Effects of Protein Level on the Production and Growth Performance of Juvenile Chinese Mitten Crab ( Eriocheir sinensis ) and Environmental Parameters in Paddy Fields

: Rice–crab co-culture systems represent integrated agriculture–aquaculture systems de-veloped in China over the last 30 years. The rice–crab co-culture area comprised approximately 1.386 × 10 5 hm 2 in 2019. However, there is no speciﬁc feed designed for Chinese mitten crab ( Eriocheir sinensis ) cultured in this system until now. In this study, we investigated feed formu-lae for the nutritional requirements of Chinese mitten crab in this mode. The control group was not fed with any artiﬁcial feed (Co), and the experimental groups were fed with three different feeds of 15% (T15), 30% (T30), or 45% (T45) protein content, respectively. Growth performance variations in E. sinensis were investigated along with water quality, phytoplankton, zooplankton, aquatic vascular plants, and benthic animals in the paddy ﬁelds to determine the effect of crabs and their diet on the paddy ecosystem. Dietary protein levels had no signiﬁcant effect on water quality. The biomass and species of phytoplankton, zooplankton, aquatic vascular plants, and zoobenthos in the paddy ﬁeld were affected by crabs and their diet. Morphological parameters of crabs were signiﬁcantly more pronounced in the high-protein group than in the other groups. However, the T45 diet negatively affected production by increasing feed costs, causing precocious puberty and inducing water eutrophication. In conclusion, adding a 15% protein compound feed can meet the nutritional needs of crabs, reduce culture costs, and improve water quality. The discharged water had low ammonia nitrogen and nitrite content and no eutrophication occurred, so the water could be recycled. These ﬁndings provide a scientiﬁc reference for supporting rice and ﬁsh co-cultivation. length of zooplankton decreases, but the average body length and length frequency distribution of zooplankton also shifted to that of smaller individuals under predation pressure. The results showed that predation


Introduction
The Chinese mitten crab (Eriocheir sinensis) is the most commonly farmed crab species in China. In 2020, the General Office of the Ministry of Agriculture and Rural Affairs proposed the implementation of "five major actions," including the promotion of ecological and healthy farming modes. The new rice-crab co-culture mode integrates the culturing of rice and crabs with ecological, economical, and social benefits [1]. The food chain in the ecosystem of this mode is quite complex, creating a more stable ecosystem than that in single-species aquaculture. Crabs are at the top of this food chain and feed on plankton, weeds, and benthic animals in rice fields, ensuring an efficient matter circulation and a smooth energy flow through the whole system [2]. Furthermore, this mode produces a double harvest of rice and crabs [3].
The ecological environment in rice-crab co-culture may be affected by several factors. For instance, high-density culture adversely affects phytoplankton and benthic animals [4].

Figure 1.
Experimental design field of the crab-rice field. Inlet is where the enclosure received water; the outlet is where the enclosure drained water. Co is the control group with no artificial feed supplied, and T15, T30, and T45 represent the treatment groups fed with experimental feeds of 15%, 30%, and 45% protein content, respectively.
The feed used in the experimental group was designed by the research group with FM and soybean meal as the main protein sources, and fish oil was used as the main fat source (Table 1). Three types of isolipid feeds with different protein contents were formulated by simultaneously increasing the FM and soybean meal content (FM: soybean meal = 2). The feed protein levels were 15%, 30%, and 45%. The various solid raw materials were accurately weighed according to the required formula ratio and, then, were fully mixed according to the principle of step-by-step enlargement. Subsequently, the artificial feed was pulverized through a 100-mesh sieve, the oil was added, and all ingredients were stirred to an even consistency. Finally, water was added (30%), and the feed was mixed again. A double helix A pellet mill (DES-TS1280, Jinan Dingrun Machinery Equipment Co., Ltd.) was used to press the feed into 3 mm diameter pellets. The pellets were naturally air-dried and packaged and sealed in plastic bags. The bags were stored in a refrigerator at −20 °C. Crabs were fed at a rate of 10% of their body weight. A detailed list of contents of each experimental feed is shown in Table 1, and the feed costs are presented in Table  A1. . Experimental design field of the crab-rice field. Inlet is where the enclosure received water; the outlet is where the enclosure drained water. Co is the control group with no artificial feed supplied, and T15, T30, and T45 represent the treatment groups fed with experimental feeds of 15%, 30%, and 45% protein content, respectively. During the experiment, physical and chemical water quality indicators were recorded and monitored in each enclosure. The indicators included temperature (oxygen dissolving instrument, YSI550-A, Vasey Instrument Company, Exton, PA, USA), pH (pH meter, PHB-1, Shanghai Thunder Magnetic, Shanghai, China), salinity (Pen salinity meter, AR-8012, Xima Instrument Co., Ltd., Dongguan, China), dissolved oxygen (oxygen dissolving instrument, YSI550-A, Vasey Instrument Company), ammonium nitrogen (visible spectrophotometer, V-1100, Shanghai Meitong Instrument Co., Ltd., Shanghai, China), and nitrite nitrogen [11].

