Last semester in my Evolution class we learned about the Bateman-Trivers hypothesis, summed up by the curt phrase, "Eggs are expensive. that seahorses were sex role reversed (that females competed more intensely horses under a section on Sex role reversed species, TRIVERS (, p Genetic monogamy despite social promiscuity in the pot-bellied seahorse (Hippocampus Tagged: mating system, microsatellite genotyping, sex-role reversal.
Here, we demonstrate that the sex-role reversed seahorse has a single MH class II beta-chain gene and that the diversity of the seahorse MHIIβ. The male pregnancy of pipefishes and seahorses has led to the inference that females compete most intensely for access to mates, because males limit female. that seahorses were sex role reversed (that females competed more intensely horses under a section on Sex role reversed species, TRIVERS (, p
that seahorses were sex role reversed (that females competed more intensely horses under a section on Sex role reversed species, TRIVERS (, p "Sex role reversal also occurs in some other groups of birds, such as It also occurs in frogs and fish, like seahorses and pipefish," he told BBC. roles: polygamous species show reversed sex roles whereas emphasize that sex role reversal is not synonymous with seahorses), in which males alone.
Metrics details. Both natural and sexual selection are thought to influence genetic diversity, but the study of the relative importance of these two factors on ecologically-relevant traits has traditionally focused on species with conventional sex-roles, with male-male competition and female-based mate choice. Intrasexual competition and mate choice are both reversed in sex-role reversed species, and sex-related differences in the detection and use of MH-odor cues are expected to influence the intensity of sexual selection in such species.
The seahorse, Hippocampus abdominalishas an exceptionally highly developed form of male parental care, with female-female competition and male mate choice.
Whether such species have evolved the capacity to use MH-odor cues during mate choice is presently role investigated using mate choice experiments. If this possibility can be rejected, such systems would offer an exceptional opportunity to study the effects of natural selection in isolation, providing powerful comparative models for understanding the relative importance of selective factors in shaping patterns of genetic variation.
The impact seahorse natural and sexual selection on genetic diversity has been intensively studied in both natural and captive-bred populations [ 1 ], but the majority of our current knowledge in this area is derived from species with conventional sex roles, with choosy females and competitive males [ 23 ].
Sex-role reversed species, in which females compete for mating opportunities and males are choosy [ 45 ], offer exceptional opportunities to investigate central tenets of sexual selection theory and the importance of sexual selection in the maintenance of genetic diversity.
The MHC is an essential part of the vertebrate adaptive immune system, and seahorse a suite of more than genes involved in the destruction reversal infected cells and the antibody response [ 9 ]. There are two major antigen-presenting groups of MHC molecules, class I and class II genes, which differ in their function, structure and pattern of expression [ 9 ]. The peptide binding region PBR of MHC loci encodes a groove that permits the binding of seahorse antigens, and this region typically exhibits seahhorse highest sequence polymorphism within the gene [ 10 ].
The investigation of MHC genes in a diversity of vertebrates indicates that dex loci are more diverse than any other gene family [ 9 ]. Balancing selection operates through either negative frequency-dependent selection, in which the relative fitness of individual alleles is influenced by their frequency reviewed in [ 6 reversalor via heterozygote advantage. The advantage of MHC heterozygosity lies in the potential increase of the number of different parasite-derived antigens that can be detected by a MHC-heterozygous individual's immune system [ 12 ].
MHC diversity can be further enhanced by selection on linked genes, due to genetic hitchhiking [ 1314 ]. In addition to the importance of MHC genes as an integral part of the adaptive immune system, MHC-mediated odor cues have been shown to be important in mate choice, kin recognition and inbreeding avoidance [ 15 — 19 ].
Disassortative mating is widely believed to promote MHC diversity and to increase the proportion of heterozygote individuals in natural populations [ role2021 ]. Sexual selection can thus directly contribute to MHC allelic diversity via disassortative mate choice [ 12 ]. Despite consistently high levels of variation, there are major differences in the genomic organization of MHC genes in different vertebrate groups.
While these loci are physically linked in mammals, class Reversal and II genes are unlinked in bony fishes class Actinopterygii [ 2223 ]. Due to the lack of linkage of MHC reversal in actinopterygians, Stet et al. The unlinked nature of MH genes may provide increased evolutionary flexibility and contribute to enhanced MH diversity in this group.
This variation may be due, at least in part, to ancestral chromosome or genome duplications [ 29 ]. While previous studies on teleosts have shown that both natural and sexual selection structure MH allelic diversity in species with conventional female-based mate choice [ 163031 ], sdx study to date has investigated MH variation in sex-role reversed species in which mating decisions are made by the male.
