Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Agricola
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Related Collections
Right arrow Other Grain Crops
Right arrow Crop Models
Right arrow Crop Genetics
Agronomy Journal 92:114-120 (2000)
© 2000 American Society of Agronomy

SMALL GRAINS

A Model Analysis of Yield Differences among Recombinant Inbred Lines in Barley

Xinyou Yina, Martin J. Kropffb, Jan Goudriaanb and Piet Stamc

a Plant Research International, Wageningen Univ. and Research Centre, P.O. Box 14, 6700 AA Wageningen, Netherlands
b Laboratory of Theoretical Production Ecology, Wageningen Univ. and Research Centre, P.O. Box 430, 6700 AK Wageningen, Netherlands
c Lab. of Plant Breeding, Wageningen Univ. and Research Centre, P.O. Box 386, 6700 AJ Wageningen, Netherlands

x.yin{at}plant.wag-ur.nl


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 
Crop models can support plant breeding if they can predict differences in performance of different genotypes. In this study, the ability of a crop model to explain yield differences among genotypes in a recombinant inbred line (RIL) population of two-row barley (Hordeum vulgare L.) was explored. Yield and model-input traits of 94 RILs and their parents, `Prisma' and `Apex', were measured in field experiments conducted in Wageningen, Netherlands, in 1996 at low and in 1997 at high N levels. The major gene, denso, with the dwarfing allele from Prisma, was segregating in this population. Short denso RILs outyielded tall types in both years, and this yield advantage was stronger in 1997, largely because the tall genotypes lodged. A crop model based on existing routines for biomass production explained only 26 to 38% of the yield variation among genotypes. The model, using input traits measured from the 1997 data, did not accurately predict growth of genotypes in 1996 because some traits varied with plant N status, which the model did not account for. Model analysis in the high-N environment showed that of the seven model-input traits examined, only lodging score, preflowering duration, and fraction of biomass partitioned to spikes had a significant effect on yield. When these three traits were used while fixing others at their across-genotype means, the model explained 65% of yield variation. To allow effective use of crop modeling in breeding, the ability of crop models to explain yield differences among genotypes has to be improved.

Abbreviations: DAE, days after emergence • DS, development stage • FPleaf, fraction of shoot biomass partitioned to leaves • FPspike, fraction of shoot biomass partitioned to spikes • LAI, leaf area index • LNC, leaf nitrogen content • Pre-F, preflowering duration • Post-F, postflowering duration • RIL, recombinant inbred line • SLA, specific leaf area • SYP-BL, simulator of yield potential for barley


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 
TO COPE WITH growing demand for food, new cultivars with increased yield potential are required. Through intensive selection, largely by selecting for yield per se on an empirical basis, breeders have been successful in increasing crop yield potential. By comparing eight wheat (Triticum aestivum L.) cultivars released between 1962 and 1988, Sayre et al. (1997) reported a linear progress in yield potential, with a rate of progress against year of release of 67 kg ha-1 yr-1. However, further progress has been increasingly difficult and has in fact declined in recent years (CIMMYT, 1995). Support from other disciplines is therefore becoming more important (Bindraban, 1997).

Donald (1968) introduced the concept of designing ideotypes with enhanced yield potential based on a composition of favorable traits. Ideotype breeding gives higher priority to sets of individual traits than selection for yield per se. While Donald (1968) focused mainly on morphological traits, ideotype breeding was broadened by including physiological traits. For example, the new plant type for rice (Oryza sativa L.), which is expected to break through the apparent yield potential barrier of current cultivars, was defined largely on the basis of studies of physiological yield-component traits (Peng et al., 1994). The need for physiological studies in plant breeding results at least partly from the fact that yield is a very complex trait controlled by numerous interacting genes. Physiological studies provide a way to dissect complex traits into simpler components that might be under separate genetic control.

A promising approach to analyzing yield is the use of ecophysiologically based crop growth models (Loomis et al., 1979; Boote et al., 1996). These models have been developed by integrating knowledge across disciplines and are increasingly applied in problem-oriented research (Boote et al., 1996). One application is the explanation of differences in yield potential of genotypes on the basis of individual physiological characteristics and the use of this knowledge to evaluate and design new plant types (Loomis et al., 1979; Kropff et al., 1995). In mechanistic crop models, the genotypic expression of genetic characteristics of plants is simulated for a specific environment. Physiological parameters in crop models, or model-input traits, are often referred to as genetic coefficients, indicating that these parameters are mainly under genetic control (Stam, 1998). Using models, many studies (e.g., Penning de Vries, 1991; Boote and Tollenaar, 1994; Kropff et al., 1995) examined opportunities to increase yield potential. Others have explored better selection environments. For example, Aggarwal et al. (1997) revealed that selection for high rice yield is possible only under high-N conditions, as many lines with higher yield potential than the standard cultivar would be eliminated under the N conditions often practiced in current breeders' screening plots.

