familiar with the fine blue and brown glazes showing starry and radiating crystalline groups that are being produced at Copenhagen, and Sèvres, Röstrand and Berlin. Some three or four years ago, in a paper I had the honour to read before this Section, I drew attention to these crystalline glazes, as well as to others of a very different type which had been produced at the Rookwood Pottery in America, and by my brother, Mr. Joseph Burton, at our works at Clifton Junction. Since then, however, the whole subject of crystalline glazes and their production with scientific precision, so that they might be used as a certain and regular means of pottery decoration rather than as the occasional triumphs of the kiln, has been engaging a great deal of our attention, and if in certain directions the results are not yet all that might be wished, such a measure of success has attended our experiments as to warrant me in laying the results before you.

It is essential, first of all, that one should try to form some conception of what a pottery glaze is, and as to how it behaves when it is melted on a piece of pottery at the requisite temperature of the kiln, and under the firing conditions which control production on a large scale.

The popular idea of a glaze, as of a sheet of window glass, is that of a uniform transparent substance, which may be either coloured or colourless, and which is melted on the surface of the pottery so as to form an impervious, protective layer. Such a conception is, however, very wide of the truth. In the first place, a glaze like any other form of glass is not, either physically or chemically, a uniform substance. Though it may be possible to produce glaze to which one can give an approximate chemical formula, no chemist would suggest for a moment that such a formula represented anything more than the ultimate percentage composition of a mixed mass which might be compounded in many different ways. A glaze, like any other form of glass, is really a complex of various silicates, very loosely combined, even if they are combined with each other.

Undoubtedly, the correct view of the nature of glaze is that which regards it as an alloy in which the various silicates play the same part as the separate metals in a metallic alloy. With the same ultimate chemical composition, the particular silicates that will be formed in any given glaze will differ within very wide limits according to the temperature and

duration of firing, and the relation between the glaze and the kind of pottery on which it is applied. With a glaze and body of the same ultimate chemical composition produced under firing conditions which are fairly constant, it is possible to obtain constant results, but if any of these conditions vary from time to time, as they will in manufacturing processes conducted on a large scale, then, of course, the results will vary too. It has been found by the accumulated experience of generation after generation of potters, that certain types of glaze have fairly wide limits of stability within which they will produce satisfactory results; while there are other types of glaze which have such narrow limits of stability that they are of very little use for commercial work, their limits being narrower than those commonly met with in every-day practice.

Thus it is that after the experience of centuries, we find certain types of glaze associated with certain types of pottery throughout the civilised world. The universal use of a felspathic glaze on hard porcelain; of tin enamel on faïence or “Delft;" of saltglaze on stoneware; and of lead glazes on earthenwares made after the English fashion, are the common illustrations that will occur to everyone. It would, of course, be possible to use glazes of other types on each of these varieties of pottery, but in every case the glaze which has become as it were, common or traditional, is the one which experience has proved to fulfil most completely all the conditions of that particular branch of manufacture. Another feature of pottery glazes which marks them out from glass and from the enamels used on metals, is the fact that the glaze is melted or produced on a bed of silica and mixed silicates analogous in composition to the glaze itself. When a transparent enamel is melted at a low temperature on a sheet of metal, there is practically no chemical action between the molten enamel and the bed on which it lies. On the other hand, when a pottery glaze of any type is melted on a piece of ware, there is a decided and in some cases a strongly-marked chemical action between glaze and ware. One of the well-marked properties of complex silicates is the readiness with which, at high temperatures, they will dissolve into each other, and in nothing is this more clearly shown than in the way which a melted glaze attacks the surface of the pottery on which it is fired.

One of the simplest experiments in a chemical laboratory is to dissolve clay by

heating it up with a sufficient quantity of lead oxide. At a fairly low temperature the lead oxide melts and dissolves the clay just as perfectly, though not so rapidly, as hot water would dissolve lump sugar. Many other metallic oxides would behave precisely like lead oxide, in fact it is only a question of getting any particular oxide in a molten or vaporous condition to enable this action to take place. In pottery manufacture this solvent action of melted oxides upon clay has been used as a means of producing glazed pottery for many centuries. The common mediæval green and yellow glazed pottery made over the whole of Western Europe was produced in this manner. When the vessel had been shaped in any common clay it was dusted over with powdered lead ore (generally galena, the native sulphide of lead), and the clay vessel thus coated was placed in the potter's kiln and fired. The first result of the firing would be to drive the sulphur out of the lead ore, which the increasing heat then slowly roasted into oxide of lead. This oxide of lead in its turn melted and attacked the clay body, now become red hot. In this way, and at one operation, the ware was hardened from clay into pottery, and its surface was coated with a brilliant glassy compound of lead oxide, alumina and silica. The well-known saltglaze of stonewares is produced by an entirely analogous method though with very different materials, and at a much higher temperature. In this case the clay vessels are put into the kiln without any glazing substance upon them. They are then fired to a white heat, at which point the silicates present in the clay itself commence to fuse. When this temperature has been reached, wet common salt is thrown into the kiln and is rapidly decomposed with the formation of vapours of oxide of sodium and hydrochloric acid. The oxide of sodium vapour coming in contact with the white hot stoneware melts some of the clay substance, just as the lead oxide would do, and forms a glass of soda, alumina, and silica on the surface of the ware, which is the well-known salt-glaze.

