THE PROBLEM OF THE ORIGIN OF SPECIES. 39 



De Vries dwells at length and with effect on this " transgressive variability," 

 as he terms it, endeavoring to make it abundantly clear that it is not really transi- 

 tional, and hence can have no value as a support for natural selection. 



The trouble with such series is that the sequence is always wholly artificial 

 and of doubtful significance, as it holds only for the extremes of single characters. 

 Moreover, if the extremes are taken from two distinct species, they can not of 

 course be interpreted as transitions looking to the origin of a new species. Such 

 prospective significance can not lie between two species already in separate existence. 



It does not follow, however, that any sharp line of qualitative distinction is to 

 be drawn between such variation and the variation (or mutation) which eventuates 

 on new species. In statistical attire, individual variations may seem to be amenable 

 to Quetelct's law, and to merely oscillate about a center which sets limits to their 

 excursions, but never follows them. Thus conceived, they may appear to be quali- 

 tatively different from variations that result in a shifting of the center of gravity 

 to a new face of the polyhedron, to use the famous simile of Galton. If gathered at 

 random and arranged arbitrarily, we may get as much chance out of them as we 

 put into them. But is it safe to conclude that the mathematical theory of chance 

 really covers all that is essential for a theory of such variations? 



It is on this assumed distinction that the whole theory of mutation rests. If 

 true, then natural selection is not directly concerned in the origin of new species. 

 Neither is there any room in such work for definite and continuous variation 

 advancing in one direction. No matter how small the difference between two allied 

 species, that difference is absolute and unmediated. Neither natural nor artificial 

 selection has any power to raise the species to a new specific level. The species 

 may be improved within its sphere, but the limit of such improvement is soon 

 reached, and retrogression to the original status is swift and inevitable if the species 

 is left to shift for itself. In the same way all human endeavor and progress is 

 tethered, with no possibility of escape through cumulative improvement. Stare 

 super antiquas vias is the utmost hope of the species. As de Vries remarks : 



A hard, a difficult, I should be inclined to say, a sad law. What is obtained with much 

 trouble and care can be preserved only by as great care and trouble. And history is witness 

 that this law is also true for mankind. Everywhere and always progress, but followed by 

 regression as soon as the effort ceases. 3 



Conclusions of so far-reaching import, based upon twenty years of observational 

 and experimental research that has not been equaled since Darwin's time, certainly 

 challenge our deepest interest. 



The distinction which threads this great work, binding it into one logical whole, 

 is one not easily brought to a decisive test. The behavior of species in hybridiza- 

 tion strongly suggests qualitative differences, such as are not apparent in ordinary 

 individual variation. The usual aversion of one species for another, and the general 

 infertility and instability of hybrids, seem to show that the behavior of species 

 towards one another is influenced by something more deeply seated than habit. 



Ordinary individual variability fluctuates in such an aimless way, and is so 

 unstable even when nursed continually in a definite direction and for long periods, 



3 "Unity and Variability." Inaugural address, University of Amsterdam, Jan. 8, 1S0S. Also, The University 

 Chronicle, Berkeley, Cal., Sept. 1898. 

 4 



