NATURE 



[May 3, 1888 



I lately discussed Murray's theory of coral formation with a 

 class of boys and girls (fourteen to sixteen years of age), and 

 th<y raised two questions which I am unable to answer, (i) If 

 sea water dissolves the coral near the surface at such a rate as to 

 form a lagoon, why does it not dissolve the limestone foundation 

 even more rapidly ? (2) After a reef has progressed a considerable 

 distance from the shore, and a channel of open water is formed 

 between, why should not the reef extend back again shoreward* ? 

 How could such a channel as exists between Australia and its 

 Great Barrier Reef ever have been kept open? These seem 

 to be valid and serious objections : will some expert be kind 

 enough to answer them ? Charles R. Dryer. 



Fort Wayne, Indiana, U.S.A., April 16. 



Density and Specific Gravity. 

 The point raised by Mr. dimming in last week's Nature 

 (vol. xxxvii. p. 584), as to the use of the words density and 

 specific gravity is, it seems to me, of some importance. For 

 many years pa-t I have, in my lectures, taken the law into my 

 own hands in this matter, and, defining density as the mass of 

 unit volume, I have defined specific gravity, in the way Mr. 

 Cumming suggests in the last paragraph of his letter, as the weight 

 of unit volume (or rather, lest I should cause any to offend 

 against the examiner, I have thus denned absolute specific gravity, 

 or specific gravity proper, and have pointed out that the defini- 

 tion commonly given was the definition of relative specific 

 gravity). We thus get the parallel relations — 



M = P V and W = .fV, 

 also 



W = gM and s = gp. 



Thus regarded, specific gravity is to density just what weight is 

 to mass. When force is expressed in absolute units of any 

 kind, specific gravity and density must of course have different 

 numerical values, just as weight and mass have. But in the very 

 large number of cases in which weights are the only forces that 

 have to be considered, and in which it is not needful to take 

 account of the small changes of weight dependent on changes of 

 geographical position, the local weight of the unit of mass may 

 be conveniently taken as the practical unit of force — that is, we 

 may take g= 1. In all such cases we have, numerically, 

 weight = mass, and specific gravity = density, though the idea 

 of weight is essent'ally different from that of mass, and the idea 

 of specific gravity from that of density. 



Of course, as Mr. dimming points out, when specific gravity 

 is defined as weight of unit volume, its numerical value for a 

 given substance depends on what is taken as unit of weight and 

 what as unit of volume. With the weight of 1 pound avoir- 

 dupois and the cubic foot as units, the specific gravity of water 

 becomes 62'5, and that of platinum I3I2'5, instead of r and 21 

 as given in the ordinary tables of (relative) specific gravities. 

 If, on the other hand, we ta! e as unit of weight the weight of 

 unit volume of the standard substance, as is done when weights 

 are expressed in grammes and volumes in cubic centimetres, or 

 weights in kilogrammes and volumes in litres, absolute specific 

 gravities and relative specific gravities become equal, and the 

 ordinary specific gravity tables can be used for practical purposes, 

 which is one of the great advantages to be gained by using the 

 metrical system of weights and measures. With any other 

 system, the numbers given in the tables require to be multiplied 

 by the specific gravity of water — that is, they must be translated 

 into absolute specific gravities — before they are of use for almost 

 any real calculation, such as oc urs either in experimental physics 

 or in engineering practice. For instance, we weigh a measured 

 length of copper wire and want to know its diameter, or we 

 weigh the quantity of mercury that fills a glass bulb of which we 

 require the capacity, or that fills a measured length of a tube of 

 which we require the bore ; or an engineer compares his pressure- 

 gauge against a mercury-manometer in order to convert its 

 indications into pounds-weight per square inch ; or he has to 

 calculate the pressure exerted by a brick wall so many feet high, 

 or the weight of a mass of rock of so many cubic feet. In all 

 these cases it is the absolute specific gravity that comes into 

 account ; it is no use to tell us that copper is 8*9 times as heavy 

 as water, and mercury 13-6 times as heavy, unless we are told 

 how heavy the unit volume of water itself is. 



I maintain, in short, that the weight of unit volume of a sub- 

 stance is a quantity of very great practical importance, for which 

 specific gravity is a very suitable name, whereas the ratio 

 usually defined as specific gravity is of little or no use outside 



examination questions, and that if it needs a name it should be 

 called relative density. 



