Going through the non-canonical readings collected by Ibn Ḫālawayh -- a student of the canonizers of the 7 readings, Ibn Mujāhid -- I ran into this unusual report: ʾAyyūb al-Saḫtiyānī read the last word of al-Fātiḥah as al-ḍaʾallīn, rather than the canonical al-ḍāllīn. 
Until recently I was convinced that this must have been a clear case of Hypercorrect hamzah use. Already in the 7th century there were varieties that lacked the hamzah, in recitation this would sometimes lead to recitation of words with hamzah that historical never had it.
Among the canonical readers there are a couple of examples of that: ʿĀṣim's yaʾjūj and maʾjūj. The latter must come from Hebrew מָגוֹג måḡoḡ, which never contained a hamzah.
Qunbul ʿan Ibn Kaṯīr reads the plural of ḍawʾ 'light' as ḍiʾāʾ rather than regular ḍiyāʾ.
Qunbul ʿan Ibn Kaṯīr reads the plural of ḍawʾ 'light' as ḍiʾāʾ rather than regular ḍiyāʾ.
But then I ran into this report by Ibn Jinnī in his al-Ḫaṣāʾiṣ, where he explicitly says that words that have an ā followed by a long consonant, may break the ʾalif up by inserting a hamzah, citing šābbah, dābbah > šaʾabbah, daʾabbah. He even cites poetry that requires it!
It seems then that our ʾAyyūb al-Saḫtiyānī (presumably the muḥaddiṯ, d. 131/748) was not inserting a hamzah into this word hypercorrectly at all, but was specifically employing a real poetic license that was condoned by grammarians in later times!
