Porcelain

It would seem that it was the beauty of the glassy surface, enhancing the brilliancy of the colouring, rather than any practical advantage connected with its use, that first led to the application of glaze to pottery. The turquoise and green glazes of the Egyptians (the colour is derived from a silicate of copper along with soda and sometimes lime) were known to the men of the Early Empire. They were applied to a fritlike mass of sand held together by silicate of soda, to which the name of porcelain has sometimes been very wrongly given. Objects of steatite, of slate, and even of rock crystal were sometimes covered with a coloured glaze of this kind, but it was never applied to the clay vessels in daily use. These were made, then as now, from the unctuous clay of the Nile bank. For this restriction there was a very good reason, namely that a glaze of this nature, composed chiefly of alkaline silicates, will not adhere to a base of ordinary clay. It was not until Ptolemaic and Roman times that, by the discovery or adoption of a glaze containing lead, the ancients were enabled to glaze their pottery. So in Assyria, the employment of glazes was almost confined to the decoration of the surface of brickwork, the bricks being of a loose and somewhat sandy texture.17

In these glazes, and indeed in much earlier examples from Babylonia, both tin and lead have been found. The respective virtues of the silicates of these metals were doubtless appreciated, that of tin to form a white opaque enamel hiding the material below, and that of lead to enable the glaze into which it enters to adhere to a paste formed of a plastic clay.

With the Chinese the aim was rather æsthetic than practical. They sought by means of the marvellous glazes that cover their ancient porcelain to imitate the surface of natural stones; their early celadons were in a measure intended to take the place of the precious green jade, so highly esteemed by them.

At the time when the manufacture of porcelain was first introduced from China there were (apart from the salt-glazed stoneware, which lies quite outside our inquiry) three classes of glaze in general use either in Europe or in the nearer East:—

1. Glazes consisting essentially of alkaline silicates without either lead or tin. Such glazes could only be applied to a fritty silicious base, and in India and Persia their employment seems to have been a survival from Egyptian and Assyrian times.18

2. Opaque enamel glazes, the opacity being due to the presence of tin; a considerable amount of lead also is generally found in these glazes. We are not concerned here with the obscure origin of this group, but in the sixteenth century this enamelled fayence was in general use for the better class of table-ware. It includes the Italian majolica, the French fayence of Nevers and Rouen, and above all the earthenware of Delft.

3. The oily-looking lead glazes with which the common earthenwares were covered. These were essentially the glazes of the Middle Ages in Europe, and their employment could probably be traced back to the lead-glazed ware sparingly used by the Romans. We have already noticed the use of a similar glaze in Egypt as far back probably as Ptolemaic times.

There were practical objections to all these glazes. It is true that at Delft, by the use of the tin enamel, a ware could be turned out closely resembling, in external aspect, the blue and white porcelain of China, but the enamel was soft and would in time chip off at the edges, showing the dark earthy clay beneath. On the other hand, the alkaline glazes of the East were not much known in Europe; they can only be used upon a very tender and treacherous base. In India and Persia, however, a ware thus glazed still competes with the hard porcelain of the Far East. In spite of the great objections to the glazes of our third class, those containing lead—objections arising from their softness and from the danger of poisoning to those employed in their manufacture—their use has tended rather to increase. Not only is lead the principal constituent of the glazes still universally used for common pottery, but it forms an important element in the glaze of our finer earthenwares as well as in that of those bone pastes which rank with us as porcelain.

The glaze which had been brought to perfection by the Chinese at an early period differs from all those yet mentioned by its hardness, its high fusing-point, and in its chemical composition. Speaking generally, the glaze of porcelain differs in composition from the paste which it covers only sufficiently to allow of its becoming completely liquid at the extreme heat of the furnace; and just as the paste of Chinese porcelain has a wider limit of variability than that made in Europe, but is on the whole of a ‘milder’ type than the latter, so we find that while the glazes of the Chinese are as a whole less refractory and not quite so hard, there is still a wide range of variation in these qualities.