Further, my experience is that the definition here advocated 

 presents considerable advantages from the point of view of 

 systematic teaching. G. Carey Foster. 



University College, London, April 21. 



Je crois que la notion de specific gravity donnee par M. 

 Cumming dans Nature du 19 avril (vol. xxxvii. p. 584) est de 

 nature a puzz'er les etudiants plus encore que la vraie definition 

 physique de la densite. 



La densite d'un corps est le rapport de sa masse a son volume — • 

 M 



Dans le systeme C.G. S. la densite doit done etre exprimee en 

 grammes masse par centimetre cube (voy. Everett, "Units and 

 Physical Constants "). Le poids specifique est le rapport du 

 poids d'un corps a son volume et devrait etre exprime, dans le 

 systeme C.G.S. en dynes par centimetre cube. Mais il y aurait 

 alors le grave inconvenient pratique a cette definition rigourcuse 

 que le poids specifique varierait avec ff> acceleration due a la 

 pesanteur, tandis que la densite resterait constante. 



La confusion provient de ce que le mot weight, comme le mot 

 poids en francais, s'applique indistinctement a la masse d'un 

 corps en grammes-masse et a la force qu'exerce la pesanteur sur 

 le corps exprimee en grammes. 



La solution logique est de supprimer le mot poids du langage, 

 a cause de son double sens, et de ne parler que de la masse ou 

 de la force exercee par la pesanteur, suivant que l'un ou l'autre 

 facteur intervient dans les calculs. 



En tout cas, exprimer le poids specifique en livres ou en 

 grammes est aussi absurde que d'exprimer les vitesses en metres, 

 et la puissance {pmver) d'une machine en ergs ou en foot-pounds. 

 1 .e respect de l'homogeneite des formules est la condition 

 essentielle des definitions des quantites physiques, et cette 

 homogeneite n'est pas respectee dans la definition donnee par 

 M. Cumming. E. Hospitalier. 



Paris, le 23 avril. 



The Ignition of Platinum in Different Gases. 



An abstract appeared a few weeks ago in Nature relating to 

 the " Occlusion of Gases by Platinum and their Expulsion by 

 Ignition," which induces me to mention some curious results 

 obtained by Mr. Lowndes and myself by the ignition of platinum 

 in different gases. We were led to the experiments by another 

 investigation on the behaviour of carbon at high temperatures 

 in various gases. We find that when a platinum wire is heated 

 to nearly melting by a current in an atmosphere of chlorine, the 

 walls of the glass vessel become covered with a yellow deposit, 

 which is insoluble in water, but dissolves in hydrochloric acid, 

 and then, after addition of a little nitric acid, gives all the re- 

 actions of platinic chloride. The yellow deposit is in fact 

 platinous chloride. At the same time the thick part of the 

 platinum wire conveying the current, and which was not heated 

 very highly, became incrusted with very fine long crystals of 

 platinum. Some of these were more than the sixteenth of an 

 inch in length, and apparently considerably more were located 

 on that end of the thick wire leading to the negative pole than 

 on the other. 



There was also a very decided but lambent flame playing 

 around the ignited and part of the cooler wire during the pas- 

 sage of the current. The arrangement used was a wide-necked 

 flask, stopped with a glass bulb, through which a delivery-tube 

 for the chlorine, and the two No. 12 platinum wires leading the 

 current, passed. The ignited parts of" the wire are little coils of 

 No. 24 wire separated by a I -inch piece of No. 12. On heat- 

 ing the flask externally up to the softening of the glass, the 

 appearance of a flame around the wire increased slightly. 



On repeating the experiment with bromine, very nearly the 

 same effects were observed. The amount of platinous bromide 

 was much less than in the case of the chloride, but the flame 

 appearance was very much more pronounced. On passing 

 chlorine into the bromine, so as to form chloride of bromine, 

 both the flame appearance and the action on the platinum were 

 largely increased. With iodine in the flask, vaporized by heat- 

 ing externally, little chemical action on the platinum was ob- 

 served, only the slightest deposit being formed of a platinum- 

 iodine compound on the glass; but, on passing chlorine into 

 this also, a still more vigorous action on the metal took place, 

 the dep:.sit containing only chlorine and platinum. The flame 



