Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Thresholds for the Perception of Angular Acceleration About the Three Major Body Axes' BRANT CLARK San Jose State College JOHN D. STEWART Ames Research Center, NASA SUMMARY This study is concerned with man's sensitivity to body rotation about his three major body axes. The specific purpose was to determine thresholds for the perception of rotation about the x-, y-, and z-axes and to compare these results for the group and for the individual observers. The thresholds of 18 men with normal vestibular function were established for the x-, y-, and z-axes by use of a pre- cision rotation device. The angular acceleration was ordered, using a random, forced-choice, double- staircase procedure, and the order of the determination of the three thresholds for each observer was established by a Latin-square method. Mean thresholds were found to be equal for the x- and z-axes. The mean threshold about the y-axis (somersaulting axis) was found to be substantially greater than those about the x- and z-axes, but these differences were both just below statistical significance. There was a great range in thresholds for all three conditions. The intercorrelations among the three thresh- olds were not significantly different from zero. It was concluded that under optimum testing con- ditions, the mean thresholds about the * â¢. y- and z-axes are essentially the same but that the threshold about one body axis does not predict the threshold about the other two axes for a given observer. INTRODUCTION Almost all the studies of the sensitivity of the semicircular canals to angular acceleration have been made with the observer rotating only about his vertical axis (yaw or z-axis) (ref. 1). Con- sequently, there is a dearth of information re- garding man's sensitivity to rotation about the other two major body axes. Indeed, comparisons of the perception of angular acceleration applied about the somersaulting axis (pitch or y-axis) and cartwheeling axis (roll or x-axis) are so extremely limited that no definitive statements on the range of canal sensitivity about these 1 The experimental work for this study was carried out at Ames Research Center under National Aeronautics and Space Administration grant NGR 05-046-002 to San Jose State College. axes are possible. On the basis of anatomical and physiological information concerning the semicircular canals, it is quite clear that the vertical and the horizontal ones are functionally different in several ways (refs. 2 to 4). For example, Lowenstein and Sand (ref. 2), using single nerve fiber preparations in the ray, re- ported that the vertical canals can be directly stimulated by rotation in any plane, whereas the horizontal semicircular canals are stimulated almost exclusively by rotation about a vertical axis. They also pointed out that the vertical semicircular canals are phylogenetically older than the horizontal semicircular canals and are subdivided by a septum which is not found in the horizontals. Consequently, upon neurological and physiological bases one might expect some differences in sensitivity in rotating about 299
300 THE ROLE OF THE VESTIBULAR ORGANS IN SPACE EXPLORATION the various axes. At the same time, there is no comparative anatomical evidence regarding the differential functioning of the three pairs of ampular systems, and, furthermore, rotation about any of the three major body axes will stimulate more than one pair of canals. The limited experimental findings are some- what conflicting with regard to thresholds for the perception of angular acceleration about the three major body axes. Some investigators have reported that there is no difference in these thresholds (refs. 5 and 6). Much evidence would suggest, however, that rotation about the z-axis will produce lower thresholds than would rota- tions about the x- or the y-axis (refs. 4 and 7). Only one study has been found that presents even limited information; direct comparisons were made of the perception of angular accelera- tion in the horizontal and vertical planes on the same observers. Meiry (ref. 8) studied the thresh- olds for the perception of angular acceleration for three nortnal men and reported thresholds about the z-axis to vary between 0.1Â°/sec2 and 0.2Â°/sec2, whereas rotation of the head about the x-axis produced thresholds of about 0.5Â°/sec2. Thresholds as high as 8.2Â°/sec2 have been re- ported about the y-axis (ref. 9), but the method involved a very complex task of operating a flight simulator, and no direct comparisons were made for the same observers for rotations about the z-axis. Data obtained by Decher (ref. 10), using nystagmus as an indicator of sensitivity, lend support to these data on the perception of rotation. Decher reported that for nystagmus, the thresholds for the vertical canals were more than twice those of the horizontal canals. On the other hand, one investigator (ref. 11) has written: "According