Here's an interesting quote from S.A. Schelkunoff, "Theory of Antennas
of Arbitrary Size and Shape", Proc. of the I.R.E., September, 1941
(footnote 17): "From the point of view developed in this paper there is
no difference between 'end effect' and radiation."

This is the paper in which Schelkunoff develops his often-quoted
approximate equations for antenna feedpoint impedance (the ones
including sine and cosine integral -- Si and Ci -- terms). He says,
basically, that an antenna acts like a transmission line -- a conical
antenna like a constant-Z line and a cylindrical (e.g., wire or tubing)
antenna like a variable-Z line -- *except at the ends*. At the ends,
modes other than TEM are excited, resulting in radiation, modification
of antenna impedance, and modification of current distribution. The
radiation, he says, can be modeled as either a terminating impedance or
as a distributed impedance (R and L) along the line. You can find an
abbreviated version of this explanation in Kraus' _Antennas_.

A transmission line is similar to an antenna in only some respects, and
assuming they act exactly the same leads to erroneous conclusions. Among
the many mistakes made in recent postings is the assumption that a
complete reflection takes place from the end of an antenna wire. As
Schelkunoff, Kraus, and others explain, this isn't correct.

Roy Lewallen, W7EL