Incorporating geographical and evolutionary rarity into conservation prioritization

Key goals of conservation are to protect both species and the functional and genetic diversity they represent. A strictly species-based approach may underrepresent rare, threatened, or genetically distinct species and overrepresent widespread species. Although reserves are created for a number of reasons, including economic, cultural, and ecological reasons, their efficacy has been measured primarily in terms of how well species richness is protected, and it is useful to compare how well they protect other measures of diversity. We used Proteaceae species-occurrence data in the Cape Floristic Region of South Africa to illustrate differences in the spatial distribution of species and evolutionary diversity estimated from a new maximum-likelihood molecular phylogeny. We calculated species richness, phylogenetic diversity (i.e., summed phylogenetic branch lengths in a site), and a site-aggregated measure of biogeographically weighted evolutionary distinctiveness (i.e., an abundance weighted measure that captures the unique proportion of the phylogenetic tree a species represents) for sites throughout the Cape Floristic Region. Species richness and phylogenetic diversity values were highly correlated for sites in the region, but species richness was concentrated at a few sites that underrepresented the much more spatially extensive distribution of phylogenetic diversity. Biogeographically weighted evolutionary diversity produced a scheme of prioritization distinct from the other 2 metrics and highlighted southern sites as conservation priorities. In these sites, the high values of biogeographically weighted evolutionary distinctiveness were the result of a nonrandom relation between evolutionary distinctiveness and geographical rarity, where rare species also tended to have high levels of evolutionary distinctiveness. Such distinct and rare species are of particular concern, but are not captured by conservation schemes that focus on species richness or phylogenetic diversity alone.     

Patterns of Genetic Diversity in Remaining Giant Panda Populations

Abstract: The giant panda ( Ailuropoda melanoleuca ) is among the more familiar symbols of species conservation. The protection of giant panda populations has been aided recently by the establishment of more and better-managed reserves in existing panda habitat located in six mountain ranges in western China. These remaining populations are becoming increasingly isolated from one another, however, leading to the concern that historic patterns of gene flow will be disrupted and that reduced population sizes will lead to diminished genetic variability. We analyzed four categories of molecular genetic markers (mtDNA restriction-fragment-length polymorphisms [RFLP], mtDNA control region sequences, nuclear multilocus DNA fingerprints, and microsatellite size variation) in giant pandas from three mountain populations (Qionglai, Minshan, and Qinling) to assess current levels of genetic diversity and to detect evidence of historic population subdivisions. The three populations had moderate levels of genetic diversity compared with similarly studied carnivores for all four gene measures, with a slight but consistent reduction in variability apparent in the smaller Qinling population. That population also showed significant differentiation consistent with its isolation since historic times. From a strictly genetic perspective, the giant panda species and the three populations look promising insofar as they have retained a large amount of genetic diversity in each population, although evidence of recent population reduction—likely from habitat loss—is apparent. Ecological management to increase habitat, population expansion, and gene flow would seem an effective strategy to stabilize the decline of this endangered species.     

Conservation Genetics in the Management of Desert Fishes

Abstract: The status and security of fishes in North American deserts have steadily declined in this century due to man's activities in this naturally fragile region. We address genetic aspects of the population structure of desert fishes as applicable to conservation and recovery programs by developing two zoogeographic models of isolation and gene flow. In the Death Valley model populations are isolated, with no chance of natural gene flow among them. Genetic diversity within populations tends to be low, but genetic divergence among populations within a species is high. In the Stream Hierarchy model, a complicated hierarchical genetic structure exists and is a function of geographic proximity and connectivity of habitats. Within-habitat genetic diversity tends to be higher, and among-habitat differentiation lower, than in the Death Valley model. These two systems must be recognized as distinct and managed differently. We also suggest three areas of experimentation needed to better understand and manage genetic stocks of desert fishes: relationships between heterozygosity and fitness, experimental mixing of similar stocks to examine effects of increased heterozygosity, and analysis of the relative roles of genetic adaptation and phenotypic plasticity in local differentiation.     

