Theory on the density-body mass (DBM) relationship predicts that the density of animal species decreases by the power of −0.75 per unit increase in their body mass, or by the power of −1 when taxa across trophic levels are studied. This relationship is, however, largely debated as the slope often deviates from the theoretical predictions. Here, we tested the ability of the DBM relationship to reflect changes in the structure of communities subjected to an anthropogenic disturbance. The slope would become less steep if smaller animals were more impacted by the disturbance than the larger ones, whereas the slope would become steeper if larger animals were more affected than the smaller ones. We tested the changes in the DBM relationship by sampling soil fauna, i.e. nematodes, Collembola and larger arthropods, from a semi-arid grassland before and after spraying diesel fuel as disturbance. We applied three different treatments: a control, a light disturbance and an intense disturbance. We found that the slopes of the DBM relationships before the disturbance were around −1 as predicted by theory. The slope became more positive (i.e. less steep) just after the disturbance, especially after the intense disturbance as smaller fauna suffered the most and early colonizers had larger body mass. Interestingly, we observed that the slopes converged back to −1 in two months post-disturbance. Our findings show that the response of soil fauna communities to anthropogenic disturbances could explain the large variation in observed slopes of the DBM relationships. We experimentally demonstrate that an animal community, when disturbed, shows a temporal pattern of DBM relationships ranging from deviations from the predicted slope to convergence to the predicted slope with time. We recommend that deviations in the DBM relationships after disturbances can provide insights in the trajectory community recovery, and hence could be used for biomonitoring.