Differences in Habitat Substrate Preference Selection Among Sexes and Populations of Leptobotia elongata

Abstract

Cascade hydropower development in the upper Yangtze River has substantially altered riverine habitat conditions. However, quantitative information on fish substrate preferences remains limited. Based on field surveys, six representative substrate types were identified in the Yibin and Jiangjin reaches, and substrate selection patterns of Leptobotia elongata were quantified under controlled flume conditions. This study examined sex-specific and population-level differences in substrate preference. A total of 50 samples were analyzed using one-way ANOVA and the Kolmogorov–Smirnov test. The results are summarized as follows: (1) Substrates of 40–20 mm: Individual female L. elongata exhibited the longest residence time on the 40 mm substrate (1640 s), while individual males also spent the longest time on the 40 mm substrate (1549 s). Within this substrate range, males showed a preference only for the 40 mm substrate (PR = 739), whereas females preferred three substrate types, with the strongest preference for the 40 mm substrate (PR = 7543). (2) Substrates of 10–2.5 mm: Female individuals spent the longest time on the 10 mm substrate (1468 s), and male individuals likewise showed the longest residence time on the 10 mm substrate (1568 s). In this range, males did not show a significant preference for any substrate (PR = −907), whereas females preferred three substrate types, with the 2.5 mm substrate exhibiting the highest preference intensity (PR = 2059). (3) Population-level patterns for 40–20 mm substrates: The population spent the longest time on the 40 mm substrate (1799 s) and preferred three substrate types, among which the 20 mm substrate had the highest preference ratio relative to the other preferred substrates (PR = 4750). (4) Population-level patterns for 10–2.5 mm substrates: The longest residence time was observed on the 10 mm substrate (1762 s). The population preferred three substrate types, with the 10 mm substrate showing the highest preference ratio (PR = 5709). Overall, L. elongata showed a general tendency to prefer larger-sized substrates. This preference is likely associated with enhanced habitat complexity, improved foraging opportunities, and the formation of low-velocity refuges. Our results suggest that maintaining or restoring coarse substrate patches in regulated reaches may improve habitat suitability for L. elongata and potentially benefit other benthic fishes with similar ecological requirements.

1. Introduction

Riverbed substrate is a fundamental physical component of riverine habitats and plays a critical role in shaping fish distribution, reproduction, and survival [1,2,3,4]. For benthic species in particular, substrate provides essential conditions for spawning, shelter, and foraging. However, human activities such as hydropower construction, channel regulation, and dredging have substantially altered riverbed structure, leading to habitat degradation and loss in many river systems [5,6,7]. Leptobotia elongata, a migratory benthic fish endemic to the upper Yangtze River, relies heavily on hydrological conditions and substrate characteristics throughout its life cycle. Extensive cascade hydropower development in this region has markedly modified riverine environments and contributed to a sharp population decline of the species [8]. L. elongata is currently listed as Vulnerable (VU) in the China Red Data Book of Endangered Animals [9]. Understanding its behavioral responses to substrate conditions is therefore essential for effective habitat restoration and conservation planning.

Habitat structure and associated physical complexity strongly influence fish assemblages by affecting food availability, shelter, and recruitment processes [10,11,12]. Numerous studies have demonstrated that fish species richness and density vary among substrate types, with structurally complex habitats often supporting higher biodiversity [13,14,15]. Some species exhibit strong associations with specific substrate types, whereas others shift habitats across life stages. These findings highlight the ecological importance of substrate characteristics in shaping fish distribution and population dynamics, underscoring the need to incorporate habitat structure into conservation and management strategies [16]. Fish habitat selection is influenced by multiple environmental factors, including substrate, flow velocity, temperature, light, and sound [17,18,19]. Among these, riverbed substrate plays a particularly important role for benthic species, and understanding substrate preference is essential for habitat modification and restoration efforts [20]. For example, gravel enhancement has been shown to improve spawning success in salmonids [21]. In addition to environmental factors, biological traits such as sex and social behavior can shape habitat use patterns. Sex-based differences in schooling behavior and aggregation preferences have been widely documented [22,23,24,25,26,27], and these behavioral differences may indirectly influence spatial distribution and habitat selection. However, the interactive effects of sex and schooling behavior on substrate preference remain poorly understood.

