Despite their generally low volume fraction, Fesingle bondTi oxides have the potential to greatly influence the eruptive style because they lower the supersaturation pressure for heterogeneous bubble nucleation. Once nucleated, bubbles respond fast to pressure changes, fostering rapid expansion and explosive behavior. Yet, oxide microlite quantifications are often absent from data of explosive products. We used new, re-analyzed, and published data to build a compilation of oxide number densities (ONDs) and vesicle number densities (VNDs) of trachytic and calc-alkaline products. Four eruptive styles were selected: 1) Vulcanian explosions from Soufrière Hills volcano, Montserrat, Lascar volcano, Chile, and Kilian volcano, France, 2) blasts from Mt. Pelée volcano, Lesser Antilles, Mount St. Helens, USA, and Merapi volcano, Indonesia, 3) a sub-Plinian explosion from Merapi volcano, and 4) lava dome effusions with intermittent collapse from Soufrière Hills and Mt. Pelée volcanoes. Natural samples were separated into two groups according to the dominant texture of the products of each event: 1) vesicular pumice clasts from explosions with a strong vertical component and 2) dense clasts with diktytaxitic textures from dome collapse event and lateral blast. Group 1 samples are either distributed alongside the 1:1 trend between VND and OND that spans from 1015 to 1017 m−3, or have a constant VND of 1016–16.5 m−3 regardless of OND. A large proportion of oxides (55–100%) touch vesicles. A more variable proportion of vesicles (16–72%) are in contact with oxides because of syn-explosive growth and coalescence. Group 2 samples have ONDs in the same broad range as group 1 samples. We also used new and published data to build a compilation of ONDs and VNDs of five series of experimental decompression of rhyolitic and phonolitic melts. In samples with OND ≈ VND, most bubbles are in contact with more than one oxide and 64–88% of the oxides are in contact with bubbles. Such high levels of connectivity suggest that the role of oxides in controlling bubble nucleation has been underestimated. When VND ≥ OND, nucleation densities of experimental vesicles can be reproduced by heterogeneous nucleation models, which we used to calculate syn-explosive decompression rates from VNDs at Merapi, Soufrière Hills, and Kilian. These rates and textural evidence suggest that the decompression front accompanying these Vulcanian and sub-Plinian explosions is responsible for syn-explosive bubble nucleation. We calculated the average pre-explosive ascent rates necessary to yield the observed ONDs at Soufrière Hills and Merapi volcanoes. The resulting rates, 0.005–20 m/s, overlap considerably with the range of critical ascent rate inferred for the effusive–explosive transition, supporting the hypothesis that this transition is primarily controlled by oxide microlites in the conduit because oxides are a primer for explosive behavior when present in sufficiently high number densities. Focusing on the older eruption of Kilian, for which no observation is available, we infer that pre-explosive ascent rates of >7 × 10−3 m/s were necessary for explosive behavior to occur.