AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 150:184–189 (2013)
Genetic Drift and the Population History
of the Irish Travellers
John H. Relethford1 and Michael H. Crawford2
1
Department of Anthropology, State University of New York College at Oneonta, Oneonta, NY 13820
2
Department of Anthropology, University of Kansas, Lawrence, KS 66045
KEY WORDS drift; population history; R matrix; Ireland
ABSTRACT The Irish Travellers are an itinerant
group in Ireland that has been socially isolated. Two
hypotheses have been proposed concerning the genetic ori-
gin of the Travellers: (1) they are genetically related to
Roma populations in Europe that share a nomadic lifestyle
or (2) they are of Irish origin, and genetic differences from
the rest of Ireland reflect genetic drift. These hypotheses
were tested using data on 33 alleles from 12 red blood cell
polymorphism loci. Comparison with other European,
Roma, and Indian populations shows that the Travellers
are genetically distinct from the Roma and Indian popula-
tions and most genetically similar to Ireland, in agreement
with earlier genetic analyses of the Travellers. However,
the Travellers are still genetically distinct from other Irish
populations, which could reflect some external gene flow
Studies of anthropological genetics examine how popu-
lation structure and population history affects genetic
variation in human populations. An interesting case
study is the genetic history of the Irish Travellers, an
itinerant population in Ireland. The Travellers (formerly
known as ‘‘Irish Tinkers’’) are a nomadic group that tra-
ditionally traveled in rural Ireland, often performing a
variety of odd jobs, including seasonal farm labor, as
well as providing other goods and services, such as horse
trading and tinsmithing. The Travellers make up less
than 0.2 percent of the population of Ireland, and they
have typically been socially isolated and endogamous. In
more recent times, the Travellers have had to cope with
urbanization and modernization (Gmelch, 1977). Even
then, the Travellers have remained socially isolated from
the rest of Irish society and are very endogamous
(Gmelch and Gmelch, 1976). This social isolation, com-
bined with small numbers, suggests that the Traveller
population has experienced considerable genetic drift.
The ethnogenesis of the Travellers has been the focus
of historical debate (Gmelch and Gmelch, 1976). One hy-
pothesis is that the Travellers were a hybrid population
resulting from gene flow from Roma (Gypsy) populations.
This hypothesis likely came about because of the superfi-
cial similarity of the nomadic lifestyles of Travellers and
Roma. Previous studies have suggested that the Roma
hypothesis should be rejected because genetic analysis
showed greater affinity of the Travellers with other Irish
populations than with Roma or Northern Indian popula-
tions (the origin point of Roma populations) (Crawford
and Gmelch, 1974; Crawford, 1975). Such studies show
that cultural identity and genetic ancestry are often
different; in this case, the genetic data point to an Irish
origin despite a cultural similarity with the Roma, such
as a nomadic lifestyle and social isolation. However, that
C 2012
V WILEY PERIODICALS, INC.
and/or the action of genetic drift in a small group that was
descended from a small number of founders. In order to
test the drift hypothesis, we analyzed genetic distances
comparing the Travellers to four geographic regions in Ire-
land. These distances were then compared with adjusted
distances that account for differential genetic drift using a
method developed by Relethford (Hum Biol 68 (1996)
29–44). The unadjusted distances show the genetic distinc-
tiveness of the Travellers. After adjustment for the
expected effects of genetic drift, the Travellers are equidis-
tant from the other Irish samples, showing their Irish
origins and population history. The observed genetic
differences are thus a reflection of genetic drift, and there
is no evidence of any external gene flow. Am J Phys
Anthropol 150:184–189, 2013. V 2012 Wiley Periodicals, Inc.
C
study was limited to four loci and further investigation
is warranted.
The other major hypothesis of Traveller ethnogenesis
is the idea that the Travellers are of Irish origin, a group
that became socially isolated at some point in the past.
There has been debate over whether this initial isolation
dates back into prehistory or is more recent (North
et al., 2000). Even given the Irish origin of the Travel-
lers, there are still some genetic differences between the
Travellers and other Irish samples (Crawford and
Gmelch, 1974; Crawford, 1975; North et al., 2000). Such
differences could indicate some admixture with other
(non-Roma) populations outside of Ireland. The other
possible explanation is genetic drift, which would act
to increase genetic distances from other populations in
Ireland.
