peared to have combined with 10 parts of water; while analysis gives 20 parts more water in the starch sugar than in the starch. It is obvious, however, that the proportion of water obtained by the first method must be too small, as in a process of that kind it is very difficult to avoid all loss, besides that which is occasioned by a commencement of roasting.

Analysis of the Sugar of Grapes.

I obtained the sugar, with which the following experiments were made, from Mr. Pautex, by whose labours the manufacture of sugar of grapes has been greatly improved.

100 parts of sugar of grapes, dried at the temperature of 53.5°, and when the hygrometer stood at 75°, when exposed to the heat of boiling water, lost 3.14 parts of water; and when burnt left a residuum weighing 0.513 parts.

0.55 centigrammes of the same sugar, dried at 53.5°, when burnt, consumed, according to the mean of two experiments, 34.21 cubic centimetres of oxygen gas, and formed 3617 cubic centimetres of carbonic acid gas. Hence 100 parts of sugar of grapes, dried at the temperature of boiling water, are composed as follows:

The result of this analysis of sugar of grapes does not differ farther from that of starch sugar, than is usual in two different experiments upon the combustion of the same body. These two sugars likewise approach so near each other in all their other properties, as to render it probable that they constitute only one species. They both melt at the temperature of boiling water; they have both the same sweet and fresh taste; they both undergo the vinous fermentation; they both crystallize confusedly in spherical crystals; they are both equally soluble in water and in weak alcohol; and all the differences which exist between them are analogous to those which we frequently find between sugar of grapes from two different varieties of grapes.

2

Sugar from the cane and from beet differs much from these two, and from all other sugars, by containing a greater proportion of carbon. According to the analysis of Gay-Lussac and Thenard, 100 parts of the sugar of the sugar-cane contain between 42 and 43 parts of carbon, and the oxygen and hydrogen are so graduated as to form water without any residue. I have obtained the same result, except a small excess of oxygen above the elementary water, which might very well be owing to an error in my experiment.

Method of conducting my Experiment.

In the preceding analyses I made use of the following method, which I employ in the analysis of gummy and woody bodies that contain very little or no azotic gas. I reduce the vegetable body to as fine a powder as possible, mix it with 50 times its weight of silicious sand, and put it into a glass tube bent in the middle at a right angle, close at one end, and furnished at the other with an iron stop cock. This tube is about a metre (39.37 inches) in length, and its width is such, that it is capable of containing rather more than 200 cubic centimetres (12-2 cubic inches) of gas. I weigh the vegetable substance by the way of substitution, in the tube itself, by means of a balance, which, when loaded, turns by the addition of one milligramme (0.015 gr.) The air is then extracted from this tube by means of the air pump; it is filled with oxygen gas, the stop cock is shut, and all its joints are covered over with mastich, or it is surrounded by a column of mercury during the burning of the vegetable matter, in order to be certain that no gas makes its escape while it is expanded by heat: for the stop cock, which incloses a portion of gas, is not always able to withstand the compression or dilatation which takes place within, which frequently acts as a kind of valve.

The stop cock being thus secured, I heat that part of the tube with which the vegetable body is in contact, to an obscure red; and for this purpose I employ a spirit lamp, which gives a flame at least one decimetre high (3.937 inches), and at least of such a diameter as to surround the whole circumference of the tube. Liquid and sooty matter speedily disengages itself, and is deposited in a neighbouring part of the tube, which is kept cool by being surrounded with moist paper, I afterwards heat this part of the tube to redness. The vegetable matter burns, and is partly volatilized and condensed in another part of the tube. This new portion is heated to redness in its turn. I go on in this manner, heating the condensed portion a great many times in succession, till the decomposition appears to be complete, and the liquid remaining to be nothing else than pure transparent water.

To measure the alteration of the volume, which the gas has undergone during the combustion, I fill a graduated tube one decimetre in length, and furnished with an iron stop cock at each end, with mercury and with oxygen gas, fix it on the first tube, plunge the end of it under mercury, and open both the stop cocks, which establishes a free communication between the two tubes, and observe the increase or diminution of gas which has taken place. get the gas out of the tube, I screw upon both the tubes a balloon, furnished with a stop cock, and filled with mercury. The mercury runs into the tube, and the gas makes its way into the balloon. The quantity of it is sufficient to make four eudiometrical experiments. on its nature.

To

In order to determine whether the vegetable substance subjected

to combustion contained azote, I wash with 30 grammes of water the large tube and the mercury with which it was filled, in order to drive out the gas, and separate this water from the mercury, by means of a funnel with a capillary tube. I then distil this water at a moderate heat over hydrate of lime. A tube moistened with: neutral nitrate of mercury, dipped into a few drops of the distilled liquid, occasions a greyish black precipitate, even when the quantity of azote in the vegetable substance analysed does not exceed one thousandth part of the whole.

By means of this reagent I estimated nearly the small quantity of ammonia which existed in the products of my analyses, afterhaving ascertained, by previous trials, how much water it was necessary to add to a known mixture of ammonia and water, to reduce it to the limit of precipitating immediately with nitrate of mercury. I found that when water has absorbed half its volume of ammoniacal gas, or, in other words, when one gramme of water has absorbed 0.0003443 gramme of ammonia, 55 times its volume of pure water may be added to it, without preventing the precipitate from immediately appearing. But this is the utmost limit of dilution; for if more water be added, the precipitate does not immediately appear; supposing that in both cases we operate upon equal doses of the liquid; and the quantity which I always take is six/ grammes. From this single datum a table may be constructed,. showing the quantity of ammonia in the liquid, according to the quantity of water necessary to add to it, in order to bring it to the limit of precipitating with the nitrate of mercury. We mix determinate quantities of the liquid and water together till we come to. this limit: suppose, for example, that I find that this limit takes place when I mix together equal quantities of the liquid and water, the table in that case shows me that one gramme of the liquid contains 0.000012297 gramme of ammonia. From this quantity of ammonia I determine the quantity of subcarbonate of ammonia formed during the experiment, and introduce the constituents of this salt into my calculation.

