[59] In this essay we are concerned only with the landholders, not with the wage workers. The relative number of persons holding land and of agricultural labourers without land is an important question on which it is not easy to get light. The surveys and rentals, a species of private census invaluable in giving information about the holders of property, tell us only the number of householders, and as the labourers employed in agriculture (like many of those employed in manufacturing industry) usually lived on the premises of their masters, they do not enable us to calculate the number of those living entirely by their labour. Still, since they include all tenants, whether holders of a cottage only or holders of land in addition, they enable us to say what proportion of heads of families held land, and what proportion had none, or none except a garden. This is of some importance. A tenant holding even as much as fifty acres can hardly have employed more than two or three agricultural labourers, and most tenants held less than this; so that in those places where the cottagers form a small proportion of the whole population we may conclude that a large proportion of the villagers were landholders (for the figures on this point see the tables given below).