Evolution, 60(11), 2006, pp. 2293–2310

EVOLUTION OF POECILOGONY FROM PLANKTOTROPHY: CRYPTIC SPECIATION, PHYLOGEOGRAPHY, AND LARVAL DEVELOPMENT IN THE GASTROPOD GENUS

RYAN A. ELLINGSON AND PATRICK J. KRUG1 Department of Biological Sciences, California State University, 5151 State University Drive, Los Angeles, California 90032-8201 1E-mail: [email protected]

Abstract. Poecilogony, a rare phenomenon in marine invertebrates, occurs when alternative larval morphs differing in dispersal potential or trophic mode are produced from a single genome. Because both poecilogony and cryptic are prevalent among sea slugs in the suborder (: ), molecular data are needed to confirm cases of variable development and to place them in a phylogenetic context. The nominal species produces long-lived, feeding larvae throughout the North Atlantic and Pacific, but in California can also produce short-lived larvae that metamorphose without feeding. We collected morphological, developmental, and molecular data for Alderia from 17 sites spanning the eastern and western Pacific and North Atlantic. Estuaries south of Bodega Harbor, California, contained a cryptic species (hereafter Alderia sp.) with variable development, sister to the strictly planktotrophic A. modesta. The smaller Alderia sp. seasonally toggled between planktotrophy and lecith- otrophy, with some individuals differing in development but sharing mitochondrial DNA haplotypes. The sibling species overlapped in Tomales Bay, California, but showed no evidence of hybridization; laboratory mating trials suggest postzygotic isolation has arisen. Intra- and interspecific divergence times were estimated using a molecular clock calibrated with geminate sacoglossans. Speciation occurred about 4.1 million years ago during a major marine radiation in the eastern Pacific, when large inland embayments in California may have isolated ancestral populations. Atlantic and Pacific A. modesta diverged about 1.7 million years ago, suggesting trans-Arctic gene flow was interrupted by Pleistocene glaciation. Both Alderia species showed evidence of late Pleistocene population expansion, but the southern Alderia sp. likely experienced a more pronounced bottleneck. Reduced body size may have incurred selection against obligate planktotrophy in Alderia sp. by limiting fecundity in the face of high larval mortality rates in warm months. Alternatively, poecilogony may be an adaptive response to seasonal opening of estuaries, facilitating dispersal by long-lived larvae. An improved understanding of the forces controlling seasonal shifts in development in Alderia sp. may yield insight into the evolutionary forces promoting transitions to nonfeeding larvae.

Key words. Cryptic species, larval development, lecithotrophy, phylogeography, planktotrophy, poecilogony, sacog- lossan, trans-Arctic exchange.

Received March 9, 2006. Accepted July 20, 2006.

Marine ecosystems have fewer morphologically distin- while increasing local adaptation and the likelihood of spe- guishable species than terrestrial habitats, and many marine ciation events (Hansen 1978; Vermeij 1982; Jablonski 1986; species have cosmopolitan distributions (May 1994). This Palumbi 1995; Sanford et al. 2003). Understanding transi- may result from the considerable dispersal ability of - tions from feeding to nonfeeding modes of larval develop- dwelling organisms, since pelagic adults and planktonic lar- ment is an outstanding goal of evolutionary biology (Strath- vae face few geographical obstacles to large-scale transport mann 1978; Raff 1987; Hadfield et al. 1995; Havenhand (Strathmann 1990; Palumbi 1992; Pechenik 1999). However, 1995; Hart et al. 1997; Duda and Palumbi 1999; Hart 2000; molecular studies of marine organisms frequently uncover McEdward 2000; Jeffery et al. 2003). However, the selective subdivided populations and cryptic species, suggesting ge- forces that drive the evolution of nonfeeding development netic divergence may commonly occur at small spatial scales remain opaque, partly for lack of robust phylogenies paired or with minimal morphological diversification (Palmer et al. with relevant ecological and developmental data for the spe- 1990; Palumbi 1992; Knowlton 1993, 2000; Lee 2000; Daw- cies involved (Havenhand 1995; Wray 1995; McEdward son and Jacobs 2001; McGovern and Hellberg 2003; Lee and 2000). Small body size has been hypothesized to favor brood- O´ Foighil 2004; Tarjuelo et al. 2004; Collin 2005; Meyer et ing or direct development (Strathmann and Strathmann 1982; al. 2005). Where biological lines are sharply drawn in what Byrne and Cerra 1996), but there is little theory accounting appears to be an easily traversed oceanic milieu, we may gain for why pelagic lecithotrophy should be more adaptive than insight into the forces that promote divergence, maintain planktotrophy (Chia 1974). boundaries between species, and drive the adaptive evolution Insight into life-history evolution may come from organ- of marine life histories. isms that exhibit poecilogony, a stable reproductive poly- Development mode strongly affects patterns of phylo- morphism in which larvae of a single species can develop geography and macroevolution in marine invertebrates by alternative pathways (Giard 1905; Chia et al. 1996). Mo- (Arndt and Smith 1998; Collin 2000; Jeffery and Emlet lecular data often reveal that putative cases of poecilogony 2003). Planktotrophic species produce feeding larvae with are cryptic species differing in development, however (Hoag- high dispersal potential, resulting in panmictic populations land and Robertson 1988; Bouchet 1989). Convincing ex- and extended ranges (Caley et al. 1996; Todd 1998; Bohonak amples of variable development are limited to polychaete 1999; Pechenik 1999). An abbreviated planktonic period worms in the family Spionidae (Levin 1984; Gibson et al. tends to decrease a species’ range and geological longevity 1999; Morgan et al. 1999; Schultze et al. 2000; Gibson and 2293 ᭧ 2006 The Society for the Study of Evolution. All rights reserved. 2294 R. A. ELLINGSON AND P. J. KRUG

TABLE 1. Populations of Alderia modesta sampled for development mode, morphology, and genetic analyses. Slugs from Russia, Belgium, Vancouver, and Humbolt were obtained as ethanol-preserved specimens. Unless otherwise noted, all populations are in California.

Location Latitude, longitude Date sampled Doel, Belgium (DB) 51Њ19ЈN, 4Њ16ЈE Aug. 2005 Amurskii Bay, Russia (RS) 43Њ09ЈN, 131Њ54ЈE June 2004 Kodiak, Alaska (AK) 57Њ35Ј50ЉN, 152Њ28Ј17ЉW June 2005 Vancouver, British Columbia, Canada (BC) 49Њ18Ј21ЉN, 123Њ05Ј32ЉW July 2003 Tillamook, Oregon (TI) 45Њ26Ј55ЉN, 123Њ50Ј57ЉW March 2005 Coos Bay, Oregon (CB) 43Њ21Ј48ЉN, 124Њ11Ј45ЉW July 2000, October 2001, February 2005 Humboldt (HU) 40Њ44Ј31ЉN, 124Њ12Ј59ЉW February 2004 Bodega Harbor (BH) 38Њ18Ј58ЉN, 123Њ03Ј24ЉW September 2003, September 2004 Walker Creek (WC) 38Њ13Ј46ЉN, 122Њ54Ј56ЉW September 2004 Cow Landing (CL) 38Њ11Ј02ЉN, 122Њ54Ј40ЉW September 2004 South Tomales Bay (TB) 38Њ06Ј53ЉN, 122Њ51Ј08ЉW September 2003, September 2004 San Francisco Bay (SF) 37Њ52Ј57ЉN, 122Њ31Ј09ЉW September 2003–August 2004 (5ϫ) Morro Bay (MB) 35Њ20Ј39ЉN, 120Њ50Ј35ЉW September 2002 Santa Barbara (SB) 34Њ24Ј01ЉN, 119Њ32Ј07ЉW July 1999, February 2000 Los Angeles Harbor (LA) 33Њ42Ј48ЉN, 118Њ17Ј02ЉW January 2004–August 2004 (4ϫ) Newport Bay (NB) 33Њ37Ј16ЉN, 117Њ53Ј32ЉW June 2003–August 2004 (6ϫ) San Diego (SD) 32Њ47Ј36ЉN, 117Њ13Ј47ЉW monthly 1996–1998; July 1999–August 2004 (4ϫ)

