Interactive effects of CO2 and water availability have been predicted to al
ter the competitive relationships between C3 and C4 species over geological
and contemporary time scales. We tested the effects of drought and CO2 par
tial pressures (pCO(2)) ranging from values of the Pleistocene to those pre
dicted for the future on the physiology and growth of model C3 and C4 speci
es. We grew co-occurring Abutilon theophrasti (C3) and Amaranthus retroflex
us (C4) in monoculture at 18 (Pleistocene), 27 (preindustrial), 35 (current
), and 70 (future) Pa CO2 under conditions of high light and nutrient avail
ability. After 27 days of growth, water was withheld from randomly chosen p
lants of each species until visible wilting occurred. Under well-watered co
nditions, low pCO(2) that occurred during the Pleistocene was highly limiti
ng to C3 photosynthesis and growth, and C3 plants showed increased photosyn
thesis and growth with increasing pCO(2) between the Pleistocene and future
CO2 values. Well-watered C4 plants exhibited increased photosynthesis in r
esponse to increasing pCO(2), but total mass and leaf area were unaffected
by pCO(2). In response to drought, C3 plants dropped a large amount of leaf
area and maintained relatively high leaf water potential in remaining leav
es, whereas C4 plants retained greater leaf area, but at a lower leaf water
potential. Furthermore, drought-treated C3 plants grown at 18 Pa CO2 retai
ned relatively greater leaf area than C3 plants grown at higher pCO(2) and
exhibited a delay in the reduction of stomatal conductance that may have oc
curred in response to severe carbon limitations. The C4 plants grown at 70
Pa CO2 showed lower relative reductions in net photosynthesis by the end of
the drought compared to plants at lower pCO(2), indicating that CO2 enrich
ment may alleviate drought effects in C4 plants. At the Pleistocene pCO(2),
C3 and C4 plants showed similar relative recovery from drought for leaf ar
ea and biomass production, whereas C4 plants showed higher recovery than C3
plants at current and elevated pCO(2). Based on these model systems, we co
nclude that C3 species may not have been at a disadvantage relative to C4 s
pecies in response to low CO2 and severe drought during the Pleistocene. Fu
rthermore, C4 species may have an advantage over C3 species in response to
increasing atmospheric CO2 and more frequent and severe droughts.