XIT.] FORBES' SCIENTIFIC WORK. 477 



effect over given areas." I think you are quite wrong 

 as In -low. 



'Article 31 : "It is in fact convertible with this. Every 

 portion, &c. will include as many stars as any other.'* 

 itur. I think you are here most absolutely wrong. 

 All that is meant by being as likely to be in one situa- 

 tion as another is this, that if the experiment of creating 

 a starry firmament were repeated a great many times, 

 and if you took the mean of the number of stars found, 

 at all the times, in each particular square degree, the 

 ter the number of such creations, the more nearly 

 would the mean contents of each square degree approach 

 to uniformity. 



' This affects several things which follow. If I might 

 presume to indicate what I think a radical fault in your 

 paper, it is this misunderstanding of the doctrine of 

 chances. Pray look to it well. 



' Note A. I believe (but will not assert) that there is 

 an error in Mitchell, as you have remarked : if so, it also 

 affects the investigation which I sent you. 



' And so end my comments. . . .' 



* ROYAL OBSERVATORY, GREENWICH, October 26th, 1860. 



4 ... I have had no further opportunity of writing, 

 and do not see the 1 -ast prospect of leisure sufficient for 

 c;ilm consideration of Struve and Mitchell. Therefore, 

 do not reckon on that. Meantime I send you my 

 remarks .-i] -plying to the three things which you 

 B .Mitchellite must believe, and you can judge whether 

 1 am Mitehellite or no. 



' 1. "That there is a definite probability against two 

 stars being within a certain distance/' 



'Certainly there is: errors in calculation always ex- 

 cept t this is to be taken in the sense which I 

 have before explained, that if it was possible to have a 

 great number of starry firmaments created, the av< i 



number of the double stars would !>< determine.! 



by the !;i\v as^-rted. 



' P.ut : finding that the chance is 10,000 to I 



