Temperature difference between screen and various heights in rice field. \*\* Different at *p*<0.01.

Table 3. Temperatures at different locations in rice field and their difference when compared to that in weather station

Plant Temperature for Sterile Alteration of Rice 175

It indicated that height of 20 cm was the location expressing the highest correlation coefficient. The result happened to meet the height of panicle at that time. Thus, 20 cm was

> 40 cm Air temperature

Line model -0.316\*\* -0.410\*\* -0.275\* -0.225\* -0.224\* -0.200 -0.180

Conic model 0.318\*\* 0.424\*\* 0.275\* 0.240\* 0.238\* 0.227\* 0.208\*

Present model 0.853\*\* 0.778\*\* 0.503\*\* 0.314\*\* 0.362\*\* 0.369\*\* 0.265\*

Sample No. 66 85 85 85 85 85 89

Table 4. Correlation of the seed setting rate of Peiai64S and temperature at different heights

Determining fertility sensitive stage was the base to establish a model of fertilitytemperature. The fertility sensitive stage can be determined by distance between the last upper two leaves or can be checked by the length of panicle. However, a method by growth days was popular for its advantage in quantitative analysis (Lu *et al* 2001). The method was given a range of growing days and it will cause difficulty for calculating temperature accumulation (Lu *et al* 2001), since it does not consider the growing difference owing to temperature. The chapter determined the initiative date (total days as well) by effective

>24°C°C effective accumulated temperature Initiative days from heading

<10°C.d 15 10-20°C.d 12 >20°C.d 10

By analyzing the relationship between the fertility and temperature, it was found that the fertility showed a rule of recovery-increase-decrease along with temperature decreasing. Thus, the chapter set the base of the model as: when the temperature decreased to an upper limit (self-fertilized seed setting rate of 0.5%), the TGMS was regarded as fertility, and at the optimum temperature, it showed the highest fertility. When the temperature

60 cm Air temperature 100 cm Air temperature

150 cm Air temperature

Screen temperature

taken as the suitable location for expressing the fertility of TGMS.

20 cm Air temperature

20 cm Stem temperature

\*, \*\* Significant at *p*<0.05 and *p*<0.01, respectively.

accumulated temperature >24. The method was set as:

(Total sensitive days)

**3.3.2 Fertility-temperature model** 

**3.3 Model of fertility-temperature** 

**3.3.1 Fertility sensitive stage** 

Item

in field

Owing to the difference in the underlay surface and growing plant, the temperature at each canopy height was lower than that of the screen of weather station in the four periods. For the periodⅠ, which represented the various weather conditions, the air temperature at height of 150 cm showed 0.69°C lower than the screen temperature of the weather station, which was at the same height. At heights of 100 cm, 60 cm, 40 cm, and 20 cm, owing to the energy absorbing and reflecting, temperature difference was enlarged at the four heights. The average temperature at heights of 100 cm, 60 cm, 40 cm, and 20 cm of periodⅠ was significantly lower than screen one by 1.10°C, 0.83°C, 1.18°C, and 1.63°C, respectively. A tendency of larger difference along with the height increase was seen, and the largest one was seen at 20 cm.

#### **3.2 Fertility sensitive position and its height**

It was reported that developing panicle was the fertility sensitive part of rice (Xu *et al* 1996). Thus, panicle height was important for determining the fertility sensitive position and the water depth for irrigation to regulate the temperature of the fertility sensitive part. The preceding studies proved that the sensitive stage was around the stage of meiosis of pollen mother cells (panicle developing stage IV – VI), and the visible morphological trait was at ±5 cm distance between the last upper two leaves (DL) (Lu *et al*  2001). Fig. 10 shows the correlation between the panicle height, length, and the DL. Result showed that during the stage, panicle height from the base was 15.3-21.9 cm with an average value of 18.5 cm, and meanwhile, the panicle length was 9.7-16.8 cm with an average value of 13.2 cm when DL was ±5 cm. It indicated that height of 20 cm was the suitable location for fertility sensitivity.

