PERSPECTIVES S ince the work of James Clerk Maxwell

that light is an electromagnetic wave. An intricate mechanism generates magnetic fi elds around the electric fi elds, and vice versa. In the optical-wavelength range, experimen-tal studies have been limited to probing only the electric-fi eld components. On page 550 of this issue, Burresi et al. ( 1) report direct mea-surements of the magnetic-fi eld components of light obtained with a nanostructured metal-lic probe at the tip of a sharp glass fi ber. The instrument used by Burresi et al. can be viewed as a variant of the scanning tunneling microscope ( 2). Rather than imaging atoms on the surface, scanning near-fi eld optical micros-copy (SNOM) ( 3, 4) collects light from a