Molecular phylogenetic analysis of life-history evolution in asterinid starfish.
The ecological and evolutionary factors that lead to diverse life histories of marine invertebrates are not well known, in part because such life-history variation is usually difficult to study using the sophisticated demographic or quantitative genetic methods that are often applied to life histories of terrestrial insects, plants, and vertebrates (e.g., Lack 1947; Istock 1967; Charlesworth 1980; Wilbur 1980; Charnov 1982; Lande 1982; Rose 1985; Werner 1986; Caswell 1989; Ebenman 1992; Roff 1992; Stearns 1992). For example, projection matrix methods (Leslie 1945) require knowledge of parameters such as offspring mortality rate that are difficult to measure for small larvae that are widely dispersed in a large ocean.
Molecular phylogenies and the comparative method (Felsenstein 1985b; Harvey and Pagel 1991) provide an alternative tool for testing ideas about the frequency, order, and timing of life-history changes in various marine groups (e.g., Lieberman et al. 1993; Hadfield et al. 1995; o Foighil and Smith 1995; Reid et al. 1996). Here we present a phylogenetic analysis of the evolution of life-history traits among species of asterinid starfish (Patiriella and Asterina) that have evolved different larval forms (reviewed by Byrne 1991). We derived a phylogeny for these species from complete mitochondrial DNA (mtDNA) sequences for an enzyme-coding gene and five transfer RNA genes. Earlier efforts to understand relationships among some Japanese asterinids were based on immunological distances (Mochizuki and Hori 1980), allozyme polymorphisms (Matsuoka 1981), or mitochondrial ribosomal RNA genes (Wada et al. 1996), which all gave poorly resolved results.
We used our mtDNA phylogeny for Patiriella and Asterina species to test a series of hypotheses about the evolution of larval forms and modes of larval development in marine invertebrates.
1. Is a complex, feeding larva ancestral for these starfish? All feeding larvae of starfish share similarities in coelom formation, gut function, ciliary structures, and feeding behavior, while nonfeeding starfish larvae are distinguished mainly by the loss of these traits and the presence of large amounts of nutritive yolk (Chia and Walker 1991). Studies of larval forms intermediate between feeding and nonfeeding development suggest that the traits characteristic of feeding larvae of echinoderms can be lost one at a time (Emlet 1986; Olson et al. 1993; Eckert 1995; Hart 1996). [TABULAR DATA FOR TABLE 1 OMITTED] If feeding larvae could be lost and regained in some lineages, by convergent atavisms of many traits, then homoplasies in feeding larval morphology should be common (Strathmann and Eernisse 1994), but such homoplasies are not known for any starfish. Thus the feeding larval form is probably ancestral for starfish and other echinoderms, and multiple losses of feeding larvae are more likely than parallel gains of the identical suite of traits for feeding in the plankton (Jagersten 1972; Strathmann 1978, 1985, 1988; McEdward and Janies 1993). Multiple origins of feeding larvae have been suggested for other groups (Reid 1990; Haszprunar et al. 1995; Reid et al. 1996). Phylogenetic information can formalize this problem by showing the number of gains or losses of a feeding larva that are implied by different assumptions about character change and polarity (Strathmann and Eernisse 1994).
2. Are evolutionary changes in larval form frequent or rare? The evolution of early development is traditionally thought to be conservative (but see Wray 1995). Demonstrating abundant changes in larval form among close relatives would falsify this assumption.
3. Do changes in life histories occur in a predictable transformation series? We tested a specific hypothesis about the association between small adult size and brooding hermaphroditism in starfish (Strathmann et al. 1984; Byrne and Cerra 1996). based on the conditional advantages of brood protection and hermaphroditism in small adults, Strathmann et al. predicted that these traits evolve in the following order: (a) reduced adult size, (b) benthic development of nondispersing offspring, and (c) self-fertilizing hermaphroditism.
We also tested a more general hypothesis about the order of evolutionary changes in larval nutrition, larval habitat, and brood protection (Garstang 1929; Strathmann 1985; McEdward and Janies 1993; Byrne and Cerra 1996; Table 1). based on shared patterns of larval form and ecological conditions, McEdward and Janies (and others) have suggested that ancestral feeding larval forms give rise to various types of nonfeeding larval development by (a) loss of larval feeding structures and gain of large, yolky eggs, followed by (b) loss of planktonic dispersal, and then (c) gain of parental brood protection.
If these transformation series cannot be demonstrated by phylogenetic inference, then our understanding of how larval forms and life histories evolve must be incomplete.
4. Are changes in larval form relatively old or recent? If some of these changes are relatively recent, then the ecological or evolutionary factors that cause such changes may still be operating in contemporary populations and might be discovered.
5. Has diversity of larval forms arisen mainly by species selection or mainly by transformation of lineages (Lieberman et al. 1993; Strathmann and Eernisse 1994)? Lineages with nonfeeding larval development could accumulate in clades by a process of species selection if this life-history shift affects rates of speciation or extinction (Hansen 1982; Jablonski 1986, 1987). Species selection would be implied for asterinids if most speciation events occurred in ancestors inferred to have nonfeeding larvae.
Life Histories of Asterinid Starfish
The genus Patiriella includes the greatest diversity of larval types and life histories known among extant starfish (Emlet et al. 1987; Chia and Walker 1991; Byrne 1996; Byrne and Cerra 1996). Embryogenesis and larval forms in this genus have been described in detail (Dartnall 1969, 1971; Crump 1971; Chia 1976; Keough and Dartnall 1977; Lawson-Kerr and Anderson 1978; Byrne 1991, 1992, 1995, 1996; Byrne and Barker 1991; Chen and Chen 1992; Byrne and Anderson 1994; Byrne and Cerra 1996). One species (P. regularis) spawns small eggs (150-170 [[micro]meter] diameter) that develop in the plankton as long-lived, feeding larvae similar to the feeding larvae of other starfish orders (Chia and Walker 1991). Four species (P. brevispina, P. calcar, P. gunnii, and P. pseudoexigua) develop in the plankton from large, yolky eggs (380-415 [[micro]meter]) without feeding by the larva, and have reduced or lost many of these distinctive larval structures, such as an open mouth, functional gut, or organized ciliated bands. In all five of these species, adults are dioecious, clutches are large ([approximately][10.sup.3]-[10.sup.6]), fertilization and development occur in the plankton, embryos and larvae receive no postzygotic care, and the dispersal potential of planktonic larvae is much greater than the dispersal capability of the benthic adults. Most of these species are relatively large as adults. Radius from the center of the disk to the tip of the arm in the largest adults (R) is 40 mm for P. gunnii and 60-90 mm for P. brevispina, P. calcar, and P. regularis, Patiriella pseudoexigua is an exception (R [less than] 20 mm).
In contrast, three Patiriella species reproduce as small, hermaphroditic adults (R [less than] 15 mm) with benthic development. Patiriella exigua deposits large, yolky eggs (400 [[micro]meter]) in masses on the substratum, where development occurs without subsequent planktonic dispersal or parental care. Patiriella vivipara and P. parvivipara retain small (135-150 [[micro]meter]) fertilized eggs and highly simplified larvae in the gonad, where metamorphosis occurs at a small juvenile size, followed by intragonadal cannibalism among siblings and live birth of large juveniles. In all three species, clutches are small ([approximately][10.sup.2]-[10.sup.3]), self-fertilization is possible, and the dispersal potential of offspring is no greater than the dispersal potential of small benthic adults. Low dispersal potential is correlated with increased population genetic differentiation (McMillan et al. 1992; Stickle et al. 1992; Hunt 1993) and reduced geographic range (Emlet 1995) in Patiriella and other echinoderms.
As a close outgroup for the phylogenetic analysis, we chose four species from the related genus Asterina, including species with three different modes of larval development. We expected these species to form a monophyletic sister group to Patiriella. Asterina pectinifera and A. miniata are large, dioecious adults (R = 70-90 mm) that develop from small eggs (165-170 [[micro]meter]) into long-lived, feeding planktonic larvae similar to P. regularis (Emlet et al. 1987; Strathmann 1987). Adults of A. gibbosa are mainly small (R [less than] 30 mm), protandrous hermaphrodites that deposit large, yolky eggs (500 [[micro]meter]) in benthic masses like those of P. exigua (Emson and Crump 1976, 1979; Crump and Emson 1978), where their offspring develop as benthic larvae inside the egg mass. Adults of A. pseudoexigua pacifica are small (R [less than] 15 min), simultaneous hermaphrodites that brood their offspring inside the ovotestis as do P. vivipara and P. parvivipara. Whereas P. vivipara and P. parvivipara brood larvae hatched from small eggs, A. pseudoexigua pacifica broods large, yolky eggs about 400 [[micro]meter] in diameter (Komatsu et al. 1990). In addition, A. phylactica (which we were unable to include in our study) is also a small (R [less than] 10 min), simultaneous hermaphrodite that deposits large eggs ([approximately]500 [[micro]meter]) in benthic masses (Emson and Crump 1979), but the adult broods the egg mass and occasional embryos are retained inside the gonad where they undergo metamorphosis (Strathmann et al. 1984). Asterina phylactica and A. gibbosa co-occur, were only recently recognized as distinct species, and are probably sister taxa.
