Outline of a History of Differential Geometry: I

Author(s): D. J. Struik
Source: Isis , Apr., 1933, Vol. 19, No. 1 (Apr., 1933), pp. 92-120
Published by: The University of Chicago Press on behalf of The History of Science
Society

Stable URL: https://www.jstor.org/stable/225188

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/225188?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

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

The History of Science Society and The University of Chicago Press are collaborating with
JSTOR to digitize, preserve and extend access to Isis

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 Outline of a history of differential
geometry (*)

I The Time before Leibniz.

It is difficult to talk of differential geometry before LEIBNIZ.

There were many applications of infinitesimals to geometry before,
but almost without exception they were quadratures and rectific-
ations of curves, cubatures and quadratures of surfaces or solids,
and studies of special curves, subjects we now exclude from
differential geometry proper, except as occasional illustrations.
Only a few topics have an immediate bearing of our subject.
Among them we have to mention the investigation of the nature
of tangency found in EUCLID'S " Elements " (last part of
4th century, B.C.), discussed only in the case of a circle. EUCLID
explains, in Book III, as a " definition " that a straight line is
tangent to a circle if it meets the circle and does not intersect
it after being continued. Also in Proposition i6 of the same
book a property of tangency is explained. The tangent to a circle,
it is said here, will fall outside the circle and no other line will fall
in the space between this straight line and the circumference.
This property of the tangent was taken up again later and general-
ized by LAGRANGE, when he developed his theory of contact
(Theorie des fonctions analytiques, Seconde partie I, 5). EUCLID
himself tries to describe the nature of contact more in detail
in the same proposition, in which he states that " the angle of the
semicircle is larger than any rectilinear acute angle, the remain-
ing angle smaller." This suggestion of extending the notion of

(*) This outline was given in a series of ten' lectures at the Massachusetts Instit
of Technology during fall and winter of 1931-32.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 93

angle to so-called horned angles, though having been the sub-

ject of long discussions, has not been carried far in differential
geometry. (i)
Similar ideas can be found in the works of ARCHIMEDES and
APOLLONIUS. APOLLONIUS (3d century B.C.) makes full use of
the normal to a plane curve in his theory of conic sections. He
goes further; he finds that the normals to a conic section have
an envelope, and he determines this envelope in the three cases
of an ellipse, parabola and hyperbola. (2) He comes so closely
to the conception of curvature that KEPLER, in a book on optics,
could talk of the circle of curvature at a point of a parabola, as
if it were well-known to all his readers. (3)
In the same treatise on conic sections (in book II) we find
the asymptotes to a hyperbola. The technical name for these
lines is also due to APOLLONIUS (Jcrvju7TTc-6-).
ARCHIMEDES (287-212 B.C.) occasionally discusses subject matter
relevant to differential geometry, as in the beginning of his books
on the sphere and the cylinder, where he defines the straight
line as the shortest distance between two points in the plane.
He states in this work the definition of curves " concave in the
same direction " and arrives at the statement that if two plane
curve segments with the same endpoints are concave in the same
direction, the curve lying between the straight line connecting
the two points and the other curve is shorter than this other
curve. He establishes similar theorems for surfaces. This paper
includes, for instance, the theorem that when a convex plane
curve lies inside another convex plane curve, its circumference
is the shorter. The curves also may be partly or entirely com-
posed of line segments, or may partly coincide. (4)
The problem of isoperimetrical figures belongs also to antiquity,
and is now included in that part of differential geometry which
utilizes calculus of variations. POLYBIUS, historiograph of the Punic

(I) Comp. F. KLEIN, Elementarmathematik vom hoheren Standpunkt aus.

(Berlin, SPRINGER) 1925, P. 222 sequ. See E. KASNER, Bulletin Amet. Math.
,Soc. (2) 17 (1910-I I) P 393.
(2) APOLLONIUS OF PERGA, ed. TH. HEATH (Cambridge, I892), P. I60-179.
(3) J. KEPLER, " Paralipomena in Vitellionem, " III,Theor. XIX, Werke II,
P. 175. (Ostwald's Klass. 198, P. 54).
(4) See e. g. ARCHIMEDES, French translation of VER EECKE, P. 4-6.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 94 D. J. STRUIK

wars, remarks that most people measure the size of towns or camps
by their perimeter, and that they can hardly believe that Sparta,
with a circumference of 48 stadia, is twice as large as Megalopolis
with a circumference of 50 stadia. (5) The mathematics of this
problem seems to go back to the time before ARISTOTLE. A series
of theorems is found in a paper by ZENODORUS (about I50 B.C.),
who stated that the circle is larger than all plane figures of the
same circumference, and the sphere larger than all solid figures
of the same area. Exact proofs, as a matter of fact, date back
only to the igth century.
Then there is the problem of mapping the earth on a plane,
a problem which offered itself to those geographers of Antiquity
who recognized the earth as a sphere. The principal contribution
is due to PTOLEMY (I50 A.D.), though we may readily believe that
his ideas were those of HIPPARCH, who lived three centuries earlier.
In PTOLEMY'S Geography, Chapter 24, we find what we now call
the stereographic projection. He takes the equator as plane of
mapping, and he not only explains the projection, but also shows
its conformal character. He also modifies the projection by
mapping the figure on a cone tangent to the sphere. This allows
a good representation of that part of the earth known to PTOLEMY.
The " map of the world according to PTOLEMY," reproduced in
many textbooks, is drawn in this projection. (6)
We cannot deal here with the reasons for the slow progress
of differential geometry in Antiquity, as it is only one aspect of
the much more general problem why antiquity did not advance
beyond ARCHIMEDES in the calculus of infinitesimals. We remark
that it has to do only, in a general way, with the fact that the
economic system on which the Mediterranean culture was based
already began to decay in the last centuries of the Roman republic.
Many new methods of mapping a sphere on a plane were
invented in the sixteenth century, when the great discoveries gra-

(S) POLYBIUS lived from 201-lI9 B.C. The place is from his Books IX, 21.
Our information on isoperimetry is taken from W. SCHMIDT, " Geschichte der
Isoperimetrie im Altertume," Bibliotheca mathematica (3)2 (1901), P. 5-8. See
also M. CANTOR, Vorlesungen I, 3( ed., p. 357.
(6) Map projections also in PTOLEMY'S " Planisphaerum " and " Analemma."
The name "stereographic projection " is due to F. D'AIGUILLON (1566-I6I7),
a Belgian Jesuit, who has Monge, central, and stereographic projection in his
" Optics " (16I3). (He calls the first two orthographic and scenographic).

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 95

dually widened the knowledge of the terrestrial sphere. In I540,

GEMMA FRISIUS, professor at Louvain, again used the stereogra-
phic projection. Of greater importance is the work of GERHARD
KRAMER, Latin MERCATOR (1512-1594), a Flemish cartographer
who lived a good part of his life at Duisburg. He used many
map projections, of which one carries his name, because he used
it for the first time in the famous map of the world of I569. This
method, the only one invented by the great cartographer, projects
meridians and parallels into straight lines. MERCATOR knew the
properties of his map very well, for instance its conformity,
and the meaning of the straight lines. He discriminates between
" plaga " and " directio," the " plaga " being the shortest con-
nection between two points on the earth, the " directio " the
shortest distance on the map. This " directio," in the words
of MERCATOR, is not straight, but " oblique curvatur." For larg
distance and high latitude there is considerable difference between
" plaga " and " directio". (7)
There was a considerable literature on MERCATOR projection
in the next decades, and connected with it we find a discussion
of the " directio." NUNES (as early as 1544, in print I573),
STEVIN, SNELLIUS, WALLIS, LEIBNIZ contributed. The name
loxodrome " is due to SNELLIUS' Typhys Batavus (I624).
Earlier than FRIsIUs and MERCATOR, J. WERNER suggested,
after JOH. STABER, a projection (I5I4), which conserves areas
It was used in I53I by 0. FINAEUS for a map of the world, and in
I538 by MERCATOR. This map, with its curious heart-like shape,
is seldom used. (8)
The many new investigations on curves and on infinitesimals
connected with the names of KEPLER, DESCARTES, FERMAT, CAVA-
LIERI and others are mostly of too special, or too general a nature to
find discussion here. A point of inflexion was first discussed by
DE SLUSE (I668) and FERMAT (I679). (9) But to the early history

(7) See H. v. AVERDUNK, GERHARD MERCATOR (1914), P. iz8 sequ. The

properties of the MERCATOR projection in the Legenda to the map of I569. In
v. AVERDUNK also discussion of the other literature.
(8) Annotationes JOANNIS VERNERIS, Nuremberg I5I4; 0. FINAEUS, De linea-
rum, superficierum et corporum dimensionibus (I 53 I). See H. v. AVERDUNK,
1. c. (7)
(g) See M. CANTOR, Vorlesungen 11 (i892), p. 840; III, P. I94, see however
G. ENESTRO)M, Bibliotheca mathematica 12 (I9I2-13), P. 156; 13 (0913-14), P. i68.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 96 D. J. STRUIK

of differential geometry belongs certainly the Horologium oscilla-

torium Of-CHRISTIAEN HUYGENS (I629-I695), his book on pendulum
clocks (I673) (io). The problem of measuring time in an exact
way suggested here a new mathematical theory. One of the
chapters of the book gives a complete theory of evolutes and
involutes in the plane. HUYGENS wanted a pendulum so con-
structed that the period of vibration would be independent of
the altitude. This is the problem of the tautochrone. The
solution is that the mass of the pendulum moves, not on a circle,
but on a cycloid. But the evolute of the cycloid is another cycloid.
We can therefore get a tautochronic pendulum by forcing the
thread of the pendulum to move along the circumrference of two
small parts of a cycloid with cusp at the point of suspension and
cusp tangent in the direction of equilibrium. To find this form
of the " cheeks" HUYGENS develops the general theory of evolutes
and involutes (" evolutae " and " evolventes", as he calls them),
in the plane and he gets an expression, in geometrical form, for
the radius of curvature. Here we also find the theorems that the
involute intersects orthogonally the tangents of the evolute and the
relation between arc-length of involute and length of the tangent
to the evolute.

2. - The First Systematic Contributions

When LEIBNIZ started his work, analytical geometry of the

plane was well under way, as was the application of infinitesi-
mals to quadratures. His main contributions to differential geo-
metry can be found in papers of I684, i686 and I692.
In the Nova methodus pro maximis and minimis, the first paper
in which LEIBNIZ published his new method (Acta Eruditorum
I684), we already find the interpretation of the equation d2y = o.
It indicates a point of inflexion, a conception, as we saw, introduced
by DESLUSE and FERMAT. In a paper of i686 (ii), we find the
circle of osculation, but not the expression for its radius in analytical
form. LEIBNIZ thought in his paper that the circle of osculation
passes through four consecutive points of the curve, because

(iO) German translation in Ostwald's Klassiker, I92.

(ii) LEIBNIZ, Meditatio nova de natura anguli contactus et osculi. Math.
Schriften, ed. GERHARDT II 3, P. 326-329.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 97

it has two contacts with it. After a remark by JACOB BERNOULLI

(I692, Acta Eruditorum) he readily recognised that only three
consecutive points come into consideration. The word " oscula-
tion " is taken from this paper of i686.
In I692, LEIBNIZ published (I2) his theory of envelopes of a
family of plane curves f (x, y, a) - O. It is necessary for this

to eliminate between f o O and , o. This early result is

the more remarkable, as only in recent times some essential
advance is made on this statement. In another paper of I692 (13)
we find a discussion of evolutes and involutes, mainly a statemnent
of HUYGENS' results, with the additional remark that the different
involutes are " parallel," the first place where this word is used
for plane curves.
To what extent differential calculus was applied to geometry in
those early days of the new method can be estimated by the recent-
ly published lectures of JOHANN BERNOULLI at Basle in the winter
of I69I-92. (I4) There we find computation of tangents to plane
curves, with cycloid, cissoid, quadratrix, as examples. Maxima
and minima are found by taking dy - o. The condition ddy o
leads to points of inflexion, as shown for the case of conchoid
and versiera. Even polar coordinates in the plane are introduced.
In the Integral calculus, written at the same time, the radius of
curvature appears (I4). L'HOSPITAL wrote his Analyse des infiniments
petits (I696), the first published textbook on the calculus, under
the influence of these lectures of BERNOULLI; L'HOSPITAL did not
add much of interest to us.
With the entrance of the BERNOULLI brothers into the field a
highly competitive race for new results begins. The principal
figures become engaged in bitter quarrels, LEIBNIz against NEWTON,
JOHANN against JACOB BERNOULLI. The net result for science
was a development of such rapidity that even modern times can

(I2) LEIBNIZ, De linea ex lineis numero infinitis ordinatim ductis inter se con-
currentenbus formata easque omnes tangente. Acta Eruditorum I692. LEIBNIZ
Math. Schriften, ed. GERHARDT, II I, p. 266-269.
(I3) LEIBNIZ. Genetalia de natura linearum, anguiloque contactus et osculi.
Math. Schriften 2e Abh. III, P. 331-337.
(I4) Differential calculus: Ostwald's Klassiker 2II, Integral Calculus (Opera III,
P. 386): Ostzvald's Klassiker I94.

This content downloaded from 7

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
98 D. J. STRUIK

scarcely break the record. Before 1700 many new curves are
discovered, many ordinary differential equations are solved, the
first elliptic integral is introduced, and the calculus of variations is
set up. In I697 and I698, the BERNOULLIS study geodesic lines
on a surface; JOHANN discovers that osculating plane and tangent
plane are perpendicular: " quod planum transiens per tria quaelibet
puncta proxima lineae quaesitae debeat esse rectum ad planum
tangens superficiem curvam in aliquo istorum punctorum." (I5)
The equation of the geodesic lines does not appear either in print
or in private letters, though JOHANN claims that he has found
it. (i6) JACOB also outlines an inquiry into the so-called isoperi-
metrical problems. Both brothers investigate orthogonal and
more general trajectories in the plane. The name " trajectory "
occurs in a letter of JOHANN to LEIBNIZ of I698. (17) An applic-
ation of this theory was found in the theory of light in a medium
of varying density, under HUYGENS' assumption that a ray of
light intersects the wave front orthogonally. An application of
JACOB lies in the finding of the orthogonal trajectories of logarithmic
curves.
The problem was taken up again in I716, when LEIBNIZ, in
the priority quarrel, tried to induce NEWTON to show the power
of his methods. He asked NEWTON (via CONTI) to find the
orthogonal trajectory to a given set of curves, for instance, all
hyperbolas of equal center and vertex. NEWTON answered, but
only in a general way and his answer does not suggest the best
method of attack. He seems to indicate that the finding of
orthogonal trajectories depends on the determination of their
center of curvatures as intersections of consecutive normals to
the given curves; this suggests a differential equation of the second
order instead of the first. (i8)
This brings us to the question of the contributions of NEWTON
to the application of analysis to geometry. Here we are unable
to find much worth mentioning except his general method. If

(I 5) For the literature see M. CANTOR, Vorlesungen III (I908), p. 229, 232,
23 5. The quoted passage in a letter to LEIBNIZ of August I698.
(i6) In letters to LEIBNIZ, see G. ENESTR6M, Sur la decouverte de l'equation
gen6rale des lignes geod6siques, Bibliotheca mathematica 13 (I899), p. 19-24.
(17) JOH. BERNOULLI, Opera I, p. 266, see CANTOR III, P. 222, 233, 443-445.
(i8) See CANTOR III, P. 444, 445.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 99

he really wrote his Theory of Fluxions, published in 1736, as early

as I671, he was the first to find an analytical expression for the
radius of curvature of a plane curve. But this is doubtful. (I9)
This dearth of investigations on differential geometry, which
continued in England long after NEWTON'S death and even now
has not disappeared, is the more remarkable as the method of
fluxions was geometrical. NEWTON'S reasoning was always
geometrical, the algorithms belonging to LEIBNIZ. But even later,
in MACLAURIN's Theory of Fluxions (1742), in which hardly any
formulae are used, we do not get more differential geometry than
the old theory of curvature for plane curves. The justifying
claim was that the book established more exact foundations of the
Newtonian way of reasoning.
We have already reported on certain publications in the early
i8th century, but they are isolated and contain no new results. The
same may be said of a series of papers by PIERRE VARIGNON (I654-
1722) (20) on evolutes and related subjects. Of more importance
is a paper by R. A. F. DE REAUMUR (I683-1757), which generalized
evolutes by considering lines intersecting a plane curve under
arbitrary angle. Then he obtains evolutes which he calls " im-
parfaites." (2I) It was a youthful production of the later thermo-
metrist and investigator of the social life of insects. But this
is almost all we can find. After 1700 the interest in differential
geometry declines sharply. The young instrument of analysis
is used for other purposes. Geometrical. problems remain almost
untouched for several decades. But the fertile days of LEIBNIZ
and the BERNOULLIS achieved a considerable result. We have
nearly the whole scheme of elementary differential geometry of
plane curves.

