564 



SCIENCE 



[N. 8. Vol. XXV. No. 641 



When the separation was so slight that the 

 star disks would have appeared in contact, 

 or nearly so, with the smaller telescopes, it 

 was found that the measures made with 

 them were uniformly about ten per cent, 

 too small. When the images were well 

 separated the large and small telescopes 

 gave identical results. 



The early measures made on /* Draeonis 

 with the very small apertures indicate that 

 the deviation is a real rather than a sub- 

 jective one, and that the star images actu- 

 ally approach one another owing to some 

 diffraction phenomenon. It would appear, 

 therefore, that future catalogues of close 

 double stars should always give the reader 

 an opportunity of knowing the aperture 

 employed with every observation. A more 

 detailed account of this investigation will 

 shortly appear in the Harvard Annals, 

 Vol. 61. 



The Tenth Satellite of Saturn: W. H. 



PiCKEEING. 



The elements of the orbit of the tenth 

 satellite of Saturn, recently published 

 {Harvard Annals, Vol. 53), were based on 

 negatives taken at Arequipa in 1904. An 

 examination of another set of negatives 

 taken in 1900 has just been completed. 

 The satellite was found upon them also, 

 but some of the elements of its orbit differ 

 materially from those found in 1904. This 

 is particularly true of the eccentricity and 

 the inclination. The orbit of 1904 was ex- 

 tremely eccentric, and it was noted in the 

 publication above mentioned that in one 

 part the satellite would pass extremely 

 near to Titan. It was further suggested 

 that the peculiarities of the orbit might be 

 due to such a recent encounter. 



The orbit of 1900 was found to be nearly 

 circular, at times passing near Hyperion, 

 however. It Avas not possible to tell from 

 the negatives just when or how the ele- 

 ments had been transformed from those of 



one orbit to those of the other, if such a 

 transformation had indeed taken place, but 

 it seemed at least as probable as that two 

 satellites should exist with practically the 

 same period. The orbit is still under in- 

 vestigation, and later results will be pub- 

 lished in the Harvard Annals, Vol. 61. 



Photographic Color Photometry of Short 



Period Variable Stars: J. A. Paekhurst 



and F. C. Jordan. 



The broad problem of astronomical color 

 photometry finds an interesting application 

 in the determination of color changes in 

 short period variables of the 8 Cephei type. 

 Various observers, by indirect methods, 

 have found that these variables show 

 stronger color at minimum than at maxi- 

 mum. As the spectroscope shows these 

 stars to be binaries, a proof of color change 

 coincident with light-variation would be of 

 the highest importance. A direct determi- 

 nation of such changes is made possible by 

 a suitable combination of color-filter, ortho- 

 chromatic plate and reflecting telescope. 

 An exposure on an ordinary plate (Seed 

 27) was followed immediately by one with 

 a filter and Cramer ' Trichi-omatic ' plate, 

 which give visual magnitudes (as described 

 by R. J. Wallace in the Astrophysical 

 Journal for November, 1906). The two 

 principal advantages of the method are: 

 (1) the simultaneous determination of 

 visual and photographic magnitudes; (2) 

 elimination of personal equation with its 

 physiological complications, from the 

 visual magnitudes of the colored stars, and 

 the substitution of the peculiarities of the 

 instrumental outfit, which can be definitely 

 stated and exactly reproduced. 



Briefly stated, the results confirm the 

 color-changes in the short period variables 

 under observation, the photographic range 

 being greater than the visual ; in other 

 words, the color- factor (photographic mimts 

 visual magnitude) is greater at minimum 



