mises, Traité, tome i, page 52. From what he there states, a different number ought to have been obtained; and the error affects many of his subsequent formulæ relating to the barometer, expansion of fluids, of gases, &c.

In common with many others, Mr. D. seems to think that boiling the mercury in the tube of a barometer is of great consequence; this is at least doubtful, but certain it is that no human art can render a mass of this fluid such mere mercury that it shall not contain something, which in all its mechanical effects may not be called air. Sir H. Davy is in the right when he says so; and it is not a little to his credit to have perhaps by induction inferred that such is the fact, when there is no reason to think him aware of certain proofs to which it were needless here to appeal.

Mr. Daniell speaks also of filtering the mercury; though performed a thousand times no good effect can follow this practice. A knowledge of this metal gained from a peculiar application of it, warrants the assertion, that the mercury of commerce is not improvable by either distillation or filteration, in so far as its application is purely mechanical, and that its fitness for barometers can be completely known by bare inspection.

There is a probable source of error in the barometer hitherto little attended to, and of which Mr. D. takes no notice; in making the correction for temperature, it has ever been taken for granted, that the expansion proceeds pari passu, or that the fraction of dilatation is, for example, the same from 20 to

30, as it is from 90 to 100. This is, however, quite a gratuitous assumption; and although there are reasons for here suspecting something different from the utmost degree of precision, it is possible that the experimental means which we at present possess are inadequate to ascertain any appreciable discrepancies.

Much learning has been brought to bear on the other corrections requisite in using the mountain barometer; and Mr. Daniell could not have conferred a more substantial benefit than by having set the question at rest, as to the absolute dilatation of mercury, without a certain knowledge of which, all other minute attentions are little better than mere drivelling.

Mr. Daniell's account of the manufacture of barometers and thermometers is most certainly not overcharged. Throughout the continent, and even in England, the business is in the hands of itinerant Piedmontese; and these artists supply not only the general public with their glittering baubles, but furnish the greater part of the most reputable instrument-makers with their whole stock of meteorological wares. Such of these as choose to graduate their own scales, must confide entirely as to the quality of their tubes and the excellence of the filling, in one who has but indirect interest in the matter, or equivocal

reputation to lose; responsibility is thus shuffled from both, and rests on neither. Such, however, are the people who by unaccountable prescription supply the city of London, and the philosophers of England, with the instruments which Mr. Daniell so well describes.

If common notoriety did not bear Mr. Daniell out in his assertions, the shameful disagreement of the thermometers used by Captain Parry in his last voyage, would fully do so. On one occasion this amounted to no less than 13 degrees; Capt. Parry could do nothing else than give a mean, though in such 48° had as good a chance of being the truth as

a case

35°.

X.

ARTICLE XII.

Instructions for the Assay of Chloride of Lime.
By M. Gay-Lussac.*

THE uncertainty which has hitherto existed in the modes of ascertaining the quality, and consequently the commercial value of chloride of lime, and in no small degree retarded its coming into general use, has determined me to publish the following instructions on the subject. I shall divide the work into two parts; in the first I shall expose the principles on which the assay of the chloride of lime is founded, and in the second I shall describe the instrument which I call a Chlorometer, and the manipulations necessary for making the assay with sufficient accuracy for the purposes of those arts in which chlorine is employed.

PART I.

Principles on which the Assay of Chloride of Lime by means of Indigo is founded.

It is known that chlorine destroys vegetable colours, by forming new compounds with their component principles. It is in consequence of this property which it possesses, whether in the state of gas, in solution in water, or in combination with an alcali, that it is employed in the arts of bleaching, calico printing, &c. The same quantity of chlorine, in either of those three states, destroys the same quantity of colouring matter; and since by combination with an alcali, it becomes fixed, has scarcely any smell, keeps better, is inore portable, and more capable of concentration, the advantages of preparing it in that form are obvious.

Caustic potash, soda and lime, and even their carbonates, combine very readily with chlorine. Its combination with the

* From the Annales de Chimie.

potash, or soda of commerce, has long been known in France by the name of eau de javelle; that with lime was called oxymuriate of lime; but it is more accurate to denominate the first, as is now generally done, chloride of potash or soda, and the latter chloride of lime.

The chlorides of potash, soda, and lime, have very little stability of composition; the two first, indeed, can only be obtained in the liquid state, in a large quantity of water. If, for instance, we pass chlorine into a concentrated solution of potash, at first chloride of potash will be formed; but this chloride will soon be decomposed, and converted into chlorate of potash, and chloride of potassium. The two latter compounds, not having the property of destroying colours, must be avoided, and the only means of preventing their formation is to employ a very weak solution of the alcali, which, at most, should not exceed the proportion of 125 grammes to a litre of water. (In round numbers, about 4 oz. potash to 24 pints of water.)

Lime has not, like potash and soda, the inconvenience of converting the chlorine into chloric acid; it may consequently be combined with the chlorine en masse.

