tinguished chemists referred to, were not adapted to an accurate determination of the absorbent power of liquids. To be capable of precise results, the absorbing apparatus must fulfill the following conditions:

First. It must provide means for maintaining the temperature uniform throughout the experiment.

Second. It must maintain the tension of the gas unaltered.

Third. It must afford means for rapid and continuous agitation of the liquid with the gas.

Fourth. The tube in which the absorption is measured by the mercurial column, must be apart from the vessel in which the absorption occurs, and the mercury must not be introduced into that vessel.

In view of these requisites we were led, after many unsatisfactory trials with other arrangements, to the form of apparatus represented in the accompanying diagram, (see next page,) which, besides greatly expediting the experiments, affords uniform and consistent results seldom varying in successive trials, to the extent of one per cent., and which is equally applicable to all liquids.

Absorption Apparatus.-This consists of a gasometer A, plunged in a large wooden reservoir B, containing water to the level indicated in the figure, adjoining which is a smaller but taller reservoir C, of glass, also containing water. In the latter is immersed, in a fixed vertical position from the strong frame above, a syphon-shaped measuring tube with a finely graduated scale between the limbs. A horizontal arm of thick barometer tube extending from the top of this, is united by a short gumelastic joint, with a similar tube which bends down over the edge of the frame and is inserted below into the actual opening of the absorption flask D. Cylinder thermometers graduated to tenths of a degree are placed in the gasometer, large and small reservoir, and flask. [Figure 2 is a larger view of the flask.]

The main reservoir, charged as indicated in the figure, contains five thousand six hundred cubic inches of water, the smaller one, of glass, six hundred cubic inches, and the gasometer three hundred cubic inches. The large volume of water in the reservoirs, serves to maintain an almost absolute uniformity of temperature in the flask and measuring tube during the experiment. The capacity of the flask usually employed by us, is 6-2 cubic inches. The measuring tube is read to th of a cubic inch.

A long winding leaden tube serves to conduct the gas from the gasometer to the flask in the process of charging the latter, and being plunged in the water of the large reservoir, secures us against any variation of temperature in the gas, which might arise from the reaction of the materials in the gasometer. The gas will thus in all cases have the temperature of the main res

ervoir, when conveyed into the flask. The end of this tube is mounted with a close cork and smaller glass tube, to be placed in the near opening of the flask when the gas is to be passed into the latter and the measuring tube.

Connected with the flask is a movable vessel r, adapted by accurate grinding to the mouth of the former, and designed to contain the liquid whose absorbent power is to be determined. This unit bottle as we will call it, has its opening below contracted to about one-eighth of an inch, so that when filled with the liquid and connected in an inverted position with the flask, as in the diagram, the liquid is in no danger of flowing out. When thus inserted in its proper position, it is confined in place by the strong pressure of a steel spring, attached to the central tube of the flask, and which by a revolving motion is then brought to press by a leather cushion upon its upper end. This secures the juncture at the stopper perfectly air-tight during the active agitation of the flask.

The central tube is in like manner ground to an air-tight joint at its insertion into the flask, and the connection is forcibly secured by a strong cord wrapped around the neck and then passed around a screw peg inserted in a wooden block which is firmly cemented to the tube above. To avoid all chance of leakage at this and the junction of the unit bottle with the flask, a very delicate coating of tenacious cement made of beeswax, rosin and tallow, is applied to the upper part of the ground surface. This from its very minute amount, and its removal from the gas in the interior, is incapable of exercising any appreciable absorbent effect. Indeed we have found that even with a large mass of this cement placed in the flask, the effect is quite insignificant.

The flask suspended by the horizontal part of the tube just described, hangs, when in its natural vertical position, at such depth as to immerse about half the length of the unit bottle. The agitation is given by a hooked rod which embracing the central neck, is moved to and fro longitudinally, and causes the flask to swing as rapidly as the operator pleases, the axis of motion being the part of the horizontal tube which is external to the flexible joint, and lies upon the wooden frame near its end. The rest of the horizontal tube is connected with the measuring tube, and firmly attached, along with the latter, to the horizontal bar of the frame. This slides up and down upon the vertical supports, and can be adjusted to a proper position by the movable pins seen in the figure, immediately beneath it, or can be lifted off, carrying the measuring tube and flask. To secure the axis tube from lateral motion during the shaking, a wooden block descends by a hinge over the flexible joint and this tube, and embracing them in a longitudinal groove retains them in place.

