Braun Tibor, Schubert András (szerk.): Szakértői bírálat (peer review) a tudományos kutatásban : Válogatott tanulmányok a téma szakirodalmából (A MTAK Informatikai És Tudományelemzési Sorozata 7., 1993)

MARTIN RUDERFER: The Fallacy of Peer Review: Judgement without Science and a Case History

177 RUDER

I ER: T

IIK I'

A

I.I ACY O

l PEER REVIEW timekeeping, from oppositely rotating clocks in the Sagnac experiment and from differential rotation of atomic clocks at the same latitude in the HafeleKeating experiment. The difference in clock rates then all derive, to second order, from equation (25) (p.399) which has its relativistic origin in equation (2) (p.

388): t

£ - tj = (0I-0I)/c

2 -(VI

2 - VJ

2)/2c

2 + t

u

v (1) where ij are rotating clocks, t is clock rate, <j> is the gravitational potential, c is the speed of light, VJJ is geocentric rotational velocity and u is Earth's orbital velocity. The Vjj (uv) term is the kinematic effect in the geocentric (heliocentric) frame. For oppositely rotating clocks with ground speed v at Earth's surface corresponding to oppositely directed light rays in the usual Sagnac experiment, Vj = Í2R - v and v

; = Í2R + v, where is angular velocity of Earth and R is distance of clocks to Earth's axis. Then the difference in clock readings due to the geocentric kinematic effect (i.e. neglecting the <*> and uv terms) after one revolution in time At = 2

ÍT

/

Í

2 is, since tjj = Atjj/At, (ti-tj)At ~ Atj-Atj = 4?rRv/c

2 (2) This is a common expression for the first-order Sagnac effect.*

2

9

' In the Hafele-Keating experiment, clock j is flown around the world at average ground speed v and clock i is stationary on Earth. Hence Vj = Í2R and Vj = Í2R + v. Then equation (1) similarly yields, after one revolution of clock i in time Ati = t

0 = 2ir/S2 at which time clock j registers time At

; = r T - T

0 = -(2i2Rv + v

2)r

0/2c

2 (3) This is exactly the kinematic relation tested by Hafele and Keating.*

3

0> Moreover, they noted (last paragraph) that a more precise evaluation should yield effects of Moon and Sun beyond their measurement precision. These include gravitational and kinematic effects, the solar contribution to the latter being identifiable with the last (uv) term in equation (1). This is small but is evaluated and listed for reference in the rejected paper as "sidereal corrections" (Table 1, p.

395). Over the six year period of the follow-up analysis these became measurable by the relative mean drift of the clocks (Fig.5, p.410) and by the long-term drift of all the clocks combined due to interaction with the geocentric effect (Fig.4, p.

409). Furthermore, these are inherently verifiable directly by the proposed double-disc Sagnac experiment (p.401). Although this simple unification of rotating clock behaviour is obvious in retrospect, the Hafele-Keating experiment was not brought up by the author (except in P) or by the refreees, nor has its straightforward relevance to atomic timekeeping, the Cannon-Jensen finding or the Sagnac effect been heretofore recognised. The reason becomes apparent from the derivation in the follow-up paper (p.

417) that classical transverse aberration suffices to explain all the results. To the precision of the data, the kinematic terms in equation (1) are properly explained by classical, not relativistic, transverse aberration. In effect, a paradigm shift in the interpretation of rotating clock behaviour is required.