252 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1919. 
this short general account of opals. A glass may be required which, 
while remaining clear in thin pieces after removal from the flame, 
will " strike " on reheating so readily that the temperature needed to 
develop the full density of the opal is not high enough appreciably 
to soften the glass and so cause deformation. One way of securing 
this behavior is to add as an ingredient a small amount of a sub- 
stance which will produce a trace of a compound insoluble in the 
glass except at very high temperatures. In the flame this compound 
persists as a slight turbidity and appears to facilitate the " striking " 
by affording nuclei on which the opal material can collect. 
Colored glasses are sometimes divided into two main groups : (a) 
Those in which the coloring matter is diffused in very small particles 
throughout the glass, and which may be likened to colloidal solu- 
tions; and (b) those in which the coloring substances are in a state 
resembling that of solution, and which may be likened more nearly 
to aqueous solutions of colored salts. Just, however, as in aqueous 
solutions there may be traced or inferred all grades of subdivision of 
the coloring matter from separate particles which can be revealed 
by their scattering action on light, and which may be seen in the 
ultra microscope through smaller and smaller particles scattering 
light less and less obviously down to those in the extreme state of 
subdivision frequently described as that of true solution, so in glasses 
similar grades of subdivision of their coloring matter may be seen 
or inferred. 
It is in fact impossible sharply to divide colored glasses into these 
two groups ; but it can be said of certain glasses that they are typical 
of group (a), and among the more common of these may be men- 
tioned those owing their colors to the presence of gold, copper, and 
selenium. It is generally considered that these coloring agents exist 
in the glasses as metallic gold, metallic copper and elementary 
selenium respectively, and the varying colors which can be obtained 
in each case appear to depend on the state of division of the coloring 
agents, or at least to be associated with it. How far there is evidence 
that selective absorption of light has also to be taken into account, 
is a question which can hardly be dealt with in a short time. 
With gold the colors most readily obtainable range from red to 
blue through varying stages of purples. With copper in the metallic 
state the common color is red; but it is possible to get variations 
very similar to those seen in gold glass, and a copper glass giving 
a definite blue by transmitted light has been obtained. It is to be 
understood that this blue was not due to copper in an oxidized condi- 
tion, but to metallic copper. Selenium glasses are also generally red, 
but again states of division of this material can be secured which 
give other colors, although it is difficult, except on a very small scale, 
