Population samples of an African C4 grass, Panicum coloratum L., were collected from two locations in the Serengeti Grasslands varying in grazing intensity, one a high—grazing location (GA = grazing—adapted), the other a low—grazing location (NGA = nongrazing—adapted). Plants were cloned, put in controlled environments mimicking the natural photo—thermoperiod, and subjected to light grazing pressure by a generalist feeding North American grasshopper, Melanoplus sanguinipes. Carbon assimilation and redistribution were measured in the short term with an infrared gas analyzer and 11C—labelled CO2, coupled with a three—compartment analytical model, and by harvesting whole plants at the end of a 12—wk regrowth experiment. Results documented several significant differences between the GA and NGA samples, suggesting the evolution of physiological traits related to C assimilation, translocation, and storage in response to previous grazing history. Pregrazing net C—fixation rates, translocation rates, and C—storage pools were identical for the two ecotypes. After grazing, overall C—fixation rates were 39% higher for the GA plants, and the regrowth data suggest they remained higher than NGA rates throughout the experiment. Removal of < 10% of initial green—leaf biomass by grazing at each grazing period produced major differences in carbon flux between the two samples. Throughout the experiment GA plants produced and stored more C in leaves, stored less C in stem sinks, had higher phloem activity, and translocated more of the labile C to roots where it was stored in higher quantities. This suggests that storage of labile C reserves in sinks or pools readily accessible to the plant, which allows rapid mobilization after grazing, is an important element of adaptation to grazing. Major storage of labile reserves in stems, characteristic of the NGA plants, may be advantageous in ungrazed habitats where there is vertical growth due to canopy closure and competition for light, but such storage makes those reserves accessible to grazers. GA ecotype plants in the regrowth experiment compensated completely for the eight weekly defoliation events by the time of the 12—wk harvest; yield of the NGA ecotype was reduced 21% by the moderate level of grasshopper grazing. Increased total yield (biomass harvested plus compensation) by the GA ecotype was expressed in both above— and belowground biomass, suggesting that suppression of the latter does not contribute to compensation by the former.
