In the time of Augustus, however, it was found that the water brought by this aqueduct was not sufficient, especially in summer; and as there was a large Roman camp which also required to be supplied with water, situated at a short distance from the city, it was determined to construct a second aqueduct.  For this purpose the springs at the head of a small river, called now the Brevenne, were tapped, and conveyed by means of an underground aqueduct (known as the aqueduct of the Brevenne) which wound round the heads of the valleys, and after a course of about thirty miles is believed by some to have arrived at the city, but by others to have stopped at the Roman camp, and to have been constructed exclusively for its supply.

I have here a diagram, after Flacheron, showing a section of this aqueduct, and this will give a very good general idea of the section of a Roman aqueduct where constructed underground.  It will be seen that the specus or channel is 60 centimeters (or nearly 2 ft.) wide, and 1m. 57c. (or a little over 5 ft.) high, and that it is lined with a layer of 3 c. (or nearly 11/4 in.) of cement.  It is constructed of quadrangular blocks of stone cemented together, and has an arched stone roof.  It will be noticed also that the angles at the lower part of the channel are filled up with cement; it appears also that this aqueduct crossed a small valley by means of inverted siphons.  But neither of these aqueducts came from a source sufficiently high to supply the imperial palace on the top of Fourvieres.

Their sources are, in fact, according to Flacheron, at a height of nearly 50 ft. below the summit of Fourvieres, and it was, therefore, considered necessary by the emperor Claudius to construct a third aqueduct.  The sources of the stream now called the Gier, at the foot of Mont Pila, about a mile and a half above St. Chamond, were chosen for this purpose, and from this point to the summit of Fourvieres was constructed by far the most remarkable aqueduct of ancient times, an engineering work which, as will be seen from the following description, partly taken from Montfalcon’s history of Lyons, partly from Flacheron’s account of this aqueduct, and partly from my own observations on the spot, reflects the greatest possible credit on the Roman engineers, and shows that they were not, as has been frequently supposed by those who have only examined aqueducts at Rome, by any means ignorant of the elementary principles of hydraulics.

To tap the sources of a river at a point over 50 miles from the city, and to bring the water across a most irregular country, crossing ten or twelve valleys, one being over 300 ft. deep, and about two-thirds of a mile in width, was no easy task; but that it was performed the remains of the aqueduct at various parts of its course show clearly enough.  It commences, as I have said, about a mile and a half from the present St. Chamond, a town on the river Gier, about 16 miles from St. Etienne.  Here a dam appears to have been constructed across the bed of the river, forming a lake from which the water entered the channel of the aqueduct, which passed along underground until it came to a small stream which it crossed by a bridge, long since destroyed.