Notices respecting New Books. 147 



lunar inequalities, and the precessional phenomena .... make it 

 highly probable that the earth has once been in a fluid or semifluid 

 state" (Conclusion to Treatise on the Figure of the Earth). And 

 Archdeacon Pratt (p. 1 60) says, " we can hardly conceive stronger 

 evidence than the foregoing pages furnish that the earth was once a 

 fluid mass." 



As the Archdeacon has brought forward an additional argument 

 on the subject, and one of great importance if sound, we will 

 briefly pass the whole in review, as stated in the present work. 



In the first place, the variations of the force of gravity on the 

 earth's surface deduced from this hypothesis exactly coincide with 

 those which are ascertained by means of experiments with pendu- 

 lums. In the next place, the moon's motion in latitude experiences 

 a perturbation which falls in exactly with the hypothesis. In the 

 third place, by making a particular assumption as to the variation of 

 the densities of the consecutive strata, a value of the ellipticity of the 

 mean surface is deduced which nearly coincides with that ascertained 

 by geodesy. In the fourth place, on the same assumption, an ex- 

 pression for the annual precession can be deduced which, when 

 compared with that which actually exists, gives an ellipticity for the 

 mean surface not differing much from that ascertained by geodesy. 



Such being the evidence, let us now consider its value. It is 

 plainly all of one kind. An hypothesis is framed, and is found to 

 lead to results coinciding with observed facts. This may be presump- 

 tive evidence of the truth of the hypothesis, but is not proof. If a 

 large number of phenomena of various kinds are connected and ex- 

 plained by an hypothesis, the presumption is strong, and may amount 

 to something equivalent to proof. But if the facts are few, the pre- 

 sumption cannot be strong ; and if to enable it to embrace these few 

 facts it has to be backed by a subsidiary hypothesis, the presumption 

 in its favour is very much weakened. And this seems to be the case 

 with the fluid hypothesis. The third and fourth " tests " depend on 

 a particular assumption as to the variation of the densities of the 

 successive layers, viz. that expressed by the formula 



pa=Qsmqa, 



where p denotes the density of the layer whose semi- major axis is a, 

 Q and q being constants. Now this formula does incidentally imply 

 that the density increases as the centre is approached, and is so far 

 unobjectionable ; but its chief recommendation is that it possesses 

 the unique property of rendering a certain differential equation inte- 

 grable. It would, to our mind, be a singularly happy coincidence 

 if a law of density assumed for such a purpose were physically 

 true. We think, therefore, that the real evidence for the fluid 

 hypothesis is reduced to that furnished by the first two " tests." 



Now, with regard to them, it is plain that any hypothetical 

 arrangement of the matter composing the earth would possess the 

 same amount of consistency with these facts as the fluid hypothesis, 

 provided that the matter when so arranged continues to exert the 

 same attraction as before on all external points. Hence in order to 



