Discussion Forum—A Way with Words, a fun radio show and podcast about language

It's "Two" and comes from the military. Why "two", I don't know. From the Marine Corps Close Order Drill Manual:

1. The purpose of the hand salute is to demonstrate mutual respect and courtesy between members of military organizations and to show respect to national colors. It is executed in one count when halted at attention, marching at quick time, or seated in a vehicle. The command is “Hand, SALUTE.” To return to the position of attention the command is “Ready, TWO.”
a. When “SALUTE” is given, raise your right hand smartly in the most direct manner until the tip of your forefinger touches the lower part of the headdress above and slightly right of your right eye. Your fingers should be extended straight and joined with the thumb along the forefinger. You should be able to see your entire palm when looking straight ahead. Your upper arm should be parallel with the deck with the elbow in line with the body and your forearm at a 45-degree
angle. Your wrist and hand should be straight, a continuation of the line made by your forearm. At the same time, if not in ranks, turn your head and eyes toward the person or colors you are saluting.
b. At the command “TWO,” return to attention. Move your hand smartly in the most direct manner back to its normal position by your side.
c. To ensure simultaneous execution of the second movement of the hand salute
when troops are in formation, the preparatory command “Ready,” will be used prior
to the command of execution “TWO.”

MadWriter said:

It's "Two" and comes from the military. Why "two", I don't know.

It's "two" because it's part of executing a command "by the numbers," which is commonly done during training.

Drill moves vary in complexity. Some of them, such as coming to attention, require only one step. Most others such as Right Face have two, and a few (often involving weapons) have more. Almost without exception, they begin with a preparatory command, for example, "Right." On hearing that, the soldier can anticipate what the command will be, but waits for the actual command before beginning to execute it, so that all troops act at one time. The preparatory commands are unique--that is, "Right" is never followed by anything but "Face". (The command "Attention", while one word, is actually delivered as though it were two; "Attent- -tion," with the second part often pronounced more like "-hut.")

Once begun, multi-part commands are typically completed automatically, according to a learned rhythm. But during training, it is useful to be able to watch and critique each action a step at a time. For example, a drill instructor might issue the preparatory command, "By the numbers, Right". The troops know that upon hearing the command "Face," they are to execute only the first part of the action, rotating to the right on the ball of one foot and the heel of the other, and then wait for a "Two" command before completing the action by bringing the feet together.

The instructor can tell by looking at the intermediate position whether individual troops are executing the action correctly. When satisfied, the instructor will continue with "Ready Two," "Ready Three," etc., as appropriate until the command is complete.

The other purpose, as you mentioned, is as a way to synchronize the execution of a recovery from a command such as a Hand Salute, held for some indeterminate length of time.

(In most cases this is handled by using a different command. For instance, when troops are under arms, the command "Present Arms" would be probably be used. The exact actions would depend on the type of weapon being carried by each individual. Those not carrying a weapon would execute a hand salute instead. The recovery action would be commanded by "Order Arms".)