Loading field map
Crop physiologists have long asked a deceptively simple question: what determines the yield of a crop, and can that knowledge be used to predict or improve it? Over nearly two centuries, the field has produced a series of frameworks that each offered a different answer—some focusing on single limiting factors, others on whole-plant growth, and still others on the intricate interplay of genes, architecture, and environment. This article traces the evolution of those frameworks, showing how each built on, replaced, or reacted against its predecessors.
The first systematic framework, Limiting-Factor Yield Physiology (1840–1920), drew on Justus von Liebig's law of the minimum and F.F. Blackman's concept of limiting factors. It treated yield as constrained by the scarcest resource—nitrogen, water, or light—and sought to identify that single bottleneck. This approach was powerful for diagnosing deficiencies but could not capture how crops integrated multiple resources over a growing season.
Classical Growth Analysis (1919–1970) superseded the limiting-factor view by measuring whole-plant growth rates, leaf area, and net assimilation rate. Pioneered by F.G. Gregory and others, it replaced the search for a single limiting factor with quantitative descriptions of crop growth over time. Using periodic harvests, researchers calculated relative growth rate, leaf area ratio, and net assimilation rate. This framework was more integrative than its predecessor, but it remained largely descriptive—it could say how fast a crop grew, but not why.
Process-Based Crop Physiology (1950–Present) shifted the focus from describing growth to explaining it through underlying mechanisms. Researchers began measuring photosynthesis, respiration, transpiration, and nutrient uptake as independent processes that together determined yield. This framework coexisted with Classical Growth Analysis for decades, gradually absorbing its methods while adding mechanistic depth. It remains a cornerstone of the field today because it provides the causal understanding needed to diagnose yield gaps and guide management.
A major breakthrough came with Crop Radiation Capture and Use Efficiency (1961–Present), which narrowed the process-based approach to a simple but powerful relationship. John Monteith showed that crop biomass is proportional to the amount of photosynthetically active radiation intercepted by the canopy, multiplied by a conversion efficiency. This framework gave researchers a straightforward way to predict potential yield from solar radiation and leaf area, and it remains widely used for benchmarking and modeling.
At roughly the same time, Source-Sink and Harvest Index Physiology (1962–Present) addressed a different dimension: how assimilates are partitioned to harvested organs. It distinguished between source tissues (leaves that produce carbohydrates) and sink tissues (grains, tubers, or fruits that store them), and introduced the harvest index—the fraction of total biomass allocated to yield. This framework complemented radiation-use efficiency by explaining why two crops with similar biomass could have very different yields. Together, these three process-based frameworks—mechanistic physiology, radiation capture, and source-sink partitioning—form the core of modern crop physiology.
Dynamic Crop Simulation Modeling (1965–Present) subsumed Classical Growth Analysis by integrating process-based knowledge into computer models that simulate crop growth day by day. The Wageningen school (led by C.T. de Wit) and later DSSAT created models that combined weather, soil, and management inputs to predict yield, water use, and nitrogen dynamics. These models absorbed the equations of classical growth analysis and process physiology, turning them into predictive tools. Dynamic simulation remains active today, used for climate impact assessment, precision agriculture, and decision support.
Crop Ideotype and Physiological Breeding (1968–Present) applied physiological understanding directly to plant breeding. Donald's ideotype concept proposed breeding for an ideal plant architecture—short stems, erect leaves, high harvest index—rather than selecting solely for yield. This framework narrowed the focus of process physiology to traits that breeders could select, and it coexists with dynamic modeling by providing the trait targets that models can evaluate. It remains influential in breeding programs for cereals and other crops.
Functional-Structural Plant Modeling (1995–Present) reacted against the aggregated nature of dynamic simulation models. Instead of treating the canopy as a uniform layer, it explicitly represents individual organs—leaves, stems, roots—and their three-dimensional arrangement. This allows researchers to simulate how local light, water, and nutrient capture affect whole-plant growth. Initially a challenge to the lumped-parameter approach of dynamic models, functional-structural modeling later influenced those models by providing more realistic sub-models for canopy architecture and resource competition.
Crop Phenomics and Systems Biology (2010–Present) represents the latest attempt to bridge scales. It combines high-throughput phenotyping (measuring traits like leaf angle, chlorophyll content, and root depth in thousands of plants) with genomic data and systems models. The goal is to understand how genetic variation translates into physiological performance and yield. This framework extends process-based physiology by adding a genetic dimension, and it interacts with both dynamic simulation and functional-structural modeling by providing data to parameterize and validate them.
Today, seven frameworks remain active: Process-Based Crop Physiology, Crop Radiation Capture and Use Efficiency, Source-Sink and Harvest Index Physiology, Dynamic Crop Simulation Modeling, Crop Ideotype and Physiological Breeding, Functional-Structural Plant Modeling, and Crop Phenomics and Systems Biology. They broadly agree that yield is determined by three factors: the capture of resources (light, water, nutrients), the efficiency with which those resources are converted into biomass, and the partitioning of that biomass to harvested organs. They also agree that understanding these processes requires a combination of measurement and modeling.
Where they disagree is on the level of detail needed. Dynamic simulation models favor aggregated representations that are computationally efficient and suitable for regional predictions. Functional-structural modelers argue that ignoring spatial heterogeneity misses critical interactions, especially under stress. Crop phenomics and systems biology advocates insist that without genetic information, physiological models cannot guide breeding. These disagreements are productive: they drive the field toward more integrated approaches that combine the strengths of each framework. The leading frameworks today are not in competition but in a state of productive tension, each best suited to different questions—prediction, mechanistic understanding, or genetic improvement.