In the following research the albedo of the earth has been determined from visual photometric measurements of the earth-shine on the moon, in comparison with light from similar sun-illumined areas on the moon's surface. T h e photometer, shown in plan in Fig. r a and in front'elevation i n I b , consists of a rectangular wooden case, having a f n l projecting shelf at S on which a lantern is placed, whose flame (furnishing the standard light) is viewed through an iris-diaphragm (I) by means of

more »

... otally reflecting prisms (PI P,) and a positive eye-piece ( E ) . At L is an achromatic lens of 34.9 cm focal length and 1.94 cm effective aperture, which, being directed towards the moon, gives an image of that object, or of the neighboring sky at M (after reflexion by the prisms P s P.). A t the common focal point of the objective and the positive eye-piece, are twin circular apertures in black card, each subtending 7' of arc on the celestial sphere, one above, and the other below the horizontal line of junction of the last pair of reflecting prisms. On removing the black card, the upper half of the field is seen occupied by the moon and surrounding sky, while the lo'wer half of the field is uniformly illuminated by light from the flame, variously modified. It is seldom necessary to point the instrument at other than moderate elevations, since the earth-shine is best observed not far from new moon and therefore the moon's altitude is low. The instrument, as constructed, is not suitable for observations at large altitudes, but could easily be adapted to such work by the addition of a diagonal mirror in front of the iris, placing the lantern at one side on a rocking frame. The ideal illumination on the right hand (as seen from behind) would have been that of a uniformly illuminated ground-glass disk whose luminosity could have been modified by the iris-diaphragm in exact proportion to the area of its aperture; but the exigencies of the work demanded a wider range of light. In some of the measures it was necessary to dispense with any ground-glass screen and to receive the light from the unobstructed flame through various colored glasses. Since the distribution of light in the flame is not uniform and the diffusion of its light not the same as that from the screen, the brightness of the lower and flame-illumined half-field is then no longer proportional to the area of the aperture, and it becomes necessary to calibrate the instrument and to prepare curves of illumination corresponding to the opening of the iris by which the readings must be evaluatedcurves which differ essentially from the typical iriscurve. A further need of several such curves of calibration will be seen when we consider the problems connected with the measurement of very faint light which also possesses distinct color. The plan finally adopted was to place one or tnore modifying glasses at G, and if necessary a plate of parallel and plane-surfaced, neutral-tinted glass could be inserted in the path of the r$ys from the lens at GI, while in some cases the moonlight was cut down by diaphragming the objective. The final adjustment to equality in the two small illuminated circles (viewed in the center of the field as if they constituted the components of a double star) was made by opening or closing the iris-diaphragm which was graduated into fifteen divisions and read by estimation to tenths of a division. The lunar measures were,made easily with this instrument. Whatever discrepancies appeared were usually readily explicable, and resulted mainly from the inevitable variations of atmospheric transparency, rather than from any Haw in the apparatus, or in the method. Some difficulties were encountered in obtaining consistent and reliable measurements of the transmission-coefficients for the differently colored glasses, but these were finally overcome. After trying various arrangements, including a flicker-photometer, the most satisfactory method for measuring the very wide range of intensities included in the several transtnitting pieces was found in a modification of the wedge-photometer, due to Dr. Charlrs H. Wi//iattis, of Boston, in which a long strip 2 0