March 17, 1905.] 



SCIENCE. 



417 



The conspicuous singles are the same in 

 the drawings of both observers. 



The Anion and Kison are convergent in 

 both and in the same direction. 



Only the faint or very close doubles show 

 differences at the two presentations. 



The double canals, then, declare their 

 own ob.jectivity on three counts, each more 

 compelling than the one before: (1) The 

 fact of showing double, (2) the relative 

 width of the double, (3^) the absolute width 

 of the double ; and they do this precisely 

 as a real object would, the certainty in- 

 creasing with the ease of observation. The 

 determination of the absolute width is very 

 difficult, and here we find the probability 

 for "reality strong but not expressible; the 

 relative width is easier to determine and 

 the probability for reality is 24 to 1; lastly 

 the defcrminatidn of the fact of being 

 double, the easiest observation of all, shows 

 the probability that it is real to be 128 to 1. 



Note on Three Solar Periods: Frank H. 



BiGELOW. 



The mean period of rotation of the solar 

 photosphere at the equator is about 26.68 

 days, as determined by solar observations. 

 There is a mean period of about 25.98 days 

 indicated by terrestrial, magnetic and 

 meteorological observations, which has been 

 regarded as a period of solar rotation. The 

 relative frequency of the solar prominences 

 and the annual variations in the earth's 

 atmosphere show that there is a short 

 cycle of about 1,004 days. These are ap- 

 parently related together b.y the equation, 



2r,^68 + 1004 = 25^ (appro.ximatel v ). 



Some discussion is given of a possible 

 physical cause for this condition, as found 

 in the interior circulation of the sun's mass. 



The Coordiiiafioii of Visual and Photo- 

 graphic Star Magnitudes: John A. 

 Parkiiurst. 



The importance of stellar pliotcme'.ry 

 among the departments of modern astron- 

 omy arises from the fact that the magni- 

 tude of a star bears immediately on the 

 star's physical condition and changes. 

 That this is of growing importance is wit- 

 nessed, among other things, by the numer- 

 ous discoveries of new variable stars, over 

 three hundred in the present year; show- 

 ing that variability must be reckoned with 

 as a factor in stellar evolution to an extent 

 that would net have been imagined a de- 

 cade ago. The relation lately shown to 

 exist between stellar variability and sun- 

 sp(.t phenomena adds at once to the inter- 

 est of the problem and the possibility of its 

 solution. 



The photometric catalogues published 

 within the last few years by the Harvard 

 and Potsdam observatories furnish a se- 

 cure basis for visual photometry, their re- 

 sults agreeing reasonably well except the 

 discordances arising from differences in the 

 star colors. No such basis for photo- 

 graphic magnitudes now exists, therefore 

 to be useful and intelligible, magnitudes 

 must be reduced or reducible to the visual 

 system. But the extension of photometric 

 work demanded by the present needs of 

 astronomy is possible only by photographic 

 means ; hence the pressing need of finding 

 some method of harmonizing visual and 

 photographic results for stars differing in 

 type of spectrum and, therefore, in color. 

 The usefulness of such a method will vary 

 somewhat in proportion as it enables us to 

 utilize the photographic magnitudes al- 

 ready obtained. 



That the great accuracy of photographic 

 methods applied to the astronomy of posi- 

 tion has as yet no counterpart in the astron- 

 omy of magnitudes, is due to the one dis- 

 turbing factor of star color. It is well 

 known that a colored star will affect dif- 

 ferently the eye and the photographic plate, 

 but it is not so well appreciated that equal 



