Intercropping for the Improvement of Sorghum Yield, Soil Fertility and Striga Control in the Subsistence Agriculture Region of Tigray (Northern Ethiopia)

Striga hermonthica is a major biotic constraint in the dry and less fertile areas of northern Ethiopia. Emphasis is being placed on improved cropping systems to address the interrelated problems of Striga and soil fertility decline. The potential benefits of intercropping were investigated at two sites representing different environments for crop yield improvement, soil fertility maintenance and Striga control. Ten food legume and oilseed crop species were compared in inter‐row arrangement with sorghum under non‐fertilized conditions. In most cases, there was no significant negative impact of intercropping on sorghum growth and development. Among the intercrops, two cowpea varieties – cv. TVU 1977 OD and cv. Blackeye bean – produced the highest supplemental yield of up to 329 and 623 kg ha^?1 grain and 608 and 1173 kg ha^?1 biomass at Adibakel and Sheraro respectively. Treatment differences on Striga infestation and measured soil fertility indicators were not significant. Nevertheless, valuable grain and biomass obtained from the legume intercrops, without seriously compromising sorghum yield, could offer multiple benefits as a source of protein, additional income, feeds for animals and manure in the subsistence agriculture regions of northern Ethiopia.     

Molecular mapping and identification of candidate gene conferring organophosphate-sensitive reaction in sorghum (Sorghum bicolor)

Some sorghum cultivars are sensitive to organophosphate insecticides, which cause red to purple injury spots and, in severe cases, leaf death. Sensitivity to organophosphates is conditioned by a single locus on chromosome 5. We constructed a high-density genetic map around the locus, termed osr, using DNA markers, and delimited osr to a 377-kb region on the short arm of chromosome 5. Sequence analysis of this region predicted 19 gene candidates. Four of the candidates are homologous to a disease resistance NBS-LRR gene. mRNA-Seq analysis of gene expression and SNPs from two resistant cultivars (‘JN43’ and ‘Greenleaf’) and one sensitive cultivar (‘Nakei MS3B’) suggest that one or more members of this NBS-LRR gene family are osr. Our results suggested that the recessive allele of osr results in sensitivity to OP insecticides.     

Relay Cropping of Sorghum and Legume Shrubs for Crop Yield Improvement and Striga Control in the Subsistence Agriculture Region of Tigray (Northern Ethiopia)

Striga hermonthica is a major constraint in the subsistence agriculture regions of northern Ethiopia. Low soil fertility and overall environmental degradation has contributed to the build up of the parasitic weed infestation. Improved cropping systems have to be introduced to address the interrelated problems of Striga and soil fertility decline. Thus, relay cropping of sorghum with legume shrubs was investigated at two locations representing different environments. Results showed that the output of the improved cropping system was dependent on ecological endowments. Relay cropping led to significant improvement in yield at Sheraro, at the site with relatively better weather and soil conditions. The legume shrubs resulted in significantly lower sorghum yield in a dryland location (Adibakel). Overall Striga infestation declined over the 3‐year period; however, treatment differences were not apparent. Among the two legume shrubs, Sesbania sesban was better adapted to the dryland areas. Relay cropping could provide a viable option for farmers in both types of environments that are characterized by accelerated decline in natural resource base. However, it could mean compromising the yield of non‐fertilized sorghum in the interest of long‐term benefits of low incidence of Striga and more rewarding crop enterprise in dry areas.     

Genetic divergence among a non-restorer collection of sorghum (Sorghum bicolor (L.) Moench) and its relationship with heterosis

Summary A genetic diversity analysis in a collection of 171 non-restorer lines of sorghum (Sorghum bicolor (L.) Moench) using D² technique and canonical variate analysis indicated that considerable variation in grain yield has been added to the collection by the addition of lines derived from random mating populations. The efficiency of D² and canonical variate techniques in distinguishing extremely diverse genotypes was confirmed. However, the two techniques showed weak correspondence in their clusters. The F₁ hybrids of 15 diverse lines exhibited no relationship between heterosis or per se performance of crosses and diversity in their parents. Therefore, traditional plant breeding methods are being advocated.Approved by ICRISAT as Journal Article no. 435.     

