Aims. We study a microflare, classified as a GOES-A1 after background subtraction, which was observed in active region NOAA 8541 on May 15, 1999. Methods. We used TRACE filtergrams to study the morphology and time evolution. SUMER spectral lines were used to diagnose the chromospheric plasma (Si ii1533 A), transition region plasma (C iv 1548, 1550 A), and coronal plasma (Ne viii 770 A). Results. In the 171 A and 195 A filtergrams, we measure apparent mass motions along two small loops that compose the microflare from the eastern toward the western footpoints. In SUMER, the microflare is detected as a small (47 Mm 2 ), bright area at the western footpoints of the TRACE loops. The spectral profiles recorded over the bright area are complex. The Si ii1533 A line is self-reversed owing to opacity, and the coronal Ne viii line profile is composed of two Gaussian components, one of them systematically redshifted. The C iv1548 A and 1550 A profiles are badly distorted because of the temporary depression of the detector local gain caused by the very high count rates reached in the flaring region and we can only confirm the presence of strong blueshifts of � ‐200 km s −1 . Few, unaffected C iv profiles show two spectral components. In the northern part of the bright area, all SUMER spectral lines have at least one blueshifted spectral component. In the southern region of the bright area the spectral lines are redshifted. Adjacent to the microflare we measure, for the first time on the solar disk, an intensity ratio of the 1548 A line to 1550 A line with values of three to four, indicating that resonance scattering prevails in the lines formation. Moreover, the scattering region is found to be cospatial to a solar pore. Conclusions. The blueshifts in the footpoints of the microflare and the apparent mass motions observed with TRACE can be explained by a gentle chromospheric evaporation triggered by the microflare. The redshifted spectral components can be explained as cooling material that is falling back on the solar surface. The presence of resonant scattering can be explained by the low electron density expected in the transition region of a solar pore, combined with the high photon flux coming from the nearby microflare. We estimate that the lower limit of the electron density in the pore lies in the range 10 8 cm −3 to 10 9 cm −3 .