Stable isotope analysis spills the beans about spatial variance in trophic structure in a fish host - parasite system from the Vaal River System, South Africa.

Stable isotope analysis offers a unique tool for comparing trophic interactions and food web architecture in ecosystems based on analysis of stable isotope ratios of carbon (13C/12C) and nitrogen (15N/14N) in organisms. Clarias gariepinus were collected from six sites along the Vaal River, South Africa and were assessed for ectoparasites and endoparasites. Lamproglena clariae (Copepoda), Tetracampos ciliotheca and Proteocephalus glanduligerus (Cestoda), and larval Contracaecum sp. (Nematoda) were collected from the gills, intestine and mesenteries, respectively. Signatures of δ13C and δ15N were analysed in host muscle tissue and parasites using bulk stable isotope analysis. Variable stable isotope enrichment between parasites and host were observed; L. clariae and the host shared similar δ15N signatures and endoparasites being depleted in δ13C and δ15N relative to the host. Differences in stable isotope enrichment between parasites could be related to the feeding strategy of each parasite species collected. Geographic and spatial differences in enrichment of stable isotopes observed in hosts were mirrored by parasites. As parasites rely on a single host for meeting their nutritional demands, stable isotope variability in parasites relates to the dietary differences of host organisms and therefore variations in baseline stable isotope signatures of food items consumed by hosts.

Introduction

Since the late 1970's analysis of natural levels of stable isotopes of nitrogen (15N/14N) and carbon (13C/12C) have been used as a unique fingerprint for studying trophic relationships of organisms (Kelly, 2000; Peterson and Fry, 1987; Post, 2002). Differences in trophic levels between organisms have been assessed through comparison of the nitrogen stable isotope ratio (δ15N), with consumers being enriched in 15N by an average of 3.4‰ per trophic level (Minagawa and Wada, 1984; Vander Zanden et al., 1997). Information regarding the food sources incorporated into the diet by consumers is delineated by comparing enrichment of the stable carbon isotope (Post, 2002). Differences in carbon isotope signatures between sources of food and consumers are slight and only account for 1–2‰ (DeNiro and Epstein, 1978; Kelly, 2000). Stable isotope analysis (SIA) provides a useful means for assessing complex ecological interactions by tracing energy flow through communities (Post, 2002; Wada, 2009). Study of trophic interactions between organisms using stable isotopes has mostly been applied for free–living organisms, whereas, comparisons incorporating macroparasites have lagged behind (Nachev et al., 2017; Sures et al., 2019).

In ecosystems, parasites are functionally important in shaping and stabilising the structure of food webs (Marcogliese, 2003; Nachev et al., 2017; Poulin, 2010; Sabadel et al., 2019). The functional importance of parasites is represented by their ubiquity in ecosystems and proportion of the global biomass (Dobson et al., 2008; Kuris et al., 2008; Selbach et al., 2020), and their ability to regulate host populations makes parasites important in determining host community structure (Marcogliese and Cone, 1997). Parasites are in turn also affected by the structure of food webs (Lafferty et al., 2008) as many parasites rely on trophic interactions between organisms for their transmission (Marcogliese and Cone, 1997). Notwithstanding the functional importance of parasites in ecosystems, they have received less attention in food web studies and as a result our understanding of the ecology of these organisms is limited (Nachev et al., 2017; Thompson et al., 2005). Until recently, due to the fact that parasites derive nutrition from a host organism, the host – parasite relationship has been likened to a predator–prey association (Lafferty et al., 2008; Marcogliese and Cone, 1997; Nachev et al., 2017; Poulin, 2010). According to this premise, parasites should occupy at least a single trophic level above their host in a manner similar to how consumer organisms are trophically distinct from their food sources (Lafferty et al., 2008). However, unlike a typical predator which feeds on multiple prey organisms, parasites only derive nourishment from a single host at a time in their life cycle (Lafferty et al., 2008). Additionally, parasites have developed different feeding strategies which encompass active feeding on host tissues, assimilation of nutrients derived from the metabolism of the host and sharing of resources with the host (Goater et al., 2014; Nachev et al., 2017). Variable isotope fractionation may further serve as an indication of selective feeding strategies developed by parasites (Xu et al., 2007).

