摘要
Gobiarian fishes, exemplified by gobies, sleepers and cardinalfishes, have radiated acrosscoastal marine and aquatic habitats worldwide, yet the biological traits deemed responsible forgenerating their great diversity, such as small body size, short generation times, andecomorphological specialization, have also hindered resolution of their phylogeny. The currentclassification of Gobiaria is based on molecular phylogenetics, and while broad relationshipsamong major groups have largely been settled by independent investigations, the placement ofall higher taxa have not been verified with multilocus data. The root topology recovered fromdifferent molecular datasets is contentious, with multilocus and phylogenomic studies resolvingnursery and cardinalfishes either in reciprocal or sequential sister clades to other gobiarians,and this deep systematic controversy questions the current two-order classification. Here I usedtwo complementary approaches to resolve the phylogeny of Gobiaria. In my first study I minedpublic databases to assemble a sparse supermatrix of 23 genes and construct a phylogenetictree with dense taxon sampling, comprised of approximately 30 percent of the more than 2,400 known gobiarian species. In my second study I generated new ultraconserved elementsequence data to assemble a phylogenomic matrix of 704 genes, construct a phylogenetic treewith comprehensive sampling of higher taxa, and estimate a timescale for diversification underthe relaxed molecular clock. Overall my studies produced more evidence in support of thecurrent two-order classification of Gobiaria, and revealed the root uncertainty is a result ofancient incomplete lineage sorting. I also discovered that collared wrigglers (Xenisthmus spp.)form a monophyletic clade separate from sleepers (Eleotridae), and recommend recognizingfamily Xenisthmidae in the clade-based classification of Gobiaria. I dated origination of Gobiariain the youngest age of the Early Cretaceous (104 Ma), found major clades of gobies, sleepers,and cardinalfishes diverged in the early Eocene (~50 Ma), and placed goby lineagediversification in the Oligocene and Miocene. In summary, my studies support the currenttwo-order classification placing nursery and cardinalfishes in Kurtiformes and the remaininggobiarian fishes in Gobiiformes, confirm the clade-based phylogenetic classification of gobiarianfamilies, and advance evidence for recognizing collared wrigglers in family Xenisthmidae.