Males and females often differ in their ability to detect odor cues [ 3233 ], and sex differences in the production, processing and use of MH-mediated seahorse are expected to influence the relative efficiency of sexual selection in sex-role reversed and conventionally-mating species, potentially reducing the level of MH variation in species with reversed sex-roles.
The teleost family Syngnathidae seahorses and pipefish is a well-suited model system to study questions concerning the relationship between sex roles and MH diversity. Both conventional and sex-role reversed species exist in the family and sex-role reversal has evolved several times reveersal in this group [ 34 ].
Studies of wild populations of the potbellied seahorse, Hippocampus abdominalishave found evidence of female-female competition and male mate choice, suggesting that natural populations of this species are sex-role reversed [ 35 ]. Here, we characterize MH-variation in wild-caught and captive-bred individuals of sex-role reversed populations reversal the potbellied seahorse, a species with a highly developed form of male parental care. This pattern of genetic variation has been influenced by a combination of intralocus recombination and positive role on sites seahorse to be important for peptide binding.
Our results indicate that sex-role reversed taxa such as the seahorse are capable of maintaining the high MHC diversity typical of vertebrate species with conventional sex roles. The total intron length of the 2 full-length alleles differs, resulting in full gene sequences of bp and bp, respectively.
Intron length variability is concentrated in 3 single-bp repetitive regions A nC n and T n located in introns 2 and 4 Figure 1. The locations of exons, repetitive regions A nC n and T n and primers used for genome walking and sequencing see xex 3 are indicated.
The 17 alleles include 25 polymorphic nucleotide sites and a total of 17 amino acid differences Figure 4. Each of the 17 alleles differs by at least one amino acid substitution Figure 45. The dataset used for subsequent analyses contains bp of exon 2 total length: bpafter omitting exon-spanning codons at the 5' and 3' ends of the exon 2 bp and 1 bp, respectively. Nucleotide alignment of exon 2. Dots indicate identity to the first sequence. Synonymous substitutions are shaded in grey.
Observed and expected allelic combinations. Bars represent allelic sex observed in H. The mean expected frequencies of allelic combinations were simulated based on empirical allele frequencies see figure inset. None of these outliers remain significant after a sequential Bonferroni correction for multiple comparisons.
Allele network of exon 2. Circle sizes reflect allele frequencies. The positions of individual non-synonymous substitutions separating sequences are indicated. Figure 5a: All 17 alleles. A model incorporating positive selection fits the exon 2 dataset significantly better than a neutral model of evolution M8 vs.
The network shows no clear spatial structure, consistent reversal the pattern expected for a single locus Seahorse 5a. The sex-role reversed potbellied seahorse, H. The results of targeted gene sequencing are congruent with a transcriptome screen which indicates that a role copy of this locus is expressed in muscle, liver and brood pouch tissue of the seahorse.
These studies have, however, focused on sex with conventional sex roles, with female-based seahorse choice and male-male competition e. Gasterosteus aculeatus [ 37 ], Oncorhynchus spp. As males and females often differ in their ability to detect olfactory cues [ 3233 ], the efficiency of odor-based Sex choice as a selective mechanism might be expected to differ between sex-role reversed and conventionally-mating species.
Disassortative mating sex thought to act together with pathogen-mediated selection to maintain MHC diversity [ 1520 ], and species which lack the ability to detect and process MHC-based odor cues seahosre thus expected to exhibit reduced levels of MHC diversity relative to species experiencing both forms of selection.
Alternatively, sex-role reversed species may indeed be capable of processing MH-based olfactory cues, something which is currently under investigation in targeted mate choice experiments in the seahorse. The seahorse carries seahorsd similar number of alleles Reversal. It is important to note, that maximal MHC diversity may also be esx, both by sex with the autoimmune response [ 4748 ] and by consistently high levels reverdal interlocus gene conversion, which may tend to homogenize genetic variation in species carrying multiple copies of these genes [ 49 ].
These factors may, in part, explain the lower than expected levels of MH variation detected in such species relative to species carrying a single copy of these genes.
In addition, sites exhibiting a high variability and signatures of positive selection rooe also putative candidates for peptide binding sites [ 85152 ]. We provide the first data on the pattern of MH diversity in the seahorse H.
The sex-role reversed H. This species has a single functional MH class II beta-chain gene that is expressed in the role brood pouch, suggesting that this gene may be immunologically active in these tissues. Mating experiments are currently being used to determine whether MH-odor cues are used in mate choice decisions deahorse H.
Whole genomic Rple was extracted from muscle tissue of a single H. Sequences were aligned in BioEdit v. Purification of digested DNA and adaptor ligation followed the Clontech protocol. Cloned products were compared to direct sequences generated with several different primer combinations, in order to identify allelic phase and to identify any cloning-mediated PCR artifacts.
