1882.] On the Disappearance of some Spectral Lines, fye. 429 



We have observed the spectrum of a block of magnesia, rendered 

 incandescent by an oxyhydrogen jet. In the visible part of this spec- 

 trum we found no discontinuity, no lines bright or dark (except the 

 inevitable D lines), no sign of the blue channelled spectrum of 

 magnesia. Drs. Huggins and Reynolds (" Proc. Roy. Soc," voL 18, 

 pp. 547, 551) have recorded the appearance of the b group under these 

 circumstances, but we failed to get sight of it. In the ultra-violet 

 region photographs still show a continuous spectrum, extending far 

 beyond the limit of the solar spectrum, in fact] as far as we have 

 hitherto observed any lines of magnesium to occur, but on this con- 

 tinuous spectrum one line, and only one, comes out, which is the 

 strongest line of burning magnesium and of the arc spectrum, at wave- 

 length 2852 (2850 Cornu). This line shows sometimes bright, some- 

 times reversed, against the continuous background. In this case, 

 where the appearance of the lines depends on their relative brightness, 

 as compared with the continuous spectrum, there is no advantage, so 

 far as the visible rays are concerned, in the photographic method over 

 that of observation by the eye ; there may be a disadvantage, as the 

 photograph presents only the mean result of a certain time. But where 

 the faint lines of a discontinuous spectrum are in question the photo- 

 graphic method has the advantage, for when the vibrations are too 

 feeble to produce any sensible impression on the retina, they may yet, 

 hj integration of their effects during a lengthened exposure, produce a 

 definite effect on the photographic plate. In general we have only 

 exposed our plates for such times as would give us the best defined 

 images, so that very faint lines are not developed in them ; but by 

 using a prolonged exposure we find that in many cases the disappear- 

 ance of lines from the arc or spark is more apparent than real, and is 

 attributable to a variation of intensity, not to an absolute cessation of 

 the vibrations corresponding to the evanescent lines. Thus of the 

 quadruple group between wave-lengths 2789 and 2802 in the spark 

 spectrum of magnesium only the stronger two lines are usually seen in 

 photographs of the arc with short exposure, but the whole four 

 produce their impressions on the plate if sufficient time be given. 

 Again, the triplet in the arc spectrum at wave-length about 2942 — 

 2937*5 is not usually seen in the spectrum of the spark, but when the 

 plate has had a lengthened exposure the strongest two lines of this 

 triplet make their appearance in the spectrum of the spark. Even the 

 triplet near M, so strong in the flame of burning magnesium, but 

 not before recognised either in the arc or spark in photographs taken 

 with short exposure/'' comes out in plates of the spectrum of the 

 Spottiswoode induction spark (if we may give this name to the method 

 of stimulating the induction coil by the intermittent current of a 



* Dr. Huggins has informed us that his old photographs of the magnesium 

 epark taken with an induction coil in the ordinary way show this triplet distinctly. 



2 i 2 



