CONFUSION OF HYPOTHESES 639 



their rounding, partly owing to internal viscosity and partly by the grind- 

 ing off of the more brittle surface irregularities. 



It is worthy of note that in all the less altered and hence better pre- 

 served examples of this structure the interstitial material consists of 

 angular fragments of glass of many sizes, but without rock-flour, and 

 there is no grooving or scratching of the glassy spheroidal surfaces. In 

 short, there is no evidence that appreciable rubbing, rolling, or other me- 

 chanical movement has occurred between the individual pillows, and it 

 seems difficult to conceive of any motion that might properly be called a 

 "rolling over among themselves" occurring in such a manner as to leave 

 no trace of it. 



Miss Eaisin's hypothesis that the pillows are the product of jointing 

 while cooling, modified by a subsequent injection of lava that penetrated 

 the whole mass and coated each individual block with glass, is inconsist- 

 ent with the occurrence of empty interstitial spaces, so commonly found 

 between the pillows, and also with the zonal arrangement of the interiors 

 and the gradual transition from these into the superficial crust of glass. 



In agreeable contrast with most of the others, Johnston-Lavis and 

 Platania (1891) offer a hypothesis that does not seem strained or im- 

 probable ; and it may well be found, when the characteristic distinctions 

 shall have been worked out, that some of the pillows that are intermin- 

 gled with the finer marine muds and oozes have been injected into the 

 semifluid and immiscible sediment and thus have formed a sort of giant 

 emulsion. This suggestion is adopted by Harker (1909) as probably of 

 wide application. Obviously, however, such a process would not be pos- 

 sible in coarse sediments nor on the land, and hence many well known 

 examples of pillow lavas would still remain unexplained. 



Many writers (Eansome, 1893; Kiimmel, 1897; Lewis, 1907; Grrabau, 

 1913) have regarded pillowy and ellipsoidal forms as merely variations 

 in the ropy flow structure of viscous pahoehoe, the circumstance of sub- 

 aerial or subaqueous eruption being by implication of minor importance. 



So many references have been made to pahoehoe and aa lavas that de- 

 scriptions of these types are included here for purposes of comparison 

 (and contrast) with pillow lavas. Descriptions are also appended of such 

 direct observations as are available on the formation of spheroidal and 

 pillow-like bodies in flowing lavas. 



Formation of Pahoehoe Lava 

 Dutton^^^ has described the typical pahoehoe lava of Hawaii as follows : 



123 c. E. Dutton : Hawaiian volcanoes. Fourth Ann. Rept. U. S. Geo!. Survey, 1884, 

 p. 96. 



XLVI— Bull. Geol. Soc. Am., Vol. 25, 1913 



