diff --git a/ChangeLog b/ChangeLog
index 51502c9686..2d2375f762 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+Sat Feb  9 18:34:45 2008  Tanaka Akira  <akr@fsij.org>
+
+	* math.c (math_cbrt): new method Math.cbrt.
+
+	* configure.in (cbrt): check for replacement functions.
+
+	* missing/cbrt.c: new file.
+
 Sat Feb  9 17:51:24 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* ext/bigdecimal/bigdecimal.c (BigDecimal_to_f): use strtod() for more
diff --git a/configure.in b/configure.in
index 28e85f39ec..0ede6e4d4b 100644
--- a/configure.in
+++ b/configure.in
@@ -649,7 +649,7 @@ esac
 AC_FUNC_MEMCMP
 AC_REPLACE_FUNCS(dup2 memmove strerror strftime\
 		 strchr strstr crypt flock vsnprintf\
-		 isnan finite isinf hypot acosh erf tgamma lgamma_r \
+		 isnan finite isinf hypot acosh erf tgamma lgamma_r cbrt \
                  strlcpy strlcat)
 AC_CHECK_FUNCS(fmod killpg wait4 waitpid fork spawnv syscall chroot fsync getcwd eaccess\
 	      truncate chsize times utimes utimensat fcntl lockf lstat\
diff --git a/math.c b/math.c
index fd10e97008..1f78d98a55 100644
--- a/math.c
+++ b/math.c
@@ -400,6 +400,20 @@ math_sqrt(VALUE obj, VALUE x)
     return DOUBLE2NUM(d);
 }
 
+/*
+ *  call-seq:
+ *     Math.cbrt(numeric)    => float
+ *  
+ *  Returns the cube root of <i>numeric</i>.
+ */
+
+static VALUE
+math_cbrt(VALUE obj, VALUE x)
+{
+    Need_Float(x);
+    return DOUBLE2NUM(cbrt(RFLOAT_VALUE(x)));
+}
+
 /*
  *  call-seq:
  *     Math.frexp(numeric)    => [ fraction, exponent ]
@@ -611,6 +625,7 @@ Init_Math(void)
     rb_define_module_function(rb_mMath, "log2", math_log2, 1);
     rb_define_module_function(rb_mMath, "log10", math_log10, 1);
     rb_define_module_function(rb_mMath, "sqrt", math_sqrt, 1);
+    rb_define_module_function(rb_mMath, "cbrt", math_cbrt, 1);
 
     rb_define_module_function(rb_mMath, "frexp", math_frexp, 1);
     rb_define_module_function(rb_mMath, "ldexp", math_ldexp, 2);
diff --git a/missing/cbrt.c b/missing/cbrt.c
new file mode 100644
index 0000000000..54db2703a0
--- /dev/null
+++ b/missing/cbrt.c
@@ -0,0 +1,10 @@
+#include <math.h>
+
+double cbrt(double x)
+{
+    if (x < 0)
+        return -pow(-x, 1/3.0);
+    else
+        return pow(x, 1/3.0);
+}
+