The fact that a society maintains vital records for demographic purposes does not necessarily mean that there is a ready-to-hand source of data for the study of migration. Such records are usually collected for some other purpose, usually administrative. In a census, while the record sheets tell the location of a given individual on a particular date, these are rarely available to investigators (in the United Kingdom not for 100 years) and the data are often engulfed in the categorization required for demographic analysis. Direct questions on migration are considered of low priority, time-consuming to administer, and are often discarded when the number of questions to be asked is limited by cost. The location may be specified in some unit of space of little interest to the enquirer and, if large, migration within it will not be indicated. The boundaries of the space units may have changed between censuses, so that the same address appears in what seem to be two different localities in consecutive censuses. Use of fixed periods of time gives no indication on multiple moves and return moves during the interval; if a question is asked on absence from the census address for, say, a year, this will not identify the seasonal migrants. Census data can reveal little on the motivations for migration or the characteristics of those who migrate. Something may however be inferred from them.