Low variation in partial cytochrome oxidase subunit I (COI) mitochondrial sequences in the coralline demosponge Astrosclera willeyana across the Indo-Pacific

Partial sequences of the mitochondrial DNA (mtDNA) gene cytochrome oxidase subunit 1 (COI) were analysed from individuals of the coralline demosponge Astrosclera willeyana sensu lato out of ten Indo-Pacific populations from the Red Sea to the central Pacific. This taxon is widely distributed in cryptic coral reef habitats of the Indo-Pacific and is regarded as a modern representative of long-extinct, formerly reef-building stromatoporoid-type sponges. The aims were to clarify phylogeographic and taxonomic relationships in this “living fossil” and to explore mitochondrial DNA sequence variation over a wide geographic range. Very low variability was observed across the Indo-Pacific, as only three COI haplotypes were identified, with a maximum p-distance of 0.418% and low nucleotide diversity (π=0.00049). Very low genetic structure was revealed among populations: Haplotype 1 was found in all specimens from nine Pacific populations (N=45), separated by distances of more than 7,000 km; haplotype 2 was restricted to the Red Sea population (N=4); haplotype 3 was only found in the Tuamoto specimens (N=7). COI data presented here do not support the hypothesis of at least two sibling species belonging to genus Astrosclera in the Pacific. Considering the maximum geographic distance of more than 20,000 km between sampled populations, mtDNA COI sequence variation observed here is among the lowest reported to date for a diploblastic taxon and adds to the growing evidence of a general mtDNA conservation in sponges. It is argued that this low mtDNA variation in A. willeyana s.l. is due to a low rate of mtDNA evolution caused by a combination of long generation time and low metabolic rate.     

Population genetic structure of North American <Emphasis Type="Italic">Ophiactis</Emphasis> spp. brittle stars possessing hemoglobin

With the discovery of previously unreported populations of hemoglobin-possessing Ophiactis from the Texas coast in the Gulf of Mexico, an investigation into its population structure, including populations of O. simplex from the Pacific coast of California and O. rubropoda from the Atlantic coast of Florida, was undertaken using DNA sequence data from the mitochondrial COI gene. The reconstructed haplotype network suggests that California populations contain the ancestral source of mtDNA variation, and there is no evidence of recent introductions into Texas. Population genetic analyses reveal the California, Florida, and Texas Ophiactis populations to each be significantly differentiated from one another. Sequence divergence among the three areas is shallower than would be predicted given biogeographic history. Texas and Florida populations are equally genetically diverged from California populations as they are to one another, despite the greater potential for gene flow between these areas. The genetic distinctiveness of the Texas populations and the concordance of this pattern with phylogeographic patterns in other brittle star systems indicate an isolated and independent evolutionary history and we hypothesize that the three geographic regions included in this study each serve as hypotheses of population-level lineages that remain to be tested with independent sources of data.     

Patterns of Rarity in the Birds of the Atlantic Forest of Brazil

Patterns of rarity in species are generally explained by several factors: evolutionary history, spatial distribution, and genetic structure of each taxon. Human intervention also leads to or increases rarity in species. The discernment of causes of rarity is essential to the understanding of extinction patterns, and thus to devising conservation strategies. I examine patterns of rarity among bird species in the Atlantic forest region in Brazil, one of the most threatened ecosystems in the world. I assigned bird species to one of eight possible categories that differ in degree of vulnerability and that are based on three parameters of rarity: geographic distribution, habitat specificity, and population size. The Atlantic forest avifauna is a highly endangered group; 68% of the species are rare. Patterns of rarity among the birds in the region likely result from their specific ecologies or evolutionary histories. In addition, human alteration of natural habitats and hunting pressures have undoubtedly influenced rarity for a number of species.     

Landscape bioacoustics allow detection of the effects of habitat patchiness on population structure

Landscape structure may affect individual dispersal abilities, thus influencing the genotypic and phenotypic composition of populations. We analyzed the interplay among landscape, behavior, and evolutionary processes by correlating habitat patchiness to the variability in vocalizations of Dupont's Lark Chersophilus duponti, one of the most habitat-selective and rare European songbirds. We tape-recorded males throughout the species distribution in Spain, analyzed the spatial patterns of territorial call variation at different scales (individuals, populations, and broad geographic areas), and related acoustic variability to patterns of isolation by geographic distance and by landscape unsuitability (calculated by building a predictive model of habitat suitability). The differentiation of spectro-temporal call features resulted from both isolation by distance and isolation by landscape unsuitability mechanisms. Landscape connectivity was often a better determinant of call differentiation than simple straight-line distance between individuals, providing the first evidence that call transmission can be limited by the presence and distribution of patches of adequate habitat, which likely mediates bird dispersal. Landscape patchiness resulted in a reduction of acoustic diversity (repertoire size) within populations, and a parallel increase in differentiation among populations. Landscape bioacoustics can represent a promising tool for estimating population structure, although the study of animal communication cannot be viewed as an alternative, but a source of complementary information to genetics, given that it provides evidence of male-male transmission and social and cultural phenomena that are currently undetectable from molecular data.     

