Neo-Newtonian Mechanics With Extension

NEO-NEWTONIAN MECHANICS
WITH EXTENSION TO
RELATIVISTIC VELOCITIES
PART 1: NON-RADIATIVE EFFECTS
DENNIS P. ALLEN, JR.

AND
JEREMY DUNNING-DAVIES
Copyright © 2019 Dennis P. Allen, Jr. & Jeremy Dunning-Davies.
All rights reserved.
Ninth Edition
ISBN:1491024895
ISBN-13:9781491024898
LCCN:2013914618
DEDICATION
This book is respectively dedicated to the Holy Spirit of

God, Source of all Wisdom and Knowledge and the
Spirit of Truth, together with His Most Chaste Spouse,
the Blessed Virgin Mary, without Whom this book
could have neither been conceived nor written.
However, any mistakes are, of course, solely the

authors’ responsibility.
CONTENTS
Acknowledgments i
Preface to the Second Edition ii
1 Introduction 1
2 Physical content and the closely related hydra effect 5
3 Causality in Mechanics 11
4 Neo-Newtonian Theory 16
5 An elementary method for the elimination of the first 65

principal difficulty in the axiomatization of classical
mechanics: circular reasoning.
6 Some results in classical mechanics for the case of a variable 69

mass.
7 Wesley, his neo-mechanics and special relativity. 80
8 Appendices – Appendix 1 88
Appendix 2 97
ACKNOWLEDGMENTS
The first author (DPA) would like to express thanks to his

late physics mentor, Thomas E. Phipps Jr., PhD (Harvard,
1951), for patiently explaining to him (at length) that lower
order physics must be made right before higher order
physics can be made right. Also, thanks to Greg Volk for
his help and encouragement, and especially his typing,
during the last few years as the neo-Newtonian mechanics
slowly came into being. Finally, thanks also to Glen Deen
for helpful math typing and for his encouragement, and to
Harvey Fiala for reading the first part of this work and
offering his suggestions.
The second author (JD-D) would like to thank all who

taught him so much over the years: Professors R. Shail, N.
B. Slater, P. T. Landsberg and G. H. A. Cole, but mostly he
wishes to thank his wife, Faith, for all her patience, love
and support over the years, when he was attempting to
produce some new work.
i
NEO-NEWTONIAN MECHANICS WITH EXTENSION TO
RELATIVISTIC VELOCITIES.
Preface to the Second Edition
The first author has struggled with Chapter 4 “Neo-

Newtonian Theory” for years, and it has now become clear
that, while the approach adopted herein of neglecting
gyroscopic precession (when small) is indeed necessary
pedagogically (because otherwise the calculations become
so extensive as to mandate the use of mathematical
software such as Derive or Mathematica); it nevertheless is
unfortunately true that this neglecting of precession (unless
there is no precession whatsoever as in Prof. A. Dmitriev’s
falling rotor experiment) often results in unacceptable
approximation error.
In Prof. O. G. Tietjens’ Preface to the Dover Edition of

the classic Fundamentals of Hydro- And
Aeromechanics, he remarks that “During the brilliant
development of theoretical hydrodynamics in the second
half of the last century, contact with reality was more and
more lost. This was due to the fact that in this so-called
classical hydrodynamics everything was sacrificed to logical
construction and no results could be obtained unless they
were deduced from the basic equations. Yet, in order to
overcome the mathematical difficulties, these equations were
simplified in a manner which often was not permissible even as an
approximation.” [Emphasis added]
Thus we see that similar problems of exposition have arisen

previously in the area of classical hydrodynamics. And,
therefore, there is nothing to do but to reiterate the above
warning that neglecting (as small) a non-vanishing precession
in gyroscopic calculations using the Neo-Newtonian theory
of Chapter 4 – while pedagogically necessary – often leads to
unacceptable approximation!
ii
ALLEN AND DUNNING-DAVIES
iii
1 INTRODUCTION
It has been some years now since the late Professor Eric Laithwaite
dropped his bombshell at a Friday Evening Discourse meeting of
Britain’s Royal Institution. It was here, one evening in 1973, that
Professor Laithwaite chose to demonstrate the possibility that the well-
established laws of motion due to Newton could be violated. He first
showed his audience a large gyroscope he had constructed. The device,
which was very heavy – in fact, it weighed approximately fifty pounds –
could be spun on a low friction bearing to high speeds by means of an
electric motor. When at rest the device proved difficult to even lift but,
once the wheel started to rotate at high speed and then was whirled
about him with the shaft nearly horizontal and with the direction of the
whirling being the same as the device would precess about the vertical
had it had a fixed pivot point and otherwise was free to move, it became
so light he was able to raise it above his head very easily. He then
explained to his audience that Newton’s laws of motion appeared to be
violated by this demonstration. A moment’s thought shows that this
sort of explanation was not to be appreciated by the type of audience
facing Laithwaite; instead of gasps of wonder and surprise, he was faced
by an involuntary hush followed by a frosty silence. Why? Surely one
would expect an audience of dedicated scientists seeking the truth about
our amazing Universe to be excited by this indication of a possible new
path to follow in research – one unhampered by the restrictions
imposed by Newton’s laws of motion?
However, the reality concerning the so-called scientific search for

truth is very different from that imagined by a gullible public and
reinforced by a compliant media. Newton’s laws of motion count as one
of the cornerstones of scientific conventional wisdom; their total
acceptance is possibly even more established than Einstein’s relativity in
the minds of those in power in science. Hence, in retrospect,
Laithwaite’s reception was hardly unexpected and his subsequent
ostracization was again something almost to be expected. On the other
hand, it is easy to see this with the benefit of hindsight and so, it is also
easy to understand Laithwaite’s dismay at the reception he received and
his subsequent treatment at the hands of powerful establishment
hierarchy. It is interesting to reflect that this isn’t the first time questions
have surfaced about the universal applicability of Newton’s laws of
1
NEO-NEWTONIAN MECHANICS WITH EXTENSION TO RELATIVISTIC
VELOCITIES.
motion. In fact, back in Newton’s own day, Robert Hooke carried out
an experiment whose result did just that but today it is difficult to find
a detailed reference to that piece of work.
Nevertheless, over the years, people have continued to work on this

problem of the applicability, or otherwise, of Newton’s laws of motion
to gyroscopic motion. Here a short selection of articles intended to
further this examination is presented. The first is a short introductory
article entitled Physical Content and the Closely Related Hydra Effect, which
sets the scene, so to speak, for much that follows in the more technical
articles. It looks briefly at an attitude of mind which seem to pervade
many involved in modern science.
The second article is about the question of causality in mechanics that

is discussed in terms of Euler's method for the numerical integration a
system of ordinary differential equations, the most elementary such
method of them all.
This short piece is followed by a considerably longer one entitled Neo-

Newtonian Theory. In this, an attempt is made to generalize classical
Newtonian mechanics to take account of recent gyroscopic studies
undertaken by first Alexander Charles Jones closely followed by Eric
Laithwaite and later by Bruce de Palma and Harvey Fiala amongst
others. It cannot be denied that the above-mentioned demonstration by
Laithwaite, as well as more recent experiments by de Palma and Fiala,
indicate a problem with Newtonian mechanics when high-speed
rotations are involved. It might reasonably be noted that Newton
himself did not have data on rapid gyroscopic motion at his disposal;
the nearest he would have come to such data is that involving the
rotation of heavenly bodies and such involve relatively low rates of
rotation. Thus, he could not sensibly have been expected to discover
the gyroscopic effects stumbled upon by Jones, Laithwaite and others.
It seems clear that Newtonian mechanics must be almost correct except
in the case of very rapid rotational motion and hence, this article
attempts to apply a small correction to that mechanics which is normally
so successful in explaining everyday phenomena. It follows that
Newton’s original laws of motion are retained, but the Second Law is
retained in its original form where force is proportional to the rate of
change of the momentum rather than the popular form which claims
force is equal to mass multiplied by acceleration. However, the big
difference is that a variable inertial mass is assumed. This is, in reality,
the only basic assumption of Newton which is relaxed and so, it is
assumed that the inertial mass may vary with motion even for non-
2
relativistic situations. The article begins by developing a generalized

Neo-Newtonian theory and then applies that theory to the case of a
gyroscope with a fixed pivot which precesses slowly but does not nutate.
It continues to consider the situation when the precessional angular
velocity suddenly experiences a constant deceleration until the
precession halts and while restrained from nutation in a way that is easily
realizable physically. These examples involve anomalous behavior and
it is hoped that analysis of them will lead to a successful prediction of
other experienced anomalous gyroscopic behavior. It is hoped also that
the analysis proffered here will eliminate the problems associated with
the present well-known Newtonian mechanics.
The fourth contribution is a short piece entitled An Elementary Method

for the Elimination of the First Principal Difficulty in the Axiomatization of
Classical Mechanics: Circular Reasoning. Here an extension of the notion of
a Euclidean-style axiomatization of a physical theory is presented by way
of the example of the resolution to the well-known difficulty in
axiomatizing Newtonian mechanics concerning the circular reasoning
involved with inertial systems and forces. An elementary, clear,
straightforward resolution of this legendary conundrum results, which
is in stark contrast with previous attempts. It might be noted also that
the method used here extends, by example, the classical Euclidean
notion of the axiomatization of a physical theory.
The fifth piece, entitled Some results in classical mechanics for the case of a
variable mass is concerned, as its title clearly implies, with considering the
extension of several results in classical mechanics to the case involving
a variable mass. In most, if not all, introductions to classical mechanics,
the mass is assumed to be constant. Usually this is mentioned and often
attention is drawn to such systems as rocket motion to indicate that, in
practice, the mass is not always a constant. In truth, many students
actually meet a varying mass for the first time when introduced to the
Special Theory of Relativity. However, varying masses do occur in
nature when relativistic effects are not important as is evidenced by
material contained in earlier contributions to this short publication. In
this particular article, an attempt is made to draw together some
common results of classical mechanics with a variable mass taken into
account. Particular attention will be drawn to a perceived change in the
expression for the kinetic energy and to crucial changes in the basic form
of Lagrange’s equations of motion. Here the traditional approach to the
derivation of Lagrange’s equations of motion is considered and
3
VELOCITIES.
extended. However, it should be noted that the other means of deriving

these equations contain the same flaws for variable mass situations as
are encountered here. Again, in addition, the derivation via Hamilton’s
Principle relies on use of the Lagrangian function which, in turn, is
dependent on the potential energy and so ensures that attention is
immediately restricted to conservative systems as well. It is possibly
because variable mass systems are encountered so rarely by most that
these restrictions on the use of Lagrange’s equations of motion go
unnoticed. However, these are important points to be noted on those
relatively rare occasions when a variable mass does come into play.
In the sixth, entitled Wesley, his Neo-mechanics and Special Relativity, an

introduction to Wesley’s neo-mechanics is presented. It is shown to
produce some of the same results as Special Relativity but without both
the mathematical and philosophical basis of that subject. As with other
work in which results associated with General Relativity are obtained
without recourse to the fundamental bases of that subject, so here too
the pre-eminent place afforded Special Relativity in modern science is
called into question. The opportunity is taken to extend Wesley’s ideas
to the case where the mass of the body when at rest is not constant.
The two appendices contain detailed calculations using the Derive 5

mathematical software of the two unknown constants C and K of the
neo-Newtonian theory (nearly) from Prof. Dmitriev’s falling rotor data
and the calculations supporting our thoughts related to our He4
mechanics speculations, respectively.
As a postscript, Hans Lehner kindly informs us that the demonstration

of Laithwaite’s mentioned above requires a wheel constructed of pure
steel without any vanadium, chrome, nickel etc., and so is a magnetic
effect since pure steel is a ferromagnetic material. We do not consider
(macro) electromagnetic effects herein following Newton, but we do
utilize his key notion of force defined in terms of effects, not causes;
and so such electromagnetic effects can then be added to the theory
later on as they become more well understood.
4
2 PHYSICAL CONTENT AND THE
CLOSELY RELATED HYDRA EFFECT
Dennis P. Allen Jr.
17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI

49456
allens10@sbcglobal.net
&
Jeremy Dunning-Davies.
Institute of Basic Research, Palm Harbor, Florida, U.S.A.

and
Institute for Theoretical Physics and Advanced
Mathematics (IFM) Einstein-Galilei,
Via Santa Gonda, 14 - 59100 Prato, Italy.
j.dunning-davies@hull.ac.uk
In Chapter 1 of “The Cambridge Companion to Newton”,

Prof. Robert DiSalle talks of the “physical content” of a
physical theory. In the authors’ opinion, this notion is not
well understood nowadays by physicists, whether they be
orthodox or alternative. At this point, a contemporary
example of this notion might serve to illustrate the point:
when the GPS became available in automobiles, the
accuracy was over 50 feet, and so, if one was led by this
system to (say) turn right; then, if there were two possible
right turns close by one another, one simply had to guess
5
as to which one to take. However, if one took the wrong

turn, it soon became apparent via the system that a wrong
turn had been made, and so one merely had to retrace one’s
route and then take the other turn when one arrived back
at the point of error. Thus, this lack of accuracy was only
really a nuisance.
Now, however, the accuracy is 15-18 inches … because

the GPS engineers have rejected special and hence, also
general, relativity but especially the “relativity of
simultaneity” and returned to Newton’s “absolute time”.
[See “A New Theory of Gravity …” (2007, Physics Essays)
by Ronald R. Hatch, a leading GPS scientist, and google
‘geo-cashing’ concerning the 15-18 inch accuracy.] Of
course, this might be viewed quite reasonably as another
way of saying that special and general relativity have less
physical content than current GPS theory.
Although it may appear that this notion is not nearly as

relevant to getting a device enabled by an alternative theory
into mass production than (say) the generality and/or the
beauty of the new alternative theory first developed, this is
nevertheless incorrect in general. To explain the thinking
here, consider the notion of the “hydra effect”. The Hydra
was a mythical monster with multiple heads; and if one
head was cut off, two grew back in its place. Thus, the
ancients knew about the fact that a physical theory with
insufficient physical content tends to lead to the occurrence
of problems in the designing of consequential devices for
mass production … in that when a design or production
problem rears its ugly head in an endeavor then, in solving
one such problem, typically two new ones arise to take its
6
place … and so on. Or, to put it another way, the hydra

effect is a consequence of the lack of physical content in
the theory that was devised to enable such a device. It
seems that this explains just why at the Tesla Tech and
Natural Philosophy Alliance meetings, for example, one
runs into truly amazing inventions that may be both
observed and examined at these events; however, for such
hardware – even though it may certainly appear very
praiseworthy and remarkable – the hydra effect often
arises, and so the desired mass production never actually
occurs in many, many cases, unfortunately. This is to say
that the device cannot be designed and built in a mass
production scenario due to the lack of convergence of the
development process. Thus, an interesting and/or
beautiful theory may well turn out to simply have too little
physical content in order to allow commercially successful
mass production of (more or less) working inventions
based upon this theory!
One might reflect on the great change in attitudes to

science over the more recent centuries which has
occasioned this unfortunate state of affairs. Even a casual
glance through Gale Christianson’s hefty tome dedicated
to the life of Newton (In the Presence of the Creator: Isaac
Newton and his Times, The Free Press, 1984) indicates the
stress Newton and many of his contemporaries, such as
Hooke, placed on experimentation and knowledge of
detailed observational data. A more detailed examination
of this work shows just how much time Newton spent
collecting data to check, and even help formulate, his
theoretical expressions and predictions. It also shows how
7
much time and effort he expended performing more and

more detailed experiments before announcing any progress
in a given field. This latter point is emphasized when one
reads of all his work on light before he produced the final
form of one of his monumental books, Opticks.
True, many of Newton’s major discoveries were made

possible by his phenomenal mathematical ability and
insight; it was probably this, as much as anything, that
placed him ahead of Hooke since, although Hooke was a
brilliantly insightful scientist, his mathematical ability and
knowledge let him down when it came to following his
ideas through to a final conclusive theory. However, in all
one reads, it seems that mathematics was really only a tool
for Newton, brilliant mathematician though he was, in
helping him achieve a true, complete understanding of any
particular topic he was wont to investigate. It does seem
that Newton recognized, either consciously or
unconsciously, the dual role mathematics plays in
academia. Firstly, it may be studied quite correctly and
profitably as a genuine subject in its own right but,
secondly, it may be both studied and used as the language
of physics. In its second role, mathematics is merely a tool
and must always remain subservient to the physics it is
being used to attempt to explain; it is the physics which is
all-important here.
Again, in its second role, it is being used to formulate a

theoretical model of a physical situation; it is not being used
to describe reality exactly. This is a point which is often
forgotten these days. Scientists create models, some of
which seem to relate very closely to physical reality but that
8
reality is never ever achieved completely. It is the lack of

realization of this point which has lead, and continues to
lead, to many problems with physical theories whose
advocates and supporters attempt to claim too much for
them. This is probably true, for example, for both the
Special and General Theories of Relativity alluded to
above. The other huge problem which arises in this
context, however, relates to the best way of showing a
theory either correct or incorrect. In many ways, the best
way to achieve either of these aims is to produce a
commercially viable device which shows quite clearly that
a particular theory is, indeed, correct. If, in producing the
said theory, due attention has not been paid to the detailed
physics involved, it is highly likely that any suggested device
either cannot be produced or so many problems arise in its
manufacture that commercial interest diminishes. This is
where these remarks link up closely with the above
mentioned Hydra effect and this is the sort of effect that
might easily arise in these circumstances. Some might
simply use such manufacturing problems to dismiss the
theory involved as incorrect but that could prove disastrous
for mankind in that useful devices could remain
unproduced but, more importantly, it could mean dubious
theories remaining in vogue because the device which
would clinch their demise had not been produced. If such
a lack of production is due to a lack of attention to physical
input in the original model, then that may only be viewed
as a tragedy from all points of view and could lead, quite
easily, to scientists continuing along the wrong path for
many more years than they should. It might be felt by some
that the present generation is seeing science following the
9
wrong path for just this reason in a variety of fields.
10
3 CAUSALITY IN MECHANICS
Dennis P. Allen Jr.
17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI 49456
There appears to be considerable confusion in classical physics, not

involving electromagnetic or gravitational phenomena, concerning
causality. The late Prof. Oleg D. Jefimenko writes near the beginning
of chapter 1 of his “Causality Electromagnetic Induction and
Gravitation” that: “One of the most important tasks of physics is to
establish causal relations between physical phenomena. No physical
theory can be complete unless it provides a clear statement and
description of causal links involved in the phenomena encompassed by
that theory. In establishing and describing causal relations it is important
not to confuse equations which we call ‘basic laws’ with ‘causal
equations.’ A ‘basic law’ is an equation (or a system of equations) from
which we can derive most (hopefully all) possible correlations between
the various quantities involved in a particular group of phenomena
subject to this ‘basic law.’ A ‘causal equation,’ on the other hand, is an
equation that unambiguously relates a quantity representing an effect to
one or more quantities representing the cause of this effect. Clearly, a
‘basic law’ need not constitute a causal relation, and an equation
depicting a causal relation may not necessarily be among the ‘basic laws’
in the above sense.”
“Causal relations between phenomena are governed by the principle of

causality. According to this principle, all present phenomena are
exclusively determined by past events. Therefore, equations depicting
causal relations between physical phenomena must, in general, be
equations where a present-time quantity (the effect) relates to one or
more quantities (causes) that existed at some previous time. An
exception to this rule are equations constituting causal relations by
definition; for example, if force is defined as the cause of acceleration,
then the equation F = ma, where F is the force and a is the acceleration,
is a causal equation by definition.”
11
VELOCITIES.
“In general, then, according to the principle of causality, an equation

between two or more quantities simultaneous in time but separated in
space cannot represent a causal relation between these quantities
because, according to this principle, the cause must precede its effect.
Therefore the only kind of equations representing causal relations
between physical quantities, other than equations representing cause
and effect by definition, must be equations involving ‘retarded’
(previous-time) quantities.”
It is evident that Jefimenko sees no way to introduce causality into

