Host-driven adaptive radiation and host usage in sponge-associated barnacles Authors: Meng-Chen Yu, Hsiu-Chin Lin, Gregory A Kolbasov, Jens T Høeg, Benny K K Chan Published: 2026 ABSTRACT AbstractBackgroundAdaptive radiation often involves ecological and morphological diversification, influenced by environmental opportunities, such as symbiotic interactions. Sponge-associated barnacles have evolved specialized traits, such as modified shells, to thrive within sponge hosts. However, the evolutionary mechanisms driving their diversification and host specificity remain unclear.ResultsWe integrate molecular phylogenetics (12S, 16S, 18S, COI, H3), detailed morphology, and ecological assessment to examine how host usage influences sponge-associated barnacle diversification. Our findings reveal two primary barnacle clades: Membranous-Base (M-Base) and Calcareous-Base (C-Base), each exhibiting distinct morphological adaptations. Most of the species exhibit strong host specificity being specialists within particular sponges. However, there is no significant correlation between morphology and host specificity, nor a clear co-evolutionary pattern with sponge hosts. These results challenge the assumption that host-driven specialization alone explains diversification.ConclusionsOur findings suggest that sponge-associated barnacles show a pattern of host-associated divergence, with generalist species occupying basal positions and specialization evolving from ancestrally broader host-use states. The last common ancestor of sponge barnacles appears to have been associated with Clionaida sponges, which may serve as permissive hosts due to their lower biochemical diversity.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12915-026-02618-3.Keywords: Co-evolution, Coral reef symbiosis, Diversity, Host-driven morphology, Macro-organism symbiosis Background Adaptive radiation is the evolution of species resulting from ecological diversification within a rapidly multiplying lineage [1–3]. Radiations occur mainly when species evolve innovative adaptations to occupy new ecological niches [4–6]. One classic and famous example of adaptive diversification of species with phenotypic variation is Darwin’s finches, which diversified from a common ancestor by adapting to different environments and feeding habits [7, 8]. Similar mechanisms of adaptive radiation have also been observed in Hawaiian honeycreeper birds [9], African cichlid fishes [10, 11] and plants in the Hawaiian silversword alliance [12–14]. Organisms can undergo adaptive radiation when ecological opportunities are open for exploitation, provided that they possess the requisite phenotypic plasticity to survive under environmental change. A symbiotic relationship is an ecological context and selective pressure that promotes adaptive radiation and accelerates major processes such as morphological and molecular diversification, and speciation within ecosystems [15]. Hosts can provide ecological opportunities for symbionts and may enhance interaction between species [5, 6, 15]. Co-evolution occurs when two or more species influence each other's evolution through natural selection, leading to interactions that can be mutualistic or antagonistic as in parasitism [1, 16–19]. In terrestrial ecosystems, a common example involves plants and herbivores, where plants produce secondary metabolites to reduce grazing [20–24], while some herbivores evolve resistance and can sequester these compounds for defense against predators [25, 26]. Sponges (Porifera) are, with approximately 10,000 known species, functionally important members of marine benthic fauna and house within and outside their bodies, a great diversity of associated organisms, including microbiota, mollusks, annelids, nemerteans, echinoderms and crustaceans [27–31]. Complex microbial communities often show strong phylo-symbiosis, co-phylogeny and co-evolutionary interactions with their host sponges [32, 33] and produce secondary metabolites that defend the host sponge from predators [33–38]. Sponge-associated invertebrates often adapt to these metabolites and some have evolved specific morphological adaptations that allow them to live within the sponge environment [39–41]. Host specificity in sponge-associated organisms ranges from generalists to specialists [42, 43]. Generalists undergo speciation while specialists partition the habitat into narrow niches [44, 45]. Due to drug development based on natural products, sponge microbiomes are well understood [46] whereas the diverse sponge-associated macro fauna have received only scant attention. Although free-living barnacles possess physical protection through their calcareous shells, this protection is often insufficient against shell-crushing predators such as crabs and certain fishes. This vulnerability is particularly pronounced in small barnacles, including newly settled and recently metamorphosed juveniles. Barnacles are common sponge associates. These highly specialized barnacles, with reduced external shell plates, embed their shells within the sponge, from whence they extend their cirri to suspension feed on zooplankton (Fig. 1a–b). Sponges and their sponge-associated barnacles appear to have mutualistic relationships. Barnacles gain protection from predation by living inside the sponge. In turn, the incorporated barnacles can strengthen the sponge skeletal structure and beating of the barnacle cirral fans generates currents that can enhance filter-feeding efficiency of the sponge, removing particles or sediment fouling its surface [47–50]. Settlement and metamorphosis of sponge-barnacle cyprids involves embedding into the sponges. Consequently, the antennular attachment organ of the cyprids is found to vary according to sponge host characteristics [50] and differ from free-living forms. Diversity of morphological characteristics and ancestral state reconstruction of selected morphological traits in sponge-associated barnacles, mapped onto a phylogenetic tree reconstructed using five genetic markers (12S, 16S, 18S, COI, and H3). a Live sponge in situ with embedded sponge-associated barnacles; b sponge partially removed to expose the position of barnacle; c isolated sponge-associated barnacles showing high variation in shell morphology; d schematic drawing showing variety of morphological characteristics of sponge barnacle shells and cirri; e–g three selected morphological traits for ancestral state reconstruction: basis