o 
 
 COLUMBIA LIBRARIES OFFSITE 
 
 HEALTH SCIENCES STANDARD 
 
 HX641 02386 
 QP101 .Sh6 The eftect of odours 
 
 RECAP 
 
 
Columbia (Mnit)em't|) 
 
 College of l^f^v^itimfi anb ^urgeonsi 
 
Digitized by tine Internet Arciiive 
 
 in 2010 with funding from 
 
 Open Knowledge Commons (for the Medical Heritage Library project) 
 
 http://www.archive.org/details/effectofodoursiOOshie 
 
L iPARTWENT OF PtfYSIOLi 
 
 CoLLEae OF Pmy^ciaws Y«p Suroi 
 
 NEW y^Y. 
 
 THE EFFECT OF ODOURS 
 
 IRRITANT VAPOURS, AND MENTAL WORK 
 
 UPON THE BLOOD FLOW 
 
 DISSERTATION 
 
 Presented to the Board of University Studies of The Johns Hopkins University 
 
 for the Degree of Doctor of Philosophy 
 
 BY 
 
 T. E. SHIELDS 
 1895 
 
 BALTIMORE 
 
 1896 
 
 From 
 
 THE JOURNAL OF EXPERIMENTAL MEDICLNE 
 Vol. I, No. i, 1896 
 
o 
 
 LC 
 
 THE EFFECT OF ODOUKS, lEKITANT VAPOURS, AND 
 MENTAL WORK UPON THE BLOOD FLOW.* 
 
 Plates I-VII. 
 
 By T. E. shields, Ph. D. 
 
 {From the Physiological Laboratory of the Johns Hopkins University.) 
 
 Introduction and Desoeiption of Appakatus. 
 
 Attempts have been made from time to time to determine the 
 efiects produced by various sensations on the circulation of the 
 blood. In 1877 a series of experiments was conducted by MM. 
 Gouty and Charpentier, under the direction of M. Vulpian,f in the 
 hope of obtaining a more exact determination of these disturbances. 
 They used curarized dogs with artificial respiration and recorded the 
 results on a kymograph in the usual manner by means of a mercury 
 manometer in connection with an artery. On stimulating the various 
 organs of special sense they obtained cardio-vascular reactions simi- 
 lar in many respects to those usually obtained by stimulating sensory 
 nerves, but ditfering markedly from these in their greater variability 
 in both form and intensity under the same stimuli. 
 
 Both pneumogastric nerves were then severed. Sensory stimula- 
 tion continued to produce its usual eifect on arterial tension, but 
 ceased to produce any effect on the heart. This convinced them of 
 the fact that the heart disturbances and vasomotor changes were 
 mediated by distinct mechanisms. 
 
 The cerebral cortex of fresh subjects was then shut out by pres- 
 sure, by injection, and by light doses of chloral hydrate. In every 
 case stimulation of the organs of sense failed to produce any effect, 
 either on the heart or on arterial pressure, while the direct stimula- 
 
 * A thesis submitted to the Johns Hopkins University for the degree of Ph. D. 
 f De I'influence des excitations des organs des sens sur le coeur et sur les vaisseaux 
 (recherches exp. faites dans le lab. de M. Vulpian), note de MM. Couty et Charpentier, pre- 
 sentee par M. Vulpian. Comptes Eendus, t. Ixxxv, No. 3, p. 161. 
 1 
 
 CoPTRiGHT, 1S96, Br D Appleton and Compakt. 
 
2 Efect of Odours, Irritant Vapours, etc., upon the Blood Flow 
 
 tion of a peripheral nerve, such as tlie sciatic, continued to produce 
 its usual effect. From these facts they concluded that the cardio- 
 vascular disturbances accompanying sensory stimulation are not due 
 directly to sensation itself, but to a contingent cerebral activity which 
 they called emotion. 
 
 Experiments of this nature, however valuable in themselves, will 
 hardly be regarded as sufficient data from which to draw conclusions 
 as to the nature and extent of the cardio-vascular changes accom- 
 panying sensations in man. Even if it be granted that cerebral ac- 
 tivity is the immediate cause of the cardio-vascular changes in both 
 man and dog, it does not follow that these changes will be the same 
 in the two cases. 
 
 In 1880 Dogiel * attempted, by means of the plethysmograph, to 
 study the effects of music on the circulation of the blood in man. 
 The apparatus used contained so many sources of error that even 
 the meagre results obtained can scarcely be regarded as reliable. It 
 is to Mosso's t researches that we are chiefly indebted for what definite 
 information we possess concerning the correlation of cerebral func- 
 tion and cardio-vascular changes in man, but our literature on the 
 whole subject of the changes produced in the circulation in response 
 to various sensations is still very meagre. This is particularly true 
 of olfactory sensations. The present research was undertaken in 
 the hope of throwing some light on this problem. 
 
 The form of plethysmograph devised by Mosso was first used, 
 and the volume changes in the arm recorded on a drum kymograph ; 
 but it soon became evident that several sources of error would have 
 to be eliminated before accurate measurements could be obtained. 
 The possibility of moving the arm into or out of the cylinder during 
 the experiment must be removed. It is not sufficient to request the 
 subject to sit quietly, nor even to support his head and arm, as Dogiel % 
 ' did. The voluntary effort required to sit immovable is itself a dis- 
 turbing factor in the circulation of the blood. Besides, it is impos- 
 
 * Archivfur Anat. und Physiol, Physiol Abthlg., 1880. S. 420 ff. 
 
 f Kreislauf des Blutes im menschlichen Gehiru. Von A. Mosso, Leipzig, 1881. 
 
 t Op. cit., pp. 419 et 429. 
 
T. E. Shields 3 
 
 sible to prevent movement of the arm for any considerable time, 
 since normal respiration moves the arm. Mosso's * plan of suspend- 
 ing the cylinder from the ceiling and allowing itjto follow the move- 
 ment of the arm is much better, but even here the inertia of the 
 cylinder filled with water offers sufficient resistance to the movements 
 of the arm to produce quite a perceptible error. I think it probable 
 that the respiratory waves f often obtained by the plethysmograph 
 are due in greater measure to this movement of the arm than to the 
 changes in its volume of blood. During the early part of the pres- 
 ent research I frequently obtained marked respiratory waves, but in 
 every case I was able to trace them to this movement of the arm, 
 and when the movement was prevented the respiratory waves were 
 very much diminished if not entirely suppressed. 
 
 A second source of error lies in the fact that every increase of 
 pressure in the cylinder tends to force the arm out of it, and every 
 diminution of pressure exerts an opposite influence. The play of 
 these opposing forces is felt very distinctly on the inclosed arm dur- 
 ing each pulse beat, and when the resulting movement of the arm is 
 prevented the record of the pulse wave is very much amplified. 
 This source of error is still more marked in sudden large vasomotor 
 changes, nor can it be eliminated by making the rubber sleeve so 
 tight that it will not slip on the arm ; for apart from the two obvious 
 objections to this mode of procedure— that the arm should be oiled :j: 
 to prevent it from absorbing water, and that a tight sleeve interferes 
 with normal circulation— the elastic sleeve itself lengthens and short- 
 ens in response to changing pressure in the cylinder. 
 
 There is still another source of error connected with this mode of 
 closing the cylinder which remains even after all movements of the 
 arm into or out of the cylinder have been prevented. If the arm 
 does not tit the mouth of the cylinder snugly the portion of the elas- 
 tic sleeve between the arm and the cylinder will yield to each change 
 of pressure in the cylinder. The panting of this membrane in Te- 
 sponse to the pulse may be seen by the unaided eye. If the mouth 
 
 Op. cit., p. 44. X The oil, of course, increases very much the tendency to slip 
 
 •f Dogiel, op. cit, p. 130. 
 
4 Lfect of Odours, Irritant Vapours, etc., upon the Blood Flow 
 
 of tlie cylinder be large enough to allow the elbow to enter, it will 
 usually be found much too large to fit the upper arm. This was 
 very conspicuously the ease in the cylinder used by Dogiel * If the 
 cylinder terminates over the large muscles of the forearm, as it does 
 in Mo^so's arrangement, it can be made to fit the arm snugly, but 
 another source of error appears. Every movement of the inclosed 
 fingers withdraws muscle from the cylinder or introduces muscle 
 into it. 
 
 /////////////y//y/////////^^^^^ 
 
 y//////////////////y///y//yy/y///^/// y/yyyyy''yyy/y//////////M. 
 
 Fig. 1. S, rubber sleeve ; R, ring cemented into end of rubber sleeve ; A, arm ring ; G, 
 
 shoulder in same ; II, bevelled face of arm ring supporting rubber sleeve ; C, D, rings mak- 
 ing arm holder rigid to cylinder; X N', rods connecting the rings C and D; W, wristlet ; 
 B, band for same ; M, e, I, smaller openings in arm cylinder ; /t, thermometer ; t, stopcock ; 
 M M', metal rods, making wristlet rigid to arm rmg. 
 
 Fig. 1 is a sketch of a device for removing these sources of error 
 The same is shown in Plate II. The lettering is the same in both. 
 
 One end of a suitable-sized rul)ber sleeve, S, is drawn over the 
 arm cylinder and fastened in position by a ligature. A metal ring, 
 R, large enough to pass readily over the flange is cemented into the 
 other end of the sleeve. By the aid of this ring the sleeve is easily 
 doubled back over the flange of the cylinder, where it remains until 
 the arm has been put into position in the cylinder. A hard-rubber 
 wristlet, provided with a hinge and clasp, tits round the wrist. A 
 soft band, B, passes from one side of this wristlet between the thumb 
 and index finger and is buckled to the other side of the wristlet as 
 tightly as comfort will permit. A second hard-rubber ring, A, pro- 
 vided witli hinge and clasp, fits the arm above the elbow. This ring 
 is rigidly connected with the wristlet by two metal rode, M, M'. A 
 
 * Loc. ci\ 
 
T. E. Shields 5 
 
 shoulder, G, is cut into one face of this ring so as to make it fit into 
 and against the mouth of the cylinder. The face, H, of the ring is 
 shaped so as to keep the rubber sleeve taut down to the point where 
 it meets the arm. "When the arm has been introduced into the cylin- 
 der and the arm ring fitted info its place, the rubber sleeve is drawn 
 down over it. Two hard-rubber rings, C and D, rigidly connected 
 by two metal rods, N, N', and provided with hinges and clasps, are 
 then fastened outside the rubber sleeve. One of these rings, C, rests 
 against the flange of the cylinder; the other, D, fits tightly against 
 the outer surface of the arm ring. 
 
 This arm holder reduces toa minimum, if it does not entirely re- 
 move, the sources of error discussed above. All movements of the 
 elastic sleeve are rendered impossible. The arm holder is made ab- 
 solutely rigid to the cylinder. The wristlet, with its tight band be- 
 tween thumb and index finger, prevents the arm from moving in 
 the holder. As the inclosed arm can not bend, no muscle can be 
 withdrawn from the cylinder or introduced into it. 
 
 The arm was now found to change in volume about one cubic cen- 
 timetre at each pulse beat. This introduced a new difficulty in the 
 recording apparatus. It will be remembered that in lilosso's^ ple- 
 thysmograph the volume changes were recorded by means of a test 
 tube suspended in a beaker of water or some liquid of suitable spe- 
 cific gravity. The test tube was counterpoised by a weight carrying 
 the recording needle and connected with the test tube by two threads 
 passing over a pulley. However well this arrangement may serve 
 for recording slow changes, its inadequacy for recording such 
 changes as we are here dealing with is evident. The introduction 
 and withdrawal from the test tube of a cubic centimetre of water fifty 
 to one hundred times a minute sets up considerable agitation in the 
 water in the beaker, where every motion is a source of erroi*. Be- 
 sides, the system has a period of its own which sometimes neutral- 
 ises the pulse effects and at other times combines with them to pro- 
 duce greatly exaggerated excursions of the test tube, as happens in 
 
 * MM. A. Mosso et P. Pellacani. Sur les fonctions de la vessie. Archives italiennes 
 de biologie, vol. i, p. 08. 
 
6 Effect of Odours^ Irritant Vapours, etc., upon the Blood Flow 
 
 the production of tone beats. The great inertia of the system also 
 prevents it from recording accurately such rapid changes in volume. 
 
