GE is grateful to LGA, that has contributed, trough its researches and observations to determine some of the data here next reported. LGA additional research work has been donated to UNESCO.

For a better understanding of this chapter it is suggested to refer to design drawings RE-06 and RE-07 in which schemes and detailed representations are provided.

9.1 Bridge reinforced masonry

One of the most interesting constructive characteristic of the former bridge was the methodology of assembling the stones, performed with traditional mortar and additional metal elements placed across joints with the aim of providing a sort of reinforcing system which should have worked for the long lasting of the structure against river flows. Bridge stone elements were linked, with the adjacent ones, trough special metal devices like cramps and dowels that were anchored to stone blocks in purposely built slots in which it was poured melted lead to lock the whole system. The ancient techniques used at the time were undoubtedly advanced and a remarkable foreknowledge should be admitted: it seems that they were aware of the fragile behaviour of stone to tensile stresses and that a sort of implementation with a more ductile material like forged iron were a conscious device to gain a more suitable and resistance structure. Nowadays we can observe that those metal elements were not necessary for the stability and resistance of the bridge and that a good mortar layer, with a periodical maintenance of the joints, would have ensured anyhow the long lasting of the bridge. Nevertheless metal dowels may have contributed for the resistance of the structure to eventual share forces due to horizontal thrusts of seismic events and of river floods. Moreover cramps, over the extrados of the load bearing arch, behaved as if metal ties were put around the vault, and they may have given a contribute to face shifts and settlements of the ground and of the springers.

The stone arch structure stability is anyhow mostly related to the shape and curvature of the arch, (it is commonly said that an arch, either collapses when centering is dismantled, either lasts forever); while the stone arch structure resistance is mostly related to the resistance of the construction materials to compression stresses.

fig.01 – axonometric cut view of the stone elements of the bridge.

Metal strengthening devices were assembled with stones following accurate criteria, some of which are clearly visible and understandable, but some others have required a more detailed analysis and inquiry. These criteria have been quite difficult to be fully determined, either for the ruined current condition of the bridge, either for the ranges in which all the dimensions vary, that may wrongly lead to think that some choices had been performed randomly and not trough a plan or a methodology.

In this chapter have been reported the results of the studies concerning the assembling technique and the refine constructive method designed by an architect that has left a really impressive example of the knowledge of the time.

9.2 Voussoirs assemblage

Being the arch voussoirs the elements of the load bearing arch, and therefore the most important portion of the bridge, many reinforcing devices were present to ensure efficient joint connections.

• the load bearing arch was composed by 111 rows of voussoirs in a thickness of cm395;
• each row could contain 2-5 voussoirs (average 3-4);
• rows of voussoirs were mounted from both side of the bridge;
• joints, among voussoirs belonging to one row, were shifted compared to the preceding row, (as it happens in an ordinary masonry layout);
• a mortar layer was used to combine adjacent voussoirs, (as it happens in an ordinary masonry work);
• additional metal elements were mounted across the joints.
• three different strengthening metal devices were used for the arch stones: dowels, side cramps and extrados cramps;
• metal elements, and purposely built carvings to host them, were of dimensions and sizes quite variable in the bridge structure.

9.2.1 Dowels in the arch stones

Each arch stone was linked to one or two voussoirs of the preceding row by one or more metal dowels. Dowels were applied to the stone blocks in purposely carved slots, (holes), and melted lead was poured in the slots to connect metal and stone elements.

fig.02 – front view of the arch during the assembling with details of the arch stones and dowels;

• slots were slightly widen at the bottom to avoid untying;
• slots in the built-in stones (A-1°) were wider to allow the positioning and adjusting of the new stone (A-2°), and matching of the dowel with related slot;
• slots were of square/rectangular section as the dowels that were supposed to be inserted;
• slots were not polished but perfectly cleaned;
• dowel position should have been planned knowing all the joint positions of both of the rows to be connected, to avoid positioning of dowels next or over the joints;
• channels should have been protected to avoid their obstruction with mortar;
• channels were V shaped or U shaped, not polished but perfectly cleaned;
• lead, while melted, in many cases, has escaped from the slot going all around in the joint thickness and filling other gaps;
• dowels were of square-rectangular section profile with heads widen by hammering, to avoid untying; dowels were made in hand forged iron;
• top arch stone row was presumably without dowels to allow correct insertion of key stones;

fig.03 – two adjacent rows (side open view); top image: built-in row; below image: row to be assembled

• dowels were positioned at about cm40, from the extrados (which is half way between extrados and intrados); this matter may have been subjected to considerable variations at the springers: two anomalous samples have been surveyed on-site;
• dowel positioning criteria are represented in figure here below reported; those criteria were aimed at preserving the integrity of the stones by placing the dowels not too close to the stone edges which would have been weakened by the carved slots;
• dowels sizes were quite variable and with no apparent criteria: most probably because of ordinary constructive imperfections:
  • dowel section: square-rectangular
  • dowel heads: (cm 4-5) × (cm 4-5); (also cm 6 has been found)
  • dowel body section (midpoint): cm 3-4
  • dowel length: cm 15-20; (high variation and difficult evaluation of still inserted dowels);
• dowels longitudinal section was continuos and variation from head size to body size was reached gradually with a curved profile;

fig.04 – dowels positioning criteria (plan view): voussoir with more dowels: a>cm17 b>cm37; voussoir with one dowel: c˜d/2; edge voussoir: e˜cm37.