to our experimental results, the two pairs of vertical canals functioning to- gether are much more effective than the hori- zontal canals in the perception of passive rotary motion of the body." Data on differences between the functioning of the horizontal and of the vertical canals are also available from cupulometric studies. Some of these studies support the notion that there are no significant differences in the perception of rotation about the x-, y-, and z-axes. For ex- ample, Benson (ref. 12) found no difference be- tween the "cupulometric thresholds" in the x- and z-axes. Jones et al. (ref. 13) have reported cupulograms which indicate similar "cupulo- metric thresholds" in pitch, roll, and yaw. On the other hand, they found different time con- stants for the duration of the aftereffects of rota- tion about the three major body axes. They reported that the duration of the aftereffects following acceleration was greatest for yaw and least for pitch, with roll in an intermediate posi- tion. Similar results have been reported by Collins and Guedry (ref. 14), Ledoux (ref. 15). and van der Vis (ref. 16). Aschan and Stahle (ref. 17) in a study of pigeons also reported that there were significantly more nystagmus beats and the duration of nystagmus was longer after stimulation of the horizontal canals than after stimulation of the vertical canals. Collins and Guedry (ref. 14) reported similar results for cats and humans. Fluur and Mendel (ref. 4) studied the habituation of the horizontal and vertical semicircular canals and found that it was more difficult to produce habituation of the vertical canals than the horizontal canals as a conse- quence of repeated stimulation. This brief survey of the literature makes it clear that definitive data regarding the sensi- tivity of man to rotation about his three major body axes are lacking. Adequate findings re- garding the sensitivity in these three body planes have implications for the theoretical formulation of the behavior of the semicircular canal system. Furthermore, they have practical implications in connection with aircraft and spacecraft flight. In flying aircraft, rotations about the z-axis are relatively small and infrequent, whereas rotations about the x- and y-axes are much more frequent and more significant for efficient flight (ref. 13). Consequently, it was the purpose of the current investigation to compare the sensitivity of normal men to angular accelerations applied around their x-, y-, and z-axes. METHOD Apparatus The observers were rotated in the Ames man- carrying rotation device (MCRD). The MCRD is a one-degree-of-freedom simulator which has
ANGULAR ACCELERATION ABOUT BODY AXES 301 been described in detail elsewhere (refs. 18 and 19). Accelerations may be produced and meas- ured in 0.01Â°/sec2 steps at the low velocities used in this study. The accelerations are pro- gramed by an analog computer, making it possible to produce changes in acceleration with a rise time of the order of 0.1 second. The simulator is essentially free of vibration that might be per- ceived by observers at the low velocities used during the threshold measurements in this study. Furthermore, the observer is unable to detect when he accelerates through zero veloc- ity. Two special seats were used to place the observers in the proper position. These seats made it possible to position the observer com- fortably in three positions in order to rotate him about the x-, y-, or z-axis of his body while the simulator rotated about an Earth-vertical axis. In each case, his head was positioned at the center of rotation, and his legs were drawn up in a sitting position. For rotation about his z-axis, he sat in a normal, erect, seated position. For rotation about his jc-axis, a horizontal chair was arranged so that he was essentially flat on his back with his legs in a seated position. This produced the same angular acceleration as would be found in a roll. To rotate the observer around his y-axis, he was rotated from the previ- ous position 90Â° so that his right ear was down. Thus, with the simulator rotating about an Earth- vertical axis, the observer was turned around his pitching axis. Observers The observers were 18 men who were in good health by their own affirmation, and a general physical examination revealed no significant abnormalities. They had normal hearing, and their responses to a caloric test were judged to be normal. Procedure The observer sat with his helmet pressed firmly back in a U-shaped headrest to maintain his head in a fixed position for rotation about the x- and z-axes; for the y-axis, the helmet was secured in position. The angular accelerations were presented for 10 seconds for all trials. The direction of the acceleration varied at random from trial to trial, and a minimum of 30 seconds elapsed between the end of one acceleration and the beginning of the next. A single series of approximately 32 trials lasted about 30 minutes, and