No CrossRef data available.
Published online by Cambridge University Press: 22 January 2016
A nonsingular algebraic surface V is called a K3 surface if i) 1 = i.e. a canonical divisor Kv on V is linearly equivalent to zero; and ii) . When the characteristic is zero, condition ii) is equivalent to ii)′ q = dimension of the Albanese variety of V = 0, and always ii) implies ii)′ as in fact ≥ q ≥ 0, but in non-zero characteristic it can happen that > q ([6], [15]). When ii)′ is true, the algebraic and linear equivalences of divisors coincide on V, because of the duality between Picard and Albanese varieties of V, [8].