Population Numbers 1901-1964

Chagos Archipelago – Births, Deaths and Population Changes - 1901 – 1964

Notes:

1. Source: Birth & Death Registers, FCO Hanslope Park - Gifford, R., Second Affidavit of R D Gifford, 2000. p. 3.

2. All data processing was carried out in ‘R: A language and environment for statistical computing’ - R Core Team, R: A language and environment for statistical computing, 2013, R Foundation for Statistical Computing: Vienna, Austria.

1. Introduction

Near complete records of births and deaths exist in the registers from three atolls in the Chagos Archipelago (Diego Garcia, Peros Banhos, and Salomon) for the period 1901-1964. Population census numbers are also known for the years 1901, 1911, 1921, 1931, 1938, 1944, 1947 (not Peros Banhos), 1948, 1949, 1950, 1951, 1952, 1953, 1960, 1961, 1962 and 1964.

By combining these two sources of information it is possible to reconstruct pictures of changes in birth and death rates over time, population changes, and net emigration/immigration in the Chagos.

1.1 Patterns in Population Numbers

To determine if population numbers in the Chagos or on any one island might be influencing the birth and death rates, changes in population over time were first examined using data from the population censuses for 1901 to 1964 . [Note: Data for the first 40 years are sparse (5 censuses) compared to 11 censuses in the last 24 years]

Both this analysis and earlier work [Gifford, R. and R.P. Dunne, A Dispossessed People: the Depopulation of the Chagos Archipelago 1965–1973. Population, Space and Place, 2014. 20(1): 37-493] had identified one of these censuses, in 1962, as recording an anomalously low figure of 200 for Diego Garcia as opposed to 517 the year earlier and 483 two years later. As discussed in Gifford & Dunne this value was treated with caution and so excluded from the analyses.

For the Chagos and all atoll groups except Salomon there was no significant increase or decrease in population over the period 1901 – 1964. On Salomon although a simple straight line model appeared to fit the data (Figure 1.1) the residuals were non-normal and a plot of the data suggested a step-change in population from about 150 prior to the 1940s to 200 thereafter. It is not known if there was a reason for this, such as an expansion of the plantations on the atoll or some other requirement.

Between 1901 and 1964, a total of 2,785 people were born in Chagos, whilst deaths during this period amounted to 1,713. In the absence of emigration/immigration the population of the Chagos would thus have more than doubled between 1901 and 1964 (from 965 to 2,037) whereas in 1964 it stood at 993.

Figure 1.1. Salomon Population 1901- 1964 showing change over time. Straight line = Ordinary Least Squares (OLS) regression ±95% Confidence Intervals.

1.2 Patterns in Births

Patterns in annual birth rates were examined on the assumption that they may reasonably have been influenced by one or both of two factors, namely (1) the population size at any one time; and (2) changes over time, perhaps as survival rates altered due to better medical facilities and care.

The statistical model used to examine the birth data therefore included both Year and Population as factors, together with an interaction term (Year*Population). To achieve the most parsimonious model, non-significant terms were progressively removed until a significant model fit was obtained.

There was a linear increase in births in the Chagos over the period 1901 to 1964 (Figure 1.2) from ~34 to ~52 births per year but there was no significant population effect. Although this may seem counter-intuitive, a plausible reason for this is that since there was no significant variation in population numbers in the Chagos over the time period, no contributory effect of population on the birth rate could be detected.

Examining data by atoll group, it can be seen that this Chagos-wide increase was driven by increases in births on Peros Banhos (Figure 1.3) over time (from ~10 per year in 1901 to ~19 in 1964) but with no population effect, and by an increase in births on Salomon (from ~6 per year to ~12) due solely to an increase in population numbers (Figure 1.4). On Peros Banhos the birth rate was approximately 5 births per 100 people whilst on both Diego Garcia and Salomon this was about 4 per 100 people.

Figure 1.2. Chagos Births 1901 to 1964 showing change in births per year over time. Straight line = OLS regression ±95% Confidence Intervals.

Figure 1.3. Peros Banhos Births 1901 to 1964 showing change in numbers of births per year over time. Straight line = OLS regression ±95% Confidence Intervals.

Figure 1.4. Salomon Births 1901 to 1964 showing increases in births due to increasing population size. Straight line = OLS regression ±95% Confidence Intervals; line forced through the origin.

For Diego Garcia there was no increase over time, nor any relationship with population numbers, with a mean of 21 births per year over the 63 year period, although the birth rate appeared to have fluctuated with a drop in the 1920s (Figure 1.5).

