66 INVASION 



The primary causes of endemism are two, lack of migra- 

 tion and presence of barriers. Since distributive agents are 

 practically universal, lack of migration corresponds essenti- 

 ally to immobility, a fact which decreases the difficulty of 

 ascertaining the immediate causes of endemism in any par- 

 ticular species. Either immobility or a barrier may pro- 

 duce endemism; extremely immobile plants, for example, 

 liliaceous species propagating almost wholly by underground 

 parts, are as a rule endemic, while alpine plants and those 

 of oceanic islands are endemic in the highest degree, re- 

 gardless of their mobility. When the two conditions act 

 concomitantly upon a species, endemism is almost inevit- 

 able. It cannot be supposed, however, that immobility or 

 natural barriers alone, or the concomitance of the two, must 

 invariably give rise to endemic species; the most immobile 

 plant may be carried into another region by unusual or acci- 

 dental agencies, or the most formidable barrier to migration 

 may be overcome by the intensity of an agent or through 

 the action of man. Endemism is also brought about by the 

 modification of species : new or nascent species are as a rule 

 endemic. Whether they will remain endemic or not will de- 

 pend upon the perfection of their contrivances for dissemin- 

 ation and upon the presence of barriers to migration or 

 ecesis. Finally, as Drude was the first to point out, the dis- 

 appearance of a polydemic species in all regions but one, 

 owing to the struggle for existence or to changed physical 

 conditions, will result in endemism. 



Endemism is readily recognized by methods of distribu- 

 tional statistics, applied to areas limited by natural barriers 

 to migration or ecesis. For political areas, it has no signi- 

 ficance whatever, unless the boundaries of these coincide 

 with barriers. It determines in the first degree the validity 

 of regions, though the latter are often recognized also by 

 the presence of barriers and by the character of the vege- 

 tation. Endemism may occur in areas of vegetation of any 

 rank from a formation to a zone. Comparisons to be of 

 value, however, can be instituted only between areas of the 

 same order, i. e. between two or more formations, two or 