The drift hypothesis makes sense given the small size
and cultural isolation of the Travellers as well as demo-
graphic and genetic data. The Travellers have very high
mean fertility with a high variance (Crawford and
Gmelch, 1974), which would act to reduce effective popu-
lation size further, thus increasing the likely impact of
genetic drift. Genetic support for the drift hypothesis
Grant sponsor: Wenner-Gren Foundation.
*Correspondence to: John Relethford, Department of Anthropol-
ogy, State University of New York College at Oneonta, Oneonta, NY
13820, USA. E-mail: john.relethford@oneonta.edu
Received 18 June 2012; accepted 16 October 2012
DOI 10.1002/ajpa.22191
Published online 26 November 2012 in Wiley Online Library
(wileyonlinelibrary.com).
 GENETIC DRIFT AND THE IRISH TRAVELLERS 185
also comes from a study of transferase-deficient galacto-
semia, a metabolic disorder that is much higher in fre-
quency among the Travellers than in the rest of Ireland
(Murphy et al., 1999). DNA analysis has shown that all
of the Traveller cases, and 89 percent of the rest of the
non-Traveller Irish cases, were due to a specific mutation
(Q188R), with an allele frequency of 0.046 in the Travel-
lers and only 0.005 among the non-Travellers. It has
been suggested that these allele frequency differences
resulted from initial founder effect among the Travellers
combined with continued genetic drift over time.
The purpose of the present article is to test the Roma
hypothesis and the genetic drift hypothesis using a
larger comparative data base than earlier studies and by
using a method developed by one of us (Relethford, 1996)
to help untangle the effects of population history (such
as admixture) and genetic drift on the pattern of genetic
distances between populations. Here, allele frequencies
from a number of red blood cell polymorphisms are used
to compute genetic distances between a sample of Irish
Travellers and samples representing the rest of Ireland
(the term ‘‘Ireland’’ is used in this article to refer to the
entire island, currently made up of two political
entities—the Republic of Ireland and Northern Ireland).
Relethford’s (1996) method is used to examine genetic
distances before and after adjustment for variation in
population size, which allows a representation of genetic
distances after controlling for genetic drift. Our hypothe- Fig. 1. Location of the four Irish provinces. Today, the island
sis is that the genetic distinctiveness of the Travellers of Ireland is composed of the Republic of Ireland (Connacht,
will be reduced or eliminated once we have controlled for Leinster, Munster, and part of Ulster) and Northern Ireland
genetic drift. (part of Ulster). Throughout this paper, ‘‘Ireland’’ refers to the
entire island.

MATERIALS AND METHODS

Blood specimens were collected by one of us (MHC) on
119 Irish Travellers during the summer of 1970 (Craw-
ford and Gmelch, 1974; Crawford, 1975). These data
were collected from three government settlements out-
side of Dublin and from Travellers on the road in Dublin,
Wexford, and Wicklow counties. Data on 33 alleles from
12 red blood cell loci are used here, which consist of six
blood groups (ABO, MNS, Rhesus, Kell, Duffy, and P),
three serum proteins (GC, HPA, and TF), and three red
cell enzymes (ACP1, PGM1, and AK1). Although data
were also available for the Lewis blood group, they were
not used here because of lack of comparative data for
other populations in Ireland. The allele frequencies used
here are those reported in Crawford (1975), with a cor-
rection for the frequency of the A1 allele for the ABO
blood group (corrected value 5 0.1647).
Our first analysis addresses the Roma hypothesis by
comparing the allele frequencies of the Irish Travellers
with Eurasian populations that included several national
samples and two Roma samples. Comparative samples
were chosen based on availability of allele frequencies
and suitability for testing the Roma hypothesis. Three
national samples were used: Ireland (the entire island),
England, and Hungary. The two Roma samples are the
Hungarian Roma and Welsh Roma. In addition, a com-
parative sample from north India (Punjab) was used, as
this is the area often cited as a likely point of origin for
some Roma populations (Mastana and Papiha, 1992;
Kalaydjieva et al., 2001). Data for Ireland were provided
by the late Don Tills and listed in his doctoral thesis
(Tills, 1975) [portions of these data have also been pub-
lished by Tills (1977) and Tills et al. (1977)]. Data for
England, Hungary, and Punjab were taken from the
work by Roychoudhury and Nei (1988). Data on Hungar-
ian Roma were taken from the work by Rex-Kiss et al.
