ami its Observers. 123 



Professor Gibbs did not count the lines, nor does he profess to 

 have counted them, but that he saw the lines separated. 



Yours truly, 

 Mr. C. Stoddek. JosiAH Curtis. 



Extracts from a letter from Dr. F. A. P. Barnard, President of 

 Columbia College, New York. 



President's Eoom, Sept. 27, 1870. 



My dear Sir, — Since the bands of Nobert's plate are so exceed- 

 ingly narrow, it seems to me not extravagant to claim that when with 

 a power not greater than a yV^^' ^^^ jjortion of the breadth of one of 

 them is truly resolved, the whole breadth should be so likewise. If 

 they were four or five times as broad, I should certainly regard Colonel 

 Woodward's assumption as too sweeping. On the other hand, in my 

 counts which I made with a filar micrometer in the sirring of 1868, 

 and which I described to you at the time, I found it extremely difiicult 

 to count the whole breadth of the higher bands, principally because the 

 manipulation of the micrometer distiu'bed, sooner or later, the sharj)ness 

 of the definition. I did not count the 19th band entire. My method 

 was this — first by means of the coarser bands used as a stage micro- 

 meter, to obtain the value of the revolution of the screw of the filar 

 micrometer with the given objective ; then by bringing the higher 

 band into the field, to coimt as long as I could, and to determine the 

 values of the intervals between the lines counted, and thus by infer- 

 ence the number of lines to the inch from the reading of the screw. 

 I actually counted from twenty to thirty lines — never more, I think, 

 than thirty. As your letter suggests, Sulivant and Wormley coimted, 

 apparently, for the purpose of verifying Nobert's statements in regard to 

 the plate, Nobert himself objected to one of Colonel Woodward's early 

 photographs, that it could not represent the band which it was claimed 

 to represent, because the number of lines in the entire band was too 

 small. This is a legitimate criticism of a photograph, because the 

 distance between the lines of this print would not indicate the distance 

 between the lines ruled on the jjlate imless the enlargement should be 

 exactly known — and about that there might be some reasonable doubt 

 — but the question, whether the lines are resolved when the real 

 distance between them is known, is settled the moment that the 

 measured distance between any two such lines distinctly seen is 

 found to accord exactly with the real distance. When, for instance, I 

 found that the value by micrometer of twenty spaces on the 19th 

 band as counted was exactly equal to the value by the same micrometer 

 of ten spaces on the 9th band, I could not doubt that the 19th was 

 • resolved. I subsequently mounted the micrometer on a detached 

 stand, so that turning the screw should not disturb the microscope, 

 but my eye or my illumination failed me, and it did not improve uj)on 

 the former performance, or even do so well. 



I do not wish to be quoted as absolutely dissenting from Colonel 

 Woodward's proposition, but I do hold that whether or not the whole 

 breadth of a band is so resolved to be coimtable, the question whether 



