of many organs and tissues—including
ulmost certainly the glycogenic function of the liver—to the reaction of
the blood, we have abandoned the idea. .

These difficulties in securing constantly low ‘normal’ values for
blood-sugar in laboratory experiments make it necessary to adopt other
standards with which to compare the results obtained during some ex-
perimental procedures. The ideal being unattainable at present, there
remains available one of two methods, either: (1) to determine the blood-
sugar level of each animal for some time before bringing about any ex-
perimental change, or (2) to use, for comparison with the experimental
animal, data secured from a sufficient number of anaesthetized animals
in which all the conditions are as similar as possible to those obtaining
in the experimental animal.

LEVEL OF BLOOD-SUGAR IN THE DOG 417

We have hitherto employed the former of these methods although, as
already stated, we have recognized the limitations due to the fact that
the initial disturbances dependent upon etherization, etc., might mask
many of the results. For further work which we contemplate, how-
_ ever, it will be necessary to employ the second method, and some of the
data which we have collected to this end are given in the present paper.

Methods. Dogs were employed in all the experiments. Most of
these were removed from the stock kennels several days before being
used and were either starved or given a liberal diet of bread broken up
inameat broth. Certain of the animals were given 5—7 grams of cane
_ sugar per kg. body weight, dissolved in water, by stomach tube, on the
evening preceding the experiment. Etherization was brought about as
quickly as possible, and the animal was then tied out on a warmed
operating table and a tracheal cannula introduced and connected with
an anaesthetic bottle. Throughout the remainder of the experiment
the concentration of ether in the inspired air was kept as constant as
possible. The arterial blood pressure, the respiration and the rectal
temperature were carefully watched.

After opening the abdomen the aorta was ligated between the coeliac
axis and renal arteries, and the vena cava tied at the same level. A
cannula was then introduced in the vena cava so that its open end lay
opposite the hepatic veins. This cannula was plugged by a pipe
cleaner. To secure samples of blood from the portal vein a similar
cannula was inserted in the pancreatico-duodenal vein with its free end
just at the vena porta. For the analyses samples of blood were removed,
usually at intervals of two minutes, by removing the pipe cleaners and
connecting a pipette to the cannula. A sufficient amount of blood was
removed to be certain that the cannula was filled with that present in
- the particular vein under observation. This pipette was then removed
and a moistened 2 cc. pipette connected with the cannula. There was
practically never any trouble with clotting, for if a clot did form it re-
mained adherent to the cleaners and was withdrawn with them.

The 2 ce. samples of blood were immediately transferred to test tubes
containing 8 cc. water, and the sugar content was ascertained by one of
the methods described elsewhere by one of us (R. G. P.) (5).

Portions of liver were also removed at the end of the experiments for
glycogen estimation.

Results. The following features of the blood-sugar level have occu-
pied our attention in the present investigation, viz: (1) the extent of the
fluctuations occurring under normal conditions; (2) the initial height of

418 J. J. R. MACLEOD AND R. G. PEARCE

the level in relationship to the amount of glycogen in the liver; (3) the
relationship between the levels in the blood of the portal vein and vena
cava.

Although we have performed a considerable number of experiments,
we do not consider that we have by any means sufficient data from which

Re.

cent

260] Slyeag iveB/ OS fo *
150 ee ea

REA
doe | NaI

Mibu lees fos Exp.

Fig. 1. Curves from seven experiments showing the fluctuations in the sugar
level in the blood of the vena cava.

all of these questions can be finally answered, but we have thought it
advisable to place what we have on record.

The extent of the fluctuations. For most of the experiments in this
connection the modified method of Bang was employed, as described

LEVEL OF BLOOD-SUGAR IN THE DOG 419

by one of us (R. G. P.) elsewhere, and the results of seven of them are
compiled as curves in figure 1. The extent of the experimental error
involved in this method, as indicated here and there along the curves by
vertical oblongs, is seen to be considerable. Usually, however, it is be-
low 5 per cent, and frequently duplicate analyses agreed exactly. After
allowing for the greatest possible error, it is plain that considerable
fluctuations occur, even when the two specimens of blood were removed
at intervals of five minutes apart. It will be noticed that the fluctua-
tions are of two types, one, sudden, occurring over a period of a few
minutes, and the other, much more gradual.

A. The sudden
fluctuations. It is
probable that these
are often due to
experimentalerror,
but this is not al-
ways the case be-
cause they are also
present when the
much more accu-
rate modified Ben-
edict method is
employed. The re-
sults of three ex-
periments of the
same type in which
the estimations
were made by this
method, are com-
piled in the curves of figure 2, from which it will be observed that the
samples of blood were removed every two or three minutes, instead
of every five. This permits us to determine with greater certainty
whether these momentary fluctuations actually exist, which, it will be
seen, they do, although usually very slight in degree. Thus, in the
fifty or so estimations used in the compilation of the curves, the greatest
fluctuations occurring in two minutes are a fall from 0.156 to 0.137 per
cent (No. 78) and a rise from 0.140 to 0.160 per cent (No. 76). Earlier
in this latter experiment two consecutive rises of about 15 per cent each
are also present. The former variation occurred immediately after the
start of the experiment, and is probably to be attributed to operative

Fig. 2. Curves showing behavior of blood-sugar level
compiled from data furnished by a modified Lewis-Bene-
dict method.

420 J. J. R. MACLEOD AND R. G. PEARCE

disturbances involving the portal circulation. We have observed simi-
- lar declines in the blood-sugar during the first few minutes of other
experiments of this type, so that we always allow a sufficient time for
it to occur before attempting to bring about any experimental change.
The sudden rises seen in experiment 76 are more serious and we are
at a loss to explain them. The animal used in the experiment behaved
under anaesthesia in a perfectly normal fashion, and there can be no
doubt about the accuracy of the estimations, since blood samples taken

from the portal vein

Ad page gave values which
Es ' .
310 [nae | agreed very closely
| | Ke | swith those taken simul-
O()
al zi =] =a taneously from the
29 ae | vena cava. Unusual

| though such sudden
fluctuations may be, it
210 is essential that their

; | :
ae j pH possible occurrence in
bl
/

work of this nature
should be allowed for

ee ba Ke by frequent repetition

0 t %

sim mie of the exper iments.

BEY Such sudden rises in
ce we ae! the blood-sugar level

og oF Lyven 975 Pst .
ae aoe
, ul rs during the course of
PP RSS| an peinent on an

all)

on 0 3b 7 D

wee PS be apparently ‘normal’

r
* ; animal should be dis-
Fig. 3. Curve from experiment in which a sudden

Ce ; inguished f imi-
fall in blood pressure occurred following a period of tingul hed ae eee
faulty anaesthesia. lar rises almost invari-

ably to be observed
when the animal becomes asphyxiated, or when there is a pronounced
fall in arterial blood pressure. In figure 3 curves plotted from the re-
sults of such an experiment are given, the animal being hyperglycae-
mic from the very start.

B. The gradual changes. Tn all the eleven experiments, save the one
already referred to (viz., No. 76), there occurred a progressive fall in the
blood-sugar level during the first part of the period of observation.
This fall lasted for at least thirty minutes, although it persisted for the

pey

LEVEL OF BLOOD-SUGAR IN THE DOG 421

whole duration of the observation in about one-half of the experiments
(viz., Nos. 6, 14, 16, 73, 78). The gradual decline is probably due to
the passing away of conditions which excite the glycogenolytic process
during the anaesthetisation and operative manipulation of the animal.

Such evidence that the preparation of the animal causes hyper-
glycogenolysis, along with the marked variation in the level to start with,
might seem to indicate that investigation of the glycogenic function
would be impossible in anaesthetized animals, since the already existent
disturbance would mask any further change, especially an increase,
which might be brought about by experimental interference. It is very
probable that conditions which would cause a slight increase of blood-
sugar in a ‘normal’ animal might fail to do so in one that was hypergly-
caemic, but this fact does not, in our estimation, justify the conclusion
of one author that ‘the dog is little suited for investigations concerning
glycaemia’ (Bang) (6). The emotional glycaemia is at least as marked
in the other animals available for such work (cat and rabbit), and,
in the case of the rabbit at least, there are much more serious objections
to their employment, such as their small size, their herbivorous habits
and the impossibility of removal of the pancreas. Unless dogs are em-
ployed, investigation of many of the problems concerning the glyco-
genic function of the liver becomes impossible, but to discount the dis-
turbing influences due to etherization and operative interference we
must know exactly the extent ‘and frequency of such disturbance.

The initial level and the subsequent rise in blood-sugar in relationship to
the percentage of glycogen in the liver. It is clear from figures 1 and 2
that an initially high level of blood-sugar is more likely to occur in ani-
mals having a high percentage of glycogen than in those in which there
is only a trace. ‘The secondary rise is also usually more marked in gly-
eogen-rich animals, although it sometimes occurs, as in experiment 11,
in those that are glycogen-poor.

The relationship between glycogen-content and the initial sugar-level
is well illustrated in the following table (I) taken from other experiments
than those used in the compilation of the curves in figure 1.

The relationship between the levels in the blood of the portal vein and
vena cava. On account of the magnitude of the blood flow through the
liver, the very smallest difference in sugar concentration in- the blood
entering and leaving it could exist only when an extreme degree of gly-
cogen formation or breakdown existed in the liver. In a subsequent
paper we shall submit results showing how such ‘percentage’ values be-
have when dextrose is injected into the portal vein. In the present

422 J. J. R. MACLEOD AND R. G. PEARCE

paper we have collected, in Table II, the results of estimations which
were made on samples of blood which were collected from the two ves-
sels either simultaneously or at intervals of not more than one minute
apart and prior to the sugar injections. :

TABLE I

Relationship between the initial sugar-level and the glycogen-content of the liver.

STARVED WELL-FED
Number of Glycogen in | Sugar in blood Number of Glycogen in | Sugar in blood
experiment liver at start experiment liver at start
per cent per cent per cent per cent
Sle geek eee trace 0.176 32 6.50 0.246
BOM a aeaee nae trace 0.098 34 0.70 0.163
Bislaihe ser ohh b ce heerekc trace 0.090 35 3.62 0.161
1 RN OP eI ee trace 0.110 37 Did} 0.181
A ee Ai eects cee trace 0.085 40 1.57 0.200
AONE «meer oy Wh trace 0.075 42 10.37 0.202
AS: Ee po trace 0.131 47 0.41 0.170
7 Nace ee trace 0.120 46 9.12 0.118
py tegers ho, ee ye trace 0.118 48 2.00 0.122
Goben ses ee oe trace 0.168 53 12.00 0.170
(\Usctaeiee an eee we trace 0.146 54 4.16 0.205
Gott yt eat oe trace 0.130 62 0.150
68 0.67 0.100
72 6.5 0.200
Max mum 0.176 Maximum 0.246
Minimum 0.075 Minimum 0.100
Average 0.120 Average 0.170
TABLE II
INTERVAL BETWEEN COLLECTION
SUGAR IN BLOOD OF PAN- ry. 4 x oF eS ee
pxpemivess| —REATICODUODENAL | YON ava
ee (1) From same}(2) From other
vein vein
per cent per cent min. min.
OO rare eee 0.170, 0.168 0.168, 0.167 1 1
IO ee eae 0.133, 0.1338 — , 0.146 1 1
ie pee eee tee 0.100, 0.099, 0.100 0.100, 0.098, 0.100 1 1
bs Ne tee Be 0.137, 0.136, 0.126 0.140, 0.140, 0.140 2 0.5
7 Nera «| 0.200, 0.200, 0.207 4 0.5
ft eee. 0.160, 0.148, 0.160 0.160, 0.160, 0.160 2 0.5
fi are stad 0.158, 0.138 0.166, 0.158, 0.158 2 0.5

LEVEL OF BLOOD-SUGAR IN THE DOG 423

A remarkable correspondence is evident between the two bloods in
most cases. When any difference exists it is always small in degree and
is, with one exception (No. 70), due to a decline in the portal blood value
(Nos. 71, 74, and 75). In the case of No. 74 it is possible that the value
0.148 is due to experimental error, but we have no reason to suspect this
in the two other experiments in which the decline occurs (Nos. 71 and
75). In two experiments (Nos. 73 and 76) the estimations were made
over longer periods without injecting dextrose into the portal circulation.
The results of these experiments are given in the curves in figure 4,
from which it will be seen that in one of the experiments (No. 73) there

Fig. 4. Curves showing blood-sugar levels in vena cava and portal vein in
_ two dogs which were only anaesthetized.

was entire correspondence, whilst in the other (No. 76) the sugar con-
centration of the ‘cava’ blood stood very considerably above that of
the portal blood during a period of nearly twenty minutes. This result
indicates a high degree of hyperglycogenolysis for which there was no
apparent cause. The behaviour of the curve for the portal blood
indicates further that the increase of sugar in the blood of the
systemic circulation did not in this experiment immediately cause a
similar increase in that of the portal vein. The excess of sugar is evi-
dently retained somewhere in the organism, probably in the muscles.
In the experiment in which artificial glycaemia was brought about by

424 J. J. R. MACLEOD AND R. G. PEARCE

injecting dextrose into the portal vein, this delay in the appearance of
the excess of dextrose in the portal blood was not noticed (see subse-
- quent communication). We cannot account for it in the present ex-
periment unless we assume that dextrose which has been set free in the
liver as a result of hyperglycogenolysis is more liable to be ‘fixed’ or used
in the tissues than injected dextrose. Although there is so far no direct
evidence (7) that glycogen-dextrose is dealt with by the tissues in any
different manner from chemically pure dextrose, yet it is possible that
it is so, and that one of the differences is with regard to its readiness to
become deposited in the tissues.

In conclusion we would point out that although the blood-sugar level
in an anaesthetised animal does not remain unchanged from time to
time, yet the changes occurring during a period of ten minutes are very
small when compared with those observed by us on blood similarly
removed from animals in which the splanchnic or hepatic nerves were
stimulated (8).

REFERENCES

(1) Scorr: This Journal, 1914, xxxiv, 271.

(2) Swarrer: Journ. Biol. Chem., 1914, xix, 297.

(3) SHAFFER AND Hupparp: Proc. Soc. Biol. Chem., 1914, iii, 31.
(4) Macteop: This Journal, 1907, xix, 388; 1908, xxii, 373.

(5) Pearce: Journ. Biol. Chem., 1915 (in publisher’s hands).

(6) Bana: Der Blutzucker. Wiesbaden, 1913, p. 31.

(7) Macterop: Journ. Biol. Chem., 1913, xv, 497.

(8) Macireop AnD Prarce: This Journal, 1911, xxviii, 403.

STUDIES IN EXPERIMENTAL GLYCOSURIA

X. THe SuGAR RETAINING POWER OF THE LIVER IN RELATIONSHIP
TO THE AMOUNT OF GLYCOGEN ALREADY PRESENT IN
THE ORGAN

J. J. R. MACLEOD anp R. G. PEARCE

From the Department of Physiology and Biochemistry, Western Reserve Medical
School, Cleveland, Ohio

Received for publication August 21, 1915

Since the time of Claude Bernard it has been usual to consider the
glycogenic function of the liver as one analogous to that of starch for-
mation in plants, that is, as a means for temporarily storing away for
future uses in the organism, carbohydrate that is not immediately re-
quired. It has been recognized, however, that the capacity of the
liver to store glycogen is quite inadequate to account for all the carbo-
hydrate that is assimilated from the intestine, and it is believed that
» some of the excess is carried to the muscles and deposited there as glyco-
gen, the remainder being either oxidized or more slowly converted into
fat. According to the above view, we should expect that the power
of the liver to retain sugar as glycogen would be more marked when
this organ is empty than when it is already well filled with glycogen.
Undoubtedly, such a view is correct when we consider the total glyco-
gen capacity; that is to say, the empty liver will hold more than the
- liver that is already partly filled. But this does not necessarily mean
that the rate with which the liver converts the sugar to glycogen will
be different in the two cases.

Capacity for glycogen and rate of glycogen-formation in the liver
need not run parallel. Indeed there are several observations on record
which would seem: to show that dextrose, absorbed from the intestine,
passes more freely into the blood of the systemic circulation in fasting,
than in well-fed animals. Thus, Bang (1) and his collaborators have
found that after giving from 2 to 10 grams of dextrose by stomach to
rabbits, the dextrose concentration of the systemic blood began to rise
in fifteen minutes, gaining a maximum in about an hour and then re-
turning to the normal level in three hours. This hyperglycaemia de-

425

426 J. J. R. MACLEOD AND R. G. PEARCE

veloped more rapidly and reached a higher level in starved rabbits than
in those that had received previous administrations of dextrose; and it
usually failed to develop at all in animals that had received and re-
covered from several previous injections.

Such a result can most easily be interpreted by assuming that the liver
more quickly removes the absorbed dextrose when it already contains
some glycogen than when it is glycogen-free. If these results can be
confirmed, it would indicate that the process which is going on in the
liver during starvation is one of glycogenolysis, or more correctly of
glyconeogenesis, and that the liver cell cannot immediately reverse it
to one of glycogen-formation (glycogenesis) when excess of dextrose
appears in the portal blood. The liver cells, in other words, cannot
produce glycogen and dextrose at the same time, either when the dex-
trose comes from their own stores of glycogen or from protein, and it
takes some time for the cells to change over from the one function to the
other.

The present investigation was undertaken to supply more direct evi-
dence for or against the above hypothesis.

Methods: Dogs were either starved or fed for some days on a diet of
bread and meat, or given cane sugar by stomach tube on the evening
preceding the experiment. After being anaesthetized with ether, can-
nulae were placed in the inferior vena cava and, in the later experiments,
in the duodenal vein, as described in a previous paper. A cannula
filled with Ringer’s solution was also placed in one of the branches of
the mesenteric vein, as far down as possible. Through the tubing which
closed this cannula passed a hypodermic needle connected through a
three-way stop-cock with tubes leading to the bottom of two graduated
cylinders, one containing Locke’s solution, the other, a 4 per cent solu-
tion of pure dextrose. The graduates were closed above with stoppers,
through a second hole in which they were connected with an 8-litre
bottle, in which an air pressure was established by means of a pump.
The large capacity of the bottle and the comparatively slow rate of
outflow through the hypodermic needle ensured a practically uniform
rate of injection during five minutes, and this rate was altered in the
different experiments, either by choosing needles of different bore or
by changing the air pressure—by means of a mercury valve—in the
bottle. Although somewhat hypotonic, the above strength of dextrose
was chosen to simplify calculation, since we intended, at the outset,
to alter the rate of injection by known degrees during each experiment.
The rate of injection varied from 8 to 27.5 ce. in the different experi-

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER 427

ments, i.e., from 0.32 to 1.2 grams of dextrose in five minutes. Taking
the average blood-flow through the portal vein as 1300 ce. in five min-
utes (calculated on the basis of 4.3 cc. per second for a liver of average
weight, 400 grams) (2), and assuming that the blood to start with con-
tained 0.15 per cent dextrose, the above amounts of injected dextrose
would give a maximal percentage of 0.25 in the portal blood, suppos-
ing none of the dextrose was removed from the blood during the five
minutes of injection. This degree of hyperglycaemia in the blood of
the portal vein is believed to be not infrequently overstepped during
the absorption of carbohydrate, so that the conditions in our experiments
in this regard can be considered as well within the normal limits.

In order to ascertain the normal sugar concentration of the blood of
each animal, several samples of blood (2 cc. each) were removed at two-
minute intervals from the vena cava, and also—in the later experiments
—from the duodenal vein. Meanwhile, in most of the experiments,
Locke’s solution was being injected into the mesenteric vein. The sugar
solution was then injected for exactly five minutes, samples of blood
being still collected every two or three minutes, and at similar inter-
vals after discontinuing the injection. A second injection period fol-
lowed after from ten to twenty minutes, unless the condition of the
animal was such as to make it inadvisable. 2

At the end of the experiment a portion of liver was removed and cut
in thin slices, which were then dried between filter paper. Duplicate
portions, of 10 grams each, were then used to determine the glycogen-
content. In some of the earlier experiments a piece of liver was re-
moved, after mass ligation, early in the experiment. It was hoped
that results could be secured, from a comparison of the glycogen-con-
tent at the beginning and the end of the experiments, that would fur-
nish information as to whether glycogenesis or glycogenolysis was tak-
ing place during the time of the observation. So far, the results in this
respect have not been satisfactory. Anaesthesia, arterial blood pres-
sure, and the respirations were practically unchanged during the experi-
ments, unless otherwise noted in the tables of results, which follow.

Discussion of results: These are compiled in Tables I and II. It
will be most convenient to consider them as a whole first of all, and later
to select one or two typical experiments for the purpose of examining
more closely the time relationships and the magnitude of the changes
produced by the sugar injections.

