PART III—
TWO PROVINCES

Up to here at both the national and the local level, the account of royal policy and agrarian change has been essentially episodic: of evolving theories of agrarian reform, of the struggles of ministers to solve social ills and fiscal crises, and of the evolution of different towns. The last chapter drew some general lessons from the histories of seven towns, but their validity is limited by the very uniqueness of these towns.

In an attempt to establish some historical regularities at a more macro level, the last part of this study deals with two provinces: Salamanca and Jaén. Rather than be fully descriptive, it looks at the structure and development of the rural society and economy in the form of variables that can be quantified and compared. Thus we may attempt to establish correlations between structures existing at the time of the catastro, the evolution of the next half century, and the changes that disentail produced.

The choice of these provinces results from the structure of the administration of the desamortización at the local level. Permanent commissioners of the Amortization and Consolidation Funds were established in every provincial capital and in the major cabezas de partido, while a number of lesser cabezas de partido had commissioners temporarily located in them. At least 137 localities appear as seats of commissioners at one time or another. For ease of collecting data, it made sense to choose the sales handled by certain commissioners, because they can be easily identified in the Madrid notarial archives. After some trial and error, I selected the documentation of two permanent commissioners, those in

the cities of Salamanca and Jaén. The commissioner at Jaén had charge of the entire province, but in Salamanca province the extensive partido of Ciudad Rodrigo in the southwest was administered by a commissioner in that city, and I did not record his sales. My study of Salamanca province therefore lacks this territory.

They proved fortunate choices. The provinces are located in different regions of central Spain, providing examples of both sides of the Salamanca-Albacete line, and each province has considerable variety of landscape and social structure, facilitating comparative analysis. Both provinces were also areas that preoccupied the royal reformers, and they had a high percentage of church property sold.

Chapter XV—
The Social and Economic Spectrum of Jaén and Salamanca

The province of Salamanca is the southernmost of the kingdom of León. It is located south of the Rio Duero, which drains the northern meseta of Castile, a frontier region that was captured from the Muslims and settled during the eleventh and twelfth centuries in the second major push of the Reconquest. To the west it borders on Portugal and to the south is separated from Extremadura by the central mountain range. To the north and east it has no clear boundary, for the flat, rich grain-bearing tableland of the northern meseta extends south into Salamanca and forms the core of the province.

The capital city, Salamanca, sits on high ground above the Rio Tormes, and the tableland stretches out in all directions around it. Under the old regime most of this plain belonged to the partido of Salamanca, the portion of the province directly under the jurisdiction of the provincial capital and thus of the crown. To the southeast, the plains bordering on Ávila province formed the partido of Alba de. Tormes, which belonged to the señorío of the Duque de Alba; and to the west, beyond the partido of Salamanca, was that of Ledesma, most of which was part of the señorío of the Duque de Alburquerque.

Farther to the southwest, the plain spreads on into the partido of Ciudad Rodrigo, second only to that of Salamanca in extent. It accounts for most of the frontier of the province with Portugal and Extremadura. As explained before, because the royal commissioner for the disentail of Carlos IV located in Salamanca city did not deal with the partido of Ciudad Rodrigo, it is not included in this study.

The sierras in the south of the province contained in the eighteenth century a number of partidos, all under seigneurial jurisdiction. These included those of Miranda del Castañar, Montemayor, and Béjar, lying in the valleys of the sierras of Peña de Francia and Béjar, most of which are still part of the province of Salamanca. The province also embraced the partidos of Barco de Ávila, Piedrahita, and El Mirón, in the valleys to the north of the massive Sierra de Gredos, which were transferred in the nineteenth century to the province of Ávila. When we observed the town of El Mirón, we formed a sense of what these sierra regions were like.

Leaving Salamanca and crossing the central sierras and the meseta of New Castile one reaches the kingdom (reyno ) of Jaén, as it was still known in the eighteenth century. It forms the northeastern corner of Andalusia. The Guadalquivir River, whose basin forms the heart of Andalusia, rises in the imposing Sierra de Cazorla, which separates Jaén from the kingdom of Murcia to the east. To the north, the Sierra Morena lies between Jaén and La Mancha, while to the south a series of sharp ranges divide it from the kingdom of Granada. Only to the west does Jaén lack a clear geographic frontier; the province of Córdoba occupies the next portion of the Guadalquivir valley. The eighteenth-century kingdom was somewhat smaller than the present province, for the partido of Orcera now belonging to it was added from La Mancha and Murcia in the nineteenth century.

When we looked at the town of Baños, we saw that the main route from Castile entered Andalusia through Jaén after crossing the Sierra Morena by the passes of El Viso or Despeñaperros. Jaén was the first Muslim kingdom conquered by the Christians after the decisive victory of Las Navas de Tolosa—northernmost town of the kingdom—in 1212. The twin cities of Baeza and Úbeda fell rapidly, and then the Christian forces gathered their strength for a generation before pushing on to Jaén, Córdoba, and Seville in the middle of the century. Granada, to the south, held out until after 1480.

For two and a half centuries the southern mountain ranges of Jaén province were the frontier between the Christians and the Muslims, as they had been earlier between rival Muslim kingdoms. Today the fortresses that dominate the towns of Jaén and the numerous towers scattered along these mountains bear witness to this history. Even the names of many towns recall the military past: Alcalá la Real, La Guardia, Torre del Campo, Torreperogil (Torre Pedro Gil), Torredonjimeno, and the simple Torres. Other names are evidence of later resettlement: Villa-

nueva del Arzobispo, Villanueva de la Reina, Mancha Real, La Carolina (in the eighteenth century), and more recently Puente del Obispo.

In order to grasp the structure of rural society and economy of these two provinces and how it was evolving at the end of the old regime, one may start by observing how individuals behaved in the disentail. The data come from the deeds of deposit (escrituras de imposición) recorded in Madrid by the two notaries Juan Manuel López Fando and Feliciano del Corral. We met these documents in Chapter 5, where they formed the basis for calculating the extent of disentail throughout the monarchy. They are copies of the acknowledgments that the crown gave to the former owners of disentailed properties and redeemed censos specifying the amount of royal indebtedness to the former owner and the annual interest payment due on it. Between 1 January 1800 and 30 January 1801, deeds of deposit were also issued by provincial notaries. My best estimate is that the deeds of the two Madrid notaries cover about 75 percent of the sales outside Madrid province. A search through the historical archives of the provinces of Jaén and Salamanca unearthed only a few examples of deeds of deposit registered at the provincial level. Since they are not complete, I judged that their inclusion would distort more than strengthen the results, and I have based the study of the provinces and their zones on the Madrid records alone.

The deeds of deposit are printed on standard forms of three pages, which quote in extenso the royal decrees and specify the nature of the royal obligation. They provide blank spaces in which the notaries recorded the location of the commissioner, the name and location of the former owner, the date of payment for the purchase, the name(s) of the buyer(s) or the redeemer(s) of censos, the amount paid and the form in which paid (hard currency, vales reales, etc.), the number of installments for payment, and usually the type and location of the property. Each document contains up to fourteen items of usable data. I found and recorded 4,642 deeds for Jaén province and 3,314 for Salamanca.

The first step in preparing the data for analysis was to group together all the purchases made by each individual. One here faced the problem familiar to demographic historians who have reconstituted families: the name of an individual often appears with more than one spelling. Spanish orthography had not yet been standardized, and notaries sometimes recorded one and sometimes two surnames (apellidos). There are also many common Christian names and surnames in Spanish, so that identical names do not necessarily refer to the same person. I used my best judgment. If similar names appeared in the same town, I usually as-

sumed that they belonged to the same individual; if in distant towns, I did not. On the whole, the likely error is of treating the purchases of one individual as if they were made by two or more, rather than vice versa.

I concluded that 2,552 individuals made purchases in Jaén (1.8 purchases per buyer) and 2,148 individuals in Salamanca (1.5 purchases per buyer). These include 67 "unnamed" individuals in Jaén and 233 in Salamanca, my solution for dealing with deeds which read "so-and-so and two others," "so-and-so and his companions," and "various persons." In the case of unspecified plurals, I assumed three buyers. In all cases of more than one buyer, I divided the amount spent equally among them, for lack of information on each buyer's share. One must remember that the identity of the buyers and the amount they spent individually were of marginal interest in the deeds, which were contracts between the king and former owners. The information essential to this study was frequently omitted or entered hastily, for the copyists of the Madrid notaries produced fifty deeds per working day on the average, and in years of intense activity many times more.

In order to proceed, I ranked the buyers in each province in ascending order according to the amount they spent. This procedure produced Lorenz curves, with the cumulative percent of buyers plotted against the cumulative percent of the amount they spent. By their nature, the curves rise slowly at first and sharply at the end. I then divided the purchasers ranked on the curves into four levels as shown in Figure 15.1. In both provinces the small buyers in Level 1 are more numerous than the large buyers in Level 4, although the amount spent by the latter totaled ten times that of the former, a clear indication of the disparity among the buyers.

2

Once one has accomplished this step, it becomes possible to compare the characteristics of the different levels of buyers and the patterns they followed in making their purchases. Three features prove revealing: the type of property the buyers acquired, the terms of payment, and the legal estate of the buyers. The months, years, and seasons in which the buyers made their payments were also compared, but with no meaningful results. The decisive factor in determining when a buyer could make a purchase was the date that the royal commissioner put the property up for sale. The data indicate that this decision was not related in

Figure 15.1.
Allocation of Buyers to Four Different Levels
NOTE: Buyers are ranked in ascending order of
the amount of each individual's total purchases:
Level 1 —buyers of the bottom 5 percent of total purchases (0–5.0 percent)
Level  2—buyers of the next 15 percent of total purchases (5.1–20.0 percent)
Level  3—buyers of the next 30 percent of total purchases (20.1–50.0 percent)
Level  4—buyers of the top 50 percent of total purchases (50.1–100 percent)

any regular pattern to the level of buyer to which the property would appeal.

Table 15.1 shows how each level of buyers of Jaén province divided the amount it spent among the different types of property and the redemption of censos. By showing the share of their money that each level of buyers devoted to each type of property, this table reveals their preferences, with arrows to identify the trends. The lower was the level of buyers, the more their money went into arable, vineyards, orchards, and redemption of censos, and there was a strong but not complete trend in this direction for urban properties (Level 2 devoted a larger share than Level 1 to this type). On the other hand, the higher the level of buyers, the more they spent on olive groves and rural estates.

Table 15.1. Property and Buyers, Jaén Province   Level of Buyers     Type of Property 1 (bottom) 2 3 4 (top) Total Province Total Amount (000s rs.) ( percent of total purchases ) Arable   32.3 12,266 Improved/irrigated   8.2 3,128 Vineyards   0.5 203 Orchards   1.4 541 Olive groves   17.8 6,761 Rural estates   25.4 9,630 Urban properties   9.5 3,588 Redeemed censos   4.8 1,824 Total percent   99.9   Total amount  (000s rs.)     37,941 SOURCE. AHPM, deeds of deposit. NOTES. The types of property in this and subsequent tables are as follows: Arable. The deeds describe many properties simply as tierras or propiedades or fincas. The pattern of deeds indicates that these unspecified terms refer in most cases to arable. I have included them in this category. The percentage of arable that is unspecified in this table is (Levels 1–4): 14.0, 10.5, 9.8, 6.5; total amount 3,192,000 rs. Improved/irrigated. Includes huertas (irrigated vegetable plots) and cortinas (in Salamanca, fields adjacent to the town nucleus enclosed by stone walls, used for the cultivation of flax, "green fodder" [verdes de centeno (rye) or . . . de cebada (barley)], and the like). On cortinas see García Fernández, "Champs ouverts," 709–14. Orchards. Fruit and nut trees, etc. Rural estates. Any rural property with buildings on it (cortijos in Jaén and alquerías and cotos redondos in Salamanca are the most important, but this classification also includes olive groves with buildings and mills, and at a more modest level huertas with a small house). Urban properties. Buildings, lots, corrals, etc., in the town nucleus. Arrows: Where the shares of the different levels vary in a regular trend, the table has an arrow indicating the direction of the increasing preference for this type of purchase. When one reading is out of line, there is a broken arrow.

The table demonstrates a feature of the disentail that we already encountered in the individual towns of the province. One may recall that the catastro revealed that nonresident owners preferred to lease properties whose cultivation required large labor inputs (intensive agriculture) while tilling directly with hired labor and the administration of stewards the properties that called for less labor or whose production was oriented toward the outside market. Large resident landowners and local religious institutions had similar preferences, although they followed them in a smaller proportion of the cases. The economic rationality became evident when one considered the costs of the various inputs for the different types of property and the net income resulting from leasing compared with direct administration. Table 15.1 shows that when it came time to buy disentailed properties, the large buyers preferred the kinds of property that large and nonresident owners had learned was most profitable to exploit directly: olive groves and rural estates (that is, cortijos). These properties produced for the market and relied primarily on seasonal hired labor. Large buyers were not eager, however, to outbid smaller buyers for labor-intensive lands and those whose produce went mostly to feed the town, with the result that the lower levels of buyers spent a larger proportion of their money on these categories. This is notably the case for arable not located in cortijos and therefore probably distributed in small plots. Although they account for only a small part of the disentail, vineyards and orchards also went proportionately more to the lower levels.

The lower levels also demonstrated a stronger preference than the upper for urban properties and the redemption of censos. In both instances the upper levels were following a comprehensible investment policy. For them, in most cases, the purchase of buildings in the town nuclei would mean buying rental properties, from which they could expect a limited income. Leasing buildings to others attracted them no more than leasing agricultural lands. For small buyers, however, the prospect of owning a house rather than renting one had enough appeal to persuade them to spend money here rather than compete for farm land. The reasoning with censos was similar. Whether large buyers were proportionately more or less encumbered with censos than small buyers one cannot know, nor does it matter, since the amounts redeemed were not enough to have represented major shares of local indebtedness. For large buyers, the prospects of investing in market agriculture was on the whole more attractive than paying off obligations bearing 3-percent in-

terest. Smaller buyers, with fewer prospects of production for the market, showed a greater preference to redeem their debts.

All in all, the evidence reveals that as a group the people of Jaén with disposable wealth at the time of the disentail had acquired a fund of comprehensible economic reasoning. The disentail of Carlos IV placed a sufficient number of properties of various types on the market for this reasoning to produce regular patterns in the purchases of them. The trends in Table 15.1 are too clear to be the product of pure chance. Only one form of property shows no trend: improved and irrigated lands. This exception is puzzling, because most of these would be labor-

Table 15.2. Property and Buyers, Salamanca Province   Level of Buyers     Type of Property 1 (bottom ) 2 3 4 (top ) Total Province Total Amount (000s rs.) (percent of total purchases ) Arable   58.8 21,390 Pastures   7.2 2,613 Improved/irrigated   1.9 694 Vineyards   1.0 365 Orchards   0.3 90 Rural estates   16.4 5,987 Urban properties   9.9 3,608 Redeemed censos   4.4 1,617 Total percent   99.9   Total amount  (000s rs.)     36,364 SOURCE. AHPM, deeds of deposit. NOTE. For types of property, see Table 15.1. Olive groves are too few to include.  Pastures are a new type. The percentages of arable that are "unspecified" properties are (Levels 1–4) 29.8 percent, 31.8 percent, 45.0 percent, 37.7 percent; total amount 14,060,000 rs.

intensive properties and in the towns we looked at large private owners preferred to lease them to others.

At first sight, the pattern of property acquired by the different levels of buyers in Salamanca province appears unlike that of Jaén (Table 15.2). Whereas in Jaén there was a clear trend for the smaller buyers to acquire proportionately more arable, Salamanca shows no such trend although there is an indication that larger buyers directed their purchases toward arable more than the smaller ones. The difference, however, is more apparent than real. Large buyers in Jaén interested in the grain market bought cortijos, classed as "rural estates." In Salamanca, as we saw in the two towns of La Armuña, large buyers acquired the collections of numerous small grain plots that the ecclesiastical institutions had built up and whose purchase the peasants could not afford. Unimproved arable was the predominant type of property in Salamanca, accounting for 59 percent of all the money spent; even though arable was also highest in Jaén, it represented only 32 percent. The difference between the provinces lay in the willingness of large buyers in Salamanca to acquire arable to lease to peasants, as the church had done. In Jaén much commercial production was conducted with administrators and hired labor, in Salamanca with tenant farmers. Rural estates in Salamanca meant primarily alquerías and despoblados, which produced both grain and livestock. As in Jaén and for the same reason, the trend in purchasing rural estates was strongly toward the top buyers.

In Salamanca the sales included a number of pastures. The table shows that they were of two kinds, large pastures that could carry extensive herds and appealed to the wealthy, and smaller ones primarily for household animals. Commercially oriented pastures went to the top level, but below Level 4, the trend was toward the smallest buyers, who were prepared to pay extravagant prices, as was the case in El Mirón.

In other respects, the two provinces were similar. More obviously than in Jaén, large buyers shied away from labor-intensive properties, especially the improved or irrigated plots, most of them walled cortinas so typical of this part of Spain. They also spent a smaller share of their money to redeem censos, and they avoided urban properties (even though the inclusion of Salamanca city in the provincial analysis means that substantial urban properties, out of reach of small buyers, figure in the table). Curiously, in both provinces, the level that showed the greatest propensity for buying urban properties was the second, not the bottom, so that this preference was not a straight linear function of wealth.

One is led to conclude that wealthy buyers had a different mentality

than more modest ones, more open to the possibilities of the market. Before accepting this conclusion, however, there is an alternative explanation that should be considered. Perhaps the difference was simply one of opportunity. In the most obvious case, wealthy buyers spent more of their money on rural estates than small buyers. The fact is that most rural estates went for high prices because of their size, and the simple act of buying one automatically placed the buyer into one of our top levels because its cost was greater than the total amount spent by a single buyer in the lower levels.

There is one way to test whether opportunity or a different economic outlook was the primary motivating factor in the different choices of the various levels. If one tabulates only those individual purchases made by all buyers that cost less than the "breaking point" between Levels 1 and 2, that is, those purchases that a person in the lowest level could have made, one can observe whether the preferences observed for all purchases are repeated here (Table 15.3). The test is not conclusive because the upper levels spent only a small percentage of their money on purchases in this price range, and the trends turn out to be fewer and weaker than when all purchases are considered. Nevertheless, those that do appear run in the direction one would expect if the big buyers had different goals than the small buyers. In Jaén the big buyers put a greater share of their money into olive groves than the small ones, and in Salamanca the small buyers preferred vineyards and orchards, in all three cases trends that were found when all purchases were tabulated. It would appear that there was indeed a difference in mentality among the levels, large buyers as a whole showing more interest in properties that were directed to the market and could be profitably cultivated with hired labor, whether they were making large or small purchases.

