diff --git a/README.md b/README.md
index 83e04d3fb..409ef7a61 100644
--- a/README.md
+++ b/README.md
@@ -105,9 +105,9 @@ sudo rm -rf /usr/local/share/meevax
 
 | Target Name        | Description
 |:-------------------|:--
-| `all` (default)    | Build shared-library `libmeevax.0.4.207.so` and executable `meevax`.
+| `all` (default)    | Build shared-library `libmeevax.0.4.208.so` and executable `meevax`.
 | `test`             | Test executable `meevax`.
-| `package`          | Generate debian package `meevax_0.4.207_amd64.deb`.
+| `package`          | Generate debian package `meevax_0.4.208_amd64.deb`.
 | `install`          | Copy files into `/usr/local` __(1)__.
 | `install.deb`      | `all` + `package` + `sudo apt install <meevax>.deb`
 | `safe-install.deb` | `all` + `test` + `package` + `sudo apt install <meevax>.deb`
@@ -122,7 +122,7 @@ __(1)__ Meevax installed by `make install` cannot be uninstalled by the system's
 ## Usage
 
 ```
-Meevax Lisp System, version 0.4.207
+Meevax Lisp System, version 0.4.208
 
 Usage: meevax [OPTION...] [FILE...]
 
diff --git a/VERSION b/VERSION
index 3e33edbc5..cab5b51e7 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-0.4.207
+0.4.208
diff --git a/include/meevax/kernel/exact_integer.hpp b/include/meevax/kernel/exact_integer.hpp
index 261666dd9..daba883ea 100644
--- a/include/meevax/kernel/exact_integer.hpp
+++ b/include/meevax/kernel/exact_integer.hpp
@@ -29,13 +29,15 @@ inline namespace kernel
   {
     mpz_t value;
 
-    explicit exact_integer() noexcept;
-
-    explicit exact_integer(mpz_t const) noexcept;
+    exact_integer() noexcept;
 
     exact_integer(exact_integer const&) noexcept;
 
-    explicit exact_integer(exact_integer &&) noexcept;
+    exact_integer(exact_integer &&) noexcept;
+
+    ~exact_integer();
+
+    explicit exact_integer(mpz_t const) noexcept;
 
     explicit exact_integer(int);
 
@@ -47,18 +49,12 @@ inline namespace kernel
 
     explicit exact_integer(external_representation const&, int = 0);
 
-    ~exact_integer();
-
     auto operator=(exact_integer const&) -> exact_integer &;
 
     auto operator=(exact_integer &&) noexcept -> exact_integer &;
 
     auto operator=(external_representation const&) -> exact_integer &;
 
-    auto swap(exact_integer &) noexcept -> void;
-
-    explicit operator bool() const;
-
     operator int() const;
 
     operator signed long() const;
@@ -68,6 +64,8 @@ inline namespace kernel
     explicit operator float() const;
 
     explicit operator double() const;
+
+    explicit operator bool() const;
   };
 
   auto operator ==(exact_integer const&, int const) -> bool;
diff --git a/include/meevax/kernel/number.hpp b/include/meevax/kernel/number.hpp
index 8e3957d7f..0363d66ff 100644
--- a/include/meevax/kernel/number.hpp
+++ b/include/meevax/kernel/number.hpp
@@ -195,9 +195,9 @@ inline namespace kernel
       }
       else if constexpr (std::is_same_v<std::decay_t<decltype(z)>, ratio>)
       {
-        if (auto i = exact_integer(mpq_denref(z.value)); i == 1)
+        if (z.denominator() == 1)
         {
-          return z.numerator();
+          return make(z.numerator());
         }
         else
         {
@@ -380,7 +380,7 @@ inline namespace kernel
       }
       else if constexpr (std::is_same_v<std::decay_t<T>, ratio>)
       {
-        return x.denominator().template as<exact_integer>() == 1;
+        return x.denominator() == 1;
       }
       else
       {
diff --git a/include/meevax/kernel/ratio.hpp b/include/meevax/kernel/ratio.hpp
index d4d97163d..1b6a04d5f 100644
--- a/include/meevax/kernel/ratio.hpp
+++ b/include/meevax/kernel/ratio.hpp
@@ -18,7 +18,6 @@
 #define INCLUDED_MEEVAX_KERNEL_RATIO_HPP
 
 #include <gmp.h>
-
 #include <meevax/kernel/pair.hpp>
 
 namespace meevax
@@ -35,6 +34,8 @@ inline namespace kernel
 
     ratio(ratio &&);
 
+    ~ratio();
+
     explicit ratio(exact_integer const&);
 
     explicit ratio(exact_integer const&, exact_integer const&);
@@ -43,11 +44,9 @@ inline namespace kernel
 