The figures in the fourth columns of the tables represent the grams
of dextrose injected into a mesenteric vein in five minutes; those of the

MACLEOD AND R. G. PEARCE

URE

Jie

, 428

‘UOTZBUITISA VUOo
*£IMDAODOI peyaegl
pr Bae ai
“dd {49Ip [etoods
qua siup ¢ pot
“UOTPBUITISI JUQo
‘DA 2 °C ‘d#
‘uorTyooful puoves
ped ‘“uorjoofur
puooes 199ye IIT4
yovq oulod JOU plo

“porp oq
“youq au09 ou pid *

A][eIyzed youq VUIB oe

‘UOTJVUITYSA OU #
“Toye Apqw

-pisuoo = posvoi0uy x

“A]jrvd Yyouqg IWIV >
“DUIMO]IO}
BIMIVvVdIA[SALIA A TTX
“yoeq vu100 you pig *K
“DUISTL A[TMO]Qo

SMUV Wo

eI

papy pwar ay? fo poo7g ay) fo 7wWa}UWOI-as0.yxap ay} WO sbop paasm7s fo uraa yo7Lod ay7 opur UOYyIalur asoujxap [oO J9a[Ja AY],

dOVIL

d0BL

OBL,
dOVL],

dBA,
dBRL,
dOBL,

OVAL,
ORAL,
OBL,

BLY,

quad Jad

a

6100
080°0

PIL 0
060° 0

920° 0
890° 0
9F0'0

660 0

P80 0

ONIUOG
WOWIXVW
aNv
ONIGUIAUd
GTOVUAAV
NUM LAE
GON AMAT ATG

II

Go 0 |§€0'0
8c1'0
¥ 8S '0 [0090 STL 0

CF9L'O |0ZZ°0 |F80°0
C69T'0 |0€2°0 |420'°0
CAS6I'0 |088 0 |€60°0
Cc9t'0 | 80 T \299°0
C8sST 0 |c20'G |80I 0
Vé0L 0 |6ZT 0 |620 0
CISTI 0 |9cF' 0 |2S0'0
COL 0 \68T' 0 |€20°0
va 0 61 0
C106 0 080° 0
dcsl 0 |’S0' 0 |S00°0
V8SI 0 |€9L'0 |8S0 0
CcErl 0 |949'0 |TE0 0
C2cl 0 \9FZ' 0 |600°0
do0€ 0) 96° T |620°0
quad dad
NOLL
-OULNI (V) |LHDIGM| ASOUL
ONIMOTION| -AGO -xada
ANGLVIG | CHL AG] “WY T
-AWWL UO aatid ud
(dq) DONIWOAG] -1n7OW | asvayo
ASVAMONI ‘od -NI
IVWIXVW
OL 6 8

vs sh OC

«00'0 | {09T°0 | OFT'0 | 09'0
ouo Ny 9FT'O | o9FT'0 | 8F'0
x 9F0 0 Fle 0 89T'D | OF 0
*2460'0 | oGSFl 0 SILO | c&0
4€0°0 c9L'O Sel 0 | SFO
090° 0 OST’ O OGIO | #90
£600 ccT 0 0&1 0 | FF 0
920° 0 LST 0 TEI 0 | ¥2°0
*,610'0 60° 0 G10 0 | ¥2 0
1c0' 0 FOL 0 €80' 0 | OF 0
G10 0 160° 0 G80 0 | 2G 0
oxO0IL 0 FIG’ 0 | ¥F0L 0 | O80
«690 0 revedh (0) NTO | 08 0
*900 0 GIL O 9010 | 06° T
GPO 0 GEL O 060°0 | G20
[€0'0 |} oZIL 0 980°0 | 00 T
*800'0 | ¥*90T°0 860° 0 | 88 0

r »

620°0 | ¥*&82'0 | o9ZT°0 | 00'T
quad Jad quad dad qua9 sad “ub
NOILOULNI | NOLLOALNI | aaLoar

ASV GUONI ONIUod auodae -NI
GDVUGAV | GsouLxad | asoOuLxad | asouL
-doo1mg -doo 1d -xaa

J

I WTA VL

9

eral

a

¥peT |08'°9
PPAIBIS OE 6G
P9AIBIS |0G'¢
poaryg |eg’s
PIAIVIG |OC’ 6
PAAIBI_ 0G’ 6
poealeyg |00' 6I
PeAtIByS |00° 6I
poareyg |9Q' OL
peareyg |JT'S8

POAIBIS JTS

PoAIvyg
POAIBIS

PpoAreyg
PoAareyg
Ppoareyg

DAS
S
—_

P9sArByg

poarvyg | @ LT

qdoost LTHOIaM

éIl
OL

CIP
If

BYE
8é
BOE
9€

I€

“ON

429

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER

“HOTIOLUTL OY} LOFJE SOPLULUL OAT, UTYIIM [BULLOU OY} OF WANJOL YOU PIP AVBNS Jo oBBJUddAed OY} YRY} OFOIPUT SySLIOYSV OY T,y

*MOTPBUITYSE OU) ¥
*poP MUO TOTYBUIT}SO OUD,» |

“1078
Buisvodout youq ouwuy*
*qno snxe [d oredoyy oy
‘A ‘d

ul astl jar[easd Ayjoexoy # | G19°0
*youq oulod you pic] #

‘ATO 831898 Jo

“BULY JO UOISNJUl Sulla}, y Ol'p
“youq ourod Jou pic[ # 00°21
*pepnypout JON
*‘]ensn uy} JOPEy, yByS],
“BIUOBOA [D10d
-Aqy Jooyjo oVBIpoUrUyT
‘s}[NSeL AB[NBIT AlOAS 00°%
G16
“RItUOBOA[DIOdGAY AT[BOT# | ZIP'O
*AIODAO001 ON # 1°01
"SSOT [[1GS posoaooey | LG" T
» *UOLJBUITISA [ o
“A[IYSLYS posroaoooas A[UGO # Loy
‘MIND aasy
*podojoaop Al 1pwe4s
BIULOVOA[DIOdAY poxieyy # Lg
(aU
*youq euros you pic] #*
“TB[NGOIT SI[NSOYo | GIL°0
G9
UUALT
NI NGO
fMUV NG -OOXTD
INGO
ud
ll él

120°0
200°0

10
020°0
810°0
000
C100
020°0
280°0
s10°0

OFL 0

sel 0
8700
££0°0
000° 0
000°0
£20°0
6S0°0
al 0

090°0
£110
_ 00°0
60° 0

620°0
quad sad

NGGAL
-2a& GoONaudssida

ONINOG WOWKIXVHK
QNY SONIGGOGNd

SOVUGAY

a

L610
(c61'0
Vo6e 0

022 0
CsIt'0

260
Cg9t'0
CU9Te’ 0
C1696 0

V88L0
CLO86 0

(092° 0
Ce9t 0
CIst 0
CiIPT 0
COLL 0
(19220
Ss 0
VbPe 0

(10920
Vos 0
Voor 0
g61°0

“MOTION
MIALVIGUN
-WI Uo (a)

OnTUOnG
ASVANONT

IVWIXVIN

Or

GCE 0
026 °0
G9'T

L¥'0
0L6°0

0820
GOT 0
GEh 0
9¢2°0

610
OST’?

OPO T
0g8°0
£860

Loe 0
ops 0
£92°0

819°0

PEL 0

080
£060
916°0

LHOTUIM
-AGOd
HL AG
aattd
-W1OW
‘od

6

2£0°0
T£0°0
1e'T

SLEo0'0

610°0
osusyo ON
0Z1 0
9600
GE0°0
9¢0°0
asva10e(T
e100

£60

990°0
8L0°0
920°0
osvaloo(]
osvol0a(]
080°0
0¢0'0
SOLO

680°0

9¢0°0
ouoN

960°0
910°0
Z10°0
quad wad

WsouLrxkad
WVuUS [ Ud
ASV GUONT

8

S100
«9100
£90°0

810°0

6100
osusyo ON
»« ¥€20°0

«800°0
1£0°0
«60° 0
OSBO100(T
« ¥900°0

4160°0

L20°0
8£0°0
610°0
OsvelooCT
# OUON
+ ¥*£20°0
« ¥ 8600
44#80°0

« ¥9¢0'0

« #®9E0°0
ouON
¥180°0

810°0
»610°0
qua? wad

G@SVGNONT
GOVUAAV

Z

T610
9210

o608°0
#8100
#ZIL'0

0960
Zt2'0
6ol'0
112'0
£920
L1Z'0
9210

18o°0

ob PI 0

£o1°0
L810

TPL 07

OLT 0
926 °0
ope 0
9180

Gos °0

116 0
2910

# POL 0

$860
£920

yuad wad

NOILOULNI
ONTHOG

asouLrxaa

aqoo1d

9

#OLT'O
CT
920

006°0

O00
96°0
69T 0
STO
9PS 0
G02'0
G30
OLL0

OFT 0

6060
o68S 0

0060
Ist 0
19t°0
ot91 0

9960
96 0

quao wad

NOILOULNI
GY Ost
asouLxaa

qoo'la

‘

G

PO | pRar1g L4°8 BPEL
GEO | pBReag L°8 bl
8h 0 | pRorgy 9 ZI BSL
Sh O | pBorg, 931 GL
P90 | pRorgy € FT 89
Ph 0
09°0
060 | PBetEe q’9 a)
96°0
88°0 | pvorEe Q'sI #¢
OF 0
SPO | pRote 676
OF 0 ivsng ST BRP
SPO | Aweng 81 8F
SPO | 1esng 6°01 BOP
610 | Awsng 6°01 oF
89°0 | 4BdNg OL BLE
08'0 | resng Or LP
#90 | AveNg 6° OL BCP
080 | avsng 6°01 a
080 | aedng LoL BOP
#90 | tesng 13°2 OF
P90 | wesng | TST le
080 | Weng | gor ce .
06 T IVINS L Gt ine
80°T IVINS | GE
ULDLb “BY

aa Lour
goer aqood |LHDIGM!] ‘ON
-xuda
b £ é I

pana puaa ay) fo poojg ay) fo juajuoa asonjrap ayy uo sbop paf-jjam fo uraa po)Lod ayy o7ur suonoalur asoaxap fo Joaff'a ay. J,

IL WTAVL

430 J. J. R. MACLEOD AND R. G. PEARCE

fifth, the percentage of dextrose in the blood of the vena cava immedi-
ately preceding the sugar injection. In practically every instance these
- figures, i.e., in column 5, are the average of determinations on several
samples of blood removed at two-minute intervals. The sixth column
gives the average percentage of blood-sugar during the injection of
the sugar, and the seventh, the average increase. This average increase
is nearly always considerably less than the maximal increase, which
is given in the tenth columns, because the values in the samples of
blood obtained immediately after the injection did not, in most cases,
show any increase over the normal. On this account, and also be-
cause the first samples taken after discontinuing the injection still
showed the increase in sugar, it might have been advisable to use,
for the computation of the average, not the values obtained during the
injection, but rather those immediately preceding and following the end
of the injection period. This method was not chosen because it was
found in many cases that the increased discharge of sugar was main-
tained for some considerable time after discontinuing the injection,
thus indicating that a hyperglycogenolysis had become established,
probably as a result of the sugar injections. This stimulation of a
more or less persistent sugar mobilization will be discussed ina future
paper; meanwhile it is important to note that it was more frequent
in the case of glycogen-laden livers than in those that were glycogen-
free, so that, to have used the figures obtained after the injection, would
have made the averages for fed animals too high. The cases in which
the sugar percentage did not return to the normal within about five
minutes after the injection are indicated by asterisks.

Since figures in the seventh columns represent increases produced
by the injection of varying amounts of dextrose into the portal vein,
they cannot be compared with one another. To make this possible
it was necessary to calculate in each case the average increase for 1
gram of injected dextrose. This has been done by direct proportion—
figures in column 8—on the assumption that the same fraction of in-
jected sugar will always be retained by the liver, an assumption for
which there is however no experimental justification, for it may well
be that, with progressively increasing injections, a greater or a smaller
fraction of the sugar is retained. We intend to investigate this pomt
at an early date.

To further reduce the figures to a common basis allowance had of
course to be made for the greater quantity of blood in the portal cireu-
lation in large, as compared with small animals. Since the total volume

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER 431

of blood in the body is practically proportional to the body weight, we
have made this correction by multiplying the figures in the eighth
column by those in the second, since the larger the animal, the greater
. must be the dilution to which the injected sugar would be subjected.
The values thus obtained are more or less arbitrary, but they are com-
parable one with another. The figures in the twelfth column give the
percentage amounts of glycogen found present in portions of liver re-
moved immediately after the animal was killed, or, in some cases, in
portions removed before the injection was started.

Although the values obtained after making the corrections as above
indicated—i.e., those in the ninth column—do not exhibit as close a
correspondence as we had hoped they would, yet, when we compare
them in Tables I and II, they seem sufficient to justify the general con-
clusion that the glycogen-free liver does not remove dextrose from the
portal blood any more quickly than one that is glycogen-rich. The
ultimate storage capacity of the glycogen-free liver is no doubt greater
than that of one that is already partly filled with glycogen, but its avidity
for dextrose is certainly not more pronounced; indeed, if anything, it
appears to be somewhat less, as a close comparison of the results in the
two tables will indicate.

It will be noted that the percentage of dextrose in the blood issuing
from the liver at the start of the experiment is very much more con-
stant in the case of the starved animals than in those that were fed.
The variation for the former is between 0.075 and 0.146 (leaving out the
values in experiments 31 and 65, which are considerably higher) and the
average 0.111. The variation for the latter is between 0.100 and 0.206
(leaving out experiment 32) and the average 0.150. This is of course
what we should expect, and is entirely in line with our previous experi-
‘ence that operative hyperglycaemia is much more likely to develop in
well-fed, as compared with starved animals. The differences, however,
render comparison somewhat uncertain, and they make it desirable
to compare the extent of the hyperglycaemia produced by 1 gram of
dextrose in animals having about the same initial blood-sugar percent-
age. This is done in the following table (Table III).

It will be seen that the dextrose tends to pass the liver more readily
in the starved group of animals.
The initial hyperglycaemia probably does not introduce any very
serious error, for the increase produced by injecting 1 gram of sugar js
at least as marked in initially hyperglycaemic, as in normal animals.
Thus, for sugar percentages up to 0.16, the increases were: 0.330. 0.00,

432 J. J. R. MACLEOD AND R. G. PEARCE

TABLE III

STARVED ANIMALS WELL-FED ANIMALS

Increase in percentage
of dextrose perl gram .
dextrose injected mul-
tiplied by the body-
weight

Increase in percentage
Original percentage of of dextrose perl gram | Original percentage of
dextrose in blood dextrose multiplied dextrose in blood
by the body weight

0.098 0.246 0.100 0.270

0.090 0.630

0.118 0.720 0.118 0.280

0.140 0.220 0.150 0.162
0.160 0.270
0.161 0.000
0.163 0.330

0.168 . 0.600 0.170 0.000
0.169 0.120

0.000, 0.283, 0.850, 0.162, 0.27, 0.270; and in those having an initial
blood-sugar percentage of 0.200 or above, they were: 0.216, 0.203, 0.618,
0.763, 0.545, 0.327, 0.000.

The uncertain nature of the results of the above experiments led us
to suspect that the differences obtained might be due to unequal amounts
of sugar entering the liver, for although we always examined to see that
the sugar solution was really going into the mesenteric vein, yet it might
not become properly mixed with the blood by the time that thehilus
of the liver was reached. ‘To control this we performed several experi-
ments in which blood was taken from the portal vein, through a cannula
inserted in the central end of the pancreatico-duodenal branch, at
about the same time as that collected from the vena cava.

The results of these experiments are given in Table IV, in the sixth,
seventh and eighth columns of which the figures standing opposite D
represent the percentages of sugar in the blood of the portal vein, and
those opposite V, the percentages in that of the vena cava. That is
to say, the figures occupying the position of numerators represent
the values of the portal blood; the denominators, those of the blood
of the vena cava. The numbers are arranged in each column in the
order in which the samples of blood were withdrawn from either vein.
\ It will be seen that, almost without exception, the percentage of
sugar in the inflowing and that in the outflowing blood of the liver are
equaél before any injection of sugar is made (see previous paper). In

colump 7 the change produced by the injection of sugar is recorded.
\

433

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER

“WUr Qf OINSserd poorg x
“AIOVV PLYOICD,,

96F°0 ,68F°0 .o2F'0 C680 ,Sée"0 oge"0 29270 A ' bie
= a rg B
“uyUr ¢°0 ‘ura g | #z¢6°0 Zee FeO | 58" OFe0 w20 weod| % 9 i a eae Deal
L81°0 | S¢1°0 .8S1°0 ,99T'0 A
“UI G9 “UU % (WOT}By1OSNsar SUuIY4BoIq peddoys) ozo} Selo ssl od 980 a7 POFTIPAL g a
8810 ,88I'0 2610 L610 ,981'0 POLLO eAal| Pe eres
“UTM ¢°9 ‘urn | SFO O20 T6F'0 Ta ga oroa| %°? ee ple I
8810 ,98I°0 ,egt'0 G61" .091°0 | o9r'o ooo oor 0 A] _. be
“urUr ¢°9 ‘amg | ZO BLO 290 I8L°0 0610] o9t'0 stro osroa| S22 5 Ea | La8oale ve
eae 80.7" ,268°0 LLE°0 0920 ,002°0 £5Z'0 .6F2'0 A
G-] Suoroofur GEE O (G06 446 0 .0G ‘ ; OF6 0 A : bre erie -
“‘uyur g'9 | oroyoq “ur Z| ZeeQ ose"0 G€8°0 8920 6F7 0 oce'0 ocwoa| 9 fee Bes TPA 8 EY es
022°0 ,002°0 2220 ,S1Z"0 ,21z'0 102'0 ,002°0 A ; ;
002'0 212°0 002°0 A an 2
“urUr ¢°Q “urUL g One Ua se 020 0260 Oza'0 (e'0) ooo a} SF 9% peas fl
P8I'0 ,961'0 0220 . (EFT 0) STO A :
eee SEM _——_- = O0BI 19) 3) 3
“urur ¢°Q “urUr g OFI'0 #810 C160 160 Foro |="? He 4] ARSE TEEN | BiB |
“""" ONT'0 {ISTO LLU'0 ,861°0 | OFT0 .OFT'0 .OFT'0 A
ee a Bh —_— _—_ OS SCS 90B1 o) o
“urut ¢°9 ‘arg | eNO STO SFO (982°0) I8t'0 | 9er0 9er0 zero a| °° ih pee | Sa ae 0
6110 SILO ,FIL‘O ae COUTMON IN Poy
A ne A — ——_— G 3 Ajoye1epo ?
earls sap eke aaa aro oe ro sprue; | et On| aunt jexexopoyy | °F | ugg
Z0T'O 8010 .£01'0 | SIT'O .IIT"O ,OIL'O ,IIT'0 | OOT"0 ,860°0 .00T'0 A ; poy
STA CSR [UE GU) —— ‘—— = ; A[o7B.10 po i
“urUr J mimi} — -80r'O 80r°O eIT'0 | giro Tio 90r0 ~~ | coro eo coro dg | 0 | 929° [YPC | SFT | 89
=< 9 gO, 6410 “8 g9T-9 9910 A
6210 ,621°0 ee wuld Ser : . dAIB 2
“uyUr G~9 ‘urmg-g | ero Zsr0 zsl0 9910 691-0 oroa | 80 | 2880 Pome | 8°88 | soz
Pei anna Feo 8S1'0 .9FT'0 ,FeT'0 del renege oat
‘ulut §6¢°Q |e 4sey coy -uTUE ! ‘ ‘ ‘ ‘ A 3 : ehar 5
gnq surat y | gg qnq “ure | FTO OST'0 6sr'0 TO BeneORarsoane| 8F Us )288.0 ea
Gia aorieurth 6610 ,922'0 ,Z12'0 1¥B'0 .S€8'0 ,29T'0 2900 8910 A] a aes '
‘urorT /e-g anq -urery | = ogrg B60 ZIZ0 “= 352-0 66L-0 $910 oro a | oF 0 uL Poet Pee eel 80
2910 .£91'0 ,82T°0 S210 .£91'0 £610 | eT'0 80 ,£T'0e .
GIT VY col 0 SAT 0 OE ra SY a ee a0vd DAIBY
“uyUr J ‘um tT | gO FEO S210 8.10 010 S10] sro ero eroa| 80 sh paar PW PL
“16 quad sad “By
Se a a re oe a ee Loe
SNIQA “NIW G
INDUCE NIGA GNVS WOlQdofur SutMo][o iT wolqoofat Satan uorqoefur o10j;agZ NI NHATT
woud WOU SaTda - aqggoar NI aooa “TM “ON
SQ1TdWwyvs -WVS NAGM ia -NI Napoox 1p
NOUGMALAE IVAUGTINI SS Le ee asoun
IVAUMALNI SATdNVS COO1s NI qsouLxaa INGO uma -xada

Or. 6

a L a > _

434 J. J. R. MACLEOD AND R. G. PEARCE

In the two-minute interval that was allowed to elapse between the be-
ginning of the sugar injection and the collection of the next sample of
blood, a detectable increase in sugar concentration had occurred in
the portal blood (i.e., numerators) in nearly every instance. In the
case of the cava blood (i.e., denominators), however, the increase in
sugar concentration in the two-minute interval was not nearly so fre-
quent (e.g., it was absent in Nos. 65, 70, 70a, 71, 71 a, 72 a and 74).
This delay in the appearance of the injected sugar in the ‘cava’ blood
is very evident in the curves which will be referred to immediately.
It seems to occur as frequently in well-fed as in starved animals.

