Mirror Images

Author(s): David Emil Thomas

Source: Scientific American , Vol. 243, No. 6 (December 1980), pp. 206-229
Published by: Scientific American, a division of Nature America, Inc.
Stable URL: https://www.jstor.org/stable/10.2307/24966483

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Scientific American, a division of Nature America, Inc. is collaborating with JSTOR to digitize,
preserve and extend access to Scientific American

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 Mirror Images
The mirror image is not always a faithful reflection: it can be

inverted, reversed in handedness or distorted in other ways. The

transformations are analyzed through a few basic curved mirrors

by David Emil Thomas

5
T
he ways in which a plane mirror ponents from which the mirrors are cre­ tween the numerals and 6. (It is as­
does (and does not) transform the ated. Indeed, it is often possible to pre­ sumed throughout this discussion that
objects it reflects are familiar to dict the appearance of images reflected the observer and the object are close to
everyone. The image of a right hand ap­ in even the most elaborate compound the axis of the mirror: the line through
pears as a left hand, but its shape and mirrors. In order to understand com­ the mirror's center of curvature that is
size are unaltered. As a visit to a fun pound mirrors and the images they gen­ perpendicular to the tangent plane at
house shows, however, not all mirrors erate, then, it is necessary to begin by that point. )
generate such faithful images. Plato ob­ discussing the light-reflecting properties This coordinate system provides a
served that a piece of reflecting metal of the six basic mirrors. means of explicitly describing the elu­
bent into a concave cylinder can form an sive distinction between handedness and
image of a right-hand glove that is itself
right-handed. More generally, by intro­ C onsider, first, a puzzling and funda­
mental question about the plane
orientation. The numerals in the plane­
mirror image of a transparent clock face
ducing various curvatures into reflect­ mirror: Why does such a mirror ex­ run in counterclockwise order, just as
ing surfaces it is possible to create mir­ change right and left but not up and the numerals viewed through the back
rors that change the shape, size, orienta­ down? In other words, why is it that of the clock face itself do. Compar­
tion and handedness of the objects they when a book is held up to a plane mirror, ing the real object with the image "face
reflect in dramatic and disturbing ways. each letter appears reversed and the se­ to face," then, it is apparent that nei­
The number of different curved mir­ quence of letters runs from right to left, ther the direction 9:00 (right) nor 3:00
ror surfaces is infinite, but they could whereas the order of the lines from the (left) nor 12:00 (up) nor 6:00 (down) is
all be assembled from just a few kinds top of the page to the bottom is un­ changed by the plane mirror. Because
of basic region. Any small region of changed? This paradox can be traced to orientation is not changed along any di­
a curved surface can be described in the confusion between the handedness rection in the plane of the mirror, the
terms of its curvature along a pair of of an image and its orientation. Actually numerals in the image of a clock face
perpendicular axes whose origin lies at for an observer facing a plane mirror the run in counterclockwise sequence. But
the region's center of curvature. The di­ directions right and left are not changed plane mirrors (and all other mirrors) do
rections of these principal axes are de­ any more than the directions up and invert the directions forward and back­
termined by the directions of minimum down. This fact can be demonstrated by ward: whereas the actual clock's front
and maximum curvature on the surface. comparing a clock face with its image in faces away from the observer, the image
The curvature of each axis can be con­ a plane mirror. clock's front faces toward the observer.
vex, concave or planar. The axes serve The familiar geometry of a clock face This single inversion creates an image
to define six basic types of reflecting sur­ makes it a useful standard object with with reversed handedness. It is for the
face: each type is a simple mirror whose which to study object-image relations same reason that the image of a right­
overall shape is defined by a pair of for differently curved mirrors. In my handed glove is left-handed. To avoid
principal curvatures. These six basic own investigations I rely on a clock face confusion in what follows I shall em­
surfaces are the plane mirror, for which that has the ring of numerals on a trans­ ploy "invert" to describe changes in the
both principal axes are planar; the con­ parent plastic backing. An observer orientation of an image and "reverse'"
vex mirror, for which both are convex; holding such a clock face toward a mir­ to describe changes in handedness.
the concave mirror, for which both are ror is able to view not only the mirror
concave; the convex cylinder, for which image but also the real object, at least hen an object is held parallel to
one axis is convex and one is planar; the from the back. Through the transparent W the surface of a plane mirror, the
concave cylinder, for which one is con­ backing the observer sees the numerals forward-backward inversion is evident.
cave and the other is planar, and the of the object clock (reversed and run­ If a .clock face is held perpendicular to
saddle, for which one is convex and the ning counterclockwise) as well as the the mirror, however, the same transfor­
other is concave. images they generate. mation has a different result. When the
Over the past five years I have been The transparent clock face is particu­ clock face is perpendicular to the mir­
studying reflecting surfaces such as bells larly useful in that its numerals serve to ror and in a vertical position, the for­
and toruses that can be constructed by define lateral directions in the plane tan­ ward-backward inversion transposes 3
joining mirrors of these six types. The gent to the mirror at its center of curva­ and 9 (as well as 1 and 1 1, 2 and 10, and
images created by such compound mir­ ture. For example, 1 1:00 identifies the so on), so that right and left appear to be
rors are diverse, but I have found that direction from the center of the clock inverted. Similarly, if the clock is held
they also have certain striking regulari­ face to the numeral 1 1; similarly, 5:30 perpendicular to the mirror and in a hor­
'
ties: underlying traits that can be attrib­ identifies the direction from the center izontal position, the forward-backward
uted to the character of the basic com- of the clock to the point midway be- inversion transposes 6 and 12 (as well

206
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
MULTIPLE IMAGES of a single object, in
this instance an uppercase F, can be viewed
in a "matrix mirror" with a complexly curved
surface, shown here in top and oblique views.
The smoothly connected surface of the mir­

F.
ror consists of convex, concave and saddle­
shaped regions, each of which reSects an im­
age of the single, fixed Owing to the differ­
ent light-reSecting properties of the various
regions som e of the images have reversed
handedness whereas others are not reversed.
The matrix mirror shown is one of an infinite
variety of compound mirrors that can be cre­
ated by joining mirrors of six basic types.