If, then, we theoretically regard porcelain as a compound of a silicate of alumina with an alkaline silicate of the same base, we may say that the glaze of porcelain is formed by the latter body alone, that it is, in fact, merely a fused felspar. But as in the case of the paste, so in the glaze there is generally present an excess of silica, derived from the quartz contained in the petuntse or pegmatite, and this silica enters into combination with some other bases which are present in the constituents of the glaze, thereby increasing its fusibility and modifying the contraction in cooling. The most important of these additional bases is lime, so that the more fusible type may be called a calcareous, as opposed to a more refractory or purely felspathic glaze. As much as 21 per cent. of lime has been found in some Chinese glazes, the amount of alumina being proportionately reduced.

There is more or less lime in the glaze of most kinds of European hard porcelain, but the exceptionally hard and refractory paste made at Sèvres since the time of Brongniart is covered by a glaze of corresponding hardness from which that earth is absent. This hard paste has, however, of late been replaced in part by one of a milder type, and with this latter a calcareous glaze has been adopted even at Sèvres, the object of the change being, as we have said, to allow of a more brilliant decoration.

There is a perceptible difference in the aspect of these two types of glazes after firing. The hard, non-calcareous glaze has a slightly milky look. The softer calcareous type is more brilliant, and approaches in transparence and limpidity to the lead glazes of soft porcelain. A glaze of this last kind was used at Sèvres for a few years after the first introduction of the hard paste, and perhaps also at Dresden in quite early days.

The principal objection to a hard refractory glaze, such as that so long in use at Sèvres, arises from the difficulty of properly incorporating the enamel colours with its body. The restriction of the number of pigments that can be employed, both under and on the surface of the glaze, in consequence of the high temperature at which the latter melts, is another drawback. The dulness, the ‘painted on’ look of so much of the decoration on European hard paste porcelain, is in great measure a consequence of the employment of a glaze that is only softened at a high temperature. As an example of a medium type of glaze we give the composition of that used at Berlin in 1836. This consisted of kaolin, 31 per cent.; quartz, 43 per cent.; gypsum, 14 per cent.; and ground porcelain, 12 per cent. A glaze long in use at Dresden is of a very similar character. Felspar, it will be seen, does not enter into its composition, and such a glaze can contain but little potash or soda. With this we may contrast the hard glaze of Sèvres, composed simply of ground pegmatite, a rock consisting mainly of felspar. This glaze yields on analysis 74 per cent. of silica, 17 per cent. of alumina, and as much as 8 per cent. of potash.

The glaze on Chinese porcelain is prepared by mixing certain special varieties of petuntse with an impure lime, prepared by burning limestone with dry fern as fuel. It contains, as we have seen, from 15 to 21 per cent. of lime, 5 to 6 per cent. of alkalis, 11 per cent. of alumina, and 66 per cent. of silica.

We give these examples to illustrate the principal types of glazes used for hard paste porcelain. It will be noticed that the constituents are drawn from widely different sources.

The glazes of soft paste porcelain always contain a large amount both of lead and of potash or soda, so that they approximate in composition to a flint glass. The alkalis, generally introduced as carbonates, necessitate a previous fritting of part at least of the materials. Boracic acid plays an important part in the glaze of most modern English wares: it is generally introduced in the form of borate of soda or borax. This acid replaces in part the silica, just as in the paste the glassy materials are replaced by bone-earth.

CHAPTER   IV

DECORATION BY MEANS OF COLOUR

IF we were treating the subject purely from a practical point of view, with the glazing and firing of a piece of porcelain the manufacture might be held to be terminated. This would be strictly true, for instance, of the white porcelain of Berlin, so largely used in the chemical laboratory; a great deal, too, of the china in domestic use receives no decoration of any kind. But for us there remains still to examine the element of colour and the way in which it is applied to the decoration of porcelain.

