Growing evidence indicates that the tendency to synchronize community dynamics may be a general phenomenon in nature related to metaecosystem stability. Synchronized dynamics reduce stability since local variations in biomass or richness are directly propagated to variations in metacommunity dynamics. Asynchronous dynamics have a strong stabilizing effect because of the cancelling out of local variations. Synchronization of biodiversity dynamics may occur due to dispersal among communities and similar responses of different communities to correlated environmental variations---the generalized Moran effect. This congruent response of different communities to environmental dynamics depends on their similar functional composition, which is determined by the similarity in local conditions and the spatial distance between them. In a metacommunity of 51 temporary ponds that were surveyed for 14 years, we evaluated the existence of synchronized dynamics in plant richness and biomass among communities, and their association with temporal stability. A path analysis based on Mantel tests was used to evaluate the hypothesized determinants of synchrony and stability. Community-community matrices for Mantel tests were constructed with synchrony in biomass and richness, and with the Euclidean distance in geographic location, environmental conditions, and functional composition. Elements of the stability matrices were the inverse of variation in richness and biomass for each pair of ponds. A wide range of dynamics from asynchronous to synchronous rhythms was observed, and stability decreased with the coupling of community dynamics. Geographic, environmental, and functional distances between pairs of communities decoupled their dynamics in richness but weakly affected biomass synchrony. The role of rare species in richness dynamics and of dominant species in biomass dynamics may explain the observed discrepancies. While evidence for biodiversity synchrony is accumulating, evaluating the role of the proposed determinants becomes essential. Our results support the main hypotheses but also identify discrepancies that may devote further attention.