18 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



obtain with a comparatively small mirrer, concentrating umch less heat than that dealt with by 

 the 3-tbot reflector of Parsoustown, much more consistent indications. 



Tiie lunar rays, reflected from the IL'-inch silvered mirror of a large siderostat, pass horizon 

 tally through an 8-inch circular ajierture in the north wall of the observatory dark chamber, and 

 fall upon a 10-iuch concave silvered glass mirror of abont .'?() inches focus, mounted on a very solid 

 tripod stand.* The bolometer, in a case specially designed for this work, is mounted on a sliding 

 carriage directed toward the center of the mirror, so that it can be adjusted to such a distance as 

 to bring the working face of the instrument into the plane of the lunar image. The bolometer 

 case referred to has a series of circular dia]ihragms of various ajtertures so disposed as to protect 

 the bolometer itself both from air currents and extraneous radiations, while just admitting the 

 cone of rays from the concave mirror. 



This mirror is inclined slightly to the incident rays, so that the bolometer can be placed a little 

 to one side of the Inrgi- ajierture in the wall and not obstruct them. The hniar image can theu be 

 adjusted by means of the uiillcd headed screws, by which the mirror is secured to the vertical plate, 

 so as to fall truly on tliM bolometer strips before observation, and afterward in actual observation 

 be cariied on or off by a motion of the sideiostat niirror outside, or by the nuxni's own motion in 

 the heavens. In either case no screen is interposed, and no alteration in the relation of the radiat. 

 ing objects around takes place in reference to the bolometer, which experiences no changes, except 

 those which come from its alternate exposure to the moon and the neighlioring sky. 



The lunar iuuige is about O.liO inch in diameter, and when properly directed is received by 

 the workiugfaceof the bolometer, which it very nearly covers; hence it will be seen that (neglect- 

 ing the absoriition of the mirror, which is very small lor invisible heat rays), remembering that only 

 eight inches of the mirror's diameter is utilized, the intensity of the lunar heat was increased from 

 780 to 1,050 times, according as the distance of the moon from the earth varied. On a clear night 

 with a full moon, and the galvanometer in its condition of then greatest sensitiveness, a deflection 

 of 300 millimeter divisions of its scale could be obtained (in 1883), but toward the close of 1884, with 

 still further improvements in the apjiaratus, this limit was much exceeded. 



The "exposures" in later and adopted measures were only made, as we have said, by moviur 

 the image of the moon on and off the strijts of the bolometer, by slightly inclining the siderostav 

 mirror, thus simply replacing the image of the lunar surface by one of the adjacent sky. This was 

 readily efl'ected by means of a pulley on the azimuth screw of the siderostat from which a cord led 

 into the building. 



This method, we repeat, leaves the radiation from the apparatus itself unchanged by the 

 introduction of the lunar heat, and avoids the disturbing inliuences which come from the interpo- 

 sition and withdrawal of a screen. Other methods, however, were tried in these earlier experi- 

 ments, such as that of displacing the image by turning one of the screws of the concave mirror 

 mount. This method, though sometimes yielding identical results, was found to be liable to errors, 

 as V as also that of the use of a screen, and in a still greater degree. It is not always easy to point 

 out the exact nature of the error introduced in these delicate determinations, but we will give some 

 of these preliminary observations to show the nature of the discrepancies presumably due to such 

 methods of exposure, t 



The character of the lunar energy, as compared to the solar, was first investigated as in Lord 

 Rosse's experiments, by detej mining the relative transmissibility of the lunar and solar rays, as a 

 whole, by certain pieces of glass, which were interposed in the path of the rays immediately in 

 front of the bolometer case. Four pieces of glass were used for this ])nrpose. The first was a 

 disk 4.2 millimeters thick, of the same glass as the i)rism made by Adam Ililger, of London, used 

 in a previous determination of wave-lengths in the infra red of the solar spectrum ; the second was 



* See Plate 1, wliere M is the concave mirror; B, the bolometer; C, the cable connectiag it with the galvauom- 

 eter; G, the place at which llic glass is iuterposed. 



tTlie appaiMtus aliovc ilcscribed, as ciiiployctl in 1883 and tlie snninicr of 1884, is snbstanlially the same as tliat 

 used in the later Inuar heat iiicasiircs, the chict iinprovcnieiits, since it was conijih ted, lia\ ing been made in the gal- 

 vanometer and other electrical ad jnncts of the bolcnu-ter. For a description of this instrument the reader is referred 

 to earlier papers by the writer in the Trocecdiugs of the American Academy of Arts and Sciences, xvi, 1881, and the 

 American Journal of Science for March, 1881, and to details given later ou in the present memoir. 



