250 
Willis and Bur kill. — Floivers and 
TABLE II. 
Class. 
Per cent, of total 
visits. 
Per cent, of Lepi- 
doptera visits. 
Per cent, of Long- 
tongued Bees. 
Per cent, of Short- 
tongued Bees. 
Per cent, of other 
Hymenoptera. 
Per cent, of Long- 
tongued Flies. 
Per cent, of Short- 
1 tongued Flies. 
Per cent, of Coleop- 
tera. 
Per cent, of others. 
Po 
7-32 
_ 
1.4 
_ 
3-33 
11. II 
II.29 
17-39 
4-54 
A 
21.74 
10-4 
1.4 
21.43 
76-66 
l6-66 
30-64 
13.04 
18.18 
AB 
1.65 
— 
— 
— 
• — 
2-22 
4-°3 
— 
— ■ 
B 
11. 11 
16.66 
15.28 
7.14 
— 
13*33 
8.87 
17-39 
— 
B' 
46-33 
64-6 
37*5 
7 M 3 
20.0 
50.0 
42.74 
39.12 
68.18 
H 
1 1. 11 
8-33 
41*66 
— 
— 
5-55 
2*42 
13.04 
9.09 
F 
2.8 
“ 
1. 11 
_ 
_ 
These two tables at once show the enormous share of visits 
obtained by class B', which is practically the order Com- 
positae. The figures will be seen to support the Mullerian 
theory of flowers very well. Taking first the Lepidoptera, 
they show a very great preference 1 (64-6-46*33) for Class B', 
which is the usual result shown in similar statistical tables 
(see Lit. 341 b). They do not appear to visit their own 
special class F, but these flowers were not watched at night, 
when the long-tongued moths, to which they are specially 
adapted, are on the wing. The only other class to which 
they show any preference is B. Most other observers have 
found them rather to avoid this class. They avoid, more or 
less, all the remaining classes. The long-tongued bees, as 
usual, show a very marked preference (41-66-11*11) for 
class H, and a small avoidance of B'. The short-tongued 
bees prefer especially B'. The long-tongued flies show very 
slight preferences for B', B, Po ; they rather avoid A, a result 
at variance with those of other workers. The short-tongued 
5 Preference of a particular class of flowers, X, by a particular group of insects, 
Y, is shown by the per cent, of Y visiting X exceeding the per cent, number of the 
total insect visitors to X. 
