Racial Reality

In August of 2003, RM renamed his site "Racial Reality", mainly in response to "Racial Myths" having been shown for the joke it is by this site. Besides changing the name, RM redesigned his site to try to make it appear more "objective" to unwary visitors (or, as RM himself recently put it, "Racial Reality" targets "uninformed web surfers"). However, most of the material used remains unchanged, and the refutations on this page still stand. Below, I respond to what little new material RM has added.

Basques are the "purest" Europeans?

RM claims Basques are:

confirmed to be the oldest, most genetically pure of all European populations . . . lacking recent non-European admixture

RM cherry-picks a single study which found no "Neolithic" male lineages in a Basque sample, and goes on to conclude that "'mixed-looking' phenotypes . . . are entirely native to Europe". Actually, Basques "are little more of a Mesolithic relict than any other European population" (M. Richards, Annu. Rev. Anthropol. 2003. 32:135–62). On the male side, Basques tend to be low in "Neolithic" ancestry compared to southern Europeans, but not especially low compared to northwestern Europeans.

Most Basques are of course not racially ambiguous in appearance, but as regards those who are it is incorrect to claim they are necessarily of "pure" European ancestry. Berber haplogroup E-M81 has been detected in at least two seperate samples of Spanish (but not French) Basques (Cruciani et al., Am J Hum Genet. 2004 May;74(5):1014-22; Semino et al., Am J Hum Genet. 2004 May;74(5):1023-34), and:

The relatively young TMRCA of 5.6 ky (95% CI 4.6–6.3 ky) that we estimated for haplogroup E-M81 and the lack of differentiation between European and African haplotypes in the network of E-M81 (fig. 2C) support the hypothesis of recent gene flow between northwestern Africa and Iberia. (Cruciani et al. 2004)

Note that in Europe, E-M81 is found only in Iberia and Italy (i.e., Berber ancestry is detected in Portugal, Spain, and Italy, but not in Britain; Cruciani et al. (2004) did detect E-M81 in (non-Basque) France, but their sample consisted of blood donors, and I assume E-M81 in France is associated with recent North African and Portuguese immigrants). So the presence of E-M81 in Spanish Basques makes them quite a lot less "pure" than most other Europeans when it comes to Berber ancestry.

It is true that there have probably been both "light" and "dark" types in Europe since the Paleolithic, as I've pointed out before. It's also conceivable that one may occasionally come across a pondersome looking European who owes his phenotype to chance recombination of native European genes, rather than non-European ancestry. However, RM is wrong to claim Basques are "pure" and therefore "ambiguous" Basques necessarily reflect indigenous phenotypes. It's even more logically unsound to draw an analogy between Basques, who overwhelmingly carry Paleolithic European Y chromosomes and who only rarely look "ambiguous", and southern Italians, most of whom carry non-European Y chromosomes and many of whom show a non-European cast.

"Celts" are closely related to Basques?

Claims such as the following have appeared in the press over the past few years:

The Welsh and Irish Celts have been found to be the genetic blood-brothers of Basques, scientists have revealed. (BBC News)

The connection is this: Western British and Irish populations share with the Basque very high frequencies of R1b Y chromosomes. Below, I'll show that Basques and "Celts" are distinct in their frequencies of autosomal and mtDNA markers, but first some background on R1b.

Y haplogroup R1b

R1b is associated with some of the earliest inhabitants of Europe who have left male line descendants down to the present. The Y chromosomes ancestral to R1b probably entered Europe over 30,000 years ago, carried by "Cro-Magnons". During the Last Glacial Maximum, climate change forced Europeans into a handful of refugia in southern Europe (R1b is associated with a refuge in southwestern Europe; I with a refuge in central Europe or the Balkans; and R1a with a refuge in the Ukraine). It is of course incorrect to assume the ice age inhabitants of these refugia were related in appearance to modern southern Europeans. According to Stephen Oppenheimer:

