Document(s) 29 of 41

The rice–azolla–fish system features an ecological three-dimensional approach to agriculture. Rice (the main crop), azolla, and fish are combined in a symbiotic complex. Rice is planted on the ridge, azolla grown on the water surface, and fish raised in the water. This new farming system was studied between 1984 and 1987.

Experiments were conducted on two plots of land, each with an area of 0.2 ha and a 6-m2 fish pit. Double-cropping rice was planted in one plot, single-cropping rice in the other. Each plot was randomly arranged with three replications and planted with local high-yielding varieties or hybrid combinations of rice and azolla (Azolla filiculoides or A. caroliniana) (6000 kg/ha). Fish species selected for mixed raising were grass carp (Ctenopharyngodon idella), tilapia (Oreochromis nilotica), lotus carp, dull carp, and Hunan crucian carp (Carassius auratus). Growth of rice, azolla, and fish was recorded. The nutrient content of azolla and fish dung was determined by standard methods of analysis.

Results and DiscussionProportion of Ridges to Ditches

Ridge width had a bearing on the yields of rice, azolla, and fish (Table 28). The yield of double-cropped rice on the 106-cm wide ridge (13 834.5 kg/ha) was 4.1% higher than in the control (conventional planting), 5.1% higher than on the 53-cm ridge, and 2.5% higher than on the 80-cm ridge. The yield of azolla grown on the water surface between the 53-cm ridges (72 930 kg/ha), was 70.3% higher than in control, 13.4% higher than between the 80-cm ridges, and 55.6% higher than between the 106-cm ridges.

Table 28. Effect of different ridge widths on yields of rice, azolla, and fish.
Ridge width (cm)	Rice (kg/ha)			Azolla (kg/ha)	Fish (kg/ha)
	Early rice	Late rice	Total
53	6463.5	6697.5	13 161.0	72 931	841.5
80	6750.0	7651.5	14 401.5	64 320	736.5
106	7072.5	6762.0	13 834.5	46 875	676.5
Control	6675.0	6616.5	13 291.5	42 825	436.5

Yields of fresh fish increased as ridge width decreased. The yield of fresh fish with the 53-cm ridges (841.5 kg/ha) was 89.3% higher than the control, 12.2% higher than with the 80-cm ridges, and 22.2% higher than with the 106-cm ridges. Therefore, wide ridges favoured rice yields, and the narrow ridge favoured growth of azolla and fish.

When ridge width was uniform and ditch width was varied, rice yields differed. With a 40-cm ditch, rice yield was 10 462.5 kg/ha, the same as in conventional planting, 3.5% more than with 46-cm ditches, and 11.4% more than with 53-cm ditches. Yields of azolla and fish were directly related to increases in ditch width (Table 29).

Table 29. Effect of different ditch widths on yields of rice,* azolla, and fish.
Ditch width (cm)	Rice (kg/ha)			Azolla (kg/ha)	Fish (kg/ha)
	Early rice	Late rice	Total
40	4965.0	5497.5	10 462.5	56 242	613.5
46	4762.5	5347.5	10 110.0	50 258	702.0
106	4425.0	4965.0	9 390.0	61 995	784.5
Control	5175.0	5385.0	10 560.0	—	—
* Early rice variety (2106); late rice variety (Wei-You 35); ridge width 53 cm.

Production practices have proven that to determine the proportion of ridges to ditches for to obtain high rice yields, it is necessary to consider soil fertility, the characteristics of the rice varieties, whether the proportions favour growth of azolla and fish, and whether rice, azolla, and fish are well coordinated. In a ricefield in which rice is planted on the ridges, azolla is grown on the water surface, and fish are raised in fish pits, the water area of the fish pit should be 5–10% of the total area of the ricefield.

