78 EEPORt OF THE SECRETARY. 



the air duct branches into two 14-inch tubes, which proceed north and south respec- 

 tively and communicate by 5-inch pipes to the interior of telescope tube. At points 

 intermediate with these other 5-inch pipes lead out of the telescope tuVje and thence 

 by return mains to the suction end of the blower, and thus the same air is continually 

 being churned aljout through the entire system. 



To prevent the blowing engine and the city traffic from communicating prejudicial 

 tremors to the apparatus, three deep and massive piers have been constructed, sup- 

 orting respectively the (;oelostat, the long-focus concave mirr or, and the spectro- 

 bolometrii- aj)paratuH used to investigate the solar image. Each pier is contained in 

 a pit originally 12 feet s(iuare and 10 feet deep, but supported by retaining walls of 

 grouting 1 foot thick, so as finally to leave a cubical-shaped pit 10 feet on a side. At 

 the bottom is a layer of sand 2 feet deep, and on this a base of grouting 2 feet thick 

 and 9 feet square, supporting the brick pier, which is built to the surface of the 

 ground 7 feet scjuare, with 18-inch walls on the four sides and a 13-inch wall north 

 and south through the center. Over all is the capstone, 8 feet north and south, 7 

 feet east and west, and 7 inches thick. In the case of the coelostat pier a two-walled 

 brick superstructure is carried up to the top of the horizontal tube to support the 

 base plate of the coelostat. In spite of all these precautions I regret to report that 

 the disturl)ance from passing traffic and even distant railroad trains has not been suf- 

 ficiently eliuiinated, and recjuires further measures to be taken to overcome the 

 almost unconquerable difficulties of the site. 



Owing to prolonged cloudy weather, the apparatus had not been fully tried between 

 the time of its installation, about June 1, and the close of the period covered by this 

 repoi-t, l)ut, so far as preliminary experiments have shown, the whole promises to be 

 a valuable equipment if the tremors due to the site can be corrected. 



The sensitive galranonieier.--'Reierrm<i to portions of my reports of preceding years 

 describing the construction and installation of a highly sensitive galvanometer, I 

 regret that attention has been diverted this year to other matters so completely that 

 comparatively little work has been done with it. In the fall of last year an appa- 

 ratus was arranged to measure the heating effect of the Ijrighter stars by its aid, but 

 unfortunately communication with the mercurial air pump had so quickly blackened 

 the silvered galvanometer mirrors that this, together with their almost microscopic 

 size, made it impossible to read the galvanometer by artificial light. After several 

 trials numerous mirrors were platinized by electrical discharge in vacuo and the gal- 

 vanometer was provided with nurrors of this kind, but immediately after the experi- 

 ments were discontinued to take up work on the provision of the great horizontal 

 telescope. It is hoped to provide for the use of this special galvanometer in spectrum 

 work on the solar image, especially in connection with sun spots, and perhaps upon 

 the heat of the stars. 



3. Invesligations relating to the solar constant of radiMion. 



Referring to my last year's report, the holographic measurements of atmospheric 

 absorption then described have been continued chiefly in the hands of the junior 

 assistant, Mr. Fowle, with improved arrangements and with more complete and 

 exact results. In connection with them the absorption of the radiation in all parts 

 of the apparatus has been determined frequently, and measures of the total solar 

 radiation by the actinometer or i)yrhe]iometer have been made also. From these 

 several kinds of data the solar constaiit of radiation, or rate of receipt of solar energy 

 at the outer limit of our atmosphere, has been computeil for a numl)er of the best 

 days. 



Imjyrovement of bolograpJts. — Before giving these values, I invite attention to Plate 

 V, which shows three superposed energy curves of the jOTsmatic solar spectrum. 

 Such holographic curves are now obtained covering the region of spectrum from 



