676 SUBANTARCTIC ISLANDS OF NEW ZEALAND. [Land Nemerteans. 



NOTE ON THE NUMBER OF PROBOSCIDEAL NERVES AND OF RESERVE STYLET- 

 POUCHES IN G. AUSTRALIENSIS AND G. NOVAE-ZEALANDIAE. 



The observations which I am about to place on record were made on a collection 

 of land nemerteans made in Australia and New Zealand by Professor Dendy, and 

 kindly placed at my disposal by him. 



Many of the descriptions of new species of animals and plants are made, as in 

 the case of the two new species described in the present paper, on solitary specimens 

 of the new forms. If this description is not soon superseded by one based on a 

 larger number of specimens, the impression that the characters of the species are 

 not subject to variation takes a hold which becomes firmer with each repetition 

 of the original description. 



One of the characters used in the classification of the species of the genus Geone- 

 mertes is the number of the proboscideal nerves. Now, a number of this kind when 

 used for classificatory purposes is valueless unless the range of variability, if any, 

 to which this number is subject is known. Let us fix out attention on a particular 

 instance, G. australiensis : this was first described in detail by Dendy,* who does 

 not actually state the number of proboscideal nerves, but publishes a drawing 

 of a transverse section of the proboscis from which this number — 18 — can easily be 

 made out. Coe, in his monograph on G. agricola,'\ has drawn up a synoptic table 

 of the species of Geonemertes, and states that the number of proboscideal nerves 

 in G. australiensis is 18. Punnett, in his account of G. arboricola from the Sey- 

 chellesj, publishes an amplification of Coe's synoptic table, in which G. australiensis 

 again appears with 18 proboscideal nerves. 



All that I am concerned with in pointing out here is that the impression that 

 the number was invariable, which the reader who did not read the original de- 

 scription would be likely to gather, would be simply a result of the fact that the 

 first description of this character was based on observations made on a single 

 specimen. My observations on Professor Dendy' s specimens prove, however, that 

 this number is by no means constant. There proved to be two specimens with 

 16 nerves, two with 17, seven with 18, and one with 19. So that the specimen 

 whose proboscis Dendy figured happens to have represented the commonest type. 



The state of affairs in G. novae-zelandiae, the number of whose proboscideal 

 nerves has not yet been recorded, is very curious. 



The two actual specimens on which I have been able to count the number of 

 proboscideal nerves are already familiar to those who are conversant with the litera- 

 ture relating to land nemerteans. The species was first named by Dendy,§ from 

 two spirit specimens — one found in a jar of " land planarians " which had been 

 collected in various unspecified localities in New Zealand ; the other (a much more 

 perfect specimen) found in a bottle labelled " Land Planarians, Toitoi, Southland ; 

 Miss T. G. Eich." Dendy soon afterwards found living specimens of this species, 

 and gave a brief account of it.'] The two specimens I have examined are the spirit 



* Proc. Roy. Soc. Victoria, vol. iv, p. 85 (1892). 

 t Proc. Boston Soc. Nat. Hist., vol. xxxi (1904). 



I Trans. Linn. Soc. London, 2nd ser., Zoology, vol. xii, p. 1 (1907). 

 § Trans. N.Z. Inst., vol. xxvii, p. 191. 



II Trans. N.Z. Inst,, vol. xxviii, p. 214. 