 The method adopted by Bowditch * of suspending the test tube 
 from a spiral spring is more convenient. This removes the error 
 which, in Mosso's plan, arises from the motion of the water in the 
 beaker. It also materially lessens the inertia of the system. But 
 the great inertia still remaining, and especially the periodicity 
 of the spring, compelled me to seek some other method of record- 
 ing. I had recourse to a water manometer of large bore con- 
 nected with the cylinder by a wide, indistensible tube. A light and 
 buoyant float f was used to carry the recording needle. This sys- 
 tem, being rigid, and having no period of its own except that due 
 to the float, gave a fairly accurate j'ecord of the changes occurring 
 in the volume of the arm; but it introduced a new source of error 
 which compelled me to abandon its use. The level of the water in 
 the manometer tube determines the pressure which the water in the 
 cylinder exerts on the arm. At the beginning of the experiment 
 the manometer may be set so as to exert the proper pressure. But 
 in the course of the experiment every change in the volume of the 
 arm will cause a change of pressure on its surface. The arm some- 
 times changes twenty or thirty cubic centimetres in volume in a 
 few seconds. This, of course, causes a corresponding change in the 
 level of the water in the manometer tube, wliich in turn alters very 
 considerably the pressure on the arm ; nor can this factor be neg- 
 lected. I found that the change of pressure due to an alteration of 
 a few centimetres in the level of the water in the manometer caused 
 a very large change in the volume of blood in the arm. That is 
 what we would expect from the low blood pressure in the capillaries 
 and veins. 
 
 Lehmann ;}: gives the following account of the apparatus used by 
 him in investigating the effects of sensory stimulation. I quote 
 
 * Proceedings of the American Academy, May 14, 1879. 
 
 \ Howell and Warfield. Studies from Biological Laboratory, Johns Hopkins Umversiiy, 
 vol. ii, p. 235. 
 
 \ Die Hauptgesetze des menschlichen Gefiihlsleben. Von Dr. Alfr. Lehmann, Leipzig, 
 1892. 
 
T. E. Shields ' Y 
 
 his description in full, as I shall have occasion to refer to the results 
 obtained later on. 
 
 " Wir haben nun die Aufgabe auf die Bestimmung der Atembe- 
 wegungen, des Herzschlages und des Yoluraens eines einzelnen 
 Gliedmasses, z. B. eines Armes, reduciert. Die erste Bestimmung 
 ist mittels des Pneumographen leicht aiiszufiihren, wahrend die 
 beiden letzten sich mittels Mosso's Plethysmographen zusammen 
 ausfiihren lassen. Dieser besteht aus einem am einen Ende ge- 
 schlossenen Eohr, das eben weit genug ist, den Arm zu umschliessen, 
 der dnrch einen Gummiarmel wasserdicht mit dem Rohr verbunden 
 wird. Durch ein Seitenrohr wird das ganze Kohr, nachdem der 
 Arm in die rechte Stellung gebracht ist, mit Wasser gefiillt, und 
 hierauf wird das Seitenrohr durch einen Gummischlauch mit einem 
 Mareyschen Schreibapparat {tambour enregisteur) in Yerbindung 
 gebracht. Jede Yeranderung des Armvolumens bewirkt eine 
 Hebuno; oder Senhuno; des Wassers, durch welche die Luft im 
 Schlauch und im Schreibapparat beeinflusst wird, und die Bewegnng 
 wird dann vom Stift des Schreibapparats auf eine rotierende Walze 
 (das Kymographion) gezeichnet." 
 
 Lehmann's apparatus seems to be a Mosso hydros phygmograph * 
 in which the bottle tor keeping the pressure constant has been 
 omitted so as to compel the tambour to register both pulse and vaso- 
 motor eifects. In both these forms of apparatus the pressure on the 
 arm is too high. The air cushion is at the top of the cylinder. 
 The whole arm is consequently under a positive pressure of some 
 centimetres of water. The under surface of the arm can scarcely be 
 under a pressure of less than twelve centimetres. If the cylinder be 
 large enough to extend above the elbow, as it should do in order to 
 escape a source of error mentioned above, this pressure is likely to 
 be still higher. 
 
 The form used by Lehmann introduces another grave source of 
 error. In dispensing with the pressure bottle, the tambour and 
 cylinder are rendered a closed system in which every change of 
 volume in the arm alters the pressure on its surface. If, as some- 
 
 * Diagnostik des Pulses. Von Dr. A. Mosso, Leipzig, 18Y9. 
 
8 Effect of Odours^ Irritant Vapours, etc., upon the Blood Flow 
 
 times happens, twenty-five or thirty cubic centimetres of blood be 
 added to the volume of the arm, the distention of the membrane of 
 the tambour necessary to make room for this increased volume of 
 blood would exert a considerable pressure on the arm. 
 
 Besides the sources of error enumerated in connection with each 
 of the forms of apparatus discussed above, there is a defect common 
 to them all. We have seen that the changes in volume of blood in 
 the arm are due to at least two distinct sources — changes in the 
 heart's action and changes in the calibre of peripheral vessels. It is 
 therefore highly desirable that the record of these two effects be 
 kept as distinct as possible. Whenever, as is the case with each of 
 these forms of apparatus, the same writing point is made to record 
 the pulse and the gross volume changes, these effects partially mask 
 each other and do not stand out as distinctly as they should. This 
 is illustrated very well by many of the curves published in Leh- 
 mann's work. I have a large number of similar curves in my own 
 possession which I obtained by the use of these forms of apparatus. 
 
 From what has been said, it is evident that an efficient apparatus 
 for recording the volume changes in the arm must at least meet the 
 following three requirements: First, the pulse and vasomotor effects 
 should be recorded separately ; second, the apparatus for recording 
 the rapid and rhythmic changes of volume due to the pulse should 
 have very little inertia and no periodicity; tliird, the pressure ex- 
 erted on the arm should be as nearly normal as possible, and must 
 remain constant during all changes of volume which may occur in 
 the arm. These requirements are met fairly well by the following 
 arrangement : 
 
 A long metal tube. A' (Plate I), is slipped over an upright, U, 
 to which it is fastened by a clamp, C. The top of the upright is 
 held in position by two metal rods, M, M', which make it rigid to 
 the wooden screen interposed between the subject and the kymo- 
 graph. A short metal tube, B, is supported in a vertical position by 
 two clamps, D, D', which make it rigid to the upper end of tube A. 
 One end of a long close spiral, S, of l^o. 5 piano wire is slipped over 
 the lower end of the tube, B, and fastened to it by a clamp, E. Two 
 
T. E. Shields 9 
 
 wires, F, F, pass down from the free end of the spiral, and are fas- 
 tened by two small binding screws to a light hard-rubber ring, K, 
 one inch in diameter. By suitable cork collars, test tubes of various 
 sizes may be suspended from this ring. The slipping of the wire 
 under clamp E permits the length of spiral to be so adjusted that 
 the level of the water in the suspended test tube will remain constant 
 while the quantity of water in the test tube varies. One arm of a 
 long glass siphon, G, is passed down through tube B, thence 
 through the centre of the spiral S, and dipped a few millimetres 
 beneath the surface of the water in the test tube T. The siphon is 
 centred and held in position by six. set screws passing through tube 
 B. An indistensible tube, H I, connects the other end of the siphon 
 with the arm cylinder. The four-way J interposed in this tube will 
 be described later. 
 
 This is essentially the same as the apparatus described by Bow- 
 ditch, to which reference has already been made. Its inertia unfits 
 it for recording the changes in volume due to the pulse. Its perio- 
 dicity combining with the rhythmic pulse at times neutralizes the 
 pulse effects, and at other times combines with them to produce 
 greatly exaggerated excursions of the test tube. The pulse and vaso- 
 motor effects also partially mask each other. The pulse effects are 
 the disturbing factor in each case ; if they be removed from the 
 test tube and recorded separately, the instrument is well adapted for 
 registering the vasomotor effects. 
 
 By the following device this separation of the pulse record from 
 the vasomotor record may be attained, the pulse wave being recorded 
 by a tambour the tension of which remains practically constant ex- 
 cept for the variation with each pulse wave and the vasomotor 
 changes being recorded by the test tube and spiral spring already 
 described, which are so arranged that only the slow vasomotor 
 changes in the volume of the arm cause any perceptible movement 
 of the system, the quicker changes due to the heart heat disappear- 
 ing from this record. A wide glass piston tube (K, Fig. 2 and Plate 
 I ; P, Plate II) is supported in a vertical position beside the arm 
 cylinder, with the bottom of which it communicates freely. The 
 
10 Efect of Odours, Irritant Vapours, etc., xipon the Blood Flow 
 
 TO TAMBOUR 
 
 Fig. 2. — K, piston tube ; * P, piston ; N, piston rod ; Q, milled nut for adjusting level 
 of piston ; Q', milled nut for adjusting indicator ; (/, metal cap of piston tube ; W, washer, 
 giving support to nut Q in lowering piston ; 0, 0', rods supporting washer ; v, tube connect- 
 ing with tambour ; Y, T-way ; A, B, stopcocks on same ; X, tambour ; e, tambour lever carry- 
 ing horizontal pivot ; k, horizontal pivot ; /t, vertical pivot. 
 
 * The vertical piston tube allows the pressure of the water on the arm to be regulated 
 at pleasure. The piston in this tube serves to keep the volume of the air between the 
 water and the tambour constant. 
 
T. E. Shields 11 
 
 upper end of the tube is incased in a metal cap, L (Fig. 2), which is 
 perforated to allow free play to the hollow piston rod, N, and free 
 communication between the air in the upper end of the piston tube 
 and the atmosphere. From this cap two metal rods, O, O', parallel 
 with the piston tube extend upward about one inch and hold in 
 place a washer, W, which encircles the piston rod. Between the 
 washer and the cap a milled nut, Q, fits on the piston rod, whi:h is 
 threaded from end to end. The piston rod, 'N, is a metal tube one 
 eighth of an inch in diameter, which extends through an otherwise 
 air-tight piston, P, in the piston tube. A small indistensible tube, Y, 
 connects the upper end of the piston rod with a Marey tambour, X. 
 A T-waj, Y, provided with two stopcocks, A, B, interposed in the 
 course of this tube, permits connection between the air cushion be- 
 low the piston and the tambour or the atmosphere to be made or 
 broken as desired. A milled nut, Q', fitting the piston rod above the 
 washer, carries an indicator, which passes down through a perfora- 
 tion in the washer, thence through a perforation in the flange of the 
 metal cap, and terminates below the level of the piston. The dis- 
 tance between the point of the indicator and the lower surface of 
 the piston measures the vertical height of the air cushion in the pis- 
 ton tube. It can be set at any desired distance by means of the nut 
 Q'. Since the indicator is carried by the piston rod, the distance 
 between its point and the piston remains constant whatever altera- 
 tions may be made in the level of the piston. The level of the water 
 in the piston tube when the air cushion is in communication with 
 the atmosphere indicates the pressure exerted by the water on the 
 surface of the inclosed arm. For convenience in registering this 
 pressure a centimetre scale is affixed to the side of the piston tube. 
 
 The arm cylinder is of the usual form (see Plate II). A flexible 
 metal band, Y, encircles the cylinder near one end and is closed over 
 the upper surface of the cylinder by a nut and bolt. To this bolt 
 are also fastened two chains which pass round the cylinder in oppo- 
 site directions, and thence through a ring, d, which is suspended some 
 distance above the cylinder. The chains are doubled back and 
 hooked on themselves at desired lengths. The other end of the 
 
12 Effect of Odovrs^ Irritant Vapours^ etc., upon the Blood Elow 
 
 cylinder is suspended in a similar manner from the same rin<r. 
 Tiiese chains serve to keep the cyh'nder in a horizontal position and 
 to prevent it from rotating. A litth chain, from which the ring d 
 is suspended, passes through a pulley fastened to the ceiling, thence 
 down to a ratchet which is supported at convenient height. By 
 turning this ratchet the cylinder is raised or lowered to suit the com- 
 fort of the subject. 
 
