**4. A second approach: a search for survey and draft step details**

Due to their representation scale and their aims, the maps could be regarded as true precursors of the modern "technical maps". But it is likely that these wide-sized maps were simply due to a scenic use of the products (e.g. for a good view during technical meetings held in the presence of political authorities) rather than to a real technical need. Unfortunately, up to now we possess neither the original field drafts nor the related field notebook that could allow us to do an enormous quality jump in the detailed inspection of the map generation process. Thus, any other analytical attempt must be taken into account to try to find out information about survey and draft phases.

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 83

82 Cartography – A Tool for Spatial Analysis

intuitive way (Jenny & Hurny, 2011).

the first and the third map.

landscape.

the scale of the ancient map), such as the rotation angle with respect to the present cartographic North, the scale variation throughout the map and the distortion of the present cartographic grid as resulting after its adaption onto the old map. The process is performed recognizing in the ancient map a proper number of still existing points, whose cartographic coordinates can be derived from the present cartography. A specifically designed software tool can result very useful for analysis of map deformations, allowing the calculation of all the aforementioned parameters and their subsequent drafting in an

The study of map deformation characterising the cartographic samples here analysed showed scale factors quite variable throughout the maps, being slightly more homogenous in F map than in P map. The average scale resulting from the calculation was 1:12,300 (1:14,300 ÷ 1:10,300) and 1:13,400 (1:16,600 ÷ 1:10,200) in F and P maps, respectively. Notwithstanding this, the former map showed two severe anomalous variation areas near the northern and southern delta lobe corners (Bitelli et al. 2009). In particular, the northern gross deformation affecting F map is supposed to derive from a shift of the drawing, intentionally made by the author for unclear reasons. L map presented a bit more constrained scale factor (mean 1:11,200, range 1:13,200 ÷ 1:11,200) and a smaller deformation than the other two, but it has to be taken into account that L map depicts a smaller area in respect to F and P maps (Figure 4). Moreover, the calculated rotation angles were 15.7°, 8.9° and 3.6° for F, P and L maps, respectively: they indicate an angular displacement of about 7° between the first and the second map, and 12° between

In this specific case, where large areas, depicted in the maps, correspond to disappeared coastal (i.e. peripheral) belts, points suitable to be used as GCPs cannot be found in the present landscape, and the insertion of GCPs all around the deltaic area becomes obviously impossible. The metric analysis that was possible to perform on the set of maps highlighted gross deformations in all maps (especially in F and P maps). Notwithstanding some differences in the results showed by the maps, this analysis alone resulted insufficient to state which map has to be considered as the most faithful to the real asset of the ancient

Thus, a question still remains open: why does exist a family of so severe deformations?

Due to their representation scale and their aims, the maps could be regarded as true precursors of the modern "technical maps". But it is likely that these wide-sized maps were simply due to a scenic use of the products (e.g. for a good view during technical meetings held in the presence of political authorities) rather than to a real technical need. Unfortunately, up to now we possess neither the original field drafts nor the related field notebook that could allow us to do an enormous quality jump in the detailed inspection of the map generation process. Thus, any other analytical attempt must be taken into account

**4. A second approach: a search for survey and draft step details** 

to try to find out information about survey and draft phases.

**Figure 4.** Graphical results from a deformation analysis performed on the three maps (see Jenny & Hurny 2011 for software description): in blue the UTM-ED50 grid (mesh size = 2 km), in red the residual vectors on the GCPs, in yellow the scale isolines, in orange some values of the calculated map scale.

In order to achieve this aim, late Renaissance land-surveying techniques (in particular the use of the *squadra zoppa* probably invented by Fabri himself) were studied and linked to technical signs preserved in the maps. Various signs, preserved either in the palimpsest of the maps or in their final drawing, were taken into account, and nine classes of evidence have been recognized (Bitelli et al. 2010): i) written information; ii) technical grids; iii) topographic measurements; iv) sighting tracks; v) "lost landmarks"; vi) preparing/correcting/updating drawings; vii) additional iconography; viii) unresolved questions; ix) restoration problems. The meaning of some signs is partly explained in the legend, whereas the meaning of others can be inferred by Fabri's methodological textbook. They will be briefly analysed in the next paragraphs.

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 85

A small-sized square mesh grid, drawn in charcoal, is distinctly visible on all the three maps (Figure 6a). The grids are characterized by different cell sizes: 10 cm on paper (about 1 km in world) in F map, 7.7 cm (about 1 km) in P map, 4 cm (about 400 m) in L map. In the first case, the cell size is exactly as long as half of the scale bar. This kind of grid covers the whole map and is drawn parallel to the North branches of the compass-card and to the map edges. Hence, the grid was probably helpful in assisting the transposition of the field-bearings onto the map support. We cannot state if the numbering of the grid squares (characterising the P map) had been used for this or other purposes, such as a mere scale reduction for copying.

