Interest in the Si/Ag(110) system, which forms highly ordered linear nanostructures coined 'silicon nanoribbons', was recently boosted by the claim that such nanoribbons may be formed by silicon atoms arranged in a 2D honeycomb structure as in graphene, i.e. silicene. Despite such a revived interest, many discrepancies still exist in the recently reported results. This paper reports on a systematic investigation by scanning tunneling microscopy and low-energy electron diffraction of the Si/Ag(110) system as a function of the amount of deposited silicon and the deposition temperature. This reveals a complex interplay between these two factors, resulting in a rich array of possible self-assembled nanostructures and surface reconstructions. Several novel findings and clarification of the contradictory results reported in the literature are discussed in this work. In particular, the deposition temperature is demonstrated to be a key parameter to control the width of the Si nanoribbons produced. Recently, massive linear nanostructures were reported to be 'multilayer silicene', forming once the deposited silicon amount exceeds full coverage. However, we show that such nanostructures are also observed at low silicon coverage, demonstrating that their formation is exclusively determined by a deposition temperature higher than 460 K. On the other hand, for Si amounts higher than one monolayer the surface presents a novel c(8 × 4) reconstruction, which is responsible for the ×4 periodicity detected by LEED measurements, previously attributed to the 1.6 nm-wide nanoribbons overlayer or to 'multilayer silicene'. Finally, the large collection of acquired data also allowed us to single out image artifacts that may explain the contradictory results appearing in previous papers.