 Besides the opening through which the arm is introduced, the 
 cylinder is provided with three other smaller orifices. Through one 
 of these, Z, situated on the end of the cylinder, communication is 
 made with the suspended test tube used to record the vasomotor 
 changes. A second orifice, e, provided with a tight-fitting rubber 
 stopper through which two perforations have been made, is located 
 on the upper surface of the cylinder. A glass stopcock, ^, is fitted 
 into one of these perforations. The bulb of a thermometer, h, is in- 
 troduced through the other. On the under surface of the cylinder 
 is a third orifice, M, into which is fitted one branch of a large T-way. 
 A second branch of this T-way communicates with the lower end of 
 the piston tube. The third branch is connected with the tube k for 
 filling and emptying the cylinder. 
 
 1 have found the following method convenient for filling and 
 emptying the cylinder, for regulating the temperature and pressure 
 of the water, and for adjusting the instrument: 
 
 The piston is raised or lowered by turning the milled nut Q until 
 the point of the indicator stands at the level to which it is desired to 
 bring the water in the piston tube. Two supply jars, P, Q (Plate I) 
 — one filled with hot water, the other with cold water — are placed 
 on a shelf attached to the screen above the level of the test-tube 
 siphon, G. Two siphons connect these jars with two branches of a 
 large three-way stopcock, w, which permits the hot and cold water to 
 be mixed in the desired proportions. The bulb of a thermometer 
 for indicating the temperature of the water is placed in the tube, 
 which conducts the water from the three-way w to one branch of a 
 second large three-way, 2. A tube connects a second branch of this 
 three-way with the arm cylinder ; another tube leads from its third 
 
T. E. Shields 13 
 
 branch to the sink. This three-way allows the water to pass from 
 the three-way w to the cylinder, or from the cylinder to the sink, or 
 from a hydrant tap to the supply jars, as may be desired. 
 
 The four-way J, each branch of which is provided with a stop- 
 cock, is made rigid to the tube A, at a level somewhat lower than 
 that of the arm cylinder. One of these stopcocks (I) opens into the 
 test-tube siphon ; a second (2) communicates with the arm cylinder; 
 a third (3) is connected by a siphon with one of the supply jars; a 
 tube from the fourth (4) leads to the sink. 
 
 When the subject's arm has been coated with vaseline and placed 
 in the arm cylinder, as directed above, stopcock t, on the upper sur- 
 face of the arm cylinder, is opened to allow the air to escape. Stop- 
 cocks A and B (Plate I and also Fig. 2) are closed to cut off connec- 
 tion between the air cushion of the piston tube and the atmosphere 
 and tambour. Stopcocks 1 and 2, on the four-way, are opened, and 
 stopcocks 3 and 4 closed. Three-way stopcocks w and s are ad- 
 justed so as to allow water of the proper temperature to enter the 
 arm cylinder. When the cylinder is within a few cubic centimetres 
 of being full the supply of water from this source is shut off, and a 
 strong clamp placed on the supply tube at the point where it meets 
 the cylinder, to prevent any possible error that might otherwise 
 arise from leakage or from distention of the supply tube. Stopcock 
 3, on the four-way, is now opened and water allowed to flow into the 
 cylinder until it is filled, when the stopcock at the top of the cylin- 
 der {t, Plate II) is closed to prevent air from re-entering the arm 
 cylinder. Stopcock 2, on the four-way, is closed at the same moment 
 to prevent increase of pressure on the arm while the test-tube siphon 
 is being filled. When the test tube is partially filled, stopcock 3 is 
 closed and stopcock 2 opened. If the test tube is too full at any 
 time, water may be drawn off" by opening stopcock 4. Stopcock B 
 is now opened, thus establishing free communication between the 
 air cushion and the atmosphere. The water in the piston tube will 
 now rise to the same level as that in the test tube. Tube A', carry- 
 ing the test tube and its appurtenances, is slowly raised or lowered 
 on the upright until the level of the water in the piston tube reaches 
 
14 Effect of Odours^ Irritant Vapours^ etc.^ upon the Blood Flow 
 
 the point of the indicator. Stopcock B is tlien closed and stopcock 
 A opened, thus putting the air cushion in communication with the 
 tambour. The lever of" the tambour will now give a magnified rec- 
 ord of the changes in the level of the water in the piston tube. If 
 the resistance offered by the long narrow tube, H I, to the rapid pas- 
 sage of water between the test tube and the arm cj'linder is not suf- 
 ficient to prevent the changes of volume due to the pulse from mov- 
 ing the test tube, this resistance is increased to the desired extent by 
 partially closing stopcocks 1 and 2 on the four-way. The small and 
 rapid changes of volume due to the pulse are consequently recorded 
 by the tambour alone, while the slower and larger volume changes, 
 freed from pulse waves, are recorded by the motions of the test tube, 
 on the same principle that the height of the tide is recorded by the 
 level of the water in the vertical pipe of the tide gauge, while the 
 separate waves are prevented from influencing this level by the 
 resistance which the holes in the bottom of the pipe offer to the rapid 
 passage of water. 
 
 This form of apparatus reduces very much the sources of error 
 discussed above. The suspended test tube is entirely freed from the 
 error due to the periodicity of the spring. The inertia of this system 
 ceases to be a grave source of error owing to the slowness of the 
 motions which it is now required to execute. The tambour, relieved 
 for the most part from the disturbing factor of large volume changes, 
 gives a fairly accurate record of the pulse waves. The chief source 
 of error remaining is the pressure exerted by the water on the arm. 
 It is, of course, impossible to have an arm immersed in water under 
 perfectly normal pressure. I have found the conditions to be most 
 favourable when the level of the water in the piston tube is that of 
 the middle of the arm, the lower half of the arm being then under a 
 positive pressure of a few centimetres of water, and the upper half 
 under a like amount of negative pressure. 
 
 Each pulse wave is accompanied by a slight change of pressure 
 on the arm, due to the resistance which the water in the cylinder 
 offers to the entrance of the wave. This resistance is made up of 
 several factors: First, the inertia of the volume of water to be 
 
T. E. Shields 15 
 
 moved ; second, the force of gravity to be overcome in slightly- 
 raising the level of the water in the piston tube ; and third, the re- 
 sistance offered by the air cushion in the piston tube to change in the 
 level of the water in the tube. The last-named resistance is due to the 
 tension of the membrane of the tambour and to the inertia and friction 
 of the recording lever. These sources of error are quite small, but 
 they still exist, and should be taken into account in calculating results. 
 The resistance offered by the air cushion has been rendered very 
 slight by reduction in the weight and friction of the recording lever, 
 and by the delicacy and low tension of the tambour membrane. The 
 inertia of the mass of water to be moved may be reduced somewhat 
 by so changing the form of the arm cylinder as to diminish the dis- 
 tance between the surface of the water in the piston tube and the 
 point where the expansion of the inclosed arm first occurs.. The 
 wider the piston tube, the less the change in the level of the water 
 caused by the pulse wave ; but the pulse recorder becomes propor- 
 tionally more sensitive to vasomotor changes. When there is a 
 sudden large change in the volume of the arm, the vasomotor 
 recorder can not take care of it rapidly enough. There is, in conse- 
 quence, a temporary variation in the level of the water in the piston 
 tube, which interferes somewhat with the pulse recorder and changes 
 the pressure on the arm. The influence of this disturbing factor, 
 which I have not yet been able entirely to eliminate, will receive 
 attention when we come to discuss the results. 
 
 In the early part of the investigation, I used a Ludwig drum 
 kymograph, with smoked paper for recording the curves. It soon 
 became evident, however, that this plan of recording presented 
 another grave difiiculty. 
 
 Spontaneous changes, or at least changes not due to sensory stim- 
 ulation, were continually appearing in both vasomotor and pulse 
 curves. It was, therefore, necessary to have a continuous record 
 of the changes going on in the arm for some time before and after 
 stimulation, in order to be able to differentiate them from any sen- 
 sory reactions that might appear. This is impossible with the 
 Ludwig drum kj-mograph, for if it be made to travel fast enough to 
 
16 Effect of Odours, Irritant Vapours, etc., upon tlie Blood Flow 
 
 bring out the pulse curve, the paper will be used up in three or four 
 minutes, and it will be necessary to stop and put on fresh paper. 
 An ink record on a continuous band of paper is evidently what is 
 needed. The chief difficulty in obtaining such a record is the 
 resistance to be overcome in moving the vasomotor and pulse pens. 
 I was unable to reduce this sufficiently in the case of vertical kymo- 
 graphs for continuous paper, and accordingly built a horizontal 
 kymograph, the general arrangement of which is shown in Plate I. 
 By special contrivances for suspending and moving the pens the 
 resistance was reduced within tlie limits of that offered by the smoked 
 paper. 
 
 A light plate of aluminium foil is cut in the form of a cross. The 
 two ends of the transverse beam widen out and are bent so as to 
 clasp tightly the tube of a glass pen. The ends of the other beam of 
 the cross are bent into hooks, from one of which a thread passes 
 over a pulley adjusted to the side of the kymograph, and is held 
 taut by a small weight. From the hook on the other end of the 
 beam a thread passes under a pulley adjusted to the other side of the 
 kymograph, and thence vertically to a violin key, which turns in a 
 cork holder cemented to the bottom of the test tube used for record- 
 ing the vasomotor changes. This key is turned until the pen rests 
 over the centre of the paper. A thread fastened to the bulb of the 
 pen passes vertically to a similar violin key, wliich is supported in 
 proper position. The tube of the pen, at a point somewhat beyond 
 where it leaves the holder, is bent downward through an angle of 
 about 45°. The key to which the vertical thread supporting the 
 bulb of the pen is fastened is turned until the point of the pen just 
 touches the paper. The pulleys are so adjusted that the upper por- 
 tion of the pen will be about horizontal. The thread passing be- 
 tween the pulleys is horizontal also, except for the sag in the middle, 
 caused by the weight of the pen. As the pen moves from either 
 pulley toward the other it describes a curve, due to the sag in the 
 horizontal thread. This curvilinear motion is neutralized in the 
 point of the pen by the pendular motion of the bulb, due to the 
 thread from which it is suspended, the factors in this adjustment 
 
T. E. Shields lY 
 
 being the extent of the sag, the length of the pendular thread, and 
 the respective distances of the point and bulb of the pen from the 
 horizontal thread. The extent of the sag is controlled by the size of 
 the weight on the end of the thread. The length of the pendular 
 thread is regulated bj raising or lowering the key holder from which 
 it is suspended. The distance of the ends of the pen from the hori- 
 zontal thread is controlled by moving the pen in its holder. 
 
 The friction between the point of this pen and the paper is very 
 small, since the point need not actually touch the paper, the capil- 
 lary attraction between the paper and the meniscus being sufficient 
 to cause the ink to flow if the pen be properly made. But even 
 when the point of the pen actually touches the paper, it does so very 
 lightly. If the point of the pen be lifted by a finger until the sag 
 leaves the horizontal thread, half of the weight of the tube of the 
 pen will rest on the finger. If the finger be now gradually lowered, 
 the weight on it will be found progressively to diminish and finally 
 to disappear, having been thus shifted to the horizontal thread. 
 This gives very great buoyancy to the point of the pen and reduces 
 friction to a minimum. 
 
 The pulleys are large and very light, and are pivoted in a manner 
 similar to the balance wheel of a watch. Their friction is thus 
 rendered very small. The inertia due to the pen, its aluminium 
 holder, and the light weight on the end of the horizontal thread, is 
 relatively a very small addition to that of the test tube partially 
 filled with water; but the inertia of this whole system, large as it is, 
 has very little influence on the accuracy of the record, owing to the 
 slowness of the motions which it describes. 
 
 The tambour, x (Fig. 2 and Plate I), of the pulse recorder is sup- 
 ported with its membrane vertical. The motions of this membrane 
 are communicated to one end of a light aluminium lever, e, which 
 just after passing through a pivot, h, whose axis is vertical' bends 
 upward at a right angle, and after running in this direction about 
 one inch supports a second pivot, h, whose axis is horizontal and 
 perpendicular to the long axis of the paper. A light glass pen, I, 
 with a long tube is passed through this pivot and fastened to it when' 
 
18 Effect of Odours^ Irritant Vapours, etc., upon the Blood Flow 
 
 the bulb of the pen is directly over the axis of the first pivot. The 
 motion of the membrane will now impart a magnified horizontal 
 movement to the point of the pen, while its bulb, being in the axis 
 of the pivot, only undergoes a very slight rotary motion. The sec- 
 ond pivot allows the pen to njove freely in a vertical plane. Ink is 
 now poured into the bulb of the pen until the point of the pen 
 barely rests on the paper. The friction here is very slight. 
 