structure, shell plate connection, and cirral armature area. R-A, anterior ramus, R-P, posterior ramus, R-S, ramal segment, Pr-A coxa, coxa of the anterior (A) protopod; Pr-P coxa, coxa of the posterior (P) protopod; Pr coxa, lateral coxa of the protopod (neither anterior nor posterior). All photographs taken by MCYu Shell shapes and bases in adult sponge-associated barnacles also vary in structure, being highly affected by the host sponge (Fig. 1c–d) [51–55]. Several hypotheses regarding the evolution of sponge-associated barnacles have been proposed in the past (Additional file 1: Fig. S1). Among them, Kolbasov [51] suggested that ancestors of sponge barnacles probably shared general structures with free-living barnacles of the archaeobalanid type, including a flat and membranous basis. The derived taxa evolved specialized structures for living with sponges, including a cup-shaped calcareous basis. Even more specialized species have reduced shell plate armature, resulting in so-called ‘windows’ of thin membrane spanning between the plates, where barnacle and sponge are in closed contact. This might enable direct communication and transfer of substances between sponge and barnacle, but whether this is the case has yet to be explored. Morphologically based taxonomy [51, 53, 56–60] classifies sponge-associated barnacles into four subfamilies, comprising the Acastinae Kolbasov, 1993 (84 spp.), the Bryozobiinae Ross and Newman, 1996 (11 spp.), and the genera Membranobalanus Hoek, 1913 (12 spp. in the Archaeobalaninae Newman and Ross, 1976) and Pyrgospongia Achituv and Simon-Blecher, 2006 (1 sp. in the Pyrgomatinae Gray, 1825). The Bryozobiinae and Acastinae are both speciose families, with the former living exclusively in sponges, while the latter also have some species associated with “gorgonian” corals [61]. Recent molecular phylogenetic studies of sponge-associated barnacles suffer from not including all relevant taxa [52, 54]. In order to achieve a full view of how barnacles evolved into sponge associates, such studies must include important sponge-associated genera such as Pectinoacasta Kolbasov, 1993, Membranobalanus, Multatria Van Syoc and Newman, 2010 and Pyrgospongia, which have not yet been studied genetically. It remains unknown to what extent morphological specialization of sponge barnacles relates to host specificity and how barnacle speciation relates to sponge phylogeny. Any advances concerning these problems must involve accurate and molecular species identification of both host sponges and their associated barnacles and the construction of robust phylogenies on which to map evolutionary events [62–64]. Here, we use multiple markers (12S, 16S, and COI and two nuclear 18S and H3 genes) to analyze the phylogeny of 34 species of sponge-associated barnacles from 8 different genera (including subfamilies Acastinae and Bryozobiinae and species in Membranobalanus (Archaeobalaninae) and Pyrgospongia (Pyrgomatidae), with their sponge hosts accurately identified using both morphological and molecular techniques. We test the following hypotheses: Host-driven morphological specialization hypothesis The morphological specialization of sponge-associated barnacles is shaped by their host sponges. Generalists retaining ancestral traits, and specialists evolving structural adaptations specific to their hosts are found in a clear pattern within sponge barnacle phylogeny. Co-evolutionary diversification hypothesis The specialization and host specificity of sponge-associated barnacles follow a co-evolutionary process. Generalist species occupy basal positions, have large populations that contribute significantly to diversity, while specialists with smaller populations contribute less. Accordingly, barnacle and host phylogenies show congruent co-evolutionary patterns. Results Origins and evolution of sponge-associations in barnacles We concatenated the sequences from five markers (12S, 16S, 18S, COI, and H3; Additional file 1: Figs. S1–S5) to obtain a high-resolution phylogenetic tree, covering all subfamilies of sponge barnacles. In both Bayesian inference (BI) and maximum likelihood (ML) analyses, datasets yielded cladograms of a similar topology (Additional file 1: Figs. S6–S9). In our analysis, sponge-associated barnacles were divided into two strongly supported clades. The first major clade consisted of the genera Membranobalanus, Archiacasta Kolbasov, 1993 and Pyrgospongia, which all possess a membranous basis (M-Base clade). The remaining genera, all having a calcareous base (Acasta, Euacasta Kolbasov, 1993, Neoacasta Kolbasov, 1993, Pectinoacasta and Bryozobia Ross & Newman, 1996), formed the second major clade (C-Base clade) (Figs. 1e–g, 2, and 3a; Additional file 1: Figs. S6, S7, S9). Co-phylogenetic analysis of sponge barnacle species and their host sponge orders (right, modified from Plese et al., 2021) with calibration time (Mya). Arrows a and b denote barnacle M-Base and C-Base clades, respectively; arrows c and d indicate the sponge orders Dictyoceratida and Haplosclerida a Ancestral state reconstruction of host spectra of sponge barnacles, showing basic specificity at the family and genus levels, and taxonomic specificity (STD). b Occurrence of barnacle species in six species of host sponges, where 40 or more individuals of each sponge species were sampled. Inset pie charts show the prevalence of barnacle individuals (percent of total barnacles collected) by species for each sponge species Within the M-Base clade, Membranobalanus was paraphyletic, since both Archiacasta pustulata Kolbasov, 1993 and Pyrgospongia stellula (Rosell, 1973) were nested among Membranobalanus species with high support by both ML and BI analyses. The Acastinae were clearly polyphyletic, with their genera widely dispersed in the C-Base clade, while A. pustulata was within the M-Base clade in our tree. Moreover, the acastine genera, Acasta, Euacasta and Pectinoacasta were also polyphyletically distributed in the C-Base clade. Surprisingly the genus Bryozobia, which belongs to the unrelated subfamily Bryozobiinae, is also nestled among “acastines” in the C-Base clade. Thus, our analysis cannot support the current morphology-based taxonomy of sponge barnacles (Additional file 1: Figs. S6–S9). To investigate phylogenetic trends, we analyzed the evolution of ten morphological characters distributed within three main traits: basis structure, shell plate connections and cirral armature (Fig. 