1
gion à géométrie variable. M. Norton écrit : « We
need
not fear that we cannot manage implications of non-coincidence of bou
2
implications of non-coincidence of boundaries, we
need
not fear those complications ; one cannot meet complexities in advanc