predominance of female converts might provide the answer. It might also explain opposed findings on Jews from Yemen, who in earlier tests matched poorly with other Jews.

Doron M. Behar, Michael F. Hammer, Daniel Garrigan, Richard Villems, Batsheva Bonne-Tamir, Martin Richards, David Gurwitz, Dror Rosengarten, Matthew Kaplan, Sergio Della Pergola, Lluis Quintana-Murci, and Karl Skorecki. "MtDNA evidence for a genetic bottleneck in the early history of the Ashkenazi Jewish population." European Journal of Human Genetics 12:5 (May 2004): 355-364. (Advance online publication on January 14, 2004.) An observer who read the study indicates that the study shows that approximately 60 percent of European Jewish maternal roots come from European sources, with the other 40 percent from Middle Eastern or Asian roots. Abstract excerpt:

To test for the effects of a maternal bottleneck on the Ashkenazi Jewish population, we performed an extensive analysis of mitochondrial DNA (mtDNA) hypervariable segment 1 (HVS-1) sequence and restriction site polymorphisms in 565 Ashkenazi Jews from different parts of Europe. These patterns of variation were compared with those of five Near Eastern (n=327) and 10 host European (n=849) non-Jewish populations. Only four mtDNA haplogroups (Hgs) (defined on the basis of diagnostic coding region RFLPs and HVS-1 sequence variants) account for approximately 70% of Ashkenazi mtDNA variation. While several Ashkenazi Jewish mtDNA Hgs appear to derive from the Near East, there is also evidence for a low level of introgression from host European non-Jewish populations.

With the exception of Ethiopian Jews, all Jewish samples show a high genetic correlation... 3. In descending order after these Middle Easterners, Ashkenazi Jews correlate best with Greeks and Turks; then with Italians; then with Spaniards; then with Germans; then with Austrians; and least of all with Russians...

L. L. Field, J. A. Lowden, and A. K. Ray. "Immunoglobulin (Gm) allotypes in a sample of Canadian Ashkenazic Jews." American Journal of Physical Anthropology 48(2) (February 1978): 159-164. This is an old study which may not be using current techniques. Abstract:

"Gm typing on the serum specimens of 507 Ashkenazic Jews (pre-dominantly of Polish-Russian ancestry) from Toronto, Canada has established the presence of haplotypes Gm3;5, Gm1;21, Gm1,2;21, and Gm1,17;5, and the absence of haplotypes Gm1;13,15,16, Gm1;5,6, and Gm1;5,6,24 which have been found in other Jewish peoples. It is suggested that Ashkenazic populations have lower frequencies of haplotype Gm1,17;5 than non-European Jewish populations, and that some eastern European Jewish populations have acquired the Gm1;13,15,16 haplotype through gene flow from Central Asia. Thus Jewish populations show differences in the Gm system; many of the differences may be in the direction of similarities to neighbouring non-Jewish populations."