diff --git a/source/nanoFramework.CoreLibrary/CoreLibrary.nfproj b/source/nanoFramework.CoreLibrary/CoreLibrary.nfproj
index 5ba082e4..bc856d7c 100644
--- a/source/nanoFramework.CoreLibrary/CoreLibrary.nfproj
+++ b/source/nanoFramework.CoreLibrary/CoreLibrary.nfproj
@@ -45,6 +45,7 @@
   </PropertyGroup>
   <Import Project="$(NanoFrameworkProjectSystemPath)NFProjectSystem.props" Condition="Exists('$(NanoFrameworkProjectSystemPath)NFProjectSystem.props')" />
   <ItemGroup>
+    <Compile Include="Friends.cs" />
     <Compile Include="System\AppDomain.cs" />
     <Compile Include="System\AppDomainUnloadedException.cs" />
     <Compile Include="System\ApplicationException.cs" />
@@ -379,4 +380,4 @@
       <ProjectConfigurationsDeclaredAsItems />
     </ProjectCapabilities>
   </ProjectExtensions>
-</Project>
+</Project>
\ No newline at end of file
diff --git a/source/nanoFramework.CoreLibrary/Friends.cs b/source/nanoFramework.CoreLibrary/Friends.cs
new file mode 100644
index 00000000..2f8c5f06
--- /dev/null
+++ b/source/nanoFramework.CoreLibrary/Friends.cs
@@ -0,0 +1,8 @@
+//
+// Copyright (c) 2017 The nanoFramework project contributors
+// See LICENSE file in the project root for full license information.
+//
+
+using System.Runtime.CompilerServices;
+
+[assembly: System.Runtime.CompilerServices.InternalsVisibleTo("System.Math, PublicKey=00240000048000009400000006020000002400005253413100040000010001001120aa3e809b3da4f65e1b1f65c0a3a1bf6335c39860ca41acb3c48de278c6b63c5df38239ec1f2e32d58cb897c8c174a5f8e78a9c0b6087d3aef373d7d0f3d9be67700fc2a5a38de1fb71b5b6f6046d841ff35abee2e0b0840a6291a312be184eb311baff5fef0ff6895b9a5f2253aed32fb06b819134f6bb9d531488a87ea2")]
diff --git a/source/nanoFramework.CoreLibrary/System/Math.cs b/source/nanoFramework.CoreLibrary/System/Math.cs
index bf707f3e..7c38ab2d 100644
--- a/source/nanoFramework.CoreLibrary/System/Math.cs
+++ b/source/nanoFramework.CoreLibrary/System/Math.cs
@@ -8,129 +8,15 @@ namespace System
 {
     using Runtime.CompilerServices;
 
-    /// <summary>
-    /// Provides constants and static methods for trigonometric, logarithmic, and other common mathematical functions.
-    /// </summary>
-    /// <remarks>
-    /// Specific for nanoFramework: this class library is split between mscorlib and it's own assembly.
-    /// If you require any of the other methods that are not available here add the NuGet package nanoFramework.System.Math.
-    /// </remarks>
-    public static partial class Math
+    internal static class MathInternal
     {
-        /// <summary>
-        /// Represents the ratio of the circumference of a circle to its diameter, specified by the constant, π.
-        /// </summary>
-        /// <remarks>The value of this field is 3.14159265358979323846.</remarks>
-        public const double PI = 3.14159265358979323846;
-        /// <summary>
-        /// Represents the natural logarithmic base, specified by the constant, e.
-        /// </summary>
-        /// <remarks>The value of this field is 2.7182818284590452354.</remarks>
-        public const double E = 2.7182818284590452354;
-
-        /// <summary>
-        /// Returns the absolute value of a 32-bit signed integer.
-        /// </summary>
-        /// <param name="val">A number that is greater than Int32..::..MinValue, but less than or equal to Int32..::..MaxValue.</param>
-        /// <returns>A 32-bit signed integer, x, such that 0 ≤ x ≤ Int32..::..MaxValue.</returns>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern int Abs(int val);
-
-        /// <summary>
-        /// Returns the absolute value of a double-precision floating-point number.
-        /// </summary>
-        /// <param name="val">A number that is greater than or equal to Double..::..MinValue, but less than or equal to Double..::..MaxValue.</param>
-        /// <returns>A double-precision floating-point number, x, such that 0 ≤ x ≤ Double..::..MaxValue.</returns>
-        /// <remarks>
-        /// This method with double-precision floating-point parameter might not be available in all platforms or with firmware images that where build with single point FPU option.