Generally, parasites which are more enriched than the host actively consume host tissue resulting in isotope enrichment reminiscent of a predator feeding on a prey organism (Sures et al., 2019). While those which are generally 15N depleted feed by assimilating metabolic by-products present in intestinal contents of the host (Deudero et al., 2002; Lafferty et al., 2008). As a result of unique feeding strategies adopted by parasites, patterns of stable isotope enrichment have similarly been shown to be variable between parasite taxa. For instance, cestodes (Behrmann-Godel and Yohannes, 2015; Boag et al., 1998; Deudero et al., 2002; McGrew et al., 2015; Navarro et al., 2014; Neilson et al., 2005; Persson et al., 2007; Pinnegar et al., 2001; Power and Klein, 2004), acanthocephalans (Nachev et al., 2017; Zhang et al., 2018) and trematodes (Doi et al., 2010; Dubois et al., 2009; Iken et al., 2001; Zhang et al., 2018) are generally depleted in the 15N isotope compared to their hosts, while some nematodes (Boag et al., 1998; Deudero et al., 2002; Neilson et al., 2005; O'Grady and Dearing, 2006), monogeneans (Sures et al., 2019) and ticks (Schmidt et al., 2011) are 15N enriched relative to their hosts. However, the pattern of stable isotope fraction between hosts and parasites is not clear cut, as in some cases stable isotope fractionation has been found to vary between related parasite taxa infecting the same host (Demopoulos and Sikkel, 2015) or between the same parasite taxa infecting different hosts (Deudero et al., 2002). Studies have also found that stable isotope enrichment is affected geographically (Gómez-Díaz and González-Solís, 2010). Along with the differences in stable isotope fractionation being related to the feeding strategies of parasites, observed variances may result from high selectivity for specific microhabitats adopted by many parasite taxa.

In South Africa, to date, two studies have been performed to compare the trophic relationship between parasites and their host fish from the Vaal Dam (Gilbert et al., 2020; Sures et al., 2019). Sures et al. (2019) found that the monogenean, Paradiplozoon ichthyoxanthon, was enriched in 15N isotope relative to the host fishes, Labeobarbus kimberleyensis and Labeobarbus aeneus. Following on, Gilbert et al. (2020) showed that the cestode, Schyzocotyle acheilognathi, also found infecting the L. kimberleyensis, was depleted in the heavier nitrogen isotope (15N) relative to the host fish. From both studies, the differences in isotopic enrichment of 15N accounted for a difference of two trophic levels above and below the host fish respectively. In the present study, stable isotope fractionation and enrichment was compared in a host–parasite model where a single fish host is infected by different parasite taxa.

We hypothesize that within the Clarias gariepinus – parasite system there will be variability in the enrichment factors of ectoparasites and endoparasites in relation to the host fish; where ectoparasites will be enriched in 15N relative to the host and the opposite will be observed for endoparasites. The variability in isotope enrichment in parasites will be related to differences in feeding patterns and mechanisms parasites use to assimilate nutrients from the host fish. Further to this, spatial geographic variability in the host fish stable isotope levels will be mirrored in parasites and given the expansiveness of the Vaal River we expect to find that 15N and 13C levels in hosts and parasites will show similar geographic patterns. To test these hypotheses, C. gariepinus were collected from six different sites along the Vaal River in South Africa and muscle tissue and parasites were analysed for stable isotopes of carbon and nitrogen. Clarias gariepinus is a known omnivore and has been found to feed on a wide variety of prey items ranging from plant material to invertebrates and small vertebrates in aquatic environments (Skelton, 2001). This varied diet has further been suggested as a factor leading to the diversity of parasites found infecting this fish species (Crafford and Avenant-Oldewage, 2009). In light of this, we therefore further hypothesize that variability in stable isotope levels in C. gariepinus will be related to the size of the fish, with larger fish being enriched in 15N isotope compared to smaller sized fish.

Methods

Host and parasite collection

A total of 49 C. gariepinus were collected from six sites along the Vaal River, South Africa (Fig. 1). The Vaal River is the third largest river in South Africa and is an economically and ecologically important aquatic ecosystem in South Africa due to its importance in providing water for agricultural, urban and industrial activities occurring within the catchment area (Braune and Rogers, 1987). The river originates in the Mpumalanga Province and flows in a south-west direction across the interior of the country and eventually joins the Orange River in the Northern Cape Province near the town of Douglas (Braune and Rogers, 1987). The Vaal River, being 1120 km long, is the largest tributary of the Orange River with six major impoundments. Our sampling sites represent three of the available impoundments (Vaal Dam (26°53′40.3″S 28°08′44.1″E), Bloemhof Dam (27°40′42.3″S 25°40′25.2″E) and Douglas Weir (29°01′23.9″S 23°53′09.7″E)) and three riverine sites (Vaal River below Grootdraai Dam (26°55′17.9″S 29°16′52.5″E), Vaal River below the Vaal Barrage (26°44′23.8″S 27°38′08.9″E) and the Vaal River below Vaalharts Weir (28°05′56.9″S 24°51′12.8″E)). The upper most site was located below the Grootdraai Dam in the Mpumalanga Province and the last site was located in the Douglas Weir before the confluence of the Vaal River and Orange River in the Northern Cape Province. During the surveys at each site, C. gariepinus were caught using gill nets (mesh size: 45–190 mm) and transported in 160 L plastic containers filled with aerated water from each site to a field laboratory. Collections of fish were performed in accordance with permits from relevant national government organisations (Mpumalanga Tourism and Parks Agency: MPB. 5555; Gauteng Department of Agriculture and Rural Development: CPE000125; the Department of Economic Development, Tourism and Environmental Affairs: 01/34287; Northern Cape Nature Conservation: FAUNA 1120/2016) and following approval from the Ethics Committee of the University of Johannesburg (reference number: 2016-5-03). Fish were then euthanized by severing the spine posterior to the head. Weight and total length of C. gariepinus were recorded for determination of condition factor (K) for comparison of the nutritional status between host individuals from the same site and between sampling locations (Heincke, 1908):where the weight of fish in kilograms (kg) and total length of the fish in centimetres (cm). Following collection of morphometric data the fish were euthanized and dissected, the intestines, mesenteries and gills were removed and assessed for parasites with a stereo microscope. Muscle tissue of the host was also collected using a ceramic knife and plastic forceps, and along with parasites, were frozen at −20 °C before returning to the laboratory. Parasites collected included Lamproglena clariae (Copepoda) from the gills, Tetracampos ciliotheca (Cestoda) and Proteocephalus glanduligerus (Cestoda) from the intestine, and larval Contracaecum sp. (Nematoda) which were found encysted in the mesenteries of the intestine.

