Remobilization of dry matter in wheat: effects of nitrogen application and post-anthesis water deficit during grain filling

Abstract

Pre-anthesis stored dry matter in wheat (Triticum aestivum L.) is important in a Mediterranean climate because grain filling greatly depends on the remobilization of pre-anthesis assimilates. A water deficit at the post-anthesis stage may increase the dry matter stored before anthesis. This field study assessed the effects of post-anthesis water deficit and the application of nitrogen (N) fertilizer on three wheat cultivars. Data collected over 2 years showed that, in wheat with a post-anthesis water deficit (WD), dry matter remobilization efficiency reached its maximum (29%) at 80 kg N ha⁻¹, but further additions of N decreased it. The contribution of remobilized dry matter to a grain ranged from 7% to 23% of the grain's dry weight and, in WD grain, was 78% more than that of well-watered (WW) grain. Grain from plots on which fertilizer had been applied had a lower proportion of remobilized dry matter than did grain from unfertilized plots. For grain from adequately fertilized treatments, limited irrigation was associated with reduced dry matter remobilization. The amount of non-structural carbohydrates (NSC) remaining in parts of the plant was much greater in the WW than in the WD treatments. The 160 kg N ha⁻¹ treatment also left more NSC unused than did the no N treatment. The active grain-filling period was shortened substantially by a water deficit, but this was countered by the application of 160 kg N ha⁻¹. Grain-filling rates for all cultivars or all N treatments were increased by inducing a water deficit. For WD grains, kernel weight was reduced when fertilizer application rates were 0–80 kg N ha⁻¹, but increased when rates were 160 kg N ha⁻¹, unlike WW grain exposed to similar fertilizer regimes. The grain yield of WD wheat was reduced by 25%, and that of grain receiving no fertilizer was 15% lower than that receiving 80 kg N ha⁻¹. Among three wheat cultivars, cv. Chamran produced the highest grain yield (19% higher than that of either cv. Shiraz or cv. Marvdasht). It was concluded that the stored carbohydrate had provided an important buffer against water stress during grain filling, in terms of yield.

Keywords:

• grain filling rate
• grain filling period
• grain yield
• kernel weight
• non-structural carbohydrate residue

Introduction

During spring, in the Mediterranean climate of southern Iran, decreased rainfall and increased evaporation and temperature coincide with the grain-filling stage of wheat (Fathi 2005). Consequently, wheat crops often experience a water deficit during grain filling, which limits subsequent grain yield (Spiertz et al. 2006; Barnabas et al. 2008).

The carbon (C) necessary for grain filling in wheat comes from three sources: current assimilation; remobilization of pre-anthesis assimilates stored in the stem and other parts of the plant; and retranslocation of assimilates stored temporarily in the stem after anthesis. To understand the reduction in grain yield arising from post-anthesis water deficits, it is necessary to identify which of these sources of C is limiting the grain-filling process (Ercoli et al. 2008).

Post-anthesis water deficits are known to reduce C assimilation and hence the availability of current assimilates for grain filling, but they are not considered to affect the translocation of C to the grain (Ercoli et al. 2006). A water deficit during grain filling increases the proportion of stored assimilates relative to current assimilates in the grain (Yang et al. 2000), but whether this reflects a larger actual mobilization of stored assimilates rather than simply a reduction in current assimilation is not known (Fan et al. 2005).

Van Herwaaden et al. (1998) showed that under dry conditions in the field, 75%–100% of the grain yield could be attributed to stored assimilates, compared with 37%–39% under high-rainfall conditions. In fact, a considerable correlation was found between the amount of stored non-structural carbohydrates (NSC) in wheat stems and yield in several wheat cultivars under drought conditions (Gavuzzi et al. 1997). Besides the stem, NSC are also stored within leaf sheaths and leaves, and fructans are probably the most abundant stored carbohydrate sources for kernel filling (Wardlaw & Willenbrink 2000).

The remobilization of assimilates in stored vegetative tissues and their transfer to the grain requires the initiation of whole plant senescence (Yang et al. 2002; Ehdaie et al. 2006). Early senescence caused by post-anthesis water deficit leads to reduced photosynthesis, shortens the grain-filling period and results in reduction of grain weight (Yang et al. 2000). In Iran, nitrogen (N) application results in an undesired delayed senescence, leading to a low kernel weight. Low grain weight can also occur, in northern Iran, when the hot and dry winds typical of the end of the growing season coincide with delayed senescence: such winds dehydrate the wheat rapidly. Early senescence induced by water stress may increase the rate of grain filling and improve kernel weight in this case.