The contribution of physiology and crop modeling to breeding so far has, however, been small to modest (Jackson et al., 1996). One reason for this could be a predefined or different focus in most physiological and modeling studies, using cultivars with limited genotypic range in a wide range of environments. An important aspect of successful interfacing of physiological and modeling research with breeding is the need to work with relevant populations and the close integration of physiological and modeling research with an active breeding program (Jackson et al., 1996). Crop models have been shown to explain yield differences of a genotype among highly different environments (Kropff et al., 1994). The main challenge for the use of crop models in breeding is to predict differences in performance of relatively similar lines in a population (McLaren, 1995).

The objectives of this study were (i) to examine the ability of an ecophysiological crop growth model to explain yield differences among genotypes in a segregating population and (ii) to analyze the relative importance of individual physiological characteristics in determining yield differences. The study has been part of an extensive research effort that explores possibilities to link crop modeling and genetic analysis for a better understanding of yield potential. Barley was chosen as a model crop because it is a diploid, self-pollinating species for which an experimental population can be developed with relative ease, and because barley has relatively few chromosomes. We have reported results of the genetic analysis elsewhere (Yin et al., 1999). Here we report the model analysis of yield differences among genotypes in the population with an attempt to identify major yield-determining physiological traits.


   Materials and methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and methods
Results and discussion
 Conclusions
 REFERENCES
 
Recombinant Inbred-Line Population
A population of 94 RILs was produced by eight generations of single-seed descent (i.e., each randomly chosen plant contributes a single offspring to the next generation) from a cross of commercial two-row spring barley cultivars, Apex and Prisma. The parents have contrasting morphophysiological characteristics and Prisma usually outyields Apex (Schut, 1992). Prisma is of shorter stature than Apex, largely due to the dwarfing allele of a major gene denso. Two denso classes of genotypes can be unambiguously identified under field conditions, as the dwarfing allele confers a distinctive prostrate juvenile growth habit (Haahr and von Wettstein, 1976).

Field Experiments
The performance of 94 RILs and the parents was examined in field experiments in 1996 and 1997, conducted on alluvial clay soil in Wageningen (52°N lat), Netherlands. Crops were grown under conditions to ensure that plants were free of pests, diseases, and weeds. To avoid any confounding effect of split fertilizer application (i.e., an application on a date that would have been given at a different physiological age for these genotypes), fertilizers were all broadcast at the date of sowing. Plants were grown in plots of 10 rows, spaced 13 cm apart and 9 (1996) or 8 m (1997) long. A randomized incomplete block design was used, with two replicates of each genotype. In 1996, however, four replicates were used for 10 preselected genotypes. Each genotype had one plot (1996) or two adjacent plots (1997) in each replicate. Two plots were used in 1997, one for individual periodic destructive samplings and one for the final harvest for grain yield. Additional information on the experiments is given in Table 1 .


View this table:
[in this window]
[in a new window]
  		Table 1 Selected details of field experiments conducted in 1996 and 1997 (Wageningen, Netherlands)

 
Three phenological stages were recorded: (i) emergence date, when first leaf tip of approximately 80% of the plants was visible, (ii) flowering date, when approximately 80% of spikes in a plot were almost fully emerged, and (iii) maturity date, when approximately 80% of spikes in a plot had turned yellow.

Leaf area index (LAI) and dry matter of individual shoot organs during crop growth were determined by destructive samplings. Each sample contained plants from 0.5-m-long sections of the two (1996) or four (1997) central rows. Dry matter was determined after oven-drying at 105°C for at least 15 h. Leaf area was measured with a LI-1000 leaf area meter (LI-COR, Lincoln, NE). In 1996, measurements for all genotypes were performed at flowering and maturity. For the 10 preselected genotypes, additional measurements were taken at 42 d after emergence (DAE) and 82 DAE. In 1997, samplings were taken at 17, 27, 38, and 50 DAE, flowering, 14 d after flowering, and maturity.

