viii PREFACE TO THE REPRINT OF 1884, 



the first rank in the chain of causes to dichogamy. Simi- 

 lar views are given in the present edition, p. 383, in 

 connection with observations by Hildebrand. 



In a subsequent paper (' Botan. Centralblatt,' 1880, 

 iv. p. 839) he describes a similar gynodicecious condi- 

 tion in some Stellarias and Cerastiums. Here there 

 are pure female, semi-female, and hermaphrodite plants, 

 the flowers of the female form being smaller than those 

 of the others. This distribution of sex he calls " gyno- 

 dimorphism," a condition which he describes ('Bot. 

 Centralblatt,' 1880, p. 1031) as occurring in Arenaria 

 ciliata and Alsine verna. 



F. Ludwig ('Kosmos,'* 1880-81, viii, 357) de- 

 scribes two forms of Erodium cicutarium. The first, 

 which is distinguished by having nectar-guides, is pro- 

 tandrous, and adapted for fertilisation by insects. The 

 second form is weakly protogynous and autogamic. 

 This form has no nectar-guides, and the petals are usu- 

 ally shed during the day on which the flowers open. It 

 resembles E. moschatum, which is homogamie (or weak- 

 ly protogynous) . The first form is more like E. macro- 

 denum, which is markedly protandrous, and where auto- 

 gamy is impossible. 



Hermann Mtiller ('Nature,' xxiii. p. 337, 1881) 

 has shown that Syringa persica is gynomoncecious, hav- 

 ing on the same inflorescence a majority of hermaphro- 

 dite flowers of large size, and a minority of small-sized 

 females. 



Stellaria glauca and Sherardia arvensis are gyno- 

 dicecious. 



H. Miiller has also written an important paper on 

 Centaurea jacea ('Kosmos,' x., and "Nature,' xxv.) in 

 which he publishes his change of opinion as to the origin 



*See also 'Irmischia,' 1881, No. 1, and 'Bot. Centralblatt,' xii. p. 

 83, and viii. p. 87. 



