Abstract

In order to identify promising wild argan trees (Argania spinisa L. Skeels), with high oil content and facilitated of crushing seeds, as a part of domestication and breeding programs, seventy five candidate plus trees were chosen from different eco-geographical regions in the southwest of Morocco. Based on several biometric characters describing trees and their fruits, uni and multivariate analysis of eighteen quantitative traits were done. Results showed significant differences (P<0.01, GLM) between trees of the same provenances and among it for all morphological fruit traits, except the tree traits. Considerable variability was found in oil production ranged from 39.19% to 57.92%. Thus, Ao, Hd, La and Bi provenances exhibited high performance yield and appeared to be the best adapted to drought conditions, contrary to Al provenance. Based on the variance components method, high broad-sense heritability was recorded for oil content (93.28%), indicating the additive gene action. Correlation analysis revealed that fruit weight, seed weight, almond weight, seed length, seed width and AW-90S are highly and positively correlated with oil content. Hence, almond weight and the number of almond per seed are positively correlated to AW/SW ratio. However, there was no correlation between crushing seed trait and oil content. In addition, promising trees which have a clear superiority relating to the oil production and the facility of crushing seed, were identified and made a first pre-selected for the oil yield improvement for further studies.

Key words: Argania spinosa (L.) Skeels, provenance, oil content, crushing seed, pre-selection.