The title problem is studied by means of the stability criterion due to hill (1957). The stationary principle associated with Hill's criterion provides precise stability bounds for the problem which are compared with the critical loads predicted by the classical Euler formula. It is found, if one uses the slenderness ratio defined in engineering practice, that the Euler formula gives excellent estimates of the critical stress for slenderness ratios exceeding 35. Furthermore, if the results are plotted in terms of the dimensions of the underformed body and the ‘ nominal stress,’ then the Euler formula gives conservative estimates of the critical stress for slenderness ratios exceeding 6. It is proved, unless passive constraints are acting, that only a ‘bending’ mode of instability is possible. This leads one to question the interpretation given by beatty and hook (1968) to some recent experimental work of theirs.