Seasonal variations in food resource partitioning among four sympatric gudgeon species in the upper Yangtze River

Abstract Knowledge of food resource partitioning among sympatric fish species is crucial for understanding the potential mechanisms of species coexistence. Gudgeons (Teleostei: Cyprinidae: Gobioninae) often dominate fish assemblages in the upper Yangtze River. However, little research has been conducted on their trophic interactions. In this paper, seasonal diet and feeding strategy variations of four sympatric gudgeon species, Coreius guichenoti, Coreius heterodon, Rhinogobio ventralis, and Rhinogobio cylindricus, were investigated by analysis of intestinal tract contents, aiming to explore whether food resource partitioning occurred among them. Fish specimens were collected during spring (April–May) and autumn (August–October) in 2010 in Hejiang, a free‐flowing stretch of the upper Yangtze River. Coreius guichenoti, C. heterodon, and R. cylindricus showed omnivorous feeding habits, while R. ventralis exhibited an obligate carnivore feeding habit. Diet overlap among the four studied species was high, especially in spring. However, changes in feeding strategies were observed in autumn. Specifically, C. guichenoti and R. cylindricus expanded their dietary niche breadth and consumed detritus, Sinopotamidae or Hydropsychidae as important complementary food resources. In contrast, C. heterodon and R. ventralis reduced their dietary niche breadth and became more specialized on mussels (Limnoperna lacustris). These results confirmed that sympatric fish species can coexist with high diet overlap, and food resource partitioning among these species may also fluctuate with the seasons.

Noteworthy, food resource partitioning patterns among sympatric fish species may change obviously across seasons, according to the seasonal variations in food availability, including food diversity and food abundance (Prejs & Prejs, 1987;Sánchez-Hernández et al., 2016;Sánchez-Hernández, Gabler, & Amundsen, 2017;Gray et al., 1997). Some researchers found that co-occurring species may specialize following their species-specific resource preferences when food resource become limited and the overlap will decrease (Gabler & Amundsen, 1999;Robinson & Wilson, 1994;Schoener, 1971;Gray et al., 1997). For example, Gray et al. (1997) observed greater trophic partitioning among sympatric fish species in April, when food resource was scarce, than in July, when prey was abundant. Deus and Petrere-Junior (2003) noted that fish species were more generalized in summer when food availability was higher and more specialized in winter when food resource was scarce. In contrast, other researchers insisted that species should be forced to converge and to exploit the same resources when the food density was low (Liem, 1984;Magalhães, 1993;Pyke, Pulliam, & Charnov, 1977;Wiens, 1993). Under this circumstance, the population trophic niche breadth will expand and the overlap will increase (Liem, 1984;Magalhães, 1993;Pyke et al., 1977;Wiens, 1993). Sánchez-Hernández et al. (2017) demonstrated that these seemingly contradictory standpoints can be solved when food diversity is taken into consideration. That is, alternatives to niche differentiation can be used to explain the coexisting of sympatric species (Amarasekare, 2003;Gabler & Amundsen, 2010;Genner, Turner, & Hawkins, 1999;Sánchez-Hernández et al., 2017). However, more studies are recommended to examine what component of food availability (prey diversity and prey abundance) affects food resource partitioning among sympatric fish species (Sánchez-Hernández et al., 2017).
The upper Yangtze River supports the highest biodiversity of the Palearctic region, with 286 fish species distributes in its mainstream and tributaries, and 124 of these species are endemic to this area (He, Wang, Lek, Cao, & Lek-Ang, 2011;Matthews, 1998;Nelson, 1994). However, little is known about the potential mechanism facilitating the coexistence of these sympatric species. In this article, we studied food resource partitioning among four abundant Gudgeons (Teleostei: Cyprinidae: Gobioninae) in the upper Yangtze River: Coreius guichenoti, Coreius heterodon, Rhinogobio ventralis, and Rhinogobio cylindricus (Figure 1). Among them, C. guichenoti, R.
ventralis, and R. cylindricus are endemic to the upper Yangtze River.
All these species show similar morphological (e.g., elongated body, inferior mouth, and small eyes) and ecological characters (e.g., inhabit running waters, bottom-feeding, and release pelagic eggs into stream currents; Zeng & Liu, 2011), which provide a unique opportunity to examine the mechanism facilitating the coexistence of sympatric species with similar niches (Wang, Liu, Lin, Yang, & Liu, 2015). Therefore, the present study aims to (a) examine the possible seasonal changes in diet composition and feeding strategy among these sympatric species and (b) better understand the coexistence phenomenon of sympatric species.