Best results are obtained if the ridge width is 53–106 cm and the ditch width is 40 cm. On each ridge, 4–8 rows of rice are planted 13–16.5 cm apart to obtain 300 000–375 000 hills of seedlings per hectare. If hybrid rice varieties are used, plant spacing should be 16.5–20 cm with 3–7 rows on each ridge to obtain 225 000–300 000 hills of seedling per hectare.

Planting Rice During Different Seasons

When ordinary rice varieties were planted in the early crop season and hybrid rice varieties were planted in the late crop season, yearly rice yield was 12925.5 kg/ha or 1405.5 kg less than that obtained from two crops of hybrid rice (Table 30). However, there were no differences in the yields of azolla and fish between ordinary rice + hybrid rice and hybrid rice + hybrid rice. To increase rice yields, it is important to determine the proper time for planting hybrid rice varieties in the early and late crop seasons.

Table 30. Effect of the proper arrangement of rice varieties on rice yield.
Treatment*		Yield (kg/ha)
		Early Rice	Late Rice	Total
I	Ordinary rice in early season
	Rice–azolla–fish	6364.8	6567.4	12 932.2
	Hybrid rice in late season
	Control (rice only)	6932.6	6092.1	13 024.7
II	Two crops of hybrid rice
	Rice–azolla–fish	7291.6	7039.6	14 331.2
	Control (rice only)	7613.4	7216.5	14 829.9
* Ridge width 53 cm; ditch width 40 cm. Rice varieties used in treatment I: early rice Xiangzao indica I; late rice Wei-You 6. Rice varieties used in treatment II: early rice Wei-You 49; late rice Wei-You 64.

When early- or medium-maturing rice varieties (growth duration 100–110 days) are used as the early crop, the rice varieties used in the late crop season should be late-maturing (growth duration about 120 days). If late-maturing rice varieties (growth duration 115–120 days) are used as the early crop, the rice varieties used in the late crop season should be early or medium-maturing (growth duration 100–110 days). When two crops of hybrid rice are grown in a year, the hybrid combinations used in the early and late seasons should be all early or medium-maturing (growth duration about 110 days). If a single crop of medium rice is planted, ordinary rice varieties or hybrid combinations that are late maturing (growth duration 130–135 days) should be used. Conditions of location, cultivation practice, and soil fertility should also be considered.

Yields of Fresh Azolla

Annual yields of fresh azolla grown in ridged ricefields that use the rice–azolla–fish system can reach 113 877–132 235 kg/ha. Yields of different azolla species vary. The yield of A. filiculoides (132 235 kg/ha) was 15.5% higher than that of A. caroliniana. In a mixed culture of A. filiculoides and A. caroliniana, yield (124 417 kg/ha) was between the yield of pure cultures of the two species.

Yields of azolla also varied with different locations (Table 31). In winter (8 November–16 March), A. filiculoides propagated more rapidly in the hilly region of Hunan and the yield of fresh azolla reached 47 010 kg/ha. On average, yields doubled in 18.8 days. Yield of A. filiculoides were 59.7% greater than yield of A. caroliniana. Mixed cultures of these two azolla species, on average, doubled yields in 19.1 days.

