Saturday, September 15, 2012

Snapshots of Old Fish Phylogenies, Part I

Phylogenetics became a science with a consistent and objective methodology after the introduction of phylogenetic systematics, or cladistics, by Willi Hennig in the 1950s and 1960s (1, 2). As a life-long student of ichthyology and a scientist who specializes in phylogenetic methods, I have been interested in hypotheses and graphical depictions of phylogenetic relationships of ray-finned fishes that were published prior to the introduction of cladistics.

There is no longer much attention paid to these early views of fish phylogeny, which I think is unfortunate. There is an opinion that with the advent of cladistics, there is no need to study and understand these pre-cladistic hypotheses of fish relationships. However, it is important to note that most biologists in the 19th Century immediately accepted Darwin’s fundamental thesis that all life on Earth shares common ancestry (3). Notable examples of ichthyologists that never accepted evolution include Louis Agassiz, a professor at Harvard University, and Albert K. L. G. Günther. The late 19th and early 20th Century ichthyologists that were thinking about how lineages of fishes were related to one another were explicitly attempting to create taxonomies that reflect hypothesized genealogical relationships. The problem is that prior to Hennig there was no standard method to infer these relationships, which meant that even when using the same type of information scientists could arrive at dramatically different conclusions about phylogeny.

Phylogeny of teleost fishes from E.D. Cope's book, Primary Factors of
Organic Evolution, 1896.
After starting to offer an ichthyology course at the University of Tennessee, and now at Yale University, I started to become interested in the history of the scientific study of fish phylogeny to augment and accentuate my lectures. My foray into pre-cladistic fish phylogenies was greatly helped by Colin Patterson’s masterly review that successfully demonstrates the influence fossil lineages had on the study of teleost relationships (4), with Patterson’s thesis that the most direct path to the inference of phylogeny is the study of living lineages (5).

It is not entirely clear to me what we can specifically learn by studying pre-cladistic efforts at fish phylogeny. Will we discover a hypothesis that is now again finding support in explicit post-Hennig phylogenetic analyses, or will we see reflections of both method and theory that will allow a more nuanced view of how we approach phylogeny inference in the 21st Century? Even if there are no obvious undiscovered gems in these old phylogenetic trees, an understanding of this history will minimally allow us to appreciate the set of objective approaches shared by most comparative biologists interested in phylogeny, regardless of the group of organisms investigated. What I think we do see in these old trees is that the approach used by different scientists to infer relationships was idiosyncratic and often limited by the patterns of biological diversity exhibited in the specific organismal lineage.

Phylogeny of fishes from Gill's 1871 work,
Arrangement of the Families of Fishes, or
Classes Pisces, Marsipobranchii, and
Among the earliest diagrammatic fish phylogenies are from 1866 (6) by Ernst Haeckel (1834-1919), who is credited with the introduction of the term “phylogeny.” One of the first comprehensive post-Darwin classifications of fishes was introduced by Edward D. Cope (7-9). Cope emphasized osteology as evidence for his conclusions on relationships of fishes (8, p. 446). Cope’s publications on ray-finned fish classification did not include any diagrammatic representation of phylogeny, but his work is noteworthy for the description of Actinopteri and the delimitation of Nematognathi that includes catfishes and sturgeons, and the decision to not recognize Müller’s Teleostei. Twenty-five years later Cope published a branching diagram that represented his views on relationships of what would encompass the teleosts in his book “Primary Factors of Organic Evolution (10).”

The earliest fish phylogeny shown here is from Theodore Gill’s very influential and informed classification of fishes that includes Cope’s Nematognathi and Müller’s Teleostei (11, p. xliii), which as mentioned above, was not recognized by Cope.

Phylogeny of ray-finned fishes from Dean's book
Fishes, Living and Fossil, 1895.
The next fish phylogeny shown here is from Bashford Dean’s (1867-1928) book “Fishes, living and fossil” published in 1895 (12). Dean does not discuss any specifics of this tree, but it appears to reflect Cope's classification because it shows a close relationship of catfish and sturgeons (Nematognathi). Dean was impressed with teleosts stating, “Teleosts have diverged most widely of all fishes from what seem to have been their primitive structural conditions.” Later, Dean notes that convergent evolution is pervasive among teleosts and this makes inferences regarding relationships within the group difficult, “Environment, like a mould, has impressed itself upon forms genetically remote, and in the end has place them side by side, apparently closely akin, similar in form and structure (12, p. 167).”

