Immunoglobulin Haplotypes: Markers of 
Reproductive Success? 
J.C. STEVENSON, 1 M.S. S C H A N F I E L D , 2 M.H. C R A W F O R D , 3 A N D P.M. E V E R S O N 4 

Abstract Immunoglobulin haplotypes are highly polymorphic and 
are useful for analyses of both macro- and microdifferentiation 
of populations. The origins of this diversity are not known, but 
recent reports suggest strong selection at this locus. Increased rates 
of first-trimester spontaneous abortions have been reported when 
parents share GM phenotypes. Reduced fertility has been observed 
in mixed European descent white and Hutterite populations when 
both parents share immunoglobulin haplotypes. Population samples 
with completed family information and GM haplotype data are rare; 
the objective here is to provide this information on another sample. 
A sample of 242 Mennonite couples with mothers older than 40 years 
was divided into 3 groups of matings based on how many haplotypes 
were shared: 0, 1, or 2. The distribution of mean completed family 
sizes for the three groups were 3.35 ± 1.85 (n = 23), 3.47 ± 1.69 (n = 
128), and 3.37 ± 1.60 (n = 91), respectively; these values were not 
significantly different (F = 0.145, p = 0.865). The log-rank test was 
used to compare the time-to-next-birth curves. The intervals between 
first and later births (2-4 births) were not significantly different for 
the three subgroups either. There is also only limited evidence for 
segregation distortion in another sample of 923 offspring (in which 
at least one parent is heterozygous). 

Inherited differences in human immunoglobulins known as allotypes or 
allotypic markers are found on IgG H chains (GM markers), IgA H chains 
(AM markers), and k-type L chains (KM markers, formerly called Inv) 
(Schanfield 1978). Each allotypic determinant is normally restricted to 

1 Anthropology Department, Western Washington University, Bellingham, WA 98225. 
2Analytical Genetic Testing Center, 7808 Cherry Creek Drive South, #201, Denver, CO 80231. 
3 Laboratory of Biological Anthropology, University of Kansas, Lawrence, KS 66045. 
4 4-Firs Research, 1305 E. Victor, Bellingham, WA 98225. 
Human Biology, August 1990, Vol. 62, No. 4, pp. 479-489. 
© Wayne State University Press, 1990 
KEY WORDS: GM, MENNONITES, SEGREGATION ANALYSIS 



4 8 0 / STEVENSON ET AL. 

one of the IgG or IgA subclasses and thus far are found on only the 
constant regions of the IgG and IgA H chains. Different combinations 
of allotypic markers are inherited as closely linked units known as 
haplotypes. Populations are readily distinguished by this system because 
there are many haplotypes that can serve as unique population markers 
(Johnson et al. 1977; Ropartz et al. 1963, 1964; Schanfield 1978, 1980; 
Steinberg 1969; Steinberg and Cook 1981). Widely separated populations 
are readily differentiated by this system, but one can also distinguish 
populations in closer proximity because of conspicuous differences in 
haplotype frequencies. 

The origins of this diversity are not yet understood, but recent re-
ports suggest strong selection at the GM loci (Ober et al. 1986; Weitkamp 
1986; Weitkamp et al. 1986). Reduced fertility has been reported in mixed 
white and Hutterite populations when both parents share immunoglobu-
lin haplotypes (Ober et al. 1986; Weitkamp 1986; Weitkamp et al. 1986). 
There are also increased rates of first-trimester spontaneous abortions 
when parents share GM phenotypes (Weitkamp et al. 1986). Population 
samples with both completed family information and GM haplotype data 
are rare. The objective here is to corroborate Weitkamp and colleagues' 
findings with data on another population. 

The association of shared HLA antigens and reduced reproductive 
performance has already been documented in Hutterite couples (Ober 
et al. 1983, 1985; Ober, Elias, O'Brien et al. 1988; Ober, Elias, Hauck 
et al. 1988) and in couples having habitual abortions [reviewed by Gill 
(1983b)]. In Hutterite couples, where large families are the rule, intervals 
between births are larger when couples share one or more HLA antigens, 
and this effect is more apparent with increasing parity (Ober et al. 1983, 
1985; Ober, Elias, O'Brien et al. 1988; Ober, Elias, Hauck et al. 1988). 

