Magma degassing during the Plinian eruption of Novarupta, Alaska, 1912

We have modeled the nucleation and isothermal growth of bubbles in dacite from the 1912 Plinian eruption of Novarupta, Alaska. Bubble growth calculations account for the exsolution of H 2 O and CO 2 , beginning with bubble nucleation and ending when bubble sizes reproduced the observed size distribution of vesicles in Novarupta pumice clasts. Assuming classical nucleation theory, bubbles nucleated with a diameter of the order of 10 −8 m and grew to sizes ranging from 10 −6 m to greater than 10 −3 m, the typical range of vesicle sizes found in Novarupta pumice. The smallest vesicles in Novarupta pumices are also the most abundant and bubbles with radii of 10 −6 m to 10 −5 m comprise almost 90% of the entire bubble population. We find that these bubbles must have nucleated and grown to their final size within a few 100 milliseconds. Despite these extremely fast growth rates, the pressures of exsolved volatiles contained within the bubbles remained high, up to about 10 7 Pa in excess of ambient pressure. Assuming a closed‐system, the potential energy of these compressed volatiles was sufficient to cause magma fragmentation, even though only a fraction of the pre‐eruptive volatiles had exsolved. Unless the matrix glasses of Novarupta pyroclasts retains a large fraction of pre‐eruptive volatiles, the majority of magmatic volatiles (80–90%) was likely lost by open‐system degassing between magma fragmentation and quenching.