456 MR. A. FOWLER ON THE SPECTRUM OF MAGNESIUM HYDRIDE. 



The magnitudes of these displacements we're determined by reference to the ordinary 

 lines of magnesium, and partly by comparison with wave-lengths derived from 

 photographs of magnesium hydride taken with one pole of iron, and presumably 

 showing no relative displacement. 



The stronger lines of the green fluting were measured on second-order photographs, 

 and where the results from different plates were in sufficient accordance, the wave- 

 lengths in the catalogue have been stated to seven figures. It is not claimed, however, 

 that the third decimal is entitled to any great weight, but it has in some cases been 

 found desirable to retain it in the discussion of series lines. 



The lines composing the blue group are, on the whole, much fainter than those of 

 the one in the green, so that, although the stronger lines were measured on second- 

 order plates, the accuracy is probably less than that for the green fluting, and the 

 resulting wave-lengths are stated to six figures only. The fainter lines of the blue 

 and green flutings, and the whole of the yellow-green, were determined from 

 photographs of the first-order spectrum, but it is hoped that the errors will rarely 

 exceed Q'02 or 0'03 tenth-metre, except possibly in the case of doubles which are 

 barely resolved. 



Retention of ROWLAND'S ticale. 



As the immediate use of the magnesium hydride wave-lengths is for comparison 

 with the sun-spot spectrum, it has been thought desirable to convert them from the 

 scale of FABRY and BUISSON, on which they were computed, to that of ROWLAND. 

 A very complete investigation of the relation between the two scales has been made 

 by HARTMANN,* who has advocated the general adoption of a slightly modified 

 Rowland scale, whicli would have the advantage of demanding only small changes in 

 the existing tables of wave-lengths. After due consideration, however, it has been 

 thought that immediate needs would be best met by expressing the magnesium 

 hydride wave-lengths as nearly as possible on the Rowland scale now in use ; that is, 

 the scale of the " Preliminary Table of Solar Spectrum Wave-lengths." Comparison 

 of ROWLAND'S solar wave-lengths with the new standards shows that this may be 

 accomplished by adding 0'170 tenth-metre to the values on the Fabry and Buisson 

 scale from 4370 to 5365, and - 218 tenth-metre from 5365 to the head of the yellow- 

 green band at 5621. The tabulated wave-lengths (in the catalogue at the end of 

 this paper) have accordingly been expressed on this system, and may, therefore, be 

 directly compared with ROWLAND'S solar wave-lengths without introducing errors of 

 inconvenient magnitude. At the same time the wave-lengths stated in this way 

 retain some of the advantages resulting from the use of the new standards, and are 

 directly available for investigations of the series relationships of the different lines, 

 provided that 0'05 be subtracted from the tabulated wave-lengths on the less 

 refrangible side of 5365 when connecting lines on the two sides of this position. 



* ' Astrophys. Jour.,' vol. 18, p. 167 (1903); ' Physikalische Zeitschrift,' 10. Jahrg., No. 4, p. 121 (1909). 



