Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

Lucy H. Underwood; Aimee van der Reis; Andrew G. Jeffs

doi:10.1002/aff2.113

Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

Underwood, Lucy H.; van der Reis, Aimee; Jeffs, Andrew G. 2023-06-01 00:00:00 INTRODUCTIONCoastal aquaculture structures operate as artificial habitats for fish through the provision of shelter and food resources (Alleway et al., 2019; Callier et al., 2018; Theuerkauf et al., 2022). Aquaculture habitats are known to attract and aggregate some fish species (e.g. Siganidae species in seaweed farms [Hehre & Meeuwig, 2016] and Sparidae species in mussel farms [Šegvić‐Bubić et al., 2011]), enabling them to reach greater abundance within farms compared to sites without aquaculture structures (Barrett et al., 2022). Some fish species are also known to settle and recruit directly into non‐fed aquaculture structures, utilising the structure as a nursery habitat (Underwood & Jeffs, 2023). It is often assumed that the aggregation of fish at aquaculture structures is a response to the input of feed, the abundance of the farmed species or the associated assemblage of potential prey items due to the natural colonisation of aquaculture structures (e.g. biofouling) (Barrett et al., 2022; Callier et al., 2018; Sanchez‐Jerez et al., 2011; Theuerkauf et al., 2022). Research in this area has largely focused on differences in the abundance of the functional groups of fish living within or in surrounding aquaculture habitats, to infer responses to the food resources which may be available (Clynick et al., 2008; Theuerkauf et al., 2022). However, few studies have directly analysed the composition of the diet of fish within aquaculture habitats, and of these most have focused on fed finfish aquaculture (Anyango et al., 2017; Fernandez‐Jover & Sanchez‐Jerez, 2015; Fernandez‐Jover et al., 2009; Hayden, 1995; Hehre & Meeuwig, 2016; Šegvić‐Bubić et al., 2011; Skog et al., 2003).In non‐fed aquaculture practices (e.g. shellfish), studies have focused on the predatory pressures that wild fish can have during the early stages of production, particularly predation of mussel spat (Peteiro et al., 2010; Šegvić‐Bubić et al., 2011; Stenton‐Dozey & Broekhuizen, 2019). For example, members of the Sparidae family (seabreams and snapper) are commonly sighted in large schools stripping recently seeded mussel lines (Gerlotto et al., 2001; Peteiro et al., 2010; Stenton‐Dozey & Broekhuizen, 2019). These fish are generalist predators and tend to feed based on prey availability rather than selectively (Šegvić‐Bubić et al., 2011; Usmar, 2012). The gut contents of seabream sampled from mussel (Mytilus galloprovincialis) farms identified mussels as the dominant prey (69.7% biomass), although a diverse range of other prey items (e.g. gastropods, fishes and macroalgae) were also found (Šegvić‐Bubić et al., 2011). This provides evidence of what a member of the Sparidae family may feed on within mussel farms, and highlights the concurrent and/or selective epibiota consumption.Methods to analyse fish diet have traditionally relied on gut dissections and morphology as a precise method to determine the possible food items that are sourced from local habitats, and to compare the differences among habitats (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Braga et al., 2012). However, several other methods exist including visual observations, stable isotope analysis, fatty acid analysis and DNA metabarcoding (Braga et al., 2012; Castro et al., 2008; Udy et al., 2019). Visual observations of the feeding behaviour of fish usually provide high‐level information on patterns of feeding but little detail on the relative components of their diet (Amundsen & Sánchez‐Hernández, 2019; Braga et al., 2012; Cole, 2010; Morton et al., 2008). Non‐lethal methods are less common and have only been trialled on few fish species (Barnett et al., 2010; Castro et al., 2008; Kamler & Pope, 2001; Rennó Braga et al., 2017; Trkov & Lipej, 2019). Non‐lethal methods are species specific, are invasive and require technical extraction experience. However, they can still result in mortality and generally are not suitable for most studies of fish diet. Stable isotope and fatty acid analyses are better suited to inferring broad‐scale trophic linkages and changes to diet over time rather than identifying specific dietary components (Boecklen et al., 2011; Budge et al., 2006; Udy et al., 2019). DNA metabarcoding is suited to the latter and can identify digested gut content items previously ‘unidentifiable’ due to the lack of morphological structure (Berry et al., 2015).A range of methods can be used to sort and quantify gut contents, including numerical frequency of prey items, gravimetric methods and volumetric composition (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Mahesh et al., 2018). Although there have been attempts to standardise the quantification metrics among visual gut content investigations, suitable metrics vary based on the objectives of the study, and each quantification metric has biases making some more suitable metrics for certain species than others. For example, numerical frequency methods are most suitable for fish that consume similar sized prey items to reduce bias towards smaller numerically dominant species, which may have limited volumetric importance (Amundsen & Sánchez‐Hernández, 2019). The relative gut fullness method is a modified volumetric method that allows for each prey grouping to be proportionately allocated, while also standardising the entire contents by the fullness of the gut. Therefore, it is an effective approach when samples have varied gut fullness, and for accommodating a range of prey types. DNA‐based methods (e.g. DNA metabarcoding) can increase taxonomic accuracy by allowing for the identification of the presence of prey items when their remains are visually unidentifiable in gut contents (O'Rorke et al., 2012; van der Reis et al., 2020). However, the results are semi‐quantitative (Deagle et al., 2019), and species‐level identification is reliant on the availability of an adequate taxonomic DNA reference library (Amundsen & Sánchez‐Hernández, 2019; van der Reis & Lavery, 2020). Therefore, the combination of visual gut analysis and DNA‐based methods is considered an effective approach for analysing gut contents.The overall aim of this research was to compare the diet of snapper (Chrysophrys auratus), a common coastal demersal fish species, within and outside established green‐lipped mussel longline farms (Perna canaliculus) in the Coromandel Harbour, Firth of Thames, New Zealand. Snapper are commonly present as adults and tend to become resident within coastal mussel aquaculture sites in New Zealand (Gibbs, 2004; Stenton‐Dozey & Broekhuizen, 2019). These fish are generalist predators that have marked ontogenetic shifts in their diets that are regulated by age and the habitat within which they reside (Usmar, 2012). Therefore, it could be expected that their diet would reflect their use of the surrounding habitat. This research aimed to test whether the diet of snapper collected from mussel farms would differ from those collected in adjacent areas, with the differences likely relating to prey items sourced directly or indirectly from the mussel aquaculture infrastructure.METHODSSite locationsSnapper were sampled from four sites in the Firth of Thames. Two were long established longline green‐lipped mussel farms and two were sites without mussel farms (i.e. control sites), all located over similar unstructured soft sediment habitat of similar depth (8–13 m) (Figure 1). The farm sites were Motukopake Island (36° 45′ 2.87″ S, 175° 25′ 22.8″ E), and Rat Island (36° 45′ 25.9194″ S, 175° 26′ 59.99″ E). The control sites were at least 500 m away from the farms and were Motukopake Island control site (36° 57′ 32.03″ S, 175° 28′ 51.59″ E) and Rat Island control site (36° 45′ 23.03″ S, 175° 27′ 39.6″ E; 36° 45′ 46.07″ S, 175° 26′ 49.2″ E). Two locations, with the same habitat features, were collectively used for the Rat Island control site due to the low numbers of fish caught at the first control site. The mussel aquaculture operations comprised a series of paired parallel backbone lines held near the surface by large plastic floats that support suspended loops of dropper ropes covered with attached mussels. The dropper ropes extended 6–8 m below the surface floats and were held several metres above the soft‐sediment seafloor.1FIGUREThe sampling sites within the Coromandel Harbour, Firth of Thames where snapper were captured using hook and line fishing methods from the seafloor within mussel farms and from nearby control sites at both Motukopake Island and Rat Island.Sample collection and processingAdult snapper only within a size range of 26–42 cm fork length were sampled in May and June 2022. Hook and line fishing methods were used, with plastic soft‐bait and lures utilised as much as possible to avoid possible contamination of gut contents by bait. Where natural bait was used, a minimal amount of bait was deployed of a readily recognisable species that were not common in the area (i.e. pilchard, squid or mullet). Crepuscular periods of snapper feeding activity were primarily targeted for the sampling to increase the probability of capturing snapper with gut contents. Sixteen fish were obtained from each of three sampling sites (Motukopake Island mussel farm, Motukopake Island control site and Rat Island mussel farm), while 13 snapper were sampled from Rat Island control site. Immediately upon capture, all fish were humanely euthanised according to animal ethics approval (NZ Animal Welfare Act 1999, UoA‐AEC Approval # 21619), labelled and put into salted ice slurry. Each snapper was measured (fork length) and weighed back on land (2–6 h after capture) before being frozen for subsequent gut analyses. Previous studies have indicated that freezing snapper gut contents was an appropriate method of preservation that facilitates both visual, molecular genetic and biochemical analysis of gut contents to be undertaken (Drummond, 2020; Supono et al., 2021; Third, 2022).Snapper gut dissectionFrozen snapper samples were thawed at room temperature and then dissected. The alimentary tract was removed via incisions at the oesophageal opening and at the anus. The foregut and hindgut were separated and each opened. A gut fullness score was estimated for both the foregut and the hindgut, with 0 being completely empty and 10 being completely full (Third, 2022). Subsequently, the gut fullness score was averaged between the hindgut and the foregut, so that there was only one score per snapper gut. Similarly, a digestion score was calculated to provide an estimate of the extent of digestion for the prey items for the hindgut and the foregut individually, with a digestion score of 0 being no digestion and 5 being fully digested (Third, 2022). Subsequently, the digestion score was averaged between the hindgut and the foregut, so that there was only one score per snapper gut. Any bait identified in the foregut was removed, and not considered further in this study other than for exclusion from any DNA signal detected. The remaining material in the foregut was weighed, followed by the foregut lining after the gut contents had been emptied. The process was repeated for the hindgut.Visual analysis of snapper gut contentsGut contents for the foregut and hindgut were combined and then spread onto a sterile tissue culture dish (2 × 2 cm grid) and sorted into groups of similar prey items, for each sample. Distilled water was used to separate and clean contents. Each prey group was classified to the lowest practical taxonomic level. An adapted version of the relative‐fullness method was used to quantify the proportions of each prey grouping in each individual snapper gut (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Binning & Chapman, 2010). The relative‐fullness method then uses proportions of each prey group and standardises by the fullness of the gut to calculate ‘points’ for each individual prey group within the snapper gut contents. This was undertaken by firstly estimating the two‐dimensional (2D) coverage of each prey group in the culture dish (e.g. Bivalvia covering two squares). All prey groups were then summed together to estimate the total coverage within each snapper gut, including the unidentifiable digested material which had its own category. This estimate was then used to calculate the proportion of each prey group within the gut of the sampled snapper. Proportions were multiplied by the average fullness estimate (foregut and hindgut calculated independently but averaged to produce one fullness score) to calculate the relative proportion of each prey group within each individual snapper, thereby standardising the relative proportion of each prey group to facilitate comparisons among samples. For example, if the proportion of the gut content of a snapper was 0.8 (i.e. 80% coverage of the gut contents in the culture dish) for green‐lipped mussel (P. canaliculus) and the average fullness estimate was 3, this would be calculated as 0.8 × 3 which equals a relative proportion of 2.4 points.Genetic analysis of snapper gut contentGut content removal and DNA extractionOnce gut contents for individual fish had been extracted and analysed for visual gut analysis, the gut contents were mixed and a 2‐mL subsample was preserved in 90% ethanol (2 mL cryovial) and stored at −20°C. All dissection tools were sterilised between handling individual snapper.DNA was extracted from the subsample using the Nucleospin Tissue DNA extraction kit (Macherey‐Nagel) following the manufacturer's instructions. Controls were included in the extractions, that is DNA extraction blank where no content was added. The DNA quality and quantity were viewed on 0.8% agarose, visualised with gel red (Biotium) using a Gel Doc XR+ (Bio‐Rad).DNA amplificationSpecies identified microscopically from the gut contents were compared to sequences available on GenBank. The results indicated that cytochrome oxidase one (COI) would have the best coverage, and thus COI was targeted for amplification during the polymerase chain reactions (PCRs). The universal primer pair selected was M1COIintF (forward; Leray et al., 2013) and jgHCO2198 (reverse; modified replacing ‘I’ with ‘N’; Geller et al., 2013). Illumina Nextra adapters were added to the primer sequences. The PCRs were done in triplicate using 6.25 μL of MyTaq Red Mix (Bioline; Meridian Bioscience), 0.25 μL of each primer (10 μM), 1 μL of bovine serum albumin (1%) and 4.75 Ultra‐Pure DNase/RNase free water (Invitrogen; Thermo Fisher Scientific). Each PCR cycle included the DNA blanks and a PCR blanks (no DNA added) to check for possible contamination. The PCR protocol was a modified version of Lobo et al. (2013): 94°C for 60 s, 35 × [94°C for 30 s; 54°C for 90 s; 72°C for 60 s] and a final extension at 72°C for 5 min (van der Reis et al., 2023). Amplification was checked by running the PCR products on a 1.6% agarose, and visualised as outlined above.Sequencing and bioinformaticsPCR triplicates were pooled per sample and then cleaned using AMPure XP (Beckman Coulter), following the Illumina 16S protocol (Illumina, 2013). The concentration of the cleaned PCR products was determined using Qubit dsDNA HS Assay Kit (Invitrogen) following the manufacturer's instructions and subsequently diluted to 2 ng/μL before sequencing. Sequencing was done by Auckland Genomics and undertaken on an Illumina MiSeq (2 × 250 paired‐end nano run; single lane used).Cutadapt v3.5 (Martin, 2011) was used for primer removal, retaining only those sequences with exact primer matches anchored at the beginning of the sequence. Qiime2 v2022.2 (Bolyen et al., 2019) was used to assess the sequence quality prior to truncation in DADA2 (Callahan et al., 2016; utilised within Qiime2). High‐quality amplicon sequence variants (ASVs; Callahan et al., 2017) were produced by filtering for sequence quality, retaining only sequences that merged and that were identified non‐chimeric.ASVs were run through GenBank's BLAST database v2022‐07 using the MegaBLAST function within the BLASTn suite (Benson et al., 2013; minimum E‐value threshold of 0.001 and a minimum percentage identity of ≥70%). Resulting taxonomic assignments were then run through the World Register of Marine Species (WoRMS) database (Horton et al., 2022) to confirm taxonomy and filter to retain only marine species (i.e. remove any spurious terrestrial sequences). R v4.0.4 (R Core Team, 2021) was used for final sequence filtering and quality assurance of the taxonomic assignments. Potential contamination was accounted for by a proportional subtraction of the controls (DNA and PCR blanks were sequenced). Sequences for snapper (C. auratus) (host species) were removed. Only sequences >280 base pairs that had ≥85% identity match to the taxonomic sequence were used. Sequences were also filtered to retain only those with ≥90% query coverage.Statistical analysesAll statistical analyses and plots were produced using R (R Core Team, 2021). All analyses used a significance of α = 0.05. All means are presented as mean ± standard error. Plots were prepared with ggplot (package version 4.0.4, ggplot2 function).Snapper metricsTo compare differences in snapper length, a two‐way analysis of variance (package version 4.0.4, aov function) was used with the main factors of Location (Motukopake Island or Rat Island) and Treatment (mussel farm or control). Data were plotted and visually assessed to confirm parametric assumptions were met. Tukey's honest significant difference (package version 4.0.4, TukeyHSD function) post hoc tests were conducted if an overall significant difference was identified. This approach was repeated to compare the differences in the wet weight (g) for each of foregut and hindgut contents separately. Only foregut wet weight data needed to be transformed prior to analyses by using a log(x + 1) transformation. A generalised‐linear model (package version 4.0.4, glm function) was used to compare the differences in the combined average digestion score for the hindgut and the foregut, with a quasi‐Poisson distribution and the main factors of Location and Treatment. If an overall significant difference was identified, ‘emmeans’ (package version 1.5.5‐1, emmeans function) post hoc analyses were used.Visual gut analysisTo compare the amount of prey groups at a broad taxonomic level in snapper gut contents among samples, non‐metric multidimensional scaling (NMDS) was conducted with a Bray–Curtis distance and a permutation value of 999, using points data derived from the relative‐fullness method. An ordination plot was produced for a 2D representation of the ordination, separated by Treatment. A permutational multivariate analysis of variance (PERMANOVA) was undertaken using points data of prey groups at the broad taxonomic level, with the Bray–Curtis distance and the main factors of Location and Treatment. Where the overall model results were significant, a Wilcoxon test was used to determine differences in the amount (quantified as ‘points’ from the relative‐fullness method) of prey groups at the broad taxonomic level consumed by snapper among sites. A Bonferroni adjustment was used for p‐values to account for potential error inflation from multiple comparisons.For prey groups at the lowest taxonomic level, only presence/absence data were used for each prey group identified within each individual snapper. NMDS was conducted with a Bray–Curtis distance and a permutation value of 999, using presence/absence data of the different prey groups at the lowest taxonomic level. An ordination plot was produced for a 2D representation of the ordination, separated by Treatment. A PERMANOVA was undertaken for prey groups at the lowest taxonomic level using presence/absence data, with the Bray–Curtis distance and the main factors of Location and Treatment. Where the overall model results were significant, a Fisher's exact test was used to determine whether the frequency of each prey group at the lowest taxonomic level found within snapper gut contents was significantly different between Treatment and Location. A Bonferroni adjustment was used for p‐values to account for potential error inflation from multiple comparisons.DNA metabarcoding analysisThe presence/absence data of prey groups identified from DNA metabarcoding at the family level and class level for gut content of all individual snapper were analysed separately. A PERMANOVA was undertaken on the presence/absence data for detected prey separately for the family level and at the class level with the Bray–Curtis distance using the main factors of Location and Treatment. If significance was detected, a Fisher's exact test was undertaken using presence/absence data of the prey groups. Family‐level comparisons among the samples of gut content excluded 13 snapper samples where no families were identified (Motukopake Island control site = 2, Motukopake Island mussel farm = 3, Rat Island control site = 3 and Rat Island mussel farm = 5). Class‐level comparisons excluded nine snapper samples where zero taxonomic classes of prey were identified (Motukopake Island control site = 3, Motukopake Island mussel farm = 1, Rat Island control site = 2 and Rat Island mussel farm = 3).RESULTSSnapper metricsSnapper length ranged from 27.4 to 41.2 cm at Rat Island mussel farm, 27.0 to 41.6 cm at Motukopake Island mussel farm, 25.8 to 30.2 cm at Rat Island control site and 26.8 to 37.4 cm at Motukopake Island control site (Figure 2). There were differences in mean snapper length for the sampled snapper for Treatment, that is mussel farm versus control site (F(1, 57) = 5.41, p = 0.024) and the interaction Treatment × Location (F(1, 57) = 9.51, p = 0.0032), but not for Location, that is Motukopake Island versus Rat Island (F(1, 57) = 1.04, p = 0.31) (Figure 2). Post hoc analyses identified that snapper sampled at the Motukopake Island control site were larger than at the Rat Island control site (p = 0.024; 31.6 ± 1.0 and 27.8 ± 0.3 cm, respectively), while snapper from Rat Island mussel farm were larger than those sampled from the Rat Island control site (p = 0.002; 32.8 ± 1.0 and 27.8 ± 0.3 cm, respectively).2FIGURESize frequency of snapper (Chrysophrys auratus) length (as fork length [cm]) (bin width = 2 cm) sampled at four sites in the Firth of Thames (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16). Red lines indicate mean snapper length for each site.The gut contents of snapper sampled at mussel farm sites had a range of 1–5 in digestion scores, and control sites had a range of 2–5 in digestion scores (Figure 3). There were overall significant differences in mean digestion scores between the two Locations and Treatments (F(57, 60) = 2.79, p = 0.04). The main factor effects of Location and Location × Treatment were significant (p = 0.04 and p = 0.01, respectively), but not the main factor Treatment (p = 0.42). Post hoc analyses identified that Rat Island control site had a significantly higher digestion score compared to Motukopake Island control site (p = 0.04; 0.9 ± 0.24 and 3.3 ± 0.16) and Rat Island mussel farm (p = 0.005; 3.9 ± 0.24 and 2.9 ± 0.29).3FIGURETop figure mean digestion score of gut contents of snapper sampled from each of the four sites (Motukopake Island control site, Motukopake Island mussel farm, Rat Island control site and Rat Island mussel farm). A digestion score of 0 represents completely undigested material and a score of 5 represents fully digested material (error bars represent standard error). Bottom figure mean wet weight (g) of snapper foregut and hindgut contents for the four sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat control site, n = 13; Rat Island mussel farm, n = 16).Total wet weight of snapper gut contents ranged from 1.26 to 30.08 g at Rat Island mussel farm, 1.29 to 17.99 g at Motukopake Island mussel farm, 3.26 to 15.33 g at Rat Island control site and 2.17 to 4.50 g at the Motukopake Island control site. Every snapper had some gut contents present; however, the majority of gut contents were in the hindgut. There were no differences in the mean wet weight of the foregut contents between Treatments (F(1, 57) = 0.16, p = 0.69), Location (F(1, 57) = 0.75, p = 0.39), or for the interacting effect of Treatment × Location (F(1, 57) = 2.56, p = 0.12) (Figure 3). Similarly, there were no differences in mean wet weight of the hindgut contents between Treatments (F(1, 57) = 1.36, p = 0.25), Location (F(1, 57) = 0.004, p = 0.95), or for the interacting effect of Treatment × Location (F(1, 57) = 0.89, p = 0.35) (Figure 3).Gut content composition in visual analysisThe gut contents of snapper were moderately to highly digested for all four sites (Figure 3), which made the identification of prey groups difficult, and required piecing together parts of the organism to identify what taxa it was. Most of the material came from the hindgut, with 25% of snapper in mussel farm sites and 17% of snapper in control sites with no material in the foregut. The Rat Island control site also had a higher total digestion score (average of hindgut and foregut) and a higher proportion of digested material compared to other sites, which is likely due to time since prey consumption rather than types of prey consumed. Overall, the digestion of the gut contents meant that the taxonomic level to which prey groups could be identified was most often limited. For example, the remains of Decapoda within the gut contents of snapper sampled at Rat Island control sites could not be identified to lower taxonomic levels, for example Caridea, Brachyura and Paguroidea.Gut content analysis for prey groups at broad taxonomic levelBivalvia were the dominant prey group in snapper gut contents, especially for the two mussel farm sites (Figure 4; Table 1). This was due to the high proportion of mussels present in the gut contents (i.e. 29% in Rat Island mussel farm and 18% in Motukopake Island mussel farm vs. 0% for both the Motukopake Island and Rat Island control site), which are within the Bivalvia category at the broad taxonomic level (Figure 4). In contrast, Decapoda were a common component of the gut contents at one control site but not at either of the mussel farm sites (i.e. 22.5% in Rat Island control site vs. 0% in Rat Island mussel farm) (Figure 4). Other Crustacea (i.e. excluding those prey groups that were classified into different Crustacea taxa such as Caridea, Decapoda, Brachyura and Paguroidea) were higher in mussel farm sites (i.e. 14.6% in Rat Island mussel farm vs. 1.4% in Rat Island control site), and Paguroidea was higher in control sites (i.e. 9.5% in Motukopake Island control site vs. 0% in the Motukopake Island mussel farm) (Figure 4; Table 1).4FIGUREMean percentage of prey groups at the broad taxonomic level present in snapper gut contents (hindgut and foregut combined), as a relative proportion of each prey group based on the total gut contents within each individual snapper. The relative proportions of each broad taxonomic group in individual snapper were averaged for each of the four sampling sites (Motukopake Island control site, Motukopake Island mussel farm, Rat Island control site and Rat Island mussel farm). Mussels (green‐lipped and blue) extracted from other Bivalvia for visual purposes only and not as part of the statistics at the broad taxonomic level.1TABLETotal amount of prey groups at the broad taxonomic level in gut contents of snapper sampled from each of four sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16)Prey groups at the broad taxonomic level in snapper gut contentsMotukopake Island control siteMotukopake Island mussel farmRat Island control siteRat Island mussel farmDigested material18.9415.3414.4813.25MolluscaBivalvia4.716.530.07512.01Gastropoda0.520.000.000.00Chitonida0.110.000.000.20ArthropodaBrachyura3.834.160.381.15Paguroidea3.370.000.830.02Decapoda0.530.945.990.00Other