The main aim of this paper was to examine whether the characteristics of the Lotka distribution of publications (in particular, the changes that the two parameters, n and c, undergo) constitute an indicator of the structure of influence in a scientific field. A quasi-experimental method was used to estimate the parameters of Lotka's law in a number of scientific areas (by means of a series of searches in the Scopus database). The study was performed on 90 sets of author productivity data (resulting from a combination of 10 areas, 14 countries, and 3 time periods). Both the exponent of the law, n (i.e., the slope of the log-log plot), and the constant c (the fraction of authors with only a single publication) were found to depend on the state of development of the scientific area, on its productivity, on the country, and on the time period being studied. A characteristic that distinguished the so-called "hard sciences" from the "social sciences and humanities" was the level of co-authorship, with the average number of authors per publication being greater in science than in the social sciences and humanities. The empirical results show a picture of the behaviour of the Lotka distribution in different situations, due to different causes. This could be interesting as a better understanding of these regularities may allow them to be incorporated into the theoretical context.