between Stars forming Binary or Multiple Groups. 423 



who has represented me as making averments altogether 

 foreign to my thoughts, and has quite overlooked the pre- 

 cise objections contained in my letter printed in the Philo- 

 sophical Magazine, and the precise statements quoted from 

 HerschePs Astronomy, which I have again cited in § 5 

 of this paper. The reviewer charges me with " a singular 

 misconception of the true incidence of the argument from pro- 

 bability." I can only say that my error, if such it be, is one 

 sanctioned by Sir John Herschel, to whom I referred as my 

 authority ; for in his " Outlines " he states Mitchell's argu- 

 ment from the Pleiades, and Struve's arguments against the 

 chance combination of a solitary double, or treble star, as I 

 have shown in the text, but makes no mention, so far as I 

 have been able to discover, of a generalized argument from 

 the number of double stars in the different classes, depending 

 upon the disproportionate number of close double stars to those 

 wider apart; and either to understand or express that argument 

 aright, it would be an insult to common sense to invoke the aid 

 of the theory of probabilities; being neither more nor less than 

 this, that the occurrence of double stars increases exceedingly 

 much slower than the area of space including two members of 

 the same pair. I never found fault with this argument : it is an 

 argumentgood as an induction, but nothing more; the argument 

 complained of by me was this, — " that the odds are 9570 to 1 

 against any two stars falling within 4" of each other, and the 

 odds against any two such stars fortuitously falling within 32" 

 of a third is not less than 173,524 to 1 ;" and Sir John Her- 

 schePs inference from these premisses, and these alone, " the 

 conclusion of a physical connexion of some kind or other is 

 therefore unavoidable." The reviewer ignores these state- 

 ments, joins issue on a point never before brought into the con- 

 troversy, and silences objection by the general assertion that if 

 this case be given up " there remains no possibility of applying 

 the theory of probabilities to any registered fact whatever," — 

 a conclusion which of course I dispute as simply an assump- 

 tion of the point in debate. 



41. I am sorry to have to allude even thus briefly to the 

 controversial article in the Edinburgh Review ; but it has 

 been brought under my notice in a manner so authoritative 

 that I could not overlook it ; although, as an argument, it ap- 

 pears to me to be destitute of weight, and in fact to proceed 

 on a misapprehension. To criticize it minutely would not 

 only prolong this paper beyond due limits, but would involve 

 me in the invidious, and by me unsought task of dwelling on 

 other mistakes of eminent authors connected with the subject 

 of these pages. The principle of Mitchell is, evidently, that 



