Effect of sowing date and seeding rate on bread wheat yield and test weight under Mediterranean conditions

Yield and test weight are attributes of particular economic importance in wheat production systems and are strongly affected by environmental conditions. This study was conducted to determine the effects of sowing date and seeding rate on grain yield and test weight of fifteen bread wheat varieties and five advanced lines from Portuguese Wheat Breeding Program (Plant Breeding Station, Elvas, Portugal) under irrigated Mediterranean systems. Field experiments were conducted at two locations of Southeast Portugal, during 2011/2012 growing season. Two seeding rates were compared (200 and 350 seeds.m) in two different sowing dates in each location. Results showed that sowing date and seeding rate affects yield and test weight under irrigation field conditions, for Mediterranean region of Southeast Portugal, but only sowing date had significant effects among the varieties. At Elvas, higher yield was obtained with the 2nd sowing date (21 December) for most of the varieties studied. In opposite, in Beja trials, the highest values for yield were found when varieties were sown earlier (1st sowing date 26 October). Test weight had a similar performance in the two sites, though top values for this trait were found with the 1st sowing date. Comparing the results obtained in the two studied locations, Beja showed, for the majority of the varieties, 3t/ha higher average yield than Elvas.


Introduction
Wheat (Triticum aestivum L.) is a major cereal crop in many parts of the world and it is commonly known as the king of cereals.It belongs to Poaceae family and globally, after maize and rice, is the most cultivated cereal (FAOSTAT, 2013).Researchers can manage wheat cultivars, fertilizer levels, irrigation regime and agricultural practices to maximize wheat crop yield under the current conditions, but environmental constraints still be the main factors affecting wheat productivity in many regions of the world (El-Maaboud et al., 2004).In the US, for instance, unfavorable environments account for over 94% of the difference between average and record yield, and less than 6% of the disparity is attributable to diseases, insects and weeds (Boyer, 1982).In Mediterranean regions rain falls mostly during autumn and winter, and the water deficit rises in spring, coinciding with the anthesis and grainfilling period.Thus, drought and heat stress usually reduce yield potential during the period of grain formation (Simane at al., 1993;Lloveras et al., 2004).The challenge to increase wheat yield is even more difficult by projected climate changes, particularly higher temperatures and changes on rainfall distribution and amount (Parry and Hawkesford, 2010;Lobell et al., 2011).Even at a single location, in addition to variation due to agronomic and genetic factors, there is often considerable year to year variation reflecting different weather patterns.It is important to recognize that for farmers, maximizing yield is not their sole objective; profitability and managing risk are the most important criteria.Test weight is especially important for several food grain crops, particularly for those on which this trait is compulsory measured.The test weight is the first measurable/weighable qualitative trait of grain cereals mentioned in history, from the 19th century.Since then a great attention has been paid to it.Although it was introduced into regulations during the 20th century, it is hardly mentioned in the seed legislation.Protic et al. (2007) defines seed test weight using its dependence on seed density, shape and size, highlighting the fact that test weight is an important quality trait and that it can be used to estimate the amount of grain in a warehouse.They also state that test weight increased in time as a contribution of plant breeding.It is used by breeders to evaluate variety adaptation to the specific local conditions.The main goal of this study was to evaluate the effect of sowing date and seeding rate on grain yield and test weight of bread wheat under irrigation, in farmers field conditions.

Trials location
Trials were conducted during the 2011/2012 cropping season at farmers field in two different environments: Elvas (Alto Alentejo region) and Beja (Baixo Alentejo region), representing the most important provinces in Portugal for bread wheat crop.Table 1 shows some important data about the field trial sites.

Wheat germplasm
Cultivars choice was based on its growth cycle and origin.Among the studied germplasm, fifteen bread wheat varieties are from different origins and the remaining five are advanced breeding lines from Portuguese Wheat Breeding Program of National Institute for Agrarian and Veterinarian Research (Elvas, Portugal).

Field experiments
Wheat germplasm was evaluated in four experiments with two different sowing dates and two seeding densities, in two locations (Elvas and Beja), under irrigation conditions (Table 3).
All treatments were conducted with: nitrogen fertilization at sowing time (150 kg ha -1 as 18-46-0) and three top-dressed fertilizations (150 kg ha -1 as Urea 46%; 937 kg ha -1 as 32N Solution and 150 kg ha -1 as ammonium nitrate 27%); two weed control (at pre-emergence and post-emergence) and three antifungal treatments (tillering, jointing and heading stages).The experimental design was a randomized complete block design with four replications using a split plot treatment arrangement.Each plot size area was 6 m 2 (5 m long and and six rows, 20 cm apart).Details of the meteorological conditions and irrigation supply, in both environments, are presented in Figure 1.

