BOAS] HANDBOOK OF AMERICAN INDIAN LANGUAGES 95 



the open position maintained in breathing- to the position required for 

 the vowel, and is written h. It is rather stronger than English A. 

 When final, the spirant is caused b}^ the sudden opening of the glottis 

 without diminution of the force of the breath, and is written ('). It has 

 been noted only where it is most prominent, or where it diflferentiates 

 one word or word-element from another. One of the spirants (?) im- 

 parts a noticeable harshness to the Hupa language. It is formed 

 rather far back in the mouth, apparently in the velar position. The 

 mouth-passage is made quite narrow, and the uvula is thrown into vibra- 

 tion. The period of these vibrations is about forty per second. The 

 resulting sound is harsh, both from the lowness of the period and 

 from its irregularity. The degree of harshness varies considerably in 

 individuals, and, indeed, in the same individual. While the sound is 

 not far removed from the velar r in its place and manner of forma- 

 tion, its effect on the ear is rather that of German cA after back vowels. 

 In Hupa, however, this sound is usually initial. There is no correspond- 

 ing sonant in Hupa. It does occur in Navaho; as, for example, in 

 the proper pronunciation of hogari house, where the first consonant is 

 nearly like the Hupa sound, and the second is its sonant. There is a 

 spirant pronounced in the palatal position, but accompanied by marked 

 labial rounding. It closely resembles wj but it is a surd, not a sonant. 

 When this sound is initial {hw), it appears to begin without rounding 

 of the lips, sounding much like English wh in who. When final (w), 

 the sound makes much less impression on the ear. It is to be distin- 

 guished from X by its lack of roughness, and from both x and h by the 

 rounding of the lips. It differs from a bilabial /in that it is accom- 

 panied by a narrowing of the mouth-passage in the palatal position. ^ 

 Another spirant {£) common on the Pacific coast, and found in Hupa, 

 causes great difficulty when first heard. It is formed at one or both 

 sides of the tongue, as is I, but differs from that sound in that the 

 breath which passes through the opening is surd instead of sonant, 

 and that the passage is narrower, causing a distinct spirant character. 

 When the passage is entirely closed and the breath must break its way 

 through to continue as a spirant, an affricative L is formed. Both of 

 these sounds, but especially the latter, impress the ear of one unac- 

 customed to them as combinations of t ov h and l. The spirant s in 

 the alveolar position is frequent in Hupa, and does not differ espe- 



iThis sound has for its equivalent in other dialects c (sh). Cf. Hupa hwa sun and hme i, me, with 

 Kato ca and ci. 



§2 



