ART. XXVI.-Upon the Atomic Volumes of the Elements ; by FRANK WIGGLESWORTH CLARKE, S.B.

IN a recent paper in this Journal I called attention to some remarkable relations which seem to exist between the atomic volumes of certain elements in the liquid condition. Since that time I have found similar relations between the same elements in the solid state, and also between the atomic volumes of some other solid elements of which no liquid compounds are known.

A few relations between the atomic volumes of different metals have previously been observed. Chief of these we find the equality of the atomic volumes of the metals of the iron group, the similar equality connecting the platinum metals, the equality between silver and gold, and that between molybdenum and tungsten. But, as far as I have been able to learn, nothing seems to have been done toward studying the more remarkable multiple relations which connect the members of certain natural groups with each other.

None of the specific gravities cited in this paper are due to my own determinations. I have simply collected the most reliable data which I have been able to find upon specific gravity, and thence calculated the atomic volumes of the elements. But in every case I shall try to give the authority to whom the experimental observations are due.

The alkaline metals most appropriately come first in order. For the first four only have we any data, since cæsium has not been isolated. The specific gravities are as follows. Lithium 0-589, Bunsen; sodium 0.972, and potassium, 0-865, Gay Lussac and Thenard; rubidium 1.52, Bunsen. Upon dividing the atomic weights of the four metals by these specific gravities we get as their atomic volumes respectively the numbers 11.9, 23-7, 45·1, and 56-2. It will at once be seen that the three highest of these values are very nearly multiples by whole numbers of the lowest. If we assume that in reality an exact multiple relation connects the members of this group, and then, adding together the atomic volumes actually found for them, we take an average in order to get the multiples with the least possible alteration of our data, we obtain for these four metals the values 114, 22.8, 45-6, and 570. These stand to each other as 1:24:5, and vary but very slightly from the numbers actually found, and, moreover, correspond to these trifling alterations in the specific gravities, respectively, 025, 037, 010, and 022. Is it not safe to assume that under strictly com

parable circumstances this apparent multiple relation would prove to be exact ?

The oxygen group, oxygen, sulphur, selenium, and tellurium, affords a still more remarkable series of values. The atomic volume of oxygen in its solid compounds, varies according to the nature of the combination. For this element in the solid state Kopp has deduced three values,-26, 52, and 104. Kopp himself has called attention to the fact that these stand to each other as 1:2:4. For this element in liquid compounds the same observer deduced the values 7.8 and 12.2. If we compare these with the values for solid oxygen we shall find that the four lowest numbers stand exactly in the ratio of 1:23: 4, and that an alteration of 0.8 in the highest, making it 130 instead of 12.2, will place it also as a fifth multiple of the lowest. I cannot but think it probable that Kopp himself must have noticed this relation, although I find no mention of it in any paper of his that I have seen. The atomic volume of sulphur varies, both with its crystalline form, and with the state of combination. The specific gravity of the octahedral modification, according to Marchand and Scheerer, is 2:045. This gives us for its atomic volume the number 156, precisely three times the second value for oxygen. In the liquid condition also, it will be remembered that the lower value for sulphur, 234, is three times the lower value for oxygen.

Many metallic sulphids possess atomic volumes very nearly the sums of those of the free sulphur and the metal. The native sulphids, syepoorite, CoS, millerite, NiS, and troilite, FeS, however, seem to vary from this rule. Their atomic volumes, deduced from the specific gravities given in "Dana's Mineralogy," are respectively 16.9, 17.6, and 18.4. But here

it must be borne in mind that the native minerals are not absolutely pure, and that the specific gravity of each mineral often varies considerably. Since the metals in these three sulphids have equal atomic volumes, and since their corresponding compounds in most cases have equal values also, it seems safe to assume that these minerals in an absolutely pure condition, would possess also equal atomic volumes. The mean of the three numbers above given is 17 6. If we subtract from this the value of the metal, 69, we get 107, as the atomic volume of the sulphur, a number varying but 0:3 from twice the second number for oxygen. As we pass on we shall meet farther confirmations of this apparent multiple relation. In the liquid state sulphur has three atomic volumes, two of which,

Buff, Kopp and Zamminer. schen Chemie."

"Lehrbuch der Physicalischen und Theoreti

ART. XXVI.-Upon the Atomic Volumes of the El by FRANK WIGGLESWORTH CLARKE, S.B.