The delay is no doubt partly due to the time occupied by the blood
in passing through the liver vessels, but this can not account for all of
it, as the following rough calculations will show. In a dog weighing
12 kg., which is about the average of those employed in the experiments,
and the liver weight 400 grams, there would be from 1300-1500 ce. of
blood passing the organ in five minutes, say 550 cc. in two minutes.
Assuming that the liver contains 30 per cent of the total blood in the
30 X 800

100
before the injection started would have been displaced at least twice
during the first two minutes of the injection. Part of the delay must,
therefore, depend on diffusion of the excess of dextrose into the lymph
and tissue juices of the liver, a diffusion which presumably goes on
until blood and tissue juices contain the same concentration of dextrose.

When we compare the sugar concentrations in the samples of blood
removed from the two vessels during the last three minutes of the in-
jection periods, they are seen to be practically identical in all of the
experiments, except 71 and 72 a. From the neighboring values we are
inclined to attribute the result in No. 71 to experimental error. That
in No. 72a is undoubtedly due to the progressive hyperglycogenolysis
present in this case.

The figures in column 8 furnish us with information regarding the
after-effects of the sugar injection, the first samples being usually taken
two minutes after. discontinuing the injection, and the others at sub-
sequent two-minute intervals. Taking the change in the cava-blood
first, it will be seen that in twelve of the experiments (Nos. 63, 65, 70,
70 a, 68, 68 a, 71, 72, 72 a, 74, 74 a, 75) there was no return to the
normal level within four or five minutes after discontinuing the injec-
tion, such a return being evident only in one experiment (No. 71 a).

Similar comparison of the portal blood reveals a more rapid return

body (i.e., = 240 cc.), it will be seen that the blood in the organ

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER 435

to the normal sugar level in five or six of the experiments (Nos. 63, 65,
70?, 71, 71 a, 74 a), but in the remainder the increase in sugar concen-
tration remained as high as in the cava-blood. These results indicate
that there can have been no retention of sugar by the muscles of the
body within the time of the observation, so that the blood entering the
portal radicles in the intestine contained a concentration of dextrose
which was almost equal to that leaving the liver. The dilution due to
admixture of this blood with blood from the upper extremities and head
was not sufficient to make an appreciable difference. On the other hand,
in the experiments in which the sugar level in the portal blood fell more
rapidly than that of the cava-blood, it must be concluded that muscle
retention of dextrose had occurred. We hope to publish further obser-
vations on this question in a subsequent communication. It is of in-
terest to compare this result with that noticed when the increase of
sugar in the cava-blood is entirely due to hepatic hyperglycogenolysis
(see previous paper).

Returning to our main question, as to whether any difference exists
between glycogen-poor and glycogen-rich livers in their avidity for
dextrose, we may say that there is no evidence, as judged from the
difference between the sugar concentrations of portal and cava-blood,
of any difference.

In order to give a clearer presentation of the results than is possible
in a table, the curves in figure 1 have been plotted from three of the
experiments (LXX, LXXI and LXXIV). The dotted line represents
the percentage of sugar in the portal blood, and the discontinuous line,
that in the portal vein. The following points should be noted:

1. The quicker rise in sugar concentration in the porta! blood than
in that of the vena cava.

2. The delay, after discontinuing the injection, in the return to the
normal level of the sugar concentration of the blood of the vena cava
and frequently also in that of the portal vein.

CONCLUSIONS

The existing glycogen-content of the liver does not demonstrably
influence the rate with which this organ removes dextrose from the
blood of the portal vein.

The authors wish to acknowledge the valuable assistance of Messrs.
M. and H. Mahrer in the preparation of this paper.

436 J. J. R. MACLEOD AND R. G. PEARCE

Exp. LXX. Starved animal (dextrose injection between arrows).

a
HE a eT ge a EOE
ee ce i

i ii “| CAE aR Neary oe aa
is a | nial a ac pee i
ee a i EUG al oo. ce a a

=
eps
bate
Sees
SESE
t-—
eee

eH
ia

Per cent. Dextrose

isa Nag GE ce ae Ha eite
cee a 7 Ht Ga
14 gree Magee TE
Shae foe BREE
133 Hi! Hatte SEU: FEREE EEO ESSES RE
1 Wey 3) ae i i a ate ia cy i be ai an oe 16 17 ‘16 19
Minutes

22 Exp. LXXI. Fed animal, but ees Peron eee pen

hobs ee ee dee i Hd _ it
eae ii ie etc ee et
HUI LEE E BH He
fo) eli eles et oe a H
GTA HiT Ret Sa

22 Le
£ AL
2 aa
OH
S Fa ie AE
© U7 i a
® 16 i
ri TG Ee
ee Te
14th soo TEE HEHTHE iit itt iT
Bhar aneeyaeenamnNGIN an
13) Hf H{! Hits * Hh staltssnetts SU isiesipseetucsteesesieestiegge tases: Stet rstestsEs Ha
aces 2 3 (Pe) 6 “rd 8 9. 10, JT 22 33 24 25 Ie aya
® Exp. LXXIV. Fed animal, with over 4 per cent. glycoyen.
8 2 ee
2 Ti a il a ce a Le
i | |
( HH He tf le? i rt
; ae a ee - ee | | a
a Cee a cs He oo i
My BHTH Ht i : H Ei
ef ee a ee i
ee 3 3 ‘& 5 9 270 ll Fa 13 14 25 2690

Fig. 1. Curves showing effect of dextrose injections into the portal vein on
the sugar content of the blood of the portal vein and vena cava.

INFLUENCE OF GLYCOGEN ON SUGAR STORAGE OF LIVER 437

REFERENCES
.

(1) Banc: Der Blutzucker, Wiesbaden, 1913; Jacopson:
1913, lvi, 471; Bo#: ibid, 1913, lviii, 106; Wet: :
und Phar., 1913, Ixxiii, 159.

(2) MactEop anp PrEarce: This Journal, 1914, xxxv, 87.

Biochem. Zeitsch.,
Arch. f. exper. Path.

THE DIFFERENTIAL EFFECTS OF ADRENIN ON
SPLANCHNIC AND PERIPHERAL ARTERIES

FRANK A. HARTMAN
From the Laboratory of Physiology in the Harvard Medical School

Received for publication August 24, 1915

The opposed action on arterial blood pressure, in the dog and cat, of
different amounts of adrenin injected intravenously—a rise, or a rise
and fall, after large doses; and a pure fall after small doses—present
phenomena of considerable theoretical interest. This is especially true
in the relation of adrenin to the sympathetic system, and in view of
present rather meager evidence of vasodilator influences effected
through that system. In studies previously made in this laboratory!
it was proved that the same amount of adrenin might have opposed
effects under different circumstances. Thus the depressor influence of
a given small dose when normal arterial pressure prevails is changed to a
pressor influence if the arterial pressure has been carried below a certain
level, as, for example, by pithing. These opposed effects indicate that
adrenin may operate on antagonistic processes, and, as an alternative
to the idea of dilator and constrictor sympathetic actions, the tentative
suggestion was offered that adrenin causes relaxation of blood vessels
when they are tonically contracted, and contraction when they are re-
laxed. Another natural opposition in the vascular system which has
been claimed by various observers? is one existing between the vessels
of the splanchnic area and those of the periphery. The present inves-
tigation was directed towards an answer to the question whether adrenin
has differential effects on these two vascular regions.

The cats used in these experiments were anesthetized with urethane
(2.0 g. per kilo, by stomach). The form of adrenin employed was fresh
adrenalin diluted with distilled water just before each experiment.
Unless otherwise stated the concentration was 1: 100,000. Injections
were made into the jugular vein from a syringe graduated to fiftieths of
a cubic centimeter. In most cases injections of 0.2 ec. were made, at a

1 Cannon and Lyman: This Journal, 1913, xxxi, 385.
2 Dastre and Morat: Systéme Nerveux Vaso-moteur, Paris, 1884, 329.
4388

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 439

uniform rate in each instance, over a period of from fifteen to thirty
seconds.

Blood pressure was registered by means of a mercury manometer,
connected usually with one of the carotid arteries. In a few cases the
pressure changes in the nasal mucosa were recorded by a membrane
manometer connected with one of the anterior nares, the other being
plugged with vaselined cotton and the posterior nares being closed with
Mendenhall’s apparatus.’ .

The vasodilator effect of small doses of dilute adrenin is, in our experi-
ence, a constant phenomenon rather than ‘‘a somewhat variable effect,”
as Dale’ describes it. Of the fifty-three animals used in this series of
experiments only five failed to show a fall of blood pressure when in-
jected with 0.2 cc. adrenalin, 1: 100,000. As will be shown later, four
of these were in an abnormal condition which would account for the
failure. Of the forty-nine normal cats, therefore, in only one did the
small standard dose of adrenalin fail to cause a drop of arterial pressure.

The response of peripheral arteries. In order to determine the re-
sponse of the peripheral arteries to this small dose of adrenalin, the cir-
culation was excluded from the splanchnic region by tying the inferior

-and superior mesenteric arteries and the coeliac axis, and sometimes
also the renal arteries. Of twenty-three animals with splanchnic ves-
sels thus tied, fourteen still showed the characteristic fall of arterial
pressure after injection of the standard dose. In four animals in which
there had been a preliminary rise of pressure followed by a fall the
same dose failed to cause the rise after the splanchnic vessels had been
excluded. In two, showing only a rise previous to the tying, the adren-
alin produced only a fall after the tying. And in another case a fall of
20 mm. (14 per cent) before splanchnic exclusion was increased to 38
mm. (32 per cent®) afterwards.

These results clearly indicate that a small dose of adrenalin (0.2 cc.,
1: 100,000) causes relaxation of the peripheral arteries. In no case did
constriction result. That in some instances the fall was not increased
after splanchnic exclusion was probably due to lessened tone in the pe-
ripheral arteries as a result of the operation. With such lessened tone
the adrenalin would probably not have so great an influence as in con-
ditions of greater tone. This explanation is supported by the obser-

3 Mendenhall: This Journal, 1914, xxxvi, 59.

4 Dale: Journ. Physiol., 1913, xlvi, 291.

5 The blood pressure which had been 143 mm. before tying the splanchnic ves-
sels was lowered to 118 mm. by the operation.

440 FRANK A. HARTMAN

vation that immediately after the operation the standard dose of ad-
renalin evoked no response, but as time elapsed it brought forth the
characteristic dilation.

The response of splanchnic arteries. To find the response of splanchnic
arteries the abdominal aorta above the iliacs, both subclavians and both
carotids were tied. There remained for the circulation the vessels sup-
plying the trunk and the thoracic viscera as well as those of the splanch-

TABLE I

Effect of excluding the limb and neck arteries upon the response to 0.2 cc. adrenalin,
1:100,000. Pressures are in mm. of mercury. The same dose of adrenalin was
given in each case.

BEFORE TYING AFTER TYING PERIPHERAL ARTERIES
Adrenalin Per cent Adrenalin A
Cat Pressure caused a fall of tall Pressure | caused ariseof| Per cent rise

mm mm. mm mm
I 104 23 22.1 150 2 13
72 7 9.7
II (54.ct Wwe 13.6 157 11(fall)| 7.0 (fall)
104 6(rise)| 5.7 (rise)
III 135 9 6.7 131 10 7.6
IV 115 30 26.0 81 12 14.8
V 113 17 15.0 99 11 1148
VI 110 20 1g. 55 14 25..4
VII 124 28 22.5 75 30 40.0
VIII 112 32 98.5 131 8 6.1
80 26 93.5
EX 109 12 11.0 140 12 8.5
54 108 16 14.8 44 17 38.6
Dat 112 24 21.4 60 8 13.3
x 106 20 18.8 108 15 13.8
LET 94 32 34.0 107 23 21.4
XIV 132 36 27 .2 95 4 4.2
XV 138 18 13.0 140 16 11.4
XVI 90 26 28.8 48 7 14.5

nic region. Adrenin is without effect on the pulmonary arteries;® it

typically causes relaxation of the coronaries,’ but since the coronary

vessels are not capacious they need not be considered. So far as the

blood supply to the trunk is concerned, it is probably of much smaller
6 Brodie and Dixon: Journ. Physiol., 1904, xxx, 488.

7 Langendorff: Zentralbl. f. Physiol., 1907, xxi, 555; Cow: Journ. Physiol.,
1911, xlii, 132.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 44]

volume than that of the large and extensive abdominal organs.® It is
justifiable, therefore, to regard the animal operated upon as described
above as carrying on mainly a splanchnic circulation, and to consider
the influence of adrenin as exerted chiefly upon the splanchnic vessels.
The limb and neck arteries were tied in twenty animals. The opera-
tion invariably resulted in a marked increase in arterial pressure.

In all but two of the twenty cases the standard dose of adrenalin
(0.2 ee., 1: 100,000), injected after the limb and neck arteries were tied,

TABLE II

Effect of excluding alternately the peripheral and splanchnic arteries upon the re-
sponse to 0.2 cc. adrenalin, 1:100,000. Blood pressure is expressed in milli-
meters of mercury.. The Roman numerals show the order of clamping and
unclamping the arteries.

PERIPHERAL SPLANCHNIC ARTERIES

SAS gue ARTERIES CLAMPED WEEE CLAMPED
ro rs) 3
= Per 2 Per Adren- a a Per
One| Jsore | ae | Genk | sure | SE | Some | ‘sure | 2024| cent | sure | 2% | Se
23 23 Ba
ao om ao
< < <
mm. mm. mm. mm. mm. mm. mm. mm.
E II III
Bae ar | 21.7 | 7 32 | 48.8] 83 19 | 22.8} 110 10 9.1
fall | fall
Ill II I
Pee AZ |} 15.3) 73 5 6.8} 68 5 7.3 | No elffect
rise | rise
Ill II I
C}] 112) 32 | 28.5 | 132 8 6.0} 118 17 | 14.4] 128 1 tab
80 26 | 32.5 fall | fall
if II Ill
D| 1385 18 13.3 | 182 9 6.8 113 36 31.8

caused a rise of arterial pressure—an effect just the opposite of that
caused by the same dose before the tying. Often when the pressure
had been greatly increased by the tying, the standard dose at first pro-
duced a fall of pressure, but later, as the pressure lessened, the fall was
changed to arise. Moreover, as the pressure decreased below normal
the percentage rise after the standard dose usually increased.

In Table I the figures show the change from a fall to a rise when the
peripheral arteries are tied.

8 Ranke: Die Blutvertheilung, Leipzig, 1871, 69.

449 FRANK A. HARTMAN

Splanchnic and peripheral effects in the same animal. In four animals
both splanchnic and peripheral responses were obtained after adrenalin
injection by clamping first one group of arteries and then the other
after the first had been released. The splanchnic arteries were clamped
first in two cases, and the peripheral arteries were clamped first in the
other two. The results are shown in Table II. In each case the dose
of adrenalin was 0.2 cc., 1: 100,000.

In every case clamping of the peripheral arteries, whether primary or
secondary to splanchnic exclusion, caused the vascular response to the
standard dose of adrenalin to change from the normal relaxation (fall)
to contraction (rise) (see fig. 1). And in every case except one (B),

Fig. 1. Effect of injecting 0.2 cc. adrenalin, 1:100,000: a. Upon the whole cir-
culation; b. Upon the arteries of the splanchnic area (peripherals clamped);
c. Upon the arteries of the limbs (splanchnic arteries clamped). Time interval
five seconds. Middle line, pressure base line (0 mm).

either primary or secondary clamping of the splanchnic arteries contin-
ued the normal drop in pressure resulting from the standard dose of
adrenalin, though the drop in some instances was not so great as before.
In D, however, the fall of 13 per cent under normal conditions was in-
creased to 31 per cent after the splanchnics were clamped. The failure
of effect in B may have been due to a too-great relaxation of the vessels
as a result of the operation In A and C after the release of the
peripheral and the splanchnic vessels adrenalin produced the usual
fall. The rise in B may be explained as a splanchnic effect, in the pres-
ence of relaxed peripheral vessels (note the low pressure, 68 mm.).
Cases in which dilute adrenalin caused a rise in blood pressure. Occa-

sionally there are animals in which no fall in arterial pressure results

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 443

from doses of adrenin which normally cause a fall, nor can a fall be ob-
tained by diminishing the dose. As already stated, five cases out of
fifty-three were found in which no fall occurred after the standard dose.
It will be recalled that all but one of these animals were in poor physical

condition.

creased in weight from 2.9 to 2.4 kilos.
of mercury; 0.2 cc. adrenalin, 1: 100,000, injected
at the same rate as in other experiments resulted in
17-18 mm. rise in blood pressure. Even 0.2 cc.
adrenalin 1: 1,500,000, caused in some tests a slight
rise. In no test was the slightest fall in pressure
obtained.

In a second case the animal was pregnant and
appeared poorly nourished, for its ribs were very
prominent. The blood pressure was about 95 mm.
of mercury, and nothing but a rise could be in-
duced, even with as little as 0.02 cc. adrenalin,
1:100,000. The arteries of the splanchnic area were
tied, whereupon five or six times as much adrenalin
was required to cause a rise equivalent to that ob-
tained before the occlusion of the splanchnic arter-
ies (sufficient time was allowed to elapse for the
recovery of vasomotor tone). Apparently the pri-
mary rise had been due largely to contraction of
the splanchnic area and the remainder of the ar-
terial system was not responsive, in the direction
of contraction, to the standard minute dose.

A third animal, whose blood pressure was but
67 mm. of mercury, was diseased; in this animal
0.2 ec. adrenalin, 1: 100,000, caused a rise of 7mm.
(fig. 2), but after the splanchnic arteries were tied the
same dose caused a fall of only 3 to 6 mm.

One animal had been suffering from urethritis for about
three weeks and had eaten insufficiently during that time.

It had de-

The blood pressure was 58 mm.

Fig. 2. Effect of
0.2 ce. adrenalin,
1:100,000 upon a
cat that was ill: a.
Upon whole circula-
tion; b. Upon pe-
ripheral arteries
(splanchnics
clamped). Nasal
plethysmograph
line just above pres-
sure base line (0
mm). Time line at
bottom, interval
half minutes. A fall
in plethysmograph
curve indicates con-
striction.

This experiment, there-

fore, indicates that the rise, with very dilute adrenin, in animals in
weakened condition, is due to constriction of the splanchnic arteries and

also to lessened dilation of the peripheral vessels.

A fourth animal which had been given alcohol, failed to show a defi-

nite fall with dilute adrenalin.

No explanation was to be had for the failure of the depressor effect in

the fifth animal.

444 FRANK A. HARTMAN |

In three of the five, however, there was clear evidence of a low blood
pressure, a condition that harmonized with the atonic state of the
animals. The indication that toneless peripheral vessels are not made
to relax further by injections of adrenin has already been mentioned.
It seems reasonable to assume, therefore, that when animals are in
poor physical condition, a small dose of adrenin fails to cause relaxa-
tion of the vessels that normally relax for such a dose, because the
tone of these vessels is already low.

Depressor action of adrenin as affected by low.arterial pressure. If low
arterial pressure was the occasion for failure of adrenin to produce a °
fall of arterial pressure in the foregoing cases, it becomes a matter of
interest to learn whether an induced low pressure will affect the action

TABLE III

Effect of lowering the arterial pressure, by hemorrhage, and then raising it, by restor-
ing the lost blood, upon the response to the standard dose of adrenalin

E IN PRE 5
BLOOD PRESSURE IN MM. CHANGE IN PRESSURE IN

OF MERCURY mE Teich ae PER CENT CHANGE
mm. mm.

102 14 fall 13.7 fall
83 from hemorrhage 6 fall 7.2 fall
71 from hemorrhage 5 rise 7.0 rise
65 from hemorrhage 8 rise 12.3 rise
77 from injection 4 rise 5.1 rise
87 from injection 4rise 4.6 rise
94 from injection 4rise’° 2 fall 2.1 fall
99 from injection 5 rise 65 fall 5.0 fall

109 from injection 2rise 12 fall 11.0 fall

of the substance. Low pressure was established in two different ways,
by hemorrhage, and by depressor stimulation.

(a) Low pressure due to hemorrhage. In seven experiments there
was one in which the primary fall of arterial pressure due to injected
adrenalin was changed to a rise as the pressure was reduced by hemor-
rhage from 102 mm. to 65mm. The drawn blood was defibrinated and
returned to the circulation by degrees. As the pressure rose in conse-
quence of the restored blood the depressor action of adrenalin grad-
ually reappeared.

Table III shows in this case the change in response to 0.2 ec. adre-
nalin, 1: 100,000, as the pressure was lowered and then raised again by
injection of defibrinated blood.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 445

In the remaining six experiments there was merely a smaller fall,
following the standard dose of adrenalin, as the blood pressure was
lowered by bleeding, until at about 40-50 mm. there was either no
effect or a slight rise and fall. That no considerable rise occurred may
have been due to a temporary abolition of the splanchnic response as a
result of hemorrhage, for very little time elapsed between the bleeding
and the injection.

