J 90 Transactions. 



Having discovered this mistake I proceeded to examine each of the 

 other characters more carefully. 



B. As to the chaetae, I find from a study of eight individuals that there 

 is a considerable range of variation, as may be seen by a study of the annexed 

 table, in which I have summarized the number of chaetae in the dorsal 

 and ventral bundles in the pre-clitellar and in the post-clitellar region of 

 the body of specimens from the Macquaries and from the Campbell and 

 Auckland Islands. It will be noted that the difference between extremes 

 such as No. 2 and No. 5 amongst specimens from the Macquaries is greater 

 than the difference between No. 2 and No. 8 from two distant islands, and 

 it is impossible to include in the diagnosis of a species a character with such 

 a wide margin of variation. 



C. It will be noted, too, that the segment in which the dorsal vessel 

 originates shows a similar variation. It is true that in the type it com- 

 mences at the hinder end of the 13th or 14th segment, while in the type of 



inter wed ins," as I can confirm from renewed examination, this point is 

 in segment 17 ; but even amongst those from the Macquaries the position 

 varies, being in two cases in the 15th, in a third in the 16th, while in one 

 that was sectionized it lies in the 17th segment. 



D. The number of the subneural glands, which are characteristic of 

 this genus, exhibits the same instability, for though usually there are 3 

 glands in segments 14, 15, and 16, there is one individual from the Mac- 

 quaries in which there are 6 glands, and in two " intermedins " there 

 are 4. 



E. Finally, I made a point of the proportion of length to breadth of 

 the funnel of the sperm-duct, for in the type of L. macquariensis I stated 

 that the length is twice the breadth, whereas in " L. intermedins " I gave 

 it as about five times the breadth. I have measured it in three funnels of 



intermedins " whose outlines I drew with the camera, two in a series of 

 longitudinal sections, and one in a bisected specimen mounted as a trans- 

 parent object. From these measurements I find that the length is 

 respectively five, five and a half, and six times the breadth. 



I am unable to give measurements for the funnel of macquariensis, as 

 it is bent in all the preparations, but the proportions given in the original 

 statement seem to be borne out. But the state of preservation of the type 

 is bad ; the worm was soft, and it is possible that the gland-cells around 

 the funnel are much swollen, just as those of the subneural glands are. In 

 my figure of the latter (1, pi. xiv, fig. 8) they are represented as much too 

 broad and too high. Without at that time having well-hardened specimens 

 for study, I did not recognize the effect of this bad preservation on the 

 gland-cells ; but a comparison of the sections with well-preserved material 

 shows at once the fact that the gland-cells are swollen, so that the whole 

 gland appears larger than it would be in life. Hence again the difference 

 between the figure of macquariensis referred to and that given for 

 ik intermedins ; " (pi. x, fig. 9). 



So I think we may take it that in the case of the funnel gland-cells the 

 same explanation may be given ; their swollen condition increases the 

 width of the funnel, and led me to give proportions which are no doubt 

 untrue in life. It is not improbable, however, that the size of the gland- 

 cells in both glands may vary according to the sexual condition of the worm, 

 and it is likely that when fully mature in the breeding season the glands 

 would be larger. I conclude, then, as a result of this comparison, that 



