114 



VARIATION IN I.KPTINOTARSA. 



resent a line of development in which the three species are steps in a certain 

 direction of evolution. It would be difficult to invent a hypothesis to account 

 for this case other than the one advanced above that would not contain obvious 

 absurdities. 



The species multitmviata, oblongata, intermedia, melanthorax, ruhicunda, 

 and decemlineata, which are very closely allied forms, are all, excepting 

 decemlineata, confined entirely in their distribution to the Mexican table-lands. 

 This is suggestive. As I have shown in the first chapter, we may regard mul- 

 titceniata as the central species in this group, variable, plastic, and able to live 

 in diverse habitats. It has spread over a wide area, and as it has been dissem- 

 inated new species have been produced. In all their variations this series of 

 species shows the same kind and direction of modification ; they differ, how- 

 ever, in the extent of their variation and in the portion of the possible scale 

 thereof which they exhibit. Thus, multitceniata, intermedia, and decem- 



OBLONQATA 

 RUBICUNDA 



DECEMLINEATA 



t 



INTERMEDIA 



MULTIT/ENIATA 



MELANOTHORAX 



SIGNATICOLLIS 



ANQUSTOVITTATA 



DIVERSA 



HYPOTHETICAL ANCESTOR 



Text-figure 4.— Scheme of phylogenetic development of the li?ieata group, 

 based upon the data of distribution and variation. 



lineata, form a series from north to south over the continent, which show 

 not only geographically arranged steps, but steps also in the variations 

 presented; from multitceniata, with variation entirely in the melanic half of 

 the possible series, through intermedia to decemlineata, with its variations in 

 the middle and albinic end of the possible variation series. Thus these three 

 species, and to the south, in like manner, oblongata, show differentiation in one 

 direction, both geographically and in their fluctuating variations. Melano- 

 thorax and rubicunda, which live in the same habitat with multitceniata, rep- 

 resent more local species than the rest, while juncta and dejecta show one and 

 the same sort of a series in their geographical distribution and their variations. 

 Is it possible logically to derive any other conclusion from the above data 

 than that the conditions of distribution and variation as found indicate the 

 direction and order of evolution and the relationship of the species? Dis- 

 carding impossible hypotheses, we are forced to one of two alternatives — 



