diff --git a/internal/ceres/cubic_interpolation.cc b/internal/ceres/cubic_interpolation.cc
index b202579..764b306 100644
--- a/internal/ceres/cubic_interpolation.cc
+++ b/internal/ceres/cubic_interpolation.cc
@@ -46,14 +46,13 @@
 // Evaluate the cubic Hermite spline (also known as the Catmull-Rom
 // spline) at a point x that lies in the interval [0, 1].
 //
-// This is also the interpolation kernel proposed by R. Keys, in:
+// This is also the interpolation kernel (for the case of a = 0.5) as
+// proposed by R. Keys, in:
 //
 // "Cubic convolution interpolation for digital image processing".
 // IEEE Transactions on Acoustics, Speech, and Signal Processing
 // 29 (6): 1153–1160.
 //
-// For the case of a = -0.5.
-//
 // For more details see
 //
 // http://en.wikipedia.org/wiki/Cubic_Hermite_spline
@@ -190,7 +189,7 @@
   const double p23 = (col < num_cols_ - 2) ? v(row + 1, col + 2) : 2 * p22 - p21;  // NOLINT
 
   // The four corners of the block require a bit more care.  Let us
-  // consider the evaluation of p00, the other four corners follow in
+  // consider the evaluation of p00, the other three corners follow in
   // the same manner.
   //
   // There are four cases in which we need to evaluate p00.
diff --git a/internal/ceres/cubic_interpolation_test.cc b/internal/ceres/cubic_interpolation_test.cc
index d3b5b49..0b4bb12 100644
--- a/internal/ceres/cubic_interpolation_test.cc
+++ b/internal/ceres/cubic_interpolation_test.cc
@@ -108,7 +108,7 @@
   EXPECT_TRUE(interpolator.Evaluate(x, &f, &dfdx));
 
   // Create a Jet with the same scalar part as x, so that the output
-  // Jet will be evaluate at x.
+  // Jet will be evaluated at x.
   Jet<double, 4> x_jet;
   x_jet.a = x;
   x_jet.v(0) = 1.0;
@@ -272,7 +272,7 @@
   EXPECT_TRUE(interpolator.Evaluate(r, c, &f, &dfdr, &dfdc));
 
   // Create a Jet with the same scalar part as x, so that the output
-  // Jet will be evaluate at x.
+  // Jet will be evaluated at x.
   Jet<double, 4> r_jet;
   r_jet.a = r;
   r_jet.v(0) = 1.0;