The archaeology shows us how the south-western refuge of the Basque country drew cultures and presumably people down from north-west Europe during the lead up to the LGM. (Out of Eden, p. 251)

Oppenheimer points out that both Y and mtDNA evidence support the view from archaeology. Following the LGM, the descendants of northern European hunters expanded back into northern Europe, leaving a strong imprint on the Basque male gene pool. Today, most of the inhabitants of northern Europe have a mix of Y chromosomes from the major refugia. But due to geographic isolation, the inhabitants of western Britain and Ireland have maintained very high levels of R1b. This makes them extremely distant cousins to the Basque, not "blood-brothers". We can also note that the Y chromosome pools of Basques and "Celts" are strongly differentiated in some ways. For example, as noted above, Spanish Basques have experienced gene flow from Berbers, while "Celts" have not. Moreover, mutations like R1b8, which mark more recent expansions from Iberia, are found in western Britain only sporadically and not at elevated frequencies compared to eastern Britain.

Autosomal markers

The lists below show genetic distances between the Irish and the Basque and selected poulations based on Cavalli-Sforza's analysis of data on 'classical' markers (from HGHG, Table 5.5.1.)

Distance from Irish

Irish 0
Scottish 29
English 30
Danish 68
Belgian 75
Dutch 76
Norwegian 79
Spanish 113
Basque 145

Distance from Basque

Basque 0
French 93
Spanish 104
Belgian 107
Dutch 118
English 119
Russian 140
Italian 141
Irish 145


The Irish are more closely related to all north-west and north-central Europeans than they are to the Basque. Basques are more closely related the English, the Dutch, and even Russians than they are to the Irish, according to autosomal data. We can also note that in the fifth PC of classical variation in Europe -- which C-S claims "corresponds to the progressive retreat of the boundary of the Basque language" -- Ireland (along with Britain, Scandinavia, and a large swathe of central Europe) is at the opposite pole from Basque country. Iberia, France, Italy, even Russia and the Caucasus are closer to Basque country on this PC than Ireland is.

mtDNA

The plot below shows that when considering mtDNA the Irish and Welsh are distinct from the Basque and these "Celts" cluster with other northern Europeans (Source: PNAS 98(9):5078-5083).

Seafaring Armenoids in Northern Europe?

Years after I originally took apart his ridiculous claims about "hamito-semitic" britons, RM continues to propagate misinformation on the subject. Despite RM's assertion that "I no longer maintain that there's a Hamito-Semitic influence in Britain", his "hamito-semitic britons" page remains online and he has actually added additional "evidence": an excerpt from a 1957 paper by Bertil Lundman on "The Problem of Ancient Oriental Shipping on the North Sea" (which, tellingly, RM has mis-labeled on his "References" page with the title of the "blog" entry by Dienekes which was RM's immediate source for this "evidence"). The facts on my page about Britain stand, and no further comment should be necessary. But since Pontikos and RM insist on circulating excerpts from this odd Lundman paper, we might as well dispose of its claims (thanks to Karl Earlson for his comments and for providing the image used below).

* Lundman tries to associate Deniker's litoral type with Armenoid invaders, but there's not much room for comparison: the Litoroid is tall, slender, long-headed, and small-nosed; the Armenoid is short, planoccipital, and large-nosed with a tendency to obesity.

* The supposed descendants of Armenoids Lundman "discovered" in Scandinavia don't sound particularly Armenoid. They are described as "tall" and "slim", not short and fat. They are said to have "large, slightly and evenly curved but scarcely fleshy noses". Needless to say, many northern Europeans have convex noses and this has nothing to do with Armenoids. Likewise, some fraction of northern Europeans are dark and broadheaded; this is part of their natural variation, and one need not seek the origin of these traits among sea-borne Armenoids. The fact that Lundman identifies supposed Near Eastern types in Ireland completely discredits his theory: modern genetics shows Near Eastern Y chromosomes are practically absent from Ireland (and any stray "Neolithic" Y chromosomes which found their way to Ireland most likely diffused northwest across Europe via land over a period of thousands of years).