IN a recent paper in this Journal I called attenti remarkable relations which seem to exist between volumes of certain elements in the liquid cond that time I have found similar relations betw elements in the solid state, and also between t umes of some other solid elements of which pounds are known.

A few relations between the atomic vol metals have previously been observed. Chi the equality of the atomic volumes of the group, the similar equality connecting t the equality between silver and gold, and denum and tungsten. But, as far as I' nothing seems to have been done tow remarkable multiple relations which certain natural groups with each oth

None of the specific gravities cite my own determinations. I have si able data which I have been able and thence calculated the atom But in every case I shall try the experimental observations an The alkaline metals most ap For the first four only have w been isolated. The specific g 0.589, Bunsen; sodium 0 Lussac and Thenard; rubid the atomic weights of th ities we get as their aton 11.9, 23·7, 45·1, and 56-2 highest of these values numbers of the lowest. multiple relation conne adding together the at we take an average in possible alteration of the values 114, 22 other as 1:2:4:5," bers actually found, alterations in the 010, and 022. IN

nst naturally expect that it- liquid compounds) ugh so extremely tical reasons refer leaving

show more ferent atomic es for these the following se determined 's actually detertent of alteration

S.

this connection to observe the remark

on connecting the atomic weights of this

ifferent allotropic conditions of sulphur and no numerical relations whatever. The atomic ismatic sulphur I find to be from 16·3 to 167, and orphous selenium to be 18.6. These values find no he series above given.

nitrogen group affords another remarkable series of In the liquid state, it will be remembered, nitrogen ree values, 2 15, 86, and 17·2, all multiples of the lowThen came boron, phosphorus, vanadium, and arsenic, la common value, three times the second number for rogen. In the solid state, however, we find a variation From this. For nitrogen itself we have few suitable data, and I have made no elaborate calculations concerning it; yet as far as I have examined, it seems to possess several values. One of these is easily found. Kopp determined the atomic volume of NO, in nitrates, as 28-6: and if we subtract from

* In the numbers which required no alteration, parentheses seemed unneces

sary.

The first of these

23.4 and 28.6, are cited in my last paper.* is three times the lower value for liquid oxygen, and of course is also a multiple of 2-6, which seems to be the starting point for this series. The second is also an exact multiple of 26. In two observations Buff obtained for the atomic volume of sulphuric anhydrid at its boiling point, the numbers 44.18 and 44-19. Regarding this compound as (SO)O, and taking Kopp's determination of the higher value for liquid oxygen, he deduced from these numbers as the atomic volume of hexvalent sulphur in its liquid compounds, the value 120. If, however, we take as the higher value of oxygen the altered number 130, we shall get as the value of sulphur in (SO)O the number 10.39, which agrees closely with the 104 above suggested for sulphur in the solid sulphids of iron, cobalt and nickel.

The approximate equality between the atomic volumes of solid sulphur and selenium has often been noticed, and in my last paper I showed that a similar equality probably existed in the liquid condition. The specific gravity of crystalline selenium, according to Hittorf, is 4.808. This gives us as its atomic volume, 16.5, which exceeds by 0-9 the value for octahedral sulphur. An exact equality between sulphur and selenium, however, seems to me doubtful, since I have repeatedly noticed that, as a rule, the atomic volumes of selenids slightly exceed those of the corresponding sulphids, and also that the selenates have values slightly in excess of the sulphates. Yet there are striking exceptions to this, and the point is by no means decided.

Like the sulphids, many selenids have atomic volumes equal to the sum of those of the selenium and the metal. For the proto-selenid of iron we have no data, but the artificial selenids of cobalt and nickel, CoSe and NiSe according to Little, have the specific gravities 7 65 and 8:46, whence we obtain for their atomic volumes the numbers 18.2 and 169, or a mean of 17·5. This is very near the mean obtained for the sulphids, and gives us as another value for selenium, the number 106, only 0-2 greater than the double multiple of oxygen.

The sp. gr. of tellurium, according to Löwe, is 6·18, and hence we get as the atomic volume of this element, the value 20.7. 208 is exactly four times the second value for oxygen.

We now see that in the solid condition the atomic volumes of these four elements are to each other very nearly in the relation of 1:3:3: 4, and since the first three preserve the

My paper in the March No. of this Journal was sent to press before I had seen Buff's calculation of the atomic volume of hexvalent sulphur. See Buff's "Grundlehren der Theoretischen Chemie "