In vitro screening and field evaluation of tissue-culture-regenerated sorghum (Sorghum bicolor (L.) Moench) for soil stress tolerance

Summary Sorghum bicolor (L.) Moench is generally quite sensitive to salt and acid (high aluminium) soil stresses, but quite tolerant of drought stress. As with any stress phenomenon, intra-specific variability exists within the genus. In vitro cell selection and somaclonal variation offer an alternative to traditional breeding methodology for generating improved breeding lines for hybrid development. A field selection protocol was developed for the three soil stresses and inter-stress evaluations were conducted in an effort to find multiple, stress-tolerant genotypes. The acid soil-drought stress, super-tolerant selections were located by the R₇ generation when exposed to a combined aluminium-drought stress field environment and when the regeneration population (number of regenerated lines from one callus source) was maintained at 15,000 plants or higher. A variant frequency of 0.1 to 0.2% for stress tolerance and acceptable agronomic traits among the surviving somaclones, provided an adequate number of phenotypes with desirable agronomic characteristics and a high level of soil stress tolerance. Subsequent research verified that the stress-tolerant regenerants had superior acid soil and drought stress tolerance to that of the donor parents, that their yield capabilities under stress were superior to their parents, and that their stress tolerance attributes were transferred in hybrid combinations. In vitro selection was not effective in increasing the number of field stress survivors. In fact, superior germplasms were developed from non-stressed callus or salt-stressed callus. In vitro selection reduced regeneration frequency and subsequent survival of plants under field stress. In vitro-stressed regenerants should be subjected only to non-stressed environments to maintain population numbers for field selection and thereafter should be subjected to stress environments during later (R₅₊) generations. The optimal strategy for the exploitation of somaclonal variation may be through short-term cell culture (< 12 months) with no attempt at in vitro selection.     

Farmer and formal breeding of sorghum (Sorghum bicolor (L.) Moench) and the implications for integrated plant breeding

Sorghum (Sorghum bicolor (L.) Moench) is the fifth most important cereal crop worldwide and it is the fourth most important crop in Ethiopia. The national average yield amounts 1302 kg/ha. In order to assess the achievement in farmer breeding various types of research were undertaken. These include survey research to quantify the trend in productivity, the level of and reasons for adoption of improved varieties, yield performance and preference evaluation of farmers’ varieties (FVs) and improved varieties (IVs). As per the trend analysis over the last four decades, total production and yield per hectare has increased by 11.63 and 14.2%, respectively. However, area allocated to sorghum has decreased over years by −2.93%. The lack of consistent productivity is attributed to the fluctuation of environmental factors. Sorghum production in Ethiopia is predominantly based on varieties developed by farmers. The share of IVs is very low. FVs and IVs are adopted by 87.3 and 12.7% of the farmers, respectively. Besides, the adoption of IVs is limited to the lowland crop ecology. The comparative yield of FVs is higher than IVs by 132%. On top of yield, farmers do prefer their varieties for other multipurpose values namely feed, fuel wood and construction material. FVs under production are identified in each wereda. Farmer breeding has been successful compared to four decades of formal breeding. On the other hand, both farmer and formal breeding are not without weaknesses; a comparative balance sheet is outlined for both. Ideotypes for the three major crop ecologies are suggested and integrated plant breeding is anticipated to develop the proposed ideotypes thereby increase sorghum productivity in the region.     

Variation of the genomic proportion of the recurrent parent in BC1 and its relation to yield performance in sorghum (Sorghum bicolor) breeding for low-input conditions

In order to define the variation of the genomic proportion of the recurrent parent [G(RP)] and its relation to yield, G(RP) of individual BC₁ plants of two sorghum populations composed of a high‐yielding cultivar as recurrent parent (RP) and a donor with superior drought resistance or grain quality, respectively, was estimated using AFLPs and SSRs. G(RP) in BC₁ ranged from 0.53 to 0.95 and averaged to 0.76 in the population (NP4453 × ‘SV‐2’) × ‘SV‐2’. G(RP) varied between 0.60 and 0.86 and averaged to 0.74 in the BC₁ of (ICV‐219 × ‘SV‐2’) × ‘SV‐2’. Results show that plants with a G(RP) equivalent to BC₂ (0.875) or BC₃ (0.938), respectively, can be selected from BC₁. Yield performance of BC₁S₁ families was tested in field trials carried out in South Africa. The correlation between yield and G(RP) in BC₁ was low. Selection according to G(RP) did not result in an effective preselection for yield.     

