Wheat seedlings were grown on soils with different resistances to root penetration and at various ambient partial pressures of CO2, pa. The partial pressure pa was varied either (a) as soon as seedlings emerged and until the end of the experiment (i.e. until c. 2 weeks after sowing, until the three leaf stage) or (b) only after the effects of high soil resistance were well established, and then for two days. In the first week after emergence, a higher soil resistance induced slower growth of both roots and shoot. This effect was independent of pa. We conclude that at this stage, with much of the plant's carbon being supplied from hydrolysis of seed reserves, growth at high soil resistance was not C-source limited. There was, however, a positive relationship between shoot growth rate and the carbohydrate concentration in that tissue, the degree of which varied with soil resistance. A given carbohydrate concentration was associated with a lower growth rate at high, compared with low, soil resistance. We deduced that this reduced sensitivity of growth to internal availability of substrate carbohydrate may be one manifestation of a sink limitation. Subsequent to the first week following emergence, roots grew faster on soil with higher resistance, while the shoot continued to grow more slowly. As seed reserves were becoming exhausted, growth became sensitive to pa, i.e. somewhat source-limited. This response to partial pressure of CO2 was mainly seen in the roots. The correlation between growth rate and carbohydrate levels was maintained. We conclude that increased soil resistance induces a factor which retards shoot growth, partly by decreasing its sensitivity to carbohydrate levels, making more carbon available for root growth. It is unclear whether or not this factor also directly affects the physiology of the roots. A higher soil resistance resulted in greater respiratory losses as a proportion, Φ, of carbon fixation. Reduced pa also induced greater Φ. Theory is developed relating Φ to carbon assimilation, allocation and maintenance. It leads to the prediction that respiratory losses, as a proportion, Φ, should be increased under most environmental conditions which reduce relative growth rate.