Fig. 1

Map of the Vaal River showing the position of sampling sites (I: below Grootdraai Dam; II: Vaal Dam; III: below Vaal River Barrage; IV: Bloemhof Dam; V: below Vaalharts Weir; VI: Douglas Weir) along the Vaal River. The block (B) indicates the position of the Vaal River within South Africa and insert A indicates the position of South Africa shaded on the African continent.

Stable isotope analysis

In the laboratory, parasites were defrosted and cleaned in fresh saline (0.092% w/v) to remove any host tissue and debris from the microhabitat. In the case of L. clariae, egg strings were removed from adult females. The parasites were then refrozen (−20 °C) and along with host tissue, dried to weight consistency in a freeze dryer (Martin-Christ Gefriertrocknungsanlagen, GmbH; Germany). Host muscle and parasites tissue were homogenised and triplicates of each sample were weighed (0.4–0.8 mg) into 4 × 6 mm tin capsules for analysis of stable isotopes of carbon (13C/12C) and nitrogen (15N/14N). Analysis of samples was performed following procedures described by Nachev et al. (2017) in C/N mode using a Vario PYRO Cube elemental analyser (EA) system (Elementar Analysensysteme, Langenselbold, Germany) coupled with to an IsoPrime 100 isotope ratio mass spectrometer (IRMS; Elementar Analysensysteme). All isotope ratios were reported in δ-notation as differences in isotopic proportion of the sample and internal reference standard (acetanilide AcAn) by equation (2).

Ratios of 13C/12C and 15N/14N isotopes are expressed as in the reference material and in the sample. Normalisation of samples and working standard, acetanilide, was done against the international scale, USGS40 and USGS41 reference materials (International Atomic Energy Agency). Normalized delta values for both stable isotopes were reported against the VPDB scale for carbon and Air for nitrogen.

Statistical analysis

Differences in trophic level (ΔTL) of hosts and parasites were assessed using isotope levels for host muscle tissue and parasites according to following equation (3):

Trophic level differences were determined using the average trophic enrichment factor (TEF = 3.4‰) by Minagawa and Wada (1984). To compare trophic enrichment of parasites, the same equation was applied.

Statistical analyses of the data was performed using SPSS version 25 for Windows (Statistical Package for the Social Sciences, SPSS Inc., USA). The normality of the data was assessed using the Shapiro-Wilk test and as data were not normally distributed all comparisons were performed using non-parametric tests. To compare the stable isotope levels between hosts and parasites the Kruskal-Wallis test was applied. In cases where differences in δ15N and δ13C values were significantly different the Wilcoxon matched pair test was used between hosts and parasite taxa. Differences in stable isotope enrichment in hosts and parasites between different sites were analysed using the Kruskal-Wallis test and where differences were significant the Mann-Whitney test was applied as a post hoc test to the Kruskal-Wallis test. Mean intensity of parasites infecting C. gariepinus was calculated according to Bush et al. (1997). Comparisons of spatial variation in host morphometrics and parasite mean intensity were performed using the Kruskal-Wallis test. To assess the magnitude of the effect of host size on the isotope enrichment and parasite intensity, and host isotope enrichment on parasite intensity the Kendall's’ Tau test was used.

