When defining a group, do we need to include closure? This is a detail that is often touched upon when the notion of a group is introduced to undergraduates. Should closure be listed as an axiom in its own right, or should it be regarded as an inherent property of the binary operation? There is no clear answer to this question, although there are firm opinions on both sides. Indeed, a very brief survey of group theory textbooks found in [1, pp. 458-459] suggests that there is a rough 50 : 50 split between authors who include closure explicitly and those who do not. In this Article, we go back to the beginning of the twentieth century to provide some historical perspective on this problem.

In 1879, the Irish astronomer Robert Stawell Ball published a slim book entitled simply Mechanics [1]. This book appeared as part of the series of ‘London Science Class-Books’, published by Longmans, Green & Co. These books were intended as elementary science texts for use in schools, and, as a consequence, their mathematical content was quite basic — even for those books on supposedly mathematical topics. In this article, I will look at Ball's handling of his subject, and compare his book to its distant ancestor: Newton's Principia.