X-RAY CRYSTALLOGRAPHY

The ne"\v knowledge of the atolnic structure of Inatter uncovered

over the past half-century by the X-ray-diffractjon technique

has led to a funclalnental revision of ideas in n1any sciences

by Sir Lawrence Bragg

F
ifty-six years ago a new branch of summed up in a "model." I have often atoms in a crystal of sodium chloride (or
science was born with the discov been asked: "Why are you always show dinary table salt) is 2.8 1 angstrom units
ery by Max von Laue of Germany ing and talking about models? Other (an angstrom is 10-10 meter), whereas
that a beam of X rays could be diffracted, kinds of scientists do not do this." The the most commonly used wavelength in
or scattered, in an orderly way by the answer is that what the investigator has X-ray analysis is l.54 angstroms.
orderly array of atoms in a crystal. At been seeking all along is simply a struc Actually crystals came into the pic
first the main interest in von Laue's dis tural plan, a map if you will, that shows ture only because- they are a convenient
covery was focused on its bearing on the all the atoms in their relative positions means to an end. The resultant scatter
controversy about the nature of X rays; in space. No other branch of science is so ing of X rays would be hopelessly con
it proved that they were waves and not completely geographical; a list of spatial fused and impossible to interpret if the
particles. It soon became clear to some coordinates is all that is needed to tell scattering units were randomly distrib
of us, however, that this effect opened the world what has been discovered. uted in all orientations. In a crystal the
up a new way of studying matter, that in The atomic structure of a crystal is de units are all similarly oriented and hence
fact man had been presented with a duced from the way it diffracts a beam scatter the X rays in the same way; as
new form of microscope, several thou of X rays in different directions. A crys a result a total scattering measurement
sand times more powerful than any light tal is built of countless small structural made with a whole crystal leads directly
microscope, that could in principle re units, each consisting of the same ar to a determination of the amount scat
solve the structure of matter right down rangement of atoms; the units are re tered by an individual unit.
to the atomic scale. The development of peated regularly like the pattern of a
X-ray crystallography since 1912 has wallpaper, except that in a crystal the The Condition for Djffraction
more than fulfilled our early expecta pattern extends in three dimensions in
tions. It not only has revealed the way space. The directions of the diffracted The easiest way to approach the opti
atoms are arranged in many diverse beams depend on the repeat distances of cal problem of X-ray diffraction is to con
forms of matter but also has cast a flood t11e pattern. The strengths of the diA:ract sider the X-ray waves as being reflected
of light on the nature of the forces be ed beams, on the other hand, depend on by sheets of atoms in the crystal. When
tween the atoms and on the large-scale the arrangement of atoms in each unit. a beam of monochromatic (uniform
properties of matter. In many cases this The wavelets scattered by the atoms in wavelength) X rays strikes a crystal, the
new knowledge has led to a fundamental terfere to give a strong resultant in some wavelets scattered by the atoms in each
revision of ideas in other branches of sci directions and a weak resultant in others. sheet combine to form a reflected wave.
ence. A culmination of sorts has been The goal of X-ray analysis is to find the If the path difference for waves reflected
reached in the past few years with the atomic arrangement that accounts for by successive sheets is a whole number
successful structural analysis of several the observed strengths of the many dif of wavelengths, the wave trains will com
of the basic molecules of living matter fracted beams. bine to produce a strong reflected beam.
the proteins-each of which consists of This brings us to the question of why In more formal geometric terms, if the
thousands of atoms held together by an X rays, of all the available forms of elec spacing between the reflecting planes is
incredibly intricate network of chemical tromagnetic radiation, are indispensable d and the glancing angle of the incident
bonds. for this method of investigation. In or X-ray beam is B, the path difference for
The purpose of this article is to go der for the interference of the diffracted waves reflected by successive planes is
back to the beginning and broadly sum beams to produce marked changes in the 2d sin B [see upper illustration on page
marize the course of X-ray crystallog amount of scattering in different direc 60). Hence the condition for diffraction
raphy over the past half-century or so. tions, the differences in the paths taken is n'\ 2d sin B, where n is an integer
=

In so doing I shall try to answer two key by reflected beams must be on the order and ,\ is the wavelength.
questions: Why X rays? Why crystals? of a wavelength. Only X rays have wave I first stated the diffraction condi
X-ray crystallography is a strange lengths short enough to satisfy this con tion in this form in my initial adventure
branch of science. The result of an in dition. For example, the distance be into research in a paper presented to
vestigation lasting many years can be tween neighboring sodium and chlorine the Cambridge Philosophical Society in

58
1968 SCIENTIFIC AMERICAN, INC
X-RAY PHOTOGRAPH of lysozyme, the second protein and first precession photograph; it is produced by manipulating the crystal,
enzyme to have its molecular structure determined by X-ray analy the phot ographic p late and an intervening screen in such a way as
sis, symbolizes the recent achievements of the X-ray technique_ The to hold one of the three indices of the diffracted beams constant
bright spots c o rrespond to various orders of diffracted waves p ro while recording the values of the other two indices in the form of a
duced by irradiating the lysozyme crystal with a beam of monochro rectilinear pattern. The lysozyme molecule contains 1,950 atoms and
matic X rays_ This particular type of X-ray photograph is called a measures approximately 40 angstrom units in its largest dimension.

