The Fire of Saint Paul Outside the Walls: a very retrospective scientific investigation

Abstract

On the night of July 15, 1823, the papal basilica of St. Paul outside the walls of Rome was seriously damaged by a fire. There is an extensive bibliography on the history of the basilica up to the fire and on its reconstruction. On the contrary, the documents aimed at investigating the causes of the fire are limited in number and characterised by a historical approach. Furthermore, at the state of archival and bibliographic knowledge, there is no relevant text carried out according to the current criteria used for fire investigations. The research, starting from literature and archive documents, aims to investigate the causes of the fire, using the method of scientific investigation of the fire based on archival documents and the knowledge acquired in recent years through research on the dynamics of wood and vegetation fires.

Full text

The Fire of Saint Paul Outside the Walls: a very retrospective
scientific investigation
Stefano Marsella, Monica Calzolari
Rome, October 2023
of 1 26

Abstract ....................................................................................................................3
1. Introduction ..............................................................................................................3
1.1. The methodology 4
1.2. The investigative problem 5
1.3. The documents 5
2. Information analysis ..................................................................................................6
2.1. The basilica at the time of the fire 6
2.2. Location of the basilica 7
2.3. Location of the basilica 8
2.4. The nature of the maintenance work 8
2.5. The environmental conditions 9
2.6. The testimonials 11
2.7. The causes of the fire according to the chronicles 13
3. The hypothesis ........................................................................................................14
3.1. The first hypothesis: construction work fire 15
3.2. The second hypothesis: arson 16
4. Test of the hypotheses .............................................................................................16
4.1. Notes on the wood smouldering 18
4.2. Comments on the first hypothesis 20
4.3. Comments on the second hypothesis 21
5. Conclusions ...........................................................................................................22
6. Aknowledgements ...................................................................................................22
7. Chronicles ..............................................................................................................23
8. Bibliography ...........................................................................................................23
9. List of figures ..........................................................................................................25
10. List of Tables ..........................................................................................................26
of 2 26

Abstract
On the night of July 15, 1823, the papal basilica of St. Paul outside the walls of Rome was seriously damaged by a fire.
There is an extensive bibliography on the history of the basilica up to the fire and on its reconstruction. On the con-
trary, the documents aimed at investigating the causes of the fire are limited in number and characterised by a histor-
ical approach. Furthermore, at the state of archival and bibliographic knowledge, there is no relevant text carried out
according to the current criteria used for fire investigations. The research, starting from literature and archive docu-
ments, aims to investigate the causes of the fire, using the method of scientific investigation of the fire based on
archival documents and the knowledge acquired in recent years through research on the dynamics of wood and veget-
ation fires.
Figure 1 - View of the Basilica of Saint Paul Outside the Walls before the 1823 fire. G. Vasi, 1767.
1. Introduction
On the night of July 15, 1823, the papal basilica of St. Paul outside the walls of Rome was seriously damaged by fire.
The religious, artistic and historical importance of the building, together with the extent of the damages, gave the fire
worldwide resonance, such as to send donations to Rome from various countries of the world to rebuild the basilica in
its current form, which also externally is significantly different from the one damaged by the fire (figures 1 and 2).
Eyewitnesses and documents of the time give a clear picture of the dismay caused by the fire. The event led to the 1
reorganisation of the fire brigade service, more precisely of the “Vigili romani Augustali” (Brigante Colonna, 1929;
Fiorillo, 2022), and to the care of making the new building safe, of which the creation of a pioneering automatic fire
detection and extinguishing system, designed at the behest of Pope Pius IX by the Jesuit physicist Father Angelo Sec-
chi (Calzolari, Marsella, 2021).
Stendhal, 1829, t. II, p. 187: «Je visitai Saint-Paul le lendemain de l’incendie. J’y trouvai une beauté sévère et une empreinte de malheur telle
1
que dans les beaux-arts la seule musique del Mozart peut en donner l’idée. Tout retraçait l’horreur et le désordre de ce malheureux événement;
l’église était encombrée de poutres noires fumantes et à demi brûlées; de grand fragments de colonnes fendues de haut en bas menaçaient de
tomber au moindre ébranlement. Les Romains qui remplissaient l’église étaient consternés».
of 3 26
17
Sulla storia della basilica no all’incendio e sulla sua ricostruzione esiste un’ampia bi-
bliograa. Al contrario, i testi volti ad approfondirne la causa, oltre ad essere limitati nel
numero, sono connotati da un approccio unicamente storico mancando, allo stato delle
conoscenze archivistiche e bibliograche, un’indagine svolta secondo i criteri delle investi-
gazioni dell’incendio. Per questo motivo, gli Autori hanno ritenuto opportuno aggiungere
alcuni elementi di riessione basati sui metodi utilizzati per le investigazioni scientiche,
che si giovano della possibilità di attingere alle conoscenze recenti sulla dinamica dell’in-
cendio del legno.
Il metodo ulizzato
Gli Autori hanno utilizzato le informazioni contenute nei documenti dell’epoca e negli
studi di carattere storico per sviluppare un approfondimento sulle cause dell’incendio,
basato sui principi dell’investigazione scientica, applicando l’impostazione individuata
dallo standard NFPA 921- 2021 della National Fire Protection Association. Poiché la man-
canza di reperti impedisce l’osservazione diretta degli eetti dell’incendio sui materiali e le
strutture, il metodo citato è stato applicato approfondendo le sole parti dell’identicazione
dell’esigenza, della denizione del problema, della raccolta ed analisi dei dati, dello svilup-
po e prova delle ipotesi, per poi giungere a formulare le ipotesi nali.
Le fonti documentali da cui sono state tratte le informazioni sull’incendio della basilica
riportano poche notizie sulle sue cause e sulla sua dinamica. Il testo che più degli altri ha
approfondito la questione, a rma di Fabio Sebastianelli (Sebastianelli, 2004), si presenta
come un’analisi storica che dedica solo marginalmente attenzione alle informazioni ne-
cessarie per comprendere quale sia stato l’innesco e come si sia sviluppato l’incendio. I
A sinistra, gura 1 - Veduta della
Basilica di San Paolo fuori le mura.
G. Vasi, 1767.
Sotto, gura 2 - Veduta della Basilica
di San Paolo fuori le mura nella
forma della basilica attuale.
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari

There is an extensive bibliography on the history of the basilica up to the fire and on its reconstruction. On the con-
trary, the texts aimed at investigating the cause, in addition to being limited in number, are characterized by a purely
historical approach, lacking, in the state of archival and bibliographic knowledge, a study carried out according to the
criteria of the fire investigation. For this reason, the Authors have deemed it appropriate to add some elements of re-
flection based on the methods used for scientific investigations, which take advantage of the possibility of drawing on
recent knowledge on the dynamics of wood and vegetation fire in the areas of wood combustion and in the propaga-
tion of vegetation fires.
Figure 2 - View of the current Basilica of San Paolo Fuori le Mura in the form of the current basilica. The position and shape
of the bell tower, the vestibule and the portico in correspondence with the transept mark the main external differences
compared to the Basilica destroyed by the fire.
1.1. The methodology
The authors used the information contained in the documents of the time and historical studies to develop an in-depth
analysis of the causes of the fire, based on the principles of scientific investigation, applying the approach identified by
the National Fire Protection Association standard NFPA 921-2021 (Guide for Fire and Explosion Investigations).
Since the lack of finds prevents the direct observation of the effects of the fire on the materials and structures, the
cited method was applied by deepening only the parts of the identification of the need, the definition of the problem,
the collection and analysis of the data, the development and testing of hypotheses, to then arrive at formulating the
final hypotheses.
The documentary sources from which the information on the fire in the basilica was drawn report little information on
its causes and its dynamics. The text that more than the others has delved into the question, signed by Fabio Sebasti-
anelli (Sebastianelli, 2004), presents itself as a historical analysis that only marginally pays attention to the informa-
tion necessary to understand what was the trigger and how it developed the fire.
The news reports of the time attributed the cause of the fire, with greater certainty first and more covertly in the fol-
lowing weeks, to the inattention of the workers in charge of repairing the roof or, according to some sources, the gut-
ters. The Chronicle of the annexed to the basilica Monastery indicates the cause of the event in a dispute that took
place between the masons and the tinsmiths or “Stagnari” (Cronaca del Monastero of San Paolo, 1823).
In all cases, the description of important aspects is missing: which equipment the workers were using on the con-
struction site area, which intervention was in progress on the evening of July 15th and, giving credit to the official
version, how hot coals have fallen on wooden elements without the workers had realized the damage that could result
of 4 26
17
Sulla storia della basilica no all’incendio e sulla sua ricostruzione esiste un’ampia bi-
bliograa. Al contrario, i testi volti ad approfondirne la causa, oltre ad essere limitati nel
numero, sono connotati da un approccio unicamente storico mancando, allo stato delle
conoscenze archivistiche e bibliograche, un’indagine svolta secondo i criteri delle investi-
gazioni dell’incendio. Per questo motivo, gli Autori hanno ritenuto opportuno aggiungere
alcuni elementi di riessione basati sui metodi utilizzati per le investigazioni scientiche,
che si giovano della possibilità di attingere alle conoscenze recenti sulla dinamica dell’in-
cendio del legno.
Il metodo ulizzato
Gli Autori hanno utilizzato le informazioni contenute nei documenti dell’epoca e negli
studi di carattere storico per sviluppare un approfondimento sulle cause dell’incendio,
basato sui principi dell’investigazione scientica, applicando l’impostazione individuata
dallo standard NFPA 921- 2021 della National Fire Protection Association. Poiché la man-
canza di reperti impedisce l’osservazione diretta degli eetti dell’incendio sui materiali e le
strutture, il metodo citato è stato applicato approfondendo le sole parti dell’identicazione
dell’esigenza, della denizione del problema, della raccolta ed analisi dei dati, dello svilup-
po e prova delle ipotesi, per poi giungere a formulare le ipotesi nali.
Le fonti documentali da cui sono state tratte le informazioni sull’incendio della basilica
riportano poche notizie sulle sue cause e sulla sua dinamica. Il testo che più degli altri ha
approfondito la questione, a rma di Fabio Sebastianelli (Sebastianelli, 2004), si presenta
come un’analisi storica che dedica solo marginalmente attenzione alle informazioni ne-
cessarie per comprendere quale sia stato l’innesco e come si sia sviluppato l’incendio. I
A sinistra, gura 1 - Veduta della
Basilica di San Paolo fuori le mura.
G. Vasi, 1767.
Sotto, gura 2 - Veduta della Basilica
di San Paolo fuori le mura nella
forma della basilica attuale.
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari

from it, are fundamental questions that still remain obscure. The authors felt the need to clarify how the fire began,
comparing what has been obtained from historical sources with the data that research has produced in recent years.
1.2. The investigative problem
In the case of the fire of Saint Paul outside the walls, the investigative problem consists in the systematic organisation
of the information available on the fire, to develop one or more hypotheses to be tested according to the criteria deriv-
ing from the application of the laws of physics and chemistry of fire. To allow for the subsequent phases of the process,
therefore, the information useful for the formulation of the hypotheses that the authors have identified are listed and
organised.
Figure 3 - Luigi Rossini. 1823. View of the Ruin of the Great Basilica of St. Paul Outside the Walls. The image shows the
extent of the damage suffered by the naves and colonnades immediately below.
1.3. The documents
The informations of historical nature were obtained above all from the chronicles published between 16 and 31 July
1823, referencing also to some other handwritten documents connected in various ways to them (see infra List of
documents used). For the interpretation of the combustion processes, the scientific literature on fire behaviour of
materials and wood in particular was examined. The information on the basilica and on the state in which it was found
has been taken, respectively, from the text by Nicola Maria Nicolai (Nicolai, 1815) and from that of Angelo Uggeri
(Uggeri, 1823).
of 5 26
19
La basilica al momento dell’incendio
La basilica danneggiata dall’incendio del 1823 era stata edicata sotto il regno degli impe-
ratori Teodosio I, Graziano e Valentiniano II intorno alla ne del quarto secolo, nel luogo
in cui la tradizione collocava la sepoltura dell’apostolo Paolo e dove l’imperatore Costan-
tino aveva edicato una basilica più piccola (Docci, 2014, p. 30).
Tra le diverse caratteristiche dell’edicio ci si limita a richiamare quelle costruttive e di
ubicazione di interesse per l’indagine. La planimetria, illustrata nella gura 4, presentava
un impianto a croce latina con aula a cinque navate separate da quattro le di colonne.
Le coperture erano realizzate con strutture in legno di cui quelle delle navate erano in
vista, diversamente da quelle del transetto, che presentava un controsotto. Le strutture
della copertura della navata centrale, chiamate dai documenti dell’epoca “incavallature”,
costituivano una realizzazione di particolare interesse dal punto di vista della tecnica co-
struttiva. Le loro dimensioni sono state accuratamente documentate pochi anni prima
dell’incendio dai rilievi eettuati da Nicola Maria Nicolai (Nicolai, 1815).
Figura 3 - Luigi Rossini. 1823. Veduta Della Rovina Della Gran basilica di S. Paolo Fuori Le Mura. L’immagine
mostra l’entità dei danni subiti dalle navate e dai colonnati immediatamente sottostanti.
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari

The issues of the Diario di Roma (a journal published in Rome dealing with local and international issues) of 16 July
1823 and 26 July 1823 provide important information on the state of the basilica and on the damage caused by the
fire. The extent of the damages can also be deduced from the huge number of images drawn in the weeks immediately
following the event, such as that of figure 3, created by painters and engravers on whom the importance of the build-
ing and the seriousness of the event had exerted a strong attraction.
In the following sections, all references marked with the inverted commas « » have been translated from Italian by the
authors.
Figure 4 - Plan of the basilica (Nicolai, 1815).
2. Information analysis
The following sections present the data and information used to understand the events, divided between the condi-
tions of the building, the environmental conditions, and the testimonies cited from archival sources.
2.1. The basilica at the time of the fire
The basilica damaged by the fire of 1823 had been built under the reign of the emperors Theodosius I, Gratian and
Valentinian II around the end of the fourth century, in the place where tradition placed the burial of the apostle Paul
and where the emperor Constantine had built a smaller basilica (Docci, 2014, p. 30). Among the different character-
istics of the building, construction and location aspects of interest for the investigation will be described.
The plan, illustrated in 4, had a Latin cross plan with a hall with five naves separated by four rows of columns. The
roofs were made with wooden structures of which those of the naves were visible, unlike those of the transept, which
of 6 26
20
Per quanto riguarda il comportamento al fuoco, di cui si analizzeranno in seguito alcuni
aspetti specici per le strutture lignee, tutti i materiali costituenti le colonne, le mura, i
rivestimenti ed i pavimenti erano incombustibili. La quantità di legname utilizzato, inve-
ce, era imponente. Le sole strutture portanti del tetto della navata centrale, che andarono
distrutte nell’incendio, secondo un calcolo degli Autori dovrebbero aver superato il peso
di 350 tonnellate. La quantità di energia termica prodotta dalla combustione di una tale
massa di legname spiega i danni delle colonne in pordo che denivano la navata centrale.
Una ulteriore informazione importante riguarda la ventilazione dell’interno dell’edicio,
che era dotato di 107 nestre (Hartmann Grisar, 1942) di cui Nicolai specica che «quat-
tro sole per parte ne sono aperte essendo le altre murate» (Nicolai, 1815, p. 305). Per com-
pletezza di informazione si aggiunge che nello studio presentato le caratteristiche della
ventilazione interna non sono state utilizzate per trarre speciche considerazioni sull’e-
voluzione dell’incendio in quanto le ipotesi prese in considerazione riguardano inneschi
manifestatisi al di sopra del tetto, in un ambiente non condizionato dal tipo e dal numero
di aperture presenti.
La basilica, attualmente inglobata nel tessuto cittadino, si trova a circa 2 km dalle Mura
aureliane, che all’epoca dell’incendio denivano il perimetro urbano (gura 5). Fino alla
ne del diciannovesimo secolo, la zona in cui essa è edicata era parte dell’Agro romano
in prossimità del Tevere e soggetta ad allagamento, in un’area che nel periodo estivo era
considerata malarica, tanto da essere abbandonata dalla maggior parte della comunità
monastica del convento adiacente alla basilica ogni anno, dal 29 giugno no alla ne di
settembre (Marocchi, 1823, p. 5). L’ubicazione della basilica è rilevante anche per valutare
il tempo occorso per dare l’allarme e consentire l’arrivo dei soccorritori. In mancanza di
altri mezzi, l’allarme poteva essere dato ai Vigili del fuoco, solo attraverso un messo che si
fosse recato alla loro sede posta nella piazza di San Macuto, adiacente alla chiesa di Sant’I-
gnazio di Loyola. Per valutare la velocità con cui ci si poteva spostare lungo il percorso, si
A sinistra, gura 4 - Planimetria della basilica (Nicolai,
1815). Sopra, gura 5 - Il tratto di via Ostiense che collegava
la basilica di San Paolo fuori le mura alla Porta Ostiense in
una stampa del 1557 di Étienne Du Pérac edita da A. Lafrery,
1557 (da Frutaz, 1962).
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari

had a false ceiling. The roof structures of the central nave, called "incavallature" in the documents of the time, had a
particularly interesting aspect under the point of view of the construction technique. Their dimensions were accur-
ately documented a few years before the fire by the surveys carried out by Nicola Maria Nicolai (Nicolai, 1815).
As far as fire behaviour is concerned, some specific aspects of the wooden structures will be analysed below. All the
materials making up the columns, walls, cladding and floors were non-combustible. The amount of wood used, how-
ever, was impressive. The load-bearing structures of the roof of the central nave alone, which were destroyed in the
fire, should have weighed more than 350 tons according to a calculation by the authors. The amount of thermal en-
ergy produced by burning such a mass of timber explains the damage to the porphyry columns that defined the central
nave.
Further important information concerns the ventilation of the inside of the building, which was equipped with 107
windows (Hartmann Grisar, 1942) of which Nicolai specifies that «only four on each side are open as the others are
walled up» (Nicolai, 1815, p. 305). For the sake of completeness of information, it should be added that in the study
presented, the characteristics of the internal ventilation were not used to draw specific considerations on the evolu-
tion of the fire, as the hypotheses taken into consideration concern ignitions which occurred above the roof, that is in
an environment where fire dynamics was not conditioned by the type and number of the windows.
Figure 5 - The stretch of via Ostiense that connected the Basilica of San Paolo Fuori le Mura to the Porta Ostiense in a 1557
print by Étienne Du Pérac published by A. Lafrery, 1557 (from Frutaz, 1962).
2.2. Location of the basilica
The basilica, currently incorporated into the city fabric, is located about 2 km from the Aurelian Walls, which defined
the urban perimeter at the time of the fire (figure 5). Until the end of the nineteenth century, the area in which it was
built was part of the Roman countryside near the river Tiber and was subject to flooding. During summer it was con-
sidered malarial, so much so that it was abandoned by most of the community monastery of the convent adjacent to the
of 7 26
20
Per quanto riguarda il comportamento al fuoco, di cui si analizzeranno in seguito alcuni
aspetti specici per le strutture lignee, tutti i materiali costituenti le colonne, le mura, i
rivestimenti ed i pavimenti erano incombustibili. La quantità di legname utilizzato, inve-
ce, era imponente. Le sole strutture portanti del tetto della navata centrale, che andarono
distrutte nell’incendio, secondo un calcolo degli Autori dovrebbero aver superato il peso
di 350 tonnellate. La quantità di energia termica prodotta dalla combustione di una tale
massa di legname spiega i danni delle colonne in pordo che denivano la navata centrale.
Una ulteriore informazione importante riguarda la ventilazione dell’interno dell’edicio,
che era dotato di 107 nestre (Hartmann Grisar, 1942) di cui Nicolai specica che «quat-
tro sole per parte ne sono aperte essendo le altre murate» (Nicolai, 1815, p. 305). Per com-
pletezza di informazione si aggiunge che nello studio presentato le caratteristiche della
ventilazione interna non sono state utilizzate per trarre speciche considerazioni sull’e-
voluzione dell’incendio in quanto le ipotesi prese in considerazione riguardano inneschi
manifestatisi al di sopra del tetto, in un ambiente non condizionato dal tipo e dal numero
di aperture presenti.
La basilica, attualmente inglobata nel tessuto cittadino, si trova a circa 2 km dalle Mura
aureliane, che all’epoca dell’incendio denivano il perimetro urbano (gura 5). Fino alla
ne del diciannovesimo secolo, la zona in cui essa è edicata era parte dell’Agro romano
in prossimità del Tevere e soggetta ad allagamento, in un’area che nel periodo estivo era
considerata malarica, tanto da essere abbandonata dalla maggior parte della comunità
monastica del convento adiacente alla basilica ogni anno, dal 29 giugno no alla ne di
settembre (Marocchi, 1823, p. 5). L’ubicazione della basilica è rilevante anche per valutare
il tempo occorso per dare l’allarme e consentire l’arrivo dei soccorritori. In mancanza di
altri mezzi, l’allarme poteva essere dato ai Vigili del fuoco, solo attraverso un messo che si
fosse recato alla loro sede posta nella piazza di San Macuto, adiacente alla chiesa di Sant’I-
gnazio di Loyola. Per valutare la velocità con cui ci si poteva spostare lungo il percorso, si
A sinistra, gura 4 - Planimetria della basilica (Nicolai,
1815). Sopra, gura 5 - Il tratto di via Ostiense che collegava
la basilica di San Paolo fuori le mura alla Porta Ostiense in
una stampa del 1557 di Étienne Du Pérac edita da A. Lafrery,
1557 (da Frutaz, 1962).
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari
Saint Paul outside the walls
Aurelian walls
Via Ostiense