Fig. 10. Correlation of distance between the uppermost two leaves and panicle height (A) or length (B) of Peiai64S.

Table 4 shows the correlation coefficients between the self-fertilized seed setting rate and air temperature at various plant heights or screen one with a liner, conic, and present model (See the part '3.3 Model of fertility-temperature'). The result showed that the coefficients between the self-fertilized seed setting rate and stem or air temperature at 20 cm height were the highest and showed significant correlation.

Owing to the difference in the underlay surface and growing plant, the temperature at each canopy height was lower than that of the screen of weather station in the four periods. For the periodⅠ, which represented the various weather conditions, the air temperature at height of 150 cm showed 0.69°C lower than the screen temperature of the weather station, which was at the same height. At heights of 100 cm, 60 cm, 40 cm, and 20 cm, owing to the energy absorbing and reflecting, temperature difference was enlarged at the four heights. The average temperature at heights of 100 cm, 60 cm, 40 cm, and 20 cm of periodⅠ was significantly lower than screen one by 1.10°C, 0.83°C, 1.18°C, and 1.63°C, respectively. A tendency of larger difference along with the height increase was seen, and the largest one

It was reported that developing panicle was the fertility sensitive part of rice (Xu *et al* 1996). Thus, panicle height was important for determining the fertility sensitive position and the water depth for irrigation to regulate the temperature of the fertility sensitive part. The preceding studies proved that the sensitive stage was around the stage of meiosis of pollen mother cells (panicle developing stage IV – VI), and the visible morphological trait was at ±5 cm distance between the last upper two leaves (DL) (Lu *et al*  2001). Fig. 10 shows the correlation between the panicle height, length, and the DL. Result showed that during the stage, panicle height from the base was 15.3-21.9 cm with an average value of 18.5 cm, and meanwhile, the panicle length was 9.7-16.8 cm with an average value of 13.2 cm when DL was ±5 cm. It indicated that height of 20 cm was the

Fig. 10. Correlation of distance between the uppermost two leaves and panicle height (A) or

Table 4 shows the correlation coefficients between the self-fertilized seed setting rate and air temperature at various plant heights or screen one with a liner, conic, and present model (See the part '3.3 Model of fertility-temperature'). The result showed that the coefficients between the self-fertilized seed setting rate and stem or air temperature at 20 cm height

was seen at 20 cm.

**3.2 Fertility sensitive position and its height** 

suitable location for fertility sensitivity.

were the highest and showed significant correlation.

length (B) of Peiai64S.


It indicated that height of 20 cm was the location expressing the highest correlation coefficient. The result happened to meet the height of panicle at that time. Thus, 20 cm was taken as the suitable location for expressing the fertility of TGMS.

\*, \*\* Significant at *p*<0.05 and *p*<0.01, respectively.

Table 4. Correlation of the seed setting rate of Peiai64S and temperature at different heights in field

### **3.3 Model of fertility-temperature**

#### **3.3.1 Fertility sensitive stage**

Determining fertility sensitive stage was the base to establish a model of fertilitytemperature. The fertility sensitive stage can be determined by distance between the last upper two leaves or can be checked by the length of panicle. However, a method by growth days was popular for its advantage in quantitative analysis (Lu *et al* 2001). The method was given a range of growing days and it will cause difficulty for calculating temperature accumulation (Lu *et al* 2001), since it does not consider the growing difference owing to temperature. The chapter determined the initiative date (total days as well) by effective accumulated temperature >24. The method was set as:


### **3.3.2 Fertility-temperature model**

By analyzing the relationship between the fertility and temperature, it was found that the fertility showed a rule of recovery-increase-decrease along with temperature decreasing. Thus, the chapter set the base of the model as: when the temperature decreased to an upper limit (self-fertilized seed setting rate of 0.5%), the TGMS was regarded as fertility, and at the optimum temperature, it showed the highest fertility. When the temperature

