Background
Stomatal density (SD; the number of stomata per unit leaf area) and stomatal index (SI; the number of stomata divided by the sum of the numbers of stomatal and epidermal cells) are often used as proxies of past atmospheric CO2 concentrations.  The evidence for the validity of this technique is said by Reid et al. to rest primarily upon paleontological data (Woodward, 1987; Retallack, 2001) and growth chamber studies (Royer, 2001).  Hence, they sought to broaden the experimental basis for the protocol by adding field studies to the mix of evidence supporting it.

What was done
The authors examined the responses of SD and SI to atmospheric CO2 concentration in 15 species of woody, herbaceous and annual plants after four years of exposure to a field CO2 gradient ranging from 200 to 550 ppm (Gill et al., 2002), as well as in three FACE studies of equal duration (DeLucia et al., 1999; Smith et al., 2000; Norby et al., 2001).

What was learned
In the words of the authors, "SI was not correlated with CO2 either for any single species or for the pooled data."  SD, on the other hand, "showed a weak positive correlation with CO2," which was in the opposite direction to that assumed in prior studies.

What it means
In view of these findings, Reid et al. conclude that SI and SD "are unlikely to decline in response to future high CO2, and may not have responded significantly to low CO2 in the past," which clearly suggests they are not good proxies for estimating paleoatmospheric CO2 concentrations.

Conflicting evidence
In spite of these observations, there are a number of other equally reputable "field" studies that have observed what Reid et al. could not detect.  Lin et al. (2001), for example, studied Scots pine (Pinus sylvestris) seedlings growing out-of-doors in open-top chambers maintained at atmospheric CO2 concentrations of 350 and 750 ppm for four years, finding that needle stomatal density was reduced by an average of 7.4%.  Studying mature white oak (Quercus ilex) trees growing in the vicinity of CO2-emitting springs in central Italy, Paoletti et al. (1998) also observed significant reductions in stomatal density (by a factor of nearly 1.5 over a CO2 concentration range of 350 to 750 ppm).  Likewise - based on stomatal density measurements made on needles of four conifer species (Tsuga heterophylla, Picea mariana, Picea glauca and Larix laricina) collected from living trees, herbarium samples and well-dated peat cores that could be assigned atmospheric CO2 concentrations corresponding to the times of the needles' creation on the basis of historical atmospheric CO2 measurements and CO2 measurements of air bubbles trapped in Antarctic ice cores - Kouwenberg et al. 2003 observed stomatal density reductions ranging from 4.5% to 8.1% in response to the 80 ppm increase in atmospheric CO2 concentration experienced over the last century.  And in a similar study, Beerling and Kelly (1997) compared the stomatal densities of fully sixty tree, shrub and herb species from temperate woodlands in the United Kingdom with similar data recorded by Salisbury in 1927, finding that the increase in the air's CO2 content of the past 70 years caused significant decreases in the stomatal densities of this vast array of plants.

The bottom line
Clearly, the issue is far from settled.

References
Beerling, D.J. and Kelly, C.K.  1997.  Stomatal density responses of temperate woodland plants over the past seven decades of CO2 increase: A comparison of Salisbury (1927) with contemporary data.  American Journal of Botany 84: 1572-1583.

DeLucia, E.H., Hamilton, J.G., Naidu, S.L., Thomas, R.B., Andrews, J.A., Finzi, A., Lavine, M., Matamala, R., Mohan, J.E., Hendrey, G.R. and Schlesinger, W.H.  1999.  Net primary production of a forest ecosystem with experimental CO2 enrichment.  Science 284: 1177-1179.

Gill, R.A., Polley, H.W., Johnson, H.B., Anderson, L.J., Maherali, H. and Jackson, R.B.  2002.  Nonlinear grassland responses to past and future atmospheric CO2.  Nature 417: 279-282.

Kouwenberg, L.L.R., McElwain, J.C., Kurschner, W.M., Wagner, F., Beerling, D.J., Mayle, F.E. and Visscher, H.  2003.  Stomatal frequency adjustment of four conifer species to historical changes in atmospheric CO2.  American Journal of Botany 90: 610-619.

Lin, J., Jach, M.E. and Ceulemans, R.  2001.  Stomatal density and needle anatomy of Scots pine (Pinus sylvestris) are affected by elevated CO2.  New Phytologist 150: 665-674.

Norby, R.J., Todd, D.E., Fults, J. and Johnson, D.W.  2001.  Allometric determination of tree growth in a CO2-enriched sweetgum stand.  New Phytologist 150: 477-487.

Paoletti, E., Nourrisson, G., Garrec, J.P. and Raschi, A.  1998.  Modifications of the leaf surface structures of Quercus ilex L. in open, naturally CO2-enriched environments.  Plant, Cell and Environment 21: 1071-1075.

Retallack, G.J.  2001.  A 300-million-year record of atmospheric carbon dioxide from fossil plant cuticles.  Nature 411: 287-290.

Royer, D.L.  2001.  Stomatal density and stomatal index as indicators of paleoatmospheric CO2 concentration.  Reviews of Paleobotany and Palynology 114: 1-28.

Smith, S.D., Huxman, T.E., Zitzer, S.E., Charlet, T.N., Housman, D.C., Coleman, J.S., Fenstermaker, L.K., Seeman, J.R. and Nowak, R.S.  2000.  Elevated CO2 increases productivity and invasive species success in an arid ecosystem.  Nature 408: 79-82.

Woodward, F.I.  1987.  Stomatal numbers are sensitive to increases in CO2 from preindustrial levels.  Nature 327: 617-618.