at least 30 minutes elapsed between sessions. A 3-minute rest period was given halfway through each session. Preliminary practice sessions preceded the regular observations of the percep- tion of rotation at each of the three body positions. The observers were given knowledge of results in the preliminary practice trials, but during the regular trials they had no knowledge of results. All data were collected for each observer at a given body position before data were collected on another. The collection of data for body position, however, was systematically ordered among observers by a Latin-square procedure to reduce sequential effects. All the observations were made in darkness with both eyes closed. The observer's task was merely to indicate the direction of rotation by pressing a switch. The angular accelerations were presented according to a forced-choice, random, double-staircase method which is the same as we have used in previous studies (refs. 18 and 19). The results of this method have been found to correlate closely with a frequency method (ref. 19). Thirty pairs of observations were made following the final level, and the mean of these accelerations was considered to be the threshold for each condition for each observer (refs. 19 and 20). RESULTS The data (table 1) show that, for the 18 observ- ers, the mean thresholds about the x- and z-axes are the same to the second decimal place. The variability about the means is essentially the same for the x- and z-thresholds; the difference is not statistically significant (P>0.20). The mean threshold about the y-axis, however, is substantially greater than the thresholds about the x- and z-axes, but the differences do not quite meet conventional criteria for statistical significance at P = 0.05 (t=1.96 and 2.08, respectively; df= 17). This is in part due to the very great differences among the observers in y-thresholds (table 1). In this regard it is note- worthy that y-thresholds of individual observers were the highest (2.24Â°/sec2) and the lowest (0.06Â°/sec2) of the thresholds. The standard
302 THE ROLE OF THE VESTIBULAR ORGANS IN SPACE EXPLORATION TABLE I. âThresholds1 for the Perception of Rotation About the x-, y-, and z-Axes of the Body for 18 Normal Men [Thresholds are in deg/sec2] Body position â 0-T- 0-C ^ Observer Threshold order Axis of rotation X y z 1 xyz 0.19 0.06 1.04 .18 .82 .64 .12 .61 .22 .58 1.03 2.24 .59 .38 .56 .68 .42 .31 1.50 0.27 .73 .49 .27 .87 .17 .60 .33 .39 .45 .47 .34 .30 .61 .41 .21 .17 .32 2 yzx xzy yxz zxy zyx xyz zxy zyx yzx yxz xzy xyz yzx zxy yxz xzy zyx .43 .36 .40 .17 .20 .33 .51 .45 .32 .55 .45 1.02 .37 .78 .26 .26 .25 3 4 5 6 7 8 9 10 11. 12 13 14 15 16 17 18 Mean thresho Median thresl Standard devi d(A'=18) iold 0.41 .37 .21 0.17-1.02 0.67 .59 .52 0.06-2.24 0.41 .38 .19 0.17-0.87 Kiin(Â£(" Thresholds co Pearson corre x-y + 0.11 xâz -0.06 y-z + 0.26 1 Threshold = mean of accelerations for each condition for each observer based on 30 pairs of observations. deviation was significantly greater about the y-axis than about the*- orz-axis (P < 0.01 in each case). Pearson correlations were also computed between the x-, y-, and z-thresholds (table 1). None of these correlations was found to be sig- nificant (P > 0.20 at N = 18 for each correlation). Illustrations of the marked deviation in thresholds among the three axes are to be found in several observers; e.g., 10, 11, 13. 18. The Latin-square design also made it possible to examine the data to determine whether prac- tice on threshold determination about any two axes would influence the threshold of the third. Thresholds were available for six observers at each of the three body positions taken first, sec- ond, or third. If experience in threshold deter- mination were an important factor, the threshold would be expected to be lower if a particular position were taken second or third in order. An analysis of these effects showed that the order effects were very small and not statistically significant for the x- and z-axes (P > 0.20 in each case). This confirms the lack of practice effects for the z-axis previously reported by Clark and Stewart (ref. 19). The thresholds about the y-axis showed a consistent decrease from the first to the third position. In fact, when the thresholds for the x- and z-axes were both taken before the y-threshold, the mean y-threshold was very near the mean of the x- and z-thresholds. However, these differences were also not sta- tistically significant (P>0.10) for the small number of observers. DISCUSSION This experiment was concerned with the sensi- tivity of normal men to angular accelerations applied about their three major body axes. It should be noted that these results are not directly comparable with earlier studies in which there was an attempt to identify the sensitivity of individual pairs of canals. While this is of interest to the clinician (e.g., in the caloric test) and from an analytical point of view, our