The objectives of this study were to assess: first, the effect of a post-anthesis water deficit on the remobilization dry matter stored in various plant tissues and its contribution to grain yield during the grain-filling period; and, second, the interacting effects of N fertilization and post-anthesis water deficit on dry matter remobilization in three wheat cultivars.

Materials and methods

The studies were conducted at Shiraz Agricultural Research Station (52°36′E, 29°33′N), for 2 years (2006 and 2007). A split-plot experimental design was used, based on a complete randomized block design with four replications. The main plots received one of two irrigation treatments: WW (well-watered) or WD (post-anthesis water deficit). Each subplot received one of three rates of N (in the form of urea) fertilizer application: 0, 80 or 160 kg ha⁻¹. Within the subplots were sub-subplots, each containing one of the three cultivars Shiraz, Marvdasht or Chamran. These cultivars, with differing anthesis dates, were chosen as representative of the winter wheat cultivars widely grown in the south of Iran. To determine basic soil physical and chemical properties, 15 samples from 30 cm depth were collected and analysed by the Shiraz Soil Testing Laboratory (Table 1). Based on the laboratory's recommendations, 150 kg ha⁻¹ of superphosphate and 50 kg ha⁻¹ of potassium sulphate were applied to all plots. The plots, 8×1.5 m, were sown on 11 November 2006 and on 14 November 2007 with a cone seeder at the rate of 350 seeds m⁻² (six rows, 0.2 m apart). Plots were ploughed and disced after the winter wheat harvest in July and were disced again before sowing in November. Irrigation of each main plot was based on volumetric data collection with six field-calibrated gypsum blocks (Delmhorst Instrument Co) that had been installed in each replication randomly. The blocks were cylindrical, with a diameter of 25 mm and a length of 35 mm. The soil water potential at 30 cm depth was kept at −0.025 MPa (WW) or −0.1 MPa (WD) from anthesis to maturity. The irrigation system was operated so that runoff did not occur.

Table 1  Soil physical and chemical properties of the experimental site.

	Year
Property	2006–07	2007–08
Texture	Silty-clay	Silty-clay-loam
pH	7.96	7.85
EC (dS m^−1)	1.88	1.25
Organic matter (%)	0.25	0.39
N (%)	0.023	0.036
P (mg kg^−1)	5.4	16
K (mg kg^−1)	340	206
Fe (mg kg^−1)	3.7	5.4
Zn (mg kg^−1)	0.64	0.1
Mn (mg kg^−1)	5.8	9.9
Cu (mg kg^−1)	0.48	0.98

To control both broad- and narrow-leaved weeds, Apirus was applied in early April. Apirus is a sulphonylurea (Su) herbicide, and therefore controls weeds through inhibition of the acetolactase synthase enzyme. The three most common active ingredients of Su herbicides are chlorsulfuron, metsulfuron-methyl and triasulfuron. Twenty main stems that headed on the same day were tagged for each treatment. As those stems reached anthesis, 10 plants in each plot (per cultivar) were removed and main stems were divided into spikes, flag leaf blades, lower leaves and stems. Tillers were discarded. At maturity, 10 additional tagged plants were removed, and the main stems were subdivided as at anthesis. Anthesis was scored when anthers in the central florets of 50% of the spikes in plants had dehisced; maturity was scored when almost all the spikes in the plots showed complete loss of green colour. Anthesis date of the cultivars differed by about 10 days in both years. Samples were dried to a constant weight at 65 °C, weighed and ground to pass through a 1 mm screen. Average kernel weight was determined by weighing 250 kernels randomly taken from the bulk grain sample in each plot. The grain filling process was fitted to the Richards (1959) growth equation as described by:

where W is kernel weight (mg), A is final grain weight, t is time after anthesis (days), and B, k and N are coefficients determined by regression (Zhu et al. 1988). The active grain-filling period was defined as that when W was from 5% (t1) to 95% (t2) of A. The average grain-filling rate during this period was calculated from t1 to t2.

