Dispersion and Selective Absorption of Fuchsln. 225 



dark, and a match may again be produced by narrowing the 

 slit S'. and if we know the original width of the slit the ratio 

 of the change in width to the original width giyes rery closely 

 the jn-oportion in which the introduction of the fuchsin has 

 diminished the intensity of the light. If instead of diminish- 

 ing the intensity at S' by narrowing the sht a reyolving 

 sector is used which cuts do^vn the light by the proper 

 amount, the error due to the lack of proportionality between 

 intensity and slit-width does not enter in. Howeyer, it is 

 found more conyenient to use a sector cut into an arbitrary 

 number of parts, as shown in fig. 6, and as with this w-e can 

 in general only make an approximate match, the varying of 

 the slit-width is used as a fine adjustment. This giyes prac- 

 tically as good results as can be obtained with a variable 

 sector. A sector of eighty notches was used, except for the 

 measurements within the absorption-band, in which case a 

 tin disk, as shown in fig. 7, was used having a slit cut in the 

 edge O'T mm. wide and another about a centim. from the 

 edge 1*5 mm. wide. This disk was rotated very rapidl}' 

 before the collimator slit by means of a small motor, and thus 

 the intensity of the light incident upon the collimator slit 

 was diminished by 99*856 or 99'682 per cent., according to 

 which of the slits of the disk w^as used. Since the edges of 

 the slits of the disk are parallel, care had to be taken to 

 have the collimator-slit at a definite distance from the centre 

 of the disk. This was done by using the disk close up against 

 the collimator-slit, and then by holding a sharp edge on a 

 line B or C scratched on the disk, the shadow of this edge 

 would fall on the centre of the collimator slit when the centre 

 of the disk was at a distance OC or OB from the rav enterino- 

 the collimator. 



The use of this tin disk, which cut down the intensity of 

 the light so greatly, made it possible to match very accurately, 

 although the light had been reduced by the fuchsin to such 

 an enormous degree. Since the slope of the absorption-curve 

 was very steep, it was necessary to have the collimator-slit S 

 quite narrow, not more than half a millimetre wide at most, 

 in order that the readings should be made with sufficiently 

 homogeneous light, otherwise the absorption as observed 

 would have too low a value. With the slit S, half a milli- 

 metre wide, S' would have been less than a five thousandth 

 part of a millimetre wide when the adjustment was made 

 usino; no sector, and accurate settino^s of the slit at this width 

 would have been impossible. This feature of the photometer, 

 together with the fact that a bare flame or direct sunlight 

 can be used instead of a flame or sunlight behind ground or 

 Phil Mag. S. 6. Vol. 6. No. 32. Aug. 1903. Q 