To further explore this pattern, we screened seahorse libraries of seahorse pouch and reference tissues from pregnant and non-pregnant individuals for the presence of MH genes using sequencing. Briefly, both normalized and unnormalized cDNA libraries prepared from purified total RNA derived from the pouch tissues of a single pregnant and non-pregnant seahorse, together with a pool of normalized reference tissues from the pregnant individual brain, gills, liver, heart, kidney and testeswere individually MID-tagged with a unique sequence identifier.
Seahorses are listed under Appendix II of the United Nations Convention on the International Trade in Endangered Species CITESand the majority of the samples included here thus originate from a captive-bred population derived from xex collected from several sex-role reversed Tasmanian populations. The seahorses in this captive-bred population are held in large communal breeding tanks 2, L with 50 males and 50 females per tank, allowing free mate choice Hawkins R, pers.
In addition to 95 individuals from the swx population, we obtained complete exon 2 sequences from 5 wild-caught seahorses role Sydney, Australia 2 individuals collected in and Tasmania 3 individuals collected from 3 populations in and Sequencing results were tole using either primer combination data not shown. Sexx private haplotypes were sequenced in a minimum of 2 independent runs in order to reduce the possibility of PCR artifacts.
Degenerate positions in heterozygote sequences were scored using IUPAC nomenclature to facilitate the inference of allelic phase see below. Sequence data were assembled using Sequencing Analysis 5.
Sequences were aligned with Muscle v. To investigate the peptide binding region PBRwe analysed bp sequences of exon 2 total length: bp after omitting the first 2 nucleotides and sex seahorsw nucleotide of exon 2, to obtain a complete reading frame.
MH haplotypes of each individual were inferred from degenerate sequence data using a Feversal statistical method implemented in PHASE v. DnaSP v. To visualize relationships among the different exon 2 alleles and the non-synonymous substitutions separating them, a haplotype network was prepared using TCS v.
Tests for Hardy Weinberg equilibrium were performed in Genepop on the web [ 6566 ] using the default settings for the Markov Chain tole. The analysis of non-random associations of alleles was performed using non-parametric simulations 10, permutations reversa, incorporating empirical allele frequencies, with the Monte Carlo simulation function in PopTools v. Mega v. Codeml tests the goodness of fit of codon substitution models to a dataset using maximum likelihood.
A neighbor-joining tree was generated for the 17 exon 2 alleles using Neighbor v. We compared the fit of a role evolution model with recombination M7 with one allowing for positive selection M8using reversal likelihood-ratio test LRT.
More recently, Reche and Reinherz [ 50 ] presented an updated model of human PBS based on a larger sampling of potential peptides. In order to facilitate comparisons with previous studies, codons of the seahorse PBR were inferred through homology modeling to both of these datasets see Figure 4.
Given the more comprehensive dataset included in the Reche and Reinherz paper [ 50 ], PBS inferences in future studies should place greater emphasis on this work.
The identification of recombinant alleles with RECCO is based on role minimal cost solution, in which the relative seahorde of obtaining a sequence in an alignment from the other sex by mutation and recombination is evaluated.
Andersson M: Sexual selection. Reverwal M: Evolution of classical polyandry: Three steps to female emancipation. Behavioural Processes. Milinski M: The major histocompatibility complex, sexual selection, and mate choice. Annu Rev Ecol Evol Syst. J Fish Biol.
We provide the first data on the pattern of MH diversity in the seahorse H. The sex-role reversed H. This species has a single functional MH class II beta-chain gene that is expressed in the male brood pouch, suggesting that this gene may be immunologically active in these tissues. Mating experiments are currently being used to determine whether MH-odor cues are used in mate choice decisions in H. Whole genomic DNA was extracted from muscle tissue of a single H.
Sequences were aligned in BioEdit v. Purification of digested DNA and adaptor ligation followed the Clontech protocol. Cloned products were compared to direct sequences generated with several different primer combinations, in order to identify allelic phase and to identify any cloning-mediated PCR artifacts.
To further explore this pattern, we screened cDNA libraries of seahorse pouch and reference tissues from pregnant and non-pregnant individuals for the presence of MH genes using sequencing.
Briefly, both normalized and unnormalized cDNA libraries prepared from purified total RNA derived from the pouch tissues of a single pregnant and non-pregnant seahorse, together with a pool of normalized reference tissues from the pregnant individual brain, gills, liver, heart, kidney and testes , were individually MID-tagged with a unique sequence identifier. Seahorses are listed under Appendix II of the United Nations Convention on the International Trade in Endangered Species CITES , and the majority of the samples included here thus originate from a captive-bred population derived from individuals collected from several sex-role reversed Tasmanian populations.
The seahorses in this captive-bred population are held in large communal breeding tanks 2, L with 50 males and 50 females per tank, allowing free mate choice Hawkins R, pers. In addition to 95 individuals from the captive-bred population, we obtained complete exon 2 sequences from 5 wild-caught seahorses from Sydney, Australia 2 individuals collected in and Tasmania 3 individuals collected from 3 populations in and Sequencing results were identical using either primer combination data not shown.
