Year and tree effect on reproductive organisation of Cedrus atlantica in a natural forest Fazia Krouchi a,* , Arezki Derridj a , Franc ¸ois Lefe `vre b,1 a De ´partement d’Agronomie, Faculte ´ des Sciences Agronomiques et Biologiques, Universite ´ M. Mammeri, BP 17, Tizi-Ouzou, RP 15000, Algeria b Unite ´ de Recherches Forestie `res Me ´diterrane ´ennes, INRA, Avenue A. Vivaldi, 84000, Avignon, France Abstract Male and female fecundity parameters (flowering intensity, pollen germination, seed yield and seed quality) were measured on 161 trees in two populations of a Cedrus atlantica forest in Algeria across 7 years. We found a huge variation among trees for all variables, e.g., the number of cones per tree ranged from 0 to 770, the number of male branches from 0 to 59, and the number of male strobili per flowering branch from 15 to 398. The two populations only differed for female fecundity and there was no year  population interaction. The variation among years was highly significant for both male and female fecundity parameters. The ranking of the trees was globally maintained over years. However, fertility variation over the whole study period was lower than the variation observed in any single year, due to temporal stochasticity and cumulative effect. As usually found in this monoecious species, we observed a diversity of sexual types, from ‘‘male’’ to ‘‘female’’ individuals with a trade-off between the two functions. This polymorphism was stable over the time scale of the study. Female fecundity was correlated to tree dimension, but the correlation for male fecundity was weak. The consequences of this pattern of variation for the dynamics of genetic diversity are discussed. # 2004 Elsevier B.V. All rights reserved. Keywords: Cedrus atlantica; Fecundity; Sex; Demography; Masting 1. Introduction Within population variation in fecundity parameters affects the genetic diversity by reducing the effective population size i.e., increased co-ancestry and inbreeding, and loss of genetic diversity (Nunney, 2000). This effect is enhanced when fecundity dis- tribution is skewed (Wilson and Levin, 1986), and when both male and female contributions vary (Kang et al., 2001). Male and female fecundity variation in forest trees has been observed both in natural stands (Linhart et al., 1981; Linhart and Mitton, 1985) and in seed orchards (Chaisurisri and El-Kassaby, 1993; El- Kassaby, 1995; Yazdani et al., 1995; Matziris, 1997; Bila et al., 1999). This variation is attributed to the genotype of individuals (Savolainen et al., 1993; El- Kassaby and Cook, 1994; Sorenson and Cress, 1994; Burczyk and Chalupka, 1997; Bila and Lindgren, 1998; Kang and Lindgren, 1998; Nikkanen and Ruot- salainen, 2000), to environmental factors (Freeman et al., 1981; Bila and Lindgren, 1998; Fogal et al., 1999) and to silvicultural practices (Fries, 1994; Bila and Lindgren, 1998). Forest Ecology and Management 197 (2004) 181–189 * Corresponding author. Tel.: þ213-26-21-15-27; fax: þ213-26-21-15-27. E-mail address: krouchi@yahoo.com (F. Krouchi). 1 Tel.: þ33-4-90-13-59-20; fax: þ33-4-90-13-59-59. 0378-1127/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2004.05.013