Effect of Nitrogen Rate and Stubble Height on Dry Matter Yield,Crude Protein Content and Crude Protein Yield of a Sorghum–Sudangrass Hybrid[Sorghum bicolor (L.) Moench × Sorghum sudanense (Piper) Stapf.] in the Three‐Cutting System

In this study, the effects of nitrogen (N) rate (60, 120, 180 and 240 kg N ha^?1 applied in three equal dressings at seeding and after the first and second cuttings) and stubble height (7, 14 and 21 cm) on the dry matter (DM) yield, crude protein (CP) content, and CP yield of a sorghum–sudangrass hybrid [Sorghum bicolor (L.) Moench × Sorghum sudanense (Piper) Stapf., cv. Pioneer 988] in the three‐cut system was investigated. The N rate had no significant effect in the first and third cuttings, but in the second cutting DM yields increased significantly with increase in N rate. The highest yield of 9.1 t ha^?1 was obtained with 80 kg N ha^?1 for the average of 2 years at the second cutting, but no significant difference was found among the 40, 60 and 80 kg N ha^?1 rates. CP content and yield were not significantly affected by N rate at the first and third cuttings, but CP content and yield were significantly affected by application of N at the second cutting. Stubble height had a significant effect on CP content at the third cutting. However, it had no significant effect on CP content at the first and second cuttings. Stubble height had a significant effect on the CP yield at the first cutting, but no significant effect on CP yield at the second and third cuttings.     

玉米和高粱用于碳同化和光呼吸的电子效率估算

为了探讨C₄植物碳同化和光呼吸的电子效率,运用Li-6400光合仪同时测定玉米和高粱在30℃和380 μmol CO₂ mol^-1下叶片的气体交换和叶绿素荧光,结果表明,直角双曲线修正模型可较好地拟合所测的光响应曲线和快速光曲线,其拟合值与实测值较为一致。在此基础上算得玉米和高粱在光呼吸条件下参与碳同化的电子流分别为198.60 μmol m^-2 s^-1和178.00 μmol m^-2 s^-1,所占比率分别为75.34%和74.81%;参与光呼吸的电子流分别为7.04 μmol m^-2 s^-1和7.84 μmol m^-2 s^-1,所占比率分别为2.67%和3.29%。而根据Valentini和Epron的方法算得玉米和高粱碳同化的电子流分别为210.45 μmol m^-2 s^-1和188.54 μmol m^-2 s^-1,所占比例分别为82.68%和79.24%;参与光呼吸的电子流则分别为45.67 μmol m^-2 s^-1和49.40 μmol m^-2 s^-1,所占比率分别为17.32%和20.76%。以前法研究表明,玉米和高粱在光呼吸条件下,来自PSII的电子除流向光呼吸和碳还原外,还存在其他消耗电子的途径,证明其他消耗电子的途径并不能被忽略或其他途径所消耗电子的量并不是常数。后法过高地估算了玉米和高粱叶片中来自PSII的电子用于光呼吸的消耗量。两法的结果相差6倍左右。这对重新评估光呼吸在植物的光保护中所起的作用提供了理论依据。     

Efficiency of secondary traits in selecting for improved grain yield in extra‐early maize under Striga‐infested and Striga‐free environments

A base index involving Striga damage, number of emerged Striga plants and ears per plant is used for selecting for maize (Zea mays L.) grain yield under Striga infestation. There are contradictory reports on the reliability of number of emerged Striga plants for selecting for Striga resistance. The objective of this study was to confirm reliability of the secondary traits for selecting for improved grain yield under Striga infestation. Ten Striga‐resistant extra‐early cultivars were evaluated for 3 years under artificial Striga‐infested and Striga‐free environments in Nigeria. Analysis of variance combined across years and locations showed significant mean squares for genotype, year, location and their interactions for most traits. Sequential path analysis identified ear aspect as the only trait with significant direct effect on yield under artificial Striga infestation, while GGE biplot confirmed ear aspect, ears per plant and Striga damage as the most reliable traits. Ear aspect should be included in the base index for selecting for improved grain yield of extra‐early maize under Striga infestation, while the number of emerged Striga plants should be excluded.     