3

Wealth not only gave access to properties oriented toward the market but reduced the ratio of price to expected return, making expensive properties attractive in more ways than one. The royal decrees organizing the disentail, one recalls, offered the buyers a choice of ways in which to pay for their purchases, the basic choice being between payment in hard currency or in vales reales. From the beginning, preference was given to bids in hard currency (metálico or efectivo), and within a year bids for less than the assessed value (the minimum bid was two-thirds of it) could be made only in hard currency. Moreover, the smallest

Table 15.3. Purchases Below the "Level 1–2. Breaking Point" (percent of purchases in this price range)   Level of Buyers Type of Property 1(bottom) 2 3 4 (top) Jaén Province   Olive groves      Urban properties   No trend            Arable 49.8 54.0 55.9 44.6    Redeemed censos 6.3 5.4 9.1 8.6    Percent of all purchases madea 100.0 13.7 7.6 2.6 Salamanca Province Vineyards      Orchards      Rural estates      No trend       Urban properties 13.1 4.8 13.3 10.2       Redeemed censos 8.9 6.8 6.2 7.1    Percent of all purchases madea 100 12.6 2.9 1.0 SOURCE. AHPM, deeds of deposit. NOTE. The level 1–2 breaking point refers to the maximum amount spent by a buyer in Level 1. a The total amount paid for all purchases in this price range by the buyers in each level divided by the total amount spent by the buyers in the level.

vale real was worth 150 pesos face value, or some 2,260 reales. Sales concluded for less than this amount should have been paid in efectivo. Since the vales were depreciated, however, one finds properties going for around 2,000 reales being paid for with a vale real, but below this price, payment was almost exclusively in hard currency.^[1] For sales concluded after 16 August 1801, the deeds of deposit provide an additional item of

[1] The deeds covering payment in vales usually say that payment was made in vales reales "and the balance in hard currency" (y su preciso en efectivo ) but do not list how much in each medium. Jaén deed C9404 (1801) went for 1,605 reales paid with a vale real, the lowest in the two provinces.

information, the relation between the price paid in hard currency and the assessed value, because the king now recognized his debt to the former owner for the full assessed value even if the buyer had paid less in hard currency. Until November 1802 the king also gave the former owner a 25-percent bonus if the sale netted more in hard currency than the assessment. Because bidding in hard currency could begin at two-thirds of the assessed value, sales for the assessed value ("en su tasa") or more in hard currency could result only from competing bids and show us that in these cases demand was high. Conversely, payment in vales reales suggests that there was little bidding and relatively cheap acquisitions, for a bidder could forestall further bids in vales by offering hard currency. While payment in vales had to be at least for the full assessment, vales were usually depreciated by over one-third, making the "minimum"

Table 15.4.   Terms of Purchase, Jaén Province (percent of total purchases)   Level of Buyers   Termsa 1 (bottom) 2 3 4 (top) Total Not statedb   3.5 Hard currency (efectivo)   16.4 Hard currency below assessmentc   34.8 Hard currency at or above assessmentd   0.6 Vales reales   44.8 Total percent   99.9 Total amount  (000s rs.)   37,956 SOURCE. AHPM, deeds of deposit. a For meaning of different terms, see text. b Applied primarily to small purchases. One assumes that most were paid for in hard currency. c A comparison of the sale price with the assessed value in two years reveals that the following proportions of purchases in this category involved active bidding among the buyers (that is, the purchase price was greater than two-thirds the assessed value, the minimum bid allowed): 1802, 40 percent (N = 145); 1806, 20 percent (N = 164). d Purchases in this category could result only from active bidding.

bid actually lower than the minimum bid in efectivo (see Figure 6.1). What terms the different levels of buyers used in making payment thus tell us something about the conditions under which they acquired their properties.

Tables 15.4 and 15.5 show the results. The trends are strong in almost all categories. The higher the level, the greater the proportion of its purchases paid for with vales reales. To a certain extent, the trend is inevitable, since all purchases below about 2,000 reales had to be paid for in efectivo. However, the lowest level included all buyers whose total purchases reached 4,000 reales in Salamanca and 3,390 reales in Jaén. Even the first level thus was not ruled out from the use of vales. Sales recorded as paid in hard currency at or above the assessed value after August 1801 were concluded after active bidding, and so were a large proportion of those paid for in hard currency at a price below the assessed value. The lower levels made more of their purchases in this way. The wealthier bidders, in other words, faced competition less fre-

Table 15.5. Terms of Purchase, Salamanca Province (percent of total purchases)   Level of Buyers   Terms 1 (bottom) 2 3 4 (top) Total Not stated   5.4 Hard currency   (efectivo)   7.9 Hard currency below assessmenta   12.4 Hard currency at or above assessment   4.5 Vales reales   69.8 Total percent   100.0 Total amount (000s rs.)   36,418 SOURCE. AHPM, deeds of deposit. NOTE. See the notes to Table 15.4. a In 1802, 43 percent of the sales in this category were concluded after active bidding (N = 112); in 1806, 42 percent (N = 31).

Table 15.6. Terms of Purchases below the "Level 1–2 Breaking Point" (percent of purchases in price range)   Level of Buyers Terms 1 (bottom) 2 3 4 (top) Jaén Province    Not Stated      Hard Currency      Vales reales   Salamanca Province      Hard currency      below assessment      Vales reales   SOURCE. AHPM, deeds of deposit. NOTE. Other terms do not show marked trends.

quently, and they paid for their purchases more often in depreciated vales. Smaller buyers competed more actively against each other and got less favorable terms as a result. This is the case even when one considers only those purchases below the "Level 1–2 breaking point," although the larger buyers had to use hard currency more often for these small purchases than for their others (Table 15.6).

The difference between upper and lower levels is more marked in Salamanca than Jáen. In Salamanca the bigger and smaller buyers seem almost to have lived in distinct economic worlds. Level 4 buyers made four-fifths of their purchases for vales, Level 1 only one-fifth (Table 15.5). Conversely, Level 1 paid hard currency proportionately more than four times what Level 4 did. Big buyers here were into an economy based on paper money, received, most likely, for commodities sold on the national market or furnished to the crown for its armed forces. Smaller buyers lived in a world of cash on the line. They could not bid for the large properties, but the rules of the game meant that they could keep each other from using vales to buy church lands. They fought each other at the auctions, not their economic superiors.

In Jaén the trends were the same, but the differences were less marked.

Vales reales dominated the economy here less, only 45 percent of payments were made in this medium, compared to 70 percent in Salamanca (Table 15.4). The biggest buyers used vales six times more than the smallest, but the small buyers used hard currency proportionately only one and a half times as much as the large buyers.

When one considers entire provinces, as here, the range of economic means of the buyers in Salamanca appears much greater than that in Jaén. We saw earlier, however, that the distinction in types of property acquired by the different levels was more marked in Jaén. Putting both sets of evidence together, one can conclude that in both provinces there were two types of economy present, one involved in production for a nonlocal market (olives and wheat being most important harvests) and the other of a variety of products—fruits, vegetables, lesser grains, flax in Salamanca, grapes in Jaén, as well as wheat—for local consumption (self-sufficiency would be a misleading term). The market economy dealt primarily in vales reales, the local economy in hard currency. Production for the former minimized the input of labor, intensive farming was directed to the latter. In actual practice, almost every locality and most individuals participated in both economies, but each was located at a different point on a continuum that ran from the limiting case of all production for the nonlocal market to the other extreme of all for local consumption. At this time, no individual would have reached the first limit, but many small farmers (although probably not many buyers) may have been at the second limit, except that their payments for rent and tithes would enter the larger market. Disentail reinforced this spectrum.

The information on terms of payment permits one to distinguish also between purchases paid for at once and those paid for over several years. Time payments were not very advantageous; they had to be made in three installments: at the time of purchase and on the next two anniversaries, with interest on the last two. As a result this method was little used: 1.5 percent of the total payments in Jaén, 0.4 percent in Salamanca. Jaén shows a slight trend for smaller buyers to use this arrangement more,^[2] Salamanca no trend. If conceived to help the less affluent to compete for the properties, the measure had a negligible effect. Again one perceives that so far as the local rural world was concerned, Spain's was still primarily a cash economy.

[2] Level 1: 2.5 percent; 2: 1.8 percent; 3: 1.6 percent; 4: 1.3 percent.

4

Finally, the data in the deeds of deposit enable one to establish a limited social profile of the people involved in the two economic worlds. First they tell us the appellation of the buyers. Some were called don or doña, but the majority not. Since these were notarized documents, it is more than likely that the use of the title depended on a legal privilege or, if not a specific document, on a well-recognized public reputation. Of the persons who made more than one purchase, a few appear at times with the don and at others without it, but they are a minority. The consistency of usage for each individual is remarkable, especially when one recalls that these were résumés of the original documents and that the name of the buyer was of marginal interest.^[3] To call oneself don was not a matter of caprice but a recognition of a legal or at least semiofficial status. Among those who regularly appear with the title don are priests, who can be identified by the abbreviation pbro. (presbítero ) after their name. I have kept them as a separate category.

Table 15.7 indicates the proportion of dons and doñas and priests among the different levels of buyers in the two provinces. There are virtually no titled aristocrats (condes, duques, and marqueses) among the buyers (most aristocrats probably hid behind the name of an agent), but those that appear are grouped with the dons. (The previous tables have been based on the amount spent by the buyers; this one counts individuals, not the amounts they spent.) The table reveals very clear trends. In both provinces 70 percent of the buyers in Level 4 (highest) were entitled don or doña. With Level 1 buyers the trends run the other way: almost 90 percent without any qualification in Salamanca province, 70 percent in Jaén.

It is not easy to know who had the privilege of displaying the title don at the end of the old regime. Hidalgos (nobles) were distinguished in this fashion, and they were undoubtedly the most important part of the group. But it is certain that not all who enjoyed the appellation were legal nobles. The census of 1797 shows 785 nobles in the province of Jaén, 7 percent fewer than the buyers identified as don; and in the portion of Salamanca under study, some 441, 8 percent fewer than the buyers.^[4] Since it is inconceivable that all hidalgos bought properties,

[3] See Appendix O.

[4] The category "noble" of the census of 1797 is equivalent to hidalgo; one can tell because the census of 1787 uses the term "hidalgos," while the census of 1797, in repeating the totals of 1787, calls them "nobles." For the province of Salamanca the census of 1787 gives a total of 567 hidalgos. In the region studied here, that is, without the partidoof Ciudad Rodrigo, the individual town returns of the census give a total of 533 hidalgos, 94 percent of the provincial total. The census of 1797 shows 470 nobles in Salamanca; the same proportion gives 441 in the region studied. (For the Censo español . . . de 1787 and the Censo de . . . 1797 , see Appendix A. The individual returns are in Real Academia de la Historia, legajos 9-30-2, 6240 to 6242 and 6259 [the last for the city of Salamanca].)

Table 15.7. Estate of Buyers, Jaén and Salamanca Provinces   Level of Buyers     Estate 1(bottom) 2 3 4 (top) Total N (percent of individuals) Jaén Province       Commoners   57.5 1,467a Don and doña   39.4 1,005 Priests   1.9 48 Otherb   1.2 31 Total percent   100.0   Total N     2,551 Salamanca Province       Commoners   71.5 1,535a Don and doña   25.4 545 Priests   2.5 54 Otherb   0.6 13 Total percent   100.0   Total N     2,147 SOURCE. AHPM, deeds of deposit. a Includes "unnamed" buyers: Jaén, 67; Salamanca, 233 (see text). b In Jaén, twenty-six religious institutions, five civil institutions; in Salamanca, thirteen religious institutions.

there must have been many buyers called don who were not included as nobles in the census. Who were they?

It is very likely that the census takers underreported the number of hidalgos. The preface to the census of 1787 says that pressure was applied to eliminate from the number of hidalgos the persons who were not entitled to the privilege, and the number in the census of 1797 is

even less. This last gives 14 percent fewer hidalgos for the province of Jaén than the census of 1768 does for the bishopric of Jaén, and for the province of Salamanca 12 percent fewer than the bishoprics of Salamanca and Ciudad Rodrigo in 1768. Since three partidos of the province—Barco de Ávila, Piedrahita, and El Mirón—were in the bishopric of Ávila, the decline in Salamanca between 1768 and 1797 was even greater.^[5] It is probable that the censuses of 1787 and 1797 counted fewer hidalgos than the persons who were reputed to be hidalgos and that if these persons bought disentailed properties, they had themselves listed as don in the deeds.

This could account in part for the number of buyers called don, but more can be explained by other categories of men addressed as don because of their vocation, whether or not they were hidalgos. Such were the priests, as already noted. The deeds reveal that high servants of the crown, officers of the armed forces, and notaries appear as don (they could, of course, also be hidalgos, as were Campomanes and Jovellanos). In the various town catastros, men with official and semiofficial positions enjoyed the title: doctors, surgeons, apothecaries, and administrators of the tobacco monopoly. The census of 1797 lists 404 notaries, doctors, surgeons, and apothecaries in Jaén (against some 785 nobles) and some 450 in our portion of Salamanca (against 441 nobles). Nevertheless, the royal government strove to limit the use of the appellation. A decree of 1775 allowed notaries (escribanos) the right to sign with the title don only if they were hidalgos.^[6] Our evidence suggests that the practice was more liberal than the letter of the law, but the consistency in the use of the appellation don in the deeds of deposit indicates that it was applied only to persons with a recognized right to it.

We can conclude that the buyers who disported the don or doña had a proper claim to the title. They formed an elite of "notables," of which the true hidalgos were the most important (and probably the most numerous) part, the top stratum of the society of the old regime, not counting the exiguous high aristocracy. Table 15.7 shows that these notables dominated the top level of buyers in both provinces, the people who could spend fortunes for church properties.

This conclusion ill fits the stock hidalgo of literature and history, the threadbare and deluded Quijote. The Cervantine image was alive and

[5] The census of 1768 appears at the beginning of the published Censo . . . de 1787 .

[6] Decreto del Consejo (14 Sept. 1775), Nov. rec., VII, xv, n. 11. A RC of 19 May 1801 established, among other imposts to support the Consolidation Fund, a fee of 550 reales "for the authorization for notaries [escribanos] who are in possession of nobility to sign themselves don" (ibid., n. 12, and in more detail in AHN, Hac., libro 8053, ff. 195–200).

well in the mentality of the eighteenth-century reformers, recognizable in the scornful portrait of the rural noble painted by José Cadalso: "He takes the air majestically in the sad square of his impoverished village, wrapped in his poor cape, admiring the coat of arms above the door of his half-ruined house, giving thanks to God for having created him don Fulano de Tal."^[7] A picture that twentieth-century historians have not abandoned,^[8] but our provincewide analysis of the disentail brings its accuracy into question. It is not the first such piece of evidence; the reconstruction of the society of Lopera and Baños in Part 2 already revealed the economic and political power of the local hidalguía. We shall return to the issue when we look at the big buyers in Chapter 19.

Priests followed the same trend as the dons, toward the higher levels of buyers. Not all priests can be identified, because the qualification pbro. was not always added in the deeds,^[9] so that the trend toward the larger buyers may be exaggerated, but it is not spurious. Part 2 has shown that in the towns of Salamanca the income of the priests was among the highest in the parishes, that in those of Jaén the priests belonged to the best families. In Chapter 19 we shall see another, higher clerical world represented among the buyers, that of cathedral canons, capellanes, and university professors. Their means permitted a significant number of them to make purchases that placed them among the upper levels of buyers.

One can also identify the gender of the buyers from the deeds, in this case relying on their first names and the title doña (Table 15.8). If one considers all women as a group, one finds no trend in the proportion of buyers who were women from level to level in either province, but when the women are divided between those who were called doña and the others, the same trends become apparent as among the men, doñas being more common in the top levels and commoners at the bottom. That is, family and social level, not gender, determined the economic behavior of women. (Because some of the "unnamed" buyers in Table 15.7 would have been women, the correct trend for commoner women would be toward an even stronger presence in the lower levels that Table 15.8 shows, especially in Salamanca.)

With this added information one can give a social definition to the economic spectrum observed earlier. The persons who engaged in the commercial agricultural economy, those who had developed a capitalist

[7] José Cadalso, Cartas marruecas, Carta 38, quoted in Domínguez Ortiz, Sociedad española, 116. Fulano de Tal is the Spanish equivalent of "so-and-so."

[8] Callahan, Honor, Commerce, and Industry, 8.

[9] See Appendix O.

Table 15.8. Women Buyers, Jaén and Salamanca Provinces   Level of Buyers     Estate of Women 1 (bottom) 2 3 4 (top) Total N (percent of individuals) Jaén Province       Commoners   2.3 59 Doña   5.2 133 Total women   7.5 192 Salamanca Province       Commoners   2.3 49 Doña   2.1 44 Total women   4.3 93 SOURCE. AHPM, deeds of deposit.

mentality, were typically (but not exclusively) hidalgos and priests and other notables with official positions. The other economy, of production for local consumption, belonged primarily to commoners. The deeds of deposit tell one nothing about their occupation, but the largest number would have been peasants, like the vecinos of La Mata and El Mirón or the labradores peujaleros of Las Navas. Others may have been muleteers or bakers or craftsmen. When we come to study the big buyers, we shall find persons who may be properly classified as bourgeois—administrators of estates and merchants—in the upper levels. One may suppose that others of this kind are unidentified in the lower levels; after all, even the lowest buyers would have had some kind of standing in their towns. But the bourgeois would have been a minority in all levels.

The cases of Jaén and Salamanca indicate that Andalusia and Castile at the end of the old regime had a dual agricultural economy and a corresponding dual society. At one extreme was a peasant society producing primarily for local consumption; marked by high labor inputs; relying on hard currency (when not on barter), which the peasants saved carefully and brought forth to bid against their neighbors when given the chance to buy plots of land of their own. For such buyers, the disentail offered the opportunity to acquire the lands they farmed. At the

other extreme, the notables and a lesser number of wealthy commoners controlled the market economy; sought to minimize the costs of labor on their lands; and accumulated depreciated vales reales, which they could use to buy those disentailed properties that could expand their commercial output. Though functionally and socially different, the two economies were not physically separate. Most individuals connected with agriculture engaged in both, to a greater or lesser extent. In the towns of Jaén those who concentrated on the market lived cheek by jowl with the others; in Salamanca, the commercially oriented were more likely absentees residing in the provincial capital or the cabezas de partido. Such is the general picture that emerges from the provincewide analysis of the deeds of deposit. To refine it, we can turn to the variations among the geographic zones within the provinces.

Chapter XVI—
Jaén Province:
The Determinants of Growth

In order to proceed, I have divided the two provinces into a number of zones that I have sought to make geographically and economically homogeneous. They have been drawn on the basis of maps and studies of the physical landscape and geology and extensive personal travels throughout the two provinces by automobile. The purpose is to establish units with different social and agrarian patterns that can be used for a comparative analysis of their evolution.

