PART IV. | GENERAL RETROSPECT. 581 
of the honey, The Composite, however, furnish several series of forms well 
fitted to demonstrate this point. Omitting some aberrant forms, the honey is 
on the whole least deeply situated in the Senecionida, more deeply placed in the 
Cichoriacew, and most deeply in the Cynaree, On (a) ten Seneciunide (Nos. 
216, 222, 224, 226, 227, 228, 232, 233, 236, 237), I have observed 335 different 
insect-visits, or an average of 33°5 to each species ; on (b) fifteen Cichoriacee 
(Nos. 259, 261—273, 275), 356 distinct visits, or an average of 23 to 24 ; 
and on (c) ten Cynarew (Nos. 240, 241, 244, 245, 249, 252, 254, 256, 257, 258), 
189, or an average of 18 to 19. The following table is similar to the preceding 
one, and shows the proportionate number of visits paid by each group of 
insects, reckoning the total number of visits observed at 1,000. 
(1) Specially long-tongued bees, Bombus and Anthophora : (a) 15 ; (b) 48; 
(c) 211. 
_ (2) Bees with abdominal collecting-brushes (and long proboscides) : (a) 27 ; 
(b) 48; (c) 181. 
(3) Other bees with long or moderately long proboscides : (a) 42 ; (b) 126; 
(c) 85. 
(4) Sphecodes, Andrena, and Halictus (proboscides moderately long) : (a) 
167 ; (b) 399; (ec) 196. 
(5) Prosopis and Oolletes (short-tongued) : (a) 30; (b) 8; (c) 5. 
(6) Other Hymenoptera (mostly short-lipped) : (a) 137 ; (b) 17; (c) 35. 
(7) Long-tongued, purely suctorial Diptera (Bombylius, Empis, Conopide) ; 
(a) 42; (b) 42; (c) 35. 
(8) Rhingia: (a) 3; (b) 3; (ce) 10. 
(9) Syrphide with shorter tongues than Rhingia (Eristalis, Helophilus, 
Volucella) : (a) 92; (b) 84; (e) 55. : 
(10) Short-tongued Diptera: (a) 242; (b) 121; (ce) 10. 
(11) Lepidoptera: (a) 80; (b) 67; (c) 171. 
(12) Coleoptera : (a) 116; (b) 34; (c) 465. 
(13) Hemiptera and Panorpa: (a) 9; (b) 3; 10. 
This table proves clearly that in Composite as the honey becomes more 
deeply placed the visits of the more highly specialised bees increase, while, in 
spite of the exposed situation of the pollen, the visits of flies fall off. This is 
true of the number of species of visitors ; but if it were possible to record the 
number of visits paid by each species, the preponderating influence of bees 
would be still more apparent. From these first stages in elongation of the 
corolla-tube and increase of bees’ visits by exclusion of short-lipped insects, we 
pass by the most gradual steps to the exceedingly long tubes whose honey is 
monopolised by a smaller and smaller number of bees, At the top of the 
series stand flowers whose tubes are 16 to 20 mm. long, and whose honey is only 
accessible to a few species of Bombus and Anthophora (Aquilegia, Delyhinium, 
Pedicularis, Lamium maculatum, etc.). 
It would not be possible without many more observations to classify our 
flowers in a long series of groups, each one restricted toa smaller circle of bees, 
The length of the tube is not of itself a measure of the proboscis necessary to 
reach the honey ; for often the mouth of the flower is widened to admit the 
insect’s head or even part of its body, and further the honey often accumulates 
to a considerable depth in the tube. Such conditions are illustrated in many 
Silenee, Boraginee, Scrophulariacee, Ericacee, etc. 
