322 DARWINISM chap. 



sandwort (Spergula), and some willow-herbs (Epilobium) ; or 

 they arch over the pistil, as in Galium aparine and Alisma 

 Plantago. The style is also modified to bring it into contact 

 with the anthers, as in the dandelion, groundsel, and many 

 other plants. 1 All these, however, may be occasionally cross- 

 fertilised. 



3. In other cases precautions are taken to prevent cross- 

 fertilisation, as in the numerous cleistogamous or closed flowers. 

 These occur in no less than fifty-five difl'erent genera, belonging 

 to twenty-four natural orders, and in thirty-t^vo of these genera 

 the normal flowers are irregular, and have therefore been 

 specially modified for insect fertilisation.^ These flowers appear 

 to be degradations of the normal flowers, and are closed up by 

 various modifications of the petals or other parts, so that it is 

 impossil)le for insects to reach the interior, yet they produce 

 seed in abundance, and are often the chief means by which 

 the species is continued. Thus, in our common dog-violet the 

 perfect flowers rarely produce seed, while the rudimentary 

 cleistogamic flowers do so in abundance. The sweet violet also 

 produces abundance of seed from its cleistogamic flowers, and 

 few from its perfect flowers ; but in Liguria it produces only 

 perfect flowers which seed abundantly. Xo case appears to 

 be known of a plant which has cleistogamic flowers only, but 

 a small rush (Juncus bufonius) is in this condition in some 

 parts of Russia, while in other parts perfect flowers are also 

 produced.^ Our common henbit dead-nettle (Lamium amplex- 

 icaule) produces cleistogamic flowers, as do also some orchids. 

 The advantage gained by the plant is great economy of 

 specialised material, since w^ith very small flowers and very 

 little expenditure of pollen an abundance of seed is produced. 



4. A considerable number of plants which have evidently 

 been specially modified for insect fertilisation have, by further 



^ The above examples are taken from Rev. G. Heuslow's paper on *' Self- 

 Fertilisation of Plants," in Trans. Linn. Soc. Second series, Botany, vol. i. 

 pp. 317-398, with plate. Mr. H. 0. Forbes has showTi that the same thing 

 occurs among tropical orchids, in his paper " On the Contrivances for insuring 

 Self- Fertilisation in some Tropical Orchids," Journ. Linn. Soc, xxi. p. 538. 



- These are the numbers given by Darwin, but I am informed by Mr. 

 Hemsley that many additions have been since made to the list, and that 

 cleistogamic flowers probably occur in nearly all the natural orders. 



^ For a full account of cleistogamic flowers, see Darwin's Forms of Flowers, 

 chap. viii. 



