The whole text may be downloaded here.
Referenced in Narratio de Itinere Navali, a chronicle of an anonymous crusader who took part in the first conquest of Silves (1189).
I include here a historical-geographical excursus of some lesser-known places of western Algarve, named in the Narratio and situated in the neighborhood of Silves. The excursus is dedicated to an audience less familiar with local themes. The major place of Silves is omitted, being already object of an extensive bibliography.
The Arab etymology Al-Buhaira (sea lagoon or lake) fits perfectly the old coastal morphology of the place, as shown in Figure 17. The medieval settlement stood in a scarped peninsula, surrounded by water on three sides, once by the sea and twice by the paleo-lagoon which formed a natural moat. The castle defended the top of the rising land neck.
It is a place-name unknown before 1189, omitted from all Arab sources before the Portuguese conquest. It appears in the Narratio as an individualized settlement, relatively to the castles dependent from Silves:
· It is not part of the list of dependent castles.
· It is worthy to suffer a seizure of property after its surrender to the Portuguese (Albafere...cujus opes in Silviam transtulit.), which means the existence of a degree of wealth.
· The inhabitants have autonomy to surrender (and therefore, theoretically, also for having resisted), contrarily to the aforementioned castles, which are considered only as sieges of garrisons, without autonomy from the city.
Albufeira thus emerges in 1189 with the characteristics of a small walled medina, more important than a mere hisn village, similar in urban status to Loulé and Tavira. This may be confirmed by the plan of its fortification in Figure 24.
The geographic situation of its territory in this period - namely its separation from the castle (and territory) of Paderne - reveals itself as a small enclave on the edge of the area of Silves, much smaller than the future "termo" (municipal territory) of Albufeira.
Everything thus points to a late development of the settlement. Smaller and less important than Almohad Tavira, it is also ignored in the Kitab Rujar of Al-Idrisi (which reflects an outlook of regional Islamic chorography c. 1100, about the beginning of Almoravid rule), not even deserving the epithet qaria as did the contemporary village of Tavira. This strongly suggests that it belonged originally to Silves, as another small dependent castle or fortified tower (as it can be inferred from the size and shape of its castle in Figure 24, similar to those of places dependent from Silves).
The subsequent expansion of the territory and place of Albufeira, with the inclusion of Paderne, should be interpreted as the combination of the increased importance of Albufeira with the concurrent decline of Silves. This was accelerated by the conquest of 1189 and especially by the Crusader raids after the Almohad reconquest (namely the raid of 1217). A similar situation may have occurred in Aljezur and Marachique (both omitted in the Narratio, no doubt by their remoteness), which would became autonomous from Silves. Aljezur is also unknown in Arabic sources before the Portuguese conquest. Along with Loulé, Tavira and presumably Albufeira, they are all places that only gained prominence in population and urban and military autonomy after the Portuguese conquest of the Tagus valley, growing by the fixation of immigrant populations, escaped from conquered areas.
Ipses has been identified with Alvor in 1987. The original pre-Roman name would be *Ipsis, Romanized as Ipsa. As far as we know, it has the privilege of being one of the two settlements in the Algarve (along with Kilibe /Cilpes) listed in Geography of Artemidorus of Ephesus, written about BCE 100. Its location, its status as a mint, and the themes of its coins (Hercules/Melqart Venus Marine/Astarte) allow us to define it as a port-sanctuary and a way-station of the Western Atlantic shipping route from Gadir/Gades.
Eight centuries later, is still the only site in the Algarve identified as a limit of the Visigothic Diocese of Ossonoba, in the document known as Divisio Wambae. Medieval archaeology identified the extant castle as originally Islamic, and also the reutilization of the pre-Roman oppidum as a small walled al-baqar, at least in the Almohad period. Our earliest sources on the chorography of Alvor are from the second half of the CE 16.
Its urbanism of Roman origin is remarkable, comprising a Hippodamian settlement (likely, initially a small castra), the pre-Roman oppidum and a small port district, where the primitive temple would stand, under or not far from the actual CE 16 church. The initial Roman occupation (final BCE 2) would be a small naval base. The presence of abundant Roman rural occupation in the ancient fringe of the inner estuary, and the situation of Ipses at the end of the access road to the Monte Sacrum (Monchique) and its thermal sanctified spa, reveal a situation of local capitality. This impression is reinforced by the highly urban, even if small settlement.
