1The frequent use of the expression “old demographic regime” in contrast to the “modern” regime, tends to mask the broad range of situations and dynamics that existed before the demographic transition. In this article, Guido Alfani illustrates this diversity using 164 annual baptismal series covering Northern Italy from 1560 to 1628, evenly distributed in terms of environment type, geographical region, and economic and social context. Based on a single indicator, the data have limitations that the author himself points out. But they are largely offset by the broad scope of the sample, which covers contrasting situations and highlights the large variety of behaviours and processes at play among the populations of the past.

2Studies in historical demography based on time series have often given the impression that it is only possible to write an account of Italian population history starting from the middle of the seventeenth century. However, if we consider research based on sources other than those traditionally used by demographers, we come upon excellent studies on the Middle Ages characterized by particular methods. Anyway, both types of study neglect almost entirely the fifteenth and the sixteenth centuries, which are decisive for the interpretation of Italian economic and social history. Clearly, there are objective limits to the documentation available and this complicates the task of applying consolidated methods of historical demography to populations in the Early Modern Age. However, this is not sufficient to account for a more generalized omission, especially considering that the history of the sixteenth century in Italy is full of unresolved questions as far as demographic trends and their relationship with the economic cycle are concerned.

3There are currently two different estimates of the size of the Italian population in the course of the century, to which reference is made from time to time in an uncritical manner. One is that of K.J. Beloch (1937-1961), and the other that of C.M. Cipolla (who presents it as a revision of the former) which appeared for the first time in the famous volume “Population in History” (Glass et al., 1965); both have been cited in various other publications [1]. Although the estimates at given points in time may seem very similar, on closer inspection, clear differences emerge which have been interpreted as the consequence of a different vision of the mechanisms of population at work. Levi (1991) noted that the Italian population estimates put forward by Cipolla for Northern Italy between the sixteenth and the eighteenth centuries were significantly lower than those of Beloch [2]. He also argued that the literature on which Cipolla based his estimates seemed inadequate for the purpose [3]; consequently he felt that the revisions were unjustified. Levi hypothesized that the two groups of estimates implied different interpretations of the ways in which cereal production can influence demographic trends. In particular, Cipolla’s estimates betray his conviction that the relationship between population and resources is of the Malthusian [4] type, while those of Beloch fit in with a Boserupian model [5]; the two theories imply a very different idea of the relationships between population, resources and technology (as explained in Section 5). Levi raised an important question, i.e. the nature of the connection between population dynamics and crop regimes. Even if difficult to tackle for the earlier periods, my task here will be to investigate this question, while also taking into account the problems posed by innovations in agricultural techniques introduced in some Italian regions during the sixteenth century. To this end, after having briefly examined the characteristics of the data set and discussed some methodological issues (Section I), I will evaluate the relevance of variables that may have caused different demographic dynamics, such as the different population patterns of the mountains and the Po plain (Section II), or of cities and countryside (Section III). I will then examine in more detail the countryside of the Po plain where the aforementioned agricultural innovations took place (Section IV) and, on the basis of the evidence, I shall present some hypotheses on the population dynamics at work and on the demographic role of technological innovation (Section V). In the last Section I will reconstruct the general trend of births in Northern Italy between sixteenth and early seventeenth centuries to check which set of population estimates (that of Beloch or that of Cipolla) is more consistent with my data.

I – Sources and methods

4This paper presents key results obtained as part of a research project whose main aim was to catalogue and collect sources, elaborate methods, and discuss reasons for resuming investigations on sixteenth century demographic dynamics and their influence on the economy (Alfani, 2004). The need to delimit the research field led me to concentrate on Northern Italy, an area characterized at the time by wide geographic, social, political and institutional variety. From a chronological point of view, the oldest data at my disposal date back to the second half of the fifteenth century and I decided to stop my investigation in 1629, eve of the most serious epidemic that Northern Italy had experienced since the return of the Black Death to the continent in 1348.

5With regard to the availability of demographic sources, the sixteenth century is a moment of transition. While the Middle Ages are characterized by a lack of sources, and in particular of those allowing time series analysis, during the sixteenth century a type of serial data fundamental for research in historical demography – i.e. baptismal series – became available in much greater abundance than commonly believed. These series constitute a close approximation to those of births (Wrigley and Schofield, 1981) and can be reconstructed from baptismal records (libri baptizatorum), for periods when other important series, such as those of deaths and marriages [6], are almost completely absent. Only starting from seventeenth century did all three types of series become widely available; this is the reason why research into Modern Age population dynamics usually does not start before 1650 or, at the earliest, 1600.

6Despite the methodological problems raised, the birth data available for the sixteenth century are so abundant that I think they are worth using, especially considering that – for Italy at least – these early parish registers are of excellent quality, often better than those found in later centuries. I was able to collect 164 series of baptisms celebrated in Northern Italy starting from the fifteenth or sixteenth century, characterized by a balanced distribution for geographic area, settlement pattern and economic and social context [7]. The geographic and chronological composition of the database are given in the Appendix. For reasons of space, it is impossible here to give full information about the sources from which I have drawn each series; in this regard see Alfani (2004). Suffice it to note that 41.5% of the series have been published in books or journal articles, 44.5% come from unpublished post-graduate theses held in a variety of Italian university archives, 7.3% are unpublished series lent to me by their authors, and 6.7% (i.e. 11) have been reconstructed by myself [8].

7Having to work exclusively on baptismal series, I was obliged to use methods capable of fully exploiting the possibilities offered by demographic data based solely on births, and I have discussed this matter elsewhere (Alfani, 2003, 2004). Here, suffice it to say that I aimed to exploit the strong points of the available data (in particular, the abundance of demographic “points of observation” i.e. the local series), in an attempt to get round the obstacles and to extract as much as possible from sources which are useful indicators not only of demographic dynamics but also of economic, social and cultural phenomena.

