- 



I 



1917] 



ROBERTSON— FLOWERS AND INSECTS 



313 



When the lower Hymenoptera together make only o . 8 per cent 

 of the visits to the non-social Papilionaceae, it is evident that they 

 have had little to do with the evolution of the Papilionaceae, even if 

 they were instrumental in their origin. To support the latter con- 

 dition it would be necessary to show that the non-social forms were 

 developed from the social forms. 



Visits to Labiatae show: 



Non-social (13) . 

 Social (12) 



Total (25) 



Lycopus, etc. (5) 



Bees 



Diptera 



Other 

 Hymen- 

 optera 



Lepidop- 

 tera 



Coleop- 



tera, 



Hemiptera 



831 



395 

 44.6 



29.6 



5-8 

 20.8 



19.0 



24.6 



0.8 

 24.6 



21.8 



34- S 



IO. O 



12.7 



12.4 



7.8 



O.O 

 2.2 



1-9 



3-2 



Total 



119 



897 

 1016 



576 



When the lower Hymenoptera show 24.6 per cent of the visits 

 to social Labiatae and only 0.8 per cent to non-social Labiatae, 

 it is hard to connect them with the origin of the Labiatae unless we 

 suppose that the non-social developed from the social. Muller 

 (Fertilisation of flowers, p. 471) says: "Delpino considers Mentha 

 and Coleus degraded forms of the labiate type; he, however, gives 

 no reason for thinking them to be such, and not rather less special- 

 ized forms, differing less from the common ancestors of the Labia- 

 tae/' If there are non-social zygomorphous wasp flowers or 

 ichneumon flowers, no doubt they should be regarded as modified 

 from bee flowers. 



The view held here, that the early flowers were non-social and 



modified in connection with the visits of bees, and 



flow 



ers in 



linly visited by other insects are later, is supported 

 by what is known of the behavior of insects and by inferences from 

 the affinities of the flowers. Of course, if it can be shown that the 

 primitive flowers were social and that the non-social flowers were 

 developed from them, this view will have to be abandoned for 



that of Muller. 



Of the total visits of the lower insects, 88.0 per cent are to social 

 flowers, and of the total insect visits to social flowers the lower 

 insects make 62.7 per cent. Xow the flowers which these insects 