Table 31. The speed of propagation of different Azolla species in different locations.
Location	Winter			Spring			Growing between ridges			Summer			Autumn
	I*	II	III	I	II	III	I	II	III	I	II	III	I	II	III
A. Changsha
A. filiculoides	47 010	6.3	18.8	33 825	4.5	8.2	33 206	4.4	9.0	4808	0.7	43.9	—	—	—
A. caroliniana	25 380	3.9	30.1	23 850	3.2	—	—	—	—	—	—	—	—	—	—
Ux mixed culture	46 410	6.2	19.1	25 800	3.4	10.8	24 144	3.3	12.4	6162	0.8	35.4	—	—	—
B. Nan County
A. filiculoides	—	—	—	18 998	2.5	16.2	26 249	3.5	12.3	1245	0.2	88.2	26 400	3.5	14.5
A. caroliniana	—	—	—	15 933	2.1	19.3	23 000	3.1	14.0	17 499	2.3	17.6	28 350	3.8	15.6
Ux mixed culture	—	—	—	14 490	1.9	21.2	20 123	2.7	16.0	6162	0.8	35.4	28 073	3.7	15.8
C. Guidong
A. filiculoides	—	—	—	36 449	4.9	8.3	42 800	5.7	7.0	29 291	3.9	11.5	—	—	—
A. caroliniana	—	—	—	21 341	2.8	14.1	43 187	5.8	7.0	60 110	8.0	5.2	—	—	—
Ux mixed culture	—	—	—	28 605	3.8	10.5	50 768	6.8	5.9	43 715	5.8	7.7	—	—	—
* I=Yield (kg/ha); II=Multiplying (-fold); III=Period of multiplying (days); A=the hilly region in middle Hunan; B=the region along Dongting Lake in Northern Hunan; and C=the mountain region in southern Hunan.

During the spring propagation stage in Changsha, in the hilly region of Hunan, A. filiculoides also propagated faster than A. caroliniana, and the propagation rate of mixed cultures was between the rate for pure cultures. In the region along Dongting Lake in northern Hunan, the yield of A. filiculoides was higher than A. caroliniana, and the yield of the mixed culture was lowest. Differences in yield were related to the slow rise in temperature in the spring. In the mountainous region of southern Hunan, the propagation rate of different azolla species was the same as in middle Hunan because the temperature rose rapidly in early spring.

When azolla was grown between ridges, yields of fresh azolla in the mountainous region of southern Hunan was highest, followed by middle Hunan and northern Hunan. The rapid temperature rise between May–June in northern Hunan affected the propagation rate of azolla. In southern Hunan, the day temperature was high and the difference in temperature between day and night was great, which favoured growth of azolla.

When azolla was cultured on the ridge to over-summer, no matter where it was grown (in middle, northern, or southern Hunan), the propagation rate of A. caroliniana was highest, A. filiculoides was lowest, and the mixed culture was intermediate. These results reflect the fact that A. filiculoides is unable to tolerate high temperatures. Therefore, in winter or spring, it is better to grow A. filiculoides, which can tolerate low temperatures; and in summer or autumn, it is better to grow A. caroliniana, which can tolerate high temperatures. To compensate for deficiencies in each species, a mixed culture is recommended.

Yields of Fish

Effect on azolla. Four fish species (grass carp, tilapia, crucian carp, and lotus carp) were raised with azolla for 110–112 days. The fish species best suited to the rice–azolla–fish system are grass carp and tilapia (Table 32). Both like to eat azolla and adapt easily to the ricefield environment. The omnivorous crucian carp and lotus carp (which are benthic and planktivorous feeders) can be raised in the ricefield in lower numbers (Table 5). Grass carp and tilapia eat over 60% of their body weight in azolla each day; whereas, crucian carp and lotus carp eat about 8% of their body weight in azolla.

Table 32. The total amount of azolla eaten by four fish species and the feed conversion efficiency (CF) (plot area 66 m²; water depth 1.2 m; 30 fish stocked; 30 fish harvested 31 July).
Fish	Stocking date	Body weight (g/fish)	Body weight (g/fish)	Days raised	Survival rate (%)	Body weight increase (g/fish)	Weight increase (%)	Azolla eaten (g)	CF
GC	9 April	54.7	228.7	112	100	174	5220	25 590	49.0
CC	9 April	75.0	110.8	112	100	35.9	1075	33 550	31.2
TL	21 April	24.7	163.1	100	100	138.7	4162	217 100	52.2
LC	9 April	96.8	92.2	112	76.7	4.6	104.2	17 810	0
Note: GC, grass carp; CC crucian carp; TL, tilapia; LC, lotus carp.