Fifteen to 20 plants were sampled from each treatment at 6-day intervals from anthesis to maturity for the measurement of non-structural carbohydrates (NSC) in the stem, flag leaf and other leaves. The method for extraction of NSC was modified according to the method by Yoshida et al. (1976). The sample was dried in an oven and ground into a fine powder. In a 15 mL centrifuge tube, 100 mg of ground sample was added to 10 mL of 8% ethanol and kept in a water bath at 80 °C for 30 min. The tube was then centrifuged at about 900×g for 20 min after cooling. The supernatant was collected and the extraction was repeated three times. The alcohol in the supernatant was evaporated on an 80 °C water bath until most of the alcohol was removed and the volume was reduced to about 3 mL. Absorption spectra(625 nm wavelength) were recorded by spectrophotometer (Model UV-160A SHIMADZU). The four central rows (of six rows) of each plot were harvested for assessment of grain yield (kg ha⁻¹).

Various parameters for dry matter, within different parts of the wheat plant, were calculated as follows (Papakosta & Gagianas 1991):

Apparent dry matter remobilization (mg plant⁻¹) = dry matter at anthesis (spike +flag leaf + stem + lower leaves) − dry matter at maturity

Apparent dry matter remobilization efficiency (%)=(dry matter remobilization/dry matter at anthesis)×100

Contribution of apparent dry matter remobilization to grain (%)=(dry matter remobilization/grain yield)×100

Data were analysed by analyses of variance using the general linear model (GLM) procedure provided by SAS (SAS Institute 1982. SAS/STAT user's guide). SAS Inst., Cary, NC. Combined analysis of variance between years assumed that replications were random, and that irrigation, N fertilizer and cultivars were fixed. When significant differences were found (P=0.05), the Duncan's multiple range test (DMRT) was carried out.

Results and discussion

Weather conditions

The amount and distribution of rainfall differed over the 2 years that they were measured (Table 2). A total of 327 mm of rain was recorded in the period October to June in 2006–07, compared with 295 mm for the same period in 2007–08. Spring rainfall (March to June) was 29 mm in the first season, and 15 mm in the second. Spring maximum temperatures were on average 2 °C lower in 2006–07 than in 2007–08. Autumn and winter temperatures were near the yearly average and spring temperatures were above average in both years.

Table 2  Summary of October–June precipitation and temperature data for 2006–08 at the experimental site.

	October	November	December	January	February	March	April	May	June
				2006–07
Precipitation (mm)	13	57	95	98	35	20	8	1	0
Temperature – max (°C)	21.9	12.9	11.4	9.9	12.4	15.1	22	25.4	30.0
– min (°C)	4.4	−1.9	−4.6	−6.4	−5.3	0.8	4.3	9.3	14.3
				2007–08
Precipitation (mm)	10	19	78	151	24	14	1	0	0
Temperature – max (°C)	23	15.6	13.2	11.6	14.1	16.7	23.0	26.2	32.3
– min (°C)	6.1	−0.7	−3.4	−5.4	−4.3	−1.1	6.7	9.6	16.1
				1978–2008 average
Precipitation (mm)	10.7	30	84	120	61	25	10	7	0.5
Temperature – max (°C)	19.6	16.2	12.0	10.8	7.5	15.7	20.1	24.4	27.9
– min (°C)	6.7	1.5	−4.1	−7.2	−6.6	−1.2	3.0	9.2	13.1

Dry matter remobilization

The apparent percentages of remobilized reserves, and their contribution to the grain weight, were significantly higher in grain exposed to the WD treatment (Tables 3a, 3b). Dry matter remobilization in all parts of the plant was increased by 14%–42% under the WD regime compared with the WW one. At maturity, 173 mg plant⁻¹ of the spike dry weight had been remobilized to the grains under the WD treatment, a 20% increase over spike dry weight of WW plants (Table 3a). For the flag leaves and stem, a dry weight of 9 and 277 mg plant⁻¹, respectively, were remobilized under the WD regime, which were 14% and 42% above those of corresponding WW treatments (Table 3a).

Table 3a  Mean values of dry matter remobilisation (DMR) and DMR efficiency (DMRE) of spike, flag leaf, stem and lower leaf of three wheat cultivars under three N applications and two water regimes in 2006–08.

