142 
MR. W. GEOFFREY DUFFIELD OX THE 
remarkable agreement with one another. Diagrams I. and II. agree in assigning a 
low value to the shift of this line. 
Line 4 also belongs to this group, but the presence of a faint line very close to it 
vitiated its measurements, which therefore have not been included in the diagram. 
Group III. consists of lines which are invariably broad and diffuse under pressure, 
rendering their measurement a matter of great difficulty. As in the case of Group I., 
it is possible that two groups are united in Group III., but here again sub-division is 
SET A. 
0 10 SO 30 40 50 60 ’ 70 80 30 100 
Pressures /n Atmospheres. 
Diagram III. 
not at present feasible, though there is some indication that the lowest members— 
the two lines dS and hO —are to be treated apart from those exhibiting the greatest 
shift —f 1, f 2, f 3, Diagram II. The difficulty of determining the true position of 
these lines is accentuated by the shape of their intensity curves, which under 
pressure are nearly flat-topped. 
Group IV.—The existence of a fourth group is suggested by the displacements 
of the lines g6, g7, but it is possible that their abnormal displacements are only 
apparent. (See p. 157.) 
