Consider using the unnumbered <bldv type="errata"> tag. Next use a feature structure for each erratum so that it is nested within this division. The method is based on the double end-point attachment technique presented in TEI P3 for critical apparatus (pp. 577-79) but has been modified for this purpose. (TEI P3 appears not to give advice on tagging errata.) Each correction in the errata section is then enclosed in <ref target=""> ... </ref>, and the target value may then be any unique string, such as err1. The cross-reference information is not otherwise tagged, but both the error in the text (the lemma of the erratum) and its suggested replacement are enclosed by two tags, <error from="" to=""> ... </error> and <correct> ... </correct>. The from attribute in the <error> tag should have the same value as the target attribute in the <ref> tag. The to attribute in the <error> tag should have a unique arbitrary value, such as a1, a2, a3, etc. The errata sheet should then look like the following (from Cotgrave's 1611 French-English dictionary), where errors appear in paragraph form, sometimes without separating punctuation, and certainly not in list form.
<bkdv type="errata"> <heading>Errata.</heading> <bkdv3 type="col" n="1"> <subhead>A</subhead> <ref target="err1">IN the first Page, col. I. <error from="err1" to="a1">Il a gaigne</error>, for, <correct>Il y a gaign‚.</correct> </ref> <ref target="err2"><error from="err2" to="a2">volunte</error>, for, <correct>volont‚.</correct></ref> and <ref target="err3">col. 2. <error from="err3" to="a3">an Adiectiue</error>, for <correct>a Nowne</correct></ref> <ref target="err4"><error from="err4" to="a4">whose tone foot is</error>, for, <correct>who hath one foot</correct></ref>. and <ref target="err5"><error from="err5" to="a5">Abbaisse</error>, for <correct>Abbaiss‚.</correct></ref> ... </bkdv>Once the errata sheet has been tagged, comparable anchor tags must be placed in the text itself. Here is part of the first page in which the first error is found. The error citation in the errata is wrong in one detail (it is uncapitalized), but the tagging makes clear that the contents of the <correct> tag within the <ref target="err1"> .. </ref> tag should be substituted for the text, whatever it is, between two anchors, those with the identifier attributes err1 and a1.
<ttdv3 type="entry" lemma="avoir">
<ttdv4 type="lemma"><form n="1">A.</form> <explan>
The third person singuler of the verbe <term>Avoir</term>, signi{\}
fies (of it selfe) hath, holdeth, possesseth, enioieth; as, </expan>
<egt n="1"> Assez a qui bon credit a: </egt> <explan> also (when it hath the lo{\}
cal Adverb <term>y</term> before it) is; as,</explan> <egt n="2">
Assez y a si trop n'y a: </egt>
<explan>But when it precedes a Participle in the preterperfect
Tense of another verbe, it is restrained vnto the first
signification, hath;</explan> as, <egt n="1"><anchor id="err1">il a gaigne<anchor id="a1"> ce qu'il a peu: </egt> <explan> In
all which (and otherwise, onely as it is a letter) it is
written without Accent. </explan>
<bkdv errata>
<heading>Errata.</heading>
<col 1>
<subheading>A</subheading>
<tt xref>IN the first Page, col. I. <tt error>Il a gaigne<tt text>,
for, <tt correct>Il y a gaign{'e}.
<tt xref ><tt error>volunte<tt text>, for, <tt correct>volont{'e}.<tt
text> and <tt xref>col. 2. <tt error>an Adiectiue<tt text>, for <tt correct>a
Nowne<tt xref>whose tone foot is<tt text>, for, <tt correct>who hath
one foot<tt text>. and
<tt xref>Abbaisse<tt text>, for <tt correct>Abbaiss{'e}.<tt ->
...