646 



SCIENCE 



[N. S. Vol. XXXVII. No. 956 



there is marked shortening of the focus. Compu- 

 tation of the "criterion" constant gives 3*^0.33, 

 placing the quality well within the highest class. 



The mean diameter of the confusion disk, when 

 expressed in seconds of arc, becomes 0".69 as 

 against 0".45 for the 40-inch Yerkes. The the- 

 oretical resolving power of a 9.4-inch is 0".52, 

 whereas for the 40-inch it is 0".12. It is sug- 

 gested that a criterion to best represent the quality 

 of the optical work should involve this constant 

 for any given aperture. This would appear to 

 favor the quality of the smaller lens. This ratio 

 might well be called a "coefficient of resolution." 

 On th^ Luminous Efficiency and Color-index of a 



Black Body at Different Temperatures: Heney 



Norms Russell. 



The curves of spectral sensitiveness given by 

 Parkhurst in his "Yerkes Actinometry" (with 

 the addition of certain data very kindly conamuni- 

 cated by Professor Parkhurst) make it possible to 

 compute the luminous efficiency of a body radi- 

 ating according to Planck 's law at any tempera- 

 ture, that is, the ratio of its actual visual or pho- 

 tographic brightness to that of a body radiating 

 the same amount of energy, but all of the wave- 

 length of greatest visual or photographic effi- 

 ciency. The results here given are provisional, 

 and may be somewhat altered when fuller data 

 regarding the spectral sensitiveness become avail- 

 able. 



The visual luminous efficiency is a maximum for 

 a temperature of about 7,500°, its value being 

 0.11. The visual surface brightness, on the 

 Yerkes scale, varies with the temperature very 

 nearly as the intensity of monochromatic radia- 

 tion of wave-length 0.541 /i, and the photographic 

 surface brightness like that of wave-length 0.428 /t, 

 the deviations averaging less than O^.O? for tem- 

 peratures between 2,000° and 25,000°. The color- 

 index can be still more closely represented by the 

 formula 



7500° 

 " T " 

 the residuals averaging only 0"'.02. 



For the Harvard visual and photographic ob- 

 servations the mean effective wave-lengths appear 

 to be 0.516' /i and 0.419^, and the color-index is 

 given by the equation 



Phot. — Vis. : 



■ 0"'.70, 



Phot. — Vis. 



6900 



- — o^.eo. 



The "black-body" temperatures corresponding 

 to the color indices of stars of the various spec- 

 tral types may now be determined. The Harvard 



and Yerkes data give effective temperatures 

 (ranging from 23,000° for Class B„ to 3,100° for 

 Class M and 2,300° for Class N), which are in 

 excellent agreement with one another and with 

 the previous determinations of Wilsing and 

 Scheiner by visual spectro-photometric methods. 



It appears also that the relative visual surface 

 brightness of any two "black bodies" at different 

 temperatures, if expressed in stellar magnitudes, 

 should be 3.8 times the relative color-index on the 

 Yerkes scale, or 4.3 times the color-index on the 

 Harvard scale. The luminous efficiency is almost 

 constant for color-indices between 0.0 (Sp. A) 

 and 0.7 (Sp. G), but falls off rapidly for bodies 

 at lower temperatures. 



If these results, which are strictly true only for 

 theoretically perfect radiators, apply approxi- 

 mately to the actual stars, and we could measure 

 the brightness of the latter by means of their 

 whole energy radiation, instead of a narrow spec- 

 tral region, stars of Class K would seem about 

 twice as bright, those of Class M four times as 

 bright and of Class N more than twenty times as 

 bright as they do now, in comparison with stars 

 of Classes A to G. This would profoundly modify 

 our estimates of the relative abundance of stars of 

 different spectral types. 



The Eclipsing Binary t Aurigw: Harlow Shap- 



LEY. 



More than five thousand observations of the 

 variable star e Aurigw, made by Schmidt through- 

 out the interval from 1843 to 1884, have been 

 studied by Ludendorff, who deduces a period of 

 light variation of 27.1 years and a light curve 

 similar to those of certain eclipsing variables. 

 Only three minima have occurred since the dis- 

 covery of the light fluctuation by Pritsch in 1821. 

 A study of the light curve by the writer shows 

 that the observations can be satisfactorily repre- 

 sented by the eclipse theory. An accurate orbit is 

 not possible, notwithstanding the large number of 

 observations, but limiting sets of elements have 

 been derived. The eclipse is total — the small 

 bright star being completely hidden for more than 

 a year behind the faint companion, whose volume 

 is one thousand times the greater. The com- 

 ponent stars are distantly separated. The mean 

 density of the smaller component, if the masses 

 are approximately equal, is not unusual in com- 

 parison to the densities already found for 

 W Crucis and EZ, Ophiuchi, whose periods of 

 198 and 262 days, respectively, are the longest 

 heretofore known. The density of the big com- 



