"The generator A consists of a cylindrical vessel with sloping bottom, provided with a sludge outlet a, a gas exit-pipe b, and a lid b' fastened by screws. In the upper part ten boxes c are installed for the purpose of receiving the carbide. The bottoms of those boxes are flaps which rest through their wire projections on a revolvable disc d, which is mounted on a shaft l. This shaft passes through a stuffing-box to the outside of the generator and can be rotated by moans of the chains f, the pulleys g and h, and the winch i. Its rotation causes rotation of the disc d. The disc d, on which the bottoms of the carbide- holders are supported, is provided with a slot e. On rotating the disc, on which the supporting wires of the bottoms of the carbide-holders rest, the slot is brought beneath these wires in succession; and the bottoms, being thus deprived of their support, drop down. It is possible in this way to effect the discharge of the several carbide-holders by gradual turning of the winch i.
"The washer B is provided with a thermometer m passing through a sound stuffing-box and extending into the water.
"The gasholder C is provided with a scale and pointer, which indicate how much gas there is in it. It is connected with the pressure-gauge n, and is further provided with a control thermometer o. The gas exit-pipe q can be shut off by a cock. There is a cock between the gasholder and the washer for isolating one from the other.
"The dimensions of the apparatus are such that each carbide-holder can contain readily about half a kilogramme (say l lb.) of carbide. The gasholder is of about 200 litres (say 7 cubic feet) capacity; and if the bell is 850 mm. (= 33-1/2 inches) high, and 550 mm. (= 21-1/2 inches) in diameter it will admit of the position being read off to within half a litre (say 0.02 cubic foot)."
The directions of the German Association for sampling a consignment of carbide packed in drums each containing 100 kilogrammes (say 2 cwt.) have already been given in the rules of that body. They differ somewhat from those issued by the British Association (vide ante), and have evidently been compiled with a view to the systematic and rapid sampling of larger consignments than are commonly dealt with in this country. Drawing a portion of the whole sample from every tenth drum is substantially the same as the British Association's regulations for cases of dispute, viz., to have one unopened drum (i.e., one or two cwt.) per ton of carbide placed at the analyst's disposal for sampling. Actually the mode of drawing a portion of the whole sample from every tenth vessel, or lot, where a large number is concerned, is one which would naturally be adopted by analysts accustomed to sampling any other products so packed or stored, and there in no reason why it should be departed from in the case of large consignments of carbide. For lots of less than ten drums, unless there is reason to suspect want of uniformity, it should usually suffice to draw the sample from one drum selected at random by the sampler. The analyst, or person who undertakes the sampling, must, however, exercise discretion as to the scheme of sampling to be followed, especially if want of uniformity of the several lots constituting the consignment in suspected. The size of the lumps constituting a sample will be referred to later.
The British Association's regulations lead to a sample weighing about 3 lb. being obtained from each drum. If only one drum is sampled, the quantity taken from each position may be increased with advantage so as to give a sample weighing about 10 lb., while if a large number of drums is sampled, the several samples should be well mixed, and the ordinary method of quartering and re-mixing followed until a representative portion weighing about 10 lb. remains.
A sample representative of the bulk of the consignment having been obtained, and hermetically sealed, the procedure of testing by means of the apparatus already described may be given from the German Association's directions:
"The first carbide receptacle is filled with 300 to 400 grammes (say 3/4 lb.) of any readily decomposable carbide, and is hung up in the apparatus in such a position with regard to the slot e on the disc d that it will be the first receptacle to be discharged when the winch i is turned. The tin or bottle containing the sample for analysis is then opened and weighed on a balance capable of weighing exactly to 1/2 gramme (say 10 grains). The carbide in it is then distributed quickly, and as far as may be equally, into the nine remaining carbide receptacles, which are then shut and hung up quickly in the generator. The lid b' is then screwed on the generator to close it, and the empty tin or bottle, from which the sample of carbide has been removed, is weighed.
"The contents of the first carbide receptacle are then discharged by turning the winch i. Their decomposition ensures on the one hand that the sealing water and the generating water are saturated with acetylene, and on the other hand that the dead space in the generator is brought under the pressure of the seal, so that troublesome corrections which would otherwise be entailed are avoided. After the carbide is completely decomposed, but not before two hours at least have elapsed, the cock p is shut, and the gasholder is run down to the zero mark by opening the cock q. The cock q is then shut, p is opened, and the analytical examination proper is begun by discharging the several carbide receptacles by turning the winch i. After the first receptacle has been discharged, five or ten minutes are allowed to elapse for the main evolution of gas to occur, and the cock p is then shut. Weights are added to the gasholder until the manometer n gives the zero reading; the position of the gasholder C is then read off, and readings of the barometer and of the thermometer o are made. The gasholder is then emptied down to the zero mark by closing the cock p and opening q. When this is done q is closed and p is opened, and the winch i is turned until the contents of the next carbide receptacle are discharged. This procedure is followed until the carbide from the last receptacle has been gasified; then, after waiting until all the carbide has been decomposed, but in any case not less than two hours, the position of the gasholder is read, and readings of the barometer and thermometer are again taken. The total of the values obtained represents the yield of gas from the sample examined."