Introduction

The red swamp crayfish, Procambarus clarkii (Girard, 1852), is an economically important species in China. In 2018, the output of P. clarkii in China was 1.64 million tonnes, so it ranked first in the output of freshwater crustaceans in China; in addition, the output was nearly double the annual output in 20161,2. In 2018, the total area for P. clarkii cultivation in China was 1.12 million hectares, and, of this, paddy fields accounted for 0.84 million hectares. Paddy fields accounted for 40% of the total area used for rice cultivation and fisheries in 20183. Rice–crayfish co-culture has a lot of economic benefits, with an increase in net income by 6302.7 USD per hectare4. According to our field survey (data not provided), farmers in some places such as Honghu City in Hubei Province and Yancheng City in Jiangsu Province often widen rice ditches and pay more attention to P. clarkii cultivation than rice production, which is against the principles of rice–fish co-culture (Technical specifications for integrated farming of rice and aquaculture animal, SC/T 1135-2017).

Since the twenty-first century, rice–fish co-culture has been widely practiced in paddy fields in Asian countries, especially in China5. The term ‘fish’ refers to a wide range of aquatic animals such as carp, crab and crayfish. Previous studies have suggested that aquaculture animals ushered into paddy fields can provide multiple services to rice ecosystems, i.e. decrease the abundance of insects and weeds, reduce agrochemicals inputs and enhance both soil and rice quality5,6,7,8. The aims of rice–fish co-culture are to ensure the stable yield of rice and achieve stable gain and income through the cultivation of aquatic animals. Therefore, it is important to study the ecological services provided by aquatic organisms, such as weed control, in a rice–crayfish co-culture system.

Weed damage has become an important factor that affects rice production, potentially reducing rice yield by 20–80%9. Chemical control is the main method used to control weeds10, however, the wide application of chemical herbicides has caused environmental problems such as pollution, drug resistance and residual toxicity11,12,13. This has created a demand for sustainable crop management. Previous field studies have shown that rice–crayfish farming has a good control effect on dominant weed species, such as Ludwigia prostrata, Ammannia baccifera, Leptochloa chinensis, Lindernia procumbens and Echinochloa crusgalli14,15. Xu et al.15 found that Dicotyledoneae and Gramineae weed densities after 2–3 years of integrated rice–crayfish farming decreased by 73.53% and 63.26%, respectively, when compared with the rice monoculture model. Nonetheless, there is still a lack of direct evidence for P. clarkii control of weeds. In this study, we hypothesized that crayfish can directly feed on weeds found in rice fields in China. We evaluated the ability of P. clarkii to feed on common weeds found in rice fields in China by direct feeding and observation of animal behaviour and determined the specific amount of different weeds fed on by P. clarkii.

Materials and methods

Animals and weeds

The experiments were conducted in the laboratory of the Eco-environmental Protection Research Institute, Shanghai Academy of Agricultural Sciences. The P. clarkii specimens used in this study were obtained from Zhuanghang Experimental Station of Shanghai Academy of Agricultural Sciences. About 500 P. clarkii (9.76 ± 4.42 g) specimens were transported to the laboratory and acclimated for 2 weeks in plastic boxes (46.4 × 34.7 × 22 cm, L × W × H) filled with tap water. In the first week of acclimatization, the P. clarkii specimens were fed excessively with pelleted commercial crayfish feed (38% crude protein) every day. In the second week of acclimatization, the P. clarkii specimens were fed excessively with weeds every 2 days for domestication. Animal welfare and experimental procedures were in accordance with the Guide for the Care and Use of Laboratory Animals [Ministry of Science and Technology of China, (2006)398].

The following plant species were selected on the basis of their status as common weeds found in local rice–crayfish farms15,14: Ludwigia prostrata Roxb., Leptochloa chinensis (L.) Nees, Echinochloa crusgalli (L.) Beauv and Eclipta prostrata L. Mature seeds of the 4 weed types were collected from a field located at the Zhuanghang Experimental Station of the Shanghai Academy of Agricultural Sciences. Botanist Dr Tian was responsible for the collection and identification of the weed seeds. The seeds were air-dried and peeled using physical percussion. After collection, the seeds were separated and purified by air separation. The purified seeds were stored in the refrigerator at 4 °C in the weed laboratory of the Eco-Environmental Protection Research Institute, Shanghai Academy of Agricultural Sciences. The voucher IDs of the 4 weeds are D201903, Q201903, B201903 and L201903. The weeds were planted in trays (length × width × height, 60 × 24 × 3 cm) 1 month before the experiment. The trays were filled with soil sterilized by dry heat and placed on a plastic chassis, distilled water was added to the soil in the trays till saturation, and the weed seeds were then evenly spread. The trays were placed in a plant culture room with a photoperiod of 14 h light: 10 h dark and relative humidity of 60%. During the cultivation, water evaporation was monitored, and water was supplemented regularly. The plant experiments were performed in accordance with relevant guidelines.

