In my first two posts about the SI unit prefixes, I described how the system originated in the French Republican metric system of 1795. Part 1 dealt with those original fractional prefixes—deci-, centi- and milli-, designating a tenth, hundredth and thousandth part of the base unit. Part 2 dealt with the multipliers—deca-, hecto-, kilo- and myria-, for tenfold, hundredfold, thousandfold and ten-thousandfold multiplication.
Inspired though the original system was, it didn’t offer enough range for scientific use, and the collection of prefixes has been steadily growing. There was also a realization that there didn’t need to be a prefix for every power of ten—that would get unwieldy very quickly. In the last century we’ve added prefixes only for integer powers of one thousand.
The next two prefixes were introduced as part of the definition of the electrical units of measurement, by a committee set up in 1861 by the British Association for the Advancement of Science, under the leadership of Lord Kelvin and James Clerk Maxwell. The new prefixes were mega-, for a millionfold multiplication, and micro-, for a millionth part of the base unit. This created an abbreviation crisis—there were now three prefixes beginning with “m”. Mega- was distinguished from milli- by the use of its capital initial (“M”) establishing a precedent for all the multiplier prefixes above kilo-. But for micro-, the Greek lower case letter mu (μ) had to be recruited. In future, care would be taken to avoid producing more than two prefixes with the same initial letter.
Both mega- and micro- are Greek, which upsets the original system of using Latin for fractions and Greek for multiples. Mega- comes from megas, “big”, and micro- from micros, “small”. And both prefixes had already been doing duty in the scientific vocabulary to designate big things and small things. The word microscope, for a device to look at small things, had been around since the seventeenth century; the less well-known megascope was a nineteenth-century projecting microscope—it threw an enlarged image on to a screen, where it could be traced and turned into a drawing. There are so many words in micro- and mega- that even I am not tempted to try to list them. Megas also gave us the medical suffix -megaly, meaning “enlargement”—acromegaly, enlargement of the extremities (the hands and feet); hepatomegaly, a swollen liver. And maybe it’s worth mentioning that microwaves are just “small waves”—their wavelength is short in comparison to radio waves, but still measurable in centimetres or millimetres, not micrometres.
After the adoption of mega- and micro-, the idea grew that we need a new prefix at each new integer power of one thousand—thousandfold, millionfold, billionfold, and so on up; thousandth, millionth, billionth and so on down. In powers of ten that goes 103, 106, 109 …, and 10-3, 10-6, 10-9 … I’m going to need to use that notation from here on in. The smaller fractions and multipliers were still useful enough to be kept in regular use, but myria- languished and died before the nineteenth century was over.
The next additions came in 1947, from the Union Internationale de Chimie. (I suppose they took a particular interest in the matter because chemistry deals fairly regularly with very large numbers of very small things.) At their fourteenth conference they adopted the following prefixes: giga- for 109 and tera- for 1012; nano- for 10-9 and pico- for 10-12. Along with mega- and micro-, this set a useful precedent for multiplier prefixes to end in “a”, while fraction prefixes end in “o”.
These new prefixes are all Greek but one. The odd one out is pico-, which is Spanish. Pico is the Spanish word for a mountain, and for a bird’s beak. But it’s also used to mean “a little”. So dos metros y pico is “a little over two metres”. (In the case of the picometre, a very little over two metres.) Pico- seems to have been in general use for some time before 1947—the Oxford English Dictionary has an example of its use from William Eccles‘s Wireless Telegraphy and Telephony from 1915, whereas all the others have their first citation from the Proceedings of the 1947 Union Internationale de Chimie conference.
Nano- comes from nanos, “dwarf”. Nanism is a state of dwarfishness, a term applied to the evolutionary change in the size of species at high latitudes or on remote islands. And nanization is the English word for the process of producing a bonsai tree—the deliberate production of a dwarf plant.
Giga- comes from gigas, “giant”, which spawned a whole collection of synonymous adjectives in English. At one time or another we’ve had gigantal, gigantean, gigantesque, gigantical, gigantine and gigantive, but only gigantic survives in common usage. Giga- was intended to be pronounced ʤaɪɡə, with the same first syllable as gigantic. Back in the 1970s I had a physics lecturer who actually pronounced it that way—and we all thought he was wrong, because (unaware of the relationship to gigantic things) we had grown up pronouncing it ɡɪɡə, with the same first syllable as giggle. And that non-etymological pronunciation is now utterly dominant.
Tera- is from teras, “monster”. Teratogenesis is any process, such as drug toxicity in pregnant women, that causes birth deformities—a word coined in a less sympathetic age than our own, for sure.
In 1960, the prefixes pico-, nano-, micro-, mega-, giga- and tera- were all adopted into the newborn Système International, at the 11th General Conference of the Conférence Générale des Poids et Mesures. From here on in, any new prefixes needed to be approved by the CGPM.
And at their next meeting, in 1964, they approved femto- (for 10-15) and atto- (for 10-18). Perhaps they were running out of classical words that didn’t conflict with existing abbreviations—these two are Danish. Femto- is from femten, “fifteen”, and atto- from atten, “eighteen”, the two powers of ten represented, and two initial letters that hadn’t been used before. I wonder if Danish was the chosen language as a nod to the strongly Danish foundations of quantum mechanics, which deals with the extremely small and the extremely fast; but I can’t find any confirmation of that idea.
This was the point at which some sort of mnemonic was becoming necessary. At school I learned, “To Give Me Kicks My Musicians Now Play For Ages”—for tera-, giga-, mega-, kilo-, milli-, micro-, nano-, pico-, femto-, atto-.
But in 1975 the 15th General Conference messed that one up, while restoring symmetry to the prefix range. The approved peta- (for 1015) and exa- (for 1018). We’re back to Greek again, but with new rules. Peta- is from pente, “five”, and exa- is from hex, “six”. Peta- could be distinguished from pico- by using a capital letter in its abbreviation, and exa- was a new initial letter, so that’s all good … but what’s the “five” and “six” about? These are powers of a thousand. 1015 = 10005 and 1018 = 10006. There’s also seems to have been a bit of word-play going on, as if pretending to continue a series in which tera- was derived from Greek tetras, “four”.
O-kay. We can still deal with this. “Ever Polite, To Give Me Kicks My Musicians Now Play For Ages.”
Well, that lasted until 1991, and the 19th General Conference, which approved zetta- (for 1021) and yotta- (for 1024), zepto- (for 10-21) and yocto- (for 10-24).
I know what you’re thinking: Really? Aren’t these some of the less famous Marx Brothers? But they’re actually a continuation of the pattern established by peta- and exa-. Zetta- and zepto- come from Latin septem, “seven”, while yotta- and yocto– are derived from octo, which means “eight” in both Latin and Greek. So the reference is to powers of a thousand again, with these prefixes designating seventh and eighth powers. The custom of ending multipliers in “a” and fractions in “o” continues, as does the sporadic use of paired consonant sounds in the fractions but never in the multipliers—so we have “ct” and “pt” in yocto- and zepto- echoing the “cr” and “mt” in micro- and femto-. As for the odd initial letters, I think they reflect the increasing difficulty of finding abbreviations that aren’t already in use in the SI’s crowded alphabet soup.
And that’s all we’ve got, so far. The Andromeda Galaxy is 24 zettametres away, the Earth has a mass of 5972 yottagrams, and … well, a picture is worth 1.66 zeptomoles of words.
Note: The SI prefixes were expanded in range in 2022, after this post was published. See my update here.