float are 4 byte, so:
a[0]= float>>(3*8);
a[1]= float>>(2*8);
a[2]= float>>(1*8);
a[3]= float>>(0*8);
that means
a[0]= float>>(24);
a[1]= float>>(16);
a[2]= float>>(8);
a[3]= float;
to rebuild it (i think):
float=0;
for (int i=0; i < 4; i++){
float=float<<8;
float = a[i];
}