22 N. ODHNER, NORTHERN AND ARCTIC INVERTEBRATES. VI. PROSOBRANCHIA. 2 SEMIPROBOSCIDIFERA. 



conditions (Ice Fiord 1908, St. 41; off Horn Sonnd etc; cfr the list of localities). Tlie 

 deep sea form of Spitzbergen resembles the soutliern var. affinis in its small size and uni- 

 form white colour, though it ne ver attains the minimum size of the last. The shell is 

 often thin and transparent, and in that particular it reminds one forcibly of Natica 

 bathybii, but it has a uniform white or darker operculum witli a verv faint darkerbrown 

 tint in the nucleus and always a closed umbilicus; a dead shell already partially decom- 

 posed, from King Cliarles Land, exhibits an open umbihcal fissure, it is true, but in living 

 forms this ne ver occurs. Though it is probable that N. bathybii has originated from such 

 deep forms of N. clausa, adapted to the conditions in great depths A^ ith accompanying 

 feebler calcification, it must be regarded as a good species, because no complete series of 

 intermediary forms seems to occur. 



The above table also exhibits the remarkable fact that N. clausa in the southern 

 parts of Greenland (Sukkertoppen to Julianehaab) is smaller-sized and more like the 

 Iceland form than it is in Godhavn (northern West Greenland). 



As a summary of the above accounts it may be stated that N. clausa attains its 

 greatest size in districts of high-arctic nature, not reached by the Gulf Stream; in warmer 

 areas it becomes smaller or gives rise (owing maybe to different causes) to dwarf forms. 



Vari.atioH of tlie Rjuliila. 



There is but little variation of the radula and this is chiefly limited to the dimen- 

 sions of the teeth: the median tooth is dilated in the breadth (Fig. 7) or more produced 

 in the longitudinal direction. Tlie wing-like processes of the base are narrower or broad- 

 er. In young or small specimens the side denticles of the median tooth are strongly 

 marked, but in larger individuals they become worn, till they wholly disappear; such 

 may be the case also with the denticles of the laterals. The large specimens from the 

 Behring Sea consequently have all the teeth wholly smooth, while in small specimens 

 (h. 7.9) from the same districts they are as evidently denticulated as in forms from the 

 Swedish coast. 



In Ice Fiord the radula seems to be broader in shallow water forms (with a wider 

 aperture), narrower in deeper water. For the specimens figured on Plate 3, Figs. 7 — 9 

 the following measurements (in millimeters) may be noticed: 



Locality h. of sliell ap. h. br. ap. lir. umb. sut. vvli. br. of radula 



St. 127 (Fig. 7) 12.5 11 12 7.4 4.8 3 5 0.50 (Pl. 5 Fig. 10). 



St. 92 (Fig. 8) 11.2 8.5 10.8 6.3 2.7 4.8 5+ 0.38 (PJ. 5 Fig. 11). 



St. 41 (Fig. 9) 11 8 10.2 C 3.5 4 <5 0.34 (Pl. 5 Fig. 12). 



In other cases the breadth of the radula is usually in direct proportion to the 

 dimensions of the shell; where these have been subject to a reduction, the radula has 

 simultaneously become narrower; thus the breadth of radula in specimens of the same 

 size may indicate, just as does the number of whorls, that a form is dwarf ed, a fact that 

 is illustrated by the following table: 