1e–g, the remaining seven morphological characters are shown in Additional file 1: Fig. S10–S17. The 10 characters include: (1) basis structure, (2) shell plate connection, (3) cirral armature area, (4) external structure on shell plates, (5) basis form, (6) aspect ratio of the scutum, (7) external structure of the scutum, (8) spur furrow of the tergum, (9) teeth on the labrum, and (10) basidorsal point of penis.). Ancestral state reconstructions (ASR) revealed that the most recent common ancestor (MRCA) of all sampled sponge-associated barnacles is ambiguous with respect to most morphological traits (Fig. 1e–g; Additional file 1: Figs. S11–S17). Only “shell plate connection” mapped consistently with a “slitless connection” representing the state of the MRCA. In contrast, both the M-Base and C-Base clades had clear patterns of shell evolution, with a membranous base being ancestral for the M-Base clade and a calcareous one for the C-Base clade, whence our informal names. The character “shell plate connections” also mapped consistently with a “slitless connection” being ancestral both for all barnacles and the two major clades (Fig. 1e–g). Our analysis showed that multiple independent events have occurred in character state evolution, without any consistent direction of change. This suggests that detailed morphology of sponge barnacles (at least for those characters being used in their systematics) does not reveal clear phylogenetic relationships or trends, but that morphology evolved by ad hoc adaptive evolution to specific host sponges (Fig. 1e–g; Additional file 1: Figs. S11–S17). Adaptive evolution of barnacles to their host sponges In our ASR, host usage for the MRCA of all sponge barnacles is ambiguous, but with highest posterior probability (PP) for the sponge order Clionaida (PP: 0.57) (Fig. 2). In more recent nodes the pattern becomes clearer. The MRCA of the M-Base clade (indicated by a with arrow in Fig. 2) is highly specific for the Clionaida (PP: 0.97), whereas the MRCA of the C-Base clade (indicated by b with arrow in Fig. 2) has an ambiguous state since host usage changed several times to associate with the sponge orders Dictyoceratida, Haplosclerida, and Poecilosclerida. Acasta cyathus, the sister taxon to all remaining C-Base clade forms, resolves as a generalist utilizing a multitude of sponge taxa. At the next higher split, one branch is predominantly (but not exclusively) found on Dictyoceratida (indicated by c with arrow in Fig. 2). The MRCA of the sister-branch was with high probability originally hosted by the Haplosclerida (indicated by d with arrow in Fig. 2) but then evolved multiple times with other sponge taxa. Comparison of the molecularly based phylogenies of the sponge barnacles and their hosts failed to reveal any clear relationship, the two trees being highly incongruent (Fig. 2; Additional file 1: Fig. S18) [65]. According to the sponge host phylogenetic tree, the class Demospongiae is estimated to have diverged from other Porifera approximately 529 million years ago (Mya) in the early Paleozoic [65]. Sponge associated barnacles diverged from other balanoid barnacles around 38 Mya in the late Eocene (87% and 1.00 branch support from ML bootstrapping and PP, respectively for these two date estimates; Additional file 1: Fig. S18). Most of the early diverging sponge-associated barnacles are hosted by the orders Clionaida and Dictyoceratida, which diverged from each other 231 and 90 Mya, respectively, in the Mesozoic (Fig. 2; Additional file 1: Fig. S18) [65]. This mismatch of divergence times and tree topologies fails to yield evidence that sponge barnacles and their hosts have a coordinated co-evolutionary relationship. Barnacle host specificity and prevalence in sponges For barnacle species with known host association, the number of sponge taxa utilized as hosts ranges from one to nine orders, one to 17 families and one to 22 genera (Figs. 2 and 3a). The value of taxonomic specificity (STD) ranges from 1 to 4.67. Most barnacle species have a high host specificity (17 species with STD < 1.6). These include those specific to a single sponge species (STD = N/A) and those specific to species from the same genus (Fig. 3a; Additional file 1: Tables S1–S2). There are few species with an STD > 4, a measure indicating that they are generalists and can live in hosts from different sponge orders or even subclasses [66–68]. Our analysis of host specificity suggests a trend in which some relatively basal lineages exhibit broader host use, whereas most derived lineages show higher host specificity (Figs. 2 and 3). Early nodes were generally reconstructed as ancestors that utilized multiple sponge taxa, while more recent nodes and most extant species have high specificity, utilizing only one to three sponge orders (Figs. 2 and 3a). However, only two extant species, Acasta cyathus and Euacasta dofleini, exhibit a generalist host-use pattern, which limits broader inference across basal lineages. Overall, the ASR suggests that specialist species evolved from ancestrally broader host-use states, and that different degrees of host specificity (STD) have evolved multiple times independently. Species with different degrees of STD can co-exist within the same clade, such as E. dofleini and E. porata (Fig. 3a). In Fig. 3b, we illustrate barnacle species number and prevalence in six host sponge species, from which at least 40 specimens were sampled. In these intensively sampled hosts, multiple barnacle species often occurred in the same sponge individual. Such co-occurrence could include both generalists and specialists, but specialist barnacles were dominant in all six sponge hosts, i.e., Neoacasta laevigata, E. porata, Pectinoacasta sculpturata, A. sandwichi, A. crucibasis, and Membranobalanus orcutti (Fig. 3b; Additional file: Table S1–S2). The above-mentioned generalist barnacles, A. cyathus and E. dofleini, can live in diverse sponges, but even so they always have a low prevalence when co-occurring with other barnacle species. Correspondence between sponge and barnacle (host specificity/host usage) Our study examined 34 barnacle species (8 genera; 4 subfamilies, Acastinae, Bryozobiinae, Archaeobalaninae, and Pyrgomatinae) found in 10 different orders of sponges (Fig. 4a; Additional file 1: Fig. S9, Tables S1–S2). All sponge hosts belonged to the class Demospongiae, except for the order Clathrinida in the class Calcarea. The number of sponge species