Of the four species, grass carp and tilapia had the highest growth rate. Their body weights increased 4.2 and 6.2 times, respectively. The body weights of crucian carp tended to decrease when they grew to a certain stage, and when lotus carp were fed only azolla, their body weight decreased. The digestibility of A. filiculoides (18.1%) was higher than A. caroliniana (12.9%) for grass carp and tilapia, respectively. Therefore, A. filiculoides is a better food for fish than A. caroliniana.

If it is assumed that fish yield was 750 kg/ha and that the amount of azolla eaten per gram of fish daily was 0.605 g/day, the amount of dung excreted by the fish would be 52.9% and it would contained 2.7% N. Therefore, when fish are raised in a ricefield for 100 days, they provide the ricefield with 85.5–103.5 kg N per hectare.

Effect of different fish species. When fingerlings bred in spring were raised, yields of rice and azolla were correlated negatively with the proportions of grass carp and tilapia, but yields of fish were correlated positively with the proportions of grass carp and tilapia. To fully use the azolla and the natural resources of the ricefield, the proportion of fish raised should be 60–70% grass-eating grass carp and tilapia and 30–40% omnivorous crucian carp and lotus carp (Table 33).

Table 33. Effect of the proportion of different fish species on yields of rice, azolla and fish.*
Treatment	Species	Yield of rice (kg/ha)			Yield of azolla (kg/ha)	Yield of fish (kg/ha)
		Early	Late	Total
I	Grass carp 45%	4636.5	6600.0	11 236.5	47 670	628.5
	Tilapia 25%
	Lotus carp 15%
	Crucian carp 15%
II	Grass carp 25%	4761.0	6529.5	11 290.5	51 975	633.0
	Tilapia 45%
	Crucian carp 15%
	Lotus carp 15%
III	Crucian carp 35%	4458.0	6301.5	10 759.5	59 700	568.5
	Lotus carp 35%
	Grass carp 15%
	Tilapia 15%
* 22 500 fingerlings bred in the spring are raised per hectare; rice variety used in early season was Zhefu 802; hybrid rice variety used in late season was Wei-You 6; ridge width 53 cm; ditch width 40 cm.

Effect of stocking density. When fingerlings bred in spring were raised, the yields of fish were correlated positively with stocking density. With 30 000 fingerlings/ha, yield of fish was 630 kg, which was 15% higher than the yield obtained from 15 000 fingerlings/ha and 5.6% higher than from 22 500 fingerlings/ha. Therefore, stocking density has an important bearing on fish yields.

Survival rate of fingerlings (e.g., grass carp) was correlated negatively with stocking density. At a density of 15 000 fingerlings/ha, survival rate was 3% higher than at 30 000 fingerlings/ha. Increases in body weight of fish followed a similar trend. When stocking density was 15 000 fingerlings/ha, body weight was 11.4 g heavier than at 22 500 fingerlings/ha and 20.6 g heavier than at 30 000 fingerlings/ha (Table 34).

Table 34. The survival rate of fingerlings raised in different densities and weight increased.
Species	Fingerlings (per ha)	No. fish raised (per ha)	No. fish harvest (per ha)	Survival rate (%)	Avg body weight (g)	Yield (kg)	Total yield (kg)	Total yield (kg/ha)
Grass carp	15 000	2220	1755	79.1	53.4	93.0	201.5	535.5
Tilapia		2220	1095	49.3	45.2	49.5
Lotus carp		555	450	81.1	61.7	27.0
Crucian carp		555	450	86.5	64.1	31.5
Grass carp	22 500	3330	2535	76.1	62.0	106.5	202.5	592.5
Tilapia		3330	540	16.2	47.2	25.5
Lotus carp		840	630	75.0	54.8	34.5
Crucian carp		840	705	83.9	51.1	36.0
Grass carp	30 000	4440	2925	65.9	32.8	96.0	229.5	630.0
Tilapia		4440	1110	25.0	43.2	48.0
Lotus carp		1110	615	55.4	56.1	34.5
Crucian carp		1110	1005	90.5	50.8	51.0