Reduced fertility resulting from shared antigens has also been re-
ported for the GM system. In preliminary studies a significant reduction 
in family size was observed by Weitkamp et al. (1986) in a sample of 437 
white couples with 2 or more children (selected originally over 15 years 
for family studies of various diseases and for linkage analysis) when the 
couples shared the same GM phenotype (3.10 ± 0.16 for like couples 
versus 3.55 ± 0 . 1 3 for unlike couples). The effect was also observed in 
the 282 couples with mothers older than 40 years (3.59 ± 0.22 for like 
couples versus 4.90 ± 0.17 for unlike couples). This difference was also 
apparent in 68 Hutterite couples (4.42 ± 0.68 versus 6.71 ± 0.57) and 
was even more dramatic among the 33 Hutterite couples with mothers 
older than 40 years (6.00 ± 1 . 2 1 versus 9.16 ± 0.17). 

Of the 304 mixed white couples who had been both HLA and GM 
typed, virtually all the differences in family size "as a function of sharing 



Immunoglobulin Haplotypes / 481 

GM phenotypes was limited to the couples who also shared one or more 
HLA-A or HLA-B antigens" (Weitkamp et al. 1986, p. 33). Weitkamp et 
al. then compared couples who had first-trimester spontaneous abortions 
(and at least three pregnancies) with control couples who had three or 
more offspring, according to whether or not they shared HLA-A or HLA-
B antigens. Spontaneous abortions increased significantly in couples who 
share GM haplotypes when compared to control couples ( x 2 = 7.2 with 
Yates correction, p < 0.01). Weitkamp et al. believe that this represents 
an interaction "between HLA and GM in susceptibility to spontaneous 
abortion and in fecundity" (1986, p. 34). 

Thus there is some evidence for an association between reduced 
completed family size and sharing of immunoglobulin markers in popu-
lation samples representing larger families. In this study we present data 
from a Mennonite sample not chosen for family size. A segregation anal-
ysis was performed to test whether a fetus identical with the mother for 
both GM haplotypes is at a selective disadvantage compared to a fetus 
that shares only one haplotype with the mother. 

Methods and Materials 
The populations studied consisted of one sample of Mennonite cou-

ples with mothers older than 40 years (women born before 1941) and a 
sample of mixed Caucasian matings and offspring. The 242 Mennon-
ite couples represent the communities of Goessel and Meridian, Kansas, 
and Henderson, Nebraska, and were recruited during an interdisciplinary 
aging study (Crawford and Rogers 1982). There are no significant differ-
ences among the communities for completed family size (Stevenson et 
al. 1989). The second sample consisted of 923 matings (in which at least 
one parent is heterozygous) ascertained from 268 paternity studies and 
221 family studies of various diseases (not selected for family size). The 
offspring of the Mennonite couples had not been allotyped and were not 
included in the segregation analysis. All samples were tested minimally 
for G1M (A, F, and X) and G3M (BO, Bl, B3, B5, G) using previously 
described methods (Schanfield et al. 1975). The nomenclature used for 
immunoglobulin allotypes is consistent with the recommendations of 
Shows e t a l . (1987). 

The Mennonite sample was divided into three groups of matings 
based on how many haplotypes were shared. There are three GM hap-
lotypes common to European Caucasians: IGHGTA G3*G, IGHG1*A,X 
G3*G, and IGHG1*F G3*B (Steinberg and Cook 1981). Each person 
carries two haplotypes. Thus "zero haplotypes shared" means that the 



4 8 2 / STEVENSON ET AL. 

parents do not share any one of the three haplotypes, and "one (two) 
haplotypes shared" means that they share one (two) of the three haplo-
types. 

Weitkamp and colleagues (Ober et al. 1986; Weitkamp 1986; Weit-
kamp et al. 1986) recognized two subdivisions for G M matings: "like" 
versus "unlike." Their category of like matings corresponds to two 
haplotypes shared, because all other matings are phenotypically unlike. 
The distinctions used here are more consistent with the recent HLA 
studies in which subdivisions of matings are based on 0, 1, or 2+ antigens 
shared (Ober et al. 1985; Ober, Elias, O'Brien et al. 1988; Ober, Elias, 
Hauck etal. 1988). 

