Previous Page Table of Contents Next Page


  1. Selected data for countries, simplified to predominant U.S. centile within age bands (Table II)

  2. Interpretation of centile data as basis for generalised height and weight curves

  3. Generalised height and weight curves in tabular form (Table III)

  4. Summary of countries represented by generalised growth curves (Table IV)

  5. Variation in stature and weight

  6. Weight for height


C E N T I L E S90    Yugoslavia   
80Germany   Germany  Netherlands
 Sweden   Sweden   
70BelgiumNorway NetherlandsBelgiumGermanyNetherlandsNorway
 Finland  NorwayFinlandNetherlands  
 Switzerland   SwitzerlandNorway  
 TurkeyNetherlandsNetherlands TurkeySwitzerlandNorwayHungary
   Norway  Yugoslavia  
  ItalySwedenSwitzerland Finland Sweden
  PolandSwitzerland  Spain Switzerland
  SwedenTurkey  Turkey UK
40 FinlandFinlandSpain CzechoslovakiaSwitzerlandBelgium
  UKPolandTurkey HungaryTurkeyCzechoslovakia
  YugoslaviaUKUK ItalyUKItaly
   YugoslaviaYugoslavia UKYugoslaviaTurkey
30CzechoslovakiaCzechoslovakiaBelgiumBelgiumCzechoslovakia BelgiumPoland
  HungaryCzechoslovakiaCzechoslovakia  CzechoslovakiaSpain
  SpainHungaryPoland  Finland 
   Spain   Hungary 
C E N T I L E S70YugoslaviaNorway  GermanyGermany  
60BelgiumGermanyTurkey BelgiumNorwayTurkey 
 FinlandNetherlands  FinlandSpain  
 GermanySpain  SwitzerlandSweden  
 PolandSwitzerland  TurkeySwitzerland  
 Sweden   UKTurkey  
 Switzerland    UK  
 Turkey    Yugoslavia  
50 BelgiumGermanyGermanyPolandBelgiumGermanyCzechoslovakia
  ItalyNetherlandsNorway CzechoslovakiaNorwayFinland
  SwedenNorwayTurkey FinlandPolandNetherlands
  TurkeyPoland  HungarySpainNorway
  UK   ItalySwedenSweden
  Yugoslavia    UKTurkey
40 CzechoslovakiaBelgiumHungary  BelgiumBelgium
  FinlandHungaryNetherlands  CzechoslovakiaGermany
  HungaryItalyPoland  HungaryPoland
   SpainSwitzerland  Italy 
   SwedenUK  Netherlands 
   SwitzerlandYugoslavia  Switzerland 
30Czechoslovakia CzechoslovakiaBelgiumCzechoslovakia Finland 
C E N T I L E S50 +Nigeria (60–70) privilegedNigeria privileged  Nigeria (60) privilegedChad  
 South Africa White (90)South Africa White  ChadUganda  
 Uganda   SomaliaSouth Africa White  
 Somalia   Uganda   
     South Africa (Food supplement)   
     South Africa White (90)   
     Egypt privileged   
40South Africa (Food supplement)ChadSouth Africa (White) South Africa (Pedi)Uganda  
  South AfricaTunisia     
30 Somalia   SomaliaNigeria privilegedSomalia
20 UgandaSouth Africa  LiberiaEthiopiaNigeria
  SudanEthiopia  BeninSomaliaUganda
   Sudan   TanzaniaEthiopia
10Liberia privilegedEthiopiaSomaliaNigeriaNigeria rural   
 AngolaLiberiaLiberia South Africa poor   
5Nigeria poorTunisia TanzaniaSudanNigeriaSouth Africa (Pedi)Angola
 Uganda poor    SudanSudan 
3South Africa poorAngolaAngolaUganda poorTunisiaAngola  
 Liberia  AngolaTogo   
<3      Angola 
C E N T I L E S50 +Nigeria privilegedNigeria privilegedTunisiaSouth Africa (White)Nigeria privilegedUgandaNigeria privilegedChad (70)*
 South Africa (Food supplement)South Africa (White)  ChadSouth Africa  
 South Africa (White) (90)*   South Africa (Food supplement)Nigeria privileged  
     South Africa (White) (90)*   
40UgandaSouth Africa (Pedi)Nigeria privileged  South AfricaUgandaUganda
   South Africa (White)     
30   NigeriaSouth AfricaChadLiberiaEthiopia
      Liberia Liberia
20 SomaliaSouth Africa (Pedi)  EthiopiaSomaliaSudan
  SudanSudan  SomaliaSudanSomalia
10NigeriaUganda poorSomaliaSudanNigeriaSudanSouth AfricaTanzania
 South Africa (poor)  Ethiopia  Ethiopia 
 Sudan  C.A.R.    
3   South Africa (Pedi)    

