148 
MR. W. GEOFFREY DUFFIELD ON THE 
Group I. Reversed lines. 1*8 
Group II. Unreversed lines .... 2'Q 
Group III. Unreversed lines . . . 1'8? 
Group III. Reversed lines. 2'0 
The ratio for the unreversed lines of Group I. is difficult to determine, since the 
readings of the plates made by different observers are not very concordant; it does 
not fall below 2. 
The value for Group III. unreversed is derived from the readings for f 1, f 2, f'3, 
in which the value for Dl9 (reversed) is seen to be 0'9 times the mean value for the 
other three plates (unreversed). Section 16, p. 151, indicates that for this group the 
reversals are displaced roughly half as much as the emission lines, consequently, if 
unreversed lines in one case could be compared with unreversed lines in the other, the 
ratio of the displacements would be 1'8 to 1. 
From Table VI. it may be seen that the photographs at 20 atmospheres exhibit 
the same phenomenon for Groups II. and III. ; the values for the different lines are 
given in the Tables I. and II., and a comparison of them shows that the same kind of 
difference exists at this pressure as was found at 25 atmospheres. 
At 15 atmospheres, also, discrepancies appear to exist between the values given by 
different plates for some of the lines, but at this pressure only the measurements for 
Group III. are discordant. There is a similar, but less marked, difference between the 
values for Group III. at 10 atmospheres; here it is the plate belonging to Set B that 
shows the higher values. 
It has been suggested that the phenomenon may be due to different parts of the 
arc having been focussed upon the slit during the two exposures. Each photograph 
is, however, the integration of a number of short exposures due to the intermittent 
nature of the arc, and what is obtained on the photographic plate appears to be the 
average effect of the whole arc; nevertheless, the fact that reversals are found on all 
the plates shows that during each exposure we have been dealing with a hot central 
core surrounded by an absorbing envelope. 
15. The Displacement and the Tendency to Reverse, 
There is some evidence that the abnormal displacements measured in the region 
between 1 5 and 30 atmospheres are connected with the great tendency of the lines to 
reverse at these pressures. The Displacement Tables, I. and II., in which the lines 
which are reversed are distinguished from the rest, show that at 25 and 20 atmospheres, 
in Sets A and B respectively, the maximum number of reversals occurs, and the 
photographs also indicate that at these pressures the reversals are specially broad 
and strong. These are precisely the pressures at which the displacements in the 
two sets show their most marked departures from a linear relationship. 