The apparent increase in birth rate on Peros Banhos as a function of time may have been influenced by increasing health and improving child survival rates, but if this was responsible it is not clear why it did not also occur on Salomon and Diego Garcia. In the alternative, it is possible that on Peros Banhos the proportion of women of child bearing age in the population increased over the time period whilst the overall population numbers remained fairly stable. At the time of writing, a detailed breakdown of population in each of the censuses was not available to verify this, but figures from the 1921 census show that adult women formed 25% of the Peros Banhos population and by 1944 and 1952 the proportion of females (not broken down into adults and children) relative to the whole population was 51% [Wenban-Smith, N., Population of the Chagos 1820-1973, in Chagos News 2012. p. 18-22.]. The same proportions for Salomon were 43% in 1921 and 45-46% in 1944/52; and for Diego Garcia 28% in 1921 and 46-50% in 1944/52.

Figure 1.5. Diego Garcia Births 1901 to 1964 varying over time. Line = 7th order polynomial ±95% Confidence Intervals.

Throughout the 1960s when detailed population data are available for 7 years, the proportion of adult females in the population of each atoll remained relatively constant at approximately 20%. The first of these population surveys by Newton in 1964 [Newton, R., Report on the Anglo-American Survey in the Indian Ocean - 23 September 1964, 1964: London. p. 1-26] was one year before detachment of the islands to form the BIOT, and the remaining 6 surveys were from 1967 to 1971 by the BIOT Administrator, John Todd (Table 1).

Table 1. Proportion of adult females in the population 1964 to 1971. 1964 data from Newton and thereafter Todd [National Archives files FCO 141/1417, 1419, 1413, 1455, 1355]

1.3 Patterns in Deaths

Analysis of deaths proceeded on the same basis as that for births, and with the anomalous 1962 census data removed.

There was no straight line relationship between Chagos deaths and population or time either for the full data set (all years) or for the data with 1919 and 1920 removed (see discussion below). The data are best described by a second order polynomial for both datasets which show a small but non-significant rise in deaths from 1901 to 1920 - 30, and a significant decline thereafter to 1964 (Figure 1.6).

Figure 1.6. Chagos Deaths 1901 to 1964. Line = 2 order polynomial ±95% Confidence Intervals. Years 1919 and 1920 removed (see text).

Two years, 1919 and 1920, had particularly high death rates compared to others (57 and 58 respectively compared to a mean of 27 per year – data not shown in Figure 1.6). This period coincides with the 1918 worldwide Spanish Flu pandemic which is known to have reached Mauritius between May and October 1919 [Kohn, G.C., Encyclopedia of Plague and Pestilance: from Ancient Times to the Present. 2007: Infobase Publishing, and Fokeer, A.F., The Spanish Influenza in Mauritius. 1921: Mauritius Indian Times]. It is highly likely that it then found its way to the Chagos. The death rate was particularly high on Diego Garcia where there were 36 and 48 deaths in each of these years (mean death rate 1901-1964 = 16), but not so acute on the other atolls (Peros Banhos 13 and 6, mean = 7; Salomon 8 and 4, mean = 4). The progression of the disease on Mauritius is described in and extract from Kohn (2007) at page 257:

Figure 1.7. Diego Garcia Deaths 1901 to 1964. Line = AR(1) (p=0.0021) ±95% Confidence Intervals. 1919 and 1920 data removed.

Deaths on Diego Garcia exhibited a straight line relationship with time only, after correction for first order autocorrelation (Figure 1.7). Deaths per year declined from 20 in 1901 to 9 in 1964. The high numbers of deaths in 1919 and 1920 (36 and 48 respectively) were excluded from this analysis and were largely responsible for the peaks in Chagos deaths for those years.

Neither Peros Banhos nor Salomon deaths showed any straight line relationship with either population numbers or time. The mean number of deaths per year on Peros Banhos was 7, and for Salomon, 4 per year.

The reason for the decline in deaths on Diego Garcia but not the other atolls is not known. There is a possibility that it could either be linked to improving old-age care which contributed to longer life spans, but why this would apply to Diego Garcia only is unclear. Another explanation might be that more of the older generation on Diego Garcia migrated to Mauritius over time as the shipping link improved, particularly after WWII; possibly to take advantage of better healthcare there. The overall fall in the death rate for the Chagos as a whole is clearly driven by the Diego Garcia data.

1.4 Patterns in immigration/ emigration

As has already been noted, between 1901 and 1964, a total of 2,785 people were born in Chagos, whilst deaths during this period amounted to only 1,713. In the absence of emigration/immigration the population of the Chagos would thus have more than doubled between 1901 and 1964 (from 965 to 2,037) whereas in 1964 it stood at 993.