It might be imagined that while vaporised or melted, metallic oxides are thus capable of attacking and dissolving the surface of a piece of pottery and forming an actual glass with it, natural or artificial silicates, such as the felspar used in glazing hard-paste porcelain, or the fritted glazes used on English earthenware and porcelain would not necessarily have the same effect. Experience proves, however, that

they have this effect. Even molten felspar, which is already a complex silicate of potash, soda and alumina, or of soda, lime and alumina, when it is melted at the high temperature of the porcelain furnace, actually dissolves some of the clay substance. A microscopic examination of a thin slice of hard-paste porcelain, prepared exactly as one would prepare a rock section, shows three clearly marked layers — (a) an outer skin of clear glaze; (b) inside that an intermediate "felted" layer; (c) the body of the ware itself. The intermediate felted layer is clearly a mixture of glaze and body where the molten glaze has attacked the body, and dissolved some of its constituents. The same effect is shown, though generally to a less marked degree, with the ordinary fritted glazes used on English earthenware. When these glazes are fired at a high temperature on earthenware, and the fire is unduly prolonged so as to give a longer time for the action to take place, it is found that their solvent action becomes quite pronounced. I have here a tile of ordinary English earthenware which has been submitted to a prolonged firing, and the glaze has so attacked the body of the tile that it has eaten into it in places, leaving certain portions of the edge exposed almost like bits of hard rock on the edge of a precipice from which all the covering soil has been washed away.

Incidentally, it may be remarked that even those scientific men who have studied the chemistry of pottery most carefully, appear to have overlooked the influence of this solvent action of the glaze upon the body, and the consequent formation of an intermediate layer, with the cracking of glazes which is known technically as "Crazing." A moment's reflection will convince anyone who is familiar with the various kinds of pottery that it is precisely the hard-fired varieties, in which the formation of this intermediate layer is strongly marked, that are least liable to this defect. In the case where a glaze has been melted on pottery at so low a temperature that it is practically unable to dissolve any of the body substance, crazing takes place most readily.

It follows from what has been said that one should always look upon a piece of glazed pottery as exhibiting a gradual progression from the outer skin of what is technically called the glaze to the true body or clay substance, with an intermediate layer within which body and glaze are in various states of

transfusion. As to the exact relation which the intermediate layer bears to the true glaze itself, the three controlling factors are-(a) the chemical nature of the glaze; (b) the chemical composition of the body, and (c) the degree and duration of the firing to which they have been subjected.

Having clearly established why a pottery glaze behaves, when it is melted, like any other fluid, there is no difficulty in imagining a molten glaze as capable of dissolving certain other substances which may be presented to it. Thus if the clay substance contains a considerable quantity of free oxide of iron, as all the common red clays do, a molten glaze of suitable composition would dissolve that oxide of iron; and if the iron oxide were present in sufficient quantity, the glaze might even become saturated with it. Everyone who has studied the formation of crystalline bodies knows that one of the readiest methods of obtaining such substances in a perfect condition is by dissolving them in fluids and allow ing them to crystallise out as the menstruum becomes more concentrated. The same results can be obtained to a less perfect degree when the solvent is saturated at a high temperature and allowed to cool slowly. Under such circumstances, when the solution is cooled, it is no longer able to retain all the dissolved substance, and if this substance be a crystallisable one, it will separate out in the crystalline form. These conditions can be obtained perfectly in the process of firing pottery glazes, and it so happens that the earliest piece of crystalline glaze which has ever been described was formed in this simple way. There is, in the British Museum collection, a Staffordshire tyg of the 17th century, the body of which is of common red clay, glazed by firing on it powdered galena in the manner already described. When the lead oxide formed from the galena was melted in contact with the clay at a bright red heat, it dissolved some of the silica and alumina of the clay so as to form a lead glaze. This lead glaze must also have dissolved in its turn some of the oxide of iron contained in the clay, and on cooling, this oxide of iron, being in excess of what the glaze could retain in solution, has crystallised out, and the glaze is filled in its under layer with brilliant sparkling crystals, giving it an appearance very mach like that of the beautiful mineral known as Aventurine." In recent years this particular fo:.n of crystalline glaze has been largely developed at the Rookwood pottery in America, and some

66

specimens which they have produced are quite remarkable, and at the same time singularly beautiful, from the silky golden sheen of the crystals seen under a layer of clear yellow glaze. V. hen glazes of this type are produced on suitable shapes, the natural flow of the glaze down the side of the vase tends to arrange the crystals in lines, with their long axes in the direction of the line of flow, with the result that in specimens that have been slowly cooled, so that the changes should have time to take place perfectly, a beautiful striated effect is produced, very similar to that given by the asbestiform threads in the mineral Crocidolite" or tiger-eye stone of South Africa. By the kindness of the authorities of the Victoria and Albert Museum I am able to exhibit here to-night, some of the most perfect specimens that I have ever seen from the Rookwood pottery, while I have also a group of vases showing crystalline glazes of the same type from my own factory. I have spoken as if the crystals obtained in this way might be simply pure oxide of iron which had crystallised out from the glaze as it cooled, but in all probability the crystals are not simply oxide of iron, but small crystalline plates apparently belonging, by their appearance and optical properties, to that large group of minerals known as Micas." This being the case, we must look upon the crystals, whatever may be their ultimate chemical composition, as one of the complex silicates formed in a glaze which separates out from the other silicates with which it is associated, in such a way as to be visible to the naked eye. We must therefore regard these crystalline glazes not as being entirely different from the clear and apparently homogeneous glazes which the potter generally obtains, but as glazes constituted like all other glazes, save that some of the mixed silicates have separated out from the rest in a crystalline form. Another point which may just be mentioned in passing, is that although these crystalline glazes were obtained in the first place accidentally by means of the oxide of iron which the glaze had dissolved from the body, they can just as well be obtained by adding the oxide of iron to the glaze mixture itself before firing. Other oxides, particularly those of chromium and uranium, readily yield glazes of a similar character. It is now nearly ten years since my brother first turned this knowledge to account in the production of our Sunstone glazes, many of which are exhibited here. These can be obtained in a variety of colours;