Growth Performance and Yield of Crabs
The megalopae of crabs were collected and weighed to ensure all enclosures contained the same number of crabs (160/g) on 25 May. The megalopae were cultured in the enclosures for 46 days until they reached the feeding phase. The experiment was initiated after the crabs were measured. The average body length, width and height was 1.08 ± 0.08, 1.16 ± 0.09 and 0.53 ± 0.05 cm, respectively, and average body weight was 0.63 ± 0.13 g. During the experiment, the body length, width, height, and weight of the crabs were measured on 10 July, 28 July, 17 August, 8 September, and 8 October. All crabs were humanely harvested at the end of the experiment. Crabs were caught using a plastic bucket inserted into a hole dug in the bottom of the enclosure. The frequency of collection depended on the number of crabs. All specimens were counted, measured, and weighed. Precocious puberty was assessed by comparing the abdomen, junction, villi, gonad, color, and crab patterns with those of the representative crab specimens [12]. Growth performance indicators were calculated using the following formulae: Survival rate/% = n t /n 0 × 100%, Weight gain rate/% = (m t − m 0 )/m 0 × 100%, Specific growth rate/%/d = (ln m t − ln m 0) /t × 100%, Total output/g·m −2 = W/S, Net output/g·m −2 = (W − W 0 )/S, where n 0 represents the initial number of crabs, n t represents the final number of crabs, m t represents the final average body weight, m 0 represents the initial average body weight, t represents the total number of days of the experiment, W represents the final total weight of crabs in an enclosure, W 0 represents the initial total weight of crabs in an enclosure, and S represents the area of the enclosure (6 m × 6.7 m = 40.2 m 2 ).

Qualitative and Quantitative Analysis of Phytoplankton
The sampling and measurement methods used to assess the phytoplankton were based on those of Zhang [6]. Briefly, 1 L of water was collected by five-point sampling at each point and mixed in a bucket. A lugol solution (10-15 mL) was then evenly mixed into the water. After 48 h, the sample was concentrated by siphonage, fixed at 100 mL volume, and then put into an iodometric bottle for qualitative analysis. The qualitative and quantitative analyses followed Li et al. [13], and Zhao [14], respectively. The specific gravity of phytoplankton is approximately 1. Therefore, the volume was directly converted into wet weight, and the phytoplankton biomass was calculated (Table A2).

Qualitative and Quantitative Analyses of Zooplankton
The sampling and measurement of the zooplankton were based on methods of Zhang [6]. The qualitative and quantitative methods followed those described in Section 2.3.3, and the zooplankton biomass was calculated (Table A3).

Qualitative and Quantitative Analyses of Aquatic Vascular Plants
The aquatic vascular plants were sampled by selecting two points that were consistent for each enclosure. A 30 cm × 30 cm iron frame was used to divide the sampling area. The plants (except rice) were uprooted, species were identified, and plant wet weight was determined. The quantification of the benthic animals was conducted at the same sampling points as those mentioned in Section 2.3.5 at a depth of approximately 10 cm using a self-made barrel dredger [15]. The benthic animals were screened using a sieve with an aperture of 0.2-2 mm and then wet-weighed, identified, and counted with precision.

Statistical Analysis
The experimental data were collated using Excel. The homogeneity of variance test and one-way ANOVA were performed using SPSS 24.0. Any significant differences between groups were further analyzed using Duncan's multiple comparison tests. The results were expressed as the mean ± standard deviation. In all analyses, a probability value less than 0.05 was considered significant (p < 0.05).
Dominance (Y) was calculated according to the formula: where Y is the degree of dominance, ni is the number of individuals of species i, N is the total number of individuals, and fi is the frequency of occurrence of species i at five sampling points within an enclosure. The Shannon-Wiener diversity index (H ) was calculated using the formula: where ni is the number of individuals of species i, and N is the total number of individuals of the species.

Growth Performance and Yield of Crabs
The morphological parameters of the crabs in the high-and low-protein diet groups varied significantly at each measurement ( Figure 2). At the end of the experiment, the carapace length, width, and height of both T30 and T45 groups were significantly higher than those of the Co group (p < 0.05), and the carapace length and width were significantly higher than those in the T15 group (p < 0.05). The final body weight and weight gain rate of the crabs in the T45 group were significantly higher than those of the Co group (p < 0.05). The growth rate of the crabs in the Co group was significantly lower than that of the T30 and T45 groups (p < 0.05).
Water 2022, 14, x FOR PEER REVIEW and significantly increased with the increase in dietary protein content (p < specific growth rate of the crabs varied from 3.30%/d to 3.85%/d and sig increased as the dietary protein content increased (p < 0.05).
The total and net yields of crabs varied from 43.86 g/m 2 to 60.41 g/m 2 and 3 to 54.64 g/m 2 , respectively. The highest total and net yields were for crabs in the T followed by those of the T15 and Co groups; the lowest was for those of the T There was no significant variation in the total and net yield of crabs in the experimental groups (p > 0.05). Towards the end of the experiment, approximate crabs in the T45 group experienced precocious puberty.   I-Co  I-T15  I-T30  I-T45  F-Co  F-T15  F-T30  F-T45 n=3; ±SE  The final body weight of the crabs ranged from 8.30 g to 17.28 g ( Table 2). The final body weight of the crabs that were fed diets increased significantly with the increase of protein content (p < 0.05). The body weight increase rate varied from 9641.86% to 20,181.73% and significantly increased with the increase in dietary protein content (p < 0.05). The specific growth rate of the crabs varied from 3.30%/d to 3.85%/d and significantly increased as the dietary protein content increased (p < 0.05). The total and net yields of crabs varied from 43.86 g/m 2 to 60.41 g/m 2 and 38.10 g/m 2 to 54.64 g/m 2 , respectively. The highest total and net yields were for crabs in the T45 group, followed by those of the T15 and Co groups; the lowest was for those of the T30 group. There was no significant variation in the total and net yield of crabs in the different experimental groups (p > 0.05). Towards the end of the experiment, approximately 10% of crabs in the T45 group experienced precocious puberty.

