142 Proceedings of*Philosophical Societies. [Fres. 
to show that the tower contained a water-clock, the vestiges of 
which still exist in conduits, which he has described with care, and 
of which he has given figures in two of his plates. 
If the tower contained a clepsydra, Varro might call it Tower of 
the Clock. He would have named it Tower of the Clock, if, be- 
sides this clepsydra, it had presented eight other clocks, or solar 
dial plates. 
_ This curious particularity for the history of gnomonics was a 
thing sufficiently remarkable for Varro and Vitruvius. We draw 
little more information from the incomplete expression of the one 
than from the silence of the other. 
_ The authors of the Historical Dictionary, in speaking of the 
architect Andronicus, say nothing of the time when he lived. 
Those of the Universal Biography say that “ we judge from the 
style of architecture of that monument already corrupted, and by 
the mediocrity of the bas reliefs, that he was after the time of 
Penclesswe ec ee ' 
{n the time of Pericles and Anaxagoras the science of gnomonics 
was too little advanced in Greece to enable them to form these eight 
dials at Athens. Historians speak of the first gnomon established 
by Anaximander at Lacedemon. ‘There was a great distance be- 
tween this gnomon, which probably only pointed out mid-day, and 
the dials declining in various figures, exhibited in the Tower of the 
Winds. It appears, then, very probable, that Andronicus, or the © 
author of the eight dials, whoever he was, lived a good while after 
Pericles, who died 429 years before our era. Nothing prevents us 
from supposing him contemporary with Hipparchus; and then the 
sun-dials at Athens will suppose nothing that was not known by the 
works of the ancient mathematicians, of whom Ptolemy has ex- 
plained and completed the doctrine in his book of Analemma. If 
there be no contrary proof, I should be inclined to assign as the date 
the first years of our era. Probably the question will never be re- 
solved. What is certain, or at least very probable, is that these sun- 
- dials suppose a knowledge of gnomonics, and consequently of tri- 
gonometry, unless we suppose them to have been traced empyrically 
by means of the concave hemisphere of Berosus. 
These sun-dials are of a form similar to those which we find in 
the Commentary of Commandin on the Analemma. Their theory 
is perfectly known. It remained to be known with what precision 
they had been drawn. 
The style is every where wanting. We see only in the marble 
the holes where it was inserted ; but the summit of the style was 
seldom in the axis of these holes, not even in the regular dials, 
which are here to the number of four. But the height of the style, 
and the place of its foot, are not indispensable data; we can de- 
duce them from some of the dimensions of the dial. The author 
has taken Care to mark on his plates the length of a considerable 
number of these lines; but the choice which he has made is not 
