CLASSIFICATION. 63 



cerned, than are the lowest Vertcbrata. In a Hnear classifi- 

 cation, therefore, the cuttle-fishes should be placed above 

 the lowest fishes — such as the lancelet — in spite of the fact 

 that the type upon which the latter are constructed is by 

 far the highest of the two. 



It is obvious, therefore, that a linear classification is not 

 possible, since the higher members of each sub-kingdom are 

 morei highly organised than the lower forms of the next 

 sub-kingdom in the series, at- the same time that they are 

 constructed upon a lower morphological type. 



In the words of Professor Allen Thomson, " it has become 

 more and more apparent in the progress of morphological 

 research, that the different groups form circles which touch 

 one another at certain points of greatest resemblance, rather 

 than one continuous line, or even a number of lines which 

 partially pass each other." In the same way the highest 

 vegetables do not approximate to, or graduate into, the 

 lowest animals ; but " each kingdom presents, as it were, 

 a radiating expansion into groups for itself, so that the rela- 

 tions of the two kingdoms might be represented by the 

 divergence of lines spreading in two different directions from 

 a common point." 



