diff --git a/traits/src/int.rs b/traits/src/int.rs
index ec88b6d..4f9221f 100644
--- a/traits/src/int.rs
+++ b/traits/src/int.rs
@@ -1,5 +1,4 @@
 use std::ops::{Not, BitAnd, BitOr, BitXor, Shl, Shr};
-use std::num::Wrapping;
 
 use {Num, NumCast};
 use bounds::Bounded;
@@ -375,106 +374,3 @@ prim_int_impl!(i16,   i16,   u16);
 prim_int_impl!(i32,   i32,   u32);
 prim_int_impl!(i64,   i64,   u64);
 prim_int_impl!(isize, isize, usize);
-
-// While this might violate the possible assumption that
-// PrimInt represents genuine primitive integer types,
-// Wrapping<T> is _practically_ one too, so this
-// shouldn't be a concern.
-impl<T: PrimInt> PrimInt for Wrapping<T> 
-    where Wrapping<T>:
-          Num + NumCast
-        + Bounded
-        + Not<Output=Wrapping<T>>
-        + BitAnd<Output=Wrapping<T>>
-        + BitOr<Output=Wrapping<T>>
-        + BitXor<Output=Wrapping<T>>
-        + Shl<usize, Output=Wrapping<T>>
-        + Shr<usize, Output=Wrapping<T>>
-        + CheckedAdd<Output=Wrapping<T>>
-        + CheckedSub<Output=Wrapping<T>>
-        + CheckedMul<Output=Wrapping<T>>
-        + CheckedDiv<Output=Wrapping<T>>
-        + Saturating
-{
-    #[inline]
-    fn count_ones(self) -> u32 {
-        self.0.count_ones()
-    }
-
-    #[inline]
-    fn count_zeros(self) -> u32 {
-        self.0.count_zeros()
-    }
-
-    #[inline]
-    fn leading_zeros(self) -> u32 {
-        self.0.leading_zeros()
-    }
-
-    #[inline]
-    fn trailing_zeros(self) -> u32 {
-        self.0.trailing_zeros()
-    }
-
-    #[inline]
-    fn rotate_left(self, n: u32) -> Self {
-        Wrapping(self.0.rotate_left(n))
-    }
-
-    #[inline]
-    fn rotate_right(self, n: u32) -> Self {
-        Wrapping(self.0.rotate_right(n))
-    }
-
-    #[inline]
-    fn signed_shl(self, n: u32) -> Self {
-        Wrapping(self.0.signed_shl(n))
-    }
-
-    #[inline]
-    fn signed_shr(self, n: u32) -> Self {
-        Wrapping(self.0.signed_shr(n))
-    }
-
-    #[inline]
-    fn unsigned_shl(self, n: u32) -> Self {
-        Wrapping(self.0.unsigned_shl(n))
-    }
-
-    #[inline]
-    fn unsigned_shr(self, n: u32) -> Self {
-        Wrapping(self.0.unsigned_shr(n))
-    }
-
-    #[inline]
-    fn swap_bytes(self) -> Self {
-        Wrapping(self.0.swap_bytes())
-    }
-
-    #[inline]
-    fn from_be(x: Self) -> Self {
-        Wrapping(T::from_be(x.0))
-    }
-
-    #[inline]
-    fn from_le(x: Self) -> Self {
-        Wrapping(T::from_le(x.0))
-    }
-
-    #[inline]
-    fn to_be(self) -> Self {
-        Wrapping(self.0.to_be())
-    }
-
-    #[inline]
-    fn to_le(self) -> Self {
-        Wrapping(self.0.to_le())
-    }
-
-    // This is (or should be ?) fine because Wrapping<T> only guarantees that
-    // _standard_ operations (Add, Mul, etc) have wrapping semantics.
-    #[inline]
-    fn pow(self, exp: u32) -> Self {
-        Wrapping(self.0.pow(exp))
-    }
-}