MATERIALS AND METHODS
Patiriella species. - We expected mtDNA genetic distances among asterinid species to be large relative to variation within species, as in other echinoderms (e.g., Palumbi and Wilson 1990, Edmands et al. 1996 on sea urchins; Arndt et al. 1996 on sea cucumbers; Hrincevich and Foltz 1996 on forcipulate starfish), so we used only one individual for most species. In two cases (P. brevispina, P. gunnii), we used one individual from each end of the species' geographic ranges, which span the southern Australian coast from New South Wales (samples labeled E) to Western Australia (samples labeled W).
Asterina species. - For the close outgroup, we sequenced mtDNA of A. gibbosa from Wales, A. miniata from British Columbia, and A. pseudoexigua pacifica from Japan. We also used mtDNA sequences for A. pectinifera from Japan (data from the published complete mitochondrial genome by Asakawa et al. 1995).
Rate Calibration. We sequenced mtDNA from two species of Oreaster (Order Valvatida, Family Oreasteridae): O. occidentalis from the eastern Pacific (Isla Floreana, Galapagos) and O. reticulatus from the western Caribbean (Nalunega, San Blas Islands). These species form a geminate pair separated from each other by the rise of the Isthmus of Panama 3.5-3.1 million years ago (Mya) (Jordan 1908; Keigwin 1978; Lessios 1990). We used the amount of mtDNA sequence divergence between Oreaster species to calibrate the rate of sequence divergence among asterinid species, and thus to estimate the ages of life-history changes among asterinids.
This calibration assumes that (1) rates of mtDNA sequence divergence are similar between asterinids and oreasterids, (2) sequence divergence is similar among asterinids, and (3) the geminate Oreaster species actually diverged at the time of the rise of the Isthmus. Rate variation is common, and some geminate species diverged before the last rise of the Isthmus (Lessios 1990; Knowlton et al. 1993; Lessios 1997). The calibration is based on a single time point, but there is no suitable fossil record for asterinid starfish that could be used for corroboration. If Oreaster species diverged before the rise of the Isthmus, then we will have overestimated the rate of mtDNA sequence divergence and underestimated the ages of life-history changes among asterinid species.
Outgroup Rooting. - We rooted the phylogenetic tree using the mtDNA sequences for Pisaster ochraceus (Order Forcipulatida) in Smith et al. (1990). Classification, collection location, and database accession numbers for taxa used in this study are listed in Table 1.
DNA Extraction, Amplification, and Sequencing
We obtained total genomic DNA from frozen or ethanolpreserved tissue using standard protease digestion, phenolchloroform extraction, and ethanol precipitation (see Arndt et al. 1996). We used tissue from gonads or tube feet where possible (A. miniata, A. gibbosa, P. brevispina, P. calcar, P. exigua, P. gunnii, P. pseudoexigua, P. regularis, P. vivipara, O. species) or whole individuals otherwise (P. parvivipara). We were given genomic DNA extracted from gonads of A. pseudoexigua pacifica by N. Satoh.
We used the polymerase chain reaction (PCR) to amplify a portion of the mitochondrial genome [approximately]2400 bp long. One PCR primer (5[prime]-CTTTGAAGGCTTTTAGTTTAGATTAAC-3[prime]) annealed to the coding strand of the tryptophan transfer RNA gene ([tRNA.sup.trp]) at positions 11725-11751 in the A. pectinifera complete mitochondrial genome (Asakawa et al. 1995). The other PCR primer (5[prime]-CCTARTTGGGTTCARTTTGCC-3[prime]) annealed to the coding strand of the cytochrome oxiduse subunit II gene (COII) at positions 14098-14078 in the A. pectinifera genome.
We performed 50 [[micro]liter] PCR reactions in a GTC thermal cycler using 0.5 [[micro]molar] final primer concentrations. Template DNA was titrated for each sample. Typical amplification conditions were: 90 [degrees] C (2:00), 50 [degrees] C (0:40), 72 [degrees] C (2:00) for 1 cycle; 90 [degrees] C (0:30), 50 [degrees] C (0:30), 72 [degrees] C (1:40) for 30 cycles; 90 [degrees] C (0: 40), 50 [degrees] C (0:40), 72 [degrees] C (7:00) for 1 cycle. Weak amplification products were purified in agarose gels, solubilized and precipitated from the agarose, and used in a second amplification of 25 cycles using a slightly higher annealing temperature (usually 55 [degrees] C).
This PCR product includes the genes for alanine, leucine (CUN), asparagine, glutamine, and proline tRNAs ([tRNA.sup.ala], [tRNA.sup.leu(CUN)], [tRNA.sup.asn], [tRNA.sup.gln], [tRNA.sup.pro]), a possible TATA promoter region, cytochrome oxidase subunit 1 (COI), arginine tRNA ([tRNA.sup.arg]), and NADH dehydrogenase subunit 4L (ND4L) (Smith et al. 1990; Asakawa et al. 1995). The transfer RNA genes are part of a larger cluster of 13 tRNAs bounded by COI and by NADH dehydrogenase subunit 1 (ND1) (Jacobs et al. 1989). We used one of the PCR primers (PAT2) and a series of 13 internal sequencing primers to sequence six of these genes: [tRNA.sup.ala], [tRNA.sup.leu(CUN)], [tRNA.sup.asn], [tRNA.sup.gln], [tRNA.sup.pro], and COI. Primer sequences are available from MWH. We sequenced the tRNAs from both strands; for most samples, we sequenced more than half of COI from both strands as well.
We directly sequenced these PCR products using the Sequenase 2.0 PCR Product Sequencing Kit (Amersham) and 33P-dATP. We modified the Amersham protocol by adding 1.0 [[micro]liter] DMSO to the annealing reaction, snap-cooling the annealing reaction in dry ice-ethanol, and adding 0.5 [[micro]liter] DMSO to the extension reaction.
Sequences were entered and edited using the Eyeball Sequence Editor (ESEE 3, Cabot and Beckenbach 1989). All COI sequences were 1554 bp long without insertion/deletions. We aligned the tRNA sequences by eye and inserted gaps at 16 sites without reference to secondary structure, resulting in 369 bp aligned sequences. Later comparison with the inferred secondary structures of the five gene transcripts (Asakawa et al. 1995) confirmed that all gaps had been placed in regions other than the stems and anticodon loops, indicating that our alignments are conservative. We treated gap sites as missing in the phylogenetic analyses. We removed from the analysis the short noncoding promoter region between [tRNA.sup.pro] and COI, leaving 1923 bp of aligned sequence. This alignment is available from MWH. We computed genetic distances and nucleotide composition using the ECOMPARE program of A. Beckenbach (Simon Fraser University).
The tRNA and COI sequences form a single mitochondrial "locus," though the rates of change and constraints on evolution of the two kinds of genes are probably different. In some preliminary analyses the tRNA sequences were more useful for resolving relationships among large groups of species but failed to resolve some relationships near the tips of the trees, while the COI sequences reliably resolved relationships near the tips of the trees but often failed to resolve deeper nodes. For these reasons, we combined all of the DNA sequence information into a single dataset rather than analyze the tRNA and COI sequences separately (Kluge 1989).
We produced phylogenetic hypotheses using these 1923 bp nucleotide sequences by maximum parsimony in PAUP 3.1.1 (Swofford 1993), by neighbor-joining in MEGA 1.02 (Kumar et al. 1993), and by maximum likelihood in PHYLIP 3.5c (Felsenstein 1993). All three methods should converge on similar tree topologies when long sequences are analyzed (Tateno et al. 1994). We also used the LogDet transformation method of Lockhart et al. (1994), but those results were the same as or less well-resolved than results using neighborjoining (results not shown). For the parsimony analyses, we used the heuristic search option with TBR branch swapping and simple addition of sequences. For the neighbor-joining analyses, we used Kimura two-parameter distances. Other distance measures, including those designed to account for different nucleotide frequencies in mtDNA (Tamura and Nei 1993), gave similar results. For the maximum likelihood analyses, we used empirical base frequencies and a transition/transversion ratio of 2. We randomized the input order of sequences 10 times (the JUMBLE option). In nine of 10 cases we obtained the same branching order and similar branch lengths, while the tenth case differed in the placement of a single species, thus input order had almost no effect on the result.
