ATOMIC THEORY. 



AUSTRALIA AND POLYNESIA. 



In each of the periodical series the capacity 

 of combining with oxygen seems to increase up 

 to a certain point, and then to decrease. The 

 series headed by silver may be taken as a type 

 of the oxygen compounds formed by the ele- 

 ments in the other periods, the formulae being 

 here doubled for the sake of uniformity : 



Ag a O; Cd,O; In,O,; Sn a O 4 ; Sb a O 6 ; Te a - 

 O.; I a O 7 ; OsO; IrO 9 ; PtO a . 



The first five members of every period but 

 one follow these types exactly. The variations 

 of affinities for chlorine and hydrogen within 

 the groups are made evident by the following 

 formula?, combinations with hydrogen being 

 confined to the last four terms of the groups: 

 Li Cl ; G C1 3 ; B Cl s ; C Cl. 

 NaCI; MgCl a ; A1C1,; Si Cl. 

 CH 4 ; NH S ; OH a ;FH. 

 SiH<; PH; S H s ; 01 H. 



Dumas, to whom the merit of grouping the 

 elements into natural families belongs, called 

 attention again to Front's neglected hypothe- 

 sis in 1879. The French chemist discovered 

 simple numerical relations between the metal- 

 loids and some of the families of metals be- 

 longing to each group. In the sulphur group, 

 for instance, at the head of which oxygen is 

 now placed, there is a progression representing 

 additions to the atomic weight of the initial 

 body of multiples of a common difference. 

 Starting with oxygen, whose atomic weight is 

 8, the next member, sulphur, has the atomic 

 weight 16, formed by the addition of the in- 

 crement 8 ; selenium has 40, corresponding to 

 the addition of four times this difference to the 

 weight of oxygen ; and tellurium 64, an incre- 

 ment of seven times the difference. In the 

 lithium and magnesium groups there are like 

 simple progressions. In the families of fluorine 

 and nitrogen he has established arithmetical 

 relations of a more complex order. 



A recalculation of atomic weights, based on 

 the determinations of Stas and other data, has 

 impelled Professor F. W. Clarke, following 

 Mallet and Dumas, to revive the abandoned 

 hypothesis of Prout, according to which the 

 atomic weights of all the elements are multi- 

 ples of the atomic weight of hydrogen. Among 

 the 65 determined elements when their atomic 

 weights are referred to that of oxygen, in order 

 to avoid the multiplication of the variation of 

 oxygen from Prout's hypothetical law, it is 

 found that 39, as calculated by Clarke, do not 

 vary more than 0*1 from exact multiples of the 

 atomic weight of hydrogen; and of the re- 

 maining 26, 3 are almost exact half-multiples ; 

 5 are rare or vaguely determined elements ; 2 

 are subject to the constant error from the oc- 

 clusion of oxygen, detected by Dumas in the 

 'case of silver, potassium, and iodine; 1, thal- 

 lium, is brought within the limit by a correc- 

 tion of Crookes's calculation ; 2, glucinum and 

 ytterbium, can also be brought by a recalcula- 

 tion within the limit; and 1, antimony, is al- 

 most an exact multiple of hydrogen, according 

 to a recent analysis of the bromide; for 4, 



mercury, chromium, vanadium, and gold, new 

 determinations are wanted ; and the remaining 

 8 are still subject to slight revision. Professor 

 Clarke concludes, then, that as three fourths of 

 the well-determined atomic weights agree with 

 Prout's hypothesis, the seeming exceptions 

 may be due to undetected constant errors, such 

 as have been brought recently to light in some 

 of the most familiar bodies in the entire list of 

 elements. 



Maximilien Gerber has sought to determine 

 common factors in the atomic weights of the 

 component members of each of the elemental 

 groups, and has determined empirically certain 

 common divisors in the several groups whose 

 multiples vary but slightly from the experi- 

 mentally-determined atomic weights. In the 

 group of mono-atomic elements the common 

 factor is 0*769. The alkaline metals, lithium, 

 sodium, potassium, rubidium, and ca3sium, 

 which combine with oxygen after the type 

 R a O, and with chlorine according to the for- 

 mula R Cl, have, excepting the last named, the 

 additional common factor 3. The non-metallic 

 halogens, fluorine, chlorine, bromine, and io- 

 dine, are another division of this class, and are 

 likewise multiples of 0-769. 



The atomic weight of hydrogen is related to 

 this number in the ratio 10:13, and that of 

 silver is an exact multiple. The di- and tetra- 

 atomic elements have the common divisor 1-995. 

 Oxygen has an atomic weight equal to eight 

 times this number, and the weights of sulphur, 

 selenium, and tellurium are multiples of that 

 of oxygen. 



The alkaline-earthy metals, magnesium, cal- 

 cium, and strontium, which have the combin- 

 ing formula RO, have the quadruple of the 

 original factor for a divisor ; but barium, which 

 belongs to the same group, does not. Carbon, 

 silicon, titanium, zirconium, and tin, have only 

 the one common factor. Mercury, molybde- 

 num, tungsten, and uranium, are also multiples 

 of this number. The tri- and penta-valent ele- 

 ments, the group of nitrogen, boron, etc., which 

 form a stable oxide of the type R 2 O 3 , and 

 chlorides of the types RCla or RC1 B , have 

 most of them the common factor 1 '559 in their 

 'atomic weights. The fourth and most numer- 

 ous class, combining into the oxides RO and 

 R 2 O 3 , have atomic weights which are approxi- 

 mate multiples of 1'245. Gerber's provisional 

 determination of common divisors is found to 

 agree with two recent corrections of atomic 

 weights : that of tellurium, which, as redeter- 

 mined by Will, is 127'8, a number which ac- 

 cords better with MendelejefFs scheme; and 

 that of glucinum, which, according to the find- 

 ings of Nilson and Petterson, should not be 

 classed among the diatomic alkaline-earthy 

 metals, as its oxide is of the type RsO 8 , as 

 originally established by Berzelius, and its 

 atomic weight must therefore be taken as 

 13-65. 



AUSTRALIA AND POLYNESIA. I. GEN- 

 ERAL STATISTICS. The area (in square kilo- 



