January 3, 1913] 



SCIENCE 



33 



brighter companion. Frequency curves showing 

 the distribution of total proper motion and of 

 motion in both right ascension and declination 

 were exhibited and certain suggestions relative to 

 the structure of the stellar system were derived 

 from them; e. g., the mean motion of these stars, 

 about 3".5 per century, is larger than has usually 

 been supposed, indicating a smaller total extent 

 than is commonly attributed to the sidereal uni- 

 verse. A general southerly drift, shown by these 

 stars, suggests that in great part they lie outside 

 a group associated with the sun and comprising a 

 large part of the brighter stars. The precession 

 constant furnished by these faint stars is shown 

 to be in excellent agreement with that found by 

 Boss from the brighter stars, implying that these 

 bright stars possess, as a whole, no motion of 

 rotation about the earth's axis that is not shared 

 by the fainter stars. 



A curious empirical relation, long known but 

 little heeded, is confirmed and considerably ex- 

 tended by these results, viz., the product, stellar 

 magnitude multiplied by proper motion, is approxi- 

 mately constant over all magnitudes from the 

 second to the eleventh, inclusive, and unless there 

 are conditions of the problem not now apparent 

 this relation would imply that in the average of a 

 large number of eases a star's distance from the 

 sun is proportional to its numerical magnitude. 

 This suggestion is probably of too revolutionary 

 character to be readily accepted, although it is 

 paralleled by the conclusion reached by Campbell 

 from spectroscopic evidence that the stars of dif- 

 ferent magnitudes are more thoroughly intermixed 

 in space than has been commonly assumed. 



The Solar Eclipse of April 16, 1912, as Predicted 

 and as Observed: Arthur Newton. 

 The Nautical Almanac Office is preparing for 

 the press the final installment of Professor Simon 

 Newcomb's "Eesearches on the Motion of the 

 Moon." Corrections therein given have been ap- 

 plied in the American Ephemeris and Nautical 

 Almanac for 1912 to the moon's position in the 

 computation of the two solar eclipses of the 

 present year. 



The many observations made on April 17 are 

 entirely confirmatory of the accuracy of these final 

 values by Professor Newcomb. The American 

 Almanac gave a predicted time of occurrence that 

 was within two seconds of the mean observed time. 

 The European almanacs were in error from 15 

 to 30 seconds. The central line of eclipse was 

 placed about one mile too far to the northwest by 



the American Almanac, and much farther in the 

 same direction by the other almanacs, with the 

 exception of the French. 



These results indicate that the moon 's position 

 as employed in the American Ephemeris for this 

 eclipse required a further correction of + 1".0 

 to the longitude and of — 0".5 to the latitude. 

 The prediction was further borne out in the phase 

 which was total in Portugal and Spain with a 

 maximum duration of one second, and annular in 

 France with a duration of three or four seconds 

 at Paris. The indicated correction to the adopted 

 semi-diameter of the moon is less than 0".2. 

 A Method of Approximating Rainfall over Long 



Periods and Some Results of its Application: 



A. E. Douglass. 



It was found by a test extending over 43 years 

 that the radial thickness of the rings of the yellow 

 pine of northern Arizona gives a measure of the 

 rainfall in that vicinity with an average accuracy 

 of over 70 per cent. By applying a simple for- 

 mula, taking into account the conservation of 

 moisture, the accuracy may be increased to about 

 75 per cent. By cross identification of rings 

 between all the hundred trees examined, the accu- 

 racy of counting rings was greatly increased. 



Five trees from the vicinity of Flagstaff were 

 measured to the number of 400 rings, and two of 

 these to 500. It was found that all the trees in 

 that locality gave very similar records. A 21-year 

 variation amounting in all to 20 per cent, of the 

 mean is shown in 400 out of the 500 years re- 

 corded. A shorter variation amounting to 16 per 

 cent, of the mean was found to have a period of 

 11.4 years. Its plot derived from 492 years shows 

 2 maxima which correspond in time with 2 maxima 

 of rainfall in the 50 years of records on the 

 southern California coast. These in turn match 

 the major and minor maxima in the temperature 

 of that region for the same period. The larger 

 maximum of the latter occurs at the time of the 

 sim-spot minimum as averaged for 125 years. 



The Algol Variable BB Draconis: F. H. Seabes. 

 The writer made an attempt to determine the 

 light curve of this star from observations with 

 the 7i-inch refractor of the Laws Observatory. 

 On account of its faintness it could not be fol- 

 lowed through the minimum. On August 7 the 

 star was followed photographically through an 

 eclipse with the 60-inch reflector of the Mt. Wilson 

 Observatory. Sixty-five exposures, covering an 

 interval of 6 hours, were obtained. The variable 

 is normally of 9.70 mag. There is an interval of 



