22 



SCIENCE 



[N. S. Vol. XXXIV. No. 862 



face by capillarity, at once begin to lose car- 

 bon dioxid, leaving a residue of sesquicar- 

 bonate; which will, without difficulty, dissolve 

 the humus of the soil, and act in other re- 

 spects precisely like the normal carbonate. 



In view of the fact that the air of the soil 

 always contains more carbonic dioxid than the 

 air, any sodic carbonate it contains in solu- 

 tion will inevitably suffer transformation 

 into bicarbonate to the extent to which car- 

 bonic acid happens to be found in the soil 

 under the existing conditions of vegetable 

 growth, temperature, moisture, and bacterial 

 action in the oxidation of organic matter. It 

 is a matter purely of seasonal accident; so 

 that, if a soil sample happens to be taken at a 

 time when carbon dioxid is abundantly form- 

 ing, the chemist may find in it exclusively 

 bicarbonate; while similar samples, taken a 

 few weeks afterwards, may uncfer the influ- 

 ence of aeration and drying, be found to con- 

 tain, in the main, the sesquicarbonate. 



I have therefore considered, and do now 

 consider, the determination of sodium, hicar- 

 honate in the soils as quite immaterial for 

 practical purposes, it being a variable and en- 

 tirely uncertain factor; and inasmuch as ulti- 

 mately the entire amount of sodic carbonates 

 may serve for the formation of sesquicarbon- 

 ate and the normal salt, I have thought best 

 to calculate the entire amount of these car- 

 bonates found to the latter salt, without ref- 

 erence to the other two. 



Of late, the official characterization of the 

 hydrocarbonate salt as no more harmful than 

 other white alkali, has led some chemists 

 analyzing irrigation waters to recommend for 

 that purpose, waters containing considerable 

 amounts of sodium bicarbonate. In this case, 

 the prospect of the accumulation of indefinite 

 amounts of black alkali in the soil irrigated 

 would be such a positive and inexcusable 

 detriment, that it seems high time to put an 

 end to the misleading statement which leads 

 chemists, as well as farmers, to expose lands 

 to serious injury. 



E. W. HiLGARD 



Beekelet, Cal., 

 Jime, 1911 ■ 



METALS ON METALS, WET 



It is generally agreed that the coefficient of 

 friction between metals (outside the physical 

 laboratory) is a very elusive quantity. It is 

 with the hope of suggesting some considera- 

 tions not mentioned by Professor HalP but 

 of great practical importance that the follow- 

 ing comments are offered. 



The difference in coefficient of friction be- 

 tween driving wheel and rail, whether the 

 latter is wet or dry, is relatively immaterial 

 so far as the effect of using sand to prevent 

 the slipping of the driving wheels is con- 

 cerned. Sand is used on dry as well as wet 

 rails, with a similar result in either case, 

 namely, to increase the coefficient of friction 

 between wheel and rail. The reason for this 

 is obvious, since the sand particles become 

 ground between the two surfaces, giving them, 

 in effect, the roughness which greatly in- 

 creases the coefficient of friction over that of 

 the unsanded smooth surfaces. 



It is also well known to railroad engineers 

 that a cleanly washed wet rail, as after a 

 heavy rain, is a " better " rail and is less 

 likely to result in slipping of wheels than a 

 perfectly dry rail. Of course a rail having 

 slimy water or any foreign matter, such as 

 coal dust, frost, etc., which can act as an 

 ungent, results in lowering the coefficient of 

 friction. It is because a wet rail is ordi- 

 narily greasy that sand is commonly used 

 when the rail is wet, whereas, a dry rail is 

 more apt to be gritty, due to the dust, etc. 

 There is therefore less necessity for using 

 sand. 



Referring to Professor Hall's paragraph 

 one, would not term " adhesion " ' apply to 

 the phenomena mentioned more accurately 

 than term " friction " ? A similar effect is 

 noticed when a thin film of water separates 

 two plates of glass. 



Professor Hall's conclusion would seem to 

 satisfactorily account for one phase of the 

 problem, but, as stated above, the phenomenon 



^ Science, May 19, 1911, p. 775. 

 - See DuBois, ' ' Mechanics of Engineering, ' ' 

 Vol. 1, p. 220. 



