Getting Through Do Carmo Without Losing Your Mind
Do Carmo is the standard reference for a reason. The problem sets are clean, the statements are precise, and the solutions—when you can find good ones—are genuinely useful. The books in question are Differential Geometry of Curves and Surfaces and Riemannian Geometry, and most students hit them around their second year of graduate work or senior undergrad. The gap between the exposition and the exercises is wider than most textbooks advertise, and that is where the frustration lives. I have worked through both books multiple times, and I have watched students try to use solution manuals in ways that actively hurt their learning. The core issue is not whether solutions exist. They do. The issue is how you approach them.
Where Do Carmo Differential Geometry Solutions Actually Help
The best solutions I have encountered are those that show the line of thought, not just the algebra. Do Carmo's problems range from straightforward computation to results that require knowing which theorem to reach for. When a problem asks you to prove something about a curvature formula, the solution should tell you why that formula matters in that context. Most PDFs floating around the internet skip that part entirely and just dump three lines of calculation. Those are worse than useless because they create the illusion that you understand the material when you do not. The ones worth using tend to come from course websites run by universities. Professors like Brian Clymer, Sigurd Angenent, and others have posted problem sets with full solutions over the years. The University of Toronto, Stanford, and MIT have all had materials online at various points. You will also find compilations on sites like Scribd and GitHub repositories where students have shared their own worked solutions. The quality there varies enormously, so you have to check. My own go-to is the Angenent notes from UT Austin combined with the problem sets from various European universities. The solutions tend to be terse but correct, which is actually preferable. Overly long solutions make you passive. A concise solution forces you to fill in the gaps yourself.
How to Work Through the Problems
Start with the curve chapter in Curves and Surfaces. The first dozen or so problems are computational. You need to calculate curvature, torsion, and the Frenet frame for various parametric curves. These are fine for building fluency with the definitions. The trouble begins around Chapter 4, where the material shifts from local curve theory to surface theory proper. When you hit the second fundamental form, stop and make sure you understand what it represents before moving forward. It is not just a matrix you compute. It measures how the surface bends in ambient space relative to its normal vector. If your intuition here is fuzzy, everything after Chapter 5 will feel like memorization. The Gauss map is another sticking point. Students routinely confuse the differential of the Gauss map with the shape operator without noticing they are the same thing up to sign. I ran into this myself when I was first working through the material. The resolution was to write out the definition of the Weingarten map directly from the derivative of the normal and see that it matches the shape operator immediately. That single derivation removes a lot of later confusion about principal curvatures and directions.
Get the Full Details
For the geodesic problems, the key insight most people miss is that geodesics are critical points of arc length, not minimizers. The existence and uniqueness theorem gives you local geodesics, but global behavior is where things get interesting. The Hadamard theorem, for instance, relates the topology of a complete surface with negative curvature to the behavior of geodesics. The solution to that problem requires you to connect several earlier results, and the book does not hand it to you on a platter. Here is a specific edge case I ran into while checking solutions for a student. Problem 14 from Section 4-3 asks about the sphere and a certain curve involving the exponential map. The naive approach is to parameterize everything and compute directly, which leads to a mess of trigonometric identities. The correct path uses the fact that the exponential map at a point on the sphere maps radial lines to great circles. Recognizing this cuts the computation from a page of algebra to about four lines. Most published solutions I found online took the brute force route and produced answers that were correct but obscenely long. That is the kind of thing you notice only after you have spent enough time on the problem yourself to see the structure.
Common Pitfalls
One frequent mistake is treating local results as if they are global. The Gauss-Bonnet theorem has a local version and a global version, and Do Carmo separates them deliberately. Students often apply the global version to surfaces with boundary without checking the assumptions. The theorem requires compactness and a specific boundary term. Skip those checks and your answer is wrong even if your computation is perfect. Another pitfall appears in the Riemannian Geometry book, particularly around the Jacobi equation. People memorize the solution formula for constant sectional curvature and then apply it blindly to variable curvature situations. The general Jacobi equation is a second-order linear ODE along a geodesic, and its solutions depend on the curvature tensor evaluated along that geodesic. There is no closed form in general. If a problem seems to require one, you are probably overcomplicating it or looking at the wrong lemma.
What Solutions Cannot Do for You
A solution manual will never teach you how to read a proof the way Do Carmo expects. The book assumes you are comfortable with multivariable calculus, linear algebra at an abstract level, and basic topology. If any of those are shaky, working through solutions alone will not fix the gap. You will recognize the steps when you see them but be unable to reconstruct them yourself. The biggest limitation of available solutions online is that many are written by students who are one step ahead of you. They know how to get to the answer but may not explain why a particular approach was chosen. That gap between procedure and reasoning is where real learning happens, and no solution set can supply it for you. You have to sit with the problem long enough to feel the pressure points. If you find yourself stuck on a particular chapter for more than a week, switch to a different source temporarily. Loring Tu's An Introduction to Manifolds covers overlapping material with a different emphasis, and Spivak's A Comprehensive Introduction to Differential Geometry Volume 1 has incredible detail on the curve theory. Sometimes reading the same concept from a different angle is the only thing that makes it click. Do Carmo is elegant, but elegance can be opaque when you are encountering the material for the first time.
The solutions exist. Find the careful ones, use them to check your work rather than replace your work, and do not skip the proofs. The book rewards patience and punishes shortcuts, which is exactly how a text at this level should behave.