This is effected in three different ways: by the employment of coloured glazes; by painting on the surface of the paste before the glaze is applied (this is the decoration sous couverte); and finally by coloured enamels applied to the surface of the glaze. These methods may be combined, but as this is rarely the case, such a division forms the basis of a convenient classification, more especially for the wares of China and Japan.

In the case of both the paste and of the glaze, we have been dealing with a restricted group of elements, with alumina, lime, potash and soda; and apart from impurities unintentionally introduced, all the combinations of these bodies are colourless. We have now to consider the effect of introducing certain of the heavy metallic bases which combine with the excess of silica to form coloured silicates.

The metals that give to Oriental porcelain its brilliant hues are few in number. Indeed, in all lands and at all times, iron, copper, cobalt, and manganese have been the principal sources of colour in the decoration not only of porcelain, but of most other kinds of pottery. As equal to these four metals in importance, but not strictly to be classed as colouring materials, we may place tin, the source of most opaque whites, and lead, which is the main fluxing element for our enamels. Next in importance to these metals come antimony, long known to the Chinese as a source of yellow, and finally, but this last only since the beginning of the eighteenth century, gold, as the source of a red pigment.19 This exhausts the list, not only for the Far East, but for all the pottery of Europe up to the end of the eighteenth century.

It was in a period of artistic decline that the advance of chemical knowledge led to the introduction of other colours, derived both from new metallic bases and from fresh combinations of those already known. By far the most important of these new colours are those derived from the salts of chromium, but uranium and other rare metals have also been called into use. As with the sister art of painting, the beauty and harmony of the effects produced have not kept pace with the enlargement of the palette—the result was rather to accentuate the decline that had already set in from other causes.

There are two metals, iron and copper, that have always been of pre-eminent importance as sources of colour. Each of them forms two series of combinations differing entirely in hue, so that were we confined to the use of these two metals, our palette would still be a fairly complete one.

The protoxide of copper, especially when a certain amount of lime and of soda is present, forms a series of beautiful blue and green silicates. When the proportion of oxygen is decreased, as happens when the surface of the ware is exposed in the kiln to a reducing flame, a suboxide of copper is formed, which gives a deep and more or less opaque red hue to the glaze. So in the case of iron, the so-called sesqui-oxide is perhaps the most abundant source of colouring matter in the mineral kingdom: the colours produced by it range from pale yellow to orange, brown, and full red. When, however, the iron is present as a protoxide, the colour given to the glaze is entirely altered; it ranges from a pale sea-green to a deep olive.

The remaining two elements that have long played an important part in the decoration of pottery are cobalt and manganese. These metals, in the form of silicates, yield the well-known series of blues and purples. One important source of the famous underglaze blue of China and Japan is a black mineral known to us as wad, which occurs in earthy to stony concretions. This wad contains oxides of both cobalt and manganese, and the quality of the blue obtained from it depends in great measure upon the proportion in which the two metals occur.

The employment of antimony is comparatively rare, but, generally in combination with iron, it is an important source of yellow. In spite of the volatile nature of most of its salts, in the presence of silica this metal is able to withstand a high temperature.

But before considering the application of colour to the glaze, we must mention briefly a method of decoration which was in great favour at Sèvres some years ago—I mean the application of colour to the paste itself. This was done long ago by Wedgwood, sometimes to the whole mass of the paste, as was the case with his jasper ware, which some authorities class as a true porcelain. At Sèvres these coloured pastes have been generally applied to the surface only, in thin layers, or even as mere coats of paint. When laid on in successive coats, as in the so-called pâte-sur-pâte, the amount of colouring matter need not be large, from 2 to 5 per cent. When larger proportions of coloured oxides are mixed with the pâte, and this is painted on with a brush, the process differs little from the ordinary decoration under the glaze, into which it indeed may be said to pass. Coloured pastes of this description have never been employed by the Chinese, and it is not possible to obtain much brilliancy or decorative effect by their use. They are, indeed, foreign to the nature of porcelain, sacrificing the brilliant white ground which should be the basis of all decorative schemes.

