204 ANGIOSPERMAE DICOTYLEDON ES 



described, but the corolla-tube is somewhat shorter than in M. pratense, though this 

 does not render the nectar legitimately accessible to shorter-tongued insects than in 

 that species. The corolla-tube rises vertically for 5-6 mm., then bends sharply and 

 runs horizontally for 7-7-2 '^' further. It is scarcely i mm. broad, though rather 

 more than i mm. high, but these dimensions increase to 2 and 4 mm. in front, and 

 it is only into this wide part that the head of a humble-bee can be inserted. A 

 proboscis at least 12 mm. long is therefore necessary to suck out all the nectar. 

 The entrance of unbidden guests is checked by the tolerably close apposition of the 

 lower lip with the hood-like upper one. 



Visitors. The following were recorded by the observers, and for the localities 

 stated. 



Herm. MuUer (in the wood at Kitzingen), a butterfly (Melitaea athalia Rott., 

 vainly trying to suck) and the humble-bee Bombus lapidarius L. 5 (proboscis 

 12-14 mm.), skg. legitimately. Schulz saw flowers perforated by humble-bees. 



2146. M. sylvaticum L. (Sprengel, ' Entd. Geh.,' pp. 315-16; Herm. 

 Miiller, ' Weit. Beob.,' Ill, pp. 39-41; Schulz, ' Beitrage,' II, p. 218; Loew, 

 ' Bliitenbiol. Floristik,' p. 399.) In this species the tube of the small dark-yellow 

 corolla is about i mm. wide. It ascends obliquely for about 3 mm., and then bends 

 horizontally to continue 5 mm. further, widening gradually as it does so and finally 

 dividing into a roof-like upper lip with margins fringed by soft dependent threads, 

 and a three-lobed under lip serving as an alighting-platform. There is no 

 constriction of the entrance to the flower as in other species of the genus, but it 

 widens fairly regularly to an aperture 3 mm. high and broad. The filaments lie 

 close against the outer wall of the corolla-tube and bend together under the upper 

 lip, so that the anthers are close together behind the hairy part of this, and turn their 

 dehisced surfaces downwards. The style lies between the filaments and behind the 

 anthers, and its end bends forwards and downwards, bringing the stigma into 

 the entrance of the flower. The proboscis of an insect probing for nectar must 

 therefore first touch the stigma and then the pollen-covered surfaces of the anthers, 

 thus effecting crossing. Hermann Miiller was unable to find any special nectary ; 

 nectar seems to be sparingly secreted by the lowest part of the ovary. Vestiges of 

 nectar-covers appear to be present on the inner wall of the corolla-tube where it 

 bends horizontally. 



Automatic self-pollination takes place should insect-visits fail, for the stigma 

 comes to lie under the anthers and gets dusted with pollen. This is not so dry and 

 powdery as it is in M. arvense and other species, and remains clinging for a longer 

 time to the under-side of the dehisced anthers. 



Visitors. The following were recorded by the observers, and for the localities 

 stated. 



Buddeberg (Nassau), the humble-bee Bombus senilis Sm. (= B. muscorum F^ 

 5, skg. Herm. Miiller (Bavarian Oberpfalz), the wasp Vespa rufa Z. J, on several 

 flowers. Schulz (Riesengebirge), perforated flowers. Loew (Alps), 3 hover-flies as 

 useless visitors i. Chrysotoxum octomaculatum Curt.', 2. Syrphus luniger Mg., 

 skg. ; 3. S. lunulatus Mg. 



