Strategies in mental arithmetic have been studied by cognitive psychologists for many years and this has led to an impressive amount of studies. Yet and quite surprisingly, the conclusions reached in the literature are not always coherent and it is difficult to draw a clear picture of what happens exactly when adults and children solve simple problems. For addition, it was commonly assumed that the transition from counting procedures to retrieval took place around the age of 10. At this age, children were indeed considered as able to solve most simple addition problems without counting. Addition facts were supposed to have already been constructed in long-term memory and easily activated in arithmetic problem solving task (Ashcraft & Fierman, 1982). However, we have recently challenged the strong consensus that very simple additions (involving operands from 1 to 4) are solved by adults through retrieval from memory (Barrouillet & Thevenot, 2013; Fayol & Thevenot, 2012). If adults still use procedural strategies to solve such simple problems, the view that children use retrieval for those problems needs to be reconsidered. This is the aim of the present project. In a longitudinal approach, we would like to test children aged from 8 to 9 (before the transition from counting to retrieval is supposed to occur) to 11-12 (when the transition is supposed to have occurred) on their ability to solve one-digit additions. As in our previous research with adults (Barrouillet & Thevenot, 2013), children will simply be asked to solve the additions and to give their answers out loud as quickly as possible. Children’s answers as well as solution times will be recorded by the computer. Five classrooms of third graders (around 125 children) will be tested. The same children will be retested one year and two years later. Moreover and in order to further investigate individual differences, children’s working memory capacities and arithmetic skills will be measured and examined in relation to their arithmetic strategies.