November 7, 1907 J 



NATURE 



mechanics, but no longer profess to be able to 

 describe, detail by detail, the ultimate moving ele- 

 ments and motions that underlie these phenomena. 



In the second part of his book the author seeks to 

 show that the salient divergences between the schools 

 simply mask the essential congruity of their views. 

 All physicists admit — in whatever idiom they may 

 describe them — the same ultimate objective data ; 

 while even if their hypotheses are only methodological 

 instruments of organisation and discovery, it must be 

 recognised that the science presents in the different 

 schools a real though not obvious unity of develop- 

 ment T. P. N. 



How to tell the Birds from the Flowers: a Manual 

 of Flornithology for Beginners. Verses and illustra- 

 tions. By Prof. R. W. Wood. Pp. 28. (San 

 Francisco and New York : Paul Elder and Com- 

 pany, n.d.) Price 50 cents net, or in cat-bird 

 cambric, 75 cents net. 

 It will come somewhat as a surprise to those of our 

 readers who know Prof. Wood only as a physicist to 

 learn that the present booklet contains nothing but 

 quaint illustrations and jest in verse. The volume is 

 obviously a satire directed against the sentimental 

 nalure-study literature which sometimes masquerades 

 as scientific teaching, particularly in the United 

 States. 



LETTERS TO THE EDITOR. 

 [The Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Winding of Rivers in Plains. 



.'\ CURIOUS obsession as to a matter of fact, to which 

 everyone is more or less liable when obfuscated by an 

 erroneous theory, has recently been noticed by mc in some 

 geological boolcs, e.g. in Le Conte's " A Compend of 

 Geology," and in Tyndall's "Glaciers of the .-VIps." I 

 noticed it first in my late colleague Prof. Watts 's recent 

 little text-book of geology ; but, indeed, I have not found 

 any book of the kind quite clear and correct on the 

 subject. 



The statement is clearly made and illustrated by a figure 

 that the flow of a bending river is most rapid on the 

 outer side, where its banks are concave ; and the well- 

 known scouring or excavating action which a stream 

 exerts on this bank is then attributed to this imaginary 

 more rapid flow. 



But the fact is that the flow is most rapid on the inner 

 or sediment-depositing side of the bend, and Prof. James 

 Thomson showed in 1S76, in a well-known communication 

 to the Glasgow meeting of the British Association — when 

 he exhibited a model, confirming calculations previously 

 made by himself — that the excavating action of a river is 

 not due to the direct scouring action of the main stream at 

 all. The explanation which he gave v^as virtually as 

 follows : — 

 The rapid flow on the inner and strongly curved side of 

 the bend piles up the water on the outer side by centri- 

 fugal force, so that near the concave bank it is nearly 

 stationary, but elev'ated ; its energy there is potential, not 

 kinetic. Now if the rapidity of flow were uniform from 

 top to bottom the slope would be in equilibrium ; but owing 

 to the retardation of the bed the flow near the bottom is 

 slower, and there is not nearly so much centrifugal force 

 exerted down below ; wherefore the piled-up water is con- 

 tinuously returning from upper to lower level, that is, 

 from the concave to the convex bank, as an undercurrent, 

 almost at right angles to the main stream, bringing with 

 it, by its undertow, silt and solid matter, which it deposits 

 near the inner side of the bend ; thus constantly increasing 

 ats own sinuosity in the well-known way. 



The stream itself, combining a progressive with a lateral 



NO. 1984, VOL. yy] 



circulating motion, may be said to screw itself like a 

 corkscrew round a bend : and it is the lateral circulation 

 which shifts the bed. 



So much for streams, now for glaciers. Prof. Tyndall, 

 as is well known, took careful measurements of the flow 

 of glaciers, and finding that their line of quickest motion 

 was more sinuous than the glacier bed, said that this was 

 another analogy lu the flow of a river. 



There, however, he was in error. The line of most 

 rapid flow of a river is less sinuous than the river itself. 

 The water flows round the bend somewhat as it would 

 flow in a vertical columnar vortex ; most rapid on the 

 inside, and almost stationary or even retrograding on the 

 outside of some bends. If ice flows otherwise — and I have 

 no reason whatever to doubt Tyndall's measurements — it 

 must be because the rate of change of momentum of so 

 slow a motion, compared with its lateral stiffness, is very 

 small ; so that we might certainly anticipate that the laws 

 of its flow would be in many respects different from — 

 though also in some respects singularly like — those of a 

 liquid of but small viscosity. Probably it obeys exactly 

 the laws of an extremely viscous liquid the viscosity of 

 which could be specified. The flow cannot be much 

 governed by inertia, as that of water is. 



But I know that glacier motion is a thorny subject upon 

 which I have no desire to tread. I would not be under- 

 stood as making any assertion concerning it, but merely 

 throw out a hint. 



As to winding rivers, however, the matter is fairly 

 simple ; and the writers of geological and geographical 

 text-books may easily amend some incautious though 

 natural statements as to matters of fact, which they some- 

 times illustrate by erroneous diagrams. 



OuvER Lodge. 



Recalculation of Atomic Weights. 



During the last few years our knowledge of the accurate 

 atomic weights of the elements silver, sodium, potassium, 

 chlorine, and bromine has been greatly extended by the 

 masterly researches of T. W. Richards and his colleagues. 

 At present, however, there is no really trustworthy value 

 for the ratio of silver to o.xygen, and a satisfactory value 

 for nitrogen has only just been obtained by Gray and by 

 Guye. 



While reading an account of the determination of the 

 ratio Ag : AgNO,, it occurred to me that this result, 

 together with others previously obtained by Richards, 

 afforded a means of calculating absolutely the atomic 

 weights of the above-mentioned elements in terms of 

 oxygen. The following values are available : — 



(i) A5 : KCI = 100 : 69'io73 o'ooo4=w 



(2) AgCl : KCI = 100 : 52011S O'ooo4=jr 



(3) Ag : AgN03= 100 : 157-479 O'ooi =m 



(4) Ag : AgCl = 100 : 32'867 O'ooo5=j/ 



(5) N0O5 : K20 = ioo : 87'232 o-oo2?=2 



We have thus five simultaneous equations, from four of 

 which the four unknown quantities Ag, K, CI, and N can 

 be calculated in terms of O. Neglecting 2, and putting 

 0=16, I worked out the atomic weight of N, and was 

 intensely surprised to find the value 13-940. 



Now it is a well-known fact that the final results of 

 an " indirect analysis " such as the above may be greatly 

 influenced by a small experimental error, and so I pro- 

 ceeded to estimate the maximum effect which could thus 

 be produced. Putting (69-1073 -I- v)Ag=ioo KCI, where v 

 is the error, instead of the first equation above, and 

 similar expressions for the last three, I obtained the 

 formula 



2422'o8 -f- eoot" - 6007 - 33624; - 2870 

 2779'94 -20oi'-f 200j'-j- I74JW-1- 1153 



From this it is evident that, if v is made positive and 

 the other three quantities negative, the numerator will be 

 increased and the denominator diminished, both these 

 facts tending to raise the value of N. Putting for the 

 symbols their values given in the table of errors, the 

 following result is obtained : — 



N = - 



:i6. 



■2779-94 -0-58 



= i395i- 