sampled within these 10 orders ranged from one to 24. Two sponge orders hosted only a single barnacle species each: Clathrinida with E. dofleini and Tethyida with A. cyathus. The Clionaida hosted six barnacle species, all with a membranous basis (Archiacasta pustulata, four species of Membranobalanus, and Pyrgospongia stellula). The highest diversity was found in the Haplosclerida, which hosted six barnacle genera from three subfamilies (Acastinae, Bryozobiinae, Archaeobalaninae). a Number of sponge-associated barnacle species in the present study hosted by sponge orders. b 3D plot derived from correspondence analysis, showing the relationship between sponge orders and barnacle species across dimensions 2–4. Percentages of variance explained by each dimension are indicated on the axes, and the full variance distribution across dimensions 1–9 is provided in Additional file 1: Fig. S20 A chi-square test demonstrated a significant association between barnacle species and host sponge orders (chi-square: 40,676.02, p-value < 2.2e − 16). To further investigate the relationship between sponge barnacles and their corresponding host sponges, we utilized correspondence analysis (CA) to compute the frequencies of barnacle species presence or absence across host sponge orders. The Haplosclerida sponges, which hosted a large number of barnacle species, may influence the resolution, with the first three dimensions of the CA explaining 57.2% of the total variation. The cumulative percentage variance explained by the first five ordination axes in the CA analysis exceeded 88.5%, indicating a high correlation between host orders and barnacle species. The ordination results revealed that most barnacle species are specific to their host orders (Fig. 4b). For example, the order Clionaida is predominantly associated with membranous base barnacles. In contrast, generalists like A. cyathus and E. dofleini are hosted by multiple sponge orders. This may lead to the grouping of Clathrinida and Tethyida with Haplosclerida, as Clathrinida and Tethyida exclusively host E. dofleini and A. cyathus, respectively. Sponge orders that host a single barnacle species are tightly clustered around that species in the CA axes (Fig. 4b; Additional file 1: Figs. S19–S22). Host usage of sponge barnacles in the Indo-West Pacific and adjacent regions Based on our samples and a review of the literature [51, 68–74], we studied the host usage of sponge barnacles across different geographical locations. It is important to note that Taiwan, our local region, was the most intensively sampled, which may have influenced our analysis. Sponge barnacles have been widely recorded in the Indo-West Pacific. The number of barnacle species collected in each region ranged from three to 31, predictably highest in Taiwan (Fig. 5a). Distribution and diversity patterns of barnacle-hosting sponges. a (Left) Encounter points of barnacle-hosting sponges across a latitudinal gradient in the Indo-West Pacific (present data plus literature review) Base sea surface temperature map was generated from the database in The National Oceanic and Atmospheric Administration (NOAA), USA. (Right) Sponge diversity color-coded by taxonomic order and arranged by locality. b β-diversity or Sørensen dissimilarity of sponge species (bSOR) computed for all 9 sites in the present study, measuring differences in species composition and partitioned into its components of species replacement or turnover (bSIM) and nestedness or subset relatedness (bSNE) Most of the reported barnacles are specific to a single species of host sponge. Among these, species such as Acasta lappa, Euacasta porata, Pyrgospongia stellula, consistently show high host specificity overall in tropical and subtropical waters. Some barnacle species appear to have shifted host taxon along latitudinal gradients, but they mostly remain within the same sponge orders (Fig. 5a; Additional file 1: Table S1). Beta-diversity of communities illustrates the relative differences in species composition between habitats [75]. It connects biodiversity at local and regional scale [76, 77]. The diversity index is calculated for each pair of samples, representing similarity or dissimilarity between the two. The β-diversity (βsor) of sponge barnacles along the latitudinal gradients was 0.76, and the turnover component (βsim) was 0.52, suggesting that there was a 50% turnover of barnacle species among sponge samples from the Indo-West Pacific. The nestedness-resultant component (βsne) was 0.25, indicating that species richness has an influence on the interpretation of dissimilarity indices (Additional file 1: Tables S3–S5). All components of β-diversity exhibited a nonlinear relationship with latitude. The total dissimilarity of barnacle species (βsor) ranges from 0.76 (South Korea) to 1.00 (Taiwan), indicating high barnacle species richness among sponge samples. The species turnover (βsim) ranged from 0.68 (South Korea) to 1.00 (Taiwan), indicating a high species replacement among sponge samples, which share few species in common. Species nestedness (βsne) also exhibited no variation with latitude and its low value of 0–0.08 implies there is not much effect of differences in species richness in the interpretation of dissimilarity indices. Thus, the composition of barnacle communities in sponge samples was mostly determined by turnover and less by nestedness (Fig. 5b; Additional file 1: Tables S3–S5). Percentages of generalists and specialists in relation to community diversity In general, host sponges monopolized by single barnacle species are common, while sponges hosting multiple barnacle species are rare. Most of the monopolizing barnacles are specialists with high host specificity to their host. In agreement with this, each sponge individual in our material usually hosted only a single barnacle species. Figure 6a shows the accumulation of barnacle species in sponge sample collections from eight locations. Seven sponges were collected from the Philippines, but all hosted only a single barnacle species, and thus this dataset was excluded. Occupancy patterns of barnacles. a Counts of sponge-associated barnacle species per host sponge across eight examined regions, illustrating that most sponges host only a single barnacle species. Bars represent the observed number of individual sponges inhabited by different numbers of barnacle species. b β-diversity between single and multiple barnacle species in sponges