The word reconnaissance often means exploration in exploration’s narrative. At a hinge time, the 1900s, the paper aims to show that, instead of reducing the meaning of reconnaissance to its explorative side, it could be fruitful to consider this term’s polysemy and to question the link between reconnnaître and connaître. The corpus studied here is the french geographical review, the Annales de Géographie, where exploration’s narrative is well present even if it’s not a specialized review. Our hypothesis is that the reconnaissance’s moment should be taken very seriously in the shaping of geographical knowledge: the reconnaissance is a time for testing hypothesis and confronting conjectures to geographical reality. The first part of the paper develops a general analysis, while the second one offers a case study on Antarctica, the last big white patch on the globe around 1900.