For both parsimony and neighbor-joining analyses, we characterized the reliability of each phylogenetic hypothesis by resampling the original dataset 1000 times using the nonparametric bootstrap (Felsenstein 1985a). Bootstrap percentages give a biased but conservative indication of confidence in tree topology (see Felsenstein and Kishino 1993; Hillis and Bull 1993; Sanderson 1995). Under some restrictive conditions in computer simulations and analysis of known phylogenies, Hillis and Bull (1993) found that bootstrap percentages above [approximately]70% provided an accurate indication of the true underlying phylogenetic relationships among taxa. To test bootstrap support for one deep node within the Asterinidae, we analysed the more conservative tRNA sequences alone.
In addition to the bootstrap analysis, we used PAUP to search for all trees less than 1% longer than the most parsimonious tree and to compute the frequency of individual clades among those trees.
We inferred the sequence of amino acids in the translated COI sequences (Himeno et al. 1987), and analyzed amino acid substitutions by the same methods applied to nucleotide data. However, we found only 31 phylogenetically informative sites among all of the 517 aa sequences, so that phylogenies based on these amino acid sequences were all poorly resolved (results not shown).
We analyzed evolution of categorical traits (such as development with feeding or nonfeeding larvae) by mapping traits onto the best phylogenetic tree and optimizing changes in traits using MacClade 3.0 (Maddison and Maddison 1992). We did not try to polarize the directions of evolutionary changes in the asterinids by reference to characters in the outgroup Pisaster (Maddison et al. 1984) because the true sister group to the Asterinidae is not known.
DNA Sequence Features
We found 747 variable sites (excluding gaps) among all of the 1923 bp aligned nucleotide sequences; 609 of these sites were phylogenetically informative (517 in COI, 92 in the tRNAs). Most of the phylogenetically informative sites in COI (430) occurred at third positions in codons.
We found biased nucleotide frequencies in all sequences, similar to the biases reported for other echinoderms (Jacobs et al. 1988; Asakawa et al. 1991; Arndt et al. 1996). The frequency of G in the coding strand sequence of COI was low (16.9% to 19.3% of all nucleotides), and G was especially rare at third positions (8.5% to 13.7%). The frequency of G in the same strand in the group of tRNA genes was also low (11.8% to 15.6%).
Genetic Distances among Taxa
Within the Asterinidae, genetic distances between COI sequences of nominal species ranged from 2.5% to 24.2% (Table 2). Distances between tRNA sequences were smaller (0.8% to 16.5%). The genetic distances between COI sequences were much larger than distances between partial COI sequences of congeneric sea urchins (Strongylocentrotus species; Kessing 1991) or sea cucumbers (Cucumaria species; Arndt et al. 1996).
We found large genetic distances between P. gunnii individuals from eastern and western Australia, which differed by 3.5% (tRNAs) and 10.4% (COI). These large distances are similar to or larger than distances that we found between sister species that are given different names and have nonoverlapping geographic distributions (e.g., A. miniata and A. pectinifera, or P. vivipara and P. parvivipara; Table 2). In contrast, P. brevispina individuals from New South Wales and Western Australia differed by only 0.8% (tRNAs) and 0.2% (COI), comparable to distances between COl sequences of other conspecific echinoderm populations (Palumbi and Wilson 1990; Lessios 1997; A. Arndt, unpublished data).
The geminate Oreaster species differed by 7.8% (tRNAs) and 17.6% (COI). Taking the earliest estimated age (3.5 Mya) of the separation of the tropical Atlantic and Pacific Oceans as a conservative estimate of Oreaster divergence times, these genes are expected to diverge among asterinids at a rate of [TABULAR DATA FOR TABLE 2 OMITTED] about 2.2% [My.sup.-1] (tRNAs) or 5.0% [My.sup.-1] (COI). These rates for COI are higher than rates estimated for COI sequences of geminate pairs of sea urchins from the same region: 1.8% to 2.8% [My.sup.-1] for Echinometra species, and 1.4% to 1.5% [My.sup.-1] for Diadema species (Palumbi 1996; Lessios 1997). The estimated divergence rates for Oreaster mtDNA sequences might be too high if populations in the western Caribbean and eastern Pacific were actually isolated from each other greater than 3.5 Mya by shoaling or other features of the gradual rise of the Panamanian land bridge, as has been demonstrated for some other geminate species pairs (Knowlton et al. 1993).
Analysis of the combined COI and tRNA sequences produced the phylogeny shown in Figure 1. This tree is 2116 steps long, with a consistency index of 0.54. Asterina gibbosa (which develops in benthic egg masses) was the sister species to all of the Indo-Pacific asterinids. We obtained the same topology when using the Oreaster species pair as the outgroup in place of Pisaster. Although Oreaster species are classified in the same order as the asterinids (Table 1), this alternative analysis resulted in shorter internal branches and lower bootstrap support for many clades. This difference may reflect current uncertainty about relationships among families and orders of starfish (Appendix I).
The Indo-Pacific asterinids consisted of two large clades. One of these clades included three nominal Patiriella species from Australia that have planktonic nonfeeding larvae (P. calcar, P. brevispina, and the sibling species of P. gunnii); the sister group to these species was A. miniata and A. pectinifera from the north Pacific that have feeding planktonic larvae. The other large clade included the three Patiriella species from Australia that have benthic development (P. exigua with development in egg masses, and the viviparous brooders P. vivipara and P. parvivipara), as well as the biologically and geographically diverse trio P. regularis (with planktonic feeding larvae), P. pseudoexigua (with planktonic nonfeeding larvae), and A. pseudoexigua pacifica (with viviparous brooding).
All three phylogenetic methods indicated the same topology shown in Figure 1. Bootstrap values for all but one of the nodes in Figure 1 exceeded 80%. Support for the large clade including P. regularis, P. exigua, and their relatives was slightly lower using both maximum parsimony (72%) and neighbor-joining (62%). However, bootstrap support for this clade increased to 98% and 99%, respectively, when we restricted the analysis to substitutions among the more conservative tRNA sequences of the asterinids alone. In addition, most (20/36) of the trees within 1% of the length of the shortest tree ([less than] 2139 steps) included this clade.
An unexpected feature of this tree is the clear demonstration that neither Asterina nor Patiriella is monophyletic. Because there are many other species assigned to Asterina, whose relationships to the species in our study are not known, it is not possible to distinguish between paraphyly and polyphyly for these genus names.
A second surprising feature of this tree is the close relationship between P. pseudoexigua from Taiwan and A. pseudoexigua pacifica from Japan. These lineages differed by just 1.4% (tRNAs) or 2.5% (COI) in the neighbor-joining analysis, or by five steps in the parsimony analysis of tRNAs alone, though these species have different types of larval development and are currently classified in different genera. Unfortunately, we were unable to confirm the morphological identification of the A. pseudoexigua pacifica specimen, though it was collected by M. Komatsu in the same area from which development was originally described in this species (Komatsu et al. 1990).
The Ancestral Larval Form. - We reconstructed ancestral modes of larval development under three different assumptions about transitions between the four character states (feeding planktonic development, nonfeeding planktonic development, development in benthic egg masses, or viviparous brooding). The results suggest that a feeding larva is not the most parsimonious ancestral character state.
If larval forms are assumed to evolve in any order and direction from one character state to another, then the ancestral mode of development for these starfish is ambiguous and cannot be inferred [ILLUSTRATION FOR FIGURE 2A OMITTED]. Any of the four ancestral states would be most parsimonious, indicating only six subsequent changes in mode of development. If some form of nonfeeding larval development was ancestral, then these six changes included two parallel origins of a complex feeding larva: in Patiriella regularis, and in the most recent common ancestor of Asterina miniata and A. pectinifera.
If larval forms of starfish are constrained to evolve in one order but either direction (feeding planktonic development [equivalence] nonfeeding planktonic development [equivalence] benthic development in egg masses [equivalence] viviparous brooding; [ILLUSTRATION FOR FIGURE 2B OMITTED]), then the ancestral mode of development was either a planktonic nonfeeding larva or development in benthic egg masses, and seven changes in larval form evolved among these species, including parallel evolution of feeding larvae and parallel evolution of viviparous brooding.
