WERNER MARCH AND 171 



THE EVOLUTION OF THE ABDOMINAL PATTERN 

 IN TABANIDAE (DIPTERA)' 



BY WERNER MARCHAND 



There are certain acknowledged and well-known principles by- 

 means of which it can be ascertained whether a given type of 

 pattern is primitive or derived, and by which, among a number 

 of given patterns, the most primitive of them may be determined. 

 A form in which the same elements of pattern are repeated on 

 all segments of the abdomen, and in the same relative position 

 on each of them, will be considered more primitive than a form 

 where this same pattern is found only on one or a few of the seg- 

 ments while the others ai'e of uniform color, or show a different pat- 

 tern. In the internal as well as in the external anatomy of the 

 segments we possess numerous instances of a specialized devel- 

 opment of single segments, a character which is always con- 

 sidered as derived or secondary. In general, though not in all 

 cases, we will then have to assume that characters which poten- 

 tially existed in all segments, have been preserved and further 

 specialized only in certain of them, while in the others every 

 trace of them has disappeared. Or it may be that the specializa- 

 tion of originally equal elements took place in divergent direc- 

 tions. In the case of abdominal patterns, fusion of spots may 

 take place in one segment, suppression of spots in others, etc. 

 The second characteristic of primitiveness is, consequently, the 

 homeotropic isotropism of pattern, as we may call it: the even 

 distribution of the elements of pattern in the segment and on the 

 whole abdomen, relatively to size, degree of pigmentation, condi- 

 tion, etc., of these elements. If the elements of pattern are evenly 

 distributed over the segment, or its dorsal and ventral surface, we 

 are confronted with a more primitive condition than in the case 

 of an uneven pattern. A pattern consisting of a definite number 

 of spots of the same size and in equal distance from one another, 

 is more primitive than a pattern consisting of large spots on the 

 median line and small or no spots on the sides. The absence of 

 any definite pattern, then, will always have to be considered 



1 From the Department of Animal Pathology of the Rockefeller Institute for 

 Medical Research, Princeton, N. J. 



TRANS. AM. ENT. SOC, XLIV. 



