where m and n are the valencies of the atoms making up each bond. In the paper "Development of 3-Dimensional Molecular Descriptors" Computers Chem. Vol. 14, No.3 pp. 237-246, 1990, the authors state that the distance matrix is viewed as input and for each CC distance the appropriate weight is found using 1(mn)^1/2. Do you multiply each entry of the geometric distance matrix by this weighting factor and proceeding in the same manner as with the adjacency matrix to obtain the atomic path numbers and molecular path number, P1, the molecular connectivity? In doing so you don't get their values for their connectivity indices.