# TOPOLOGY

FROM THE
DIFFERENTIABLE VIEWPOINT
By John W . Milnor
Princeton University

Based on notes by David W . Weaver

The University Press of Virginia Charlottesville

PREFACE

THESE lectures were delivered at the University of Virginia in December 1963 under the sponsorship of the Page-Barbour Lecture Foundation. They present some topics from the beginnings of topology, centering about L. E. J. Brouwer’s definition, in 1912, of the degree of a mapping. The methods used, however, are those of differential topology, rather than the combinatorial methods of Brouwer. The concept of regular value and the theorem of Sard and Brown, which asserts that every smooth mapping has regular values, play a central role. To simplify the presentation, all manifolds are taken to be infinitely differentiable and to be explicitly embedded in euclidean space. A small amount of point-set topology and of real variable theory is taken for granted. I would like here to express my gratitude to David Weaver, whose untimely death has saddened us all. His excellent set of notes made this manuscript possible.

J. W. M.
Princeton, New Jersey March 1965

vii

CONTENTS

Preface
1. Smooth manifolds and smooth maps Tangent spaces and derivatives Regular values The fundamental theorem of algebra

v11
1 2 7 8

2. The theorem of Sard and Brown Manifolds with boundary The Brouwer fixed point theorem
3. Proof of Sard’s theorem
4. The degree modulo 2 of a mapping Smooth homotopy and smooth isotopy

10 12 13
16

I

20 20 26 27 32
42 50

J

5 . Oriented manifolds The Brouwer degree
6. Vector fields and the Euler number 7. Framed cobordism; the Pontryagin construction The Hopf theorem
8. Exercises

52
55
59

Appendix: Classifying l-manifolds Bibliography Index

63

ix

Any particular diffeomorphism g : U --f W A Af is called a parametrixation of the region W A M . (The inverse diffeomorphism W A AI + U is called a system of coordinates on W A M . The identity map of any set X is automatically smooth. More generally let X C Rk and Y C R' be arbitrary subsets of euclidean spaces. . . want to look a t completely arbitrary sets X .+ V ) is called smooth if all of the partial derivatives ay/ax.Q1. If f : X + Y and g : Y -+ 2 are smooth. --f DEFINITION. -- ---f DEFINITION. xk) of real numbers. We can now indicate roughly what diferential topology is about by saying that it studies those properties of a set X C Rhwhich are invariant under diffeomorphism. The following definition singles out a particularly attractive and useful class.. however. SMOOTH MANIFOLDS AND SMOOTH MAPS FIRST let us explain some of our terms. A map f : X Y is called a di\$eomorphisnz if f carries X homeomorphically onto Y and if both f and f-' are smooth. ax.. A map f : X Y is called smooth if for each x E X there exist an open set U C Rk containing x and a smooth mapping F : U --f R 1 that coincides with f throughout U A X . thus a point x E Rk is an k-tuple x = (xi. . A mapping f from U to V (written f : U . Rk denotes the k-dimensional euclidean space. A subset ilf C Rkis called a smooth inanifold of dimension m if each x E Af has a neighborhood W A M that is diffeomorphic to an open subset U of the euclidean space R". exist and are continuous. note that the composition g o f : X + 2 is also smooth. ) . Let U C Rk and V C R' be open sets. We do not. .

y ) E R2 with x # 0 and y = sin(l/x).Y . y . z. then T M . Y ) + (x. is defined to be the entire vector space Rk. if U C U' are open sets and i:U--+U' I): .) Similarly one thinks of the nonhomogeneous linear mapping from the tangent hyperplane a t x to the tangent hyperplane a t y which best approximates f .1.(h) = Sometimes we will need to look a t manifolds of dimension zero. y . (Compare Figures 1 and 2.Q 1. Parametrization of a region in IIf origin that is parallel to this. to every commutative triangle U w 9 Of r/\ r v \ TANGENT SPACES AND DERIVATIVES To define the notion of derivative df. and z < 0. x. are called tangent vectors to llil a t x. Smooth manifolds Tangent spaces 3 1 Rm t Figure 1 . evaluated a t 2. of first partial derivatives. we first associate with each x E M C Rk a linear subspace T M . If I i s the identity map of U . A somewhat wilder example of a smooth manifold is given by the set of all ( x . x < 0. (In fact d f . for a smooth map f : M + N of smooth manifolds.) with x: = 1 is a smooth manifold of dimension n . i s the identity map of R k . I n fact the diff eomorphism + + + for x E U . The unit sphere S2.. to TN. Rnkthere corresponds a commutative triangle of linear maps R f Rd 2 49 R Y c f)= 2. For example SoC R' is a manifold consisting of just two points. Then d f . lim ( f ( x t-0 + th) .Clearly df. Intuitively one thinks of the m-dimensional hyperplane in Rk which 1 4 near x. f(x) = y. By definition.consisting of all ( 2 . we obtain similar parametrizations of the regions x > 0. and changing the signs of the variables. with (x.Y".f(x))/t EXAMPLES. For any open set U C Rkthe tangent space T U . parametrizes the region z > 0 of S2. M is a manifold of dimension zero if each x E M has a neighborhood W n M consisting of x alone. More generally the sphere Sn-' C R" consisting of all (xl. I f f : U + V and g : V + W are smooth maps. y < 0. one obtains dfz.it follows that S2 is a smooth manifold. z ) E R3 with x2 y2 z2 = 1 is a smooth manifold of dimension 2. for x2 y2 < 1. d 1 .x2 . 0 dfz. then d(g 0 f& = dg. we must study the special case of mappings between open sets. is just that linear mapping which corresponds to the 1 X k matrix (af J a x i > . h E Rk.) Here are two fundamental properties of the derivative operation: 1 (Chain rule). Elements of the vector space T M . Translating both hyperplanes to the origin. Before giving the actual definition. will be a linear mapping from T M . where y = f ( x ) . . R'.. Since these cover S2.By iritcrchanging the roles of x. then d I . For any smooth map f : U + V the derivatiue dfz . .(h) is a linear function of h. y > 0. I n other words. More generally. Rkof dimension m called the tangent space of dd a t x. is the hyperplane through the best approximates 1 of smooth maps between open subsets of Rk. Rk+ R' is defined by the formula df.

It follows immediately that d F .(v) for all v e TM. : TAT. Replacing U by a smaller set if necessary. we may assume that g ( U ) C W and that f maps g ( U ) into h(V). v = h-'(y). I f f :M N i s a difleonzoyphism.. C Thrz with irrclusion m u p di. carries TAT. then d(g 0 f>r = dgu 0 dfz. Furthermore the resulting map df. we obtain a commutative diagram of linear mappings RE dgu dFZ >R1 dha. -+ T N . for we can obtain the same linear * This restriction will be removed in \$2. the derivative operation has two fundamental properties: 1.+ N be a smooth map between manifolds of the same dimension. : T M . (Compare Figure 2. R m d(h-' T 0 f 0 g > u >R" REGULAR VALUES Let f : ill . = Image (dg. if M C N with inclusion m a p i.) into TN. More generally..* We say that x e n/r is a regular point of j if the derivative where U = g-'(x). If I i s the identity m a p of M . It follows that h-' o f o g : U -+ V is a well-defined smooth mapping.) is defined as follows. De&e df. 7 of smooth mappings between open sets. then T M . Taking derivatiws. Consider the commutative diagram 9 f P h-1 Figure 2. In payticular the dimension of ill must be equal to the dimension of N . -+ T N .6 51. 2. does not depend on the particular choice of F . i s an isomorphism of vector spaces.(v) belongs to T N . . 3 TN. That is: f:M+N with f(x) = df. y. (Chain rule). = Image (dh. As before.+ N C R ' for neighborhoods g ( U ) of x and h ( V ) of y. *4s before. 0 d(h-' This completes the proof that 0 f 0 g)u 0 (dgJ*. these two properties lead to the following: ASSERTION.).at' ions g:U+MCRE and h : V . is a well-defined linear mapping. then d I z i s the identity m a p of TM.. Since J is smooth there exist an open set W containing x and a smooth map F:W4R1 that coincides with f on W (7 M. and that it does not depend on the particular choice of F .(v) to be equal to dF. If f : M + N and g : N --+ P are smooth. Choose parametri~ . with f(s)= y.. df. The derivative dfz : T M . To justify this definition we must prove that dF. 7'he tangent space of a submanifold fo U gJh V The proofs are straightforward. Smooth maniJolds Regular values 7 smooth map transformation by going around the bottom of the diagram. = dh. then df.

If df. and there are only finitely many zeros since P' is not identically zero. then a short V = V. * . To see this we introduce the stereographic projection h. Thus each y E N is either a critical value or a regular value according as f-'(y) does or does not contain a critical point. 0. 0. 0 . x k be the points of f . Hence the locally constant function #f-'(y) must actually be constant on this set.-i jzl-'.. - . Since #f-'(y) can't be zero everywhere. Observe that i f M i s compact and ?J E N i s a regular value. and the polynomial P must have a zero.\$ alzn-l computation shows that &(z) = = z / 1 ~ 1= ~ -- 1/Z. . Next observe that f has only a finite number of critical points.. 0. V .' ( y ) . then f-'(y) is afinite set (possibly empty). For f-'(y) is in any case compact. 0 . -1) and set &(z) = h-fhI'(z). 1)) -+ R2 X 0 C R3 from the "north pole" (0.I z"/(a0 + d. THE FUNDAMENTAL THEOREM OF ALGEBRA As an application of these notions.8 01. and it follows that f = hI'Qhis smooth in a neighborhood of (0. by elementary geometry. The set of regular values of f .A V. even in a neighbor- . 1) It is well known that this resulting map f is smooth. For a smooth f : M --\$ N .((0.. .f(M . in N . where f(x) f ( 0 .U. Consider the unit sphere X2 C R3 and the stereographic projection h+ : Sz. since f is one-one in a neighborhood of each x P f-'(y). Smooth manifolds Fundamental theorem of algebra 9 df. - Kotc. 0. space M . . we prove the fundamental theorem of algebra: every nonconstant complex polynomial P(z) must have a zero. we conclude that it is zero nowhere. and choose pairwise disjoint neighborhoods U. .hI'(z) S o w if P(z) = a. 1 ) . 0. . is singular. I..from the south pole (0. The point y E N is called a regular value if f-'(y) contains only regular points. + + . In this case it follows from the inverse function theorem that f maps a neighborhood of x in M diffeomorphically onto an open set in N . with a. 1) = = Thus Q is smooth in a neighborhood of 0. there i s a neighborhood V C N of y such that # f ' ( y ' ) = #f-'(y) for any y' E V . and the image f ( x ) is called a critical value. being a sphere with finitely many points removed. Ur of these which are mapped diffeomorphically onto neighborhoods V. 1). 1) of S2.e.z + * * * + an?). with M compact.'Ph+(x) (0. For the proof it is first necessary to pass from the plane of complex numbers to a compact manifold. Thus f is an onto mapping. A A V. + a. . . We may then take Figure 3. for x # (0.(See Figure 3. . Stereographic projection hood of the north pole. we define #f-'(y) to be the number of points in f-'(y).) We will identify R2 X 0 with the plane of complex numbers. for P fails to be a local diffeomorphism only a t the zeros of the derivative polynomial P'(z) = a. The polynomial map P from R2 x 0 itself corresponds to a map f from S2to itself. is therefore connected. [Let x l . The first observation to be made about #f-'(y) is that it is locally constant as a function of y (where y ranges only through regular values!)..UiJ . that h. and a regular value y E N . and f-'(y) is discrete. being a closed subset of the compact. is nonsingular. h. then x is called a critical point of f . # 0.

