144 ANGIOSPERMAEDICOTYLEDONES 



automatic self-pollination the rule. The other form is large -flowered, chiefly 

 visited by humble-bees, and strongly protandrous with the stigmatic branches 

 always projecting beyond the anthers, thus making cross -pollination necessary. 

 The latter form, provided with red or violet-red nectar-guides above the nectar- 

 passages, was also observed by Kirchner in the Tyrol and MacLeod in Belgium. 

 Indeed MacLeod found no less than four distinct forms growing on the dunes 

 at Blankenberg. (i) The one described above after Sprengel and Hermann 

 Miiller, (2) A large -flowered form 35 mm. in diameter, with the corolla of 

 a bright reddish colour above, and a broad red wavy band surrounding its 

 central and lower yellowish part. (3) A small-flowered form similarly coloured 

 and marked, with two curved appendages at the base of each filament, though 

 these may be reduced or entirely absent. The small stiff projections on the 

 edges of the filaments are ill- developed, besides which the filaments may be 

 interwoven so as to close the nectar- openings. (4) A female form with short 

 stamens and reduced anthers. 



Schulz also observed, though rarely, gynomonoecism and gynodioecism. 



According to E. Heckel (quoted in Ludwig's 'Lehrbuch d. Biologic,' p. 30), 

 a smut- fungus (Thecaspora hyalina Fingerh. = T. capsularum Desm.) induces 

 floral dimorphism. In the most diverse parts of France this is associated with 

 the presence of a spider, Thomisius onustus Walck., which destroys the insect 

 visitors. Obviously plants which are on this account forced to resort to self- 

 fertilization will become weakened by in-breeding and more exposed to the attacks 

 of the parasitic fungus that causes modification of the flowers. 



The investigations of Schilberszky do not support Heckel's view. Besides 

 normal macrandrous flowers he distinguished abnormal micrandrous ones, it being 

 a case of heterandry, though of purely pathological or teratological nature. 



In the micrandrous flowers Schilberszky found the conidial form of a smut- 

 fungus on the angular nectary and anthers, not only of open flowers, but also 

 of many buds, even those in a very immature condition, which would be impossible 

 if Heckel's theory were correct. If, continues Schilberszky, the supposed weakening 

 of plants by in -breeding were a fact, all the flowers on the same stock would 

 necessarily be of the same kind. But this is not the case, for normal and 

 infected ones are borne on the same stock, especially if there are many branches 

 Infection takes place when the seed begins to germinate, for it is then oftei 

 surrounded by a perfect crust of germinating spores. The mycelium then penetrates! 

 into the growing tissue of the stem, and ultimately makes its way along the peduncle 

 to the flower-buds, where it develops first conidia and then chlamydospores. 



With reference to the presence of the spider Thomisius onustus in Convolvulus^ 

 Schilberszky remarks that it lies in wait for insects inside the flowers of varioi 

 Hungarian plants. He concludes that its presence has nothing to do with cai 

 the appearance of micrandrous flowers in this species. 



Warnstorf also found the conidial form of Thecaphora capsularum on manj 

 anthers of convolvulus at Neu-Ruppin. In this case all the pollen -grains 

 destroyed, while the anthers are of a dirty-brown colour and are borne on shot 

 filaments at the bottom of the corolla. Warnstorf has so far only noticed the 

 infestation on the small-flowered variety, but since he has found many hundred 



