[190] Throughout this book there are occasional references to boundaries, considered sometimes historically and sometimes, implicitly at least, as natural phenomena. As they are always potentially important, and sometimes crucially so, a few pages may be taken profitably to consider them.
 

In Iberia international boundary lines have been drawn, for the most part, through zones of limited desirability. Roughly nine-tenths of the Spanish-Portuguese border is located in such areas, and the reflection of this fact is to be seen in the sparseness of population nearly everywhere along the frontier (Fig. 13). ⁽¹⁾ The one important exception is the area of the lower [192] Minho River, which is certainly not a region characterized by undesirability. ⁽²⁾

It should be remembered that the lower Minho River was the northern limit of the "desolation" of Alfonso I. Cities to the south had been eliminated. The filling-in of the routes of communications and the restoration of cities up to that line was accomplished by a southern group working northward from Porto. In the area of the lower Minho River a group of settlers that was oriented politically and economically toward the south met one of Galicia, where ancient connections eastward toward Spain had never been broken for more than short periods of time. ⁽³⁾

Less than fifty miles upstream from the mouth of the Minho River the boundary turns sharply southward, and the population density falls off rapidly as the boundary rises along the slopes up to the high, winter-cold Serra de Laboreiro. After crossing the sharp, deep gorge of the Lima River, the border turns eastward again, following near the crests of other high mountains, the serras of Gerez and Larouco. To the east of Larouco is another exception to the general unattractiveness of locale along the border. This is the upper valley of the Tâmega River, between the Portuguese city of Chaves and the Galician [194] town of Verín. A fertile valley, protected from the worst of the winter cold by uplands to the west, north, and east, it supports a modest cluster of population, and there is no break between settlement on the Portuguese side and in Galicia. In addition, the mountains, virtually encircling Verín, separate it clearly from the rest of Galicia. Its normal associations should be with Chaves and southward into Portugal, and one would have expected the boundary line to have been established along the barren uplands lying between the upper Limia (Lima) and the middle Miño (Minho) rivers, thus putting Verín into Portugal. However, another factor was more important than geography in the matter. Verín lies on the early pilgrimage road from Zamora to Orense, which skirts the north border of Portugal, and from there leads into Santiago de Compostela. ⁽⁴⁾ At Orense, this road is joined by that from León. The importance of Santiago as a religious and pilgrimage center after the ninth century was a sufficient reason for Verín to have remained under Spanish control. The area was remote from the center of authority of either León or Portugal. It is a relatively unimportant area to either state. An arbitrary political decision, although seemingly in violation of local economics, could have been enforced without difficulty, for actually political boundaries had, and have, very little effect upon such distant and self-sufficient communities. A Portuguese author recently referring to the area of Soajo, in the northeast of the Minho Province near the Spanish border, described the inhabitants thus: "Soajeiros are very independent. They are not irritated by the laws. They just do not pay much attention to them. Their 'leading men' are the authorities recognized by the people. Their customs are their laws . . ." ⁽⁵⁾ No doubt life in Verín continued much in the same way that it would have done had the border been drawn elsewhere.

Beyond the valley of the Tâmega River the northern boundary line is again to be found at sufficient elevation so that the [195] factor of undesirability is marked. This is the higher, more mountainous part of the province of Trás-os-Montes, where the limitation of population is imposed, not only by slope, but by duration of winter cold.

The east boundary of Trás-os-Montes is drawn through an extension of the Leonese plateau, high and sufficiently exposed to the winter cold of interior Iberia to inhibit agriculture. For all of Trás-os-Montes another fact is of importance. The province lies in the lee of the mountains that separate it from the "green Minho." The orographic barrier causes a rainshadow condition. Drought thus adds another factor to the impoverishment of cultivation. ⁽⁶⁾ All of this "explains the relative isolation and the tenuous social and economic relations between the human groups on the two sides of the frontier. This isolation is particularly marked along the Leonese frontier. In almost all [197] of the territory between the Macas and the lands of Vinhais, the character of the frontier is much like that of a marca (a remote frontier province), so slight is the human occupation. He who goes from Bragança to Puebla de Senabria [sic] has the bleak sensation of travelhng in 'Terra nullius domini' . . . ," so spoke a geographer born and raised in the region. ⁽⁷⁾