Distributions of completed family sizes were checked for skewness 
and kurtosis, and an analysis of variance was performed. 

Life table analyses provide another means of comparing the effects 
of shared haplotypes on mean completed family size (Anderson et al. 
1980, pp. 199-234; Cox 1972). Reproductive outcome was assessed 
indirectly by estimates of the length of the interval from first birth to 
each successive birth. Time from year of first birth was used rather than 
from date of marriage because there was no information on the date 
of marriage for the women of the Meridian and Goessel congregations. 
Couples were considered to be at risk from the time of the first birth 
through the time of menopause (or the surgically induced end of the 
fertility period). The log-rank test was used to compare the time-to-next-
birth curves. 

The distribution of offspring from various kinds of matings formed 
the basis of a segregation analysis. Ratios based on Mendelian inheritance 
were used to predict expected frequencies of offspring; these ratios 
were then compared in a chi-square analysis to actual distributions of 
phenotypes in the offspring. 

A minor adjustment was made for phenotype GM A,X G. The phe-
notype GM A,X G consists of two genotypes (IGHG1*A G3*G/G1*A,X 
G3*G and IGHG1*A,X G3*G/G1*A,X G3*G). Therefore the transmis-
sion likelihood for the IGHG1*A,X G3*G allele will be greater than 0.5. 
The estimate of the transmission likelihood can be calculated using the 
expected values from the Hardy-Weinberg equation. The formula for the 
calculation of the probability of transmission for ambiguous situations is 

t = [(*A,XGHt/*A,XGT) x 0.5] + [(*A,XGHm/*A,XGT) x 1.0], (1) 
where *A,X G Ht is the Hardy-Weinberg estimate of the frequency of 
heterozygous *A,XG, *A, X G Hm is the Hardy-Weinberg estimate of 
the frequency of homozygous *A,XG, and *A,XGT is the sum of het-



Immunoglobulin Haplotypes / 483 

Table 1. Summary Statistics for Completed Family Size 
Completed Number ofHalotypes Shared 
Family Size 0 % 1 % 2 % Totals 
0 3 0.130 9 0.070 6 0.066 18 0.074 
1 0 0 5 0.039 4 0.044 9 0.037 
2 4 0.174 12 0.094 11 0.121 27 0.112 
3 5 0.217 38 0.297 26 0.286 69 0.285 
4 4 0.174 39 0.305 30 0.330 73 0.302 
5 4 0.174 16 0.125 5 0.055 25 0.103 
6 3 0.130 4 0.031 6 0.066 13 0.054 
7 0 0 1 0.008 2 0.022 3 0.012 
8 0 0 2 0.016 1 0.011 3 0.012 
9 0 0 2 0.016 0 0 2 0.008 
N 23 0.999 128 1.001 31 1.001 242 0.999 
Mean number of 3.348 ± 1.849 3.469 ± 1.688 3.374 ± 1.596 3.421 ± 1.664 
offspring 
Median number of 3 3.5 3 3 
offspring 
Modal number of 3 4 4 4 
offspring 
Skewness - 0 . 3 7 4 ± 0.481 0.339 ± 0.214 0.051 ± 0.253 0.160 ± 0.156 
Kurtosis - 0 . 5 4 9 ± 0.935 1.678 ± 0.425 0.690 ± 0.500 1.057 ± 0.312 

erozygous and homozygous frequencies. The probability of transmission 
estimated using US white frequencies (Dykes 1982) is 0.5995. 

Results 
The distribution of completed family sizes for the total Mennonite 

sample and for the three subgroups of the Mennonite sample are pre-
sented in Table 1. Completed family sizes for the total sample range from 
0 to 9 with a mean of 3.42 (±a, a = 1.66), a median of 3, and a mode of 4 
children. The distribution is leptokurtic. The three subgroups consist of 
23, 128, and 91 families for 0, 1, and 2 haplotypes shared, respectively, 
with corresponding mean completed family sizes of 3.35 ± 1.85, 3.47 
± 1.69, and 3.37 ± 1.60. The variances of the three subgroups are not 
significantly different (Bartlett's Box F = 0.361, p = 0.697; Cochrane's 
C = 0.380, p = 0.414). 

