able to gain a more distinct vision of them, we turn our eyes back, and again and again pursue the flying shade.

But this rolling of the eyes, after revolving till we become vertiginous, cannot cause the apparent circumgyration of objects, in a direction contrary to that in which we have been revolving, for the following reasons. 1.  Because in pursuing a spectrum in the sky, or on the ground, as above mentioned, we perceive no retrograde motions of objects. 2.  Because the apparent retrograde motions of objects, when we have revolved till we are vertiginous, continues much longer than the rolling of the eyes above described.

3.  When we have revolved from right to left, the apparent motion of objects, when we stop, is from left to right; and when we have revolved from left to right, the apparent circulation of objects is from right to left; yet in both these cases the eyes of the revolver are seen equally to roll forwards and backwards.

4.  Because this rolling of the eyes backwards and forwards takes place during our revolving, as may be perceived by the hand lightly pressed on the closed eyelids, and therefore exists before the effect ascribed to it.

And fifthly, I now come to relate an experiment, in which the rolling of the eyes does not take place at all after revolving, and yet the vertigo is more distressing than in the situations above mentioned.  If any one looks steadily at a spot in the ceiling over his head, or indeed at his own finger held up high over his head, and in that situation turns round till he becomes giddy; and then stops, and looks horizontally; he now finds, that the apparent rotation of objects is from above downwards, or from below upwards; that is, that the apparent circulation of objects is now vertical instead of horizontal, making part of a circle round the axis of his eye; and this without any rolling of his eyeballs.  The reason of there being no rolling of the eyeballs, perceived after this experiment, is, because the images of objects are formed in rotation round the axis of the eye, and not from one side to the other of the axis of it; so that, as the eyeball has not power to turn in its socket round its own axis, it cannot follow the apparent motions of these evanescent spectra, either before or after the body is at rest.  From all which arguments it is manifest, that these apparent retrograde gyrations of objects are not caused by the rolling of the eyeballs; first, because no apparent retrogression of objects is observed in other rollings of the eyes:  secondly, because the apparent retrogression of objects continues many seconds after the rolling of the eyeballs ceases.  Thirdly, because the apparent retrogression of objects is sometimes one way, and sometimes another, yet the rolling of the eyeballs is the same.  Fourthly, because the rolling of the eyeballs exists before the apparent retrograde