Although Brazil nut (B. excelsa) is often touted as one of the most economically successful NTFPs little is known about the population structure of this species within its natural range in Southwestern Amazonia or ecological factors that affect fruit production. Since these are considered fundamental for sustainable resource management we examined a natural Brazil nut stand in an extractive reserve in Acre Brazil posing the following questions: (1) What is the spatial distribution species density and size-class structure of B. excelsa? and (2) What tree-level factors influence Brazil nut production? In a 420 ha census 568 trees [greater-than-or-equal]10 cm diameter at breast height (dbh) were counted resulting in a density of 1.35 trees ha-1. Based on the nearest-neighbor method an index of aggregation (R) of 0.77 indicated a rejection of the null hypothesis of a strictly random distribution pattern. Yet this value suggests a much greater tendency toward randomness than either clumping or uniformity. Our data do not show the commonly reported existence of groves referring to clearly defined clusters of 50 to several hundred trees separated from similar clusters by great distances. Almost 1/4 of the population (23%) was composed of non-reproductive juveniles. Maximum R2 improvement analysis applied to four distinct diameter classes provided insight into the dynamics of production-related variables over the species life cycle. While dbh explained 1/3 of production variance (R2 = 0.3360) in the smallest diameter class (10 cm = dbh < 50 cm) which included those in the process of reaching reproductive maturity crown form best explained production variance of very large trees (dbh = 100 cm). Results also demonstrated a significant negative correlation between crown vine load and production of trees = 50 cm dbh (r = -0.13 P = 0.008) suggesting the need for further study on vine cutting as a possible silvicultural treatment for enhancing nut yields.