is thus obtained for every flame-coloring metal; this spectrum consists either, as in the case of baryta, of a number of colored lines lying side by side; or, as in the case of lithia, of two separate, differently-colored lines; or, as in the case of thallium, of a single green line. These spectra are characteristic in a double sense-viz., the spectrum lines have a distinct color, and they occupy also a fixed position.

It is this latter circumstance which enables us to identify without difficulty, in the spectrum observation of mixtures of flame-coloring metals, every individual metal. Thus, for instance, a flame in which a mixture of potassa, soda, and lithia salts is evaporated, will give, side by side, the spectra of the several metals in the most perfect purity.

KIRCHHOFF and BUNSEN have constructed two kinds of apparatus, which are both of them suited for spectrum observation, and enable the operator to determine by measure the positions in which the spectrum lines make their appearance. Both are constructed upon the same principle. A description, with illustration, of the larger of the two, which is also the most perfect one, has been published in POGGENDORFF'S Annalen, 113, 374, and in the Zeitschrift für Analytische Chemie, 1862, 49. The smaller, more simple, and accordingly cheaper apparatus, which suffices for all common purposes, and is used most in chemical laboratories, we will describe here. It is shown in fig. 29 a.

A is an iron disk, in the centre of which a prism, with circular refracting faces of about 25 mm. diameter, is fastened by a bow, which presses upon the upper face of the prism, and is secured below to the iron plate by a screw. The same disk has also firmly fastened to it the three tubes, B, C, and D. Each of these tubes is soldered to a metal block, of which fig. 29 b gives an enlarged representation. This block contains the nuts for two screws, which pass through wider openings in the iron plate, and are firmly secured beneath when the tube has been adjusted in the proper position. B is the observation telescope; it has

a magnifying power of about 6, with an object-glass of 20 mm. diameter. The tube C is closed at one end by a tin-foil disk, into which the perpendicular slit is cut through which the light is admitted.* The tube D carries a photographic copy of a millimetre-scale, produced in the camera obscura on a glass plate of about one-fifteenth the original dimensions. This scale is covered with tin-foil, with the exception of the narrow strip upon which the divisional lines and the numbers are engraved. It is lighted by a gas or candle-flame placed close behind it.

The axes of the tubes B and D are directed, at the same inclination, to the centre of one face of the prism, whilst the axis of the tube C is directed to the centre of the other face. This arrangement makes the spectra produced by the refraction of the colored light passing through C, and the image of the scale in D produced by total reflection appear in one and the same spot, so that the positions occupied by the spectrum lines may be read off on the scale. The prism is placed in about that position in which there is a minimum divergence of the rays of the sodium line; and the telescope is set in that direction in which the red and the violet potassium lines are about equidistant from the middle of the field of view.

The colorless flame into which the flame-coloring bodies are to be introduced, is placed 10 cm. from

the slit. BUNSEN's lamp, shown in fig. 22, gives the best flame. The lampis adjusted so as to place the upper border of the chimney about 20 mm. below the lower end of the slit. When this lamp has been lighted, and a bead of substance say of sulphate of potassa-introduced into the fusing zone by means of the holder shown in fig. 27, the iron disk of the spectrum apparatus, which, with all it carries, is moveable round its vertical axis, is turned until the point is reached where the luminosity of the spectrum is the most intense.

To cut off foreign light in all spectrum observations, the centre part of the apparatus is covered with a black cloth or card-board box.

REXROTH'S spectroscope, fig. 30, may be safely recommended as useful and economical.f It is very different in its arrangement from the appa

Fig. 30.

ratus just described. bb is a cylinder fixed onto the heavy disk a, and closely covered by c. de is a telescope, and il a bent tube,

This arrangement is not very strong. For laboratories a more solid material such as brass would be much better than tin-foil.

+ Zeitschrift f. anal. Chem., 3, 443.

which are fastened to c; il is so drawn that its internal construction may be plainly exhibited, is a moveable slit, k is a reflecting prism, and m is a lens. The rays coming through the slit are made parallel by m, and then reflected from the mirror g onto the flint-glass prism f, by which they are refracted up the telescope. By means of the screw h it is possible to move g round a horizontal axis, and thus to bring the different points of the spectrum one after the other into the middle of the field of view. Opposite that side of ƒ from which the rays leave the prism, is a side tube, o, containing the photographed scale and n being a lens by which the rays coming from o are made parallel; the rays are thus reflected from the upper face of the prism into the telescope, and become visible simultaneously with the spectrum. The scale is fixed, and being lighted by a small oil-lamp or a gas-flame, is equally visible throughout the whole field of view. The sodium line is set by means of h, so that in the case of lines which are remote from each other, as for instance the blue potassium lines, we are enabled to read off their positions very conveniently on the scale, by moving the sodium line. 20 divisions towards the left, and thus bringing the lines to be measured more into the middle of the field of view. Under these circumstances we should have to add 20 to the number read off, and on the other hand, if the sodium line were moved 20 divisions towards the right, in order to bring a red line more into the middle, 20 would have to be subtracted from the number read off.

