Antonio Vernucci wrote:

1. The feedline Z0 isn't exactly 50 ohms. The Z0 of coax easily varies
+/- 5 ohms from nominal, and sometimes closer to +/- 10 -- it's seldom
exactly 50. If you connect a perfect 50 ohm load to your transmitter
via a 45 ohm line, the impedance seen by the transmitter will change
with line length. Consequently, the SWR meter reading will also
change. The actual SWR on the line will not, except as dictated by
loss, described next.

2. The feedline has loss. The SWR will improve as the line becomes
longer due to line loss. If the line is long enough to be very lossy,
the transmitter will see nearly the line's Z0 regardless of what load
is connected to the other end. The actual SWR on the line will be
greatest at the load, decreasing as you get farther away.

3. There is current on the outside of the coax shield (common mode
current). When this happens, the feedline becomes part of the antenna.
Consequently, changing the feedline length actually changes the
effective antenna length, which in turn changes the feedpoint impedance.

I would say that in case no. 1 the meter measures an apparent SWR,
whilst in case no. 2 it measures the real SWR existing at the
measurement point. I am not sure what it measures in case no. 3

No, the meter is measuring the exact same thing in all cases. it always
(indirectly) measures the SWR within itself, which is directly related
to the impedance connected to the output end of the meter. It never
measures the SWR on any transmission line outside itself. The three
cases only explain reasons the impedance connected to the output of the
meter -- hence the meter reading -- changes as the transmission line
length is changed.

Roy Lewallen, W7EL