38 Mr. J. K Lockyer. On the Wave-length of the [May 2, 



" I find, on reference to my note-book, that the words written at 

 the time are £ map showed reading between lines of tin.' It was in 

 expanding this sentence for my paper that the words ' about midway ' 

 came in. I noticed when the paper was in proof that the expression 

 was not strictly accurate, but I allowed it to pass, as the sentence 

 was obviously merely a rough indication of position for the selection 

 of the metal for comparison. There is absolutely no ground for the 

 suggestion of any shift, as the position of the aurora line from the 

 direct comparison with tin agrees with the former one (No. I)." 



From the above, it appears that the statement that the aurora line 

 fell " about midway between " the lines of tin was not strictly 

 accurate, and this consequently led to a misunderstanding] . 



The importance of actually confronting the spectrum of the aurora 

 with that of manganese, and definitely determining the coincidence or 

 non-coincidence of the brightest fluting of manganese with the chief 

 aurora line is therefore obvious, and no opportunity of making the 

 observations should be lost. 



In support of the suggestion that the chief aurora line occupied the 

 same place in the spectrum as the brightest fluting of manganese, I 

 •quoted the observations of Smyth and Proctor, which placed the line 

 considerably less refrangible than Angstrom's observation (5567). I 

 attached no special weight to these observations beyond the fact that 

 I knew the observers to be absolutely trustworthy, and there can now 

 be no doubt that they were justified in differing from Angstrom. 

 Gyllenskiold's statement of Proctor's value (5595 + 25*0) was based 

 upon a curve constructed from the data contained in a small sketch 

 of the spectrum with an empirical scale given in ' Nature ' (vol. 3, 

 p. 346), and is, therefore, not so important as Proctor's own state- 

 ment : " My own measures give me a wave-length very slightly 

 greater than those of Winlock and Angstrom."* 



Winlock's value being 557, it seems pretty certain that the line is 

 decidedly less refrangible than 557. Gyllenskiold estimated the 

 probable error of Proctor's value as equal to one division of the scale 

 (the distance between C and F being about 96 divisions) or more 

 than four times the distance between the D lines. 



With reference to Krafft's observations at Bossekop in Alten, Dr. 

 Huggins says : "I have already pointed out that Krafft's measures 

 were not made under circumstances which assured to them a high 

 degree of accuracy, and Krafft's own words, which I have quoted, 

 disclaim expressly any special attempt on his part to redetermine the 

 position of the principal line with a higher degree of accuracy than 

 the observers who preceded him." Dr. Huggins does not appear to 



* < Nature,' vol, 3, p. 384. 



