Abstract
The magnetic powder patterns found on polished iron crystals (which may contain a few percent of silicon) have been studied more carefully by using a macroscopic model and by investigating the forces on the powder particles. Reasons are given for preferring one of two simple models to explain the patterns; some interesting implications of the proposed model are then considered. Experiments are described which indicate that the structures under discussion originate during the polishing operation, but that other than magnetic causes must be responsible for the remarkable regularity of the patterns.