theta in Greek; [we may adopt th to express this in English.] as th in thus; [a d with a dot under it may be used to express this letter.] like the n in nut; & like the Р in papa; as bin but; as m in mug; as y in yard; is sounded like r in run; & like lin lunch; & like v in vulture; [we see no objection to v or w being indiscriminately used to represent this letter.] like s in sun; ∞ like h in hunt; sounds the same as C, but it is said that where greater force than usual is intended to be given to this sound, e is used, which is a lingual, whereas the is a dental [this is an unnecessary distinction]; and o has the sound of n in the French termination mon, and may be expressed in English by an n with a dot under it.

The above consonants may be divided thus;

Of the last 8 characters, may be included in the 1st class, in the 2nd, o in the 3rd, es and in the 4th, & in the 5th, and the o is a nasal. We have adopted the above classification in conformity with a dictum of the Commentator to the Sidath' Sangarawa, who in reference to the union letters, , and, says, that they are formed by the coalition of the letters, 4, and 2 with the last of the cerebrals, dentals, and labials respectively. Vide infra, note 4.

Before concluding we may notice a difficulty to which allusion is made by the Rev. B. Clough, in the Introduction to his Dictionary. He says

"The want of an acknowledged standard of spelling has, notwithstanding the perfection of the alphabet, created a

,, and (which may be expressed in English by "g, nd, nd, mb) are compound letters in which two sounds are melted into one sound.

degree of orthographical irregularity and confusion which will require much attention effectually to correct. Almost every writer seems to have adopted a system of his own, having been solely guided by the manner in which the sound of the word struck his ear. But in a language and alphabet like the Singhalese, in which there is so nice a discrimination of sound distinguishing words of totally different meanings, such a practice could not fail to prove most fatal to correct spelling; hence, in familiar correspondence especially, scarcely two persons will be found to spell alike. These irregularities originate chiefly in the misapplication of the bindu o; the five nasals, or te,, with their corresponding symbols; the three sibilants, and ~, and the symbols and substituted for the letter 6."

and

The student will perceive that these anomalies proceed from an inattention to the derivation of words, and the different powers of the constituent parts of some of the compound letters; e. g.

is compounded of

and ; and where its full sound

is lost in compounding words, people are apt to substitute

or o for; as

compounded are written මේ එල්,

lotus,' and

'flowers,' when

නෙළු මල, or නෙ

for I;' and

; but the last alone is correct.

is sometimes written for ; as

if we look to the root of this word we must at once perceive that

is here incorrect.

is compounded of o and ; and therefore the o alone must be used when this letter loses in composition its compound sound; as ∞ becomes mo 'river-water,' and not ගන්දිය.

in composition frequently leaves merely the with which is compounded; as q 'to break,'

'broken.'

It would therefore be incorrect to use o or, to express the න in බූන්.

is compounded of and ; and alone should

therefore be used where it is alone retained in composition, or where it is mute before; as a pandit. '

A distinction exists between and. The first is a dental, the second a cerebral. In composition where the sound of n occurs together with a dental in the same syllable, the former is preferable, and when it occurs in conjunction with a cerebral the latter should be employed. But where n is associated with any other organ-letters the student may be guided by the usage of Sanscrit writers.

The distinction between and can only be learned by an attention to usage; although we may remark that the latter is more frequently used with gutturals.

With reference to @, mę, bę, ∞, ∞ and, see our remarks in the Introduction, p. lxiii.

2 is the symbol for the suppressed sound of ∞.

on the top of a letter, is a Sanscrit symbol for the suppressed, and it is used for the sake of brevity, as 'doctrine.'

The other difficulties attending a correct spelling in Singhalese, are easily obviated by an attention to the powers of the letters; see Introduction: also chapter I, and notes at p. p. 17, 25, &c. &c.

Note 4.-page 2.

The author of the Elu Prosody says in the beginning of his work, that the letters,, c, and, are respectively formed by the coalition of the simple consonants ∞,,, and, with the final letter of each of the classes to which they respectively belong. That is to say; dividing the consonants into 5 classes as in Sanscrit (see Wilson's Grammar, p. 2), and the guttural & being coalesced with the final nasal of its species, produces, thus ganga 'river;' the cerebral being united with the final nasal of its species, produces, thus dandu 'sticks;' the dental & being united with the final nasal of its species, produces ę, as

in e sanda 'moon;' and the labial being blended with the final nasal of its species, produces, as in q amba 'mango.' It would thus seem that is a letter proper to the Singhalese. This is however a mistake, arising from the circumstance of many Sanscrit and Pali characters being included in the Singhalese alphabet, and therefore from an adoption by Sanscrit scholars (see Vadankaripota) of the Sanscrit division of the consonants into Gutturals, Palatines, Cerebrals, Dentals, and Labials. The only Grammar extant in the Singhalese, and which labours to redeem the Singhalese alphabet from being improperly amalgamated with Sanscrit and Pali characters, is the Sidath' Sangarawa. Its author in shewing the 10 vowels and 20 consonants proper to the language, has proved that the 5 long vowels and the last consonant o are necessary characters in the Singhalese, and that the long must be considered distinct from the short vowels. He has shewn this necessity by producing examples; 1st, where the 5 long vowels are separately used, as in &, &, cé, de, and (see p. 2); and 2ndly, where they are inflected with consonants, as in eso, e,,, and oes. If the long vowel sounds occur in the language; and, moreover, they are also inflected with consonants by other signs; it appears but reasonable that the student should be informed of their formation. So with respect to the o, the grammarian has proved the existence of its sound both singly, and in union with other letters in the Singhalese language, and thence its necessity to be treated in the alphabet. When we consider, therefore, the urgent necessity there exists for such a course, especially in view of other systems of ancient grammars, which give in their alphabets long as well as short letters for the same radical sound (e. g. then in Greek); we may pronounce the grammarian's labours misapplied, but that we are told by his commentator, that this part of the grammar was written in reference to the opinion of certain philologers, (probably the writer of the Elu Prosody was one

amongst the number) that 'it was unnecessary to treat of the five long vowels, and the last consonant ..'

The commentator says [we here state the substance of his remarks], 'Grammarians think that the five long vowels are unnecessary to be treated as separate characters, because they are inherent in, and are produced from the 5 short; and that in a manner similar to the formation of q, q, &c, (vide supra) the co is formed by a union of the letter ∞ with the last guttural. Now it is to be observed;—1st, that by giving the genus as in other instances, the species would not here be indicated by giving the 5 short vowels alone; nor, in the grammars of Maghada and Sanscrit languages, &c., is the genus given in this respect to indicate the species. It is therefore desirable to shew the long vowels separately;-2ndly, to suppose that is formed by a union of and, is to suppose that exists in the Singhalese language, which the very disputants do not shew; because they give only 24 letters, i. e. the same characters that we have given, minus the 5 long vowels and the o. And if we include @ as a letter proper to the Singhalese, we shall be introducing a superfluous character, because its scund in a positive state is not found in our language like the cerebral, which by a union of produces ; or the dental, which by a union of produces, or the labial, which by a union of a produces. And for these reasons we affirm that the is formed [not by a union of and, but] by a coalition of 。 and hence, therefore, the necessity for the o being shewn as a separate character. And if it be objected that o is a mute, and that its sound does not occur except in its coalition with another character; our answer is simple: that the o as a mute alone occurs in divers systems of grammar. Furthermore, to employ, which is a sonant, and has a separate independent existence as a mute only (for we have seen that it is of no use in the Singhalese), would