Floating point numbers are represented using a sign, a mantissa, and an exponent. This is not unlike scientific notation. Very large and very small numbers can be represented but there is a limit to the precision.