Water Quality
The physical and chemical parameters of water quality in each enclosure during the experimental period was assessed ( Figure 3). Generally, the parameters of water quality in each enclosure were consistent with the fishery water quality standard of the People's Republic of China (GB 11607-89). During the experiment, the water temperature ranged from 24.5 • C to 28.2 • C, and the salinity from 0.493‰ to 1.743‰. No significant difference was found in dissolved oxygen, ammonia, and nitrate levels in all enclosures ( Figure 3).

Phytoplankton Biodiversity
A total of 54 species of phytoplankton from seven phyla were detected in four treatments during the experiment (Table A4). Seven phyla were present in all treatments ( Figure 4). Bacillariophyta was the dominant group, with 19 species present, and accounted for 35.19% of the phytoplankton species observed. Chlorophyta was the phylum with the second highest number of species (18 species) and accounted for 33.33% of the total number of species. Other groups included Cyanophyta (9 species; 16.67%) and Euglenophyta (5 species; 9.26%). The phyla Cryptophyta, Chrysophyta, and Pyrrophyta were represented by one species each and accounted for 1.85% of the total species.

Phytoplankton Biodiversity
A total of 54 species of phytoplankton from seven phyla were detected in four treatments during the experiment (Table A4). Seven phyla were present in all treatments ( Figure 4). Bacillariophyta was the dominant group, with 19 species present, and accounted for 35.19% of the phytoplankton species observed. Chlorophyta was the phylum with the second highest number of species (18 species) and accounted for 33.33% of the total number of species. Other groups included Cyanophyta (9 species; 16.67%) and Euglenophyta (5 species; 9.26%). The phyla Cryptophyta, Chrysophyta, and Pyrrophyta were represented by one species each and accounted for 1.85% of the total species.

Phytoplankton Biodiversity
A total of 54 species of phytoplankton from seven phyla were detected in four treatments during the experiment (Table A4). Seven phyla were present in all treatments ( Figure 4). Bacillariophyta was the dominant group, with 19 species present, and accounted for 35.19% of the phytoplankton species observed. Chlorophyta was the phylum with the second highest number of species (18 species) and accounted for 33.33% of the total number of species. Other groups included Cyanophyta (9 species; 16.67%) and Euglenophyta (5 species; 9.26%). The phyla Cryptophyta, Chrysophyta, and Pyrrophyta were represented by one species each and accounted for 1.85% of the total species.

Variation Trends in Phytoplankton Biomass over Time
The phytoplankton biomass of different treatment groups over times was analyzed statistically. The overall average biomass of each experimental group showed a downward trend with time. However, an abnormal increase occurred for those in Co group on 15 August ( Figure 6). On 25 May, the biomass in T45 was significantly highe than that in Co (p < 0.05). On 30 June, the biomass in T15 was significantly lower than tha in Co and T45, while those in Co and T15 was significantly higher than that in T45 on 15 August (p < 0.05). The phytoplankton biomass in Co group was the lowest among al groups on 15 October (p < 0.05).

Variation Trends in Phytoplankton Biomass over Time
The phytoplankton biomass of different treatment groups over times was analyzed statistically. The overall average biomass of each experimental group showed a downward trend with time. However, an abnormal increase occurred for those in Co group on 15 August ( Figure 6). On 25 May, the biomass in T45 was significantly higher than that in Co (p < 0.05). On 30 June, the biomass in T15 was significantly lower than that in Co and T45, while those in Co and T15 was significantly higher than that in T45 on 15 August (p < 0.05). The phytoplankton biomass in Co group was the lowest among all groups on 15 October (p < 0.05).
Water 2022, 14, x FOR PEER REVIEW 9 of 31 Figure 6. Phytoplankton biomass in different treatment groups during the experiment. Different lowercase letters in each group represent significant differences (p < 0.05).

Succession and Population Changes in the Dominant Phytoplankton Species
The dominant phytoplankton species were calculated, and the results are presented in Figure 7. If Y > 0.02, then the species was considered dominant. After the quantification and calculation analysis, there were five phyla of dominant zooplankton. The species composition and dominance of each phytoplankton species varied within different sampling time intervals in each treatment group (Table A5).