3.- The Eighteenth Century

For many years we have practically the work of two men,

but they were great geniuses: CLAIRAUT and EULER.

(i9) L. c., p. 171, 172.

(20) VARIGNON, Memoires pres. par div. sav. 1700-I713.
(zi) REAUMUR, Methode generale pour determiner le point d'intersection
de deux lignes droites infiniment proches qui rencontrent une courbe. Me'm.
pres. par div. sav. 1709, P. 149-I6I.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 100 D. J. STRIJIK

ALEXIS CLAUDE CLAIRAUT (1713-1765), when still a boy of

sixteen, wrote his Recherches sur les courbes a double courbure (I 73 I),
which brought him, at eighteen, into the Academie des Sciences.
The book is mainly analytic geometry of space, a subject new
in those days. Space curves enter as intersection of surfaces,
not as independent entities. CLAIRAUT uses algebra, differential
and integral calculus in a study of these curves. In differential
calculus he considers tangents, subtangents and subnormals. Only
those normal lines to space curves are considered that are normal
to the surface on which the space curve lies, which implies
the knowledge of the existence of the tangent plane to a surface.
CLAIRAUT also finds the locus of the points of intersection of the
tangents to the space curve with the plane of projection, and the
same for the normals.
The integral calculus gives the possibility of rectification and
cubature. Here we find at the same time the development
of a curve on its projecting cylinder, " si l'on imagine que la surface
cylindrique... s'etende le long du plan RAP et se developpe pour
ainsi dire."
CLAIRAUT's examples are algebraical curves, as the intersection
of y2 ax and z2 by, or x2 + y2 a2 ana y2 + z2 a2; some-
times he considers a transcendent curve, as the cycloid.
In the last part of the book he asks for the curve on a surface,
de'crite en faisant tourner dessus un compas dont une pointe
est attachee 'a un point fixe C," but he does not think so much
of geodesic circles as of the intersection of a simple algebraic
surface (as the sphere) with a cone. Another set of problems is
created by having a curve roll on another (congruent) curve in
a plane perpendicular to the plane of this curve, which he specifies
for parabola on parabola and circle on circle.
Of importance is the name of the book, through which " courbe
a double courbure " became the recognized technical term.
CLAIRAUT got the name from HENRI PITOT (I695-177I), in his
time a famous hydraulical engineer, who used it in a paper of
1724 dealing with the helix. (22) Neither PITOT nor CLAIRAUT,
however, expressed by their choice of the name any knowledge

(22) H. PITOT, Quadrature de la moitie d'une courbe des arcs, appellee la com-
pagne de la cycloide. Histoire de I'Academie de Sciences, 1724, publ. 1726,
P. 107-113.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 10I

of first and second curvature. " J'ai cruf devoir appeller ces sortes
de courbes, courbes 'a double courbure, parce qu'en les conside6rant
de la fagon qu'on vient de dire elles participent pour ainsi dire
toujours de la courbure de deux courbes, et c'est m'eme le nom
qu'on leur donne dans un memoire de l'Academie Royale des
Sciences ou on les propose comme un objet digne des recherches
des geometres ", are CLAIRAUT 'S words.
CLAIRAUT soon became interested in geodetic work, and in
a paper of 1733 on this subject he came to the theorem on surfaces
of revolution bearing his name; this theorem states that along
a geodesic line C
p sin a const,
where p is the radius of the parallel circle and a the angle of C
with that circle (23). Later he came to integrability conditions
of differential equations in studies on hydrostatics. It deserves
mention as a first step in what we now call PFAFF'S problem. (24)
He found that
MUdx + Ndy + Pdz o
is exact, when and only when,

/8 +P N 8 +M P8 - -N i
M ( 3N Jr N (P- Mz (My AxN

He actually proves that the condition is not only necessary, but

sufficient.
LEONARD EULER'S (1707-1783) work is so varied that it is hard,
in this outline, to do him justice. From early youth he con-
stantly turned to the application of the calculus to geometry, from
work done in 1727 on parallel curves in the plane, (25) inter-
secting under constant angle to his paper of 1782, on the
differential geometry of space curves. In a series of papers
between I728 and 1732 he takes up the problem of the geodesics

(23) A. C. CLAIRAUT, Determination geometrique de la perpendiculaire a la

meridienne tracee par M. CASSINI, lb. 1733, publ. 1735, p. 406.
(24) The work dates from I739 and 1740, see CANTOR III, p. 856, 86i. See
also CLAIRAUT'S " Theorie de la figure de la terre, tiree des principes de l'Hydro-
statique. Paris, 1743. German translation in Ostwald's Klassiker, no. I89.
(25) See for the literature F. MOLLER, tber bahnbrechende Arbeiten L. EULERS
aus der reinen Mathematik. Abh. zzur Geschichte d. Math. Wiss. 25 (1907),
p. 63-1I6, esp. Io8-II3, or CANTOR III and IV.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 102 D. J. STRUIK

on a surface. JOHANN BERNOULLI had attracted his attention

to it, probably through the aid of his nephew DANIEL, who was
at St. Petersburg with EULER. A result was the equation of the
geodesics, in the form
Qddx + Pddy dx ddx + dy ddy,
Qdx + Pdy dt2 +dx2 + dy2
where Pdx -Qdy + Rdt connects the variables of the surface. (26)
He made applications to several types of surfaces, for instance
cones. In I732 he uses the coordinates (x, s) in a discussion
of the cycloid as a tautochrone. In the same number of the
Commentarii this is also done by G. W. KRAFT. In I736, in a
paper on the tractrix, he introduces among other new coordinates
arc length s and radius of curvature p as coordinates of a plane
curves, and so opens the series of papers on intrinsic geometry.
He shows how x and y can be found when p and s are given.
In the Mechanics of I736 he proves that mass points on a surface
without a force field move along geodesics. In I740 he studies
evolutes and involutes, a study leading him, in I764, to the curious
result already announced by JOHANN BERNOULLI, that the nth
involute of a curve for increasing n tends to become a cycloid.
With all these results it is rather astonishing that the Introductio
in analysin infinitorum, the standard textbook EULER published in
I748, contains so little differential geometry. It may have been
the intention of EULER to write a special book on this subject; (27)
if so, it was never accomplished. The Introductio contains
only some remarks on singular points and asymptotes of plane
curves, and some osculation properties. EULER, writing the
equation of plane curves in the form
o At + B u + Ct2 + Dtu + Eu2 + F t3 +.
A, B, C, D, ... constants
is first led to introduce an osculating conic section at the origin,
which he approximates by a parabola, but then changes to the
osculating circle. There are also some remarks on concavity
and convexity in relation to the ambiguity of the sign of the
radius of curvature. (z8)

(26) JOHANN BERNOULLI had the equation also in I728, printed in his Opera IV
(I744) P. io8. Here the word " planum osculans " for the first time.
(27) See CANTOR III, P. 784.
(28) L. EULER, Introductio in analysin infinitorum, II, ch. XIV.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I03

It is a remarkable fact that only in I760 EULER opens an entire

new field in differential geometry. All his previous work has
been more in the way of elaboration of old results of LEIBNIZ
and the BERNOULLIS, with the possible exception of the intro-
duction of natural coordinates. Even CLAIRAUT'S work has merit
for differential geometry only as a statement of the problems.
We must, however, make reservation for EULER'S fundamental
work on the calculus of variations, culminating in his Methodus
inveniendi lineas curvas maximi minimive proprietate gaudentes of
I744, in which he not only states and gives methods of solutions
to isoperimetrical problems, but finds interesting geometrical
properties of curves. The best known perhaps is the theorem
that the catenoid is a minimal surface. (29)
EULER'S paper of I76o, Recherches sur la courbure des surfaces, (30)
written during his Berlin residence, contains the first important
contribution to surface theory, and also to three-dimensional
differential geometry in general. So far only the existence of
the tangent plane at a point of a surface had been established,
and that in not a very satisfactory way (for instance, by CLAIRAUT).
EULER here takes a definite step forward, and arrives at the so-called
EULER theorem on curvature of surfaces. It states, in EULER'S
terms)

r= 2 f g -
f + g + (f-g) cos 2a
where f and g are the extreme values of r, the radius of curvature
of a normal section, and a is the angle of this normal section
with one of the normal sections of extreme curvature. The
form under which we know the theorem is due to DUPIN, but
the name " section principale " is due to EULER, as well as the
theorem that the two sections of extreme curvature are normal
to each other. His demonstration starts with an arbitrary plane
section through a point of the surface, then proceeds to an expres-
sion of the radius of curvature for this section, the expression
being gradually simplified.
Shortly afterwards (I762) LAGRANGE (I736-I8I3), then a young

(29) L. EULER, Methodus inveniendi V, Ex. VII 47, German translation in

Ostwald's Klassiker, 46, p. i iI.
(30) Histoire de l'Acad6mie royale des Sciences (Berlin), I760, p. II9-I4I.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 I04 D. J. STRUIK

professor at Turin, published his famous paper on the calculus

of variations, the main results of which he had already shown
to EULER in 1755. In an appendix, he found the differential
equations of the minimal surfaces in the form that p and q must
be found under the condition that

p dx + q dy, pdy - qdx

are exact differentials. (3 I)
In I770 EULER continued his study of surfaces, and began
to investigate developables. He represents the x, y, z of a point
on a surface as functions of two variables t and u (the first time
the so-called Gaussian variables are introduced), and writes down
the conditions that
dx2 + dy2 + dz2 dt2 + du2;
12 +m2 +n2== I, A2+p2 +V2 =I lA +mH ?n
ex ay bz A x ay az
I- -- m - -, n~ --; A~ -- -) v =--->
or, as he states, that
" Une consideration tout 'a fait singuliere m'a conduit 'a la solu-
tion de ce probleme," he writes to LAGRANGE. EULER is able to
integrate the equations and to show that the tangents to an arbi-
trary space curve form such a developable surface. (32) From
his integral he does not seem, however, to draw the conclusion
that such surfaces are the only real solution.
Differential geometry had advanced thus far when an entirely
new development started. With the exception of EULER'S papers
and occasional work of LAGRANGE, very little had been done

(3 I) J. L. LAGRANGE, Essai d'une nouvelle m6thode pour determiner les maxi

et les minima des formules int6grales ind6finies. Miscellanea Taurinensia I760-6I,
publ., I762, p. I73-I95. Giuvres I, p. 335-362. German translation in Ostwald's
Klassiker, 47, p. 23.
(32) L. EULER, De solidis quorum superficiem in planum explicare licet. Nova
Comm. Petrop. i6, I77I, p. 3-34. See EULER'S letters to LAGRANGE of Jan. i6
and March 9, I770, CEuvres de Lagrange XIV, p. 2I7, 2i8, 22I-223, 224.
There is another paper of EULER'S hand on surface theory, written in this time,
between I766 and I775, but only published in i862, Opera posthuma I, p. 494-496:
" Problema invenire duas superficies quarum alteram in alteram transformare
licet, ita, ut in utraque singula puncta analoga easdem inter se teneant distantias."
Here we find that surfaces are applicable if, in modern notation, E, F, G are equal,
and the remark that a closed surface cannot be bent.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 105

for a long time. Feudalism, in decay, could not send auxiliary

forces to the aid of the lone genius. EULER, in many respects,
represented this last period of the feudal system, which disappeared
intellectually with such undeniable elegance. EULER'S creations
perhaps may find a counterpart in those of MOZART.
LAGRANGE felt it, " Ne vous semble-t-il pas " he wrote to D'ALEM-
BERT in I772, " que la haute geometrie va un peu en decadence ? "
He expresses the same view at other places and D'ALEMBERT's
answers are sceptical. (33)
We can interpret the new life, which was developing at the
military academy of Mezieres, as the beginning of the influence
of the French revolution on geometry. Here GASPARD MONGE
(I746-I8I8) was professor since I768, and began in that early
time to show that fecundity in geometrical invention which
made him the real creator of differential geometry, of descriptive
geometry, and directly and indirectly, of modern geometry in
general. His starting point was a series of questions on fortification,
which led him to descriptive geometry, but he also knew how
to use analysis. His first publication, in I77I, already showed
the master. It deals with space curves, the first paper on this
subject since CLAIRAUT treating this subject for its own sake. (34)
It contains a broad exposition of the whole differential geometry
of space curves. It is shown how such curves admit an infinity
of evolutes, that they all lie on a developable surface, the polar
developable, and that they are geodesics of this surface. He also
introduces what we would call the rectifying developable and shows
that the original curve is a geodesic on this surface. Here appears
the normal plane, the. radius of first curvature, the osculating
sphere. Two types of inflexion exist: inflexion caused by (what
we call) torsion zero, and inflexion caused by (what we call)
curvature zero. In the first case, the " points de simple inflexion,"
four consecutive points of the curve lie in one plane, in the second,
the " points de double inflexion," three consecutive points lie
on a straight line. Several terms, since adopted, appear here

(33) CEuvres de LAGRANGE XIII, P. 229, 232, 237.