In this study, L. elongata was selected as the focal species. Based on field surveys of riverbed substrate composition, experimental flumes were constructed to simulate habitats with different substrate types. Video tracking was used to quantify substrate selection by female and male individuals under solitary and group conditions. A substrate preference index was calculated to evaluate preferred substrate types for each sex and for grouped fish. We aimed to (1) determine whether substrate preference differs between males and females, and (2) assess whether group composition modifies substrate selection patterns. Understanding these behavioral responses provides critical insights for habitat conservation and riverbed restoration efforts in regulated freshwater systems.

2. Materials and Methods

2.1. Preliminary Survey

Based on preliminary fish resource surveys, L. elongata was found to be primarily distributed in the Yibin and Jiangjin sections of the upper Yangtze River (Figure 1). Field habitat surveys for this study were conducted in May 2025 across these two areas. At each sampling site, geographic coordinates were recorded, and data on substrate type and particle size were collected using measuring tapes and photographic documentation. Substrate types were classified following the “Standards for Sediment Particle Analysis in Rivers” [28]. Based on the substrate index (the volumetric proportion of small and medium gravel in the substrate composition) [29], habitats in the Yibin section were classified into three substrate types: 40 mm, 30 mm, and 20 mm. While habitats in the Jiangjin section were classified into another three substrate types: 10 mm, 5 mm, and 2.5 mm. Relevant substrate parameters are detailed in

Table 1. Habitat substrate types in the Yibin and Jiangjin sections of the upper Yangtze River, China.

Sampling Sites	Sediment (<2 mm)	Gravel (2–16 mm)	Small and Medium Pebbles (16–250 mm)	Smooth Type-Large Pebbles and Large Stones (>250 mm)	Bottom Mud Attachment-Large Pebbles and Large Stones (>250 mm)
1 (Yibin)	*	/	/	/	/
2 (Yibin)	+	+	+	*	*
3 (Yibin)	+	*	*	/	/
4 (Yibin)	+	*	*	/	+
5 (Jiangjin)	*	*	*	/	/
6 (Jiangjin)	*	+	/	/	/
7 (Jiangjin)	/	+	*	/	+

Note: “*” represents a high proportion; “+” represents a low proportion; “/” represents absence this substrate type. Substrate types were classified following the “Standards for Sediment Particle Analysis in Rivers”.

.

2.2. Experimental Setup

A rectangular flume (1.5 m long, 0.8 m wide, 0.6 m high and slope 0°) was used to simulate riverbed substrate habitats (Figure 2). An infrared network camera (DS-2CD2T87SWDV2-L, Hikvision, Hangzhou, China) was mounted 3.0 m above the flume to record the substrate selection behavior of L. elongata. The camera’s field of view covered the entire flume, ensuring complete behavioral observations. The experimental water was aerated tap water, maintained for 5 days, with dissolved oxygen levels kept above 7 mg/L. The water temperature was maintained at 24 °C and the depth at 0.4 m. A water pump (QDX1.5-16-37, Shanghai People’s Enterprise Group Pump Co., Ltd., Shanghai, China) provided a flow with an inflow velocity of 0.3 m/s, which exceeds the known flow perception threshold of L. elongata [30]. Two experimental substrate configurations were tested: one with substrates of 40–20 mm and another with substrates of 10–2.5 mm. In both setups, the flume was divided lengthwise into three equal sections, with each substrate type (occupying 1/3 of the total area) arranged sequentially according to particle size.