Another very large sized grid exists on F map alone, and it is characterized by cells about five times larger than those of the previous one (Figure 6b). The grid orientation does not fit with the magnetic North, but coincides with a topographic direction clearly identifiable with the most rectilinear street reach located along a wide artificial canal (*Rettinella*) between the towns of *Loreo* and *Tornova*. It can be supposed that the canal was the starting baseline for the survey of the central part of the mapped lands, according to the forward intersection scheme showed by Fabri in his textbook (Figure 2c), assuming the still existing *Loreo* and *Tornova* bell towers as main reference points (Figure 2c, A and B points). Therefore, this second grid could have probably been used by the map-maker to draw on the map the

azimuths measured in field and the sighting tracks toward all the noticeable points.

**Figure 6.** Linedraw of the technical grids visible in F map, superimposed on the map. In (b) the

Topographic measurements can be subdivided into three classes (Figure 7):

*Rettinella* canal is roughly drafted in a marginal area of the map).

*Rettinella* canal is highlighted in red.

**4.3. Topographic measurements** 

It is not possible to state whether the field survey strategy was the same for the other maps, as in them a similar large grid does not exist. If it was, the same baseline would have probably been used in P map (where it is clearly recognizable) but not in L map (where the

**4.2. Technical grids** 

#### **4.1. Written information**

In their cartouche, F and L maps state that the documents were made by merging some previous maps or other cartographic drawings and partly by means of direct topographic measurements (Figure 5). This kind of capital information helps us to understand that pregeodetic maps can be composite products (at least in the case of large scale maps), therefore they can be very difficult to be analyzed and their usual georeferencing could be partially inappropriate. L map cartouche states that different surveying techniques were adopted according to the different interest level of the various topographic domains (see Paragraph 4.3). Unfortunately, due to the low resolution of the digital image, on P map we cannot read the text content; moreover it seems to be an unfinished product regarding the cartouche decorations. Besides, the fundamental message is that the maps can be considered a sort of patchwork probably generated by merging some local maps, previously surveyed by the authors themselves or other colleagues, or partially *ad hoc* surveyed.

**Figure 5.** The cartouche in the three maps.

## **4.2. Technical grids**

84 Cartography – A Tool for Spatial Analysis

**4.1. Written information** 

**Figure 5.** The cartouche in the three maps.

In order to achieve this aim, late Renaissance land-surveying techniques (in particular the use of the *squadra zoppa* probably invented by Fabri himself) were studied and linked to technical signs preserved in the maps. Various signs, preserved either in the palimpsest of the maps or in their final drawing, were taken into account, and nine classes of evidence have been recognized (Bitelli et al. 2010): i) written information; ii) technical grids; iii) topographic measurements; iv) sighting tracks; v) "lost landmarks"; vi) preparing/correcting/updating drawings; vii) additional iconography; viii) unresolved questions; ix) restoration problems. The meaning of some signs is partly explained in the legend, whereas the meaning of others can be inferred by Fabri's

In their cartouche, F and L maps state that the documents were made by merging some previous maps or other cartographic drawings and partly by means of direct topographic measurements (Figure 5). This kind of capital information helps us to understand that pregeodetic maps can be composite products (at least in the case of large scale maps), therefore they can be very difficult to be analyzed and their usual georeferencing could be partially inappropriate. L map cartouche states that different surveying techniques were adopted according to the different interest level of the various topographic domains (see Paragraph 4.3). Unfortunately, due to the low resolution of the digital image, on P map we cannot read the text content; moreover it seems to be an unfinished product regarding the cartouche decorations. Besides, the fundamental message is that the maps can be considered a sort of patchwork probably generated by merging some local maps, previously surveyed by the

methodological textbook. They will be briefly analysed in the next paragraphs.

authors themselves or other colleagues, or partially *ad hoc* surveyed.

A small-sized square mesh grid, drawn in charcoal, is distinctly visible on all the three maps (Figure 6a). The grids are characterized by different cell sizes: 10 cm on paper (about 1 km in world) in F map, 7.7 cm (about 1 km) in P map, 4 cm (about 400 m) in L map. In the first case, the cell size is exactly as long as half of the scale bar. This kind of grid covers the whole map and is drawn parallel to the North branches of the compass-card and to the map edges. Hence, the grid was probably helpful in assisting the transposition of the field-bearings onto the map support. We cannot state if the numbering of the grid squares (characterising the P map) had been used for this or other purposes, such as a mere scale reduction for copying.

