In the preceding chapter we have dealt with the processes of manufacture employed in the production of both the crudest and the most perfect forms of flat glass as used for such purposes as the glazing of window openings. The products now to be dealt with are of an intermediate character, sheet-glass possessing many of the properties of polished plate, but lacking some very important ones; thus sheet-glass is sufficiently transparent to allow an observer to see through it with little or no disturbance—in the best varieties of sheet-glass the optical distortion caused by its irregularities is so small that the glass appears nearly as perfect as polished plate—but in the cheap glass that is used for the glazing of ordinary windows, sheets are often employed which produce the most disturbing, and sometimes the most ludicrous, distortions of objects seen through them. It is a curious fact that even in good houses the use of such inferior glass is tolerated without comment, the general public being, apparently, remarkably nonobservant in this respect. In another direction sheet-glass has the great advantage over plate-glass that it is very much lighter, or can at least be produced of much smaller weight and thickness, although this advantage entails the consequent disadvantage that sheet-glass is usually much weaker than plate, and can only be used in much smaller sizes. In recent times the production of relatively thin plate-glass has, however, made such strides that it is now possible to obtain polished plate-glass thin enough and light enough for almost every architectural purpose. Finally, the most important advantage of sheet-glass, and the one which alone secures its use in a great number of cases in preference to plate-glass, is its cheapness, the price of ordinary sheet-glass being about one-fourth that of plate-glass of the same size.

The raw materials for the manufacture of sheet-glass are sand, limestone, salt-cake, and a few accessory substances, such as arsenic, oxide of manganese, anthracite coal or coke, which differ considerably according to the practice of each particular works. In a general way these materials have already been dealt with in Chapter III., and we need only add here that the sheet-glass manufacturer must keep in view two decidedly conflicting considerations. On the one hand the requirements made in the case of sheet-glass as regards colour and purity render a rigorous choice of raw material and the exclusion of anything at all doubtful very desirable; but on the other hand the chief commercial consideration in connection with this product is its cheapness, and in order to maintain a low selling price at a profit to himself the manufacturer must rigorously exclude all expensive raw materials. For this reason sheet-glass, works such as those of Belgium and some parts of Germany, which have large deposits of pure sand close at hand, possess a very considerable advantage over those in less favoured situations, since sand in particular forms so large a proportion of the glass, and the cost of carriage frequently exceeds, and in many cases very greatly exceeds, the actual price of the sand itself. The same considerations will apply, although in somewhat lesser degree, to the other bulky materials, such as limestone and salt-cake; but both these are more generally obtainable at moderate prices than are glass-making sands of adequate quality for sheet manufacture.

Ordinary “white” sheet-glass is now almost universally produced in tank furnaces, and a very great variety of these furnaces are used or advocated for the purpose. It would be beyond the scope of the present book to enter in detail into the construction of these various types of furnace or to discuss their relative merits at length. Only a brief outline of the chief characteristics of the most important forms of sheet-tank furnaces will therefore be given here.

Sheet tanks differ from each other in several important respects; these relate to the sub-division of the tank into one, two, or even three more or less separate chambers, to the depth of the bath of molten glass and the height of the “crown” or vault of the furnace chamber, to the shape and position of the apertures by which the gas and air are admitted into the furnace, and the resultant shape and disposition of the flame, and finally to the position and arrangement of the regenerative appliances by which some of the heat of the waste gases is returned into the furnace.

Taking these principal points in order, we find that in some sheet tank furnaces the whole furnace constitutes a single large chamber. In this type of furnace the whole process of fusion and fining of the glass goes on in this single chamber, and an endeavour is made to graduate the temperature of the furnace in a suitable manner from the hot end where the raw materials have to be melted down to the colder end where the glass must be sufficiently viscous to be gathered on the pipes. It is obvious that this control of the temperature cannot be so perfect in a furnace of the single chamber type as in one that is sub-divided. Such sub-divided furnaces are, as a matter of fact, much more frequent in sheet-glass practice; but this practice differs widely as to the manner and degree of the sub-division introduced. In the extreme form the glass practically passes through three independent furnaces merely connected with one another by suitable openings of relatively small area through which the glass flows from one to the other. If it were possible to build furnaces of materials that could resist the action of heat and of molten glass to an indefinite extent, it is probable that this extreme type would prove the best, since it gives the operator of the furnace the means of controlling the flow of glass in such a way that no unmelted material can leave the melting chamber and enter the fining chamber, and that no insufficiently fined glass can leave the fining chamber and find its way into the working chamber. But in practice the fact that this extreme sub-division introduces a great deal of extra furnace wall, exposed both to heat and to contact with the glass, involves very serious compensating disadvantages—the cost of construction, maintenance and renewal of the furnace is greatly increased, while there is also an increased source of contamination of the glass from the erosion of the furnace walls. It is, therefore, in accordance with expectations to find that the most successful furnaces for the production of sheet-glass are intermediate in this respect between the simple open furnace and the completely sub-divided one. In some cases the working chamber is separated from the melting and fining chamber by a transverse wall above the level of the glass, while fire-clay blocks floating in the glass just below this cross wall serve to complete the separation and to retain any surface impurities that may float down the furnace.