207
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 1
as and 5, 11and 7, and so on), so that light when it is reflected from the sur­ convex mirrors, which do not bring
the mirror seems to invert up and down. face of a curved mirror. The incident wave fronts to convergence, cannot in­
A plane mirror creates what is tradi­ wave front is spherical and diverging, troduce lateral inversions. Moreover,
tionally called a virtual image: an im­ but after reflection it can be either when an observer is close to a curved
age that seems to be behind the surface spherical or asymmetrical, and it can be mirror, no matter what the inverting
of the mirror. In other words, the re­ either diverging (possibly at a rate dif­ properties of the mirror are, the image is
flected light waves that form a plane­ ferent from that of the incident wave always similar to a plane-mirror image.
mirror image appear to emanate from a front) or converging, that is, parts of the The wave fronts do not have space in
point behind the mirror. The fundamen­ wave may be coming together. It is the which to converge before they reach the
tal optical law of reflection states that passing of one part of the wave through observer's eye, and so the image is up­
the angle of reflection of a light ray must another that creates a lateral inversion. right, reversed and (compared with a
be equal to the angle of incidence. From The crossing takes place at the point plane-mirror image) neither much mag­
this law it follows that the image of an where the mirror focuses incoming light nified nor much reduced. An image sim­
object in a plane mirror seems to be as waves to form what is called the real ilar to a plane-mirror image is also in­
far behind the mirror as the object is in image of the object. The directions variably formed when the object (rather
front of it. The image of an object three along which the mirror causes parts of a than the observer) is close to the surface.
feet in front of the mirror appears with wave to converge determine which lat­ The explanation in this case, however, is
the same size, shape and orientation (but eral directions are inverting. If orienta­ that for an object up against a mirror the
with the opposite handedness) as an ob­ tion is laterally inverted along both prin­ section of the reflecting surface that ac­
ject placed three feet behind the mirror. cipal axes of a curved basic mirror, an tually forms the image is almost planar.
Almost the same analysis can be ap­ image with reversed handedness is gen­
plied to convex mirrors, which also give erated. If orientation is inverted along
rise to virtual images. Like a plane mir­ only one lateral axis, a nonreversed im­ Aconcave mirror can give rise to at
least three distinct kinds of mirror
ror, a convex mirror inverts orientation age is generated. The lateral inversions image. To begin with, it can generate
only along the forward-backward direc­ are in addition to the forward-backward either a real or a virtual image, depend­
tion and not along any lateral direction; inversion common to all mirrors. In ing on the position of the object with
hence it reflects upright, reversed im­ general if the number of inversions (in­ respect to the mirror's focal point: the
ages. Because the surface has a convex cluding the forward-backward one) is point where incoming planar wave
curvature, however, the reflected light odd, the image has reversed handed­ fronts either converge or appear to di­
waves seem to emanate from a point ness; if the number is even, the image verge. For an object between the focal
that is not as far behind the mirror as the has nonreversed handedness. point and a concave mirror the reflected
point where a plane-mirror image is ap­ The position of the observer is an im­ wave fronts diverge, seemingly from a
parently positioned. The convex-mirror portant factor in the formation of mir­ point behind the mirror, so that the im­
image is therefore reduced in size com­ ror images. An observer can see an age is a virtual one. For an object placed
pared with its plane-mirror counterpart. inverted, real image only if the light re­ beyond the focal point of the mirror, re­
Any lateral direction along which a flected from the mirror is brought to a flected wave fronts converge along all
curved mirror is planar or convex can focus at a point between him and the lateral directions to focus at a point on
be called a noninverting direction. What surface of the mirror, that is, the observ­ the observer's side of the mirror. The
constitutes an inverting direction? Con­ er must intercept the wave fronts only image in this case is real.
sider what happens to a wave front of after they have converged. Plane and Therefore when an object is between

PLANE CONVEX CONCAVE CONVEX CYLINDER CONCAVE CYLINDER SADDLE

(color)
BASIC CURVED MIRRORS from which compound mirrors are tile convex cylinder, for which one curvature is convex and the other
constructed are described in terms of a pair of perpendicular axes is planar; the concave cylinder, for which one curvature is concave

F
that pass through each mirror's center of curvature. The di. and the other is planar, and the saddle, for which one curvature is con­
rections of the principal axes are determined by the directions of min· cave and the other is convex. The image-forming properties of the
imum and maximum curvature on the surface; the shape of each mirrors are illustrated in the lower panels, where an uppercase
axis can be convex, concave or planar. The axes serve to define six serves as the standard reflected object. In som e cases a single mirror
types of reflecting surface: the plane mirror, for which both curva. can give rise to diff�ent images in different circumstances. It is as·
tures are planar; the convex mirror, for which both curvatures are sumed that both the observer and the object are positioned on the axis
convex; the concave mirror, for which both curvatures are concave; line: a perpendicular line through the mirror's center of curvature.

208
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 © 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
The Turbulence Parameter
Energy-efficient operation of the internal combustion engine
requires the highly turbulent movement of fuel and air in
the chamber. Recent advances at the General Motors Research
Laboratories provide a new basis for determining what degree
of turbulence will get the most work from each drop of fuel.

Flame Speed Behavior � HOUT TURBULENCE,

the highly agitated motion of cylin­
Drs. Frederic Matekunas and
Edward Groff. Their investigative
der gases, combustion would take results have been incorporated into
place too slowly for the gasoline a model that successfully predicts
engine to function. Predicting the effect of engine design and
combustion behavior in order to operating conditions on power and
design engines with greater fuel fuel economy.
efficiency depends upon under­ The researchers separated
standing the relationship between their experiments into two phases.
vital, turbulent gas motions and In the first phase, they measured
burning rate. The challenge is to turbulen�e in the engine cylinder;
quantify this relationship-a com­ in the second phase, they deter­
plex task made more difficult by mined flame speeds over a broad

60 80
40 (mm)
20 Radius
the requirements of measuring a range of operating conditions. Test­
transient event occurring in a few ing t o o k p l a c e i n a s p e c i a l l y
Flame mil liseconds within a small, designed, single-cylinder engine
Burning velocity plotted as a function of flame confined space. equipped with a transparent piston
radius. Combustion stages are indicated by
roman numerals. New knowledge of how tur­ to permit high-speed filming of the
bulence affects flame speed has combustion event.

(II);
High·speed photographs showing flame evolution
(lasting six milliseconds) through four stages:
been revealed in fundamental stud­ Hot-wire anemometry was
initiation (I); flame growth full development i e s c o n d u c t e d at t h e G e n e r a l applied to measure the turbulent
(Ill); termination (IV). Motors Research Laboratories by fl o w s w h i l e t h e e n g i n e w a s
operated without combustion.
Instantaneous velocities were cal­
culated from the anemometer sig­
nals and simultaneous mea­

400,000
surements of gas temperature and
pressure. More than pieces
of data were processed for each
ten-second measurement period.
The significant measure of
turbulence is its "intensity," defined
as the fluctuating component of
velocity. Because conditions in the
cylinder are both transient within
cycles and variant between cycles,
separating the fluctuating and
mean components of velocity
is inherently difficult . T h e re­
searchers overcame this problem
by using a probe with two orthog­
onal wires properly aligned with
the direction of the mean flow.

© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 In the combustion phase,
tests were performed at over one
SL, the laminar flame speed-a
known function of pressure, tem·
THE
hundred operating conditions of perature and mixture composi· MEN
varied spark timing, spark plug tion-is the flame speed that would BEHIND
location, engine speed and intake
valve geometry. Detailed thermo·
exist without turbulence. The vari·
able u' is the turbulence intensity.
THE
dynamic analyses were applied to PR repr.esents a pressure ratio WORK
the recorded cylinder pressures to accounting for combustion· Drs. Matekunas
calculate flame speeds throughout i n d u c e d c o m pr e s s i o n o f t h e and G r o ff a r e
combustion. High·speed films were unburned mixture. The dimension· senior engineers
analyzed frame by frame to vali· less factor f3 accounts for the effect in t h e En g i n e
date flame speeds and to charac· of spark timing on geometric dis· Research De­
terize how gas motions influence tortion of the flame which occurs partment at the General Motors
the initial flame. during the first combustion stage Research Laboratories.
The researchers used these and persists into the later stages. B o t h r e s e a r c h e r s h o ld
measured flame speeds, turbulence T h e r e s e a r ch e r s also un d e r g r a d ua t e a n d g r a d u a t e
intensities, and the conditions observed that the burning velocity degrees i n the field o f mechanical
under which they occurred to for· in the second stage increases in engineering.
mulate a burning law for engine proportion to flame radius, and that Dr. Matekunas (right) re­
flames. They divided the combus· in predicting the energy release ceived his M.S. and Ph. D. from
tion event into four stages. The rate from the burning velo city Purdue University, where he com­
initiation stage begins with igni· equation, it is necessary to account pleted graduate work in advanced
tion and ends as the flame grows to for the finite flame·front thickness. optics applications.
consume one percent of the fuel "The form of our burning Dr. Groff (left) received an
mass. In the second stage, the equation," says Dr. Matekunas, M.S. from California Institute of
flame accelerates and thickens in "shows a satisfying resemblance to Technology and a Ph. D. from The
response to the turbulent field. The expressions for non·engine flames. Pennsylvania State University. His
third stage exhibits peak flame This helps link complex engine doctoral thesis explored the com­
speed. In the final stage, the thick combustion phenomena to the bustion of liquid metals.

1973,
flame interacts increasingly with existing body of knowledge on tur­ General Motors welcomed

1977.
the chamber walls and decelerates. bulent flames." Dr. Matekunas to its staff in
"We see this extension," adds and Dr. Groff in

O
Dr. Groff, "as a significant step

RAN
toward optimizing fuel economy in
VER THE GE of turbu· automotive engines."
lent intensities encountered in
engines, the researchers were able
to describe the turbulent burning
velocity, Sr, during the critical
third stage of combustion with the

+ 1.2 u' PR082 f3

expression:

Sr = 4.0 SL

© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 His business is world travel.
His watch is Rolex.
From bomber airline executive.
squadron duty in the His manner is precise,
flak-filled skies of World exact. . .like the Rolex he
War II to the challenging chooses to wear.
office of Chairman and Precision and
Chief Executive Officer exactness are Rolex
of Pan Am, W illiam T. guidelines. What you
Seawell's firm hands at might expect from a
the controls have watch that is handcrafted
proved over and over again in 192 individual steps
his qualities as a leader from a solid block of gold
of men. or stainless steel. The
Leadership. It's what unique Oyster case is
it took to accomplish the world famous for its
largest airline merger in unparalleled durability.
history. Leadership. It's The handsome movement
what sets Bill Seawell's world renowned for
sights on horizons more its reliability.
distant than most can Pan Am: An airline
see. His personal atten­ that goes to the world's
tion has meant billions of most prestigious places.
dollars in expanding Rolex: A watch that goes
tourism and international commerce. on the world's most prestigious wrists.

�
His calm, take-charge style prevails,
whether as an Air Force general or an
ROLEX

Write for brochure. Rolex Watch, U.S.A., Inc., Dept. 144,RolexBuilding, 665
Fifth Avenue, New York, NYlO022.
World headquarters in Geneva. Other offices in Toronto and major cities around the world.

© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
the focal point and a concave mirror, the

d
mirror' generates a virtual image for an
observer at any distance. The image is
an upright, reversed reflection that is
magnified compared with a plane-mir­
ror image. When the object is beyond ts:rrr
the focal point, there are a number of
possibilities. For example, when the ob­
£Jlo�Oo�
z,,'t •
server is between the mirror surface
and the point at which the real image is
focused, he intercepts the converging
wave fronts before they have actually
converged. In this case it is the observer
who must bring the wave fronts into fo­
cus, and the result is a blurred image
that is upright, reversed and larger than
a plane-mirror image would be.
When the observer is beyond the posi-
tion of a real image, the concave mirror
introduces. inversions along all lateral
directions. In other words, when both
the observer and the object are suffi­
ciently far from the surface of the mir­
ror, a transparent clock face is reflected
so that the image of each numeral ap­
pears in a position 180 degrees (or six

1
hours) around the clock face. The nu­
meral appears at 7:00 and 7 at 1:00, 2
appears at 8:00 and 8 at 2:00, and so on.
Since the directions 12:00 (up) and 6:00
(down) are inverted, the image is upside
down. The positions of the object and
the observer determine whether the im­
age is magnified, reduced or the same
size compared with its plane-mirror
counterpart. Since orientation is invert­
ed along all lateral directions, it is in­
verted along both principal axes of cur­
vature; these two inversions and the
front-to-back one make an odd num-

f
ber, and so the image is not only invert-

c
ed but also reversed.
With a concave mirror and an object
placed beyond the focal point it is possi­
ble for the observer and the mirror im­
,.,''
age to be at about the same point on the
mirror axis. In this instance the result is aJ·OJ0
JIS' S
a totally magnified image: an unrecog­
nizable image that "lights up" the entire
surface of the mirror. The same kind of
image can be seen on any curved mirror
that can generate a real image.