The late development of the luxurious villa of Abicada (mid CE 4), then at the apex of a wedge north of the Alvor estuary, suggests the residence of an important personage associated with the status of Ipses and not a mere villa possessor. In Late Antiquity, the identification of Ipsa as the western boundary of the Diocese of Ossonoba (CE 7) should lead to its existence as a paroecia of a vast territory that would include Cape St. Vincent and Monchique. Ipsa would thus possess a paleo-Christian basilica with a certain importance in the place of the usual pagan temple.
The change of name from Ipsa to al-Bur is unique in the Algarve, where all other relatively important towns kept their previous names. The fact is more peculiar when we consider the choice of an extremely common and rural choronym ("well"), not adjustable to a settlement with some Christian importance and pre-existing urban traces from Roman times that had to be still conspicuous. Even more peculiar is the absence of a toponymical designation: it is just "the well", without further qualifiers.
I think the simplest explanation for these anomalies is the reputation and unique character of this "well", so well known by the Arabic speakers that they started to designate the whole place by it. By the end of CE 16 Sarrão describes the shrine of "Senhora da Ajuda" (Our Lady of Good Help), a place of much devotion, with its "well of a very unique water". The plan of Massai - somewhat later - shows the situation of the well and chapel, as well as the beach road to Lagos and the ferry. The old sacred well of miraculous medical and maritime devotion and renowned throughout the Algarve is also referenced by Athayde de Oliveira, who noted its destruction by the tsunami that followed the great earthquake of 1755.
This was certainly the successor of an older Mozarab shrine, likely of pre-Christian origin, and originally located on the barrier island that still partially closes the lagoon. In my opinion the dedication to a "Lady of Good Help" will date from CE 16; it is a later Marian title, subsequent to the expulsion of Jews and Moors (1497) and the extinction of the mixed sanctuaries of Mozarab tradition. The earlier dedication is unknown, but was perhaps to Santa Maria, as in Faro, an acceptable patroness for such a mixed shrine.
The parallel of a sacred source upon an island by the sea, such as in Gades and Ossonoba, is striking. It is quite possible that in Roman times the island was already connected to land in some way and the ferryway from Portus Magnus to Laccobriga by Ipses was already there, a situation maintained until the second half of the CE 18. (The first whole overland main road to Lagos was only built c. 1800 and required a long detour).
The Islamic fortification of Alvor (Figure 24) stands over the infrastructure of the Roman settlement. The al-baqar is a reutilization of the Iron Age oppidum and the castle joins the probable perimeter of the Roman walls. The Roman street system somehow survived through the centuries, undoubtedly due to Alvor stagnation until late CE 19. Its use during Islamic times is unknown.
The site of Lagos has been identified with the port of Al-Zawia, a waystation on the naval itinerary of al-Idrisi between Silves and Sagres. The indicated distances are however far from worthy and there is not any positive reason for this identification. Moreover, there exists a much better alternative: Al-Zawia is described by al-Idrisi as a port and village. The name corresponds transparently to "Zavial", a hamlet and beach located between Lagos and Sagres. The full Arab name in al-Idrisi is Halk al-Zawiya. Halk means a cove (cf. Halq al-Wadi, "La Goulette", in Tunisia) which translates directly to Mozarab/Late Latin Angra, whose diminutive would be *Angrina > Angrinha. This is a perfect match for the unusual place-name "Ingrina", the actual name of the conspicuous little cove adjacent to "Zavial". See Figure 23.
Lagos has also been traditionally identified with the Roman Laccobriga. In fact there are two distinct Lacobriga in Greco-Roman Iberian sources. The first is a civitas of the vaccei identified by Pliny and by Ptolemy, and also in the Antonine Itinerary and the Ravennate Cosmology. The second is the Laccobriga in the coast of Southern Lusitania, positively identified only by Pomponius Mela in the area of the Sacred Promontory, between Ossonoba and Portus Hannibalis if we follow the enumeration order of this author.
