266 ON THE ORIGIN OF SPECIES 



as they grew, of many long intervals of time and changes 

 of level during the process of deposition, which would 

 never even have been suspected, had not the trees 

 chanced to have been preserved : thus Messrs. Lyell 

 and Dawson found carboniferous beds 1400 feet thick 

 in Nova Scotia, with ancient root-bearing strata, one 

 above the other, at no less than sixty-eight different 

 levels. Hence, when the same species occur at the 

 bottom, middle, and top of a formation, the probability 

 is that they have not lived on the same spot during the 

 whole period of deposition, but have disappeared and 

 reappeared, perhaps many times, during the same geo- 

 logical period. So that if such species were to undergo 

 a considerable amount of modification during any one 

 geological period, a section would not probably include 

 all the fine intermediate gradations which must on my 

 theory have existed between them, but abrupt, though 

 perhaps very slight, changes of form. 



It is all-important to remember that naturalists have 

 no golden rule by which to distinguish species and 

 varieties ; they grant some little variability to each 

 species, but when they meet with a somewhat greater 

 amount of difference between any two forms, they rank 

 both as species, unless they are enabled to connect them 

 together by close intermediate gradations. And this 

 from the reasons just assigned we can seldom hope to 

 effect in any one geological section. Supposing B and 

 C to be two species, and a third, A, to be found in an 

 underlying bed ; even if A were strictly intermediate 

 between B and C, it would simply be ranked as a third 

 and distinct species, unless at the same time it could be 

 most closely connected with either one or both forms by 

 intermediate varieties. Nor should it be forgotten, as 

 before explained, that A might be the actual progenitor 

 of B and C, and yet might not at all necessarily be 

 strictly intermediate between them in all points of 

 structure. So that we might obtain the parent-species 

 and its several modified descendants from the lower and 

 upper beds of a formation, and unless we obtained 

 numerous transitional gradations, we should not recog- 



