OF ARTS AND SCIENCES. 167 



" When a character has several sounds, it has a special and an ac- 

 cidental power, the former usually found in its alphabetic name. It 

 seems plain, that the accidental power should have a new or modified 

 character, and not those which have always been written and recog- 

 nized. Mr. Ellis assigns to the vowel O the character w closed at the 

 top. He should then, at least, have supplied that in not with a modi- 

 fied character. A character formed like the Greek a (the Latin u) is 

 perverted to a diphthongal power, as if to justify and perpetuate a 

 false pronunciation of Greek. 



"Mr. Ellis (in conjunction with Mr. Pitman) has proposed, it is said, 

 fifteen alphabets, in which there is a gradual deterioration, the last 

 being the worst. There is a certain relation between the primary 

 vowel in meal and the secondary one in viill, which was recognized 

 by Mr. Ellis in 1844, when the former was represented by I with a 

 horizontal medial line, and the latter by I. The related vowels in 

 dale, dell, were represented by E with the line continued across, and 

 E without a medial line. Now, meal and mill are spelt ' msl, mil ' ; 

 whilst dale and dell stand ' dal, del'; and A, the capital of 'a' is 

 reserved for the rare Welsh vowel in fat. Having reached its lowest 

 point of deterioration, this alphabet ' is brought to a satisfactory con- 

 clusion,' and fault is found with those who will not adopt the later 

 changes, termed 'improvements' (p. 220, and Phon. Journ.), When 

 the pure vowel in meal is short (without falling into that of mill), i is 

 employed ; which is correct, but inconsistent. 



"Mr. Ellis's ethnical alphabet contains 56 characters, including a 

 with a line through it, which is omitted in the table on p. 126. Some 

 of his analyses are very minute, as the 'middle sound' (pp. 3, 7) 

 between the consonant and vowel of see. On the other hand, his 

 ideas of the relation between the open (and usually long) vowels in 

 paio* fur, pool, lo, and their close (and usually short) condition in 

 naught* worth, full, obey, are very confused. At present, he makes 



c had " undoubtedly several sounds" in Latin. He should have informed his 

 unclassical readers, that in these words, according to the ancient grammarians, the 

 I and u had not tfieir true poioer, but an allied one, for which Claudius proposed a 

 character. Consult Velius Longus, Priscian, and Donatus. The power in ques- 

 tion was not the French u, as that was represented by Y. The aperture of the 

 I in fin is nearer that of U than of I. 



* To form the latter, a longer pipe is required than for the former, according to 

 the experiments of Wheatstone. Herschel (Encyc. Metrop.) confesses himself 

 unable "to detect any shade of difl'erence " between them. 



