Rev. F. D. Morice on Andrcno^ taraxaci, etc. 251 



no such connection, at worst we should find it out, and 

 have to drop our group, and look out (with the advantage 

 of an experience gained) for a fresh one. 



Whether or not this method is possible in dealing with 

 the subdivision of Andrena into groups, the other at least 

 has not led us, and seems unlikely to lead us, to satis- 

 factory results. All kinds of characters have been tried 

 separately, as the basis of groups; and all alike lead to 

 results, which it is impossible to accept as corresponding 

 to the relations between the species which really exist. 

 The categories of "Apidie Europsege," based on pilosity- 

 characters, throw cdbicrus, Kirby, into one section and 

 argentata, Smith, into another, and remove pilijjes, F., from 

 its place near himaculata, etc., to group it with species 

 differing essentially in structure both from it and from 

 one another — morio,fiessa3, cineraria, and so forth. Surely 

 most unnatural results ! Yet this is the nearest approach 

 that has been made to a natural grouping of the Paloearctic 

 species. It might be thought that a system could be 

 founded on some one obvious difference of structure — the 

 male head and mandibles, the sculpture of the thorax, etc. 

 But species which are evidently nearly related would be 

 thrown into different sections if we made, say, the form of 

 the mandibles, or the size of the head, or the length of the 

 antennal joints, our sole criterion. The character of a 

 strongly rugose and enclosed propodeum, no doubt, associ- 

 ates a certain number of species which may reasonably be 

 thought near allies. Yet it is at least open to question 

 whether ephippium or albicans have really such an affinity 

 with tibialis, bimacidata, etc., as undoubtedly exists between 

 the latter species. Neither ephip2num nor albicans has the 

 characteristic tibialis armature or ventral segments, nor in 

 this respect do they at all resemble each other, and the 

 unlikeness of their puncturation and general appearance 

 is extreme. I have dwelt on this case as the one -in which, 

 on the whole, the attempt to rest a group on a single 

 character leads to the least undesirable results. But even 

 if we acknowledged it to have been successful, it would be 

 the sorb of " exception that proves the rule ": for no other 

 tolerably natural grouj) has yet been founded on a detail 

 of structure. And as to colour, whether of pilosity or 

 integument, it cannot be trusted even to separate species 

 — far less groups. 



Such being the present state of the classification of 