If changes in mode of development are ordered and irreversible (feeding planktonic development [right arrow] nonfeeding planktonic development [right arrow] benthic development in egg masses [right arrow] viviparous brooding), then a total of nine evolutionary changes are indicated [ILLUSTRATION FOR FIGURE 2C OMITTED]. Note that some reconstructed transformations (for example, from feeding planktonic larvae to nonfeeding benthic larvae in the ancestor of P. exigua, P. vivipara, and P. parvivipara) imply that two changes occurred (loss of larval feeding, followed by the loss of planktonic dispersal), though a planktonic nonfeeding larva is not found among the three extant members of the clade. Thus the assumption of ordered and irreversible loss of an ancestral feeding larval form for these starfish implies a large increase (from six to nine steps) in the number of inferred changes in modes of larval development.
Number of Life-History Changes. - Our results suggest a large number of independent, parallel changes in life histories. If a feeding larva is ancestral for starfish (Strathmann 1978; Chia and Walker 1991), then the adaptations for feeding in the plankton were lost at least four times among Patiriella and Asterina species [ILLUSTRATION FOR FIGURE 3A OMITTED]. Similarly, dispersal by planktonic larvae was lost at least three times and viviparous brooding evolved at least twice [ILLUSTRATION FOR FIGURE 3B, C OMITTED].
Small adult size is associated with the evolution of derived larval forms in echinoderms (Strathmann and Strathmann 1982), but small adults may be ancestral in asterinids and large adults derived in two clades [ILLUSTRATION FOR FIGURE 4A OMITTED]. Hermaphroditism evolved from dioecy three times with independent evolution of simultaneous hermaphrodites in two lineages, and internal fertilization evolved twice [ILLUSTRATION FOR FIGURE 4B, C OMITTED].
Life histories appear to have evolved more often among some species than among others. In the large clade including A. miniata and P. calcar, we found only one change in larval type (from feeding to nonfeeding planktonic larvae). In the other large clade including P. regularis and P. exigua, we found at least four major changes in larval form and larval habitat, including two losses of feeding larvae and two parallel appearances of viviparous brooding.
Order of Life-History Changes. - Is the evolution of selffertilizing hermaphroditism preceded by small adult size and nondispersing larval development (Strathmann and Strathmann 1982; Strathmann et al. 1984)? Our data support some aspects of this hypothesized order of changes. Small adult size (R [less than] 30 mm) appears to be ancestral in these starfish. All three cases of benthic development and both cases of simultaneous hermaphroditism appeared only in clades with small adults [ILLUSTRATION FOR FIGURE 3B, 4A, B OMITTED]. If A. phylactica is the sister species to A. gibbosa, then small adult size evolved before simultaneous hermaphroditism in this case as well. However, in only one of two cases (the viviparous Patiriella species) did we find a pattern consistent with the loss of dispersing larvae before the evolution of simultaneous self-fertilizing hermaphroditism. In A. pseudoexigua pacifica, brooding hermaphroditism may have evolved directly from a dioecious ancestor with planktonic larvae similar to P. pseudoexigua, perhaps by retention and internal fertilization of large eggs in the gonad (see Byrne 1996).
We looked for evidence supporting the more general hypothesis of a four-stage ordered transformation: feeding planktonic development [right arrow] nonfeeding planktonic development [right arrow] benthic development in egg masses [right arrow] viviparous brooding. We could not find an example of this transformation series, though all four stages in the series are found among species in this study. We inferred four losses of feeding larvae, but in only two of these cases was the descendant larval form inferred to be a planktonic nonfeeding larva [ILLUSTRATION FOR FIGURE 2C OMITTED]. We inferred three evolutionary gains of benthic development, but in only one case (A. pseudoexigua pacifica) did we find a pattern consistent with descent from an ancestor with planktonic nonfeeding larvae (as in P. pseudoexigua). Finally, we found two independent origins of brood protection, but only one in which brooding was preceded by external benthic development in egg masses. A second case might be A. phylactica, if this brooder is the sister species of A. gibbosa.
Dating Life-History Changes. - Both cases of viviparity evolved recently. If live-bearing evolved in the most recent common ancestor of P. vivipara and P. parvivipara rather than independently in these two species, then this event occurred 2.1-1.3 Mya based on COI sequence divergence (Table 2). The same evolutionary rate calibration suggests that livebearing in A. pseudoexigua pacifica evolved at most about 0.5 Mya. Other changes in modes of development are older. For example, the loss of a feeding larva in the most recent common ancestor of P. calcar, P. gunnii, and P. brevispina occurred at least 4.1-3.4 Mya, based on tRNA sequence divergences among these species, and at most 6.0-4.5 Mya, based on divergences between these species and the sister group consisting of A. miniata and A. pectinifera.
An alternative interpretation of these differences is that P. calcar, A. miniata, and their close relatives are all recently diverged from each other and few life-history changes have appeared in this young clade, but rates of mtDNA evolution are even higher in this clade than we have estimated from comparisons of geminate Oreaster species, resulting in large genetic distances among members of this clade. We have no reason to suspect large differences in rates of mtDNA evolution between clades of asterinids, but cannot discount this hypothesis without other data, perhaps from fossils or from comparisons of nuclear DNA sequences.
Species Selection versus Lineage Transformation. - Jablonski (1986) showed that gastropods with feeding planktonic larvae have lower background rates of speciation and extinction than their relatives with various forms of nonfeeding larval development, and he suggested that this kind of species selection could contribute to patterns of diversity in modes of larval development. We inferred the frequency of speciation events among asterinid species with feeding or nonfeeding larvae under the assumption of irreversible loss of a feeding larva [ILLUSTRATION FOR FIGURE 2C OMITTED]. Among these 12 species are 11 speciation events. For each of those 11 internal nodes in the phylogeny, the larval form for the ancestor at that node must have been a feeding planktonic larva if any of the extant descendants of that ancestor also develop as a feeding larva. Thus 6 of 11 speciation events occurred in an ancestor with a feeding planktonic larva (open branches in [ILLUSTRATION FOR FIGURE 2C OMITTED]), and 5 speciation events occurred in an ancestor with some form of nonfeeding or nonplanktonic larva. Rates of speciation do not appear to be higher among asterinids with nonfeeding development, though our sample of lineages within this family is small. Species selection may contribute to the diversity of modes of development in these starfish if extinction rates vary among lineages with different larval forms, but we have no information on extinction rates from a phylogeny of extant taxa.
Patterns of Life-History Evolution
Life-history traits of asterinids appear to evolve in a liberal fashion under few constraints. We found a large number of independent, parallel changes in larval form and type of larval development among Patiriella and Asterina species (e.g., four losses of feeding by larvae, three losses of planktonic dispersal by larvae). Some life-history features, such as internal development with small eggs and sibling cannibalism in the viviparous Patiriella species, may be reliable indicators of close relationships (Byrne 1996), but other kinds of lifehistory changes appear to evolve easily in parallel. Recent molecular phylogenetic analyses of life histories in snails (Lieberman et al. 1993; Rumbak et al. 1994; Reid et al. 1996), clams (6 Foighil and Smith 1995), tunicates (Hadfield et al. 1995), sea urchins (Wray 1996), and sea cucumbers (Arndt et al. 1996) have identified multiple parallel changes in larval development and breeding system (see also Normark 1996). Nonmolecular approaches have also identified large numbers of parallel evolutionary changes in life histories of sea urchins and copepods (Emlet 1990; Poulin 1995).
If these results are typical of life-history evolution among closely-related marine invertebrates with complex life cycles, then the ecological features of early development may be less constrained by history and relatedness than is generally supposed (see also Wray 1995). In addition, larval form and type of larval development may give poor indications of phylogenetic relatedness (e.g., Rouse and Fitzhugh 1994). Phylogenetic hypotheses based in part on these kinds of lifehistory traits or on morphological traits tightly associated with larval type may be inaccurate or poorly resolved, due to parallel loss of feeding larval characters as well as convergent evolution of adaptations for life as a nonfeeding larva (Emlet 1994). Wray (1996) illustrated this problem for the evolution of nonfeeding larval development in sea urchins.
We found substantial variation in the rate or age of lifehistory changes among clades and among types of life-history change, based on a provisional calibration from mtDNA divergence of Oreaster species. One large clade included only one identifiable change in larval form, whereas another large clade included at least four substantial changes in larval form, habitat, and brood protection. All four losses of planktonic feeding occurred a relatively long time ago, but both cases of viviparous brooding evolved more recently. The recent evolution of viviparity in these starfish suggests that the factors promoting such changes in modes of development might be discovered in field studies of contemporary populations.
