interretialia: interretialia:Latin actually prefers -o/-onis. The majority of n-stems in -o behave
interretialia: interretialia: Latin actually prefers -o/-onis. The majority of n-stems in -o behave that way and their inflection was taken as the model whenever the question of “how does this word inflect?” was brought up. The n-stems in -o/-inis, comprise only a few small groups of nouns, and their inflection was not taken as a model for other, unrelated words. Look at the words we have: Apollo, -inis is a very early borrowing from Greek, and even then the genitive Apollinis stands for Apollonis (second o short), but even then we actually see Apollonis and Apolonis sometimes. Then there are the suffixes -tudo, -ago, -ugo, and -igo. There are some other words in -udo such as hirudo and testudo (where the -do may just be related to -tudo). There are fewer words have stems in -[vowel][consonant other than d or g or n]in-; homo and nemo are probably the only ones. Then there are the words with stems that end in -[consonant]in- such as cardo and turbo. One may balk at this vowel/consonant business, but it is important because nearly all of the words with the -[vowel][consonant]in- structure are using the suffixes -(t)udo,-ago, -ugo, and -igo. Homo (stem homin-) and ordo (stem ordin-) are not good analogs for a stem avocadin- because its stem ends in -[vowel]din-, not -[vowel]min- or -[two consonants]in-. Cupido and gravedo would be better analogs, but your problem then is the fact that no native Latin word ends in -ado and has a stem of -adin-. Since no actual Latin word declines like -ado/-adinis, and because the -o/-onis inflection is the tie-breaker, I cannot agree that the avocado/avocadinis inflection is the winner here. To the final paragraph I should add: “and because homo and ordo are not great analogs.” -- source link