NEW MAI.I.OI'HAGA. 207 



anterior part of the tube they are entirely kicking, the 

 tiil>e being here transparent. Beginning at the pos- 

 terior end tlie tube runs straight forward to near tlie 

 anterior end of the transverse muscles. In the longi- 

 tudinal muscles it beconies folded and a loop may 

 project from the latter. It is here of greatest diameter. 

 It becomes narrowed in the part that has no internal 

 processes and gradually passes into the comparatively 

 narrow ejaculatory duct. To its posterior end the 

 chitin rod is attached. This is circular in transverse 

 sections, rather long since it reaches into the fourth 

 segment, and tapering anteriorly. It is curved, having 

 the convexity to the right and dorsally. 



The relationships of these different parts is very 

 clearlv shown bv transverse and longitudinal sections 

 (plate XV, figs. 2-5). The posterior end of the last (the 

 ninth) abdominal segment is deeply invaginated, tlie 

 invagination running forward into the fifth segment. 

 The walls of the tube thus formed (plate xv, figs. 2, 3, 4, 

 ti) are very thin and transparent, being entirely non- 

 chitinized. In the anterior part of the sixth segment, 

 however, they begin to increase in thickness and con- 

 tinue to become thicker from here forward to their 

 anterior ends in the tifth segment. Here they turn 

 awav from the axis of the tube they form and then 

 curve backwards a short distance. Then thev become 

 thin again and turn sharply forward and inward, 

 closely following the inner face of the first bend. This 

 is the beginning of an evagination which extends back- 

 wards almost to the beginning of the first. The walls 

 of this are likewise very thin and transparent, and are 

 closely applied, except posteriorly, to the inner surface 

 of the first tube. There is thus formed a double-walled 

 tube open in front and having a recurved anterior edge. 