(34) G. MONGE, Memoire sur les developpees, les rayons de courbures et les
differents genres d'inflexion des courbes 'a double courbure. Mem. div. savans
1785, p. 5I I-550 (written I77I), also last chapter of the " Applications de l'Analyse
h la Geometrie," where only a part is reprinted.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 io6 D. J. STRUIK

for the first time, as " ligne des pales," " arete de rebroussement,"
" developpee." Many applications to plane and space curves
illustrate the general theorems.
In 1780 MONGE published a second paper, written in I775 (35),
in which he took up EULER'S theory of developables. MONGE
intends to simplify EULER'S results. But in his hands the whole
theory takes another shape. The geometrical part is treated
in such a way as to make the great author of the Recherches
sur la courbure des surfaces and of many more contributions to
geometry more analyst than geometer. Nevertheless there is a
good deal of the analyst in MONGE. But the formulas always
follow the dynamics of geometrical development, so that the
integration of a partial differential equation becomes the gradual
building up of a geometrical system in space. Nobody except
LIE ever equalled MONGE in this direction.
MONGE points out the essential difference between general
ruled surfaces and developables, sets up the differential relation
rt -S2 = o and finds as first integral that there is an arbitrary
relation between p and q, which means that a developable is always
tangent surface to a space curve. It is also the envelope of a two
parameter family of planes. Application is made to the tangent
developable of two surfaces, which was already partly elucidated
in EULER'S work, but which as a problem of " ombres et pe-
nombres " had a great attraction for the inventor of descriptive
geometry. We also find here the differential equation of the
third order for the ruled surfaces, with the solution of the problem
of finding the ruled surface passing through three space curves.
The volume of the Memoires des savans etrangers, of 1785,
which contains MONGE'S first paper, contains another classic of
differential geometry, MEUSNIER'S Memoire sur la courbure des
surfaces, written in I776. (36) The title already shows the
indebtedness of the author to EULER. MONGE had, indeed,

(35) G. MONGE, Sur les propri6t6s de plusieurs genres de surfaces cou

particulierement sur celles des surfaces developpables, avec une application 'a
la theorie des ombres et des penombres. Mem. div. savans IX, I780, p. 593-624
(written 1775).
(36) Mem. sav. etrangers 1785, P. 477-5IO. An exposition of MEUSNIER'S
paper not only in CANTOR IV, P. 547-550, but also in DARBOUX, Theorie generale
des surfaces I (i887), P. 260-27I.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 107

recommended the paper of EULER to one of his pupils, JEAN

BAPTISTE MEUSNIER DE LA PLACE (1754-1793), and, working under
the direction of his teacher, the young officer not only found
EULER'S results in a new way, but added the results which for ever
carry his name. In one of the principal sections of the surface
at a point he draws a circle, tangent to the surface, with radius
equal to the " rayon de courbure " at that point in that direction.
This circle is rotated about an axis in its plane parallel to the
tangent plane and at a distance equal to the second principal
radius of curvature. In this way MEUSNIER gets a torus which
has the first and second derivatives in common with the surface
at the point. Then he takes this torus as representative of the
surface and gets not only EULER'S theorem, but also " MEUSNIER'S
theorem," which he interprets with the aid of a sphere tangent
to the surface and with radius equal to the normal radius of cur-
vature of the section in the arbitrary direction on the surface.
MEUSNIER uses his torus in finding the condition under which
a surface be a minimal surface. LAGRANGE had already found
the differential equation. MEUSNIER interprets it by showing
that it means that the sum of the radii of principal curvature
is constant. Then he interprets this equation by simple geo-
metrical methods in two cases, and finds the twisted helicoid
and the catenoid, which for many years were the only minimal
surfaces known. EULER already had found the catenoid, but
MEUSNIER, it seems, found his solution independently.
This paper remained the only contribution of MEUSNIER to
mathematics. He published it under the best auspices; D'ALEM-
BERT, feeling the new spirit, said " MEUSNIER commence comme
je finis." But MEUSNIER went to other spheres of activity, where
he also did excellent work. He collaborated with LAVOISIER
to separate water in its constituents (paper of 1784) and wrote
important papers on the new subject of aeronautics. " Apres
avoir consacre sa trop courte vie aux recherches les plus neuves,
les plus difficiles, les plus f6condes, il a trouve devant l'ennemi,
au siege de Mayence, la mort la plus heroique." (37)

(37) G. DARBOUX, Notice historique sur le g6neral MEUSNIER, I909. In " Eloges
academiques et discours, Paris", HERMANN, I9I2, p. 2i8-262. GOETHE describes
how he watched the French soldiers, defeated, leaving Mayence. They carried
the body of MEUSNIER away with them.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 io8 D. J. STRUIK

CHARLES TINSEAU (1749-I822) was also a graduate of the Me-

zieres academy, class of 1771. He presented a paper to the
Academy in I774, which contains, among several fundamental
contributions to the analytic geometry of space, the equation
of the osculating plane to a space curve, the surface of the tangents
to a curve (already introduced by CLAIRAUT), and the theorem
that the orthogonal projection of a space curve on a plane has
a point of inflexion if its plane is perpendicular to the osculating
plane. (38)
In this period falls a charming paper by EULER, in which he
investigates what we now call curves of constant breadth, in EULER'S
terms: " orbiformes." He gets them as involutes of " triangular
curves," that are closed curves with three cusps. (39)
In the meantime MONGE had continued his productivity, of
which we shall say more in the next chapter. We only mention
a paper of I 78I, Memoire sur la theIorie des deblais et des remblais, (40)
which takes as starting point the engineering problem of moving
a heap particle after particle from one place to another in a
minimum of effort. This leads to line congruences, which admit
two sets of developable focal surfaces. When these are normal,
the congruence is normal to a surface, and cuts it along the lines
of curvature. In this original way the lines of curvature were in-
troduced into literature. Then there is another paper of I784, in
which he integrates the equation of the minimal surfaces. (4I)
There are more papers, equally fundamental; but as all are col-
lected in his book of i8o8, we may discuss them together. Since
1780 he had been living in Paris for six months a year, where
he taught hydraulics at the Louvre, but after BEZOUT's death, 1783,
he settled there permanently.
About this time the aging EULER again wrote a fundamental

(38) C. M. T. TINSEAU, Solution de quelques problemes relatifs 'a la theorie

des surfaces courbes et des courbes 'a double courbure. Mim. div. savans IX,
1780, P. 593-624.
(39) L. EULER, De curvis triangularibus. Acta Petr. 2 (1778), P. 3-30.
(40) Mem. div. sav. 178I (publ. 1784). See P. APPELL. M6morial des Sciences
math6matiques XXVIII, also our footnote (50).
(41) G. MONGE, Une m6thode d'integrer les 6quations aux diff6rences ordinaires.
Mem. div. sav. 1784, P. i i8. An improvement was suggested by LEGENDRE,
l'Int6gration de quelques 6quations aux diff6rences partielles. Mem. div. sav.
1787, P. 311-12.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY lO9

paper (42). It sets forth the first analytical treatment of

the differential geometry of space curves. MONGE had treated
the subjects from a geometrical point of view, but had not given
an analytical frame. EULER provides for this by taking x, y,
z as functions of the arc length s, and the direction coefficients
of the three axes of the moving trihedron. For this purpose
EULER introduces the spherical image, using the unit sphere as
GAUSS did forty years later. The equation of the osculating
plane is here given in the symmetrical form x (rdq-qdr) +
y (pdr-rdp) + z (qdp-pdq) t, where t is determined by
the condition that the plane must pass through a given point
of the curve. This symmetrical way of treating coordinates
also characterizes other papers of EULER, as, for example, one
of 1779, in which he writes the equation of the geodesic lines
on a surface,
d2x(qdz-rdy) + d2y(rdx-pdy) + d2Z(pdy-qdz) o,
where pdx + qdy + rdz= o
is the differential equation of the surface. In this paper the
integration is carried out for rotation surfaces, partially repeating
thereby results of CLAIRAUT. (43)
An account of the important work on map projection done in
the i8th century, and again through the efforts of EULER and LA-
GRANGE, supported by LAMBERT is still missing in our report. All
this work is carried out after I770, the time of the revival of differ-
ential geometry in general. In I777, EULER introduced complex
numbers in his study of conformal projection, which LAGRANGE,
in the same year, used for the more general problem of mapping
meridians and parallels of a sphere into an arbitrary orthogonal
system of plane curves. There is a discussion of this work by
V. KOMMERELL in CANTOR'S Vorlesungen IV, p. 572-576 and
in books on cartography. Here, on p. 508-5 II, is also a discussion
of work on parallel curves done in the same period (EULER, NIEU-
PORT, KASTNER, and others).

(42) L. EULER, Methodus facilis omnia symptomata linearum curvarum non

in eodem plano sitarum investigandi. Acta Petr. I782, I, P. 19-57 (publ. 1786).
(43) L. EULER, Accuratior evolutio problematis de linea brevissima in superficie
quacumque ducenda, Nov. Act. Petr. XV, 1779, P. 44-54.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 110 D. J. STRUIK

4. - Monge and the tcole Polytechnique.

The French Revolution influenced scientific thought in all

directions. Under its influence modern geometry was born. In
algebraic geometry the strict prescriptions of Greek thought were
discarded, and an entirely new school of reasoning was created.
In differential geometry the mathematicians at last learned to
apply the century-old thoughts of LEIBNIz and the BERNOULLIS
and to establish a science on collective work where EULER SO long
had pioneered alone.
GASPARD MONGE was uncontested leader. His political work
was reflected in his scientific activity. During the revolution
he joined the Jacobins, but like many of his political friends
he later supported the Empire, which they interpreted as the
executor of the will of the French Revolution. NAPOLEON entrus-
ted to MONGE many important functions; he even made him
for a while secretary of the navy. On the expedition to Egypt
he had with him MONGE as well as many other famous scholars.
But MONGE'S life work became the organization and scientific
leadership of the Ecole Polytechnique, of which he was the director
from its beginning, in I794, till the fall of the Empire. But the
Restoration and the old Republican were irreconcilable, and
MONGE had to resign. He died a few years later, closing a life
not only crowded with scientific achievements but characterized
by a unity of thought and deeds seldom found among scholars.
The importance of the Ecole Polytechnique for the development
and the organization of science has so well been treated by FELIX
KLEIN in his lectures on the history of mathematics in the igth
century (44), that we need not discuss it here. From the beginning
MONGE'S teaching was an integrating part of the instruction.
Here his remarkable geometrical intuition went hand in hand
with practical engineering applications to which his whole manner
of thinking was always inclined. He also started a collection
of models, later continued by TH. OLIVIER. He taught descriptive
geometry as a new subject, and he collected his lessons in the

(44) F. KLEIN, Vorlesungen iiber die Entwicklung der Mathematik im i 9. Jahr-

hundert, I. Berlin, SPRINGER, 1926, Ch. II. KLEIN refers to JACOBI'S paper.
Werke 7, p. 355-370.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I I I

Gdomdtrie descriptive, which still is a standard treatise. He taught

differential geometry in the same way; these lessons appeared first
as loose pamphlets, the Feuilles d'Analyse appliquee a' la Geometrie
(since I795), then, in I807, with little modification, as a book,
Applications de l'Analyse a la Geometrie. The lessons show all
the characteristics of MONGE'S genius. (45)
Those who are interested in a discussion of the Feuilles d'Analyse
can find the material in KOMMERELL'S paper in CANTOR IT. We
will here discuss the main line of the Applications, which however
differ but slightly from the Feuilles.
The leading thought of the book is the geometrical interpretation
of partial differential equations and the interpretation of geo-
metrical facts into the language of partial differential equations.
For this MONGE develops the theory of envelopes, characteristics,
and edges of regression. At the same time he shows what the
integration process means in space.
As the simplest example let us take Chapter II on cylindrical
surfaces. These surfaces can be considered in different ways.
If we look at them as surfaces of which the tangent plane is parallel
to the generating line and therefore parallel to the direction
x az, y bz, we get as the equation ap + bq - i, (p anc 8

are the symbols p -, q , ).

But cylindrical surfaces are also surfaces of which the generating
line is always parallel to the line x _ az, y bz. This gives
as equation y - bz .p (x- az), where cp is an arbitrary function.
In this way we get the integral of ap + bq i. From this we
can solve several other problems, as the determination of the
cylindrical surface if the direction of the generators and a space
curve directing their motion is known, or the determination of
a surface that envelops a surface along a given curve. At the
same time this shows the manipulations which may be effected
with partial differential equation.
In this way every chapter of the Applications is built up.
MONGE classifies his problems into those leading to partial differen-
tial equations of the first order, of the second order, and the third

(45) The most interesting edition is the fifth, with notes of J. LIOUVILLE, i850.
The " Feuilles d'Analyse " appeared again in i8oi.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 112 D. J. STRUIK

order. To the first category belong the cylinders, the cones,

the canal surfaces; to the second, the developables (rt -S2 _ )
the ruled surfaces with generators parallel to a given plane, and
to the third category, the ruled surfaces.
But to the second category belongs also that whole class of
surfaces whose curvatures satisfy certain simple conditions.
MONGE is, starting in I784, the first to introduce the lines of
curvature and their properties (4I) In a clever way he integrates
(Ch. i6) the lines of curvature on an ellipsoid, by increasing the
order of the differential equation. He then solves the following
four problems:
i) surfaces with one set of lines of curvature plane (Ch. I7)
2) surfaces with R1 const (Ch. i8)
3) surfaces with R1 R2 (Ch. i9)
4) surfaces with R1 R2 (Ch. 2o)
Problem i) leads to the molding surfaces (" surfaces moulures "),
z) to the tube surfaces, 4) to the minimal surfaces. But in 3)
MONGE finds the paradoxical result that the sphere alone isasolution.
In 3) and 4), he integrates the equations in full, and this shows
him the explanation of the paradox. There is, indeed, an extended
class of families answering problem 3). But the sphere is the
one real surface. All other surfaces are imaginary with one real
curve on each. These surfaces, says MONGE, are really curves,
with area everywhere zero. This is the first full discussion of
imaginaries in geometry.
MONGE, however, does not give any other special cases of his
general equations of the minimal surfaces than the two known
since MEUSNIER.
Another example of a problem leading to a partial differential
equation of the third order is the problem of the spheres of variable
radius with centers on a space curve (Ch. z2). If the radius
is constant the problem is of the second order, if the curve, besides,
is plane, it is of the first. The methods used, in all cases, follow
the same line of thought.
The last chapters deal with another problem at which MONGE
arrived by noting that the normals to a surface along the lines
of curvature form a developable surface. This new problem
is the inverse one; to find the surfaces of which the normals are
tangent to a given surface.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY II3

From these lectures many words remained standard expressions.

Besides those mentioned, we have: " lignes de courbure, "
" enveloppe," " characteristique," and the notation p, q, r, s,
t for the partial derivatives. But not only MONGE's differential
geometry, also his descriptive geometry has a bearing on our
subject. It deals with curves and surfaces, but in a purely con-
structive way, without formulas. Two ways are thus indicated
as methods of attacking geometrical problems on curves and
surfaces, the geometrical and the analytical. We see this separation
clearly in work of MONGE'S pupils, as for example DUPIN, who
proves many theorems twice, both by geometry and by analysis.
In our present day differential geometry we still show that influence,
when we define lines of curvature, asymptotic lines, conjugate
lines in two different ways. From this " descriptive " type of
geometry, which MONGE taught, projective geometry emanates
in the hands of his pupils.
MONGE'S general idea of connecting partial differential equations
with geometry of space is still a leading method in differential
geometry, especially in France. In a modernized way, though
only through indirect influence, it dominates the work of SOPHUS
LIE.

5. - Monge's Pupils

A galaxy of brilliant men supported MONGE at the Ecole Poly-

technique either as colleagues or as pupils. A school of mathe-
matics was the result, in which analytical geometry and differential
geometry flourished, and in which projective geometry was created.
A stimulating influence had the necessity of teaching courses
on advanced subjects, and there was a regular output of textbooks
on mathematics and mechanics, many of which established stan-
dards valid till today. Almost all geometry was threedimensional.
Our theory of quadric surfaces dates from those times, and a
considerable attention was paid to plane and space transforma-
tions.
Common to all these scientists was their contact with practice,
either in the abstract form of mathematical physics and mechanics,
or in the direct form of engineering and economic or political
activity. They represented one phase of the emancipation of

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC 8
All use subject to https://about.jstor.org/terms
 114 D. J. STRUIK

the French bourgeoisie. NAPOLEON expressed also their ideas

when he wrote to LAPLACE: " L'avancement, le perfectionnement
des mathematiques sont lies 'a la prosperite de 1'e6tat."
Among the colleagues of MONGE of mature age we find LAGRANGE
and LAZARE CARNOT; among his younger colleagues and pupils,
FOURIER, AMPiERE, POISSON, PONCELET, RODRIGUES, LANCRET,
CORIOLIS, MALUS, DUPIN, FRESNEL, CAUCHY, SADI CARNOT, SOPHIE
GERMAIN. For our purpose we must examine closely the work
of AMP'ERE, LANCRET, MALUS, RODRIGUES, and especially that
of DUPIN.
A. M. AMPtRE'S (I775-I846) mathematical discoveries are less
remembered than his physical, though a certain type of partial
differential equation carries his name. He commenced his main
physical work only after OERSTED'S discovery in i 82o of the
influence of the electric current on a magnetic needle, when
he was already famous as a mathematician. For us a paper
on osculating parabolas (46) is of importance, because it contains
the notion (if not the name) of the differential invariant. AMPiERE
recognises the importance of p and s as instrinsic coordinates of
a plane curve, but remarks that s still depends on an arbitrary
constant. Therefore higher derivatives are necessary for truly
intrinsic coordinates. He chooses for this the osculating parabola.
When its equation with respect to tangent and normal at a point
of the curve is u2 = pt, (p the parameter), the curve can be given
as a function between u and t. He shows that u, p, t are differential
invariants under rotations and translations. They depend on
third derivatives; and AMPE'RE approaches affine differential geome-
try, when he finds the condition for points with parabola osculating
in five consecutive points (affine curvature zero, as we say now) (47).
To similar contact with affine conceptions came LAZARE CARNOT
(I753-I823) who, in his Geometrie de position of I803, defined
what we now call the affine normal. He also proposed intrinsic
coordinates, the radius of curvature, and the angle of affine normal
with the ordinary normal, as in AMPERE'S case a mixture of affine