basilica every year, from 29th June to the end of September (Marocchi, 1823, p. 5). Therefore, when the fire des-
troyed the Basilica, only a few people inhabited the monastery.
The location of the basilica is also relevant for assessing the time required to raise the alarm and allow the arrival of
rescuers. In the absence of other means, the alarm could be given to the firefighters only through a messenger who
had gone to their headquarters located in the square of San Macuto, adjacent to the church of Sant'Ignazio di Loyola.
To evaluate the speed with which one could move along the route, it can be taken into account that the Via Ostiense
was paved with stones for about two thirds of its extension (Esposito, 2011, p. 299), but further insights can be car-
ried out on this topic. Furthermore, public lighting, created with oil lamps, was present only at the intersections of the
built-up area (Benocci, 1985).
2.3. Location of the basilica
On the state of the roof of the basilica (figure 6), in the text by Angelo Uggeri (Uggeri, 1823, p. 13) there is an inter-
esting description of the state before the fire, which clarifies the aspects of degradation that made the works neces-
sary: «Who does not see that all these innumerable nests formed of straw, fronds of brushwood and other combustible
materials, could, as the fire of some coal that was transported by the wind or by negligence out of its receptacle, ignite
the tinder and let the flame run?". In another passage, Uggeri explicitly writes of the «abandonment of that part of the
building, which nevertheless had to draw the care and attention of man for its conservation».
On the construction and maintenance techniques of the roof, the same text clarifies various aspects: «The ordinary
method of covering buildings with flat tiles and with curved tiles, which the ancients have left us, is the same one that is
adopted by the moderns, with the difference only that they abuse the method adopted by those and this case such an
unreasonable arbitrariness was fatal to the roof of St. Paul. The flat tiles were placed on top of the timbers that were to
support them, as was the custom in this immense roof, the curved tiles were placed on the approach of the straight lines,
and thus the great contignazione was protected from top to bottom from water and damp. Such was the way that the
Greeks and Romans have handed down to us, with the only difference, that they used these tiles of marble and metal
in their temples, while we, out of our poverty, make them of terracotta. It is a commendable custom to place between the
wood and the tile a layer of well-laid bricks, and also plastered so that if some tiles break by accident due to the harsh-
ness of the season or other causes, the waters can find an outlet without harming the wood underneath. But the un-
reasonable arbitrariness often causes a method to be practiced the method to place the flat tiles at various distances and
without order on the flat ones, and on top of them flat lay the curved ones, which is often lets voids and holes that
cause to nest for birds that in numerous families stayed in this vast roof far from the town and in the middle of an al-
most deserted and unhealthy region».
The text clearly denounces the presence of birds' nests and vegetation between the tiles, which in summer may have
been dry and therefore liable to participate in the fire. Furthermore, the construction methods of the roof made it
permeable to water and even inside the basilica the roof structures were used by birds as a nest. Basically, the roof was
in a state of disrepair which made it urgent to carry out the work for reasons of decorum and to safeguard the wooden
structures. As for the materials present in the basilica, the combustible ones were the load-bearing structures of the
roofs. The rest of the structural elements consisted of non-combustible material. In particular, the walls were in brick,
the columns in marble or porphyry, the floors in stone material. No systems of any kind were installed inside the ba-
silica.
2.4. The nature of the maintenance work
The text reported in the previous paragraph describes the need to carry out maintenance works and it specify the part
of the building where the construction site was open and of the type of work carried out. This is extremely important
information for reconstructing the dynamics of the fire. Moreover, the Diario di Roma (Diario di Roma n. 59, p. 3)
of 8 26

refers to a repair of the copper channels of the gutters, placing the building site area on the third "incavallatura" near
the facade, on the right side (i.e., the southern one). Figure 6 shows the section of the roof wooden structure. Gi-
useppe Marocchi (Marocchi, 1823, p. 3), with respect to the Diario of 26 July, adds that a work was in progress also
on a "staffone" that is a metal reinforcement of a truss and the Chronicle of the Monastery of Saint Paul also reports
works on the copper eaves (Cronaca di San Paolo, 1823).
Figure 6 - The wooden structure covering the central and lateral naves is reproduced in an image resulting from an accurate
survey (Nicolai, 1815).
Angelo Uggeri's text is less interested in the location of the works but confirms the causal relationship with the fire.
The only discrepancy with respect to the other two sources, also reported by the Journal des débats politiques et lit-
téraires, concerns the citation of the lead as material used to form the eaves: «On ne varie point sur le cause de cet in-
cendie. On l'attribute à la negligence d'un ouvrier qui travallot sur la toiture de l'édifice à reparer les plombs des
canaux pour l'écoulement des eaux. Cet homme a laissé tomber, sans s'en apercevoir, un charbon allumé d' un réchaud
qui était nécessaire à son ouvrage di lui» .
2
However, despite the quantity of details provided, the above information is not sufficient to reconstruct important
aspects of the works carried out at the time of the fire, as it does not clarify whether the tiles and pantiles have been
removed, leaving the layer of flat elements bricks in place or if this layer has also been removed. Such information,
indeed, would be essential to understand whether the wood of the structures may have been exposed to contact with
burning coals that accidentally fell from the pan used for the work or thrown during an altercation between bricklayers
and tinsmiths to which the Chronicle of the Monastery of San Paolo attributes the cause of the fire.
2.5. The environmental conditions
Some considerations on the environmental conditions are necessary for a better understanding of the facts. In particu-
lar:
We do not vary on the cause of this fire. It is attributed to the negligence of a worker who worked on the roof of the building to repair the leads of
2
the channels for the flow of water. This man dropped, without realizing it, a lit coal from a stove which was necessary for his work.
of 9 26
21
può tenere conto che la via Ostiense era pavimentata per circa due terzi della sua estensio-
ne (Esposito, 2011, p. 299), ma ulteriori approfondimenti possono essere svolti su questo
tema. L’illuminazione pubblica, inoltre, realizzata con lampade ad olio, era presente solo
agli incroci dell’area edicata (Benocci, 1985).
Le condizioni del teo
Sullo stato del tetto della basilica (gura 6), nel testo di Angelo Uggeri (Uggeri, 1823, p. 13)
si trova un’interessante descrizione dello stato ante incendio, che chiarisce bene gli aspetti
di degrado che rendevano necessari i lavori: «Chi non vede che tutti questi innumerevoli
nidi formati di paglie di fronde di sterpi e di altre materie combustibili, poterono all’avvici-
narsi del fuoco di qualche carbone che dal vento o dall’incuria fosse trasportato fuori del suo
ricettacolo accendervi l’esca e far correre la amma». In un altro passaggio l’abate scrive
esplicitamente dell’«abbandono di quella parte dell’edicio, che pur tutta doveva trarre la
cura e l’attenzione dell’uomo per la sua conservazione».
Sulle tecniche costruttive e di manutenzione della copertura lo stesso testo chiarisce diver-
si aspetti: «Il metodo ordinario di coprire gli edicj con tegole piane, e con tegole curve, che
gli antichi ci hanno lasciato, è quello stesso che viene adottato da moderni, colla dierenza
soltanto che questi abusano del metodo da quelli adottato ed in questo caso fu fatale al Tetto
di S. Paolo, un così irragionevole arbitrio. Collocate le tegole piane sopra i legnami che le do-
vevano reggere, come in questo immenso tetto si è usato di fare, venivano collocate le curve
sull’avvicinamento delle rette, e così da cima a fondo era difesa la grande contignazione dalle
acque e dall’umido. Tale era il modo che i Greci ed i Romani ci hanno tramandato, colla sola
diversità, che essi usavano queste tegole di marmo e di metallo ne’ loro templi, mentre noi per
nostra povertà le facciamo di terra cotta. È costume lodevole nelle piccole dimensioni de’ tetti
il frapporre fra il legno e la tegola uno strato di mattoni bene incalciati, ed anche intonacati
onde al rompersi per accidente qualche tegola per crudezza di stagione od altra causa, le ac-
que possano trovare uno sfogo senza nuocere ai sottoposti legni. Ma l’irragionevole arbitrio
fa sovente praticare un metodo e cioè di porre in varie distanze e senz’ordine le tegole piane
sopra le piane, e sopra di esse piane posarvi le curve, ciò che fa bene spesso de’ vacui e de’ fori
capaci da annidarvi de’ volatili che in numerose famiglie soggiornavano in questo vastissimo
Figura 6 - La struttura lignea
di copertura della navata
centrale e di quelle laterali è
riprodotta in una immagine
derivante da un accurato
rilievo (Nicolai, 1815).
L’incendio di San Paolo fuori le mura:
un’invesgazione scienca molto a posteriori
Stefano Marsella
Monica Calzolari

