Gysin Sequence

In mathematics, a common and powerful strategy is to understand complex objects by breaking them down into simpler components. But how does one reverse this process? If we construct a new, intricate space by weaving together simpler ones—a process known as creating a fiber bundle—can we predict the properties of the final creation from its ingredients? This fundamental question poses a significant challenge in geometry and topology. The Gysin sequence provides a remarkably elegant and powerful answer for a crucial class of such constructions: orientable sphere bundles. This article serves as a guide to this essential tool. In the "Principles and Mechanisms" chapter, we will dissect the algebraic machinery of the sequence, revealing how a single geometric invariant, the Euler class, drives its computations. Following that, the "Applications and Interdisciplinary Connections" chapter will showcase the sequence's power in action, from unraveling the structure of classical spaces like projective spaces to its revolutionary role in the discovery of exotic spheres.

Imagine you have a collection of spheres, and you decide to attach one sphere to every single point of some other space, say, another sphere or a torus. You don't just place them next to each other; you weave them together into a new, larger, and often much more complicated space called a ​​fiber bundle​​. A natural, and rather difficult, question arises: if we know the properties of the base space and the fibers, can we deduce the properties of the grand, composite space? Answering this is like trying to predict the exact texture and flavor of a cake just by knowing the ingredients and the mixing instructions. The "mixing instructions" in our case correspond to the way the fibers are "twisted" as they are attached over the base.

The ​​Gysin sequence​​ is a remarkable piece of mathematical machinery that provides a surprisingly precise answer to this question for a specific, yet common, setup: when the fibers are spheres and the bundle is ​​orientable​​. It presents us with a long, interconnected ladder of relationships between the cohomology groups of the base space ($B$) and the total space ($E$). In essence, it allows us to climb from the known world of the base to the unknown world of the total space.

For an oriented bundle with an $(n-1)$-sphere fiber $S^{n-1}$ over a base $B$, the Gysin sequence in cohomology is a long exact sequence that looks like this:

Let's not be intimidated by the symbols. Think of this as a perfectly balanced assembly line. Each group is a station, and the arrows are conveyor belts. The term ​​exact​​ means that at every station, the material arriving from the previous belt is precisely the set of items that the next belt is designed to discard. This perfect balance creates a rigid structure, constraining the possible nature of the unknown groups $H^k(E)$. The sequence links the cohomology of the base space to itself at different degrees, and cleverly inserts the cohomology of the total space in between.

But what drives this machine? What is this mysterious arrow labeled "$\cup e$"? This is where the physics, the geometry, and the magic truly reside.

The map $\cup e$ represents the ​​cup product​​ with a very special cohomology class $e \in H^n(B)$, called the ​​Euler class​​ of the bundle. This single element is the mathematical embodiment of the "twist" in our bundle. If you were to glue circles ($S^1$) over another circle ($S^1$), you could do it in the straightforward way to get a torus (no twist), or you could add a half-twist as you go around to create a Klein bottle (a non-orientable twist). For orientable bundles, the Euler class measures a more general notion of twisting. A zero Euler class means the bundle is, in a cohomological sense, "untwisted." A non-zero Euler class tells us the bundle has some global topological complexity.

This connecting map, $\cup e$, is the engine of the Gysin sequence. It's the crucial link that dictates how the base space's properties influence the total space. Let's see it in action.

Consider a bundle of circles ($S^1$, so $n=2$) over a 2-sphere ($S^2$). The Euler class $e$ lives in $H^2(S^2; \mathbb{Z})$, which is just the integers $\mathbb{Z}$. So, the "twist" is simply captured by an integer, let's call it $n_E$. The Gysin sequence can be run like a program. When we do, we find that the cohomology of the total space $E$ depends critically on this integer:

• $H^1(E; \mathbb{Z})$ is trivial if $n_E \neq 0$, but is $\mathbb{Z}$ if $n_E = 0$.
• $H^2(E; \mathbb{Z})$ is the cyclic group $\mathbb{Z}_{|n_E|}$ if $n_E \neq 0$, but is $\mathbb{Z}$ if $n_E = 0$.
• Remarkably, $H^3(E; \mathbb{Z})$ is always $\mathbb{Z}$, no matter what the twist $n_E$ is!

