Of course no source is perfect, but, as several researchers have pointed out, pedigrees are not as representative of the population as descending genealogies which strive to include all descendents of the apical ancestors whether they left descendents or not and thus childless individuals are included (see Dupâquier, 1993). Descending genealogies, however, have other problems: because a number of people leave the sample before they die and people are observed only when they experience a vital event, the life histories are less apt to be complete, a problem which can be corrected, at least in part, by searching for these “censored” individuals in other sources. My remarks will center on the question of representativeness of Pooley and Turnbull’s pedigrees and ask to what extent their conclusions depend upon their having used this particular type of genealogy.

But first I would like to mention features of their data that are unique in other ways. Assembled from calls to members of genealogical societies, respondents were asked to send in forms summarizing their ancestors’ biographies. Pooley and Turnbull broke new ground by asking the descendents to provide more than just a list of their ancestors’ residences. They asked for work histories as well and thus have produced a “triple biography” composed of work, migration and vital events ( see Courgeau, 1999). All changes of address were to be included, even those within the same locality, providing valuable information on “moving house” which they link to changes in transportation, especially numbers of commuters. And, they illustrate most of their findings with actual case studies taken from a set of diaries for the period which they also examined.

The authors are well aware of the biases in pedigrees and discuss them in the second chapter. There are really four problems.

Fewer moves are recorded for more distant ancestors. Other information such as the reasons for a move also is less abundant for the people the descendent did not know personally. But that is undoubtedly true as well of other sources, including descending genealogies, especially if they were compiled by descendents themselves.

There are relatively few biographies for persons who died childless. I was surprised to find any at all, but apparently some were provided. About 3% of their sample died single, presumably maiden aunts and uncles.

There are fewer biographies of people in lower class occupations and an overrepresentation of professionals. This is attributed to professionals being more interested in pursuing family history as a hobby than are members of other classes coupled with the fact that they were more apt to have ancestors who were also professionals.

Geographical clustering would be another possible problem if they had responses only from a few regions of Britain.

Pooley and Turnbull have done a good job on the last question. They ap-proached a large number of genealogical societies in several locations throughout Britain and their sample mirrors the actual distribution of the population as recorded in British censuses in the xix^th and xx^th centuries surprisingly well. Also the first problem—a shortfall of moves in the earlier cohorts—is not severe. It was mainly short distance moves that were lacking, moves within the same settlements. The rate of longer distance moves seems quite comparable to the later cohorts.

I will concentrate on how compensa-ting for the other two biases—the lack of people who left no descendents and the overrepresentation of professionals—might change their main conclusion of “gentle change”, that is, that industrialization and its attendant changes did not cause great disruption in people’s lives. Unfortunately, the two problems push the sample in opposite directions so that compensating for them could well cancel out, but let me first talk about each in more detail.

To Pooley and Turnbull their sample is a collection of individual biographies which happen to have been compiled through genealogical research. In contrast to many other studies of migration based upon genealogical data, they do not use relationships between individuals in their sample in their research even though several descendents did provide biographies of related individuals. Respondents were asked, however, to note the “companions” for each move, from which Pooley and Turnbull could discover whether the individual moved alone or in a family group. They found that most migration occurred in family groups and this is one reason they concluded that migration was not disruptive. Table 6.8 shows that only 22% of all first moves were made alone before 1850 and even fewer—16%—afterwards. But in every stable people who moved away from family groups, as single individuals, were more mobile and moved greater distances than people in families. In their conclusions, Pooley and Turnbull criticize earlier studies for using only such lone individuals and thereby emphasizing the disruption migration caused rather than its normality. Surely, if their figures were increased to account for the missing childless individuals the rate of migration in family groups would be lower, but would it be low enough to invalidate their finding that “most” migration occurred in family groups?

Pooley and Turnbull rightly point out that there are really two kinds of single people: children and adults. Children are single but they will eventually marry and these people are already in their sample. It is those who never marry that would change their findings. Unfortunately the two are usually counted together in their tables. Pooley and Turnbull cite various estimates of the number of unmarried adults from other sources ranging from 7 to 16% of the population. But in addition to people who did not marry, one would have to include people who, while having married, produced no living descendents. If such people exist in their sample, they are hard to discern from the way their tables are arranged. They are undoubtedly merged with couples who had children but did not move with them, so one cannot discover, from their tables, whether the migration patterns of the two sorts of couples were different nor the frequency of permanently childless couples in the study population.

