^{page 149 note 1} See Techmer, Internationale Zeitschrift, i, 1, Tab. iv.

^{page 149 note 2} See an article by Prof. Sheldon and myself, called ‘Phonetic Compensations,’ in Mod. Lang. Notes, iii, 6. This kind of compensation is, I think, illustrated in the chart that accompanies Dr. Techmer's pamphlet ‘Zur Veranschaulichung der Lautbildung’ (Barth, Leipzig, 1885): if I remember his pronunciation rightly, the author forms a, as I do, with the tongue lying nearly flat in the bottom of the mouth; but in the drawing, which represents a man uttering a with his mouth stretched open to its widest extent, the middle of the tongue is violently raised, evidently to compensate for the enlargement of the mouth-cavity through unnatural jaw-lowering. Similar compensations are to be noted in P honetische Studien, ii, 2, ‘On the Bell Vowel-System.

^{page 149 note 3.} How far a really good observer may be led astray by “sensation” is sadly apparent in some parts of the article ‘On the Bell Vowel-System,’ by the late W. R. Evans, Phouetische Studien, ii, 1.

^{page 149 note 4} See ‘Physiologie der menschlichen Sprache,’ 1866, pp. 68, 82, 85, 86, 89, 91, 93, 98, 103. See also, however, Vietor, ‘Phonetik,’ 1887, p. 36; and Brücke, ‘Grundzüge der Physiologie und Systematik der Sprachlaute,’ 1876, pp. 37, 38.

^{page 149 note 5} In Phonetische Studien, iii, p. 114, Swert says: “The only observations that can bé fully relied on are those made by trained observes on themselves.”

^{page 149 note 6} The “long” and “short” marks are used in this article merely to distinguish different vowel-qualities: they have no reference to quantity. The correspondence of my symbols with those used by the American Dialect Society is as follows: my u— Am. Dial. Soc. u, u—u, o=o, a=o, o—, u=e, ë—ë. on=o, a=a, e=, i=i, i—i, é=ě, en=e, an= e.

^{page 149 note 7} See Swert, ‘Primer of Phonetics', 1890, p. 75.

^{page 149 note 8} See Phonetische Studien, i, 2, p. 171 7 and Swert, ‘Primer of Phonetics,’ 1890, pp. 76 and 85.

^{page 149 note 9} Merkel ('Physiologie der menschlichen Sprache', p. 103) notes a very decided rise and fall of the larynx. Techmer (Internationale Zeitschrift, i, 1, Tab. iii) indicates something similar.

^{page 149 note 10} See Mod. Lang. Notes, iii, 6, p. 364.

^{page 149 note 11} See end of this article.

^{page 149 note 12} Merkel ('Physiologie der menschlichen Sprache', p. 103) makes the difference between t and a. Passy (Phonetische Studien, i, 1, p. 24) gives diagrams which seem to indicate a little more jaw-lowering. Western ('Englische Lautlehre', 1885, pp. 5 and 83) implies that the distinction between i and a is almost entirely a matter of jaw-position.

^{page 149 note 13} In his introduction to the Revue des Patois gallo-romans, i, 1, the Abbé Rousselot says, in the course of an ‘Analyse des sons', in speaking of the vowels (p. 13): “Je ne tiendrai compte ici que des mouvements de la langue et de ceux des lèvres. L'explorateur que j'emploie est tout simplement le doigt.”

^{page 149 note 14} I have made use of a small glass bulb enclosing a one-candle-power incandescent burner connected by two thin wires with a three-cell battery.

^{page 149 note 15} The soft palate can readily be trained to take the proper positions. It is well to begin by watching its movements in natural speech, and then to try holding the tongue down with the finger and uttering the vowels mentally. Before long the tongue will stay down of its own accord, a nd the soft palate wilt move independently of it.

^{page 149 note 16} For the sake of greater accuracy (as these data are of the highest importance), we may make some supplementary measurements. Open the mouth wide; determine the exact position of a (Figure 2) with reference to b; then measure af. The positions of a and b, and the distances ab, af, and bf being known, we can find the exact location of f. Similarly we can, if necessary, calculate the positions of e and d.

^{page 149 note 17} Great care should be taken, especially in measuring i, é, and en, lest the card-board sink into the back of the tongue and thus indicate a false position. If this digging into the tongue cannot be avoided, some allowance must be made for it. Much care is required, also, to keep the oval perpendicular to the tongue.

^{page 149 note 18} To admit the finger the mouth must, of course, be opened wider than usual; but this jaw-lowering, which amounts to three-quarters of an inch at the teeth, is far less perceptible at the hack of the mouth.

^{page 149 note 19} My drawings appear to show a regular gradation from an to i and from a to u: nearly all German phoneticians have maintained that this was the case with their vowels.

^{page 149 note 20} See Jespersen, ‘Articulations of Speech Sounds', 1889, p. 17. Sweet himself says, ‘Primer of Phonetics', 1890, p. 18: “The distinction between narrow and wide is not so clear in the back vowels.”

^{page 149 note 31} Sweet: ‘Handbook of Phonetics', 1877, p. 16; ‘History of English Sounds', 1888, p. 3; ‘Primer of Phonetics', 1890, pp. 71, 72, 73. In the last work, p. 73, Sweet says of his “mid-back-narrow” u: “This vowel is slightly advanced.”

^{page 149 note 22} See Proceedings of the American Philological Society for 1884, pp. xxxviii-xl.

^{page 149 note 23} With persons who round the a, the mouth-cavity for that vowel is probably somewhat smaller than with me. Bell, however, says ('Speech Reading and Articulation Teaching', 1890, p. 13): “Enlarge the cavity of the mouth to the utmost …. Emitted … voice will then have the quality of what is called the ‘Low Back’ vowel”. I do not see how Vietor can say ('Phonetik', 1887, p. 15): “Bei u ist der Resonanzraum im Munde am grössten.” It seems to me that no back vowel can have a smaller cavity than ü. The low pitch of this vowel is evidently caused by rounding. Cf. Sweet, ‘Primer of Phonetics', 1890, p. 26.

^{page 149 note 24} I ought, perhaps, to say that I made hundreds and hundreds of preparatory measurements before I thought myself sufficiently skilled to begin on the final experiments, the results of which are set forth in this article. The ticklishness of the soft palate, which, at first, is apt to produce choking and retching, can easily be overcome by a little practice; but the sensitiveness of the pharynx, which, if exploration in that region be long continued, is liable to develop into sore throat and coughing, I have never been able to cure. The difference in the effect of contact on the parts touched sometimes affords a clue to the whereabouts of the end of the exploring finger, when that member is not sensitive enough to distinguish, by its own sensation, the soft palate from the inner wall of the pharynx.