Introduction to Khuddakapāṭha
with an Analysis of its Metre


The Metre

In recent times we have gained much knowledge in regard to the correct form of the metres that are used in Pāli metrical composition, which is due mainly to the labours of Helmer Smith, A. K. Warder and K. R. Norman. In establishing a verse text it is, of course, essential that the parameters of the prosody are understood.

As I have stated elsewhere See my An Outline of the Metres in the Pali Canon elsewhere on this website.01 it seems possible to identify three phases of canonical Pāli verse composition, which for convenience we may designate the early, the middle, and the late. It may be stated here that although canonical Pāli metrical composition stretches over a period of several centuries, it seems that the first two periods described below have to fall within the lifetime of the Buddha.02 Briefly, the early period concerns the two main metres used in Pāli, the Siloka and the Tuṭṭhubha. The Siloka in the early period is characterised by the regular inclusion of the Anuṭṭhubha variation in the prior lines. In the middle and late periods this occurs only sporadically (and can nearly always be ‘corrected’ to the pathyā, or normal cadence, which makes one believe that the current readings may simply be corruptions). In the late period the pathyā predominates over the other variations to a marked degree, sometimes reaching as much as 85% See Warder, Pāli Metre (London 1967)( = PM) p. 19803.

The Tuṭṭhubha in the early period is normally used as an independent metre, without admixture of Jagatī lines, which occur only very occasionally. In Aṭṭhakavagga of Suttanipāta, for instance, there are only 4 Jagatī lines among 99 vs of Tuṭṭhubha (there is also one Jagatī verse, no 836 in Ee). In Pārāyanavagga, the Jagatī lines amount to approx 7% of the lines in the Tuṭṭhubha verses.04 In the middle period mixing is not only common, but normal. In the later period, the Tuṭṭhubha becomes restricted to the classical Upajāti form, and Jagatī to Vaṁsaṭṭhā.

The middle period also saw the emergence of the so-called new metres, the mattāchandas and gaṇacchandas. In Mettasuttaṁ, which appears as the last of the texts in this collection, we are dealing with what is probably a transitional metre between these two, the Old Gīti. This is the metre referred to by Alsdorf in his monograph Die ĀryāStrophen des PaliKanon (Mainz 1967) as Old Āryā. Norman in Group of Discourses II (Oxford 1992)(= GD II) also used this name, but later in his essay on The Origins of the Āryā Metre in Collected Papers Vol 4 (Oxford 1993)(= CP), preferred the name Old Gīti. The latter seems in every way preferable, as the structure of the metre is in fact a primitive form of Gīti, which has the same pādayuga structure repeated to make up a verse, whereas Āryā has two different pādayugas to the verse.05 The structure of the metre is rather primitive and unsettled, as will be seen from the description that follows. Towards the close of the later period both of these type of metres were superceded by their fixed classical counterparts.

On the basis of this description we can fairly confidently It may be stated here that these periods are by no means hard and fast, but run over into each other as we might expect from an evolving culture.06 ascribe the Managlasuttaṁ, Ratanasuttaṁ, and Mettasuttaṁ to the middle period; while the Tirokuḍḍasuttaṁ and the Nidhikaṇḍasuttaṁ belong to the late period.


The conventions used in this paper are as follows:

Sarabhatti (partial vowels which do not count metrically) are represented by the sarabhatti vowel being written in superscript e.g. from Maṅgalasuttaṁ (10a): tapo ca brahmacariyañ-ca.

In the analysis of the metre:

= light syllable; = heavy syllable; = light or heavy; = 2 shorts or one heavy; = 2 shorts or one heavy or one short.

Resolved syllables are underlined e.g. from Nidhikaṇḍasuttam (15ab):

⏑⏑−⏑−¦⏑−−−¦¦−⏑−⏑¦⏑−⏑− pathyā
paṭisambhidā, vimokkhā ca, ~ yā ca sāvakapāramī,

As in the above example the variation (where appropriate) has been identified and indicated. I have been greatly helped in compiling the notes that accompany this edition by Helmer Smith's metrical analysis of the 3 suttas that also appear in Suttanipāta (in Paramatthajotikā Vol III, 1918, reprinted Oxford 1997); and for the same suttas by K R Norman's comprehensive notes in GD II.07

The analysis of the metres in the text is based on the following description:

1: Siloka (Maṅgalasuttaṁ, Tirokuḍḍasuttaṁ, Nidhikaṇḍasuttaṁ)

Here is an analysis of the pathyā (normal) structure of the Siloka:

Odd line:


Even line:

⏓⏓⏓¦⏑−⏑⏓ x 2

In the 2nd & 3rd positions two light syllables are normally avoided.

In the first half of the line 7 variations (vipulā) occur, besides the normal structure, they are:














−⏓−¦−−⏑⏓ ~ (very sporodic)


2: Tuṭṭhubhajagatī (Ratanasuttaṁ)


−−¦⏓⏑⏓¦−⏑−(⏑)⏓   x 4

The normal opening is ⏓−⏑−, but occasionally we find ⏓−−−

The normal break is the bhagaṇa −⏑⏑, but others also occur e.g. −⏑−, ⏑⏑⏑, etc.


3: Upajāti (Tirokuḍḍasuttaṁ vs 13)


⏓−⏑−¦−⏑⏑¦−⏑−(⏑)⏓   x 4


4: Old Gīti (Mettasuttaṁ) This description is based on Norman, The Origins of the Āryā Metre in CP Vol 4.08


⏔−¦⏑−⏑¦⏔−¦⏓,⏔¦−−¦⏑−⏑¦⏑⏑−¦⏓   x 2

Resolution occasionally produces different patterns e.g. −− > ⏑⏑−

Replacement sometimes produces different patterns e.g. ⏑−⏑ > −−

Note that −⏑⏑ is not normally found in any gaṇa.