John Ferrell wrote:

Do I misunderstand? My intuition is to place the radials I use over as
much area as I can. The way I read your note is that they should be
installed in a small sector... I think I am misunderstanding that
point.

You want to cover the maximum area...but using the minimum amount of
wire to do so effectively...which is what 0.025 wavelengths spacing
between the endpoints yields.

I am inclined to pursue the advice presented in the article
http://www.ncjweb.com/k3lcmaxgainradials.pdf

The problem with all of K3LC's articles is that he only uses NEC-4
modeling with no empirical measurement data. N6LF's article below
compares his measurements to the same NEC-4 model (see Figure 5 and
last paragraph on page 5). Modeling programs are notorious for poorly
handling radials on ground. Although NEC-4 is supposedly the best for
this, the results N6LF reported (i.e. NEC-4 saying no improvement for
more than 16 radials) are inconsistent with all measurement data going
back to the 30's (including that used by the FCC for broadcast station
commissioning).

http://www.antennasbyn6lf.com/files/...asurements.pdf
(see Fig. 5 & bottom paragraph, p. 5)

73, Bill W4ZV