59

1968 SCIENTIFIC AMERICAN, INC

1912, and it has come to be known as ent ways; three possible arrangements condition sets a limit on how many or
Bragg's law. It is, I have always felt, a of the sheets in a crystal of sodium chlo ders of diffracted waves can be obtained
cheaply earned honor, because the prin ride are indicated in the illustration on from a given crystal using an X-ray beam
ciple had been well known for some time the opposite page. The equation for re of a given wavelength.
in the optics of visible light. flection can be satisfied for any set of In the case of an optical diffractio
The atoms of a given crystal can be planes whose spacing is greater than half grating with an interlinear spacing a,
arranged in sheets in a number of differ- the wavelength of the X rays used; this the orders of the diffracted waves are de
fined by a single integer n in the equa
tion nA = a sin fJ; the diffracted waves
are referred to as first-order waves, sec
ond-order waves and so forth [see lower
illustration at left J. In the case of a crys
tal, on the other hand, the pattern re
peats in three dimensions, and so the
order of the diffracted waves must be
defined by three integers, which are rep
resented generally by the letters h, k
and I.
In the structural diagrams of sodium
chloride on the opposite page the axes of
the structure are denoted by the letters
OA, OB and OG, these being the inter
vals at which the pattern repeats. In the
diagram at the left the first reflection to
appear from the planes perpendicular to
OA will be one for which there is a path
BRAGG'S LAW, first formulated by the author in 1912, states the condition for diffraction
of an incident beam of monochromatic X rays by the successive sheets o f atoms in a crystal.
difference of the two wavelengths
In general terms the law states that i f the p ath differ ence for waves reflected by successive between 0 and A, since there are two
sheets of atoms i s a whole number o f wavelengths, the wave tr ains will combine to pr oduce sheets of atoms in this distance. vVith re
a strong r e flected beam. I n more formal geometric terms, if the spacing between the reflect spect to the spacing OA, then, this ini
ing planes of atoms is d an d the glancing angle of the incident X-ray beam is fJ, the path dif tial reflection is a second-order reflec
ference for waves reflected by successive planes is 2d sin fJ. In this diagram the extra path tion; with respect to the spacings OB
followed by the lower ray (heavy colored line at bottom) i s four wavelengths long, which i s and OG, however, the same reflection is
exactly equal t o the path differ ence of 2 d s i n fJ between the two diffracted r ays (upper right).
a zero-order reflection, since the reflect
ing planes are parallel to both OB and
OG. Therefore this type Qf reflection, or
diffraction, is assigned the order (200),
indicating h = 2, k = 0 and I = O. Sim
ilarly, the initial reflection to appear from
the planes in the diagram at the center
is (220), whereas for the diagram at the
right it is (Ill). Higher orders of reflec
tion would of course have higher integer
values of 11, k and I.

An Example of X-Ray Analysis

DIFFRACTION GRATIN

The structure of sodium chloride is a

simple arrangement of cubic symmetry
in which sodium and chlorine atoms oc
cur alternately in three directions at right
angles, like a chessboard in three dimen
sions. How was this structure derived?
The analysis will be gone into in some
detail here, because it is generally repre
sentative, even though this particular
case is so very simple. A glance at the
structural diagrams of sodium chloride
shows that the planes represented there
are of two kinds. The reflections, or or
DIFFRACTION ORDERS are illustrated here for the comparatively simple case of a ruled
optical diffraction gratin g_ In this case the diffracted waves are defined by a single
ders, designated (200), (400), (600) and
integer n i n the equation nA = a sin fJ, where A i s the wavelen gth of the incident radiation
so on, and those designated (220), (440),
and a i s the spacing between the lines of the gr ating_ I n the case of a crystal, on the other (660) and so on, arise from sheets of at
hand, the pattern repeats i n three dimensions, and so the or der o f the diffracted X-ray waves oms that are identical, each containing
must be defined by three integers, which are represented generally by the letters h , k and I. equal numbers of sodium and chlorine

60

1968 SCIENTIFIC AMERICAN, INC

 A o A

THREE POSSIBLE ARRANGEMENTS of the reflecting sheets of this case h = 2, k = 0 and I = 0.) Similarly, the initial r eflection to
atoms in a sodium chloride crystal ar e indicated by the colored appear fr om the p lanes in the diagram at center i s designated
planes. The axes o f this simple cubic crystal are denoted by the (220), whereas for the diagram at right it i s ( 1 1 1) with respect to
letters OA , OB and OC . In the diagram at left the fir st reflection to chlorine planes. I n general, order s with even indices arise fr om
appear fr om the planes p erpendicular to OA i s assigned the order sheets that are identical and hence result in strong, in.phase reflec
(200), since there is a p a th difference of two wavelengths between tions, whereas orders with odd indices arise fr om alternately oc
o and A while the reflecting planes are parallel to OB and OC . (In cupied sheets and hence result in weak, outofphase reflections.

atoms. One would expect the sequence flected from crystal faces with crystal there is a reduction in intensity owing
of successive orders to fall off regularly lographic indices (100), (1 10) and (Ill). to interference.
in intensity. As the diagram at the right As the curves show, the order (I ll) for It is important to note that F(hkl) is
on this page shows, however, the reflec sodium chloride is anomalously small, a dimensionless ratio, characteristic only
tion (Ill) comes from a more complex whereas (222) fits into the same sequence of the crystal structure. It is independent
set of planes, in that the sheets are alter with (200) and (220). For potassium of the wavelength of the X rays. If a
nately occupied by sodium and chlorine chloride, on the other hand, the scatter smaller wavelength is used, the orders
atoms. Since for (I ll) there is a path dif ing powers of the potassium atoms and appear at lower angles and path differ
ference of one wavelength for the strong the chlorine atoms are so nearly the same ences are reduced, but phase differences
ly reflecting chlorine planes, the waves that thc order (I ll) is too weak to be ob remain the same. Thus F(hkl) depends
reflected from the weaker sodium planes served. It was on the basis of such evi only on the distribution of scattering
halfway between them will be opposite dence that the sbuctural arrangement of matter in the unit cell, which it is the
in phase. The order (Ill) will be weak, both of these alkaline halides was con object of X-ray analysis to determine.
since the sodium conbibution partially firmed. The theoretical basis for measuring
offsets the chlorine contribution. On the Although the preceding analysis is values of F(hkl) was laid down by C. C.
other hand, for (222) the contributions somewhat simplified, it is a typical ex Darwin in two brilliant papers soon af
will be in phase and the order will be ample of the method used in the early ter the discovery of X-ray diffraction. In
sbong. determinations of crystal structure. A those early days the experimental obser
In this type of space lattice, as it is number of orders of diffracted waves vations were too approximate for a test of
called, there are identical pOints at the were measured, either with the ioniza his theory, and a number of years elapsed
face centers as well as at the comers of tion spectrometer or on a photographic before it could be applied.
the cube; this implies that the indices plate, and an attempt was then made Darwin's first calculation assumed the
must be either all odd or all even. These to find an atomic arrangement that ac crystal to be "ideally perfect." Rough
observations can be generalized by stat counted for the relative intensities of the tests showed, however, that the efficien
ing that orders with even indices, such various orders. cy of reflection was many times stronger
as (200), (220) and (222), should form a than his theory indicated. Darwin cor
strong sequence, whereas those with odd The Significance of F(hkl) rectly reasoned that the cause of the dis
indices, such as (Il l), (1 13) and (333), crepancy was the departure of the crystal
should be comparatively weak. The quantity F(hkl) is the cornerstone from perfection. It is a curious paradox
This is the effect that is actually ob of X-ray analYSiS, and its determination that imperfect crystals reflect more effi
served. The illustration on page 63 shows is the final aim of all experimental meth ciently than perfect crystals. In the lat
a very early set of measurements of so ods. This quantity is a measure, for each ter case the reflection, which is almost
dium chloride and potassium chloride order (hkl), of the intensity of the beam complete over a few seconds of arc,
made with the ionization spectrometer, a scattered by the whole unit of a pattern comes from a thin superficial layer only;
device invented by my father, W. H. expressed in terms of the amount scat planes at greater depths cannot contrib
Bragg, in 19 13. The abscissas measure tered by a single electron as a unit. For ute because the uppermost layers have
the glancing angle, the ordinates the instance, the quantity F(OOO) is scattered robbed the radiation, so to speak, of the
strength of the reflection. The two peaks in the forward direction through zero component the lower layers would other
seen on each order are the Ka and Kf3 angle, so that there are no path differ wise have reflected. Actual crystals, how
"lines" in the spectrum of the palla ences to cause interference; F(OOO) is ever, are in general far from perfect.
dium anticathode, the Ka line being the therefore the total number of electrons They are like a three-dimensional crazy
stronger of the two. The orders are re- in the unit of pattern. For higher orders quilt of small blocks that differ slightly

