EEPOET OF THE SECKETAEY. 81 



The bolometric examination of sun spots shows that they, too, exhibit much 

 greater absorption at the violent end of the spectrum than in the infra-red. Thus, 

 if our eye were like the bolometer, and could view sun spots by homogeneous 

 rays of different wave lengths of the infra-red as well as violet, we should see 

 the same spot four times as dark in violet light as when viewed by extreme 

 infra-red rays. 



In all the observations and reductions involved in the work described above 

 Mr. Fowle has taken by far the greatest share. 



The temperature data plotted in Plate V are reduced from the Internationaler 

 Dekadenberichte, published by the Kaiserliche Marine Deutsche Seewarte. The 

 reductions were made partly by R. Norris and partly by J. Dwyer. 



(2) MISCELLANEOUS WOEK. 



Radiation of the stars. 



Preliminary preparations were made for the detection of the radiation of 

 the brighter stars. It was at first thought practicable to mount the bolometer 

 in the center o^ the tube of the 50-centimeter diameter mirror of 1-meter focus 

 and to point the mirror directly upon the star to be exammed, but it was 

 quickly found that the disturbances due to exposure to outside air were too 

 great to be permissible with the refined sensitiveness of the bolometric appa- 

 ratus. Afterwards the mirror and bolometer were placed within the inner 

 chamber of the observatory, and the starlight was reflected in from a 30-incb 

 plane mirror on the coelostat. The galvanometer employed was the' one de- 

 scribed at pages 91-92 of your report for the year ending June 30, 1902. 



The sensitiveness available depends largely on reducing the damping of the 

 needle, and a long time was spent in making the galvanometer case air-tight, 

 so that a pressure of 1/1000 atmosphere or less could be maintained without 

 rapid change. In this we were at length so successful that the change of 

 pressure was hardly appreciable in several months. Great difficulty was en- 

 countered in balancing the bolometric apparatus at the high sensitiveness em- 

 ployed on account of small electromotive forces in the galvanometer circuit 

 and its shunts. Thus a balance would be obtained with a certain shunt across 

 the galvanometer, and on passing to the 'next shunt a very great deflection would 

 be found, due to a new electromotive force in the new shunt circuit. At lengtii 

 it proved necessary to discard shunts wholly, and to employ instead a variable 

 resistance in series with the galvanometer. Great difficulty was still ex- 

 ]:)erienced in balancing, but not so great as to render it impossible, as before. 

 When once balanced the apparatus was well behaved. The sensitiveness ap- 

 peared from tests on candle flames to be A^ery much greater than that obtained 

 by previous experimenters on stellar radiation. Unfortunately the difficulties 

 encountered were not surmounted until early in January, and bad weather pre- 

 vented a trial on the stars until more important work displaced the investi- 

 gation for the present. 



ISlew apparatus. 



Bolometer. — In our previous construction of bolometers we have been guided 

 more by experience than by any theorj"- in their design. While preparing for 

 the research on stellar variation it seemed A^ery desirable to determine the 

 conditions which would insure the highest sensitiveness. Accordingly the 

 subject was studied from the standpoint of Fourier's analytical theory of heat, 

 and numerous experiments AA^ere made to further enlighten it. As a result a 

 complete theory of bolometer construction was reached, and it is now possible 