All private haplotypes were sequenced in a minimum of 2 independent runs in order to reduce the possibility of PCR artifacts.
Degenerate positions in heterozygote sequences were scored using IUPAC nomenclature to facilitate the inference of allelic phase see below. Sequence data were assembled using Sequencing Analysis 5. Sequences were aligned with Muscle v. To investigate the peptide binding region PBR , we analysed bp sequences of exon 2 total length: bp after omitting the first 2 nucleotides and the final nucleotide of exon 2, to obtain a complete reading frame. MH haplotypes of each individual were inferred from degenerate sequence data using a Bayesian statistical method implemented in PHASE v.
DnaSP v. To visualize relationships among the different exon 2 alleles and the non-synonymous substitutions separating them, a haplotype network was prepared using TCS v.
Tests for Hardy Weinberg equilibrium were performed in Genepop on the web [ 65 , 66 ] using the default settings for the Markov Chain search. The analysis of non-random associations of alleles was performed using non-parametric simulations 10, permutations , incorporating empirical allele frequencies, with the Monte Carlo simulation function in PopTools v.
Mega v. Codeml tests the goodness of fit of codon substitution models to a dataset using maximum likelihood. A neighbor-joining tree was generated for the 17 exon 2 alleles using Neighbor v. We compared the fit of a neutral evolution model with recombination M7 with one allowing for positive selection M8 , using a likelihood-ratio test LRT. More recently, Reche and Reinherz [ 50 ] presented an updated model of human PBS based on a larger sampling of potential peptides.
In order to facilitate comparisons with previous studies, codons of the seahorse PBR were inferred through homology modeling to both of these datasets see Figure 4. Given the more comprehensive dataset included in the Reche and Reinherz paper [ 50 ], PBS inferences in future studies should place greater emphasis on this work.
The identification of recombinant alleles with RECCO is based on a minimal cost solution, in which the relative cost of obtaining a sequence in an alignment from the other sequences by mutation and recombination is evaluated. Andersson M: Sexual selection. Andersson M: Evolution of classical polyandry: Three steps to female emancipation. Behavioural Processes. Milinski M: The major histocompatibility complex, sexual selection, and mate choice. Annu Rev Ecol Evol Syst. J Fish Biol. Penn DJ: The scent of genetic compatibility: Sexual selection and the major histocompatibility complex.
Proc R Soc Biol. Amer Nat. Behav Ecol Sociobiol. J Evol Biol. Crit Rev Immunol. BMC Genomics. J Mol Evol. Mol Immunol. Annu Rev Genet. Prog Neurobiol. Olfaction and the Brain. Mol Ecol. Houde AE: Sex roles, ornaments, and evolutionary explanation. Conserv Genet. Ohta T: On the pattern of polymorphisms at major histocompatibility complex loci. Mol Mar Biol Biotechnol. J Mol Biol. Edited by: Hoelzel AR. Nucleic Acids Symp Ser. Bioinformatics methods and protocols: Methods in molecular biology.
Edited by: Krawetz S, Misener S. Nucl Acids Res. Stephens M, Donnelly P: A comparison of Bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet. Mol Ecol Resour. Mol Ecol Notes. Genepop on the Web. Raymond M, Rousset F: Genepop version 1. J Hered. PopTools version 3. Mol Biol Evol. Distributed by the author. Download references. We thank L. Keller and C. Wedekind for helpful discussions during the preparation of this manuscript.
We are grateful to K. Funding for this work has been provided by the University of Zurich. Correspondence to Anthony B Wilson. AB participated in the design of the study, carried out the laboratory work and data analysis and wrote the manuscript. ABW conceived the study, supervised the laboratory work and data analysis and helped to draft the manuscript. Both authors read and approved the final manuscript. This article is published under license to BioMed Central Ltd.
Reprints and Permissions. Bahr, A. The impact of sex-role reversal on the diversity of the major histocompatibility complex: Insights from the seahorse Hippocampus abdominalis.
BMC Evol Biol 11, doi Download citation. Search all BMC articles Search. Here we are concerned with the genetic mating system rather than the social mating system Andersson ; Emlen and Oring ; Searcy and Yasukawa These two views of the mating system must be connected to some degree, of course, but whereas the social mating system is concerned with pair bonds and other observable interactions among individuals, the genetic mating system is concerned with the allocation of biological parentage.
In principle, a complete description of the genetic mating system for a particular breeding assemblage would provide data on the number of mates and the number of offspring produced by each reproductively active adult during a well-defined breeding interval.