The division of the province of Jaén into zones is not obvious. Except in the mountains, the region has a low rainfall, weather stations reporting a mean annual precipitation ranging from 487 millimeters at Cazalilla near the Guadalquivir River to 601 millimeters at Jaén. Conversely, evaporation is high everywhere. There is not much difference in annual mean temperature either, ranging from 16° to 18° centigrade. Climate, in other words, is relatively similar throughout the study region, with a cool, moist winter and spring, and a hot, dry summer.^[1]

In the eighteenth century and still today, agriculture predominated over grazing, even in the mountain valleys; and agriculture was, and is, marked everywhere by a division between wheat fields and olive groves. The marls that predominate are excellent for olives but less favorable to

[1] Information on the geography and climate of Jaén province comes from Higueras Arnal, Alto Guadalquivir, and Lázaro, Elias, and Nieves, Regímenes de humedades. In addition, I was given invaluable help by don José María Ontañon and doña María del Carmen Cid of the Centro de Estudio Hidrográfico, Madrid, whom I consulted at length and who traced maps of the soils of Jaén for me.

the cultivation of wheat, and over the centuries olives have been gradually replacing wheat. Today there are districts where the rolling hillsides of red or brown earth are dotted as far as one can see with silvery green olive trees set out in precise geometric patterns, columns of plumed soldiers marching up hill and down. In few places of the world has reason tamed nature so successfully. Yet, lest we forget that this is Andalusia, the two or three trunks that make each tree, taunting imposed discipline, become circling dancers with arms flung back in their own flamenco ritual. But go on a few kilometers and the unbroken rows of trees give way to rectangles of alternating olive grove and grain fields, the latter verdant and waving in the spring, sprinkled with brilliant flowers, then brown and parched after the June harvest.

There are few level plains, for the basin of the Guadalquivir undulates in waves of different lengths, heights, and directions. On reaching the southern ranges, the land rises sharply, and a short way up the olive trees give way before thin soil or barren limestone rock, the stark mountains wrapped in flowing green polka-dotted skirts, Gypsy maidens decked out for a perpetual fiesta of the local Virgin. The richness of Jaén's rolling basin, with its broken pattern of olive groves, wheat fields, and stark white towns, the towering ranges on the south and east, the wild, wooded Sierra Morena on the north, make the province a constantly varied vision of beauty.

As is common in southern Spain, the inhabitants of Jaén are gathered into large settlements. In the census of 1786, Jaén had seventy-four cities and towns, with an average population of 2,380. Both for military defense and to escape the heat of summer and the danger of epidemics, most towns of Jaén stand white and nucleated at high points in their territories. Baños is a good example. In the Guadalquivir basin many sit on hilltops, and at the southern and eastern edges of the basin they are located up the base of the mountains, with their fields stretching out beneath them and the barren slopes above serving as protection from attack. The capital city, Jaén, at the foot of the Sierra de Jabalcuz, is a perfect example, dominated by the Castillo de Santa Catalina, which commands the two approaches from Granada, the older western road that circles the higher ranges and the newer road south through the sierra. On the northern limits of the province, the ranges are lower and several towns nestle in saddles where roads pass through the hills, as does Navas.

Close by the towns, where water flows either from streams or wells, lie the irrigated huertas with their vegetable plots and fruit trees. In the

eighteenth century, and still today, the closest land to the nucleus was called the ruedo, the most intensively cultivated zone that is not irrigated. Beyond is the campiña of the town, not to be confused with the term campiña applied generally to the basin of the Guadalquivir as opposed to the sierra. In the empty spaces between the towns of the basin there were, and are, scattered residences and farm buildings of the large estates, the cortijos that roused the ire of Carlos III's reformers and of those of the Second Republic and the casas de campo in the major olive groves. In the case of Baños, we saw that this land use reflects the pattern known as von Thünen's rings, with the intensity of cultivation depending not only on quality of soil and availability of water but on the distance from the town nuclei.

I divided the province into nineteen zones (Map 16.1). Five are located fully in the valley of the Guadalquivir, in the western part of the province. Three have rich soil, centering on the cities of Andújar (Zone JA); Arjona (JB), which includes the town of Lopera, studied in Part 2; and Linares (JE). In the ensuing analysis, they are referred to as rich basin zones. The two other zones in this region are further south, in a rolling plain where salt deposits have reduced the fertility of the soil. Santiago de Calatrava zone (JC), made up of three distinct blocks because of the way the municipal boundaries run, is near the Córdoba frontier, while Mengíbar zone (JD) is in the center of the province. They are called poor basin zones. South and east of this region are four zones whose towns have their lands in the valley of the Guadalquivir and their town nuclei against the foothills of the southern and eastern ranges. Identified here as south basin zones, they are Martos (JF), Jaén (JG), Mancha Real (JH), and Cazorla (JI). The Sierra Morena zone, JJ (Baños), is in the foothills of this range and is made up of the town we studied and a similar one called Vilches. Two rolling ridges or hog's backs (lomas ) rising between rivers in the northeastern part of the province are the location of four zones, the two to the west rich loma zones: Baeza (JK) and Ubeda (JL); and the other two poor loma zones: Santisteban (JN), which includes the town of Las Navas of Part 2, and Villanueva (JM). Finally five zones are located south of the basin of the Guadalquivir. Three of them are sierra zones, Bédmar (JO), Pegalajar (JP), and Cambil (JQ); and two lie in broad valleys in the south, isolated from the rest of the province, Alcalá la Real (JR) and Huelma (JS), the southern valley zones. Appendix P provides a detailed description of the geography, population structure, and political status of these zones.

Map 16.1.
Zones of Jaén Province

2

Few readers will have been surprised to learn from the previous chapter that two types of economy existed in the provinces—market and local—and that they can be associated with different social groups, although the prominent role of nobles and clergy in commercial agriculture may seem paradoxical at first. If we turn to the data for the individual zones within the provinces, we can refine this simple observation, for the zones show that the different factors associated with the two types of economy come together in a variety of ways to produce distinct

subpatterns, which in turn help us to understand the economic forces at work in rural Spain at the end of the old regime.

When added to the deeds of deposit, the midcentury catastro and various censuses provide a diachronic perspective. The detailed information of the catastro on individual property holding in each town, which made possible the studies in Part 2, is unmanageable at the provincial level, at least for this study. The royal bureaucracy in the eighteenth century, however, extracted much of the information from the reports and their results have fortunately been preserved in most cases. The Archivo General de Simancas has a precise listing of the property of the largest owner in each town, the hacendado mayor, including his, her, or its (if it was an institution) name and residence. Simancas also has a copy of the responses to the forty questions that formed the first part of each town survey, with data on the number of vecinos, houses, people in various occupations, wages, prices, types of harvest, tithes, taxes, and similar matters ("Respuestas generales" or "Interrogatorio general"). Even these questionnaires are not readily usable in extenso—they may run to fifty folio sheets for one town—but their contents were extracted and tabulated in large volumes now in the Archivo Histórico Nacional. This series normally has three volumes for each province, the first ("Estado seglar") devoted to the property and other income of laymen and secular institutions (including the town council, that is, town property); the second ("Estado eclesiástico") for that of ecclesiastical institutions, foundations, and funds; and the third ("Eclesiástico patrimonial") for that of individual clergymen. For purposes of the present study, I have assembled these data, already quantified town by town, into totals for the zones within the province. Unfortunately, for Jaén only the volume summarizing the information of the secular estate has survived.^[2]

A number of eighteenth-century censuses have been preserved that give the population (vecinos or total inhabitants) of individual towns. The "Vecindario general de España," located in the Biblioteca Nacional and already described in Chapter 1, gives the vecinos around 1712. The number of vecinos at the time of the catastro (ca. 1752) can be obtained from the respuestas generales, question 21, but it has also been tabu-

[2] For most towns the information is drawn from AHPJ, Catastro, maest. segl. The volumes with this information are missing for the following towns, and I obtained it from the copy in AGS, Dirección General de Rentas, Única Contribución, libros 323–27: Albánchez, Alcalá la Real, Cabra de Santo Cristo, Castillo de Locubín, Fuente del Rey, Higuera de Calatrava, Jabalquinto, La Guardia, Los Villares, Lupión, Marmolejo, Tobaruela, Torre del Campo, and Torrequebradilla.

lated from the towns of Salamanca province in a document in the Real Academia de la Historia.^[3] The same place houses the original individual town returns of the census taken in 1786 and published the next year. These record the number of each sex, single, married, and widowed, in each of six age groups, as well as the number of individuals in various legal estates and occupations.^[4] Finally, I have made use of the number of vecinos and inhabitants recorded for each town in 1826 by the royal geographer Sebastían de Miñano.

The information on the sales has been described in the last chapter. It is now, however, broken down by zone, and within each zone by level of buyers.^[5] The level of each buyer is determined by his location on the Lorenz curve of the zone.^[6] Thus in a zone where purchases were priced low, a buyer who appeared in Level 3 or even 2 in the province may be placed in Level 4, and vice versa. Buyers in Level 4, or any other level, do not spend equal amounts of money from zone to zone and are not equally wealthy. Table 16.1 gives a sense of the difference among the zones in Jaén province in this respect. It shows the mean purchases of Level 1 and Level 4 buyers and of all buyers in each zone. If one compares the mean amount spent by all buyers in each zone, as shown in this table, with the mean town population in each zone (Table 16.2), one finds a rough, direct relationship (r = .52). One can deduce that, as a general rule, in larger towns, the upper levels of society, those that furnished the buyers of disentailed properties, had more money than the corresponding levels in smaller towns. The correspondence is far from exact, however, r² = .275, meaning that only about one-fourth of the variation in the amount spent can be attributed to the size of the towns in the zones.

By using the commercial market–local consumption spectrum that emerged from the global look at the two provinces, one may distinguish the zones in Jaén province by the strength of the orientation of their agriculture toward a wider market, as revealed by the nature of the prop-

[3] Real Academia de la Historia, leg. 9-30-3, 6258, no. 13. It is dated 30 Jan. 1760, but it comes from the returns of the catastro. Ibid., no. 14, is a later census of vecinos dated 14 Aug. 1772. I entered it in my analysis, but it proved of little use.

[4] Ibid., legajos 9-30-2, 6228 and 6240–42; 9-30-3, 6259.

[5] All sales but one (in an isolated town not included in any zone) could be assigned to zones. Where the location of the property was not specified in the deed of deposit, I assigned the sale to the zone in which the religious institution that was the former owner was located. No doubt some lay elsewhere, but not enough to invalidate the analysis. The number of sales in each zone is JA, 186; JB, 147; JC, 22; JD, 28; JE, 179; JF, 283; JG, 911; JH, 289; JI, 263; JJ, 116; JK, 762; JL, 482; JM, 273; JN, 208; JO, 47; JP, 160; JQ, 59; JR, 146; JS, 80; Total 4,641.

[6] See Chapter 15, section 1, esp. Fig. 15.1.

erties disentailed. For instance, although some olive oil was consumed locally, one can posit that most owners and buyers of olive groves were conscious of a national and even an international market and involved in it. Unfortunately, arable land, the largest type of purchase, involving 32 percent of the money spent, could produce wheat and other grains either for local consumption or for the broader market, so that the amount of it disentailed in each zone does not help one distinguish the economy of the zone. On the other hand, cortijos and other rural estates, which can be identified, did produce primarily for the broader market. Table 16.3 and Map 16.2 distinguish the zones by the percent of the total purchases devoted to olive groves and to rural estates. The higher the readings, the more the zone can be assumed to produce for the market economy.

Olive groves are strong in the north and west of the province, cortijos in the south and east. The only zones with strength in both types of cultivation are Andújar (JA), bordering Córdoba province on the Guadalquivir River, and Mancha Real (JH), between the Guadalquivir and the southern ranges. The two zones in the northeast (JM, JN) and a belt of

zones in the center and south (including the zone of Jaén city) do not show up as high in either type of clearly commercial property (JD, JF, JG, JK, JP, JQ, JR).

To understand better the economy of the various zones, one can distinguish them according to the orientation toward the market of the different levels of buyers. The purchases of each level show if it was primarily interested in buying properties directed toward market production and was able to acquire them. I shall divide the zones into three

categories according to these criteria, using the premise that olive groves and cortijos were engaged more in market production than other types of property. Some grain grown in small plots was also destined for the outside market, of course, but since much of it would be for local consumption, the marketed proportion of the grain not grown in cortijos would be less than that of the cortijos. Vines would also be directed toward an outside market, but most fruits and vegetables would be too perishable.

The three categories of zones are as follows:

Type A. All levels of buyers into the market. Those zones where both Level 4 and Level 1 buyers directed their purchases toward market-oriented properties. To qualify for this type, Level 4 buyers (highest) must devote the largest share of their purchases to olive groves or rural estates or both together, and Level 1 buyers (lowest) must either devote a larger share of their purchases to olive groves than to arable or must

Map 16.2.
Jaén Province, Rural Estates and Olive Groves in Disentail, 1799–1807

have olive groves as their next choice after arable. The second alternative for Level 1 buyers is necessitated by the fact that in most zones sales of arable outweighed considerably sales of olive groves, so that Level 1 buyers, even if they preferred olive groves, had more arable available to purchase. Zones where Level 1 buyers preferred urban properties, improved or irrigated plots, or the redemption of censos to olive groves do not qualify for this type.^[7]

Type B. Top levels of buyers into the market. Those zones where Level 4 is into the market but Level 1 is not, under the terms defined above. These are primarily zones strong in cortijos, where Level 1

[7] In Zone JO, Level 1 preferred improved or irrigated plots to olive groves (24 percent to 21 percent), but they also bought many vineyards (14 percent). Since vineyards produced for the market, I have included this zone in this category. (It is impossible for Level 1 buyers to have rural estates as their preferred type of purchase, since most of these were cortijos, whose price was greater than the total amount spent by someone in Level 1, and thus rural estates are not used as a criterion for considering the market orientation of these buyers.)

buyers did not have available olive groves that would put them into the market.^[8]

Type C. No levels of buyers into the market. Those zones where neither Level 4 nor Level 1 is into the market as defined above. Although these zones obviously sold some produce on the outside market, the nature of the properties that were disentailed indicates that this commerce was not conducted on a major scale and was not the primary objective of local agriculture.^[9]

Table 16.4 shows the distribution of the zones into these three types.

3

One may now proceed to compare these different types of zones on the basis of other known characteristics. For instance, the mean amount spent by the buyers in each zone proves to be related to the extent that the buyers were into the market. If one ranks the zones according to the mean amount spent and draws a line between numbers 11 and 12, as is done in Table 16.5, one finds that seven of the nine zones in Type A (all levels into the market) and four of five zones in Type B (top levels into the market) are above this line. All the zones in Type C (no levels into the market) are below the line. Zones JH and JO, belonging to Type A, fall below the line. The considerably smaller mean amount spent by their buyers than those of other zones in Type A indicates that these two zones had a distinct economic level and deserve to be grouped separately. I have therefore divided Type A into Types A₁ and A₂ . The same is true for Zone JQ in Type B, which becomes the only zone in Type B₂ . These subcategories are already incorporated in Table 16.4. Henceforth most of the discussion will exclude Type A₂ and Type B₂ zones, which are too few in number to draw conclusions about the significance of their characteristics. With these exceptions, being into the wider market, whether by all levels of buyers or by the top levels, meant that the properties sold in the zone were more expensive and the average buyer spent more than in the zones that were not into the market.

[8] Zone JP shows a high demand among Level 1 buyers for improved or irrigated plots (23 percent). Some of these were probably devoted to market gardening for nearby Jaén city, but I do not count this as production for a wider or national market.

[9] This categorization ignores the purchases of Levels 2 and 3. In most cases Level 2 preferences reflect Level 1, while Level 3 is closer to Level 4. Thus Type A is properly called "All levels of buyers market oriented" and Type B "Top levels market oriented."

Map 16.3 shows that these types reflect geographic affinities. Except for the two unclassified poor zones, JC and JD, marked by infertile soil, small towns, and few purchases, most of the central and west of the Guadalquivir valley consists of Type A zones (all levels into market). Type B zones (top levels into market) form a ring around this block on the east and south, while Type C (no levels into market) are peripheral in the northeast and southwest. Only one Type A zone is not contiguous to the others, JR (Alcalá la Real), made up by the three rich valleys in the southwest. Type A₂ zones (all levels into the market, small purchases) lie next to each other on the eastern edge of the A zones, and the B₂ zone (top levels into the market, small purchases) is nearby, in the long valleys above Jaén city.

Because of the way in which being into the market is determined, with the bottom levels depending on the purchases of olive groves and the top levels on either olive groves or rural estates, one would expect Map 16.3 to reflect closely Map 16.2, which shows the zones of heavy purchases of olive groves and rural estates. There is a similarity but by no means an identity. All the zones that had a large proportion of their

Map 16.3.
Jaén Province, Types of Zones

sales in olive groves do, in fact, fall into Type A (except Zone JC, not classified for lack of sufficient sales), but so do four other zones that were not strong in olives. In three of these, the small buyers sought the available olive groves within their means, while the upper levels went for rural estates. In the fourth, JO (a Type A₂ zone), Level 1 got into the market by buying vineyards as well as olive groves. Three of the six zones strong in rural estates thus fall into Type A, leaving only three others for Type B, along with two others not strong in rural estates. Those zones that had a large share of their sales in either rural estates or olive groves belong to one of the types that were into the market—and quite properly so—but by concentrating their purchases on the available rural estates (JP) or olive groves (JQ), the upper level of buyers placed two other zones among those where the upper levels were into the market.

One may question whether the division of the zones into these three types in reality represents different orientations toward the wider mar-

ket. It could, for instance, be simply a reflection of the size of their towns, with Type A zones containing the largest towns and Type C the smallest. (One recalls that we have found a rough correlation between the mean size of the towns and the mean amount spent by the purchasers.) If one treats these three types of zones as random samples drawn from an imaginary larger population of zones with similar characteristics and applies a test of significance to the differences in the size of their towns in 1786, one finds no significant relation between town size and the three types. The mean town size has the expected trend, but the difference in town size varies so widely within each type of zone that the difference is not significant.^[10] The three types of zones thus reflect something other than town size.

Another possibility would be that the types offer a measure of general economic activity and not simply the market orientation of agriculture. To explore this possibility, one must go back fifty years to the information in the catastro. Table 16.6 shows the proportion of artisans and craftsmen among the vecinos in each zone in the early 1750s. On the surface, the three types of zones do seem related to artisan production. The mean number of all craftsmen per thousand vecinos declines from Type A₁ , to B₁ , to C and, disaggregating the crafts, so do the numbers of carders and weavers and of tailors. On the other hand, the number of sandalmakers (alpargateros) per thousand vecinos is highest in Type C zones and lowest in Type A₁ , responding perhaps to the greater isolation of Type C towns. Tests show that these trends are not significant, however, because of the wide variation among the readings within each type of zone (again treating the three types of zones as if they were random samples drawn from a larger population). The same is true of the income of the retail food stores per vecino, although here there is a hint of a possible relationship.^[11]

In fact, other factors can be more readily associated with at least two types of crafts. The concentrations of carders and weavers, for instance, are found in the zones with the largest mean town population, and indeed, they are located in the large towns. Jaén, Ubeda, Baeza, Alcalá la Real, and Andújar, the only municipalities in the province with over two thousand vecinos, had 109 of the 115 carders in the province and 148 of the 156 weavers. Alpargateros also have a different association.