mechanics other than by definition. And Prof. A.P. French, in his
widely used “Newtonian Mechanics” beginning physics text, also
appears to be similarly confused as he says in his section on gyroscopic
nutation; “However convincing the analysis of gyroscopic precession
may seem, one may still wonder how a gyroscope can possibly defy
gravity in the way it appears to do. The answer is that this immunity is
indeed only apparent. If a flywheel is set spinning about a horizontal
axis, with both ends of the axle supported, the first thing that happens
if the support at one end A … is removed is that this end does begin to
fall vertically. Immediately thereafter, however, the precessional motion
in a horizontal plane begins, and as this happens the falling motion slows
down, until the point A is moving in a purely horizontal direction. It
does not stay like this; what happens next is that the precession slows
down and the end of the axle rises again, ideally to its initial level. The
whole sequence is repeated over and over … The process is called
nutation…”
Thus French also seems to fall short of demonstrating causality …

although he seems to allude to the idea that first in this gyroscopic
situation (after the gyroscope at t = 0 suddenly becomes unsupported
at one end) nutation begins which then immediately causes precession
to commence – a sort of causality that is apparently not completely
definitial as in Jefimenko’s just given quotation; but the difficulty is that
this simultaneity is shown by the exact solution to the system of two
second order ordinary differential equations describing the ensuing
precession and so on. However, this difficulty is easily obviated as
follows:
12
First notice that empirical physics has the property that since
measurement of physical variables is only approximate to just so many
significant figures, this means mathematically that one begins by
“making the continuum discrete” in that (say) the relevant physical
variables can only be measured to one significant figure, then if we
truncate (rounding is much the same) our numbers in (for example)
French’s nutation case (just quoted), then all numbers x with 2 ≤ x < 3
will then assigned the one significant figure 2 … and so on. [In the case
of (say) 0 < x < 1, we note that if we write x scientifically as k (10)^n,
then clearly the absolute value of n is bounded in our experimental
work.) Thus, when we assign measured numbers to this gyro situation
and then numerically integrate the system of two second order ODE’s
(while it may appear that French has one first order and one second
order ODE, nevertheless, just above the first numbered first order
ODE is the second order ODE it came from via integration) by Euler’s
method (the most elementary and straight-forward method) [1] after
choosing a sufficiently small time step Δt > 0; instead of referring to
French’s solution, we see that the nutation angle (measured from the
horizontal) together with its time derivative and also the precession
angle together with its first two time derivatives are all zero at t = 0 (the
initial conditions); but when the one support is removed, nevertheless,
the second nutation angle time derivative does not vanish as it is
accelerated by gravity instantly. This results in the initial values of all
but the second time derivative of the nutation vanishing at t = 0, but
after a time step of Δt, we see that the first time derivative of the
nutation then also becomes non-zero, and, of course, the nutation
second time derivative remains non-zero too as a time step of Δt occurs
… and the precession second time derivative may now become non-
zero too after this one step. But the other three quantities remain zero
here. Further, after another such time step, the nutation angle then
becomes non-zero too, just as the nutation first and second time
derivatives are non-zero as well. However, what about the precession
angle? We find that the precession angle is still zero after two time steps
… although the nutation angle is not! Thus, in making the continuum
discrete, one sees here that the nutation precedes the precession, and so
it can then be said in the sense of Jefimenko above that there is a true
13
VELOCITIES.
causal relation here with the nutation causing the precession as the
physical process develops from t = 0!
It should be noted that the continuum is dearly beloved by

mathematicians, and even the late Prof. Errett Bishop, in his
monumental “Foundations of Constructive Analysis,” mentions that
Luitzen Brower (of the Brower fixed point theorem and an important
earlier constructivist as well as one of the founders of modern topology)
seemed to feel that the continuum would [constructively] turn out to be
discrete “if he did not personally intervene”! But continuum
mathematics, nevertheless, obscures causality in mechanics, and that is
rather unfortunate, of course! This clearly illustrates that the over-
mathematization of physics nowadays is certainly not without its
deleterious foundational effects!
References
[1] Wilfred Kaplan, Ordinary Differential Equations, 400-1, Addison-

Wesley Publishing, 1958; (but the author does not mention Euler’s
name).
Gyroscope Coordinate Diagram
14
15
NEO-NEWTONIAN THEORY
Dennis P. Allen, Jr.
17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI
49456-9273
1. Introduction
We first draw the attention of the reader to Max
Jammer’s classic and prize winning work, Concepts of Mass in
Classical and Modern Physics [26] … in as much as the present
chapter is primarily about (variable) inertial mass … and
also somewhat about gravitational mass as well.
To begin generalizing Newton’s theory in Part One

(Sections 1 & 2), let us reconsider our assumptions about
the Laithwaite Effect [5]. Do the inertial and gravitational
masses of a particle differ by much (or at all) in a highly
rotational situation? Where I once assumed that the two
would, in general, differ; I now admit that they may be
nearly equal in certain important situations. This is
because:
1. Our previous theory predicts for a slowly
precessing gyro that only the weight of the
gyroscope changes, but the formula for the spin s
of Hay's gyro [1] is approximately (in the realm of
Newtonian mechanics)
mgl
s (1.1)
I3p
and is exact for horizontal precession, where l is the
distance of the center of mass of the flywheel from
the pivot, m the mass of the flywheel and hence its
dead weight, g is the acceleration of gravity, I 3 is the
moment of inertia of the gyro about its axis, and p is
the flywheel angular precession velocity assumed
small. According to our old theory only m would
16
change, but this formula has undoubtedly been

thoroughly checked experimentally and so this
theory does not seem to be correct,
2. It's hard to see how the example of a toy train being
moved by precession of a gyro at the end of
Laithwaite’s paper, Mass Transfer (which he
undoubtedly checked with a real physical model)
could only depend on the flywheel becoming less
gravitationally massive because of its rotation, and
3. R. Wayte's paper [2] uses successfully (and in a very
important way) the equivalence principle from
General Relativity in his analysis of his gyro without
a fixed pivot, and, of course, General Relativity
contains the assumption that inertial and
gravitational mass are the same.
2. Neo-Newtonian Theory
To begin, it’s necessary to include the time derivative of
the vector acceleration in the functional dependence of the
fractional mass change. Here we are loosely following
Jackson [4] who also uses the time derivative of the vector
acceleration in his treatment of “radiative reaction”.
However, we do not introduce the time derivative of the
vector acceleration explicitly as does Jackson in view of the
fact (as he mentions) that such a course is fraught with
mathematical difficulty and leads to undesirable “runaway
solutions” in even the simplest cases. Further, Jackson
points out that “The equation can be criticized on the
grounds that it is second order in time, rather than first,
and therefore runs counter to the well-known requirements
for a dynamical equation of motion.”, where he is referring
to the second time derivative of the velocity vector, not of
the position vector. (see page 784 of [4]).
17
Now, to continue our analysis, we consider a (neutral

point) particle of (rest) mass m and assume its inertial mass
changes in a rotational situation. We aim to keep as much
of classical (Newtonian) mechanics as possible with his
three laws in their entirety except that we slightly change
his second law to the (original) more general law
conserving momentum (which reduces to the modern
version of Newton's second law if dmr/dt ≡ 0; that is, if mr
is constant in time at the time in question):
dmr
F v  mr a (2.1)
dt
where mr is the (inertial) "r-mass" of the particle, which --
for all we know -- is given by
mr = f(x, v, a, da/dt, m) (2.2)
and x, v, a, da/dt, and m are, respectively, the position
vector, the velocity vector, the acceleration vector, the time
derivative of the acceleration vector, and the rest mass of
the particle in question, where we tentatively omit the
higher order than the third time derivatives of the particle
position vector in order to get started here … although they
very likely do come in to some extent … for high tech
purposes anyway. (Note that rest inertial mass is already
known to change with velocity contrary to Newton's
assumption, but we tentatively restrict our self here to non-
relativistic velocities.) We first observe that it seems to be
that we must remove m from the arguments of f and write
mr = f(x, v, a, da/dt) m (2.3)
since the linearity of force with respect to mass is, we
believe, fundamental to Newtonian mechanics; this will, of
course, have to be experimentally verified, but it appears to
follow directly from dimensional analytic considerations
(see below in section 3). Then we note that both x and v
must be removed from the arguments of f since (1) the
inertial phenomena don't depend on x because if we move
our inertial system by a simple translation, that won't affect
18
things and (2) if we move our inertial system to a point

moving with a constant (non-relativistic) velocity with
respect to our original system, then that can't affect inertia
either since consider elementary spring experiments and
changing from a given inertial system to a new one moving
at a constant (vector) velocity with respect to it … as then
in equation (2.44) cos(α) = ± 1. Note that in case (2), the
acceleration vector and its time derivative do not change
since the position change per unit time is assumed
constant. In addition, we also assume that forces on a
particle add as vectors as in Newtonian mechanics. Now,
it might be thought that in view of the above generalization
of Newton's second law that we would go on to assume
that
Fg = G m = mr ag + (dmr /dt) v (2.4)
where Fg is the gravitational force on the (point) particle

due to a (Newtonian) gravitational field, where for a
particle of rest mass m , we take G to be given by
m G  x  x 
G  x  x   3
(2.5)
x  x
at a point particle whose rest mass is m, its r-mass is mr ,
and its position vector is x , which is due to a particle of
rest mass m whose position vector is x with G not being
the magnitude of G , but rather a (very small) gravitational
constant; however, since gravitational force in the case of a
point particle is directly proportional to gravitational mass
which we take equal to rest inertial mass here (as mentioned
just above), we consider an inertial system that is in free fall
… in which case (2.4) will not apply, but (2.5) will continue
to apply. (The author obtained this idea of treating
gravitational forces by going to an inertial system in free
fall from his old friend T.E. Phipps [18].) Note that the
19
gravitational force on a particle would evidently not be

influenced by its being in circular motion in view of our
neo-Spearsian gravitational theory [12] which shows that
gravitational forces between two atoms are a consequence
of the large proton-electron mass ratio in that it is a higher
order electrostatic effect, and the Newtonian gravitational
constant G is obtained from electromagnetic theory and its
various constants. But it might be objected that our
analysis here is tainted in view of the fact that we only
consider the gravitational force between fixed atom pairs,
and here we are talking about rotational motion. However,
it is a well-known fact that in the case of two charged
particles moving at very non-relativistic velocities with
respect to each other, that the Coulomb forces dwarf the
other (magnetic) induction forces. Thus such an effect
would be expected to be small.
We retain Newton's third law (action equals reaction).

Note, for example, that the acceleration due to gravity of a
point particle need not have magnitude equal to the usual g
on the surface of the earth if mr ≠ m, but if mr = m at the
time in question, then they do have the same magnitude, of
course [11].
We might remark that Herbert A. Simon in [6] discusses

his operationalist axiomatization of Newtonian mechanics
in which time and position together with the first and
second time derivatives of the position (vector) are
primitive terms (as they are the observables) and then mass
and force are defined in terms of them, while we
additionally use the third time derivative of the position as
an (observable and hence) primitive term. Then we define
rest mass much as does Simon (without using the word
“rest”), and then utilizing a “continuity” assumption we go
on to define the mass change function f to be a function
of only the second and third time derivatives of position
(both observable and thus primitive). Finally, we conclude
20
by defining (generalized) force to be the time derivative of

the momentum is an inertial system [20] (as did Newton
also); where, since inertial mass varies, this introduces a
term in the first time derivative of position and where we
then have a point particle theory in which extended objects
are treated by summation or integration (over elements of
mass volume) in as much as the generalized force is
assumed to be (vector) additive as in classical mechanics.
And we retain Newton’s three laws in their original form.
Next, we note that our trouble generalizing Newton's

physics concerning systems of particles will always and
everywhere come from the fact just mentioned that we
have to generalize his second law for changing mass. Thus,
for example, in Hay's Chapter 3 on page 62, "Application
of Vectors to Mechanics", (Hay follows Synge and Griffith
[8] fairly closely, and references them in his discussion of
principle axes of inertia) everything goes well up to Section
31: Newton's laws. Then, on page 72, we come to his
discussion of the motion of a system of particles beginning
in Section 35: The center of mass of a system of particles.
We define the center of r-mass of a system of particles
having rest masses m j , r-masses m rj , and position vectors
x j by
N
m x
1
x rc  rj j (2.6)
mr
j 1
where
N
mr  m
j 1
rj (2.7)
21
which, of course, reduces to the usual center of mass xc in

the case where all mrj  mj for all j.
The usual discussion of the moments and products of
inertia of a system of particles needs only be changed by
replacing m j everywhere by m rj , where we note that these
new r-moments and r-products of inertia will, in general,
depend on all the variables of our mass change function f.
Concerning the Kinematics of a Rigid Body, everything

goes as usual, where we begin with the discussion of linear
and angular momentum and where, of course, we expect
trouble as mentioned above. We define the vector r-
momentum of the system of particles as
N
Mr  m v
j 1
rj j (2.8)
and then we would like to have
Mr  mr vrc (2.9)
where
dxrc
vrc  (2.10)
dt
but this does not seem to hold in general. However, we do
have
d  mr x rc 
N

d mrj x j  N
 dmrj 
dt
 
j 1
dt 
j 1
dt
x j  mrj v j 


N
dmrj
 
j 1
dt
x j  Mr
(2.11)
and also
22
d  mr xrc  dmr
 xrc  mr vrc (2.12)
dt dt
Thus we have
Mr = mr vrc + dmr/dt xrc - ∑𝑁𝑗=1 𝑑𝑚𝑟𝑗 /𝑑𝑡𝐱𝑗 (2.13)
Hence, a necessary and sufficient condition for (2.9) to

hold is evidently that the last two terms of (2.13) sum to
zero.
Next, we turn our attention to the angular momentum

of a system of particles which form a rigid body rotating
about a point O, which is fixed in the frame of reference S,
and we generalize this so that the r-angular momentum h r
of a system of particles constituting a rigid body about a
point O which rotates with an angular velocity ω having a
line of action which passes through O is given by
 m x  ω  x   m ωx 

N N
hr  rj j j rj j
2

 xj xj ω (2.14)
j 1 j 1
where x j is the magnitude of x j . Thus, the components

of h r in a rectangular Cartesian system with origin at O are
given by
hr 1   Ir 11  Kr 32  Kr 23
hr 2  Kr 31  Ir 22  Kr 13 (2.15)
hr 3  Kr 21  Kr 12  Ir 33
where Ir 1 , Ir 2 , and Ir 3 are the r-moments of inertia of the
particles about the rectangular Cartesian i1 , i 2 , and i 3
axes (which have O as origin but whose directions need not
be fixed in S) and where K r 1 , K r 2 , and K r 3 are the r-
products of inertia with respect to the three coordinate
23
planes, taken in pairs. We note (as mentioned above) that

these r-moments and r-products of inertia are defined as
usual (see, for example, [1, p.74]) except that one replaces
the inertial mass m j everywhere by the r-mass m rj .
Now, we consider a rigid body moving in a general
fashion relative to a frame of reference S. Let us introduce
coordinate axes whose origin is at the center of r-mass of
the body which we denote by C r , where the directions of
these axes need not be fixed in the body. We may consider
the body to have a velocity of translation vrt plus an
angular velocity ω about a line through C r . Then we see
that the velocity of the jth particle can be expressed in the
form
v j  vrt  ω  x j (2.16)
and so the r-angular momentum about the center of r-mass

having position vector xrc is given by
N N
hr  m x   v
rj j rt  ω xj    mrj x j  vrt   hr 0 (2.17)
j 1 j 1
where hr0 is just the term on the right of equation (2.18)

below. But the first sum is just mr xrc by definition of xrc ,
and since the origin of the rectangular Cartesian system is
assumed to be at the center of r-mass, xrc  0 and
N
hr    m x  ω  x 
j 1
rj j j (2.18)
which does not seem to use the above necessary and

sufficient condition for Mr  mr vrc . The right side of this
equation is the same as that of our formula for h r displayed
above which means that the equations for hr 1 , hr 2 , and
hr 3 obtained following that formula may be used for the
24
determination of the Cartesian components of the r-

angular momentum h r of a rigid body about either a fixed
point O in the body or the center of r-mass of the body. In
the two cases the origin of the coordinate axes is at O and
C r , respectively, the directions of the coordinate axes being
quite general.
By definition (2.8), we have
𝐌𝑟 = ∑𝑁
𝑗=1[𝑚𝑟𝑗 𝐯𝒋 ] (2.19)
so that
N
 dmrj 

dMr (2.20)
 v j  mrj a j 
dt dt 
j 1 
but
dmrj
Frj  mrj a j  vj (2.21)
dt
and so
N
F
dM r
 rj  Fr (2.22)
dt
j 1
as desired if there are no internal forces between the

particles. If Fjk is the internal force exerted on the jth
particle by the kth particle, then by Newton’s third law we
have
Fjk = - Fkj (2.23)
And by his second law since we assume with him that

forces add as vectors
25

d mrj v j  F N
dt
rj  F
k 1
jk (2.24)
and so
dMr
 Fr (2.25)
dt
since we retain Newton's third law and so
N N
 F
j 1 k 1
jk 0 (2.26)
jk
Thus r-momentum is conserved, but this is quite different

than classical momentum conservation, in general. And so
neither conservation law implies the other. (Hence inertial
propulsion devices are not ruled out here.)
Further, since
hr = ∑𝑁𝑗= 1 𝑚rj xj × vj (2.27)
and so
dhr
 Ar  B (2.28)
dt
where
N
 dx j 
Ar   m
j 1
rj
dt
 vj   0


(2.29)
dx j
since  v j , and
dt
26
  
N  N 
B  x F 

j 1 
j
 j
 
Fjk    G  H (2.30)
 k 1
k j

  
where we (ambiguously) use G here as not being any

gravitational force, but instead we follow Hay’s notation
for clarity:
N
G x
j 1
j  Fj  (2.31)
and
N N N N
H  x  F    x
j 1 k 1
j jk
k 1 j 1
k  Fk j  (2.32)
k j jk
where we note that G is equal to the total moment about