(1973), Tauszik et al. (1985), and Benkmann and Goedde
(1991), and data on Welsh Roma were taken from the
work by Harper et al. (1977). Because of missing data,
the comparative Eurasian analysis was based on only 10
of the 12 loci (excluding the P and GC loci). In addition,
the comparative Eurasian analysis used only the M and
N alleles for the MN blood group and not the MS, Ms,
NS, and Ns haplotypes. These exclusions resulted in a
total of 27 alleles for the comparative Eurasian analysis.
The second analysis compared the Irish Travellers to
data from Tills (1975) for the four traditional historical
provinces of Ireland—Connacht, Leinster, Munster, and Ul-
ster. These provinces correspond roughly to ancient king-
doms in Ireland (today, the province of Ulster is partly in
the Republic of Ireland and partly in Northern Ireland, but
is treated as a single unit here). The location of these four
population aggregates are shown in Figure 1. Although
these are not ‘‘local’’ populations, they serve here to provide
comparative information on geographic variation within
Ireland, as the provinces roughly quarter the island into
western, eastern, southern, and northern components.
Sample sizes vary by locus: Connacht 5 311–355, Leinster
5 630–696, Munster 5 372–419, and Ulster 5 464–526.
For both the comparative Eurasian and Irish analyses,
genetic distances between populations were derived
using the R-matrix method of Harpending and Jenkins
(1973). For g populations, the R matrix has g rows and g
columns, with elements
ðpi
p Þðpj
p
Þ
rij ¼ ð1Þ
 ð1
p
p Þ

American Journal of Physical Anthropology

186 J.H. RELETHFORD AND M.H. CRAWFORD
where pi and pj are the allele frequencies for populations
i and j respectively, and p̄ is the mean allele frequency
over all populations, weighted by population size:
X
¼
p wi pi ð2Þ
where summation is over all g groups and where the
population weights (w) are computed as for each popula-
tion as
.X
wi ¼ Ni Ni ð3Þ
where Ni is the size of population i and summation is
over all g groups. The weighting by population size gives
a mean allele frequency that represents the frequency if
all populations in the analysis formed a single panmictic
population. Ideally, effective population sizes would be
used for weighting, but these values are seldom avail-
able for human populations. Instead, it is conventional
to use census population sizes to compute the relative
sizes under the assumption that census sizes are roughly
proportional to effective population sizes. The census
population size for the Travellers throughout all of Ire-
land was 5,880, taken from a survey conducted in 1961
(Commission on Itinerancy, 1963). For the comparative
Eurasian analysis, population size estimates for Ireland
in 1961 were taken from the work by Vaughan and Fitz-
patrick (1978), for Hungarian Roma from the work by
Rex-Kiss et al. (1973), for Welsh Roma from the work by
Niner (2006), and for England, Hungary, and Punjab for
1961 from Wikipedia (articles on ‘‘Demography of Eng-
land,’’ ‘‘Demographics of Hungary,’’ and ‘‘Demographics
of Punjab (India)’’ respectively). For the Irish analysis,
census sizes for the provinces for 1961 were taken from
the work by Vaughan and Fitzpatrick (1978).
The final R matrix was obtained by averaging Eq. (1)
over all alleles. The diagonals of the R matrix (rii) were
corrected for sampling bias by subtracting the quantity
X
1 
k ð4Þ
2nij
from the diagonal element for population i. Here, nij is the
sample size for population i and allele j, k is the total num-
ber of alleles over all loci, and summation is over all alleles
and loci. If the corrected rii value is negative, it is set equal
to zero (Workman et al., 1973; Relethford et al., 1997).
Squared genetic distances between populations i and j
were then derived from the corrected R matrix using the
Harpending and Jenkins (1973) transformation
d2ij ¼ rii þ rjj
2rij ð5Þ
and the genetic
qffiffiffiffiffiffidistances taken as the square root of this
value: dij ¼ d2ij . If any distances were negative after bias
correction, these distances were truncated to zero.
