which should be dilute, with the above reagent, ab-phenacetylparatolylurea was obtained. When recrystallised from alcohol, in which it is rather sparingly soluble even at the boiling point, it formed long, hair-like, flexible, white needles, insoluble in water, nearly so in cold spirit, and melting at 189-189-5° (corr.).

Analytical data :—

0.2024 gave 184 c.c. at 12° and 759 mm. N = 10-78. CHIN2O2 requires N 10-47 per cent.

n-Phenacetyl-v-phenylbenzylthiourea, C.HCH,CON:C(SH)•N(CH)CH, CH.

Prepared as before, using alcoholic benzylaniline; the resultant clear, pale brown liquid after two days' standing showed no sign of crystallisation, but, on pouring out from the containing flask into a dish, it at once solidified to a mass of fine needles; the yield was almost quantitative. By recrystallisation from hot benzene, microscopic prisms were obtained, possessing, when seen in mass, a faint lemon-yellow tinge, and melting at 127-5-128-5° (corr.) with decomposition.

0.2014 gave 0.1348 BaSO4. S = 9.20.

C22H20N2SO requires S = 8.90.

The compound is insoluble in water, rather sparingly soluble in ether or boiling alcohol, much less so in cold, easily in hot benzene or acetone, very freely in chloroform. Its alcoholic solution is almost neutral; if mixed with silver nitrate, a creamy-white precipitate forms, and the liquid becomes sharply acid; this precipitate (or that produced by the addition of ammoniacal silver nitrate) is not blackened by boiling, neither is the thiourea desulphurised by boiling with alkaline lead tartrate.

Phenacetylthiocarbimide and Ammonia.

The former, in benzene, was mixed with a considerable excess of alcoholic ammonia; much heat was evolved, and, on cooling, large, ice-like, colourless plates were deposited. They were free from sulphur, moderately easily soluble in hot water, sparingly in cold, more freely in alcohol, and melted at 156-157°. The warm solution readily dissolved freshly precipitated yellow oxide of mercury, and on cooling a white solid was deposited, melting at 207.5-208-5°. Consequently, the product was not the expected phenacetylthiourea, but phenylacetamide, formed, probably, in the following manner CH,Ph•CO NCS + 2NH = CH,Ph·CONH, + CSN.H..

=

*CH,Ph-CO-NH2 melts at 158°; its HgO compound (Reimer, Ber., 1880, 13, 741) at 208°.

This action is by no means exceptional; thus Miquel (Bull., 25, 104), by passing dry ammonia into an ethereal solution of acetylthiocarbimide, obtained, instead of acetylthiourea, an oil (which he thought might possibly be acetamide thiocyanate); moreover, succinyldithiocarbimide, when treated with alcoholic ammonia, yielded (Trans., 1895, 67, 573) succinamide; and propionylthiocarbimide (ante), under similar conditions, gave a brownish oil, from which no thiourea could be isolated.

Other Acidic Thiocarbimides.-Several attempts were made to obtain from picryl chloride the corresponding picrylthiocarbimide, C&H2(NO2)3 NCS. Both lead and mercuric thiocyanates were employed, and as solvents, benzene, toluene, or cumene; but, under no circumstances, was there any indication of the desired union.

Experiments with the chlorides of phenylsulphonic and ethylsulphuric acids are now in progress, the results of which, if successful, will be communicated later on.

In conclusion, I wish to express my thanks to Mr. R. E. Doran for the help which he has given with a portion of the work above recorded.

Chemical Department,

Queen's College, Cork.

LVIII.-Carbon Dioxide. Its Volumetric Determina

tion.