it follows that the complement R" .it is possible to cover f(C) by a sequence of cubes in R" having total n-dimensional volume less than E . and let C = { X E U I rankdf. IN GENERAL. Brown in 1935. If M Rkl choose a linear map I. Now define c F :M by F(E) = -+ N X R"-" (f(t).Thc derivative d F . (Compare Whitney [38]. Thus d F . . Hence F maps some neighborhood U of x diffeoinorphically onto a neighborhood I' of (y. More generally consider a smooth map f : M -+ N .. P. and i f y E N i s a regular value. Then C will be called the set of critical points. is then a vector space of dimension na .' ( y ) . then clearly C = U . and the smooth manifold f-'(l) is the unit sphere. Let f : U + R" be a smooth m a p .m' called the space of normal vectors to M' in M at x. is not onto). Brown). given any B > 0.Kote that f-'(y) corresponds. [361. U V ) ) . : T M . deJined on a n open set U C Rm.) Since M can be covered by a countable collection of neighborhoods each diffeomorphic to a n open subset of R"'. to the hyperplane y X R"'-".L ( t ) ) . T h e n the image f(C) C R" has Lebesgue measure zero. hence the theorem says simply that f ( U ) has measure zero.n. + TN. df. In fact F maps f-'(y) r\ U diffeomorphically onto (y X R"-") A V.f(C) must be everywhere denset in R". will therefore be an (nz. it has already been noted that T M : is a subspace of T M . then the set f . Let x E f .(. from a manifold of dimension m to a manifold of dimension n.'l in the sense indicated by the next theorem. : Rk -+ R'''-n that is nonsingular on this subspace (31 C T M . I n order to exploit this corollary we will need the following: Lemma 1. Consider the function f : R" --+ R defined by f(.n)-dimensional vector space. Sard in 19-12 following earlier work by A. (References [30].e. + * * * + xi..) I f M' is a manifold which is contained in &I. * In other words. which was proved by A.f (C) the set of repular values of f . The null space 91 C T M . If f : AT --+ N i s a smooth m a p between manifolds of dimension m 2 n.* Since a set of measure zero cannot contain any nonvacuous open set. of dfL. t Proved by Arthur B. As an example we can give an easy proof that the unit sphere Sm-' is a smooth manifold. B. But this set will be LLsmall. onto 2'". T h e set of regular values of a smooth m a p --f N i s everywhere dense in N . It is essential for the proof that f should have many derivatives. T H E THEOREM has rank less than n (i. If m < n.) We will be mainly interested in the case m 2 n. and the complement N . The orthogonal cornplement of TM: in T M . (This agrees with our previous definitions in the case n~ = n. Since y is a regular value. Illorse. C Rk. Any y # 0 is a regular value. < n ) . is nonsingular. f(C) the set of critical values. (References [51. we have: f :M O F SARD AND BROWN Corollary (A. This result was rediscovered by Dubovickii in 1953 and by Thom in 1954. the derivative djl must map TAI. Let C be the set of all x E M such that P ROOF. This proves that f ' ( y ) is a smooth manifold o f dimension m . for x E M'.) Theorem.n. In particular let M' = f ' ( y ) for a regular value y of f : M -+ N .) = x: + x. under F . it is too much to hope that the set of critical values of a smooth map be finite.Regular values 11 82. L ( x ) ) .' ( y ) C M i s a smooth manifold of dimension m . The proof will be given in \$3. [HI. is clearly given by the formula dPz(v) = (dfdv). [S].

The unit disk Dm. maps the space o j normal vectors to M' isomorphically onto TN. : T M . e .n)-manifold with boundary.n. Therefore the set g-'(y) f l H" = f-'(y) A U . we see that df. MANIFOLDS WITH BOUNDARY The lemmas above can be sharpened so as to apply to a map defined on a smooth "manifold with boundary. H" = { ( x 1 . then as before f-'(y) is a smooth manifold in the neighborhood of 2. we may assume that g has no critical points. is a full m-dimensional vector space. The interior X . so that T X . ~ . Hence d f l maps the orthogonal complement of T M : isomorphically onto TN. even if x is a boundary point. 2 O}." Consider first the closed half-space PROOF. it suffices to consider the special case of a map f : H" + R". Hence g-'(y) is a smooth manifold of dimension m . Let Z E f-'(y). both for f and for the restriction is a smooth (m . The null space of df. consisting of all x E g-'(y) with n ( x ) 2 0.) E R" I X. 0 . of the submanifold M' = f-'(y). Furthermore the boundary a ( f .dX is a smooth manifold of dimension m. is a smooth manifold. Let X be a compact manifold with boundary. by Lemma 3. PROOF. i s precisely equal to the tangent space T M : C T M .. The boundary aX is the set of all points in X which correspond to points of dH" under such a diffeomorphism. Suppose that 3 is a boundary point.. It is not hard to show that dX is a well-defined smooth manifold of dimension nz . The tangent space T X . . with regular value y E R". T h e set of x in M with g(x) 2 0 i s a smooth manifold. The boundary dH" is defined to be the hyperplane R"-' X 0 We claim that IT has 0 as a regular value.16 Lemma 2. n(x1. The proof is just like the proof of Lemma 1. with boundary equal to n-'(O). : R" + R". Counting dimensions we see that df. dg. T N . Choose a smooth map g : U + R" that is defined throughout a neighborhood of 3 in R" and coincides with f on U A H". C R".1. This completes the proof. to zero. where m > n. THE BROUWER FIXED POINT THEOREM We now apply this result to prove the key lemma leading to the classical Brouwer fixed point theorem. but the hypothesis that f I aH" is regular at x guarantees that this null space cannot be completely contained in R"-' X 0. Since we have to prove a local property. consisting of all x 1- R'" with cx: 2 0.' ( y ) ) i s precisely equal to the intersection of f . x. * . Let M be a manifold without boundary and let g : M ---f R have 0 as regular value. smooth manifold.. Let IT : g-l(y) --f R denote the coordinate projection. A subset X C Rk is called a smooth m-manifold with boundary if each x E X has a neighborhood U n X diffeomorphic to an open subset V n H" of H". Now consider a smooth map f : X -+N from an m-manifold with boundary to an n-manifold.' ( y ) with aX. = d f . 2. then f-'(y) C X Lemma 4. M' i is a.) = x. For the tangent space of g-'(y) at a point x c r .' ( O ) is equal to the null space of DEFINITION. If y c N i s a regular value. Replacing U by a smaller neighborhood if necessary. maps the subspace TA!!: C T M . . Sard-Brown theorem Brouwer fixed point theorem 13 E EXAMPLE. Here is one method for generating examples. with boundary equal to g-'(O) . If z is an interior point. with boundary equal to A S " ' . f 1 ax. Lemma 3. is defined just as in 51. From the diagram - 52.

However. Fiyrtre But f-'(y) is also compact.u)z. (Compare 58. Then the continuous function 1 IG(x) . Given t > 0. There i s no smooth map f : X wise jixed.e. This contradiction establishes the lemma. let f(x) E 8'-' be the point nearer x on the line through x and g(x). x:. where Then clearly P maps D" into D" and I IP(x) . We reduce this theorem to the lemma by approximating G by a smooth mapping.\$1 I must take on :L minimum p > 0 on W .) the expression under the square root sign being strictly positive. (To see that f is smooth we make the following explicit computation: f(x) = x tu. Suppose there were such a map f. + D" has aJixed point (i. Since y is certainly a regular value for the identity map f I a X also. . f See for example Dieudonn6 [7. a point P ROOF. Here x .G(x)II < t for 2 E D". Choosing 1' : D"+ I)" as above. aX that leaves a X point- P ROOF (following M.u +d 1 . Thus P is a smooth map from D" to itself without a fixed point. PROOF. + + The procedure employed here can frequently be applied in more general situations: to prove a proposition about continuous mappings. z P(x> = P'(X)/(l + 4. and the only compact 1-manifolds are finite disjoint unions of circles and segments. 1331. . Let y E ax be a regular value for f. In particular the unit disk 4 D"= (xER"Ix:+ +z:< 1) of D" into points outside of D".* so that aff'(y) must consist of a n even number of points.) and subsequently llxll denotes the euclidean length d Brouwer Fixed Point Theorem. Suppose that G ( x ) # x for all x E D".G ( x )I I < p for all x. I + U = IIX x . Problem 4. P Lemma 6. For x E D". we clearly have P ( x ) # x.g(4ll ' t = -x. Sard-Brown theorem + Brouwer jixed point theorem 15 Lemma 5.G(s)1 I < 2e for x E D". Hence as a special case we have proved that the identity map of A"-' cannot be extended to a smooth map D" --f AY-'. we first establish the result for smooth mappings and then try to use an approximation theorem to pass to the continuous case. Hirsch). and completes the proof. according to the Weierstrass approximation theorem. with I IP(x) . This contradicts Lemma 6 . it follows that f-'(y) is a smooth 1manifold. (See Figure 4. p. with boundary consisting of the single point f-'(y) n ax = {v). A n y smooth map g : D" D" with g(x) = x ) . Any continuous function G : D" + D" has a Jixed point.g(x> . .) Then f : D" --+ s"-' is a smooth map with f(x) = x for x E s"-'.x ' x + (x. may send points * A proof is given in the Appendix. Suppose g has no fixed point. To correct this we set is a compact manifold bounded by the unit sphere x"-'. : R" -+ R" with IIP1(x) . which is impossible by Lemma 5 . . P .14 \$2.t there is a polynomial function P .

is nonsingular. (REMARK. The image f(C .) Q3. ) . is vacuous unless f is constant on an entire component of U . (Compare de Rham [29. Sincc f # C. C C denote the set of all x E U such that the first derivative d f z is zero. and let C be the set of critical points. which is not zero a t 3.) has measure zero. since C = C. ..Step 1 17 the C. Sincc dh. We may assume that p 2 2. (By definition Ro consists of a single point. give a unified proof which makes these cases look just as bad as the others. Consider the map h : U --f R” defined by < p. 51-52]. p.) has measure zero. the theorem is certainly true for n = 0. The proof will be by induction on n. If f happens to be real analytic.) To start the induction.C. c 3 c. The image f(C.C. . STEP 2. Note that the statement makes sense for n 2 0. R EMARK. 101. p 2 1. P R O O F O F P ROOF O F STEP 1. Thus we have a descending sequence of closed sets h(x) = ( f l ( X > . however. STEP 3. h maps some neighborhood V of 3 diffeomorphically onto a n open set V’. Hcnce in this case it is sufficient to carry out Steps 1 and 2. * For an easy proof (as well as an alternative proof of Sard’s theorem) see Sternberg [Xi. Note that the set C’ of critical points of g is precisely h ( V n C). with U open in Rn. The image f(C. but the general case follows easily from this special case. pp.C. we will find an open neighborhood V C R” so that f ( V (7 C) has measure zero. c R” has measure zero.. Let f : U .) We will. The coniposition g = f o h-’ will then map V’ into R”. for i 2 1. denote the set of x such that all partial derivatives of f of order < i vanish a t x. is covered by countably many of these neighborhoods.+R” be a smooth map. We will need the well known theorem of Fubini* which asserts that a measurable set A SARD’S THEOREM* C R” = R‘ X R”-’ iriust have measure xero i f it intersects each hyperplane in a set of ( p . More generally let C. xz.. say dfl/dzl.1)-dimensional measure zero. * * 1 2 . there is some partial derivative. Sternberg assumes that A is compact.C. 3 . . The details are somewhat easier since we assume that f is infinitely differentiable. The cases where n _< p are comparatively easy. (constant) X Rp-’ For each f E C . * * The proof will be divided into three steps as follows: STEP 1. when p = 1.) has measure zero for k sufficiently large. Since C . hence the set g(C’) of critical values of g is equal to f ( V n C). Construction of the map g * Our proof is based on that given by Pontryagin [28]. Let C. 3 c. . This first step is perhaps the hardest.) has measure FIRST let us recall the statement: Theorem of Sard. 3 c. this will prove that f (C . then the intersection of I Figure 5 .C. that is the set of all x E U with rank d f z Then f(C) zero.

