King Zulu wrote:
The FW dipole I used for years had a 1.2:1 SWR, fed with RG59 and no balun.

It still may have had a feedpoint impedance of ~100 ohms. RG59's
Z0 is 75 ohms, close to what you would need to transform 100 ohms
to 50 ohms. An odd number of 1/4WLs of RG59 will transform 100 ohms
to 57 ohms, neglecting losses.

Never saw an indication that the feed line was radiating.

We can mount a logical argument that since the feedpoint
impedance looking in each direction is different, the
currents have to be unbalanced resulting in feedline
radiation.

The 2db gain was what the ARRL Antenna Book (1964 vintage, p. 142) was
calling out for a FW collinear array gain. Actually, the FW dipole fed at
the current node would be like a 1-wavelength long longwire (a short
longwire?), and the Antenna Book shows a negligible 1/2 db gain (p. 170 on
"Long-Wire Antennas") over the dipole gain. The cloverleaf assertion was
based the Antenna Book radiation patterns shown on p. 39 (Fig 2-16) and the
discussion of "How Patterns are Formed", as well as p. 59 (Fig 2-74b)
showing the FW pattern distortion caused by end feeding.

EZNEC says that it is somewhat of a square pattern where the "corners"
of the square are just barely starting to form a cloverleaf. The "nulls"
off the ends are only 3 dB down. I suspect there was a lot of guessing
about radiation patterns in 1964. Some say there still is a lot of
guessing. :-)
--
73, Cecil http://www.qsl.net/w5dxp