Another very large sized grid exists on F map alone, and it is characterized by cells about five times larger than those of the previous one (Figure 6b). The grid orientation does not fit with the magnetic North, but coincides with a topographic direction clearly identifiable with the most rectilinear street reach located along a wide artificial canal (*Rettinella*) between the towns of *Loreo* and *Tornova*. It can be supposed that the canal was the starting baseline for the survey of the central part of the mapped lands, according to the forward intersection scheme showed by Fabri in his textbook (Figure 2c), assuming the still existing *Loreo* and *Tornova* bell towers as main reference points (Figure 2c, A and B points). Therefore, this second grid could have probably been used by the map-maker to draw on the map the azimuths measured in field and the sighting tracks toward all the noticeable points.

It is not possible to state whether the field survey strategy was the same for the other maps, as in them a similar large grid does not exist. If it was, the same baseline would have probably been used in P map (where it is clearly recognizable) but not in L map (where the *Rettinella* canal is roughly drafted in a marginal area of the map).

**Figure 6.** Linedraw of the technical grids visible in F map, superimposed on the map. In (b) the *Rettinella* canal is highlighted in red.

#### **4.3. Topographic measurements**

Topographic measurements can be subdivided into three classes (Figure 7):


In the latter case, the mapmaker clearly highlighted which topographic reaches are the best surveyed, by means of red lines drawn onto the map. In those reaches pickets were certainly located and their related distances measured. According to that, each of these points had to be further sighted from a standpoint lying on the opposite riverside.

Therefore, this class of evidences states the fundamental technical character of the maps as a whole or for some selected subareas.

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 87

Some sighting tracks (sometimes identified by a number and a letter) are clearly visible, in particular in L map (Figure 8) and occasionally in P map. These tracks are preserved in sea areas lying between the Po delta and the continental coast reaches in the northern and southern part of the delta lobe. They try to define the correct distance existing between selected points lying on the fast-prograding delta coast and well-visible landmarks lying on the opposite coastal domain, e.g. *Mesola* defense towers and *Chioggia* bell towers. The length of the longest tracks is about 10 km, a plausible value for field surveys in marine environment free of the vegetative mantle (thus characterized by a good inter-visibility of

This kind of information could suggest the author's need to control the existence of gross

Other signs of unclear value are visible in F map. They are not preserved in today landscape, thus they should have not been true *landmarks* (Figure 9): for this reason we prefer to use the term *lost landmarks* for their identification. They consist in few small crosses and a large one located at the diffluence point of two river branches of the *Po di Levante*. The large cross was probably a real reference-point as this object gave its name to a salt marsh (*Polesine de la Crose*) and to a still preserved building. The small crosses are located in desert areas of the E and SE realms of the delta lobe, along a red double curvilinear track. The meaning of the curvilinear tracks has been deduced from the comparison with another map of the same area dating back to the year 1608: they define the boundaries of the lands

This class of evidences highlights how difficult the ancient land surveying was, due to lacking of tall and fixed benchmarks in high-speed changing areas. On the other hand, it highlights how much important, from an economic point of view, those new generated areas were for the local government, thus suggests a possible reason to start the survey

**4.4. Sighting tracks** 

the points).

errors in the conclusive survey phase.

**Figure 8.** Examples of sighting tracks visible in L map.

confiscated by the Venice Government in 1588.

**4.5. "Lost landmarks"** 

**Figure 7.** Topographic measurements in L map.

#### **4.4. Sighting tracks**

86 Cartography – A Tool for Spatial Analysis

mouth of the *Tramontana* branch.

whole or for some selected subareas.

**Figure 7.** Topographic measurements in L map.

i. explicit length measurements, which are clearly written as numbers of ancient Venice paces (1 Venice pace = 1.738674 m, as stated in Martini, 1883), e.g. in the area of the

ii. implicit length measurements, e.g. along the riverside where the points are graphically

iii. local bathymetric surveys and some selected riverbed transects, in particular at the river

In the latter case, the mapmaker clearly highlighted which topographic reaches are the best surveyed, by means of red lines drawn onto the map. In those reaches pickets were certainly located and their related distances measured. According to that, each of these points had to

Therefore, this class of evidences states the fundamental technical character of the maps as a

located in order to define the width and the geometry of the riverbanks;

*Pertegado* ditch, whose name properly means "measured";

be further sighted from a standpoint lying on the opposite riverside.

