386 T. C. CHAMBERLIN 



elongation in the hour or two preceding its entrance upon the 

 Roche sphere. (2) After entrance upon the Roche sphere, an 

 explosive elongation actuated by the elastic energy then remain- 

 ing in the body unrestrained by self-gravity in the axis of elon- 

 gation. (A portion of the original elastic energy had been 

 consumed in the previous elongation and a corresponding amount 

 of momentum had been acquired, the larger component of which 

 would be effective along the changed line of elongation.) (3) 

 After passing out of the Roche sphere, the restraints of gravity 

 begin again to be felt and rapidly increase as A and B retire 

 from each other, but the distance to which the extremities of A 

 have already been projected, and the new relations thereby 

 assumed to the remaining mass of A, and to B, render the 

 renewed gravitative influence far less effective than the original, 

 and the projection must continue until the momentum acquired 

 is overcome. (4) Coincident with this projection a constantly 

 increasing rotation toward B has been generated, which possibly 

 reached an effectiveness comparable to that of the solar system. 

 The effects of explosive projection combined with concwTe?it rotation 

 must obviously give rise to a spiral f of m. 



It seems clear from the nature of the case that there would 

 be a certain brief period when the climax of projective effects 

 would be reached, and that a stream of material of much greater 

 mass and velocity than at other instants would at this time be 

 projected from the extremities of the elongated mass in both 

 directions. There should therefore be two chief arms to the 

 resulting spiral starting from the opposite points of the central 

 mass and extending outward to the limits of the spiral — indeed 

 constituting the most outlying portions of the spiral. These 

 must be curved in a common direction by the rotation of the 

 mass. Such predominant arms are notable features in the typi- 

 cal spiral nebulae. They are well shown in Nos. i, 2, 3, 4, 5, and 

 6, Plate I, all of which are reproductions from photographs 

 furnished by the late Professor Keeler, 



In the illustrative case that has just been discussed the solid 

 body B was made to represent a convenient possible case but 



