sultant of the uncovered portions and partly of the wire, the former appearing as the continuation of the latter.

We may I think conclude from these experiments that the coincidence of the wire with the resultant of the inclined lines, instead of proving this resultant to be impressed on the eyes as a single line, merely shows that in respect to relief both it and the wire are viewed under the same visual conditions.

It is to be remarked that for these experiments the lines of the diagram ought to be as slender as is consistent with great distinctness and the wire as thin as it can be made without a loss of rigidity. When the lines are broad or the wire thick there is an appearance of singleness near the point of intersection due to the overlapping of the two images which in the case of slightly inclined lines may extend through much or all of their length.

14. Tapering form of the perspective resultant.

In repeating these experiments it will be observed that while the wire coinciding with the perspective resultant appears of unchanging thickness throughout its whole length, this resultant seems to increase slightly in diameter from the near to the remote end. The cause of this difference although obvious, is not unworthy of remark. In the case of the wire the retinal picture in each eye is of course wider for the near than for the remote end, and would if uncorrected, convey the impression of an object tapering towards the farther extremity. But the different distances at which through varying convergence and other adjustments we see the different parts of the wire, serve to rectify this impression, and under the perception we obtain that of an object of uniform thickness throughout. In the case of the perspective image, on the other hand, the retinal picture in each eye is of uniform breadth, because the lines of the diagram are so and are practically in every part of their length at the same distance from the eyes. Thus the diameter of the resultant disc or lines actually subtends the same angle in both eyes for every part of the length. But in virtue of its perpective position, its different parts seem to be placed at different distances. Hence the diameter of any part of the resultant must appear greater in proportion to the distance at which it seems to be placed, and it will appear to grow broader as it is traced from the near to the remote end.

The effect here described is scarcely appreciable with most of the stereoscopic drawings, owing to the small range of distance through which the perspective extends. By using two inclined lines each three or four inches long and combining them when placed on the upper stage of the stereoscope I obtain a resultant whose remote end is more than twice as distant as the near one, and in this case the enlargement towards the farther extremity is very obvious.

A simple mode of obtaining the same effect in a yet more manifest degree is the following. Adjust two slender cylindrical rods six or seven inches long, so that, while they nearly touch at one end, they spread apart about an inch at the other. Hold them before the eyes R, L, at the distance of distinct vision in a vertical plane transverse to the line of view and with the angle downward (fig. 26). Then form the binocular resultant by directing

27.

L

R

the axes beyond them with varying convergence. This resultant stretching away from the angle obliquely towards the wall will be seen to increase in diameter regularly from its near to its remote end.

From these facts it may be inferred that in order with the stereoscope to obtain a perspective figure, all the lines of which shall be of precisely equal thickness compared with one another, and each of them of the same thickness at every point, it would be necessary so to construct the drawings that the breadth of the corresponding parts of the component lines should be proportioned inversely to the distances at which they are to be united.

(To be continued.)

ART. XXI.-On the identity of Sanguinarine and Chelerithrine, and on the direct determination of Nitrogen; by Dr. JAMES SCHIEL of St. Louis, Mo.

I HAVE used a small portion of hydrochlorate of Chelerithrine which I received in 1843 from my friend, Prof. H. Will of Giessen, and which the discoverer of that alcaloid had prepared himself, to make a few analytical experiments as to its composition.

As the salt was not found to be free from impurities, it was dissolved in water, precipitated by ammonia, washed, dried, dissolved in ether, the filtered solution treated with animal charcoal, and the chelerithrine precipitated with a solution of pure sulphuric acid in ether; the sulphate of chelerithrine was washed with ether, dried and dissolved in water, when it yielded with ammonia a precipitate of pure Chelerithrine, which absolutely showed the same properties and behavior as Sanguinarine.

Of the pure substance dried at 105° C.,

0-356 grammes burned with oxyd of copper and oxygen, furnished 0.918 carbonic acid 70-34 per ct. of carbon and 0.167

==

water 521 per ct. of hydrogen.

=

The direct determination of nitrogen gave 5:07 per ct.

0-3925 gramm. of the double salt of the hydrochlorate of chelerithrine and chlorid of platinum yielded 0·707 of platinum: thence the atomic weight = 341.74.

Accordingly the composition of Chelerithrine is:

The atomic weight was found in three determinations 322-7; 362-7; 346-4, or taking the mean =343.9.

From all this it appears that Chelerithrine and Sanguinarine are one and the same substance corresponding to the formula C38 H.NO. This formula gives:

It will be perceived that the above formula includes one equivalent of carbon more than the one I formerly deduced from the analysis of Sanguinarine.

For the preparation of one or the other of these alcaloids the following method is the simplest and cheapest.

Digest the root of the plant Sanguinaria Canadensis (or Chelidonium majus) with water strongly acidulated with sulphuric acid and precipitate with ammonia, wash and dry the precipitate, dissolve in ether, and treat with animal charcoal. After filtration the alcaloid is precipitated with a solution of pure sulphuric acid in ether. It is pure sulphate of sanguinarine.

* Annalen d. Chemie und Pharmacie by Liebig and Wöhler, B. XLIII, p. 233.

For the determination of nitrogen I do not use the bicarbonate of soda but the following apparatus, from which a current of car bonic acid is passed through the combustion tube.

IN

39

B

A is a Woulfe's bottle containing diluted muriatic acid.

B, a similar bottle with pieces of chalk and water.

C, Chlorid of calcium tube, part of it filled with chalk; a piece or two of pumice stone separates the two substances. T, Combustion tube.

Condensing the air in bottle A by blowing through the mouthpiece a, some of the acid passes into B and sets carbonic acid free, which passing through A washes the atmospheric air contained in B mostly out, then close the mouth-piece with a cork and open the little stop-cock D. When a part of the gas passing throughthe combustion tube is entirely absorbed by the chalk, the stop cock D is closed, the cork taken out from the mouth-piece, and the combustion is performed in the common way. After the combustion, another current is passed through the combustion tube by opening the stop-cock and closing the mouth-piece, having previously driven some acid into B. In case of need the stopcock could be omitted, but the little expense hardly justifies such a simplification.

ART. XXII.-On an apparent Perturbation of the Law of Defin ite Proportions observed in the Compounds of Zine and Antimony; by JOSIAH P. COOKE, Jr., Cambridge.*-With a plate.

In a former paper in this Journalt I described two new compounds of zinc and antimony Sb Zns and Sb Zna which I named respectively, Stibiotrizincyle and Stibiobizincyle, because they resemble in their composition the metallic radicals of organic chemistry, and because the first decomposes water rapidly at 100° C. I there stated that crystals of Sb Zns could be obtained containing a much larger amount of zinc than that required by the law

* Abstract from a Memoir of the American Academy, New Series, vol. v, p. 337. This Journal, vol. xviii, p. 284.