When the colouring matter is subjacent to the glaze it must be of a nature to withstand the full heat of the subsequent firing; we are restricted therefore to colours ‘à grand feu.’ This practically confines us to cobalt and to certain combinations of iron and copper, as far as the ‘old palette’ is concerned. At Sèvres and elsewhere other metals have been made use of whose silicates withstand the extreme temperature of the kiln. By the use of chromium we have command of many shades of green. If to an oxide of tin we add a minute quantity of the sesqui-oxide of chromium, we can obtain, in the presence of lime, many shades from rose to purple; and a mixture of cobalt and chromium produces a fine black. There is, however, as yet no satisfactory yellow pigment known that will withstand the grand feu. At the best we can get a straw colour from certain ores of tungsten and titanium, and from uranium a yellow deeper in tint but uncertain in application.

The majority of the colours we have mentioned require a more or less oxidising flame for their full development. There are, however, two most important groups of coloured glazes, long the monopoly of the Chinese, but now successfully imitated in France and elsewhere, which require, for a term at least, to be subjected to a reducing flame.

The first of these glazes is the well-known Celadon, using that term in its proper and restricted sense, for certain shades of greyish green. The celadon of the Chinese is produced by the presence of a small quantity, about two per cent., of protoxide of iron in the glaze. An oxidising flame would change this protoxide to the yellow sesqui-oxide. We may note that a celadon of good tint can only be produced when a considerable quantity of lime is present in the glaze.

The other group, depending also upon a reducing flame, is constituted by the famous Sang de bœuf and Flambé glazes.

The colour of the first is given by the red sub-oxide of copper, chiefly suspended in the glaze. In the case of the flambé or ‘transmutation’ glazes, the strange caprices of colour have their origin, in part at least, in the contrast of the red sub-oxide and the green silicate of copper. In the case of both these glazes everything depends on the regulation of the draught of the furnace in which they are fired. The French have lately been at great pains to master the difficulties attendant upon the development of the effects sought after, and some success has been attained not only on a porcelain ground as at Sèvres, but these glazes have also been applied to fayence at the Golfe St. Juan and elsewhere. It has been proved by some experiments made at Sèvres, that in the firing, the critical period, during which so much depends upon the regulation of the draught, is just before the melting of the glaze. Once melted the glaze not only forms an impervious cover which prevents the smoky flame from discolouring the paste below, but the glaze itself is no longer sensitive to the action of the gases which surround it. It is therefore only during a short period preceding the moment when the glaze begins to melt, that it is necessary to promote a smoky and reducing flame. This is a point of considerable practical importance.20

The application of the Decoration under the Glaze is essentially a Chinese method. To it we owe the important family of ‘blue and white’ ware. The superiority of the Chinese in the management of the blue colour has been attributed to various causes. The result is no doubt influenced not only by the constitution of both paste and glaze, but also by the fact that the colour is painted upon the raw paste.

An important factor also is the care exercised by the Chinese in the selection and preparation of the blue pigment, by which not only the desired intensity but the richness of hue is secured. The quality of the blue depends in great measure upon the presence of a small quantity of manganese in the cobalt ore employed.

The only other colour that the Chinese have succeeded in using under the glaze is the red derived from the sub-oxide of copper. The full development of this colour has for long been a lost art, but a less brilliant red from this source, often little better than a buff colour, is sometimes found in later examples combined with the blue.

In the application of colours under the glaze there is one difficulty that the Chinese have surmounted even in their commonest ware, and this is the tendency of the cobalt blue to dissolve and ‘run’ in the glaze, giving to the design a blurred and indistinct appearance. It would seem that the sharpness of outline depends upon the consistency of the glaze at the moment when it first melts. At that point the glaze should be viscous and not inclined to flow, and this is what occurs in the case of the highly calcareous glazes of the Chinese.

