diff --git a/closed_ricci_flow_curvature_blowup_v2.txt b/closed_ricci_flow_curvature_blowup_v2.txt
index d3da9b8..5f8275f 100644
--- a/closed_ricci_flow_curvature_blowup_v2.txt
+++ b/closed_ricci_flow_curvature_blowup_v2.txt
@@ -7,35 +7,8 @@
 ================================================================================
 
 
-0.  HOW TO READ THIS
 
 
-Sections 1 to 3 need only calculus, vectors, and the idea of an average. They
-contain the whole story: what is being asked, why the obvious guess is wrong,
-and what the answer is.
-
-Sections 4 to 9 are the proof. Every step of the argument is stated. Most are
-computations you can check by differentiating. A few use a standard tool from
-analysis -- a heat kernel estimate, a maximum principle, Shi's derivative
-estimate -- and each of those is named at the moment it is used, with the
-precise place where it is proved. When this document says "imported", that is
-exactly what it means: proved elsewhere, cited exactly, not re-proved here.
-That is the only kind of gap in this document, and it is never hidden.
-
-Section 10 says honestly what is new and what is borrowed. Section 11 is a map
-of which lemma feeds which. Section 12 is a glossary of every symbol. Section
-13 is a list of things you can test yourself on.
-
-Two habits make the proof much easier to follow.
-
- (a) Almost every estimate below has the same shape: "something is at most a
-     constant times a power of r, or a power of delta". Watch the powers. The
-     powers are the plot.
-
- (b) There are exactly two length scales in the problem, sqrt(delta) and
-     theta, with theta much smaller. Almost every lemma says which of the two
-     it is about. Mixing them up is the main way to get lost.
-
 
 1.  THE QUESTION