FROM THE SOCIETY S GARDENS. 



687 



subterminal on the dorsal surface. Fi-om between the parameres 

 a.nd continuous at that point with the basal plate is a small 

 endonieral plate continued backwards as a long, nai-row, thick, 

 parallel-sided, chitinons strip whicli supports the sac on its lower 

 side along- the middle line. The ductus can be seen through the 

 wall of the sac running ba,ckvvards into the abdomen, where it lies 

 on the doi'saJ side of the bahal plate. 



Text-ti"ure 31, 



Lipeurus furficulatiis. $. 'J'raiis\erse section throiifrli the aluloiiien sbovvins the 

 copulatoiy sac lying retracted within the body-cavitj' aljove the basal plate. 

 (Diagrauimatic.) 



M. rectum. T. trachea. 8. sac. FL. narrow (diitinous strip on ventral surface oi 

 sac. M. uiuicles. ]?C. fat-cells. SF. basal plate, showing the rift. 



The A-pparatus lohen retracted. — Dining I'etraction the sac is 

 continuously invagina,ted until the distal end with its denticles 

 comes to lie farthest fprward within the abdominal cavity a little 

 anterior to the fore end of the basal plate. The thick endomeral 

 strip on its lower wall, of course, curls upwards and travels in 

 with the rest, so as to form in the retracted state an enigmatic 

 loop difficult to interpret until an extruded sac is examined. In 

 cross-section, therefore, the endomeral strip forms the lower wall 

 of the inner tube (see text-fig. 32). A similar endomeral loop 

 w^ith a similar history was described in Part I., p. 271, for Tricho- 

 dectes latus. The diagram should make the i-elation of the parts 

 quite clear. 



It should be clearly understood that the parameres are at no 

 place rods or appendages discrete from the sac. At their distal 

 end (text-fig. 33) a membrane crosses between them dorsally and 

 another membrane crosses ventrally. If the parameres became 

 shorter these two membranes would become continuous with 

 one another and with the dorsal wall of the outer tube, and if, 

 finally, they disappeared we should have a simple exsertile tube. 

 Sections anywhere across the length of the parameres all show 



