chloride. This was made by twice crystallising the ferrous chloride, passing chlorine through a dilute solution, and removing excess of chlorine by heating gently for some time. The water was determined by careful ignition, the absence of chlorine being proved by mixing a portion of the oxide with pure carbonate of sodium, moistening with distilled water, drying and gently igniting. The mass was boiled with water, and on testing the solution only a very faint trace of chlorine was in any case detected. 1. Some solution of ferric chloride was precipitated by ammonia, the precipitate thoroughly washed with boiling water, and then boiled with distilled water for 112 hours, loss by evaporation being prevented by a condensing apparatus. The oxide became dense and lost its gelatinous appearance. It was again washed, and dried at 100° C. On ignition 17.85 grains lost 1·03 grain, or 5·77 % 2. The solution was precipitated as before and then boiled 100 hours without filtering from the chloride of ammonium, thoroughly washed and dried at 100° C. 13.805 grain lost 56 grain, or 4.05 %. 3. On using hydrate of potassium in slight excess, and treating as in No. 2, not filtering, and boiling for 100 hours, the dehydration was not so complete, contrary to expectation, as the oxide in the former cases contained traces, not to be estimated, of ammonia. 25.77 grains lost 1.69 grains, or 6.55 %. 4. Fearing that the loss might take place during the drying, a portion of washed oxide was dried at 100° C. for at least an equal time, being finely pulverised when dry, and finally heated until it no longer lost weight after several hours' interval. 15.64 grains lost 1.43 grains, or 9.14 %, being slightly below one equivalent. 5. Boiling the oxide for long periods being inconvenient on * Fe = 56. account of the bumping which occurs when the oxide has undergone the physical change, I resolved to try the effect of gently heating for a very prolonged period. The oxide was precipitated in the cold by ammonia, washed with cold water and heated for 1004 hours at a temperature varying from 50-60° C., with frequent agitation. In a few days it became dense, and brick-red in colour. It was washed with water at 50° C., and dried at the same temperature. 20-50 grains lost 84 grain, or 4.09 %. 6. As the oxide in No. 5 contained a trace of ammonia, hydrate of sodium was substituted as a precipitant, the oxide was washed as before, heated at 50-60° C. for 2,000 hours. The flask was closed and heated in an oil-bath. It thus appears to be impossible to drive off all the water, 4 to 5 per cent. adhering with extraordinary tenacity. The oxide thus prepared is brick-red, very dense, having a specific gravity of 4-545, that of red hæmatite being 47. It dissolves very slowly in nitric acid, more readily in hydrochloric acid. Under the microscope the dried oxide presents angular masses, translucent when very thin. Hoping to remove the unchanged hydrate with nitric acid, 24-18 grains of oxide prepared as in No. 6 were heated to 50° €. for one hour with 2 oz. of dilute nitric acid (1 part acid to 3 parts water). 4-77 grains dissolved, and the residue was washed until no longer acid, and dried at 50° C. 18-48 grains lost 65 or 3-517 %. These experiments show that the immense beds of hæmatite found in our own and other countries do not demand the suppcsition of great heat, to account for their anhydrous state. Probably much lower temperatures than that employed in the foregoing experiments, acting in presence of water for a long period would bring about the same change. The various statements respecting the amount of water in ferric hydrate are thus explained, that amount depending on the length of time during which the hydrate was exposed to the action of water, and a more or less elevated temperature. Chromic and aluminic hydrates were also operated upon by precipitating them with ammonia, washing and boiling for 100 hours. No action appeared to take place, aluminic hydrate retaining 3 eqs. of water, and chromic hydrate 5 eqs. The hydrates retained their gelatinous condition. X.-Tables for the Calculation of Vapour-density Determinations. By JAS. T. BROWN. Ir under the head of vapours we include 'gases, there are several tables which may be used for lessening the labour of calculating a vapour-density. There is one by Gerhardt* giving approximately the values of the expression 760 (1 + 0·00367 T) for every integral value of T from 0 to 30. There is another by Bunsen† indicating the values of 1 + 0·00366 T and log. (1 + 0·00366 T) for every T=0·1 from 2 to + 40. There is also a table calculated by Mr. Greville Williams, giving the values of the formula for every T 1 from 0 to 150; and 1 1 +0.00367 T lastly Gerhardt gives the values of log. (1+0.00367 T) for every integral value of T from 0 to 299. But I found that by substituting the weight of a cubic centimetre of air for that of a cubic centimetre of nitrogen in the formula with which I headed a table for the calculation of direct nitrogen determinations, I obtained an expression which occurs three times in the formula for the determination of a vapour-density according to the method of Dumas. For in this formula, which is generally written but which, written without any abbreviation and modified, becomes * Gerhardt, Traité de Chimie Organique, vol 1, page 50. Gerhardt, vol 1, page 115. D 0.0012932 P' − P + (760(1 + 0·00367t) × V x H) − (760(1 + 0·00367t) 0.0012932 0.0012932 [ V (1 + KT − e) ) − v (1 + 1 + 0.00367(t))]760(1 +0-00367T) − we see the repetition of the expression 0.0012932 × H From this expression I have calculated a table of its values for every T = 0·5 from — 20 to +50 and for every T = 1 from 51 to 350. I have also added a column of differences for calculating the fractional values of the expression. In order to find the weight, in grammes, of a given volume of air, at any pressure and temperature up to 350° C., multiply the number in the table headed 0.0012932 T·) 760 (1 + 0·00367 T-) corresponding to the temperature by the volume expressed in cubic centimetres, and also by the pressure expressed in millimetres. In substituting [(1+ 0.00367 (T-1)] for 1+0.00367T 1+0·00367 ť I have made no actual alteration, as the two expressions are equal; but, as I have given a table indicating the values of the 1 +0.00367 t expression is reduced to the simple form of 1 + k (T − t). In calculating by this table, it is only necessary to use it to the fifth decimal place, unless the value of v is very large. In calculating the table for correcting for the expansion of the glass, viz., that headed 1 + k (T − t) I have adopted the values of k as given by Regnault, taking the average value of t at 20. Gerhardt gives a table, for the same purpose, headed log. [1 + k (Tt)] for every T t = 10 from 100 to 290 and takes the value of k at 0.000027. The three following examples will illustrate more fully the use of the tables. Ex. 1.-To find the weight of 90 c.c. of air at a temperature of 15° C., and at a pressure of 746 mm. 0-00000161279 x 90 x 746 = 0·108282 gramme. Ex. 2.-To find the density of a vapour from the following data (Dumas). Regnault's Elements of Chemistry, translated by Betton, vol 2, page 420. † Gerhardt, vol 1, page 117. |