of tube at widest part, 1 inch, tapering in 4 inches to inch; length of tube from junction with the scrubber to top of bend, 6 inches.

University of

The scrubbers consist of two U-tubes of the form shown, fitted with draining tubes closed by clips, and filled with glass beads. The beads used were in the form of 3-inch lengths of quill tubing, and were easy to wash. A few large round beads were placed at the top, to keep the others in place. The tube (D) is fitted into the scrubber by means of a ring of india-rubber, and the corks are of the same substance. The dimensions of the U-tubes are: diameter of tube, 1 inch; length of short limb, 10 inches; length of inclined tube, 4 inches; width across, 4 inches.

K is a small bottle containing solution of silver nitrate, and serves to indicate the completeness of the absorption, and the rate of the current. It is connected with a Sprengel pump (water). A small weighing tube is required, of such a size that it will easily slide down the burner into the bulb. It must be stoppered, but not contracted at the mouth, otherwise its contents will not readily flow out into the bulb.

The mode of conducting an experiment is the following: the scrubbers are charged with 50 c.c. each of a solution of pure caustic soda, prepared by dissolving 1.5 grams of sodium in 100 c.c. of water, and a current of air is drawn through the apparatus, just sufficient to allow of the Bunsen flame burning steadily within the conical tube. A disc of card may be placed round the burner, so as nearly to close the mouth of the cone, to neutralise the effects of draughts. The liquid to be analysed is then weighed out in the small stoppered tube (0.1-0.25 gram), and the gas having been previously turned on for some seconds, the weighing tube is quickly unstoppered, and dropped down the burner, the gas instantly lit, and the flame placed within the cone. If the liquid has not already run out into the bulb, it is made

to do so, and its evaporation is then aided by warming the bulb with hot water, or otherwise. The temperature of the bulb is raised or lowered, so that the presence of the substance may always be detected in the flame, and yet not in sufficient quantity to cause the flame to become luminous. With some bodies the flame is tinged green; with others, e.g., amyl chloride, a slight luminosity at the tip of the flame is all that is perceptible. The success of the experiment depends upon the regular volatilisation of the compound; on the one hand this must not be unduly hurried, or the flame becomes luminous, and the combustion may be incomplete; on the other, the evaporation should not be too slow, or, before it is complete, the soda in the absorbers will have become carbonated, and therefore incapable of stopping every trace of the free halogen and their acids. The silver nitrate solution will of course indicate when this point has been reached.

Usually the combustion occupies 15-20 minutes, and the silver solution remains clear. As soon as the liquid has disappeared from the bulb and tube, the former is warmed gently with a Bunsen burner, and then more strongly, so as to drive off the last traces.

We have found that the easiest way of securing a slow and regular evaporation in the case of liquids boiling below 100° C., is to place the stoppered weighing tube with its stopper within the bulb; on warming with hot water, sufficient vapour escapes round the stopper. This plan was adopted with good results in experiments made with carbon. tetrachloride, 76°; ethyl iodide, 71°; ethyl bromide, 39°; and ethylene chloride, 82°. In the case of ethyl iodide, indeed, it is impossible to use an open tube, for the iodide is so little inflammable and so volatile that the gases refuse to burn at all.

Liquids with boiling points between 100° and 135° are weighed out in a tube from which the stopper is removed on dropping it into the burner. The most convenient way of heating the bulb is to surround it with water kept hot by a current of steam.

With liquids boiling above 135° cautious heating with a Bunsen may be resorted to; the flame must be kept moving, and taken away directly the substance shows its presence in the flame. When the combustion is at an end, the soda-solution is run out from the U-tubes into a flask of about a litre capacity, and the wash-water from the same, together with the rinsings of the corks and connecting tubes, added. Generally the total quantity of liquid amounts to about half a litre. Some sulphurous acid is now added, and the contents of the flask boiled down for 10 minutes or more over a large Bunsen, so as to reduce any chlorates, &c., that may have been formed. Nitric acid is then added, the solution precipitated with silver nitrate, and the estimation finished in the usual way. The precipitate may easily be washed from the flask on to the filter.