There is also a reference in Plutarch to the Lacobriges (Λακοβρίγας), allies of Sertorius whose town (πóλει) was located at two days' march from Q. Caecilius Mettelus' camp in Conistorgis. This would most probably be the Laccobriga of Algarve, by its nearness of Conistorgis.
In historic-linguistic terms, Laccobriga fits perfectly the topography of Lagos (see Figure 20). The termination –briga gives linguistic and geographic information. It indicates a foundation of a Celtic-speaking population, common in the Portuguese southwestern littoral, where Mirobriga and Cetobriga are also known. Briga also indicates both a site on a height and a human settlement which is a fort. It fits therefore the Latin concept of oppidum. The radix Lacco- is considered Celtic or already the result of Roman hybridization, always with the meaning of lake.
A paleo-physiographic reconstitution reveals a large inner lagoon in the Northern zone of actual Lagos (with an estimated area of c. 1000 acres), parts of which remained more or less flooded up to Early Modernity. The area had a regular Roman occupation disposed around the ancient lagoon margin and the surrounding hills. This dispersed rural habitat is manifested by common building structures and materials, sepultures and some mosaic floors. There are also extended topographic marks of a presumable fossil field centuriation in the zone of "Espiche", a few miles to the West.
Amongst this occupation, three centers stand out:
· A pre-Roman oppidum has been identified at "Monte Molião", a small hill by the old lagoon, founded in the end of BCE 4 and also having later Roman occupation, peaking in middle CE 2, even if the known remains do not correspond to a proper urban settlement. While no specific Sertorian military epigraphy has been found, the defensive structures and materials of "Monte Molião" are contemporary of the Sertorian war, and therefore chronologically and functionally compatible with the Plutarchian Lacobriga.
· In the area of "Quinta do Paul" and "Jardim" were found the most important surviving Roman remains in the whole area of Lagos, recognized as such for centuries. The tradition of the location of Laccobriga comes from fanciful narratives based on these findings, still partially surviving in late CE 19. They probably belonged to a late palatial villa complex, whose condition is presently unknown.
· On both sides of the channel of Lagos, especially in the area of the present city, between the ancient brooks of "Touro" and "Naus", was recently discovered a complex of fish-processing factories, with a dedicated necropolis and remains of residential quarters. It likely corresponded to a marine industrial vicus. More important remains were found further north, including the Roman dam of "Fonte Coberta", part of the water supply system to the riverside quarters.
In synthesis, both the right littoral position between Faro and Sagres and the geomorphologic and archaeological scenarios of Lagos are compatible with the Laccobriga of Mela and Plutarch. However, it must be said that no positive epigraphic evidence exists associating Laccobriga with Lagos, either in coins or monuments. In fact Laccobriga is the only proto-urban center of Algarve without its own mint (all the other five centers of Algarve, including Myrtilis had mints at the end of BCE 2 or in the first half of BCE 1). This omission is quite plausible if this is the place referenced by Plutarch: as the ally of Sertorius in the defeat of Aquinus, they would certainly have been punished after the victory of Pompey the Great in BCE 72, by political extinction or demotion.
The post-Augustan Roman habitat of Lagos follows a pattern found in other lagoonal estuaries of the Algarve: a complex of scattered sites, either farming or specialized marine stations, some of these with considerable amenities (with a climax in the first half of CE 3), with several "Italic style" villas, among which stands out a very large and luxurious palatial villa from mid CE 4, but without any formal urban center, either monumental or residential.
It is a pattern totally distinct from that found at places with complex and canonical Roman urban landscapes and rich civic epigraphs, like Balsa or Ossonoba. As far as is known, thus it seems that Laccobriga never acquired an urban status after BCE 70, and that could be one reason for its omission from the roll of the civitates of Lusitania compiled by Pliny the Elder.
To conclude, it must also be said that the location of the Lagus castle remains unknown.
MONTE AGUDO Montagut
Considering the high accuracy of the text in the orderly enumeration of the castles from Silves, by longitude, from west to east, Montagut ought to be located between "Monchique" and "Carvoeiro" (Map 5.1). The only known Islamic castle in the area, laid out as a "pointed mountain" is "Alferce". Castle "Belinho" would be the alternative, but its location does not justify the designation Montagut.