The twist tangles up the first and second cohomology groups, but leaves the third untouched. In the case where the twist $n_E$ is non-zero, it can even create ​​torsion​​—elements which, when added to themselves a finite number of times, become zero. Think of it as a rope that, when twisted, develops kinks ($H^2(E)$) and tightens up so certain loops can no longer be formed ($H^1(E)$). We see a similar phenomenon for a circle bundle over a torus, where a non-zero Euler class introduces a torsion component into the cohomology of the total space, with the order of the torsion directly related to the integer defining the class. The Euler class isn't just an abstract symbol; it's a number we can use to compute concrete properties.

This raises a deeper question. Where does this Euler class come from? Is it just an abstract topological invariant, or does it relate to something more tangible, something a physicist or a geometer might measure? The answer, discovered through the magnificent theory of ​​Chern-Weil​​, is that the topological twist is a shadow of geometric curvature.

For ​​complex vector bundles​​ (spaces where each fiber is a complex vector space $\mathbb{C}^r$), there is a whole family of characteristic classes called ​​Chern classes​​. For a rank $r$ complex bundle, which can be viewed as an oriented real bundle of rank $2r$, the Euler class is nothing but the top Chern class, $c_r$. This provides a powerful dictionary for translating problems between real and complex settings.

The Chern-Weil homomorphism takes this one step further. It states that you can compute these purely topological Chern classes by integrating polynomials of the ​​curvature​​ of a connection on the bundle. Curvature is a local geometric quantity—it tells you how much the space is bending at each point. This is profound: by measuring local bending everywhere and adding it all up, you can determine a global, topological property of the bundle that is fundamentally discrete (like an integer).

A beautiful illustration comes from considering a complex line bundle ($r=1$) over a closed, oriented surface $\Sigma_g$ of genus $g$. Here, the Euler class is the first Chern class, $c_1(L)$. The Chern-Weil correspondence tells us its de Rham cohomology class is represented by the form $\frac{i}{2\pi}F$, where $F$ is the curvature 2-form. If the curvature is given as $F = -2\pi i d \omega$ for some integer $d$ (the degree of the bundle), the connecting map in the Gysin sequence acting on the fundamental class $1 \in H^0(\Sigma_g)$ simply produces the class $d[\omega]$. The topological sequence is powered by a geometric invariant!

What if the Euler class is zero? This is the "untwisted" case. The engine of the sequence is turned off. The connecting map $\cup e$ becomes the zero map, sending everything to zero. The long exact sequence then shatters into a collection of simple ​​short exact sequences​​:

For vector spaces, this means the dimension of the middle group is just the sum of the dimensions of the outer two. The calculation becomes wonderfully simple. A striking example is the unit tangent bundle of the 2-sphere, $T_1S^2$. This is an $S^1$-bundle over $S^2$ whose Euler class corresponds to the integer 2. However, if we compute cohomology with coefficients in $\mathbb{Z}_2$ (the field with two elements, where $1+1=0$), the Euler class becomes $2 \equiv 0 \pmod 2$. The twist becomes invisible to $\mathbb{Z}_2$-cohomology. The Gysin sequence splits, and we can easily compute the cohomology of the total space, which turns out to be the same as that of the 3-dimensional real projective space $\mathbb{R}P^3$. This demonstrates that the "twist" we detect depends intimately on the coefficients we use to probe the space.

Like any great principle in physics or mathematics, the Gysin sequence obeys a principle of universality, or ​​naturality​​. This means that it respects maps between bundles. If you have a map from one bundle to another, you get a corresponding ladder-like diagram of their Gysin sequences where every square commutes. This isn't just an aesthetic feature; it's an incredibly powerful computational tool when combined with algebraic results like the ​​five-lemma​​.

For instance, if we have an $S^3$-bundle over an $S^4$ and we pull it back via a map $f: S^4 \to S^4$ of degree $-1$, the naturality of the Gysin sequence gives us a diagram relating the original bundle to the new one. Since a map of degree $-1$ induces isomorphisms on the homology of $S^4$, the five-lemma forces the conclusion that the induced map on the total spaces, $\tilde{f}_*: H_k(f^*E) \to H_k(E)$, must also be an isomorphism for all $k$. We can deduce a property of the total spaces just by knowing a property of the map on the base.