Crustacea0.353.330.426.07Caridea0.510.000.400.11AnnelidaPolychaeta0.040.880.000.52ChordataTeleostei0.000.002.420.00Ascidiacea0.000.320.000.00Shell debris0.10.000.000.16Note: Prey were classified into a broad taxonomic level, and the relative proportion measure (points) identifies the contribution of each group standardised by the gut fullness score, as calculated by the modified relative‐fullness method. Bolded values represent the highest prey group (excluding digested material) per site.The ordination plot of the amount of prey groups at the broad taxonomic level (derived from relative proportion data) in the snapper gut contents identified a partial separation between control and mussel farm sites; however, there is still considerable overlap between treatments (Figure 5). PERMANOVA identified significant differences in the composition of gut contents for Treatment (pseudoF(1, 57) = 6.39, p = 0.001) and Treatment × Location (pseudoF(1, 57) = 3.12, p = 0.01), but not for Location (pseudoF(1, 57) = 1.46, p = 0.18). Post hoc analyses indicated that the prey groups at a broad taxonomic level driving the differences were the categories Bivalvia, other Crustacea, Decapoda and Paguroidea (Table 2). Specifically, snapper sampled at Rat Island control site had consumed fewer Bivalvia compared to snapper sampled from all other sites (Table 2). Additionally, snapper from Motukopake Island mussel farm consumed more other Crustacea than both control sites. Rat Island control site snapper consumed more Decapoda compared to both mussel farm sites. Motukopake Island control site snapper consumed more Paguroidea compared to Motukopake Island mussel farm (Table 2).5FIGURENon‐metric multidimensional scaling (NMDS) plot of the relative proportion measure data of each of 13 prey groups at a broad taxonomic level found in snapper gut contents at four sites and presented for the two control and two mussel farm sites (see Table 1). Confidence interval (50%) ellipses displayed over Treatments.2TABLEWilcoxon test post hoc significant differences in the amount of prey groups at a broad taxonomic level consumed by individual snapper among the sites Motukopake Island mussel farm, Motukopake Island control site, Rat Island mussel farm and Rat Island control sitePrey group at a broad taxonomic levelComparisonMean and standard error (points)Adjusted p‐valueBivalviaRat Island control site < Motukopake Island control site0.006 ± 0.006 < 0.29 ± 0.070.01Rat Island control site < Rat Island mussel farm0.006 ± 0.006 < 0.75 ± 0.250.03Rat Island control site < Motukopake Island mussel farm0.006 ± 0.006 < 0.41 ± 0.130.04Other CrustaceaMotukopake Island control site < Motukopake Island mussel farm0.02 ± 0.02 < 0.21 ± 0.060.006Rat Island control site < Motukopake Island mussel farm0.02 ± 0.02 < 0.21 ± 0.060.04DecapodaMotukopake Island mussel farm < Rat Island control site0.06 ± 0.06 < 0.46 ± 0.170.02Rat Island mussel farm < Rat Island control site0 ± 0 < 0.46 ± 0.170.004PaguroideaMotukopake Island mussel farm < Motukopake Island control site0 ± 0 < 0.21 ± 0.080.04Note: Prey were classified into a broad taxonomic level, and the relative proportion measure identifies the relative contribution of each prey group calculated from the modified relative‐fullness method. Mean and standard error are from the raw data.Gut content analysis for prey groups classified to lowest taxonomic levelPrey identified to the lowest taxonomic level within snapper gut contents produced two sets of data. Firstly, the relative proportion measure derived from the proportions of prey items in gut content combined with the relative‐fullness measure was used to identify the relative amount of each prey group consumed among snapper (Table 3). Secondly, presence/absence data were used to calculate the percentage of individual snapper within a site that had the prey group present (Table 3). The highest amount of prey consumed by Motukopake Island mussel farm snapper was green‐lipped mussels (P. canaliculus); however, the most frequently present prey was the triangle barnacle (Balanus trigonus) (Table 3). At Rat Island mussel farm, the highest amount and most frequently present prey was green‐lipped mussels (Table 3). The highest amount of prey consumed by Motukopake Island control site snapper was wrinkled swimming crab (Liocarcinus corrugatus) and the most frequently present prey was hermit crabs in the Paguridae family (Table 3). At the Rat Island control site, the highest amount and most frequently present prey in snapper gut contents was Decapoda (Table 3).3TABLETotal amount (derived from relative proportion measure data) of prey groups classified to the lowest taxonomic level consumed by snapper, and total percentage of individual snapper with a prey group present (derived from presence/absence data), among the sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16)Prey groups at the lowest taxonomic level in snapper gut contentsMotukopake Island control siteMotukopake Island mussel farmRat Island control siteRat Island mussel farmDigested material18.94 and 100%15.34 and 100%14.48 and 100%13.25 and 93%MolluscaBivalviaUnidentified Bivalvia0.48 and 6%0.000.000.04 and 6%Anomia trigonopsis1.28 and 25%0.03 and 6%0.00Atrina zelandica2.11 and 37%0.000.001.07 and 6%Austrovenus stutchburyi0.34 and 6%0.000.000.00Dosina spp.0.59 and 19%0.000.000.00Mytilus galloprovincialis planulatus0.000.60 and 13%0.004.04 and 31%Ostreidae spp.0.000.000.075 and 8%0.00Paphies australis0.08 and 6%0.000.000.00Perna canaliculus0.005.90 and 56%0.006.85 and 56%GastropodaTrochidae spp.0.52 and 13%0.000.000.00ChitonidaUnidentified chiton0.11 and 6%0.000.000.043 and 6%Cellana ornata0.000.000.000.16 and 6%ArthropodaBrachyuraUnidentified Brachyura0.1 and 6%0.61 and 19%0.38 and 8%0.17 and 6%Halicarcinus innominatus0.001.49 and 6%0.000.00Halicarcinus spp.0.000.18 and 6%0.000.00Liocarcinus corrugatus3.70 and 13%0.000.000.00Nepinnotheres novaezelandiae0.000.20 and 6%0.000.25 and 6%Notomithrax spp.0.04 and 13%0.86 and 13%0.000.72 and 6%Pilumnus novaezelandiae0.000.77 and 13%0.000.00Portunidae0.000.057 and 6%0.000.00PaguroideaPaguridae spp.2.86 and 38%0.000.83 and 23%0.02 and 6%Lophopagurus spp.0.50 and 19%0.000.000.00DecapodaUnidentified decapod0.53 and 13%0.94 and 6%5.99 and 62%–CrustaceaBalanus trigonus0.003.33 and 68%0.42 and 15%5.76 and 44%Epopella plicata0.000.000.000.32 and 6%Meiura0.35 and 6%0.000.000.00CarideaUnidentified Caridea0.02 and 6%0.000.33 and 15%0.11 and 6%Alpheus richardsoni0.000.000.07 and 8%0.00Biffarius filholi0.48 and 6%0.000.000.00AnnelidaPolychaetaUnidentified polychaete0.04 and 6%0.000.000.05 and 6%Eulalia microphylla0.000.45 and 6%0.000.00Serpulidae spp.0.000.43 and 38%0.000.47 and 13%ChordataTeleosteiForsterygion spp.0.000.002.42 and 8%0.00AscidiaceaStyela clava0.000.32 and 6%0.000.00Shell debrisMisc. shell debris0.1 and 6%0.000.000.16 and 6%Note: The relative proportion measure identifies the relative contribution of each prey item calculated by the modified gut fullness method. The percentages represent the proportion of snapper samples that had the prey item present, out of the total number of individuals within each site. Bolded values represent the highest value (excluding digested material) per site.Statistical analysis for prey groups classified to the lowest taxonomic level was only completed on presence/absence data. The ordination plot for presence/absence data of prey groups at the lowest taxonomic level identified separation between control and mussel farm sites (Figure 6). PERMANOVA identified significant differences in the composition of snapper gut contents for Treatment (pseudoF(1, 57) = 13.58, p = 0.001) and Treatment × Location (pseudoF(1, 57) = 3.46, p = 0.001), but not for Location (pseudoF(1, 57) = 2.11, p = 0.05). Post hoc analyses indicated that the main prey groups at the lowest taxonomic level driving the differences were triangle barnacle, green‐lipped mussel and other Decapoda (Table 4). Triangle barnacles and green‐lipped mussels occurred more frequently in Motukopake Island mussel farm compared to the Motukopake Island control site (Table 4). Motukopake Island mussel farm also had a higher frequency of green‐lipped mussels present in snapper gut contents compared to Rat Island control site. Similarly, the Rat Island mussel farm had a higher frequency of green‐lipped mussels present in snapper gut contents compared to Rat Island control site and Motukopake Island control site. The Rat Island control site had a higher frequency of other Decapoda compared to the mussel farm site (Table 4).6FIGURENon‐metric multidimensional scaling (NMDS) plot of presence/absence data for prey groups classified to the lowest taxonomic level within snapper gut contents at control and mussel farm sites (see Table 3). Confidence interval (50%) ellipses displayed over Treatments.4TABLEFishers exact test post hoc significant differences in the presence/absence of prey groups classified to the lowest taxonomic level within snapper gut contents between the sites Motukopake Island mussel farm, Motukopake Island control site, Rat Island mussel farm and Rat Island control sitePrey group at lowest taxonomic levelComparisonPercentAdjusted p‐valueTriangle barnacle (Balanus trigonus)Motukopake Island control site < Motukopake Island mussel farm0% vs. 68%0.002Green‐lipped mussel (Perna canaliculus)Motukopake Island control site < Motukopake Island mussel farm0% vs. 56%0.02Rat Island control site < Rat Island mussel farm0% vs. 56%0.04Motukopake Island control site < Rat Island mussel farm0% vs. 56%0.03Rat Island control site < Motukopake Island mussel farm0% vs. 56%0.04Decapoda spp.Rat Island mussel farm < Rat Island control site0% vs. 62%0.01Gut content composition in genetic analysisAt Motukopake Island, there were 20 families identified at the four sites (Figure 7). The greatest proportion of species at both Treatments was within the Malacostraca (Crustacea) class (Figure 7). Key differences were the presence of a high proportion of Mytilidae (mussels) and Caprellidae (Amphipoda) in mussel farm sites and none in the control site (Figure 7). Sesarmidae (Brachyura) had the highest proportion in control sites with none present at mussel farm sites (Figure 7).7FIGUREPresence/absence plot for DNA sequences identified in snapper gut contents at Motukopake Island mussel farm (n = 16) and Motukopake Island control site (n = 16), and Rat Island mussel farm (n = 16) and Rat Island control site (n = 13) categorised by family level (left‐hand side) and class level (right‐hand side). The size of the circles indicates the number of individuals where the prey group was identified, and the colour is the relative read abundance (RRA) which is a semi‐quantitative metric to understand dominance of the prey group within the snapper sample.At Rat Island, there were 15 families identified at the control site and 29 families identified at the mussel farm (Figure 7). The greatest proportion of families at the control site was within the Actinopteri (ray‐finned fish) and Malacostraca (Crustacea) class, and in the mussel farm it was Malacostraca (Figure 7). Mytilidae (mussels) and Skeletonemaceae (microalgae) were most present in the mussel farm, whereas Semilidae (clams) and Lithodidae (king crabs) were represented in greater number of samples in the Rat Island control site (Figure 7). The most dominant prey group (based on relative read abundance) in the mussel farm was the Dotillidae (crabs) and Moridae (cod fish) at the control site (Figure 7).PERMANOVA identified significant differences in the presence/absence of prey at the family level in the gut contents of snapper for Treatment (pseudoF(1, 44) = 5.66, p = 0.001), Location (pseudoF(1, 44) = 2.07, p = 0.007) and Treatment × Location (pseudoF(1, 44) = 2.12, p = 0.009). However, p‐adjusted values in post hoc analyses could not detect any significant differences among the four sites. Non‐adjusted p‐values showed some differences in families between Treatments. For example, the higher presence of Mytilidae in snapper gut contents sampled from Motukopake Island mussel farm (54%) and Rat Island mussel farms (55%) was significantly greater than the control sites (0%) (p = 0.002 and p = 0.01, respectively). Sesarmidae had a higher presence in snapper gut contents from Motukopake Island control site (50%) compared to Motukopake Island mussel farm and Rat Island mussel farm (0%) (p = 0.006 and p = 0.04 respectively). Motukopake Island mussel farm had a higher presence of Balanidae (barnacles) (46%) compared to the Motukopake Island control site and Rat Island control site (0%) (p = 0.006 and p = 0.02, respectively). Also, Caprellidae (Amphipoda) were present in snapper gut contents sampled from Motukopake Island mussel farm (38%) but absent at the Motukopake Island and Rat Island control sites (0%) (p = 0.02 and p = 0.04, respectively).PERMANOVA identified significant differences in the presence/absence of prey items at the class level in the gut contents for Treatment (pseudoF(1, 48) = 3.74, p = 0.007) and Treatment × Location (pseudoF(1, 48) = 2.48, p = 0.04), but not Location (pseudoF(1, 48) = 1.13, p = 0.34). However, p‐adjusted values in post hoc analyses could not detect any significant differences among sites. Non‐adjusted p‐values showed some differences in the presence/absence of the various classes of prey items in the gut contents between Treatment and Location. For example, snapper gut contents sampled from Motukopake Island mussel farm and Rat Island mussel farm had a greater presence of Bivalvia (including mussels) (47% and 46%) compared to Motukopake Island control site (0%) (p = 0.007 and p = 0.01, respectively). Motukopake Island mussel farm also had a greater presence of Thecostraca (e.g. barnacles) compared to both control sites (40% vs. 0%, p = 0.02). Snapper gut contents from the Motukopake Island control site had a greater presence of Conoidasida (parasitic alveolates) (35%) compared to Motukopake Island mussel farm (0%) (p = 0.03). Rat Island control site had a greater presence of Chromadorea (round worms) (36%) compared to Rat Island mussel farm (0%) and Motukopake Island control site (0%) (p = 0.03).DISCUSSIONThis study compared the difference in snapper gut contents between mussel farms and control sites to see whether prey groups identified in mussel farm snapper were sourced directly or indirectly from the mussel farm infrastructure. The results indicate that the snapper sampled from mussel farm habitat are consuming a distinctly different diet compared to those sampled from control sites. The differences in diet are due to the snapper utilising prey species made available through the presence of the mussel farm habitat, such as harvested and common biofouling species within the farm habitat, which are not available in nearby unstructured soft‐sediment habitats (i.e. the control sites).Differences in snapper gut contentsMussel farm versus control site snapperThe key differences in the gut contents of snapper between mussel farm and control sites were the significant contribution of mussels (green‐lipped and blue) and barnacles in mussel farms and not in control sites. Genetic analyses also confirmed that there were differences in the presence of caprellid amphipods which were present in mussel farms but not control sites, but these soft bodied Amphipoda were not picked up in visual analyses. There were some differences with Brachyura families between mussel farm and control sites, such as for Sesarmidae which were more commonly present in snapper gut contents at both control sites. The crab family Sesarmidae is not a common Brachyura family in New Zealand and therefore likely represents a distinct Brachyura species not yet available in the DNA reference databases (Wilkens & Ahyong, 2015). The other key difference in control sites compared to mussel farms was the presence of Paguridae (hermit crabs). Paguridae are abundant across a range of soft‐sediment and rocky reef locations, and have shown an association with mussel farm habitats overseas and in soft‐sediment mussel reefs in New Zealand (D'Amours et al., 2008; McLeod et al., 2014; Sean et al., 2022). Therefore, it is unclear why the abundance of Paguridae was higher in control sites and not mussel farms. It is possible that the abundance of shells which Paguridae rely on for host shells was more abundant in the control site, as Paguridae prefer host shells from Gastropoda species rather than Bivalvia, and Bivalvia shells (from mussels) are most abundant on the seafloor underneath mussel farms (McLaughlin & Thiel, 2015; Wilding & Nickell, 2013).The snapper consumption of/predation on cultured shellfish species (i.e. farmed green‐lipped mussels) is consistent with the general global analyses of the diet of other farm‐dwelling Sparidae species (Gerlotto et al., 2001; Hayden, 1995; Peteiro et al., 2010; Šegvić‐Bubić et al., 2011; Tsuyuki & Umino, 2017). However, most of these studies focused on the spat stage rather than mature mussels on a grow‐out cycle (Gerlotto et al., 2001; Hayden, 1995; Peteiro et al., 2010; Stenton‐Dozey & Broekhuizen, 2019; Tsuyuki & Umino, 2017). Sparid species are known to be significant predators of farmed shellfish species, especially mussels. For example, the diet of gilthead sea bream (Sparus aurata) sampled from a Croatian mussel farm consisted of 70% cultured mussels, which was considerably higher than for the snapper sampled in the mussel farm sites in the current study, with up to 19% consumption of cultured green‐lipped mussels. Snapper in the current study were also captured on the seafloor and so are likely consuming mussels which have fallen off the dropper lines as well as those present on the dropper lines. Additionally, the Croatian study showed no consumption of Crustacea by gilthead sea bream in mussel farms, which is surprising given that Crustacea are a dominant food resource for Sparidae and are commonly associated with mussel farms (Callier et al., 2018; McKindsey et al., 2011; Šegvić‐Bubić et al., 2011; Usmar, 2012; Woods et al., 2012). Nonetheless, gilthead sea bream did consume Teleostei and more Gastropoda compared to the current study, suggesting that the site‐specific differences in biofouling or the differences in diet between the two Sparidae species have impacted their respective diet compositions (Šegvić‐Bubić et al., 2011).Snapper in mussel farms also consumed blue mussels (up to 10.1%) and barnacles (up to 14%), which are key nuisance biofouling species in New Zealand green‐lipped mussel farms (Forrest & Atalah, 2017; Woods et al., 2012; Zazzaro et al., 2018). Since 2009, there has been a dramatic increase in the amount of fouling of the triangle barnacle (B. trigonus) across mussel farms in the Hauraki Gulf (Zazzaro et al., 2018). Globally, it is estimated that aquaculture operations spend 5%–10% of total profits managing biofouling, and so if snapper are selectively predating on these nuisance fouling organisms, this could be of benefit to mussel farm operations (Zazzaro et al., 2018). Consequently, there is a potential trade‐off between snapper consuming harvested green‐lipped mussels, but also having some benefit from reducing nuisance biofouling within the farm. Given that the overall consumption of harvested mussels by snapper was a lot lower than previously documented for another sparid species (19% vs. 70%), the predation may not be significant enough to impact mussel production at the intermediate to final grow out stages.Caprellids found in mussel farm snapper gut contents were likely present within the mussels on dropper lines as these are common biofouling species that associate with biogenic structure, for example mussels, hydroids and tunicates (Lim & Harley, 2018; South et al., 2019; Woods et al., 2012, 2014). Caprellids are particularly abundant on dropper lines during seed out of juvenile mussels (South et al., 2019). The presence of caprellids, mussels and barnacles within mussel farm snapper gut contents suggests that snapper are feeding on the dropper lines as well as the benthos, as these species are much more likely to be present on the structural substrate of the dropper lines (South et al., 2019; Woods et al., 2012; Zazzaro et al., 2018). Research that investigated the abundance and diversity of snapper within Coromandel mussel farms showed that snapper were observed at the surface 34% of the time and at the benthos 66% of the time, based on 89.6 h of footage (Underwood, 2023). These observations show that although snapper were present on the benthos for the majority of the time, they still visited the dropper lines presumably for feeding opportunities, as also confirmed by the composition of prey items in snapper gut contents.There were other key indicator species that were present in mussel farm snapper but not control sites which confirmed that mussel farm snapper are consuming biofouling within the farm environment. These included Notomithrax and Halicarcinus Brachyura species which are both common inhabitants of mussel dropper lines (Woods et al., 2012). The pea crab (Nepinnotheres novaezelandiae) was also present in mussel farm snapper which is typically a parasitic crab only within blue and green‐lipped mussels (Trottier & Jeffs, 2015a, 2015b; Trottier et al., 2012; Wilkens & Ahyong, 2015). Additionally, some Bivalvia which were present in control site snapper and not mussel farms, such as Dosina spp., cockle (Austrovenus stutchburyi) and pipi (Paphies australis), are more commonly found within soft, muddy and fine sand benthic habitats with no direct association with mussel farm benthic habitat where organic loading from the shellfish farming is typically higher (Powell, 1979; Wilding & Nickell, 2013).Comparison to broader Hauraki Gulf snapper populationPrevious analyses on snapper gut contents within the Hauraki Gulf have covered broader habitats and snapper size ranges which has resulted in a wide range of species and taxonomic groups identified in gut contents (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). Snapper have been observed to consume nearly 100 different species across a range of soft‐sediment sampling stations in the Hauraki Gulf, which is much higher than the 34 species identified in the visual gut analysis in the current study (Colman, 1972; Godfriaux, 1970). For example, higher proportions of Polychaeta (5%–10% vs. 0%–2%), Teleostei (5%–10% vs. 0%–4%) and Gastropoda (5%–15% vs. 0%–1.6%) have been previously identified in the gut contents of snapper living in natural habitats in the Hauraki Gulf (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). Additionally, Echinodermata and Porifera were present in broader diet studies, with Echinodermata making up a large proportion of gut contents (9.6%–28.3%) (Colman, 1972; Godfriaux, 1970; Russell, 1983). These comparisons have confirmed the plasticity of snapper diet even within different habitats found within one marine region.Overall, there is general consensus that mostly free‐living Crustacea are the dominant prey group targeted by snapper, with all previous diet studies recording this taxa as the dominant component (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). This is generally consistent with the current study; however, the dominance of Bivalvia within mussel farm snapper gut contents exceeds the proportion of Crustacea at the Rat Island mussel farm (32% Bivalvia vs. 18.4% Crustacea), and is high but does not exceed the proportion of Crustacea within Motukopake Island mussel farm (18% Bivalvia and 26% Crustacea). Both control sites had a dominance of Crustacea—29% at Rat Island control site and 21.4% at Motukopake Island control site. Of the Crustacea, Brachyura was found to be the dominant component in four previous studies of snapper diet in New Zealand (Colman, 1972; Godfriaux, 1970; Russell, 1983; Usmar, 2012); however, Paguroidea and Caridea were more dominant in two studies that focused on a smaller size range of adult snapper of 25–55 cm (Drummond, 2020; Third, 2022). Snapper typically start consuming Brachyura from 10 cm (fork length) and their diet becomes more varied from 25 cm to include Bivalvia, Paguroidea, Polychaeta, Gastropoda and Teleostei (Usmar, 2012). This is mostly due to the jaw crushing strength of the snapper which develops with age/increase in size, allowing snapper to successfully consume more hard‐shelled organisms (Usmar, 2012). In the current study (which also used a refined size class of 26–42 cm), the dominant Crustacea component within mussel farm snapper were barnacles, mostly triangle barnacle (B. trigonus). This is an important finding as no other previous snapper study within the Hauraki Gulf has identified barnacles as a diet component (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). For control sites, Paguroidea were dominant in Motukopake Island, aligning with the recent soft‐sediment snapper studies (Drummond, 2020; Third, 2022). Decapoda spp. were dominant in Rat Island which could be from any of the Brachyura, Caridea and Paguridae families.Of the Brachyura identified by our genetic and visual gut analysis methods, only four species had been previously recorded. The dwarf swimming crab (Liocarcinus corrugatus), identified in control sites, was present in two previous snapper gut studies and dominant in one (Drummond, 2020; Third, 2022). Other Brachyura found in mussel farm snapper were from the Majidae (decorator crab) family, including Notomithrax species which were previously found in most Hauraki Gulf snapper gut studies (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). Similarly, Halicarcinus spp. were present in mussel farm snapper and are commonly consumed by snapper sampled from natural habitats in the Hauraki Gulf (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022; Usmar, 2012). Even though these crab species are present in natural soft‐sediment habitats, there is likely to be an increased abundance of them within the mussel farm habitat (Wilkens & Ahyong, 2015; Woods et al., 2012).The overall proportion of Bivalvia within the gut contents of mussel farm snapper largely exceeded the proportions in previous studies, for example 32%–18% in the current study versus 8%–13% (Godfriaux, 1970; Third, 2022). The Bivalvia species previously identified in the gut contents of snapper include Dosina spp., Anomia spp., A. stutchburyi, Atrina zelandica, Tawera spissa, Soletellina nitida, Purpurocardia purpurata, Neilo australis, Gari stangeri and Linucula hartvigiana (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). These Bivalvia species previously observed in the gut contents of snapper sampled from natural habitats have four species in common with the control sites over soft‐sediment natural habitat in this study, and only two in mussel farm sites. Interestingly, Motukopake Island control site snapper also had a high proportion of Bivalvia (excluding mussels) in gut contents (16%) exceeding previous studies (Godfriaux, 1970; Third, 2022). This is likely due to the availability of Bivalvia in the benthic habitat at this specific site, as snapper sampled from Rat Island only included 0.15% Bivalvia in their gut contents. Green‐lipped mussels have been rarely reported in the gut contents of snapper sampled from natural habitats (Third, 2022; Usmar, 2012), but snapper are known to predate on them when they are available (Alder et al., 2021, 2022a, 2022b). Blue mussels have never been recorded in previous snapper diets in New Zealand (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). However, the detection of the consumption of mussels relies on the presence of this species in sufficient quantities within the habitat.Comparison of visual and genetic gut analysis resultsOverall, there was good alignment between the visual and genetic gut analyses for the key species, such as mussels and barnacles. Additionally, many of the Brachyura families identified by the molecular genetic analysis precisely matched species of Brachyura identified by the visual gut analysis, such as Majidae (Notomithrax spp.), Hymenosomatidae (Halicarcinus spp.), Pilumnidae (e.g. Pilumnus novaezelandiae) and Pinnotheridae (Nepinnotheres spp.). The genetic analysis was also able to identify a broader range of taxa, including soft body and small organisms such as Ascidiacea (tunicates), Holothuroidea (sea cucumbers), Polychaeta and Hoplonemertea (worms) and amphipods (small crustaceans). Identifying these types of soft‐bodied organisms to an informative level is often not possible using visual gut analysis due to their rapid digestion once consumed, masking their morphological structure needed for identification (O'Rorke et al., 2012; van der Reis et al., 2020). However, there were insufficient numbers of individual snapper for which DNA was detected from these prey groups to result in a significant difference between snapper from control versus mussel farm sites. Although amplification was achieved for the majority of samples, the DNA extracted from the gut contents was degraded (visible when running a 0.8% agarose DNA gel), likely hindering better amplification, and may be a limitation to identifying the full scope of taxa within the gut content in this study (O'Rorke et al., 2012; van der Reis et al., 2020). Although appropriate protocols were followed, the logistic constraints of field sampling and multiple sampling of gut contents (i.e. visual gut analysis and sampling for biochemical analysis) meant that gut contents could not be immediately preserved with chemicals (i.e. ethanol) upon capture of the snapper which would have been optimum for molecular genetic analyses (van der Reis & Lavery, 2020).Taxonomic assignment of DNA sequences is reliant on a match that reaches the required minimum accuracy threshold that is typically established on the basis of the availability of sequences in the reference databases for the suspected prey types (van der Reis & Lavery, 2020). In New Zealand, there is a lack of references for Crustacea species in the DNA databases (particularly noticeable for crabs) (van der Reis et al., 2020). Therefore, a BLAST 85% identity match was chosen to be able to assign closest matches for these species (van der Reis et al., 2020). The lack of New Zealand Crustacea families sequenced has resulted in some crab misidentifications (five Malacostraca families) and thus few common coastal species were identified (Forest & McLay, 2001; Wilkens & Ahyong, 2015). Regardless, distinct differences were found for crab families identified in one treatment and not the other.Implications for Hauraki gulf snapper populationOverall, the results show that snapper within mussel farms are consuming a unique diet, with prey groups linked to the mussel farm habitat, either underneath mussel farms and/or on dropper line habitats. In particular, the key prey groups which snapper were consuming in mussel farm habitats (mussels and barnacles) are highly abundant within mussel farms (Forrest & Atalah, 2017; Skelton et al., 2022; Woods et al., 2012; Zazzaro et al., 2018). Firstly, the green‐lipped mussel industry produces approximately 1.78 billion adult mussels per year, with 30% of this production within the Coromandel region, which sums to approximately 530 million adults per year (Skelton et al., 2022). Secondly, blue mussels and barnacles are key nuisance biofouling species in Coromandel mussel farms as these can foul surfaces at high densities across mussel dropper line habitats (Forrest & Atalah, 2017; Woods et al., 2012; Zazzaro et al., 2018). Abundant prey groups can especially be targeted by snapper because of their omnivorous and highly diverse diets (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). Snapper have been observed at much higher densities within mussel farms compared to adjacent control sites (i.e. soft‐sediment without mussel farms), with snapper up to five times more abundant underneath mussel farms (Underwood, 2023). Therefore, the higher abundances of snapper within mussel farms are likely to be partially due to the availability of prey given the direct link to prey species present in mussel farm habitats. A stable prey supply is important as it reduces the energy partitioned for foraging, so that there is more energy available for other biological processes such as growth and reproduction (Wootton, 2012). Furthermore, the prey groups targeted in mussel farms are energy rich, with high lipid content present in mussels and barnacles (Barclay et al., 2006; Barnes & Achituv, 1976). Consequently, snapper are likely receiving a nutritional benefit from feeding within mussel farm habitats. The highly abundant mussel farm habitats within the Coromandel have the potential to support the Hauraki Gulf snapper population through the supply of food resources. This is a direct outcome that can be used towards restorative aquaculture frameworks, to utilise farm habitats in a way that provides net positive ecological outcomes for coastal environments (Alleway et al., 2019; Froehlich et al., 2017; Gentry et al., 2020; Theuerkauf et al., 2019).CONCLUSIONSThe results from this study show that snapper living within mussel farms in the Hauraki Gulf are consuming a distinctly different diet to snapper in adjacent natural habitats without a mussel farm. The diet that snapper have consumed within mussel farms can be directly linked to the cultured species (i.e. green‐lipped mussels) and the biofouling (i.e. blue mussels and barnacles). There was good alignment between the visual gut and genetic analyses for the key species identified. The DNA degradation within individual snapper samples has likely impacted the lack of additional differences identified. Overall, the results show that the highly abundant prey groups consumed by snapper in mussel farm habitats are likely to be beneficial to the snapper population, reducing foraging effort and potentially supplying more nutritious prey. These findings provide evidence towards the supporting services of mussel farm habitats through the provision of food resources. This evidence can support the restorative aquaculture objectives for the region, which aims to provide net positive ecological outcomes for wild fish populations.AUTHOR CONTRIBUTIONSLucy H. Underwood: Conceptualization; formal analysis; investigation; methodology; writing—original draft. Aimee van der Reis: Conceptualization; methodology; writing—review and editing. Andrew G. Jeffs: Funding acquisition; project administration; supervision; writing—review and editing.ACKNOWLEDGEMENTSWe thank Gold Ridge Marine farms for providing access to the mussel farm sites and for logistic support. We also thank the Coromandel Marine Farmers’ Association (CMFA) and their members for their support. Thanks to Esther Stuck, Brad Skelton, David DuPavillon, Bruce Underwood and Joe Dennehy for fieldwork support. Thanks to Ash Heaphy for laboratory support. We are grateful to the Nature Conservancy for financial support for this research and to the Ministry for Primary Industries’ Sustainable Food and Fibre Futures Fund (#22074).Open access publishing facilitated by The University of Auckland, as part of the Wiley ‐ The University of Auckland agreement via the Council of Australian University Librarians.ETHICS STATEMENTThis research was conducted in accordance with animal ethics approval under New Zealand's Animal Welfare Act 1999 (UoA‐AEC Approval # 21619), which ensures the humane treatment of animals for research.CONFLICT OF INTEREST STATEMENTThe authors declare no conflicts of interest.DATA AVAILABILITY STATEMENTThe raw sequence data is available on NCBI's Sequence Read Archive (http://www.ncbi.nlm.nih.gov/bioproject/971684) with associated metadata. The underlying data for statistical analyses will be shared on reasonable request to the corresponding author.PEER REVIEWThe peer review history for this article is available at: https://publons.com/publon/10.1002/aff2.113REFERENCESAlder, A., Jeffs, A.G. & Hillman, J.R. (2022a) The importance of stock selection for improving transplantation efficiency. Restoration Ecology, 30(4), e13561. Available from: https://doi.org/10.1111/rec.13561Alder, A., Jeffs, A. & Hillman, J. (2022b) Timing mussel deployments to improve reintroduction success and restoration efficiency. Marine Ecology Progress Series, 698, 69–83. Available from: https://doi.org/10.3354/meps14157Alder, A., Jeffs, A. & Hillman, J.R. (2021) Considering the use of subadult and juvenile mussels for mussel reef restoration. Restoration Ecology, 29(3), e13322. https://doi.org/10.1111/rec.13322Alleway, H.K., Gillies, C.L., Bishop, M.J., Gentry, R.R., Theuerkauf, S.J. & Jones, R. (2019) The ecosystem services of marine aquaculture: valuing benefits to people and nature. Bioscience, 69(1), 59–68. https://doi.org/10.1093/biosci/biy137Amundsen, P. & Sánchez‐Hernández, J. (2019) Feeding studies take guts – critical review and recommendations of methods for stomach contents analysis in fish. Journal of Fish Biology, 95(6), 1364–1373. https://doi.org/10.1111/jfb.14151Anyango, J.O., Mlewa, C.M. & Mwaluma, J. (2017) Abundance, diversity and trophic status of wild fish around seaweed farms in Kibuyuni, South Coast Kenya. International Journal of Fisheries and Aquatic Studies, 5(3), 440–446.Baker, R., Buckland, A. & Sheaves, M. (2014) Fish gut content analysis: robust measures of diet composition. Fish and Fisheries, 15(1), 170–177. https://doi.org/10.1111/faf.12026Barclay, M.C., Irvin, S.J., Williams, K.C. & Smith, D.M. (2006) Comparison of diets for the tropical spiny lobster Panulirus ornatus: astaxanthin‐supplemented feeds and mussel flesh. Aquaculture Nutrition, 12(2), 117–125. https://doi.org/10.1111/j.1365‐2095.2006.00390.xBarnes, H. & Achituv, Y. (1976) The utilization of various biochemical entities in gonadally mature Balanus balanoides (L.) under starvation and feeding in the absence of copulation. Journal of Experimental Marine Biology and Ecology, 22(3), 257–262.Barnett, A., Redd, K.S., Frusher, S.D., Stevens, J.D. & Semmens, J.M. (2010) Non‐lethal method to obtain stomach samples from a large marine predator and the use of DNA analysis to improve dietary information. Journal of Experimental Marine Biology and Ecology, 393(1), 188–192. https://doi.org/10.1016/j.jembe.2010.07.022Barrett, L.T., Theuerkauf, S.J., Rose, J.M., Alleway, H.K., Bricker, S.B., Parker, M., et al. (2022) Sustainable growth of non‐fed aquaculture can generate valuable ecosystem benefits. Ecosystem Services, 53, 101396. https://doi.org/10.1016/j.ecoser.2021.101396Berry, O., Bulman, C., Bunce, M., Coghlan, M., Murray, D.C. & Ward, R.D. (2015) Comparison of morphological and DNA metabarcoding analyses of diets in exploited marine fishes. Marine Ecology Progress Series, 540, 167–181. https://doi.org/10.3354/meps11524Benson, D.A., Cavanaugh, M., Clark, K., Karsch‐Mizrachi, I., Lipman, D.J., Ostell, J. & Sayers, E.W. (2013) GenBank. Nucleic Acids Res, 41, D36–42. https://doi.org/10.1093/nar/gks1195Binning, S.A. & Chapman, L.J. (2010) Is intraspecific variation in diet and morphology related to environmental gradients? Exploring Liem's paradox in a cichlid fish. Integrative Zoology, 5(3), 241–255. https://doi.org/10.1111/j.1749‐4877.2010.00209.xBoecklen, W.J., Yarnes, C.T., Cook, B.A. & James, A.C. (2011) On the use of stable isotopes in trophic ecology. Annual Review of Ecology, Evolution, and Systematics, 42(1), 411–440. https://doi.org/10.1146/annurev‐ecolsys‐102209‐144726Bolyen, E., Rideout, J.R., Dillon, M.R., Abnet, C.C., Al‐Ghalith, G.A., Alexander, H., et al. (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature Biotechnology, 37(9), 1091. https://doi.org/10.1038/s41587‐019‐0252‐6Braga, R., Bornatowski, H. & Vitule, J. (2012) Feeding ecology of fishes: an overview of worldwide publications. Reviews in Fish Biology and Fisheries, 22(4), 915–929. https://doi.org/10.1007/s11160‐012‐9273‐7Budge, S.M., Iverson, S.J. & Koopman, H.N. (2006) Studying trophic ecology in marine ecosystems using fatty acids: a primer on analysis and interpretation. Marine Mammal Science, 22(4), 759–801. https://doi.org/10.1111/j.1748‐7692.2006.00079.xCallahan, B.J., McMurdie, P.J. & Holmes, S.P. (2017) Exact sequence variants should replace operational taxonomic units in marker‐gene data analysis. The ISME Journal, 11(12), 2639–2643. https://doi.org/10.1038/ismej.2017.119Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.A. & Holmes, S.P. (2016) DADA2: high‐resolution sample inference from Illumina amplicon data. Nature Methods, 13(7), 581–583. https://doi.org/10.1038/NMETH.3869Callier, M.D., Byron, C.J., Bengtson, D.A., Cranford, P.J., Cross, S.F., Focken, U., et al.(2018) Attraction and repulsion of mobile wild organisms to finfish and shellfish aquaculture: a review. Reviews in Aquaculture, 10(4), 924–949. https://doi.org/10.1111/raq.12208Castro, A.L.C., de Farias Diniz, A., Martins, I.Z., Vendel, A.L., de Oliveira, T.P.R. & de Lucena Rosa, I.M. (2008) Assessing diet composition of seahorses in the wild using a non‐destructive method: Hippocampus reidi (Teleostei: Syngnathidae) as a study‐case. Neotropical Ichthyology, 6(4), 637–644. https://doi.org/10.1590/S1679‐62252008000400012Clynick, B.G., McKindsey, C.W. & Archambault, P. (2008) Distribution and productivity of fish and macroinvertebrates in mussel aquaculture sites in the Magdalen islands (Québec, Canada). Aquaculture, 283(1), 203–210. https://doi.org/10.1016/j.aquaculture.2008.06.009Cole, A. (2010) Cleaning to corallivory: ontogenetic shifts in feeding ecology of tubelip wrasse. Coral Reefs, 29(1), 125–129. https://doi.org/10.1007/s00338‐009‐0563‐zColman, J.A. (1972) Food of snapper, Chrysophrys auratus (Forster), in the Hauraki Gulf, New Zealand. New Zealand Journal of Marine and Freshwater Research, 6(3), 221–239. https://doi.org/10.1080/00288330.1972.9515422D'Amours, O., Archambault, P., McKindsey, C. & Johnson, L. (2008) Local enhancement of epibenthic macrofauna by aquaculture activities. Marine Ecology Progress Series, 371, 73–84. https://doi.org/10.3354/meps07672Deagle, B.E., Thomas, A.C., McInnes, J.C., Clarke, L.J., Vesterinen, E.J., Clare, E.L., Kartzinel, T.R. & Eveson, J.P. (2019) Counting with DNA in metabarcoding studies: how should we convert sequence reads to dietary data? Molecular Ecology, 28(2), 391–406. https://10.1111/mec.14734Drummond, L. (2020) How does diet affect snapper (Chrysophrys auratus) movement? (Masters dissertation). Auckland: University of Auckland.Fernandez‐Jover, D. & Sanchez‐Jerez, P. (2015) Comparison of diet and otolith growth of juvenile wild fish communities at fish farms and natural habitats. ICES Journal of Marine Science, 72(3), 916–929. https://doi.org/10.1093/icesjms/fsu153Fernandez‐Jover, D., Sanchez‐Jerez, P., Bayle‐Sempere, J.T., Arechavala‐Lopez, P., Martinez‐Rubio, L., Jimenez, J.A.L. & Lopez, F.J.M. (2009) Coastal fish farms are settlement sites for juvenile fish. Marine Environmental Research, 68(2), 89–96. https://doi.org/10.1016/j.marenvres.2009.04.006Forest, J. & McLay, C.L. (2001) The biogeography and bathymetric distribution of New Zealand hermit crabs (Crustacea: Anomura: Paguridea). Journal of the Royal Society of New Zealand, 31(4), 687–720. https://doi.org/10.1080/03014223.2001.9517670Forrest, B.M. & Atalah, J. (2017) Significant impact from blue mussel Mytilus galloprovincialis biofouling on aquaculture production of green‐lipped mussels in New Zealand. Aquaculture Environment Interactions, 9, 115–126. https://doi.org/10.3354/aei00220Froehlich, H.E., Gentry, R.R. & Halpern, B.S. (2017) Conservation aquaculture: shifting the narrative and paradigm of aquaculture's role in resource management. Biological Conservation, 215, 162–168. https://doi.org/10.1016/j.biocon.2017.09.012Geller, J., Meyer, C., Parker, M. & Hawk, H. (2013) Redesign of PCR primers for mitochondrial cytochrome c oxidase subunit I for marine invertebrates and application in all‐taxa biotic surveys. Molecular Ecology Resources, 13(5), 851–861. https://doi.org/10.1111/1755‐0998.12138Gentry, R.R., Alleway, H.K., Bishop, M.J., Gillies, C.L., Waters, T. & Jones, R. (2020) Exploring the potential for marine aquaculture to contribute to ecosystem services. Reviews in Aquaculture, 12(2), 499–512. https://doi.org/10.1111/raq.12328Gerlotto, F., Brehmer, P., Buestel, D. & Sanguinede, F. (2001) A method for acoustic monitoring of a mussel longline ground using vertical echosounder and multibeam sonar. In: ICES Annual Science Conference, 26–29 September 2001, Oslo.Gibbs, M.T. (2004) Interactions between bivalve shellfish farms and fishery resources. Aquaculture, 240(1), 267–296. https://doi.org/10.1016/j.aquaculture.2004.06.038Godfriaux, B.L. (1970) Food of predatory demersal fish in Hauraki Gulf. New Zealand Journal of Marine and Freshwater Research, 4(4), 325–336. https://doi.org/10.1080/00288330.1970.9515351Hayden, B.J. (1995) Factors affecting recruitment of farmed greenshell mussels, Perna canaliculus (Gmelin) 1791, in Marlborough Sounds (Doctoral dissertation). Christchurch, New Zealand: University of Canterbury. Available from: http://hdl.handle.net/10523/9115.Hehre, E.J. & Meeuwig, J.J. (2016) A global analysis of the relationship between farmed seaweed production and herbivorous fish catch. PLoS ONE, 11(2), e0148250. https://doi.org/10.1371/journal.pone.0148250Horton, T., Kroh, A., Ahyong, S., Bailly, N., Bieler, R., Boyko, C.B., et al. (2022) World Register of Marine Species (WoRMS). Ostend, Belgium: World Register of Marine Species (WoRMS).Illumina. (2013) 6S metagenomic sequencing library preparation: preparing 16S ribosomal RNA gene amplicons for the Illumina MiSeq system. Report no. 15044223 Rev. B. Available from: https://support.illumina.com/documents/documentation/chemistry_documentation/16s/16s‐metagenomic‐library‐prep‐guide‐15044223‐b.pdfKamler, J.F. & Pope, K.L. (2001) Nonlethal methods of examining fish stomach contents. Reviews in Fisheries Science, 9(1), 1–11. https://doi.org/10.1080/20016491101663Leray, M., Yang, J.Y., Meyer, C.P., Mills, S.C., Agudelo, N., Ranwez, V., et al. (2013) A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: application for characterizing coral reef fish gut contents. Frontiers in Zoology, 10(1), 34. https://doi.org/10.1186/1742‐9994‐10‐34Lim, E.G. & Harley, C.D.G. (2018) Caprellid amphipods (Caprella spp.) are vulnerable to both physiological and habitat‐mediated effects of ocean acidification. PeerJ, 6, e5327. https://doi.org/10.7717/peerj.5327Lobo, J., Costa, P.M., Teixeira, M.A., Ferreira, M.S., Costa, M.H. & Costa, F.O. (2013) Enhanced primers for amplification of DNA barcodes from a broad range of marine metazoans. BMC Ecology, 13(1), 34. https://doi.org/10.1186/1472‐6785‐13‐34Mahesh, V., Gop, A. & Nair, R.J. (2018) Stomach content analysis techniques in fishes. In: Nair, R. J., Gop, A. P. & Mahesh, V. (Eds.) Recent advances in fishery biology techniques for biodiversity evaluation and conservation. Kochi, India: Central Marine Fisheries Research Institute, pp. 104–115.Martin, M. (2011) Cutadapt removes adapter sequences from high‐throughput sequencing reads. EMBnet.Journal, 17(1), 10–12. https://doi.org/10.14806/EJ.17.1.200McKindsey, C.W., Archambault, P., Callier, M.D. & Olivier, F. (2011) Influence of suspended and off‐bottom mussel culture on the sea bottom and benthic habitats: a review. Canadian Journal of Zoology, 89(7), 622–646. https://doi.org/10.1139/Z11‐037McLaughlin, P.A. & Thiel, M. (2015) Crustaceans in mobile homes. The Natural History of the Crustacea, 2, 145–162.McLeod, I., Parsons, D., Morrison, M., Van Dijken, S. & Taylor, R. (2014) Mussel reefs on soft sediments: a severely reduced but important habitat for macroinvertebrates and fishes in New Zealand. New Zealand Journal of Marine and Freshwater Research, 48(1), 48–59. https://doi.org/10.1080/00288330.2013.834831Morton, J., Platell, M. & Gladstone, W. (2008) Differences in feeding ecology among three co‐occurring species of wrasse (Teleostei: Labridae) on rocky reefs of temperate Australia. Marine Biology, 154(3), 577–592. https://doi.org/10.1007/s00227‐008‐0951‐xO'Rorke, R., Lavery, S. & Jeffs, A. (2012) PCR enrichment techniques to identify the diet of predators. Molecular Ecology Resources, 12(1), 5–17. https://doi.org/10.1111/j.1755‐0998.2011.03091.xPeteiro, L.G., Filgueira, R., Labarta, U. & Fernández‐Reiriz, M.J. (2010) The role of fish predation on recruitment of Mytilus galloprovincialis on different artificial mussel collectors. Aquacultural Engineering, 42(1), 25–30. https://doi.org/10.1016/j.aquaeng.2009.09.003Powell, A.W.B. (1979) New Zealand mollusca: marine, land, and freshwater shells. Auckland: Collins.R Core Team. (2021) R: a language and environment for statistical computing. Version 4.1.0; R studio version 1.4.1106. Vienna: R Foundation for Statistical Computing.Rennó Braga, R., Ribeiro, V., Bornatowski, H., Abilhoa, V. & Vitule, S.J. (2017) Gastric lavage for dietary studies of small fishes: efficiency, survival and applicability. Acta Ichthyologica Et Piscatoria, 47(1), 97–100. https://doi.org/10.3750/AIEP/02079Russell, B.C. (1983) Food and feeding habits of rocky reef fish of north‐eastern New Zealand. New Zealand Journal of Marine and Freshwater Research, 17(2), 121–145. https://doi.org/10.1080/00288330.1983.9515991Sanchez‐Jerez, P., Fernandez‐Jover, D., Uglem, I., Arechavala‐Lopez, P., Dempster, T., Bayle‐Sempere, J.T., et al. (2011) Coastal fish farms as fish aggregation devices (FADs). In: Bortone, S. A., Pereira Brandini, F., Fabi, G. & Otake, S. (Eds.) Artificial reefs in fishery management. Boca Raton, FL: CRC Press Taylor & Francis Group, pp. 187–208.Sean, A., Drouin, A., Archambault, P. & McKindsey, C.W. (2022) Influence of an offshore mussel aquaculture site on the distribution of epibenthic macrofauna in Îles de la Madeleine, Eastern Canada. Frontiers in Marine Science, 9, 859816. https://doi.org/10.3389/fmars.2022.859816Šegvić‐Bubić, T., Grubišić, L., Karaman, N., Tičina, V., Jelavić, K.M. & Katavić, I. (2011) Damages on mussel farms potentially caused by fish predation—Self‐service on the ropes? Aquaculture, 319(3), 497–504. https://doi.org/10.1016/j.aquaculture.2011.07.031Skelton, B.M., South, P.M. & Jeffs, A.G. (2022) Inefficiency of conversion of seed into market‐ready mussels in New Zealand's Greenshell™ mussel (Perna canaliculus) industry. Aquaculture, 560, 738584. https://doi.org/10.1016/j.aquaculture.2022.738584Skog, T., Hylland, K., Torstensen, B.E. & Berntssen, M.H.G. (2003) Salmon farming affects the fatty acid composition and taste of wild saithe Pollachius virens. Aquaculture Research, 34(12), 999–1007. https://doi.org/10.1046/j.1365‐2109.2003.00901.xSouth, P.M., Floerl, O. & Jeffs, A.G. (2019) The role of biofouling development in the loss of seed mussels in aquaculture. Biofouling, 35(2), 259–272. https://doi.org/10.1080/08927014.2019.1596261Stenton‐Dozey, J. & Broekhuizen, N. (2019) Provision of ecological and ecosystem services by mussel farming in the Marlborough Sounds: a literature review in context of the state of the environment pre and post‐mussel farming. Prepared for Marine Farming Association, National Institute of Water & Atmospheric Research. Available from: https://www.marinefarming.co.nz/site_files/24792/upload_files/Fullreport_28.07.2021update.pdf?dl=1Supono, S., Dunphy, B. & Jeffs, A. (2021) Role of nutritional history in the attachment and retention of green‐lipped mussel spat. Aquaculture, 533, 736094. https://doi.org/10.1016/j.aquaculture.2020.736094Theuerkauf, S.J., Barrett, L.T., Alleway, H.K., Costa‐Pierce, B.A., St Gelais, A. & Jones, R.C. (2022) Habitat value of bivalve shellfish and seaweed aquaculture for fish and invertebrates: pathways, synthesis and next steps. Reviews in Aquaculture, 14(1), 54–72. https://doi.org/10.1111/raq.12584Theuerkauf, S.J., Morris, J.A., Waters, T.J., Wickliffe, L.C., Alleway, H.K. & Jones, R.C. (2019) A global spatial analysis reveals where marine aquaculture can benefit nature and people. PLoS ONE, 14(10), e0222282. https://doi.org/10.1371/journal.pone.0222282Third, G. (2022) Morphological and diet variation in Chrysophrys auratus populations (Masters dissertation). Auckland: University of Auckland.Trkov, D. & Lipej, L. (2019) A non‐destructive method for assessing the feeding habits of coastal fish. Mediterranean Marine Science, 20(2), 453–459. https://doi.org/10.12681/mms.20234Trottier, O. & Jeffs, A.G. (2015a) Recruitment of the parasitic pea crab Nepinnotheres novaezelandiae into green‐lipped mussels Perna canaliculus. Diseases of Aquatic Organisms, 112(3), 199–205. https://doi.org/10.3354/dao02809Trottier, O. & Jeffs, A.G. (2015b) Mate locating and access behaviour of the parasitic pea crab, Nepinnotheres novaezelandiae, an important parasite of the mussel Perna canaliculus. Parasite, 22, 13. https://doi.org/10.1051/parasite/2015013Trottier, O., Walker, D. & Jeffs, A.G. (2012) Impact of the parasitic pea crab Pinnotheres novaezelandiae on aquacultured New Zealand green‐lipped mussels, Perna canaliculus. Aquaculture, 344–349, 23–28. https://doi.org/10.1016/j.aquaculture.2012.02.031Tsuyuki, A. & Umino, T. (2017) Spatial movement of black sea bream Acanthopagrus schlegelii around the oyster farming area in Hiroshima Bay, Japan. Fisheries Science, 83(2), 235–244. https://doi.org/10.1007/s12562‐016‐1058‐9Udy, J., Wing, S., O'Connell‐Milne, S., Kolodzey, S., McMullin, R., Durante, L. et al. (2019) Organic matter derived from kelp supports a large proportion of biomass in temperate rocky reef fish communities: implications for ecosystem‐based management. Aquatic Conservation, 29(9), 1503–1519. https://doi.org/10.1002/aqc.3101Underwood, L.H. & Jeffs, A.G. (2023) Settlement and recruitment of fish in mussel farms. Aquaculture Environment Interactions, 15, 85–100. https://doi.org/10.3354/aei00454Underwood, L.H. (2023) Habitat value of green‐lipped mussel (Perna canaliculus) farms for local fish populations in northern New Zealand (Doctoral dissertation). Auckland: University of Auckland.Usmar, N. (2012) Ontogenetic diet shifts in snapper (Pagrus auratus: Sparidae) within a New Zealand estuary. New Zealand Journal of Marine and Freshwater Research, 46(1), 31–46. https://doi.org/10.1080/00288330.2011.587824van der Reis, A.L & Lavery, S.D. (2020) Marine DNA metabarcoding. In: Goldstein, M. I. & Della Salla, D. A. (Eds.) Encyclopedia of the world's biomes, vol. 4. Amsterdam: Elsevier, pp. 612–618.van der Reis, A.L., Beckley, L.E., Olivar, M.P. & Jeffs, A.G. (2023) Nanopore short‐read sequencing: a quick, cost‐effective and accurate method for DNA metabarcoding. Environmental DNA, 5(2), 282–296. https://doi.org/10.1002/edn3.374van der Reis, A., Jeffs, A. & Lavery, S. (2020) From feeding habits to food webs: exploring the diet of an opportunistic benthic generalist. Marine Ecology Progress Series, 655, 107–121. https://doi.org/10.3354/meps13511Wilding, T.A. & Nickell, T.D. (2013) Changes in benthos associated with mussel (Mytilus edulis L.) farms on the west‐coast of Scotland. PLoS ONE, 8(7), e68313. https://doi.org/10.1371/journal.pone.0068313Wilkens, S.L. & Ahyong, S.T. (2015) Coastal crabs: a guide to the crabs of New Zealand (Version 1). Wellington, New Zealand: The National Institute of Water and Atmospheric Research (NIWA).Woods, C., Floerl, O. & Hayden, B. (2012) Biofouling on Greenshell™ mussel (Perna canaliculus) farms: a preliminary assessment and potential implications for sustainable aquaculture practices. Aquaculture International, 20(3), 537–557. https://doi.org/10.1007/s10499‐011‐9484‐2Woods, C., Williams, R. & Heasman, K. (2014) First record of the caprellid amphipod Caprella andreae Mayer, 1890 (Crustacea, Amphipoda, Caprellidae) from New Zealand. BioInvasions Records, 3(2), 97–102.Wootton, R.J. (2012) Ecology of teleost fishes. Dordrecht: Springer. https://doi.org/10.1007/978‐94‐009‐0829‐1Zazzaro, D., Ruggiero, K. & Jeffs, A. (2018) Use of extract from adults of the triangle barnacle, Balanus trigonus, for reducing fouling in mussel farms. Aquaculture, 483, 223–229. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Aquaculture Fish and Fisheries Wiley http://www.deepdyve.com/lp/wiley/diet-of-snapper-chrysophrys-auratus-in-green-lipped-mussel-farms-and-CTqWx11Kb2