The application of adaptive cluster sampling for rare subtidal macroalgae

Adaptive cluster sampling (ACS) is a targeting sampling method that provides unbiased abundance estimators for populations of rare species that may be inadequately sampled with simple random sampling (SRS). ACS has been used successfully to estimate abundances of rockfish and sardine larvae from shipboard surveys. In this study, we describe the application of ACS for subtidal macroalgae. Using SCUBA, we measured abundances of Codium mamillosum, C. pomoides, and Halimeda cuneata at three islands and two levels of wave exposure. The three species were relatively patchy and could be sampled with ACS at one site per dive. Their distributions differed among islands and with exposure to wave energy, with H. cuneata found at only one island. ACS is a useful tool for understanding the spatial distribution and abundance of populations of rare benthic species, but, as was the case in this study, may not be as efficient as sampling with SRS with comparable replication.     

Phylogenetic Reanalysis of the Saudi Gazelle and Its Implications for Conservation

Abstract: The identification of taxonomically appropriate populations of endangered species for captive breeding and reintroduction programs is fundamental to the success of those programs. The Saudi gazelle (   Gazella saudiya ) was endemic to the Arabian peninsula but is now considered extinct in the wild and is potentially a candidate for captive breeding and reintroduction. Using 375 base pairs of mitochondrial DNA (mtDNA) cytochrome b gene derived from museum samples collected from the wild prior to the presumed extinction of this species, we show that G. saudiya is the sister taxon of the African dorcas gazelle (  G. dorcas ). Reciprocal monophyly of G. saudiya mtDNA haplotypes with G. dorcas , coupled with morphological distinctiveness, suggests that it is an evolutionarily significant unit. These data indicate that captive populations identified previously as potential sources of G. saudiya for captive breeding appear incorrectly designated and are irrelevant to the conservation of G. saudiya. The polymerase chain reaction–restriction fragment length polymorphism ( PCR-RFLP) analysis of several private collections of living gazelles in Saudi Arabia provides no evidence for the survival of G. saudiya. We recommend that field surveys be undertaken to establish whether G. saudiya is indeed extinct in the wild and that other private collections within the Arabian peninsula be screened genetically. We urge caution when captive animals of unknown provenance are used to investigate the phylogenetics of cryptic species groups.     

Genetic structure of natural populations of California red abalone (<Emphasis Type="Italic">Haliotis rufescens</Emphasis>) using multiple genetic markers

Mitochondrial cytochrome oxidase subunit one (COI) sequence, nuclear microsatellites, and amplified fragment length polymorphisms (AFLPs) were used to evaluate connectivity among nine red abalone (Haliotis rufescens) populations sampled between August 1998 and November 2003 along approximately 1,300 km of California coastline from Crescent City (41°46′N, 124°12′W) to San Miguel Island (34°02′N, 120°22′W). COI sequences and microsatellite genotypes did not show significant genetic divergence among nine sampled populations. A subset of five populations spanning the geographic range of the study was scored for 163 polymorphic AFLP markers. Of these, 41 loci showed significant divergence (P < 0.001) among populations. Still, no AFLP markers were diagnostic for any of the study populations, and assignment tests did not consistently assign individuals to the correct population. Although the AFLP data are the first to suggest there is significant genetic differentiation among California red abalone populations, the discordance between the different genetic markers needs further study before unambiguous conclusions can be drawn with respect to connectivity among the populations. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Desert-dwelling Mountain Sheep: Conservation Implications of a Naturally Fragmented Distribution