| Study area
The Yangtze River is the largest river in China and the third longest river in the world, with a length of ~6,380 km and a drainage area of 1.8 × 10 6 km 2 (Hydrology Bureau of Changjiang Water Resources Committee, 2003). The present study was conducted along a 30 km stretch nearby the Hejiang County, Sichuan Province, which is ~100 km upstream of the backwater of the Three Gorges Reservoir ( Figure 2). The width of the sampled stretch ranges from 500 to 1,000 m, characterized by a substrate composed of bedrock, boulders, and sand. The maximum water depth measured to ~60 m. The climate is a typical subtropical monsoon, with the air temperature and rainfall vary drastically among seasons. The water temperature usually peaks at ~25.0°C in July or August and drops to ~10.0°C in Results showed that macroinvertebrate community in this stretch was dominated by Perlidae, Hydropsychidae, Gammaridae, and Ephemeroptera. Additionally, the diversity and abundance of macroinvertebrate varied significantly with seasons . In spring, the macroinvertebrate community showed a high diversity and the average density peaked with 165.50 ind/100 m 2 . However, due to the washout of monsoon floods and emergence of some aquatic invertebrates, the diversity of macroinvertebrate reduced significantly in autumn with some groups (e.g., Trichoptera, Diptera, and Neuroptera) almost disappeared from the community, and the average density decreased to only 27.9 ind/100 m 2 . Fish assemblage in this stretch was dominated mainly by the four studied gudgeon species, with the relative biomass reached 60-80% of the total catches collectively (see Liu, Wang, & Cao, 2012 for more details).

| FIS H SAMPLING AND D IE T E X AMINATION
Sampling for dietary analysis was conducted in spring (April-May) and autumn (August-October) in 2010, generally representing the dry season (low water level, abundant food resource) and the wet season (high water level, low food resource) in this area. Fishes were collected using drift gill net, an active fishing gear, with a mesh size of 3-5 cm (100-200 m long × 1.0-2.3 m high). All samplings were conducted in the daytime. Nine sampling sites were distributed along the study reach with a length of 1-2 km, representing all accessible habitat units (Liu et al., 2012). Specimens were measured to determine standard length (to the nearest mm) and body weight (to the nearest g), and dissected immediately in order to reduce postmortem digestion. The foregut (the section of the intestine from the esophagus to the posterior end of the first loop) was removed and the contents weighted (to the nearest 0.0001 g) and then preserved in 4% formalin for taxonomic assessment and quantification (Herder & Freyhof, 2006). In the laboratory, prey from each gut were identified to the lowest possible taxonomic level under a dissecting microscope, and weighted (to the nearest 0.0001 g) and counted. Unidentifiable prey categories were quantified but not used in further analyses. The cestodes were not quantified because they might be parasites.
The importance of each food category was calculated in terms of frequency of occurrence of prey F i , relative abundance of prey A i , and prey-specific abundance P i (Amundsen, Gabler, & Staldvik, 1996): where N i is the number of fishes with prey i in their gut, N is the total number of fishes with gut contents of any kind, S i is the total weight of prey i, S t the total foregut content of all foreguts examined, and St i is the total weight of foregut content with prey i in their foregut.
The degree of dietary overlap among each pair of species was calculated using Schoener's (1970) similarity index: where p i and q i represent the proportions by weight of different gut content categories of any two species, respectively. D varies between 0 and 1, representing no to complete food overlap. Diet overlap is usually considered significant when D exceeds 60% (Wallace, 1981).
To determine diet specialization of each species, diet breadth (B) was calculated using Levin's (1968) index: where p i is the proportion of each prey i in the diet.
The modified Costello (1990) graphical method (Amundsen et al., 1996) was used to assess the seasonal variations in feeding strategy of different fish species. In this method, the prey-specific abundance (P i ; y-axis) of each prey was plotted against the frequency of occurrence (F i ; x-axis) in a two-dimensional graph.
Information on prey importance, feeding strategy, and phenotype contribution to the niche width can be obtained according to the distribution of points along the diagonals and axes of the diagram.  Amundsen et al. (1996).