Leaf nitrogen content (LNC) was determined by micro-Kjeldahl digestion and distillation on green leaf blades after oven-drying at 70°C to constant weight. This was done for all genotypes at flowering in both years, and for the parents at other sampling times in 1997.

Grain yield, expressed at 140 g kg-1 moisture content, was determined from all plants in a complete plot (1997) or the remaining plants in a plot (1996). Kernel number per hectare was calculated from yield and kernel weight, which was determined from 250 kernels.

As some genotypes, especially the tall denso class, lodged after flowering in 1997, lodging severity was scored during grain filling, based on relative area and height of lodged plants. The lodging score was valued between 0 and 1, with 0 for an upright canopy and 1 for a completely lodged canopy.

The Model
The model used in our study, referred to as SYP-BL (simulator of yield potential for barley), is based on routines for simulating biomass production in widely used models SUCROS (Goudriaan and van Laar, 1994) and ORYZA1 (Kropff et al., 1994).

The model quantifies barley growth as affected by radiation, temperature, and plant N status. Leaf photosynthesis is calculated based on radiation flux and specific leaf N [the ratio of LNC to specific leaf area (SLA)]. The vertical distribution of both radiation and specific leaf N in the canopy is described by an exponential function. Total daily crop photosynthesis is calculated by integrating instantaneous photosynthesis rates over the LAI and over the day. Daily growth rate is calculated by subtraction of dark respiration rates from daily photosynthesis rates. The biomass produced is distributed among the organs based on partitioning coefficients that are the function of the development stage (DS). The DS is defined as 0 at emergence, 1 at flowering, and 2 at maturity. Daily development rates are assumed to increase proportionally with the effective temperature between 0 and 26°C. The model simulates LAI development in two phases. Before canopy closure, LAI increases exponentially as a function of temperature. The relative growth rate of the leaf area is obtained by linear regression of the log-transformed LAI against the accumulated daily effective temperatures. After canopy closure, the increase in LAI is estimated from SLA and the increase in leaf biomass. Reserve carbohydrates temporarily stored in leaves and stems are remobilized and added to the assimilates available for kernel growth. To simulate the effect of lodging, a new subroutine was developed. As the detrimental effect of lodging is due to self-shading and reduction in canopy photosynthesis (Setter et al., 1997), a completely lodged canopy is assumed to be a single big horizontal leaf, and a single-leaf photosynthesis model is used to estimate assimilates produced by a lodged canopy. Assimilates produced in the actual crop are calculated by linear interpolation between values of estimated photosynthesis for normal and completely lodged canopies, using lodging scores observed at different times.

The model has 10 critical parameters determining genotypic differences: lodging score, preflowering duration (Pre-F), postflowering duration (Post-F), LNC, SLA, relative growth rate of leaf area, the fraction of remobilized reserves in leaves and in stems, and the fraction of shoot biomass partitioned to leaves (FPleaf) and to spikes (FPspike). The fraction of partitioning to stems is calculated in the model as 1 - FPleaf - FPspike. As root weight was not measured in the experiments, biomass partitioning between root and shoot was based on Penning de Vries et al. (1989) and Schut (1992). The fraction of remobilized reserves was estimated from the weight loss of stems and leaves after maturity (Kropff et al., 1994). The maximum weight is assumed to occur at flowering for leaves and at 14 d after flowering for stems (Gallagher et al., 1975; Austin et al., 1980).

Analytical Approach
Analysis of variance was carried out on yield and individual traits using the General Linear Model (SAS Inst., 1988). A covariate was included in the analysis whenever necessary.

The SYP-BL model was evaluated using both 1996 and 1997 experiments. Model parameters were all estimated from the 1997 experiment, because of more detailed measurements in 1997. As the observed developmental pattern of LNC did not differ much between the parents, the average pattern of LNC measured for the parents was used for all RILs, using the observed LNC at flowering as the reference value.

Model analysis was performed to identify critical traits for high yield potential. This was achieved by examining yield variation explained by the SYP-BL model when observed values of a parameter for genotypes were used in the model, while fixing other parameters at their across-genotype mean value. Results were compared with those from linear regression of yield against individual parameters. As FPleaf and FPspike are DS-dependent, their value at the stage when their genotypic difference is most evident (Tables 2 and 3) was used in regression.


