The current study aimed to examine how the recognition of grey-scale photos of fearful or angry female bodies would be affected by three conditions: social situation (single person vs. facing vs. nonfacing dyad) the orientation of figures (upright vs. inverted); emotion complementarity (same vs. complementary). We hypothesized that the recognition of emotions would be the most accurate when either single or facing body pairs were presented, while the inversion would impair the perception of affective expressions. Facing bodies in fact had an advantage over nonfacing ones, same emotion condition also had higher accuracy than complementary, as well as the overall accuracy was higher for anger than fear, thus context was an important factor in differentiating between these two negative emotions. Inversion effect was not confirmed for emotions conveyed by bodies, therefore our results demonstrated that not only configural, but part-by-part analysis is also required for emotion recognition. Keywords: body inversion effect, social interactions, bodily emotional expressions