Ecological variation drives morphological differentiation in a highly social vertebrate

A free Plain Language Summary can be found within the Supporting Information of this article. Adaptations to ecological conditions may include morphological and physiological attributes, behavioural traits, social structure or a combination thereof (DeWitt et al., 1999; Heynen 2017; Koenig, 1981; Willmer 2005). These adaptations reinforce compensate each other Marshall & Wund, Mushagalusa 2019). For instance, under high predation pressure, freshwater snails (Physa gyrina) evolved narrow shell apertures rendering them less vulnerable shell-entry predation. Along with resulting reduced vulnerability, individuals were also bolder show more antipredator behaviour 1999). In animals, variable factors directly affect interactions by changing costs benefits group living (Arnold Owens, Curry, 1989; Koenig 2011; 2016). addition, variation trigger in phenotypic which may, turn, feed back on organization (Cannizzo 2019; Zuk 1998). example, parasite infections lead changes female red junglefowl's Gallus gallus comb size, is accompanied shift hierarchy so that infected females lose their dominance consequently attain lower status (Zuk Therefore, it seems appropriate study effects different attributes organisms understand mutual influence striking diversity structures. local ecology consistent differences between populations. Two distinct processes are commonly assumed responsible for such divergence. First, physical geographical restrictions gene flow cause genetic populations through drift, result differentiation as by-product (Bolnick Otto, 2013; Slatkin, 1993; Wright, 1943). Second, from divergent natural selection populations, leading (Nosil 2005; Schluter, 2000; Sexton 2014; Wang Summers, 2010). However, fluctuating habitats, fitness phenotypes considerably diverge generations (Bradshaw, 1965; Miner Ruehl DeWitt, West-Eberhard, 2003). Fluctuating pressures should hence diminish maintaining heritable Scheiner, 2004; Nosil 2018). The ability plastically respond particular environment expressing DeWitt McCollum Van Buskirk, 1996) happen either life span an individual (Young 2003) across (Agrawal any case, adjustments sort typically involve responses strongly (Cantor 2020; Groenewoud Morphological seem plastic than, behaviour. They might vary primarily due (Falconer, 1981). Still, has been well established plasticity similarly important adaptation (Krueger Dodson, Rohner Moczek, 1989). Fishes one best studied groups showing remarkable morphology response various factors, linked (e.g. Eklöv Jonsson, 2007; Imre 2002; Olsson Eklöv, Riesch 2018; 2011). Among driving risk habitat complexity, induce body shape (Brönmark Miner, 1992; Frommen Ghalambor Stauffer van Snik Gray, Taylor McPhail, 1985, 1986). risk, fishes often deep potential benefits, including handling success gape-size limited predators Nilsson Brönmark, 1995; Weber 2012), higher burst swimming performance enhanced turning rate (Domenici 2008; Langerhans Reznick, 2010; Webb, 1982, 1984). Improved especially advantageous highly structured environments (Langerhans, 2009; Schrank Walker, 1997; 1984; Webb 1996; Weihs, On contrary, species relying access shelters hide breed, depth constrained size available (Gashagaza Takahashi 2009). renowned impressive organization, ranging solitary assembling loose shoals long-lasting, complex societies (Godin, Hara Pitcher, 1986; Taborsky, This biotic pressure) abiotic type; review, see Taborsky Wong, 2017). long-lasting adjust (Bergmüller 1985; Zöttl 2013) (Groenewoud 2016; Tanaka 2016) adapt prevailing conditions. Such coincide divergence traits varying conditions, morphology. covariance exposed deviating hitherto not understood. cooperatively breeding cichlid Neolamprologus pulcher offers great opportunities interplay sociality animals. N. endemic Lake Tanganyika where breeds colonially ecologically diverse benthic habitats along sublittoral zone (Balshine 2001; Heg Colonies stable over several years (Heg Stiver 2004) dispersal distances rarely exceed 20 m (Stiver 2004, 2007). Groups consist pair, average assisted five six helpers age cohorts Within groups, engage direct brood care, territory maintenance defence (Balshine-Earn 1998; Bergmüller Territories usually contain clefts rocks, burrows stones empty gastropod shells providing shelter serving chambers. defended against hetero- conspecific intruders, mainly done large members (Bruintjes Jungwirth 2015). Variation settings (including