Alice Fernandes says the identification of Montagut and "Alferce" is reinforced by the maintenance of the same toponymical designation in both languages ("peak" or "peaky"). Following this explanation, the Arab form appeared probably later by translation, and bilingualism kept both forms alive for a while. However, the Arab form did survive (as did many others) and the Mozarab one disappeared after the Portuguese conquest. Luis Oliveira also accepts the possibility of the location in "Alferce".
Alternatively, Rosa Gomes identifies a medieval watchtower in a hill named precisely "Monte Agudo" near "S. Bartolomeu de Messines" (See Figure 21). Towards a positive identification with Montagut this location has the same name, a perimeter situation and a strategic position in the control of the northern road to Silves.
On the other hand, its misplacement in the enumeration order, the confirmed inexistence of any castle but only of an isolated tower (as there were very many others) and the excessive proximity with the real castle of "Messines" in Penedo Grande, which also controlled the same roadway (as in Figure 21). It seems quite odd that the author of the Narratio would confound a simple watch tower with a real castle, in contradiction with the remaining known castles.
Besides, "Monte Agudo" must have been a common place-name, a generic designation of peak. Its multiple occurrences should not be therefore strange. This "Monte Agudo" of "Messines" was much more probably one of the several watch towers that controlled accesses to the major "Messines" crossroad, dependent upon the homonym castle.
I present here both cases so that the reader may decide.
The first etymological correspondence of Portus Magnus to "Portimão" is by David Lopes, and the subject was convincingly developed by José Machado . The oldest local attestation is in an Arabic source of CE 11, where it appears as the qaria (village, hamlet) of Burtimûn. In 1189, the Narratio refers to the castle Porcimunt.
The likely Mozarabic etymon is *Portiman, an evolution of Portus Magnus. The toponym Portus Magnus is recorded in several places of the Roman Empire, notably in Mauretania Caesariensis (Bettioua in Algeria) and Hispania Terraconensis in the area of Almeria, where appears also in CE 12 in the Arabic form Bortmán, according to al-Idrisi.
The name referred initially to the water space (such as the Portus Magnus of Alexandria) or to a port settlement, subsequently to designate the river. In fact, the estuary is still designated in CE 19 as "Rio Portimão" (without preposition), distinct from river Arade which ended at the confluence of the streams of "Boina" and "Odelouca". The very name "Vila Nova de Portimão" (New town of Portimão) reveals the existence of the previous toponym.
In Roman times there would be at least five places assignable to port functions (Figure 22):
· "Portimões". A Roman fish-processing site. The site of "Portimões" (or Estrumal, or Casinhas dos Mouros) is a choronym derived from a Portuguese family anthroponym in the plural form, common in the Algarve ("Giões", etc.). This family name was certainly originated from their living at the estuary.
· "Ferragudo". A likely Roman port vicus, with former occupation from the Iron Age, and later, apparently, a fortified village in the Islamic period. There was found a Roman fishing processing factory and building materials.
A source from the end of CE 16 gives an interesting description, which may be interpreted as the ruins of a Roman riverfront horrea complex: "... [here] are some buildings they call Ferragudo, which were very imposing, and from these buildings, even if now ruined, it shows the greatness of them, because the place was surrounded by walls, and with houses of seven or eight residents [sic], which connected to one another, made in a very curious way, and the lower of them ran along the river. In many parts the walls are overthrown and the houses flattened. Their original condition was likely more impressive than it is now".
This could be the qaria of Burtimûn, later promoted to hisn, as the place is already referred as castle in 1189. The most probable location for a castle would be under the fort of "S. João do Arade" (started in 1634). However, several other alternatives exist.
· The inlet of "Mexilhoeirinha" ("Mexilhoeira of Carregação") is the best natural pier of the Algarve, capable of mooring the biggest ships of the time and the main transshipment site for dried figs up to CE 19 (therefore the name carregação = stowage). Situated at the end of the Roman roads from Faro and "Messines", here was certainly the main port facility of the estuary, or the Portus Magnus itself, by the later historical landing site of the river ferry to the right margin. The place retained this ferry assignment until the end of CE 18, when Bishop Gomes de Avelar built the "causeway of the barge", ancestor of the road to the older Arade bridge.