concern was simply with thresholds of rotation about the three major body axes. Our results are more directly comparable to observer's standards of reference for judgments of his orientation and motion in everyday life situations, e.g., in walking about and in flying in aircraft or spacecraft, although such activities may be quite complex. Consequently, it is obvious that each of the thresholds reported involves the stimulation of all three pairs of semicircular canals. The mean thresholds for these 18 observers turned out to be the same for the x- and z-axes. and the standard deviations were nearly equal. No definitive data on variability have been found in the literature, but the lack of difference be- tween the mean x- and z-thresholds would appear
ANGULAR ACCELERATION ABOUT BODY AXES 303 to be in agreement with some earlier reports (e.g., ref. 5) and at variance with others (e.g., refs. 8 and 11). It should be noted, however, that the range of thresholds is considerable about both the x- and z-axes even in this relatively small group of observers (table 1) and that with even a smaller number of observers, one might find the mean threshold would fall at any point from about 0.17Â°/sec2 to 1.0Â°/sec2. With larger numbers of observers, one would, of course, expect the range of thresholds to increase somewhat beyond the levels reported here. Indeed, some of our un- published data on about 50 normal men for the z-axis show a range of thresholds very close to that of the y-axis reported above. It is of importance to note that although the mean threshold about the y-axis is substantially greater than the thresholds about the x- and z-axes, both of these differences are just below conventional levels of statistical significance (/) = 0.05). A second consideration here is that comparisons of the threshold levels for each observer show that the y-thresholds are greater in only 12 instances for the 18 observers than the x- and/or z-thresholds. This, too, is below con- ventional levels of significance. A third result is that the individual variability about the y-axis is substantially greater. Finally, the y-thresholds tend to be substantially smaller if the observer has had the x- and z-thresholds determined first, despite the practice trials which, in general, amounted to some 60 to 70 observations before threshold data were counted. What interpreta- tion can be placed on these somewhat ambigu- ous results regarding the thresholds about the y-axis? It is suggested that the observations are more difficult for some observers to make when they are placed on their side. Indeed, several observers reported this. After additional trials, however, this difficulty tended to dis- appear, and the x-, y-, and z-thresholds became much more alike. Consequently, it can probably be said that the stimulus thresholds about the x-, y-, and z-axes are much the same under optimum test conditions. The very low correlations between the thresh- olds about the x-, y-, and z-axes are also note- worthy. These correlations show that no ac- curate prediction can be made from the thresholds customarily determined with observer seated in an erect position as to his sensitivity about other body axes. Consequently, measure- ments about the z-axis can reveal little regarding a particular pilot's sensitivity to the very complex angular accelerations produced in operating an aircraft or spacecraft, and no doubt even less prediction is possible for the complex Coriolis accelerations produced in moving about a rotating room or a rotating space platform (refs. 21 and 22). It should be emphasized that this experiment has been concerned with the perception of rota- tion as a manifestation of a highly complex vestibular system rather than as a hypothetical threshold for the semicircular canals or more specifically, the cupula (ref. 23). For any par- ticular threshold, whether the perception of rotation, the oculogyral illusion, or nystagmus is used as the indicator, the sensitivity tested does not reflect cupular activity in any simple way. It has been well established from many observations that the vestibular system may be stimulated by extremely low angular accelera- tions. For example, with the oculogyral illusion as the indicator about half of the observers have been found to have thresholds below 0.10Â°/sec2. But when the visual target is removed, the thresh- old of the system increases to about three times that value (refs. 1 and 18). It has also been shown repeatedly that the vestibular system interacts in complex ways with other sensory systems (ref. 23). For example, visual targets affect vestibular sensitivity: general level of alertness influences vestibular effects; body tilt influences the perception of the visual ver- tical; and angular acceleration affects visual and auditory localization. With these facts in mind, the findings of this experiment become understandable in terms of a sensory system whose operation is influenced by complex functions and interactions. Thus, the high variability and the possibility of higher thresholds about the y-axis may be considered to be related to the unique tactual and propriocep- tive information in this body position which may lead to different spatial orientation in different observers. These effects may be considered