2.3. Fish Species Husbandry and Trials

Sexually mature and cultured L. elongata (Figure 3) specimens (n = 50) were supplied by Jingzhou Breeding Base of the Yangtze River Fisheries Research Institute, Chinese Academy of Fishery Sciences (Animal Welfare and Ethics Approval Number: YFI2025YULIXIONG 02) on 28 July 2025 (

Table 2. Physiological parameters of experimental fish.

Experimental Setting	Sample Size	Length (cm)	Weight (g)
Substrate of 40–20 mm ♀	5	20.67 ± 3.78	209.5 ± 14.08
Substrate of 40–20 mm ♂	5	21.7 ± 1.7	201.1 ± 16.7
Substrate of 10–2.5 mm ♀	5	21.5 ± 2.0	199.6 ± 12.8
Substrate of 10–2.5 mm ♂	5	22.0 ± 2.1	204.3 ± 14.0
Substrate of 40–20 mm ♀:♂ = 2:1	15	22.2 ± 2.3	208.3 ± 30.1
Substrate of 10–2.5 mm ♀:♂ = 2:1	15	22.9 ± 1.4	209.1 ± 26.2

Note: “♀” represents female L. elongata; “♂” represents male L. elongata.

) and reared in recirculating holding tank (3000 L) sterilized with permanganate to minimize the risk of disease. Water temperature was maintained at 23.0 ± 1.0 °C using the thermostatic system, dissolved oxygen was kept above 7.0 mg/L, and approximately 20% of the water volume was renewed daily. To minimize the impact of light on L. elongata, shading cloth was used to provide light shielding. Experimental fish were fed to satiation once daily with the same feed (protein content ratio was 35%). After 30 min feeding, uneaten feed and feces were removed using a siphon. To allow L. elongata to recover from transport stress and regain physical condition, the fish were acclimated for at least 10 days prior to the start of the experiments [31].

This study involved both individual and group experiments. Since schooling behavior is readily observable in groups of three fish, the group size was set to three individuals [32]. The experiments were conducted from 2 to 25 August 2025 at the behavioral research facility in Jianli city, daily between 09:00 and 18:00. Before each trial, healthy, active experimental fish were acclimated in the experimental flume’s holding zone for 1 h. After acclimation, cameras were activated to record fish behavior in the flume. Each trial lasted 30 min and was repeated five times.

2.4. Statistical Analysis

To quantify the substrate preference of L. eptobotia, the residence time of fish on different substrate types was recorded. Habitat preference was assessed based on both residence time and a calculated substrate preference index (PI). Residence time on a given substrate was determined by frame-by-frame tracking (25 frames·s⁻¹) in SwisTrack (version 4.0.0; The Distributed Intelligent Systems and Algorithms Laboratory (DISAL) and the LPM Vision Group at EPFL, Lausanne, Switzerland), using the position of the fish head as the reference. The interval from when the fish head entered a specific substrate area until it exited was recorded as one residence event. For each 30 min trial, all such intervals within a given substrate area were summed to obtain the total residence time. The proportion of residence time for each substrate was calculated as its residence time divided by the total residence time across all substrates. The substrate preference index (PI) was calculated as the difference between the mean cumulative area-weighted residence value for a substrate and the actual area proportion of that substrate, with PI > 0 indicating preference. When PI ≥ 2, the preference ratio (PR) was used to express the relative preference intensity of the most preferred substrate compared to other preferred substrates [33]. Data were processed using Microsoft Excel 2018. Statistical analyses were performed in R (version 4.3.3), and figures were generated with Origin 2022 (64-bit). Differences in residence time among substrate types were analyzed using one-way ANOVA and the Kolmogorov–Smirnov (K–S) test. Data are presented as mean ± standard deviation (mean ± SD). A p-value < 0.05 was considered statistically significant.

3. Results

3.1. Substrate Selection Preferences of L. elongata Individuals Among Different Substrate Types

The residence times of both female and male L. elongata differed significantly among substrate types (F = 4.087, p < 0.05, Appendix A.1 

Table A1. Male and female L. elongata residence time of both substrates.

