If the object moves from =X= to =O=, Fig. 9, in the direction of the radius, all the images will likewise move towards =O=, and the patterns will have the appearance of being absorbed or extinguished in the centre. If the motion of the object is from =O= to =X=, the images will also move outwards in the direction of the radii, and the pattern will appear to develop itself from the centre =O=, and to be lost or absorbed at the circumference of the luminous field. The objects that move parallel to =X O= will have their centre of development, or their centre of absorption, at the point in the lines =A O=, =B O=, _a_ =O=, _b_ =O=, etc. where the direction in which the images move cuts these lines. When the object passes across the field in a circle concentric with =A B=, and in the direction =A B=, the images in all the sectors formed by an even number of reflexions will move in the same direction =A B=, namely, in the direction β _b_, _a_ α; while those that have been formed by an odd number of reflexions will move in an opposite direction, namely, in the directions _a_ =B=, =A= _b_. Hence, if the object moves from =A= to =B=, the points of absorption will be in the lines =B O=, α =O=, and _b_ =O=, and the points of development in the lines =A O=, _a_ =O=, and β =O=, and _vice versa_, when the motion of the object is from =B= to =A=. If the object moves in an oblique direction _m n_, the images will move in the directions _m t_, _o n_, _o p_, _q t_, _q p_, and _m_, _o_, _q_, will be the centres of development, and _n_, _p_, _t_, the centres of absorption; whereas, if the object moves from _n_ to _m_, these centres will be interchanged. These results are susceptible of the simplest demonstration, by supposing the object in one or two successive points of its path _m n_, and considering that the image must be formed at points similarly situated behind the mirrors; the line passing through these points will be the path of the image, and the order in which the images succeed each other will give the direction of their motion. Hence, we may conclude in general, 1. That when the path of the object cuts both the mirrors =A O= and =B O= like _m n_, the centre of absorption will be in the radius passing through the section of the mirror to which the object moves, and in every alternate radius; and that the centre of development will be in the radius passing through the section of the mirror from which the object moves, and in all the alternate radii: and, 2. That when the path of the object cuts any one of the mirrors and the circumference of the circular field, the centre of absorption will be in all the radii which separate the sectors, and the centre of development in the circumference of the field, if the motion is towards the mirror, but _vice versa_ if the motion is towards the circumference.