[10] Mean town population: Type A = 3,230, B = 2,960, C = 2,680. Despite the apparent trend, the null hypothesis, that all three types come from the same parent population, is not disproved by an analysis of variation, whether one uses the mean town population for each zone, or the individual population readings of all the towns.

[11] An analysis of variation shows F = 1.75. For a 20 percent probability, F = 1.87.

The high concentrations (over ten per thousand vecinos) all occur in the four loma zones and Cazorla, contiguous in the east. Theirs was a regional specialty.

On the other hand, one kind of activity recorded in the catastro correlates well with the three types of zones. This is the income of muleteers (arrieros). Table 16.7 shows the income per vecino from muleteering in the zones, and Map 16.4 locates the towns that reported income from

muleteering. Type A₁ zones have a mean per vecino income from arrieros of 19.5 reales; Type B₁ , 5.8; Type C, 0.9. An analysis of variation shows that there is less than a 5-percent probability that these differences occurred by chance. Even more conclusive is a comparison of the number of towns in each type of zone that had a major sector of muleteering (here defined as having a minimum of one thousand reales total annual income for all the arrieros in the town). Eighteen of twenty-seven towns in Type A₁ , zones reach this figure, three of twelve in Type

Map 16.4.
Jaén Province, Income from Muleteering, ca. 1750

B₁ , and only one of eleven in Type C. The difference in proportions is significant at the 1-percent level.

Of course, active muleteering could be independent of the nature of local agricultural production, as is shown by zone JQ, the deviant case labeled Type B₂ . It had the largest arriero income per vecino of all the zones. It lay on the new road from Madrid to Granada through the sharp valleys south of Jaén city, and all its arriero income came from the two towns nearest the border with Granada, Campillo de Arenas and Noalejo. Relatively small mountain towns, they specialized in transport independently of their agricultural activity. Zone JD was a similar

case, second in income per vecino from arrieros only to JQ. All of its arrieros were located in one town, Mengíbar, site of the main ford across the Guadalquivir River on the way south to Jaén and Granada. A small town, it was nevertheless the largest of five in this poor valley zone and prospered from its location. Surrounded by zones actively into the market (Type A₁ ), it stood in a natural spot to develop muleteering.^[12] Elsewhere the variation in the amount of muleteering income among the three types of zones lends support to the proposition that the types do in fact represent different orientations toward commercial agriculture. Although positive proof is lacking, all the information adduced leads to the conclusion that the three types of zones differed in the market orientation of their agriculture but not significantly in other forms of economic activity except muleteering, which can be seen as a forward linkage of commercial agriculture.

4

Having determined the extent of the involvement in the market of the agriculture of the different zones, we may proceed by observing how this factor is related to their social structure. To begin with, I shall use the information provided by the sales during the disentail. One statistic is especially revealing: the proportion of the buyers who were notables (as defined in Chapter 15). Table 16.8 gives these figures for Levels 1 and 4 and all buyers, broken down by type of zone. Among the three types of zones, there is a significant difference in the proportion of buyers who were notables. Zones with all levels of buyers into the market (Type A₁ ) had the highest proportion (46 percent), zones not into the market (Type C) the lowest proportion (30 percent), while those where only the top levels of buyers were into the market were in between (35 percent). This trend appears also among Level 1 (small) buyers, but not to a statistically significant extent among Level 4 (large) buyers. The previous chapter showed that there were many more notables among the top buyers in the province as a whole, and the table shows that to be the case in the individual zones (except the zones not included in the three types because they had few buyers). Not here, however, but in the lower levels does a clear pattern emerge of the proportion of buyers who were notables directly correlated to the extent of involvement in the wider market.

[12] For its location at the main ford, see Biblioteca Nacional, Tomás López, "Atlas particular," and Laborde, View of Spain 2 : 108.

The census of 1786 reveals that the pattern was not peculiar to buyers alone but characterized the population as a whole. The zones where all levels of buyers were into the market (A₁ ) were those with the largest proportion of nobles among the residents. The census called for an enumeration of the hidalgos in each town, a piece of information not included in the questionnaire or provincial summaries of the catastro. Table 16.9 shows the number of hidalgos per thousand adult males in each zone. In many respects the accuracy of the census for Jaén is suspect, as already noted in the case of Baños, so that the number of hidalgos reported may not be correct. The trend is clear, however. The ratio of nobles to adult males is distinctly higher in all but two Type A₁ zones than in Type B₁ , zones (the two deviant A₁ , zones are contiguous in

the north center). Two of the Type C zones show a lower proportion of hidalgos than the large majority of Types A₁ and B₁ zones, but the third, JM in the northeast, has the highest ratio of all zones. (The high count is in the two newer towns Villacarrillo and Villanueva del Arzobispo, the other older, pre-conquest town Iznatoraf reported only one noble. One cannot tell from the data whether the deviance of JM is real or the result of an enthusiastic census taker.) Except for these three cases the pattern reflects the picture drawn from the sales: the more fully a zone was involved in the outside market for agricultural products, the larger was its class of hidalgos.

This information permits us to look again at the conclusion reached in the last chapter about the relation between social class and commercial agriculture. The data in Tables 16.8 and 16.9 indicate an apparent paradox. The zones of local economy (Type C), where the average amount spent by the buyers was low, mostly located on the periphery of the province, had the fewest nobles and the most distinctly stratified group of buyers (small buyers having a far smaller proportion of nobles than large buyers). By contrast, the wealthy zones, where all levels of buyers were into the market and the average amount spent by buyers was high, had the most nobles and reveal the least stratification among

the buyers. The percentage of nobles is higher and more evenly distributed among the levels. Zones in which only the top levels of buyers were into the market lie in between.

If one recalls from Part 2 the contrast between Baños and Lopera on one hand (both located in zones where all levels of buyers were into the market, JJ and JB), and Navas on the other (zone JN, no levels into the market), one can posit a conclusion. The first two had a wealthy resident elite headed by a number of hidalgo families. Their elite formed a distinct layer at the top of the social pyramid (8 percent of the vecinos in Baños, 10 percent in Lopera, see Tables 11.20 and 12.14), which ran their towns and extracted wealth from the land without getting their hands dirty. But in Navas there was only one landowning don out of 214 male vecinos, the notary. The information provided by the present chapter indicates that this contrast can be generalized to the entire province. One can conclude that there were two distinct typologies of social structure in Jaén province. Towns in zones actively engaged in commercial agriculture attracted more wealth and developed a strong resident economic, social, and political elite of which hidalgos formed a major part. Nonnobles were a vast majority of the vecinos, but among the buyers they were a bare majority (a minority in two zones), and even at the lowest level of buyers they were not an overwhelming majority. The conclusion of the last chapter is reinforced: the presence of commercial agriculture and a strong resident hidalgo class went hand in hand.

At the other extreme, areas involved primarily in a local economy did not produce or attract a resident elite of this nature. A few hidalgos were present and formed the apex of the social pyramid. When they bought disentailed lands, they sought the best properties, having the most money to spend, and dominated the top level of buyers, where they might be joined by outsiders. But they left to local commoners the largest amount of land put on sale. Although there was greater disparity in the composition of the buyers, this reflected a smaller and weaker, not a stronger, local elite.

Type B zones, where the top levels of buyers were into the market, are a halfway stage between the two extremes. Disentailed property here was split more evenly between what was commercially oriented and what was produced for local consumption. The elite of notables, smaller than in Type A zones, stressed the purchase of market-oriented properties, leaving the less expensive properties to the noncommercialized, nonnotable sector of the towns.

5

This analysis of the relationship between the different types of agricultural activity and social structure offers a static picture of the rural world of Jaén at the end of the old regime. Our next task will be to make the picture move, to demonstrate evolution during the half century under study and discover the forces at work. The data will not always permit the sequence to be fully in focus, but the picture will be instructive, nonetheless.

One may start by determining in which zones the market economy was expanding. For this purpose, one would like to compare the proportion of land devoted to harvests for the market at the time of the catastro with that shown by the sales, but the data available do not permit so precise a comparison. The catastro of each town states the area devoted to each type of cultivation and the value of the annual harvest,^[13] while the records of the sales give the price but not always the size of the ecclesiastical properties that were sold. The extent of land devoted to the various harvests at the time of the sales is unknown, but from the available data one can compare the value of land devoted to cereals and olive groves at the time of the catastro with that shown by the sales. If the ratio of olive groves to arable shows an increase between the two dates, one can assume increasing production of olive oil for the market. There is unfortunately no way of determining an absolute increase in grain production for the market, but because the demand for olive oil was outdistancing that for grain, as will be explained below, greater participation in the market was likely to show up as an increasing ratio of olive groves to arable.

Even here there are difficulties. One saw in Chapter 5 that the ratio of catastro value to sales price varied widely among the properties devoted to different cultivations. One must adjust for the different markup of the two types of property when comparing figures from the two sources. Table 16.10 shows the results.

My comparison of the figures (Column D of Table 16.10, which measures the change in olive tree cultivation vis-à-vis arable in each zone) is based on three assumptions, all of them open to some error: (1) that the sales represent a true cross section of the cultivated land in each zone between 1799 and 1807; (2) that the ratio of the markup of olive groves

[13] This information is in the introduction to the survey of property owed by laymen in each town, the libro maestro seglar. Where maest. segl. is lost, I used the earlier estimate in AGS, resp. gen. QQ 10, 12, 13, 14 (see n. 2 above).

to arable between the catastro and the sales is everywhere similar (that is, 43 : 24); and (3) that the lands sold whose use is not specified are on the whole arable. How open to error this procedure can be is seen in the figures for zones JK and JN, both of which show a considerably smaller ratio of olive groves to arable in the sales than in the catastro. If one assumes that olive cultivation was expanding at least as fast as grain production, the readings for these zones suggest that one should allow a margin of error of at least 50 percent in comparing the two ratios. Greater growth than 50 percent in Column D provides a strong presumption of expansion of olive groves in the second half of the century. Zones JR, JL, JP, JS, JH, JO, JQ, and JC show such expansion.

Map 16.5 makes the result visual. Assured growth occurred primarily in the south of the province: the three southern sierra and the two southern valley zones. Two other zones were contiguous to older olive regions, JC in the west and JL on the Loma de Ubeda, next to Baeza zone. The other zone to show growth, JH (Mancha Real), was the only Guadalquivir basin zone with an established olive production at mid-

Map 16.5.
Jaén Province, Concentration of Olive Groves, ca. 1750,
and Expansion of Olive Groves, 1750–1800

century that definitely revealed expansion. The traditional olive areas in the north and west, Andújar, Arjona, Linares, Baños, and Baeza, either did not expand olive groves or expanded so little as not to show up in this analysis. (Zone JN, also an olive zone in 1751–53, showed an apparent decline in olives and in the sales appeared as a region not into the market. It was a marginal zone as far as commercial agriculture was concerned. Las Navas was located here.) On the other hand, five of the eight zones of confirmed growth had few olive trees in 1750.

This apparent paradox may be partially the effect of the statistical procedure: it is easier to achieve a high proportional increase when starting from a low base. The range of the estimates in Column D is so wide, however, as to indicate that the difference in rate of growth was real. For all the shakiness of the table, it suggests that the expansion of olive production in Jaén province took place in waves. A first wave, antedating 1750 and not observed in this study, had affected the rich basin zones JA (Andújar), JB (Arjona), JE (Linares) as well as the two loma zones JK (Baeza) and JN (Santisteban) and the Sierra Morena zone JJ (Baños). In the half century of this study growth here either slowed or came to a halt. Other zones that had turned toward olives before 1750 continued to expand, JH, JO, and JP. JH (Mancha Real) was in the basin south of the river, JO (Bedmar) and JP (Pegalajar) in the sierras nearby. Finally, a new wave of expansion after 1750 hit many of the zones not previously affected, leaving only two in the east and a belt of three in the center still strongly committed to grain as the cash crop in 1800. These last, oddly, include Martos (JF) and Cazorla (JI), today among the areas in the province most intensively dedicated to olive cultivation. Apparently their specialization did not begin until after the fall of the old regime.

That most zones strongly devoted to olive cultivation in 1750 give little evidence of growth in production over the next half century presents a puzzle. During this period the price of olive oil rose somewhat faster than that of wheat. According to the price tables published by Earl Hamilton for New Castile, if the mean of annual prices for the decade 1751–60 is assigned the index number 100, during the decade 1791—1800 the mean price of wheat was 196, of olive oil 230.^[14] At the end of the century olive oil had a wider market than wheat. It was being exported from Spain to northern Europe, whereas the market for wheat was limited to central Spain, since the periphery imported grain from

[14] Hamilton, War and Prices, Appendix 1.

abroad.^[15] Why should landowners already engaged in producing olive oil not continue to push into an expanding market? The answer may be economically rational if one considers that the expansion of olive groves could entail taking over arable land. The use of labor for the cultivation of olives and grain is complementary. The harvests are about six months apart, with olives in winter and grains in the summer, and the plowing of olive groves and the pruning and care of the trees can be done during the slack seasons for grain growing (Figure 16.1). So long as the local labor did the cultivation of olives, the opportunity costs of expanding olive production were low. Although the chapters on Baños and Lopera show that migratory labor was used for grain harvests and thus presumably was available for olive harvests as well, the marginal cost of migratory labor would be higher. The differential rise in the price of olive oil was not great enough to encourage unbalancing the established economy by turning grain fields to olive groves. Pastures and wastes would be sacrificed, but most of these were public lands and could be turned to private use only by corrupt means. Expansion into pastures and wastes was probably going on, but if a balance of wheat and olives was being maintained, it would not show up in Table 16.10. In other words, under prevailing circumstances, there was a limit to the proportion of land that could be economically devoted to olives, and the older olive regions of the Guadalquivir basin appear to have approached this limit by 1750. Only a sharp drop in the price of wheat vis-à-vis olive oil, such as occurred at the end of the nineteenth century when American wheat invaded the world market and the international demand for olive oil rose sharply, would create conditions that would encourage the expansion of olive groves beyond this limit.^[16]

If this were the full explanation, the proportion of land devoted to olive production would have stopped in all zones at about the same level. This was not the case, however. In the zones that showed no clear growth between 1750 and 1800 (JA, JB, JE, JJ, JK), the ratio of the value of olive harvest to arable harvest ranged from 34 to 225 percent in the 1750s. The ratio was highest in Andújar (JA), a zone that in many ways was unique. The mean amount spent here by buyers was twice as high

[15] According to Laborde, View of Spain 2 : 75, 132–33, Cádiz and Málaga exported olive oil to northern Europe (along with wines and dried fruits). Townsend, Journey Through Spain 3 : 30, corroborates the export of oil from Málaga. Laborde says Andalusia also shipped some oil (together with much wheat) to other parts of Spain (2 : 131). For wheat imports to Spain: Anes, "Agricultura española," 258.

[16] For the expansion in demand of olive oil after 1880, especially in South America, see Vicens Vives, Historia 5 : 238–39.

Figure 16.1.
 Seasonal Demand for Labor in Grain Fields and Olive Groves
SOURCE . Martínez Alier,  La estabilidad del latifundismo.  Appendix 1,
no. 3, pp. 350–51 (based on Consejo Económico Sindical de Córdoba,
"Bases para un plan de desarrollo económico de la provincia," June 1962).
NOTE : The greatest demand for labor comes in February and March. In
February the end of the olive harvest overlaps with fertilizing the arable
(a task perhaps not performed in the eighteenth century). In March, pruning
and cultivating the olive trees overlaps with weeding the grain fields. Much
of the harvesting of olives and weeding of grain fields is done by women.

as in any other zone (Table 16.1), and the total land disentailed that was olive groves and cortijos was higher than in any other zone (76 percent, Table 16.3). One can see some reasons for its peculiarities. Within it lay the wide alluvial basin of the Guadalquivir River, the most fertile soil in the province. The English traveler Joseph Townsend commented on its location "in a rich and highly cultivated plain," and the Frenchman Alexandre de Laborde noted that "the land about is very fertile and produces a great deal of corn, oil, wine, honey, fruit and game."^[17] Equally important, no doubt, was its location on the highway from Ma-

[17] Townsend, Journey Through Spain 2 : 296–97; Laborde, View of Spain 2 : 111.

drid to Seville and Cádiz. The continual traffic on this road, making it easy to import grain and migrant labor, combined with its fertility, could encourage its landowners to develop olive production beyond the level of the balance with arable that would be favored elsewhere. Evidence for the importance of the Andalusian highway is also provided by Baños (JJ) zone. Also on the highway, it was the other zone that had stabilized at a high ratio of olives to arable (107 percent).

6

If we return now to the categorization of the zones into three types of orientation toward the outside market at the end of the century, we find a clear and suggestive relationship between these categories and the evidence of increase in olive cultivation (Table 16.11). Type A₁ zones, where all levels of buyers were into the market, show little expansion (only one zone of seven). These are the zones of relatively large, established hidalgo elites. At the other extreme, all three Type C zones, not into the market, also show no evidence of significant expansion of olive production. Expansion occurred in Type B₁ zones (three of four), those with only the top levels of buyers into the market and thus with a narrower elite than A₁ zones. Four other zones also gave evidence of expansion. Three were the deviant cases of types A and B, those that were into the market but with average purchases much lower than the majority of the zones of their types, and the fourth was zone JC, not included in any type because of its few sales. The B₂ zone was acting like B₁ zones, expanding its olives. The difference that calls for explanation is between A₁ and A₂ zones, since the former were not expanding their production. Table 16.8 shows that the A₂ zones (the two contiguous zones east of the zone of Jaén city, JH, Mancha Real and JO, Bedmar) had a similar percentage of their buyers who were noble as the A₁ zones. There was also much income from arrieros in this region, concentrated in JH zone (see Table 16.7 and Map 16.4), another characteristic of A₁ zones. Although similar in these respects to A₁ zones, the A₂ zones had not seen their wave of expansion recede by 1750 and were still in full growth. Perhaps their distance from the major arteries of traffic explains their late blooming and the modest fortunes of their elites. If so, they provide another example of the influence of communications on economic growth.

The main conclusion to draw from Table 16.11, however, is that the zones where the top buyers were devoting the largest proportions of their capital to acquiring cortijos were those where the expansion of

olive production was most pronounced. (Compare Table 16.4, Types A₂ and B₁ .) Commercial agriculture was old here, but it had been directed to grain production, hence the cortijos; now the rural entrepreneurs were discovering the advantages of olives, for they were nowhere near the limit of efficient use of labor in the olive groves.

