这里有一个例子:

#include <iostream>

void by_val(int arg) { arg += 2; }
void by_ref(int&arg) { arg += 2; }

int main()
{
    int x = 0;
    by_val(x); std::cout << x << std::endl;  // prints 0
    by_ref(x); std::cout << x << std::endl;  // prints 2

    int y = 0;
    by_ref(y); std::cout << y << std::endl;  // prints 2
    by_val(y); std::cout << y << std::endl;  // prints 2
}

It's a way how to pass arguments to functions. Passing by reference means the called functions' parameter will be the same as the callers' passed argument (not the value, but the identity - the variable itself). Pass by value means the called functions' parameter will be a copy of the callers' passed argument. The value will be the same, but the identity - the variable - is different. Thus changes to a parameter done by the called function in one case changes the argument passed and in the other case just changes the value of the parameter in the called function (which is only a copy). In a quick hurry:

Java only supports pass by value. Always copies arguments, even though when copying a reference to an object, the parameter in the called function will point to the same object and changes to that object will be see in the caller. Since this can be confusing, here is what Jon Skeet has to say about this. C# supports pass by value and pass by reference (keyword ref used at caller and called function). Jon Skeet also has a nice explanation of this here. C++ supports pass by value and pass by reference (reference parameter type used at called function). You will find an explanation of this below.

代码

因为我的语言是c++，我将在这里使用它

// passes a pointer (called reference in java) to an integer
void call_by_value(int *p) { // :1
    p = NULL;
}

// passes an integer
void call_by_value(int p) { // :2
    p = 42;
}

// passes an integer by reference
void call_by_reference(int & p) { // :3
    p = 42;
}

// this is the java style of passing references. NULL is called "null" there.
void call_by_value_special(int *p) { // :4
    *p = 10; // changes what p points to ("what p references" in java)
    // only changes the value of the parameter, but *not* of 
    // the argument passed by the caller. thus it's pass-by-value:
    p = NULL;
}

int main() {
    int value = 10;
    int * pointer = &value;

    call_by_value(pointer); // :1
    assert(pointer == &value); // pointer was copied

    call_by_value(value); // :2
    assert(value == 10); // value was copied

    call_by_reference(value); // :3
    assert(value == 42); // value was passed by reference

    call_by_value_special(pointer); // :4
    // pointer was copied but what pointer references was changed.
    assert(value == 10 && pointer == &value);
}

Java中的一个例子不会有什么坏处:

class Example {
    int value = 0;

    // similar to :4 case in the c++ example
    static void accept_reference(Example e) { // :1
        e.value++; // will change the referenced object
        e = null; // will only change the parameter
    }

    // similar to the :2 case in the c++ example
    static void accept_primitive(int v) { // :2
        v++; // will only change the parameter
    }        

    public static void main(String... args) {
        int value = 0;
        Example ref = new Example(); // reference

        // note what we pass is the reference, not the object. we can't 
        // pass objects. The reference is copied (pass-by-value).
        accept_reference(ref); // :1
        assert ref != null && ref.value == 1;

        // the primitive int variable is copied
        accept_primitive(value); // :2
        assert value == 0;
    }
}

维基百科

http://en.wikipedia.org/wiki/Pass_by_reference#Call_by_value

http://en.wikipedia.org/wiki/Pass_by_reference#Call_by_reference

这家伙说得很准:

http://javadude.com/articles/passbyvalue.htm

下面是一个例子，演示了通过值-指针值-引用传递之间的区别:

void swap_by_value(int a, int b){
    int temp;

    temp = a;
    a = b;
    b = temp;
}   
void swap_by_pointer(int *a, int *b){
    int temp;

    temp = *a;
    *a = *b;
    *b = temp;
}    
void swap_by_reference(int &a, int &b){
    int temp;

    temp = a;
    a = b;
    b = temp;
}

int main(void){
    int arg1 = 1, arg2 = 2;

    swap_by_value(arg1, arg2);
    cout << arg1 << " " << arg2 << endl;    //prints 1 2

    swap_by_pointer(&arg1, &arg2);
    cout << arg1 << " " << arg2 << endl;    //prints 2 1

    arg1 = 1;                               //reset values
    arg2 = 2;
    swap_by_reference(arg1, arg2);
    cout << arg1 << " " << arg2 << endl;    //prints 2 1
}

“参照传递”方法有一个重要的局限性。如果参数声明为引用传递(因此前面有&号)，其对应的实际参数必须是变量。

引用“按值传递”形式参数的实际参数一般可以是表达式，因此它不仅可以使用变量，还可以使用文字甚至函数调用的结果。

该函数不能将值放在变量以外的东西中。它不能将新值赋给文字或强制表达式更改其结果。

PS:你也可以在当前的线程中检查Dylan Beattie的答案，它用简单的语言解释了它。

“按值传递”将发送存储在指定变量中的数据的副本，“按引用传递”将发送到变量本身的直接链接。

因此，如果你通过引用传递一个变量，然后改变你传递给它的块中的变量，原始变量将被改变。如果只是按值传递，原始变量将不能被传递到的块所改变，但您将获得调用时它所包含的任何内容的副本。

简而言之，通过值传递的是它是什么，通过引用传递的是它在哪里。

如果你的值是VAR1 = "Happy Guy!"，你只会看到"Happy Guy!"如果VAR1变成了“Happy Gal!”，你就不会知道了。如果它是通过引用传递的，并且VAR1发生了变化，那么您就会这样做。

当通过引用传递时，基本上是传递一个指向变量的指针。通过值传递，即传递变量的副本。

在基本用法中，这通常意味着通过引用传递，对变量的更改将在调用方法中看到，而在通过值传递时则不会。