CLASSrPICATIOW. 47 



The orders and classes of moUusca have already been referred 

 to ; those now in use are in most cases natural. 



It has been sometimes asserted that these groups are only 

 scientific contrivances, and do not really exist in nature Jbut 

 this is a false as well as a degrading view of the matter. The 

 labours of the most eminent systematists have been directed to 

 the discovery of the subordinate value of the characters deriv- 

 able from every part of the animal organisation ; and, as far as 

 their information enabled them, they have made their systems 

 expressive "of all the highest facts or generalisations in 

 natural history." — (^Owen.) 



M. Milne Edwards has remarked, that the actual appearance 

 of the animal kingdom is not like a well-regulated army, but 

 like the starry heavens, over which constellations of various 

 magnitude are scattered, with here and there a solitary star 

 which cannot be included in any neighbouring group. 



This is exceedingly true ; we cannot expect our systematic 

 groups to have equal numerical values,* but they ought to be 

 of equal structural importance ; and they will thus possess a 

 symmetry of order, which is superior to mere numerical regu- 

 larity. 



All the most philosophic naturalists have entertained a belief 

 that the development of animal forms has proceeded upon som^ 

 regular plan, and have directed their researches to the discovery 

 of that "reflection of the Divine mind." Some have fancied 

 that they have discovered it in a mystic number, and have ac- 

 cordingly converted all the groups into fivesA We do not 

 undervalue these speculations, yet we think it better to describe 

 things so far only as we know them. 



Great difficulty has always been found in placing ' groups 

 according to their affinities. This cannot be efiected in — the 

 way in which we are compelled to describe them — a single 

 series ; for each group is related to all the rest ; and if we 

 extend the representation of the affijiities to very small groups, 

 any arrangement on a plane surface would fail, for the 

 affinities radiate in all directions, and the " network" to which 

 Fabricius likened them, is as insufficient a comparison as the 

 " chain" of older writers. J 



* The numerical development of groups is inversely proportional to the bulk of the 

 iudividuals composing them. — {Waterhouse.') 



t The quinarians make out five molluscous classes by excluding the tunicata ; the 

 same end would be attained in a more satisfactory manner by reducing the ptercpoda 

 to the rank of an order, which might be placed next to the opistho-branchs. 



X The quinary arrangement of the molluscous classes reminds us of the eastern 



