326 MESSES. G. A. J. COLE AND J. W. GREGOEY 



the mass. Moreover, we probably find preserved for us only the 

 lower layers of the volcanic cauldron, and any pumiceous matter 

 of the final surface must have been removed by early denudation. 



It is only natural that spherulites should occur in the glasses of 

 the Sandwich Islands, although they may not be on so bold a scale 

 as at Mt. Genevre. Mr. E. S. Dana* has recently described 

 spherulitic structure in a lithoidal lava from Mauna Loa ; the rock 

 contained olivine, and the spherulites were of two kinds, light-brown 

 ones being set in a " nearly opaque spherulitic ground-mass." On 

 pp. 451 and 459 of the volume quoted, Mr. E. S. Dana mentions 

 similar structures in other rocks from Mauna Loa and Kilauea. 

 Moreover, his beautiful "fan-shaped or feather-like" groups of 

 augite in the "clinkstone-like basalt "f appear to be sections 

 of interesting, if imperfect, spherulitic aggregations. 



Cohen J has also described a tachylyte from Hawaii with a 

 spherulitic ground-mass ; and a specimen in the collection of the 

 jS^ormal School of Science and Koyal School of Mines shows small 

 brown spherulites, clustering more and more closely until the glass 

 passes into the almost opaque matrix of the basalt. 



It would have been a harmonious conclusion to this comparison 

 if we could have classed our agglomerates as ancient act lava-streams, 

 that is to say, as brecciated, rugged, and scoriaceous flows. But 

 neither the arrangement of the masses nor the globe-like bombs 

 correspond with the features so clearly described and figured by 

 Prof. Dana §, and we are led to treat these deposits as produced by 

 true explosive action. The compact non-scoriaceous character of 

 the vast majority of the ejecta is paralleled, curiously enough, in 

 the " stones " of the tuffs on Kilauea ||. 



While the extent of these fragmental rocks does not seem to have 

 been sufiiciently insisted on in the past, authors have often stated 

 their views as to the origin of the variolite itself. The great 

 advance in the discussion of this question appears to us to have 

 been made by Lory^, when he so clearly recognized the variolite as 

 a product of the rapid cooling of an igneous mass. After this 

 statement the various accounts of its discovery as a selvage to 

 diabase dykes in other districts have seemed in the highest degree 

 natural, and have met with ready acceptance. 



Finally, if it is granted that the variolite of Mt. Genevre 

 represents a glassy lava-crust which has been devitrified by 

 slow secondary action, is there anything in the nature of the 

 compact diabases that will explain its relative abundance? If 

 the lavas were unusually fusible, the glassy condition might be 

 retained over large areas ; but it is unfortunately impossible to 

 argue as to the exact original characters, chemical or physical, 



* " Contributions to tbe Petrography of the Sandwich Islands," Amer. 

 Journ. Sci. vol. xxxvii. (1882) p. 452. 

 t Ihid. p. 443. 



\ Neues Jabrb. fiir Min. &c. 1880, vol. ii. p. 51. 



§ Amer. Journ. Sci. vol. xxxiv. (1887) pp. 3()2-364. || Ibid. p. 360. 



*j[ Bull. Soc. geol. France, 2^ ser. t. xviii. (1861) p. 782. 



