wires of the telescope have been previously made to coincide with some one of the Fraunhofer lines, and the dark band be now brought to cover it, the reading on the graduated disc will give the angle of rotation for that particular ray. In a similar manner the amounts of rotation can be determined for other portions of the spectrum.

§ 63. Instead of solar light, the use of which is of course very much restricted, artificial light may also be employed for this method of observation. V. von Lang's method of using lithium, sodium, and thallium light, is first to adjust the cross-lines of the telescope upon the lines produced by volatilizing a salt of the respective metal in a Bunsen burner; then, replacing the Bunsen by the luminous flame from an Argand, to bring into view the continuous spectrum necessary for the production of the dark absorption-band. By rotating the Nicol until the band is made to coincide with the cross-threads the position of which must, of course, remain undisturbed during the experiment—the angle of rotation for the given ray is obtained. In the same way the light from a hydrogen Geisslertube can also be used. By these means it is possible to determine the angle of rotation for six different rays of known wave-length. In the subjoined table these artificial lines are collated with the Fraunhofer lines so as to indicate their position in the spectrum.2

v. Lang: Pogg. Ann. 156, 422.

2 The wave-lengths of the Fraunhofer lines given above (of which C' coincides with

The arrangement of the apparatus is shown in Fig. 43. It consists of a Mitscherlich polariscope, with its polarizing Nicol at

a, and its analyzing Nicol, with rotatory motion, at c; in front of it a spectroscope, consisting of a collimator, d, a prism, e, and a movable telescope, f. Instead of the usual cross-threads, the telescope is fitted Ha, D with Na, and F with Hg), and also of line Hy, are the means of the results obtained by Fraunhofer, van der Willigen, Ditscheiner, Ångström, Stefan, and Mascart (Wüllner's Lehrbuch der Phys. 3 Aufl. 2, 136 and 141). For Lia and Tl Ketteler (Beobachtungen über die Farbenzerstreuung der Gase, Bonn, 1865) gives the values 6706 and 5345, that of Na being taken as 5888; so that putting Na

obtain the values Li= 6712, and T1 = 5350.

5893, we

The lines of other metals might probably be made available, as the following:

Hurion (Pogg. Beiblätter, 2, 83). Lecoq de Boisbaudran (Spectres lumineux. Paris,

1874).

with two parallel vertical threads, separated from each other by a distance somewhat greater than the breadth of the black absorptionband. The determination of the angle of rotation of a given substance comprises the following operations:-1. The determination of the zero-point of the analyzer. The spectroscope is removed, and the experimental tube left empty or filled with water. The Argand lamp i serves as the source of light. 2. The spectroscope is. set up in its place, and in front of the Nicol a, the source of light h, which gives the lines (a Bunsen burner, with a bead of salt or a hydrogen tube); the analyzer is then rotated till the light can pass through. After widening pretty considerably the slit of the collimator d, the telescope f is moved horizontally until the bright line under examination lies accurately between the parallel threads, when the clamp-screw g is tightened. This position of the telescope can be more readily found when the Argand lamp i, with a small flame, is placed behind the Bunsen burner h, so that the threads stand out against a bright background. 3. The tube containing the active liquid is then laid in its place, the Argand light turned full on, and the analyzer moved on its axis until a broad dark band, with faint edges, appears in the field of the telescope. This also is brought between the parallel threads, and the angle read off on the graduated disc. Lastly, by reproducing the bright line again, it can be ascertained whether the position of the telescope has been disturbed in the meantime.

In observing the parts of the spectrum which lie near the extremities, as the red lithium-line, the Drummond lime-light should be used for the production of the continuous spectrum, the Argand being too weak in red rays sufficiently to illumine the edges of the dark band.

The above method does not admit of the same degree of concordance of results as when the polariscope is used alone, the position of the dark absorption-band being less definite, owing to the indistinctness of its edges. Thus von Lang, in the paper referred to, gives the following table as showing the varying positions of the analyzing Nicol in a determination of the rotatory powers of a quartz plate.

Observation-series 1 and 2 were obtained with an Argand lamp, 3 and 4 with a Drummond light, the position of the polarizer in the second being different from that in the first two series. The temperatures of the quartz plate are also given in the table; these exhibit

certain variations, but too small in amount to account for the differences in the observed positions.

§ 64. The experimental tubes of polariscopes should invariably be made of glass. They are generally 2 decimetres (about 8 in.) in length, but shorter ones, 1 decimetre, and longer ones, 3 decimetres and upwards in length, can be used where the construction of the instrument will admit of it. The internal diameter varies from 6 to 10 millimetres (-in. to 3-in.), and the thickness of the glass walls should be about 2 millimetres. The extremities are ground flat with great care, the ground surfaces being kept as nearly as possible perpendicular to the axis of the tube, otherwise it will be impossible to determine the length of the tube with any exactness. Moreover, it is convenient to have the internal walls of the tube ground with coarse emery powder, so as, by dulling the surface, to prevent disturbing reflections.

The mode of closing the ends of the tubes usually adopted is, as shown in Fig. 44, by plane parallel glass plates, which can be fixed down by a screw-cap, having a washer of india-rubber or soft leather between. As Scheibler1 first observed, this mode of closing

Scheibler: Ber. d. deutsch. chem. Gesell. 1868, 268.

the tubes may prove a source of error owing to the tendency of the glass plates under pressure to become double-refracting, whereby light passing through them is circularly-polarized. Moreover, when differences of tension exist in the body of the glass, owing to imperfect annealing, the same results will appear altogether apart from pressure, although the latter intensifies them. Indeed, by applying sufficient pressure to the glass plates, the errors thus arising may even amount to several degrees. It will be evident then how essential it is that all new glasses should be carefully tested, before they are taken into use.

For this purpose a series of observations should be made of the zero-point of the instrument, first with a tube open at the ends, then with a glass plate applied to one end under moderate pressure. Moreover, as the glass may be differently affected at different parts, the tube should be turned about on its axis so as to test every part. Generally speaking, if one glass is found to be circular-polarizing, all others of the same lot, cut at the same table, will prove to be so likewise.

§ 65. The tubes supplied with the apparatus are usually fixed within a simple brass tube, or they may be left uncovered; in either case the liquid inside will be exposed to the temperature of the surrounding air. Now as (see § 22) the rotatory power of most substances is materially affected by heat, it is necessary to be able to control the temperature of the solutions during the continuance of the observations. This can be done by enclosing the tube in a jacket of brass, 4 to 5 centimetres wide, and allowing water to flow between, supplied by a reservoir in which it has been previously raised to the desired temperature. Fig. 44 represents a tube enclosed in a

K