We have had today for the first time a confrontation of the views of many astronomers about the use of ultraviolet spectroscopic and photometric data in the field of spectral classification.

Dr. Jaschek recalled us that much work has still to be done from the ground since only 14% of the stars up to the tenth visual magnitude have M.K. spectral types. The current accuracy is around one tenth of a spectral type and 0.6 of a luminosity class. Jaschek notes that the luminosity classification from spectral inspection is superior to what is obtained from photometric data, a fact which is not surprising if we recall that lines are much more sensitive to pressure effects than the continuum. However, the introduction of a third parameter seems necessary, although it is not clear what this parameter should be.