(46) A. M. AMPE'RE, Sur les avantages qu'on peut retirer, dans la theorie des
courbes, de la consideration des paraboles osculatrices, avec des reflexions sur
les fonctions differentielles dont la valeur ne change pas lors de la transformation
des axes (presented I803). Journal Ec. Polyt. I4e cah. (i8o8), P. 159-181.
(47) See p. 178 of AMPERE's paper.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 115

and metric conceptions, which could not be very fertile. (48)

That there were many discussions on intrinsic coordinates in
those days is also seen in S. F. LACROIX' much used textbook
in three volumes on differential and integral calculus, which
in its first volume has more than 250 pages on curves and surfaces.
He deals faithfully with all results obtained by EULER, MONGE,
MEUSNIER, LAGRANGE and others. He also discusses the sug-
gestions of CARNOT and AMPERE, and mentions other ways to
study curves independent of their position in the plane. (49)
E. L. MALUS (1775-i8i2) is famous as the discoverer of the
polarization of light (i8o8). His investigations in optics lead
him, as later HAmILTON, to the study of line congruences. This
theory dates from MONGE'S work of I78I on " deblais and rem-
blais " (5o), but MALUS establishes again with his methods the
theorem that in such a congruence each line is in general cut
by two other lines, so that the lines are the intersection of two
families of developpable surfaces. The application to normals
to a surface establishes the theorem on lines of curvature. MALUS
studies the behavior of line congruences under reflection and
refraction. He also studies what we now call line complexes.
Through a mistake he fails to obtain entirely the "MALUS-DUPIN"
theorem. (5 I)
With MICHEL ANGE LANCRET (I 774- I 807) we have a young,
promising scholar, who died too early to fulfil his promises.
He belonged with MONGE, FOURIER, AMPtRE, GEOFFROYST.HILAIRE
and many others, to the scholars who accompanied NAPOLEON
on his Egyptian expedition. Later he became a member of the
commission appointed to publish the results, but he died at

(48) L. N. M. CARNOT, Giom&trie de position. Paris, I803, 489 p., see

Probleme LXXVI, art. 433, p. 477, and art. 432, p. 475-476.
(49) S. F. LACROIX, Trait? du calcul difThrentiel et du calcul int6gral. Tome I
Paris. Seconde 6d. I8Io, 652 p., espec. no. 255, p. 484-485. From AMPRE3's
paper it seems that LACROIX in the first edition of his book started the discus-
sion. The work was continued by GERGONNE, Annaks de mathSmathiques, 4
(iB83-14) p. 42-55. GERGoNNE, p. 372 of the same volume, in restating some
sults of DUPIN, frames the term "1 tangentes principales ".
(So) See note (41), also C. SEGRE, MONGE e le congruenze generali di r
Bibliotheca mathematica 3 x8 (1907-08).
(5I) MALUS, Optique, Jour. Ec. Polyt. 14 cah., I8O8, p. I-4; Dioptrique,
ib., p. 84-129; Traite d'Optique, Mimoires prisentis a t'InstitUt 2 (i8ii), P. 2I4-
302, cont. p. 303-5o8 as " Th4orie de la double refraction."

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 D. J. STRUIK

33 years of age. He wrote two mathematical papers, (52) the

first on the general theory of space curves, the other (published
after his death) on " developpoides," space curves whose tangents
are lines cutting a given space curve under a constant angle different
from go9. His first paper is of a more general nature. It contains
the two fundamental quantities of the space curve, which he
calls " premiere flexion " and " seconde flexion." The first is
the angle dj of two consecutive normal planes, the second the
angle dv of two consecutive osculating planes. (53) Curvature
and torsion appear therefore as differentials, and are not written
as finite quantities until CAUCHY. There is a third quantity,
the angle dwo of two " plans rectifians " (this name and conception
also appears here), and the " equation of LANCRET" exists
d,12 + dV2 = dCO2
which shows that only two of the three quantities are independent.
As an application we learn that the first flexion of the " d6velop-
pante " is equal to the second flexion of the " developpee," and
vice versa, a theorem only correct for the differentials.
LANCRET is therefore the first to take up the systematic theory
of space curves after EULER, but it seems in an independent way.
The line of progress goes here from CLAIRAUT via EULER and
LANCRET to CAUCHY and FRENET.
OLINDE RODRIGUES (I794-I85I) did some work on lines of
curvature, simplifying MONGE'S results. A set of important for-
mulas still is called after him (54), and so a formula in the theory
of LEGENDRE functions. He found also what we now call the
Gaussian curvature of a surface by comparing an element of
surface with its spherical image; he missed, however, the " theor-
ema egregium." He later became acquainted with ST. SIMON

(52) M. A. LANcRE'r, Memoire sur les courbes A double courbure. Memoires

presentes d l'Institut i (I8o6), P. 4I6-454 (presented 1802); Memoire sur les
d6veloppoides des courbes planes, des courbes i double courbure et des surfaces
d6veloppables, ib., 2 (i8I I), p. I-79, presented i8o6.
(53) LANcurr says that he got his ideas from FOURIER (p. 42o). FOURER did
not publish it himself.
(54) 0. RODRIGUES, Recherches sur la theorie analytique des lignes et des rayons
de courbure des surfaces, et sur la transformation d'une classe d'int6grales doubles,
qui ont un rapport direct avec les formules de cette thdorie. Correspondance
sur l'Ec. Polyt. 3, i8I5, P. I62. Also a paper in Bull. Soc. Philomatique Paris,
(3) 2, I815, P. 34-36: ((Sur quelques proprietes des integrales doubleset des rayons
de courbure des surfaces." Here the name is given as Rodrigue.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY II7

and did considerable work to propagate the socialistic theories of

his master, which prevented him from work in mathematics.
CHARLES DUPIN (1784-I873), since T824 Baron DUPIN (many
eminent mathematicians were given titles of nobility in those days,
a custom introduced by NAPOLAON; MONGE, for instance, became
" comte de Pelouse ") is the most important differential geometer
among the direct pupils of MONGE. Many of his discoveries
were made long before he published them in this books, the delay
being partly due to his many duties as a naval officer, some of
which carried him " dans des pays presque barbares," such as Corfu.
At the age of sixteen he discovered the " cyclide of DUPIN ", (55)
about I807 the " DUPIN theorem " on orthogonal surfaces.
His results were collected in the Developpements de g6om6trie
(I8I3), later followed by the more applied mathematical Applica-
tions de geomdtrie et de mdcanique of I822. (56) The De'veloppe-
ments, he explicitly stated, have been written as a sequel to MONGE'S
books. MONGE'S two methods of approach, the descriptive and
the analytical, appear as different currents in DUPIN's book, half
of which is purely geometrical, and half of which uses analysis.
These tendencies were soon afterward to grow into entirely
different branches, into projective geometry and differential geo-
metry proper. DUPIN therefore in almost all cases proves his theo-
rems in two different ways, geometrically and analytically, again fol-
lowing MONGE who defined lines of curvature geometrically as
lines along which the normals form a developable surface and
analytically as lines along which the normal curvature has extreme
value.
The D6veloppements are divided into two sections. The first
section contains the theory of the indicatrix, the second of the
orthogonal systems of surfaces. The main discoveries of the
first section are asymptotic lines and the conjugate sets; that
of the second section, " DUPIN'S theorem " on the lines of inter-
section of triply orthogonal systems. This enables the author

(55) See J. BERTRAND, Vloges acad6miques. Paris, Hachette, I890, P. 22I-2

(56) CH. DUPIN, Developpements de geometrie, avec des applications & la
stabilit6 des vaisseaux, aux d6blais et remblais, au d6filement, & l'optique, etc.
Th6orie. Paris I8I3.
CH. DUPIN, Applications de g6ometrie et de mecanique i la marine, aux ponts
et chauss6es, etc., pour faire suite aux d6veloppements de geometrie. Paris, I822.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 II8 D. J. STRUIK

to give a new treatment of MONGE'S lines of curvature on an

ellipsoid by introducing confocal quadrics. The general theory
of the indicatrix leads to a discussion of elliptic, hyperbolic and
parabolic points, and throws new light upon the lines of curvature
and the umbilics. DUPIN even enters into a discussion of the
simplest cases of lines of curvature through an umbilic (guided
by their behavior on an ellipsoid), and he gives the first geometrical
proof of MONGE'S theorem that the sphere is the only real surface
with only umbilics.
From DUPIN's book date several names, as asymptotic lines,
and conjugate directions, and also the modern form of writing
EULER'S equation,
I cos2a sin2a
_.+

R R1 R2
DUPIN continued his research in the Applications, where he
attacked many problems in applied fields, as stability of floating
bodies, optics, and " deblais et remblais." Here, moreover, we
also find the correction of " MALUS-DUPIN'S theorem" on normal
systems of straight lines, and the " cyclide " of DUPIN. (57)
DUPIN lived to a ripe age, but did not continue his work on
geometry. His travels led him to many countries, and he became
especially interested in the growth of capitalism in England,
which he liked to propagate in France. This he did in a great
number of papers and books on social subjects. He entered
politics, became " pair de France " under the Restauration and
"senateur."
As representative of the purely geometrical school of MONGE
in the time of the Restoration we have Louis L. VALLAE (I804-
I864), a prominent civil engineer, who wrote a Traite' de gdom6trie
descriptive in I8I9, reprinted in I825, and dedicated to MONGE.
It contains the theory of space curves and surfaces, showing
how their theory is built up by geometrical reasoning. Here
we find for the first time the word " angle de courbure " together
with " angle de torsion." VALLAE reveals to us why geometry
should be studied, quoting MONGE: " Pour faire fleurir l'industrie
fran9aise, il faut diriger l'education nationale vers la connaissance

(57) MALUS-DUPIN'S theorem on p. I9I, with a criticism of MALUS. See for

the history DARBOUX, Surfaces II, p. 280.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY II9

des objets qui exigent l'exactitude," and remarks that DUPIN

" indique, comme une cause remarquable des succes manufacturiers
de l'Angleterre, les soins qu'on donne a l'instruction des ouvriers
anglais." (58) It is no accident that we see so many geometers
of the Napoleonic time,- MONGE, DUPIN, RODRIGUES, VALLEE-
interested in industrial problems. JACOBI'S contention that the
only goal of science is the honor of the human mind belongs to
a generation already emancipated from the revolution.
We must devote a few remarks to the Thdorie des fonctions
analytiques, in which LAGRANGE, in the first years of the Ecole
Polytechnique, tried to build up calculus without the use of
infinitesimals (I797). As his main object is the study of " deri-
vees " f'(x), f" (x), etc. (notation and name appear here for
the first time) he is attracted by the contact of curves and surfaces.
In this book therefore we find for the first time an elaborate
analytical theory of osculation, illustrated by many examples. It
served as model to all later expositions of the subject, together
with that of CAUCHY.
As the last representative of this school we have therefore to men-
tion A. CAUCHY (1798-I857), who became professor at the Ecole
Polytechnique as a royalist, in i8i6. He wrote on differential
geometry in one of his textbooks (i826) " destin6 'a faire suite
au Resume des lecons sur le Calcul infinitesimal." (59) This
textbook contains a beautiful exposition of the theory as it stood
in CAUCHY'S time at Paris, and shows in many respects the
characteristics of CAUCHY'S genius. Like LAGRANGE, another
analyst, CAUCHY devotes special attention to the contact of curves
and surfaces, and gives a first geometrical definition of the contact
of two curves. He takes a point in common with two curves, draws
a small circle of radius i with this point as center, considers the
angle w intercepted on the circumference by the two curves,
and compares w to a power of i. As source of inspiration in
many details he mentions " les lumieres de M. M. AMPERE et
CORIOLIS." AMPERE's investigations on electric currents, indeed,
inspired CAUCHY'S discrimination between a right-handed and

(58) L. VALLAE, Trait6 de g6om6trie descriptive. Sec. ed. I825, P. 296-287.

(59) CAUCHY, Lecons sur les applications du calcul infinitesimal A la geom6trie.
Paris, I (I826), 400 P.; II (I828), I23 P. In this book the word " normale prin-
cipale " is used (I, p. 285, 298).

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
 120 D. J. STRUIK

a left-handed system of coordinate axes. Other innovations are

the use of polar coordinates for the computation of the radius
of curvature, the introduction of the " premi&re courbure
I1 I
-and the " seconde courbure " - for space curves, a systematic
r
treatment of space curves and the use of the " rayon vecteur."
It is entirely clear that CAUCHY used to the fullest extent all previous
sources, especially MONGE and DUPIN. (6o)

(to be continued.)
Mass. Institute of Technology February 1932
Cambridge, Mass, D. J. STRUIK.

(6o) Of a certain importance is also the work of TH. OLIVIER (I793-I853),

a lieutenant of artillery, who was called to Sweden to found a school after the
pattern of the Ecole Polytechnique, and later became professor of descriptive
geometry at Paris. He continued MONGE'S tradition in his collection of geometrical
models, and wrote several papers on differential geometry, e.g., a " M6moire
de g6ometrie," J7ourn. Ec. Polyt., cah. 24 (I835), p. 6I-9I, in which he studies
the circular helix having, at a point of a space curve, curvature and torsion equal
to the corresponding quantities of the curve.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:21 UTC
All use subject to https://about.jstor.org/terms
Outline of a History of Differential Geometry (II)
Author(s): D.J. Struik
Source: Isis , Nov., 1933, Vol. 20, No. 1 (Nov., 1933), pp. 161-191
Published by: The University of Chicago Press on behalf of The History of Science
Society

Stable URL: https://www.jstor.org/stable/224886

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/224886?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

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

The History of Science Society and The University of Chicago Press are collaborating with
JSTOR to digitize, preserve and extend access to Isis

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 Outline of a History of differential
Geometry (*)

(II)

6. - GAUSS

If the French drew the most consistent economic and political

consequences from their revolution, the Germans on the other
hand were more profoundly stimulated by the accompanying
intellectual upheaval. France had NAPOLEON, but Germany had
BEETHOVEN, KANT and GAUSS. The small town philosopher of
K6nigsberg and the small town mathematician of G6ttingen
represent as truly the essential aspects of the new era as the little
corporal himself.
Germany, after the death of LEIBNIZ, had participated only
in the construction of new mathematical ideas in so far as its
despots had imported genius to adorn their throne. The same
causes that bring the revolution to France revivify the German
intellect. Its clearest expression in mathematics is CARL FRIEDRICH
GAUSS (1777-I855).
For us the activity of GAUSS is threefold: as inventor of non-
Euclidean geometry, as inventor of intrinsic differential geometry,
and as a theoretical geodesist. This geodetical work, as the
practical one, was at the bottom of all his geometrical discoveries.
Most of them were made in the period between I815 and I830.
From i82I to '25 GAUSS was busy surveying the Kingdom of
Hanover.
We shall not discuss GAUSS' work on non-Euclidean geometry,
though it profoundly affected later differential geometry. It has
been done in considerable detail. We shall deal only very

(*) See Isis, I9, 92-I20.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i62 D. J. STRUIK

briefly with his geodetical papers. Our main subject has to be

GAUSS' direct intervention into the theory of surfaces.
GAUSS' mathematical investigation of the problem of map pro-
jection was written in i822 under the title " Allgemeine Aufl6sung
der Aufgabe die Theile einer gegebnen Fliche auf einer andren
gegebner. Flache so abzubilden dass die Abbildung dem Abgebil-
deten in den kleinsten Theilen ahnlich wird." (i) It is the general
theory of the conformal representation of arbitrary surfaces, and
it represents a continuation of papers by EULER and LAGRANGE.
As EULER and LAGRANGE, GAUSS uses " Gaussian coordinates,"
taking x, y, z, equal to functions of t and u, but where EULER
and LAGRANGE represent a surface only on a plane, GAUSS repre-
sents two arbitrary surfaces upon each other. The solution, which
uses complex functions, is applied to different cases, for instance,
that of mapping a rotation ellipsoid on a sphere.
Later GAUSS came back to this subject and sought that conformal
representation of the rotation ellipsoid on a sphere which is best
for certain purposes. (2)
This work led GAUSS to the general theory of surfaces. A
manuscript on this subject dates from i 825. He published a
revised paper on this theory in I827. (3)
These Disquisitiones generales circa superficies curvas of I827
mean an entire departure from the French methods. (4) A new
linie of approach is found in the investigation of the intrinsic
properties of surfaces, depending only on the linear element,
and not the properties of the surface as imbedded in Euclidean
three-space. Such intrinsic properties are unchanged by bending

