diff --git a/CGAL_Core/include/CGAL/CORE/BigFloat.h b/CGAL_Core/include/CGAL/CORE/BigFloat.h
index 5969c1b5e28..de9ed2f5508 100644
--- a/CGAL_Core/include/CGAL/CORE/BigFloat.h
+++ b/CGAL_Core/include/CGAL/CORE/BigFloat.h
@@ -513,32 +513,6 @@ inline long minStar(long m, long n) {
 }
 /// \name Functions for Compatibility with BigInt (needed by Poly, Curves)
 //@{
-/// isDivisible(a,b) = "is a divisible by b"
-/**         Assuming that a and  b are in coanonized forms.
-        Defined to be true if mantissa(b) | mantissa(a) &&
-        exp(b) = min*(exp(b), exp(a)).
- *      This concepts assume a and b are exact BigFloats.
- */
-inline bool isDivisible(const BigFloat& a, const BigFloat& b) {
-  // assert: a and b are exact BigFloats.
-  if (sign(a.m()) == 0) return true;
-  if (sign(b.m()) == 0) return false;
-  unsigned long bin_a = getBinExpo(a.m());
-  unsigned long bin_b = getBinExpo(b.m());
-
-  BigInt m_a = a.m() >> bin_a;
-  BigInt m_b = b.m() >> bin_b;
-  long e_a = bin_a + BigFloatRep::bits(a.exp());
-  long e_b = bin_b + BigFloatRep::bits(b.exp());
-  long dx = minStar(e_a, e_b);
-
-  return isDivisible(m_a, m_b) && (dx == e_b);
-}
-
-inline bool isDivisible(double x, double y) {
-  //Are these exact?
-  return isDivisible(BigFloat(x), BigFloat(y));
-}
 
 /// div_exact(x,y) returns the BigFloat quotient of x divided by y
 /**        This is defined only if isDivisible(x,y).
@@ -551,7 +525,6 @@ inline bool isDivisible(double x, double y) {
 // normalizing it then we get zero.
 inline BigFloat div_exact(const BigFloat& x, const BigFloat& y) {
   BigInt z;
-  CGAL_assertion (isDivisible(x,y));
   unsigned long bin_x = getBinExpo(x.m());
   unsigned long bin_y = getBinExpo(y.m());
 
diff --git a/CGAL_Core/include/CGAL/CORE/BigRat.h b/CGAL_Core/include/CGAL/CORE/BigRat.h
index df60c54b9bf..cc0c8d5ed17 100644
--- a/CGAL_Core/include/CGAL/CORE/BigRat.h
+++ b/CGAL_Core/include/CGAL/CORE/BigRat.h
@@ -67,20 +67,6 @@ namespace CORE {
   return BigRat(n,d);
   }
 
-  // Chee: 8/8/2004: need isDivisible to compile Polynomial<BigRat>
-  // A trivial implementation is to return true always. But this
-  // caused tPolyRat to fail.
-  // So we follow the definition of
-  // Expr::isDivisible(e1, e2) which checks if e1/e2 is an integer.
-  inline bool isInteger(const BigRat& x) {
-    return denominator(x) == 1; // AF: does that need canonicalize?
-  }
-  inline bool isDivisible(const BigRat& x, const BigRat& y) {
-    BigRat r;
-    r = x/y;
-    return isInteger(r);
-  }
-
   /// BigIntValue
 inline BigInt BigIntValue(const BigRat& br)
 {