The study aims at examining whether and when morphological information is activated when French children read words. Although previous studies have indicated that children do use morphological structure, there is a need for clarity regarding when this information is activated in the word reading process. The priming paradigm, contrasting various prime durations associated with lexical decision, allows tracking activation of information. The aim of the present study is to testify whether and when morphological priming can be evidenced in child readers and in what extend it differs from orthographic priming. In the expert literature, the priming paradigm has been extensively used, especially with low frequency target words. In the child reading field, less is known about frequency effects. The second aim of the study is to compare time course of activation of both morphological and orthographic priming according to target frequency. Participants were 75 third- and 75 fifth-graders. They performed a lexical decision task with visual priming, with three prime durations: 55 ms, 80 ms, 250 ms. Targets were 45 base words. Targets were base words. Three kinds of primes were paired with each target item: morphological (collage-COLLE 'gluing-glue'), orthographic (college-COLLE 'college-glue'), control (seringue-COLLE, 'syringe-glue'). Three versions of the experiment were constructed, with items counterbalanced across conditions. Facilitation effects were observed in the morphological priming condition only, in the 80 and 250 ms prime duration conditions, and in both grade levels, without any effect of orthographic control priming. Morphological priming at 55 ms was observed only in a more advanced fifth graders subgroup. When considering separately priming for high vs low frequency targets, some significant differences raised, with a stronger quantity of priming for low frequency target words and a later morphological activation in high frequency target words.