12 THE DISPERSAL OF PLANTS [pt. i 



from the parent. This dispersal impHes the concurrence of various 

 circumstances, and when all of these are external to the plant 

 it is spoken of as occasional or accidental, while when some are 

 inherent in the nature of the plant itself, it is said to take place 

 by aid of the regular "mechanisms." As instances of "irregular" 

 dispersal, we have such cases as the carriage of heavy seeds by 

 a hurricane, or their casual attachment to a log which is acci- 

 dentally floated across the sea to a new country; whilst it is 

 "regular" in the case of seeds so hght that they will always be 

 carried by wind to some little distance, or fleshy fruits which 

 are eaten by birds and the seeds subsequently dropped. It may 

 prove of more interest if an account be given of some actual 

 researches carried out upon this subject, rather than a mere 

 enumeration of the various mechanisms, etc. (54, 71). 



My chief pleasure in life being travel, I have always been 

 interested in the movement of plants, and in 1893, with Mr I. H. 

 Burkill, published (137) a study of the flora foimd in the bowl- 

 like tops of the pollard willows that line the banks of the Cam, 

 especially from Cambridge to Ely. We examined about 4000 of 

 these trees, and counting each occurrence of one species in one 

 tree, whether represented by few or many individuals, as 1, and 

 only as 1, we obtained 3951 records. The tops of the trees being 

 about six feet above the ground, it is clear that without some 

 assistance seeds would be quite unable to reach them, though 

 Avhen once reached, a Millow top presents a virgin area of soil, 

 with no other species groAving there. There were some 200 to 240 

 species in the neighbourhood wliich if planted in the willows 

 would probably have been able to grow there, but of these we 

 found that only 80, or about a third, actually occurred, showing 

 that the presence of a barrier even so trifling as the height of a 

 willow Avas sufficient to exclude very many. Most of the plants 

 with well-marked "regular" mechanisms Avere among the 80, 

 though one missed Cornus (dogAvood), Salix, the avUIoav itself 

 (possibly it Avould not groAV in its oAvn humus), Pojmlus, the 

 poplar (jjossibly for the same reason, it belonging to the same 

 family), and a fcAV Compositae and the orchids. The commonest 

 plant in the tops AA-as Galium Aparine, the goose-grass, found in 

 644 trees, or over 16 per cent, of the total records. The fruit of 

 this plant has little hooks, so that it may easily chng to an 

 animal or a bird for time enough to be carried to a Avillow. But 

 it Avas also found to be largely used by birds in nest-making, 

 and probably the bulk of the records are due to this, for ripe 



