1905 - 6 .] Dr J. Halm on the Spectro-heliometer. 85 
tances for annual and lunar shift, and collect the observations 
in groups according to the hour-angle of the sun. The diurnal 
displacement, as the formula shows, depends on the declination of 
the sun. We may assume, however, that in all the groups the 
factor cos 8 is the same, viz. 0’957. Table III. shows the 
observed distances for different hours of the day (true Edinburgh 
time). 
Table III. 
Hour. 
Distance 
in t. m. 
Hour. 
Distance 
in t.m. 
h. 
21*71 
0*3727 
h. 
0*45 
0*3694 
22*46 
*3718 
0*65 
*3691 
22*75 
*3727 
0*88 
*3691 
23*02 
*3715 
1*26 
*3676 
23*28 
*3700 
2*18 
*3673 
23*51 
*3712 
2*54 
*3676 
23*67 
*3703 
2*87 
*3667 
23*87 
*3703 
3*24 
•3658 
0*06 
*3700 
3*72 
*3664 
0*26 
*3697 
4*60 
•3652 
5*68 
•3649 
Representing these values by a formula 
A + b. sin 
we find for A and b the values 
A = 0*3700 t.m. and * b = — 0*00502 ± 0*00017 t.m. 
or in km. per sec. b= — 0*240 ± 0*008. 
The true value should be 0*245 km. per sec. Hence we find 
again an agreement within the limits of the probable error of the 
observed quantity. The observed annual and diurnal shifts are 
graphically represented in the figs. 4 and 5. The smooth curves 
indicate in each case the shifts due to the actual motions. 
Ho doubt a more strictly scientific method, but also a far more 
troublesome one, would have been to determine the two shifts 
simultaneously from the measurements. I am confident, how- 
ever, that the result would have been the same. 
The outcome of the preceding calculations seems to be interest- 
ing from two points of view. Hot only have the observations 
