In this paper, I shall discuss how national income should be measured in an 'imperfect' economy, where feasible policy instruments such as taxes, tariffs, quotas, and quantitative controls do not operate in a lumpsum manner, and may be far from their optimum level. In particular, I want to examine the meaning of national income, and its parts, when they are measured in terms of 'world prices'; or, more precisely and more generally, in terms of the accounting prices that might be computed as a guide to particular production decisions, in the public sector and elsewhere. I have found this set of questions confusing, and it may be that others do too, and will welcome an attempt at clarification.