The Roma hypothesis was tested using a modification
of a method developed by Bertorelle et al. (1995), which
tests the hypothesis of common origin by comparison of
genetic distances with control groups that are geographi-
cally proximate with the group of interest. Starting with
the Hungarian Roma sample, we examine the genetic
distance between the Irish Travellers and the Hungarian
American Journal of Physical Anthropology
Roma relative to the distance between the rest of Ireland
(as a control group) and the Hungarian Roma. Following
Bertorelle et al., the null hypothesis is that the Travel-
lers are no more similar to the Hungarian Roma than
Ireland, and the alternative hypothesis is that the Trav-
ellers are more similar to the Hungarian Roma, which
would support the Roma origin hypothesis. In order to
test this hypothesis, Bertorelle et al.’s bootstrap method
was used. Randomized genetic distances (‘‘pseudodis-
tances’’) were generated by sampling with replacement
from the 10 loci used in the comparative Eurasian data-
set, and this sampling was repeated 10,000 times. The
proportion of the 10,000 runs where the pseudodistance
between the Travellers and the Hungarian Roma is
greater than the pseudodistance between Ireland and
the Hungarian Roma provides an estimate of the proba-
bility that the null hypothesis is true. The method was
also used to compare the Travellers and Ireland to both
the Welsh Roma and to the Punjab sample. In addition to
these formal hypothesis tests, multidimensional scaling
was used to examine graphically the genetic relationships
between all samples. Variation within Ireland, comparing
the Travellers to the four provinces of Ireland, were based
on the entire 12 loci and also used Relethford’s (1996)
method to examine the potential role of genetic drift.
RESULTS
Under the model of Roma ancestry, we would expect
the genetic distance between the Travellers and the
Roma to be less than betweenqIreland
ffiffiffiffiffiffi and the Roma.
The observed genetic distance d2ij between the Travel-
lers and the Hungarian Roma (0.253 6 0.057) is slightly
more than the distance between Ireland and the Hun-
garian Roma (0.237 6 0.039) (the standard errors of the
distances were obtained from the bootstrap distribution).
These two distances are not significantly different (Z 5
0.232). The bootstrap test of Bertorelle et al. (1995) gives
a probability of P 5 0.606 supporting the null hypothesis
that there is no Roma ancestry in the Irish Travellers.
For the Welsh Roma comparisons, the genetic distance
between the Travellers and the Welsh Roma is 0.185 (6
0.065) is slightly more than the genetic distance between
Ireland and the Welsh Roma (0.169 6 0.033), but not
significantly different (Z 5 0.219), and the bootstrap dis-
tribution gives a probability of P 5 0.612. Finally, the
genetic distance between the Travellers and the Punjabi
(0.255 6 0.047) is slightly more than the genetic dis-
tance between Ireland and the Punjabi (0.238 6 0.037),
but not significantly different (Z 5 0.284), and the boot-
strap distribution probability is also P 5 0.612. These
results show that the Irish Travellers are not more simi-
lar than the rest of Ireland to Roma populations, and
therefore do not support the Roma origin hypothesis.
These results were confirmed from visual inspection of
multidimensional plots. Figure 2 shows the two-dimen-
sional plot of the genetic distances among the Irish Trav-
ellers and other Eurasian populations using non-metric
multidimensional scaling. Several points regarding the
Roma populations are clear from this plot. First, both
Hungarian Roma and the Welsh Roma are genetically
distinct from corresponding non-Roma populations in
Hungary and the United Kingdom. Second, the Hungar-
ian Roma show the greatest level of genetic affinity with
the Punjabi, consistent with an Indian origin of the
Roma as shown in other analyses of Roma population
 GENETIC DRIFT AND THE IRISH TRAVELLERS
Fig. 2. Non-metric multidimensional scaling plot of genetics
distances between the Irish Travellers and selected Eurasian
populations based on 27 alleles for 10 red blood cell loci. Krus-
kal’s stress 5 0.042, which indicates an excellent fit.
genetics (e.g., Rex-Kiss et al., 1973; Kalaydjieva et al.,
2001). Third, the separation of the two Roma populations
points to the genetic heterogeneity found in other studies
of Roma genetic variation (e.g., Mastana and Papiha,
1992; Kalaydjieva et al., 2001), specifically the separa-
tion between western European Roma (Welsh) and east-
ern European Roma (Hungary). The most relevant find-
ing seen in Figure 2 is that the Irish Traveller popula-
tion is most similar to the rest of Ireland and most
different from the Roma populations, confirming the
rejection of the Roma hypothesis of Traveller origins.