(i) Printed in Astronomische Abh. I825. See GAUSS Werke IV, p. I89-2I6.
The paper was a solution to a problem posed by the Academy at Copenhagen
in I822.
(2) C. F. GAuss, Untersuchungen uiber hohere Geodisie, I, II, Gdttingen Abh.
2 (I844), 3 (I847), GAuss' Werke IV, p. 259-300, 30I-340.
(3) The older ms. has the title: " Neue allgemeine Untersuchungen uiber
die krummen Flatchen " (I825), Werke VIII, p. 408 (with remarks of STACKEL).
The " Disquisitiones " were published in Commentationes soc. reg. sc. G6ttingensis,
6 (i828), Werke IV, p. 2I7-258. English translations of both papers by
J. C. MOREHEAD and A. M. HILTEBEITEL. Princeton Univers. Library, I902,
P. I26, with bibliography of papers using GAUSS' method, and many notes.
(4) That GAUSS was acquainted with MONGE'S contributions to geometry is
shown by his favorable review of MONGE'S " Geometrie descriptive," Gott. gel.
Anzeiger i8I3, Werke IV, p. 359-360.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I63

the surface without tearing or stretching. GAUSS writes the linear

element

ds = V (Edp2 + 2Fdpdq + Gdq2%y

where p and q are " Gaussian " curvilinear coordinates on the

surface. The main theorem is that the quantity k R

a result from EULER'S paper of 1760, depends only on E, F,
G, and their first and second derivatives. GAUSS proves this
" theorema egregium " in more than one way, first by taking
the general linear element, and then by assuming ds2=dp2+Gdq2.
This involves a discussion of geodesics, and the discovery of
another outstanding theorem, the one dealing with the-sum of the
angles in a geodesic triangle. This theorem is one of the first
examples of so-called differential geometry in the large, and was
partly anticipated by the corresponding theorem of LEGENDRE
on the sum of the angles of a spherical triangle. (5) GAUSS
proves the theorem by a transformation of the surface integral
of curvature k, the " curvatura integra," into a line integral.
The paper ends with expansions in series of -different functions
on the surface, in a coordinate system ds2 - dp2 + G dq2.
These results constitute the essentials of the intrinsic theory
of surfaces, linear element, geodesics, Gaussian curvature, curvatura
integra, and geodesic triangle. Later investigations have not been
able to add many more essential elements; they have shown only
that with linear element and Gaussian curvature the theory is
complete. One simple conception seems to be lacking, that
of geodesic curvature, but in GAUSS' unpublished paper o
this also appears, as " Seitenkriimmung."
GAUSS makes much use of the spherical representations of a
surface, an idea he derived from astronomy, and indeed he defines
the curvature k as the limit of the quotient ot a small area of the
surface and the corresponding area of the unit sphere. This
raises, perhaps, the question of GAUSS' originality. Spherical
representation, geodesics, curvilinear coordinates are all to be
found in EULER-why admire GAUSS for their discovery? RODRI-

(5) A. M. LEGENDRE, Mem. div. sav. 1787, P. 358; also in the Appendix,
Section 5, to his trigonometry. LEGENDRE's theorem was known to G.Auss
Disquisitiones, Sec. 27.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I64 D. J. STRUIK

GUES even had found the ratio between the surface elements
of surface and unit sphere. But, though they were known to
EULER, nobody had sensed their importance before GAUSS, and
that after fifty years of productivity in France. And the light
that GAUSS throws on both old and new conceptions as intrinsic
properties of the surface was altogether new.
The invariance of the curvature under bending has been found
in papers of GAUSS dating back to i822. In these papers he uses
the linear element of his paper on map projection, ds2 = m2 (dt2 +
du2).
However, nothing in the Disquisitiones, as KLEIN remarks, (6)
shows that GAUSS. kept his boldest ideas to himself. His manu-
scripts show that he was in possession of the non-Euclidean
geometry when writing his paper on surfaces, but he did not
hint at it even remotely. But it throws a new and better light
upon his theorem on the sum of the angles of a geodesic triangle.
It was a theorem enabling the physicist to test the nature of
physical space. It was a step in the direction later consistently
followed by RIEMANN and HELMHOLTZ. GAUSS indeed measured
geodesical triangles, but did not find confirmation of his belief
that there might be a perceptible curvature (negative, as GAUSS
thought, in agreement with the later theories of LOBATSCHEWSKY
and BOLYAI) of space.
Of GAUSS' contribution to notation and nomenclature we mention
the symbols E, F, G, D, D', D" for what we now call the coefficients
of the first and second fundamental differential form, and the
word " conformal." (6a)
GAUSS became the teacher of the entire learned world, but
he created no direct school as MONGE had done. Conditions
in Germany were different from those in France. The only
exception, so far as differential geometry is concerned, was FERDI-
NAND MINDING (I806-I885) in far-away Dorpat, who filled the

(6) F. KLEIN, Vorlesungen uiber die Entwicklung d. Mathem. I, p. i6. Here

are excellent descriptions of the work of GAUSS.
(6a) In the first paper on higher geodesy, I844: " ich werde daher dieselben
conforme Abbildungen oder tlbertragungen nennen, in dem ich diesem sonst
vagen Beiworte eine mathematisch scharf bestimmte Bedeutung beilege," Werke
IV, p. 262. The word is, indeed, already used by F. T. SCHUBERT, " De pro-
jectione sphaeroidis ellipticae geographica," Nova Acta Petr., p. I30-I46, see
CANTOR IV, P. 575.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I65

gap left by GAUSS and published a paper on the geodesic cur-

vature (6b). Some years later MINDING published other papers
of outstanding importance, all showing a deep understanding
of GAUSS' ideas. In them he took up the " problem of MINDING,"
that is the question of the applicability of surfaces on each
other. One of his theorems states that two surfaces of constant
curvature are always applicable. (6c) He determined the rota-
tion surfaces of constant positive curvature and established
the existence of the helicoids of constant curvature. He also
remarked (as EULER had done) that a closed convex surface
cannot be bent.
JACOBI also grasped immediately the importance of GAUSS'
work in geometry and explained it in his lectures. C. G. J. JACOBI
(1804-i85I) was at that time at K6nigsberg. In a paper of I836
he generalized GAUSS' theorem on geodesic triangles on a sur-
face. (7) But JACOBI'S contributions to differential geometry
much more important than this. He beautifully combined
GAUSS' ideas with his own. Through his discoveries on
Abelian integrals he was able to integrate the geodesic lines
on an ellipsoid, which led to hyperelliptic integrals, a counterpart
to MONGE'S integration of the lines of curvature. (8) In the
calculus of variations he established the existence of the conjugate
points on the geodesics passing through a point on a surface.

(6b) F. MINDING, Bemerkung uiber die Abwicklung krummer Linien und

Flachen. Crelle 6 (I830), p. I59.
(6c) F. MINDING, Wie sich entscheiden lasst ob zwei gegebene krumme
Flachen auf einander abwickelbar sind oder nicht. Crelle i9 (i839), p. 370. In
Crelle, from I838-40, are several papers of MINDING on this subject, the remark
on convex surfaces in i8 (i838), p. 365-368.
Through MINDING, his colleague K. E. SENFF (I8IO-I849) and their pupil,
KARL PETERSON (i828-i88i), Dorpat figures as a minor center in the history of
differential geometry. SENFF found several results later discovered by SERRET.
PETERSON, who later became teacher at Moscow, wrote a book " Ueber Curven
und Flachen" (i868) with remarkable results, and in an unpublished " Kandi-
datenschrift" of i853 found BONNET'S theorem of i866 on a surface being
determined by first and second fundamental form and equations equivalent to
those of MAINARDI. See P. STACKEL, Bibl. mathematica (3) 2 (I90I), P. I22-I32.
(7) C. G. J. JACOBI, Demonstratio et amplificatio nova theorematis Gaussiani
de curvatura integra trianguli in data superficie e lineis brevissimis formati, Crelle
i6 (i836) p. 344-358; see also note io.
(8) C. G. J. JACOBI, C. R. 8 (i839), p. 284; Crelle i9 (I837), p. 309, Vorlesungen
uber Dynamik, p. ii6.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i66 D. J. STRUIK

On an ellipsoid these conjugate points form a curve with four

cusps. (g)
In this work JACOBI touches on differential geometry in the
large. Another example is his theorem that the spherical image
of the principal normals to a closed continuously curved space
curve divides the surface of a sphere into two equal parts, (Io)
a corollary of his generalization of GAUSS' theorem on geodesic
triangles.
JACOBI regularly gave lectures on the theory of curves and
surfaces, some of which still exist in manuscript. In K6ENIGS-
BERGER's book on JACOBI excerpts are given. They show a vivid
interest in the production of the French school and GAUSS, and
they present, with JACOBI'S own work, much of this material
in an original way.
An early pupil of GAUSS and JACOBI in Germany was F. JOA-
CHIMSTHAL (i8I6-i86I) to whose work we have to come back.
A pupil of JACOBI at K6nigsberg was HEINRICH FERDINAND
SCHERK (I798-I885), later teacher of mathematics at Kiel and
Bremen, who found a minimal surface, differing from the two
found by MEUSNIER more than fifty years earlier. (ii)
One of the reasons for the small number of workers on differential
geometry in those days was, of course, the great attraction exercised
by the great algebraic geometers, STEINER, M6BIUS, PLU4CKER,
CHASLES, PONCELET, and others. STEINER and M6BIUS occasionally
ventured into infinitesimals, and the result was always interesting.
For instance, STEINER found, in his original, thoroughly geometrical,
way, the properties of the curvature centroid of closed plane
curves (I838), and renewed in classical way the isoperimetrical
problem (I841). A brief paper of I839 deals with involutes
of space curves; it is related to JACOBI'S paper on space curves
of that year. (I2) MOBIUS, in his Barycentrische Calcil (I827)

(9) C. G. J. JACOBI, Posthumous paper: Ober die Curve, welche alle von
einem Punkte ausgehenden geodfitischen Linien eines Rotationsellipsoides beruhrt.
Ges. Werke VII, p. 72-80.
(Io) C. G. J. JACOBI, Ober einige merkwuirdige Curventheoreme, Astron.
Nachr. 20 (I842), p. 115-120. Ges. Werke VII, p. 34-39.
(I I) H. F. SCHERK, Bemerkungen uber die kleinste Flache innerhalb gegebener
Grenzen. Crelle I3, x835, p. I85; earlier in a Latin paper in Acta Soc. ablo-
novianae 4 (I830).
(12) J. STEINER, Einfache Beweise der isoperimetrischen Hauptsatze, Crelle

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I67

has three chapters on curves and surfaces, treated for the first
time in an invariant symbolism. His work leads to DUPIN'S
results on the indicatrix, and he claims to be an independent disco-
verer of its properties; it was only after he had his results, he
writes, that he discovered them in DUPIN's Developpements. (I3)
Accordingly, not only in Germany was there little immediate
response to GAUSS' work, but also in France, where, moreover,
the MONGE tradition was still very much alive. We may even feel
a certain polemical note in a paper written by SOPHIE GERMAIN,
(at that time already fifty-five), in Crelle I83I, recommending the

"9 mean curvature ( - +-- as the measure of curvature

of a surface instead of GAUSS' R The mean curvature is

R1 R2'
indeed of importance in elasticity, a field in which SOPHIE GERMAIN
had won her laurels.

7. - The French School of the Forties.

The new generation in France considered GAUSS and JACOBI

its leaders. GALOIS, in his last letter to a friend, expressed it
in this way: " Tu prieras publiquement JACOBI ou GAUSS de
donner leur avis non sur la verite, mais sur l'importance des theo-
remes." But there was also a second influence, working in
France, namely the development of physics in general and mathe-
matical physics in particular, - a direct result of the tremendous
changes in the European industrial system.
During the Restoration this development, begun during the
Empire, reached with FRESNEL and AMPERE a point where it
veritably revolutionized scientific thought. Optics, electricity,
elasticity, the theory of heat, and astronomy were affected. MALUS
and CHLADLI were among the first venturing into these unknown
regions; then followed POISSON, FOURIER, FRESNEL, ARAGO,

i8 (I836), P. 28i-296, Ges. Werke II, p. 75-9I. Von dem Krimmungsschwer-

puncte ebener Curven. Crelle 2I (I838), P. 33-63, IOI-I33, Ges. Werke II,
P. 97-I59. Ueber einige allgemeine Eigenschaften der Curven von doppelter
Krummung. Monatsber. Akad. Berlin, I839, P. 76-80, Ges. Werke II, p. I63-I65.
(I3) A. F. M6BIUS, Ges. Werke I, p. I-388. Differential geometry in Ch. VI.
VII, VIII. See Vorrede p. 8.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
I68 D. J. STRUIK

AMPiRE, SADI CARNOT, CLAPEYRON, and many others. With

NAVIER and SOPHIE GERMAIN, DE SAINT VENANT, and LAMA, the
field of theoretical elasticity is opened, - an event which was to
exert considerable influence on the application of analysis to geo-
metry.
For our purpose the most important figure of this group is
GABRIEL LAMA (I795-I870), a pupil of the Ecole Polytechnique,
who was sent, in the company of CLAPEYRON, to Russia to assist
in the improvement of methods of technical instruction. After
the July revolution of I830, he became professor at Paris. This
engineering work in the development of new railroads exhibits
clearly the close relationship which characterized the progress
of natural science and the rapidly growing French capitalism
of his time; compare, for instance, also the work of CRELLE on
Prussian railroads. In asking for solutions of the equations of heat
and elasticity in other than rectangular solids LAMA developed
a theory of rectangular curvilinear coordinates, in which the line
element of space is written
ds2 = H2 dp2 + H12 dp,2 + H22 dp22
and where the quantity
J (dF)2 (dF)2 ? (d)2

is what LAMA called the "parametre differentiel du premier

ordre " of the function F (x, y, z). There are also "parametres
diff6rentiels du second ordre "
d2F d2F d2F
dx2 dy2 dz2
LAMA'S six partial differential equations of the second order
connecting the H form the basis of his investigations, which
carry him deep into interesting analytical developments.
Here we find not only " coordonnees curvilignes " in spac
but also invariants under rotations and translations. This invariance
is indicated by the introduction of the symbols A11 F and A12 F for
first and second differential parameters of F (x, y, z). A single
infinity of surfaces F = const, for which A11 F - o, is called
"isotherme," the function F a " parametre thermometrique"
of the family of surfaces.
LAMA introduced these notions in a series of publications, the
first of which, in I837, on isothermic surfaces, received immediate

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I69

attention. It was followed in I840 by a paper on curvilinear

coordinates. (I4) After many years of teaching he collected the
fruits of his research in a series of textbooks, which are, even
now, very readable. (I5)
Another mathematician, who entered the field of differential
geometry through his work in applied mathematics, was
A. J. C. BARRA DE SAINT VENANT (I797-I886), well known as
an early contributor to the theory of elasticity. He wrote, in
I846, a paper which supplied the impetus to the final completion
of the elementary theory of space curves. (i6)
It was under such teachers that the younger mathematicians
developed. The leading figure in pure mathematics became
JOSEPH LIOUVILLE (I809-I882).
LIoUVILLE'S fame is the result, on the one hand, of his diversi-
fied mathematical investigations, and on the other, of his success
as the founder and editor of the " Journal de mathematiques
pures et appliquees." This Journal, first published in I836,
has published a great number of papers on geometry; its parallel
in Germany was CRELLE'S " Journal " of the same title. CRELLE
and LIOUVILLE'S Journals constituted the principal agencies for
the progress of differential geometry in the middle part of the
igth Century, and as such their importance is not merely of the
past.
Among the younger contributors to LIOUVILLE'S Journal were
J. A. SERRET (I8I9-I885), V. PuIsEux (I820-I883), 0. BONNET
(I8I9-I892), and J. BERTRAND (I822-I900), all of whom became
professors at Paris, together with F. FRENET (i8i6-i868), professor
at Lyons, and the Belgian E. CATALAN (I8I9-I894), who studied
at the Ecole Polytechnique and became professor at Liege. Their
work is based not only on that of LAMA and MONGE but also
of GAUSS and JACOBI. We shall try to summarize, without going

