CH. XX] THE ORIGIN OF SPECIES 219 



original species, which we must look upon as the probable parent 

 {e.g. Dillenia indica, p. 159), often survives in the same locality. 

 A few cases like Acrotrema dissectum (Trimen's Flora ofCeijlon,\ 

 p. 9), where intermediate forms {possibly hybrids) occur, have 

 been noticed, but more usually the local species is fairly well 

 distmguished from the wide-ranging form. And in some instances 

 transitions are impossible, as, for example, with Coleus elongatus 

 (detailed characters given on p. 152). I may quote here what 

 has already been said about it in a paper of 1907 (118): 



''The species is too entirely different from the other species 

 ot Coleus, whether we take C. barbatus or one of the others, for 

 evolution by means of continuous variations to have been 

 possible. To take some of the characters, especially those that 

 are most prominent, how is the one type of innorescence goin^ 

 to develop into the other by any possible continuous variation? 

 Ihe mind cannot conceive of such a process, unless it be by dis- 

 continuous variation. Still more, how is a calvx with one big 

 tooth on top and four small ones below going to develop into 

 one with five equal teeth? The study of infinitesimal variation 

 shows that the maximum change to be expected in one generation 

 would be a mere fraction of the width of a tooth, and how is this 

 to prove of sufficient advantage or disadvantage to be of any 

 material import in the struggle for existence? The question is 

 equally hard if we suppose a common ancestor, for what kind of 

 calyx or inflorescence will be intermediate?" 

 And cf. above, p. 209, as to changes in caUx teeth. 



In this case the species that one must regard as ancestral, 

 C. barbatus, is also found in the same locality; it is as frequent 

 on the summit of Ritigala as C. elongatus, and grows in similar 

 spots on the exposed rocks. Both suit the same conditions, and 

 if they have descended from a common ancestor, not one from 

 the other, it is very remarkable that one should be confined to 

 Ritigala, one common to tropical Asia and Africa. 



Nearly seventy years ago, Lyell (69, p. 39-^) said "Might not 

 the births of new species, like the deaths of old ones, be sudden? " 

 and it appears to me, that when one puts together the facts of 

 distribution as understood in the light of Age and Area, and the 

 still more surprising fact of the agreement of the type of dis- 

 ]3ersal of species, both by area and into genera, and of genera by 

 area and into families, etc., as more fully described above, one 

 can hardly arrive at any other conclusion. Ad\-antage as a cause 

 in evolution seems to be ruled out with practical completeness, 

 though it will determine whether the newly evolved form will 

 survive or not; and if advantage cannot be adduced, then one 

 can hardly conceive of the changes that distinguish one species 