Part II will begin with George A. Boulenger.


1.  Hennig, W. 1950. Grundzüge einer Theorie der phylogenetischen Systematik. Berlin: Deutscher Zentralverlag.
2.  Hennig, W. 1966. Phylogenetic systematics. Urbana: University of Illinois Press.
3.  Darwin, C. 1859. On the origin of species. London: John Murray.
4.  Patterson, C. 1977. The contribution of paleontology to teleostean phylogeny, in Major patterns in vertebrate evolution, P.C. Hecht, P.C. Goody, and B.M. Hecht, Editors. Plenum Press: New York. p. 579-643.
5.  Patterson, C. 1981. Significance of fossils in determining evolutionary relationships. Annual Review of Ecology and Systematics. 12:195-223.
6.  Haeckel, E. 1866. Generalle morphologie der organismen. Berlin: G. Reimer.
7.  Cope, E.D. 1871. Observations on the systematic relations of the fishes. American Naturalist. 5:579-593.
8.  Cope, E.D. 1871. Contribution to the ichthyology of the Lesser Antilles. Transactions of the American Philosophical Society. N.S., 14:445-483.
9.  Cope, E.D. 1872. Observations on the systematic relations of the fishes. Proceedings of the American Society for the Advancement of Science. 20:317-343.
10.  Cope, E.D. 1896. Primary factors of organic evolution. Chicago: The Open Court Publishing Company.
11.  Gill, T.N. 1872. Arrangement of the families of fishes, or classes Pisces, Marsipobranchii, and Leptocardii. Smithsonian Miscellaneous Collections. 11:i-xlvi, 1-49.
12.  Dean, B. 1895. Fishes, living and fossil. New York: Columbia University Press.

Friday, September 14, 2012

Nuclear Genes Result in a Resolved Squirrelfish-Soldierfish Species Tree

Fifteen years ago Wayne Maddison articulated a vision for phylogenetics, which was realized with the introduction of methods that incorporate coalescent theory into molecular phylogenetic inferences (1, 2). These “species tree” methods attempt to estimate the optimal tree that contains the set of sampled gene trees, using simple coalescent models to account for ancestral polymorphism (3).

There is a growing frequency of fish phylogenetic studies that are using multi-locus DNA sequence datasets analyzed with species tree methods (e.g., 4). Most molecular phylogenetic studies of closely related animals have relied on mitochondrial DNA because there is no recombination in the mitochondrial genome and the rate of nucleotide substitution is substantially higher than nuclear DNA. Also, universal primers to amplify specific regions of the mtDNA genome have been around since the late 1980s (5). This historical reliance on mtDNA means that we do not really know if we can expect phylogenetic resolution for clades of closely related species when sampling DNA sequences from a modest number of nuclear genes.

A print of Holocentrus ascensionis sampled
from Wikipedia
My graduate student Alex Dornburg is working on the phylogenetic relationships and patterns of diversification in the teleost clade Beryciformes. The first paper from his work on these fishes was published this week in MPE (6). Alex’s paper focuses on the phylogenetic relationships of the beryciform lineage Holocentridae (Squirrelfishes and Soldierfishes), a clade of 84 large eyed and primarily nocturnal species. The most recent common ancestor of living holocentrid species is minimally dated at 50 million years ago based on the phylogenetic relationships of two fossil holocentrid species from the Pesciara beds at Bolca, Italy (7-10). The Eocene age of the clade indicates that many holocentrid species likely share a fairly recent common ancestry.

