Szabó Miklós, Petres F. Éva: Decorated weapons on the La Tene Iron Age in the Carpathian Basin. (Inventaria Praehistorica Hungariae 5; Budapest, 1992)
II. DRAGON-PAIR SWORDS
125,1), both of which, however, are surpassed by the scabbard from grave 47 of Breáice (Cat.no. 106; Pl. 102,1) whose diagonal Sword Style ornamentation springs from under the "body" of the beasts. The grave from Zemun-Gardoä (Cat.no. 142; Pl. 125,3) suggests the survival of the motif into the late La Tène period. 190 The above call for a revision of proposals based on the results of find statistics according to which the appearance of De Navarro's Type III dragon-pairs should be linked to a Middle La Tène workshops active in Switzerland. 191 A growing number of finds from the Carpathian Basin illustrates the transition between Type I and III and the subsequent development which differs from the "Swiss" one, and which was no doubt influenced by the Hungarian Sword Style. 192 Moreover, the scabbard from grave 40 of Piskolt (Cat.no. 97; PL 96) indicates — insofar as the published drawing is accurate — that the appearance of Type III dragon-pairs in the Carpathian Basin can be dated to the La Tène B2 phase. However, the conservation work done on the occasion of the Venice exhibition has shown that the dragon-pair ornamented with circles can be assigned to Type I. 193 The history and development of dragon-pair swords can broadly be fitted into the following chronological framework. Type II makes its first appearance in the La Tène lb period in the Marne region. The presently known earliest Italian specimens cannot be dated earlier than 320/310. Even though scabbards which can be directly linked to the latter have been found in the Carpathian Basin, Type II dragon-pairs with authentic find associations do not occur prior to the La Tène B2 phase. 194 The evolution of Type I dragon-pairs falls within the latest phase of the early La Tène period: and there are several scabbards which reflect a transition between Types II and L 195 Type II dragonpairs survived into the beginning of the middle La Tène period. 196 Most Type III dragon-pair scabbards can be assigned to the La Tène C period even if some specimens exhibit expressly early La Tène traits, 197 which would imply that the three types (and their transitional variants) could well have been contemporaneous. At the same the uncertainties in the archaeological definition of the early and middle La Tène transition does not allow a finer chronology. 198 This is best illustrated by the scabbard from Singen which can be assigned to the early La Tène period on morphological grounds, but which decorated with a Type III dragon-pair, and whose associated grave goods include fibulae of both early and middle La Tène types. 199 A scabbard fragment from a disturbed grave of Gödöllő (Cat.no. 13; PL 14) illustrates eloquently the complexities in the history of these types. As far as typology and morphology are concerned this scabbard can be assigned to the very end of the early La Tène or even to the beginning of the La Tène C period. 200 At the same time, its Type II zoomorphic lyre reflects the genetic link between the lyre palmette and the "dragon-lyre" far better than the late 4th century B.C. scabbards discussed in the above. 201 Also, the two medallions of the scabbard entry reinforcement are fitted into the animal heads and are encrusted with gold carrying repoussé triskeles. 202 The distribution of dragon-pair swords from England to Romania definitely supports De Navarro's proposal that the motif is best regarded as an inter-Celtic "currency". 203 This is furthermore supported by an art historical observation, namely that all three dragon-pairs occur in combination with "Hungarian" and "Swiss" Sword Style designs. 204 At the same time, the chronological analysis of the distribution of dragon-pair swords indicates "well-established" contacts over the entire La Tène culture province at the turn of the 4th and the 3th centuries, in whose background the migration of warriors or the activity of itinerant craftsmen can both be assumed. 205
