REFRACTIVE INDICES. 97 
terminations can be made, the observer is able to select roughly those 
obliquely transmitted waves which vibrate normal to a principal section, 
and these remain the same even for different angles of inclination. In this 
respect, the Becke method, if applied to a vertical edge as illustrated in 
Fig- 57 is superior to the method requiring greater oblique illumination. 
Other methods of refractive index determination of mineral plates in the 
thin section, as that of Due de Chaulnes* and its later modifications, by 
E. Becquereland A. Cahours.f by A. Bertin.J by C. Viola, || and by W. O. 
Hotchkiss, are not sensitive nor accurate enough for satisfactory work 
and, although useful in certain instances, will not be considered further in 
this paper. 
In practical work with refractive liquids the writer has found it con- 
venient to use a set of liquids of refractive indices ranging from i . 450 to 
i .840, the refractive index of each successive liquid differing from that of the 
foregoing by 0.005. The refractive indices of the liquids are determined 
directly on an Abbe- Pulf rich total refractometer and their constancy 
checked every three months at least. ^[ Experience has shown that the in- 
dices of the liquid mixtures do not vary over o . 002 in a year, while the aver- 
age change in refractive index of a liquid for temperature rise is about o . ooi 
decrease for every 3 C. 
The following is the set of liquids at present used in the Geophysical 
Laboratory for this purpose. It differs from previous sets** slightly, in 
that the number of liquids employed is smaller and their dispersion rela- 
tively low. The refractive indices were measured on a total refractometer 
and those above 1.74 by the hollow-prism method in sodium light, ff The 
glass hemisphere of the Abbe-Pulfrich total refractometer (w Na = 1.907) is 
soft and so readily attacked by chemicals that it should not be exposed to 
the action even of methylene iodide, as with time this liquid attacks its 
surface and destroys the polish; the same holds true for the more highly 
refracting liquids of the list below. The liquids are kept conveniently in 
small dropping bottles with ground-glass dropper and cap, which interposes 
two ground joints to prevent evaporation. The set is so prepared that 
the refractive index of each successive liquid is 0.005 higher than that of 
the one preceding it. The mixtures are shown in table 4, page 98. 
In the preparation of the mineral grains for examination, the following 
details have been found by experience to be useful. Larger fragments are 
broken up and reduced to powder 0.05 mm. or less by tapping them (not 
rubbing) with a pestle in an agate mortar. A short cylinder of brass, into 
*Hist. Acad. R. Sci.. Paris, 1767; 1770, 162; Mem. Acad. R. Sci.. 1767, 423. 
tComptes Rendus, II, 867-871, 1840; Pogg. Ann., 51, 437-432, 1840. 
tAnn. Chim. Phys.. 26, 288-296, 1849; Pogg. Ann., 76, 611-612, 1849. 
NT. M. P. M., 14, 554-562, 1895; 16,150-154, 1897; also Rendiconti, R. Acad. Line. 5, 212-216, 1896. 
{Am. Geol.. 36, 305-308, 1905. 
^Refractive index indicators have been suggested by Michel-Levy (Etude sur la determination des 
feldspaths, Paris, 62, 1894); and also by Souza-Brandao (Centralblatt f. Min.. 14-18. 1904). 
**Amer. Jour. Sci. (4). 27, 35, 1909; for the measurement of the dispersion of the liquids used in the prepa- 
ration of the present set, the writer is indebted to Dr. H. E. Merwin. of this Laboratory. 
flAnton Pauly (Zeitschr. f. Wiss. Milcrosk.. 22, 344-348. 1905) has recently described an ingenious 
method for measuring the refractive index of a liquid by means of sections of calctte and siderite cut parallel 
with the axis. By noting the angle 2<f> between the positions at which the mineral plate disappear* in the 
liquid, he calculates the refractive index of the liquid from the equation 
I i sin* *+* cos* * 
where <> and are the principal refractive indices of the mineral. Pauly claims an accuracy of * or 3 in the 
fourth decimal place for this method. 
