A modified FORTRAN code for R-mode correspondence analysis is presented for large-scale digital image analysis problems. It reduces the required memory capacity by more than a factor of 4 M, where M is the number of variables considered per sample. The original code, CORSPOND, requires 4 NM memory, where N is the total number of samples that are considered. The modified code requires only an N-dimensional vector to form the matrix S. For a PC having 128 MB RAM, the code can be used for a problem with up to N=253,000 and M=30. The code automatically detects, then shifts negative values to zeros or positive values. Applications to grey-level co-occurrence matrix parameter data are used to demonstrate the efficiency of the code.