Manifolds 41 | Measurable Sets and Null Sets
Based on The Bright Side of Mathematics's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.
A subset’s measurability on a manifold is determined by whether its chart-image is Lebesgue-measurable in R^n.
Briefing
Measurable sets and null sets are the missing pieces that make integration on manifolds work beyond a single coordinate chart. Once a subset’s image in Euclidean space is measurable (in the usual Lebesgue sense), its “generalized volume” can be assigned consistently, and if that image has Lebesgue measure zero, the subset can be ignored without changing any integral. This is what turns the manifold integral from a chart-local construction into a global tool.
The discussion starts with the familiar case of the 2-sphere S2. Because S2 is an orientable 2-dimensional manifold, it carries a canonical volume form that measures areas. Using spherical coordinates, a parameterization F maps an open domain U in (θ, φ)-space into R3 and onto the sphere. To compute the area of a subset, the integral of the manifold volume form Ω over the subset is rewritten as an integral in Euclidean coordinates: it becomes an integral of the pullback F*Ω over the parameter domain. In this setup, the pullback introduces a Jacobian factor equal to sin θ, reflecting how area elements scale under the parameterization.
A key practical point follows: the parameterization does not cover the entire sphere—only a line of points is missed. Despite that omission, the area integral still comes out to 4π, the total surface area of S2. The missing line is treated as a null set: it has Lebesgue measure zero in the relevant coordinate representation. Since null sets contribute nothing to the integral, the “gap” in coverage does not affect the result. This example motivates the formal definitions that come next.
With the sphere as motivation, the text fixes the notions of measurable and null sets for an orientable n-dimensional manifold equipped with a volume form. A set A is called measurable if, for any chart (U, φ), the intersection A ∩ U maps under the chart into a measurable subset of R^n (measurable with respect to Lebesgue measure). Similarly, A is a null set if, under the chart map, the corresponding intersection has Lebesgue measure zero. Crucially, these properties must hold regardless of which chart is used to test them, so the definitions are compatible with the manifold’s atlas.
These definitions then extend the integral. Inside a chart, the integral of Ω over a measurable set A is well-defined. Moreover, if two sets differ only by a null set, their integrals agree: the integral over A can be defined by subtracting the null set from A, and the result is unchanged. That is exactly how the sphere computation justifies ignoring the parameterization’s missing line.
The final takeaway is that integration is now robust for measurable subsets, even when they don’t fit neatly into one chart. The remaining challenge—handled in a later step—is building a general theory for sets that require multiple charts to cover, which demands technical work to ensure the global integral stays consistent.
Cornell Notes
Integration on manifolds needs a way to talk about which subsets are “integrable” and which can be safely ignored. A set A is measurable if, in any chart, the image of A ∩ U inside R^n is Lebesgue-measurable. A set is a null set if that image has Lebesgue measure zero. The sphere S2 example uses spherical coordinates: the pullback of the canonical area form introduces a sin θ Jacobian, and the integral still gives 4π even though the parameterization misses a line. That missed line is a null set, so it contributes nothing—explaining why integrals can ignore measure-zero sets and still be well-defined globally.
Why does the area integral on S2 still equal 4π even though the parameterization misses some points?
What makes a subset A of a manifold “measurable” in this framework?
How is a “null set” defined, and why does it let integrals ignore parts of a set?
How does the pullback F*Ω connect manifold integration to ordinary integration in R^n?
What does it mean to say the integral over a measurable set is well-defined inside a chart?
Review Questions
- In the S2 computation, what role does sin θ play, and where does it come from?
- How do the definitions of measurable and null sets depend on charts, and what consistency requirement prevents chart-dependent answers?
- Why does subtracting a null set from a measurable set not change the value of the integral?
Key Points
- 1
A subset’s measurability on a manifold is determined by whether its chart-image is Lebesgue-measurable in R^n.
- 2
A null set is a subset whose chart-image has Lebesgue measure zero, so it contributes nothing to integrals.
- 3
The sphere S2 example shows the pullback of the canonical volume form turns a manifold area integral into an ordinary double integral with a sin θ Jacobian.
- 4
Parameterizations may miss points on the manifold, but if the missed set is null, the integral value remains unchanged.
- 5
If two sets differ only by a null set, their integrals of the volume form agree.
- 6
Inside a single chart, integrals over measurable sets are well-defined; extending to sets spanning multiple charts requires additional technical work.