R. J. Lawn

Towards the Reliable Prediction of Time to Flowering in Six Annual Crops. I. The Development of Simple Models for Fluctuating Field Environments

Despite numerous altempts, the development of generalixed models capable of accurate predictions of the times from sowing to flowering ( f ) of crop plants in field environments has remained elusive. Models which seek to correlate ;ʃ with various states of environmental factors such as photoperiod, P , and temperature, T , using formal statistical procedures arc often complex. Here, we describe a family of photothermal responses (involving unambiguous parameters and limits) which quantify the linear, non-interacting effects of P and T not on ʃ but on 1/ ʃ , i.e. on the rate of progress towards flowering. Based on these relations we suggest that the modelling of crop phenology will be simplified, more reliable and more biologically plausible.

Improving genotypic adaptation in crops - a role for breeders, physiologists and modellers

Approaches using breeding, physiology and modelling for evaluating adaptation of plant genotypes to target environments are discussed and methods of characterizing the target environments outlined. Traditional approaches, and their limitations, to evaluation of genotypic adaptation using statistical and classificatory techniques with a phenotypic model are discussed. It is suggested that a simple biological model is the most appropriate framework in which to integrate physiology and modelling with plant breeding. Methods by which physiology and modelling may contribute to assessment of adaptive traits and to selection for adaptation in a breeding programme are considered.

Agronomic and Physiological Constraints to the Productivity of Tropical Grain Legumes and Prospects for Improvement

Growth and development of the tropical grain legumes are generally highly sensitive to photo-thermal regime, so that seasonal and regional effects on phenology and yield potential can be large. Yet failure adequately to recognize and fully exploit the consequences of genotype × latitude/sowing date × density interactions has frequently constrained both agronomic and genetic advance with these species. Thus there is opportunity for short term productivity improvements through agronomic strategies which accept the implications of phenological plasticity, and seek to optimize management practices such as sowing date and sowing density, to exploit more effectively the yield potential and broaden the adaptation of existing cultivars. The greatest physiological potential for genetic improvement in the productivity of the tropical grain legumes lies not with increasing total biomass, but with increasing the proportion of bio-mass partitioned into seed, i.e. with improved harvest index (HI). There are difficulties in selecting for higher HI, including its association with phenology, although the latter provides the most powerful tool for manipulating HI in the short term. As with the cereals, more productive genotypes are likely to be characterized by reduced sensitivity to photothermal conditions, shorter growth duration, a more synchronous reproductive ontogeny, and greater HI, than is typical of traditional cultivars. Accordingly, optimal sowing densities will be higher, and crops will need more inputs, better management and more effective protection, than afforded to them in subsistence systems of production. Moreover, the increases in productivity will be achieved, at least partially, at the expense of the potential for yield homeostasis in adverse environments, and of non-seed components such as forage, fuel or tubers - attributes that are often highly valued in subsistence agriculture where many of the tropical food legumes are presently grown. Improvement programmes will therefore need to adopt strategies in particular situations which reflect local perceptions of the importance of productivity improvement relative to that of minimizing inputs and risk.

Towards the reliable prediction of time to flowering in six annual crops. VI. Applications in crop improvement

Variation in time from sowing to flowering (f) was examined for 44 cultivars of soyabean, mungbean, black gram, ricebean, cowpea, chickpea, lentil and barley, when grown in up to 21 diverse environments obtained by making one or more sowings at each of six locations spanning tropical, sub-tropical and temperate climates in Australia. The utility of simple linear models, relating rate of development (1/f) towards flowering to mean photoperiod and temperature prevailing between sowing and flowering, was evaluated. The models were highly efficient, explaining most (86.7%) of the variation observed across species, cultivars and environments. They were particularly efficient in describing responses where cultivars were relatively well-adapted, in agronomic terms, and least efficient where cultivars were exposed to unfavourable temperature and, to a lesser extent, photoperiod. Opportunities for exploiting the models in applied crop improvement include their use in interpretation of G × E interaction, genotypic characterization and selection of parental genotypes, selection of test environments, designing screening procedures, and more efficiently matching genotypes to target environments. The main strengths of these linear, additive rate models in crop improvement are their wide applicability across species and genotypes, their relative simplicity, and the requirement for few genotype-specific response parameters. Their main weakness is their lack of precision in describing responses when plants are exposed to unfavourable photothermal extremes, albeit in circumstances that are sometimes unrealistic for cropping those particular genotypes

Use of field observations to characterise genotypic flowering responses to photoperiod and temperature: a soyabean exemplar

Thirty-nine accessions of soyabean [Glycine max (L.) Merrill] and 1 of wild annual soyabean (Glycine soja L.) were sown at two sites in Taiwan in 1989 and 1990 and on six occasions during 1990 at one site in Queensland, Australia. On two of the occasions in Australia additional treatments extended natural daylengths by 0.5 h and 2 h. The number of days from sowing for the first flower to appear on 50% of the plants in each treatment was recorded (f), and from these values the rate of progress towards flowering (1/f) was related to temperature and photoperiod. In photoperiod-insensitive accessions it was confirmed that the rate is linearly related to temperature at least up to about 29°C. In photoperiod-sensitive genotypes this is also the case in shorter daylengths but when the critical photoperiod (Pc) is exceeded flowering is delayed. This delay increases with photoperiod until a ceiling photoperiod (Pce) is reached. Between Pc and Pce, 1/f is linearly related to both temperature (positive) and photoperiod (negative), but in photoperiods longer than Pce there is no further response to either factor. The resulting triple-intersecting-plane response surface can be defined by six genetically-determined coefficients, the values of which are environment-independent but predict time to flower in any environment, and thus quantify the genotype x environment interaction. By this means the field data were used to characterise the photothermal responses of all 40 accessions. The outcome of this characterisation in conjunction with an analysis of the world-wide range of photothermal environments in which soyabean crops are grown lead to the following conclusions: (1) photoperiod-insensitivity is essential in soyabean crops in temperate latitudes, but such genotypes flower too rapidly for satisfactory yields in the tropics; (2) photoperiod-sensitivity appears to be essential to delay flowering sufficiently to allow adequate biomass accumulation in the warm climates of the tropics; (3) contrary to a widely held view, some degree of photoperiod-sensitivity is also needed in the tropics if crop-duration homeostasis is required where there is variation in sowing dates (this is achieved through a photoperiod-controlled delay in flowering which counteracts the seasonal increase in temperature that is correlated with increase in day-length); and (4) a greater degree of photoperiod-sensitivity is necessary to provide maturity-date homeostasis for variable sowing dates — a valuable attribute in regions of uncertain rainfall. Since the triple-intersecting-plane response model used here also applies to other species, the use of field data to characterise the photothermal responses of other crops is discussed briefly.

