variable composition; Endeman gives the formula of one variety as 16 Si O2, 3 Al1⁄2 O3, 5 Na, O. But all this is open to great doubt. The most competent authorities consider that the green ultramarine is not a true chemical compound, but a combination of the blue with sodium salts; because, by simply boiling with water, it is converted into the blue ultramarine, while soluble sodium salts are found in the water; on the other hand, by heating the blue ultramarine with sodium sulphate and charcoal, it is converted into the green ultramarine.

Gueckelberger, one of the most recent writers on the subject, confirms the figures given by Hoffmann; but considers that the ultramarines are derived from a typical compound containing Si18; thus, for the variety rich in silica he proposes the formula Si18 Al12 Na20 S 962; while the variety poor in silica has the formula Si18 Al18 Na20 S O71. It is doubtful whether ultramarines have the complex composition here assigned to them and, moreover, no light is thrown on their constitution.

ASSAY AND ANALYSIS.-As there is so much difference between various makes of ultramarines, it is necessary to assay for colour, fineness, body, &c. Those ultramarines which are to be used by paper-makers should be tested for their power of resisting the action of alum; this can be done by taking about 5 grammes and boiling in a solution of alum of about 5 per cent. strength. To see what change of colour may have taken place, 5 grammes of the colour should be shaken up with clean water and the two wet samples compared together; any change brought about by the alum can then be readily detected.

It is rarely that a complete analysis of ultramarine is required; in such an event, the following scheme can be adopted

:

For Water.-Heat 2 grammes in a weighed crucible for about half an hour over the Bunsen flame; the loss in weight is the amount of water present.

For Silica, Si O2-Treat 2 grammes with hydrochloric acid until the colour is completely destroyed; evaporate the mixture to dryness and gently ignite the residue; treat the dry mass with hydrochloric acid, filter off the insoluble silica, well wash it, then dry, and burn in a weighed crucible; the increase in weight minus the weight of the filter-paper ash is the weight of the silica.

For Alumina, Al, O.—To the filtrate from the silica add ammonia in slight excess, boil gently, then filter, and treat the precipitate of alumina as the silica.

For Soda, Na, O.-To the ammoniacal filtrate from the alumina add sufficient sulphuric acid to neutralise the ammonia, then evaporate to dryness in a weighed basin and ignite the residue

until all ammoniacal fumes have been given off; weigh the residue of sodium sulphate, and multiply this weight by 0.4366 to ascertain the weight of the soda, Na2 O.

Total Sulphur.-Treat 2 grammes of the ultramarine with a mixture of 2 parts of nitric acid and 1 part of hydrochloric acid, until the colour is completely decomposed and only a transparent mass of silicate is left; filter this off, and to the filtrate add a solution of barium chloride in excess, boil and filter, wash the precipitate well, dry, burn, and weigh it in a crucible. To find the weight of sulphur, multiply the weight of barium sulphate so found by 0.13734; from this deduct the weight of sulphur present as sulphuric acid to find the quantity of sulphur present as sulphide.

For Sulphur as Sulphuric Acid.-Weigh out 2 grammes of ultramarine, treat with dilute hydrochloric acid, filter off the precipitated sulphur and silica, and precipitate the filtrate with barium chloride, treat the precipitate as in the last. To find the amount of sulphur triaxide present, multiply the weight of the barium sulphate so found by 0.34335.

ULTRAMARINE DERIVATIVES.-It has been stated above that some of the constituents of the blue and green ultramarines can be substituted by other analogous bodies, such as selenium for the sulphur or the sodium by silver; in this way other ultramarines can be prepared, but, as a rule, they are only of scientific interest, as their colour is of no technical moment, as will be seen later on; hence they are not made on a large scale ; still their production may ultimately throw light upon the question of the chemical constitution of ultramarine, and the fact of their formation must be faced by all chemists who essay to deal with this question. There are one or two coloured derivatives of ultramarine which are used to a limited extent; these are the violet and red ultramarines.

VIOLET ULTRAMARINE.--This product can be made from either the green or blue ultramarines, from which it differs by containing less sulphur and more alumina. It can be made in several ways. Zeltner makes violet ultramarine by submitting either the blue or green varieties to the temperature of about 300° C., and passing chlorine gas over them; at first the colour is, if the blue is used, turned green, then this becomes dark red; at this point the operation is stopped and the red product is boiled in an alkaline solution until it turns violet; after which it is washed. Instead of using dry chlorine at a temperature of 300° C., the green or blue varieties may be heated in a mixed current of steam and chlorine at a temperature of 160° to 180° C.,

until the colour is developed; after which it is washed with water to free it from the sodium chloride formed, and dried.

In another method, devised by Hoffmann, blue ultramarine is 'mixed with about 2.5 per cent. of ammonium chloride; the mixture heated to a temperature of 200° C., is exposed to the air until the violet colour is properly developed, and the mass allowed to cool slowly; when cold, it is washed thoroughly and dried.

Violet ultramarine has very similar properties to the blue variety, and is similarly decomposed by acids; boiling in alkalies changes the colour to blue. The shade of the pigment is a very pale reddish-violet. This pigment is not used to any great extent owing to its want of colouring power. In composition it resembles the blue varieties; a sample analysed by the author contained :

RED ULTRAMARINE.-Zeltner prepares the red ultramarine by exposing the blue variety at a temperature of 130° to 150° C. to the action of the vapours of nitric acid, when he obtains deep or dark red, or light rose or pink shades, according as the acid vapours are dilute or strong. Hoffmann prepares the red ultramarine by passing dry hydrochloric acid gas over either the blue or violet varieties until the proper colour is developed, when the mass is washed and dried.

