XLIII.] CLASSIFICATION OF THE SCIENCES. 217 



It is one of the many remarkable harmonies between Hamiony 

 the mind of man and the nniverse of which it is the l^j^f of 

 noblest product, that this distribution of the products of man with 

 nature, both in space and in time, is that which appears . . .' 

 beautiful to the artistic sense. It is a maxim in art that rule of 

 the highest beauty should be introduced in relatively i^efnty in 

 small quantity : thus, in architecture, which is perhaps least 

 the best instance, such parts as cornices and capitals, which 

 are at once small and conspicuous, can scarcely be too 

 richly ornamented ; but the effect would be very bad if 

 the ornament which suits a cornice were spread over a 

 wall, or if that which suits a capital were continued down 

 the shaft of the column. In all art whatever, the effect 

 of an equal distribution of beauty over every part is not 

 good. In all art whatever, any part of a composition which 

 rises above the general level of the whole in dignity or 

 beauty will add dignity or beauty to the whole, pro- 

 vided that it is properly placed ; wliile if any part sinks 

 below the general level, it lowers the character of the 

 whole. These principles are applicable alike to those arts 

 which address themselves to the eye, and those which 

 address themselves to the ear. But in the latter — that is 

 to say in poetry and music, in which the parts of a com- and, in 

 position are not simultaneous but successive — this further mugjif at" 

 maxim is to be observed, that the highest beauty not only ^^^ end. ^ 

 ought to be small in quantity, but ought to come last; 

 and every previous part of the composition ought to lead 

 up to it. In thus arranging his work, the artist, without 

 knowing it, follows the example of nature, which, as I 

 have already said, produces the liighest results at the end 

 of long ages of slow preparation, and then in compara- 

 tively small quantity. 



If it is asked whether I believe that all things which are Not all 

 capable of being known are capable of being included in is'hicluded 

 the two tabular series that I have drawn out in tliis in the 

 chapter, I reply that I have no such notion. In this work enume- 

 I have confined myself so far as possible to what is called ^f**^*^ ^° 

 positive science ; that is to say, science having its data in chapter. 



