In Tanzania rainfall scarcity during the vegetative or reproductive growth phases of rainfed rice has been reported to cause yield losses under farmers’ field condition, where they depend totally on rain-fed cultivation (URT, 2014). In that way varieties which are tolerant to vegetative and/ or reproductive moisture stresses are required. However, little is known on the critical growth stages of the selected rice varieties that need more moisture for attaining their yield potentials, and the information on their drought tolerance is not well known in Tanzanian conditions. Therefore, the present study analyzed the effects of moisture stresses during the vegetative and reproductive growth stages on growth and yields of selected rice varieties for establishing the most drought tolerant varieties among the selected rice varieties in each ecosystem. Critical growth stages of each variety that demands more moisture for achieving maximum productivity were also analyzed.

Experimental design and treatment

The experiment was carried out under field conditions in a split  plot design layout, in which the main plots were the stages at which irrigation withdrawals were induced and the sub-plots were the selected rice varieties. Three upland rice varieties (NERICA 1, NERICA 2, and NERICA 4) were grown in an upland rainfed condition, while three lowland rice varieties (TXD 306, Tai and Komboka) were grown under lowland rainfed condition. For both the upland and lowland rice varieties, five lines each with six hills at the spacing of 20cm x 20cm were used; the treatment was 3 varieties x 3 irrigation withdrawals in 3 replications. Five seeds were sown directly per hill.  After germination the seedlings were reduced to one plant per hill. The soil fertility before the experiment was analysed and summarized in Table 1.  Table 1 shows slight differences among upland and lowland fields in terms of soil texture, PH, EC, OC, Total N, available P, and K cation. All the varieties and the two sites were fertilized with the same fertilizer rate of 80kgN ha^-1 , applied during sowing and all other field managements were maintained similarly.

Moisture stress management

All plots were irrigated uniformly to field capacity up to 30 days after seedling emergence. Then, the irrigation water was withdrawn to create moisture stresses at the vegetative growth stage; that is, from 31 to 52 days after seedling emergence, and at the reproductive growth phase the irrigation water withdrawals started from the flowering initiation stage; that is, 52 to 71 days after seedling emergence. The soil moisture content of plots was recorded at the beginning and at the end of stress period using gravimetrical method; soil samples at 3 soil level depths in each treatment (0-5 cm, 6-10 cm and 11-15 cm) were taken. The fresh soil samples were weighed and sun dried to constant weight. Then the samples were re-weighed, and the weights of dry soil samples were subtracted from the weights of fresh soil samples to obtain the weights of moisture in the soil.

Data recording and procedures

For both upland and lowland field trials, the recorded parameters included:  Number of tillers per plant and plant heights, which were recorded at harvest stage by physical counting.  The plant height (cm) was measured using a metre ruler; shoot dry weight (SDW) at maturity stage was recorded using a weighing balance after oven drying at 80 ºC to a constant dry weight.

Relative water contents RWC (%) of leaves

Relative leaf water content (RWC %) was recorded at the end of each respective water withdrawals.  The leaf water content was calibrated using a gravimetrical method where by fresh leaves of one gram were harvested for each treatment and weighed to get the fresh weight (Wf). The leaf disks were then placed in a test tube containing  distilled  water  for  24 h at  room  temperature to get the turgid weight (Wt); subsequently the disks were dried in an oven at 80 °C until a constant weight was obtained to get the oven dry weight (Wd). The relative water content (RWC) in leaves was calculated using the formula by Karrou and Maranville (1995) and Coombs et al. (1985) as follows:   

During moisture induced stress, leaf drying rate and leaf rolling were evaluated using an IRRI Standard Evaluation System (SES) for rice (2014). The values are as follows; 0 = healthy leaves or having no symptoms; 1= leaves starting to fold (shallow v-shape) or tip slightly drying; 3= leaves folding (deep v-shape) or tip drying extended up to ¼ length in most leaves; 5 = leaves fully cupped (u-shape) or ¼ to ½ of all leaves fully dried.

The yield and yield components

The grain yield and yield components were measured at maturity (harvest). An area of about 1 m² was sampled for upland field trial; while in the lowland trial 3 hills from each water regime treatment were harvested for yield and yield components analyses. Plants were cut 4 cm above the ground and sun dried for 3 days to get the total biomass weights above the ground then threshed to get the grain only. The straws were separately dried at 80°C until a constant dry weight was attained. The number of panicles per plant m^-2, number of fertile and sterile spikelets per panicle and 1000 grains were recorded by physical counting from the threshed grains. Then 1000 grains weight were measured to get the weight of 1000 grains at 14% moisture content by the procedures described by Gomez (1972). Panicle length (cm) was measured using metre ruler. The grain yields of the selected rice varieties were obtained from the relationship by Yoshida (1981) as follows:

GY= (P x SP x FS x 1000GW x 10^-5)                                             (2)   

Where, GY=grain yield (tha^-1), P= number of panicles (m^-2), SP= number of spikelets per panicle, FS=    percentage filled spikelet or grain and GW= 1000-grain weight (g)

And the harvest index was calculated using the relationship by Fageria et al. (2011) as follows,

Statistical data analysis

The obtained data in both trials were subjected to analysis of variance (ANOVA) using Genstat (2011) and Excel (Microsoft). The irrigation and variety treatments means were separated using Tukey’s significance difference test at 5% level.

