For this operation time must be allowed if the operator's hands are steady enough to permit of it; the shoulder should form partly by the glass sinking in and partly by the process of drawing the hot glass out. A shoulder properly prepared is shown in the sketch. Beginners generally make the neck too thin on large tubes, and too thick on smaller ones. There ought to be no great difference in thickness of glass between the neck on the larger tube, and the smaller tube. The diameters should be as nearly as possible alike. Having drawn down the larger tube to a neck, take it out of the flame, and as it cools pull and turn till the neck is of the right thickness and is perfectly straight, i.e. make the final adjustment outside the flame, and to that end have the neck rather too thick (as to glass) before it is taken out. It is not necessary to wait till the neck gets cold before the end can be cut off. Make a scratch as before--this will probably slightly damage the temper of the file knife, but that must be put up with. Hold the tube against the edge of the table, so that the scratch is just above the level of the rim, and strike the upper part a smart blow with the handle of the glass knife rather in the direction of its length. [Footnote: A bit of hoop iron nailed against the side of the table is a very convenient arrangement, and it need not project appreciably above the general level of the rim.] Of course this applies to a tube where economy has been exercised and the end is short. If the tail is long enough to form a handle, the tube may be pulled apart as before. As a rule a temporary joint between a tube and a rod is not strong enough to enable the shoulder to be broken at the scratch by mere pulling. The ends to be welded must be broken off very clean and true. Subsequent operations are to be carried out as already described. Sec. 23. The above operations will be easily performed on tubes up to half an inch in diameter, if they are not too long. It is the length of tube, and consequent difficulty in giving identity of motion with the two hands, which make the jointing of long tubes difficult. There are also difficulties if the tubes are very thin, have a very fine bore or a very large diameter. All these difficulties merely amuse a good glass-blower, but to an experimenter who wants to get on to other things before sufficient skill is acquired (in the movement of the hands and arms) the following method is recommended. First, use flint glass. Then, assuming that any drawing down has to be done, do it as well as possible, for on this the success of the method to be described especially depends. Be sure that the tubes to be welded are cut off clean and are as nearly as may be of the same size at the point of junction.