approximation to the atomic weight of the corresponding member of column 4.

Returning to the second and third series, we find that the corresponding degrees of each have the same atomicity. Only the last four members of each column can combine with hydrogen, forming respectively RH, RH,, RH2, RH. Thus we see that the power of uniting with hydrogen, absent in the first members of each series, becomes high in the middle and declines again to the end. The permanence or instability of their compounds, their acidity and power to exchange hydrogen for metals, change according to the same gradation. Thus HCl is a well-marked acid, H,S a much feebler acid, H,P can no longer be called an acid, and in H,Si there is no residue of the acid character. In the corresponding members of the first column, again, HF1 is manifestly acid, H2O neutral, H,N basic, and H,C neither acid nor basic.

All the members of the second column combine with oxygen, forming salifiable bodies, as

Na2O, Mg2O2(MgO), Al2O,, Si2O4(SiO2), P2O5,

S2O (SO), Cl2O,;

two atoms of the element thus combining with a number of atoms of oxygen, increasing from 1 to 6. This arrangement corresponds with the decrease of basicity and the increase of acidity. At the beginning of the series is a decided base, soda; at the end a powerful acid, the chloric; and in the middle is alumina, which may be either a feeble base or a feeble acid.

A regular connection may be traced, not merely in the combining numbers, but in other chemical and physical characteristics. At the beginning of the first, second, third, and fourth series we find in each case bodies unmistakably metallic; at the opposite end come non-metallic substances, whilst those which stand on the boundary between these two classes take an intermediate position.

Again, if we consider the same set of elements with reference to specific gravity and atomic volume, we find periodic changes. Thus in specific gravity we have

Na=0'97, Mg=1'75, Al=267, Si=2'49, P=184.

Here the numbers, as will be seen, increase up to aluminium, and then decrease again, save that sulphur is anomalous, and has a higher specific gravity than phosphorus, as the

theory would require. With the atomic volumes the periodic changes are in the opposite direction; there is first a decrease and then an increase ;

Na=24, Mg=14, Al=10, Si=11, P=16, S=16, Cl=27.

In their oxides, also, the gradation of specific gravity and atomic volume is to be traced. The specific gravities are

Na2O=2'8, Mg2O2=3'7, Al2O ̧=4'0, Si2O=2·6,
P2O5=27, S2O6=1*9.

Here again there is a rise, succeeded by a fall. The series of atomic volumes is

Na2O 22, Mg,O,=22, Al,O,=25, Si2O, 45, P2Os=55,

S206=82.

=

Here is progressive increase without a return to the original number.

Speaking of alloys, M. Mendeleeff remarks that phosphorus and sulphur have not quite lost the metallic character of the elements at the beginning of the series, since the phosphides and sulphides have still the aspect of alloys, This feature is quite lost in the chlorides.

Turning to another series we find the atomic weights increasing as follows:

Ag=108, Cd=112, In=110, Sn=118, Sb=122,

Te 125, I=126.

Here again we have a decided metal, silver, at the head of the group; an equally decided non-metallic body, iodine, at the end; and in the middle an element, tin, whose metallic character-paradoxical as it may sound-has been questioned on totally different grounds, and which, according to M. Mendeleeff's scheme, should bear to silicon the same analogy which antimony does to phosphorus, tellurium to sulphur, or iodine to chlorine. The foretold-but as yet undiscovered-element, eka-silicium, is in this case to bridge over the gap, as in the three parallel instances is done by arsenic, selenium, and bromine.

In the silver group (No. 7) we find the specific gravities regularly decreasing as follows:

Ag=10'5, Cd=86, In=7'4, Sn=7'2, Sb=6'7,

Te=62, I=4'9.

Further instances of "small series," each consisting of seven members, known or unknown, are the following:—

Thus M. Mendeleeff considers that all the functions by which the dependencies of the properties of the elements on the atomic weights are expressed appear as periodic. The properties change in accordance with the increasing atomic weights, and are then repeated in a new period with the same regularity as in the former. If we examine his scheme we must admit that it brings admitted relations into a very prominent light. Such groups as fluorine, chlorine, bromine, and iodine; as sulphur, selenium, and tellurium; as nitrogen, phosphorus, arsenic, and antimony; as calcium, strontium, and barium; or as potassium, rubidium, and cæsium, though their respective members are placed in different series,-fall into positions which well agree with their respective analogies. Even the trinity of the commercial world-gold, silver, and copper-fall in the same horizontal line across the series; gold being a representative of its companion metals in a higher series. A further study of Table II. will bring to light many more curious instances of such representation, which we submit lend a powerful support to M. Mendeleeff's arrangement.

As a matter of course the mere arranging the elements in one arithmetical series would thrust these analogies totally out of sight. As completely would they be hidden if the series were made to consist of any other number of elements. M. Mendeleeff's classification is shown in two tables taken from the translation of his memoir in Liebig's “Annalen.” He admits that not all the known elements can be introduced into the small series. Thus in Table I. we see iron, cobalt, nickel, ruthenium, rhodium, palladium, osmium, iridium, and platinum forming an appended eighth group, concerning which the author remarks that they are all grey, all very sparingly fusible, their fusibility increasing in each line from iron to nickel, from ruthenium to palladium, and again from osmium to palladium. Between the corresponding members of the even and the odd small series (except the two first) there is a distinct difference, whilst the corresponding members of the even and again of the odd, series show a close analogy.

Thus if we recur to the four series above mentioned :—

Here it is at once evident that there a decided analogy prevails between the corresponding members of Nos. 4 and 6, and again of Nos. 5 and 7. But between Nos. 4 and 5, 5 and 6, or 6 and 7, there is no such analogy.

Table II. shows the elements as arranged in "great series." In Series I. stands hydrogen alone. This reminds us of a remark made lately by Prof. F. Guthrie, that very generally the commonest bodies are the most anomalous. Series II. contains the seven elements from lithium to fluorine inclusive, these two first groups being designated typical." Series III. contains also seven elements. Series IV. contains fourteen bodies, which, when the three blanks indicated by M. Mendeleeff have been filled up, as one of them already is by the discovery of gallium (ekaaluminium), will be raised to seventeen.

This "great series " consists of the two" small series," No. 4 and No. 5, with three of the non-serial elementsiron, cobalt, and nickel-interposed between them. The great series No. 5 is composed of sixteen known elements, and a blank space is left for a body not yet discovered. The small series here embraced are Nos. 6 and 7, and again three of the non-serial elements-ruthenium, rhodium, and palladium-are interpolated. The great series VI. is very imperfect, comprehending only four elements, which are placed parallel with the first four bodies in Nos. IV. and V. Great series VII. is more perfect; it contains parts of two small series, and again three non-serial elements-osmium, indium, and platinum. Here, however, we are struck with a difficulty: between the atomic weights of tungsten (W=184) and osmium (Os=195) there is the gap of II. Now if M. Mendeleeff concluded that two undiscovered elements, eka-aluminium and eka-silicium, must exist to fill up the interval of 10 between Zn-65 and As=75, in great series IV., why may we not with equal right calculate on the discovery here also of two unknown bodies? But should this be hereafter found to be the fact, the correspondence between the members of Series V. and VII. will be deranged. Gold will occupy the position corresponding not with silver, but with cadmium; mercury VOL. VII. (N.S.)

Y