Tests have been performed on the optical, resonance and levels segments. A number of misprints and erroneous coding have been detected and corrected.

Several RIPL participants tested the preliminary version of the levels database by using the data in calculations. A new simple test was worked out for checking nuclear temperature (T) derived from the analysis of cumu­lative plots of discrete levels, yielding temperature values which are remark­ably similar to the T(A) function obtained in the global fitting procedure. Ignatyuk tested the performance of the T(A) function by comparing the results with the temperature obtained by Gilbert and Cameron; he found reasonable agreement and recommended the use of T(A) in cases for which no direct estimation is possible. Herman has extensively tested Nmax values for nearly 500 nuclei using the Gilbert-Cameron procedure and level densi­ties specific to the EMPIRE code. Perfect fits were obtained for about 50% of all analyzed cases, fair agreement was found for about 25%, and poor for the remaining 25%. The quality of the fit depends on the model used for level densities. No formatting errors were detected while reading files with discrete levels.

Global testing of the RIPL-2 database has been performed in three sep­arate exercises. Large numbers of nuclear reaction cross sections were calcu­lated by means of the nuclear model codes EMPIRE-II, UNF and TALYS. Herman performed calculations for the most important neutron-induced re­actions on 22 targets from 40Ca up to 208Pb in the energy range from 1 keV up to 20 MeV. The 2-17-beta version of the statistical model code EMPIRE — II has been used with all default parameters except those differentiating the 3 series of runs. In all cases TUL MSD and Heidelberg MSC models were used for pre-equilibrium emission of neutrons, and exciton model (DEGAS) for pre-equilibrium emission of protons and ys. These studies were comple­mented with Hauser-Feshbach calculations including widths fluctuations at incident energies below 5 MeV (HRTW model). The results were converted into ENDF-6 format and compared with experimental data available from the EXFOR library. Three sets of calculations were performed in order to test new levels segment, Koning’s global optical potential and HF-BCS level densities. No problems were encountered while processing the new RIPL-2 files, which indicates that the files are formally correct. Comparison with experimental data shows reasonable overall agreement for most of the calcu­lations. There is a clear indication that calculations using the new RIPL-2 files fit experimental data better than those with default EMPIRE-II param­eters, which demonstrates the improvements brought about by RIPL-2. The HF-BCS microscopic level densities were found to perform comparably to the phenomenological level densities and in some cases even better. However, significant discrepancies among the results of the three sets of calculations were observed in a number of cases. These findings illustrate the importance of the model parameters and prove the practical usefulness of the RIPL-2 library for basic research and applications. The second exercise was carried out by the Beijing group, using the recently developed UNF code to study 103 nuclei from the mass region 69-160 in the incident energy range from

0. 1 to 20 MeV. All input parameters were taken from the RIPL database. Agreement with the experimental data was found to be very good for total and elastic cross sections (within 3%). For other main reaction channels, calculations reproduced the shape, but some parameter adjustments were necessary in order to fit the absolute cross sections. TALYS calculations were performed for various neutron-induced reactions on 5 isotopes from 52Cr to 208Pb. Default input parameters originated from RIPL-2. This exercise concentrated on the comparison of Ignatyuk-type and microscopic level densities and provided very reasonable agreement with experimental data for both formulations.