Treatments	DMR of spike (mg plant^−1)	DMR of flag leaf (mg plant^−1)	DMR of stem (mg plant^−1)	DMR of lower leaves (mg plant^−1)	DMRE of spike (%)	DMRE of flag leaf (%)	DMRE of stem (%)	DMRE of lower leaves (%)
Irrigation
WW	139	4	186	30	22.57	20.57	13.91	19.322
WD	173	9	277	39	27.09	26.49	18.21	22.99
LSD (0.05)	19	0.9	21	5	3.2	3.2	1.6	2.1
Nitrogen (N) kg ha^−1
0	143	2.6	212	33	28.13	23.60	14.80	20.51
80	146	16	227	49	24.18	23.94	16.97	23.59
160	190	16	271	37	22.01	23.04	16.42	19.36
LSD (0.05)	21	1.1	20	5.1	2.8	3.1	1.4	2
Cultivar (C)
Shiraz	158	16	269	34	22.03	26.38	16.48	19.58
Marvdasht	173	18	200	43	24.26	23.92	13.52	21.78
Chamran	153	1.5	241	37	28.20	20.29	18.18	22.10
LSD (0.05)	21	1.3	19	4.6	2.6	3.1	1.1	1.9
WW = well-watered treatment.
WD = post-anthesis water deficit.
LSD = least significant difference.

Table 3b shows that NSC were substantially reduced after the water deficit was imposed, with the reduction being greater under the 0 and 80 kg N ha⁻¹ regimes than under the 160 kg N ha⁻¹ one. The amount of NSC remaining in parts of the plant was much greater under the WW treatment than under the WD treatment (Table 3b). The 160 kg N ha⁻¹ treatment also left more NSC unused than did the 0 kg N ha⁻¹ treatment. Acreche & Slafer (2009) and Alvaro et al. (2007) demonstrated that stomata begin to close 5 days after anthesis, when leaf water potential fell to −1.5 MPa. This would reduce the amount of net assimilation and the accumulation of stem water-soluble carbohydrates, which occurs between anthesis and the onset of the linear phase of grain filling (Ercoli et al. 2006). Dry matter remobilization was also found to vary between cultivars. Dry matter remobilization in Marvdasht was 11% and 25% more than those of Shiraz and Chamran, respectively (Table 3b). Water deficit greatly enhanced the grain-filling rate (Table 3b).

Table 3b  Total dry matter remobilisation (DMR) and DMR efficiency (DMRE), and mean values of contribution of dry matter to grain, active grain-filling period, grain-filling rate, NSC residue, kernel weight, grain yield and harvest index, of three wheat cultivars under three N applications and two water regimes in 2006–08.

Treatments	Total DMR (mg plant^−1)	Total DMRE (%)	Contribution of DMR to grain (%)	Active grain-filling period (days)	Grain filling-rate (mg day^−1)	NSC residue (mg g^−1 DW)	Kernel weight (g)	Grain yield (kg ha^−1)	Harvest index (%)
Irrigation
WW	377	17.01	7.88	37	0.8	232	33.78	5897	31.7
WD	513	22.01	14.54	25	1.9	120	27.57	4410	28.0
LSD (0.05)	35	0.9	1.7	4.8	0.2	19	4.7	1177	5
Nitrogen (N) kg ha^-1
0	427	19.96	14.35	29	1.9	67	31.08	5135	33.9
80	448	20.51	10.86	35	1.6	188	35.94	5832	34.0
160	459	18.06	8.42	41	1.1	222	36.02	5823	38.7
LSD (0.05)	37	1	1.8	4.9	0.3	22	4.9	1185	5.1
Cultivar (C)
Shiraz	482	19.05	13.77	42	0.7	219	30.83	4799	28.5
Marvdasht	427	18.18	14.69	38	1.3	162	27.49	4908	28.7
Chamran	525	21.30	9.17	27	1.8	113	33.72	5853	32.5
LSD (0.05)	37	1	1.8	5	0.2	23	4.7	1298	5.1
NCS = non-structural carbohydrates.
WW = well-watered treatment.
WD = post-anthesis water deficit.
LSD = least significant difference.

