Babies born small or large are more likely to have health and development problems than those of a 'normal' size. But in reality it is difficult to know just what a small or large size is. A common measure that has been used in lots of studies is to take account of how far on the mother was in her pregnancy at the time their child was born. This allows us to tell the difference between, for example, a baby who is small because they delivered too early but might have been growing normally in the womb and one who is small because they were undernourished and did not grow properly in the womb. Similarly it allows us to tell the difference between a baby who might be large because they stayed in the womb for longer than average and one who grew too rapidly because they were overnourished. Calibrating birth size against how long a woman has been pregnant is done using growth charts, similar to those used to plot an infants growth after they are born. Until recently these charts have reflected average growth patterns for a specified population. For example if a baby today was plotted on a UK 2000 growth chart of birth weight for gestational age and they were on what was called the 50th percentile. This would indicate that their weight was the same as the average for all other babies of the same sex and who had had the same time in the womb who were born in the UK in 2000. Recently it has been suggested that instead of those total population averages 'customised growth charts' should be used to define whether a baby is too heavy or light after accounting for how long they have been in the womb. As well as time in the womb these 'customised growth charts' take account of what are assumed to be characteristics that affect fetal growth in a 'biological' (non-pathological) way: the mother's weight and height at first clinic, her ethnicity, whether this is her first, second, etc child. These charts are somewhat controversial as on the one hand they might be correctly allowing for 'biological' differences in size but on the other they might (for example) be assuming that it is normal for a particular ethnic group to always be smaller than another ethnic group which might not be the case.