Results

Enrichment of stable isotopes in host fish and parasites, parasite mean intensities and host length and weight varied among sampling sites in the Vaal River (Table 1). There were significant differences in the composition of δ15N (Kruskal–Wallis, df = 5, Z = 21.54, P = 0.001) and δ13C (Kruskal–Wallis, Z = 28.96, df = 5, P < 0.001) among the C. gariepinus samples along the Vaal River. Host fish collected from the Vaal River Barrage were most enriched with 15N compared with fish from the other sites, whereas, hosts collected below the Grootdraai Dam had lowest δ15N values. Fish collected from Douglas Weir showed highest depletion in 13C and those collected from Bloemhof Dam (δ13C = 18.49‰) and below Grootdraai Dam (δ13C = 18.74‰) were the most enriched in the 13C isotope. To better understand the observed geographic differences in stable isotope composition in C. gariepinus hosts; host size, weight and condition factor were compared as a means of determining if dietary differences were a contributing factor. Host fish size (total length) and weight were significantly different between different sites (Kruskal-Wallis H = 11.5, df = 5, P = 0.042), while the condition factor showed no differences between sites (Kruskal-Wallis H = 8.44, df = 5, P = 0.134) or host sex (Mann-Whitney U = 101, P = 0.220). Comparison of host size (length and weight) showed that smaller hosts were less enriched in 15N than larger hosts and the opposite was observed for δ13C which was higher in smaller hosts than larger ones. Nitrogen stable isotope enrichment in C. gariepinus was positively but not significantly correlated with total length (Kendall's tau, τ = 0.135, P = 0.247) and weight (Kendall's tau, τ = 0.214, P = 0.071) of host fish. Correlations between host length and weight, and 13C enrichment were significant and negative (total length: Kendall's tau, τ = −0.361, P = 0.002; weight: Kendall's tau, τ = −0.382, P = 0.001).

Table 1

Stable isotope composition, host morphometry and mean intensity (±SE) of parasites for selected host – parasite system for six sampling sites along the Vaal River.

Site		Number of hosts	TL [cm]	W [kg]	K	Parasite mean intensity (±SE)	δ15N	δ13C	ΔhE (parasite – host)
									Δ15N	Δ13C
Vaal River below Grootdraai Dam	Clarias gariepinus	5	71.7(±26.5)	2.92(±2.76)	0.01(±0.01)		14.1(±2.66)	−18.74(±0.76)
	Lamproglena clariae (adult)					6.67(±3.15)	16.28(±1.05)	−18.76(±1.17)	2.18	0.02
	Lamproglena clariae (egg)						16.38(±0.64)	−19.73(±1.29)	0.1a	0.97a
	Tetracampos ciliotheca					2.00(±2.00)	13.19(±0.00)	−19.08(±0.00)	−0.91	0.34
	Contracaecum sp. (larvae)					6.00(±1.00)	15.79(±0.06)	−19.27(±0.92)	1.69	0.53
Vaal Dam	Clarias gariepinus	10	53.8(±21.7)	1.36(±1.43)	0.01(±0.01)		15.77(±2.06)	−19.97(±0.82)
	Lamproglena clariae (adult)					8.57(±4.47)	16.35(±2.58)	−19.82(±0.51)	0.58	−0.15
	Lamproglena clariae (egg)						16.49(±2.58)	−20.21(±0.58)	0.14a	0.39a
	Tetracampos ciliotheca					3.00(±0.00)	13.49(±0.27)	−19.94(±2.94)	−2.28	−0.03
Vaal River below Vaal River Barrage	Clarias gariepinus	10	61.9(±13.4)	1.99(±1.29)	0.01(±0.01)		24.16(±1.01)	−23.19(±1.15)
	Lamproglena clariae (adult)					3.80(±1.13)	23.59(±1.33)	−23.46(±1.86)	−0.57	0.27
	Lamproglena clariae (egg)						23.58(±1.55)	−23.63(±2.32)	−0.01a	0.17a
	Tetracampos ciliotheca					2.50(±0.681)	14.36(±1.70)	−26.86(±2.78)	−9.8	3.67
Bloemhof Dam	Clarias gariepinus	11	50.3(±17.1)	1.20(±1.60)	0.01(±0.01)		16.24(±1.16)	−18.49(±0.64)
	Lamproglena clariae (adult)					3.91(±1.96)	15.28(±0.00)	−17.42(±0.00)	−0.96	−1.1
	Tetracampos ciliotheca					10.3(±6.60)	13.15(±1.26)	−20.69(±0.64)	−3.09	2.2
Vaal River below Vaalharts Weir	Clarias gariepinus	8	53.03(±9.53)	0.89(±0.36)	0.01(±0.01)		15.65(±0.37)	−20.85(±1.51)
	Lamproglena clariae (adult)					4.29(±1.44)	16.91(±0.00)	−19.45(±0.00)	1.26	−1.4
	Lamproglena clariae (egg)						21.8(±0.00)	−22.2(±0.00)	4.89a	2.75a
	Tetracampos ciliotheca					1.50(±0.50)	11.77(±0.01)	−20.04(±1.77)	−3.88	−0.81
	Proteocephalus glanduligerus					1.00(±0.00)	12.14(±0.00)	−18.38(±0.00)	−3.51	−2.47
	Contracaecum sp. (larvae)					13.0(±3.49)	12.66(±0.48)	−20.64(±1.35)	−2.99	−0.21
Douglas Weir	Clarias gariepinus	5	73.5(±4.34)	2.67(±0.68)	0.01(±0.01)		15.71(±0.38)	−24.57(±0.28)
	Tetracampos ciliotheca					4.67(±2.73)	12.2(±0.06)	−28.09(±1.29)	−3.51	3.52
	Proteocephalus glanduligerus					3.00(±0.00)	13.35(±0.00)	−26.33(±0.00)	−2.36	1.76
	Contracaecum sp. (larvae)					29.0(±3.62)	11.41(±1.07)	−22.47(±1.45)	−4.3	−2.1

Delta values for15N and13C calculated from difference in isotope signatures between eggs and adults of L. clariae.

