If all magnetic quantities are defined via fields, with no use of poles and with magnetic moment defined as the volume integral of magnetization, energy considerations yield the torque on a magnet immersed in a permeable fluid in an external field. Although the torque on a disk magnet is proportional to B, the torque on a needle magnet is proportional to μ0H, as reported experimentally by Whitworth and Stopes‐Roe. The experiment does not prove either A. E. Kennelly [J. Eng. Educ. 27, 290 (1935)] or Sommerfeld to be right or wrong, but demonstrates that magnetic theory based onpolescannot be applied tomagnets, and thereby strengthens the philosophy that electromagnetic theory should be based on fields.