Succession and Population Changes in the Dominant Phytoplankton Species
The dominant phytoplankton species were calculated, and the results are presented in Figure 7. If Y > 0.02, then the species was considered dominant. After the quantification and calculation analysis, there were five phyla of dominant zooplankton. The species composition and dominance of each phytoplankton species varied within different sampling time intervals in each treatment group (Table A5).

Succession and Population Changes in the Dominant Phytoplankton Species
The dominant phytoplankton species were calculated, and the results are presented in Figure 7. If Y > 0.02, then the species was considered dominant. After the quantification and calculation analysis, there were five phyla of dominant zooplankton. The species composition and dominance of each phytoplankton species varied within different sampling time intervals in each treatment group (Table A5). Chromulina pygmaea and Chlorella pyrenoidosa were highly dominant species in different treatment groups during the experiment. The diversity of dominant phytoplankton species in the Co group increased over time. Comparatively, the dominant species of each feeding group were relatively simple. In the later stage of the experiment, the degree of dominance of Chroococcus in the T30 and T45 groups was higher than that in the Co and T15 groups ( Figure 8). On 2 September, the degree of dominance of Chroococcus in T45 was significantly higher than in the other groups (p < 0.05).
Water 2022, 14, x FOR PEER REVIEW 10 of 31 species of each feeding group were relatively simple. In the later stage of the experiment, the degree of dominance of Chroococcus in the T30 and T45 groups was higher than that in the Co and T15 groups ( Figure 8). On 2 September, the degree of dominance of Chroococcus in T45 was significantly higher than in the other groups (p < 0.05).

Zooplankton Species Diversity
A total of 50 species of zooplankton were detected during the experiment ( Figure 9). Protozoa had the highest number of species, with 23 species detected, and accounted for 46% of the total number of species. Rotifera had 15 species detected, accounting for 30% of the total species. Cladocera had eight species, accounting for 16% of the total species. Copepoda had four species, accounting for 8% of the total species.

Zooplankton Species Diversity
A total of 50 species of zooplankton were detected during the experiment ( Figure 9). Protozoa had the highest number of species, with 23 species detected, and accounted for 46% of the total number of species. Rotifera had 15 species detected, accounting for 30% of the total species. Cladocera had eight species, accounting for 16% of the total species. Copepoda had four species, accounting for 8% of the total species.

Zooplankton Species Diversity
A total of 50 species of zooplankton were detected during the experiment (Figure 9). Protozoa had the highest number of species, with 23 species detected, and accounted for 46% of the total number of species. Rotifera had 15 species detected, accounting for 30% of the total species. Cladocera had eight species, accounting for 16% of the total species. Copepoda had four species, accounting for 8% of the total species. The zooplankton species in different treatment groups during the experiment are listed in Table A6. The average biomass of Copepoda was the largest at 31.61 mg/L and accounted for 50.00% of the overall zooplankton biomass. That of Cladocera was 29.29  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  T45  Co…  T15  T30  The zooplankton species in different treatment groups during the experiment are listed in Table A6. The average biomass of Copepoda was the largest at 31.61 mg/L and accounted for 50.00% of the overall zooplankton biomass. That of Cladocera was 29.29 mg/L and accounted for 46.33% of the total biomass. These two groups accounted for 96.32% of the total average biomass. The average biomass of Protozoa was 1.76 mg/L, accounting for 2.79% of the total. The zooplankton group with the lowest average biomass was the rotifers, with only 0.56 mg/L, accounting for 0.89% of the total average biomass.
The zooplankton diversity in each group was analyzed using the Shannon-Wiener diversity index ( Figure 10). The overall average value was 1.69. The diversity index was from 1. 19    The total average zooplankton biomass in the paddy field fluctuated over tim was an upward trend from the beginning of the experiment to 15 July, wh decreased to 15 August and increased to September 15 before decreasing agai 11). The biomass of zooplankton in T45 was significantly lower than those in T15

Variation Trends of the Zooplankton Biomass over Time
The total average zooplankton biomass in the paddy field fluctuated over time. There was an upward trend from the beginning of the experiment to 15 July, which then decreased to 15 August and increased to September 15 before decreasing again ( Figure 11). The biomass of zooplankton in T45 was significantly lower than those in T15 and T30 on 30 June (p < 0.05), while the biomass in T45 was significantly higher than those in other groups on 29 July (p < 0.05).

Variation Trends of the Zooplankton Biomass over Time
The total average zooplankton biomass in the paddy field fluctuated over time. T was an upward trend from the beginning of the experiment to 15 July, which decreased to 15 August and increased to September 15 before decreasing again (Fi 11). The biomass of zooplankton in T45 was significantly lower than those in T15 and on 30 June (p < 0.05), while the biomass in T45 was significantly higher than those in o groups on 29 July (p < 0.05).

