-0 Trantiactlons. 



white rioNVPrs, wliicli are scented, and mostly open during the 

 evening. Now, the coniinon ancestor of the two species may 

 have had red flowers, but with a tendency (by reversion to still 

 earlier conditions) to produce white flowers occasionally, just as 

 we see happens with Lychnis floscucidli in our own district. 

 This tendency would probably be most frequent in those flowers 

 which opeii towards evening through the absence of strong 

 sunlight, and such white, late opening flowers would be best 

 visible to evening moths, and be frequently crossed between 

 themselves. The red day flowers of diiirna would be visited, as 

 now, by buinble-bees, which retire to rest before the evening 

 moths come out, and hence the flowers of each incipient variety 

 would be constantly crossed with each otiier, and but seldom with 

 tiiose of the other variety. White colour, late hours, and scent 

 being all directly of service to moth flowers, would be ti.xed by 

 natural selection, and the two incipient species would diverge 

 more and more widely, and liave room to produce the other veiy 

 minor distinctions which now separate them. Again, why should 

 Lepvjouum or Spergidaria rubra be pink 1 It is most unusual 

 in the section of Caryophyllece to which it belong.s, and I could 

 not have answered this question before this summer. I now And 

 it is visited not infrequently by hive-bees, and the pink is 

 obviously to assist it against the strong competition of Arnieria 

 vulyaris, which often accompanies it. One must, however, be 

 careful not to apply the principle too universally. Most flowers 

 cannot rely entirely on one class of visitors, and though Sir John 

 Lubbock has siifliciently proved by direct experiment that bees 

 do prefer red to yellow and blue to red, still they do not by any 

 means confine their attention to red and blue flowers. Thus the 

 hive-bee visits the common yellow buttercup, and bumlfle-bees 

 often gather pollen from St. John's Wort and the Willows in 

 early spring. None of these three plants are of course specially 

 given up to bees. Even in cases like that of Lamium album, one 

 must remember that one has maiidy to explain why some floweis 

 have turned red or purple ; it is not necessary to supjjose that all 

 bee-flowers must do so, as they may use tJieir surplus material in 

 other ways. 



Everyone knows that one of the great subdivisions of botany is 

 that of CoroUiflorai — flowers, that is to say, in which the petals 

 are united to form a corolla. I think, however, few can have 



