486 



SCIENCE. 



of similar implements does not seiTe the purpose of 

 ethnological collections. From a collection of string 

 instruments, flutes, or drums of ' savage ' tribes and 

 the modern orchestra, we cannot derive any conclu- 

 sion but that similar means have been ajsplied by all 

 peoples to make music. The character of their music, 

 the only object worth studying, which determines 

 the form of the instruments, cannot be understood 

 from the single instrument, but requires a complete 

 collection of the single tribe. Here, however, it can 

 be seen that each ethnological collection affords only 

 very fragmentary instruction ; that its real use is 

 only to illustrate descriptions of the tribes. For a 

 study of native art and its development, they are in- 

 dispensable. For this i3uri30se, duplicates, of which 

 the superficial visitor of ethnological museums fre- 

 quently complains, are absolutely necessary. They 

 are the only means of determining what is character- 

 istic of a tribe, and what is merely incidental. 



Mason's method takes a place in ethnology similar 

 to the former ' comparing method ' in geography. 

 A mere comparison of forms cannot lead to useful 

 results, though it may be a successful method of 

 finding problems that will further the progress of 

 science. The thorough study must refer to the his- 

 tory and development of the individual form, and 

 hence jsroceed to more general phenomena. 



Db. Feanz Boas. 

 New York, May 13. 



Explosions in coal-mines. 



In Science for May 6, is a review of the report of 

 the Atkinsons on explosions in coalmines. One or 

 two statements therein seem to convey an erroneous 

 impression ; notably, " At the working faces the 

 dust is not often a serious evil." and, under remedial 

 rneasures, that " watering the roadways ... is of 

 little avail as a means of preventing explosions, since 

 the upper dust in every instance is left undisturbed." 



The first quotation is manifestly an error, as dust- 

 explosions can generally be traced to the firing of the 

 dust in the working faces by blown-oub shots, espe- 

 cially when such shots react against a tamping of 

 coal-slack. In the main body of the article the argu 

 ment seems to be in favor of the dust in the gang- 

 ways as the proximate cause of explosion, while it is 

 but the ultimate cause. The dust formed by cutting 

 or breaking down coal has very little to do with the 

 formation of an explosive mixture, because it is not 

 impalpable enough. As the article states, the gang- 

 way dust is ground to an impalpable powder, and 

 carried away by the air ; but such dust would do 

 little hai-m, did the return currents through the 

 working places not lose their velocity and depo=iit 

 this impalpable dust on the walls nearest the face, 

 from the fact that the ventilating currents must 

 sweep the faces free from smoke and foul air. In 

 the only known American accident due to dust (the 

 Pocahontas explosion), there was little evidence of 

 initial explosive force along the gangways, but in 

 the headings of dusty rooms there occurred a series 

 of explosions that made the fact evident that fine 

 dust in headings is the cause of so many accidents. 



The Prussian commission showed that a certain 

 percentage of volatile matter was necessary for an 

 explosive dust, and experiments made at this place 

 show that the temperature at which coals give up 

 their volatile ingredients vary ; so that a blown-out 

 shot, from a hole tamped with coal slack, projected 



into a chamber whose walls are thic^ 

 with fine dust, will have its flame pmloij^, 

 dust of the tamping and the gases from tho, 

 and the case will be analogous to those stateu 

 Professor Abel, where a small admixture of gt 

 would render inert dusts explosive. 



Finally, it has been found that watering dusty 

 roads with brine at intervals of thirty days made the 

 dusts less ready to rise in clouds, and stopped the 

 formation of ' upper dust.' But the best way of all 

 is to keep the roads clean by ballasting them at the' 

 outset with rock or shale free from carbon, and by 

 picking off the coal shaken from cars. In the Poca- 

 hontas accident the explosive phenomena ceased as 

 soon as the current left the region where the roads 

 were ballasted with coal-slack, and the action in the 

 parts ballasted with slate was a simple burning of 

 the dust brought there by the current. In spite of 

 an explosive coal, the parts of that mine last men- 

 tioned were comparatively free from dust, and the 

 props nearly all standing, while in the former por- 

 tion there was explosive action in nearly every work- 

 ing place, so that the tracks were torn to pieces and 

 the props down. In mines of this nature, black 

 powder should be avoided, and the coal should be 

 wedged, or, if the coal must be shattered, the dyna- 

 mite cartridge with water casing can be used with 

 impunity. Edward H. Williams, Jr. 



Lehigh univ., May 13. 



"Water-filtration. 



It may be of interest to notice in connection with 

 your note on the results of Dr. Swarts's experiments 

 on the relation of water-filtration to bacterial devel- 

 opment, that Dr. J. H. M. Munro, in his experiments 

 on the nitrification of well-waters, discovered and 

 called attention to the fact that a well-water nitrified 

 more rapidly after filtration through a Lijiscombe's 

 charcoal filter in common use, than did an unfiltered 

 sample of the same water (Journ. chem. soc, 1886, 

 p. 666). William Feeae. 



State college, Penn., May 16. 



The fact that an increase of micro-organisms would 

 take place in a filter constantly in use, had already 

 been demonstrated by Percj' Fraukland, England, 

 and by several German investigators. 



The series of experiments conducted by myself 

 differed from theirs merely in using the filters found 

 upon the local market, and in imitating as near as 

 possible their use in ordinary family water-supply ; 

 cleansing in the first use by allowing the sujDply to 

 pass through the filter to wash away detritus, and 

 not by absolute sterilization, as in usual laboratory 

 experiments. The filters used were variously packed 

 with bone or animal charcoal, quartz, the two com- 

 bined in layers, felt, and unglazed porcelain. 



Such a result as that spoken of in Science (ix. p. 

 457) is to be expected, when we consider the me- 

 chanical work we have to do in filtration ; for no 

 amount of chemical change is expected except in 

 Clarke's process, which is precipitation, not filtra- 

 tion. If the filter is a successful strainer, the sus- 

 pended matters within the water are held back upon 

 the surface of the strainer and ivithin the interstices 

 of the filtering media, whether it be gauze, asbestos, 

 iron shavings, felt, or porcelain. The great mistake 

 seems to be in believing, that by use of a current of 

 water, or by removing the media and scrubbing the 

 surface with a brush, all the filtrate is removed, for- 



