On this account Malthus entirely disagreed with Hume, who supposed that the Roman world was probably most populous during the long peace under Trajan and the Antonines. Wars, he says, do not depopulate much while industry continues in vigour: and peace will not increase the number of people when they cannot find means of subsistence. “The renewal of the laws relating to marriage under Trajan indicates the continued prevalence of vicious habits, and of a languishing industry, and seems to be inconsistent with the supposition of a great increase of population.” Hume also thought that the population of the ancient world was greater than in modern times, because, he said, there were hosts of domestic servants in modern States remaining unmarried. But the contrary inference, says Malthus, seems to be the more probable. When the difficulties attending the rearing of a family are very great, and, consequently, many persons of both sexes remain single, we may naturally suppose that the population is stationary, but by no means that it is not absolutely great; because the difficulty of rearing a family may arise from the very circumstance of a very great absolute population, and the consequent fulness of all the channels to a livelihood; though the same difficulty may undoubtedly exist in a thinly peopled country, which is yet stationary in its population. The number of unmarried persons in proportion to the whole number, says Malthus, may form some criterion by which we may judge whether population is increasing, stationary, or decreasing; but will not enable us to determine anything respecting absolute populousness. Yet even in this point we may be deceived, since, in some southern countries early marriages are general, and very few women remain in a state of celibacy, yet the people not only do not increase, but the actual number is perhaps small. In this case the removal of the preventive check is made up by the excessive force of the positive check. The sum of all the positive and preventive checks taken together, forms, undoubtedly, the immediate cause which represses population; but we never can expect to obtain and estimate accurately this sum in any country; and we can certainly draw no safe conclusion from the contemplation of two or three of these checks taken by themselves, because it so frequently happens that the excess of one check is balanced by the defect of some other. Causes which affect the number of births or deaths may or may not affect the average population, according to circumstances; but causes which affect the production and distribution of the means of subsistence must necessarily affect population; and it is therefore on these causes, besides actual enumerations, on which we can with any certainty rely. “All the checks to population, which have been hitherto considered in the course of this review of human society, are clearly resolvable into moral restraint, vice, and misery.”