Inasmuch as it has been claimed that inductance plethysmography can measure cross-sectional area changes, we tested this assumption. We present experimental and computed relationships between self-inductance (L) of coils and areas (A) included inside for a coil with a well-defined side wavy pattern (triangular or sinusoidal) and for a real belt (Respitrace) placed on elliptical or rectangular configurations. The results are applied to the physiological field using measurements obtained from a computed tomography experiment. We demonstrate that the L-A relationships vary not only with shape or ellipticity of the cross section but also with the wavy pattern shape. This last parameter is critical because it is difficult to actually control. When the coil wavy pattern remains steady, there are some physiological situations where inductance plethysmography can more accurately estimate area changes: when the configuration shape is constant, the correspondence between delta L and delta A is almost linear with a shape-dependent sensitivity; when the configuration is nearly circular (ellipticity in the range 0.8-1), the relative error in delta A estimation is less than 5%.