Biological Journal of the Linnean Society

Understanding the evolutionary history of species within a geographic context is key to historical biogeography, as it reveals how geological and climatic changes shaped biodiversity. This is especially important in ecologically significant regions like the Amazon and Orinoco basins. Together, they host the worlds greatest freshwater fish diversity ([~]3,500 species), sharing a common but not yet fully understood evolutionary history. The gilded catfish (Brachyplatystoma rousseauxii), an ancient species widely distributed as a metapopulation in Neotropics, is an important model for studying past connectivity, divergence, and historical processes shaping fish diversity between these basins. This study analyzed the genetic structure, connectivity routes, and demographic history of B. rousseauxii using nuclear (microsatellite and ddRADseq) and mitochondrial DNA. Population structure analyses and coalescent models indicate that B. rousseauxii populations from the Orinoco and Amazon basins are genetically distinct, with no evidence of current gene flow. However, our results support the occurrence of a possible secondary contact event after the divergence, with the Boa Vista population retaining the genetic signal of this process. The ancestral population split occurred at the Rupununi Portal around 2.54 Ma (ddRAD) or 1.31 Ma (mtDNA). Then, the species colonized the Branco and Orinoco Rivers [~]1.90 Ma (ddRAD) or 0.6 Ma (mtDNA), rapidly expanding in the Orinoco (>1.3 or >0.29 Ma), while colonization of the Amazon from the Branco River was more recent ([≤]1.0 or [≤]0.15 Ma). Population expansion signal was detected in the Orinoco ([~]0.20 Ma), whereas the Amazon remained stable. Our findings suggest that the rise of the Vaupes Arch in the Late Miocene does not explain the observed genetic divergence. Likewise, the Casiquiare Canal and Japura-Guaviare headwaters are not connectivity routes between basins. Instead, the Rupununi Portal, including the recent capture of the Branco River by the Negro River, was the last point of connection and played a key role in shaping B. rousseauxiis distribution. These findings provide insights into Neotropical fish biogeography and the historical configuration of the Orinoco and Amazon basins.

Machine learning models can be used to analyze large bioacoustics datasets and explore variation due to geography or habitat. We find that in the monotypic Lark Sparrow (Chondestes grammacus), both environmental variables and geographic distance influence song variation in this species. Bird song is an important method of communication within avian species. The variation in bird song within a species can be due to a variety of factors, including genetic drift and isolation by distance. However, it remains unclear in species with wide ranges how environmental factors in particular can cause changes to the song. In this study, the song C. grammacus was analyzed via machine learning to determine if it had significant variation based on multiple geographical metrics. We trained a convolutional neural network to segment individual syllables of 91 C. grammacus recordings, then extracted song characteristics. We found that ecoregion and state explain variation in C. grammacus songs. Our results demonstrate the efficacy of using machine learning models to analyze large datasets, as well as the impact that ecogeographic variation has on song variance.

Animal personality is the expression of consistent among-individual variation in a behavioural trait across time and context. The application of this theory to behavioural research provides a valuable framework to investigate the among- and within-individual variation in behaviour. The animal personality theory is particularly relevant to understand and diagnostic fish welfare. It can be integrated within the nature-based welfare framework, which emphasizes the expression of species-specific behaviours and the maintenance of consistent behavioural patterns over time. Because behavioural trait such as boldness may be closely linked to other functional phenotypes, such associations reflect the broader concept of animal personality, whereby a behavioural trait can covary with another phenotypic trait to support an adaptive responses to environmental conditions. Although these relationships are both species and context dependent, they are consistently shaped by environmental conditions and environmental complexity tend to promote species specific behaviours and reduce maladaptive traits. Here we examined how structural environmental complexity shapes personality in Arctic charr and their covarying functional phenotypes, specifically growth rate and brain size. We propose that environmental complexity promotes alternative behavioural and functional phenotypes through multivariate phenotypic plasticity. We found that environmental complexity did not influence mean boldness between both treatments, but repeatability of boldness in the complex environment was remarkably consistent over a longer-term period, while estimates collapse after seven days in the plain treatment. Our findings are a clear evidence that environmental complexity foster stable behavioural trait expression and that a plain environment may suppress personality. Our results provide compelling evidence that behavioural structure and dynamics are embedded within patterns of behavioural variance. Although we found no support for behavioural covariation or associations with growth rate and brain size, we suggest that the animal personality framework may offer a valuable approach for diagnosing fish welfare issues through the partitioning of behavioural variance.