Then any point of I1 can be written as x h. A I". Here c is a coiistant wliich depends oiily on f and I". with + WlI I c + g(C'> = j(V has measure zero.. the compactness of I". and the definition of ck. .*+. the set of critical values of 0 has measure zero in R". From Taylor's theorem. the set of critical values of g t has measure zero in t X R"-'. 2.1 to be precise) we will prove that f(Ck A I") has measure zero. For each z E c k .. since it can be expressed as a countable union of compact subsets. where 1) + h) ll&. EIence f(ck I") is contained in a union of a t most 1 ' cubes having total volume 2) n-(k+l)p V 5 rn(a/l'k+l)n = a D1 ' akf. n In) must have measure zero. . If k is sufficiently large ( k > n / p . . .. Therefore the set of critical values of g intersects each hyperplane t x R"-' in a set of measure zero. Thus the function W(X) = + llhll I di(6h). 22. where a = 2c (di 6 ) k + 1 is constant. A V ) is certainly a critical point of 0 (since all derivatives of order _<k vanish).J belongs to the hyperplane t X R"-' C R": thus g carries hyperplanes into hyperplanes. xpl .which is not zero. * . = f(x) + R(z1 h) llhll"' According to the induction hypothesis./ax. ax./ax8. be a cube of the subdivision which contains a point x of c k . Since c k can be covered by countably many such cubes. x h E I".+. Hence.* 1 z . = vanishes at but aw/dx.az. Therefore gh(Ck n V ) = f(Ck n V ) has measure zero. Now subdivide I" into Y* cubes of edge 6/r. C) PROOF O F S TEP 2.C. Suppose for definiteness that s. Ck+l) has measure zero. so f(C..) has measure zero.. Let I" C U be a cube with edge 6. . . Then the map h : U -+ R" defined by 1. I'rom 1) it follows that / ( I . By induction. by Fubini's theorem. Since c. it follows that f(C.+. ) lies in a cube of edge a / ~ ~ ccntered +' about f(z).) E V' note that g(t. Proof of Surd's theorem Step 3 19 For each ( t .. This completes the proof of Sard's theorem.. + h(z>= (W(Z>. But each point in h(C. -. > carries some neighborhood V of diffeomorphically onto an open set V'. does not. and Step 1 is complete. there is some (Ic I)-'' derivative ak*'f. . x2. This set g ( C ' ) is measurable. Let gt : ( t X Rn-l) A V' -+ t X R"-' 1 denote the restriction of g. x. Note that a point of t X R"-' is critical for gt if and only if it is critical for g. . then evidently V tends to 0 as 1 ' -+ m . Let I . If k 1 > n / p . we see that f(. for the matrix of first derivatives of g has the form PROOF O F S TEP 3. . the set I for x E C. Let 0 : (0 X Rn-') n V' -+ R" denote the restriction of g. this will prove that f(C. Note that h carries Ck A V into the hyperplane 0 X R"-'. Again we consider g = f o h-' : V' -+ R".4 18 \$3.Ck+1 is covered by countably many such sets V .

More generally this same definition works for any smooth (For example. and both manifolds have the same dimension. Fb. This map F is called a smooth homotopy between f and g. We will prove that the residue class modulo 2 of #f-’(y) does not depend on the choice of the regular value y. t ) = g(x) - 5 for Q 5 t 5 + 1. If f and g happen to be diffeomorphisms from X to Y . 0) I = fb). y is also a regular value for F . we can also define the concept of a “smooth isotopy” between f and g. since otherwise the mod 2 degree would necessarily be zero. t ) = F ( x . let p(t) = X(t . where M i s compact and without boundary. 1 1 -+ Y from f to g so that. I] 3 [0. The diffeomorphism f is smoothly isotopic to g if there exists a smooth homotopy F : X X [O. it is easy to prove that f h. 11 + N be a smooth homotopy between f and g. This also will be an equivalence relation.) For the proof we introduce two new concepts. Let F : M X [0. g : M --+ N be smoothly homotopic m a p s between manifolds of the same dimension. * If M is a smooth manifold without boundary. where X ( T ) = 0 for T 5 0 and X(7) = exp( -. 11 is a smooth manifold bounded by two “copies” of M . g It will turn out that the mod 2 degree of a map depends only on its smooth homotopy class: Homotopy Lemma. - G(z.t ) ) .) Given a smooth homotopy F between f and g. THE DEGREE MODULO 2 O F A MAPPING for all z E X . This residue class is called the mod 2 degree of f . I f y E N i s a regular value for both f and g . Then F-’(y) . PROOF. . with the aid of this construction. N is connected. 11.-I) for 7 > 0.Homotopy and isotopy 21 smooth map F : X X [O.0 Y. (We may as well assume also that N is compact without boundary. p ( t ) ) defines a smooth homotopy G with + G(z. then #f-’(y> = #9-’(y> :x+ Y are called smoothly homotopic (abbreviated f - g) if there exists a (mod 2). I] denote the subset* of Rk” consisting of all (x. I] + Y with F ( x . Two mappings f. the formula G(x. CONSIDER a smooth map f : S“ + S”. Given X C Rk. To see that it is transitive we use the existence of a smooth function rp : [0. Note that the relation of smooth homotopy is an equivalence relation.8 ) A(\$ . Let f. let X X [0. If y is a regular value.Q ) / ( X ( t . then. I] with p(t) = 0 for 0 5 t 5 for \$ 5 p(t) = 1 5 t 5 1. then M X [0. t) maps X diffeomorphically ont. the correspondence z -+ SMOOTH HOMOTOPY AND SMOOTH ISOTOPY F(z. 1) = 9(z) 84. for each t E [0. DEFINITION. recall that #f-’(y) denotes the number of solutions x to the equation f(z) = y . t ) = f(x) for 0 5 t I f:M--+N where M is compact without boundary. 13. First suppose that.t) with z E X and 0 5 t 5 1. Now if f g and g h. Boundary points of M will give rise to “corner” points of M X [O.

l l ~ 1 1 ~ where ) A ( t ) = 0 for t 5 0 and A(t) = exp(-t-') for t > 0. 2) F"(3) = 3. We will use the notation x ( t ) = F t ( 2 ) for this solution. Thus #j-'(y) #g-'(y) is even. ( 2 ) is defined for all t and 2 and depends smoothly on t and 2. 3) F . so that (For example let p(z) = A(l . C Figure 7 .). # f ' ( Y ' > for all y' E = #f-W . and therefore #f-'(y) = #g-'(Y>T(mod 2 ) .. Let y and z be arbitrary interior points of the smooth. z. Then #f-'(Y> = #f-'(z) = #g-%) = #s-YY)). V. We will also need the following: Homogeneity Lemma. I ~ * Compare [22. consisting of regular values of f . + (For the special case N = x" the proof is easy: simply choose h to be the rotation which carries y into z and leaves fixed all vectors orthogonal to the plane through y and z. #g-'(u') = #g-'(Y) c N so that For any 2 E R" these equations have a unique solution x = x ( t ) . 0 F. - * * . C N of y .22 \$4. n. Dejorming the unit ball Mx 0 Mxl Figure 6 .) The proof in general proceeds as follows: We will first construct a smooth isotopy from R" to itself which 1) leaves all points outside of the unit ball fixed. Recall (from 51) that #f-'(y') and #g-'(y') are locally constant functions of y' (as long as we stay away from critical values). Choose a regular value z of F within V . consider t. Then clearly 1) F . A V. for all y' E V.41.he differential equations dxi d -t -.) Given any fixed unit vector c E S"". Degree modulo 2 Homotopy and isotopy 23 is a compact 1-manifold. Now suppose that y is not a regular value of F. connected manifold N . The number of boundary points o n the left i s congruent to the number o n the right modulo 2 Let cp : R" --+ R be a smooth function which satisfies &) q(x) >0 = . with boundary equal to ~-yyn ) (M x o v ~I/Ix #f-'(y) 1) = f-'(y) x o v g-l(y) x 1. Thus the total number of boundary points of F-'(y) is equal to + #s-'(y). and 2 ) slides the origin to any desired point of the open unit ball.for 11x1 I <1 0 for 11x11 2 1. + . Then there exists a diffeomorphism h: N +N that i s smoothly isotopic to the identity and carries y into z.. which completes the proof. .. defined for all* real numbers which satisfies the initial condition x(0) = 2. 02. Thus there is a neighborhood V..(Z). and there is an analogous neighborhood V . i = 1. -. But we recall from \$2 that a compact 1-manifold always has an even number of boundary points. ( 3 ) = F .ci&.

