The paper discusses the duality symmetries of type II string theory compactified on a six-torus, which leads to N = 8 supergravity. It shows that the symmetry group E₇(7) is broken by quantum effects to a discrete subgroup E₇(ℤ), which includes both T-duality and S-duality groups. The authors conjecture that E₇(ℤ) is an exact 'U-duality' symmetry of type II string theory, unifying S and T dualities and mixing sigma-model and string coupling constants. They argue that certain extreme black hole states should be identified with massive modes of the fundamental string. The gauge bosons from the Ramond-Ramond sector couple to solitons rather than string excitations. The paper also discusses similar issues in other compactifications, including K₃ × T² and eleven-dimensional supermembrane theory. The effective field theory for compactified string theories includes a spacetime sigma model with target space as the moduli space of the torus. The constants g_ij and b_ij are the expectation values of the scalar fields. The group O(n,n) acts on the moduli space, and a discrete subgroup O(n,n;ℤ) takes a string theory into an equivalent one. The T-duality group is O(n,n;ℤ), and the true moduli space is the moduli space of the torus factored by the T-duality group. For the heterotic string compactified on a six-torus, the T-duality group is O(6,22;ℤ). The effective field theory is N = 4 supergravity coupled to 22 abelian vector multiplets, giving 28 abelian vector gauge fields. The effective field theory has an SL(2;ℝ) × O(6,22) invariance, which is broken to the discrete subgroup SL(2;ℤ) × O(6,22;ℤ). The O(6,22;ℤ) factor extends to the full string theory as the T-duality group, and it is conjectured that the SL(2;ℤ) factor also extends to the full string theory as an S-duality group. The paper also discusses the Bogomolnyi bound for states carrying electric or magnetic charges, which is believed to be unrenormalized to arbitrary order in the string coupling constant. The Bogomolnyi bound is U-duality invariant, and the spectrum of Bogomolnyi states should also be duality invariant. These states include winding and momentum modes of the fundamental string and those found from quantization of solitons. The authors argue that soliton states in the type II string fall into representations of the U-duality group because this is a symmetry of the equations of motion. They also discuss the identification of soliton states with extreme black holes andThe paper discusses the duality symmetries of type II string theory compactified on a six-torus, which leads to N = 8 supergravity. It shows that the symmetry group E₇(7) is broken by quantum effects to a discrete subgroup E₇(ℤ), which includes both T-duality and S-duality groups. The authors conjecture that E₇(ℤ) is an exact 'U-duality' symmetry of type II string theory, unifying S and T dualities and mixing sigma-model and string coupling constants. They argue that certain extreme black hole states should be identified with massive modes of the fundamental string. The gauge bosons from the Ramond-Ramond sector couple to solitons rather than string excitations. The paper also discusses similar issues in other compactifications, including K₃ × T² and eleven-dimensional supermembrane theory. The effective field theory for compactified string theories includes a spacetime sigma model with target space as the moduli space of the torus. The constants g_ij and b_ij are the expectation values of the scalar fields. The group O(n,n) acts on the moduli space, and a discrete subgroup O(n,n;ℤ) takes a string theory into an equivalent one. The T-duality group is O(n,n;ℤ), and the true moduli space is the moduli space of the torus factored by the T-duality group. For the heterotic string compactified on a six-torus, the T-duality group is O(6,22;ℤ). The effective field theory is N = 4 supergravity coupled to 22 abelian vector multiplets, giving 28 abelian vector gauge fields. The effective field theory has an SL(2;ℝ) × O(6,22) invariance, which is broken to the discrete subgroup SL(2;ℤ) × O(6,22;ℤ). The O(6,22;ℤ) factor extends to the full string theory as the T-duality group, and it is conjectured that the SL(2;ℤ) factor also extends to the full string theory as an S-duality group. The paper also discusses the Bogomolnyi bound for states carrying electric or magnetic charges, which is believed to be unrenormalized to arbitrary order in the string coupling constant. The Bogomolnyi bound is U-duality invariant, and the spectrum of Bogomolnyi states should also be duality invariant. These states include winding and momentum modes of the fundamental string and those found from quantization of solitons. The authors argue that soliton states in the type II string fall into representations of the U-duality group because this is a symmetry of the equations of motion. They also discuss the identification of soliton states with extreme black holes and