422 



SCIENCE 



[N. S. Vol. XL. No. 1029 



mometer bulb J cm. in snow) ^^ Y.°0 F. 

 External air several degrees warmer. In tliia 

 case the snow temjjerature is adopted. Rela- 

 tive tumidity, 86 per cent, at 8^ 58™, 74 per 

 cent, at lO'' 48". Hoar frost early a.m. Calm. 

 Good blue sky. Cirro stratus bands S.W. 

 Thin mist in valley below. 



(3) Feb. 3, 1909. Evening, 8»» 25™ to 8^ 35™ 

 P.M. External temperature = ■+ ll-°9 ^• 

 Dew-point, + 1.°7 F. Relative humidity, 65 

 per cent. Little wind. Sky, quite clear. 



Small corrections to the galvanometer read- 

 ings are needed in order to reduce them to a 

 standard time of vibration. 



In summer, the successive readings are apt 

 to agree better, not so much on account of any 

 improvement in the sky, but because the 

 thermal and instrumental conditions are more 

 conducive to accuracy. The following are 

 examples for summer: 



July 5, 1909 (4) Early a.m. 

 r^ 30" to 7" 40° 

 External temperature -{-66°.7F. 



Dew-poinit -|- 57. 



Eelative humidity 72;^ 



Little wind — Deep blue sky. 



Zenithal sky 



Galvanometer deflections . . 



Mean 



The loss of heat by radiation from oceanic 

 areas is much smaller than from land, because 

 of the great absorption of this radiation by 

 the very moist air which constantly hangs over 

 the water. The value of 40 per cent, trans- 

 mission which I have adopted is for land con- 

 ditions. In ascribing this value to me without 

 any restriction in his " Note on the Trans- 

 mission of the Atmosphere for Earth Radia- 

 tion," Mr. Angstrom has overlooked a pas- 

 sage in my paper which I will quote: 



In the very moist tropics, nocturnal cooling is 

 only about half as great [as in temperate regions], 

 while over the ocean the total diurnal change of 

 temperature of the water is less than ^° C. 



A computation by Lowell's method gave me 

 for the transmission of terrestrial radiation 

 in the tropics this result; that 



whereas about 60 per cent, of surface heat may be 

 emitted as radiation from temperate regions, only 

 one third as much heat escapes in this way in the 

 tropics [over land surfaces]. It is possible that 

 the low value of 10 per cent. . . . may apply to 

 saturated air over the tropical oceans, where the 

 moisture is in an especially absorbent f orm.4 



These illustrations will be sufBcient to show 

 the very great variability in the coefficient of 

 transmission of terrestrial radiation. 



A further statement in Angstrom's note, to 

 the effect that 



only a very weak part of this radiation [namely, 

 from the air at 3,000 m. altitude] reaches the 

 earth's surface, 



seems to imply that there is supposed to be 

 some interchange of radiation between bodies 

 of air thus widely separated, although actually 

 an air layer only radiates efficiently from a 

 depth of a few meters. The method employed 

 by Mr. Angstrom consists in equating one 

 half of the difference of radiation for black 

 bodies at the temperatures found at top and 

 bottom of an air layer 3,000 m. thick, to the 

 absorption in this layer. The result obtained 

 in this way in the lower air depends entirely 

 upon the thickness assumed for the " effective 

 radiating layer." But this layer, as I have 

 shown elsewhere,'' can not possibly be 3,000 m. 

 deep, nor is it even 1/100 of that depth, as the 

 investigations of Hutchins and Pearson abun- 

 dantly prove," and the depth of 3,000 meters 



i Astrophysical Journal, Vol. XXXIV., p. 376, 

 December, 1911. 



5 See ' ' Atmospheric Radiation, ' ' Bulletin G, U. 

 S. Weather Bureau, where the efficient radiating 

 layer for carbon dioxide is given as 90 em., and 

 that for illuminating gas hardly exceeds 20 cm. 

 (Op. cit:, p. 62.) 



s American Journal of Science (4), Vol. IS, 

 pp. 277-2S6, October, 1904. For air "some 

 60 per cent, of its own radiation is absorbed 

 by a column as thin as 245 cm." (op. cit., 