61

1968 SCIENTIFIC AMERICAN, INC

in orientation; as a result the crystal re school, culminating in the paper by mation has obviated the tedium of mak
flects over an appreciable angular range. Reginald James, Ivar Waller and Haltree ing numerous individual measm:ements,
Within this range rays penetrate into the on the zero-point energy of the rock-salt and the most advanced analyses are now
crystal until they encounter a block at lattice. Amplitudes of thermal vibration performed with counters as recorders..
the correct angle fm reflection, and the can be measured by their effects in re The original X-ray spectrometer "de
contributions from all such blocks add to ducing F values; by eXh'apolating to ab signed by W. H. Bragg is a typical ex
the total reflection. solute zero it was found that the atoms ample of the first method [see illustmtion
Darwin's second fmmula, therefore, still had a vibration corresponding to a below]. A collimated beam from the
applies to what is called an "ideally im half-quantum, as theoretical studies had X-ray tube fell on the face of the crystal
perfect" crystal, and it is the formula al indicated. and was reflected through slits into the
ways used. The intensity of the incident recording ionization chamber, which was
beam, or the amount of radiation per Experimental Measurements filled with a heavy gas (methyl bromide)
unit of time, is compared with the total to increase ionization. The outer case was
amount of radiation received by the re When a diffracted X-ray beam is re at a potential of several hundred volts,
corder as the crystal is swept through the corded by an ionization chamber, a Gei and the ionization was measured by driv
reflecting range at a constant angular ger counter or a proportional counter, ing the charge onto a coaxial wire con
rate; this enables all elements of the mo the orders are recorded one by one, by nected to a tilted gold-leaf electroscope.
saic to make their contribution to the re setting the crystal and the chamber at It was with this instrument that my fa
flection. suitable angles. Alternatively the beams ther made his pioneer investigations on
When calculating a value of F(hkl) can be recorded as spots on a photo the X-ray spectra from anticathodes of
for a postulated atomic arrangement, it graphic plate or film by turning the crys a number of different metals, a project
is necessary to know the contributions tal during the exposure so that a number" that formed the basis for H. G. J. Mose
from individual atoms, which depend on of planes can reflect. In the early crystal ley's subsequent work 011 atomic number;
the characteristic distribution of elec determinations the ionization spectrom the early determinations of crystal struc
trons in each atom. These distributions eter measured orders individually. As ture, fm which I was mainly responsible,
were calculated by Douglas R. Hartree more complex crystals were attempted were also made with this instrument.
in 1925 and are expressed as "F curves" and more orders had to be measured, Considering the crudeness of the ap
typical of each atom. Intensity measure the photographic method was favored paratus by modern standards, it gave sur
ments were the subject of an extensive because a single exposure registered a prisingly accurate results. A main trou
study by the University of Manchester large number of orders. Recently auto- ble arose from the vagaries of the X-ray

LEAD SHIELD
IONIZATION CHAMBER

ANTICATHODE

('\
" j
SLIT

u
X-RAY TUBE z

<{

CATHODE :
CRYS AL TABLE /
CRYSTAL

ELECTROSCOPE

T ELESCOPE

IONIZATION SPECTROMETER was the instrument used by the the X-ray spect... a from various metallic anti cathodes and later by
author' s father, W. H . Bragg, to conduct the first investigations of the author to make the early determinations of crystal structure.