Succession of Dominant Zooplankton Species and Community Changes
The dominant zooplankton species in the paddy fields during the experiment are shown in Table A7 and Figure 12. When dominance value (Y) > 0.02, the species was considered dominant. The dominant species in different treatment groups consisted of 32 species belonging to 4 zooplankton groups (i.e., Rotifera, Copepoda, Cladocera and Protozoa), respectively. The biomasses of the dominant zooplankton species were relatively low in the four treatment groups on 25 May and 7 October. The number of dominant species in T15 was the lowest on 15 August. The numbers of dominant species were 6 to 10 in Co group, 4 to 11 in T15 group, 4 to 9 in T30 group, and 6 to 9 in T45 group, respectively, from 30 June to 26 September.

Species, Quantities, and Changes in Aquatic Vascular Plants
The species diversity and biomass of the aquatic vascular plants in the enclosures of each treatment group are shown in Figure 13. Seven aquatic vascular plants were detected in the four treatment groups, i.e., Vallisneria spiralis, Monochoria vaginalis, Sparganium stenophyllum, Spirodela polyrhiza, Potamogeton sp., Elodea nuttallii, and Scirpus validus.
The changes of the submerged-plant biomass in different treatment groups over time are shown in Figure 14. The statistical analysis revealed significant variations in the biomass of submerged plants in each treatment group with different times. The overall submerged-plant biomass rapidly decreased to 30 June before rapidly increasing to 25 July and then increased gradually to 29 July and 15 August. The biomass in the T45 group was significantly higher than the other groups (p < 0.05). Afterwards, submerged-plant biomass rapidly decreased again. No submerged plant was found in all treatment groups from 15 September to 7 October.
The biomass of the emergent plants in different treatment groups over time are shown in Figure 15. The statistical analysis revealed significant variations in the biomass of emerged plants at different times. Generally, the biomass increased throughout the experiment and only decreased in the last sample collection. The emergent plants biomass in the Co group was significantly higher than that in the T45 group on July 15 and July 29, while the biomass in the T15 and T30 groups was significantly higher than that in the T45 group on July 29 (p < 0.05). The dominant zooplankton species in the paddy fields during the experiment are shown in Table A7 and Figure 12. When dominance value (Y) > 0.02, the species was considered dominant. The dominant species in different treatment groups consisted of 32 species belonging to 4 zooplankton groups (i.e., Rotifera, Copepoda, Cladocera and Protozoa), respectively. The biomasses of the dominant zooplankton species were relatively low in the four treatment groups on 25 May and 7 October. The number of dominant species in T15 was the lowest on 15 August. The numbers of dominant species were 6 to 10 in Co group, 4 to 11 in T15 group, 4 to 9 in T30 group, and 6 to 9 in T45 group, respectively, from 30 June to 26 September.

Species, Quantities, and Changes in Aquatic Vascular Plants
The species diversity and biomass of the aquatic vascular plants in the enclosures of each treatment group are shown in Figure 13. Seven aquatic vascular plants were detected in the four treatment groups, i.e., Vallisneria spiralis, Monochoria vaginalis, Sparganium stenophyllum, Spirodela polyrhiza, Potamogeton sp., Elodea nuttallii, and Scirpus validus.    and then increased gradually to 29 July and 15 August. The biomass in the T45 gro significantly higher than the other groups (p < 0.05). Afterwards, submerge biomass rapidly decreased again. No submerged plant was found in all treatment from 15 September to 7 October. r 2022, 14, x FOR PEER REVIEW experiment and only decreased in the last sample collection. The emergent plan in the Co group was significantly higher than that in the T45 group on July 15 a while the biomass in the T15 and T30 groups was significantly higher than that group on July 29 (p < 0.05).

Variations in the Benthic Animal Species and Quantities with Time
The diversity and biomass of benthic animals in different treatment g shown in Table A8 and Figure 16. Five taxa were found in the nine samples fro treatment groups, i.e., Gyraulus sp., Euconulus sp., Limnodrilus sp., Branchiur Insecta.

Variations in the Benthic Animal Species and Quantities with Time
The diversity and biomass of benthic animals in different treatment groups are shown in Table A8 and Figure 16. Five taxa were found in the nine samples from the four treatment groups, i.e., Gyraulus sp., Euconulus sp., Limnodrilus sp., Branchiura sp., and Insecta.
The biomass of benthic animals in different treatment groups changed over time. Generally, they initially increased from 23 May to 29 July, then decreased rapidly to 15 August, followed by a gradual decline through September until the end of the experiment ( Figure 17). On 2 September, the biomass of the benthic animals in the Co group was significantly higher than that in the T45 and T30 groups (p < 0.05). No significant difference was found among various treatment groups at the same sampling time.

Variations in the Benthic Animal Species and Quantities with Time
The diversity and biomass of benthic animals in different treatment groups are shown in Table A8 and Figure 16. Five taxa were found in the nine samples from the four treatment groups, i.e., Gyraulus sp., Euconulus sp., Limnodrilus sp., Branchiura sp., and Insecta. The biomass of benthic animals in different treatment groups changed over time. Generally, they initially increased from 23 May to 29 July, then decreased rapidly to 15 August, followed by a gradual decline through September until the end of the experiment ( Figure 17). On 2 September, the biomass of the benthic animals in the Co group was significantly higher than that in the T45 and T30 groups (p < 0.05). No significant difference was found among various treatment groups at the same sampling time.

