This fact and the peculiar character of the houses must have given to Silchester rather the appearance of a village with scattered cottages, each in its own plot facing its own way, than a town with regular and continuous streets.

6. Industries. - Shops are conjectured in the forum and elsewhere, but were not numerous. Many dyers' furnaces, a little silver refinery, and perhaps a bakery have also been noticed.

Fig. 6.  Plan of supposed Inn and Baths at Silchester. Fig. 6. - Plan of supposed Inn and Baths at Silchester.

7. Streets, Roads, etc. - The streets were paved with gravel: they varied in width up to 28½ ft. They intersect regularly at right angles, dividing the town into square blocks, like modern Mannheim or Turin, according to a Roman system usual in both Italy and the provinces: plainly they were laid out all at once, possibly by Agricola (Tac. Agr. 21) and most probably about his time. There were four chief gates, not quite symmetrically placed. The town-walls are built of flint and concrete bonded with ironstone, and are backed with earth. In the plans, though not in the reports, of the excavations, they are shown as built later than the streets. No traces of meat-market, theatre or aqueduct have come to light: water was got from wells lined with wooden tubs, and must have been scanty in dry summers. Smaller objects abound - coins, pottery, window and bottle and cup glass, bronze ornaments, iron tools, etc. - and many belong to the beginnings of Calleva, but few pieces are individually notable. Traces of late Celtic art are singularly absent; Roman fashions rule supreme, and inscriptions show that even the lower classes here spoke and wrote Latin. Outside the walls were the cemeteries, not yet explored. Of suburbs we have as yet no hint.

Nor indeed is the neighbourhood of Calleva at all rich in Roman remains. In fact, as well as in Celtic etymology, it was "the town in the forest." A similar absence of remains may be noticed outside other Romano-British towns, and is significant of their economic position. Such doubtless were most of the towns of Roman Britain - thoroughly Romanized, peopled with Roman-speaking citizens, furnished with Roman appurtenances, living in Roman ways, but not very large, not very rich, a humble witness to the assimilating power of the Roman civilization in Britain.

The country, as opposed to the towns, of Roman Britain seems to have been divided into estates, commonly (though perhaps incorrectly) known as "villas." Many examples survive, some of them large and luxurious country-houses, some mere farms, constructed usually on one of the two patterns described in the account of Silchester above. The inhabitants were plainly as various - a few of them great nobles and wealthy landowners, others small farmers or possibly bailiffs. Some of these estates were worked on the true "villa" system, by which the lord occupied the "great house," and cultivated the land close round it by slaves, while he let the rest to half-free coloni. But other systems may have prevailed as well. Among the most important country-houses are those of Bignor in west Sussex, and Woodchester and Chedworth in Gloucestershire.

The wealth of the country was principally agrarian. Wheat and wool were exported in the 4th century, when, as we have said, Britain was especially prosperous. But the details of the trade are unrecorded. More is known of the lead and iron mines which, at least in the first two centuries, were worked in many districts - lead in Somerset, Shropshire, Flintshire and Derbyshire; iron in the west Sussex Weald, the Forest of Dean, and (to a slight extent) elsewhere. Other minerals were less notable. The gold mentioned by Tacitus proved scanty. The Cornish tin, according to present evidence, was worked comparatively little, and perhaps most in the later Empire.

Lastly, the roads. Here we must put aside all idea of "Four Great Roads." That category is probably the invention of antiquaries, and certainly unconnected with Roman Britain (see Ermine Street). Instead, we may distinguish four main groups of roads radiating from London, and a fifth which runs obliquely. One road ran south-east to Canterbury and the Kentish ports, of which Richborough (Rutupiae) was the most frequented. A second ran west to Silchester, and thence by various branches to Winchester, Exeter, Bath, Gloucester and South Wales. A third, known afterwards to the English as Watling Street, ran by St Albans Wall near Lichfield (Letocetum), to Wroxeter and Chester. It also gave access by a branch to Leicester and Lincoln. A fourth served Colchester, the eastern counties, Lincoln and York. The fifth is that known to the English as the Fosse, which joins Lincoln and Leicester with Cirencester, Bath and Exeter. Besides these five groups, an obscure road, called by the Saxons Akeman Street, gave alternative access from London through Alchester (outside of Bicester) to Bath, while another obscure road winds south from near Sheffield, past Derby and Birmingham, and connects the lower Severn with the Humber. By these roads and their various branches the Romans provided adequate communications throughout the lowlands of Britain.