Vegetative dry matter losses from anthesis to maturity have also been observed by other investigators and have varied greatly with environmental factors (Ehdaie & Waines 2001; Robert et al. 2001). Dry matter remobilization increased with increasing amounts of added N on WD plots (Table 4) At the highest level of N fertilizer, the limited water available appeared to reduce dry matter remobilization. By contrast, when only 80 kg N ha⁻¹ was applied, 18% of the dry matter was partitioned into the grains under the WD treatment, while only about 8% was allocated to the grains under the WW treatment, at maturity (Table 4), indicating that more dry matter had been remobilized and deposited into the grains as a consequence of the water deficit. The plants that had received 80 kg N ha⁻¹ remobilized 10% more dry matter than the plants that had received 0 N, and therefore the former had a higher grain yield (Table 4).

Table 4  Mean values of dry matter remobilisation (DMR) and DMR efficiency (DMRE) of spike, flag leaf, stem and lower leaf; contribution of dry matter to grain, active grain-filling period, grain-filling rate, NSC residue, kernel weight, grain yield and harvest index, as affected by N application and irrigation regime in 2006–08.

Treatments		DMR of spike (mg plant^−1)	DMR of flag leaf (mg plant^−1)	DMR of stem (mg plant^−1)	DMR of lower leaves (mg plant^−1)	DMRE of spike (%)	DMRE of flag leaf (%)	DMRE of stem (%)	DMRE of lower leaves (%)	Total DMR (mg plant^−1)
Irrigation	Nitrogen (kg ha^−1)
WW	0	147	1.1	234	19	23.95	20.71	14.77	15.60	401
	80	139	13	135	58	21.70	19.36	9.78	25.75	338
	160	146	13	204	27	22.05	21.63	9.78	16.60	391
WD	0	149	5	307	47	24.38	26.48	19.16	25.42	462
	80	240	18	288	41	34.92	28.51	19.80	21.42	495
	160	145	20	250	47	21.96	24.45	15.65	22.12	581
	LSD (0.05)	22	1.6	27	4.6	2.9	2.8	1.8	2.6	37
Source of variation
	Irrigation (I)	*	**	**	*	*	**	**	*	**
	Nitrogen (N)	NS	*	*	NS	NS	NS	*	NS	*
	I×N	NS	*	*	NS	NS	NS	NS	NS	*
CV,%		22.3	9.3	25.6	5	6.7	9.1	11.2	10.9	17.9
		Total DMRE (%)	Contribution of DMR to grain (%)	Active grain- filling period (d)	Grain- filling rate (mg day^−1)	NSC residue (mg day^−1)	Kernel weight (g)	Grain yield	Harvest index (%)
WW	0	18.10	8.16	29	1.7	106	35.13	5931	35.45
	80	16.33	8.05	38	1.2	232	33.56	5884	29.25
	160	16.58	7.42	41	0.9	263	32.66	5873	30.31
WD	0	21.80	13.55	19	1.4	73	29.02	4337	29.16
	80	24.68	18.63	30	0.9	184	26.30	4320	24.79
	160	19.53	11.42	33	0.8	203	27.38	4572	30.33
	LSD (0.05)	2.5	1.4	3.1	0.2	25	2.9	1201	4.7
Source of variation
	Irrigation (I)	*	*	**	*	*	*	*	*
	Nitrogen (N)	*	NS	*	*	*	*	*	*
	I×N	NS	NS	*	*	*	*	*	NS
CV,%		13.5	5.3	9.9	8.5	14	10.2	13.1	17.5
NSC = non-structural carbohydrates.
WW = well-watered treatment.
WD = post-anthesis water deficit.
LSD = least significant difference.
*,** significant at 0.05 and 0.01 probability levels, respectively.
NS = non-significant at P>0.05.
CV=Coefficient of variation.

The increase in dry matter remobilization caused by the added nitrogen was high for Marvdasht and low for Chamran (Table 5) Under the WD treatment, however, all three cultivars allocated less dry matter to the grains under the highest N application rate (160 kg N ha⁻¹; Table 6). The application of N fertilizers resulted in significantly higher above-ground dry matter at anthesis and maturity and greater grain yields (Table 3b). Grain yield was positively related to dry matter at anthesis. Correlation coefficients, although low, were significant for all three cultivars (Shiraz, 0.61; Marvdasht, 0.50; and Chamran, 0.40) (data not shown). The active grain-filling period was shortened substantially by water deficit (Tables 3b–6), which was countered by the addition of 160 kg N ha⁻¹. When grain-filling rates are compared for all cultivars or under all N treatments, rates were higher in WW plants compared with the respective WW ones (Tables 4–6).

