292 ANGIOSPERMAE—DICOTYLEDONES 



The variety (b) nivale (=the species T. nivale Sieb.), which lives for the most 

 part above the limit of trees, and bears dirty white instead of red flowers, was seen 

 by Hermann Miiller to be visited in the Alps by 7 humble-bees and 17 Lepidoptera, 

 i. e. 7 1 % of the visitors were of the latter kind. The same investigator noticed that 

 the ordinary variety was visited in the same region by 1 5 humble-bees, 2 1 Lepido- 

 ptera, and a beetle, i. e. 55 % of the visitors were Lepidoptera (' Alpenblumen,' 

 pp. 241-3). 



I may here introduce an observation of Hermann Miiller's on a well-known 

 series of inferences. — ' The more cats there are, the fewer are the mice ; the fewer 

 mice there are, the more abundant are the humble-bees; the more numerous the 

 humble-bees, the more prolific is the red clover; therefore, the more numerous 

 the cats, the more prolific is the red clover.' It is true that humble-bees are the 

 chief pollinators (though not, as Darwin thought, the only ones) of the red clover, 

 but when they are excluded there are numerous nectar-sucking and pollen-collecting 

 insects that can effect the cross-pollination necessary for complete fertility. Hence 

 one link in the above chain of inferences will not hold, for it is wrong to say ' the 

 more numerous the humble-bees, the more prolific is the red clover '. The red 

 clover of New Zealand became wonderfully prolific after the introduction of about 

 100 humble-bees, there being no indigenous species (Dunning, Proc. Ent. Soc, 

 London, 1886). 



690. T. incamatum L. — According to Kirchner (' Flora v. Stuttgart,' 

 pp. 491-2), the flower mechanism of this species essentially agrees with that of 

 T. pratense. The length of the corolla-tube of the bright blood-red flowers is 

 8-9 mm., and that of the calyx-tube 5 mm. The vexillum is folded up, and there- 

 fore serves to guide the proboscis of the longer-tongued insects to the nectar. The 

 base of its limb also completely embraces the claws of the alae and carina, while its 

 own claw is free. The alae possess well-developed vesicular processes, gripping the 

 upper side of the sexual column ; also a longitudinal fold, united internally with 

 the epidermis of the carina. 



Automatic self-pollination is much less productive than cross-pollination. 



Visitors. — I observed the long-tongued bees Bombus lapidarius L. 5, an Eucera 

 longicomis L. 5 and $, in Mecklenburg; both skg. (' Bloemenbiol. Bijdragen'). 



Hoppner saw three humble-bees in Bremen. — i. Bombus agrorum F,; 2. B. 

 muscorum F. ; 3. B. variabilis Schmiedekn. 



691. T. alpestre L. (Herm. Miiller, 'Weit. Beob.,' II, pp. 247-8 ; Schulz, 

 'Beitrage,' p. 209.) — The flower-heads of this species, according to Hermann 

 Miiller, are larger and more brightly coloured than those of T. pratense, with which 

 the flower mechanism agrees in most respects. But while in the latter species the 

 corolla-tube is 7 mm. in length up to the point where carina and vexillum separate, 

 and II mm. long measured to the tip of the carina, in the former species these 

 dimensions are 1 1 mm. and 1 4 mm. respectively. A great many humble-bees are 

 therefore prevented from sucking nectar legitimately. The carina and alae being 

 shorter than the vexillum, introduction of the proboscis into the flower is rendered 

 more difiicult for bees and easier for Lepidoptera. Lastly, the carina is considerably 

 higher than the corolla-tube, and curves sharply upwards, so that the proboscis of 

 a Lepidopterid cannot reach the base of the flower except through the open cleft, 



r 