The point of "Carregação" defines the southern end of a long narrow bowed cove. This localization would have, in my view, the ideal conditions for an inland Roman port, sheltered from storms from the sea in all wind directions. At the time the river had here (south of the the hill of "Garcias") a width about three times the present one, covering today's dried marshes on the right bank, at the confluence with the stream of "Boina".
Between the point of "Carregação" and "Quinta do Parchal" a second port area is defined, whose most protected part has silted up, which remains surrounded by derelict warehouses and piers about the current margin.
· The islet of "Rosário". A Small habitat on the left bank of the "Odelouca" stream at its confluence with river Arade, with a Roman occupation from BCE 2 onwards, including ruins of small buildings. Its value as a port is small but occupies a strategic position in the control of river traffic, either as a trading station, sanctuary and/or toll. Leite de Vasconcellos noted in 1917 the existence upon the islet of a beverage shop!. The presence of a chapel on the island with the image of Our Lady of the Rosary (the current place name), a post-Trentine dedication, reveals an earlier religious location and thus also a probable even earlier function as river sanctuary.
I speculate that the river name Arade (already existing in 1189) may be derived from such sanctuary, with one or more islet altars, from where the Latin adjective *ARATUS may have originated.
· The site of Cilpes (peninsula of "Rocha Branca") lay in the stream of Arade, also identified in the papyrus of Artemidore as Kilibe. A Proto-Historic oppidum with oriental influence, perhaps founded in BCE 5, the seat of the *Cilibitani. Minted coin with the legends Cilpes and Cilbes . The original name was the turdetanian *Cilipis.
The site corresponds to the central place of a pre-Roman civitas which apparently was maintained during the Empire, whose name would became later Xilb in the Islamic era and then Silves. The urban center corresponding to the Roman civitas capital is unknown. Although there are some Roman findings in actual Silves, a trivial fact in all the Algarve, nothing points so far to an urban pre-Islamic origin.
Portus Hannibalis was identified in Algarve only by Pomponius Mela: "In Cuneo sunt, Myrtili, Balsa, Ossonoba; in Sacro Lacobriga, et Portus Hannibalis; in Magno, Ebora".
André de Resende initiated of the controversy over the location of Portus Hannibalis in "Alvor", due to the identification of the ruins of "Alvor-Velho", which he attributed to the Carthaginians (See Alvor above). Other authors preferred a location in the estuary of river "Arade", nearby "Portimão" (See Portimunt above). Others still locate it in the area of Sagres. The issue has been debated ever since, because no later data clarified the issue significantly. Here I posit the alternative of Sagres, a bay and settlement near Cape Saint Vincent, the ancient Sacred Promontory (Figure 23).
Portus Hannibalis may have been the waiting-port of Sagres. Crossing the Sacred Promontory from the south could force to long delays in a western port of the Algarve. The use of the Sagres Bay as a waiting port in Antiquity was therefore virtually inevitable, due to its unique situation by the Cape Saint Vincent. The difficult sea conditions here would occasion the need to beach the ships, granting full sense to the etymology Terçanabal (* tarsana'Hannibal) as a place with some shipyard facilities.
Its military control would be of strategic importance in accessing the Atlantic, including the Tagus valley. Sources from CE 15 confirm the waiting port function and the creation of infrastructure by Prince Henry the Navigator in the site of "Terçanabal", today associated with the inlets of "Mareta" and "Belixe".
A Portuguese etymology "terçã naval" (due to alternate spellings "terçanaavall" in 1459, and "terça naval" in 1505) is anachronistic because the toponym already appears in the Narratio as tarphanabal, thus pre-dating any Portuguese creation of a place-name. On the other hand, the evolution *tarafnabal < * tarf-anabal, with the Arab meaning of "cape" (corresponding to the Point of Sagres), is inappropriate because we know that this point was named in the Islamic period as *tarf al-minar from its earlier name Trasfalmenar, still used in CE 15.
