Projectile Motion: The “Coming and Going”

Phenomenon
Cite as: Phys. Teach. 59, 168 (2021); https://doi.org/10.1119/10.0003656
Published Online: 01 March 2021

Williams J. M. Ribeiro, and J. Ricardo de Sousa

ARTICLES YOU MAY BE INTERESTED IN

Have Astronomers Found a Star Older Than the Universe?

The Physics Teacher 59, 154 (2021); https://doi.org/10.1119/10.0003653

Demonstration to Show Resonant Oscillations of a Simple Pendulum Coupled to a Mass-

Spring Oscillator
The Physics Teacher 59, 166 (2021); https://doi.org/10.1119/10.0003655

Hands-on Experiment for Modeling the Baumgartner Jump Using Free-Fall Kinematics
with Drag
The Physics Teacher 59, 111 (2021); https://doi.org/10.1119/10.0003464

Phys. Teach. 59, 168 (2021); https://doi.org/10.1119/10.0003656 59, 168

© 2021 American Association of Physics Teachers.

Projectile Motion: The “Coming and Going”
Phenomenon
Williams J. M. Ribeiro, Instituto de Física, Universidade de S­ão Paulo, São Paulo, Brazil
J. Ricardo de Sousa, Departamento de Física, Universidade Federal do Amazonas, Manaus, Brazil

D
uring elementary physics courses, many examples
in mechanics are studied without considering air
resistance effects. We can mention, for example, pro-
jectile motion and free fall, among many others. Only brief
studies of such systems in classical mechanics courses con-
sider linear air resistance1 (Fdrag ~ v), which is the simplest
model to account for the drag effect by modeling objects
moving in viscous media with small velocities (more precise-
ly, small Reynolds number2). Quadratic air resistance (Fdrag
~ v2), which applies for objects moving in resistive media
with high velocities (high Reynolds number), is hardly men-
tioned, but it can be found in many papers.3-6 However, in
nature air resistance plays a big role in the motion of bodies Fig. 1. Projectiles moving along their trajectories with different
and so it must be included when studying any kind of phe- launch angles (scales in meters).
nomena related to mechanics.
Years ago, Walker7 worked on an interesting phenome-
non related to projectile motion. He showed that, for launch
angles greater than a determined critical angle, the projectile
moves away from the origin, moves back toward it, and then
moves away again, going against our common sense that a
projectile should only move away from the origin. Also, he
studied an important geometrical property displayed by pro-
jectiles, namely the return ellipse, which is deeply connected
to the aforementioned phenomenon. His work was devel-
oped considering projectile motion without air resistance,
which we shall reference as “ideal case” from now on.
The main goal of this paper is to generalize Walker’s work, Fig. 2. Radial distance of the projectiles with different launch
demonstrating if the “coming and going” of the projectile angles (scales in meters).
motion also occurs in presence of a linear air resistance force finally moves away again. This property might be promptly
and if there are any changes in the value of the critical angle, verified by the reader using a ruler in Fig. 1, for example, to
mapping how it is related to the air resistance coefficient. We measure the distance from the origin as the projectile follows
also discuss the effects of linear air resistance in the return its trajectory.
ellipse. In order to analyze this phenomenon, we define the radial
distance as the distance between the origin and any point of
Summary of projectile motion and the “com- the projectile’s trajectory. Mathematically, the radial distance
ing and going” phenomenon can be defined as
Starting with the equation describing the motion of a pro- (2)
jectile without air resistance,8
(1) Now we can see graphically what happens to the radial dis-
tance, while the projectile is moving, if we take a look at Fig. 2.
As we can see, the radial distance increases monotonically for
where is the launch angle, g is the acceleration of gravity, the projectile with launch angle = 65o, while for the one with
and v0 is the launch speed, let us consider a specific situation, = 75o the radial distance increases, decreases, and increases
without loss of generality, where v0 = 10 m/s, g = 9.8 m/s2 , again next to the end of the trajectory. This is a beautiful result
and two launch angles = 75o and = 65o. A graphic is dis- that unveils a remarkable phenomenon.
played in Fig. 1. It can be shown7 that there is a critical angle, unique in the
There is a property displayed by the projectile with launch ideal situation since it does not depend on v0, from which this
angle = 75o not possessed by the one with = 65o. The phenomenon begins to appear. This critical angle is exactly
second one always moves away from the origin while the first
one moves away from the origin, then moves closer to it and

168 THE PHYSICS TEACHER ◆ Vol. 59, March 2021 DOI: 10.1119/10.0003656
 We are now ready to analyze if there is a relation between
the critical angle and the air resistance coefficient in the linear
speed drag force case.