* Figures in brackets, e.g. (70), indicate actual centile when above 50th.

C E N T I L E S70Taiwan   Taiwan   
60 India privileged      
50  India privileged  India privilegedIndia privilegedIndia privileged
30Hong KongTaiwanJapanIndia privilegedHong KongTaiwan  
 Japan  JapanSingapore   
 Singapore   Thailand   
20PhilippinesHong KongHong Kong PhilippinesJapanHong KongHong Kong
 ThailandJapanTaiwan  New GuineaJapan 
  Thailand    Taiwan 
10SingaporeIndia   Hong Kong Japan
  New Guinea   Singapore Taiwan
  India   Thailand  
5 PhilippinesIndonesiaHong Kong PhilippinesPhilippinesIndonesia
  Indonesia Japan  IndonesiaIndia
3India Philippines IndiaIndia  
<3  ThailandBurma  ThailandBurma
   IndonesiaPhilippines  New GuineaPhilippines
    New Guinea  IndiaNew Guinea
C E N T I L E S70Taiwan   Taiwan   
60 India privileged      
50  India privileged  India privilegedIndia privileged 
40  Japan     
30ThailandJapan India privileged  JapanJapan
        India privileged
20JapanTaiwan  JapanJapanTaiwanTaiwan
 New Guinea  ThailandTaiwan   
10Hong KongHong KongHong KongJapanHong KongHong KongHong KongHong Kong
 SingaporeThailandIndonesia SingaporeIndonesiaIndonesia 
5 PhilippinesPhilippinesHong Kong New GuineaNew Guinea 
3India  India ThailandThailand 
<3New Guinea IndiaIndia  Thailand 
 India     New Guinea 
C E N T I L E S70Colombia   Colombia   
60 Puerto Rico   Colombia  
      Puerto Rico  
50 ColombiaPuerto Rico Neth. AntillesCubaPuerto Rico 
  Cuba   Jamaica  
      St Kitts - N.A.  
40Neth. AntillesGuatemala well off Puerto Rico Neth. AntillesJamaica 
  Neth. Antilles    St Kitts - N.A. 
  St Kitts - N.A.      
30 JamaicaCubaGuatemalaJamaicaUruguayCubaPuerto Rico
  UruguayUruguay  Caribbean combinedGuatemala well offSt Vincent
   Guatemala   Neth. Antilles 
   Neth. Antilles     
   St Kitts - N.A     
20JamaicaGuyanaJamaica UruguayGuyanaUruguayCuba
 UruguaySt Vincent  St VincentSt VincentSt VincentUruguay
  Caribbean combined    Caribbean combinedGuatemala
        well off
10  GuyanaCubaChile Chile 
   St VincentUruguay  Guyana 
   Caribbean combinedGuyana    
5       Colombia
3     Chile  
<3GuatemalaGuatemala ColombiaGuatemala  Chile
C E N T I L E S70ColombiaPuerto RicoPuerto Rico ColombiaPuerto Rico  
60Neth. AntillesColombia   ColombiaPuerto Rico 
  Guatemala well off   Guatemala well off  
50   Puerto Rico UruguayGuatemala well off Uruguay
40UruguayCubaUruguayUruguayUruguayNeth. AntillesCubaPuerto Rico
  UruguayGuatemalaGuatemalaNeth. AntillesSt Kitts - N.A.Uruguay 
  Neth. Antilles    Neth. Antilles 
30 JamaicaCuba JamaicaJamaicaJamaicaColombia
  St. Kitts - N.A.   Caribbean combinedSt. Kitts - N.A.Cuba
        St Vincent
20JamaicaSt VincentJamaicaCubaSt VincentGuyanaSt Vincent 
  Caribbean combinedNeth. AntillesJamaica St VincentCaribbean combined 
   St Kitts - N.A.     
10 GuyanaGuyanaColombia  GuyanaGuatemala well off
   St Vincent     
   Caribbean combined     
5GuatemalaGuatemala Guyana Guatemala  
    Caribbean combined    
3    Guatemala   
C E N T I L E S50 W. Australia  W. AustraliaW. AustraliaNew ZealandNew Zealand
  New Zealand  New ZealandNew Zealand  
40  W. AustraliaNew Zealand    
   New Zealand     
30Fiji  FijiFijiFijiW. Australia 
20 Fiji     Fiji
10  Fiji   Fiji 
C E N T I L E S70    W. AustraliaNew Zealand  
     New Zealand   
60W. AustraliaW. Australia    New ZealandFiji
 New ZealandNew Zealand      
50Fiji W. Australia FijiW. Australia  
   New Zealand     
40 Fiji   FijiW. Australia 
30   Fiji    
20  Fiji   Fiji 



