Veneridae is the most diverse Recent bivalve family, comprising over 800 extant, presumably valid, species in approximately 170 genera. William Healy Dall (1902: 336) called venerids ‘the culmination of pelecypod evolution’ in terms of their morphology, distribution, and bathymetric range. Its members arguably include the most familiar of all bivalves, such as the hardshell clam Mercenaria (‘quahogs’, ‘cherrystones’), Pismo clams, littlenecks, butterclams, and Manila clams, which form key components of the world’s clam fisheries. They are circumglobally distributed in temperate to tropical waters, and are adapted to a wide range of environments (Kondo, 1998).

Although their numbers and economic relevance have focused attention on certain species, this has not translated into broader systematic studies on Veneridae, nor have the many published single-species studies been placed into phylogenetic context. Venerid classification has been historically unstable in terms of taxon placement and higher-order arrangement. Numerous family-group taxa have been introduced and used in various classification schemes over time (Appendix 1). In his classic systematic compendium, Thiele (1934) declined using venerid subfamilies (then based on hinge teeth), considering them unnatural. The 12 nominal subfamilies in the Treatise on invertebrate paleontology (Keen, 1969: N670) were used ‘for convenience of arrangement’ without ‘necessarily reflect[ing] genetic relationships’. Nevertheless, this classification has since become widely accepted. Abbott (1974: 521) summarized that ‘classification of the family... has been one of continual debate and rearranging for some years’, a controversy that originated in the original broad concept of the genus Venus (Dodge, 1952). Some alternative subfamilial arrangements have been adopted by a few more recent authors (e.g. Hikida, 1996; Shimamoto, 1996), although without gaining widespread acceptance.

Modern phylogenetic studies on other bivalve groups have shown that morphological traits frequently do not support widely accepted classifications (e.g. Graf, 2000; Lydeard, Minton & Williams, 2000) and are probably influenced by evolutionary convergence (Canapa et al., 1996). Wagner (2000: 365) postulated that morphological character states are subject to exhaustion – ‘when character [state] change is more likely to yield homoplasy than novelt[y]’ – in large deeply rooted families. These tenets also seem to apply to Veneridae, which dates back to the Cretaceous (Skelton & Benton, 1993); there are no recognized synapomorphies for Veneridae or any of its recognized subfamilies. Veneridae is usually distinguished by a single rather generalized hinge character – the presence of three cardinal teeth in each valve – with all other shell characters (of lateral teeth, pallial sinus, lunule, escutcheon, sculpture) varying greatly (Keen, 1969; see below). A good example of the level of morphological variation in venerids relative to the present classification is Harte’s (1998b) informal organization of the subfamilies into two groups:

Weakly ornamented (Clementiinae, Dosiniinae, Meretricinae, Pitarinae, Sunettinae): weak surface ornamentation, smooth margins, well-developed pallial sinuses, well-developed anterior lateral teeth.

Ornamented (Chioninae, Gemminae, Samarangiinae, Venerinae): strong surface ornamentation, crenulate margins, small or absent pallial sinuses, weak or absent anterior lateral teeth.

Numerous exceptions to this dichotomy are evident, including members of Gemminae with smooth shells, of Sunettinae with crenulate margins, and of Dosiniinae and Clementiinae with weak or absent lateral teeth. Three nominal subfamilies (Cyclininae, Gouldiinae, Tapetinae) exhibit too strong a mixture of features to have been categorized by Harte (1998b) in this scheme. Perhaps most importantly, many of the listed features have been considered as ecophenotypic adaptations against predators: well-developed pallial sinus (= long siphons) for deep infaunal burrowing, strong surface ornamentation for anchorage, and marginal crenulations for tighter closure. Morphological convergence is one potential reason for the difficulty of resolving venerid relationships; paedomorphosis is another. F. R. Bernard (1982) used the degree of reduction of the outer demibranch, the relative length and separation of siphons, and the presence/absence of an adult byssus to develop an evolutionary scenario that grouped small, paedomorphic, brooding venerids together. The resulting taxonomy from this study has been criticized (Lindberg, 1990) as lacking a phylogenetic framework.

Despite the apparent lack of morphological synapomorphies, higher-order molecular analyses of the Bivalvia using portions of slowly evolving nuclear [18S rRNA, 28S rRNA, histone 3 (H3)] and faster evolving mitochondrial [cytochrome oxidase I (COI)] genes (summarized in the following sections) have supported the monophyly of Veneridae, although none so far has included more than four venerid taxa. Within-group relationships of Veneridae are likewise unresolved. Most molecular studies have focused on single species or genera (e.g. Dillon & Manzi, 1989a, b; Passamonti, Mantovani & Scali, 1997, 1999) rather than relationships within and among subfamilies, and those with wider foci have had limited taxon sampling. Even so, some studies have suggested potential sister relationships between certain subfamilies (see the following sections for summaries of individual studies).