O of the external forces. We have
N N
2H    x
j 1 k 1
j  Fjk  x k  Fk j  (2.33)
k j
But it follows from Newton's third law that Fjk  Fkj so

that
N N
2H    x j 
 x k  Fjk  (2.34)
j 1 k 1
k j
but the last cross product vanishes as x j  x k and F jk are

parallel in view of Newton's third law. So H  0 and hence
27
x  F
dhr
 j j (2.35)
dt
j 1
and the rate of change of the angular momentum of the

body about O is equal to the total moment about O of the
external forces in the general case too.
Now, this is fine as far as it goes, but we need to relate
it to the center of (inertial) r-mass because we want to have
all motions of a rigid body in terms of this center of r-mass
plus a rotation about it, and in view of (2.13) this will
involve some difficulty. Let us consider now the case
where the coordinate system is at the center of r-mass Crm
which, however, is moving with velocity vrc which is a
function of time. Then the total velocity of the jth particle
vtj = vrc + vj and since we have
hrc = ∑𝑁𝑗= 1 𝑚rj xj × vtj (2.36)
then the time rate of change of the r-angular momentum

about Crm is:
dhrc/dt =
𝑑𝑚𝑟𝑗
∑𝑁
𝑗=1 𝐱𝑗 × 𝐯𝑡𝑗 + ∑𝑁 𝑁
𝑗=1 𝑚𝑟𝑗 𝐯𝑗 × 𝐯𝑡𝑗 + ∑𝑗=1 𝑚𝑟𝑗 𝐱𝑗 × 𝐚𝑡𝑗
𝑑𝑡
𝑑𝑚𝑟𝑗
= ∑𝑁𝑗=1 𝑑𝑡
𝐱𝑗 × 𝐯𝑡𝑗 − ∑𝑁 𝑁
𝑗=1 𝑚𝑟𝑗 𝐯𝑡𝑗 × 𝐯𝑡𝑗 + ∑𝑗=1 𝑚𝑟𝑗 𝐯𝑗 × 𝐯𝑟𝑐 +
∑𝑁
𝑗=1 𝑚𝑟𝑗 𝐱𝑗 × 𝐚𝑡𝑗 (2.37)
(the last equality by substituting using the equation

embedded in the text in the line just above (2.36) when
solved for vj after setting vj × vtj = -vtj × vj ), and so the
third to last term on the right vanishes identically. Also,
from equations (2,20), (2.21), and (2.31), we have
𝑑𝑚𝑟𝑗
G = ∑𝑁𝑗=1 𝑚𝑟𝑗 𝐱𝑗 × 𝐚𝑡𝑗 + ∑𝑁𝑗=1 𝑑𝑡
𝐱𝑗 × vtj (2.38)
28
And, in view of (2.20) and (2.22), we have
dhrc/dt = G + Mr × vrc (2.39a)
Thus, using (2.13), if (both sums being over the index j)
xdrc = (∑ dmrj/dt xj)/(dmr/dt) (2..39b)
is the center of r-mass change with
dmr/dt = ∑ dmrj/dt (2.39c)
being the total r-mass change in the system of particles,
then
dhrc/dt = G + vrc × ((dmr /dt ) xdrc). (2.39d)
Thus, since G is the moment of exterior r-forces about O,

we obtain a similar result to the result that r-angular
momentum is conserved, but the last term on the right
need not vanish. And note that this result we have just
obtained is quite different from the classical angular
momentum conservation result and neither implies the
other, in general. Thus, classical conservation of angular
momentum need not hold if variable inertial mass is
significant.
Note that Hay’s (usual) discussion (on pages 79 – 80)

about a rectangular Cartesian coordinate system having a
fixed point O as its origin and rotating with angular velocity
ω all holds true in our theory also. Thus we have for an
arbitrary time differentiable vector field b
db/dt = δb/δt + ω × b
29
with
δb/δt = db1/dt i1 + db2/dt i2 + db3 /dt i3 (2.40)
where ω is the angular velocity vector of the rotating (about

its origin) orthonormal system.
Finally, let us consider the function for a point particle

of rest mass m and position vector x relative to an inertial
system S:
 da 
f  f  x , v, a , , m   mr (2.41)
 dt 
where we have explained above that x , v , and m need to

be taken out of the arguments of f and we also need
 da 
mr  f  a , m (2.42)
 dt 
However, a vector is uniquely determined by its

magnitude and direction, so we may write
 da a da dt 
f  f  a, , ,  (2.43)
 dt a da dt 
where we denote (by an abuse of notation) the magnitude

of da dt by da dt . But by symmetry, only the angle 
between a and da dt can matter in f, so we write
 da 
f  f  a, ,  (2.44)
 dt 
where  is the angle between the acceleration vector and

its time derivative and hence f may be taken to be an even
function of  ranging from  to  . And we must have
30
da
f  1 as 0 (2.45)
dt
throughout a non-trivial interval of time t as then v
approaches a constant vector … which must then vanish
(see the above discussion of f and inertial systems in this
section 2).
Now, the alert reader has doubtlessly noticed that it

follows from our using Newton’s original definition of
force as being the time derivative of the momentum that
force is an invariant in our theory because the r-mass of a
particle in not assumed a function of its velocity at non-
relativistic velocities, in general.
Note that all electrodynamic forces between two

(possibly) charged particles, which are moving with non-
relativistic velocities with respect to each other, are
according to Jackson [4] (section 6.1) relativity theories in
as much as Maxwell’s equations are then approximately
Galilean invariant; and Wesley’s [9] equation (5.1) for
induction says this too (even though it gives different
numbers than the just mentioned Maxwell theory).
Further, as we mentioned after (2.45), then it follows that
gravitational force between two particles, being higher
order electrostatic in nature, is a Galilean relativity theory
too. In fact, we believe all matter and forces resulting from
matter interacting with matter are fundamentally
electrodynamical in origin [13] just as Poincare thought.
Our plan is to thoroughly analyze Hay's Example 3 (that

we reproduce in full here for those who do not have Hay’s
book): A gyroscope is mounted so that one pivot point on
its axis is fixed, and we investigate those motions in which
the axis of the gyroscope makes a constant angle with the
vertical (i.e., no nutation). Hay treats this example using
Euler's equations of motion completely obtaining the
31
equations of motion and, of course, these equations

determining the motion of the gyroscope have doubtlessly
been thoroughly verified. Hence we must obtain them very
nearly in analyzing it using our revised neo-Newtonian
theory, as well as any weight loss data from
experimentation, which will impose certain conditions on
our function f  mr m and hopefully suggest a good trial
function for it. It appears from Wayte's paper that some
weight loss does occur in the case of a gyro not having a
fixed pivot even in pure Newtonian theory, but we will see
later that this won’t also be detected in the fixed pivot case
during steady precession without nutation. Of course, we
have had to tamper with that part of Newtonian mechanics
which assumes the inertial mass of a particle is constant in
time but otherwise pretty much left it alone. We think our
neo-Newtonian theory is now in its final form but "There's
many a slip between cup and lip!" Now for our fixed pivot
gyro analysis.
3. Application to a Slowly Precessing Gyro
Now it's time to consider Part One of our neo-

Newtonian theory in the case of Hay's gyro (but having a
washer shaped disc rotor) with a fixed pivot point. In Part
one, we considered that we were in an inertial system. In a
(so-called) inertial system on the surface of the earth, we
have to make the change that in f we write
f = f(|a + G + ah|, da/dt, α) (3.1)
where G  G x  is the gravitational field strength at the
particle having position vector x as explained in Part one,
where ah is the constant deceleration of the precession
angular velocity, where the magnitude of da dt , which we
denote (by an abuse of notation) as da dt does not change
since dG(x)/dt ≡ 0 on the earth's surface, and α is now the
32
angle between (a + G + ah) and da/dt. Now we can

proceed.
Consider the case of a gyroscope (see the coordinate

diagram at the end of the last chapter) with a thin washer
shaped fly-wheel and a fixed pivot point, that is, let the
situation be pretty much as in our diagram at the end of
our Chapter 3 (but using a thin disc rotor that is washer
shaped) and assume that we have entered rectangular
Cartesian coordinates î1 , î2 , and î3 such that î3 is at an
angle  with the vertical, î1 is at right angles to î3 (both
pointing towards the sky, but not usually vertical) where
both î1 and î3 are in the plane of the page, and the other
unit vector i2 completes the triad and is, of course,
horizontal and out of the page. We introduce a constant
unit vector ĵ in the direction of the vertical so that it makes
a (tilt) angle of  with î3 . Then we assume that the plane
containing ĵ and î3 rotates at a slow rate of p radians per
unit time about the former so that p is the precession about
the vertical. Now
ˆj  sin iˆ  cos iˆ (3.2)
1 3
Let O be the origin of the triad and let C be a point on

the positive î3 axis with OC being of length l. Let O be
the origin of the triad and let C be a point on the positive
î3 axis with OC being of length l (lower case L). Suppose
that a thin and washer shaped disc of uniform thickness s
and outer and inner radii r and r’ is mounted on a shaft of
negligible diameter and mass which is along the î3 axis
with its bottom end at the point O and such that the disc's
33
center is at the point C. Then we have (where φ = ω t) that

if r’ ≤ ρ ≤ r, the typical point on the disc at radius  is
given by the equation [note that two different symbols for
the Greek letter rho appear here; this is a Microsoft Word
problem due to using different versions of that word
processing program]:
x   sin  iˆ1   cos iˆ2  l iˆ3 (3.3)
dx
 v   cos iˆ1   sin  iˆ2 (3.4)
dt
dv
 a    2 sin  iˆ1   2 cos iˆ2 (3.5)
dt
da
   3 cos iˆ1   3 sin  iˆ2 (3.6)
dt
Then since on the surface of the earth, G  g

a + G = a – g j = (-g sin θ – ρω2 sin φ) i1 - ρω2 cos φ i2
- g cos θ i3 (3.7)
and so we have with ah = h i2 since the rotor disc’s center

C moves in a circle lying in a horizontal plane (no nutation)
as the gyro precesses and so is parallel to i2 where the
orthonormal system rotates with the gyro as it precesses:
a + G + ah = (-g sin θ – ρ ω2 sin φ) i1 + (h - ρ ω2 cos φ) i2

- g cos θ i3 (3.8)
whence
(a + G + ah)2 = (-g sin θ – ρ ω2 sin φ)2 + (h - ρ ω2 cos φ)2
+ (-g cos θ)2 (3.9)
34
But
dt
 1 3 
ˆj  da  sin  iˆ  cos iˆ    3 cos iˆ   3 sin  iˆ
1 2 
   3 cos sin 
(3.10)
Thus since a • da/dt = 0, we have:
cos(α) = [(a + G + ah) • da/dt] / [|a + G + ah ||da/dt|]
= (-g sin θ i1 + h i2 - g cos θ i3) • (- ρ ω3 cos φ i1 + ρ ω3 sinφ
i2)/ [(ρ ω3 ) {ρ2 ω4 – 2 ρ ω2 (-g sin θ sin φ + h cos φ) + (g2 +
h2)}1/2]
= [ g sin θ cos φ + h sin φ] / {ρ2 ω4 – 2 ρ ω2 (-g sin θ sin φ
+ h cos φ) + (g2 + h2)}1/2
= [ g sin θ cos φ + h sin φ] / ρ ω2 (nearly) for ω >> -h, g, & ρ
but rather
= sin φ (nearly) if – h >> ω, g, & ρ (3.11)
Note that, in both cases, cos(α) is (nearly) independent

of the shaft length l since we are assuming slow precession
angular velocity. Also note that should ω < 0, then the
algebraic sign of da/dt will have the opposite sign than
when ω > 0, and so the angle α will also have a sign reversal;
but cosine is an even function and so cos(α) will be
unaffected. Thus, only (ρω3) will undergo a sign change
depending upon whether ω is positive or negative;
however, (ρω3) is set above to da/dt ≥ 0, meaning that
equation (5.2) above needs a reversal of sign for ω < 0
35
when we are in the steadily precessing or steadily precessing

followed by stopping dead gyro cases discussed above.
Alternatively, we can replace ω by its absolute value
throughout our gyro analyses herein.
Remark (3.1): It might be objected here that the

acceleration – g j of gravity on the typical flywheel particle
should be omitted in the discussion above on the grounds
that this particle of a non-nutating gyro does not
experience this acceleration by definition of nutation;
however, experiment shows that this is not the case … in
that a gyro with a steadily spinning rotor that is not
precessing, one with a fixed pivot, and one with a non- zero
tilt angle θ, when allowed to move freely, first dips as the
gravitational force takes effect, but then it begins to precess
and also to nutate slightly as the dipping rebounds to rising
and then to dipping again and so on [10]. Thus the
gravitational force is felt! And, in fact, such a spinning gyro
whose shaft is supported so that the rotor does not feel
gravity fails to precess! This situation reminds the author
of the concept of virtual work in that gravity exerts a sort
of virtual acceleration on the rotor particles. (Alternatively,
we can go to a true inertial system that is in free fall … with
respect to which the gravitational force does result in this
acceleration g being exhibited for each point particle
belonging to the rotor using the equivalence principle [2].)
Finally, we note that our notion of force as the time
derivative of momentum requires that are taking this time
derivative relative to an inertial system as in Newtonian
mechanics [6].
Note that we are neglecting the precession rotation of

the gyro here on the grounds that the precession p is
assumed small. (See pages 688-691 of [10] for an
illuminating analysis of such a horizontally precessing fixed
pivot gyro without nutation, which analysis includes the
36
precession angular velocity as well … assuming the validity

of Newtonian mechanics, of course).
Our next dimensional analytic arguments lean heavily

upon [15] and especially on Chapter 3, and the author
strongly recommends this work. Now (forgetting about
gyroscopes for a moment) we assume that our function f =
f(a, da/dt, cos(α)) is well behaved and expand it in a Taylor's
series
f = 1 + k1 a + k2 da/dt + k3 cos(α) + higher order terms.

(3.12)
Note that cos(α) is not defined when either these two

arguments of f vanish, but since we must have all terms of
the above expansion of it are of the form [am1 (da/dt)m2
cos(α)m3] with each of the mj being an integer at least unity,
any numerical value of absolute value between one and
zero (including both one and zero) may be assigned to
cos(α) without any problems. And that all terms must be
of this form follows from [25] (see Figure 2) since if a
spinning rotor has a vertical axis, then the gravitational
force is uncoupled from the spinning of the rotor.
Further, let us consider the matter of units and dimensional

analysis which, surely, we don’t want to abandon at this
point – it having served the physics community well for
hundreds of years!
Now our function f is dimensionless since it equals mr/m.

Thus the question comes up: how are we going to make
our formula for f dimensionally correct? Each additive
term must be dimensionless. We evidently have no choice
but to introduce for each such term a constant with the
appropriate dimensions to force this. Thus we write
37
f = 1 – ∑ b(m1, m2, m3) am1 (da/dt)m2 cos(α)m3 . (3.13)

Note (again) that if either m1 or m2 were negative, then we
would be in a singular situation at “a” or “da/dt” vanishing,
so we rule this out. Also, if either vanished, we would be
in a discontinuous situation unless f was a trivial function
of α, so we can additionally rule out this by experimental
data showing the Laithwaite effect being (primarily) a
rotational one, whence f depends nontrivially on the angle
α!
Now, we have computed above cos(α) (for Hay's gyro)

and we plan on using it (God willing) as follows. Since –π
≤ α ≤ π, f is an even function of α. And since f (α) is
evidently a periodic function of α having period 2 , and
since f is assumed well behaved bounded and possessing
continuous partial derivatives of all orders and being
analytic at α = 0 [and hence at cos(α) = 1, cos(α) taken as a
general variable] with a radius of convergence (talking
about α now as the angle between a and da/dt) at least 
and also converging at  and - π, we can truncate the
formula (3.11) as follows: suppose ε > 0, and let
RN = the terms of (3.11) of order higher than
N = m 1 + m2 + m3 (3.14)
Suppose further that N is the smallest positive integer such
that
| RN | < ε (3.15)
for all  [which must exist as equation (3.13) is assumed
absolutely convergent where needed and is assumed well
behaved]. Thus we can replace f by [ f - RN ] if we choose
ε sufficiently small so that we obtain sufficient accuracy.
If  is large, we have approximately [in view of (3.11)]:
38
f = 1 – ∑ b(m1, m2, m3) (ρ ω2)m1 (ρ ω3)m2 [( g sin(θ)
cos(φ) + h sin(φ)) / ( ρ ω2)]m3 (3.16)
But if – h >> ω, then we have instead
f = 1 - ∑ b(m1, m2, m3) (-h)m1 (ρ ω3)m2 sin(φ)m3 . (3.17)
also by equation (3.11).
Now it will be shown to follow from the assumption

that the r-products of inertia vanish if the (Newtonian)
products of inertia vanish is a physical situation (where the
r-products of inertia are defined in the same way as the
products of inertia except using the r-mass instead of the
mass of the system particles, see below) that the positive
integer m3 must be even, and it will follow from [11] that
m1 & m2 must be such that both
0 ≤ 2 m1 + 3 m2 - 2 m3 ≤ 2
and
m1 + m 2 – m3 ≥ 0 . (3.18)
The latter is because lim(ρ f) as r  0 must itself be zero if

r is the outside radius of Prof. Dmitriev’s falling rotors
(hollow cylindrical objects of outside and inside radii r and
r’ and spinning about their axes of symmetry, which axes
are assumed horizontal [11]) and since ρ is the Jacobian for
cylindrical coordinates. The former is because the last 28
points of Prof. Dmitriev’s 56 data points (i.e. the ones with
the largest spin) can be well least squares fitted to a
quadratic thereby eliminating the noise in his data; this then
means that each individual term of f must be either directly
39
proportional to ω2 or ω, or else not a function of ω. (We

choose these particular 28 data points because in [7] it is
remarked that the rotor spin angular velocity ω must be at
least about 100 times the precession angular velocity p or
else the thrust effect drops off, but that otherwise, the
thrust is more or less independent of the spin ω; however,
now, we recalculate using all these points together with 57
others obtained by noting that only the magnitude of the
spin is relevant since we should have left-right symmetry in
spin here. See Appendix 1 at the book’s end.) Note also
that the larger the rotor spin ω, the smaller the gyro
precession p in view of equation (1.1) ... all else being equal.
4. The Flywheel
Now, we are finally in a position to compute the r-mass
mr of the flywheel. We assume (- h) is small enough so that
we can set it to zero in (3.16). Now consider an element of
its volume
dV  s d d (4.1)
where s is (ambiguously, as in (1.1) s denote spin instead)
the very small thickness of the cylindrical flywheel having
outer radius r and inner radius r’ and also  is between r’
and r. Let  be the (constant) mass volume density of the
flywheel material when at rest so λ = m / [π (r 2 – r’ 2) s].
Then we have because m = π (r 2 - r’ 2) s λ, that (holding
time t fixed in order to sum over the gyro rotor geometry
only)
d(mr) = λ dV f = [m / (π (r 2 - r’ 2) s)] (s ρ dφ dρ f) =
= (m ρ dφ dρ / (π ( r 2 - r ’ 2)) {1 – Σ(ρ ω2)m1
(ρ ω ) 2 b(m1, m2, m3) am1 (da/dt)m2 cos(α)m3 }
3 m (4.2)
Hence, because
40
𝑟 𝜋
{m / [π (r 2 – r’ 2)]} ∫𝜌=𝑟 ′ ∫𝜑=−𝜋 𝜌 𝑑𝜌 𝑑𝜑
= {m / [π (r 2 – r’ 2)]}[2 π [(r 2 - r’ 2) / 2] = m (4.3)
we have
mr =m – {m / [π ( r 2 - r’ 2)]}{Σ b(m1, m2, m3) ∫ ∫ ρ
(ρ ω2)m1 (ρ ω3)m2 cos(α)m3 dρ dφ} (4,4)
where we utilize the usual limits of integration in the two

integrals. The key integral, with t = m1 + m2 + 2 – m3,
contains the factor
𝑟
∫𝜌=𝑟 ′ 𝜌𝑡−1 𝑑𝜌 = (r t – r’ t) / t if t ≠ 0, (4.5)
where we temporarily use “t” to denote an integer, not
time; and we also do this in (4.9), just after (4.14), (4.15),
and (4.17) below … by an abuse of notation.
Otherwise, we have
[ln(r) - ln(r’ )] = ln[r / r’ ]. (4.6)
But then if r’ vanishes, this term is infinite and so it is non-

physical and must be omitted.
Now consider the first case where we have the integral

 cos  d
k
ck = (4.7)

If k = 0 then it is evidently equal to 2 . If k  0 and

even, then the integrand is positive except at     2 , and
so the integral is a positive constant. Note also (by the way)
that if we consider
41
𝜋
∫𝜑=−𝜋(𝑐𝑜𝑠 𝑘1 𝜑𝑠𝑖𝑛𝑘2 𝜑)𝑑𝜑, (4.8)
and if k2 is even, then the sink2(φ) term can be expanded in

even powers of cos(φ), and so should k1 also be even, it
follows that each term in this expansion will be non-zero
by the just above. On the other hand, if k2 is odd, then
using the identity sin2(φ) = 1 – cos2(φ), one can assume that
k2 = 1, then the substitution y = cos(φ) yields immediately
that (4.8) vanishes. Thus the integral (4.8) does not vanish
if and only if k1 and k2 are both even. Now, clearly, from
equation (4.7), we have:
mr = m –{m / [π (r 2 – r’ 2) ]} ∑ b(m1, m2, m3) cn [g sin(θ)]m3
ω2m1+3m2- 2m3 ((r t – r’ t) / t),
where
t = m1 + m2 – m3 + 2 and n = m3. (4.9)
We make the convention that we do not divide by zero any

term in the sum; and we note that if h = 0, then we have
that mr is nearly independent of the gyro shaft length l as
well as the precession angular velocity p. Note also that we
can show that
dmr/dt ≡ 0 (4.10)
since φ = (ω t) does not appear in (4.9) and as the rotor spin
angular velocity ω is assumed nearly constant.
Now, since we have φ = ω t:
d(dmr)/dt = {m / [π (r 2 – r’ 2}{𝜌 𝑑𝜌 𝑑𝜑(𝜌𝜔2)𝑚1 (𝜌𝜔3)𝑚2

