On the Spectra of Pigments. 349 



Emerald green affords a spectrum from which the purple is 

 excluded, while a delicate shade overspreads the green. A 

 dark shade is seen in the least refrangible red, and a lighter 

 shade in the rest of the red and yellow. 



Yellow ochre, purple, and blue shut out, dark shade in green 

 nearest the blue. 



Rose madder, shadow in purple and blue, band in green. 



By pursuing these experiments with a variety of coloured 

 objects, an insight will be gained into Nature's method of 

 painting, and some curious optical facts may be ascertained. 

 When water-colour paints are laid on white paper, if trans- 

 parent, they allow a certain portion of white light to be 

 reflected from the paper ; and if opaque, it will generally be 

 found that certain roughnesses in the paper have remained 

 comparatively unaffected by the pigment. More white light 

 will also be reflected at certain angles than at others. Allow- 

 ance must be made for these facts ; and with some colours the 

 experiments may be conveniently varied by transmitting a 

 bright pencil of light through the paper on which they are 

 placed, and contrasting the effects of transmitted and reflected 

 light. 



Having tried the spectrum of cobalt as a water-colour 

 pigment, take a piece of the blue glass with a violet tint 

 commonly used in decoration, and said to be coloured with 

 cobalt; this gives a fine spectrum with transmitted light. 

 On trying a piece rather more than 1 — 16" thick, a broad 

 dark band was seen at the edge of the red and yellow, 

 next to which was a band of dingy red, only visible under 

 strong illumination, occupying part of the place of the yellow, 

 then an apple green space, and then a dark green band, with a 

 paler green space leading to the blue. Another piece of blue 

 glass, rather thinner and with less of the violet tinge, gave a 

 similar spectrum, with some slight though interesting diffe- 

 rences. The dark tint between the two bands in the red and 

 yellow was paler, and less opaque. Experiments should be 

 tried with various tints of smalt glass, of different thicknesses. 