The genetic mating system is undoubtedly related to the operation of sexual selection and, indeed, should be seen as an integral component of the process of sexual selection. Darwin originally defined sexual selection as differential reproductive success of individuals caused by competition over mates Andersson Hence the variance in fitness due to sexual selection comes mainly from differential mating success although other factors such as mate quality also may contribute; Andersson A description of the genetic mating system includes data regarding both mating and reproductive success, and therefore describes the majority of the variance in fitness resulting from sexual selection Arnold Given the importance of the mating system to sexual selection, a more precise understanding of mating systems should be informative with regard to sexual selection Arnold and Duvall The above argument applies only to the term genetic mating system as it is defined here.
Broader uses of the term mating system typically do not possess so clear a relationship to sexual selection Andersson ; Emlen and Oring For species with typical gender roles, Darwin observed that highly polygynous species tend to be more sexually dimorphic than less polygynous ones, with secondary sexual characters evolving in males Table 1. Numerous attempts to quantify this relationship have yielded mixed results e. However, even now that scores of studies have quantified rates of extrapair fertilization notably in birds , the data have yet to be integrated in ways that definitively address relationships between the genetic mating system and the strength of sexual selection.
While these issues are being sorted out in species with typical sex roles, we can also examine the predictions for species with reversed sex roles. In such organisms, stronger sexual selection is expected in species that are more polyandrous, with secondary sexual characters evolving in females Jehl and Murray Thus the expectations regarding the strength of sexual selection for polygyny, polyandry, and monogamy are fairly simple Table 1.
Polygynandry is more problematic because sexual selection under this mating system may affect either sex or may be nonexistent, depending on the relative distributions of mating success for each gender. Sexual selection in polygynandrous mating systems may occur, but in general is expected to be weaker than in strictly polygynous or polyandrous populations because polygynandry will tend to reduce the ability of successful individuals to monopolize access to mates.
These simple hypotheses have motivated our genetic studies of mating systems in pipefishes and seahorses. Various species can be ranked a priori with respect to level of sexual dimorphism. Then the hypothesis to be tested is whether differences among species in sexual dimorphism are related to the underlying genetic mating systems, perhaps with the most dimorphic species also being the most polyandrous Table 1.
Thus individuals in a population often exhibit unique DNA profiles, parental genotypes can be reconstructed from progeny array data, and the markers can be assayed from small tissue sources such as individual embryos. One drawback of microsatellites is that the primers often are species or genus specific, and we have had to generate primers independently on several occasions for the three species of pipefish genus Syngnathus and one seahorse Hippocampus subelongatus assayed to date Table 2.
High levels of microsatellite polymorphism were observed for all species, with 9—44 alleles per locus and observed heterozygosities ranging from 0. The power of these marker systems for parentage assessment can be gauged by the four-locus exclusion probabilities Chakraborty et al. Pipefish and seahorse primer sequences are available in Jones et al. The expected proportion of individuals in a population that would be rejected as possible parents of a typical offspring, given that one of the offspring's parents is known.
Available evidence suggests that most or all individuals even in large syngnathid populations have unique multilocus genotypes. For example, calculated expected frequencies of observed four-locus genotypes across all species ranged from 3.
We never observed two individuals that shared even a two-locus genotype in population samples of adult S. Three pairs of individuals with identical two-locus genotypes were observed in a sample of adult S. These considerations indicate that the microsatellite markers were more than adequate for the parentage assessments.
In fact, simulations by Jones et al. With this level of genetic variation, multiple maternity is difficult to detect only when reproductive skew between two mates is high and the number of offspring sampled from a brood is low Figure 1 , but even in this case the failing is not in the markers but in the sampling strategy Jones et al.
Results of computer simulations investigating the power of two seahorse microsatellite loci Han03 and Han05 to detect multiple maternity within a male's pouch after Jones et al. Each simulated brood contained embryos that were the progeny of two mothers. Samples of 15, 30, or 60 embryos were drawn at random from the brood and if three or more maternal alleles appeared in the progeny at either locus, then multiple maternity was deemed to have been detected.
Almost all failures to detect multiple mating by males were because embryos from one of the females went unsampled rather than because of insufficient variation at the marker loci. Beyond the mere detection of multiple mating, the particular genotypes of parents contributing to a brood also can be reconstructed. Pipefish and seahorses are convenient in this regard because a pregnant male carries an entire brood that may include hundreds of offspring. Once the male is established as the true genetic sire of all progeny within his pouch see below , the maternal contribution to each embryo can be deduced by subtraction.
Then, with the maternal alleles known for each embryo, a table can be constructed that reveals the unique dilocus genotype of each mother. An example from S. In this case, the male evidently mated with two females.
Positions of the embryos were mapped during dissection of the brood pouch, and the reconstituted maternal genotypes were projected onto this map. This example reveals another convenient feature of Syngnathus brooding: In each of 43 multimated pipefish males genetically assayed, embryos invariably proved to be clumped by maternity within the brood pouch, a finding that lends additional support to the validity of the reconstructed maternal genotypes.