Σ b(m1, m2, m3) [[g sin(θ) cos(φ)/(ρ ω2)]m3 (m3 d(cos(ω
t))/dt /cos(φ))]} =
42
={m / π (r 2 – r’ 2)]} {𝜌 𝑑𝜌 𝑑𝜑(𝜌𝜔2 )𝑚1 (𝜌𝜔3 )𝑚2 Σb(m1, m2,

m3) [- m3 ω sin(φ)/cos(φ)] [g cos(φ) sin(θ)/(𝜌𝜔2 )]m3. (4.11)
In Part One, we obtained a formula in terms of the r-

moments and r-products of inertia for the angular
momentum of a system of particles which formed a rigid
body both under the assumption that the rotation is about
a fixed point O in the body or about the center of r-mass
of the body C r , where in these two cases the origin of the
coordinates axes is at O and C r (which we have shown
above is here just C, the center of rest mass), respectively,
the directions of the coordinate axes being quite general.
This is all that's needed to obtain our generalized Euler's
equations of motion under the assumption that all three of
the r-products of inertia vanish, where these following
three equations are almost (see below) the same as in our
complete discussion of Hay’s gyro with fixed pivot except
that his G becomes our G ' , and the I j moments of inertia
are replaced by the I rj r-moments of inertia for j  1,2,3 ,
and further Ω  pj with p being the precession angular
velocity in view of equations (2.15) & (2.36). Thus Euler’s
equations of motion have the r-generalized form [note that
should the last two terms of (2.13) not sum to zero, then
this sum cross vrc must be added to G in order to obtain G′,
which equals the right hand side of equation (2.39), and in
view of (2.38) … where we note that the term Mr ×vrc in
(2.39) vanishes here since, in the gyro rotors treated herein,
Mr = mr vrc (and vrc × vrc = 0) provided there is no nutation
and rotor spin remains constant]:
43
Ir1 d(ω1)/dt + ω1 d(Ir1)/dt – Ir2 ω2 Ω3 + Ir3 ω3 Ω2 = G′1

(4.12)
in our Neo-Newtonian theory, where we are considering
the three components of the vectors ω , Ω, and
G' 
 x  dF
V
r (where V is the flywheel volume), G' is the
total moment about O of the external forces Fr , [see (2.21)

and (2.39)]. But we can compute the r-moments as follows
in view of our formula (3.3) for x  x1i1  x2i2  x3i3
Ir3 = ∫𝑓𝑙𝑦𝑤ℎ𝑒𝑒𝑙 𝑣𝑜𝑙𝑢𝑚𝑒 𝑉[(𝑥12 + 𝑥22 )𝑑𝑚𝑟 ] 𝑑𝑉 (4.13)
And one gets the integrals for j = 0, 1, 2, …


 cos
2j
c2j =  d (4.14)

are the same positive constants that appeared in our

development of our formula for mr from integrating dmr
over the flywheel. The only difference between our
expansion for mr and our forthcoming expansion for Ir 3
will be that we get the new integral [instead of (4.6) where
t is given as just before (4.5)], that is, we have (4.5) with t
= m1 + m2 + 4 – m3 ≠ 0, and m3 is even.
Now, we continue to consider first the case where h =

0, that is, where there is no deceleration, and so the gyro is
precessing steadily at angular velocity p. Thus in view of
(4.4) we have
Ir3 = [(r 2 + r’ 2) m / 2] – [m / (π (r 2 – r’ 2))] ∑ b(m1, m2,

m3) [g sin(θ)]m3 (r t – r’ t) / t] (ω2m1 + 3m2 + 4 – 2m3) cn, (4.15)
44
where we have (again) set h = 0, where n = m3, and where

t = m1 + m2 - m3 + 4.
Thus d(Irj)/dt ≡ 0 for j = 1,2,3 [see also equation (4.10)],

whence we obtain the same Euler’s equations as Hay’s
example of a gyro with fixed pivot ... except that m
becomes mr and Ij becomes Irj for j = 1, 2, 3. [That is, the
second terms on the left in our generalized Euler’s
equations (4.12) all vanish.]
Now, Prof. Dmitriev’s free falling and spinning rotor

data [11] appears to show (m - mr) being slowly decreasing
on the average (but oscillating) as ω increases in magnitude,
where mr is the r-mass. Note that we have {mr g’(ω)} = {m
g}, where g’(ω) is the observed free falling acceleration in
the earth’s gravitational field in Prof. Dmitriev’s
experiment, and is a function of ω which is the spin angular
velocity in radians per second, but his paper uses the unit
of Hertz, not radians per second. [And, of course, g’(ω)
would depend on r and r’ (the outside and inside radii of
the hollow cylinder used) as well.] Note that the left term
{mr g’(ω)}is Newton’s second law [the dmr/dt containing
term does not appear as explained just below equation (2.5)
as this term is due to gravity and the right term {m g} is the
force of gravity according to Newton as m is the rest (equals
gravitational) mass (see page 36 above) and g is the
acceleration due to gravity for point particles at the earth’s
surface. More generally, in examples not involving rising
or falling of the r-center of mass, we need in our gyroscopic
work (with the usual moving triad coordinate system) to
replace the addition to the total acceleration sans gravity of
the vector (0, 0, -g) by the addition of (0, 0, -g (m /mr)) = (0,
0, (-g / f) ) instead. But then this vector will not, in general,
have a vanishing total time derivative, and so the total surge
45
(time total acceleration derivative) vector of the particle will

no longer be independent of g ; instead, we will have f
defined implicitly by equation (5.2) below since then f
appears not only alone on the left here, but also on the right
and its total time derivative appears there also. This then
yields a first order ordinary equation in f as a function of
time, but we are at a loss to determine the corresponding
arbitrary constant in its solution. And so we must employ
“iteration” [23] to obtain a solution here. [We might use
the function f(t) ≡ 1 to begin the iteration process … as this
corresponds to using Newton’s Universal Law as a first
approximation.] Note that if we set h = 1 / f, then the left
side of equation (5.2) is 1 / h, and so if we take the
reciprocal of both sides of our equation, we can then
iterate for h; once h is obtained explicitly, we can then set f
= 1 / h. And it appears to follow from [24] on the upper
part of page 30 that these two iteration methods need not
have the same regions of convergence nor the same rates
of convergence, if convergent. [Here we are abusing
notation in that this h is a function of time, not a constant
deceleration as elsewhere in the chapter.] And we see that
f(t) cannot vanish here, and so by (assumed) continuity, it is
either a positive function of time over a closed interval of
time or else a negative one there, and it must be bounded
in absolute value over a closed interval of time also.
We have noted earlier in this chapter that gravitational

forces are treated herein in a special way as the term
involving the time derivative of the inertial mass in the
generalized equation (2.4) fails to appear for such forces
[see equation (2.5)]; however, the first author feels that this
iteration technique just mentioned will remedy this
asymmetry … which might concern the careful reader …
since it means that the earlier procedure of adding on the
gravitational field vector G to the system acceleration
vector can then be viewed as a (strictly) first order
approximation.
46
We might also mention that in [19] Ph. M. Kanarev

explains the observations of a difference between the free
falling accelerations of a vertically spinning gyro depending
upon spinning clockwise or counterclockwise using
electromagnetic effects, and we, of course, do not consider
such effects here (although they do fit in much as in the
case in Newton’s mechanics as there force is not defined
by its causes, but rather by its effects … as we do also).
At this point, we note that it follows from equations

(4.10), and (4.13) that d(Irj)/dt ≡ 0 for j = 1, 2, and 3. This
means that our r-generalized Euler’s equations of motion
(4.12) reduce in our situation with h zero to the Newtonian
Euler’s equations of motion except that the Ij moments of
inertia become the Irj r-moments of inertia. We can also
show that when - h >> ω, g, and r, we have d(Irj)/dt ≡ 0.
We now assume that the r-products of inertia all vanish

as the products of inertia do (as mentioned by Hay) since
that would seem to extend to our neo-Newtonian theory
as our aim is to change as little as possible of classical
mechanics in this our straight forward generalization of it.
(We note that in Hay’s equation (41.26) the minus sign
should be a plus sign, with thanks to Prof. Chris Provatidis
for this errata.)
Next we drop our assumption that h = 0 and assume

instead that - h >> g (as then the non-positive h is negative).
Now we utilize our remark after equation (3.17) to re-
compute (4.4) as follows (where in this case, a = - h >> ω
& g, and where additionally da/dt = (ρ ω3) with both
equalities holding approximately) :
d(mr) = λ dV f = [m / (π (r 2 - r’ 2) s)] (s ρ dφ dρ f ) =
47
= [m ρ dφ dρ / (π ( r 2 - r’ 2))] [1 – Σ b(m1, m2, m3) (-h)m1
(ρω3)m2 sin(φ)m3], (4.16)
whence
mr = m – {m / [π (r 2 - r’ 2)]}Σ b(m1, m2, m3) [(r t – r’ t) / t]
(ω3m2) (-h)m1{∫ sin(φ)m3 dφ} . (4.17)
using the usual limits of integration, and where t = m2 –

m3 + 2.
Now, since we have φ = ω t: we have
d(d(mr))/dt = {m / [π (r 2 – r’ 2)]} { ρ dφ dρ}Σ b(m1, m2, m3)
(-h)m1 (ρ ω3)m2 [sin(φ)m3-1 (m3 cos(ω t) ω)] (4.18)
Thus, at this point, we can utilize (2.13) to compute Mr, the

linear momentum of the spinning rotor (with steady
angular velocity ω = s) which is also precessing with angular
velocity p using as we have shown dmr/dt ≡ 0 [equation
(4.10)], whence dmr/dt xrc ≡ 0 also. Thus it only remains
to compute
∫𝑜𝑣𝑒𝑟 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑟𝑜𝑡𝑜𝑟[d(d𝑚𝑟 )/dt 𝐱]dV (4.19)
Here we are utilizing the following correspondences [see

equation (2.13)]:
48
dmr <---> mrj (because t is held fixed in the

differential)
x <-----> xj
d(dmr)/dt <---> dmrj/dt
xrc <----> xrc (it corresponds to itself). (4.20)
Now x is given by equation (3.3) and d(dmr)/dt from (4.18),

and we can show that the first two coordinates of x result
in no contribution to the volume integral in question by
using the fact that – h >>  >> g, and we use our previous
assumption that the r-products of inertia are assumed to all
vanish here (these integrals vanishing imply the vanishing
of the first two coordinates of the volume integral). But
there may be a contribution from the third coordinate of
x, namely, (l i3), but a computation shows that this
coordinate of the volume integral vanishes too since
𝜋
∫𝜑=−𝜋 𝑐𝑜𝑠 𝑞 𝜑𝑠𝑖𝑛𝑝 𝜑𝑑𝜑 > 0 (4.21)
if p and q are both even, but otherwise it vanishes. [See

equation (4.8) and its discussion.] Thus
∫𝑉[𝐱 𝑑(𝑑𝑚𝑟 )⁄𝑑𝑡] 𝑑𝑉 = 0 (4.22)
again in view of our earlier assumption that the r-products

of inertia in Hay’s gyro situation vanish (but we may be able
to prove this directly as f is an even function of α of period
2) and since this integral involves the term [cos(φ)m3-1
sin(φ)] in φ under the integration sign; and so by (4.21), this
follows.
49
It should be remarked that in our theory, the inertial mass

of a rigid body (that is, resistance to force) depends in
general on the vector force applied to it. And so we cannot
think that even if the r-inertial mass mr = m at first order if
h = 0, that this means that the same holds also if - h >> 
>> g.
Now, it follows from just above this paragraph that:
Mr = mr vrc (4.23)
but with assuming slow precession. Eric Laithwaite in his 1974

Christmas lecture [5] claims to show that a steadily
precessing gyro has reduced impact force … which seems
to correspond to the term by term integrals of the trig
functions of (4.16) being negative so that impact mass is
lost in general. Note that this implies that we must only
consider real forces in our theory, not what French [10]
names fictitious forces such as centrifugal and Coriolis
forces which only occur in rotating systems … unless these
forces can be derived in terms of real forces as (for
example) is done by French in the purely Newtonian case!
[See equations (2.40) above.]
Now, the rate of change of the angular momentum hrc

of the rotor about its r-center of mass is given by
𝑑𝒉𝑟𝑐 𝑑𝑚𝑟 𝑑𝑚𝑟𝑗

𝑑𝑡
= 𝐆 − 𝒗𝑟𝑐 × 𝐌𝑟 = 𝐆 − 𝒗𝑟𝑐 × (
𝑑𝑡
𝒙𝑟𝑐 − ∑𝑁
𝑗=1 𝑑𝑡
𝒙𝑗 ) = 𝐆 (4.24)
in view of equations (2.39) and (4.22), where G is the total

moment of the external forces about the r-center of mass
of the rotor (the same as the center of mass of the rotor by
symmetry) and equation (2.13), since vrc is in the same
direction as – i2 as p is non-negative and equal to p0 (1 – t /
e), where t is time and where e is the time that it takes to
50
decelerate the precession to zero, it having begun at the

constant value p0 at t = 0. Thus, the term on the right of
(4.21) is the usual Newtonian term, and the second is
anomalous but vanishes identically here. It can be shown
using the first of equations (4.12) [the generalized Euler
equations] that we have the equation
Ir1 sin(θ ) dp/dt = - F2 l + G1' (4.25)
where G1' is the total moment of forces [see equation

(2.38)] about i1 except for the (new) such moment due to
this force F2 which is a force acting on each rotor particle
and parallel to i2 to insure that the corresponding torque
about the origin O of the coordinate system is along i1 … in
order to mandate this constant decelerating of the
precessional angular velocity. [Here, for pedagogical
reasons, in the discussion below, we neglect G1', the
moment of the forces about i1 of any rotor particle
acceleration due to gravity and also due to the rotor particle
forces corresponding to either spinning or precessing; and
similarly for i3. (See the preface on to this work.)] Also,
we must introduce a force F1 acting on the rotor center of
mass and parallel to i1 to exactly prevent any nutation of
the gyro; and this situation can be physically realized, of
course. Further, we have here that the rotor center of mass
is also the rotor center of r-mass by symmetry
considerations.] Since F2 is the force parallel to i2 needed
to insure that the precession p decelerates constantly with
deceleration h from t = 0 to t = e, we note that during this
time interval, dp/dt = - p0 /e = h and that d(Ir1)/dt ≡ 0 in
view of (4.10) and (4.15). But notice that it follows from
the third of the generalized Euler equations (4.12) that (not
neglecting p)
Ir3 [dω/dt + dp/dt sin(θ)] = 0, (4.26)
51
and so ω varies slightly during deceleration of p {see

equation (41.26) of [1]} . But in view of the fact that we
have all along been assuming that it is nearly constant, we
must then add a torque on the gyro rotor (i.e. along i3) to
exactly cancel the torque on the rotor generated by this
deceleration of ω (similar to the addition of the force F1
along i1 giving rise to a torque along i2 that just exactly
prevents any nutation during this deceleration resulting in
a moment of force about the origin O, but only along i2,
that must be added as a vector to the other moments of
force about O [16].) And, since this torque (to prevent spin
up) is only along i3, this will not affect the other two
equations obtained from these generalized Euler equations!
Further, we assert that m, h, g, l, θ and ω are constant in this

time interval because it follows from the third of equations
(4.12) [after introducing the just mentioned torque to
prevent rotor spin up] that the spinning angular velocity 
(also denoted as s) is constant as its time derivative
vanishes, and also h is assumed constant [16].
Therefore, we have from (4.25) that
F2 = (Ir1 / l ) sin(θ) dp/dt = - Ir1 p0 sin(θ) / (e l ) (4.27)
Thus the magnitude of the total counter angular

momentum needed to force this constant precession
deceleration is
𝑒
∫𝑡=0 𝐹2 𝑙 dt = - Ir1 p0 sin(θ) (4.28)
Hence, Fiala’s notion of ‘quality’ [7] in the stop dead impact

case is seemingly given by (1 – Qa) [but Fiala – if this
author understands him correctly -- objects to this as he
seems to feel that quality shouldn’t depend upon spin
angular velocity, but rather follows Laithwaite here] where
52
Qa = {- Ir1 p0 sin(θ ) / (- I1 p0)} = Ir1 sin(θ) / I1

(4.29)
But it is not true that Ir1 = I1 (mr / m) since the series in
(4.16) of sine functions when integrated term by term will
differ from the term by term integrals of the series
multiplied by either cos2(φ) or sin2(φ) [and so the r-
moments Ir1 and Ir2 may differ, unlike I1 and I2, as cos2(φ) =
1 – sin2(φ), but we follow Hay in not requiring their
equality, although Synge and Griffith do, as we only use
principal axes of inertia to get Euler’s equations here] for
some constants C and K … as we are now going to a three
termed approximation of function f to fifth order in cos(α):
Ir1 = m [(r2 + r’ 2) / 2 + l 2] – [(m / (π (r2 - r’ 2))] [ ∫ ∫ {((ρ
cos(φ))2 + l 2)(b(1,1,2) (-h) (ρ ω3) cos(α)2 + b(2,2,4) (-h)2 (ρ
ω3)2 [cos(α)]4)}ρ dρ dφ] =
= m {(r2 + r’ 2) / 2 + l 2} + (m / (π (r2 – r’ 2)) (-h){[C (π /
4) ((r5 – r’ 5) / 5) (ω3) + K (-h) ((π / 8) ((r6 – r’ 6) / 6) ω6] +
l 2 [C (π) ((r3 - r’ 3) / 3) ω3 + K (-h) (3 π / 4) ((r4 – r’ 4) / 4)
ω6 ]} (4.30)
where φ is integrated from –π to π, where ρ is integrated

from r’ to r, where cos(α) = sin(φ) here in view of (3.11),
and where b(1,1,2) = C and b(2,2,4) = K by way of
definition. {The conditions (3.18), even though obtained
from [11], apply as the b(i,j,k) are universal constants of
physics (if our theory is correct)}. And we note the facts
that I1 = m [(r2 + r’ 2) / 2 + l 2] which follows from the
53
parallel axis theorem [but this theorem may not apply to r-

moments] and the formula for the moment of inertia of a
disc having outside radius r and inside radius r’ and that I1
/ m = (r2 + r’ 2) / 2, and so (4.30) seem to be correct. We
are assuming that the gyro is precessing steadily until
deceleration begins at t = 0. Thus our ‘quality’ is given by
(1 – Qa) where Qa is from (4.29) after substitution from
(4.30) since vrc = vc by symmetry.
Note that then the expression on the right of the

consequential formula for our quality (1 – Qa) is not a
function of m or g, but it is a function of h as the right side
of equation (4.30) is a quadratic function of h so that when
-h  ∞, we have Ir1  ∞ also. This seems unphysical …
and we used to suspect that just as there is trouble in
Newton’s mechanics when very large velocities enter the
picture (resulting in the introduction of a γ-factor of mass
change), there is similarly trouble in our theory when
excessively large accelerations come up, but now think that
this may be incorrect since, for example, in the case of a
He4 atom, in directly impacting a plane barrier, h will never
really be minus infinity since the billiard ball model fails to
apply! It is the h being minus infinity in this situation that
is the true unphysical ideal, not our theory! Also, in this
situation considering billiard ball collision, for example, we
see that if we used a constant deceleration to model their
impact the surge (i.e. da/dt) would almost vanish, and so
consequently their r-masses would nearly equal their rest
masses and the situation would then be essentially
Newtonian! And, additionally, it may then appear that there
is a problem here as these two parameters do appear in (4.4)
or (1.1), but recall that we are assuming – h >> ω >> g (for
impact analysis) and the term on the right becomes infinite
as l  0 so we must have l > 0 here. Finally, m > 0 (as m
= 0 corresponds to nothing) should not enter into the
quality as it is the fractional rest mass change and, as such,
should be independent of rotor inertial rest mass.
54
Now note that C and K are constants not depending

upon any of our parameters [including the tilt angle θ and
also g ], and so it can be determined and computed by
experiments involving a Hay type gyro (with thin washer
shaped rotor) in steady precession when stopped dead
from precessing and measuring the resulting impact. It will
be interesting to see if this result is reproduced by these
type of experiments using different values for the
parameters of equation (4.30) ... as there is only one pair of
values (C, K) that must hold in each and every such
experiment; further, these values cannot depend on the
particulars of any experiment of this type as both C and K
are (if we are correct here) universal constants of physics
(not depending upon even g)!
However, in view of the fact that we only have Prof. A. L.