Finally, a crucial word of caution. Our whole discussion has relied on the bundle being ​​orientable​​. What if it's not, like a Möbius strip? Then, the fibers are glued together with orientation-reversing maps. The standard Gysin sequence breaks down. To salvage it, one must pass to a more general theory using "twisted coefficients" or deploy the even more powerful ​​Serre spectral sequence​​, which takes the monodromy action of the fibration into account. This limitation does not diminish the Gysin sequence's power; rather, it beautifully delineates the border between two regimes of topology—the oriented and the non-oriented worlds. Within its domain, the Gysin sequence remains a testament to the deep and often surprising connections that knit the fabric of geometric spaces.

We have seen the algebraic machinery of the Gysin sequence, this marvelous chain of maps linking the topology of a sphere bundle to its base. But what is it good for? A tool is only as impressive as what it can build or dissect. It is here, in its applications, that the Gysin sequence truly reveals its power and beauty. It is not merely a computational curiosity; it is a lens through which we can understand the structure of some of the most important spaces in geometry and physics, a bridge connecting disparate mathematical fields, and even a factory for constructing new and unexpected geometric worlds.

Let's take our new machine for a spin, starting with some of the most classic and elegant structures in topology.

Perhaps the most celebrated family of fiber bundles are the Hopf fibrations. Let's first look at the bundle that started it all: the fibration of the 3-sphere over the 2-sphere, with a circle as the fiber ($S^1 \to S^3 \to S^2$). We can use the Gysin sequence to compute the cohomology of the total space, $S^3$, from the base, $S^2$. For this bundle, the fiber is one-dimensional ($n-1=1$, so $n=2$), and its Euler class $e$ lies in $H^2(S^2; \mathbb{Z})$. A relevant portion of the cohomology Gysin sequence is: $\dots \to H^3(S^2; \mathbb{Z}) \to H^3(S^3; \mathbb{Z}) \to H^2(S^2; \mathbb{Z}) \xrightarrow{\cup e} H^4(S^2; \mathbb{Z}) \to \dots$ We know that $H^3(S^2; \mathbb{Z}) = 0$ and $H^4(S^2; \mathbb{Z}) = 0$. The sequence thus simplifies to: $\dots \to 0 \to H^3(S^3; \mathbb{Z}) \to \mathbb{Z} \to 0 \to \dots$ For this to be an exact sequence, the map from $H^3(S^3; \mathbb{Z})$ to $\mathbb{Z}$ must be an isomorphism. This forces $H^3(S^3; \mathbb{Z}) \cong \mathbb{Z}$, correctly recovering a known fact. The machine works! It correctly predicts the topology of the total space from its components.

This success emboldens us. Let's try something more ambitious. The complex projective spaces, $\mathbb{C}P^n$, are the arenas for much of algebraic geometry and quantum mechanics. They are constructed by identifying lines through the origin in complex space. While their CW complex structure gives clues about their homology, the Gysin sequence provides a stunningly elegant, inductive proof. For each $n$, there is a Hopf fibration $S^1 \to S^{2n+1} \to \mathbb{C}P^n$. The Gysin sequence relates the known homology of a high-dimensional sphere to the homology of $\mathbb{C}P^n$ and $\mathbb{C}P^{n-1}$. For most dimensions, the sequence creates a direct isomorphism: $H_k(\mathbb{C}P^n) \cong H_{k-2}(\mathbb{C}P^n)$. Starting with $H_0(\mathbb{C}P^n) \cong \mathbb{Z}$ and $H_1(\mathbb{C}P^n) = 0$, this isomorphism machine churns out the entire homology structure: a copy of the integers $\mathbb{Z}$ in every even dimension up to $2n$, and zero everywhere else. The intricate structure of this infinite family of spaces is unraveled step-by-step by our sequence.

From these well-behaved spaces, let's turn to something a little more twisted: the lens spaces, $L(p,q)$. These can be defined as quotients of the 3-sphere by a cyclic group action. But they can also be viewed as total spaces of circle bundles over the 2-sphere. From this perspective, the Gysin sequence tells a remarkable story. The integer $p$ from the group action materializes as the Euler class of the bundle. The sequence then immediately predicts that the second cohomology group must be $\mathbb{Z}/p\mathbb{Z}$. This is a beautiful example of a topological invariant—torsion—arising directly from the "twist" of the bundle, which itself encodes the group action.

The Gysin sequence does more than just analyze abstractly defined spaces; it illuminates concrete geometric objects. Consider the unit tangent bundle of a surface, like the 2-sphere $S^2$. This is the space of all positions on the sphere paired with all possible directions (unit tangent vectors) at each position. This space, $T_1S^2$, is naturally a circle bundle over $S^2$. What is its topology?

