Species and provenance trial of Prosopis
Dry weight could only be assessed for P. juliflora and P. pallida, but the total basal areas indicate that these species were also the most productive species.
Compared to other trials at Bebedouro the best provenances in this trial had a relatively large production of biomass (Rębild et al. 2001a, c, d). If we ignore Punjab9b of P. juliflora, which could be biased due to the mix with another species, the provenance Peru13 of P. pallida had the largest growth with almost 2.1 t dry weight ha-1 y-1. In two other Prosopis trials at Bebedouro (Trial nos. 3 and 4 in this series), the best provenances of P. juliflora attained dry weight productions of 1.6 and 1.1 t ha-1 y-1. In a trial of Acacia species (Trial no. 1), the best provenance of A. nilotica had a production of 2.0 t ha-1 y-1.
At the same time, height growth was impressive, being approximately 1 m per year for Peru13 - larger than in any of the trials mentioned above.
It appears that part of the reason for the large production in the current trial is that there are favourable micro-site conditions compared to the other trials mentioned. Brazil2 is present in all the Prosopis trials, and the growth is clearly larger in this trial than in the others.
Differences between species
Interpretation of species differences based on trials with few provenances of each species may be dangerous unless there is certainty that the provenances represent the best provenances for the site. It is in this light that the considerations below should be seen, and there is reason to be cautious with the conclusions.
In many variables the differences between species were at the border of significance. For the variables height, crown area, basal area of the mean tree and total basal area there were signs that P. juliflora and P. pallida were the best. The multivariate analysis confirmed that these species were separated from the other species, even though the provenance Punjab9b was different from the other provenances of the two species.
The potential of the species P. alba may warrant more investigation, as one of the provenances had basal areas that were if not as high then at least comparable with the basal areas in the two best species.
Differences between provenances
In the univariate analyses there were only few signs of differences between provenances. For P. alba, the analyses indicated that there were provenance differences in crown area, basal area of the mean tree and total basal area. The provenance Argentina3 was generally more productive compared to the provenance Argentina2, which is of the variety panta. In P. chilensis there were signs of differences in total basal area.
However, the multivariate analysis showed a somewhat different picture. First, the two provenances of P. juliflora, Punjab9b and Brazil2, were far from each other although they were hard to separate in the univariate tests. Second, Argentina4 and Chile05, the provenances of P. chilensis, could not be separated in the multivariate test. Third, the data indicated that there were differences between the provenances Chile09 and Argentina5 of P. flexuosa, a fact that was not detected in the univariate analyses. Even though one should be careful not reading too much into the multivariate test because it does not take variance heterogeneity into account, at least the first two conclusions mentioned here seem rather robust.
Finally the multivariate test confirmed that the two provenances of P. alba were separated, thus acknowledging the fact that there are differences between the two provenances. This is interesting, not only because they represent different varieties, but also because they were collected at the same site.
Giving provenance recommendations it appears that the best provenances are found in P. juliflora and P. pallida. Even though the average growth of the provenance Punjab9b of P. juliflora appeared to be larger than the other provenances, there is so much uncertainty linked to this provenance that it should be avoided before further testing has taken place. In the trial number 3, Punjab9b had a much inferior growth than the local landrace, Brazil2.
Thus, for purposes of wood production, it appears that there are three, possibly four, provenances to be recommended: Brazil2 (P. juliflora), Peru05 and Peru13 (both P. pallida). The provenance Argentina3 of P. alba may also be considered, even though not as fast growing as the provenances just mentioned.