Analysis of variance was performed to determine the effects of 
shared haplotypes on completed family size, and the results are presented 
in Table 2. The low F value (0.145) shows that mean completed family 
size does not vary significantly {p = 0.865) among the subgroups. 

The results of the birth interval analyses are presented in Table 3. 



4 8 4 / STEVENSON ET AL. 

Table 2. A n a l y s i s of Variance: C o m p l e t e d Family S i z e by H a p l o t y p e Subgroup 
Sum of Degrees Mean F F 

Source Squares of Freedom Squares Ratio Probability 
Total 664.847 241 
Between groups 0.806 2 0.403 0.145 0.865 
Within groups 664.041 239 2.778 

Table 3. Intervals ( Y e a r s ) from First Birth to S u c c e s s i v e B i r t h s 
( U p p e r Quartile)* 

Haplotype Subgroups Chi-Square Degrees of 
Parity 0 1 2 Value Freedom Probability 
2 4.08 4.03 3.67 0.465 2 0. 75 < p < 0.90 
3 8.25 8.05 8.50 0.929 2 0.75 < p < 0.90 
4 11.75 12.20 13.25 0.093 2 0.95 < p < 0.975 
a. Upper quartiles: 75% of couples completed the interval from first birth to this birth 

within the time period (number of years). Upper quartiles are estimated from time-to-
birth curves generated from survival analysis; P values (two-tailed) are based on the log-
rank test and compare the full time-to-birth curves for each parity. 

The length of the interval from the first to the fourth birth does increase 
slightly, as seen in Figure 1. However, the length of the intervals from first 
birth to second, first to third, and first to fourth were not significantly 
different for any of the three subgroups. 

The results of the segregation analyses are presented in Table 4. 
There are two statistically significant deviations from Mendelian ex-
pectations. Mothers with phenotype GM F B mated to fathers with 
phenotype GM A,X G have a slight excess of GM A,F,X,Z B,G and 
slight deficiency of GM A,F B,G phenotypes in the offspring. This is 
most likely an artifact of the fact that the GM A,X G phenotype is 
a combination of two genotypes, IGHG1 *A,X G3*G/Gl *A G3*G and 
IGHG1*A,X G3*G/G1 *A,X G3*G; neither can be distinguished with-
out further pedigree analysis. In addition, mothers with phenotype GM 
A G mated to fathers with GM A,F B,G produce a slight excess of 
GM A,F B,G phenotypes and a slight deficiency of GM A G pheno-
types in the offspring. Small sample size (28) may account for this dis-
crepancy. This last deviation from Mendelian expectations, if biologi-
cally real, would represent low-level selection against the homozygous 
IGHG1 *A G3*G/G1 *A G3*G genotype. In the 64 matings of moth-
ers with phenotype GM A,F,X B,G with fathers with GM A,F B,G, the 
results are not statistically significant given four outcomes of equal prob-
ability. The chi-square value could also be calculated as a 1:3 ratio of GM 



Immunoglobulin Haplotypes / 485 

1 .00 

0 . 9 0 
0 . 8 0 
0 . 7 0 
0 . 6 0 

0 . 5 0 
0 . 4 0 
0 . 3 0 
0 . 2 0 

0 . 1 0 

0 . 0 0 • T 

? i 

-I t- H 1 1 f- H H H h 
0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 

Time (in Years) t o F o u r t h Birth 

Figure 1. Time to third child by haplotype subgroup. Filled circles, 0 shared haplotypes; 
open circles, 1 shared haplotype; squares, 2 shared haplotypes. 

F B versus the other three phenotypes. Observed (to expected values) of 
24 (16):40 (48) result in a chi-square value of 5.33, which for 1 degree of 
freedom is now significant (0.2 < p < 0.5). Thus there is a slight excess of 
GM F B phenotypes, or in other words, evidence for low-level selection 
in favor of the IGHG1 *F G3*B/G1 *FG3*B homozygotes. 