(I. and completes the proof. the sphere is not a smooth ((retract" of the disk. EXAMPLES. as required. Lemma 5 . Assume that M is compact and boundaryless. the diffeomorphism F . then it has a neighborhood diffeomorphic to R". Degree modulo 2 Homotopy and isotopy 25 Therefore each F . f is a smooth homotopy invariant. But clearly. hence the above argument shows that every point sufficiently close to y is "isotopic" to y. = f(tz). we see that each F . But the interior of N is connected. We can now prove the main result of this section. g now shows that deg. that N is connected. ) For such a map f would give rise to a smooth homotopy I ! F ( z . F : S" X [0. so that #(h O f)-'(z) = #f-'(y>. I n the case M = s". I n other words. Call this common residue class deg2(f). which i s called the mod 2 degree of f . is smoothly isotopic to the identity under an isotopy which leaves all points outside of the unit ball fixed. If y and z are regular values of f then # f ' ( y ) = #f-'(z) (modulo 2). f = #f-'(y) = #g-'(y) = deg. c But (h 0 f)-'(z) = f-'h-'(z) = f ' ( y ) . The congruence (mod 2) deg. depends only on the smooth homotopy class of f. This is clearly a n equivalence relation. will carry the origin to any desired point in the open unit ball. PROOF. T h i s common residue class. which is a regular value for both f and g.this result implies the assertion that no smooth map f : Dn+' -+ S" leaves the sphere pointwise fixed. the Homotopy Lemma asserts that #(h 0 f)-'(z) = #f-'(z) (mod 2). is a diffeomorphism from R" onto itself. This completes the proof. and that f : A4 -+ N is smooth. Then z is a regular value of the composition h o f . Now consider a connected manifold N . Letting t vary. Call two points of N "isotopic" if there exists a smooth isotopy carrying one to the other. let h be a diffeomorphism from N to N which is isotopic to the identity and which carries y to z. with suitable choice of c and t. Since h o f is homotopic to I .Now suppose that f is smoothly homotopic to g. Hence the identity m a p of a compact boundaryless manifold i s not homotopic to a constant. Therefore #f-'(y) = #f-'(z) (mod 21. Compare \$2. By Sard's theorem. The identity map I of M has odd degree. A constant map c : M -+ M has even mod 2 degree. Theorem.e. 1 1 -+ Sn. each "isotopy class" of points in the interior of N is an open set. there exists a n element y P N .24 \$4.. and the interior of N is partitioned into disjoint open isotopy classes. hence there can be only one such isotopy class. If y is an interior point. t) between a constant map and the identity. Given regular values y and z.

The vector space R” has a standard orientation corresponding to the basis (1. I f M is connected and orientable. . 1I Figure 8. . carries the specified orientation for T M . If M is compact and N is connected. 6 9 if b: = a. Then (u2. 0).T h e Rrouwer.b. y) = zrf-’(u) sign df. . 1). this integer deg(f. 0... U . so that df. is an ‘(outward” vector. . with det(a. 0. U. into the standard orientation for R”. be a smooth map. Thus each positive dimensional vector space has precisely two orientations. : T M .. we can distinguish three kinds of vectors in the tangent space T M . Define the sign of df. 2) there are the (‘outward” vectors. I n the case of the zero dimensional vector space it is convenient to define an ‘(orientation” as the symbol 1 or . How to orient a boundary . in the sense that for each x E U the isomorphism dh. . is a linear isomorphism between oriented vector spaces. b.) < 0. to be 1 or . forming an (m . 0. . these are required to fit together as follows: For each point of M there should exist a neighborhood U C M and a diffeomorphism h mapping U onto an open subset of R“ or H”’ which is orientation preserving. then it has precisely two orientations. An oriented smooth manifold consists of a manifold M together with a choice of orientation for each tangent space TM. THE BROUWER DEGREE Now let M and N be oriented n-dimensional manifolds without boundary and let f :M+N 1) there are the vectors tangent to the boundary. . 3) there are the (‘inward” vectors forming a complementary half space.- -- - + As an example the unit sphere Sm-’ C R” can be oriented as the boundary of the disk D”.) > 0. .. An orientation for a finite dimensional real vector space is an equivalence class of ordered bases as follows: the ordered basis (bll . -+ TNt(. O R I E N T E D MANIFOLDS Each orientation for M determines an orientation for dM as follows: For x E dM choose a positively oriented basis (ul. 0 ) )(0. For any regular value y E N define + deg(f.. ) for T M . preserves or reverses orientation.) determines the same orientation as the basis (6.. If the dimension of M is 1. It is defined on a dense open subset of N . 1. I f M has a boundary.1. v2) . y) is a locally constant function of y.. * . U . If m 2 1. DEFINITIONS.1 according as df. then each boundary point x is assigned the orientation ..1)dimensional subspace T ( d M ) . at a boundary point: --.. in such a way that v2.. are tangent to the boundary (assuming that m 2 2 ) and that v. As in \$1. then the degree of f is defined as follows: Let x E M be a regular point of f . forming an open half space bounded by T(dM). . It determines the opposite orientation if det(a..1 or 1 according as a positively oriented vector a t x points inward or outward. . degree 27 Q5.) determines the required orientation for dM a t x.) -- ) + IN ORDER to define the degree as an integer (rather than an integer modulo 2) we must introduce orientations. C TM. . . (0. (See Figure 8.

= +l. and v. T h e degree deg(g. negative. as in \$4. X I . . Yo) This proves Lemma 1. as well as for f = F I M .. y ) is constant within some neigliborhood U of y. between Lemma 2. The proof will be essentially the same as that in \$4. y ) deg(f. i f and only if d F . Now consider a smooth homotopy two mappings fb) = F(O.28 \$5.' ( y ) deg(f. Oriented manifolds T h e Brouwer degree 29 Theorem A. signdf. but not for F . The complex function z -+ zk. y) is equal to deg(f. Therefore deg(f. Let v1 ( x ) denote the positively oriented unit vector tangent to A a t x. y ) does not depend o n the choice of regular value y . deg(f. It follows immediately that signdf. we can choose a regular value y for F within U and observe that d 4 f . T h e n U. and + and hence (summing over all such arcs) that deg(/ . I] x Ai? -+ N g b ) = F(1. Y) . suppose that yo is a regular value for f. hence deg(h 0 The orientations for X and N determine an orientation for A as follows: Given z e A . Y ) = deg(h O f. . Then h will preserve orientation. If f : M -+ N extends to a smooth m a p F deg(f. Adding up over all such arcs A . First suppose that y is a regular value for F . It will be called the degree of f (denoted deg f). y ) = 0. then PROOF.) The degenerate mapping f :M --+ constant e N . determines the required orientation for T A . and will then have boundary consisting of 1 X M" (with the correct orientation) and 0 X M" (with the wrong orientation).It is only necessary to keep careful control of orientations. More generally. The remainder of the proof of Theorems A and B is completely analogous to the argument in 54. Thus the degree 1 X M") a t a regular value y is equal to the difference of F I a([O. = x. tangent to A. then deg f = deg g. z) - by Lemma 2. let (vl. z). choose a diffeomorphism h : N -+ N that carries y to z and is isotopic to the identity. carries (vz. or zero. Y ) = 0.. If f i s smoothly homotopic to g . = -1. The manifold [0. * * . Let A C F-'(y) be one of these arcs. The function deg(f. Theorem B. First consider the following situation: Suppose that M is the boundary of a compact oriented manifold X and that M is oriented as the boundary of with sum zero. Y>* According to Lemma 1 this difference must be zero.+') into a positively oriented basis for TN. . (Here k may be positive. v. : X --f N . f . h ( d ) by inspection...+J be a positively oriented basis for T X . z) = = deg(f. y) for a n y common regular value y . If y and z are both regular values for f : M ---\$ N . PROOF. I] X M" can be oriented as a product. . But f is homotopic to h o f . maps the unit circle onto itself with degree k .' ( y ) is a finite union of arcs and circles.deg(f. with aA = { a ) U ( 0 ) . The compact. Figure 9. y ) = 0 for every regular value y . 1 deg(g. T h e integer deg(f. Clearly U. with only the boundary points of the arcs lying on M = a X . 1-manifold F . y) = 0. U. EXAMPLES. which completes the proof. F : [0.(x) points outward a t one boundary point (say b) and inward a t the other boundary point a. z ) . we have proved that deg(f.z # 0. How to orient F .. is a smooth function. Lemma 1. We will show that sign dfa sign dfb = 0. with U. Hence.

incidentally. (Compare smooth field of nonzero tangent vectors i Figures 10 and 11. 2n+1) = ( ~ 1 .(zl.23. G+J.l)"". As an application. then we may as well suppose that 2) 31 has degree zero. For in any case 8(z) = v ( ~ ) / ~ ~ would v ~ ~ )be ~ ~a vector field which does satisfy this condition. Now define a smooth homotopy The antipodal map of S" has degree (.-. Figure 10 (above). . 7r] by the formula F(z.2r(x) = 1 for all z E S".X = 0 for all 2E S".. . as we can see by noting that the antipodal map is the composition of n 1 reflections: + F : S" X [0. . a - . : S" ---f x". e). 7r) = -2. I n 57 we will prove a more general result which implies Hopf's theorem.+~(X). One example of an orientation reversing difi'eomorphism is provided by the reflection r . .0) i = z.30 \$5.) F(z. On the other hand. If v(z) is nonzero for all z. This completes the proof. F ( X . we show that S" admits a f and only i f n i s odd. T h u s a n orientation reversing difeoinorphism of a compact boundaryless manifold i s not smoothly homotopic to the identity. r. Attempts for n = B DEFINITION. A nonzero vector jield on the 1-sphere Figure 11 (below). L2k) = (22. .. . A diffeomorphism f : Jl + N has dtgree 1 or .. the explicit formula U(X1. Oriented manifolds T h e Brouwer degree using the euclidean inner product. 7 X2ki -22k-1) defines a nonzero tangent vector field on S". Thus we can think of v as a smooth function from x" to itself. . that the antipodal map of S" i s homotopic to the identity for n odd. .* . a fact not detected by the dcgree modulo 2. following Brouwer.1. A famous theorem due to Heinz Hopf asserts that two mappings from a connected n-manifold to the n-sphere are smoothly homotopic i f and only i f they have the same degree. 54. OT. * . the antipodal m a p of x" i s not smoothly homotopic to the identity. where + v(z).. for each x E M . if n = 2k . Computation shows that = -z = r1 o r 2 o 0 . -L. Rut for n even we have seen that this is impossible. I t follows. Thus the antipodal map of S" is homotopic to the identity. 6) = z + S" cos 0 + v ( x ) sin 6. I n the case of R " " this is clearly equivalent to the condition the sphere S" c 1) V(X). * . . A smooth tangent vector jield on M C Rk is a smooth map v : M -+ Rksuch that ~ ( x E) T M . .1 according as f preserves or reverses orientation. e) and that Y 1 T h u s if n i s even.cl. F(z.F(z.

with a vector field on each manifold. The function fib> = U(. on U' under a diffeomorphisiii f : U + U'. To explain what this means. 0. = df 0 2. VECTOR FIELDS AND THE EULER NUMBER t=I A s a further application of the concept of degree. It is these curves which are actually sketched in Figure 12. and the function zk defines a vector field with a zero of index -k. T h e n the index of v at a n isolated ' at f(z). 1. let us consider the more general situation of a map f : M --+ N .)/ I lv(4 I I I maps a small sphere centered a t z into the unit sphere. carries V ( X ) into v ' ( f ( x ) ) for each x E M .(xl. df o of-' * Each sphere is to be oriented as the boundary of the corresponding disk. with indices . 0 f-1 will be used. Some examples. zero z i s equal to the index of U . . Lemma 1. xJ.The index 33 86. Suppose that the vector jield v on U corresponds to V' = DEFINITION. The vector fields v on A I and U' on N correspond under f if the derivative df. Consider first an open set U C R" and a smooth vector field v :U--+R" with an isolated zero a t the point z E t = O U . we study vector fields on other manifolds. then clearly U' is uniquely determined by The notation 2)' U.1. (Intimately associated with v are the curves "tangent" to v which are obtained by solving the differential equations dx. Examples of plane vector fields If f is a diffeomorphism. We must prove that this concept of index is invariant under diffeomorphism of U. ./dt = v. .* The degree of this mapping is called the index L of v a t the zero z. are illustrated in Figure 12. + C=+l c =+2 Figure 12. 2.) A zero with arbitrary index can be obtained as follows: In the plane of complex numbers the polynomial zk defines a smooth vector field with a zero of index k a t the origin.

