230 ANNUAL EEPOET SMITHSONIAN INSTITUTION, 1916. 



into general terrestrial mechanics only as a subsidiary feature. It 

 seems necessary to limit liquid and viscous lacunae — if there are lacunae 

 in any proper sense at all — to such moderate dimensions that they do 

 not seriously kill out distortional waves passing through the outer 

 half of the globe in various directions, for seismic instruments show 

 that these waves retain their integrity with surprising tenacity 

 through long traverses. It seems equally necessary to limit the 

 liquid and viscous factor rather severely if the interior structure is 

 to be susceptible of so prompt a response to twelve-hour stress pulses 

 as is implied by its almost complete elastic fidelity. 



In the light of these determinations, strengthened not a little by 

 their concurrence with the later geological determinations, the work- 

 ing hypotheses of the earth student can scarcely fail to take shape 

 according to the dynamic tenets implied by a rigid earth. 



The limitation of liquid and viscous matter thus imposed quite 

 radically conditions all tenable views of magmas and of vulcanisra, 

 and thus bears upon the origin of igneous matter. No small part 

 of petrologic effort in past decades has been spent on the differentia- 

 tion of magmas. To a notable degree these efforts have proceeded 

 on the assumption, conscious or unconscious, that differentiation took 

 its departure from an original homogeneous magma such as might 

 arise from residual portions of a molten earth. Indefinite lapses of 

 time, and such conditions of quiescence as are naturally assignable 

 to residual reservoirs of lava, have been freely assumed as working 

 conditions without much question as to their reality. Under the 

 hypothesis of a molten earth passing slowly into a partially solid 

 earth and retaining residual lacunae of molten matter as an incident 

 of the change, these assumptions are quite natural. On the other 

 hand, under the hypothesis of a pervasively rigid earth, affected by 

 stress conditions that are constantly varying in intensity and in dis- 

 tribution — and subject to more radical changes at times of periodic 

 readjustment — the existence of such residual magmas becomes at least 

 questionable, perhaps improbable. Still more questionable is the 

 assumption that the multitude of little liquid spots supposed to arise 

 within the elastico-rigid mass always have conformed to one type oi- 

 to one set of types. The inherent probabilities of the case seem to 

 point strongly to a wide variation in nature of these local bodies due 

 to selective solution or to differential fusion. The liquefying action 

 that brings magmas into being under this view is presumably con- 

 trolled by the same chemical and physical principles as the solidifying 

 phases of the same cycle. The logical presvimption is that at all 

 stages of a magma's career from its inception through its growth, 

 climax, and decline to its final solidification, selective action will be 

 in progress more or less and that no stage Avill be entitled to be 

 regarded as original or parental in a special sense, such a sense, for 



