In the last five years, much attention was devoted to the signless Laplacian of a graph by the scientific community. One of the main reasons for this interest is the intuition, shared by many researchers on the basis of studies concerning small graphs, that more graphs are determined by their signless Laplacian spectrum than by those of the adjacency and Laplacian matrices. Results presented in this thesis brought new elements on the informations hidden in the spectrum of this matrix. On the one hand, we present relations between structural graph invariants and an eigenvalue of the signless Laplacian. On the other hand, we present families of extremal graphs for two of its eigenvalues, with and without additional constraints on the shape of...