There is currently a debate concerning the definition of the taxon Tetrapoda, which is part of a wider debate dealing with the definiton of phylogenetic groups. Many papers over recent years (eg. Gauthier et al., 1989, Rowe, 1988, Rowe et al., 1992, de Queiroz et al., 1992) have suggested that a phylogenetic, or node-based, definition, is more rigorous than one which employs a "key-character", with both a "crown-group" definition (Altangerel et al., 1993, Norell and Novacek, 1992, Norell et al., 1993, Lebedev and Coates 1995) and a "total-group" definition (Patterson, 1993, Coates, 1996) having its adherents. In the case of the Tetrapoda, the crown-group would encompass the clade containing all the modern representatives, namely the Amphibia and the Amniota, plus any fossil taxa falling above the node in the cladogram representing their nearest common ancestor. In that case, Tetrapoda would exclude any fossil taxon which falls outside this grouping, independent of the possession of any key character such as limbs with digits. It has been stressed elsewhere (Clack and Coates, 1995) that as a key character, "limbs with digits" is unsatisfactory since it can be broken down into many separate characters. The "total-group" definition would include not only all members of the crown group but also its complete stem-lineage back to a node defined by the nearest common ancestor of the crown group and its closest living sister-group (in the case of tetrapods, that would be either lungfishes or coelacanths).
There are some major practical problems with node-based definitions. First, almost inevitably, by the "crown-group" definition, many fossil taxa which indisputably have limbs with digits are excluded. Thus Acanthostega and Ichthyostega would not belong within the Tetrapoda, since they are stem tetrapod plesions outside the crown group. A strict application of the crown-group method not only forbids the use of the term "Tetrapoda" for these animals, but disqualifies them as "tetrapods". Conversely, by the "total-group" definition Tetrapoda would include many animals which clearly do not have limbs with digits, not only "fish" such as Eusthenopteron and Panderichthys but the rest of the osteolepiform lineage.
Both node-based definitions are equally rigorous, defensible and valid phylogenetic definitions of the group Tetrapoda but in both cases, the result is counterintuitive in respect of some animals. The problem arises not with node-based definitions per se, but with the logical difficulty of applying to a node-defined group a name whose origins are etymologically founded upon an innovative "key-character" as so many higher taxonomic group names are. Lee and Spencer (1997) have pointed out other problems concerned with stability of node-based groupings.
Either way in this case, the origin of limbs and the origin of terrestriality are distanced from the origin of tetrapods, which may be an advantage in that it separates our perceptions of these three independent evolutionary strands.