June 20, 1895] 



NATURE 



18; 



ii|i|)roximately equal angles with the mesial plane. In all cases 

 which I have examined, the angle between the posterior canal 

 and the mesial plane is somewhat larger than that between the 

 superior canal and the mesial plane. 



From the bilateral symmetry, therefore, the superior canal of 

 Ihe one side is nearly, but not quite, parallel to the posterior 

 canal of the other side. In the discussion of the way in which 

 the system of canals may Ije supposed to act, I shall for con- 

 venience assume that these canals are parallel, as the deviation 

 from exact parallelism only complicates, but does not at all 

 vitiate, the argument. 



(4) In man, and in a large number of other animals, the three 

 canals are very nearly at right angles to one another. But, in a 

 good many of the animals I have looked at, the superior and 

 posterior canals make with one another an angle considerably 

 greater than a right angle. 



Looking at the six canals as forming one system, we see that 

 we have three axes, that at right angles to each axis 

 there are two canals, one in the one internal ear, the other 

 in the other ; these two canals having their ampulla; at 

 opposite ends, so that if rotation takes place about the 

 axis, the ampulla in the one case precedes the canal, in 

 the other follows it. The vertical axis, as we may call that at 

 right angles to the two external (or horizontal) canals, is pretty 

 nearly vertical in most animals, in the usual position of the head 

 when the animal looks to the horizon ; in man it is lot exactly 

 so, we must bow our head a little to make this axis vertical. If 

 we suppose we are looking north, the other two axes are north- 

 east and south-west and north-west and south-east respectively. 

 In man they pass from the eye of one side to the mastoid ])ro- 

 cess of the other side, and are nearly at right angles to one 

 another. As already stated, in some animals they are inclined 

 and are nearer the right and left than the fore and aft line in the 

 head. 



In order to see how such a system can work as a hydro- 

 dynamical instrument, let us first consider one canal. 



Here we have two watery liquids, the endolymph within the 

 membrantfus canal, its amj^ulla and the utricle, the perilymph 

 between these and the bony case. How will these behave when 

 rotation takes place about an axis normal to the plane of the 

 canal ? The inertia of the liquids will tend to produce a flow 

 through the canal in the sense opjxisite to that of the rotation. 



Let the rotation take place so that the ampulla precedes the 

 canal. Here the endolymph will tend to flow from the utricle 

 into the .ampulla, and thence through the canal to the utricle 

 again. But, as Mach has pointed out, the canal has too small 

 a bore to allow of any sensible flow through it, so that the 

 effect of this rotation will be to increase the pressure within the 

 membranous ampulla. But (and this is a point to which, as far 

 a,s I kn(jw, no one has hitherto called attention) as there will 

 also he a tendency of the pcrilynq)h to circulate, so in its 

 circle there is also a narrow place, namely at the ampulla : 

 for as the membranous ampulla nearly fills its bony case, there 

 is not nmch room there for the perilymph to pass from the 

 vestibule into the space .surrounding the membranous canal. 

 There will, therefore, be a diminution of pressure of perilynq>h 

 at the ampullary end of the canal, so that the ampullary walls 

 will be stretched by the increase of pressure within and the 

 diminution of pressure withtmt. Of course when the rotation 

 is kept up uniformly for some lime the pressure inside and 

 outside of the membranous ampulla is soon equalised, and the 

 stretching or relaxation ceases. With the cessation of the 

 stretching the sensation must also cease. 



If now the rotation is stopped the |ierilympli and endolymph 

 will lend to move on, .and pressure will be produced inside the 

 membranous anqiulla of that canal, which during the rotation 

 moved with anqiulla following the canal. 



All this will of course be reversed when the rotation takes 

 place with the ampulla following the canal ; the pressure inside 

 the membranous auqmlla will be diminished, that without 

 increased, and the walls will become flaccid. 



In each membranous ampulla there is a so-called crista acustica 

 where nerves terminate in hair-cells, and it is not difficult to 

 sup|Kise that stretching of the ampullar)' walls will irritate these 

 nerve-endings, while a relaxation of the ampullary walls will 

 I>roduce no irritation. If this be so, then we have three axes 

 each with an organ sensitive to rotation about it in either sense, 

 and caixxble of discriminating between the two ; and as every 

 rotation of the head can be resolved into component rotations 

 about these three axes, we have the means of perceiving the 



NO. 1338, VOL. 52] 



axis and what we may call the intensity of the rotation, or 

 perhaps more correctly the rotational acceleration. 



