May 14, 1908J 



NATURE 



of 55 years, in a plane inclined 1° 34' to the ecliptic ; the 

 n.^xt return of which it was impossible to observe from the 

 earth's position; which about 1779, August 23, approaclied 

 Jupiter within o-oi of the earth's mean distance from the 

 sun ; and which has not been again observed. 



At that distance Jupiter's attraction exceeds that of the 

 sun in the proportion of at least 200 to i, and the distance 

 from one of the satellites may have been very small. In 

 this way it is not altogether impossible that the comet may 

 have been diverted into an elliptic orbit round Jupiter, and 

 a retrograde motion round the planet would be as likely 

 as a direct motion to ensue. The intervention of a satel- 

 lite is essential, and this, combined with the observed large 

 angular distance from Jupiter in one part of its orbit, 

 implies a large eccentricity. Should further observations 

 reveal a moderate eccentricity, the impossibility of identity 

 between this satellite and Lexell's comet will be proved. 



The Athenjeum, May 6. George Forbes. 



The Corrosion of Iron and Steel. 



In' a letter to Nature of April 16, Dr. Frank Clowes 

 states his bdief that pure lead will dissolve to a slight 

 extent in perfectly pure water " as iron did in the experi- 

 (iients made by Whitney and repeated by the American 

 investigators, when they brought iron into contact with 

 water under conditions similar to those which I had 

 secured." 



The American investigators here referred to are W. H. 

 Walker, Cederholm, and Bent, who have recently studied 

 the process of the rusting of iron (Journ. ,\mer. Chem. 

 Soc, 1907, .xxix., p. 1251), and agree with Whitney in 

 regarding the whole subject as an electrochemical one. 



With reference to the behaviour of lead in contact wiili 

 water, I have nothing to say, for although the study of 

 the corrosion of metals has been pursued by me for several 

 nionths, my observations have not as yet been extended 

 to lead. I wish to point out, however, that if this metal 

 does dissolve to an appreciable extent in water as Dr. 

 Clowes suggests, then the analogy adopted by this gentle- 

 man is open to criticism. I have made a very careful 

 study of the process of the rusting of iron, and mv results 

 point conclusively to the fact that it is primarily tiie result 

 of acid attack. These results will, I hope, shortly appear 

 in print ; it is unnecessary, therefore, to enter into any 

 detail here. Nevertheless, one or two general remarks may 

 prove of interest. 



Walker and his co-workers found that if water in which 

 iron has been boiled is concentrated to a few drops, tlic 

 presence of iron can be detected by chemical means. It 

 is clear that if the concentration of the iron in solution can 

 t)e increased by the evaporation of the water, as these 

 authors say, the same result should be obtained by pro- 

 tracted boiling of the water and iron, keeping the volume 

 of the former constant by using a reflux condenser. 



My experiments show, however, that such is not the 

 case. The presence of traces of iron discovered by Walker, 

 Cederholm, and Bent is not, therefore, to be ascribed to 

 the action of pure water alone, but to that of traces of 

 dissolved carbon dioxide. 



In a very interesting series of investigations, Leduc 

 iComftes rendus, 1906, cxlii., p. 149) has recently shown 

 that all the dissolved air cannot be expelled from water 

 by the mere process of boiling. He further calculates that 

 at least i c.c. of gas remains in a litre of water even 

 after thorough boiling. Since carbon dioxide is not only 

 much more soluble than oxygen and nitrogen, but also 

 combines with the water to form carbonic acid, it is not 

 unreasonable to suppose that a considerable percentage of 

 this residual gas is carbon dioxide. This will effect the 

 solution of a trace of iron, .\lthough this quantity may 

 bi' too minute to detect at once by chemical means, its 

 concentration can be greatly increased by evaporating to 

 small bulk, when the characteristic reactions for iron may 

 l.>e obtained On the other hand, protracted boiling with 

 leflux condenser cannot increase the concentration of the 

 iron, since the amount of carbon dioxide remains the same. 