Dmitriev’s falling rotor experimental data [11] to use in
order to calculate our two constants C and K, and since he
only used two identical GMS-1 aviation rotors (may be
Googled) in this work; it still could be that C and K depend
also upon the exact rotor material used in manufacturing
analogous to the dielectric constant varying in EM theory.
We can now define the ‘effective r-mass’ mer the rotor

(much as in [16]) by setting it equal to (m Qa) for a Hay’s
type gyro having rest mass m and Qa as in (4.29) & (4.30)
above, and this is just the mass with which the gyro
discussed above impacts the stop dead mechanism when
considered as a point particle (spinning about itself or not
… as a point particle has no extension with which to
exhibit rotation about its own point location) having
velocity at impact of (l p0), the precession linear velocity at
impact. We note [17] that Newton built his theory of
continuous forces upon the theory of impact forces,
55
whence the notion of effective r-mass mer is well within the

spirit of Newton’s Principia.
We might have expected that the two gyroscopic cases

discussed just before equation (5.1) would have turned out
to be the same as far as the concept of our quality is
concerned. However, the example in [16] has impact force
parallel to the rotor center of mass linear velocity vector;
but in the case of centrifugal force, the centrifugal force
vector is along the gyro shaft and so is perpendicular to the
impact force vector just mentioned. And, of course, this
mer depends upon vector acceleration direction as well as
upon vector magnitude! Clearly further work has to be
done here to get to the bottom of this matter, and we
certainly need to remove our assumption that the
precession angular velocity is small!
Finally, if the effective rotor mass in the stop dead case

were negative, then the gyro impact would have to be zero;
but the more negative it were, the sharper the recoil
tendency would be after the (zero) impact that would be
expected! However, this recoil tendency would be in the
opposite direction than that of the precession tendency
thus making the situation murky. And if the gyro should
begin to nutate during recoil, then energy physics would
enter in … since then the rotor falls or rises thereby
changing its gravitational potential energy (and see the
energy appendix of this chapter).
5. Conclusion
Now, we have approximated our function f to fifth
order in cos(α), and we have obtained what appear to be
good results. So we suspect that this approximation is quite
correct, but experimental verification is certainly required
here! And, in this regard, the author would urge
experimentalists that are into gyroscopic phenomena to
view the 1974 Royal Society's Christmas lecture given by
56
Eric Laithwaite concerning his anomalous gyroscope

experiments, and, in particular, the last (#17) segment (the
lecture video being subdivided into short segments) which
may be viewed individually at [5]. It’s a lecture to children,
of course, but very enlightening!
This segment #17 is a demonstration of the puzzling

phenomena of precessing gyroscopes with a fixed pivot
which are steadily precessing without nutating (one
situation of our Part Two) apparently having reduced
angular momentum. However, the author has calculated
the two (universal) constants C and K mention above using
Prof. Dmitriev’s data [11] (Appendix 1 at its end), and they
turn out to be about zero and about (-1.786189927 10-13),
respectively. {However, it should be mentioned that it
might be that K also depends upon the actual rotor material
composition (assumed homogeneous here) … although
Prof. Dmitriev appears to think otherwise [25, 27].} Thus,
if his data is good, we see that for a gyroscope of the type
we are discussing, there is a possible increase of inertial
mass, not always a decrease as claimed by Laithwaite; and,
moreover, our theory then might possibly predict the
observed direction of the net motion of Fiala’s working
model of his inertial propulsion device featured in his 2012
TeslaTech talk [21]. (Fiala also has an interesting paper
concerning inertial propulsion and star travel [3].)
That case would certainly appear not to be the same as

the stopping dead impact one (treated above). If these cases
could be successfully addressed completely using the
above, then one would appear to be off to a good start in
developing a consistent and viable generalization of
Newtonian theory so as to include all the old applications
and also the gyro applications that don't quite fit. But, of
course, what one is really shooting for is to find the exact
formula for
57
f = f(a, da/dt, α) = 1 - {∑ b(m1, m2, m3) am1 (da/dt)m2
cos(α)m3} (5.1)
which is our Eq. (3.13) where is an even function of α of

period 2.
Further, we see that the coefficients b(m1, m2, m3)

vanish for m3 odd since the functions sinj(θ) are linearly
independent over our region from – π to π, and so if we
divide this region into a (p - 1) number of subintervals of
equal length by θ1, θ2, …, θp, where θ1 = -π & θp = π, then
looking at the condition that the r-products of inertia all
vanish for each such θi , it is readily seen that for p >> n0,
there are too many equations in too few unknowns if b(m1,
m2, m3) ≠ 0 for m3 odd , and so the conclusion follows.
Thus, in view of (3.18), we have as a first approximation
that
f = f(a, da/dt, α) = 1 – {C a (da/dt) cos2(α) + K a2
(da/dt)2 cos4(α)} (5.2)
(again, with da/dt standing for |da/dt|) to fifth order in

cos(α) for some non-zero constants C and K; and this may
well actually be exact. Also, the term containing K with K
omitted is just the square of the term containing C with C
omitted. Thus, to sum all this up, it seems reasonable to
assume that, even if (5.2) is not exact, it is close enough for
virtually all ordinary (low-tech) purposes!
The above discussion in this chapter is concerning

“momentum mechanics” only; but in engineering texts
58
such as Greenwood’s [16], almost the whole modern

approach to dynamics is rather an energy one involving
Lagrange’s equations of motion … heavily. So, then, the
question arises, can all this be extended to a similar
treatment involving r-mass instead of ordinary inertial
(rest) mass? And the second author (JD-D) in his (below)
Chapter 6 -- and under the heading “Modified Lagrange
Equations” -- shows clearly that the usual Lagrange
equations go over completely to the variable (r-mass) case
… subject only to the (obvious) condition that the
components of force therein then change from being the
usual inertial mass components to the corresponding r-
mass force components instead!
This is, of course, very fortuitous … because Lagrange’s

equations are central to obtaining solutions to mechanical
engineering problems today … as Greenwood points out
in his classical engineering dynamics text.
We refer the reader interested in the first author’s

world view to [14], a memoir.
References
[ 1 ] G. E. Hay, Vector and Tensor Analysis, pp. 62-101
(Dover, 1953). [Just recently back in print at Dover
Books.]
[ 2 ] R. Wayte, ”The Phenomenon of Weight-Reduction

of a Spinning Wheel”, Meccanica 42: 359–364 (2007).
[ 3 ] Harvey Fiala, ”Inertial Propulsion: The Holy Grail of

Travel to the Stars”, ExtraOrdinary Technology, 4 (3):
29-44 (2006).
[ 4 ] J. D. Jackson, Classical Electrodynamics, 2nd Ed., (John
59
Wiley & Sons, 1975).
[ 5 ] Eric Laithewaite, ”The Jabberwock: The Royal

Institution’s 1974-75 Christmas Lecture” (1974),
http://gyroscopes.org/1974lecture.asp.
[ 6 ] Herbert A. Simon, The Axiomatization of Classical

Mechanics, Phil. Of Sci, 340-343 (1954).
[ 7 ] Fiala et al (2011) U.S. Patent No. 7,900,874 B2

Washington DC: U.S. Patent and Trademark Office.
[ 8 ] J. L. Synge and B. A. Griffith, Principles of Mechanics,

(McGraw—Hill Book Co., 1942).
[ 9 ] J.P.Wesley, Scientific Physics, (Benjamin Wesley, 2002)
[10] A.P. French, Newtonian Mechanics, (W.W. Norton &

Company, 1971).
[11] Alexander L. Dmitriev, “Frequency Dependence of

Rotor’s Free Falling Acceleration and Inequality of Inertial
and Gravity Masses”, http://arXiv.org:1101.4678v1
[physics. gen-ph], pdf(2011).
[12 ] Dennis P. Allen, Jr., Foundations of Neo-Spearsian

Gravitational Theory with Application to Earthquake Early
Warning Systems, 2nd Ed., (CreateSpace. 2017)
{Available from Amazon.com}.
[13] David L. Bergman & Dennis P. Allen, Jr., “Electron

in the Ground State – Part 1”, Foundations of Science 16 (1)
(February, 2012).
[14] Dennis P. Allen, Jr., The Reality Oriented

Mathematician, 3rd Ed., (CreateSpace 2017). {Available
from Amazon.com}
60
[15] M. A. Burnstein and W. A. Freidman, Thinking About

Equations (John Wiley & Sons, 2009).
[16] Donald T. Greenwood, Classical Dynamics, (Dover,

1997).
[17] I. Bernard Cohen & George E. Smith (Editors), The

Cambridge Companion to Newton (Cambridge University
Press, 2002).
[18] T.E. Phipps, Old Physics for New, (C. Roy Keys Inc.,
2012).
[19] Ph. M. Kanarev, “The Law of Conservation of

Momentum”, Journal of Theoritics, 4-4, (2002).
[20] Dennis P. Allen Jr., “An Elementary Method for the

Elimination of the First Principal Difficulty in the
Axiomatization of Classical Mechanics: Circular
Reasoning”, Unpublished (2013).
[21] Harvey E. Fiala, “An Inertial Propulsion Patient &

Working Model”, Presentation, July 29, 2012, Tesla Tech,
Inc., Marriot Pyramid North, Albuquerque, New Mexico.
[22] Dennis P. Allen Jr. and Jeremy Dunning-Davies,

“Neo-Newtonian Mechanics With Extension To Relativistic
Velocities; Part 2: Radiative and Inductive Effects”, 7th Ed.,
(CreateSpace.com, 2017). {Available from Amazon.com}
[23] Victor Bryant, Metris Spaces: Iteration and Application,

(Cambridge Univ. Press, 1985).
[24] Erwin Kreyszig, Advanced Engineering Mathematics, 2nd
61
Ed., (John Wiley & Sons, 1962).
[25] Alexander L. Dmitriev, “Measurements of the

Influence of Acceleration and Temperature of Bodies on
their Weight”, CP969, Space Technology and
Applications International Forum – STAIF 2008.
[26] Max Jammer, Concepts of Mass in Classical and Modern

Physics, (Dover, 1997).
[27] Alexander L. Dmitriev, “Weighing of a Mechanical

Gyroscope with Horizontal and Vertical Orientations of
the Spin Axis,” Measurement Techniques, 44(8), 831-833
(2001).
Appendix: Neo-Newtonian Energy Methods
In view of equation (2.1), we have:

2
∫1 𝐅 𝐝𝐨𝐭 d𝐫 =
2 d(𝑚𝑟 ) 1 2 d(𝐯 𝐝𝐨𝐭 𝐯)

= ∫1 ( 𝐯 𝐝𝐨𝐭 d𝐫) + ∫1 𝑚𝑟 dt =
dt 2 dt
𝟐 d𝑚𝑟 d𝐫 1 2
∫𝟏 dt
(
dt
dot d𝐫
dt
) dt + ∫1 𝑚𝑟
2
d(𝐯 𝐝𝐨𝐭 𝐯)
dt
dt =
2
∫1 (𝐯 𝐝𝐨𝐭 𝐯)
d𝑚r
dt
dt + 12 ∫12 𝑚𝑟 d(𝐯 𝐝𝐨𝐭 𝐯)
dt
dt =
mr2 v22 - mr1 v12 − ∫𝟏𝟐 𝟐 𝑚𝒓 (𝐯 𝐝𝐨𝐭 d𝐯

dt
) dt + 12 ∫12 𝑚𝑟 d(𝐯 dt
𝐝𝐨𝐭 𝐯)
dt =
1 2 d(𝐯 𝐝𝐨𝐭 𝐯)
= mr2 v22 - mr1 v12 - ∫ 𝑚𝑟
2 1 dt
dt =
1 2 d(v2 )
= mr2 v22 - mr1 v12 - ∫ 𝑚𝑟 dt
2 1
dt
62
(integrating by parts). Thus we have finally (since we think

we want the work-energy theorem to continue to hold in
our neo-Newtonian theory):
Work12 = ∫12 𝐅 𝐝𝐨𝐭 d𝐫 =

1 2 d(v2 )
= mr2 v22 - mr1 v12 - ∫ 𝑚𝑟 dt
2 1
dt
Note that should dmr/dt vanish identically, then this

generalized work-energy formula reduces to the usual
K12 = ½ mr2 v22 - ½ mr1 v12.
as expected … since our neo-Newtonian theory reduces

to Newtonian mechanics when d(mr)/dt vanishes
identically! (Note that mr = m in Newtonian mechanics …
by definition.)
It may be worth mentioning here that our above revision

of the formula for the kinetic energy would seem to call
into question all the various proofs that so called
“perpetual motion devices” of any kind violate
“conservation of energy,” and thus must be hoaxes. But,
of course, a wrong proof need not necessarily mean a
wrong proposition; that is just not true.
And we conclude the chapter by noting that the

electromagnetic pendulum anomalous non-conservation
of energy results of (the late) Prof. Peter Graneau and his
son Neal (see their excellent book: Newtonian
Electrodynamics) may now finally be explainable using our
above generalization of Newton’s mechanics … and then
also of Leibnitz’s kinetic energy mechanical ideas (see the
first author’s web page on the research web site,
ResearchGate.com).
63
64
5 AN ELEMENTARY METHOD FOR THE
ELIMINATION OF THE FIRST PRINCIPAL
DIFFICULTY IN THE
AXIOMATIZATION OF CLASSICAL
MECHANICS: CIRCULAR REASONING
Dennis P. Allen, Jr.
17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI 49456-9273
The principal difficulty in the beginning of the axiomatization of

Newtonian mechanics is that there appears to be circular reasoning in
the development of his important fundamental notion of an inertial
system and his key notion of force, with force being the time derivative
of the (vector) momentum. Of course, he used neither the notion of a
vector nor that of an inertial system, but both were certainly implicit in
his “Principia”. The problem here may be expressed as follows: on the
one hand, in view of his first law of motion, an inertial system is one in
which point particles move in unaccelerated motion unless acted on by
a force of some kind but, on the other hand, the force on a point particle
is the time derivative of its momentum in an inertial system. Hence,
circular reasoning is involved in that an inertial system presupposes the
notion of a force while the notion of a force presupposes the notion of
an inertial system. The complex, usually accepted resolution of this
difficulty is expounded in an excellent article [1] by Eisenbud and it is
this article which is relied on quite heavily in, for example, the well-
known text by Resnick and Halliday [2].
In order to motivate our alternative resolution of the problem,

reference is made first to pages 164-6 of Poincaré’s book Science and
Method [3], where he criticizes Peano’s five axiom system for the natural
numbers, by which is meant the non-negative integers. His argument
reduces to saying that Peano’s main axiom (his number 5) – the principle
of complete induction – together with his other four axioms form a
system that may only be shown to be consistent by using induction.
Hence, circular reasoning is employed, The incompleteness theorem
due to Kurt Gödel came later and so does not impact the analysis here.
However, the criticism may be removed quite easily as follows:
65
VELOCITIES.
Let us view this argument in three dimensions rather than two. Suppose
this circular reasoning is viewed as the unit circle in the x-y plane, using
Cartesian coordinates. Let the third dimension be intuitive thinking on
the positive z-axis and formal thinking on the negative z-axis. Then
Peano’s axiomatization may be viewed as a spiral going from the
intuitive notion of induction on the positive z-axis down to the formal
and precise notion expressed in his fifth axiom on the negative z-axis
rather than as a circle, but with the spiral projected onto the x-y plane
as a circle, just as Poincaré was apparently visualizing the reasoning of
Peano.
With this in mind, consider the circular reasoning mentioned in

connection with Newtonian mechanics. Note that, although the notion
of an inertial system does rely on the notion of force, it does so only in
that the intuitive and non-quantitative notion of force is required …
such as a “nudge”. And having obtained a formal, rigorous notion of an
inertial system in this way, the notion of force may then be formalized
and, hence, quantified, using this newly obtained formal, rigorous
notion of an inertial system.
Hence, it seems that, while the idea of a Euclidean style axiomatization

of classical mechanics, such as outlined by Eisenbud [1], involves
explaining the notion of an inertial system in psychological terms, if the
notion of an axiomatization is itself revised and generalized as indicated
above to include the extra dimension of intuition-formal, then the
situation concerning an inertial system becomes considerably clearer
and elementary as well as more straightforward. The notion of
Euclidean-style axiomatization for a physical theory is simply too rigid
and narrow and needs to be revised to include the process of going from
intuitive thinking to formalization and, hence, quantification!
As a final remark, it might be noted that Simon [4,5 and note his
discussion of 6 in the latter] has published articles on the axiomatization
of classical mechanics and he refers to, on the one hand, an isolated
system and, on the other hand, finds he has to refer to a coordinate
system that is un-accelerated with respect to the fixed stars. Needless to
say, this involves something of a contradiction since these stars are
evidently outside his system. He has to do this since, no doubt, he has
noticed the well-known fact that Newton’s own argument involves the
above mentioned ‘strictly speaking’ circular reasoning involving inertial
systems and force. The interested reader might usefully refer to Chapter
1 of Prof. Robert DiSalle’s in The Cambridge Companion to Newton for an
66
excellent treatment of Newton’s thinking with regard to the foundations

of his mechanics [7].
References.
[1] L.Eisenbud; American Journal of Physics, March 1958.
[2] R. Resnick & D. Halliday; Physics for Students of Science and Engineering,
Part1, (John Wiley & Sons, New York, 1960).
[3] H. Poincaré; Science and Method, ( Dover, New York, 1914).
[4] H Simon; Phil. of Sc., December 1947.
[5] H. Simon; Phil. of Sc.,October 1954.
[6] J. C. McKinsey, A. C. Sugar and P. Suppes; J.Rat.Mechs,& Anal.,