Gibson 2004) and herbivorous sea slugs in the gastropod ther planktotrophic larvae with a month-long maturation pe- suborder Sacoglossa (West et al. 1984; Clark 1994; Krug riod, or lecithotrophic larvae competent to settle upon or 1998). The adaptive value of expressing multiple dispersal before hatching (Krug 1998, 2001). In this life-history trade- strategies from one genotype is well documented for adult off, lecithotrophy limits adult fecundity by an order of mag- morphs of terrestrial insects (Denno et al. 1980; Harrison nitude and doubles benthic development time, but greatly 1980; Langellotto and Denno 2001); it is therefore perplexing reduces the planktonic period and hence the mortality rate that poecilogony is so rare among marine invertebrates. of larvae. Although prior evidence indicated this was a case Equally puzzling are the phylogenetic constraints that may of variable development, the northern limit of poecilogony limit this phenomenon to polychaetes and opisthobranch mol- was not established, nor was the evolutionary relationship luscs. between strictly planktotrophic and southern, poecilogonous Sacoglossan sea slugs are suctorial feeders that specialize populations delineated. on coenocytic or large-celled, filamentous algae (Jensen The geographically restricted expression of poecilogony in 1983, 1996; Trowbridge 2002). Adults feed, mate, and de- an otherwise planktotrophic taxon provided the opportunity posit eggs on their algal host and are generally unable to to explore an evolutionary origin of variable development. switch hosts after settlement (Jensen 1989; Trowbridge 1991; We used molecular, developmental, and morphological char- Trowbridge and Todd 2000). Furthermore, planktonic larvae acters to assess the global phylogeography of the genus Al- preferentially metamorphose in response to host-specific deria. Sequences from two mitochondrial loci were used to chemical cues (West et al. 1984; Krug and Manzi 1999; Trow- test for genetic breaks between strictly planktotrophic versus bridge and Todd 2000; Krug 2001). Host acceptance via se- poecilogonous populations, and to examine population struc- lective metamorphosis may thus determine both assortative ture in the eastern Pacific. A molecular clock was calibrated mating and ecological specialization in these marine herbi- with geminate sacoglossans to date major divergence events vores, as in some insect host races (Feder et al. 1994; Caillaud between and within lineages. Our data support a model of and Via 2000). In addition to their host specificity, sacog- cryptic speciation via peripheral isolation, accompanied by lossans display a high degree of reproductive plasticity, with a rare life-history shift to environmentally cued changes in many independent transitions to nonfeeding larval develop- development mode. We discuss features of adult and larval ment and several documented cases of poecilogony (West et ecology that may have favored seasonal lecithotrophy in al. 1984; Clark 1994; Jensen 1996; Krug 1998). Such de- southern estuaries. velopmental flexibility may facilitate studies of the selective pressures behind, and consequences of, alternative larval MATERIALS AND METHODS morphs (Krug and Zimmer 2004). Study System and Population Sampling The sacoglossan Alderia modesta is found in temperate estuaries throughout the Northern Hemisphere with its ob- The genus Alderia was described from Ireland (Thompson ligate hosts, yellow-green algae in the genus Vaucheria (Har- 1844; Allman 1845), and the sole recognized species, A. mo- tog 1959; Trowbridge 1993). Atlantic populations of the desta, from Sweden (Love´n 1844). Specimens superficially monotypic genus Alderia are planktotrophic, as are Pacific resembling A. modesta were collected between 1999 and 2005 populations from Monterey, California, north to Canada, and from 15 sites along the northeastern Pacific coast of the Unit- from Russia (Hand and Steinberg 1955; Hartog 1959; See- ed States and Canada (Table 1). Slugs were collected at low lemann 1967; Gibson and Chia 1995; Chernyshev and Cha- tide from exposed patches of Vaucheria species, examined ban 2005). However, the nominal taxon A. modesta exhibits morphologically under a dissecting microscope, and placed poecilogony in southern California, where adults produce ei- individually in petri dishes of seawater for egg mass depo- EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2295 sition. Development mode was typed one to two days later Alderia from the eastern Pacific (n ϭ 233), western Pacific according to egg size within the first deposited clutch (Krug (n ϭ 3), and Atlantic (n ϭ 7). Suitable sequence data was 1998). Typed adults were frozen at Ϫ80ЊC or stored in 95% obtained from all specimens for 480 bp of COI, correspond- ethanol prior to DNA extraction. Specimens were collected ing to positions 139 to 618 of the complete Aplysia californica and preserved from the western Pacific in Amurskii Bay, Sea COI gene (GenBank accession number AY569552), and of Japan, Russia, in June 2004, and from the Atlantic in the aligned with SeqScape 2.1.1 software (Applied Biosystems). Scheldt estuary, Doel, Belgium, in August 2005 by col- A 450-bp fragment of the 16S gene was sequenced from a leagues (see Acknowledgments). subset of specimens (n ϭ 49), resulting in 12 haplotypes from No closely related genus exists in the temperate eastern A. modesta (six each from the Atlantic and Pacific), and eight Pacific, and the phylogenetic position of Alderia is contro- haplotypes from Alderia sp. (see Results). All 16S sequences versial (Evans 1953; Gascoigne 1976; Jensen 1996). The At- were aligned in ClustalX 1.83 (Thompson et al. 1997), with lantic genus Limapontia was united with Alderia in the family adjustments made by eye using three models of gastropod based on distinctively sabot-shaped radular 16S secondary structure (Lydeard et al. 2000; Medina and teeth (Jensen 1996), and was closely related to Alderia in a Walsh 2000; Valde´s 2003). Amino acid substitutions, mean preliminary molecular phylogeny of the (R. base composition, and transition/transversion (s/v) ratio were A. Ellingson and P. J. Krug, unpubl. data). For outgroup determined in Mega 3.0 (Kumar et al. 2004). comparisons, L. depressa was therefore collected from Gal- A neighbor-joining (NJ) tree of all eastern Pacific COI way Bay, Republic of Ireland (53Њ13ЈN, 08Њ54ЈE) in August haplotypes was constructed using PAUP* 4.0b10 (Swofford 2005. 2001). For NJ analysis, the Tamura-Nei model of sequence evolution was used (Tamura and Nei 1993), and robustness DNA Extraction, Amplification, and Sequencing of the topology was assessed with 2000 bootstrap replicates. Subsequent analyses included Atlantic and western Pacific Genomic DNA was extracted from whole frozen slugs us- specimens, and used both COI and 16S sequence data from ing the QIAamp DNA Mini Kit (Qiagen, Inc., Valencia, CA) 49 individuals to determine (1) interspecific relationships and and stored in extraction buffer at Ϫ20ЊC prior to amplifica- (2) intraspecific divergence between ocean basins. A parti- tion. Polymerase chain reactions (PCR) were used to amplify a 710-base pair (bp) fragment of the mitochondrial cyto- tion-homogeneity test as implemented in PAUP* indicated chrome oxidase subunit I (COI) gene, using primers no significant conflict between COI and 16S datasets; we LCO1490 and HCO2198 (Folmer et al. 1994), and a 480-bp therefore performed Bayesian and parsimony analyses on the fragment of the mitochondrial 16S rRNA (large ribosomal combined data (Cunningham 1997; Wiens 1998). For Bayes- subunit) gene using primers 16Sar-5Ј and 16Sbr-3Ј (Palumbi ian analysis, COI and 16S datasets were partitioned so that 1996). A negative control (no template) was included in each a separate substitution model could be applied to each. The run. Amplification used a reaction volume of 50 ␮l containing best-fit model for the COI dataset selected using the Akaike 1ϫ Promega (Madion, WI) PCR buffer (10 mM Tris-HCl pH information criterion as implemented in Modeltest 3.7 (Po- sada and Crandall 1998) was TrN ϩ⌫, with gamma distri- 9.0, 50 mM KCl, 0.1% Triton X-100), 2.5 mM MgCl2, 0.2 mM dNTPs, 0.2 ␮M of each primer, 1 U Promega Taq DNA bution shape parameter of 0.1397. For 16S, the best-fit model ϩ ϩ⌫ Polymerase and 50–200 ng of template DNA. To amplify the was HKY I , with gamma shape 0.2932. Bayesian COI fragment, the following thermal cycler profile was used: analysis used the metropolis-coupling Markov chain Monte denaturation at 94ЊC for 2 min followed by 35 cycles of (94ЊC Carlo method as implemented in MrBayes 3.1.1 (Ronquist for 30 sec, 40ЊC for 45 sec, and 72ЊC for 60 sec). For 16S, and Huelsenbeck 2003). The analysis was run for 3,000,000 the thermal cycler profile began with denaturation at 94ЊC generations with a tree saved every 1000 generations; a con- for 2 min followed by 40 cycles of (94ЊC for 30 sec, 50ЊC sensus tree was generated in PAUP* after a burn-in period for 30 sec, and 72ЊC for 60 sec), followed by an extension of 300. For parsimony analysis, the maximum number of trees step at 72ЊC for 7 min. saved was limited to 5000 due to computational time con- Polymerase chain reaction products were visualized by straints, and bootstrap values (1000 replicates) were obtained electrophoresis on a 1% agarose gel, and were purified with using the ‘‘fast’’ stepwise-addition option. the Wizard SV Gel and PCR Clean-Up System (Promega). Both strands of purified products were directly cycle-se- Within-Clade Genetic Structure and Expansion Estimates quenced in both directions using the amplification primers Genetic structures of the two major COI clades from the and Big Dye Terminator 3.1 Cycle Sequencing chemistry, eastern Pacific were characterized in Arlequin 2.000 (Schnei- and electrophoresed on an ABI 3100 Avant Capillary Se- der et al. 2000). Haplotype diversity (H), nucleotide diversity quencer (Applied Biosystems, Foster City, CA). Sequencing (␲), and mismatch distributions were calculated for each both strands and resequencing of unique haplotypes revealed clade, hereafter referred to as A. modesta and Alderia sp. (see no evidence of PCR error. Sequences used for this study have Results). The proportion of genetic variation partitioned been deposited in GenBank (accession numbers DQ364252– among estuaries was estimated by analysis of molecular var- DQ364426). iance (AMOVA; Excoffier et al. 1992), using both Slatkin’s (1991, 1995) linearized F and ⌽ values. Sites within Phylogenetic Analyses ST ST Tomales Bay (south Tomales Bay, Cow Landing, and Walker To determine phylogenetic relationships and population Creek) were not significantly different in preliminary AMO- structure, COI sequences were amplified from specimens of VA analysis and, being located in the same estuary, were 2296 R. A. ELLINGSON AND P. J. KRUG

grouped together for subsequent analyses. Based on Model- thus determined for the first codon position of COI, for which test results, the best-fit model of sequence evolution for each log-likelihood ratio tests could not reject molecular clock ⌽ ␲ clade was used to calculate ST distance matrices and (A. assumptions. Based on Modeltest, the TrNef (Tamura-Nei, modesta, Tamura-Nei; Alderia sp., Tamura three-parameter) equal base frequencies) model of sequence evolution was (Tamura 1992; Tamura and Nei 1993). Tajima’s D and Fu’s used to calculate first position distances (Tamura and Nei FS were calculated by comparing observed values against a 1993; Huelsenbeck and Crandall 1997; Posada and Crandall null distribution based on 1000 simulations to determine sig- 1998). To correct for intraclade variance, net nucleotide di- ␳ ϭ␳ Ϫ ␳ ϩ␳ nificance levels for each test statistic. Tajima’s D compares vergences were calculated as AB(net) AB 0.5( A B), ␳ ␳ ␳ the number of segregating sites with the number of pairwise where A and B are mean intraclade distances, and AB is differences in a population, and should be near zero for an the mean interclade distance (Nei and Li 1979; Edwards and equilibrium population; negative values indicate a selective Beerli 2000). The rate determined with E. crispata–E. di- sweep or population expansion. Large negative values of Fu’s omedea was used to estimate divergence times for (1) A. modesta versus Alderia sp., and (2) Pacific versus Atlantic FS suggest population expansion (Fu 1997; Excoffier and Schneider 1999). populations of A. modesta. Parameters ␶ (time since expansion, expressed in units of ␪ ␪ Copulatory Behavior and Reproductive Isolation mutation rate), 0 (population size before expansion), and 1 (population size after expansion) were estimated from mis- Reproductive isolation between slugs was assessed by match distributions using a generalized nonlinear least- monitoring mating behavior and egg mass production be- squares approach (Schneider and Excoffier 1999). Confidence tween pairs in the laboratory. Because field-caught Alderia intervals were determined by a parametric bootstrap proce- store sperm for weeks, virgin slugs were reared by individ- dure, assuming a sudden expansion model. The validity of ually metamorphosing lecithotrophic larvae of Alderia sp. the model was tested using the sum of square deviations from San Diego parents on filaments of Vaucheria longicau- (SSD) between observed and expected mismatch distribu- lis. Virgin slugs were raised in isolation for three to four tions as a test statistic. The SSD distribution was generated weeks, by which time all were reproductively mature; control via coalescent simulations of expansion using the estimated slugs raised in pairs began reproducing within two weeks of ␶ ␪ ␪ parameters , 0 and 1; the P-value was the proportion of metamorphosis. Virgin A. modesta were not available, due times the difference between simulated and expected mis- to difficulties in culturing their long-lived planktotrophic lar- match distribution exceeded the observed SSD (Schneider vae. Laboratory-reared virgins (n ϭ 7) were paired with in- and Excoffier 1999). To express expansion time in years, dividual specimens of the northern clade from Bodega Har- values of ␶ were multiplied by the mutational time scale bor, and observed under a dissecting microscope over 6 h. (generation time/2u, where u is the number of substitutions The members of each pair were then isolated and monitored per year for the fragment of interest) (Rogers 1995; Excoffier for egg production for 48 h. Controls were paired virgin and Schneider 1999). The substitution rate per nucleotide site Alderia sp. put together at the same time as the northern- was calculated from our COI data, and then multiplied by southern pairs. the length of the COI fragment used in mismatch analyses, Size and Fecundity Measurements yielding a rate of 2.62 ϫ 10Ϫ5 substitutions/yr. Based on larval period and laboratory rearing data, we estimated gen- For wet weight determinations, freshly collected slugs eration times (egg-to-egg period) of two months for A. mo- were first blotted dry with a paper towel and weighed to the desta, versus one month for Alderia sp. (averaging the gen- nearest 0.1 mg, then returned to individual dishes for egg eration time from a lecithotrophic larva, 2.5 weeks, and a laying. Size differences between northern and southern spe- planktotrophic larva, 5.5 weeks). cies (see Results), and between sites within each species, were tested with a nested ANOVA using site as a random effect nested within species as a fixed effect (Sokal and Rohlf Divergence Estimates 1995). For fecundity measurements, the first egg mass de- We calibrated a molecular clock using the only well-es- posited by each preweighed adult was maintained in a petri tablished geminate sacoglossans. Specimens of the Caribbean dish with 4 ml of 0.22 ␮m filtered seawater until hatching Elysia (ϭ Tridachia) crispata were collected December 2004 (three days for planktotrophic, six days for lecithotrophic from Bocas del Toro, Panama, with permission of the Pan- eggs), and larvae per clutch were counted under a dissecting amanian government; preserved specimens of E. (ϭ Trida- microscope. Relationship between size and fecundity was chiella) diomedea from the eastern Pacific were obtained from determined by regressing egg number for the first clutch de- the Los Angeles County Museum of Natural History. Any posited after collection on adult weight (Krug 2001). Re- geminates may have been isolated prior to final closure of gressions were run for three populations of A. modesta and the Panamanian seaway (Knowlton and Weigt 1998; Marko for two populations of Alderia sp. (one for each larval morph). 2002). However, E. crispata and E. diomedea live in shallow Mean fecundity was measured for an additional two popu- A. modesta Alderia mangrove fringe, probably the last habitat sundered by the lations of and five populations of sp. Isthmus; we therefore used a minimum divergence time of RESULTS 3.1 million years ago (MYA) for rate calibration (Coates and Obando 1996). Plots of s/v ratio against genetic distance Cryptic Species and Variable Development in Alderia indicated third positions of COI were saturating at distances In a survey of estuaries along the northeastern Pacific greater than 15% (mean ratio ϭ 1.5); substitution rate was coastline, specimens of A. modesta from Bodega Harbor, Cal- EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2297