| Diet composition
A total of 16 prey categories were identified (

| Dietary niche breadth
The

| Diet overlap
Considerable dietary overlap among the four studied gudgeons was detected in spring, with all Schoener's indexes exceeded 60% (

| Feeding strategy
In spring, all studied species exhibited a remarkable similarity in their feeding strategies (Figure 4)

39.59
Notes: F i , the frequency of occurrence of each prey item; A i , the relative abundance of each prey item.

| Diet composition
The analysis of dietary composition revealed that the four studied gudgeon species fed predominantly on aquatic macroinvertebrates, such as L. lacustris, Hydropsychidae, Perlidae, Tubificidae, Sinopotamidae, Bellamya, Gammarus, Ephemeroptera, and Sphaerium. These results were broadly in line with previous studies (Huang & Deng, 1990;Xu et al., 1981;Zhou & He, 1992). For example, Huang and Deng (1990) found that C. heterodon fed mainly on aquatic insects, Chironomidae larvae, and L. lacustris. Xu et al. (1981) observed that the diet of C. Heterodon was mainly composted of L.
lacustris, Sphaerium, and Gammarus. Zhou and He (1992)  cylindricus as an important supplemental food resource, despite of the low nutritional and energetic value (Bowen, 1979(Bowen, , 1987 whereas mollusks constituted a relatively small proportion (Liu, 2009;Liu et al., 2009). Because of the high flow velocity in the upper Yangtze River, it would be very difficult for these benthic-feeding fish to catch the drift aquatic insects. Therefore, they would prefer to select the benthic, sessile, and relative slow-moving prey categories, in order to maximize their net rate of energy gain, as predicted by the optimal foraging theory (Emlen, 1966;Gerking, 1994).
Among the four studied species, C. guichenoti showed the broadest dietary niche. Some large-sized prey categories, such as Sinopotamidae, Macrobrachium, and fish, were exploited exclusively by C. guichenoti. Morphologically, C. guichenoti has a relative larger mouth than other species. This large mouth might allow C. guichenoti to capture large-sized prey more efficiently. Thus, the low utilization of crabs, shrimps, fish, and other large-sized prey by C. heterodon, R. ventralis, and R. cylindricus was probably a result of morphological constraints of feeding apparatus, similar as in other species (Magalhães, 1993).
Seasonal variations in the diet composition of studied fish species were observed. Some aquatic invertebrates, such as Macrobrachium, Bellamya, Ephemeroptera, and Chironomidae, were commonly consumed by fish species in spring but little consumed in autumn. Many investigations have revealed that the availability of macroinvertebrate in the upper and middle Yangtze River changed significantly across seasons (Chen, Xia, Pan, Xu, & Ni, 2017;Jiang, Xiong, & Xie, 2017;Liu et al., 2009). In spring, the macroinvertebrate showed a high diversity and abundance, benefited from the moderate water temperature and stable flow regime (Jiang et al., 2017;Liu et al., 2009). However, with the washout of monsoon floods and the emergence of some species (e.g., Trichoptera, Diptera, and Neuroptera) in late summer and autumn, the diversity and abundance of aquatic invertebrates both decreased significantly (Jiang et al., 2017;Liu et al., 2009). Therefore, the seasonal changes in diet composition were probably resulted from the decreased availability in the environment, as observed by other studies (Magalhães, 1993;Martin & Genner, 2009).

