This reference to a review:
a trendy new English cookbook devoted to the preparation of offcuts, snouts, rectii, marrow, and bladders of all description.contains a pseudo-Latin plural. What is evidently intended is the plural of rectum, which is properly recta. rectii would be correct if the stem ended in i, that is, were recti, and if its gender were not neuter, in which case the nominative singular would be rectius. It's interesting where these things come from. I suppose that people who don't actually know Latin but think that a word should have a Latin plural work by analogy from other Latin plurals they have heard. In this case, the analogy must be quite indirect, since no Latin noun ending in -um has a plural in -ii. (Such a noun would have to be a non-neuter ending in ium, which to my knowledge does not exist.)
It's easy enough to see how someone who doesn't know Latin could fail to realize that certain plural endings go with certain singular endings. That would account for someone deciding that the plural ending was i, not realizing that this was true only of masculine nouns, not neuters. So recti would not be a surprising error. But it looks like this form is derived by adding the ending ii to the stem rect. Where does this ii come from?
Some plurals do end in ii, but they are all plurals whose singulars end in ius, e.g. gladii "swords", singular gladius. Why don't people assume, as seems natural, that what is invariant, that is, gladi, is the stem, and that the plural ending is therefore just i? And how do they decide that the stem of rectum is rect? My best guess is that they have come up with a generalization about the form of the case/number endings, namely that they consist of one or more vowels possibly followed by a consonant. This is true of nominatives singular and plural of all nouns other than some third declension consonant stems. Someone who doesn't actually know Latin will generally encounter nouns in the nominative case, so that much is plausible. What I can't quite explain is where they get the idea that the ending can contain more than one vowel. Maybe they pronounce ii and i the same and so treat both as a single vowel, leading them to treat ii as a single morpheme and forcing them to conclude that the ius of words like gladius is also a single morpheme.
[Update: A convenient summary of Latin declension and conjugation is available on-line here.]