Table II was created by converting the reported mean heights and weights from different countries into the nearest U.S. centiles, within the age groups 0–4.9; 5–9.9; 10–14.9 and 15-adult. The steps involved are described in the section entitled “Outline of method”. The final allocation of a country to a given centile was often a matter of judgement as to how much weight should be given to each of a number of differing values in the literature and therefore the results are approximate only.

However, approximation is essential to this stage of the study whose aim is to produce a relatively small number of curves which will give a reasonable representation of the heights and weights of all the populations studies. Towards this end a further approximation was made.

It was assumed that the U.S. 50th centile curve is a reasonable representation of the true growth curves of those countries from which the most reliable reported means lie consistently between the 40th and 60th centiles (inclusive). This is justified by the fact that the 40th and 60th centiles differ from the 50th by only 1.8cm at age 18 and by smaller amounts at earlier ages. The available data from many countries do not permit estimation of the population mean within greater limits of accuracy than this. On the basis of a similar argument it will be assumed that observed differences between other adjacent pairs of centiles (e.g. 30th and 20th) do not reflect real population differences unless there is clear evidence to the contrary.

The interpretation of table II in relation to separate regions and countries is as follows.


Male stature

The majority of European countries lie consistently between the 40th and 60th centiles and are therefore adequately described by the US 50th centile curve, henceforth referred to as curve MS 50 (male stature 50th centile).

Children under five years of age in Finland, Germany, Sweden and Yugoslavia are distinctly taller on average than the Americans. These countries might be better described by a curve beginning at the 80th centile and gradually descending to meet the 50th centile at age 5. However, with this reservation, curve MS 50 may be used.

The males of Belgium show a progressive drop from the 80th centile in early childhood to the 30th at 10 years and beyond and may require a separate curve (MS 80/30) for their adequate description. The curve for Poland shows a similar but less marked decline. The data do not provide a satisfactory basis for deciding whether Poland would be better described by curve MS 50 or MS 80/30 and do not justify the creation of a separate curve.

The statures of Czechoslovakian males are best described by a curve following the 30th centile (MS 30). This curve would probably also be reasonably representative of Spain and Hungary.

There is no reason to think that the US standard deviation (adjusted as described in section 4.v.) would not be appropriate to all the above curves, with the possible exception of MS 30 as applied to Spain. The tallest and shortest samples of boys in Europe have been derived from the Burgos and Guadalajara regions respectively of Spain. This suggests that the variation within the Spanish population may be greater than in other European populations. However, any attempt, based on existing data, to calculate an alternative standard deviation for Spanish males would not be justified.

