19th century population explosion!

Well, a lot happens in four years, but enough about me...

David Ellis of Natick MA recently posted to the LitvakSIG listserv a question that related to population growth in the shtetlach in the 19th century.  His question related to Butrimonys.  It led me to wonder whether we can find out anything interesting about shtetl populations using the datasets we have available.

I looked at Salakas for which we have a range of data sets - census, revision lists and family lists.

(An article by Boris Feldbyum explains the background of much of this information   http://www.bfcollection.net/fast/articles/ruscensus.pdf )

The 1784 Grand Duchy of Lithuania (GDL) census has 80 lines for Salakas, the 1845 revision list has 650 lines or so, the 1874 family list has 3080 entries and the 1887 family list about 8,500.

Plotted on a chart:

This looks exponential: and plotting on a log chart we see the trend very clearly.

This is a steadyish trend of close to 4.5% every year.  In other words the population doubled about every 16 years.   If this had continued Salakas would now be a metropolis of more than 1 1/2 million people.

There was a mix of immigration and organic growth.  Comparing the 1845 and 1887 lists, 1845 has 72 unique family names, 1887 about 200. Using this as a proxy for net immigration (which may or may not be reasonable) it looks as though about 2.5% annual growth was through net immigration with another 2% or so “organic” growth. 

There is a myth of the unchanging shtetl.  Salakas was in ferment with a doubling in size in less than a generation.  With this pace of growth the number of young people massively outweighed the old.

While the population expanded, opportunities became more and more limited as the Russian state failed to industrialise and limited Jewish access to higher education and the professions. The pressures on resources was clearly a key factor in the explosive emigration that started in the 1880s.

How healthy was interwar Lithuania?

My conclusion: it was about as healthy as living in Switzerland. But before 1945 Switzerland was not such a great place to live: their economic miracle came after 1945. And life expectancy was below that for the US.

Last year I wrote a post that mentioned a 499 record dataset of deaths between 1922 and 1939 for one district ("dataset A"). I have also obtained another dataset of 963 deaths for the same period for an adjacent district ("dataset B"). The deaths are almost exactly split between male and female. Deaths of infants of less than one year in dataset A were 25, and dataset B 63. I have anonymised the information due to Lithuanian privacy laws for events within 100 years.

Grouping the deaths by age in 5 year blocks and adjusting for the different numbers in each dataset ("normalising") we can see the overall pattern of age at death in the two datasets.

The chart on the left (click to expand) shows a good eyeball consistency between the two districts in the pattern of age at death. This gives us some confidence in the quality of both datasets.

Note that I have excluded deaths of infants aged one or under.

What pattern can we see?

• Up to the age of about 50 roughly similar numbers die annually and then from there the numbers rise sharply.


• There is an interesting twin peak pattern: in the range 71-75 years the number of deaths dips when we might expect a peak. What might be the reason for this? This is really very odd - we are looking at a 17 year period and there is a consistent gap in this age range. I'll look at this again later.


• The number of deaths then drops as the number of people left alive at each higher age falls. The oldest death in Dataset A was 102 and in Dataset B was 105. Now there may be some doubters out there, but I have looked at the 102 year old and traced the person back into mid 19th century revision lists and the age does seem to be correct.

How does the pattern of age at death compare to the rest of the world? I chose Switzerland as a benchmark. Switzerland is more formally known as the "Confoederatio Helvetica" or "CH" for short. I made this choice for a number of reasons. It was largely unindustrialised at the time, similar to Lithuania; it was unaffected by Great War deaths, which completely changed the demographics of France, the UK, and Germany; and reliable data is readily available from the Swiss Federal Statistical Office. I used data for 1928 and compared Dataset A with Swiss male and female deaths.

The chart on the right (click to expand) shows this comparison. (Again the total number of deaths has been normalised to facilitate this)

The eyeball consistency is remarkable: and particularly with CH female deaths.

We see the "twin peak" in the Swiss data. This means that our Litvak datasets' twin peaks were not aberrant or an artefact of my amateur analytical methods. There was a cross european phenomenon which meant that there were fewer deaths of 71-75 year olds of both sexes than one might expect.

