A Phylogeny and a New Classification of the Corophiidea Leach, 1814 (Amphipoda)

Abstract

Based on a phylogenetic analysis of 104 genera and 156 species of corophiidean amphipods, we present a new phylogeny and higher-level classification for the suborder Corophiidea Leach, 1814. The phylogeny divides the corophiideans into two infraorders, the Corophiida and the Caprellida, based on a hypothesis of the evolution of different feeding strategies. Members of the Corophiida are derived from bottomfeeding detritivores, whereas members of the Caprellida are derived from ancestors that fed on material suspended in the water column. Within the Corophiida there are unspecialized clades such as the aoroids, whose members are mainly detrital feeders, as well as specialized forms feeding on living algae (family Ampithoidae) and on wood (superfamily Cheluroidea). In the Caprellida, members of the podocerid and caprellid clades often climb organisms such as hydroids in order to get their antennae as far up in the water column as possible to suspension-feed, whereas ischyrocerins build nests and then suspension-feed. Specialized forms include cyamids that attach to whales where they ‘‘feed on whale skin.’’ Barnard and Karaman (1984) divided the Corophiidea into two superfamilies (Corophiidea and Caprelloidea) within which were nine families. The classification presented here includes those two infraorders, with 11 superfamilies, 21 families, 13 subfamilies, and 5 tribes. Barnard and Karaman (1984) proposed an untested scheme, which placed the infraorders Caprellida (as Caprelloidea) and Corophiida (as Corophioidea) in the suborder Corophiidea Leach (1814). It should be noted that Stebbing (1906) attributed the Corophiidae to Dana (1849), but this is in error. Leach (1814: 405) erected the Corophiidae (as the family Corophiini) where it has page placement priority over the family Caprellini. This error has since been perpetuated by all authors, including Barnard and Karaman (1984, 1991). Within the corophiideans (sensu lato), the internal phylogenetic relationships have always been uncertain. Barnard and Karaman (1991) attempted no phylogenetic classification of their infraorder Corophiida, commenting that ‘‘the practical identification of family or subfamily groups in this conglomerate is very difficult.’’ Laubitz (1993) and Takeuchi (1993) both proposed untested schemes for the Caprelloidea (in the traditional sense). The lack of clarity that surrounds the relationships of this group is not unique within the Amphipoda, but it has been particularly intractable in the corophiideans. We should not, however, be surprised. The punctuated nature of evolution may always result in low-resolution trees, which are based on extant taxa. Whenever a new innovation occurs during the course of evolution, it is probably followed by explosive adaptive radiation as the innovator invades new niches, further developing and adapting its innovation. Most of the descendants of the early innovator will fail and become extinct; only a few will succeed to form new clades. Such bursts of evolutionary innovation make it difficult, if not impossible, to distinguish the dichotomous nature of character-state acquisition by examining only extant taxa. Interpretation of phylogeny through a cladistic study of living species will result in trees with unresolvable polychotomies. The precision of a cladistic analysis will vary with the hierarchical taxonomic level analyzed (and hence to temporal scale). Analyses of species groups should generally yield rather clear relationships, because there are fewer missing links, but the clarity will decrease, perhaps exponentially as earlier and earlier relationships are investigated. This means that we should, perhaps, be rather skeptical of a tree pertaining to family relationships that purports to be fully resolved. The best we can expect, and for the present with which 443 JOURNAL OF CRUSTACEAN BIOLOGY, 23(2): 443–485, 2003