August 4, 1916] 



SCIENCE 



169 



pressures of water vapor varying between 

 4.62 and 9.99 mm. on September 20, and be- 

 tween 2.21 and 7.49 mm. on September 21. 

 The total quantity of precipitable water in the 

 atmosphere on September 20, as determined 

 by Fowle's spectroscopic method, varied be- 

 tween 3.32 at low-sun observations to 8.6 mm. 

 at high-sun observations, and on September 21 

 between 3.8 and 8.3 mm. Thus there was be- 

 tween two and three times as much water vapor 

 present in the midday air as there was at low- 

 sun observations. Since the transmissivity of 

 the atmosphere is known to diminish with the 

 increase of aqueous vapor, other things re- 

 maining equal, would it be at all likely that 

 Mr. Abbot's assertion that the transmissive 

 quality of the atmosphere above Mount Wilson 

 remained unchanged throughout these days, 

 should turn out to be true? And do not the 

 partial transmissions which I have derived 

 from his own figures point to a contrary con- 

 clusion ? 



In his second paragraph, Mr. Abbot tries to 

 discredit my measurements of the distribution 

 of intensity in the spectrum of the earth- 

 shine, because my statement that the night 

 sky at Flagstaff in the early morning of 

 August 9 and 10, 1912 (civil reckoning), was 

 exceptionally clear, appears to him incompati- 

 ble with the experience of himself and others 

 that the " skylight near the sun in daytime 

 notably increased " during that month. My 

 statement rests upon the following evidence: 



The spectrograms of the earth-shine were 

 made for me at Dr. Lowell's observatory by Dr. 

 V. M. Slipher. I had asked Dr. Slipher to 

 place the slit of his spectroscope half on and 

 half off the dark limb of the moon. In this 

 way there were obtained juxtaposed spectro- 

 grams of precisely the same duration of expo- 

 sure and photographic development, one of the 

 earth-shine plus diffuse skylight from inter- 

 vening air, illuminated by the light passing 

 through it from the bright crescent of the 

 moon, and the other of the skylight alone, from 

 which the true earth-shine was obtained by 

 difference. Dr. Slipher had given me his im- 

 pression from eye estimate that the sky on 

 August 8 (astronomical date) was " good," and 



on August 9 " excellent " ; but my quantitative 

 measurements are far superior to any eye esti- 

 mates, and these tell the following story : 



Without going into the minutise of the 

 photographic corrections, I will merely record 

 that all necessary corrections of this sort have 

 been applied. Those interested will find the 

 details given in my paper on " The Photo- 

 graphic Spectrography of the Earth-shine and 

 a Spectrophotometric Comparison of the Earth- 

 shine with Moonlight, Skylight and Sunlight, 

 together with a Study of the Difficulties of 

 Photographic Comparisons." 1 



The ratios of exposure durations for earth- 

 shine (tE) and for moonlight (iji/) were 



August 8, 1912, t E : t u = 4800 : 1, 

 August 9, 1912, t E : t M = 2840 : 1. 



The average of the ratios of photographic 

 opacities on the spectrograms for earth-shine 

 and moon (Jjl/Jjsi) and for earth-shine and 

 sky (Je/J s) were 



August 8, 1912, JeIJm = 1.360 : 1; Je/Js = 3.62 : 1, 

 August 9, 1912, Je/Jm = 1.062 : 1; Je/Js = 8.49 : 1. 



The ratios of moonlight to the skylight just 

 outside of the extreme border of the moon's 

 dark limb were therefore 



„ ,„,„ tE Jm Je 4800X3.62 , o __ o 

 Augusts, 1912, - X-X^ = — 3 6Q - - = 12,776: 1, 



August9,1912,^X^X^= 2 ? 4 ^^ = 22,704:l. 

 til Jjs Js 1.062 



For comparison I give corresponding values 

 of the ratio of moonlight to skylight, obtained 

 at Westwood, Massachusetts, during my visual 

 measures of the earth-shine, which give an 

 idea of the variation which is to be anticipated 

 in skies ordinarily reputed " clear " : 1911. 

 Sept. 28, 52 : 1 (sky hazy) ; Sept. 30, 3095 : 1 

 (clear) ; Oct. 2, 1149 : 1 (clear, followed by 

 cirro-stratus) ; Oct. 26, 3033 : 1 (clear) ; Oct. 

 29, 3626:1 (clear); Nov. 16 (a.m.), 1871:1 

 (clear); Nov. 17 (a.m.), 8579:1 (exceptionally 

 clear) ; Nov. 27, 1358 : 1 (clear to hazy) ; Dec. 

 14 (a.m.), 9380:1 (exceptionally clear). 1912. 

 Feb. 20, 2476: 1 (faint cirrus bars). 



Here the greatest degree of clearness at this 

 station about 200 feet above sea level, gave a 



i Astronomische Nachrichten, Nr. 4819-20, No- 

 vember, 1915. 



