3. The mouth parts are developed for sucking only, 



4. The attachment of the abdomen to the thorax in some 

 flies shows that they once possessed a pedunculated abdomen, 

 similar to that of Hymeuoptera (p. 251). 



Of these features, the first is the most weighty. Had not 

 its importance been overlooked, the order could never have 

 been thought inferior to the Lepidoptera, of which the mem- 

 bers have while larvte thoracic legs and usually abdominal 

 ones also. Among the Hymenoptera, the Tenthredinidse 

 have thoracic legs and even more numerous abdominal ones 

 than the Lepidoptera. The Uroceridae also have rudimen- 

 tary thoracic legs, although the larvae are borers in wood. 



The second and third arguments are essentially one in 

 principle. In the lower winged insects, we find both pairs 

 of wings of equal size and importance. The Hymenoptera 

 show a condition in which the hind wings are much smaller 

 and so of less use. Now, why do not the Diptera represent 

 the extreme of this series ? The question is not whether two 

 pairs of wings or one pair are in themselves " higher; " it is 

 rather. Which type shows the greater departure from the 

 forms universally acknowledged as ancestral ? So regarding 

 the mouth development: If the mandibular mouth of Thy- 

 sanura, Odonata, etc., be admitted as representing the ances- 

 tral form, then surely the mouth combining mandibular and 

 suctorial apparatus is intermediate, and that with only suc- 

 torial organs is the ultimate degree of specialization. The 

 recent researches of Dr. John B. Smith (Trans. Am. Ent. 

 Soc, XVII.) show that true mandibles are almost never 

 present in Diptera (he found them only in Simulium). Al- 

 though his conclusions in this respect, as well as in regard 

 to the homologies of the dipterous mouth in general, are 

 -widely different from those of earlier investigators, they are 

 probably correct. In summing up, he says (p. 339), "The 

 development required is simply a further development of the 

 line started in the Hymenoptera." 



An argument that strongly reinforces the first one above 

 is found in the fact that the embryo in Diptera, at least in 

 the higher forms, does not develop any traces of legs, diff'er- 

 ing in this respect from even the highest Hymenoptera, which 

 first develop the legs and then reabsorb them before hatching 

 {Psyche, June, 1891, p. 98). 



The subject of mimicry also throws some light on these 

 relations. As is well known, the Diptera afford many in- 

 teresting cases of mimicry, and it is important to our theory 

 to notice that they generally imitate the Hymenoptera, es- 

 pecially the very highest forms, such as wasps, humble-bees, 

 and even honey-bees. One of the most widespread of all 

 species, Eristalis tenax Linn., is such a good imitation of 