• dimensions of the smaller-slots, (in stone A-2°), were quite close to the dowel dimensions (cm 1-2 of tolerance: globally not per each side); of course the slot opening should have been wide enough to allow the entrance of the dowel head;
• dimensions of the wider-slots, (in stone A-1°), were remarkably bigger compared to the dowel size and were not regularly and accurately shaped; dimensions depended on needed adjustments: in the available samples cm1-2 more than the other slots have been surveyed, but variability is high;
• slots had their bottom widen in order that hardened lead would have avoided untying: lateral sides were accurately inclined of about 5° (few samples available);
• slots depth was about half of the dowel length, and, in the first stone in which it was inserted (A-2°), the dowel-head was touching the bottom of the slot in a way that sometimes it was not covered by melted lead; gradient of slots have been found in some cases but samples were not enough to work out a possible criteria;

fig.05 – plan view (section): a detail of a dowel

• channels were V and U shaped, but it seems that in the voussoirs were more U shaped: about cm2-2.5 wide and 1-1.5 depth. really very few samples are still available for inspection, (most of them are still assembled);
• channels were long enough to connect the extrados to the slot: about cm 40.

fig.06a – some recovered dowels winded by lead; in the first sample it is possible to note the different size and type of lead cover depending on the different slots; in the second sample it is possible to note the portion of lead pertaining to the channel; in third and fourth samples the dowel is half visible; in fourth sample the lead has spread all around in the joint.

fig.06b – some recovered dowels winded by lead; in the first sample it is possible to note the different size and type of lead cover depending on the different slots; in the second sample it is possible to note the portion of lead pertaining to the channel; in third and fourth samples the dowel is half visible; in fourth sample the lead has spread all around in the joint.

fig.07 – some recovered stones: the first one is stone A2 and the second is stone A8: it is possible to see on both of them the carved channels; in the third sample a very peculiar case has been found: the slot was really close to the edge and among the side cramps: channel (even if scarcely visible) had a gradient (about 30°) in order not to interfere with the cramps.

fig.08 – a dowel still inserted in the stone slot: it is visible its curved profile and its widen head presumably obtained by hammering (hand forging);

9.2.2 Side cramps in the arch stones

Adjacent voussoirs of each row were connected with double side cramps crossing the joints; even in this case cramps were linked to the stone by melted lead, poured in purposely built slots. Side cramps were quite regular, of similar shape and size and they had not so many variations along the load bearing arch.

assembling procedure:

• slots were carved over the side of each stone block across the joints among two adjacent voussoirs;
• for each joint four slots were carved, (two per side), and joint was in the middle;
• slots were carved only on the side of the row that looked towards the keystone;
• between two twin-slots, located on the opposite sides of the joint, a trace was carved to host the thickness of the cramp and to avoid any interference with the thickness of the joints among adjacent rows.
• once the voussoirs were correctly placed on site, the side cramps could be applied by pouring melted lead in the slots;

fig.09 – axonometric view of the load bearing arch during the assembling stage;

technical observations:

• slots were slightly widen at the bottom to avoid untying;
• slots were not polished but perfectly cleaned;
• slots were of rectangular section as the cramps edges that were supposed to be inserted;
• cramp positions depended only on the position of the joint to be connected;
• two rows of cramps were provided for each joint (presumably to avoid rotations);
• lead pouring was quite easy for all the rows of the arch, special devices might have been used for the rows close to the key stone because of the gradient of the slots; similar problems were probably present also for the channels of the dowels: there, the work was much easier at the top of the arch, (because of gravity), than at the springer where the channels had a very low gradient (almost horizontal);
• cramps were of rectangular section profile with extreme edges widen by hammering, to avoid untying; cramps were made in hand forged iron;
• top arch stone row was presumably without cramps to allow correct insertion of key stones;

fig.10 – plan view of a portion of the load bearing arch: side cramps are visible crossing the inner joints;

criteria and dimensions:

