130 CARNEGIE INSTITUTION OF WASHINGTON. 



the fact that all the streams draining the valley contain much more water 

 in their upper courses than at their outlets. (3) Decidedly the smallest 

 amount of water in any of the gases was found in those issuing on the west 

 slope of Novarupta a short distance from the only point where the lava body 

 reaches the surface. (4) The fumaroles below the surface of the ground appear 

 to be accessible to air; if so, they should be accessible to water-vapor as well. 



It is also shown that in none of the fumaroles are combustible gases burning 

 at the surface. The observed negative temperature gradients were in all 

 probabilitj'' due to hot lateral gas currents. 



No relation between temperature and gas composition was observed, 

 except in the case of the soluble gases. These are usually very small 

 in amount in fumaroles of low temperature and relatively high as compared 

 with the "insoluble" gases in the hotter fumaroles. A survey of the literature 

 indicates that magmatic gases are not necessarily uniform in composition 

 and that the original gas mixtures are modified by a number of secondary 

 agencies only partially dependent on temperature. The generalization of 

 Deville and Fouqu^ that magmatic gases are everywhere uniform and that 

 their composition is subsequently modified by no variable agency except 

 temperature seems untenable. 



The second volcano study was begun in the second year (1915) of the recent 

 eruptive activity of Lassen Peak in California and has been continued inter- 

 mittently since that time, partly in the field and partly in the Laboratory. 



Compared with other volcanic outbreaks of which we have a record, Lassen 

 Peak has not during its recent activity displayed unusual phenomena or 

 excessive violence. This activity may be summarized as a series of explosions 

 in the summit crater, beginning in May 1914 and continuing at intervals of 

 2 or 3 days until the same month of 1915, during which period both the 

 violence and the duration of the explosions generally increased. The out- 

 breaks culminated in 3 days of nearly continuous explosions, during which 

 red heat was once visible on the top of the mountain (May 19), the lava plug 

 forming the floor of the crater was pushed up 300 feet or more to the level of 

 the crater rim, and finally two horizontal blasts of extreme intensity broke 

 through on the northeast flank of the mountain beneath the plug, laying 

 waste the valleys of Lost Creek and Hat Creek for a distance of 4^ miles. 

 These two blasts occurred on May 19 and 22, 1915, and swept out every 

 vestige of vegetation in the center of their path, including nearly 5,000,000 feet 

 of original forest timber. At the sides of the valley, down which the blasts 

 passed, the trees were blown down but were not carried away, so that they 

 remained lying upon the steep sides of the inclosing valley, pointing directly 

 away from the crater, thereby showing clearly the point of origin of the blast 

 which blew them down. There was no eye-witness of the horizontal blasts 

 either on May 19 or May 22. 



Following these outbreaks, the activity of the volcano slowly subsided, the 

 last eruptions of any magnitude occurring in May of 1917. 



The mechanism of these horizontal blasts is interesting. At the beginning, 

 or shortly before the activity of the 19th, the crater floor was raised from 

 below until it stood nearly level with the crater rim. Had it gone higher, the 

 plug would undoubtedly have been dislodged entirely from the vent and a 

 flow of lava would have resulted. As it was, the enchained pressure of 



