Bridge near Limyra

The Bridge near Limyra (in Turkish: Kırkgöz Kemeri, »Bridge of the Forty Arches«) is a late Roman bridge in Lycia, in mod­ern south-west Turkey, and one of the old­est segmented arch bridges in the world. Lo­cated near the an­cient city of Limyra, it is the largest civil en­gi­neer­ing struc­ture of antiquity in the re­gion, span­ning the Alakır Çayı river over a length of 360 m (1,181.1 ft) on 26 seg­men­tal arches. These arches, with a span-to-rise ratio of 5.3:1, give the bridge an un­usu­ally flat pro­file, and were un­sur­passed as an ar­chi­tec­tural achieve­ment until the late Middle Ages. Today, the struc­ture is largely buried by river sed­i­ments and sur­rounded by greenhouses. De­spite its unique fea­tures, the bridge re­mains rel­a­tively un­known, and only in the 1970s did re­searchers from the Istanbul branch of the German Ar­chae­o­log­i­cal Institute carry out field examinations on the site.

Contents

 

Field examinations

Ancient paving stones. View to the east, along the greenhouses

No in­for­ma­tion on the bridge sur­vives from ancient sources. The first de­scrip­tions ap­pear in Eu­ro­pean travellers' ac­counts from the 19th century. The British ar­chae­ol­o­gist Charles Fellows was the first to ex­plore the re­gion of Lycia, and vis­ited the bridge in May 1840. Fel­lows, as well as T.A.B. Spratt and Ed­ward Forbes, who vis­ited the site two years later, de­scribe it as hav­ing 25 arches.^[1] In 1882, an Aus­trian ex­pe­di­tion, in­clud­ing Otto Benndorf, in­ter­preted the struc­ture as part of an ancient road that con­nected Limyra with the city of At­taleia (mod­ern Antalya) to the east. How­ever, this mis­sion failed to pro­duce any plans or sketches of the site.^[1]
The first, and as of 2008 only sci­en­tific ex­am­i­na­tion of the bridge was un­der­taken by the Ger­man ar­chae­ol­o­gists Wolf­gang W. Wurster and Joachim Ganz­ert in two suc­ces­sive days in September 1973, and com­pleted through fur­ther vis­its in sub­se­quent years.^[1] Their find­ings were pub­lished in 1978 in the Archäologischer Anzeiger jour­nal of the German Ar­chae­o­log­i­cal Institute, with the ex­press in­tent of bring­ing the im­per­iled state of the hith­erto al­most in­tact bridge into the spot­light:^[2]

Recently, cit­rus fruit plan­ta­tions were cre­ated in this fer­tile bot­tom­land; green­houses for early veg­eta­bles are now being built east of the bridge. The bridge is greatly en­dan­gered through the be­gin­ning of in­ten­sive cul­ti­va­tion in the vicin­ity. The lo­cals strip the still in­tact bridge sur­face for stones; bull­doz­ers em­ployed to build ir­ri­ga­tion canals rip the struc­ture apart and crush the stone pave­ment with their caterpillar tracks.

In 1993, the civil en­gi­neer Colin O'Connor sum­ma­rized their re­port in his English-language monograph on Roman bridges, and also em­pha­sized the ex­cep­tional char­ac­ter of this bridge.^[3] No fur­ther sci­en­tific ex­am­i­na­tions of the Limyra Bridge are known.

Location and situation

The rather modest Alakır Çayı, flowing under the 1st segmented arch.

