H.—ANTHROPOLOGY. 145 
possible reasons are three. Claudius may quite well have been cheated 
by his freedmen ; secondly, it was certainly risky to place two or three 
channels upon arches that were only built to carry one; and thirdly, the 
channels seem frequently to have been flooded by the Anio, if we are to 
judge by the foul deposit which is frequently found in the channels of 
the Aqua Marcia and the Aqua Claudia, instead of the crystalline deposit 
which is proper to them. 
It will be convenient to follow the course of all the four aqueducts 
together, observing their principal remains as they occur. Before doing 
so I should mention two things by which my task has been greatly 
facilitated. One was the careful scientific levelling of all the remains of 
these four aqueducts that could be found, which was carried out in 1915 
by the late Professor Vincenzo Reina and his assistants, Ingegneri 
Corbellini and Ducci.® This, among other advantages, rendered it far 
easier to assign the ruins correctly (where the course of two or more 
aqueducts was almost identical) than it had hitherto been. The second 
_ was the visit which I paid to the remains of the aqueducts in the spring 
_ of the present year in company with Dr. Esther van Deman (whose 
__ tesearches on the chronology of the various types of construction employed 
in the aqueducts and their frequent restorations are of the greatest import- 
ance and have been drawn upon in what follows) and Mr. G. R. Swain, 
Photographer of the Near East Expedition of the University of Michigan, 
_ towhom I am indebted for a most valuable series of fine photographs. Nor 
ean any student of the Roman aqueducts forget Professor Lanciani’s pioneer 
work in 1880: and my own researches were begun under his guidance. 
The Anio Novus originally drew its water from the river four miles 
_ above the springs of the Aqua Claudia, at the forty-sécond mile of the 
Via Sublacensis; but as the water was apt to become turbid, Trajan 
carried out a project of Nerva, according to which the three lakes used by 
Nero for the adornment of his villa above Subiaco were used as filtering 
tanks. This increased its length considerably, the new intake being some 
six or seven miles further up. Considerable remains of the dam still 
exist on the way up to the far-famed monasteries ; but the centre of it 
‘collapsed in a flood in 1305—as the story goes, partly owing to the malice or 
imprudence of some of the monks who began to tamper with it. Otherwise 
there are no remains of any particular interest until we reach the gorge of 
§. Cosimato, some fifteen miles further down. It lies a couple of miles 
above Vicovaro, where the road from the valley of the Digentia and Horace’s 
‘Sabine farm joins the main road down the Anio valley, the ancient Via 
Valeria. 
Here the Anio Novus is on the left bank of the Anio and the Aqua 
Marcia and Aqua Claudia are on the right bank, as they have been all the 
way from their respective beginnings. At the beginning of the gorge 
the Claudia is vertically above the Marcia, and there is a shaft by which 
water could be run from it into the lower channel in case of need. 
i All through the gorge the two channels may be followed one above 
: the other, cut in the rock, which is here somewhat rotten, so that it has 
_ more than once been necessary to make a new channel further in and 
abandon the outer one. 
® See Livellazione degli antichi acquedotti Romani in Memorie della societa italiana 
delle scienze detta dei XL, serie 3. vol. xx. (1917). 
1925 L 
