460 



SCIENCE 



[Vol. LV, No. 1426 



Hg (+80) (2x1) (8x3) (6x3) (8x2) (6x3) (2) 



and (+80) (2x1) (8x3) (6x3) (8x3) (6x2) 

 Nt (+86) (2x1) (8x3) (6x3) (8x3) (6x3) or 



(+86) (2x1) (8x3) (6x3) (8x2) (6x3) (4x2) 

 These and similar formuliB for the other ele- 

 ments express very satisfaetorih' their known 

 chemical, physical and crj'stallogi-aphic prop- 

 erties. By applying this theory to crystal 

 structnres, it has been found possible to deter- 

 mine the arrangements of electrons in nearly 

 all crystals for which the arrangements of 

 atomic centers were already known, and also 

 the atomic and electronic structures in many 

 cases in which even the atomic marshalling 

 ivas previously unknown. These structures 

 furnish incontrovertible proof that this theory 

 of atomic structure, fundamentally and in 

 many of its details, is correct. 



An important part of the theory is the idea 

 that a single bond may be formed not only by 

 the attraction between two atoms, each of which 

 contains an unpaired electron in its valence 

 shell, the two single electrons forming a pair, 

 but also by the attraction of an atom containing 

 a "lone eleetronpair" — one not acting as a 

 bond — for another capable of holding on to 

 this lone pair. The following are typical re- 

 actions of this type : 



Zn++ + 4 : NH 



F : B 

 " : F 



H H 



= N N 2 



H^N : Co : NH^ 



N" 'N 



H N : Zn : NH 



In the last two cases the lone pairs of the 

 nitrogen atoms become bond pairs, assuming 

 positions at octahedron corners opposite the six 

 faces of the distorted cube of the cobalt kernel 

 and at tetrahedron corners opposite the four 

 previously unoccupied faces of the zinc kernel 

 octahedron (or, what is the same thing, oppo- 

 site the four faces of the zinc kernel tetra- 

 hedron). We thus have an entirely satisfac- 

 tory picture of Werner's "auxiliary valencies" 

 and "coordination numbers." In some cases 

 (e. g., in Ag(NH )*) all the faces of the kernel 

 polj'hedron are not occupied. (In the silver 

 iodide crystal, each silver kernel is surrounded 

 by four electronpairs at tetrahedron corners, 

 showing its true coordination numljer to be 

 four. ) 



This theory has not yet been applied to the 

 explanation of spectra; nor is it possible to 

 give the exact positions of the electrons in each 

 atom. These positions may in fact be merely 

 the centers or foci of electronic orbits. In 

 these and other respects the theory is stiU 

 incomplete. 



Maurice L. Huggins 



University of California 



A SIMPLE BUBBLING HYDROGEN 

 ELECTRODE 



The electrode described in this paper is the 

 result of an attempt by the writer to combine 

 the principles of the bubbling type of elec- 

 trode with simplicity of construction and the 

 necessity for only a small amount of solution. 

 That this has been accomplished, seems to be 

 apparent from a study of the accompanying 

 diagram and the behavior of the electix)de in 

 numerous tests. 



Four models sunilar to the one shown in the 

 diagram were constructed by the writer and 

 compared with each other and a Bailey elec- 

 trode. Various standard buifer solutions were 

 tested and it was found that all electrodes gave 

 results that agreed within .3 of a millivolt, 

 which was the limit of accuracy of the gal- 

 vanometer in the set up. 



While the models constructed by the writer 

 require only about 1.5 c.c. of solution, there 

 seems to be no good reason why, with pi'oper 