304 THE ROLE OF THE VESTIBULAR ORGANS IN SPACE EXPLORATION to be comparable to the modification of the observer's perception of the visual vertical and horizontal by changes in body position (ref. 24). The near-zero correlations between the thresh- olds for the x-, y-, and z-axes and the nonsig- nificant decrease in the y-threshold with order of presentation may be understood in the same way. The notion of a complex vestibular system is also supported by our results showing that the x- and z-axes appear to be more alike than the x- and y-axes. On the basis of the orientation of the vertical canals in the head, rotation about the x- and y-axes results primarily in the stimu- lation of these two pairs of canals while the horizontal canals are less affected. The z- thresholds, on the other hand, result primarily from stimulation of the horizontal canals. Con- sequently, this implies that all the processes in the vestibular system must be involved in thresh- old determination rather than the relatively simple characteristics of the transducer mech- anism. This notion would appear to be in general agreement with the point of view developed by Groen (ref. 25) and others which emphasizes the importance of the central nervous system in vestibular processes. REFERENCES 1. CLARK. B.: Thresholds for the Perception of Angular Acceleration in Man. Aerospace Med., vol. 38, 1967, pp. 443-450. 2. LOWENSTEIN, O.; AND SAND, A.: The Individual and Integrated Activity of the Semicircular Canals of the Elasmobranch Labyrinth. J. Physiol., vol. 99, 1940. pp. 89-101. 3. WENDT, R. G.: Vestibular Functions. Handbook of Experimental Psychology, S. S. Stevens, ed., Wiley, 1951. 4. Fl.UUR, E.: AND MENDEL, I..: Habituation, Efference and Vestibular Interplay. IV. Rotary Habituation of the Vertical Semicircular Canals. Acta Oto-Laryngol.. vol. 57, 1964, pp. 459-464. 5. VAN EGMOND, A. A. J.: GROEN, J. J.; AND JONGKEES. L. B. W.: The Function of the Vestibular Organ. Pract. Oto-Rhino-Laryngol., vol. 14, suppl. 2, 1952. 6. DE VRIES, HL.: AND SCHIERBECK, P.: The Minimum Perceptible Angular Velocity Under Various Condi- tions. Pract. Oto-Rhino-Laryngol., vol. 15. 1953, pp. 65-72. 7. DOHLMAN, G. F.: On the Case for Repeal of Ewald's Second Law. Acta Oto-Laryngol., suppl. 159, 1960. pp. 15-24. 8. MEIRY, J. L.: The Vestibular System and Human Dy- namic Space Orientation. NASA CR-628. Oct. 1966. 9. SADOFF. M.: MATTESON. F. H.: AND HAVILL, C. D.: A Method for Evaluating the Loads and Controllability Aspects of the Pitch-up Problem. N AC A RM A55D06. 1955. 10. DECHF.R. H.: Neuees zur Labyrinthphysiulogie: die Drehreizschwellen der vertikalen Bogengange. Arch. Hals.-Nas.-u. Kehlkopfheilk., vol. 181, 1963. pp. 395-407. 11. TRAVIS, R. C.: The Effect of Varying Position of the Head on Voluntary Response to Vestihular Stimulation. J. Exptl. Psychol., vol. 23, 1938. pp. 295-303. 12. BENSON. A. J.: Spatial Disorientation in Flight. A Text- book of Aviation Physiology, J. A. Gillies, ed., Perga mon Press, 1965. 13. JONES, G. M.; BARRY, W.: AND KOWALSKY, N.: Dynamics of the Semicircular Canals Compared in Yaw. Pitch, and Roll. Aerospace Med., vol. 35, 1964, pp. 984-989. 14. COLLINS. W. E.; AND GUEDRY. F. E.. JR.: Duration of Angular Acceleration and Ocular Nystagmus From Cat and Man. I. Responses From the Lateral and the Vertical Canals to Two Separate Durations. Acta Oto-Laryngol.. vol. 64,1967, pp. 373-387. 15. LEDOUX. A.: Les Canaux Semi-circulaires. Etude Elec- trophysiologique; Contribution a I'Effort d'Uniformisa- tion des Epreuves Vestibulaires, Essai d'lnterpreta- tion de la Semiologie Vestibulaire. Acta Otorhinolar.. Belg., vol. 12. 1958, pp. 111-346. 16. VAN DER Vis, K.: Cupulometry of the Vertical Semicir- cular Canals. Pract. Oto-Rhino-Laryngol., voL 20, 1958, pp. 250-260. 17. ASCHAN. G.; AND STAHLE, J.: Cupulometric Studies on the Pigeon. A Comparison Between the Functions of the Horizontal and Vertical Ampullae. Acta Oto- Laryngol.. vol. 46,1956. pp. 91-98. 18. CLARK. B.; AND STEWART, J. D.: Comparison of Sen- sitivity for the Perception of Bodily Rotation and the Oculogyral Illusion. Perception and Psychophysics. vol. 3.1968. pp. 253-256. 19. CLARK. B.: AND STEWART. J. D.: Comparison of Three Methods to Determine Thresholds for the Perception of Passive. Bodily Angular Acceleration. Am. J. Psychol., vol. 81,1968, pp. 207-216. 20. CORNSWEET, T. N.: The Staircase-Method in Psycho- physics. Am. J. Psychol., vol. 75, 1962, pp. 485^91. 21. CLARK, B.: AND GRAYBIEL. A.: Human Performance During Adaptation to Stress in the Pensacola Slow Rotation Room. Aerospace Med., vol. 32. 1961. pp. 93-106.