62

1968 SCIENTIFIC AMERICAN, INC

tubes in those days. The tube was ener
gized by a Rumkorff coil, first with a
hammer switch and later with a mercury
switch, which gave a steadier discharge.
If the X rays from the tube got too hard,
one held a match under a fine palladium
tube attached to the main tube, which
allowed some gas to diffuse in; if they
got too soft, one sparked to a bunch of
mica sheets inside, which absorbed some
gas. The gold-leaf electroscope was also
a tricky instrument for accurate work. I
think that one of the main reasons why
X-ray analysis developed in my father's 111
laboratory at the University of Leeds,
even though the fundamental discovery
__ --A=._ ..
=
__ _
_ ====
had been made in Germany, was that
my father had so much experience in
making accurate ionization measure
ments with the primitive apparatus then
available.
When the diffracted beams are mea
sured one by one, the indexing presents
no difficulty because the crystal orienta
tion that produces each beam is known.
When many beams are recorded at the
same time on a photographic plate, how
ever, each of them must be identified.
A number of ingenious methods have
been devised for this purpose.
In general two types of X-ray photo
graph have been widely used. One type
is called a "rotation" photograph [see top o 5 10 15 20 25 30 35 40
illustmtion on next page]. In this method GLANCING ANGLE (DEGREES)
the X rays fall on a small crystal, which
is rotated around an axis that coincides EARLY MEASUREMENTS of the intensity of the reflected X rays from sodium chloride
with one of its principal crystal axes, and (top) and potassium chloride ( bottom) were made with the ionization spectrometer. The
orders were reflected from crystal faces with crystallographi c indices (00), (1l0) and (Ill).
the diffracted beams are recorded on a
The two peaks seen on each order are the KOI and Kf3 "lines" in the spectrum of the p a l
cylindrical film. The images of the dif ladium anticathode, the KOI line heing the stronger of the two. For sodium chloride the
fracted beams all lie on "layer lines"; for order OIl) i s anomalously small, because the weak sodium contribution partially offsets the
instance, if the crystal axis is along C, strong chlorine contributi on, whereas for p otassium chloride the order OIl) is too weak to
the layer lines correspond to 1 0, 1 I,
= = be observed, because the scattering p owers of the potassium atoms and the chlorine atoms
1= 2, and the spots have all values of are so nearly similar. This comparison confirmed the structures assi gned to the crystals.
hand k. If the spots are very numerous,
it may be too difficult to sort them out
and a Weissenberg camera is used. In The second general method of X-ray atoms are so similar. In addition potas
this technique one layer line is singled photography is the powder method, de sium chloride has a larger spacing than
out by a slit, and the film is translated veloped independently in 1 9 1 6 by Peter sodium chloride, hence the displacement
as the crystal turns. If the film has been J. W. Debye and Paul Scherrer in Switz of the arcs to smaller angles. The powder
translated horizontally, the displacement erland and by Albert W. Hull in the U.S. method has found its main use in the
of a spot along the horizontal axis tells [see bottom illllstmtion on next page]. study of alloys.
the angle of setting of the crystal when The powder method is used when the
it was recorded and so defines the other material is available only in microcrys Symmetry
indices [see illllstmtion on page 66]. talline form. The X rays fall on a mass
Another elegant device in this cate of tiny crystals in all orientations, and The pattern of every crystal has cer
gory is the precession camera. Crystal, the beams of each order (hkl) form a tain symmetry elements that form a
photographic plate and screen perform a cone. Arcs of the cones are intercepted three-dimensional scaffolding on which
sinuous dance in such a way that those by a film surrounding the specimen. In the atoms are arranged, and these ele
spots, for instance, with a definite value the powder photographs of sodium chlo ments can be uniquely determined by
of 1 and all values of hand k are recorded ride and potassium chloride on page 65 the X-ray-diffraction method. In the ear
as a rectilinear net [see illustmtion on one can see that in sodium chloride there ly days of X-ray analYSis, when the ionic
page 59]. This method is particularly is a weak series for odd (hkl) values, compounds that were being studied of
suitable for crystals with large unit cells whereas in potassium chloride these ten had a high symmetry, this fact was
and consequently numerous values of the disappear because the scattering powers of great assistance in arriving at a solu
indices. of the potassium atoms and the chlorine tion. The possible schemes of symmetry

63

1968 SCIENTIFIC AMERICAN, INC

 OC AXIS OC AXIS are limited by geometry, just as the pos
sible number of regular solid figures are
limited, although in the case of crystal
1= 2 symmetries the number, 230, i quite
1=2 large. Symmetry axes and symmetry
planes can be identified by noting reg
1=1 ular absences of diffracted-wave orders;
1=1
the presence or absence of symmetry
centers can be determined by a statistical
1=0
1=0
survey of intensities, as was first shown
by A. J. Wilson; crystals with symme
1= -1
try centers characteristically have many
1= -1 more weak reflections than crystals with
1= -2 no symmetry centers.
Finally, X rays can tell "which way
1= -2
around" a structure is. Optically active
molecules can have two forms, one of
which is the reflection of the other (the
dextro and levo forms of the chemist).
ROTATION PHOTOGRAPHS ARE MADE by aiming the X rays at a small crystal, which In general when the waves scattered by
i s r otated around a n axis that coincides with one o f its principal symmetry axes; the dif the atoms have phases as if coming from
fr acted beams are recorded on either a flat plate (left) or a cylindrical film (right). The atomic centers, these two forms give
images of the diffracted beams all lie on "layer lines"; in this case the crystal axis is along identical X-ray diffraction, that is, the
DC, the layer lines correspond to l = 0, l = 1, l = 2 a n d the spots have all values of hand k.
reflection from the right-hand side has
the same amplitude as that from the left,
although the phase is reversed. vVhen
the wavelength of the X rays is close to
an absorption edge of an atom, however,
there is an appreciable phase change.
The atom scatters as if at one location for
the one side and at another location for
the other side, so that the resultant am
plitudes are different. This enables dex
tro and levo to be distinguished; for
instance, in the classic case of a tetra
hedron with four different corners one
could tell for each orientation whether
one was looking at an apex or a base.
J. M. Biyvoet was the first to distinguish
between dextro and levo forms of the
tartrate ion. There was a 50 : 50 chance
that the traditional chemical conven
tion for representing dextro and levo
was correct; luckily it turned out to be
right!

- ---
1IU.l.-.l-iL
Inorganic Compounds
MICROCRYSTALLINE POWDER

The first crystals to be analyzed by

'NCIDENT B EAM I means of X rays were simple types. An
approximate measure of the complexity
--,--
of a crystal is the number of parameters
that must be determined in order to de

LEFT
I)) ).) I I ( (( ( (( 0 )) ) )1 I I ( (( (( 1 RIGHT
fine the positions of the atoms. In the
case of an atom at a symmetry center,
for instance, no parameters are needed;
180 90 o 90 180 it must be exactly at the center. If the
ANGLE (DEGREES) atom lies on an axis, its position along
the axis is fixed by one parameter; if on
POWDER PHOTOGRAPHS ARE MADE by aiming the X rays at a mass of tiny crystals
a reflection plane, by two; if in a position
i n all orientation s. The diffracted beams of each order (hkl) will then form a cone_ If r e
of no symmetry, by three.
cor d e d on a plate p erpendicular to the incident beam, each diffraction order will appear as
a ring surr ounding the central spot (top); the position s of the rings shown are typical o f a
The early determinations were limited
face-centered-cubic crystal lattice. It is usually more convenient to employ a cylindrical to one or two parameters; in fact, it was
photographic film whose axis is perpendicular to the incident r a diation (bottom). Arcs of doubted whether more complicated crys
the cones are intercepted at all angles up to nearly 180 degrees; the film is then unrolled. tals would ever be analyzed. The break-

64

1968 SCIENTIFIC AMERICAN, INC

POWDER PHOTOGRAPHS of the diffracted X rays from sodium chloride there is a weak series for odd ( hkl) values, whereas in
chloride (top' and potassium chloride ( bottom) confirm the ear potassium chlor ide th ese orders disappear. Potassium chloride has
lier fi n di n gs m a d e with the ion ization spectrometer: I n so dium a larger spacing; hence the ar cs ar e di splaced to smaller angles.

