854 



SCIENCE 



[N. S. Vol. XXVIII. No. 728 



used for improving the agreement of the cal- 

 culated and observed curves. The application of 

 the method to a considerable number of cases can 

 be facilitated by the use of tables. 

 The Temperature Gradients of the Atmosphere 



and am Attempt to Account for the Upper 



Inversion: W. J. Humphreys. 



During the past few years many sounding bal- 

 loons, equipped with registering apparatus, have 

 been sent up from different places and under 

 different conditions to altitudes ranging from 

 twelve to twenty kilometers, and observations 

 have been secured indicating a division of the 

 explored atmosphere into the following three 

 regions : 



1. A layer some 3,000 meters thick next the 

 surface, in which the change of temperature with 

 elevation is irregular and often locally reversed. 



2. A region of fairly uniform and rapid tem- 

 perature decreases with elevation extending from 

 the top of the first layer to the high cirrus clouds. 



3. The upper inversion, or the region above 

 the cirrus clouds, when the temperature slowly 

 increases with elevation. 



The upper inversion is due primarily, but not 

 wholly, to long wave radiation from the earth 

 as a planet, and to which water vapor is a black 

 body. When the temperature of the high atmos- 

 phere is 218° C. absolute, it can be shown that 

 the temperature of the effective radiating surface 

 is about 260° C. absolute. 



Effect of Increasing the Slit Width on the Accu- 

 racy of Radial Velocity Determinations : J. S. 

 Plaskett. 



Experiments at Ottawa have shown that ex- 

 posure time is almost proportionally decreased 

 VTith increase in slit width to about 0.075 mm. 

 Six spectra of /3 Orionis were made for each of 

 four slit widths — 0.025, 0.038, 0.050 and 0.075 ram. 

 — at three different dispsrsions. The accidental 

 errors of setting, as measured by the probable 

 errors of the velocity from a single line obtained 

 from the six plates in each series, are only slightly 

 increased, scarcely at all in the higher dispersion, 

 by increase in slit width to 0.075 mm. A further 

 discussion shows that the increase of systematic 

 error, so far as it may be determined from the 

 limited number of plates, is also small with high 

 dispersion, although quite marked for a slit 

 0.075 mm. wide with low dispersion. These some- 

 what unexpected results show that for stars with 

 single lines the slit may be much widened without 

 much loss of accuracy and with a considerable 

 saving in exposure time. 



The Algol System, RT Persei: R. S. Duqan. 



With a Pickering polarizing photometer with 

 sliding achromatic prisms, a series of 14,048 

 settings on RT Persei was completed in February, 

 1908. Most of the observations were made with- 

 out a recorder. 



The period is obtained from nineteen more or 

 less thoroughly observed minima. 



The mean curve shows a primary minimum of 

 1.33 magnitudes' range and a secondary of 0.16, 

 each lasting about four and one quarter hours. 



The curve between minima is not a straight 

 line. After recovering from primary minimum 

 the curve rises steadily until it reaches the point 

 midway between the minima, and from that point 

 decreases very little until secondary minimum 

 begins. 



The residuals show a constant correction for 

 the night, depending on the average hour angle 

 at which the observations were made. 



The eccentricity and inclination are small, the 

 two stars are nearly of the same size, one has 

 six times the intrinsic luminosity of the other, 

 and the radius of the orbit is three times the 

 radius of the stars. 



Observations of the Total Solar Eclipse of Jan- 

 uary 3, 19^8: W. W. Campbell. 



The paper summarized the more interesting 

 results of the work of the Crocker Expedition 

 from the Lick Observatory, with lantern-slide 

 illustrations copied from the original negatives. 



The corona was rather remarkable for the great 

 number of long, straight and slender streamers 

 extending in all directions. There was a con- 

 spicuous conical pencil of radiating streamers, 

 near position angle 75°, whose vortex, if on the 

 sun's surface, would lie within the largest sun- 

 spot group visible on January 3. 



By far the greater part of the coronal light 

 came from the areas lying within two inches of 

 the sun's edge. 



Polarigraphic negatives showed the existence of 

 strong polarization in the coronal light, even up 

 to the very edge of the sun. 



A spectrogram with continuously moving plate 

 recorded the changing spectrum of the sun's edge 

 as the edge was gradually uncovered by the moon. 



The spectrum of the corona was essentially free 

 from absorption lines. Several new coronal bright 

 lines were discovered. The maximum of contin- 

 uous spectrum of the corona was displaced toward 

 the red from that of the solar spectrum, indi- 

 cating a lower temperature in the corona than in 

 the photosphere. 



