344 Ecology, 82(2), 2001, pp. 344–359  2001 by the Ecological Society of America INTEGRATION OF DEMOGRAPHIC ANNUAL VARIABILITY IN A CLONAL DESERT CACTUS MARı ´A C. MANDUJANO, 1,4 CARLOS MONTAN ˜ A, 2 MIGUEL FRANCO, 1 JORDAN GOLUBOV, 1 AND ARTURO FLORES-MARTı ´NEZ 3 1 Instituto de Ecologı ´a, Universidad Nacional Auto ´noma de Me ´xico, Apartado Postal 70-275, 04510 Coyoaca ´n, D.F., Me ´xico 2 Instituto de Ecologı ´a, Apartado Postal 63, 91000 Xalapa, Veracruz, Me ´xico 3 Departamento de Bota ´nica, Escuela Nacional de Ciencias Biolo ´gicas, Instituto Polite ´cnico Nacional, Carpio y Plan de Ayala, 11340 D.F., Me ´xico Abstract. Spatial and temporal variability influence the structure and dynamics of perennial plant populations. In order to investigate the consequences of this environmental heterogeneity on population and life history traits of a perennial plant with a complex life history, four size-classified population matrix models were employed (i.e., annual, mean, periodic, and stochastic simulations) in a clonal cactus, Opuntia rastrera, in the Chihuahuan Desert, Mexico. This species was studied over a seven-year period in two contrasting, neighboring habitats (nopalera and grassland). The specific aims were: (a) to assess the effect of annual environmental variability on the long-term dynamics of both populations; and (b) to estimate the relative contribution of sexual reproduction and clonal propagation to the populations’ rates of increase. Projections from each model provided complementary information on population dynamics and life history in the two habitats. The finite rate of population growth () varied among matrix projections. Mean, periodic, and stochastic projections yielded  1 for nopalera and   1 for grassland. The relative contributions of size classes and demographic processes to changes in  differed widely between years and habitats for annual matrices. In contrast, elasticities of periodic matrices showed a stable habitat-dependent pattern. The proportional change in  produced by sexual recruit- ment and clonal propagation showed wide spatial variation in which the most distinctive difference between habitats was the predominance of clonal recruitment at the nopalera and seedling recruitment at the grassland. Elasticities also showed temporal variation where- by clonal propagation decreased as precipitation increased, while sexual reproduction and growth tended to increase with precipitation. The striking spatial and temporal differences found in the structure, dynamics, and life history traits of O. rastrera are aptly summarized by the periodic analyses. These differences reflect both the varying influence of the selective pressures operating on this species and the ways in which demographic plasticity deals with them. Key words: Cactaceae; Chihuahuan Desert, Mexico; clonal propagation; demographic param- eters; demographic plasticity; elasticity; life history; Opuntia; periodic matrix; prickly pear; sexual reproduction; stochastic environments. INTRODUCTION The importance of having different types of recruits (sexual vs. clonal) in plants with complex life cycles inhabiting heterogeneous environments is not yet clear. The dynamics of these populations can be adequately described using matrix models (Caswell 1985, 1989, van Groenendael and Slim 1988, van Groenendael et al. 1988, Cochran and Ellner 1992, Eriksson 1994, Franco and Silvertown 1996). Specifically, these mod- els could be employed with plants whose clonal off- spring develop independently and propagate the pa- rental genes in time and space and to analyze the dy- namics of organisms in which demographic processes depend more on size and/or developmental stage of the 4 E-mail: mcmandu@miranda.ecologia.unam.mx organisms than on their age (Lefkovitch 1965, Caswell 1985, Cochran and Ellner 1992, Caswell et al. 1997). They are also useful when the age of each individual is unknown or difficult to assess or when populations have different types of recruits (e.g., sexual and clonal newborns of different sizes) (Bazzaz and Harper 1977, Bierzychudek 1982, Huenneke and Marks 1987, Coch- ran and Ellner 1992, Nault and Gagnon 1993, Eriksson 1994, Damman and Cain 1998). The recruitment of individuals into populations with complex life cycles may occur by either sexual repro- duction or clonal propagation. While sexual reproduc- tion promotes genetic diversity, clonal propagation maximizes the fitness of established genets (Williams 1975, Cook 1979, Willson 1979). Shifts in resource allocation to each kind of recruit will have strong ef- fects on the population dynamics of the species and on its life history (Cook 1979, Caswell 1985, Nault and