	Df	Sum sq	Mean sq	F Value	p Value
Time	1	11.52	11.521	4.087	0.0478
Residuals	58	163.48	2.819

). In the experiment with 40–20 mm substrates, the total residence time for both sexes was longest on the 40 mm substrate, followed by the 20 mm and then the 30 mm substrate. The time spent on the 40 mm substrate was significantly longer than that on the other two substrates (male: F = 8.974, p < 0.05; female: F = 7.347, p < 0.05) (Appendix A.2 

Table A2. Male L. elongata residence time of 40–20 mm substrates.

	Df	Sum sq	Mean sq	F Value	p Value
Time	1	4.084	4.084	8.974	0.0103
Residuals	13	5.916	0.455

and

Table A3. Female L. elongata residence time of 40–20 mm substrates.

	Df	Sum sq	Mean sq	F Value	p Value
Time	1	3.611	3.611	7.347	0.0178
Residuals	13	6.389	0.491

). Similarly, in the 10–2.5 mm substrate experiment, the total residence time was greatest on the 10 mm substrate, followed by the 2.5 mm and then the 5 mm substrate, and residence time on the 10 mm substrate was significantly longer than on the others (male: F = 7.306, p < 0.05; female: F = 4.801, p < 0.05) (Figure 4; Appendix A.3 

Table A4. Male L. elongata residence time of 10–2.5 mm substrates.

	Df	Sum sq	Mean sq	F Value	p Value
Time	1	3.598	3.598	7.306	0.0181
Residuals	13	6.402	0.492

and

Table A5. Female L. elongata residence time of 10–2.5 mm substrates.

	Df	Sum sq	Mean sq	F Value	p Value
Time	1	2.697	2.6969	4.801	0.0473
Residuals	13	7.303	0.5618

).

The substrate preference index (PI) was calculated for female and male L. elongata across different substrate types. Results showed that both sexes exhibited preferences for multiple substrates within each substrate-size category. In the 40–20 mm experiment, male preference intensity followed the order 40 mm (PI = 2685) > 20 mm (PI = 1732) > 30 mm (PI = 1210), with a preference ratio (PR) of 739 for the 40 mm substrate relative to others. Similarly, female preference intensity was ordered as 40 mm (PI = 2581) > 20 mm (PI = 1869) > 30 mm (PI = 1302), with a PR of 7543 for the 40 mm substrate. In the 10–2.5 mm experiment, male preference intensity ranked as 2.5 mm (PI = 6377) > 5 mm (PI = 6245) > 10 mm (PI = 5437); however, the negative preference ratio (PR = −907) indicated no overall substrate preference. Conversely, females showed a clear preference ranking of 2.5 mm (PI = 6931) > 5 mm (PI = 6877) > 10 mm (PI = 5890), with the 2.5 mm substrate having a positive PR of 2059 relative to other preferred substrates (Figure 5).

3.2. Substrate Selection Preferences of Schooling L. elongata Among Different Substrate Types

The residence time of L. elongata groups (female-to-male ratio 1:2) was quantified across different substrates. In the 40–20 mm experiment, total residence time followed the order 40 mm > 30 mm > 20 mm, and was significantly longer on the 40 mm substrate than on the others (p < 0.05, Appendix A.4 

Table A6. Substrates of 40–20 mm of group L. elongata.

Comparison	Z Value	P.undaj	P.adj
40–30 mm	4.559	5.127 × 10−6	1.538 × 10−5
40–20 mm	4.823	1.419 × 10−6	4.228 × 10−6
30–20 mm	0.264	7.916 × 10−1	1.000

). In the 10–2.5 mm experiment, residence time followed the order 10 mm > 5 mm > 2.5 mm, with time on the 10 mm substrate also being significantly longer (Figure 6; p < 0.05, Appendix A.5 

Table A7. Substrates of 10–2.5 mm of group L. elongata.