Towards the reliable prediction of time to flowering in six annual crops. V: Chickpea (Cicer arietinum)

Four genotypes of ‘desi’ and two of ‘kabuli’ chickpea (Cicer arietinum) were sown at six locations in Australia on various dates between 1986 and 1988, giving 22 combinations of site and sowing date with diverse photothermal environments. Times from sowing to first flowering (f) varied from 30 to 162 d, mean pre-flowering temperatures from 10.8° to 29.3°C and mean photoperiods from 11.3 to 15.6 h d−1. There was no evidence that any observation had been obtained in photoperiods shorter than the ceiling photoperiod (Pcc) or longer than the critical photoperiod (Pc). This suggests that, in typical agricultural environments, chickpea crops experience photoperiods (P) which should satisfy the condition Pcce < P < Pc. In one ‘kabuli’ and two ‘desi’ genotypes, 1/f was influenced by both temperature and photoperiod. The coefficient of determination (R2) for a linear, additive rate of development model ranged from 0.74 to 0.80, with no significant difference in either temperature sensitivity or photoperiod sensitivity among these three genotypes. In the remaining three genotypes, no significant response to temperature was detected between 10.8° and 29.2°C, so rate of progress to flowering was influenced solely by photoperiod. There was no significant difference in the sensitivity of 1/f to P among these three genotypes. The linear, additive rate model found here to be so satisfactory as a predictive tool for phenology, is also shown to have much wider general application.Prediccion del liempo de floracion

Time to flowering of mung bean (Vigna radiata) genotypes and their hybrids in response to photoperiod and temperature.

(...) The analyses suggest that all the genotypes and F 1 s tested were quantitative short day plants, with the possible exception of one line which may be day neutral. The analyses further suggest that time to flowering was influenced by four genotype-specific attributes: minimum time to flowering, critical photoperiod (P C ), responsiveness to photoperiod longer than P C , and responsiveness to temperature. Among the genotypes tested, different values were observed for each attribute (...)

Towards the Reliable Prediction of Time to Flowering in Six Annual Crops. IV. Cultivated and Wild Mung Bean

Eight genotypes of cultivated mung bean, black gram and rice bean ( Vigna mungo, Vigna radiata ssp. radiata and Vigna umbellata , respectively) were sown at six sites in Australia on various dates in order to provide a range of photothermal environments. In addition, four accessions of the related wild species Vigna radiata ssp. sublobata were sown on five occasions. Times from sowing to first flowering ( f ) varied between environments from 34 to 317 d; pre-flowering temperature and photoperiod means ranged from 12.7° to 29.1°C and from 11.8 to 15.5 h d −1 . No effect of photoperiod was detected on rate of progress towards first flowering (1/ f ) in four genotypes, but in each case a significant positive relation was detected between 1/ f and mean temperature. These simple thermal time relations did not differ significantly among these four genotypes; the common base temperature was 7.9°C. In two genotypes observations were well described by a thermal response plane when the mean photoperiod was less than 13 h d −1 (p −1 delayed flowering. In each of the remaining genotypes the observations were best described by photothermal planes, that is, 1/ f was modulated by temperature and photoperiod. Predictions from the models based on our data were in good agreement with the times to first flowering observed in three genotypes in an earlier controlled environment study.

Environmental Modulation of Flowering in Mung Bean ( Vigna radiata ): a Reappraisal

The phenology of mung beans is extremely plastic; responsiveness to both photoperiod and temperature is known to modulate flowering. A previous conclusion, perpetuated uncritically now for almost forty years, has been that both (quantitative) short- and long-day flowering responses exist in the mung bean germplasm. However, our re-analysis of the original data leads us to an alternative conclusion: that genotypes of mung bean are quantitative short-day plants with different optimum mean diurnal temperatures for flowering. This alternative interpretation (which is plausible biologically and in evolutionary terms) is discussed.

Environmental Modulation of Flowering in Mung Bean ( Vigna radiata ): Further Reappraisal for Diverse Genotypes and Photothermal Regimes

Genotypes of mung bean commence flowering at very different times depending on sowing date and location, but relatively little is known about the modulation of flowering by environmental factors. Previous and frequently cited conclusions have been that: (a) genotype, photo-period and temperature all interact to determine relative earliness to flower, and that (b) the genetic control of these photothermal responses is seemingly complex. However, our reanalyses of original data in terms of rates of progress towards flowering, 1/ f , rather than the traditional approach based on days from sowing to flowering, f , show that the photothermal modulation of flowering in mung bean can be described by a series of simple, linear models, and that interaction terms involving photoperiod and temperature are often insignificant. The merits and implications of this alternative analysis and interpretation of original data are discussed.