The principles of glass-making

Contents

Page

e Principles of Glass Making, and the Manufacture of Hollow Ware.. . . 1

the Manufacture of Crown and Sheet Glass . .101

Preface

T HE different branches of the manufacture of glass are of so distinct a nature that it has been thought de¬ sirable to secure the services of gentlemen who have had practical experience of the more important divisions of the business in the preparation of this handbook. By this means it is hoped that a fuller and more accurate account of the details of the various processes has been insured than might perhaps have been the case had the volume been the work of a single author.

The first part of the work, by Mr. Powell, of Whitefriars Glass Works, is introductory to the whole. It treats of the general principles of the manufacture, methods of analysis, raw materials, apparatus, tools, &c. It also includes the more special divisions of the manufacture of “ hollow ware” and “glass mosaic” or stained and painted glass. The second part, by Mr. H. Chance, of Birmingham, is confined entirely to the manufacture of crown and sheet glass; and the third part, by Mr. Harris, the Secretary to the British Plate Glass Company, to the manufacture of plate glass.

An appendix contains the examination papers set in the subject of glass-making in the Technological Examinations of the City and Guilds Institute, and a second appendix gives a list of works on glass for the use of students in these (examinations and others. This list does not profess to be a

VI

PREFACE.

complete bibliography of glass-making ; it is only intended as a guide to those who wish to extend their studies beyond the necessarily narrow limits of an elementary handbook.

The Publishers have to acknowledge their indebtedness to Messrs. Spon for permission to reproduce many of the illustrations used in the excellent article on Glass in the Encyclopaedia published by that firm.

H. T. W.

THE PRINCIPLES OF GLASS MAKING, AND THE MANUFACTURE OF HOLLOW WARE.

BY H. J. POWELL, B. A.

INTRODUCTORY.

PRACTICAL knowledge of the manufacture of glass is generally acquired through apprenticeship, and the study of hereditary recipes. There is danger that the proficiency, which may be thus attained, will be based on a groundwork of “ rule of thumb ” rather than of science. The want has often been felt of a work illustrating the present condition of the manufacture, and at the same time defining the principles upon which experience has proved the several processes to be based. Such a work should be valuable as a guide amidst disjointed recipes, as well as a basis for the accumulation of new facts and methods. It must, however, be remembered that a manual can never supply the place of experiment and experience.

The order of contents observed in this division of the work is as follows :—

1. The discovery of the constituents of a given sample

of glass.

2. The chemical and physical qualities of glass.

B

PRINCIPLES OP GLASS MAKING.

3. The raw materials used in the manufacture of

glass.

4. The means employed for utilizing the qualities of

glass for manufacturing purposes, namely, cru¬ cibles, furnaces, and auxiliary furnaces.

5. The special treatment of the several branches into

which the manufacture is divided ; namely,

The manufacture of hollow ware, and the pro¬ cesses by which decorative effects are obtained. The manufacture of flat, white, and coloured, glass for glazing, and the manufacture of optical glass.

CHAPTER I. Analysis of Glass

T HE word “ glass” suggests a substance which is trans¬ parent, solid, hard, and brittle. This description is far from complete, and will require considerable amendment. The manufacturer must have a more accurate acquain¬ tance with this substance, and if a specimen be handed to him for imitation he must be able to determine of what the specimen is composed, and how a similar material can be produced. Part of the specimen must be subjected to chemical analysis, both qualitative and quantitative. The processes employed in the analysis of glass differ in some respects from those of simple analysis owing to the virtual insolubility of the material in all acids except hydric fluoride. It will not be out of place to indicate a simple method of procedure, omitting such particulars as are to be found in elementary works on practical chemistry.

ANALYSIS OF GLASS.

The Discovery of the Constituents of a given Sample of Glass.

Reduce a fragment of the glass to impalpable powder, and divide the powder into five parts. One of these parts should be expended in preliminary tests, with the blow¬ pipe flame, using charcoal and borax-beads as carriers. By these tests the presence or absence of certain metals may be determined. Of the remainder of the powder, set apart two parts for the qualitative, and two for the quantitative examination.

Qualitative Analysis (Part A).