Quantitative feeding experiment

A total of 30 plastic boxes and 120 P. clarkii specimens of the same size (Table 1) were used for the quantitative feeding experiment. On the basis of the density of P. clarkii in rice–crayfish fields, 4 P. clarkii (about 4 times the field density) were raised in a plastic box (38 × 26 × 12.8 cm, L × W × H). P. clarkii males and females were fed the 4 weeds separately; each group was fed with 1 weed, with a pelleted commercial crayfish feed (38% crude protein) as the control. The experiment had 10 treatments, with 3 replicates of each treatment.

Table 1 Summary statistics (mean ± SD) of feeding related parameters for P. clarkia after different treatments.
Full size table

At the beginning of the experiment, the healthy P. clarkii specimens were weighed and randomly divided into plastic boxes. Each box contained 4 plastic plants (3 cm in diameter and 35 cm in length) separately fixed on a sinker as shelter. The boxes were filled with approximately 5 L of aerated tap water to a depth of 5 cm. The experiment lasted for 2 weeks. No water exchange occurred during the experiment. The pH was 8.02 ± 0.12; dissolved oxygen, 8.82 ± 0.17 mg/L; residual chlorine, < 0.05 mg/L and water temperature, 22.2 ± 0.96 °C. A photoperiod of 12 h light (8:30 to 20:30):12 h dark was maintained during the experiment, and light intensity at the water surface was 108–254 lx (light meter LM-332, Japan).

The weeds were fed to P. clarkii from 15:00 to 16:00. Before feeding, the weed weight (Wi) was measured using an electronic balance. The weeds were offered to the specimens every 2 days, and the feeding amount was calculated according to the feeding conditions of P. clarkia; the feeding amount was 0.77 ± 0.12 g of each weed. Before each feeding, the remaining weeds were removed with a net, placed in a glass culture dish and dried in an oven at 60 °C for 48–72 h. The remaining weeds after drying (Wd) were weighed. In addition, 1.0 g (fresh weight) of each weed was weighed and dried in the oven, and the dry weight of 1.0 g weed (Wc) was calculated and used to measure the actual weed feed intake of P. clarkii. The P. clarkii commercial feed was fed to the specimens excessively at 16:00 every day. Before each feeding, the commercial feed was weighed with an electronic balance, and the feed quantity was recorded. At 9:00 every day, the remaining commercial feed was removed with a net, placed in a glass culture dish and dried in an oven at 60 °C for 48–72 h. In addition, 1.0 g of the commercial feed was weighed and placed in a glass culture dish, covered with water and maintained from 16:00 to 9:00, and then dried in an oven, and the dry weight of 1.0 g commercial feed was calculated and used to measure the actual commercial feed intake of P. clarkii. During the experiment, the dead P. clarkii and moulted shells were removed at 9:00 every day.

Behaviour observation experiment

To avoid cannibalism among the crayfish, 1 P. clarkii (10.03 ± 0.87 g) was placed in a plastic box (38 × 26 × 12.8 cm, L × W × H), and water depth in the box with no shelter was 3 cm. Each group was fed with 1 weed (1.08 ± 0.11 g), with 8 treatments in the experiment and 8 replicates of each treatment. The behaviour of the crayfish in the plastic box was observed and analysed (ViewPoint Zebralab3.3, France). The feeding rhythm of P. clarkii occurs mainly at night and in the morning16,17, so the behaviour observation time was from 15:00 to 23:59 and from 00:00 to 10:00.

The behaviour data were analysed using videos (Stomer player, China). The P. clarkii specimens caught the weeds with their claws and first appendages: single feeding behaviour. The following parameters were recorded during the experiment: feeding frequency, time of feed intake and duration of P. clarkii feeding.

Statistical analysis

The daily feed intake (FI) was calculated as follows:

$$ {\text{FI }}\left( {{\text{g}}/{\text{ind}}.} \right) \, = \, \left[ {W_{i} - \, \left( {W_{d} /W_{c} } \right)} \right]/n/d $$

where n is the total number of P. clarkii, and d is days between feedings.

The percentage of daily feed intake (PFI) was calculated as follows:

$$ {\text{PFI }}\left( \% \right) \, = { 1}00 \, *FI/W $$

where W is the mean initial body weight of P. clarkii in a weed treatment group.