8This approach is well-known to those who are familiar with the great theses of economic and social history of the 1960s and 1970s, in particular those of the French school known as Annales. In the early 1960s, Baehrel (1961) showed how, for the purposes of an economic historian of the Early Modern period, the pattern of births is an acceptable proxy for complete population data [9]. Indeed, using birth data to evaluate general population trends is not new, even to those historical demographers more oriented towards statistics, when faced with a serious lack of documentary information about deaths (for instance, Perez Moreda, 1999). The alternative would be to lose entire centuries to historical demography. Recently, another kind of series, i.e. grain prices, has been used to study the relationship between macro-economic trends and demographic responses (Breschi et al., 2002). However, I shall not use grain price series here because, while available since the second half of the sixteenth century at least, they are much less abundant than birth series and almost exclusively concern urban environments. Furthermore, their use raises theoretical problems that would unnecessarily complicate my analysis.

9Using birth trends as a proxy for population change means hypothesizing that birth rates are constant, or at least that they change little and slowly in time. While this is usually the case for pre-transitional long-term trends, many studies have revealed major short-term fluctuations in the birth rate (e.g. Wrigley and Schofield, 1981). Given the sources available, this problem cannot be completely resolved, but can be somewhat circumscribed by using the comparative method and focusing data analysis and interpretation on short periods, i.e. by asking questions such as: at the moment of the terrible famine that struck much of Europe in the 1590s, which communities suffered most? What kind of environment proved most sensitive to the crisis, probably suffering from a bad population-resources ratio? and so on. Replacing original indexes with series of moving averages is also of some help.

10When considering a famine for example, it is obvious that a reduction in births can be caused by many different factors: a diminished population (because people died from hunger or emigrated), reduced fertility etc. However, if we are interested in the balance of population and resources, this is a minor issue because a (comparatively) greater or smaller reduction in births, whatever its causes, suggests a (comparatively) worse or better balance, the reaching of the carrying capacity or the availability of surplus resources (or, as I shall explain, of a range of resources).

11While much of the analysis focuses on short periods and especially on years of crisis, sometimes I shall analyse long-term trends [10]. In this case, it is possible that the series were substantially influenced by a change in birth rates, of which we have no proof as yet, but given the status of our knowledge this is unavoidable. Furthermore, as shall be seen, serial data suggest that the available estimates relative to a particular year are not especially reliable, so I am convinced that by using birth series we can significantly improve our knowledge of sixteenth century Italian demographic trends, even if much will remain hypothetical and will require further research.

12The aggregate series I shall use have been reconstructed from the local series using a statistical method developed for analogous purposes, also used to process the series found in Histoire de la Population Française (Dupâquier et al., 1988). The method was introduced and discussed by Biraben et al. (1980, 1982, 1985) [11]. It can be used to assess the problem of series which progressively disappear as one moves back towards earlier periods (a particularly important problem in my case, given that the majority of series begin during the period under examination). When possible, small gaps in the original local series (up to a maximum of 6 years) have been closed before aggregation by using the method developed by Scalone (Scalone, 2001; Del Panta et al., 2002) for solving problems of “partial completeness” of a data set (the gap is closed by interpolation based on data concerning neighbouring communities or communities having similar characteristics).

13When building aggregate series, I weighted the local series to give the maximum possible importance to each single place as a demographic point of observation. In practice, I first replaced the real values of each local series with indexes based on the average of the years 1610-1614 in such a manner that the numerically more consistent series were not given a greater weight than the others. I then calculated the simple average of the indexed data, giving a weight of one to each series. Some empirical tests proved that this method, when applied on my data, did not distort results as compared to the alternative one of using “natural” (non-indexed) data, while guaranteeing better territorial and environmental representativity.

II – Plain, coast or mountain: the demographic weight of environmental factors

14In the case of “old regime” populations, constantly (or almost so) in precarious equilibrium with the available resources, some environmental factors play a primary role in determining the sustainable demographic load, i.e. the carrying capacity of the system. It would be wrong, nevertheless, to believe that rigid, unchanging relationships exist over time between population, resources and territory. The reflections of Boserup (1981), who stressed the role of demographic pressure in stimulating the introduction of micro-innovations capable of increasing food output, as confirmed by recent studies on the population of Alpine areas (Mathieu, 2000; Fornasin and Zannini, 2002), lead to a distrust of “naïve Malthusianism”, i.e. a fatalistic view that populations, being unable to check their growth, are bound to periodically face disaster. This view, while still widespread among social and economic historians, has actually been moderated and largely redefined by recent Malthusian approaches (e.g. Livi Bacci, 1991). I shall return to this point later [12].

15To investigate the weight of environmental factors, I will make use of some aggregate series created using the most complete and reliable local series. The general principle I follow is that of territorial representativity. One of the series is related to the Alps, one to the Tyrrhenian and Adriatic coasts and one to the Po plain [13]. In this phase I excluded urban series from the analysis so that I could concentrate on smaller places for which the importance of environmental factors is easier to determine. The results are shown in Figure 1.

Figure 1

Parishes of lowland, mountain and coastal areas in comparison, Northern Italy, 1560-1628. Averages of the indexes of baptisms, base 100 = average of years 1610-1614

Parishes of lowland, mountain and coastal areas in comparison, Northern Italy, 1560-1628. Averages of the indexes of baptisms, base 100 = average of years 1610-1614

16On a first reading, the three curves show similar trends, though significant differences can be observed [14]. First, both the coastal areas and the inland plains are severely stricken, and in similar measure, by the 1590-1593 famine (the worst Italian episode of the crisis that hit Europe in the 1590s (Clark, 1985)), particularly in its acute phase (1591-1592), while mountain settlements seem to be much less affected. The other important crisis that struck Northern Italy in this period, i.e. the plague of 1576-1577, does not have an evident impact on the series. This is mainly because plague does not have a geographic spread comparable to that of famine. With the exception of the most serious outbreaks (such as that of 1629-1630), plague strikes some places and spares others, so is better appreciated at the level of local instead of aggregate series. Second, in the 1620s, characterized by food shortages which, in a certain sense, are a prelude to the pandemic of 1629-1630, the communities on the plain seem to suffer more than those in the mountains or on the coast.

17The most interesting comparison is, without doubt, that between the mountain series and the other two. As already mentioned, the former is subject to smaller fluctuations; moreover, during the years 1591-1592 (the peak of the famine) the drop in births is limited and the voids that it creates are quickly refilled. For a better appreciation of the trend, it is useful to resort to five-year moving averages (Figure 2).

Figure 2

Parishes of lowland, mountain and coastal areas in comparison, Northern Italy 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

Parishes of lowland, mountain and coastal areas in comparison, Northern Italy 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

18The famine of 1590-1593 was caused by several consecutive years of bad weather which directly damaged the crops and also favoured crop diseases such as blight, which spread from Italy in the following years to France and other parts of Europe. The crisis was very acute in coastal areas and on the inland plain, but much less serious in the mountains. The fact that the drop in births in these areas is noted only in 1592, at the peak of the crisis, while in the previous year births were stationary, is significant. One can hypothesize that a lesser dependence on the grain harvest enabled the population to resist the famine for a certain period but not indefinitely. It is also probable that mountain dwellers suffered less than others from typical complications of famine such as typhus epidemics, more severe in crowded areas such as cities and the rich lowland countryside. The famine was also acute on the coast, suggesting that fishing was unable at that time to make up for a severe grain shortage, at least not in areas as densely populated as the Ligurian and Northern Adriatic coasts [15].

19From a demographic point of view, the 1590-1593 famine was by far the most important event in the period under consideration. To better appreciate its extent, it is useful to refer to a map on which the presence of the crisis has been hypothesized in many places. The reference localities are those which were certainly hit by the famine, as confirmed by historical research (Map). In these localities I found an average reduction in births of 44% [16] and on the basis of this I formulated hypotheses about the occurrence of the crisis elsewhere [17].

Map

The 1590-1593 famine in Northern Italy

The 1590-1593 famine in Northern Italy

20The map well illustrates the generalized extension and the gravity of the famine: in those years, finding wheat in Italy must have seemed well-nigh impossible. Looking at the few blank circles corresponding to places where the reduction in births was modest (below 4%) or non-existent, we note that the places spared are concentrated at the edge of the Po plain, along the Alpine arc and in certain areas of the Ligurian Apennines. In almost every case, these are mountain villages, especially if we ignore some parishes of the suburbs of Bologna where the evaluations may have been distorted by the small number of annual baptisms (the lower the number of baptisms per year, the greater the influence of random factors on the demographic trend).

21The geographic position explains why these places escaped the crisis: the relationship with the land established by mountain dwellers is very different from that of plain dwellers, being characterized by a lower population density and a different crop regime. The second point is decisive, since the 1591-1592 crisis was essentially due to a grain shortage. Mountain populations relied much more on animal husbandry, horticulture, and specialized crops than the populations in the lowlands and obviously suffered less from a lack of grain. Furthermore, tree crops typical of low and mid-mountain areas were also an important food source, above all the chestnut, which is very resistant to adverse climatic factors and whose fruits are highly nutritious.

22Returning to the graphs, looking at the series as a whole we note that the trend in the mountainous areas is towards slow, but constant growth which leads to an increase in births of approximately 35-40% in the period between 1565 and 1605. This tendency does not seem consistent with Braudel’s interpretation of the mountain as a “factory producing men”, a place closed in on itself, condemned to remain in a state of underdevelopment “without history” characterized by scarce resources and harsh living conditions, making it an exporter of manpower to the lowlands and the city (Braudel, 1949). My observations, on the other hand, correspond very well to some recent studies of Alpine demography, which conclude that over the long term, the history of the Alps has been governed by a pattern of slow but continuous growth, much more than by a process of expulsion of a population surplus (Mathieu, 2000; Fornasin and Zannini, 2002). Crucial to achieving such a result was the early establishment in the mountains of a “low pressure” demographic regime, characterized by moderate birth and mortality rates, not greater than 30-35 per thousand (Viazzo, 1990). This phenomenon has been described many times in research covering the period from the first years of the eighteenth century (e.g. Albera et al., 1988) but the few available studies for earlier periods suggest that it would be possible to backdate it at least to the second half of the sixteenth century (Maggi, 2002). Its interpretation still has some unclear aspects, but many elements have been identified, such as the regulatory role of nuptiality, the forms of organization of the family, the way the work force was used (with large-scale seasonal migration), etc [18].

23In the case of the mountains, therefore, environmental peculiarities require us to focus particular attention on the relationship with the available resources which favours the establishment of social customs that reduce the birth rate [19]. The result is a more gradual, but steadier demographic growth, which, along with a diet, crop regime and settlement pattern different from that of the plain, guarantees less vulnerability to subsistence crises (at least those caused by a shortage of grain), and mitigates their effects. My data support such conclusions, confirming the applicability of this model to the sixteenth century.

24On the whole, the series presented suggest that in the first two decades of the seventeenth century the population reaches a new ceiling on the plain as well as on the coast and in the mountains: as revealed by the flat lines in the graphs, there is a threshold that the population is not able to cross. Recently it has been suggested that this ceiling represents the maximum carrying capacity of Italy from the times of the Roman Empires. The population fell below this level during the early Middle Ages, reached it again around 1300 then fell below again due to the plague outbreak in 1347-1348, in the same way as sixteenth century growth was cancelled out by the plagues of 1629-1630 and 1656-1657 (Lo Cascio and Malanima, 2005), but my data do not allow me to judge such a hypothesis. What I can say is that the population ceiling reached at the beginning of seventeenth century does not represent a situation of “stable” equilibrium everywhere. After the recovery from the 1591-1592 subsistence crisis, there is a risky reduction in resources with respect to population, which seems to affect the lowlands more seriously than other areas. After a decline in the years 1618-1623, which also concerned the coastal areas (in that period there were frequent food shortages caused by bad weather conditions) the coastal areas recovered quickly while the plains seemed unable to get out of the negative spiral, with a consequent erosion of much of the recovery achieved at the turn of the century. In fact, the crisis seems to have hit certain types of settlement models and crop regimes, based on cereal growing. We shall return to this point later [20].

III – Cities and countryside

25In the Modern Age, cities and countryside are closely linked by complex social, economic, institutional and demographic relationships. Historiography provides us with a consolidated image of the city as a “man eater”, because of its high mortality. In the cities, growth closely depends on the force of attraction exerted on rural populations in the plains, the hilly areas and the mountains. I now propose to examine whether, in sixteenth century Northern Italy, the urban settlement context was characterized by a birth trend different to that of non-urban habitats, the countryside in particular.

26Many believe that, for the cities in general, the Early Modern Age was a time of crisis (e.g. Braudel, 1949). In sixteenth century Northern Italy, this judgment is not appropriate. Indeed, thanks to the political fragmentation of the territory, that century was often a period of great vitality for the urban centres, many of which, recently promoted to the rank of capitals, began to grow vigorously (Alfani, 2006). Things changed in the following century, and numerous local studies have highlighted a relative decline in Northern Italy over that period, due mainly to the crisis of the urban economy (Belfanti, 1990).

27The identification of the status of “city” is a particularly delicate matter in the case of Northern Italy: due to the high level of political fragmentation and other factors, many places performed typically urban functions, but had a population of modest dimensions [21]. A definition of city must take into consideration legal, functional and social factors as well as questions of size; at the same time, it must not be so complex as to lose all operational utility. In the analyses that follow, although the local series have been aggregated on the basis of dimensional criteria, the choice of which ones to include also considered the above factors.

28It is evident that cities like Ivrea (4,467 inhabitants in 1612) and Venice (148,637 inhabitants in 1586) were incomparably different places, with very different provisioning problems. Rather than building a single generic urban series, I have therefore introduced further subdivisions, corresponding to significant size bands [22]: 4,000-7,000 inhabitants (the small cities so abundant in Northern Italy), 7,001-12,000 inhabitants, 12,001-20,000, and finally the cities of more than 20,000 inhabitants, the metropolises of the era. Figure 3 shows the demographic trends of the centres belonging to each size band.

Figure 3

The movement of births in cities, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

The movement of births in cities, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

29There are two moments in which the size of the urban centres seems to influence the course of the series. First, the initial phase: comparing the level of births after the peace of Cateau Cambrésis (1559) with that of the first decade of the seventeenth century, it would appear that the bigger the urban population, the lower the starting point. Second, during the famine of the early 1590s, the drop in births is more pronounced in the bigger cities.

30As regards the first point, one can suppose that the bigger cities suffered more from the negative situation of the first half of the century, during which Italy was a theatre of war in the conflict between France and Spain. Not only were big cities the main targets of military action, but it was also more difficult for their provisioning authorities to maintain food supply in times of disruption of trade, etc. Thus, when peace returned, they found themselves in conditions of relative depopulation [23]. It is also true that bigger cities have a notable potential for growth. In a fifty-year period, cities with more than 20,000 inhabitants registered increases in births of about 60%, against about 16% in cities in the 7,001-12,000 size band and 13% in those in the 4,000-7,000 band. For cities with 12,001-20,000 inhabitants, the only possible comparison is with the early 1570s, when the increase was about 29% [24].

31Figures of this kind certainly do not correspond to a picture of urban decline. It seems, rather, that the cities, and the bigger ones in particular, were experiencing a phase of tumultuous expansion, reaching a high population level at the beginning of the new century and not experiencing crisis until the 1620s.

32Such results, nevertheless, are not consistent with the most accredited estimates of the state of the population in the principal cities, which give more modest increases or even slight decreases [25]. The remarkable degree of concordance of the series collected by myself convinced me that, at least in this case, they are more reliable than the population estimates used so far.

33As regards the 1590-1593 crisis, the damage seems to become more serious as the size of the city increases. Comparing the average level reached in the most negative phase (1591-1592) with that of the years 1585-1589, I found a drop in births of about 29% in cities with more than 20,000 inhabitants, and of 39.5%, 28% and 24% in those of the bands 12,001-20,000, 7,001-12,000 and 4,000-7,000, respectively.

34My data suggest that the pattern of births is altogether more dynamic in the big cities, both in phases of growth and of crisis. Given the rates observed, it is likely that growth is due to the influx of populations from outside: from the countryside, from marginal areas, perhaps from smaller cities, certainly from those former capitals of which there are so many in sixteenth century Italy. However, while the metropolises, those colossal “man-eaters”, attract large numbers of immigrants, they are also quick to reject them. During the crisis, when it is clear that the provisioning authorities are unable to cope with the shortage of resources, the ranks of the refugees are swollen by the expulsion of the forestieri (outsiders) from the cities, who often try to return to their places of origin. As we shall see in the comparison with the countryside, the rate of re-entry also has the effect of prolonging the low level of total births.

35Figure 4 compares the series related to the countryside areas on the plain, already used previously, with an urban series that has approximately the same territorial representativity. To build the urban series only cities with more than 7,000 inhabitants are considered. The series are interpolated with the five-year moving averages.

Figure 4

City and country comparison, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

City and country comparison, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

36Although, in the acute phase of the crisis (1591-1592), the countryside areas of the Po plain reached a much more pronounced negative peak than the cities, it is also true that, in the latter, the depression phase began earlier, and ended later. This depends both on the “food reserves cycle” [26], and on the rate of re-entry. In fact, when the crisis has passed, those who fled the city linger on in the places where they found shelter, slowing down the restoration of the “normal” levels of births. Apart from these details, the two series show a good degree of concordance both as regards the trends, and the levels reached.

37In reality, to better compare urban and rural demographic trends we should control more precisely the geographical areas actually being compared [27]. For this purpose, a comparison of birth trends in the cities with those of their contado (surrounding countryside) has proved particularly useful. For reasons of brevity, I will not report the results obtained here. I note simply that the examined cases suggest a remarkable conclusion: at times of relative decline in the urban centre, we systematically find a demographic increase in the contado, thanks to a reduction in the direct migratory flows towards the city from the surrounding countryside. Analogous phenomena have been observed in other areas and in other eras, but for the sixteenth century they have rarely been studied. Such circumstances assume great importance when we consider that, until now, all available evaluations of population trends during the century have been based almost exclusively upon data of urban origin.

IV – Which countryside? Demographic dynamics, settlement models and crop regimes

38Up to now I have treated the extensive plain created during the course of millennia by the river Po and its tributaries as a homogeneous whole. Nevertheless, during the Modern Age, the Po plain, far from constituting a vast area endowed with common characteristics, presented a fragmented picture: from political-institutional (territory divided between many states), social-cultural and physical-morphological viewpoints; and in terms of settlement patterns and crop regimes in their widest sense.

39Here, I shall consider the physical characteristics, settlement patterns and crop regimes of areas of the Po plain for which I have both demographic data, and studies of agricultural history concerning the period under examination. Going from west to east, the countryside areas in question are the Canavese, the Pavese, the Milanese-Lodigiano, eastern Emilia (particularly the countryside of San Felice) and part of the Bolognese contado. It is worth noting some peculiar characteristics of each of these areas:

• Eastern Emilia (Sanfeliciano): no major innovations in agricultural techniques or type of production took place in this period. Compared with the priority given to bread and wine production, the importance attached to livestock breeding was negligible (Cattini, 1978). The irrigation system of the area was extremely complex but did not work in a satisfactory manner: from the first half of the sixteenth century there were signs of environmental damage which prompted innumerable disputes rendered insoluble by the political fragmentation of the territory (Cattini, 1984). The series used here relate to the countryside in the communes of San Felice and Massa Finalese where, in the sixteenth century, the territory was subdivided into chiusure which, towards the turn of the century, evolved into piantate, permitting a more intensive form of farming [28]. This type of land division, which in other areas of Emilia was associated with sharecropping, was associated in the countryside of San Felice with small- and medium-scale land ownership.
• Bolognese: although the broad band of suburban land around Bologna is not homogeneous from a morphological, settlement model or crop regime viewpoint, here I shall limit my analysis to the part characterized primarily by estates organized according to the sharecropping principle [29]. As in the San Felice area, in the Bolognese there were no significant developments in farming techniques or in the crops grown. In both places bread and wine were the main products.
• Milanese and Lodigiano: they represent the heart of the irrigated area of Lombardy, with the most sophisticated irrigation networks of the period. From the agricultural viewpoint, these areas are characterized by the large number of meadows (23.7% of the general territory in the lower Milanese, over 35% in the Lodigiano) and, at least in the Milanese, of paddy fields (over 7% of the territory), almost totally absent in the Lodigiano (Chittolini, 1988). In terms of land division, large leased estates where wage workers provided the workforce were very widespread, and this state of affairs was “singularly suited to irrigated crop farming, and able to fully exploit its productive capacities” (Chittolini, 1988, p. 213). Alongside the leasing of large areas of land, the settlement model of the cascina (a sort of large farmhouse) began to develop in the course of the sixteenth century (Chittolini, 1984).
• Pavese: compared with the central part of the state of Milan, in much of the Pavese there was less effective control of the water supply. Because of its proximity to the Ticino and the Po, the terrain was sandy, and liable to swamping and flooding. In these areas, meadows constituted about 18% of the territory. The northern part of the province, on the borders of the Milanese, was part of the irrigated Lombardian plain: meadows accounted for 24.5% of the territory and paddy fields for 2.2% (Chittolini, 1988; Crosia Fiocchi, 1980). To this variety of crop regimes corresponded a similar variety of modes of land use, with the diffusion of forms of agricultural labour and large-scale leasing of the most fertile land, while, in less productive areas, more traditional production systems prevailed. The Pavese thus represents a sort of intermediate case between the Milanese and the areas of the lower Po plain where, in that period at least, irrigated farming was not widespread.
• Canavese: the plain that extends around the city of Ivrea, bordered by the morainic hills produced by the Valle d’Aosta glacier is an example of the “upper” Po plain, characterized by land which is much less fertile than the areas mentioned above and situated at a markedly higher altitude. We know that rice was never successfully grown in this area and that wheat remained predominant, at least until the spread of maize. Other food resources, such as chestnuts typical of the Alpine and pre-Alpine areas were also available. The land registers indicate the prevalence of small fragmented properties, as a rule directly farmed by the owners, although sharecroppers are not totally absent [30]. The settlement pattern of the area is characterized by a myriad of small towns and villages where the majority of the population reside: the housing model of the cascina, so well adapted to rice-growing areas, is rarely seen here.

40Figure 5 shows the five-year moving averages of the aggregated series of baptisms for the five rural areas [31] described above. Comparison of the series reveals, in the first place, the limited impact on the Canavese of the food crises of the 1590s and the 1620s, a fact which differentiates it from all the other areas examined. This is probably due to a number of causes, such as the serious damage suffered during the Italian Wars which led to depopulation; the demographic attraction of Turin, that limited the possibilities of immediate recovery and helped to maintain a more balanced relationship between population and resources; and finally, the geographical position close to mountains and hills where the food resources were less sensitive than wheat to adverse climatic conditions.

Figure 5

The movement of births in the countryside, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

The movement of births in the countryside, Northern Italy, 1562-1626. Five-year moving averages of the indexes of baptisms, base 100 = average of years 1610-1614

41Secondly, concentrating on the subsistence crisis of 1590-1593 and the years that followed, it would seem that the demographic damage was more serious in the Emilian countryside than in the Lombardian countryside. This does not emerge so much from observing the reduction of the births in the acute phase 1591-1592 as from examining the capacity for recovery after the crisis. In the countryside of Pavia, Milan and Lodi, births recovered much more quickly to previous levels than in the Bolognese and, above all, in the lower Modenese. An explanation can be found in the geographical position of such areas. Emilia was one of the parts of Italy most acutely affected by the food shortage, which remained generally severe throughout 1593. It is possible to hypothesize that the share of the decline in births attributable to temporary migration is greater in the Lombardian countryside than in the Emilian countryside, where many episodes of high mortality are known to have occurred. This would give Lombardy an increased capacity for recovery once the refugees had returned to their homes.

42The geographical position cannot explain what happened during the 1620s. In the Milanese and the Lodigiano, the crisis seems to predate and to be more severe than in Sanfeliciano, in the Bolognese and also in the bordering Pavese. This circumstance constitutes the key to a preliminary interpretation of the incidence of crop regimes and settlement models on general population trends in the countryside.

43Among the areas under examination, the Milanese and the Lodigiano are undoubtedly those characterized by the most modern agriculture, just as the Canavese has the most “archaic” [32]. It is not just by chance, I believe, that the birth trends in these two areas at the beginning of seventeenth century lie at opposite extremes. One could hypothesize that the introduction of new crop regimes, together with the expansion of irrigation and the intensification of animal husbandry improved the diet and reduced the susceptibility of the local population to crises caused by cereal shortages. The data at my disposal suggest that such a hypothesis is completely unfounded and that indeed the opposite conclusion should be drawn: the expansion of irrigated farming increased vulnerability to cereal shortages.

44For the moment this is just a hypothesis which could be verified only by extending the investigation to at least the whole of the seventeenth century [33] and expanding the database. However, I can already put forward some considerations in support of my conclusions. First of all, few cattle were raised directly on the farms [34], even those with good meadows, and did not provide a regular flow of alternative food supplies. Second, in the words of a contemporary observer, “it would be better, especially for the poor, to lack meat and cheese rather than bread. Before the Lodigiana had water brought by the Muzza [canal] there was an abundance of all types of cereals; now that the majority of fields have been transformed into meadows because water has become so easily available, if the harvest is slightly less abundant than normal the city does not have bread for the entire year; so the richer this territory has become in terms of income, the scarcer is the primary food resource of its inhabitants…” [35]. The expansion of irrigated meadows accelerated between the sixteenth and seventeenth centuries during that “Indian summer” of the Italian economy in which urban capital flowed plentifully into the Lombardian countryside and helped to change its appearance (Bolognesi, 1984). This chronology corresponds well to the pattern of births: in their incompleteness, the series nevertheless show a more serious disquiet in the 1620s than in the aftermath of the 1590-1593 famine, the period when agriculture in Lombardy had achieved its greatest “progress”.

45It is a very strange picture that emerges from observing these populations: pursuing the principles of maximization of profit (a rare situation in the countryside of the period), they create for themselves the conditions of their own downfall. This phenomenon also raises important questions of a theoretical order, so it deserves particular attention.

V – The model of population change: Malthus or Boserup?

46The two macro-demographic theories that constitute essential points of reference for those who deal with the relationship between population and resources, were elaborated in their original forms by Malthus (1798) and (much more recently) by Boserup (1965, 1981). Such theories are based on radically different ideas of the relationship between levels of population and agricultural technologies in use.

47In extreme synthesis, for Malthus, the level of population corresponds to a given technological situation, as well as to social and cultural factors. Despite the fact that imbalances periodically occur between population and resources due to their different rates of growth (geometrical for the former, arithmetical for the latter), overall the system oscillates around an equilibrium, maintained through repeated episodes of super-mortality. So in order to bring about lasting population growth, there must be a substantial improvement in agricultural technology.

48For Boserup, on the other hand, the population itself, by promoting technological micro-innovations to traditional practices in response to demographic pressure, permits slow population increase by starting a chain reaction, and a periodic re-equilibrium of population and resources is unnecessary. According to this theory, therefore, demographic pressure stimulates technical innovation in a continuous manner; population is not a dependent variable whose maximum amount is determined by technology and environmental conditions (as in the Malthusian model), but in its turn exerts an influence on technology.

49In this simple formulation I chose to underline the main differences between the two theories, but it should be remembered that they can be thought of as complementary rather than contradictory, and that efforts have been made to reconcile the two positions (Lee, 1986). Surely my data seem to confirm this complementarity, given that they do not perfectly match with either one of the two theories. Rather, according to the place, time and circumstances, they suggest that the underlying logic is either Malthusian or Boserupian: with a clear precedence, at least in the region and time I am dealing with, of the former over the latter.

50Even if more cautious positions are not lacking (Souden, 1985), historians generally agree that the serious 1590-1593 famine was a Malthusian-type crisis in the sense that the demographic pressure on resources had become so strong that a few consecutive years of bad, at times very bad, harvests caused the whole system to collapse, not only in demographic terms, but also from institutional, economic, social and cultural points of view (Cattini, 1984; Romani, 1975).

51Nevertheless, comparing different territorial aggregates, it turns out that, despite the impressive extension of the crisis, some people were spared: first of all, the inhabitants of the mountains, the Alps particularly (the Apennines were stricken by famine to a lesser extent, but they were by no means spared). In these areas, the lower housing density, linked to the fact that food supplies depended less on cereals than in the plain, enabled them to overcome bad years without too much distress.

52Up to here, nothing is unusual: indeed, the precautionary measures adopted by the mountain-dwellers to limit their own fertility (for example, deferment of marriages) would certainly have met with Malthus’ approval. Nevertheless, if we consider the slow, but constant increase in the Alpine population, that led to its trebling between 1500 and 1900 (Mathieu, 2000) and which can already be intuited in the period under consideration, Malthusian logic no longer works very well. It was not an “Alpine agricultural revolution”, some sudden innovation in farming techniques, that produced such a result. Rather, there seems to have been a progressive improvement in crops, better exploitation of, and adaptation to, the meagre resources available. The slow population increase suggests therefore that Boserupian forces were at play.

53Why were such different results obtained in the mountains and on the plain? One causal factor seems decisive: the relative rapidity of demographic growth as shown by the birth indexes. In the Alps, thanks to the social conventions which moderate birth and mortality rates, there was slower growth. The population grew therefore at a much slower pace than the geometric increase suggested by Malthus. On the plain, and particularly in the most fertile areas, the rates of growth were rapid, as can be seen from the graphs for the years 1560-1589. Probably, there was not enough time to increase the available resources by improving the crops through endeavours that, as a rule, require years and great quantities of labour. Furthermore, the results were not always positive (the countryside of the Po plain abounds with stories of “water disasters” due to badly designed canals). Certainly in the case of the 1590-1593 subsistence crisis, the population of the Po plain unknowingly brought the disaster upon itself.

54The only way out would probably have been to follow the precepts of Malthus and to imitate the customs of the “uncivilized mountain-dwellers”. Evidently this was not a practicable choice, since it required a lot of time, even more than that required to dig new canals and farm new lands.

55This doesn’t mean that the farmers of the Po plain did not respond to the ever growing pressure on their resources. For example, in the countryside of San Felice, the transition from the chiusura to the more intensive piantata seems linked to the need to produce greater quantities of food, even at the cost of sacrificing variety, by focusing on two fundamental resources: grapes and wheat (Cattini, 1984). This transformation, which lasted decades and involved a change in mentalities and social customs (for example, about what to eat, and how), is the kind of “demand-induced” agricultural innovation postulated by Boserup (1981) in times of shrinking supplies of land and food resources.

56However, technical innovations in farming techniques did not always favour increased population density. In the countryside of the lower Milanese and Lodigiano, indeed, there appears to have been a worsening of the population/resources ratio at a time of major investment and rapid transformation of the agrarian landscape. This is due, in substance, to the fact that people can’t eat hay.

57As noted earlier, the number of irrigated meadows was increased at the expense of wheat fields. Why do we encounter such different processes in two areas of the Po Plain (Sanfeliciano and Milanese-Lodigiano) at the same time? Probably, the reason must be sought in the people who took the decisions about crop innovations. In Emilia, these were the small-holding farmers who, obviously, had a direct interest in not starving and therefore, were careful about subsistence. In Lombardy, they were the new landed gentry anxious to exploit their land investments to the utmost, according to a logic of profit, bearing in mind the proximity of Milan, the biggest outlet for meat and dairy products from the Po plain.

58Some interesting conclusions can be drawn from this. In the first place, it does not necessarily follow that modernizing agriculture and adopting the best available techniques leads to a greater availability of food resources: much depends upon the purpose for which such techniques were adopted (profit or subsistence). Secondly, in this case at least, the stereotype of the stubborn farmer adverse to innovation becomes confounded with the image of a “rational actor” who knows full well what is best for himself and his family and sets himself a fundamental objective – to survive [36].

59The implicit rationality of some traditional agricultural structures has become a topic of discussion in certain branches of economic science, in particular in that known as “neo-institutionalism”. Researchers have found that such traditional structures operate according to a system of moderation and division of risks [37]. It is not possible to explore this matter here but suffice it to recall the words of Innocenzo Malvasia who, at the beginning of the seventeenth century, justifies as follows the preference given to sharecropping over other farming methods: “[land farmed by sharecropping] although it does not give a big yield, almost never fails to give a moderate yield, and of everything [38]”.

60On the basis of what is shown in the preceding pages, in sixteenth and seventeenth century Northern Italy a Mathusian-type pattern seems to be prevalent. However in well defined areas, and notably in the mountains, Boserupian theory seems more apt at describing population dynamics. The prevalence of one population model or the other does not depend simply on the introduction of new techniques as opposed to a lack of innovation, but is due to the complex interaction of factors such as the type of technical innovations introduced, the social characteristics and economic aims of the innovators, the environment and the available resources, the prevalent social and reproductive customs, etc.

VI – The general pattern of births in Northern Italy from 1500 to 1628

61Comparing different areas and environments, I found a considerable variety of population “histories” (just compare the Alps with the Emilian plain). Estimating birth trends for the whole of Northern Italy appears to be a very delicate task which would require, as a starting point, a minute division of the territory into homogeneous areas according to profiles liable to influence demographic trends. It would also be necessary to know the precise boundaries of each area and its population density.

62In the current state of knowledge, we do not have enough data to proceed in such a way. I can only present an approximate result as a basis for outlining certain key features. However I think that my data, simply because of their quantity, enable me to propose an estimation of the number of births in the period and in the area which is more solidly based than those available to us in the past [39]. Indeed, if, for a moment, we suppose that the estimates of the population of Northern Italy made by Beloch for the year 1600 (5,412,000 inhabitants) were accurate, and bearing in mind the fact that, between 1599 and 1601, an average of about 22,300 baptisms were celebrated in the parishes of my samples, we obtain an overall quota of births of between 11.8% (assuming a birth rate of 35 per thousand) and 9.2% (corresponding to a birth rate of 45 per thousand): a decidedly consistent sample.

63For the sake of brevity, in this paper I shall not describe in detail the methods used to estimate the total number of births, intended to take account of what emerged during the study of the separate localities and of the homogenous aggregates I mentioned above [40]. I shall thus limit myself to presenting the results (Figure 6).

Figure 6

Reconstruction of births in Northern Italy, 1550-1628. Index of number of births, base 100 = average of years 1601-1628

Reconstruction of births in Northern Italy, 1550-1628. Index of number of births, base 100 = average of years 1601-1628

64Combining the analysis of the above figure with what emerged earlier and with the results of further investigations which I am unable to present here (in particular as regards the first half of the sixteenth century; see Alfani, 2004), I propose the following division into periods:

• 1500-1559: demographic stasis. The stimuli to growth, although present, are inoperative due to a simultaneous combination of war, plague and famine. Rather than producing a contraction of the population, in the mid-term this prevents a substantial breaching of the levels already reached at the end of the preceding century.
• 1560-1589: rapid demographic growth, which has already levelled off by around 1585 and is followed by a brief period of stasis prior to the subsequent crash. Birth series are flat in the years 1550-1559, as shown by the graph, and all available data suggest this had been the case since about 1500, so the growth since 1560 was not due to any cohort effect.
• 1590-1593: serious subsistence crisis which brings about a lasting contraction in births of about 25-30%. Furthermore, famine has a redistributive effect on local populations.
• 1594-1619: after a phase of gradual recovery, which ends around 1610, the population of Northern Italy returns to levels corresponding approximately to those preceding the famine. This fact suggests that precise limits of population had been reached, closely linked with the availability of food resources.
• 1620-1628: signs of a tension between population and resources become evident again. Famines and the fact that the “depleted generations”, born in the 1590s, reach reproductive maturity combine to produce a contraction in births, which represents a sort of prologue to the 1629-1630 tragedy.

65The large birth cohorts in the 1580s helped to sustain the level of births in the first years of the seventeenth century (at least until about 1615), while the reduced cohorts of the 1590s subsequently held back demographic growth in the 1620s. This alternating demographic pattern is evident and suggestive. Nevertheless, the “baby boom” of the years 1560-1589 is not a logical sequel to the dynamics of preceding generations. It may have been due to an increase in the birth rate, related perhaps to a temporary respite from epidemics and a series of favourable crop cycles, and probably also to an increase in the marriage rate once the wars were over.

Conclusion

66In this paper, I have described a rather complex “old regime” demographic situation, whose various aspects can only be brought to light by a continuous two-way analysis from the general to the particular and vice versa. On the one hand, the single series constitute useful indicators of important local demographic dynamics, linked to political, military and institutional events, etc. On the other hand, it is only by overcoming localism (intended without any negative meaning, since criticism of the local series constitutes an essential starting point) that is it possible to show, by extended comparisons of several series, the long-term importance of variables often shared by localities that do not belong to the same territory.

67After a stagnation due to difficult conditions in the first half of the century, rapid demographic growth began in the 1560s. Within thirty years, this “baby boom” led to a population ceiling which could not be breached: the Malthusian crisis of 1590-1593 only temporarily re-balanced the situation. The creeping crisis of the 1620s seemed to herald the need for a further adjustment, which arrived with the terrible 1629-1630 pandemic.

68While triggered by climatic or epidemiological factors, famine and plague are nevertheless strongly influenced, in their development and in their severity, by population density and size, and also (in the case of famine at least) by agricultural techniques and crop regimes. Consequently, while they are partially random events, they can be used to compare how different communities coped with the crisis. The 1590-1593 famine, in particular, constitutes an exceptional opportunity to show the dynamics of population at work. Comparing the capacity of response to the crisis in different parts of Northern Italy, we obtain the image of populations that are prey to the known “positive Malthusian checks”, with the exception of mountain-dwellers. These latter, thanks to a different relationship with their living environment, possess the rare ability to adjust their own numerical growth to the speed of increase in the food supplies available, according to mechanisms entirely analogous to those postulated by Boserup. Besides, although all the populations of the Po plain suffered a drastic drop in their food supplies during famine, some seemed to cope with the crisis more successfully than others. In this respect, the comparison between populations faced with the creeping crisis of the 1620s is particularly significant. The results, in some ways surprising, show that the best-performing areas were those that had remained faithful to traditional crop regimes; the worst were in areas where some of the most advanced agronomic techniques in Europe had been introduced, and where production choices were becoming increasingly market-oriented. This leads us to the question of land tenure (leasing, sharecropping or ownership of small plots of land) and consequently suggests the importance of clearly identifying who the decision makers were. If my preliminary results are confirmed by other data, the comparison with studies by neo-institutional economists on many developing countries would certainly constitute an interesting avenue of future research.