FIG. 1. Seasonal distribution of development modes exhibited by sea slugs in the genus Alderia along the northeastern Pacific coast. Data are the proportion of adult slugs at each site that produced planktotrophic (black) or lecithotrophic (white) larvae, with sample size given on each pie graph. Sampling dates are in Table 1; summer sampling was conducted from June through September, winter sampling from November through February. ifornia, northward laid only planktotrophic eggs; in contrast, Both morphotypes were present within Tomales Bay. In south of Bodega, slugs produced mostly lecithotrophic planktotrophic egg masses from northern slugs, eggs were clutches in summer and both development modes in winter coiled into a spiral, a plesiomorphic character in the super- (Fig. 1). Concordant morphological differences were evident: family Limapontioidea. In contrast, both planktotrophic and slugs from Bodega north had a smooth yellow dorsum dotted lecithotrophic clutches of southern slugs lacked any spiral, with brown speckles, whereas south of Tomales Bay, most with eggs haphazardly arranged inside the egg mass (Krug had a dark background color with a raised dorsal hump, split et al. 2007). down the midline by a band of yellow (Krug et al. 2007). To determine whether there were genetic differences across 2298 R. A. ELLINGSON AND P. J. KRUG the Bodega Harbor breakpoint, a portion of the mitochondrial sp., no among-estuary comparisons were significant based on ⌽ COI gene was sequenced from slugs previously scored for ST. No population structure was evident in Pacific A. mo- ⌽ development mode of their larvae. The mean percent base desta by either FST or ST. composition was 37T:23A:21C:19G. This fragment was We subsequently included specimens of A. modesta from highly polymorphic, yielding 146 unique haplotypes with 155 the Atlantic and western Pacific in a broader phylogeographic variable positions (21 first, 2 second, 132 third codon posi- analysis. In addition to COI, a 450-bp fragment of the more tion). Nearly all substitutions were synonymous, excepting conserved 16S gene was sequenced from a subset of speci- four haplotypes that had one conservative amino acid sub- mens. In both Bayesian and parsimony trees based on com- stitution compared to all others, each at a different site. bined COI and 16S sequence data, A. modesta from the North A neighbor-joining (NJ) tree of eastern Pacific COI hap- Pacific and Atlantic formed a clade that was reciprocally lotypes showed a genetic divide between populations north monophyletic with Alderia sp. (Fig. 4). The sister species of and south of Bodega Harbor, concordant with developmental Alderia were a mean 20.6% divergent in COI (Tamura-Nei and morphological differences (Fig. 2). The COI haplotypes distance) and 3.1% in 16S (uncorrected distance). Within A. formed two reciprocally monophyletic clades with 100% modesta, Pacific and Atlantic populations were also recip- bootstrap support. Intraclade divergence was a maximum of rocally monophyletic and substantially divergent (11.0% in 5% but the two clades were 18–24% different from each other COI, 1.3% in 16S). Three haplotypes from the western Pacific (Tamura-Nei distance), exceeding the interspecific diver- (Sea of Japan) fell within the larger A. modesta clade, but gence between many sacoglossans (R. A. Ellingson and P. did not group together. Within Alderia sp., bootstrap values J. Krug, unpubl. data). Further, there were fixed differences and posterior probabilities supported a clade composed pri- between clades at 27 positions (Krug et al. 2007). Based on marily of haplotypes from southern California (Fig. 4). differences in morphology, development, and DNA sequence A geminate pair of elysiid sacoglossans yielded a rate of data, the southern clade represents a cryptic species hereafter 1.57% change per million years for the first codon position referred to as Alderia sp. of COI. Based on this rate, Alderia species diverged about Within Alderia sp. there was no concordance between de- 4.1 MYA, in the Miocene, whereas Atlantic and Pacific pop- velopment and tree topology, and 12 different haplotypes ulations of A. modesta diverged about 1.7 MYA, in the early were shared by specimens that produced different types of Pleistocene. larvae (Fig. 2). The southern Alderia sp. is thus poecilogon- ous, expressing alternative larval developmental modes with- Overlap Zone in a single evolutionarily independent lineage. There were To refine the range limits of Alderia spp., five sites were two weakly supported intraspecific clades that did not cor- sampled between San Francisco Bay and Bodega Harbor in relate with geographic origin of haplotypes. Northern pop- 2003–2004 (Fig. 5). Individuals were typed by morphology, ulations were morphologically consistent with the original development, and mitochondrial DNA analysis. In both sum- description of A. modesta, which is planktotrophic in all mers, only A. modesta was found in Bodega Harbor, and the known Atlantic and Pacific populations. southern end of Tomales Bay contained only Alderia sp. The The more broadly distributed A. modesta had higher hap- two species co-existed in central (Cow Landing site) and lotype and nucleotide diversities and more segregating sites northern (Walker Creek site) Tomales Bay (Fig. 5). A few than its geographically restricted congener (Table 2). The A. modesta were found in San Francisco Bay in December northern species A. modesta also had a significantly higher 2003 and August 2004. Genetic analysis confirmed that all mean number of differences between haplotypes (Fig. 3 and slugs (n ϭ 80) from mixed populations were correctly typed Ͻ results of a permutation test, P 0.0005). In contrast, Alderia by morphology (haplotypes included in Fig. 2). There was sp. had fewer pairwise differences and a common haplotype no evidence of mitochondrial DNA introgression in either shared throughout its range (Fig. 3). Tajima’s D and Fu’s FS direction where the two species coexisted. The host alga at were significantly negative for both species, suggesting re- the south Tomales Bay site (nominally V. longicaulis) was cent expansion (Table 3), and a model of sudden population similar in appearance to Vaucheria from southern California, expansion could not be rejected for either species (A. modesta, forming dense, discrete patches with a mosslike texture. The ϭ ϭ ␶ ␪ P 0.89; Alderia sp., P 0.95). Parameters and 0, es- algae in San Francisco, the rest of Tomales Bay, and Bodega timated from mismatch distributions, suggested A. modesta Harbor had a more filamentous morphology not typically began expanding at an earlier date and from a population of observed in southern California. larger size (Table 3). The mutational time scale calculated for each species placed these expansions at the end of the Copulatory Behavior and Reproductive Isolation Pleistocene. To assess reproductive isolation between the sibling spe- Alderia n ϭ Phylogeography of Alderia spp cies, laboratory-reared virgin sp. ( 7) were paired with individual A. modesta from Bodega and observed for 6 Based on FST values, a significant percentage of the total h. All Bodega slugs reacted to contact with a virgin Alderia genetic variance was distributed among estuaries in Alderia sp. by immediately extending the penis; at least three were Ͻ sp. (AMOVA: P 0.05). All significant pairwise FST com- observed to inseminate a virgin, based on flow of sperm parisons involved Morro Bay, the most geographically iso- through the translucent penis. All virgin slugs extended their lated site (Table 4). Only Morro Bay contained multiple pri- penis upon contact with a Bodega slug and one aberrantly vate haplotypes shared by four or more individuals. In Alderia discharged sperm into the seawater, but none were observed EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2299

FIG. 2. Phylogenetic relationships among slugs in the genus Alderia from the northeastern Pacific, based on COI sequences. A neighbor- joining (NJ) tree was derived from all nucleotide sites; branches with bootstrap support Ն70 are bold. Unmarked haplotypes were obtained from adults producing planktotrophic larvae, those labeled with a dark circle from adults producing lecithotrophic larvae, and those marked with a triangle were shared by adults that produced larvae differing in development mode. 2300 R. A. ELLINGSON AND P. J. KRUG

TABLE 2. Molecular diversity in Alderia species from the northeastern Pacific. Sequence data for COI were collected for A. modesta and its cryptic congener, Alderia sp., and analyzed in Arlequin 2.000. ‘‘Unique’’ haplotypes were sampled only once in the study. Haplotype and nucleotide diversities are given Ϯ variance.