As regards the depth of glass in the tank, practice also varies very much. The advantages claimed for a deep bath are that the fire-clay bottom of the furnace is thereby kept colder and is consequently less attacked, so that this portion of the furnace will last for many years. On the other hand the existence of a great mass of glass at a moderate heat may easily prove the source of contamination arising from crystallisation or “devitrification” occurring there and spreading into the hotter glass above. Also, if for any reason it should become necessary to remove part or all of the contents of the tank, the greater mass of glass in those with deep baths becomes a formidable obstacle. On the whole, however, modern practice appears to favour the use of deeper baths, depths of 2 ft. 6 in. or even 3 ft. being very usual, while depths up to 4 ft. have been used.

The question of the proper height of the “crown” or vault of the furnace is of considerable importance to the proper working of the tank. For the purpose of producing the most perfect combustion, it is now contended that a large free flame-space is required. The earlier glass-melting tanks, like the earlier steel furnaces, were built with very low crowns, forcing the flame into contact with the surface of the molten glass, the object being to promote direct heating by immediate contact of flame and glass; the modern tendency, however, is strongly in the direction of higher crowns, leaving the heating of the glass to be accomplished by radiation rather than direct conduction of heat. There can be little doubt that up to a certain point the enlargement of the flame-space tends towards greater cleanliness of working and a certain economy of fuel, but if the height of a furnace crown be excessive there is a decided loss of economy. Flame-spaces as high as 6 ft. from the level of the glass to the highest part of the crown have been used, but the more usual heights range from 2 ft. to 5 ft.

The “ports” or apertures by which pre-heated gas and air enter the furnace chamber differ very widely in various furnaces. In some cases the gas and air are allowed to meet in a small combustion chamber just before entering the furnace itself, while in other cases the gas and air enter the furnace by entirely separate openings, only meeting in the furnace chamber. The latter arrangement tends to the formation of a highly reducing flame, which is advantageous for the reduction of salt-cake, but is by no means economical as regards fuel consumption. On the other hand, by producing a perfect mixing of the entering gas and air in suitable proportions, the other type of ports can be made to give almost any kind of flame desired, although their tendency is to form a more oxidising atmosphere within the furnace. The latter type of ports, although widely varied in detail, are now almost universally adopted in sheet tank furnaces.

All modern tank furnaces work on the principle of the recovery of heat from the heated products of combustion as they leave the furnace, and the return of this heat to the furnace by utilising it to pre-heat the incoming gas and air; but the means employed to effect the application of this “regenerative” principle differ considerably in various types of plant. Perhaps the most widely-used form of furnace is the direct descendant of the original Siemens regenerative furnace, in which four regenerator chambers are provided with means for reversing the flow of gas and air in such a way that each pair of chambers serves alternately to absorb the heat of the outgoing gases and subsequently to return this heat to the incoming air that passes through one, and the incoming gas that passes through the other of these chambers. In these furnaces, the regenerator chambers themselves are generally placed underneath the melting furnace, and they are built of fire-brick and filled with loosely-stacked fire-bricks, whose function it is to absorb or deliver the heat. In the most modern type of furnaces of this class, the gas-regenerators are omitted entirely, the air only being pre-heated by means of regenerators, while the gas enters the furnace direct from the producer, thus carrying with it the heat generated in the producer during the gasification of the fuel. While this arrangement is undoubtedly economical, it has the serious disadvantage, especially in the manufacture of sheet-glass, that the gas, rushing direct from the producer into the furnace, carries with it a great deal of dust and ash, which it has no opportunity of depositing, as in the older types of furnace, in long flues.