The genetic distances in Figure 2 support an Irish ori-
gin of the Irish Travellers. Of course, these findings do
not mean that there has been no non-Irish gene flow;
comparisons of the Travellers to variation within Ireland
are needed to provide further insight. Figure 3 presents
the two-dimensional plot of the genetic distances based
on 12 loci among the Irish Travellers and the four Irish
provinces using non-metric multidimensional scaling.
This plot shows clearly the genetic divergence of the
Irish Travellers from the rest of Ireland. Figure 3 also
shows some spatial variation within Ireland, where
there is some slight separation of the eastern provinces
(Leinster, Ulster) from the western provinces (Connacht,
Munster). This east–west difference has been seen in
other studies of Irish population history (e.g., Dawson,
1964; Hooton et al., 1955; Tills et al., 1977; Relethford
and Crawford, 1995; Relethford, 2008), and correlates
with the history of population settlement.
What is most striking in Figure 3, however, is the
divergence of the Irish Travellers. Although Figure 2
shows that the Irish Travellers have the greatest affinity
with other Irish populations when examined on a conti-
nental basis, the Travellers are clearly divergent when
considered relative to genetic variation within Ireland.
This divergence of the Travellers could result from
genetic drift and/or divergence due to gene flow outside
of Ireland. As noted by Relethford (1996), genetic drift
can distort inferences made about population history
from genetic distances. For example, two historically
related populations may appear genetically more differ-
ent if one (or both) is small and has experienced consid-
erable genetic drift. In some cases, the confounding
aspects of population history and drift can be untangled
if adequate demographic and historic data are available.
It is also possible to obtain some insight from a set of
genetic distance measures that have been scaled to
187
Fig. 3. Non-metric multidimensional scaling plot of genetic
distances between the Irish Travellers and the four provinces of
Ireland based on 33 alleles for 12 red blood cell loci. Kruskal’s
stress 5 0.017, which indicates an excellent fit.
reflect differential population size (and drift). This type
of scaling was suggested by Relethford and Harpending
(1994) based on intuitive reasoning, and then later dem-
onstrated mathematically by Relethford (1996). This
method uses an extension of Rogers and Harpending’s
(1986) migration matrix method to estimate migration
patterns from the observed R matrix among local popu-
lations and then use these to estimate the R matrix
under a model of equal population sizes (and hence equal
expectations of genetic drift). This method is equivalent
to running a simulation of known migration patterns
under the assumption of equal population sizes.
Although the method of extracting the underlying migra-
tion matrix and then estimating the simulated distances
under the assumption of equal drift involves complex
matrix algebra, the process reduces to a simple algebraic
solution (see details in Relethford, 1996). Here, the ele-
ments of this scaled R matrix (reflecting equal effects of
drift) are computed from the observed R matrix and the
population weights as
pffiffiffiffiffiffipffiffiffiffiffiffi
r0ij ¼ g wi wj rij ð6Þ
A scaled genetic distance is then computed using the
scaled R matrix following Eq. (5) as
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
d0ij ¼ r0ii þ r0jj
2r0ij ð7Þ
Relethford (1996) provides validation of the method
with a simulated example and application to actual
genetic distances between Jewish and non-Jewish popu-
lations.
Figure 4 presents the two-dimensional plot of the
scaled genetic distances among the Irish Travellers and
the four Irish provinces using non-metric multidimen-
sional scaling. Here, the Travellers are no longer diver-
gent, but plot near the center of the four provinces. This
observed pattern is exactly what is expected under the
Irish origin hypothesis, where the Travellers represent a
subset of the overall Irish population, having no particu-
lar affinity to one geographic region or another. Compar-
ison of the scaled distances in Figure 4 with the
unscaled distances in Figure 3 shows that the observed
genetic distinctiveness of the Irish Travellers reflects
genetic drift and not external gene flow.