A. modesta Alderia sp. n (no. of individuals) 104 129 k (no. of haplotypes) 84 62 % unique haplotypes 94.0% 74.2% No. of segregating sites (S) 103 72 Haplotype diversity (H) 0.9860 ϩ 0.0062 0.9119 ϩ 0.0217 Nucleotide diversity (␲) 0.0209 ϩ 0.0107 0.0141 ϩ 0.0074 Mean no. of pairwise differences (95% CI) 9.68 (5.86–12.29) 6.66 (3.29–10.28) to inseminate their partner even after prolonged contact (Ͼ10 cies can be differentiated by fixed nucleotide differences in min). Only one of seven inseminated virgins produced egg the COI gene, allozymes at the PGI locus, and dorsal mor- masses over the next 48 h, whereas all paired control virgins phology (Krug et al. 2007). Whereas A. modesta is exclu- reciprocally inseminated each other and produced fertilized sively planktotrophic, the poecilogonous Alderia sp. express- clutches. One former virgin produced two fertilized clutches es either planktotrophy or lecithotrophy depending on the that developed normally; however, it was subsequently found season; specimens differing in development cannot be dis- that Alderia sp. can self-fertilize, leaving the parentage of tinguished by any other criteria, share COI haplotypes, and offspring from the virgin slug ambiguous (N. Smolensky and switch development mode under laboratory conditions (Krug P. J. Krug, unpubl. data). 1998, 2007). The Californian endemic Alderia sp., likely evolved at the Size and Fecundity southeastern range limit of a broadly distributed ancestor. An Overall, specimens of A. modesta were significantly larger emerging paradigm for macroevolution of intertidal gastro- than Alderia sp. (Fig. 6, Table 5). This size difference initially pods is speciation via peripheral isolation. Intertidal organ- drew attention to rare individuals of A. modesta in San Fran- isms are effectively confined to one-dimensional strips of cisco Bay. Within each species, there was highly significant shoreline, and may be susceptible to transient isolation variation in size between sites (Table 5). Specimens from caused by climate change (Valentine and Jablonski 1983; Coos Bay, Oregon were larger than slugs from all Californian Reid 1990). For instance, Pleistocene conditions forced many sites (Fig. 6, and results of a post-hoc Scheffe´ test: P Ͻ 0.05). taxa into refugia in the Southern California Bight, followed At Cow Landing, specimens of A. modesta were significantly by Holocene expansion northward (Hellberg 1994; Marko larger than sympatric Alderia sp. (post-hoc Scheffe´ test: P Ͻ 1998; Dawson et al. 2001; Hellberg et al. 2001; Marko et al. 0.05). Size did not consistently discriminate between species, 2003; Jacobs et al. 2004). However, in no previous case has however; A. modesta from Walker Creek 2004 were not sig- a major life-history shift been potentially linked to transient nificantly larger than Alderia sp. from that or any other site isolation at the edge of a species’ distribution in the eastern (Fig. 6). Interannual differences were found at San Francisco, Pacific. As lecithotrophy is expressed by the southern species where the mean size in Aug 2004 was significantly larger primarily in summer months, it is likely an adaptation to than in Aug 2003 (post-hoc Scheffe´ test: P Ͻ 0.05). Within warm conditions. The data for Alderia sp. are consistent with Alderia sp., there was no increase in size with latitude, and speciation by peripheral isolation from a planktotrophic an- the smallest specimens came from south Tomales Bay. cestor. Lecithotrophy is an apomorphy of the southern spe- Fecundity (eggs per clutch) was positively related to size cies, as planktotrophy is pleisiomorphic in the Limapontiidae in both species and, in Alderia sp., for both development (Jensen 1996). Whereas A. modesta retained the ancestral modes (Table 6). The mean fecundity of A. modesta ranged development mode, Alderia sp. incorporated an extraordinary from 267 Ϯ 34 to 702 Ϯ 98 planktotrophic eggs per clutch degree of flexibility into its life history, seasonally expressing (n ϭ 5 populations), whereas specimens of Alderia sp. pro- alternative larval morphs and bet-hedging dispersal strategies duced 166 Ϯ 15 and 311 Ϯ 25 planktotrophic eggs per clutch (Krug 2001). in two populations. The mean number of lecithotrophic eggs In the absence of paleontological data for shell-less mol- per clutch ranged from 32 Ϯ 2to325Ϯ 71 (n ϭ 5 popu- luscs, molecular methods may permit reconstruction of their lations), but the maximum value was atypically high; four of evolutionary history. A molecular clock, calibrated with gem- the five populations had a mean fecundity of less than 120 inate sacoglossans, suggested a coalescence time of 4.1 MYA lecithotrophic eggs per clutch. for A. modesta–Alderia sp. In the late Miocene and early Pliocene, coastal geomorphology and high productivity in DISCUSSION the eastern Pacific contributed to radiations of many marine taxa (Jacobs et al. 2004). High sea level also created isolated Cryptic Speciation and Phylogeography in Alderia estuaries in the Central Valley of California, providing op- Based on molecular, morphological, and developmental portunities for allopatric speciation among wetland organ- differences, most specimens previously regarded as A. mo- isms such as Alderia (Hall 2002; Jacobs et al. 2004). desta south of Bodega Harbor, California, comprised a cryptic Pleistocene climate change likely severed trans-Arctic species (Dayrat 2005; DeSalle et al. 2005). The sibling spe- gene flow in A. modesta. The molluscan fauna of the North EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2301

FIG. 3. Histogram showing the distribution of pairwise differences between COI haplotypes for (A) Alderia modesta, ranging from Kodiak, Alaska, south to San Francisco, California, and (B) Alderia sp., ranging from Walker Creek, California, south to San Diego.

Atlantic and Pacific evolved independently for about 100 million years until the opening of the Bering Strait, as early as 5.5 MYA or as recently as 3.1 MYA (Marincovich and Gladenkov 2001). Formation of the Strait initiated a trans- TABLE 3. Comparative estimated population expansion parameters Arctic migration lasting until the late Pliocene. In this asym- for Alderia species, based on mismatch distributions of COI se- quences from the northeastern Pacific. Time since expansion, ␶,is metric invasion, molluscs of Pacific origin disproportionately given in units of the mutational time scale, or as t in 1000 years colonized the North Atlantic by an 8:1 ratio (Vermeij 1991). (KY; Excoffier and Schneider 1999). Calculations were based on u Species that participated in the trans-Arctic exchange show ϭ 2.62 ϫ 10Ϫ5 substitutions per year and generation times of two either genetic differences suggesting prolonged isolation months (A. modesta) versus one month (Alderia sp.). Raw values ␪ (Grant and Stahl 1988; McDonald and Koehn 1988; Zaslav- of 0 were divided by 2u to estimate population size at expansion ␪ skaya et al. 1992; Collins et al. 1996), or low intraspecific (given as no. of individuals), and after expansion ( 1, given as 1000 ␪ ␪ individuals). The 95% confidence intervals around t, 0 and 1 are variation indicating recent gene flow (Graves and Dizon given in the corresponding units. *P Ͻ 0.05; ***P Ͻ 0.0001. 1989; Palumbi and Wilson 1990; Palumbi and Kessing 1991). A divergence time of 1.7 MYA for Atlantic and Pacific A. A. modesta Alderia sp. modesta suggests early Pleistocene glaciation removed suit- Tajima’s D Ϫ1.67* Ϫ1.58* able habitat from the Arctic, and conditions have since re- Ϫ Ϫ Fu’s FS 24.49*** 24.92*** mained unfavorable for larval transport between basins. At- ␶ 11.76 11.23 t (KY) 37 18 lantic and Pacific slugs have not diverged in morphology or 95% CI 18–59 6–27 development, perhaps due to ecological similarity of their ␪ 0 (individuals) 24,000 40 habitats (Schneider et al. 1999); however, they represent evo- 95% CI 0–72,000 0–100,000 lutionarily independent lineages separated by Quaternary cli- ␪ 1 (1000 individuals) 405 207 mate change. Speciation predated the trans-Arctic exchange, 95% CI 220–1710 100–830 indicating a Pacific origin for the genus Alderia. 2302 R. A. ELLINGSON AND P. J. KRUG

TABLE 4. Pairwise differences among estuaries in Alderia sp. Data are Slatkin’s linearized FST values, with sample size given in parentheses. Nucleotide diversity Ϯ variance is given on the diagonal. Significant pairwise differences were determined by AMOVA in Arlequin with 10,000 permutations of the data. *P Ͻ 0.05, **P Ͻ 0.01.

Tomales Bay San Francisco Morro Bay Los Angeles Newport Bay San Diego (n ϭ 29) (n ϭ 32) (n ϭ 26) (n ϭ 14) (n ϭ 14) (n ϭ 14) TB 0.013 Ϯ 0.007 SF Ϫ0.000 0.015 Ϯ 0.008 MB 0.037* 0.036* 0.012 Ϯ 0.006 LA 0.024 0.010 0.082** 0.018 Ϯ 0.010 NB Ϫ0.002 0.001 0.017 0.049 0.014 Ϯ 0.008 SD Ϫ0.008 Ϫ0.011 0.034 0.019 Ϫ0.014 0.015 Ϯ 0.008

Both species show evidence of recent population expan- this strictly planktotrophic species was panmictic in the sion, the parameters of which were estimated from mismatch northeastern Pacific, and haplotypes from the Sea of Japan data. Assuming Atlantic and Pacific A. modesta diverged 1.7 grouped within the eastern Pacific clade. The dispersal po- MYA, we derived a substitution rate of 5.47% per site per tential of long-lived, feeding larvae is evidently realized in million years for COI, or u ϭ 2.62 ϫ 10Ϫ5 substitutions per A. modesta, genetically homogenizing distant estuaries. With- year for a 480-bp fragment. The average generation time of in Alderia sp., FST results indicated reduced gene flow be- A. modesta was estimated as twice that of Alderia sp., due tween Morro Bay and three other estuaries. This could result to obligate planktotrophy and colder temperatures during lar- from the geographically isolated nature of Morro Bay, and val maturation. Dividing their respective generation times by also from the limited dispersal potential of lecithotrophic 2u gave a mutational time scale of 3180 for A. modesta and larvae produced by Alderia sp. for much of the year (Krug ␶ ⌽ 1580 for Alderia sp.; multiplying by converts this parameter 2001; Krug and Zimmer 2000, 2004). Although ST analysis to years since expansion (Rogers 1995). From these esti- of the COI data found no phylogeographic structure in Alderia mates, Alderia sp. began expanding about 18,000 years ago sp., a clade with primarily southern distribution was evident (95% CI: 6000–27,000), whereas A. modesta expanded about in phylogenetic studies combining COI and 16S data. ␪ 37,000 years ago (95% CI: 18,000–59,000). Values of 0 were ␪ ϭ also higher for A. modesta ( 0 24,000 slugs) than Alderia Present Distribution of the Sister Species ␪ ϭ sp. ( 0 40 slugs), but the confidence intervals for both were too broad (95% CI: 0–100,000) to draw conclusions about Historical processes do not explain the current biogeog- pre-expansion population sizes. However, the higher haplo- raphy of Alderia. The range endpoints of Alderia species fall type diversity and mean number of pairwise differences in neither at Point Conception, a major faunal transition zone, A. modesta are also consistent with a larger pre-expansion nor in central California, where some coastal taxa have phy- population. The ability of long-lived larvae to disperse across logeographic breaks (Valentine 1966; Burton 1998; Wares et the Pacific may have allowed the cold-adapted A. modesta to al. 2001; Dawson 2001; Sotka et al. 2004). The northern limit migrate between viable patches during glacial maxima, es- of Alderia sp. was Bodega, and A. modesta was rare south caping population crashes associated with regional glaciation of Tomales Bay. Evidence from diverse taxa suggests gene (Warner et al. 1982; Marko 2004; Hickerson and Cunningham flow may be interrupted between Bodega and sites to the 2005). south of Point Reyes, California. In the supralittoral copepod In contrast, low sea levels and a concomitant loss of es- Tigriopus californicus, a northern clade with reduced genetic tuarine habitat across central California may have restricted diversity ranges from Alaska to Bodega, suggesting T. cal- Alderia sp. to refugia in southern California throughout the ifornicus expanded from refugia south of Point Reyes (Ed- Pleistocene; reduced nucleotide and haplotype diversities, the mands 2001). The intertidal crab Petrolisthes cinctipes has a ␪ single common COI haplotype, and estimates of 0 all suggest genetic discontinuity between Bodega and the mouth of To- Alderia sp. experienced a more dramatic bottleneck than its males Bay, only 10 km distant (R. Toonen and R. Grosberg, congener (Grant and Bowen 1998; Caudill and Bucklin unpubl. data). In the urchin Strongylocentrotus franciscanus, 2004). A selective sweep, the alternative basis for reduced only Bodega is differentiated from other populations (Moberg diversity, is unlikely given the predominance of silent sub- and Burton 2000). The concordant break at Bodega across stitutions in our COI dataset. The expansion of Alderia sp. diverse taxa may reflect (1) historical processes such as con- dates to the beginning of the Holocene, when higher sea levels tracting ranges during glaciation events; (2) differences in may have allowed it to recolonize estuaries as they reformed the abiotic environment imposing divergent selection on ei- across central California, eventually coming into secondary ther side of Bodega, affecting the distribution and frequency contact with its sister species (e.g., Marko 1998). of linked loci; (3) unusual hydrographic features that affect Both Alderia species are specialists on Vaucheria species, patterns of dispersal around the mouth of Tomales Bay; or which grow only in protected back bays separated by unin- some combination thereof. Further study of the factors con- habitable stretches of open coast. The fragmented nature of trolling the demographics of Alderia species around Point their estuarine habitat was expected to inhibit gene flow. Reyes may shed light on why this area produces phylogeo- However, no phylogeographic structure was evident in A. graphic patterns in diverse invertebrates. modesta within the Pacific, despite its considerable range; Although larvae of Alderia species can disperse hundreds EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2303