C urved mirrors that are not sections

of a sphere generally reflect images
CLOCK FACE with a transparent backing serves to define lateral directions in the plane tan­
that are transformed in an asymmetric­
gent to a mirror at its center of curvature. For example, 2:00 identifies the lateral direction
al way. For example, the mirror images
from the center of the clock to the numeral 2 and 10:30 identifies the direction from the center
observed in oblate convex and concave
to the point midway between the numerals 10 and 11. An observer holding such a clock up to a
mirrors are, compared with the corre­

(a),
mirror is able to see the real numerals (reversed and running counterclockwise) as weIl as their
sponding plane- or spherical-mirror im­ mirror images. When such a clock is held paraIlel to the surface of a plane mirror and the real
ages, elongated or shortened in certain and image clocks are compared "face to face" it becomes evident that the plane mirror
lateral directions. does not invert orientation along any lateral direction: the 2 in the image appears at 2:00, the
When the curvature along both prin­ 10 at 10:00 and so on. Neither 12:00 (up) nor 6:00 (down) nor 9:00 (right) nor 3:00 (left) is an in­
cipal axes of a nonsymmetrical mirror is verting direction. Although the plane mirror introduces no lateral inversions, the handedness
known, it is possible to predict the gen­ of the image is reversed. Moreover, a plane mirror, like any other mirror, does invert orienta­

(b) (c) (d)

tion along the forward-backward axis, so that an observer looks into the mirror to see objects
eral appearance of the images. Consid­
in front of it. If the clock face is held perpendicular to the surface of the mirror, either right
er the mirror that is shaped like a con­

if (e),
and left or up and down are inverted. A convex mirror does not invert orientation
vex cylinder. This basic curved mirror along any lateral direction either, and so it too creates an upright, reversed image. For a con­
acts approximately like a convex mirror cave mirror, both the observer and the object are sufficiently close to the surface orienta­
along directions parallel to the convex tion is not inverted along any lateral direction, so that once again an upright, reversed image is
curvature and like a plane mirror along formed. If the observer and the object are sufficiently far from the concave mirror, however,
directions parallel to the plane curva- orientation is inverted along all lateral directions. An inverted and reversed image is formed.

213
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
ture, creating images that are reduced in that would appear in a convex cylinder. server or the object is sufficiently close
size along the first axis but neither re­ Similar predictions can be made for to the surface of a concave cylinder
duced nor magnified along the second. the images reflected in a concave cylin­ (close enough so that in the same situa­
By applying the rules of plane geometry der, a basic mirror that acts approxi­ tion a simple concave mirror would cre­
it is possible to predict from the convex­ mately like a concave mirror along one ate an upright, reversed and magnified
and plane-mirror images a fairly accu­ principal curvature and like a plane mir­ image), the image reflected by the con­
rate representation of the single image ror along the other. When either the ob- cave cylinder is upright, reversed and

INCIDENT WAVE FRONTS

OBJECT

/ •
/
IMAGE
REFLECTED WAVE FRONTS

d e f

(c).
SIZE, ORIENTATION AND HANDEDNESS of a mirror image flected wave fronts appear to diverge from a point farther behind the

(gray) (color) (d),

are determined by the way in which wave fronts of light are trans­ mirror than the point where a plane-mirror image would be formed

(a)
formed by reflection. The wave fronts emanating from the object As a result an upright, reversed image that is larger than its plane­
are spherical and diverging. The reflected wave fronts mirror counterpart is created. When the object is at the focal point

(black dot) (colored dot)

can be diverging or converging. For a plane mirror the reflected the reflected wave fronts are flat and the observer must focus them,
wave fronts appear to diverge from a point far behind the surface
as creating an npright image that is magnified compared with a plane­
of the mirror the object
as is in front of it. Hence the ob­ mirror image. When the object is beyond the focal point, the reflect­

(b),
server views an upright, reversed image that is neither ed wave fronts converge in front of the mirror. An observer between

(e).
larger nor smaller than the real object. The situation is similar for a the mirror and the focal point must focus the wave fronts before they
convex mirror althongh the reflected wave fronts appear to di­ have converged, creating an upright, reversed and magnified image
verge from a point that is not as far behind the surface and the image An observer who is beyond the focal point intercepts the wave
is therefore smaller. For a concave mirror a number of cases must be fronts after they have converged, and the image is reversed and in­
considered. When the object is between the mirror and the focal point verted (J). The exact position of the observer determines whether the
(where incident rays parallel to the mirror axis are focused), the re- image is magnified, reduced or the same size as a plane-mirror image.

214
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
magnified along the concave axis. When CONVEX MIRROR �".11111.".
the observer and the object are suffi­
ciently far from the mirror (far enough
so that a simple concave mirror would
reflect an inverted, reversed image), the
result is quite different.
Wave fronts reflected from the sur­
face of the concave cylinder, instead of
being inverted in all lateral directions as CONVEX-CYLINDER MIRROR
they are after reflection from a concave
mirror, are inverted laterally only along
PLANE MIRROR
'\
the concave curvature. Because the re­
sulting image has a total of two inver­
sions (one lateral and one front-to­
back) it is not reversed, so that a left
hand is reflected as a left hand. Depend­
ing on the positions of the object and
the observer, the image can be magni­
fied, reduced or the same size along di­
rections parallel to the concave curva­
ture. Moreover, the lateral orientation
of the concave curvature determines the
lateral orientation of the mirror image; - --- - --- -- - t---+--+--+-_
hence by rotating the axis of the mirror
it is possible to give the image any orien­
tation. When the axis of concave curva­ - - - - - - -- - - - '---'---'--'-
ture is positioned horizontally, the non­
reversed image appears upside down.
The wave fronts emanating from a
point on an object are spherical until

at,
they reach the mirror surface of a con­
cave cylinder. After reflection from the
surface they become increasingly asym­

"-,,- "­
metrical until they finally come to a

, "­ "­
focus. Since the converging wave fronts CONVEX-CYLINDER MIRROR

,,
are lacking in symmetry, however, they

"-
cross over not through a single point
but through a line parallel to the con­
cave cylinder's planar curvature. An ob­
"-,,-
X "-
server farther away from the mirror

"­ ,
than this focal line is able to see a non­ '
CONVEX MIRROR ''-
reversed mirror image by focusing the ,

asymmetrical, inverted wave fronts, but

the resulting image is generally blurred "-
and distorted. The blurriness diminishes
as the observer moves away from the
mirror' because as the wave fronts move
past the focal line they approach a pla­
nar (and thus symmetrical) shape. It is
not only the concave cylinder that has
this property. Any curved mirror that
lacks rotational symmetry can focus in­
coming light from a point source to a
locus of points rather than to a single PLANE MIRROR

point. Hence the images reflected in mir­

rors such as the convex cylinder are also
affected by this blurring, which is relat­
ed to astigmatism in the eye.

T ofulclarify these matters I find it help­

to distinguish between two types
of mirror image. What I call a primary
image is created by a mirror alone and
"hangs" in space, appearing at the mir­
ror's focal point or (in the case of asym­

I and anda
metrical mirrors) at the locus of focal

plane-
points. What call a secondary image is
CONVEX CYLINDER acts somewbat like a plane mirror along one lateral axis like
created by the mirror in conjunction

is andcan
reduced
convex mirror along tbe otber lateral axis. By projecting from tbe corresponding
with an observer and can be said to be

the
convex-mirror images, it is possible to predict tbe appearance of tbe asymmetrieaUy trans­
positioned in the eye of the observer (on formed image reflected in tbe convex cylinder. Tbe image uprigbt, reversed In
surface of the retina or on a photo­ size along tbe axis of convex curvature. Tbe same kind of geometric construction be em­
graphic plate). The concept of primary ployed to analyze tbe image formed wben tbe convex cylinder is tipped to the right or the left.