Discussion and Conclusions 
It is HLA or GM compatibility (rather than incompatibility) that 

is correlated by Weitkamp and colleagues with a reduction in fertility 
in couples not selected for excessive abortions (Ober et al. 1983, 1986; 
Weitkamp 1986; Weitkamp et al. 1986) and in couples with abortions 
(Gill 1983b). It is likely that both immunologic and genetic mechanisms 
are operating (Gill 1983a; Gill and Repetti 1979; Gill, MacPherson et al. 
1987; Hedrick 1988; Hedrick and Thomson 1988; Schacter et al. 1984; 
Weitkamp et al. 1986). There is growing evidence that histoincompati-
ble fetuses elicit a maternal response to paternal antigens that increases 
the frequency of successful implantation and may prevent rejection of 
the implanted ovum (Gill, MacPherson et al. 1987). It is likely that an 
unidentified class 1 antigen is the primary stimulus in eliciting the im-
mune response to the conceptus, and in humans the antigen is coded by 



4 8 6 / STEVENSON ET AL. 

Table 4. Segregation Analysis 
GM Phenotypes 

of Parents 
(at least one parent GM Phenotypes of Offspring: 

heterozygous) Observed (Expected) 
Offspring Mother Father FB A,FB,G A,F,XB,G AG A,XG X2 
28 A G A,FB,G 0 20 0 8 0 5.1429* 

(14) (14) 
3 A G A,F,X B,G 0 1 0 0 2 ID 
4 A G A , X G 0 0 0 2 2 ID 
10 A,FB,G A G 0 4 0 6 0 ID 
128 A,F B,G A,FB,G 33 65 0 30 0 0.1719 

(32) (64) (32) 
21 A,FB,G A,F,X, B,G 5 9 6 0 1 ID 
22 A,F B,G A,X G 0 7 4 4 7 ID 
152 A,F B,G F B 79 73 0 0 0 0.2368 

(76) (76) 
2 A,F,X B,G A G 0 0 0 0 2 ID 
64 A,F,X B,G A,F B,G 24 14 13 0 13 5.3750 

(16) (16) (16) (16) 
42 A,F,X B,G A,F,X B,G 10 0 24 0 8 1.0476 

(10.5) (21) (10.5) 
17 A,F,X B,G A , X G 2 0 8 0 7 4.7647 

(4.25) (4.25) (8.5) 
102 A,F,X B,G F B 49 53 0 0 0 0.1569 

(51) (51) 
6 A,X G A G 0 0 0 4 2 ID 
16 A,X G A,F B,G 0 5 6 2 3 ID 
16 A,X G A,F,X B,G 0 4 4 0 8 ID 
6 A,X G A,X G 0 0 0 3 3 ID 
40 A,X G F B 0 16 24 0 0 0.0000 

(16.02) (23.98) 
123 F B A,F B,G 68 55 0 0 0 1.3740 

(61.5) (61.5) 
68 F B A,F,X B,G 30 0 38 0 0 0.9412 

(34) (34) 
53 F B A,X G 0 13 40 0 0 5.3182* 

(21.2265) (31.7735) 
923 300 339 167 59 58 
ID = Insufficient data. 
*0.02 < P< 0.05. 



Immunoglobulin Haplotypes / 487 

gene(s) in the HLA A,B,C region of the major histocompatibility com-
plex (MHC) [Gill 1983b; see also Faulk and Mclntyre (1981), Hedrick and 
Thomson (1988), Mclntyre and Faulk (1979, 1982), Ober et al. (1985), 
Ober, Elias, O'Brien et al. (1988), and Ober, Elias, Hauck et al. (1988)]. 
In addition, there may be lethal recessive genes (such as the t haplotypes 
in the mouse and the grc in the rat) epistatically linked to the HLA loci 
that may come together more frequently in homozygous matings, result-
ing in increased frequencies of spontaneous abortions (Gill et al. 1983; 
Gill, Wegmann et al. 1987; Hedrick 1988; Schacter et al. 1984). 

Whatever the mechanism(s), there is little evidence here for reduced 
fertility in matings homozygous for GM haplotypes. Statistically signif-
icant differences were not demonstrated among the three subgroups of 
this study for mean completed family sizes or mean times from first to 
second, first to third, and first to fourth births. The segregation analy-
sis provided little evidence for statistically significant deviations from 
Mendelian expectations. 

