unnecessary, with the arrangement as here represented, to attach more than the first four pieces, to and including the bisulphate tube G: for reasons which will be obvious without explanation. Some five or six cubic feet of gas transmitted from a preparatory train, will dry out these first four members sufficiently to allow of correct weighing. This will be accomplished by setting and keeping the apparatus at work during one night, say from 5 P. M. till 8 A. M., or fifteen hours, at the rate of about 0.3 or 0.4 foot per hour. The rate of flow is readily regulated by a few observations, of ten minutes each, of a meter at the tail of any train. A little practice enables the rate to be judged of closely enough, by the size of the flame burning from a small straight glass tube at the tail of the apparatus, as shown in Fig. 4.

The preparation for the final weighings is here the important part, and indeed the only troublesome part of this analysis. Fig. 3 represents, as aforesaid, the mode of attaching the preparatory train to Fig. 1, for this final operation, or rather succession of operations. D here represents a new tube introduced at this time, which is an ordinary' form of CaCl tube, but having its smallest extremity bent at an obtuse angle, as shown, and which contains, near its upper end, a slip of turmeric paper, and has, moreover, running through it from end to end, and projecting an inch at each end, a small flexible copper wire the object of which latter is to clear a passage through the crystals of naphthaline which condense in this tube at a subsequent stage of operations.*

After this disposition of the apparatus has been made -D having been of course weighed with its appendages; and the jar of ice N (which is no longer of use) removed, the meter being connected directly with M-the first step is to distil over, at a very low temperature (unless saving time is important, at ordinary temperature) in the current of gas from the preparatory train-which

*This wire should have shown more distinctly in the cut, but this is only one of several unfortunate defects in these cuts,

SUPPLEMENTARY ARRANGEMENT FOR COMPLETION OF PROCESS IN THE TRAIN
SHOWN BY FIG. 1.

generally in this case may flow even as rapidly as two feet per hour-all the volatile ammonic carbonate and sul-. pho-hydrate that has condensed in the flask C and cotton tubes E and F. The volume of gas required to effect this may be double (or even more) that from which the liquid was condensed. It ought to have been stated that the total amount of crude gas analyzed by this apparatus should not exceed ten cubic feet. This volume is within the limits of manageability, but generally somewhat less is probably advisable. With a rapid current of gas, this complete transfer of the volatile ammonia compounds will be found to have been effected

during fifteen or eighteen hours; but sometimes more time is required; and it is better to allow twenty-four hours for this part of the process. If an attempt to hasten it is made, by immersion of the flask in warm water, constant watching is necessary, as naphthaline then enters upon the scene, and may at once clog up all passage. The complete transfer of the ammonia is indicated by the return of the slip of turmeric paper to its natural color. As, in my experience so far, this liquid always contains chloride of ammonium, and possibly sulphite and other non volatile ammonia salts, those who wish to transfer all the ammonia to the bisulphate tube G, must introduce into the flask a weighed quantity of dry hydrate of soda.

When the ammonia transfer has been completed, the flask C is immersed-the gas-current being continuedin water sufficiently hot to volatilize the naphthaline, which will condense in crystalline form, together with water, in the tube D. The wire running through D now comes into play; and, in fact, without it, naphthaline would here become as great a cause of difficulty as it often does in the distribution of coal gas. When clogging occurs, the gas-current is stopped, the rubber connector between D and E detached, and the wire moved up and down or twisted round. With careful management, two hours' work and some two feet of gas from the preparatory train will suffice to transfer the naphthaline to D, which is then detached and weighed again. To remove the intermixed water from the naphthaline, the lower end of D is stopped, the cork and tube removed from the upper end, the wire coiled within, and this upper end connected by a wide short rubber connector with the open end of a test tube containing some granulated CaCl. When dry, the cork and tube restored, the stopper removed, and the weight again ascertained; the variation now from the first weight is the naphthaline, so far as it can be saved by

this mode; which of course is imperfect, but is the best hat I have so far succeeded in devising.*

The flask C, still containing tar and water, is then removed likewise from the arrangement-the stoppers DD (Fig. 1.) being applied-and then H of the preparatory train is connected directly with the first cotton tube E; the gas-flow being resumed. Some five feet more of dry gas may be now required to transfer all the moisture to the CaCl tube between G and H (omitted from the cut). The water and tar in the flask C must be separated by careful distillation at a very moderate temperature, and in estimating the weight of the residual tar, account must be taken of the sodic hydrate that was added, as above, to eliminate the fixed ammonia.

The point at which the whole train becomes dry and ready for the final weighings, is readily perceived by a practised eye. One indication is a faint superficial whitening of some unaltered blue granules (if any) of cupric sulphate in the tube HJ. Several other indications might also be cited.

* Since the above was written, I have discovered that a very appreciable amount of naphthaline condenses, and remains throughout the final desiccation, in the cotton tubes E and F, and has therefore been rated as solid matter mechanically suspended-W.