Effects of Diets with Different Protein Levels on the Growth Performance and Yield of Juvenile Crabs
Protein is one of the most important nutritional components in the diet of crabs, and the level required varies depending on growth stages [16]. The five separate measurements of morphological parameters of the crabs revealed that the high-protein compound feed resulted in significantly higher crab carapace length, width, and height and body weight compared with those in the low-protein group. Feeding with a highprotein compound feed had a significantly positive effect on the weight gain rate, final body weight, and specific growth rate of juvenile crabs. These results support the findings of Zhang [6]. However, the T30 group in this experiment may have escaped and/or had "milky disease" [17]. The rate of disease varied according to the original health status of the crabs and may have resulted in the observed differences in survival rates; however, there was no significant difference in the crab yield between the different diet treatments. These results do not correspond with the growth rate results. Therefore, the contribution of natural food to crab growth may be underestimated in the rice-crab co-culture mode.
Integrated agricultural and aquaculture systems can effectively contribute to green and sustainable agricultural development and ensure food security [18]. In 2017, the "General Principles of Technical Specifications for Rice and Fishing Integrated Planting

Effects of Diets with Different Protein Levels on the Growth Performance and Yield of Juvenile Crabs
Protein is one of the most important nutritional components in the diet of crabs, and the level required varies depending on growth stages [16]. The five separate measurements of morphological parameters of the crabs revealed that the high-protein compound feed resulted in significantly higher crab carapace length, width, and height and body weight compared with those in the low-protein group. Feeding with a high-protein compound feed had a significantly positive effect on the weight gain rate, final body weight, and specific growth rate of juvenile crabs. These results support the findings of Zhang [6]. However, the T30 group in this experiment may have escaped and/or had "milky disease" [17]. The rate of disease varied according to the original health status of the crabs and may have resulted in the observed differences in survival rates; however, there was no significant difference in the crab yield between the different diet treatments. These results do not correspond with the growth rate results. Therefore, the contribution of natural food to crab growth may be underestimated in the rice-crab co-culture mode.
Integrated agricultural and aquaculture systems can effectively contribute to green and sustainable agricultural development and ensure food security [18]. In 2017, the "General Principles of Technical Specifications for Rice and Fishing Integrated Planting and Culture," issued by the Chinese Agriculture and Rural Affairs Bureau, highlighted the fact that animals raised in aquaculture should make full use of natural bait present in the environment (in this case the rice paddies), reducing the use of fish feed. The results of this experiment strongly support this statement. The lower temperature in the paddies is more favorable to the growth of crabs compared with the temperature in monoculture crab systems, which can reduce crab sexual precocity [19].
In this experiment, the sexual precocity rate in the T45 group was approximately 10%, which may be due to the high protein content. Chen et al. [20] demonstrated that when the protein content in the feed is too high, excess protein is converted into fat and stored in the hepatopancreas, resulting in sexual precocity in crabs. Sexual precocity during the breeding process reduces culture efficiency. Studies have shown that the survival rate of adult crabs cultured with precocious crab species in the second year is already extremely low [21]. Therefore, it is important to prevent the sexual precocity of crabs in production.

Changes in Physical and Chemical Properties of Paddy Water Environment
There were no obvious changes in water temperature, pH, salinity, ammonia nitrogen, or nitrite nitrogen over the course of this experiment. There were also no significant differences between the experimental groups. The ammonia nitrogen and nitrite content of the water was low. However, there was a downward trend in dissolved oxygen levels in the water, which may have been caused by a variety of factors. The daily photosynthesis of plants is the main source of dissolved oxygen in water [22]. Animal respiration in the water releases large amounts of organic matter. Additional organic matter is produced during decay (after death) and after feeding, which leads to an increase in the respiration in water and sediments, and is also the main destination of dissolved oxygen [23]. The crabs were placed in the experimental enclosures on 29 May. On the same day, the dissolved oxygen in the water began to decrease, indicating that crab respiration was the main factor causing the low dissolved oxygen in the rice-crab co-culture. As the experiment progressed, the shading effect of large vascular plants (including the rice) led to a decrease in phytoplankton photosynthesis. This is also one of the reasons for the decrease in dissolved oxygen levels.
In the later stages of the experiment, the plankton species and biomass and the biomass of zooplankton increased, while the biomass of the phytoplankton decreased. Consequently, there were more aerobic biological factors and less oxygen-producing organisms in the environment, resulting in a decrease in the dissolved oxygen levels in the water. The levels of dissolved oxygen ranged from 3.04 to 8.75 mg/L, which is lower than the normal dissolved oxygen requirements of crabs (5 mg/L). In low dissolved oxygen conditions, crabs tend to escape. Crabs also crawl to the shore in the later culture stage. The dissolved oxygen content of the water in rice-crab co-culture is lower than that in conventional rice fields [24,25]. This may cause a stress response in the crabs and is, therefore, one of the shortcomings of breeding crabs in rice paddies.