 The pens for the pneumograph (to record the respiratory move- 
 ments), the time marker, and the signal marking the time of applica- 
 tion of the stimulus, are pivoted in a slightly different manner, but 
 one which involves the same principle. The bulb of the pen with 
 the ink which it contains acts in each case as a counterpoise to the 
 long tube of the pen, thus causing the point to rest on the paper as 
 lightly as desired. A horizontal movement is imparted to the point 
 of the pen by a fork of the tambour lever or of an armature of an 
 electro-magnet, as the case may be. For this purpose the pen must 
 be free to move round two axes, one vertical and the other hori- 
 zontal. 
 
 The points of the five pens are adjusted so as to fall in a line per- 
 pendicular to the long axis of the paper. The test tube used in the 
 vasomotor recorder was of such a calibre that three cubic centimetres 
 of water moved it one centimetre. The paper travelled from right to 
 left; the time record is nearest the test tube, and consequently at the 
 bottom of the paper. Movement of vasomotor or pulse pen toward 
 the bottom of the paper represents diminishing volume of arm, and 
 conversely. I used the time record as base line in calculating the 
 volume changes. If 6? represents the initial distance in centimetres 
 between the point of the vasomotor pen and the base line, and d' its 
 distance in any subsequent position, then (tZ' — oT) X 3 = F", where F" is 
 the volume change in cubic centimetres. If at any time during the 
 experiment the test tube is becoming overfilled, stopcock 2, Plate I, 
 on the four-way, is closed and stopcock 4 opened until the desired 
 amount is drained off. Stopcock 4 is then closed and stopcock 2 
 opened, and the fact noted on the paper. The point of the pen, of 
 coarse, records the volume drawn out, which must be added in cal-' 
 
T. E. Shields 19 
 
 cnlating subsequent changes. If the test tube is in danger of being 
 emptied, water is introduced through stopcock 3 in a similar man- 
 ner, and the amount subtracted in all subsequent calculations. 
 Specimens of the curves traced by this instrument may be seen in 
 Plate in. 
 
 Being thus enabled to obtain what I believed to be a fairly accu- 
 rate record of the changes which actually occur in the volume of 
 the arm, I proceeded to study the influences which determine these 
 changes in the hope of being able to eliminate some of them and to 
 differentiate the changes which would remain from any odour reac- 
 tions I might obtain. 
 
 I first determined empirically what temperature of water and 
 what pressure on the arm would maintain the most constant volume 
 of blood in the arm. The most favourable pressure, as has already 
 been stated, is that which is exerted when the level of the water in 
 the piston tube was about that of the middle of the arm. When I 
 used pressure lower or higher than this the arm exhibited a tendency 
 to increase or diminish in volume. The proper temperature of water 
 I found to be between 30° and 34° C, according to the temperature 
 of the room and of the arm of the subject. I found that this tem- 
 perature kept the volume of the arm more constant than any other, 
 and I found also that the temperature of the water in the cylinder 
 remained constant throughout the experiment, whereas, if the tem- 
 perature be higher at the beginning, it will sink during the experi- 
 ment, and conversely, if the initial temperature be lower. 
 
 After the experiment had been continued for about half an hour, 
 large and frequent changes of volume usually occurred, which I at- 
 tributed to the discomfort of the subject, who up to this time had 
 been seated on a wooden chair. These disturbing factors were re- 
 moved by putting the subject in an invalid chair which could be ad- 
 justed to any desired position. He could rest here comfortably, with 
 all his muscles as completely relaxed as if he were in bed. This ren- 
 dered the volume of the arm much more constant. But the most 
 conspicuous of the disturbing factors still remained. This is the 
 mental reaction. Apparently, every thought that flits across the 
 
20 Effect of Odours^ Irritant Vapours, etc., upon the Blood Flow 
 
 subject's mind records its presence b}" a change of volume in the 
 arm. Talking in the room, the entrance of a stranger, the striking 
 of a clock, or a noise in an adjoining room, is usually accompanied 
 by a change of volume in the arm, and this sometimes even when 
 the subject is asleep. This is what was to be expected from the 
 correlation which Mosso has shown to exist between cerebral func- 
 tion and the volume of the arm.* Nor do I see any way in which 
 this disturbing factor can be eliminated while working M^th sensory 
 stimulation. The only course open seems to be to confine it within 
 as narrow limits as circumstances will permit, and to endeavour by 
 control experiments to differentiate what mental reactions may re- 
 main from the sensory reactions under investigation. 
 
 The odours were at first administered to the subject by holding a 
 bottle under his nose and allowing him to smell its contents. This 
 was the plan pursued by Lehmann,f but it is open to several obvious 
 objections. The disturbance of mental equilibrium occasioned by 
 such a procedure is itself sufficient to condemn it. Any sudden or 
 unexpected movement of the operator produces a mental reaction 
 in the subject. On the other hand, I soon found that when he an- 
 ticipated my movement a reaction appeared. When he saw me 
 reaching for the odour bottle, even when I did not apply the stimu- 
 lus, a reaction appeared which was often very similar to that accom- 
 panying the sensation itself. It is evident, then, that the subject 
 should be isolated, and everything that could attract his attention 
 as far as possible removed. To attain this end a screen was placed 
 between the subject and the kymograph, so as to prevent him from 
 perceiving the recording apparatus or any of my movements, and a 
 special apparatus was devised for administering the odour without 
 exciting or disturbing the subject. 
 
 The apparatus used for applying the odours is shown in Plate I. 
 By means of an aspirator, a, attached to a hydrant tap, air and cold 
 water are mixed together and poured into a large bottle, t. The 
 water is drawn off by a siphon. The air is forced out through a 
 
 * Kreislauf des Blutes im mensch. Gehim. f Op. cit,, p. 82. 
 
T. E. Shields 21 
 
 small tube, 5, which connects with a coil of lead pipe immersed in a 
 warm bath of constant temperature. This gives us a constant cur- 
 rent of washed air of any desired size at a constant temperature and 
 in constant hygrometric condition. A tube, ^o, conducts this current 
 of air into a metal pipe, y, fastened to the front of a drawer in the 
 kymograph table. A series of thirteen metal stopcocks open along 
 the upper surface of this pipe. In the front of the drawer is ar- 
 ranged a series of thirteen bottles, each containing a small portion of 
 odoriferous substance. A pair of small glass tubes perforate the stop- 
 per of each bottle and terminate at its lower surface. One of each 
 of these pairs of tubes connects with a corresponding stopcock on 
 the metal pipe. Each of the other tubes communicates with a cor- 
 responding glass tube which passes down through the bottom of the 
 drawer, bends over to the screen, and runs up to about the level of 
 the subject's head, where it terminates after passing through the 
 screen. Thirteen flexible tubes connect these glass tubes with thir- 
 teen short metal tubes which are screwed into a terminal plate, the 
 odour plate, Z', and end on a level with its distal surface. A tube, 
 14, which does not make connection with any odour bottle, leads di- 
 rectly from the stopcock on the end of the metal pipe to a four- 
 teenth metal tube in the odour plate. This permits a current of 
 odourless air to be administered when desired. 
 
 The odour plate (Z', Plate I and Fig. 3) is a copper disk, eight 
 inches in diameter, supported from its centre on a universal holder, 
 which allows it to be adjusted in any desired position. Through 
 this plate are bored five series of fourteen holes each, arranged in 
 five concentric circles. The odour tubes terminate in the inner circle 
 of holes. The lumen of a short portion of the terminal end of these 
 metal tubes is less than half the diameter of that of their proximal 
 end. Thirteen of these tubes are provided with electric valves con- 
 structed so as to work noiselessly and to leave the face of the plate 
 as free as possible. This prevents the subject from perceiving the 
 opening or closing of the valves, except by his sense of smell, and 
 removes from the face of the plate all objects to which odours might 
 cling. 
 
22 Effect of Odours^ Irritant Vapours^ etc., upon the Blood Flow 
 
 Fig. 3 shows the construction of one of these valves. A metal 
 disk, A, faced with rubber dam is placed in the large end of the 
 tube. The diameter of the disk is less than the large lumen of the 
 tube and greater than its small lumen. A small wire, J, passes from 
 the centre of this disk out through the small lumen of the tube, and 
 a few millimetres from the face of the plate is fastened to a metal 
 rod, c, which, after running outward parallel with the face of the 
 plate for a short distance, bends back at a right angle, passes through 
 a perforation in the plate, and is adjusted by two nuts to one end of 
 an armature lever, d. This armature lever runs outward parallel 
 with the plate, and terminates over an electro-magnet, e. It is sup- 
 ported on a pivot, h\ situated a short distance centrally to the mag- 
 net. A small spiral spring,^, fastened to the odour plate is adjusted 
 
 h ^MMB 
 
 
 1\\\<\\\^ ,^^ z' 
 
 ^^ 
 
 Fig. 3. — Z, terminal plate (odour plate) ; H, holder ; A, metal disk for closing odour 
 tube ; 6, wire from centre of disk A to rod C ; C, rod connecting wire of valve to armature ; 
 d, armature lever; e, electro-magnet; 6', pivot of armature lever; ff, spiral spring from 
 armature lever to terminal plate ; A, insulated binding screw. 
 
 to the central arm of the lever by a small nut and bolt, which allow 
 the tension of the spring to be regulated at pleasure. One wire of 
 the electro-magnet makes contact with the plate ; the other passes to 
 an insulated binding screw, h, fastened to the plate beside the coil. 
 The tension of the spring is so adjusted as to hold the valve closed 
 except while a current is passing through the magnet, and to pre- 
 vent the armature from touching the magnet at any time. This 
 prevents noise, and the rubber-dam face renders the valve itself 
 noiseless. The movement of the wire on the face of the plate in 
 opening or closing the valve is so small that it escapes the notice of 
 the subject. 
 
 A wire from one pole of a battery makes contact with the metal 
 
T. E. Shields 23 
 
 pipe on the face of the kymograph drawer. Each of the thirteen 
 stopcocks on the upper surface of this pipe is provided with a plati- 
 num contact, which, when the stopcock is open, dips into a corre- 
 sponding mercury well situated on the upper surface of the drawer 
 between the stopcock and its odour bottle. A wire passes from each 
 of these mercury wells to a corresponding binding screw, thence 
 through the coil of the electro-magnet to the terminal plate. A 
 wire in contact with the copper terminal plate passes to the magnet 
 of the signal pen, and thence back to the other pole of the battery. 
 
 The opening of any of these thirteen stopcocks, therefore, sends 
 a constant current of air through the corresponding odour bottle, and, 
 by completing the electric current, opens the terminal end of the 
 corresponding odour tube near the subject's nose, and marks the in- 
 stant of stimulation on the paper of the kymograph. 
 
 Opening the fourteenth stopcock at the end of the metal pipe 
 sends a current of odourless air to one of the odour tubes of the 
 odour plate. 
 
 To secure quiet and exclude light stimulation, the subject was at 
 first placed in a separate and darkened room. I soon found that 
 movements of any part of the body might produce a change in the 
 volume of the inclosed arm. It was necessary, therefore, to bring 
 him into a position where I could observe any movement he might 
 make. A screen was placed so as to prevent him from seeing me or 
 the recording apparatus. The room was of course kept quiet, and 
 visitors excluded except where their presence is noted. All the 
 records discussed here were taken by the improved form of appa- 
 ratus described wabove. 
 
 Discussion of Results. 
 