Comparison	Z Value	P.undaj	P.adj
10–5 mm	3.232	1.228 × 10−3	0.003
10–2.5 mm	4.274	1.912 × 10−3	5.74 × 10−5
5–2.5 mm	1.042	2.971 × 10−1	0.891

).

Analysis of the substrate preference index (PI) revealed that the L. elongata population exhibited preferences for multiple substrate types within each substrate-size category. In the 40–20 mm experiment, population-level preference intensity decreased in the order: 20 mm (PI = 1650) > 40 mm (PI = 1531) > 30 mm (PI = 1511). The 20 mm substrate had a preference ratio (PR) of 4750 relative to other preferred substrates. Similarly, in the 10–2.5 mm experiment, the preference order was 10 mm (PI = 1970) > 5 mm (PI = 1902) > 2.5 mm (PI = 1828), with the 10 mm substrate yielding a PR of 5709 (Figure 7).

4. Discussion

Substrate type plays an important role in shaping fish community structure and habitat use [34,35], as habitat characteristics strongly influence fish distribution and abundance [36]. Fish assemblages associated with sandy substrates are often distinct from those inhabiting harder substrates, likely due to differences in habitat complexity and prey availability. Species richness, abundance, and biomass are generally lower in sandy habitats and higher on hard substrates, with intermediate values observed in structurally complex habitats such as seagrass beds. The reduced structural complexity of unvegetated sandy bottoms limits the availability of refuges and food resources, thereby constraining fish community development [37]. Field surveys conducted in this study further revealed that Leptobotia elongata predominantly occupies habitats composed of gravel and small- to medium-sized pebbles. Substrates with greater structural complexity provide abundant interstitial spaces that support diverse benthic invertebrate assemblages, which constitute an important component of the diet of L. elongata [38]. As a carnivorous-leaning omnivorous species, L. elongata mainly feeds on benthic macroinvertebrates and small aquatic organisms. Therefore, substrate type can indirectly influence its habitat selection by mediating prey distribution, accessibility, and foraging efficiency. Consequently, clarifying substrate preferences is critical for understanding habitat requirements and for guiding riverine habitat restoration and conservation efforts [39,40].

In the present experiments, both sexes exhibited preferences for three substrate types within the 40–20 mm range, with the strongest preference for the 40 mm substrate (females: PR = 7543; males: PR = 739) (Appendix A.6 

Table A8. Summary of cumulative PI and PR for males, females, and groups across different substrate sizes.

Sex/Group	Substrate Size	Cumulative PI	Cumulative PR
male	40 mm	2685	739
male	30 mm	1210	−770
male	20 mm	1732	−230
male	10 mm	5437	−907
male	5 mm	6245	−744
male	2.5 mm	6377	−718
female	40 mm	2581	7543
female	30 mm	1302	3706
female	20 mm	1869	5407
female	10 mm	5890	1747
female	5 mm	6877	2043
female	2.5 mm	6931	2059
Group	40 mm	1531	4392
Group	30 mm	1511	4333
Group	20 mm	1650	4750
Group	10 mm	1970	5709
Group	5 mm	1902	5505
Group	2.5 mm	1828	5282

). This shared preference may be attributed to the ability of larger substrates to create low-velocity zones and visual refuges on their downstream sides, which reduce locomotor energy expenditure and facilitate spawning-related behaviors, thereby enhancing reproductive success [41,42]. In addition, coarse gravel substrates may function as effective ambush sites, increasing prey encounter rates and capture success for L. elongata. By contrast, in the 10–2.5 mm substrate range, substrate selection patterns diverged between sexes. Males showed no overall substrate preference (PR = −907), whereas females preferred three substrate types, with the highest preference for the 2.5 mm substrate (PR = 2059). This female-specific preference for finer sediments is likely related to sex-specific ecological roles, potentially associated with reproductive behavior or spawning-site selection, rather than feeding efficiency alone [43]. Overall, these results indicate that substrate preference in L. elongata reflects a balance between foraging opportunities, energetic efficiency, and reproductive ecology shaped by substrate-mediated prey availability and habitat complexity in natural riverine environments.

