Hemidonax, AND THE VENEROIDEA

Healy, John M., Mikkelsen, Paula M. & Bieler, Rüdiger, 2008, Sperm ultrastructure in Hemidonax pictus (Hemidonacidae, Bivalvia, Mollusca): comparison with other heterodonts, especially Cardiidae, Donacidae and Crassatelloidea, Zoological Journal of the Linnean Society 153 (2), pp. 325-347 : 339-343

HEMIDONAX AND THE VENEROIDEA

Several non-cardioid members of the Veneroida have been examined for sperm ultrastructure (but still only a fraction of the known genera) especially among the Veneridae ( Pochon-Masson & Gharagozlou, 1970; Gharagozlou-Van Ginneken & Pochon-Masson, 1971; Nicotra & Zappata, 1991; Reunov & Hodgson, 1994; Sousa et al., 1998; Healy et al., 2006). As part of a ongoing survey of sperm ultrastructure within the Veneroidea (and especially Veneridae ), we have also examined many genera (J. M. Healy, P. M. Mikkelsen & R. Bieler, unpubl. data) and include results for a few previously unstudied examples in this account for comparison with Hemidonax . Lioconcha annettae and Antigona chemnitzii (Hanley, 1844) show similar acrosomal dimensions to Hemidonax pictus , but in both, the acrosomal complex rests in an anterior depression of the nucleus and is angularly tilted, and in neither of these venerids is the acrosomal vesicle invagination narrow anteriorly (see Figs 8, 9). In addition, there are nuclear differences (narrow and curved in both) and at least in L. annettae five rather than four mitochondria. As mentioned previously in this discussion, L. annettae exhibits an electronlucent layer at the apex of the acrosomal vesicle, ·

similar to that observed in H. pictus . While this may prove significant, the lability of this region of the acrosomal complex (i.e. acrosome reaction stages) may be a factor and we reserve our judgement on the nature of this layer. In Dosinia nedigna (Iredale, 1930) the nucleus and midpiece are essentially as observed in H. pictus , but the acrosomal vesicle is tilted and exhibits a wide invagination (see Figs 7, 8). In most other investigated venerids, acrosomal, nuclear or midpiece features (usually a combination of these) do not closely tally with results obtained for H. pictus (see Pochon-Masson & Gharagozlou, 1970; Gharagozlou-Van Ginneken & Pochon-Masson, 1971; Reunov & Hodgson, 1994; Sousa et al., 1998). Only in Callista chione (Linnaeus, 1758) does the acrosomal vesicle approach that of H. pictus in shape, size and narrowness of the invagination, but even in this species, an axial rod is present within the subacrosomal material, the nucleus is slightly elongate and strongly curved, and the midpiece exhibits five mitochondria (see Nicotra & Zappata, 1991). However, despite various sperm differences between investigated Veneridae and Hemidonax , it remains impossible, at least at this stage, to rule out unequivocally a relationship between these two taxa, especially in view of the large number of venerid genera awaiting sperm study.

TAXONOMIC AND PHYLOGENETIC CONSIDERATIONS: THE AFFINITIES OF HEMIDONAX

‘After examination and comparison of the shell and of the anatomy (both external and internal) of Hemidonax to both cardiids and donacids, I cannot justify placing Hemidonax as a member of the Cardioidea. However, neither can I place Hemidonax within the Donacidae . Instead, I favor placing Hemidonax as incertae sedis within the order Veneroida , until a phylogenetic analysis of the Veneroida is undertaken’ ( Schneider, 1992: 145). With those words, Jay Schneider – a recognized authority on cardioid systematics and phylogeny (see also Schneider, 1995, 1998a, b) – effectively reopened the debate concerning the affinities of Hemidonax . His rejection of a relationship between Hemidonax and the Donacidae (based on his own anatomical observations) was in accordance with the views of Boss (1971) and Ponder et al. (1981), and it can be said with confidence that all available sperm ultrastructural data ( Hodgson et al., 1990; Pal, 1992, 1996; Sousa & Oliveira, 1994; Healy, 1995a; present study) likewise argue strongly against any connection between these two taxa. The same conclusion was reached by Schneider & Carter (2001) based on comparative shell microstructure. Schneider’s other conclusion – that Hemidonax is non-cardioidean – is all the more remarkable when it is considered that he was not persuaded by the arguments of either Boss (1971) or Ponder et al. (1981), who strongly supported cardioid affinities for Hemidonax . Certainly, in terms of its anatomy, Hemidonax shows a number of features not consistent with placement in the Cardiidae as outlined in some detail by Ponder et al. (1981; who argued for retention of a separate family Hemidonacidae ). Perhaps significantly, neither Boss (1971) nor Ponder et al. (1981) identify any affiliations between Hemidonax and any cardioidean genus, lending some degree of credence to Schneider’s (1992) decision to leave Hemidonax as incertae sedis among the Veneroida . It is of interest to note that Keen (1980), while accepting Boss’s (1971) subfamily Hemidonacinae within the Cardiidae , did not choose to discuss the relationships of Hemidonax [being evidently influenced by Wilson & Stevenson’s (1977) decision not to include the genus in their review of the Western Australian Cardiidae ]. Keen, however, grouped the Fraginae and Hemidonacinae in her tabulation of cardiid shell features and also in her taxonomic keys, and by so doing perhaps was hinting at the possibility of some relationship between the two subfamilies. Most recently, Schneider & Carter (2001) have argued, largely on the basis of comparative shell microstructure, that Hemidonax shows much closer affinity with the tellinoidean family Psammobiidae than with the Cardiidae (or any other ·

group of heterodonts). Whereas it is true that shell microstucture has proven a very valuable source of characters for phylogenetic analysis, Schneider & Carter (2001) have not offered any explanation for the key anatomical differences between Hemidonax and the Tellinoidea, particularly the absence of the cruciform muscle (its presence is a synapomorphy of the Tellinoidea – see also Boss, 1971, 1982). In this connection it is worth noting that Pharus and the Novaculininae (both originally included in Solecurtidae ) were often cited as tellinoidean taxa lacking a cruciform muscle ( Yonge, 1949, 1959; Ponder et al., 1981; Boss, 1982) but are now both placed within the Solenoidea (e.g. Morton, 1984; von Cosel, 1990; Willan, 1998).