92 



NA TURE 



\_May 22, 1884 



volcanic dust will be the result. The same might be said of the 

 Krakatoa dust which has been collected far from its source. 

 That which fell August 27 at Batavia, about sixty miles from 

 Krakatoa, according to Renard, consists chiefly of glass parti les, 

 with plagioclase, augite, rhombic pyroxene, and magnetite, 

 giving the general composition found in some augite-andesites 

 In Krakatoa dust and pumice obtained from various localities 

 in the vicinity of the Java coast, I have always found glass the 

 most abundant constituent. The rhombic pyroxene, hypersthene, 

 predominates largely over augite, and as Mr. Iddings has already 

 shown, the ejected material belongs to hypersthene-andcsite very 

 like the pumiceous variety of the same rock upon the south- 

 western slope of Mount Shasta in the Cascade Range. 



While it is evident that all kinds of ejected material, from the 

 finest dust to the coarsest fragments, may be found upon their 

 parent cone, yet it is true that all ashes which have been trans- 

 ported by the winds for distances greater than one hundred miles 

 are composed chiefly of glass fragments distinguished by their 

 pumiceous character. Volcanic glass may be considered an 

 almost inevitable product of violent eruptions ; and of all the im- 

 portant constituents of sand and dust formed in this way, it is 

 the lightest. Furthermore, for a reason easily explained, it is 

 blown to much finer dust-particles than any of the products of 

 crystallisation. In a magma where crystallisation is taking place, 

 the absorbed gases and uncombined water under enormous ten- 

 sion are gradually accumulated in that portion which is most 

 liquid and least individualised. In this way the portion of the 

 magma which upon solidification yields glass becomes stored 

 with the energy that will cause its distension and perhaps blow 

 it to atoms when the mass is relieved .from the antagonising 

 pressure. The individualised and unindividualised portions of 

 the magma may be irregularly commingled, or they may arrange 

 themselves, as is frequently the case in obsidians from Oregon, 

 in more or less regular alternating bands. The streams of 

 microlites must necessarily imprison less uncombined water or 

 absorbed gases than the bands between them, so that when the 

 pressure is relieved the latter will suffer the greatest amount of 

 distension. If the tension of the confined water and gases is 

 great enough, the amorphous portion of the magma may be 

 blown to glassy dust, while the individualised portion, pulverised 

 rather by external than internal forces, is not reduced to so 

 high a degree of fineness. Volcanic dust is the extreme term of 

 a series which begins in compact lava, and has for its middle 

 members pumice in different stages of inflation. It appears to 

 be a significant fact, at least as far as I have had an opportunity 

 to observe, that effused pumice, i.e. pumice which occurs in 

 places as froth upon a stream of obsidian into which it gradually 

 passes, is highly microlitic. The glassy partitions which bound 

 the more or less rounded vesicles are crowded with microlites 

 and crystallites, while in ejected pumice where long, distended 

 vesicles prevail, or in volcanic glass dust, the products of crys- 

 tallisation are comparatively few or entirely wanting. 



Prof. F. W. Clarke, chief chemist of the United States Geo- 

 logical Survey, has determined the amount of silica in the sand 

 which fell at Unalaska, as well as that from the north-eastern 

 slope of Mount Shasta. The former contains 52-48 Si<_>... As 

 should be expected, it is more basic than hornblende-andesite. 

 and indicates that the acid portion of the magma — the volcanic 

 glass in the form of dust — was blown away from the sand. It 

 is well known that the glassy base is in general the most acid 

 portion of a rock, and any mechanical means by which the 

 magma is separated into crystal sand and glass dust divides the 

 rock into a basic and acidic portion. This division is in certain 

 degrees indicated by the analysis of Renard and Iddings of 

 material from the recent eruption of Krakatoa. The pumice 

 contained only 62 per cent, of SiOo, while the dust which fell at 

 Batavia, according to Renard, contained 65per cent, of SiG 2 . The 

 few observations I have been able to gather indicate that even 

 under favourable circumstances volcanic sand is not carried a 

 hundred miles from its source, while dust from the same vent 

 may be distributed over many hundreds of miles beyond the sand. 