The Gysin sequence requires an Euler class. For a tangent bundle, the Euler class is profoundly linked to the geometry of the base manifold—it is none other than the Euler characteristic, $\chi(S^2) = 2$. Plugging this into the sequence, we find a segment that looks like: $\dots \to H^0(S^2; \mathbb{Z}) \xrightarrow{\cup e} H^2(S^2; \mathbb{Z}) \to H^2(T_1S^2; \mathbb{Z}) \to H^1(S^2; \mathbb{Z}) \to \dots$ This simplifies to a short exact sequence $0 \to \mathbb{Z} \xrightarrow{\times 2} \mathbb{Z} \to H^2(T_1S^2; \mathbb{Z}) \to 0$. The astonishing result is that $H^2(T_1S^2; \mathbb{Z}) \cong \mathbb{Z}/2\mathbb{Z}$. The Euler characteristic, a number you can compute by triangulating the sphere and counting vertices, edges, and faces, manifests as torsion in the cohomology of its tangent bundle! This principle is general: for many sphere bundles, the integer value of the Euler class directly determines the order of a torsion group in the homology or cohomology of the total space. The sequence provides a direct algebraic translation of a geometric twist into a topological twist.

The utility of the Gysin sequence extends far beyond spheres. What if the base space is more complex, like a surface of genus $g$ (a donut with $g$ holes), $\Sigma_g$? Or the product of two spheres, $S^2 \times S^2$? These spaces have richer cohomology rings of their own. The Gysin sequence shows how the Euler class interacts with this structure. For a circle bundle over $\Sigma_g$, if the Euler class is non-trivial, the sequence shows that the first Betti number of the total space is precisely $2g$, inheriting it directly from the base. For a circle bundle over $S^2 \times S^2$, if the Euler class is a generic combination of the two natural cohomology generators, the sequence shows how the dimension of the second cohomology group is reduced by exactly one.

This points toward the deep connections with algebraic geometry. Consider the Grassmannian $Gr(k, \mathbb{C}^n)$, the space of all $k$-dimensional planes in an $n$-dimensional space. These are fundamental objects in modern geometry. Over them lives the "tautological bundle," where the fiber over each point (a plane) is that plane itself. The Euler class of this bundle is a Chern class, a central invariant in algebraic geometry. The Gysin sequence can be deployed in this advanced setting to compute the homology of the associated sphere bundles, providing a crucial tool for exploring the topology of these intricate spaces. The sequence acts as a Rosetta Stone, translating the language of characteristic classes in algebraic geometry into concrete statements about homology groups.

So far, we have used the sequence to analyze spaces. But its most astonishing application is in synthesis—the construction of new mathematical objects. In 1956, John Milnor used this framework to make a revolutionary discovery: the existence of "exotic" spheres.

The construction is deceptively simple. An $S^3$-bundle can be built over $S^4$ by taking two trivial pieces, $D^4 \times S^3$, and gluing their boundaries together with a "twist." This twist can be parameterized by an integer, $k$. For each integer $k$, we get a 7-dimensional manifold, let's call it $M_k$. The question is: what are these manifolds?

This is where the Gysin sequence makes its dramatic entrance. The integer $k$ determines the Euler class of the bundle. By feeding this Euler class into the sequence, we can compute the cohomology of $M_k$ for any given $k$. The calculation reveals a surprise. For most values of $k$, the resulting manifold has torsion in its cohomology, so it certainly isn't a simple sphere. But for a few special values, namely when $k+2$ is $0$, $1$, or $-1$, the torsion vanishes! The cohomology of these specific $M_k$ manifolds looks just like the cohomology of the standard 7-sphere.

Are they the standard 7-sphere? Milnor proved the unbelievable: at least one of them is not. It is a new manifold, homeomorphic to $S^7$ (it is a sphere from a topological point of view) but not diffeomorphic to it (it has a different smooth structure). It was the first "exotic sphere." A simple-looking long exact sequence, when applied to a simple gluing construction, had revealed the existence of a new geometric universe, hiding in plain sight.

From confirming basic facts about spheres to discovering entirely new ones, the Gysin sequence demonstrates the profound principle at the heart of algebraic topology: that by translating geometry into algebra, we gain a tool of unimaginable power, capable of not only describing the world we know, but creating worlds we never thought possible.