These trends in the age of different life-history changes might indicate that these changes occur in one sequence. However, support for such transformation series was variable. In no case could we trace the evolution of larval form from feeding in the plankton to internal viviparous brooding via planktonic nonfeeding larvae and benthic external development (Table 1; [ILLUSTRATION FOR FIGURE 2C OMITTED]). There are at least four reasons why we might not have found good evidence for this transformation series. First, such transformations could be difficult to identify if we have not included all of the close relatives of the asterinid species in this study. For example, other Asterina species with planktonic, nonfeeding larvae (such as A. batheri from Japan) might form the sister group to Patiriella species with benthic development (Matsuoka 1981). Patiriella and Asterina are not reciprocally monophyletic; therefore, we do not know how many close relatives of Patiriella species might have been missed by our cursory sampling of Asterina species.
Second, some transitions from one larval type to another may occur rapidly without leaving evidence of past transitional forms. For example, morphological and functional intermediates between feeding and nonfeeding larvae have been found among echinoderms and some other marine invertebrate groups (e.g., Okazaki and Dan 1954; Emlet 1986; Bosch 1989; Olson et al. 1993; Hart 1996) but are rare relative to lineages with either obligate feeding or nonfeeding larvae. This rarity suggests that the transition from feeding to nonfeeding larvae is rapid. Other evolutionary changes in larval development may be rapid as well (Wray and Raft 1991). Some branches in the phylogeny along which life-history changes must have occurred [ILLUSTRATION FOR FIGURE 1 OMITTED] are short, implying possible rapid change along these branches.
Third, some transitional steps may be skipped. If viviparous brooding in A. pseudoexigua pacifica was derived [less than] 0.5 Mya via external benthic development (now extinct in this clade), then the series of changes from planktonic lecithotrophy (as in P. pseudoexigua) to viviparity occurred in a surprisingly short time. Alternatively, viviparity in A. pseudoexigua pacifica could be derived directly from a planktonic nonfeeding larva merely by the evolution of internal selffertilization (Byrne 1996). Our phylogenetic results are consistent with this prediction. The recent evolution of brooding in this species is also consistent with the retention of a specialized attachment structure in these larvae: they do not require such a structure for settlement and metamorphosis, but little time has passed in which this adaptation might have been reduced or lost, as in the viviparous Patiriella species (Raft 1987; Byrne and Cerra 1996).
Some of these arguments depend on the accuracy of our estimate of mtDNA sequence divergence rate based on Oreaster species. This estimate was more than twice as high as estimated rates for geminate sea urchin species (see above). If the true rate of sequence divergence for asterinids is lower than we have estimated, then some life-history traits (such as viviparous brooding) among Patiriella and Asterina species evolved less recently than we have suggested.
Finally, the hypothetical transformation series from feeding larvae to live bearing may be wrong altogether. For example, de Fraipont et al. (1996) concluded that (1) external egg guarding and (2) viviparity have both evolved repeatedly among lizards and snakes as alternative tactics for brood protection, and not as sequential stages of a transformation series leading to viviparity. Similarly, external development in egg masses and internal viviparous brooding of starfish larvae may also be alternative adaptations for protection of developing young. Other Asterina species that develop in benthic egg masses (e.g., A. minor, A. atyphoida, A. scobinata) may be phylogenetically distant from any lineages in which viviparity occurs.
The Ancestral Larval Form
Is a complex feeding larva ancestral for asterinid starfish and other marine invertebrates? Strathmann and Eernisse (1994) showed that this question cannot be answered using the phylogenetic approach, even with perfect knowledge about evolutionary relationships among species with different modes of development, because inferring the larval form of an ancestor requires an assumption about the likelihood of particular evolutionary changes (e.g., Haszprunar et al. 1995). Other information is required in order to weigh the likelihood, for example, that identical feeding larval forms could evolve in parallel among distantly related species. Reid (1989, 1990) identified a likely case in which feeding larvae have reevolved in several lineages of littorinid snails, based on similarities between the earliest shells of some feeding larvae and the distinctive larval shells of littorinids and other snails that have nonfeeding development.
Analogous observations of asterinid starfish do not suggest parallel origins of similar feeding larvae. For example, the embryos and larvae of P. regularis appear to resemble by descent the embryos and larvae of a diverse array of starfish families (Byrne and Barker 1991; Chia and Walker 1991), though obvious homoplasies should be found if larval feeding has evolved in parallel. Thus we have no basis for suggesting that a feeding larva might have reappeared during the recent evolutionary history of asterinid starfish, even though that is a most parsimonious reconstruction of evolutionary changes under an assumption of unordered character transformations [ILLUSTRATION FOR FIGURE 2A OMITTED].
However, phylogenetic information can be used to frame the problem of the likelihood of different character transformations. If the irreversible loss of a feeding larva is merely twice as likely as the parallel gain of such a larva, then similar numbers of evolutionary steps are inferred if feeding is ancestral (nine steps; [ILLUSTRATION FOR FIGURE 2C OMITTED]) or derived (eight steps; [ILLUSTRATION FOR FIGURE 2A OMITTED]). If loss is three times more likely than parallel gain, then parallel gain implies more evolutionary steps (nine vs. 10). Other studies of morphological development and gene expression are needed to evaluate these transformation probabilities.
If a feeding larva has been lost but not regained among starfish, then many starfish clades should eventually consist of only species with nonfeeding development. Differences in rates of speciation or extinction might explain the persistence of feeding larvae in some of these clades, but we found no apparent difference in speciation rates for ancestors inferred to have feeding or nonfeeding development. Differences in extinction rates, estimated from the fossil record, could help to explain the long evolutionary lives of ancestral feeding larval forms (Hansen 1982; Jablonski 1986).
Taxonomy of Asterinidae
Our mtDNA phylogeny conflicts with the morphological taxonomy of Patiriella and Asterina species (see also Kessing 1991; Foltz et al. 1996). Clark (1983) and Clark and Downey (1992) also noted numerous ambiguities in the morphological taxonomy of adult Asterinidae. From both perspectives, a complete taxonomic revision of the Asterinidae is needed. Asterina gibbosa (Pennant, 1777) is the type species of the genus (see Clark 1983). In order to retain the genus name Asterina and apply that name to a monophyletic group of species, we must know the identity of species descended with A. gibbosa from a common ancestor. Species outside this clade (however it is eventually defined) could then be assigned to Patiriella or other genera.
Until this revision is complete, we recommend the following taxonomic steps. (1) Revise P. gunnii, possibly splitting it into two or more species. (2) Reassign A. pseudoexigua pacifica to Patiriella and refer to it as P. pacifica. Other starfish taxonomists have suspected that A. pseudoexigua pacifica should be classified as a species of Patiriella on morphological grounds (F. Rowe, personal communication), and our surprising phylogenetic result tends to confirm this suspicion. (3) Survey genetic structure among populations of P. exigua. This species is among the most widely distributed starfish, known from St. Helena in the South Atlantic to the southeastern coast of Australia (Clark 1983; Clark and Downey 1992). How has a small, brooding starfish without planktonic larvae achieved such a wide geographic distribution? Adults and juveniles may be capable of rafting, and migration in surface ocean currents may keep these populations genetically connected (Fell 1962; Johannesson 1988; Helmuth et al. 1994; Byrne 1995). Alternatively, dispersal may be rare but occasionally successful (Palumbi and Wilson 1990), resulting in widely separated and evolutionarily divergent populations in the South Atlantic, South Africa, and Australia.
The early life histories of asterinid starfish show substantial variation among closely related species. Numerous cases of superficially similar, highly simplified larval forms have evolved in parallel, and some changes in breeding system and larval dispersal capability have evolved recently. This variation contradicts the common assertion that evolutionary change is most frequent in the later stages of development (Ridley 1996). This variation is also surprisingly common among marine invertebrates with complex life cycles. Wider application of molecular systematics methods may eventually lead to wider appreciation of the labile nature of evolutionary changes in embryogenesis, larval forms, and associated ecological traits in marine animals.
We are grateful to many colleagues who generously helped us to obtain specimens: J. Boom, L. Cannon, A. Cerra, S.-m. Chao, C.-P. Chen, R. Crump, J. Griffith, A. Hodgson, A. Hoggett, L. Marsh, S. McKillup, B. O'Connor, D. O Foighil, K. Sewell, and H. and C. Sowden. We are especially grateful to H. Lessios for samples of Oreaster species, and to N. Satoh for genomic DNA of Asterina pseudoexigua pacifica. A. Arndt, H. Lessios, and S. Palumbi provided unpublished data or observations. A. Arndt and K. Beckenbach helped with PCR and sequencing. B. Crespi, R. Grosberg, P. Marko, R. Raff, R. Ydenberg, and two anonymous reviewers commented on the manuscript. We were supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) postdoctoral fellowship (MWH), a grant from the Australian Research Council (MB), and a NSERC operating grant (MJS). MWH was also supported by a National Science Foundation grant to R. Grosberg.