Female stature

For a number of countries, the mean statures of females remain consistently within the limits of the 40th and 60th centiles and may be adequately represented by a curve, FS 50, which follows the female 50th centile. This curve would also be a reasonable approximation to the data from Yugoslavia, Sweden and Switzerland.

The females of Germany, the Netherlands and Norway would be better represented by a curve following the 70th centile (FS 70) than by FS 50.

Belgium and Finland might be better represented by a curve FS 70/30/50, beginning at the 70th centile and dropping to the 30th at age 12 with a rise thereafter to reach the 50th at age 16.

A curve following the 40th centile (FS 40) would seem to be the best generalisation to cover Czechoslovakia, Hungary and Italy. The error would probably be slight if this curve were also applied to Poland and Spain.

Male weight

A curve following the 50th centile for weight (MW 50) is an adequate description of the mean weights of nearly all the European populations which have been studied.

Curves beginning at the 70th centile and joining the 50th at age 5 or 10 are required for Yugoslavia and Norway respectively but MW 50 is a sufficient approximation.

Belgium and Finland would be represented by a curve (MW 60/30) beginning at the 60th centile and reaching the 30th at age 15.

Czechoslovakia requires a curve (MW 30) which follows the 30th centile.

Female weight

A curve lying mainly on the 50th centile but dropping temporarily to the 30th between ages 10 and 15 would seem to be appropriate for Finland but curve FW 50, following the 50th centile, represents the majority of the population.

All the remaining European countries for which we have data would be adequately described by curve FW 50.


Male stature

White South Africans and privileged Nigerians are probably adequately represented by curve MS 50.

Uganda, Somalia and Tanzania have children at the 40th or 50th centile but fall to lower levels with increading age whilst the adults are in the region of the 3rd, 5th or 10th centiles. These populations might be collectively represented by a curve dropping (MS 50/5) gradually from the 50th to the 5th centile.

A curve (MS 10) following the 10th centile represents, as far as it is possible to judge from the available data, a reasonable approximation to the statures of males in Liberia, Sudan, Benin, Ethiopia, Liberia and Nigeria. A curve on the 3rd centile (MS 3) might be more appropriate to Angola and the poorer parts of Uganda. However, the Angolan data are based mainly on groups with very short stature and may not represent the whole population.

The data on Tunisia are conflicting and difficult to interpret. It seems unlikely that there is a change during childhood from the 5th to the 40th centile and these two levels are suggested by different samples. There may be two distinct segments in the population, one of which follows approximately MS 50 and the other MS 10.

Female stature

Curve FS 50 may be applied to the white population of South Africa. They are nearer the 90th centile in infancy but rapidly fall to the 50th. The people of Uganda also appear to follow this curve to about age 15 when they drop to the 20th centile.

The more privileged members of the Nigerian population and the females of Somalia might be represented by a curve following the 30th centile (FS 30) although they are close to the 50th in infancy.

A curve following the 20th centile (FS 20) would seem to give a reasonable representation of the female populations of Liberia, Benin, Tanzania, Ethiopia and possibly Nigeria. The Sudan would probably be better represented by a curve on the 5th centile (FS 5).

The Pedi of South Africa are best represented by a curve dropping gradually from the 50th centile to reach the 5th at about age 11 (FS 50/5). The Angolan population may be best represented by a curve on the 5th centile (FS 5) but this is uncertain because, as in the case of males, we lack evidence about the taller members of the population. The short stature of some Angolan groups will probably result in the statures of the whole population being distributed in a non-Gaussian way. A bimodal distribution is most likely and it may be correct to regard the population at each age as being made up of two distinct normal distributions. One of these would represent the “pygmy” groups and another would represent those with statures more typical of other African populations. In the absence of direct information as to the growth curve of this latter population it might be reasonable to adopt FS 20 which seems to have the widest applications in Africa.

Male weight

Curve MW 50 provides a satisfactory description of the more privileged Nigerians and White South Africans and the limited data which are available suggest that it might also apply to Tunisia and privileged Ugandans.