The physical conditions of eastern Trás-os-Montes extend beyond the Douro until south of the Serra das Mesas, a Portuguese continuation of the great central mountain system of Spain. Both elevation and rainshadow continue to limit productivity along the boundary zone. South of this, the lack of rainfall and the winter cold continue to be the obstacles to a more productive use of the land, but the causes are somewhat different. Although elevation decreases, outbreaks of cold from the Iberian meseta are a threat to crops through the winter. The growing season is longer than that of Trás-os-Montes, but so is the period of summer drought and, unfortunately, in a land of little rain, the evaporation rate is high during the spring and fall when there is maximum precipitation. ⁽⁸⁾

There are minor areas of increased population near the eastern border of Middle Portugal, those of Sabugal, Portalegre, and Elvas. In each case there is a higher rainfall, due to somewhat greater elevation. In no case, however, could the population be called dense. ⁽⁹⁾ Beyond Elvas, toward the south, the [199] obvious lack of appeal of the boundary zone continues to be shown in the sparseness of population.

The length of the summer season and the evaporation rate both increase as total rainfall remains low down to the area of the lower slopes of the Serra do Caldeirão, which separates the Alentejo Province from the Algarve. Rainfall is higher here, but this advantage is more than offset by the nature of the soil materials. These schists do not allow an easy penetration of water. Rainfall mostly drains off by surface flow, and in these now deforested mountains each winter season sees the removal of another sheet of surface. Two generations ago, directly following upon the clearing of the natural vegetation, the harvests were copious. Now after years of erosion, the yield is but a fraction of that of the early years. Only at the extreme southeast is there a slight increase in the population of the border zone, but this is limited to the Portuguese side. There are two reasons for it: first, the mine of Sto. Domingos, producing copper and sulphur, supports several thousand people gathered immediately around it;⁽¹⁰⁾ and second, the recently developed area of early vegetable production near Vila Real de Santo António at the mouth of the Guadiana River supports another concentration of population. Truck gardens here supply Lisbon with early vegetables. This phenomenon developed during recent generations as a result of improved transport. This population area, like that of Sto. Domingos, appears as merely a spot against the border, for it is limited to the coastal sands and does not follow the Guadiana upstream. Across the Guadiana, the Spanish area lacks the economic advantage of Vila Real. Spanish growers would not be able to compete in the Lisbon market due to import restrictions, and there is no nearby home market to support such an enterprise.

National states require more than a frontier zone. They require a boundary line. This line in Iberia, more often than not, is a river; such is the case for over 60 percent of the Portuguese land boundaries and for over 70 percent of the eastern border. Along the eastern border, rivers are signally useful, for they not only traverse land of little attraction, but have carved deep, steep-sided canyons into the plateau surface, which effectively cut communications. Fifty-four per cent of the eastern border is along such canyons. A striking example is that of the international Douro between Paradela and Barca d'Alva, where the river flows through a canyon, at times with vertical walls several hundred feet high, for over 76 miles, falling 1600 feet in the distance. The great descent takes place through a series of falls and rapids, making navigation impossible. ⁽¹¹⁾ A left-bank tributary of the Douro, the Agueda River, and farther south its affluent, the Tourões, both act as boundaries. The Agueda cuts a deep canyon; the Tourões does so in part of its course.

South of the Tourões, in the relatively high country between Ciudad Rodrigo and Guarda, the boundary does not follow rivers but is drawn across the headwaters of several small streams. It follows approximately the divide between the drainage of the Agueda and Côa rivers. Beyond this area, in the Tejo drainage, streams again are used to mark the boundary line; the Torto and the Erges both run well below the surface of the country on either side of them. Where the Erges meets the Tejo, the boundary line turns sharply west, to follow the larger stream in a deep canyon to its juncture with the Sever. Here it again turns sharply, in an acute angle, to follow the canyon of the latter almost to its source in the Serra de S. Mamede. South of the Serra there is another stretch of boundary that is erratic and oblivious of hydrography. This is part of the territory passed back and forth in the marriage portions of the [202] seventeenth and eighteenth centuries whose ultimate possession was decided arbitrarily, although in amity, and largely in disregard of physical factors.

A water course again becomes the boundary along the lower Caia River just before it joins the Guadiana. The boundary follows this river and the Guadiana almost to the latitude of Mourão, where there are other lands of frequent exchange between the Iberian monarchs. These too were finally apportioned amicably, although arbitrarily in terms of physical factors. The boundary of Portugal here runs well to the east of the Guadiana. At the latitude of Serpa the boundary meets and follows the Chança (Chanza) River to its juncture with the Guadiana, and then again follows the Guadiana. This is the ultimate stretch of the international boundary southward to the ocean.