215
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
CONCAVE CYLINDER acts like a concave mirror along one axis become inverting directions, so that the nonreversed image is upright.
and like a plane mirror along the other, so that when the observer When the axis of concave" curvature is vertical, as is shown at the
and the object are sufficiently distant, orientation is inverted along right, up and down become inverting directions, and so the noore­
just one axis and a nonreversed image is formed. When the axis of versed image is upside down. The positions of the object and the ob­
concave curvature is horizontal, as is shown at the left, right and left server determine whether the image appears magnified or reduced.

and secondary images serves to explain image act as if it were in two places at nonreversed images generated by a con­

A
how a mirror shaped like a concave cyl­ once? This puzzle of depth perception, cave cylinder are the result of a single
inder can combine the properties of a known as the parallax paradox, is re­ reflection of light. pair of plane mir­
plane mirror, which reflects virtual im­ solved by the idea that the image is nei­ rors joined at right angles can also cre­
ages, and a concave mirror, which (for ther in front of the mirror nor behind it ate nonreversed images, but they arise
sufficiently distant observers and ob­ but is in the eye of the beholder. from two separate reflections and thus
jects) reflects real images. How can an It is important to understand that the from two separate inversions of light.

SADDLE MIRROR acts like a convex mirror along one axis and the axis of concave curvature is horizontal (le/t), the mirror inverts

(right),
like a concave mirror along the other. When the observer and the ob­ 9:00 and 3:00 but not 6:00 and 12:00, and so an upright, nonreversed
ject are sufficiently distant, the saddle surface inverts orientation image is formed. When the mirror is turned 45 degrees clockwise
along directions that pass through the center of a hiD but not along the mirror inverts the directions 1:30 and 7:30 but not 4:30
directions that pass through the center of a valley. The orientation of and 10:30. As a result the image clock is rotated 90 degrees clock­
the image is determined by the orientation of the lateral axes. When wise (or twice the rotation of the mirror) and appears to be on its side.

216
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
Indeed, when an object is held up to
such a. configuration of mirrors, three
images are observed: an upright, re­
versed image appears in each of the
plane mirrors and a nonreversed image
resulting from a double reflection ap­
pears at the boundary between the two.
The concave cylinder is not the only
basic curved mirror in which nonre­
versed images result from a single re­
flection of light. Consider the saddle­
shaped mirror, which acts approximate­
ly like a convex mirror along one of its
principal axes of curvature and like a
concave mirror along the other. For an
observer or an object sufficiently close
to the surface, the saddle mirror, much
like the concave cylinder, generates up­
right, reversed images that are reduced
along the axis of convex curvature and
magnified along the axis of concave cur­
vature. The images created for distant
observers and objects differ according
to the orientation of the mirror.
The surface of a saddle mirror can be
divided into two hills and two valleys,
arranged so that the centers of the hills
lie on the axis of concave curvature and
the centers of the valleys lie on the axis
of convex curvature. Hence each lateral
direction that passes through the center
of a hill is inverting, whereas each later­
al direction that passes through the cen­
ter of a valley is noninverting. If the sad­
dle mirror is held so that its hills lie on a
horizontal line, 9:00 and 3:00 become
inverting directions, whereas 12:00 and
6:00 become noninverting ones. Since
12:00 and 6:00 are not inverted, the im­
age of an object is upright; since orienta­
tion is inverted along only one lateral
direction, the image is nonreversed. In
addition, compared with a plane-mirror
image, the nonreversed image is usually
red uced in all lateral directions. (It is For the photographer, astronomer, naturalist,
possible, however, for the image to ap­ educator, or casual observer Celestron has
- es­
pear magnified along the direction of
tablished a new standard in superb telescopes and
concave curvature.)
telephoto lenses at affordable prices (from $295).
Dotating the saddle mirror clockwise All feature large observatory mirror/lens type
R 45 degrees brings the centers of the optics folded into a compact, lightweight, portable
hills to 4:30 and 10:30, so that these di­ telescope or telephoto lens. Close-up detail borders
rections become inverting, whereas 1:30 on the fabulous from celestial objects light years
and 7:30 become noninverting. In this away, to ships, planes, flowers, wildlife or people.
case 12 appears at 3:00 and 3 at 12:00, 6 Here is truly an instrument that allows you to observe
appears at 9:00 and 9 at 6:00. The image or photograph the world around you. Celestron
is still nonreversed, but it has been rotat­ also has a full line of binoculars of astronomical
ed 90 degrees so that it appears to be
quality.
lying on its side. When the mirror is ro­

Send $2.00 for 32-page fullandcolor

tated another 45 degrees, so that the cen­
ters of the hills lie on a vertical line and
the centers of the valleys on a horizontal catalog on how to select use lens.
a Celestron telescope or telephoto
one, 12:00 and 6:00 are inverted but not
9:00 and 3:00. In other words, the nonre­
versed image is upside down. In general
if a saddle mirror is rotated by any given Celestron International
2835 Columbia St., Box 3578-SA
amount, the image of a fixed object is
Torrance, Calif. 90503, U.S.A.
rotated in the same direction by approx­
Telephone: (213) 328-9560
imately twice that amount.
The properties of the saddle mirror
can be made clearer by imagining how
such a mirror would reflect a transpar-
(Dealer Inquiries invited.)

217
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
ent clock face equipped with an hour motion of the image hand due to the basic types it is possible to create an
hand. As the real hand makes one 12- rotation of the mirror and the motion of endless variety of compound mirrors
hour journey around the face, the image the image hand due to the rotation of the with curious light-reflecting properties.
hand also makes one complete rotation object hand cancel each other. Indeed, For example, I have designed and pat­
about the nonreversed image face, but the image of the moving hour hand ented a series of mirrors, beginning with
(from the point of view of the observer) appears to be stationary, whereas the a saddle mirror, that can generate a
it moves in the opposite direction. If the image of the stationary clock face com­ nonreversed image or several such im­
mirror itself turns in the counterclock­ pletes one counterclockwise rotation ages of a single object. Although the im­
wise direction so that it completes one every 12 hours. ages are smaller than the corresponding
rotation every 24 hours. however, the By joining together mirrors of the six plane-mirror images, they exhibit strik-

MULTILOBE MIRROR, such as the three-lobe version shown, is a that another 12 appears at 12:00, another 8 at 8:00 and another 4 at

is
compound mirror generated by rotating a saddle, or two-lobe, sur­ 4:00. Two nonreversed images of each numeral are generated. The

2:00, 6:00 andis

face about a plane surface. Like the saddle, the muitilobe mirror gen­ result is a nonreversed 24-hour clock face. When the three-lobe mir­
erates nonreversed images and inverts orientation along directions ror rotated 60 degrees clockwise, as is shown at the upper right,

clock is divides
moved
that pass through the center of a hill but not along directions that 12:00, 4:00 and 8:00 become inverting directions and
pass through the center of a valley. When the mirror is positioned as 10:00 become noninverting ones. The 24-hour image clock face
is shown at the upper left, 2:00, 6:00 and 10:00 are inverting direc­ then rotated 90 degrees clockwise. When the ob e t clock
j c
tions, and so an 8 appears at 2:00, a 12 at 6:00 and a 4 at 10:00. On away from the mirror axis, the reflected. 24-hour face
the other hand, 12:00, 4:00 and 8:00 are noninverting directions, so into two nonreversed 12-hour clocks, as is shown at the lower left.

