T H E C O L L E C T I O N O F R E C O R D S F O R ANALYSIS S I R S , -- Y o u r correspondent, Mr. T. C. Smout, raises issues of fundamental importance (antea, vol. xlix, pp. 287-288). If a paper is put forward for publication in a scientific form and purporting to have some scientific value, the author or authors should be prepared to abide by the canons of scientific behaviour. In particular, when the data available to the authors are known to be or are strongly suspected of being incomplete, they must either state so in their publication and be prepared for their conclusions to be upset by more complete data or withhold publication until they have obtained most, if not all, of the necessary facts on which their conclusions are to be based. This is particularly true of work on problems relating to migration where considerable inferences are often made on comparatively few data. There are two main courses of action open to the worker who is unable to obtain all the data he requires within a given time. One is to wait, as suggested above, for there is " n o justification for the publication of hasty or slipshod work or for flooding the literature with trivia" (Mayr et al., 1953). The other is to publish, but, because of the lack of complete data, the treatment should be very brief and the conclusions