American Journal of Physical Anthropology
188 J.H. RELETHFORD AND M.H. CRAWFORD
Fig. 4. Non-metric multidimensional scaling plot of scaled
genetic distances [Eqs. (6) and (7)] between the Irish Travellers
and the four provinces of Ireland based on 33 alleles for 12 red
blood cell loci. These scaled distances control for variation in
potential genetic drift. Kruskal’s stress 5 0.005, which indicates
an excellent fit.
DISCUSSION AND CONCLUSION
Populations can be genetically distinct under different
scenarios of population history. For example, populations
can be genetically divergent when they experience high
levels of gene flow from outside the local region, as
shown by Relethford and Blangero (1990) for two west
coast Irish populations that, because of historical factors,
experienced higher levels of English admixture. Popula-
tions can also be genetically distinct due to genetic drift,
particularly if the population under question is small
and has been isolated.
In the case of the Irish Travellers, their cultural dis-
tinctiveness had at one time been considered a reflection
of a different pattern of ethnogenesis, perhaps an origin
or admixture from Roma sources. As shown here (Fig.
2), the Travellers are more closely related to Irish popu-
lations than to Roma populations as suggested by Craw-
ford (1975). However, the Travellers are still somewhat
distinct from other Irish populations, as shown here in
Figure 3, and this distinctiveness could reflect some
other source of gene flow or genetic drift. Although the
demographic and social structure of the Travellers sug-
gests genetic drift, genetic analysis was needed to con-
firm this hypothesis. Here, we used the scaling method
outlined by Relethford (1996) to weight genetic distances
by their census population sizes. This method reduces
the divergence of a population from the centroid propor-
tional to population size. Smaller populations, which are
those that would be affected more by genetic drift, are
brought in closer to the population centroid, while larger
populations are affected less. The results of this scaling
analysis when applied to the Travellers (Fig. 4) controls
for differential drift, and shows the Travellers to lie
squarely in the geographic center of Ireland. This pat-
tern is exactly what we would expect given a definitive
Irish origin of the Travellers, and whose subsequent
genetic history has been shaped by founder effect and
intergenerational genetic drift. Studies of population his-
tory must consider genetic drift as a potential confound-
ing factor.
Although the overall effect of genetic drift on the rela-
tionship of the Travellers to the rest of Ireland has been
demonstrated, the historical nature of that drift remains
unknown. Specifically, it is not clear how much drift is
American Journal of Physical Anthropology
related to an initial founder effect and how much is
related to continued intergenerational drift. As with our
analysis of the relationship of the Travellers to Roma
populations, there is limited insight available from clas-
sical genetic markers. Future research on the population
history of the Irish Travellers would benefit from com-
parison of numerous DNA markers and application of
phylogeographic methods.
ACKNOWLEDGMENTS
Authors thank the many Travellers who participated
in this study. This article is dedicated to a Traveller ‘‘el-
der statesman,’’ the late Bon Connors. Without his help,
this research would not have been possible.
LITERATURE CITED
Benkmann H-G, Goedde HW. 1991. Genetic markers, serum
protein and erythrocyte enzyme polymorphisms. In: Czeizel A,
Benkmann H-G, Goedde HW, editors. Genetics of the Hungar-
ian population. Berlin: Springer-Verlag. p 124–149.
Bertorelle G, Bertranpetit J, Calafell F, Nasidze IS, Barbujani G.
1995. Do Basque- and Caucasian-speaking populations share
non-Indo-European ancestors? Eur J Hum Genet 3:256–263.
Commission on Itinerancy. 1963. Report of the Commission on
Itinerancy. Dublin: The Stationery Office.
Crawford MH. 1975. Genetic affinities and origin of the Irish
Tinkers. In: Watts ES, Johnston FE, Lasker GW, editors. Bio-
social interrelationships in population adaptation. The Hague:
Mouton. p 93–103.
Crawford MH, Gmelch G. 1974. The human biology of the Irish
Tinkers: demography, ethnohistory and genetics. Soc Biol
21:321–331.
Dawson GWP. 1964. The frequencies of the ABO and Rh(D)
blood groups in Ireland from a sample of 1 in 18 of the popu-
lation. Ann Hum Genet 28:49–59.
Gmelch G. 1977. The Irish Tinkers: the urbanization of an itin-
erant people. Menlo Park, CA: Cummings.
Gmelch SB, Gmelch G. 1976. The emergence of an ethnic group:
the Irish Tinkers. Anthropol Quart 49:225–238.
