[8] Extraordinary are the single finds of 1890 at the Prehistoric pile dwelling settlement Alpenquai – the so-called Potin lumps whose largest weights 59.2 kilograms (131 lb) consist of about 18,000 of used Celtic coins which date to around 100 BC.
The hilltop area dominates the city of Zurich alongside the eastern Limmat riverbank, and its northern slope called Sihlbühl towards the former Sihl delta marked the northern boundary of the Helvetic and Roman settlement – where the structures of the medieval Oetenbach Nunnery, Waisenhaus Zürich and later the Urania Sternwarte were erected at the present Uraniastrasse, and therefore important historical archaeological excavations never were done.
The largely flattened Lindenhof area elevates at 428 metres (1,404 ft) above sea level, and rises about 25 metres (82 ft) above the level of the Limmat at the Schipfe–Limmatquai area in the west; probably some Roman buildings were built at the site of the Zunfthaus zur Zimmerleuten on the other riverbank, and the Roman settlement may stretched towards the present Münsterbrücke which is crossing the Limmat between Grossmünster (remains of graves) and Wasserkirche, and the Münsterhof plaza.
The earliest record of the town's name is preserved on the 2nd-century tombstone found in 1747 AD on the Lindenhof hill, referring to the Roman Vicus as "STA(tio) TUR(i)CEN(sis)" as customs station for goods going to and coming from Italy at the same location as the Celtic Oppidum.
At the present Zunfthaus zur Zimmerleuten at Limmatquai opposite of the Lindenhof hill, the area was stabilized with embankments; some of these mounds date back to the Roman settlement era.
[8] As a Vicus, Turicum was not secured by town walls, but the buildings grouped around the customs station (Quadragesima Galliarum) where the clearance of goods and travelers prior to transfer between the provinces of Gallia Belgica and Raetia took place, mainly on the water route (from and to the Roman heartland over the mountain passes of the Swiss Alps) Walensee-Obersee-Zürichsee passing Centum Prata (Kempraten) towards the Limmat, Aare and Rhine.
Archaeologically excavated are the remains of public baths (Thermengasse), graves and traces of craft enterprises, residential buildings, as well as everyday objects and jewelry, but also of cult equipment.
[15] The ancient name Turicum, along with the indication of a Roman customhouse, is first attested in the epitaph for Lucius Aelius Urbicus, an infant son of the p(rae)p(ositus) sta(tionis) Turicen(sis), "head of the toll-station at Zurich",[16] that was found on Lindenhof hill in 1747 and dates from 185/200 AD.
[17] The place name reappears in the Early Middle Ages as Turicum, Turico, Doricum, Torico, Turigo, Turegum, and in its Old High German forms Ziurichi, Zurih,[18] with regularly shifted consonants; it is unanimously seen as a Gaulish formation, *Turikon, though vowel quantities and accentuation have been a matter of debate: Stress on the second syllable in Rumantsch Turitg, Turi, as well as in the partially reshaped Italian form Zurigo, have been taken as evidence establishing originally long -ī-, that would have drawn the accent to the penult in Latin pronunciation; the constituting elements of the toponym have been identified as a Celtic personal name Tūros and a suffix -īko- forming relational adjectives.
[19][20] This analysis has been contested by arguing that the spelling Turegum, widely attested in documents from the 9th century onward, seems to reflect lenition of the intervocalic stop consonant, as well as lowering of short Latin ĭ to ĕ, both common to most Western Romance languages, and that numerous other place names of Celtic origin, as Autricum, Avaricum or Aventicum, are undoubtedly formed with a suffix -ĭko-, and are usually derived from a hydronym; thus the basis of Turicum supposedly being Turos or Tura, should rather be an ancient name of one of the watercourses around Lindenhof hill, either a distributary of the Sihl river, or possibly the Limmat river.
The Romansh and Italian forms may likely have been taken from medieval written records, with the accent determined by analogy within the borrowing languages, e. g. Turitg after amitg, 'friend', which bears its stress on the second syllable, too.
[21][22][23] The diverging evolution of several place names of Gaulish origin, as Bourges < Bitúriges, Berry < Bituríges; Condes < Cóndate, Condé < Condáte, suggests that shifting accent and unsettled vowel quantity may not have been exceptional.