ordinates are radii of concentric circles the curvature of the base-line introduces no error in the records. With sufficiently numerous concentric circles for comparison it appears to be just as easy to determine where a curve rises from a circular base-line as it would be to determine where it rises from a straight line.

A DEVICE FOR CONTROLLING THE TIME OF EXPOSURE IN THE DODGE TACHISTOSCOPE

BY H. R. CROSLAND

University of Oregon

The carrying through of several projects directed at the legibility of printed characters has necessitated the supplanting of the Whipple tachistoscope with a vertical type of Dodge mirror-tachistoscope. This latter apparatus, however, lacks a device for the controlling of the time of exposure, problem the solution of which with the Whipple tachistoscope had been effected by a gravity-propelled disc carrying a calibrated aperture through which light passes to the exposure field for a fraction of a second. Having made the change from the Whipple to the Dodge apparatus, we set about working out a device for time-control. Correspondence with Professor Dodge himself, who has had much experience with such apparatus, furnished us with several designs of shutter; one of these was tentatively selected and efforts at construction were begun. The type selected, however, presented several difficulties to the mechanician who was engaged to carry out the plans, and gave promise of presenting difficulties after construction with reference to the shutter's passage between the source of light and the frosted windows of the tachistoscope. Such circumstances forced the author to attempt to construct an apparatus which, by making and breaking the light current, would obviate the objections to the shutter arrangement and would give us light exposures for fractions of a second. Consequently, he designed a rotating disc, much like Whipple's, but of 3-ply veneer wood, inch thick, to carry copper contact

'Mr. Frank Ackley, of the Chemistry Department of the University of Oregon, deserves credit for the fine workmanship exhibited in the apparatus. The drawings were made by Mr. Linn Forrest, of the School of Architecture. That 3-ply veneer wood could conveniently serve as our rotating-disc was suggested by Mr. T. D. Cutsforth, of the Psychology Department. Help in calibrating the time of exposure of the light itself and in controlling voltage has come from the Physics Department.

plates, so arranged that the lengths of the plates, wired through the disc to a continuous plate, would give a light through each of the two windows of the tachistoscope for exactly the sigma of time required. This disc is actuated in its rotation by weights on lever-arms, as in the Whipple apparatus; and the principle of contact-making has been borrowed

from a form of brush-contact common to several pieces of apparatus no one of which was specifically in mind when these plans were formulated.

The three plates which accompany this exposition will serve to make clear the parts of the apparatus and their arrangement.

A-N represents a wooden standard which rests upon a Tshaped base which screws firmly to the top of a table, A being

the upright and N the base. At B is located a non-sliding but rotating screw-socket holding the meter-stick lever-arms E, F in position. Behind B is the axle of the disc, made of handmilled steel as are also the bearings, giving a minimum of play or lost motion or untrue rotation of the disc. The leverarms are held by fulcrum B; and C and D are adjustable weights with center screws holding them tightly in position. K is the rotating wooden disc. H and G are the copper contact-plates, which, in the fall of the lever and the rotation of the disc, engage the metal spring contacts L and M. The latter, L and M, are wired in circuit with one of the wires (W1) from a wall-socket furnishing 125-v a.c. energy, whereas H and G are wired, through the wooden disc, with metal-plate Q (see rear view, below), which is a continuous band always in contact with the spring P. The latter, P, is supplied with the other wire (W) from the wall-circuit. R and S represent the two 25-watt tungsten lamps which are placed in the two separated compartments leading to the two frosted windows. in the rear of the Dodge tachistoscope. J represents a rubberband which gives leverage to P against the contact-plate.

The rear and side plates, shown herewith, present the complete circuits of wiring, which are so arranged, with their contact-plates and spring-contacts, that the light in each compartment goes on, with a part-rotation of the disc, for 100 sigma, and just as it goes off, the second light, in the other compartment, goes on for 100 sigma.

Two possible sources of error, from this arrangement, might arise from two sets of inertias, one for the metal contacts and the other for the light-bulbs themselves since it is wellknown that the initial lag of a bulb is greater than its terminal or final lag. Two procedures have been followed to eliminate these sources of error. One of these consists in using nitrogenfilled bulbs, and the other consists in calibrating the exposuretime of the light itself directly at the eye-piece of the tachistoscope by filming the shadow of a hair supported by one tyne of a 100 d.v. tuning fork. Adjustments and corrections are experimentally instituted with respect to the lengths of the copper-plates, G and H, as a consequence of the measurement