By WILLIAM H. SYMONS, D.P.H. (Oxon), and F. R. STEPHENS. CARBON dioxide appears to have been first recognised as a regular constituent of the atmosphere by MacBride in 1764 (Experimental Essays), and he estimated the amount gravimetrically after drawing a measured volume of air over caustic alkali, a process which, with some modifications, still remains the standard. It was not, however, until Pettenkofer described his well known volumetric process for its estimation before this Society, in 1857, that its measurement as an index of the respiratory impurity of enclosed spaces became practically available. Since then many chemists have been attracted by the subject, and, in looking over the Chemical Society's Journal, we find over 120 papers or abstracts of papers bearing on this subject; and among the contributors are many illustrious chemists. As a rule, some modification of Pettenkofer's process has been used for estimating the carbon dioxide in the air of rooms, and the older gravimetric method for the air of open spaces; but, among special processes, the following may be mentioned. Wanklyn suggested a

"nephalometric method," imitating a cloud of barium carbonate; Dupré and Hake used a similar process, but with oxyacetate of lead; Winter Blythe absorbs by sodium hydroxide, and then estimates by Schrotter's carbon dioxide apparatus; Angus Smith used manganate of sodium, and observed the change to permanganate. He also suggested as a household test his minimetric method, which depends on the formation of a cloud in a given volume of lime water. It is, however, very difficult to exclude all adventitious carbonates, and this considerably detracts from the value of any test in which the absence of carbonates in the reagents is assumed. Kapussten (J. Chem. Soc. Abstr., 1880, 420) used alcoholic soda solution, and noted the amount of water required to dissolve the carbonate.

In 1884, Blockman (Ber., 1884, 17, 1017, 1019) suggested phenolphthalein as an indicator, lime water as an alkali, and a 500 c.c. flask in place of the larger vessel of Pettenkofer. He showed that the test could be made either by adding the solution to a measured quantity of air until decolorised, or by using a fixed quantity of alkaline solution, re-filling the flasks with air, and agitating as often as necessary to decolorise the indicator; he found the latter method the better. Rosolic acid was suggested as an indicator by Pettenkofer in 1876. Ballo (Ber., 1884, 17, 1097, 1101) stated that lime water was sluggish in its action, and he recommended a mixture of potassium hydroxide and barium chloride, and at least three minutes' shaking. Since then, Lunge and Zeckendorf have introduced a method which is in principle identical with Blockman's, but they use a small vessel (110 c.c.), the air being drawn in by an indiarubber hand pump, supposed to remove 70 c.c. of air from the vessel after each compression; and a N/500 Na2CO, solation with phenolphthaleïn. Wolpert has patented a still smaller apparatus on nearly the same principle, and works on quantities of air varying from 10 to 50 c.c. It is obvious that such methods must be open to great error, and that the results obtained by their means have very little scientific value; but we are of opinion that Wolpert's form of apparatus is useful for making rough estimations.

To return to Pettenkofer, it is unfortunate that most of our text books describe as his a process which lacks many of the safeguards which he adopted; for example, in his later process he added a small quantity of barium chloride to the barium hydroxide to counteract the influence of small quantities of alkalis which might be present; he also used a vessel closed by caoutchouc and titrated direct through a small tube, letting down turmeric paper from time to time by means of a wire through another tube; whereas most of the text books direct the titration to be done in an open vessel, with part of the contents, either before or after filtration, and do not mention the

need of precautions against contamination from the operator, who is all the time giving out a stream of air containing at least 4 per cent. of carbon dioxide. Pettenkofer used a dry bottle, and drove the air into it by means of a pair of bellows, the bellows being furnished with a bent tube, which was pointed in the direction from which it was desired to sample the air; but as the length of the tube was not, from his illustration, apparently more than 6 inches, a turn in one direction or another could not have much influence.