Effects of Diets with Different Protein Levels on the Phytoplankton in Paddy Fields
As primary producers, phytoplankton also act as a natural food source for crabs in ricecrab co-culture systems. Through the experiment, the aquatic organisms and water quality factors affected and correlated with each other. Species diversity is a basic characteristic of biological communities and is an important indicator of a healthy system [26]. In this experiment, the overall average Shannon-Wiener diversity index of the phytoplankton in the paddy field was 1.07, indicating that the phytoplankton diversity in the paddy field environment was not extremely diverse but superior to that in polluted waters. Studies have shown the Shannon-Wiener diversity index for phytoplankton in the rice-crab coculture mode is higher than that in conventional rice fields [5,27]. This is due to the reduction in the use of chemical fertilizers and pesticides in rice-fish symbiosis [27,28]. Thus, biodiversity in the paddy fields has been well protected [29,30].
The rice growth and the subsequent shading effect of the rice reduced the lightreceiving area of the water in the paddy field. This weakened the photosynthesis by phytoplankton. Furthermore, the physical and chemical factors involved in water quality and organisms in the water environment interact with each other [30,31]. Individual phytoplankton are small and varied, and different species can affect the environment in diverse ways. Adaptability of the species differs, and most can intuitively reflect the changes in water physicochemical factors after environmental changes. Rice will absorb nutrients and ions in the paddy field, effectively regulating the physical properties and chemicals in the environment, and can inhibit the absorption of nutrients by phytoplankton. Therefore, it was expected that the total average biomass of phytoplankton would show a decreasing trend with time. When the phytoplankton productivity is low, the consumption of phytoplankton by zooplankton is an important factor affecting phytoplankton growth [32]. For example, the sample collected on 15 August revealed that the phytoplankton biomass had abnormally increased. When the data were combined with the analysis of changes in zooplankton biomass, the zooplankton biomass had decreased significantly at that time. Reduced grazing pressure on phytoplankton results in abnormally elevated phytoplankton biomass. In the natural environment, there are many reasons for an increase in phytoplankton biomass. For example, an increase in nutrient concentration can lead to similar results. The water quality monitoring results showed that the nutrient index in the water was not significantly different from that in other periods; therefore, there was no increased nutrient concentration.
The results of this experiment revealed that the diversity of the dominant phytoplankton species in the control group presented an increasing trend. However, it did not vary among the feeding groups. This may be because the addition of exogenous nutrients in the feed led to the eutrophication of the water body, resulting in a decrease in biodiversity. Water eutrophication destroys the ecosystem balance and can even lead to the collapse of the entire aquatic system [33]. When the nutrients in the water increased, Cyanophyta phytoplankton gradually became the dominant species at the expense of other species, indicating that Cyanophyta are indicator organisms for water eutrophication [34]. In this experiment, the dominance of Chroococcus in the T45 group was significantly higher than that of the other three groups on 2 September, after which it decreased with no significant difference, indicating that high protein levels could cause water eutrophication. However, in a paddy field environment, water eutrophication is not a concern owing to the self-purification function of rice.

Effects of Different Protein Levels in the Crab Diet on Zooplankton in Paddy Fields
In the rice-crab co-culture environment, zooplankton can feed on phytoplankton, which is also the main food of crabs. Zooplankton is an important link between energy flow and material cycling in the ecosystem [35]. Zooplankton species and community structure are affected by environmental factors. A total of 51 species of zooplankton were identified in this experiment, and the Shannon-Wiener diversity index was 1.73. Previous research has shown that the Shannon-Wiener diversity index of Cladocera and Rotifers in a water environment under rice-crab co-culture is higher than that of conventional rice fields [36]. In addition, owing to the purification effect of rice, zooplankton in crab paddy fields is highly diverse.
The results of this experiment showed that the zooplankton biomass initially increased and then decreased before increasing again. Combined with the analysis of the changes in phytoplankton in the paddy field, the biomass of the phytoplankton was relatively high in the early stage of the experiment. Then, the zooplankton fed on the phytoplankton and grew rapidly; its biomass also increased. As crabs grew, they preyed on the zooplankton, and the zooplankton biomass decreased. Horn et al. [37] tracked zooplankton body length and found that not only the maximum body length of zooplankton decreases, but the average body length and length frequency distribution of zooplankton also shifted to that of smaller individuals under predation pressure. The results showed that predation pressure on zooplankton by crabs led to the miniaturization of zooplankton. In the later stage of this experiment, the miniaturization of zooplankton, combined with the larger size and mouthparts of crabs, reduced the crab predation on zooplankton, so the biomass of zooplankton increased.
The dominant zooplankton species in the early stages of the experiment, on 25 May and 30 June, were rotifers, especially Polyarthra trigla, which is consistent with Zhang's [6] results. Rotifers, Cladocera, and copepods all competitively feed on phytoplankton in paddy fields. According to Gilbert [38], when there is competition between Cladocera and rotifers, Cladocera has an advantage. Therefore, the existence of Cladocera affects the diversity and quantity of rotifers. As the experiment progressed, some Cladocera species gradually became dominant. In the later stage of this experiment, the dominant species of zooplankton in the crab paddy field were small and mainly existed in the state of copepod nauplii, with the dominant species being mainly protozoa.

