Nitrogen fixation estimated by the 15N natural abundance method in Acacia mangium Willd. inoculated with Bradyrhizobium sp and grown in silvicultural conditions

A field experiment was established at Port-Bouet in Cote d'Ivoire (West Africa) on quaternary coastal sandy soils in order to test the effects of rhizobial inoculation and soil fertility on the growth of Acacia mangium and the percentage of atmospheric N-2 fixed by this species. Three inoculation treatments, all replicated in three blocks having different fertility levels, were tested within a complete randomized block design. Twenty-one months after transplanting to the field, we observed a positive effect of the inoculation with the Bradyrhizobium strain Aust13c on tree growth in all blocks with an overall significant increase of 10% in height and 15% in diameter compared to trees inoculated with the Bradyrhizobium strain CB756 or uninoculated control trees, the two latter treatments being not significantly different. The gradient of fertility determined by a soil analysis and observed from block to block had a significant effect on tree growth as well. Concomitantly, the total nitrogen content (%N) measured in sampled A. mangium leaves did not vary according to the block or inoculation treatment. In contrast, the %N contained in leaves of inter-planted Eucalyptus urophylla, which were used as non-N-2-fixing reference trees for the measurement of nitrogen fixation, decreased gradually from block I to III and proportionally to a decline in the different forms of soil N and other elements. The percentage of nitrogen derived from atmospheric N-2 (%Ndfa), estimated by comparing the natural N-15 abundance (delta(15)N) in A. mangium leaves with that of inter-planted E. urophylla, varied greatly and ranged from 19 to 91%, according to the block and inoculation treatments. The %Ndfa calculated in the inoculation treatments Aust13c, CB756 and uninoculated controls reached 62, 55, and 41% respectively, all blocks combined. On the other hand, combining all inoculation treatments, %Ndfa was found to reach a mean of about 65% in each of the two least fertile blocks (blocks II and III) while with 27%, it was 2.4 times lower in trees located in the most fertile block (block I). This study shows that the N-2-fixing potential of A. mangium, which is subject to large variations within a relatively restricted area, is mostly due to the gradient of fertility observed. The quantification of total N fixed per tree remains to be determined in further in situ studies to explain the positive effect of inoculation with Aust13c strain on A. mangium growth. (C) 2002 Elsevier Science Ltd. All rights reserved.