“Coming and going” projectile motion subject

to linear air resistance force
Considering drag force of the form Fr = –m v, the equa-
tion for the trajectory of a projectile is given by9

(3)

Fig. 3. Critical angle (in degrees) as a function of the dimension-

where is the linear air resistance coefficient (it depends on
the projectile’s shape) and m is the projectile’s mass. The initial
conditions are x(0) = y(0) = 0, vx(0) = v0 cos , and
vy(0) = v0 sin .
By using the dimensionless variables t  t, x  2x/g,
y  2y/g, and   v0/g, we can rewrite Eq. (3) as
(4)
where we maintained our notation without primed variables
for convenience. Also, we can evaluate the projectile’s position
using the same dimensionless variables, obtaining9
x(t) =  cos (1 – e–t) (5)
and
y(t) = –t + ( sin + 1)(1– e–t). (6)
less air resistance coefficient.
limit using a routine, because if the program finds a transition
of dr/dt from positive to negative, it immediately associates
the value of  with the value of the critical angle c. The evo-
lution of the critical angle is displayed in Fig. 3. As expected,
when  tends to zero we recover the value of the critical angle
in the ideal case, and when  tends to infinity the critical angle
decreases monotonically.
Although this problem is intractable analytically in its
complete form, we can find a solution by considering the ap-
proximation  << 1 for small air resistance coefficient. This is
an interesting approximation since the experimental values of
 are usually small as we can check in Table I. By setting
dr/dt = 0 in Eq. (7), together with a Taylor expansion, one
finds, after some algebra,
(8)

As we are considering linear drag, it is useful to Table I. Estimates for  in different media, assuming values for properties of
compare it to Stokes’ law,10 F = 6 Rv (Stokes’ law con- different projectiles. All values are in SI units.
siders the force of viscosity F on a sphere of radius R Projectile Medium h (Pa.s)11 m (kg) v0 (m/s) R (m)
in a fluid with dynamic viscosity ), so that the reader
Marble ball Olive oil 0.081 0.005 4.0 0.015 1.87
can picture the practical significance that drag has on
the results. In Table I, we depict some estimates for , Marble ball Motor oil 0.065 0.005 2.0 0.015 0.75
-4
considering different objects and media (actually, the Marble ball Water 8.9 3 10 0.003 3.0 0.015 0.026
objects and media considered are better modeled by a (25 oC)
quadratic drag due to their high Reynolds number). Tennis ball Air (27 ­­oC) 1.86 3 10-5 0.06 5.0 0.05 1.5 3
An attempt to find how the critical angle varies with 10-4
the dimensionless resistance coefficient  using Walker’s Basketball Air (27­­ oC) 1.86 3 10-5 0.62 3.0 0.12 2.1 3
procedure is found not to be effective in order to find an 10-5
analytical solution. However, performing a numerical
analysis gives us a good idea of what happens when
linear air resistance is taken into account. By writing the radial
distance as a time-dependent variable and taking its derivative,
one obtains

(7)

If the derivative in Eq. (7) is equal to zero, then for the chosen
value of  the angle considered in the problem is the critical
angle c (in fact, the critical angle is specifically the smallest
angle for which the derivative ever equals zero, since it equals
zero twice for all angles larger than critical, as we shall see). Fig. 4. Critical angle (in degrees) as a function of the dimen-
sionless air resistance coefficient for  << 1. As we can see, the
Bearing that in mind, it is easier for the computer to find this
numerical solution matches the analytical approximate solution
[Eq. (9)] when  gets close to zero, as expected.
THE PHYSICS TEACHER ◆ Vol. 59, March 2021 169