* Lundman asks if certain "dark" navigators of English history "were the descendants of the ancient Tyrrhenian seafarers from Tyre?" More on these "dark" navigators below, but the fact that Lundman brings up the Phoenicians in the context of Britain is enough to discredit his theory. There is no archaeological evidence for Phoenicians ever having set foot in northern Europe. Furthermore, the exploitation of metal deposits in Britain began thousands of years before the Phoenicians existed, and where outside involvement in British mining is indicated the evidence points to west-central Europe, not the Near East.

* Lundman admits that "most of England's great navigators were conspicuously blond", but he then attempts to single out three, Drake, Dampier, and Cook, who he believes were of his pseudo-Armenoid type. Of course, like members of all northern European populations, Englishmen vary in pigmentation; so it would be surprising if Lundman could name no examples of dark English historical figures. But it's interesting that at least one of Lundman's three daring navigators of supposed Near Eastern ancestry is not particularly dark. As shown in the portrait below, Drake is ruddy, grey-blue-eyed, and brown-haired with blond beard.

mtDNA haplogroup M1

RM claims M lineages in the Mediterranean are "proto-Asian". Actually, coalescence analysis shows the age of haplogroup M in eastern Africa is similar to that in India, and researchers have been unable to conclusively determine where M originated. Regardless, the M lineages which occur in Italy and Greece mainly belong to M1, which most likely traces an expansion from East Africa within the past 10,000-20,000 years:

The presence (~20%) of the 'Asian' mtDNA haplogroup M, defined by the 10400 CT transition, is unique in eastern Africa7. As in Asia, the haplogroup M in Ethiopia belongs to a superhaplogroup (L3), which encompasses almost all Eurasian and a considerable number of African mtDNAs (Refs 8,9).

[. . .]

The two groups are considerably different (Table 1). The eastern-African group, named M1, is characterized within M by a consensus HVS-I motif defined by four transitions at nt 16,129, 16,189, 16,249 and 16,311. This HVS-I signature motif is not found either in our or other10 Indian samples, whereas it characterizes most of the M types sporadically observed in the Mediterranean area11, 12 and 7% of Nile Valley sequences13. . . . The African and Indian M phylogenies are highly divergent.

[. . .]

We calculated the coalescence time for haplogroup M mtDNAs in eastern Africa and India using the statistic (ref. 14) by RFLP and HVS-I. The age of eastern-African haplogroup M, calculated using RFLP data (48,000+-15,000), is compatible with that of Indian haplogroup M (56,000+-7,000; Table 2), whereas a lower age is obtained using HVS-I data. In fact, the root of the eastern-African M1 clade is separated from the root of M by four transitions (Fig. 3), which, even if hypervariable, needed some time to be accumulated (a few thousand or a few ten-thousand years). A somewhat more recent coalescence of haplogroup M in eastern Africa (that is, in the area of the Oromos) than in Asia can be explained by coalescent theory: a small localized population cannot maintain the same level of diversity as a population that expanded and spread over an entire continent. Thus, it is plausible that haplogroup M already existed in eastern Africa approximately 60,000 years ago. The four subgroups (I–IV) of M1 (Fig. 1) may have similar ages of approximately 10,000–20,000 years. These values may correspond to local expansions of these subclades followed by dispersals, leading to the sporadic presence of M along the Mediterranean.

[. . .]

Thus, in the context of the out-of-Africa model1, the following scenario can be envisaged: haplogroup M originated in eastern Africa approximately 60,000 years ago and was carried toward Asia. This agrees with the proposed date of an out-of-Africa expansion approximately 65,000 years ago10. After its arrival in Asia, the haplogroup M founder group went through a demographic and geographic expansion. The remaining M haplogroup in eastern Africa did not spread, but remained localized up to approximately 10,000–20,000 years ago, after which it started to expand.