View this table:
[in this window]
[in a new window]
  		Table 2 The intercept (a), slope (b), and r2 for simple linear regression of yield in barley against each of six plant traits observed in 1997

 

View this table:
[in this window]
[in a new window]
  		Table 3 Linear regression of barley yield (Y) against lodging (L) and another parameter trait observed in 1997 (n = 96; for definition of these traits, see Table 2)

 

   Results and discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
Conclusions
 REFERENCES
 
Yield Differences
Yields were generally higher in 1997 than in 1996 (Fig. 1) . There was no yield difference between replicates in 1997 (P > 0.05), but significant differences between replicates occurred in 1996 (P < 0.001). This significance can be explained largely by differences in plant N status, because LNC of replicates as a covariate significantly affected yield in 1996 (P < 0.001). Differences in plant N status in 1996 reflected heterogeneity of soil fertility in the field as applied fertilizer levels were low (Table 1). In 1997, no difference in LNC between replicates was observed (P > 0.05) and LNC as a covariate did not affect yield (P > 0.10).



View larger version (42K):
[in this window]
[in a new window]
  		Fig. 1 Yield of recombinant inbred lines (RILs) of barley in replicate plots in (A) 1996 and (B) 1997. `Apex' (solid diamonds) and `Prisma' (solid triangles) are the parents

 
Yields of genotypes differed significantly in both 1996 and 1997 (P < 0.001). These differences were greater in 1997 than in 1996 (Fig. 1), supporting the simulation findings of Aggarwal et al. (1997) for rice that selection for high yield should be conducted in high-N environments. The larger genotypic differences in 1997 than in 1996 were due to the larger differences between the two denso classes in 1997. Among 94 RILs, 47 belonged to the short denso class and the other 47 to the tall denso class. In 1997, yield of the short class (8.249 t ha-1) was significantly higher (1.284 t ha-1; P < 0.001) than the tall class. In 1996, the yield difference between the two classes was very small (0.185 t ha-1; P = 0.055). This indicates that the yield advantage of dwarfing genes is more fully expressed in high-N environments. The advantage of the short genotypes in high-N environments was due at least in part to their greater resistance to lodging compared with the tall ones, which lodged heavily in 1997. A further experiment would be needed to explore the extent to which the short class could have an advantage over the tall class if the latter had not lodged in high-N environments.

Model Performance
Of the 10 parameters determined, relative growth rate of leaf area and the fraction of remobilized leaf and stem reserves did not differ significantly among genotypes (P > 0.05). The use of measured values for these parameters created great noise in predicting yields (data not shown). Therefore, across-genotype means were used in the model (0.0148 for relative growth rate of leaf area, 0.45 for fraction of reserves in leaves, and 0.30 for the fraction in stems.)

Model performance in predicting total shoot biomass, LAI, and yield in 1997 is shown in Fig. 2 . For biomass measured at different stages, the model did not predict the genotypic differences accurately (Fig. 2A). This could be due to inaccurate simulation of LAI (Fig. 2B). When observed LAI was used as inputs, the model better simulated biomass production for stages before maturity (Fig. 3) , but the explained percentage of observed variability gradually decreased with DS. This trend suggests that inaccuracies obtained at each modeling step accumulated, resulting in greater inaccuracy in predicting final biomass. The model explained 37.8% of the variation in final yield (Fig. 2C), but the use of observed LAI did not improve yield prediction (result not shown).



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 2 Comparisons between observed and simulated values of (A) shoot biomass, (B) leaf area index (LAI), and (C) grain yield of barley for 1997 (for each correlation, n = 96). The diagonal line is the 1:1 line

 


View larger version (25K):
[in this window]
[in a new window]
  		Fig. 3 Comparisons between observed and simulated values of barley shoot biomass for 1997, using observed leaf area index as model inputs (for each correlation, n = 96). The diagonal line is the 1:1 line

 
For simulating yield in the 1996 experiment, first, all parameter values (except LNC) established from the 1997 experiment were used. Because LNC in 1996 differed between replicates for a given genotype, simulation was conducted on individual replicate basis with measured LNC as input. The model accurately predicted time to flowering in 1996, but considerably overestimated time to maturity (results not shown). This could be due to the difference in plant N status between the two years, because Post-F depends on plant N status as a result of N translocation from vegetative organs to meet N requirement for grain growth (Sinclair and de Wit, 1975). In the subsequent simulation, observed Post-F was used as input to avoid any artifacts induced by inaccurate simulation of maturity dates. Model performance with the observed Post-F of 1996 is given in Fig. 4 .