Fig. 5. Return curves (dashed) for different dimensionless air resistance (which are two times for projectiles with launch an-
coefficients along with projectile trajectories (x and y are dimensionless gles greater than the critical angle).
coordinates).
Fig. 6. Return curves (blue) for different dimensionless air resistance coeffi-
cients along with loci of points for maximum height (red) of projectiles (x and
y are dimensionless coordinates).
For angles smaller than the critical one, the solutions are
imaginary, which means that the projectile steadily moves
away from the origin. Real solutions appear at the angle where
the term in the square root is equal to zero, in which one ob-
tains
(9)
whose graph is displayed in Fig. 4.
The return ellipse
Besides the analysis of the critical angle, Walker also
analyzed an interesting geometrical property of projectiles
related to the critical angle: the return ellipse. This is defined
as the locus of points (x, y) for which dr/dt = 0, that is, the
transition points in the radial distance when varying (with
 constant). In the ideal case, Walker showed that this locus
170 THE PHYSICS TEACHER ◆ Vol. 59, March 2021
of points corresponds to an ellipse centered at x = 0
and y = ymax/2, and that the physical significance of
the ellipse is that projectiles moving with trajectories
crossing the ellipse when moving toward the ground
(dy/dt < 0) are also moving toward the origin (dr/dt <
0), which means that, for these projectiles, the launch
angle is greater than the critical angle.
In order to see what happens to the return ellipse
when linear air resistance is taken into account, we
plotted in Fig. 5 the corresponding return “ellipses”
obtained numerically for six values of  along with
the trajectories of projectiles for different launch
angles. As we can see, the curves are not ellipses any-
more due to the drag, which is why we shall call them
“return curves” from now on. We obtained these
curves by evolving Eq. (7) and saving the positions
at the transition points where dr/dt changes its sign
As expected, for  = 0.01 and  = 0.1, projectiles
with launch angles = 85o, = 80o, and = 75o are
the only ones passing inside the return curve while
moving toward the ground. However, for  = 0.5
projectiles with smaller launch angles (below the crit-
ical angle in the ideal case) begin to enter the return
curve. The new projectile falling to the ground inside
the return curve is the one with launch angle = 65o.
This is in agreement with Fig. 3, in which we see that
the critical angle for  = 0.5 is smaller than 65o. The
subsequent graphics also agree with Fig. 3. For exam-
ple, we see that for = 1.0 the projectile with launch
angle = 60o is entering the return curve, for  = 1.5
the projectile with launch angle = 55o is about to
enter the return curve, and for  = 2.0 the projectile
with launch angle = 55o has already entered the
return curve.
Walker also demonstrated that, in the ideal case,
the locus of points corresponding to dr/dt = 0 is coin-
cidentally the same locus of points corresponding to
dy/dt = 0 (maximum height), a fact that he stated as
“rather unexpected” since these two requirements correspond
to very different conditions. Surprisingly, when air resistance
is taken into account, this property is not satisfied anymore.
In order to see this, for plotting the locus of points for dy/dt
= 0, let us use the equations for the dimensionless maximum
height of the projectile considering air resistance, obtained by
eliminating the time variable with the condition dy/dt = 0 ap-
plied to Eq. (6) together with Eq. (5). One obtains9
(10)
and
=  sin – ln(1 +  sin ).
ymax (11)
The plots are shown in Fig. 6. As we can see, for  = 0.01,
the two curves still agree with each other, but when we go to
higher values of  they do not match anymore, and the differ-
ence between them increases for larger . Therefore, the con-