In these analyses I only burn five or six centigrammes of the vegetable body. The error, resulting from weighing, amounts; only, in consequence of the accuracy of my scales, toth of the analysed body. I endeavour to remove this error by repeating the analysis several times, and only taking the results when they are sufficiently accordant.

The uncertainties attending the eudiometrical processes are much more considerable. By means of them alone can the im portant question be answered, whether in saccharine, gummy, resinous, and starchy bodies the oxygen and hydrogen exist in the

* The reagent consists of 64 parts of crystallized nitrate of mercury, dissolved in 100 parts of cold water. All the results which I give are founded on such a state of the solution. But as it is probable that others may not prepare in exactly the same manner as I do, it will be necessary for every one to determine the data respecting the solution of nitrate of mercury for himself.

exact proportion necessary to constitute water. The quantity of oxygen above this proportion, which I found in my analyses, does not appear to be sufficiently great to destroy the law established by Gay-Lussac and Thenard, that a vegetable body is acid when it contains an excess of oxygen above what is necessary to constitute water, since it may be ascribed to errors in my experiments, or in my method. But some other of my analyses have given so great an excess of oxygen, that it cannot be ascribed to errors in the experiment. Thus, for example, the analysis of gum arabic, which I repeated five times, with very little difference in the results, gives its composition as follows: 100 parts of gum arabic, dried at the temperature of boiling water, and abstracting the ash, are composed of

Gum tragacanth gave me very nearly the same result; and sugar. of milk contains five or six per cent. of oxygen above what is necessary to constitute water, and 39 or 40 per cent. of carbon, which is nearly the quantity given by Gay-Lussac and Thenard.

On the other hand, I have found vegetable bodies neither of a resinous, oily, nor alcoholic nature, which yet contained no excess of hydrogen above what was necessary to constitute water. To such belong purified sugar of manna, precipitated from boiling alcohol, 100 parts of which, dried at the temperature of boiling water, contain

Carbon..... 47.82-
Oxygen... 45.80

or Water

Hydrogen 6.06
Azote

51.86

.....

0.32

100.00

The alterations in the laws respecting the combination of the constituents of vegetable bodies resulting from my experiments are not remarkable. We see that all gummy, saccharine, and starchy.. vegetable bodies consist, in fact, of little else than carbon united. to a portion of water reduced to its elements. However, the excess of the oxygen or hydrogen in these bodies, above what is requisite to constitute water, must be greater than is indicated by my analyses. Because the boiling heat of water, at which the vegetable bodies were dried, is not sufficient to dry them completely, and to remove all the water retained by capillary attraction this adventitious water figures in the analyses as so much elementary water, and renders the quantity of it found in the substance too great, when compared with the other constituents.

ARTICLE X.

Answer to Mr. Prevost's Queries respecting the Explanation of Mr. B. Prevost's Experiments on Dew. By Wm. Charles Wells, M.D. F.R. S.

SIR,

(To Dr. Thomson.)

HAVING seen in the last number of your Journal an indirect application to me by the acute and learned Mr. Prevost, of Geneva, I request permission to inform that Gentleman, through the same channel, that the explanation which he has given, in his work on Radiant Heat, of Mr. Benedict Prevost's observations on dew, is regarded by me as being neither referrible to the whole of them, nor altogether satisfactory with respect even to those to which it applies.

In the first place, he takes no notice whatever of a whole class of Mr. B. Prevost's observations; those, namely, which relate to what happened, when glass vessels partly filled with various substances were exposed by him to the influence of the causes of dew. In these experiments the lower parts of the vessels remained dry, though other parts of them, which were above the level of the contained substances, were covered with dew. The author adds, that the distance between the upper surface of the contained substance, and the part of the vessel at which dew began to appear, varied according to the nature of the substance; it being greater, for example, in a vessel containing mercury, than in another of the same size containing water.

In the second place, Mr. Prevost, of Geneva, supposes dew to form in circumstances, in which, I venture to say, it cannot occur. If a thin plate of a bright metal be fastened to a pane of glass in a window of a room, the air in which is warmer than that without, dew, according to his representation, will be deposited on the outside of this piece of glass, as the metal covering its inside is a screen against the heat, which is radiated towards it by the walls and contents of the warm chamber. Now it is manifest, that the utmost effect which can be produced in this way will not occasion the outside of the glass to be as cold as the external atmosphere; for the metal will admit into itself some part, however small, of the heat which is radiated to it, and will communicate this to the glass, along with that which it acquires at the same time, by conduction, from the contiguous warm air. But, if the outside of the glass be warmer than the air, dew will not form upon it; since, according to my experience, bodies will not receive dew unless they be colder than the air.

I think, Sir, that I need say nothing more in justification of the intention, which I formerly entertained, of offering an explanation of Mr. B. Prevost's observations on dew, though one had already been given of a part of them by Mr. Prevost, of Geneva. Much