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 4 Comparisons between observed and simulated values of (A) leaf area index (LAI), (B) shoot biomass, and (C) yield of barley for 1996 (n = 40 for measurements at 42 and 82 DAE, and n = 212 for measurements at flowering and maturity). The diagonal line is the 1:1 line

 
The model considerably overestimated LAI (Fig. 4A). This could be due to differences in SLA between years (Fig. 5) , perhaps because leaves are thicker in a lower-N environment (Grindlay, 1997). However, the model did not strongly overestimate biomass (Fig. 4B) and final yield (Fig. 4C). This is because effects of overestimation of LAI were canceled out by effects of underestimation of leaf photosynthesis, which is related to SLA. For shoot biomass measured at different times, model assessments showed similar results as in 1997. For yield, the model explained 25.8% of observed variation (Fig. 4C).



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 5 Relation between specific leaf area and leaf N content of barley measured at flowering in 2 yr

 
Contribution of Individual Physiological Traits to Yield Potential
To identify the traits that determine yield potential, effects of individual model parameters on yield were analyzed using the 1997 data, because yield potential of genotypes was more fully realized in 1997. Simple single-factor regression based on data of all genotypes (Table 2) revealed that yield correlated strongly and positively with Pre-F, FPleaf, DS = 0.475, and FPspike, DS = 1.15, slightly with Post-F, and negatively with LNC at flowering. No correlation was found with SLA at flowering.

The strong positive relationship of yield with Pre-F, FPleaf, DS = 0.475, and FPspike, DS = 1.15 is likely caused by the denso gene, as the significance of correlation became very minor or even disappeared when the analysis was conducted separately for each denso class (Table 2). Within genotypes with the dwarfing denso allele, yield increases only slightly in response to Pre-F and FPspike, DS = 1.15. Within the group of tall genotypes, yield correlated with none of the traits. The denso gene affects a large number of agronomic traits (Powell et al., 1985). Our results suggest that physiological model-input traits are associated with the denso gene as well, as confirmed by our genetic mapping study (Yin et al., 1999).

Because the denso gene largely determined lodging occurrence in 1997 and lodging score correlated strongly with Pre-F, FPleaf, DS = 0.475, and FPspike, DS = 1.15 (r = -0.89, -0.82, and -0.61, respectively; P < 0.001), regression was conducted relating yield to each parameter using lodging score as a covariate (Table 3). When the effect of lodging was included in the regression model, yield correlated significantly only with Pre-F and FPspike, DS = 1.15.

Effects of individual traits on yield were further analyzed using SYP-BL. Simulations showed that the model explained 53.6% of yield variation when only observed lodging score was used in the model while fixing all other parameters at their respective across-genotype means (Table 4) . Subsequently, other parameters were input in the model. Yield variation explained by the model was improved only when genotype-specific values for Pre-F or FPspike were used in the model. Use of genotype-specific values of other parameters reduced the percentage of model's explanation, indicating that these traits were not responsible for yield differences among genotypes at high N inputs. Best model performance was achieved when averaged values for Post-F, SLA, LNC, and FPleaf and genotype-specific values for lodging, Pre-F, and FPspike were used, explaining 64.7% of yield variation (Table 4).


View this table:
[in this window]
[in a new window]
  		Table 4 Percentage of the observed yield variation in barley explained by the SYP-BL model when different sets of values for seven parameters were used in the model (1997 data; n = 96)

 
The above results show that single-factor regression overestimated the role of individual traits. The relative importance of these traits identified by regression using lodging as a covariate, however, agreed well with the results obtained from the SYP-BL model analysis. The SYP-BL model relates traits to yield in an ecophysiological way and considers interactions among growth processes and nonlinear responses of these processes to environments; the regression analysis does not. Overall, source-related parameters, determining LAI and leaf photosynthesis, were not important. The negative response of yield to LNC contradicts the expectation that high leaf N gives high yields by favoring a high photosynthesis. When leaf N is at a high level, as was the case in 1997, N does not limit yield and yield may not be affected by N uptake.

A strong positive effect of Pre-F and a small effect of Post-F on yield were also reported by Dofing (1997), based on a path analysis of data for 24 barley cultivars. Dofing (1997) indicated that the importance of Pre-F was due to its strong association with kernel number per unit area (sink), as long Pre-F provides sufficient time for spike differentiation. Our data support this association between Pre-F and kernel number per unit area (r = 0.77, P < 0.001). The role of Pre-F was identified here with the SYP-BL model, however, which does not account for kernel formation. The results indicate that this process should be added to the model. However, current physiological understanding does not allow an accurate modeling of kernel formation and other sink-related processes (Bindraban, 1997).


