Magnetic particles, especially iron rust and magnetic fraction of placer deposits, although granular and dry, form spikes and troughs pointing directly to or away from (parallel to) magnetic fields, and collapse into granules while magnetic field is withdrawn. Why these spikes form? I have somewhat qualitative college-knowledge on Magnetism on iron and rocks. For ferrofluids, the spikes and trough increases surface energy (and often also gravitational potential energy) as magnetic energy is applied on the sample. How does this rationale hold for already granulated materials, which do not attract one another in absence of magnetic field?