(14) G. LAME, Memoire sur les surfaces isothermes, Journ. de math. (I0)2 (I837),
p. 149; Memoire sur les coordonnees curvilignes, ib. 5 (I840), p. 313-347.
(I5) G. LAME, Le,ons sur la theorie mathematique de 1'dlectricit6 (I852);
Le,ons sur les fonctions inverses des transcendantes et les surfaces isothermes
(I857); Le,ons sur les coordonnees curvilignes et leur diverses applications (I859);
Le,ons sur la theorie analytique de la chaleur (i86i).
(i6) SAINT-VENANT, Mdmoire sur les lignes courbes non planes. J7ourn. Ec.
Polyt. cah. 30 (I846), p. I-76.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I70 D. J. STRUIK

too far into an estimation of individual merits, the results of this

school, which flourished from I840 to I850.
i) Work on Gauss' theorems. In many papers and by various
methods the theorem of GAUSS on the invariance of curvature and
the sum of the angles of a geodesic triangle is proved again (I7).
It is in this work that modern nomenclature evolves more and
more, as, R1 and R2, ds2 Edu2 + 2Fdudv + Gdv2, etc. BONNET
names the "courbure geodesique," (i8) and derives the " formula
of GAUSS-BONNET." LIOUVILLE introduces as a counterpart to
GAUSS' ds2 in geodesic polar coordinates the ds2 = A (du2 + dv2),
already studied by GAUSS in his paper on map projections, and
found in LAMA'S investigations on heat (isothermal element).
This leads LIOUVILLE to conformal representation and to " Liou-
VILLE's surfaces," for which the geodesics can be found by quadra-
tures. JACOBI's determination of the geodesics on an ellipsoid
is a special case. (I9) LIOUVILLE also determined all rotation
surfaces of constant curvature, completing MINDING's results.
2) Special problems in the Monge tradition. MONGE'S tradition
remained entirely alive. As an example of work accomplished
under its influence we mention several studies on surfaces with
lines of curvature, plane or spherical. JOACHIMSTHAL, the German
geometer, first at Halle, later at Breslau, shared in these investiga-
tions, (2o) and found, among other results, the " theorem of
JOACHIMSTHAL " on plane lines of curvature. In this connection
we may also mention ABEL TRANSON (I805-I876), who continued
the work of AMP'ERE and CARNOT and who belongs to the early
students of affine differential geometry.
3) The greater prominence of the notion of invariance. The
frequent use of curvilinear coordinates emphasizes GAUSS'

(I7) E.g. J. LIOUVILLE, Sur un theorbme de M. GAUSS concernant le produit

de deux rayons de courbure principaux en chaque point d'une surface. Journ.
de Math. I2 (I847), p. 291-304.
(i8) 0. BONNET, Memoire sur la theorie generale des surfaces. J7ourn. Ec.
Polyt. ig, cah. 32 (I848), p. 1-146.
(Ig) J. LIOUVILLE, Sur quelques cas particuliers otu les equations du mouvement
d'un point materiel peuvent s'integrer, _ourn. de Math. ii (I846), p. 315. De
la ligne geodesique sur un ellipsoide quelconque, _ourn. de Math. g (I844), p. 40I,
also appendix of the fifth ed. of MONGE'S " Applications."
(20) F. JOACHIMSTHAL, Demonstrationes theorematum ad superficies curvas
spectantium, Crelle 30 (I846), p. 347-350.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I71

standpoint on the invariance of quantities under coordinate trans-

formations. LAMA's differential parameters A11 F and z2 F are
a first effort to create a special notation for differential invariants
on the surface. This development is contemporary with that in
algebraic invariants, a field in which CAYLEY and ARONHOLD
had already started their investigations. These geometers,
however, did not arrive at the relationship between the GAUSS
coordinates on a surface and the LAMAl coordinates in space.
This could be done only after RIEMANN'S fundamental work
on the nature of space.
4) Development of the general theory of space curves. The theory
of space curves lacked both elegance and easy access. DE SAINT
VENANT, who wrote an extensive study on space curves in i 846, (i 6)
in which he collected the available material, added several new
theorems, introduced the word " binormale," and gave a good
historical review, concluding his paper with several pages of
formulas. This approach could be abandoned when F. FRENET,
in his Toulouse dissertation of I847, found the key to an easy
control of the computations in this field. (2i) His " FRENET
formulas " are the result of a belief that we should not only find
the derivatives with respect to the arc of the direction cosines of
the tangent, but also those of the principal normal (rectifying
plane) and the binormal (osculating plane). A few years later
J. A. SERRET, (22) ignorant of FRENET'S dissertation, arrived at
the same formulas; this led FRENET to a renewed publication of
his results in the same periodical, Liouville's Journal. Their
importance was not generally recognized very soon. The textbook
of F. JOACHIMSTHAL on differential geometry, which was the
result of lectures at Breslau during the winter term of I856-57,
does not give them, (23) not even the third edition of I890,
edited by L. NATANI. Neither do they appear in a book on space

(2I) F. FRENET, Sur les courbes a double courbure. These Toulouse, I847.
Abstract in Yourn. de Math. I7 (I852), p. 437.
(22) J. A. SERRET, Sur quelques formules relatives a la theorie des courbes
a double courbure. Journ. de Math. i6 (I851), p. I93.
(23) F. JOACHIMSTHAL, Anwendung der Differential- und Integralrechnung
auf die allgemeine Theorie der Fliichen und der Linien doppelter Krummung.
Leipzig I872. This excellent introduction, one of the first to appear in Germany,
gives a symposium of the results of the French school and those of GAUSS and
JACOBI.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I72 D. J. STRUIK

curves, written in i86o by another SERRET (PAUL) (I827-I898),

a professor at the Universite Catholique at Paris, who gave an
interesting exposition of the whole theory as it was known in I86o,
with special applications to spherical curves. (24)
5) Solution of special problems on space curves. PuIsEux proved
that the ordinary helix is the only curve for which curvature
and torsion are constant. (25) Then BERTRAND showed that the
helix on a general cylinder is the only curve for which the ratio
of curvature and torsion is a constant. SERRET found the equation
of the curves of constant curvature and that of curves of constant
torsion, and BERTRAND discovered the" BERTRAND curves," curves
for which a linear relation exists between curvature and torsion.
After SERRET had published his formulas, it was recognized that
these results could be obtained much more easily from the FRENET
formulas. PAUL SERRET's book of i86o has already been men-
tioned.
6) Original results. It would be entirely wrong to think of these
mathematicians only as pupils of great masters. As an example
of a new and surprising result we mention LIOUVILLE'S theorem,
that conformal transformations in space are only inversions,
similarity and congruency transformations. LIOUVILLE found this
theorem in an investigation of the conformal representation of
surfaces; he published it, together with many other results of his
own and his associates, in the fifth edition of MONGE's Applications.
LIOUVILLE also added a reprint of GAUSS' paper and as a result
we get from this fifth edition a fair idea of the progress of differentia
geometry to the fifties of the nineteenth century. (26)

(24) PAUL SERRET, ThAorie nouvelle g4om6trique et m6canique des lignes

a double courbure, I86o; also Th6se Paris, I859, without some notes. The
book also contains the theory of curves on surfaces, and has several new
results, as e.g. the theorem that asymptotic lines of a ruled surface are cut by
the generators in 4 points of constant anharmonic ratio (p. I69).
(25) V. PUISEUX, ProblWme de g6om6trie. Yourn. de Math. 7 (1842), p. 65.
(26) A textbook showing the regular class programs of those days in France
is C. F. A. LEROY, Analyse appliqu6e a la G6om6trie des trois dimensions. Paris,
Mallet Bachelier, 4th ed. I854, 408 p., of which almost one half is devoted
to differential geometry. It is almost exclusively based on the texts of MONGE
and DUPIN, influenced by the presentation of these texts by CAUCHY and POISSON
(Journ. Ec. Pol. XXI). The word " torsion " is used here (p. 292, 299). There
is no influence of GAUSS.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I73

8. - RIEMANN

Outside of France there was more interest than initiative. In

England the interest in differential geometry could only be awakened
after the fossilized epigones of NEWTON had been replaced by
younger men, acquainted with LEIBNIZ'S methods. This renais-
sance was initiated by CHARLES BABBAGE and his friends in i 8i i.
But scarcely any differential geometry was investigated till the
publication of the Cambridge Mathematical Journal (I837), which
soon became The Cambridge and Dublin Mathematical Journal
(I846), in which geometry and analysis were cultivated in the
continental style. Differential geometry, however, appeared mainly
as an aspect of the theory of conic sections and quadrics. Among
the authors in this journal we mention D. F. GREGORY, W.
WALTON, A. CAYLEY, and W. THOMSON. GREGORY (I8I3-I844),
who died young, left a well written Treatise on the application
of analysis to solid geometry, edited by WALTON (I845). It shows
the French influence even in its terminology. The importance
of GREGORY is shown by the fact that his friends published his
collected works. The centers of this new life were at Cambridge
and Dublin, and at Dublin we find the leading figure of W. R.
HAMILTON (I805-I865), who published between i828 and I837
his four papers on systems of rays, which carry on the work
of MONGE and MALUS and lead to his fundamental theorems on
partial differential equations. (27) He also presented curve and
surface theory in his quaternion lectures, published in I853 and
i866. (28) They contain a complete theory of curves and surfaces,
with many new results. He studies, for instance, space curves
up to elements of the fifth order with respect to the arc-length.
To his circle belong JOHN THOMAS GRAVES (I806-I870), a professor
of law at London, who studied mathematics at Dublin, and the
twin brothers MICHAEL (I8I7-I882) and WILLIAM ROBERTS (I8I7-
I883), professors of mathematics at Dublin, who contributed to
differential geometry. In the theory of confocal curves on an

(27) W. R. HAMILTON, On systems of rays, Trans. Roy. Irish Academy I828;

Supplements I, II, ib. I833; III, ib. I837.
(28) W. R. HAMILTON, Lectures on Quaternions, I853; Elements of Quaternions
I866.

I2

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I74 D. J. STRUIK

ellipsoid a theorem bears GRAVES' name. From this Dublin center

come also the famous textbooks of the theologian-mathematician
GEORGE SALMON (I819-I904), who deals to a considerable extent
with the application of analysis to geometry, as in Higher plane
curves (i852) and Analytic geometry of three dimensions (i862).
England, though responsible for little creative work in our field,
had, for a time, the better of the textbooks. In the forties we
find in Germany only the work of F. JOACHIMSTHAL, an investi-
gator who understood both the French and the German masters.
In Italy differential geometry found excellent teachers in BORDONI
and CHELINI, but it was their pupils who were to contribute
first class results. Even in France little work of outstanding
value is produced in the fifties. (29) A new impulse had to
come.
This new impulse comes again from Germany, and again from
G6ttingen. In I854, BERNARD RIEMANN (i826-i866) addresses
an audience in the little town as part of the requirements for
a " Dozentur." He talks " on the hypotheses which serve as
foundation to geometry." GAUSS was in his audience. The paper
was not printed till i867, (30) but its publication set in motion
an influence which has lasted till the present day.
The actual mathematical facts in RIEMANN'S address can be
stated in a few words. He develops the conception of a general
n-dimensional manifold in the sense of the analysis situs and
then introduces into it a quadratic linear element ds2. He describes
how the curvature of such a manifold can be measured, and
specializes manifolds of constant curvature. In this way both
Euclidean geometry and the non-Euclidean geometry of BOLYAY
are obtained as special cases of a general " Riemannian " geometry.
As these two cases are represented by a zero and negative constant
curvature, the question arises as to a manifold of positive constant
curvature, the " Riemannian" non-Euclidean spaces. The results
of GAUSS on intrinsic properties of surfaces covered RIEMANN'S
special case n=2.
But the meaning of RIEMANN'S work lies far deeper than the

(29) 0. BONNET is an outstanding exception.

(3Q) Uber die Hypothesen, welche der Geometrie zugrunde liegen, GMtt. Abh.
I3 (I867), Werke No. XIII, new edition by H. WEYL, Berlin, SPRINGER, 2 Aufl.,
I92I.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 175

purely mathematical structure which resulted from it. We best

see it perhaps when we compare MONGE and RIEMANN. MONGE
studies surfaces in ordinary space. This space is for him a given
entity, about which no discussion can exist. It is given and the
task of the geometer is the same as that of the carpenter: he takes
bodies in that space and works with them. There is a perfect
separation of surface and space, because there is such a separation
between rigid body and space. MONGE in this respect is the
representative of the metaphysical materialism of the eighteenth
century with its sharply defined conceptions.
RIEMANN does not only ask how rigid bodies behave in space,
but also how space is affected by rigid bodies. Space is not the
undiscussed given entity of MONGE, but rather an entity subject
to the same scientific investigation as are the bodies themselves.
Starting with RIEMANN, dialectical materialism begins to replace
metaphysical materialism in the foundations of geometry. Space
works on bodies, bodies influence space. There may be in the
first " Anschauung " an undefined conception of space, but
experience alone projects into it the properties which make it
the space of our geometry and of our physics, the projective rela-
tions, the metrical relations and even the relations of analysis
situs. From the abstract space of MONGE which relates bodies
at finite distance from each other, we pass to the " field " theory
of space as RIEMANN gives it. FARADAY and MAXWELL took
the same step for the field of electricity (as WEYL remarks) (3 I).
Also in other branches of science similar steps were taken at the
same time, opening the passage from the metaphysical classification
of eighteenth-century science to modern conceptions. The same
dialectical methods characterize LYELL'S work in geology, DARWIN'S
in biology, MARX' in sociology. From RIEMANN to EINSTEIN
is but one step-a step however that had to wait sixty years.
RIEMANN actually suggested, at the end of his paper, the
possibility of determining the metric of space by the physical
masses. This is the more striking as an example of almost
superhuman divination, as RIEMANN'S actual research in electricit
and gravitation followed the classical lines of his day.
RIEMANN wrote but few formulas in his address. He had a

(3I ) See note (3), " Vorwort des Herausgebers."

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
176 D. J. STRUIK

chance to work out some of his ideas mathematically in a paper

of i86i on the distribution of electricity on cylinders. (32) Here
he had to study the question of bringing the partial differential
equation
a a u au au'
ax (all ax, + a a2 BX + a u3 ) +
/ Sit Sit SU'
a- -S - + a22 + a23 +
5X2 - -- 23 3 BX3
a j3 axu + a32 + a33 u= a

aik functions of the x, into the simplest form. For a general

number n of xi this problem is equivalent to that of transforming
the quadratic differential form 2 aii, dxci dx. RIEMANN finds
the necessary and sufficient conditions that this form can be
reduced to a sum of n squares in the form of the vanishing of the
four index symbols.
(W ) i"i'iM) =Q.
Here, therefore, we have the principal concomitant of Riemannian
geometry.
It is, however, underestimating RIEMANN'S influence on differen-
tial geometry by referring only to his work on the space problem.
His work on function theory opened new ways of attacking geome-
trical problems, as is shown, for instance, by the later investigations
of SCHWARZ on minimal surfaces. RIEMANN'S introduction of the
connectivity of a manifold plays an important role in later differen-
tial geometry in the large.
Another study on n-dimensional geometry, from an entirely
different point of view, appeared at the same time, or better,
reappeared. That was GRASSMANN's Ausdehnungslehre. HER-
MANN GRASSMANN (i 809- I 877), who was a high school teacher
(Gymnasiallehrer) at Stettin, had written a first edition of this
work in i 844, which remained almost entirely unknown. He,
therefore, rewrote it entirely, and published it again in i862. (33)

(32) Commentatio mathematica qua respondere tentatur quaestioni ab Ill.