Relative leaf water content (RWC %) 

There was a significant decrease in leaf water contents at the end of irrigation withdrawals (Figure 1).   A  significant difference (P £ 0.05) among the varieties, moisture stress treatment and stages of water induced stress was observed. Moisture stress caused a significant reduction in leaf water content; the highest reduction among the varieties occurred during reproductive growth stage compared to vegetative growth stages (Figure 1). The upland rice varieties NERICA1, NERICA2 and NERICA4 recorded the highest leaf water content as compared to the lowland rice varieties Komboka, Tai and TXD306 when subjected to moisture stress treatment during vegetative and reproductive stages (Figure 1).

Observation on leaf rolling and leaf drying

Table 2 shows that NERICA1, NERICA2, NERICA4 and Tai had the highest degree of rolling ability compared to the lowland rice varieties Komboka and TXD306 (Table 2). However, there were no differences in the drying rate among the tested varieties at vegetative and reproductive moisture stress inductions (Table 2).

Soil moisture condition during stress

The soil moisture contents during the vegetative and reproductive moisture stress in the fields decreased significantly (Figure 2). At vegetative stress the soil moisture contents at the depth of 0-5 cm decreased significantly from 100 to 19.4%, while at reproductive stress soil moisture decreased to 1.2%.  At the depth of 6-10 cm, the soil water content decreased significantly from 100 to 21.4% at vegetative stress, and at reproductive stress the soil moisture decreased from 100 to 7.9%. However, at the depth of 11-15 cm, the soil moisture content decreased significantly from 100 to 22.3% at vegetative stress, while at reproductive moisture stress the soil moisture contents decreased from 100 to 12.2% (Figure 2). As a result, there was a reduction in growth and yields of all varieties with varying intensity.

There was a significant decrease in leaf water content as moisture stress treatments took effect (Figure 1). These findings are in line with the findings of Cruz et al., (1986) in rice. Decline in leaf water content in our study may be attributed to loss through evapotranspiration and decreased water uptake by roots when the soil water was limiting (Figure 2). As a result, the rice growth and productivity reduced in all varieties. These findings are in line with the observation of Fukai et al. (1995, 1999b), who reported that in moisture stress conditions as water is rapidly lost from the soil surface layers, plant growth and productivity are restricted through reduced availability of water and nutrients. Based on Sah and Zamora (2005), relative water content is an important measure of plant water status.

The water content in leaf relative to maximum amount that the leaf can take under full turgidity was considered as suitable gauge of normal tissue physiological functioning and growth processes (Sikuku et al., 2012). In this study, moisture stress at reproductive growth caused higher reduction in leaf water content compared to moisture stress at vegetative growth stage (Figure 1). This situation constrained the growth and plant function, which was reflected in decreasing number of effective tillers (Figure 4), plant height (Figure 5), shoot dry weights (Figure 6) and grain yields and yield components of all selected rice varieties, with different declining intensity (Table 3). Therefore, higher relative moisture contents in leaves are crucial for suitable growth and function of plants. 

In the present study, varietal differences in relative leaf water content were significant between the upland and lowland rice varieties at vegetative and reproductive moisture stress treatments (Figure 1). NERICA2 and NERICA4 had the highest moisture contents in leaves among the upland rice varieties, while NERICA1 showed the lowest leaf water contents. These results may imply that  NERICA2   and   NERICA4  had  the  highest  stress tolerance characteristics toward moisture stress at vegetative growth stages (Figure 1).  Tai variety had significantly high leaf moisture content among the tested lowland rice varieties and hence was the most tolerant than others in lowland ecosystem. The observations from the present study are in line with the findings of Sinclair and Ludlow (1985) under moisture deficit conditions, who found that the varieties that are tolerant to drought have more relative water content at any stage of stressing. They suggested that high leaf relative water content can be employed in selecting high yielding varieties that uphold cell turgidity under moisture stress and confer relative high grain yield.  Moreover, during stress the upland rice varieties NERICA1, NERICA 2, NERICA4 and lowland rice Tai varieties had the highest rolling ability compared to the lowland rice varieties Komboka, and TXD306 (Table 2). Leaf rolling characteristics in rice minimize evaporative water loss through leaf surfaces (O’Toole et al., 1979), and consequently cause a high degree of tolerance to water deficit stress. High leaf rolling in upland rice used in this study implies that they have relatively high tolerance characteristics to water stress at vegetative and reproductive growth stages compared to the lowland rice varieties investigated. However, among the lowland rice Tai variety had significant higher leaf rolling ability than Komboka and TXD306. For this reason, Tai was the most tolerant variety among the used lowland rice varieties.  These results are also in line with the findings of O’Toole and Garrity (1984) who reported that a rice variety with high leaf rolling ability is regarded as drought resistant.