Agronomic performance of three cocoyam (<Emphasis Type="Italic">Xanthosoma violaceum</Emphasis> Schott) genotypes grown in Nicaragua

Cocoyam (Xanthosoma spp.), the third most important starch food crop in Nicaragua, can be cultivated countrywide. However, very little information about the field performance and genetic background of the different genotypes is available. In this study, the agronomic performance of three purple genotypes established in four locations with different climatic conditions, were evaluated during 2 years. Phenotypic characteristics, yield, and virus incidence were assessed and the time at which the different genotypes reached their physiological maturity was estimated. The trials were based on randomized complete block design with three or four blocks. Genotype × location interaction was found both for phenotypic and yield traits. A differential response of the genotypes to the varying climatic conditions at the locations is suggested to be one of the causes to the interaction. The unpredictable and variable climate in the area where cocoyam traditionally has been grown is one possible explanation to the frequent genotype × year interaction at that location. Other locations with more stable climate only showed an interaction between genotype and year in a few traits. The genotypic differences regarding the time when the area of the largest leaf reached its maximum and the variation in presence of sprouts and roots on the cormels at harvest, indicate differences in optimal harvest time between genotypes. The percentage of plants infected with dasheen mosaic virus (DMV) differed across locations but not between genotypes.     

Agronomic performance of winter wheat grown under highly divergent soil moisture conditions in rainfed and water‐managed environments

Even in the temperate climates of Europe, increasing early season drought and rising air temperature are presenting new challenges to farmers and wheat breeders. Sixteen winter wheat (Triticum aestivum L.) genotypes consisting of three hybrids, six line cultivars and two breeding lines from Germany as well as five line cultivars from France, Austria, Slovakia, Hungary and the Ukraine (referred to as “exotic” lines) have been included in this study. The genetic materials were evaluated over three growing seasons under a range of soil moisture regimes at the three North German sites Braunschweig (irrigated and drought‐stressed), Warmse (rainfed) and Söllingen (rainfed). The average grain yields in the twelve growth environments (water regime × season combinations) ranged from 6.1 to 13.5 t ha^?1. The exotic lines showed little evidence of specific phenological adaptation to drought although they are frequently faced with water scarcity in their countries of origin. The hybrids and German lines exhibited higher regression coefficients (b_i) to environmental means than the exotic lines, indicating particular adaptation to favourable growing conditions. The phenotypical correlations of grain yield between the various environments were high, ranging for instance from 0.6 to 0.8 for the irrigated and drought‐stressed environments at Braunschweig. It is thus expected that in the foreseeable future continued selection aiming at high yield potential will suffice as a means to counter the expected increase in droughts.     

Evaluation of Mustard (Brassica compestris Var. Toria) and Legume Intercropping under 1 : 1 and 2 : 1 Row‐Replacement Series Systems

Field experiments were conducted at Agricultural Experimental Farm, Giridih, Bihar during the winter seasons of 1992–93 and 1993–94 to compare 1 : 1 and 2 : 1 row‐replacement series intercropping of mustard (Brassica campestris Var. Toria) with pea (Pisum sativum), with lentil (Lens culinaris) and with gram (Cicer areitinum L.). Intercropping systems were assessed on the basis of new indices termed ‘actual yield loss’ (AYL) and ‘intercropping advantage’ (IA), as well as several existing intercropping indices such as LER, RCC and aggressivity. Intercropping reduced the yield of component crops, compared with respective pure stands. Mustard + pea in 1 : 1 row‐replacement series gave the highest mustard yield equivalent value (1.00 t/ha). When the actual sown proportion was considered, mustard + lentil (1 : 1) resulted in maximum AYL (+0.438) and IA (+2.889) values. We conclude that AYL and IA are more appropriate, particularly when per plant yield is considered.     

Mega‐environment identification for soybean (Glycine max) breeding and production in Brazilian Midwest region

Mega‐environment (ME) identification is the first step for evaluating, selecting and recommending genotypes within a target region (TR). The present study aimed to (a) identify MEs, using GGE biplot methods, in Brazilian edaphoclimatic region (ECR) 402 of soybean cultivation, located in the Mato Grosso State (the TR) and (b) compare the performance of genotypes within the TR and in each ME using fixed and mixed models. Data from three years of soybean yield trials, 19 genotypes and 22 environments were used. The biplots GGE, GGL + GGE and GGS + GGE were implemented to identify the MEs. Two MEs were identified in the TR. ME1 presents a higher altitude, farms which use a higher level of fertilizer inputs and a higher occurrence of the soybean cyst nematode (SCN) than ME2. When selection and recommendation are made based on MEs, genotypes with both broadly and specific adaptation can be selected. This action can improve grain yield in the entire target region.     

