A second method, and the one employed on this occasion, was that of pressing into our service an old turnstile counting apparatus. This was attached to the axis of the revolving sec tor, and its index watched for a set time, say io seconds, the number read off and divided by to to give the number of revolutions per second.
Behind the sector wheel is seen the lens and camera arranged in the same plane with the other parts of the apparatus, having a special chamber in the rear, holding a metal drum five inches in diameter, extending transversely across the interior of the camera to its full width. The drum turns about a horizontal axis that is parallel to the slit in the spectroscope tube and perpen dicular to the optic axis of the line of the apparatus. One end of the axis projecting through the side of the camera has a pulley disk on the end. The cylinder may be seen through the broken portion of the camera. The camera lens is adjusted to throw a sharp image of the slit in the spectroscope, upon the center of the surface of the cylinder in the camera. The cylinder is cov ered with a strip of sensitive (e. g., bromide) paper held in position by elastic bands.
In place of the cylinder of bromide paper a circular sensitized glass plate can be fixed in a vertical plane in the back of the camera and made to rotate or whirl at a given speed, the flashes of light through the slit impressing it in the shape of radial lines, their number determining the rapidity of the shutter.
If now light from the lantern passes through the slit lens, etc., on to the front of the re volving paper on the drum, it would trace on the paper a rectangle image of the same width and length as the slit, the vertical length depending upon the rate of the drum revolutions, i. e., how much paper turned past the image of the slit. If the sector be set revolving, every time a spoke or bar came in front of the lens it would cut off the light while it was passing. On developing the paper we should have a dark rectangle crossed by bars of light corresponding to the transits of the spokes. Suppose now the shutter to be of that form which causes a rectangular opening to pass in front of the lens. As the beginning of the opening was commencing to travel across the slit we should get an image of a portion only of the slit formed on the front of the revolving drum, and similarly as the opening commenced to close it would cut off more and more light from the slit, and so on, the image dwindling from a line to a point.
In Fig. 403 we have some such result. If we suppose the drum and sector to be stationary, we should getan image of the slit as a straight line as H2. If the drum revolved and the sector was stationary, this straight line would be drawn out into a rectangle. If the sectors re volved, we should find this interspaced with clear parts, but suppose the shutters to commence slowly opening at the end H„ and go on until the whole slit H, H 2 were fully covered by light for a time and then begin to close up again from H, toward the sector and drum revolving at the time, we should have some such figure as N M K L. The triangular part, M M' N, corre sponding to the time the shutter took to get fully open, the rectangular part, M M' L L', being the time that it remained fully open, and the triangular part, L L' K, the period of closing. On fur ther examing this diagram we note two complete bar spaces, and a little portion outside each, together equal to about that between two clear spaces. If now the sector were revolving at the rate above supposed and described, i. e., second between each bar space, we should say that the shutter took then something between two and three such intervals to open, say to sec ond to get fully open ; remained fully open about two spaces, i. e., say, and closed in a little more than one space, to second. The ideal or theoretical perfect shutter is one which takes no time to become fully open and to close as quickly, but as this is as yet not a practical thing, we have to accept this as a mental conception, and compare the actual performance of any shutter with it by contrasting its practical efficiency with its theoretically possible efficiency ; we thus need but compare the area, M M' L L', i. e., the period of full open ing, with the corresponding area, N N' K K', i, e., the base, N K', with base, L M'.