• Area lighting. At the time, the streets of the town inside the walls were equipped with oil lamps (Benocci, 1985).
The authors, however, have not found documents or images from which it is possible to deduce the presence of
such lamps in the area surrounding the basilica. After midnight the moon, whose disk was half visible in the previ-
ous hours, had already set, since the astronomical data indicate that in Rome on 15 July 1823 the moonset took
place at 10.24 pm and the moonrise took place the following day at 1.16 pm. The lunar phase allowed the view of
49% of the disc. Dawn on July 16 saw the sun rise at 4:49. In essence, between 10.30pm and approximately
4am, it was completely dark in the area.
Figure 7 - Route between the Basilica of San Paolo Fuori le Mura and the headquarters of the "Augusta Roman Vigili” in
piazza San Macuto. The map of Rome is based on a print by Ettling - Day & Son - 1860.
• Wind. The only source which, in addition to the presence of wind during the early stages of the fire specifies its
direction, also giving an interesting indication of its intensity, is the Chronicle of the Monastery of Saint Paul
(Cronaca del Monastero di San Paolo, 1823): «The westerly wind that blew carried the flame to the central cross
vault, and to the other two naves towards Rome, and its strength was such that in less than three hours all the
beams, and the cross vault no longer existed ...». We’ve already found that Uggeri has taken up this information
when he describes the conditions of the roof. The other sources do not provide useful information on ventilation,
nor can be found in the archives the only meteorological data recorded at the time in Rome, i.e. those of the Me-
teorological Station of the Collegio Romano. To make up for this lack, reference can be made, by way of example,
to the values recorded in recent times, according to which the probability that in Rome in the month of July the
wind blows with a speed that falls within the interval of 1-12 km /h (0.28 - 33.87 m/s) are around eighty percent.
of 10 26
Mappa di Roma Ettling - Day & Son - 1860
1 km
Via Ostiense
P.za San Macuto

For the air temperature, recent surveys indicate that in July it varies between 19° and 32° C degrees . The only
3
source which, in addition to the presence of wind during the early stages of the fire specifies its direction, also
giving an interesting indication of its intensity, is the above mentioned Chronicle of the Monastery of Saint Paul
(Cronaca del Monastero di San Paolo, 1823).
• Fire alarm and rescue operation times. The description of the environmental conditions must also include a brief
reference to the organization of the technical rescue services in 1823 and in particular to two relevant aspects for
estimating the time which may have elapsed between the alarm and the arrival of the firefighters.
Table 1 - The count of the hours on July 15th more interesting for the analysis of the dynamics of the fire according to
the two methods.
The headquarters from which the rescue services departed was located in Piazza San Macuto (figure 7), inside the
urban perimeter defined by the Aurelian Walls, about 5 km from the basilica. Moreover, the fire-7hting services did
not have horses, a circumstance which implied the Fire Brigade to tow the carts by hand when it was not possible to
requisition them on the spot. The horses, as actually happened on the morning of July 16, according to the testimony
of Marocchi (Marocchi, 1823, p. 5): «The alarm was given to their station of S. Ignazio, and they brought with them
three wagons with horses taken from carriages, two of which wagons carried two pumps, and the third was heavily
loaded with various necessary implements». For the purposes of the investigation, the time needed to travel about five
kilometers from the basilica to the site of San Macuto (adjacent to the church of Sant'Ignazio), to organize the rescue
operations and to reach the basilica by the first rescuers, at best should have been more than 150 minutes, consider-
ing an immediate departure and an average speed of 4 km/h.
2.6. The testimonials
Before reading the chronology of the events narrated in the documents examined, some words about the method of
subdividing the hours used at the time in Italy are needed, called "Italico" (Italian). According such method, the day
began at sunset while, in some cities, as in Rome, the day started half hour after sunset (Paltrinieri, 2022, p. 5). This
method is clearly different from the one currently in use, in which the beginning of the day is fixed at midnight (the
instant placed 12 hours from the one that divides the duration of daylight hours into two equal parts, called noon).
Therefore, the starting time of the day and the duration of the hours varied throughout the year, according to the vari-
ation of the moment in which the sun sets.
The calculation of the moment of abandonment of work by the workers can be used as an example of conversion of
hours used in the sources used. The Diario di Roma sets the time for abandoning the construction site at 11.00 pm
according to the Italian method. To convert this value into the current count, one must start from the fact that on 15
July the sun sets in Rome at 20:43 according to the current method. To define the hours of 24:00 of the Italian meth-
https://www.meteoblue.com/it/tempo/historyclimate/climatemodelled/roma_italia_3169070, accesso 13.03.2023, https://it.windfinder.
3-
com/windstatistics/roma_urbe accesso 13.03.2023, https://it.weatherspark. com/y/71779/Condizioni-meteorologiche-medie-a-Roma-
Italia-tutto-l’anno
of 11 26
Italian
count
23.00
24:00
01:00
02:00
03:00
04:00
05:00
06:00
07:00
8:00
Current
count
20:13
21:13
22:13
23:13
00:13
01:13
02:13
3:13
4:13
5:13

od of that day we must add 30 minutes, which bring the start of the new day to 21:13 of the current method. On July
15th, therefore 11 pm of the Italian method therefore correspond to 20:13 of the current count.
In the days immediately following the fire, the chronicles describe the sequence of events indicating the time at which
they allegedly occurred:
•8:13 pm (before) - The workers leave the area: «the “Stagnari” took the road to Rome, before 11 pm» (Postilla
4
anonima, 1823, c. 1r; Diario di Roma, n. 59).
• 8:13 pm (after) - Access of three "ultramontani" on the “incavallature” authorized by the custodian to finish the
5
building measurements started a few days earlier: «At 23:00, once the tinsmiths had left, three people from bey-
ond the mountains (“ultramontane”) went to beg the custodian to give them access to the comfortable mounts to
take some dimensions to complement the others that had been carefully provided for in recent days » (Postilla an-
onima, 1823).
•9:13 pm (after) - First testimony that a combustion with flame was not in progress: «The Sacristan D. Isidoro
Ferri, who was responsible for the guard on the aforementioned day, went as usual to recite the Hail Mary after
midnight, and to do this he had to cross the entire large nave» (relazione, 1823, fol. 1v; Diario di Roma, n. 59, p.
4).
• 10:13 pm (approximately) - Second, third and fourth testimony that there was no combustion with a flame in
progress: «About the hour of night the merchant Mr. Laici returned from the Tor di Valle estate, and passed under
S. Paolo and he also asserts that he was not aware of any fire on the roof of the basilica» (Relazione manoscritta,
1823, c. 1v; Diario di Roma, n. 59, p. 4). «Nor is it enough. After the hour of night, the under-Curate D. Martino
Testa returned to the Monastery from the Ponticello road, and with him the Cleric Mariano Cotognoli in a group
of other people who were taking themselves into Rome: and no one likewise noticed the slightest sign of fire above
the basilica: and it certainly had to be seen by them since they were coming towards that part where the fire
began» (Relazione manoscritta, 1823, c. 1v; Diario di Roma, n. 59, p. 4). «Finally, the same applies to the
gardener of the Pietro Battisti Monastery, who watered the garden in the evening with his assistant, and finished
the work around 1:00 am.» (Relazione manoscritta, 1823, c. 2r; Diario di Roma, No. 59, p. 5).
• 11:13 pm (approximately) - Detection of the fire according to the Chronicle of the monastery of Saint Paul:
«about two o'clock in the night, a farmer, who was sleeping in the nearby meadow, saw a fire start on the side of
the garden (...) immediately ran to bring notice to the priests and clerics».
•01:30 am (approximately) - Detection of the fire according to the Relazione manoscritta «Only at four and a
quarter of the night a “buttaro” belonging to the merchant Giuseppe Perna, who kept cattle in the large meadow
under the walls of the Monastery, saw fire on the large roof of the basilica» (Relazione manoscritta, 1823, c. 2r)
and Marocchi «On the evening of the 15th of July, about four and a quarter of the night began to manifest a small
flame which, seeing a “Butaro” of the Merchant Perna» (Marocchi, 1823, p. 4).
•03:30 am - Arrival of the firefighters according to the Rome Diary: «In fact, at 6.30 am, six firefighters arrived in
S. Paolo with a pump» (Diario di Roma, n. 59, p. 7).
• 04:00 am approximately - Detection of the fire according to Francesco Fortunati. The reference (Fortunati,
1823) differs significantly from the previous one: «On July 16, 1823. Recurring the Feast of Maria
S(antississi)ma Carmelitana, on the same night around seven o'clock, in the Church of St. Paul outside the walls,
The term “Stagnari” is translated in the paper with tinsmiths
4
The term “ultramontani” is no more used in the current Italian language and means “people from beyond the mountains”. At the time of
5
the document the terms was applied to French people.
of 12 26