An apparently extraneous factor may clarify the forces at work, the jurisdiction over the towns. Spanish municipalities could be under any of four types of jurisdiction: (1) directly under the crown (royal or realengo ); (2) under lay señores, that is held by a hereditary grant to a noble family (señorío lego ); (3) under ecclesiastical señores, such as episcopal sees and monasteries (señorío eclesíastico ); and (4) under the military orders founded in the Middle Ages (señorío de orden militar ). The last were effectively under the crown since the king of Castile became the permanent grand master of the orders in 1523. Table 16.12 shows the percentage of each type of zone under each jurisdiction. It is based on the population rather than the number of towns, so as to avoid giving undue weight to small places, of which many were under señorío lego. Towns in A₁ zones, which experienced the wave of growth in olive production prior to 1750, were overwhelmingly under royal jurisdiction (89 percent of the population), either in royal municipalities or in the territory of the Order of Calatrava. In Type C zones, not into the market and not expanding, only 46 percent of the population lived under royal jurisdiction. Except for Type B₂ (zone JQ), those zones experiencing the current wave of expansion (Types A₂ , B₁ ) were closer to Type C than to Type A₁ , zones, with respect to the jurisdiction under which they lived.

Why was this so? Did señorío retard economic growth, or did the kings of Castile, when they gave towns to new lords, consciously retain for themselves the ones that were most economically evolved? The rulers had granted seigneurial jurisdiction over the towns of Jaén at various times (Map 16.6). Since the province formed a part of the frontier of Granada until the conquest of this kingdom, strategic considerations were central to the early decisions on jurisdiction. In 1231, before the conquest of the kingdom of Jaén was complete, Fernando III assigned the strategic fortress of Quesada, on the southeastern frontier, to the archbishop of Toledo to capture and defend. Quesada soon fell, and in the next twenty years Cazorla and Iznatoraf and the villages in their region were added to what became known as the Adelantamiento de Cazorla, under the jurisdiction of that prelate, who ruled through a local official called the adelantado until the seventeenth century. The Muslims captured Quesada in 1303, and after it was recovered, Alfonso XI kept it under the crown, subject to the city of Ubeda. The rest of the adelantamiento remained with the archbishop through the eighteenth century. Cazorla was the main center south of the Guadalquivir, but to the north Iznatoraf was superseded by two villas created by the archbishop in the late fourteenth and fifteenth centuries out of villages within this town's limits: Villanueva del Arzobispo and Villacarrillo.^[18] Elsewhere in Jaén, the usual royal policy was to keep the main towns and those along the frontier under the king's jurisdiction. Of the latter, only Bedmar received a señor before 1400. Various places in the north of the province were given out before 1410, notably the jurisdiction of Santisteban del Puerto and some towns between Bailén and Baeza, but not these two important centers. As the war against Granada proceeded in the fifteenth century, the two frontier fortresses of Alcaudete and Huelma received señores, as well as Bailén. The remaining seigneurial jurisdictions were granted by the Habsburgs, except for the small town of Los Villares in the southern sierra apparently not given out until Felipe V.^[19]

In the eighteenth century the king still had the towns along the main highways to Seville and Granada except Bailén on the Andalusian highway, Alcaudete on the old road to Granada, and Noalejo on the new

[18] Rivera Recio, Adelantamiento. I am indebted to Stanley Brandes for this reference.

[19] See Ladero Quesada, Andalucía, 56–64, for the fifteenth century and earlier, and Guia oficial de España, 1902 , 244–322, for the rest. The latter gives the date of the original grant of the title. It may not coincide with the grant of señorío over a specific town, but it is the best evidence readily available.

Map 16.6.
Jaén Province, Jurisdictions and Residences of Señores

one. Through the Order of Calatrava he also held the west. Although strategic considerations entered into the alienation of jurisdiction, conscious economic planning does not seem to have been involved. The economic evolution of the towns under lay señorío thus appears to have been more an effect than a cause of the type of jurisdiction.

Señores received tribute from their towns, but the case of Navas, one of the first given out, unless widely atypical, indicates that this tribute was not heavy. One has to look for indirect rather than direct effects of seigneurial domination to account for the relative backwardness of their towns. Evidence for such effects is not hard to find. One feature is the residence of the señores. Of twenty-five towns in all zones under señorío

lego, seventeen had señores who resided in Madrid at the end of the eighteenth century and three others lived elsewhere outside the province (Córdoba, Granada).^[20] Only three resided in the province, but none of them in the town of which he was señor. (The residence of the señores of two towns is unknown.) Lay señores were thus by the middle of the eighteenth century overwhelmingly absentees (if they had ever been residents), most of them living in Madrid. The lord of the towns of ecclesiastical señorío was the archbishop of Toledo, also nonresident.

Benign neglect by an absentee señor or señora might not have been prejudicial, were it not for the fact that in most cases he was also the owner of extensive properties in the towns under his jurisdiction, indeed he was frequently the largest property holder (hacendado mayor). This was so in seventeen of the twenty-five towns of señorío lego, in fifteen of the twenty towns whose lay señor lived outside the province (Map 16.6).^[21] In two other towns under señorío lego the largest property owner, who was not the señor, also lived outside the province. The hacendado mayor was much more likely to be an absentee in towns under señorío lego than under other forms of jurisdiction (Table 16.13). Eighty-two percent of the population under royal jurisdiction lived in towns where the largest property owner was either a vecino or the municipal council, and an additional 13 percent where he lived somewhere else in the province. For the population under lay señorío the situation was almost the reverse: only 18 percent lived in towns where a vecino was the hacendado mayor; 72 percent were in towns where he lived outside the province (for 57 percent, in Madrid). Towns of the military Order of Calatrava had a pattern close to that of realengo towns, except that more hacendados mayores lived elsewhere in the province (in Baeza, Jaén, Martos, and Huelma) than in the town itself. The six towns in the eastern part of the province under the archbishop of Toledo were very similar to royal towns. Lay señorío stands out as the only jurisdiction where the largest owners were heavily absentees. In conceiving the effect of this fact on economic development, it seems proper to assume that absenteeism among other large owners was more likely where the largest owner was an absentee.

A second potential disadvantage of lay señorío appears in the relation between the property in the hands of the hacendado mayor and the total property in the town (Table 16.14). In towns of this jurisdiction the

[20] The residence of twelve is identified either by resp. gen. Q 2 or by the list of hacendados mayores in AGS, Dirección General de Rentas, Única Contribución, libro 328; and five others by a survey made for Napoleon in 1808: ANP, AF IV, 1608 , 2 : 46.

[21] List of hacendados mayores (see above, n. 20).

largest owner had a far greater proportion of all the property (a mean of 35 percent) than elsewhere. The mean for realengo is 10 percent, and for the other two types even lower. Nor is this the result simply of the smaller size of towns under señorío lego;^[22] the hacendados mayores had on the average much larger properties than elsewhere. The mean income from their properties was over seventy thousand reales, in other jurisdictions forty thousand or less (see the table). Thus towns under señorío lego had a much higher concentration of property than other

[22] The mean town population in 1786 for the different types of jurisdiction was realengo, 3,410; military order, 2,360; señorío lego, 1,280; señorío eclesiástico, 2,760.

towns, with the largest property frequently in the hands of the señor himself. Furthermore these large owners were far more likely to be absentees than those elsewhere. In the distribution of property and nonresidence of the largest owners, towns of lay señorío formed a class by themselves.

If these characteristics explain why towns under lay señorío had responded less than others to the lures of the market, they should correlate with the extent that the zones were into the agricultural market, but in fact they do not. If concentration of property in the hands of hacendados mayores were a cause of economic backwardness, one would expect a trend of less to greater concentration from Type A₁ to Type C zones, but no such trend shows up. Nor should this be surprising. Large properties were more rather than less commercially oriented than others. Absenteeism of the hacendado mayor does not provide such a trend either.

The towns of señoro lego had a different characteristic that more likely accounts for their slower economic growth. They did not attract a strong hidalgo elite. The census of 1786 indicates that towns under lay señorío averaged 3.5 hidalgos per thousand adult males, while those under other forms of jurisdiction averaged between 17.6 and 25.0 (Table 16.15). In almost three-fourths of the towns under lay señorío there were no hidalgos present; the proportion of towns without hidal-

gos was between one-fifth and one-third for other jurisdictions. Small towns as a rule had a smaller proportion of hidalgos than big ones, but the difference between towns under lay señorío and others cannot be accounted for only by their smaller size.^[23] When all towns under lay señorío with populations over one thousand are compared with royal towns in the same population range, the former still have a strikingly smaller proportion of nobles, as the table shows. In the six towns under the archbishop of Toledo in the eastern part of the province, hidalgos were as dense as in towns under the crown. The incompatibility between señorío and hidalgo residents applied only to señores who were laymen.

One of the characteristics of Type A₁ zones was a prosperous hidalgo elite, as shown by the census of 1786 and the characteristics of the purchasers (Tables 16.8 and 16.9). We see now that even though royal towns had such an elite, those under lay señorío did not, or did to a much lesser extent. For reasons that may have been largely social, hidalgos preferred to live in towns where they were directly under the king rather than under another noble.^[24] In royal towns they had the government effectively in their hands and the church as well, to judge from Baños and Lopera. Their relative absence surely was related to the undeveloped state of the economies of towns of señorío lego.

Although lay señorío appears to have been a major factor contributing to economic backwardness in Jaén province, ecclesiastical señorío was not, according to the evidence developed here. Only one of the six towns under the archbishop of Toledo had an absentee hacendado mayor, they did not have a high concentration of land ownership (the proportion owned by hacendados mayores was even lower than in royal towns), and three had a large contingent of hidalgos. In all these characteristics they resembled royal towns. Yet they were all located in zones that showed no commercial expansion in our period.^[25] The only appar-

[23] The increasing proportion of hidalgos as town size grows is as follows (total hidalgos over total males 25 and over, census of 1786): towns with populations 30–499, 3.2 hidalgos per 1,000 adult males (N = 13 towns); 500–1,499, 10.6 (N = 22); 1,500–2,999, 9.8 (N = 17); 3,000–14,999, 16.8 (N = 16); over 15,000, 48.4 (N = 1, Jaén city). The concentration of hidalgos in the larger cities of Jaén (and the rest of Andalusia) was already the case at the end of the sixteenth century (Molinié-Bertrand, "'Hidalgos,'" 80).

[24] Domínguez Ortiz, Sociedad española, 326, remarks that under the Habsburg rulers, "when a town fell under señorío, the more illustrious families expatriated themselves so as not to be subject to someone they considered of equal or lesser rank than they." Vassberg, Land and Society, 98–99, also reports a strong preference in the sixteenth century for royal jurisdiction over señorío.

[25] They were two of the three towns in Zone JI, the only Type B zone not to give evidence of commercial expansion, all three of the towns in Zone JM, and one of four in Zone JN (both Type C).

ent reason for the backwardness of these regions was their geographic isolation in the northeast and east of the province, away from any major transportation artery. The 1787 map of Tomás López shows no direct road from either Cazorla or Villanueva del Arzobispo west to Jaén, Ubeda, or the main Andalusian highway, or east to Murcia. The only connection was a local north-south road. Even today it takes a considerable detour from any regularly traveled route to reach these towns.

7

One might anticipate a relationship between the different types of zones and their demographic structure. Certain connections have already appeared involving the size of towns. Town size was related to jurisdiction and to the proportion of hidalgos in the population, but it was not significantly related to the type of zone. Population changes between the four main census dates, 1712, 1751–53, 1786, and 1826, would give an indication of the different demographic development of the various zones, provided of course that the data are sufficiently reliable. Unfortunately, we have plenty of evidence that the data are not very good. The towns were big, census takers hardly went from door to door, and they had to accept the information available from the municipal officials and the clergy. The information in the catastro is probably the best, but this gives only the vecino count. The 1712 census of vecinos is everywhere notoriously low, for the respondents feared that it would be used to impose taxes or military conscription. Rather than attempt to establish real growth rates, I have compared the figures in the consecutive censuses, looking for different patterns of growth among the types of zones and hoping that errors in reporting would balance out (Table 16.16). While the different types of zones show markedly different growth rates, the variation among the zones within each type is so great that these differences are not statistically significant. One does find a meaningful pattern, however, for the middle time period, from the catastro to the census of 1786, if one takes not the types of zones but the individual zones. Comparing the population change of those zones that showed an expansion of olive cultivation against those that did not, one finds a more rapid population growth among the former, statistically significant at the 1-percent level.^[26]

[26] Since the vecinos reported by the catastro do not include the clergy, I subtracted the number of clergy in 1786 from the total population before comparing it with the vecino count of 1751–53. Zones JC and JD omitted.

This information provides convincing evidence that the zones that we have determined were expanding their commercial agriculture were also undergoing a more rapid population growth than either those zones already heavily commercialized by 1750 or those not yet significantly commercialized in 1800. Before 1751–53 and after 1786, the population was also increasing, but no distinct patterns emerge among the types of zones or between those zones expanding commercially and those not doing so between 1750 and 1800. If population growth was related to the expansion of commercial agriculture, as the data strongly indicate, then the distinct wave of growth that we have identified in certain parts of the province did not begin much before 1750 or extend much after 1800. Commercial growth would certainly have been present in the province, especially after 1800, but following a different pattern. Indeed, I propose that the disentail changed the pattern by giving an opportunity for commercially oriented elites of all zones to expand their activities.

The demographic mechanisms whereby the expanding zones grew faster than the others between 1751–53 and 1786 cannot be satisfac-

torily determined by the available data. The information provided in Table 16.16 reveals no statistically significant difference in family size in 1751–53 among Types A₁ , B₁ , and C, nor does any turn up between the zones growing commercially and those that were not. Likewise, the census of 1786 shows no significant correlations for the proportion of the population under sixteen (that is, for a younger and presumably more rapidly expanding population) or for the proportion of males in the labor force (over sixteen). One may assume that part of the differential growth was due to migration and part to natural increase, but the available data do not enable us to sort out the factors.

8

The analysis of the nature and evolution of the different zones of Jaén province has been complex, but when all the features are assembled a clear pattern emerges. The transition from agricultural production for local consumption to that for a wide market proceeded in various stages. Sometime before 1750 the richer sectors of the Guadalquivir basin and some adjacent loma and foothill regions had developed a strong commercial agriculture based on the balanced cultivation of wheat and olive groves. Other regions to the east of these, the zones of Ubeda, Cazorla, and Huelma, had numerous cortijos devoted to commercial grain production but few olives. Some of the province was only weakly oriented toward the market at this time: a belt of the Guadalquivir basin marked by saline soils unpropitious for grain, running from Menjíbar to Martos and including at this time the zone around Jaén city, and certain outlying regions: the southern valleys of Alcaudete and Alcalá la Real, the sierra valleys south of Jaén, and the loma around Villanueva del Arzobispo in the northeast.

How long this pattern had been in existence this study cannot say. The flimsy demographic data covering the first half of the century do not reveal any pattern of evolution in this period. The province had been known for its olives since at least the sixteenth century, but we do not know the area and extent of olive production at that time or the changes that had occurred since.

The towns of advanced commercial agriculture in 1750, those with a balance of grain and olives, had as a rule strong sociopolitical elites, of which hidalgos were the outstanding sector, including clergy of hidalgo background. Hidalgo families, in fact, stood out as the commercial agricultural class par excellence in Jaén. These towns had other distinctive

characteristics. Many had a prosperous sector of muleteers, and their largest landowners, the hacendados mayores, were more likely to be local residents than in other regions. They were also more likely to be under royal jurisdiction. Towns under lay seigneurial jurisdiction, in contrast, on the whole lacked both a strong hidalgo elite and advanced commercial agriculture at the beginning of our period.

A wave of expansion of olive cultivation swept over portions of the province in the second half of the century—the only stage of economic growth that this study can identify and describe. In these regions a demographic spurt responded to the improving economic conditions. The regions of the north and west that had already achieved a balance of arable and olive groves did not experience this wave, but the region east and south of the city of Jaén, although already actively involved in the cultivation of olives in 1750, did expand this crop. In addition most of the zones that had specialized in grain production for the market in 1750—the regions of cortijos and few olive groves—participated in this wave. Together they formed a semicircle around the east and south of the older olive area. These were less wealthy zones, with a weaker hidalgo elite and extensive lay señorío. The general expansion of economic activity in Spain in the second half of the century drew these sectors into olive production, overcoming their less propitious social structure and relative geographic isolation.

By 1800 only three zones in the east and northeast—Cazorla (JI), Villanueva del Arzobispo (JM), and Santisteban (JN)—and Martos zone (JF) in the west lacked evidence of active commercial production. (All regions, of course, shipped some products out for sale, the question is one of degree.) The first three contained the six towns of ecclesiastical señorío, but this does not appear to be the cause of their backwardness, which our information can assign only to their geographic isolation. Martos, in contrast, was on the old road to Granada and belonged to the Order of Calatrava. The explanation for its failure to take off before 1800 would appear to be its soil. A saline marl, it is not conducive to good arable, with the result that the zone had not developed the export of grain. Although its soil is ideal for olives—as one can see today from its virtually seamless blanket of silver-green trees—it had only a weak hidalgo elite in the eighteenth century (Table 16.9) and little commercial orientation. Its case indicates that entry into commercial agriculture in the old regime came first via the sale of grain and not directly with olives. Oil production required a significant capital investment: ten to fifteen years' wait after planting for the production of the first harvests,

and the building of olive mills. The commercial growth of grain could provide the needed savings. A later wave of olive expansion would turn Martos into one of the richest olive zones in Spain, as it would Cazorla in the east, but not until after the old regime.

Wide commercialization of agriculture thus begins in most parts of Jaén well before the end of the old regime. Four factors interacted to determine the rhythm of the different regions: the quality of the land, the accessibility of major highways, the nature of jurisdiction, and the presence of a prosperous, largely hidalgo elite, which ideally included the hacendado mayor. Originally, the development of this elite depended greatly on the other three factors, most notably the type of jurisdiction.

The example of Jaén province indicates that lay señorío—feudalism, if one wishes—retarded the transformation of agriculture into commercial production by removing the largest owners from the towns and discouraging the presence of hidalgos and no doubt by other subtler means not detected through our data. The economic takeoff of the late eighteenth century, however, affected towns of señorío lego as well as others, moving them into olive oil, where they were not inhibited by factors such as poor soil or poor communications.

Although señorío was an obstacle, hidalguía was not. Hidalgo families formed the core of the sector that developed the local agricultural potential. Hidalgos too might be considered feudal, especially because a good part of their property was tied up in vínculos and mayorazgos, but they were far from being anticapitalist as a group. The reformers of Carlos III's reign, one recalls, roundly condemned entail as harmful to the economy and hidalgos as an idle, parasitic class. Jovellanos claimed that no one in Andalusia was investing in the improvement of agriculture because entail made owners neglectful absentees. Cortijos too they found evil; they wanted these estates divided into prosperous homesteads. In their eyes, the laws of property in Andalusia had created a class of poderosos who were taking over whole towns and impoverishing the laboring poor.

The present study indicates that the reformers misread the situation. Vinculos in some instances were surely neglected, as were other types of property, but as a group the notables were interested in agricultural development. Cortijos offered them the wherewithal to pursue further commercialization through olive groves. They included many poderosos, for we have seen them in control of municipal governments and they exploited the large and powerless class of jornaleros. But what eco-

nomic initiative existed was largely in their hands. More like members of the English gentry than the royal reformers appreciated, very different from the popular perception of the lazy impoverished hidalgo stereotyped by the authors of the Golden Age and accepted uncritically by later historians, they seized a good possibility when it arose.