However, in the Itinerarium in Terram Sanctam by friar Mauritio, dating from 1270, the name of the Point de Sagres (identified as "nose") is Trafalnaban: Post Cabo sant Vicente venitur ad nasum, qui dicitur Tarfanaban. Within this dilemma I am inclined to accept the local Portuguese source as more knowledgeable, even if 173 years later.
Excluding the tarf- root, the most probable etymology becomes *tarsana'Hannibal < *tarfanabal < Terçanabal, based on a proposition of Professor Adel Sidarus, here presented with the spelling tarsana. The original Arabic etymon would *Daar'sinnat Hannibal (according to Abdallah Khawli) and represents a remarkable literal Arabic translation of Portus Hannibalis. This toponym indicates the local presence of a military port (arsenal), and may be based on extinct elements or just be a later Arab interpretation.
The exact location of the castle of Tarphanabel is still unknown.
 David 1939, 634-5.
 Partially based in (Fraga da Silva, Ipses 2005a) and (Fraga da Silva 2006a).
 T. J. Gamito 1994.
 Faria 2000, 134.
 (Fraga da Silva 2006b), (Faria 2009) .
"Oxonoba teneat de Ambia usque Sallam, de Ipsa usque Turrem". (A. d. Fernandes 1968, 134:138).
 A temporary walled refuge, shelter for people and livestock in case of attack. (Bazzana, Cressier and Guichard 1998, 119).
 R. V. Gomes 2002, 133.
 (Resende 1996 , IV, 186): "The river.. of Alvor is a little bit more difficult, because of the meanders and sands that stilt up its mouth, similarly to what happens in Sirte. But, once entered into the widened estuary, one comes across en excellent anchorage port. Inland there is a little peninsula, quite elevated and flat on the top, where in other times an oppidum existed. There still subsist walls everywhere, from the foundations to middle height in masonry and from then up in taipa like the Punic construction. Everywhere on the top of this flat terrain, full of rubbish, walls, pieces of pottery and tiles there can be seen ruins of buildings. Because it has disappeared a long time ago did our ancestors build Alvor just near by in the inner shore of the estuary. We think that, instead of Vila Nova, town really recent was here the so called Port of Hannibal, because of the situation of the place, where ships could be perfectly defended...".
 (Teichner 2008, 417-47), Cf. (Ellis 2002, 43).
 Meaning a parochial church and an administrative center of a rural territory.
 (Sarrão 1983 , 150). An acceptable way at the time for referring to a source of water with miraculous properties, but not officially recognized as such by the post-Trentine Church. Post-Trentine Catholic doctrine was a major result of Counter Reformation c. mid CE 16. Several of its principles were applied earlier in Iberia, since 1493-97.
 (Almeida 1948, v. 3, p. 490).
 (F. A. Oliveira 1907, 97).
 A much compacted but still very long note is necessary to summarize the historical geographical issues of regional geo-conditioned cult places, across religious shifts and reforms.
It is worth mentioning the existence, current until the second decade of CE 12, of mixed sanctuaries attended by Muslim and Christian Mozarabs (and possibly also by Jews). The regional importance of these sanctuaries was based on the influence of popular Sufism in rural and urban lower class people. Popular Sufism, formally an Islamic belief, was shared by local Muladi (converted) and populations of Berber origin (information of Abdallah Khawli). Its influence allowed the wide regional support to the muridun insurrection, leaded by Ibn Casi, a self-proclaimed Mahdi (1144-1151) (Picard 2000, 92-100). Popular Sufism had numerous points in common with Mozarabic Christianity in respect to naturalistic beliefs and rituals, overwhelmingly from a pre-Christian tradition.
It should be recalled that the Algarve existed under a powerful Catholic hierarchy only during the short period of Visigothic rule (From Suintila/Swinthila, the victor of the Byzantines c. 624, to the Islamic conquest of 711, therefore less than a century!). Rural Christianization was consequently identitarian and superficial. Christian ideology may have been reinforced in lower social strata until CE 10, as a political reaction to the Islamic conquest, but without effective coercive power and under progressive social relegation. This context contributed to create an extended Christian hagiologic interpretatio of a dominantly polytheistic naturalism. Older shrines were then renamed with the names of later Byzantine saints, such as Sancti Blasius (Saint Blaise), Arabized as Xanbaras (Kahwli, Fraga da Silva and Fernandes 2007, 49), whose barely Christianized naturalistic tradition survived until recently (Fraga da Silva, São Brás de Alportel na Antiguidade 2002, 80-1).
