LEWIS. — POLYCH^TE ANNELIDS. 235 



seventy segments, the difficulty of obtaining entire worms has prevented 

 such a study of the remaining segments as would give reliable topograph- 

 ical results. In these twenty segments I find the following conditions. 

 Leydig's fibre appears in the sub-oesophageal ganglion as a pair of tubes 

 which, passing caudad, very soon unite to form a single fibre extending 

 backward along the mid-dorsal region of the nerve cord. It increases 

 somewhat in diameter as it runs caudad, but the increase is slight in 

 proportion to the size and number of the cell-processes which enter it. 

 In this worm there is to be found in the first twenty segments only one 

 Leydig's fibre. 



In Axiothea torquata the fibre of Leydig consists in the sub-oesopha- 

 geal ganglion of two tubes, precisely as in C. producta, as is shown by a 

 series of cross sections of the worm through this region. These two 

 tubes in passing caudad soon unite, thus producing a single fibre, which 

 extends backward along the mid-dorsal region of the nerve cord as in 

 C. producta. But this fibre continues single only as far as through the 

 seventh segment. Beyond the seventh segment there are two fibres lying 

 side by side almost in contact, and extending as far as the nineteenth seg- 

 ment. In the nineteenth segment the two again unite, and from this 

 point onward there is again only one Leydig's fibre ; this continues 

 undivided, and ends in the twenty-second segment. 



As long as Leydig's fibre remains a single structure, it shows the same 

 relation to the giant cells as in the case of Clymene producta ; but in the 

 region where there are two fibres side by side, I find that some of the 

 giant cells send their process into the fibre of the corresponding side of 

 the cord, whereas the process of others crosses the median plane to enter 

 the more I'emote fibre. 



The two Leydig's fibres of Axiothea torquata extend throughout most 

 of their course parallel to each other, lying side by side, but in certain 

 places they cross each other. The crossing may be a single isolated one, 

 or there may be two crossings in quick succession, so that an exchange of 

 positions is quickly followed by a return of both to their original relations 

 to the body. In these successive crossings, whether close together or far 

 apart, a fibre, first passing above (dorsad of) its mate, may then pass back 

 to the side on which it originally lay, either dorsad or ventrad of the other 

 fibre. The first condition results in a simple crossing, the second in a 

 real twisting of the fibres. In either crossing or twisting no anasto- 

 mosing or branching of the two fibres has been discovered. 



While the crossings of Leydig's fibres show in the diS^erent segments 

 considerable variation in plan and number, there seems always to be at 