ANGULAR ACCELERATION ABOUT BODY AXES 305 22. NEWSOM, B. D.; AND BRADY, J. F.: A Comparison of Per- formances Involving Head Rotations About y and z Cranial Axes in a Revolving Space Station Simulator. Aerospace Med., vol. 37, 1966, pp. 1152-1157. 23. GIBSON, J. J.: The Senses Considered as Perceptual Systems. Houghton Mifflin Co., 1966. 24. MILLER, E. F.. II; FREGLY, A. R.; VAN DEN BRINK, G.; DISCUSSION Waite: How were the subjects restrained in this appa- ratus? In the y-axis, where you obtained the highest threshold, was there the least amount of body-surface area touching the device as opposed to the other two axes, as it would appear from your presentation? Clark: We are a little sensitive about safety in rotating devices; therefore, the observers were carefully strapped in, in every case. I would say that there was more weight, let us say, on the back, if that is what you have in mind, when they were lying and rotating around the x-axis. But on the other hand, they were strapped in very firmly, comfortably, but firmly. In the y-position there was a zipper arrangement that held their thighs in position and their feet were clamped into position; they would have to be in order to maintain the position. There was a lot of contact, really. I do not know that I could say precisely how much more. Money: Could some of the variability in the results using the y-axis be accounted for by a difference in threshold for the two directions of rotation? This would fit in very nicely with observations that have been made in dogs, cats, monkeys, and men. in which the vertical nystagmus with the fast component down is much more easily elicited. Clark: Yes. I think that this is a possibility. I have dis- cussed this with Dr. Guedry who has similar notions. I would say yes, possibly, but I have some reservations. On the basis of a large amount of data, I believe that predictions from nystagmus measurements do not necessarily give us prediction for sensation measurements. I do not know that we have adequate data on nystagmus thresholds, especially in normal human beings, but certainly there is a difference between perception of rotation thresholds and oculogyral illusion thresholds, for example. Money: Was the threshold the same in the two directions? Clark: I do not know, but I am going to find out. The reason we did not do so previously was that we tried many right-left comparisons and failed every time. Guedry: You indicated that some people were disturbed when the right side was down. Did you mean to imply that others were not disturbed? AND GRAYBIEL, A.: Visual Localization of the Hori- zontal as a Function of Body Tilt up to Â± 90Â° from Gravi- tational Vertical. Naval School of Aviation Medicine Report NSAM-942, NASA Order No. R-47, Pensacola, Fla., Aug. 1965. 25. GROEN, J. J.: Central Regulation of the Vestibular System. Acta Oio-Larvngol., vol. 59, 1965, pp. 211-218. Clark: Yes. Guedry: Did you then turn them on the other side and did they have comparable difficulties? Clark: No. We have no data for both sides. This might be worth doing. The right-sidedness, I would say, is merely due to their body position itself rather than the particular side down. Lowenatein: Is your appliance capable of coping with rotation in the plane of the anterior-vertical and posterior- vertical canals? Clark: No. It rotates only about the Earth-vertical axis. Lowenstein: Could you orientate it? Clark: No. We could not. Lowenstein: It would be extremely interesting to see what the thresholds would be in those positions. Steer: Some measurements made by Dr. Meiry in our own laboratory a few years ago were done for the y-axis with the head tilted to the side rather than the entire trunk tilted to the side. In that case he found, with a very small number of samples, that there was no real significant difference in thresh- old between the horizontal and lateral canals. I would agree that it is extra information from the body being tilted that is causing this additional delay. Clark: Yes. For four subjects in a rather limited number of trials, we measured thresholds in different body positions, one on the back, and then with the subject seated erect and leaning forward with the nose down. We did not test it on the side. The threshold with the nose up was 0.42Â°/sec2; with the nose down, it dropped to 0.23Â°/sec2. This would apparently be similar to the findings by Dr. Meiry. We then turned them over so that the head was in the same position, and we found the same threshold again, which complicates the matter, I am afraid. This may very well be related to how the subjects are restrained. Interestingly enough, when we repeated this a second time on these four subjects, the threshold came out to be 0.44Â°/sec2 compared to 0.42Â°/sec2. Unfortunately, these data were not properly counterbalanced, and so on. We just made a sort of trial run, but it is in- teresting to speculate on what this might turn out to mean.