. Hence the index of v = vo a t 0 must be equal to the index of U' = U . We may assume that f(0) can be defined by = 0. . then w i s required to point outward at all boundary points. for many values of m there exists an orientation prcserving diffeomorphism of the sphere S'" which is not smoothly isotopic to the identity.34 \$6. * See for example [22. t) F ( x . g m are suitable smooth functions. 4041. Thus f is isotopic to the linear mapping dfn. If f preserves orientation. * + xmgm(tx) for all values of 1. If M has a boundary. We may assume that z = f ( x ) = 0 and that U is convex. (In contrast. A n y orientation preserving diffeomorphism f of R" i s smoothly isotopic to the identity. even as t f(x) = where g . This proves Lemma 1 for orientation preserving diff eomorphisms. which is clearly isotoyi(* to the identity.=O (-1)' rank H . + R" by the formula F(x. These vector fields are all defined and nonzero on a sufficiently small sphere centered a t 0. and f. a t 0.. . so the associated function e'(x) = v'(x)/l lv'(z)lj on the +sphere satisfies 0 p-1. First consider the special case of a compact domain in R"'. Then v' = p o v 0 p-l. 5' = p 0 0 P ROOF. and note that F(xi t ) = + . LEMMA 1. ( M ) . Vector jields The index sum we construct a one-parameter family of embeddings f t : U+R" 36 Assuming Lemma 1. proceeding exactly as above. * Here H i ( M ) denotes the i-th homology group of M . To consider diffeomorphisms which reverse orientation it is sufficient to consider the special case of a reflection p . t-0 it is natural to define an isotopy F : R" X [O. We will study the following classical result: Let M be A compact manifold and w a smooth vector field on M with isolated zeros. x1g1(x) 0. we can define the concept of index for a vector field w on an arbitrary manifold M as follows: If g : U + M is a parametrization of a neighborhood of z in M . * * In particular this index s u m i s a topological invariant of M : it does not depend on the particular choice of vector jield. To prove that F is smooth. It clearly will follow from Lemma 1 that L is well defined. f l = f . which completes the proof of Lemma 1. The full theorem was proved by Hopf [14] in 1926 after earlier partial results by Brouwer and Hadamard.(O) = 0 for all t. + <t 1. This will be our first and last reference to homology theory. Let X C R" be a compact m-manifold with boundary. See [20. 51. The s u m C L such a vector jield i s equal to the Euler number* d f o ( x ) = lim f ( t z ) / t . The proof of Lemma 1 will be based on the proof of a quite different result: Lemma 2. OF PROOF ax . ( U ) . then.+ xmgm(x). p. then the index L of w a t x is defined to be equal to the index of the corresponding vector field dg-' 0 w o g on U a t the zero g-'(z).+ sm-' assigns to each x E aX the outward unit normal vector a t x.' on f . The Gauss mapping 9 : Xlgl(tx) + * . . Since the derivative a t 0 Evidently the degree of 8' equals the degree of fi.) We will prove part of this theorem. and sketch a proof of the rest. 1 1. p. we write f in the form* . which corresponds to v on U .) with f o = identity. m x(M) = . (A 2-dimensional version of this theorem was proved by Poincare in 1885. of the indices at the zeros of Poincare-Hopf Theorem. Let U . denote the vector field d f t o v of. This proves Lemma 2. 0) = = f(tx)/t for 0 dfo(x>.

c does not depend on the choice of U . By definition U = d h . to itself. This integer is of course equal to the Euler number of X .) Hence the index is certainly equal to +l.)t' + c:v. * . Before extending this result to other manifolds. A n example with index sum +1 PROOF. : R"' + R"' is defined.) in terms of the derivatives of v a t z. The vector field v is nondegenerate a t z if the linear transformation dv. If v : X + R" i s a smooth vector Jield with isolated zeros. We must compute the image of t' = t'(u) under the linear transformation dw. The function 8 ( x ) = v(z)/ilv(x)ll maps this manifold into S"-'. t" form a basis for the tangent space TllIhc. and the degrees on the other boundary components add up to -E L . is nonsingular. I f U preserves orientation.) . Let h U 4 A I be a parametrization of some neighborhood of z . But 6 I a X is homotopic to 9. Lemma 4. hence L = -1. It is natural to try to compute the index of a vector field v at a zero z so that the vectors i'. +1 or . (The minus sign occurs since each small sphere gets the wrong orientation. L Lemma 5. = 0 REMARK. + PROOF. The degree of g is also known as the "curvatura integra" of d X . then L = + l . In particular. and if v points out of X along the boundary. actually carries T M . It follows that z is a n isolated zero. Consider first a vector field U on an open set U C R"' and think of v as a mapping U -+ R". Figure IS.e' be the vector field on U which corresponds to the vector Let = field w on M . -+ Rk is defined. For example. of z into R".1 according as D i s positive or negative. and hence can be cmsidered as a linear transformation from T M . Vector . : T M . we have seen that vlUo can be deformed smoothly into the identity without introducing any new zeros.. T h e derivative dw. Removing a n €-ball around each zero. .(at'/azc.). then similarly v can be deformed into a reflection.(e') = aw(h(u))/au. (Compare Figure 13./au.36 \$6. The index of v at a nondegenerate zero z i s either according as the determinant of dv. Think of v as a diffeomorphism from some convex neighborhood U . First note that - . Hence the sum of the degrees of 8 restricted to the various boundary components is zero.. since it can be expressed as a constant times the integral over a X of the Gaussian curvature. into the subspace T M .1 PROOF. = E: (av. 2. (See Lemmas 1..(v) = Therefore 2) Vati- aw(h(u))/au. If this linear transformation has determinant D # 0 then z i s a n isolated zero of w with index equal to 1 or . if a vector field on the disk D" points outward along the boundary.(. Think of w as a map from llf to Rk so that the dcrivative dw. Let e' denote the i-th basis vector of R" and let tt = dh. then the index s u m L i s equal to the degree of the Gauss inapping from dX to S"-'. we obtain a new manifold with boundary. i s positive or negative.. We may assume that z = 0. so that dv.fields The index sum 37 Lemma 3 (Hopf). dwh(-)(t') = d(w 0 h). c DEFINITION. C Rk.) Therefore deg(g) as required. If v reverses orientation.(e') = ah/au. 1) v. More generally consider a zero z of a vector field w on a manifold A l C Rk. For tn odd it is equal to half the Euler number of a X . some more preliminaries are needed. SO that w(h(u)) = dh.' ~ w 0 h.

implies that L = 0 * For example. [9].r(x)) arid u ( r ( x ) ) are mutually orthogonal. Now consider a compact.) for 172 odd.( p ./au. (See \$8... For any odd-dimensional. At the north pole the vectors converge inward. v can be defined by the formula u(z) = p .l)m. For E sufficiently small one can show that N . and then evaluating a t the zero h-’(z) of obtain the formula 3) we We will also consider the squared distance function p(x) = 1 1 2 dwz(t*) = xi(dv. for each point 2 of the level surface d N . Kow according to Lemma 3 the index sum L is equal to the degree of g. is equal to the determinant of dv.r(z)) Theorem 1.). On the sphere S“ there exists a vector field v which points “north” a t every point. dv. thus yielding an m-dimensional version of the classical Gauss-Bonnet theorem. Problem 12. by setting w(z) = (x .)t’. -+ Sk-’. T h e index s u m 39 Combining 1) and 2). For a n y vector field v on M with only nondegenerate zeros. hence the index is (.) . + TAI.U .yII 5 for some y E M ) . Hence. maps T M . Problem 11.) 2(x . Kote that w can vanish only a t the zeros of v in M : this is clear since the two summands (x . This gives a new proof that every vector field on an even sphere has a zero. This proves Theorem 1.* At the south pole the vectors radiate outward. (References [l].(h) = = PROOF.(h) dw. i s equal to the degree of the Gauss mapping* g : d N . the index s u m L +u(44).r(z))/e. then each index is and the equality multiplied by (. * A different interpretation of this degree has been given by Allendoerfer and Fenchel: the degree of g can be expressed as the integral over M of a suitable curvature scalar. z ) z . Thus the invariant L is equal to 0 or 2 according as m is odd or even. is a smooth manifold with boundary.r ( x ) is perpendicular to the tangent space of i l l a t ~(x).r(z)). For if the vector field v is replaced by . boundaryless manifold M C Rk. h for h E TM:. then the function r(z) is smooth a11d well defined. is equal to the determinant of the matrix (av. I n particular this sum does not depend on the choice of vector j e l d . boundaryless manifold the invariant L is zero. For x E N . Together with Lemma 4 this completes the proof. + Figure 14. (Compurc \$8.Let N . The eneighborhood of M c L = (-1)” c 1. If E is sufficiently small. . See also Chern [S]. into itself. Computing the derivative of w :it a zero z E AI. denote the closed eneighborhood of M(i. since the inner product w ( x ) . where p is the north pole.g ( x ) is equal to e > 0.r(2)1j2. (See Figure 11.) Note that the vector x . unit normal vector is given by the outward g(z) = grad p/l lgrad p [I Extend U = (z . the set of all x E Rk with 1 1 5 .(h) for all h E TA1.38 \$6. to a vector field w on the neighborhood N . Then w points outward along the boundary.e. Hence the index of U’ a t the zero z is equal to the index L of v a t x . Thus the determinant of d w .l)m./au. p An easy computation shows that the gradient of grad p = is given by Thus dw. EXAMPLES. for otherwise ~ ( x would ) not be the closest point of M . = p-’(cz). let ~ ( x E)M denote the closest point of AI. we see that dw. hence the index is 1. Vector jields U. and the determinant L) of this linear transformation T M .

Proving the theorem /or a vector Jield with degenerate zeros. 11 takes the value 1 on a small neighborhood N . Manifolds with boundary. Morse showed that the sum of indices associated with such a gradient field is equal to the Euler number of M . Furthermore.e. For details of this argument the reader is referred to Milnor [22. Vector fields The index s u m 41 L = 0 on a connected manifold M . Identification of the invariant It is sufficient to construct just one example of a nondegenerate vector i equal to x ( M ) . + this is the following: According to Marston Morse.40 56.field with isolated zeros can be replaced by a nondegenerate vector jield without altering the 1. . . it is always possible to find a real valued function on M whose ('gradient" is a nondegenerate vector field.X(x)y is nondegenerate* within N . * Clearly d is nondegenerate within N I . Consider first a vector field v on an open set U with an isolated zero at z. Applying this argument locally we see that any vector . then any vector field v which points outward along aM can again be extended over the neighborhood N . More generally consider vector fields on a compact manifold M . 361. if defined as before by w(z) = v(r(z)) z . If x : U--+ [ O . integer C STEP 3. and hence does not change during this alteration. 29. three further steps are needed. differentiable of class Cm) manifold. The most pleasant way of doing field on M with but only a C'-manifold.r ( z ) . L with the Euler number x(M). I n order to obtain the full strength of the PoincarBHopf theorem. STEP 1. STEP 2.N I . so as to point outward along a N . of z and the value 0 outside a slightly larger neighborhood N . The extension w. then the vector field d ( z ) = U)(. will only be a continuous vector field near aM. and if y is a sufficiently small regular value of U . I f of Hopf asserts that there exists a vector field on M with no zeros a t all. But if y is sufficiently small. then v' will have no zeros a t all within N . Thus N . is not a smooth (i. The sum of the indices a t the zeros within N can be evaluated as the degree of the map 0 : aN + X"-'. If M C Rk has a boundary. pp. then a theorem REMARK. . However. there is some difficulty with smoothness around the boundary of M . The argument can nonetheless be carried out either by showing that our strong differentiability assumptions are not really necessary or by other methods.

