evaporated is not decomposed, and, being mixed with the gas, it is not deposited in the water, but is carried probably in the form of gas into the tubes and pneumatic trough. In consequence of this circumstance, the analysis of this acid always gave a loss too great to be ascribed to oxygen. I thought at first that my analyses of the benzoates were inaccurate; but having repeated them with the same result, it became obvious that the cause of the anomaly must be something different. After having made analytical experiments on benzoic acid in a free state, and in that of benzoate and subbenzoate, I found that the loss diminished in proportion as the acid was more strongly retained by the base; so that the subbenzoates gave a less loss than the other salts. This induced me to examine whether there was really an evaporation of undecomposed acid during the experiments. I observed that the vessels through which the gas had passed, and in which it remained, had diminished in transparency, though not very remarkably. As these experiments were mostly made in the very severe winter of 1813-14, it happened sometimes that the temperature of the mercury in the trough was at 3° or at zero. On this account a greater quantity of the acid had been deposited in the part of the tube which entered the mercury, where I found sufficient to be collected and examined. It possessed the characters of benzoic acid; but its odour was a little empyreumatic.

The analysis of a quantity of subbenzoate of lead, containing 0317 of acid, gave 01414 water and 0.8645 carbonic acid. Hence the acid is composed of

Hydrogen
Carbon
Oxygen.

5.16

74.41

20:43

100.00

The most simple ratio in which these bodies can be combined is 30+ 12 H + 15 C, or (which comes to the same thing) O + 3 H + 5 C. By turning these volumes into numbers we have the composition of the acid per cent, as follows:

Now 6·69 × 3 = 20 07. Hence it follows that the analysis of the subbenzoate has given an exact result; while neither the benzoate nor fused benzoic acid ever gave more than 48 hydrogen and 70 carbon.

9. Tannin from Nutgalls.-I prepared a cold infusion of nutgalls in water, which I mixed with ammonia till it ceased to act as an acid, I then added a little of the infusion not mixed with

ammonia till I restored the property of reddening litmus paper, that there might be no excess of ammonia. The tannate of ammonia thus formed mixed with a solution of muriate of barytes lets fall abundance of tannate of barytes. I separated it by a filter, and washed it with care. From the experiments of Sir H. Davy, we know that the gallate of barytes is soluble in water; so that in this experiment it is only the tannate which is precipitated. The tannate of barytes mixed with diluted sulphuric acid furnished a super-tannate of barytes soluble in hot water. To this solution I added diluted sulphuric acid till almost the whole was decomposed. I filtered the liquid thus obtained, which had a very astringent taste, and strongly reddened litmus paper. This last property was not owing to the presence of sulphuric acid, for the liquid still retained tannate of barytes in solution. I now mixed it with caustic ammonia till the tannate of barytes began to precipitate. This solution of tannate of ammonia, with a small excess of tannin, was mixed with neutral nitrate of lead as long as any precipitate fell. The tannate of lead thus obtained had a yellowish colour, and became somewhat brown while washed. On examining the combination of tannin with oxide of lead, I found that the oxide has a strong propensity to combine with an excess of tannin, and that the precipitate obtained was in reality a mixture of tannate and supertannate of lead. I found that the super-tannate might be rendered neuter by boiling it in water as long as that liquid separates tannin. The water dissolves no part of the oxide; and the tannate which remains undissolved is neutral. If the precipitation of the tannate takes place at 212° there is very little super-tannate formed.

Neutral tannate thus prepared was dried in a vacuum, not to be exposed to the influence of the oxygen of the air, by which it is a little altered,

Tannate of lead obtained at different times, though by the same means, was analyzed by combustion. It gave 34 12, 3421, 34.56 per cent. of oxide of lead, probably as the excess of tannin was more or less completely removed. Hence this tannate is composed of

Now 52 oxide of lead contain 3.718 of oxygen.

When I attempted to obtain a subtannate of lead by means of the action of ammonia on the neutral tannate, I obtained a tannate of a browner colour, and very mucous; but which, after being washed and dried in a vacuum, gave 346 per cent. of oxide of lead. Hence it would appear that ammonia does not decompose the neutral tannate; just as happens when we add an excess of alkali to the tannate of the same alkali.

There is, however, a subtannate which I obtained by precipitating tannate of ammonia with a boiling solution of subnitrate of lead.

This compound is white; and, when dried, it becomes greenish ; but I have never obtained it of the same degree of saturation when prepared by different processes. The tannin in it was always combined with more than 1 times as much base as in the neutral tannate; though it never contained so much as twice that quantity.

I analyzed tannin by employing for the combustion both supertannate and tannate as neutral as possible, and the analyses furnished the same result: 04 of tannin produced 0·1425 water and 0.7625 carbonic acid. Hence tannin of nutgalls is composed of

But we have seen that 100 of tannin combine with 52 oxide of lead, the oxygen in which is 3.718. Now 3.718 x 12 = 44·616. Hence we may conclude that tannin contains 12 volumes of oxygen. But there is no doubt that there must be an analogy between gallic acid and tannin, and this analogy can scarcely be any thing else than the same compound radicle combined with different volumes of oxygen. Supposing, then, that tannin, like gallic acid, contains equal volumes of carbon and hydrogen, and that it is 120 + 18 C+ 18 H, or (which comes to the same thing) 4 O + 6 C + 6 H, its composition ought to be per cent.

