26 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



cular aperture 4.86 millimeters in diameter, in the center of a cap covering the object-glass of a small 

 telescope, D, of about 520 millimeters focus. On leaving the eye-piece of this telescope the sunlight 

 is spread out into a diverging cone of rays, which at the distance of the photometer slit *S', 2610 

 millimeters beyond the eye-piece, has a diameter of 652 millimeters. Its intensity has therefore 

 been weakened about 18,200 times (independently of the absorption of the glass, wliich is not a 

 a factor in our qualitative determination). 



S is the slit of a grating spectroscope. The collimator G has a focal length of 1254 millimeters 

 and an aperture of 57 millimeters. The observing telescope T is much shorter, having a focal 

 length of but 400 millimeters (in order tbat the head of the observer may not interfere), and is set 

 at a fixed angle of about 49° to the collimator. A holder within the case at G carries a flat Row- 

 land grating with a" ruled surface 51.6 by 35.0 millimeters, and with the number of lines per milli- 

 meter equal to 568.4. Tliis grating, which gives very brilliant and very perfect spectra, was used 

 in such a position that the normal to its surface fell between the two telescopes, the comparisons 

 being made in the brighter first spectrum. Its angular position is indicated by a divided circle and 

 vernier, reading to minutes on the outside of the case. 



The lower part of the photometer slit is covered by a totally reflecting prism, P, which cuts 

 off the sunlight entering there and substitutes for it the light from the standard lamp, L. Two 

 spectra in close juxtaposition are therefore seen in the eye-piece of T, the upper belonging to the 

 lamp and the lower to the sun. By means of a 2-millimeter blackened cardboard slit in the com- 

 mon focus of the object-glass and eyepiece, the range of wave lengths included in the field of view 

 was limited to 0''.0048, or about eight times the interval between the 1> lines. 



The lamp has already been partially described. It was fastened to a slider, which could be 

 drawn by the observer to and fro along a graduated scale, at right angles to the collimator so as 

 to approach or recede from the slit, by pnlling a cord. A heavy screen which hung down to the 

 level of the photometer scale concealed the lamp from the observer, who was thus unaware of its 

 position while making a measurement (excei>tfrom the ajipearance of its spectrum in the eye-piece 

 of the telescope) until the index had been read, and thus any bias resulting from a preconceived 

 opinion as to the proper position of the lamp was avoided. The range of the scale was 20 deci- 

 meters, and its zero-point was so adjusted that the reading of the index of the lamp-carriage was 

 the distance of the center of the flame from the slit of the photometer. On account of the great 

 difference in the quality of the lights compared, the range of the scale proved to be insufiBcient, 

 and the wheel photometer, an instrament i)resently to be described, was used to diminish the more 

 intense light by a given ratio. 



When moonlight, instead of sunlight, was compared with the standard, the diminishing tele- 

 scope was removed, and a telescope of 1,054 millimeters focus and 77 millimeters aperture, with the 

 eye-piece removed, was placed on the axis of the beam from the siderostat, so as to form an image 

 of the moon on the upper half of the photometer slit. 



The wheel-photometer, referred to above, consists of two circular disks of sheet-zinc about 

 twenty inches ia diameter, each pierced near the circumference by eighteen radial apertures 

 separated by spaces of the same width. The two disks may be rotated past each other with con. 

 siderable friction, enough to hold them firmly in relative position, and are held by an axis passing 

 through their centers, by means of which, and a multiplying wheel connected with it, they may 

 be rotated in a vert.cal plane with great velocity as a single wheel. If they are adjusted to coin- 

 cide, and rotated in front of a source of light, they diminish its brilliancy one-half, although, on ac- 

 count of the persistency of vision, the eye does not perceive any flickering or unsteadiness caused 

 by the interruptions of the spokes. A graduated arc is attached to one of the disks, and an 

 index to the other, so that the apertures may be adjusted to any width from the full opening 

 down to zero. Thus the intensity of a luminous source may be diminished to any fraction less than 

 one-half of its original value. 



On looking into the eye-piece of this apparatus (Plate 2) two nearly square patches of light 

 were s^eii, the lower belonging to the sun or moon, and the upper to the lamp. The color of the 

 light would depend, of course, upon the position of the grating. The observations were made 