complexity substrate quality) pronounced impact Here we ask whether multidimensional niches relate pulcher. We used three approaches. quantified wild-caught eight structural complexity. compared data assess explain differentiation. tested neutral (i.e. variance 13 microsatellites distributed genome unlikely have effects) these examine restricted drift. If arose flow, would expect find correlation Third, disentangle variation, collected specimens wild-caught, F1 F2 fish raised common garden laboratory. selected at southern tip Tanganyika, near Mpulungu, Zambia, according marked rock cover predator abundance (for details Section 2.2; Table 1; Figure 1b). Throughout manuscript, refer numbers 1–8, correspond scores niche axis ascending manner details, 2). 140 21 km apart (Table S1). geographically clearly separated others, whereas cluster two (Figure 1a). Wild-caught n SL (mm) 5F (26–30) 10M (25–37) 9F (50–73) 7M (55–66) (28–42) 6M (29–40) 28F (31–56) 25M (31–65) (50–61) 8M (55–70) (29–43) 9M (31–42) 4F (38–54) (34–56) (44–62) (53–70) (30–39) 0M (NA) (33–49) 1M (39) (48–69) (54–71) (30–40) 4M (40–42) 8F (31–50) (33–52) 11F (29–45) 7F (43–64) (54–78) 13F (30–42) (29–41) 0F 11M (57–70) 3F (27–49) (37) (55–71) (53–80) 14F (28–41) (32–43) analysis, tissue samples 34 breeder-sized population September November 2014 2015. Individuals caught haphazardly using fence nets complete colony, small part (approx. 1 mm 2 mm) dorsal anal fin was removed preserved 98% ethanol. Afterwards, they released again. geometric morphometric analyses, 151 2012 2013 nets. all parts measured standard length (SL) specimen determined its sex visual inspection genital papilla. Males only opening (urinary pore), pore oviduct; Popma Masser, euthanized overdose KOI MED® Sleep buffered 4% formalin. Determining possible individual, therefore excluded further analyses. December brought (populations 1, 3, 4, 5, 7 8) laboratory Ethologische Station Hasli (Bern, Switzerland). breed first subsequently individuals, (see ‘Appendix S1’ detailed description housing conditions). sampled 119 originating 75 families (specimens per family: range = 1–4, median 2) 102 49 noted before fixing Sexing because had reached sexual maturity, From 2012–2015, parameters SCUBA diving. focussed shown cichlids Domenici methods described detail al. (2016) Josi (2018). brief, four transects × 10 laying sisal fibre ropes lake bottom, starting centre population. Predation assessed counting number predatory (Lepidiolamprologus elongatus, Lepidiolamprologus attenuatus Lamprologus lemairii) transects. Each transect surveyed 6 times days, account activity. mean (>10 cm) L. elongatus served proxy most species, posing sizes. sand measure sandy stretches being flat little third dimension, while rocky patches exhibit much dimension. percentage every square meter transects, calculated means site. randomly inhabited (ntot 127, range: 10–21 population) entrances. Following (2016), transparent sheet rolled cylinder inserted into entrance. There, slowly unrolled, until diameter equalled entrance, ruler 1). constructed model combining principal components (PCs). Given strong 3.1), PC one-dimensional analysis ‘ecological niche’ following. created set radiographs Faxitron LX 60 X-ray device. visible bone fractures deformations shapes (13 18 F1, 15 F2; final sample sizes). All observed occurred skull, except distorted spine. To quantify morphology, 19 landmarks placed homologous skeletal structures S1) TpsDig 2.3.2 software (Rohlf, R (R Core Team, 2018) packages geomorph (Adams 2019) Morpho (Schlager, 2017) perform statistical performed generalized Procrustes correct position orientation radiograph. Then, ran ANOVA random residual permutation procedure (1,000 iterations) identify overall fish. explanatory variables. Subsequently, applied post-hoc pairwise comparison disparity estimated significance iterations). conducted component (PCA) main axes shape. PC1 explained 25.2% accounted head caudal peduncle S3) known typical artefacts fishes. biologically meaningful, arise during fixation slight posture captured PC1s intraspecific comparisons dorsolateral bending body; Fruciano, Larochelle Valentin 2008). contained dorsoventral pelvic insertion S3), signal likely represents current condition stomach fullness spawning condition). Although subtle signatures anteroposterior elongation shortening relevant, cannot disentangled artefacts. contrast, PC2 (explaining 21.9% variance; covariation describing depth) considered informative respect focus study, S3). sharp