Academia Parisiensi propositae, I86I, Werke, p. 39I-423, only published
posthumously.
(33) H. GRAssmANN, Die lineale Ausdehnungslehre. Ein neuer Zweig der
Mathematik, I844. Ges. Schriften I (Leipzig, Teubner), p. 1-139. Die Aus-
dehnungslehre I862. Ges. Schriften II, p. I-5II.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFERENTIAL GEOMETRY I77
Here we find an admirable treatment of the elementary Euclidean
(and affine) n-dimensional geometry in the form of a (far from
elementary) formal calculus of points and vectors, as " Aus-
dehnungsgr'ossen," a consistent direct calculus invariant under
the affine group and partly under the rotational group. Not
only vectors are introduced, but also what we now call tensors.
This work has not only bearing on differential geometry in so
far as it inspired other geometers to extend their investigations
to more dimensions, but also through the direct application
GRASSMANN himself made of his apparatus to the problem of
PFAFF, through which this problem first took geometrical form.
RIEMANN certainly is the originator of more-dimensional differen-
tial geometry, GRASSMANN however of the symbolical methods
introduced for the study of this geometry. Here GRASSMANN'S
work is supplemented by that of HAMILTON on quaternions.
HAMILTON also introduces a symbolical method, of which the vector
and scalar product, and the nablaoperator, have been utilized
in the modern treatment of differential geometry. The invariant
character of differential geometry, clearly expressed by GAUSS
and the LIOUVILLE school, has its formal expression here, as
is immediately seen by the easy way in which LAME'S differential
parameters fit into the HAMILTON scheme. HAMILTON, however,
has no " Ausdehnungslehre."
RIEMANN and GRASSMANN together form a remarkable example
of the ways of development of human knowledge, where one
mind, essentially dialectical in nature, hews down barriers and
discovers new relationships, and the other, operating more formally,
builds up new symbolisms to control the field again through rigid,
even frozen, methods. This reveals again the greatness of
LEIBNIZ, who, in this respect at any rate, combined the merits
of RIEMANN and GRASSMANN.

9. - The beginning of modern times.

From now on the development of differential geometry does

not follow one main line. In the second part of the nineteenth
century different centers and different schools arise, in which
the influence of their leaders extends far beyond their immediate
location. Three sources of creative study draw immediate atten-

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I78 D. J. STRUIK

tion, though many different nations participate. These three sources

are France, Germany and Italy. Their paramount positions did
not arise accidentally. France with its long established traditions
*was the oldest center and continued uninterruptedly its economic
development under Bonapartism, and later under the Third
Republic. In Germany and Italy we see, at the same time, the
birth of the national state, completed in i 870, and, as a result,
an enthusiastic expansion of the capitalist system. The creative
influence of this development on mathematical thought is clearly
seen in Italy.
Here the national revival was known as the Risorgimento.
After a series of wars against Austria and its own despots, the
Italian people established the kingdom of Italy.
For a long time there had been much interest in differential
geometry, mainly in the North, at the University of Pavia, where
it was especially fostered . Here A. M. BORDONI (I789-I860)
taught for many years, and published, beginning in I82I, papers
on the applications of analysis to geometry, papers which show
the influence of GAUSS and the LIOUVILLE school. But his
important qualities lie mainly in his power to create interest.
As colleague he had G. MAINARDI (I800-I879), as pupils or
colleagues, D . CODAZZI (I824-I875), F. BRIOSCHI (I824-I897),
L. CREMONA (I830-I903), F. CASORATI (I835-I890), and above
all, E. BELTRAMI (I835-I900). Another early geometer of influence
was the priest D. CHELINI (I802-I878), who was professor of
mathematics and physics at different places, and who stimulated
interest through his publications (commencing I845) and his
teaching.
Productive activity rises to a higher level in the fifties, when
MAINARDI publishes a paper (34) in which he inquires as to
what relations must exist between the six functions E, F, G,
D, D', D" introduced by GAUSS into the theory of surfaces. It
is clear that there must be such relations, because the coordinate
representation of a surface shows that three functions determine
the surface. GAUSS found one relation, and the task is to find
others. This leads MAINARDI to four equations, in which we now

(34) G. MAINARDI, Su la teoria generale delle superficie, Giorn. Istit. Lombardo

9 (i856), P. 385-398.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY 179

recognize the two " Mainardi-Codazzi " equations. At the same

time the younger generation begins to publish, BRIOSCHI in
I852, CREMONA in I855, and CODAZZI in I858.
All these younger mathematicians were not only scholars, but
also organizers and prominent in political activities. They were,
accordingly, in a position to influence not only their pupils, but
also the instruction of the whole country and build up, in a relati-
vely short time, remarkable scientific enthusiasm. In geometry,
Italy became one of the leading countries.
In the sixties France, Germany and Italy begin to participate
equally in the development of differential geometry. A problem
set by the Paris Academy acted as a stimulus: to study the applica-
bility of surfaces upon each other (I86o). This problem, a conse-
quence of GAUSS' theory, had so far commanded the attention of
MINDING. and BONNET in France, who had framed the problem as
that of recognizing when two given surfaces are applicable, that is,
when they have the same ds2. Answers to the Paris question
came from BOUR and BONNET in France and from CODAZZI in Italy;
WEINGARTEN in Germany wrote a related paper. Here we find
the problem as that of finding all surfaces which are applicable
to a given surface. EDOUARD BOUR (I832-I866), after whom
this problem is called, and who got the first prize, was a young
mathematician at Paris, who became professor at the Ecole Poly-
technique; his paper, in the words of LIOUVILLE, " peut 6tre
pris comme un beau memoire de LAGRANGE." (3 5) BONNET,
whose work covers the whole period in which differential geometry
took its modern shape, wrote more than one important paper during
this period; in I867 he reached the fundamental result that a
surface is perfectly determined, but for its position in space,
by its first and second fundamental form, if these forms are
related by CODAZZI's equations. (36) CODAZZI'S contribution
to surface theory (37) was followed by other papers in which

(35) E BOUR, Theorie de la deformation des surfaces. Yourn. Ec. Polyt.,

22, cah. 39 (I862), p. 1-148.
(36) 0. BONNET, Memoire sur la theorie des surfacesapplicablessurunesurface
donnee, _'ourn. Ec. Polyt., 24 (cah. 41), i865, p. 209-230; 25 (cah. 42, I867
See on BONNET the necrology by P. APPELL. Comptes rendus I17 (I893) p.
IOI4-I024.

(37) D. CODAZZI, Memoire relatif a l'application des surfaces les unes sur les
autres. Mem. prds. par div. sav. 27 (I882), Nr. 6; it was written in I859.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i8o D. J. STRUIK

he went deeper into the nature of certain equations used by

him in his Paris answer, and he showed how they led to the
three " Mainardi-Codazzi equations." He did not know that
MAINARDI had reached similar results before. The problem of
the applicability of surfaces leads naturally to these equations.
BOUR gave them, but only for geodesic polar coordinates.
JULIUS WEINGARTEN (I836-I9IO), from I873 to I903 professor
at the Technische Hochschule of Berlin-Charlottenburg, developed
in his early papers the theory of the so-called W-surfaces, for which
a relation exists between the principal radii of curvature. As
the problem of finding all surfaces isometrical to a given rotation
surface can be reduced to that of finding all W-surfaces of the same
class, WEINGARTEN gave an example of a complete set of isometrical
surfaces that can be found only by elimination and quadrature.
Before that time only the developable surfaces were known. All
the later work of WEINGARTEN is related to this problem of the
applicability of surfaces. (38)
The important papers of this period are all of the type character-
ized by the names LAMA and GAUSS, and consciously or
unconsciously reflecting the ideas of RIEMANN. The BOUR-
MINDING type of problem set by GAUSS is indeed an investigation
into the nature of two-dimensional RIEMANN manifolds; the same
is true of the BONNET problem on the congruency of surfaces,
and the investigations of THEODORE MOUTARD (I827-I90I) on
infinitesimal isometry. Then we have the fundamental papers
of BELTRAMI, who, between the years I864 and i868, starting
from LAMA, led directly into the problems left open by RIEMANN.
Finally we have papers immediately due to RIEMANN, as HELM-
HOLTZ' inquiry into the foundations of geometry, and CHRISTOFFEL
and LIPSCHITZ' study of the transformation of quadratic forms.
Several German mathematicians whose main line of work lies

(38) D. CODAZZI, Sulle coordinate curvilinee d'una superficie e dello spazio.

I,1I,JII, AnnalidiMatem. (2) I (I867-68), p. 310,2(I868-69), p. 101-119, 26
" Codazzi-equations ", p. 273-274. See on this subject R. v. LILIENWFHAL, Encykl.
d. math. Wiss. III, 3, p. 159.
J. WEINGARTEN, Uber eine Klasse aufeinander abwickelbarer Flachen, Crelle
59 (i86i), p. 382.
See also J. WEINGARTEN, Ueber die Oberflachen, fur welche einer der beiden
Hauptkriimmungshalbmesser eine Function des andern ist, Crelle 62, p. I64;
Crelle 59, p. 382.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I8I

outside this field contribute in these years important papers to

differential geometry. KUMMER writes on line congruences,
WEIERSTRASS on minimal surfaces, and HELMHOLTZ on the axiom-
atics of geometry.
EUGENIO BELTRAMI is the most brilliant representative of the
group of Italian geometers who were trained during the Risorgi-
mento. He was, from i862 on, professor at Bologna, Pisa, Bologna,
Rome (I873-I877), Pavia, and Rome (I89I-I900). One of his
first papers was a translation of a work of GAUSS. During his
time at Pisa he often met RIEMANN, and his work shows the deep
influence of both GAUSS and RIEMANN. In an astonishing tempo
he published his fundamental research on differential geometry;
in I864, he produced his extension of LAMA'S differential parameters
to curvilinear coordinates with references to the applicability
problem; in I865, several beautiful theorems on the bending
of surfaces; in I867, the generalization of complex functions to
surfaces; and in i868, his theorems on non-Euclidean spaces, in
which he happily combined RIEMANN'S ideas with the older
notions of GAUSS and LOBATCHEVSKI. Here he showed that there
is a representation of non-Euclidean geometry on the pseudosphere,
so that there is no possibility of a contradiction in non-Euclidean
geometry, and he proved that in manifolds of constant RIEMANN
curvature the geodesics can be expressed by linear equations. (39)
SCHLAFLI (I8I4-I895), from I852-I89I professor at Bern, then
proved the inverse theorem. (40) This brought RIEMANN'S ideas
definitively out of the realm of speculation into the main body
of mathematical investigation, and it resulted in the combination
of the ideas of LAMA, GAUSS and RIEMANN into one solid theory.
The theory of quadratic forms of n variables was taken up
independently by CHRISTOFFEL and LIPSCHITZ in papers published
in I869 in the same periodical, CRELLE'S " Journal ". ERWIN

(39) Some of these papers are E. BELTRAMI: Ricerche di analisi applicata alla
geometria, Giorn. di mat. 2 (1864), 3 (I865); Sulla teoria generale delle superficie,
Atti Ist. Veneto (2I5 (I86o); Delle variabili complesse sopra una superficie
qualunque, Ann. di Mat. (2) i (1867); Teoria fondamentale degli spazi di cur-
vatura constante, Ann. di Mat. (2) 2 (I868-69); Saggio di interpretazione della
geometria non-euclidea, Giorn. di Mat. 4 (I868). All these papers in his
Opere," I, II.
(40) L. SCHLAFLI, Nota alla memoria del sig. BELTRAMI. Ann. di mat. (2)
5 (i87I-73), P. 178-I93.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i82 D. J. STRUIK

BRUNO CHRISTOFFEL (I829-I90I), at that time at Zurich, became,

after the Franco-Prussian war, professor at the new German
university of Strassburg, where he taught with REYE. He also
published some papers on related subjects, as on geodesic trian-
gles on a surface and the determination of a surface by its
mean curvature. RUDOLF LIPSCHITZ, (I832-1903) since I864
professor at Bonn, who is also well remembered in analysis,
continued his investigations on Riemannian manifolds during
the next years and built up a theory of these manifolds; since
Ricci's work these papers are somewhat antiquated. Both geo-
meters established theorems on the equivalence of quadratic
forms and derived the so-called Riemann-Christoffel tensor. To
this purpose CHRISTOFFEL constructed his CHRISTOFFELsymbols."
HERMANN HELMHOLTZ' (I82 1-I894) paper is Ueber die Tat-
sac/en, die der Geometrie zum Grunde liegen (i868). (4I) The
name already shows the influence of RIEMANN'S address of I854,
that was published a short time before HELMHOLTZ composed his
paper. He emphasizes the empirical aspects of our space concep-
tion, in the way RIEMANN had indicated. HELMHOLTZ, a physio-
logist, had been led to the same idea by his own investigations.
Experience, he says, gives us several measurable quantities,
continuous and of more dimensions, such as space, the system
of colors, and the field of vision. Each more-dimensional conti-
nuum is characterized by its own special properties, and our
task is to analyze these properties. HELMHOLTZ now gives four
postulates by which he can get RIEMANN'S geometry with the
quadratic line element; they are: i) the existence of n dimensions
and of continuity; 2) the existence of moving rigid bodies; 3)
the free movability of a rigid body; 4) no dependence of the form
of a rigid body on rotations (the monodromy). He then proves
as a purely mathematical theorem that RIEMANN'S geometry is
the only case in which these conditions are satisfied.
This paper by HELMHOLTZ (later corrected by LIE) was of
great importance for the final adaptation of non-Euclidean geo-

(4I) Gott. Nachrichten I868, Nr. 9; Wiss. Abhandlungen II, p. 618-639. It

was in I866 preceded by an address " Ueber die tatsaichlichen Grundlagen
der Geometrie," Wiss. Abh. II, p. 6IO-617. Later HELMHOLTZ defended his
thesis against the philosopher LAND: " Ueber den Ursprung und Sinn der geo-
metrischen Satze." Mind io (1878), P. 212-224, Wiss. Abh., II, p. 640-660.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I83

metry. But, like RIEMANN'S paper it was more than that. It

was one of the outstanding contributions to an understanding
of the nature of space, which was brought from the field of specula-
tion into that of experimental physics. As such it is one of the
outstanding contributions to a materialistic conception of space.
DARBOUX' work will be discussed in the next paragraph. To
this period belong also investigations of ALFRED ENNEPER (I830-
I885), from I859 Dozent, after I870 professor at G6ttingen.
He gave his name to a minimal surface and to a theorem on the
torsion of the asymptotic lines. The first papers by H. A. SCHWARZ
(i843-1921) on minimal surfaces appeared after I865; his final
solution of the isoperimetrical problem for the sphere dates from
I 884.
The ideas of LAME were taken up in a different way by French
geometers who now began to study triply-orthogonal systems.
LAMA, following DUPIN, had used only such systems as enabled
him to integrate his differential equations of physics. The
question of the construction of such systems was still unsolved.
BONNET, in a paper of i862, opened it anew; then follow A. RIBAU-
COUR (I845-1923), E. COMBESCURE (I824-I889), and DARBOUX
(thesis), (42) who showed respectively the relation of these systems
to circle congruences, the way in which such systems can be
obtained by a transformation of a surface into other surfaces
with parallel tangent planes, and their dependence on a differential
equation of the third order.
To the work done in the older line.we mention papers written
by Louis AOUST, a priest, " chanolne de Montpellier" (I814-
I885), who was professor of mathematics at Marseilles since I854
and contributed largely, in papers and textbooks, especially between
i86o and i88o, to the theory of curves and surfaces; in these
his " courbure inclinee " plays a large, too large, a role.
REINHOLD HOPPE (I8I6-I900), in I859 Dozent, and after I870
professor at Berlin, also published many papers, most of them
antiquated now; several however are of importance as first ventures
into the differential geometry of n dimensions. His work on
intrinsic coordinates of curves deserves mention.