Parasite mean intensity was variable between the different sampling sites. Highest mean intensity of L. clariae was recorded at the Vaal Dam. Mean intensities of T. ciliotheca and Contracaecum sp. were highest at Bloemhof Dam and Douglas Weir respectively. Proteocephalus glanduligerus were only collected from fish below Vaalharts Weir and Douglas Weir where, in both instances, only a single cestode was recovered from either site. Intensity of parasites infecting C. gariepinus were compared with host length and weight, and showed intensity for all taxa collected was positively related with total length and weight of the host fish. However, only the intensity of T. ciliotheca correlated significantly with host size (Kendall's tau, length: τ = 0.297, P = 0.019; weight: τ = 0.324, P = 0.012).

Similarly to host fish, 15N and 13C were variably enriched and depleted in the parasite taxa collected in the present study and the stable isotope composition in parasites reflected those of the host fish along the Vaal River. Generally, endoparasites were depleted in 15N and enriched in 13C compared to C. gariepinus hosts. For cestodes, T. ciliotheca Δ15N varied from 0.91‰ below the Grootdraai Dam to 9.8‰ at the site below the Vaal Barrage and in P. glanduligerus Δ15N ranged from 2.36‰ to 3.51‰ at Douglas Weir and below Vaalharts Weir, respectively. Contracaecum sp. larvae collected below the Grootdraai Dam were enriched in 15N by 1.69‰ relative to the host fish, whereas, specimens collected below Vaalharts Weir and in Douglas Weir were depleted in 15N compared to the host. For 13C, enrichment patterns mostly showed that cestodes and Contracaecum sp. were enriched relative to C. gariepinus hosts. Except for T. ciliotheca collected in the Vaal Dam and below Vaalharts Weir; P. glanduligerus from below Vaalharts Weir and Contracaecum sp. collected from below Vaalharts Weir and Douglas Weir sites which were depleted in 13C compared with the hosts.

Adult female L. clariae were variably enriched and depleted in both 15N and 13C isotopes. In instances where adult female L. clariae were enriched in 15N isotope, differences accounted for 0.58–2.18‰ higher than in host muscle tissue. Samples from below the Vaal Barrage and Bloemhof Dam were depleted in 15N by 0.57‰ and 0.96‰ respectively. Differences in 15N stable isotope indicate that despite the mixed fractionation patterns, adult L. clariae and C. gariepinus hosts share similar trophic levels. For differences in 13C isotope, L. clariae were depleted at Vaal Dam, Bloemhof Dam and below Vaalharts Weir by 0.15–1.4‰, whereas, adult copepods collected from below the Vaal Barrage were enriched by 0.27‰. Differences in nitrogen and carbon stable isotopes between adult female copepods and hosts were not significant (δ15N: Wilcoxon matched pair test, Z = −0.314, P = 0.754; δ13C: Wilcoxon matched pair test, Z = −1.098, P = 0.272). Eggs of adult L. clariae were analysed and compared with isotope signatures in adult parasites as larval stages derive nutrients from the adult organism and not from the host fish. Stable isotope fractionation patterns between adult L. clariae and egg strings were variable; with Δ15N for egg strings being enriched from 0.1‰ below Grootdraai Dam to 4.89‰ below Vaalharts Weir or shared similar signatures for 15N signatures relative to adult parasites as in the case of samples collected below the Vaal River Barrage. Differences in δ15N between egg strings and adult L. clariae were not significant (Wilcoxon matched pair test, Z = −0.459, P = 0.646). For the Δ13C, egg strings were consistently enriched compared to adult L. clariae by 0.17–2.75‰.

Geographic differences in 15N enrichment in parasites were not significant for adults (Kruskal-Wallis Z = 8.5, df = 4, P = 0.075) and eggs (Kruskal-Wallis Z = 7.36, df = 3, P = 0.061) of L. clariae, and T. ciliotheca (Kruskal-Wallis Z = 10.1, df = 5, P = 0.073). For Contracaecum sp. differences in 15N were significant (Kruskal-Wallis Z = 7.12, df = 2, P = 0.028) among the sampling sites. In the case of P. glanduligerus, the low number of specimens collected (n = 2) hampered any statistical comparison. The specimen collected at Douglas Weir showed higher enrichment for both stable isotopes analysed compared with the specimen collected below Vaalharts Weir. In terms of enrichment of 13C, geographic differences were only significant for T. ciliotheca (Kruskal-Wallis Z = 14.4, df = 5, P = 0.013). Differences among L. clariae adults, eggs and Contracaecum sp. larvae were not significant.

