The C library function scanf() reads formatted input from the keyboard and stores each value at a memory location. Two things must be correct for it to work: the format specifier must match the variable's data type, and scanf must be given the address (a pointer) of the variable, not the variable's value, because scanf writes the input into that memory location.

The correct statement is:

• %f — the proper format specifier for a float in scanf.
• &x — the address-of operator supplies the location of x, so scanf knows where to store the value it reads.

Why the other statements fail:

• Using %d with &x tells scanf to interpret the input as an integer and write an integer-sized bit pattern into a float variable, giving a wrong/garbage result.
• Writing *x dereferences x as if it were a pointer, but x is a plain float, not a pointer — this is a type error and leads to undefined behaviour.
• Passing x without the & hands scanf the current (uninitialised) value of x instead of its address, so scanf tries to store data at an invalid location — again undefined behaviour, often a crash.