80 REPORT OF THE SECRETARY. 



and thence to the wall of the chamber by convection. Accordingly, if the measured 

 heating from the coil agrees closely with the known heating api-iied to the coil, 

 much more closely will the measured heating from the solar rays represent the actual 

 rate of solar radiation. 



Repeated trials have showed that the new form of pyrheliometer is capable of 

 accurate measurement of the heating of the coil, so that confidence is felt in its meas- 

 ures of the more advantageously applied solar radiation. Preliminary comparisons 

 with the new instrument seem to indicate that the mercury pyrheliometer hereto- 

 fore used as a standard at this Observatory reads somewhat too high. 



Horizontal telescope.— Tests have been made to determine how completely the defects 

 of bad seeing are removed by churning the air column in the great horizontal tele- 

 scope, according to the i)lan initiated by you and mentioned in my last year's report. 

 It has been found that though the cliurning is of great advantage, and generally indis- 

 pensable to any work whatever on the solar image, yet the definition of an artificial 

 star, whose beam travels twice through the whole tube, is far from perfect even with 

 churning. Much of the disturbed seeing is found to be caused by the heating of the 

 poorly protected tube in the sun, and a ventilated canvas tent has been ordered to 

 screen the tube thoroughly. Plans are also formed for making additional tests on 

 other methods of churning the air in the tube. Recalling the more perfect results 

 secured in experiments on a smaller scale in 1902, and the evident improvement of 

 definition obtained in the present large tube with the churning device now installed, 

 no doubt is felt that bad seeing within the tube itself may at length be wholly 

 removed. 



A second serious defect in definition was found to be caused by warping of the 

 large plane mirrors of the coelostat, one of which is inclined forward and was at first 

 supported by a ring in front, while the other mirror, made originally for much less 

 severe work, was too thin to keep flat with an ordinary system of support. Both 

 mirrors have been almost entirely cured of these defects by the introduction of the 

 Ritchey supporting system composed of numerous balanced plates.« For the mirror 

 which is inclined forward, Mr. Ritchey's original design had to be somewhat modi- 

 fied because the mirror must be stuck to the plates instead of resting on them by its 

 weight as in mirrors supported face up. We have heretofore employed here ground 

 brass plates to which the mirror is stuck with rubber cement, but it would probably 

 be better to make the plates slightly concave and connect them by flexible tubes to 

 a large reservoir from which the air is partially exhausted, so that the mirror would 

 be held to the plates by suction. But even as we have used it, the Ritchey system 

 has wonderfully improved the definition secured on the solar image. All the work 

 on these support systems was done in the Observatory shop. 



A third serious defect in the definition of the horizontal telescope is due to the 

 tremor of the mirrors continually kept up by the city traflic, notwithstanding the 

 costly and massive piers on which the apparatus rests. Very great imi)rovement in 

 steadiness has come from placing li-inch rubber blocks under the coelostat and 

 under the concave mirrors. 



Before the improvements noted, the solar focal image, 40 cm. in diameter, was an 

 ill-defined circle at a focal distance varying often 10 feet during a single day, and with 

 the sides at different focal distances from the top and bottom. Now the image is 

 pretty sharply defined, comes to focus on all sides in the same plane, and stays within 

 less than a foot of the same focus all day; while its wanderings rarely reach much 

 over a millimeter in amplitude. It is now possible to observe the absorption in the 

 solar envelope with accuracy at within 1 or 2 per cent of the radius from the sun's 

 limb. But further improvements of the horizontal telescope are in progress, notably 

 the provision of well-protected shelters over the coelostat, the concave mirror, and 



«Astrophysical Journal, v, 143, 1897. 



