A tale of four stories: soil ecology, theory, evolution and the publication system

Abstract

Background: Soil ecology has produced a huge corpus of results on relations between soil organisms, ecosystem processes controlled by these organisms and links between belowground and aboveground processes. However, some soil scientists think that soil ecology is short of modelling and evolutionary approaches and has developed too independently from general ecology. We have tested quantitatively these hypotheses through a bibliographic study (about 23000 articles) comparing soil ecology journals, generalist ecology journals, evolutionary ecology journals and theoretical ecology journals.

Findings: We have shown that soil ecology is not well represented in generalist ecology journals and that soil ecologists poorly use modelling and evolutionary approaches. Moreover, the articles published by a typical soil ecology journal (Soil Biology and Biochemistry) are cited by and cite low percentages of articles published in generalist ecology journals, evolutionary ecology journals and theoretical ecology journals.

Conclusion: This confirms our hypotheses and suggests that soil ecology would benefit from an effort towards modelling and evolutionary approaches. This effort should promote the building of a general conceptual framework for soil ecology and bridges between soil ecology and general ecology. We give some historical reasons for the parsimonious use of modelling and evolutionary approaches by soil ecologists. We finally suggest that a publication system that classifies journals according to their Impact Factors and their level of generality is probably inadequate to integrate "particularity" (empirical observations) and "generality" (general theories), which is the goal of all natural sciences. Such a system might also be particularly detrimental to the development of a science such as ecology that is intrinsically multidisciplinary.

Figure 1. Use of models and evolutionary thinking in ecology journals and link with Impact Factors.

Top panel, relation between the percentage of articles dealing with soil ecology and (1) the percentage of articles using a model, R² = 0.62, F = 25.8, P<0.0001; (2) the percentage of articles dealing with evolution, R² = 0.87, F = 110.5, P<0.0001. Bottom panel, relation between the percentage of articles dealing with soil ecology and the Impact Factor, R² = 0.43, F = 12.0, P = 0.0032. Each point corresponds to one of the eighteen journals investigated. Axes have a logarithmic scale. The log-log linear regression is highly significant for each of these relations with a negative slope in each case. For each relation both the raw data and the regression line are displayed. See Table 1 for raw data.

Figure 2. Citation practices in soil ecology.

Distribution of articles cited by Soil Biology & Biochemistry (left panel) and citing articles published in Soil Biology and Biochemistry (right panel) according to the category of journal they have been published in. Issues 10, 11 and 12 of the volume 35 (2003) of Soil Biology & Biochemsitry have been scanned in 2007 using the bibliographic data base ISI Web of Science, so that article citing these issues of Soil Biology & Biochemistry have been published between 2003 and 2007 while all the articles cited in these issues have been taken into account whatever their year of publication. Categories of journals are described in the text.