Effects of pesticides. In general, pesticides are applied in the rice–azolla–fish system to control rice pests and diseases. The routine doses of pesticides such as methamidophos, dimethoat, dichlorphos, chlordimeform, trichlorfon, MIPC, and kasugamysin did not harm grass carp, tilapia, crucian carp, and lotus carp. Malathion and EBP were safe for crucian carp and dull carp, but had lethal effects on tilapia. Phenthoate was harmful to all the fish tested. The toxicity of various pesticides to fish was in the order: kasugamucin < methamidophos < trichlorfon < dimethoat < chlordimeform < tetra chlorvinphos < dichlorphos < malathion < phenthoate (Table 35).

Table 35. The safe concentration of different pesticides to four fish species in 72 hours.
Pesticides	Crucian carp		Tilapia		Dull carp		Lotus carp
	I*	II	I	II	I	II	I	II
Methamidophos	23–30	110	18–22	50	23–27	78.5	18–20	79.5
Dimethoat	25–30	53	23–26	21.2			23–25	53
4049	21–24	6.9	22–25	2.1	23–27	17.6	23–26	10.6
Dichlorphos	25–28	12.9	24–26	9.2	24–30	11.7	22–23	10.7
Tetrachlorvinphos	25–28	25.2	21–23	31.5			21–24	31.5
Trichlorfon	25–28	50	22–24	8.3			21–24	25
Chlordimeform	21–24	55.5	22–24	22.2	23–27	92.5	23–27	22.2
MIPC	23–24	1250	23–30	420	23–27	416	23–25	665
EBP	25–28	6.1	24–26	3			22–27	9
Kasugamucin	21–24	4590	22–26	2750	23–27	3500	23–25	4440
* I = water temperature ( C); II = safe concentration (ppm).

Conclusion

In a rice–azolla–fish system that includes ridge culture, the ricefield should be constructed with: ridges 53–106 cm wide, ridge ditches 40 cm wide and 20–25 cm deep, a main ditch (50 cm wide and 50 cm deep) in the centre of the field, and deep ditch (50 cm wide and 50 cm deep) surrounding the ricefield. A fish pit (80–100 cm deep) should occupy 3–10% of the total area of the ricefield. Of the fingerlings raised in the pit, 3–5% are bred in spring and 8–10% are overwintering fingerlings. Soybean and melon can also be planted around the pit.

High-yielding rice varieties are used. Two crops of hybrid rice are planted with wide row spacing and narrow plant spacing to increase the border effect and enhance the use of light. In a ricefield of ordinary rice, 300 000–375 000 hills of seedlings per hectare (5–7 seedlings/hill) are planted; if hybrid rice is used, 225 000–300 000 hills of seedlings per hectare (1–2 seedlings/hill) are planted. More P and K fertilizers and less N fertilizer are applied in deep placement. Late rice seedlings can be transplanted without tillage. Attention must be paid to field management. The water level on the ridge surface should be regulated according to the growth stage of the rice.

Low-temperature tolerant A. filiculoides and high-temperature-tolerant A. caroliniana can be grown. In general, 300–500 kg of A. filiculoides are planted in the field in early or middle March, and 200–400 kg of A. caroliniana are planted 7 days after transplanting early or medium rice. Mixed culture of these two azolla species is possible.

Suitable fish species are grass carp, tilapia, crucian carp, and lotus carp. All four species like to eat azolla, grow rapidly, and adapt themselves to the ricefield environment. The fingerlings (grass carp and tilapia 60–70%; crucian carp and lotus carp 30–40%) are released into the ricefield in early May. Generally, 6000–12 000 overwintered fingerlings or 30 000–45 000 fingerlings bred in the spring are raised per hectare.

Routine dosages of common pesticides are safe to these four fish species. However, malathion and EBP are lethal to tilapia, and phenthoate is harmful to all of the fish.

Document(s) 29 of 41