REPORT OF THE SECRETARY 



85 



lie measured at any sinjjle hitatioii high or low. Acrordingly we must suppose tliat 

 our solar constant values are (liminished from one or both the above causes, and 

 hence that the atmospheric absorption here estimated is too little rather than too 

 nuu'h. f ' ^^ 



TR.ANSMISSION OF SOLAR ENVELOPE. 



Table 3. — Prelimmarij ralueti of coeffirienU of transiiiii^Sivn in tlw aolar envelope. 



Wave k'ligth. 



Coefficients for vertical transmission. 



Turning now to Table 3, which purports to give the vertical transmission for dif- 

 ferent wave lengths in the solar enveloi:)e, this depends but little on our own atmos- 

 phere, and much is hoi)ed from a continuation of the work with tlie great horizontal 

 telescojte on which tlu^se results rest. .\s this work is but newly developed Here, a 

 short description of the procedure and assumptions involved in estimating the trans- 

 mission of the solar envelope will not l)e out of place. A spectro-bolonieter, provided 

 with a slit only about 5 millimeters high, is set so that rays of a certain known wave 

 length reach tiie Ijolometer from the slit. Then with the i)rism stationary, but with 

 the automatic recording plate moving before the galvanometer as usual, the solar 

 image is allowed to (h'ift by the earth's diurnal motion across the slit of the spectro- 

 bolometer. Thus is jiroduced a curve like those shown in Plate VIII, in which hori- 

 zontal distances from the center of the figure are proportional to distances along the 

 radius of the solar disc, and the height of the curve is i)roportional to the intensity of 

 the radiation of the given wave length at the corresponding point on the solar disc. 

 Ai^parently the figure would be "fiat-topped," if there was no absorption of the rays 

 by the solar envelope, and the curvature of the figure gives a rough indication of the 

 amount of the absorption." If the curve for a given day's observation has steeper 

 sides than that of another, greater absorption would be indicated for the former day, 

 and such was the case for the two days' work represented in Table 3. 



But in order to estimate quantitatively the change in absorption in the solar 

 envelojie, it is necessary to make two assumptions whose truth can as yet only be 

 verified by the accuracy with which they represent the experimental results. First, 

 it may be supposed that the radiation of the absorbing envelope is negligible com- 

 pared with that of the photosphere, and that the absorption is, like that of the 

 earth's atmosphere, represented by such an exponential formula as would apply to a 

 homogeneous atmosphere. Second, the thickness of the absorbing layer l)eing 

 unknown, its relative thickness compared with the solar radius must be assumed in 

 order to compute the exponent of the formula. 



In the first computations made here it was assumed that the thickness of tlie 

 absorbing envelope was very small as comjjared with the sun's radius. Upon this 



« Yon have suggested the possibility that owing to a columnar structure of the solar 

 surface we might find the limb of the sun brighter than the center of the disc in the 

 absence of an absorbing envelope, by reason of our seeing near the linil) only the 

 supposedly In'ighter tops of the columns. No allowance for this is made in these 

 preliminary computations. ' 



