50 NORTH AMERICAN DII'TERA. 



cles of the thorax are provided with 'vocal cords', and a 

 considerable part of the humming sound is produced by 

 those structures. The abdominal spiracles of some flies 

 are as primitive as are to be found among insects, being 

 simply unlipped openings. 



The condition of the nervous system varies greatly 

 within the order. In the elongate, more generalized ne- 

 matocerous forms there are five or six abdominal ganglia, 

 and three distinct thoracic ganglia. From this condition 

 to that shown by the Muscidae, where all the thoracic 

 and abdominal ganglia are united into one large gang- 

 lion in the thorax, a most instructive series of gradatoi v 

 forms is presented. In the Empididae, which stand in- 

 termediately as regards the concentration of the ventral 

 cord, the two anterior thoracic ganglia are fused into 

 one ; this condition is radically different from that shown 

 by insects of other orders, as the Ccleoptera, Lepidcp- 

 tera and Hymenoptera, which have but two thoracic 

 ganglia. In these insects, however, it is the two poste- 

 rior ganglia (meso- and metathoracic) which are fused 

 into one. 



The studies of Child on the so-called Johnston's organ 

 (located in the second anteunal joint), an elaborate struc- 

 ture of fine chitin rods connected with special nerve cells 

 and these connected by fine nerves with the main anten- 

 ual nerve, seem to ascribe definitely an auditory function 

 to the antennae. Kellogg has shown that the divided 

 eyes of such flies as the Blepharoceridae and others ( in- 

 dicated externally by the division of the corneal facets 

 into two regions in which the facets are of different size ) 

 are in reality made up two types of ommatidia, one type 

 being much larger and much less strongly supplied with 

 retinal pigment than the other type. This condition pro- 

 duces a certain sort of differentiation of the visual func- 

 tion, one part of the eye being better adapted for seeing 



