170 



SCIENCE 



[N. S. Vol. XLIV. No. 1127 



ratio of not over 10,000 : 1, which falls consid- 

 erably short of the Flagstaff conditions on 

 either of the given dates. 



It seems to me that I am fully justified in 

 calling the mornings of August 9 and 10, 1912 

 (civil date), exceptionally clear, even for 

 Flagstaff ; and I submit that exact quantitative 

 measurements, such as I have given, are to be 

 preferred to Mr. Abbot's vague estimate that 

 " skylight near the sun in daytime notably in- 

 creased." If the discrepancy is regarded as 

 'sufficiently noteworthy, I would suggest that 

 it indicates that the " dust cloud from 

 Katmai " was not as universal as Mr. Abbot 

 supposes. Mr. Abbot has inferred from the 

 consistent agreement of his observations with 

 those of some other observers, that the obscura- 

 tion which he attributes to the eruption of 

 Katmai was world-wide and continuous; but 

 this is a mere hypothetical conjecture, in the 

 absence of anything known to the contrary, 

 which a single good opposing observation can 

 overthrow. 



"While the presence of a clear and uniform 

 sky is an advantage in such delicate measures 

 as those of the spectrum of the earth-shine, it 

 is not an indispensable one, because my method 

 of observation permits accurate measurement 

 of and correction for the interfering skylight; 

 and it is not quite exact to say that " Mr. Very 

 hangs the merit of his work on the exceptional 

 clearness of August 8 and 9, 1912," because I 

 have given these observations no greater weight 

 in the final result than is assigned to other 

 dates when the skylight was considerably 

 stronger than the earth-shine. Being freed 

 from the variable effect of skylight, my meas- 

 ures are sufficiently exact to show not only the 

 variation of the earth-shine from day to day 

 with the changing phase of the illuminating 

 earth, but they also detect variations in the 

 quality of the light which are attributable to a 

 variable proportion of blue " skylight," i. e., 

 sunlight scattered upward by the clear air in 

 the same way that skylight is scattered down- 

 wards, and varying in amount according to 

 the cloudiness of the earth's hemisphere facing 

 the moon. 



Coming to Mr. Abbot's third point, in which 



he defends the conclusions of Mr. A. Angstrom, 

 who finds a mean atmospheric transmission of 

 terrestrial radiation by clear air of about 15 

 per cent., where I obtain about 40 per cent., 

 I anticipated Mr. Angstrom's curve of instru- 

 mental radiation to limited areas of sky at 

 different zenith distances, and obtained a sim- 

 ilar, but more accurate curve; 2 but I did not 

 make his mistake of confounding this purely 

 instrumental result with the radiation of the 

 earth's surface to outer space. It is true that 

 the radiation from a small surface so circum- 

 scribed that the rays can only escape through 

 a narrow aperture, pointing to the sky in a 

 direction but little elevated above the horizon, 

 so that the path through the lower moisture- 

 bearing layers of the atmosphere is equivalent 

 to a passage through a considerable depth of 

 water, is usually so impeded that scarcely any 

 gets through. But the radiation of the indefi- 

 nitely extended surface of the earth, free to 

 radiate vertically through a comparatively 

 shallow layer of moist air, escapes readily. 

 For such radiation there is an extensive region 

 of the spectrum between 8.5 and 12.8 /x, where 

 the transmission averages something like 80 

 per cent. Yet even the maxima, or spectral 

 regions of comparatively free transmission, 

 are almost obliterated in the long road through 

 the air in a pointing not much above the 

 horizon. This is an important fact, and its 

 explanation has seemed to me to lie in the 

 presence of multitudes of excessively faint ab- 

 sorption lines in the parts of the spectrum 

 where the maxima reside — lines which are too 

 fine and too faint to be individually discrimi- 

 nated by the bolometer, but which increase in 

 intensity and finally produce a somewhat gen- 

 eral obscuration of the spectrum, even in its 

 more transmissible portions, when the air path 

 becomes excessive. The recognition of the 

 existence of these faint lines by Mr. Abbot 

 would go a long way towards removing the 

 discrepancy between our points of view. 



1 will not trespass on your space to point 

 out the numerous errors in Mr. Angstrom's 



2 See my paper, ' ' Sky Eadiation and the Iso- 

 thermal Layer, ' ' Am. Jour. Sci., Vol. XXXV., Fig. 

 2, p. 383, April, 1913. 