The effect of the standard dose of adrenalin was next studied in ani-
mals with splanchnic circulation excluded and with blood pressure low-
ered by hemorrhage. In the five cases studied nothing but a fall in
pressure occurred even when the pressure had been lowered to 40 or 50
mm. (fig. 3). At 20 to 30 mm. the dose had no effect. The figures in

TABLE IV

Response to the standard dose of adrenalin in an animal with the splanchnic arteries
tied, when the pressure is lowered by hemorrhage

ae | Faroe 20,3 PeR CENT ALL
mm. mm.
161 12 fall 7.4
135 12 fall 8.8
129 18 fall 13.9
110 13 fall 11.8
97 15 fall 15.4
84 13 fall 15.4
71 18 fall 25.3
50 10 fall 20.0
34 6 fall 17.6

Table IV present a good example of the response to adrenalin (0.2 cc.,
1: 100,000) when in an animal with splanchnic circulation excluded, the
blood pressure is decreased by hemorrhage.

Judging from the experiments described earlier in this paper one
would expect, in an animal with the peripheral arteries tied, that the
rise in pressure, in response to adrenin would continue as the pressure
was lowered by bleeding. This was found to be the case in the only
animal in which it was tried. The rise with 0.2 cc. adrenalin persisted
with the pressure as low as 10 mm. (fig. 4). At 70 mm. pressure the
dose caused 8 mm. or 11.4 per cent rise. At 29 mm. the same dose
caused 8 mm. or 27.5 per cent rise. At 10 mm. the same dose caused
3 mm. or 30 per cent rise.

446 FRANK A. HARTMAN

Fig. 5.

Fig. 3. Effect of 0.2 ec. adrenalin, 1: 100,000 after hemorrhage, with arteries
of splanchnic area tied: a. Before bleeding; b. 65 ce. of blood removed; c. 35 ce.
additional blood removed. Time half minutes. Lower line pressure base line
(0 mm).

Fig. 4. Effect of 0.2 ec. adrenalin, 1: 100,000 after hemorrhage with peripheral
arteries tied: a. Before tying peripheral arteries; b. After tying peripheral arter-
ies; c. 35 ec. of blood removed; d. 28 ec. additional blood removed. Time inter-
val half minutes. Base line zero pressure.

Fig. 5. Effect of 0.2 cc. adrenalin, 1: 100,000 when blood pressure is lowered
by stimulation of the depressor nerve: a. Without depressor stimulation; b. Pres-
sure lowered by depressor stimulation, showing rise with adrenalin; ec. Pressure
lowered by depressor stimulation, showing fall with adrenalin (different animal).
Adrenalin injected at ad. wx shows effect of same dose of adrenalin with normal
pressure. Time half minutes.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 447

(b) Low pressure due to depressor stimulation. When blood pressure
is lowered by stimulation of the depressor nerve, 0.2 cc., adrenalin,
1: 100,000, may cause, as shown by Cannon and Lyman,’ a fall, a rise
and fall or a pure rise depending to a certain extent upon the height
of the pressure when injected (fig 5). In cases where a rise occurs, it
disappears if the splanchnics are tied. Seven animals were used in

TABLE V

Effect of stimulation of the depressor nerve upon the response to the standard dose of
adrenalin (0.2 cc., 1:100,000). Pressures are expressed in mm. of mercury.

NORMAL AFTER DEPRESSOR STIMULATION
Ba: Splanchnics tied
Per be =I Per
Animal es pet eres gas pe peet hie ae Pressure from Effect of
re sfepreseor | adrenalin
eal =]
mm. mm. mm. mm. mm.
Jii,o-Lise
se ts fall| 12.0} 65| 9rise| 13.8 74 | Noeffect
B 115 | 30 fall | 26.0 85 | 14 rise | 16.4 114 14 fall
(158 mm.
after tying
splanch-
nics)
C 129) 10 fall Sl 67 19 rise | 28.3 80 No effect
D 120 | 19 fall | 15.8 93 1l5rise| 5.3
tall es
E 89 | 28 fall | 31.4 43 5rise | 11.6
F 126 | 12 fall] 9.5 90 13 fall | 14.4
(90 {| 2 rise [ 2.2
13 fall || 14.4
G 102 | 15 fall | 73 | 4rise 5.4 42 No effect
| 4fall || 5.4
\49 || Srise}| 6.1

studying the effect of dilute adrenalin when pressure was lowered by
depressor stimulation. In every case both vagi were cut. Standard
doses of adrenalin (0.2 cc., 1: 100,000) were invariably injected. Table
V shows the results of these experiments.

As the results in Table V indicate, exclusion of the splanchnic area
- causes adrenin to have no pressor effect when the general blood pressure

9 Cannon and Lyman: Loc cit., 387.

448 FRANK A. HARTMAN

has been lowered by depressor stimulation to a point where the standard
dose would naturally produce a rise.

In the foregoing experiments the evidence obtained by simple tying
out of the splanchnic region or of the neck and limb vessels has been
supported. If arterial pressure has been lowered either by hemorrhage
or by depressor stimulation the standard dose of adrenin, instead of
producing a fall of pressure, produces a rise or has no effect. If now the
splanchnic area is excluded adrenin fails to have any pressor influence
and may indeed drop the pressure still further. If, on the contrary, the
peripheral arteries are tied, the same dose causes an elevation of the
lowered pressure, even when 10 mm. measures the arterial tension.

Latent period and duration of the adrenin effects. Since the splanchnic
and peripheral arteries respond in opposite directions to minute doses of
adrenin, and yet the end result, when both portions of the circulation
are affected, is the characteristic fall of pressure which is seen when the
peripheral vessels alone are involved, the time relations of the response
in the two vascular regions becomes a question of considerable interest.
Tests were made on eight animals, and by records written on a rapidly
moving drum the latent period and the duration of the adrenin effect
were carefully estimated. The results are presented in Table VI.

In the cases of a pure drop in pressure the latent period varied from
12.5 to 22 seconds, with an average of 15.7 seconds. The average
latent period for the splanchnic rise (after tying the peripheral arteries)
was 18.6 seconds. It seemed probable, therefore that with a minute
dose of adrenin, peripheral dilation precedes splanchnic constriction and
consequently masks it. There is a possibility, however, that the oper-
ation of tying or clamping off part of the circulatory system tends to
lengthen the latent period—a suggestion that receives support from an
observation that a fall of pressure which occurred in 12.6 seconds before
operation did not occur after splanchnic exclusion until 26 seconds
had elapsed.

If the foregoing suggestion is correct, the discrepancy between the
latent periods of the peripheral relaxation and the splanchnic contrac-
tion would not be so great as the figures in Table VI indicate. Indeed
in two of the cases there was first a rise of blood pressure and then a
fall. With somewhat larger doses of adrenin than those here given
this is a usual result. Since the standard dose causes both a splanchnic
contraction and a peripheral relaxation of arteries, the common occur-
rence of a pure fall of pressure, when both these parts of the circulatory
system are affected, indicates that this small amount of adrenin usually
affects first the peripheral vessels.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 449

The duration of the fall in pressure after the standard dose (0.2 ce.
adrenalin, 1: 100,000) averaged in five animals 59 seconds. The aver-

age duration of the splanchnic rise in five animals, however, was only ©

37.8 seconds. In no single case did the splanchnic rise last more than
three-fourths as long as the normal fall. With the amount of adrenin
used in these experiments, therefore, the dilation of the peripheral
arteries begins earlier and lasts longer than the constriction of the
splanchnic arteries. The result is that the splanchnic rise is masked
by the peripheral fall of pressure.

TABLE VI

Latent period and duration of adrenin effects. In each case 0.2 cc. adrenalin, 1:100,-
000, was injected. Records were made on a rapidly moving drum

LATENT PERIOD

A y
Ss WEIGHT - TENE BEBTOD DURATION OF oF peel DURATION OF
ANIMAL IN KILOGRAMS OF THE FALL SPLANCHNIC RISE
cs re NORMAL FALL : RISE (PERIPH-

ERALS TIED)

sec sec sec sec

7 i pa 13 70 17 42
Bao... 2.0 12.6 53 1lf/ 5 38
Ce ae sss: 3.2 15 46 18 31
|b) 2 2.5 PGs 66 23 41
Ds be Qe ali (rise 8)

fall 15 (52) 18 37
LR cts eee PAV 20 ;
(Co cadena 3.0 (rise 14)

fall 22 (48)
i 3.2 16 59
AVOTAPE:....:.. 15.7 59 18.6 37.8

Effect of dilute adrenin on the blood vessels in the nasal mucosa. In
view of the opposite effects of dilute adrenin upon the splanchnic and
peripheral arteries, the question arose as to whether the dilation of the
peripheral vessels extended to the skin or was limited to the more deeply
lying vessels. No convenient method was found to determine the ac-
tion of vessels in the skin areas over the general body surface, but the
vascular changes in the nasal mucosa were easily studied. A mem-
brane manometer was connected to one of the nasal openings in the
manner described earlier in this paper. The nasal plethysmograph
record was made simultaneously with the tracing of general blood
pressure.

450 FRANK A. HARTMAN

Doses of adrenalin (0.2 cc., 1: 100,000) which caused dilation of the
peripheral arteries always caused constriction of the nasal mucosa
(fig. 2). Amounts as small as 0.05 ec. of adrenalin, 1: 100,000, caused
constriction. Martin and Mendenhall'® produced constriction of the
nasal mucosa by injecting 0.5 cc. adrenalin, 1: 1,000,000. The reaction
of the nasal mucosa to adrenin therefore is of the splanchnic type rather
than the peripheral type.

The threshold dose of adrenin changing a fall of blood pressure to a rise.
In determining the threshold dose it is necessary to consider the rate"!
of injection as well as the concentration. And since the sensitiveness
of tissues to adrenin decreases if the injections are too frequent or too
concentrated,” care must be exercised to avoid these sources of error.

ATT TTT TTT TUTTI

UTA TTT MTT MMMM TTT
a b c d

Fig. 6. Threshold of the change from a fall to a rise with adrenalin when
splanchnic arteries are tied: a. 0.1 cc. adrenalin, 1: 10,000, fall only; b. 0.25 ce.
adrenalin, 1: 10,000, fall and rise; c. 0.3 cc. adrenalin, 1: 10,000, slight fall, large
rise; d. 0.35 ec. adrenalin, 1: 10,000, rise only. Time seconds.

The rate of injection was kept nearly constant, being uniform and
over a period of from ten to twenty seconds, For convenience two con-
centrations of adrenalin solution were kept ready, viz., 1: 100,000 and
1: 10,000. To prevent dilution in the process of changing the solu-
tions one solution was injected into the external jugular vein while the
other was injected into the femoral vein. In order to avoid the lowering
of the sensitiveness of the blood vessels to adrenalin the smallest possi-
ble number of doses were injected.

As the evidence already presented has shown that adrenin causes con-
striction of the arteries of the splanchnic area and relaxation of those of
the periphery, the question of the threshold dose of adrenin, changing

'0 Unpublished work done in this laboratory.

11 Cannon and Lyman: Loc cit., 382.

12 Mlliott: Journ. Physiol., 1905, xxxii, 443.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 451

a fall to a rise, is concerned solely with the effect on the peripheral

vessels.

The threshold was determined for three animals, the splanchnic

arteries being tied off in each case.

The threshold for two of these was

between 0.2 cc. and 0.3 cc. of 1: 10,000 adrenalin while in the third it

was about 0.1 cc. of 1: 10,000 adrenalin.

threshold of the change of a fall to a rise apparently
varies with the individual. The experiments were
too few in number to permit definite conclusions
but they indicate the magnitude of the threshold.
Reversal of the adrenin effect after ergotoxine.
Dale’s® discovery, that after ergotoxine a dose of
adrenin that would normally cause an increase of
arterial pressure causes a decrease, is of consider-
able interest in connection with the reversed effects
of adrenin on peripheral and splanchnic arteries.
Since the alteration induced by ergotoxine is from
contraction to relaxation it was necessary to make
the test on animals with the peripheral arteries tied.
Under these circumstances does ergotoxine cause
the usual increase of pressure resulting from the
standard dose of adrenin to change to a decrease?
It was impossible to obtain the “‘reversal effects”
after the injection of small doses of ergotoxine
phosphate (e.g., 0.56 mgm.). Indeed these small
doses seemed to render the splanchnic response more
sensitive, for the rise of pressure after adrenalin in-
jection was thus increased (see fig. 7). Reversal
or inhibition of the splanchnic constriction was ob-
tained only after several larger doses (5 to 10

The third animal weighed
but 2.0 kilos while the others weighed 3.7 and 3.4 kilos.
hold was just passed a preliminary fall preceded the rise (fig. 6).

As the thres-
The

Fig. 7. Effect of
small doses of ergo-
toxine phosphate in
increasing response
of splanchnic arter-
ies to adrenalin (pe-
ripheral arteries
tied): a. Effect of
0.2 cc. adrenalin,
1:100,000, before
giving ergotoxine;
b. Effect of same
dose of adrenalin af-
ter giving two doses
of ergotoxine 0.5
mgm. and 0.6 mgm.,
respectively. Time
half minutes.

mgm.) of ergotoxine, such as Dale used, had been administered (see

fig. 8).

Thus after such large doses, adrenalin (0.4 cc., 1: 10,000)

caused, in an animal with limb and neck arteries tied, a fall of arterial

pressure from 86 mm. to 65 mm.

Recovery required six minutes.

When 1 ce. of this solution was injected there was a similar fall, which

was recovered from after eleven minutes.

13 Dale: Journ. Physiol., 1905, xxxii, lix.

452 FRANK A. HARTMAN

Discussion of the opposed action of splanchnic and peripheral arteries
in response to dilute adrenin. Cannon and Lyman" proposed the idea
that adrenin causes relaxation of the blood vessels when they are toni-
cally shortened,—contraction when they are relaxed. The evidence in
this research indicates that the opposite action of dilute adrenin depends
rather upon opposite effects produced in the splanchnic and periperal
arteries. The rise of pressure when the vessels are relaxed would thus
be explicable on the ground that the standard dose of adrenin cannot
relax further the peripheral vessels already relaxed but can still have
its constrictor influence on the splanchnic area. In certain conditions,
however, the state of tonicity may play a part, e.g., when the blood

oly

a

Fig. 8. Reversal of splanchnic arterial response to adrenalin after giving large
doses of ergotoxine phosphate: a. Effect of 0.2 ce. adrenalin 1: 100,000 before tying
arteries; b. Effect of 0.2 ce. adrenalin 1: 100,000 after tying peripheral arteries;
c. Effect of 0.2 cc. adrenalin 1: 100,000 after 21.5 mgm. ergotoxine had been given.
d. Effect of 0.4 cc. adrenalin, 1:1,000 after a total of 27.5 mgm. ergotoxine had
been given.

pressure has been rendered unusually high by clamping the peripheral
arteries, for dilute adrenin then often causes dilation. But even in this
case, the dilation might be due to the arteries supplying the trunk, as it
is impossible to tie them off.

Although there is no convincing proof that there are vasodilator
nerves in the sympathetic system, their presence in the peripheral blood
vessels and their absence from, or relatively slight development in, those
of the splanchnic area, would offer the most plausible explanation of the
differential effects of dilute adrenin. An alternative explanation would

144 Cannon and Lyman: Loc cit., 398.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 453

be that the arterial muscle differed in the two regions and that adrenin
produced its effect by direct stimulation of the plain muscle. The first
seems impossible, while the second is contradicted by evidence of Dixon
and Brodie that adrenin produces its effect through the sympathetic
nerve ending.

If the ‘‘sympathetic vasodilator” hypothesis is accepted it is neces-
sary to assume that the vasodilators are more sensitive to adrenin than
are the vasoconstrictors, in order to account for the change from a fall
to a rise in the response of the peripheral arteries to increasing doses of
the drug. The rise which occurs as the vasoconstrictor threshold is
passed is preceded by a fall (see fig. 6). This would be expected from
adrenin slowly injected because the first amounts reaching the nerve
endings might be insufficient to stimulate the less sensitive constrictor
endings yet strong enough to affect the vasodilator endings and then as
the amount of adrenin increased the vasoconstrictors would be brought
into action and overwhelm the vasodilator effect.

The idea that vasodilator and vasoconstrictor nerves can be brought
into action in turn by different strengths of stimulation receives sup-
port from the observation that weak sensory stimulation may cause a
lowering'® of blood pressure while strong sensory stimulation (20 to
200 times stronger than the stimulation which produces a depressor
effect) usually produces a rise.!7_ Bowditch and Warren!’ found that a
slow rate of stimulation with an induced current caused vasodilation
while a more rapid rate caused vasoconstriction. Vasodilators may
have been stimulated in the former case and vasoconstrictors in the
latter case.

Ostroumoff!® was one of the earliest to believe that the sympathetic
contained vasodilator nerves. He was supported in this belief by
- Puelma and Luchsinger.?® Dastre and Morat* thought that they had
evidence of the existence of vasodilators in the sympathetic. Dale” is
inclined to accept the theory of the existence of an admixture of vaso-
dilators and vasoconstrictors in the sympathetic and, as he points out,

16 Dixon and Brodie: Loc cit., 494.

16 Knoll: Sitzungs. a. Akad. d. Wissensch. zu Wien, Math.-Naturwiss. Klasse,
1885, xcii, Abtheilung, 3, 449.

17 Martin and Lacey: This Journal, 1914, xxxiii, 222.

18 Bowditch and Warren: Journ. Physiol., 1886, vii, 447.

19 Ostroumoff: Pfliiger’s Arch., 1876, xii, 219.

20 Puelma and Luchsinger: Pfliiger’s Arch., 1878, xvili, 489.

21 Dastre and Morat: Loc cit., 247.

22 Dale: Loc cit., 299.

454 FRANK A. HARTMAN

this would account for the fact that no reversal can be obtained in the
rabbit, vasodilators in the sympathetic being theoretically absent.

The utility of the simultaneous peripheral dilation and splanchnic con-
striction. It may be that the amount of adrenin, which is poured into
the blood stream during times of stress or excitement, is at first of the
order of that used in the foregoing éxperiments. Hoskins and McClure”*
estimated that the amount of adrenin secreted as a result of splanchnic
stimulation was at first of this order. Ellott?4found that the sensitive-
ness of the arrectores pilorum to adrenin varied with the functional use by
each animal. Ina similar mannerit is possible that response of the blood
vessels to adrenin is in accordance with their functional use in times of
excitement, there being an active dilation of the peripheral arteries and
simultaneously a constriction of the arteries of the splanchnic region.
Such an arrangement would assure the motor organs an abundant blood
supply for their most efficient action.

SUMMARY

1. Dilute adrenalin slowly injected caused a fall in general blood
pressure in 48 out of 53 animals tried. Three of the animals in which a
fall did not occur, were in poor physical condition and one was just
recovering from the effects of alcohol.

2. Dilute adrenalin caused dilation of the peripheral arteries even
after extremely low pressures had been produced by hemorrhage. The
same dose of adrenalin caused constriction of the splanchnic arteries.

3. When the blood pressure was lowered by depressor stimulation the
same dose of adrenalin caused a fall, a rise and fall or a pure rise de-
pending somewhat upon the height of the pressure.

4. The average latent period for the peripheral fall in the blood pres-
sure (from doses of 0.2 cc. adrenalin, 1: 100,000) was 15.7 seconds; the
latent period for the splanchnic rise was 18.6 seconds. The duration of
the splanchnic rise was 37.8 seconds, while the duration of the peripheral
fall was 59 seconds.

5. The threshold for the change of a fall to a rise in the peripheral
arteries, when adrenalin was injected (in three animals only) over a
period of from ten to twenty seconds, was between 0.1 and 0.3 ee. of a
1: 10,000 solution.

23 Hoskins and McClure: Arch., Internal Med., 1912, x, 352.
24 Elliott: Loc cit., 416.

EFFECT OF ADRENIN ON SPLANCHNIC AND PERIPHERAL ARTERIES 455

6. Large doses of ergotoxine phosphate inhibit the splanchnic re-
sponse to dilute adrenalin. The existence of sympathetic vasodilator
nerves in the peripheral arteries and their absence in the splanchnic
arteries would account for the opposed action of like doses of dilute
adrenalin upon the peripheral and splanchnic arteries.

I wish to thank Dr. W. B. Cannon for suggesting this research and for
his advice and criticism.

THE OSMOTIC PROPERTIES OF CALCIUM AND MAG-
NESIUM PHOSPHATE IN RELATION TO THOSE
OF LIVING CELLS

EDWARD B. MEIGS
From the Wistar Institute of Anatomy and Biology

INTRODUCTION

In 1867 Traube suggested that living cells might owe their semi-
permeable properties, at least in part, to precipitates of inorganic sub-
stances.! In recent times this suggestion has been nearly lost sight of
and the attention of physiologists has been chiefly centered on the ques-
tion whether or not the semi-permeable surface of the cell is composed
of lipoid.

Most of the work on this question has been directed toward discover-
ing to what substances the cell surface is or is not permeable, it being
tacitly assumed that we know or can predict a priori the semi-perme-
able properties of ‘‘lipoids.”” But the little work that has been done
with artificial lipoid membranes is far from justifying this assumption.

