PHYSICS: NICHOLS AND HOWES 
309 
This is obviously a composite curve due to the superposition of phosphor- 
escence of both types. It does not follow however that it is a true example of 
a single substance brought to both types of phosphorescence by photo-excita- 
tion. The evidence is to the contrary. 
1. The two varieties of willemite are commonly associated. One of our 
specimens contains parallel veins of the persistent form and of willemite of 
short duration in the same matrix. 
2. Willemite is associated with the Franklin Furnace calcite. 
That the vanishing phosphorescence in the curve in figure 5 may be due to 
an admixture of calcite is suggested by the following observations. 
a. When the disk of the phosphoroscope, coated with the powdered will em- 
ite and exposed to the light of the iron spark, was driven rapidly, the region 
nearest the spark in the direction of revolution but shielded from the direct 
light was a brilliant yellow-green, the regions approaching the spark but 
60 1 
WILLEMITE. U-V. EXC. DISK. 
40 
/% 
20 
1. 
.10 .20 .30 SECONDS .40 
FIG. 5 
shielded, were of a dark blue-green. This is the effect which would be expected 
were calcite present, since its red-yellow phosphorescence would give the yel- 
lowish caste during the first process of decay but would vanish before the revo- 
lution of the disk was completed. 
b. When the collimator of a spectroscope was directed to the rapidly revolv- 
ing disk the spectrum of the phosphorescent willemite appeared as a broad 
continuous band extending from the extreme red to the blue. The crest was 
approximately at 0.54 /x and there was suspicion of a weak region in the yellow. 
At a somewhat lower speed, which allowed a longer interval of time between 
excitation and observation the red end of the spectrum disappeared and at a 
much lower speed the yellow and brightest part of the green vanished leaving 
a much narrower band in what had been a comparatively feeble region, with 
its crest in the blue green at about 0.52 /x. 
The former brilliant crest at 0.54 fi was now very dim and lay near the edge 
of the persistent band. This persistent crest coincides in position with that 
of the fluorescence spectrum of a willemite of long duration determined spec- 
trophotometrically some years ago. 9 
