Extracting Non-timber Forest Products is sustainable only if the harvest does not cause negative population growth. Leaves of Chamaedorea palms are harvested from Mexican forests and exported for floral greenery. We examined the effect of alternative harvest schedules, and livestock browsing, on demography of Chamaedorea radicalis Mart., a dioecious understory palm, in El Cielo Biosphere Reserve (2308'N, 9909'W), Tamaulipas, Mexico. In Jan. 1999 we assigned 100 adult palms to each of 5 treatments: Control (no leaf removal), 1X/year (marketable leaves removed each Aug.), 2X/year (Aug., Feb.), 4X/year (Feb., May, Aug., Nov.), and Modified 4X/year (max. 1 leaf per palm removed, and no palms defoliated). We calculated stage transition probabilities and fecundities of adults over 1 year (Aug. 1999 - Aug. 2000). We parameterized a stage transition matrix for each treatment using these values and transition probabilities for seeds, seedlings, and juveniles obtained from the same site. Eigenanalysis of each matrices yielded l (finite rate of increase) for each treatment. To quantify effects of browsing by free-range livestock, we randomly assigned 2 100m2 plots to each of 3 treatments; No browse, 1X browsed (Feb. 2000) by 1 burro for 15 min., and 2X browsed (Aug. 1999, Feb. 2000). The 4X harvest treatment reduced adult survival and reproduction, and caused some adults to regress to "younger" stages, resulting in l=1.00, while the control was projected to grow with l=1.06. Browsing elevated mortality of most stages. Incoporating this mortality into the control's matrix, l declined to 0.96 for 1X browse and 0.82 for 2X browse. Furthermore, adults that survived defoliation by harvest or browse showed a developmental response; their next leaf was frequently smaller than previous leaves. In Jan. 2001 the youngest fully-expanded leaf on the 4X treatment averaged 34 cm long vs. 45 cm on controls (T=5.99, df=136, P<0.001). Collectors only take leaves > 40 cm long, so defoliated palms have a respite from harvest. Treatments are continuing, allowing assessment of their impact beyond the first year. Present results suggest occasional browsing more negatively effects populations than periodic leaf harvest.

Key words: browse, Chamaedorea radicalis, leaf harvest, Mexico, palm, transition matrix