The most serious disadvantages of the ordinary types of regenerative furnaces are due to the considerable dimensions of the regenerative apparatus, necessitating a costly form of construction and occupying a large space, while the necessity of periodically reversing the valves so as to secure the alternation in the flow of outgoing and incoming gases requires special attention on the part of the men engaged in operating the furnace, as well as the construction and maintenance of valves under conditions of heat and dirt that are not favourable to the life of mechanical appliances. It is claimed that all these disadvantages are overcome to a considerable extent in one or other of the various forms of furnace known as “recuperative.” In these furnaces there is no alternation of flow, and the regenerator chambers are replaced by the “recuperators.” These consist of a large number of small flues or pipes passing through a built-up mass of fire-brick in two directions at right-angles to one another; through the pipes running in one direction the waste gases pass out to the chimney, while the incoming gas and air pass through the other set of pipes. A transference of heat between the two currents of gas takes place by the conductivity of the fire-brick, and thus the outgoing gases are continuously cooled while the ingoing gases are heated—the transference of heat being somewhat similar to that which takes place in the surface condenser of a steam engine. Theoretically this is a much simpler arrangement than that of separate regenerator chambers, and to some extent it is found preferable in practice, but there are certain disadvantages associated with the system which arise principally from the peculiar nature of the material—fire-brick—of which the recuperators must be constructed. In the first place, the heat-conductivity of fire-brick is not very high, so that, in order to secure efficiency, the recuperators must be large, and while the individual pipes must be of small diameter, their area as a whole must be large enough to allow the gases to pass through somewhat slowly. Next, owing to the tendency of fire-brick to warp, shrink and crack under the prolonged effects of high temperatures, it becomes difficult to prevent leakage of gases from one set of pipes into the other. If this occurs to a moderate extent its only effect will be to allow some of the combustible gas to pass direct to the chimney, and at the same time a dilution of the gases entering the furnace by an addition of products of combustion from the waste-gas flues. This, of course, will materially reduce the efficiency of the furnace and require a higher fuel consumption if the temperature of the furnace is to be maintained at its proper level. If, however, the leakage should become more serious, a disastrous explosion might easily result, particularly if the nature of the leakage were such as to allow the incoming gas and air to mix in the flues. It follows from these considerations that, although the recuperative furnace is somewhat simpler and cheaper to construct, it requires, if anything, more careful maintenance than the older forms of regenerative furnace.

Tank furnaces for the production of sheet-glass in this country are generally worked from early on Monday morning until late on Saturday night, glass-blowing operations being suspended during Sunday, although the heat of the furnace must be maintained. On the Continent, and especially in Belgium, the work in connection with these furnaces goes on without any intermission on Sunday—a difference which, however desirable the English practice may be, has the effect of handicapping the output of a British furnace of equal capacity by about 10 per cent. without materially lessening the working cost.

The process of blowing sheet-glass in an English glassworks is generally carried out by groups of three workmen, viz., a “pipe-warmer,” a “gatherer” and a “blower,” although the precise division of the work varies according to circumstances. The pipe-warmer’s work consists in the first place in fetching the blowing-pipe from a small subsidiary furnace in which he has previously placed it for the purpose of warming up the thick “nose” end upon which the glass is subsequently gathered. The sheet-blower’s pipe itself is an iron tube about 4 ft. 6 in. long, provided at the one end with a wooden sleeve or handle, and a mouthpiece, while the other end is thickened up into a substantial cone, having a round end. Before introducing the pipe into the opening of the tank furnace, the pipe-warmer must see that the hot end of the pipe is free from scale or dirt and must test, by blowing through it, whether the pipe is free from internal obstructions. He then places the butt of the pipe in the opening of the furnace and allows it to acquire as nearly as possible the temperature of the molten glass. When this is the case the pipe is either handed on to the gatherer, or the pipe-warmer, who is usually only a youth, may take the process one step further before handing it on to the more highly skilled workman. This next step consists in taking up the first gathering of glass on the pipe. For this purpose the hot nose of the pipe is dipped into the molten glass, turned slowly round once or twice and then removed, the thread of viscous glass that comes up with the pipe being cut off against the fire-clay ring that floats in the glass in front of the working opening. A small quantity of glass is thus left adhering to the nose of the pipe, and this is now allowed to cool down until it is fairly stiff, the whole pipe being meanwhile rotated so as to keep this first gathering nicely rounded, while a slight application of air-pressure, by blowing down the pipe, forms a very small hollow space in the mass of glass and secures the freedom of the opening of the pipe. When the glass forming the first gathering has cooled sufficiently, the gatherer proceeds to take up the second gathering upon it. The pipe is again introduced into the furnace and gradually dipped into the molten glass, but this must be done with great care so as to avoid the inclusion of air-bells between the glass already on the pipe and the new layer of hotter glass that is now taken up. This freedom from air-bells is secured by a skilful gatherer by a gradual rotation of the pipe as it is lowered into the glass, thus allowing the two layers of glass to come into contact with a sort of rolling motion that allows the air time to escape. When completely immersed, the pipe is rotated a few times and is then withdrawn and the “thread” again cut off. The mass of glass on the end of the pipe is now considerably larger than before and requires more careful manipulation to cause it to retain the proper, nearly spherical, shape. During the cooling process which now follows the pipe is laid across an iron trough, kept brimful of water; this serves to cool the pipe itself, and also allows the pipe to be readily rotated backwards and forwards by rolling it a little way along the trough. When the whole mass of glass has again cooled sufficiently to be manipulated without risk of rapid deformation, a third gathering of glass is taken up, in precisely the same manner as that already described for the second gathering, and if the quantity of glass required is large, or the glass itself is so hot and fluid that only a comparatively small weight adheres at each time of gathering, the process may be repeated a fourth or even a fifth time, but as the weight of pipe and adhering glass increases with each gathering, each step becomes more laborious, while the hot glass, being now held on a much larger sphere, tends to flow off more readily, so that greater skill is required to avoid “losing” the gathering.

