The idea I’m currently on is the fact that there are 81 letters in the Morse Code phrases and 25 E’s mixed in.  This gives us 106 letters in total, not a very useful number.  You can factor it 1×106 or 2×53.  So I got to thinking, if it’s a “matrix” and we know it’s really unlikely that there’s substitution then it would have to be some really basic transposition.  After some prayer and cogitation, I returned to the issue of all those E’s.  They have to serve some purpose.

I think there’s two squares, one 5×5 and one 9×9.  There is a 9×9 square of letters that make up the classic Kryptos Morse code phrases and then there’s a square of E’s.  These have been combined.

My favorite explanation for how to solve this would be to build the 5×5 grid of displaced letters and then grid out the leftover letters and the E’s.  My idea is that the E’s are actually locations to insert the displaced letters in order to restore the message that was originally written.

Why E’s?  Most null ciphertext is X’s or Q’s or Z’s or something?  Well, it helps baffle cryptanalysis for that very reason but far simpler is the fact that E’s in Morse Code are just dots.  The only null text that they could use that wouldn’t make the message unreadable were E’s.  Add a bunch of dots around the words of a Morse Code message and it’s still readable.

I haven’t been able to get it to work, but it’s a nice idea that’s worth some effort.  I’m definitely considering the possibility that there’s a different solve-method than the one I outlined above.