Loading next page...

References (63)

J. Udy, S. Wing, S. O’Connell-Milne, S. Kolodzey, R. McMullin, L. Durante, Russell Frew (2019)
Organic matter derived from kelp supports a large proportion of biomass in temperate rocky reef fish communities: Implications for ecosystem‐based management
Aquatic Conservation: Marine and Freshwater Ecosystems
A. Alder, A. Jeffs, J. Hillman (2022)
Timing mussel deployments to improve reintroduction success and restoration efficiency
Marine Ecology Progress Series
Jacques Forest, C. McLAY (2001)
The biogeography and bathymetric distribution of New Zealand hermit crabs (Crustacea: Anomura: Paguridea)
Journal of the Royal Society of New Zealand, 31
B. Deagle, Austen Thomas, Julie McInnes, L. Clarke, E. Vesterinen, E. Clare, Tyler Kartzinel, J. Eveson (2018)
Counting with DNA in metabarcoding studies: How should we convert sequence reads to dietary data?
Molecular Ecology, 28
A. Alder, A. Jeffs, Jenny Hillman (2020)
Considering the use of subadult and juvenile mussels for mussel reef restoration
Restoration Ecology, 29
Myriam Callier, C. Byron, D. Bengtson, P. Cranford, S. Cross, U. Focken, H. Jansen, P. Kamermans, A. Kiessling, T. Landry, F. O’Beirn, E. Petersson, R. Rheault, Ø. Strand, K. Sundell, T. Svåsand, G. Wikfors, C. McKindsey (2018)
Attraction and repulsion of mobile wild organisms to finfish and shellfish aquaculture: a review
Reviews in Aquaculture, 10
D. Fernández-Jover, P. Sanchez‐Jerez (2015)
Comparison of diet and otolith growth of juvenile wild fish communities at fish farms and natural habitats
Ices Journal of Marine Science, 72
O. Trottier, A. Jeffs (2015)
Mate locating and access behaviour of the parasitic pea crab, Nepinnotheres novaezelandiae, an important parasite of the mussel Perna canaliculus
Parasite, 22
D. Fernández-Jover, P. Sanchez‐Jerez, J. Bayle-Sempere, P. Arechavala‐Lopez, L. Martínez-Rubio, J. Jiménez, F. Lopez (2009)
Coastal fish farms are settlement sites for juvenile fish.
Marine environmental research, 68 2
O. Trottier, Dion Walker, A. Jeffs (2012)
Impact of the parasitic pea crab Pinnotheres novaezelandiae on aquacultured New Zealand green-lipped mussels, Perna canaliculus
Aquaculture, 344
R. Braga, H. Bornatowski, J. Vitule (2012)
Feeding ecology of fishes: an overview of worldwide publications
, 22
H. Barnes, Y. Achituv (1976)
The utilization of various biochemical entities in gonadally mature Balanus balanoides (L.) under starvaton and feeding in the absence of copulation
Journal of Experimental Marine Biology and Ecology, 22
J. Kamler, K. Pope (2001)
Nonlethal Methods of Examining Fish Stomach Contents
Reviews in Fisheries Science, 9
C. Woods, Rissa Williams, K. Heasman (2014)
First record of the caprellid amphipod Caprella andreae Mayer, 1890 (Crustacea, Amphipoda, Caprellidae) from New Zealand
BioInvasions Records, 3
E. Bolyen, J.R. Rideout, M.R. Dillon, C.C. Abnet, G.A. Al‐Ghalith, H. Alexander (2019)
Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2
, 37
B. Clynick, C. McKindsey, P. Archambault (2008)
Distribution and productivity of fish and macroinvertebrates in mussel aquaculture sites in the Magdalen islands (Québec, Canada)
Aquaculture, 283
D. Trkov, L. Lipej (2019)
A non-destructive method for assessing the feeding habits of coastal fish
Mediterranean Marine Science
I. Mcleod, D. Parsons, M. Morrison, S. Dijken, Rb Taylor (2014)
Mussel reefs on soft sediments: a severely reduced but important habitat for macroinvertebrates and fishes in New Zealand
New Zealand Journal of Marine and Freshwater Research, 48
Seth Theuerkauf, J. Morris, Tiffany Waters, L. Wickliffe, Heidi Alleway, Robert Jones (2019)
A global spatial analysis reveals where marine aquaculture can benefit nature and people
PLoS ONE, 14
James Anyango, C. Mlewa, J. Mwaluma (2017)
Abundance, diversity and trophic status of wild fish around seaweed farms in Kibuyuni, South Coast Kenya
International Journal of Fisheries and Aquatic Studies, 5
S. Supono, B. Dunphy, A. Jeffs (2020)
Role of nutritional history in the attachment and retention of green-lipped mussel spat
Aquaculture
(2015)
Crustaceans in mobile homes
W.J. Boecklen, C.T. Yarnes, B.A. Cook, A.C. James (2011)
On the use of stable isotopes in trophic ecology
, 42
Alice Reis, A. Jeffs, S. Lavery (2020)
From feeding habits to food webs: exploring the diet of an opportunistic benthic generalist
Marine Ecology Progress Series, 655
Paul South, O. Floerl, A. Jeffs (2019)
The role of biofouling development in the loss of seed mussels in aquaculture
Biofouling, 35
N. Usmar (2012)
Ontogenetic diet shifts in snapper (Pagrus auratus: Sparidae) within a New Zealand estuary
New Zealand Journal of Marine and Freshwater Research, 46
B. Russell (1983)
The food and feeding habits of rocky reef fish of north-eastern New Zealand
New Zealand Journal of Marine and Freshwater Research, 17
LH Underwood, A. Jeffs (2023)
Settlement and recruitment of fish in mussel farms
Aquaculture Environment Interactions
Anne-Sara Sean, A. Drouin, P. Archambault, C. McKindsey (2022)
Influence of an Offshore Mussel Aquaculture Site on the Distribution of Epibenthic Macrofauna in Îles de la Madeleine, Eastern Canada
, 9
P. Amundsen, J. Sánchez‐Hernández (2019)
Feeding studies take guts - critical review and recommendations of methods for stomach contents analysis in fish.
Journal of fish biology
(2012)
GenBank
M.C. Barclay, S.J. Irvin, K.C. Williams, D.M. Smith (2006)
Comparison of diets for the tropical spiny lobster Panulirus ornatus: astaxanthin‐supplemented feeds and mussel flesh
, 12
B.J. Callahan, P.J. McMurdie, M.J. Rosen, A.W. Han, A.J.A. Johnson, S.P. Holmes (2016)
DADA2: high‐resolution sample inference from Illumina amplicon data
, 13
L. Peteiro, R. Filgueira, U. Labarta, M. Fernández-Reiriz (2010)
The role of fish predation on recruitment of Mytilus galloprovincialis on different artificial mussel collectors
Aquacultural Engineering, 42
O. Berry, C. Bulman, M. Bunce, M. Coghlan, Dáithí Murray, R. Ward (2015)
Comparison of morphological and DNA metabarcoding analyses of diets in exploited marine fishes
Marine Ecology Progress Series, 540
O. Trottier, A. Jeffs (2015)
Recruitment of the parasitic pea crab Nepinnotheres novaezelandiae into green-lipped mussels Perna canaliculus.
Diseases of aquatic organisms, 112 3
O. D'amours, P. Archambault, C. McKindsey, L. Johnson (2008)
Local enhancement of epibenthic macrofauna by aquaculture activities
Marine Ecology Progress Series, 371
T. Šegvić-Bubić, L. Grubišić, N. Karaman, V. Tičina, Krstina Jelavić, I. Katavić (2011)
Damages on mussel farms potentially caused by fish predation—Self service on the ropes?
Aquaculture, 319
E. Hehre, J. Meeuwig (2016)
A Global Analysis of the Relationship between Farmed Seaweed Production and Herbivorous Fish Catch
PLoS ONE, 11
R. Baker, A. Buckland, M. Sheaves (2014)
Fish gut content analysis: robust measures of diet composition
Fish and Fisheries, 15
H. Froehlich, R. Gentry, B. Halpern (2017)
Conservation aquaculture: Shifting the narrative and paradigm of aquaculture's role in resource management
Biological Conservation, 215
L. Drummond (2020)
How does diet affect snapper (Chrysophrys auratus) movement?
Seth Theuerkauf, L. Barrett, Heidi Alleway, B. Costa‐Pierce, Adam Gelais, Robert Jones (2021)
Habitat value of bivalve shellfish and seaweed aquaculture for fish and invertebrates: Pathways, synthesis and next steps
Reviews in Aquaculture
Aimee Reis, L. Beckley, M. Olivar, A. Jeffs (2022)
Nanopore short‐read sequencing: A quick, cost‐effective and accurate method for DNA metabarcoding
Environmental DNA
S.A. Binning, L.J. Chapman (2010)
Is intraspecific variation in diet and morphology related to environmental gradients? Exploring Liem's paradox in a cichlid fish
, 5
A. Cole (2010)
Cleaning to corallivory: ontogenetic shifts in feeding ecology of tubelip wrasse
, 29
B. Callahan, Paul McMurdie, S. Holmes (2017)
Exact sequence variants should replace operational taxonomic units in marker-gene data analysis
The ISME Journal, 11
A.L.C. Castro, A. Farias Diniz, I.Z. Martins, A.L. Vendel, T.P.R. Oliveira, I.M. Lucena Rosa (2008)
Assessing diet composition of seahorses in the wild using a non‐destructive method: Hippocampus reidi (Teleostei: Syngnathidae) as a study‐case
, 6
M. Gibbs (2004)
Interactions between bivalve shellfish farms and fishery resources
Aquaculture, 240
S.M. Budge, S.J. Iverson, H.N. Koopman (2006)
Studying trophic ecology in marine ecosystems using fatty acids: a primer on analysis and interpretation
, 22
J. Geller, C. Meyer, M. Parker, H. Hawk (2013)
Redesign of PCR primers for mitochondrial cytochrome c oxidase subunit I for marine invertebrates and application in all‐taxa biotic surveys
, 13
J. Colman (1972)
Food of snapper, Chrysophrys auratus (forster), in the Hauraki Gulf, New Zealand
New Zealand Journal of Marine and Freshwater Research, 6
Heidi Alleway, C. Gillies, M. Bishop, R. Gentry, Seth Theuerkauf, Robert Jones (2018)
The Ecosystem Services of Marine Aquaculture: Valuing Benefits to People and Nature
BioScience, 69
Bradley Skelton, Paul South, A. Jeffs (2022)
Inefficiency of conversion of seed into market-ready mussels in New Zealand's Greenshell™ mussel (Perna canaliculus) industry
Aquaculture
A. Alder, A. Jeffs, Jenny Hillman (2021)
The importance of stock selection for improving transplantation efficiency
Restoration Ecology, 30
A. Barnett, K. Redd, S. Frusher, J. Stevens, J. Semmens (2010)
Non-lethal method to obtain stomach samples from a large marine predator and the use of DNA analysis to improve dietary information
Journal of Experimental Marine Biology and Ecology, 393
B. Godfriaux (1970)
Food of predatory demersal fish in Hauraki Gulf: 3: Feeding relationships∗
New Zealand Journal of Marine and Freshwater Research, 4
Emily Lim, C. Harley (2018)
Caprellid amphipods (Caprella spp.) are vulnerable to both physiological and habitat-mediated effects of ocean acidification
PeerJ, 6
L. Barrett, Seth Theuerkauf, J. Rose, Heidi Alleway, S. Bricker, M. Parker, D. Petrolia, Robert Jones (2022)
Sustainable growth of non-fed aquaculture can generate valuable ecosystem benefits
Ecosystem Services
Davide Zazzaro, K. Ruggiero, A. Jeffs (2018)
Use of extract from adults of the triangle barnacle, Balanus trigonus, for reducing fouling in mussel farms
Aquaculture, 483
T. Wilding, Thom Nickell (2013)
Changes in Benthos Associated with Mussel (Mytilus edulis L.) Farms on the West-Coast of Scotland
PLoS ONE, 8
R. Gentry, Heidi Alleway, M. Bishop, C. Gillies, Tiffany Waters, Robert Jones (2020)
Exploring the potential for marine aquaculture to contribute to ecosystem services
Reviews in Aquaculture
Barrie Forrest, J. Atalah (2017)
Significant impact from blue mussel Mytilus galloprovincialis biofouling on aquaculture production of green-lipped mussels in New Zealand
Aquaculture Environment Interactions, 9

