Here, we investigated how root age and mode of death influenced their subsequent turnover and rate of C loss from soil. Young white-coloured and older pigmented roots of Cistus monspeliensis were excised (to simulate death by mechanical severance) or frozen (to simulate death by cell rupture) and immediately buried in soil. CO2 loss from soil was then measured over time. In a parallel experiment, the rate of CO2 loss from severed or ruptured roots in the absence of soil was determined. Our results revealed large differences in root chemistry related to age, with young roots having a lower C:N ratio and a greater nutrient content (soluble C, N, P and K). Both root age and mode of death resulted in very different temporal patterns of C release from soil. The amount of C lost from soil followed the series: severed white roots (42.6 ± 3.3 mg C) > ruptured pigmented roots (27.7 ± 0.4 mg C) = ruptured white roots (27.1 ± 0.5 mg C) > severed pigmented roots (10.1 ± 1.0 mg C) > soil only (3.0 ± 0.2 mg C). Therefore, depending on the treatment, 7 to 41% of the total root-derived C was lost as CO2 over the duration of the experiment. Comparison with soil-free treatments revealed that the CO2 release from the severed roots buried in soil was not associated with microbial breakdown but caused by root-induced autophagy in an attempt to keep themselves metabolically active. Ruptured roots also induced a rapid loss of CO2 which we ascribe to the diffusive loss of root solutes into the soil and subsequent microbial mineralization. Surprisingly, the rate of C loss from soil was greater from the severed root tips than those that were ruptured. Our results imply two distinct routes of C loss dependent on how roots die, one which bypasses the microbial community and one which flows through it.