FIG. 4. Bayesian tree showing phylogenetic relationships among Alderia species from the Pacific and Atlantic, based on combined COI and 16S mtDNA haplotypes. Bayesian posterior probabilities (above branches), and parsimony bootstrap values based on 1000 replicates (below branches), are shown for clades that received Ն80% support in both analyses. Haplotypes from the western Pacific are indicated with asterisks. Collection sites: Doel, Belgium (DB); Sea of Japan, Russia (RS); Kodiak Island, Alaska (AK); Vancouver, British Columbia (BC); Coos Bay, Oregon (CB); Bodega Harbor, California (BH); Tomales Bay (TB); San Francisco (SF); Morro Bay (MB); Los Angeles (LA); Newport Bay (NB); and San Diego (SD). of km, they were partitioned over spatial scales of less than also express divergent host-specific adaptations limiting their 15 km in Tomales Bay. Distributions of Alderia species may ability to use different Vaucheria species but few studies have reflect those of the host algae if larvae only settle in areas tested whether local coadaptation occurs in marine herbivores where a preferred species of Vaucheria grows. Proximal pop- and host algae (Sotka 2003; Sotka et al. 2003). Alternatively, ulations can remain differentiated due to larval habitat choice a gradient in the physical environment may limit where Al- behavior, an emerging mechanism by which genotypes seg- deria species live if they differ in physiological tolerance regate without extrinsic barriers to gene flow (Appelbaum et (Theison 1978; Koehn et al. 1980; Schmidt and Rand 2001; al. 2002; Gilg and Hilbish 2000, 2003a,b; Hilbish et al. 2003; Bierne et al. 2002; Brind’Amour et al. 2002; Riginos et al. Baird et al. 2003; Bierne et al. 2003). The sister species could 2002). 2304 R. A. ELLINGSON AND P. J. KRUG

FIG. 5. Demography of Alderia species in a 100-km overlap zone. Data are the proportion of A. modesta (hatched bars) versus Alderia sp. expressing planktotrophic (black) or lecithotrophic (white) development. Sample size is given below each pie graph.

There was no evidence of hybridization or introgression spite insemination, and could be a pleiotropic by-product of where Alderia species were sympatric, yet both species at- the different genetic programs that direct larval development tempted to inseminate each other in laboratory crosses. There in the sister species. is no premating isolation between Alderia species; slugs mate year-round via hypodermic insemination, injecting sperm Evolution of Poecilogony anywhere on the recipient’s body (Hand and Steinberg 1955; Angeloni 2003). Divergence in sperm chemoattractants or Transitions between planktotrophy and lecithotrophy have gamete recognition proteins may promote prezygotic isola- occurred frequently in the evolutionary history of many taxa tion (Vacquier 1998; Lyon and Vacquier 1999; Swanson and and can substantially alter the population structure and mac- Vacquier 2002; Zigler et al. 2003; Riffell et al. 2004), but roevolutionary fate of a lineage (Raff 1987; Jeffery and Swal- this has not been tested in a marine organism with internal la 1992; Hadfield et al. 1995; Hart et al. 1997; Duda and fertilization. Postzygotic isolation between Alderia species Palumbi 1999; Hart 2000; McEdward 2000; Jeffery et al. was implicated by the failure of most laboratory crosses de- 2003, Collin 2004). However, for any given species the se- EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2305

FIG. 6. Interspecific versus intraspecific differences in body weight for Alderia species. Data are mean wet weights ϩ SE. All sites are in California except Coos Bay, Oregon. lective forces that favored the evolution of nonfeeding larvae cardia proboscidea (Gibson et al. 1999; Gibson and Gibson remain unclear. Due to the rarity of poecilogony, studies 2004), Streblospio benedicti (Levin 1984; Schulze et al. contrasting different development modes usually employ in- 2000), and Pygospio elegans (Morgan et al. 1999). Identifying terspecific comparisons (e.g., Villinski et al. 2002), which the selective forces favoring poecilogony, and the phyloge- may be confounded by other differences between species. As netic constraints that limit it to a few taxa, may substantially a case of variable development, Alderia sp. should be a model advance our understanding of marine life-history evolution. system with which to study developmental evolution from In southern California, slugs produced mostly lecithotroph- ecological and biochemical perspectives. ic clutches from May through September, with a varying There are few confirmed examples of poecilogony among proportion expressing planktotrophy the rest of the year. Fur- marine invertebrates, despite the adaptive value of dispersal thermore, some adults switch from lecithotrophic to plank- polymorphisms (Giesel 1976; Denno et al. 1980; Harrison totrophic clutch production upon transport to the laboratory 1980; McPeek and Holt 1992; Toonen and Pawlik 1994; Chia (Krug 2007). Development in Alderia sp. is thus to some et al. 1996; Hopper 1999; Langellotto and Denno 2001). Most extent phenotypically plastic. Populations of the poecilogon- claims of poecilogony were resolved as cryptic species dif- ous polychaete S. benedicti did not change development sea- fering in development upon detailed molecular or morpho- sonally, nor did individuals ever change the trophic mode of logical study (Hoagland and Robertson 1988; Hirano and their offspring (Levin and Huggett 1990; Levin and Creed Hirano 1992; Sisson 2002). In molluscs, multiple larval types 1991; Levin and Bridges 1994). The adaptive value of sea- within a species have been confirmed by breeding studies or sonal polyphenism has been studied for morphological char- molecular data for three sacoglossans: Elysia chlorotica acters (Brakefield and French 1999; Nijhout 1999), diapause (West et al. 1984), Alderia sp. (Krug 1998; present study), strategies (Shapiro 1976), and sex determination (Conover and ocellifera (R. A. Ellingson and P. J. Krug, and Heins 1987), but has not been described for larval de- unpubl. data). Among polychaetes, examples include Boc- velopment in a marine .

TABLE 5. Results of a nested ANOVA on weights of Alderia sp. and A. modesta from sites in California and Oregon, with site as a random effect nested within species as a fixed effect.

Effect Type df SS MS FP Intercept fixed 1 16,493.25 16,493.25 24.50 0.0005 Site(species) random 10 7011.50 701.15 42.16 0.0000 Species fixed 1 6015.15 6015.15 8.93 0.0136 Error 318 5288.61 16.63 2306 R. A. ELLINGSON AND P. J. KRUG

TABLE 6. Linear regressions of the effect of weight fecundity in Alderia modesta and Alderia sp.

Species Population Development F-ratio r2 P Regression equation A. modesta ϭ ϭ ϩ Bodega planktotrophic F1,18 9.07 0.34 0.008 y 0.007x 5.45 ϭ ϭ Ϫ San Francisco planktotrophic F1,9 9.86 0.52 0.01 y 35.25x 101.33 ϭ ϭ ϩ Tillamook planktotrophic F1,18 7.86 0.50 0.02 y 0.435x 0.01 Alderia sp. ϭ ϭ ϩ San Francisco planktotrophic F1,18 18.03 0.50 0.0005 y 0.003x 0.19 ϭ ϭ ϩ San Diego lecithotrophic F1,27 4.53 0.14 0.04 y 0.01x 1.74