Indonesian mungbean landraces II. Comparison of performance,components of variance and selection for grain yield under wide row-spacing in monoculture and when intercropped with maize

Summary The possibility of testing mungbean for grain yield in monoculture in single wide-spaced row plots as an alternative to selection in intercropping with maize was investigated. Results indicated no significant genotype x planting pattern interactions and genetic correlation coefficients between grain yields in the two planting patterns were, in most cases, close to unity. It was concluded that, as far as the mungbean component in the mixture is concerned, yield selection in either of the two planting patterns is justified. Effects of single-plant and progeny selection on mungbean characteristics were also studied empirically. Lines derived from single-plant selection in monocropping or intercropping showed differences in yield, number of pods per plant and seed weight. Except for seed weight, no differential effects of progeny selection in monoculture or intercropping were observed for any characteristic.     

Generation mean analysis of horticultural traits in mid‐late cauliflower (Brassica oleracea L. var. botrytis) under sub temperate conditions of Western Himalayas

Suitable breeding procedure for crop improvement mainly depends on the type and relative amount of genetic components and presence of non‐allelic interactions. Therefore, the study was carried out during three successive rabi seasons 2013‐14, 2014‐15 and 2015‐16 in mid hill zone of North‐western Himalayas, India with objective to estimate the main genetic effects including digenic non‐allelic interactions controlling horticultural traits in cauliflower. The generation mean analysis was carried out on six generations (P₁, P₂, F₁, F₂, B₁ and B₂) of six crosses involving four diverse parents of cauliflower. The results obtained revealed that the nature and magnitude of gene effects differed in different crosses and showed importance of duplicate epistasis in the inheritance of curd diameter and curd size index, whose effect can be wiped out by following sophisticated selection procedure such as reciprocal recurrent selection and/or biparental mating in early segregating generation. Complementary type of epistasis was recorded in most of the cross combinations for the important yield traits, which indicates the importance of heterosis breeding for the development of high yielding varieties.     

QTL mapping for six ear leaf architecture traits under water‐stressed and well‐watered conditions in maize (Zea mays L.)

Morphological traits for ear leaf are determinant traits influencing plant architecture and drought tolerance in maize. However, the genetic controls of ear leaf architecture traits remain poorly understood under drought stress. Here, we identified 100 quantitative trait loci (QTLs) for leaf angle, leaf orientation value, leaf length, leaf width, leaf size and leaf shape value of ear leaf across four populations under drought‐stressed and unstressed conditions, which explained 0.71%–20.62% of phenotypic variation in single watering condition. Forty‐five of the 100 QTLs were identified under water‐stressed conditions, and 29 stable QTLs (sQTLs) were identified under water‐stressed conditions, which could be useful for the genetic improvement of maize drought tolerance via QTL pyramiding. We further integrated 27 independent QTL studies in a meta‐analysis to identify 21 meta‐QTLs (mQTLs). Then, 24 candidate genes controlling leaf architecture traits coincided with 20 corresponding mQTLs. Thus, new/valuable information on quantitative traits has shed some light on the molecular mechanisms responsible for leaf architecture traits affected by watering conditions. Furthermore, alleles for leaf architecture traits provide useful targets for marker‐assisted selection to generate high‐yielding maize varieties.     

Detection of additive and additive×environment interaction effects of QTLs for yield‐component traits of rice using single‐segment substitution lines (SSSLs)

Additive effects (A) and additive‐by‐environment interactions (A×E) for five rice yield components were analysed using 20 SSSLs under mixed linear model methodology. Thirty‐one QTLs were detected. Different yield components have different QTL‐by‐environment (Q×E) interaction patterns. No A×E interaction effects were detected for the four QTLs for panicle number (PN). Four QTLs detected for spikelets per panicle (SPP) had A×E interactions. Five of seven QTLs detected for grains per panicle (GPP), two of 10 QTLs detected for 1000‐grains weight (GWT) and three of six QTLs detected for seed set ratio (SSR) showed significant A×E interaction. Most of these QTLs were distributed in clusters across the genome. The complexity of linkage and pleiotropy of these QTLs plus environmental effect may result in the diversity of the yield phenotype in the SSSLs. Only S19 exhibited a significant increase in yield with a predicted gain by 281.58 kg ha^?1. The results may be useful to design a better breeding strategy that takes advantage of QTL‐by‐environment interaction effects in each of the SSSLs.     

