Mesocosm experiments

Detail in mesocosm experiments 

For the rice-carp mesocosm experiment, five hills of rice plants (one seedling per hill) were planted in each mesocosm. One week after rice transplanting, five fish fry (ca. 10 g each) were added to each mesocosm. A water level of 25 cm was maintained during the experiment, with pH 6.0–6.5, NO₂^- < 0.01 mg L^-1, and NH₄⁺ < 0.2 mg L^-1. No fertilizer was applied. At 4 days after the fry were introduced and on every day of the experiment thereafter, the ¹⁵N-labeled feed was applied twice per day (at 7:00 and 17:00). Throughout the experiment, the total amount of feed applied was 126 g m^-2.

For the rice-crab mesocosm experiment, four hills of rice plants (three rice seedlings per hill) were transplanted into each mesocosm. One week after rice transplanting, 50 juvenile crabs (ca. 1.22 g each) were released into each mesocosm. A water level of 25 cm was maintained during the experiment, with pH 8.0-8.5, NO₂^- < 0.1 mg L^-1, and HN4⁺ < 0.5 mg L^-1. At 3 days after crab release and on every day of the experiment thereafter, the ¹⁵N-labeled feed was applied once per day at 18:00. Throughout the experiment, the total amount of feed applied was 124 g m^-2.

For the rice-turtle mesocosm experiment, four hills of rice plants (three rice seedlings per hill) were transplanted into each mesocosm. One week after rice transplanting, 14 young turtles (ca. 3.5 g each) were released into each mesocosm. A water level of 30 cm was maintained during the experiment, with pH 8.0-8.5, NO₂^- < 0.1 mg L^-1, and NH₄⁺ < 0.5 mg L^-1. Four days after turtle release and on every day of the experiment thereafter, the ¹⁵N-labeled feed was applied twice each day (at 6:00 and 16:00). Throughout the experiment, the total quantity of feed applied was 130 g m^-2.

Method of obtaining soybean powder labeled with¹⁵N

Soybean plants were planted in pots in a green house. On 3 days during the growing period, ¹⁵N-enriched (NH₄)₂SO₄ (99.7% ¹⁵N) was added to the soil at the rate of 5 mg N kg^-1 soil. After they were harvested, soybeans were oven-dried at 65℃ and ground into powder, which was used as the ¹⁵N labeling ingredient in the ¹⁵N-enriched feed. 

Sample collecting and preparing for rice, aquatic animals, and soil

Animal collection and preparation for ¹⁵N. Before and after the experiments, carp, crabs, or turtles were randomly sampled and stored at -20℃. Muscles were peeled from every animal sample, oven dried at 50℃, and separated into two sub-samples. One sub-sample was directly used for ¹⁵N analysis, and the other was used for ¹³C analysis after being pretreated with 2:1 chloroform and methanol for lipid removal.

Biomass and yield. At harvest, the aboveground portions of rice plants on one hill from each mesocosm were separated into grain and straw and were oven-dried at 65℃ to a constant weight. Samples of carp, crabs, or turtles were randomly collected from each mesocosm. These aquatic animal samples were placed in clean water for 48 hours to empty their guts and were stored at -80℃ after being freeze-dried. Soil samples (0-20 cm depth) were collected from each mesocosm before and after the experiment and were air-dried. 

In the dietary reconstruction, the discrimination factor (△) is the difference in δ¹³C or δ¹⁵N values between food sources and food consumers (Peterson and Fry, 1987). In our study, △¹⁵N and △¹³C were 2.73‰ and 1.71‰, respectively, for carp(Guo et al., 2015), and were 2.75‰ and 0.75‰, respectively, for crabs and turtles (Caut et al., 2009).