218
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 © 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 ing clarity. These compound mirrors. are
derived from the basic saddle mirror by
rotating (in a complex manner) a saddle
mirror about a planar region, possibly
one as small as a single point. In this way
a saddle surface with any number of
lobes, or hills, can be generated, and so I
call these reflecting surfaces multilobe
mirrors.
As I have already demonstrated for
the saddle mirror, or two-lobe mirror,
on a multilobe mirror lateral directions
from a center of curvature that pass
through the center of a hill are inverting,
whereas those that pass through the cen­
ter of a valley are noninverting. When a
three-lobe mirror is placed with its hills
at 2:00, 6:00 and 10:00, it inverts those
directions but not 12:00, 4:00 and 8:00,
which are the directions taken by the
valleys. Because of the inversions an 8
appears at 2:00, a 12 appears at 6:00 and
a 4 appears at 10:00. Because of the non­
inverting reflections, however, another
12 appears at 12:00, another 4 at 4:00
and another 8 at 8:00. Indeed, provided
the surface is continuous and smooth,
two images appear for each of the 12
numerals on the object clock face. Be­
cause each of the two sets of image nu­
merals appears in clockwise order, the
image of the 12-hour clock face is a 24-
hour clock face.
If the three-lobe mirror is rotated 60
degrees, it inverts 12:00, 4:00 arid 8:00
but not 2:00, 6:00 and 10:00. In this
case the image is still that of a 24-hour
clock, but it has one 12 at 3:00 and an­
other at 9:00, one 10 at 4:00 and anoth­
er at 10:00, and so on. In general if the
three-lobe mirror is rotated by a given
amount, the mirror image of a fixed ob­
ject is rotated in the same direction by
one-and-a-half times that amount.
An hour hand on a transparent clock
face reveals a more striking manifesta­
tion of the same phenomenon. As the
hour hand rotates about the transpar­
ent 12-hour clock face two image hour
hands are visible on the nonreversed 24-
hour image clock face. Each hand ro­
tates 180 degrees in 12 hours. At the end
of 12 hours, then, the two image hands
have exchanged positions, and only af­
ter the real hour hand has completed
two 12-hour rotations has each of the
image hands returned to its original po­
sition. It is also interesting to note that
when a transparent clock face is moved
to one side of the three-lobe mirror, so
that the center of the face no longer lies
on the mirror axis, the nonreversed 24-
hour image clock face divides into two
nonreversed 12-hour clock faces.
A saddle-shaped mirror with lobes
n
(where is equal to or greater than 2)
n
can create - 1 nonreversed images of
n
a single object. Thus the four-lobe mir­
ror can generate three nonreversed im­
FOUR-LOBE MIRROR, shown in top and oblique views, generates three nonreversed im-' ages, reflecting a 12-hour clock face as
ages of an object. Here the mirror is positioned so that it inverts 1:30, 4:30, 7:30 and 10:30 but a 36-hour clock face. In addition, when
not 3:00, 6:00, 9:00 lind 12:00. The result is a nonreversed image of a 36-hour clock face. a mirror with n lobes is rotated by a

220
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 Compare the relative mer­ ha n d li n g , J e t t a w i l l n e v e r

dElia its of t oday's family cars,

and you'll find that a trip to
embarrass you at the gas puml2:..
It g ets an EPA estimated I2.SJ
mpg, 40 mpg highway estimate.

41N THE FIiONT.���;f�J

(Use "estimated mpg"for compar­
isons. Your mileag e may var y
with weather, speed and trip

81N THE BACKa

Jetta. length. Actual highway mileage
In the front, will probably be less.)
for example, And it gives you front-wheel
• there's room drive, rack-and-pinion steering,
enough for family of 4 to ac- all-independent suspension and
tually stretch out and enjoy the classic European styling.
ride. Impressed with Jetta's relative
In back, there's a t r u nk big merits?
enough to handle 8 suitcases. Your relatives will be, too.

VOlRSWAOEN
(Something you don't even get
with a Rolls-Royce.)
And under the hood, there's a

DOESn
CIS fuel-injected engine powerful
enough to take you from 0 to 50
in just 9.2 seconds. As well as

AOAIN
around any trucks, onto any high­
w ays and up any mountains
you meet along the way.
Whats more, even though Road

~
& Track has said it "will embar­
rass a lot of cars costing a lot
more" with its performance and

©'960 IIOll(SWAGEN OF AMERICA

© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 CON" "·C' UNDR<.c R

� Z
_ _________
__

.tL-_
given amount, the image of a fixed ob­
ject is rotated in the same direction by
n / (n - 1) times that amount. Converse­
ly, if an object reflected in a stationary
� n-Iobe mirror is rotated by a given

II / II/
CONVEX REGION
amount, its image is rotated in the oppo­
site direction by l / (n - 1) times that

II + II
amount.

� -- + Saddle-shaped mirrors are only one of

the many kinds of compound mir­

�
CONVEX-
�
-
CYLINDRICAL rors that can be made to form more than

-l' -lI -- � one image of a single object. Indeed,

-- -
REGION

,, , _______ I
BOUNDARY
PA RALLELS
with the endless possibilities for con­

,-+---
necting basic mirrors, the compound
BOUNDARY
mirrors that reflect multiple images
PERPENDICULARS
would seem to defy classification. The

PLANAR
I , situation is not as hopeless as it appears,
REGION however. The key to the solution is to
consider not only the basic regions that
make up a compound mirror but also
SIMPLE BOUNDARY -
the boundaries that separate one region
from another. (It is assumed that the
"join" between two basic regions of a
:-- compound mirror is smooth, with no
discontinuities.) The boundaries of a
compound mirror fall into three main

\
categories, and if the category of a par­

\
ticular boundary can be identified, the
appearance of mirror images on each

(solid
side of it can be predicted.

black lines)
The best way to describe a boundary

(broken black lines)

BASIC REGIONS IN A COMPOUND MIRROR are joined along smooth boundaries
on a compound mirror is by considering

