The mineral might therefore be regarded as consisting of 1 atom soda elæolith (according to the older formula), and of 1 atom calcareous spar, and should have the following composition:

Silica

38-23

But if my formula be adopted as the correct one for elæolith, cancrinite would be expressed by the formula Na Si + 2 Al Si + Ca C, and calculated accordingly, should have

A more satisfactory agreement of the formula with the composition found can scarcely be expected in analysis. They appear merely to differ in the formula giving the amount of lime about 1 per cent. higher; but this difference disappears almost entirely if we admit that the potash replaces a portion of the lime. That lime and potash replace one another is evident from the analyses 10, 11, and 12 of nepheline. This latter is, indeed, only essentially different from the other elæoliths by the greater quantity of lime it contains; but it seems that principally the amount of potash has been diminished by its occurrence, while the soda does not at all differ in proportion. Cancrinite, therefore, actually contains 1 at. of nepheline of a composition, such as was formerly adopted for the Vesuvian nepheline, but which has hitherto not been found in an isolated mineral.

On closer consideration, it will be found that the cancrinite affords even a better test for the correctness of my formula than the elæolith itself. The atomic value of the elæolith, according to the older formula (if we add 4 Si, 3 Äl, 2.4 Na, and 0.6 Ka together), equals 5529, according to the new one (if 3 Si, 2 Al, 1.6 Na, and 0.4 Ka are added) = 3877. Now since the weight of 1 at. of carbonate of lime is 632, it results that considerable differences must originate by these different atomic values on the per centage calculation of the constituents of a combination of both atoms. In fact, were the atomic weight of elæolith = 5529 the cancrinite ought not to contain more than about 10 per cent. carbonate of lime; while, according to my formula, it should contain 14 per cent. But the latter number is likewise confirmed by the two analyses of G. Rose.

IV. COLOUR OF THE ELÆOLITHS.

Only two of the analysed varieties of this mineral possess colour, those from Fredricksvärn, the one green, the other brown. When both are finely powdered and decomposed with concentrated hydrochloric acid, the silica separated has the same colour, only in a less degree. This is especially evident in the green elæolith, which is more intensively coloured than the brown, and with which I therefore chiefly performed experiments. Even when the concentrated hydrochloric acid is evaporated, and the dry mass again moistened with acid, and treated with water, the filtered silica still retains its colour, which, however, immediately disappears on heating the silica. It is further destroyed when the silica is heated with nitric acid, or the mineral decomposed by fuming nitric acid. This latter action sufficiently proves the colour to be of organic origin. But the colouring substance must certainly be of a peculiar nature, as it withstands the action of fuming, nay, even chloriferous hydrochloric acid.

The following are the main results afforded by these examinations of the elæoliths. 1. The formula for elæolith

Na2

and nepheline must be altered to Si + 2 Al Si. The

2

atomic relation of soda and potash is as 4:1. Both minerals are perfectly identical, the latter being merely characterized by a somewhat greater amount of lime.

2. Elæoliths, from the most varied localities, exhibit traces of hydrochloric and sulphuric acids, especially of the former. 3. The amount of water in elæoliths varies considerably, and must be considered as accidental.

It is probably only hygrometric, and is perhaps prevented from evaporating at 80° R. by a force similar to that which several porous bodies exercise on gaseous substances. 4. The colour of some elæoliths is of organic origin.

5. The specific gravity of the elæoliths is exceedingly near XVIII. On the Theoretical Constitution of the Compounds of Ammonia. By ROBERT KANE, M.D., M.R.I.A.

I

N the course

of the investigations to which I have subjected

the various classes of compounds that ammonia is capable of forming, it has been my lot to submit to the consideration of chemists a great number of theoretical views to which I had been led by my experimental results, and by which I conceived that the mutual connexion of the different classes of ammoniacal compounds could be explained, and their origin and properties accounted for more satisfactorily than it was possible to do by means of the ideas that had been previously received in science. In advancing this new theory of the nature of ammonia and its compounds, I was not so sanguine as to expect that our ideas of a department of chemistry so complex and so important could be immediately or easily modified, or that the adoption of my views could take place without much conflicting reasoning and discussion. In this respect I have had cause to be very much gratified. All principles that can be considered as really vital to my theory have been adopted by the most eminent chemical philosophers, and in place of being dissatisfied that in the collateral parts of the theory some portions have been thought not positively proved, and which have hence been criticized and left for the time aside by Graham and by Rose, I was at once surprised and pleased to see how little had appeared in the eyes of these acute-minded chemists unfit for being at once adopted into science.

