two plano-parallel plates, which are set in brass frames. One of these plates is fixed, while the other has a horizontal sliding motion, for which purpose it is provided at the bottom with a side to side rack and pinion motion, worked by the milled-head M. The frame on the upper edge of the movable plate carries the divided scale, and that of the fixed plate a vernier. For reading the latter an inclined mirror s throws the image of the scale along the axis of the tube, which is fitted with a lens K. The piece, HI, contains the analyzer and the telescopic lenses, the eye-glass being movable. At H in the Fig. is shown a screw-head, which admits of the analyzer being adjusted to its proper position in relation to the fixed polarizer. It is only to be used, however, either by the makers in the first adjustment of the instrument, or when it has got out of proper adjustment-when, that is, we cannot by any movement of the quartz wedges bring a uniform tint into the two halves of the field. When this is the case, either the polarizer or analyzer must have been disturbed. To readjust the instrument, the two quartz wedges along with the fixed plate are removed, and the screw H turned until uniformity of tint in the field is attained. To fix the zero-point, another screw, not shown in the Fig., is attached to the quartz wedge frame which carries the scale, and allows of, a certain amount of the adjustment of the latter. Lastly, as regards the length of the solution tubes, the instrument is generally arranged for tubes 2 decimetres in length; but saccharimeters capable of taking 4 and 6 decimetre tubes are made, and are of service in the analysis of weak saccharine solutions.

§ 80. In using the saccharimeter, it is set up against a gas or paraffin lamp furnished with a metal chimney having a side opening (see Fig. 25, p. 102), care being taken to keep it a distance of at least 5 centimetres (2 inches) from the flame, to prevent the Nicol becoming heated. The manipulation is then as follows:

1. A tube, either empty or filled with water, is laid in the instrument, and the eye-piece of the telescope drawn out until the vertical joining line bisecting the circular bi-quartz appears sharply defined; next taking the milled-head M (Fig. 56), the movable quartz wedge must be adjusted, until over both halves of the field an approximately uniform colour prevails.

2. The regulator is then moved by means of the milled-head L of the side gearing, so as to bring into view that sensitive tint

which with the slightest displacement backwards or forwards of the movable quartz wedge produces the most distinct colourdifferentiation in the two semi-circles. Thus it will be found that if we start with a uniform deep-red or deep-blue tint, the quartz wedge has to be moved much further before we can distinctly recognize a difference of colour in the halves, than if we had started from a uniform bright violet tint. For most eyes the best position of the regulator is that which gives a sensitive tint most nearly approaching to white. Here a very slight movement of the quartz wedge will cause one half to exhibit a faint greenish tinge, and the other a flesh-pink, which on further turning of the milled-head, M, passes into green and orange-red respectively. In arranging the regulator for this position we do not get exactly the ordinary sensitive tint-the reddish-purple, which denotes most perfect extinction of the yellow light. Some of the yellow rays are allowed to pass, thereby making possible the green and orange colours which appear. It is, however, in other respects immaterial which tint is chosen for purposes of observation.

3. Having once decided upon the sensitive tint, the quartz wedge is adjusted to give the greatest possible similarity of colour in the halves of the field and the corresponding reading on the scale noted. This should be several times repeated, the mean of the readings being taken as the zero-point of the instrument. Should this not coincide with the zero-point of the scale, the latter must be adjusted, by slightly moving the screw referred to in § 79 as being fixed on the brass frame of the wedge. The position of the zero-point must be verified from time to time. Moreover, it is found to vary considerably for different eyes.

4. If the tube be now filled with a solution of cane-sugar and laid in its place, colour-dissociation at once takes place, and may be made to vanish again by moving the quartz wedge in such a way that the 100 point on the scale is made to approach the zero-point of the fixed vernier. If the solution be perfectly colourless, the screw L, Fig. 56, which manipulates the regulator, need not be touched; but if, on the contrary, as frequently happens, it has a yellowish tinge, we must disturb the screw a little to the right or left, until we obtain, as nearly as possible, the tint used in fixing the zero-point. By making several such adjustments of the quartz wedge to the position in which colouruniformity is shown, we arrive at an accurate determination of the rotation. With a little practice, it will be found easy to get observa

M

tions not varying from one another by more than 0.4 division of the scale at the most, provided that the instrument is well constructed. Where greater differences occur, as will be the case with coloured solutions, the number of observations must be increased. As a rule,

five are enough to determine the mean value within ± 0·1 of a division.

This instrument is, of course, unsuited for the colour-blind.