Publisher: Wiley
Copyright: © 2023 The Authors. Aquaculture, Fish and Fisheries published by John Wiley & Sons Ltd.
eISSN: 2693-8847
DOI: 10.1002/aff2.113
Publisher site: See Article on Publisher Site

Abstract

INTRODUCTIONCoastal aquaculture structures operate as artificial habitats for fish through the provision of shelter and food resources (Alleway et al., 2019; Callier et al., 2018; Theuerkauf et al., 2022). Aquaculture habitats are known to attract and aggregate some fish species (e.g. Siganidae species in seaweed farms [Hehre & Meeuwig, 2016] and Sparidae species in mussel farms [Šegvić‐Bubić et al., 2011]), enabling them to reach greater abundance within farms compared to sites without aquaculture structures (Barrett et al., 2022). Some fish species are also known to settle and recruit directly into non‐fed aquaculture structures, utilising the structure as a nursery habitat (Underwood & Jeffs, 2023). It is often assumed that the aggregation of fish at aquaculture structures is a response to the input of feed, the abundance of the farmed species or the associated assemblage of potential prey items due to the natural colonisation of aquaculture structures (e.g. biofouling) (Barrett et al., 2022; Callier et al., 2018; Sanchez‐Jerez et al., 2011; Theuerkauf et al., 2022). Research in this area has largely focused on differences in the abundance of the functional groups of fish living within or in surrounding aquaculture habitats, to infer responses to the food resources which may be available (Clynick et al., 2008; Theuerkauf et al., 2022). However, few studies have directly analysed the composition of the diet of fish within aquaculture habitats, and of these most have focused on fed finfish aquaculture (Anyango et al., 2017; Fernandez‐Jover & Sanchez‐Jerez, 2015; Fernandez‐Jover et al., 2009; Hayden, 1995; Hehre & Meeuwig, 2016; Šegvić‐Bubić et al., 2011; Skog et al., 2003).In non‐fed aquaculture practices (e.g. shellfish), studies have focused on the predatory pressures that wild fish can have during the early stages of production, particularly predation of mussel spat (Peteiro et al., 2010; Šegvić‐Bubić et al., 2011; Stenton‐Dozey & Broekhuizen, 2019). For example, members of the Sparidae family (seabreams and snapper) are commonly sighted in large schools stripping recently seeded mussel lines (Gerlotto et al., 2001; Peteiro et al., 2010; Stenton‐Dozey & Broekhuizen, 2019). These fish are generalist predators and tend to feed based on prey availability rather than selectively (Šegvić‐Bubić et al., 2011; Usmar, 2012). The gut contents of seabream sampled from mussel (Mytilus galloprovincialis) farms identified mussels as the dominant prey (69.7% biomass), although a diverse range of other prey items (e.g. gastropods, fishes and macroalgae) were also found (Šegvić‐Bubić et al., 2011). This provides evidence of what a member of the Sparidae family may feed on within mussel farms, and highlights the concurrent and/or selective epibiota consumption.Methods to analyse fish diet have traditionally relied on gut dissections and morphology as a precise method to determine the possible food items that are sourced from local habitats, and to compare the differences among habitats (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Braga et al., 2012). However, several other methods exist including visual observations, stable isotope analysis, fatty acid analysis and DNA metabarcoding (Braga et al., 2012; Castro et al., 2008; Udy et al., 2019). Visual observations of the feeding behaviour of fish usually provide high‐level information on patterns of feeding but little detail on the relative components of their diet (Amundsen & Sánchez‐Hernández, 2019; Braga et al., 2012; Cole, 2010; Morton et al., 2008). Non‐lethal methods are less common and have only been trialled on few fish species (Barnett et al., 2010; Castro et al., 2008; Kamler & Pope, 2001; Rennó Braga et al., 2017; Trkov & Lipej, 2019). Non‐lethal methods are species specific, are invasive and require technical extraction experience. However, they can still result in mortality and generally are not suitable for most studies of fish diet. Stable isotope and fatty acid analyses are better suited to inferring broad‐scale trophic linkages and changes to diet over time rather than identifying specific dietary components (Boecklen et al., 2011; Budge et al., 2006; Udy et al., 2019). DNA metabarcoding is suited to the latter and can identify digested gut content items previously ‘unidentifiable’ due to the lack of morphological structure (Berry et al., 2015).A range of methods can be used to sort and quantify gut contents, including numerical frequency of prey items, gravimetric methods and volumetric composition (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Mahesh et al., 2018). Although there have been attempts to standardise the quantification metrics among visual gut content investigations, suitable metrics vary based on the objectives of the study, and each quantification metric has biases making some more suitable metrics for certain species than others. For example, numerical frequency methods are most suitable for fish that consume similar sized prey items to reduce bias towards smaller numerically dominant species, which may have limited volumetric importance (Amundsen & Sánchez‐Hernández, 2019). The relative gut fullness method is a modified volumetric method that allows for each prey grouping to be proportionately allocated, while also standardising the entire contents by the fullness of the gut. Therefore, it is an effective approach when samples have varied gut fullness, and for accommodating a range of prey types. DNA‐based methods (e.g. DNA metabarcoding) can increase taxonomic accuracy by allowing for the identification of the presence of prey items when their remains are visually unidentifiable in gut contents (O'Rorke et al., 2012; van der Reis et al., 2020). However, the results are semi‐quantitative (Deagle et al., 2019), and species‐level identification is reliant on the availability of an adequate taxonomic DNA reference library (Amundsen & Sánchez‐Hernández, 2019; van der Reis & Lavery, 2020). Therefore, the combination of visual gut analysis and DNA‐based methods is considered an effective approach for analysing gut contents.The overall aim of this research was to compare the diet of snapper (Chrysophrys auratus), a common coastal demersal fish species, within and outside established green‐lipped mussel longline farms (Perna canaliculus) in the Coromandel Harbour, Firth of Thames, New Zealand. Snapper are commonly present as adults and tend to become resident within coastal mussel aquaculture sites in New Zealand (Gibbs, 2004; Stenton‐Dozey & Broekhuizen, 2019). These fish are generalist predators that have marked ontogenetic shifts in their diets that are regulated by age and the habitat within which they reside (Usmar, 2012). Therefore, it could be expected that their diet would reflect their use of the surrounding habitat. This research aimed to test whether the diet of snapper collected from mussel farms would differ from those collected in adjacent areas, with the differences likely relating to prey items sourced directly or indirectly from the mussel aquaculture infrastructure.METHODSSite locationsSnapper were sampled from four sites in the Firth of Thames. Two were long established longline green‐lipped mussel farms and two were sites without mussel farms (i.e. control sites), all located over similar unstructured soft sediment habitat of similar depth (8–13 m) (Figure 1). The farm sites were Motukopake Island (36° 45′ 2.87″ S, 175° 25′ 22.8″ E), and Rat Island (36° 45′ 25.9194″ S, 175° 26′ 59.99″ E). The control sites were at least 500 m away from the farms and were Motukopake Island control site (36° 57′ 32.03″ S, 175° 28′ 51.59″ E) and Rat Island control site (36° 45′ 23.03″ S, 175° 27′ 39.6″ E; 36° 45′ 46.07″ S, 175° 26′ 49.2″ E). Two locations, with the same habitat features, were collectively used for the Rat Island control site due to the low numbers of fish caught at the first control site. The mussel aquaculture operations comprised a series of paired parallel backbone lines held near the surface by large plastic floats that support suspended loops of dropper ropes covered with attached mussels. The dropper ropes extended 6–8 m below the surface floats and were held several metres above the soft‐sediment seafloor.1FIGUREThe sampling sites within the Coromandel Harbour, Firth of Thames where snapper were captured using hook and line fishing methods from the seafloor within mussel farms and from nearby control sites at both Motukopake Island and Rat Island.Sample collection and processingAdult snapper only within a size range of 26–42 cm fork length were sampled in May and June 2022. Hook and line fishing methods were used, with plastic soft‐bait and lures utilised as much as possible to avoid possible contamination of gut contents by bait. Where natural bait was used, a minimal amount of bait was deployed of a readily recognisable species that were not common in the area (i.e. pilchard, squid or mullet). Crepuscular periods of snapper feeding activity were primarily targeted for the sampling to increase the probability of capturing snapper with gut contents. Sixteen fish were obtained from each of three sampling sites (Motukopake Island mussel farm, Motukopake Island control site and Rat Island mussel farm), while 13 snapper were sampled from Rat Island control site. Immediately upon capture, all fish were humanely euthanised according to animal ethics approval (NZ Animal Welfare Act 1999, UoA‐AEC Approval # 21619), labelled and put into salted ice slurry. Each snapper was measured (fork length) and weighed back on land (2–6 h after capture) before being frozen for subsequent gut analyses. Previous studies have indicated that freezing snapper gut contents was an appropriate method of preservation that facilitates both visual, molecular genetic and biochemical analysis of gut contents to be undertaken (Drummond, 2020; Supono et al., 2021; Third, 2022).Snapper gut dissectionFrozen snapper samples were thawed at room temperature and then dissected. The alimentary tract was removed via incisions at the oesophageal opening and at the anus. The foregut and hindgut were separated and each opened. A gut fullness score was estimated for both the foregut and the hindgut, with 0 being completely empty and 10 being completely full (Third, 2022). Subsequently, the gut fullness score was averaged between the hindgut and the foregut, so that there was only one score per snapper gut. Similarly, a digestion score was calculated to provide an estimate of the extent of digestion for the prey items for the hindgut and the foregut individually, with a digestion score of 0 being no digestion and 5 being fully digested (Third, 2022). Subsequently, the digestion score was averaged between the hindgut and the foregut, so that there was only one score per snapper gut. Any bait identified in the foregut was removed, and not considered further in this study other than for exclusion from any DNA signal detected. The remaining material in the foregut was weighed, followed by the foregut lining after the gut contents had been emptied. The process was repeated for the hindgut.Visual analysis of snapper gut contentsGut contents for the foregut and hindgut were combined and then spread onto a sterile tissue culture dish (2 × 2 cm grid) and sorted into groups of similar prey items, for each sample. Distilled water was used to separate and clean contents. Each prey group was classified to the lowest practical taxonomic level. An adapted version of the relative‐fullness method was used to quantify the proportions of each prey grouping in each individual snapper gut (Amundsen & Sánchez‐Hernández, 2019; Baker et al., 2014; Binning & Chapman, 2010). The relative‐fullness method then uses proportions of each prey group and standardises by the fullness of the gut to calculate ‘points’ for each individual prey group within the snapper gut contents. This was undertaken by firstly estimating the two‐dimensional (2D) coverage of each prey group in the culture dish (e.g. Bivalvia covering two squares). All prey groups were then summed together to estimate the total coverage within each snapper gut, including the unidentifiable digested material which had its own category. This estimate was then used to calculate the proportion of each prey group within the gut of the sampled snapper. Proportions were multiplied by the average fullness estimate (foregut and hindgut calculated independently but averaged to produce one fullness score) to calculate the relative proportion of each prey group within each individual snapper, thereby standardising the relative proportion of each prey group to facilitate comparisons among samples. For example, if the proportion of the gut content of a snapper was 0.8 (i.e. 80% coverage of the gut contents in the culture dish) for green‐lipped mussel (P. canaliculus) and the average fullness estimate was 3, this would be calculated as 0.8 × 3 which equals a relative proportion of 2.4 points.Genetic analysis of snapper gut contentGut content removal and DNA extractionOnce gut contents for individual fish had been extracted and analysed for visual gut analysis, the gut contents were mixed and a 2‐mL subsample was preserved in 90% ethanol (2 mL cryovial) and stored at −20°C. All dissection tools were sterilised between handling individual snapper.DNA was extracted from the subsample using the Nucleospin Tissue DNA extraction kit (Macherey‐Nagel) following the manufacturer's instructions. Controls were included in the extractions, that is DNA extraction blank where no content was added. The DNA quality and quantity were viewed on 0.8% agarose, visualised with gel red (Biotium) using a Gel Doc XR+ (Bio‐Rad).DNA amplificationSpecies identified microscopically from the gut contents were compared to sequences available on GenBank. The results indicated that cytochrome oxidase one (COI) would have the best coverage, and thus COI was targeted for amplification during the polymerase chain reactions (PCRs). The universal primer pair selected was M1COIintF (forward; Leray et al., 2013) and jgHCO2198 (reverse; modified replacing ‘I’ with ‘N’; Geller et al., 2013). Illumina Nextra adapters were added to the primer sequences. The PCRs were done in triplicate using 6.25 μL of MyTaq Red Mix (Bioline; Meridian Bioscience), 0.25 μL of each primer (10 μM), 1 μL of bovine serum albumin (1%) and 4.75 Ultra‐Pure DNase/RNase free water (Invitrogen; Thermo Fisher Scientific). Each PCR cycle included the DNA blanks and a PCR blanks (no DNA added) to check for possible contamination. The PCR protocol was a modified version of Lobo et al. (2013): 94°C for 60 s, 35 × [94°C for 30 s; 54°C for 90 s; 72°C for 60 s] and a final extension at 72°C for 5 min (van der Reis et al., 2023). Amplification was checked by running the PCR products on a 1.6% agarose, and visualised as outlined above.Sequencing and bioinformaticsPCR triplicates were pooled per sample and then cleaned using AMPure XP (Beckman Coulter), following the Illumina 16S protocol (Illumina, 2013). The concentration of the cleaned PCR products was determined using Qubit dsDNA HS Assay Kit (Invitrogen) following the manufacturer's instructions and subsequently diluted to 2 ng/μL before sequencing. Sequencing was done by Auckland Genomics and undertaken on an Illumina MiSeq (2 × 250 paired‐end nano run; single lane used).Cutadapt v3.5 (Martin, 2011) was used for primer removal, retaining only those sequences with exact primer matches anchored at the beginning of the sequence. Qiime2 v2022.2 (Bolyen et al., 2019) was used to assess the sequence quality prior to truncation in DADA2 (Callahan et al., 2016; utilised within Qiime2). High‐quality amplicon sequence variants (ASVs; Callahan et al., 2017) were produced by filtering for sequence quality, retaining only sequences that merged and that were identified non‐chimeric.ASVs were run through GenBank's BLAST database v2022‐07 using the MegaBLAST function within the BLASTn suite (Benson et al., 2013; minimum E‐value threshold of 0.001 and a minimum percentage identity of ≥70%). Resulting taxonomic assignments were then run through the World Register of Marine Species (WoRMS) database (Horton et al., 2022) to confirm taxonomy and filter to retain only marine species (i.e. remove any spurious terrestrial sequences). R v4.0.4 (R Core Team, 2021) was used for final sequence filtering and quality assurance of the taxonomic assignments. Potential contamination was accounted for by a proportional subtraction of the controls (DNA and PCR blanks were sequenced). Sequences for snapper (C. auratus) (host species) were removed. Only sequences >280 base pairs that had ≥85% identity match to the taxonomic sequence were used. Sequences were also filtered to retain only those with ≥90% query coverage.Statistical analysesAll statistical analyses and plots were produced using R (R Core Team, 2021). All analyses used a significance of α = 0.05. All means are presented as mean ± standard error. Plots were prepared with ggplot (package version 4.0.4, ggplot2 function).Snapper metricsTo compare differences in snapper length, a two‐way analysis of variance (package version 4.0.4, aov function) was used with the main factors of Location (Motukopake Island or Rat Island) and Treatment (mussel farm or control). Data were plotted and visually assessed to confirm parametric assumptions were met. Tukey's honest significant difference (package version 4.0.4, TukeyHSD function) post hoc tests were conducted if an overall significant difference was identified. This approach was repeated to compare the differences in the wet weight (g) for each of foregut and hindgut contents separately. Only foregut wet weight data needed to be transformed prior to analyses by using a log(x + 1) transformation. A generalised‐linear model (package version 4.0.4, glm function) was used to compare the differences in the combined average digestion score for the hindgut and the foregut, with a quasi‐Poisson distribution and the main factors of Location and Treatment. If an overall significant difference was identified, ‘emmeans’ (package version 1.5.5‐1, emmeans function) post hoc analyses were used.Visual gut analysisTo compare the amount of prey groups at a broad taxonomic level in snapper gut contents among samples, non‐metric multidimensional scaling (NMDS) was conducted with a Bray–Curtis distance and a permutation value of 999, using points data derived from the relative‐fullness method. An ordination plot was produced for a 2D representation of the ordination, separated by Treatment. A permutational multivariate analysis of variance (PERMANOVA) was undertaken using points data of prey groups at the broad taxonomic level, with the Bray–Curtis distance and the main factors of Location and Treatment. Where the overall model results were significant, a Wilcoxon test was used to determine differences in the amount (quantified as ‘points’ from the relative‐fullness method) of prey groups at the broad taxonomic level consumed by snapper among sites. A Bonferroni adjustment was used for p‐values to account for potential error inflation from multiple comparisons.For prey groups at the lowest taxonomic level, only presence/absence data were used for each prey group identified within each individual snapper. NMDS was conducted with a Bray–Curtis distance and a permutation value of 999, using presence/absence data of the different prey groups at the lowest taxonomic level. An ordination plot was produced for a 2D representation of the ordination, separated by Treatment. A PERMANOVA was undertaken for prey groups at the lowest taxonomic level using presence/absence data, with the Bray–Curtis distance and the main factors of Location and Treatment. Where the overall model results were significant, a Fisher's exact test was used to determine whether the frequency of each prey group at the lowest taxonomic level found within snapper gut contents was significantly different between Treatment and Location. A Bonferroni adjustment was used for p‐values to account for potential error inflation from multiple comparisons.DNA metabarcoding analysisThe presence/absence data of prey groups identified from DNA metabarcoding at the family level and class level for gut content of all individual snapper were analysed separately. A PERMANOVA was undertaken on the presence/absence data for detected prey separately for the family level and at the class level with the Bray–Curtis distance using the main factors of Location and Treatment. If significance was detected, a Fisher's exact test was undertaken using presence/absence data of the prey groups. Family‐level comparisons among the samples of gut content excluded 13 snapper samples where no families were identified (Motukopake Island control site = 2, Motukopake Island mussel farm = 3, Rat Island control site = 3 and Rat Island mussel farm = 5). Class‐level comparisons excluded