Methods of approximation, and successive approximation, are often used in more advanced mathematics, but it is not widely realised that they can with advantage be introduced at quite an early stage in the teaching of school mathematics. This note is a plea that these methods should be taught earlier; indeed it is written in the belief that a wider use of these methods would have not only a salutary effect on teaching and learning alike, but would help to eliminate some of the widespread troubles with which teachers of science are generally confronted and with which they often ask us to help. How often are we told by science colleagues that pupils calculate a physical quantity, for instance, a specific heat, to four decimal places when it is likely to be in error in the secorld place! Similarly those of us who teach practical mechanics, know full well how the resources of four- or five-figure logarithmic tables are invariably used to the full, when the data are correct only to two significant figures.