I completely agree with the previous answer. It is not true (as stated in the OP) that most appendices could be bona fide separate papers. Typically, when separate from the main paper they would be very hard to sell (they have no introduction, etc...). In most of the cases I am aware of, appendices are:
1) lemmas proved by someone who was not part of the main effort behind the paper, and which turn out to be important for some part of the paper (this is the typical example that would be rejected if separate from the paper: "it's just a lemma").
2) new approaches or alternative proofs which came up after a first version of the paper was circulated. After all, it's not that easy to publish papers entitled "A new proof of...", especially if the result you re-prove is not a time-honoured one.