Statistical analysis
Statistical analysis was performed on SPSS software (IBM, version 17.0).Means were compared using Tukey Student's test (significance level P < 0.05).Analyses of variance were done across sowing date and seed rate for each location.

Evolution of soil water availability and climatic conditions
The evolution of water stored in the soil differed between the two studied locations.At Elvas, a shallow silt loam soil with about 50 cm of deep and with infiltration problems, applied irrigations were of low allocation to avoid losses by flooding.The irrigation had an effect only on the first 10 cm of soil deeper.With raising temperatures in March, it was observed an intensification of water consumption by plants and, consequently, a reduction of its availability on the soil (Figure 1).This situation was slightly reversed due to rainfall occurred during April and May.At Beja, with a deep and clay soil, irrigation had a visible effect in the 20-30 cm of soil deep, allowing a better "hydric comfort" to plants during the growth cycle.Only in March (when the maximum temperature increased), water consumption augmented due to crop evapotranspiration and water evaporation from the soil surface, resulting in a decrease on the amount of available water to the crop.This situation was quickly reversed by the occurrence of precipitation at the end of March (Figure 1).The soil differences of the two sites were of great importance in the wheat development.The higher water holding capacity of the soil at Beja allowed genotypes to grow under hydric comfort during grain filling period, as it can be seen in general on yield and test weight values.

Elvas experiments
Variety, sowing date and seeding rate affected yield significantly (Table 4).A significant variety x sowing date interaction for grain yield was found result of the cultivars different growth habits (winter/facultative/spring).
Significant differences were found for grain yield among varieties, when sowed with different seeding rates and at different sowing times.This experiment also showed an overall yield advantage for the late sowing time (Table 5).Ingenio and Flycatcher"s" obtained the highest values for yield in the 2 nd sowing date differing significantly from other varieties.Linha 2 obtained the smallest yield value for both sowing times and seeding rates with a significant difference from others.The greatest increase on yield between the 1 st and the 2 nd sowing date was observed in Ingenio, Inoui and Aguilla.These results are in disagreed with other authors (Cutforth et al., 1990, Ozturk et al., 2006) who reported that delaying sowing date leads to a decrease in wheat yield.Our results revealed that facultative and winter wheat varieties showed an advantage for grain yield when sown later as, in this specific crop season (2011/2012), temperatures during the grain filling period (April and May) where moderate as shown in Figure 1.Thus, it was possible for the varieties to elongate the cycle and increase the individual thousand grain weight with a positive response on yield.Nevertheless, spring wheat varieties (Ardila and Badiel) showed a decrease in grain yield for 2 nd sowing date, data that are in accordance with several authors (Cutforth et al., 1990;Lloveras et al., 2004;Ozturk et al., 2006).Grain yield increased with an increase in seeding rate (Table 5).Similarly, Ozturk et al. (2006) found that an increasing seeding rate up to 525 seeds m -2 , increased spikes per square meter at harvest, resulting in increased grain yield.Seeding rate effect was less important than sowing date in maximizing grain yield in Mediterranean environments.
Table 6 shows great differences between minimum and maximum yield values obtained at Elvas.The minimum yield value was obtained with Badiel with the lower seeding rate and at the 2 nd sowing data.In opposite, the maximum was obtained with Ingenio with the higher seeding rate and also at the 2 nd sowing data.An high coefficient of variation confirm this finding.Test weight was significantly affected by variety, sowing date and seeding rate.Furthermore, interactions between variety and sowing date and variety and seeding rate were found to be statistically significant for test weight (Table 7).
Test weight depends on grain size, shape and density and indicates the adaptability of a variety to environment.Nabão, TE0205, Roxo and Pata-Negra, showed the highest value for test weight and reveal remarkable stability across the two sowing times (Table 8).
A test weight advantage with the 1 st sowing date was observed when compared with the 2 nd date (Table 8).These results are in accordance with Protic et al. (2007) who concluded that test weight of winter wheat decreased with later sowing, as a consequence of compensatory effects among yield components (Borghi et al., 1995).Portuguese variety Nabã o and advanced line TE0205, both developed at Wheat Breeding Program (INIAV-Elvas), had the highest test weight values.Linha2, Linha3 and Linha1 had the lowest values for this trait (Table 8), with non-significant differences between seeding rates neither sowing dates.Genetics has an important role in regulating test weight but it can be affected also by climatic and edaphic factors.
Results of the whole data set showed that, test weight ranged from 61,84 to 83,62 kg.hl -1 at Elvas with a coefficient of variation showing low data dispersion (Table 9).