Discussion

In the present study distinct spatial-geographic differences in stable isotope fractionation patterns were observed for C. gariepinus and some parasites from the Vaal River. In host fish, highest 15N levels were recorded from samples collected below the Vaal River Barrage, whereas, samples collected from the site below Grootdraai Dam showed lowest stable nitrogen isotope levels. The opposite was observed for 13C enrichment in hosts among the different sites, with lowest levels observed in C. gariepinus collected below the Vaal River Barrage. These differences may be related to the variability in the diet of C. gariepinus as a result of availability of prey items, behavioural differences of C. gariepinus and ecological differences at each sampling site along the Vaal River. Clarias gariepinus is a typical omnivore, feeding on a wide variety of organic matter with prey items varying from birds, reptiles, other fish species, including smaller C. gariepinus, to macroinvertebrates and plant material (Skelton, 2001; Willoughby and Tweddle, 1978). Willoughby and Tweddle (1978) demonstrated through analysis of the stomach contents of C. gariepinus from Shire Valley, Malawi, were highly variable regardless of whether or not C. gariepinus were inhabiting the same and different environments. As stable carbon isotope ratios serve as a representation of the nutrient sources in food webs (Fry and Sherr, 1984), comparison of differences in the enrichment of 13C isotope can be considered a good indicator of spatial variation in baseline isotope levels (Gómez-Díaz and González-Solís, 2010; Hobson, 2005). The variability of δ13C in muscle tissue of C. gariepinus between sampling sites suggests that a wide range of dietary items are consumed by this fish species. In addition to a highly diverse diet, C. gariepinus have been found to undergo a shift in diet as they grow (Kadye and Booth, 2012; Willoughby and Tweddle, 1978). The correlations observed in C. gariepinus stable isotope levels and host size (length and weight) support a shift in the foraging habits and food items consumed by this fish species with growth. In addition to dietary variation with growth, Kadye and Booth (2012) showed that there is a high dietary overlap between fish of different size classes which can be related to the omnivorous feeding strategy of this fish species. Clarias gariepinus serves as a host for a wide variety of endoparasite and ectoparasite taxa (Crafford and Avenant-Oldewage, 2009) and this could likely result from the variable diet and wide distribution of this fish species. Comparison between host size (length and weight) and parasite intensity indicate that larger fish harbour more parasites than smaller ones. This concept has been well documented in previous studies (see Patterson and Ruckstuhl, 2013).

Isotopic discrimination of 15N indicated that all nematode larvae and cestodes collected from C. gariepinus were depleted relative to the host fish from all sites. Depletion of the 15N stable isotope in cestodes and larval nematodes is in line with previous findings for other cestodes and larval nematodes (Iken et al., 2001; Pinnegar et al., 2001) infecting fish hosts, and for some cestodes infecting rabbit hosts (Boag et al., 1998). Differences in δ15N between larval nematodes and cestodes with host muscle tissue accounted for shifts of approximately one to two trophic levels, respectively. These observed differences are in the range for other host – cestode and larval nematode systems analysed (Boag et al., 1998; McGrew et al., 2015; Navarro et al., 2014; Persson et al., 2007; Pinnegar et al., 2001; Power and Klein, 2004) and can be related to the mode of nutrient acquisition. For cestodes, lack of a digestive system has meant that the tegument of these organisms has become specially modified for the accumulation of molecules derived from the hosts' metabolism (Smyth and McManus, 1989). As a result of transamination of complex proteins, the 15N stable isotope is retained in the tissues of the host (Macko et al., 1986) and as a result endoparasites are depleted in heavier isotopes.

Results for isotopic discrimination between host muscle tissue and Contracaecum sp. larvae in the present study are in line with previous studies for other larval nematodes encysted in the peritoneal cavity of host fish (Deudero et al., 2002; Nachev et al., 2017). In the case of larvae of Contracaecum sp. infecting C. gariepinus, nematodes were encysted in the mesenteries of the viscera in the region of the intestinal tract. According to Moravec et al. (2016) these larval stages exhibit low pathogenicity and along with the current 13C and 15N isotope data indicate that the nematodes are not actively feeding on fish hosts. Rather Moravec et al. (2016) suggests that these larval nematodes develop to the third stage in the egg in the water column which is then ingested by a fish which functions as a paratenic host. Thus, it is plausible that the lack of a relationship between host stable isotopes and larval nematodes is the result of C. gariepinus being a paratenic host to Contracaecum sp. larvae which are not actively feeding on or assimilating nutrients from the host. Additionally, Nachev et al. (2017) indicated that the slow growth rate of some nematode larvae is a contributing factor resulting in the lack of a relationship between host stable isotopes and larval nematodes. In order to determine if these larval stages derive some nutrition from paratenic hosts, comparisons of stable isotopes for different larval stages will have to be performed. Stable isotope fractionation observed in the present study indicated that endoparasite taxa infecting C. gariepinus occupy relatively similar trophic levels. These findings are comparable to other reports for similar host – parasite systems (Behrmann-Godel and Yohannes, 2015; Deudero et al., 2002; Power and Klein, 2004) and indicate that nutrients assimilated by the host and parasites are from similar sources (Power and Klein, 2004).

