0. C. Lester — Oxygen Absorption Bands. 149 



The measurements of the stronger lines were made as usual 

 with a micrometer microscope. Some of the weak lines had 

 to be strengthened by a fine mark placed on the back of the 

 negative. Others were measured from Rowland's charts, care 

 being taken to set on the center of density of the lines. This 

 last method is better for weak lines, although even here a line 

 mark must sometimes be used, as any magnification by the 

 micrometer causes some of them to melt into the background. 



The group a' was first noted by Jewell.* In the " head " or 

 first band of this group, many of the lines appear double and 

 some foreign lines seem to be present. The line 5T89"40, 

 which is the " chief line " of this band corresponding to simi- 

 lar lines in A, B, and <z, has been assumed double, as it is in all 

 the other bands. The only indication of being double actually 

 shown is its greater intensity and a certain flatness of the 

 intensity curve characteristic of close doubles. The uniformity 

 of the two series also calls for a close double at this place. 



The approximate positions of many of the lines in the 

 group called a" were calculated from relations established 

 between the other groups, the observed values differing by 

 less than *2 of a unit from the calculated. The lines are all 

 extremely weak. Some, though not all of them, appear on 

 negatives taken in zero weather, which indicates that they are 

 not water- vapor lines. The first band of the group begins as 

 usual with a double line, possesses a chief line and a final pair 

 in its proper position, as a glance at the groups as shown in 

 Plate III will show. Probably not all the lines present can be 

 seen. Many are so faint as to be visible only on the charts 

 and then only when they are held in particular positions with 

 respect to the light. Some of the lines are stronger on the 

 corresponding chart of Rowland's first series, which is more 

 intense than the second. No attempt has been made to meas- 

 ure many of the lines of this group closer than the nearest 

 half-tenth, which is readily done by estimation. Blunders and 

 mistakes in calculation for all groups except A have been prac- 

 tically eliminated by the use of verniers made to fit Rowland's 

 charts, which, in spite of irregularities in the map scales, enabled 

 any but small mistakes to be detected at once. 



* The Absorption Spectrum of Oxygen ; Astron. and Astrophys., xii, 1893. 