Source: Quintana-Murci et al. Genetic evidence of an early exit of Homo sapiens sapiens from Africa through eastern Africa. Nature Genetics 23(4):437-441 1999.

Dark-haired Germanic nobility?

RM picked up this theme from Greek CompSci student "Dienekes Pontikos", who some time ago decided (apparently based on a fictional movie he saw) that northern European aristocrats have dark hair and therefore he has something in common with them. Pontikos clings to this idea and uncritically appeals to the authority of H.S. Chamberlain to support his view. Here, Karl Earlson responds:

RM states: "even the old Germanic nobility was predominantly dark-haired." Here, he is ultimately citing Chamberlain, via Hankins, via Pontikos:

http://www.dienekes.com/blog/archives/000157.html

The text in question, concerning "Fair Hair", can be found in Chamberlain's "Foundations of the Nineteenth Century", Vol. I, pp. 522-526:

http://www.hschamberlain.net/grundlagen/division2_chapter6.html

But there is nothing to any of this. I'll pick it apart just to show how empty these statements are, and reveal how nothing is proven.

[1] Chamberlain states: "Eckermann was struck by the brown eyes of Wellington."

But Wellington WASN'T brown-eyed! Please read this quote, describing Wellington's appearance in 1814:

"He was just forty-five, in the pride of manhood: lean, springy, his hair still brown, his eyes whether frozen over or sparkling as blue as ever, [...] and his profile impressively Roman."

Lady Elizabeth Longford, Wellington: The Years of the Sword (London: Weidenfeld & Nicolson, 1972), p. 415.

Lady Longford was not only a descendant of Wellington's, but her biography of the Duke is considered to be a standard work on the subject. Incidentally, German anthropologists (e.g. Guenther and Fischer) routinely classified Wellington as a Nordic type.

[2] Chamberlain states: "poets from the extreme north of Germany pretty frequently speak of dark hair as a characteristic feature not only of the nobility but also of the people".

But the examples he provides, Theodor Storm's "Hans und Heinz Kirch", as well as the works of Goethe, are VERY late in composition (18th & 19th Century works). Similarly, the Storm examples he gives decribe "defiant Germanic seamen" and "another daring figure, Hasselfritz"; in other words, COMMON people of recent times, not the "old Germanic nobility". Chamberlain goes on to state that "those genuine Teutons therefore resemble Achilles with his 'brown hair'." But Achilles was depicted by the ancient Hellenic authors as yellow-, red-, and even golden-haired! See Sieglin (1935) for the details about this. [Sieglin, W. (1935) Die blonden Haare der indogermanischen Völker des Altertums (Munich: J. F. Lehmanns Verlag).] Equally, Chamberlain relies on anonymous authority when he says: "How often, too, in the folksongs do 'dark brown eyes' occur!" No examples are cited, so the statement lacks any real value. In direct contrast to this, Biehahn (1964) has collected a range of descriptions from various German poets, which talk approvingly of blond hair and blue eyes, praising such features as beautiful, truly Germanic, heroic characteristics, etc. Indeed, Biehahn even refers to a "cult of blondness" in German literature.

Biehahn, E. (1964) "Blondheit und Blondheitskult in der deutschen Literatur." Archiv für Kulturgeschichte, 46, 309-333.

[3] Chamberlain states that the predominant type "in German Tyrol, whose inhabitants Henke says 'represent the true type of the primeval Teuton'", for the most part, have "dark and often black hair".

However, this is wrong on several points. The ancient Germanics were fair-haired and long-headed. On the contrary, the modern Tyroleans are largely dark-haired and broad-headed. This question is discussed and proved at length in Schwidetzky (1979). Even Chamberlain acknowledges this in his next section, titled "The Shape of the Skull". He mentions that "Germanic peoples, wherever they have not crossed with others or only to a small extent, as in the north, are long-skulled and fair" and in contrast to this, the "Un-Germanic, short-skulled, dark type" reveals that "in these phenomena we see the effects of the infiltration of an Un-Germanic race, a race which does not belong at all to the Indo-European circle, but to the raceless chaos". I wouldn't go as far as Chamberlain in his polemic, but the point is made nonetheless.

