322 The Mechanism of Evolution" in Leptinotaesa 



SOME RESULTS OF THESE ANALYSES. 



The analysis of the condition in the population covering a series of genera- 

 tions in the same location has not often been attempted, and when it has, not 

 infrequently the outcome was to arrive at incorrect results, owing to the short 

 time that the observations were continued, or to the wide gaps that intervened 

 between different determinations. One of the best examples of this is the experi- 

 ence of Kellogg with Diabrotica soror in California, who examined the popula- 

 tion of this beetle, finding certain conditions in the population as expressed in 

 the variations of the elytral pattern, and later, with an interval of several genera- 

 tions, a second set of determinations gave results that looked like movement of 

 the population, which was strengthened by the determination in one or two 

 succeeding generations without change ; and so the conclusion was reached that 

 the population had changed in one direction with respect to this character, 

 between the first and the second censuses, and the results were described as 

 determinate evolution. Later determinations showed that the population had 

 moved back to the first position with respect to the same wing characters. What 

 happened in this instance is that the lack of continuity in the observations gave 

 false appearance in the determinations taken at random times. Out of the 

 data that I have had at my disposal, any number of instances of this sort could be 

 realized by precisely these methods. 



The data that I have presented in the preceding pages would, by the statisti- 

 cian and his measuring-stick, be massed by force into curves of error, interpreted 

 in terms of some hypothesis, and then put forth as " proof " of the same hy- 

 pothesis. So, also, the f aunistic and ecological worker and the geographic varia- 

 tion worker could and would utilize the arrays presented as examples of the 

 direct action of environment upon an extremely labile species, and derive 

 various agents that may or may not have been operative in the actual conditions 

 in nature. These plausible relations, easy to create, soon become fixed in the 

 minds of the worker and are soon " established facts " in support of the especial 

 proposition with regard to the action of geographic influences. 



Some, no doubt, conceive of the population in terms of minor groups which 

 the census does not reveal, and can therefore be interpreted upon the basis of the 

 presence in the population of numerous elemental conditions "without inter- 

 grades," but whose " fluctuations " overlap and thereby produce the appearance 

 of continuity, wherein there is no real " continuity " excepting in the " fluctua- 

 tions of the minor groups." No census can show anything with regard to the 

 truth of this proposition. 



With these materials I have gone carefully over the series and have applied 

 the census methods, but nowhere have I been able to discover that the real nature 

 of the materials and of the processes at work was even indicated. These 

 are problems for the laboratory and the experimentalist and not for the 

 statistician. 



I, for one, do not care to discard either the method or the point of view of the 

 statistician, as some would do, because the method does give information con- 

 cerning the mass of the population that is of real value in the attempt to experi- 

 mentally analyze its constitution and the processes at work; but more, it gives 

 accurate information that could never be provided by the laboratory and the 

 experimental garden with respect to the sum total result produced in and the 

 aspect of the population after it has been subjected to the operations of nature. 



