S5003149

38

tkM to holp thtir reduction ln cylindrical fumaccs, of a dlarootcr clearly less t han those of Luatln! Ncvcrtheless, elsewhere Iron ora roaatlng furnncos ara boglnnlng to ba recogniscd and daacrlbad aa auch. Thua tha Poli-sh evpert A Pawłowski (1979) hns racant ly recogniscd a do zen furnacaa of thls type associated with a Romnn-period alta at Dobrr.cn Mn ly (Communliy of Dobrzeń Wielki, province of Opole, Po land). Two hcnps of roaatad ora accompanied che fumaccs as woli as flva raduction fumacat and twenty hearths. Similarly Laube (1980) hns daacrlbad four roaatlng furnacaa of the third to fifth centuries AD at Waltanadorf (Kr Ktłnlga WuatarHausen,

Ge imany) wherc Chara wara 51 araas pavad with atone showlng tracas of heatlng, and a fcw fragments of charcoal. An excursion of the Erzberg Symposium hałd ac Vordranbarg in Uppar Styria (Erz, Holzkohle, Schlacke und das Prflhc Eisen, 12-15 October 1981) anablad us to sec tha elghteenth and nlneteenth ecntury roasting Inatallatlons not far from the thlrteenth century ironworks.

Anothar examplo, of ninataenth century dato, lles near to the fara of Koncerny, te the west of tha connune of tha Coulonche (departnent Orne, France). Thls furaice probably roasted Iron carbonate ores from the area of tha Ferte-Mac<* and of the Perrlere-aux-Etangs not far dlstant. Blocks of armorican sandstone wara found wcighing between 2 and 10 kg. Sonie of thea have ona face greenish and vitrified, together forming the botton of an lnvartad eona ona aetre in depth, mixed with burnt clay strongly coloured red. Thls is all ln the aound, 10 m in diameter at the base and 6 a ln helght, covered by copplca which extends for about a hectare. A raised road about 5-6 a wide runs east-west through thls coppice, which used to lead to Iron carbonate workings at the Perriere-aux-Etangs on one side and the Fercć-Mace on the other. The wood probably came from the forests near to Daaplerre, Messei and Andalnes.

Two fragmonts of the sandstone, one of which had a vltrified face, have been analysed (Table 4). One can see from these flgures that the vitrifleation of the surface of sonie of the sandstone blocks is a result of the actlon of the a 1 kalinę ashes on the silica of the blocks with the foraatlon of potassium sllicate ln the naln, which proves that wood was the fuel used with the ala of heatlng at relatlvely Iow temperatures compared to thoae at which sleg ls foraed.

Petrologlcal study of the sandstone has shown the structure to be that of a type of local armorican sandstone of Silurian age. Thls quarzltic sandstone resists heat well, it does not usually spllt ln the fire, and ls well-sulted to the constructlon of roasting furnaces. Some of the blocks show areas somewhat reddened beneoth the vltrlfleation layer resultlng from the actlon of heat on the smali quantitles of Iron occurring naturally ln thls sandstone. The reddened clay shows a variety of characteristic tints which show that lt was heated to differlng temperatures. The ceramlst Salvetat has eatablished a table of temperature-related tints from orange-yellow (850*C) to purple-red (1000*C), as glvan in the following seąuence: orango-ye1 Iow, flesh-coloured, caralne, lacquer-red, red-brown, and purple-red. These taras although old-fashioned may be used to eatlaate the temperaturę to which the furnace clay was exposed.

We Bse from these conslderatlons that the scientific study of roasting procesees ln foraai tlmes would deserve to be completed by careful scientific examlnatlon of the remalns found on metal working sltes when excavated furthar.

Table 4

	llnmodified sandstone %	Vitrified surface X
loss of ignition		0
Si02	93.60	83.50
**2»3		2.98
Pe20j	1.10	0.95
CaO	0.40	0.30
MgO	0.45	0.28
Rz0	-	9.6
Na20	-	1.9
P203	-	0.40

Discussion

We will finish with a discussion of certain fumaccs of which traces have been recovered, even complete examples, but which unhappily were often destroyed by the workers at the tiae of their accidental discovery owing to the fragility of their burnt clay walls.

In about 1860 a group of approximately ten furnaces was discovered in the slag heaps at Lavardin, about a dozen kilometres northwest of Hans (department of the Santhe). The furnaces were of unusual dimension and unlike the smelting furnaces in the sante region. The ore used was calied Roussard locally (Harąuis n. d.) and definitely must hawe been roasted in these furnaces before being reduced not far from the o ld chapel of Saint -Nicolas (Jeanne Chapin, personal communication). Their form is siailar to the furnaces discovered in 1907 and described by the minlng engineer L Davy de Chateaubriant (1913), and to that of Blars uncowered in 1912 by the Marquis de Tryon-Montalembert on the lands of the commune of Fontaines near to Toucy (department Yonne). The interior diaweter ot 1.85 ■ greatly exceeded that of the smelting furnaces usually employed in the Gallo-Roman period (Tryon-Montaleabert 1955).

A bank of four pear-shaped furnaces of the same dinensions was uacov-ered in the hamlet of Puits-d'Havenas in the commune of Chatillon-sur-lorre (department Loiret) under one of the five slag heaps covering about a sąuare kilometre, and described by Ł Dumuys (1898).

To conclude, G Vallois (1884) calied atteation to furnaces of badly burnt clay mixed with sand and easily crushec by pressure. This : ragi lity is again well-attested by F de Hely (1906) for a furnace at Nesni 1-Gemain (department Calvados). Tryon-Nontalembert (1955) expresses doubts on the •'urpose of thls type of badly burnt furmace, easily crushod between the