After the workmen’s day (7:00 A.M. to 4:30 P.M., with a half-hour for lunch) was over, there was still much for the staff to do. Pottery, spread out on trestle tables, had to be examined, joins made where possible, types distinguished. (Careful attention at Cosa to plain Roman black glaze has led to an arrangement of types in a dated series which will be useful for future dating on other sites.) Evenings were devoted to writing up the journal, studying the manuals, making drawings, planning the next day’s dig, and shop talk. The results of a typical season’s work, in 1950 on the _arx_ at Cosa (Fig. 4.4), were to isolate a second temple at right angles to the Capitolium, restore on paper the design of several sets of terracotta revetments, follow the line of the Via Sacra from the _arx_ gate to the Capitolium, clear the _arx_ wall, get down to bedrock beside the Capitolium, discover a terracotta warrior who was part of the pedimental sculpture of an older temple under the excavation shack, and in general get a pretty clear idea of the religious center of the colony as it was, perhaps, in the time of the elder Cato, in the early second century B.C. [Illustration: FIG. 4.4 Cosa, _arx_. (F. E. Brown)] [Illustration: FIG. 4.5 Cosa. (J. B. Ward Perkins, _loc. cit._, Fig. 8)] In the two seasons preceding the discoveries on the _arx_ just described, much work had been done. In the survey to set up the fifty-meter grid, Cosa’s own ancient rectangular grid of streets, with pomerial street running just inside the wall as at Marzabotto, came out clear enough to be plotted on the plan (Fig. 4.5), together with the standard blocks of housing, like the identical “ribbon-development” apartment blocks of a welfare state, which compensated the pioneers for whatever fleshpots they had given up in the metropolis or elsewhere. Housing was found to occupy two-thirds of Cosa’s thirty-three acres, while public buildings took just over twenty per cent, and streets the rest. The site, which is waterless, was found to be honey-combed with cisterns: over sixty-five were plotted. The mile-and-a-half of walls, with their eighteen towers, spaced an effective bowshot apart, had been closely examined. They were found to be built with two faces and a rubble fill. The outer face was handsomely finished, with tight mortarless joints, and sloped seven degrees back--this is called “batter”--from the perpendicular; the inner face was left rough. Potsherds of the Etrusco-Campanian style found in the rubble fill were of a period matching Livy’s date of 273 B.C. for the colony. It was clear that the walls, which show throughout no difference in technique, were built all at one go, at the time the colony was founded. Those impatient of the Roman reputation for perfect engineering will be pleased to know that the ancient craftsmen, when they came to close the ring of the wall, found they had made an error of from two to four Roman feet.