TM growth rates and resistive layer size for different numerical schemes. Left: growth of the (2D) TM instability for different treatments of the parallel current and a fixed number of points per current sheet width p = 6. We display our standard method for treating J_|| (i.e., with a fourth-order accurate discretization of the derivatives, solid lines), extreme cases in which J_|| = 0 (dotted lines, following Yu 2011), and the case in which the algebraic condition D ⋅ B = 0 is not enforced (dashed lines). Different colors correspond to distinct reconstructions (see legends). In the case of no enforcement of the D ⋅ B = 0, the results are indistinguishable for the three MP reconstructions. Right: size of the resistive layer L_R normalized to the grid spacing Δz in the direction perpendicular to the current sheet for all the reconstruction methods (see legends) and models (A, circles and B, squares).