Plant Temperature for Sterile Alteration of Rice 177

Location A B TL T0 TH P0 R n

Average 1 0.24 21.6 22.3 22.5 10.82 0.755\*\* 68 Maximum 1 0.55 27.9 28.1 28.2 10.82 0.371\*\* 68 Minimum 1 1.12 17.0 17.7 18.5 10.82 0.964\*\* 34

Average 1 0.37 21.7 22.5 22.8 10.82 0.853\*\* 66 Maximum 1 0.55 28.5 28.7 28.8 10.82 0.470\*\* 66 Minimum 1 1.33 17.0 17.5 18.1 10.82 0.550\*\* 66

Average 1 19.35 22.2 22.4 26.0 10.82 0.586\*\* 64 Maximum 1 0.62 28.0 28.2 28.3 10.82 0.344\*\* 64 Minimum 1 0.63 18.7 18.9 19.0 10.82 0.610\*\* 64

Average 1 1.17 21.9 22.1 22.3 10.82 0.715\*\* 83 Maximum 1 1.50 26.6 26.8 27.1 10.82 0.360\*\* 83 Minimum 1 27.27 17.8 18.0 23.0 10.82 0.659\*\* 76

Average 1 12.14 21.5 21.8 26.0 10.82 0.778\*\* 85 Maximum 1 0.60 26.8 27.0 27.1 10.82 0.356\*\* 85 Minimum 1 5.02 17.7 17.9 19.1 10.82 0.684\*\* 83

Average 1 0.88 22.0 22.3 22.6 10.82 0.491\*\* 86 Maximum 1 0.97 26.7 26.8 26.9 10.82 0.408\*\* 86 Minimum 1 0.56 19.1 19.3 19.4 10.82 0.361\*\* 85

Average 1 0.22 22.9 23.4 23.5 10.82 0.559\*\* 41 Maximum 1 0.67 29.5 29.6 29.7 10.82 0.280 41 Minimum 1 0.55 17.7 17.9 18.0 10.82 0.426\*\* 37

Average 1 12.61 22.8 23.0 26.0 10.82 0.594\*\* 42 Maximum 1 0.34 28.5 28.8 28.9 10.82 0.548\*\* 42 Minimum 1 15.56 17.5 17.8 22.9 10.82 0.686\*\* 41

Average 1 9.15 22.9 23.2 26.0 10.82 0.504\*\* 43 Maximum 1 0.08 27.1 28.3 28.4 10.82 0.223 43 Minimum 1 0.31 18.7 19.0 19.1 10.82 0.623\*\* 42

Table 5. Simulation of stem or air temperature with three parameters and three durations

20 cm Stem temperature (Tp20)

20 cm Air Temperature (Ta20)

150 cm (ck) Air Temperature (TA)

One day

Three days

Five days

One day

Three days

Five days

One day

Three days

Five days

decreased furthermore to a lower limit, the self-fertilized seed setting rate returned to zero.

Thus, equation (8) was established as:

$$P = P\_0 (\frac{T - T\_L}{T\_0 - T\_L})^A (\frac{T\_H - T}{T\_H - T\_0})^B \tag{8}$$

Here, P indicated self-fertilized seed setting rate, while P0 indicated its maximum value. T denoted the averaged temperature of the sensitive stage, and TH, TL indicated the upper and lower limit temperature, respectively, when fertility was zero. T0 denoted the optimum temperature when the TGMS showed maximum fertility. A and B were undetermined parameters and were both>0. Equation (8) shows the following parameters:

The derivative equation:

$$P' = P\_0 \frac{A(T - T\_L)^{A-1} (T\_H - T)^B}{\left(T\_0 - T\_L\right)^A \left(T\_H - T\_0\right)^B} - P\_0 \frac{B(T - T\_L)^A \left(T\_H - T\right)^{B-1}}{\left(T\_0 - T\_L\right)^A \left(T\_H - T\_0\right)^B}$$