-+ z'(sp)y maps the subspace Tf-'(y)=to zero and maps its orthogonal complement Tf-'(y)t isoinorphically onto T(Sv)II. . 1 U of X . . If y' is another regular value off and b' is a positively then the framed manifold ( f ' ( y ' ) .The Pontryagin cmstruction @ 87. . b) is called a framed submanifold of M . but they all belong to a single framed cobordism class: F DEFINITION. . 1 1 can be extended to a compact manifold X C M X 10. framed cobordant to ( f ' ( y ) . Two mappings from M to s" are smoothly homotopic i f and only if the associated Pontryagin inanifolds are framed cobordant.. . THE degree of a mapping M -+ M' is defined only when the manifolds M and M' are oriented and have the same dimension. The map f induces a framing of the manifold f-'(y) as follows: Choose a positively oriented basis b = (U'. corresponding to different choices of y and b. uE(z. f*b) will be called the Pontryagin manifold associated with f . Let N and N' be compact n-dimensional submanifolds of M with aN = aN' = aM = 0. Of course f has many Pontryagin manifolds. b) are framed cobordant if there 1 between N and N' and a framing exists a cobordism X C A1 x [0. of M X [0. of normal vectors to N in M at x. which is defined for a smooth map f:M--+S" from a n arbitrary compact. 1 1 so that that maps into U' under df.) DEFINITION. t) = (U'(. Theorem B. Two framed submanifolds ( N . Hence there is a unique vector DEFINITION. E) U N' X (1 - E. . due to Pontryagin. . so that u'(x. A framing of the submanifold N C M is a smooth function t. .n is called the codimension of the submanifolds. I] w'(x) E Tj-'(y): c TM. First some definitions. vrn-'(x)) for the space TN: C T M . N is cobordant to N' within M if the subset N X [ 0 . boundaryless manifold to a sphere. ) .:Pastiny together two eobordisms within M Figure 16. Again this is an equivalence relation. and so that X does not intersect M X 0 U M X 1 except at the points of ax. f*d) is oriented basis for T(SI))III. . U") for the tangent space T ( S Y ) v . = 0) for (2. t ) E N X [0.) The pair ( N . . FRAMED COBORDISM MxO MX I MX2 T H E PONTRYAGIN CONSTRUCTION Figure 15. 0) for (z. which assigns to each x E N a basis b(x) = ( U ' . F Theorem A. f*b). Clearly cobordism is an equivalence relation. This framed manifold (f-'(y). For each x E f-'(y) recall from page 12 that dfb : TAP. (See Figure 15.). a x = N x OUN' x 1. Now consider a smooth map f : AP + S" and a regular value y E S". E) E. . The difference of dimensions m . (See . t ) E N' X (1 - I]. ~ " ( x of ) ff'(y). b) and (N'. It will be convenient to use the notation tm = f*b for the resulting framing w'(x). We will study a generalization. .t ) (w'(x).

1 ) = r. t) = PROOF.e < t I 1. then f-'(z) is framed cobordant to f-'(y). h ~ ) o j codimension p in M occurs as Pontryagin manifold for some smooth mapping f : M S". Theorem C. If f and g are smoothly homotopic and y i s a regular value for both. f ( z ) 0 5 t < e. 2) rt equals r 1 for 1 . R ) of matrices with positive determinant. Choose a homotopy F with Lemma 1.'(z) lies on the great circle from y t o z .f(z). F(z. . --f Thus the homotopy classes of maps are in one-one correspondence with the framed cobordism classes of submanifolds. Framed cobordisin The Pontryagin construction 45 can be identified with the space GL'(p. f*b'). an isotopy) r r : S" S" so that rl(y) = z. Given z with I[z . 1 . Lemma 3. This proves Lemma 3. f*b) i s framed cobordant to (f-'(y). Framed submanifolds and a framed cobordism is a regular value for the mapping F . If y i s U regular value off. PROOF. This framed manifolds f-'(z) and (rl o f ) . Such a path gives rise to the required framing of the cobordism f-'(y) X [0. and 3) each r." Lemma 2. The proof of Theorem A will be very similar to the arguments in \$04 and 5 . and z i s suficiently close to y.e. choose a smooth one-parameter family of rotations (i. then the Pontryagin manifold (f-'(y). For each t note that z is a regular value of the composition rt 0 \ : ill -+ S". This is possible since this space of bases is framed cobordant to f-'(y) and so that g .' ( z ) is framed cobordant to g-'(y). It follows a fortiori that Figure 16. we can choose > 0 so that the e-neighborhood of y contains only regular values. and hence is a regular value of f. If b and b' are two dij'erent positively oriented bases at y.. and hence is connected. I].' ( z ) = f-'rF'(z) proves Lemma 2. e --f Define the homotopy F : M X [0. Choose a smooth path from b to b ' in the space of all positively oriented bases for T(S"). A n y compact /rained suhmanifold ( N . By abuse of notation we will often delete reference to f*b and speak simply of "the framed manifold f-'(y). t ) = g(z) Choose a regular value z for F which is close enough to y so that f-'(z) F ( z . then f-'(y) i s framed cobordant to g-l(y). PROOF.e' < 1 5 1. Since the set f(C) of critical values is conipact. 11 is a framed manifold and provides a framed cobordism between the = f-'(y). and so that 1) r t is the identity for 0 5 t < E'.yI( < e. Herice F-'(z) C M X 10. It will be based on three lemmas. 1 1 --f Sv by F(z.44 \$7. Then F-'(z) is a framed manifold and provides a framed cobordism between f-'(z) and g-'(z).

Assume that N is compact and that aN = aM = @. y ) ) = y . The remainder of the proof proceeds exactly as before. o . PROOF OF THEOREM C. Given any two regular values y and z for f. denotes the e-neighborhood of 0 in R”. Constructing a m a p with given Pontryagin manifold U> = dx’. hence g maps some neighborhood of (2. More precisely let g(z..) be the endpoirit of the geodesic segment of length I ltld ( x ) * . Framed cobordism The Pontryagin construction 47 PROOF OF THEOREM A.. This completes the proof of Theorem A. for e sufficiently small.. (rl 0 f)-’(z) = f-’r. is diffeomorphic to the full euclidean space R” under the correspondence u . t.. dlffeomorphically onto an open set. The reader who is familiar with geodesics will have no difficulty in checking that g:NXU. say to xo. . . x + t1U1(X) + * * .) Since g(x. = Since N is compact. + t”U”(X). ) is nonsingular.’(z) f-’(y). . . framed submanifold. 1 . . and with U -+ 0 and U ’ --+ 0. REMARK. and define the projection Figure 17. Thus f is homotopic to rl 0 f . Consider the mapping g : N X R” -+ M . First suppose that M is the euclidean space R”’”. For otherwise there would exist pairs (z. f + t.) Clearly 0 is a regular value. this proves the Product Neighborhood Theorem for the special case M = E’”. (See Figure 18. 0) E N X R” diffeomorphically onto an open set. Let N C AI be a compact. we could choose a sequence of such pairs with x converging. But U . t”) = by s ( g ( x .. that g is one-one in a neighborhood of (xo. Product neighborhoods do not exist for arbitrary subnianifolds. t.-+M is well defined and smooth. Choose a product representation g:NXIZ”-+VCM for a neighborhood V of N . where U . 0 ) E N x R” and so that each normal frame ~ ( x corresponds ) to the standard basis for R”. as above. The proof of Theorem C will be based on the following: Let N C M be a framed submanifold of codimension p with framing b. o . of N X 0. and n-’(O)is precisely N with its given framing.llui12/e2). Then clearly xo = x. U ) # (x’. and since dg. Some neighborhood of N in M i s diffeomorphic to the product N x R”. boundaryless. * . (Compare Figure 17..+ u / ( l . we can choose rotations r t :SP-+S” so that ro is the identity and r l ( y ) hence f-’(z) is framed cobordant to = z.46 \$7. Thus g maps N X U . . 0) = x . and we have contradicted the statement 0).. We will prove that g is one-one on the entire neighborhood N X U . . defined by g ( x . An unJramable submanifold T : V-+R” PROOF. with x’ converging to x. t. For the general case it is necessary to replace straight lines in R”’” by geodesics in M . Clearly d g ( z .U”(X)/~ It#’(X) + + + * * f + t g V ( x )I 1. . Furthermore the diffeomorphism can be chosen so that each x E N corresponds to (2.u”(x)I I in M which starts a t x with the initial velocity vector tlul(x) + . U’). does what is expected of it. t. U’) in N X R” with llull and I/u’I( arbitrarily small and with ! & l N Figure 18. providing that e > 0 is sufficiently small. Product Neighborhood Theorem...

. g : M be smooth maps with a common regular value y.t). t) F(z. In order to prove Theorem B we must first show that the Pontryagin S” manifold of a map determines its homotopy class. That is.@ \$?. Hence I(F(.so.u). &(x)) for x E V for x # V .N + proves that f is smoothly homotopic to g. where V is small enough so that f ( V )and + X(U)~G(X. Let h : S” .where h is the stereographic projection from SO and where X is a smooth monotone decreasing function with X ( t ) > 0 for t < 1 and X ( t ) = 0 for 1 2 1.X ( U ) ~ ] F ( U Z ).with a similar inequality for G. U ) = [l .u). Choose a product representation for 0 < \lull < Min(c.u 2 . 0 F ( x .c1 [lull3> 0 = f(4 for x e V = h-’[t*h(f(x)) (1 . G : N X R” with --f R”. N X R” + V C M for a neighborhood V of N . f:M-+S” by f(z) = F-‘(O) = G-’(0) = N X 0. then f i s smoothly homotopic to g. Then the homotopy bet. The hypothesis that f*b = g*b means that df.weeiiF and G will not map any new points into 0.u > 0. the points F(z. G(z. g*b).y -+ R” be stereographic projection. = dg. this clearly completes the proof of Lemma 4. (2) coincides with F for ( / U \ ( 2 c. We will first find a constant c so that n-’(O> F(z.U) + tG(x. and maps the open unit ball in R pdiffeomorphically* onto 8” . and the point cp(0) is a regular value of f . U) * For example. = f(x>= s o dG(. It will be convenient to set N = f-l(y). deforms F = Fo into a mapping F . that (1) coincides with G in the region \ ( U ( ( < c/2. = (projection to R”) Clearly f is smooth. and with dF. To avoid moving distant points we select a smooth map X : R p -+ R with X(U) X(U) = = 1 for l l u l l 5 c/2 ljull 2 c.o.u). being careful not to map any new points into y during the deformation. --f for z E N and 0 < llul[ < c. Define I g ( V ) do not contain the antipode fj of y. 0 for Now the honiotopy F t ( z .u >0 is precisely equal to the framed manifold N .. and (3) has no new zeros. f*b) i s equal to (g-’(y). for all 2 E N .uI I I/U113 llU/j2 and w. Thus is suffices to deform f so that it coincides with g in some small neighborhood of N . PROOF.5 with R”. U ) and G ( x . Identifying V with N X R” and identifying S” .. a t least for I Iu[ I < c.h(g(x))] for x E M . c1 llull I 1. U ) belong to the same open half-space in R”. this completes the proof of Theorem C. By Taylor’s theorem IIF(z. Since the corresponding Pontryagin manifold f-l(P(0)) = for all 5 E N . .. So the honiotopy (1 .u) -~ lI l c1 l l ~ l l ~ for . U) Lemma 4. Making a corresponding deformation of the original mapping f . Let f.t)F(z. I f the framed manifold (f-’(y). p(z) = h-l(z/A(\\z1\2))... First suppose that f actually coincides with g throughout an entire neighborhood V of N . we obtain corresponding mappings F . Framed cobordisrn The Pontryagin construction 49 Now choose a smooth map cp : R” 3 x” which maps every x with 11x1 2 1into a base point so. 4 .’. l).u).

