Very little of the tower was found in the excavation. About half of the lowest course was uncovered from the point the wall abuts the tower, going around a rock outcrop, together with parts of two-three not very well built additional courses (the tower seen today is mostly reconstructed). Three things, however, were clear even from the little that was found:
a. What was found of the tower was constructed mostly (but not exclusively) of headers.
b. It was a complete and separate structure; the wall abutted it and was not joined to it.
c. It was built directly on the soft chalk that makes up the hill of Gamla, without any discernible foundation.
The first two points hint at an early date for the construction of the tower, perhaps in the Hellenistic period. Header construction, especially in military architecture, was common in this period, as illustrated by the splendid tower at Samaria (Crowfoot et al. 1942:Pl.XXXVI). Josephus’ builders then used it, incorporating it as the highest terminus of their fortification .

The third point brings us to a passage that is usually looked upon as fantasy. It happened just before the second Roman assault on the city:
‘Working in silence, the [three] soldiers [of the 15th legion] rolled away five stones forming the base. As they jumped out of the way the tower fell with a resounding crash, bringing the sentries down with it’ (9, §62–69). 
While by no means conclusive, the fact that the tower had no foundations and that its entire northern side was missing, lends some credibility to this story, as it would have been relatively easy to dislodge stones from a structure built directly on soft chalk.

[1]  Though not strictly relevant to the topic under discussion, the numismatic evidence (and it alone, so far) points to an appreciable presence at Gamla as early as the days of Antiochus IV (175–164 BCE). The campaign of Jannaeus to capture the ‘strong fort’ of Gamla indicates a military presence there in his days (War I, 4, 8 §(103–106), Ant. XIII, 394) and perhaps the tower was constructed hastily to face his attack. Contrary to the belief of Gutmann, who saw in Gamla a Jewish foundation from the start, I now tend to accept the suggestion made by Moti Aviam, that Gamla may have been founded as a Seleucid outpost in one of the later Syrian wars of the 3^rd century BCE. For a suggested line of Hellenistic fortifications in the Golan see M’aoz 1983.