The following are some of the results obtained by the above method :

Three analyses of an impure specimen of benzal chloride yielded, 43.3, 43-6, 43.3 per cent. chlorine; a determination by Carius' method gave 43.6 per cent. The above experiments, made with compounds of very various composition and boiling points, will suffice to show that the method described is capable of yielding accurate results. The time occupied by a complete estimation is about two and a half hours.

The experiments here described were made in the Laboratory of University College, London.

XVI.-A Modified Liebig's Condenser.

By W. A. SHENSTONE, Lecturer on Chemistry at Clifton College.

FOR effective working the original form of Liebig's condenser leaves nothing to be desired. It has, however, the disadvantage of occupying a great deal of space, and when used as part of an arrangement for digesting substances which are subsequently to be distilled, as in the preparation of formic acid by heating oxalic acid with glycerol, or in the preparation of ethyl iodide by digesting together red phosphorus, alcohol and iodine, there is a certain amount of trouble and loss of time in arranging it for distilling after the digestion is completed. These disadvantages, and especially the former, as I frequently want to have several operations in which condensers are employed simultaneously on a small lecture table, led me to try the modification I now describe. It is, as I expected, superior to the original form in taking

up but little space, and in its ready adaptability to the purposes of digesting and distilling without rearrangement; and the only objection I find to it is that, as the distillate is delivered rather near to the source of heat, it would be somewhat dangerous to employ it in distilling highly volatile and inflammable substances such as ether and benzene.

A tube A, which may be of any convenient size, is fitted in the ordinary way inside a larger glass tube, for which purpose I think the old-fashioned method of employing corks answers better than the modern plan, in which the outer tube is contracted at its ends and connected with the smaller tube by pieces of india-rubber tube. At B the tube A has a small projecting tube F, about 10 mm. in length. Into the side of A, so as to be below the level of the open end of F, a side tube C, having a slight dip at c', is joined on, its open end D being about 5 mm. below the level of the other end. A small stopper is ground into C at D. D may conveniently be 2 mm. to 3 mm. in diameter.

The end E can be attached to flasks by means of corks, or as might occasionally be desirable, can be ground, to fit the neck of a flask.

When the condenser is in action, the vapour passes up through F, is condensed in A, and falls into the annular space around F at B. If the stopper is fixed at D, this space soon fills, and then, as further condensation occurs, the products overflow into the flask through F. The amount that collects at B with a well-made tube is very small, as the

air in C usually prevents the liquid from flowing into it when the stopper is fixed. When the apparatus is to be used for distilling, the stopper D' is removed, and the distillate then flows out at D. The bend at c' was made in order that the first portions condensed might flow there, and so prevent any vapour from escaping by C at the early part of the operation; from what I see of the action of the tube, I think, however, that it is not necessary.

The outer tube should be brought down as near to the joint as possible for the vapour coming up through F slightly warms the liquid at B, so that it is delivered at a slightly higher temperature than that of the water in the condensing jacket, and it is desirable to avoid this as far as possible.

When the temperature of the vapour must be known, I hang a thermometer inside E by means of a little hook of platinum wire. The length of the tube therefore from B to E should be such that there shall be room for an ordinary thermometer between the point B and the level of the liquid in the flask below. It is perhaps an advantage that in thus using this condenser, the actual temperature of the vapour is observed, as the stem of the thermometer is entirely immersed in it.

Mr. Cetti, the maker of the tube from which this is drawn, tells me that for fractional distillation, if, as might be the case, the form of tube proposed by Dr. Armstrong were employed, there would be no difficulty in grinding the end E to fit the upper end of the fractionating tube.

I have found this form of condenser so compact for moving about, and so easy and convenient to work with, especially for lecture purposes, that, as I believe such a modification has not been described before, I have thought it might be worth while to bring it under the notice of the Society.

XVII.-On Some Fluorine Compounds of Uranium.

By ARTHUR SMITHELLS, B.Sc. (Dalton Scholar in the Laboratory of The Owens College).

I. SINCE the preparation of uranium oxyfluoride by Berzelius, the fluorine compounds of uranium have been investigated by Carrington Bolton (Zeitschr. f. Chem. [2], 353; Bull. Soc. Chim., 1866, 2, 450) and more recently by Ditte (Compt. rend., 91, 115). As the results obtained by these two chemists do not agree, the present experiments