There are, however, a number of estimates of the proportions of people who did not leave descendents from other sources. For example, Smith (1987, 262), in a series of computer simulations, found that the proportion with living descendents peaked at age 55. At this age, in the simulation of a “modern English population”, 18.7% had no living children. The figure was 16.3 % for the simulation of a “pre-modern” population. However, in a later simulation replicating the conditions of the xviii^th century, Smith and Oeppen (1993, 314) reported a higher proportion, approximately 25%. The higher rate seems closer to other estimates. Dupâquier and Kessler (1992, 30) report that in the descending genealogies they used beginning in the early xix^th century, 10 to 15% of married couples left no descendents and this would be in addition to what is probably a similar proportion of those who never married. In the data we use for the American North, the proportion of men who left no descendents increased from a low of 12% in the earliest birth cohorts (born before 1780 when birth rates were at their highest and death and celibacy rates at their lowest) to 22% in later years (men born between 1820 and 1840). Pooley and Turnbull report the frequency of only one of the two childless groups in their study population, those who never married, and their proportion is much less, only 3%. As a guess, I would say that the actual proportion of people who left no descendents (including as well the other childless group, people who married but left no descendents) for the period they are studying fell between 20 and 30%, a considerable proportion of the population.

Were the migration patterns of people who left no descendents comparable to those who did? There is one table summarizing the information on the small number of people in their sample who never married, but numbers are so small it is hard to judge. In our data for the American North, men who had not married by the age of 35 were rather less apt to have left their birth places during their lives (70% did) than were those who married and left descendents (83%). Men who married but had no children fell in between (74% ). And the distances they traveled in their lifetimes were different as well: men who never married who did leave their birth places migrated longer distances during their lives than did men who married and had children; but, men who married and did not have children displayed a different pattern: they moved much shorter distances over their lifetimes than either men who never married or men who married and had children. There are clearly two different groups on the basis of their migration patterns and while Pooley and Turnbull talk about the permanently single, they say little about men who married but left no descendents.

It is not possible to go further with the exercize of trying to compensate for these two sorts of missing people in Pooley and Turnbull’s data. Perhaps they would cancel each other out as they do somewhat in our materials. But I don’t think the migration rates or distances would go up so much that the generalization “Most people moved in families” would no longer be correct. If it doubled the percentages moving alone—a rather generous correction—it is only in the first period that the rates would go above 50% for first moves; later the generalization —“most”—would still be true. The same can be said for the distances traveled. It would take a lot more long distance moves to contradict their finding that “most” moves were at shorter distances.

Rather than trying to arrive at exact rates and distances characterizing the entire population, however, it is more interesting to think about what studying the rather large minority of people who did not leave descendents might reveal about connections between migration and other events in the life course. We hare already seen that there are really two such groups to be considered and each of them might shed light on constraints and opportunities for moving among the majority who did leave descendents. There may be even more subgroups, for example, two types of never married adults—those very tied to home—hence less apt to migrate than men who married—and those who did move and were somehow freer to go farther. In our American data, children not only increased the probability of a married couple’s moving but led them to go farther than did married couples without children. In our case, it is probable that having children meant one would not have to hire labor on the frontier. Perhaps the single men who went very long distances were just those hired laborers. Or maybe they tacked onto their siblings’ households enabling them to move earlier, before their own children were old enough to clear land. One could also imagine scenarios of hardship—the young unmarried person sent off early in life to help the household survive—only to return later and be unable to marry. Or it could be one of greater prosperity, the addition of another wage earner, without children of his or her own, perhaps, making it less likely that the household would have to move. Would such singles, then, be found in areas where it was easier to find work?

The census which did at least aim to include everyone could provide further insight into the frequency of these types along with information about where such single people were living and a bit of information about their migration histories which could be extrapolated from information on their birth places. Thus one could discover whether adults who never married were as likely as those who did to be living away from their birth places at particular ages.

