(Sachs, 1940: Figure 33, p. 102)

A beautiful sketch from Sachs (Sachs, 1940: Figure 33) details the primitive lute of the 18th dynasty in Egypt; a wooden and oval body with a covering skin as soundboard. A long handle pieced the body, either stuck in and out through the skin or protruding from the lower end. The strings were wound around the top of the handle and attached in place by thongs with tassels on them. The average number of strings was two, with sometimes three or four. The right hand plucked the strings with a plectrum while the left hand stopped them by pressing them onto one of the frets tied around the neck, (Sachs, 1940: p. 102).

Nearly all of the lutes talked about up until this point in (non-) history were long-necked. The first short-neck lute to pop up in the near east was in Iran, the same country which afterwards became the center of the short-neck lute. Clay figures from the 8th century BC show these lutes in rough outlines. There is no more evidence of the short-neck lute until, many hundreds of years later, it arose again in the Islamic Near East. Its pegbox was bent backwards in a sickle shape and had pegs sticking out laterally. The strings were attached to the lower end of the body, and a skin served as the soundboard. (Sachs, 1940: p. 251)

The first instrument that can truly be called the ‘ud is what Sachs names ‘the lute with a frontal string-holder.’
The Asiatic countries and Egypt designate it by its classical, Arabic name ‘ud. The principal meaning of this noun is not ‘wood,’ as generally supposed, but ‘flexible stick.’ This corroborates our assertion that the earliest lute came from the musical bow; but, obviously, the name must have belonged to a long lute whose handle was derived from a ‘flexible stick,’ before it was given to the short lute, which probably was not, or at least not directly, connected with the musical bow. (Sachs, 1940: p. 253)

Sachs criticizes the prevailing theory that the ‘ud’s name derives from the Arabic for ‘wood.’ It is important that he make the distinction – because of the connection with ‘musical bow’ or ‘bow’ in general, because the first lutes were probably derived from the hunting bow.

The most accepted theory is, however, that the name ‘ud started to be used when the skin that covered the face of primitive lutes was replaced by a wood surface, in around the 7th century AD.

Iraqi lusterware bowl of the tenth century. http://trumpet.sdsu.edu/M151/Arab_Music1.html

This ancestor to the modern ‘ud was developed in Persia, and called the barbat. It had a body and a neck with two crescent-shaped sound holes, like many East Asian lutes, which might suggest its origin in West Asia. The present form of the ‘ud probably emerged in Iraq, and further during the Muslim occupation of Spain, with a round sound hole that had a carved wooden rose, and a separate neck from the body. (http://easyweb.easynet.co.uk/~orpheus/ref3.htm)

The short lute that developed in Muslim Spain traveled west into southern Europe. In the 10th century, a very similar instrument was carved on a wooden pulpit in a small church in Arcetri at Florence. See (Sachs, 1940: Plate XIVa.)

(Sachs, 1940: Plate XIVa.)

Because the depictions of the lutes were all artistic and not scientific, there is great debate among music historians is whether the lutes of ancient times had frets or not. Polin talks about the use of frets in primitive lutes:
Frets were first equidistantly placed, but since this was unsatisfactory except for octaves, the division principle of dividing the strings into half, quarter, etc. in order to attain desirable tones, gained vogue. (Polin, 1954: p. 46)

Valentin writes that frets arose in the 2nd millennium BC. (Valentin, 1954: P. 208) Sachs, however, denies the fact that the ancient lutes might have had frets:
Lutes seem to have had no frets, either in old times or today, in spite of the constant use by the theorists of the word dasatin, plural of Persian dast or hand, which is used to indicate frets. And it would have been difficult to string them securely around the sloping end of a pear-shaped lute. Very probably, the frets existed only theoretically to symbolize the positions of the stopping fingers. (Sachs, 1940: p. 253)

Bill Bradley, in a talk in 1997, (a good 60 years after Sachs), had the following to say:
Graphic images from the Andaluz kindom make it clear that… until the fourteenth and 15th centuries it does appear to have had frets, meaning that it cannot have been used to play quarter-tonal music, unlike the unfretted instrument in use today.
(http://userpage.fu-berlin.de/~telehist/ambass/ud.htm)
Bradley’s remarks are puzzling because it is clear that an instrument could easily have been tuned to the quarter tone intervals used in Arabic music. Nevertheless, the present argument is about frets, not about quarter-tones.