#### **3.3.3 Analysis of temperature indices for sterile alteration by the model**

Table 5 shows the simulation of stem or air temperatures with self-fertilized seed setting rate by three parameters as average, maximum, and minimum temperatures, with three durations of 1, 3, 5 days, and by 0.1°C of temperature step length for the simulation. Result showed that there were significant correlations between self-fertilized seed setting rate and the temperatures of 20 cm stem, air temperature or 150 cm air temperature with the three temperatures and three durations. It also implied that the effect of average temperature was better than that of the maximum or minimum temperature, and three-day duration was better than one-day or five-day durations. The result also showed that the stem and air temperatures at 20 cm were better than 150 cm air temperature, and the stem temperature at 20 cm showed the best effect of simulation. Thus, stem temperature at 20 cm was selected as the best simulating parameter for the model.

Equations (9), (10), and (11) are the statistic models of the self-fertilized seed setting rate simulated with the actual self-fertilized seed setting rate and temperatures of 20 cm stem, air, and 150 cm air. With the self-fertilized seed setting rate of 0.5% as initial (upper limit) and that of zero as the lower limit, and with stem temperature at 20 cm for three days, using equation (9), upper limit temperature and lower limit temperature was determined to be 22.8°C and 21.7°C, respectively. When compared with air temperature at 150 cm, there was 1.2°C and 1.1°C lower, respectively. With air temperature at 20 cm for three days, using equation (10), upper limit temperature and lower limit temperature was determined to be 23.2°C and 21.5°C, respectively. When compared with air temperature at 150 cm, there was 0.8°C and 1.3°C lower, respectively.

decreased furthermore to a lower limit, the self-fertilized seed setting rate returned to

*TT T T P P TT T T*

Here, P indicated self-fertilized seed setting rate, while P0 indicated its maximum value. T denoted the averaged temperature of the sensitive stage, and TH, TL indicated the upper and lower limit temperature, respectively, when fertility was zero. T0 denoted the optimum temperature when the TGMS showed maximum fertility. A and B were undetermined

0 0 ( )( ) *L H A B L H*

1 1

*A B AB L H L H A B AB L H L H*

0 00 0 ( ) ( ) ( )( ) ( )( ) ( )( )

*TT T T TT T T* 

i. When P' =0, the optimum temperature T0 was calculated, and the highest self-fertilized

Table 5 shows the simulation of stem or air temperatures with self-fertilized seed setting rate by three parameters as average, maximum, and minimum temperatures, with three durations of 1, 3, 5 days, and by 0.1°C of temperature step length for the simulation. Result showed that there were significant correlations between self-fertilized seed setting rate and the temperatures of 20 cm stem, air temperature or 150 cm air temperature with the three temperatures and three durations. It also implied that the effect of average temperature was better than that of the maximum or minimum temperature, and three-day duration was better than one-day or five-day durations. The result also showed that the stem and air temperatures at 20 cm were better than 150 cm air temperature, and the stem temperature at 20 cm showed the best effect of simulation. Thus, stem temperature at 20 cm was selected as

Equations (9), (10), and (11) are the statistic models of the self-fertilized seed setting rate simulated with the actual self-fertilized seed setting rate and temperatures of 20 cm stem, air, and 150 cm air. With the self-fertilized seed setting rate of 0.5% as initial (upper limit) and that of zero as the lower limit, and with stem temperature at 20 cm for three days, using equation (9), upper limit temperature and lower limit temperature was determined to be 22.8°C and 21.7°C, respectively. When compared with air temperature at 150 cm, there was 1.2°C and 1.1°C lower, respectively. With air temperature at 20 cm for three days, using equation (10), upper limit temperature and lower limit temperature was determined to be 23.2°C and 21.5°C, respectively. When compared with air temperature at 150 cm, there was

*AT T T T BT T T T P P <sup>P</sup>*

(8)

0

parameters and were both>0. Equation (8) shows the following parameters:

0 0

**3.3.3 Analysis of temperature indices for sterile alteration by the model** 

zero.