(broken colored lines)

that can be classified according to their effect on two sets of curves on the surface
of the mirror. The curves are the boundary parallels and the boundary per­
its relation to two sets of curves that lie
pendiculars that lie in the regions on each side of the boundary. The hox­
on the mirror surface: the curves that
like mirror with rounded edges shown here is made up of three kinds of region: planar, convex run parallel to the boundary and those
and convex cylindrical. In all cases, as one moves from one side of a boundary to the other, the that run perpendicularly across it. Like
inverting properties of both the boundary parallels and the boundary perpendiculars are un­ any curve on the surface of a mirror, a
changed. For example, at the top of the front face, where a plane region joins a convex cylindri­ boundary parallel or a boundary per­
cal region, the houndary parallels are planar and hence are noninverting on hoth sides of the pendicular can be described as nonin­

is
boundary; the boundary perpendiculars change from planar to convex, but both of those cur­ verting (if its curvature is either convex
vatures are noninverting. A boundary that does not change the inverting properties of either its
or planar) or inverting (if its curvature is
parallels or its perpendiculars called a simple boundary, and a compound mirror that includes
only simple boundaries reflects a single, connected image of a fixed object. The size of the
concave). A boundary between two ba­
parts of the image, however, may vary according to the regions in which they are reflected.
sic regions, then, can be categorized ac­
cording to the way the inverting proper­

�
ties of its parallels and perpendiculars
change as one travels from one side of
the boundary to the other.
For example, I call a boundary simple
CONV"' R 'G'ON OOUNDA", "R"ND 'CU CAR S
if, as one goes from one side to the other,
neither the parallels nor the perpendic­
ulars are transformed from inverting

" \\ \
curves to noninverting ones (or vice ver­

II
sa). Compound mirrors whose regions

I \ ., ___-- I
\ //..;-
are all joined by simple boundaries gen­

," , I
BOUNDARY PA RA LLELS
erate a single, connected image of an

I-
object. Parts of the image in various re­
// gions of the mirror may differ in size

I\
- - ------
depending on the curvature of those re­

..... '\
gions but the parts are attached to one
another just as they are attached in the

/ -- - - - - - - _ ..---- \\,
real object.

/I , '-
A boundary where the perpendiculars
change from inverting to noninverting
but the parallels do not change charac­
/ � SADDLE REGION
ter is what I call a bell boundary. As one
might expect, a mirror shaped like a

(broken color) (solid black)

bell, which consists of a convex region

(broken black).
joined to a saddle region, is an example
BELL BOUNl>ARY changes the inverting properties of boundary perpendic­
ulars but not of boundary parallels A boundary of this kind
' of a compound mirror with a boundary
joins a convex region to a saddle region in a bell mirror. Since the inverting properties of the of this kind. Because the inverting prop­
boundary perpendiculars are changed, a pair of images is created for each pair of regions. erties of the perpendiculars change at

222
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 © 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
the boundary, a pair of images is creat­ the mirror in the opposite direction will mirror orientation, the nonreversed im­
ed for each pair of regions. The images then "create" the pair of images appar­ age that appears in the saddle region
are enantiomorphic (they have opposite ently out of nothing. When a bell mirror completes two full rotations for each ro­
handedness), and they are arrayed sym­ is rotated about the forward-backward tation of the mirror. In this way the non­
metrically on opposite sides of the bell axis, however, the enantiomorphic im­ reversed and reversed images always
boundary. ages react quite differently. On the one present corresponding sides to the bell
hand, since reversed images are not af­ boundary as well as to each other.
fected by changes in mirror orientation,
W hen a bell mirror (or any com­
pound mirror that creates pairs of the reversed image reflected in the con­
An array of four or more regions all
joined by bell boundaries forms what I
enantiomorphic images) is turned about vex region retains the same orientation call a matrix mirror. The simplest mir­
a lateral axis parallel to the bell bounda­ through a complete rotation of the mir­ ror of this kind is the two-by-two matrix
ry, the images may annihilate each other ror. On the other hand, since the orienta­ consisting of two saddle regions joined
by merging and disappearing. Turning tion of nonreversed images depends on at a point and separated by a convex

ENANTIOMORPHIC IMAGES (having opposite handedness) ap­ about a lateral axis: the two images are first annihilated, merging and
pear symm etrically on opposite sides of bell boundary. Tbe bell mir­
a disappearing from !he surface, and then re-created. The lower se­

tbe
ror re8ects an uprigbt, reversed image in it. convex region and an in­ quence sbows the effects of turning tbe mirror clockwise: tbe re­
verted. nonreversed one in its saddle region. The upper sequence of versed image remains upright, whereas tbe nonreversed one makes
pbotographs shows the effect on the image when the mirror is turned two full rotation. in the coorse of single rotation of
a mirror.

224
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 How can you earn more
from your stocks than
they earn on their own ?

Merrill Lynch says,

consider your options.
Right. Options. Stock options. they wire every Merrill Lynch office
You can earn extra income on an Options Analysis Report featur­
stocks you own-on top of your divi­ ing opinions on the most overvalued
dends-by selling options on the and undervalued options of the day.
stock. The option buyer pays you a And, most important, our
premium that gives him the right, fur options specialists are instantly acces­
a given period of time, to buy the sible to your Merrill Lynch Account
stock at a predetermined price. Executive to answer your questions
You can use options as partial on options and help develop your

A breed apart
protection against a market decline, portfolio strategy.
or for leverage in buying additional Options may or may not be
shares at advantageous prices. suitable for you. But if you're willing
Obviously, options don't guarantee a Timing is crucial. You have to to explore new avenues of invest­
profit; but they can be a big help in know the market and constantly mon­ ment, you may find they give you just
a wide range of investment strategies. itor both the options and the underly­ the flexibility you've been looking
And that accounts for part of ing stocks. for. Find out more about the rewards
the reluctance of many investors to That
's something Merrill Lynch and risks from "The Merrill Lynch
investigate options: their very versa­ is well equipped to help you do. Guide to Writing Options" and the
tility implies complexity. But that's Merrill Lynch specialists work exclu­ prospectus of the Options Clearing
not always the case. sively on the options needs of our Corporation.
Actually, the hardest part of retail customers. They keep an eagle We'll send you both if you call
understanding options is the element eye on the market. They generate toll-free 800-2 2 3 -1910 (in New York
of time. Unlike stocks, options expire. actionable ideas and timely informa­ State, call 800-S2 2-560S) . Or visit
So you can't just put your money into tion. Each day the market is open, your local Merrill Lynch office.
them and leave it there to grow. And find out just what your
options are.