Table 5  Mean values of dry matter remobilisation (DMR) and DMR efficiency (DMRE) of spike, flag leaf, stem and lower leaf; contribution of dry matter to grain, active grain-filling period, grain-filling rate, NSC residue, kernel weight, grain yield and harvest index of three wheat cultivars as affected by N-fertiliser application rate.

Treatments		DMR of spike (mg plant^−1)	DMR of flag leaf (mg plant^−1)	DMR of stem (mg plant^−1)	DMR of lower leaves (mg plant^−1)	DMRE of spike (%)	DMRE of lag leaf (%)	DMRE of stem (%)	DMRE of lower leaves (%)	Total DMR (mg plant^−1)
Nitrogen (kg ha^-1)	Cultivar
0	Shiraz	152	9	293	32	23.16	25.85	17.27	19.25	488
	Marvdasht	145	11	278	47	21.26	27.89	17.27	24.55	483
	Chamran	145	0.2	241	20	28.09	17.05	16.35	17.74	473
80	Shiraz	153	16	214	47	20.78	25.28	14.88	23.31	432
	Marvdasht	227	28	159	43	30.89	22.23	11.37	21.70	437
	Chamran	190	4	262	57	33.26	24.30	18.13	25.74	509
160	Shiraz	169	34	299	25	22.15	28.01	17.27	16.18	527
	Marvdasht	145	15	163	38	20.65	21.61	11.91	19.09	361
	Chamran	123	0.9	219	48	23.23	19.51	20.05	22.81	391
	LSD (0.05)	23	1.2	26	3.9	2.5	2.6	1.8	2.1	35
Source of variation
	Nitrogen (N)	NS	*	*	NS	NS	NS	*	NS	*
	Cultivar (C)	*	*	*	NS	NS	*	*	NS	*
	N×C	NS	*	NS	NS	NS	*	NS	NS	NS
CV,%		8.4	4.7	11.4	9.2	15.6	16.2	8.9	13.5	12.1
		Total DMRE (%)	Contribution of DMR to grain (%)	Active grain-filling period (day)	Grain-illing rate (g day^–1)	NSC residue (mg g^−1 DW)	Kernel weight (g)	Grain yield (kg ha^−1)	Harvest index (%)
0	Shiraz	20.07	11.53	35	1.2	280	32.44	4946	32.39
	Marvdasht	19.83	11.57	28	1.5	223	28.93	5091	32.25
	Chamran	19.95	9.46	19	1.9	150	34.85	5365	34.05
80	Shiraz	17.52	11.99	38	0.9	311	30.44	4882	26.11
	Marvdasht	20.23	15.33	32	1.3	254	26.76	4783	25.67
	Chamran	23.77	12.71	23	1.6	173	32.60	5641	29.23
160	Shiraz	19.55	11.55	42	0.8	343	29.58	4568	26.91
	Marvdasht	14.48	8.39	36	1.1	285	26.77	4847	28.24
	Chamran	20.15	8.32	27	1.3	197	33.69	6253	34.05
	LSD (0.05)	2.1	0.9	3.8	0.2	29	3.5	1211	4.9
Source of variation
	Nitrogen (N)	*	NS	*	*	*	*	*	*
	Cultivar (C)	NS	NS	*	*	*	*	*	*
	N×C	NS	NS	*	*	NS	NS	NS	NS
CV,%		7	5.6	13.6	9.5	17.3	11.5	14.2	17.3
NSC = non-structural carbohydrates.
WW = well-watered treatment.
WD = post-anthesis water deficit.
LSD = least significant difference.
*,** significant at 0.05 and 0.01 probability levels, respectively.
NS = non-significant at P>0.05.
CV=Coefficient of variation.

Table 6  Mean values of dry matter remobilisation (DMR) and DMR efficiency (DMRE) of spike, flag leaf, stem and lower leaf; contribution of dry matter to grain, active grain-filling period and grain-filling rate of three wheat cultivars as affected by N-fertiliser application rate and irrigation treatments.