Mix one part with about four parts of a mixture of the anhydrous sodic and potassic carbonates ; the weights of the ingredients being in the proportion of 53 of the sodic to 69 of the potassic carbonate. The mixture of glass and carbonates is placed in a platinum crucible and heated until the fusion is complete, and bubbles are no longer given off. The crucible is removed from the flame, and the contents, when cold, are placed in a porcelain dish and treated with dilute hydric chloride, a gentle heat being applied until everything is dissolved except certain white flakes, which remain floating in the liquid. If the pre¬ liminary tests have indicated the presence of lead in the glass, the contents of the crucible should be treated with hydric nitrate instead of hydric chloride. Separate the white insoluble flakes from the liquid by filtration; the liquid which passes the filtering medium and is known as the filtrate, holds in solution all the metallic oxides present in the glass under examination. What metals are present will be discovered by the ordinary analytical tests ; it will, however, be useless to look for sodium and potassium, as the carbonates of both these metals have intentionally been added. Wash thoroughly, and dry the white in-

PRINCIPLES OP GLASS MAKING.

soluble flakes remaining on tbe filtering medium. Place the dry powder detached from the filtering medium, in a platinum dish, and warm gently with an excess of perfectly pure hydric fluoride. The substance combines with the fluorine of the hydric fluoride and passes off in a gaseous form, leaving no residue. The substance is pure silica.

Place another part of the powdered glass in a platinum dish. Cover the powder with pure hydric fluoride, and stir into a thin paste with a platinum wire. Evaporate the paste to dryness over a water-bath, and to the dry mass add a small quantity of hydric chloride. Evaporate once more to dryness, in order to expel the volatile fluorides and silicic fluoride, and dissolve the residue, when cool, in a large quantity of distilled water. Into this solution pass hydric sulphide gas for a considerable time, filter off any precipitate, boil the filtrate in order to expel the excess of hydric sulphide, and reoxidize with a few drops of hydric nitrate. Add ammonic hydrate, filter off any precipitate, and boil the filtrate until it is highly concen¬ trated. Add a strong solution of ammonic sesquicarbonate, then ammonic hydrate, and allow the mixture to stand. Filter off any precipitate. The filtrate may retain potassic and sodic chloride together with certain ammonic salts. Evaporate the liquid to dryness, and ignite carefully until every trace of ammonia has been dispelled. Dissolve the residue in a small quantity of distilled water, acidu¬ lated with hydric chloride, and divide into five parts— ct, &, c, 6.

(a) A yellow precipitate, when platinic chloride is added, indicates potassium.

(b) Add hydric tartrate and stir with a glass rod. If potassium be present, and the solution be sufficiently concentrated, a crystalline precipitate of potassic tartrate will be found.

ANALYSIS OF GLASS.

(c) Add hydric silico-fluoride. A white opalescent preci¬ pitate indicates potassium.

(d) Add hydric potassic metantimoniate. A white crystalline precipitate indicates sodium.

(e) Evaporate to dryness, mix the residue with alcohol, and ignite. A yellow flame indicates sodium. Examine the flame through a thick sheet of dark-blue glass. If potassium be present the characteristic violet flame will become visible, even in the presence of the intense of the sodium flame.

Quantitative Analysis.

The qualitative analysis will have discovered x&jW presence of silica, together with certain metallic oxides^ but the quantity by weight of each substance in a given weight of the glass remains to be determined. Dry the remainder of the impalpably powdered glass at 100° C. and place it in a desiccator to cool. Weigh a convenient quantity, and place it in a shallow platinum dish, moisten the powder with pure hydric sulphate, and place the platinum dish on a leaden or platinum tripod in a leaden box, having previously covered the bottom of the box with a stiff paste made with powdered fluor spar and hydric sulphate. Cover the box loosely with a sheet of lead, and place it where fumes can pass away without causing injury or discomfort. A gentle heat is applied and con¬ tinued until the glass is completely decomposed. When decomposition is complete, which will be in from four to five hours, the platinum dish with its contents is placed in a larger dish of the same metal, and pure hydric sulphate is added drop by drop in more than sufficient quantity to convert all the metallic oxides into sulphates. The outer dish is heated in order to drive off from the mixture the volatile fluorides and silicic fluoride, and the mixture is subsequently ignited in order to expel the excess of free

PKINCIPLES OF GLASS MAKING.