§ 81. The graduation of the saccharimeter scale, according to Ventzke, is made by laying in the instrument a tube 2 decimetres long, filled with an aqueous solution of pure sugar, having at the temperature of 17-5° Cent. a specific gravity of 1·1, and marking the point of the observed deviation as 100. A solution of the above density contains in 100 cubic centimetres exactly 26.048 grammes of sugar, so that it can be more easily prepared by weighing this quantity and adding the proper amount of water. The space between the 100 point and the zero-point is then divided into one hundred equal parts, and the graduation extended some way (30 or 40 divisions) on the other side of the zero-point. As cane-sugar is dextro-rotatory, the latter divisions of the scale will indicate lævo-rotation, and must be marked with the negative sign.

Assuming that deviation is exactly proportional to concentration, each division of Ventzke's scale will indicate a sugar value of 0.26048 gramme in 100 cubic centimetres of solution (or 2.6048 grammes per litre). Accordingly the process for estimating a saccharine solution is as follows:

To determine concentration (the number of grammes of sugar in 100 cubic centimetres of solution), we fill a 2 decimetre tube with the solution, note the deviation on the scale, and multiply the number of degrees by 0-26048. The solution must not, of course, contain more than 26.048 gramines of sugar in 100 cubic centimetres, otherwise the indication on the scale would pass outside the 100 point.

To determine per cent. composition (the number of grammes of sugar in 100 grammes of solution), weigh out 26·048 grammes of the solution, dilute to the mark in a 100 cubic centimetre flask, with this latter solution fill a 2 decimetre tube, and note the degree of deviation on the scale. This will indicate directly the percentage of sugar in the original solution.

The analysis of solid saccharines may be effected similarly :-To

estimate the value of a crude sugar, for instance, 26-048 grammes of substance are dissolved in water, diluted to 100 cubic centimetres, and observed in a 2 decimetre tube. The degree of deviation expresses directly the percentage of pure sugar in the substance.

If instead of 26-048 grammes we had used some other weight P of liquid, or solid saccharine matter, in preparing the 100 cubic centimetre solution, then, taking the degree of deviation observed in a 2 decimetre tube as a, the percentage composition will be

Moreover, the percentage weight of a sugar solution can be calculated from the concentration as determined directly, provided we know the specific gravity—that is, the weight of 100 cubic centi

metres.

When a tube 1 decimetre in length is used instead of a 2 decimetre tube, the scale-readings must, of course, be doubled; similarly, in using a 4 decimetre tube, they must be halved.

For the method of preparing sugar solutions for the saccharimeter, see § 92.

The original method as used by Ventzke was to prepare, by means of an areometer, a solution of the saccharine substance to be analyzed of a specific gravity 1.1, and to examine this in a 2 decimetre tube. The observed deviation was then taken as giving directly the weight per cent. of sugar in dry substance. This method, however, which was intended to dispense with all weighings, does not yield accurate results, as the salts contained in natural sugars always have a specific gravity different from that of the sugar itself. It has therefore been abandoned, but the awkward normal weight of 26.048 grammes continues in use. It would be much more convenient to substitute some simpler figure- as, for example, 20 grammes-in which case, however, a re-calculation of the tables prepared for Ventzke's instrument would be necessary.

§ 82. Correction of Saccharimeter-Readings for slight Disproportionality between Rotation and Concentration.-As shown by the researches of Schmitz and Tollens, already described § 37, the specific rotation of sugar is not constant for solutions of different concentrations, but increases inversely as the concentration. Thus, given that a solution containing 26-048 grammes of sugar in 100 cubic centimetres records 100 on the scale of the saccharimeter, the reading

50 on the same scale will not be recorded exactly by one containing 13.024 grammes, but by a solution containing a somewhat smaller sugar percentage. Hence the need, at least in the case of exact observations, of correcting the saccharimetric readings.

Schmitz1 has calculated the required correction, on the basis of the following observations:

According to these experiments the increase of specific rotation with decrease of concentration may be expressed by the following interpolation-formula, from which the calculated values in the table were obtained :

Putting [a], as the specific rotation calculated from this formula, for a normal solution containing 26.048 grammes of sugar in 100 cubic centimetres of solution, and [a], as that for some other solution of

[a]c lower concentration c, then represents the proportion in which the [a]n

concentration and per cent. composition will appear too high when estimated on the supposition that the rotation varies uniformly therewith. To obtain the true values the results must, therefore, be divided by the above fraction. This calculation is exemplified in detail in the table annexed.

Col. 1. N gives the scale-reading at every ten divisions.

2.

c, the number of grammes of sugar in 100 cubic centimetres of solution, assumed to correspond.

3. [a] and [a]c, the specific rotation of sugar for the above-mentioned concentrations (26.048 and c), calculated from the interpolation-formula.

1 Schmitz: Zeitsch. des Vereins für Rübenzuckerindustrie, 1878, 63.