This glyph offers two problems: what is the left part and what is the right part? As regards the first problem, I remember how my very first exercise with rongorongo - now more than six years ago - was to see if I understood and could use the classification system of Barthel; my selfappointed task was to express the glyphs of the first line in Tahua (Aa1) in his code numbers. I would after that compare with his own version of this 'translation' and see if I had grasped his system.

This exercise made me realize that Barthel's system was not good enough for practical use, it lacked the precision needed for a 'true' transkription of glyphs into numerical values. The glyph that made me stop, think, and realize this was Aa1-12:

It could not possibly be the same type of glyph as the much slimmer Aa1-49:

And neither looked as the prototype (for no. 4) in Barthel's catalogue.