 To illustrate the character of the records actually obtained, 
 several specimens (reduced in size) of the tracings upon the kymo- 
 graph are given in Plate III, Figs. 1, 2, 3. Fig. 1 illustrates the 
 vasomotor reaction obtained in a certain instance by the odour of 
 extract of heliotrope, and Fig. 2 shows the effect of formic acid when 
 inhaled into the nose. The lowermost line gives the time record in 
 
24 Effect of Odours^ Irritant Vapours^ etc.^ upon the Blood Floxo 
 
 seconds; above this is the pulse record taken by the tambour; and 
 above this the volume curve of the arm upon which the pulse move- 
 ments are not perceptible. A fall of this curve indicates a shrinkage 
 of the arm due to vaso-constriction. At the top of the tracing the 
 respiratory movements are recorded ; and above, the signal pen 
 marks the time of application of the sensory stimulus — i. e., in these 
 cases heliotrope and formic acid. Fig. 3 is a portion of a tracing 
 in which the effect of isobutyric acid was studied. This tracing 
 may.be taken for special description, as it illustrates one of the main 
 difficulties met with in interpreting the curves. The subject's arm 
 was in the cylinder about fifteen minutes when this portion of the 
 record was begun. Curve E gives the time in seconds. Ordinates 
 are drawn at every fifth second. The base line F is drawn parallel 
 to the curve E. The pulse curve, D, shows very little change in 
 the heart rate. There are nineteen heart beats in the first fifteen 
 seconds, and about eighteen in each succeeding fifteen seconds. The 
 respiratory curve, B, shows no very great irregularities. The vaso- 
 motor curve, C, shows an increase of 1'2 cubic centimetres in the 
 volume of the arm during the first twelve seconds. Curve A indi- 
 cates that the odour of isobutyric acid was turned on at the tenth 
 second, but it is evident from the respiratory curve (phase of expira- 
 tion) that it could not have reached the olfactory meml)rane much 
 before the twelfth second. At this point, the arm ceases to increase 
 in volume, and three seconds later it begins to decrease. At the 
 twentieth second it shows a diminution of 1*5 cubic centimetres. 
 The curve then slowly returns to the base line, which it almost 
 touches at the thirty-eighth second. It indicates a second diminu- 
 tion of one cubic centimetre at the forty -fifth second. After this the 
 curve rises slowly and crosses the base line at the sixtieth second. 
 This is followed at intervals of about ten seconds by three positive 
 waves and one negative wave, not shown in the figure, of somewhat 
 less than five cubic centimetres. This curve would seem to indicate 
 that the odour of isobutyric acid produces a reflex stimulation of the 
 vasoconstrictor fibres of the arm, but I do not think it furnishes 
 suflicient evidence for such a conclusion. The curve in other places 
 
T. E. Shields 25 
 
 frequently shows just the opposite effect, although the conditions 
 remained the same. To be at all certain of the effect of any par- 
 ticular stimulation we must, at least, know the general movement 
 going on for some time before and after the stimulation. To avoid 
 the inconvenient length of the original records needed for such a 
 study, I have condensed their data into plotted curves, several of 
 which (Plates IV, Y, YI, and YII) accompany this article. In these 
 curves the distance between the ordinates marks intervals of ten sec- 
 onds each. Curve A, at the bottom of the plates, gives the changes 
 in the amplitude of the pulse waves. The absolute value of this 
 amplitude might easily be determined empirically, but this I have 
 not done. Nor can I calculate it exactly, since I failed to keep a 
 record of the magnification due to the leverage of the writing appa- 
 ratus. But as this leverage was not varied during any of the indi- 
 vidual records, its effect could have no influence on the changes which 
 occur during the same record. The vertical distance between tbe 
 abscissae represents about two-tenths of a cubic centimetre in volume. 
 Curve T> gives the heart rate. For this curve the distance between 
 each two abscissae represents one beat per minute. Curve C gives the 
 changes in the volume of the arm in cubic centimetres. The vertical 
 distance between the abscissae represents a change of two-tenths of a 
 cubic centimetre. The pulse amplitude and the volume changes are 
 calculated at every fifth second. The heart rate is reckoned for every 
 ten or fifteen seconds. The time of stimulation by the various odours 
 is indicated in the plates. The respiratory curve was taken in all the 
 original records, and whenever it presented any marked irregularities, 
 this fact is noted on the plates. 
 
 In order to study the effects of repetition and of individual varia- 
 tion in response to the same odours, the experiments were arranged 
 in two series. The first series of experiments were all made on 
 the same subject ; the second series were made on twelve different 
 individuals. 
 
 J^irst /Series. 
 
 Mr. George Bill, who acted as subject in the first series of ex- 
 periments, is a mechanic, thirty-three years of age, in good health, 
 
26 Effect of Odours, Irritcmt Vapours, etc., upon the Blood Flow 
 
 of fair muscular development, and not particularly susceptible to 
 odours. His arm was kept in the cylinder about an hour daily from 
 the middle of October to the middle of January. He was trained to 
 keep his mind as blank as possible. The monotony of daily repeti- 
 tion soon removed the disturbances which in untrained subjects 
 arise from the novelty of the surroundini^s. The odours of heliotrope, 
 wood violet, wintergreen, musk, skatol, and indol were used daily ; 
 a few others were used occasionally. 
 
 Defects in the earlier forms of apparatus used render the experi- 
 ments made during October and November unsatisfactory. Plates 
 TV to VI, inclusive, give the results of three experiments of this 
 series, which were made during December and January, when all the 
 conditions were most favourable, and may serve as examples of the 
 o-eneral results obtained from this series. 
 
 One of the most obvious features of these curves is the general 
 increase in the volume of the arm which they exhibit during the 
 course of the experiment. In one experiment there was an increase 
 of 39 cubic centimetres in the volume of the arm between the 4:50th 
 and the 2350th seconds; in another, an increase of 30 cubic centi- 
 metres between the 1st and the 1360th seconds; a third gave an in- 
 crease of 35 cubic centimetres between the 20th and the 1300th 
 seconds ; and in a fourth the arm increased 37 cubic centimetres in 
 volume between the 1st and the 1900th seconds. 
 
 Several facts indicate that this increase in the volume of the 
 arm is due not to the temperature or pressure of the water in the 
 arm cylinder, but to the mental condition of the subject ; for in 
 other experiments on the same subject, though the temperature and 
 pressure of the water were the same as in the series just described, 
 no such general increase in volume occurred, nor did it take place 
 in experiments made on the other subjects. Second, the volume 
 of the arm decreased immediately when anything attracted the sub- 
 ject's attention. In Experiment Y (Plate IV) the subject was rest- 
 ing quietly with his arm in the cylinder for ten or fifteen minutes 
 before I began to take the record. During this time the arm in- 
 creased several cubic centimetres in volume. One hundred seconds 
 
T. E. Shields 27 
 
 after the record was begun Dr. K. entered the room quietly, but was 
 perceived by the subject. The volume of his arm decreased 11"5 
 cubic centimetres in the following 50 seconds. One hundred and 
 fifty seconds later, as Mr. G. entered the room, a further decrease 
 of 4 cubic centimetres in 20 seconds ensued. After both had left 
 the room the arm immediately began to increase and continued to 
 increase to the end of the experiment. In Experiment X, I spoke to 
 him at the 240th second. This occasioned a fall of 5 cubic centi- 
 metres in a few seconds. Twice in the course of this experiment I 
 turned on the odour of formic acid for a few seconds ; each time it 
 occasioned a sudden diminution in the volume of the arm. 
 
 The subject had grown familiar with his surroundings ; the chair 
 was very comfortable, the room quiet ; there was nothing to attract 
 his attention, and as a consequence his mind quieted down, and he 
 was frequently dozing toward the end of the experiment. In Ex- 
 periment YII he slept from about the 300th second to the end of 
 the experiment. It was found, too, that in these experiments the 
 heart rate became slower and steadier as the arm increased in vol- 
 ume — an indication that the subject was gradually quieting down or 
 going to sleep. These facts, I think, show pretty clearly that this 
 general increase in the volume of the arm is due to diminished mental 
 activity ; but, whatever may be its cause, it must be taken into 
 account in calculating the efiect of any particular sensory stimulation. 
 To check an upward movement of the curve is practically the same 
 thing as to cause a downward movement of a curve that is running 
 horizontally. A horizontal line evidently can not be taken as a base 
 line in calculating local changes. The base line must run in the 
 general direction of the curve. It is not an easy matter to deter- 
 mine accurately what this line should be ; but in most cases it may 
 be approximated and a line obtained which will be much nearer the 
 truth than a horizontal would be. 
 
 The curves also exhibit two species of local variation which do 
 not seem to be connected with sensory stimulation. In most of the 
 curves small oscillations occur, which suggest Traube-Hering waves. 
 Besides these, larger and slower changes occur at irregular intervals, 
 
28 Effect of Odours^ Irritant Yaponrs^ etc.^ upon the Blood Flow 
 
 which seem to be due to variations in the mental condition of the 
 subject. These changes are the most annoying factors in the curves. 
 I have been unable to eliminate them ; and it is very difficult, if not 
 impossible, to determine the part they play in the volume changes 
 which take place during sensory stimulation. 
 
 Some peculiar features appear in Plate Y, which are probably 
 due to the fact that the subject was asleep during most of the ex- 
 periment. The arm increased rapidly in volume from the begin- 
 ing of the record to the 1300th second, and gradually decreased from 
 this to the end of the experiment. The subject had been resting 
 quietly with his arm in the cylinder for some time before the record 
 was begun. He seemed to be asleep about the 300th second ; at the 
 430tb second I spoke to him and found that such was really the 
 case. He did not wake, but his arm decreased 3 cubic centimetres 
 in the following 20 seconds, and then increased more rapidly than 
 before; 75 and 120 seconds later two other falls of about 4 cubic 
 centimetres occurred which did not seem to be connected with any 
 external stimulus. At the 1930th second, 50 seconds after a musk 
 stimulation had ceased, there was a fall of 5*5 cubic centimetres in 30 
 seconds. At the 2250th second, during an indol stimulation, there 
 was a fall of Q'Q cubic centimetres in 35 seconds, followed by an in- 
 crease of 9 cubic centimetres in 265 seconds, when, 35 seconds after 
 the indol stimulation had ceased, there followed another fall of 7*6 
 cubic centimetres in 30 seconds. The odour of wintergreen was now 
 turned on for 150 seconds, during which time the volume of the arm 
 increased 4 cubic centimetres ; 5 seconds later there was a fall of 5 
 cubic centimetres in 35 seconds. During the following 400 seconds 
 four or five similar falls of less extent occurred. At the 3150th second, 
 during askatol stimulation, the subject woke with a deep inspiration ; 
 the heart rate changed from 56 to 80 ; the volume of the arm in- 
 creased 20 cubic centimetres in 70 seconds, and presently decreased 
 almost as rapidly. He spoke to me at the 3250th second, and seemed 
 asleep 100 seconds later. This sudden increase in the volume of the 
 arm upon awaking is noteworthy, as ordinarily the reverse result is 
 obtained, the arm decreasing in size as the subject returns to a con- 
 
T. E. Shields ' 29 
 
 scions condition. The probable explanation of the unusual result in 
 this case will be given presently in speaking of the changes in heart 
 rate. 
 
 How are these changes in the volume of the arm to be interpreted ? 
 The curve throughout its entire extent exhibits rhythmic changes 
 which suggest Traube-Hering waves. Are the larger changes men- 
 tioned above of a similar nature ? There is a certain periodicity 
 about them, and, if we except the sudden increase in volume which 
 occurred at the 3150th second, they resemble the small changes in all 
 but extent, and, even in this respect, it will be seen that they gradu- 
 ally shade off into each other. It is possible that they are caused by 
 variation in cerebral activity. This explanation would harmonise 
 with Mosso's observations on the changes in cerebral circulation dur- 
 ing sleep.* Changes in mental activity invariably cause changes in 
 the volume of the arm ; the play of the sleeper's imagination may 
 account for the changes with which we are here dealing. Some facts 
 in the curve itself point in this direction. The fall at the 430th sec- 
 ond was apparently caused by my speaking to the subject. The 
 large increase in the volume of the arm at the 3150th second upon 
 awaking appears to be due chiefly to the very great acceleration in 
 heart rate, which occurred at that moment. This was during a skatol 
 stimulation, but it will be seen from the other records that skatol 
 does not, itself, quicken the heart rate. Did the foul odour occasion 
 an unpleasant dream in which some spasm of emotion caused the 
 sudden increase in heart rate and woke the subject ? Speculation 
 concerning the origin of these changes, however, will be of little 
 value until we have more data at our disposal. The question which 
 immediately concerns us here is whether or not, when all due allow- 
 ances are made for changes produced by other causes, the curve still 
 presents any evidence of sensory reactions. We will be in a better 
 position to answer this question after we have made a comparative 
 study of the other curves. The odour of formic acid was administered 
 once for a few seconds in Experiment IX and twice in Experiments 
 
 * Mosso. Kreislauf des Blutes, Leipzig, 1881, S. 74 fE. 
 