hydric sulphate. The residue, when cold, is moistened with hydric chloride, diluted with distilled water, and gently warmed. If lead and barium be present in the glass a white deposit of the mixed sulphates of these metals remains undissolved. If this be the oase, concentrate the liquid by evaporation, add an excess of hydric sulphate, and continue the application of heat until the hydric sulphate itself begins to volatilize. By this treatment the hydric chloride is expelled and the sulphates of lead and barium are entirely precipitated. The strongly acid mix¬ ture is allowed to cool, distilled water is added, and the clear solution (a) is separated from the precipitate by filtration. The precipitate still remaining on the filtering medium is washed first with water acidulated with hydric sulphate, and then with alcohol: the filtrate in each case being added to the solution (a,). The precipitate and filtering medium are placed in a small glass beaker and a solution of ammonic sesquicarbonate is added in order to convert the sulphate of lead into carbonate. This process requires considerable time and a low temperature. The whole mixture is transferred to a filter, and the solid residue is washed first with solution of ammonic sesqui¬ carbonate, and then with distilled water. The carbonate of lead is now separated from the sulphate of barium by careful treatment with hydric acetate, which dissolves out the carbonate of lead, but leaves the sulphate of barium undissolved. The residue is well washed, dried with the filtering medium at 100° C., and ignited in a porcelain crucible. The porcelain crucible when cool is transferred to a desiccator. The weight of the crucible and its contents, less the ascertained weight of the crucible to¬ gether with that of the ash of the filtering medium (the weight of the latter having been ascertained by igniting and weighing an equal quantity of an identical material) gives the weight of sulphate of barium. The clear solution containing the carbonate of lead dissolved in hydric acetate is evaporated with an excess of hydric sulphate until the latter begins to volatilize. By this treatment the lead is reprecipitated as sulphate of lead. This precipitate is separated by filtration, and dried, ignited, and weighed in the same manner as the sulphate of barium. The acid solution (<x), separated by filtration from the sulphates of lead and barium, contains as sulphates whatever other metals are present in the sample of glass. These metals must be separated by the ordinary methods. In order to determine the weight of silica another part of the powdered glass must be used.

CHAPTER II. The Chemical and Physical Qualities of

T HE analysis of every sample of glass discovers the presence of the substance known as silica. For this reason silica (Si0 2 = 60), more accurately termed the di¬ oxide or anhydride of silicon, is recognized as the essential constituent of glass. There are many varieties of glass, which differ one from another in the metallic oxides asso¬ ciated with the silica, every true glass containing at least two metallic oxides. The oxides which occur most fre¬ quently are those of potassium, sodium, lead, calcium, barium, aluminium, and magnesium. Traces of other metals are found which are present as impurities, or have been added to produce effects of colour. Analysis deter¬ mines the quality and quantity of the constituents of glass, but throws little light upon the manner in which they are associated. It is known that the association of the silica with the metallic oxides is effected by heat, but the re¬ lationship by which they are bound together requires further investigation. Taking the result of an analysis of flint or lead glass as an instance, namely, silica 53*17,

PRINCIPLES OF GLASS MAKING.

potassic oxide 13 - 88, plumbic oxide 32"94, the problem to be solved is, what chemical or physical changes developed heat have converted these ingredients into an amor¬ phous transparent body ? It sometimes happens when red lead (Pb 3 () 4 ) is subjected to prolonged heat in an ordi¬ nary fire-clay crncible, that the oxide of lead finds its way through the bottom of the crucible by uniting with the free silica in the clay to form silicate of lead. If sand (the commercial representative of silica) be mixed with red lead in various proportions and the mixtures be heated, silicates of lead will be formed differing from each other in quality according to the difference of the proportions of the ingredients. Potassic carbonate consists of potassic oxide associated with carbonic anhydride, and potassic sul¬ phate of potassic oxide associated with sulphuric anhydride. If potassic carbonate be heated in a platinum crucible until it becomes perfectly fluid, and dry sand be gradually intro¬ duced, an effervescence will be set up and a new substance will be found in the crucible composed of various potassic silicates. The nature of the disturbance caused by the intro¬ duction of the sand may be ascertained by placing the pla¬ tinum crucible containing the fluid carbonate at the bottom of a deep vessel, adapted for the collection of the evolved gas, and subjecting this gas to the ordinary tests for carbonic anhydride. By simple experiments of a similar character it is found that sodic, calcic, baric, and other silicates may be obtained by the indirect decomposition of the respective carbonates caused by the combined action of heat and silica. It has also been found that the energy excited in silica by heat is sufficient to decompose the sulphates of sodium and potassium, the decomposition being accom¬ panied by the evolution of sulphuric anhydride, and the formation of sodic and potassic silicates. The heat required for the expulsion of the sulphur in the form of sulphuric anhydride is very intense; but if the sulphate be mixed with about one-tenth of its weight of charcoal or coal-dust, which robs the sulphate of a part of its oxygen to form carbonic

