September 8, 1898J 



NATURE 



451 



salt, strong brine, or with wet salt, provided that a piece of 

 linen is used as the rubber ; but if the finger be substituted for 

 the linen to rub on brine, a smell is observed with copper and 

 German-silver, this smell, however, being rather like that of 

 soda ; and whether dry salt, brine, or wet salt be rubbed on 

 aluminium, a smell is noticed if the finger be used as the 

 rubber, this smell being very marked in the case of the brine 

 or wet salt. Again, although even when linen soaked in brine, 

 or having wet salt on it, is used to rub tin, iron, or steel, a 

 faint smell is noticed, this is much increased when the finger is 

 substituted for the piece of linen. 



As a further illustration of the part played by the skin in 

 causing metallic smells, it may be mentioned that the explana- 

 tion of certain entirely contradictory results, which were ob- 

 tained in the early part of the investigation, when linen soaked 

 in strong brine was rubbed on aluminium, was ultimately traced 

 to one layer of moist linen of the thickness of a pocket- 

 handkerchief, allowing the finger to act through it, so that an 

 odour was sometimes noticed on rubbing aluminium with the 

 piece of linen soaked in brine. For it was found that when 

 two or more layers of the same linen soaked in the same brine 

 were employed to separate the finger from the aluminium during 

 the rubbing, no smell could be detected. 



From the preceding it seems that the smell in these cases is 

 evolved partly by contact with the finger, partly by the action 

 of the solution of salt, and partly by the rubbing of the solid 

 particles of salt against the metals. That the friction of solid 

 particles against metals is operative in evolving smells is also 

 illustrated by the smell noticed when iron is filed, or when alu- 

 minium, iron, or steel is cleaned with glass-paper or emery- 

 paper in the air. Indeed, the smell thus evolved by aluminium 

 Mrs. Ayrton finds particularly offensive. A slight smell is also 

 noticed if iron or steel be rubbed in the air with even a clean piece 

 of dry linen, and each specimen of the copper group, with the 

 exception of the phosphor-bronze, which was tried in this way, 

 gave rise to a faint, rather agreeable smell. No indication of 

 odour could, however, be thus produced with aluminium or zinc 

 when both the metals and the linen rubber were quite clean. It 

 should, however, be borne in mind that all these experiments, 

 where very slight smells are noticed, and especially when the 

 odour rapidly disappears on the cessation of the operation that 

 produced it, are attended with a certain amount of doubt, for 

 the linen rubber cannot be freed from the characteristic smell of 

 " clean linen," no matter how carefully it may be washed. 



Before, then, a metal can evolve a smell, chemical action 

 must apparently take place, for rubbing the metal probably frees 

 metallic particles, and facilitates the chemical action to which I 

 shall refer. All chemical actions, however, in which metals take 

 part do not produce smell ; for example, no smell but that of 

 soda, or of sugar, respectively, can be detected on rubbing any 

 single one of the series of metals that I have enumerated with a 

 lump of wet soda, or a lump of wet sugar, although chemical 

 action certainly takes place. Again, no metallic smell is 

 observable when dilute nitric acid is rubbed on copper, 

 German-silver, phosphor-bronze, tin, or zinc, although the 

 chemical action is very marked in the case of some of these 

 metals. Weak vinegar or a weak solution of ammonia are also 

 equally inoperative. On the other hand merely breathing on 

 brass, copper, iron, steel, or zinc, which has been rendered 

 practically odourless by cleaning, produces a very distinct smell, 

 while a very thin film of water placed on iron or steel evolves 

 a still stronger odour. Such a film, however, produces but little 

 effect with any of the metals except these two, and if the whole 

 series is lightly touched in succession with the tongue, the iron 

 and steel smell as strongly as when breathed on, the German- 

 silver more strongly than when breathed on, or covered with 

 a water-film, and the other metals hardly at all. 



Now, as regards the explanation of these metallic smells, 

 which have hitherto been attributed to the metals themselves. 

 This, I think, may be found in the odours produced when 

 the metals are rubbed with linen soaked in dilute sulphuric 

 acid. For here, apart from any contact of the metal with the 

 skin, the aluminium, tin, and zinc are found to smell alike ; 

 the copper group also smell alike ; and the iron and steel 

 give rise to the characteristic "iron" smell, which, in this 

 case, can be detected some feet away. Now, it is known 

 that when hydrogen is evolved by the action of sulphuric 

 acid on iron, the gas has a very unpleasant smell, and this. 

