AUCUST 2, 1906] 



NA rURE 



34: 



whenovi-T the liquid contents of the foam cells contracted 

 on solidification, or when the walls and the contents of 

 the foam cells contracted differently as they cooled. (6) By 

 the bounding surfaces of the doubly refracting crystals 

 (glacier-grains), which are differently orientated in neigh- 

 bouring foam cells, (c) On illumination with sunlight or 

 electric light, or on warming, when the doubly refracting 

 contents of the foam cells melt and are transformed into 

 singly refracting liquid, (i) By lens-shaped masses, foam 

 flakes or air bubbles, suspended in the foam walls, (c) By 

 the furrows, or network of lines on the solidified surface 

 formed by the intersection with that surface of the foam 

 walls in the interior of the solidified mass. (/) By polish- 

 ing or etching the natural or artificial surface, in cases 

 when the walls and the contents of the foam cells differ 

 in hardness or in the rapidity with which they are attacked 

 by chemical reagents. 



The surfaces "of solidified drops of pure molten metals. 

 show a network of straight lines or arcs of circles (usually 

 inclined to one another at: 120° or 00°), or foam walls with 

 embedded lens-shaped masses. This is so in the case of 

 gold, silver, platinum, palladium, iridium, indium, copper, 

 zinc, iron, nickel, cobalt, bismuth, sodium, potassium and 

 mercury. Similar phenomena are to be observed on the 

 surface of solidified drops of sulphur and selenium, or on 

 the surface of carbon which has been distilled with the 

 electric arc in a magnetic field, and deposited on the 

 kathode. 



The shapes of the bounding surfaces of molten metals, 

 and the circular arcs in the network of lines on the surface 

 of metals raised to red or white heat, show that these 

 bounding surfaces must be regarded, not as they have 

 hitherto been, viz. as crystalline faces, but as solidified 

 oily foam walls, which, as in the glacier-grains of ice, 

 enclose foam cells with contents differing from the walls. 

 Just as the glacier-grains of ice run together and enlarge 

 by the bursting of the foam walls, so also larger foam 

 cells with fewer foam walls are formed in metals heated 

 nearly to melting point. 



Pure molten metals after solidification exhibit on 

 artificial polished and etched surfaces n network of lines 

 or foam cells (similar to the glacier-grains of ice), which 

 are bounded by thin foam walls. Those thin foam walls 

 themselves contain still smaller foam cells, as is proved 

 by the visible lens-shaped masses embedded in them, and 

 the wave-like furrows on their surface, which are capable 

 in reflected light of giving diffraction colours like mother- 

 of-pearl. Tliis foam structure of pure metals when solidified 

 after fusion has been demonstrated in the case of bismuth, 

 cadmium, cobalt, copper, gold, iron, indium, iridium, 

 lead, manganese, mercury, nickel, palladium, platinum, 

 potassiimi, rhodium, sodium, tin, and zinc. 



Molten metals solidify on cooling to a liquid jelly, and 

 later to a solid jelly. The walls and contents of the foam 

 cells of such a jelly still consist of viscous liquid, i.e'. the 

 jelly itself is still liquid — like ice — at temperatures lower 

 than the melting points of the respective metals. The weld- 

 ing of two pieces of metal corresponds to the running 

 together of the cell walls and cell contents of two lumps 

 of jelly, or the regelation of ice. 



All the other substances in nature behave like these 

 metals. The soft, plastic condition, which all bodies 

 assume for a larger or smaller interval of temperature on 

 the transition from the solid to the liquid state, proves the 

 presence of jelly, i.e. of oily, visible or invisible foam walls, 

 over this interval of temperature. 



The heterogeneous oily liquid, which as solidification 

 occurs becomes visible in all substances in nature in the 

 form of thin foam walls of different surface tension, must 

 also appear as a thin liquid skin on the surface of solidify- 

 ing drops. This explains the variations in the measure- 

 ments of the surface tension of molten metals and salts, and 

 of liquids in general. 



The walls and contents of the foam cells consist of hetero- 

 geneous substance. That foreign matter in very small 

 quantities — i.'ioooooo per cent, and even less — does form 

 oily layers and foam walls in pure liquids is proved by 

 the author's observations on ice and benzene. Traces of 

 foreign matter (gases, carbon, metals, &c.) too small to 

 be shown in any other way are present even in the purest 



NO. 19 18, VOL. 74] 



liquids, and are sufficient to e.xplain the observed foam, 

 structure of all solidified substances in nature. 



