Praudhamanorama p196

p196

On "tṛjvatkroṣṭuḥ" (Pāṇini 7.1.95): kruś- in the sense of calling. From this, with the affix tun by "sitani...", comes the word kroṣṭu; but with tṛc comes the word kroṣṭṛ. Both denote a jackal. There, when the use of both would indistinctly apply, the group of three sūtras is initiated for the sake of restriction: only the word ending in tṛc occurs before a sarvanāmasthāna and in the feminine; both occur before a vowel-initial case ending beginning with the instrumental; elsewhere, only the word ending in tun occurs. "kroṣṭuḥ" is the nominative singular from that which ends in tun. ["tṛc"] is the nominative singular from that which ends in tṛc. "tṛjvat" is formed with the affix vati from a word ending with the instrumental. Although transfer is of six kinds—having as its scope the cause, the designation, identity, the grammatical rule, the operational effect, or the form—here, on account of primacy, only the grammatical form is transferred. And that does not belong to any arbitrary root such as pac-, but to kruś- alone, because it is immediately present and closest in meaning; intending this, he states the resulting sense: "in the place of the word kroṣṭu".
"For the sake of restriction": it means: for the purpose that it should not apply to other proper names. "kroṣṭā": the substitute anaṅ. Lengthening of the penultimate takes place because it is subsequent to the elision by halṅyābilopa. And although that [lengthening] can be obtained by "sarvanāmasthāne cāsamvuddhau", still, being subsequent, "aptṛntṛc..." alone is proper. Even when anaṅ has been applied, it still possesses tṛc-hood, because that which is altered in one part is not different. Otherwise, even when guṇa has been made, lengthening would not take place—this is the point.
On "rātsasya" (Pāṇini 8.2.24): here, the letter t is also joined by coalescent elision—this is another view. By that, when the sūtra beginning with halṅyāp is simplified as "ṅyāpaḥ saḥ", forms like abibhar bhavān are established. But in ajarghart, because the substitution of a hard consonant is not yet established, elision of t does not occur. If it is objected: In the case of kvip after kīrtayati, would kīrt not result here? No; there is the Mahābhāṣya statement: "With respect to what is not in actual usage, [derivation proceeds] strictly according to the grammatical rule." By this, avarvart is explained.
"Because it is subsequent": and one should not say that the augment nuṭ is invariable; for once the treatment as ending in tṛc has occurred, nuṭ is hard to obtain on account of the metarule of harmonious union—this is the point.