Abstract: Mountain sheep (Ovis canadensis) are closely associated with steep, mountainous, open terrain. Their habitat consequently occurs in a naturally fragmented pattern, often with substantial expanses of unsuitable habitat between suitable patches; the sheep have been noted to be slow colonizers of vacant suitable habitat. As a result, resource managers have focused on (1) conserving "traditional" mountainous habitats, and (2) forced colonization through reintroduction. Telemetry studies in desert habitats have recorded more intermountain movement by desert sheep than was previously thought to OCCUT. Given the heretofore unrecognized vagility of mountain sheep, we argue that existing corridors of "nontraditional" habitat connecting mountain ranges be given adequate conservation consideration. Additionally, small areas of mountainous habitat that an? not permanently occupied but that may serve as "stepping stones" within such corridors must be recognized for their potential importance to relatively isolated populations of mountain sheep. We discuss the potential importance of such corridors to other large, vagile species.     

Genetic Diversity and Population Size in Four Rare Southern Appalachian Plant Species

Allozyme diversity was examined in four rare, high-montane plant species from the Appalachian Mountains of southeastern North America. These species may represent relictual members or descendants of an alpine community that was more widespread during the late Pleistocene. We sampled five populations of Geum radiatum (Rosaceae), Carex misera (Cyperaceae), Trichophorum cespitosum (Cyperaceae), and the four known populations of Calamagrostis cainii (Poaceae). Genetic diversity was low for all species but was typical of that found for plant species with limited ranges. Low genetic diversity may reflect historical events associated with changes in the species' biogeography. As the Pleistocene climate warmed, suitable habitat decreased in areal extent and became fragmented, probably resulting in smaller, more-isolated populations. In recent times these species, which co-occur in fragile rock outcrop habitats, have been adversely affected by human activities. Genetic analyses revealed reduced diversity in populations of decreasing size for three species. Estimates of gene flow were low ( Nm < 1.0) in all four species. Positive associations between genetic diversity and population size, evidence of recent population declines, and the low estimates of gene flow suggest that genetic drift may play a prominent role in shaping the present-day genetic composition of these species. Furthermore, these data suggest that the genetically depauperate populations are unlikely to regain genetic variation without human intervention.     

Consequences of Low Mate Availability in the Rare Self-Incompatible Species Brassica insularis

Abstract:  Self-incompatibility systems prevent self-fertilization in angiosperms. Although numerous S alleles are usually maintained by negative frequency-dependent selection, the number of S alleles can be low in small populations, which limits mate availability and reduces fecundity in endangered populations of self-incompatible plants. Despite the increasing evidence of the negative effect of self-incompatibility in small populations, the direct link between the number and the distribution of S alleles and their reproductive consequences has been rarely reported. Brassica insularis is a rare self-incompatible species with medium to very small populations. Results of a previous study showed that the smallest population has very few S alleles. We investigated whether reduced mate availability affects reproduction in this species. We compared the pollination success and the fruit set in 4 populations differing in population size and number of S alleles. Our results suggest that reproduction may be negatively affected by the low S-allele diversity in the smallest population. Nevertheless, other populations also had reduced fruit set that could not be attributed to self-incompatibility alone.     

Critique of the Evolutionarily Significant Unit as a Definition for "Distinct Population Segments" under the U.S. Endangered Species Act

The U.S. Endangered Species Act grants protection to species, subspecies, and "distinct population segments" of vertebrate species. Historically, Congress included distinct population segments into endangered species legislation to enable the U.S. Fish and Wildlife Service to implement a flexible and pragmatic approach in listing populations of vertebrate species. Recently, the U.S. Fish and Wildlife Service and the National Marine Fisheries Service have proposed a policy that would narrowly define distinct population segments as evolutionarily significant units based on morphological and genetic distinctiveness between populations. Historically, the power to list species or populations as distinct population segments has been used to tailor management practices to unique circumstances; grant varied levels of protection in different parts of a species' range; protect species from extinction in significant portions of their ranges as well as to protect populations that are unique evolutionary entities. A strict redefinition of distinct population segments as evolutionarily significant units will compromise management efforts because the role of demographic and behavioral data will be reduced. Furthermore, strictly cultural, economic, or geographic justifications for listing populations as threatened or endangered will be greatly curtailed.     

