Exclusive selection for yield raises, the harvest index of self-pollinated crops with little or no gain in total bipmass. In addition to selection for yield, it is suggested that efficient breeding for higher yield requires simultaneous selection for yield's three major, genetically controlled physiological components. The following are needed: (1) a superior rate of biomass accumulation. (2) a superior rate of actual yield accumulation in order to acquire a high harvest index, and (3) a time to harvest maturity that is neither shorter nor longer than the duration of the growing season. That duration is provided by the environment, which is the fourth major determinant of yield. Simultaneous selection is required because genetically established interconnections among the three major physiological components cause: (a) a correlation between the harvest index and days to maturity that is usually negative; (b) a correlation between the harvest index and total biomass that is often negative, and (c) a correlation between biomass and days to maturity that is usually positive. All three physiological components and the correlations among them can be quantified by yield system analysis (YSA) of yield trials. An additive main effects and multiplicative interaction (AMMI) statistical analysis can separate and quantify the genotype × environment interaction (G × E) effect on yield and on each physiological component that is caused by each genotype and by the different environment of each yield trial. The use of yield trials to select parents which have the highest rates of accumulation of both biomass and yield, in addition to selecting for the G × E that is specifically adapted to the site can accelerate advance toward the highest potential yield at each geographical site. Higher yield for many sites will raise average regional yield. Higher yield for multiple regions and continents will raise average yield on a world-wide basis. Genetic and physiological bases for lack of indirect selection for biomass from exclusive selection for yield are explained.

The hypothesis tested was that lack of photoperiod gene activity allows inherent partitioning of photosynthate to continued growth of the earliest potential buds, flowers, pods, and seeds (the organs that give rise to the yield). Alternatively, and competitively, photoperiod gene activity causes the photosynthate to be partitioned predominantly toward continued growth of new vegetative organs plus later initiation of more reproductive (yield) organs. This hypothesis was tested by comparing an insensitive and a photoperiod-sensitive bean (Phaseolus vulgaris L.) cultivar and their F1 with F2 segregates of undetermined genotype. Randomly derived homozygous F8 segregates were also compared. The F8 generation included one photoperiod-insensitive and one photoperiod-sensitive genotype in a 1:1 ratio, which verified control by one photoperiod gene. Under long daylength (LD), in addition to early versus late flowering and maturity, the two genotypes expressed opposite levels of 23 other traits that would be changed by competitive partitioning of the photosynthate. In contrast, under short daylength (SD), both genotypes flowered and matured early, and both expressed the levels for all 25 traits that the photoperiod-insensitive genotype expressed in both SD and LD. The photoperiod gene interacted with daylength to control the levels of all three major physiological components of yield: the aerial biomass, harvest index, and days to maturity. Included among the other traits with levels altered by daylength-modulated photoperiod gene activity were: the number of branches, nodes, leaves and leaf area, the rate of yield accumulation, and sink activity.

Number of days to flower (DTF) of 78 bean (Phaseolus vulgaris L.) genotypes was measured in tropical fields at various elevations. The associated 18 mean temperatures varied between 12 and 28C. Daylength was natural 12 or 13 hours of sunlight with or without incandescent light for a total of 18 hours. A statistical analysis with additive main effects and multiplicative interaction effects (AMMI) quantified the effects on the deviation from the DTF grand mean caused by each genotype, plus those caused by each daylength and by each temperature. The more photoperiod-sensitive the genotype (factor 1), the more a longer daylength (factor 2) increased DTF and the more a higher temperature (factor 3) synergistically increased DTF. These three factors interacted to delay the node to flower. An additional control over DTF occurred as the same higher temperature (factor 3) reduced the days required to develop a node (factor 4). Thus, a higher temperature tended to decrease DTF by enhancing the rate of vegetative development, at the same time that it tended to increase DTF by enhancing the photoperiod gene activity. This four-factor interaction resulted in a U-shaped curve of DTF in response to temperature. The smallest DTF on the U-shaped response was interpreted as occurring when the simultaneous effects of temperature toward earlier and later DTF exactly cancelled. At all temperatures below this optimum for flowering, a change of temperature changed DTF predominantly by altering the days required to develop a node. At all temperatures above the optimum, a change of temperature changed DTF predominantly by altering the photoperiod-gene-caused delay of the node to flower. The optimum temperature for flowering was lowered by higher sensitivity of the genotype to photoperiod and also by longer daylength.