A necessidade de se averiguar ocorrência de dormência em sementes de acácia amarela é importante para produção de mudas sadias e com desenvolvimento desejável, principalmente para recomposição florestal. O experimento foi realizado no IFES, com sementes extraídas manualmente de vagens recém-colhidas de plantas localizadas no instituto, as mesmas foram imersas em água destilada, solução de giberelina 0,3 % (v/v) e solução de cloreto de potássio 5 % (v/v), por 30 minutos, água quente (90 ºC) e água com gelo (entre 0 ºC e 2 ºC) por 30 minutos. Foi avaliado a porcentagem de germinação; índice de velocidade de germinação; tempo médio de germinação; número de folhas; altura da parte aérea; comprimento da raiz; massa fresca e massa seca das plântulas. The treatments with gibberellin and potassium chloride were superior in all the variables related to germination and green mass production, evidencing the importance of the pre germinating treatment of the seed. The treatment with ice, although not improving the germination, presented superior results to the control in the characteristics related to root growth and height of the seedling. Treatment with gibberellin or KCl, and ice treatment may be recommended for acacia seeds.

The large amount of degraded areas and productive potential of the legal reserves in Brazil make restoration an environmental demand and a commercial opportunity. We modelled the diameter growth as a function of age of eight tree species in restoration plantations in the Brazilian Amazon. From 14 years of annual forest inventory data, for each species, we tested variations of logistic function: simple logistic, logistic with covariant (plant area at the time of planting), logistic with random effect, logistic with random effect and covariant. Amongst the studied species, Schizolobium parahyba var. amazonicum, Tectona grandis and Simarouba amara showed the highest growth rates while Cordia alliodora, Cedrela odorata and three species of the genus Handroanthus showed slower growth. The gains from using the covariant in modeling were small for both fixed and mixed-effect models. Gains from the inclusion of the random effect were substantial. Mixed-effect models had the best performance in modeling the growth of the species. Our results provide basis for a critical view of the criteria and possibilities for degraded areas restoration and management practices in legal reserves of the Amazon. An economic analysis is required to ensure the viability of these areas’ sustainable exploitation.