Schwidetzky, I. (1979) "Rassengeschichte von Deutschland." I. Schwidetzky [ed.], Rassengeschichte der Menschheit, 7. Lieferung. Europa V: Schweiz, Deutschland, Belgien und Luxemburg, Niederlande (Munich: R. Oldenbourg Verlag), 45-101.

[4] Chamberlain mentions several times the so-called "Moltke type (or, as the English say, the Wellington type)", to which he attributes dark hair. Wellington has already been dealt with above. As for Moltke, he was fair-complexioned, and various German physical anthropologists (Guenther, von Eickstedt, etc.), classified him as Nordic. See the picture of Moltke here (Fig. 300):

http://www.white-history.com/earlson/hfk/reoehchap10.htm

[5] Chamberlain refers to: "the prevalence of dark colour among the members of the most genuine old Germanic nobility." Once again, this is simply an opinion, and since no evidence is adduced in its favour, none is needed to refute it. Chamberlain also claims that: "In England this is quite striking." However, Chamberlain was not a professional physical anthropologist, and in the opinion of a genuine anthropologist, namely Coon:

"In general, differences in social level and in occupation reflect racial differences, which show themselves to a certain extent in pigmentation. The upper social strata, being on the whole blonder, follow the pigment pattern of the Saxons, Danes, and Normans. This differentiation may well have been even stronger in the Middle Ages, when social lines were more strictly and overtly drawn than today. The Englishman who travels abroad and is seen by foreigners, and the one whose photograph frequently appears in the London Illustrated News, is more likely to be blond than the general run of his more obscure compatriots who stay at home, and whose faces are publicly depicted only when they have committed crimes."

http://www.fikas.no/~sprocket/snpa/chapter-X3.htm

Wurm (1993) has collated a list of the earliest descriptions of the highest Germanic nobility, and from these he has compiled a composite portrait of the typical German nobleman. The nobles are described as being tall and athletic in form, with handsome facial features, bright and flashing eyes, a fair or florid complexion, and fair hair and beard. Wurm mentions specific references to white-blond, golden-blond and reddish-blond hair colours, and only one (possible) brown-haired individual. Wurm concludes by noting that the Germanic nobility belonged to the tall, muscular, long-headed, sharply-profiled, Nordic-Germanic racial type of the Reihengräber.

[Wurm, H. (1993) "Menschentyp und Macht im Früh- und beginnenden Hochmittelalter (750 bis 1000 n. Chr.)." Würzburger medizinhistorische Mitteilungen, 11, 235-260.]

One must always bear in mind the fact that Chamberlain was in many senses a polemicist: all Germans are equally Germanic, both blond and brunet, Celts and Slavs are Germanic, Jesus was an Aryan, so Christianity is also Germanic, etc. His work undoubtably contains much of value, but his opinions must always be carefully sifted, and not accepted at face value. Finally, note this quote by Chamberlain, on the early Germanen:

"the great radiant heavenly eyes, the golden hair, the gigantic stature, the symmetrical muscular development, the lengthened skull (which an ever-active brain, tortured by longing, had changed from the round lines of animal contentedness and extended towards the front), the lofty countenance, required by an elevated spiritual life as the seat of its expression". [p. 535]

http://www.hschamberlain.net/grundlagen/division2_chapter6.html

"Cambro-Britannic Hebraisms"
The most amusing new piece of evidence RM presents is a list of "Cambro-Britannic Hebraisms" apparently published in 1675 by a Welsh scholar named Charles Edwards, which RM copied from a website (http://www.geocities.com/Athens/8431/welsh.html) promoting the work of Yair Davidy, an Israeli proponent of the theory that Lost Hebrew Tribes are to be found in Britain. Of course, this sort of "evidence" doesn't merit a serious response. This absurd line of argument just further proves RM's complete lack of objectivity (and/or lack of analytic ability).