Alex worked with a closely with a Yale College student, Rachel Webster, to compile a nice dataset of DNA sequences of six protein coding genes sampled from 39 holocentrid species. A single mtDNA gene was sequenced as well because data for this gene was available on Genbank for Corniger spinosus, which we did not have access to specimens for DNA sequencing. The phylogenetic analyses of the concatenated nuclear genes, as well as the species tree inference, were well resolved with very impressive bootstrap and Bayesian posterior clade support. We included the mtDNA gene and the 6 nuclear genes in a subsequent set of phylogenetic analyses, including a species tree inference. Inclusion of the mtDNA tree did not appear to alter the phylogenetic inferences, nor did it increase the posterior clade support in the *BEAST inferred species tree. These results indicate that DNA sequences sampled from a set of nuclear genes result in a strong inference of phylogenetic relationships among the closely related species of Holocentridae.
Phylogenetic relationships of Holocentridae inferred from six nuclear genes. A. Optimal tree from RAxML anlaysis, black and open circles are strong bootstrap values. B. Maximum clade credibility tree from a MrBayes analysis, black and open circles denote nodes with a Bayesian posterior equal to or greater than 0.95. C. *BEAST inferred maximum clade credibility species tree, black and open circles denote nodes with a Bayesian posterior equal to or greater than 0.95.

*BEAST inferred maximum clade credibility species tree inferred from 
six nuclear genes and a single mtDNA gene, black and open circles
denote nodes with a Bayesian posterior equal to or greater than 0.95.
The inferred phylogenies are orthodox in the resolution of the two major holocentrid taxonomic groups, Holocentrinae (Squirrelfishes) and Myripristinae (Soldierfishes), as monophyletic; however, two of the three genera of Holocentrinae (Sargocentron and Neoniphon) are dramatically non-monophyletic in these phylogenies. We use the phylogenetic trees to show that many of the characters used to diagnose and delimit these genera have a complex evolutionary history. A new taxonomy for Holocentrinae is proposed where several species previously classified as Sargocentron are treated as species of Neoniphon, Sargocentron is restricted to the least inclusive clade containing S. spiniferum and S. rubrum, and the subgenus Flammeo is elevated to contain F. marianus. Holocentrus is monophyletic in the phylogenies.

These results are exciting, as they indicate that molecular datasets being collected by ichthyologists in the early 21st Century provide well-resolved phylogenies for closely related lineages, as well as the deepest branches in the Actinopterygian Tree of Life (e.g., 11). Also, these well resolved phylogenies reveal substantial incongruence between the formal classifications of species and inferred phylogenetic relationships. Perhaps most exciting are the numerous evolutionary studies, aimed at understanding the mechanisms that have generated the abundant species diversity of fishes, which will be guided by these well resolved species level phylogenies.


1.  Maddison, W.P. 1997. Gene trees in species trees. Systematic Biology. 46:523-536.
2.  Edwards, S.V., L. Liu, and D.K. Pearl. 2007. High-resolution species trees without concatenation. Proceedings of the National Academy of Sciences of the United States of America. 104:5936-5941.
3.  Edwards, S.V. 2009. Is a new and general theory of molecular systematics emerging? Evolution. 63:1-19.
4.  Hollingsworth, P.R. and C.D. Hulsey. 2011. Reconciling gene trees of eastern North American minnows. Molecular Phylogenetics and Evolution. 61:149-156.
5.  Kocher, T.D., W.K. Thomas, A. Meyer, S.V. Edwards, S. Paabo, F.X. Villablanca, and A.C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in animals:  amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences of the United States of America. 86:6196-6200.
6.  Dornburg, A., J.A. Moore, R. Webster, D.L. Warren, M.C. Brandley, T.L. Iglesias, P.C. Wainwright, and T.J. Near. 2012. Molecular phylogenetics of squirrelfishes and soldierfishes (Teleostei: Beryciformes: Holocentridae): reconciling more than 100 years of taxonomic confusion. Molecular Phylogenetics and Evolution. 65:727-738.
7.  Sorbini, L. 1975. Gli Holocentridae di Monte Bolca. II: Tenuicentrum pattersoni nov. gen. nov. sp.  Nuovi dati a favoure dell’origine monofiletica dei beryciformi (Pisces). Studi e Ricerche sui Giaciamenti Terziari di Bolca. 2:456-472.
8.  Sorbini, L. 1979. Gli Holocentridae di Monte Bolca. III. Berybolcensis leptacanthus (Agassiz). Studi e Ricerche sui Giaciamenti Terziari di Bolca. 4:19-35.
9.  Sorbini, L. 1975. Gli Holocentridae di Monte Bolca.  I: Eoholocentrum, nov. gen., Eoholocentrum macrocephalum (de Blainville) (Pisces-Actinopterygii). Studi e Ricerche sui Giaciamenti Terziari di Bolca. 2:205-228.
10.  Stewart, J.D., Taxonomy, paleoecology, and stratigraphy of the halecostome-inoceramid associations of the North American Upper Cretaceous epicontinental seaway. 1984, University of Kansas: Lawrence. p. 201.
11.  Near, T.J., R.I. Eytan, A. Dornburg, K.L. Kuhn, J.A. Moore, M.P. Davis, P.C. Wainwright, M. Friedman, and W.L. Smith. 2012. Resolution of ray-finned fish phylogeny and timing of diversification. Proceedings of the National Academy of Sciences of the United States of America. 109:13698-13703.

