ANATOMY OF THE GOUT-FLY OF BARLEY. 805 



liypodei-mis at a later stage, but it has only been actually 

 observed in the (lorso-Literal regions. The abdominal imaginal 

 discs, of course, never become invaginated, and therefore have 

 no imaginal sac and no peduncle. 



llie Dorsal Vessel, 



The dorsal vessel is a tubular orgnn lying clo-se beneath the 

 dorsM,l hypoderm in the median line, its posterior end lying at 

 about the junction of segments 10 and 11. From here it runs 

 forwards as fn.r ns the anterior region of segment G, where it dips 

 ventrally to the brain to take the course already described in 

 connection with the central nervous system, etc.*^ The part of 

 tlie tube lying in segment 10 is relatively wide, but it narrows 

 considerably as it passes forwards into segment 9. The wide 

 posterior region may be defined as the heart ; the portion 

 anterior to segment 6 may be termed the aorta. The ostia are 

 confined to the cardiac region, but I have been unable to make 

 certain as to their number and position. The posterior end of 

 the heart is slightly anterior to the posterior transverse com- 

 missure. ThH latter is suspended from the dorsal integument by 

 a delicate double transverse membrane, the space between the 

 two layers being Avide dorsally but narrowing ventrally where 

 the membrane fuses with the commissure. From the middle of 

 the anterior membrane a delicate ligament passes forwards to 

 become attached to the ventral wall of tlie posterior end of the 

 dorsal vessel. At about the level of each pair of pericardial cells 

 (see below) in segments 10-6 inclusive the dorsal vessel is 

 attaclied to the dorsal hj^poderm by a delicate ligamentous 

 strand. 



Pericardial Nejjhrocytes, and Alar Muscles. 



On each side of the dorsal vessel there are 26 pericardial cells 

 (nephrocytes) situated as follows : — 



In each of segments 10, 9, and 8, four cells on each side. 



In segnionta 7 and 0, five cells on en,ch side. 



In segment 5, four cells on each side. 



The cells lie in pairs, one on each fide of the vessel, exactly 

 Opposite each other. Seen in surface view in the livino- larva 

 the cells are perfectly circular ; in transverse or longitudinal 

 section tliey are oval. They are all uninuclear. In the majority 

 of cyclorrhaphous larva) the pericardial nephrocytes are of two 

 kinds ; posteriorly there are a few large cells somewhat separated 

 from each other, while anteriorly the cells, though similar to the 

 posterior ones in structtne, are smallei^, more numerous, and 

 closely crowded (see Keilin, 1917, for a review of the subject). 

 In Chlorops, however, there is no such distinction, the cells beinir 

 all alike. ' ^ 



In connection with each pericardial cell (with the exception of 

 the four a.nterior pairs which lie alongside the ventrally-directed 

 portion of the vessel) there is an alary muscle. The alary muscles 



Proc. Zool. Soc— 1923, No. LIII. 53 