across eight sites, showing that single barnacle species contribute more to β-diversity than multiple barnacle species The β-diversity of single barnacle species in sponges (SS) and multiple barnacle species in sponges (MS) was estimated for the same eight regions. The β-diversity components βsor and βsim present similar values in both SS and MS across the regions. Furthermore, the βsne values were low, ranging from 0 to 0.05, suggesting that high diversity originated from differences in species composition (βsim) rather than being related to nestedness (βsne). Generally, β-diversity is higher for SS than for MS, except in Hong Kong, where the number of specialist barnacles is lower than that of generalist barnacles. This result indicates that SS specialist barnacles likely contribute more to the diversity of sponge barnacles (Fig. 6b; Additional file 1: Table S5). Discussion Our study provides new insights into the evolutionary history and host specialization of sponge-associated barnacles. By integrating molecular phylogenetics with detailed morphological analyses, we investigate their diversification in relation to sponge hosts. However, our findings do not support the hypothesis that the morphological specialization of sponge barnacles is closely linked to their host sponges or that co-evolutionary processes drive their diversification. Molecular phylogenetic analyses reveal that sponge barnacles are divided into two major clades: the M-Base clade and the C-Base clade. Furthermore, our results suggest that the subfamily Acastinae is polyphyletic, its species sharing closer relationships with Pyrgomatidae, Archaeobalaninae, and Bryozobiinae. These molecular findings challenge existing taxonomic classifications and do not entirely align with morphological evidence. Ancestral state reconstructions of ten morphological traits indicate that probably only the basis structure and shell plate connection are consistently mapped across sponge barnacle lineages, highlighting its evolutionary significance. Host-driven divergence and adaptive evolution Our phylogenetic analyses suggest that sponge-associated barnacles show a pattern of host-associated divergence, with generalist species occupying basal positions and specialization evolving from ancestrally broader host-use states. This pattern is consistent with broader evolutionary trends in symbiotic systems, where ecological opportunities and phenotypic plasticity drive adaptive radiation [5, 6]. The availability of diverse sponge hosts likely created ecological niches that facilitated barnacle speciation, akin to classic cases of adaptive radiation such as Darwin’s finches [7, 78] and African cichlid fishes [10]. Our molecular phylogenetic analysis of sponge barnacles reveals significant incongruence with the phylogeny of their sponge hosts, and divergence time estimates suggest that sponges evolved much earlier than barnacles. Sponge-associated barnacles split from other balanoid barnacles around 38 Mya in the late Eocene, a period of increasing ecological complexity in reef systems. These findings indicate that barnacle diversification has not followed a strict co-evolutionary pattern with their hosts. Instead, the symbiotic relationship likely originated when ancestral barnacles first settled on Clionaida sponges, which possess a remarkable ability to incorporate foreign materials. While this strategy is generally assumed to strengthen the skeleton, its exact function and underlying mechanisms remain poorly understood [79]. By integrating foreign materials, Clionaida sponges increase their weight, enhance structural stability, and anchor more effectively to substrates [47, 48], traits that may have inadvertently facilitated barnacle settlement and persistence. Furthermore, the literature review of this study suggests that Clionaida sponges possess fewer biochemical types than other sponge orders, potentially exerting a minimal biochemical influence on barnacle settlement, development, and long-term symbiosis (Additional file 1: Fig. S23, Tables S2, S6–S7). This could explain why the basal M-Base clade of sponge barnacles remains closely associated with Clionaida sponges, whereas the C-Base clade exhibits more diverse host associations, possibly reflecting different ecological or physiological constraints in host selection. Sponge-associated barnacles display a spectrum of morphological specializations influenced by host interactions. Generalist species retain ancestral traits, while specialists develop distinct structural adaptations suited to their particular hosts. These findings align with previous morphological studies [51, 52, 55], reinforcing the idea that host-imposed selection pressures shape barnacle morphology. Features such as reduced shell plate armature and ‘window’ formations between barnacle plates and sponge tissues suggest functional integration between host and symbiont. This evolutionary trajectory underscores how symbiotic relationships can drive niche specialization and morphological adaptation [1, 15], ultimately fostering diversification and functional integration between interacting species. Host specificity and ecological impact Sponge-associated barnacles exhibit a remarkable degree of host specificity, with most species forming tight associations with particular sponge hosts, often at the species level. Only a few barnacle species are generalists, capable of inhabiting a broader range of sponge hosts. However, despite their broader host range, generalist barnacles tend to have lower prevalence within their host sponges, whereas specialist species, having successfully adapted to specific hosts, dominate in sponge communities. This pattern remains consistent across various geographical locations, from tropical to subtropical waters, where specialists maintain high host specificity and prevalence, while generalists remain relatively rare. The pronounced host-driven divergence and adaptive evolution observed in sponge barnacles reinforce their competitive advantage within specific host sponges and contribute to the broader diversity and complexity of marine ecosystems. By integrating into sponge communities with high fidelity, these barnacles enhance species interactions and influence community structure, ultimately playing a critical role in shaping the ecological dynamics of sponge-associated marine environments. Conclusion This study investigates how host associations influence the diversification