256 CARNEGIE INSTITUTION OF WASHINGTON. 



case. The indicated velocity is greatest at the highest level and falls 

 to a minimum and rises again, as does the pressure on the previous 

 assumption. 



The graph of intensities and velocities shows a point of inflexion 

 between intensities 4 and 5. This indicates that pressure begins to be 

 effective at this level. By extending the uninfiected course of the graph, 

 it appears that the downward velocity dies away at about the level of 

 intensity 3. Such an extrapolation is hazardous, but offers a means 

 of separating the two effects. The conclusion is that the solar vapors 

 are descending with a decreasing velocity which approaches zero at the 

 level of lines of solar intensity 3, that the pressure does not differ greatly 

 from zero above the level of lines of intensity 5, and that at the level of 

 lines of intensity 4 it is approximately one terrestrial atmosphere and 

 increases rapidly with depth. 



Lines of groups c and d with positive pressure shifts are displaced to 

 the violet under the decreased pressure in the solar atmosphere, as the 

 observations show. Lines of group e, with a negative pressure shift, are 

 displaced to the red by the decrease of pressure. 



Added evidence that the pressure in the solar atmosphere is approxi- 

 mately zero at the level of lines of intensity 5 is given by obtaining the 

 velocities required to produce the observed effects in the several groups 

 whose average intensities are all nearly 5. These are in very close 

 agreement for all groups. No estimation of the pressure at lower levels 

 can yet be made, but the sharpness in the solar spectrum of diffuse arc 

 lines shows that it is moderate. The observations will be extended to 

 weaker lines as rapidly as possible. The great mass of the vapors 

 forming the solar atmosphere is found within a very thin laj^er in which 

 the pressure gradient is high. The observations appear to support the 

 growing view that the sharp boundary of the sun may be explained 

 without the assumption of a cloudy photosphere. The known down- 

 ward drift of the rare high-level vapors, now found to be shared by 

 those at lower levels, is analogous to the slow settling down of volcanic 

 dust in the terrestrial atmosphere. The upward currents are narrowly 

 localized and find their origin probably in the prominences. 



This investigation will be continued with the reconstructed 60-foot 

 tower telescope, using arc and spark in vacuo for comparison spectra, 

 apparatus for simultaneous exposure of solar and comparison sources, 

 and a Fabry-Perot interferometer with fused quartz plates, in invar 

 mounting. 



SOLAR ROTATION. 



Observations of the solar rotation are being made bj'' Mr. Adams and 

 Mr. St. John with the 150-foot tower telescope and 75-foot spectro- 

 graph. A distinct gain in precision over previous work is hoped for, 

 because of the better quality of the grating and the enlarged solar image, 



