diff --git a/constants/52a.md b/constants/52a.md
index 796250a..63173b7 100644
--- a/constants/52a.md
+++ b/constants/52a.md
@@ -32,11 +32,11 @@ $\liminf_{n \to \infty} r_{3,n}$.
 | 4.758 | [KMPS1995] ||
 | 4.643 | [DB1997]   ||
 | 4.602 | [KKKS1998] ||
-| 4.506 | [DBM2000]  ||
 | 4.596 | [JSV2000]  ||
 | 4.571 | [KKSVZ2007]||
-| 4.453 | [MS2008]   |Under an extra hypothesis,<br> see Additional Comments (1) below  |
-| 4.490 | [DKMP2009] ||
+| 4.506 | [DBM2000]  ||
+| 4.490 | [DKMP2009] | Unconditional record among the tabulated upper bounds |
+| 4.453 | [MS2008]   |Under an extra hypothesis,<br> see Additional Comments (1) below. Not an unconditional record.  |
 
 
 
@@ -55,7 +55,7 @@ $\liminf_{n \to \infty} r_{3,n}$.
 
 
 ## Additional comments
-1. The extra hypothesis used in [MS2008] is about the satisfying assignments of formulas with density below and close to the threshold. 
+1. The extra hypothesis used in [MS2008] is about the satisfying assignments of formulas with density below and close to the threshold. That 4.453 bound is therefore listed after the unconditional 4.490 record.
 
 2. In [CF1986], the probability of satisfiability is shown to be only a positive constant. However, this, by Friedgut's result of 1999, implies that the probability is actually 1.