596 



NATURE 



[February 5, 1920 



it left a very large residue adhering to the valley floor. 

 Subjected to heat from below, the mud dried, (.ilied, 



Photo\ [/. /('. Skifley. 



Fig. 2. — Cross-section of horizontal tunnel about fifteen feet below the 

 surface of the mud-flow. Note the cleavage planes of the mud 

 above the vent. One hundred feet from this is another horizontal 

 tunnel seventy-five feet below the surface, and large enough to 

 drive a team and wagon through. These have most probably 

 been formed by the solvent action of superheated steam and 

 hydrofluoric acid. 



and was eventually baked. The 

 contracting, hardening mass split 

 and cracked according to the strains 

 and stresses set up by the irregulari- 

 ties of the valley floor beneath. The 

 Katmai volcanic ash lies conform- 

 ably on top of this mud-flow. Ap- 

 parently the order of the sixty-hour 

 eruption of June, 1912, was :- 

 (i) Novarupta explosion, followed 

 by the great flow of mud ; (2) Kat- 

 mai eruption ; and (3) the up- 

 thrusting of the lava plug of 

 Novarupta. No ash covers thi.s 

 plug. It was the last major event 

 to happen. 



The mud-flow occupies the floor 

 of the Valley of Ten Thousand 

 Smokes described in this journal by 

 Dr. Griggs ("The Eruption o'^f 

 Katmai," Nature, .August 22, 1918, 

 vol. ci., p. 497). The volcanic gases 

 force their way upward through this 

 superincumbent detrital material, 

 using the existing cracks and 

 fissures, and dissolving out new 



channels where openings were not already available. 



As chemist of the 19 17 Katmai Expedition, jt fell to 

 NO. 2623, VOL. 104] 



my lot to examine these gases and the encrustations 

 deposited around the fumaroles. 



The gases contain some of the strongest dis- 

 integrating agents known. Hydrofluoric acid and 

 hydrochloric acid, together with superheated steam, 

 proved to be the most common constituents of thri 

 outpouring gases, frequently issuing at a temperature 

 above 400° C. Many fumaroles were so impregnated 

 with these acids that it was impossible to breathe the 

 vapours. The surface of the mud-flow surrounding 

 some of the more acid vents was covered with ferrous 

 chloride and impregnated with free hydrochloric acid. 

 The presence of hydrofluoric acid in the emanations 

 was accompanied by a deposit of amorphous silica 

 around the vents, almost completely closing the 

 orifices and forcing the gases to issue through cracks 

 in the hot, baked silica. These deposits, sometimes 

 1)8 per cent, of pure silica and altogether anhydrous, 

 formed dykes several feet high around the hotter 

 vents. 



This association of hydrofluoric acid and silica is 

 not accidental. Hydrofluoric acid decomposes silicates, 

 setting free the silicon as gaseous silicon tetrafluoride. 

 and this in turn is decomposed in the presence of 

 water, forming SiO, and free acid. The majority of 

 the samples of encrustants brought back gave a quali- 

 tative test for fluorine, while quantitative results ran 

 as high as 7 per cent. 



Realgar and orpiment were found in conjunction 

 with deposits of sulphur. Haematite in 'the form of 

 "Venetian red" and .small crystals of ovrites em- 

 bedded in a matrix largely silica were common 

 secondary reaction products of the volcanic gases. 

 Hydrogen sulphide was almost ubiquitous. 



Hygroscopic iron and aluminium salts formed in 

 the throats of vents protected from the weather and 

 at a temperature above 100° C. Unfortunately, these 

 were highly deliquescent, and lost their crystalline 

 form on exposure to the air. 



The throats of several vents near Novarupta were 

 lined with quantities of ammonium chloride crystals, 

 almost 90 per cent. pure. 



One of the most interesting deposits was a tarrv 

 substance found in the proximity of the ammonium 

 chloride fumaroles, which proved on analysis to 

 contain hydrocarbons of an asphaltic character. 



Main- of the fumaroles contained ammonium com- 



I'Ih4o\ 

 Fig. 3. — ^The surface of 

 to distinguish from 

 Trident Volcano. 



I/. )('. Shipley. 

 the mud is very deeply eroded by watercourses, and these are frequently hard 

 the fissures. These gullies are in a very active area on the northern slope of 



pounds in the issuing gases, strikingly indicated by 

 the growth of algae. Wherever blue-green algae were 



