Moulding by Pressure of the Air.—­The plaster mould having been put in place upon the mould board, and the liquid paste having been long and thoroughly stirred in order to make it homogeneous, and get rid of the air bubbles, we open the cock that puts the paste reservoir in communication with the lower part of the mould, care having been taken beforehand to pour a few pints of water into the bottom of the mould.  The paste in ascending pushes this water ahead of it, and this slightly wets the plaster and makes the paste rise regularly.  When the mould is entirely filled, the paste is still allowed to flow until it slightly exceeds the upper level, and, spreading out over the entire thickness of the plaster, forms a sort of thick flange.  The absorption of the liquid begins almost immediately, and, consequently, the level lowers.  A new quantity of paste is introduced, and we continue thus, in regulating its flow so as to keep the mould always full.  This operation is prolonged until the layer is judged to be sufficiently thick, this depending upon the dimensions, form, or construction of the vessel.  The operation may take from one to five hours.

The desired thickness having been obtained, it becomes a question of allowing the paste to descend and at the same time to support the piece by air pressure.  The flange spoken of above is quickly cut, and the paste is made to rise again for the last time, in order to form a new flange, but one that this time will be extremely thin; then a perforated disk designed for forming the top joint, and acting as a conduit for the air, is placed upon the mould.  This disk is fastened down with a screw press, and when the apparatus is thus arranged the eduction cock is opened, and the air pump maneuvered.

If the flange did not exist, the air would enter between the mould and the piece at the first strokes of the piston, and the piece would be inevitably broken.  Its object, then, is to form a hermetical joint, although it must at the same time present but a slight resistance, since, as soon as the liquid paste has flowed out, the piece begins to shrink, and it is necessary that at the first movement downward it shall be able to disengage itself, since it would otherwise crack.

As soon as the piece begins to detach itself from the mould the air enters the apparatus, and the pressure gauge connected with the air pump begins to lower.  It is then necessary, without a moment’s loss of time, to remove the screw press, the disk, and the upper part of the mould itself, in order to facilitate as much as possible the contraction of the piece.  Finally, an hour or an hour and a half later, it is necessary to remove the lower part of the mould, this being done in supporting the entire affair by the middle.  The piece and what remains of the mould are, in reality, suspended in the air.  All these preparations are designed to prevent cracking.

Moulding by Vacuum.—­The operation by vacuum follows the same phases as those just described.  It is well, in order to have a very even surface, not to form a vacuum until about three hours after the paste has been made to ascend.  Without such a precaution the imperfections in the mould will be shown on the surface of the object by undulations that are irremediable.