The Pedi of South Africa apparently drop through the centiles with increasing age (curve MW 50/3).

A curve following the 10th centile (MW 10) would represent, within a reasonable approximation, the male populations of Somalia, Sudan, Liberia and Ethiopia. It might also be appropriate to Tanzania in the Central African Republic (Rwanda) as well as the poorer sections of the populations of Nigeria, Uganda and South Africa.

Female weight

The weights of those populations of African females, for whom data exist, can be approximately represented by two curves.

FW 50 would apply to White South Africans, although their mean stature is well above the 50th centile in the first five years; to privileged Nigerians and to Ugandans. This curve might also be an adequate representation of the people of Chad although the data indicate a mean stature below the 40th centile in the 5–10 age group and a mean above the 50th centile for adults.

Curve FW 20, on the 20th centile, describes the people of Somalia and Sudan although their mean statures are nearer the 5th centile during infancy. Ethiopia, Liberia and possibly Tanzania also fit this curve reasonably well.


Male stature

None of the Asian male populations for which we have data follows curve MS 50.

Taiwan requires a curve dropping gradually from the 70th to the 5th centile. However, curve MS 50/5 is a sufficient approximation within the limits of the data. Some privileged males of India follow MS 50 up to about the age of 14 years and then gradually drop to the 30th centile, following a curve MS 50/30. Poorer Indian populations however would be represented by a curve beginning at the 3rd centile, rising to the 10th and then dropping to the 5th. However, given the limitations of the data which are available it would seem to be a legitimate approximation to replace this by a curve following the 5th centile throughout (MS 5). This curve might also be valid for Indonesia.

At a similar level of approximation, Hong Kong and Japan would be represented by a curve MS 30/5, following the 30th centile up to age 13 and then gradually dropping to the 5th.

The Philippines require a curve (MS 20/3) beginning at the 20th centile and dropping gradually to the 3rd. This would probably represent a satisfactory approximation to the curves for Singapore, Thailand, Indonesia and New Guinea. Burma, for which we have only adult data, might also be fitted by this curve.

Female stature

The mean statures of privileged Indian females at all ages appear to be adequately described by curve FS 50. The major part of the Indian population may be best described by a curve (FS 3) on the 3rd centile but the data clearly do not adequately represent this very varied population. The standard deviation is almost certainly greater than that of the U.S. standards although we have no basis for even an approximate estimate of its value.

Curve FS 20, following the 20th centile, is the most appropriate simplified representation of the available data from Hong Kong, Japan and Singapore.

Taiwan requires a curve, FS 70/10, dropping gradually from the 70th to the 10th centile whilst one dropping from the 20th to the 3rd (FS 20/3) would be the best simplification of the data from Thailand, Indonesia, the Philippines and New Guinea.

Male weight

Privileged Indian males follow curve MW 50 up to about the age of 14 years and then fall to the 30th centile whilst the majority of the population would seem to be better described by MW 5, with the same reservations as are stated above in relation to stature.

A curve, falling gradually from the 70th to the 10th centile is required for Taiwan, but only infants are at the 70th centile and curve MW 50/3 may be used in conjunction with the latter part of MW 10.

The 10th centile curve MW 10 is the best simplified representation of the data from Hong Kong, Singapore, the Philippines and Indonesia. The Japanese have mean weights at the 20th centile during infancy and at the 10th in adult life but, between the ages of 5 and 14 years they rise gradually to the 40th centile. They therefore require a special curve MW 20/40/10.

Thailand seems to require a curve dropping gradually from the 30th centile to the 3rd. Curve MW 30 may be applied up to age 8 years and MW 50/3 at later ages.

Female weight

Privileged Indians follow curve FW 50 up to the age of about 14 years and then drop to the 30th centile whilst the remainder of the Indian female population, subject to the reservations stated above, is best represented by curve FW 3.

The 20th centile curve (FW 20) gives an adequate representation of Japan, Hong Kong, the Philippines and Indonesia. It is also appropriate to Taiwanese over the age of 5 years although infants are said to be at the 70th centile.