drop after PC2. proceeded following analyses considering avoid results. corresponds practice morphometrics Albert Hudson Hüllen 2021). effect analysed linear mixed-effect (LMM) package lme4 (Bates (PC2) variable, included effect. distance drift (Hereford, 2009), fitting multiple regression matrices (MRM) r ecodist (Goslee Urban, geographical, distance, Euclidean shortest water 1a; FST values 2B). Significance running 1,000 permutations. test persists generations, projected PCA morphospace if correlated separate models (LM). LM, fitted score variable. second corrected testing implementing Benjamini–Hochberg method (Benjamini Hochberg, 1995). processing material followed (2019): microsatellite markers polymorphic loci (UNH154, UNH106 (Lee Kocher, 1996); NP007, NP773, ULI2 (Schliewen 2001); TmoM11, TmoM13, TmoM25, TmoM27 (Zardoya Pzeb4 (Van Oppen 1997); UME003 (Parker Kornfield, UNH1009 (Carleton 2002); Ppun21 (Taylor 2002)). Some sequences optimized (Kotrschal 2012). Using QIAGEN® Multiplex PCR Kit, amplified DNA allowing co-amplification locus-specific, fluorescently labelled primer pairs single PCR. sets containing pairs, respectively. PCRs took place µl cocktail consisting genomic DNA, 5 2x QIAGEN Master Mix, 3 H2Odd 10x mix. Primer mix fluorescent-labelled forward non-labelled reverse end concentrations 0.04–0.06 µM each, intensity respective amplification products. fluorescent dyes 6-FAM (blue), Yakima Yellow (green), HEX ATTO550 (yellow), ATTO565 (red) (Microsynth), PET VIC (green) (Thermo Fisher). GeneAmp® 9700 System (Applied Biosystems) cycling parameters: min 95?, 35 cycles 95? 30 s, 57? 72? s step min. Fluorescent fragments visualized capillary electrophoresis ABI 3100® Genetic Analyser Biosystems). GeneScan 500 LIZ Fisher) internal GeneMarker® Analysis version 2.4.0 (SoftGenetics). GenAlEx 6.5 (Peakall Smouse, 2006, 2012) conduct marker Hardy–Weinberg equilibrium (HWE) After correcting Bonferroni (Bland Altman, 1995), 11 remained HWE population, subsequent S2). computed global index (FST) molecular (AMOVA) permutations significance. measures population: loci, alleles locus, effective alleles, expected heterozygosity, loci. Bayesian clustering assignment STRUCTURE 2.3.4 (Falush 2003, Hubisz Pritchard 2000). independent runs K 10, burn-in period 100,000 iterations Markov Chain Monte Carlo algorithm. identified optimal Harvester (Earl, Evanno (Evanno Additionally, MRM purpose, above. interest (predation cover) (Spearman correlation, 8 populations; risk—shelter size: 0.83, p 0.015; risk—sand cover: ?0.69, 0.069; cover—shelter ?0.90, 0.004). combined single, axis. 83.5% habitat. loaded (|loading| > 0.52), loadings positive, negative. continuously Body significantly differed (Procrustes ANOVA, df 7, Z 9.91, < 0.001; tended deeper than females, though statistically significant (df 1.33, 0.095). Comparing showed 4 shallow bodies, bodies S4). Populations 6, eye (LMM, intercept: ?0.0002, ? ± SE: 0.0047 0.0018, t 2.59, 0.014). relationship (MRM, 0.0095, ?: 0.0022, 0.09; 3a), but neither (?: ?0.1157, 0.35; 3b) nor 0.0003, 0.34; 3c). Average generation (LM, ?0.0016, 0.648 0.190, 3.41, 0.013; 4a). ?0.0076, 0.823 0.330, 2.49, 0.027; 4c), there no ?0.0091, 0.462 0.323, 1.43, 0.115; 4b). AMOVA revealed among (FST 0.037, 0.001). Population differentiated another, non-significant trend 0.004 [populations 7] 0.071 4]; 2b). Microsatellite moderate levels allelic evidence allele 2a). suggested clusters S5 results Ks calculated). similar frequencies, 2, compositions frequencies 1c). 0.0313, 0.0008, 0.51; 5a) ?0.0009, 0.76; 5b). Animals levels, environmental level potentially constrain enhance adaptive strategies. Our cichlid. inherited suggesting reflects link differentiation, indicating simple consequential increased representing Deep predicted evolve counter-selected sizes Larouche finding deepest featuring pressure suggests serve protection hampered body. crucial pulcher, defend territories crevasses, excavated chambers underneath Interestingly, largest located differ settings. largely absent, rocks scarce. intense, crevasses cavities shelters. shows spatial scale absence (cf. Conover 2006). Comparable lamprologine Telmatochromis temporalis, inhabits namely rocky-sandy bottoms beds (Takahashi, apparently induced evolution morphotypes dependence niches. morph occurs shelters, dwarf (Takahashi present positively other, do so, necessary did appear our indicate suitable scarce general. increase increasing variation. amounts corroborating appearing homogeneous still exhib