42 ILLINOIS BIOLOGICAL MONOGRAPHS [436 



form — and its absence is associated with extreme reduction of the 

 markings in general. Thus patterns made up of a row of dots on the 

 costal side of the elytron are the most numeroiis in Iowa localities and 

 probably those just east of the Missouri River. Thus the selected 

 classes of individuals are geographic in their relations and hence true 

 classes. Further evidence for this statement is shown in plate XXXIV 

 where the color differences are indicated, showing that the immaculate 

 forms are further divided into races on the basis of color. Those of 

 the humid southern states are green, and those of the western steppe, 

 with its dry early summer following early spring rains, are red. 



In full accord with the experimental results cited above are certain 

 differences in patterns of two localities from which collections were 

 made often. The larvae used^ in experiments were collected from a 

 point just north of the village of Miller, Indiana, from a small area of 

 oak dimes about an acre in extent. Adults were collected from this 

 same locality dviring several years at various times in the season and 

 differences in color and pattern were noted. Graph 10 is the distribu- 

 tion of classes in 200 individuals belonging to the generations of 1904 

 and 1905. This same graph is repeated above on a smaller scale with 

 graph 11 added, which shows the distribution of classes in .51 speci- 

 mens collected from the same area in April, 1906. Graph 12 shows the 

 distribution of classes in a series of 60 specimens collected in the north- 

 western part of Gary, (600 ft.) (Pine Station, Indiana,) in April, 

 1906, showing the modal class to be o instead of q and a small percent- 

 age of individuals with markings joined. Graph 13 shows the distribu- 

 tion of classes in a series of 37 specimens collected in September, 190S, 

 in which the same difference is shown. A difference in the distribution 

 of classes is indicated by a comparison of Graphs 12 and 13. These 

 differences are striking for one who is familiar with them. The dif- 

 ferences between the Gary and the Miller locality were noted while 

 collecting the species in the two localities during several years. The 

 specimens collected in Gary showed those with markings joined as very 

 rare. The entire series from the Gary locality show the same thing. 

 There are also similar differences from generation to generation, in the 

 catches from Miller. The difference in the conditions at Miller and in 

 the Gary locality is striking particularly during the larval and pupal 

 periods. The area in Gary is covered with scattered pines and in 

 places from which some of the specimens were collected cottonwoods 

 occur. The area is one of lake sand on which cottonwoods grow up 

 and are succeeded by pines and the pines by oaks. The Miller locality 

 is an oak dune area with well-established growth of oaks. One mile 

 south of the Gary locality are oak covered ridges. Specimens from 

 jiere are of the usual type taken in the Miller locality. Many of the 



