How the short type is represented by the Visual C++ compiler?

Question

I know that most significant bit means the sign of a number in signed types. But I found one strange (for me) thing: if the number is negative and we use short type, this number will look like 0xffff####. How can this be? Short contains only 2 bytes and in 0xffff#### we see 4 whole bytes. Why do 16 more bits become one in binary representation. Explain me, please, how does it works.

For example,

short s = 0x8008;
printf("%x", s);

Output:
>>> ffff8008

Answer 1

As @Pete Becker says, the problem is implicit conversion to int. If you try the same thing with C++ iostreams you will get the output you expect though

#include <iostream>

int main() { 

    short foo = 0x8008;
    std::cout << std::hex << foo << "\n";

    short bar = -32760;
    std::cout << std::hex << bar << "\n";

}