diff --git a/Number_types/include/CGAL/CORE_coercion_traits.h b/Number_types/include/CGAL/CORE_coercion_traits.h
index 1fb9392675d..dd636bbf58b 100644
--- a/Number_types/include/CGAL/CORE_coercion_traits.h
+++ b/Number_types/include/CGAL/CORE_coercion_traits.h
@@ -84,7 +84,7 @@ struct Coercion_traits<CORE::BigFloat , ::CORE::BigInt>{
         Type operator()(const CORE::BigFloat& x)  const { return x;}
         Type operator()(const ::CORE::BigInt x) const {
             CORE::BigFloat result;
-            result.approx(x,CORE::defRelPrec.toLong(),1024);
+            result.approx(x,CORE::defRelPrec.toLong(),LONG_MAX);
             // Do not use MakeFloorExact as it changes the Bigfloat
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()-result.err(),0,result.exp())) <= x );
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()+result.err(),0,result.exp())) >= x );
@@ -104,7 +104,7 @@ struct Coercion_traits<CORE::BigFloat , ::CORE::BigRat>{
         Type operator()(const CORE::BigFloat& x)  const { return x;}
         Type operator()(const ::CORE::BigRat x) const {
             
-            CORE::BigFloat result(x,CORE::defRelPrec.toLong(),1024);
+	  CORE::BigFloat result(x,CORE::defRelPrec.toLong(),LONG_MAX);
             // Do not use MakeFloorExact as it changes the Bigfloat
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()-result.err(),0,result.exp())) <= x );
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()+result.err(),0,result.exp())) >= x );
@@ -123,7 +123,7 @@ struct Coercion_traits<CORE::BigFloat , ::CORE::Expr>{
         typedef Type result_type;
         Type operator()(const CORE::BigFloat& x)  const { return x;}
         Type operator()(const ::CORE::Expr x) const {
-            CORE::BigFloat result(x, CORE::defRelPrec.toLong(),1024);
+            CORE::BigFloat result(x, CORE::defRelPrec.toLong(),LONG_MAX);
             // Do not use MakeFloorExact as it changes the Bigfloat
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()-result.err(),0,result.exp())) <= x );
             CGAL_postcondition( ::CORE::BigRat(::CORE::BigFloat(result.m()+result.err(),0,result.exp())) >= x );