I have no idea why this might have happened. If any historical demographer chances along here please comment. Anyone in fact: please post any ideas you might have.

Despite this gap in understanding, it is clear from the overall pattern that Litvaks and Swiss had very similar mortality patterns during the interwar years - and it's reasonable to conclude that overall health was therefore similar.

The consistency of the Litvak and Swiss data for the interwar period suggests that looking at Swiss data for earlier periods might be suggestive for the pattern of age at death for Litvaks in earlier periods, where we have much less information for Litvaks.

The chart on the left (click to expand) shows Swiss data for 1876. We see that the pattern is quite different: there are many more deaths at younger ages and many fewer very old people. The average age at death was under 50 in 1876, compared to 64 in dataset A.

This change in longevity has a genealogical impact. Litvak naming traditions were that children were named, when possible, for deceased ancestors. In the first half of the 19th century one can often see names repeated every other generation - by the time the first grandson was born the grandfather was often already dead and their name was therefore available for reuse in the family. But as time passed the grandparents were increasingly still alive - possibly for the birth of every grandchild - and so the first opportunity to reuse a name might be for a great grandchild or even a great great grandchild. The neat naming patterns break down and our task becomes that much more interesting.

To illustrate how things are today I also show 2002 deaths. People now live much longer. But note there is also a dip at 71-75. This suggests we may be seeing a culling effect with slightly more infirm people dying in their late 60s leaving a slightly healthier group with a slightly enhanced chance of survival.

These datasets have a lot more information to offer - for example, in most cases a cause of death is given and analysis of this information could be interesting. This sort of analysis, which can give us the possibility of new insights into the pattern of the lives of our ancestors is only possible with the translation of complete runs of records: this is a new sort of reason to support LitvakSIG's work.

More on marriage

I am very pleased to be able to acknowledge the donation by Maria Krane, a former district co-ordinator, of an interesting set of records. The dataset contains records for 242 marriages and 7 divorces of people connected with Salakas from 1877 to 1915. This will be made available to qualified donors once editing is completed.


I have carried out a preliminary analysis of the data: and this posting is to share some of the more interesting findings. Note that the data before 1886 and after 1913 is sparse.


Age at marriage

There is data on age on marriage for most of the women and about half the men.

The average age of men on marriage was about 26 and for women about 23 1/2. The chart on left (click to expand) shows the annual average age for men and women and the rolling 5 year averages. The annual variability is due to the small number of marriages in each year. But the pattern over the period becomes clear when we look at the rolling average: the average age increases by about 1 year every decade. This appears to be the continuation of a longer term trend. From about 1909 the average age drops for men and from 1911 for women. This may be because mass emigration had by those times eliminated many older men and women from the marriage pool, as the eldest generally left first: only younger sisters and brothers were still in Russia.

The youngest age for men was 19 and for women 18. The oldest man was 40 - and he was a widower. The oldest first marriage for a man was 32. The oldest woman on first marriage was 34.

The chart on the right (click to expand) shows the frequency of each age for men and women.



You'll see that the male age curve is about 2 1/2 years to the right of the female curve.





Spouses' hometowns



For most of the marriages we can identify the hometowns for both the bride and groom. In only five cases did a woman from Salakas marry someone from elsewhere - and in every case the groom was from Dvinsk/Daugavpils. So was this also the case for the brides of Salakas men?




The chart on the left (click to expand) shows the home towns of 186 brides of Salakas men over the 25 years from 1886 t0 1910 in five year blocks. I grouped the data together to see trends more easily.

The answer is no.

It is interesting that the total number of marriages rises over 20 years and then drops in 1905-10. This may be an effect of mass emigration - or it may reflect differential survival of records.