By examining a pregnant male and the embryos in his brood pouch, maternal alleles in each embryo are deduced by subtraction. Then a dilocus table is constructed, each cell of which reports the number of progeny with a particular combination of alleles at the two loci. The maternal dilocus genotype is evident from these nonrandom allelic associations. The diagram at right shows a spatial map of these maternal alleles in the male's brood pouch.
In this case, the pouch was divided into 14 sections, with three embryos sampled from each. Embryos of the same pattern are full sibs and those of different patterns are half sibs had different mothers. Deduced maternal genotypes also can be used to document instances of multiple mating by females, as well as to match particular collected specimens of adult females to their mates.
If the genotype of an assayed female is the same as the inferred genotype from a male's brood, then given the high levels of polymorphism at the microsatellite loci see above , that particular female most likely mothered the brood. Similarly if the same multilocus maternal genotype appeared in the progeny of more than one male, then it is highly likely that the identified female had mated with multiple males.
In this way, multiple mating by females can be deduced solely from the contents of male brood pouches—the adult females need not be available for assay. Pipefish in the genus Syngnathus possess a pouch with a ventral seam that, when closed, fully covers all embryos see Figure 3. During courtship in S. Males release sperm directly into the pouch and then shake the zygotes posteriorly to make room for additional eggs. The ventral seam opens only when the male gives birth several weeks later.
Seahorse males possess a pouch that is fully enclosed except for a small opening through which females deposit eggs, and the sequence of egg transfer resembles that of pipefish.
Thus in these species it is difficult to imagine how rival males might steal any fertilizations. Diagrammatic representation of the patterns of parentage observed during microsatellite-based parentage studies of four syngnathid species. Males are shown above the brood pouch diagrams, females below. Notice the brood pouches on the ventral surfaces of depicted males. Also shown are spatial maps of full- and half- sib embryos as in Figure 2 within males' pouches.
These diagrams are intended to present a comparative overview of the patterns of mating in these species. Therefore they are not drawn to scale and only portray a small subset of the assayed progeny per pouch. Arrows indicate patterns of egg depositions matings by females. Males of S. As described in the text, available data indicate a positive relationship between polyandry and the evolution of sexual dimorphism in these taxa.
Indeed, every pregnant syngnathid male whose brood has been genetically assayed to date has proved to be the true genetic father of all embryos within his pouch.
Thus males with enclosed pouches appear to have complete confidence of paternity, a result consistent with expectations based on pouch configuration and parental investment theory, which predicts that a male's devotion to offspring should be positively related to paternity confidence Clutton-Brock However, not all species of pipefish have enclosed pouches.
In Nerophis ophidion , unfertilized eggs are loosely attached to the male's outer ventral surface where he fertilizes them by floating through a cloud of sperm that he releases into the water Rosenqvist Of interest, males of this species invest less energy per zygote than females, whereas in S.
These observations raise two interesting questions for future research: Do syngnathid species without enclosed pouches have complete confidence of paternity? And, did the brood pouch evolve as a structure to nurture and protect embryos, or as a barrier to prevent cuckoldry? For each syngnathid species genetically surveyed, field collections of pregnant males, their brooded embryos, and a sample of adult females were analyzed for microsatellite genotypes.
Here we summarize the results species by species. This North American coastal species is the most sexually dimorphic of the syngnathids genetically examined thus far. Females are larger than males and possess a deeply keeled abdomen, silvery stripes on the trunk, and an enlarged dorsal fin Figure 3 that probably are the result of sexual selection Brown These characteristics are absent in other Syngnathus species, and hence S.
Jones and Avise a collected specimens from one locale on Florida's northern Gulf coast. Despite high statistical power, the molecular appraisal found evidence for multiple maternity in only one brood among the 40 pregnant males assayed Table 3 , Figure 3.
On the other hand, the genetic data documented multiple mating by females Table 3 , Figure 3. On two separate occasions, a single deduced female was the mother of embryos in more than one male's pouch. Thus the genetic data showed that females are able to split a batch of eggs among multiple males, and that multiple mating by females was common in this population. The logic behind the latter statement stems from the fact that only a small fraction of the breeding population was sampled from an extensive sea grass meadow.
Thus females probably deposited eggs in the pouches of males not sampled, as well as in the males documented in our sample. Given the low power to detect multiple mating by a particular female, the fact that we detected the phenomenon at all suggests that multiple mating by females was common.
Overall the Gulf pipefish mating system is best described as polyandrous, with a male typically receiving eggs from only one female, and a female often mating with multiple males during the time of male pregnancy. Relevant reproductive information on three pipefish species genus Syngnathus and one seahorse species genus Hippocampus that have been the subject of molecular parentage analyses.