The same factors that favored lecithotrophy in the ancestor Millen. Special thanks to T. Van Der Knocker and R. Dekker of Alderia sp. could favor its present-day expression during for Atlantic specimens, and A. Chernyshev and A. Shukalyuk warm months. Understanding the ecology of Alderia sp. may for western Pacific specimens. We thank S. Escorza-Trevino thus pinpoint the selective pressures behind the evolution of and E. Torres for advice and access to laboratory facilities. nonfeeding development. Lecithotrophy could be an adap- Access to field sites was provided by I. Kay (Natural Reserve tation to decrease larval mortality, either from offshore ad- Office of the University of California, San Diego), K. Brown vection during upwelling or from a seasonal reduction in and P. Connors (Bodega Marine Reserve), B. Shelton (Upper phytoplankton abundance. However, the transition away from Newport Bay Ecological Reserve), and M. Schaadt (Los An- planktotrophy occurs in May and June in southern California, geles Harbor). Helpful comments on earlier drafts were pro- after the peak in upwelling and when phytoplankton abun- vided by R. Collin, D. Eernisse, M. Hellberg, D. Jacobs, R. dances are high (Bakun and Nelson 1977; Reid et al. 1985; Kelly, C. Meyer, D. O´ Foighil, A´ . Valde´z, and J. Wares. This P. J. Krug, unpubl. data). The environmental parameter best study was supported by awards from the National Science correlated with lecithotrophy is freshwater input from winter Foundation (OCE 02-42272 and HRD 03-17772) to PJK, and rain and spring snowmelt. Historically, estuaries in southern by a Lerner-Grey award to RAE. California were closed by sand berms when runoff ceased in summer and fall; lecithotrophy may thus be an adaptive re- LITERATURE CITED sponse to seasonal habitat closure. Runoff could then flush Allman, G. 1845. On a new genus of nudibranchiate . p. planktotrophic larvae out of back bays and allow alongshore 65 in Notices and abstracts of miscellaneous communications to transport in winter and spring, facilitating dispersal. Alter- the British Association for the Advancement of Science held at natively, lecithotrophy could allow exploitation of stable the York meeting, September 1844. John Murray, London. adult food resources if Vaucheria is less seasonal in the south, Angeloni, L. 2003. Sexual selection in a simultaneous hermaph- rodite with hypodermic insemination: body size, allocation to which could be tested with monthly field surveys of host sexual roles and paternity. Anim. Behav. 66:417–426. algal abundance. Appelbaum, L., Y. Achituv, and O. Mokady. 2002. Speciation and Size and fecundity differences between Alderia species the establishment of zonation in an intertidal barnacle: specific suggest another hypothesis for the evolution of poecilogony, settlement vs. selection. Mol. Ecol. 11:1731–1737. integrating biochemical with ecological constraints. The Arndt, A., and M. J. Smith. 1998. Genetic diversity and population structure in two species of sea cucumber: differing patterns ac- southern species Alderia sp. is smaller than its congener even cording to mode of development. Mol. Ecol. 7:1053–1064. where they coexist, and produces fewer larvae. Due to high Baird, A. H., R. C. Babcock, and C. P. Mundy. 2003. Habitat se- planktonic mortality (Morgan 1995), there may be a threshold lection by larvae influences the depth distribution of six common clutch size below which no long-lived larvae are likely to coral species. Mar. Ecol. Prog. Ser. 252:289–293. Bakun, A. and C. S. Nelson. 1977. Climatology of upwelling-related survive, leading to selection against obligate planktotrophy processes off Baja California. CalCOFI Reports 19:107–127. in Alderia sp. If Vaucheria species are of lower nutritional Bierne, N., P. David, A. Langlade, and F. Bonhomme. 2002. Can value during the summer in southern waters, Alderia sp. may habitat specialization maintain a mosaic hybrid zone in marine be unable to produce enough planktotrophic offspring to off- bivalves? Mar. Ecol. Prog. Ser. 245:157–170. set larval mortality. The adult energy budget and odds of Bierne, N., F. Bonhomme, and P. David. 2003. Habitat preference and the marine-speciation paradox. Proc. R. Soc. Lond. 270: larval survivorship could thus interactively drive evolution 1399–1406. of development mode. Quantifying the caloric content of Bohonak, A. J. 1999. Dispersal, gene flow, and population structure. Vaucheria spp. should indicate whether host value fluctuates Q. Rev. Biol. 74:21–45. with latitude or over the course of a year. Presently being Bouchet, P. 1989. A review of poecilogony in gastropods. J. Mol- luscan Stud. 55:67–78. described (Krug et al. 2007), Alderia sp. is an excellent model Brakefield, P. M., and V. French. 1999. Butterfly wings: the evo- organism with which to explore the factors controlling ex- lution of development of colour patterns. BioEssays 21:391–401. pression of alternative developmental pathways, providing Brind’Amour, A., E. Bourget, and R. Tremblay. 2002. Fecundity, insight into how adult and larval ecology interactively shape growth rate and survivorship at the interface between two con- life-history evolution. tiguous genetically distinct groups of Semibalanus balanoides. Mar. Ecol. Prog. Ser. 229:173–184. Burton, R. S. 1998. Intraspecific phylogeography across the Point ACKNOWLEDGMENTS Conception biogeographic boundary. Evolution 52:734–745. Byrne, M., and A. Cerra. 1996. Evolution of intragonadal devel- Key logistical support was provided by D. and O. Tilton opment in the diminutive asterinid sea stars Patiriella vivipara and T. and S. Kreines. For help with collections, we thank and P. parvivipara with an overview of development in the As- D. Willette, N. Smolensky, G. Botello, V. Rodriguez, and S. terinidae. Biol. Bull. 191:17–26. EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2307

Caillaud, M. C., and S. Via. 2000. Specialized feeding behavior lations do not show signs of Pleistocene demographic expan- influences both ecological specialization and assortative mating sions. Proc. Natl. Acad. Sci. USA 96:10597–10602. in sympatric host races of pea aphids. Am. Nat. 156:606–621. Excoffier, L., P. Smouse, and J. M. Quattro. 1992. Analysis of Caley, M. J., M. H. Carr, M. A. Hixon, T. P. Hughes, G. P. Jones, molecular variance inferred from metric distances among DNA and B. A. Menge. 1996. Recruitment and the local dynamics of haplotypes: application to human mitochondrial DNA restriction open marine populations. Annu. Rev. Ecol. Syst. 27:477–500. data. Genetics 131:479–491. Caudill, C. C., and A. Bucklin. 2004. Molecular phylogeography Feder, J. L., S. B. Opp, B. Wlazlo, K. Reynolds, W. Go, and S. and evolutionary history of the estuarine copepod, Acartia tonsa, Spisak. 1994. Host fidelity is an effective premating barrier be- on the Northwest Atlantic coast. Hydrobiologia 511:91–102. tween sympatric races of the apple maggot fly, Rhagoletis po- Chernyshev, A. V., and E. M. Chaban. 2005. The first findings of monella. Proc. Natl. Acad. Sci. USA 91:7990–7994. Alderia modesta (Love´n, 1844) (Opisthobranchia, Ascoglossa) Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. in the Sea of Japan. Ruthenica 14:131–134. DNA primers for amplification of mitochondrial cytochrome c Chia, F.-S. 1974. Classification and adaptive significance of de- oxidase subunit I from diverse metazoan invertebrates. Mol. velopmental patterns in marine invertebrates. Thalassia Jugosl. Mar. Biol. Biotechnol. 3:294–299. 10:121–130. Fu, Y.-X. 1997. Statistical tests of neutrality of mutations against Chia, F.-S., G. D. Gibson, and P-Y. Qian. 1996. Poecilogony as a population growth, hitchhiking and background selection. Ge- reproductive strategy of marine invertebrates. Oceanol. Acta 19: netics 147:915–925. 203–208. Gascoigne, T. 1976. The reproductive system and classification of Clark, K. B. 1994. Ascoglossan (ϭ Sacoglossan) molluscs in the the Stiligeridae (Opisthobranchia: Sacoglossa). J. Malacol. Soc. Florida Keys: rare marine invertebrates at special risk. Bull. Mar. Aust. 3:157–172. Sci. 54:900–916. Giard, A. 1905. La poecilogonie. CR. 6th Congr. Int. Zool. Berne Coates, A. G. and J. A. Obando. 1996. The geological evolution of 1904:617–646. the Central American Isthmus. Pp. 21–56 in J. B. C. Jackson, Gibson, G., and F.-S. Chia. 1995. Developmental variability in the A. F. Budd and A. G. Coates, eds. Evolution and environment poecilogonous opisthobranch Haminoea callidegenita: life-his- in tropical America. Univ. of Chicago Press, Chicago, IL. tory traits and effects of environmental parameters. Mar. Ecol. Collin, R. 2000. Phylogeny of the Crepidula plana (Gastropoda: Prog. Ser. 121:139–155. Calyptraeidae) cryptic species complex in . Can. Gibson, G., and A. J. F. Gibson. 2004. Heterochrony and the evo- J. Zool. 78:1500–1514. lution of poecilogony: generating larval diversity. Evolution 58: ———. 2004. Phylogenetic effects, the loss of complex characters, 2704–2717. and the evolution of development in calyptraeid gastropods. Gibson, G., I. G. Paterson, H. Taylor, and B. Woolridge. 1999. Evolution 58:1488–1502. Molecular and morphological evidence of a single species, Boc- ———. 2005. Development, phylogeny, and of Bostry- cardia proboscidea (Polychaeta: Spionidae), with multiple de- capulus (Caenogastropoda: Calyptraeidae), an ancient cryptic velopment modes. Mar. Biol. 134:743–751. radiation. Zool. J. Linn. Soc. 144:75–101. Giesel, J. T. 1976. Reproductive strategies as adaptations to life in Collins, T. M., K. Frazer, A. R. Palmer, G. J. Vermeij, and W. M. temporally heterogenous environments. Annu. Rev. Ecol. Syst. Brown. 1996. Evolutionary history of northern hemisphere Nu- cella (Gastropoda, Muricidae): molecules, ecology, and fossils. 7:57–79. Evolution 50:2287–2304. Gilg, M. R., and T. J. Hilbish. 2000. The relationship between allele Conover, D. O., and S. W. Heins. 1987. Adaptive variation in en- frequency and tidal height in a mussel hybrid zone: a test of the vironmental and genetic sex determination in a fish. Nature 326: differential settlement hypothesis. Mar. Biol. 137:371–378. 496–498. ———. 2003a. Spatio-temporal patterns in the genetic structure of Cunningham, C. W. 1997. Can three incongruence tests predict recently settled blue mussels (Mytilus spp.) across a hybrid zone. when data should be combined? Mol. Biol. Evol. 14:733–740. Mar. Biol. 143:679–690. Dawson, M. N. 2001. Phylogeography in coastal marine : ———. 2003b. Patterns of larval dispersal and their effect on the a solution from California? J. Biogeogr. 28:723–736. maintenance of a blue mussel hybrid zone in southwestern Eng- Dawson, M. N., and D. K. Jacobs. 2001. Molecular evidence for land. Evolution 57:1061–1077. cryptic species of Aurelia aurita (Cnidaria, Scyphozoa). Biol. Grant, W. S., and B. W. Bowen. 1998. Shallow population histories Bull. 200:92–96. in deep evolutionary lineages of marine fishes: insights from Dawson, M. N., J. L. Staton, and D. K. Jacobs. 2001. Phylogeog- sardines and anchovies and lessons for conservation. J. Hered. raphy of the tidewater goby, Eucyclogobius newberryi (Teleostei, 89:415–426. Gobiidae), in coastal California. Evolution 55:1167–1179. Grant, W. S., and G. Stahl. 1988. Description of electrophoretic Dayrat, B. 2005. Towards integrative taxonomy. Biol. J. Linn. Soc. loci in Atlantic cod, Gadus morhua, and comparison with Pacific 85:407–415. cod, Gadus macrocephalus. Hereditas 108:27–36. DeSalle, R., M. G. Egan, and M. Siddall. 2005. The unholy trinity: Graves, J. E., and A. E. Dizon. 1989. Mitochondrial DNA sequence taxonomy, species delimitation and DNA barcoding. Philos. similarity of Atlantic and Pacific albacore tuna (Thunnus alal- Trans. R. Soc. Lond. B 360:1905–1916. unga). Can. J. Fish. Aquat. Sci. 46:870–873. Denno, R. F., M. J. Raupp, D. W. Tallamy, and C. F. Reichelderfer. Hadfield, K. A., B. J. Swalla, and W. R. Jeffery. 1995. Multiple 1980. Migration in heterogenous environments: differences in origins of anural development in ascidians inferred from rDNA wing forms of the dimorphic planthopper, Prokelisia marginata sequences. J. Mol. Evol. 40:413–427. (Homoptera: Delphacidae). Ecology 61:850–867. Hall, C. A. 2002. Nearshore marine paleoclimate regions, increasing Duda, T. F., and S. R. Palumbi. 1999. Developmental shifts and zoogeographic provinciality, molluscan extinctions, and paleo- species selection in gastropods. Proc. Natl. Acad. Sci. USA 96: shorelines, California: Late Oligocene (27Ma) to Late Pliocene 10272–10277. (2.5Ma). Geol. Soc. Am. Spec. Pap. 357:1–489. Edmands, S. 2001. Phylogeography of the intertidal copepod Ti- Hand, C., and J. Steinberg. 1955. On the occurrence of the nudi- griopus californicus reveals substantially reduced population dif- branch Alderia modesta (Love´n, 1844) on the central California ferentiation at northern latitudes. Mol. Ecol. 10:1743–1750. coast. Nautilus 69:22–28. Edwards, S. and P. Beerli. 2000. Gene divergence, population di- Hansen, T. A. 1978. Larval dispersal and species longevity in Lower vergence, and the variance in coalescence time in phylogeo- Tertiary gastropods. Science 199:886–887. graphic studies. Evolution 54:1839–1854. Harrison, R. 1980. Dispersal polymorphisms in insects. Annu. Rev. Evans, T. J. 1953. The alimentary and vascular system of Alderia Ecol. Syst. 11:95–118. modesta (Love´n) in relation to its ecology. Proc. Malacol. Soc. Hart, M. 2000. Phylogenetic analyses of mode of larval develop- Lond. 29:249–258. ment. Semin. Cell Dev. Biol. 11:411–418. Excoffier, L. and S. Schneider. 1999. Why hunter-gatherer popu- Hart, M. W., M. Byrne, and M. J. Smith. 1997. Molecular phylo- 2308 R. A. ELLINGSON AND P. J. KRUG

