As the idea that G-quadruplex nucleic acid structures are involved in cellular processes is gaining support, it becomes important to develop ligands that specifically target G-quadruplexes. However, ligand design is complicated because there are multiple G-quadruplex target sequences, some sequences are polymorphic, and very few ligand-quadruplex structures in solution were solved to date. Further, structure alone does not reveal the driving forces for ligand binding. To know why a ligand binds, the thermodynamics of binding must be characterized. Electrospray mass spectrometry makes it possible to detect and quantify each specific stoichiometry in terms of number of strands, number of specific cations, and number of ligands, and thus allows one to simultaneously determine the equilibrium constants for the formation of each complex. We designed and built a temperature-controlled nano-electrospray source to monitor thermal denaturation by mass spectrometry ("MS-melting"). We studied the thermal denaturation of G-quadruplexes, including the c-myc promoter and several telomeric sequence variants, and their complexes with popular ligands (Phen-DC3, TrisQ, TMPyP4, Cu-ttpy). From the temperature dependence of the equilibrium constants, we determined the enthalpic and entropic contributions to the formation of each stoichiometric state. In absence of ligand, we untangled the potassium-induced G-quadruplex folding thermodynamics, one potassium ion at a time. The formation of each quartet-K + -quartet units is strongly enthalpy driven, with entropy penalty. In contrast, the formation of quartet-K + -triplet units is entropically driven. For this reason, such misfolded structures can become more abundant as the temperature increases.In the presence of ligands, mass spectrometry also revealed new states at intermediate temperatures. For example, even in cases where only a 1:1 (ligand:quadruplex) is observed at room temperature, a 2:1 complex predominates at intermediate temperatures. Mass spectrometry also makes it easy to distinguish ligand bound to the 2-quartet structures (containing 1 K + ), the 3-quartet structures (containing 2 K + ) and to the unfolded strand (no specific K + ). We confirm that TrisQ binds preferably, but not exclusively, to 3quartet structures, Phen-DC3 binds to a 2-quartet structure, while the porphyrin ligand TMPyP4 is characterized as non-selective, because it binds to all forms including the unfolded one. The thermodynamics of ligand binding to each form, one ligand at a time, provides unprecedented detail on the interplay between ligand binding and changes in G-quadruplex topology.