Disclosed herein is method and apparatus for determining the nature of an unknown substance, which includes apparatus for entering a peak table of the spectrum of an unknown substance into computing apparatus, apparatus for adjusting the peak table to a first preselected standardized format, computing apparatus for comparing the so standardized peak table of the unknown with a first library of chemical structural units contained in memory in the computing apparatus, apparatus for making a list of the possible chemical structural units most closely corresponding to the unknown substance, apparatus for readjusting the peak table to a second preselected standardized format, apparatus for forming a file for the unknown substance including data corresponding to the readjusted peak table and to the list of possible chemical structural units, computing apparatus for comparing the file of the unknown substance with a second library containing files of known substances contained in memory in the computing apparatus, each known substance having a file containing data corresponding to its respective peak table in the second standardized format and to its respective chemical structural unit, and apparatus for outputting a list of known substances which most closely correspond to the unknown substance. In a form of the invention, apparatus is provided for adding data corresponding to a known substance to the second library contained in memory in the computing apparatus.