above the ceiling, or perhaps in the large vault in the middle, which was all made of wood (...) a huge fire broke
out, and within 8 hours it reduced all that huge building to ashes of walls, a very large construction of the Roof,
and almost of all those columns (...)».
The damage caused by the fire led to the destruction of the central nave, two lateral naves, some load-bearing walls and
numerous columns. Furthermore, frescoes and various works of art dating back to the time of construction of the ba-
silica were damaged. The impression aroused by the event was enormous, as evidenced, for example, by a writing con-
served in the Vatican Secret Archive (Piazza, 1823): «However, took place the fire of the aforesaid basilica, a sac-
rosanct asylum always protected by the Lord for the long space of so many centuries, and in the midst of the greatest
dangers, a sign that he has never [...] faced earthquakes, lightning, floods or other fires, and not even the fury of the
Barbarians similar devastation, indeed He wanted, that it be respected, and preserved by Alaric himself from the
cruelty of the Goths, and by Genserico from the barbarism of the Vandals».
A detailed list of the damage suffered by the basilica is reported in the Diario di Roma (Diario di Roma, 59 pp.
15-17), which also recalls the rich production of engravings and paintings that the ruins stimulated, but also all the
other chronicles agree in stating that the damage caused by the fire affected a large part of the basilica, starting from
the roofing of the central nave, the left side nave, the transept, including the granite and marble columns.
Table 2 - The sequence of the events reported listed according the Italian and the current count of hours.
2.7. The causes of the fire according to the chronicles
Almost all sources agree in attributing the fire to the recklessness of the workers engaged in the execution of works on
the southern side of the main nave of the basilica. In particular, the Diario di Roma of 26 July on page 3: «Due to a
very fatal misfortune resulting, as it appears, from the fact of some tinsmiths, who yesterday, while carrying out vari-
ous works on the roof of the Basilica of San Paolo Fuori le Mura, dropped some lit coals from a brazier, the fire was set
last night on the ceiling of the said basilica..». This position is confirmed by the Journal des débats politiques et lit-
téraires of 31 July 1823 and the text by Angelo Uggeri (Uggeri, 1823, p. 12), whose architectural skills are recalled
and, due to his knowledge of the basilica, in 1825, was appointed secretary of the special Congregation for the con-
struction of the Basilica of San Paolo destroyed by the fire of 1823.
of 13 26
Italian
count
Current
count
Event
reference
23:00
8:13 pm
the workers leave from the site
Relazione manoscritta, 1823, c. 1r
Diario di Roma, n. 59, p.3
23:00
8:13 pm
access of three “ultramontani"
Postilla anonima
24:00
9:13 pm
first testimony that, inside the basilica, a flame
combustion was not taking place
Relazione manoscritta, 1823, c. 1r Diario di
Roma, n. 59, p.4
01:00
10:13 pm
second, third and fourth testimony that there was no
flame combustion taking place
Relazione manoscritta, 1823, c. 1v
Diario di Roma, n. 59, pp. 4 e 5
02:00
11:13 pm
fire detection
Cronaca del Monastero di San Paolo
04:30
01:30 am
fire detection from outside, on the roof
Relazione manoscritta, 1823, c. 2r
Diario di Roma, n. 59, p. 5
Marocchi, 1823, p. 4
Postilla anonima, 1823, p. 3
06:30
03:30 am
arrival of the firefighters
Diario di Roma, n. 59, p. 7
Postilla anonima, 1823, p. 3
07:00
04:13 am
fire detection
Fortunati, 1828, p.3

The Chronicle of the monastery of Saint Paul attributes the cause of the fire to the «altercation that the bricklayers
and the “Stagnari” the latter threw the frying pan full of fire at the former», while the text by Giuseppe Marocchi (Ma
6-
rocchi, 1823, p. 4), in describing the context of the works, seems to distance itself from attributing the cause to the
workers as it explains that: «and we don't believe we are going to go into the cause of the fire, since the very vigilant
Government has tried and is trying to investigate it».
The anonymous annotation, written in the margin of a volume conserved in the National Library of Rome adds another hy-
pothesis. In volume n. 68 of the Thesaurus Romanus, a repertoire created by the scholar Costantino Maes, a copy of the
Diario di Roma n. 59, in correspondence with the article concerning San Paolo, presents a synchronous handwritten mar-
ginal note (Postilla anonima, 1823) which introduces important details. In this text, different statements: can be found
•Confirmation of the moment of detection and of the place of origin of the fire: «At 4 1/4 o'clock Battisti [the guardian]
saw the head of the beam where the tinsmiths had worked burning ...»
•Indication of the time needed for the fire to destroy the basilica: «So in two and a quarter hours... in three hours some
spark, some lit coal, which the tinsmiths are supposed to have neglected to extinguish, or were careless if they had it car-
ried by the wind on a beam, have destroyed, and an immense Temple was consumed, built so solidly that it resisted the
ravages of time for several centuries!»
•Introduction of the hypothesis of arson: «At 11 pm, once the tinsmiths had left, three people from beyond the mountains
(“ultramontani”) went to beg the custodian to give them access to the structures to take some measures to complement the
others which they had carefully taken in the previous days. This is a premise. Those who began to ring the bells clearly
saw that the fire was coming from four points in a square… . If this is true, the tinsmiths cannot be blamed for the ruin that
occurred shortly afterwards on the entire immense ceiling; but from the prompt release of them from the prisons immedi-
ately by the Police, from the progress that the diabolical sect of "carbonari incendiari” is currently making and more from
the express prohibition that subsequently came to all custodians from allowing anyone access to the dome of St. Peter's,
and in the other conspicuous monuments of the City, without the written permission of the Secretariat of State accompan-
ied by two trusted persons[,] it is easy to recognise the true cause of such irreparable ruin».
3. The hypothesis
In the immediate aftermath of the fire, all the chronicles agree on the fact that maintenance work was underway on the
roof, which was the only combustible structure of the building. Therefore, following the information provided by the
chronicles, the main and secondary wooden framework of the roof is the construction element from which the authors
have investigated the dynamics of the early stages of the fire.
In general, the main causes of accidental ignition in a building where, like the basilica of 1823, there are were no sys-
tems fueled by electricity or by flammable or combustible materials, can be traced back to lightning or to the improper
use of open flame lighting devices (torches, lamps, candles). No documents of the time report meteorological per-
turbations, while according to the anonymous “Postilla”, three people would have gone to the roof structure to carry
out some surveys, an activity that after sunset would have needed the use of torches.
Giuseppe Marocchi (1797-1865), born in Massa Lombarda and lived in Rome, but defined as Imola, blacksmith turner, and also author of
6
numerous works including a Chiaramonteide, in honor of Pope Pius VII. So, his testimony is worthy of the utmost consideration as Maroc-
chi was very close to Secretary of State Consalvi and therefore also informed of the facts even if required to respect the will of the protector
of him. After the death of Pius VII he lost his place at the papal court because he belonged to the party opposed to that of the zealous who
had elected Pope Leo XII (from: Antonio Castronuovo, Giuseppe Marocchi, alias Morocco, poet and turner, in «Pagine di vita e storia
imolesi», 15, 2015, pp. 79-88).
of 14 26

Therefore, taking into account the increased risk of fire that maintenance work entails , two hypotheses can be formu
7-
lated: accidental fire caused by maintenance work (on or under the roof) or intentional fire.
Figure 8 - Hypothesis of roof structure drawn by the authors on the basis of Uggeri's text (Uggeri, 1823, p. 13) and Nicolai's
drawings (Nicolai, 1815).
3.1. The first hypothesis: construction work fire
About maintenance work on the roof, Giuseppe Marocchi (Marocchi, 1823, p. 3) mentions those on the metal brack-
et of a beam and on the eaves, the Diario di Roma of 16 July refers to the problem in a more generic way, while the
Diario of 26 July on page 3 identifies the location in a precise point: «there two tinsmiths worked to place the copper
channels on the gutters of the roof of the great nave, located to the west, and precisely in the third beam which looks
towards the garden of the Monastery next to the facade of the basilica». The hypothesis that connects the cause of the
fire to the works carried out on the roof of the basilica (whose structure is hypothesised by the authors as represented
in figure 8) is referred to in various sources: the Chronicle of the Monastery of Saint Paul, the Diario di Roma n. 56
and no. 59, which explicitly refer to the presence of braziers necessary for the maintenance of the gutters of the cent-
ral nave or for the repair of lead elements.
Angelo Uggeri accurately describes the need to carry out the work due to the deterioration of the roof (Uggeri, 1823,
p. 13). The scenario described leads us to hypothesise that work was taking place above the wooden load-bearing
structures. In particular, accidentally fallen burning coal that ignite a beam needs to deposit it, so it must fall from
above, a circumstance which presupposes a work carried out after the removal of the brick elements when the beams
were exposed to contact with incandescent elements. In this case, a possible coal accumulation point could be the gap
“Buildings under construction or renovation are at their most vulnerable and weakest condition. Accumulation of waste combustibles, limited
7
access, minimal water supply and hazardous operations increase the challenge. Add to this the effects of firefighting operations, increased
water weight, weakened metal and support structures, and hidden hot spots, and you have a formula for disaster waiting
to!happen" (https://www.usfa.fema.gov/blog/cb-031919.html last accessed 03.06.2023)
of 15 26