The 3rd centile curve, FW 3, would adequately represent the greater part of the populations of Thailand and New Guinea although infants from the former country are at the 20th centile and those from the latter are below the 3rd.


Male stature

Curve MS 50 is probably an adequate description of the statures of males in Puerto Rico.

The Netherlands Antilles require a curve at a slightly lower level, following the 40th centile for most of the first ten years and then dropping to the 30th. This curve might also be applied to St Kitts, Nevis and Anguilla. Curve MS 50/30 may be taken as a convenient approximation.

The best approximation to the remaining Caribbean islands and Guyana appears to be provided by curve MS 20; following the 20th centile.

A curve dropping gradually from the 70th to the 5th centile would give an approximate representation of the data from Colombia, Cuba and Uruguay. Given the limitations of the data and the fact that only small children are at the 70th centile, curve 50/5 may be used. Well off Guatemalans also appear to fit this curve but the rural poor are closer to the 3rd centile and would be represented by a curve MS 3. This curve may also be the most appropriate to the rural poor of other South American countries. In these countries the standard deviation for the whole population is probably much greater than the corresponding SD represented in the US standards.

Female stature

Curve FS 70/10 gives a reasonable description of the data from Colombia, the Netherlands Antilles and Cuba. The same curve may be applied to Jamaica and Uruguay after the age of 5 years but the infants of these countries are at the 30th and 20th centile respectively.

St Kitts-Nevis-Anguilla may be fitted by curve FS 50 although satisfactory data are lacking at both ends of the age range. The curve for Puerto Rico drops to the 30th centile after the age of 15 years and therefore curve FS 30 may be used after this age.

Curve FS 20 may be applied to St Vincent, Guyana and to the Caribbean Islands in general. It is also appropriate to the well off sector of the Guatemalan population.

FS 5, following the 5th centile, is probably the most appropriate curve for Chile where the data suggest very wide variation.

Male weight

A curve dropping gradually from the 70th to the 20th centile is the best generalised representation of Colombia, Cuba and the Netherland Antilles. Curve MW 60/30 is an approximation to this.

The males of Uruguay and well off Guatemalans are reasonably represented by curve MW 50 but Puerto Rico requires a curve following the 70th centile until the age of 15, after which it drops to the 50th (MW 70/50).

Curve MW 20 applies to Jamaica, St Kitts-Nevis-Anguilla and St Vincent on the assumption that the pattern would not be altered if suitable data were available from the latter two countries for the final age group. This may not be the case because both Guyana and the combined Caribbean data show a drop to the 5th centile after the age of 15 years. St Kitts-Nevis-Anguilla and St Vincent may follow the same pattern.

The poorer section of the Guatemalan population probably follows curve MW 5, on the 5th centile.

Female weight

A curve FW 70/10, dropping gradually from the 70th to the 10th centiles, may be applied to Colombia, Cuba and well off Guatemalans.

Curve FW 50 would apply to Uruguay and the Netherlands Antilles, whilst FW 20 could be used to represent Jamaica, St Vincent, Guyana and the combined Caribbean islands.

St Kitts-Nevis-Anguilla lies on the borderline between FW 50 and FW 20. The poorer part of the Guatemalan population follows curve FW 5.


Male stature

Curve MS 50 represents Australia and New Zealand whilst MS 20 represents Fiji.

Female stature

Curve FS 50 represents New Zealand and this curve is probably also the best representation of Australia although girls in Western Australia drop to the 30th centile between the ages of 10 and 15.

Fiji is represented by curve FS 20.

Male weight

Australia and New Zealand follow curve MW 50 but Fiji is better represented by a curve dropping gradually from the 50th to the 20th centile. MW 60/30 is an approximation to this.

Female weight

New Zealand follows the 70th centile (curve FW 70). Western Australia and possibly the whole of Australia may be adequately represented by FW 50 although West Australian infants are at the 70th centile.

Fiji follows an unusual pattern with a gradual drop from the 50th to the 20th centile during the first 15 years and then a rise to the 60th centile FW 50/20/60.


