I was excited to see a recent meta-analysis dealing with plant reproduction under conditions of elevated CO2 (Jablonski et al., 2002). A highly needed assessment. Literature-based syntheses of this type have the potential to contribute more to our understanding and the development of theory than often costly individual experiments. However, I was disappointed to find that this assessment will be of much less help than I had hoped because the analysis did not strictly stratify data based on the fertility of the growth conditions. A meta-analysis on CO2 responses provides a rather limited advance unless we have a completely clear picture of the resource status of the test plants.

With adequate moisture, treatment variables other than nutrition are of almost negligible interest. For instance, CO2 enrichment technology does not usually emerge as a major driver of responses. Pot size may matter, but it is not only the pot or its absolute size, but the pot size/plant size ratio and the amount of nutrients flushed through which are critical. Small pots (with a soft substrate) may become functionally large pots if highly fertilized, and species may show contrasting responses (McConnaughay et al., 1996). So, selecting pot vs nonpot data misses the key issue, nutrition.

Furthermore, the analysis is based on 75% crop data and 25% wild species data. Because half of the wild species were grown like crops, and since a selection criterion was for species which completed a full life cycle (or most of it) under CO2 enrichment, the results are, at a level of up to 88%, based on short-lived plants grown under nonnutrient limited conditions. Therefore the statistical power was insufficient to allow any conclusions to be drawn regarding nonfertilized, perennial wild plants.

The authors are not to blame for this bias in the experimental work available. However, it would have been better if the paper had made it obvious that this problem exists and that we have very little data from which to draw valid conclusions about reproductive efforts under elevated CO2 for plants tested where they grow with no amendments other than CO2 availability included (Thürig et al., 2003).

I hope that it becomes more widely acknowledged that meta-analysis on aspects of ecological CO2 research must account for the resource status of the test plants – otherwise I see little advance in our understanding. This was also a shortcoming of earlier attempts (Curtis & Wang, 1998; Wand et al., 1999; Kerstiens, 2001).

The next most important criterion by which data should be grouped is plant age. I suggest a shift in emphasis in data treatment from technology-oriented or taxonomy-oriented criteria to those which control sink activity of plants – nutrition, moisture, developmental stage (Körner, 2001). As an example, the legume vs nonlegume separation of CO2 responses does not often lead to differentiation in cases where plants are grown on substrates with natural (poor) P supply. The ‘legume CO2 effect’ is an agricultural phenomenon at best rarely seen in nature, unless P is added or is naturally high – not a very common situation in late successional plant communities (Stöcklin & Körner, 1999).


  1. Top of page
  2. References
  • Jablonski LM, Wang X, Curtis PS. 2002. Plant reproduction under elevated CO2 conditions: a meta-analysis of reports on 79 crop and wild species. New Phytologist 156: 926.
  • Kerstiens G. 2001. Meta–analysis of the interaction between shade-tolerance, light environment and growth response of woody species to elevated CO2. Acta Oecologica 22: 6169.
  • Körner C. 2001. Experimental plant ecology: some lessons from global change research. In: PressMC, HuntlyNJ, LevinS, eds. Ecology: achievement and challenge. Oxford, UK: Blackwell Science, 227247.
  • McConnaughay KDM, Nicotra AB, Bazzaz FA. 1996. Rooting volume, nutrient availability, and CO2-induced growth enhancements in temperate forest tree seedlings. Ecological Applications 6: 619627.
  • Stöcklin J, Körner Ch. 1999. Interactive effects of CO2, P availability and legume presence on calcareous grassland: results of a glasshouse experiment. Functional Ecology 13: 200209.
  • Thürig B, Körner Ch, Stöcklin J. 2003. Seed production and seed quality in a calcareous grassland in elevated CO2. Global Change Biology 9: (In press.)
  • Wand SJE, Midgley GF, Jones MH, Curtis PS. 1999. Responses of wild C4 and C3 grass (Poaceae) species to elevated atmospheric CO2 concentration: a metaanalytic test of current theories and perceptions. Global Change Biology 5: 723741.