By a curious historical confusion, two different systems for naming large numbers exist, one in the US and the other in Britain and other parts of the English-speaking world. This could be troublesome, but these days such large numbers are much more often given in unambiguous scientific notation, and the US usage is in any case coming to dominate, as it has almost completely with billion.
The original scheme, invented in France in the sixteenth century, started with million and multiplied 1 by that number the required number of times. The name of the unit was then based on the number of multiplications, using Latin numerals. So a sextillion was 1 multiplied by a million six times, making a number expressed by 1 followed by 36 zeroes (1036 in scientific notation).
In the eighteenth century French mathematicians changed to multiples of a thousand instead, but took over the existing number names; the Latin numbers then marked one less than the number of multiplications, so that trillion was 1 multiplied by a thousand four times. The US system was based on the thousands scheme, but the British stayed with the older millions one.
The following table gives some names and values in the two systems:
billion (Latin bi-, twice)
trillion (Latin tres, three)
quadrillion (Latin quattour, four)
quintillion (Latin quinque, five)
sextillion (Latin sex, six)
septillion (Latin septem, seven)
octillion (Latin octo, eight)
nonillion (Latin nonus, ninth)
decillion (Latin decem, ten)
Two parallel sets of prefixes for number multiples exist, one derived from Latin, the other from Greek. These appear widely in compounds but are no longer much used to create new words, the job having been largely passed to the SI method described below, especially in scientific usage.
The standard system of prefixes for multiples these days is that laid down in SI units (Système International D’Unités), an international agreement dating from 1960, which defines standard units for quantities and the names for the decimal prefixes to use with them. These are widely used.
The standard SI prefixes for multiples are:
Greek deka, ten
Greek hekaton, hundred
Greek khilioi, thousand
Greek megas, great
Greek gigas, giant
Greek teras, monster
Greek penta-, five
Based on Greek hexa-, six
Based on Latin septi‑, seven
Based on Latin octo‑, eight
The standard SI prefixes for fractional quantities are:
Latin decimus, a tenth
Latin centum, a hundred
Latin mille, thousand
Greek mikros, small
Greek nanos, dwarf
Spanish pico, literally a little bit
Danish or Norwegian femten, fifteen
Danish or Norwegian atten, eighteen
Adapted from septi-, seven, on the pattern of other multiples
Similarly adapted from octo-, eight
The prefixes hecto-, deca-, deci-, and centi- are generally avoided in scientific work.
The use of decimal prefixes to describe the similar — but not identical — binary multiples used in computing (such as megabyte or terabit) has caused confusion — as a result of various conventions, a megabyte can be 1,048,576 or 1,024,000 or 1,000,000.
In 1998 the International Electrotechnical Commission (IEC) agreed an international standard for a new group of prefixes that removes the ambiguity; names use the first two letters of the SI decimal prefix, followed by the letters bi, for binary. These are only slowly coming into use.