The North American data have been used as the basis for comparison and, therefore, both sexes in North America follow by definition the 50th centile curves for both height and weight in MS 50; FS 50; MW 50 and FW 50.


Male Stature Curves — Mean Values (cms)
Age (years)MS 50MS 80/30MS 30MS 50/5MS 10MS 5MS 50/30MS 30/5MS 20/3MS 20 
Female Stature Curve - Mean Values (cms)
Age (years)FS 50FS 70FS 70/30/50FS 40FS 30FS 20FS 5FS 50/5FS 70/10FS 20/3FS 3
Male Weight Curves - Mean Values (Kg)
Age (years)MW 50MW 60/30MW 30MW 50/3MW 10MW 20/40/10MW 70/50MW 20MW 5
Female Weight Curves - Mean Values (Kg)
Age (years)FW 50FW 20FW 3FW 70/10FW 5FW 70FW 50/20/60


Summary of countries to which curves may be applied
 AustraliaBelgiumCzechoslovakiaColombiaBeninIndiaAngolaIndia (privileged)Hong KongBurmaMost Caribbean
 CanadaPoland CubaEthiopia BoliviaNetherlands AntillesJapanIndonesia 
 Finland  SomaliaLiberia BrazilSt Kitts New GuineaFiji
 Germany  TaiwanNigeria Guatemala  Philippines 
 Italy  TanzaniaSudan Uganda (poor)  Singapore 
 Netherlands  Uganda     Thailand 
 New Zealand  Uruguay       
 Nigeria (privileged)          
 Puerto Rico          
 South Africa (white)          
 AustraliaGermanyBelgiumCzechoslovakiaNigeriaBeninAngolaSouth AfricaColombiaIndonesiaIndia
 CanadaNetherlandsFinland SomaliaCaribbean combinedChile CubaNew Guinea 
 India (privileged)NorwayHungary   Indonesia JamaicaPhilippines 
 New Zealand Italy  EthiopiaPhilippines Netherlands AntillesThailand 
 Puerto Rico Poland  FijiSudan    
 St Kitts Spain  Guyana  Taiwan  
 South Africa (white)    Hong Kong  Uruguay  
 Sweden    Japan     
 Switzerland    Liberia     
 Turkey    Nigeria     
 Uganda    St Vincent     
 U.K.    Singapore     
 U.S.A.    Tanzania     
Summary of countries to which curves may be applied
 AustraliaBelgiumCzechoslovakiaSouth AfricaCentral African RepublicJapanPuerto RicoJamaicaGuatemala
 CanadaColombia TaiwanEthiopia  St KittsIndia
 GermanyCuba ThailandHong Kong    
 Guatemala (privileged)Fiji  Indonesia    
 HungaryFinland  Liberia    
 IndiaNetherlands  Nigeria (poor)    
 ItalyAntilles  Philippines    
 Netherlands   Singapore    
 New Zealand   Somalia    
 Nigeria   South Africa (poor)    
 Norway   Sudan    
 Poland   Tanzania    
 South Africa (White)   Uganda (poor)    
Summary of countries to which curves may be applied
 AustraliaCaribbean combinedIndiaColombiaGuatemalaNew ZealandFiji
 BelgiumEthiopiaNew GuineaCuba   
 CanadaGuyanaThailandGuatemala (well off)   
 ChadHong Kong     
 India (privileged)Philippines     
 ItalySt Vincent     
 Netherlands AntillesSudan     
 Nigeria (privileged)Taiwan     
 South Africa (White)      
 Uganda (privileged)      


The standard deviations of stature and weight reported for each age group in different countries are shown in table 1.

