must be regulated so that the brush flame produced is free from every sign of incomplete combustion,[5] which may be known by its outer zone being only faintly visible in daylight, and quite free from luminous streaks (see Fig. 4, p. 9). When a suitable flame has been produced, try it by rotating a piece of lead glass at or near the end of the inner blue part of the flame (A Fig. 4); the appearance of the glass will quickly indicate reduction. When this occurs move the glass forward to the end of the outer zone B, but keep it sufficiently within the flame to maintain it at a high temperature. If all is right the metallic reduction will quickly disappear, the glass will become perfectly transparent once more, and will present the appearance previously observed in the experiments with the pointed flame, or, if very hot, assume a brownish-red appearance. If this does not occur, the supply of air must be increased or the supply of gas diminished until the proper effects are secured.

In working upon lead glass with the highly oxidising brush flame, it is a good plan to heat it in the reducing part of the flame A for thoroughly softening the glass, and to remove it to the oxidising flame B to burn away the reduced metal. In prolonged operations, in order that reduction may never go too far, hold the glass alternately in the hot reducing flame and in the oxidising flame. The inferiority of the outer oxidising flame to those portions nearer the inner blue zone for softening the glass, may perhaps be accounted for by the presence of a larger proportion of unconsumed air in the former, which being heated at the expense of the hot gases produced by combustion, thereby lowers the temperature of the flame. At or near A (Fig. 4) where the combustion is nearly complete, but no excess of air exists, the temperature will naturally be highest.

If a very large tube be rotated in the oxidising flame at B (Fig. 4) it may happen that the flame is not large enough to surround the tube, and that as it is rotated those parts of it which are most remote from the flame will cool down too considerably to allow all parts of the tube to be simultaneously brought into the desired condition. This difficulty may be overcome by placing two blow-pipes exactly opposite to each other, at such a distance that there is an interval of about an inch between the extremities of their flames, and rotating the tube between the two flames. It may be necessary to provide two blowers for the blow-pipes if they are large.

Again, if a very narrow zone of a tube of moderate size is to be heated, two pointed flames may be similarly arranged with advantage. Occasionally more than two flames are made to converge upon one tube in this manner.

Another method of preventing one side of a tube from cooling down whilst the other is presented to the flame, is to place a brick at a short distance from the extremity of the flame. The brick checks the loss of heat considerably. A block of beech wood may be used for the same purpose, the wood ignites and thereby itself becomes a source of heat, and is even more effective than a brick.

Fuller details of the management of lead glass under various circumstances will be found in the subsequent descriptions of operations before the blow-pipe.

Before proceeding to work with soda glass, the student should not only verify by experiments what has been already said, but he should familiarise himself with the action of the blow-pipe flame on lead glass by trying the glass in every part of the flame, varying the proportions of gas and air in every way, repeating, and repeating, his experiments until he can obtain any desired effect with certainty and promptitude. He should practice some of the simpler operations given in Chapter III. in order to impress what he has learned well on his mind.

Management of Soda Glass.—In working with soda glass the following points must be constantly kept in mind. That as it is much more apt than lead glass to crack when suddenly heated, great caution must be exercised in bringing it into the flame; and that in making large joints or in making two joints near each other, all parts of the tube adjacent to that which, for the moment, is being heated, must be kept hot, as it is very apt to crack when adjacent parts are unequally heated. This may be effected by stopping work at short intervals and warming the cooler parts of the tube, or by the use of the brick or block of wood to check radiation, or even by placing a supplementary blow-pipe or Bunsen burner in such a position that its flame plays upon the more distant parts of the work, not coming sufficiently into contact to soften the glass, however, but near enough to keep it well heated. Lastly, to prevent the finished work from falling to pieces after or during cooling, the directions given under the head of annealing must be carefully carried out.

In very much of his work the glass-blower is guided more by the feel of the glass than by what he sees. The power of feeling glass can only be acquired by practice, and after a certain amount of preliminary failure. As a rule I have observed that beginners are apt to raise their glass to a higher temperature than is necessary, and that they employ larger flames than are wanted. If glass be made too soft it may fall so completely out of shape as to become unworkable except in very skilful hands. The following rules, therefore, should be strictly adhered to. Always employ in the first instance the smallest flame that is likely to do the work required. In operations involving blowing out viscous glass, attempt to blow the glass at low temperatures before higher ones are tried. After a little experience the adoption of the right-sized flame for a given purpose, and the perception of the best condition of glass for blowing it, become almost automatic.

I may add that glass which is to be bent needs to be much less heated than glass which is to be blown.

Annealing.—If apparatus, the glass of which is very thin and of uniform substance, be heated, on removal from the source of heat it will cool equally throughout, and therefore may often be heated and cooled without any special precautions. If the glass be thick, and especially if it be of unequal thickness in various parts, the thinner portions will cool more quickly than those which are more massive; this will result in the production of tension between the thicker and thinner parts in consequence of inequality in the rates of contraction, and fractures will occur either spontaneously or upon any sudden shock. Thus, if a hot tube be touched with cold or wet iron, or slightly scratched with a cold file, the inequality of the rate of cooling is great, and it breaks at once. It is therefore necessary to secure that hot glass shall cool as regularly as possible. And this is particularly important in the case of