254 
CAPTAIN W. DE W. ABNEY ON THE TRANSMISSION 
maximum intensity. The aperture was then closed to read 80°, then 70°, and so on 
till 20°, after which every 5° were used till the final opening was itself 5°. The 
coloured glass was then placed in front of the comparison-light, or the plain glass 
substituted for the mirror, and the readings recommenced from 90° until 5° was 
again reached. The light from the spectrum at this point would ordinarily be so 
feeble that it would be impossible to read rays of feebler luminosity. Three readings 
of each setting of the sectors were made and the mean taken. This involved thirty- 
three readings, or sixty-six in all, and the reverse order of the observations was 
again carried out, i.e., commencing with 5° aperture and the absorbed (or partially 
reflected) light up to 90° aperture, and then from 5° back to 90° with an unabsorbed 
comparison-light. The time of the commencement of each cycle of readings was 
noted, and the mean taken as the correct time applicable to the mean of the whole 
set. Twenty-five minutes generally sufficed to make the double set of measures. 
For two hours on each side of noon the alteration in the sun’s zenith distance during 
twenty-five minutes is not sufficient to make any very material alteration in the 
relative proportions of the rays, and the mean of the two sets may be taken to be the 
reading of all the rays at the mean time of observation. 
Later in the day, when the sun’s zenith distance rapidly increases, the mean values 
derived from the cycles of observations may not exactly correspond to the mean time, 
though, perhaps, not very far from it. 
§ Y. Comparison of Results. 
I am not at all disposed to think that observations taken during a whole day are 
likely, as a rule, to give a true value for the coefficients of transmission of the 
different rays, more particularly in a climate such as that of England. The 
atmospheric conditions often vary greatly between the evening and morning, and I 
have come to the conclusion that by combining observations of the sun at different 
altitudes, but at approximately the same hour, the minimum values of absorption for 
each ray are more likely to be correctly determined. As a rule, just the contrary 
mode of proceeding has been taken. In the determination of atmospheric absorption 
of stellar or lunar light by Bouguer, Seidel, Pritchard, Muller, Pickering, and 
Langley the days’ observations have been compared together. It is probable that the 
atmospheric conditions obtaining at night are more equable than those in the day. 
§ VI. Atmospheric Conditions most suitable at the Time of Observation. 
A hazy day is essentially an unfavourable day for taking such observations as I had 
in view, and I have only used, in my final result for minimum absorption, observa¬ 
tions made on such days as appeared suitable from a meteorological aspect. A still 
day is usually a slightly hazy day, and I preferred, where possible, to utilise those 
