The sac lies in a somewhat irregular cavity in a bone at the side of the head, in the walls of which are five openings leading into curved tunnels in the bone in which lie the membranous loops. The planes in which the three semicircular canals lie are nearly at right angles to each other, and they are called respectively the horizontal, the superior, and the posterior. The two latter unite at one end before they reach the sac; hence there are five, and not six, openings into the cavity. At one end of each semicircular canal is a swelling, or ampulla, in each of which is a ridge, or crest, abundantly supplied with hair-cells. And in a little recess in the sac there is, occupying its floor, its front wall, and part of its outer wall, a patch of hair-cells covered by a gelatinous material with numerous small crystalline otoliths. The only other point that calls for notice is that the membranous sac does not fit closely in the bony cavity in which it lies, while the diameter of the membranous semicircular canals is considerably less than that of their bony tunnels, except at the ampullæ, or swellings, where they fit pretty closely. Both the bony cavity and the membranous labyrinth (as it is called) are filled with fluid. From its close connection with the organ of hearing, this apparatus was for long regarded as in some way auditory in its function, and it was surmised that it enabled us to perceive the direction from which the sound came. But how it could do so was not clear. In 1820 M. Flourens made the observation that the injury or division of a membranous canal gave rise in the patient to rotatory movements of the animal round an axis at right angles to the plane of the divided canal; and he, therefore, suggested that the canals might be concerned in the co-ordination of movement. They are now regarded as the organs of a sense of rotation or acceleration. That we have such a sense of rotation has been proved experimentally.[FA] Let a man, blindfolded, sit on a smooth-running turn-table. When it begins to rotate he feels that he is being moved round, but if the rotation be continued at the same rate, this feeling quickly dies away. If the rotation be increased, he again feels as if he were being moved round, but this again soon dies away. Further increase gives a fresh sensation, which in turn subsides, and the man may then be spinning round rapidly, and be perfectly unconscious of the fact. He is only aware that he has been gently turned round a little two or three times. Now let the speed of rotation be slackened. He has a sensation of being gently turned round a little in the opposite direction. Each time the speed is lessened he has this sense of being turned the reverse way. From these experiments we see that what we are conscious of is change of rate of rotation, or, in technical language, acceleration, positive or negative. [Illustration: Fig. 30.--Diagram of semicircular canals.