The care and skill with which these operations of gathering are carried out determine, to a large extent, the quality of the resulting sheet of glass; any want of regularity in the shape of the gathering leads inevitably to variations of thickness in different parts of the sheet, while careless gathering will introduce bubbles or “blisters” and other markings. During the intermediate cooling stages the glass must be protected from dust and dirt of all kinds, since minute specks falling upon the hot glass give rise to an evolution of minute gas bubbles which become painfully evident in the sorting room.

When the last gathering has been taken up and the mass cooled so far as to allow of its being carried about without fear of loss, the glass forms an approximately spherical mass, with the nose-end of the pipe at or near the centre of the sphere. The next stages of the process consist in the preliminary shaping of this mass in such a way as to bring the bulk of the glass beyond the end of the pipe, and then in forming just beyond the end of the pipe a widened shoulder of thinner and therefore colder glass, of the diameter required for the cylinder into which the glass is to be blown. This is done by bringing the glass into the successive shapes shown in Fig. 12, the forming of the glass being effected by the aid of specially shaped blocks and other shaping instruments in which the glass is turned and blown. The final shape attained at this stage is a squat cylinder containing the bulk of the glass at its lower end, and connected to the pipe by the thinner and colder neck and shoulder already mentioned.

At this point of the process the pipe with its adherent glass is handed over to the blower proper. This operator works on a special stage erected in front of small furnaces, called “blowing holes,” although in some works these are dispensed with, and the stages are erected in front of the melting furnace itself. The sheet-blower’s stage is simply a platform placed over or at the side of a suitable excavation which gives the blower the necessary space to swing the pipe and cylinder freely at arm’s length. The blowing process itself involves very little actual blowing, but depends rather upon the action of gravitation and on centrifugal effects for the formation of the large, elongated cylinder from the squat cylinder with which the blower commences. The process consists in holding the thick, lower end of the cylinder in the heating-furnace, and when sufficiently hot, withdrawing it and swinging the pipe with a pendulum movement in the blower’s pit. The cylinder thus elongates itself under its own weight, and any tendency to collapse is counteracted by the application of air-pressure by the mouth, the pipe being also, at times, rotated rapidly about its own axis. The re-heating of the lower end of the cylinder is repeated several times, until finally the glass has assumed the form of a cylinder of equal thickness all over, but closed with a rounded dome at the lower end (Fig. 13). This rounded end is now opened. In the case of fairly thin and light cylinders this is done by holding the thumb over the mouthpiece of the pipe in such a way as to make an air-tight seal, and then heating the end of the cylinder in the blowing-hole. The heat both softens the glass at the end and at the same time causes considerable expansion of the air enclosed in the cylinder, with the result that the end of the cylinder is burst open. After a little further heating, during which the glass at the end of the cylinder becomes very soft, and takes a wavy, curly shape, the blower withdraws the cylinder from the furnace, and holding it vertically downwards in his pit, spins it rapidly about its longitudinal axis. The soft glass at the lower end immediately opens out under the centrifugal action, and the blower increases the speed of rotation until the soft glass has opened out far enough to form a true continuation of the rest of the cylinder, and in this position it is allowed to solidify. With thick, heavy cylinders, the first opening of the end is done in a different way. A small quantity of hot glass is taken up by an assistant on an iron rod, and is laid upon the centre of the closed end of the cylinder. The heat of this mass of hot glass softens the glass of the cylinder, and the operator, with the aid of a special pair of shears, cuts out a small circle of this softened glass, thus opening the end of the cylinder. The final operation of straightening out the opened end is carried out in the same way as described above for lighter cylinders.