QUALITIES OF GLASS.

oxide, the sulphur is expelled in the form of sulphurous anhydride and at a lower tern peratui’e. These experiments supply sufficient data for settling the question as to the constitution of flint and of the other varieties of glass. If potassic silicate results from the fusion of sand with a substance containing potassic oxide, and plumbic silicate from the fusion of sand with an oxide of lead, it will not be too much to assume that if the same ingredients be simul¬ taneously fused, the result will be a mixture of the same sili¬ cates. Flint glass is therefore regarded as a mixture of the silicates of lead and potassium ; Bohemian glass, which is made up of silica with the potassic and calcic oxides, as a mixture of the silicates of potassium and calcium; and Plate or Sheet glass, the constituents of which are silica and the calcic and sodic oxides, as a mixture of the silicates of cal¬ cium and sodium. In fact, every variety of true glass is found to be a mixture of at least two metallic silicates. The word “ mixture ” is used in preference to “compound,” as there is difficulty in determining the molecular composition of consti¬ tuent silicates. The difficulty is caused, in great measure, by the fact that silica unites with the same metallic oxide in several different proportions, which is well illustrated by the formulae of the silicates of sodium (Ha 2 0, Si0 2 : 2Na 2 0, 5Si0 2 : Na 2 0, 3Si0 2 : Na 2 0, 4SiO). Experience, however, has ascertained a proportion of ingredients for each variety of glass, which produces a material possessing the requisite qualities. The results of analyses of the glasses already referred to have been represented by formulae resembling that of felspar (K 2 0, A1 2 0 3 , 6Si0 2 ), felspar being a natural glass. The accepted formula of flint or lead glass is K 2 O, PbO, 6Si0 2 : of Bohemian combustion tube, K 2 0, CaO, 6Si0 2 , and of plate or sheet glass Ha 2 0, CaO, 6Si0 2 . As the proportions indicated by each formula are adhered to, so the quality of the glass is insured, pro¬ vided the raw materials are sufficiently pure. The quality of a glass is the resultant of the qualities of its constituent silicates. It has already been shown that silicates are

PRINCIPLES OF GLASS MAKING.

compounds formed by the direct or indirect combination of silica with metallic oxides, and that their constitu¬ tion is exceedingly varied. Silicates are more or less durable according to the greater or less proportion of silica which they contain. The silicates of lead are de¬ composed by acids, and the sodic and potassic silicates dis¬ solve in boiling water. The plumbic sesquisilicate (2PbO, 8Si0 2 ), a pale yellow transparent substance, and the potassic or sodic tetrasilicate (Na 2 0, (or K 2 0) 4Si0 2 ), which have been commonly known as glasses, are so unstable as to be useless for those purposes for which true glass is especially valuable. The latter, which has received the name of soluble glass, is obtained by fusing 8 parts of sodic or 10 parts of potassic carbonate with 15 parts of sand and 1 part of charcoal. The charcoal, by robbing the carbonate of oxygen, facilitates its decompo¬ sition. The fused mass is dark in colour, but transparent. The different silicates vary greatly in fusibility. The aluminic disilicate (A1 2 0 3 , 2Si0 2 ) and the calcic and magnesic silicates can only be fused with the greatest difficulty. The ferrous and manganous sesquisilicates (2FeO, (or 2MnO) 3Si0 2 ) fuse readily, and crystallize on cooling, and the plumbic sesquisilicate fuses with even greater facility. A valuable quality belonging to mixtures of silicates is the fact that the fusing point of a mixture is considerably lower than the mean of the fusing points of the silicates constituting the mixture. Thus, although the aluminic, calcic, and magnesic silicates are practically, when separate, infusible, a mixture of two of the same silicates fuses at a comparatively low tempera¬ ture. It is also found that a glass containing three or more silicates is more fusible, but less perfect in structure and in purity, than a glass containing two silicates only. The qualities contributed to the different glasses by their constituent silicates are as follows :—Potassic silicate is free from colour, and contributes brilliancy and fusibility. Sodic silicate produces a sea-green colour, but adds to the