Cover Caption

Cover: Tree pincushion (Leucospermum conocarpodendron viridum), a member of the family Proteaceae and the largest member of its genus, Cape of Good Hope National Park, South Africa. On pages 593‐601, Tucker et al. explore differences in the species richness, phylogenetic diversity, and evolutionary distinctiveness of Proteaceae species in the Cape Floristic Region of South Africa. They found that species richness and phylogenetic diversity typically were highly correlated, but sites with high phylogenetic diversity were more spatially extensive than sites with high species richness. Rare species generally had high levels of evolutionary distinctiveness. Sites near the southern edge of the Cape Floristic Region were occupied by species with high levels of evolutionary distinctiveness and limited geographic ranges, but those sites were not assigned high conservation‐priority rankings on the basis of species richness or phylogenetic distinctiveness. Photographer: Photographer and conservationist Jaime Rojo ( http://www.jaime‐rojo.com ) believes visual storytelling is a powerful tool for conservation. Since moving to Mexico from his native Spain 8 years ago, Jaime has worked with diverse environmental nongovernmental organizations while developing his career as a photographer. He has participated in numerous conservation initiatives including the promotion of El Carmen Big Bend Conservation Corridor, a transboundary protected area along the border between Mexico and the United States; a campaign to protect the San Pedro Mezquital, the last free‐flowing river in the western Sierra Madre; and the celebration of WILD9, the ninth World Wilderness Congress, for which he was executive director in Mexico. Rojo is a trustee of Th e WILD Foundation and an associate of the International League of Conservation Photographers.     

Bathymetric patterns of genetic variation in a deep-sea protobranch bivalve, Deminucula atacellana

The origin of the deep-sea benthic fauna is poorly understood and represents an enormous gap in our understanding of basic evolutionary phenomena. One obstacle to studying evolutionary patterns in the deep sea has been the technical difficulty of measuring genetic variation in species that are typically minute, rare, and must be recovered from extreme depths. We used molecular genetic techniques to quantify variation in the 16S rRNA mitochondrial gene within and among populations of the common protobranch bivalve Deminucula atacellana (Schenck, 1939). We analyzed 89 individuals from nine samples collected in the 1960s along a depth gradient from 1100 to 3800 m in the western North Atlantic. Genetic variability within populations is much lower than between populations, and peak haplotype numbers occur near the center of its depth distribution. Continental slope (<2500 m) and rise (>2500 m) populations were genetically distinct despite the lack of any obvious topographic or oceanographic features that would impede gene flow. These findings indicate that the deep-sea macrofauna can have strong population structure over small (134 km) spatial scales, similar to that observed in shallow-water and terrestrial organisms. This surprisingly high biodiversity at the genetic level affords the potential for adaptation and evolutionary diversification, the ultimate historical causes of high species diversity in the deep-sea benthos. Received: 24 July 1997 / Accepted: 26 January 1998     

A Hierarchical View of Genetic Structure in the Rare Annual Plant Clarkia springvillensis

Genetic structure at several spatial scales was examined in the rare California annual, Clarkia springvillensis . Using seven isozyme-encoding loci as genetic markers, we assessed the amount and distribution of genetic variation among three populations and eight subpopulations. Total genetic variation was lower than in species with similar life history traits but equivalent to that of other endemic plants. Spatial autocorrelation showed some evidence for very limited differentiation within subpopulations at a scale of 1–2 m. The subpopulations, separated by tens of meters, were found to be more differentiated from each other ( F _sp = 0.084) on average than were populations ( F,pt = 0.017). This local genetic differentiation was not correlated with physical distance between subpopulations. The low F_pt estimates suggest that substantial gene flow is occurring among populations. However, the lack of correlation between genetic and geographic distances and the significant differentiation of subpopulations suggest that genetic drift is occurring within populations. Therefore, we believe the apparent homogeneity of populations is due to each population's gene frequencies' being an average of several divergent subpopulations. If drift is causing differentiation within populations, it may eventually cause differentiation between populations. The importance of using a hierarchical approach to evaluating genetic structure is clear. Patterns occurring at one spatial scale may not be evident at others. One should not necessarily conclude that gene flow is substantial and that the risk of genetic erosion via drift is negligible just because differentiation between populations is small; the system may not be at equilibrium. This lesson is particularly important when recent changes in climate or land use are apparent.     

