Fischer is of the opinion that line Nb5 is the beginning of the text and that the end is in line Na1:

"Because the end of RR 23a1 has been 'squeezed' into the remaining space, RR 23a1 must actually be the end of inscription. This would probably make RR 23b5 the beginning of the inscription. (This is not contradicted by parallel passages occuring elsewhere.)"

I do not find Fischer's argumentation of any value whatsoever.

Probably, instead, the text begins on side a with line a1 and ends on side b with line b5. There is a continuation in glyph types (most clearly seen in GD21, poporo) from side a to side b. Moreover, there is a continuation over all five lines on side a, expressed in GD52 (kai). Other signs also point to a continuation of the text from line Na1 up to and including line Nb6.

As every line has more or less destroyed beginnings and ends I have documented each line in a separate page.