Now we shall pass to the Ivese-Stilwell experiment. Note that Ivese himself was
a SRT opponent and explained the experiment from the ether theory viewpoint
(which means that such an interpretation is also possible). Generally, it is
characteristic of SRT to "put" everything into a personal "pile" (probably,
in order to look more solid) or to "tie up" SRT with all theories (even not
completely verified), pretending that if SRT "sinks", then "all science will
also sink". Generally speaking, unlike the elementary theory of the Doppler
effect, determination of a frequency dependence in some arbitrary configuration
is a prerogative of experiments (and an implication of an additional hypothesis
for time here is rather doubtful).
Actually, the Ivese-Stilwell experiments, even in the ideal
case (with neglecting real features of a process) would determine not the
transversal Doppler effect, but the Doppler effect for two directions close to
and , i.e. the effects close to longitudinal ones.
These experiments are indirect, since the value of a relativistic correction
is a calculated quantity (which is compared, in addition, from various
regions, which results in the additional asymmetry). The experiments [22] have
shown essential systematic deviations from the relativistic expression (up to
6010). Therefore, the effect can be determined not so much by the
Doppler expression, as by the feature of reactions in beams. In addition to
mentioning the other alternative experimental data [22,120], we shall give
some criticism of considered experiments. Relativists describe the experiment
in such a manner, as if the transversal Doppler effect is perceived from one
point of an installation at some certain time instant (the time of passage
through the middle perpendicular). Actually, the perceived signal is an
integral sum from various regions of radiation for various time, and these
regions are, in addition, not perpendicular to the motion (where, for example,
the aberration has gone?). That is, the studied effect represents some
"composite mean value" between two longitudinal Doppler effects. Besides,
the theory (and the formulas) in SRT are presented for __plane-parallel__ waves,
but in fact we have point-like sources, i.e. the __spherical__ waves at these
distances. We write lengths of sides in a triangle: 1) the first side describes
a way of the signal along the axis Y from the source to the origin of the
reference system O, where the receiver was situated at the moment of
emission of the signal: Y0=ct; 2) the second side describes a passed way
of the receiver along the axis X from the moment of emission to the moment of
the receipt of the signal: X1=vt'; 3) the third side (diagonal) describes a
way of the signal from the source to the point of the receipt: ct'. Then, from
the relation of sides in a triangle it can be found the change of a time delay
as compared to the case at rest:
. In reality, we obtain the transversal
Doppler effect for spherical waves which also exists both for light and
in acoustics as well! As a result, for the real source the displacement into
the red area will be observed
(a greater time of action of such a displaced line), and the effect should
depend on the distance to the observation point. And who could prove that the
classical Doppler effect for plane-parallel waves must be applicable for light? This effect possesses the
classical form in the case of __pure wave__ motion only, you know.
But if light is not entirely a wave, other expressions could be obtained,
including the relativistic ones [60].
Thus, the given experiment can not be unconditionally attributed to the
experiments confirming the relativistic time slowdown in SRT.

Some relativists [38,107] distinguish three key experiments (by Michelson, Kennedy-Thorndike and Ivese-Stilwell) which should unambiguously result in the Lorentz transformations (a basis for SRT). We see, however, that all these three experiments are not evidential. SRT "hangs in the emptiness" even from the experimental point of view.