The stop-cock at the bottom of the measuring tube is used for adjusting the level of the mercury in the two limbs at the beginning of each experiment, and for removing the mercury which is poured in through the funnel above, to maintain the columns at the same height during the progress of the absorption.

It will be seen from the figure, that while the flask and measuring tube are both constantly immersed as the experiment is going on, and are thus kept at an invariable temperature, the connecting tube, between the two levels of water, is necessarily out of the liquid, and must be influenced by the temperature of the ambient air. The capacity of this exposed part of the tube was found to be very nearly 7ths of a cubic inch, the expansion of which for 10 is equal to cubic inch. As the temperature of the apartment seldom differed from that of the apparatus by more than some four or five degrees during our experiments, the entire error would be within th of an inch, while as before mentioned, the smallest reliable reading of the measuring scale isth cubic inch. It has therefore been thought useless to attempt any correction for the temperature of this part of the enclosed gas.

Mode of Manipulating.-From the description just given of the several parts of the apparatus, the general method of operating with it will be readily inferred, and but few words need be added on this head.

Bringing the entire apparatus to the required temperature, (60° in most of our experiments,) the unit bottle, charged with the liquid to be used, is hung in the large reservoir, to attain exactly the same temperature. The back of the gasometer pipe is inserted in the flask air-tight, and a brisk stream of carbonic acid is suffered to flow through the apparatus for five minutes. Some mercury is now poured by a long funnel into the measuring tube to arrest the current. The flask being raised so as to lay bare the mouth, the cork is withdrawn, and at the same moment, while the stream of gas is pouring out and overflowing from the flask, the unit bottle is secured in the opening and fastened by the spring above. After swinging the flask down to its vertical position, the level of the mercury in the measuring tube is carefully adjusted, and the agitation is now commenced. The liquid at first descends only by drops, but soon begins to flow more rapidly. The vibratory movement of the flask is of that sudden kind which effectually brings the gas and liquid into intricate contact and the absorption rapidly proceeds. Two operators are necessary in conducting the experiment, one to keep up the shaking, and the other to supply the outer limb of the measuring tube with mercury as the column on the other side ascends.

With water, we have found the absorption to be completed in about five minutes. The oils, dense saline solutions and sul

phuric acid, require a longer time, but even with the last named substance which is one of the most sluggish in its action, the absorption reaches its limit in less than thirty minutes.

Purity of the Carbonic Acid.-The gas used in our experiments was supplied by the reaction of dilute hydrochloric acid and fragments of calc spar, contained in the self-regulating apparatus figured in the preceding diagram. For sometime, after charging the vessel with water and acid, the gas evolved contains a marked proportion of atmospheric air, derived from the air originally present in the water, and which is slowly disengaged as the carbonic acid is absorbed by the liquid. This admixture with air was found to continue until the solution became well charged with the gas, and this result, in the ordinary use of the apparatus, was very slowly attained. To hasten the saturation, and thus bring the materials into a condition to furnish unmixed gas, the action of the acid liquid on the carbonate was renewed at short intervals, by opening the stop-cock of the reservoir, and in this way in a few hours the gas evolved was almost absolutely exempt from atmospheric air.

At the commencement of each set of experiments, a specimen of the gas, two cubic inches, was passed into a tube over mercury and tested by a moist fragment of caustic potash. When the contents of the reservoir were in a proper condition, the residuum of unabsorbed gas in this experiment was a mere globule, rarely more thanth of an inch in diameter, and therefore indicating from th toth of gaseous impurity.

1

Thus assured of the almost total absence of atmospheric air in the gas supplied under these conditions, our next precaution was to determine the degree of purity it retained, when transferred, as in our experiments, by simple displacement into the flask and measuring tube. For this purpose a V shaped tube, eighteen inches in each leg, with a stop-cock at the bend, was attached temporarily to a flask like that of the absorbing apparatus, and was charged by allowing a stream of CO, to pass through the vessel and tube steadily for five minutes. The stop-cock of the tube was then closed, the open end stopped with the finger, and the tube detached and inverted over mercury. The contents were now examined in the usual way with caustic potassa. In a number of such trials, made at different stages of our investigation, we found the amount of residual gas to range from about

ath to ath of the entire volume employed. In the absorption apparatus, the charge of gas is probably less, and certainly not more contaminated with atmospheric air than in the trials just mentioned, and can therefore involve no sensible error from this

source.

It remains to ascertain how far the CO,, escaping from the gasometer, might be mingled with hydrochloric acid. The pres