The bridge crosses the Alakır Çayı stream, whose an­cient name is un­known, 3.2 km (2.0 mi) east of the ruins of Limyra (dis­tance mea­sured from the city's the­atre) and 3.8 km (2.4 mi) north of the mod­ern coast­line, close to the mod­ern road from Turunçova to Kumluca.^[1] The local topog­ra­phy is dom­i­nated by the foothills of the Toçak Dağı moun­tains, which blend here with the al­lu­vial bot­tom­lands of the Bay of Finike.^[1] In this tran­si­tional area, the bridge was built im­me­di­ately above the point where the nar­row river val­ley opens into the wide es­tu­ary plain, and where the cross­ing of the river dur­ing the rain pe­riod would be ob­structed by high water.^[4] While the east­ern end of the bridge lev­els off at the gravel plain, the west­ern end abuts di­rectly to the ris­ing rock wall of the moun­tain, pre­sum­ably for pro­tec­tion against floods.^[5] The re­sult­ing sharp bend in the course of the road could also be ex­ploited in case of need to block the road. A mod­ern dam, Alakır Barajı, was con­structed up­stream for ir­ri­ga­tion pur­poses and flood pro­tec­tion.^[5]
Lycia did not, in con­trast to other Roman provinces, pos­sess a very de­vel­oped road system. While the north–south traf­fic was con­ducted pri­mar­ily through the few river val­leys, the east–west routes led, un­like today, mostly on and along the moun­tain ridges.^[6] This par­tic­u­lar road, lead­ing from Limyra over the Alakır Çayı and into the neigh­bour­ing re­gion of Pamphylia and At­taleia, must have been of spe­cial im­por­tance, since the two re­gions were united in a sin­gle province, Lycia et Pamphylia, until the 4th century. In com­par­i­son to the main ar­te­r­ial roads of the Roman Em­pire how­ever, the roads of Lycia were, with 3–4 m (9.8–13.1 ft) width, rather mod­est and con­fined to pedes­trian and pack an­i­mal traf­fic.^[6] This is fur­ther cor­rob­o­rated by the fact that no wagon ruts are ev­i­dent on the paving of the Limyra bridge, nor any traces of a para­pet or breast­work.^[7]

Construction

The 1st arch seen from the south.

With a length of 360 m (1,181.1 ft), the Limyra Bridge qual­i­fies as the largest sur­viv­ing en­gi­neer­ing achieve­ment of an­tiq­uity in Lycia.^[8] The bridge stands on 26 uni­form seg­men­tal arches con­sist­ing of a dou­ble, ra­di­ally laid course of bricks.^[9] At the east­ern end, the orig­i­nal 27th arch has been re­placed by two smaller, semi­cir­cu­lar arches of later con­struc­tion. The lat­ter are built with a sin­gle course of bricks. Traces of the orig­i­nal, flat be­gin­ning of the col­lapsed arch are still vis­i­ble on the piers.^[9]
At the time of Wurster's and Ganzert's visit to the site, the en­tire bridge was buried by river sed­i­ments up to the spring­ing line of the vaults. No ef­forts to dig them up were un­der­taken by Wurster and Ganz­ert.^[5] Only two of the 28 arches were ex­posed enough so that di­rect mea­sure­ments of the clear span and the pier width could be un­der­taken. It was, how­ever, pos­si­ble to cal­cu­late the di­men­sions of the re­main­ing bays from their ex­posed sec­tions.^[10]

Architecture

Arches and piers

Dimensions of a typical segmental arch

The spans of the arches ranged from 11.60 m (38.1 ft) to 14.97 m (49.1 ft) (arches 2 and 26 re­spec­tively).^[9] Based on the size of the spans, one can dis­tin­guish four groups, with the fol­low­ing median val­ues:^[9]

• 11.60–12.30 m (38.1–40.4 ft) for four arches (Nos. 2, 3, 7, 21)
• 12.75 m (41.8 ft) for fourteen arches (Nos. 5, 9–15, 17–19, 22–24)
• 13.10 m (43.0 ft) for four arches (Nos. 1, 4, 6, 8)
• 13.60 m (44.6 ft) for three arches (Nos. 16, 20, 25)