genetic analysis of life-history evolution in asterinid starfish. sociated carbohydrates as settlement cues for larvae of the spe- Evolution 51:1848–1861. cialist marine herbivore Alderia modesta. Biol. Bull. 197: Hartog, C. D. 1959. Distribution and ecology of the slugs Alderia 94–103. modesta and Limapontia depressa in the Netherlands. Beaufortia Krug, P. J., and R. K. Zimmer. 2000. Developmental dimorphism 7:15–36. and expression of chemosensory-mediated behavior: habitat se- Havenhand, J. 1995. Evolutionary ecology of larval types. Pp. 79– lection by a specialist marine herbivore. J. Exp. Biol. 203: 122 in L. McEdward, ed. Ecology of marine invertebrate larvae. 1741–1754. CRC Press, Boca Raton, FL. ———. 2004. Developmental dimorphism: consequences for larval Hellberg, M. E. 1994. Relationships between inferred levels of gene behavior and dispersal potential in a marine gastropod. Biol. flow and geographic distance in a philopatric coral, Balano- Bull. 207:233–246. phyllia elegans. Evolution 48:1829–1854. Krug, P. J., R. A. Ellingson, R. Burton, and A´ . Valde´s. 2007. A Hellberg, M. E., D. P. Balch, and K. Roy. 2001. Climate-driven new poecilogonous species of (Opisthobranchia: Sa- range expansion and morphological evolution in a marine gas- coglossa) from California: comparison with the planktotrophic tropod. Science 292:1707–1710. congener Alderia modesta (Love´n, 1844). J. Moll. Stud. In press. Hilbish, T. J., J. Timmons, V. Agrawal, K. R. Schneider, and M. Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated soft- R. Gilg. 2003. Estuarine habitats protect hybrid mussels from ware for molecular evolutionary genetics analysis and sequence selection. J. Exp. Mar. Biol. Ecol. 292:177–186. alignment. Brief. Bioinformatics 5:150–163. Hickerson, M. J. and C. W. Cunningham. 2005. Contrasting Qua- Langellotto, G. A., and R. F. Denno. 2001. Benefits of dispersal in ternary histories in an ecologically divergent sister pair of low- patchy environments: mate location by males of a wing-dimor- dispersing intertidal fish (Xiphister) revealed by multilocus DNA phic insect. Ecology 82:1870–1878. analysis. Evolution 59:344–360. Lee, C. E. 2000. Global phylogeography of a cryptic copepod spe- Hirano, H. J., and Y. M. Hirano. 1992. Poecilogony or cryptic cies complex and reproductive isolation between genetically species? Two geographically different development patterns ob- proximate ‘‘populations.’’ Evolution 54:2014–2027. served in Cuthona pupillae (Baba, 1961) (Nudibranchia: Aeo- Lee, T., and D. O´ Foighil. 2004. Hidden Floridean biodiversity: lidoidea). J. Moll. Stud. 57:133–141. mitochondrial and nuclear gene trees reveal four cryptic species Hoagland, K. E., and R. Robertson. 1988. An assessment of poe- within the scorched mussel, Brachidontes exustus, species com- cilogony in marine invertebrates: phenomenon or fantasy? Biol. plex. Mol. Ecol. 13:3527–3542. Bull. 174:109–125. Levin, L. A. 1984. Multiple patterns of development in Streblospio Hopper, K. R. 1999. Risk-spreading and bet-hedging in insect pop- benedicti Webster (Spionidae) from three coasts of North Amer- ulation biology. Annu. Rev. Entomol. 44:535–560. ica. Biol. Bull. 166:494–508. Huelsenbeck, J. P., and K. A. Crandall. 1997. Phylogeny estimation Levin, L. A., and T. S. Bridges. 1994. Control and consequences and hypothesis testing using maximum likelihood. Annu. Rev. of alternative developmental modes in a poecilogonous poly- Ecol. Syst. 28:437–466. chaete. Am. Zool. 34:323–332. Jablonski, D. 1986. Larval ecology and macroevolution of marine Levin, L. A., and E. L. Creed. 1991. Effect of temperature and food invertebrates. Bull. Mar. Sci. 39:565–587. availability on reproductive responses of Streblospio benedicti Jacobs, D. K., T. A. Haney, and K. D. Louie. 2004. Genes, diversity, (Polychaeta: Spionidae) with planktotrophic or lecithotrophic and geological process on the Pacific coast. Annu. Rev. Earth development. Mar. Biol. 92:103–113. Planet. Sci. 32:601–652. Levin, L. A., and D. V. Huggett. 1990. Implications of alternative Jeffery, C. H., and R. B. Emlet. 2003. Macroevolutionary conse- developmental reproductive modes for seasonality and demog- quences of developmental mode in temnopleurid echinoids from raphy in an estuarine polychaete. Ecology 71:2191–2208. the Tertiary of Southern Australia. Evolution 57:1031–1048. Love´n, S. L. 1844. Om nordiska Hafs-Mollusker. Ofvers Kongl. Jeffery, C. H., R. B. Emlet, and D. T. J. Littlewood. 2003. Phylogeny Vetensk-Akad. Forh. Stockholm, 1:49. and evolution of development mode in temnopleurid echinoids. Lydeard, C., W. E. Holznagel, M. N. Schnare, and R. R. Gutell. Mol. Biol. Evol. 28:99–118. 2000. Phylogenetic analysis of molluscan mitochondrial LSU Jeffery, W. R. and B. J. Swalla. 1992. Evolution of alternative rDNA sequences and secondary structures. Mol. Biol. Evol. 15: modes of development in ascidians. Bioessays 14:219–226. 83–102. Jensen, K. R. 1983. Factors affecting feeding selectivity in herbiv- Lyon, J. D., and V. D. Vacquier. 1999. Interspecific chimeric sperm orous Ascoglossa (Mollusca: Opisthobranchia). J. Exp. Mar. lysins identify regions mediating species-specific recognition of Biol. Ecol. 66:135–148. the abalone egg vitelline envelope. Dev. Biol. 214:151–159. ———. 1989. Learning as a factor in diet selection by Elysia viridis Marincovich, L., and A. Y. Gladenkov. 2001. New evidence for (Montagu) (Opisthobranchia). J. Moll. Stud. 55:79–88. the age of Bering Strait. Q. Sci. Rev. 20:329–335. ———. 1996. Phylogenetic systematics and classification of the Marko, P. B. 1998. Historical allopatry and the biogeography of Sacoglossan (Mollusca, Gastropoda, Opithobranchia). Philos. speciation in the prosobranch snail genus Nucella. Evolution 52: Trans. R. Soc. Lond. B 351:91–122. 757–774. Knowlton, N. 1993. Sibling species in the sea. Annu. Rev. Ecol. ———. 2002. Fossil calibration of molecular clocks and the di- Syst. 24:189–216. vergence times of geminate species pairs separated by the Isth- ———. 2000. Molecular genetic analyses of species boundaries in mus of Panama. Mol. Biol. Evol. 19:2005–2021. the sea. Hydrobiologia 420:73–90. ———. 2004. ‘‘What’s larvae got to do with it?’’ Disparate patterns Knowlton, N. and L. A. Weigt. 1998. New dates and new rates for of postglacial population structure in two benthic marine gas- divergence across the Isthmus of Panama. Proc. R. Soc. Lond. tropods with identical dispersal potential. Mol. Ecol. 13: B 265:2257–2263. 597–611. Koehn, R. K., R. I. Newell, and F. Immermann. 1980. Maintenance Marko, P. B., and G. J. Vermeij. 1999. Molecular phylogenetics of an aminopeptidase allele frequency cline by natural selection. and the evolution of labral spines among eastern Pacific Ocean Proc. Natl. Acad. Sci. 77:5385–5389. brine gastropods. Mol. Biol. Evol. 13:275–288. Krug, P. J. 1998. Poecilogony in an estuarine opisthobranch: plank- Marko, P. B., A. R. Palmer, and G. J. Vermeij. 2003. Resurrection totrophy, lecithotrophy and mixed clutches in a population of of Nucella ostrina (Gould, 1852), lectotype designation for Ne- the ascoglossan Alderia modesta. Mar. Biol. 132:483–494. marginata (Deshayes, 1839), and molecular genetic evidence of ———. 2001. Bet-hedging dispersal strategy of a specialist marine Pleistocene speciation. Veliger 46:77–85. herbivore: a settlement dimorphism among sibling larvae of Al- May, R. M. 1994. Biological diversity: differences between land deria modesta. Mar. Ecol. Prog. Ser. 213:177–192. and sea. Philos. Trans. R. Soc. Lond. B 343:105–111. ———. 2007. Poecilogony and larval ecology in the gastropod McDonald, J. H., and R. K. Koehn. 1988. The mussels Mytilus genus Alderia. Am. Malacol. Bull. In press. galloprovicialis and M. trossulus on the Pacific coast of North Krug, P. J., and A. E. Manzi. 1999. Waterborne and surface-as- America. Mar. Biol. 99:111–118. EVOLUTION OF VARIABLE LARVAL DEVELOPMENT 2309