The evolutionary patterns of trait diversification provide insights into the function of the trait. Early burst of trait evolution is often associated with adaptive radiation, rapidly diversifying the trait in response to vacant niches followed by the slowdown of the diversification with niche filling, whereas late burst is more likely to be associated with sexual selection, possibly contributing to reproductive barriers between closely related species. Here, we studied the diversification of tail fork depth through time in hirundines to infer its function, which remains unclear due to the competing two alternative hypotheses: the sexual selection hypothesis, which is a classic explanation of deeply forked tails, proposed that this trait has evolved via sexual selection, which was then challenged by the viability selection hypothesis, which proposed that deeply forked tails have mainly evolved via viability selection for enhancing aerodynamic performance during aerial foraging on large prey. We found a late burst of tail fork depth, but not of bill length, i.e., an index of prey size. The observed pattern is consistent with the sexual selection hypothesis but not with the viability selection hypothesis.

Honey bees can swim on the water surface toward dark regions, a behavior known as scototaxis that may facilitate escape from water. Although this behavior has been reported in both honey bees and solitary bees, variation among honey bee species remains poorly understood. We compared scototaxis during swimming in four honey bee species representing two nesting types: open-nesting (Apis florea and A. dorsata) and cavity-nesting (A. cerana and A. mellifera). Individual bees were released into a water-filled arena containing a dark sector, and their landing angles were recorded. All species exhibited significant orientation toward the dark sector. However, open-nesting species showed significantly stronger orientation than cavity-nesting species. No significant differences were detected between replicate colonies within species or between species within the same nesting type, whereas differences between nesting types were highly significant. Hierarchical clustering based on orientation strength placed Osmia, a solitary cavity-nesting bee from our previous study, in the same behavioral cluster as the two open-nesting Apis species rather than the cavity-nesting honey bees. We also measured swimming duration, distance, and velocity, but found no consistent differences between nesting types. These results demonstrate substantial interspecific variation in swimming scototaxis. The behavioral clustering is consistent with the hypothesis that strong scototaxis represents an ancestral trait that has been reduced in the derived A. cerana/A. mellifera lineage.

EU has reclassified the sika deer (Cervus nippon) as an undesirable invasive species based on reports that hybridization with the indigenous red deer (C. elaphus) may produce fertile offspring. Since sika-derived DNA previosuly introduced into the red deer population (introgression) cannot be removed, the crucial question is whether new (F1) hybridisation occur. To address this, we analysed the chromosomes in 56 sika and 22 red deer. All red deer had a chromosome number 2n=68. In contrast, the chromosome number in sika ranged from 64 to 67, due to the variable presence of two sika-specific Robertsonian translocations (ROB1,ROB2). In the free-ranging sika population in Jutland, >90% of the sika deer were homozygote for at least one of these ROBs, excluding that they could be F1-hybrids. Moreover, ROB2 was in Hardy-Weinberg equilibrium, further supporting the absence of gene flow between the two species. In contrast, ROB1 was in Hardy-Weinberg disequilibrium, suggesting negative fitness of heterozygotes, including potential F1-hybrids. In Jaegersborg Deer Park, the eight examined sika deer had the same genotype (absence of ROB1, homozygosity of ROB2), supporting that it is a founder population which may have been isolated for [~]100 years. Again, none of these can be F1-hybrids due to the homozygosity of ROB2. We conclude that F1-hybridisation between sika and red deer either does not occur or occur very rarely in Denmark. The study establish the Danish sika-populations as unique models for adressing important biological questions: What underlies the absence of hybridisation? Why are ROBs frequent in sika deer but not in the closely related red deer? How fast do new species/subspecies develop in isolated founder populations? Which factors determine, that some ROBs have little heterozygous effects, whereas others are selected against, with implications for the role of ROBs as genetic barriers promoting speciation, and for fertility problems in some human ROB carriers.

