xvi PREFACE. 



+r 



heterostyled genera are added to the family. Dr. Koehne 

 does not believe that any species of Lagerstrcemia is, or 

 has been, heterostyled and trimorphic. He has also 

 sent me an outline of an important view, well worth fol- 

 lowing out, namely, that heterostylism has arisen 

 through the modification of plants which were tend- 

 ing to become polygamous or dioecious. 



It is stated at p. 187 that Mr. Leggett felt some 

 doubt whether Pontederia cordata was trimorphic and 

 heterostyled; but he has since written to me that his 

 doubts are removed: see also to this effect, in 'Bull. 

 Torrey Bot. Club,' vol. vi., 1877, p. 170. All three 

 forms of this Pontederia appear to be highly variable. 

 He informs me that humble-bees are the fertilisers. 



With respect to the origin of the dioecious state, 

 which is discussed in the beginning of the seventh 

 chapter, Hermann Miiller has given some interesting 

 remarks in < Kosmos,' 1877, p. 290. The same author 

 shows (Ibid., p. 130) that Valeriana dioica exists under 

 four forms closely allied to the four presented by 

 Rhamnus, as described in this same chapter. It is 

 much to be desired that some one should experimentise 

 on these forms and make out their meaning. Bernet 

 has published ('Bull. Soc. Bot. France,' torn, xxv., 

 1878) a paper entitled, ' Disjonction des sexes dans 

 VEuonymus Europwus,' which may be compared with 

 my observations on the same plant. I have stated 

 at p. 297 that I could never find an hermaphrodite 

 plant of the common holly, but according to Mr 

 Hibberd (< Gard. Chron.,' 1877, pp. 39 and 776) such 

 occur among the many cultivated varieties. The evi- 

 dence, however, is far from conclusive, for it does not 

 appear that Mr. Hibberd ever observed under the mi- 

 croscope pollen taken from a plant known to produce 

 berries. Trees of Juglans cinerea in the TJ. States are 



