174 ANGIOSPERMAE—DICOTYLEDONES 



thinks it may well be supposed that the latter was founded upon exceptional cases. 

 Jordan says that the two anterior lobes of the nectary are more strongly developed 

 than the two posterior ones. 



The length of the spur debars short-tongued bees from sucking the nectar, 

 while the personate character of the flowers excludes flies, Lepidoptera, and beetles. 

 Only long-tongued bees can suck legitimately and effect crossing. They press 

 down the lower lip and creep into the flowers, which are adapted in size to those 

 able to suck the nectar. As the style and stamens lie against the upper lip such 

 visitors when sucking brush the dorsal side of their bodies against the simultaneously 

 mature anthers and stigma. Since the latter is placed between the two pairs of 

 anthers cross- and self-pollination are effected with equal facility. Automatic self- 

 pollination also easily takes place, but Darwin says that it is always ineffective. 

 Warnstorf describes the pollen-grains as yellow in colour, smooth, and almost 

 spherical when examined in water. 



As I have elsewhere pointed out (' Blutenbiol. Notizen '), there is an admirable 

 arrangement for conducting the proboscis of an insect to the nectar. When a humble- 

 bee opens the flower by depressing the lower lip it finds an orange-coloured nectar- 

 guide on either side of this, which is not simply, as usual, a mere colour-streak 

 running into the nectar-containing spur. It is, in fact, a ridge composed of closely 

 packed, almost erect orange-coloured hairs, and a smooth hairless space i mm. broad 

 is left between the two guides. A bee or humble-bee cannot force its proboscis 

 through these hairy barriers, but is obliged to thrust it along the smooth median 

 groove, and in doing this rubs the upper side of its head, prothorax, and mesothorax 

 against the stigma and anthers. 



Visitors. — Herm. Miiller observed various species of ants, freq., skg., and also 

 records the following bees. — 



I. Apis mellifica Z. 5, very freq.; 'To suck, it creeps almost entirely into the 

 flower and thrusts its head into the wide entrance of the spur, which it empties down 

 to a depth of 2-3 mm. It creeps out again with its back covered with pollen, and 

 proceeds more frequently to flowers at the same height on neighbouring plants than 

 to higher flowers on the same. In other cases I have seen the honey-bee bite a hole 

 in the spur, and empty it, as Sprengel describes. Sprengel has correctly described 

 its behaviour while collecting pollen ; ' It slightly separates the lower lip of the corolla 

 from the upper, and thrusts its head so far in as to reach the anthers and obtain their 

 pollen.' 2. Bombus terrester L. j, skg. legitimately ; ' It inserts its head, thorax, and 

 forelegs into the flower, then thrusts its proboscis (7-9 mm.) almost to the tip of the 

 spur, and emerges with the upper surface of its head, pro- and mesothorax thickly 

 covered with pollen. Sometimes it sweeps off part of this pollen with the brushes on 

 its fore- and midlegs, and places it on the hindlegs. Sprengel's idea that the large 

 humble-bees do not enter the mouth of the flower is accordingly erroneous.' 

 3. B. hortorum Z. 5, 5 and i; '1 have very frequently seen this bee empty the flowers 

 of their honey, which it can do more quickly, owing to the length of its proboscis 

 (17-21 mm.), than the preceding species. Even the males sometimes swept the 

 pollen off their heads with their forelegs, and always had a number of pollen-grains 

 on all their tarsal brushes.' 4. Megachile maritima K. 5 (prob. 8-9 mm.), skg. 

 5. Osmia aenea Z. 5 (prob. 9-10 mm.), freq., skg. and po-cltg. 6. O. leucomelaena 

 K. ( = 0. parvula Zzf/!) 5 (prob. 2-| mm.), po-cltg. 7. Anthidium manicatum Z. 

 and t (prob. 9-10 mm.), both skg. and 5 po-cltg. 8. Andrena gwynana Z. 

 (prob. 2^ mm.), po-cltg. 



