• CO2 uptake;
  • controlled experiments;
  • methodology;
  • nutrient uptake;
  • relative addition rate;
  • relative growth rate;
  • relative uptake rate;
  • steady-state;
  • transpiration

The generally accepted norm for controlled experiments with plants has traditionally been restricted to a simplified maintenance of a constant state of the surrounding medium. The ‘treatments’ and a ‘control’, as a reference for comparisons, have been applied without a clear specification of causality and function. However, for pure mathematical reasons, plants are in a constant physiological state (steady-state) only when the internal concentrations of carbon and nutrients and the ratios between them remain constant. Consequently, experimental conditions are controlled in a meaningful sense only when carbon and nutrients are taken up with a constant relative uptake rate, Ru, that is equal to the relative growth rate, RG. The successful application of these fundamental rules to experimental procedures has been verified in numerous experiments. As for acclimatized plants in steady-state, responses to limiting factors have been graded relative to maximum performance, which constitutes the ultimate reference value. Despite the comprehensive and straightforward material of evidence, the methodology is still a matter of argument.

It is concluded that the driving variable commonly used for uptake, external concentration, should be abandoned since its form (lack of time dimension) makes it incompatible with the uptake rate. This calls for a shift of paradigm since the failure of recognition both of fundamental relationships and of experimental shortcomings has been harmful to the development of a sound scientific basis for specification of plant properties. Although fragmented knowledge of basic mechanisms may be good, understanding of causes and effects is poor in relationships between environmental, internal and inherited conditions on the one side, and plant performance on the other. A pseudo-science has in an empirical ‘trial-and-error’ research been based upon statistical evaluation of data of low quality. This is a problem of major significance that needs open discussion and public attention.