Here’s the other half of that list that I started in (this post). These are the class of consonantal adjectives that tend not to operate as quasi-nouns, and therefore tend to take -i in the ablative, rather than -e.
āmēns, āmentis, frantic, crazed
anceps, ancipis, double, doubtful
concors, concordis, agreed, joint
dēgener, dēgeneris, low-born, weak
hebes, hebetis, dull, blunt
ingēns, ingentis, huge, vast
inops, inopis, needy, helpless
memor, memoris, mindful of
pār, paris, alike, equal to
perpes, perpetis, lasting
praeceps, praecipitis, headlong
praepes, praepitis, nimble, winged
teres, teretis, smooth
Sorry for the relative obscurity of these last two posts. The thing is, they cover a few footnotes in Allen and Greenough that I feel should be out there on the Internet, preferably with a basic entry format and definition.
Consider section 121a4, which lists a variety of consonant stem adjectives that do not take -i in the ablative singular. They are completely regular, and the entry is really there only to keep you from having second guesses.
I’ll list them here so they get some web mileage, despite not being especially interesting, however rare:
caeles, caelitis relating to the heavens or their Gods
compos, compotis possessing control of
dēses, dēsidis lazy
dīves, dīvitis wealthy
hospes, hospitis amicable, relating to guest-friendship
particeps, participis participating in
praepes, praepitis nimble, winged
pauper, pauperis poor, destitute
prīnceps, prīncipis princely, noble
sōspes, sospitis safe and sound
superstes, superstitis surviving
As someone pointed out in a comment, the general but non-binding idea is that those adjectives which most often operate as nouns (like these) take the ablative in -e, whereas those that are properly adjectival take the ablative in -i. You find hints of this throughout the consonantal and i-stem entries in Allen and Greenough (see 121a1-2), but they make no effort to propagate it as a formal rule.
So, unfortunately, I stated earlier that the neuter i-stem class is entirely regular. In fact, a footnote in A&G reveals about ten exceptions.
These nouns are almost regular, except that with consonantal stems –al and –ar they also add the ending -e to the nominative and accusative singular. Note that because of this ending, the -ā-is long in all cases.
Where the singular is uncommon or does not exist, I have used the plural.
A&G define the vocative as “the case of Direct Address.” (35f)
Generally speaking, the vocative and the nominative are the same.
However, in certain nouns of the second declension (those with nominative -us or -ius) have two exceptional variations. All nouns in -us feature an -e in the vocative (mūrus…mūre). Those ending in -ius (Vergilius, fīlius, genius, etc.) take a vocative -ī (Vergilī, filī, genī).
[Highly attentive readers should note that this vocative does not shift its accent, rendering Vergílī, and not *Vérgilī, as one might expect.]
That’s how it stands for nouns. There’s a slight variation in policy for adjectives, though luckily the same general rule (same as the nominative) holds true for all but the second declension (bonus…bone). However, the one catch is that adjectives ending in -ius change to -ie and not -ī. Therefore, when calling to a Spartan son, we might say O fīlī Lacedaemonie! (not *Lacedaemonī).
If anyone has a better understanding of vocative plurals, which I assume are all identical to their nominative forms, feel free to say more in the comments below. A&G are totally silent on this issue, which I assume signals that listing the vocatives would be redundant (with respect to the nominatives).