The rea­son for this group­ing is un­clear; they cer­tainly can­not be at­trib­uted to an at­tempt to match the lay of the river bed. The dif­fer­ences how­ever could point to the re­peated use of differently-sized falsework struc­tures in the con­struc­tion of the barrel vaults (see an­i­ma­tion below).^[11]
Only in a sin­gle case, be­tween arches 26 and 27, were Wurster and Ganz­ert able to de­ter­mine the breadth of a pier: 2.10 m (6.9 ft). Sub­tract­ing this value from the com­mon arch span of 12.75 m (41.8 ft), a clear span of 10.65 m (34.9 ft) re­mains.^[12] Since all arches have a rise of ca. 2 m (6.6 ft), the Limyra Bridge has an un­usu­ally large span-to-rise ratio of 5.3 to 1. Such flat­tened arches were very rare at the time for stone bridges, and were not matched and sur­passed until the reap­pear­ance of seg­men­tal arch bridges in 14th century Italy. For the largest arch in the Limyra Bridge, the ratio is even greater, span­ning a width 6.4 times its height.^[12] The two later arches, in con­trast, are, with a ratio of 2.7:1, quite or­di­nary semi-circular arches.^[12]

Height

Arches 8 and 9, south side, silted up to the level of the arch abutments.

The total height of the bridge could not be de­ter­mined be­cause most of the struc­ture is silted up. How­ever, the dis­tance be­tween the arches' abut­ments and the pave­ment sur­face could be de­ter­mined at only 3.25 m (10.7 ft).^[13]
The sur­face level is al­most hor­i­zon­tal: the road­way lies at a height of 20.05–20.55 m (65.8–67.4 ft) above sea level be­tween arches 1 to 20, and falls slightly in its east­ern sec­tion, be­tween arches 21 and 26 to a level of 19.94 to 19.66 m (65.4 to 64.5 ft).^[14] Since there is no in­di­ca­tion of a later sink­ing of the struc­ture, this nearly uni­form height in­di­cates care­ful lev­el­ling and the ex­is­tence of solid pier foun­da­tions.^[15] In con­trast, the length­wise axis of the bridge dis­plays often sig­nif­i­cant de­vi­a­tions from arch to arch.^[15]

Statics

Remarkably, the sup­port line of the structure's dead load is al­most iden­ti­cal with the curve of the vault arch.^[16] The statical analysis of the bridge proves the structure's great load ca­pac­ity:

In mod­ern clas­si­fi­ca­tion, the bridge would be able to sup­port loads of a Class 30 bridge (per DIN 1072); this would mean, that it could sup­port a 30-tonne ve­hi­cle on one arch and ad­di­tion­ally on the re­main­ing sur­face of the arch a load of 500 kp/m². The bridge was thus very safe for an­cient traf­fic.^[17]

Materials

The Limyra bridge was built in a com­pos­ite man­ner of brick, stonemasonry, and rubble.

Arches

Arch closeup: double layer of bricks bound with mortar.

The bricks of the seg­men­tal arches are made of yellow-red clay, mixed with fine brick frag­ments. The bricks are rec­tan­gu­lar plates, mea­sur­ing ca. 40 cm × 50 cm (15.7 in × 19.7 in) and 5 cm (2.0 in) thick. They are placed with the shorter side up­right fac­ing out­wards, so that the total thick­ness of the arches mea­sures ca. 80 cm.^[12] The bricks are bound with grout of lime mor­tar mixed with brick shards and fine gravel, form­ing 4 cm (1.6 in)-thick joints. The two later semi-circular arches were built with smaller bricks, al­though sur­viv­ing bricks from the orig­i­nal arch were re-used at places.^[12] The abut­ment stones are dressed lime­stone ashlar blocks, and form a slop­ing sur­face to sup­port the arches.
The double-layer tech­nique fa­cil­i­tated a more ef­fec­tive use of the false­work, since they could be re­moved and used else­where as soon as the lower course was fin­ished:

The con­struc­tion of the two lay­ers of the dou­ble brick-arches in two sep­a­rate phases had two ad­van­tages. In the first phase, the false­work only had to carry the weight of the lower course, so that it could it­self be of lighter con­struc­tion. In the sec­ond phase, the weight of the upper layer could be al­ready borne by the lower vault; the false­work was thus al­ready avail­able for use in a dif­fer­ent arch.^[17]

Hypothetical workflow on the bridge: the falsework was moved to another opening as soon as the lower arch rib had been completed.