Harper PS, Williams EM, Sunderland E. 1977. Genetic markers
in Welsh gypsies. J Med Genet 14:177–182.
Harpending H, Jenkins T. 1973. Genetic distance among South-
ern African populations. In: Crawford MH, Workman PL, edi-
tors. Methods and theories of anthropological genetics. Albu-
querque: University of New Mexico Press. p 177–200.
Hooton EA, Dupertuis CW, Dawson H. 1955. The physical
anthropology of Ireland. Papers of the Peabody Museum,
Vol.30, Nos. 1–2. Cambridge, MA: Peabody Museum.
Kalaydjieva L, Gresham D, Calafell F. 2001. Genetic studies of
the Roma (Gypsies): a review. BMC Med Genet 2:5.
Mastana SS, Papiha SS. 1992. Origin of the Romany gypsies—
genetic evidence. Z Morph Anthropol 79:43–51.
Murphy M, McHugh B, Tighe O, Mayne P, O’Neill C, Naughten
E, Croke DT. 1999. Genetic basis of transferase-deficient gal-
actosaemia in Ireland and the population history of the Irish
Travellers. Eur J Hum Genet 7:549–554.
Niner P. 2006. Accommodation needs of Gypsy-Travellers in
Wales. Report to the Welsh Assembly Government. United
Kingdom: Center for Urban and Regional Studies, School of
Public Policy, University of Birmingham.
North KE, Martin LJ, Crawford MH. 2000. The origins of the
Irish travellers and the genetic structure of Ireland. Ann
Hum Biol 27:453–465.
Relethford JH. 1996. Genetic drift can obscure population his-
tory: problem and solution. Hum Biol 68:29–44.
Relethford JH. 2008. Geostatistics and spatial analysis in bio-
logical anthropology. Am J Phys Anthropol 136:1–10.
Relethford JH, Blangero J. 1990. Detection of differential gene
flow from patterns of quantitative variation. Hum Biol 62:5–25.
 GENETIC DRIFT AND THE IRISH TRAVELLERS
Relethford JH, Crawford MH. 1995. Anthropometric variation
and the population history of Ireland. Am J Phys Anthropol
96:25–38.
Relethford JH, Crawford MH, Blangero J. 1997. Genetic drift
and gene flow in post-Famine Ireland. Hum Biol 69:443–465.
Relethford JH, Harpending HC. 1994. Craniometric variation,
genetic theory, and modern human origins. Am J Phys
Anthropol 95:249–270.
Rex-Kiss B, Szabó L, Szabó S, Hartmann E. 1973. ABO, MN,
Rh blood groups, Hp types and Hp level, Gm(1) factor investi-
gations on the Gypsy population of Hungary. Hum Biol 45:
41–61.
Rogers AR, Harpending HC. 1986. Migration and genetic drift
in human populations. Evol 40:1312–1327.
Roychoudhury AK, Nei M. 1988. Human polymorphic genes:
world distribution. New York: Oxford University Press.
Tauszik T, Friss A, Gyódi E, Santora Z, Takács S, Kotvász A,
Tóth AM, Horváth M, Tarján L, Petrányi G, Hollán SR,
189
Simonvits I. 1985. Genetic polymorphism of the Gypsy popu-
lation in Hungary as based on studies of red blood cell anti-
gens. Haematologia 18:205–217.
Tills D. 1975. A genetical study of Irish populations. PhD thesis,
Serological and Biochemical Genetics Unit, British Museum
(Natural History).
Tills D. 1977. Red cell and serum proteins and enzymes of the
Irish. Ann Hum Biol 4:35–42.
Tills D, Teesdale P, Mourant AE. 1977. Blood groups of the
Irish. Ann Hum Biol 4:23–34.
Vaughan WE, Fitzpatrick AJ. 1978. Irish historical statistics:
population, 1821–1971. Dublin: Royal Irish Academy.
Workman PL, Harpending H, Lalouel JM, Lynch C, Niswander
JD, Singleton R. 1973. Population studies on Southwestern
Indian tribes. VI. Papago population structure: a comparison
of genetic and migration analyses. In: Morton NE, editor.
Genetic structure of populations. Honolulu: University of
Press of Hawaii. p 166–194.
American Journal of Physical Anthropology