Low-level selection is not ruled out. Birth histories collected from 
Mennonite women reveal that miscarriages and stillbirths were reported 
for 22.7%, 20.9%, and 25.3% of women who share 0, 1, and 2 haplo-
types, respectively, with their husbands. Plus, the trend to increasing in-
tervals between first and fourth births as the number of haplotypes shared 
increases, although not statistically significant, is suggestive. Future re-
search should include more information on birth and marriage. More 
sensitive detection of fetal loss would also be helpful [see Ober, Elias, 
Hauck et al. (1988)]. Finally, detection may be possible only in large 
samples or samples consisting primarily of couples who are maximizing 
family size (e.g., Hutterites). 
Acknowledgments This research was supported in part by the National 
Institutes of Health under grant AGO 1646-03. We also wish to thank the Goessel, 
Meridian, and Henderson congregations and the anonymous reviewers. 

Received 17 January 1989; revision received 28 November 1989. 

Literature Cited 
ANDERSON, S., A . AUQUIER, W . W . HAUCK, D . OAKES, W. VANDAELE, H . I . 

WEISBERG, A.S. BRYK, AND J. KLEINMAN 1980 Statistical Methods for Compar-
ative Studies: Techniques for Bias Reduction. New York: Wiley. 

COX, D.R. 1972 Regression models and life-tables. J. R. Statist. Soc., ser. B, 34:187-202. 
CRAWFORD, M.H., AND L. ROGERS 1982 Population genetic models in the study of aging 

and longevity in a Mennonite community. Soc. Sci. Med. 16:149-153. 



4 8 8 / STEVENSON ET AL. 

DYKES, D.D. 1982 The use of frequency tables in parentage testing. In Probability of 
Inclusion in Paternity Testing: A Technical Workshop, H. Silver, ed. Arlington, Va.: 
American Association of Blood Banks, 15-44. 

FAULK, W.P., AND J.A. MClNTYRE 1981 Trophoblast survival. Transplantation 32:1-5. 
GILL, T.J., III 1983a Immunogenetic aspects of the maternal-fetal interaction. In Im-

munology of Reproduction, T.G. Wegmann and T.J. Gill, III, eds. N e w York: Oxford 
University Press, 55-76. 

GILL, T.J., III 1983b Immunogenetics of spontaneous abortions in humans. Transplanta-
tion 35:1-6. 

GILL, T.J., III, AND C.F. REPETTI 1979 Immunologic and genetic factors influencing 
reproduction. Am. J. Pathol 95:465-570. 

GILL, T.J., III, S. SLEW, AND H.W. KUNZ 1983 Major histocompatibility complex 
(MHC)-linked genes affecting development. J. Exp. Zool. 228:325. 

GILL, T.J., III, T.G. WEGMANN, AND E. NlSBET-BROWN 1987 Immunoregulation and 
Fetal Survival. London: Oxford University Press. 

GILL, T.J., III, T.A. MACPHERSON, H.N. HO, H.W. KUNZ, A.C. HASSETT, K.S. 
STRANICK, AND J. LOCKER 1987 Immunological and genetic factors affecting im-
plantation and development in the rat and in the human. In Immunoregulation and 
Fetal Survival, T J . Gill, III, T.G. Wegmann, and E. Nisbet-Brown, eds. London: 
Oxford University Press, 137-155. 

HEDRICK, P.W. 1988 HLA-sharing, recurrent spontaneous abortion, and the genetic hy-
pothesis. Genetics 119:199-204. 

HEDRICK, P.W., AND G. THOMSON 1988 Maternal-fetal interactions and the mainte-
nance of HLA polymorphism. Genetics 119:205-212. 

JOHNSON, W.E., P.H. KOHN, AND A.G. STEINBERG 1977 Gm and Km (Inv) frequencies 
in two Roumanian populations. Hum. Genet. 39:199-211. 

MClNTYRE, J.A., AND W.P. FAULK 1979 Antigens of human trophoblast: Effects of 
heterologous anti-trophoblast sera on lymphocyte responses in vitro. J. Exp. Med. 
149:824. 

MClNTYRE, J.A., AND W.P. FAULK 1982 Allotypic trophoblast-lymphocyte crossreactive 
(TLX) cell surface antigens. Hum. Immunol. 4:27-35. 

OBER, C.L., L. WEITKAMP, AND S. ELIAS 1986 Gm mating types and family size differ-
ences among Hutterite couples. Abstr. Am. J. Hum. Genet. 39:A242. 