The statistical analyses were performed using SPSS 23.0, and Origin 2017 (Origin Lab, USA) was used for plotting. FI and PFI were compared between weeds and P. clarkii gender by using two-way ANOVA, followed by the LSD multiple comparison test.

Results

Weed consumption

A summary of the feed intake experiment results is presented in Table 1. The survival rate of both P. clarkii females and males in the P. clarkii feed group was 100%, whereas P. clarkii specimens died in all the other treatment groups. The two-way ANOVA showed that the P. clarkii gender and weed species had no interactive effects on FI and PFI (P > 0.05; Table 2). No significant differences in FI and PFI were observed between the P. clarkii females and males (P > 0.05), but the FI and PFI values of different weed species were significantly different (P < 0.05; Table 2). The FI and PFI values of both P. clarkii females and males were significantly higher in the P. clarkii feed group than in the weed treatment groups (P < 0.05; Table 1). Both FI and PFI were significantly higher in the L. chinensis group than in the other weed treatment groups. No significant differences in FI and PFI values were observed between the L. prostrata and E. prostrata groups. The FI and PFI of the P. clarkii females and males were significantly lower in the E. crusgalli group than in the other groups (Table 1).

Table 2 Statistical analysis (two-way ANOVA) of the effects of gender and different weeds on the daily feed intake (FI) and proportion of daily feed intake (PFI) of P. clarkii.
Full size table

Feeding characteristics

The commercial feed was basically consumed within 4 h after feeding in the feed intake experiment, so the control group was not used for the behaviour observation experiment. The behavioural analysis results are listed in Table 3. The feeding frequency and duration of P. clarkii on the different weed groups are as follows: L. chinensis > E. prostrata > L. prostrata > E. crusgalli. No significant differences in feeding frequency and feeding duration were observed between the P. clarkii females and males. Both P. clarkii females and males were more active from 17:00 to 20:00 (Fig. 1). Both feeding frequency and feeding duration were significantly higher in the L. chinensis group than in the L. prostrata and E. crusgalli groups (P < 0.05, Table 3). No significant differences in feeding frequency and feeding duration were observed between the L. prostrata and E. crusgalli groups (P > 0.05). The feeding activity of both P. clarkii females and males was the least in the E. crusgalli group (Table 3).

Table 3 Behavioural statistics of P. clarkii in different weed treatment groups (mean ± SD).
Full size table
Figure 1
figure 1

Feeding activity of P. clarkii during the observation period.

Full size image

The P. clarkii females and males of the L. chinensis group ate every hour during the observation period, and the feeding duration was longer from 19:00 to 20:00 and 4:00 to 5:00 for the males and from 18:00 to 19:00 and 2:00 to 4:00 for the females (Fig. 1). The P. clarkii males in the L. prostrata group fed frequently from 17:00 to 19:00, whereas the P. clarkii females fed frequently from 16:00 to 17:00. The P. clarkii males in the E. prostrata group fed frequently from 17:00 to 23:00, whereas the P. clarkii females fed frequently around 5:00. Both P. clarkii females and males in the E. crusgalli group showed the lowest feeding frequency and duration, but both P. clarkii females and males exhibited a relatively long feeding time from 21:00 to 22:00.

P. clarkii exhibited 2 living states: active and static. When it started to move, it may have been looking for food. When it fed on the weeds, it used 2 positions: lying on its side for feeding (Fig. 2a) and feeding on its front (Fig. 2b). When P. clarkii fed on the weeds, it first used its cheliped to catch the weeds and then used its mouthparts to complete feeding (Fig. 2a, b). P. clarkii would lie on its side when it was still, for example, for sleeping (Fig. 2c).

Figure 2
figure 2

Feeding behaviour of P. clarkii (using L. chinensis as an example). (a) Feeding in the side direction, (b) Feeding in the front direction, (c) no motion.

Full size image

Discussion

P. clarkii is one of the most important invasive species found worldwide, and its ecological plasticity allows it to live in different types of environments. The great ecological plasticity of P. clarkii is also expressed in its feeding habits attributable to its polytrophic feeding behaviour18. Previous studies have shown that P. clarkii can establish a food chain suitable for its growth needs on the basis of the food sources in its living environment19. Therefore, is the fact that crayfish can reduce weed biomass in paddy fields related to this ecological characteristic? In this study, P. clarkii showed a strong appetite for some weeds, such as L. chinensis,and the PFI of L. chinensis was more than 2%. The results of the quantitative feeding and behaviour observation experiments were highly consistent. The P. clarkii specimens mostly preferred to eat L. chinensis, but hardly ate E. crusgalli. No significant differences were observed in the feeding amount with respect to L. prostrata and E. prostrata,however, according to the behaviour experiment results, the P. clarkii specimens preferred E. prostrata.