through into much more complex struc (Be;jAI2SiG01s), for instance, is of a high ing to the negative ion from all its posi
tures was made in the early 1920's by order, with sixfoid, threefold and two tive neighbors just balances its charge.
the Manchester school, where analysis fold axes, symmetry planes and symme This might seem at first sight a simple
was extended to cases of 10 or 20 param try centers. An atom cannot overlap it rule, but it is powerful in excluding im
eters, a great advance at that time. It self, so that it must either lie exactly on probable sbuctures. Its significance is
was made possible by quantitative mea one of these symmetry elements or be that when Pauling's rule is obeyed, the
surements and increasing experience in just off it. This restriction is so demand lines of force between positive and nega
the nature of inorganic compounds. ing that the structure of beryl could be tive ions stretch only over the very short
One of the first successes of X-ray immediately deduced once the symmetry distances between nearest neighbors, so
analysis was to show that these com was determined. The top illustration on that the energy of the electric field is at a
pounds are not built of molecules. They page 67 shows the only possible way of minimum and the structure is stable.
are ionic in character, with a regular al packing the atoms of the beryl formula The study of inorganic compounds
ternation of positive and negative ions into the network of symmetry elements. culminated in the determination of all
held together by electrical attraction. For The laws governing the structure of in thc common mineral forms, in particular
instance, in the sodium chloride structure organic compounds, established by Linus the silicates. These compounds obey
there are no sodium chloride groups but Pauling in 1929, afford another guide Pauling's rule rigorously, because they
rather a chessboard pattern of positive in seeking a solution. They also explain must be very stable in order to exist as
sodium ions and negative chlorine ions. why some compounds are stable, where minerals. The explanation of their com
It was difficult in the early days to rec as others that seem equally plausible position proved to be very interesting.
oncile the new view of ionic compounds from a chemical point of view do not Their nature is determined by the silicon
with classical chemical ideas, but once actually exist. Pauling's rule is based on to-oxygen ratio, ranging from Si04 in the
accepted the ionic view afforded a much the requirement that for stability the en basic rocks to SiO in quartz. Although
fuller understanding of the construction ergy of the compound must be as low the ratio varies widely, the silicon ion
of such compounds. as possible. Each small positive ion lies is always surrounded by four oxygen
In an ionic compound the ions pack inside a cluster of larger negative ions; atoms; the silicon-oxygen tetrahedrons
together as if they had characteristic for instance, very small positive ions such may, however, share no corners, one cor
sizes. Their dimensions are not complete as beryllium or boron are each surround ner, two corners, three corners or four
ly fixed, but they vary only over a small ed by three oxygen atoms; silicon is corners by having an oxygen atom in
range. On the whole the negative ions surrounded by four oxygens, magnesium common. In the Si03 silicates, for exam
are by far the largest, because their elec and iron by six oxygens and still larger ple, there are long strings of Si04 groups
trons are more loosely held. ions by eight or more oxygens. If we pic that run endlessly through the structure,
The packing of ions of characteristic ture electric fields in terms of lines of representing infinite linear negative ions
size is a very useful concept when postu force, suppose the number of lines repre bound laterally by the positive ions. The
lating various atomic arrangements, par senting the charge on the positive ion silicate groups occur as sheet ions in the
ticularly when combined with a knowl is divided equally between the negative micas and clays, and as three-dimension
edge of the symmetry elements. The ions coordinated around it. Pauling's rule al-network ions with metals in their
hexagonal symmetry of the crystal beryl states that the total number of lines com- iuterstices in the feldspars. This unex-

65

1968 SCIENTIFIC AMERICAN, INC

pected feature of minerals explains their alloy structure provided the foundation trons as standing waves, the system has a
composition in a simple and elegant way; on which a theory of alloy chemistry low energy if the electrons of shortest
it is one of the important new concep could be developed. When two metals wavelength are just too long to be re
tions introduced by X-ray analysis. unite in varying composition, they form flected by the most marked reflecting
a series of phases. These compounds are planes in the phase structure. In the sta
Alloys non-Daltonian, that is, they are not com ble phase the atoms take up arrange
posed of some simple ratio of elements; ments that create the strongly reflecting
After the nature of inorganic com on the contrary, each exists over a range planes required for low energy. To put
pounds had been clarified, the next of composition. William Hume-Rothery it broadly, an alloy is not a compound
achievement of X-ray analysis was to ex first pointed out that in different binary between one metal and another but
plain the nature of metallic alloys, which systems phases with very similar physical rather a compound between all the metal
had hitherto been so mysterious. The pi properties tend to have the same ratio of atoms on the one hand and all the free
oneer investigations into alloy structure free electrons to atoms in their composi electrons on the other. It is perhaps
were made by Arne F. Westgren in Swe tion. Structure determinations showed not too much to say that the X-ray deter
den. They were developed by Albert J. that such phases have a closely similar mination of alloy structures led for the
Bradley and his pupils at Manchester, for atomic arrangement but with curious first time to a rational theory of metal
the most part between 1925 and 1935 . characteristics. The essential similarity chemistry.
The powder method was used perforce, lies in the positions occupied by atoms, Equilibrium systems, or the phases
because the material was in microcrys not in the way the kinds of atoms are dis produced by variations of composi
talline form, and in Bradley's hands it tributed among the places. Apparently tion, had previously been deduced from
reached a peak of perfection that has the relation of atom to atom is relatively studies of polished and etched speci
hardly been equaled since. The ground immaterial; it is the position of the atoms mens, but they can be mapped out far
covered was so extensive that it is only that is all-important. This in turn was ex more directly by X-ray analysis. Phases
possible to give the briefest summary. plained by theoretical phYSicists in terms can be recognized by their powder
In the first place, the determination of of Brillouin zones. Treating the free elec- photographs and the composition of the
phase in any given case can, after some
preliminary trials, be fixed by noting
which spacings of the unit cell vary over
the range. Ternary and even quaternary
systems, which would hardly be amena
ble to metallographic methods, can be
tackled by X-ray methods.
An interesting application in this area
was found in the "order-disorder" sys
tems. An example of such a system is the
copper-gold alloy CUaAu, first studied by
Gudmund Borelius of Sweden. At high
temperatures all the points of the face
centered-cubic lattice of this alloy are
occupied at random by gold or copper
atoms; at low temperatures after slow
annealing the gold atoms segregate into
the cube corners, leaving the face centers
to the copper atoms. The progress of the
segregation can be followed by the ap
pearance of new lines, corresponding to
greater spacings, in the powder photo
graphs. The variation of segregation with
temperature presents interesting thermo
dynamic problems of second-order phase
change, and the X-ray work did much
to stimulate the study of corresponding
changes in other systems.
Another phenomenon studied inten
sively by Bradley was the splitting of a
phase into regions with slightly different
composition, which were nonetheless
united in having a continuous crystal lat
tice. Such segregation sets up intense in
ternal strains. It is characteristic of alloys
used as strong permanent magnets, be
WEISSENB ERG PHOTOGRAPH is a type of rotation photograph that is used when th e
cause the strains give the material a high
spots are too numerou s to sort out by the conventional method . In th is technique one layer magnetic retentivity.
line is sin gled out by a slit, and the film is translated as the crystal turns. The displacement In general, X-ray analysiS has provid
o f a spot along the translation axi s (in this case the horizontal axis ) tells the an gle of setting ed a powerful new tool for examining the
of the crysta l when it w a s recor ded and so defines th e other refl ection i n d i c e s o f the eryst " 1 . properties of alloy systems, an achieve-

