1 6 B. E. Moore 



eral of these lines may be associated in close groups. Some of 

 these groups have s- and p- nearly equal. Other groups will have 

 p- stronger than .?- or vice versa. The behavior of these lines led 

 to placing them in a separate table (Table XIX), and to record 

 the intensities for the no-field lines (designated "no f") as well 

 as the intensities of the p- and ^-components. Special citations 

 from this table would be scarcely necessary unless one could col- 

 lect them into definite, well-defined, and related groups. A fur- 

 ther characteristic is that this list of lines begins very suddenly 

 at its red end, near the beginning of the carbon band. The blue 

 end of the list is nearly reached at 3700 A, but there seem to be a 

 few scattering lines belonging to the class as far as 3656 A. There 

 are very few symmetrically separated lines toward the red end 

 of the group, but before reaching the end of the group the sym- 

 metrical lines predominate. 



There were some of these lines of zero separation which did 

 not seem to have exactly the same position upon the no-field .t- 

 and />-plates. The lines upon the no-field plates were, in general, 

 exceptionally sharp. But these unsymmetrical lines were always 

 broadened. Take lines 3864.04 A and 3751.22 A. When the corre- 

 sponding p- and j-components are compared with each other, one 

 sees an appreciable difference in position. When they are com- 

 pared with the corresponding no-field lines, the ^-components fall 

 upon one side and the /'-components upon the other side of these 

 broad lines. Two adjacent lines, 3870.30 A and 3870.13 A, have 

 the ^-components of the former displaced toward the red, and of 

 the latter displaced toward the blue. The no-field lines are again 

 broad in both cases, and the s- and /'-components sharp. Also, 

 these ^-components do not fall outside of the no-field line. Some 

 other illustrations might be taken from the table. The error nat- 

 urally increases here, and one should not attach much importance 

 to the recorded values of the latteral displacements. But that 

 there are displacements of some of the s- and /'-components with- 

 out separation, as well as a variation in the ratio of their respective 

 intensities from line to line, seems probable. This' same broadness 

 or diffuseness upon one side occurs for a few no-field lines which 

 are found to have one of the corresponding ^-components broader 



104 