Beja experiments
Variety, seeding rate and sowing date showed a significant effect on yield.Interaction between sowing date and variety was also statistically significant for grain yield (Table 10).
Average yield of the top 5 varieties/ advanced lines in the 1 st sowing date was almost 2 t/ha higher than the 2 nd sowing date.Table 11 shows that in Beja trials, the higher yielding varieties increased grain yield when sown earlier.Results showed that sowing with a higher seeding rate did not outcome higher yield.Grain yield obtained with higher seeding rate was slightly superior (Table 11).Moreover, under favorable edafoclimatic conditions (irrigation, soil, etc.) a higher seed density results in high-biomass production, high number of spikes per square meter though with smaller spikes and consequently with no increase in grain yield.Peltonen-Sainio (1991) showed that a higher seed rate usually produces high-biomass and the genotypes often mature late, which is usually undesirable.
For the 1 st sowing date, the higher values for yield were obtained with Nogal, Inoui, Bologna, Flycatcher"s", Eufrates, Aguilla and Linha1 (Table 11), cultivars with facultative growth cycle (excepting Linha1) for which heading time occurred after April's 10 (data no shown).In Mediterranean conditions of Portugal the optimum heading time must occurred ± 10 days around April 1 st .For this facultative or winter wheat, the earlier sowing date (26 October) promoted a higher expression of grain yield potential.These results are according with Malcolm et al. ( 2013) who reported that in many countries where only spring wheat is cultivated, the highest wheat yield of over 15 t/ha have been achieved for winter wheat grown with a long growing season at higher latitudes.Woodruff et al. (1983) also reported that the large differences on the yield of genotypes having different development cycles, within a group and from a given sowing date, were primarily due to the interactions between growth duration, water use and evaporative demand conditions around anthesis.For 1 st sowing date, the lowest yield varieties were Pata-Negra, Alabanza, Mané-NicK, Ardila, Siena and Linha2 (Table 11).Except for Linha2 (winter variety), other varieties have short growth cycles, with earlier heading time, before April´s 1 st .Consequently, the late sowing on November 29 th , showed to be an advantage for grain yield of these spring wheat varieties.Ingenio, Roxo, Nabã o, TE0206, Badiel and TE0205 revealed remarkable yield stability concerning sowing date with an optimum heading time around April 1 st , as referred.However, yield of these varieties was always below the trial average mean.
Table 12 shows a large gap between the minimum and maximum yield values.This is in accordance with the big coefficient of variation found.The minimum yield value was obtained with Linha2 with the higher seeding rate and for the 2 nd sowing date.This performance points out the importance of duration of growth cycle, indicating that this variety is not adapted, once is a very late variety (data not shown).The maximum value was obtained with Nogal with the higher seeding rate and for the 1 st sowing date (Table 11).
Wheat variety, sowing date and seeding rate affected significantly test weight.Interaction between variety and sowing date was also found to have a significant effect on this trait (Table 13).Data showed that the higher and steady test weight values (including two sowing dates) were obtained with Ardila, Roxo, Nabã o, Eufrates, TE0205 and TE0206, with an increase when sowed later (Table 14).Varieties with longer growth cycle presented a significant reduction on test weight in 2 nd sowing date.Spaner et al. (2000) and Ozturk et al. (2006) reported that a delayed in sowing tends to decrease test weight in facultative wheat.
Maximum test weight was recorded Roxo, Bologna, Nabã o, Eufrates, Nogal, TE0205 and TE0206, in the 1 st sowing date (October 26 th ) with a significant decrease in the 2 nd sowing date (November 29 th ) as shown in table 14.This behaviour is similar to the observed with the grain yield at Beja.This performance reflects an important adaptation of these varieties to Mediterranean conditions predominant in Portugal.In opposite, varieties developed in different environmental conditions (longer growth cycles) showed worse adaptation resulting on lower test weights.
Results showed that test weight ranged from 60,14 to 83,74 kg.hl -1 in Beja with a small coefficient of variation indicating low dispersion of the data.