Theorems A. However. then two maps M are homotopic i f and only i f they have the same mod 2 degree.and similarly F1 g. then two maps M -+ S"' are smoothly homotopic i f and only i f they have the same degree.* Pontryagin's construction is. With considerably more difficulty he was able to show that there are just two homotopy classes also in the case nz = p 2 2 4. oriented. The composition operation in ?rp(M. An easy argument then proves the following: whose Pontryagin manifold (F-'(y). B. m). an argument completely analogous to the proof of Theorem C constructs a homotopy F : M X [0. and boundaryless. If f and g are smoothly homotopic.) + + THE HOPF THEOREM * See for example S.1 according as the preferred basis determines the right or wrong orientation. Framed cobordism T h e Hopf theorem 61 PROOF OF T HEOREM B.50 \$7. note that the maps F . Let sgn(z) equal +1 or . Homotopy As an example. Setting F . (Compare \$8. Hu. it conversely enables us to translate any information about homotopy into manifold theory.1]-+x" Theorem of Hopf. Thus we have proved the following. we can slide x around M in a closed loop so as to transform the given basis into one of opposite orientation. ( z ) = F(z. Some of the deepest work in modern topology has come from the interplay of these two theories. [all. let M be a connected and oriented manifold of dimension m = p. and f have exactly the same Pontryagin manifold. The theory of framed cobordisni was iiitroduced by Pontryagiii in order to study homotopy classes of mappings S" + x" am --\$ (S".p > 2 this approach to the problem runs into manifold difficulties. for m . - - - Theorem. using framed 2-nianifolds. Theory. and C can easily be generalized so as to apply t o a manifold M with boundary.so) are in one-one correspondence with the cobordisni classes of franicd submanifolds N C Interior(M) of codimension p . If M i s connected. given a framed cobordism ( X .) + with m > p. then this set of homotopy classes can be given the structure of an abelian group. (References [36]. then Lemma 3 asserts that the Pontryagin manifolds f f ' ( y ) and g-l(y) are framed cobordant. F 4 ) is precisely equal to ( X . Problem 17. -+ S'" 4 . For example if m = p 1 2 4. more algebraic methods. which completes the proof of Theorem B. A framed submanifold of codimension p is just a finite set of points with a preferred basis at each. Hence F . Then sgn(x) is clearly equal to the degree of the associated map M -+ S"'. r ~ ) between f-'(y) and 9-l (y). Conversely. Then given a basis for T M . On the other hand. there are precisely two hoiiiotopy classes of mappings S"' + S". But it is not difficult to see that the framed cobordisni class of the 0-manifold is uniquely determined by this integer sgn(z). Therefore f g . If M i s cmnected but nonorientable. REMARKS. t ) . however. If p 2 am 1. Ren6 Thorn's work on cobordism is an important example of this. suppose that M is not orientable. The essential idea is to consider only mappings which carry the boundary into a base point so. The homotopy classes of such mappings (AI.-T. It has since turned out to be easier to eriunierate homotopy classes of mappings by quite different. f by Lemma 4. d M ) corresponds to the union operation for disjoint framed submanifolds of Interior (M). Pontryagin proved this result by classifying framed 1-manifolds in S"'. called the p-th cohomotopy group ?rP(Ml d M ) . a double-edged tool. It not only allows us to translate information about manifolds into homotopy theory. .

PROBLEM 3. PROBLEM 2. Using this retraction r. . y) E M X N with f(z) = y. PROBLEM 1. is equal to the graph of the linear map df. Show that any map S" different from (. PROBLEM 8. Given M TAI = C Rk. X TN. PROBLEM 4. . show that the correspondence (x. of M in Rk. If two smooth maps X -+ x" are continuously homotopic. Define r : N .is equal to T M .+ Sk PROBLEM 7. show that the tangent space T ( M x N ) c . PROBLEM 11. PROBLEM 9. prove that Z(M. Similarly show that the normal bundle space E = ((~. with total dimension m the degree of X is called the linking number l ( M . then oriented.Exercises manifold and that the tangent space 53 T r ( z . + + + PROBLEM 13. Show that r(z v) is closer to x U than any other point of M . n = k. Show that any map s" -+ S" of odd degree must carry some pair of antipodal points into a pair of antipodal points. If M and N are compact.g(x)11 < 2 for all x.y)/llx .M ) = (-l)(m+l)(n+l) w 1NI. If two maps f and g from X to x"satisfy I If(x) . Show that every complex polynomial of degree n gives rise to a smooth map from the Gauss sphere S2 to itself of degree n. PROBLEM 5.. show that the tangent bundle space {(x. y) = (x . prove that f is homotopic to g.l)"+' must have a fixed point. . 88. N ) . If m to a constant.yII. d(g 0 f ) = (dg) o ( d f ) . EXERCISES PROBLEM 10. N -+ c R'".4 H z from E to Rkmaps the E-neighborhood of M X 0 in E diffeomorphically (Compare the Product Neighboronto the eneighborhood N . Show that any smooth map f : M gives rise to a smooth map df : TM -+ N TN where IN CONCLUSION here are some problems for the reader. d(identity) = identity. X : M X N . show that every map fiI" --\$ S" is homotopic S" with degree PROBLEM 12. If M is compact and boundaryless. N ) = 0. Define the linking number for disjoint manifolds in the sphere X"+"". hood Theorem in \$7. prove the analogue of Problem 4 in which the sphere S" is replaced by a manifold M . Given smooth manifolds M C Rk and N C R'.If X is compact. show that every continuous map X -+ S" can be uniformly approximated by a smooth map. Show that the degree of a composition g 0 f is equal to the product (degree g) (degree f). and boundaryless. Prove that + l(N.) + < p . (Brouwer).U ) E AI X Rk 1 v E Tillz) -+ is also a smooth manifold. show that they are smoothly homotopic. Show that I ' is a smooth If M bounds a n oriented manifold X disjoint from N . T K x TN.) is a smooth manifold. the homotopy being smooth if f and g are smooth. the linking map PROBLEM 6. vc .~~)rnIXRk~~~TTnI. The graph r of a smooth map f : M -+ N is defined to be the set of all (5. is defined by X(z. -+ M by r ( x v ) = 2. Given disjoint manifolds M .. .

Prove that the resulting intersection is a smooth manifold. use the disjoint union operation to make an abelian group. A brief discussion of the classification problem for higher dimensional manifolds will also be given. 2. + PROBLEM 17. (An interval is a connected subset of R which is not a point.) Since any interval is diffeomorphic* either to [0. PROBLEM 16.. for each s E I . The proof will make use of the concept of “arc-length. Any given local parametrization I’ -+ M can be transformed into a parametrization by arc-length by a straightforward change of variables. IIY .) PROBLEM 15. 11. closed. Let f : I M and g : J --\$ M be parametrizations by arc-length. 2.41 = f-’W G’(Y). prove that H ( g 0 f) is equal to H ( f ) multiplied by the square of the degree of g. has unit length. If the dimension p is odd. 1). t Thus I can have boundary points only if M has boundary points. then f can be extended to a parametrization by arc-length of the union f(I) U g ( J ) .” Let I be an interval. f-’(zN = Vg-l(Y). (Compare p.. 50. Ifit has only one component. hence the linking number l(f-’(y). prove that K f 1 ( Y ) . Then f(I) A g ( J ) has at most two components..’ ( z ) ) depends only on the honlotopy class of f . g-‘(4). Prove that H ( a ) = 1. for each x E N f l N’.3. it follows that there are only four distinct connected 1-manifolds. It may be finite or infinite. c) Prove that l(f-’(y). If i t has two components.. For W E WILL prove the following result. and T N J together generate TM. 11. 1 . (Reference [15]. Lemma. open. A map f : I 4 M is a parametrization by arc-length if f maps I difleomorphically onto an open subset t of M .’ ( z ) ) is called the Hopf invariant of f.54 \$8. then M must be diffeomorphic to X ’ . This integer H(f) = l(f-’(y).9(”)1/ < for all x. (0. a) Prove that this linking number is locally constant as a function of y. --f If p 2 inz 1.2 2 . then the manifolds f-’(y). the subspaces T N . Exercises PROBLEM 14.’ ( z ) can be oriented as in \$5.) + nP(M) into * For example. and if the “velocity vector” d f . Let IIp(M) denote the set of all framed cobordism classes of codimension p in M. use a diffeomorphism of the form f(t) = a tanh ( t ) + b. ( l ) E TM. which has been assumed in the text. If y # z are regular values for a map f : S”-’ 4 S”. Two submanifolds N and N‘ of A f are said to intersect transversally if. f . . A n y smooth.) = h-’((. The Hopf Jibration a : S3--+ S2is defined by a ( z l . or (0. where APPENDIX CLASSIFYING 1-MAN1F O LD S Ilf(4 . b) I f y and z are regular values of g also. and does not depend on the choice of y and z. (If n n’ < in this means that N f l N’ = @. prove that H(f) a composition S2P-l = 0. f-’(z)) is defined. or half-open.) I f N is a framed submanifold. Use the transverse intersection operation to define a correspondence rIP(M)x rI*(M)3 rI”+“(M). THE HOPF INVARIANT.cl + ix2)/(x3 + i d ) i where h denotes stereographic projection to the complex plane. f . connected 1-dimensional manifold i s difeomorphic either to the circle S‘ or to some interval of real numbers.8”4 8” Theorem. prove that it can be deformed slightly so as to intersect a given N‘ transversally. DEFINITION. f .

