October 6, 1916] 



SCIENCE 



495 



tific investigation of anilines, the working out 

 of processes, and the study of the constitution 

 of color, particular regard is to be paid to coal 

 tar distillation and the industrial application 

 of cellulose. Another feature will be an experi- 

 mental dyehouse. Mr. G. H. Frank, M.Sc, 

 and Dr. Oesch, a Swiss expert, are retained on 

 the staff, and with them will be associated Mr. 

 P. E. King, Lieutenant A. E. Woodhead, 

 M.Sc, Professor E. E. Watson, D.Sc, of 

 Dacca College, and, as outside lecturers, Mr. 

 H. P. Hird and Mr. C. F. Cross, both special- 

 ists engaged in allied industries. 



DISCUSSION AND CORRESPONDENCE 



ATMOSPHERIC TRANSMISSION 



To the Editor of Science : On page 168 of 

 your issue of August 4, 1916, Mr. Very is un- 

 fair to himself, to your readers, and to me. He 

 points out that the Smithsonian Mount Wilson 

 observations of September 20 and September 

 21, 1914, indicate greater transparency of the 

 atmosphere for the complete, complex solar 

 beam made up of energy of all wave-lengths 

 the greater the air mass. From this he tries 

 to lead your readers into the conclusion that 

 the atmosphere gradually decreased in clear- 

 ness during our period of observations. No- 

 body knows better than Mr. Very of Langley's 

 mathematical proof that a complex beam 

 traversing a medium the transmissive power of 

 which varies with the wave-length must nec- 

 essarily behave in this manner even though 

 the medium is perfectly homogeneous. Pure 

 water or glass would show the same effect. 

 The transmission would continually increase 

 for each successive layer traversed. This is 

 because the less transmissible rays are con- 

 tinually becoming a smaller proportion of the 

 intensity of the whole complex beam the far- 

 ther it goes through the medium. If our 

 pyrheliometric observations had not shown the 

 phenomenon which Mr. Very mentions they 

 would have proved that the sky was growing 

 clearer. The question then only remains 

 whether the effect they do show is of the 

 right magnitude or not. This is settled affirm- 

 atively by the results obtained with the 

 spectro-bolometer. 



For monochromatic rays the atmospheric 

 transmission should be constant for all air 

 masses, if the atmosphere neither grows clearer 

 nor more opaque. Our spectro-bolometric work 

 shows that this condition was closely fulfilled 

 on the two days in question, as Mr. Very well 

 knows. Having no comfort from the spectro- 

 bolometric work, he omits mention of it, and 

 tries to carry his point with the uninformed by 

 paradoxing. 



Mr. Very, however, draws attention to the 

 increase of atmospheric humidity during the 

 observations as indicated by Fowle's measure- 

 ments. It may be remarked that between air- 

 masses 11.0 and 7.2 on September 20 no appre- 

 ciable change occurred. Yet that part of the 

 observations gives the same result as the rest, 

 showing that the effect of such small increase 

 of humidity as occurred during the rest of the 

 morning was negligible. Those who consult 

 the original derivation of Fowle's method of 

 estimating atmospheric humidity, are, how- 

 ever, aware that it rests on laboratory experi- 

 ments extending only to 5 millimeters of pre- 

 cipitable water. For the exceptionally large 

 air masses occurring on September 20 and 

 21 it was applied to the estimation of over 

 65 millimeters. It seems as likely that 

 this extreme extrapolation involved inaccu- 

 racy, increasing with increasing air-mass 

 rather than that the atmospheric humidity 

 really increased from 3.3 to 4.0 millimeters 

 during so short a time as the first 8 minutes 

 after sunrise. I therefore incline to think that 

 there was very little or no increase at all in 

 atmospheric humidity on September 20 be- 

 tween air masses 19 and 3, although a small 

 increase from 3.3 to 5.2 is indicated by Fowle's 

 results. Later on there was really a small in- 

 crease of humidity, but it appears to have 

 been insufficient to produce appreciable error 

 in the solar-constant values as calculated from 

 small air masses. 



As to the clearness of the sky at Flagstaff, 

 Arizona, in August, 1912, Mr. Very shows 

 that it was clearer there, at 7,000 feet eleva- 

 tion, than he is accustomed to find it near 

 Boston, but he does not show that it was clear 

 sky at Flagstaff. If it was really exceptionally 



