The review is devoted to an analysis of the influence of typical factors present in a real experiment (impurities, phonons, domain structure, etc.) on the behavior of thermodynamic quantities near a second-order phase transition point. Chapter II discusses equilibrium thermodynamic perturbations. Perturbations that differ in nature are considered from a unified point of view in which the idea of isomorphism is used. In the first stage, in each concrete case, thermodynamic variables called ``isomorphic'' are introduced, in which the ordinary second-order phase transition takes place. It is frequently difficult or impossible, however, to carry out an experiment in which the ``isomorphic'' variables remain constant. The second stage therefore consists of a transition from ``isomorphic'' variables to ``experimental'' variables, which can be conveniently maintained constant in the experiment. In terms of the "experimental" variables, the picture of the phase transition can differ strongly from the usual picture of a second-order phase transition. Thus, in the case of impurities, if the measurements are carried out at constant concentration, a third-order phase transition takes place, and if the ordering parameter interacts with longitudinal or transverse phonons (at constant pressure), the phase transition is of first order. Chapter III discusses the influence of nonequilibrium perturbations (mainly immobile impurities and inhomogeneities) on the thermodynamics of the phase transitions. In this case there may exist several (at least two) points that are nonanalytic in the temperature and have weak singularities of the thermodynamic quantities. A temperature interval of the ``smearing'' of the transition is indicated for each type of perturbation. The theory is compared with experiment.