Pavement with projecting corbel stone at the 3rd arch, south side

Superstructure

As far as can be de­ter­mined, be­tween arches 2 to 21, the outer fac­ing of the superstructure con­sists of four lay­ers of brick, fol­lowed by lay­ers of rough stone­ma­sonry bound with mor­tar.^[18] In con­trast, be­tween arches 22 and 26, as well as in both ramps on ei­ther end, the fac­ing con­sists of ash­lar blocks.^[7] The two re­paired arches 27a and 27b are dif­fer­en­ti­ated from the ear­lier work through the use of smaller un­worked stones and the in­cor­po­ra­tion of ir­reg­u­larly placed brick tiles.^[7] In arch 26, the lower side of the arch still fea­tures a pro­ject­ing bear­ing used for the place­ment of the false­work.^[18]
The in­te­rior of the su­per­struc­ture con­sists of a com­bi­na­tion of un­worked ma­sonry blocks and large river boul­ders, bound with lime mor­tar.^[18]

Pavement

The bridge sur­face is lo­cated only 30–40 cm (11.8–15.7 in) over the arches' apex, is 3.55–3.70 m (11.6–12.1 ft) wide, in­creas­ing to­wards ei­ther end at 4.30 m (14.1 ft) and pro­jects ca. 10 cm (3.9 in) over the su­per­struc­ture on ei­ther side.^[7] It is paved with large and ir­reg­u­lar lime­stone cobblestones, al­though smaller peb­bles have been used in the re­pairs at the two semi­cir­cu­lar arches.^[7]

Dating

The exact chrono­log­i­cal place­ment of the Limyra bridge is dif­fi­cult due to its unique place within Roman engineering tra­di­tion and the lack of com­pa­ra­ble struc­tures.^[17] Wurster and Ganz­ert used fol­low­ing char­ac­ter­is­tics as a start­ing point:^[19]

• multiple arches of same width, and horizontal deck level with only small-gradient ramps on either end
• very flat segmental arches built with double, radially layered brick courses
• the use of mortar in the stonework
• exterior facing primarily rough masonry with interspersed bricks, partly dressed ashlar blocks
• exceptionally large paving stones.

In con­trast to these, most Roman stone bridges were faced with ash­lar and rested on voussoir arches,^[20] a method which is dom­i­nant in other vaulted struc­tures in Lycia too.^[21] Com­pared to the mas­sive and tall semicircular-arched bridges that are typ­i­cal of Roman architecture, the Limyra bridge, with its flat seg­mented arches, of­fers a markedly lower and more elon­gated ap­pear­ance,^[22] so that Wurster and Ganz­ert offer a »pro­vi­sional« late dat­ing, ap­prox­i­mately dur­ing the reign of em­peror Justinian I in the 6th century, dur­ing which the use of mixed brick and stone is at­tested in other struc­tures of the re­gion.^[23]
On the other hand, this mixed tech­nique is al­ready pre­sent in the 3rd-century AD aqueduct of Aspendos, and the Ro­mans cer­tainly knew how to build segmented-arched bridges, as at­tested by sur­viv­ing ex­am­ples, three of which are listed by Wurster and Ganz­ert them­selves.^[24] An ear­lier con­struc­tion date in the 2nd or 3rd cen­turies AD can­not there­fore be ex­cluded.^[25] The case for an ear­lier con­struc­tion date has been strength­ened in re­cent years with the dis­cov­ery of fur­ther seven Roman seg­mented arch bridges.^[26] The rem­nants of the nearby Kemer Bridge over the river Xanthos, which dates to the 3rd century and has some ar­chi­tec­tural sim­i­lar­i­ties with the Limyra bridge, could fur­ther re­in­force that case.^[27]