30 Effect of Odours^ Irritant Vapours^ etc., upon the Blood Flow 
 
 YI, YIII (Plate YI), and X. Each time it prevented full respira- 
 tion, and caused a sudden and quite large decrease in the volume of 
 the arm. The same result was invariably obtained whenever I used 
 formic acid. Lehmann ^ obtained similar results with ammonia and 
 bisulphide of carbon, which he used as typical unpleasant odours. I 
 think, however, that the reactions in these cases are due in greater 
 measure to the irritating action of these substances on the termina- 
 tions of branches of the trigeminal nerve in the mucous membrane of 
 the nose than to their odours. 
 
 In Experiment YIII, Plate YI, acetic acid was administered at 
 the 1260th and again at the lllOth second. Each time it produced 
 a diminution of 9 cubic centimetres in about 50 seconds. It was also 
 administered in Experiment YI at the 1400th second, in Experiment 
 X at the 1450tli second, and at the end of Experiment XI ; but in 
 none of these cases did it produce any marked reaction. The cause 
 of this discrepancy in results is not apparent. 
 
 Propionic, butyric, isobutyric, and valeric acids were adminis- 
 tered on several occasions without any very decisive reactions 
 appearing. 
 
 In Experiment YI propionic acid was administered at the 2225th 
 second. A fall of 4 cubic centimetres occurred during the 66 sec- 
 onds of the stimulation ; but it is not at all certain that this fall wa& 
 due to the acid, since similar falls occur throughout this curve with- 
 out any apparent dependence on sensory stimulation. In Experi- 
 ment YIII, Plate YI, at the 2160th second the volume of the arm 
 increased 3 cubic centimetres during the first 40 seconds of a propi- 
 onic acid stimulation, and decreased 1-8 cubic centimetres during the 
 remaining 20 seconds. But this stimulation followed immediately 
 upon a fall of 11 cubic centimetres produced by formic acid. The 
 curve usually rises in this manner in such cases without any sensory 
 stimulation. The fall during the latter part of the stimulation was 
 possibly due to the acid. Propionic acid was administered again at 
 the 1870th second in Experiment X. There was a rise here of 1*5 
 
 * Op. cit., p. 84 
 
T. E. Sliields 31 
 
 cubic centimetres during the 20 seconds of tiie stimulation ; but this 
 rise is in the general direction of the curve, and presents no evidence 
 whatever of a sensory reaction. In Experiment YI butyric acid was 
 administered from the lYlOth to the 1890th second, and valeric acid 
 from the 2000th to the 2175th second. Falls occurred in both in- 
 stances ; but from the general movement of the curve it would 
 appear that they are rhythmic contractions, and independent of 
 the stimulations. At the 1920th second in Experiment YIII, Plate 
 YI, the curve was rising rapidly and continued to rise, apparently 
 uninfluenced by the butyric acid which was turned on during the 
 following 50 seconds. At the end of the stimulation the curve took 
 a horizontal direction, which it kept during the following 130 sec- 
 onds. Yaleric acid was administered during 75 seconds of this time 
 without producing any apparent effect. Butyric acid was also ad- 
 ministered at the 1715th second in Experiment X. A fall of 2 cubic 
 centimetres occurred during the first 30 seconds, but the curve shows 
 that this also may be a rhythmic contraction. Isobutyric acid was 
 used once in each of the Experiments YI, YIII, IX, X, and XI with- 
 out any apparent effect. 
 
 Besides formic acid and the two instances of acetic acid men- 
 tioned above, there are three other causes which, when present, usu- 
 ally produce a marked diminution in the volume of the arm. These 
 are a deep inspiration, muscular movement, and mental activity. 
 A typical effect of a deep inspiration occurred at the 1070th second 
 in Experiment X. This diminution in the volume of the arm seems 
 to be due in part to the increased negative pressure in the pleural 
 cavity, but I think it probable that a factor in the production of this 
 reaction is the change in the mental condition of the subject, which 
 usually accompanies an inspiration of this kind ; for on such occa- 
 sions the subject is usually drowsy, and with the deep inspiration 
 the mind brightens temporarily. It would seem that there is a con- 
 striction of the peripheral arteries, causing an increased flow of 
 blood to the brain. When a deep inspiration occurs while the mind 
 is bright there is seldom any very large decrease in the volume of 
 the arm. It is possible that when the change takes place during the 
 
32 Effect of Odours, Irritant Vapours, etc., upon the Blood Flow 
 
 odour stimulation, the stimulation also plays a part in its production ; 
 but if so, the effect of the odour can not well be isolated. It is better, 
 therefore, for present purposes, to discard such reactions. At the 
 2940th second in Experiment YI, I requested the subject to move his 
 foot, which he did during the following fifty seconds. The heart 
 beat immediately changed from 72 to 92, and after thirty seconds 
 gradually returned to its former rate. This sudden increase in 
 heart rate caused a transitory increase of 6*5 cubic centimetres in the 
 volume of the arm, followed by a decrease of 13 cul)ic centimetres 
 in fifty seconds. In this case it is probable that the effect was due 
 in part to the muscular work and in part to increased mental ac- 
 tivity. Whenever the subject moved his foot unconsciously, as he 
 sometimes did, as at the 1670th second in this experiment, it pro- 
 duced a slight decrease in the volume of the arm, but of course there 
 was not nearly the same amount of muscular work in these cases as 
 in the instance referred to above. 
 
 The curves furnish numerous illustrations of the effect of mental 
 work. This factor always causes a decrease in the volume of the 
 arm. A typical instance of this occurs at the 100th second in Ex- 
 periment Y, Plate lY. Sometimes the fall is preceded by a transi- 
 tory rise, as at the 1690th and 2940th seconds in Experiment YI ; 
 but in all such cases the rise is evidently due to a sudden accelera- 
 tion of heart rate. Besides these changes, which can be definitely 
 accounted for, the records of this series of experiments show a num- 
 ber of other variations in the volume of the arm. The general 
 movement of the curve has already been discussed. In all the curves 
 small oscillations occur every few seconds which are very much more 
 pronounced in some records than in others, but which remain pretty 
 much the same throughout the entire extent of the same record. 
 These facts indicate that they are influenced by the condition of the 
 subject rather than by any particular stimulation. 
 
 When due allowance is made for all these changes, do the records 
 of this series of experiments present any evidence of pure odour re- 
 actions ? 
 
 White heliotrope, wood violet, wintergreen, and musk were 
 
T. E. Shields 33 
 
 chosen as specimens of pleasant odours ; skatol and indol as unpleas- 
 ant odours. The subject enjoyed the first four very much and dis- 
 liked the latter two. 
 
 Heliotrope was administered at the 1365th second in Experiment 
 Y, Plate lY ; at the 220th and 270th seconds in Experiment YI ; 
 at the 210th and 2800th seconds in Experiment YII, Plate Y ; at 
 the 4:5th second in Experiment YIII, Plate YI ; at the 850th and 
 1910th seconds in Experiment XI. In none of these instances is 
 there any clear evidence of a reaction. There is not one feature 
 common to these various curves. There is scarcely a resemblance 
 between any two of them. The only approach to a resemblance 
 among them is the upward tendency of the curve, but this is in the 
 general direction of the curve, and evidently is not due to the stimu- 
 lation. 
 
 What has here been said of heliotrope applies with equal force 
 to the other five odours mentioned. Yiolet was administered at the 
 430th second in Experiment YI. The volume of the arm increased 
 2 cubic centimetres in the first thirty seconds and decreased 4'5 
 cubic centimetres in the following thirty seconds ; it then increased 
 to the end of the stimulation. Indol was administered at the 920th 
 second in the same experiment, with very similar results. At first 
 sight these look like odour reactions. But it was found that five 
 other similar falls occur in the first 1,000 seconds of this experiment, 
 at approximately equal intervals of time. The first two falls 
 occurred before any odour was administered ; the third, fifth, 
 and sixth in intervals between stimulations. The latter portion of 
 the curve presents a similar series of falls. Besides, if the falls oc- 
 curring during the violet and indol stimulations were odour reactions, 
 we would expect that they would be repeated with some regularity 
 in numerous experiments on the same subject ; but such is not the 
 case. It is altogether probable, therefore, that we are here dealing 
 not with sensory reactions, but with rhythmic contractions of the 
 blood vessels. A comparison of the curves obtained by repeated 
 applications of any one of these six odours will make it clear that, if 
 the stimulation had any effect on the volume of the arm, it is very 
 
34 Efect of Odours^ Irritant Vapours, etc, njpon the Blood Flow 
 
 thoroughly masked by other movements of the curve. If the odours 
 really produced any characteristic effects they would come out in a 
 composite curve made from a large number of experiments. I 
 have not yet a sufficient number of experiments to give any value to 
 such a composite curve. 
 
 The experiments presented here are essentially similar to a num- 
 ber of other experiments made on the same subject, all of which 
 were conducted with great care; but the results obtained do not 
 justify any such conclusions as those which Lehmann deduces from 
 his experiments. 
 
 " Jeder lusterregende Eindruck erzeugt eine Vergrosserung des 
 Yolumens des Armes und der Hohe der einzelnen Pulsschliige nebst 
 einer Vergrosserung der Tiefe des Atemholens." ^ 
 
 " Einfache, unlusterregende Sinneseindriicke rufen, wenn sie 
 schwach sind, sogleich eine Yerminderung des Armvolumens und 
 der Hohe der einzelnen Pulsschlage hervor. Das Yolumen nimmt 
 bald wieder zu, trotz der Yerkleinerung der Pulsschliige, und iiber- 
 schreitet gewohnlich die Norm, wenn die Pulsschlage ihre vorige 
 Grosse erreicht haben, die iibrigens im allgemeinen ebenfalls iiber- 
 schritten wird. Bei stiirkeren, aber doch nicht schmerzhaften Ein- 
 driicken treten diese Yeranderungen mehr hervor und werden zu- 
 gleich unmittelbar nach Anfang der Reizung von einigen tiefen 
 Atembewegungen begleitet." f 
 
 " Einfache lustbetonte Sinnesempfindungen werden von einer 
 Gefiisserweiterung begleitet und vielleicht auch zugleicli von einer 
 Yergrosserung des Urafanges der Herzkontraktionen in Yerbindung 
 mit einer Erhohung der Innervation der willkiirlichen Muskeln, je- 
 denfalls der Atmungsmuskeln.";}; 
 
 An inspection of the curves of this series will satisfy any one 
 that none of these statements is true of them. Heliotrope and wood 
 violet are certainly pleasant odours and were enjoyed by the subject ; 
 yet the volume of the arm diminished quite as often as it increased 
 during their application. Indol and skatol are unpleasant odours, yet 
 
 * Op. cit., p. 8-2. f Ibid., p. 89. :}: Jbld. 
 
T. E. Shields 35 
 
 the volume of the arm frequently increased during the first few 
 seconds of their application, and then decreased. It is not certain, 
 however, that any of these changes were pure odour effects. It is 
 but proper to say that the results obtained in this series with the 
 perfected apparatus were all obtained from the subject after he had 
 been repeatedly used in similar experiments involving the same set 
 of odours. There were some indications that the vasomotor effects 
 in the beginning were more pronounced than toward the end, when 
 the odours had lost somewhat of their pleasant or unpleasant effects 
 and the subject endured them in a rather perfunctory manner, but 
 it is not possible to speak definitely as to this point, since the earlier 
 apparatus employed was subject to so many errors that the records 
 then obtained were not trustworthy. 
 
 Second Series. 
 
 The second series of experiments was made on twelve students, 
 ranging from twenty to thirty years of age. Each experiment 
 lasted about an hour. The same odours and acids were used as in the 
 former experiments. There are some features common to all these 
 experiments. 
 
