Vaikhānasamantrapraśna, 5-8 · Chapter 5.1.16

imé bījā́ vrī́hi-yavā́ ruhantūṣṇā́ṃśu-prasādā́t
sómenā́rpitā grahaír nirvyā́dhikāś śuddhā́ś śubhā́ bhavantu

prároha-bījā́n nírvapāmi bhūmaú yé brahma-sr̥ṣṭā́ dáśa sapta caíva. té vardhitā vr̥ṣṭi-váśena sárve saṃpā́lyamānā hy amŕ̥tena lokā́n

saptá grāmyās sáptá cāraṇyakā́ś ca yé gharma-naṣṭā́ himá-varṣa-sr̥ṣṭā́s té vr̥ṣṭi-bhūtā́ hy amŕ̥taṃ rakṣantu

dīrṇā́rdra-saṃśuṣká-vivárdhamānā mū́laiḥ phalaíś śā́kais sámr̥ddha[*30]-rūpaíḥ
annéná martyānāṃ[*31] havíṣā surāṇā́n té vai ruhantā́ṃ tr̥ṇávat paśúbhyaḥ

tā́n bhūmi-śuddhyaí nírvapāmi bījā́n devá-sthiti-kā́raṇāya

yé vyoma-bhūmy-agni[*32]-jalásya vāyós sáṃyoga-bhūtā́ḥ práruhanti sárve
tā́n daivike śuddhi-kárāya bījā́n nírvapāmi devá-sthiti-kā́raṇāya
[*30] 1946 edition shows the incorrect reading pamr̥ddhaiḥ which was corrected in the 1984 edition. In Telegu script, pa and sa are very similar.
[*31] 1946 text reads martyānṃ. The 1984 text reads martyān, assuming that what appears to be a second r is actually a symbol for final n with one loop broken off. Otherwise as it stands, the 1984 edition would read marrtyā. I have corrected all these unintelligible forms to martyānāṃ, which combined with annena provides an intelligible reading parallel to haviṣā surāṇān.
[*32] The text places an anudatta under the ū of bhūmy and then a svarita over the a in agni, seemingly an impossible feat of recitation.