April 1953.
[7] B. Cohen and G. E. Smith (eds.), The Cambridge Companion to

Newton, (C.U.P., Cambridge, 2002).
67
68
6 SOME RESULTS IN CLASSICAL
MECHANICS FOR THE CASE OF A
VARIABLE MASS
Jeremy Dunning-Davies,
Institute of Basic Research, Palm Harbor, Florida, U.S.A.
and
Via Santa Gonda, 14 - 59100 Prato, Italy.
Introduction.
It was at a PIRT meeting in London in the 1990’s I first

met Ruggero Santilli and discussed the well-known
Lagrange equations of motion and the derivation normally
presented to undergraduates. Since I had always presented
the general form in undergraduate lectures before
proceeding to discuss the special case concerning
conservative fields, I was a little surprised to find that he
found a great many were unaware of the fact that the
special case is just that and regarded Lagrange’s equations
of motion as being the form depending on the so-called
Lagrangian, which is the difference between the kinetic and
potential energies of a system and, hence, purely relevant
to the case of conservative fields of force. However, that
common particularization is not the only one being, at least
tacitly, fed to students in classical mechanics. In most, if
not all, introductions to classical mechanics, the mass is
assumed to be constant. Frequently, when Newton’s
famous Second Law is introduced, attention is drawn to
such systems as rocket motion to indicate that, in practice,
the mass is not always a constant but often this is as far as
69
discussion of variable mass systems goes. In truth, many

students actually meet a varying mass for the first time
when introduced to the Special Theory of Relativity.
However, varying masses do occur in nature when
relativistic effects are not important. Here an attempt is
made to draw together some common results of classical
mechanics with a variable mass taken into account.
Particular attention will be drawn to a perceived change in
the expression for the kinetic energy and to crucial changes
in the basic form of Lagrange’s equations of motion but
first the situation pertaining to rotating frames of reference
will be considered before proceeding to the main two
topics to be considered here.
Rotating frames of reference.
Consider the well-known situation concerning rotating

frames of reference. Suppose a vector A has components
a, b, c with respect to a set of axes along which the unit
vectors are i, j, k. Then
A = ai + bj + ck
and
𝑑𝑨 𝑑𝑎 𝑑𝒊 𝑑𝑏 𝑑𝒋 𝑑𝑐 𝑑𝒌
= 𝒊+𝑎 + 𝒋+𝑏 + 𝒌+𝑐 .
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
But, as is well-known
𝑑𝒊
𝑑𝑡
= 𝝎x𝒊, etc.
Hence, finally,
𝑑𝑨 𝑑𝑎 𝑑𝑏 𝑑𝑐
= 𝒊+ 𝒋 + 𝒌 + 𝑎𝝎x𝒊 + 𝑏𝝎x𝒋 + 𝑐𝝎x𝒌
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑑𝑎 𝑑𝑏 𝑑𝑐
= ( , , ) + 𝝎x(𝑎𝒊 + 𝑏𝒋 + 𝑐𝒌)
𝑑𝑡 𝑑𝑡 𝑑𝑡
𝜕𝑨
= + 𝝎x𝑨 = 𝑨̇ + 𝝎x𝑨.
𝜕𝑡
The notation in the final equations might be termed a
convention since a partial derivative in the usual sense is
70
not being discussed. Also, the dot refers to differentiation

with respect to time of the components of the vectors A.
In this case of variable mass, the equation symbolizing
Newton’s second law becomes
𝑑 𝑑𝑚
𝑭 = 𝑑𝑡 (𝑚𝒗) = 𝑑𝑡 𝒗 + 𝑚𝒂,
where F represents the force acting, v the velocity and a
the acceleration.
But, if a rotating frame is involved, this becomes
𝜕
𝑭 = (𝑚𝒗) + 𝝎x(𝑚𝒗)
𝜕𝑡
that is
𝑭 = 𝑚̇𝒗 + 𝑚𝒗̇ + 𝝎x[𝑚(𝒓̇ + 𝝎x𝒓)]
= 𝑚̇(𝒓̇ + 𝝎x𝒓) + 𝑚(𝒓̈ + 𝝎̇x𝒓 + 𝝎x𝒓̇ )
+ 𝝎x[𝑚(𝒓̇ + 𝝎x𝒓)]
(𝒓̇
= 𝑚̇ + 𝝎x𝒓) + 𝑚(𝒓̈ + 𝝎̇x𝒓 + 𝟐𝝎x𝒓̇ + 𝝎x(𝝎x𝒓))
or
𝑚𝒓̈ = 𝑭 − 𝑚̇𝒓̇ − 𝑚𝝎x(𝝎x𝒓) − 𝑚̇(𝒘x𝒓) − 𝑚(𝝎̇x𝒓)
− 2𝑚(𝝎x𝒓̇ )
Here, once again, the dot refers to differentiation with
respect to time of the components of the various vectors.
Kinetic energy.
In many approaches to classical mechanics, the starting

point is to assume a constant acceleration so that
𝑑2 𝑠
= 𝑎,
𝑑𝑡 2
where s is distance, t time, and a the constant acceleration.
Integrating once and assuming the velocity, v, initially has
the value u, we get
𝑑𝑠
𝑣= = 𝑢 + 𝑎𝑡.
𝑑𝑡
Integrating a second time and assuming s is zero initially,
we get
71
1
𝑠 = 𝑢𝑡 + 2 𝑓𝑡 2 .
These are very well-known straightforward equations but
are included here, together with their derivation, to
highlight the fact that they do not depend in any way on
the mass of an object. In fact, all assumptions made are
clearly stated.
Finally, if s is eliminated between the latter two equations
𝑣 2 = 𝑢2 + 2𝑎𝑠
results and it follows from this equation that, if an object is
brought to rest (v = 0) by a retardation a then this equation
gives
1
0 = 𝑢2 − 2𝑎𝑠  2 𝑢2 = 𝑎𝑠
If this equation is multiplied throughout by m, one obtains
1
2
𝑚𝑢2 = mas.
With a taken to be g, the acceleration due to gravity, this
equation takes on the popular interpretation of kinetic
energy being equivalent to potential energy. Here the
interpretation of the right-hand side follows from
considering Newton’s Second Law with the mass assumed
constant; that is
𝑑
Force = P = 𝑑𝑡 (𝑚𝑢) = 𝑚𝑎.
Then, since work done is force and x distance, mas is a work
term and is regarded as being equal here to the kinetic
energy. This is how the notion is introduced in textbooks
but, since the acceleration is taken to be constant, it follows
that the force, P, must be constant also.
If the mass is assumed variable, then Newton’s Second

Law takes the form
𝑑 𝑑𝑚
𝑃 = (𝑚𝑢) = 𝑚𝑎 + 𝑣 .
𝑑𝑡 𝑑𝑡
The definition of the kinetic energy of a body is usually
taken to be the energy the body possesses by virtue of its
72
motion and is measured by the amount of work which it

does in coming to rest. The introduction of the second
term on the right-hand side of the above equation to take
account of the fact that the mass is assumed to vary with
time obviously can have no bearing on the actual definition
but, if you consider the work done in a displacement, s, in
this case, the work done might, at first sight, seem to be
given by
𝑑𝑚 1 𝑑𝑚
work done = 𝑃𝑠 = 𝑚𝑎𝑠 + 𝑣𝑠 = 𝑚𝑢2 + 𝑢𝑠 =
𝑑𝑡 2 𝑑𝑡
1
2
𝑚𝑢2
+ 𝑢𝑠,
and so, in this case, there would appear to be an extra term
appearing in the expression for the kinetic energy.
However, this last step is incorrect because, in this case of
varying mass, the force is not constant. In this case, the
final step must be replaced by
2 2 𝑑𝑣 𝑑𝑚
work done = ∫1 𝑃𝑑𝑠 = ∫1 (𝑚 𝑑𝑡 + 𝑑𝑡
𝑣) 𝑑𝑠 =
2 𝑑𝑣 𝑑𝑠 𝑑𝑚 𝑑𝑠 𝑑𝑠
∫1 (𝑚 𝑑𝑡 𝑑𝑡 + 𝑑𝑡 𝑑𝑡 𝑑𝑡
) 𝑑𝑡
2 𝑑𝑣 2
=∫1 (𝑚𝑣 𝑑𝑡 𝑑𝑡 + 𝑣 𝑑𝑚)
2 𝑑𝑣 2 𝑑𝑣
=∫1 𝑚𝑣 𝑑𝑡 𝑑𝑡 + [𝑚𝑣 2 ]12 − ∫1 𝑚2𝑣 𝑑𝑡 𝑑𝑡
2 𝑑𝑣
=[𝑚𝑣 2 ]12 − ∫1 𝑚𝑣 𝑑𝑡 𝑑𝑡,
where it is assumed that the work being done is between
states 1 and 2. Note also that this expression reduces to the
familiar one for kinetic energy when the mass, m, is
supposed constant. In accordance with what has preceded
it in this section, this final result is derived using scalar
quantities at all points; the slight generalization using vector
quantities is a trivial extension. It should be noted that
attention has already been drawn to this very point by one
of us (D.A.) in an article entitled Neo-Newtonian Theory [1].
Again, it should be remembered that, here, all that has
been assumed about the variability of the mass is that it
varies with time. No mention is made of how! The said
73
variation could be through being a direct function of time;

through being a function of varying position coordinates
or varying velocity components or varying acceleration
components, etc. The exact form of this dependence
doesn’t matter at this juncture. However, when it comes to
deriving Lagrange’s equations of motion, the exact form of
the dependence does assume genuine significance.
Modified Lagrange Equations.
For the sake of this discussion it will be assumed that the

mass does not depend on either position or velocity; that
is, m is independent of both 𝑞 and 𝑞̇ in all that follows.
To follow, but generalize, the basic outline in Synge and
Griffith [2], suppose (x, y, z) are the Cartesian coordinates
of a typical particle of a system and suppose we have a
holonomic system of n degrees of freedom described by
generalized coordinates qi, i = 1, 2,…., n. Then,
𝜕𝑥 𝜕𝑥
𝑑𝑥 = ∑𝑛𝑖=1 𝑑𝑞𝑖 , 𝑥̇ = ∑𝑛𝑖=1
𝜕𝑞𝑖
𝑞̇𝜕𝑞1
with similar equations for both 𝑦̇ and 𝑧̇ .
From the second equation, it is seen immediately that
𝜕𝑥̇ 𝜕𝑥
𝜕𝑞̇ 𝑖
= 𝜕𝑞 .
𝑖
𝑑 𝜕
It is straightforward to show that the operators 𝑑𝑡 and 𝜕𝑞
𝑖
commute. Then
𝑑 𝜕 1 2 𝑑 𝜕𝑥̇ 𝜕𝑥̇ 𝑑 𝜕𝑥̇
( 𝑥̇ ) = (𝑥̇ ) = 𝑥̈ + 𝑥̇ ( )
𝑑𝑡 𝜕𝑞̇ 𝑖 2 𝑑𝑡 𝜕𝑞̇ 𝑖 𝜕𝑞𝑖̇ 𝑑𝑡 𝜕𝑞̇ 𝑖
𝜕𝑥 𝑑 𝜕𝑥 𝜕𝑥 𝜕
= 𝑥̈ + 𝑥̇ ( ) = 𝑥̈ + 𝑥̇ (𝑥̇ )
𝜕𝑞𝑖 𝑑𝑡 𝜕𝑞𝑖 𝜕𝑞𝑖 𝜕𝑞𝑖
𝜕𝑥 𝜕 1 2
= 𝑥̈ + ( 𝑥̇ )
𝜕𝑞𝑖 𝜕𝑞𝑖 2
That is
𝑑 𝜕 1 2 𝜕 1 𝜕𝑥
( 𝑥̇ ) − 𝜕𝑞 (2 𝑥̇ 2 ) = 𝑥̈ 𝜕𝑞 *
𝑑𝑡 𝜕𝑞̇ 𝑖 2 𝑖 𝑖
74
with similar equations for y and z.
The next step is to multiply these equations by m, sum over

all particles of the system and add the three resulting
equations together. However, here m is not constant and so
it cannot be taken inside the differentiation signs with
impunity. Let us suppose now that m is independent of
both position and velocity. If that is the case, it may be
taken inside all differentiation signs except d/dt. Hence, if
we carry out the above sequence of procedures but note
this final point, after some algebra one arrives at
𝑑 𝜕 1 𝜕 1
( 𝑚𝑥̇ 2 ) − ( 𝑚𝑥̇ 2 )
𝑑𝑡 𝜕𝑞̇ 𝑖 2 𝜕𝑞𝑖 2
𝜕𝑥 𝑑𝑚 𝜕 1 2
= 𝑚𝑥̈ + ( 𝑥̇ )
𝜕𝑞𝑖 𝑑𝑡 𝜕𝑞̇ 𝑖 2
again with similar equations for y and z.
Note that the right-hand side of this equation may be

written
𝜕𝑥 𝜕𝑥̇ 𝜕𝑥 𝜕𝑥 𝜕𝑥
𝑚𝑥̈ 𝜕𝑞 + 𝑚̇𝑥̇ 𝜕𝑞̇ = 𝑚𝑥̈ 𝜕𝑞 + 𝑚̇𝑥̇ 𝜕𝑞 = (𝑚𝑥̈ + 𝑚̇𝑥̇ ) 𝜕𝑞 .
𝑖 𝑖 𝑖 𝑖 𝑖
Now note that, for variable mass
Force = 𝑚(𝑥̈ , 𝑦̈ , 𝑧̈ ) + 𝑚̇(𝑥̇ , 𝑦̇ , 𝑧̇ ) = (𝑚𝑥̈ + 𝑚̇𝑥̇ , 𝑚𝑦̈ +
𝑚̇𝑦̇ , 𝑚𝑧̈ + 𝑚̇𝑧̇ )
Therefore, the x- component of force is
𝑚𝑥̈ + 𝑚̇𝑥̇
It follows immediately that, as in the case of constant mass,
the right-hand side of these modified Lagrange equations
is equal to the coefficient of 𝑑𝑞𝑖 in the equations of virtual
work. This follows because
75
𝛿𝑊 = 𝑋𝛿𝑥 + 𝑌𝛿𝑦 + 𝑍𝛿𝑧

𝑛
𝜕𝑥
= 𝑋∑ 𝛿𝑞
𝜕𝑞𝑖 𝑖
𝑗=1
𝑛 𝑛
𝜕𝑦 𝜕𝑧
+𝑌∑ 𝛿𝑞𝑖 + 𝑍 ∑ 𝛿𝑞
𝜕𝑞𝑖 𝜕𝑞𝑖 𝑖
𝑗=1 𝑗=1
where X, Y, Z are the components of the force; that is
(𝑋, 𝑌, 𝑍) ≡ (𝑚𝑥̈ + 𝑚̇𝑥̇ , 𝑚𝑦̈ + 𝑚̇𝑦̇ , 𝑚𝑧̈ + 𝑚̇𝑧̇ ).
Hence, it follows that, as far as the case of a mass varying
with time, but not with either position or velocity, the
formal form of Lagrange’s equations of motion remains
unaltered. It is just the form of the components of force
that change. However, it should be noted also that the
expression being differentiated in both terms on the left-
hand side of these modified Lagrange equations is still
1 1
2
𝑚𝑥̇ 2 or 2 𝑚𝑣 2 ; it seems this part of the equations remains
unaltered. However, given that the starting point for these
1
considerations was an examination of a derivative of 𝑥̇ 2 ,
2
this is possibly not too surprising. Again, as in the usual
derivations of Lagrange’s equations of motion, the system
concerned is assumed holonomic. The only difference here
is that the mass is assumed variable and to be dependent
on the time t.
It might be noted also that this form of the Lagrange
equations will hold as long as the mass remains
independent of both position and velocity; that is, it may
depend on acceleration or on even higher time derivatives
of position. This latter point follows because the only
partial derivatives appearing in the derivation are those
with respect to position and velocity. It is also important to
note at this point that no mention has been made of
potential energy; that concept simply hasn’t been
introduced nor has it been needed. As with the usual
derivation of Lagrange’s equations of motion, the right
76
hand side is dependent on the generalized components of

force; the introduction of potential energy and, hence, the
Lagrangian function defined by L = T + V, where T is
kinetic energy and V potential energy, is a later
development in the theory which restricts attention from
then on to conservative fields of force.
Generalization to the cases when the mass depends on

position and/or velocity.
Now suppose the mass depends on the velocity as well; in

other words the mass is dependent on 𝑞̇ . In this case, it is
seen immediately that
𝑑 𝜕 1 𝑑 1 𝜕𝑚 𝜕𝑥̇
( 𝑚𝑥̇ 2 ) = ( 𝑥̇ 2 + 𝑚𝑥̇ )
𝑑𝑡 𝜕𝑞̇ 2 𝑑𝑡 2 𝜕𝑞̇ 𝜕𝑞̇
𝜕𝑚 1 2 𝑑 𝜕𝑚 ̇ 𝑑 𝜕 1 2 𝜕𝑥̇ 𝑑𝑚
= 𝑥𝑥̈ 𝜕𝑞̇ + 2 𝑥̇ 𝑑𝑡 𝜕𝑞̇ + 𝑚 𝑑𝑡 𝜕𝑞̇ (2 𝑥̇ ) + 𝑥̇ 𝜕𝑞̇ 𝑑𝑡 .
In this case, if equation * is multiplied throughout by m, the
resulting equation may be written
𝑑 𝜕 1 𝜕 1
( 𝑚𝑥̇ 2 ) − ( 𝑚𝑥̇ 2 )
𝑑𝑡 𝜕𝑞̇ 2 𝜕𝑞 2
𝜕𝑥 𝜕𝑚
= (𝑚𝑥̈ + 𝑚̇𝑥̇ ) + 𝑥̇ 𝑥̈
𝜕𝑞 𝜕𝑞̇
1 2 𝑑 𝜕𝑚
+ 𝑥̇
2 𝑑𝑡 𝜕𝑞̇
or
𝑑 𝜕𝑇 𝜕𝑇 𝜕𝑥 𝜕𝑚 𝑇 𝑑 𝜕𝑚
− = (𝑚𝑥̈ + 𝑚̇𝑥̇ ) + 𝑥̇ 𝑥̈ +
𝑑𝑡 𝜕𝑞̇ 𝜕𝑞 𝜕𝑞 𝜕𝑞̇ 𝑚 𝑑𝑡 𝜕𝑞̇
where, as previously 𝑇 = 𝑚𝑥̇ 2 /2.
Hence, as soon as the mass becomes dependent on the
velocity, two more terms appear on the right-hand side of
the relevant Lagrange equations of motion and the
equations become more complicated. Considering this is
the situation when the mass is dependent on the velocity,
77
it raises questions about the apparent use of the more

familiar orthodox form of Lagrange’s equations of motion
in some texts concerned with Special Relativity where the
mass is velocity-dependent.
If, further, the mass depends on position q, then
𝜕 1 𝜕 1 1 𝜕𝑚
( 𝑚𝑥̇ 2 ) = 𝑚 ( 𝑥̇ 2 ) + 𝑥̇ 2
𝜕𝑞 2 𝜕𝑞 2 2 𝜕𝑞
𝜕 1 2 𝑇 𝜕𝑚
= 𝑚 ( 𝑥̇ ) +
𝜕𝑞 2 𝑚 𝜕𝑞
and, in this case, the final form of the Lagrange equations
of motion is
𝑑 𝜕𝑇 𝜕𝑇 𝜕𝑥 𝜕𝑚 𝑇 𝑑 𝜕𝑚 𝜕𝑚
𝑑𝑡 𝜕𝑞̇
− = (𝑚𝑥̈ + 𝑚̇𝑥̇ ) + 𝑥̇ 𝑥̈
𝜕𝑞 𝜕𝑞 𝜕𝑞̇
+ (𝑚 𝑑𝑡 𝜕𝑞̇
+ ). 𝜕𝑞
Concluding remarks.
Some of the above results may be familiar to some

readers but they are not, as far as we are aware, to be found
commonly in existing mechanics texts; neither are these
extensions to familiar theory taught in many undergraduate
courses. In general, it seems that, at most, only a cursory
mention is made of variable mass situations in most
courses and then usually only reference to rocket motion.
Here various applicable results have been derived and, in
the case of the examination of the Lagrange equations of
motion, a distinction has had to be drawn among various
individual cases with the resulting equations becoming
more complicated if the mass is dependent on velocity
and/or position. The case when the mass is independent
of velocity and position is interesting since it is possible to
interpret the resulting right-hand side of the equations in a
manner similar to that for the constant mass situation. It
should be noted also that, here, attention is not restricted
to conservative forces and so a potential energy appears
nowhere. Attention was drawn to this point re the usual
derivation of the Lagrange equations of motion presented
78
to most undergraduates in the introduction and it is no less

valid a point here. It is possibly worth noting that the
derivation of the Lagrange equations presented here is, in
common with most derivations, for a purely mechanical
system. Any extension to other systems and situations
would need to be addressed afresh, starting from
fundamental principles and seeing if the approach could be
adapted successfully.
References.
[1] D.P. Allen Jr., Neo-Newtonian Theory – see article under

chapter 4.
[2] J.L. Synge and B.A. Griffith, Principles of Mechanics

(McGraw-Hill, New York, 1959)
.
79
7 WESLEY, HIS NEO-MECHANICS AND
SPECIAL RELATIVITY
Jeremy Dunning-Davies,
Via Santa Gonda, 14 - 59100 Prato, Italy
and
Institute for Basic Research,
Palm Harbor, Florida, U.S.A.
Introduction.
Qualms concerning at least some aspects of the Special