June 28. — " On the Ultra-violet Spectrum of Ytterbium." 

 By Sir William Crookes, F.R.S. 



The rare earth, ytlerbia, was discovered in 1878 by 

 Marignac (Comptcs rendus, vol. Ixxxvii., p. 578). In 1880 

 Nilson (Bcr., vol. xii., p. 554), in purifying Marignac's 

 ylterbia, found that it contained another earth, which he 

 named scandia. Cleve, and more recently his daughter 

 Astrid Cleve, have worked much on ytterbia, and within 

 the last few years M. Urbain has taken up the subject, 

 and has succeeded in purifying ytterbia in larger quanti- 

 ties. During the author's own work on the fractionation 

 of the rare earths he also has prepared and worked with, 

 vtterbia. 



M. Urbain's ytterbia was prepared by the fractional 

 crystallisation of the ethyl-sulphates of crude gadolinite 

 earths {Comptes rendus, vol. cxxxii., p. 136). The sub- 

 sequent separation is by the fusing nitrate method. This 

 after twenty series of fusions gave in the least basic por- 

 tions a mixture of ytterbia and thoria, which are easily 

 separated by Wyrouboft and Vcrneuil's method. 



The examination for absorption bands in a strong solu- 

 tion is a fairly good test for an earth such as erbia and 

 thulia giving absorption spectra, but it is not so delicate 

 as an examination of the spark spectrum photographed 

 through a quartz train, for dominant lines, which most 

 elements show in some part of their spectrum. For 

 instance, the dominant lines of yttrium are at w-ave-lengths 

 36009, 3710-4, 3774-5, 4177-7. and 4375-1. The dominant 

 lines of erbium are at 3499-3, 3692-8, and 3906-5. They 

 are, however, not strong, and fortunately the absorption 

 bands of this element are striking and characteristic. The 

 spark spectrum of thulium has only been slightly examined 

 by the author, and he does not think it has any strong lines. 

 Its absorption spectrum, as with erbium, is a very character- 

 istic one. The spark spectrum of ytterbium has strong 

 dominant lines at 3289-5 .and 36944. Scandium has 

 dominant lines at 3572-7, 3614-0, 3630-9, 36429, and 

 4247-0. 



The author's photographs were taken with the quartz 

 apparatus already described, the spectrum of pure iron being 

 used as a standard. The ytterbium spark was taken from 

 a strong solution of the nitrate between platinum poles, 

 sufficient self-induction being introduced to eliminate nearly 

 all the air lines. The ytterbium, by this very severe spec- 

 trum test, is seen to be not absolutely free from impuri- 

 ties — thulium, copper, and calcium being present. Thulium 

 is seen by its lines at 30207, 3131-4, 34^5-2. .1441-6. 3462-4. 

 and 3848-2. Copper is seen by its dominant lines at 3247-7 

 and 3274-1, and calcium by its dominant lines at 3933-8 

 and 3968-6. 



The platinum lines which are present are easily recog- 

 nised, and are useful as an additional measure of identifi- 

 cation. Besides these, a number of fainter and indistinct 

 lines are seen. These may be due to ytterbium or to 

 traces of hitherto unrecognised impurities. 



The wave-lengths of all the recognisable lines of 

 vtterbium are given on the photograph, and also those of 

 thulium, calcium, and copper, but the platinum lines are 

 not marked. 



Paris. 



Academy of Sciences. July 16. — M. H. Poincare in th 



chair. — The absorption of nitrogen by organic substances, 

 determined at a distance under the influence of radio-active 

 materials : M. Berthelot. The action of air upon cellu- 

 lose in the presence of a radium salt has been studied ; 

 the effects are comparable with those produced by the 

 silent discharge. — A photometer specially designed for 

 measuring the circumsolar light. Its use during the total 

 eclipse of August 30. 1905 : H. Deslandres and A. 

 Bernard. The standard light used in the comparisons 

 was a small osniium lamp. Two diagrams are given show- 

 ing the arrangement of the photometer and telescope. 

 The apparatus was used at Burgos during the last total 

 eclipse, but the meteorological conditions were unfavour- 

 able. — Study of an apparatus designed by M. Lippmann for 

 the photographic measurement of right ascensions : W. 



