Nitrate dosing is commonly used to control hydrogen sulfide production in sewer systems. However, quick rebound of the sulfide concentration after nitrate depletion has been observed and results in more serious odor and corrosion problem. To investigate the mechanism of sulfide regeneration in the nitrate-free period, a laboratory-scale sewer reactor was run for 30 days to simulate sulfide production and oxidation with intermittent nitrate addition. The results show that nitrate addition substantially reduced the sulfide concentration, but the produced elemental sulfur was then quickly reduced back to sulfide in nitrate-free periods. This induced more and more sulfide production in the sewer reactor. Elemental sulfur and polysulfide reductions were found in the sewage in nitrate-free periods, showing their contributions to the sulfide regeneration. Through batch tests, polysulfide was confirmed as the key intermediate for accelerating sulfur reduction during the nitrate-free period in the sewer. Sulfide production rates significantly increased by 65% and 59% in the presences of tetrasulfide and sulfur with sulfide, respectively, at the beginning of the test. While polysulfide formation was prevented by the ferrous chloride addition, the sulfur reduction rate remarkably decreased from 12.8 mgS/L-h to 1.8 mgS/L-h. This indicates that direct sulfur reduction was significantly slower than the indirect sulfur reduction via polysulfide; the latter process could be the cause for the quick rebound of the sulfide concentration in the sewer with intermittent nitrate dosing. Thus, the pathways of sulfur transformations in a sewer, both in the presence and absence of nitrate, were proposed. Microbial community analysis results reveal that some common sulfate-reducing bacteria (SRB) genera in sewer sediment were possible sulfur reducers. According to this finding, the effect and strategy of nitrate dosing for hydrogen sulfide control in sewers should be re-evaluated and re-considered.