September 15, 192 1] 



NATURE 



lOI 



bered. We halted one golden evening at the little 

 inn of Petronell, and praised the wholesome quiet of 

 the spot. "Yes," replied the ostler, "Marcus Aurelius 

 loved it well." It was not until the next morning 

 that we saw the ruins of Carnuntum by the road. 



As an example of European study we may note 

 that Fritz Machatschek (vol. 60, pp. 235 and 274) 

 describes the southern flanks of the Erzgebirge. This 

 chain must no longer be regarded as an Armorican 

 anticlinal ridge, but has been determined by succes- 

 sive bulgings of the floor from Lower Oligocene times 

 onwards. Terrace-regions on the south represent 

 portions of the main gently curved surface that have 

 been lowered by faulting, and the well known vol- 

 canoes that produce such variety in the landscape 

 are manifestations of earth-movements that con- 

 tinued into the Ouartary (Quaternary) era. The 

 evidence of such movements is seen in the stages by 

 which the Eger \'^alley has been deepened. To turn 

 to farther fields, economists and politicians will find 

 much of interest in H. Leiter's review (vol. 62, pp. 

 A14 and XII,, 1910) of the present railway-system of 

 Africa. The unfinished Cape-to-Cairo route now 

 offers few attractions in view of the proposed cross- 

 cut from Bukama, its present terminus, to the Congo 

 mouth, a point quickly reached by sea from England. 

 The numerous recent developments from the coasts 



of the Mediterranean, the Red Sea, and the Indian 

 Ocean are described, and the dry details soon link 

 themselves together as a romance. '" Un tourbillon 

 de sable," we once said to our guide, looking south 

 over the Sahara, where a cloud rose among the 

 dunes. "Pardon, monsieur," he replied, "c'est le 

 chemin de fer. " The railway was going forward to 

 Touggourt. 



The comprehensive paper by C. Diener on " Die 

 Groszformen der Erdoberflache " (vol. 58, p. 329, 

 19 15) concludes that continental blocks and ocean- 

 basins have shown permanence, at any rate since 

 Permian times. Such work takes us into wider fields 

 than Eugen Oberhummer's interesting discussion of 

 imperialistic policies (vol. 63, p. 101, 1920) or H. 

 Hassinger's well reasoned attempt to define the boun- 

 daries and the qualities of a Middle European region 

 (vol. 60, p. 437, 1917). Both these essays in political 

 geography provide much food for thought, and Has- 

 singer's remark on the mixture of races between the 

 Straits of Dover and the Black Sea is worth quoting : 

 " Reiner Nationalstaat und Panslawismus sihd Mitt^l- 

 europa fremd und wirken hier unnatiirlich.^' 



More illustrations are needed for papers dealing 

 with natural features ; but this defect must not be 

 charged against those who have done their best to 

 advance knowledge through the war-time. 



Valence. 



IN an article on "Types of Valence " in Science for 

 July 22 Dr. Irving Langmuir develops still further 

 the views on this subject which are associated with 

 his name. He points out that the term "valence" 

 has been used to describe (i) positive valence, or the 

 number of electrons an atom can give up ; (2) nega- 

 tive valence, or the number of electrons an atom can 

 take up; and (3) covalence, or the number of electrons 

 an atom can share with its neighbours. He brackets 

 he first two under the name of "electrovalence," and 

 iccounts for them by the tendency to form a complete 

 layer of 2, 8, 8, 18, 18, or 32 electrons (First postu- 

 late). The outer incomplete layer of electronegative 

 elements, such as oxygen and the halogens, which 

 tend to take up electrons, is distinguished as a 

 sheath, and this term is used even when the atom has 

 taken up electrons, so that the outer layer has be- 

 come for the time a' completed sheath in the nega- 

 tively charged ion. The same term is also used in 

 the case of the electropositive elements to describe 

 the small excess of electrons which are lost when 

 the atom becomes a positively charged ion. 



The simplest complete layer consists of two electrons 

 (e.g. in He or H+) in close proximity to a nucleus; 

 such a pair of electrons is called a duplet, and this 

 term is extended to include any pair of electrons which 

 is rendered stable by its proximity to one or more 

 positive charges. The third type of valence is then 

 controlled by the second postulate that " two atoms 

 may be coupled together by one or more duplets held 

 in common by the completed sheaths of the atoms." 



Since each duplet eliminates two electrons as com- 

 pared with those required to form a pair of complete 

 sheaths, it follows that the algebraic sum of the 

 electrovalences and covalences for all the atoms in 

 any complete compound (i.e. a compound in which 

 the sheaths of all the atoms are complete) is zero. 

 Complete compounds without covalence include NaCl, 

 BaBr,, K.S ; MgO, BN, ALO, : A1F„ PCI,, SF^, etc. 

 In these last compounds the halogens have seven elec- 

 trons on the sheath of the atom, and therefore an 

 excess of seven positive charges on the "'kernel " (i.e. 



NO. 2707. VOL. 1 08] 



the atom minus the sheath), which creates a very 

 strong attraction and enables these atoms to drag 

 away as many as five electrons from an atom of 

 phosphorus and six from an atom of sulphur. When 

 no positive atom is present the electrovalence of everv 

 atom must be negative ; no atorh can lose electrons, 

 and only covalence can exist. In this case the ordinary 

 use of structural formulae is legitimate ; when elec- 

 trons are transferred instead of being shared it is not. 

 Since the sheaths of atoms of atomic number less 

 than about 25 never contain more than eight electrons, 

 the covalence of these atoms can never exceed four. 

 With heavier atoms larger covalences mav occur occa- 

 sionally, e.g. in Ni(CO)„ Fe{CO)„ and ■Mo(CO), the 

 number of atoms in the sheath of the metallic atom 

 is 10, 8, and 6, or 8, 10, and 12 less than the number 

 required to make a complete sheath of 18 electrons. 

 ^^'e therefore have negative valences of 8, 10, and 12, 

 as indicated by the formulae of the three carbonyls. 



Incomplete compounds, containing atoms with in- 

 complete sheaths, are particularly abundant in inter- 

 metallic compounds where only electropositive atoms 

 are present, each with a sheath very far from com- 

 plete. Thus in the two long periods 18 electron? are 

 needed to complete the sheath, whilst, as a rule, onlv 



4 can be lost (maximum positive valencv equals 4) 

 in order to remove it ; there are therefore 14 elements 

 which are compelled to retain electrons in the outer 

 layer, and very few of these can complete their 

 sheaths. These elements are, therefore, exclusivelv 

 metallic in character, and even their compounds with 

 electronegative elements usually contain loose electrons 

 in incomplete sheaths and often exhibit metallic con- 

 ductivity. The small and irregular positive valences 

 of the transition-elements of the two long periods 

 depend on the fact that their sheaths contain from 



5 to 13 electrons, of which, as a rule, onlv 2, 3, or 4 

 can be detached to make a kathion. 



The article is ingenious and suggestive, and repre- 

 sents a distinct advance in the process of explaining 

 whv and how combination between atoms takes place. 



T. M. L. 