The completed cylinder, still attached to the pipe, is now carried away from the blowing-stage and laid upon a wooden rack; then the blower takes up a piece of cold iron, and placing it against the neck of glass attaching the cylinder to the pipe, produces a crack; a short jerk then serves to completely sever the pipe from the cylinder. A boy now takes the pipe to a stand where it is allowed to cool and where the adhering glass cracks off from it prior to passing it back to the pipe-warmer for fresh use. On the wooden rack the cylinder of glass is allowed to cool to a certain extent, and then the remaining portion of the neck and shoulder (see Fig. 13) are removed. This is done by a boy who passes a thread of soft, hot glass around the cylinder at the point where it is to be cut off; the thread of hot glass merely serves to produce intense local heating, for as soon as it has become stiff, the thread of glass is pushed off and a cold or moist iron is applied to the cylinder at the point where it had been heated by the thread. As a rule a crack immediately runs completely round the cylinder along the line of the thread, and the “cap” is thus removed. The glass is now in the form of a uniform cylinder open at both ends, but it must be opened out into a flat sheet before it can assume the familiar form of sheet-glass.

The first stage in the opening-out process is that of splitting. For this purpose the cylinders are carried to a special stand, upon which they are laid in a horizontal position, and here a crack or cut is made along one of the generating-lines of the cylinder. This may be done either by the application of a hot iron, followed, if necessary, by slight moistening, or by the aid of a cut from a heavy diamond drawn skilfully down the inside of the cylinder. It will be seen from the account of the process so far given, that the glass has as yet undergone no real annealing, although the blower is expected to “anneal” his cylinder during the blowing process, as far as possible, by never allowing it to cool too suddenly, and this degree of annealing is usually sufficient to save the cylinder from breaking under its internal stresses when left to cool on the racks. The surface of the glass, however, is left in a decidedly hardened condition, especially on the outside, which has necessarily been most rapidly cooled. For this reason—among others—the splitting cut is always made on the inside of the cylinder. The difference between the rates of cooling of the outside and inside of the cylinder has a further effect, which becomes evident as soon as the cylinder is split. The outside having become hard while the inside was still relatively soft, the outer layers of glass are in a state of compression and the inner layers in a state of tension in the cold cylinder. As soon as the cylinder is split, however, these stresses are to some extent relieved, the inner layers being then free to contract and the outer layers to expand; the result is an increase in the curvature of the cylinder, which slightly decreases in diameter, the cut edges overlapping. If the cylinder has been cooled rather too quickly, or if the glass itself has a high coefficient of expansion, this release of internal stresses at the moment of splitting becomes very marked, and each cylinder splits with the sound of a small explosion, while if the internal stresses are still more severe, the cylinders may even fly to pieces as soon as they are cut.

The next stage in the manufacture of a sheet of glass is the flattening and annealing process. For this purpose the split cylinders are taken to a special kiln, generally known as a “lear,” or “lehr,” where they are first of all raised to a dull red-heat; they are then lifted, one at a time, on to a smooth stone or slab placed in a chamber of the kiln where the heat is great enough to soften the glass. Here the cylinder is laid down with the split edges upwards, and by means of a wooden tool the glass is slowly spread out, being finally rubbed down into perfect contact with the slab or “lagre.” From the flattening slab, the sheet as it now is passes into the annealing kiln, which communicates with the flattening chamber. This consists, similarly to other continuous annealing kilns already described in connection with other varieties of glass, of a long tunnel, heated to the temperature of the flattening kiln at one end and nearly cold at the other. The sheets are moved down this tunnel at a uniform slow rate by the action of a system of grids which, at intervals, lift the sheets from the bottom of the kiln, move them forward by a short distance, and again deposit them on the bottom, the grids themselves returning to their former position by a retrograde movement made below the level of the kiln-bottom, and therefore not affecting the glass.

On leaving the annealing kiln the sheets of glass are sometimes covered with a white deposit arising from the products of combustion in the kiln and their interaction with the glass itself. This deposit can be removed by simple mechanical rubbing, but it is usual to dip the glass into a weak acid bath, which dissolves the white film and leaves the glass clear and bright, ready for use.

From the annealing kiln the finished sheets of glass are taken to the sorting room, where they are examined in a good light against a black background, and are sorted according to their quality for different purposes.

The defects which are found in sheet-glass are of a very varied nature, as would be anticipated from the long and complicated process of manufacture which the material undergoes in the course of its transformation from the raw materials into the finished sheet of glass. A full enumeration of all possible defects, with their technical names, need not be given here, but a description of the more important and frequent ones will be useful. The defects may be conveniently grouped according to the stage of the process from which they originate.