Thus, equation (8) was established as:

The derivative equation:

seed setting rate, Pmax=P0

ii. When *T T <sup>H</sup>* and *T T <sup>L</sup>* , then *Pmin*=0

the best simulating parameter for the model.

0.8°C and 1.3°C lower, respectively.


Table 5. Simulation of stem or air temperature with three parameters and three durations

Plant Temperature for Sterile Alteration of Rice 179

plant temperature would be increased by 2°C, which was effective for safeguarding the sterility of TGMS (Lu *et al* 2004, Zou *et al* 2005). So far, the forecast of sterile alteration is only based on the temperature information from weather station; there is lack of any study on field microclimate or plant temperature, and especially no research has focused on the temperature scale of plant or air around it. In researches of wheat, some researchers used plant temperature as the parameter for freeze injury and grain growing speed (Feng *et al* 

The chapter put forward a method to conclude the fertility of TGMS by stem or air temperature at 20 cm height, which takes various factors including microclimate and location of field into account. It is more direct and exact than the traditional method. In rice production, we can use it to estimate any field or representative plot of a large field, and to monitor directly the result of regulation for safeguarding seed production in two-line hybrid rice. The technique is: when attacked by lower temperature weather with average temperature lower than 24°C during 5-15 days before TGMS heading, using infrared or thermosensor temperature indicator to determine plant stem temperature at 20 cm height or air temperature around it at 02:00, 08:00, 14:00, 20:00 (or only 08:00 and 20:00) every day. If the averaged value is lower than the line of 22.8°C for stem temperature or air temperature is lower than 23.2°C, it implies that the TGMS will transform its sterility to fertility. For safeguarding its sterility, it is necessary to irrigate by warmer water higher than 25°C, and

The present chapter also established a statistic model for stem or air temperature at 20 cm height. By the inflow and outflow water temperatures of any field, and screen temperature and cloud cover from the local weather station, stem or air temperature at 20 cm height can be concluded. For an application example, if one day, the average temperature and cloud cover was 22°C and 9, respectively, and the actual water temperature of the field was 23°C, by (12) and (13), stem or air temperature at 20 cm height was calculated to be 22.55°C and 22.74°C, respectively, both of which were lower than the above temperature index. By irrigating warmer water from river, the inflow and outflow water temperature were measured as 26°C and 24°C, by (12) and (13), the stem and air temperature at 20 cm height was calculated to be 23.74°C and 24.23°C, respectively, since both are higher than the above

Tp:plant temperature. Tp10, Tp20, Tp30, Tp40 denotes rice stem temperature at plant heights of

Tw-A: temperature difference between water temperature and air temperature at height of

by depth of 15 cm, until the temperature is higher than the above index.

temperature index, it will be effective for safeguarding the sterility of TGMS.

ΔTp: plant temperature difference between water irrigated and non-irrigated. Ta20, Ta40 and Ta: air temperature at heights of 20cm, 40cm and 150cm, respectively.

DL: distance between the last upper two leaves of rice plant.

2000, Liu *et al* 1992).

**4. Abbreviation** 

Tw: water temperature.

LAI: leaf area index.

150cm.

Tin: temperature of inflow water. Tout: temperature of outflow water.

10cm, 20cm, 30cm and 40cm, respectively.

$$T\_{p20} = 10.82 \times (\frac{T - 21.7}{22.5 - 21.7})(\frac{22.8 - T}{22.8 - 22.5})^{0.37} \text{ (\$R = 0.853^{\*\*}\$ : } n = 66\$) \tag{9}$$

$$T\_{a20} = 10.82 \times (\frac{T - 21.5}{21.8 - 21.5})(\frac{26.0 - T}{26.0 - 21.8})^{12.14} \text{ ( $R = 0.778^{\*\*}$  :  $n = 85$ )}\tag{10}$$

$$T\_A = 10.82 \times (\frac{T - 22.8}{23.0 - 22.8})(\frac{26.0 - T}{26.0 - 23.0})^{12.61} \text{ (R = 0.594^{\*\*} ; ; n = 42)}\tag{11}$$

