48 ROYAL SOCIETY OF CANADA 
the light about the fibre. The intensity of these bands can be varied 
at will from total blackness to disappearance by increasing the dis- 
tance of the fibre from the focus. This change of distance can be read 
off on the scale, and constitutes an arbitrary scale of intensity. 
From a consideration of the aqueous vapour lines in Rowland’s 
tables, it was decided to use the portion in the neighbourhood of the 
D lines, as the vapour lines there are very pronounced. The important 
ones whose wave lengths are 5901.682 and 5919.S60 were at last se- 
lected. 
Specimen observations upon the intensity are plotted in the 
accompanying curves. The variations in 5901.682 are represented 
in the left-hand curves, those in 5919.860 in the right-hand ones. The 
abscissæ are devoted to the hours of the day at which the readings 
were taken, while the ordinates are the readings on the millimeter 
scale. Each ordinate is the mean of four observations at one time. 
CONCLUSIONS. 
1. This method of measuring the intensity of the lines is capable 
of considerable accuracy. When the spectrum is ordinarily bright 
the greatest and least readings at any one time differ by less than 
two millimetres; however, in order to obtain the best results, the eye 
and also the fibre must be kept as steady as possible, and all extraneous 
light should be eliminated. A further precaution is necessary, viz., 
a specific portion of one diffraction band should be chosen to compare 
with an adjacent portion of the line; otherwise the possible lack of 
perpendicularity in the fibre will prove a source of error. If these 
precautions are taken, a difference of about one millimetre may be 
found in the readings. 
2. An inspection of the curves shows that the intensity varies 
considerably throughout the day and also from day to day. Especially 
intense are they before storms. This fact is made clear by the obser- 
vations on January 20th. Although the day was fine, the lines were 
about the darkest in the whole winter’s observations. The same 
statement would be true of their number, for one or two lines in the 
band of Angstrôm’s map were observed, which have not since appeared. 
Usually one sees in the spectrum given by the grating that is used 
three lines between the D lines, but on days when the lines are very 
intense as many as seven may be detected and many new lines appear — 
on both sides. The phenomena detailed above were followed by a 
severe sleet storm. 
3. Generally the noon readings are less than the morning or 
evening; however, exceptions are noted on February 12th and March 
a a 