In the case of stature, the standard deviation is a valid measure of the variation within each sample. However, many of the reported samples contain very small numbers of subjects and are unrepresentative of their countries. They therefore do not give a reliable indication of the variation within the populations which they represent. The publications, collectively, do not give enough information to permit the calculation from their separate standard deviations of a single combined standard deviation which could be taken as a reliable estimate for each country. The reader therefore has two alternatives. He may make a subjective judgement for each country, based on the data in table 1. Alternatively, he may make use of the United States data which have been used as a basis for comparison throughout this study. These data include standard deviations which are at least reliable estimates for one varied population. These might be applied to the growth curves for other populations if the mean stature, rather than age, is used to determine the point at which each standard deviation is applied.

Table V shows the standard deviations corresponding to the US 50th centile at 10cm intervals of stature. This may be used in conjunction with the generalised growth curves (Table III) by applying to any point on the curve, that standard deviation which corresponds to the nearest US 50th centile stature.

For example, consider the male population of Somalia (East Africa) whose mean statures are represented by curve MS 50/5. At age 9, the mean stature on curve MS 50/5 is 129.5 cms. By reference to table V, the standard deviation corresponding to a mean stature of 130 cm is 5.5 cms. This value, in the absence of more appropriate data, may be applied to the population of Somalia. Gallo and Mestriner (1980) quote a standard deviation of 6.7 from their sample of 49 ten year old Somalian boys. However their standard deviations at ages eight and ten were 5.9 and 5.6 respectively. The value of 5.5 is probably nearer than 6.7 to the true population value for ten year olds. Some cases in which the above procedure would not be valid are noted in the section on the interpretation of table II. Other such cases may be known to individuals with good knowledge of the structures of given populations. However, for many populations this procedure may give a closer approximation to the truth than may be obtained from limited local data. Use of the latter would often give a false impression of accuracy which does not actually exist.

The standard deviation is not a satisfactory measure of variation in weight within a sample or population. Weight usually has a skewed distribution which is more appropriately described by centiles than by the standard deviation. Unfortunately, very few authors present their data as centiles and, as in the case of stature, valid estimates for whole countries are seldom possible. Therefore a procedure analagous to that used in the case of stature may be employed, taking the 10th and 90th centile corresponding to given values of the 50th centile. The relevant centiles are shown in table IV. The assumptions and restrictions in the use of this method for weight are similar to those which apply in using the United States standard deviations for stature.

There is at least one further source of error in the case of weight. The U.S. 10th and 90th centiles have been calculated in relation to the corresponding 50th centile and not to the mean. However, the generalised curves for weight, presented in this study, are based on means because means and not 50th centiles are reported by most authors. An error is therefore introduced by substituting the 50th centile for the mean. However, in view of the unavoidable approximations made in constructing the generalised curves of the MW and FW series, the additional error should be negligible.

In the case of both height and weight, the variation increases at the ages when different individuals may be at different stages in their adolescent growth spurts. This may occur at a different range of ages, statures or weights in different populations. If it is found necessary to make allowance for this in the case of specific countries the standard deviations in table I provide a basis for doing so.


Standard Deviations corresponding to Mean Statures
Mean Stature (cms)Male SDFemale SD


10th and 90th Centiles corresponding to 50th Centiles for Weight
50th Centile (Kg)Male 10thMale 90thFemale 10thFemale 90th


There are very few studies that report the relationship between weight and height in individuals. Usually only the mean weights and heights of samples at given ages are presented. These values for samples from separate countries are shown in table I.

Table VII summarises the data for the different geographical regions. Within 10cm groups by stature it shows the mean of the separate sample means, together with the standard deviation of these sample means. These standard deviations illustrate the variation in mean weights amongst samples of a given stature for each geographical region. They do not represent the standard errors of the individual means. If a large number of samples were taken from a given region and the mean weight were calculated for each stature group, we should expect 95 per cent of these means to fall within ± 2 standard deviations of the overall mean.

It appears from the table that the variation in mean weight for height within each of the major geographical regions is greater than the variation between the regions. However, the data are insufficient to justify formal statistical testing of this hypothesis as many populations within each region are not represented.


Mean and Standard Deviation of Reported Mean Weights at 10cm intervals of Stature
 EuropeAfricaAsiaSouth America
170–179.964.0114.10   59.042.7263.051.48

Previous Page Top of Page Next Page