It has been shown, for instance, by Nathansohn? and Ruhland* that

‘ artificial lipoid membranes, when soaked with water are quite perme-
able to substances which are soluble in water, whether or not the sub-
stances in question are soluble in lipoids. On the other hand, water-
free lipoid membranes are impermeable to water as well as to dissolved
substances.

It was shown by Pfeffer that semi-permeable membranes can be
made of calcium phosphate, and it is well known that both calcium and
inorganic phosphates are present in most living cells. Since Pfeffer’s
time there has accumulated a considerable body of evidence which shows
that calcium plays a very important part in the activities of living tis-
sues, and particularly that the permeability of the cell surfaces is af-
fected by its presence or absence.

1 Traube: Arch. f. Anat. Physiol., und wissensch. Med.; 1867, xxxiv, 146.

2 Nathansohn: Jahrb. f. wissensch. Bot., 1904, xxxix, 607.

3 Ruhland: Jahrb. f. wissensch. Bot., 1909, xlvi, 1.

4 Pfeffer: Osmotische Untersuchungen, Leipzig, 1877, 11.

456

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 457

Thus Loeb’ finds that CaCl, can protect the fish, Fundulus, against
the toxic effects of KCl. He gives reasons for thinking that the calcium
protects by forming an insoluble compound with some ‘‘organic anion”’
at the surface of contact between the fish and the solution (Loc. cit., pp.
320 and 321).

Somewhat more direct in their bearing on this question perhaps are
certain experiments of Osterhout on plants. This author gives a series
of experiments which indicate that the surface of spirogyra filaments is
decidedly more permeable to NaCl than to CaCls, and that the addition
of CaCl, to an NaCl solution may prevent the latter salt from entering.*
In another series of experiments he shows that the resistance of the leaf
cells of the common kelp (laminaria) to the passage of an electric cur-
rent may be increased by the action of CaCl, and gives reasons for think-
ing that the salt acts by decreasing the permeability of the plasma
membrane.’

Both Loeb and Osterhout find that the effects of the CaCl; can be to a
greater or less extent imitated by substituting for it certain other salts
such as SrCl,, BaClh, MgCle, and Lag(NO3)¢. The kations of all these
salts form insoluble compounds with the phosphates, and it seems,
therefore, not impossible that they may all tend to decrease the permea-
bility of the cells in the same way, namely by impregnating their sur-
faces with layers of insoluble phosphate.

Both calcium and magnesium are known to be present in most living
cells. The physiological experiments make it appear that calcium is
decidedly the more important of these two, but a consideration of the
ash of certain tissues of which the permeability has been much studied,
points to a different conclusion. Abderhalden, for instance, finds that
the blood corpuscles of a number of the domesticated mammals con-
- tain small and rather variable quantities of Mg, but no Ca. And Katz
has shown that samples of striated muscle taken from the human being,
pig, ox, calf, deer, rabbit, dog, cat, hen, frog, haddock, eel, and pike
contain larger and more constant quantities of Mg than of Ca. The
Mg content of the muscle in this series of animals varies from 0.017 per
cent of the weight of the fresh tissue in the haddock to 0.037 per cent in
the hen; while the Ca content varies from 0.002 per cent in the ox to
0.04 per cent in the pike.?®

5 Loeb: Biochem. Zeitschr., 1911, xxxii, 308.

6 Osterhout: Science, N.S8., 1911, xxxiv, 187.

7 Osterhout: Science, N.S., 1912, xxxv, 112; ibid., xxxvi, 350.

* Abderhalden: Zeitschr. f. physiol. Chem., 1898, xxv, 106.

® Katz: Arch. f. d. gesammt. Physiol., 1896, Ixiii, 1.

458 EDWARD B. MEIGS

An interesting observation which indicates that magnesium phosphate
is present in considerable quantities at the surfaces of the striated muscle
fibers is recorded by Hiirthle.!° He finds that if fresh frog’s muscle
fibers are treated with ammonia, they “cover themselves” with crystals
of magnesium ammonium phosphate. It is difficult to see how such erys-
tals can be formed, unless magnesium phosphate is previously present.

In view of the facts given above, it has seemed to me worth while to
make a study of the osmotic properties of calcium and magnesium phos-
phate. I had hoped to make a more or less extended comparison be-
tween the permeability of these precipitates to various substances and
that of living cells. But.some two years of work on this subject have con-
vinced me that the time is not yet ripe for such a comparison. One
begins by thinking of a semi-permeable membrane as a more or less
simple and stable kind of filter by means of which water can readily be
separated from most substances that can be dissolved in it. But this
view is far from the truth. The studies that have been made on the
copper ferrocyanide membrane, which is at present the best known
semi-permeable membrane, show that its permeability varies greatly
with the physical conditions under which it is formed and to which it is
afterward subjected—with the electrical conditions at the time of its
formation, with the time which elapses between its formation and the
experiment, with temperature, with the presence or absence of elec-
trolytes, and with the nature of the electrolytes which happen to be
present.!! If one wishes, therefore, to determine whether one osmotic
membrane is more or less permeable to a given substance than another,
one must be careful to control all these conditions. To do this in the
case of artificial membranes formed in the laboratory is by no means
easy; it is obviously impossible when one of the membranes to be studied
is that of a living cell. The experiments to be reported, therefore, bear
only in a general way on the question whether the semi-permeable prop-
erties of the phosphate membranes are similar to those of the surfaces
of animal and plant cells. But they do throw some light on the semi-
permeable properties of these precipitates and on certain fundamental
questions regarding the nature of semi-permeable membranes and of
osmotic reactions in general. I am compelled by external circumstances
either to publish this work now or to lay it aside for an indefinite period,
and these considerations impel me to publish it now, though the results
are in many respects fragmentary and incomplete.

10 Hiirthle: Arch. f. d. gesammt. Physiol., 1903, C, 451.

11 Morse: The osmotic pressure of aqueous solutions; Publication No. 198 of
the Carnegie Institution of Washington, 1914, Chapter IV.

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 459

THE CONDITIONS ON WHICH DEPEND THE CRYSTALLIZATION OF CALCIUM
AND MAGNESIUM PHOSPHATE

Morse believes that true semi-permeability is an attribute of colloids
only, and gives many cogent reasons for holding this belief.2. Experiences
of my own, which will be described later, lead me to concur in this be-
lief; and I have spent a considerable amount of time in inquiring under
what circumstances the phosphates of calcium and magnesium fail to
crystallize.

I have confined my inquiry to the orthophosphates and monohydro-
phosphates of the two metals, as the dihydrophosphates can exist only
in acid solutions. Abegg describes the crystallization of both the
orthophosphate [Mg3(PO,)2] and monohydrophosphate [MgHPO,] - of
magnesium and of the monohydrophosphate of calcium [CaHPO,].'8
With regard to Ca3(POx.)2, however, he says,'* ‘‘ This amorphous precip-
itate has never been observed to become crystalline.”

CaHPO, precipitates were prepared by mixing together equimolec-
ular solutions of CaCl, and KzHPOx,; Ca3(PO;)2 precipitates, by mix-
ing solutions of CaCle with solutions of KzHPO, to which KOH had
previously been added. The KOH and K,HPO, mixtures were made
by adding together equal portions of equimolecular KOH and K2_HPO,
solutions.

It was found that under these conditions the CaHPO, crystallized'®
within twenty-four hours, while the Ca3(PO,)2 never crystallized, though
its precipitates were kept sometimes for more than four months.

Precipitates of MgHPO, and Mg3(POx,)2 were prepared in a manner
corresponding to that which has just been described, MgCl: being sub-

stituted for the CaCl. The rapidity with which both these precipi-

12 Morse: Loc. cit., pp. 87 and 209.

13 Abegg: Handuch der anorganischen Chemie, Leipzig, 1905, vol. ii, Abt. 2,
pp. 65 and 151.

14 Thid., p. 152.

15 Concentrations are, throughout this article, given in the terms of what
Morse calls ‘‘weight-normal solutions.’’ See p. 479 of this article and ‘“‘Osmotic
pressure of aqueous solutions,’’ Chapter V.

16 Tt was determined whether or not the various precipitates crystallized by
examining them microscopically. Such examination cannot, of course, show
whether or not the precipitates are in a colloidal state in the sense of Morse, and
capable of showing true semi-permeability. Morse finds, for instance, that
precipitates of zine ferrocyanide may become granular and lose their semi-per-
meable properties without becoming actually crystalline (Osmotic pressure of
aqueous solutions, p. 91). But it may at least be said that crystalline precipi-
tates do not exhibit semi-permeability.

460 EDWARD B. MEIGS

tates crystallized was extremely variable. The MgHPO; sometimes
crystallized in four days, and sometimes failed to crystallize at all,
‘though its precipitates were kept for more than four months. I did not
make any study of the conditions which control the crystallization of
this precipitate; as it is so soluble that it can hardly be supposed to play
a part in modifying the semi-permeable properties of living cells.

With regard to Mg3{PO,)2 it was found that its crystallization de-
pended on a number of conditions, of which the following were more or
less extensively investigated. (1) Amount of alkali present. (2)
Temperature. (3) Presence or absence of Ca. (4) Concentration
of salts in supernatant fluid.

The influence of alkalinity on the crystallization of Mg3(PO,)2 was
investigated by mixing solutions of MgCl and K,HPO, together with
varying quantities of KOH, and noting the time required for crystal-
lization to occur (experiments 1 and 2). These experiments show that
when varying proportions of MgCle, KzHPO,, and KOH are mixed to-
gether, the crystallization of the resulting precipitates depends on the
relative proportions of the three ingredients. When the final mixture
contains a higher molecular concentration of K2.HPO, than of KOH
and enough Mg to combine with all the PO: as Mg;(POu,)s, the resulting
precipitates remain for the most part amorphous, for many weeks at
least. But when the final mixture contains a higher molecular concen-
tration of KOH than of KsHPO, or not enough Mg to combine as
Meg;(PO,)ewith all the PO, present, the resulting precipitate crystallizes
more or less rapidly.

It is probably not correct to speak of the precipitates which have
just been described as Mg3(PO,)2. Such precipitates are probably
mixtures of this compound with varying amounts of MgHPO, and
Mg(OH)s. There are many reasons for thinking that the quantities of
the last two named substances present are small, but no special investi-
gation of this question has been undertaken. For my purpose the
main point of interest is the fact that magnesium phosphate combina-
tions, which have a high degree of insolubility, may, under certain cir-
cumstances, remain amorphous indefinitely, or, at least, for a long time;
for convenience the precipitates will in future be spoken of simply as
magnesium phosphate.

Experiments 3 and 4 show the influence of temperature on the erystal-
lization of magnesium phosphate. When formed above a certain criti-
cal temperature, which is not far from 23° the precipitates fail to erys-
tallize, whereas they crystallize readily at lower temperatures. These

’

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 461

experiments indicate also that precipitates which have been formed at
temperatures above 23° and have*therefore failed to crystallize, show
little or no tendency to crystallize, when later kept at lower tempera-
tures. A comparison of experiment 3 with experiment 4 shows finally
that, other things being equal, the precipitates have a greater tendency
to crystallize when formed from more concentrated solutions. The
precipitate formed from the less concentrated solutions at 20° in experi-
ment 3 was only about half crystalline after 21 hours, while that formed
from the more concentrated solutions in experiment 4 was completely
crystalline at the end of the same period at the same temperature.

Experiment 5 shows that calcium has an inhibiting effect on the
crystallization of magnesium phosphate precipitates.

THE OSMOTIC PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE AND
OF COPPER FERROCYANIDE

The osmotic properties of calcium and magnesium phosphate and of
copper ferrocyanide have been studied by making small, unsupported
membranes of them according to the method of Traube!’, and by pre-
cipitating them on porous clay cups. The electrolytic method of pre-
cipitation of Morse!® has not been employed in these experiments. It
is, of course, highly desirable that the properties of phosphate mem-
branes precipitated by the electrolytic method should be studied, but
it would be out of the question to follow the technique of Morse in the
rough preliminary survey which constitutes the subject matter of this
article.

EXPERIMENTS WITH PHOSPHATE AND FERROCYANIDE MEMBRANES
PRECIPITATED ON POROUS CLAY CUPS

Experimental methods

It has been my plan to determine as far as possible for what sub-
stances the phosphate membranes are impermeable and for what sub-
stances they are permeable, and to compare the rapidities with which
substances of the latter class diffused through the membranes. To do
this, it is not necessary to measure the osmotic pressure of the solutions
used. The substance, of which the osmotic properties are to be studied,

17 Traube: Arch. f. Anat., Physiol., und wissensch. Med., 1867, xxxiv, 123, 133,
et seq.
18 Morse: Loe. cit., 83-84.

462 EDWARD B. MEIGS

may be precipitated on the inner surface of a porous cup, which may
then be provided with suitable attachments for determining how rap-
idly fluid passes through it to the interior against a slight hydrostatic
pressure. The cup is then filled with a solution of known composition
and concentration and immersed for a given period in distilled water or
in another known solution. At the end of the experimental period it
is determined how much fluid has passed through the walls of the cup
and the membrane; and the fluids within and without the cup are
analyzed to determine how much, if any, of the solute experimented
with has passed through the membrane from within the cup to the
exterior.

I supposed at first that almost any kind of porous earthenware vessel
would do for experiments of this class, and I spent a good deal of time
in attempting to precipitate phosphate membranes and membranes of
CuzFe(CN)s on alundum filters and on various other kinds of porous
clay vessels which allow a ready passage of water through their walls,
but are supposed to hold back undissolved solids. I found, however,
that all of the precipitates mentioned are easily forced through the
walls of most of these vessels under pressures of a metre of water or
less.

The porous cups which I finally found satisfactory were furnished me
through the kindness of Prof. B. E. Livingston of the Johns Hopkins
University. They are the regular cylindrical atmometer cups supplied
by the Plant World, Tucson, Arizona; and are formed by pouring
liquid kaolin and quartz mixture into dry plaster of paris moulds, and
subsequently drying and burning the cups so formed. These cups are
about 13 em. high with an inner diameter of about 2.2 cm. and en-
tirely unglazed. For the purposes of my experiments I filled the pores
of the upper rims (about 2 cm. in height) with paraffin. This left an
inner surface of about 70 sq. em. on which the membrane was precipi-
tated. The walls of the cup are about 0.25 em. thick. These cups
were connected by means of rubber attachments to an upper piece of
glass provided with a stop-cock and an upright glass tube about one
metre high. The whole cell so formed held from 70 to 80 ce. of fluid.
The amount of fluid passing through the walls of the cup and the
semi-permeable membrane was determined by noting the rise or fall
of the meniscus in the upright tube. I used rubber attachments
instead of the sealing wax used by Pfeffer'® in his experi-

19 Pfeffer: Osmotische Untersuchungen, Leipzig, 1877, 5 et seq.

}

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 463

ments, because I wished to work with alkalies in many cases, and be-
cause it was not necessary for me to make arrangements to withstand
any considerable pressure. I need not describe my attachments in de-
tail. They are such as will readily suggest themselves to anyone seri-
ously interested in the subject. But it may be worth while to say that
I found it much harder than I had supposed to entirely prevent leakage
through rubber attachments, and overcame this difficulty only by using
pretty tight fittings and binding the joints tightly with stout cotton
thread which had previously been wet.

I removed the air from the porous cups by allowing first boiled dis-
tilled water and then a solution of one of the membrane-formers to
seep through them under a pressure of about 70 cm. of water. Each of
these fluids was allowed to seep through for a period of three or four
days. After the cup had been thoroughly infiltrated with the outer
membrane-forming solution, it was emptied of this fluid, rinsed out
quickly with distilled water, dried slightly, filled with the inner mem-
brane-forming solution and allowed to stand for twenty-four to forty-
eight hours in the outer membrane-forming solution, there being no
difference in the pressures within and without the cup. Subsequently
a pressure of about 70 cm. of water was put on the inner membrane-
forming solution, and the cup was allowed to stand for another twenty-
four to forty-eight hours in the outer solution. In this way the semi-
permeable membrane was precipitated on its inner surface.?°

RESULTS OF THE EXPERIMENTS

I had little difficulty in making by these methods membranes of
CusFe(CN). which showed osmotic activity. Experiment 6 shows the
behavior of such a membrane under various conditions, and experiment

7 gives the results of somewhat similar procedures with another copper

ferrocyanide membrane.

These two experiments add a little to the already existing evidence
for the view that the CuzFe(CN).s membrane is impermeable for sugar?!
and highly permeable for NaCl.” Perhaps their most interesting re-
sult, however, is the rapid osmotic intake of fluid caused by the NaCl
solution in spite of the facts that its osmotic pressure (as calculated
from freezing point determinations) was nearly the same as that of

20 Compare Pfeffer: Osmotische Untersuchungen, Leipzig, 1877, 4 et seq.
21 Morse: Loe. cit., pp. 92 and 93.
22 Traube: Arch. f. Anat. Physiol. und wissensch., Med., 1867, xxxiv, 137-141.

464 EDWARD B. MEIGS

the sugar solution used, and that it escaped so rapidly from inside the
cell to the exterior. In the first experiment the osmotic intake caused
by the NaCl was nearly twice as rapid as that caused by: the sugar;
and, as the NaCl period fell between two sugar periods in which the
osmotic intake was nearly the same, it can hardly be supposed that the
result was due to any irreversible change taking place in the membrane.
In the second experiment the osmotic intake caused by the NaCl was
slower than that caused by the sugar, but it seems to me reasonable to
explain this as caused by the decidedly greater leakiness of the second
membrane.

For the experiments with membranes of Ca3(POx)2 the cups were
provided with the same fittings as in the experiments which have just
been described, and subjected to the same preliminary treatment.
They were then impregnated with a 0.09M CaCl: solution and filled
with a 0.075M K,HPO, + 0.075M KOH solution, the details of the
procedure being the same as in the case of the ferrocyanide membrane.
Experiments 8 and 9 show that semi-permeable membranes can be
made of Ca3(PO,)2, and give an idea of the permeability of this mem-
brane with respect to cane sugar and potassium hydroxide. The mem-
brane is impermeable to cane sugar, but quite permeable to potassium
hydroxide.

An interesting result of experiment 8 is shown in the period from
March 13 to March 24. In the first four days of this period the cell
contained a 0.22M C,2H_201, solution; in the next three days, a KOH
solution of about the same calculated osmotic pressure; and in the last
four days, the same sugar solution as at first. It was immersed during
the whole period in a 0.009M CaCl. solution. The cell showed con-
siderable osmotic activity in both the first and second sugar periods,
fluid passing to the interior against a moderate hydrostatic pressure.
But during the alkali period there was practically no osmotic activity
after the first hour, in spite of the fact that the alkali solution had about
the same calculated osmotic pressure as the sugar solution. It is true
that the alkali escaped fairly rapidly from the cell to the exterior. But
the rate of its escape was not many times greater than that of the NaCl
in experiment 6 between the dates March 16 and 20, and yet the NaCl
solution caused a more rapid intake of fluid in this case than did a sugar
solution of the same calculated osmotic pressure. Further, the results
obtained in experiment 11 between November 3 and 11 show that a
sedium chloride solution may cause greater osmotic activity in a mag-

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 465

nesium phosphate membrane than a sugar solution of the same calcu-
lated osmotic pressure, in spite of the fact that in this case also the salt
escapes to a considerable extent through the membrane while the sugar
does not. KOH does not cause the crystallization of the Ca3(POx)>
membrane, nor does it produce any known chemical change in it; the
membrane is formed in the presence of a pretty strongly alkaline solu-
tion. Further, the fact that the membrane showed decided osmotic
activity in the presence of a sugar solution subsequent to its treatment
with the alkali indicates that this latter produced no destructive irre-
versible change in it. It seems likely, therefore, that alkalies and salts
produce entirely different osmotic responses when applied to the sur-
face of a semi-permeable membrane. It would be easy to find some
explanation for this difference if it may be supposed that all such mem-
branes are colloidal in nature, but difficult on any other basis.

Experiment 9 shows the same sort of result with sugar on the Ca3(PO,).
membrane as experiment 8; and, in addition, the result of filling the
cell with a mixture of KCl and KOH. A rather slow intake of fluid is
produced. Several features of the alkali effects shown in experiment
8 are reproduced, but it would be rash to attempt a detailed interpreta-
tion of these results at present.

Experiments 10, 11, and 12 show some of the properties of the mag-
nesium phosphate membrane; the general method of making these mem-
branes was the same as in the case of the ferrocyanide and calcium phos-
phate membranes. These experiments show that magnesium phos-
phate may retain its osmotic activity for two months or more in the
absence of its membrane formers; that it is quite impermeable to cane
sugar; somewhat permeable to KCl] and NaCl; and apparently about
equally so to each of the two salts. It seems to be decidedly less per-
- meable to either NaCl or KCl than is the copper ferrocyanide membrane.
It is highly permeable to ethyl] alcohol.

In addition, two interesting points come out which bear on the na-
ture of osmotic action in general. In experiment 11 the results obtained
between November 3 and 11 indicate that a sodium chloride solution
sets up a greater osmotic activity in the magnesium phosphate mem-
brane than does a cane sugar solution of the same calculated osmotic
pressure, in spite of the fact that the salt escapes through the membrane
fairly rapidly and the sugar not at all. This result is similar to that
obtained in experiment 6 in the case of the CusFe(CN), membrane (see
page 464); it indicates that solutions of neutral electrolytes have the

466 EDWARD B. MEIGS

general property of setting up a more rapid osmotic action in semi-
permeable membranes than solutions of non-electrolytes of the same
- calculated osmotic pressure.