The spectra produced by the alkalies, the alkaline earths, thallium and indium are shown in the picture at the commencement of the book. The solar spectrum has been added simply as a guide to the position and bearings of the lines. The spectra are represented as they appear in the apparatus furnished with an astronomic telescope. In the third section, in the chapters treating of the several bodies, attention will be called to the lines which are inost characteristic for each metal. Here I will simply state the manner in which the highest degree of certainty is imparted to spectrum analysis. This is done by exposing the beads of the pure and unmixed metallic compounds to the flame, and marking on copied scales the position which the most striking spectrum lines occupy on the scale of the apparatus, in the manner shown, by way of illustration, in fig. 2 of the picture, with regard to the strontium spectrum. It is self-evident that the spectrum of an unknown substance can only pass for the strontium spectrum, if the characteristic lines not only agree with those of the latter in point of color, but appear also in exactly the same position where they are marked on the strontium scale.

pur

The drawings of such scales every operator must, of course, make for his own apparatus; and they become useless for the intended pose if any alteration is made in the position of the prism or the scale. It is therefore always advisable to set the apparatus so that it can be easily readjusted to its original position, which is most readily done by making the left border of the sodium line coincide with the number 50 of the scale.

With the introduction of spectrum analysis a new era has, in many respects, begun for chemical analysis, as by means of this discovery we can detect such minute quantities of bodies as by no other method. Spectrum analysis is marked moreover by a certainty above all doubt, and gives results in a few seconds, which could formerly be obtained only, if at all, in hours or days.

APPENDIX TO SECTION I.

§ 18. APPARATUS.

As many students of chemical analysis might find some difficulty in the selection of the proper apparatus, I append here a list of the articles which are actually required for the performance of simple experiments and investigations, together with instructions to guide the pupil in the purchasing of them.

1. A BERZELIUS SPIRIT LAMP (§ 16, fig. 14).

2. A GLASS SPIRIT LAMP (§ 16, fig. 16). Or, instead of these two, where coal-gas is procurable, a BUNSEN'S Gas-lamp, best one with chimney (§ 16, figs. 17, 18, and 22).

3. A BLOWPIPE (see § 15).

4. A PLATINUM CRUCIBLE.

Select a crucible which will contain

about a quarter of an ounce of water, with a cover shaped like a shallow dish; it must not be too deep in proportion to its breadth.

5. PLATINUM FOIL, as smooth and clean as possible, and not very thin length about 40 mm.; width about 25 mm.

6. PLATINUM WIRE (see pp. 18 and 24). Three stronger wires and three finer wires are amply sufficient. They are kept most conveniently in a glass filled with water, most of the beads being dissolved by that fluid when left in contact with it for some time; the wires may thus be kept always clean.

7. A STAND WITH TWELVE TEST-TUBES-16 to 18 cm. is about the proper length of the

tubes, from 1 to 2 cm. the proper width. The tubes must be made of thin white glass, and so well annealed that they do not crack even though boiling water be poured into them. The rim must be quite round, and slightly turned over; it ought not to have a lip, as this is useless and simply prevents the tube being closely stopped with the finger, and also shaking the contents. The stand shown in

Fig. 31.

fig. 31 will be found most suitable. The pegs on the upper shelf serve for the clean tubes, which may thus be well drained.

8. SEVERAL BEAKERS AND SMALL FLASKS of thin, well annealed

glass.

9. SEVERAL PORCELAIN EVAPORATING DISHES, AND A VARIETY OF SMALL PORCELAIN CRUCIBLES. Those of the royal manufacture of Berlin are unexceptionable, both in shape and durability. Meissen and Nymphenburg porcelain will also answer the purpose.

10. SEVERAL GLASS FUNNELS of various sizes. They must be inclined at an angle of 60°, and merge into the neck of a definite angle.

11. A WASHING BOTTLE of a capacity of from 300 to 400 c.c. (see § 7). 12. SEVERAL GLASS RODS AND GLASS TUBES. The latter are bent, drawn out, &c., over a Berzelius lamp or gas-lamp; the former are rounded at the ends by fusion.

13. A selection of WATCH-GLASSES.

14. A small AGATE MORTAR.

15. A pair of small STEEL or BRASS PINCERS, about four or five inches long.

16. A WOODEN FILTER STAND (see § 5).

17. A TRIPOD of thin iron, to support the dishes, &c., which it is intended to heat over the small spirit or gas lamp.

18. The Colored Glasses described in § 17, especially blue and green.

SECTION II.

REAGENTS.

§ 19.

A VARIETY of phenomena may manifest themselves upon the decomposition or combination of bodies. In some cases liquids change their color, in others precipitates are formed; sometimes effervescence takes place, and sometimes deflagration, &c. Now if these phenomena are very striking, and attend only upon the action of two definite bodies upon one another, it is obvious that the presence of one of these bodies may be detected by means of the other: if we know, for instance, that a white precipitate of certain definite properties is formed upon mixing baryta with sulphuric acid, it is clear that, if upon adding baryta to any liquid, we obtain a precipitate exhibiting these properties, we may conclude that this liquid contains sulphuric acid.

Those substances which indicate the presence of others by any striking phenomena are called reagents.

According to the different objects attained by the application of these bodies, we make a distinction between general and special reagents. By general reagents we understand those which serve to determine the class or group to which a substance belongs; and by special reagents those which serve to detect bodies individually. That the line between the two divisions cannot be drawn with any degree of precision, and that one and the same substance is often made to serve both as a general and a special reagent, cannot well be held a valid objection to this classification, which is in fact simply intended to induce a habit of employing reagents always for a settled purpose-viz., either simply to find out the group to which the substance belongs, or to determine the latter individually.

Now whilst the usefulness of general reagents depends principally upon their efficiency in strictly characterizing groups of bodies, and often