McEdward, L. 2000. Adaptive evolution of larvae and life cycles. 2003. Local selection and latitudinal variation in a marine pred- Semin. Cell Dev. Biol. 11:403–409. ator-prey interaction. Science 300:1135–1137. McGovern, T. M., and M. E. Hellberg. 2003. Cryptic species, cryp- Schmidt, P. S., and D. M. Rand. 2001. Adaptive maintenance of tic endosymbionts, and geographical variation in chemical de- genetic polymorphism in an intertidal barnacle: habitat- and life- fenses in the bryozoan Bugula neritina. Mol. Ecol. 12: stage-specific survivorship of MPI genotypes. Evolution 55: 1207–1215. 1336–1344. McPeek, M. A., and R. D. Holt. 1992. The evolution of dispersal Schneider, C. J., T. B. Smith, B. Larison, and C. Moritz. 1999. A in spatially and temporally varying environments. Am. Nat. 140: test of alternate models of diversification in tropical rainforests: 1010–1027. Ecological gradients versus rainforest refugia. Proc. Natl. Acad. Medina, M., and P. J. Walsh. 2000. Molecular systematics of the Sci. USA 96:13869–13873. order Anaspidea based on mitochondrial DNA sequence (12S, Schneider, S. and L. Excoffier. 1999. Estimation of past demo- 16S, and COI). Mol. Biol. Evol. 15:41–58. graphic parameters from the distribution of pairwise differences Meyer, C. P., J. B. Geller, and G. Paulay. 2005. Fine-scale ende- when the mutation rates vary among sites: application to human mism on coral reefs: Archipelagic differentiation in turbinid gas- mitochondrial DNA. Genetics 152:1079–1089. tropods. Evolution 59:113–125. Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin: a soft- Moberg, P. E., and R. S. Burton. 2000. Genetic heterogeneity among ware for population genetics data analysis. Ver. 2.000. Genetics adult and recruit red sea urchins, Strongylocentrotus franciscan- and Biometry Laboratory, University of Geneva, Switzerland. us. Mar. Biol. 136:773–784. Schulze, S. R., S. A. Rice, J. L. Simon, and S. A. Karl. 2000. Morgan, S. G. 1995. Life and death in the plankton: Larval mortality Evolution of poecilogony and the biogeography of North Amer- and adaptation. Pp. 279–321 in L. McEdward, ed. Ecology of ican populations of the polychaete Streblospio. Evolution 54: marine invertebrate larvae. CRC Press, Boca Raton, FL. 1247–1259. Morgan, T. S., A. D. Rogers, G. L. J. Paterson, L. E. Hawkins, and Seeleman, U. 1967. Rearing experiments on the amphibian slug M. Sheader. 1999. Evidence for poecilogony in Pygospio elegans Alderia modesta. Helgol. Wiss. Meeresunters. 15:128–134. (Polychaeta: Spionidae). Mar. Ecol. Prog. Ser. 178:121–132. Shapiro, A. M. 1976. Seasonal polyphenism. Evol. Biol. 9:259–333. Nei, M., and W.-H. Li. 1979. Mathematical modeling for studying Sisson, C. G. 2002. Dichotomous life history patterns for the nu- genetic variation in terms of restriction endonucleases. Proc. dibranch Dendronotus frondosus (Ascanius, 1774) in the Gulf of Natl. Acad. Sci. 76:5269–5273. Maine. Veliger 45:290–298. Nijhout, H. F. 1999. Control mechanisms of polyphenic develop- Slatkin, M. 1991. Inbreeding coefficients and coalescent times. Gen- ment in insects. BioScience 49:181–192. et. Res. 58:167–175. Palmer, A. R., S. D. Gayron, and D. S. Woodruff. 1990. Repro- ———. 1995. A measure of population subdivision based on mi- ductive, morphological, and genetic evidence for two cryptic crosatellite allele frequencies. Genetics 129:555–562. species of northeastern Pacific Nucella. Veliger 33:325–338. Sokal, R. R., and F. J. Rohlf. 1995. Biometry: he principles and Palumbi, S. R. 1992. Marine speciation on a small planet. Trends practice of statistics in biological research. W. H. Freeman, San Francisco, CA. Ecol. Evol. 7:114–118. Sotka, E. E. 2003. Genetic control of feeding preference in the ———. 1995. Using genetics as an indirect estimator of larval herbivorous amphipod Ampithoe longimana. Mar. Ecol. Prog. in dispersal. Pp. 369–387 L. McEdward, ed. Ecology of marine Ser. 256:305–310. invertebrate larvae. CRC Press, Boca Raton, FL. Sotka, E. E., J. P. Wares, and M. E. Hay. 2003. Geographic and ———. 1996. Nucleic acids. II. The polymerase chain reaction. genetic variation in feeding preference for chemically defended Pp. 205–247 in D. Hillis, C. Moritz, and B. Mable, eds. Mo- seaweeds. Evolution 57:2262–2276. lecular systematics. Sinauer, Sunderland, MA. Sotka, E. E., J. P. Wares, J. A. Bath, R. K. Grosberg, and S. R. Palumbi, S. R., and B. D. Kessing. 1991. Population biology of the Palumbi. 2004. Strong genetic clines and geographical variation trans-Arctic exchange: mtDNA sequence similarity between Pa- in gene flow in the rocky intertidal barnacle Balanus gladula. cific and Atlantic sea urchins. Evolution 45:1790–1805. Mol. Ecol. 13:2143–2156. Palumbi, S. R., and A. C. Wilson. 1990. Mitochondrial DNA di- Strathmann, R. R. 1978. The evolution and loss of feeding larval versity in the sea urchins Strongylocentrotus purpuratus and S. stages in marine invertebrates. Evolution 32:899–906. droebachiensis. Evolution 44:403–425. ———. 1990. Why life histories evolve differently in the sea. Am. Pechenik, J. A. 1999. On the advantages and disadvantages of larval Zool. 30:197–207. stages in benthic invertebrate life cycles. Mar. Ecol. Prog. Ser. Strathmann, R. R., and M. F. Strathmann. 1982. The relationship 177:269–297. between adult size and brooding in marine invertebrates. Am. Posada, D., and K. A. Crandall. 1998. Modeltest: testing the model Nat. 119:91–101. of DNA substitution. Bioinformatics 14:817–818. Swanson, W. J., and V. D. Vacquier. 2002. The rapid evolution of Raff, R. A. 1987. Constraint, flexibility, and phylogenetic history reproductive proteins. Nat. Rev. Genet. 3:137–144. in the evolution of direct development in sea urchins. Dev. Biol. Swofford, D. L. 2001. PAUP*: phylogenetic analysis using parsi- 119:6–19. mony (*and other methods). Sinauer Associates, Sunderland, Reid, D. G. 1990. Trans-Arctic migration and speciation induced MA. by climate change: the biogeography of Littorina (Mollusca: Tamura, K. 1992. Estimation of the number of nucleotide substi- Gastropoda). Bull. Mar. Sci. 47:35–49. tutions when there are strong transition-transversion and G ϩ Reid, F. M., C. B. Lange, and M. M. White. 1985. Microplankton C content biases. Mol. Biol. Evol. 9:678–687. species assemblages at the Scripps Pier from March to November Tamura, K. and M. Nei. 1993. Estimation of the number of nucle- 1983 during the 1982–84 El Nin˜o event. Bot. Mar. 28:443–452. otide substitutions in the control region of mitochondrial DNA Riffell, J. A., P. J. Krug, and R. K. Zimmer. 2004. The ecological in humans and chimpanzees. Mol. Biol. Evol. 10:512–526. and evolutionary consequences of sperm chemoattraction. Proc. Tarjuelo, I., D. Posada, K. A. Crandall, M. Pascual, and X. Turon. Natl. Acad. Sci. USA 101:4501–4506. 2004. Phylogeography and speciation of colour morphs in the Riginos, C., K. Sukhdeo, and C. W. Cunningham. 2002. Evidence colonial ascidian Pseudodistoma crucigaster. Mol. Ecol. 13: for selection at multiple allozyme loci across a mussel hybrid 3125–3136. zone. Mol. Biol. Evol. 19:347–351. Theison, B. 1978. Allozyme clines and evidence of strong selection Rogers, A. R. 1995. Genetic evidence for a Pleistocene population in three loci in Mytilus edulis L. (Bivalvia) from Danish waters. explosion. Evolution 49:608–615. Ophelia 17:135–142. Ronquist, F. and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and phylogenetic inference under mixed models. Bioinformatics 19: D. G. Higgins. 1997. The ClustalX windows interface: flexible 1572–1574. strategies for multiple sequence alignment aided by quality anal- Sanford, E., M. S. Roth, G. C. Johns, J. P. Wares, and G. N. Somero. ysis tools. Nucleic Acids Res. 24:4876–4882. 2310 R. A. ELLINGSON AND P. J. KRUG

Thompson, W. 1844. Report of the fauna of Ireland: Div. Inver- time and space: the emergence of the biosphere. Academic Press, tebrata. Report of the British Association Advancement Science . for 1843. No. 250. Vermeij, G. J. 1982. Phenotypic evolution in a poorly dispersing Todd, C. D. 1998. Larval supply and recruitment of benthic inver- snail after arrival of a predator. Nature 299:349–350. tebrates: Do larvae always disperse as much as we believe? ———. 1991. Anatomy of an invasion: the trans-Arctic inter- Hydrobiologia 375:1–21. change. Paleobiology 17:281–307. Toonen, R. J. and J. R. Pawlik. 1994. Foundations of gregariousness. Villinski, J. T., J. C. Villinski, and R. A. Raff. 2002. Convergence Nature 370:511–512. in maternal provisioning strategy during developmental evolu- Trowbridge, C. D. 1991. Diet specialization limits herbivorous sea tion of sea urchins. Evolution 56:1764–1775. slug’s capacity to switch among food species. Ecology 72: Wares, J., S. Gaines, and C. Cunningham. 2001. A comparative 1880–1888. study of asymmetric migration events across a marine biogeo- ———. 1993. Local and regional abundance patterns of the As- graphic boundary. Evolution 55:295–306. coglossan (ϭ Sacoglossan) opisthobranch Alderia modesta (Lov- Warner, B. G., R. W. Mathewes, and J. J. Clague. 1982. Ice-free e´n, 1844) in the northeastern Pacific. Veliger 36:303–310. conditions on the Queen Charlotte Islands, British Columbia, at ———. 2002. Northeastern Pacific sacoglossan opisthobranchs: the height of Late Wisconsin glaciation. Science 218:675–677. natural history review, bibliography, and prospectus. Veliger 45: West, H. H., J. F. Harrigan, and S. K. Pierce. 1984. Hybridization 1–24. of two populations of a marine opisthobranch with different Trowbridge, C. D., and C. D. Todd. 2000. Host-plant change in developmental patterns. Veliger 26:199–206. marine specialist herbivores: ascoglossan sea slugs on intro- Wiens, J. J. 1998. Combining data sets with different phylogenetic duced macroalgae. Ecol. Monogr. 71:219–243. histories. Syst. Biol. 47:568–581. Vacquier, V. D. 1998. Evolution of gamete recognition proteins. Wray, G. A. 1995. Evolution of larvae and development modes. Science 281:1995–1998. Pp. 413–448 in L. McEdward, ed. Ecology of marine invertebrate Valde´s, A´ . 2003. Preliminary molecular phylogeny of the radula- larvae. CRC Press, Boca Raton, FL. less dorids (Gastropoda: Opisthobranchia), based on 16S Zaslavskaya, N. I., S. O. Sergievsky, and A. N. Tatarenkov. 1992. mtDNA sequence data. J. Molluscan Stud. 69:75–80. Allozyme similarity of Atlantic and Pacific species of Littorina Valentine, J. W. 1966. Numerical analysis of marine molluscan (Gastropoda: Littorinidae). J. Molluscan. Stud. 58:377–384. ranges on the extratropical northeastern Pacific shelf. Limnol. Zigler, K. S., E. C. Raff, E. Popodi, R. A. Raff, and H. A. Lessios. 2003. Adaptive evolution of bindin in the genus Heliocidaris is Oceanogr. 11:198–211. correlated with the shift to direct development. Evolution 57: Valentine, J. W., and D. Jablonski. 1983. Speciation in the shallow 2293–2302. sea: general patterns and biogeographic controls. Pp. 201–226 in R. W. Sims, J. H. Price, and P. E. S. Whalley, eds. Evolution, Corresponding Editor: D. O´ Foighil