Regarding the copepod, L. clariae, fraction of both 15N and 13C showed that the parasites were variably enriched and depleted in both stable isotopes compared with host muscle tissue. Based on Δ15N values this did not correspond to differences in trophic position and therefore host and female copepods likely occupy similar trophic positions. Instances where adult female L. clariae were 15N enriched relative to host muscle tissue correlates with the feeding pattern of the parasite and corroborates stable 15N isotope fractionation patterns in other haemophagous ectoparasites (Boag et al., 1998; Doucett et al., 1999; Schmidt et al., 2011; Sures et al., 2019; Voigt and Kelm, 2006). The variability in 15N fractionation between C. gariepinus and L. clariae observed in the present study further corroborates findings for other parasitic copepods, which parasitise the gills of the host fishes and share similar feeding biology as L. clariae (Deudero et al., 2002; Iken et al., 2001; Pinnegar et al., 2001). Iken et al. (2001) found that a copepod infecting gills of Coryphaenoides armatus was enriched by 2.7‰ in 15N relative to the host. Pinnegar et al. (2001) found that Lernaeocera branchialis, infecting the gills of the host fish, Platichthys flesus, were depleted in 15N stable isotope by 0.81‰ relative to the host. Unlike L. clariae which feeds on blood from the gill filaments, L. branchialis feeds on blood from the bulbous arteriosus of the heart of infected fish hosts, where it macerates host tissue with its mandibles and feeds on host blood (Brooker et al., 2007). In an assessment of isotope fractionation between several copepods infecting the gills of their hosts, Deudero et al. (2002) showed high interspecific variation in 15N enrichment between copepod species feeding on the same host fish. For instance, they found that Clavella adunca feeding on both cod and whiting were depleted in 15N stable isotope but L. branchialis feeding on whiting or haddock were enriched in 15N but were depleted in 15N when parasitising cod. Similarly with L. clariae, the copepods analysed by Pinnegar et al. (2001) and Deudero et al. (2002) infected the gills and feed on the blood of the host fishes. However, Shotter (1971) indicated that the mandibles of Clavella uncinata, a similar species to C. adunca, were too weak to tear tissue but instead function in gathering material by scraping superficial tissue toward the mouth, with little blood comprising the diet and intestinal contents. In some instances, studies have indicated adult female copepods are significantly enriched in 15N relative to the host organism (Baud et al., 2004; Goedknegt et al., 2018; Gretsy and Quarmby, 1991). Gretsy and Qyarmby (1991) found that adult Mytilicola intestinalis were enriched by 3‰ relative to the intestine of European blue mussel host (Mytilus edulis). Goedknegt et al. (2018) similarly found that adult Mytilicola orientalis were enriched in 15N stable isotope relative to the adductor muscle of the host mussel, M. edulis, by 1.2‰. In both instances the higher 15N enrichment of both species could be related to the parasites feeding directly on the intestinal tissue of the host. In a seasonal study on the gut ultrastructure and contents in conjunction with stable carbon and nitrogen isotope analysis of Neoergasilus japonicus, Baud et al. (2004) showed that adult female parasites infecting the fins of the host fish, Perca fluviatilis, were enriched in the 15N stable isotope by 3.7‰ relative to the host muscle tissue, indicating that the parasite feeds on host tissue. An additional aspect to consider for parasitic copepods is the effect of changes in the life style of these parasites from juvenile stages to adults. During their life cycles many parasitic Copepoda undergo drastic morphological changes from free swimming larval stages which are able to move between hosts to sedentary parasitic adults. It is possible that during the free swimming larval stages, copepodite stages may feed on hosts of variable isotope enrichment as well as incorporate other sources of nutrition before becoming parasitic on a host fish. The variation in the diet during the free swimming larval stages may then further result in variations in stable isotope enrichment observed in adult organisms.

