The extent of the lastglacial icecap over the northern Rocky Mountains
of Montana has been inferred only from cursory correlation of its mar
ginal deposits, which leaves several significant unresolved controvers
ies. Those controversies all involve the maximum altitude and shape of
that interconnected ice mass, and include (1) the source of the Two M
edicine piedmont lobe (local or Cordilleran ice), (2) correlation of t
erminal deposits of the Flathead lobe (late lastglacial, early lastgla
cial, or pre-lastglacial), and (3) regional paleoclimate at last glaci
al maximum (dominant westerlies or local easterlies). Theoretical reco
nstructions of the glacial surface along major flowlines, constrained
by ice-marginal features, nunataks and breached divides throughout the
region, tentatively resolve those issues. The Two Medicine lobe was d
ominantly composed of northeast-flowing local ice, not ice of northwes
tern (Canadian cordillera) origin. Flathead lobe deposits include last
glacial deposits both supported by and without a component from the Sw
an Valley. There is evidence of only a regional westerly flow of moist
ure across the region. In addition, a proglacial lake 60 km long occup
ied the Swan Valley during the early stages of deglaciation. Although
subject to revision from fieldwork, glacial process modelling remains
one of the few geological arenas in which rigorous process modelling c
an be used to predict the evolution of form.