and General Theories of Relativity seem to have existed
ever since both broke into modern science. However, in
more recent years, some very real queries have been raised
concerning the true place those theories should occupy in
science. For example, as far as General Relativity is
concerned, three crucial tests are normally discussed in
most, if not all, textbooks on the subject. However, it has
been shown[1] that, as far as the test associated with the
gravitational red shift is concerned, although the final result
draws on modern results due to Planck and Poincaré, it is
fundamentally deduced from notions of Newtonian
mechanics alone. Further, in a 2005[2] article, Lavenda
showed that the deflection of light and the advance of the
perihelion of Mercury may be treated as diffraction
phenomena on the basis of Fermat’s principle and the
modification of a Bessel function in the short wavelength
limit. He has also shown, in the same article, that the
problem of time delay in radar echoes may be treated by
the same means. Hence, where does that leave the General
Theory of Relativity? It must be admitted that that theory
80
does seem to lead to correct results but it is now clear that

those same results may be obtained by other means which
do not involve any of the ideas of General Relativity. This
certainly denies the General Theory of Relativity the pre-
eminent position it has seemingly occupied in physics for
many years.
However, this says nothing of the Special Theory of
Relativity and possibly even more qualms have been
harbored about that theory for many years and have led to
the almost complete ostracizing of such as Herbert Dingle.
As far as Dingle’s particular case is concerned, the
complete history is chronicled in detail in the book Science
at the Crossroads[3], which may be found on the internet. It
is interesting, therefore, that manipulations by J. P. Wesley
[4] seem to have gone unnoticed.
Although the main object is to make more people
acquainted with the ideas of Wesley, the opportunity has
been taken to extend that gentleman’s results to the case
when the mass of the body when at rest is not a constant
but may vary with time.
Wesley’s Neo-mechanics.
Although rarely mentioned, it is a long known fact that

there is a strong relationship linking mass and energy or, in
other words, the idea of mass/energy equivalence was
known and used long before Einstein’s 1905 article in
which, to many, it was introduced into science for the first
time. In the 1800’s, the topic was investigated at length and
the notion that the two were linked via a coefficient of the
order of c2 was established by electrodynamic
considerations. One pre-1905 reference to this is to be
found in J. J. Thomson’s book Electricity and Matter [5],
which is a printing of a series of lectures Thomson gave at
Yale University in May 1903, but the history of the relation
goes back much further in time. Although there was some
81
doubt for a time over the exact coefficient linking mass and
energy, it became apparent that that coefficient was c2 and
that has been verified experimentally to a high degree of
accuracy. Wesley’s so-called neo-mechanics is based on this
vitally important principle of mass/energy equivalence as
expressed in
E = mc2 (1)
and has absolutely no connection with the space-time of
Special Relativity. For those unacquainted with Wesley’s
work, it is worth examining his deduction.
Where Wesley diverges from traditional
Newtonian mechanics is a result of his noting that, since
mass/energy equivalence is an established fact, if this
applies to any form of energy, it follows that there must be
a mass equivalent for kinetic energy. This fact has to be
included, therefore, in traditional Newtonian mechanics as
a modification. Consider a body at rest whose measured
mass is m. Suppose this same body then moves and when
in motion, possesses a kinetic energy T then, the mass
equivalent is T/c2 and, therefore, the total momentum of
the body is given by
p = (m + T/ c2)v
From Newton’s Second Law, it follows that the force
acting, P, is given by
𝑑
𝑷 = (𝑚 + 𝑇⁄ 2 ) 𝒗
𝑑𝑡 𝑐
Since the rate of working of the force equals the rate of
increase of the kinetic energy, it follows that
𝑑𝑇 𝑑
= 𝒗. 𝑷 = 𝒗. (𝑚 + 𝑇⁄ 2 ) 𝒗
𝑑𝑡 𝑑𝑡 𝑐
2 −1/2
Next note that, if 𝛾 = (1 − 𝑣 ⁄ 2 ) ,
𝑐
𝑑𝛾 𝛾 3 𝑑𝒗
= 𝒗.
𝑑𝑡 𝑐 2 𝑑𝑡
and so
82
𝑑𝑇 𝑑
= 𝒗. (𝑚 + 𝑇⁄ 2 ) 𝒗
𝑑𝒗 1 𝑑𝑇
= (𝑚 + 𝑇⁄ 2 ) 𝒗. + (𝒗. 𝒗) 2
𝑐 𝑑𝑡 𝑐 𝑑𝑡
2 2
𝑐 𝑑𝛾 𝑣 𝑑𝑇
= (𝑚 + 𝑇⁄ 2 ) 3 +
𝑐 𝛾 𝑑𝑡 𝑐 2 𝑑𝑡
Rearranging leads to
𝑑𝑇 1 𝑑𝛾
= (𝑚𝑐 2 + 𝑇)
𝑑𝑡 𝛾 𝑑𝑡
or
1 𝑑𝑇 1 𝑑𝛾
2
=
(𝑚𝑐 + 𝑇) 𝑑𝑡 𝛾 𝑑𝑡
Integrating both sides with respect to t and noting that T =
0 when v = 0 leads to
𝑇 = 𝑚𝑐 2 (𝛾 − 1),
a result normally associated with the Special Theory of
Relativity.
A modification of Wesley’s neo-mechanics.
In the above m, the mass of the body when at rest, is

tacitly assumed to be constant. However, it is not
unreasonable to wonder what modifications would need to
be made to the above if m was not constant; it can be
imagined to alter with time for example. Consequently,
assume m to be a function of time. Under these
circumstances, the equation for dT/dt above becomes
𝑑𝑇 𝑑
= 𝒗. (𝑚 + 𝑇⁄ 2 ) 𝒗
𝑑𝒗
= (𝑚 + 𝑇⁄ 2 ) 𝒗.
𝑐 𝑑𝑡
𝑑𝑚 1 𝑑𝑇
+ (𝒗. 𝒗) ( + )
𝑑𝑡 𝑐 2 𝑑𝑡
83
𝑐 2 𝑑𝛾 𝑣 2 𝑑𝑇 𝑑𝑚
= (𝑚 + 𝑇⁄ 2 ) 3 + + 𝑣2 ,
𝑐 𝛾 𝑑𝑡 𝑐 2 𝑑𝑡 𝑑𝑡
that is,
𝑑𝑇 (𝑚𝑐 2 +𝑇) 𝑑𝛾
𝑑𝑡
= 𝛾 𝑑𝑡
+
𝑑𝑚
𝑣 2 𝛾 2 𝑑𝑡 . (2)
But,
1 𝑑𝑇 𝑇 𝑑𝛾 1 𝑑𝑇 𝑑𝛾 𝑑 𝑇
( − ) = 2 (𝛾 − 𝑇 ) = ( ).
𝛾 𝑑𝑡 𝛾 𝑑𝑡 𝛾 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝛾
Hence, the above equation (2) becomes
𝑑 𝑇 𝑚𝑐 2 𝑑𝛾 𝑑𝑚
( )= 2 + 𝑣 2𝛾 .
𝑑𝑡 𝛾 𝛾 𝑑𝑡 𝑑𝑡
Then, integrating throughout with respect to t,
remembering that m is a function of t, leads to
𝑇 𝑚
= 𝑐 2 ∫ 2 𝑑𝛾 + ∫ 𝑣 2 𝛾𝑑𝑚 + 𝑐𝑜𝑛𝑠𝑡.
𝛾 𝛾
1 𝑚
= 𝑐 2 [∫ 𝑑𝑚 − ∫ 𝑑 ( )] + ∫ 𝑣 2 𝛾𝑑𝑚 + 𝑐𝑜𝑛𝑠𝑡.
𝛾 𝛾
𝑐2 𝑚𝑐 2
= ∫ ( + 𝑣 2 𝛾) 𝑑𝑚 − + 𝑐𝑜𝑛𝑠𝑡.
𝛾 𝛾
𝑚𝑐 2 (𝑐 2 + 𝑣 2 𝛾 2 )
=− +∫ 𝑑𝑚 + 𝑐𝑜𝑛𝑠𝑡.
𝛾 𝛾
Substituting for  and simplifying leads to
𝑇 𝑚𝑐 2
=− + ∫ 𝑐 2 𝛾𝑑𝑚 + 𝑐𝑜𝑛𝑠𝑡.
𝛾 𝛾
Once again T = 0 when v = 0 and so the value of the
constant is given by
𝑚𝑐 2 − [∫ 𝑐 2 𝛾𝑑𝑚]
𝒗=𝟎
Hence, the final result may be written symbolically as
84
𝒗
𝑇 = 𝑚𝑐 2 (𝛾 − 1) + ∫ 𝑐 2 𝛾𝑑𝑚.
𝒗=𝟎
Comments.
It should be noted that none of the above derivation in

the section purely devoted to an introduction to Wesley’s
neo-mechanics is new; it may be found, albeit in a
somewhat abbreviated form, in Wesley’s own book[4].
However, it does appear that this work is not readily
available and it seems largely unknown, or judiciously
ignored, in scientific circles. Although the introduction of
the usual Special Relativity factor  is crucial and it is not
clear how or why that factor is introduced, it is seen that
the expression for the kinetic energy normally associated
with Special Relativity is derived without recourse to the
methods or philosophy of that subject. This is basically the
starting point for Wesley’s so-called neo-mechanics. True;
it does have some of the features of Special Relativity but
that is where any similarity ends. Neo-mechanics has a
totally different base from Relativity and is without the
awkward results which accompany that popularly accepted
theory. Those awkward results of the mathematical theory,
which have caused so many problems over the years and
have led to literally reams of paper devoted to numerous
publications, all attempting to justify a physical
interpretation for what are purely mathematical results, are
missing. However, the useful, physically justified results
remain.
Yet again then, results associated in most peoples’ minds
exclusively with relativity are found to be derivable by
alternative methods. Just as the results alluded to in the
introduction raise grave questions about the place afforded
General Relativity, so this discussion here, highlighting
ideas of Wesley, raises similarly serious questions about the
85
pre-eminent place afforded Special Relativity in modern

science. In the case of General Relativity, no-one was
claiming that subject incorrect; they were simply pointing
out that a lot, maybe all, of the results on which the various
crucial tests are based were obtainable by alternative
methods which relied on mathematical techniques which
had been in place for many years; they did not draw on, for
example, results of Riemannian Geometry. In that case
though, if it became clear that a result was obtainable only
via the new methods that might indicate a need for caution
when using the said result until an alternative derivation is
found. The position in Special Relativity is, however,
possibly more serious because, in that case, there are
several results – to which reference has been made already
– which cause grave problems when any attempt is made
to fit a seemingly feasible physical explanation to them.
With Wesley’s approach, the equations of the Lorentz
transformation simply do not appear and that in itself
removes many awkward aspects of the currently favored
theory. However, the truly crucial result, equation (1)
above, which should not really be regarded as the exclusive
property of Einstein and his Special Relativity, remains and
forms the cornerstone for this new approach.
In the section devoted to a modification of Wesley’s
work, the mass of the body when at rest is assumed to be
non-constant and to be, in fact, a function of time. It seems
it cannot depend on anything else since all effects due to
subsequent motion are contained in the kinetic energy.
However, one can envisage effects which might cause the
mass of a body while at rest to alter over time and so this
slight modification is, it seems, worth considering. Once
again though, it should be emphasized that, although the
familiar factor  is very much in evidence in all the
expressions, there has been absolutely no recourse to the
methods or philosophy normally associated with Special
Relativity.
86
The crucial point to be drawn from all this is that, as with

the examples mentioned associated with General Relativity,
a practically important result usually regarded as being
solely a consequence of Special Relativity has been
obtained without utilizing any of the basics of that subject.
This does not even hint at Special Relativity being incorrect
but does show that some of its results, at least, are
obtainable by other means. This, in turn, casts doubt on the
position and usefulness of those results which have led to
the awkward paradoxes that have caused so much
intellectual effort to be spent on attempting to explain
them; in fact, to find physical explanations for results
which, quite simply, may not admit such explanations since
the results are themselves merely non-physical
consequences of the particular mathematical techniques
initially used to investigate some genuine physical
phenomena.
References.
1. R. F. Evans and J. Dunning-Davies; Galilean

Electrodynamics 18 No. 4 (2007), 71
http://arxiv.org gr.qc/0403082
2. B. H. Lavenda; Three Tests of General Relativity via

Fermat’s Principle and the Phase of Bessel Functions, J. App.
Sciences 5 (2005) 299
http://arxiv.org math-ph/0310054
3. H. Dingle; Science at the Crossroads,

(Martin, Brian & P’Keefe, London, 1972)
87
4. J. P. Wesley; Selected Topics in Scientific Physics,

(Benjamin Wesley, Blumberg, 2002)
5. J. J. Thomson; Electricity and Matter,

(Constable & Co., Westminster, 1904)
88
89
8 APPENDICES
Appendix 1.
It's now clear that formula (3.11) from our neo-Newtonian

theory Chapter 4 is what should be used to explain Prof.
Dmitriev's falling rotor data, but with  = /2, h = 0, and  >>
r2 (the outside radius) and also  >> g. Then we have
approximately cos = g cos /(2), where  is the angle made
by the radius  of an arbitrary point of the rotor with the vertical.
#1: x = [ sin,  cos, l ]
This just above is the position vector of a typical point on the

rotor [note that the last coordinate is "l" (lower case "L"), not
unity], where we are actually really considering a cylindrical shell
which is parallel to the axis of symmetry of the rotor, the rotor
having inside and outside radii r1 and r2, respectively. We will
integrate dMr over the volume of the (hollow) cylindrical rotor to
get the total r-mass Mr of the rotor. Then we SHOULD have the
key equation M + Mc= (Mr +Mc)(g"/g) since the less the inertial
mass Mr compared to the rest (and also gravitational) mass M,
the larger the falling acceleration as there is then less resistance
to the gravitational force, as the gravitational force is unaffected
by spinning at non-relativistic velocities (see pages 19-20 of
Chapter 4 above), where g is 9.806 m/sec2 (the gravitational
acceleration at the earth's surface) and M is the rotor rest mass
0.25 kilograms. Here the graph of Prof. Dmitriev’s is on the
horizontal axis the variable h (the rotor spin in Hertz which is
[2  h] =  radians per second ...after changing units from
Hertz to radians per second) and the vertical axis is in Gal', the
unit of one cm/sec2. Positive Gal' here means that the rotor fell
faster than g. It will be shown (at the appendix end) that the
quadratic least squares fit to the last 28 points on right of his
graph has the result (see #21 below) for Gal':
#2: - 1.377073589x10-6 ω2 +0.0045257932 ω – 3.914102168
90
But we note that Prof. Dmitriev’s falling rotor experiment

involved a case (its interior having the air evacuated from it for
no air friction) that enclosed his rotors during their falling which
allowed the rotors to spin freely, but did not spin itself. Let the
rest inertial mass of this case be Mc (equaling 0.77 kilograms) so
that the total inertial mass of the rotor-case combo is (Mc + Mr),
where Mr is the r-mass of the rotor depending upon , the rotor
spin. If  = 0, then Mr = M. So we see that we have by Newton's
laws that [g''(Mc + M)] = [g Mc + g Mr]. (Here, again, M = 0.25
kilograms.)
Then in Chapter 4 above, we have also equations (3.4) through

(3.11), but we take the special case (as mentioned above) of  =
/2, h = 0, and  >> g and r2 [again with ω = (2 π) ωh]. Thus,
cos = [- g cos /2], whence equation (5.2) yields
approximately):
2 4
#3: f = 1 − 𝐶(𝜌𝜔2 )(𝜌𝜔3 ) (𝑔𝑐𝑜𝑠𝜙
𝜌𝜔2
𝑔𝑐𝑜𝑠𝜙
) − 𝐾(𝜌𝜔2 )2 (𝜌𝜔3 )2 ( 2 )
𝜌𝜔
This just above is the new and revised mass change ratio f. I have
dMr =  dV f = [M/(((r22-r12)l)(l  d d f) = (M  d d / ((r22
– r12 ))(1 - C g2 (cos2) - K ω2 g4 cos4) . Here,  is the rotor
material density (assumed constant), and we evidently have  =
M / ( ( r22 – r12) l )
#4: dV = l  d d
This just above is a volume element of the above described rotor

of length "l", radius differential d, and angle differential d.
Then Mr is the volume integral of dMr which is the  integral and

the  integral of (Mf). But first we note the following:
#5: g" (Mc + Mr) = g Mc + g M
#6: SOLVE (g" (Mc + Mr) = g Mc + g M, g")
91
#7: g" = (g Mc + g M) / (Mc + Mr)
This is the addition of forces on the rotor-case combo as Mc

and M are the rest masses of the case and the rotor,
respectively, while Mr is the r-mass of the rotor. The variable g''
is the falling acceleration of the combo, and so the term on the
left is the combined force on the combo, and the term on the
right is the sum of the individual forces, one on the case and the
second on the rotor ... as the gravitational mass is fixed, but the
inertial mass varies with rotation. (Here, we divide Gal' by 100
because it is cm per second squared, not m per second squared
as we had thought in editions 5 and prior editions; the unit Gal
is only used -- in all of physics – in gravimetry.)
#8: Gal'/100 + g = g"
#9: g - (5.436468 x10 ωh2 + 0.028436397 ωh – 3.91410216)/100
= (g Mc + g M) / (Mc + Mr)
We change from Hertz to radians per second, with ω h being

replaced by (ω/(2 π)), and then:
#10: SOLVE(g - (- 1.377073589x10-6 ω2 +0.0045257932 ω –

3.914102168)/100 = (g Mc + g M) / (Mc + Mr) , Mr)
This is the addition of forces on the rotor-case combo as M and

M are the rest masses of the case and the rotor, respectively, while
Mr is the r-mass of the rotor. The variable g'' is the falling
acceleration of the combo, and so the term on the left is the
combined force on the combo, and the term on the right is the
sum of the individual forces, one on the case and the second on
the rotor ... as the gravitational mass is fixed, but the inertial mass
varies with rotation.
Then we solve for Mr which is formulated from gravitational and

inertial mass considerations, and using the 28 points of Prof.
Dmitriev’s data points having largest frequency.
Next we calculate Mr from the neo-Newtonian theory, where the
92
just below is from our above formula for the mass change
function f.
#11: M  (1 - C g2 (𝜌𝜔2 )(𝜌𝜔3 )(cos2/(2)2 - K (2)2(3)2(g cos
/2)4)) )/( (r22 – r12))
4
#12: ∫𝑟𝑟12 𝜋(𝑟𝑀𝜌 2 2
2 −𝑟 2 ) [1 − 𝐶𝜔𝑔 𝑐𝑜𝑠 𝜙 − 𝐾(𝜌𝜔
2 )2 (𝜌𝜔3 )2 𝑔𝑐𝑜𝑠𝜙
( 2 ) ] 𝑑𝜌
𝜌𝜔
2 1
𝐾𝑀 𝜔2 𝑔4 𝑐𝑜𝑠 4 𝜙 𝐶𝑀 𝜔 𝑔2 𝑐𝑜𝑠 2 𝜙 𝑀
#13: − − +
2𝜋 2𝜋 2𝜋
𝜋 𝐾𝑀 𝜔2 𝑔4 𝑐𝑜𝑠 4 𝜙 𝐶𝑀𝑔2 𝜔 cos(𝜙)2 𝑀

#14: ∫−𝜋 [− 2𝜋
−
2𝜋
+
2𝜋
] 𝑑𝜙
𝑀[4𝐶𝑔2 𝜔+3𝐾𝑔4 𝜔2 − 8]
#15: −
8
This just above is Mr.
This two termed (not counting the constant term) Taylor's series
is exact. Note that r1 and r2 fail to appear here; however, they
do appear when M is written as (constant) density times volume.
Next we consider the square of the difference of the above two

expressions for Mr, and then integrate this expression from the
lowest frequency of our 28 points of Prof. Dmitriev which have
the highest frequency of all his 56 points of data to the highest
frequency of these data points, but converting from Hertz to
radians per second. This integral is a function of the two
constants C and K, and so we minimize this integral so as to
obtain a closest fit by taking the partial with respect to C and also
the partial with respect to K of it, and then set both partials to
zero and solve for C and K. We thereby obtain the unique
solution:
#16: C = 7.007700329 x10-8 K = -5.345870791x10-13
Note that C is positive, but K is negative; and, of course the C

term in our formula for the mass change function f usually
dominates the K term in as much as it is seven orders of
magnitude greater. Having obtained C and K, we can substitute
93
back into the integral from which they came by way of

minimization, and then divide by the upper less the lower limit
of integration in order to normalize our result and obtain an
average squared error, which turns out to be 1.014935148×10-7 ,
and the square root of this is 3.185804683×10-4, which is quite
small relative to the second coordinates of the 28 points of Prof.
Dmitriev’s that have the highest frequencies as can be seen by
inspection from the least squares calculation below. However,
since the unit Gal is 1 cm per second2, we must multiply this last
by 100 to bring it into line with the second coordinates of
Dmitriev’s data, and then we see that our averaged error of about
0.032 is still fairly good.
#17: We now calculate the least squares quadratic fit of Prof.