ARNDT, A., C. MARQUEZ, P. LAMBERT, AND M. J. SMITH. 1996. Molecular phylogeny of eastern Pacific sea cucumbers (Echinodermata: Holothuroidea) based on mitochondrial DNA sequences. Mol. Phylogen. Evol. 6:425-437.
ASAKAWA, S., Y. KUMASAWA, T. ARAKI, H. HIMENO, K. MIURA, AND K. WATANABE. 1991. Strand-specific nucleotide composition bias in echinoderm and vertebrate mitochondrial genomes. J. Mol. Evol. 32:511-520.
ASAKAWA, S., H. HIMENO, K.-I. MIURA, AND K. WATANABE. 1995. Nucleotide sequence and gene organization of the starfish Asterina pectinifera mitochondrial genome. Genetics 140:1047-1060.
BLAKE, D. B. 1981. A reassessment of the sea-star orders Valvatida and Spinulosida. J. Nat. Hist. 15:375-394.
-----. 1987. A classification and phylogeny of post-Paleozoic sea stars (Asteroidea: Echinodermata). J. Nat. Hist. 21:481-528.
BOSCH, I. 1989. Contrasting modes of reproduction in two antarctic asteroids of the genus Porania, with a description of unusual feeding and non-feeding larval types. Biol. Bull. 177:77-82.
BRUSCA, R. C., AND G. J. BRUSCA. 1990. Invertebrates. Sinauer, Sunderland, MA.
BYRNE, M. 1991. Developmental diversity in the starfish genus Patiriella (Asteroidea: Asterinidae). Pp. 499-508 in T. Yanigasawa, I. Yasumasu, C. Oguro, N. Suzuki, and T. Motokawa, eds. Biology of Echinodermata. Balkema, Rotterdam, The Netherlands.
-----. 1992. Reproduction of sympatric populations of Patiriella gunnii, P. calcar and P. exigua in New South Wales, asterinid seastars with direct development. Mar. Biol. 114:297-316.
-----. 1995. Changes in larval morphology in the evolution of benthic development by Patiriella exigua (Asteroidea: Asterinidae), a comparison with the larvae of Patiriella species with planktonic development. Biol. Bull. 188:293-305.
-----. 1996. Viviparity and intragonadal cannibalism in the diminutive sea stars Patiriella vivipara and P. parvivipara (Family Asterinidae). Mar. Biol. 125:551-567.
BYRNE, M., AND M. J. ANDERSON. 1994. Hybridization of sympatric Patiriella species (Echinodermata: Asteroidea) in New South Wales. Evolution 48:564-576.
BYRNE, M., AND M. F. BARKER. 1991. Embryogenesis and larval development of the asteroid Patiriella regularis viewed by light and scanning electron microscopy. Biol. Bull. 180:332-345.
BYRNE, M., AND A. CERRA. 1996. Evolution of intragonadal development in the diminutive asterinid sea stars Patiriella vivipara and P. parvivipara with an overview of development in the Asterinidae. Biol. Bull. 191:17-26.
CABOT, E. L., AND A. T. BECKENBACH, 1989. Simultaneous editing of multiple nucleic acid and protein sequences with ESEE. CABIOS 5:233-234.
CASTELL, H. 1989. Matrix population models. Sinauer, Sunderland, MA.
CHARLESWORTH, B. 1980. Evolution in age structured populations. Cambridge Univ. Press, Cambridge.
CHARNOV, E. L. 1982. The theory of sex allocation. Princeton Univ. Press, Princeton, NJ.
CHEN, B.-Y, AND C.-P. CHEN. 1992. Reproductive cycle, larval development, juvenile growth and population dynamics of Patiriella pseudoexigua (Echinodermata: Asteroidea) in Taiwan. Mar. Biol. 113:271-280.
CHIA, F.-S. 1976. Reproductive biology of an intraovarian brooding starfish Patiriella vivipara Dartnall, 1969. Am. Zool. 16:181.
CHIA, F.-S., AND C. W. WALKER. 1991. Echinodermata: Asteroidea. Pp. 301-353 in A. C. Giese, J. S. Pearse, and V. B. Pearse, eds. Reproduction of marine invertebrates. Vol. 6. Echinoderms and lophophorates. The Boxwood Press, Pacific Grove, CA.
CLARK, A. M. 1983. Notes on Atlantic and other Asteroidea. 3. The families Ganeriidae and Asterinidae, with description of a new asterinid genus. Bull. Br. Mus. Nat. Hist. (Zool.) 45:359-380.
CLARK, A. M., AND M. E. DOWNEY. 1992. Starfishes of the Atlantic. Chapman and Hall, London.
CRUMP, R. G. 1971. Annual reproductive cycles in three geographically separated populations of Patiriella regularis (Verrill), a common New Zealand asteroid. J. Exp. Mar. Biol. Ecol. 7:137162.
CRUMP, R. G., AND R. H. EMSON. 1978. Some aspects of the population dynamics of Asterina gibbosa (Astroidea). J. Mar. Biol. Assoc. UK 58:451-466.
DARTNALL, A. J. 1969. A viviparous species of Patiriella (Asteroidea, Asterinidae) from Tasmania. Proc. Linn. Soc. NSW 93: 294-296.
-----. 1971. Australian sea stars of the genus Patiriella (Asteroidea, Asterinidae). Proc. Linn. Soc. NSW 96:39-49.
DE FRAIPONT, M., J. CLOBERT, AND R. BARBAULT. 1996. The evolution of oviparity with egg guarding and viviparity in lizards and snakes: a phylogenetic analysis. Evolution 50:391-400.
EBENMAN, B. 1992. Evolution in organisms that change their niches during the life-cycle. Am. Nat. 139:990-1021.
ECKERT, G. L. 1995. A novel larval feeding strategy of the tropical sand dollar, Encope michelini (Agassiz): adaptation to food limitation and an evolutionary link between planktotrophy and lecithotrophy. J. Exp. Mar. Biol. Ecol. 187:103-128.
EDMANDS, S., P. E. MOBERG, AND R. S. BURTON. 1996. Allozyme and mitochondrial DNA evidence of population subdivision in the purple sea urchin Strongylocentrotus purpuratus. Mar. Biol. 126:443-450.
EMLET, R. B. 1986. Facultative planktotrophy in the tropical echinoid Clypeaster rosaceus (Linnaeus) and a comparison with obligate planktotrophy in Clypeaster subdepressus (Gray) (Clypeasteroida: Echinoidea). J. Exp. Mar. Biol. Ecol. 95:183-202.
-----. 1990. World patterns of developmental mode in echinoid echinoderms. Pp. 329-335 in M. Hoshi and O. Yamashita, eds. Advances in invertebrate development. Vol. 5. Elsevier, New York.
-----. 1994. Body form and patterns of ciliation in nonfeeding larvae of echinoderms: functional solutions to swimming in the plankton? Am. Zool. 34:570-585.
-----. 1995. Developmental mode and species geographic range in regular sea urchins (Echinodermata, Echinoidea). Evolution 49:476-489.
EMLET, R. B., L. R. McEDWARD, AND R. R. STRATHMANN. 1987. Echinoderm larval ecology viewed from the egg. Pp. 55-136 in M. Jangoux and J. M. Lawrence, eds. Echinoderm studies. Vol. 2. A. A. Balketa, Rotterdam, The Netherlands.
EMSON, R. H., AND R. G. CRUMP. 1976. Brooding in Asterina gibbosa (Pennant). Thalass. Jugoslav. 12:99-108.
-----. 1979. Description of a new species of Asterina (Asteroidea), with an account of its ecology. J. Mar. Biol. Assoc. UK 59:77-94.
FELL, H. B. 1962 West-wind drift dispersal of echinoderms in the southern hemisphere. Nature 193:759-761.
FELSENSTEIN, J. 1985a. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791.
-----. 1985b. Phylogenies and the comparative method. Am. Nat. 125:1-15.
-----. 1993. PHYLIP (phylogeny inference package). Vers. 3.5c. Distributed by the author. Department of Genetics, Univ. of Washington, Seattle.
FELSENSTEIN, J., AND H. KISHINO. 1993. Is there something wrong with the bootstrap on phylogenies - a reply. Syst. Biol. 42:193-200.
FOLTZ, D. W., W. B. STICKLE, E. L. CAMPAGNARO, AND A. E. HIMEL. 1996. Mitochondrial DNA polymorphisms reveal additional genetic heterogeneity within the Leptasterias hexactis (Echinodermata: Asteroidea) species complex. Mar. Biol. 125:569-578.