#### **3.4 Stem and air temperatures at 20 cm**

Considering energy exchange and balance, stem and air temperatures in rice canopy were connected with two energy sources: solar radiant energy and water heat energy. To avoid trouble in actual operation, the chapter established effective statistic models as equations (12) and (13), which included the screen temperature (TA), cloud cover (N, from 1 to 10), and the water temperature at inflow (Tin) and outflow (Tout). Only by determining Tin and Tout, stem or air temperature at 20 cm could be calculated by screen temperature and cloud cover, which were both offered by the local weather station. It was helpful to estimate the damage of lower temperature weather and the effects of the adjusting measures in actual seed production.

$$T\_{p.20} = \frac{N}{15.2} \frac{(T\_{in} + T\_{out})}{2} + (1 - \frac{N}{15.2})T\_A \text{ ( $R = 0.812^{\*\*}$ ,  $n = 46$ )}\tag{12}$$

$$T\_{a20} = (1 - \frac{N}{34.9})\frac{(T\_{in} + T\_{out})}{2} + \frac{N}{34.9}T\_A \text{ ( $R = 0.975$ ",  $n = 46$ )}\tag{13}$$

#### **3.5 Techniques by used plant temperature to safeguard sterile of Peiai64S**

Peiai64S was a widely used TGMS in two-line hybrid rice breeding. Its fertility was controlled by temperature during its sensitive stage. Its sterility usually fluctuated owing to the frequent fluctuation caused by monsoon, which resulted in damage to the seed purity of seed production in China southern rice area (Lu *et al* 2001, Yao *et al* 1995). The first so-called super hybrid rice in China, Liangyoupeijiu, which was released by the authors' research group, was popularized over 7 million ha, and has been the major planted rice for its largest planting area of China from 2002. However, in lower reaches of Yangtze River, in the past five years, there was twice lower temperature weather (daily averaged temperature for three days lower than 24°C) in August, during which was the sterility sensitive period of TGMS, that caused damage to the seed production of Liangyoupeijiu and other two-line hybrid rice. It was a hidden problem for seed production of two-line hybrid rice. In the seed production practices of Liangyoupeijiu, it was found that when attacked by lower temperature weather, the seed purity exhibited a large difference even if the TGMS was in a same weather condition, owing to the individual differences at landform and water treatment and so on. Researches showed that the fertility of rice TGMS was affected directly by plant temperature, which was infected by the microclimate of the field (Hu *et al* 2006). When attacked by lower temperature weather, by irrigating warmer water from river or deep pool,

Considering energy exchange and balance, stem and air temperatures in rice canopy were connected with two energy sources: solar radiant energy and water heat energy. To avoid trouble in actual operation, the chapter established effective statistic models as equations (12) and (13), which included the screen temperature (TA), cloud cover (N, from 1 to 10), and the water temperature at inflow (Tin) and outflow (Tout). Only by determining Tin and Tout, stem or air temperature at 20 cm could be calculated by screen temperature and cloud cover, which were both offered by the local weather station. It was helpful to estimate the damage of lower temperature weather and the effects of the adjusting measures in actual seed

