The Microscope. 105 



wick of an alcohol flame and bringing this light into the instru- 

 ment by reflection from the small mirror. This will enable the 

 observer to produce a map in which the bands will coincide in 

 relative position to such spectra as may be mapped in daylight. 



If the spectrum of a very small object is to be examined, 

 the tube containing the prisms should be removed and the slit 

 opened wide. The object should then be placed upon the mi- 

 croscope stage and focussed as under ordinary conditions. The 

 objective used should be as low a power as will show the object 

 properly. The latter should then be brought to the center of the 

 field, and the slit closed ; also the side shutter should be closed 

 by the two small levers, so that no light can pass through the 

 slit except that which passes through the object. Then by 

 throwing the object slightly out of focus and replacing the tube 

 containing the prisms, its spectrum will be seen. 



For fluids small glass cells should be used, and it is fre- 

 quently convenient to have them made wedge-shaped, so that 

 different thicknesses of the solution may be examined. At 

 times, when the solution contains but little color, it becomes 

 necessary to use very deep cells, so that a considerable thick- 

 ness may be used. For this purpose glass tubing may be cut in 

 short pieces, the desired length, and after grinding the ends off 

 flat, cementing them to the ordinary slides by marine glue. 



The thickness of the solution will make a great difference 

 in the spectrum. No. 5 of the plate shows the result of a thick 

 solution of eosin. Only a little of the less refrangible end of 

 the spectrum is transmitted, while in No. 6, which is produced 

 by a thinner solution, a large part of the blue end is trans- 

 mitted. If a still thinner solution be used, the result will be 

 a narrow band at E. If a thicker solution of permanganate of 

 potash than the one which produces the spectrum shown at No. 

 9, be used, the five comparatively narrow bands become so 

 wide as to run into each other, and thus produce one very broad 

 band somewhat similar in position and width to that of the thin 

 solution of eosin. Thus it will readily be seen that varying 

 thicknesses of the substance should be examined until the char- 

 acteristic bands are obtained. When the bands are very faint 

 a prism giving a short spectrum should be used ; as they then 

 appear very iutense. In no case, however, is it of any very 



