Phreatic eruptions are common in the recorded history of White Island, New Zealand. Although the larger eruptions have been described in literature, little attention has yet been given to the smaller, more frequent phreatic activity. In addition, the style in which steam bubbles are released during phreatic eruptions at the surface can be highly variable and is poorly understood. Throughout 2013, multiple episodes of phreatic bubble bursts in a mud-sulphur pool occurred at White Island. The first of these episodes, lasting from January 15 to February 7, was well preserved in the video record by GNS staff, tourists and tour operators. Analysis of these videos showed that the mud surface expressions of rising bubbles varied over this period as the apparent high water fraction of the mud pool was first depleted and then regained. Here, these expressions are classified into four regimes, which progressed from 1) low ~8 m, highly fluidal structures to 2) brittle 'starbursting' of hemispheres and heaves of ~40 m height to 3) rapid gas jets followed by high heaves up to ~102 m and finally 4) dry mud venting up to ~67 m height with the complete pool desiccation. Regression back through these regimes from January 30 coincided with the pool returning to the initial fluidal state and a similar depth. Experimental modelling of bubble bursts in mud was conducted in order to identify the influence of viscosity and bubble shape, length and depth on the White Island regimes. An analogue to the White Island mud was created by mixing of kaolinite powder and water to controlled ratios. Bursts were performed at depths of 0.5, 5 and 10 cm in 9 different mud ratios corresponding to an increasing viscosity. Results of these experiments show that viscosity has a negative influence on heave heights but controls the transition from fluidal to brittle structures. Importantly, both a shallowing of bubble depth and an increase to bubble length are shown to increase the height of heaves. Relating experimental results to White Island observations provides insight to bubble processes not observable in the video record. The higher heaves observed in conjunction with viscosity increase imply that bubble morphology or depth must be varying throughout the episode. Here, increases in bubble length towards long, conduit-controlled slug-shaped bubbles is shown to be a possible mechanism for the increasing ejection heights. Decrease of the mud pool level by desiccation, results in shallower bubble depths, and is also considered to be an influence to the observed increased explosivity. The more brittle behaviour of bubble bursts and mud heaves is shown to be related to the increasing viscosity of the mud pool. These results ultimately fill an absence in knowledge of phreatic processes occurring at White Island. The evolution of the White Island eruption to more explosive regimes ejected ballistics to increasing distances. Similar bubble processes are associated with other volcanic and geothermal systems such as lava lakes and geysers. Evolution of eruptions at these systems could therefore result in similar ballistic hazards. Understanding the processes responsible for the White Island evolution may provide insight and improve hazard assessments at White Island and other systems around the globe.