66
1968 SCIENTIFIC AMERICAN, INC
O X Y G E N ----i

B E RY L L I U M --='.It-,

A L U M I N U M --+----i

V E RT I C A L - R OTAT I O N A X E S

o TWO-FOLD A X I S
S I L I C O N ----.zr=.=-i.

L T H R E E - FO L D A X I S

o S I X - FO L D A X I S

STRUCTURE OF BERYL ( right ) , a complex inorganic compound fraction. Since an atom must l i e exactly on o n e of the symmetry
with the formula Be H AI:!Si 6 0 1 8 , was deduced from a knowledge elements or be just off it, there could be only one way of packing
of the packi n g of the constituent ions ( le/t ) as soon as the basic the atoms of the beryl formula into the network o f symmetry ele
hexa gonal symmetry of the crystal was determined by X.ray dif ments. A key to the symmetry axes of the crystal is given at bottom.

FOURIER REPRESENTATION o f the e lectron.density distribu pable o f scattering X rays. The density distribution of the molecule
tion in a molecule of phthalocyanine ( le/t ) was used to construct as a whole is obtained by adding together the terms of a Fourier
the atomic model of the molecule (right ) . The Fourier "density series, a mathematical expression that can be used to represent the
map" i s arrived at by treating a molecular structure not as a cluster periodic variation o f sets of electron.density sheets in all direc
of individual atoms but as a continuous electron distribution ca tions. The Fourier method i s ideal for analyzing organic molecules.

67

1968 SCIENTIFIC AMERICAN, INC

ment that has great technical importance even when the structure was approxi tity that varies periodically. Since a
as well as scientific interest. mately known. The refining of the struc crystal is a periodic pattern in three
ture to get the best fit became extremely dimensions, the electron density can be
The Fourier Method laborious. represented by a three-dimensional Fou-:
The Fourier method is in a sense a rier series. Each element of the series is
Another method of X-ray analysis at complete reversal of this process. A struc a set of electron sheets, or strata, that
tacks the solution of the crystal structure ture is treated not as a cluster of atoms vary periodically in density, and if the
from a quite different angle. So far the but as a continuous electron distribution amplitudes and phases of these sheets
crystal has been regarded as a pattern of capable of scattering X rays. The investi (which crisscross in all directions) are
atoms, each of which scatters X rays as if gator seeks to map this continuous dis known, they can be added and the re
from its center with an efficiency deter tribution, and, if he is successful, he can sult is a p]ot of the dens;ty distribution.
mined by Hartree's F curves. The result then recognize the positions of the atoms My father first pointed out, in his Ba
ant of the waves scattered by these atoms by noting where the electron density kerian Lecture to the Royal Society of
is then compared with the observed am rises to peak values. There is no juggling London in 19 15, that each of these peri
plitude of reflection, the position of the with the positions of atoms one by one; odic components reflects one correspond
atoms being adjusted to give the best fit. the density map shows the best position ing order and only that order; moreover,
This method was successful as long as for all of them, however large their the amplitude of the reflected waves is
the number of atoms in the unit cell was number. proportional to the amplitude of the Fou
small. As increasingly complex crystals The density distributions are mapped rier component. Since changes in the
were studied, however, it became more by adding together the terms of a "Fou phase of the reflected waves can still
and more difficult to try adjustments rier series," a mathematical expression give the same X-ray effects, however,
of so many parameters simultaneously, that can be used to represent any quan- the only way of choosing the right phase

CARBON

N IT R O G E N

0 OXYG E N

0 PHOSPHORUS

e C O B A LT

STRUCTURE OF VITAMIN B.12, solved by D orothy Crowfoot ments o f what might be called the classical methods of Xray analy.
Hod gkin in 1955, represented one of the outstanding achieve. sis. The formula o f vitamin B -12 molecule is C63H S 4 N 1 4 0 14PCo.

68

1968 SCIENTIFIC AMERICAN, INC

 is to introduce some criterion of reality
leading to a picture with the right num
ber of atoms of the right kind in the unit
cell. The Fourier method therefore cen
ters around a "phase hunt." Once the
phases are known the structure is "in the
bag."