Such a possibility came with the desamortización of Carlos IV. These hidalgos and other notables associated with them were eager to break entail—the entail of the church—if they could get more land for themselves, and this meant members of the hidalgo families who had entered the clergy as well as the heads of families. Notables were of course not the only buyers, but they dominated the top levels and were proportionately much more numerous among all buyers than in the population as a whole. Disentail strengthened their class and pushed it toward greater commercial orientation.

Everywhere commoners also bought lands, in most zones in larger numbers than the notables. Properties that were not market oriented were more likely to go to them, and those that needed intensive labor input. Commoners with available capital also bought commercial properties, especially olive groves, and thus established a common interest with the notables. This was especially the case in those zones of grain culture, where the hidalgo elite was exiguous. Desamortización did not mean the displacement of an older landed class by a new class of capitalists but rather the invigoration of the older class by adding new recruits and by giving its more driving individuals a chance to excel. In the process, it accelerated the commercialization of agriculture.

Chapter XVII—
Salamanca Province:
Hardening the Patterns

The patterns discovered in Jaén most likely existed elsewhere in the Guadalquivir valley, but they were not representative of Castile as a whole. This truth is proclaimed when one turns to Salamanca province, for almost none of the characteristics of Jaén are found here. Most of the people who lived in the countryside of the portion of the province covered by this study inhabited small towns and villages that were set relatively close together and enjoyed more egalitarian societies than the large agrarian towns of Jaén.

Geographically and socioeconomically there is much greater contrast among the regions of Salamanca than of Jaén. The portion of the Old Castilian plain that surrounds the city of Salamanca is another world from the hill towns and upland valleys of the southern sierras. The sharp contrasts facilitate comparison, with the result that I have divided the province into only nine zones, while Jaén has seventeen (Map 17.1).

Four zones make up the plain. Its communities are larger and more prosperous to the north and east, where the earth is rich and deep, than to the south and west, a land of thin, sandy soil, better suited to pastures than to fields of grain. In our time, interspersed among the villages like La Mata and Villaverde were many despoblados like Narros and Pedrollén. Armuña zone (SA) north of Salamanca city and Alba de Tormes zone (SB) in the northeastern corner of the province can be classed as rich plains zones. The other two zones, south and west of the city, will be identified as poor plains zones. Charro zone (SC), which includes Pedrollén, lies on either side of the road to Ciudad Rodrigo and Por-

Map 17.1.
Zones of Salamanca Province

tugal, and Ledesma zone (SD) to the north of it borders on the province of Zamora.

There are three sierra zones. On the east (and now in Ávila province) is Piedrahita zone (SI), occupying the valleys north of the Sierra de Gredos. El Mirón is on its northern edge. Just west of it, Béjar zone (SH) is dominated by the sierra of that name and lies astride the main pass between Old Castile and Extremadura. The large town of Béjar was already an important woolen center in the eighteenth century. Finally the zone of Miranda (SG) lies beneath the Sierra de la Peña de Francia. The towns through here are large by comparison with those of the plain. Steep, narrow, cobbled streets lined with two- and three-story granite houses, many still boasting coats of arms from past ages, provide human features that combine with the majesty of the sierras to make this one of the most picturesque corners of Spain.

The two remaining divisions do not fit either of these types. Aldeadávila zone (SE) is the northwestern corner of the province, bordering on

Portugal, where the Salamanca plain gives way to rolling hills and glaciated outcroppings that furnish the material for abundant stone walls. Reminiscent of the sierras, its towns spread out around granite churches with imposing square towers. Salvatierra zone (SF) lies in the center of the province, a relatively poor upland region that juts out into the plain from the ranges to the south. Its thin soil barely covers a shale base, and the small villages, abundant in slate, resemble those of the Charro plain to the west. Although these two zones are very different from each other, they have the common label of hilly zones. Appendix Q gives a detailed description of the geography, population, and political status of the zones.

How unlike Jaén Salamanca province is, one can judge from a simple comparison. In Jaén the mean amount spent by the buyers in each zone was directly proportional to the mean size of the towns in the zone: the bigger the towns, the more each buyer spent. In Salamanca precisely the opposite correlation appears. If one compares Table 17.1 with Table 17.2, which gives the mean town size, one sees that the larger the size of the towns, the smaller amount each buyer was likely to spend (r = –.58). The correspondence again is loose, only about one-third of the variation in the amount spent can be attributed to the size of the towns,

and the correlation is significant only at the 10-percent level. As the present chapter proceeds, however, it will become apparent that the relationship is not fortuitous. The challenge is to produce a different concept of both structure and change to apply to this province.

The analysis of Jaén province was based on a distinction among the zones according to the extent of the orientation toward the outside market of the lands purchased by the different levels of buyers. At first sight the information provided by the sales in Salamanca province does not allow one to draw a similar distinction here.^[1] Table 17.3 gives the basic information on the purchases made by Level 1 and Level 4 buyers. We can assume that the purchasers of rural estates were investing in com-

[1] Many sales in Salamanca province could not be used for analysis of the individual zones. Salamanca city, unlike Jaén, has a small término, so that the 233 sales within its limits were very largely urban properties. I have not included the city in any zone. In addition, the records of many sales of rural properties owned by religious institutions of the city did not state where the properties were located. The vast majority would have been in nearby towns in the partido of Salamanca city, like La Mata, Villaverde, and Pedrollén. But the partido of Salamanca is divided among three zones, SA, SB, and SC, and there being no way to determine in which zone each unlocated sale belonged, they had to be excluded. These are 657 sales (out of a total of 3,314 for the province). Finally, a few towns named in the Salamanca deeds could not be found, and some towns in different zones had the same name, so that 68 sales could not be assigned to any zone. The final number of sales that could be allocated is 2,356: SA, 368; SB, 554; SC, 97; SD, 207; SE, 200; SF, 102; SG, 185; SH, 242; SI, 401.

mercial property, but only in the plains zones did estates form a sizable part of the disentail. These were alquerías and despoblados similar to Pedrollén or to Narros before it was colonized from La Mata in the 1780s, large properties dedicated to both arable and grazing usually leased for payment in money and kind to the farmers who exploited them.^[2] Level 4 purchasers here, who bought most of the estates, were into the outside market. So, of course, were the buyers of large collections of arable plots that would be rented to small farmers for specified payments in wheat and other products. The large outside buyers of the arable plots whom we observed in La Mata and Villaverde were among these people. The plots they purchased, however, were identical to those bought by local farmers found in Level 1, who would be more oriented to domestic consumption than the market so that purchases of arable land in Salamanca are not useful as an indication of the extent of the buyers' involvement in the market.

In contrast, big buyers in the mountain zone of Béjar (SH) devoted the largest share of their purchases to pastures. Here Level 4 buyers were very different from their counterparts in the plains: they spent on the average thirty-three thousand reales; those of the plains four to ten times as much. Béjar zone had many sheep, twenty-two per vecino engaged in agriculture, but five others had more sheep per vecino (Table 17.4). The zone had few cows, yet the disentail indicates that more of its land was devoted to pastures than anywhere else. To what purpose? One may guess that they were leased to the owners of transhumant sheep. The main north-south sheepwalk (cañada) of the Mesta crossed the central sierras by the pass of Béjar, and the largest and best share of the wool shorn in Salamanca province still came from transhumant sheep.^[3] One may logically conclude that many of the pastures acquired by the top buyers of Béjar zone were rental properties to be leased to sheep owners, most of whom resided in other parts of Castile. These were capitalist investments. The buyers were indirectly involved in the market for wool, and they probably had some sheep of their own.

The sales reveal little about the commercial orientation or lack of it of the smallest buyers. Everywhere except in Béjar zone they spent the largest share of their money on arable, whose orientation is indeterminate. This was true, of course, in Jaén too; but unlike Jaén, the second choice of the Level 1 buyers also offers few clues to the extent of their commercial orientation, except in the sierra zones.

[2] See Chapters 7 and 9. Chapter 18 will take up the subject of the despoblados and alquerías.

[3] Larruga, Memorias 34 : 307–9.

In Béjar zone (SH), while the largest buyers were looking for pastures to rent, the smallest buyers had their eyes turned to the outside agricultural market. Although they spent more in paying off censos than on any other item, they devoted almost a third of their capital to their second choice, improved and irrigated plots. All of the plots that Level 1 buyers acquired were linares, enclosed plots planted with flax and usually irrigated. Flax, like grain, could be grown either for domestic use or for sale to centers of linen weaving. According to the contemporary author of a survey of economic activity in Spain, Eugenio Larruga, the cultivation of flax was widespread in Salamanca province, but only certain areas harvested it commercially. He lists 131 towns in our zones where it was cultivated (Table 17.5), but only 16 grew more than a small amount for local use. According to Larruga's list, which of course may not be complete, a larger proportion of the towns in Béjar zone grew flax than in any other zone, and they sold much of it to Extremadura or traded it to outsiders for soap and olive oil.^[4]

[4] Ibid., 35 : 1–19.

The raising and preparation of flax for linen was highly labor intensive. Farmers selected the most fertile fields in their towns, plowed them five or six times, fertilized them heavily and scattered seed liberally, using three times as much seed as they would for grain. Once the flax was up they weeded by hand, and they devoted great care to the harvest. The process of turning the harvested flax stalks into linen was further time-consuming. The linseed for next year's planting must be beaten out, the stalks soaked for a week or ten days in a stream, dried, and gently pounded to remove the skin. The women then took over, beating the flax with a wooden swingle, combing and cleaning it, making it ready for spinning.^[5]

The purchase of linen fields agrees precisely with the pattern described in Chapter 15. Highly labor intensive, the cultivation of flax interested the smaller buyers, those prepared to exploit the land person-

[5] Ibid., 19–21.

ally. Thirty-nine percent of the "improved" fields disentailed in Béjar zone were flax fields (linares). Level 4 buyers devoted only 5 percent of the money spent for improved fields to flax fields, Level 1 buyers 100 percent.^[6] A significant proportion of the small buyers in Béjar zone were commercially oriented, one is led to conclude, but at the cost of great personal input.

The lowest level of buyers in the other two sierra zones also revealed distinctive preferences related to the nature of local agriculture, but it is less clear that they were oriented toward an outside market. In Miranda del Castañar zone (SG), only 29 percent of Level 1 purchases were arable, while 23 percent were orchards. All of these were chestnut groves, a regional specialty,^[7] but most chestnuts were probably consumed locally, since grain was scarce. Level 1 buyers did make 13 percent of their purchases in vineyards and 8 percent in olive groves, no doubt with an outside market in mind, but as a group these buyers cannot be classified as commercially oriented.

Much the same can be said of the smallest buyers in Piedrahita zone (SI). Although half their purchases were arable fields, 22 percent were pastures. At midcentury this zone had the largest number of sheep and cattle per vecino in agriculture of all the zones in the province. Our study of El Mirón showed that pastures were the most sought-after purchase in the region, but they were much more likely to be used for cattle than for sheep, which could graze on the hills and stubble. Some cows and steers were produced for the market, but again, the smallest buyers in this zone did not belong structurally to commercial agriculture.

These inferences exhaust the information presented by the type of property purchased by different levels. The examples of La Mata and Villaverde in Armuña indicate that the nucleated towns of the rich plains zones (but not their despoblados) had a virtual grain monoculture, a conclusion reinforced by the high percentage of arable land in the disentail in these zones (Table 17.3).^[8] Here every farmer, and thus every level of buyers, had to be involved to a greater or lesser extent in the sale of grain.

One is left with a sense that the picture is incomplete and selective. Reviewing the conclusions developed in Chapter 15, one discovers another approach. A much greater proportion of the purchases were paid

[6] Level 2, 79 percent; Level 3, 72 percent.

[7] Larruga, Memorias 35 : 21.

[8] On the grain monoculture of this region, see García Fernández, "Champs ouverts," esp. 695–701.

for in paper currency—vales reales—in Salamanca than in Jaén (70 percent compared to 45). Large buyers everywhere used vales reales more than small buyers and were more oriented toward the national market. Furthermore, a separate analysis of Salamanca province shows that properties oriented toward commercial agriculture were paid for in vales much more than those for home production. This at least is the logical conclusion from a comparison of the sales of rural estates and improved or irrigated fields. Eighty-six percent of the payments for rural estates—most of which would be oriented toward market production—were made in vales reales. The latter account for only 35 percent of payments for improved or irrigated fields—labor-intensive huertas, enclosed fields for herren (fodder harvested green), flax fields, and the like, not all directed to domestic consumption, of course, but of all the categories the one most fully of this kind.

Basically the argument being developed is that, although arable formed the largest proportion of the sales in most zones, and one cannot distinguish between arable bought for commercial purposes and arable bought for subsistence agriculture from the deeds' descriptions of the properties, one can predict reasonably well the orientation of the purchases by the currency used in payment. Table 17.6 provides this information for Level 1 and Level 4 buyers. In all the plains zones and both hilly zones Level 4 buyers paid for more than 70 percent of their pur-

Map 17.2.
Salamanca Province, Types of Zones

chases with vales. This includes the four zones where Level 4 buyers put a quarter or more of their money into rural estates. In the three sierra zones Level 4 buyers paid for about half their purchases in vales, and were probably much less market oriented than their counterparts elsewhere. While Level 1 buyers used vales much less, those of SA, SB (rich plains), SD (poor plains), and SE (hilly) paid for 18 percent or more of their purchases in vales and were probably involved in the market to a significant extent.

On the basis of this analysis, which is less conclusive than one might wish, one can now assign the Salamanca zones to the categories developed for Jaén (see Table 17.1 and Map 17.2).

Type A. All levels of buyers involved in the outside market

A₁ . Large purchases

Rich plains zones: SA (Armuña), SB (Alba de Tormes)

Poor plains zone: SD (Ledesma)

Hilly zone: SE (Aldeadávila)

A₂ . Small purchases

Sierra zone: SH (Béjar)

Type B. Top levels of buyers involved in the outside market

B₁ . Large purchases

Poor plains zone: SC (Charro)

B₂ . Small purchases

Hilly zone: SF (Salvatierra)

Type C. No levels of buyers involved in the outside market

Sierra zones: SG (Miranda), SI (Piedrahita)

2

With no more than four zones in any one type, statistical correlations become almost impossible to establish, yet one can seek apparent relationships between the types and the social and economic characteristics of the zones. Certain obvious features show no correlation. This is true of mean town size (Table 17.2). Type A₁ zones span the whole range from the next to the lowest (Ledesma) to the next to the highest (Aldeadávila). Industrial activity is also not related to the extent into the agricultural market (Table 17.7). There is less cloth making (fewer carders and weavers) in the commercially oriented Types A₁ , B₁ , and B₂ zones than in Types A₂ and C.

The table is drawn up so that geographic comparisons can also be observed, and it shows that the pattern of crafts can be explained much better by geographic affinities than by the extent of commercial agriculture. Cloth making was concentrated in the sierra zones, with SH (Béjar) far ahead of any other, obviously because the city of Béjar was a major woolen center of Castile. This region, one recalls, grew much flax, while SI (Piedrahita) had more sheep per peasant family than any other zone. Much of the wool and linen was woven locally, a task performed by professional carders and weavers, such as those we met in El Mirón.^[9]

Table 17.7 shows that the total proportion of craftsmen in the popu-

[9] Zones SD, SE, and SF also grew flax, and this may explain their slightly larger proportion of clothmakers. All zones did some weaving, but the catastro summary does not distinguish between woolen and linen weavers.

lation reflects closely the number of carders and weavers. SD (Ledesma) is a deviant case, however; with few involved in cloth making it still has one of the highest proportions of craftsmen. This deviance is misleading, for it reflects a problem inherent in the study of the zones of Salamanca. The plains zones were dominated by three cities: Salamanca (population 16,438 in 1786), Ledesma (1,844), and Alba de Tormes (2,293). Salamanca city lies at the junction of three zones, and I excluded it from them. Alba de Tormes is in zone SB and Ledesma in SD. All these cities had a concentration of crafts serving their residents and the richer members of their hinterlands. If Ledesma is excluded from zone SD, the proportion of craftsmen becomes similar to that of the other plains zones. Alba de Tormes had less impact.

In sum, craftsmen tended to concentrate in the cities, but only Béjar had a specialized industry. Elsewhere, the sierra zones had more crafts-

men than the plains zones, in part no doubt because the town size was larger in the sierras. But in part one can also observe that the proportion of craftsmen was negatively correlated to the commercialization of agriculture. We can recall that the craftsmen in La Mata and Villaverde, both in the rich plain of Armuña, were hard pressed and declining, faced by the competition of outside products brought in along the lines of communication that served to export their harvests. Artisans in the larger and more isolated sierra towns fared better, and their zones had more craftsmen than the plains, with the hilly zones in between. One may conclude that the concentration of crafts responded to two factors: positively to the production of local wool and flax and negatively to the commercialization of agriculture.

Because the greater concentration of cloth making in the sierra zones does appear in a region of noncommercial agriculture, it exhibits one of

the characteristic features of the putting-out system.^[10] The appearance is misleading, however. Domestic weaving does not give evidence of being a new development in the sierras in the eighteenth century or of being organized as proto-industry. Nor did it offer an opportunity for economic growth. Béjar continues until now to be a center for woolen manufacture, but in the transition to modern factories, it destroyed the artisan production of the towns around it.^[11] The case of the sierras thus does not invalidate my earlier conclusion that the lack of the necessary commercial infrastructure precluded the emergence of a proto-industry in central Spain similar to that in northwest Europe or the Spanish periphery. One cannot extrapolate directly from Salamanca to all Castile, of course, but historians who have been looking at early modern Spanish industry tend to support this conclusion.^[12]

In Jaén muleteering was associated with market-oriented agriculture. Table 17.7 reveals no similar association in Salamanca. It is true that the two zones with the most transport activity, Armuña (SA) and Béjar (SH) have all levels of buyers involved in market-oriented agriculture (Types A₁ and A₂ ), but the concentration of arrieros in other Type A zones was among the lowest in the province. Again, geographic location, not commercialized agriculture, was the dominant factor. It worked in two ways. First, the sierra zones engaged in muleteering as they did in crafts, as compensation for poor agricultural potential. Second, towns that sent out arrieros were located on or near the main north-south highway from Andalusia and Extremadura through Béjar and Salamanca to León, Burgos, and the north (Map 17.3). It passed through Zones SH (Béjar), SF (Salvatierra), and SA (Armuña). A number of towns with arrieros were clustered in the vicinity of Salamanca. The city drew heavily on the food production of the surrounding plain, but our study of La Mata and Villaverde showed that their muleteering activity did not depend on transporting grain to the city, done rather by the farmers themselves. Muleteers served the city in other ways, in long-range operations to the north coast, Extremadura, and Madrid, that is, along the old Roman road. Although the arrieros in the region may have got their start from furnishing Salamanca city, their clientele was now located throughout the northern half of Castile.

