एरनेकाचोऽसंयोगपूर्वस्य

A semi-vowel is substituted before an affix beginning with a vowel for the final of इ or ई of a root, not preceded by a conjunct consonant forming part of the root, when the अङ्ग (stem) is mono-syllabic.
The word धातोः is understood here, and the word संयोग is qualified by that: i. e. the इ or ई which is not preceded by conjunct consonant forming part of the root is called an asamyoga parva इ ॥ Thus निन्यतुः, निन्युः, उन्न्यौ, उन्न्यः, ग्रामण्यौ, ग्रामण्यः ॥ All the above examples are of the root नी preceded by the gati prepositions नि and उत्, or a Karaka-upapada ग्राम ॥ The rule will not apply however if the preceding word is neither a gati nor a Karaka, but an adjective, as परमनी, its dual and plural will be परमनियौ and परमनियः by इवङ् ॥ Why do we say 'of इ or ई'? Observe लुलुवतुः and लुलुवः from लुलू which ending in ऊ takes उवङ् substitution. This is also shown in the next sutra. Why do we say the stem should be of more than one syllable? Observe नी; its dual and plural are नियौ and नियः by इयङ् ॥ Why do we say the इ or ई should not be preceded by a conjunct consonant? Observe यवक्री dual यवक्रियौ plural यवक्रियः by इयङ् ॥ Why do we say "forming part of the root"? So that the rule may apply to उन्नी also. Here though ई is preceded by a conjunct न्न्, yet the latter is not part of the root, one न् being part of the upasarga उत् ॥ Thus we have उन्न्यौ and उन्न्यः ॥ The phrase असंयोगपुर्व should in fact be taken as qualifying the letter ई, and not as qualifying the word अङ्ग ॥