O rganic Molecules

From 1930 onward the Fourier meth

od was recognized as being ideal for
analyzing organic molecules, and many
were determined in this way. The first
organic structures, naphthalene and an
thracene, had been outlined by "V. H.
Bragg in 1922 at the same time that the
study of inorganic structures was pur
sued at Manchester, and his laboratory
at the Royal Institution concentrated on
the organic field.
Initially most of the studies dealt with
crystals that had symmetry centers. This
is a great simplification of the problem,
because the phases of the Fourier com
ponents with respect to such a center
must by symmetry be either 0 or 7T; in
other words, F(hk l) must be either plus
or minus. The advantage here is that cal
culations on the basis of a quite rough
ADDITION OF A HEAVY ATOM at a specific place among the molecules in a protein
approximation of the structure generally crystal provided the key that has led to the recent solution of several protein structures. In
make it clear which is the right sign, par this illustration of the method two precession photographs of the same crystal of lysozyme
ticularly in the case of the strong and are superposed with a slight relative displacement. The set of spots due to the native pro
'
therefore important orders. A Fourier tein i s printed in black. The set of spots due to the protein with the heavy atom i s p rinted
series can then be calculated, the posi- in color. Numerous changes in the intensities of corresponding spots can be detected
. tion of the atoms can be improved, fur
ther signs that were formerly doubtful
can be fixed and a new Fourier can used in a two-dimensional Fourier series. but also conclusions arrived at on chemi
be summed. This process of refinement The result is shown as a contour plot of cal grounds were actually misleading.
is rapidly convergent, and after a few the electron densities. The molecule, of formula C63Hs4NHOH
stages the structure is accurately deter In most cases the results of X-ray anal PCo, is shown in the illustration on the
mined. ysis confirmed the topography assigned opposite page.
A three-dimensional Fourier series is to the molecules by organic chemistry, The solution illustrates the curious
a formidable affair, and in these attacks but the X-ray findings determined the and unique character of X-ray analysis,
on organic structures the less ambitious bond distances and bond angles with which is reminiscent of the solution of a
task of using two-dimensional series and great accuracy and so cast much light on code, or of an ancient form of writing
getting a projection of the structure on a the nature of the bonds. such as Egyptian hieroglyphics or Mi
plane was more usually undertaken. Pro The next stage in the study of organic noan Linear B. It was first assumed
jections on the three principal planes de molecules was to tackle far more com that the phases of the F's were those
fine the positions of the atoms in space. plex structures, some of which had de of the cobalt atom (Co) at the center of
A pretty example of this was the analYSis feated the efforts of organic chemists to the molecule and a Fourier series was
of phthalocyanine with 60 parameters, elucidate their stereochemistry. A fore formed on this hypothesis. Although it
where the signs of the F's were found runner of this stage was the solution of turned out that this is far from true, the
without any guesswork as to the nature strychnine by Biyvoet in 1948, published phases are, as it were, weighted in this
of the structure [see bottom illustration independently at almost the same time direction because the cobalt atom is so
on page 67]. that Sir Robert Robinson's researchers at heavy compared with the other atoms.
It is possible to substitute a heavy the University of Oxford arrived at an Fourier series have a surprisingly oblig
atom at the center of the phthalocyanine identical structure by purely chemical ing way of trying to tell the investigator
molecule without any alteration in the reasoning. The outstanding examples are something with the most sketchy basis of
crystal lattice. If a spot becomes stronger the solution of penicillin and vitamin information, and in this case the series
when the heavy atom is introduced, its B-12; in each case Dorothy Crowfoot outlined in a shadowy way the molecular
original F value with respect to the cen Hodgkin of Oxford was the leader of the structure immediately surrounding the
ter of the molecule must have been posi research. The latter investigation was a cobalt atom. The information was used to
tive; if the spot becomes weaker, the F saga of X-ray analysis that took eight adjust the phases, and a further series
value must have been negative. All the years. Not only was much chemical in was formed and so the structure gradu
signs were determined in this way and formation about the molecule lacking ally began to emerge from the cobalt

69

1968 SCIENTIFIC AMERICAN, INC

 atom outward. The calculations would caused by the addition of a heavy atom
have been impossibly onerous had it not can be seen when two precession photo
been for the availability of electronic graphs made with the same crystal are
computing of structure factors and three superposed with a slight relative pis
dimensional Fourier series. The solution placement [see illustmtion on preceding
of vitamin B-12 represented the highest page J . The black set of spots is due to
flight of what might be called the classi the native protein, the colored set to
cal methods of X-ray analysis, and the the protein with a heavy atom, and a
Nobel prize awarded to Mrs. Hodgkin in close examination will show numerous
1964 was a well-deserved acknowledg changes in the intensities of correspond
ment of her achievement. ing spots.
It is necessary to find some three or
Biochemical Molecules four heavy atoms that can be attached to
definite sites on the molecule. Although
We now come to a most dramatic turn the structure is initially unknown, direct
ing point in the history of X-ray analysis. methods are available for finding the
When vitamin B-12 was analyzed, with relative positions of these "staining"
1 8 1 atoms in the molecule, it seemed atoms. The phase difference between the
hard to imagine that much more complex F value H due to the heavy atom and the
structures could ever be tackled; it had F value P due to the protein can be
taken eight years to complete and the found by comparing F(P) with F(P + H),
difficulties increase as a high power of because F(P + H) must be the vectorial
the number of atoms. And then, as the resultant of F(P) and F(H); the knowl
result of an investigation that had lasted edge of the phase difference for several
for some 20 years, a way was finally heavy atoms pins down the position of
19 out of 2 found to solve the structure of the im the Fourier component.
hi-fi experts do. mensely more complicated molecules of Because the solution is direct it can be
They appreciate ,
living matter, the proteins. The first of found by giving instructions to a com
as you wilL
Dual's smooth,
these to be analyzed, by John C. Ken puter. The computer is essential because
quiet operation drew in 1955, was myoglobin, which has the complexity is so great. Some 100,000
and many precision 2,500 atoms in its molecule. or 200,000 intensities must be measured
features. One is By a curious paradox, the very size of accurately by means of an automatic
the flaw less lI2
such a molecule has opened up a new machine that sets the crystal and the re
gram tracking of
its friction-free
line of attack that is not possible in the corder at the right positions for one or
tonearm. Others case of simpler types, leading to a direct der after another and lists the results. A
are constant-speed determination of phases without any ele corresponding number of equations must
motors, accurate ment of guesswork or trial and error. The be solved to find the phases, and the
adjustments for
principle is the same as in the case of Fourier series of many thousands of
stylus force and
a nti-skating , auto phthalocyanine described above, where telms must be formed. This long series
matic cueing and the substitution of a heavy atom in the must be summed at perhaps a quarter of
variable pitch molecule enabled the signs of the F val a million places in the unit cell to give
control.
ues to be found. A discovery made by the density at each point. The informa
The Dual 10 19,
$ 1 29.50.
M. F. Perutz in 1953 made it possible to tion is then automatically turned into
Other models generalize this method for the proteins. contours, which are plotted on stacked
from $69.50. He found that heavy atoms such as mer transparent sheets, and the investigator
For full -
information and
cury and gold, or complexes containing has then to translate the density distribu
lab test reports. such atoms, can be incorporated at spe tion into atomic arrangement.
write United Audio. cific places in the framework of the The final result is impressive. Some
535 Ma dison A protein crystal without affecting the ar half-dozen protein structures have so far
New York.
rangement of the molecules, which are been analyzed and they are already be
N.Y.10022
so large and loosely packed that the ginning to yield valuable information on
added groups find places in the inter such vital biochemical processes as the
stices. Further, Perutz showed that the operation of enzymes. The second pro
added heavy atom produces changes in tein molecule to be analyzed successful
the F values large enough to be accurate ly (after myoglobin) was the enzyme ly
ly measurable. It might at first sight ap sozyme (by David C. Phillips). The most
pear strange that the addition of one recent success has been hemoglobin (by
heavy atom of mercury to a protein mole Perutz); the model of this protein con
cule with 2,500 atoms of carbon, oxygen tains 10,000 atoms. I confess that when I
and hydrogen should make an appre contemplate one of these models, I can
ciable difference; it does so because the still hardly believe that it has been pos
scattering comes from one center, where sible to work out all its details by the
as the random contribution from n atoms optical principles of X-ray analysis,
ranging over all phases is proporti9nal which half a century ago claimed sodium
to yn, not to n. The changes in intensity chloride as its first success.

