212-214] Ejection in Filaments 215 



further predicts that- the emitted matter ought to proceed mainly from two 

 antipodal points on the equator of the nebula. 



To examine whether these latter predictions of theory are verified, we 

 must obviously study our nebulae from a line of sight near to the axis of 

 rotation ; the photographs shewn in Plate III, all being takerTedgewise, can- 

 not be expected to give any information on the question. But there is no 

 reasonable doubt that the last three nebulae on Plate III are nothing but 

 ordinary spiral nebulae seen edgewise ; although we cannot study the cross- 

 sections of these by their equatorial planes, there are innumerable other spiral 

 nebulae whose orientation in space is such that we look almost directly on to 

 their equatorial planes. Typical examples will be found on Plates II and V ; 

 other types will be found in any collection of photographs of nebulae. The 

 general characteristic of all these nebulae is that the two arms proceed ap- 

 proximately from antipodal points on the equator. That these arms really 

 represent an ejection of matter from the central nucleus is almost proved by 

 the two instances of M. 51 and M. 101 already discussed in 4. All this is 

 quite in accordance with theory. 



Two non-typical spiral nebulae are illustrated on Plate IV. The peculiarity 

 of the first is that the spiral arms have given place to an almost continuous 

 cloud of gas or dust, the separation of the arms being shewn only by the 

 faint rifts or lanes between them. The peculiarity of the second is that the 

 spiral arms are almost circular in shape, so that the whole figure is nearly 

 symmetrical about its axis of revolution. In this nebula we appear to have 

 a close approach to the manner of evolution imagined by Laplace. It will be 

 seen that our tentative hypothesis is able to account for these exceptional 

 nebulae as well as for those of more normal type. 



Thus our conjectural interpretation of all spiral nebulae is that they are 

 masses of gas or clouds of dust in rotation, this rotation being so rapid that 

 no figure of statical equilibrium is possible. In the earlier stages of their 

 evolution, they must have passed through a series of figures of equilibrium of 

 the pseudo-spheroidal type discovered in Chapter VII, until a sharp edge was 

 formed. After this, matter was ejected along two arms originating from this 

 sharp edge. At first the points of origin of the arms were determined by the 

 infinitesimal tidal forces set up by the rest of the universe ; subsequently the 

 tidal forces from the symmetrical arms themselves would suffice to confine 

 the emission to two antipodal points. 



214. It is worth noticing that on this interpretation of the spiral nebulae 

 the mean density of the nucleus can be approximately determined when the 

 period of rotation is known, the angular velocity &> being connected with the 

 mean density p of the nucleus by the approximate relation 



a) 2 = 0-35 x 2-7T7/5 = 2-27Jo, 

 7 being the gravitation constant. 