Dmitriev's falling rotor data. The function FIT is a function of
Derive 5 that has as its first argument the type of curve to be least
squares fitted, in this case a quadratic, and the second argument
is the (28) data points of highest frequency h in Hertz.
94
388 − 1.947
375 − 1.545
362 1.231
353 0.846
342 − 0.4785
332 − 1.035
320 − 2.209
308 − 0.853
298 0.105
287 0.451
277 0.044
270 − 1.0935
260 − 0.464
#18: FIT (𝜔ℎ , 𝑎𝜔ℎ 2 + 𝑏𝜔ℎ + 𝑐) 248 − 0.3545 ,
240 − 0.303
234 0.648
225 0.1725
213 − 1.1275
208 − 0.459
200 0.9035
193 − 0.652
185 0.091
176 − 1.215
163 − 0.88
158 − 1.835
155 − 0.0655
( ( 145 − 0.659 ))
#19: - 5.43646825x10-5 𝜔ℎ 2 + 0.02843639734 𝜔ℎ - 3.9141102168
This is our closest fit to the 28 points of Alex's falling rotor data
having highest frequency h in Hertz. [Actually, my old friend
Tom Phipps counted the points above and noticed that there
were only 27, not 28, and the first author apologizes for this
error.] We convert units from Hertz to radians per second:
#20: h = /2
Here,  is the rotor frequency in radians per second. Now we

calculate our Derive 5 quadratic least squares fit:
#21 - 1.377073589x10-6 ω2 +0.0045257932 ω – 3.914102168
95
This agrees with our quadratic fit introduced above without

derivation, and we have set it equal to Gal' there.
As a final remark, if all 56 of Prof. Dmitriev’s points are used

together with these same data points but with a minus sign
affixed to each frequency – since this experiment actually
employed two identical coaxial rotors spinning in opposite
directions but at the same angular velocity [27] -- and the second
coordinate of the data points being the same as before … while
adding the obvious data point (0, 0), then we have a total of 113
data points reflecting mechanical right-left symmetry that may be
used much as the above 28 points earlier in this appendix to
calculate universal constants C and K, with the results being:
#22 C = 0 and K = -1.786189927 10-13
This 113 point least squares quadratic fit detailed calculation of

C and K (and our corresponding fairly good averaged error term
of about 0.019) so as to include all the data points reported by
Prof. Dmitriev’s and to also reflect mechanical right-left
symmetry may be freely downloaded from the first author’s page
on the ReseachGate.com web site.
96
97
Appendix 2
Here we investigate our relativistic velocity related

mechanical ideas via some calculations. In particular, we
begin by ASSUMING that we have a point particle (viewed
here, however, as having a very small spherical shape) of
rest mass "m0" which is rotating with constant (large)
angular velocity  at radius  > 0 and which is in free fall
at gravitational acceleration downwards of g', where the
acceleration of a point particle not acted upon by non-
gravitational forces is the usual g. We ARE considering
relativistic velocities here, and it seems clear that the total
(relativistic and also neo-Newtonian) measured inertial
mass "m" is given by:
#1: m = m0 + (m0 - m0) + (mr – m0),
which simply says that the total measured mass m is the

sum of the rest mass m0, the relativistic mass increase, and
also the neo-Newtonian mass increase [which, in our work,
is usually in high rotation situations … although even in
straight line motion, there may arise Neo-Newtonian
effects … as there cosα = ± 1 in equation (5.2) of Chapter
4]. To see this, note that (referring to Chapter 7 above) m
= (mr + T/c2) if and only if T = c2 (γ – 1)m0, where T is the
low rotation kinetic energy of a point particle moving with
velocity v. But there (m0 + T/c2) = m0 γ is the relativistic
inertial mass for motion at velocity v, not highly rotational
in nature. And so one might well expect that, in the high
rotation case, equation #1 just above for the total inertial
measured mass m then follows using Wesley’s above
mentioned approach to the low rotation case, but extended
in the obvious way to high rotation by simply replacing m0
by mr (as mr reduces to m0 in the low rotation situation) …
since (again) then mr + T/c2 = m, the total measured inertial
98
mass (including the rotational inertial mass mr). And it

makes sense that the relativistic inertial mass and the
rotational inertial mass increases should be “independent”
like this in as much as the former is strictly a velocity effect
while the latter is strictly an acceleration and its time
derivative effect, and is thus velocity independent (at any
given time)!
#2: m0 + (m0 - m0) + (mr – m0)
Here, mr is the r-mass, and  the relativistic mass increase

factor:
#3: c :  Real (0, )
#4: 1
⁄√(1 − 𝑣 2⁄ )
𝑐2
This just above is .
#5: 𝑚𝑟 = 𝑚0 (1 − 𝐶 𝑎 𝑠 𝑐𝑜𝑠 2𝛼 − 𝐾 𝑎2 𝑠 2𝑐𝑜𝑠 4𝛼)
Here the variable cos is [g cos/2] for the steadily

precessing case (including zero precession angular velocity
as a special case). [See equation (3.11) of chapter 4.] Here,
in this case, we have s = 3 and in the first case, we also
have a = 2 nearly since we assume that  >> g and  (with
h = 0) in the first case we are considering.
#6: (− 𝐾 𝑔4 𝜔2 𝐶𝑂𝑆(𝜑)4 − 𝐶 𝑔2 𝜔 𝐶𝑂𝑆(𝜑)2 + 1) 𝑚0
This just above is "mr", the particle r-mass.

𝑐
#7: 𝑚 = ( − 𝐾 𝑔4 𝜔2 𝐶𝑂𝑆(𝜑)4 − 𝐶 𝑔2 𝜔 𝐶𝑂𝑆(𝜑)2 +
√𝑐 2 −𝑣 2
) 𝑚0
This just above is the total measured mass m in this case.
99
We integrate it from  = -  to  and then divide the result

by (2) radians to obtain the average over one revolution.
𝜋 𝑐
#8: m0 ∫−𝜋 (− 𝐾 𝑔4 𝜔2 𝐶𝑂𝑆(𝜑)4 − 𝐶 𝑔2 𝜔 𝐶𝑂𝑆(𝜑)2 +
√𝑐 2 −𝑣 2
)𝑑𝜑
𝜋 𝑚0 [√𝑐 2 −𝑣 2 (4 𝐶 𝑔2 𝜔 + 3 𝐾𝑔4 𝜔2 )− 8 𝑐]
#9: −
4 √𝑐 2 −𝑣 2
1 𝜋 𝑚0 [√𝑐 2 −𝑣 2 (4 𝐶 𝑔2 𝜔 + 3 𝐾𝑔4 𝜔2 )− 8 𝑐]
#10: [− ]
2𝜋 4 √𝑐 2 −𝑣 2
𝑚0 [√𝑐 2 −𝑣 2 (4 𝐶 𝑔2 𝜔 + 3 𝐾𝑔4 𝜔2 )− 8 𝑐]
#11: [− ]
8 √𝑐 2 −𝑣 2
This just above is the average (relativistic) inertial mass

"mav" of this particle as it rotates with constant angular
velocity  about its circle's center, which is the point
obtained by dropping a perpendicular to the spin axis of
the atom that has this point particle as a nucleon. Now the
formula governing this particle in free fall is g' =(g m0/mav),
where g' is the measured gravitational acceleration of our
rotating (almost) point's particle's center of rotation in free
fall, and "mav" is the average free fall inertial mass of this
(almost) point particle over one revolution. (However,
Ronald R. Hatch reports that hard GPS data indicates that
in the case of relativistic velocities, the gravitational mass
changes from m0 to (m0 / γ) in his monumental “A New
Theory of Gravity …” (2007, Physics Essays.)
#12: g' = g m0 /mav
8 𝑔 √𝑐 2 −𝑣 2
#13: g' =
𝑚0 [√𝑐 2 −𝑣 2 (4 𝐶 𝑔2 𝜔 + 3 𝐾𝑔4 𝜔2 )− 8 𝑐]
This just above is g' in the free falling case. It is not a

function of the rest mass m0.
100
Next, we investigate the stop dead case of impact with the

bottom (or the top) of the container, and we need to
assume that the axis of rotation may not be perpendicular
to the spin axis; however, we CAN assume that the impact
deceleration vector is in the plane containing the unit
vectors i1 and i3 and is perpendicular to the vertical unit
vector j and pointing to the right in Hay's Figure 50 on page
92. This means that the rotor, not nutating nor precessing,
is traveling to the left in that Figure; and so, to stop it, one
must push to the right to counter the assumed initial
motion to the left. We assume that our coordinate system
is that of Hay's Figure 50 (see the diagram at the end of
Chapter 3 above), but we assume zero precession so that
this coordinate system does NOT rotate and so i1 and i3
stay in the plane of the page and i2 is perpendicular to the
page, where i1, i2, and i3 are our triad of orthogonal unit
vectors. The deceleration vector is h = - h [cos, 0, -sin ]
where CONSTANT h < 0 is the magnitude of h except
that we take it negative since we want this deceleration to
be negative and also we want h to point to the right and be
perpendicular to the vertical unit vector j.
#14: m0 + (m0 - m0) + (mr – m0)
This just above is the measured mass "m" ... recalled from
above.
#15: 𝑚0(1 − 𝐶 𝑎 𝑠 𝑐𝑜𝑠 2𝛼 − 𝐾 𝑎2 𝑠 2𝑐𝑜𝑠 4𝛼)
This just above is mr recalled from above, and we substitute

for a, s, and cos according to this stop dead case. We must
now compute a, s, and cos.
#16: [−𝜌𝜔2 𝑠𝑖𝑛𝜑, −𝜌𝜔2𝑐𝑜𝑠𝜑, 0]
101
This just above is the acceleration vector a due to the

spinning of the rotor ONLY. Its magnitude is 2.
#17: - g [sin, 0, cos ]
This just above is the gravitational vector G whose

magnitude is g and points vertically downward.
#18: - h [cos, 0, sin ]
This just above is the deceleration vector h. It points to

the right in Hay's Figure 50. Now we calculate the total
acceleration vector taa.
#19: a + G + h
#20: [-2 sin - h cos - g sin, - 2 cos, -g cos -h sin ]
This just above is the total acceleration vector "taa" for a

point particle spinning at angle  to the vertical, and at
radius  from the origin, and finally at angle  with respect
to the vertical measured clockwise.
Next, our surge vector "s" is the time derivative of a and,

since , , g, and h are constant in time, we have (since  =
 t):
𝒕𝒂𝒂.𝒔
#21: 𝑐𝑜𝑠𝛼 = √(𝒕𝒂𝒂.𝒕𝒂𝒂).(𝒔.𝒔)
#22: [- 3cos,  3sin, 0]
This just above is the total surge vector "s". This just above
is exactly the cos, where  is the angle between "taa" and
"s". We now assume that - h >> ,  and g, and we get:
#23:  :  Real [0, )
102
#24:  :  Real [0, )
#25: h :  Real (-, 0)
#26:
𝑐𝑜𝑠𝜑 (ℎ 𝑐𝑜𝑠𝜃+𝑔 𝑠𝑖𝑛𝜃)
√2𝜔2 𝜌 𝑠𝑖𝑛𝜑(ℎ 𝑐𝑜𝑠𝜃+𝑔 𝑠𝑖𝑛𝜃)+4𝑔 ℎ 𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃+𝑔2 +ℎ2 +𝜔4 𝜌2
This just above is exactly cos, where  is the angle between

vectors taa and s. We ASSUME - h >> , , and g, and get
approximately (by setting  =  = g = 0):
#27: cos cos
This just above is "cos" for -h >> , , and g, where  is

the angle between vectors s and taa. The fact that in our
(3.11) we have sin instead arises from the fact that the
deceleration vector was [0, h, 0] there, but here it is (- h
[cos , 0, -sin ]).
#28: m0 + (mo - m0) + (mr – m0)
𝑚0 2𝜋
#29: 𝑚0 + ( 2
− 𝑚0 ) + (𝑚0 (1 +
𝑐
𝑎 𝑠 𝑐𝑜𝑠 2 𝛼 +
√1−𝑣 ⁄ 2
𝑐
8𝜋2
𝑎2 𝑠 2 𝑐𝑜𝑠 4 𝛼) − 𝑚0 )
𝑐
The just above is the total measured mass "m" first recalled
from above. But in it, "a" is the magnitude of the total
acceleration taa (NOT of the vector a), and this is an
AMBIGUITY as we have said above that "a" is the
magnitude of the vector a. Thus, we calculate the
magnitude of vector taa and then substitute this magnitude
for the variable "a" in our formula for "m" just above:
103
#30: √𝒕𝒂𝒂. 𝒕𝒂𝒂
#31:
√2𝜔 2 𝜌 𝑠𝑖𝑛𝜑 (ℎ 𝑐𝑜𝑠𝜃 + 𝑔 𝑠𝑖𝑛𝜃) + 4𝑔 ℎ 𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃 + 𝑔2 + ℎ2 + 𝜔 4 𝜌2
Next we ASSUME that - h >> g, , and , and

approximate by setting  =  = g = 0:
#32: -h
This just above is our approximation for magnitude of total

acceleration vector "taa", where the minus sign is because
we assume h < 0. We now substitute in the formula above
for m:
#33: - m0(K h2 ω6 ρ2√(c2 – v2) cos4(φ) cos4(θ) – C h ω3 ρ √(c2 –

v2) cos2(φ) cos2(θ) – c) / √(c2 – v2)
We now integrate φ over one revolution and then divide by

(2) radians to obtain the average over one revolution:
#34: - m0 (3 K h2 ω6 ρ2√(c2 – v2) cos4(θ) – 4 C h ω3 ρ √(c2 – v2)

cos2(θ) – 8 c) /(8 √(c2 – v2))
This just above is the average (upon impact) inertial mass

"mav" of this particle as it rotates with constant angular
velocity  about its circle's center, which is the point
obtained by dropping a perpendicular to the spin axis of
the atom that has this (almost) point particle as a nucleon.
That is, "mav" is the average inertial mass upon impact that
we just computed above. However, the atom WILL recoil,
and we have NOT included this in our analysis which has
been a stop dead analysis, but this can be treated by the
conservation of linear r-momentum, and we leave this to
the interested reader. Thus, we see that "mav" is a quadratic
function of h, and then that we have
104
mav -->  as -h --> ,
unless  = /2, of course.
105
ABOUT THE AUTHORS
Dennis P. Allen Jr. earned his doctorate, master’s and

bachelor’s degrees from the University of California at
Berkeley. He has done research work for Bell Telephone
Laboratories and taught mathematics at Michigan
Technological University. In 2011, he published
Foundations of Neo-Spearsian Gravitational Theory with
Application to Earthquake Early Warning Systems (now in its
third edition); and more recently, the fifth edition of The
Reality Oriented Mathematician, a memoir. Also, Part 2 of this
series of two volumes on Neo-Newtonian mechanics has
now been published, and is in its seventh edition. And he
has published Why Does Classica Mechanics Forbid Inertial
Propulsion Devices When Thy Evidently Do Exist? (now in its
fifth edition) concerning mechanical inertial propulsion
devices Additionally, he has published a mathematical
book, “The Lebesgue Measure, A New Point of View” in 2016,
and it goes into the first digit phenomenon and the
continuum problem. See Amazon.com, Researchgate.com
or Academia.edu for the author’s further work. He is
divorced and has three sons and a daughter.
Jeremy Dunning-Davies gained his bachelor’s degree at

Liverpool University before moving to Cardiff University
where he gained a doctorate under the supervision of P. T.
Landsberg. He is the author of three books – Exploding a
Myth, Concise Thermodynamics and Mathematical Methods for
Mathematicians, Physical Scientists and Engineers - as well as
more than 150 academic articles. He is married with a son
and a daughter.
106
107
108

Neo-Newtonian Mechanics With Extension

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Neo-Newtonian Mechanics With Extension

Uploaded by

Copyright:

Available Formats

NEO-NEWTONIAN MECHANICS

PART 1: NON-RADIATIVE EFFECTS

DENNIS P. ALLEN, JR.

All rights reserved.

This book is respectively dedicated to the Holy Spirit of

However, any mistakes are, of course, solely the

Preface to the Second Edition ii

2 Physical content and the closely related hydra effect 5

5 An elementary method for the elimination of the first 65

6 Some results in classical mechanics for the case of a variable 69

7 Wesley, his neo-mechanics and special relativity. 80

The first author (DPA) would like to express thanks to his

The second author (JD-D) would like to thank all who

Preface to the Second Edition

The first author has struggled with Chapter 4 “Neo-

In Prof. O. G. Tietjens’ Preface to the Dover Edition of

Thus we see that similar problems of exposition have arisen

However, the reality concerning the so-called scientific search for

Nevertheless, over the years, people have continued to work on this

The second article is about the question of causality in mechanics that

This short piece is followed by a considerably longer one entitled Neo-

relativistic situations. The article begins by developing a generalized

The fourth contribution is a short piece entitled An Elementary Method

extended. However, it should be noted that the other means of deriving

In the sixth, entitled Wesley, his Neo-mechanics and Special Relativity, an

The two appendices contain detailed calculations using the Derive 5

As a postscript, Hans Lehner kindly informs us that the demonstration

Dennis P. Allen Jr.

17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI

Institute of Basic Research, Palm Harbor, Florida, U.S.A.

In Chapter 1 of “The Cambridge Companion to Newton”,

as to which one to take. However, if one took the wrong

Now, however, the accuracy is 15-18 inches … because

Although it may appear that this notion is not nearly as

place … and so on. Or, to put it another way, the hydra

One might reflect on the great change in attitudes to

much time and effort he expended performing more and

True, many of Newton’s major discoveries were made

Again, in its second role, it is being used to formulate a

reality is never ever achieved completely. It is the lack of

wrong path for just this reason in a variety of fields.

Dennis P. Allen Jr.

17046 Lloyds Bayou Drive, Apt. 322, Spring Lake, MI 49456

There appears to be considerable confusion in classical physics, not

“Causal relations between phenomena are governed by the principle of

“In general, then, according to the principle of causality, an equation

It is evident that Jefimenko sees no way to introduce causality into

Thus French also seems to fall short of demonstrating causality …

It should be noted that the continuum is dearly beloved by

[1] Wilfred Kaplan, Ordinary Differential Equations, 400-1, Addison-

Gyroscope Coordinate Diagram

To begin generalizing Newton’s theory in Part One

change, but this formula has undoubtedly been

Now, to continue our analysis, we consider a (neutral

things and (2) if we move our inertial system to a point

Fg = G m = mr ag + (dmr /dt) v (2.4)

where Fg is the gravitational force on the (point) particle

gravitational force on a particle would evidently not be

We retain Newton's third law (action equals reaction).

We might remark that Herbert A. Simon in [6] discusses

by defining (generalized) force to be the time derivative of

Next, we note that our trouble generalizing Newton's

which, of course, reduces to the usual center of mass xc in