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern double Abs(double val);
-
-        /// <summary>
-        /// Returns the absolute value of a single-precision floating-point number.
-        /// </summary>
-        /// <param name="val">A number that is greater than or equal to Double..::..MinValue, but less than or equal to Double..::..MaxValue.</param>
-        /// <returns>A single-precision floating-point number, x, such that 0 ≤ x ≤ Double..::..MaxValue.</returns>
-        /// <remarks>
-        /// This method with single-precision floating-point parameter is exclusive of nanoFramework. It doesn't exist in the .NET API, only the double-precision floating-point version.
-        /// It might not be available in all platforms or with firmware images that where build with double point FPU option. 
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern float Abs(float val);
-
-        /// <summary>
-        /// Returns the larger of two 32-bit signed integers.
-        /// </summary>
-        /// <param name="val1">The first of two 32-bit signed integers to compare. </param>
-        /// <param name="val2">The second of two 32-bit signed integers to compare. </param>
-        /// <returns>Parameter val1 or val2, whichever is larger.</returns>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern int Max(int val1, int val2);
-
-        /// <summary>
-        /// Returns the larger of two double-precision floating-point numbers.
-        /// </summary>
-        /// <param name="x">The first of two double-precision floating-point numbers to compare. </param>
-        /// <param name="y">The second of two double-precision floating-point numbers to compare. </param>
-        /// <returns>Parameter x or y, whichever is larger. If x, y, or both x and y are equal to NaN, NaN is returned.</returns>
-        /// <remarks>
-        /// This method with double-precision floating-point parameter might not be available in all platforms or with firmware images that where build with single point FPU option.
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
         [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern double Max(double x, double y);
-        
-        /// <summary>
-        /// Returns the larger of two single-precision floating-point numbers.
-        /// </summary>
-        /// <param name="x">The first of two single-precision floating-point numbers to compare. </param>
-        /// <param name="y">The second of two single-precision floating-point numbers to compare. </param>
-        /// <returns>Parameter x or y, whichever is larger. If x, y, or both x and y are equal to NaN, NaN is returned.</returns>
-        /// <remarks>
-        /// This method with single-precision floating-point parameter is exclusive of nanoFramework. It doesn't exist in the .NET API, only the double-precision floating-point version.
-        /// It might not be available in all platforms or with firmware images that where build with double point FPU option. 
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern float Max(float x, float y);
+        internal static extern int Abs(int val);
 
-        /// <summary>
-        /// Returns the smaller of two 32-bit signed integers.
-        /// </summary>
-        /// <param name="val1">The first of two 32-bit signed integers to compare. </param>
-        /// <param name="val2">The second of two 32-bit signed integers to compare. </param>
-        /// <returns>Parameter val1 or val2, whichever is smaller.</returns>
         [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern int Min(int val1, int val2);
+        internal static extern int Min(int val1, int val2);
 
-        /// <summary>
-        /// Returns the smaller of two double-precision floating-point numbers.
-        /// </summary>
-        /// <param name="x">The first of two double-precision floating-point numbers to compare. </param>
-        /// <param name="y">The second of two double-precision floating-point numbers to compare. </param>
-        /// <returns>Parameter x or y, whichever is smaller. If x, y, or both x and y are equal to NaN, NaN is returned.</returns>
-        /// <remarks>
-        /// This method with double-precision floating-point parameter might not be available in all platforms or with firmware images that where build with single point FPU option.
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
-        [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern double Min(double x, double y);
-        /// <summary>
-        /// Returns the smaller of two single-precision floating-point numbers.
-        /// </summary>
-        /// <param name="x">The first of two single-precision floating-point numbers to compare. </param>
-        /// <param name="y">The second of two single-precision floating-point numbers to compare. </param>
-        /// <returns>Parameter x or y, whichever is smaller. If x, y, or both x and y are equal to NaN, NaN is returned.</returns>
-        /// <remarks>
-        /// This method with single-precision floating-point parameter is exclusive of nanoFramework. It doesn't exist in the .NET API, only the double-precision floating-point version.
-        /// It might not be available in all platforms or with firmware images that where build with double point FPU option. 
