DISTRIBUTION. 401 



water, and somewhat more fitted to it; seals living almost 

 exclusively in the sea, and having the mammalian form 

 greatly obscured; whales wholly confined to the sea, and 

 having so little the aspect of mammals as to be mistaken for 

 fish. Conversely, sundry inhabitants of the water make ex- 

 cursions on the land. Eels migrate at night from one pool 

 to another. There are fish with specially-modified gills and 

 fin-rays serving as stilts, which, when the rivers they inhabit 

 are partially dried-up, travel in search of better quarters. 

 And while some kinds of crabs do not make land-excursions 

 beyond high-water mark, other kinds pursue lives almost 

 wholly terrestrial. 



Guided by these two classes of facts, we must regard the 

 bounds to each species' sphere of existence as determined 

 by the balancing of two antagonist sets of forces. The tend- 

 ency which every species has to intrude on other areas, other 

 modes of life, and other media, is restrained by the direct 

 and indirect resistance of conditions, organic and inorganic. 

 And these expansive and repressive energies, varying con- 

 tinually in their respective intensities, rhythmically equili- 

 brate each other maintain a limit that perpetually oscillates 

 from side to side of a certain mean. 



106. As implied at the outset, the character of a region, 

 when unfavourable to any species, sufficiently accounts for the 

 absence of this species; and thus its absence is not incon- 

 sistent with the hypothesis that each species was originally 

 placed in the regions most favourable to it. But the absence 

 of a species from regions that are favourable to it cannot be 

 thus accounted for. Were plants and animals localized wholly 

 with reference to the fitness of their constitutions to surround- 

 ing conditions, we might expect Floras to be similar, and 

 Faunas to be similar, where the conditions are similar; and 

 we might expect dissimilarities among Floras and among 

 Faunas, proportionate to the dissimilarities of their conditions. 

 But we do not find such anticipations verified. 