21.7 22.8 10.82 ( )( ) 22.5 21.7 22.8 22.5 *<sup>p</sup>*

21.5 26.0 10.82 ( )( ) 21.8 21.5 26.0 21.8 *<sup>a</sup>*

( ) (1 ) 15.2 2 15.2 *in out p A*

( ) (1 ) 34.9 2 34.9 *in out a A*

**3.5 Techniques by used plant temperature to safeguard sterile of Peiai64S** 

Peiai64S was a widely used TGMS in two-line hybrid rice breeding. Its fertility was controlled by temperature during its sensitive stage. Its sterility usually fluctuated owing to the frequent fluctuation caused by monsoon, which resulted in damage to the seed purity of seed production in China southern rice area (Lu *et al* 2001, Yao *et al* 1995). The first so-called super hybrid rice in China, Liangyoupeijiu, which was released by the authors' research group, was popularized over 7 million ha, and has been the major planted rice for its largest planting area of China from 2002. However, in lower reaches of Yangtze River, in the past five years, there was twice lower temperature weather (daily averaged temperature for three days lower than 24°C) in August, during which was the sterility sensitive period of TGMS, that caused damage to the seed production of Liangyoupeijiu and other two-line hybrid rice. It was a hidden problem for seed production of two-line hybrid rice. In the seed production practices of Liangyoupeijiu, it was found that when attacked by lower temperature weather, the seed purity exhibited a large difference even if the TGMS was in a same weather condition, owing to the individual differences at landform and water treatment and so on. Researches showed that the fertility of rice TGMS was affected directly by plant temperature, which was infected by the microclimate of the field (Hu *et al* 2006). When attacked by lower temperature weather, by irrigating warmer water from river or deep pool,

12.14

22.8 26.0 12.61 10.82 ( )( ) 23.0 22.8 26.0 23.0 *<sup>A</sup>*

20

20

**3.4 Stem and air temperatures at 20 cm** 

20

20

production.

0.37

*T T <sup>T</sup>* (*R*=0.853\*\*;*n*=66) (9)

*T T <sup>T</sup>* (*R*=0.778\*\*;*n*=85) (10)

*T T <sup>T</sup>* (*R*= 0.594\*\*;*n*=42) (11)

*N N T T T T* (*R*=0.812\*\*, *n*=46) (12)

*N N T T T T* (*R*=0.975\*\*, *n*=46) (13)

plant temperature would be increased by 2°C, which was effective for safeguarding the sterility of TGMS (Lu *et al* 2004, Zou *et al* 2005). So far, the forecast of sterile alteration is only based on the temperature information from weather station; there is lack of any study on field microclimate or plant temperature, and especially no research has focused on the temperature scale of plant or air around it. In researches of wheat, some researchers used plant temperature as the parameter for freeze injury and grain growing speed (Feng *et al*  2000, Liu *et al* 1992).

The chapter put forward a method to conclude the fertility of TGMS by stem or air temperature at 20 cm height, which takes various factors including microclimate and location of field into account. It is more direct and exact than the traditional method. In rice production, we can use it to estimate any field or representative plot of a large field, and to monitor directly the result of regulation for safeguarding seed production in two-line hybrid rice. The technique is: when attacked by lower temperature weather with average temperature lower than 24°C during 5-15 days before TGMS heading, using infrared or thermosensor temperature indicator to determine plant stem temperature at 20 cm height or air temperature around it at 02:00, 08:00, 14:00, 20:00 (or only 08:00 and 20:00) every day. If the averaged value is lower than the line of 22.8°C for stem temperature or air temperature is lower than 23.2°C, it implies that the TGMS will transform its sterility to fertility. For safeguarding its sterility, it is necessary to irrigate by warmer water higher than 25°C, and by depth of 15 cm, until the temperature is higher than the above index.

The present chapter also established a statistic model for stem or air temperature at 20 cm height. By the inflow and outflow water temperatures of any field, and screen temperature and cloud cover from the local weather station, stem or air temperature at 20 cm height can be concluded. For an application example, if one day, the average temperature and cloud cover was 22°C and 9, respectively, and the actual water temperature of the field was 23°C, by (12) and (13), stem or air temperature at 20 cm height was calculated to be 22.55°C and 22.74°C, respectively, both of which were lower than the above temperature index. By irrigating warmer water from river, the inflow and outflow water temperature were measured as 26°C and 24°C, by (12) and (13), the stem and air temperature at 20 cm height was calculated to be 23.74°C and 24.23°C, respectively, since both are higher than the above temperature index, it will be effective for safeguarding the sterility of TGMS.