Al-Idrisi notes the famous example of the rich "Church of the Crow" (kanisat-al-gurab), where there was also a mosque and where the Christian monks were to provide meals to travelers of both religions (Bresc, Nef and Jaubert 1999, 263). Shrines of medical nature - particularly aquatic - should be much more common and popular, as demonstrated by examples of their survival after the Portuguese conquest, until the end of CE 15. Afterwards they were systematically eliminated or purged of religious expressions of unorthodox pre-Trentine cult.
At least three important aquatic sanctuaries by the sea are known in Algarve: Alvor, Faro and Luz de Tavira. The one in Faro gave rise to the famous miracle, popularized in the poetry of Alfonso X of Castile, and that led to the renaming of the city of Ossonoba (Arabized as Uksunuba) to Santa Maria! After 1775 this sanctuary was dedicated to "Our Lady of the O", protector of pregnancy and good delivey (Paula and Paula 1993, 235).
In the old sanctuary of Luz de Tavira (a place named "Holy Source", on the outskirts of the ancient Roman city of Balsa) the ancient ritual of healing baths was extinguished and the cult moved in early CE 16 to a brand new church under strict canonical ecclesiastical control, rededicated to Our Lady of Light.
Other sanctuaries did probably exist, always dedicated to a female Entity Christianized as St. Mary, with attributes of shining light and with broad powers over themes of the third Dumézilian function: sea, land and womb fertility and water medical healing for men and animals. It is only the latter that remain in the expressions of worship after the Portuguese conquest, progressively restricted to popular veterinarian and feminine cults in gynecological and obstetric care (About popular Marian cults Cf. (Espirito Santo 1990). An example of the denaturing "Marianization" of cults is the shrine of "Our Lady of the Afflicted" by the ancient shore of the lake of Lagos at the basis of the oppidum of Molião. With a tradition of field masses for the blessing of animals, the newer orthodox devotion to Mary was unable to obliterate the previous dedication to "São Pedro do Pulgão" i.e. St. Peter of the "Big Flea", from Mozarab "pulcon" meaning phylloxera and other insect pests! (Lapesa and Garcia 2003, 523).
 J. S. Lopes 1841, 273, 475.
 Partially based on (Fraga da Silva 2005d). See also (Mantas 1997) for further bibliography and a general approach of Roman Algarve.
 Al-Idrisi in (Mizal 1989, n. 183 p. 180).
 Resende 1996 , 186.
 Modern Carrión de los Condes, Palencia, Spain.
 Respectively (Pliny, Naturalis Historia, 3,3,26), (Ptolemy, Geographia, 2,6,49), (It. Antonini Aug., 449.3, 454.1) and (Ravennate, 318.15).
 Mela, De Chorographia, III,1,7.
 (Plutarch, Lives, VIII, Sertorius, 13,7). The specific reference to Conistorgis (maybe Beja, Portugal or further South) is in Sallust (Historiae, I, Frag. 119. M). I follow here the interpretation of (Maia 1987, vol. 1, p. 171-173).
 García Alonso 2006.
 Garcia Alonso 2003, 95-96, n.19.
 Hoz 2010, 311.
 Hence the local place-names of Lago (lake), Paul (moor), Sargaçal (seaweed land) and Lagos itself.
 The archaeology of Lagos has had a strong development in the last years. Several partial reports are being presented to the annual meetings of archeology of Algarve that take place in Silves (proceedings published in Xelb journal since 2003 to the present day). A partial archaeological digest was published in 2007 (Arruda 2007).
 Fraga da Silva 2007a, map p. 76-7.
 Rocha 1991 , 15-22.
 Edification complex is an adequate descriptive term of a number of neighboring individualized edifications and built spaces in a single settlement, without consideration of status, function and inner layout structure. This concept is useful to describe imperfectly know sub-urban habitats.
