The essentiel oil composition of 16 wood samples from Erromango and Aniwa has shown important differences between a group where the individual variability is high and 2 individual samples. Considering that alpha-santalol and beta-santalol are the most important molecules in the Indian sandal oil, this group presents a surprisingly low content of these components (20%-30%) whereas one of the individual samples has a content of 70%, similar to the New Caledonian S. #austrocaledonicum# and the standard #S. album# sandal oil. There are two significantly distinct chemotypes. This difference leads to several questions to which the present sampling can only reply partially : where does this difference come from ? (particularities of the stands, environmental characteristics, characteristics of the tree...) If the chemotypes cannot be linked to the stand, is the 1/15 ratio significative of the population ? The answers to these two questions are of prime importance for future germplasm conservation and strategies of tree propagation for a planting policy. The selection of germplasm, the sandal oil content of which is not similar to the standard one presently known on the market, would lead to lower prices for the wood although the best pieces will always find buyers at a good price. A large and systematic survey of all the stands on all the islands is therefore recommended in order to identify the real oil variation determinism.