10 FlTTrilH IV)T.\XTST. 



fact, that it is not iio^sible to 1)0 in two places at the same time. 

 I had no hope, therefore, of doing our 'JOO species in a thorough 

 manner, so I have simply studied about 270, with much help from 

 ]Miss Ilannaj, Mr Armstrong-, and others, as well as I could manage. 

 The result is that I am firmly convinced that a flower's shape 

 and every detail of colour, scent, mode of ripening, &c., is entirely 

 dependent on the insects which carry its pollen. Thus, in the 

 order Labiatse of the fourteen species studied, I found bumble-bees 

 in every single case, except Mentha arvensis, where I should not 

 have expected them. 



The colour and two-lipped condition ai'o entirely suited to 

 these l)umble-bees, and this suitability is found in ijuite minute^ 

 details. 



But it is not safe to draw tables or to generalise in our 

 present knowledge of the question. For instance, on the common 

 bramble I found, with Miss Ilannay's assistance, the following- 

 insects — the cabbage white butterfly, hive bee, no less than four 

 bumble-bees (B. muscorum, ten-estris, Derhamellus and pratorum), 

 and only two flies or diptera, and these latter wei-e not couunon 

 sorts, but of a complicated and intelligent type (Eristalis pertinax 

 and sericomyia borealis). I should have expected the sort of simple 

 and stupid type of fly which one finds, e.g., on the strawberry, to 

 which the bramble flower is not so very different. The bramble 

 enjoys this select set of visitors, probably because the flowers 

 appear so late in the season that these bees are not tempted to 

 visit other forms, but no one would have expected such a result. 

 In spiti^ of the \ast amount of oliservation yet rer|uii';'cl. it is safe 



