THE HEESCHELS AND THE STAR-DEPTHS. 2$ 



is educed is exceedingly simple ; but the result is of 

 the utmost importance. 



Each of the Magellanic Clouds, as we have said, is 

 nearly round. Now when an object appears round, the 

 most probable opinion we can form respecting the 

 object's shape is that it is globular. An object which 

 is not globular may appear circular, as for instance, an 

 egg, a roller, or the like, looked at endwise, or a coin 

 looked at in a direction square to its flat surfaces. But 

 we know that if an egg, or a roller, or a coin, were 

 held in a random position, the chances would be 

 against that position being such that the egg, or roller, 

 or coin would present its round aspect, so to speak. 

 And clearly, therefore, if we know nothing about a 

 certain object but that it appears round, we must 

 accept as probable the belief that it is globular.* This 

 conclusion, which would be justly arrived at in the 



* We have an instance of this sort of reasoning in the case of the 

 moon. "VVe know nothing, certainly, about the shape of the moon re- 

 garded as a solid, for we only see her under one aspect. So far as 

 abstract possibilities are concerned, the moon, as seen under certain 

 aspects from Venus, might present the shape of an egg, or even of a 

 diamond. Still we conclude that the moon is a globe, because she pre- 

 sents the aspect which a globe, and a globe only, presents in all posi- 

 tions. (Lately astronomers have indeed seen reason for questioning 

 this conclusion ; but my present argument is not affected by the cir- 

 cumstance.) But now let us conceive a case directly illustrating the 

 argument dealt with above. Suppose a certain fruit of unknown nature 

 is held in such a position, and at such a distance, that all we can recog- 

 nise of its aspect is its seeming outline, and that this outline is round. 

 We should regard it as probable that the fruit is globular. Now if a 

 second specimen were similarly held up (in a random position) and seen 

 to be also round, we should be very strongly confirmed in our opinion, 

 and the mathematical theory of probabilities shows us that this natur- 