nine snapper samples where zero taxonomic classes of prey were identified (Motukopake Island control site = 3, Motukopake Island mussel farm = 1, Rat Island control site = 2 and Rat Island mussel farm = 3).RESULTSSnapper metricsSnapper length ranged from 27.4 to 41.2 cm at Rat Island mussel farm, 27.0 to 41.6 cm at Motukopake Island mussel farm, 25.8 to 30.2 cm at Rat Island control site and 26.8 to 37.4 cm at Motukopake Island control site (Figure 2). There were differences in mean snapper length for the sampled snapper for Treatment, that is mussel farm versus control site (F(1, 57) = 5.41, p = 0.024) and the interaction Treatment × Location (F(1, 57) = 9.51, p = 0.0032), but not for Location, that is Motukopake Island versus Rat Island (F(1, 57) = 1.04, p = 0.31) (Figure 2). Post hoc analyses identified that snapper sampled at the Motukopake Island control site were larger than at the Rat Island control site (p = 0.024; 31.6 ± 1.0 and 27.8 ± 0.3 cm, respectively), while snapper from Rat Island mussel farm were larger than those sampled from the Rat Island control site (p = 0.002; 32.8 ± 1.0 and 27.8 ± 0.3 cm, respectively).2FIGURESize frequency of snapper (Chrysophrys auratus) length (as fork length [cm]) (bin width = 2 cm) sampled at four sites in the Firth of Thames (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16). Red lines indicate mean snapper length for each site.The gut contents of snapper sampled at mussel farm sites had a range of 1–5 in digestion scores, and control sites had a range of 2–5 in digestion scores (Figure 3). There were overall significant differences in mean digestion scores between the two Locations and Treatments (F(57, 60) = 2.79, p = 0.04). The main factor effects of Location and Location × Treatment were significant (p = 0.04 and p = 0.01, respectively), but not the main factor Treatment (p = 0.42). Post hoc analyses identified that Rat Island control site had a significantly higher digestion score compared to Motukopake Island control site (p = 0.04; 0.9 ± 0.24 and 3.3 ± 0.16) and Rat Island mussel farm (p = 0.005; 3.9 ± 0.24 and 2.9 ± 0.29).3FIGURETop figure mean digestion score of gut contents of snapper sampled from each of the four sites (Motukopake Island control site, Motukopake Island mussel farm, Rat Island control site and Rat Island mussel farm). A digestion score of 0 represents completely undigested material and a score of 5 represents fully digested material (error bars represent standard error). Bottom figure mean wet weight (g) of snapper foregut and hindgut contents for the four sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat control site, n = 13; Rat Island mussel farm, n = 16).Total wet weight of snapper gut contents ranged from 1.26 to 30.08 g at Rat Island mussel farm, 1.29 to 17.99 g at Motukopake Island mussel farm, 3.26 to 15.33 g at Rat Island control site and 2.17 to 4.50 g at the Motukopake Island control site. Every snapper had some gut contents present; however, the majority of gut contents were in the hindgut. There were no differences in the mean wet weight of the foregut contents between Treatments (F(1, 57) = 0.16, p = 0.69), Location (F(1, 57) = 0.75, p = 0.39), or for the interacting effect of Treatment × Location (F(1, 57) = 2.56, p = 0.12) (Figure 3). Similarly, there were no differences in mean wet weight of the hindgut contents between Treatments (F(1, 57) = 1.36, p = 0.25), Location (F(1, 57) = 0.004, p = 0.95), or for the interacting effect of Treatment × Location (F(1, 57) = 0.89, p = 0.35) (Figure 3).Gut content composition in visual analysisThe gut contents of snapper were moderately to highly digested for all four sites (Figure 3), which made the identification of prey groups difficult, and required piecing together parts of the organism to identify what taxa it was. Most of the material came from the hindgut, with 25% of snapper in mussel farm sites and 17% of snapper in control sites with no material in the foregut. The Rat Island control site also had a higher total digestion score (average of hindgut and foregut) and a higher proportion of digested material compared to other sites, which is likely due to time since prey consumption rather than types of prey consumed. Overall, the digestion of the gut contents meant that the taxonomic level to which prey groups could be identified was most often limited. For example, the remains of Decapoda within the gut contents of snapper sampled at Rat Island control sites could not be identified to lower taxonomic levels, for example Caridea, Brachyura and Paguroidea.Gut content analysis for prey groups at broad taxonomic levelBivalvia were the dominant prey group in snapper gut contents, especially for the two mussel farm sites (Figure 4; Table 1). This was due to the high proportion of mussels present in the gut contents (i.e. 29% in Rat Island mussel farm and 18% in Motukopake Island mussel farm vs. 0% for both the Motukopake Island and Rat Island control site), which are within the Bivalvia category at the broad taxonomic level (Figure 4). In contrast, Decapoda were a common component of the gut contents at one control site but not at either of the mussel farm sites (i.e. 22.5% in Rat Island control site vs. 0% in Rat Island mussel farm) (Figure 4). Other Crustacea (i.e. excluding those prey groups that were classified into different Crustacea taxa such as Caridea, Decapoda, Brachyura and Paguroidea) were higher in mussel farm sites (i.e. 14.6% in Rat Island mussel farm vs. 1.4% in Rat Island control site), and Paguroidea was higher in control sites (i.e. 9.5% in Motukopake Island control site vs. 0% in the Motukopake Island mussel farm) (Figure 4; Table 1).4FIGUREMean percentage of prey groups at the broad taxonomic level present in snapper gut contents (hindgut and foregut combined), as a relative proportion of each prey group based on the total gut contents within each individual snapper. The relative proportions of each broad taxonomic group in individual snapper were averaged for each of the four sampling sites (Motukopake Island control site, Motukopake Island mussel farm, Rat Island control site and Rat Island mussel farm). Mussels (green‐lipped and blue) extracted from other Bivalvia for visual purposes only and not as part of the statistics at the broad taxonomic level.1TABLETotal amount of prey groups at the broad taxonomic level in gut contents of snapper sampled from each of four sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16)Prey groups at the broad taxonomic level in snapper gut contentsMotukopake Island control siteMotukopake Island mussel farmRat Island control siteRat Island mussel farmDigested material18.9415.3414.4813.25MolluscaBivalvia4.716.530.07512.01Gastropoda0.520.000.000.00Chitonida0.110.000.000.20ArthropodaBrachyura3.834.160.381.15Paguroidea3.370.000.830.02Decapoda0.530.945.990.00Other Crustacea0.353.330.426.07Caridea0.510.000.400.11AnnelidaPolychaeta0.040.880.000.52ChordataTeleostei0.000.002.420.00Ascidiacea0.000.320.000.00Shell debris0.10.000.000.16Note: Prey were classified into a broad taxonomic level, and the relative proportion measure (points) identifies the contribution of each group standardised by the gut fullness score, as calculated by the modified relative‐fullness method. Bolded values represent the highest prey group (excluding digested material) per site.The ordination plot of the amount of prey groups at the broad taxonomic level (derived from relative proportion data) in the snapper gut contents identified a partial separation between control and mussel farm sites; however, there is still considerable overlap between treatments (Figure 5). PERMANOVA identified significant differences in the composition of gut contents for Treatment (pseudoF(1, 57) = 6.39, p = 0.001) and Treatment × Location (pseudoF(1, 57) = 3.12, p = 0.01), but not for Location (pseudoF(1, 57) = 1.46, p = 0.18). Post hoc analyses indicated that the prey groups at a broad taxonomic level driving the differences were the categories Bivalvia, other Crustacea, Decapoda and Paguroidea (Table 2). Specifically, snapper sampled at Rat Island control site had consumed fewer Bivalvia compared to snapper sampled from all other sites (Table 2). Additionally, snapper from Motukopake Island mussel farm consumed more other Crustacea than both control sites. Rat Island control site snapper consumed more Decapoda compared to both mussel farm sites. Motukopake Island control site snapper consumed more Paguroidea compared to Motukopake Island mussel farm (Table 2).5FIGURENon‐metric multidimensional scaling (NMDS) plot of the relative proportion measure data of each of 13 prey groups at a broad taxonomic level found in snapper gut contents at four sites and presented for the two control and two mussel farm sites (see Table 1). Confidence interval (50%) ellipses displayed over Treatments.2TABLEWilcoxon test post hoc significant differences in the amount of prey groups at a broad taxonomic level consumed by individual snapper among the sites Motukopake Island mussel farm, Motukopake Island control site, Rat Island mussel farm and Rat Island control sitePrey group at a broad taxonomic levelComparisonMean and standard error (points)Adjusted p‐valueBivalviaRat Island control site < Motukopake Island control site0.006 ± 0.006 < 0.29 ± 0.070.01Rat Island control site < Rat Island mussel farm0.006 ± 0.006 < 0.75 ± 0.250.03Rat Island control site < Motukopake Island mussel farm0.006 ± 0.006 < 0.41 ± 0.130.04Other CrustaceaMotukopake Island control site < Motukopake Island mussel farm0.02 ± 0.02 < 0.21 ± 0.060.006Rat Island control site < Motukopake Island mussel farm0.02 ± 0.02 < 0.21 ± 0.060.04DecapodaMotukopake Island mussel farm < Rat Island control site0.06 ± 0.06 < 0.46 ± 0.170.02Rat Island mussel farm < Rat Island control site0 ± 0 < 0.46 ± 0.170.004PaguroideaMotukopake Island mussel farm < Motukopake Island control site0 ± 0 < 0.21 ± 0.080.04Note: Prey were classified into a broad taxonomic level, and the relative proportion measure identifies the relative contribution of each prey group calculated from the modified relative‐fullness method. Mean and standard error are from the raw data.Gut content analysis for prey groups classified to lowest taxonomic levelPrey identified to the lowest taxonomic level within snapper gut contents produced two sets of data. Firstly, the relative proportion measure derived from the proportions of prey items in gut content combined with the relative‐fullness measure was used to identify the relative amount of each prey group consumed among snapper (Table 3). Secondly, presence/absence data were used to calculate the percentage of individual snapper within a site that had the prey group present (Table 3). The highest amount of prey consumed by Motukopake Island mussel farm snapper was green‐lipped mussels (P. canaliculus); however, the most frequently present prey was the triangle barnacle (Balanus trigonus) (Table 3). At Rat Island mussel farm, the highest amount and most frequently present prey was green‐lipped mussels (Table 3). The highest amount of prey consumed by Motukopake Island control site snapper was wrinkled swimming crab (Liocarcinus corrugatus) and the most frequently present prey was hermit crabs in the Paguridae family (Table 3). At the Rat Island control site, the highest amount and most frequently present prey in snapper gut contents was Decapoda (Table 3).3TABLETotal amount (derived from relative proportion measure data) of prey groups classified to the lowest taxonomic level consumed by snapper, and total percentage of individual snapper with a prey group present (derived from presence/absence data), among the sites (Motukopake Island control site, n = 16; Motukopake Island mussel farm, n = 16; Rat Island control site, n = 13; Rat Island mussel farm, n = 16)Prey groups at the lowest taxonomic level in snapper gut contentsMotukopake Island control siteMotukopake Island mussel farmRat Island control siteRat Island mussel farmDigested material18.94 and 100%15.34 and 100%14.48 and 100%13.25 and 93%MolluscaBivalviaUnidentified Bivalvia0.48 and 6%0.000.000.04 and 6%Anomia trigonopsis1.28 and 25%0.03 and 6%0.00Atrina zelandica2.11 and 37%0.000.001.07 and 6%Austrovenus stutchburyi0.34 and 6%0.000.000.00Dosina spp.0.59 and 19%0.000.000.00Mytilus galloprovincialis planulatus0.000.60 and 13%0.004.04 and 31%Ostreidae spp.0.000.000.075 and 8%0.00Paphies australis0.08 and 6%0.000.000.00Perna canaliculus0.005.90 and 56%0.006.85 and 56%GastropodaTrochidae spp.0.52 and 13%0.000.000.00ChitonidaUnidentified chiton0.11 and 6%0.000.000.043 and 6%Cellana ornata0.000.000.000.16 and 6%ArthropodaBrachyuraUnidentified Brachyura0.1 and 6%0.61 and 19%0.38 and 8%0.17 and 6%Halicarcinus innominatus0.001.49 and 6%0.000.00Halicarcinus spp.0.000.18 and 6%0.000.00Liocarcinus corrugatus3.70 and 13%0.000.000.00Nepinnotheres novaezelandiae0.000.20 and 6%0.000.25 and 6%Notomithrax spp.0.04 and 13%0.86 and 13%0.000.72 and 6%Pilumnus novaezelandiae0.000.77 and 13%0.000.00Portunidae0.000.057 and 6%0.000.00PaguroideaPaguridae spp.2.86 and 38%0.000.83 and 23%0.02 and 6%Lophopagurus spp.0.50 and 19%0.000.000.00DecapodaUnidentified decapod0.53 and 13%0.94 and 6%5.99 and 62%–CrustaceaBalanus trigonus0.003.33 and 68%0.42 and 15%5.76 and 44%Epopella plicata0.000.000.000.32 and 6%Meiura0.35 and 6%0.000.000.00CarideaUnidentified Caridea0.02 and 6%0.000.33 and 15%0.11 and 6%Alpheus richardsoni0.000.000.07 and 8%0.00Biffarius filholi0.48 and 6%0.000.000.00AnnelidaPolychaetaUnidentified polychaete0.04 and 6%0.000.000.05 and 6%Eulalia microphylla0.000.45 and 6%0.000.00Serpulidae spp.0.000.43 and 38%0.000.47 and 13%ChordataTeleosteiForsterygion spp.0.000.002.42 and 8%0.00AscidiaceaStyela clava0.000.32 and 6%0.000.00Shell debrisMisc. shell debris0.1 and 6%0.000.000.16 and 6%Note: The relative proportion measure identifies the relative contribution of each prey item calculated by the modified gut fullness method. The percentages represent the proportion of snapper samples that had the prey item present, out of the total number of individuals within each site. Bolded values represent the highest value (excluding digested material) per site.Statistical analysis for prey groups classified to the lowest taxonomic level was only completed on presence/absence data. The ordination plot for presence/absence data of prey groups at the lowest taxonomic level identified separation between control and mussel farm sites (Figure 6). PERMANOVA identified significant differences in the composition of snapper gut contents for Treatment (pseudoF(1, 57) = 13.58, p = 0.001) and Treatment × Location (pseudoF(1, 57) = 3.46, p = 0.001), but not for Location (pseudoF(1, 57) = 2.11, p = 0.05). Post hoc analyses indicated that the main prey groups at the lowest taxonomic level driving the differences were triangle barnacle, green‐lipped mussel and other Decapoda (Table 4). Triangle barnacles and green‐lipped mussels occurred more frequently in Motukopake Island mussel farm compared to the Motukopake Island control site (Table 4). Motukopake Island mussel farm also had a higher frequency of green‐lipped mussels present in snapper gut contents compared to Rat Island control site. Similarly, the Rat Island mussel farm had a higher frequency of green‐lipped mussels present in snapper gut contents compared to Rat Island control site and Motukopake Island control site. The Rat Island control site had a higher frequency of other Decapoda compared to the mussel farm site (Table 4).6FIGURENon‐metric multidimensional scaling (NMDS) plot of presence/absence data for prey groups classified to the lowest taxonomic level within snapper gut contents at control and mussel farm sites (see Table 3). Confidence interval (50%) ellipses displayed over Treatments.4TABLEFishers exact test post hoc significant differences in the presence/absence of prey groups classified to the lowest taxonomic level within snapper gut contents between the sites Motukopake Island mussel farm, Motukopake Island control site, Rat Island mussel farm and Rat Island control sitePrey group at lowest taxonomic levelComparisonPercentAdjusted p‐valueTriangle barnacle (Balanus trigonus)Motukopake Island control site < Motukopake Island mussel farm0% vs. 68%0.002Green‐lipped mussel (Perna canaliculus)Motukopake Island control site < Motukopake Island mussel farm0% vs. 56%0.02Rat Island control site < Rat Island mussel farm0% vs. 56%0.04Motukopake Island control site < Rat Island mussel farm0% vs. 56%0.03Rat Island control site < Motukopake Island mussel farm0% vs. 56%0.04Decapoda spp.Rat Island mussel farm < Rat Island control site0% vs. 62%0.01Gut content composition in genetic analysisAt Motukopake Island, there were 20 families identified at the four sites (Figure 7). The greatest proportion of species at both Treatments was within the Malacostraca (Crustacea) class (Figure 7). Key differences were the presence of a high proportion of Mytilidae (mussels) and Caprellidae (Amphipoda) in mussel farm sites and none in the control site (Figure 7). Sesarmidae (Brachyura) had the highest proportion in control sites with none present at mussel farm sites (Figure 7).7FIGUREPresence/absence plot for DNA sequences identified in snapper gut contents at Motukopake Island mussel farm (n = 16) and Motukopake Island control site (n = 16), and Rat Island mussel farm (n = 16) and Rat Island control site (n = 13) categorised by family level (left‐hand side) and class level (right‐hand side). The size of the circles indicates the number of individuals where the prey group was identified, and the colour is the relative read abundance (RRA) which is a semi‐quantitative metric to understand dominance of the prey group within the snapper sample.At Rat Island, there were 15 families identified at the control site and 29 families identified at the mussel farm (Figure 7). The greatest proportion of families at the control site was within the Actinopteri (ray‐finned fish) and Malacostraca (Crustacea) class, and in the mussel farm it was Malacostraca (Figure 7). Mytilidae (mussels) and Skeletonemaceae (microalgae) were most present in the mussel farm, whereas Semilidae (clams) and Lithodidae (king crabs) were represented in greater number of samples in the Rat Island control site (Figure 7). The most dominant prey group (based on relative read abundance) in the mussel farm was the Dotillidae (crabs) and Moridae (cod fish) at the control site (Figure 7).PERMANOVA identified significant differences in the presence/absence of prey at the family level in the gut contents of snapper for Treatment (pseudoF(1, 44) = 5.66, p = 0.001), Location (pseudoF(1, 44) = 2.07, p = 0.007) and Treatment × Location (pseudoF(1, 44) = 2.12, p = 0.009). However, p‐adjusted values in post hoc analyses could not detect any significant differences among the four sites. Non‐adjusted p‐values showed some differences in families between Treatments. For example, the higher presence of Mytilidae in snapper gut contents sampled from Motukopake Island mussel farm (54%) and Rat Island mussel farms (55%) was significantly greater than the control sites (0%) (p = 0.002 and p = 0.01, respectively). Sesarmidae had a higher presence in snapper gut contents from Motukopake Island control site (50%) compared to Motukopake Island mussel farm and Rat Island mussel farm (0%) (p = 0.006 and p = 0.04 respectively). Motukopake Island mussel farm had a higher presence of Balanidae (barnacles) (46%) compared to the Motukopake Island control site and Rat Island control site (0%) (p = 0.006 and p = 0.02, respectively). Also, Caprellidae (Amphipoda) were present in snapper gut contents sampled from Motukopake Island mussel farm (38%) but absent at the Motukopake Island and Rat Island control sites (0%) (p = 0.02 and p = 0.04, respectively).PERMANOVA identified significant differences in the presence/absence of prey items at the class level in the gut contents for Treatment (pseudoF(1, 48) = 3.74, p = 0.007) and Treatment × Location (pseudoF(1, 48) = 2.48, p = 0.04), but not Location (pseudoF(1, 48) = 1.13, p = 0.34). However, p‐adjusted values in post hoc analyses could not detect any significant differences among sites. Non‐adjusted p‐values showed some differences in the presence/absence of the various classes of prey items in the gut contents between Treatment and Location. For example, snapper gut contents sampled from Motukopake Island mussel farm and Rat Island mussel farm had a greater presence of Bivalvia (including mussels) (47% and 46%) compared to Motukopake Island control site (0%) (p = 0.007 and p = 0.01, respectively). Motukopake Island mussel farm also had a greater presence of Thecostraca (e.g. barnacles) compared to both control sites (40% vs. 0%, p = 0.02). Snapper gut contents from the Motukopake Island control site had a greater presence of Conoidasida (parasitic alveolates) (35%) compared to Motukopake Island mussel farm (0%) (p = 0.03). Rat Island control site had a greater presence of Chromadorea (round worms) (36%) compared to Rat Island mussel farm (0%) and Motukopake Island control site (0%) (p = 