• the two side cramps divided the arch stone height in three equal parts: the distance from upper cramps to the extrados was approximately equal to the distance among the cramps, and equal to the distance among lower cramp and intrados: about cm 25-27; this matter may have been subjected to considerable variations at the springers, (two anomalous samples have been surveyed on-site);
• the side cramps positioning criteria was quite simple: two cramps per joint of similar dimensions;
• cramps sizes were slightly variable with no apparent criteria: most probably because of ordinary constructive imperfections:
  • cramp length: cm 30-35
  • cramp hooking-edges length: cm 6-8
  • cramp height: cm 4-5
  • cramp height (widen edges by hammering): cm 6-7
  • cramp thickness: cm 0.5-1
  • cramp thickness (thinner edges by hammering): cm 0.5-0.8
• cramps were quite straight and hosted in the carved traces;
• dimensions of the carved traces were as the cramp profile to be settled in;
• dimensions of the slots were quite close to the cramps dimensions; of course the slot opening should have been wide enough to allow the entrance of the cramp widen edge;
• slots had their bottom widen in order that hardened lead would have avoided untying;
• slots depth was enough to be suitable for the cramp edge to get in it;

Extrados cramps were located over the extrados of the load bearing arch: there were five rows of cramps that ran in parallel directions following the curved profile. The extrados cramps acted like tying chains, or at least they were conceived with this purpose, being one adjacent to each other. These cramps were different from all the others that have been found in the bridge stone structure, because they were dimensioned depending exclusively on the arch stone measures: where one cramp ended, the other started, sharing the same stone-slot to allow continuity of the tying action.

fig.11 – axonometric view of the load bearing arch during the assembling stage; extrados cramps are visible;

assembling procedure:

• slots were carved on the extrados of the voussoirs, following five parallel directions which ran at an equal distance;
• side joints of the voussoirs were carefully avoided, and because of this, sometimes the cramps direction slightly bent following a sort of a sinuous path;
• slots were carved approximately in the axe of each arch stone row;
• mostly for what concern the outer "chains" of cramps, (where lower cornices were supposed to be mounted), a sort of rough smoothing of the extrados was performed, and traces were carved to host the thickness of the cramps and to avoid any interference with the above stone elements to be assembled; for the inner chains of cramps this was partially performed;
• once the voussoirs were correctly placed on site, the extrados cramps could be applied by pouring melted lead in the slots;
• each slot was shared by two edges of adjacent cramps;
• cramps were slightly curved and adapted to the arch extrados profile; might have been also slightly hammered on site;

fig.12 – plan view of a portion of the load bearing arch: extrados cramps are visible;

fig.13 – a detail of the previous image;

technical observations:

• slots were slightly widen at the bottom to avoid untying;
• slots were not polished but perfectly cleaned;
• slots were of rectangular section as the cramps edges that were supposed to be inserted;
• cramp positions followed mainly five parallel directions almost regardless of the stones that were connected, (except for the position of the joints that were accurately avoided);
• five rows of cramps were provided as tying chains;
• lead pouring was quite easy for all the rows of the arch, special devices might have been used next to the arch springers;
• cramps were of rectangular section profile with extreme edges widen by hammering, to avoid untying; cramps were made in hand forged iron;
• cramps were all of different sizes depending on the voussoir dimensions, (different one to each other);
• cramps had to be slightly bent to match the extrados profile;

fig.14 – cut view: extrados cramps

criteria and dimensions:

• the side cramps positioning criteria was quite simple: the extrados cramps were assembled in five parallel chains, and distance among them was approximately cm80; the chains of cramps, located closer to the vault edges, were at a distance of about cm37 from the outer profile;
• extrados cramps had their slots approximately in the middle of the voussoirs;
• the slots were wide enough to host two adjacent cramps edges; of course the slot opening should have been wide enough to allow the entrance of the cramp widen edges;
• cramps were all of different lengths, being related to the voussoirs dimensions; other cramp sizes were variable with no apparent criteria: most probably because of ordinary constructive imperfections:
  • cramp length: variable
  • cramp hooking-edges length: cm 6-8
  • cramp height: cm 4-5
  • cramp height (widen edges by hammering): cm 6-7
  • cramp thickness: cm 0.5-1
  • cramp thickness (thinner edges by hammering): cm 0.5-0.8
• cramps were considerably bent and hammered to match the extrados profile; they were hosted in carved traces;
• dimensions of the carved traces were as the cramp profile to be settled in;
• slots had their bottom widen in order that hardened lead would have avoided untying;
• slots depth was enough to be suitable for the cramp edges to get in it;

fig.15a – cramps with their edges still covered by lead

fig.15b – cramps with their edges still covered by lead

fig.16a – view of portion of the vault (extrados side): extrados cramps are visible;

author of the text: arch. Manfredo Romeo – other contributes have been mentioned in related paragraphs

drawings and researches: arch. Paola Marrone  arch. Alessio Talarico    -   supervision: arch. Manfredo Romeo