Estimates of the rate of immigration/emigration can be computed by comparing the annual differences in the births and deaths to the periodical census data, both at atoll level and for the Chagos overall. Population census data for each of the main atolls are shown in Figure 1.8.








Figure 1. 8. Population of the Chagos by atoll: 1901 – 1964 (1962 census omitted).

Table 2. Population of other atolls 1901 to 1931

Note: from 1901 to 1931 Six Iles and Trois Freres were also occupied but data are sparse and incomplete as shown in Table 2.

This analysis shows that during the first two decades of the 20th century there must have been net immigration to the Chagos, of about 6 persons per year from 1901 – 1911 and 1 person per year from 1911 – 1921. Over this time the population of the Chagos increased from 965 to 1,190, partly as a result of immigration and partly from births on the islands. Inspection of the data for the individual atolls of Diego Garcia, Peros Banhos and Salomon (Figure 1.9) show in particular the net immigration to Peros Banhos throughout this period which together with the island births increased the island population from 184 to 346 (Figure 1.8).

Figure 1. 9. Patterns in net Immigration (positive values)/Emigration (negative) - Chagos Atolls 1901 – 1964.

From 1921 until 1947 there was a drop in the Chagos population from 1,190 to 980 accompanied by emigration of between 11 and 24 persons per year. However, on the 3 main atolls there was initially little change in the 1920s to be followed by departures from Peros Banhos during the 1930s (~11 per year) but arrivals on Diego Garcia (~7) and Salomon (~10). The peaks in immigration on Diego Garcia in the late 1940s/ early 1950s and fluctuations on the other two atolls might have been an influence of WWII (it is also not known whether the census during the war years include servicemen stationed on the island). Immediately post War, 1948-1950, the Chagos population rose to 1,141 with the arrival of 126 immigrants.

The rising birth rate (34 in 1901 increasing to 52 in 1964) and declining death rate (post 1930 from ~30 per year to ~15 per year by 1964), together with a slight decrease in the population from about 1,150 in the early 1950s to 993 in 1964, is likely to have contributed to the pattern of net emigration of 49-55 persons per year from 1953 until 1964. It is possible that this was driven by the available employment in the Chagos plantations and also by the lure of Mauritius and the improved transport links post WWII.

The largest emigration during this decade appears to have been from Diego Garcia (about 34 per year for a ten year period: Figure 1.9) but if the census figures are correct there was then an influx of 74 persons in 1960-1 followed by a departure of 338 persons in 1961-2 and then a further influx of 170 person between 1962-4. These large fluctuations are not mentioned in the report by Newton when he visited the archipelago in 1964 and yet had they occurred they must have caused the plantation managers on Diego Garcia some concern and would have been expected to be relayed to him by them.

On the other two atolls, Peros Banhos and Salomon, population numbers and arrivals and departure were more stable, with emigration of about 7 and 9 persons per year respectively between 1953 and 1960. Both islands, however also appeared to show a spike in departures (52 and 21 respectively) in 1960-1.

Clearly this period in the early 1960s would deserve further investigation to ascertain whether these figures are correct and if so then what caused the larger than normal departures and arrivals. If, for example the census data for 1961 and 1962 are ignored then a different picture emerges (effectively ‘smoothing’ out these large swings): namely that of emigration of about 55 persons per year during the 1950s followed by a decrease to 37 per year in the early 1960s. At the island level this involved emigration of just 2 persons per year from each of Diego Garcia and Salomon in the early 1960s, but a much larger emigration from Peros Banhos of 33 per year (Figure 10).

Figure 1. 10. Patterns in net Immigration (positive values)/Emigration (negative) - Chagos Atolls 1901 – 1964. (1961 and 1962 census data omitted).

1.5 The BIOT

After the creation of BIOT in 1965 net emigration rose dramatically to 269 the following year, followed by an influx of 120 persons in 1967, before successive reductions of 124 in 1968, 168 in 1969, 405 in 1972, and finally 279 in 1973 when the islands were closed. In 1970 there was a small net influx of 15 persons.

1.6 Other Observations

Between 1901 and 1973 when the islands were finally cleared of their population, there had been a total of 2,918 births on the Chagos. It is likely that many of these would still have been alive in 1973 and would thus have been able to count Chagos as their birthplace. From 1914 until 1973, the British Nationality Acts then in force meant that anyone born in the Chagos acquired British citizenship or ‘Citizenship of the UK & Colonies’ which entitled them to a UK passport.


Page last edited 22 Nov 2018