Angus Smith suggested that the air should be drawn from the bottle by means of bellows, this is undoubtedly an improvement; but in either case, it is a matter of considerable labour to insure a complete exchange of air. Lebedingebb (J. Chem. Soc. Abstr., 1891, 290) found it necessary to aspirate 200 litres through a bottle holding 7-12 litres, whereas Dr. Gill, of the Laboratory of Sanitary Chemistry and Gas Analysis, Mass. Inst., U.S.A., considers it sufficient to fill the bottle four times by 15 strokes of his bellows. Dr. Louis Parkes, in his Manual of Hygiene, second edition, p. 260, recommends a glass stoppered vessel of 2 litres capacity to be filled with mercury and emptied in the place, the air of which is to be examined; but 27 kilograms of mercury is certainly an inconvenient quantity to deal with, and he mentions the alternative method of forcing air in by means of bellows. Dr. Notter, who contributes the article on air in Stevenson and Murphy's standard work on hygiene, says: "Perhaps the best plan is to fill a jar with clean water and empty it in that part of the air space it is desired to examine, taking care to allow it to drain." The capacity of the vessel is to be about 1 gallon; obviously no amount of draining will remove all the water, yet no allowance is made for this in the subsequent titration of half the volume of alkali used for absorption; the outflowing water, particularly the portion which slowly drains out, will absorb some carbon dioxide; and tap water, which presumably is used, always contains that gas. It is only fair to add, that he concludes his remarks with the following sentence: "Although the method for determining carbon dioxide in air does not give quite accurate results, this is the most convenient for ordinary use, and sufficiently accurate for all practical purposes." A process conducted in this way would be open to an error of about 10 per cent., and if such inaccuracy as that will satisfy, we should prefer to use some such apparatus as that suggested by Lunge and Zeckendorf, or Wolpert, rather than one which needs a cab or other vehicle for its transmission. But recent legal proceedings have shown that the estimation of carbon dioxide in air may have an important bearing as evidence of over crowding, and we have taken the liberty of quoting from some of the most prominent text books on hygiene, in order

to show the necessity for a simple accurate process such as we propose.

We use for the collection of the samples of air flasks containing only water vapour, and varying in capacity from 0.5 to 3 litres, according to the nature of the air. Vacuous flasks have been used by others, and their advantages are obvious; they will remain vacuous for months with good rubber tube and efficient clamps or stoppers; they may therefore be kept ready for use, and are extremely portable; a couple of litre flasks may be easily carried in the coat pockets, and a 3-litre flask concealed under the cape of a mackintosh or great coat. The samples of air may be taken quite unobserved; we have taken them in a police-court while sitting in the barristers' enclosure, within a few feet of the magistrate. The position from which the air is taken may be selected exactly, while the time occupied in taking the sample is only a few seconds.

We have tried various kinds of pumps for exhausting flasks, and have found it much simpler and quicker to fill the vessel with steam under pressure, and then, after closing, allow the steam to condense; the small quantity of water thus introduced does not interfere with subsequent operations. A common steam steriliser makes a good boiler, and is now found in most laboratories. Distilled water should be used, as tap water gives up its "loosely combined carbon dioxide" very slowly. The flask we use is drawn out to a neck some inch internal diameter, but it is obvious that any round-bottom flask, furnished with a caoutchouc stopper perforated by a glass tube, would answer the same purpose, and in our earlier experiments such a flask was used. The neck is continued outwards by a short caoutchouc tube in such a manner that it can be controlled by a strong screw clamp or plugged by a stopper. To remove the air, the inverted flask is placed vertically over a 4-inch copper tube, connected with the boiler, and sufficiently long to reach nearly up to the bottom of the flask. The steam, at about 20 lbs. pressure, is sent through, and the flask kept in position for about two minutes, until no more steam is seen to condense on its surface, when it will be quite free from air if the steam was so; it is then lifted off, and the tube at the same time closed with the fingers, protected if necessary by means of a glove. Working sometimes at 80 lbs. steam pressure, we have been able to do this without injury to the fingers. The rubber tube is then permanently closed and the flask allowed to cool. If desired, the tube can be again secured by the fingers above the water which has collected, and the water removed without allowing any air to enter, but we have not found this necessary. The flask is now ready to take a sample of air, its capacity having previously been ascertained; an allowance of 1 c.c. for each litre, is made for condensed moisture.