[10] See Jones, "Agricultural Origins of Industry."

[11] González Enciso, "Protoindustrialización en España," 35. His Cuadro 5 gives the number of looms in four towns of the region, based on Larruga, Memorias. The catastro, however, shows weavers in all but four of the thirty-two towns of the partido, 525 altogether, although it does not specify what material they worked (AHN, Hac., Catastro, Salamanca, estado seglar).

[12] González Enciso, "La protoindustrialización en España."

Map 17.3.
Salamanca Province, Income from Muleteering, ca. 1750

In Jaén too, muleteering was a function of highways, those between Castile and Seville or Granada. Zones JD in the poor basin and JQ in the southern sierras owed their income from transportation to their location on the Madrid-Jaén-Granada road, for they, like Salvatierra and Béjar zones in Salamanca, were in poor agricultural districts. In Jaén we found that the presence of a major road also fostered commercial agriculture, especially specialization in olives. The main road of Salamanca had no such effect. Comparative advantages of location and type of soil determined the production of grain and livestock for the market, factors stronger than a secondary road system.

Despite such marked differences, at least one attribute of commercial agriculture was common to both provinces. The Madrid records of the sales show only about two-thirds as many men and women entitled to call themselves don and doña among the buyers in Salamanca as in Jaén, but here too they were associated with commercialized agriculture. In all zones, as we have come to expect, a far higher percentage of

big buyers were notables than small buyers (Table 17.8). Furthermore a greater proportion of buyers was notable in Type A₁ and B₁ zones than in the others, that is, zones where the top levels or all levels of buyers were into the market and spent a large amount per capita. A₂ and B₂ zones, where the amounts spent were small, had considerably fewer buyers called don or doña, similar in this respect to Type C zones, which were not into the market.

The census of 1786 supports this association. The zones where all levels of buyers were into the market (Types A₁ and A₂ ) had more hidalgos per thousand adult males than other zones (Table 17.9), with one apparent exception: SA (Armuña) had slightly fewer than SG (Miranda). The reason is easy to perceive. Salamanca city was the cabeza de partido for SA, as it was for zones SC and part of SB. The hidalgos associated with partido administration or who lived off the income of its lands were to be found in the city. In the sierra zones, including SG, the hidalgos lived in the zones; and they did also in SD (Ledesma), the zone with the highest proportion of hidalgos. The town and village structure was much

different from that of Jaén, but the association of hidalgos with commercial agriculture appeared here too.

Yet there was a difference even in this respect. The hidalgos of Jaén lived by and large in the towns where they owned and exploited their properties, forming the dominant portion of the local political and economic elites. In Salamanca province, with its myriad small towns and hamlets, this was not the case. Hidalgos clustered in the larger places: the mean population of towns with hidalgos was between 1.6 and 4.2 times as great as the mean population of all towns in the zones. Even if

one considers only towns that reported at least ten vecinos in the catastro, no zone had hidalgos in more than 40 percent of the towns. The zones identified as having all levels of buyers into the market (Types A₁ and A₂ ) had their hidalgos more widely scattered, that is, present in a larger percentage of their towns, than other zones. Nevertheless, one cannot escape the fact that most towns in all parts of the province, especially the smaller places, had no resident hidalgos. If nobles had a morethan-proportionate role in commercial agriculture and in making purchases during disentail, it had to be as absentees.

In Jaén we found that the presence of hidalgo residents was negatively correlated with lay señorío. In Salamanca lay señorío differed in important ways from that in Jaén, and its impact on the economy is far

Map 17.4.
Salamanca Province, Seigneurial Jurisdictions
NOTE : Based on Mateos, Salamanca,  Map 9.

less obvious. Much more of the province was under this type of jurisdiction, as one can observe from Table 17.10 and Map 17.4. Three-quarters of the population in the study area lived under lay señorío (two-thirds if one includes Salamanca city, which was royal).^[13] The only administrative division primarily under royal jurisdiction was the large partido of Salamanca city, while nine smaller partidos were entirely under lay seigneurial control.^[14] The situation was not so clear cut as this, however, because throughout the province there were "exempt towns" (villas eximidas), outside the jurisdiction of the partido within which they lay geographically.^[15] The only zones with a majority of the population

[13] For the whole province the figure was 66.3 percent (Mateos, Salamanca, 39).

[14] Partidos of Alba, Barco de Ávila, El Mirón, Piedrahita, and Salvatierra, all under the Duque de Alba; Béjar under the Duque de Béjar; Ledesma under the Duque de Alburquerque; Miranda under the Conde de Miranda; and Montemayor under the Marqués de Castro Monte.

[15] The situation is spelled out in detail in España dividida en provincias. On the grant of exempt status in the sixteenth century, see Vassberg, Land and Society, 165–69.

under royal jurisdiction were those lying mainly in the partido of Salamanca: SA and SC.

The striking difference between señorío in Jaén and Salamanca was that the former was accompanied in two-thirds of the towns by the ownership of the largest properties in the town. This was not the case in Salamanca; of 419 places with separate catastro returns that were under señorío lego, the señor held the greatest share of real property in only 29 places (7 percent).^[16] All señores hacendados mayores were absentee owners; most of them resided in Madrid.^[17]

The three largest señoríos in the province reveal the nature of señorío here. The señores had the rights of jurisdiction, with the corresponding income from fines and appointments to municipal offices, and the crown at some time in the past had given them the alcabalas, or sales tax, and the tercias reales, the royal two-ninths of the tithes. These sources could produce considerable income, but they were called alienated rights (derechos enajenados ) because the loss was to the crown, not the subjects, since the latter would have had to make the payments in any case.^[18] In fact the crown got very little from these territories, as we saw in the case of El Mirón.

The Duquesa de Alba, who lived in Madrid, was señora of five partidos covering the eastern sector of the province, all zone SI and most of SB and SF, 129 places of the catastro. In only 12 was she the largest owner of real property, including five cotos redondos, in which she was virtually the sole owner. In 11 other places she was hacendada mayor by virtue of the payments she received, and the catastro record of hacendados mayores permits us to appreciate how much of the local economy was absorbed by these payments. In 8 towns it ranged from 7 to 22 percent of the annual income from the land; in 3 cabezas de partido from 29 to 46 percent. These figures are misleading, however, because agriculture provided only part of the income of the towns. Cabezas de partido had many craftsmen and paid more alcabalas as a consequence, while towns in the sierra drew much income from livestock, yet neither of these types of income can be readily calculated from the catastro. The share of the gross town product going to the Duquesa de Alba was therefore less than these percentages, perhaps between 5 and 15 percent.

[16] In sixty-three places the señor was also the hacendado mayor, but in many of these his "hacienda" or property consisted not of land but of the complex feudal dues, taxes, and tithes paid to him. For a discussion of the wide variety of taxes, dues, and rights accruing to señores in Salamanca province, see Mateos, Salamanca, 39–53.

[17] Eight señores, hacendados mayores in fifty-eight towns, lived in Madrid.

[18] AHN, Hac., Salamanca, libro 7479.

The economic impact of the señorío of the Duque de Béjar was about the same. Lord of the partido of Béjar (the major portion of zone SH), he was hacendado mayor in 25 of the 32 towns of the partido, but in 20 of them because of his seigneurial rights. His income ranged from 3 to 76 percent of the income from agriculture in the town. These figures are even less indicative of the duke's true share, because income from livestock and from cloth making was high here. In 3 towns where the duke's income was over 60 percent of the harvests, it came mostly on alcabalas levied on local manufactures.^[19]

The position of the Duque de Alburquerque was very different. Señor of the partido of Ledesma in the northwest (most of zone SD and part of SE), he had jurisdiction over 186 places but was hacendado mayor only in Ledesma, cabeza de partido, because of payments received as señor. Since Ledesma was a town of craftsmen, the weight of his señorío was comparable here to that of Alba and Béjar in their partido capitals. Outside Ledesma his presence was nowhere as onerous as that of the hacendados mayores.

Since seigneurial jurisdiction in Salamanca seldom entailed extensive ownership of real property and it absorbed most of the income that would have gone to the crown otherwise, it could have had little adverse economic effect. It may even have been helpful. Lay señorío in Jaén discouraged the presence of hidalgos. Not so in Salamanca, to judge from Table 17.11. Hidalgos were concentrated in the provincial capital and the cabezas de partido, but outside these places, regions of seigneurial jurisdiction had almost twice as many hidalgos proportionately as regions under the crown. Towns under lay señores were more than twice as likely to have a noble in their midst as royal towns. Since the presence of hidalgos is a good indication of commercialized agriculture, señorío in Salamanca does not appear to have discouraged it. Among the Type A₁ zones, those most commercialized, Zone SD had the highest proportion of nobles and of noble buyers; it also was the zone with the greatest percentage of purchases made with vales reales, our best evidence of market-oriented farming (Tables 17.6, 17.8, and 17.9). It was dominated by the señorío of Alburquerque.

Why were the effects of lay señorío so different in the two provinces? One might expect that, being an older region of Christian Spain, Salamanca received an earlier form of señorío, less burdensome to its subjects. The major jurisdictions, however—Valdecorneja (the partidos of

[19] Baños, Béjar, Candelario.

Piedrahita, El Mirón, and Barco de Avila), Alba de Tormes, Béjar, Miranda del Castañar, and Ledesma, were given to señores in the fourteenth and fifteenth centuries, no earlier than major portions of Jaén.^[20] Although one cannot rule out the influence of earlier practices in the two regions, physical and human geography appear to account for the different nature of señorío. In Salamanca the towns were small, close together, peopled by self-employed peasant farmers. They gave little scope for building up large seigneurial properties in the fashion of the cortijos that could be carved out of the wastes that surrounded the large Andalusian towns. The lords chose to be satisfied with little real property and pushed instead to take over the royal income, the alcabalas and royal share of the tithes. As we shall see in the next chapter, large exploitations did exist in Salamanca, the cotos redondos comprising entire census units. They too were the product of geographic forces: many were in royal lands, and few of those lying in seigneurial regions belonged to their señor. The example of Salamanca suggests that seigneurial assumption of jurisdiction and royal revenues did not drive

[20] Enciclopedia universal ilustrada europeo-americana ("Espasa Calpe"), s.vv. names of cabezas de partidos. For Valdecorneja: Lunas Almeida, Historia de Valdecorneja, 24.

away hidalgos and affect economic development adversely, as did the connection of señorío with large absentee ownership, found in Jaén.^[21]

3

Except for the relationship between hidalguía and commercial agriculture, this static picture of the province has developed few leads toward explaining the extent of involvement of the zones in production for outside markets. Geographic characteristics keep coming up as the most important, indeed the decisive, variable. More enlightening could be the evolution of the zones between the time of the catastro and the first disentail. One cannot, however, proceed to compare the cultivation in the 1750s with that around 1800, as we did in Jaén. Even though the sales permit one to infer the different land uses at the turn of the century, the catastro is unmanageable for this purpose. With over eight hundred towns and villages, each with its individual survey in various volumes, an analysis of the catastro to determine land use was beyond my resources. In addition, Salamanca had no obviously expanding crop, such as olives were in Jaén, that could serve to distinguish areas moving into the national market from those that were not.

I am thus not able to observe changes taking place in the second half of the century. However, because of the large number of towns involved, it proved possible to compare the process of disentail with the conditions prior to its inception or, more correctly, with the conditions revealed by the catastro. This is a relationship that we could infer only indirectly in Jaén.^[22]

In following the process of disentail in the various towns in Part 2, we repeatedly found that the groups that benefited most from the desamortización were those that already had a strong position within the agricultural economy. The process itself seemed to favor such an outcome. Because the sales were at auction, they tended to put the land into the economically strongest hands. In a rural economy one could expect that the groups with the most disposable capital would be those that owned or controlled the land. It can thus be expected that the sales would strengthen existing patterns of control of the land—except, of course, that ecclesiastical owners would lose out. The Salamanca data allow one to test this hypothesis in two different ways.

The first involves the concentration of landowning. If the sales re-

[21] See Appendix R.

[22] An earlier version of the following analysis appeared as Herr, "Vente des propriétés de mainmorte."

inforced existing patterns, the acquisition of land should be concentrated among the buyers in proportion to the concentration already existing. The most obvious test of this proposition would be a comparison of two sets of Gini concentration coefficients (drawn from Lorenz curves) for each of the zones, one for landowning as shown by the catastro and one for the purchases. Neither is available, however, the first because of the prohibitive time that would be needed in the archives, the second because the Gini coefficients of the zones vary substantially according to the way the researcher decides to divide purchases made by more than one buyer among the different buyers (information on how they were in fact divided not being given in the Madrid records).^[23]

Using the kinds of data with which we are familiar, however, one can develop surrogate indexes of concentration that permit a satisfactory test of the proposition. For the index of concentration of ownership at the time of the catastro, I used the product of two statistics. One was the percent of the total property in the zone that was owned at midcentury by the hacendados mayores (the largest property owner in each town). The catastro records in Simancas give the income of the hacendados mayores, and the provincial summaries of the catastro in Madrid provide data for obtaining the total income from land in each zone.^[24]

Although this statistic alone is a simple index of concentration, I decided to multiply it by a second statistic: the ratio between the mean income of the properties of the hacendados mayores and the mean income of all properties. The reason for this step is that the size of not only the largest properties but the others as well determines the extent of concentration. There is more concentration (as measured by a Lorenz curve) where there are one or a few large properties in the midst of many small ones than where all properties are relatively large. The mean income of the properties of the hacendados mayores could be obtained directly, but that of other properties could only be approximated. For this purpose, the income of the properties of the hacendados mayores was subtracted from the total income in the zone, and the remainder was divided among the estimated number of other properties in the zone.^[25]

The index of prior concentration is the product of these two statis-

[23] See Appendix S.

[24] The provincial summary in Madrid: AHN, Hac., Catastro, Salamanca, estados seglar, eclesiástico patrimonial, and eclesiástico, under "letra D" of each volume, gives the number of measures of land in each town belonging to each type of owner, broken down according to the annual income from each measure of land. The calculations are extensive, but made easier with a computer.

[25] See Appendix S.

tics. It can be looked on as the area of a rectangle. The vertical side is the percentage of property in the zone owned by the hacendados mayores, and the base is the ratio of the mean income of the properties of the hacendados mayores to the mean income of the other properties (Table 17.12).

The index of the effect of sales on concentration of landholding is a similar rectangle derived from the sales records. The base of the rectangle also represents a comparison of large to small properties, in this case the ratio of the mean amount spent by the buyers in Level 4 (the largest buyers) to that of the buyers in Level 1 (the smallest buyers). Rather than use the simple ratios, which run from 23 : 1 to 210 : 1, I sub-

stituted their logarithms. The reason is that the range is very high, the largest reading being over nine times the smallest. The other statistic in this index, to be described next, has a range of only 2.6. Multiplying them together would give an excessive weight to the first. The logarithm reduces the range to 1.7, less than that of the other statistic but not so dissimilar.

The vertical side of the rectangle is the percentage of buyers in Level 4 who bought more than one property. At first sight this figure may not seem directly comparable to the percentage of land owned by hacendados mayores used for the earlier index. It was chosen because this index was intended as a measure of change due to the sales, not of the already existing concentration of the lands being sold. Properties being farmed as single units were not ordinarily divided before their sale. Therefore the lowest concentration effect (the greatest increase in the number of landowners through the sales) would result if each property went to a different buyer. The percentage of buyers who bought more than one property is thus a measure of how much concentration was effected above this minimum. It is proper to use only Level 4 buyers in each zone because, smallest in number (they ranged from 3 to 12 percent of all buyers), they had the greatest effect on the pattern of landholding: by definition they accounted for 50 percent of the purchases. The index and its components are given in Table 17.13. (One should stress that even though this index measures the extent to which the sales created new large holdings in the hands of laymen, it is not a measure of the absolute change in concentration. Some religious endowments consisted of various properties that were sold off separately, thus tending to decrease the existing concentration of ownership. The absolute change was the difference between these two contrary effects, which I have not attempted to calculate.)

When the two indexes or sets of rectangles, one of the prior concentration of landowning and the other of the effect of sales on concentration, are compared, they give a high coefficient of correlation, r = .89. Statistically this means that the correlation could have occurred by chance less than once in a hundred cases (it is significant at the 1-percent level). Although the indexes are the products of circuitous calculations, they offer strong support for the validity of the proposition.

Figure 17.1 graphs the result, the regression line of the effect of the sales on the prior concentration. The resulting pattern supports both the division of the zones into types of participation in the market and into geographic regions. Type A₁ zones (SA, SB, SD, SE) fall in the middle of

the pattern, with Type B₁ (SC) far out on its own and Type C (SG, SI) together at the other extreme. Zones in the different geographic regions are even more closely associated, poor plains zones (SC, SD) with high concentration, sierra zones (SG, SH, SI) with low concentration, and zones in the other two geographic regions are also proximate to each other.^[26]

[26] The statistical strength of the correlation depends greatly on the outlier SC. Without it, the correlation coefficient drops to 0.59, and consequently the correlation line swings clockwise. The pattern loses its statistical strength, being valid only at between the 10-percent and 20-percent level. What weakens the correlation is the deviance of Zone SB, which shows a much higher effect of concentration from the sales than the regression would predict.

Figure 17.1.
Salamanca Province, Regression of Concentration
Effect of Sales on Concentration of Prior Ownership

4

The nature of the data makes possible another test of the proposition that the sales at auction strengthened existing patterns of landholding, by comparing the amount of absentee ownership in each zone before the sales with the amount of property bought by nonresidents. We know already that absentee ownership was a strong characteristic of the province.^[27] In the studies of individual towns we observed that around Salamanca city the large blocks of property were in the hands of outsiders and that nonresidents were the biggest buyers, a fact confirmed above for the province as a whole, since we saw that notables were the major buyers and they were largely absentees. Most hacendados mayores at

[27] Absentee ownership is defined as ownership of the property in a place by persons or institutions domiciled outside the place; nonresident buyers similarly are those domiciled outside the place where the property is located.

midcentury were also outsiders (Table 17.14). Some relationship between prior absentee ownership and absentee buying is apparent; the data permit us to test statistically the strength of the correlation.

To obtain an index of prior outside ownership (Table 17.15), I used three figures. The first (Column A) was the proportion of the land in each zone owned by absentee hacendados mayores. It is an accurate statistic of how large a share was owned by certain absentees, namely those who were the largest owner in the town. To get an approximation of absentees among other landowners, I used the ratio of the total value of the land in towns with absentee hacendado mayores to the total value of all the land in the zone (Column B), on the assumption that towns with absentee hacendados mayores would also have a larger share of smaller absentee owners. The third statistic (Column C) is drawn from information on ecclesiastical ownership, since much land belonged to religious institutions and foundations located outside the small places.

