Difference bewtween these two definition of derived class constructor

[manojg]

I have one base class and one derived class.

class Base
{
  protected:
    int a, b;  
  public:
    Base() {}
    Base(int aa, int bb) { a = aa; b = bb; }
};

class Derived : public Base
{
  protected:
    int c
  public:
    Derived() {}
    Derived(int aa, int bb, int cc) { Base(aa, bb); c = cc; }
    // OR
    Derived(int aa, int bb, int cc) : Base(aa, bb), c(cc) {}
};

The two definition of the derived class constructor

(Derived(int aa, int bb, int cc) { Base(aa, bb); c = cc; } 
and 
Derived(int aa, int bb, int cc) : Base(aa, bb), c(cc) {}

looks identical but first one does not work.

So, I wounder what the difference between them?

Thanks.

[keskiverto]

The member variables are constructed before the body of the class constructor is evaluated.

This one uses initializer list:

Derived(int aa, int bb, int cc)
 : Base(aa, bb),  // these are constructor calls, initialization
   c(cc)
{
}

The other can be written more explicitly:

Derived(int aa, int bb, int cc)
 : Base(),
   c()
{
  Base(aa, bb); // error.
  // This probably creates a temporary unnamed Base object,
  // which is different from the Base-part of Derived

  c = cc;  // this is an assignment
}