21. BEFORE quitting the subject of the measures of sensible qualities, we may observe that there are several other such qualities for which it would be necessary to have scales and means of measuring, in order to make any approach to science on such subjects. This is true, for instance, of Tastes and Smells. Indeed some attempts have been made towards a classification of the Tastes of sapid substances, but these have not yet assumed any satisfactory or systematic character; and I am not aware that any instrument has been suggested for _measuring_ either the Flavour or the Odour of bodies which possess such qualities. 22. _Quality of Sounds._--The same is true of that kind of difference in sounds which is peculiarly termed their _Quality_; that character by which, for instance, the sound of a flute differs from that of a hautbois, when the note is the same; or a woman's voice from a boy's. 23. _Articulate Sounds._--There is also in sounds another difference, of which the nature is still obscure, but in reducing which to rule, and consequently to measure, some progress has nevertheless been made. {355} I speak of the differences of sound considered as _articulate_. Classifications of the sounds of the usual alphabets have been frequently proposed; for instance, that which arranges the _Consonants_ in the following groups,: Sharp. Flat. Sharp Aspirate. Flat Aspirate. Nasal. p b ph (_f_) bh (_v_) m k g (hard) kh gh ng t d th (sharp) th (flat) n s z sh zh It is easily perceived that the relations of the sounds in each of these horizontal lines are analogous; and accordingly the rules of derivation and modification of words in several languages proceed upon such analogies. In the same manner the _Vowels_ may be arranged in an order depending on their sound. But to make such arrangements fixed and indisputable, we ought to know the mechanism by which such modifications are caused. Instruments have been invented by which some of these sounds can be imitated; and if such instruments could be made to produce the above series of articulate sounds, by connected and regular processes, we should find, in the process, a _measure_ of the sound produced. This has been in a great degree effected for the Vowels by Professor Willis's artificial mode of imitating them. For he finds that if a musical reed be made to sound through a cylindrical pipe, we obtain by gradually lengthening the cylindrical pipe, the series of vowels I, E, A, O, U, with intermediate sounds[38\4]. In this instrument, then, the length of the pipe would determine the vowel, and might be used numerically to express it. Such an instrument so employed would be a measure of vowel quality, and might be called a _Phthongometer_. [Note 38\4: _Camb. Trans._ vol. iii. p. 239.]