QUALITIES OF GLASS.

fusibility and brightness of the glass. Plumbic sil in excess, causes a yellow tint in the glass, and dimit its durability ; the excess rendering the glass liable to ; tion by acids, and to lustre-giving decay by combinatioi with sulphur. Plumbic silicate increases both the fusi¬ bility and ductility of glass, and by raising its density adds greatly to its refractive power, and consequently to its brilliancy. The density of the silicate, however, is a constant source of trouble. Flint glass is a mixture of the plumbic and potassic silicates, of two materials differing vastly in specific gravity. If oil and water be stirred to¬ gether a veiny and streaky effect is produced ; the veins and cords which so commonly disfigure ordinary flint glass, and which can only be obviated in the optical flint glass by great and prolonged labour, are, in like manner, due to the inequality in density of the two constituent silicates. Baric silicate, without imparting any colour to the glass, increases the resultant density and brilliancy, though in a less degree than the plumbic silicate. Calcic silicate, when the calcic oxide is in excess, produces a milky appearance, but other¬ wise is chiefly valuable as supplying stability. The calcic and sodic silicates in comparison with other constituent sili¬ cates are the least removed from each other in specific gra¬ vity, and consequently yield a glass of great homogeneity.

Bottle glass, a mixture of the silicates of sodium, alu¬ minium, and calcium, and Yenetian glass, a mixture of the silicates of sodium, potassium, and calcium, are readily fusible, but are deficient in purity and lustre ; the latter defect, however, and the small specific gravity of V enetian glass, are reckoned as advantages.

PRINCIPLES OP GLASS MAKING.

CHAPTER III. The Chemical and Physical Qualities of Glass. ( Continued .)

l HE most important property belonging to mixtures

1 of silicates, but one which is not equally developed in all mixtures, is that curious condition of viscosity, in¬ termediate between solidity and liquidity, which is induced and is maintained by heat, and which renders possible the chief processes of the manufacture of glass. After the decomposition of the raw materials of glass in the cru¬ cible, after the expulsion of the gases, and the combination of the silica with the metallic oxides, the glass which is formed passes by imperceptible stages into a liquid condition. This condition the glass maintains as long as it is subject to the full heat of the furnace in which the crucible rests, but if the heat be slightly diminished, or the glass chilled by exposure to the air, the property of ductility, viscosity, or tenacity is acquired. In the liquid state glass can be poured or ladled directly from the crucible ; in the viscous state, it can be “gathered” or coiled on the heated end of an iron rod, very much in the same way in which thiok treacle can be coiled round the bowl of a spoon. In this condition, if the gathering-rod be hollow, the solid mass col¬ lected at the end of the rod can be expanded by the skilful use of the workman’s breath into a bulb, which is the ger¬ minal or primitive stage of all manufactured glass except that which derives its form from manual or mechanical pressure. The form of the bulb may be modified by caus¬ ing it to expand and assume the shape of a resisting envi¬ ronment, or may be either flattened or elongated by the force of gravitation. If to a part of the bulb remote from the gathering-iron a second iron be attached by a seal of glass, the bulb may be prolonged into tube. If a solid mass of glass be treated in the same manner it may be drawn out to an almost indefinite length, and to a state of extreme tenuity. If the bulb be severed from the gather-

QUALITIES OF GLASS.

ing-iron by the unequal contraction of a part of the glass caused by the application of a cold substance, and a solid iron rod be attached to a point remote from the fracture, and the bulb be reheated, and if after reheating, the rod to which the severed bulb is attached be rapidly “trundled,” the glass, like the head of a trundled mop, will expand into a disc by the action of centrifugal force, but will be prevented from fracture by cohesion. When a heated vessel of unequal thickness is left by the work¬ man to cool in the open air, the thin parts become cool sooner than the parts which are comparatively thick, and the outer layers of the thick parts sooner than those nearer the centre. By unequal cooling, tension is excited between the thick and the thin parts of the vessel, as well as between the external and internal layers of its sub¬ stance. A heated vessel which is very thin and of equal substance, may be left to cool with impunity, as the cool¬ ing is equal throughout. Sudden cooling applied to one point of a vessel, as by the application of a moistened tool, or by the fall of a drop of water, causes fracture at that point by exciting tension. So also the impact of hot iron upon cold glass, or the introduction of boiling water into cold glass vessels is a common cause of breakage. When molten glass is ladled into water, it becomes brittle and crackled throughout its substance. When a small quan¬ tity of glass is gathered from the crucible on an iron rod, and a drop is allowed to fall from the iron into boiling water, and is removed from the water as speedily as possible, it is found to possess curious properties. The shape of the drop, as shown at a, fig. 1, is that of a tear with a long tail. The head of the tear is exceedingly hard, and resists a heavy blow without fracture; it also can scarcely be marked by file or diamond. If, however, the tail be frac¬ tured by a sudden blow, by pliers, or by the corrosion of hydric fluoride, the entire mass will be disintegrated, as shown in fig. 1, b, c, d , and may be crumbled in the hand without injury. If a tear, made in the same way, be buried in sand and reheated to the point of ductility, and then