Fig. 7. Approximate return curves (blue) for different dimensionless air ball rolled in Scotch tape or even water from a hose.
resistance coefficients along with loci of points for maximum height (red) of The main goal could be trying to reproduce Fig. 5 by
projectiles (x and y are dimensionless coordinates).
ditions dr/dt = 0 and dy/dt = 0 correspond to the same locus of
points only in the ideal case  = 0.
As a consistency check, it is interesting to see if our approx-
imate solution (8) is still valid for the construction of the locus
of points for dr/dt = 0 at low . By substituting Eq. (8) into
Eq. (5) and Eq. (6), one finds the plots displayed in Fig. 7. For
small values of  (0.0001 and 0.001), we do not see visible dif-
ferences between the curves, while for  = 0.01 we can see that
our approximate solution is not valid anymore.
Conclusions
The aim of this paper was to check how an interesting
phenomenon that occurs in projection motion, the “coming
and going,” is affected by linear air resistance force. First, we
summarized how this phenomenon occurs for an ideal situ-
ation, without air resistance, and then showed how linear air
resistance affects the phenomenon.
In the ideal situation, it was demonstrated that the critical
angle is unique and does not change even if other variables
are changed, such as the initial velocity of the projectile. By
including air resistance, it was possible to show that the crit-
ical angle for the phenomenon is not unique anymore, but it
changes with the linear air resistance coefficient.
Figure 3 shows that the curve relating the critical angle to
the dimensionless air resistance coefficient is decreasing. This
is an expected result, since with an increasing air resistance
coefficient the projectile suffers an increasing resistance force,
reducing the value of the critical angle. Another interesting
observation coming from Fig. 4 is that for a null resistance co-
efficient the critical angle tends approximately to = 70.528o,
the exact result found theoretically in the ideal projectile mo-
tion, further confirming the numerical analysis.
Our approach turned out to be effective for three reasons:
when considering the  = 0 case in Eq. (9), we recovered the
result for the ideal case for c; Fig. 4 shows that the approximate
and numerical solutions match in the limit  << 1; Fig. 7 shows
that the approximate and analytical solutions also match in the
limit  << 1.
We also demonstrated that the matching between
the return ellipse and the locus of points that define the
maximum height of projectiles does not exist anymore
when air resistance is accounted for, which is why we
quoted this geometric pattern as “return curves” in-
stead of “return ellipses.” This is a beautiful result that
is certainly not trivial at first sight.
In Walker’s paper, he outlines an experiment that
could be designed by students to probe the “coming
and going” phenomenon, in which one videotapes the
trajectory of a basketball and thus measures, frame
by frame, the radial distance. In our case, a basketball
would not be a good choice since air resistance would
not have much influence in the trajectory if initial
velocity is small (see Table I). Therefore, one could
substitute a basketball by an object that could suffer in-
fluence of air resistance more directly such as a paper
changing the initial velocity and the launch angle for
the chosen projectile. As an extension to this work,
it would be interesting to consider more realistic situations
such as quadratic drag, lift, wind, etc., retaining a numerical
approach in order to analyze these cases.
References
1. S. T. Thornton and J. B. Marion, Classical Dynamics of Particles
and Systems, 5th ed. (Thomson, 2004), p. 58.
2. J. P. Owen and W. S. Ryu, “The effects of linear and quadratic
drag on falling spheres: An undergraduate laboratory,” Eur. J.
Phys. 26, 1085–1091 (2005).
3. S. Ray and J. Fröhlich, “An analytic solution to the equations of
the motion of a point mass with quadratic resistance and gener-
alizations,” Arch. Appl. Mech. 85, 395–414 (2015).
4. C. H. Belgacem, “Range and flight time of quadratic resisted
projectile motion using the Lambert W function,” Eur. J. Phys.
35, 055025 (2014).
5. K. Yabushita, M. Yamashita, K. Tsuboi, “An analytic solution of
projectile motion with the quadratic resistance law using the
homotopy analysis method,” J. Phys. A Math. Theor. 40,
8403–8416 (2007).
6. G. F. L. Ferreira, “The ballistic problem with square velocity air
resistance,” Revista Brasileira de Ensino de Física 23, 271–275
(2001).
7. J. S. Walker, “Projectiles: Are they coming or going?” Phys.
Teach. 33, 282–284 (May 1995).
8. D. Halliday, R. Resnick, and J. Walker, Fundamentals of Physics
(Extended version), 9th ed. (Wiley, 2011), p. 67.
9. S. M. Stewart, “On the trajectories of projectiles depicted in ear-
ly ballistic woodcuts,” Eur. J. Phys. 33, 149–166 (2012).
10. L. D. Landau and E. M. Lifshitz, Fluid Mechanics, Vol. 6 (Elsevi-
er, 2013), p. 66.
11. D. R. Lide, “CRC Handbook of Chemistry and Physics,” J. Am.
Chem. Soc. 128, 16 (2006).
Instituto de Física, Universidade de S­ão Paulo, São Paulo, Brazil
williamsribeiro92@gmail.com, jsousa@ufam.edu.br
THE PHYSICS TEACHER ◆ Vol. 59, March 2021 171