diff --git a/CGAL_Core/include/CGAL/CORE/BigFloat_impl.h b/CGAL_Core/include/CGAL/CORE/BigFloat_impl.h
index b6eb9054c8e..7bf616e8fda 100644
--- a/CGAL_Core/include/CGAL/CORE/BigFloat_impl.h
+++ b/CGAL_Core/include/CGAL/CORE/BigFloat_impl.h
@@ -214,7 +214,7 @@ CGAL_INLINE_FUNCTION
 void BigFloatRep::approx(const BigFloatRep& B,
               const extLong& r, const extLong& a) {
   if (B.err) {
-    if (1 + clLg(B.err) <= bitLength(B.m))
+    if (static_cast<unsigned>(1 + clLg(B.err)) <= bitLength(B.m))
       truncM(B, r + 1, a);
     else //  1 + clLg(B.err) > lg(B.m)
       truncM(B, CORE_posInfty, a);
diff --git a/CGAL_Core/include/CGAL/CORE/BigInt.h b/CGAL_Core/include/CGAL/CORE/BigInt.h
index 4dd1f1e8ecd..a4b17f28937 100644
--- a/CGAL_Core/include/CGAL/CORE/BigInt.h
+++ b/CGAL_Core/include/CGAL/CORE/BigInt.h
@@ -87,15 +87,15 @@ inline bool isDivisible(long x, long y) {
 
 }
   /// get exponent of power 2
-inline auto getBinExpo(const BigInt& z) {
+inline unsigned getBinExpo(const BigInt& z) {
     if (z.is_zero()) {
-        return (std::numeric_limits<unsigned long>::max)();
+        return (std::numeric_limits<unsigned>::max)();
     }
     return lsb(abs(z));
 }
 
   // bit length
-inline auto bitLength(const BigInt& a) {
+inline unsigned bitLength(const BigInt& a){
     if (a.is_zero()) {
         return 0;
     }
diff --git a/Number_types/include/CGAL/CORE_BigFloat.h b/Number_types/include/CGAL/CORE_BigFloat.h
index 81e9fcbaca6..b7da346a31b 100644
--- a/Number_types/include/CGAL/CORE_BigFloat.h
+++ b/Number_types/include/CGAL/CORE_BigFloat.h
@@ -178,7 +178,7 @@ public:
 
             // shift such that err.m()+err.err() fits into long
             int digits_long = std::numeric_limits<long>::digits;
-            if(::CORE::bitLength(err.m()+err.err()) >= digits_long){
+            if(::CORE::bitLength(err.m()+err.err()) >= static_cast<unsigned>(digits_long)){
                 long shift = ::CORE::bitLength(err.m()) - digits_long + 1 ;
                 //std::cout << "shift " << shift<< std::endl;
                 CORE::BigInt bi = (err.m() + err.err());