Some sighting tracks (sometimes identified by a number and a letter) are clearly visible, in particular in L map (Figure 8) and occasionally in P map. These tracks are preserved in sea areas lying between the Po delta and the continental coast reaches in the northern and southern part of the delta lobe. They try to define the correct distance existing between selected points lying on the fast-prograding delta coast and well-visible landmarks lying on the opposite coastal domain, e.g. *Mesola* defense towers and *Chioggia* bell towers. The length of the longest tracks is about 10 km, a plausible value for field surveys in marine environment free of the vegetative mantle (thus characterized by a good inter-visibility of the points).

This kind of information could suggest the author's need to control the existence of gross errors in the conclusive survey phase.

**Figure 8.** Examples of sighting tracks visible in L map.

#### **4.5. "Lost landmarks"**

Other signs of unclear value are visible in F map. They are not preserved in today landscape, thus they should have not been true *landmarks* (Figure 9): for this reason we prefer to use the term *lost landmarks* for their identification. They consist in few small crosses and a large one located at the diffluence point of two river branches of the *Po di Levante*. The large cross was probably a real reference-point as this object gave its name to a salt marsh (*Polesine de la Crose*) and to a still preserved building. The small crosses are located in desert areas of the E and SE realms of the delta lobe, along a red double curvilinear track. The meaning of the curvilinear tracks has been deduced from the comparison with another map of the same area dating back to the year 1608: they define the boundaries of the lands confiscated by the Venice Government in 1588.

This class of evidences highlights how difficult the ancient land surveying was, due to lacking of tall and fixed benchmarks in high-speed changing areas. On the other hand, it highlights how much important, from an economic point of view, those new generated areas were for the local government, thus suggests a possible reason to start the survey

works. A question remains open: were those landmarks the only ones existing or are they simply a selected set?

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 89

original analogical document in ASVe. We could ask ourselves why these symbols were stuck on the map and not directly drawn on it, whereas a very beautiful fish image (a

The question is not out of sense because it can suggest chronological problems, also related

**Figure 11.** Examples of additional iconography in F map: appliqués showing ancient vessels and the

Each map derives from the assemblage of single sheets on a canvas, clearly identifiable thanks to the visible seams (Figure 12). As a result, the homologue tracts of the drawing do not always fit well together. Hence, a problem arises concerning the age of the last sheet assemblage: does the assemblage date back to the original edition of the map or to a more

sturgeon) was drawn in front of the *Tramontana* branch mouth (Figure 11b).

to tracking of the reference grid (Figure 11a).

**Figure 10.** Examples of updating riverbed details in P and L maps.

image of a sturgeon.

**4.8. Unresolved questions** 

recent one, being related to a series of restorations?

**Figure 9.** Signs of unclear value (*lost landmarks*) drawn in F map: a large cross at the diffluence point of two river branches, and a small cross lying on a red double curvilinear track.

#### **4.6. Preparing/correcting/updating drawings**

In each map, former drawings of topographical elements can be observed at various places. They represent details of riverbeds or local ways and were probably drawn as a first approximation of topographic elements. In the last definitive drawing, they appear corrected in their shape and/or location (Figure 10).

It may be believed that they represent the first attempt at defining the topographic design in those parts of the map derived from previous cartographic documents (as stated in the cartouche of F and L maps, see Paragraph 4.1). In some other cases, these kinds of correction probably referred to real environmental changes developed over time, as in the case of the minor branches of the *Scirocco Po* branch near *Mea* place (e.g. in P map, see Figure 10a).

It is difficult to conceal that these features represent an interpretative problem, as they suggest a high approximation degree in the drafting of the topographic details.

#### **4.7. Additional iconography**

In F map a very particular detail consists in a number of appliqués showing ancient vessels (Figure 11a). It was possible to notice such a detail only thanks to a direct consultation of the original analogical document in ASVe. We could ask ourselves why these symbols were stuck on the map and not directly drawn on it, whereas a very beautiful fish image (a sturgeon) was drawn in front of the *Tramontana* branch mouth (Figure 11b).

The question is not out of sense because it can suggest chronological problems, also related to tracking of the reference grid (Figure 11a).

**Figure 10.** Examples of updating riverbed details in P and L maps.

**Figure 11.** Examples of additional iconography in F map: appliqués showing ancient vessels and the image of a sturgeon.

#### **4.8. Unresolved questions**

88 Cartography – A Tool for Spatial Analysis

simply a selected set?

works. A question remains open: were those landmarks the only ones existing or are they

**Figure 9.** Signs of unclear value (*lost landmarks*) drawn in F map: a large cross at the diffluence point of

In each map, former drawings of topographical elements can be observed at various places. They represent details of riverbeds or local ways and were probably drawn as a first approximation of topographic elements. In the last definitive drawing, they appear

It may be believed that they represent the first attempt at defining the topographic design in those parts of the map derived from previous cartographic documents (as stated in the cartouche of F and L maps, see Paragraph 4.1). In some other cases, these kinds of correction probably referred to real environmental changes developed over time, as in the case of the minor branches of the *Scirocco Po* branch near *Mea* place (e.g. in P map, see Figure 10a).