In summary, for the whole period to 1913 the statistics are :

• Salakas 17%
• Other Zarasai district 22%
• Vilkomir district 9% (about 1/2 from Utian)
• Other Kaunas Gub 4%
• Dvinsk district 20%
• Other Vitebsk Gub 3%
• Vilna Gub 11% (about 1/2 from Sventysyan district)
• Other Pale Guberniyas 6% (Minsk, Grodno, Mogilev, Podolia, Suwalki)
• Latvian provinces 5% (Courland and Livland)
• Other Russian Empire 5%

Perhaps unsurprisingly 3/4 of brides were from the district and the immediately adjacent districts. But there are some very far found brides including Tiflis in Georgia , Dagestan in the Caucasus and Kamanets in Podolia.

Interestingly there were very few brides from the major Lithuanian cities. Over the period there were only 4 from Kovno city and 2 from Vilna. By comparison, there were 38 from Dvinsk and 4 from Vitebsk city. This is reflective perhaps for most people of effective local contact networks (perhaps a local shadchan), but for a few, much broader contacts across Russia. These may have been family, trade, educational or religious affiliations. In one case, in my family, the groom had already emigrated internally to a farm colony in Minsk Guberniya and found a wife there.

In the vast majority of cases the marriage took place at or near to the bride's hometown. And with kest in mind, we can surmise that perhaps 80% or more of the new families started their marital lives outside Salakas. Many indeed in this time period will have left for the US, Canada, the UK or South Africa within a few years after the marriage. But even if they stayed in Russia the associated vital records (such as children's births) will probably be found elsewhere, while the family registration and the list entries will be mainly in Salakas.

The rabbi

For many marriages the officiant (and in some cases the names of witnesses) are given. For Salakas from 1891 the rabbi is G. Lopaiko. This is probably Gershon Lopayko described in the 1912 box taxpayers list as a "wealthy peddler". It's probable that this is the Movsha Gershon Lopayko whose father Binel brought the family from Onikchty/Anysksciai by 1880 (according to the 1887 Family List). On the 1901 town plan Gershon owns 2 properties, number 2 and number 19. Number 2 is a building in the middle of the northern part main square. On the map it is shown with a balance drawn on top - so this seems to be the market. Number 19 is a property just off the northwest corner of the main square. Neither of these buildings now survive.

Does anyone know anything more about the Lopayko family?

Some statistics: mortality in reverse

A warning to readers: this is a somewhat morbid discussion. I guess a consultancy at a life insurance company in 2006 has left its mark.

I recently was given a database of deaths containing about 500 deaths covering a district for 17 years in the 1920/30s. I found 2 people on the list who were from one family I'm researching. Only 2 people for an area that included that family's home shtetl! What was going on? Could I extract any more information from the database?

The average age of death was about 65 (excluding deaths over infants less than 12 months old). This means the overall annual death rate should be about the reciprocal of this: about 1.5% annually. This is a fairly crude basis: any actuary interested should let me know how this ought to be done.

Over 17 years this means that 25.5% of adults might be expected to die. So for any number of deaths the one might expect that they come from a group that is about 4 times that number. So 2 deaths in a family means that the size of the family should be about 8.

But this is only a statistical expectation, the real number could be higher (the family were healthy/wealthy) or lower (the family were poorly/poor). There is a strong correlation between wealth and mortality. What could be the range be? Statisticians use confidence intervals to define a range of likely results. I determined a 90% confidence range, using a roundabout method. I concluded that 2 deaths might mean between 2 and about 22 in the family.

I then looked at the numbers for other numbers of deaths over 17 years using the same overall death rate. My estimates are below [sorry about the layout]:

No of deaths .....Expected family size ....Lower limit ........Upper limit

5 ........................20 ....................9 ..................36
10 .......................39 ...................25 ..................58
20....................... 78................... 61.................. 98


Is this useful information? It told me at least that with high probability very few people in the target family had stayed in the home shtetl over this period - probably around 8 and almost certainly fewer than 22. This suggests that my failure to find significant numbers of people from this family in the home district in the interwar period and in Holocaust records (compared to the hundreds elsewhere at the time) is not a result of poor research technique - almost everyone had gone.

If anyone can suggest a better actuarial approach to this problem, please let me know.