Over the last two centuries,the rates of disturbances affecting soils have considerably increased.Modifications concem physical,chemical and biological properties.The gradient of human impact on soils induces in parallel increasing disturbances from natural soils,to Anthrosols and Technosols.Up to now,scientific literature has mainly focused on a few selected examples and case studies of highly anthropized soils (e.g.public park soils,shunting yard soils,settling tank soils,brownfield soils) and no real global statistical description is available.The aim of this work is i) to describe the heterogeneity of highly anthropized soils (HAS) and ii) to highlight their differences with non-or slightly-anthropized soils (NOSAS).We have selected HAS as soils with,on the one side,urban,military or industrial use and on the other side with forest or agricultural use when physical degradation or strong addition of artifacts occur.Results of fifteen years of pedological campaigns have been compiled to create two databases of soil characteristics:(NOSAS∶ n=378,HAS∶ n=338).Samples from the upper layer of soils were characterized for agronomic parameters (pH,CEC,clay,organic carbon,organic nitrogen,C∶N ratio,CaCO3,and available P2O5),as well as parameters specific for anthropization:i) metallic pollution (total Cd,Cu,Ni,Pb and Zn) and ii) human activity indicators (content of transported material,artifacts,and coarse fraction (>2mm)).The distribution of soil characteristics was presented and compared on box plots of each variable by running Wilcoxon non parametric tests.Our work highlighted the fact that HAS mainly presented coarse textures.They contained more coarse fraction (>2mm) and sand (>50μm) than NOSAS.Average pH values of HAS were higher than 7.5,as 75% percent of them were carbonated,and their variability range was larger.Extreme organic carbon and nitrogen concentrations were measured resulting in considerable variations of C:N ratio.Concentrations of total Cd,Cu,Pb and Zn presented almost the same distribution and in half of the samples,they were much more concentrated (2 or 3 magnitude orders) in HAS than in NOSAS.The total content of Ni was lower in HAS.Most of the presented heterogeneity can be explained by the addition and the recent transformations of organic and/or mineral artifacts in HAS (e.g.addition of wastes,by-products or rocks,cutting of soil surface,mixing of soils).Highly anthropized soils are mainly young soils where only early pedogenetical processes take place.Therefore,their characteristics are strongly contrasted with more natural and evoluted NOSAS.The contribution of our work to the classification of soils strongly impacted by human activities will be presented.Are HAS significantly different from NOSAS? Could it be possible to classify some HASs using soil types classically proposed in soil classifications or is it always necessary to describe them as Anthrosols or Technosols?