Testing the evolution of increased competitive ability (EICA) hypothesis in a novel framework

The "evolution of increased competitive ability" (EICA) hypothesis proposes that escape from natural enemies, e.g., after transcontinental introductions, alters the selection regime because costly defenses no longer enhance fitness. Such an evolutionary loss of defenses enables resources to be directed toward growth or other traits improving performance. We tested the EICA hypothesis in a novel framework in which the natural enemy is the traveler that follows its widespread host by accidental or deliberate (biocontrol) introductions. In a greenhouse experiment we used populations of Senecio vulgaris from North America, Europe, and Australia that differ in the history of exposure to the rust fungus Puccinia lagenophorae. Contrary to what is predicted by EICA, we found no evidence for increased levels of resistance to the rust fungus in plant populations with a longer history of rust exposure (Australia). Similarly, there was no evidence for reduced fecundity in these populations, although vegetative vigor, measured as secondary branching and growth rate, was lower. The maintenance of high rust resistance in populations with no (North America) or only a short history (Europe) of rust exposure is surprising given that resistance seems to incur considerable fitness costs, as indicated by the negative association between family mean resistance and fitness in the absence of disease observed for all three continents. The comparison of population differentiation in quantitative traits with estimates of differentiation in amplified fragment length polymorphic (AFLP) markers suggests that a number of fitness-related traits are under divergent selection among the studied populations. The proposed framework to test changes in the evolutionary trajectory underlying EICA can be employed in an expanded range of systems. These may include investigations on a cosmopolitan weed or crop when an antagonist is expanding its geographic range (such as our study), studies along a chronosequence of introduction time with expected increasing accumulation of natural enemies over time, or comparisons between introduced plant populations that differ in exposure time to biocontrol organisms.     

Genetic structure of dictyoceratid sponge populations on the western Coral Sea reefs

Allozyme variation at six polymorphic loci was examined in foliose dictyoceratid sponges from isolated reefs in the western Coral Sea. Four major genetic groups corresponding to the species Phyllospongia lamellosa, P. alcicornis, Carterospongia flabellifera and Collospongia auris were examined. A further two rare morphotypes from individual reefs formed genetic outliers to the P. lamellosa group, and may represent further taxa related to P. lamellosa. Gene frequencies in individual reef populations were largely in Hardy-Weinberg equilibrium, suggesting that random mating occurred in local populations of all four common species. Genetic variability was high and observed heterozygosities within populations ranged from 0.13 to 0.40. All four taxa showed significant genetic differentiation among populations (F _ST=0.05 to 0.36). Genetic distances (Nei's D) among populations within species ranged from 0 to 0.723 and increased with increasing geographical separation. There was evidence that genetic differentiation between populations to the north and to the south of the southern limit of the South Equatorial Current (SEC) divergence was greater than expected on the basis of their geographical separation. The SEC divergence may form a partial barrier to gene flow among populations of these ecologically important sponges on the submerged Queensland Plateau. Levels of migration among populations of three of the species was less than those required to prevent divergence of the populations through genetic drift (Nm<1). Restricted migration among populations may provide a mechanism to explain the occurrence of highly divergent populations of dictyoceratid sponges whose specific identity is not clear, and may allow them additionally to develop partial reproduction isolation from other populations.     

Ecological commonalities among pelagic fishes: comparison of freshwater ciscoes and marine herring and sprat

Systematic comparisons of the ecology between functionally similar fish species from freshwater and marine aquatic systems are surprisingly rare. Here, we discuss commonalities and differences in evolutionary history, population genetics, reproduction and life history, ecological interactions, behavioural ecology and physiological ecology of temperate and Arctic freshwater coregonids (vendace and ciscoes, Coregonus spp.) and marine clupeids (herring, Clupea harengus, and sprat, Sprattus sprattus). We further elucidate potential effects of climate warming on these groups of fish based on the ecological features of coregonids and clupeids documented in the previous parts of the review. These freshwater and marine fishes share a surprisingly high number of similarities. Both groups are relatively short-lived, pelagic planktivorous fishes. The genetic differentiation of local populations is weak and seems to be in part correlated to an astonishing variability of spawning times. The discrete thermal window of each species influences habitat use, diel vertical migrations and supposedly also life history variations. Complex life cycles and preference for cool or cold water make all species vulnerable to the effects of global warming. It is suggested that future research on the functional interdependence between spawning time, life history characteristics, thermal windows and genetic differentiation may profit from a systematic comparison of the patterns found in either coregonids or clupeids.