It is difficult to conceal that these features represent an interpretative problem, as they

In F map a very particular detail consists in a number of appliqués showing ancient vessels (Figure 11a). It was possible to notice such a detail only thanks to a direct consultation of the

suggest a high approximation degree in the drafting of the topographic details.

two river branches, and a small cross lying on a red double curvilinear track.

**4.6. Preparing/correcting/updating drawings** 

corrected in their shape and/or location (Figure 10).

**4.7. Additional iconography** 

Each map derives from the assemblage of single sheets on a canvas, clearly identifiable thanks to the visible seams (Figure 12). As a result, the homologue tracts of the drawing do not always fit well together. Hence, a problem arises concerning the age of the last sheet assemblage: does the assemblage date back to the original edition of the map or to a more recent one, being related to a series of restorations?

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 91

landmarks), or, finally, the involvement of more than one person in the surveying works. As explicitly stated in some cartouches, the morphology of some depicted land areas can have been taken from other previously developed cartographies; those areas are not explicitly highlighted by the authors, but it is possible to infer their locations by excluding zones where explicit signs refer to specific surveying techniques (i.e. sighting tracks, length measurements, red reaches, bathymetric surveys). It can be hypothesized that some topographical elements characterized by former drawings (e.g. in L map, see Paragraph 4.6) were derived from previous cartographic documents. In fact, some visible former drawings (e.g. Figure 10b) are placed in peripheral areas where residuals from the georeferencing

Other kinds of former drawings probably refer to corrections made by the authors over time: they can denote real environmental changes in shape and/or location of the topographic elements, or also intentionally-made corrections. The first ones can be highlighted by a roughly deletion of previously drawn elements (see Figure 10a), whereas the second ones seem to have been hidden by the authors by means of a brush stroke covering. This is the case of an astonishing map "correction", well recognizable in F map (Figure 1a), probably done at an early drawing stage. For that detail it was thought that the northern cluster of islands had been northward-shifted by the author for an unclear reason. This could be supported by both the comparison with the other maps and the existence of a wide area with no clear topographic detail lying immediately south of the delta coast. Hence, a possible verisimilar original map appearance could be that shown in Figure 13: it is a true "false case", generated by means of a simple "surgical" operation performed in a digital environment, in the opposite direction compared to the alleged author intervention. It must be noticed that the southward-shifted cluster of islands coincides perfectly in its

The above described operation could now be attempted in all the other areas of F map with already evidenced anomalies (Cremonini & Samonati, 2009), allowing a complete regeneration of this ancient map. This could be an interesting way of approaching the study

The presence of intentionally-made corrections could explain why some areas, differently represented in F map in respect to P map (whereas in L map the represented land is smaller), maintain an unclear post-georeferencing deformation in F map. In order to demonstrate this assumption, a comparison of the new product (the "false case") with present cartography via modern georeferencing methods was made. In Figure 14 the overlay of the "false case" on present high resolution satellite images (*Bing Maps*TM environment) is reported; the map was georeferenced by means of a second order polynomial transformation, as done for the original map (see Figure 3b). An analysis of residuals of georeferencing could be the best way to check a real improvement in metric quality of the new cartographic product in respect to the original one, but in this particular

process are higher than those obtained in other areas of the map.

morphology with the part of islands remained untouched.

**5.1. Georeferencing of the "false case"** 

of pre-geodetic maps.

**Figure 12.** Visible seams in F and L maps, due to the assemblage of single sheets on the canvas: the homologue tracts of the drawing do not always fit exactly together.

As what we can see today is the product of the long lifespan of the maps, a fundamental question arises concerning type, number and entity of the restorations performed on them. Each restoration probably slightly changed the map almost imperceptibly. This may be true also for the smallest details, and it could partially deny the possibility of performing a correct diachronic study of the original document.

