The most common letters in Morse code, and why E is one dot

The share column is from Cornell University's count of letters in a sample of 40,000 English words. The units column is each letter's length counted in dots: a dot is 1 unit, a dash 3, and the gap inside a letter 1. Press a code to hear it, or open a letter for its own page.

The top six letters, E T A O I N, make up just over half of all letters in the sample (51.18%), and five of them last a third of a second or less.

Other counts of English put the letters in a slightly different order below the top few, because every sample of text is different, but E comes first in the usual counts. For the same 26 codes in alphabetical order, with sound and a printable chart, see the Morse code alphabet A to Z.

Samuel Morse first planned to send only numbers, each standing for a word in a codebook. His partner Alfred Vail turned it into a letter code around 1840, and the rule behind it was simple: the more often a letter is used, the shorter its signal.

The usual account is that Vail judged how common each letter was by counting the pieces of type in a printer's type cases at a newspaper in Morristown, New Jersey. A printer stocks the most of the letters it sets most often, so the type case was a ready-made frequency count. No record of Vail's own tally is known, so treat the details as tradition; the principle is not in doubt.

Short codes for common letters make every message shorter on average. Engineers now call this idea variable-length coding, and it is the same idea behind modern data compression.

Wikipedia sums it up: Vail made the length of each symbol "approximately inverse to the frequency" of its letter in English. For the full story of who did what, see who invented Morse code and when.

You can measure the fit with the table above. Weight each letter's length by how often it is used and the average letter in real English text lasts 6.1 units.

If all 26 letters were equally common, the average would be 8.2 units. So the frequency trick saves about a quarter of the sending time.

It is close to the best these codes allow. Give the shortest of the 26 codes to the most common letter, the next shortest to the next, and so on, and the average drops to 5.7 units.

Hand them out the worst way round, with E on one of the 13-unit codes, and it rises to 10.6 units. Today's International Morse code is within about 7% of the best case.

The gap comes from a few letters that ended up longer than their rank deserves:

• O is the fourth most common letter, but its code, three dashes, lasts 11 units, as long as C and P. Wikipedia names the long O as the one real defect of Friedrich Gerke's 1848 code, the version International Morse grew from.
• Y (16th) shares the longest letter length, 13 units, with the rare J and Q.
• M (14th) and U (12th) are each 7 units, the same as R and H, which are used about twice as often.

The original American Morse code kept shorter forms for some of these letters. You can hear the differences in American Morse code.

The short, common letters are also the easiest to mix up, because they differ by a single dot: E, I, S and H are one to four dots, and T and M are one and two dashes.

The letters are sent one at a time with a short silence between them, three units long. How those silences work, and how to write them, is in spaces between letters in Morse code.

Type a word into the translator and press Play. Words made of common letters are quick: THE takes 17 units, while JAZZ, one letter longer, takes 49.