The conservation of magnetic flux demands a magnetic field fine structure as small as 15 km to explain the net circular polarization observed in penumbrae (Sánchez Almeida cite{san98a}). A recent work by Martínez Pillet (cite{mar00}) claims that modeling penumbrae as a collection of fluxtubes 100 km wide suffices. We identify several shortcomings of such modeling so that it is presently unclear whether it really contradicts the former conclusion.