'Doc, I agree wm5k was suggesting a vertical array of two 5/8 waves, but I'm not sure that is what is happening.
I've referenced this work before and I've noted that I had my doubts even though he claims his real world setup was successful. I've done models of the Imax with slanted radials before, but I never had the idea to make the radials 22.5' feet long. So, I decided to do a model of his idea at 27.205 mhz using less the 20 segments...similar to what he did using Eznec Demo.
I set the model at 40' feet to the hub, and added 4 x 22.5' slanted down radials at about 45* degrees. I did not get the exact same results, but here is what my model produced. I diddled with the model to see if I could get the results closer, but to no avail. I also added a 1' wire to the bottom of my model in order to get the feed point source close to the bottom like his model shows when using just 3 segments for a 22.5' long radiator. He also modeled at 28.5 mhz and that could be the difference. I think his work is also quite old and he may have been using a different Eznec too.
[ATTACH]10186[/ATTACH]
I'm in the process of adding a mast and the feed line to this model to see what Eznec predicts under those circumstances too, and this may suggest that the common mode currents produced by such out-of-phase currents and imbalance at the feed point will need to be effectively mitigated somehow as well.
I don't think this model suggested by nm5k will produce the results he discusses. Right off the bat I don't think we will see a collinear array or in phase results from this setup either, and you can see by the red line on the antenna the currents on the radials are quite heavy. In the past I've suggested that the radial currents for such a setup are going to be very heavy on the radials, and that may well dominate this antenna modification idea.
Here are the currents for this simple model: [ATTACH]10187[/ATTACH]
You will note that the top portion of wire #2 shows a - sign for its phase and the rest of the antenna is a + sign. That means the top 1/2 wave of the radiator is not in phase with the rest of the antenna. This is not an array of two 5/8 wave in phase elements working together.
Note: sorry I got Mark's call nm5k wrong in this post and in some pdf files I've posted.