As Faldo the Great (doesn't that just make you think of a circus clown?) points out, the letters etaoin are the five most common in the alphabet. Subsequent research has cast doubt on everything following that, so shurdlu may or may not be the right order (or letters) for the second five most used. Depends who's counting, I guess. Mergenthaler fell for it hook, line and sinker anyway, and I, as a machine typographer (i.e. linotype operator) had to live with that for ten years on linos.
On the linotype machine (hear that soap-box scraping across the floor to the front of the room), the first two columns on the left-hand and right-hand keyboard sections consisted of those letters, ie:
e s c v x
S t h m b z
P a r f g fi
A o d w k fl
C i l y q ffi
E n u - - - ? Memory lapse after 20 years!
constitutes the first section of the keyboard. The right-hand section of the board had more or less the same layout, but in capital letters, and the middle keypad had 25 "other" characters, the exact mixture depending on the font that was being used.
As was pointed out above (prolly Faldo the Great again) you would sometimes see a line of type in a column that had something like "and the weatetaoin shrdlu shurdlu etaion". What would happen would be that the operator (comme moi) would have either typed the wrong words first or for some other reason decided that the line wasn't worth completing.
Rather than take the matrices out of the assembler by hand and putting them back on the distributor bar (which was called "dissing the mats" and took time and effort), you'd just complete the line by running a finger down the first two columns of the left hand keyboard, pressing the spacebar in between. The spacebar caused something called a spaceband to fall down between the two sets of matrices.
The spaceband was two pieces of high tensile steel joined together by a tongue and groove arrangement. One was short, and sat between the matrices in the assembler, the other was tapered from very thin to about quarter of an inch wide. This was pushed up on the slide (the thin piece) in the mould to force the matrices that formed the words together, and to provide proportional spacing.
The operator (moi, encore) SHOULD have thrown the line back into the pot when it was dropped into the stick. Failing that, the proof readers SHOULD have noted the etaoin/shrdlu lines in the stick of type on the proof and marked them for removal. The stone hand (compositor) SHOULD have removed the line. But sometimes that didn't happen, et voila! etaoin shrdlu's your uncle!
HTH!
