In appreciation of the technical and technological developments during the last decades, the international commerce of cotton is gradually moving from the manual and visual classification to classification based on results from instrumental testing. Now, going from one technique to the next requires studying their respective modalities and possibilities of application. Thus, commercial practices based on manual and visual classing came out onto the use of an arbitration process and onto the of arbitral tolerances allowing the settlement of possible litigations between sellers and buyers. It is then necessary to do the same for instrumental classing data, in particular in the cotton production conditions in Africa. The variability study of the fiber technological characteristics is a crucial step forward in the definition of the conditions of good realization of instrumental testing in order to limit the litigation risk between cotton producers in Africa and their international customers. It is also required to be more specific and to adapt the actual arbitral procedures to the instrumental classing. To perform these within-bale and between bales variability studies, Dr. Everina LUKONGE, post-doc employee, in Eastern and South-Eastern Africa, and Modeste ABOE, at that time a PhD student, in West and Central Africa were involved. These two persons were taking hundreds of samples in several ginning mills per country in fourteen African countries, while the same ginning mills were continuing to sample bales on the long-run during two crop seasons. All collected samples were tested in the laboratories of the respective Regional Technical Centers (RTCs) in Tanzania and in Mali. Thousands of results of samples tests from SITC were and are statistically looked at. The objective is to define operating methods for bale sampling and for testing collected cotton samples in order to warrant gained results and thus to limit the litigation risk between seller and buyer of the prod