Praudhamanorama p127

p127

"śarpare visarjanīyaḥ" (Pā. Sū. 8.3.35): This is an exception to the substitution of sa and so forth. But if it were stated merely as "[not] when followed by śar", then even when the contextual sa-substitution is prohibited, jihvāmūlīya and upadhmanīya would occur before ku and pu: vāsaḥ kṣaumam, adbhiḥ psātam. However, by explicitly mentioning "visarjanīyaḥ", every modification whatsoever is blocked; with this intention, he says: "na tv anyad iti" (but not anything else). He clarifies that very point: "iheti" (here).
On "so'padādau" (Pā. Sū. 8.3.38): On "apadādyoḥ": in the sūtra, however, the singular is used by transposition. If, however, khar is qualified as "before a khar that is not the beginning of a pada", then it is correct as read. But [in that case] "khari" has to be continued by frog-leap (maṇḍūkapluti), which is a hardship. Nor is continuation necessary in subsequent rules either, since without a khar a visarga is difficult to obtain, so its continuation would be purposeless. "pāśakalpakakāmyeṣviti vṛttiḥ": this is an illustration of what is possible, not an exhaustive enumeration, because nothing else is possible.
On "prātaḥkalpam": even of the word prātar, which primarily expresses the power of location, here in the domain of complex formation there is reference to that which possesses the power, just as in "doṣābhūtam ahaḥ" (the day which became night) and "divābhūtā rātriḥ" (the night which became day).
In "gīḥ kāmyati": here the substitution of ṣa by "iṇaḥ ṣaḥ" (Pā. Sū. 8.3.39) does not occur, because even there the phrase "kāmye ror eva" is continued. If, however, this were taught right there, then even upon the prohibition of ṣa-substitution alone, sa-substitution would occur by the previous rule.