240 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
stripes. Hitherto only spectral rays of greater wave length have been 
attempted; mixed colors have not been attempted, nor have white 
and gray, to which most natural colors owe their luminosity, been 
attempted. 
The examination of these points appeared extremely interesting, 
as it seemed possible to account for special phenomena, which mathe- 
matical calculations can not explain. Thus the disappearance of 
white, but not colored, portions through overexposure, the general 
tendency toward red or dirty yellow, the appearance of white with 
excessive intensification, the general shift of the colors toward the 
more refrangible end of the spectrum when the pictures are rubbed, 
the frequent wart of the complementary colors by transmitted light, 
the appearance of black or violet on rubbing the white, the extinction 
of the colors, except white and black, in varnishing, and so on. 
EXPERIMENTAL METHODS. 
The methods employed by the author are briefly as follows: 
1. The plate is soaked in water and the film scraped or stripped 
off with the edge of a freshly broken piece of glass. If the film is 
not very thin this always takes place from the glass or in that part 
of the film which contains no laminz. The author also uses collodion- 
ized glass. 
2. The stripped film is immersed in alcohol and water, then in 
absolute alcohol, and finally for a few minutes in celloidine. 
3. Fine sections are cut at right angles to the film and laid in water 
to swell. 
Sometimes the water is replaced with glycerine, and the film stained 
with an aniline dye insoluble in water. After some experience one 
may use a still simpler plan, and that is to hack the damp film along 
and across with a sharp scalpel, to then cover the cut places with a 
cover glass, and examine in this way, when one or more pieces showing 
the laminve will be easily seen. 
THE GRAIN OF PLATES. 
Lippmann and others who work the process contend that the trans- 
parent emulsion in albumen or gelatine has no grain, or only such 
that as regards the wave length of hight it can be neglected. Neu- 
hauss, however, proved the existence of a grain almost invisible before 
exposure, but which after development varied between 0.1 and 0.3 p. 
The author considers this far too high an estimate, as it would be 
hardly possible with such a grain to register the half-wave length of 
violet light (4A=0.171 »). From various experiments he believes that 
he is not far out in putting the size of the grain at 0.02 to 0.05 p.¢ 
4 He speaks here of the emulsion which will register all colors up to violet. 
That which will only record red and yellow has a much coarser grain. 