GALE, A. S. 1987. Phylogeny and classification of the Asteroidea (Echinodermata). Zool. J. Linn. Soc. 89:107-132.
GARSTANG, W. 1929. The origin and evolution of larval forms. Br. Assoc. Adv. Sci. Rep. 1928:77-98.
GIBSON, G. D., AND E-S. CHIA. 1995. Developmental variability in the poecilogonous opisthobranch Haminaea callidegenita: life-history traits and effects of environmental parameters. Mar. Ecol. Prog. Ser. 121:139-155.
HADFIELD, K. A., B. J. SWALLA, AND W. R. JEFFFRY. 1995. Multiple origins of anural development in ascidians inferred from rDNA sequences. J. Mol. Evol. 40:413-427.
HANSEN, T. A. 1982. Modes of larval development in Early Tertiary neogastropods. Paleobiology 8:367-377.
HART, M. W. 1996. Evolutionary loss of larval feeding: development, form and function in a facultatively feeding larva, Brisaster latifrons. Evolution 50:174-187.
HARVEY, P. H., AND M. PAGEL. 1991. The comparative method in evolutionary biology. Oxford Univ. Press, Oxford.
HASZPRUNAR, G., L. V. SALVINI-PLAWEN, AND R. M. RIEGER. 1995. Larval planktotrophy - a primitive trait in the Bilateria? Acta Zool. 76:141-154.
HELMUTH, B., R. R. VEIT, AND R. HOLBERTON. 1994. Long-distance dispersal of a subantarctic brooding bivalve (Gaimardia trapesina) by kelp-rafting. Mar. Biol. 120:421-426.
HILLIS, D. M., AND J. J. BULL. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42:182-192.
HIMENO, H., M. HARUHIKO, T. KAWAI, T. OHTA, I. KUMAGAI, K-I MIURA, AND K. WATANABE. 1987. Unusual genetic codes and a novel gene structure for tRNAserAGY in starfish mitochondrial DNA. Gene 56:219-230.
HRINCEVICH, A. W., AND D. W. FOLTZ. 1996. Mitochondrial DNA sequence variation in a sea star (Leptasterias spp.) species complex. Mol. Phylogen. Evol. 6:408-415.
HUNT, A. 1993. Effects of contrasting patterns of larval dispersal on the genetic connectedness of local populations of two intertidal starfish Patiriella calcar and Patiriella exigua. Mar. Ecol. Prog. Ser. 92:179-186.
ISTOCK, C. A. 1967. The evolution of complex life cycle phenomena: an ecological perspective. Evolution 21:592-605.
JABLONSKI, D. 1986. Larval ecology and macroevolution in marine invertebrates. Bull. Mar. Sci. 39:565-587.
-----. 1987. Heritability at the species level: analysis of geographic ranges of Cretaceous mollusks. Science 238:360-363.
JACOBS, H. T., D. J. ELLIOTT, V. P. MATH, AND A. FARQUHARSON. 1988. Nucleotide sequence and gene organization of sea urchin mitochondrial DNA. J. Mol. Biol. 202:185-217.
JACOBS, H. T., S. ASAKAWA, T. ARAKI, K.-I. MIURA, M. J. SMITH, AND K. WATANABE. 1989. Conserved tRNA gene gluster in starfish mitochondrial DNA. Curr. Genet. 15:193-206.
JAGERSTEN, G. 1972. The evolution of the metazoan life cycle. Academic Press, New York.
JOHANNESSON, K. 1988. The paradox of Rockall: why is a brooding gastropod (Littorina saxatilis) more widespread than one having a planktonic larval dispersal stage (L. littorea)? Mar. Biol. 99: 507-513.
JORDAN, D. S. 1908. The law of geminate species. Am. Nat. 17: 73-80.
KEIGWIN, L. D. 1978. Pliocene closing of the Isthmus of Panama, based on biostratigraphic evidence from nearby Pacific Ocean and Caribbean Sea cores. Geology 6:630-634.
KEOUGH, M. J., AND A. J. DARTNALL. 1977. A new species of viviparous asterinid asteroid from Eyre Peninsula, South Australia. Rec. South Austr. Mus. 17:407-416.
KESSING, B. D. 1991. Strongylocentrotid sea urchin mitochondrial DNA: phylogenetic relationships and patterns of molecular evolution. Master's thesis, Univ. of Hawaii, Honolulu.
KLUGE, A. 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Syst. Zool. 38:7-25.
KNOWLTON, N., L. A. WEIGT, L. A. SOLARZANO, D. K. MILLS, AND E. BERMINGHAM. 1993. Divergence in proteins, mitochondrial DNA, and reproductive compatibility across the Isthmus of Panama. Science 260:1629-1632.
KOHN, A. J., AND F. E. PERRON. 1994. Life history and biogeography: patterns in Conus. Clarendon Press, Oxford.
KOMATSU, M., Y. T. KANO, AND C. OGURO. 1990. Development of a true ovoviviparous sea star, Asterina pseudoexigua pacifica Hayashi. Biol. Bull. 179:254-263.
KOZLOFF, E. N. 1987. Marine invertebrates of the Pacific northwest. Univ. of Washington Press, Seattle.
KUMAR, S., K. TAMURA, AND M. NEI. 1993. MEGA: molecular evolutionary genetics analysis. Vers. 1.02. Pennsylvania State Univ., University Park.
LACK, D. 1947. The significance of clutch size. Ibis 89:302-352.
LANDE, R. 1982. A quantitative genetic theory of life-history evolution. Ecology 63:607-615.
LAWSON-KERR, C., AND D. T. ANDERSON. 1978. Reproduction, spawning and development of the starfish Patiriella exigua (Lamarck) (Asteroidea, Asterinidae) and some comparisons with P. calcar (Lamarck). Austr. J. Mar. Fresh. Res. 29:45-53.
LESLIE, P. H. 1945. On the use of matrices in certain population mathematics. Biometrica 33:183-212.
LESSIOS, H. A. 1990. Adaptation and phylogeny as determinants of egg size in echinoderms from the two sides of the Isthmus of Panama. Am. Nat. 135:1-13.
-----. 1997. The first stage of speciation as seen in organisms separated by the Isthmus of Panama. In D. Howard and S. Berlocher, eds. Endless forms: species and speciation. Oxford Univ. Press, Oxford.
LEVIN, L. A., H. CASWELL, K. D. DEPATRA, AND E. L. CREED. 1987. Demographic consequences of larval development mode: planktotrophy vs. lecithotrophy in Streblospio benedicti. Ecology 68: 1877-1886.
LEVIN, L. A., J. ZHU, AND E. CREED. 1991. The genetic basis of life-history characters in a polychaete exhibiting planktotrophy and lecithotrophy. Evolution 45:380-397.
LIEBERMAN, B. S., W. D. ALLMON, AND N. ELDREDGE. 1993. Levels of selection and macroevolutionary patterns in the turritellid gastropods. Paleobiology 19:205-215.
LOCKHART, P. J., M. A. STEEL, M. D. HENDY, AND D. PENNY. 1994. Recovering evolutionary trees under a more realistic model of sequence evolution. Mol. Biol. Evol. 11:605-612.
MADDISON, W. P., AND D. R. MADDISON. 1992. MacClade: analysis of phylogeny and character evolution. Vers. 3.0. Sinauer, Sunderland, MA.
MADDISON, W. P., M. J. DONOGHUE, AND D. R. MADDISON. 1984. Outgroup analysis and parsimony. Syst. Zool. 33:83-103.
MATSUOKA, N. 1981. Phylogenetic relationships among five species of starfish of the genus Asterina; an electrophoretic study. Comp. Biochem. Physiol. 70B:739-743.
MCEDWARD, L. R., ED. 1995. Ecology of marine invertebrate larvae. CRC Press, Boca Raton, FL.
MCEDWARD, L. R., AND D. A. JANIES. 1993. Life cycle evolution in asteroids: what is a larva? Biol. Bull. 184:255-268.
MCMILLAN, W. O., R. A. RAFF, AND S. R. PALUMBI. 1992. Population genetic consequences of developmental evolution in sea urchins (genus Heliocidaris). Evolution 46:1299-1312.
MOCHIZUKI, Y., AND S. H. HORI. 1980. Immunological relationships of starfish hexokinases: phylogenetic implication. Comp. Biochem. Physiol. 65B: 119-125
NORMARK, B. B. 1996. Phylogeny and evolution of parthenogenetic weevils of the Aramigus tessellatus species complex (Coleoptera: Curculionidae: Naupactini): evidence from mitochondrial DNA sequences. Evolution 50:734-745.
O FOIGHIL, D., AND M. J. SMITH. 1995. Evolution of asexuality in the cosmopolitan marine clam Lasaea. Evolution 49:140-150.
