Holographic exposure (i.e., exposure of a photoresist coated substrate in an interference field of two light beams), followed by development in alkaline solution and subsequent ion-beam etching, remains to be the most important technique for fabricating diffraction gratings, especially large-area gratings, despite the advance of other techniques. In this process as an intermediate product the photoresist grating serves as a mask for ion-beam etching. The shape and critical dimensions of the photoresist mask directly determine the shape and critical dimensions of the etched end product, which in turn determine the performance parameters of the grating. In a crude and yet often effective approximation the shape of the mask can be taken as rectangular and the critical dimension is duty cycle (ratio of ridge width to period). The groove depth is not critical, as long as it is large enough, and it can be controlled easily by adjusting the photoresist layer thickness during spin coating and by detecting the critical turning point of the diffraction efficiency curve of one of the grating’s dispersive orders during development when the photoresist in the trough is completely removed. Once the maximum groove depth has been reached the following development process predominantly manifests as reduction of duty cycle. While the efficiency monitoring method is effective for detecting the point of trough clearing, it is ineffective for measuring duty cycle. A one-dimensionally periodic grating is an optically anisotropic structure when the groove period and light wavelength are comparable. Diffraction efficiencies of TE and TM polarizations are different and their ratio is a monotonic function of duty cycle in the range of duty cycle of interest. We propose to use this property to estimate the duty-cycle of surface-relief holographic gratings during development. We will present our theoretical simulation results and experimental results.