70

1968 SCIENTIFIC AMERICAN, INC

 C h efs re ly o n U O P i d eas .

s h o u l d m u n ch e rs 1
M a n y food processors a re u t i l i z i n g f l a v o r fo r
m u l a t i o n s f r o m U O P t o e n h a n c e t h e s u cc u
l e n ce, ri c h n ess a n d p i q u a n c y o f t h e i r p r e p a ra
t i o n s . So d o g o u r m e t resta u r a n t c h efs. Tod a y ,
t h e U O P F o o d P rod u cts D i v i s i o n a s s i s t s f o o d
t ec h n o l o g i sts w i t h c o n c e n trated s a u c e b a ses,
meat t e n d e r i z e rs, s e a so n i n g s , s p i ces a n d s n a c k
flavori ngs. F o r yea rs, UOP h a s e x t e n d e d i ts
t ec h n i c a l c o m pete n c e beyo n d i t s n o t a b l e c o n
t r i b u t i o n s t o p e t ro l e u m refi n i n g . F o r o n e, U O P
p i o n eered t h e eve ryd a y u s e o f B H A a s a food
a d d i t i ve t o i n h i b i t ra n c i d i ty . N o w, the UOP
C h e m i c a l D i v i s i o n s u p p l i es a d o z e n S u sta n e B H A form u l a t i o n s f o r
t h e food i n d u stry. Every b o d y c r u n c h e s c r i s p i e r potato c h i ps bec a u se
B H A is a d d e d to t h e c o o k i n g o i l , a n d oft e n to t h e p a c k a g e as w e l l .

U n i versa l O i l P ro d u cts C o m p a n y / D e s P l a i n es, I l l i n o i s 6 00 1 6 , U . S . A .

Petroleum Processes Transportation Equipment Metals & A lloys fragrances Air & Water Management Plant Construction Forest Products Chemicals & Plastics

1968 SCIENTIFIC AMERICAN, INC

 The crew of the Aluminaut had to go 4,000 feet down
to photograph this exploded torpedo.

The Alu m i n a u t ( the w o r l d 's

first aluminum subm arine) was
de s i g ne d t o dive a s deep a s 3
miles on its exploratory missions.
To document its findings, it
carries many specialized cameras :
8mm and 1 6mm motion picture
cameras, 35mm still reflexes
(for color transparencies), an
externally attached camera
(for close-inspection shots).
And the non-automatic
Polaroid Model 180 Land
camera (for first shots).
The first re a s o n for
taking the first shot with
our camera, is to guar
antee that the Alumi
naut will bring home
at le a s t one perfect
photograph of each subject.
Undersea photography of
a torpedo, or of anything else, is
tricky. The water's turbidity af
fects light reflection and refraction.
And even the most sophisticated
exposure meter readings (or elec
tric eye shutter settings) can pro
duce inaccurate exposures.
So i t ' s c o m f o r t i n g to have a
Polaroid
camera on board which makes a Corporation
finished print in just 15 seconds.
The Aluminaut photographer
takes his light reading, sets our
camera manually and, in 15 sec All he needs to do is trans
onds, he knows if he has to com pose the Model 1 80 settings
pensate for undersea conditions to his other cameras, and he
in order to get a perfect print. has the satisfaction of being sure
Once he has that print, he usu of all of his settings.
ally turns to his specialized cam And we have the satisfaction of
eras. Which brings us to the knowing that, down in the deeps,
second reason for h i s taking the a Polaroid camera is helping some
first shot with a Polaroid camera : other cameras take good pictures.

1968 SCIENTIFIC AMERICAN, INC

They took their first picture with a Polaroid camera.
For two reasons.

1968 SCIENTIFIC AMERICAN, INC

 How Renault scrimp s
on gas without

scrimping on car.

The Renault 10 can scrimp at minimum. No vibration, no pow Without those pleats, the same
the rate of 35 miles a gallon. er waste, no gas waste. paper would never even support
One of the chief reasons for another piece of paper. That, in
this remarkable mileage is the principle, is the kind of extra
Renault 1O's engine. Quite sim wrinkle we design into the steel
ply, if you don't waste power, you structure of the Renault 10. It
don't waste gas. Which isn't all allows us to increase strength
that simple.
In practice, it meant develop
ing a new casting technique to
get a more compact engine block.
By keeping the block size down, and rigidity without increasing
we keep the car size down, which sheer dead weight. And by not
keeps the gas consumption down. increasing dead weight, again, we
I t meant anchoring the crank don't increase gas consumption.
shaft with 5 main bearings (as To explain the next bit of gas Now, with your permission,
much as most V-8's) to cut vi saving we've set up a little dem a little documentation on how
brations down to an irreducible onstration . (above) little scrimping there is on car.
Included in the $1,745* price :
Seats that have been compared
with the seats in a Rolls -Royce .
The RenaultlO And which convert into twin
P . D . E . EAST COAST.
AUTOMATIC TRANSM I S S I O N beds . 11 cubic feet of trunk . Top
O P T I O N A L . F O R I N FO R M AT I O N
W R I T E R E N A U LT,I N C . ,
B O X 1 2 , 750 3 R D AV E , NYC
speed, 85. Self- adj usting 4-wheel
disc brakes. 4-wheel independ
ent suspension. 4 doors . 4 - speed
synchromesh transmission.
If there' s any scrimping on
the Renault 10 at all, it's on the
gas tank. It only holds ten gal
lons . (About 350 miles worth. )

R ENAU LTt

1968 SCIENTIFIC AMERICAN, INC