250 Rev. F, D. Morice on Andrena taraxaciy etc. 



(e.g.) as have most of the characters mentioned above, but 

 dark $ scopae, etc. (schniicdehiccliti, Magr., is one of 

 these). On the other hand it probably does not include 

 species Y\kQ falva go, Christ, rufida, Perez, rufoMspida, Dours, 

 etc., which have a short 3rd antennal joint, a different type 

 of male genitalia (see Fig. G in Plate V), and a different 

 style of surface and puncturation. Braunsiana and 

 another species which Herr Friese has determined for me 

 as leucolippa agree in several points with hnmilis, etc., but 

 their surface seems to me different, and they have quite 

 another type of $ armature resembling that of hattorfiana, 

 (Plate V, Fig. 18); circinata, Dours, is very like a small 

 humilis super ticially, but the punctures are utterly different 

 and the abdomen clearly fasciated ; it is probably nearer 

 to labicdis than to the present group, though not very 

 near to that. 



It may seem to some readers illogical to discuss whether 

 or not particular species should be admitted to a group, 

 which I own myself unable to define. But I do not think 

 it is really so. There seem to be two conceivable methods 

 of breaking up a genus into subgenera. One is deductive 

 throughout. Some particular characters are assumed to be 

 of superior importance, and on these categories are based, 

 and species parcelled out accordingly. The other is at 

 first inductive. We observe a number of species to be 

 more or less similar. We might examine into and take 

 stock of their points of agreement and difference, not so 

 much asking " In what one point do all these agree ? " as 

 " What are the points in a majority of which they agree ? " 

 and " Among these points which, or what combination of 

 them, can we think likely to indicate any real connection 

 between the species, and to separate them from the 

 general mass ? " Thus, by degrees, we might arrive at the 

 conception of a group, founded not on one character, but 

 on a preponderance of characters. And we should then 

 have to consider whether some of these species which had 

 given us our first crude conception of the group should 

 not after all be rejected from it, and the conception of it 

 modified accordingly. Subgenera formed in this way 

 must at first be tentative and elastic, all species being, as 

 it were, admitted to them on probation. But, if there were 

 any real connection between any of the latter, step by step 

 we might hope to trace it out, and define our group more 

 and more in accordance with the facts. And if there were 