It is unfortunate that we do not possess more definite and 

 detailed knowledge with reference to the source and distribution 

 of the volcanic sand from Unalaska. Mr. Applegate, in his 

 letter of transmittal dated October 22, says, " I forward by this 

 mail a sample bottle of sand that fell during the storm of 

 October 20, 1883. At 2.30 p.m. the air became suddenly 

 darkened like night, and soon after a shower of mixed sand and 

 water fell for about ten minutes, covering the ground with a thin 

 layer. The windows were so covered that it was impossible to see 



through them. This sand is supposed to have come either from 

 the Makushin or the new volcano adjacent to Bogeslov. The 

 former is a distance of about nineteen miles to the south-west, 

 us has only issued forth smoke or steam. The latter 

 is a new one which made its appearance this summer, and burst 

 out from the bottom of Behring Sea. It has been exceedingly- 

 active, and has already formed an islam] from -Soo to 1200 feel 

 high. Bogeslov is about sixty miles from here (Unalaska) in a 

 west direction. The new volcano is about one-eighth of a mile 

 north-west of it." Judging alone from the size of the grains ol 

 sand it seems probable that it may have been brought from 

 Bogeslov. Its paucity in glass fragments as compared with the 

 coarse sand from a dozen miles north-east of Shastina, indicates 

 that it was carried a considerable distance from its source, so as 

 to allow a pretty complete assorting of the material by the 

 wind. 



Grewingk, who has given us the most important contribution 

 to the geology of Alaska, more especially of the Aleutian 

 Islands, has prepared a geological map of Unalaska, and reports 

 the volcanoes there as emitting basic lavas which, from his 

 meagre description, appear to be similar to those poured out by- 

 volcanoes of the Cascade Range. It seems very probable that 

 the volcanic sand ejected by Makushin must be of the same 

 general composition as that which lately fell at Unalaska, 

 Grewingk's work contains a description of the island of 

 Bogeslov, but in it are contained no petrographic notes of im- 

 portance. Of the rucks on the new volcanic island north-west 

 of Bogeslov, for which the name Grewingk has been proposed, 

 we have no information. If the sand under consideration really 

 came from Grewingk, as seems most probable, we should 

 expect the island to be made up of hornblende-andesite. 



In connection with the all-absorbing topic, the peculiar sun- 

 set phenemena, much has been said of volcanic dust from 

 Krakatoa. A surprisingly wide distribution has been assigned 

 to it, and there is doubtless considerable scepticism concerning 

 its identification. The forms of glass particles in volcanic dust 

 are peculiar, and this, taken in connection with their isotropic 

 character, renders them easily recognised under a polarising 

 microscope. In the annexed figures Series No. I gives the out- 

 lines of glass fragments in Krakatoa dust. Series No. 2 is taken 



<£) 



a 



^=SJ 



<^ <\ ED *=»' 



d 



<=> U T^ 



v?s? 



/f ^=ss 



*} 



ty 



from the volcanic dust which fell in Norway and Sweden, 

 March 29 and 30, 1875. Series No. 3 represents the curious 

 fragments found in an old quartz-porphyry tufa at Breakheart 

 Hill, in Sangus, north of Boston. The forms of glass particles 

 seen in the volcanic dust collected by Mr. Russell near the 

 Truckee River are represented in Series No. 4. The fragments 

 represented in Series No. 3 are now chiefly quartz, but were once 

 particles of volcanic glass, and show that in the early geological 

 history of Eastern Massachusetts there were volcanoes belching 

 forth volcanic ashes like that of recent times, and flooding the 

 country with acid lavas whose beautiful and regular fluidal band- 

 ing has puzzled many observers and led them to suspect its 

 sedimentary origin. It seems reasonable to suppose that the 

 Grewingk crater must have yielded dust as well as sand, and that 

 the former can, with a high degree of probability, be distin- 

 guished from the dust of Krakatoa. The glass dust from 

 Grewingk, judging from that seen in the Unalaska sand, is less 

 clouded than that from Java. In the Krakatoa dust it is hyper- 

 sthene which is associated with the feldspar, augite, and mag- 

 netite, but in Grewingk dust we should expect to find hornblende. 

 A complete knowledge of the distribution of volcanic sand 

 and dust has such an important bearing upon meteorological con- 

 ditions, as well as upon volcanic phenomena, that it is hoped 

 accurate and continuous observations may be made upon this 

 subject at suitable meteorological stations. When we consider 

 the dust in cities rising from the ground into the air during dry 

 weather, as well as contributions, frequently glassy, made by 