Phenotypic differences among populations can arise through heritable genetic divergence, phenotypic plasticity, or both, making it difficult to determine whether trait-environment correlations observed in nature reflect adaptive evolution. Within threespine stickleback (Gasterosteus aculeatus) studies, numerous document morphological differences among allopatric-, parapatric-, and even sympatric populations. These phenotypic differences among populations are often correlated with diet and lake habitat (e.g., lake size), suggesting an adaptive value to the population differences. However, many studies of ecomorphological divergence in stickleback use wild-caught stickleback, which may differ due to evolution or plasticity. Although common garden experiments have confirmed that population differences can be heritable, such experiments typically entail small numbers of populations. Consequently, we still do not know to what extent well-known trait-environment correlations in stickleback are a result of evolution. To address this gap, we reared stickleback embryos from 27 lake populations on Vancouver Island, in a laboratory environment. Morphological differences among populations persist in common-garden fish, confirming a large role for divergent evolution. These heritable differences were associated with environmental variation among lakes, implying an adaptive value. However, some well-known trait-environment relationships in stickleback did not persist in common-garden fish and may be primarily plastic.

Polytypic species with large ranges may harbor unrecognized diversity because taxonomy ranks populations differing subtly in plumage as subspecies. The Ruddy Foliage-gleaner (Clibanornis rubiginosus) exemplifies this problem. It ranges from Mexico to Brazil, with 15 subspecies, and forms a non-monophyletic complex with two congeners, yet its songs had not been compared. We measured ten spectral and temporal variables on 104 recordings covering 14 of 15 subspecies. Bayesian linear mixed models showed three song groups: eight subspecies west of the Andes share a single-note song, whereas Amazonian and Guianan populations add a short introductory note and sing longer, lower-pitched songs. Within this group, watkinsorum sings the lowest-pitched and longest song and is phylogenetically closer to C. cinnamomeigula than to its Amazonian neighbors. The third group is C. cinnamomeigula alone, a white-eyed taxon in an otherwise dark-eyed group. Its high-pitched, vibrato song resembles none other in the genus. One-note and two-note songs differ in kind without intermediates, and every two-note taxon sequenced to date falls in one clade, so C. rubiginosus is paraphyletic. We recognize four species, C. rubiginosus sensu stricto, C. cinnamomeigula, C. watkinsorum, and C. obscurus. This raises Clibanornis from five species to eight and divides its only polytypic species.

Developing robust methods to quantify how animals allocate time across behaviours is essential for understanding energy use, habitat requirements, and responses to environmental change. For cryptic, semi-aquatic mammals such as the platypus, direct observation is difficult, creating a reliance on remote biologging approaches that can reliably infer behaviour in the wild. However, aquatic environments can both smooth acceleration signals through hydrodynamic damping and introduce noise from water movement, turbulence, and drag, potentially obscuring behavioural differences of similar magnitudes. We tested whether progressively incorporating biomechanical and frequency-domain (FFT-derived) predictors improved behavioural classification in hydrodynamically challenging aquatic environments. Tri-axial accelerometers were deployed on four ex situ platypuses, with synchronised video observations used to validate behaviour. From the acceleration data, we derived three predictor classes of increasing complexity: summary statistics describing activity level, engineered biomechanical variables capturing posture and body orientation, and FFT-derived features describing movement rhythm. These predictors were progressively incorporated into Random Forest models to classify five behaviours: burrow resting, surface resting, grooming, travelling/foraging, and diving. Model performance improved with increasing predictor complexity, although gains were behaviour specific. FFT-derived features substantially improved classification of rhythmic behaviours such as diving and foraging, while engineered biomechanical predictors improved grooming detection. In contrast, resting behaviours, particularly surface resting, showed little improvement. Overall accuracy increased from [~]75% to [~]88% when frequency-domain features were included. Misclassification was greatest among behaviours with overlapping or low-amplitude signals, and cross-individual validation revealed reduced model generalisability, indicating that individual variation in movement patterns constrained transferability. Incorporating frequency-domain features substantially improved behavioural classification in platypuses, particularly for rhythmic behaviours such as diving and foraging. This study provides the first validated accelerometry-based behavioural classification framework for the species and highlights the importance of matching predictor selection to behavioural mechanics. More broadly, the approach offers a transferable framework for aquatic and semi-aquatic taxa.

This study investigates the impact temperature-induced vertebral anomalies have on the C-start escape response of Astyanax mexicanus, a model species in evolutionary developmental biology. Employing three temperature treatments to induce varying degrees of skeletal anomalies, we assessed their effects on key swimming performance metrics including, C-start time, curvature coefficient, head displacement distance, and displacement velocity. Through the use of linear mixed models and generalized linear mixed models, our results reveal that specific anomalies such as vertebral fusions and anomalous haemal and neural spines affected the curving ability of C-start escape responses. However, these did not negatively impact other performance parameters, with velocity, distance, and response time showing no significant impacts from any anomaly types, when assessed individually. This suggests a complex interplay between structural deformities and compensatory physiological mechanisms that maintain functional performance. Other variables measured had a stronger and significant impact on swimming performance, including standard length, vertebral number, and temperature treatment, which influenced escape speed, curving ability, and overall locomotor performance. Our findings challenge conventional perceptions about the debilitating impact of vertebral anomalies, indicating that many affected fish can still effectively perform escape maneuvers critical for survival.