Application of the Cropping System Model (CSM)‐CROPGRO‐Soybean for Determining Optimum Management Strategies for Soybean in Tropical Environments

The determination of optimum crop management practices for increasing soybean production can provide valuable information for strategic planning in the tropics. However, this process is time consuming and expensive. The use of a dynamic crop simulation model can be an alternative option to help estimate yield levels under various growing conditions. The objectives of this study were to evaluate the performance of the Cropping System Model (CSM)‐CROPGRO‐Soybean and to determine optimum management practices for soybean for growing conditions in the Phu Pha Man district, Thailand. Data from two soybean experiments that were conducted in 1991 at Chiang Mai University and in 2003 at Khon Kaen University were used to determine the cultivar coefficients for the cultivars CM 60 and SJ 5. The CSM‐CROPGRO‐Soybean model was evaluated with data from two experiments that were conducted at Chiang Mai University. The observed data sets from farmers’ fields located in the Phu Pha Man district were also used for model evaluation. Simulations for different management scenarios were conducted with soil property information for seven different soil series and historical weather data for the period 1972–2003 to predict the optimum crop management practices for soybean production in the Phu Pha Man district. The results of this study indicated that the cultivar coefficients of the two soybean cultivars resulted in simulated growth and development parameters that were in good agreement with almost all observed parameters. Model evaluation showed a good agreement between simulated and observed data for phenology and growth of soybean, and demonstrated the potential of the CSM‐CROPGRO‐Soybean model to simulate growth and yield for local environments, including farmers’ fields, in Thailand. The CSM‐CROPGRO‐Soybean simulations indicated that the optimum planting dates from June 15 to July 15 produced maximum soybean yield in a rainfed environment. However, the planting date December 15 produced the highest yield under quality irrigation. Soybean yield was slightly improved by applying nitrogen at a rate of 30 kg N ha^?1 at planting. Soybean yield also improved when the plant density was increased from 20 to 40 plants m^?2. The results from this study suggest that the CSM‐CROPGRO‐Soybean model can be a valuable tool in assisting with determining optimum management practices for soybean cropping systems in the Phu Pha Man district and might be applicable to other agricultural production areas in Thailand and southeast Asia.     

Seed Number and 100‐Seed Weight of Pearl Millet (Pennisetum glaucum L.) Respond Differently to Low Soil Moisture in Genotypes Contrasting for Drought Tolerance

Water stress after flowering, one of the major factors limiting yields of pearl millet, affects both seed setting and grain filling and is a consequence of more/less water used prior to anthesis. However, whether genotypes have different sensitivities for seed setting and filling under drought, if exposed to similar stress intensity, is unclear. Experiments were conducted in two pairs of pearl millet genotypes, that is, PRLT2/89‐33 and H77/833‐2, 863B and 841B, contrasting for terminal drought tolerance, and two genotypes, ICMR 01046 and ICMR 01029 (IL‐QTLs), introgressed with a terminal drought tolerance QTL from PRLT2/89‐33 into H77/833‐2. Total seed weight, panicle number, 100‐seed weight, seed number and stover biomass were measured at different soil moistures and throughout grain filling. Sensitive H77/833‐2 had higher seed number and yield under well‐watered (WW) conditions than in PRLT2/89‐33 and IL‐QTLs. Upon increases in water stress intensity, H77/833‐2 suffered losses mostly in stover biomass (45 %) and seed number (60 %) at 0.3 FTSW whereas the biomass and seed number of PRLT2/89‐33 decreased little (20 % and 25 %). The 100‐seed weight of H77/833‐2 decreased only 20 % under stress. Tolerant 863B also maintained a higher seed number and biomass under water stress than 841B. Grain filling duration in PRLT2/89‐33 and IL‐QTLs was similar to that of H77/833‐2 under WW conditions but lasted longer than in H77833‐2 under water stress (WS). Similarly, seed growth of 863B was longer than 841B under WS. It is concluded that the higher seed yield of tolerant parents PRLT2/89‐33 and 863B, and of IL‐QTLs under WS was explained by the retention of a higher number of seeds than in sensitive lines, while the decrease in the 100‐seed weight was proportionally less than the decrease in seed number. Phenotype with lesser number and larger size of panicles and larger grain size, like genotypes PRLT2/89‐33 and 863B, withstood post‐anthesis water stress better. IL‐QTL inherited part of these characteristics, indicating a role for the terminal drought QTL in maintaining larger seed number and higher 100‐seed weight. The continuous stover biomass increase under WW in H77/833‐2, due to tillering, might indicate that tiller growth and grains are in competition for resources after anthesis, and this may relate to the relatively shorter grain‐filling period.