and morphology of sponge-associated barnacles. Diversification of sponge-associated barnacles is primarily shaped by host-driven adaptations rather than strict co-evolution with their sponge hosts. Molecular data reveal that sponge barnacles cluster into the M-Base clade and the C-Base clade, challenging existing taxonomic classifications and highlighting the polyphyletic nature of Acastinae (at least its generic assignments). Furthermore, our ancestral state reconstruction suggests that probably only the shell plate connection is consistently mapped across sponge barnacle lineages, underscoring its role in host specialization. Despite the absence of strict co-evolution, our findings support the hypothesis that generalist species represent the ancestral state from which more specialized species evolved. The availability of diverse sponge hosts likely provided ecological opportunities that facilitated barnacle speciation. The high degree of host specificity of sponge-associated barnacles highlights the ecological significance of sponge barnacles in shaping the complex marine community structures. While our analysis could have benefited from an even broader taxon sampling, it includes as many species as was possible at this time. Obviously, it is up to future more penetrating studies to test whether the conclusions presented here can be supported. Methods Field collection of sponge-associated barnacles and sponge hosts We collected sponges and their associated barnacle specimens from a total of 162 diving sites distributed across tropical and subtropical waters including Korea (Jeju Island, 8 sites), Japan (Kochi, 7 sites; Okinawa, 7 sites; Fukuoka, 2 sites), Taiwan (North Coast, 2 sites; Northeast Coast, 3 sites; Kueishan Island, 2 sites; Penghu Archipelago, 6 sites; Liuqiu Island, 9 sites; Green Island (Ludao), 28 sites; Orchid Island, 9 sites; Kenting, 1 site; Dongsha Atoll, 3 sites), Hong Kong (17 sites), Papua New Guinea (7 sites), Malaysia (18 sites), Thailand (12 sites), and Vietnam (21 sites; see sites details in Additional file 1: Fig. S24, Table S1). Collections were made by SCUBA diving at depths of 2 to 40 m. During each dive, whole sponges with barnacles were collected in situ. Morphological features of the sponges were documented and photographed according to the methods described in Yu et al. [80]. In the laboratory, sponges were dissected, and all barnacle individuals were extracted, identified, and enumerated. Our dataset includes four families of sponge-associated barnacles, 8 of 12 known genera, and 34 of 107 described species. For molecular comparison, 38 species are available in NCBI GenBank but not all of GenBank samples have sequences for multiple genes (see Table S1). Morphological and molecular data We used DNA barcoding techniques and diagnostic morphological characters to identify sponge and barnacle species (Additional file 1: Fig. S10, Tables S8–S9). Both sponges (skeletal structures and spicule types) and barnacles (shell plates, opercular plates, cirri, penis, and mouthparts) were dissected under stereomicroscope for morphological identification and subsequent molecular analysis. To avoid contamination from symbionts, we extracted DNA from the central part of the sponge. The remaining sponge material was cut into thin sections (~ 0.1 mm thickness) and subjected to acid digestion (10% HCl) for skeleton and spicules observations, respectively. For barnacles, the adductor muscles were dissected for DNA extraction, while the shell, cirri, penis, and mouthparts were used for morphological character observations (Additional file 1: Fig. S10). Sample collection, DNA extraction, and species identification for sponge barnacles followed the protocols outlined in Yu et al. [31, 50, 80]. DNA sequences for three mitochondrial DNA markers (12S rRNA, 16S rRNA, and cytochrome c oxidase subunit I [COI]) and two nuclear markers (18S rRNA and histone 3 [H3]) were obtained to reconstruct the phylogenetic relationships of sponge-associated barnacles (Additional file 1: Table S8). These markers are widely used in barnacle phylogenetic analyses [50, 81, 82]. Polymerase chain reaction (PCR) was performed using primers and annealing temperatures as described in Yu et al. [80]. Direct sequencing of the purified PCR products was conducted using the ABI 3730XL Genetic Analyzer with BigDye Terminator cycle sequencing reagents (Applied Biosystems, Foster City, CA, USA). Data analyses Barnacle molecular phylogenetic reconstruction and co-phylogeny analysis Barnacle DNA sequences were assembled, edited, concatenated and aligned with MUSCLE using the default parameter settings and subsequently verified visually in Geneious [83]. Phylogenetic relationships of sponge-associated barnacles were reconstructed using Bayesian inference (BI) and maximum likelihood (ML) methods. The best-fit evolutionary model for each partition was determined using jModeltest 2.1.7 [84] and ModelFinder [85]. BI and ML analyses were conducted with MrBayes v.3.2.1 [86] and IQ-TREE v.1.6.8 [87], respectively. Ancestral state reconstruction (ASR) and character mapping were performed in RASP v.4.2 [88] using the Bayesian phylogenetic tree of sponge barnacles inferred from five markers, which was used as the basis for all ASR analyses in the study (Additional file 1: Fig. S9). We included host usages and ten representative morphological traits suggested to have phylogenetic importance from previous traditional morphological-based phylogenetic studies (Additional file 1: Fig. S25) [51, 89–91]. These traits include four related to shell structures, three to opercular valves, and three to the somatic body (Additional file 1: Fig. S10). Morphological traits and molecular data were collected from the same specimens to create a comprehensive dataset (Additional file 1: Tables S1, S9). The divergence times of sponge-associated barnacles were estimated using MEGA X 10.2.4 [92] under the RelTime method [93]. Calibration points from Pérez-Losada et al. [94] were used for this analysis, including Austrobalanus imperator (C14) from the Middle Eocene (38.0–47.8 million years ago (Mya)), Tetraclitella sp. cf. purpurascens (C11) from the Lower Miocene (20.4–23.0 Mya), and Palaeobalanus lindsayi (C12) from the Middle Eocene (38.0–47.8 Mya). Node C12 refers to the positions of Neoacasta laevigata and Pyrgopsella annandalei in Pérez-Losada et al. [94] (Additional file 1: Table S10). Co-phylogenetic analysis was