Among Madagascar primates, the sportive lemurs (family Lepilemuridae) have seen their species diversity increase from eight in 2005 to 26 in 2009 mostly by applying the phylogenetic species concept to DNA barcode data. Despite the genus being speciose, only one case of sympatry is known from northern Madagascar, where two sportive lemur species described based on low mtDNA divergence, Lepilemur ankaranensis and Lepilemur milanoii, were found to co-occur at the center of their joint distribution range. Here, to clarify the taxonomy of these two species and examine their sympatry, we apply an integrative taxonomic framework to genomic and morphological data from 84 individuals of L. ankaranensis and L. milanoii, encompassing their entire distribution range and the forest of Analafiana, beyond their southernmost limit. Using clustering, multivariate, and isolation by distance analyses, we find no evidence of a sympatric zone and show that despite clear genetic differentiation between regions, the genomic and morphological diversity of the L. ankaranensis, L. milanoii-Analafiana group is clinal and explained by geographic distance. These results clarify that L. milanoii is a junior synonym of L. ankaranensis and that the Analafiana forest population belongs to L. ankaranensis, extending its distribution. It further implies that the 'sympatric' zone, the Andrafiamena forest, hosts conspecific individuals with slightly differentiated mtDNA backgrounds, rather than slightly differentiated sympatric species. Lastly, we re-evaluate the IUCN conservation metrics of L. ankaranensis, which continue to qualify as Endangered (EN) under the B1ab(i-v) criteria.

The ability of prey to eavesdrop on predator vocalizations is expected to increase survival by reducing detection and capture. Unfortunately, most research has been conducted in vertebrates, and little is known about this ability in invertebrates. We measured latency to emerge, overall activity, and shelter visits in wild-caught fall field crickets (Gryllus pennsylvanicus) in response to acoustic playback. Stimuli included multiple predator vocalizations, non-predator vocalizations, white noise, and a control. We predicted that crickets would reduce activity, spend more time in shelter, and freeze in response to stimuli representing greater risk. Contrary to our predictions, crickets traveled greater distances, spent more time moving, and spent less time in shelter in response to predator vocalizations versus controls. We did not, however, find clear differences in responses between predator vocalizations and other treatments. Our results suggest that crickets may not differentiate between the vocalizations of predators, non-predators, and other abrupt sounds. Consequently, eavesdropping may not be a viable method of assessing predation risk for this species and its general use remains unclear.

Parental qualities can influence the development and fitness of their offspring via genetic and non-genetic effects. Although these effects are often linked to parental phenotypes, the effect of parental genetic variation linked with relevant phenotypes is less well understood. We performed full factorial crosses based on parental genotypes for an age-at-maturity-related gene, vgll3, to investigate how the parental genotypes influence Atlantic salmon (Salmo salar) offspring survival, growth, and development in their early life. Beyond the connection with age at maturity, the additional association between vgll3 and body condition in Atlantic salmon offers a potential pathway by which the maternal vgll3 genotype could influence offspring early life fitness. Combined with measurements of maternal phenotype and egg characteristics, the crossing design therefore allowed us to disentangle the maternal and paternal genetic and non-genetic contributions to variation in offspring survival and phenotypic traits. The phenotypic traits measured were hatching length and yolk sac area, growth, and yolk sac consumption and conversion efficiency. Parental vgll3 genotype did not influence the majority of our measured egg traits or alevin traits except for a genetic effect of paternal vgll3 genotype on offspring survival, whereby the paternal late maturation allele was associated with higher survival. Maternal effects were strongest for survival and for traits associated with hatching and weaker for alevin growth and yolk sac usage. Paternal effects on the measured alevin traits were negligible. The results from our study demonstrate that both maternal and paternal effects have the potential to influence offspring early life fitness traits.

