4 BULLETIN 679, TJ. S. DEPARTMENT OE AGRICULTURE. 
position of which depends on the nature of the substance to be 
examined. With the alkaloids studied, solutions of potassium- 
mercuric iodid in glycerol and water were found to be fairly satis- 
factory. 
The stock solution is prepared as follows: Mix equal volumes of 
water and pure glycerol. In this dissolve 2.88 grams of potassium 
iodid and 3.38 grams of mercuric iodid per cubic centimeter of 
liquid; evaporate the solution on the water bath to the point of 
incipient crystallization; at 20° its index is approximately 1.71. 
Dilute this with glycerol (index 1.42) in the calculated proportions 
to make liquids with indices differing by a uniform amount, pref- 
erably 0.01. Measure the true indices of these liquids on arefractom- 
eter, or by other means, and correct the indices by the addition 
of glycerol or of stock solution, as necessary. The liquids are best 
kept in small dropping bottles, protected by glass caps. 
A few crystal grains are immersed in a drop of one of the liquids 
on a slide, and the relative indices of grain and liquid determined 
thus: The light entering the microscope is diminished by means of 
the substage diaphragm, and a grain is brought into focus. Then, 
on raising the microscope tube slightly, a band of light passes into 
that substance — crystal or liquid — having the higher refractive index 
in that direction. The more divergent the indices of the two, the 
more brilliant the band; while when they are nearly matched, the 
band becomes faint, and, if white light is used, shows pale blue and 
orange margins. When the two have the same indices, no light 
moves in either direction. For accurate matching, monochromatic 
light, such as that from a sodium flame, should be used. White light 
passed through a ray filter is convenient. The writer uses a com- 
bination of Wratten E-red No. 23 and B-2 Extra Light No. 57-A 
filters, 3 inches square, in glass. 
Crystals belonging to the cubic system have but one index, and 
their position is of no significance. Those of other systems, however, 
possess more than one, and the position in which they are observed 
must be taken into consideration. The grain to be studied should 
first be turned until it becomes dark or " extinguished" between 
crossed nicols. If the lower nicol prism, the polarizer, is inserted 
alone, two indices can be observed on any one grain, one when it is 
parallel, the other when it is perpendicular, to the direction of 
vibration of this nicol. If the elongation of the grain is positive, 
the highest index is shown when it is parallel, the lowest when per- 
pendicular; if negative, the reverse relation holds. When the sub- 
stance crystallizes in one of the less symmetrical systems, and has 
three different indices, grains lying in different crystallographic 
directions must be observed in order to measure all three. 