Social interaction and aggregation are common in natural ecosystems [44,45]. In our study, video recordings showed that male L. elongata exhibited following behaviors, including accompanying and chasing females. In the 40–20 mm substrate experiment, group substrate preferences differed from those of individual fish. This discrepancy likely reflects a leader–follower mechanism, in which one individual’s action is imitated by others while maintaining distance and moving in parallel [46,47]. In contrast, during the 10–2.5 mm substrate experiment, group preferences mirrored those of individual females. This pattern may result from reproductive strategies: oviposition cues from females can synchronize male substrate selection, overriding innate male preferences [24]. Group exploratory tendencies and leadership by key individuals also likely influence substrate selection [48,49]. Less active individuals may inhibit exploration, while stronger social connectivity facilitates coordinated movement [50]. Occasionally, individual fish temporarily separated, then the group followed, highlighting the role of following behavior in group behavior.

Overall, substrate preferences of L. elongata appear shaped by ecological habits, available substrate types, and response intensity. A clear pattern emerged: the 2.5 mm substrate was rarely selected, while larger gravel was strongly preferred. Coarse substrates create low-velocity zones and visual refuges, enhancing predator avoidance, reducing energy expenditure, and providing ambush sites that improve foraging success. Based on PI and PR values, these results can guide riverbed restoration: wherever possible, multiple large substrates should be provided, and substrates ≤ 2.5 mm should be avoided.

The present study represents a first attempt to examine potential differences in substrate preference between female and male L. elongata and to assess the influence of group behavior on substrate selection in this species. Our results suggest that sex did not influence substrate selection in the 40–20 mm size class, whereas sex-related differences may exist in the 10–2.5 mm size class. In the 40–20 mm size class, substrate preferences of groups appeared to differ from those of solitary individuals, while in the 10–2.5 mm size class, group substrate selection was generally consistent with that of individuals. The specific interactions among color, shape, and substrate size may influence individual substrate preference responses and represent an important avenue for future research. Nonetheless, this does not negate the general pattern that L. elongata predominantly selects larger substrates. Due to sample size limitations, the population-level experiments in this study were conducted with only three fish per group (female-to-male ratio of 1:2), and potential differences in substrate selection under varying group densities and sex ratios were not investigated.

5. Conclusions

Using an open flume with different substrate types, this study examined sex-specific and population-level substrate preferences of L. elongata. The results consistently showed that substrate size strongly influenced habitat use. Both females and males spent longer times on larger substrates, particularly the 40 mm substrate in the 40–20 mm treatments and the 10 mm substrate in the 10–2.5 mm treatments. Preference index analyses further indicated that individuals of both sexes displayed preferences for multiple substrate types, with the strongest preferences occurring on larger substrates. At the population level (female-to-male ratio of 1:2), group behavior modified substrate use patterns, but the overall tendency toward larger substrates remained evident. Populations spent more time on the larger substrates within each size category, while preference index results suggested shifts in preferred substrate types depending on substrate size. Overall, these findings demonstrate that L. elongata, at both individual and population levels, shows a clear tendency to prefer larger-sized substrates. This preference is likely driven by the formation of low-velocity zones and visual refuges downstream of coarse substrates, which reduce locomotor energy expenditure, enhance predator avoidance, and provide effective ambush sites that improve foraging efficiency and prey capture success.

These findings provide practical guidance for habitat restoration in the upper Yangtze River. Maintaining or restoring coarse substrate patches (e.g., 20–40 mm gravel) in regulated reaches is recommended to enhance habitat complexity and foraging opportunities for L. elongata, while substrates ≤ 2.5 mm should be avoided in restoration designs. Incorporating multiple large substrate types can also accommodate sex-specific habitat use and population-level behavioral plasticity. Such measures are expected to improve riverbed restoration effectiveness and benefit other benthic fishes with similar ecological requirements.