## **5. A discussion and a proposal**

Notwithstanding some still open questions, the whole set of the discussed details highlights a very interesting class of information related to the hidden steps of pre-geodetic map construction. In particular, these details highlight the presence in the map of areas surveyed by means of different methodologies. This fact can explain the different entity of georeferencing residuals characterising different areas as well as the presence of high deformations in some particular areas rather than in others. The use of different survey methodologies can be due to various reasons, such as different levels of interest in the representation of different zones (e.g. areas confiscated by the Venice government, or areas planned for the forthcoming Po river diversion), or difficulties in surveying some others (e.g. areas too far from a selected reference baseline, or without well-recognizable landmarks), or, finally, the involvement of more than one person in the surveying works. As explicitly stated in some cartouches, the morphology of some depicted land areas can have been taken from other previously developed cartographies; those areas are not explicitly highlighted by the authors, but it is possible to infer their locations by excluding zones where explicit signs refer to specific surveying techniques (i.e. sighting tracks, length measurements, red reaches, bathymetric surveys). It can be hypothesized that some topographical elements characterized by former drawings (e.g. in L map, see Paragraph 4.6) were derived from previous cartographic documents. In fact, some visible former drawings (e.g. Figure 10b) are placed in peripheral areas where residuals from the georeferencing process are higher than those obtained in other areas of the map.

Other kinds of former drawings probably refer to corrections made by the authors over time: they can denote real environmental changes in shape and/or location of the topographic elements, or also intentionally-made corrections. The first ones can be highlighted by a roughly deletion of previously drawn elements (see Figure 10a), whereas the second ones seem to have been hidden by the authors by means of a brush stroke covering. This is the case of an astonishing map "correction", well recognizable in F map (Figure 1a), probably done at an early drawing stage. For that detail it was thought that the northern cluster of islands had been northward-shifted by the author for an unclear reason. This could be supported by both the comparison with the other maps and the existence of a wide area with no clear topographic detail lying immediately south of the delta coast. Hence, a possible verisimilar original map appearance could be that shown in Figure 13: it is a true "false case", generated by means of a simple "surgical" operation performed in a digital environment, in the opposite direction compared to the alleged author intervention. It must be noticed that the southward-shifted cluster of islands coincides perfectly in its morphology with the part of islands remained untouched.

The above described operation could now be attempted in all the other areas of F map with already evidenced anomalies (Cremonini & Samonati, 2009), allowing a complete regeneration of this ancient map. This could be an interesting way of approaching the study of pre-geodetic maps.

#### **5.1. Georeferencing of the "false case"**

90 Cartography – A Tool for Spatial Analysis

**Figure 12.** Visible seams in F and L maps, due to the assemblage of single sheets on the canvas: the

As what we can see today is the product of the long lifespan of the maps, a fundamental question arises concerning type, number and entity of the restorations performed on them. Each restoration probably slightly changed the map almost imperceptibly. This may be true also for the smallest details, and it could partially deny the possibility of performing a

Notwithstanding some still open questions, the whole set of the discussed details highlights a very interesting class of information related to the hidden steps of pre-geodetic map construction. In particular, these details highlight the presence in the map of areas surveyed by means of different methodologies. This fact can explain the different entity of georeferencing residuals characterising different areas as well as the presence of high deformations in some particular areas rather than in others. The use of different survey methodologies can be due to various reasons, such as different levels of interest in the representation of different zones (e.g. areas confiscated by the Venice government, or areas planned for the forthcoming Po river diversion), or difficulties in surveying some others (e.g. areas too far from a selected reference baseline, or without well-recognizable

homologue tracts of the drawing do not always fit exactly together.

correct diachronic study of the original document.

**5. A discussion and a proposal** 

The presence of intentionally-made corrections could explain why some areas, differently represented in F map in respect to P map (whereas in L map the represented land is smaller), maintain an unclear post-georeferencing deformation in F map. In order to demonstrate this assumption, a comparison of the new product (the "false case") with present cartography via modern georeferencing methods was made. In Figure 14 the overlay of the "false case" on present high resolution satellite images (*Bing Maps*TM environment) is reported; the map was georeferenced by means of a second order polynomial transformation, as done for the original map (see Figure 3b). An analysis of residuals of georeferencing could be the best way to check a real improvement in metric quality of the new cartographic product in respect to the original one, but in this particular

case the test results impossible, due to the lack of still existing reference points in the deltaic area. Therefore, only a comparison between coeval maps is allowed. In Figure 15 the comparison between the "false case" and the georeferenced P map (by means of a second order polynomial transformation) is showed. In this case, the increase in correspondence for the northern cluster of islands details between the two maps is very high (see Figure 15b). For the purpose, a vectorization of P map was used in order to better visualize the overlay. Unfortunately, L map covers a smaller land area, and other coeval maps depicting the same deltaic area do not appear to exist: thus, the comparison with P map only is possible and significant.

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 93

the location of the same coastline (2 up to 3.5 km) in the same year 1592 remains between F and P maps, therefore neither reliable coastal erosion rates (Figure 16) nor a correct pointby-point response of the ancient beach line to the wind-induced long shore currents (Figure

**Figure 14.** Overlay of the "false case", georeferenced by means of a second order polynomial transformation, on present high resolution satellite images (*Bing Maps*TM environment).