The results obtained in experiment 11 from November 3 to 7, 11 to
20, 1914, and from November 25, 1914 to February 12, 1915 indicate
that the rapidity with which NaCl escapes through the magnesium
phosphate membrane depends, among other things, on the concentra-
tion of the outer solution. During these periods the salt uniformly
escaped more rapidly when the outer solution was more dilute.

In regard to the experiments on the osmotic properties of magnesium
phosphate, it must be added that there were a number of other attempts
to precipitate semi-permeable membranes of this substance on porous
cups, besides the three which are described in the protocols. In some
of these attempts the precipitations were carried out at temperatures
below 23°; in others, the proportions of alkali added to the phosphate
solutions were such as would induce crystallization of the magnesium
precipitates. It was found that the precipitates formed under these
conditions showed no osmotic activity. The experiments with mag-
nesium phosphate taken all together, therefore, are in accord with the
view that precipitates must be in a colloidal condition if they are to
exhibit osmotic acitivty.

EXPERIMENTS WITH UNSUPPORTED MEMBRANES

Traube studied the osmotic properties of a number of substances
by precipitating thin layers of them across the mouths of small glass
tubes. He filled tubes, for instance, with solutions of K,Fe(CN)., and
immersed the lower open ends in solutions of copper salts. Under these
circumstances a layer of CusFe(CN). forms across the open mouth of
the tube. I have attempted to form magnesium phosphate membranes
in a similar way, but without any very satisfactory results. Ifa small
glass tube be filled with a solution of K2.HPO, and KOH and dipped
into a solution of MgCls, a layer of magnesium phosphate is, of course,
precipitated across its mouth. But this layer rapidly becomes thicker,
and the least disturbance suffices to detach it from the walls of the tube;
so that it cannot be used even for the rough kind of osmotic experi-
ment described by Traube.

Membranes of Ca3(PO,)2, however, produced in this way are quite
as resistant and satisfactory for experimentation as the CueFe(CN)s.
membranes. To work with such membranes it is necessary in the first

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 467

place to use a rather small glass tube. Further, the phosphate and
calcium solutions used in making the membrane must have somewhere
near the same osmotic pressure; otherwise, the membrane will be very
soft, and will rapidly become thicker.

In my experiments I have used tubes which had an inner cross sec-
tional area of about 0.14 sq. cm. (inner diameter, about 0.42 cm.).

These were closed at the top with a rubber tube and pinch cock; the
lower ends were inserted through bored corks into small cylindrical
vials.23 In some experiments a few cubic centimeters of 0.186M K,HPO,
+ 0.186M KOH* were placed in the vials and the lower ends of the
tubes were filled with 0.666M CaCl. solution. Under these circum-
stances there is a considerable tendency for fluid to pass through the
membrane from the phosphate to the calcium solution. As the cal-
cium solution has a decidedly higher calculated osmotic pressure than
the phosphate solution, there can be little doubt that this phenomenon
is the result of osmotic action on the part of the calcium phosphate
membrane.

It is perhaps not generally realized by biologists that impermeability
to NaCl and KCl is by no means a usual characteristic of semi-perme-
able membranes. Traube* asserts that a CueFe(CN)s membrane when
infiltrated with AgCl becomes impermeable to KCl. But he finds the
CuzFe(CN).s membrane by itself highly permeable to KCl and probably
also to NaCl (loe. cit., pp. 137-141). Morse** describes experiments in
which KC] solutions were allowed to act on electrolytically deposited
CuzFe(CN)smembranes. He found that old (and probably thick) mem-
branes at first exhibited a high resistance to the passage of KCl, but
that the salt had a tendency to render the membrane permeable to
itself.

In experiments of my own I have found the CusFe(CN)s, membrane
highly permeable both to KCl and to NaCl. In order to compare the
permeability of this membrane to NaCl with that of the Ca3(PO,).
membrane, the following experimental procedure was employed. Four
vials, and four glass tubes with inner diameters of 0.42 cm. were ar-
ranged as described above. Into the lower end of each of the tubes
was drawn 0.28 ec. of 0.125M NaFe(CN). + 0.5M NaCl solution;

23 Compare Traube: Arch. f. Anat., Physiol. u. wiss. Med., 1867, xxxiv, 123 and
133.

24 Made up by adding 32.4 grams KpHPO, and 10.4 grams KOH to liter of water.

25 Traube: Loc. cit., p. 146.

26 Morse: Loc. cit., pp. 211, et seq.

468 EDWARD B. MEIGS

and 10 ce. of 0.25M CuSO, solution was placed in each of the vials.
The lower ends of the tubes containing the chloride and ferrocyanide
mixture were then immersed in the copper solution, and 0.14 ce. of the
chloride and ferrocyanide mixture was forced through the lower end of
the tube and appeared below it as a drop covered by the CusFe(CN)¢
membrane. This drop was then drawn back into the tube carrying
the membrane behind it and finally leaving the latter in a much folded
state across the tube’s mouth.2” The tubes were then immediately
withdrawn from the vials and the solution left in the latter was ana-
lyzed for Cl. The amount found was taken to represent that which
escaped from the inner solution during the formation of the membranes.
The same procedure was then exactly repeated up to the point at which
the drop of chloride and ferrocyanide mixture was drawn back into
the tubes, leaving the folded membrane across its mouth. At this
point the tubes, instead of being withdrawn, were left for twenty min-
utes with their lower ends immersed in the CuSO, solution. As the
chloride and ferrocyanide mixture has a decidedly higher osmotic pres-
sure than the CuSO, solution, fluid passed from the latter to the former
during this period, more rapidly at first and more slowly afterward; at
the end of the twenty minutes the membrane was completely distended
and filled with a drop of diluted chloride and ferrocyanide mixture,
about 0.14 ec. of fluid having passed through it during the interval.
The drops of chloride and ferrocyanide mixture were now again drawn
back into the tubes, the tubes were withdrawn from the CuSQ, solution,
and this last was analyzed for Cl. In both cases, of course, the Cl must
represent NaCl which has escaped from the chloride and ferrocyanide
mixture.

It was found in the first case when practically no time was allowed
for diffusion that the 40 ec. of CuSO, solution contained 0.0015 gram of
NaCl. In the second case, after the diffusion had gone on for twenty
minutes the 40 ec. of CuSO, solution contained 0.0115 gram of NaCl.
It may be supposed that 0.0115 minus 0.0015 or 0.01 gram of NaCl
diffused through the four CusFe(CN), membranes in the course of
twenty minutes. The four glass tubes contained originally 0.28 x 4
or 1.12 ec., of 0.125M NasFe(CN).s + 0.5M NaCl solution or 0.03248
gram of NaCl. Of this 0.0015 gram was lost during the formation of
the membrane, leaving 0.031 gram at the time the diffusion began.

27 Compare Traube: Arch. f. Anat., Physiol. u. wiss. Med., 1867, xxxiv, pp. 136
and 137.

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 469

And of this 0.031 gram 0.01 gram or about 32 per cent diffused through
the four membranes in the course of twenty minutes.

A similar experiment was carried out to test the permeability of the
Ca3(PO,)2 membrane to NaCl. In this case the four tubes were filled
each with 0.28 ce. of 0.24M CaCl, + 0.875M NaCl solution and im-
mersed in 10 ce. of 0.186M K.HPO, + 0.186M KOH solution. It was
found that under these circumstances it required about two hours for
as much fluid to pass through the Ca;(POs)2 membranes as passed
through the CuzFe(CN)s membranes in the previously described experi-
ment in twenty minutes. In this case, therefore, the diffusion was
allowed to go on for two hours instead of for twenty minutes; other-
wise the procedure was the same as in the previously described experi-
ment.

When practically no time was allowed for diffusion the 40 cc. of
0.186M K,HPO, + 0.186M KOH solution contained 0.0021 gram NaCl.
When the diffusion was allowed to go on for two hours, the 40 ce. of
alkaline phosphate solution contained 0.0034 gram NaCl. It may be
supposed therefore that 0.0013 gram of NaCl diffused through the four
Caz (POx)2 membranes in the course of two hours. The four tubes of
0.24M CaCh + 0.875M NaCl solution contained originally 0.0243
gram of NaCl, and of this 0.0021 gram was lost during the formation of
the membranes leaving 0.0222 gram at the time the diffusion began.
0.0013 is about 5.8 per cent of 0.0222, and it appears, therefore, that
only about 5.8 per cent of the NaCl contained in the four tubes diffused
through the Ca3(PO,)z membranes in the course of two hours. That is
about five times as much NaCl diffused through the Cu,Fe(CN),. mem-
brane in twenty minutes as diffused through the Ca3(PO.)2 membrane
under more or less similar conditions in two hours. I have confirmed
these results by other experiments which it is not necessary to describe
in detail, and I am prepared to assert, therefore, that, under the condi-
tions of formation which have been described, the CusFe(CN)s mem-
brane is much more permeable to NaCl than the Ca3(PO;)2 membrane.

PROPERTIES OF THE CELLOIDIN MEMBRANE

In my earliest experiments I used celloidin membranes on which lay-
ers of calcium phosphate were precipitated. Celloidin or collodion
membranes have been used for many years for the study of osmotic
phenomena; Smith?’ has suggested their use as a basis for the precipi-

28 Smith: Science, 1913, N.S., xxxvii, 379.

470 EDWARD B. MEIGS

tation of other forms of semi-permeable material. Since Traube’s
time it has been known that the permeability of sheets of celloidin va-
ried with the circumstances of their formation. Such sheets are formed
by pouring an alcohol-ether solution of celloidin on a flat surface, al-
lowing the alcohol and ether to evaporate to a certain extent, and then
immersing the sheet so formed in water. If the water be applied be-
fore much of the alcohol and ether have evaporated, the membrane will
be found highly permeable; otherwise, it becomes quite impermeable
to water as well as to dissolved substances.?®

I formed my celloidin membranes by pouring the alcohol-ether solu-
tion over the inner surface of a beaker, and carried out several experi-
ments to determine the properties of the membranes so formed under
different conditions of drying. The celloidin sacs were tied around
large rubber corks, and the cells so formed were fitted with long upright
outlet tubes, filled with various solutions, and immersed in others.
When the celloidin membrane was immersed in water less than an hour
after its formation it was highly permeable to water and dissolved salts;
it was easy to show that a solution of K,.HPO, could be rapidly forced
through its walls by the very moderate hydrostatic pressure of 50 to
100 cm. of water. But if the membrane was not wet with water until
after the alcohol-ether mixture had dried out for twenty-four hours or
more, it became highly impermeable, though it showed some osmotic
activity. The following experiment shows this. A celloidin mem-
brane was allowed to dry for twenty-four hours, and then washed for
twenty-four hours in tap water. At the end of this time it was used
in making a cell, which was filled with 0.075M K,HPO, and immersed
in distilled water. The contents of the cell were put under a pressure
of 78 em. of water. The meniscus immediately began to rise in the
outlet tube and continued to do so steadily for the twenty days during
which the experiment was continued at a rate indicating that a little
more than 0.1 cc. of fluid per day was passing into the cell.2° At the
end of the twenty days the water in which the cell had been immersed
was tested for K,HPO, and found to contain none.*! In another simi-
lar experiment, similar results were obtained.

In still another experiment a membrane was used which had been

29 See Traube: Arch. f. Anat. Physiol., und wissensch. Med., 1867, xxxiv, 106°

30 The area of semi-permeable surface was involved in this experiment was
about 60 sq. em.

31 The escape of 1 per cent of the KegHPO, originally contained in the cell could
easily have been detected by means of the test used.

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 471

dried for somewhere between one and twenty-four hours before being
wet with water. The cell made with this membrane was filled with
0.075M K.HPO, solution and immersed in distilled water. The menis-
cus immediately began to rise in the outlet tube, indicating a rapid in-
take of fluid; about 2 cc. passed through the membrane into the -
cell in the first half hour. In the next two hours only about 1.5 ce. of
fluid passed into the cell, and at the end of that time a good deal of
K2HPO; was found to have escaped to the distilled water surrounding
it.

These experiments seem to me interesting, because they show so
clearly that a given chemical substance may have enormously different
semi-permeable properties under different physical conditions. It was
found possible to alter the osmotic properties of the celloidin membrane
by precipitating calcium phosphate on it; but, as it was not possible
accurately to control the variable osmotic properties of the celloidin
itself, the experiments carried out along these lines are of very doubtful
significance, and I do not consider it worth while to give any description
of them.

DISCUSSION

The colloidal character of semi-permeable membranes and the nature of
osmotic action

The results of the experiments described in the preceding pages lend
support to the two hypotheses advanced by Morse—the hypotheses,
namely, that semi-permeable membranes are always colloidal, and that
the passage of water through such membranes is the result of unequal
hydration at the opposite surfaces of the colloid.

The first of these hypotheses is supported by the facts that a mem-
brane highly impermeable to K2zHPO, and yet showing some osmotic
activity can be made of an undoubted colloid, celloidin (pp. 469-471) ;
that semi-permeable membranes can readily be made of Ca3(POx,)2, which
never crystallizes (p. 459 and pp. 464, 465) that semi-permeable mem-
branes can be made from magnesium phosphate only under such con-
ditions as prevent its crystallization (p.466) ;and that the semi-permeable
properties of the various membranes are altered by such influences as
would be likely to change the physical state of colloids (pp.464—466, and
pp.469-471). Finally this first hypothesis is supported by whatever evi-
dence tends to support the second, for the view that osmotic action de-
pends on unequal hydration at the two surfaces of a colloid membrane
presupposes the existence of a colloid membrane to start with.

472 EDWARD B. MEIGS

The hypothesis that the passage of water through a semi-permeable
membrane is the result of unequal hydration of the two surfaces is sup-
- ported by the anomalous results obtained with regard to the amount
of osmotic activity set up by salt, sugar, and potassium hydroxide solu-
tions. A sodium chloride solution causes a more rapid passage of water
through the ferrocyanide and magnesium phosphate membranes than
does a cane sugar solution of the same calculated osmotic pressure,
while an equally concentrated potassium hydroxide solution causes prac-
tically no osmotic activity in the calcium phosphate membrane. The
salt and alkali escape through the membranes with considerable ra-
pidity, while the sugar does not escape at all (pp. 464-466). If the hy-
pothesis just stated is correct, these facts might be explained as the
result of the well known rule that neutral electrolytes have a greater
dehydrating effect on colloids than non-electrolytes while alkalies tend
to increase their power of holding water; I do not see at present how
they can be explained on any other basis.

Certain experiments reported many years ago by Graham® are in-
teresting in this connection. Graham found that the osmotic activity
set up in a plece of pig’s bladder by sugar solutions was very small in
comparison to that set up by salt solutions, and that when the bladder
separated a dilute solution of acid from distilled water, a considerable
quantity of fluid passed through the membrane from the acid solution
to the other side. Graham’s results with the sugar and salt solutions
are similar to mine, though not so clear cut, because the salt solutions
used by him had a much higher calculated osmotic pressure than the
sugar solutions. The result with the acid, however, is very striking.
It seems quite inexplicable from the ordinary conceptions of osmosis and
osmotic pressure, but can be readily explained on the hydration hy-
pothesis from the well known fact that acids tend to hydrate colloids
even more strongly than pure water.

If semi-permeable membranes are really hydrated colloids, it seems
probable that their semi-permeable properties would vary with the de-
gree of their hydration. My experiments furnish evidence for the view
that this is the case, and also for the view that the membranes become
more readily permeable both to water and dissolved substances the
greater the degree of their hydration. Perhaps the most direct evi-
dence for this view is the series of experiments with celloidin membranes

32 Quoted by Girard: Journ. d. Physiol. et de path. générale, 1911, xiii, pp.
365-367.

=

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 473

(pp.469-471). But on p.466 it is shown that the magnesium phosphate
membrane also is more permeable to salts when it is in contact with a
more dilute outer solution and, therefore, presumably in a state of
greater hydration.

Much of the peculiar behavior of living cells would find a ready ex-
planation from the considerations above set forth and from the addi-
tional consideration that in the case of living cells we are dealing prob-
ably always with leaky membranes. The experiments of Morse on
the osmotic pressure of potassium chloride and of lithium chloride
show that we have yet much to learn regarding the osmotic behavior
of electrolytes even with respect to the most ideal semi-permeable mem-
branes that have yet been produced. The conditions become much
more complicated; and the results, correspondingly less predictable,
when we have to deal with such membranes as may be supposed to cover
the surfaces of living cells.

DO MAGNESIUM AND CALCIUM PHOSPHATE PLAY ANY PART IN THE OSMOTIC
PROPERTIES OF LIVING TISSUES?

It is well known that calcium, magnesium, and inorganic phosphates
are constant or nearly constant constituents of living tissues; and there
is much evidence (reviewed at the beginning of this article) for the view
that calcium, at least, plays a very important physiological role, and
has the property of rendering the surfaces of cells less permeable for
other ions. The experiments reported in this article are at least suf-
ficient to show that both calcium and magnesium phosphate are capable
of forming semi-permeable membranes.

Beyond this, however, the experiments do not go very far in showing
either that these substances do or do not play an important part in
giving to living cells their semi-permeable properties. It is hardly pos-
sible that they should, because we know at present so little, on the one
hand, of the influences which affect the permeability of osmotic mem-
branes, and, on the other, of the conditions which obtain in living cells.
Still, it seems worth while to review very briefly those of the results
which bear on this question.

Both the calcium and magnesium phosphate membranes are highly
impermeable to cane sugar, potassium phosphate, calcium chloride, and
magnesium chloride. Under certain conditions they may show a con-

33 Morse: Loc. cit., Chapter XI.

474 EDWARD B. MEIGS

siderable degree of impermeability to sodium chloride and potassium
chloride. The calcium phosphate membrane is permeable to potassium
hydroxide, when that substance is present in considerable quantities.
The magnesium phosphate membrane is quite permeable to ethyl al-
cohol.

The surfaces of living cells are commonly more or less impermeable
to cane sugar and to the salts mentioned above. They are always, so
far as I know, quite permeable to ethyl alcohol. Most cells are killed
by potassium hydroxide when it is present in any considerable quantity,
and, at the same time, their surfaces become permeable to it as well
as to crystalloids generally. Striated muscle cells and kidney cells are
apparently decidedly more permeable to potassium chloride than to
sodium chloride.*t In this respect they seem to differ from the mag-
nesium phosphate cells. But the surfaces of red blood corpuscles are
equally impermeable to both these salts.* These facts and a great many
others®®* show how complicated is the question of the permeability of
living cells.

It is perhaps just to close this part of the discussion by saying that
the semi-permeable properties of magnesium and calcium phosphate, so
far as they have been studied, are as much like those of living tissues as
are the semi-permeable properties of any other artificial osmotic mem-
branes that have been examined up to this time.

In conclusion I wish to extend my most heartfelt thanks to Dr. J ohn
Marshall and to the staff of the Hare Chemical Laboratory of the Uni-
versity of Pennsylvania, without whose kind assistance and support the
work reported in the foregoing article would have been difficult or
impossible.

SUMMARY

1. Semi-permeable membranes can be formed both from calcium and
from magnesium phosphate.

2. In order to form semi-permeable membranes from magnesium
phosphate, it is necessary to precipitate it under such conditions that
it does not crystallize.

34 Siebeck: Arch. f. d. gesammt. Physiol., 1912, exlviii, 443; 1914, cl, 316.
Meigs: Journ. Exper. Zool., 1912, xiii, pp. 518. 5?0.

35 Hamburger: Osmotischer Druck und Ionenlehre, Wiesbaden, 1902, vol. 1,
pp. 208 and 209.

36 The surfaces of human red blood cells and of these of certain other animals,
for instance, appear to be quite permeable to dextrose; see Kozawa: Biochem.
Zeitschr., 1914, lx, 231; Masing: Arch. f. d. gesammt. Physiol., 1914, clix, 476.

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 475

3. The calcium phosphate membrane is impermeable or nearly imper-
meable to cane sugar, dipotassium phosphate and calcium chloride, only
slightly permeable to sodium chloride, and quite permeable to potas-
sium hydroxide.

4. The magnesium phosphate membrane is impermeable or nearly
impermeable to cane sugar, dipotassium phosphate and magnesium
chloride, somewhat permeable to sodium and potassium chloride, and
highly permeable to ethyl alcohol.

5. Anomalous results are obtained when the calcium phosphate,
magnesium phosphate, and copper ferrocyanide membranes are sub-
jected to the action of neutral electrolytes and alkalies. Sodium chlo-
ride solutions cause a more rapid osmotic action in the copper ferrocya-
nide and magnesium phosphate membranes than do cane sugar solu-
tions with the same calculated osmotic pressure, in spite of the fact
that the salt escapes through the membrane to a considerable extent,
while the sugar does not. Potassium hydroxide produces no lasting
osmotic activity in the calcium phosphate membrane, though it appar-
ently causes no marked irreversible change in the membrane, and does
not escape through it very much faster than does sodium chloride
through the copper ferrocyanide membrane.