Rosa gallica L., the French rose, is a perennial, tetraploid, heterozygous species that naturally propagates by seed and sucker. It occurs in the wild, primarily in Europe, and also exists as cultivated varieties. R. gallica cultivars were extensively bred and cultivated in France at the beginning of the 19th century. Although several hypotheses have been proposed regarding the species expansion based on historical records, none have been assessed using molecular data. Indeed, its genetic diversity has so far been investigated only at local or regional scales, hindering the identification of the evolutionary factors shaping its present-day distribution. Using 29 sequenced microsatellites, we genotyped a comprehensive sample of 1618 individuals, including wild R. gallica from 219 sites across the species range, rose cultivars, and specimens from other Rosa species. We then detected clonal lineages and characterized the range-wide genetic diversity and structure, aiming to disentangle the roles of natural and human factors in shaping the distribution of R. gallica, with particular focus on France. French diversity appears particularly structured compared to the rest of the range, suggesting multiple origins within France. Populations in South Alps, Central Eastern Europe, and Eastern France appear to have recolonized naturally from a single southern glacial refugium. In contrast, populations in the western part of France likely resulted from more recent natural or human-mediated dispersal. Finally, clonal lineages containing both wild and cultivated individuals were predominantly found in France, highlighting the role of human-mediated dispersal in 28 of the 98 French sites studied. These findings show that the present-day natural range of R. gallica was shaped primarily by post-glacial recolonization, but also reveal a contribution of human activities to its recent dispersal, particularly in France, where cultivated varieties were intensively bred and exchanged.

Arcellidae is a family of testate amoebae within Arcellinida (Amoebozoa), comprising three recognized genera: Arcella, Galeripora, and Antarcella. Although species in the family have been studied for nearly two centuries, many historically described taxa and major morphological groups remain unsampled at the molecular level. Here, we provide a comprehensive review of Arcellidae and generate new cytochrome c oxidase subunit I (COI) sequences for arcellid species from Canadian peatlands, focusing on tall-shelled Arcella historically classified in section Altae sensu Deflandre. COI phylogenetic analyses recover a strongly supported monophyletic clade corresponding to North American representatives of Altae, providing the first molecular corroboration of this morphologically defined group. Within this clade, we redescribe Arcella leidyana based on modern material from Eeyou Istchee (Quebec). We further describe Galeripora purdoni sp. nov. from a calcareous fen in eastern Ontario, representing a novel terrestrial lineage within the genus, and redescribe Galeripora artocrea, which we transfer to Arcella based on congruent molecular and morphological evidence. Phylogenomic analyses of Arcellidae isolates from the Protist 10,000 Genomes Project reveal an additional deep lineage basal to Arcella and Galeripora. Together, these results highlight hidden diversity and demonstrate the importance of integrative approaches for resolving arcellid systematics and refining its classification.

Catenulids are free-living flatworms, common in eutrophic freshwaters such as ponds, ditches, or peatbogs, with most of the diversity described from tropical regions to date. Although the majority of the species have been described from warmer climates, most molecular studies have been done on specimens from temperate zones in Europe. We addressed this gap by sampling for exotic species in localities available in a temperate climate. In this study, we investigated catenulid diversity in the greenhouses at the University of Warsaw Botanic Garden and recorded two species known only from tropical areas (Stenostomum paraguayense and Suomina evelinae) and one exotic species recorded previously from a greenhouse in Poland (Stenostomum corderoi). Additionally, in the latter species, we provide evidence for environmentally induced coloration of sensory pits, which has not been reported thus far. We placed the collected species on a phylogeny using barcoding of 18S, 28S, and COI genes and retrieved paraphyly of the family Catenulidae, with S. evelinae forming a sister group to the genus Paracatenula, and hence we propose a revision of its systematic position. In total, we recorded six species, including three with a wide cosmopolitan distribution (C. turgida, S. grande and S. tuberculosum), and provided sequences for five of them, three of which had no previous molecular records (S. paraguayense, S. evelinae and S. corderoi). Thus, we confirm that greenhouses represent an important source of exotic species for taxonomic work on microscopic invertebrates.

