We examine the effects of date of birth on state selective education using the 1944 Education Act in England and Wales as a natural experiment. We compare the probabilities of gaining selective school entry - which in our study period meant attending a grammar school - before and after the Act using a difference-in-difference approach. Before 1944, grammar school entry was achieved either noncompetitively through fee-paying or free based on a competitive 11+ exam. After 1944, all children were required to take a competitive 11+ exam and about one-third gained a grammar school place. Pre-1944 we find the children born in the middle or late in the school year (January to August) fared significantly worse in gaining a grammar school place than those born at the beginning of the school year (from September to December). Post-1944, the prospects of grammar school entry among children born in the middle of the school year (January to April) improved considerably. We argue that a greater recourse to age standardisation of 11+ test scores may well have accounted for this outcome. The youngest "summer children" (those born at the end of the school year from May to August) remained significantly disadvantaged, however. A strong influence was the practice of streaming (or tracking) junior school children at age 7 into classes delineated by average ability.