Effects of Crabs on Aquatic Vascular Plants in Paddy Fields
Large plants control zooplankton, provide habitats for fish that feed on zooplankton, and provide shelter for phytoplankton [39]. In the early stages of this experiment, the main submerged plant in the paddy field environment was S. polyrhiza. The growth of the submerged plants, including V. spiralis and Potamogeton sp., increased, and the biomass of submerged plants also increased. Shading also increases with the rapid growth of emergent plants [40]. Crabs began feeding on the submerged plants, which reduced their growth until no submerged plants were detected from 15 September. On 7 October, the emergent plants decayed and died, decreasing the biomass.
Farmers have traditionally used chemical weeding machines to remove large weeds from rice fields. Over time, weed resistance and herbicide damage have become increasingly serious problems [41], since crabs are omnivorous and feed on large plants, such as aquatic vascular plants [42]. Even if there is an excess of animal food in the environment, crabs will still consume aquatic plants, especially submerged plants [43]. However, crabs rarely feed on emergent plants, which allows the emergent plants to grow and absorb fertilizer from the crab pond sediment [44]. Lv et al. [45] found that the fresh and dry weights of weeds in the experimental group without crabs being provided artificial feed were significantly lower than those in the crab feeding group. Other studies have also shown that weed control by rice crabs is more effective than traditional weed control methods used in rice production [19,46].

Changes in Benthic Animals in Rice-Crab Co-Culture
Benthic animals are the main food source for crabs [47]. Xu et al. [48] found that crabs affect habitat structure in two ways: feeding and reducing competition with aquatic plants by preying on attached organisms, thereby promoting the growth of aquatic plants. At the beginning of the experiment, the local benthic animals were in the culture period, and the biomass showed an upward trend, which is consistent with Li et al. [46]. As crabs grew, the predation of benthic animals by crabs increased, and the biomass of the benthic animals decreased. The stress of predation by crabs caused Branchiura sp., Limnodrilus sp., and other benthic animals that reproduce via burrowing, to increase in numbers, causing an overall increase in the benthic animal biomass.
In this experiment, the crabs fed with high-protein compound feed were larger and had a stronger predation ability on benthic animals than the representative crabs. On 2 September, the biomass of benthic animals increased with the level of dietary protein.
The biomass of benthic animals in the control group was significantly higher than that of the three experimental groups. Xu et al. [48] found that the stocking of crabs reduced the benthic animal diversity in the environment based on the lakes where crabs were stocked, outside the lake enclosure, in natural lake waters, and in lakes where fish were stocked. Benthic species diversity decreased significantly, and the production volume and density decreased by more than 60% compared with those of the control water body. The results vary from the results in our experiment, which may be caused by the different environments, sampling times, and stocking densities of the crabs. First, lake and paddy field environments are quite different. Second, unlike Xu et al.'s [48] experiment, which took four samples twice a year for two years, we monitored the dynamics of zoobenthos nine times over a period of nearly five months (from May to October). By comparing the culture densities, Xu et al. [48] showed that the over farming of crabs causes the high variations. Conversely, our experiment used normal crab culture density.

Conclusions
In this study, we evaluated how different levels of protein in crab feed could affect the performance of crabs and the environment in rice-crab co-culture in paddy fields. The results showed that feed with 15% protein level compound diet can not only meet the nutritional requirements of crabs but also reduce the cost of cultivation and improve the water quality of the paddy field. The discharged water had low ammonia nitrogen and nitrite content, and no eutrophication was observed. Consequently, the water could be recycled. These findings provide a scientific basis for feed formulation for juvenile crabs in rice-crab co-culture. Institutional Review Board Statement: Our study did not involve endangered or protected species. In China, breeding and catching Chinese mitten crabs, Eriocheir sinensis, in rice fields does not require specific permits. All efforts were made to minimize animal suffering and discomfort. The animal study protocol was approved by the Animal Ethics Committee of Shenyang Agriculture University.

Data Availability Statement:
The data presented in this study are not publicly available but are available upon request from the corresponding author.

Acknowledgments:
The author would like to acknowledge my mentor Li for imparting knowledge to me and helping me revise the article carefully. We would also like to thank all employees of the Panjin Guanghe Crab Industry Co., Ltd. for patiently helping us and providing us with an excellent test environment. We thank teacher Hu, who is a teacher and a friend, for helping me to revise the article and overcome many difficulties. We are very grateful to Tian for his suggestions on the revision of the article and teaching us a lot of knowledge. Thanks to Jiang for helping me contact the company to polish the article. Thanks to my employees at the company's R and D center for their experimental help and guidance. The author would also like to acknowledge Zuo and Liu from Dalian Ocean University for their great help with the formula and production of the feed required for the experiment. Thanks also to Liu and Zheng who gave me practical guidance and help with the culture. I would also like to thank MSA Bi who provided guidance on testing techniques. Thanks to my junior Liang for assisting me in data processing and analysis.