At the outset of a pāda—whether it is the first pāda or one that follows another—certain rules guide how syllables are treated, especially in the presence of conjuncts and avagraha. Even at the beginning of a pāda, if there is no separation, the syllable remains at the end of a conjunct. When a short vowel remains, or when the letter ‘hṛ’ appears, or when an indistinct pair arises at the beginning, these are to be spoken in a low tone. Anything that is low remains low, and whatever is found in a cluster also stays low. There are specific words—‘priya’, ‘dūta’, ‘ghṛta’, ‘citta’, and ‘ati’—which are always pronounced in the low tone. Yet, among the hundreds and among the pure, the recitation rises from the low to the high tone. The syllable ‘viśvānara’ and the letter ‘o’ are pronounced with the svarita accent; the rest, O men, are spoken likewise. Wherever a double accent is applied, a short vowel should be pronounced as if it were long. The measure of time for a syllable, called a mātrā, is likened to the blink of an eye; some say it is like the flash of lightning. The accents of the Ṛk, or those with equal conjunction, are declared thus by some. The application of the accent begins from the initial syllable and continues to the end of the word. There are four kinds of openings to be understood—this is my opinion. Following a short vowel, and at the beginning, the sound that follows the ‘calf’ (the preceding element) is also at the front. Among these four openings, the interval should be measured by a mātrā. For some, it is half a mātrā; for others, an aṇu-mātrā, a still smaller unit. In the case of ‘pa’, the following letter is of the same class; for stops, the highest lord is applied. It is said that the letter ‘a’ is red, and with the letter ‘ta’, it is colored. The half-measure of the preceding syllable is extended by a further minimal measure. The ‘repha’, the ‘raṅga’, and elision, as well as sometimes nasalization, may occur. This sound is both soft and two measures in length, as exemplified by the word ‘dadhanvā’. Thus should ‘raṅga’ be applied—this is my view, O Nārada. Of the four types of metrical feet, ten kinds of word-endings are described. Consonants do not occur where only a vowel remains. One should understand the consonant as like a gem, and the vowel as the thread that holds it. In such cases, the consonant is weak, while the vowel is strong. The root of the tongue, the labial, and the eightfold movement of the sibilants should be known. There, the visarga must be considered, and the palatal is produced as well. The expansion is called ‘consonantal’, and close contact is called ‘vocalic’. Wherever there is expansion, it is designated as ‘vocalic’. One should know as ‘vowel-ending’ that which is of such a kind, except for the indistinct sibilants. If the ‘sha’, ‘ṣa’, or ‘sa’ affixes reach the second position, one should not acknowledge them, just as a fish does not notice a razor in the water. The Ṛg verses have free meter, and their feet are measured by the number of syllables. The affix, together with the ‘repha’, is measured by the division of the vowel. One should recognize a light ‘ṛ’ if it is joined with a sibilant. The guru syllable is to be understood; here, the example is the threefold verse. The fifth is Nirṛti; it is marked by the syllable ‘ṛ’, and there is no doubt about this. Sometimes, lacking both ‘i’ and ‘u’ vowels, a verse is separated from the sequence of single words. With the consideration of prescribed syllables, both vowel and consonant sounds are established. Among the ‘ṣa’ and ‘sa’ letters, the open form is known; with ‘ha’, the closed form is recognized.