between the perimeter walls and the beam heads. It must be added that it was necessary to heat the metal also for the
work on the copper of the eaves, with the use of incandescent charcoal braziers.
The fire spreading above the roof should also be taken into consideration due to the presence of dry shrubs and nests,
which probably had to be eliminated during the same maintenance work. According to Uggeri's testimony, the birds
had also nested on the ceiling inside the basilica: «When it was not difficult for me to climb up to that height, and to
walk the path practiced there from top to bottom on the double ropes of the scaffolding, I have often pointed out to
some of my assistants the numerous families of the doves, which all that region they filled with their lamentable voices,
as well as with other minor and major birds. These and those with their rapid flights very often distracted the devotees
from the concentration due to the holy place...» (Uggeri, 1823, p. 14). In this regard, coals that fell near a nest or
some dry vegetable element could have triggered a fire on the roof and towards the underlying structures, but the fire
would have been noticed by the workers or by the various witnesses mentioned in the Diario di Roma n. 59, as the
exposure to the wind would have immediately started a combustion.
A variant of the first scenario shifts the focus on the work in progress under the roof covering on the monumental
trusses, in particular for the maintenance of the metal brackets used to bind the wooden elements to strengthen the
horizontal element of a truss which «could soon threaten to ruin» (Marocchi, 1823, p. 4). Also in this case, the work
would have justified the use of braziers.
A description of this hypothesis could be the following: one or more workers are at work under the cover of the central
nave. To carry out the work, a brazier containing burning coals is used. The works were suspended shortly before
sunset and the construction material was left in place, without any particular precautions. For some reason, some
coals from the brazier fall on the roof frame. The contact of the wood with the coal causes the ignition of the fire,
which spreads to the rest of the roof.
3.2. The second hypothesis: arson
Arson hypothesis is based essentially on the the application of the data examined about smouldering combustion and
on the “synchronous handwritten marginal note” preserved by Costantino Maes Postilla anonima (1823). According
to the data examined, the time interval between the accidental ignition which occurred at the end of the works and the
moment in which the fire was detected is considerably greater than what the scientific literature found by the authors
indicates as the necessary value for the transition from smouldering combustion to flame combustion. A second aspect
is linked to the testimony of the two priests who climbed «at great risk to their lives (...) on the bell tower to ring the
bells, to ask for help» (Diario di Roma, 59, p. 6) and who saw the fire spread from four different points on the roof.
We can also mention the obligation, imposed after the fire, to visit the dome of St. Peter's only with specific authoriz-
ation and with the necessary companion.
A further source to be mentioned about the hypothesis of a malicious cause can be found in the note dated 23 July
1823 by Father Piazza (Piazza, 1823). This document, dated seven days after the fire, refers both to ongoing invest-
igations and to the recognised innocence of the "Stagnari": «At such a time, therefore, and in such circumstances,
having read the Report presented to me for the press, and observing that no mention was made of the cause of such a
disastrous event, other than the negligence of the Stagnari (saying however that they were declared innocent) or the
work of Other persons, who are under Government investigation» .
4. Test of the hypotheses
To test the hypotheses defined in the previous section and, in particular, to evaluate the correspondence of the times
narrated by the chronicles in the light of the times required by the wood ignition process, it is necessary to recall some
information on the wood ignition mechanisms. It should be noted that the literature data found do not concern the
of 16 26

specific case of ignition by incandescent coals fallen on the wood but, in the cases closest to that of the study, they deal
with the behaviour of incandescent embers (of wood or coal) of variable dimensions while burning on a bed of wood.
When incandescent elements are deposited on a combustible surface, combustion can start when the heat transfer is
sufficient to bring the temperature to the value necessary to trigger an oxidation reaction. In order to create condi-
tions for the propagation of combustion, the determining factors are the wind speed, the mass of the incandescent
elements and the density of the combustion (Salehizadeh, 2021 pg. 2). Wind speed, in particular, appears to be the
most relevant factor in this process, as it guarantees the influx of oxygen in quantities suitable for the area in which
combustion takes place.
As the speed of the air flow increases, it has been shown that the oxidation rate increases, resulting in temperatures
and higher heat fluxes on the test surface (Manzello, 2009). In particular, burning embers have the ability to trigger
combustion on combustible construction products when they enter the cracks. Of these processes, the correlation
between the mass of material and the time necessary to reach the peak combustion temperature has also been demon-
strated (Salehizadeh, 2021, p. 28). Once the flame combustion phase has started, recent research has shown that as
the mass increases, the maximum temperature reached also increases, with average temperatures above 700 °C, up to
values of 900 °C with higher wind speeds ( Salehizadeh, 2019, p. 35).
The third element that has a decisive influence on the ignition of the fuel is its density. Whether it is material pro-
duced by burning trees and deposited by the wind or accidental fall of coals, the method of accumulation has chaotic
aspects, which can be studied with theoretical models but which finds in experimentation an easiest way to understand
the underlying mechanisms. A series of experiments with pine twigs with a diameter of 6–8 mm and a length of 40 ± 2
mm in the combustion phase deposited on a wooden bed, allowed to verify the minimum quantity of accumulation
capable of generating the combustion of the bed (Matvienko, 2022, p. 10). Such experiments, in which various para-
meters such as their quantity, wind speed and type of wood material were tested, showed that regardless of the amount
of pine twigs used in the study, the ignition of wood materials does not it has been observed in the wind speed range of
0–1 m/s. Small increments of airflow in the contact patch did not result in a transition to a flaming phase, but smoul-
dering of the embers continued until they were burnt out. The cited tests show that the ignition time decreased with
increasing wind speed and with an increase in the number of embers.
The consulted literature does not give the certainty on the actual age of the wood of the roofing placed to cover the
main nave of the basilica. In the Journal des débats politiques et littéraires the structures destroyed by the fire, the ori-
ginal ones (therefore from the fifth century AD) were considered. Between the eighth and nineteenth centuries, how-
ever, the time interval is such as to make it doubtful that the same structures have not been replaced in whole or in part
(Uggeri, 1823, p.11). The basilica is located in an area which at the time was subject to flooding from the Tiber river
(Uggeri, 1823, p. 9) and, as the author himself refers, they were erected without particular care in protecting them
from the infiltration of water rain and aggression of biological agents. In any case, several difficulties are encountered
in defining the effect of aging on the fire behavior of the basilica structures.
The literature data found by the authors about the effect of aging on fire behaviour does not go beyond 400 years and,
more generally, the variety of physical and chemical processes occurring in wood during its service life does not allow
for a highly accurate artificial reproduction of all the transformations occurring in it (Aseeva, 2014, p. 239). Fur-
thermore, given the environmental conditions, it is not easy to define the variations that occurred during the period of
service of the wooden structures of the basilica, nor have documents been found regarding laboratory tests on speci-
mens of the specific essence (Cedar of Lebanon - Cedrus (figure 9) subject to ageing.
of 17 26

Figure 9 - The image (Zachar, 2021), illustrates the results of the absolute burning rate of the oak wood for different aging
(from 0 to 120 years).
The experiments cited seem to indicate that, as the useful life of wooden structures increases, the carbon content de-
creases, while three the hydrogen and oxygen contents increase proportionally. Furthermore, both from calculations
and from experimentation, the lower value of energy produced by combustion complete shows an increase over the
useful life of the timber. By determining the flammability parameters of wood samples, the cited text highlights that
the ignition delay time and the value of the critical heat flux of ignition of the sample increase as the density of the
wood increases. More generally, keep taking into account the uncertainties on the applicability of this study to a very
different context, it would seem that wood with a long service life modifies the behavior fire in a non-linear way and
therefore difficult to consider for the purposes of the study.
Still in relation to the age of the wood, understood as the period between the cutting of the tree and the moment in
which the analyzes were carried out, data on the minimum energy primer have been found on essences other than the
Cedar of Lebanon. For example, oak wood has shown that in 120 years this energy decreases by 8.8% compared to
that initial (Aseeva, 2014, p. 246). The experiments cited seem to indicate that, as the useful life of wooden struc-
tures increases, the carbon content decreases, while three the hydrogen and oxygen contents increase proportionally.
Furthermore, both from calculations and from experimentation, the lower value of energy produced by combustion
complete shows an increase over the useful life of the timber.
4.1. Notes on the wood smouldering
Before explaining the hypotheses that the authors consider most justified regarding the causes of the fire, some basic
aspects of the wood combustion process of smouldering are summarized in this paragraph.
Smouldering, according Collins Dictionary is “to burn slowly without flame, usually emitting smoke”. This phase of
8
the combustion process, linked to the presence of reactions in the solid phase only, in relation to the environmental
https://www.collinsdictionary.com/dictionary/english/smoulder last accessed 28/08/2023
8
of 18 26

conditions and the configuration of the materials involved can be followed by the one with flames, known as pyrolysis,
which is the mechanism that allows the development of flammable gases. Pyrolysis will take place when the combust-
ible material subject to ignition has reached a certain temperature and gases will allow the flames to manifest them-
selves and increase the heated area, favouring the progression of combustion.
Figure 10 - Relationship between wind speed and heat flow (redrawn from: Salehizadeh, 2019, Figure 4.1). Recorded flows
are shown with 16 g of smoking birch embers deposited above the sensor under different wind speeds. The shaded area
represents the standard deviation between tests. The wind speed is 0.5 m/s in the top left graph, 0.8 m/s in the top right
graph, 1.2 m/s in the bottom left graph and 1.4 in the bottom right.
Research has shown that even incandescent elements with a lower calorific value than carbon can give rise to combus-
tion with smoke or flames, highlighting that the incubating combustion is a relatively slow, low-temperature, flameless
exothermic reaction in which oxidation occurs on the surface of the fuel in the condensed phase (Nazar, 2021, p. 9).
In the case of flame combustion, on the other hand, the reaction takes place above the material that generates the fuel,
in an area where mixing with the air is more complete. Therefore, smouldering requires a solid fuel porous (or a per-
meable aggregate, as in the case of carbon fallen on the wood) which allows the transport of oxygen up to the fuel bed,
favouring its surface oxidation. This process is self-sustaining, when the generation of heat within the incandescent
mass above exceeds the heat losses to the surrounding environment.
of 19 26
500
1000
1500
2000
2500
3000
3500
0
4000
0
10
20
30
40
50
60
Heat flux (kW/m2)
Time (s)
500
1000
1500
2000
2500
3000
3500
0
4000
0
10
20
30
40
50
60
Time (s)
Heat flux (kW/m2)
Heat flux (kW/m2)
500
1000
1500
2000
2500
3000
3500
0
4000
0
10
20
30
40
50
60
Heat flux (kW/m2)
Time (s)
500
1000
1500
2000
2500
3000
3500
0
4000
0
10
20
30
40
50
60
Time (s)

