As far as I know, pervasion can be expressed more or less three ways in Sanskrit:

1. Iteration (āmreḍana): vṛkṣaṃ vṛkṣaṃ siñcati

2. Avyayībhāva compound: with the particles prati, pari and anu when compounded as the first members, as in pratyartham, pratyekam, pratidinam, anudinam etc. 

3. The secondary suffix (taddhita) śas: This suffix is used with numerals and words in the singular number. Such as ekaśaḥ (one by one), pañcaśaḥ (five by five), varṇaśaḥ (to each Varṇa). The words with the suffix śas behave like adverbs. 

The Pāṇinian sūtra for karmapravacanīyas prati, pari and anu (lakṣaṇetthambhūtākhyānabhāgavīpsāsu pratiparyanavaḥ 1.4.90) also says that these adnominals can be used in these senses, one of which is vīpsā (pervasion):

I looked into the commentaries and found that Jayāditya says these meanings are expressed through the context (viṣayabhūta) and the function of these adnominals (karmapravacanīya) is to establish a relation. This example, vṛkṣaṃ vṛkṣaṃ prati siñcati was given by Jayāditya in the Kāśikā. Jinendrabuddhi says that the pervasion (vīpsā) is already expressed by iteration (āmreḍana or dvirukti), the adnominal (karmapravacanīya) prati is to express the relationship between the action (watering) and the trees. But then it goes against the fundamental definition of adnominals (connected to nominals) and behaves like an upasarga (preverb) [upasargāẖ kriyāyoge]. But then the commentators leave it as a special case saying that although it is connected to the action we won't call it an upasarga because upasargas would involve other kinds of changes. Had it been the upasarga it would have been vṛkṣaṃ vṛkṣaṃ pratiṣiñcati. The non-retroflexion of the sibilant here allows us to consider it a karmapravacanīya and not an upasarga.

I think that Pāṇini made things a bit complicated here stating all these 4 meanings of the adnominal prati, pari and anu. All these we can translate with the English preposition 'towards' or 'to'. But I think when Pāṇini thought of a directional use, he thought of physical objects. In the case of the example, sādhur devadatto mātaram prati (Devadatta is good to his mother) we cannot perceive a physical direction. So, he postulated these meanings and later commentators had trouble in explaining those. So, I would translate vṛkṣaṃ vṛkṣaṃ siñcati as "s/he waters every tree" and vṛkṣaṃ vṛkṣaṃ prati siñcati as "s/he waters towards every tree" (if this makes sense at all). 

I am wondering if there are secondary materials available on these issues. I checked Joshi and Roodbergen's and Rama Nath Sharma's translations of the Aṣṭādhyāyī. There I did not find anything more than what Sanskrit commentators said. But it would be great to find some new researches on this issue. 

Deepro