It represents an attempt to calculate the percentage of land in the hands of outside ecclesiastical owners. To obtain it, the share of land owned in each zone by ecclesiastical institutions is multiplied by the share of all the land in the zone that was located in towns where the hacendado mayor was an ecclesiastical institution. The final index is a simple mean of these three statistics. One should stress that the index does not represent an estimate of the proportion of land owned by outsiders; its purpose is to permit a comparison among the zones.

The index of sales made to outside buyers also had to be derived indirectly, for the Madrid notaries' deeds of deposit did not usually state the residence of the buyers, although where it was given I noted it. A search of local notarial records unearthed the residence of many of the largest buyers. In addition, unlocated persons who bought property in more than one town were considered residents of the town where they made the most purchases and nonresidents elsewhere. Those persons with names preceded by don not identified as priests, who bought in towns where the census of 1786 showed no hidalgos, were also considered nonresidents. (Not all laymen called don were hidalgos, but hidalgos were probably the most numerous group among them.)^[28] The purchases thus identified as made by outsiders were totaled and divided by the total amount spent in the zone. The result, in Table 17.16, shows that the share of the sales in each zone that appears to have gone to outsiders ranges from 9 to 82 percent.

When the index of prior absentee ownership is compared to that of nonresident buyers, the two give a surprisingly high coefficient of correlation, r = .97. Figure 17.2 illustrates the close relationship between the two, r ² = .94, implying that the amount of land that was bought by outsiders was 94 percent determined by the amount owned by outsiders before the sales.^[29]

Where the land was in nonresident hands (much of it ecclesiastical), potential outside buyers would know about the possibilities, the channels of information were good. They could take over existing leases and tenants from the present ecclesiastical owner; the transactions costs of the purchase were low. Furthermore, where local farmers owned little, they also were likely to have little capital. As in La Mata and Villaverde, they might bid for small collections of arable plots but not for holdings

[28] See above, Chapter 15, section 4.

[29] The correlation coefficient in this case does not depend heavily on one outlying point.

Figure 17.2.
Salamanca Province, Regression of Sales to
Nonresident Buyers on Prior Outside Ownership

of many plots spread over various villages, and certainly not for a coto redondo like Pedrollén. But where the land was in local hands—in the sierras and to a lesser extent in the hilly zones—outsiders knew less about it and the residents had proportionately more wherewithal. Wealthy investors had less information and less interest in these regions, and vecinos were the main buyers.

As one can expect, outside owners and buyers were present in greater numbers where agriculture was oriented toward the national market (Type A₁ and B₁ zones). They used more vales reales than elsewhere to make their purchases. The zones with fewer outside owners and buyers not only paid much more in hard currency, but buyers also bid more against each other—there were more payments in specie above the minimum bid (Table 17.17). Chapter 15 distinguished in both provinces two types of agriculture: market oriented and local. The former was charac-

terized by greater use of vales reales, extensive cultivation, larger participation of hidalgos, and, we can now add in the case of Salamanca province, greater absenteeism. Regions of local agriculture relied more on hard currency, had more intensive farming and less land in the hands of outsiders, and were more socially egalitarian.

Chapter 15 concluded that all areas had some of both kinds of agriculture, indeed that most individuals were involved in both kinds. The more wealthy the individual, the greater his participation in the market was likely to be. What we find in Salamanca is that the distinction is not only social but regional, much more so than in Jaén. The plains around Salamanca City, including the partidos of Ledesma and Alba de Tormes, were strongly marked by commercial agriculture. The sierra zones were much less defined by external contacts. This was true even of Béjar zone (SH), which, because of its livestock and flax, was oriented toward the wider market. Its buyers had comparatively little money and were local people, like those of the sierra zones on either side. In the plain, the poor zones had more outside buyers and less active bidding at all levels than the rich zones. The latter had a more prosperous peasantry that offered an impediment to complete domination from without, peasants who were themselves participants in the greater economy.

The close relationship between prior outside ownership and the extent of purchases by nonresidents, together with the high correlation between prior concentration of ownership and concentration in buying, offers convincing evidence in support of the proposition that in an area where the economy was basically rural, sales of entailed land at auction, as provided by the desamortización of Carlos IV, perpetuated and accentuated existing patterns of landholding; they did not create new ones.

The two provinces do not offer evidence on how the result might differ in an area that was industrializing or whose economy was changing in other ways, but one can predict that the correlation between prior conditions and the redistribution of land would be weaker. Buyers with capital from sources other than agriculture could be expected to accentuate concentration and absentee ownership. The regions around Salamanca city offer some support for this view. The three zones surrounding it were SA, SB, and SC. Figure 17.1 shows that zones SB and SC are above the regression line, indicating that the effect of the sales was greater than the regression predicted. The deviance is especially notable in SB. If the correlation is recalculated without these three zones, the coefficient of correlation, r, rises from .89 to .95 (significant at the 1-percent level). Figure 17.1 shows the new regression line. Salamanca

city had no modern industry, and most of its income came ultimately from the plains around it. The university, the cathedral, and its many parishes and religious orders owned land and received tithes from farmers in this region. The city's merchants dealt in grain. Its craftsmen furnished their products to clerics, señores, hacendados mayores, and merchants who resided in the city and to the vecinos of the surrounding towns, drawing indirectly on the agriculture of the region. But Salamanca was also the capital of the province, receiving some of royal revenues that came from beyond the partido, and its artisans no doubt sold their wares throughout the province, the silver jewelry of the charro costume, for instance. Local grain was shipped to Madrid and the north, and the city's merchants received a share of the profits. The university drew students from throughout Castile.^[30] Its leading citizens thus had available more capital than that produced by local agriculture, and this can explain the greater concentration of purchasing than that predicted by a regression line based on the structure of prior landholding. Its example offers support for the proposition that nonagricultural sources of income could hasten the concentration of landowning in a period of desamortización, provided, of course, that the recipients of such income chose to invest it in agriculture, and especially in commercially oriented farming.

Disentail encouraged the commercialization of agriculture by offering lands to men and women interested in a profitable investment, whatever the source of their capital. In this respect it carried out the wishes of Jovellanos and the royal reformers. Nevertheless, its impact in a country like Spain, where agriculture was by far the major occupation, was determined primarily by existing conditions, which it served to accentuate rather than mitigate. And this, of course, is not what Jovellanos had predicted, at least in the long run.

[30] See Kagan, Students and Society, 202–210, Map 5.4, and Appendix A, Table 4. The latest date for the geographic origin of students given for the University of Salamanca is 1690, but the pattern would be similar in 1800.

Chapter XVIII—
Salamanca Province:
The Despoblados and Alquerías

Because eighteenth-century reformers believed that the Spanish countryside suffered from a lack of population, they found the cause of the agricultural backwardness of the province of Salamanca in its large number of despoblados, a verdict that recent historians have echoed.^[1] Several despoblados have played a role in our story. Royal plans to repopulate them led to the migration of settlers from La Mata to the despoblado of Narros de Valdunciel. The vecinos of Villaverde did not move their residence, but they broke grazing land in the neighboring despoblados of La Cañada and La Cañadilla, increasing their harvests in this way. Sixty percent of Pedrollén was sold in the disentail, but it had been and remained a single exploitation, rented by one tenant, so far as one can tell from the census returns.^[2]

Despoblados resembled the cortijos of Andalusia, another abomination of the reformers, but they differed in essential ways. Pedrollén, although a single exploitation, had several owners and the shares could be sold separately. Many despoblados, legally single properties called cotos redondos or términos redondos, were owned in shares by various people, but I have found no such case of a cortijo. The latter were carved out of vacant municipal land and may have had their origin before the Reconquista. The despoblados were the result of a historical process since the

[1] See García Zarza, Despoblados, 88 and passim; and Domínguez Ortiz, "Ruina de la aldea," 110. Domínguez Ortiz does not consider specifically the case of Salamanca.

[2] Although Pedrollén was called an alquería in the official index (nomenclátor ) of 1789, it was listed as a despoblado in 1769 (García Zarza, Despoblados, 146).

Reconquista that removed the inhabitants from an established village, or so at least the name implied. The Christian resettlement of Salamanca took place after the middle of the eleventh century, apparently in the form of many small villages and hamlets placed in close proximity. A despoblado arose when the population of a village was reduced to one or two families or disappeared altogether, but even when all fixed population vanished, the name and administrative unit persisted. Cortijos were not separate political or census units.

Why despoblados had declined is not clear. The times of greatest loss of population would appear to be the hundred years following the Black Death and the seventeenth century, periods of plague and demographic crisis throughout western Europe. Two authors who have studied similar deserted villages in France argue, however, that plagues alone cannot account for the complete disappearance of villages. The destruction and military levies of wartime played their part in driving away the inhabitants. Permanent depopulation did not result, however, unless other causes prevented resettlement once the hardships had passed. One might be adverse physical conditions, another the desire of a landowner to build up his estate. The smaller the hamlet, the more likely it was to disappear, but it was rare for a place to be abandoned completely. Some farmstead, some group of people remained to keep alive its name.^[3]

The story in Salamanca is similar, suggesting that the causes were not unique to this region or to Castile. Antonio Domínguez Ortiz believes that small settlements disappeared in the centuries after the reconquest because the insecurity of the open countryside drove the peasants to seek safer and more healthy locations. After 1250 there was a great demand for settlers in Andalusia, and many may have been attracted by this opportunity.^[4] Eugenio García Zarza, who has studied the despoblados of Salamanca most closely, points to the action of noble lords, desirous of creating large estates as a basis for political power during the struggles of aristocratic bands in the fifteenth century. They forcibly moved peasants to larger towns, a procedure renewed in the seventeenth and early eighteenth centuries, when the crown was again weak. Seventeenth-century wars with Portugal, destructive and demanding in money,

[3] Pesez and Le Roy Ladurie, "Deserted Villages," 95. A fuller version of this study appears as Pesez and Le Roy Ladurie, "Le cas français." This volume includes Cabrillana, "Villages désertés en Espagne," 461–512, which attempts to identify the despoblados in much of Spain but does not cover the territory of the crown of Castile north of the central sierras. Anes, Crisis, 181 n. 29, questions its figures.

[4] Domínguez Ortiz, "Ruina de la aldea," 106–7. On migration to the south, see MacKay, Spain in the Middle Ages, 70–72.

and the increasing tax burdens, which hit the remaining peasants ever more heavily as the others fled or died, completed the desertion of the villages.^[5] Curiously, these historians make no mention of the Black Death as a contributing cause to the depopulation, although students of medieval France and England have stressed its central role in the redistribution of population. Plagues of the fourteenth and fifteenth centuries are known to have laid waste Catalonia, Aragon, and Navarre.^[6] In 1350 the Black Death had spread so widely in Castile that Alfonso XI wrote to Pedro IV of Aragon, expressing his concern about the loss of population and the number of his subjects going to Rome to obtain a papal indulgence.^[7] The following year, the Cortes of Castile heard complaints that the great mortality had raised the price of agricultural labor to such heights that owners were being forced to abandon cultivation.^[8] The plague may well have been the event that triggered the early desertion of the smaller places in Salamanca.

What is unique about Salamanca is the large number of despoblados that appeared, far above anything found in France or the rest of Spain. The catastro identified 113 in the study area, while the nomenclátor of 1789 raised the number to 197, plus another 93 in the partido of Ciudad Rodrigo.^[9] Most lay in the poor plains southwest and west of Salamanca city, that is, on the large property side of the Salamanca-Albacete line, but a number were also in the rich plains north and east of the city. Almost indistinguishable from the despoblados were places called alquerías, large properties tended by one or several families. Although this name, derived from the Arabic word for a small settlement, does not imply that they were ever larger, the available censuses show that many had had more people in the past.^[10] The catastro does not identify alquerías as such, but the nomenclátor names 151 of them in the study area, located in the same regions as the despoblados, and 30 more in the partido of Ciudad Rodrigo.

[5] García Zarza, Despoblados, 36–51, 65–82.

[6] Vilar, Catalogne 1 : 464–66; Vicens Vives, Manual, 163–64; Cabrillana, "Villages désertés en Espagne," 493–95, 504.

[7] Letter referred to in Pedro IV of Aragon to Alfonso XI of Castile, 30 Mar. 1350, in López de Meneses, Documentos acerca de la peste, 377–78.

[8] García Fernández, "Champs ouverts," 707–8.

[9] AHN, Hac., libro 7476, letra D: Partido of Salamanca, 78; of Alba, 11; of Ledesma, 24. García Zarza, Despoblados, 149–51, lists these.

[10] Instituto geografico y estadístico, Nomenclátor . . . de 1900 2 : 356, defines an alquería in Salamanca as "a rural house inhabited by the owners or tenants of surrounding fields used for the cultivation of cereals."

Why this concentration of deserted villages in Salamanca? The activities of noble estate builders and the destruction of war and taxes were not peculiar to this region. One must look for a scenario that fits the local situation. From the ninth to the eleventh century, the region was a no-man's-land between the Christian monarch of León and the Muslim states south of the central sierras and was virtually without permanent residents. After the reconquest of Toledo in 1085, the future province of Salamanca came permanently under Leonese rule. Salamanca city was fortified and given a cathedral, and the smaller places, Alba de Tormes and Ledesma, also became strongholds. Rural settlement was done in small nuclei, villages and hamlets of only a few families. Some settlers came as organized groups originating in a town in the north, others had been collected under the leadership of some enterprising individual of higher social status. Frequently the name of the new village reflected the origin of the settlers or the person who founded it. Such toponymic evidence indicates that the new inhabitants came from all over: Galicia, León, the Basque lands, and Navarre to the north as well as a large number of Mozarabes (Christians who had been living in Muslim territories) and Moriscos (converted Muslims) from the south.^[11]

Settlement tended to follow the old Roman road from León to Salamanca and south through Béjar to Mérida, spreading out to the east and west from it. The rich Armuña plain north of Salamanca city was repopulated before the mid-twelfth century; by 1200 its towns were substantial nuclei. The poorer zone south of the city filled up a generation or two later, and the southern sierras between about 1190 and 1230. One can picture these early hamlets as a half-dozen or so houses and a few sheds around a small unpretentious church. The nearby arable had been distributed in fields to the settlers in a reparto, while the common pastures and woodland lay beyond. It is possible that some places were private estates with only a few permanent hands, protoalquerías, especially those in the poor plains bearing the name of the founder. Many villages were required to make payments to the cathe-

[11] Four places named Naharros (Narros), a Naharillos, and two Naharra reflect the immigration from Navarre, while three Gallegos, two Galleguillos, and a Nava Gallega, that from Galicia (six of these thirteen places were despoblados or alquerías in 1789). There was also a Córdoba, a Cordobilla, and a Mozárvez (Mozárabes), a Morisco, and a Castellanos de Morisco. Founders' names appear in Aldehuela de los Guzmanes, Gutierro Velasco, Martín Pérez, Andrés Bueno, Gimén Gómez, Velasco Muñoz, and so forth (all of these being despoblados in the eighteenth century). See Julio González, "Repoblación de la 'Extremadura,'" 241–45, and Julio González, "Reconquista y repoblación," 176–81. These are the sources for this and the next paragraph.

dral of Salamanca, but jurisdiction lay with the council of one of the fortified centers, Salamanca, Alba de Tormes, or Ledesma; the basis was already established for the later partidos. Except for a few places under ecclesiastical jurisdiction, señorío did not come until later; the cities and villages were royal.

When hardship came in the form of plague and civil strife in the fourteenth century and later, those settlements located where the natural conditions were least favorable, notably in the thin topsoils south and west of Salamanca city, suffered most. The villagers still alive in the smallest villages, places with less than ten families, some as few as two, chose to go to bigger towns or were forced to do so by local magnates who desired unobstructed use of the fields. But the places did not disappear from official records. The owners chose to preserve their names and identity so that they would not be taken over by neighboring towns as common wastes and pasture.^[12] These places had never flourished under farming, and the rise of migratory and sedentary sheep raising after the fourteenth century made them profitable as pastures, turning them into permanent despoblados.

Angel Cabo Alonso, the leading historical geographer of the province, emphasizes the importance of geographic conditions in explaining the appearance of despoblados. They were found, he says, in the plains where the underlying rock formation lay closest to the surface of the sedimentary deposits, leaving little soil cover for crops. He theorizes that after the reconquest the larger settlements were along the rivers and streams, while the future despoblados were smaller hamlets that appeared in the hilly interfluvial spaces and never became populous villages.^[13] One needs a case-by-case study to test this hypothesis. It undoubtedly applies to many places, but not all, because there are authenticated accounts of forcible removal of villagers in the fifteenth century,^[14] while the first census available, from 1534, still shows numerous vecinos in many places that would be despoblados in the eighteenth century.^[15] What seems indisputable is the importance of the physical characteristics of the plains, first in attracting early settlers into numerous small hamlets and then in proving unsuited to the permanent arable culture needed to support so many nuclei. The settlers had set up life in the

[12] See García Zarza, Despoblados, 28–29.

[13] Cabo Alonso, "Antecedentes históricos," 65–66. Anes, Crisis, 181–86, also questions the belief that all despoblados had once been populated centers.

[14] García Zarza, Despoblados, 36–48.

[15] Tomás González, Censo . . . siglo XVI, Appendix 5.

way familiar to them, only to discover too late that the physical reality would not accept it here. Theirs was not the last such experience, witness the American Dust Bowl of the 1930s.

2

The data collected for this study permit one to look at the despoblados in several new ways. Most of the following discussion is limited to the despoblados for which there are individual surveys of the catastro, but, unlike previous studies, it also considers the alquerías. Table 18.1 shows what percentage of the catastro units in each zone fell into these two categories. Those with the greatest concentration of despoblados and alquerías were the four plains zones, SA to SD, and the hilly zone SF that juts between zones SB and SC. The following analysis is limited to these five zones.

Most of the despoblados were smaller in area than the units that had survived as towns; in three of the five zones they had a mean size less than half that of the regular places. According to the returns of the catastro, their mean size ranged from 490 fanegas (about 220 hectares) in

zone SD (Ledesma) to 1,150 fanegas (about 510 hectares) in zone SB (Alba de Tormes; Table 18.2). In fact 43 percent of them contained less than 250 hectares, a figure that has been considered the minimum size of a large property or latifundio in twentieth-century Spain.^[16] Alquerías tended to be somewhat larger, although still smaller than settled towns. A community located in such a place, especially where the land was poor, was at a decided disadvantage when faced with climatic catastrophes or human violence.

The evidence from censuses permits one to assign the decline in population of most despoblados and alquerías to one of three periods: prior to the census of vecinos of 1534, between 1534 and the census of vecinos of 1712, and the period between 1712 and 1826, for which we have the catastro (ca. 1751–53), the census of 1786, and Miñano's dictionary with figures for 1826. Table 18.3 shows the distribution among these three periods of those places for which some census data are available. Since the first census is for 1534, we can know only whether a place was at its demographic low point by that date. We do not know that it had ever been larger, and in view of the doubts of Cabo Alonso on this question, one should leave it open. Some certainly had been larger at some previous time.

[16] See Malefakis, Agrarian Reform, 21, Map 1.