6. It is shown that the semi-permeable properties of celloidin mem-
branes depend to a very great extent on their physical state, and reasons
are given for thinking that this is true of semi-permeable membranes
in general. .

7. Evidence, in addition to that already given by Morse, is adduced
for the view that semi-permeable membranes are always colloidal, and
that the passage of water through them is the result of unequal hydra-
tion of the colloid at its two surfaces.

476

12.05 to 12.30 p.m., {12.05 to 12.30 p.m.,/12.05 to 12.30 p.m.,|12:05 to 12.30 p.m.,

April 30, 1914.
10=cic) OF. 5) Mi
KesHPO, mixed
with 8.4 cic:
0.267M MgCl,

12.05 p.m., May 2.

Precipitate is
now almost en-
tirely crystalline

11.40 a.m., June 6.

Precipitate now
entirely crystal-
line.

2.05 p.m., Oct. 14, 1914.

EDWARD B. MEIGS

Experiment 1

April 30, 1914.
LO Meic=) “Or SINE
Ke»HPO, mixed
with!) 8.4 ce.
0.267M MgCl. and

0.8 ee. 0.86M KOH

April 30, 1914.
10 ec. 0.15M
K,HPO, mixed
with8.4cc.0.267M
MgCl, and 1.6 ce.
0.86M KOH

April 30, 1914.
10 ec, 10-15
K,HPO, mixed
with8.4cc.0.267M
MgCl: and 2.0 ce.
0.86M KOH

12.15 p.m., May 2./12.15 p.m., May 2./12.15 p.m., May 2.

Precipitate con-
tains a few crys-
tals, but is still
for the most part
amorphous

Precipitate con-
tains a good
many crystals,
but is still, for
the most part,
amorphous

Precipitate is still
entirely amor-
phous

11.50 a.m., June 6./12.00 m., June 6./12.00 m.,

Precipitate — still
entirely amor-
phous

Experiment 2

10 ec. 0.15M

K2HPO, mixed with 5 ec. 0.2M MgCl.

Precipitate is now
chiefly made up
of crystals

June 6.
Precipitate now
almost entirely
crystalline

3.00 p.m., Oct. 14, 1914. 5 ce. 0.15M

K2HPOg

mixed with 10 ce.

0.2M

and 0.5 cc. 0.86M KOH

3.95. Precipitate contains a few small
crystals, but is still, for the most part,
amorphous

10.30 a.m., Oct. 15. Precipitate is still
largely amorphous but contains a
good many crystals

11.15 a.m., Oct. 16. Precipitate is now
about half crystalline and half amor-
phous

MgCl, and 0.5 ec. 0.86M KOH
4.00. Precipitate still entirely amor-

phous.

10:35xe.m., Oct. 15.
entirely amorphous.

Precipitate still

MS paenicg OC. m0.
entirely amorphous

Precipitate still

Experiment 3

3.40 p.m., Dec./4.10 p.m., Dec./12.45p.m., Dec./11.50 a.m., Jan.|2.54 p.m. Dee.

4,1914. 5cc.| 1,1914. 5ce.| 10, 1914.5 ce.
0.3M MgCl,; 0.3M MgCl} 0.3M MgCl,
mixed with 5| mixed with 5) mixed with 5
cc: 0.134M| cc.0.134M K,| ce. 0.134M
K2HPO, +) HPO;+0.096) K-HPO,; +
0.096M KOH.| M KOH.) 0.096M KOH.
Temp. 10° Temp. 18° Temp. 20°
11.27.a.m., Dec.}10.10 a.m. Dec.|9.25 a.m., Dec.
5. Precipi-| 2. Precipi-| 11. Precipi-
tate entirely} tate entirely} tate about
erystalline| crystalline.| half crystal-
Temp. 10° Temp. 18° line and half

amorphous
Temp. 20°.

12. Precipi-
tate is about
two - thirds
erystalline
and one-third
amorphous.
Temp. 19°

19,1915. Pre-
cipitate now
entirely crys-
talline. Temp.
21° <2)

Dec. 10, 1914,
and Feb. 19,
1915, the tem-
perature va-
ried between
10° and 25°

477

23, 1915. 5c.
03M MgCl,
mixed with 5
ce. 0.134M
K2,HPO, +
0.096M KOH.
Temp. 23°

10.20 a.m., Dec.|9.30 a.m., Jan.

25. Precipi-
tate contains
a fair number
of crystals,

but is still,
for the most
part, amor-
phous. Temp.
14°

19,1915. Pre-
cipitate con-
tains a good
many crys-
tals, but is
still, for the
most part,
amorphous.
Temp. 21° (1)

Between
Jan. 25 and
Feb. 19 the
temperature
varied be-

tween 10° and
25°

11,1914. 5ce.
03M MgCl,
mixed with 5
ce. 0.134M
KeHPOx, —
0.096M KOH.
Temp. 27°

| | ee) SSS

10.30 a.m. Dec.

12, Precipi-
tate still en-
tirely amor-
phous. Temp.
21.5°

ee

11.47 a.m. Dec.

15, Precipi-
tate still en-
tirely amor-

_ phous. Temp.

15°

12.30 p.m., Feb.|2.30 p.m., Feb.|12.47 p.m. Feb.

19, 1915. Pre-
cipitate con-
tains a few
crystals, but
is still, for
the most part,
amorphous.
Temp. 21° (1)

(1) Between/(1) (1) Between

Dec. 12, 1914
and Feb. 19,
1915 the tem-
perature va-
ried between
10° and 25°

478

EDWARD

Experi

B. MEIGS

ment 4

3.40 p.m., Dec. 4,/12.45 p.m., Dec. 10,|11.55 a.m., Jan. 23,/2.54 p.m., Dec. 11,

1915. 5 cc. 0.5M
MgCl mixed
with 5 cc. 0.2M
K2HPO, + O0.14M
KOH. Temp. 10°

11.35 a.m., Dec 5.
Precipitate en-
tirely crystalline.
Temp. 10°

1914. 5 cc. 0.5M
MgCl. mixed
with 5 cc. 0.2M
K.HPO, + 0.14M
KOH. Temp. 20°

9.30 am., Dec: 11.
Precipitate en-
tirely crystalline.
Temp. 20°

1915. 5 cc. 0.5M
MgCl. mixed
with 5 cc. 0.2M
K.HPO, + 0.14M
KOH. Temp. 23°

10.10 a.m., Jan. 25.
Precipitate con-

tains a large num-
ber of crystals,
but is still chiefly
amorphous.
Temp. 14°

1914. 5 cc. 0.5M
MgCl, mixed
with 5 cc. 0.2M
K.HPO, + 0.14M
KOH. Temp. 27°

10.32 a.m., Dec. 12.
Precipitate still
entire ly amor-
phous. Temp.
21.5°

11.53 a.m., Dee. 15.
Precipitate still
entirely amor-
phous. Temp. 15°

2.33 p.m., Feb. 19,/12.50 p.m., Feb. 19,

1915. Precipi-
tate contains a
large number of
crystals, but is
still chiefly amor-
phous. Temp. 21°
(1)

1915. Precipitate
contains a few
crystals but is
still chiefly amor-
phous. Temp.
21° (4)

(1) Between Jan. 23/(1) Between Dec.

and Feb. 19 the
temperature has
varied between
10° and 25°

12, 1914 and Feb.
19, 1915, the tem-
perature has va-
ried between 10°
and 25°

T— » So

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 479

Experiment 5

12.00-12.05 p.m., Dec. 7,/12.00-12.05 p.m., Dec. 7,/12.00-12.05 p.m., Dec. c

1914. 5cc.0.2M MgCl.) 1914. 5.15 ce. 0.19M} 1914. 5.8 ee. 0.17M

mixed with 5 cc. 0.134M| MgCl, + 0.019M CaCl.) MgCl.+ 0.086M CaCl. +

K:H PO, + 0.096M| mixed with 5.15 cc.) 5 cc. 0.134M K,HPO,+

KOH. Temp. 20° 0.134M K2,_HPO4+0.096M| 0.096M KOH. Temp.
KOH. Temp. 20° 20°

3.15 p.m. Precipitate en-|3.22 p.m. Precipitate|3.25 p.m. Precipitate en-
tirely crystalline.| about half crystalline| tirely amorphous. Temp.
Temp. 21° and half amorphous.} 21°

Temp. 21°

10.00a.m., Dec.8. Precip-|10.05 a.m., Dec. 8. Pre-

itate about half crystal-| cipitate contains a very

line and half amorphous.| few crystals, but is still

Temp. 17° almost entirely amor-
phous. Temp. 17°

2.20 p.m., Feb. 19, 1915./2.22 p.m., Feb. 19, 1915.
Precipitate about half} Precipitate contains a
crystalline and half} very few crystals, but is
amorphous. Temp. 21°) still almost entirely
(1) amorphous. Temp. 21°

(1)

(1) Between Dec. 7, 1914/(1) Between Dec. 7, 1914
and Feb. 19, 1915 the} and Feb. 19, 1915 the
temperature varied be-| temperature varied be-
tween 10° and 25° tween 10° and 25°

EXPLANATION OF EXPERIMENTS 6 TO 12 INCLUSIVE

These experiments were carried out on membranes of CusFe(CN).,
Ca3(PO,)2 and magnesium phosphate (see p. 460) precipitated on the in-
ner surfaces of such porous cups as are described on p. 462. The dimen-
sions and capacity of the cups are given on p. 462. The preliminary
treatment of the cups and the general method by which the membranes
were precipitated are given on pp. 461 to 463. The succeeding protocols
give a condensed history of the behavior of the osmotic cells from the
period when the actual membrane formation was begun.

The solutions described are what Morse calls ‘“weight-normal so-
lutions” (Osmotic pressure of aqueous solutions, Chapter V). That is,
a “molecular” solution of cane sugar would mean a solution made by
adding 342 grams of cane sugar to a litre of water. In the same way,

480 EDWARD B. MEIGS

a “0.5M NaCl + 0.5M KCI solution” would mean a solution made by
adding 29.2 grams of NaCl and 37.2 grams of KCl to a litre of water.

The amounts of dissolved substances which passed through the walls
of the cells were determined by using solutions of known strength at
the beginning of the experiments, and making a chemical analysis of
either the inner or the outer solution at the end. Which of these pro-
cedures was adopted in each particular case is shown in the individual
protocols. In some cases both inner and outer solutions were analyzed;
and it was then found, as was to have been expected, that more solute
disappeared from the inner solution than could be recovered in the
outer solution. In experiment 12, for instance, at the end of the period
between January 28 and February 3, 1915, it was found that 0.18 gram
NaCl had disappeared from the inner solution while only 0.04 gram was
recovered in the outer solution. The 0.14 gram left unaccounted for
was no doubt held in the membrane itself and in the pores of the earthen-
ware cup.37. The difference observed in this case was no doubt extreme.
It would be much less, for instance, in experiment 6, March 16 to 20,
where the membrane and pores of the earthenware cup already con-
tained much NaCl when the experimental period in question was begun.

Sugar was determined by means of the saccharimeter. Under the
conditions of the experiment this method made possible the determina-
tion in the outer solution of about 1 per cent of the sugar originally con-
tained in the inner solution. Alcohol was determined by means of the
pyknometer. The method adopted made possible the determination
of the escape of about 0.5 per cent of alcohol from the inner solution.
Sodium and potassium chloride were determined by the Vollhard-Ar-
nold method of chlorine determination (see Hawk: Practical Physiologi-
cal Chemistry, 3d edition, 1910, pp. 390-391); and potassium hydrox- .
ide, by titrating against HCl with azo-litmin as an indicator. These
determinations are decidedly more accurate than either the sugar or
alcohol determinations.

Most of the experiments were carried out at room temperature,
which underwent considerable variations; and it was noted, of course,
that.changes in temperature caused changes in the position of the menis-
cus in the outlet tube independent of the osmotic changes of the quan-
tity of fluid in the cell. The temperature changes, however, were small
in comparison to the osmotic changes. Where small osmotic changes
are recorded throughout the experiments, it may be taken for granted
that the readings on which they depend were taken at the same tem-
peratures.

37 Compare Morse: Osmotic pressure of aqueous solutions, p. 213.

481

M PHOSPHATE

MAGNESIU

AND

PROPERTIES OF CALCIUM

06

09

O&T

08

08

UALVAM TO
“WO NI NGAIO 1109
JO HOIMALNI NI
auossaud ADVUAAV

“uSI6 SNUIUL 94} AG PoxBOIPUT SB ‘UOTFN]OS 109NO OY} OF [[99 OY} UIYIIA WoIJ podvose ping *99 2°) z
“smOrgNn[os 199no ay} SurzAjeue Aq poutuliojop ‘yuoultedxe sity} JNoYsNoyy ‘o19M ‘pedwose YOY O4N]Os Jo sorytzUENDH oT, 7
*,9% PUB ZI UEdMJoq poeTIVA YOIYM o1nyv1i9duls} WOOL 4B INO PoTIIVO SBM JUSUITIEdXe SIT],

0 S7&

IOBN Jo % 8°¢
peuraniezep JON og
pourmieyep JON oh 0—

0 LY
poauluiejep JON 80%
pouruiejeap ION 40°0

0 (1p)

sooed J, 0

0 poururiajap JON

0 peururieyap JON
yaddVOSa HOIKM ‘00 NI

‘V1d0 NI GANIVINOO
ATIVNIDIVO ALATOS
JO ADSVINGDOUudd

MOAINI G1n1a

FOSNO W2400°0

FOS") WZ00'0

Fosn) W200'0

FOSNO W200°0
+IOVN W2r'0

FOSND W200°0

FOS") W2L00°0

FOS D W200°0
+NO@HAD WZS'0

FOS"OD WZI0'0

FOS"D WZI0°0

FOSNOD Ws0'°0

‘OSU INGO 0

NOILOTOS UdALoo

tI6El ‘g hupnigag 9 quawmisaday

(NO)9d'M E000
+NOZHAD WZs'0

9(NO)94'BN IWE00'0
+I0®N NZT'0

%NO)24"8N INE00'0
+IO®N Wet0

(NO)°4°8N WE00°0
+IO®N WEE 0

%(NO)@M W £00°0
+HO*@H*D WZ2'0

°(NO)°4'M WE00'0
+"O#@HAD W2s'0

°NO)°a"S IWG00°0
+NOR@HAD NZ '0

°NO)®4'*M W80'0
°(NO)2a'M IWGO'0
%(NO)94'M IWS0°0

%NO)°4" IN¢0'0

NOILOTOS HANNI

#Z 97:06 “ABW

0@ 07 OT “1B

OT 09 GT “1B

61 07 OF “TFN

01°F € “ABI

€ “TB 9796 “F9A

96 94 €3 “99H

& OF LT “F9A

11996 “9d

(ij CNG) GEN

FIGT “9 099 “Gog

WIL

MEIGS

EDWARD B.

482

‘UZIS SNUIUL 94} AQ PdzBOIPUI 8B ‘UOTYN]OS JONO OY4 OF [[90 OY} UIYIIM WOT possed png “00 J’) ¢
“SUOIJNIOS 109NO OY} SulzA[vue Aq poululiojop ‘yuoulliedxe S14} JNOYsNOIY4 ‘d19M podwose YOIYA 9N[OS Jo selqIgUBNb oyy, +
*.9Z PUB .Z] W9EMJoq palIVA YOY ‘dIN}e19dUI9} WOOL 4B YNO poliied SBA JuaUITIOdxe BIYT,

9(NO)94'°M WE00'°0

oor 0 O-LT FOS") W000 +NOZ@HAD W220 FI OF OF “ABN
9(NO)°a"EN IWE00°0
06 1OBN JO %9'9 oF Ss OS") WZ00'0 +10®N W2I 0 01 939 “ABT
%NO)°4'S IWE00'0
021 NORHAD JO %1°S é 81 ’OSnD WZ00°0 +NO@HAD W220 9 03 Z “ABW
9(NO)9a"S INE00 0
0ZT peutulieyap JON 9°8T FOS") WZ00'0 +NO@HAD W2S'0 G ‘ABW 93 9% “G94
¥OS"D WZ00°0 °NO)°4'M WE00'0
OF 0 2L0—- +MO%2H7D W220 +NO@HD WZZ'0 9% 93 $6 “G9
0 0 peulul1azp JON, OS") WE0'0 9(NO)°4'M IWS0'0 £% 9F 16 “G94
&6 0 9¢€ ’OsnD WS0 0 NO)94'M W800 1Z 93 0% “994
¢9 0) L410 YOSUD WSs 0 9(NO)°4°M IN80°0 0G 93 61 “99d
06 0 99 FOS"O Ws 0 %NO)°4'M IN80°0 61 94 9T “G94
£ 0 peururieyep JON FOS") W220 9(NO)94'°M W80°0 PIGI ‘OT 99 FT “990
UadLvyM AO yaadVOSa HOIHM “09 NI
‘WO NI NGAID I100 ‘71a0 NI GANIVLNOO
dO MOIMALNI NI XTVIVNIDINO @LAI108 MOTANI GIn14 NOILATIOS HALNO NOILO IOS UANNI GWIL
Gunssadud ADVUAAV HO GADVINGOUNGd
% YI6L ‘yT haonsgag ‘4 quaumruadxy

483

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE

“sunoy 1% 484y 9Y} Ut
00 Zp'0 ‘anoy ysuy OY} UT UT pessed ping JO “09 GTQ “UOTQN]OS Bsns oY AQ poov[doi SBVA\ T]BH]B OY} Jogye Ajpayoipawwud AJIATZO’ OLYOUISO POMOYS [99 OUT, »
*[[99 049 WOIy PNY Jo sso] FYSI[S AIA v SBM J1OY} ‘IOAVMOY ‘SAP INO} ZUIpoooons 9Y} UT “MOY J/BY puodes oY} UT “90 #0°0 pue inoy sey sig
oY} UT UT passed “90 gp'9 !pxBAUT JNOYIIM WO pny Jo odussed Suryse]-qsoys YSNoy} prdva ojinb *B 4s1y 4B SVM O19Y} ‘1[BYTB YITM [09 oY} SurI[[Y Jo 4[Nso1 BSW ¢
“uSIS SNULUL AY} AQ PayBoIPUT SB UOTZNIOS 109NO 94} OF [[29 OY} UIYIIM UWOIJ possed pny Jo *90 CTT z
*SUOTJN[OS 19jNO oy} BurzAjeue Aq poululsoejep ‘uouITIEdxe sty) YNoYSNo1YY ‘o19M podvose YOIYM o4N]Os Jo sorqztyUBNh oy, +
*.9% pusw ,G] WseMJoq polIVA YOIYM ‘oinjvs9dUIe} WOOL 4B 4NO poliiGo SBA JUSUITIOdXO SITY],

HOM W200'0
+'OdH*s WS200°0

s8 0 621 *1[08) W600 °0 +NO@HAD Wes '0 86 03 FG “IBIN
HOM WS200°0
; +'OdH*SM WS200°0

$9 0 198° *TO®) W600'0 +NO@H*) W260 #6 99:06 “ARN

'OdH*M WE200°0
09 HOM J° %02 10 TO) W600'0 +HOM WL21'0 0¢ OF LT “ABN

HOM WS200°0
'OdH*M WE200'0

00T 0 68 *[08) W600 °0 +NO@HAD W220 LT °F €T “T8W
HOM W20°0 :
&¢ 0 alt TO8) W600 +'OdH*M W100 £1 09 IT “AW
HOM W¢20°0
0 0 peurur1e,ep JON *TO®) W60'0 +'OdH*M WS20°0 POF SIN
UGLY M LO 11d VOSH HOTA "20 NI
‘WO NINQAID 1100 ‘77100 NI GaNIVINOO F
nO AV OTaIRD IE ee re MIaTEG TLEOTIGG MOTANI a1n 1a NOILA'I0S uaLno NOILA108 YANNI ANIL
aUASSaud GOVUAAV 410 AOVINGOUG

FI6I ‘6 yotopy 8 quamisaduy

“UOTINIOS IVsNs ay} Aq pooe[dos sv UOTINJos OUT[eH]B OY} JoO4Je AjoJVIpoMUT g JUOUTTIOdXY] U1 se ose SITY UT APIATPVOV ONOUISO POMOYS [[99 OT, »
“INOY PUodes OY} UI "99 10") PUB INOY 4s1y 94} UI UI pessed pring Jo ‘90 Z7'9 !paBMuT 4yno
“{}IM Wosy pny jo osussed prdvr Ayyensnun uw ‘g yuowTodx| Ul SB OsBd SIYY UT ‘4SIY 9B SLA O14} ‘UOIJNI[OS OUT[eYLe OY} YALA [[90 94} ZUT|[Y JO F[NSeI BSW ¢
“USIS SNUTUL OY} A po}PVoIpUT SB MOTFN]OS 10}NO OY} OF [[99 OY} UITIIM UIOIy possed PMY jo *99 ZI]
*SUOTINIOS JOINO 94} SuizATeuE Aq poulUJojzop ‘yUoUUNTIEdxe S1Y} JNOYZnosY} ‘aoa ‘podvose Yor oyNyjos Jo saryyUENb eu.

