332 AUTOGAMY. 



As the above conclusion forms the main basis of the theory concerning the origin 

 of species to be presently unfolded, some preliminary account of the observations on 

 which it is founded must now be given, but the task is not an easy one. For five- 

 and-twenty years I have been studying, with special reference to the phenomena in 

 question, the flowers of many more than a thousand species in all stages of develop- 

 ment from budding to fruiting, some growing wild in their original habitats, some 

 in the Gardens under my direction, and my notes are so numerous that even the 

 briefest outline of the cases observed would fill several volumes. I must, therefoz-e, 

 content myself with presenting the different instances in groups according to their 

 degrees of similarity. Even the number of groups, however, is surprisingly large, 

 and only their general description can be attempted. The best way will be to 

 select a representative species for each kind of adaptation that we have to deal 

 with, and to point out in a few words its more important characteristics. As the 

 processes to be described as taking place during the flowering period all tend to 

 the same result though diftering greatly in other ways, and as the terms used must 

 be applied over and over again to the phenomena which are common to all the 

 cases, it is impossible to avoid a certain monotony in the descriptions that follow, 

 and the reader is begged to exercise some patience in the perusal of this chapter. 



The simplest case of autogamy occurs as follows. The flower opens revealing 

 the stigma stationed in front of the entrance to the receptacle and already mature, 

 whilst the anthers are closely adherent to the stigma but are still closed. Autogamy 

 is, for the time, impossible, whereas cross-pollination may be effected through the 

 intervention of wind or insects. In the second half of the flowering-period the 

 anthers adjacent to the stigma open, and the stigma is instantly covered with the 

 pollen set free from them. There are only a few varieties of this simplest case of 

 autogamy worth mentioning. It has been observed to occur in particular in annual 

 plants with small flowers (e.g. Centunculus minimus, Geranium pusillutn, Litho- 

 spermum arvense), and again in several bulbous plants belonging to the section of 

 Liliflorese, e.g. several species of Fritillaria and Narcissus, all the species of 

 Trillium and Uvularia, and in a few species of Crocus. In Trillium grandiflorum 

 and Uvularia grandiflora, two anthers are stationed in each of the three angles of 

 the spreading stigmas, and in the process of autogamy only the pollen from that 

 half of the anther which is turned inwards is used, whilst the pollen from the out- 

 ward-facing halves of the anthers may be carried away by insects, even after auto- 

 gamy has taken place. In the Crown Imperial {Fritillaria imperialis) only the 

 pollen from the three longer stamens falls upon the stigma of the same flo^.er. 

 There are six stamens in the flowers of this plant, three of which are longer than 

 the rest and alone have their anthers appressed to the tridentate stigma. These 

 anthers open a day later than the others. Dehiscence is accompanied by a very 

 considerable shrivelling and shortening of the anthers, and the result of this con- 

 traction is that the liberated pollen is scraped off" the anthers by the edges of 

 tlie stigmatic lobes. In Crocus albifiorus, which covers alpine meadows in early 

 spring with its blossoms, the anthers at first rest with their arched backs in 