Conclusions
Results clearly confirmed that intrinsic genetic yield potential is not enough to obtain high wheat yield.Overall, in Beja, with the same varieties, the average wheat yield was around 3t/ha higher than in Elvas (Tables 6 and 12).This fact indicates that several limitations to the expression of yield potential exist, related with agronomy (i.e., depth and physical structure of the soil and crop management practices) and climate such as frost during flowering and high temperatures during grain filling, which could cause irreversible damage to wheat crop yield.In this context, the fastest and most practical ways to increase yield are to improve agronomy in conjunction with continuing genetic improvement (Costa et al. 2012).At Beja, the highest value for grain yield was obtained by Nogal with 7,8 t/ha and heading time at April 1 st .On the other hand, Linha2 showed the lowest performance with 3,6 t/ha and heading time out of adequate window (May 1 st ).At Elvas, Flycatcher"s", an advanced breeding line, was the best genotype with 3,2 t/ha and Linha2 revealed the end of the varieties ranking with 2 t/ha.The highest test weight was obtained with the 1 st sowing date, in both locations.Increasing sowing rate did not significantly influenced test weight at Elvas and Beja, for the majority of cultivars.These results showed that both sowing date and seeding rate influence grain yield and test weight in the majority of cultivars, but the effect of sowing date was greater than that of seeding rate.Results also indicated that according with the climatic conditions occurred during 2011/2012 season in both places (Elvas and Beja), where a strong Mediterranean pattern drives wheat development, the germplasm evaluation and selection is paramount to better determine and characterize the ideotype wheat plant that breeders should strive to develop.New advanced breeding lines like Flycatcher"s", TE0205

Figure 1 .
Figure 1.Evolution of water stored in the soil at Elvas and Beja.

Table 1 .
Table 2 presents origin and growth habit of the germplasm used in the experiments.Geographic location, soil type, sowing and harvesting dates of the trials.

Table 2 .
Origin and growth habit of bread wheat varieties and advanced lines.

Table 3 .
Location, sowing date and sowing rate of field experiments.

Table 5 .
Yield differences between two sowing dates with two seeding densities for 20 varieties/advanced lines of bread wheat in Elvas trials.

Table 6 .
Descriptive statistics values of yield in 20 bread wheat at Elvas.

Table 7 .
Analysis of Variance for test weight of 20 bread wheat sown at Elvas (Alto Alentejo, Portugal).

Table 8 .
Test weight differences between two sowing dates with two seeding densities for 20 varieties/advanced lines of bread wheat in Elvas trials.
Varieties are listed in test weight decreasing order of 1st date and usual seed rate (350 seed.m-2).Different letters in the same column indicate significant difference (P<0.05).

Table 9 .
Descriptive statistics values of test weight in 20 bread wheat on Elvas.

Table 10 .
Analysis of Variance for yield of 20 bread wheat sown at Beja (Baixo Alentejo region, Portugal).

Table 12 .
Descriptive statistic values for test yield in 20 bread wheat varieties at Beja.

Table 13 .
Analysis of variance for test weight of 20 bread wheat sown at Beja (Baixo Alentejo region, Portugal).

Table 14 .
Test weight differences between two sowing date with two seeding densities for 20 varieties of bread wheat in Beja trials.
Varieties are listed in test weight decreasing order of 1st date and usual seed rate (350 seed.m-2).Different letters in the same column indicate significant difference (P<0.05)

Table 15 .
Descriptive statistic values for test weight in 20 bread wheat varieties at Beja.
, TE0206, obtained by Cereal Breeding Program in Plant Breeding Station (Elvas, Portugal) are excellent examples resulting from this kind of approach.Simane, B., J. M. Peacock and P. C. Struik.1993.Differences in development plasticity and growth rate among drought-resistant and susceptible cultivars of durum wheat (Triticum turgidum L. var.durum).Plant Soil.157:155-166.Spaner, D., A. G. Todd and D. B. McKenzie.2000.The effect of seeding date, seeding rate and N fertilization on winter wheat yield and yield components in eastern Newfoundland.Can.J. Plant Sci.80:703-711.Woodruff, D. R. and Tonks, J. 1983.Relationship between time of anthesis and grain yield of wheat genotypes with differing developmental patterns.Aust.J. Agric.Res.34:1-11.