Monday, September 3, 2012

An Adaptive Radiation on Ice

 In February of 2012 we published a set of comparative analyses to investigate the adaptive radiation of Antarctic notothenioids, a clade of percomorph teleosts. The authors included a graduate student in my lab, Alex Dornburg, a former postdoc, Kristen Kuhn, the venerable master of Antarctic fishes, Joseph T. Eastman, a former Near Lab Yale undergraduate, Jillian Pennington, my valued Italian colleagues, Tomaso Patarnello and Lorenzo Zane, a fish physiologist from Argentina, Daniel A. Fernández, and my oft partner in high seas Southern Ocean mischief, Christopher D. Jones who was our cruise leader.

Artedidraco skottsbergi (9 cm) collected in the Bransfield Strait.
Photograph by Thomas J. Near
Dissostichus mawsoni,
Antarctic Toothfish
The term adaptive radiation has been applied to very few marine fish lineages. The notothenioids exhibit anti-freeze glycoproteins (AFGP) to avoid freezing in the subzero waters of the Southern Ocean around Antarctica. Antarctic notothenioids exhibit substantial morphological and ecological differences among closely related species, ranging from the pelagic Antarctic Toothfish, Dissostichus mawsoni, which can reach 1.75 m in length, to the entirely benthic Artedidraco skottsbergi that has a maximum size of only 11 cm. All notothenioids lack a swim bladder, the primary buoyancy organ of ray-finned fishes and homologous to our lungs; however, several lineages of notothenioids have evolved modifications of buoyancy through reduction of ossification and lipid deposits. Some notothenioids are neutrally buoyant, meaning they have no weight in seawater. The differences in buoyancy among closely related species is nearly as substantial as differences in their morphology.

The field work for this study encompassed four expeditions (2001, 2003, 2006, and 2009) to the Southern Ocean onboard the Russian research vessel Yuzhmogeologia. On these cruises I was able to make buoyancy measurements from more than 1,000 specimens, and I collected more than 2,000 specimens and tissue samples that are now housed in the fish collection of the Peabody Museum of Natural History.

Eleginops maclovinus, the non-Antarctic sister species to the entire notothenioid 
Antarctic Clade. Maximum size is 90 cm.
There are approximately 100 species that comprise an unnamed Antarctic Clade that is considered an adaptive radiation, but there are also three species-poor lineages that are early branching and endemic to non-Antarctic cold-temperate near-shore habitats in South America, Falkland Islands (Islas Malvinas), Tristan de Cunha, Australia, Tasmania, and New Zealand. These lineages include Bovichtidae, Pseudaphritis urvillii, and Eleginops maclovinus. These lineages were never exposed to the freezing conditions of the Southern Ocean and provide an important contrast to understand the origins of the adaptations to polar conditions, as well as the ecological and morphological diversity observed in the Antarctic Clade.
Distribution of Eleginops maclovinus

Previous work that I had published, using relaxed molecular clock analyses of mitochondrial DNA rRNA gene sequences, indicated that the Antarctic Clade originated around 24 mya. This is well after the onset of global cooling and appearance of the first Antarctic ice sheets at the Eocene-Oligocene boundary, some 35 mya. Subsequent molecular age estimates using primarily DNA sequence data from nuclear genes have confirmed this result. However, there was no analysis that assessed if Antarctic notothenioids exhibit an explosive diversification subsequent to their origin, a pattern expected in an adaptive radiation.