The concept may be restricted to a larger scale (i.e., to a shorter range), as non-habitat structural complexes, such as those found in extractive industries and multiple shrine sanctuaries. In this sense, a necropolis may be considered a sepulture/mausolea complex.
The concept may also be extended to a smaller scale (i.e., to a larger range) as a settlement complex: a neighboring group of individualized settlements or habitats, without consideration of status or specialization profile. Settlement complexes typically originate around or along a physiographical feature (such as an estuary or inner lagoon), a geological district (mining zone) or as suburban habitats of a recognizable central place.
 The first identification of Munchite with "Castelo da Nave" near Monchique and Montagut with the castle of "Pedra Branca" near Alferce is by (Gascon 1955, 61-2).
 (L. Oliveira 1999, n. 22 p. 45). I thank the author for the information about this reference.
 (R. V. Gomes 2002, 126). The distance measured in Google EarthTM is c. 2700 meters.
 Silva Lopes contradicts himself: "one can see they are mentioned in the direction from west to east, and therefore also in that direction we will seek to discover them" (Lopes and Gazzera 1844 , n. 24, p. 99) but later (Idem, 105) proposes an identification with the site "Montagudo" in the parish of "S. Estevão (Tavira)", which breaks the west-east rule and has a geographical location absolutely unrealistic and far too distant from the territory of Silves, jumping over the territories of "Loulé" and "Faro"!
 Still surviving, but with a tendency in later toponymy for "monte" to be replaced by "cerro" and "agudo" by "pico" or "picota".
 Partially based in (Fraga da Silva 2005c).
 Machado 2003, entry Portimão, p.1201.
 Umdat aTTabib fî ma´rifat annabat, traslatable as "The key to the medical knowledge of plants" (Rabat, 1990, T. 2, p. 656), from the botanist Abu al-Khayr al-Ichbili (CE 11), who says: "I saw this type [of a plant named Qaqilli] in the region of Silves, in a qaria called Burtimûn". (Information of Abdallah Khawli).
 Cf. Philological Supplement.
 Also known as Sinu Urcitanum in Graeco-Roman geography: (Mela, De Chorographia II, 94).
 Bresc, Nef and Jaubert 1999, 135-137.
 "Moors cottages": An early popular designation for remains of Roman cetariae, necropolis and silos. (Santos 1971-72, vol.1, p.128).
 Information of Alice Fernandes.
 Santos 1971-72, vol.1, p.128.
 Sarrão 1983 , 133-182.
 Such as the placement of the Fort of Santa Catarina (Magalhães 2008, 183-8) or of the Convent of São Francisco/Senhora da Esperança (Marado 2006, 75-9).
 Vasconcellos 1927, v. 2, p. 257-9.
Cibilitani in (Pliny, Naturalis Historia, IV, 35).
 Faria 2006, 219.
 Partial based in (Fraga da Silva 2005b).
 Mela, De Chorographia, 3,1,7.
 Promunturium Sacrum/ Ίερŵ άκρωτηρίω. Mentioned by (Strabo, Geographia 3.1.3 to 3.1.6) , (Mela, De Chorographia, 3,1,7), (Pliny the Elder, Naturalis Historia, IV, 113, 115, 116) and (Ptolemy, Geographia 2.5.2).
 Ships had to wait until the winds allow a broad outline off the cape, avoiding both the facing winds before the cape as well as the northern winds after its passage.This situation persists for smaller sailing ships even in the present days.
 In BCE 212 the Carthaginian admirals Mago Barca and Hasdrubal Gisco wintered their squads, respectively, in the Tagus and the Algarve (Polybius, The Histories, X, 7, 5).
 Iria 1960, 36, 49.
 Contra (Iria 1960, 49-50).
 Contra Fernandes Lopes, in (Iria 1960, 49).
 In a document of 1443, the regent of Portugal, D. Pedro, gives to his brother D. Henrique the Cape of Trasfalmenar and a league around it to settle the new Vila do Infante, nowadays Sagres (Iria 1960, 21-3).
 Storm 1880, I, 6-7.
 Alice Fernandes defends that the tarf form was a later contamination (See Philological Supplement, below).
 According to (Corriente 1999, 231, entry “arsenal”) .