Comparison between the parasites infecting C. gariepinus showed that L. clariae were significantly and constantly enriched in 15N stable isotope compared to the endoparasites. The shift in δ15N observed for endoparasites and L. clariae overall accounted for a difference of approximately two trophic levels and likely relates to differences in nutrient acquisition strategies by each parasite taxon analysed. Variability in the signatures reported between the different parasite taxa analysed are similar to other studies which found similar co-infections (Deudero et al., 2002; Iken et al., 2001; Pinnegar et al., 2001). Mature, adult female L. clariae consume whole blood and epithelial cells which they acquire from the secondary gill filaments of the host fish (Moll and Avenant-Oldewage, 2017; Tsotetsi et al., 2005). Female copepods attach to the gill filaments using their maxillipeds and feed on gill epithelium and blood using the maxillae which scrape cellular material toward the mouth and along with a negative pressure created by the muscular oesophagus, blood and cellular debris are sucked into the buccal cavity (Moll and Avenant-Oldewage, 2017; Molnár et al., 2018; Tsotetsi et al., 2005). Unlike helminth endoparasites, some copepod ectoparasites are not able to accumulate simple amino acids across the keratinised body surface and instead must break down complex proteins which are accumulated in blood meals (Deudero et al., 2002). As a result these parasites are generally enriched in heavier 15N stable isotopes in a manner similar to consumer organisms which resembles a predator–prey relationship. Whereas, in the case of both cestodes and larval nematodes, feeding can be likened to that of a commensalistic scavenger, whereby, the parasites feed on left over by-products of the hosts metabolism and in doing so pose little harm during feeding toward the host.

In the present study, variable stable isotope fractionation patterns of parasites infecting C. gariepinus hosts corresponded or mirrored geographic isotope enrichment of the host among the collection sites along the Vaal River. According to predictions of stable isotope fractionation between consumers and prey items, geographic variation in the ratios of stable isotopes of carbon and nitrogen should be reflected in a predictable manner in relation to the isotope signatures of the host (Caut et al., 2009; DeNiro and Epstein, 1981, 1978). As such variability in the diet of the host should be reflected more prominently in the stable isotope composition of endoparasites than in ectoparasites (Deudero et al., 2002). However, from the results of the present study, variations in the host diet were more closely mirrored in the isotope enrichment of L. clariae than in their endoparasites. Iken et al. (2001) observed no similarity in isotope enrichment in trematodes feeding on intestinal tissue of the host fishes, Chalinura profundicola and Chalinura leptolepis, while a copepod feeding on a gastropod host, Oneirphanta mutabilis, did exhibit similarity in isotope fractionation. Deudero et al. (2002) similarly noted that stomach nematodes did not reflect isotopic differences observed in host fishes. Sures et al. (2019) showed that in a monogenean – host fish system from the Vaal Dam, the micropredatory nature of P. ichthyoxanthon resulted in mirroring the isotope fractionation between two yellowfish hosts.

Spatial geographic differences observed at various trophic levels can be related to host – specific differences in ecology and behaviour (Deudero et al., 2002; Gómez-Díaz and González-Solís, 2010). Geographic differences observed in stable isotope enrichment patterns of parasite taxa in the present study mirrored those of the host and are likely related to differences in food items utilised by the host fish as feeding strategies of the parasites do not change between the different sites. The spatial variability in stable isotopes in parasites infecting C. gariepinus therefore reflects spatial differences in the baseline isotope signatures across the distribution of the host. Geographical differences in δ13C and δ15N levels of organisms has previously been documented (Hobson, 2005). In parasites, Gómez-Díaz and González-Solís (2010) similarly observed spatial differences in the isotopic signatures of ectoparasites infecting two closely related shearwater hosts, Calonectris diomedea and Calonectris borealis, across the Mediterranean Sea and Northeast Atlantic. Low variation in stable isotope composition of parasites collected from the Vaal River can be related to the fact that parasites do not change the resources they utilise from the host and rather spatial differences in parasite stable isotope enrichment is related to the variance in host diet. This was similarly observed by Riascos et al. (2015) for Hyperia curticephala infecting the scyphomedusa, Chrysaora plocamia. It should also be noted that unevenness of the distribution of parasites along the Vaal River likely lead to a lop-sided sampling design and as such spatial differences in parasite stable isotope enrichment observed in the present study should be confirmed following more even sample collection.

Conclusions

Results of the analysis of stable isotopes of carbon and nitrogen in C. gariepinus and ectoparasite and endoparasite taxa collected from the Vaal River are mostly consistent with patterns already described for similar host-parasite systems. General trends in stable isotope fractionation show that the ectoparasite, L. clariae and host shared similar δ15N values, whereas, the endoparasite, T. ciliotheca, P. glanduligerus and Contracaecum sp. were depleted in the 15N isotope compared to their host. These trends were in line with previous studies on similar parasite taxa. Geographic and spatial differences in 15N and 13C isotope levels in host fish and some parasite species reflect differences in baseline isotope signatures across the host distribution. The differences observed in stable isotopes of C. gariepinus tissue further reflected the omnivorous feeding biology of the host fish. In the case of the parasite taxa analysed, variation in 15N and 13C isotope levels mirrored those of the host fish at each site and related to the fact that parasites only derive nutrition from a single source, being the host. For specific parasite taxa, comparison between sites showed a degree of variation in stable isotope levels, which in turn may be related to variable resource partitioning in the hosts as well as difference relating to the feeding biology of the host fish. Unlike their hosts, parasites may reflect more accurately baseline isotope levels for specific sampling sites given that they acquire all nutrition from a single host organism. Along with a shorter lifespan than the host, large changes in isotope levels would likely not be reflected in parasites unlike host fish which could move large distances between sites.