there would be a limit point z of f ( I ) in M . and Poknaru in Fundamenta Mathematicae. being compact and open in AI. Hence has a t most two components. '1 The image h ( S ' ) . Since r is closed and 9-l 0 f is locally a diffeoniorphism. The study of %dimensional manifolds is very much a topic of current research. Then J? is a closed subset of I X J made up of line segments of slope f1. If has two components.f ( I ) .a ) . must be the entire manifold M . This contradicts the assumption that f is maximal and hence completes the proof. PROOF O F CLASSIFICATION THEOREM. there can be a t most one of these segments ending on each of the four edges of the rectangle I x J .* But for high dimensional simply connected manifolds there has been much progress in recent years. but must extend to the boundary. p ) if necessary. 8 -I A U Figure 19. Consider the graph r C I x J . This proves the lemma. as exemplified by the work of Smale [31]and Wall [37]. . (See l'apakyrialtopoulos [as]. Any parametrization by arclength can be extended to one f :I+M which is maximal in the sense that f cannot be extended over any larger interval as a parametrization by arc-length: it is only necessary to extend f as far as possible to the left and then as far as possible to the right. the required difieomorphism h:S'-+M is defined by the formula h(cos 0. -+ R.66 Appendix Classifying I-manifolds 57 PROOF. For manifolds of higher dimension the classification problem becomes quite formidable. < t < 8. * See Markov [19]. these line segments cannot end in the interior of I x J . consisting of all (s. we may assume that y = c and 6 = d. (See Figure 19. Furthermore the derivative of g-' o f is equal to i l everywhere. For if the open set f ( I ) were not all of M . Now setting 8 = 2?rt/(a . Since g-' o f is one-one and single valued. then g-l o f extends to a linear map L : R Now f and g o L piece together to yield the required extension F : I U L-'(J) + f ( I ) U g ( J ) . the two must have the same slope. Clearly 9-l 0 f maps some relatively open subset of I diffeomorphically onto a relatively open subset of J . Three of the possibilities for r i I f r is connected. and also a forthcoming paper by Boone. and hence is a diffeoniorphism. Parametrizing a neighborhood of x by arc-length and applying the lemma. they must be arranged as in the left-hand rectangle of Figure 19. with slope say 1.) lcor coinpact nianifolds of dimension 2 4 the classification problem is actually unsolvable. so that a + <b l c < d < a <p. if there are two components. For 2-dimensional manifolds. t ) with f(s) = g(t). If M is not diffeoniorphic to A". REMAHKS. Haken. a thorough exposition has been given by KerBkjArtb [17]. sin 0) = = f(t) g(t) for a for c < t < d. Translating the interval J = (y. we would see that f can be extended over a larger interval. we will prove that f is onto.) Furthermore.

379-394.) Fenchel. S. L. New York: Academic Press. A. London Math. Annalen 71 (1012). Auslander. and Sternberg [35]. S. C. 1960. Hu [16]. 38 (1935).) Chern. de Rham [29]. Lang [18]. J. (1953).” Trans.. T. New York: Harper & Row.“On total curvatures of Riemannian manifolds. Hoinology Theory. “Uber Abbildung von ?tlannigfaltigkeitcn”. 15 (1940). Foundations of Modern Analysis. (In Russian. and S. I h l w numlw of a Ricmann m:inifold. . 13.itltlison-Wcslcy. Brouwer. 443-464. Goffman.. For background knowledge in closely related fields. Sbornik N. 1960.. ilrlath.“Functional dependence. W... Dubovickii. and R.” Jour. Introduction to DiJerentiable Jlanilolds. U. Amer.. Hcwling. Hilton.S. C.. and Pontryagin [as]. Soc. Steenrod [34]. (See Theorem 3-111.I nier. Jour. Ya. Brown. Mass.. 1963. 97-115. Dieudonnb. The survey articles [23] and [32] should also prove useful. E. Mathematical Analysis. 243-248. For the reader who wishes to pursue the study of differential topology. “Tht. P. let me recommend Milnor [22]. 62 (1!140). 1. Soc.5-22.” Mat..” .Wiylie. Calculus of Several Variables. illcndocrfcr. Math. let me recommend Hilton and Wylie [ 1I]. “A simple intrinsic proof of the Gauss-Ronnet formula for closed Riemannian manifolds.BIBLIOGRAPHY THE following is a nliscellaneous list consisting of original sources and of reconimended textbooks. : . A. Press. .” Annals of ddath. L.4postol. MacKenzic. 45 (1944). 32 (74).. Cambridge Univ. J.. “On differentiable mappings of an n-dimensional cube into a k-dimensional cube. Alunkres [25]. 747-752. New York: McGraw-Hill. Jlath. M. 1057. 1965.

Vorlesungen uber Topologie. Congress of Math. “Smooth manifolds and their applications in homotopy theory. 1311 Smale.. New York: Interscience.” Proceedings Intern. Soc. 514-517. R. [26] Papakyriakopoulos. The Topology of Fibre Bundles. 64 (1956).. D. de. Congress of Math. Introduction to Differentiable Manifolds. [25] Munkres. 163-189.) [29] Rham. Calculus on Manifolds. 1321 .. 74 (1961). Amer. Cambridge Univ. 17-86. Jour.” illath. Saaty. hlath. pp. 1951. 28 (1954). “A proof of the nonretractibility of a cell onto its boundary.” L’Enseignement math. “Differential topology.” Lectures on Modern Mathematics.60 Bibliography Bibliography 61 [12] Hirsch. C. [35] Sternberg. “The measure of the critical points of differentiable maps.. 1958. 433-440.. [34] Steenrod.” Proc. New York: Benjamin. ~ 4 1__ . Lectures on Diflerential Geometry. ‘‘Insolubility of the problem of homeomorphy. 300-306. H. T. M. Annalen 96 (1926). Math. I1 (1959). “The theory of three-dimensional manifolds since 1950. [38] Whitney. Morse Theory. (Translated from Trudy Inst. 1161 Hu. Princeton Univ. New York: I’renticeHall. L. L.” Doklady Akad. 1923. 1959.S. Annalen 96 (1926). [27] Pontryagin. A. Varidte‘s diffkrentiables.. 1171 KerBkjdrt6. 2. Cambridge Univ. (Annals Studies 51. [19] Markov. 1964. Princeton Univ. 225-250.” Annals of Math. 69 (1963). 1964. 62-70. 1962.. 391-406. . [36] Thom.” Bull. 1963. 16-23.” Proceedings Intern. 427-440.-T.) [2O] Milnor. A.” Annals of Math. Amer. S. Nauk. “Vektorfelder in n-dimensionalen Mannigfaltigkeiten. Press. 48 (1942). N. . . pp.” Fundamenta Mathematicae 25 (1935). Translations. Homotopy Theory. C. 1 (1935).” Math. 1963. “On manifolds homeomorphic to the 7-sphere. Press. Steklov 45 (1955). SOC. “A classification of continuous transformations of a complex into a sphere. S. Berlin: Springer.. “Abbildungsklassen n-dimensionaler Mannigfaltigkeiten. “Quelques propriBtBs globales des vanet& difErentiables. Paris: Hermann. [30] Sard. “The behavior of a function on its critical set. [24] Morse. (In Russian.” Bull.R. 1960. “A survey of cobordism theory. G. 209-224. Math.” Duke Math. etl.” Annals of Math..” [28] Amer. [33] Spivak. 1 1 . “Generalized PoincarB’s conjecture in dimensions greater than four. 14 (1963). pp.” Commentarii Math. 40 (1939). 1955. 147-149. Press. 8 (1962). J. 364-365. I’. Soc. S. A. 165-183. [ 131 Hopf. New York: Wiley. [IS] Lang. S. Math. A. 1221 . 75 (1962). R. H. 1-114. M. v.” dnnals of Math. J.) Princeton Univ. (Comptes Rendues) 19 (1938). (Annals Studies 54). Soc. 131-145. [37] Wall. U.S. S. [21] . Ser.. 1965. “Classification of (n ... Press. Press. 1960. T.. Z€eZvet. “A function not constant on a connected set of critical points. S. “Uber die Abbildungen von Spharen auf Spharen niedrigerer ~ 5 1Dimension... 883-890... 399-405. 1955. Amer. Elementary Differential Topology. New York: Academic Press.. 1231 ----. C. “A survey of some recent developments in differential topology..1)-connected 2n-manifolds.

11 degree of a map. 55 Poincark. 50 coordinate system. 2 by arc-length.. Thomas Nelson Page (n6e Barbour) and the Honorable Thomas Nelson Page for the purpose of bringing to the University of Virginia each session a series of lectures by an eminent person in some field of scholarly endeavor.. 42. 44. 35. 8. 54 index (of a zcro of a vector field). 53 Morse. 13. 28 fixed point theorem. 42 cohomotopy. 40. 38 half-space. 12 . 52 degree. 1 with boundary. 1963. 14 Euler number. 59 dimension of a manifold. 40 nondegencrate zcro (of a vector field) 37. L. 36. 40 framed cobordism.INDEX PAGE-BARBOUR LECTURE SERIES The Page-Barbour Lccture Foundation was foundcd in 1007 by a gift from Mrs. 10. 20. Milnor in December.20. 14 homotopy. A. 11 critical value. H. B. A. 40 normal bundle. 42 orientation. 21. 42. L. 51 (dimension.27. 17 fundamental theorem of algebra. 22. 35 Pontryagin.. 7 disk. 14 Brown. The materials in this volume were presented by Professor John W. 2s outward vector. 12. as the forty-seventh series sponsored by the Foundation. 16. 8 Gauss-Bonnet theorem. 2-7 diffeomorphism.40. 46 Fubini theorem. 51. M. 53 normal vectors. 31. 8. 35-41 inverse function theorem. 32 critical point. H. 10. M. 26 isotopy. 11. 43 product neighborhood theorem.. 11.43 Sard. 7 regular value. 1 corresponding vector fields. 35. 38 Gauss mapping. 52 boundary. 3. 1 differential topology. 42 Pontryagin manifold. 28 mod two. 12 Hirsch.14. 35. 7 cobordism. 30.. 1. 43 framed submanifold.. 1. 36. 20. E. 5. 8. 4. 35.. 43. antipodal niap. 46 regular point. 16 smooth manifolds. A. 12 Brouwer. 10 Morse.. J.13. 11 chain rule for derivatives. 26 parametrization. 34 linking. 8 inward vector. 3234 index sum. 26 of a boundary. I>.. 1. 21. 52 Hopf. 27 of F-'(y). 20 derivative of a map.

64
smooth manifolds (cont.) the classification problem, 57 of dimension zero, 2 of dimension one, 14, 55 oriented, 26 smooth maps (= smooth mappings), 1 sphere, 2

lnrrcz
stereographic projection, 9, 48 tangent bundle, 53 tangent space, 2-5 a t a boundary point, 12, 26 tangent vector, 2 vector fields, 30, 32-41 Weierstrass approximation theorem, 13

-~
INDEX O F SYMBOLS
deg (f; Y), 27 df,, 2-7
Dm,

s o ! ,
Hm,
Rk,

1

ax,
!*a,
#P(Y)l

14

12
43
8

12 1 *-1, 2 TM,, 2-5 IIXII, 14

TOPOLOGY FROM T H E DIFFERENTIABLE VIEWPOINT was composed and printed by The William Byrd Press, Inc., Richmond, Virginia and bound by Russell Rutter Co., Inc., New York, N. Y. The types are Modern Number 8 and Century Schoolbook, and the paper is Oxford's Bookbuildcrs Plate. Design is by Edward Foss.