conducted using eMPRess GUI Version 1.0 [95] to reconcile the barnacle phylogeny with host sponge phylogeny under the DTL (duplication-transfer-loss) model. The input tree topologies for sponge barnacles and sponges were based on the molecular phylogenies from this study and Plese et al. [65], respectively. The analysis was performed using eMPRess default parameters (duplication cost = 1, transfer cost = 1, and loss cost = 1). Host specificity, phylogenetic, and ecological relationships We investigated the range of host usage in barnacles and their prevalence in sponges using specimens collected during this study and supplementary data from the literature on sponge barnacles and their hosts (Additional file 1: Table S1). Host usage was assessed using basic and taxonomic specificity indices. The basic specificity index is defined as the total number of host taxa (family and genus levels) used by each barnacle species in the collected samples [96, 97]. Taxonomic specificity (STD) was measured following the method of Poulin and Mouillot [66]. Species with broader host usage will have higher STD scores. Ancestral state estimates of basic and taxonomic specificity indices and STD on the phylogenetic tree were performed using the R packages ggtree, phytools, ape, RColorBrewer, ggimage, ggplot2, and treeio [98–102]. Taxonomic specificity indices for barnacles were analyzed at the levels of order, family, and genus of host sponges, while species-level analysis was avoided to reduce potential biases from misidentification (Additional file 1: Table S1) [96, 97]. To explore the relationship between the diversity of sponge barnacles and their corresponding host sponges, we used correspondence analysis (CA) to compute the frequencies of sponge barnacle species present or absent in host sponge orders, utilizing the R packages FactoMineR and factoextra [103, 104]. The influence of diversity and host specificity (composition of single and multiple species per sponge) on sponge barnacles across latitudinal climate zones (β-diversity) was analyzed using the R package betapart [105]. Since the Australian specimen data are derived from a literature review, the lack of specimen quantity information prevents the calculation of β-diversity. Additionally, our Philippine specimen collection contains only a single barnacle species per sponge, making it impossible to analyze the β-diversity between single and multiple barnacle species in sponges. In betapart, spatial patterns of β-diversity were assessed using multiple-site dissimilarity (Sørensen and Simpson indices) to evaluate heterogeneity among sites across the latitudinal gradient. Species replacement (turnover) and species loss or gain (nestedness) in each latitudinal region were also assessed. The relevance of turnover (βsim) and nestedness (βsne) components of total dissimilarity (βsor) can provide insights into the adaptive and survival strategies of symbiotic life. Graphical representations of contingency tables, chi-square test results, and CA three-dimensional plots were generated using the R packages gplots, corrplot, and plotly [106–108]. Supplementary Information Additional file 1. Figures S1–S25 and Tables S1–S10. FigS1– [Maximum likelihood topology inferred from 12S]. FigS2– [Maximum likelihood topology inferred from 16S]. FigS3– [Maximum likelihood topology inferred from 18S]. FigS4– [Maximum likelihood topology inferred from COI]. FigS5– [Maximum likelihood topology inferred from H3]. FigS6– [Maximum likelihood phylogeny of sponge barnacles based on five markers 12S, 16S, 18S, COI and H3]. FigS7– [Bayesian phylogeny of sponge barnacles based on five markers 12S, 16S, 18S, COI and H3]. FigS8– [Maximum likelihood phylogeny of sponge barnacles based on six markers including 28S from this study and Hosie et al., 2021]. FigS9– [Bayesian phylogeny inferred from concatenated dataset]. FigS10– [Ten morphological traits of sponge barnacles]. FigS11– [Ancestral states of shell plate external structure inferred on Bayesian tree].. FigS12– [Ancestral states of basis form inferred on Bayesian tree]. FigS13– [Ancestral states of scutum aspect ratio inferred on Bayesian tree]. FigS14– [Ancestral states of scutum external structure inferred on Bayesian tree]. FigS15– [Ancestral states of tergum spur furrow inferred on Bayesian tree]. FigS16– [Ancestral states of labrum teeth inferred on Bayesian tree]. FigS17– [Ancestral states of penis basidorsal point inferred on Bayesian tree]. FigS18– [Divergence time of sponge barnacles estimated using RelTime]. FigS19– [Contingency between sponges and barnacles shown as balloon plot]. FigS20– [Percentage of variance explained by each dimension]. FigS21– [Quality of representation and contributions of barnacles and sponges]. FigS22– [Quality of representation and contributions of barnacles and sponges by dimension]. FigS23– [Literature review of natural products from marine sponges]. FigS24– [Map of sampling sites]. FigS25– [Main evolutionary hypotheses on morphology and phylogeny of sponge barnacles]. Table S1– [Host affiliation, prevalence, distribution and specificity indices of sponge barnacles analysed in this study]. Table S2– [Use of demosponges by sponge-associated barnacles]. Table S3– [Beta diversity of sponge barnacles along latitudinal gradients]. Table S4– [Beta diversity of sponge barnacles by country]. Table S5– [Beta diversity of single and multiple sponge barnacle species by country]. Table S6– [Literature review of biochemical compounds identified in marine sponges]. Table S7– [Reference list for literature review of biochemical compounds in marine sponges]. Table S8– [Literature review of primers used in barnacle phylogenetic studies]. Table S9– [Specimen information and sequences used for phylogenetic reconstruction]. Table S10– [Fossil species and ages used for divergence time calibration]. Acknowledgements We thank P. C. Tsai, Y. F. Tsao, M. H. Ho and W. P. Hsieh (Academia Sinica) for assistance. We thank W. Kim (Seoul National University, South Korea), T. Metzaki (Kuroshio Institute, Japan), M. Ganmanee (King Mongkut’s Institute of Technology Ladkrabang, Thailand), J. C. H. Tan (Universiti Malaysia Terengganu, Malaysia), and Y. Nozawa (University of the Ryukyus, Japan) for assisting with and arranging for field work in South Korea, Japan, Thailand and Malaysia. We thank C. Schönberg (National Sun Yat-sen University, Taiwan) and S. C. Lim (Lee Kong Chian Natural History