Italian genetics
Do studies on the Y-chromosome suggest "gene flow from North Africa in historical times has been minimal"?

RM writes:

A subclade of North African Y-chromosome haplogroup 21 (labeled 25.2), which distinguishes Berbers, was detected in Lombardy at a frequency of 5.6%, Sardinia at 2.3% and Sicily at 2.1%, suggesting that gene flow from North Africa in historical times has been minimal. (Scozzari et al., Hum Immun, 2001)

The study RM cites certainly proves no such thing. First of all, "HG25.2, was identified by using a marker from the pseudoautosomal region", NOT by using a marker from the non-recombining portion of the Y-chromosome. The authors suggest the crossover event that defines HG25.2 happened only once, but it is impossible to prove this, and if they are wrong, the "subclades" they have identified are effectively useless.

Moreover, even if it is informative to divide HG21 using a pseudoautosomal marker:

1. This study only deals with Morrocans. There is no evidence here that "HG25.2" (which is of course not an actual NRY haplogroup) is especially prevalent in other North Africans.

2. Even within Morocco, "Arabs" have much lower levels of "HG25.2" than Berbers. We would expect the Saracens who invaded southern Italy to be mainly "Arabs", not Berbers. And, indeed, southern Italians share with Moroccan Arabs relatively high levels of "HG25.1".

Haplogroup 25.1 was found at its highest frequency (41%) in the Arabs from Morocco, but it was also very common in the Sicilians, the Sardinians from Bitti, and in some groups from continental Italy. The Berbers from Morocco could be clearly differentiated from the Arabs of the same region by the frequency of HG25.2 (71% and 29%, respectively), a haplogroup distinguished from its ancestor (HG25.1) by the G allele at XY275Y. Interestingly, HG25.2 in Europe was found only in western populations, with a particularly high incidence (42%) in the Pasiego from the Pas valleys (Figure 4B).

Do autosomal studies show there has been little gene flow between North Africa and Southern Europe?
RM states:

An autosomal analysis of 10 ancestry-informative markers in Southern Europeans (including Italians, Sicilians and Sardinians) and various Middle Eastern/North African populations revealed a "line of sharp genetic change [that] runs from Gibraltar to Lebanon," which has divided the Mediterranean into distinct northern and southern clusters since at least the Neolithic period. The authors conclude that "gene flow [across the sea] was more the exception than the rule," attributing this result to "a joint product of initial geographic isolation and successive cultural divergence, leading to the origin of cultural barriers to population admixture." (Simoni et al., Hum Biol, 1999)

Amusingly, in an attempt to give himself credibility RM has picked up a "scientific"-sounding catch-phrase from the AncestryByDNA people and misapplied it. "Ancestry Informative Markers", in ABD-terminology, are SNPs specially chosen because they appear at markedly different frequencies in different racial groups and therefore are supposedly especially useful in inferring race. In fact, the Simoni et al. study is based on "classical" genetic markers -- blood groups and protein markers; not SNPs, much less SNPs specially selected to be "ancestry informative".

For each of the 39 regions 1 or more allele frequencies for 6 blood group systems (ABO, Duffy, Kell, MN, Rh, P; 8 independent alleles overall) and 2 serum protein systems (GC, HP) were available, for 10 alleles overall. In practice, 1 allele was discarded at each locus to approximate statistical independence of the data. When more than 1 sample had been typed in 1 region, median allele frequencies were calculated. In this way we obtained an almost complete data matrix of 10 allele frequencies in 39 regions (Table 1).

