WERNER MARCHAND 173 



segments one to four, while on segments five to eight they have been 

 swallowed by the dark background. Phenomena of this kind are 

 so frequently observed that they do not detract anything from 

 the soundness of the assumption that the five-spotted segment 

 represents the primitive condition. It seems that processes of 

 reduction of pattern always begin at both ends of the abdomen, 

 but progress more rapidly from the posterior end while the middle 

 region is more conservative. On the other hand it is likely that 

 sometimes in a fresh specimen the presence of vestigial spots may 

 be ascertained on more segments than in dry specimens or from 

 illustrations. 



The type of pattern of figure 1 is met with quite frequently in 

 the genus Haematopota, and all the varieties of forms of pattern 

 in Tabanidae can apparently be shown to be onlj' modifications 

 of this Haematopota-type. As stated, there are certain species of 

 the genus Tabanus in which this pattern occurs in identical form. 

 These are all species of small, at most medium size, among them 

 the American species T. pumilus, spams, vivax, etc.; it seems 

 that this group comprises the smallest of all known Tabanidae. 

 T. sufis (fig. 2), an African species, is of the same very small size 

 and appearance as H. hrunnescens (smaller than a house-fly!), and 

 shows in its venation a character which is not usually encountered 

 in Tabanus, but forms the rule in Haematopota. The abdominal 

 pattern likewise recalls Haematopota. Here we observe a tendency 

 of the median and subdorsal light spots to become enlarged, the 

 subdorsal ones placed somewhat ol)liquely, and the median ones 

 assuming a triangular shape with the bases of the triangles rest- 

 ing on the posterior border of the segments, while the lateral 

 spots seem to have shifted to the upper corner of the segments. 

 A series of further modifications frequent in species of Tabanus is 

 already indicated here in its initial stages. 



T. mordax (fig. 3) affords an example where the lateral spots 

 have wholly disappeared; it is this condition from which the 

 three-banded patterns can- be derived, inasmuch as the spots of 

 neighboring segments tend to enter a relation with each other in 

 order to form bands over the whole length of the abdomen. In 

 Tabanus kingi (fig. 4) this tendency is already mark(Ml; at the 

 same time in segments five to seven a further reduction has taken 



TRANS. AM. ENT. SOC, XLIV. 



