25. DISCERNERE. Gronovius argued (Obseruationes III xiii) that DECERNERE (MI1) should be read here, since decernere has the required sense 'uel decertare uel iudicare et certum statuere', whereas discernere means 'separare, dirimere, distinguere, diuidere'. On the evidence of the lexica, however, Gronovius' distinction breaks down, since discernere meaning 'decide, determine, make out' is common enough: compare Sallust Cat 25 3 'pecuniae an famae minus parceret haud facile discerneres', Cic Rep 2 6 'ne nota quidem ulla pacatus an hostis sit discerni ac iudicari potest', Varro LL VII 17 'quo discernitur homo mas an femina sit', and Livy XXII 61 10 'quid ueri sit discernere'. I therefore let discernere stand.
29-30. ET PVDET ET METVO SEMPERQVE EADEMQVE PRECARI / NE SVBEANT ANIMO TAEDIA IVSTA TVO. Compare EP III vii entire (an apology to his friends for the monotony of his verse), and especially the opening lines: 'Verba mihi desunt eadem tam saepe roganti, / iamque pudet uanas fine carere preces. / taedia consimili fieri de carmine uobis, / quidque petam cunctos edidicisse reor'.
30. SVBEANT ANIMO. Subire animo occurs also at Tr I v 13. Ovid uses subire with the dative several times in the poetry of exile (Tr I vii 9, II 147, III iii 14 & V vii 58; EP I ix 11, II x 43 & IV iv 47), but not beforehand; earlier he has the accusative (Met XII 472) or the simple verb (Met XV 307). The dative construction is taken up by the author of the later Heroides (XVI 99, XVIII 62).
31. RES IMMODERATA CVPIDO EST. Cupido similarly called immoderata at Apuleius Plat II 21; elsewhere qualified as immodica (Livy VI 35 6) and immensa (Aen VI 823, Tac Ann XII 7).
33. DELABOR. Cicero uses the word for moving from one subject to another (OLD delabor 5b); here the metaphorical sense 'fall' is still active.
34. IPSA LOCVM PER SE LITTERA NOSTRA ROGAT. This line as it stands is clearly corrupt. I do not understand Wheeler's 'my very letters of their own accord seek the opportunity'; André's 'c'est la lettre qui, d'elle-meme, demande le sujet' seems equally difficult, although locus can certainly have the meaning 'subject, topic of discussion' (OLD locus 24b).
The only parallel I have found is Fast II 861 'iure uenis, Gradiue: locum tua tempora poscunt'. If littera is retained in the present passage, this parallel is of little assistance, since locum there means 'a place within a larger work', and Ovid's poetry cannot ask for a locus in that sense. Taking the passage from the Fasti as a parallel, I once thought that Ovid wrote ipsa locum pro se tristia nostra rogant (or petunt); for the noun triste compare Fast VI 463 'scilicet interdum miscentur tristia laetis', Ecl III 80-81 'triste lupus stabulis, maturis frugibus imbres, / arboribus uenti, nobis Amaryllidis irae', and Hor Carm I xvi 25-26 'nunc ego mitibus / mutare quaero tristia'. I now consider this unlikely, since the personal adjective nostra with tristia seems unidiomatic; but I still believe that littera is the key to the corruption.
Professor R. J. Tarrant has tentatively suggested something like inque locum ... redit, but questions whether in locum, even just after eodem, can have the sense in eundem locum. Professor Tarrant also points out to me the possible relevance of locus in the sense locus communis (compare Sen Suas I 9 'dixit ... locum de uarietate fortunae'); Ovid might be saying that his poetry had made rather frequent use of the locus de exilio. In this case, rogat would require emendation.
One of Heinsius' manuscripts read per se ... facit, which is just possibly correct. Heinsius proposed pro se ... facit, which I do not understand.
35. HABITVRA is a good instance of the future participle used to express what is inevitably destined to happen (with Parca balancing in the pentameter); for the sense, see Tarrant on Sen Ag 43 'daturus coniugi iugulum suae'.