When these conditions are met, smouldering propagates without external heat input. For the sake of completeness of
the picture, it must be kept in mind that, for very prolonged periods of exposure, in some woody species this mechan-
ism can be started from the temperature of 160°C (White, 2001).
Since the smouldering reaction rate depends on the heat transfer and the availability of oxygen, it must however be
taken into account that the permeability of the medium can have opposite effects, as the effect of higher reaction
speed generated by a higher air flow is counteracted by the effect of volume cooling. Other relevant factors of smoul-
dering combustion are the moisture content of the fuel, the increase of which corresponds to a greater dissipation of
the heat, the size and configuration of the fuel, which change the propensity for internal heating and heat storage, and
environmental conditions such as humidity relative, temperature and exposure to external heating (Nazar, 2021, p.
16).
Applying the results of the studies mentioned to the case in question, it should be taken into account that the role of
embers in the transition from the smouldering combustion process to the flaming process is not well understood, but
appears to be influenced by ember size and composition, wind speed, moisture content, spatial arrangement, and fuel
bed continuity. Therefore, with the same contact area, the probability of transition from smouldering combustion to
flaming combustion for an accumulation of incandescent embers depends on the wind speed (figure 10). In particu-
lar, a higher wind speed is required for a smaller stack, because more oxygen is needed to produce the amount of heat
to initiate the transition from smouldering to flaming combustion. One of the most important factors for the investiga-
tion is the evaluation of the time required for the transition from smouldering to flaming combustion. In this respect,
as mentioned above, it is crucial the role of air flow, which has two competing effects on the smouldering combustion
of the fuel substrate. Increased oxygen supply can increase the rate of coal oxidation and pyrolysis, resulting in higher
heating and higher pyrolysis production until when the flame is lit. However, the cooling effect on surface temperatures
due to convective motion and the dilution of flammable gas mixtures due to airflow can delay flame ignition: in a wind speed
range of 0.5 m/s to 2.5 m/s (Salehizadeh, 2019 p. 14) the increase in oxidation rates in solid fuel substrates is consider-
ably greater than the cooling effect.
4.2. Comments on the first hypothesis
In the documents examined the authors have found no reference to the working techniques used in Saint Paul in the
days before the fire. The hot coals could have generated the fire only if they fell on the wooden structures, and this
hypothesis presupposes access to the upper part of the roofs, with the connected condition of having removed tiles
and the brick mantle and perhaps part of the secondary roof structure. As regards the risk posed by the numerous
nests and by the dry plants grown between the tiles in the presence of work on the roof, it must be highlighted that in
case of a fire involving such materials, the flames would have been noticed immediately and that they would have
spread quickly, not after some hours as the sources clarify. Two more areas on uncertainty concern the lack of re-
searches on the fire propagation of combustible material placed between roof tiles and the fact that, despite the ex-
posure of the wood structural elements to humidity due to inadequate maintenance of the roof covering, it is not clear
whether the load-bearing elements were those dating back to the 7th century or they have been completely replaced or
partly as a result of subsequent maintenance work (Uggeri, 1823, p. 11). Given that only the Journal des débats poli-
tiques et littéraires attributes explicitly about fifteen hundred years to the crossbars, the question of the service life of
the wooden load-bearing elements is worthy of further study because, normally, the degradation of the wood due to
the action of humidity makes it necessary to periodically replace the load-bearing elements (Uggeri, 1823, p. 11).
A pivotal issue of the available information is related to the time the fire started. According to the sequence of events
reported in the chronicles, the evening suspension of work took place at 20:13. No one noticed smoke or flames until
about 11.15 pm, while by 10.00 pm (or midnight, depending on the source) the fire was clearly visible. Therefore,
of 20 26

after 8:13 pm, if there was smouldering, the smoke was produced in such quantities that it could not be perceived,
either by smell or by sight.
It should be noted that, although the first detection of the fire took place within two hours after the daily end of the
works (Chronicle of the monastery of Saint Paul) or four hours (Diary of Rome n. 56 and n. 59), this does not imply
that combustion with flames has started at that time. In fact, it is not specified in any document whether the “buttero”
(cowboy) was awake all night or if the fire was well advanced at the time he noticed it. In this regard it must be con-
sidered that at the time no public lighting system was operating near the building and that the moon had not risen on
the night of July 15th. Total darkness, therefore, would have facilitated the view of the flames or their glare by the
“buttero” or, before him, by the greengrocer, who was near the work area. Conversely, smoke in the absence of a
flame could be seen more easily before dark, then by about 8:30 pm but, inside the building, the smoke should have
been felt by the sacristan (who had crossed the aisle at 9:15 pm), or also by the greengrocer and his assistant who had
been working near the building until about 10.15 pm.
The hypotheses exposed (accidental fall of incandescent coals from the brazier used for machining on the truss, con-
sequent accumulation on the wood of the structures) are not demonstrable due to the lack of information on the type
of construction site in progress. Moreover, the hypothesis of a smouldering fire caused by the fall of incandescent
coals is linked to the simultaneous occurrence of some specific conditions (adequate but not excessive ventilation,
consistent accumulation of embers). In the case of the Saint Paul fire, this effect makes it probable that a pile of incan-
descent coal gave rise to combustion that smouldered for several tens of minutes, until it manifested itself with flames
and gives rise to the fire.
Still referring to the same hypothesis, it must be considered that the time interval between the initiation and manifest-
ation of the flames (as we know when the buttero noticed the flames, but it is not known how long the flame fire had
been going on before the detection). This value, which varies between 60 and 180 minutes, must be compared with
the results of the tests carried out (Manzello, 2012) on the flame start-up time, defined as the interval between the
moment when the first burning element comes into contact with the surface and the moment when a sustained flame
regime has been observed. These tests provide values that can be estimated in a few hundred seconds. Furthermore,
the time for the ignition of the flame increases with the increase in the mass of the embers, while the mass necessary
for ignition decreases significantly with the increase in the speed of the air flow in the range between 6 and 8 m/s
(Nazar, 2021, p. 22). The same text reports the average flame ignition times which are 437 s for low density cedar
(361 kg/m3), 758 s for sequoia (437 kg/m3) and 934 s for Douglas fir (534 kg/m3).
4.3. Comments on the second hypothesis
As regards the historical sources analyzed by the Authors, the hypothesis of malicious cause can be traced in the in-
terpretation of the note dated 23 July 1823 by Father Piazza (Piazza, 1823), referred to above, and in the content of
the anonymous writing preserved by Costantino Maes (Postilla anonima, 1823). Two aspects can support this hypo-
thesis.
The first, the anomalous interval of time elapsed between the abandonment of the work area and the moment in which
the fire was detected, that can be explained in the possibility that the burning coals have set in motion the combustion
mechanism without flame (smouldering) but keeping in mind that the interval reported by the chronicles regarding
the transition from flameless to flame combustion considerably exceeds the time ranges found in the specific literature
for this process.
The second reason that supports this hypothesis is linked to the fact that the chronicles have cited the name and sur-
name of various eyewitnesses (the sacristan D. Isidoro Ferri, the merchant Signor Laici, the under-curate D. Martino
Testa, the cleric Mariano Cotognoli, the greengrocer of the Pietro Battisti monastery, the "buttaro" of the merchant
of 21 26

Giuseppe Perna, Niccola Antoniacci (assistant of the greengrocer, sent to Rome to ask for help), while no source spe-
cifies the name of the "Stagnari", whose number varies from one to two, which the same sources indicate as respons-
ible for the devastation. More over Marocchi two weeks after the fire reports of police investigation still going on (Ma-
rocchi, 1823, p. 4).
5. Conclusions
According to the chronicles of the time, the events that led to the fire on the roof of the basilica would have started
with an accidental trigger around 8:15 pm (current count). Until 10 pm (if not later), the combustion produced no
smoke or flame. These elements do not allow a trigger and propagation mechanism to be identified with sufficient
certainty. The scientific literature on the trigger mechanisms of wood by contact with incandescent elements and on
the smouldering fire seems to indicate the compatibility of the facts described with two different sequences:
• Accidental. An accidental fall of the burning coals used during the works has occurred before the working area
has been abandoned by the worker. It happened in such a way as to determine an adequate accumulation of the
incandescent elements on the wood of the structures starting the oxidation processes which, with an adequate
wind speed, gave rise to smouldering combustion;
• Arson. A trigger due to an intentional act carried out after the daily end of the works, gave rise to a combustion on
one or more parts of the roof that spread to the wooden structures of the roofs.
The lack of testimonies given by the directly involved workers and of information on the type of work in progress and
on the extent of the fire at the time the “buttero” saw it, prevent us from corroborating both hypotheses. The com-
monly accepted cause, which can be described as a smouldering fire, is partially compatible with the delay between the
embers falling and the first evidence of combustion with a flame only if the amount of coal falling on the wood that
could generate the energy necessary to develop the flames has occurred in an adequately ventilated area. Bt, even in
this hypothesis, which does not fully explain the spread from the ignition point to the rest of the roof, the delay
between the time of abandonment of the building site and the time of detection of the fire significantly exceeds what
the data found indicate as an interval for the transition from smouldering to flaming combustion
The alternative scenario to smouldering combustion, introduced by the “Postilla anonima” is an intentional act, put in
place after the workers abandoned the roof (approximately 20:15 on 15 July). This hypothesis is supported by the fact
that chronicles have quoted punctually the names of the witnesses, without a mention of the name of the workers
guilty of such devastation. Another inconsistency of the first hypothesis is he fact that, according to Father Piazza,
Marocchi and the anonymous drafter of the annotation, a few weeks after the event the investigations were still in pro-
gress and that, according to Father Piazza, the innocence of the “Stagnari” had been recognized.
To narrow the field of the hypotheses described, studies on the spread of the fire between the elements of the roof and
further archival research on the investigations carried out at the time could be useful.
6. Aknowledgements
The authors would like to thank Mr. Francesco Cardillo (retired CNR) for finding numerous texts, Dr. Luigi Iafrate
(CREA) for researching the meteorological data of 1823, the Arch. Paolo Mirabelli (CNR-ITC) for finding texts on
wood priming, Dr. Sabrina Palanti (CNR-IVALSA) and Dr. Nicola Macchioni (CNR-IVALSA) for insights into the
effects of aging and humidity on wood, the engineer. Fabio Alaimo Ponziani (CNVVF) and Eng. Alberto Tinaburri
(CNVVF) for insights into the dynamics of combustion and smoldering fire, Prof. Richard Wittman (University of
California) for support in interpreting the historical context, P. Jean Bosco of the Monastery of St. Paul Outside the
Walls for consulting the documents kept in the Library Archives.
of 22 26