(42) See the discussion by E. SALKOWSKI. Encykl. d. math. Wiss. III, 3, p. 541I.
See for the work done in Fiance J. BERTRAND. Rapport sur les progrbs les plus
recents de l'analyse mathematique. Paris, I867.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 I84 D. J. STRUIK

Several textbooks appear at this time. We already have men-

tioned PAUL SERRET'S book. In i868 appeared PETERSON'S mono-
graph UIber Kurven und Flachen, in which the Moscow mathe-
matician published several interesting results on applicability of
surfaces (see footnote 6c). In 0. BOKLEN'S Analytische Geometrie
des Raumes of i86i is a discussion of infinitesimal geometry.
WILHELM SCHELL'S (I826-1904) Allgemeine Theorie der Kurven
doppelter Krummung of I859, was written under the influence of
JACOBI, whose lectures on differential geometry he had heard in
the winter of i849-50. Some of these books were more mono-
graphs on certain subjects than complete textbooks. A good com-
plete textbook on differential geometry had not yet been written.

I0 - Differential geometry from i 870 to 1900.

From i870 to the world war there were no important economic

disturbances in Western Europe. Life at the expanding and
flourishing universities went on very regularly, and mathematicians
could devote their full attention to professional problems. A
result of this long period of quiet development was a great progress
in differential geometry, mainly along paths blazed by the previous
generations. At the same time geometry became more and more
ain abstract science, which did no longer easily reveal its origin
in practical problems. Most geometers of this period adopted
the attitude of JACOBI that science exists for the glory of the hum
mind and that its primary importance lies outside the field of the
applications. To this attitude we owe the development of differen-
tial geometry along lines of abstract beauty often comparable to that
of the most unpractical of all sciences, the theory of numbers.
France, Germany and Italy remained the chief centers, where at the
universities differential geometry was taught and professed as
a regular part of the schedule. Other countries participated in
the development, sometimes in a very important way, as Norway
with S. LIE, Sweden with A. BACKLUND, and Russia with P. L.
TSCHEBYCHEF. By the end of the century the United States begins
to exert some influence. It is impossible to give a discussion of all
the different subjects taken into consideration, of the theories
developed, and of the papers published during this period. This
would, moreover, repeat in an unsatisfactory way what volume 33

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I85

of the Encyklopddie der mathematischen Wissenschaften in a series

of monographs has done very effectively. We shall only point
out some very general trends.
At the opening of the period we have the early work of DARBOUX,
KLEIN and LIE, influenced by C. JORDAN and the so-called Erlangen
Program of KLEIN (I872). (43) Here the consequence is drawn
from the geometrical investigations of the nineteenth century,
and the group concept is found to be the underlying idea. A
geometry is defined as the theory of invariants belonging to a
given continuous group of transformations. Also in differential
geometry this idea of transformation comes more and more into
prominence. It finds its nucleus in EULER, and later in the work
of MONGE, CARNOT and AMPiRE, it arises again in the work
of GAUSS, LAMA and RIEMANN, and finally in that of BELTRAMI.
The Erlangen Program became the program indeed of almost
all further work on geometry in the nineteenth century, revealing
itself not only in opening of new fields of investigations, but also
in the introduction of formal invariant methods into the exposition
of the material.
Nevertheless, the Erlangen Program did not exhaust the field
of geometry as a whole, and so of differential geomnetry. There
have always been tendencies leading outside of the vigorous
scheme by which KLEIN so successfully tried to lead the develop-
ment of geometry. RIEMANN'S manifold conception was broader,
and Riemannian geometry can only fit into a wider frame than
that of the Erlangen Program.
The three great geometers whose work dominates the period
after I 870 are GASTON DARBOUX, SOPHUS LIE and their younger
contemporary, LUIGI BIANCHI. LIE'S work consciously lies in the
school of the Erlangen Program, for which he is partly responsible.
BIANCHI and DARBOUX, though their work belongs to the type
considered by KLEIN, did not find their field of work in problems
especially suggested by it. The conscious cultivation of the differ-
ential geometry of different continuous groups belongs principally
to the twentieth century.
G. DARBOUX'S (I824-1917) career is analogous to that of manv

(43) V. KLEIN, Vergleichende Betrachtungen uber neuere geometrische Fo

schungen, Programm Erlangen I872, Math. Annalen, 43 (I893), p. 63-IOO.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i86 D. J. STRUIK

French geometers of the nineteenth century, his life being that

of scholar and teacher at Paris. A remarkable fact is that for
his education he was one of the first to prefer the Ecole Normale
to the Ecole Polytechnique. In his later years, as secr6taire
perpetuel of the Academy he exercised influence far beyond
the limits of his special field of work. His work, even more than
that of L. BIANCHI (I856-1928), professor of mathematics at
Pisa and pupil of BETTI and DINI, follows the classical lines and
shows the enormous variety of problems left unsolved by MONGE,
LAME and others. DARBOUX'S early work (beginning i866) deals
with orthogonal sets of surfaces, each family of surfaces necessarily
depending on a partial differential equation of the third order.
'I'hen he passes to a study of the deformation of surfaces, to the
applicability of surfaces, and to surfaces of constant curvature. BIAN-
CHI's productive career begins with the applicability of surfaces
(i878) and in rapid succession follow papers on surfaces of constant
curvature, orthogonal surfaces, Weingarten surfaces, minimal
surfaces, and congruences. BIANCHI also turns to non-Euclidean
geometry, taking up the ideas of BELTRAMI; the surfaces of cur
vature zero in such geometry draw his special attention. Following
BXCKLUND he invents transformations to pass from one set of
surfaces of special character to another. In the work of DARBOUX
and BIANCHI partial differential equations play an important role.
As a new aspect of the theory of curves on surfaces, DARBOUX
introduces the coordinates x, y, z, of a surface as solutions of
an equation
820 86 86
8 af+ A , + B - \8- 8 4- CO4 o, A, B, C functions of a,

a=-const, ,=-const appear then as conjugate parameter lines.

In DARBOUX'S work another characteristic appears, the " triedre
mobile," by which kinematical considerations are introduced into
the study of curves and surfaces. This field, opened by CODAZZI,
was especially cultivated by A. MANNHEIM (I83I-I906).
The principal fame of DARBOUX and BIANCHI lies in their
beautiful textbooks, in which they combined their own results
with those of their predecessors. BIANCHI'S Lezioni di geometria
differenziale (I893), a new edition of autographed lectures published
in i886, is a systematic treatise of the theory of curves and sur-
faces, with special attention paid to more-dimensional geometry.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I87

It became known in wider circles through the German translation

of M. LUKAT (I899). DARBOUX'S Lefons sur la thdorie gdndrale
des surfaces (four volumes, I887, I889, I894, I896) is much more
than the title professes; it is an exposition of a vast number of
theories related to the theory of surfaces, connected in a loose
way, so as to give the author an opportunity to demonstrate the
enormous breadth of his knowledge and the elegance of his
methods.
For SOPHUS LIE (I842-I899) the transformation group is not
an important means of investigation alone; it is rather the central
part of geometry. His work is closely connected with that of
KLEIN. He was a clergyman's son from a village in Norway, and
he visited Paris in I869-70 together with KLEIN, where he ex-
changed ideas with C. JORDAN and DARBOUX. JORDAN, at that
time, had taken up GALoIs' result and published the Traitd des
substitutions. During their travel the fundamental importance of
the group became clear to KLEIN and LIE. A first result was ajoint
paper on W-curves, plane curves invariant under a projective trans-
formation of the plane. The Erlangen Program cast their ideas into
permanent form. LIE then started out on his life work, the study
and classification of continuous transformation groups, an inquiry
which at the same time opened a new road to the study of partial
differential equations. Regularly he reached results of importance
to differential geometry. We already had occasion to compare
his combination of geometrical intuition with analytical skill to
that of MONGE. In his hands more-dimensional considerations
grew into as natural a part of geometry as the classical methods.
Of his special contributions we may mention the duality between
line and sphere geometry leading to the duality between asymptotic
curves and lines of curvature, the integration of the FRENET
formulas if a relation between curvature, torsion and arc length
is given, his work on minimal surfaces and contact transformations.
He improved HELMHOLTZ' analysis of RIEMANN'S problem of space.
Of his older pupils we mention F. ENGEL and W. KILLING.
More-dimensional differential geometry found also investigators
following the directives of RIEMANN, CHRISTOFFEL and BELTRAMI.
Through a series of investigations undertaken by many authors
a complete theory of curves and higher manifolds was developed.
This study was considerably simplified by the invention of the

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 i88 D. J. STRUIK

" calculo differenziale assoluto," a symbolism to express the

invariants of Riemannian geometry. The inventor was GREGORIO
RICCI-CURBASTRO (I853-1924), as BIANCHI pupil of BETTI and
DINI at Pisa, since i88o professor at Padua. Ricci published his
discoveries in a series of papers beginning in i886.
An entire new field is represented by the work of HENRI POIN-
CARE' (I854-I9I2). In a series of papers under the title: Memoire
sur les courbes definies par une equation differentielle, published
in the Yournal de mathe'matiques from i88i-i886, the great physicist
astronomer and analyst undertook the study of the properties
of integral curves of ordinary differential equations, their singular
points and their behavior in the large. (44) Again differential
geometry was enlarged by influences from outside fields, as in the
times of GAUSS and of LAMA. POINCARI was led to his theory
by his investigation on problems similar to the three-body problem
and related questions of dynamics, where geodesics appear as
trajectories of moving particles. Here appear singularities like
the " nceud," the " col," the " centre ", and the " foyer," and
relations connecting them. Analysis situs becomes definitely
connected with differential geometry. Differential geometry in
the large, so far appearing only in isolated remarks of EULER,
GAUSS, JACOBI, MINDING, turns into an important and difficult
science of its own. The question of surfaces with closed lines of a
certain kind arises.
The main tendencies of the years I870-I900 can be summarized
as followvs
a) consequent study of continuous transformation groups;
b) triply orthogonal systems;
c) surfaces of constant curvature;
d) applicability and deformation of surfaces;
e) renewed study of partial differential equations and their
geometrical interpretation.
f) integral curves of ordinary differential equations;
g) more-dimensional geometry;.
h) adaptation of invariant methods, as vector analysis or absolute
differential calculus;
i) conception of comprehensive textbooks related to the general

(44) See the discussion by H. LIEBMANN, Encykl. d. math. Wiss., III, 3, p. 503.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I89

acceptance of differential geometry in the mathematical programs

of the universities.
There are, of course, many aspects to this rich development
that we have failed to emphasize or even to mention. The work
of E. CESARO (I859-I906), professor at Naples, on natural geo-
metry, resulting in a well-known textbook (I896) belongs to these.
It may be characterized as one of the attempts to introduce invariant
methods. Other mathematicians began to introduce vector
methods into differential geometry, as G. PEANO (I887), or
C. BURALI FORTI (I897); other attempts, like that of J. KNOBLAUCH
(i888) may also be classified under this heading.
A regular output of textbooks showed how general differential
geometry was taught. Almost all the " Cours d'analyses " and
corresponding books in other languages devoted chapters to
curves and surfaces. Since the seventies special books on the subject
appeared regularly, from the books by the Abbe AOUST to those
of L. RAFFY (I 897), G. RIccI (I 898) and W. DE TANNENBERG
(i899). By the end of the century differential geometry had
begun to break up into special branches, each of which had its
spccialists.
This leads us to the beginning of the twentieth century, when
other tendencies begin to exert influence. We shall not deal
with them in this present outline.

Ii. - Sources.

The material, necessary for a history of differential geometry,

is scattered over many publications, and so far as we know has
never been reviewed as an historical whole. Up to i8oo we have
V. KOMMERELL'S excellent paper on Raumkurven und Fliachen
in the fourth volume of M. CANTOR'S Vorlesungen iiber die Geschichte
der Mathematik (I908). After i8oo we have F. KLEIN'S monu-
mental book on Entwicklung der Mathematik im I9. Jahrhunde
(1926), but this deals with only a few aspects of differential geo-
metry. These are the only larger expositions on the subject.
Of importance also are A. HAAS, Versuch einer Darstellung der
Geschichte des Kriimmungsmasses (Diss. Tilbingen, i88i), not always
correct, and S. A. CHRISTENSEN, Om den historiske udwikling af
theorien for fladers og rumkurvers krumning (Tidskrift f. Math., i,

I3

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 I90 D. J. STRUIK

I883, p. 97-127). Then there are several monographs, as

P. STXCKEL, Bemerkungen zur Geschichte der geodiitischen Linien,"
(Leipz. Berichte 45, I893, p. 444-467), (45) and many biographies,
of which the "Eloges " of ARAGO, BERTRAND and DARBOUX
contain fascinating details, and are both valuable and charming.

I2. - Final remarks.

From time to time we have been able to refer, in a few words,

to the influence of external factors in the development of differential
geometry. There is a general tendency here.
Nobody seems more free than the mathematician in the selection
of his problems; yet even he in this respect has no complete
liberty. He generally follows the directives of a certain school
of thought, opened by a leader in the field, and is guided in his
work by established traditions. The leader, however, is himself
under certain influences in the selection of his material, and
the direction of his labor is, as a rule, determined by external
factors. In all cases his topics of research are in some way or an-
other connected with the material conditions under which his
generation lives, either directly by the necessity of the mathematical
treatment of a technical problem, or indirectly by certain philo-
sophical principles estimated as valuable in his time. Even when
freedom of selection seems the very nature of his mathematical
work, it is material circumstances which allow the required amount
of abstraction from daily needs, as in the later part of the nineteenth
century under the " science for science's sake "attitude prevalent
at many universities, and reaction is bound to follow closely. We,
therefore, find a subtle, but nevertheless definite, relationship
between the general economic problems that humanity has had
to solve and the investigations of the mathematician. It is, therefore,
possible to indicate these connections even in the case of one of
the most abstract fields of mathematics, the application of analysis
to geometry. Though this field is too large to admit more than
a very insufficient treatment in a few pages, we were nevertheless

(45) Also a master's thesis at the Mass. Inst. of Technology by J. L. LAWSON.

The history of the theory of curves and surfaces as contained in papers presented
before the French Academy of Science during the entire eighteenth century,
33 P., I93'.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms
 OUTLINE OF A HISTORY OF DIFFERENTIAL GEOMETRY I9I

able to give some indications of how the different schools of geo-

meters have been influenced by the technical requirements of
their age.
Differential geometry owes its results to such problems as are
set by map projection, surveying, measurement of time (HUYGENS),
fortification and other problems of warfare (MONGE); these are
supplemented by problems taken from potential theory, elasticity,
light and vision. Its very origin is due to the invention of the
calculus, a result of a long series of involved technical problems
in a period when geometry was still the main form of mathematics.
The French revolution opened enormous sources of energy for
its promotion, and influenced that renewed study of space and
time connected with the names of KANT, GAUSS and LOBATCHEV-
SKY. The steady penetration of materialistic ideas into the study
of separate branches of science in the course of the nineteenth
century resulted in the realistic analysis of the space problem
by RIEMANN and HELMHOLTZ, and removed the previously existirig
barriers to the study of n-dimensional geometry. For the organ-
ization and further cultivation of differential geometry the
nationalistic revival of the sixties was of great importance, as is
shown, for example, by the Risorgimento in Italy. And the
unhampered development of science in the latter part of the last
century, which resulted in a steady growth of our knowledge of
curves and surfaces, also finds its source in the undisturbed material
conditions of Europe during that time.
These are a few of the tendencies that have been responsible
for the development of one of the most fascinating branches of
modern mathematics. In the preceding exposition we had to
make a selection, which was to a certain extent arbitrary, and
which depended on our personal preferences. Nevertheless, we
hope we have followed what most mathematicians will consider
the main lines of development.
MIassachusetts Institute of Technology February I932.
Cambridge, Mass. D. J. STRUIK.

This content downloaded from

143.107.128.45 on Sat, 25 Mar 2023 09:40:23 UTC
All use subject to https://about.jstor.org/terms