This is not the best sort of data to be working with, though I'll give the authors the benefit of the doubt and assume that their findings are meaningful, if inexact. But, the authors themselves admit:

These data are far from uniformly distributed. Several European populations have been typed at many loci, and the Asian and African populations have been typed much less so. Also, the data matrix is largely incomplete, because different loci have been described in different populations. In the first step of data analysis, the spatial autocorrelation analysis, this did not present a problem. However, in the second phase, the identification of genetic boundaries, several samples had to be collapsed on the basis of their geographic or linguistic affinities to avoid statistical problems arising from missing data. Moreover, some loci were discarded, decreasing the number of allele frequencies to 886.

The authors also mention:

Note, however, that according to the original sources, only 2 of the North African samples considered in this study included individuals defined as Arabs.

This point may be important, since one expects that gene flow during periods of Moorish occupation in southern Europe, for example, would have come more from North African "Arabs" than from the more isolated "Berbers". And, as we saw above, Y-chromsomal evidence supports this idea.

As for what the authors do claim, it is mostly in keeping with what we already know. Unfortunately, the authors make no attempt to quantify or date admixture (though, considering the quality of their data, it is probably wise that they refrain), so the paper adds little new information.

The most evident feature of the gene frequency patterns described in this study is the presence across the Mediterranean of a longitudinal line of sharp genetic change. This line, supported by analyses based on 2 different measures of genetic distance, runs from Gibraltar to Lebanon; it includes all 5 sharpest genetic boundaries identified on the basis of Prevosti's distances (Figure 2) and 4 of the 5 sharpest boundaries identified using Nei' s distances (Figure 3). Other genetic boundaries, surrounding groups such as Basques, Jews, Sardinians, and Corsicans, are weaker. This strong north-south differentiation clearly suggests that the sea has contributed to isolating the populations of the northern and southern coasts. Localities along the same coast are separated by barriers of lesser importance or by no barrier at all.

This is common sense. It would be surprising if they found otherwise. Of course there will be more genetic exchange between neighboring areas of Europe than between Europe and Africa.

Had the sea distance been the main isolating factor, we would have expected greater genetic differences in the eastern part of the Mediterranean. Exactly the opposite appears to be the case. The genetic boundary separating northern Morocco and Algeria from southern Spain appears as the zones of sharpest genetic change using Prevosti's distances and as the second most significant zone of change (after the zone separating Israeli Jews from all their neighbors) using Nei's distances. The ranking of the genetic boundaries recognized in this study tends to decrease as one moves east, despite sea distances being greater there. Thus it is the presence of the sea rather than the sea distance between localities that seems to have exerted the main isolating effect.

This finding is in keeping with Y-chromsomal variation. SW Europe generally has less Middle Eastern and North African ancestry than SE Europe.

The authors do claim "gene flow [across the Mediterranean] was more the exception than the rule for a substantial period of time." (RM neglected to include the end of that sentence.) Of course, the authors are in no position to quantify how long this "substantial period of time" has been. 1000 years? 2000? 10000? When discussing the clines in their results, the authors mention potential links to the Neolithic expansion, the Arab expansion in north Africa, and possible Paleolithic origins. But:

Because dating allele frequency patterns requires a number of untestable assumptions, we prefer not to draw any strong conclusions at this stage.

Nor can they quantify levels of gene flow in absolute terms. Their data shows only that lately there has been relatively more gene flow between neighboring populations in southern Europe than between southern Europe and Africa. The authors make no attempt to quantify admixture, nor is it likely their data is suitable for that task. Here is the quote in context:

What kind of evolutionary process or combination of processes can most likely account for the pattern of genetic diversity described in this study? The distribution of the most significant genetic boundaries indicates that the demographic history of the populations of the 2 main coasts were largely independent. Although population movements across the sea are historically well documented and have certainly played an evolutionary role by mixing alleles of different origins (Sokal 1991), the existence of sharp genetic differences between the northern and the southern coasts means that gene flow was more the exception than the rule for a substantial period of time. Note that only in the eastern Mediterranean are many islands present; this may have favored population contacts by reducing the sea distances to be covered.