Within the Tsallis thermodynamics framework, and using scaling properties of the entropy, we derive a generalization of the Gibbs-Duhem equation. The analysis suggests a transformation of variables that allows standard thermodynamics to be recovered. Moreover, we also generalize Einstein's formula for the probability of a fluctuation to occur by means of the maximum statistical entropy method. The use of the proposed transformation of variables also shows that fluctuations within Tsallis statistics can be mapped to those of standard statistical mechanics. DOI: 10.1103/PhysRevLett.88.020601 PACS numbers: 05.70.Ln, 05.20.Gg, 05.40. -a During the past few years there has been a great deal of interest in studying nonextensive thermodynamics [1][2][3]. This results from the assumption of nonadditive statistical entropies and the maximum statistical entropy principle, following the information theory formulation of statistical mechanics proposed by Jaynes [4]. Indeed, besides its relevance in many nonequilibrium problems, nonextensivity is of interest for systems of particles which show longrange interactions [5], as occurs in a number of ferroic materials [6] such as ferromagnetic, ferroelastic, ferroelectric solids, and astrophysical systems [7]. Within this framework, the Tsallis statistical entropy [8] has proven to be the only nonadditive generalization of Gibbs-Shannon entropy which satisfies the following properties: (i) positivity (it takes zero value for absolute certainty), increasing monotonously with increasing uncertainty, and (ii) concavity. However, many fundamental features regarding the connection between the formulation of statistical mechanics and thermodynamics remain unclear. For instance, the identification of adequate generalized thermodynamic forces and the computation of statistical fluctuations are still controversial [9]. In this Letter we clarify such problems and provide robust arguments showing the equivalence of the present formulations of nonextensive Tsallis thermodynamics with the standard extensive equilibrium formulation.Within the Tsallis formalism, the lack of information associated with any probability distribution ͕p͑i͖͒ defined on a set of microstates V ͕i͖ [10] is given bywhere the parameter q, determining the degree of nonextensivity, is positive in order to ensure the concavity of S . The q-logarithmic function is defined as ln q f ͑ f 12q 2 1͒͑͞1 2 q͒. In the q ! 1 limit, Eq. (1) reduces to the Gibbs-Shannon entropy S 2 P i p͑i͒ lnp͑i͒. In order to simplify the notation, we will indicate the set V only when necessary. The novelty of the statistical entropy (1) is that it does not satisfy additivity. Instead, for two systems A and B described by independent probability distributions,For a given physical system the equilibrium probability distribution ͕p ء ͑i͖͒ corresponds to the distribution that maximizes S under the normalization condition1͔ and the relevant constraints imposed by the available statistical information on the system. The thermodynamic equilibrium entr...