I believe, however, that even in those portions of the theory to which Graham and Rose have not acceded, some of the difficulties arise from a want of clearness and detail in the description of my views, into which error I fell from being too anxious to avoid prolixity. As also since that period some additional evidence has been obtained which corroborates my opinions, I shall now advert to those points which are yet debated, and perhaps place them in a clearer point of view than had been done in my former paper. So far as regards the action of ammonia without water, all my ideas have been adopted; but in the relation of the ammonia and water in the common ammoniacal salts, where the ammonium theory of Berzelius comes into question, the evidence for my theory has not appeared so perfect. In fact, in order to see the true relation of the Berzelian theory to mine, it is necessary to contemplate the common salts of ammonia under two different points of view,-1st, their position as alkaline salts, and 2nd, their position as compound bodies, without reference to any other circumstance. In the first the proper theory of the salts of ammonia is that of Berzelius, but for the second purpose it is necessary to adopt the ideas on which mine is founded.

For in fact the question is, What is sal-ammoniac? Its most striking philosophical character is its equivalence to chloride of potassium. It has the same crystalline form. It enters into combination subject to the same laws. They are two bodies formed decidedly upon the same plan. But chloride of potassium contains only two elements, while sal-ammoniac contains three. There is one common to both. The residual elements are equivalent, and Cl K and Cl NH4, as well as Kand NH4, are bodies which correspond to each other. Ammonium when isolated, as it has, in the amalgam, all but been, appears to possess the properties of an alkaline metal; it markedly resembles potassium. That is the Berzelian theory, in which to the full I believe as well as Berzelius. The equivalency of sal-ammoniac and chloride of potassium is a fact, and the equivalency of the K in the one and of the NH1 in the other, is the natural inference from it. When therefore the equivalency of the ammoniacal and potash salts is under question, the ammonium theory is correctly used: it is not ammonia, it is not amidide of hydrogen which replaces potash, but it is to be called oxide of ammonium in compari

son.

But if we for a moment cease to consider the relation of the ammoniacal and the potash compounds, and taking sal-ammoniac by itself, proceed to examine what light can be derived from other sources towards illustrating its internal constitution, the question presents itself, can we believe the ammonium to be ready-formed in sal-ammoniac? Can we consider the ammonium, which in the amalgam gives up its hydrogen spontaneously, to retain it so strongly when in contact with iodine or chlorine, and to enter into combination only as a single and perfect group? The answer to this question, in the framing of which all the classes of ammonia compounds require to be taken into account, led me to the development of my theory.

The combinations of ammonia with the anhydrous chlorides of copper, zinc, and mercury, resemble in all essential characters sal-ammoniac, and moreover, like it, those which are volatile or soluble without decomposition are found to belong to the regular system of crystallization. They further combine with the metallic chlorides of the magnesian class to form double chlorides. Thus there are

1.

2.

3.

CI H4N

and Cl Cu + Cl H1 N.

Cl Cu NH3 and Cl Cu + Cl Cu H, N.
Cl Zn NH3 and Cl Zn + Cl Zn H, N.

4. Cl Hg NH3 and Cl Hg + Cl Hg H3 N.

The completeness of the analogy thus indicated is acknowledged by Graham, who proposes to extend the Berzelian theory so as to include these cases. He assumes, that in the compound radical ammonium the hydrogen may be replaced by a metal, and thus a cuprammonium NH3, Cu, a zincammonium NH3 Zn, a hydrargammonium NH3 Hg, may be capable of individual existence. If Hg and Hare replaceable, then Hg, H, N is equivalent to H1 N, and thus he agrees with me that the type of sal-ammoniac and white precipitate (Cl H. H Ad and Cl Hg + Hg Ad on my theory) is the same.

But why have we not Cl Hg3 HN, or Cl Cu3 HN, or Cl Hg4 N and Cl Zn, N, taking their place among these bodies generated by ammonia? For Cu4 N or Zn4 N would also be equivalent to ammonium. The replacement stops when there remain yet two equivalents of hydrogen to the nitrogen; and it is only by a temperature such as destroys completely the constitution of these bodies that a metallic azoturet can be produced.

The compounds containing oxygen acids are precisely similar to those chlorides just described. The bodies

3

3

3

SO3. O Cu NH3 and SO3. O Cu + S O3 . O Cu NH3 SO. OZn NH3 and SO3. O Zn + SO3. Ο ΖηΝΗ SO.OH NH3 and SO.OH + SO. OΗΝΗ and SO3. NH3+SO.OHΝΗ are so obviously similar in constitution, that the one explanation of their internal structure must be admitted.

There exist thus two sorts of compounds, which are ordinary ammoniacal salts with metallic oxide in place of water, or in the words of the theory of Graham, that contain ammonium in which hydrogen is replaced by a metal; those as Cl. Hg NH3, in which one equivalent, and those as Cl. Hg NH, Hg, in which two equivalents have been thus replaced. The substitution stopping there shows that N H2 is fixed, and thus that even if these various sorts of ammonium be admitted, the amidogene must be considered as pre-existing in them; and as

Graham admits my formula for ammonium NH1= Ad so his metallic ammoniums become

H
Η'

Under this form the ammonium theory is capable of being