This is necessary because the image of the spectrum of a star on the photograph is only a thin line in which it is impossible to see the spectral lines; the spectrum must be broadened, and this is accomplished by making the star image “trail” to a certain degree on the plate. This trailing is accomplished by means of the clock, the rate of which is made to vary. In this way the trail of a spectrum of a star on the photographic plate is always obtained of the same width, while the density of the image is made fairly constant by increasing the rate for bright stars and decreasing it for fainter ones. In this way spectra of the brighter stars rivalling in perfection and detail those obtained of the spectrum of the sun itself thirty years ago have been obtained. Such photographs have rendered a minute chemical classification of the stars possible. One of the most interesting applications of photography to spectrum analysis during the latter part of the century has been the utilization by Messrs. Deslandres and Hale of a suggestion made by Janssen, that by employing photography images of the sun and its surroundings can be obtained in light on one wave length. In this way we can study the distribution of any one of the chemical constituents of the sun separately, and note its behavior, not only on the sun itself, but in the atmosphere which enfolds the disk. It is strange that, in spite of the suggestions of Faye, and others after him, one of the great advantages of the employment of photography in astronomical work, namely, the abolition of “personal equation,” has so far been almost entirely neglected. What “personal equation” is can be perhaps illustrated by considering an observer who is observing the transit of a star over the wires in a transit instrument. His object is to note the exact time, to a fraction of a second, when a star passes each wire; and this is done by listening to the beats of a clock near at hand and estimating the fractions. Some observers constantly note the time either a little in advance or a little later than the actual time, and this small distance between the observer and the true times is more or less constant for each observer. This difference has to be taken into account for every observation. Even the use of the chronograph in transit work, by which the observation is electrically recorded, does not entirely eliminate the error. The photographic method of transit work has been experimented on, but, so far as I know, it has not yet been used at more than one or two observatories. It will doubtless eventually rid us of “personal equation” entirely, for the star image may be photographed and the time recorded by the same current of electricity.