GEOLOGY. 309 



such a bilateral decentralizing force is assigned to the explosive energy 

 of very eruptive stars when stimulated and supplemented by the 

 cooperation of tidal elongation. With this combination of tidal and 

 explosive action, the resources of dynamic encounter appear competent 

 to give a very wide deployment to giant suns when the approach of 

 the two bodies is very close. There is reason to think that clusters of 

 germ-suns, some thousands in number, may arise from such coopera- 

 tive agencies when the stars are very massive; but without this eruptive 

 factor it is not clear that even the utmost resources of dynamic encoun- 

 ter are adequate to give to great star systems, already widely deployed, 

 more than a spiraloidal deformation. But it is worth considering 

 whether the very great deployment previously attained may not replace 

 the explosive factor in the planetary case, for such deployment gives 

 great efficiency to the differential effects of tidal action. 



As already remarked, the forms of spiral nebulae seem clearly to 

 imply an acti^'e transiUonal state. Spiral nebulae seem to be about 

 as far as possible from an ideal equilibrium or final state, or any 

 measurably permanent state referable to Newtonian mechanics. In 

 contrast with this, equilibrium states of a high order of maturity seem 

 to be represented by the eighty-odd globular clusters that form sub- 

 features of our galactic system. These globular clusters imply that the 

 age of our stellar system is so great as to have given time for the evolu- 

 tion of steady states of interaction in stellar assemblages vast enough 

 to embrace tens of thousands of stars. Such mature organizations 

 equally imply the effective and protracted working of the law of parti- 

 tion of energy resulting in an appropriate distribution of velocities 

 and of orbital dimensions and attitudes. Now, if spiral nebulae are 

 to be interpreted as dim, distant star-clusters of the highest order, 

 ''island universes," comparable to our whole galactic aggregation, and 

 if the generalization that most of the white nebulae have the spiral 

 form is also to be accepted, the revived Herschellian hypothesis seems 

 forced to carry the corollary that most of the stellar "universes" 

 are as yet scarcely past their infantile states. Even if only the majority 

 of white nebulae are spiral, the corollary has much significance. It 

 is, under this interpretation, quite remarkable that so few of the 

 nebulae interpreted as star-clusters show by their forms more than the 

 very earliest steps toward a steady state of star aggregation such as 

 should be evolved in time under the law of partition of energy. This 

 is the more remarkable since the law of partition of energy is so beau- 

 tifully expressed in the globular clusters. To interpret all the white 

 nebulae as star-clusters seems thus to carry implicitly the very radical 

 question, How^ can such singular and seemingly nascent states of organ- 

 ization have arisen in so large a proportion of all the observable aggre- 

 gations of this major class? This in turn raises, in a very acute form, 

 the question whence comes an adequate frequency of action — whatso- 



