Intelligence and Miscellaneous Articles. 155 



second order of the spectrum was taken by it. Theoretically 

 speaking, the time necessary for taking a photograph of the first 

 order is \ that for taking one of the second, and -Jy for that of the 

 .third. Practically this is not quite true, for reasons which it is 

 unnecessary to enter into here. This induced me to think that 

 if I could get a grating with double the number of lines to the 

 inch, its first order would give me the same dispersion as the 

 second order of the grating I was using, and at the same time the 

 exposure ought to be more than halved. Mr. Lockyer kindly lent 

 me such a grating ; and the third photograph was taken by it. 

 You will see that it gives the lines from C to A almost perfectly. 

 A positive copy of this photograph I sent to Professor Piazzi 

 Smyth, as he drew the map of this region of the spectrum, which 

 he published in the last volume of the ' Edinburgh Astronomical 

 Observations ;' and I cannot do better than quote his words re- 

 garding the accuracy of the photograph : — " As the size of the 

 glasses approximated closely to my standard-camera size, 2*25 inches 

 X 4:'2o inches, 1 was able at once to view them with any magnify- 

 ing power in a compound microscope long since arranged for such 

 things ; and the effect was astounding. I almost thought I was 

 back again in Lisbon, viewing the Sun's spectrum itself as I used 

 to see it." 



This acknowledgment of the value of the photograph was 

 particularly gratifying, as it came from an astronomer who 

 had paid special attention to this particular part of the 

 spectrum. 



The next plate shows the region of the ultra red as taken with 

 the same grating. The most conspicuous group has a wave-length 

 of about 8400 tenth metres ; and the extreme line visible has a 

 wave-length of about 9200 tenth metres. At this point the 

 closer-ruled grating seemed to fail me, and I could not get beyond 

 this point. It then struck me that the glass on which the lines 

 were ruled might absorb the rays beyond (for I must explain 

 that this grating was ruled on glass and silvered at the 

 back), or that the red glass might absorb them. I then 

 reverted to the first grating and adopted a different method of 

 proceeding. 



This was Praunhofer's method, to which I was practically new — 

 though I believe I have been credited (though inaccurately) with 

 using it for other experiments ; and after various attempts the 

 following is an outline of the arrangement adopted. The slit 

 was placed horizontally ; a prism of 60° occupied a position at the 

 end of the collimator next the lens; the rellection-grating then 

 received the rays and reflected them into the camera-lens, the 

 camera being tilted at an angle to make them fall on the sensitive 

 plate. To show what sort of effect can be got, I exhibit a pho- 

 tograph in which from the first to the fourth order of spectra 

 were impressed on the same plate in a small camera. When this 



