The Theory of Evolution ] 303 



cultural bias may be removed; at least the computer presumably 

 does not have a deep fear of miscegenation. 



A short example follows to show how numerical methods may 

 reduce the number of subjective decisions the investigator must 

 make in describing organic diversity. It must be emphasized that 

 the pilot study described below is not given in support of any 

 sweeping generality about the structure of nature. It is merely a 

 very brief example of one method that may be profitably used to 

 investigate that structure. The specimens studied are not treated as 

 a sample from a larger population; they are the entities to be com- 

 pared. It will be obvious to the reader that many more individuals 

 from many populations would have to be compared before the gen- 

 eral structure of even the Euphijdnjas edifha-chalccdona complex 

 could be elucidated. 



Thirteen male individuals of checkerspot butterflies {Euphijdnjas) 

 were compared on the basis of 75 characters of their skeletal mor- 

 phology and color patterns. By the standard methods of butterfly 

 taxonomy, eight of these individuals had been identified as belong- 

 ing to the species Euphijdnjas editha and three of them to the spe- 

 cies E. chalcedona. Two of the specimens came from a population 

 (Kings Canyon) which, on the basis of classic procedures, could 

 not be assigned with certainty to one species or the other. Therefore, 

 since the techniques of taxonomy did not permit such intermediacy, 

 these two individuals had been assigned by the taxonomists to what 

 seemed the most similar species, E. editha. The specimens and their 

 coded designations are listed in Table 13.1. 



In some areas of California E. editha and E. chalcedona are found 

 together, and no intermediate individuals have been observed. In 

 others, such as the Kings Canyon area, "problem" populations are 

 found, and the apparent distinctness blurs. At this point, note that 

 the taxonomist has, in essence, done the following: recognized that 

 the editha individuals are more similar to each other than to the 

 chalcedona specimens (and vice versa). Although recognizing that 

 the two Kings Canyon specimens are different from the editha, he 

 has decided that they are more like the editha than the chalcedona. 



The first step in the numerical analysis involved the determination 

 of the states of the various characters used in the comparisons. The 

 dimensions of markings on the wings, distances between the tips of 

 certain wing veins, sizes of structural features, etc., were measured 

 under a microscope. Other characters which did not lend them- 

 selves to direct measurement were quantified by coding into states. 



When all the characters had been evaluated for each specimen, a 

 digital computer was employed to calculate product-moment corre- 