17) can be believably proposed.

**Figure 13.** A true "false case": the first attempt at restoring the possible original appearance of the former real map configuration before the alleged author intervention (in yellow the part translated and rotated from the white original location). North is right.

#### **5.2. Towards possible future developments**

The quite simple analysis above proposed could be considered as a stimulus toward a new way of approaching the study of pre-geodetic cartography: an approach capable to deeply enter the factual map genesis processes. In fact, a mere mathematical approach could not be satisfying for situations similar to the above discussed. As shown in Figures 16 and 17, the attempt of mutual comparison among the georeferenced coastline locations for the three studied maps could remain widely questionable. In these case, in fact, a gross discrepancy in the location of the same coastline (2 up to 3.5 km) in the same year 1592 remains between F and P maps, therefore neither reliable coastal erosion rates (Figure 16) nor a correct pointby-point response of the ancient beach line to the wind-induced long shore currents (Figure 17) can be believably proposed.

92 Cartography – A Tool for Spatial Analysis

significant.

case the test results impossible, due to the lack of still existing reference points in the deltaic area. Therefore, only a comparison between coeval maps is allowed. In Figure 15 the comparison between the "false case" and the georeferenced P map (by means of a second order polynomial transformation) is showed. In this case, the increase in correspondence for the northern cluster of islands details between the two maps is very high (see Figure 15b). For the purpose, a vectorization of P map was used in order to better visualize the overlay. Unfortunately, L map covers a smaller land area, and other coeval maps depicting the same deltaic area do not appear to exist: thus, the comparison with P map only is possible and

**Figure 13.** A true "false case": the first attempt at restoring the possible original appearance of the former real map configuration before the alleged author intervention (in yellow the part translated and

The quite simple analysis above proposed could be considered as a stimulus toward a new way of approaching the study of pre-geodetic cartography: an approach capable to deeply enter the factual map genesis processes. In fact, a mere mathematical approach could not be satisfying for situations similar to the above discussed. As shown in Figures 16 and 17, the attempt of mutual comparison among the georeferenced coastline locations for the three studied maps could remain widely questionable. In these case, in fact, a gross discrepancy in

rotated from the white original location). North is right.

**5.2. Towards possible future developments** 

**Figure 14.** Overlay of the "false case", georeferenced by means of a second order polynomial transformation, on present high resolution satellite images (*Bing Maps*TM environment).

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 95

**Figure 16.** Coastline shape from the georeferencing of F, P and L maps on Italian modern cartography

(Cremonini and Samonati, 2009; Cremonini 2010, Fig. 2).

**Figure 15.** Overlay of a georeferenced vectorization of the P map (in blue) on the "false case".

To overcome this impasse, a complex methodological approach would be adopted: it would be capable to check and take into account the metric information still preserved in the ancient maps, though partially hidden, and finally to regenerate the lacking topographic details or relocate them in a more reliable way. Only at that stage, a new phase of the study, based on a new and more reliable ancient map, will be triggered for various classes of researchers.

researchers.

**Figure 15.** Overlay of a georeferenced vectorization of the P map (in blue) on the "false case".

To overcome this impasse, a complex methodological approach would be adopted: it would be capable to check and take into account the metric information still preserved in the ancient maps, though partially hidden, and finally to regenerate the lacking topographic details or relocate them in a more reliable way. Only at that stage, a new phase of the study, based on a new and more reliable ancient map, will be triggered for various classes of

**Figure 16.** Coastline shape from the georeferencing of F, P and L maps on Italian modern cartography (Cremonini and Samonati, 2009; Cremonini 2010, Fig. 2).

Analysis of Pre-Geodetic Maps in Search of Construction Steps Details 97

Finally, the study shows that pre-geodetic maps usually should not be regarded as completely original technical products but rather as an assemblage of various data-sets coming from previous sources. Thus, a study of ancient cartography should start from a philological analysis of the maps (i.e. search for origin, target, models, cultural atmosphere of that time) and it should subsequently become a "stratigraphic" reading of both relationships existing among the drawing details and accidents affecting the maps. A thorough knowledge of the ancient field surveying techniques (i.e. pathways, distances, contemporaneous use of various kinds of instruments, previous cartographic sources, etc.) is essential in order to understand genesis and size of the surveying errors recorded in the map, and therefore it is also essential to direct the georeferencing techniques towards the definition of the real topography and morphology of those areas that today no longer exist, e.g. due to natural environmental processes. In other words, the technical approach to ancient cartography cannot be reduced to a simple georeferencing attempt; the georeferencing should rather act as an essential support for a deep analysis of the maps, that

We are grateful to the Venice National Archive, for the consultation of the map by Fabri and the map by Lorini et al., and to the Engineering Faculty Library *Gian Paolo Dore* of the

Adcock E.P, Varlamoff M.T, Kremp V (2004) *Principi dell'IFLA per la cura e il trattamento dei materiali di biblioteca.* Bari: IFLA (International Federation of Library Associations and

Balletti C (2006) Georeference in the analysis of the geometric content of early maps. e-

Benavides J., Koster E. (2006) Identifying surviving landmarks on historical maps, e-

in this way can become a more useful tool to various classes of researchers.