OTHER ULTRAMARINES.-By using boracic acid instead of silica a boron ultramarine of a blue colour is obtained. Yellow ultramarine is made by heating the blue ultramarine with a solution of silver nitrate, in sealed tubes, at a temperature of 120° C. for 15 hours; the sodium is replaced by silver, and the new pigment contains 46.5 per cent. of silver.

The blue ultramarine heated with silver chloride turns green, taking up silver in the process. The yellow silver ultramarine heated with sodium chloride loses some of its silver, turning green; if the sodium chloride is replaced by potassium chloride a bluish-green potassium ultramarine is formed. If barium chloride is used a yellowish-brown barium ultramarine is obtained; in the same way zinc chloride yields a violet zinc ultramarine, and

magnesium chloride a grey ultramarine.

have any technical value.

None of these products

The sulphur can be replaced by selenium, when brown and purple ultramarines are obtained.

PRUSSIAN BLUE.

such

Prussian blue, or Berlin blue, or Chinese blue, is, next to ultramarine, the most valuable blue pigment in use for painting and other purposes for which pigments are used. It was discovered in the early part of the last century (about 1704) by a Berlin colour-maker, named Diesbach, by accident, as many discoveries have been made. Diesbach was making Florentine lake, and for this purpose he used a solution of cochineal, which he mixed with alum and copperas (ferrous sulphate) and precipitated with an alkali; in the particular instance which led to the discovery of Prussian blue he used an alkaline solution (which had been used to purify some Dippel's oil made by distilling ox blood), and instead of getting a red lake he got a blue. Diesbach followed up this discovery and found that the blue could be got by calcining blood with alkali, and, after lixiviating the mass, precipitating the liquor with a solution of copperas.

The technical manufacture of this pigment was further developed by a London colour-maker, named Wilkinson, from whom the colour was named "Wilkinson's blue," a name which is now obsolete. Wilkinson prepared the colour by first deflagrating a mixture of tartar and saltpetre, and calcining the residue with dried blood; the fused mass was lixiviated with water, and to the lye so obtained a solution of alum and copperas was added; the resulting pale blue precipitate was treated with hydrochloric acid to develop the blue.

Since the time of Diesbach and Wilkinson the composition of Prussian blue has been the subject of numerous researches by chemists, so that very little remains to be learnt as to the composition and constitution of this blue.

Prussian blue is a compound of iron, carbon, and nitrogen; the carbon and nitrogen are combined together in the form of the radicle cyanogen, CN, which is the characteristic element of a group of compounds, of which Prussian blue is a member, known as cyanides. The iron exists in the blue in two forms; one in combination with the cyanogen in an acid condition, the other in the basic condition. When Prussian blue is boiled with a solution of potash it yields oxide of iron (which remains as an insoluble red mass), and a yellow solution, which, on being

4

allowed to crystallise, deposits yellow tabular crystals of a compound of potassium, iron, carbon, and nitrogen, originally known as yellow prussiate of potash; the iron in this exists in combination with the carbon and nitrogen in an acid state, forming the radicle known as ferrocyanogen, Fe Co N. The chemical name of the yellow prussiate of potash is potassium ferrocyanide, K Fe Co N. Besides the yellow prussiate there is another, the red prussiate or potassium ferricyanide, K, Fe Co Ng. These two compounds differ from one another in their colour and in the reactions which they give with iron salts. With ferrous salts the ferrocyanide gives a bluish-white precipitate of ferrous ferrocyanide; while with ferric salts a deep blue (Prussian blue) precipitate of ferric ferrocyanide is obtained. With ferrous salts the ferricyanide gives a deep blue precipitate of ferrous ferricyanide (Turnbull's blue); with ferric salts no precipitate is obtained, but the colour of the solution becomes a little darker. The production of Prussian blue is a most characteristic reaction of iron, no other metal is capable of producing it, and very minute traces of iron in a solution can be detected by adding a few drops of a solution of potassium ferrocyanide.

It has been ascertained by various observers that the precipitates obtained by adding solutions of the prussiates to solutions of iron salts contain potassium as an essential part of their composition, and it is difficult, although possible, to rid them of this potassium. Thus the bluish-white precipitate is really potassium ferrous ferrocyanide, K, Fe Fe CN; Prussian blue is potassium ferric ferrocyanide, K, Fe, 2 Fe Co Ng; Turnbull's blue is potassium ferro-ferricyanide, K Fe Fe Co Ng

2

6

*

Prussian blue and Turnbull's blue have exactly the same composition, but their constitution is different; the one being a ferrocyanide and the other a ferricyanide. Skraup, Reindel,† Kekule,‡ and other authorities consider that they are identical; but that they are different is proved by the fact that when the alkali is eliminated from them the residual blues are of different composition, Prussian blue having the composition Fe, C18 N189 and Turnbull's blue the composition Fe, C12 N128; then the shade of the two blues is different, Prussian blue is a greenishblue, while Turnbull's blue is a violet-blue.

commerce

COMMERCIAL PRUSSIAN BLUE. In several varieties of Prussian blue are sold under the names of

§ Williamson, Mem. Soc. Chem., iii., p. 125; Reynolds, Jour. Chem. Soc., li., p. 644. See also E. J. Parry on Blues," Analyst, September, 1896.

Composition of Commercial Prussian