^{page 257 note *} When rotation has continued for some time, friction of the periosteum of the bony canals against the perilymph, and fluid friction in the perilymph, gives to the perilymph, and, of course, also to the membranous canal, the same rotation as the bony canal has; the perception of rotation will thus cease. If we now stop the rotation of the head the bony canal stops, but the perilymph and the membranous canal move on, and a pull takes place at the opposite ends of the semicircular canals, causing a perception of rotation round the same axis in the opposite direction.

The members of the three pairs of semicircular canals are not always accurately parallel to each other, and in some animals the three axes are not accurately at right angles, so that in the most general case we have two systems of co-ordinates, not necessarily rectangular, which we may call x, y, z, and ξ, η, ϼ, — each of these six axes having an organ capable of being influenced by rotation about the axis in one direction. But in all cases, as far as I know, these six axes and the corresponding organs are so placed that a different set of impressions will bo produced by each form of rotation, that is, by each combination of axis, direction, and rate.