We introduce the divergence and the gradient for functions defined on a mea
sure chain, and this includes as special cases both continuous derivatives
and discrete forward differences. It is shown that in one dimension, subjec
t to Dirichlet boundary conditions, the divergence and the gradient are neg
ative adjoints of each other and that the divergence of the gradient is neg
ative semidefinite. These are well-known results in the continuous theory,
and hence, mimic those properties also for the case of a general measure ch
ain. (C) 2001 Elsevier Science Ltd. All rights reserved.