Reducing or draining rice fields at either vegetative or reproductive phases caused significant yield loss (Castillo et al., 1992), and some researchers reported that effects of different periods of moisture stress at various growth stages would reduce yield (Salam et al., 2001; Sikuku, 2012). In the present study, moisture stress at vegetative growth stages significantly reduced grain yields by 26, 32 and 36% in NERICA1, NERICA2 and NERICA4, respectively in the upland rice varieties compared to the control moisture regime; while in the lowland rice varieties a reduction of 38, 44, and 46% in grain yield was observed for Komboka, Tai and TXD 306, respectively (Table 3). These findings are in agreement with those reported by Pirdashti et al. (2008) and Sikuku et al. (2012). In their study they found reduction in grain yield by 21 and 26% respectively, due to water deficit at vegetative growth stage.  However, Boonjung and Fukai (1996) reported that grain yields could be considerably reduced to about 60% if drought occurs during flowering time. These findings are in line with the observation revealed by the present study whereby moisture stress at reproductive growth stages highly reduced the grain yields of all varieties to more than 50% of the control moisture regime. NERICA2 and NERICA1 had the lowest reduction in grain yields (66%) compared to control; NERICA4,  which  was  highly susceptible to reproductive moisture stress, had 76% reduction in grain yields.  Among the lowland rice Tai variety had the lowest reduction among all the varieties tested in reproductive moisture stress with reduction of 58%, while Komboka variety had the highest reduction in grain yield of about 79%. All were used to compare the control moisture regimes (no stress).  The higher reduction in grain yield of all the varieties tested in this study are in line with the grain reduction reported at reproductive growth stages by Pirdashti et al. (2008) and Sikuku et al. (2012) who found reduction of 50 and 67%, respectively depending on the severity of stress. The lower grain yield in moisture stressed plots in the present study might be due to decreased panicle m^-2, spikelet panicle^-1, fertile spikelet's panicle^-1 and percentage fertility ratio caused by reduced number of effective tillers; it caused remobilization of carbohydrates reserves in shoot straw to the grains as the plants compete for moisture (Table 3). The lowest yields were recorded in the reproductive moisture stress regimes compared to the vegetative moisture stress regimes (Table 3). Low relative leaf water contents at reproductive growth stage significantly reduced the fertile grain number, and thus the grain yield was significantly reduced.

It was concluded that, the most critical growth stage that needs more moisture for the selected rice varieties to retain their higher yield in this study was the reproductive growth stage. Moisture stress at this stage had significantly high impact on growth, grain yield and yield components of both upland rice and lowland rice varieties investigated. Low relative water contents in leaves due to moisture stress inhibited growth and normal plant function and resulted in significant reduction in plant height, shoot dry weight, number of effective tillers, panicles, spikelets per panicle, fertile spikelets, 1000 grain weights and harvest index in all varieties tested, though with different decline intensity.  The most tolerant varieties among the upland rice investigated were NERICA2 and NERICA1. While NERICA4 was relatively a susceptible rice variety to reproductive moisture stress among the upland rice varieties tested because of higher reduction in grain yield. In the lowland rice varieties Tai was found to be the most tolerant variety than all other lowland rice varieties investigated; while Komboka and TXD 306 were relatively highly susceptible to both vegetative and reproductive moisture stress regimes. 

The author recommends that where possible adequate moisture should be available in soils at all developmental stages in order to achieve optimum yields of selected rice varieties. In case the area has been experiencing short period moisture stress of between 7 to 21 day at vegetative growth, all upland rice NERICA1, NERICA2, NERICA4 and Tai variety from the lowland rice varieties are   recommended   for   production.   In  case  the  area experiences short time moisture stress at reproductive growth stages, then NERICA2 for upland and Tai variety for lowland are recommended.

The authors have not declared any conflict of interests.

This work was supported by the Joshua’s Family fund that provided research and 50% of tuition fee, and the University of Dar es salaam (UDSM) for providing 50% of tuition fee funds for the first author to study at the university of Dar es Salaam.