Treatments			DMR of flag leaf (mg plant^−1)	DMR of stem (mg plant^−1)	DMRE of flag leaf (%)	DMRE of stem (%)	Active grain-filling period (day)	Grain-filling rate (mg day^−1)
Irrigation	Nitrogen (kg ha^−1)	Cultivar
WW	0	Shiraz	8	229	23.43	13.40	34	1.2
		Marvdasht	7	311	24.95	19.02	30	1.4
		Chamran	0.1	163	13.76	11.90	24	1.7
	80	Shiraz	13	80	24.45	5.86	38	1.0
		Marvdasht	28	62	13.52	5.17	34	1.2
		Chamran	1	263	20.12	18.35	29	1.4
	160	Shiraz	29	284	25.36	16.41	41	0.8
		Marvdasht	11	129	21.27	10.05	37	1.0
		Chamran	0.5	198	18.27	25.05	31	1.2
WD	0	Shiraz	11	358	28.27	21.15	26	1.2
		Marvdasht	15	245	30.83	15.52	25	1.4
		Chamran	0.1	320	20.33	20.81	18	1.7
	80	Shiraz	19	348	26.12	23.91	33	0.9
		Marvdasht	27	256	30.93	17.57	30	1.2
		Chamran	8	260	28.47	17.93	23	1.4
	160	Shiraz	38	314	30.65	18.13	35	0.8
		Marvdasht	18	196	21.96	13.77	32	1.1
		Chamran	3	239	20.76	15.06	26	1.3
	LSD (0.05)		1.3	25	2.7	2	3.6	0.3
Source of variation
	I×N×C		*	*	*	*	*	*
CV,%			5.1	12.3	18.9	11.1	14.8	10.2
WW = well-watered treatment.
WD = post-anthesis water deficit.
LSD = least significant difference.
*,** significant at 0.05 and 0.01 probability levels, respectively.
NS = non-significant at P>0.05.
CV=Coefficient of variation.

Dry matter remobilization efficiency

Water deficit markedly affected the dry matter remobilization efficiency, causing it to increase by 29% (Table 3b). With the WD treatment, dry matter remobilization efficiency increased with the application of 80 kg N ha⁻¹ but decreased with further addition of N (Table 4). Cultivars markedly differed in dry matter remobilization efficiency (Table 4)—Chamran showing the highest remobilization efficiency (23%). One explanation for differences in remobilization efficiency is that during the grain-filling period, the plant retains some of the dry matter accumulated at anthesis for survival and various biological functions, while the remainder is available for remobilization. It appears that the amount of retained dry matter depends on the cultivars and prevailing growth conditions, although genetic variability in dry matter remobilization has been reported (Reynolds & Trethowan 2007; Saint Pierre et al. 2008). Remobilization efficiencies in this study are considerably greater than those reported by Papakosta & Gagianas (1991).

Contribution of dry matter to the grain

The contribution of dry matter remobilization to the grain ranged from 7% to 23% of grain dry weight (Tables 3–6) and differed among cultivars. Similar values were reported for barley (Hordeum vulgare L.) by Krček et al. (2008), who concluded that pre-anthesis storage of carbohydrates is very important for grain yields. Contribution of dry matter remobilization to the grain under WD was 78% more than that under WW conditions (Table 3b). In the Mediterranean climate, rising temperatures and declining soil moisture prevailing at the post-anthesis period limit net assimilation rates; therefore, the contribution of post-anthesis dry matter to the grain is greater. In this study, well-fertilized plants showed a lower contribution of stored dry matter to grain than did unfertilized plants (Table 3b). Greater dry matter at anthesis resulted in a greater proportion of remobilized dry matter. Thus, cultivars Shiraz and Marvdasht showed high proportions of dry matter allocated to grains—13.77% and 14.69%, respectively. Plants that had received 160 kg N ha⁻¹ showed a lower contribution of pre-anthesis assimilates than unfertilized plants (Table 3b). With an application rate of 160 kg N ha⁻¹, plants are expected to have a continuous supply of N during the post-anthesis stage. This condition is presumably conductive to higher rates of photosynthesis and, in turn, to a larger supply of assimilates for grain filling, thus reducing the need for remobilization of pre-anthesis assimilates. Kirda et al. (2001) reported that large movements of assimilates can occur under low soil fertility conditions. No correlation was found between harvest index (HI) and dry matter remobilization efficiency or the contribution of pre-anthesis assimilates to grain (data not shown).

