TORNATELLINIDJE. 67 



sp. d., duct of spermatheca ; vag., vagina. Fig. 1, opposite 

 side of the penis of same. Fig. 3, penis of Auriculella cerea 

 Pfr. Molokai. Fig. 6, genitalia of Auriculella westerlundi- 

 ana Auc. Glenwood, Hawaii. Figs. 4, 5, living animals of 

 Auriculella castanea, Mt. Tantalus. The markings on the 

 mantle show through the shell. 



The radula is like that of Achatinellidcc, having lost all cen- 

 tral and lateral teeth, marginal teeth alone remaining (cf. 

 Vol. XXII, pp. x, xi). 



A further relationship with the two families mentioned is 

 seen in the shell, which both by its general form, and by the 

 fold, sinuosity or oblique truncation of the columella, is a 

 good deal alike in the three families. The Tornatellinida? 

 differ from the others in having a parietal lamella, which no 

 Achatinellid or Amastrid snail has. In this feature the Tor- 

 natellinidce resemble most Pupillid<r, and are remarkably 

 similar to the genus Leptinaria in the Acliatinidcz. 



The foot, when the animal is in movement, is about as long 

 as the shell in Auriculella but decidedly shorter in Tornatellina. 

 It is quite narrow, and in progression shows muscular waves, 

 as in Helicida: and many other snails (pi. 22, fig. 4). The 

 sole often becomes areolate, when the animal is not moving, 

 as though the muscles were contracted in large, irregular, ver- 

 tical bundles. This has been noticed in both Auriculella and 

 Tornatellides macrom,phala. In Auriculella the eye peduncles 

 are bulbous at the ends, as usual in land snails, but in Torna- 

 tellides they are not enlarged distally. 



Snails of this family are always small. In most of them 

 the shell is three to five mm. long ; but some species of Auricu- 

 lella reach a length of 10 or 11 mm., being larger than some 

 Leptachatinas. The largest species of Tornatellina are slightly 

 longer than the smallest Auriculellas. 



The Tornatellinidtr might, from their wide distinction, be 

 thought ancestral to the very local Achatinellidiz and Amas- 

 tridcc ; yet it does not seem possible that the simple, general- 

 ized radula of Amastridce could have been derived from the 

 very-highly modified radula of Tomatellinidce. And so 

 far as the shell is concerned, the Achatinellida: are nearer to 