   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
REFERENCES
 
This study explored the ability of a crop model to explain yield differences among diverse genotypes in an RIL population of spring barley, and analyzed the physiological basis of these differences. The following conclusions were formulated.

  1. The denso dwarfing gene segregating in the population was a major factor causing yield differences among genotypes. The advantage of short denso genotypes was expressed most strongly in high-N environments. This gene largely determined lodging occurrence in high-N environments, and it affected some physiological model-input traits, including Pre-F, FPleaf, and FPspike.
  2. Besides lodging, the two most important modeled traits to explain yield differences at high N levels were Pre-F and FPspike, which directly and positively affected yield. Other traits, including Post-F, LNC, SLA, and FPleaf, did not affect yield potential, since use of their measured genotype-specific values reduced accuracy of the model's yield prediction compared with the use of their across-genotype means. These conclusions from modeling analysis were similar to those obtained from multiple regression analysis. When measured values of lodging, Pre-F, and FPspike were used in the model with other parameters fixed at their across-genotype means, the model explained 65% of yield variation among genotypes.
  3. One difficulty in using crop models to explain genotypic differences is that parameters of current models may vary with environment. Two parameters (Post-F and SLA) varied with plant N status. For example, SLA, an important parameter for estimating both leaf photosynthesis and LAI, was lower in the year of low N inputs (Fig. 5). Therefore, the model using SLA values as measured in high-N environments was not capable of predicting the growth of crops with low N inputs although the measured LNC was used. To better assist breeding programs, models have to be improved such that their parameters are solely genetically determined.
  4. Most current crop models simulate yield as determined by source availability. The generally poor performance of the source-based SYP-BL model and no or only slight effect of many source-determining parameters (e.g., Post-F, SLA, LNC, and FPleaf) may reflect that yields with high N inputs are also affected by factors that have not yet been incorporated in the model, especially those related to sink. Some apparent interaction between source and sink was highlighted in this study. For example, an increase in kernel number can be achieved by a prolonged Pre-F. To improve the ability of models to predict yield differences among genotypes, quantitative knowledge on the interaction between source availability and sink strength is required. This would allow more effective uses of models to identify options for breeding cultivars with high yield potential.

SAS Institute 1988


   ACKNOWLEDGMENTS
 
We thank Dr. J.W. Schut for developing the RIL population, Dr. C. Grashoff for his valuable suggestions, Drs. P.S. Bindraban and M. van Oijen for their critical reading of the manuscript, and UNIFARM staff of Wageningen University for their assistance during field measurements.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 
Contribution of The C.T. de Wit Graduate School for Production Ecology of Wageningen Univ. Funded by The Netherlands Organization for Scientific Research.

Received for publication February 1, 1999.
   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 J Exp BotHome page
X. Yin, P. C. Struik, J. Tang, C. Qi, and T. Liu
Model analysis of flowering phenology in recombinant inbred lines of barley
J. Exp. Bot., March 1, 2005; 56(413): 959 - 965.
[Abstract] [Full Text] [PDF]

Home page
 J Exp BotHome page
X. Yin, P. C. Struik, F. A. van Eeuwijk, P. Stam, and J. Tang
QTL analysis and QTL-based prediction of flowering phenology in recombinant inbred lines of barley
J. Exp. Bot., March 1, 2005; 56(413): 967 - 976.
[Abstract] [Full Text] [PDF]

Home page
 ANN BOT (LOND)Home page
H.-P. YAN, M. Z. KANG, P. DE REFFYE, and M. DINGKUHN
A Dynamic, Architectural Plant Model Simulating Resource-dependent Growth
Ann. Bot., May 1, 2004; 93(5): 591 - 602.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
L. A. Hunt, M. P. Reynolds, K. D. Sayre, S. Rajaram, J. W. White, and W. Yan
Crop Modeling and the Identification of Stable Coefficients that May Reflect Significant Groups of Genes
Agron. J., January 1, 2003; 95(1): 20 - 31.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
X. Yin, P. Stam, M. J. Kropff, and A. H. C. M. Schapendonk
Crop Modeling, QTL Mapping, and Their Complementary Role in Plant Breeding
Agron. J., January 1, 2003; 95(1): 90 - 98.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Agricola
Right arrow Articles by Yin, X.
Right arrow Articles by Stam, P.
Related Collections
Right arrow Other Grain Crops
Right arrow Crop Models
Right arrow Crop Genetics

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome