90 



Y.I. KANTOR, A. MEDINSKAYA AND J.D. TAYLOR 



Although teeth with the wishbone form are found in the sub- 

 families Turrinae, Clavatulinae and Cochlespirinae, most variation 

 and the most extreme forms are found within taxa classified as 

 Crassispirinae. Although rather a disparate range of radular teeth 

 was found amongst the species we studied, all can probably be 

 derived from the basic wishbone form. Thus, the derivation of the 

 rather extreme forms of wishbone teeth seen in Hindsiclava and 

 Haedropleura can be envisaged by changes in the relative propor- 

 tion of the elements. In the large, paddle-shaped teeth of Funa and 

 Vexitomina, a thin, splint-like secondary limb lies along the shaft of 

 the main tooth. In the most-derived, harpoon-like teeth of 

 Cheungbeia, a similar small splint-like secondary limb lies along the 

 edge of the shaft. Finally, the hollow awl-shaped teeth of Ptychobela 

 consist of two components and can probably be derived from 

 flattening and fusing of the two wishbone components. 



Congruence of shell characters 



Shell characters are also a rather poorly correlated with radular 

 morphology or characters of foregut anatomy. A good example of 

 this problem is seen in the genera Inquisitor, Funa, and Ptychobela. 

 These have rather similar shells but the radulae are quite different, 

 Inquisitor having the standard crassispirine wishbone form, Funa 

 has bladed, paddle-shaped teeth and Ptychobela possesses awl- 

 shaped hollow teeth. The allocation of shells to these genera is 

 difficult without radular evidence and many species have been rather 

 arbitrarily assigned to genera (e.g. Wells, 1994). 



A striking example of the similarity of shell characters in different 

 families is seen between the pairs of West American species 

 Crassispira (Striospira) tepocana (see Keen, 1971 fig. 1701) and 

 Strictispira ericana (Keen, 1971 fig. 1727) and C. (Striospira) xanti 

 (Keen, 1971 fig. 1702) with Strictispira stillmani (Keen, 1971 fig. 

 1728). These pairs of similar gastropods have quite different anato- 

 mies. The Strictispira species (family Strictispiridae) lack the venom 

 gland, have the buccal mass located at the proboscis tip and possess 

 very distinctive radula teeth (Kantor & Taylor, 1994). 



Parallel evolution of hypodermic feeding 

 mechanism 



In two genera, the radular teeth are quite different from the rest of the 

 Crassispirinae, which for the most part consist of variations on the 

 robust wishbone form. In Cheungbeia species, the teeth are long, 

 slender and harpoon-like, with distal barbs, whilst in Ptychobela the 

 teeth are pointed, awl shaped and hollow. Although most conoidean 

 teeth, including the wishbone types can be used at the proboscis tip 

 for the penetration of prey, the hollow, barbed teeth are regarded as 

 the more efficient. Taylor et al. (1993) demonstrated that hollow, 

 barbed teeth had evolved several times within the Conoidea and 

 these two taxa may represent further evolutionary pathways to the 

 hypodermic feeding mechanism. The teeth in Cheungbeia and 

 Ptychobela although quite different in morphology can both be 

 derived from the crassispiran wishbone form. Although Cheungbeia 

 is long and gutter-shaped, a small and very thin accessory limb lies 

 along one side of the shaft (Fig. 26). A similar reduction in the size 

 of the accessory limb is seen in Funa and Vexitomina which have 

 paddle-shaped teeth, with the thin secondary limb lying along the 

 shaft (Figs 23 b-d). In Ptychobela, each tooth is formed from two 

 more or less equal parts which are fused along one edge and loosely 

 enrolled. Neither of these two tooth types can be considered as 

 precursors of the enrolled barbed teeth found in Conus (Coninae, 

 sensu Taylor et al. 1993) because their foregut anatomy is different. 

 Both Cheungbeia and Ptychobela, for example, have the modified 

 epithelium of the rhynchodeum, the elongated oesophagus, and the 



modified venom gland. None of these features is found in Conus. 

 Moreover, Conidae have lost the radular membrane and possess a 

 radular caecum for the storage of teeth prior to use. 



Comparison between conoidean subfamilies 



Three other subfamilies within the Turridae, the Turrinae, 

 Clavatulinae and Cochlespirinae possess the wishbone type of radular 

 teeth and their features should be compared with those of the 

 Crassispirinae. 



Members of the Turrinae have wishbone teeth which differ in 

 morphology from those of the other taxa. The proximal half of each 

 tooth is divided like a clothes-peg into two more or less equal units 

 (Fig. 18), with the limbs not detached. This contrasts with the 

 wishbone teeth in the other subfamilies which have the limbs 

 unequal in size and thickness and the secondary limb not in struc- 

 tural continuity with the major limb. Some species of Turrinae also 

 have a quadrate central tooth with a spine-like central cusp. Species 

 of Turrinae also differ in a number of anatomical characters; there is 

 usually no elongation of the oesophagus, no change in the histology 

 of the venom gland anterior of the nerve ring and the rhynchodeum 

 is uniformly glandular along its length (Taylor et al., 1993; Taylor, 

 1994). 



The Clavatulinae have wishbone teeth with a large, bladed, major 

 limb and the secondary limb inserted into a 'V shaped groove. 

 Additionally, central teeth are usually present. They also have a 

 medio-lateral nucleus to the operculum rather than the terminal 

 position found in the other turrid subfamilies. The buccal mass often 

 lies within the proboscis, but as in Crassispirinae, the oesophagus is 

 often elongated between the nerve ring and buccal mass and the 

 venom gland changes to a ciliated duct anterior to the nerve ring 

 (Kantor, 1990; Taylor et al., 1993). 



The Cochlespirinae, represented by Aforia, Antiplanes, and 

 Cochlespira, have similar wishbone teeth to the Crassispirinae, with 

 some species possessing a central tooth, and some having plate-like 

 lateral teeth (Sysoev & Kantor, 1987, 1988; Kantor & Sysoev, 

 1 99 1 ). In Cochlespira and Aforia, the venom gland joins the oesopha- 

 gus some way to the posterior of the buccal mass, but no details are 

 available for other taxa. 



We have found a great variation in the structure of the foregut in 

 the Crassispirinae and some of these features are shared with the 

 Turrinae, Clavatulinae and Cochlespirinae. However, rather few 

 species have been studied from the latter three groups compared 

 with the more extensive survey of the Crassispirinae. A reappraisal 

 of the boundaries and relationships of these four subfamilies is 

 required, but this cannot be attempted before more anatomical 

 details are available from a much wider range of taxa. 



Systematic conclusions 



ANTIGURALEUS and associated genera 



On the basis of his study of southern African species, Kilburn ( 1994) 

 suggested that two genera, Anacithara and Antiguraleus, which had 

 previously been referred to the subfamily Mangeliinae (as in Powell, 

 1966), should be transferred to the Crassispirinae on the basis of 

 radular characters. We studied one of the South African species, 

 Antiguraleus morgan! and confirm that it has a crassispirine type of 

 radula (Fig. 28a) and also has many foregut characters consistent 

 with other members of the Crassispirinae. Additionally, we also 

 sectioned the eastern Australian species, Guraleus costatus, and this 

 also has an operculum, with a radula and foregut anatomy of the 