OBER, C.L., S. ELIAS, W.W. HAUCK, AND D. KOSTYU 1988 Evidence for genes in more 
than one MHC region influencing fertility in Hutterite couples. Abstr. Am. J. Hum. 
Genet. 43:A219. 

OBER, C.L., S. ELIAS, E. O'BRIEN, D.D.' KOSTYU, W.W. HAUCK, AND A. BOMBARD 
1988 HLA sharing and fertility in Hutterite couples: Evidence for prenatal selection 
against compatible fetuses. Am. J. Reprod. Immunol. Microbiol. 18:111-115. 

OBER, C.L., W . W . HAUCK, D . D . KOSTYU, E. O'BRIEN, S. ELIAS, J.L. SIMPSON, A N D A.O. 
MARTIN 1985 Adverse effects of human leukocyte antigen-DR sharing on fertility: 
A cohort study in a human isolate. Fertil. Steril. 44:227-232. 

OBER, C.L., A.O. MARTIN, J.L. SIMPSON, W.W. HAUCK, D.B. AMOS, D.D. KOSTYU, 
M. FOTINO, AND F.H. ALLEN, JR. 1983 Shared HLA antigens and reproductive 
performance among Hutterites. Am. J. Hum. Genet. 35:994-1004. 

ROPARTZ, C., L. RIVAT, AND P.-Y. ROUSSEAU 1964 Les a n t i g l o b u l i n s humaines: Leur 
specificite, leur frequence et les mecanismes probables de leur apparition. Nouv. Rev. 
Fr. Hematol. 4:803-824. 

ROPARTZ, C., L. RIVAT, P.-Y. ROUSSEAU, H . BAITSCH, AND J. VAN LOGHEM 1963 Les 
systemes Gm et Inv en Europe. Acta Genet. (Basel) 13:109-123. 



Immunoglobulin Haplotypes / 489 

SCHACTER, B., L.R. WEITKAMP, A N D W.E. JOHNSON 1984 Parental H L A compatibility, 
fetal wastage and neural tube defects: Evidence for a T/t-like locus in humans. Am. 
J. Hum. Genet. 36:1082-1091. 

SCHANFIELD, M.S. 1978 Genetic markers on human immunoglobulins. In Review of 
Basic and Clinical Immunology, 2d ed., J.L. Caldwell, H.H. Fudenberg, J.V. Wells, 
and D.P. Stites, eds. Los Altos, Cal.: Lange Medical Publications, 59-65. 

SCHANFIELD, M.S. 1980 The anthropological usefulness of highly polymorphic systems: 
HLA and immunoglobulin allotypes. In Current Developments in Anthropological 
Genetics. M.H. Crawford and J.H. Mielke, eds. New York: Plenum Press, 65-85. 

SCHANFIELD, M.S., H.F. POLESKY, AND E.B. SEBRING 1975 G m a n d Inv testing. In 
Paternity Testing, H.F. Polesky, ed. Chicago, 111.: American Society of Clinical 
Pathologists, Division of Educational Media Services, 45-54. 

SHOWS, T.B., P.J. MCALPINE, C. BOUCHEIX, ET AL. 1987 Guidelines for human gene 
nomenclature: An international system for human gene nomenclature (ISGN, 1987). 
Cytogenet. Cell Genet. 46:11-28. 

STEINBERG, A.G. 1969 Globulin polymorphisms in man. Annu. Rev. Genet. 3:25-52. 
STEINBERG, A.G., AND C.E. COOK 1981 The Distribution of the Immunoglobulin Allo-

types. London: Oxford University Press. 
STEVENSON, J.C., P.M. EVERSON, AND M.H. CRAWFORD 1989 Changes in completed 

family size and reproductive span in Anabaptist populations. Hum. Biol. 61:99-115. 
WEITKAMP, L.R. 1986 G m a n d H L A m a t i n g types in relation to reproductive success. 

Abstr. Am. J. Hum. Genet. 39:A248. 
WEITKAMP, L.R., C. OBER, AND B. SCHACTER 1986 G e n e t i c a n d i m m u n o g e n e t i c mark-

ers of reproductive success. In Reproductive Immunology, D.A. Clark and B.A. Croy, 
eds. New York: Elsevier, 27-35.