Cruising speed is a key factor affecting prey-search efficiency and migration range in continuously swimming animals. Tunas and lamnid sharks (e.g., white sharks) have convergently evolved traits for high-speed cruising, including the ability to maintain slow-twitch, aerobic red muscle (RM) warmer than ambient water, known as RM endothermy. Despite their well-known high cruising speeds, kinematic features underlying their elevated speeds remain unclear. Swim speed is the product of tailbeat frequency (TBF; Hz) and stride length (SL, the absolute distance traveled per tailbeat; m). RM endothermy is expected to elevate TBF by enhancing muscle contraction performance. Furthermore, within RM-endothermic fishes, tunas and lamnid sharks may exhibit distinct kinematic features because of differences in caudal fin morphology and tailbeat amplitude. Here, we compiled kinematic parameters from 20 fish species, including five RM-endothermic species, measured in the wild using animal-borne sensors. Comparative analyses showed that, for a given body mass and water temperature, RM-endothermic fishes exhibited 1.9 times higher cruising speed and TBF than ectothermic fishes, while SL remained similar. Within RM-endothermic fishes, tunas exhibited 2.3 times higher TBF than similar-sized lamnid sharks, whereas lamnid sharks showed 1.7 times longer SL than similar-sized tunas. These results indicate that RM endothermy is generally associated with higher TBF, while significant kinematic differences remain between tunas and lamnid sharks. This divergence may be partly explained by the greater caudal fin area and tailbeat amplitude in lamnid sharks. It may also reflect contrasting skeletal types of teleosts and elasmobranchs, which potentially influence body stiffness and swimming kinematics.

Over the past decade there has been increasing interest in the use of microhaplotype markers in autopolyploid taxa. This has been driven by theoretical and observed improvements in signals of allelic dosage, linkage, and heritability. Yet, to date there has been little investigation into the suitability of microhaplotype markers for estimating kinship. Here, we develop the theory of kinship estimation from microhaplotypes, introduce the MCHap microhaplotype caller for autopolyploid populations, and apply these methods to a highly diverse germplasm population of mixed-ploidy Actinidia (kiwifruit and relatives). We find that microhaplotype-based kinship estimates are generally superior to equivalent single nucleotide variant based estimates. This is because microhaplotypes minimize the coalescent signal among alleles which may bias estimates within the context of a recent reference population. Hence, kinship estimates from microhaplotypes more accurately capture the recent demographic history of a population. These findings are supported by both coalescent simulations and the analysis of real data. Our findings are relevant to organisms of any ploidy, but most actionable in highly heterozygous taxa such as Actinidia.

Chorus howls play a central role in wolves' (Canis lupus) social communication, shaping both intra- and inter-pack interactions. While synchrony has been shown to influence perceived threat from outgroups in birds and humans, whether the synchrony of an outside pack's chorus howl plays a role in wolf pack behaviour remains unclear. This study investigated behavioural responses of 19 zoo-housed wolf packs to playbacks of synchronous and asynchronous chorus howls from unfamiliar packs, in comparison to a dove control call. We hypothesized that packs would show stronger intra-pack cohesion and greater spatial withdrawal in response to more synchronous, and thus cohesive - and potentially more formidable - sounding vocalisations. Alternatively, asynchronous howls might elicit stronger responses due to their ambiguity and the "Beau Geste" effect (where asynchronous signals make it harder to infer the number of callers). Synchronous howls resulted in more distance from the sound source and less activity compared with asynchronous howls. This could be interpreted as a more cautious response towards a perceived threat. However, since there was no effect on pack-directed behaviour and no difference between the synchronous and control sound, lower interest or motivation to engage with the stimulus cannot be excluded. In contrast, wolves spent more time near the sound-source after hearing asynchronous howls, approached it more, and oriented towards it for longer compared to synchronous playbacks, a pattern consistent with threat-investigation. In how far these reactions result from a more difficult localisation of synchronous howls or more ambiguity in pack size in the asynchronous playbacks, as well as how this translates to wild wolf behaviour, remains to be investigated.

Understanding how divergence in reproductive traits can promote speciation remains a fundamental question in evolutionary biology. Sperm morphology and kinematics diverge rapidly across species. However, the effect of hybridization between recently diverged species on sperm traits remains unclear, limiting our understanding of how reproductive isolation evolves. Here, we evaluated sperm morphology, kinematics, and trait integration in species of a young (~10,000 years), sympatric Cyprinodon pupfish radiation from San Salvador Island, Bahamas, as well as fertile advanced-generation hybrids between two of these species. We found significant divergence in flagellum length, midpiece area, and sperm velocity among species. In contrast, hybrids displayed transgressive kinematic profiles defined by high velocities, reduced path curvature, and distinct patterns of sperm kinematic integration. Our findings suggest that hybridization between recently diverged species may reorganize the underlying control of sperm locomotor mechanisms, generating novel phenotypes that could contribute to reproductive isolation in the early stages of speciation.