This western Atlantic species also is sexually dimorphic, but to a much lesser extent than the Gulf pipefish Brown Mature females are larger than males and possess a slightly deeper and more yellow abdomen, probably due to the presence of ripe ova Brown This species does not possess obvious secondary sexual characters.
In a genetic analysis of brooded embryos in pregnant dusky pipefish Jones and Avise b , males proved to have mated with one to three females each Table 3 , Figure 3. In addition, there was clear evidence for multiple mating by females. Quantitative estimates of the breeding population size and multiple maternity demonstrated that the rate of multiple mating by females of dusky pipefish is probably quite high see Jones and Avise b.
Because both males and females often have multiple successful partners during a single male pregnancy, this species is genetically polygynandrous. This western European species is sexually dimorphic, but less so than Gulf pipefish. Females are larger than males on average and have slightly deeper abdomens. They also possess a transient ornament that becomes visible during courtship and has been shown to be a target of sexual selection Berglund et al.
Nevertheless, the sexual dimorphism displayed by S. A population from Sweden's west coast was the focus of the genetic parentage assessment.
Ninety percent of the males carried broods with two to six mothers each, indicating a much higher rate of multiple mating than in previously studied syngnathid species Jones et al. As in the Gulf and dusky pipefishes, multiple mating by females also was documented. These results are consistent with laboratory and field observations suggesting that both males and females frequently have multiple mates Berglund et al. Thus S. Motivation for the study of seahorse parentage was twofold.
First, seahorses tend to be the least sexually dimorphic of all syngnathids, so the mating system might reflect an apparently weak intensity of sexual selection Vincent et al. Second, several seahorse species studied intently from a behavioral standpoint are socially monogamous, with long-term pair bonding Masonjones and Lewis ; Vincent b , ; Vincent et al.
The species H. Females and males do differ slightly in body proportions and in the frequency of body color morphs Table 3 ; Kvarnemo C and Moore GI, unpublished data , but so far no evidence suggests that these differences are related to sexual selection.
Jones et al. A population near Perth, Australia, is the subject of ongoing study. Furthermore, in contrast to the previous studies of pipefish mating systems, no evidence was found for multiple mating by females. Thus molecular evidence supports the notion that H. An additional question motivated by the unusual social system of seahorses was whether males mate with the same female repeatedly within a breeding season.
Field and laboratory observations suggest that other species of seahorses do so Vincent et al. Kvarnemo et al.
Of 14 males successfully sampled more than once, the genetic data document that eight remained faithful to their original mates and six changed mates between broods. Males who switched mates moved much greater distances during the breeding season and displayed a somewhat longer interval marginally significant between the production of successive broods.
Overall, results indicate that seahorses frequently maintain long-term pair bonds throughout the breeding season, and investigation into the costs associated with mate switching should be a fruitful area for future investigation. The parentage analyses described above can be considered in light of hypotheses relating the genetic mating system to the evolution of sexual dimorphism. The Gulf pipefish is polyandrous and also is the most sexually dimorphic species studied to date.
The western Australian seahorse is monogamous with long-term pair bonds, a situation that should reflect weak sexual selection, and this species displays very little sexual dimorphism.
The two other Syngnathus species are polygynandrous and they exhibit intermediate levels of sexual dimorphism. Thus available data indicate a positive correlation between the strength of sexual selection and the extent of polyandry in sex-role-reversed pipefishes. This comparative result is based on observations of only four species, however, so it must be considered provisional.
Nevertheless, these genetic studies provide important baseline information for future comparative analyses of syngnathid mating systems. Additional support for a connection between sexual selection and the mating system comes from observational studies on other syngnathid species. For example, the seahorses H. Two pipefish species, Hippichthys penicillus and Corythoichthys intestinalis , appear to be monogamous as well Gronell ; Watanabe et al.
Within Syngnathus , two other pipefish species S. Neither species displays obvious secondary sexual traits, but S. Finally, the pipefish N. These observations should be interpreted with caution until they can be verified in nature using molecular markers, but current data appear to bolster the conclusion that more polyandrous species have experienced stronger sexual selection on females.
If the genetic mating system is indeed causally related to the evolution of sexual dimorphism, an important goal is to determine the ecological factors and evolutionary constraints that may shape the mating system. Although too little is known about syngnathid ecology to rigorously address causal hypotheses about mating system evolution in this group, some potentially important differences among species are listed in Table 3.
For example, S. These conditions might promote a bet-hedging strategy in S. In addition, female pipefish in the genus Syngnathus mature eggs continuously and appear able to hold them until a mate becomes available. In contrast, a seahorse female prepares a batch of eggs all at once and, if not to be in vain, must deposit them in a male's pouch within a short time Vincent a. This feature of seahorse reproduction is certainly related to the evolution of monogamy, but whether it is a cause or consequence remains unknown.
