The characterization of silk properties requires a reliable measurement of stress-strain curves from tensile tests, which calls for a detailed analysis of what is considered the cross section of the sample and how it varies during the experiments. Here, spider silk fibers from the major ampullate gland (MAS) of Argiope trifasciata spiders are tensile tested, and the cross-sectional area is measured under different strained configurations. It has been found that the fiber volume remains practically constant during stretching, and deformation proceeds homogeneously in all the fibers. The conservation of volume is validated independently of the type of fiber and the strain level. This result, applied to compute true stress-strain curves for different MAS fibers, shows that the description of their properties depends noticeably on which set of tensile parameters is chosen (true or engineering), and that engineering values could lead to misinterpretation of experiments that combine results from different strain ranges.