THE HERSCHELS AND THE STAR-DEPTHS. 29 



star-systems ? Take any really parallel case and the 

 answer to this question will be obvious. Suppose a 

 botanist had asserted his belief that all the plants pre- 

 senting certain characteristic features were poisonous, 

 no evidence beyond the existence of those features 

 being at the time available, and that at length some 

 person made actual experiment on ten or twelve orders 

 of plants having such features, and found that they 

 certainly were not poisonous would not this demon- 

 strated fact dispose entirely of the reasoning, however 

 ingenious it might be, on which the general theory of the 

 poisonous nature of such plants had been supposed to 

 be established ? Would it not be a fair inference that 

 the untried orders were at least probably innocuous? 

 And would it not be thought strange if a botanist, 

 commenting on the discovery that all the as yet tried 

 orders of plants having certain characteristics were 

 innocuous, were to say, ' This demonstrated fact must 

 inspire some degree of caution in admitting as certain 

 the conclusion that the remaining orders of such plants 

 are poisonous'? I yield to none in my respect for 

 the great astronomer whose loss science is now deplor- 

 ing. I entertain most strongly the opinion that he 

 was far the greatest astronomer of our time ; but truth 

 compels me to say that in his mode of dealing with 

 demonstrated facts, and especially in this particular 

 instance, he was, to say the least, not so happy as his 

 father. He seems almost to have regretted to see 

 certain questions pass beyond the field of controversy 

 into the domain of the known. 