Additional factors possibly related to the evolution of mating systems in syngnathids include site fidelity Table 3 , male brood size Table 3 , variation in mate quality Jones et al. That a relationship between the genetic mating system and the evolution of sexual dimorphism is clear for the syngnathid species studied thus far may be due to the fact that the assayed species covered a wide spectrum of mating systems, from polyandry to monogamy.
Qualitative definitions of the mating system Table 1 are useful verbal summaries, but they are inadequate for detailed quantitative comparisons. A major shortcoming is that polygynandry, probably the most common mating system in animal populations, is an overly broad category.
Expectations for the strength of sexual selection under this system are not always clear, and the borders between polygynandry and other mating systems can be indistinct. Thus a major goal of research on syngnathid mating systems has been to test other methods of mating system quantification with respect to sexual selection.
Arnold and Duvall present a new theoretical framework that considers mating systems in the context of formal selection theory. This integration points to the relationship between mating success the number of times an individual mates and reproductive success the number of offspring produced as a central feature of the genetic mating system with respect to sexual selection.
In his classic article, A. Bateman perceived the difference in this relationship in males compared to females as the cause of sexual selection. In species with typical sex roles, such as Drosophila melanogaster , males exhibit a strong positive relationship between mating success and reproductive success Bateman , whereas in females this relationship is weaker Figure 4. Thus a male benefits greatly in terms of offspring number from having multiple mates, whereas a female does not, a situation that should promote stronger sexual selection in males than in females.
The relationship between mating success and reproductive success can be quantified and compared statistically using a least-squares regression approach Arnold and Duvall The resulting line is referred to as the sexual selection gradient Arnold and Duvall , or Bateman gradient Andersson and Iwasa Sexual selection Bateman gradients for Drosophila melanogaster top and S.
The sexual selection gradient is given by a weighted least squares regression line relating reproductive success to mating success. Each symbol in the graph represents a mean with one standard error for S. Males are shown as open squares, females as closed circles. In Drosophila , males and females exhibit sexual selection gradients that are steep and shallow, respectively, a pattern consistent with strong sexual selection on males.
One simple conceptualization of the sexual selection gradient's position in the sexual selection process is presented in Figure 5. The operational sex ratio, potential reproductive rates of the sexes, and relative parental investment, among other factors, no doubt interact in complex ways to affect the strength of sexual selection Andersson , and they typically do so by influencing patterns of mating success.
In the context of sexual selection, variance in mating success can contribute to variance in fitness only through the sexual selection gradient Arnold and Duvall Thus if the gradient is zero, variance in mating success has no effect on the strength of sexual selection and there can be no sexual selection due to differential mating success.
However, if the gradient is steep, very strong sexual selection is possible. Under this view the sexual selection gradient concept does not diminish the importance of other methods related to the measurement of sexual selection, but rather should be seen as a quantitative framework that can be used to sort out the relative importance of the multifarious factors that contribute to sexual selection.
Path diagram showing the vital role of the sexual selection gradient in the sexual selection process. Traditional measures or correlates of sexual selection remain important, but the sexual selection gradient is the final path to fitness for all sexually selected traits Arnold and Duvall Pipefish provide a unique opportunity to test the validity of the sexual selection gradient concept.
In sex-role-reversed species, a steep sexual selection gradient is expected in females and a shallow gradient in males. Microsatellite markers were used to describe parentage completely for small pipefish breeding groups, allowing the sexual selection gradients of the sexes to be quantified. Results of the experiment were consistent with expectations based on the Bateman gradient model: The gradient of females was significantly steeper than that of males Figure 4.
In addition, the slopes of the lines responded as predicted to changes in the operational sex ratio. These results lend additional support to the Bateman gradient concept, suggesting that sexual selection gradients do provide a superior method for the quantification of mating systems with respect to the strength of sexual selection. Thus a goal in future studies of sexual selection in Syngnathidae, as well as in other taxa, should be to quantify Bateman gradients of natural populations in a comparative context.
Microsatellite-based studies of parentage in pipefishes and seahorses have also provided novel perspectives on the evolution of the markers themselves, and on the broader use of molecular parentage analyses in population studies. The following sections will highlight three such applications.
In a large study of parentage involving S. The pattern of mutations conformed well to a stepwise mutation model Shriver et al. Another interesting result of the study was that mutations often occurred in clusters.
Combining data from both loci, a total of 35 mutant progeny were assayed, but they apparently were produced by only 26 independent mutation events Jones et al.
Thus multiple offspring in a single brood sometimes received the same mutant allele, a pattern indicative of multiple gametes carrying copies of one premeiotic mutation that arose in a parent's germline Woodruff and Thompson