OKAZAKI, K., AND K. DAN. 1954. The metamorphosis of partial larvae of Peronella japonica Mortensen, a sand dollar. Biol. Bull. 106:83-99.
OLSON, R. R., J. L. CAMERON, AND C. M. YOUNG. 1993. Larval development (with observations on spawning) of the pencil urchin Phyllacanthus imperialis: a new intermediate larval form? Biol. Bull. 185:77-85.
PALUMBI, S. R. 1996. What can genetics contribute to marine biogeography? An urchin's tale. J. Exp. Mar. Biol. Ecol. 203:75-92.
PALUMBI, S. R., AND A. C. WILSON. 1990. Mitochondrial DNA diversity in the sea urchins Strongylocentrotus purpuratus and S. droebachiensis. Evolution 44:403-415.
PERRON, F. E. 1981. Larval biology of six species of the genus Conus (Gastropoda: Toxoglossa) in Hawaii, USA. Mar. Biol. 61: 215-220.
-----. 1986. Life-history consequences of differences in developmental mode among gastropods in the genus Conus. Bull. Mar. Sci. 39:485-497.
POULIN, R. 1995. Clutch size and egg size in free-living and parasitic copepods: a comparative analysis. Evolution 49:325-336.
RAFF, R. A. 1987. Constraint, flexibility, and phylogenetic history in the evolution of direct development in sea urchins. Dev. Biol. 119:6-19.
REID, D. G. 1989. The comparative morphology, phylogeny and evolution of the gastropod family Littorinidae. Philos. Trans. R. Soc. Lond. B Biol. Sci. 324:1-110.
-----. 1990. A cladistic phylogeny of the genus Littorina (Gastropoda): implications for evolution of reproductive strategies and for classification. Hydrobiologia 193:1-19.
REID, D. G., E. RUMBAK, AND R. H. THOMAS. 1996. DNA, morphology and fossils: phylogeny and evolutionary rates of the gastropod genus Littorina. Philos. Trans. R. Soc. Lond. B Biol. Sci. 351:877-895.
RIDLEY, M. 1996. Evolution. 2d ed. Blackwell Science, Cambridge, MA.
ROFF, D. A. 1992. The evolution of life histories. Chapman and Hall, New York.
ROSE, M. R. 1985. Life-history evolution with antagonistic pleiotropy and overlapping generations. Theor. Popul. Biol. 28:342-358.
ROUSE, G., AND K. FITZHUGH. 1994. Broadcasting fables: is external fertilization really primitive? Sex, size, and larvae in sabellid polychaetes. Zool. Scrip. 23:271-312.
RUMBAK, E., D. G. REID, AND R. H. THOMAS. 1994. Reconstruction of phylogeny of 11 species of Littorina (Gastropoda: Littorinidae) using mitochondrial DNA sequence data. Nautilus 108:91-97.
SANDERSON, M. J. 1995. Objections to bootstrapping phylogenies: a critique. Syst. Biol. 44:299-320.
SMITH, M. J., D. K. BANFIELD, K. DOTEVAL, S. GORSKI, AND D. J. KOWBEL. 1990. Nucleotide sequence of nine protein-coding genes and 22 tRNAs in the mitochondrial DNA of the sea star Pisaster ochraceus. J. Mol. Evol. 31:195-204.
SPENCER, W. K., AND C. W. WRIGHT. 1966. Asterozoans. Pp. U4-U107 in R. C. Moore, ed. Treatise on invertebrate paleontology. Part U, Echinodermata 3. Univ. of Kansas Press, Lawrence.
STEARNS, S. C. 1992. The evolution of life histories. Oxford Univ. Press, Oxford.
STICKLE, W. B., D. W. FOLTZ, M. KATOH, AND H. L. NGUYEN. 1992. Genetic structure and mode of reproduction in five species of sea stars (Echinodermata: Asteroidea) from the Alaskan coast. Can. J. Zool. 70:1723-1728.
STRATHMANN, M. F. 1987. Reproduction and development of marine invertebrates of the northern Pacific coast. Univ. of Washington Press, Seattle.
STRATHMANN, R. R. 1978. The evolution and loss of feeding larval stages of marine invertebrates. Evolution 32:894-906.
-----. 1985. Feeding and nonfeeding larval development and lifehistory in marine invertebrates. Ann. Rev. Ecol. Syst. 16:339-361.
-----. 1988. Functional requirements and the evolution of developmental patterns. Pp. 55-61 in R. D. Burke, P. V. Mladenov, P. Lambert, and R. L. Parsley, eds. Echinoderm biology. A. A. Balkema, Rotterdam, The Nethrlands.
-----. 1990. Why life histories evolve differently in the sea. Am. Zool. 30:197-207.
STRATHMANN, R. R., AND D. J. EERNISSE. 1994. What molecular phylogenies tell us about the evolution of larval forms. Am. Zool. 34:502-512.
STRATHMANN, R. R., AND M. F. STRATHMANN. 1982. The relationship between adult size and brooding in marine invertebrates. Am. Nat. 119:91-101.
STRATHMANN, R. R., M. F. STRATHMANN, AND R. H. EMSON. 1984. Does limited brood capacity link adult size, brooding, and simultaneous hermaphroditism? A test with the starfish Asterina phylactica. Am. Nat. 123:796-818.
SWOFFORD, D. L. 1993. PAUP: phylogenetic analysis using parsimony. Vers. 3.1.1. Computer program distributed by the Illinois Natural History Survey, Champaign.
TAMURA, K., AND M. NEI. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10:512-526.
TATENO, Y., N. TAKEZAKI, AND M. NEI. 1994. Relative efficiencies of the maximum-likelihood, neighbor-joining, and maximumparsimony methods when substitution rate varies with site. Mol. Biol. Evol. 11:261-277.
THORSON, G. 1946. Reproduction and larval development of Danish marine bottom invertebrates. Medd. Komm. Hav. Ser. Plank. 4:1-523.
-----. 1950. Reproduction and larval ecology of marine bottom invertebrates. Biol. Rev. Cam. Philos. Soc. 25:1-45.
WADA, H., M. KOMATSU, AND N. SATOH. 1996. Mitochondrial rDNA phylogeny of the Asteroidea suggests the primitiveness of the Paxillosida. Mol. Phylogen. Evol. 6:97-106.
WERNER, E. E. 1986. Amphibian metamorphosis: growth rate, predation risk, and the optimal size at transformation. Am. Nat. 128:319-341.
WILBUR, H. M. 1980. Complex life cycles. Ann. Rev. Ecol. Syst. 11:67-94.
WRAY, G. A. 1995. Punctuated evolution of embryos. Science 267: 1115-1116.
-----. 1996. Parallel evolution of nonfeeding larvae in echinoids. Syst. Biol. 45:308-322.
WRAY, G. A., AND R. A. RAFF. 1991. Rapid evolution of gastrulation mechanisms in a sea urchin with lecithotrophic larvae. Evolution 45:1741-1750.
APPENDIX: NOTES ON TAXONOMY AND OUTGROUPS
The Asterinidae have usually been classified in the Order Spinulosida (Spencer and Wright 1966). Blake (1987) and Gale (1987) reassigned Asterinidae to the Order Valvatida on the basis of morphological characters. Gale favored the abolishment of the Spinulosida as a paraphyletic assemblage of "primitive and derived valvatid families." Some (Kozloff 1987) but not all (Brusca and Brusca 1990) recent texts have followed this taxonomic scheme.
Wada et al. (1996) used partial sequences from the two mitochondrial ribosomal RNA genes to examine relationships among some orders and families of starfish from Japan. Their results suggest a close relationship between four asterinid species (including Asterina pectinifera and A. pseudoexigua pacifica) and a "spinulosid" (Crossaster papposus, Family Solasteridae), to the exclusion of some other species reliably assigned to the Order Valvatida. Thus the assignment of the Asterinidae to the Order Valvatida may be incorrect.
If relationships between Asterinidae and other starfish families are uncertain, then the calibration of evolutionary rates based on divergence between Oreaster species may be inaccurate. Our results may not be biased by these possible complications, but we require a better phylogeny for starfish families and orders before the complications can be resolved.
|Printer friendly Cite/link Email Feedback|
|Author:||Hart, Michael W.; Byrne, Maria; Smith, Michael J.|
|Date:||Dec 1, 1997|
|Previous Article:||The maintenance of a cline in the marine snail Littorina saxatilis: the role of home site advantage and hybrid fitness.|
|Next Article:||The effects of winter length on the genetics of apple and hawthorne races of Rhagoletis pomonella (Diptera: tephritidae).|