0.03).DISCUSSIONThis study compared the difference in snapper gut contents between mussel farms and control sites to see whether prey groups identified in mussel farm snapper were sourced directly or indirectly from the mussel farm infrastructure. The results indicate that the snapper sampled from mussel farm habitat are consuming a distinctly different diet compared to those sampled from control sites. The differences in diet are due to the snapper utilising prey species made available through the presence of the mussel farm habitat, such as harvested and common biofouling species within the farm habitat, which are not available in nearby unstructured soft‐sediment habitats (i.e. the control sites).Differences in snapper gut contentsMussel farm versus control site snapperThe key differences in the gut contents of snapper between mussel farm and control sites were the significant contribution of mussels (green‐lipped and blue) and barnacles in mussel farms and not in control sites. Genetic analyses also confirmed that there were differences in the presence of caprellid amphipods which were present in mussel farms but not control sites, but these soft bodied Amphipoda were not picked up in visual analyses. There were some differences with Brachyura families between mussel farm and control sites, such as for Sesarmidae which were more commonly present in snapper gut contents at both control sites. The crab family Sesarmidae is not a common Brachyura family in New Zealand and therefore likely represents a distinct Brachyura species not yet available in the DNA reference databases (Wilkens & Ahyong, 2015). The other key difference in control sites compared to mussel farms was the presence of Paguridae (hermit crabs). Paguridae are abundant across a range of soft‐sediment and rocky reef locations, and have shown an association with mussel farm habitats overseas and in soft‐sediment mussel reefs in New Zealand (D'Amours et al., 2008; McLeod et al., 2014; Sean et al., 2022). Therefore, it is unclear why the abundance of Paguridae was higher in control sites and not mussel farms. It is possible that the abundance of shells which Paguridae rely on for host shells was more abundant in the control site, as Paguridae prefer host shells from Gastropoda species rather than Bivalvia, and Bivalvia shells (from mussels) are most abundant on the seafloor underneath mussel farms (McLaughlin & Thiel, 2015; Wilding & Nickell, 2013).The snapper consumption of/predation on cultured shellfish species (i.e. farmed green‐lipped mussels) is consistent with the general global analyses of the diet of other farm‐dwelling Sparidae species (Gerlotto et al., 2001; Hayden, 1995; Peteiro et al., 2010; Šegvić‐Bubić et al., 2011; Tsuyuki & Umino, 2017). However, most of these studies focused on the spat stage rather than mature mussels on a grow‐out cycle (Gerlotto et al., 2001; Hayden, 1995; Peteiro et al., 2010; Stenton‐Dozey & Broekhuizen, 2019; Tsuyuki & Umino, 2017). Sparid species are known to be significant predators of farmed shellfish species, especially mussels. For example, the diet of gilthead sea bream (Sparus aurata) sampled from a Croatian mussel farm consisted of 70% cultured mussels, which was considerably higher than for the snapper sampled in the mussel farm sites in the current study, with up to 19% consumption of cultured green‐lipped mussels. Snapper in the current study were also captured on the seafloor and so are likely consuming mussels which have fallen off the dropper lines as well as those present on the dropper lines. Additionally, the Croatian study showed no consumption of Crustacea by gilthead sea bream in mussel farms, which is surprising given that Crustacea are a dominant food resource for Sparidae and are commonly associated with mussel farms (Callier et al., 2018; McKindsey et al., 2011; Šegvić‐Bubić et al., 2011; Usmar, 2012; Woods et al., 2012). Nonetheless, gilthead sea bream did consume Teleostei and more Gastropoda compared to the current study, suggesting that the site‐specific differences in biofouling or the differences in diet between the two Sparidae species have impacted their respective diet compositions (Šegvić‐Bubić et al., 2011).Snapper in mussel farms also consumed blue mussels (up to 10.1%) and barnacles (up to 14%), which are key nuisance biofouling species in New Zealand green‐lipped mussel farms (Forrest & Atalah, 2017; Woods et al., 2012; Zazzaro et al., 2018). Since 2009, there has been a dramatic increase in the amount of fouling of the triangle barnacle (B. trigonus) across mussel farms in the Hauraki Gulf (Zazzaro et al., 2018). Globally, it is estimated that aquaculture operations spend 5%–10% of total profits managing biofouling, and so if snapper are selectively predating on these nuisance fouling organisms, this could be of benefit to mussel farm operations (Zazzaro et al., 2018). Consequently, there is a potential trade‐off between snapper consuming harvested green‐lipped mussels, but also having some benefit from reducing nuisance biofouling within the farm. Given that the overall consumption of harvested mussels by snapper was a lot lower than previously documented for another sparid species (19% vs. 70%), the predation may not be significant enough to impact mussel production at the intermediate to final grow out stages.Caprellids found in mussel farm snapper gut contents were likely present within the mussels on dropper lines as these are common biofouling species that associate with biogenic structure, for example mussels, hydroids and tunicates (Lim & Harley, 2018; South et al., 2019; Woods et al., 2012, 2014). Caprellids are particularly abundant on dropper lines during seed out of juvenile mussels (South et al., 2019). The presence of caprellids, mussels and barnacles within mussel farm snapper gut contents suggests that snapper are feeding on the dropper lines as well as the benthos, as these species are much more likely to be present on the structural substrate of the dropper lines (South et al., 2019; Woods et al., 2012; Zazzaro et al., 2018). Research that investigated the abundance and diversity of snapper within Coromandel mussel farms showed that snapper were observed at the surface 34% of the time and at the benthos 66% of the time, based on 89.6 h of footage (Underwood, 2023). These observations show that although snapper were present on the benthos for the majority of the time, they still visited the dropper lines presumably for feeding opportunities, as also confirmed by the composition of prey items in snapper gut contents.There were other key indicator species that were present in mussel farm snapper but not control sites which confirmed that mussel farm snapper are consuming biofouling within the farm environment. These included Notomithrax and Halicarcinus Brachyura species which are both common inhabitants of mussel dropper lines (Woods et al., 2012). The pea crab (Nepinnotheres novaezelandiae) was also present in mussel farm snapper which is typically a parasitic crab only within blue and green‐lipped mussels (Trottier & Jeffs, 2015a, 2015b; Trottier et al., 2012; Wilkens & Ahyong, 2015). Additionally, some Bivalvia which were present in control site snapper and not mussel farms, such as Dosina spp., cockle (Austrovenus stutchburyi) and pipi (Paphies australis), are more commonly found within soft, muddy and fine sand benthic habitats with no direct association with mussel farm benthic habitat where organic loading from the shellfish farming is typically higher (Powell, 1979; Wilding & Nickell, 2013).Comparison to broader Hauraki Gulf snapper populationPrevious analyses on snapper gut contents within the Hauraki Gulf have covered broader habitats and snapper size ranges which has resulted in a wide range of species and taxonomic groups identified in gut contents (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). Snapper have been observed to consume nearly 100 different species across a range of soft‐sediment sampling stations in the Hauraki Gulf, which is much higher than the 34 species identified in the visual gut analysis in the current study (Colman, 1972; Godfriaux, 1970). For example, higher proportions of Polychaeta (5%–10% vs. 0%–2%), Teleostei (5%–10% vs. 0%–4%) and Gastropoda (5%–15% vs. 0%–1.6%) have been previously identified in the gut contents of snapper living in natural habitats in the Hauraki Gulf (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). Additionally, Echinodermata and Porifera were present in broader diet studies, with Echinodermata making up a large proportion of gut contents (9.6%–28.3%) (Colman, 1972; Godfriaux, 1970; Russell, 1983). These comparisons have confirmed the plasticity of snapper diet even within different habitats found within one marine region.Overall, there is general consensus that mostly free‐living Crustacea are the dominant prey group targeted by snapper, with all previous diet studies recording this taxa as the dominant component (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). This is generally consistent with the current study; however, the dominance of Bivalvia within mussel farm snapper gut contents exceeds the proportion of Crustacea at the Rat Island mussel farm (32% Bivalvia vs. 18.4% Crustacea), and is high but does not exceed the proportion of Crustacea within Motukopake Island mussel farm (18% Bivalvia and 26% Crustacea). Both control sites had a dominance of Crustacea—29% at Rat Island control site and 21.4% at Motukopake Island control site. Of the Crustacea, Brachyura was found to be the dominant component in four previous studies of snapper diet in New Zealand (Colman, 1972; Godfriaux, 1970; Russell, 1983; Usmar, 2012); however, Paguroidea and Caridea were more dominant in two studies that focused on a smaller size range of adult snapper of 25–55 cm (Drummond, 2020; Third, 2022). Snapper typically start consuming Brachyura from 10 cm (fork length) and their diet becomes more varied from 25 cm to include Bivalvia, Paguroidea, Polychaeta, Gastropoda and Teleostei (Usmar, 2012). This is mostly due to the jaw crushing strength of the snapper which develops with age/increase in size, allowing snapper to successfully consume more hard‐shelled organisms (Usmar, 2012). In the current study (which also used a refined size class of 26–42 cm), the dominant Crustacea component within mussel farm snapper were barnacles, mostly triangle barnacle (B. trigonus). This is an important finding as no other previous snapper study within the Hauraki Gulf has identified barnacles as a diet component (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). For control sites, Paguroidea were dominant in Motukopake Island, aligning with the recent soft‐sediment snapper studies (Drummond, 2020; Third, 2022). Decapoda spp. were dominant in Rat Island which could be from any of the Brachyura, Caridea and Paguridae families.Of the Brachyura identified by our genetic and visual gut analysis methods, only four species had been previously recorded. The dwarf swimming crab (Liocarcinus corrugatus), identified in control sites, was present in two previous snapper gut studies and dominant in one (Drummond, 2020; Third, 2022). Other Brachyura found in mussel farm snapper were from the Majidae (decorator crab) family, including Notomithrax species which were previously found in most Hauraki Gulf snapper gut studies (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). Similarly, Halicarcinus spp. were present in mussel farm snapper and are commonly consumed by snapper sampled from natural habitats in the Hauraki Gulf (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022; Usmar, 2012). Even though these crab species are present in natural soft‐sediment habitats, there is likely to be an increased abundance of them within the mussel farm habitat (Wilkens & Ahyong, 2015; Woods et al., 2012).The overall proportion of Bivalvia within the gut contents of mussel farm snapper largely exceeded the proportions in previous studies, for example 32%–18% in the current study versus 8%–13% (Godfriaux, 1970; Third, 2022). The Bivalvia species previously identified in the gut contents of snapper include Dosina spp., Anomia spp., A. stutchburyi, Atrina zelandica, Tawera spissa, Soletellina nitida, Purpurocardia purpurata, Neilo australis, Gari stangeri and Linucula hartvigiana (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Third, 2022). These Bivalvia species previously observed in the gut contents of snapper sampled from natural habitats have four species in common with the control sites over soft‐sediment natural habitat in this study, and only two in mussel farm sites. Interestingly, Motukopake Island control site snapper also had a high proportion of Bivalvia (excluding mussels) in gut contents (16%) exceeding previous studies (Godfriaux, 1970; Third, 2022). This is likely due to the availability of Bivalvia in the benthic habitat at this specific site, as snapper sampled from Rat Island only included 0.15% Bivalvia in their gut contents. Green‐lipped mussels have been rarely reported in the gut contents of snapper sampled from natural habitats (Third, 2022; Usmar, 2012), but snapper are known to predate on them when they are available (Alder et al., 2021, 2022a, 2022b). Blue mussels have never been recorded in previous snapper diets in New Zealand (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). However, the detection of the consumption of mussels relies on the presence of this species in sufficient quantities within the habitat.Comparison of visual and genetic gut analysis resultsOverall, there was good alignment between the visual and genetic gut analyses for the key species, such as mussels and barnacles. Additionally, many of the Brachyura families identified by the molecular genetic analysis precisely matched species of Brachyura identified by the visual gut analysis, such as Majidae (Notomithrax spp.), Hymenosomatidae (Halicarcinus spp.), Pilumnidae (e.g. Pilumnus novaezelandiae) and Pinnotheridae (Nepinnotheres spp.). The genetic analysis was also able to identify a broader range of taxa, including soft body and small organisms such as Ascidiacea (tunicates), Holothuroidea (sea cucumbers), Polychaeta and Hoplonemertea (worms) and amphipods (small crustaceans). Identifying these types of soft‐bodied organisms to an informative level is often not possible using visual gut analysis due to their rapid digestion once consumed, masking their morphological structure needed for identification (O'Rorke et al., 2012; van der Reis et al., 2020). However, there were insufficient numbers of individual snapper for which DNA was detected from these prey groups to result in a significant difference between snapper from control versus mussel farm sites. Although amplification was achieved for the majority of samples, the DNA extracted from the gut contents was degraded (visible when running a 0.8% agarose DNA gel), likely hindering better amplification, and may be a limitation to identifying the full scope of taxa within the gut content in this study (O'Rorke et al., 2012; van der Reis et al., 2020). Although appropriate protocols were followed, the logistic constraints of field sampling and multiple sampling of gut contents (i.e. visual gut analysis and sampling for biochemical analysis) meant that gut contents could not be immediately preserved with chemicals (i.e. ethanol) upon capture of the snapper which would have been optimum for molecular genetic analyses (van der Reis & Lavery, 2020).Taxonomic assignment of DNA sequences is reliant on a match that reaches the required minimum accuracy threshold that is typically established on the basis of the availability of sequences in the reference databases for the suspected prey types (van der Reis & Lavery, 2020). In New Zealand, there is a lack of references for Crustacea species in the DNA databases (particularly noticeable for crabs) (van der Reis et al., 2020). Therefore, a BLAST 85% identity match was chosen to be able to assign closest matches for these species (van der Reis et al., 2020). The lack of New Zealand Crustacea families sequenced has resulted in some crab misidentifications (five Malacostraca families) and thus few common coastal species were identified (Forest & McLay, 2001; Wilkens & Ahyong, 2015). Regardless, distinct differences were found for crab families identified in one treatment and not the other.Implications for Hauraki gulf snapper populationOverall, the results show that snapper within mussel farms are consuming a unique diet, with prey groups linked to the mussel farm habitat, either underneath mussel farms and/or on dropper line habitats. In particular, the key prey groups which snapper were consuming in mussel farm habitats (mussels and barnacles) are highly abundant within mussel farms (Forrest & Atalah, 2017; Skelton et al., 2022; Woods et al., 2012; Zazzaro et al., 2018). Firstly, the green‐lipped mussel industry produces approximately 1.78 billion adult mussels per year, with 30% of this production within the Coromandel region, which sums to approximately 530 million adults per year (Skelton et al., 2022). Secondly, blue mussels and barnacles are key nuisance biofouling species in Coromandel mussel farms as these can foul surfaces at high densities across mussel dropper line habitats (Forrest & Atalah, 2017; Woods et al., 2012; Zazzaro et al., 2018). Abundant prey groups can especially be targeted by snapper because of their omnivorous and highly diverse diets (Colman, 1972; Drummond, 2020; Godfriaux, 1970; Russell, 1983; Third, 2022; Usmar, 2012). Snapper have been observed at much higher densities within mussel farms compared to adjacent control sites (i.e. soft‐sediment without mussel farms), with snapper up to five times more abundant underneath mussel farms (Underwood, 2023). Therefore, the higher abundances of snapper within mussel farms are likely to be partially due to the availability of prey given the direct link to prey species present in mussel farm habitats. A stable prey supply is important as it reduces the energy partitioned for foraging, so that there is more energy available for other biological processes such as growth and reproduction (Wootton, 2012). Furthermore, the prey groups targeted in mussel farms are energy rich, with high lipid content present in mussels and barnacles (Barclay et al., 2006; Barnes & Achituv, 1976). Consequently, snapper are likely receiving a nutritional benefit from feeding within mussel farm habitats. The highly abundant mussel farm habitats within the Coromandel have the potential to support the Hauraki Gulf snapper population through the supply of food resources. This is a direct outcome that can be used towards restorative aquaculture frameworks, to utilise farm habitats in a way that provides net positive ecological outcomes for coastal environments (Alleway et al., 2019; Froehlich et al., 2017; Gentry et al., 2020; Theuerkauf et al., 2019).CONCLUSIONSThe results from this study show that snapper living within mussel farms in the Hauraki Gulf are consuming a distinctly different diet to snapper in adjacent natural habitats without a mussel farm. The diet that snapper have consumed within mussel farms can be directly linked to the cultured species (i.e. green‐lipped mussels) and the biofouling (i.e. blue mussels and barnacles). There was good alignment between the visual gut and genetic analyses for the key species identified. The DNA degradation within individual snapper samples has likely impacted the lack of additional differences identified. Overall, the results show that the highly abundant prey groups consumed by snapper in mussel farm habitats are likely to be beneficial to the snapper population, reducing foraging effort and potentially supplying more nutritious prey. These findings provide evidence towards the supporting services of mussel farm habitats through the provision of food resources. This evidence can support the restorative aquaculture objectives for the region, which aims to provide net positive ecological outcomes for wild fish populations.AUTHOR CONTRIBUTIONSLucy H. Underwood: Conceptualization; formal analysis; investigation; methodology; writing—original draft. Aimee van der Reis: Conceptualization; methodology; writing—review and editing. Andrew G. Jeffs: Funding acquisition; project administration; supervision; writing—review and editing.ACKNOWLEDGEMENTSWe thank Gold Ridge Marine farms for providing access to the mussel farm sites and for logistic support. We also thank the Coromandel Marine Farmers’ Association (CMFA) and their members for their support. Thanks to Esther Stuck, Brad Skelton, David DuPavillon, Bruce Underwood and Joe Dennehy for fieldwork support. Thanks to Ash Heaphy for laboratory support. We are grateful to the Nature Conservancy for financial support for this research and to the Ministry for Primary Industries’ Sustainable Food and Fibre Futures Fund (#22074).Open access publishing facilitated by The University of Auckland, as part of the Wiley ‐ The University of Auckland agreement via the Council of Australian University