The first class of defects accordingly embraces those that arise from the condition of the glass as it exists in the working-end of the furnace. Chief of these are white opaque enclosures, known as “stones.” These may arise from a variety of causes within the furnace, such as an admixture of infusible impurities with the raw materials, insufficient heat or duration of melting, leading to a residue of unmelted raw material in the finished glass, or from defective condition of the interior of the furnace, leading to contamination of the glass with small particles of fire-brick. Further, if any part of the furnace has been allowed to remain at too low a temperature, or if the composition of the glass is unsuitable, crystallisation may occur in the glass, and white patches of crystalline material may find their way into the finished sheets. Another defect that may arise from the condition of the glass in the furnace is the presence of numerous small bubbles, known as “seed” in the glass. By the blowing process these are drawn out into pointed ovals, and they are rarely quite absent from sheet-glass. They arise from either incomplete fining of the glass in the furnace or from allowing the glass to come into contact with minute particles of dust during the gathering process. Another possible defect to the glass itself may be found at times in too deep a colour. This is only seen readily when a sheet of some size is examined edgewise, as most varieties of ordinary sheet-glass are too free from colour to allow this to be judged by looking through the sheet in the ordinary way. It follows from this fact that for practical purposes, where the light always traverses one thickness of the glass only, a slight difference of colour should be regarded as a very minor consideration, at all events as compared with freedom from other defects.

The gathering process in its turn is responsible for further defects of sheet-glass. Some of these, such as defects arising from the use of a dirty pipe, are never allowed to pass beyond the sorting-room, and are therefore of no interest to the user of glass. Of those whose traces are seen in the glass that passes into use, “blisters” and “string” are the most important. “Blisters” are somewhat larger, flat air-bells, arising from the inclusion of air between successive layers of the gathering. “String” is a very common defect in all sheet-glass. To some extent it may arise from want of homogeneity in the glass itself. If this consists of layers of different densities and viscosities, the gatherer will take these up on his gathering, and ultimately they will form thickened ridges of glass running around the cylinders and across the sheets. Such striÆ, due to want of homogeneity in the glass, are much more common in flint glass than in the soda-lime glasses used for sheet manufacture, but are not unknown in the latter. On the other hand, even if the glass be as homogeneous as possible, the gatherer can produce these striÆ if he takes up his glass from a place close to the side of the fire-clay ring that floats in the furnace in front of his working opening. Glass always acts chemically upon fire-clay, gradually forming a layer of glass next to the fire-clay that contains much more alumina than the rest of the contents of the furnace. Such a layer is formed on the surface of each ring in a sheet tank, but if the gathering is taken from the centre of the ring, this thick viscous layer of aluminiferous glass remains undisturbed. If, however, the gatherer brings his pipe too near the side of the ring, the glass will draw some of this different layer on to the gathering, and this glass will form thick ridges and striÆ running across the sheet in all directions. Another defect for which the gatherer is generally responsible is that of variation of thickness within the same sheet. The blower, however, can also produce this defect.

During the blowing proper, a further series of defects may be introduced, principally by allowing particles of glass derived from certain stages of the process to fall upon the hot glass of the cylinder and there become attached permanently. More serious, and also more frequent, is the greater or less malformation of the cylinder. If the glass as it leaves the blower is of any shape other than that of a true cylinder, it becomes impossible to spread it into a truly flat sheet in the flattening kiln. Sometimes, in practice, the “cylinder” is wider at one end than at the other, or, worse still, it is of uneven diameter, showing expanded and contracted areas alternately. When such a cylinder comes to be spread out on the slab it cannot be flattened completely, and various hollows and hillocks are left, which mar the flatness of the sheet and interfere with the regular passage of light through it when in use.

Finally, the process of flattening is apt to introduce defects of its own. The most common of these are scratches arising from marks left by the flattening tool; indeed, in all sheet glass it is quite possible to see, by careful examination of the surfaces, upon which side the flattening tool was used. Sheet-glass thus has one side decidedly brighter and better in surface than the other, the better side being that which rested upon the “lagre” during the flattening process. On the other hand, if the slab itself be not quite perfect, or if any foreign body be allowed to rest upon it, that side of the glass will be marked in a corresponding manner.

In the account of the manufacture of sheet-glass given above, we have outlined one typical form of the process, but nearly every stage is subject to modifications according to the practice and particular circumstances of each works. We will now describe one or two special modifications that are of more general importance.

First, as regards the melting process, although the tank-furnace has almost entirely superseded the pot furnace for the production of ordinary sheet-glass, there are still some special circumstances under which the pot furnace is capable of holding its own. Thus, where for special purposes it is desired to produce a variety of sheet-glass which, as regards all defects arising out of the glass itself, and especially as regards colour, is required to be as perfect as possible, melting in pots is found advantageous, and for some very special purposes even covered (hooded) pots are used. For such special purposes, too, sulphate of soda is eliminated from the raw materials and carbonate of soda (soda ash) substituted. For the production of tinted glasses also, whether they are tinted throughout their mass, or merely covered with a thin layer of tinted glass (“flashed”), manufacture in pot rather than tank furnaces is generally adopted, the exact nature and composition of the glass being far better under control in the case of pots.

The blowing process is also subject to wide variations of practice. The most important of these variations concerns the shape and dimensions of the cylinders. In English and Belgian works the dimensions of the cylinders are so chosen that the length of the cylinder constitutes the longest dimension of the finished sheet, the diameter of the cylinder forming the shorter dimension. In some parts of Germany, however, the practice is the reverse of this, the cylinders being blown shorter and much wider, so that the circumference of the cylinder constitutes the longest dimension of the finished sheet. It is, however, pretty generally recognised that the latter method has very serious disadvantages, although it is claimed that somewhat more perfect glass can be obtained by its means. For the production of a special variety of glass, known as “blown plate glass,” this method of blowing short wide cylinders is still adhered to. This is a very pure form of sheet-glass, blown into thick, small sheets which are subsequently ground and polished in the same manner as plate-glass. Here the great thickness of glass required seems to render the blowing of long cylinders very difficult, and the other form is therefore adopted. On the other hand, English patent plate-glass, which is made by grinding and polishing the best quality of ordinary sheet-glass, is made from glass blown into long narrow cylinders in the manner described in detail above.

The process of blowing described above is capable, with slight modifications, of yielding glass with surfaces other than the plain smooth face of ordinary sheet-glass. Thus fluted and “muffled” glass are produced in a very similar manner to that described above for ordinary sheet, except that the fluting or the irregular surface markings which constitute the peculiarities of these two varieties of glass, are impressed upon the surface of the cylinder at an early stage in the process.

From the outline description given above of the usual method of manufacture of sheet-glass, it will readily be seen that this is a long, complicated, and laborious process, involving the employment of much skilled labour, and involving the production of a relatively complicated form, viz., the closed cylinder, as a preliminary to the production of a very simple form, viz., the flat sheet. It is therefore by no means surprising to find that a great many inventors have worked and are still working at the problem of a direct mechanical method of producing flat glass possessing a natural “fire polish” at least equal to that of ordinary sheet-glass. The earlier inventors have almost uniformly endeavoured to attain this object by attempting to improve the process of rolling glass, with a view to obtaining rolled sheets having a satisfactory surface. We have already indicated why these efforts have never met with success and what reasons there are for believing that they are never likely to attain their object. A totally different line is that taken by Sievert, to whose inventions we have already referred in connection with the mechanical production of blown articles. This inventor has endeavoured to utilise his process for blowing large articles of glass for the direct production of sheets of flat glass. His method is to blow, by the steam process described in another chapter, a large cubical vessel, having flat sides, the flatness of these sides being ensured by blowing the vessel into or against a mould having flat sides. This flat-sided vessel is ultimately to be cut up into five large sheets. This process also appears to involve some of the main difficulties of rolling as regards the means of transferring the glass from the furnace to the plate of the blowing machine, and in practice the inventor has not yet succeeded in producing glass of sufficiently good surface for the purposes of sheet glass.

Another class of processes entirely avoid all means of transferring molten glass from the furnace to any machine, by working on glass direct from the molten bath itself. Some of these processes are in actual use in America, and others are being experimented with in Europe, but their complete technical and commercial success has yet to be proved; there can, however, be little doubt that they have overcome the greatest of the many difficulties that stood in the way of the mechanical production of sheet-glass, and that they are therefore destined very shortly to solve the problem completely, in which case they would, of course, rapidly supersede the hand process.

One of the earliest of these direct processes proposed to allow the molten glass to flow out from the furnace, downward, through a narrow slit formed in the side or bottom of the tank. The impossibility of keeping such a narrow orifice open and at the same time regulating the flow of glass made this proposal impracticable, although the use of drawing orifices has been revived in one of the latest processes.

The American process, which is said to be at work under commercial conditions, is not entirely satisfactory in this respect—that it is a mechanical process for the production of cylinders and not of flat sheets, so that the subsidiary processes of splitting and flattening still remain to be carried out as before. In this process an iron ring is lowered into the bath of molten glass through an aperture from above; the glass is allowed to adhere to the ring which is then slowly raised by mechanical means, drawing a cylinder of glass with it. If left to itself, such a cylinder, owing to the effects of surface tension in the glass, would soon contract and break off, but the American invention avoids this action by chilling each bit of the cylinder as soon as it is formed. This is done by the aid of air blasts delivered upon both sides of the glass as it emerges from the bath, and it is claimed that by this means cylinders of any desired length and diameter may be drawn direct from the bath. The obviously great mechanical difficulties connected with these operations have probably been overcome, but not without sacrificing much of the simplicity of the arrangement, and the relative economy of this process as a whole, compared with the hand process, has yet to assert itself.

The inventions of Fourcault, which are at present being developed on the Continent by a syndicate of glass manufacturers, aim at a much more direct process. Here also the glass is drawn direct from the molten bath by the aid of a drawing-iron that is immersed in the glass and then slowly raised, but in this case the piece immersed is simply a straight bar, and the aim is to draw out a flat sheet. In this case the tendency, under surface tension, is to contract the sheet into a thread, and apparently the simple device of chilling the emerging glass is not adequate to prevent this in a satisfactory manner, and subsidiary devices have been added. Those that have been patented include a mechanism of linked metal rods so arranged as to be immersed and drawn out of the glass continuously with the emerging sheet, in such a manner as to support the vertical edges of the glass and so aid in resisting the tendency of the glass to contract laterally. Another device consists in the use of a slit or orifice formed in a large fire-brick that floats on the surface of the glass. Through this orifice the glass is drawn, of the desired thickness and width. The use of this orifice, however, interferes markedly with the perfection of the product, and in fact all the glass produced in this way shows quite plainly a set of longitudinal striations due to the inevitable irregularities in the lips of the drawing slot. Further, it appears to be impracticable to draw thin glass in this way, a thickness of from 2½ to 3 millimetres (about 1/8 inch) being the least that is practicable, on account of the large amount of breakage that occurs with weaker sheets. This process, in its present stage of development, however promising, does not appear to have solved the problem of mechanical manufacture of sheet-glass, since it is just in the thinner, lighter kinds of glass that the advantages of sheet are most pronounced. On the other hand, it is quite possible that this drawing process, or some development arising from it, may shortly supplant the casting process in the production of polished plate-glass, although for the largest sizes of this product also, the difficulty and danger of handling the weights involved may prove a serious obstacle.

Crown Glass.—Although this is a branch of manufacture that is nearly obsolete, it deserves brief notice here, partly because it is still used for the production of special articles, and also because it illustrates some interesting possibilities in the use and manipulation of glass.

The process of blowing crown glass may be briefly described as that of first blowing an approximately spherical hollow ball, then opening this at one side and expanding the glass into a flat disc by the action of centrifugal forces produced by a rapid rotation of the glass in front of a large opening in a special heating furnace. The actual process involves, of course, the preliminary of gathering the proper quantity of glass, much in the manner already described in connection with sheet-glass manufacture. This gathering is then blown out into a hollow spherical vessel. This vessel is now attached to a subsidiary iron rod by means of a small gathering of hot glass, applied at the point opposite the pipe itself, the glass being thus, for a moment, attached to both the pipe and the “pontil” or “punty” (as the rod is called). The pipe is, however, detached by cracking off the neck of the original glass, which now remains attached to the pontil in the shape of an open bowl. This bowl is now re-heated very strongly in front of a special furnace, the open side of the bowl being presented to the fire. The pontil is meanwhile held in a horizontal position and rotated. As the glass softens the rotation spreads it out, until finally the entire mass of glass is formed into a simple flat disc spinning rapidly before the mouth of the furnace. This flat disc or “table” of crown glass is allowed to cool somewhat, is detached from the pontil by a sharp jerk, and is then annealed in a simple kiln in which the glass is stacked, sealed up, and allowed to cool naturally. It is obvious that by this process no very large sheets of glass can be produced; tables 4 ft. in diameter are already on the large side, and these can only be cut up into much smaller sheets on account of the lump of glass by which the table was originally attached to the pontil, and which remains fixed in the centre of the finished disc. For certain ornamental purposes, where an “antique” appearance is desired, these bullions are valued, but for practical purposes they interfere very seriously with the use of the glass. As a matter of fact, even several inches away from the central bullion itself, crown glass is generally marked with circular wavings, which render it readily recognisable in the windows of older buildings, but which decidedly detract from the perfection of the glass. On the other hand, crown glass is still valued for certain purposes, such as microscope slides and cover glasses, where entire freedom from surface markings, such as those found in sheet glass as a result of the flattening operations, is desirable. While, therefore, the process has merely an historical interest so far as ordinary sheet-glass purposes are concerned, it is still used in special cases.