-        /// </remarks>
-        /// <exception cref="NotImplementedException"/>
         [MethodImpl(MethodImplOptions.InternalCall)]
-        public static extern float Min(float x, float y);
-
+        internal static extern int Max(int val1, int val2);
     }
 }
diff --git a/source/nanoFramework.CoreLibrary/System/String.cs b/source/nanoFramework.CoreLibrary/System/String.cs
index 92b4d212..ab5579ec 100644
--- a/source/nanoFramework.CoreLibrary/System/String.cs
+++ b/source/nanoFramework.CoreLibrary/System/String.cs
@@ -701,7 +701,7 @@ public static string Format(string format, params object[] args)
                     }
                     else if (alignment < 0)
                     {
-                        output += token.PadRight(Math.Abs(alignment));
+                        output += token.PadRight(MathInternal.Abs(alignment));
                     }
                     else
                     {
diff --git a/source/nanoFramework.CoreLibrary/System/Text/StringBuilder.cs b/source/nanoFramework.CoreLibrary/System/Text/StringBuilder.cs
index 1bd9fdbb..ae4adb71 100644
--- a/source/nanoFramework.CoreLibrary/System/Text/StringBuilder.cs
+++ b/source/nanoFramework.CoreLibrary/System/Text/StringBuilder.cs
@@ -222,7 +222,7 @@ public StringBuilder(int capacity, int maxCapacity)
             if (capacity > maxCapacity) throw new ArgumentOutOfRangeException("capacity");
             if (maxCapacity < 1) throw new ArgumentOutOfRangeException("maxCapacity");
             if (capacity < 0) throw new ArgumentOutOfRangeException("capacity");
-            if (capacity == 0) capacity = Math.Min(0x10, maxCapacity);
+            if (capacity == 0) capacity = MathInternal.Min(0x10, maxCapacity);
             _maxCapacity = maxCapacity;
             _chunkChars = new char[capacity];
         }
@@ -711,8 +711,8 @@ public StringBuilder Replace(char oldChar, char newChar, int startIndex, int cou
             var num4 = startIndex - chunkPrevious._chunkOffset;
             if (num3 >= 0)
             {
-                var index = Math.Max(num4, 0);
-                var num6 = Math.Min(chunkPrevious._chunkLength, num3);
+                var index = MathInternal.Max(num4, 0);
+                var num6 = MathInternal.Min(chunkPrevious._chunkLength, num3);
                 while (index < num6)
                 {
                     if (chunkPrevious._chunkChars[index] == oldChar)
@@ -940,7 +940,7 @@ private void ReplaceInPlaceAtChunk(ref StringBuilder chunk, ref int indexInChunk
             while (true)
             {
                 //int num = chunk.m_ChunkLength - indexInChunk;
-                var length = Math.Min(chunk._chunkLength - indexInChunk, count);
+                var length = MathInternal.Min(chunk._chunkLength - indexInChunk, count);
                 //ThreadSafeCopy(value, ref valueIndex, chunk.m_ChunkChars, ref indexInChunk, num2);
                 Array.Copy(value, valueIndex, chunk._chunkChars, indexInChunk, length);
                 indexInChunk += length;
@@ -977,9 +977,9 @@ internal void MakeRoom(int index, int count, out StringBuilder chunk, out int in
             }
             else
             {
-                var builder = new StringBuilder(Math.Max(count, 0x10), chunk._maxCapacity, chunk._chunkPrevious);
+                var builder = new StringBuilder(MathInternal.Max(count, 0x10), chunk._maxCapacity, chunk._chunkPrevious);
                 builder._chunkLength = count;
-                var length = Math.Min(count, indexInChunk);
+                var length = MathInternal.Min(count, indexInChunk);
                 if (length > 0)
                 {
                     Array.Copy(chunk._chunkChars, 0, builder._chunkChars, 0, length);
@@ -1016,7 +1016,7 @@ internal void AppendHelper(ref string value)
         internal void ExpandByABlock(int minBlockCharCount)
         {
             if (minBlockCharCount + Length > _maxCapacity) throw new ArgumentOutOfRangeException("requiredLength");
-            var num = Math.Max(minBlockCharCount, Math.Min(Length, 0x1f40));
+            var num = MathInternal.Max(minBlockCharCount, MathInternal.Min(Length, 0x1f40));
             _chunkPrevious = new StringBuilder(this);
             _chunkOffset += _chunkLength;
             _chunkLength = 0;