Librarians.ETHICS STATEMENTThis research was conducted in accordance with animal ethics approval under New Zealand's Animal Welfare Act 1999 (UoA‐AEC Approval # 21619), which ensures the humane treatment of animals for research.CONFLICT OF INTEREST STATEMENTThe authors declare no conflicts of interest.DATA AVAILABILITY STATEMENTThe raw sequence data is available on NCBI's Sequence Read Archive (http://www.ncbi.nlm.nih.gov/bioproject/971684) with associated metadata. The underlying data for statistical analyses will be shared on reasonable request to the corresponding author.PEER REVIEWThe peer review history for this article is available at: https://publons.com/publon/10.1002/aff2.113REFERENCESAlder, A., Jeffs, A.G. & Hillman, J.R. (2022a) The importance of stock selection for improving transplantation efficiency. Restoration Ecology, 30(4), e13561. Available from: https://doi.org/10.1111/rec.13561Alder, A., Jeffs, A. & Hillman, J. (2022b) Timing mussel deployments to improve reintroduction success and restoration efficiency. Marine Ecology Progress Series, 698, 69–83. Available from: https://doi.org/10.3354/meps14157Alder, A., Jeffs, A. & Hillman, J.R. (2021) Considering the use of subadult and juvenile mussels for mussel reef restoration. Restoration Ecology, 29(3), e13322. https://doi.org/10.1111/rec.13322Alleway, H.K., Gillies, C.L., Bishop, M.J., Gentry, R.R., Theuerkauf, S.J. & Jones, R. (2019) The ecosystem services of marine aquaculture: valuing benefits to people and nature. Bioscience, 69(1), 59–68. https://doi.org/10.1093/biosci/biy137Amundsen, P. & Sánchez‐Hernández, J. (2019) Feeding studies take guts – critical review and recommendations of methods for stomach contents analysis in fish. Journal of Fish Biology, 95(6), 1364–1373. https://doi.org/10.1111/jfb.14151Anyango, J.O., Mlewa, C.M. & Mwaluma, J. (2017) Abundance, diversity and trophic status of wild fish around seaweed farms in Kibuyuni, South Coast Kenya. International Journal of Fisheries and Aquatic Studies, 5(3), 440–446.Baker, R., Buckland, A. & Sheaves, M. (2014) Fish gut content analysis: robust measures of diet composition. Fish and Fisheries, 15(1), 170–177. https://doi.org/10.1111/faf.12026Barclay, M.C., Irvin, S.J., Williams, K.C. & Smith, D.M. (2006) Comparison of diets for the tropical spiny lobster Panulirus ornatus: astaxanthin‐supplemented feeds and mussel flesh. Aquaculture Nutrition, 12(2), 117–125. https://doi.org/10.1111/j.1365‐2095.2006.00390.xBarnes, H. & Achituv, Y. (1976) The utilization of various biochemical entities in gonadally mature Balanus balanoides (L.) under starvation and feeding in the absence of copulation. Journal of Experimental Marine Biology and Ecology, 22(3), 257–262.Barnett, A., Redd, K.S., Frusher, S.D., Stevens, J.D. & Semmens, J.M. (2010) Non‐lethal method to obtain stomach samples from a large marine predator and the use of DNA analysis to improve dietary information. Journal of Experimental Marine Biology and Ecology, 393(1), 188–192. https://doi.org/10.1016/j.jembe.2010.07.022Barrett, L.T., Theuerkauf, S.J., Rose, J.M., Alleway, H.K., Bricker, S.B., Parker, M., et al. (2022) Sustainable growth of non‐fed aquaculture can generate valuable ecosystem benefits. Ecosystem Services, 53, 101396. https://doi.org/10.1016/j.ecoser.2021.101396Berry, O., Bulman, C., Bunce, M., Coghlan, M., Murray, D.C. & Ward, R.D. (2015) Comparison of morphological and DNA metabarcoding analyses of diets in exploited marine fishes. Marine Ecology Progress Series, 540, 167–181. https://doi.org/10.3354/meps11524Benson, D.A., Cavanaugh, M., Clark, K., Karsch‐Mizrachi, I., Lipman, D.J., Ostell, J. & Sayers, E.W. (2013) GenBank. Nucleic Acids Res, 41, D36–42. https://doi.org/10.1093/nar/gks1195Binning, S.A. & Chapman, L.J. (2010) Is intraspecific variation in diet and morphology related to environmental gradients? Exploring Liem's paradox in a cichlid fish. Integrative Zoology, 5(3), 241–255. https://doi.org/10.1111/j.1749‐4877.2010.00209.xBoecklen, W.J., Yarnes, C.T., Cook, B.A. & James, A.C. (2011) On the use of stable isotopes in trophic ecology. Annual Review of Ecology, Evolution, and Systematics, 42(1), 411–440. https://doi.org/10.1146/annurev‐ecolsys‐102209‐144726Bolyen, E., Rideout, J.R., Dillon, M.R., Abnet, C.C., Al‐Ghalith, G.A., Alexander, H., et al. (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature Biotechnology, 37(9), 1091. https://doi.org/10.1038/s41587‐019‐0252‐6Braga, R., Bornatowski, H. & Vitule, J. (2012) Feeding ecology of fishes: an overview of worldwide publications. Reviews in Fish Biology and Fisheries, 22(4), 915–929. https://doi.org/10.1007/s11160‐012‐9273‐7Budge, S.M., Iverson, S.J. & Koopman, H.N. (2006) Studying trophic ecology in marine ecosystems using fatty acids: a primer on analysis and interpretation. Marine Mammal Science, 22(4), 759–801. https://doi.org/10.1111/j.1748‐7692.2006.00079.xCallahan, B.J., McMurdie, P.J. & Holmes, S.P. (2017) Exact sequence variants should replace operational taxonomic units in marker‐gene data analysis. The ISME Journal, 11(12), 2639–2643. https://doi.org/10.1038/ismej.2017.119Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.A. & Holmes, S.P. (2016) DADA2: high‐resolution sample inference from Illumina amplicon data. Nature Methods, 13(7), 581–583. https://doi.org/10.1038/NMETH.3869Callier, M.D., Byron, C.J., Bengtson, D.A., Cranford, P.J., Cross, S.F., Focken, U., et al.(2018) Attraction and repulsion of mobile wild organisms to finfish and shellfish aquaculture: a review. Reviews in Aquaculture, 10(4), 924–949. https://doi.org/10.1111/raq.12208Castro, A.L.C., de Farias Diniz, A., Martins, I.Z., Vendel, A.L., de Oliveira, T.P.R. & de Lucena Rosa, I.M. (2008) Assessing diet composition of seahorses in the wild using a non‐destructive method: Hippocampus reidi (Teleostei: Syngnathidae) as a study‐case. Neotropical Ichthyology, 6(4), 637–644. https://doi.org/10.1590/S1679‐62252008000400012Clynick, B.G., McKindsey, C.W. & Archambault, P. (2008) Distribution and productivity of fish and macroinvertebrates in mussel aquaculture sites in the Magdalen islands (Québec, Canada). Aquaculture, 283(1), 203–210. https://doi.org/10.1016/j.aquaculture.2008.06.009Cole, A. (2010) Cleaning to corallivory: ontogenetic shifts in feeding ecology of tubelip wrasse. Coral Reefs, 29(1), 125–129. https://doi.org/10.1007/s00338‐009‐0563‐zColman, J.A. (1972) Food of snapper, Chrysophrys auratus (Forster), in the Hauraki Gulf, New Zealand. New Zealand Journal of Marine and Freshwater Research, 6(3), 221–239. https://doi.org/10.1080/00288330.1972.9515422D'Amours, O., Archambault, P., McKindsey, C. & Johnson, L. (2008) Local enhancement of epibenthic macrofauna by aquaculture activities. Marine Ecology Progress Series, 371, 73–84. https://doi.org/10.3354/meps07672Deagle, B.E., Thomas, A.C., McInnes, J.C., Clarke, L.J., Vesterinen, E.J., Clare, E.L., Kartzinel, T.R. & Eveson, J.P. (2019) Counting with DNA in metabarcoding studies: how should we convert sequence reads to dietary data? Molecular Ecology, 28(2), 391–406. https://10.1111/mec.14734Drummond, L. (2020) How does diet affect snapper (Chrysophrys auratus) movement? (Masters dissertation). Auckland: University of Auckland.Fernandez‐Jover, D. & Sanchez‐Jerez, P. (2015) Comparison of diet and otolith growth of juvenile wild fish communities at fish farms and natural habitats. ICES Journal of Marine Science, 72(3), 916–929. https://doi.org/10.1093/icesjms/fsu153Fernandez‐Jover, D., Sanchez‐Jerez, P., Bayle‐Sempere, J.T., Arechavala‐Lopez, P., Martinez‐Rubio, L., Jimenez, J.A.L. & Lopez, F.J.M. (2009) Coastal fish farms are settlement sites for juvenile fish. Marine Environmental Research, 68(2), 89–96. https://doi.org/10.1016/j.marenvres.2009.04.006Forest, J. & McLay, C.L. (2001) The biogeography and bathymetric distribution of New Zealand hermit crabs (Crustacea: Anomura: Paguridea). Journal of the Royal Society of New Zealand, 31(4), 687–720. https://doi.org/10.1080/03014223.2001.9517670Forrest, B.M. & Atalah, J. (2017) Significant impact from blue mussel Mytilus galloprovincialis biofouling on aquaculture production of green‐lipped mussels in New Zealand. Aquaculture Environment Interactions, 9, 115–126. https://doi.org/10.3354/aei00220Froehlich, H.E., Gentry, R.R. & Halpern, B.S. (2017) Conservation aquaculture: shifting the narrative and paradigm of aquaculture's role in resource management. Biological Conservation, 215, 162–168. https://doi.org/10.1016/j.biocon.2017.09.012Geller, J., Meyer, C., Parker, M. & Hawk, H. (2013) Redesign of PCR primers for mitochondrial cytochrome c oxidase subunit I for marine invertebrates and application in all‐taxa biotic surveys. Molecular Ecology Resources, 13(5), 851–861. https://doi.org/10.1111/1755‐0998.12138Gentry, R.R., Alleway, H.K., Bishop, M.J., Gillies, C.L., Waters, T. & Jones, R. (2020) Exploring the potential for marine aquaculture to contribute to ecosystem services. Reviews in Aquaculture, 12(2), 499–512. https://doi.org/10.1111/raq.12328Gerlotto, F., Brehmer, P., Buestel, D. & Sanguinede, F. (2001) A method for acoustic monitoring of a mussel longline ground using vertical echosounder and multibeam sonar. In: ICES Annual Science Conference, 26–29 September 2001, Oslo.Gibbs, M.T. (2004) Interactions between bivalve shellfish farms and fishery resources. Aquaculture, 240(1), 267–296. https://doi.org/10.1016/j.aquaculture.2004.06.038Godfriaux, B.L. (1970) Food of predatory demersal fish in Hauraki Gulf. New Zealand Journal of Marine and Freshwater Research, 4(4), 325–336. https://doi.org/10.1080/00288330.1970.9515351Hayden, B.J. (1995) Factors affecting recruitment of farmed greenshell mussels, Perna canaliculus (Gmelin) 1791, in Marlborough Sounds (Doctoral dissertation). Christchurch, New Zealand: University of Canterbury. Available from: http://hdl.handle.net/10523/9115.Hehre, E.J. & Meeuwig, J.J. (2016) A global analysis of the relationship between farmed seaweed production and herbivorous fish catch. PLoS ONE, 11(2), e0148250. https://doi.org/10.1371/journal.pone.0148250Horton, T., Kroh, A., Ahyong, S., Bailly, N., Bieler, R., Boyko, C.B., et al. (2022) World Register of Marine Species (WoRMS). Ostend, Belgium: World Register of Marine Species (WoRMS).Illumina. (2013) 6S metagenomic sequencing library preparation: preparing 16S ribosomal RNA gene amplicons for the Illumina MiSeq system. Report no. 15044223 Rev. B. Available from: https://support.illumina.com/documents/documentation/chemistry_documentation/16s/16s‐metagenomic‐library‐prep‐guide‐15044223‐b.pdfKamler, J.F. & Pope, K.L. (2001) Nonlethal methods of examining fish stomach contents. Reviews in Fisheries Science, 9(1), 1–11. https://doi.org/10.1080/20016491101663Leray, M., Yang, J.Y., Meyer, C.P., Mills, S.C., Agudelo, N., Ranwez, V., et al. (2013) A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: application for characterizing coral reef fish gut contents. Frontiers in Zoology, 10(1), 34. https://doi.org/10.1186/1742‐9994‐10‐34Lim, E.G. & Harley, C.D.G. (2018) Caprellid amphipods (Caprella spp.) are vulnerable to both physiological and habitat‐mediated effects of ocean acidification. PeerJ, 6, e5327. https://doi.org/10.7717/peerj.5327Lobo, J., Costa, P.M., Teixeira, M.A., Ferreira, M.S., Costa, M.H. & Costa, F.O. (2013) Enhanced primers for amplification of DNA barcodes from a broad range of marine metazoans. BMC Ecology, 13(1), 34. https://doi.org/10.1186/1472‐6785‐13‐34Mahesh, V., Gop, A. & Nair, R.J. (2018) Stomach content analysis techniques in fishes. In: Nair, R. J., Gop, A. P. & Mahesh, V. (Eds.) Recent advances in fishery biology techniques for biodiversity evaluation and conservation. Kochi, India: Central Marine Fisheries Research Institute, pp. 104–115.Martin, M. (2011) Cutadapt removes adapter sequences from high‐throughput sequencing reads. EMBnet.Journal, 17(1), 10–12. https://doi.org/10.14806/EJ.17.1.200McKindsey, C.W., Archambault, P., Callier, M.D. & Olivier, F. (2011) Influence of suspended and off‐bottom mussel culture on the sea bottom and benthic habitats: a review. Canadian Journal of Zoology, 89(7), 622–646. https://doi.org/10.1139/Z11‐037McLaughlin, P.A. & Thiel, M. (2015) Crustaceans in mobile homes. The Natural History of the Crustacea, 2, 145–162.McLeod, I., Parsons, D., Morrison, M., Van Dijken, S. & Taylor, R. (2014) Mussel reefs on soft sediments: a severely reduced but important habitat for macroinvertebrates and fishes in New Zealand. New Zealand Journal of Marine and Freshwater Research, 48(1), 48–59. https://doi.org/10.1080/00288330.2013.834831Morton, J., Platell, M. & Gladstone, W. (2008) Differences in feeding ecology among three co‐occurring species of wrasse (Teleostei: Labridae) on rocky reefs of temperate Australia. Marine Biology, 154(3), 577–592. https://doi.org/10.1007/s00227‐008‐0951‐xO'Rorke, R., Lavery, S. & Jeffs, A. (2012) PCR enrichment techniques to identify the diet of predators. Molecular Ecology Resources, 12(1), 5–17. https://doi.org/10.1111/j.1755‐0998.2011.03091.xPeteiro, L.G., Filgueira, R., Labarta, U. & Fernández‐Reiriz, M.J. (2010) The role of fish predation on recruitment of Mytilus galloprovincialis on different artificial mussel collectors. Aquacultural Engineering, 42(1), 25–30. https://doi.org/10.1016/j.aquaeng.2009.09.003Powell, A.W.B. (1979) New Zealand mollusca: marine, land, and freshwater shells. Auckland: Collins.R Core Team. (2021) R: a language and environment for statistical computing. Version 4.1.0; R studio version 1.4.1106. Vienna: R Foundation for Statistical Computing.Rennó Braga, R., Ribeiro, V., Bornatowski, H., Abilhoa, V. & Vitule, S.J. (2017) Gastric lavage for dietary studies of small fishes: efficiency, survival and applicability. Acta Ichthyologica Et Piscatoria, 47(1), 97–100. https://doi.org/10.3750/AIEP/02079Russell, B.C. (1983) Food and feeding habits of rocky reef fish of north‐eastern New Zealand. New Zealand Journal of Marine and Freshwater Research, 17(2), 121–145. https://doi.org/10.1080/00288330.1983.9515991Sanchez‐Jerez, P., Fernandez‐Jover, D., Uglem, I., Arechavala‐Lopez, P., Dempster, T., Bayle‐Sempere, J.T., et al. (2011) Coastal fish farms as fish aggregation devices (FADs). In: Bortone, S. A., Pereira Brandini, F., Fabi, G. & Otake, S. (Eds.) Artificial reefs in fishery management. Boca Raton, FL: CRC Press Taylor & Francis Group, pp. 187–208.Sean, A., Drouin, A., Archambault, P. & McKindsey, C.W. (2022) Influence of an offshore mussel aquaculture site on the distribution of epibenthic macrofauna in Îles de la Madeleine, Eastern Canada. Frontiers in Marine Science, 9, 859816. https://doi.org/10.3389/fmars.2022.859816Šegvić‐Bubić, T., Grubišić, L., Karaman, N., Tičina, V., Jelavić, K.M. & Katavić, I. (2011) Damages on mussel farms potentially caused by fish predation—Self‐service on the ropes? Aquaculture, 319(3), 497–504. https://doi.org/10.1016/j.aquaculture.2011.07.031Skelton, B.M., South, P.M. & Jeffs, A.G. (2022) Inefficiency of conversion of seed into market‐ready mussels in New Zealand's Greenshell™ mussel (Perna canaliculus) industry. Aquaculture, 560, 738584. https://doi.org/10.1016/j.aquaculture.2022.738584Skog, T., Hylland, K., Torstensen, B.E. & Berntssen, M.H.G. (2003) Salmon farming affects the fatty acid composition and taste of wild saithe Pollachius virens. Aquaculture Research, 34(12), 999–1007. https://doi.org/10.1046/j.1365‐2109.2003.00901.xSouth, P.M., Floerl, O. & Jeffs, A.G. (2019) The role of biofouling development in the loss of seed mussels in aquaculture. Biofouling, 35(2), 259–272. https://doi.org/10.1080/08927014.2019.1596261Stenton‐Dozey, J. & Broekhuizen, N. (2019) Provision of ecological and ecosystem services by mussel farming in the Marlborough Sounds: a literature review in context of the state of the environment pre and post‐mussel farming. Prepared for Marine Farming Association, National Institute of Water & Atmospheric Research. Available from: https://www.marinefarming.co.nz/site_files/24792/upload_files/Fullreport_28.07.2021update.pdf?dl=1Supono, S., Dunphy, B. & Jeffs, A. (2021) Role of nutritional history in the attachment and retention of green‐lipped mussel spat. Aquaculture, 533, 736094. https://doi.org/10.1016/j.aquaculture.2020.736094Theuerkauf, S.J., Barrett, L.T., Alleway, H.K., Costa‐Pierce, B.A., St Gelais, A. & Jones, R.C. (2022) Habitat value of bivalve shellfish and seaweed aquaculture for fish and invertebrates: pathways, synthesis and next steps. Reviews in Aquaculture, 14(1), 54–72. https://doi.org/10.1111/raq.12584Theuerkauf, S.J., Morris, J.A., Waters, T.J., Wickliffe, L.C., Alleway, H.K. & Jones, R.C. (2019) A global spatial analysis reveals where marine aquaculture can benefit nature and people. PLoS ONE, 14(10), e0222282. https://doi.org/10.1371/journal.pone.0222282Third, G. (2022) Morphological and diet variation in Chrysophrys auratus populations (Masters dissertation). Auckland: University of Auckland.Trkov, D. & Lipej, L. (2019) A non‐destructive method for assessing the feeding habits of coastal fish. Mediterranean Marine Science, 20(2), 453–459. https://doi.org/10.12681/mms.20234Trottier, O. & Jeffs, A.G. (2015a) Recruitment of the parasitic pea crab Nepinnotheres novaezelandiae into green‐lipped mussels Perna canaliculus. Diseases of Aquatic Organisms, 112(3), 199–205. https://doi.org/10.3354/dao02809Trottier, O. & Jeffs, A.G. (2015b) Mate locating and access behaviour of the parasitic pea crab, Nepinnotheres novaezelandiae, an important parasite of the mussel Perna canaliculus. Parasite, 22, 13. https://doi.org/10.1051/parasite/2015013Trottier, O., Walker, D. & Jeffs, A.G. (2012) Impact of the parasitic pea crab Pinnotheres novaezelandiae on aquacultured New Zealand green‐lipped mussels, Perna canaliculus. Aquaculture, 344–349, 23–28. https://doi.org/10.1016/j.aquaculture.2012.02.031Tsuyuki, A. & Umino, T. (2017) Spatial movement of black sea bream Acanthopagrus schlegelii around the oyster farming area in Hiroshima Bay, Japan. Fisheries Science, 83(2), 235–244. https://doi.org/10.1007/s12562‐016‐1058‐9Udy, J., Wing, S., O'Connell‐Milne, S., Kolodzey, S., McMullin, R., Durante, L. et al. (2019) Organic matter derived from kelp supports a large proportion of biomass in temperate rocky reef fish communities: implications for ecosystem‐based management. Aquatic Conservation, 29(9), 1503–1519. https://doi.org/10.1002/aqc.3101Underwood, L.H. & Jeffs, A.G. (2023) Settlement and recruitment of fish in mussel farms. Aquaculture Environment Interactions, 15, 85–100. https://doi.org/10.3354/aei00454Underwood, L.H. (2023) Habitat value of green‐lipped mussel (Perna canaliculus) farms for local fish populations in northern New Zealand (Doctoral dissertation). Auckland: University of Auckland.Usmar, N. (2012) Ontogenetic diet shifts in snapper (Pagrus auratus: Sparidae) within a New Zealand estuary. New Zealand Journal of Marine and Freshwater Research, 46(1), 31–46. https://doi.org/10.1080/00288330.2011.587824van der Reis, A.L & Lavery, S.D. (2020) Marine DNA metabarcoding. In: Goldstein, M. I. & Della Salla, D. A. (Eds.) Encyclopedia of the world's biomes, vol. 4. Amsterdam: Elsevier, pp. 612–618.van der Reis, A.L., Beckley, L.E., Olivar, M.P. & Jeffs, A.G. (2023) Nanopore short‐read sequencing: a quick, cost‐effective and accurate method for DNA metabarcoding. Environmental DNA, 5(2), 282–296. https://doi.org/10.1002/edn3.374van der Reis, A., Jeffs, A. & Lavery, S. (2020) From feeding habits to food webs: exploring the diet of an opportunistic benthic generalist. Marine Ecology Progress Series, 655, 107–121. https://doi.org/10.3354/meps13511Wilding, T.A. & Nickell, T.D. (2013) Changes in benthos associated with mussel (Mytilus edulis L.) farms on the west‐coast of Scotland. PLoS ONE, 8(7), e68313. https://doi.org/10.1371/journal.pone.0068313Wilkens, S.L. & Ahyong, S.T. (2015) Coastal crabs: a guide to the crabs of New Zealand (Version 1). Wellington, New Zealand: The National Institute of Water and Atmospheric Research (NIWA).Woods, C., Floerl, O. & Hayden, B. (2012) Biofouling on Greenshell™ mussel (Perna canaliculus) farms: a preliminary assessment and potential implications for sustainable aquaculture practices. Aquaculture International, 20(3), 537–557. https://doi.org/10.1007/s10499‐011‐9484‐2Woods, C., Williams, R. & Heasman, K. (2014) First record of the caprellid amphipod Caprella andreae Mayer, 1890 (Crustacea, Amphipoda, Caprellidae) from New Zealand. BioInvasions Records, 3(2), 97–102.Wootton, R.J. (2012) Ecology of teleost fishes. Dordrecht: Springer. https://doi.org/10.1007/978‐94‐009‐0829‐1Zazzaro, D., Ruggiero, K. & Jeffs, A. (2018) Use of extract from adults of the triangle barnacle, Balanus trigonus, for reducing fouling in mussel farms. Aquaculture, 483, 223–229.

Journal

Aquaculture Fish and Fisheries – Wiley

Published: Jun 1, 2023

Keywords: aquaculture; fish nutrition; habitat

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

Diet of snapper (Chrysophrys auratus) in green‐lipped mussel farms and adjacent soft‐sediment habitats

References (63)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies