BALSAC data

Genealogical information on the historical Québec population is digitalized in the BALSAC population database (http://balsac.uqac.ca) with a remarkably high degree of completeness [19]. BALSAC includes data from the 17^th and 18^th centuries in Québec that were initially compiled within the Early Québec Population Register of the Programme de recherche en démographie historique (PRDH) of the Université de Montréal [28] (https://www.prdh-igd.com/en/home). The genealogies were primarily reconstructed from Catholic church records about life-history events such as marriages, births, or deaths. Moreover, BALSAC and PRDH collected valuable information on genealogical connections beyond Québec, such as data on European and Acadian ancestors or emigrant marriages outside Québec (e.g., [7,29]). We define immigrants as the first generation of settlers to marry in Quebec. Most immigrants introduced new maternal and paternal lines, representing new mitochondrial DNA and Y-chromosome lineages, respectively, and are considered lineage founders. However, some immigrants were born in Quebec to parents who married before coming to Québec. Also, some were related to each other upon their arrival in Québec (e.g., siblings, cousins). In those cases, using BALSAC data from outside of Québec, the most recent common ancestor (maternal or paternal) of these related immigrants were identified as the lineage founder to avoid counting a lineage more than once (S1 Fig). The population of married couples between 1931 and 1960 was set as the reference for the 20^th-century population of Québec. The 30-year period approximates the time of one generation [30].

Demographic parameters

From BALSAC, we extracted all married individuals and categorized them either as immigrants or as Québec-born (QB) descendants of the immigrants (c.f. S1 Fig). Immigrant males and females were counted independently to separate maternal and paternal lineages and pooled into waves w_i corresponding to their marriage years t_i, starting in 1621 and up to 1930. According to their marriage year ti, QB were grouped into corresponding strata s_i. N_i represents the count of immigrants from wave w_i in the year t_i (both sexes, or only either males or females, according to the context). Nw_i denotes the number of their male or female descendants married between 1931 and 1960 (here, only maternal lineages in married women are counted: male mtDNA carriers are excluded because they do not transmit mtDNA further). The ancestors of Nw_i are referred to as contributing individuals: contributing immigrants Nc_i and contributing Québec-born Nc_QBi. Consequently, N_QBi represents the count of s_i QB married in year t_i and Ns_QBi that of their descendants married in Québec between 1931–1960.

To assess the overall rate of the demographic growth k in Québec, we used historical estimates of the Québec population size from Larin [31], census reports [32], and BALSAC data (only considering married individuals). In a growing population, N_t = N_t-x (1+k)^x, where N_t is the population size at the year t, N_t-x, its size at x years earlier, and k is the population growth rate per year. Because k is small, (1+k) ^x ~ e ^{k x}, it follows that ln(N_t)-ln(N_t-x) = k·x, such that the population growth rate k can be readily estimated from the slope of ln N_t versus time t (S2 Fig).

New descriptors of descending genealogies

Remarkably, descending genealogies come with additional information not captured in the ascending genealogies. It includes the number of initial founders and their descendants; the surviving (and lost lineages). We introduce three measures to describe this information on population demographic history, namely: (i) the survival ratio (SR_i), (ii) the net growth (dc_i), and (iii) net extinction (a_i) rates. SR_i = Nc_i /N_i for immigrants and Nc_QBi/N_QBi for Québec-born individuals, thereby describing the proportion of contributing individuals within w_i or s_i. The net growth rate dci and extinction rate ai, between historical t_i, and the descendant population of the 1931–60 Québec is estimated using t = 1947.5 (t_1947.5) as an average target year of marriages between1931-60. Both dc_i and a_i are evaluated using the same rate equation we used to estimate k, except that they only describe a net growth or extinction of the population fraction from time t_i down to the average t = 1947.5. Thus, we write Nw_i = Nc_i(1+dc_i)^(t_1947.5^-ti), whereby dc_i = ln(Nw_i,/Nc_i)/(t_1947.5-t_i) and, likewise, dc_QBi = ln(Ns_QBi/Nc_QBi)/(t_1947.5-t_i). As for the net extinction rate for immigrants: Nc_i = N_i(1-a_i)^(t_1947.5^-ti) and a_i, = ln(N_i/Nc_i)/(t_1947.5-t_i) and likewise for the Québec-born a_QBi, = ln(N_QBi/Nc_QBi)/(t_1947.5-t_i). Note that dc_i and a_i complement SR_i and 1-SR_i, respectively, and consider the time factor (t-t_i). We pooled data in bins of five or ten years to plot the data to reduce the variance, yet the values shown are averages per year.

Progeniture

We refer to the number of married children of a couple as their effective family size (EFS) [33] or the number of potentially fertile children (i.e., grand-child bearing children) [6,8,33]. While BALSAC data does not allow to calculate Darwinian fitness of individuals directly, fortunately, the EFS is strongly correlated to fitness in the French-Canadian population and was used here as a proxy [6,33]. However [34], see the danger of conflating selection and inheritance with such a fitness proxy, although the strong correlation limits this problem. We measured EFS sex-wise by counting the number of married daughters (EFS-daughters) or sons (EFS-sons) for mothers and fathers, respectively, or pooling both sexes’ maternal and paternal children. We separately examined the EFS of contributing and non-contributing parents from immigrants or their Québec-born descendants.

Genetic contribution

In genealogical studies, genetic contribution (GC) describes the expected number of copies of a neutral allele contributed by a population founder to their descendants [11,35,36]. By default, it relates to the autosomal inheritance, such that in a complex genealogy, a founder allele can be transmitted through different paths, all of which have to be taken into account when computing GC [16]. Estimation of the contribution of uniparentally inherited maternal or paternal lineages is straightforward. We count the number of lines attributed to every contributing founder (Nw_i above). Maternal lines follow the mother-to-child path, propagated further through her daughters and ending at each son, and paternal lines follow father-to-son transmission. Conversion of counts to frequencies gives relative contribution of uniparental lines founders, tantamount to frequencies of the corresponding mtDNA and Y-chromosomal founder lineages. Estimating “autosomal” GC explores the same genealogical paths, exploring all their combinations because they are crossed down, with a probability of transmission of 0.5 from a parent to a child, disregarding their sex. In other words, in contrast to maternal and paternal lines only, an autosomal founder allele can descend to a target population through all possible maternal and paternal pathways linking the founder to its descendants. Autosomal GC was estimated using the gen.gc function of the R package GENLIB [37]. Genealogical error rate, resulting from false paternity or maternity (e.g., due to undeclared adoption) and errors in parish records and their transcription is estimated to be below 1% [24]. Given the amount of genealogical data we processed, and the nature of our findings, errors at such a level can be neglected, as they are not expected to affect our results and conclusions.

Population founders and their descendants in 1931–60 Québec

The European founding of the Québec population started with the establishment of Québec City in 1608, followed by urban centers at Trois-Rivières (1634) and Montreal (1642), upstream of the Saint Lawrence River. The settlements in Beaupré (1636) and Baie-Saint-Paul (1673), north-east of Québec City, largely contributed to the peopling further downstream. In 1763, the Treaty of Paris sealed the British conquest of Québec and Canada. Around 1800, the Québec population had grown to about 220,000 people [31], eventually reaching 2,874,662 in 1931 [32]. In BALSAC (version as of March 2015), we find genealogies of 3,340,072 individuals married in Québec before 1961 (2,018,038 before 1931). This includes 455,687 immigrants, of whom 276,946 settled before 1931. Of the latter, 29,668 immigrant women and 32,608 immigrant men contributed to the pool of uniparentally-inherited lines of 1931–60 Québec (Table 1). In its early years, the Québec population primarily grew due to the inflow of immigrant settlers (Fig 1A and Table 1 and S2 Fig). The first marriage documented in Québec Catholic registers took place in 1621.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Immigrants, their Quebec-born descendants and related statistics as a function of time.

The number of immigrants and their contributing portion at linear scale (inset shows early history at a magnified scale) (A). The number of all immigrants, Québec-born and their total in logarithmic scale (B). The number of immigrants’ descendants, Nw_i, within the 1931–60 population as a function of immigrants’ settlement time (C). Evolution of the survival ratio SR of immigrants and Quebec-born (D). Corresponding plots of the net growth (dc_i) and net extinction (a_i) rates for the immigrants (E) and the Quebec-born (F).

https://doi.org/10.1371/journal.pone.0266079.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Numbers of immigrants and Québec-born individuals who married in Quebec in five historical periods.

Numbers of distinct and novel (newly introduced, i.e., in addition to those already present and introduced earlier), uniparentally inherited lineages (separately maternal and paternal) are reported as well. Contributing immigrants are those who transmitted their descendant lineages in the 1931–60 population.

https://doi.org/10.1371/journal.pone.0266079.t001

Historical periods and population expansion

Based on historical characteristics, we divided the European settling of the Quebec territory into five immigration periods. From 1621 to 1680, the first period was marked by the arrival of the Filles du Roy in the years 1663–73. They were French women whose immigration to New France was sponsored by King Louis XIV (thus “King’s daughters”), designed to boost New France’s population. They represent the first immigration peak in Fig 1A. About twelve hundred women and sixteen hundred men immigrants were registered at that period. Out of these, 708 women and 830 men left descendant lineages (Table 1) that account for 66.6% of maternal and 47.3% paternal lineages in the 1931–60 Québec population l (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Contribution of immigrants from different periods to the 1931–60 Québec, both in the number of their married descendants and in the number of distinct maternal and paternal lineages.

https://doi.org/10.1371/journal.pone.0266079.t002

The second immigration cycle in Quebec, from 1681 to 1750, took place after the arrival of Filles du Roy and before the British conquest (1760). This period contributed 13.9% and 30.6% of maternal and paternal descendants to the 1931–60 population. The paternal contribution prevailed over the maternal lineages. Altogether, the uniparentally inherited lines introduced by the immigrants before 1751 are found in ~467,000 women and ~431,000 men married between 1931 and 1960 (80.5% and 77.9% of all married descendants from before 1931, respectively; Table 2). At the turn of the 18^th century, QB individuals surpassed the immigrants in numbers and dominated the bulk of the population (Fig 1B).

From 1751 to 1800, the third colonization period was marked by a remarkable peak of male immigrants (Table 2 and Fig 1A) due to the demobilization of French soldiers [31], who settled in Québec after the British takeover in 1760. They were joined by descendants of French pioneers from Acadia (present-day Nova Scotia and New Brunswick), who were survivors of the British deportation campaign of 1755. Acadian settlers were followed by British Loyalists fleeing the United States after the US Declaration of Independence. Immigration dramatically increased after 1800 (Fig 1A and 1B). It slowed down in the middle of the 19^th century and resumed at the turn of the 20^th century. About 55,000 immigrants arrived in Quebec during the 1801–1850 period (4th period) and more than 200,000 from 1851 to 1930 (5th period).

Up to 1680, the Québec population growth mainly reflects an increasing number of new settlers. Estimates of k for this period, based on Larin’s data (k = 0.084) and BALSAC data (k = 0.11), concord very well with each other, considering the relatively small number of individuals involved (S2 Fig). During the 1681–1850 period, the population growth rate is estimated at 0.026 with both data sets (Larin and BALSAC) and k~0.015 following 1850. The population growth after 1680 is primarily due to QB, while the contribution of the new immigrants dropped to about 15% (Fig 1B). In comparison, the average population growth rate in Europe between 1500 and 1900 was about 0.006 per year [38].

Survival ratio, net growth, and extinction of lineages

While focusing on maternal and paternal lines, we counted how many lineages survived, how many were lost, how fast they grew, and what was their extinction rate; all this relative to the 1931–60 Québec population. The demographic outcome of an immigration wave i among the 1931–60 descendants, Nw_i = N_i[(1- a_i)(1+dc_i)]^(t_1947.5^-ti) is a function of the immigrants’ number N_i, their arrival time t_i, the extent of their growth (dc_i), and line loss (a_i). The maximum Nw_i coincides with the immigration peak of the Filles du Roy (Fig 1C). Note that an excess of females Nw_i preceded and was within this maximum. This was followed by much greater male than female Nw_i, reflecting the prevalence of male immigrants. In contrast to the 17^th-18^th century immigrant contributions (Tables 1 and 2), the Nw_i of the 19^th-century immigrants (Fig 1C) appears strikingly low for both sexes, despite an upsurge in immigrant numbers (Fig 1A and 1B) as judged by the relatively low level of descendants in 1931–60 Québec (S3 Fig).

Before 1681, the survival ratio of immigrants was around 0.6 (Fig 1D). Its subsequent decrease can be associated with a drop in net growth, dc_i, and an increase in net lineage extinction, a_i. In the first half of the 19^th century, there was a sudden dc_i drop and a sharp rise in the immigrant a_i. This combination explains the minimum immigrant survival ratio with a nadir of 0.1 at around 1835. Around 1825, a_i became larger than dc_i, (i.e., ln(N_i/Nc_i) > ln(Nw_i/Nc_i), such that N_i > Nw_i). It means that from that time on, the number of immigrants exceeded the number of their descendants among the 1931–60 marrieds (compare data in Tables 1 and 2).

Effective family size—determinants of the outcome

The demographic outcome of immigrants depends on their reproductive performance (reproductive value [11]) and that of their descendant generations. The QB survival ratio follows a similar trajectory but with higher survival ratios and ranges from about 0.8 to a minimum of only 0.4 in the middle of the 19^th century (Fig 1D). The extinction rate a_i of QB paralleled that of the immigrants, albeit at a lesser pace. QB appear more demographically stable but still succumb to the same external factors, decreasing the survival ratio and slowing their growth. QB did not go through the 19^th-century dc_i downfall that affected immigrants (Fig 1E and 1F). EFS values are the highest for married individuals in the 17^h-century. They are followed by a minimum around 1750, a rebound at the turn of the 19^th century to eventually decline. Contributing QB and immigrants show similar ups and downs in EFS (Fig 2). However, overall, QB have a greater EFS than the immigrants (Fig 2 –middle curves). Non-contributing individuals had a much lower EFS throughout the study period. Immigrants showed the lowest EFS with an average of ~0.5, i.e., one married child per couple, either a boy or a girl (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Effective family size of maternal and paternal lineages at different time points.

Families of immigrants (A), and Quebec-born families (B), showing EFS of contributing, non-contributing descendants and their total (All). Maternal EFS-daughters are shown in dotted lines and paternal EFS-sons in solid lines.

https://doi.org/10.1371/journal.pone.0266079.g002

Notably, the survival ratio of immigrants follows their EFS (Fig 3A and 3B). The survival ratio of the maternal and paternal lineages of the immigrants also strongly correlates with their corresponding EFS (R² = 0.91 and 0.90, respectively; p << 0.001). In other words, the proportion of the contributing immigrants’ lineages from any time i mirrors their EFS. This correlation also stands when we consider 1-year rather than 5-year intervals, despite a larger variance (S4 Fig). An immigrant with an EFS = 1 (1 child, either daughter or son) predicts that its descending lineage (maternal or paternal) has a ~0.3 probability of surviving up to ≥1931 (insets in Fig 3A and 3B). The correlation between the survival ratio and EFS is perpetuated over QB generations descending from these immigrants (Figs 3C and 3D and S4), although the correlation is less straightforward (R² = 0.65 and 0.69). And indeed, by the end of the 18^th century, QB EFS and SR plots start to diverge and then cross by the end of the 19^th century, thereby inverting the relation (see Discussion).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Correlation between effective family size and survival ratio.

Overlapping plots of total EFSs and EFSd and their survival ratios for the immigrants (A and B) and the Quebec-born lineages (C and D). Their correlation plots are shown in the insets.

https://doi.org/10.1371/journal.pone.0266079.g003

Sex-asymmetry among immigrants

During the French Regime, there were fewer immigrant women than immigrant men. Before 1681, the immigrants’ sex ratio (men to women) was 1.3 (Table 1). It went up to higher values in the middle of the 18^th-century (S5 Fig) with an average of 2.1 between 1681 and 1800. At the colony’s beginning, women married much younger than men and remarried more often [17,39]. In Fig 4, we show that before 1681 remarriages of women prevailed and that this tendency reverted after 1681. During the following years, given the shortage of immigrant women, immigrant men married Québec-born daughters of previous immigrant couples. This created a disequilibrium between EFS-daughters and EFS-sons of QB parents. According to BALSAC with the PRDH data included (see M&M), during the early time of Quebec (parental marriages before 1700), Québec-born males were less nuptially successful than their sisters (Table 1), as shown by the parents’ EFSs that was lower than EFSd (Figs 2 and S6). It is plausible that many QB males were lost from the “radar” of Québec church records; as explorers of new territories themselves, they became emigrants.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Distribution of marriages and remarriages among marrieds at four time periods.

Different categories of marrieds are specified in the above histograms by their corresponding color codes. The contributing male and female immigrants are in A and B, respectively, whereas all (immigrants and Quebec-born) married men (C) and women (D).

https://doi.org/10.1371/journal.pone.0266079.g004

Consequently, we observe the preferential impact of the matrilineal lineages from before 1681 (Fig 1C) because successive immigrant males marrying QB daughters promoted the expansion of maternal lines introduced earlier. (Note that years on the x-axis identify parental marriages, yet the statistics of the reporting children marriage, such as EFS, are taking place one generation later). At the same time, these male immigrants enriched the paternal lineages’ genetic pool relative to that of the ill-represented immigrant maternal lines (Figs 1C and S7). In the 1931–60 population, we find 630 individuals per single maternal lineage introduced before 1681, compared to only 330 per single patrilineal line (Table 2). Considering all descendants from before 1801, we find three times more paternal than maternal lineages in the 1931–60 population, 3514 and 1185, respectively.

In contrast, over the 19^th century, down to 1930, there were 24 883 paternal and 22 837 maternal contributing immigrants (Table 2). The sex ratio of 1.09 concords with an excess of male remarriages compared to the period before 1681 when female remarriages prevailed (ratio of 0.92). Male immigration bias profoundly affected the frequency spectra of male and female lineages introduced before 1801 (S7 Fig). Lower female lines diversity was observed at the molecular level by contrasting the Y chromosome versus mtDNA diversities in Gaspésie [40]. That is also known from the demographic comparisons of the frequency of patronyms and matronyms [41].

Differences in the contribution of the immigrant women and men as a function of settlement time and recorded through uniparentally inherited lines also apply to their respective “autosomal” GC’s. We find that the contribution of women and men immigrants’ maternal and paternal lines, respectively (Fig 1C), reflects that of their “autosomal” GC (Fig 5). Slight differences between uniparental and “autosomal” GC plots can be ascribed to all immigrants being considered autosomal founders. In contrast, some maternal or paternal founders were shared among related immigrants (see M&M). Computing “autosomal” GC assumes an equal probability of transmission of one of the two parental alleles, disregarding the variance due to linkage within the inherited chromosomal segments [10,36]. It is thus not unexpected that “autosomal” GC ideally correlates with the uniparental contributions (Fig 5). If genuine autosomal genetic diversity data were used, we would observe a significant variance [10] in the autosomal plots due to linkage, selection or admixture.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Comparison of “autosomal” GC of immigrant women and men with GC of their maternal and paternal lines (Nw_i) within the 1931–60 population.

The fraction of the immigrant contribution (y-axis) is expressed per year averaged over five years (x-axis).

https://doi.org/10.1371/journal.pone.0266079.g005

Effective family size and Québec evolutionary dynamics

Our study finds that the effective family size, EFS, remains a major player in the evolution of human populations. EFS represents the average number of children likely to increase offspring propagating generations down. The SR_i declines in lockstep with a decreased EFS (Figs 2 and 3), revealing a very close correlation between the EFS and their SR_i (Fig 3A and 3B). This correlation remains straightforward from the colony start until the end of the 19^th-century: EFS determines the demographic outcome for many generations ahead.

The correlation between EFS and SR_i is not fortuitous. It reflects the fate of the introduced lineages that can be readily followed in the dc_i and a_i plots (Fig 1E and 1F). These plots reveal the interplay between prosperous and extinct lineages, respectively. The immigrants a_i soars dramatically at the turn of the 19^th century, whereas their dc_i drops (Fig 1E). The immigrants became less and less successful during the early 19th century, and their number Ni eventually exceeded their 1931–60 descendants (Nw_i). This starts when the a_i and dc_i plots cross when Nw_i = Nc_i. After 1850, immigrants’ dc_i increases and a_i plateaus, but the balance is negative given N_i>Nw_i (Fig 1E). Among QB, there is less variance in the corresponding plots (Fig 1F). QB dc_i falls at the turn of the 18^th century (matching that of the immigrants), levels off after this time and eventually rises during the 19^th century (Fig 1F). QB a_i never tops QB dc_i, yet the portion of their extinct lineages steadily rises, with a_i noticeably increasing by the end of the 18^th-century (Fig 1F, but see also Fig 3C and 3D).

Favored were immigrant parents who left plentiful married progeny. EFS, a fitness proxy, appears as the most important variable determining parental demographic success through future generations [11]. However, for this success to persist, the QB progeny must continue to be reproductively successful as overall SR is ultimately a function of successive EFS among the descendants. And indeed, a correlation of family size between generations was described in various populations [8,11,12,33], including species other than humans [11]. The reproductive success appears to be heritable, as shown in Québec and other preindustrial societies [6,13,43,52]. This might imply local selection [51] acting on new mutations and/or standing variation [53–55]. High EFS creates a surplus of new combinations among existing variants and increases new haplotypic arrangements. Non-recombining haplotypes of mtDNA and Y- chromosome could partake in these new combinations. High EFS continued over many generations suggesting that some of these combinations may present increased adaptive values and were favored [6,43]. By such traits, we understand biological (evolutionary), social, cultural, and economic benefits conducive to population expansion. In Québec, despite harsh environmental conditions, it seems to include many winning factors supporting survival.

Factors affecting EFS estimates in Québec

We cannot ignore couples who failed to have children for various biological reasons, such as when one of the partners was sterile [56,57]. Sterility affects about 15% of contemporary couples [58] and was estimated to be about 10% in historical times. Therefore, the sterility of one of the partners must have contributed to the “zero fertility” peak in plots of the number of effective children that was published previously [6,33]. The overall childlessness reported in our plots corresponds to a cumulative effect of biological (see above) and other reasons that results in effective children being absent in the BALSAC records. If no married progeny is found in the BALSAC database, it can mean that the parents did have children, but these children died before reaching reproductive age; they did not marry or move outside of Québec [31,59]. If the childlessness were only due to biological causes, we would expect this to be relatively even among different Québec regions. However, we observed that zero peaks (no children among married couples) markedly differed among the regional populations, and often, this difference exceeded 10–15% (S9 Fig).

Furthermore, the proportion of parents with no recorded children is more significant among immigrants than among QB (S10 Fig). This is consistent with the emigration of the first-generation immigrants or their children, who moved out of Québec [60] and disappeared from the BALSAC horizon (Fig 1E and 1F). In early colonial times, it also happened when some immigrant settlers moved outside Québec borders southwards to the United States or western Canada [31,59]. This phenomenon became exceptionally substantial in the 19^th-century, engulfing many QB [59–61]. Emigration was motivated by unsatisfactory local conditions and tempting opportunities to prosper elsewhere. Local economic and social conditions started to worsen due to the rise in population density, military conflicts, famines, and epidemics [31,39,62,63]. This emigration affected the newcomers considerably more than the previous settlers. Therefore, when immigrants and their children left Québec, then “lost or extinct” does not necessarily imply that lineages absent from BALSAC records did not prosper elsewhere. In our calculations, we count only those that remained within the Québec borders.

Québec population was not the first colony benefitting from over-Atlantic European expansion. Europeans re-colonized different areas of the globe in historical times, often founding new nation-states, currently populated with their descendants, with various demographic and genetic outcomes. In Meso- and South America, a skewed sex ratio among European invaders, led to an asymmetric male-biased admixture with Native populations. European Y-chromosomes and Native mtDNA prevail in descendants [64–67]. Also, in Quebec, during the 17^th-century colonization of the St. Lawrence Valley by the French, male settlers were more numerous than female immigrants [39]. Genetic studies indicate that at that time, admixture with Natives was limited [68], agreeing with Catholic Church records. Only a few Native females contributed to the French-Canadian genealogy [7,40,69]. Note, however, that many mixed marriages went presumably unrecorded, explaining thus the shortage of EFS-sons at the early colony years (S6 Fig). BALSAC database covers most Québec genealogies, but not all Québec diversity. Missing information includes some unnoted Natives and many Protestant Church records.

Québec population history and founder events

In genetic studies of the Québec population, it was usually presumed that French-Canadians descended from about 8,500 immigrants [17] who settled in Nouvelle France before the British conquest in 1760 (e.g., [10,33,35,38,39]. We found that this premise is only partially correct. We show (Figs 1 and 2; Tables 1 and 2) that immigrants after the 1750s added many new matrilineal and patrilinear lines that were still present in the 1931–60 genetic pool. They exceeded the contribution of the first settlers in terms of the number of new lines but contributed much less to the mass of the genetic make-up of the post-1930 population (Table 2 and S5 Fig). And indeed, there is a striking difference between the 17^th and the early 19^th-century immigration. The population of the first immigrants was thrifty and thriving, populating a sparsely inhabited territory. In contrast, the subsequent immigrants were numerous, but their immigration success declined: the territory was already packed, and many immigrants failed to settle. It could be that during the 17^th century, there was more social and economic cooperation, despite men’s rivalry to find a spouse [17,39]. Although there were local conflicts with the Natives and the British, the environment was challenging but healthy since epidemics arrived only by the late 17^th-century [63]. Therefore, being first and successful would have given these individuals the edge over subsequent generations in a quickly growing population [6,70,71]. The earliest immigrants benefited from demographic advantage reproducing when the population was small and given their EFS, they boosted their genetic representation in future generations relative to latecomers [43,70,72,73]. In the early 19^th century, Québec was already occupied by the well-established and prospering QB individuals. For newcomers, a suitable economic niche was more challenging to find. It persuaded many immigrants to temporarily settle in Québec and eventually move elsewhere [59]. Their EFS was dropping. Also, QB families were affected, albeit less dramatically. The fraction of reproductively prosperous families decreased (Fig 2). Interestingly, however, this occurred not everywhere, since the late 19^th-century population of Saguenay–Lac-Saint-Jean was demographically as successful as that of the 17^th century Québec settlers, where again the environment was harsh and challenging [6,74].

Québec colonization can be described as a series of successive founder events. It was accompanied by the founding of genetic mutations causing regionally distributed rare and endemic hereditary diseases (“medical founder effects”), well-documented in the Saguenay-Lac-St.-Jean and Charlevoix regions [73–77]. However, besides this medical interest, Québec also provides a founder population model in terms of the original concept developed by Ernst Mayr [78]. He stated that a relatively small group of migrants establishes a new population. Due to a sampling process, the overall diversity was reduced, and many variants were moved to different classes of the allelic frequency spectrum (S7 Fig). Likewise, while many rare variants were left behind, some, including defective, disease-causing alleles, could locally increase in frequency, thereby explaining “medical founder effects” [70,75,79,80]. Familial bonds led to non-random migration, when an individual would leave France to immigrate to Québec where his sibling had already settled [29], or when deported Acadians returned to New France to reunite with their families [61]. Such immigration bias could have favored the establishment of particular lineages. Sex-bias and founder effects may be typical to expanding human populations settling in an empty territory or conquering already occupied areas [81,82]. Because population dispersals occurred multiple times in the history of the human species [83], the question is to what extent Quebec demography, at a scale of fewer than 20 generations, may serve as an example of how different human populations evolved? A recent study describes the peopling of the Pacific Islands [84] that occurred over a similar time frame as the European colonization of the Americas. In Saint-Lawrence Valley, the population expansion went through a serial founder effect following a step-wise occupation of new territories. Likewise, the first arrivals from the Asiatic mainland to isolated Polynesian islands have experienced rapid initial growth, continued through a succession of founding bottlenecks, with resulting populations dominated by the genetic contribution of their founders [84]. We may ask weather Québec history provides a good model to analyze genetic diversity of expanding human populations undergoing founder effects, or presents only its particular example. We note, however, that founder effects are not reserved to the expanding humans and are also observed in other species [85].

Humans tend to keep contact with their founder and neighboring populations, even when separated by miles of ocean water. New immigrants constantly enrich the Quebec population. Tracing their origins should add to the genealogically inferred population diversity, especially when accompanied by population-wide genomic screening programs [86,87]. Such efforts help to understand better the history of the Québec population and its changing structure, assisting important epidemiological studies that impact health care programs.