 In the first couple of experiments with each subject the curve 
 exhibited very little of the tendency to rise which constituted such a 
 prominent feature of the series already described; but this tendency 
 became more pronounced in the third and fourth experiments. 
 Where the general tendency of the curve is upward the heart rate 
 also becomes slower and steadier. This supports the view advanced 
 above, that the gradual increase in the volume of the arm and the 
 slowing of the heart rate are due to the diminished mental activity, 
 and is therefore more pronounced when the subject has become ac- 
 customed to the plethysmograph and ceases to feel the mental excite- 
 ment naturally attendant upon a novel experience. The small 
 oscillations in the curves are, as in the former series, characteristic of 
 the experiment. They are much more pronounced in some experi- 
 ments than in others, but remain pretty constant throughout the 
 whole extent of the same experiment. 
 
36 i^ftct of Odours. L^riUiut Vapours, eh\, upofi the Blood Fhw 
 
 The reactions in all the experiments are in the same direction, 
 but thev varv in extent with the different subjects, with the same 
 subject at different times, and with the diff'erent stimuli. 
 
 Plate YII (Experiment XII) will serve as an illustration of the 
 reactions obtained froiu subjects who are sensitive to odours. This 
 was the third experiment on Mr. D., who is a student twenty-two 
 years of a^e. of rather slight build, and nervous temperament. 
 
 The amplitude of the pulse wave diminishes to about one half at 
 the tirst odour stimulation,* and does not again return to its original 
 size. The heart rate slows up considerably during the experiment. 
 The general tendency of the volume curve is slightly upward. 
 
 The subject was resting quietly with his arm in the cylinder for 
 ten or fifteen minutes before the record was begun. During this 
 time, and for the first 370 seconds of the record itself, the volume 
 of the arm increased slowly, without exhibiting any very marked 
 changes. Heliotrope was administered from the 3T0th to the 725th 
 second. During the first 50 seconds of this period the volume of the 
 arm decreased 19*5 cubic centimetres. It increased about 1 cubic 
 centimetre during the following 50 seconds, and li cubic centimetres 
 during the next 125 seconds. This was followed bv a fall of 4*5 
 cubic centimetres in 25 seconds, and a rise of 7*5 cubic centimetres 
 in 100 seconds. The prompt fall at the beginning of the stimulation, 
 and the fatigue effect after about 100 seconds are pretty good evi- 
 dence that this is an odour reaction, especially when it is taken in 
 connection with the character of the curve previous to the stimula- 
 tion. This conclusion is confirmed, I think, by the fact that a 
 similar reaction was obtained every time I used this odour on Mr. D. 
 Two other heliotrope stimulations occur in this record with results 
 very similar in everything but extent ; and this difference is at least 
 partially accounted for by the fact that at the 740th second I moved 
 the terminal plate farther from the subject's nose so as to diminish 
 considerably the strength of the stimulation. 
 
 The fall of IS cubic centimetres which occurred between the 
 740th and the 790th second is a mental reaction caused by my 
 
 * The original record of part of this stimulation, reduced in size, is given in Plate IH, Fig. 1. 
 
T. E. Shields 37 
 
 speaking to the subject and readjusting the odour plate. When the 
 volume of the arm had returned to about its former condition and 
 seemed inclined to remain pretty steady I turned on wood violet. It 
 will be seen that the reaction is very similar to the previous heliotrope 
 reaction, only less extensive. There is a fall of 8*4 cubic centimetres 
 in the first 30 seconds. It remains down about 25 seconds, and in- 
 creases 6*5 cubic centimetres in 70 seconds, when a second fall of 
 3*7 cubic centimetres occurs. Wintergreen was turned on from the 
 1265th to the 1440th second ; the reaction is very similar to the pre- 
 vious violet reaction in every respect but the extent of the second 
 fall. I had requested the subject before the experiment began to 
 leave his mind as blank as possible, and not even to speculate about 
 the nature of the odours. So great, however, was the down sweep 
 of the curve that I suspected the presence of a mental reaction, and 
 accordingly at the 1500th second I turned on a current of odourless 
 air strong enough to be felt on his face. When the experiment was 
 over I remarked to him that he had not obeyed my instructions, but 
 had been indulging in speculations concerning the odours. He as- 
 sured me that such was the case on only one occasion for a few 
 seconds, when he felt the air coming from the tube but could per- 
 ceive no odour. This indicates that the reaction in the other in- 
 stances are connected in some way with the odours. At the 1650th 
 second heliotrope was again administered. There was a prompt fall 
 of 10 cubic centimetres in 35 seconds, followed by two quite large 
 falls during the course of the stimulation. After an interval of 150 
 seconds, during which time there was very little change in the volume 
 of the arm, musk was turned on. It occasioned a prompt fall of 8 
 cubic centimetres in 20 seconds. The two secondary falls were very 
 slight. Skatol was turned on from the 2640th to the 2740th second 
 without any apparent effect. Heliotrope was administered for the 
 third time at the 2760th second. The reaction is essentially similar 
 to the two former instances. I spoke to him at the 3060th second. 
 An acquaintance entered at the 3150th second. When the subject 
 had quieted down I turned on formic acid for 35 seconds. This 
 caused a fall of 18 cubic centimetres in 50 seconds. 
 
38 Effect of Odours, Irritant Vapours, etc., upon the Blood Flow 
 
 This record, as has ah'eadj been said, was chosen as a specimen 
 of very pronounced odour reactions, but each of the twelve subjects 
 on whom this series of experiments was made gave unmistakable 
 evidence of reactions to odours and irritant vapours. The reactions 
 were always in the same direction, although seldom as pronounced 
 as in the case of Mr. D. In the mental reactions obtained in this 
 series it was found that the volume of the arm decreased very mark- 
 edly notwithstanding a simultaneous quickening of the heart rate. 
 This is characteristic of the mental reactions in all my experiments, 
 A second striking peculiarity of the mental reactions is their long 
 continuance. The effect of odours or irritant vapours passed off in 
 three or four minutes at most, whereas mental work frequently kept 
 the heart beating at a high rate and kept the volume of the arm low 
 for many times that period. 
 
 The relation existing between the effects of sensory stimulation 
 and of mental work will be discussed in a future paper. 
 
 Stjmmaky. 
 
 The most important outcome of this investigation has been the 
 completion of various improvements in the construction and use of 
 the plethysmograph, by means of which numerous errors attending 
 the use of the instrument have been eliminated. 
 
 The results of the work show that all olfactory sensations, so far 
 as they produce any effect through the vasomotor system, tend to 
 diminish the volume of the arm, and therefore presumably cause a 
 congestion of the brain. Whenever the stimulation occasions an in- 
 crease in the volume of the arm, as sometimes happens, it seems to 
 be due to acceleration of the heart rate, which, of course, tends also 
 to increase the supply of blood to the brain. The effect of odours 
 varies in extent with differeiit individuals, and with the same indi- 
 vidual at different times. It was most marked in subjects sensitive 
 to odours. Irritant vapours, such as formic acid, have a marked effect 
 in the same direction — that is, they cause a strong diminution in the 
 volume of the arm. The experiments give no support to the view 
 that pleasant sensations are accompanied by a diminution of the 
 
T. E. Shields 39 
 
 blood supply to the brain and unpleasant sensations by the reverse 
 effect. In all my experiments mental work caused a marked and 
 prolonged diminution in the volume of the arm. This vasomotor 
 effect was sometimes preceded by a transitory increase in the volume 
 of the arm caused by acceleration of heart rate. 
 
 Acknowledgment. 
 
 I wish to return my sincere thanks to Prof. Howell, at whose 
 suggestion this work was undertaken, and whose kind supervision 
 and advice was a constant help and encouragement at every step of 
 its progress. I am further indebted to him for his generosity in pro- 
 viding a subject for the first series of experiments. I am also in- 
 debted to Prof. E. A, Pace, of the Catholic University, for many 
 valuable suggestions on the psychological aspects of the problem, and 
 to the Rev. C. A. Pamm for suggestions in the constructions of the 
 apparatus, and for the three accompanying drawings. 
 
 1 must also take this opportunity of thanking the young gentlemen 
 of St. Mary's Seminary, who generously placed their time at my dis- 
 posal in acting as subjects for the second series of experiments. My 
 thanks are due in a special manner to Messrs. Riedel and Beavan, 
 who acted as subjects in a number of experiments and who assisted 
 me in the laborious task of tabulating the results and plotting the 
 curves. 
 
 June 1, 1895. 
 
 Explanation of Plates. 
 
 Plate I. — General Aerangement of Apparatus. 
 
 A, B, support for test tube and siphon ; C, D, D', clamps for same ; E, clamp for spiral; 
 F, F', wires from spiral to rubber ring; H, I, tube connecting siphon with arm cylinder; J, 
 four-way; 1, 2, 3, 4, stopcocks on same; K, piston tube; S, metal cap for same; M, M', 
 stay rods to upright ; N, piston rod ; 0, 0', rods connecting metal cap with washer ; P, Q, 
 supply jars ; Q, milled nut for adjusting level of piston ; Q', milled nut for adjusting indi- 
 cator ; R, rubber ring carrying test tube ; S, spiral suspending test tube ; T, test tube ; TJ, 
 upright ; W, washer over milled nut on piston rod ; X, tambour ; T, T-way in tube con- 
 necting piston tube with tambour ; A, B, stopcocks on same ; w, Z, three-way stopcocks 
 on supply tubes ; A, aspirator ; t, bottle for washing air ; U, bath for warming air ; ?/, air 
 tube on face of kymograph drawer; z\ terminal plate; e, electro-magnet; d, armature; Ji, 
 binding screw ; d, ring suspending arm cylinder. 
 
40 Efect of Odours, Irritant Vapours, etc., upon the Blood Flow 
 
 Plate II. — Arrangement of Cylinder and Method of holding the Arm Rigid. 
 
 A, arm ring ; B, wristlet band ; C, D, rubber rings for making arm holder rigid to arm 
 cylinder ; G, H, faces of arm ring ; M, M', rods making arm ring rigid to wristlet ; R, ring 
 cemented into rubber sleeve ; S, rubber sleeve ; e, I, m, orifices in arm cylinder ; h, ther- 
 mometer ; t, stopcock to let air escape from arm cylinder ; Y, metal band for supporting 
 cylinder ; K, supply tube ; rf, ring from which arm cylinder is suspended. 
 
 Plate III. — Specimens of Tracing (reduced one half in size) taken upon the 
 
 Kymograph. 
 
 1. Heliotrope stimulation (subject, Mr. D.). 2. Formic-acid stimulation (very light current 
 of vapour). 3. Isobutyric-acid stimulation (subject, George Bill): A, signal — isobutyric acid 
 turned on from 10th to end of record ; B, respiratory curve — expiratory phase from 10th to 
 12th second shows that the stimulation could not have reached olfactory membrane before the 
 12th second; C, vasomotor curve; shows increase of 12 cubic centimetres in volume of 
 arm between the 1st and 12th second, a decrease of 15 cubic centimetres at the 20th sec- 
 ond, a return to base line at 38th second and a second decrease of 1 cubic centimetre at 
 45th second, and a return to base line at 60th second; D, heart rate. It shows 19 heart 
 beats in first 15 seconds, and about 18 in each succeeding 15 seconds ; E, time in seconds ; 
 F, base line ; ordinates are drawn at every fifth second. 
 
 Explanation of Plotted Curves. Plates IV, V, YI, and VII. 
 
 A, changes in amplitude of pulse wave ; vertical distance between abscissae 0"2 cubic cen- 
 timetre. D, heart rate ; vertical distance between each two abscissae represents one beat 
 per minute. C, changes in volume of arm ; vertical distance between each two abscissae 
 represents 0-2 cubic centimetre. The distance between the ordinates makes intervals of 10 
 seconds each. The pulse amplitude and volume changes are calculated at every fifth sec- 
 ond. The heart rate is reckoned for every 10 or 15 seconds. The respiratory curve was 
 taken in all the original records. Whenever it presents any marked irregularities this fact 
 is noted on the plates. Time of stimulation by the various odours is indicated on the plates. 
 
 Plate IV. — December 3d (subject, Mr. George Bill). Resting with arm in cylinder 15 
 minutes before record began. Volume of arm increasing slowly. Heart rate, 60. 100th 
 second, Dr. K. entered. Volume of arm decreased 115 cubic centimetres in 60 seconds. 
 250th second Mr. G. entered. Volume of arm decreased 4 cubic centimetres in 20 seconds. 
 Heart rate changed to 68, 64, 71, 64, 72, 68. 490th second. Dr. K. and Mr. G. leave room. 
 Heart rate, 60. Volume of arm increases rapidly to 700th second, and more slowly to end 
 of record. Subject dozing toward end of experiment. Heart rate gradually sinks to 52 at 
 end of experiment. Total increase of volume of arm between 450th second and end of 
 experiment, 39 cubic centimetres, due to diminished mental activity. 
 
