A number of subjective experiments designed not only to assess the tonal and playing qualities of trombones but also to aid the interpretation of objective data is described. Quantifying multi-dimensional quantities such as timbre or quality is usually attempted either by multi-dimensional scaling (MDS) or by semantic differential scaling (SDS) and the advantages and disadvantages of each for the present purpose are briefly reviewed. For both methods additional information is needed either (for SDS) to aid the selection of scales, or (for MDS) to interpret the final results of the computer analysis; in this study such information is available from the investigations of Edwards . A preliminary experiment testing the ability of subjects to rate trombone timbre using SDS showed that the intersubject variance was very large but that the order of factors governing timbre is (i) instrument, (ii) player and (iii) mouthpiece. A second experiment shows that pitch and loudness have a much greater influence on listener's ratings than do player, mouthpiece of instrument. Players' attitudes tested in further experiments show that the task of discriminating instruments when extraneous cues are removed is very hard indeed for most players, but a very few can be of quite remarkable discriminatory ability using either SDS or MDS. Results from one such player indicate clearly that timbre is the predominant acoustic factor deciding discrimination.