| Trophic partitioning
We observed a considerable high dietary overlap among the studied species, especially in spring. However, high degree of dietary overlap may not always indicate competition (Deus & Petrere-Junior, 2003;Gabler & Amundsen, 2010;. When food resource availability is high, sympatric fish species may become more generalist, which can also result in high dietary overlap (Gabler & Amundsen, 2010). In the present study, the four studied gudgeon species have similar feeding apparatus and all prefer to utilize the same aquatic invertebrate (Zeng & Liu, 2011). Moreover, the macroinvertebrate community commonly shows the highest diversity and density in spring , which may not lead to interspecific competition for food resource in this season. Therefore, it is possible that the high prey availability in spring enables fish species to share the same food resources, and hence the observed high diet overlap (Gabler & Amundsen, 2010 (Hammerschlag, Ovando, & Serafy, 2010;Jardas, Santic, & Pallaoro, 2004;Tyler, 1971;Gray et al., 1997). Numerous studies have demonstrated that seasonal fluctuation in food resource availability may affect the trophic relationships among sympatric fish species (Sánchez-Hernández et al., 2016Gray et al., 1997). When food resource availability is high, fish species may become more generalist (Gabler & Amundsen, 2010). However, with the decline of food resources, species may specialize or generalize in resource use, according to the extent of food resource limitation (Genner et al., 1999;Amarasekare., 2003;Gabler & Amundsen, 2010;Sánchez-Hernández et al., 2017). The present study confirmed that interspecific trophic relationships of fish assemblage may be more complex than we have expected. Even in the same community, divergence and convergence in resource use among species can occur simultaneously, as proposed by Genner et al. (1999) (Deus & Petrere-Junior, 2003). Furthermore, segregations in microhabitat use (Magalhães, 1993), prey size ( also alleviate the interspecific competition among sympatric fish species and thereby facilitate their coexistence. Therefore, future studies should pay more attention to these aspects, in order to enhance our understandings of coexistence mechanism of these sympatric gudgeon species in the upper Yangtze River.

| CON CLUS IONS
The present study revealed seasonal differences in food resource utilization among four sympatric gudgeons. High diet overlap among studied species was observed due to their common utilization on the abundant aquatic invertebrate, especially in spring. However, changes in feeding strategies were observed in autumn. Specifically, on previous studies about the macroinvertebrate community of the study area (Liu, 2009;Liu et al., 2009) to discuss the seasonal changes in food resource partitioning. Additionally, the present study was focused only on the four most dominated species. In order to have a comprehensive understanding of the interspecific relationships of the whole fish assemblage, more attentions should be put into the changes of food availability and more species should be included in further studies. Despite the above-mentioned problems, the present study provides valuable information for understanding the resource partitioning of sympatric species in the upper Yangtze River.

ACK N OWLED G EM ENTS
Thanks are given to Shengguo Dan, Xiufang Zhu, Yuting Hu, and other colleagues for their assistance in field sampling. We appreciate constructive advice from Dr. Xiong Zhang and Dr. Dan Yu, which helpfully improved the quality of the manuscript. Thanks are also given to reviewers for their valuable comments and suggestions that helped improve the manuscript. This study received financial support from the National Natural Science Foundation of China (NSFC 31400359), the Research Project of China Three Gorges Corporation (0799570, 0799574) and Sino BON-Inland Water Fish Diversity Observation Network.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
Fei Liu contributed to field sampling, data analyses, and writing of the manuscript. Jianwei Wang and Huanzhang Liu contributed to research design and writing of the manuscript.

DATA ACCE SS I B I LIT Y
The data supporting the results, such as the number and size range of fish specimens and the diet composition of the studied fish species, have been listed in the tables.