Before passing to the enamel colours, we must say something of a class of glazes which may be looked upon as to some extent of an intermediate character. These are the glazes associated with the ‘San tsai,’ the ‘three colours’ first used in combination by the Chinese.

These coloured glazes were applied, not, as is usually the case in China, to the raw paste, but they were, it would seem, painted on the surface after a preliminary firing. Being applied with a brush, the whole surface of the biscuit was not necessarily covered, and glazes of all these colours could be used upon the same piece of porcelain. Glazes of this class were rendered more fusible by the addition of a certain quantity of lead, and on this ground, and still more in their historical relation, as we shall see later on, these ‘painted glazes’ may be considered as a link connecting the old refractory glazes of the monochrome and ‘blue and white’ wares on the one hand, with the fusible enamels which were at a later time superimposed upon the glaze on the other.

The three colours which are applied in this way by the Chinese are: (1) A turquoise blue derived from copper with the addition of some soda or potash. (2) The manganese purple, often described as aubergine. (3) A yellow prepared from an iron ore containing some amount of antimony. None of these colours would stand the full heat of the furnace, and for a reason which will be explained further on, they are known as the colours of the demi grand feu.21

Coloured Enamels. We have now to describe

PLATE II. CHINESE MING PORCELAIN, BLACK GROUND

the decoration that is applied to the surface of the glaze. In these coloured enamels the colouring matter is dissolved in a flux which contains a large quantity of lead. The comparatively gentle heat at which such enamels fuse allows of the use of a much larger palette than is available for the decoration under the glaze.

It is well to point out at the outset the marked distinction in composition and in appearance between the brilliant enamels of the Chinese and the dull tints of the ‘porcelain colours’ found in the hard pastes of Meissen and Sèvres. To make clear the cause of this difference it will be necessary to enter into some little detail.

The colouring matter in the European enamels may amount to as much as a third part of the total amount of the flux with which they are incorporated. As there is not enough of this flux to dissolve the whole of the oxides, the enamel remains dull and opaque after firing. The flux, in fact, is only used as a vehicle to attach the colour to the surface of the porcelain. The effect in consequence is inferior in brilliancy to that obtained by the Chinese with their transparent enamels in which the metallic oxides, present in much smaller quantity, are thoroughly dissolved to form a glass. There is, unfortunately, a practical obstacle to the application of these glassy enamels to the hard pastes and glazes of Europe. It is impossible to ensure their firm adhesion to the subjacent glaze. The Chinese, however, do not appear to find any difficulty in effecting this. The following explanation has been given to account for the difference of behaviour:—the tendency of the enamel to split off in cooling, as has been proved by experiment, arises from the small amount of contraction at that time of the highly kaolinic paste, compared with that of the superimposed glassy enamel. The more silicious paste used by the Chinese contracts, on the contrary, at the same rate approximately as the enamels that it carries, and these enamels may therefore be laid on in sufficient thickness without any risk of their subsequently splitting off.22 To appreciate the difference in the decorative value of these two classes of enamels it is only necessary to compare the brilliant effect, say, of a piece of Chinese egg-shell of the time of Kien-lung with the tame surface of a contemporary Meissen plate, elaborately painted with landscapes or flowers.

The glassy enamels used by the Chinese resemble the pastes used for artificial jewellery. They are essentially silicates of lead and an alkali. The composition of the flux has to be modified to ensure the full development of the colour of the different metallic oxides which are either made up with it or added subsequently. But in a general way we may say that the colourless fluxes which form the basis of the coloured enamels are prepared by melting in a crucible a mixture of pure quartz sand and red lead, and adding more or less alkali. In certain cases the lead predominates, as when it is proposed to make an emerald green enamel by means of copper, or when the flux is to serve as a basis for the ruby colour given by a minute quantity of gold. On the other hand, if copper be added to a flux containing an excess of either soda or potash, we obtain a turquoise blue. A fine purple, again, can only be obtained from manganese with an alkaline flux; if too much lead is present only a brown tint is obtainable.