diff --git a/include/picrin.h b/include/picrin.h
index 0d250f7b..a2fafe9e 100644
--- a/include/picrin.h
+++ b/include/picrin.h
@@ -33,6 +33,7 @@ extern "C" {
 #include <setjmp.h>
 #include <stdio.h>
 #include <stdint.h>
+#include <limits.h>
 #include <assert.h>
 
 #include "xvect/xvect.h"
diff --git a/include/picrin/value.h b/include/picrin/value.h
index 69bf2139..6fa2b75d 100644
--- a/include/picrin/value.h
+++ b/include/picrin/value.h
@@ -156,6 +156,8 @@ typedef struct pic_blob pic_blob;
 static inline enum pic_tt pic_type(pic_value);
 static inline const char *pic_type_repr(enum pic_tt);
 
+static inline bool pic_valid_int(double);
+
 static inline pic_value pic_nil_value();
 static inline pic_value pic_true_value();
 static inline pic_value pic_false_value();
@@ -256,6 +258,12 @@ pic_type_repr(enum pic_tt tt)
   return 0;                     /* logic flaw */
 }
 
+static inline bool
+pic_valid_int(double v)
+{
+  return INT_MIN <= v && v <= INT_MAX;
+}
+
 static inline pic_value
 pic_nil_value()
 {
diff --git a/src/number.c b/src/number.c
index a3e201f4..e2f4ebb0 100644
--- a/src/number.c
+++ b/src/number.c
@@ -20,10 +20,10 @@ gcd(int a, int b)
   return b;
 }
 
-static int
+static double
 lcm(int a, int b)
 {
-  return abs(a * b) / gcd(a, b);
+  return fabs((double)a * b) / gcd(a, b);
 }
 
 static pic_value
@@ -439,7 +439,7 @@ pic_number_lcm(pic_state *pic)
 {
   size_t argc;
   pic_value *args;
-  int r;
+  double r;
   bool e = true;
 
   pic_get_args(pic, "*", &argc, &args);
@@ -457,7 +457,7 @@ pic_number_lcm(pic_state *pic)
       pic_error(pic, "lcm: number required");
     }
   }
-  return e ? pic_int_value(r) : pic_float_value(r);
+  return e && pic_valid_int(r) ? pic_int_value(r) : pic_float_value(r);
 }
 
 static pic_value