These caps can be made of different materials. It is most advisable to have them made of copper or brass; those which are formed of tin plate (white iron), and which are commonly used in chemical laboratories, are the worst kind of all. They soon become covered with grease or soot, which either completely closes up the orifices, or, at least, very soon alters the circular form which is necessary to the production of a good fire. Glass caps are less liable to get dirty, and are much cheaper than the above; but, on the other hand, they have the disadvantage of being easily melted. This can to a certain extent be remedied by making the points of very thick glass, and by always keeping them at some distance from the flame. Moreover, as you can make them yourself when you are at leisure, their use is very commodious. If they are to be used with the blowpipe described in this work, they must be fixed in the cork that closes the passage through which the current of air arrives. _C c_ and _C´ c_ (pl. 1, fig. 19) are two glass beaks, _c c_ are the corks, which can indifferently be adapted to _c_, in the wooden vice, by which the various parts of the blowpipe are connected when it is in action. Of whatever material the beak may be made, its orifice must be perfectly round, and the _size_ of the orifice, as we have before observed, must have a relation to the size of the wick which is to be used with it. You can ascertain the diameters of the orifices by inserting into them a little plate of brass, having the form of a long isoceles triangle, such as is represented by pl. 1, fig. 2. It should be an inch long, the twelfth of an inch wide at one end, and diminish to nothing at the other. When divided into eight equal parts, it will give, at the divisions, the respective proportions of 1, 2, 3, 4, 5, 6, 7 _eighths_ of the diameter at the wide end, as is exemplified by the figure above referred to. We have stated in the following table the relative diameters which long experience has recommended to us, as being adapted to produce the greatest effect; yet it is not to be imagined that these proportions are mathematically correct and indispensable for the obtaining of good results. A sensible difference of effect would be perceived, however, were these proportions departed from in a notable manner.