“093 PUB GT UdEMJoq patIVA YOY ‘omyvs9duI} WOOI 4v 4NO patizvO SBM JuOUTTIOdxe stTyT,
SSS SS a Ee eS eee eee eee

HOM W¢200'0
+'OdH*M W¢200°0

29 0 6°0 10%) W600°0 +NO#HD W220 8% 09 9g “UN
o See Sr coe See Sel ee
oS 'OdH*M WE200°0
gl : +HOM W2e0'0
a 29 HOM J° %0z e8°0 “10% W600°0 +IOM 1W980°0 9% 04 1 “AUN
. HOM W¢200°0
a +'OdH*M W¢200°0
5 gL 0 91 10%) W600°0 +"O#H"D W220 1 07 OT “18
= a
a HOM WE20'0
zi eg 0 201 I— “10% NGO'0 +'0dH*M W200 QT 0 Sp “By
HOM W$20°0
0 0 pouturse}9p ON 10%) W60'0 +'OdH*M W200 G1 0 IT “Ie
UaALY AM fO 1yaaddvoOsad HOIHM “00 NI
Bt heacaer caae Gilde antenenlae Me MOTANI GIA TA NOILOIOS “HaALNO NOILQN10OS HANNI GWILL
G2YOsSsaud ADVUAAV JO GVINGudd

a

TIGL ‘TT yowyy °6 Wawaday

484

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE 485

Experiment 10. May 4, 1914

AVERAGE
PRESSURE
IN INTER-
TIME INNER SOLUTION OUTER SOLUTION FLUID INFLOW | !0R OF CELL
GIVEN IN
ws co. | SENET
WATER
May 4to6 | 0.2M MgCl, | 0.134M K,HPO,+ | Not determin- 0
0.092M KOH ed
May 6 to 8 0.2M MgCl, 0.134M K,HPO.+ -1.12! 40
0.092M KOH
May 8 to 0.2M MgCl, |} 0.0134M K:HPO,+
June 8 0.00922M KOH 10.3 75
June 8 to 29 0.2M MgCl.| 0.0134M K.HPO,+ 12.0 100

0.00922M KHO

This experiment was carried out at room temperature which varied between
18° and 29°. The passage of dissolved substances through the walls of the cell
was not followed at all.

11.12 ce. fluid passed from within the cell to the outer solution, as indicated
by the minus sign.

MEIGS

EDWARD B.

486

8¢

poeurulie}ep JON

pauruseyap JON

TOM 3° %e't

IOBN JO %G'0

‘OdH*M WL0°0

*103W WSI10'0
+TIOM WHIT'0

0€ 94 LZ “AON

103 WS10'0
+IO®N Wel'0

peuruiia39p JON

poeuruliejap JON

pouturiey9p JON

paururiezep JON

peururieyap JON

PeurUIIEJOp JON

poeurtuiejap JON

poeurUtiej}ap JON

HOM W9600°0
+'OdH*S WFE10°0

HOM WW960°0
+'OdH*M WET ‘0

'OdH*M WSL0°0
'OdH*M W200 0
‘Od H*M WS20°0

HOM W8¥00'0
+'OdH*M W900 °0

s
HOM Wst00°0
+'Od HH? W2900°0

7103 We'0

ZUOI}NIOS 8,1OSUTYY

ZUOI}N[OS 8, 19sUTYY

ZUOTINIOS 8,19BUIAT

7103 WST0'0
+MO@H*D IGS'0

10310 WST0'0
+IO®N WeT'0

&@ 07 1G “AON

TZ 03 02 “AON

0G 03 9T “AON

9T 9} TT “AON

UagLyYM 10
“WO NI NGAID 1100
40 UOIMALNI NI
GuoOssaud ADVUAAV

peururieyap JON

NOILO1TOS HALNO
WOuUd T1140 GA4YaLING
HOIHM ALOIOS
40 ADVINGOUad

yaaddvOsad HOIHM

“ITa0 NI GANIVINOO

ATIVNIDIUO ALATOS
40 GADSVINGONdd

HOM W600 °0
+'OdH*M WFET0'0

HOM W60°0
+'OdH*M Wel '0

HO We60°0
+'Od HM WHET 0

11 032 “AON
103 G “AON
€ 0% ‘4 “AON

G “AON 073 08 “990

MOTANI GInTs

NOILQO10O8S HaLoAo

NOILOIOS WANNI

TI6I ‘6 4290799

‘TT quauisaday

487

MAGNESIUM PHOSPHATE

PROPERTIES OF CALCIUM AND

“UDIS SNUIUL OY} AG PO}PBOIPUL SB UOTPNIOS 109NO OY} OF []99 OY} UIYIIM Woy possed pny ¢

“1OPBM “09 NOT OF *OOHBN Wess Z0°9 pur ‘TOO wes ¢70'0 ‘TOM Wes 70'0 ‘TOVN wWras ¢9'0 SuIppe Aq opr ;

*SUOTFN]OS 19jNO oY} BurzA[vue Aq poutulsojop ‘yusuITIOdxo S14} JNOYSNOIYY ‘919M podweose YOIYA O4N]OS JO sotgIQUBNDH o,f, 7

“BW Weis 6E0'0 PUB BD OU UTBZUOD OF puNo}] pu ‘pezA/euR ‘poz09]]00 SBM 4]

“19}BM JO “UD 0% enone jo ounssoid @ aepun BAP CT JNoqe 10j dno snosod oy} YRno1Y} does 07 peMoT]"e SEM [OH 3U0O sod J*g YuoUITIOdxO SIqy JO pus 9y9 IV

“99% PUB cy] USBMJOq POTIBA YOIYM ‘oinyB19dUIe} WOOL 7B INO polio SVM FUOWUTIOdXE OY], :

06 ~- MOM) JO %ZO1 00°61 TOYVM Pa [siq NOMAD WSO 1g 09 ST “IRIN
HO'HN IW8#0°0 “103 INZ0°0

rad pouruiieyap JON pouruioyop JON 8¢°0 +*Od H*3 W290 °0 +1OBN Wor'0 | 8t “RW OF ST “49
HO'HN W8#0'0 7103 W200

e¢ peululseyap JON peurutiejop JON 0 +'OdH*3 W290 °0 +10®N W2r'0 8ST 09 ¢T “G9
HOM Wst0'0 7103 W200

8¢ HOM Jo %9°0 0 0 +'OdH*M WL90'0 +IO®N W2r'0 GI 09 31 “420

HOM Ws8t0'0 7103 WST0'0 SI6L ‘ZT
99 HOM Jo %F's 1OBN Jo %9°9 e'0 +'OdH*M W290°0 +IO®8N WZL'0 |'494 OF FI6E ‘8 99

“1030 WS10'0
39 PeUTUrTe7{op AON, IO®N JO %F'S 806 0— "OdH®M WE20°0 +IO®N WZT'0 | 8 99d 04:08 “AON

pourursajop JON

sHO'H*O Jo %G" ez

HOM Ws#0'0
+'Od HM W90'0

*1O4W WI'0
+HO'H*) WST6'0

8% 0F Gg “URe

HOM Wst0'0
+OdH*s W290°0

HOM W8t0°0
+'OdH*™ W290'0

7103 WI'0
+IO®N W80'0

HOM W8#0'0
+OdH*S W290°0

7103 WI'0
+10®N W800

MEIGS

6g

EDWARD B.

poururiajap ON

poeururezOp JON

TOM JO %6'%

TIOBN JO %8'S

HOM W800
+'OdH*M W900

HOM W800
+'OdH*M W290'0

O38 WI'0
+1OM IN80°0

71030 WI'0
+IOM W80°0

a

HOM W8#0'0
+'OdH*M W90'0

HOM Ws8t0°0
+OdH*M W290°0

HOM W8?0'0
+'OdH* W90'0

O3W WI'0
+IO®N W800

“103 WI'0
+10%N W80'0

a ne ee ee

11996 “ae
6 039 ‘ue
9 OFF ‘uRe
SI6l ‘b “uBe

| hd OC | |S | Os

UaALYM AO
“WO NI NGAIDN 1100
JO HOLMALNI NI
FUOsSsdud ADVUAAV

488

peurUrIe}ap JON

poeururIejep JON

NOILOIOS UALnAoO

WoUud T1990 CAYAINGA

HOIHM WLO'1OS
HO ADVINGONAd

daadvosa HOIKM

‘T1090 NI GUNIVINOO
ATIVNIDINO ALNIOS

40 ADVINGOUNdd

HOM W960°0
+'OdH*M WET '0

HOM W960°0
+'OdH*M Wel '0

MOTANI GI0TA

NOILAIOS YALNO

TI6I ‘8I “aquasaq $= ‘é] quawmisaday

NOILOIOS UANNI

489

PROPERTIES OF CALCIUM AND MAGNESIUM PHOSPHATE

“UdIS SNULUL oY} Aq pozvoIpUT SB ‘MOIFN]OS 109N0 oy 09 dno oy UuITIIM ulO1 possed PING *99 ZOD ¢
“WOTINIOS IOUT ey} SuizAyeue Aq pouruliojop podvase yoryM o4nyos Jo AyQUEND z
“WOTINIOS 1OINO ey4 SuizAyeue Aq pourusojop podvose yoryM ynJos jo Ay1ZUENY 7
“096 PUB ,FT WOOMJoq PABA YOrYAr ‘oMzvsI9d W194 WOOT 4v 4 doy SBM [199 BY} GIG] ‘Ff OUNL pus ayep STy4 UddAQog “CI6T ‘CZ “Goq [19uNn .¢g pue
08% WOEMJod doy SBM 41 YOY 109jB ‘9g puw ZZ USaKjoq PalIVA 41 gz “oaqE “WI-e EEG puve oul} JY} UdIMJo *,¢°ZZ OF UIT][B} pwYy oInqvioduiey oy} FZ ‘00q
“Ul"8 0F6 FV ‘oCf PUB ,6% U2BMJeq SaeInqeIaduIE ye ydoy seM juSUTIedxe Sty} UT pesn []29 94} FIG] ‘EZ ‘00q ‘uI'd O¢'e pue gt ‘90q ‘Ul'e NET] UseMyog

——— Se a ee ee a ee

oor — 0 Fos TOPE Pol [st NO@AAD WS'0 p ounf 0}, “Ady

GL > 0 (amt JOFEM PIT[TSIG NOH) WS Lady 0} gy “rey
HOM Ws8t0°0

GL PeurUneyep ON 0 80°T +OdH*M W900 703 We'0 8T “TRIN 99 6 “49
HOM W8t0°0 7O3W WI'0

+9 PouTrurie}ep JON, zHO'H“D 3° %1'9% 8&°0 +OdH*M W900 +HO'H*O WIT 69 € “FH
HOM Ws80°0 7103 WI'0

GL PeuraTIE};9p FON zlOBN 3° %L'6 8¢°0 +'OdH*M W900 +10®N W9'0 £997 07 8g “uer

ee ee eee

“44. -, ' creiyiigint ahs
we 4 : ” eats vail : * The ti ot Viti ae

; Dt an Ta eo) ae ee aah lane

ie oan ae bo rhicngh
> * ‘1 eats ve

: 3 ae .
+ in ’ .
: J : ¢ ape: "
“
z ra % , J
tet Way
7 t wk owe
.
iy A
&
A
>
‘ ‘
;
'
- t
.
' )
é
:
‘
:
= —_
.
+ i .
‘
j
‘ \ .
.
:
*
*
.
1.
i
‘
.
' *
.

INDEX TO VOLUME XXXVIII

DRENALIN, effect of on the heart-
rate, 62.
Adrenin, effects of on splanchnic and
peripheral arteries, 4388.
Anaesthesia, spinal in the cat, 108.
Anasthetics, effect on coagulation of
blood, 33.
Antithrombin, origin of, 233.
Arteries, effects of adrenin on, 438.
Auricle, effect on ventricular output,
406.

BACTERIA, killed by exposure to
ultra violet, 401.

Brenepict, F. G. and L. E. Emmgss.
A calorimetric calibration of the
Krogh bicycle ergometer, 52.

Blood, coagulation of, 33.

pressure, influence on membrane
manometer curves, 209.

——, sugar in, 269.

—— -sugar, in the dog, 415.

Blowfly larva, reactions to opposed
beams of light, 313.

Burce, W. E. and Nett, A. J. The
comparative rate at which fluores-
cent and non-fluorescent bacteria are
killed by exposure to ultra-violet,
399.

CALCIUM, effect on cardiac and

: respiratory centers, 200.

—— phosphate, relation to osmotic
properties of living cells, 456.

Cardiac center, effect of calcium and
potassium on, 200.

Cartson, A. J. and H. GInspure.
Contributions to the physiology of
thestomach. XXIV. The tonus and
hunger contractions of the stomach
of the new born, 29.

Carison, A. J.. H. Hacer and M. P.
Rogers. Contributions to the phys-
iology of the stomach. XXV. A
note on the chemistry of normal
human gastric juice, 248.

CarTEeR, E. P. See Pearce and Car-
TER, 350.

Cerebrospinal fluid, secretion of, 93.

Chenopodium, influence of on the cir
culation and respiration, 67.

CHILLINGWoRTH, F. P. See HENDER
SON, CHILLINGWORTH and WHITNEY
iI

Circulation, effect of oil of chenopod-
ium on, 67.

Coagulation of blood, 33.

Crite, G.W. See MENTEN and CRriLE,
225.

DAs E. C. Photoelectric currents
in the eye of the fish, 369.
Dead space, respiratory, 1, 20.
Denny, G. P. and G. R. Minor.
origin of antithrombin, 233.
Dog, blood-sugar in, 415.
Duodenum, acidity curves of, 191.

The

FRMMES, L. E. See Benepicr and
EmMEs, 52.

Ergometer, Krogh bicycle, calorimetric
calibration of, 52.

Eye, photoelectric currents in the fish,
369.

Eyster, J. A. E. See Merrx and
EyYstTrEr, 62.

FRAZIER, C. H. and M. M. Pzer.

Influence of diiodotyrosine and

iodothyrine on the secretion of cere-
bro-spinal fluid, 93.

491

492

AINES, W. L. A contribution to
the physiology of lactation, 285.
Gastric juice, chemistry of, 248.
GESSEL, R. The effects of change in
auricular tone and amplitude of au-
ricular systole on ventricular output,
406.

GinsBuRG, H.
GINSBURG, 29.

Glycosuria, Studies in experimental,
415, 425.

See CARLSON and

HIAGER, H. See Caruson, Hacer
and Rogrrs, 248.

Haupane, J. S. The variations in
the effective dead space in breath-
ing, 20.

Harpt, L. L. J. See Rogers and
Harpvt, 274.

Hartman, F. A. The differential ef-
fects of adrenin on splanchnic and
peripheral arteries, 438.

Heart-rate, effect of adrenalin on, 62.

HENDERSON, Y., F. P. CHILLINGWORTH
and J. L. Wuirney. The respiratory
dead space, 1.

Hooker, D. R. The perfusion of the
mammalian medulla: The effect of
calcium and of potassium on the
respiratory and cardiac centers, 200.

Hydrogen ion concentration in blood,
225.

, methods of determining, 180, 186.

Hypophysis, influence on the secretion
of saliva, 339.

KIDNEY, gaseous metabolism of,
350.

LACTATION, physiology of, 285.

Lewis, M.R. Rhythmical contraction
of the skeletal muscle tissue observed
in tissue cultures, 153.

Light, reactions of blowfly larva to,
3138.

INDEX

Livineston, A. E. See Sauant and
LiviINGstTon, 67.

Liver, sugar retaining power of, 425.

Lommen, P. A. See Sotem and Lom-
MEN, 339.

MACLEOD, 3. J. Re» itd Tie
Pearce. Studies in experimental
glycosuria. IX. The level of the
blood-sugar in the dog under labora-
tory conditions, 415.

Studies in experimental glyco-
suria. X. The sugar retaining power
of the liver in relationship to the
amount of glycogen already present
in the organ, 425.

Magnesium phosphate, relation to os-
motic properties of living cells, 456.

Mammalian medulla, perfusion of, 200.

Manometer membrane interpretation
of, as affected by variations in blood
pressure, 209. ;

Martin, E. G. and W. L. MEnpDEN-
HALL. The response of the vasodi-
lator mechanism to weak, interme-
diate, and strong sensory stimula-
tion, 98.

McCuienpon, J. F. Acidity curves in
the stomachs and duodenums of
adults and infants, plotted with the
aid of improved methods of measur-
ing hydrogen ion concentration, 191.

A direct reading potentiometer

for measuring hydrogen ion concen-

trations, 186.

New hydrogen electrodes and

rapid methods of determining hydro-

gen lon concentrations, 180.

The action of anesthetics in pre-

venting increase of cell permeability,

173.

The preservation of the life of
the frog’s egg and the initiation of
development, by increase in perme-
ability, 163.

Meek, W. J. and J. A. E. Eyster.
The effect of adrenalin on the heart-
rate, 62.

INDEX

Metics, E. B. The osmotic properties
of calcium and magnesium phosphate
in relation to those of living cells,
456.

MENDENHALL, W. L. Factors affecting
the coagulation time of blood. VII.
The influence of certain anesthetics,
33.
98.

Menten, M. L. and G. W. CRriLeE.
Studies on the hydrogen ion concen-
tration in blood under various ab-
normal conditions, 225.

Minot, G. R. See Denny and Minor,
233.

Muscle, rhythmical contraction of, in
tissue cultures, 153.

See Martin and MENDENHALL,

NEILL, A. J. See Burce and Nett,
399.

QsMoOTIC properties of living cells,

456.
AIN, nervous paths of, in spinal
cord, 128.
Patten, B. M. An analysis of certain

photic reactions, with reference to
the Weber-Fechner Law. I. The re-
actions of the blowfly larva to op-
posed beams of light, 313.

Pearce, R. G. See Mactreop and
PEARCE, 415.

See Macitrop and PEARCE, 425.

— and E. P. Carter. The influence
of the vagus nerve on the gaseous
metabolism of the kidney, 350

Pret, M. M. See Frazier and PEEt,
93.

Permeability, action of anesthetics in
preventing increase of, 173.

, of frog’s egg, 163.

Photoelectric currents, in the eye of
the fish, 369.

Pitcuer, J. D. An interpretation of
the membrane manometer curves as
affected by variations in blood pres-
sure, 209.

493

Porter, W. T. See Smirx and Por-
TER, 108.
Potassium, effect on cardiac and re-

spiratory centers, 200.

RANSOM, S. W. and C. L. Von Hess.
The conduction within the spinal
cord of the afferent impulses pro-
ducing pain and the vasomotor re-
flexes, 128.

Respiratory center, effect of calcium
and potassium on, 200.

—, dead space, 1, 20.

Respiration, effect of calcium and po-
tassium on, 200.

—., effect of oil of chenopodium on,
67.

Roecers, F. T. and L. L. J. Harpr.
Contributions to the physiology of
the stomach. XXVI. The relation
between the digestion contractions
of the filled, and the hunger contrac-
tions of the ‘‘empty” stomach, 274.

Rogers, M. P. See Carison, HAGER
and Roggrs, 248.

GALANT, W. and A. E. Livineston.

The influence of the oil of cheno-

podium on the circulation and respi-
ration, 67.

Saliva, secretion of, 339.

ScuEar, E.W.E. The content of sugar
in the blood of cats under influence
of cocain, 269.

Smita, G. G. and W. T. PorTER.
Spinal anaesthesia in the cat, 108.
Sotem, G. O. and P. A. Lommen. The
influence of the extract of the pos-
terior lobe of the hypophysis upon

the secretion of saliva, 339.

Stomach, acidity curves of, 191.

—, hunger contractions of, 274.

——, the tonus and hunger contrac-
tions of, 29.

Sugar, content of in the blood of cats
under cocaine, 269.

HYROID, changes in the blood flow
through the gland, 356.

494 INDEX

AGUS, influence of, on the gaseous
metabolism of the kidney, 350.

Vasodilator mechanism, response to
weak, intermediate and strong sen-
sory stimulation, 98.

Vasomotor reflexes, nervous paths of,
in spinal cord, 128.

Ventricle, effect of auricular tone on
output of, 406.

Von Hess, C. L. See Ransom and
Von Hess, 128.

WATTS, C. F. Changes in iodine
content of the thyroid gland fol-
lowing changes in the blood flow
through the gland, 356.
Wuitney, J. L. See HENDERSON,
CHILLINGWORTH and WHITNEY, 1.

6695

—

7 Se

- oi)

“<<

Le

-

- @ :

~
—- a
re ee

a

naa
a
>
—
_
ny =
a .
—
coal

Y Suen

re en Evo

~

err oe

a acer ‘

ai

< pa 4

Tee ee ee Be CO. ee un 906

QP American Journal of Physiology
1

A5

v.38

cop.ea

1G/S-

Biological

& Medica]

Serials

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY

STORAGE

Sot er EP © Bee
me = a

Sh we ee ey

NL Re ome
ee

z eee
Samer = SC on
mae ae oe.

~

Soest
tpl anton yt

oes
oo od

er oem
yet eee
Saree

Aes Syste mee Sen ay Net