Kernel weight

In the 0 and 80 kg N ha⁻¹ treatments, kernel weight under the WD regime was less than that of plants under the WW regime, indicating that the decrease in photosynthesis could not be compensated for by an increased remobilization of carbon reserves (Table 4). When a high amount of N was applied, kernel weight was increased under the water deficit. The obvious explanation for such a result is that, when N was heavily used, delayed senescence led to a slow grain filling, and a poor remobilization and partitioning of assimilates into the grain. For the high N application rate, the gain from accelerated grain filling and increased remobilization of pre-anthesis assimilates outweighed the loss of reduced photosynthesis and early senescence as a result of water stress. Kernel weight was high for cv. Chamran and low for cv. Marvdasht. For all varieties studied in this experiment, N addition increased kernel weight (Table 5). Kernel weight did not show any significant interaction between cultivars and N application rate.

Grain yield

Water deficit markedly affected grain yield (Tables 3b, 4), causing a reduction of 25%. The reduction resulted from a decrease in grain size, but not in grain number. Grain yield increased in response to the applied N, with the grain yield of the plants that did not receive N being 15% lower than those receiving 80 kg N ha⁻¹ (Table 3b). Average winter wheat yields were 5135 and 5832 kg grain ha⁻¹ for the 0 and 80 kg N ha⁻¹ treatments, respectively (Table 3b). Grain yield was reduced by water deficit under 80 kg N ha⁻¹ compared with 0 kg N ha⁻¹, but increased under 160 kg N ha⁻¹ when compared with respective WW treatments (Table 4). For the WD treatment, water became more limiting than fertility, and there was no yield response to the applied N. The N response data may be used in determining N fertilizer rates for specific yield goals. Water deficit at the grain-filling period induces early senescence, reduces photosynthesis and shortens the grain-filling period (Inoue et al. 2004; Waines 2006). Such responses would result in the reduction of kernel weight, straw yield and grain yield (Alvaro et al. 2007; Karam et al. 2009).

The Chamran cultivar at any nitrogen level produced the maximum grain yield compared with the two other cultivars and ranged from 5365 to 6253 kg grain ha⁻¹ (Table 5).

Nitrogen fertilizer is one of the major input costs in winter wheat production in Iran. Our soil tests indicated that the experimental field was low in available soil N. It was not surprising that the response of winter wheat to N application was dramatic under nitrogen application. Our results indicate that there is great potential to increase winter wheat yields by properly managing nitrogen fertilization.

Harvest index

Water stress during the grain filling reduced the harvest index by reducing grain yields after most of the vegetative growth had been completed. The harvest index also varied with N treatment (Table 3b). These results could be explained by the separate effects of N treatment on grain yields and straw dry matter at maturity. Cv. Marvdasht, which had the lowest grain yields, also had the lowest harvest (8%–14%). Furthermore, since Marvdasht did not always produce the lowest amount of total dry matter, its low yields were due, at least in part, to a low harvest index. Harvest index increased with increasing nitrogen, from 34% at 0 kg N ha⁻¹ nitrogen application to 39% at an application rate of 160 kg N ha⁻¹. Harvest index was reduced by water deficit under 80 kg N ha⁻¹ compared with 0 kg N ha⁻¹, but increased under 160 kg N ha⁻¹ (Table 4). Among the three wheat cultivars, Chamran produced the highest harvest index at either N level (Table 5).

Conclusion

Hot and dry winds (temperature > 30 °C, relative humidity < 30% and wind speed>10 km/h) are common in early June to July in this area and can dehydrate wheat before it matures. Water stress may accelerate senescence so that the wheat can mature before the adverse condition occurs. It may be possible to take advantage of such responses and improve yields in situations where slow grain filling, as a result of delayed senescence, is a problem.

This study has demonstrated that the grain-filling rate of wheat, while affected by water deficit, is maintained above what is expected from post-anthesis dry matter accumulation because remobilization of assimilates to the grain continued despite a reduction in C assimilation.

Stems were more important than the other parts of the plant in the remobilization of dry matter to the grain during the grain-filling period.

The excessive use of N in Iran results in unfavourably delayed senescence. Early senescence induced by water stress could increase the rate of grain filling and improve kernel weight in this case.