 The extreme difficulty of removing all traces of this gas 

 from water is not generally realised by chemists, and the 

 precautions adopted by Walker and his co-w'orkers were 

 not suiificiently refined — hence their results. .\ single 



NO. 201 1, VOL. 78] 



molecule of carbonic acid is sufficient, theoretically, to 

 cause the corrosion of iron (see my remarks. Nature, 

 September 27, 1906), and I have shown that in practice 

 a very few molecules are operative. 



There can be no reasonable doubt, therefore, that the 

 electrolytic theory is wholly inadequate. 



In a letter to Nature of October 11, 1906, Mr. Richard- 

 son asks if the rusting of iron is not caused by bacterial 

 agency. As no reply has been given, it may not be out 

 of place to deal with the question here. The suggestion 

 is not new. Schorler, Beythien, Adler, Raumer, and 

 others have directed attention to a bacterium, Gallionella 

 femiginea, which obtains its life energy by oxidising 

 ferrous carbonate or organic ferrous salts with the pre- 

 cipitation of rust. But no organism has been found capable 

 of feeding on metallic iron. Indeed, if one such were 

 found, it would have to effect the solution of the iron by 

 some presumably acid secretion, and this brings us back 

 to the acid theory of rusting. 



J. Newton Friend. 



Fault Lines in the Atlantic. 



In Prof. J. Milne's discourse at the Royal Institution 

 which appeared in Nature of .April 23 is given an interest- 

 ing map on p. 593 showing the folds and probable direc- 

 tion of fault lines in the .Atlantic. In that map is shown 

 the mid-.Atlantic " rise " extending to about 40° S. The 

 map, however, would have been more interesting had Prof. 

 Milne included in it the recent bathymetrical researches of 

 the Scotia, which w'ere described in a preliminary paper in 

 the Scottish Geographical Magazine in .August, 1905. Here 

 it is shown that the Scottish expedition sounded out this 

 " rise " for a thousand miles further south than Prof. 

 Milne has marked it, and also that that " rise " has prob- 

 able connection with another running in a more or less 

 E.N.E. and W.S.W. direction from the south of South 

 America through South Georgia and the Sandwich Group, 

 and also from Graham Land, .Antarctica, through the 

 .South Orkneys and the Sandvifich Group, thence passing 

 eastward through Bouvet Island to Madagascar and the 

 east coast of Africa. The extension of the mid-Atlantic 

 " rise " continues the reflection of the South .American 

 continent, and again the transverse " rise " reflects the 

 hypothetical and known portions of the coast-line of 

 -Antarctica. Wm. S. Bruce. 



The Pollination of the Olive. 



.As Knuth's great work on flower pollination, compiled 

 from all available sources, says nothing whatever about 

 the olive, I recently asked Prof. J. E. Coit, of Arizona, to 

 look out for insects upon the flowers. He carefully 

 examined many olive trees in flower at Tucson, and did 

 not find a single insect at the flowers, with the exception 

 of a thrips (apparentlv Euthrips occidcntalis, Pergande), 

 which occurred in great numbers. He kindly sent me some 

 twigs with the Euthrips upon them, and I -was able to 

 observe that these insects were profusely dusted with 

 pollen. No bees were seen at all. Prof. Coit adds : — 

 " Olive pollen is formed and shed in such enormous quanti- 

 ties that I think the wind among the branches is the chief 

 agent in pollination. If you jar a large branch of olive 

 while it is in full bloom, a perfect cloud of green pollen 

 will be seen to float away on the breeze." 



Arizona, however, is not the original home of the olive, 

 and the purpose of this note is to suggest that those who 

 have an opportunity to see the plant in bloom in Mediter- 

 ranean and other countries should make some observations. 



T. D. .A. Cockerell. 



University of Colorado, May 2c 



The Coloration of Birds' Eggs. 



I SHOULD be verv grateful if anyone could tell me 

 whether there is anything known with regard to the egg- 

 colour of birds. Is there here, by any chance, a field for 

 investigation which may throw light upon the so-called 

 Mendelian phenomena? Or is there yet any explanation 



of how the egg becomes coloured, or wh\ 

 66 Hallam Street, W 



R. L. Leslie. 