This hydrokinetic theory of the function of the semicircular 

 canals was propoun<led at very nearly the same time by I'rof. 

 -Mach of Prague, Dr. Breuer of Vienna, and myself. I give 

 the names in the order of publicali<m. The views expressed 

 by us were not exactly the same, and the statement of the 

 theory I have just given is any one of them with additions and 

 corrections from the other two. 



I have not thought it necessary to refer to the hydrostatic 

 theor)' of Goltz, or, indeed, to give any details of the literature of 

 the subject. .\ very full and accurate digest of almost every 

 thing that has been written on the functions of the several 

 parts of the labyrinth of the ear has been published in Russian 

 by Dr. Stanislaus vcm .Stein, and translated into German by Dr. 

 C. von Krzywicki. 



The theory as I have just described it might perhaps have 

 been developed, as I have here developed it, from a considera- 

 tion of the strvicture and position of the canals. But, as a 

 matter of fact, this was not the historical order. It was the 

 experiments of Flcurens that first directed attention to these 

 organs as having something to do with the equilibrium of the 

 body. 



In reference to these experiments and those made since by 

 many able physiologists and skilled operators, I shall only say 

 that the results seem to me to be consistent with the hydro- 

 kinetic theory. Certain of de Cyon's experiments, in which he 

 increased the pressure in the canals by inserting in them small 

 tangle plugs without producing any nystagmus or rotatory 

 movements of the head, appear to contradict the theory. But 

 increase of pressure in the bony canal can have no tendency to 

 stretch the walls of the membranous ampulla, and therefore 

 could not be expected, if the theory as I have stated it is correct, 

 to produce a sensation of rotation ; what is required, is that the 

 pressure inside the membranous ampulla should be greater than 

 that outside of it. 



The symptoms observed in cases of disease of the internal ear 

 also appear to support this hydrokinetic theor)-. 



But the position of the canals in close anatomical relation to 

 the organ of hearing had impressed on the minds of physiologists 

 so obstinate an opinion that they must be connected with the 

 perception of sound in some way or other, that even now many 

 will not admit that they are the peripheral organs of a sense 

 of rotation. 



A favourite theory was (and there are still some who hold it) 

 that the semicircular canals give us information as to the direction 

 in which sovmd comes to us. There are two ways in which we 

 can show that this view is erroneous. 



(ll By considering the physical conditions. 



The shortest soimd wave which we can hear is so long com- 

 pared with the dimension of the ear, that every part of the ear 

 nnist be at any instant in the same phase of the wave. We must 

 assume that, as far as Ihe effect of such sound waves is concerned, 

 the liquid contents of the internal ear are inconqiressible. It is 

 as absurd to speak of sound-waves travelling round one of the 

 canals as to say that it is high water at one end of a dock and 

 low water at the other, at the same time. 



(2) By experiments on the way in which we really do perceive 

 the direction of sound. I shall describe two such experiments. 

 {a) Let the observer close his eyes — for security it is best to 

 bandage them — seat liimself in a chair, and keep his head steady. 

 Now let an assistant prodiu:e a sharp short sound. In showing 

 this experiment to Section I) of the British A.ssociation, at its 

 meeting at Belfast in 1874, I used three coins in the w.ay I show 

 you now. The observer can tell with really astonishing ac- 

 curacy whether the sound comes from the right or fri>m the left, 

 because he hears it louder in the nearer ear, but he is without 

 any knowledge at ail as to whetheril comes from above or below, 

 from the front or the back. Me forms a judgment indeed on 

 this point, but his judgment is usually wrong, often very 

 ludicrously so. 



The experiment is most striking when the click is produced in 

 the mesial plane of his head, in which case he has not the binaural 

 eflect to ln'ip him. In this connection I may s;iy that I know no 

 experimeiU which illustrates so well the marvellous ileltcacy of 

 our sense of relative loudness of sound, a very small deviation 

 from the mesial plane being quite certainly recognised. 



We have then with one ear no means of ascertaining the 

 direction of sound if we keep the head fixed. How then do 

 we ascertain the direction of sound ? for we all know that we 