It would also be important to study their occupations to see whether having more singles in the sample would increase the differences in migration between professionals and laborers and unskilled workers already so striking in their data and it is to these I now turn. Professionals were much more apt to move long distances than were agricultural laborers, who were very tied to a locality or even unskilled laborers whose moves averaged 25 km compared to 70 km for professionals. Among the underrepresented lower classes only domestic servants went far but even they did not match the professionals. Clerks traveled quite far as youngsters. If we tried to compensate for the bias towards professionals, we would have less migration and less migration over longer distances overall than Pooley and Turnbull report. I also wonder how the rates of celibacy varied by occupation and how many of each married and left no descendents.

But what does this all have to do with the main thrust of the book, the hypothesis of gentle change? Pooley and Turnbull include changes of address within localities, not usually counted in migration studies, and this definitely adds to the impression of stability. But the format of their tables make it easy to recalculate how removing them would affect the distances traveled. I arrived at the figure of 10 to 20 km for the median distance moved before 1920, a short distance in any case. Compare this with our data on the American North for example where the median distance moved (for men born between 1650 and 1840) was 48 kilometers (29 miles). Again, I don’t think that the addition of more footloose childless and people who never married would lengthen the distances very much and correcting for the overrepresentation of professionals would shorten them.

I do think, however, that Pooley and Turnbull view the occupational differences they found in too rosy a light. They find a link between longer distance moves and upward mobility. If so, the lower classes may have been “locked in” to a locality and could take less advantage of opportunity elsewhere. They see the local nature of most moves as a brake against “disruption” caused by industrialization, but the less mobile lower classes also paid for this in not being able to take the better jobs located at greater distances. Perhaps the un-skilled and agricultural laborers did not move as far as professionals because they were more embedded in their kin and local networks while the young clerical workers and professionals had a different sort of orientation towards their families. Such class cum cultural differences would also contradict another one of their conclusions: the uniformity of migration during this period of change.

A large number of return and short distance moves also contributes to their conclusion of gentle change: if the move did not work out people could and did return to where they had lived before. But how many had family or even friends to return to? Given higher mortality rates and communicable diseases like tuberculosis which struck family groups, could unmarried people or couples count on a network of kin in their old places? Now, by its very nature, their sample lacks examples of families which died out. Would not people from these sorts of families have been more uprooted by moving than those who found their way into Pooley and Turnbull’s sample? They would have had less opportunity for chain migration as well because their families were already very small and due to become smaller. In this connection the large proportion of people in the general population who left no descendents is sobering. After all, even if people moved short distances, if their families were dying out, this in itself would be a disruption. This question might have been approached rather differently if the entire kin network were available for study.

In sum, their work is largely a description of the lines which made it into the professional classes. Others were certainly less successful or left no descendents to send forms to be included in the study. Whether these peoples’ moves were more disrupted or different in other ways from the more prolific people who were the subject of this book awaits further study. But even if more isolated individuals could have been included I would suppose them not to have comprised more than say 30% of the population so the general conclusions would not be much altered. Still, this is a large sub population whose migration histories certainly differed from the rest.

Pooley and Turnbull have provided a monumental analysis of a large number of biographies. Their “gentle change” hypothesis fits with work such as Rosental’s (1999) who argues that absolute distances tell us little about family territory which stretched over long distances but was able to overcome them. Together they emphasize that the image of floods of uprooted migrants comes from the receiving areas, the cities, which also had the most literate and vocal populations and where migrants might have indeed caused disruption for some of the people who were already there. But a sample like theirs which mirrors the distribution of the entire population at the time, the sending areas, provides a different view. Still a major task of integration remains: how to reconcile their microscopic picture of microscopic changes with the macroscopic view from sources at a larger scale, like the census, which describe a flow out of rural areas of “large” proportions for the very same period.

Migration stands at the crossroads of so many fields. In this study it is viewed largely through a geographical lens (the maps notwithstanding, which are very difficult to see, let alone interpret), but through it I have been reminded that migration is one of the three main topics in demography. A demographic transition was occurring in England during the period they are studying; its role in increasing or decreasing the size of a family safety net or resource pool that could be drawn upon in a local area has yet to be described. One would need to combine the fine grained picture of lifetime moves Pooley and Turnbull have provided with the history of the expansion and contraction of families. If so, one might find lives disrupted and less gentle change than pictured here. If migration were studied within the context of the epidemiological and demographic transitions that occurred at the same time, the causes of disruption might shift from migration itself to fewer family members available for support and differences between occupations that excluded some from the migration that was necessary for upward mobility at the time.