Hydrological cycle is expected to become the primary cause of ecosystem's degradation in near future under changing climate. Rain manipulation experiments under field conditions provide accurate picture on the responses of biotic processes to changed water availability for plants. A field experiment, mimicking expected changes in rain patterns, was established in a Mediterranean shrub community at Porto Conte, Italy, in 2001. In November 2011 Cistus monspeliensis, one of the dominating shrub species in the Mediterranean basin, was 13C labelled on plots subjected to extended rain shortage period and on control non manipulated plots. Carbon (C) allocation was traced by 13C dynamics in shoots, shoot-respired CO2, roots, microbial biomass, K2SO4-extractable C and CO2 respired from soil. Most of the recovered 13C (60%) was respired by shoots within 2weeks in control plots. In rain shortage treatment, 13C remained incorporated in aboveground plant parts. Residence time of 13C in leaves was longer under the rain shortage because less 13C was lost by shoot respiration and because 13C was re-allocated to leaves from woody tissues. The belowground C sink was weak (3-4% of recovered 13C) and independent on rain manipulation. Extended rain shortage promoted C exudation into rhizosphere soil in expense of roots. Together with lowered photosynthesis, this "save" economy of new C metabolites reduces the growing season under rain shortage resulting in decrease of shrub cover and C losses from the system on the long-term.