THE president's ADDRESS. 173^ 



In the second place, if a group be defined by certain characters, 

 it is nearly always found that particular forms, though clearly 

 shown by features of secondary importance to belong to the 

 group in question, nevertheless lack entirely the characters we 

 have chosen as diagnostic of the group. Living beings, adaptable 

 in every direction, will always evolve themselves out of any 

 definition that can be laid down. Thus it would be natural and 

 reasonable to separate vertebrate animals into two main groups, 

 one in which the paired limbs are of the type of fins, as in fishes, 

 and one in which the paired limbs are pentadactyle in type, as 

 in mammals, birds, reptiles and amphibia ; but then we must 

 include in our pentadactyle group the snakes and some other 

 creatures that have no paired limbs at all, not to mention horses, 

 in which the five digits originally present in each limb have- 

 become reduced to one in the course of evolution. Again in 

 siliceous sponges, we can recognise a great natural group, the 

 Tetraxonida, characterised by skeletal spicules with four rays, 

 generally combined with star-like flesh-spicules, so-called asters ; 

 but in such a form as the common Tethya * li/ncurium the 

 skeletal spicules are all reduced to simple needle-like forms ; in 

 the genus Chondrilla the skeletal spicules are gone altogether, 

 and only the asters remain ; and finally in the genus Chondrosia^ 

 there are no spicules at all. The structural peculiarities, however ,^ 

 of Tethya, Chondrilla and Chondrosia are such as to leave no 

 doubt whatever that they are closely allied to true Tetraxonida, 

 and must be classified in this group ; and thus we are brought to 

 the apparently illogical result, that after having defined a group 

 by certain characters, we are forced to place in that gi-oup forms 

 which do not possess those characters. 



A third limitation imposed upon us in the classification of 

 living beings is that there are many possible systems of classifica- 

 tion which may be flawless from the purely logical point of view, 

 but which will not serve our purpose from the biological stand- 

 point. We require what is termed a natural classification, that 

 is to say, one which exhibits the true affinities and relationships 

 by descent, so far as they can be ascertained, of the organisms 

 dealt with. From the standpoint of strict logic there is nothing 

 against subdividing animals by their habitat into terrestrial, 

 aquatic and aerial ; but then whales would be grouped with 



* Now generally named Donat'ta li/ncnrium by nomenclature purists. 



