I have tried separately processing data for Currier A and B. Again, I only considered pages that include at least an image-break. The pages are almost equally split between the two languages, so both sets are large enough to be meaningful. I cannot see anything particularly illuminating in the new graphs.
As already observed, with the exception of -am, suffixes before an image break behave similarly to those before a line-break. In B, -dy is also different, being considerably more frequent before an image-break than a line-break: but nothing as dramatic as what happens with prefixes.
The drop in q- after an image-break remains the most striking feature, and it is stronger in Currier A. In Currier B, several other prefixes behave differently after an image-break or a line-break: e.g. ok-, yk-, she-, sa-, so-.
Tentatively, one could say that an image break affects the suffix of the preceding word like line-breaks do. But the prefix on the other side of the break behaves differently. This seems to point out that the phenomenon is not entirely the same in the two cases, but it is not independent either. For instance, it seems that one can exclude that words are typically split into two halves in one of the two cases but not in the other: this would not explain why suffixes before the two breaks are so similar.
As already observed, with the exception of -am, suffixes before an image break behave similarly to those before a line-break. In B, -dy is also different, being considerably more frequent before an image-break than a line-break: but nothing as dramatic as what happens with prefixes.
The drop in q- after an image-break remains the most striking feature, and it is stronger in Currier A. In Currier B, several other prefixes behave differently after an image-break or a line-break: e.g. ok-, yk-, she-, sa-, so-.
Tentatively, one could say that an image break affects the suffix of the preceding word like line-breaks do. But the prefix on the other side of the break behaves differently. This seems to point out that the phenomenon is not entirely the same in the two cases, but it is not independent either. For instance, it seems that one can exclude that words are typically split into two halves in one of the two cases but not in the other: this would not explain why suffixes before the two breaks are so similar.


![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)