 Dr. Tilden tells me, is due to the presence of hydrocarbons, 

 and especially of paraffin. I have been, therefore, led to think 



NO. 1506, VOL. 58] 



that the smell of iron or steel when held in the hand is really 

 due to the hydrocarbons to which this operation gives rise ; and 

 it is probable that no metallic particles, even in the form of 

 vapour, reach the nose or even leave the metal. Hence, although 

 smell may not, like sound, be propagated by vibration, it seems 

 probable that particles of the metal with which we have been 

 accustomed to associate the particular smell may no more come 

 into contact with the olfactorj' nerves than a sounding musical 

 instrument strikes against the drum of the ear. 



And the same sort of result may occur when a metal is rubbed, 

 for, although in that case particles may very likely be detached, 

 it seems possible that the function of these metallic particles 

 may be to act on the moisture of the air, and liberate 

 hydrogen similarly contaminated ; and that in this case also 

 it is the impurities which produce the smell, and not the 

 particles of the metal with which we have been accustomed 

 to associate it. 



This view I put forward tentatively, and to further elucidate 

 the matter I am about to begin a series of smell tests in various 

 gases, artificially dried, with metals as pure as can be obtained. 



I next come to the diffusion of smell. From the experience 

 we have of the considerable distance at which a good nose can 

 detect a smell, and the quickness with which the opening of a 

 bottle of scent, for example, can be detected at a distance, I 

 imagined that tubes not less than 15 or 20 feet in length would 

 be required for ascertaining, even roughly, the velocity at which 

 a smell travels. But experiment soon showed, that when the 

 space through which a smell had to pass was screened from 

 draughts, it diffused with surprising slowness, and that feet could 

 be replaced by inches in deciding on the lengths of the tubes to 

 be used. These are made of glass, which is relatively easy to 

 free from remanent smells. 



When the room and tube had been freed from smell by strong 

 currents of air blown through them, the tube was corked up at 

 one end and taken outside to have another cork, to which was 

 attached some odoriferous substance inserted at the other end. 

 The tube was now brought back to the odourless room, and 

 placed in a fixed horizontal or vertical position, and the unscented 

 stopper was withdrawn. As a rule, immediately after the 

 removal of the stopper, a smell was observed, which h'd been 

 transmitted very quickly through the tube by the act of corking 

 up the other end with the stopper carrying the odoriferous 

 material. This first whiff, however, lasted only a very short 

 time, and then a long period elapsed before any further smelk 

 could be detected at the free end of the tube, whether that end' 

 was left open or closed between times. Finally, however, after, 

 for example, about eighteen minutes in the case of a three- foot 

 horizontal tube, having a large cotton-wool sponge saturated 

 with oil of limes attached to one cork, the smell became definite 

 and recognisable. 



It would, therefore, appear that the passage of smell is gener- 

 ally far more due to the actual motion of the air containing it 

 than to the diffusion of the odoriferous substance through the air. 

 And, as a striking illustration of this, the following is interest- 

 ing : — After the stopper had been in contact with the odoriferous 

 substance for some time, it, of course, acquired a smell itself, 

 which gradually spread in the room in which the experiment was 

 made. And although this smell was due simply to the ex|x>sed 

 part of the stopper, while the air inside the tube was at one end 

 m contact with a mass of the odoriferous subtance itself, the only 

 place where the smell could not be detected during the course of 

 the experiment was the space inside the open end of the glass 

 tube. And, what seemed very surprising, it was found neces- 

 sary, in several cases, to blow air through the room to clear out 

 the smell which emanated from the outside of the stopper before 

 the smell coming along the tube from the mass of odoriferous 

 substance which was inside it at the other end could be detected. 

 A further proof of the important part played by the motion of 

 the air in diffusing smell was the fact that a strong smell at the 

 free end of the tube could at any time be caused by merely 

 loosening the stopper to which the scented sponge was attached ; 

 for sniffing at the free end then made a draught through the tube 

 which brought the scent with it. 



Further, although the glass tubes were coated outside with a 

 thick layer of non-heat-conducting material, so as to check the 

 formation of convection currents, due to difference in the inside 

 and outside temperature, caused by handling, the rate of travel 

 of a smell from a given odoriferous material was found to be 

 much quicker when the tube was vertical than when it was 

 horizontal. But this, I am inclined to think, may have been 