Museum, National University of Singapore, Singapore) for their support in sponge species identification and discussion on sponge ecology. Jens T. Hoeg thanks Prof. BKK Chan and Academia Sinica for funding a visit to Taipei. Abbreviations Bayesian inference Maximum likelihood Membranous basis Calcareous base Ancestral state reconstructions The most recent common ancestor Oosterior probability Million years ago Taxonomic specificity Correspondence analysis Sørensen dissimilarity in β-diversity Simpson dissimilarity (= turnover component of Sørensen dissimilarity) in β-diversity Nestedness component of Sørensen dissimilarity in β-diversity Single barnacle species in sponges Multiple barnacle species in sponges Duplication-transfer-loss Authors’ contributions All authors conceived the ideas and designed the methodology. MCY, GAK, BKKC conducted the sampling. MCY, HCL, BKKC analyzed the data. MCY prepared the photo plates. MCY, BKKC wrote the first draft. JHT, GAK reviewed and edited the manuscript. All authors read and approved the final manuscript. Funding GAK was financially supported by the Russian Foundation for Basic Research (grant no. 21–54–52003 MNT_a). The study was also conducted under the state assignment of Lomonosov Moscow State University. BKKC was supported by the Senior Investigator Award of Academia Sinica, Taiwan (AS-IA-105-L03). MCY was supported by the Academia Sinica Postdoctoral Scholar Program and the Doctoral Fellowship of the Doctoral Degree Program in Marine Biotechnology, National Sun Yat-sen University and Academia Sinica, Taiwan. Data availability The DNA sequences generated in this study are available in GenBank at National Center for Biotechnology Information with accession numbers listed in Additional file 1: Table S9. Voucher specimens and associated information are provided in Additional file 1: Table S9 and deposited at Academia Sinica, Taipei, Taiwan. Please also see the file "Data Accessibility" — Supplementary Material for Peer Review. This supplementary file has been prepared to support the peer review process for our manuscript. It includes the following materials: Four R scripts used to generate specific figures presented in the manuscript, as indicated by the corresponding annotations within each script. Detailed information on the nucleotide sequence data newly reported in this study, which were submitted to the NCBI database in 2022. Although accession numbers have been assigned, public access links have not yet been provided by NCBI. All relevant accession numbers and associated data are included in this document for the editors’ and reviewers’ reference. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Additional file 1. Figures S1–S25 and Tables S1–S10. FigS1– [Maximum likelihood topology inferred from 12S]. FigS2– [Maximum likelihood topology inferred from 16S]. FigS3– [Maximum likelihood topology inferred from 18S]. FigS4– [Maximum likelihood topology inferred from COI]. FigS5– [Maximum likelihood topology inferred from H3]. FigS6– [Maximum likelihood phylogeny of sponge barnacles based on five markers 12S, 16S, 18S, COI and H3]. FigS7– [Bayesian phylogeny of sponge barnacles based on five markers 12S, 16S, 18S, COI and H3]. FigS8– [Maximum likelihood phylogeny of sponge barnacles based on six markers including 28S from this study and Hosie et al., 2021]. FigS9– [Bayesian phylogeny inferred from concatenated dataset]. FigS10– [Ten morphological traits of sponge barnacles]. FigS11– [Ancestral states of shell plate external structure inferred on Bayesian tree].. FigS12– [Ancestral states of basis form inferred on Bayesian tree]. FigS13– [Ancestral states of scutum aspect ratio inferred on Bayesian tree]. FigS14– [Ancestral states of scutum external structure inferred on Bayesian tree]. FigS15– [Ancestral states of tergum spur furrow inferred on Bayesian tree]. FigS16– [Ancestral states of labrum teeth inferred on Bayesian tree]. FigS17– [Ancestral states of penis basidorsal point inferred on Bayesian tree]. FigS18– [Divergence time of sponge barnacles estimated using RelTime]. FigS19– [Contingency between sponges and barnacles shown as balloon plot]. FigS20– [Percentage of variance explained by each dimension]. FigS21– [Quality of representation and contributions of barnacles and sponges]. FigS22– [Quality of representation and contributions of barnacles and sponges by dimension]. FigS23– [Literature review of natural products from marine sponges]. FigS24– [Map of sampling sites]. FigS25– [Main evolutionary hypotheses on morphology and phylogeny of sponge barnacles]. Table S1– [Host affiliation, prevalence, distribution and specificity indices of sponge barnacles analysed in this study]. Table S2– [Use of demosponges by sponge-associated barnacles]. Table S3– [Beta diversity of sponge barnacles along latitudinal gradients]. Table S4– [Beta diversity of sponge barnacles by country]. Table S5– [Beta diversity of single and multiple sponge barnacle species by country]. Table S6– [Literature review of biochemical compounds identified in marine sponges]. Table S7– [Reference list for literature review of biochemical compounds in marine sponges]. Table S8– [Literature review of primers used in barnacle phylogenetic studies]. Table S9– [Specimen information and sequences used for phylogenetic reconstruction]. Table S10– [Fossil species and ages used for divergence time calibration]. Data Availability Statement The DNA sequences generated in this study are available in GenBank at National Center for Biotechnology Information with accession numbers listed in Additional file 1: Table S9. Voucher specimens and associated information are provided in Additional file 1: Table S9 and deposited at Academia Sinica, Taipei, Taiwan. Please also see the file "Data Accessibility" — Supplementary Material for Peer Review. This supplementary file has been prepared to support the peer review process for our manuscript. It includes the following materials: Four R scripts used to generate specific figures presented in the manuscript, as indicated by the corresponding annotations within each script. Detailed information on the nucleotide sequence data newly reported in this study, which were submitted to the NCBI database in 2022. Although accession numbers have been assigned, public access links have not yet been provided by NCBI. All relevant accession numbers and associated data are included in this document for the editors’ and reviewers’ reference.