Previous studies on the diversity of weed communities in rice fields have shown that the density of L. prostrata and E. prostrata in the rice–crayfish co-culture system has significantly reduced when compared with rice monoculture, and the biomass of these 2 weeds continues to decrease with the increase in the duration of rice–crayfish co-culture14,15,20. In this study, the results showed that the feeding ability of L. prostrata and E. prostrata by P. clarkii was better and probably achieved by direct ingestion. However, the results obtained for L. chinensis and E. crusgalli were inconsistent with those of previous field studies. The biomass of L. chinensis and E. crusgalli decreased first (< 4 years) and then increased (7–8 years) with an increase in the duration of rice–crayfish co-culture, and both weed densities have been found to be lower in the rice–crayfish co-culture system than in the rice monoculture system15,20. This showed that P. clarkii can also control L. chinensis and E. crusgalli. However, in the present study, almost no P. clarkii specimen fed on E. crusgalli. In our study, increased feeding on L. chinensis was observed, relative to all the other weeds examined. Previous field investigations have reported that the biomass of L. chinensis was still significantly higher than that of L. prostrata and E. prostrata15,20, suggesting decreased consumption of L. chinensis. Therefore, it is unclear how P. clarkii controls weeds, especially L. chinensis and E. crusgalli, as it seems that the weeds were not controlled by direct feeding. P. clarkii can directly feed on agricultural seeds such as rice seeds, which contain high protein and/or energy21,22. Therefore, another possibility is that P. clarkii inhibits weed growth by ingesting weed seeds or suppressing weed seed germination through burrowing.

P. clarkii is an opportunistic, omnivorous feeder, and its diet includes submersed macrophytes, algae, invertebrates and detritus18. Generally, P. clarkii likes to feed on aquatic plants, but there as a few studies on P. clarkii feeding on terrestrial plants. During P. clarkii cultivation, fishermen generally grow aquatic plants such as Hydrilla verticillata and Elodea nuttallii. The large amount of L. chinensis consumed by P. clarkii in the present study shows the unlimited potential of P. clarkii to control weeds in rice fields. Both L. chinensis and E. crusgalli are gramineous plants, but the feeding selectivity of these 2 weeds with respect to P. clarkii was very different. L. chinensis seedlings are tender23, which may make its taste closer to that of aquatic plants. In addition, olfaction plays an important role in the feeding process of P. clarkii24, and it is possible that the odour of L. chinensis attracts P. clarkii to a greater extent than that of E. crusgalli. The results of the quantitative feeding experiment showed no significant differences in the feeding amount of the P. clarkii males and females in the L. prostrata and E. prostrata groups,however, the results of the behaviour observation experiment showed that the feeding frequency of the P. clarkii males was higher than that of the females in the L. prostrata and E. prostrata groups. P. clarkii males are more aggressive than the females25, so it is possible that the differences in their behaviour are because the P. clarkii males frequently move, search for food and eat.

All the weeds used in this study were newly grown seedlings. The appearance of weeds changes greatly in different growth stages. Therefore, the conclusions drawn on the basis of the P. clarkii specimens feeding on the weed seedlings in this study are not necessarily applicable to weeds in other growth stages. Freshwater crayfish have a dietary protein requirement of at least 30–35% for optimal growth26. The percentage of crude protein content in dry L. chinensis and E. crusgalli is 8.44% and 11.84%, respectively27,28. Although moulting of P. clarkii was observed in the weed groups in the present study, the P. clarkii specimens in the weed groups died with the extension of culture time, whereas the P. clarkii specimens in the feed group continued to grow well. Obviously, it is impossible for P. clarkii to feed on only weeds in a rice field, as the nutrition in weeds cannot completely meet the requirements for important life history events of P. clarkii, such as moulting and reproduction. However, the high ecological plasticity of P. clarkii makes its use in controlling weeds in rice fields possible. In a rice field, which has high amounts of plant material and low macroinvertebrate diversity, animal food is less important18. Therefore, the use of P. clarkii in controlling weeds in a rice field needs to be studied further.

Conclusions

This study provides direct evidence that crayfish feed on common weeds in paddy fields, suggesting that crayfish may be used to reduce the weed biomass in paddy fields. P. clarkii preferred to eat L. chinensis and hardly fed on E. crusgalli. The mean percentage of daily feed intake to body weight per P. clarkii for L. chinensis was more than 2%.