The eruption of Mount St. Helens from 2004 to 2006 has comprised extrusion of solid lava spines whose growth patterns were shaped by a large space south of the 1980-86 dome that was occupied by the unique combination of glacial ice, concealed subglacial slopes, the crater walls, and relics of previous spines. The eruption beginning September 2004 can be divided (as of April 2006) into five phases: (1) predome deformation and phreatic activity, (2) initial extrusion of spines, (3) recumbent spine growth and repeated breakup, (4) southward extrusion across previous dome debris, and (5) normal faulting of the phase 4 dome to form a depression, a shift to westward extrusion and overthrusting of earlier phase 5 products. Overall, steady spine extrusion gradually slowed from 6 m³/s in November 2004 to 0.6 m³/s in February 2006. Thermal camera data show that phase 1 activity included low-temperature thermal features, such as fumaroles, fractures, and ground warming related to rapid uplift, as well as deformation in the south moat of the crater. The relatively cold (<160°C) phreatic eruptions of early October heralded activity at a subglacial vent situated along the south-sloping margin of the 1980–86 dome. Thermal infrared imagery, documenting increased heat flow, presaged phase 2 extrusion of the October 11–15, 2004, lava spine. The thermal images of the extruding spine revealed a hot basal margin and highest temperatures of 600–730°C. During phase 3, a recumbent whaleback-shaped spine with a low-temperature shroud of fault gouge and a hot, U-shaped basal margin extruded. This spine pushed southward along the bed of the glacier until it encountered the south wall of the 1980 crater, whereupon it broke up, decoupled, and regrew. Continued southward growth of the recumbent spine pushed cold deformed rock, hot dome rubble, and glacier ice eastward at a rate of 2 m/d. In April 2005, breakup of the whaleback and growth of a lava spine across previous dome rubble heralded phase 4 spine thrusting over previous spine remnants. During phase 4, the active spine pushed southward with an increasingly vertical component and increasing incidence of large rockfalls. In late July, the spine decoupled from its source, the vent reorganized, and a new spine began to grow westward at right angles to the previous growth direction, defining phase 5. Dome migration again plowed glacier ice out of the way at a rate of about 2 m/d, this time westward. In early October, the spine buckled near the vent and thrust over the previous one. A massive spine monolith had been constructed by December 2005, and growth of spines with increasingly steep slopes characterized activity through April 2006. The chief near-surface controls on spine extrusion during 2004-6 have been vent location, relict topographic surfaces from the 1980s, and spine remnants emplaced previously during the present eruption. In contrast, glacier ice has had minimal influence on spine growth. Ice as thick as 150 m has prevented formation of marginal angle-of-repose talus fans but has not provided sufficient resistance to stop spine growth or slow it appreciably. Spines initially emerged along a relict south-facing slope as steep as 40° on the 1980s dome. The open space of the moat between that dome and the crater walls permitted initial southward migration of recumbent spines. An initial spine impinged on the opposing slopes of the crater and stopped; in contrast, recumbent whaleback spines of phase 3 impinged on opposing walls of the crater at oblique angles and rotated eastward before breaking up. Once spine remnants occupied all available open space to the south, spines thrust over previous remnants. Finally, with south and east portions of the moat filled, spine growth proceeded westward. Although Crater Glacier had only a small influence on the growing spines, spine growth affected the glacier dramatically, initially dividing it into two arms and then bulldozing it hundreds of meters, first east and then west, and heaping it more than 100 m higher than its original altitude.