*Department of Civil, Environmental and Materials Engineering (DICAM),* 

University of Bologna, for the consultation of the textbook by O. Fabri.

*Alma Mater Studiorum - University of Bologna, Bologna, Italy* 

*Department of Earth and Geological-Environmental Sciences, Alma Mater Studiorum - University of Bologna, Bologna, Italy* 

Perimetron, Vol. 1, No. 1. pp 32-42. Available: http://www.e-perimetron.org/Vol\_1\_1/Vol1\_1.htm.

Perimetron, Vol. 1, No. 3. pp 287-296. Available: http://www.e-perimetron.org/Vol\_1\_3/Vol1\_3.htm.

**Author details** 

Stefano Cremonini

**Acknowledgement** 

**7. References** 

Institutions).

Gabriele Bitelli and Giorgia Gatta

**Figure 17.** Long shore current directions induced by Bura and Sirocco winds on F map coastline profile (Cremonini and Samonati 2009; Cremonini 2010, Fig. 1).

### **6. Conclusion**

In the study, careful recovery and analysis of some data concealed in three sixteenth century maps were done, and various technical signs preserved in them were identified and discussed. All these details show interesting information concerning the original survey performed to generate the maps, and help to comprehend the differences in the represented coastal areas among the analysed coeval maps. Notwithstanding the analysis cannot comprehensively explain all questions concerning the eroded coastal areas, due to the fact that existing landmarks cannot be found in disappeared areas, an attempt to explain the difference in the morphology of the coast was made. On the base of the recognition of a gross deformation in one map, a good example of a possible alleged author's intervention was given, attempting to restore the original appearance of the possible former map configuration. Finally, a comparison of the new product with present cartography and other coeval maps, via modern georeferencing methods, was able to support the hypothesis.

Finally, the study shows that pre-geodetic maps usually should not be regarded as completely original technical products but rather as an assemblage of various data-sets coming from previous sources. Thus, a study of ancient cartography should start from a philological analysis of the maps (i.e. search for origin, target, models, cultural atmosphere of that time) and it should subsequently become a "stratigraphic" reading of both relationships existing among the drawing details and accidents affecting the maps. A thorough knowledge of the ancient field surveying techniques (i.e. pathways, distances, contemporaneous use of various kinds of instruments, previous cartographic sources, etc.) is essential in order to understand genesis and size of the surveying errors recorded in the map, and therefore it is also essential to direct the georeferencing techniques towards the definition of the real topography and morphology of those areas that today no longer exist, e.g. due to natural environmental processes. In other words, the technical approach to ancient cartography cannot be reduced to a simple georeferencing attempt; the georeferencing should rather act as an essential support for a deep analysis of the maps, that in this way can become a more useful tool to various classes of researchers.

#### **Author details**

96 Cartography – A Tool for Spatial Analysis

**Figure 17.** Long shore current directions induced by Bura and Sirocco winds on F map coastline profile

In the study, careful recovery and analysis of some data concealed in three sixteenth century maps were done, and various technical signs preserved in them were identified and discussed. All these details show interesting information concerning the original survey performed to generate the maps, and help to comprehend the differences in the represented coastal areas among the analysed coeval maps. Notwithstanding the analysis cannot comprehensively explain all questions concerning the eroded coastal areas, due to the fact that existing landmarks cannot be found in disappeared areas, an attempt to explain the difference in the morphology of the coast was made. On the base of the recognition of a gross deformation in one map, a good example of a possible alleged author's intervention was given, attempting to restore the original appearance of the possible former map configuration. Finally, a comparison of the new product with present cartography and other coeval maps, via modern georeferencing methods, was able to support the hypothesis.

(Cremonini and Samonati 2009; Cremonini 2010, Fig. 1).

**6. Conclusion** 

Gabriele Bitelli and Giorgia Gatta *Department of Civil, Environmental and Materials Engineering (DICAM), Alma Mater Studiorum - University of Bologna, Bologna, Italy* 

Stefano Cremonini *Department of Earth and Geological-Environmental Sciences, Alma Mater Studiorum - University of Bologna, Bologna, Italy* 
