This report discusses the results of a series of tests specific to the Thermal Neutron Analysis (TNA) detector being used by Computing Devices Canada (CDC) during the Vehicular Mounted Mine Detector (VMMD) ATD demonstrations conducted at Aberdeen Proving Ground (APG), Maryland, in June 1998 and at Socorro, New Mexico, in July 1998. A TNA-specific test plan was devised to address performance issues in a thorough and systematic manner; unfortunately, there were sufficient constraints such that the test plan was not implemented as designed. Instead, a much-abridged version was conducted at APG, and a still more limited version at Socorro. In addition to performance evaluation for both sites, we present detailed analysis of the APG data, where we found that the detectability of a mine did not seem to depend on either its burial depth or its nitrogen content. Furthermore, target signature variability dominated the test results, rather than background variability. Surprising results such as these must be treated with caution, however, given the extremely small data set that was available. We believe that it is in the Army's interest to pursue these issues with a more thorough test, such as the one originally proposed which appears in the Appendix of this report.