My colleagues and I investigated the tempo of diversification in notothenioids through analyses of a time-calibrated phylogeny that was inferred from DNA sequences of five nuclear genes (myh6, sh3px3, tbr1, zic1, and the first intron of the S7 ribosomal protein) and two mitochondrial genes (ND2 and 16S rRNA). The phylogenetic analysis included 83 of the approximately 120 recognized notothenioid species.

The inferred molecular phylogeny is quite orthodox, relative to previous inferences from morphological and molecular data. For example, Bovichtidae (containing Bovichtus and Cottoperca) and Pseudaphritis urvillii are successive early diverging non-Antarctic lineages, and the South American and Falkland Island distributed Eleginops maclovinus (pictured above) is the sister lineage of the Antarctic Clade. Within the Antarctic Clade the only noteworthy result is the paraphyly of Nototheniidae, but support for the inferred relationships among the earliest diverging lineages of the Antarctic Clade is limited.

Estimating the ancestral character states of AFGP on the notothenioid time tree supports a single origin of the trait along the branch subtending the most recent common ancestor of the Antarctic Clade. A method that detects changes in lineage diversification rate identified the Antarctic Clade and the common ancestor of Artedidraconidae, exclusive of Artedidraco skottsbergi, as exhibiting high diversification rates relative to the estimated background diversification rate. However, when we performed additional temporal analyses, it was clear that there are pulses of lineage diversification among much younger lineages in the Antarctic Clade. In other words, a shift of diversification rate leading to the Antarctic Clade is consistent with the expectation that AFGP was a trigger to the notothenioid adaptive radiation, we also observed temporal periods of high diversification that are at least 10 million years after the origin of AFGP. It does not seem that the origin of AFGP tells the whole story of the notothenioid adaptive radiation.

It is important to note that unlike any other set of near-shore marine habitats on Earth, the fish fauna of the Southern Ocean is completely dominated by the closely related notothenioids. This was not always the case. A fossil fish fauna from Seymour Island at the tip of the Antarctic Peninsula clearly shows that about 40 mya Antarctica was the home of a temperate and phylogenetically diverse teleost fauna that was not much different from what I can see off shore from New Haven in Long Island Sound. As polar conditions developed in the Southern Ocean this fauna was extirpated, opening up a set of niches that were ultimately exploited by notothenioids. In this paper we propose that the particularly harsh environmental and physical conditions of near-shore habitats in Antarctica, particularly starting about 14 mya, may have led to a cycle of lineage extirpation, recolonization, and diversification within the notothenioid Antarctic Clade. This hypothesis is supported by our finding of pulses of diversification associated with the most species-rich notothenioid lineages (e.g., Artedidraconidae, Trematomus, and Channichthyidae) that correspond to a time around 10 to 5 mya.

Vomeridens infuscipinnis, a species of Bathydraconidae, or dragonfishes.
This species is very light in water, with the lowest buoyancy among measured
bathydraconids. Photograph by Thomas J. Near
Trematomus newnesi, a semi-pelagic species of Nototheniidae
that has an intermediate buoyancy. Photograph by Thomas J. Near
When we examined the pattern of diversification of buoyancy in notothenioids, there is a clear pattern of substantial among clade disparity early in the diversification of the Antarctic Clade, just as expected in an adaptive radiation. However, two of the most species-rich notothenioid subclades, Channichthyidae and Trematomus, show substantial within lineage disparity throughout their respective histories. This indicates that unlike the Antarctic Clade as a whole, closely related species in these lineages are evolving substantial differences in buoyancy.

In this paper we pointed out that regions of the Southern Ocean are among the fastest warming regions on Earth. Such dramatic thermal changes may be cause for concern when considering the future of these cold-adapted fishes. From the perspective of studying the notothenioid adaptive radiation, future work should aim to provide greater phylogenetic resolution among the early diverging lineages of the Antarctic Clade, assess the relationships between habitat and dietary resource utilization, and continue to refine our understanding of species diversity in the clade. Notothenioids are a most fascinating lineage of animals and the fact they are fishes just adds that much to their allure and ensures that we will be paying attention to them for the foreseeable future.