In Iberia, particularly, another fact makes rivers useful as boundaries. This is the great difference between flood and low water, and the resulting economic uselessness, which makes them barriers rather than means of communication. The bare, deforested Iberian meseta has a quick run-off, and the floods of the rainy season are sudden and devastating wherever the streams are not incased between high canyon walls. This factor makes the use of the streams difficult for virtually any purpose. ⁽¹²⁾

Rivers have served Portugal and Spain as boundaries as far back in time as we have knowledge of the peninsula. For Rome it was standard practice to use them, and this was especially notable in Iberia. As stated above, it was the Roman intent, for reasons beneficial to herself in the matter of control, to follow custom where possible, and her choice commonly mirrored the established habits of the local peoples. It should not be surprising that later boundary makers and present governments have found rivers to be equally serviceable.

[203] Nevertheless, no matter how useful the rivers have been as political boundaries, the most important factor in the political separation of the Iberian countries is the distribution of population. Some of the distribution is, no doubt, induced by the fact of the boundary, but it is clearly obvious that the physical nature of the land has limited population density along the line of the present frontier.

1. Figure 13 is taken from Figure 4 of Luis de Hoyos Sáinz, La Densidad de población y el acrecentamiento en España. The statistics used by Hoyos Sáinz were those of the 1940 census. Unfortunately the statistics from the 1950 census are not available in such form as to be used for a new map. However, a comparison of the 1950 statistics for the sparsely settled border region of Spain where it touches eastern Portugal with those used by Hoyos Sáinz shows that there has been no material change in the situation in the decade. That border area is still, as it was in 1940, a land of slight attraction for population.

For a discussion of the distribution of population along the international boundary, see Artur de Magalhães Basto, "A Fronteira Hispano-Portuguesa," O Instituto, LXX (1923), 62-63.

2. This fact is made manifest also by the Hoyos Sáinz map of population, but the end map of J. Dantin Cereceda, Distribución geográfica de la población en Galicia, makes it even more obvious. Compare this with the map of Distribüição da população de Portugal based on the 1940 census, published by the Centro de Estudos Geográficos of Lisbon under the direction of Orlando Ribeiro. Both of the latter references are more detailed than the work of Hoyos Sáinz, but unfortunately they are concerned with limited areas and do not serve for a comparison with the rest of Iberia.

3. Although the river splits a population cluster, it has served usefully as an administrative boundary as far back as prehistory.

4. António López Ferreiro, Historia de la Santa A. M. Iglesia de Santiago de Compostela, V, p. 91.

5. Basto, "A Fronteira Hispano-Portuguesa," O Instituto, LXX, 104.

6. Vergilio Taborda, Alto Trás-os-Montes, p. 9. Also see Jorge Thas, Rio de Onor, especially pp. 79-85. The village of Rio de Onor, which Dias describes, straddles the international boundary at the extreme northeast of Trás-os-Montes. It is a small oasis of fertility in an otherwise barren land.

7. Taborda, Alto Trás-os-Montes (author's translation), p. 21.

8. The sparseness of population of the Alentejo has been attributed to the system of latifundia, the great estates. This may be accurate but one cannot be sure. We do not know that exploitation of a territory so limited by physical factors could be effectively accomplished in small units. Absentee ownership, also typical of the area, is, of course, another matter.

9. The equation of increased population and rainfall is obvious in a comparison of the rainfall in H. Amorim Ferreira, Carta Pluviométrica de Portugal of 1943, with population density as shown on the Ribeiro map (see Note 2 of this chapter). One anomaly should be noted. The population density of the Spanish province facing the Sabugal area is virtually as low as any in Spain, although rainfall there, as in Sabugal, is a little higher than in the adjacent areas in Spain. This condition, in part at least, may be ascribed to the effects of the Mesta and the discouragement of agriculture in favor of sheep herding.

10. This cluster appears on the Ribeiro map as a spot of settlement in an otherwise meagerly inhabited area. Such an agglomeration of population probably exists also in the area near the mines of Rio Tinto, which are geologically akin to those of Sto. Domingos and represent the Spanish counterpart. Unfortunately there is no Spanish map of sufficient detail available to substantiate such an assumption.

11. None of the Portuguese boundary streams are navigable for useful distances except at the lower extremities.

12. The Guadiana at Mértola increases over eighty feet (25 meters) above low water during flood. The Tajo at Alcántara increases up to nearly one hundred feet (30 meters). Hermann Lautensach, "Lebensraumfragen der Iberischen Völker," Lebensraumfragen Euro paischer Völker, I: Europa, p. 505; Pedro M. González Quijano, Mapa pluviométrico de España, pp. 277 et seq.