� MerriU Lynch
© 1980 Merrill Lynch Pi Fe & Smith Inc. Member SIPC.
erce nner

© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
region on one side and a concave region can meet. This would explain why no single object to meet at the same point.
on the other. By continuing this pattern boundaries seem to change the invert­ The saddle and planar regions of mul­
it is possible to construct a matrix mir­ ing properties of both boundary paral­ tilobe mirrors are also joined by bell
ror with any number of rows and col­ lels and boundary perpendiculars. boundaries that give rise to some inter­
umns. For example, a five-by-five ma­ Like bell mirrors, matrix mirrors can esting images. For example, although a
trix can be put together that reflects as be made to form enantiomorphic pairs three-lobe mirror normally reflects two
many as 25 images of an object. The of images. When a matrix mirror is tilt­ nonreversed images of an object, if an
four bell boundaries that separate the ed obliquely, it can be made to generate observer holding the mirror with one
regions in the two-by-two matrix mir­ double pairs, which can also be simulta­ hill at 12:00 begins to rotate the top of
ror come together at a single point, and neously annihilated or created. On no the mirror toward him, the two nonre­
it appears it is only at such a point that compound mirror, however, is it pos­ versed images move toward the mirror's
a convex region and a concave one sible for more than four images of a planar region. When the two images are
close enough to the planar region, one of
them divides into three new images, two
of them nonreversed and one reversed.
If the mirror is then turned again, the
reversed image may annihilate any one
of the three nonreversed images, leav­
ing again only two nonreversed images.

T pound
he third kind of boundary on com­
mirrors is one that changes
the inverting properties of the bound­
ary parallels but not of the perpendicu­
lars. I call this kind of boundary a torus
boundary because it separates the con­
vex outer region and the saddle-shaped
inner region that make up a torus­
shaped mirror. There are actually two
torus boundaries on a torus mirror, one
boundary joining the convex and saddle
regions around the bottom of the torus
and the other joining the two regions
around the top. For the purposes of this
discussion, however, it will suffice to
consider the torus from the top view
only, so that the two basic regions form
a pair of concentric rings, with the sad­
dle region on the inside, the convex re­
gion on the outside and a single torus
boundary separating the two.
Viewed axially, this section of a torus
mirror presents two images of a single
object: a reversed image on the outer
convex region and a nonreversed image
on the inner saddle region. Although
these images have opposite handedness,
they differ from the enantiomorphic
pairs reflected in bell or matrix mirrors
in that they are not placed symmetrical­
ly with respect to the boundary between
them. Instead the two torus images ap­
pear 180 degrees apart on their respec­
tive rings. For example, when the object
is a clock face, the torus mirror gen­
erates two concentric images, one re­
versed and running counterclockwise in
the convex region and one nonreversed
but also running counterclockwise in the
saddle region. Hence if the 12 on the
outer image appears at 12:00, the 12 on
the inner image appears at 6:00. In effect
the nonreversed image of the clock is
turned inside out; the nonreversed nu­
merals run counterclockwise and their
tops point inward. The connection be­
tween the reversed and the nonreversed
images on a torus mirror is unusual.
MATRIX MIRROR consists of four or more basic regions joined by bell boundaries. In this
For example, the reversed 12 appears
example, convex, concave and saddle-shaped regions are brought together in a repeated two­
by-two pattern. Because there are bell boundaries on at least two sides of each basic region, it is close to the nonreversed 6.
possible to generate as many as four images in each two-by-two cell of the matrix. By tilting If objects are moved over the torus
the mirror obliquely double pairs of enantiomorphic images can be annihilated and created. mirror, their images undergo peculiar

226
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 Search no further. Of all the lavish gifts on this page. we
t q
most recommend B and B. the classic French liqueur. Even when
price is no object. you c a n' give a l i u eu r that's distinctive more
or more impressive than B and B . It's priceless .
Julius Wile Sons & Co. New Hyde Park, N. Y. Imported from France. 80 proof.
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 TORUS BOUNDARY

(solid black)black)
BO U N DARY PERPENDICULARS

TORUS BOUNDARY
ties of boundary parallels
perpendiculars
(broken
(broken color).
changes the inverting proper­
but not those of boundary
On the torus-shaped mirror shown in
gions create pairs of images that have opposite handedness and are
positioned 180 degrees apart. In the photograph at the right a clock
face is reflected as two concentric image faces, one reversed and run­
the diagram at the left, the torus boundary separates the outer convex ning counterclockwise and the other nonreversed and also running
region of the mirror from the inner saddle-shaped region. These re- counterclockwise. The inner clock appears to be turned inside out.

motions near the torus boundary. In­ of compound mirrors it should be clear destroyed, reversed and nonreversed
stead of annihilating each other, like en­ that reversed and nonreversed images images always appear and disappear
antiomorphic images on a bell mirror or differ in more than handedness. For ex­ in equal numbers, even in the multilobe
a matrix mirror, the images flow togeth­ ample, all nonreversed images c im ro­ and torus mirrors. For a given com­
er to form an extremely distorted image tate with respect to fixed objects, where­ pound mirror the number of reversed
at the torus boundary. The reversed and as their reversed counterparts are con­ images minus the number of nonre­
nonreversed images of a single object strained to remain either upright or versed images is a constant "magic num­
combine to form a ring-shaped image; inverted. Moreover, although on bell ber," no matter how many images exist
the ring is composed of parts of four and matrix mirrors reversed and nonre­ altogether. For example, the five-by-five
images (two reversed and two nonre­ versed images are always affected equal­ matrix mirror can give rise to 13 re­
versed). The new images flow out of the ly by pair and double-pair creations and versed images and 12 nonreversed im­
original nonreversed image much as annihilations, on multilobe and torus ages; its magic number is 13 - 12, or 1.
they do in multilobe mirrors. mirrors a nonreversed image may divide If the mirror is tipped so that four rows
The peculiarities of the torus mirror into three new images (two nonreversed of five images each annihilate one an­
bring me to the point with which I shall and one reversed), whereas its reversed other, five single images remain; three
end this examination of curved-mirror mate remains intact. are reversed and two are nonreversed,
images. From the preceding discussion Although images can be created and so that the magic number still equals 1.

IMAGE REACTIONS on a toroidal mirror a r e demonstrated b y the gion. The exchange of .one image for the other on the toroidal mirror
changing forms of a reflected hand. When the hand is moved over is quite unlike the creation and annihilation of paired images on a bell­
the mirror, the nonreversed image in the saddle-shaped region under­ shaped mirror: on the torus the reversed and nonreversed images of a
goes a complex exchange with the reversed image in the convex re- single object are not adjacent but are 180 degrees from each other.

228
© 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms
 © 1980 SCIENTIFIC AMERICAN, INC

This content downloaded from

108.52.192.130 on Sat, 20 Jun 2020 03:12:47 UTC
All use subject to https://about.jstor.org/terms