Here we find a little more hydrogen and a little less carbon than the analysis indicates. This is a necessary consequence of the dishydrogenation which tannin undergoes by exposure to the air, in consequence of which the colour of its combinations becomes darker and darker; so that we never can procure this substance in a perfect state, excepting in fresh nutgalls in which the tannin is not yet coloured.

(To be continued.)

ARTICLE IV.

Observations on some Points connected with the Atomic Theory. By Thomas Thomson, M.D. F.R.S.

PROFESSOR BERZELIUS, having in his important dissertations on this subject published in the second, third, fourth, and fifth

volumes of the Annals of Philosophy, pointed out some difficulties which present themselves when we apply the atomic theory to the salts, and having in a recent paper refused to admit Mr. Dalton's attempts to remove these difficulties as valid-perhaps the reader will not be displeased if 1 state here, in as few words as possible, how I have been in the habit of viewing these difficulties when they happened to present themselves during my examination of the different genera of salts. When a difficulty occurs in any branch of chemical investigation, the greater number of persons there are who attempt to explain it, so much the sooner, in all probability, will it be removed. Even should my explanations not prove perfectly satisfactory, they may serve to convey a lucky thought to some other person, who may be more fortunate in his endeavours.

1. It appears from the tables of the sulphates, carbonates, and nitrates, published in the second and third volumes of the Annals of Philosophy, that the yellow oxide of lead combines always with two integrant particles of acid when it constitutes neutral sulphate, carbonate, and nitrate of lead. The same law holds in the phosphate, borate, oxalate, and all the other neutral salts of lead. Nitrate of lead is composed of. Sulphate of potash

41.580

......

2 n + 1 l
1 s + 1 p

....

....

11.000

52.580

Now if we mix 41.58 grains of nitrate of lead with 11 grains of sulphate of potash a double decomposition will take place, and two neutral salts will be produced, namely, nitrate of potash and sulphate of lead, composed as follows:

Now it is obvious to the eye that if the two salts be composed as we have supposed, such a double decomposition is impossible. The first two salts contain two integrant particles of nitric acid; the last two, only one integrant particle; while, on the contrary, there is only one integrant particle of sulphuric acid in the first two salts, but two in the last two. The weight of the first two salts is 52.58 grains; that of the last two, 50 777 grains. Thus about two grains of weight are lost by the decomposition; while a particle of nitric acid must be transmuted into a particle of sulphuric acid. The same absurdities and contradictions will be found to take place whenever we attempt to reduce any double decomposition, by means of a salt of lead, to calculation.

It is very obvious, from the appearance of these contradictions and absurdities, that there must exist some error in our tables; that the salts of lead cannot be constituted as we have supposed them. Now a very slight alteration will remove all the anomalies,

and render the composition of the salts of lead quite conformable to experiment. Reduce the weight of an integrant particle of yellow oxide of lead to one-half the weight which I have given it in my original table, published in the second volume of the Annals of Philosophy. Let the weight be 13·987 instead of 27 974. In that case all the salts of lead will be composed of one integrant particle of acid and one integrant particle of oxide of lead; we have

Here the double decomposition is obviously possible. The weight of the first two salts is just the same as that of the last two, and the number of integrant particles is the same in both.

It seems to me to be absolutely necessary to remove this anomaly from the salts of lead; nor do I see any other method of doing so except the one I have just now proposed; but if we reduce the weight of an integrant particle of yellow oxide of lead to one-half, it is obvious that it must no longer be considered as a compound of I atom lead + 2 atoms oxygen, but of 1 atom lead + 1 atom oxygen. It will be a protoxide instead of a deutoxide. If so, provided it be true that the brown oxide contains just twice as much oxygen as the yellow oxide, numbers adopted on the authority of Berzelius, the brown oxide must be a deutoxide of lead composed of 1 atom lead + 2 atoms oxygen. The red oxide, which is intermediate, must be considered as a compound of one integrant particle of yellow oxide and one integrant particle of brown oxide. This notion was first suggested by Proust, and appears to have been adopted by Mr. Dalton. I own that in consequence of the anomaly which I have just pointed out I have been for some time inclined to the same opinion; but a desire previously to examine the properties of the red oxide of lead under this point of view has hitherto prevented me from making any alteration in the weight of an atom of lead. Berzelius refuses to adopt this opinion. There is, no doubt, another method of getting rid of the anomaly which has been pointed out above without having recourse to it; namely, to double the weight of an atom of all the other metals; but that method would be attended with much greater inconveniences, and could not therefore be adopted with propriety. I do not see any alternative, therefore, in the present state of our knowledge, but that of adopting the opinion of Proust respecting the nature of red lead, and considering yellow oxide of lead as a protoxide.

If any person has taken the trouble to study the tables of the