7. Chronicles
Diario di Roma 56 (1823)
Annunzio del fatale incendio della basilica di San Paolo, in «Diario di Roma», n. 56, 16 luglio 1823, p. 4.
Cronaca di san Paolo (1823)
Cronaca del Monastero di San Paolo, c. 329-331, 16 luglio 1823 (Roma, Archivio storico dell’Abbazia di San Paolo
fuori le mura).
Relazione manoscritta (1823)
Relazione esatta, e veridica delle circostanze che precedettero il fatale incendio della Basilica di S. Paolo fuori le Mura,
16 - 22 luglio 1823, manoscritto inserto nella Cronaca del Monastero di San Paolo tra c. 331 e c. 332 (Roma, Archi-
vio storico dell’Abbazia di San Paolo fuori le mura).
Journal des débats politiques et littéraires (1823)
Extrait d’une lettre particulière [16 luglio 1823], in «Journal des débats politiques et litéraires», 31 luglio 1823, p.
1.
Piazza (1823)
Copia di una Lettera Scritta dal P. M. Piazza, 23 luglio 1823 (Città del Vaticano, Archivio Segreto Vaticano, Segret-
eria di Stato, Spogli dei Cardinali, Della Somaglia Giulio, b. 2B, fasc. E) “Carte riguardanti la riedificazione della Ba-
silica di S. Paolo in Roma dell’anno 1828”, cc. 1-2).
Diario di Roma 59 (1823)
Relazione esatta e veridica delle circostanze che precedettero il fatale incendio della Basilica di S. Paolo fuori le mura,
con alcune notizie intorno alla sua fondazione e ai danni che ora ha sofferti, in «Diario di Roma», n. 59, 26 luglio
1823, pp. 3-4.
Postilla anonima (1823)
Postilla anonima marginale manoscritta sincrona alle pp. 3-4 della copia del n. 59, del «Diario di Roma», del 26 luglio
1823, conservata nel vol. 68 del Thesaurus Romanum di Costantino Maes (Roma, Biblioteca Nazionale Vittorio
Emanuele).
Marocchi 1823 (1823)
Dettaglio del terribile incendio accaduto il dì 15 luglio 1823 della famosa Basilica di S. Paolo di Roma fuori di porta
Ostiense scritto da Giuseppe Marocchi Imolese, Roma, nella stamperia Ajani, 1823.
Fortunati (1823)
Avvenimenti sotto li Pontificati di Pio VII e Leone XII dell’Anni 1800 al 1828 raccolti dalla bona memoria di
Francesco Fortunati, parte II, c. 727v (Città del Vaticano, Biblioteca Apostolica Vaticana, Lat. 10731).
Uggeri (1823)
Della Basilica di San Paolo sulla via Ostiense. Nota dell’Abbate Angelo Uggeri, Roma, 1823.
8. Bibliography
R. Aseeva, B. Serkov, A. Sivenkov,
Fire Behavior and Fire Protection in Timber Buildings, Springer, Heidelberg, New York, London, 2014
C. Benocci,
L’illuminazione a Roma nell’Ottocento in Storia dell’urbanistica, in Lazio/1,Edizioni Kappa, Roma 1985
C. Birbhushan, T. Yates, A. Lewry
Effect of fire damage on natural stonework, in Construction and Building Materials, vol. 10, 7 (ott. 1996) pp.
539-544.
of 23 26

G. Brigante Colonna
La basilica di San Paolo completata con monumentale quadriportico, Capitolium, VII, (ago. 1929) pp. 385-392.
M. Calzolari, S. Marsella
Secchi’s Science in the Service of the Public, in I. Chinnici,
G. Consolmagno (Editor),
Angelo Secchi and Nineteenth Century Science The Multidisciplinary Contributions of a Pioneer and Innovator,
Springer, Heidelberg, 2020, pp. 261-276.
N. Camerlenghi,
St. Paul’s Outside the Walls. A Roman basilica, from Antiquity to the Modern Era, Cambridge University Press, Cam-
bridge, 2018.
M. Docci,
San Paolo fuori le mura. Dalle origini alla basilica delle origini, Gangemi, Roma, 2006.
D. Esposito
La Campagna romana nel secolo XVI: infrastrutture e insediamenti nel suburbio, in Roma Le trasformazioni urbane
nel Cinquecento, II. Dalla città al territorio, a cura di G. Simoncini, Leo S. Olschki Editore, Firenze, 2011, pp. 289-
310.
A. Fiorillo,
I Vigili di Roma tra l’era napoleonica e il Risorgimento, Quaderno di Storia Pompieristica, 37, (lug. 2022)
A. P. Frutaz
Le piante di Roma, a cura di Amato Pietro Frutaz, 3 voll., Roma 1962.
P. Hartmann Grisar
Roma alla fine del mondo antico secondo le fonti scritte e i monumenti, 2 voll., Desclee & Editori pontifici, Roma,
1943.
N. Kaye and D. Nguyen
Improving Disaster Resilience by Quantifying WUI Community Ember Exposure. Presented at the 2020 NIST Dis-
aster Resilience Symposium, 28-29 July 2020.
S. L. Manzello, S. Park, T. G. Cleary
Investigation on the ability of glowing firebrands deposited within crevices to ignite common building materials, in
Fire Safety Journal, 44 (2009), pp. 894-900.
S.L. Manzello, S. Suzuki, Y. Hayashi
Enabling the study of structure vulnerabilities to ignition from wind driven firebrand showers: A summary of experi-
mental results, in Fire Safety Journal, 54, 2012, pp. 181-196.
O. Matvienko, D. Kasymov, E. Loboda, A. Lutsenko, O. Daneyko
of 24 26

Modeling of WoodSurface Ignition by Wildland Firebrands, in Fire 38, 5, 2022: https://doi.org/10.3390/
fire5020038
S. Nazar, I. Leventson, R. Davis
Ignitibility of Structural Wood Products Exposed to Embers During Wildland Fires: A Review of Literature - NIST
Technical Note 2153 (2021)
N. M. Nicolai
Della basilica di San Paolo, Opera di Nicola Maria Nicolai Romano, nella Stamperia De Romanis, Roma, 1815
G. Paltrinieri
Fine Settecento: ora italiana e ora francese, in Quaderni di gnomonica, 5 (2002)
H. Salehizadeh
Critical ignition conditions of structural conditions of structural materials by cylindrical firebrands Master of Science,
Department of Fire Protection Engineering, University of Maryland, College Park, MD, 2019.
H. Salehizadeh, R.S.P. Hakes, M.J. Gollner
Critical Ignition Conditions of Wood by Cylindrical Firebrands. Front. Mech. Eng., 2021.
F. Sebastianelli
L’incendio della basilica di S. Paolo fuori le mura, in Roma moderna e contemporanea, XII, (2004), 3, pp. 539-566.
Stendhal
Promenades dans Rome, 2 voll., Paris, Delaunay, 1829 [fonte: gallica.bnf.fr /Bibliothèque nationale de France].
P. Testini
Archeologia cristiana: nozioni generali dalle origini alla fine del sec. 6.: propedeutica, topografia cimiteriale, epi-
grafia, edifici di culto, Edipuglia, Bari, 1980.
R.H. White, M.A. Dietenberger
Wood Products: Thermal Degradation and Fire in Encyclopedia of materials: science and technology, Elsevier Sci-
ence Ltd, S.l., 2001, pp. 9712-9716.
M. Zachar, I. Cabalova, D. Kacivova, T. Jurczykova
Effect of natural aging on Oak wood Fire Resistance, in Polymers, 13, (2021), p. 2059
9. List of figures
Figure 1 - View of the Basilica of Saint Paul Outside the Walls before the 1823 fire. G. Vasi, 1767. .........................3
Figure 2 - View of the current Basilica of San Paolo Fuori le Mura in the form of the current basilica. The position and
shape of the bell tower, the vestibule and the portico in correspondence with the transept mark the main external dif-
ferences compared to the Basilica destroyed by the fire. .......................................................................................4
Figure 3 - Luigi Rossini. 1823. View of the Ruin of the Great Basilica of St. Paul Outside the Walls. The image shows
the extent of the damage suffered by the naves and colonnades immediately below. .................................................5
Figure 4 - Plan of the basilica (Nicolai, 1815). ....................................................................................................6
of 25 26

10. List of Tables
Figure 5 - The stretch of via Ostiense that connected the Basilica of San Paolo Fuori le Mura to the Porta Ostiense in a
1557 print by Étienne Du Pérac published by A. Lafrery, 1557 (from Frutaz, 1962). .............................................7
Figure 6 - The wooden structure covering the central and lateral naves is reproduced in an image resulting from an
accurate survey (Nicolai, 1815). ........................................................................................................................9
Figure 7 - Route between the Basilica of San Paolo Fuori le Mura and the headquarters of the "Augusta Roman Vigili”
in piazza San Macuto. The map of Rome is based on a print by Ettling - Day & Son - 1860. ...................................10
Figure 8 - Hypothesis of roof structure drawn by the authors on the basis of Uggeri's text (Uggeri, 1823, p. 13) and
Nicolai's drawings (Nicolai, 1815). .................................................................................................................15
Figure 9 - The image (Zachar, 2021), illustrates the results of the absolute burning rate of the oak wood for different
aging (from 0 to 120 years). ............................................................................................................................18
Figure 10 - Relationship between wind speed and heat flow (redrawn from: Salehizadeh, 2019, Figure 4.1). Record-
ed flows are shown with 16 g of smoking birch embers deposited above the sensor under different wind speeds. The
shaded area represents the standard deviation between tests. The wind speed is 0.5 m/s in the top left graph, 0.8 m/s
in the top right graph, 1.2 m/s in the bottom left graph and 1.4 in the bottom right. .............................................19
Table 1 - The count of the hours on July 15th more interesting for the analysis of the dynamics of the fire according to
the two methods. ............................................................................................................................................11
Table 2 - The sequence of the events reported listed according the Italian and the current count of hours. ..............13
of 26 26