NO. I INFRA-RED SOLAR SPECTRUM ABBOT AND FREEMAN 3 



in which a stretched wire was displaceable over a milk-glass back- 

 ground. Lines were considered provisionally veridical when found 

 as deflections of similar form and similar setting in three or more 

 holographs. After completing this preliminary study, the positions of 

 all deflections considered possibly veridical were measured on three 

 holographs with the excellent Warner and Swasey comparator de- 

 scribed on page 64 of Volume I of the Annals. Mean values were 

 computed of positions on three (or in some cases two) of these 

 holographs on which the deflections were found. When found on only 

 two of the three they were questioned, and rejected unless supported 

 by further evidence. 



In assigning intensities, Mr. Freeman used practically the same 

 system that was used in Volume I of the Annals. Grades a, b, c, d, 

 and d? were employed. All lines falling within great bands like 

 A, par, (fi, ij/, and C! are joined in a bracket and designated as a whole 

 with " a." Bands hardly reaching this first-class prominence are 

 similarly bracketed and marked " b." Individual deflections, or com- 

 posites of several small deflections which altogether make a depres- 

 sion of 5 mm. or more in holographs are marked *' c." Smaller indi- 

 vidual deflections, whether in the bands or outside of them, are marked 

 " d." When considerable doubt remains as to the veridical character 

 of such a deflection, it is marked " d ?." We do not guarantee that 

 all the lines included in the table are veridical, but we believe a very 

 large proportion of them are so. The curves are very free from 

 accidental deflections as deep as a single half millimeter, and the 

 repetition on several holographs of similar configurations of intensity 

 '■' d " is regarded as strong presumptive evidence of reality of corre- 

 sponding solar or terrestrial absorption lines. 



To determine the wave-lengths of the lines observed, the advice of 

 Dr. Babcock was sought. From his studies of all existing laboratory 

 determinations of infra-red line spectra, amplified by his own extensive 

 photographic work in the upper infra-red spectrum as far as A= 

 1.1018 Angstroms, he sent a list of 112 identifications of wave-length 

 places, given according to our holographic work in Volume I of the 

 Annals, as compared to more recent determinations. A curve of cor- 

 rection to the wave-lengths given in Volume I of the Annals has been 

 prepared from this material. In summary it is as indicated in table i. 



The data for corrections beyond 1.18/A are so scanty and so conflict- 

 ing that there seemed no justification for applying any corrections in 

 that region. 



