150 SECTIONAL ADDRESSES. 
the earlier channel in the Valle della Mola di §. Gregorio the intrados is 
168°86 metres above sea-level, while at the last shaft of the newer channel, 
after the bridge which Hadrian built, the intrados is about 172 metres 
above sea-level ;” the floor of the specus would be about 2°80 metres lower 
in each case. The level of the water of the Fosso dell’ Acqua Nera is 155:19, 
and the bottom of the specus at the Ponte Taulella is 155°61. The distance 
between the last two points in a straight line would be not much over 
two kilometres ; but along the line of the aqueduct it should be 21,120 
Roman feet, or more than 7 kilometres according to the cvppi.”* 
It is thus quite impossible that the Anio Vetus should not have been 
above-ground at the Acqua Nera, unless it was carried under it by a 
syphon. There is, about 17 metres below Ponte Lupo, on the right bank 
of the stream, a massive concrete buttress, faced with opus reticulatum, 
probably belonging to the time of Augustus, and containing a shaft, 
which, though blocked up, certainly seems as if it went down at least as 
far as the level of the stream, and might very well reach down to a channel 
passing under it. 
There is no inherent impossibility in the use of a syphon under these 
circumstances. That the Romans were familiar with the principle is well 
known," and another example has recently been discovered near Avezzano, 
where it is cut in the limestone rock. 
The question why syphons were not made use of to take the aqueducts 
in a straight line (like the modern Aqua Marcia) across the Campagna 
from Tivoli to Rome (in which they would have come to intermediate 
levels considerably lower than that which they reach at Porta Maggiore), 
in order to avoid the long detour which we are now following, has been well 
answered by M. Germain de Montauzan in his book on the four Roman 
aqueducts of Lyon,18 in which no less than ten syphons have been observed. 
The Romans did not trust their concrete and cement for making syphons, 
though they might have done so. They were unable to make a large metal 
pipe that would stand pressure; and at Lyon the contents of a channel 
0:58 by 1:75 metre are transferred to nine or ten lead pipes with a bore of 
0-20 when the syphon is reached. We have only to calculate the enormous 
quantities of lead that would have been required to take the water from 
four channels, the largest of which measured nearly 1:20 metre wide and 
3 metres high, and to remember that small-bore pipes would have been 
choked almost at once by the heavy calcareous deposit, to realise how im- 
possible it would have been to adopt this method here. On the other hand, 
all the building material required was quite easy to obtain on the spot or not 
far off. But there is no objection to its use in a rock-cut channel for a short 
15 Livellazione, p. 74. 
16 T saw No. 733a, some 150 metres downstream from Ponte Lupo; while No. 645 
is still in situ some 500 m. before Ponte Taulella, and No. 626 was seen lying by the 
path a little beyond it though not in situ, the intervals being 240 feet. The windings 
of the Anio Vetus must have been very considerable at this point tunnelling being 
avoided as far as possible. 
See Bull. Comunale, 1899, 38; Eph. Epigr. ix. 968 for Nos. 626, 645 ; No. 733 is 
unpublished. It must be remembered that the numbering ran from Rome to the 
source. 
17 For examples see Lanciani, I Comentari di Frontino in Mem. Lincei Ser. III., 
vol. iv. 554 sqq. 
18 Germain de Montauzan, Les Aqueducs Antiques de Lyon, Paris, 1909, 176 sqq. 
aA ES Ge © sir be 