     explicit ratio(external_representation const&, int = 10);
 
-    ~ratio();
-
-    auto denominator() const -> value_type;
+    auto denominator() const -> exact_integer;
 
-    auto numerator() const -> value_type;
+    auto numerator() const -> exact_integer;
 
     explicit operator double() const;
   };
diff --git a/src/kernel/exact_integer.cpp b/src/kernel/exact_integer.cpp
index 62ab132cb..38e20a53d 100644
--- a/src/kernel/exact_integer.cpp
+++ b/src/kernel/exact_integer.cpp
@@ -31,11 +31,6 @@ inline namespace kernel
     mpz_init(value);
   }
 
-  exact_integer::exact_integer(mpz_t const z) noexcept
-  {
-    mpz_init_set(value, z);
-  }
-
   exact_integer::exact_integer(exact_integer const& other) noexcept
   {
     mpz_init_set(value, other.value);
@@ -47,6 +42,16 @@ inline namespace kernel
     mpz_swap(value, other.value);
   }
 
+  exact_integer::~exact_integer()
+  {
+    mpz_clear(value);
+  }
+
+  exact_integer::exact_integer(mpz_t const z) noexcept
+  {
+    mpz_init_set(value, z);
+  }
+
   exact_integer::exact_integer(int rhs)
     : exact_integer(static_cast<signed long>(rhs))
   {}
@@ -75,20 +80,15 @@ inline namespace kernel
     }
   }
 
-  exact_integer::~exact_integer()
-  {
-    mpz_clear(value);
-  }
-
   auto exact_integer::operator=(exact_integer const& rhs) -> exact_integer &
   {
     mpz_set(value, rhs.value);
     return *this;
   }
 
-  auto exact_integer::operator=(exact_integer && rhs) noexcept -> exact_integer &
+  auto exact_integer::operator=(exact_integer && other) noexcept -> exact_integer &
   {
-    swap(rhs);
+    mpz_swap(value, other.value);
     return *this;
   }
 
@@ -104,11 +104,6 @@ inline namespace kernel
     }
   }
 
-  auto exact_integer::swap(exact_integer & rhs) noexcept -> void
-  {
-    std::swap(*value, *rhs.value);
-  }
-
   exact_integer::operator bool() const
   {
     return (*value)._mp_size;
diff --git a/src/kernel/library.cpp b/src/kernel/library.cpp
index cd0cb8110..e4bb7750d 100644
--- a/src/kernel/library.cpp
+++ b/src/kernel/library.cpp
@@ -571,12 +571,12 @@ inline namespace kernel
 
       library.define<procedure>("ratio-numerator", [](let const& xs)
       {
-        return car(xs).as<ratio>().numerator();
+        return make(car(xs).as<ratio>().numerator());
       });
 
       library.define<procedure>("ratio-denominator", [](let const& xs)
       {
-        return car(xs).as<ratio>().denominator();
+        return make(car(xs).as<ratio>().denominator());
       });
 
       library.define<procedure>("floor", [](let const& xs)
diff --git a/src/kernel/ratio.cpp b/src/kernel/ratio.cpp
index 06a1192b1..fc5880337 100644
--- a/src/kernel/ratio.cpp
+++ b/src/kernel/ratio.cpp
@@ -63,14 +63,14 @@ inline namespace kernel
 
   ratio::ratio(external_representation const& token, int radix)
   {
-    std::regex static const pattern { "([+-]?[0-9a-f]+)/([0-9a-f]+)" };
+    // std::regex static const pattern { "([+-]?[0-9a-f]+)/([0-9a-f]+)" };
 
-    if (mpq_init(value); not std::regex_match(token, pattern) or mpq_set_str(value, token.c_str(), radix))
+    if (mpq_init(value); mpq_set_str(value, token.c_str(), radix))
     {
       mpq_clear(value);
       throw error();
     }
-    else // TEMPORARY!!!
+    else
     {
       mpq_canonicalize(value);
     }
@@ -81,19 +81,19 @@ inline namespace kernel
     mpq_clear(value);
   }
 
-  auto ratio::denominator() const -> value_type
+  auto ratio::denominator() const -> exact_integer
   {
-    return make<exact_integer>(mpq_denref(value));
+    return exact_integer(mpq_denref(value));
   }
 
-  auto ratio::numerator() const -> value_type
+  auto ratio::numerator() const -> exact_integer
   {
-    return make<exact_integer>(mpq_numref(value));
+    return exact_integer(mpq_numref(value));
   }
 
   ratio::operator double() const
   {
-    return static_cast<double>(numerator().as<exact_integer>()) / static_cast<double>(denominator().as<exact_integer>());
+    return mpq_get_d(value);
   }
 
   auto operator <<(std::ostream & os, ratio const& datum) -> std::ostream &