PRINCIPLES OF GLASS MAKING.

allowed to cool exceedingly gradually with the sand in a closed oven, it may be easily marked or scratched, and, may be broken with the same facility and the same fracture as ordinary glass. A shaped vessel dropped from the work¬ man’s rod into boiling water, whilst still red hot and retain¬ ing part of its ductility, is immediately broken up, but the liability to fracture is reduced by substituting for the water liquids which can be raised to higher temperatures than the boiling point of water. This principle is the founda-

Prince Rupert’s Drops.

tion of M. de la Bastie’s invention of hardened or “ toughened ” glass. The use of boiling fat in the place of water for immersing the heated glass renders possible the hardening of vessels of simple form. Several difficul¬ ties attend the process, which at present have been only partially overcome. It is essential that the whole surface, and as far as possible the whole substance of a vessel, should be simultaneously affected by the change of tempera¬ ture produced by immersion in the liquid. It will readily be

QUALITIES OF GLASS.

understood that if one part is chilled sooner than another the whole vessel will suffer. In order that simultaneous cooling may be attained, every part of a vessel must be at the same temperature, a condition which is almost impos¬ sible to be fulfilled if the vessel consists of several parts ; the substance of the vessel must be equal throughout, and the whole surface of the vessel must at the same time come in contact with the liquid. The last condition im¬ plies that the air contained by a hollow vessel can be ex¬ pelled through a strongly resisting liquid in an inappreci¬ able space of time. The nature of glass which has been immersed in the molten fat, closely resembles that of the tears of glass dropped into boiling water. The hard¬ ness of the surface of vessels treated by this process, like that of the tears, is sufficient to withstand considerable violence, and the fracture of both is identical; moreover, a microscopic examination of a fragment of the thicker parts of a hardened vessel discovers the substance to be divided by a porous and striated band, which is in marked contrast to the solidity of the superior and inferior crust.

The disadvantage of the practical use of the hardening process, in addition to the difficulties which have been already referred to, is the unreliability of the glass pro¬ duced. Reliable strength and durability can at present only be obtained in glass by allowing it, after manipula¬ tion, to cool as gradually and as regularly as possible. Gradual cooling, or “annealing,” is practically attained by exposing the red-hot glass to a source of considerable heat, and either gradually withdrawing the glass from the heat, or allowing the source of heat gradually to become extinct. The best explanation of the various phenomena connected with the rapid and gradual cooling of glass is afforded by the hypothesis of the repulsive force excited between the molecules of matter by heat. Under the influence of this force bodies tend to expand, and finally to change their condition of aggregation. This hypothesis assumes the existence of physical pores, which increase in

PRINCIPLES OF GLASS MAKING.

size as the body passes from the solid to • the liquid and from the liquid to the gaseous condition. If therefore the liquidity of glass be caused by the increase in the size of its inter-molecular pores, the stages of viscosity, ductility, and solidity will result from the gradual diminution of the same. When glass is allowed to cool gradually, its pores will become regularly closed or diminished in size through¬ out the entire substance, and it will acquire regular and reliable strength ; whereas when glass is rapidly cooled the crust will be intensely solidified, whilst the molecules of the interior will be withheld, by adhesion to the crust, from uniting and reducing the internal porosity. Glass is a bad conductor of heat, and this property tends to aggra¬ vate the difference in condition between the crust and the interior of the material when rapidly cooled. It is only the intense solidity or hardness of the crust of rapidly cooled glass which withholds the mass from collapse, and as soon as the solidity is weakened, disintegration necessai-ily ensues.