Lime, if perfectly dry, does not absorb chlorine, but it combines with it rapidly when in the state of hydrate, that is, after it has imbibed a sufficient quantity of water from a moist atmosphere, to split and fall to powder. Supposing it to be in the state of hydrate, it forms, according to M. Welter, a subchloride only, which is composed of

When mixed with water it is immediately decomposed; one half of the lime is precipitated, and the other half remains in solution, combined with the whole of the chlorine, and consequently forming a neutral chloride. Hence there are two combinations of chlorine with lime, a sub-chloride, and a neutral chloride. The sub-chloride is obtained by saturating hydrate of lime with chlorine, and the neutral chloride by dissolving the sub-chloride in water, or by saturating lime, dispersed through water, with chlorine.

The neutral chloride, or more simply the chloride, is very soluble; it may, however, be made to crystallize in small prisms. Its solution, left in contact with the air, is gradually decomposed, one part of the lime combines with the carbonic acid of the atmosphere, whilst its chlorine is disengaged. This decomposition of the chloride is retarded by constantly keeping an excess of lime in the solution. From these properties of the

chloride, the advantage of manufacturing the sub-chloride only is obvious; its preservation and transport are much more easily effected.

The quantity of chlorine in combination with water, or a base, may be estimated by several processes; but in the arts, in which dispatch is important, the preference has been given to M. Descroizilles' process, founded on the property of chlorine to discolour indigo. One part of indigo dissolved in 9 parts of concentrated sulphuric acid, and then diluted with 990 parts of water, forms the coloured liquid usually employed to ascertain the quality of the chlorine.

Under the same circumstances, chloride of lime discolours a quantity of this solution proportionate to its own; but if they vary, the results also are very variable. Thus, if we pour the chloride slowly into the indigo, a much smaller quantity of it is necessary to effect the discoloration than if we proceed differently. The minimum of discolouring effect, is obtained by pouring the indigo very slowly into the chloride, and the maximum by pouring the chloride very slowly into the indigo. Repeated trials have proved that the best process for obtaining constant and comparable effects, is to pour the solution of indigo rapidly into the solution of chloride, or the latter into the former. I shall explain the mode of operating by and bye.

If the indigo of commerce were pure, or always of the same quality, the quantity of its solution employed in each assay would give the relative quality of the chloride; but since its quality is very variable, the results of trials made with different indigos cannot be compared together. To avoid these inconveniences, I have followed the example of M. Welter, and taken as unity of discolouring power that of pure, dry, chlorine, at the barometrical pressure of 0.76 m. (29.92 inches,) and temperature of 0°. (32 Faht.) I prepare a solution of any of the best indigos of commerce of such a strength that the chlorine discolours exactly ten times its volume of it, and I call this solution the proof tincture; and each volume of proof tincture that is discoloured I call a degree, and I divide the degree into ten parts. Thus, if we take 10 grammes of chloride of lime and dissolve it in such a quantity of water as to form 1 litre of solution, the number of degrees, or volumes of indigo discoloured by one volume of the solution of chloride, will indicate the number of tenths of a litre of chlorine that the solution contains. Consequently, 1 kilogrammet of chloride of lime, whose quality had been determined by this method, and found to be of 7.6° orths, would contain 76 litres of chlorine. Each degree therefore is equal to 10 litres, per kilogramme of chloride, and each tenth of a degree to 1 litre. Supposing the sub-chloride

* Or 1 decagramme, Tr.

+ Or 100 decagrammes, Tr.

of lime to be perfectly pure, and formed as stated in page 219, it contains per kilogramme 101.21 litres of chlorine.

The base I have adopted appears to deserve the preference, from the simplicity and precision of expression that it admits of in chlorometry, which may remain unchanged, whatever means may be used to measure the strength of the chlorine.

We obtain more precision in general with a weak solution of chloride, marking for instance 4 or 5 degrees, than with a very concentrated solution. If, therefore, on the first trial we find that the chloride much exceeds 10°, we must add a known volume of water to the solution, for instance, twice its bulk; we then make a fresh trial, and triple the number of degrees obtained to get the true value of the chloride.

Assay of the Oxide of Manganese.

The purity of the oxides of manganese, employed in preparing the chlorine, is very variable, and consequently that of any particular ore must be ascertained by experiment, which may be easily done in the following manner.

Pure peroxide of manganese is formed of,

and furnishes 4.4265 gram. of chlorine, or 1.3963 litre, at the temperature of 0°, and under a pressure of 0.76 m; consequently 3.980 gram. would produce 1 litre of chlorine, and 1 kilogramme would produce 251-23 litres.

We take, therefore, 3.98 gram. of the oxide of manganese which we wish to assay, and treat it with muriatic acid, with a gentle heat, receiving the disengaged chlorine in rather less than a litre of milk of lime; towards the end of the operation we make the acid boil, to drive the chlorine from the vessels into the milk of lime, and add water to make its quantity just one litre. The quality of this chloride will exactly give that of the oxide of manganese.

The value of the manganese does not depend wholly on the quantity of chlorine it is capable of furnishing, but also on that of the muriatic acid required for its production. But the operation is delicate, and the low price of muriatic acid makes it unnecessary. I shall only remark, that the peroxide of manganese often contains the carbonates of lime, barytes, and iron, which saturate to mere loss a portion of the muriatic acid; moreover, as the manganese is not always in the state of peroxide, the quantity of muriatic acid required will not in that case be proportionate to that of the chlorine obtained.