 Plate V. — December 12th (subject, Mr. George Bill). Resting quietly with arm in 
 cylinder for some time before record began. Asleep from 300th second to end of experi- 
 ment. Heart rate 60 during most of the experiment. No odour reaction. Volume of arm 
 increases 30 cubic centimetres between 1st and 1360th seconds. Small rhythmic contrac- 
 tions every 10 or 15 seconds. Spoke to him at the 430th second, which occasioned a fall 
 of 3 cubic centimetres in 20 seconds. 530th second, fall of 4 cubic centimetres. 650th 
 second, fall of 4 cubic centimetres. 1930th second, fall of 5*5 cubic centimetres in 30 sec- 
 onds. 2250th second, fall of 6'6 cubic centimetres in 35 seconds. 2515th second, fall of 
 7*6 cubic centimetres in 30 seconds. 2710th second, fall of 5 cubic centimetres in 35 sec- 
 onds. Four or five similar falls of less extent occur during the following 400 seconds. 
 None of these falls seem to be due to odour stimulation. They are probably rhythmic con- 
 
JOURNAL OF EXPERIMENTAL MEDICINE. VOL. ! 
 
TH E JOURNA L OF EXPERIM ENTAL MEDICINE. VOL. I. 
 
 fiV! !C 
 
 1 
 
 
 
 
 
 
 i 
 
 P 
 
 i 1 
 j-fii 
 
 ' 
 
 ) 
 
 1 - 
 
 "4 
 
 
 1 IM 
 
 
 ^1; 
 
 "i 
 
 
 i'il 
 
 
 Hi 
 
 III: 
 
 u 
 
 ii;! 
 
 ',: I 
 
 Ijl.^ „■> 
 
 "S! 
 
 
 
 -^O/r 
 
 -J 
 
 
 ---'^ -'"' 
 
 T ,>.ri , ,--r 
 
 P; STr !V jyprp,..jy,-r V. 
 
 
 
 
 
 
 
 
 ■-S 
 
 
 
 
 * 
 
 
 
 
 
 
 .v^ 
 
 f«^' 
 
 / 
 
 
 
 \J'^\ i 
 
 iW:-i;:i^ 
 
 
 
 
 
 
 
 1 i 
 
 
 ''■f:|-.-k-|"^^-'^-^^"'^ 
 
 
 I ; • \ 
 
 ! 
 
 
 
 a; ,v n- 
 
 
 
 W ■ . 
 
 
 
 
 
 
 
 
 
 
 '••^^"•^■w^ » I^ ,-\/r^.vrv5j 
 
 
 ;is 
 
 
 ll 
 
 ±t 
 
 ' 
 
 1 
 
 -1 
 
 :t 
 
 1? 
 
 ! 
 
 
 
 
 
 
 
 iHifM^-h: 
 
 ''"^t+trti-ttjf 
 
 ■r^ 
 
 
THE JOURNAL OF EXPERIMENTAL MEDICINE. VOL. I. 
 
 V. EXPERIMENT VII. 
 
THE JOURNAL OF EXPERIMENTAL MEDICINE. VO L. 1. 
 
 ■iG\ i j rh 
 
 2;g^-— ^^°« t=^-.3ah^t T^guttjttgaint^^ i' >%; 
 
 ;\ ; p., ',^fv 
 
 M 
 
 -,.,.nWiV 
 
 PLATE V. EXPERIMENT VM. 
 
 
 
 V 
 
 /^ i;'^i' 
 
 r1/^' 
 
 1' 
 
 ■17 
 
 ^ 
 
 u 
 
 .VXJ^^ 
 
 'S 
 
 , / 
 
 
 Iff 
 
 
 «. 
 
 
 'iv 
 
 V'W'f 
 
 
 
 i 
 
 - 
 
 ^ 
 
 A 
 
 
 
 ~ 
 
 % 
 
 ; : 
 
 
 
 
 
 
 
 
 
 
 
 i - : l^^ ' r^ 
 
 '*'***nr\juriA. 
 
 ,-|- • -v ; --i '\->v,'^ I ~\,-'|~'\,,-;^-.V^.,^^^. S_V/i 
 
 -Vv'^r^^g'^^>->.^^^ 
 
 ifl-'.- U-!»l0»fttr»tii 
 
 MlM^ 
 
THE JOURNAL OF EXPERIMENTAL MEDICINE. VOL. I. 
 
 . EXPERIMENT VMI, 
 
THE JOURNAL OF EXPERIMENTAL MEDICINE. VOL. I. 
 
 --F-=;F=^p^T.-7pT.p^^^ 
 
 PLATE VI. EXPERIMENT VIM. 
 
 r^i/'V" 
 
 \^f' 
 
 ^wl 
 
 \pj 
 
 '1 (hfi 
 
 ■ \r'- 
 
 (U ■ -K 
 
 
 
 i ' \i ! 
 
 
 
 
 
 1 
 
 
 
 
 4? 
 
 '■"? -. , ^ 
 
 \ I 
 
 y 
 
 iV'V 
 
 y: 
 
 
 1 
 
 HAJ n 
 
 ;^ii 
 
 !'35!lTfEPP 
 
 
 T1 
 
 
 
 " - 19 
 
 1—- 
 
 1 '■ 
 
 T=a 
 
 - 3 
 
 
 -^^ 
 
 
 1 
 
 
 
 
 
 -^^^^^^^ 
 
 __- 
 
 ^35 
 
 -:s 
 
 ri 
 
 ^"3 
 
 
 .-— 
 
 
 - -r+ 
 
 
 
 "5 
 
 -^ 
 
 
 -q^ 
 
 iaBaiiwM5teaajai3meM» j :tfflifeitfi^ psypgr^wg 
 
THE JOURNAL OF EXPERIiVlENTAL MEDICINE. VOL. I. 
 
 II. EXPERIMENT XI 
 
THE JOURNAL OF EXPERIMENTAL MEDICINE, VOL. I. ^^^^^ 
 
 PLATE Vlt. EXPERIMENT XII 
 
 \mm\wmmmmmmm mimpmmwmieMi^^ 
 
 f'S-', 
 
 A 
 
 r\ 
 
 1 
 
 1 
 
 n 
 
 j 
 
 r U^ 
 
 
 .A/-n 
 
 
 ;/' 
 
 
 
 
 
 
 
 
 
 ■!^-U 
 
 -i^ 
 
 n- f 
 
 mr 
 
 
 
 
 
 
 
 
 
 
 
 p .- 
 
 
 
 
 
 tfzJ-- 
 
 
 
 
 
 
 
 
 
 
 ^T 
 
 U\ ^.^^^J\X 
 
 -'-^=l-^----^-^^ 
 
 ' — —H — 1 — i — '■ ■ 
 
 — t'--'!-±r-h-iT;]=f= 
 
 -^-14 
 
 --=FH 
 
 
 .^S 
 
 
 --^1 1 
 
 
 
 
 
 
 
 
 
 
 
 ^^^v..*-^^ 
 
T. E. Shields 41 
 
 tractions. 3150th second, subject woke with deep inspiration. Heart rate changed from 
 56 to 80, probably due to a spasm of emotion caused by unpleasant dream. Volume of arm 
 increased 20 cubic centimetres in '70 seconds, apparently due to quickened heart rate. 3250th 
 second, subject spoke and seemed asleep 100 seconds later. Arm returned to its former 
 volume and heart rate to 60. Amplitude of pulse wave diminishes from beginning to end 
 of experiment. 3250th second, about one third its initial amplitude. 
 
 Plate VI. — December 14th (subject, Mr. George Bill). Volimie of arm increases 12 
 cubic centimetres during first 1100 seconds. 1260th second, acetic-acid stimulation, dimi- 
 nution 9 cubic centimetres in volume of arm in 30 seconds. 1410th second, acetic acid, with 
 similar result. 1580th second, formic-acid stimulation for 15 seconds, occasioning a fall of 
 13 cubic centimetres in 80 seconds. 2100th second, formic-acid stimulation for 30 sec- 
 onds, with a fall of 11 cubic centimetres in 60 seconds. Odour stimulations apparently 
 without effect throughout the curve. Amplitude of pulse wave diminishes gradually through- 
 out the experiment. 
 
 Plate VII. — January 31st (subject, Mr. D). Subject resting quietly with arm in cyl- 
 inder 15 minutes before record began, during which time and for the first 3Y0 seconds of 
 the record itself volume of arm increased slowly without any marked changes. Heart rate 
 gradually slows during experiment. Volume of arm exhibits a slight upward tendency 
 during most of the experiment. Amplitude of pulse wave diminished to about half at the 
 first odour stimulation, and does not return to its original size. 3'70th second to 725th, 
 heliotrope stimulation. Volume of arm diminishes 19'5 cubic centimetres in first 50 sec- 
 onds, increases 1 cubic centimetre during the following 50 seconds, and 14 cubic centi- 
 metres during the next 125 seconds. 595th second, fall of 4-5 cubic centimetres in 25 sec- 
 onds. 740th second, spoke to subject; fall of 18 cubic centimetres. 1020th second, 
 wood-violet stimulation, fall of 84 cubic centimetres in first 30 seconds ; second fall, 3-7 
 cubic centimetres at 1145th second. 1270th second, wintergreen stimulation, fall of 8'5 
 cubic centimetres in 30 seconds; a second slight fall at 1370th second. 1500th second, 
 current of air; subject speculates as to its nature; fall of 4*5 cubic centimetres in a few sec- 
 onds. 1650th second, heliotrope stimulation; fall of 10 cubic centimetres in 35 seconds, fol- 
 lowed by two quite large falls in course of the stimulation. 2350th second, musk stimula- 
 tion; fall of 8 cubic centimetres in 20 seconds, followed by two slight secondary falls. 
 2640th and 2740th seconds, skatol stimulation, no apparent effect. 2760th second, helio- 
 trope ; fall of 10 cubic centimetres in 30 seconds. 2870th second, fall of 2"5 cubic centi- 
 metres in 15 seconds. 2985th second, fall of 6 cubic centimetres in 30 seconds. 3060th 
 second, subject spoke; fall of 11 cubic centimetres in 90 seconds. 3150th second, S. en- 
 tered room; fall of 2'5 cubic centimetres in 30 seconds. 3270th second, fall of 3 cubic 
 centimetres without obvious cause. 3450th second, formic-acid stimulation for 30 seconds ; 
 faU of 18 cubic centimetres in 50 seconds. 
 
 BIOGRAPHY. 
 Thomas Edward Shields was born on the 9th of May, in the 
 jear 1862, at Mendota, one of the suburbs of St. Paul. Minn, In 
 1882 he entered the sophomore class of St. Francis College, Mil- 
 waukee, and ^-emained there until 1885, completing its classical cur- 
 riculum. In September, 1885, he began his philosophical studies in 
 the theological seminary of St. Thomas Aquinas at St. Paul. Here 
 
42 Effect of Odours^ Irritant Vapours^ etc., vpon the Blood Flov) 
 
 he passed six years : two years in the study of mental philosophy, 
 ethics, and the physical sciences, and four years in the study of the- 
 ology — dogmatic and moral — and the accompanying branches of 
 Holy Scripture, Church history, and canon law. He was ordained 
 priest on March l-i, 1891, and in the following June was assigned 
 as curate to the Cathedral of St. Paul. After fourteen months in 
 the active ministry there he came to Baltimore (in September, 1892), 
 received the degree of Master of Arts from St. Mary's University, 
 and began his studies in the Biological Department of the Johns 
 Hopkins University in October of the same year. 
 
COLUMBIA UNIVERSITY LIBRARIES 
 
 This book is due on the date indicated below, or at the 
 expiration of a definite period after the date of borrowing, as 
 provided by the rules of the Library or by special arrange- 
 ment with the Librarian in charge. 
 
 DATE BORROWED 
 
 DATE DUE 
 
 DATE BORROWED 
 
 DATE DUE 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 C28(i!41)m100 
 
 
 
 1