Click for next page ( 89


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 88
Chapter VI! ANALYSIS OF THE SEX RATIO DATA THE characteristic for which the most data are available is the sex of the offsprings. 7.1 The trait. Among the various charac- teristics of the infants examined in connection with this study, sex is undoubtedly that at- tribute in whose determination there is the smallest margin of error. Although the sexing of prematurely borne infants presents certain well-recognized problems (cf. Tietze, 1948), by the twentieth week of gestation, the earliest at which pregnant women were permitted to register, there is seldom any room for doubt con- cerning the sex of a prematurely born infant." 7.2 The genetic argument for radio~tion-ir;- d~ced charges in the sex ratio. Lethal muta- tions, i.e., mutations resulting in the death of the organism sometime prior to reproduction, are an important type of mutation arising in consequence of irradiation of experimental ani- mals. Induced lethal mutations may be dominant or recessive to their normal alleles. Such mutants are of concern to us in this chapter only inso- far as they affect the sex ratio. The heteroga- metic nature of the human male results in (a) a differential distribution to his offspring of his sex chromosomes, and (b) a difference in the dominance relationships between the two sexes. Thus lethal mutations borne on the X- or Y-chromosomes afford ample opportunity for an alteration of the sex ratio. Accordingly, our discussion shall concern itself with this class of mutants, i.e., the X- and Y-borne lethals. Let us examine separately the possible effects on the sex ratio of maternal irradiation, paternal irradiation, and conjoint parental exposure. ~ Because of unavoidable uncertainties in determin- ing the duration of gestation, a few women may have registered prior to the twentieth week of gestation. However, the contribution of infants of less than twenty weeks of gestation in age to the sex ratio may be regarded as negligible. Sex ratio is here defined as the proportion of male births among all births. 88 Maternal irradiation would be tantamount to the irradiation of X-chromosomes alone. Since there is no differential distribution of the ma- ternally derived X-chromosomes among the off- spring, induced sex-linked dominant lethals would be expected to result in comparable re- ductions in the frequency of male and female births. The same would hold true for all par- tially sex-linked lethals if such exist. Induced sex-linked recessive lethals would find expres- sion only in the hemi~ygous male or the homo- ~ygous female. The much more frequent occur- rence of the former would lead to a greater diminution in male than in female births with a subsequent decrease in the sex ratio. Partially sex-linked lethal genes would result in distor- tion of the sex ratio within specific families, but at the population level, the sexes would be equally affected. Thus the net effect of maternal irradiation insofar as it affects the sex ratio would appear to be a reduction in the frequency of male births. In the event of paternal exposure, the situa- tion would be appreciably altered. Induced sex- linked lethal mutants, be they recessive or domi- nant, would be distributed only to the female. Thus sex-linked lethals arising in the differential segment of the paternal X-chromosome would lead to a reduction in the frequency of female births, or an augmentation of the sex ratio. Similarly, Y-borne lethals would be distributed only to sons of irradiated fathers; however, the evidence for the existence of any Y-borne genes in man is of the poorest order. The net effect of paternal exposure, all types of lethal mutants considered, would be a reduction in the fre- quency of female births. If both parents were exposed, the net effect on the sex ratio would presumably be the sum of the parental effects considered separately. Since the effect on the sex ratio is different and in opposite directions for the two parents, it

OCR for page 88
Analysis of the Sex Ratio Data would appear that if both parents were exposed there would be either no appreciable effect on the sex ratio, or if an effect existed it would be at the expense of the males. The latter is con- ceivable because sex-linked recessive lethals are more frequent in occurrence than sex-linked dominant lethals. So far as irradiation effects are concerned, the sex ratio is unique among the indicators of genetic damage. Its uniqueness arises from the fact that, as has been indicated, maternal ex- posure would be expected to produce an effect different from paternal exposure. 7.3 Concomitant variation ir~5z~encing the indicator. Before we turn to a consideration HE: EiERTT>s of the data regarding the effect of parental ex- 4OI . posure on the sex ratio, a brief consideration of the sex ratio as a variable is indicated. While it is true that in some respects the sex ratio may be the most specific of the indicators of genetic damage herein studied, in other respects it is a less satisfactory variable. A variety of factors have been alleged to influence the outcome, with regard to sex, of a pregnancy. Among these factors are maternal age (Lowe and McKeown, 1950), paternal age (Novitski, 1953), birth order (Ciocco, 1938), race (Strandskov, 1945), urban versus rural origin of the parents (Ciocco, 1938), and social strata (Bernstein, 1948~. In the main, the effects of these variables are small. For example, Ciocco (1938) has shown that the difference in the sex ratio between first and fifth or higher born children is the difference between 0.5153 and 0.5124. The relationship in these data between maternal age and parity, on the one hand, and sex ratio, on the other hand, is illustrated in Figures 7.1 and 7.2. The fact that these concomitants exert such small effects and the difficulty in classifying some of them, has led us not to attempt any adjustment for maternal age, paternal age, birth order, social strata, or origin of the parents in the Japanese data. 7.4 The data.-In Table 7.1 are given the frequencies of male births by parental exposure and cities. Inspection of this table reveals no striking differences in the sex ratio among the classes of parental exposure. The analysis of these data is given in Tables 7.2a and 7.2b. Analysis fails to reveal a significant effect of cities, mother's exposure, or father's exposure, and there exists no evidence from the interac- tions of significant heterogeneity among these data. 89 As we have previously indicated, the sex ratio is the one indicator of irradiation damage wherein genetic theory specifies, with the great- est precision, the direction of change anticipated if irradiation exerts an effect on the sex ratio. Accordingly, we may ask what these data would show under a more restrictive set of alternatives to the null hypothesis, that is to say, under a one-tailed significance test. This alternative - 04 30. . . I ~1 - - I I I 5 80 as ~35 40, ~ 5 MOTi1ER S AGE FIGURE 7.1-The distribution of the frequency of male births by age of mother at the birth of the infant with parity ignored. 85 701 60 MALE EARTHS (PER CENT) 50 40 1 2 3 4 S 6 7 B. PARITY FIGURE 7.2 The distribution of the frequency of male births by parity with maternal age ignored. analysis could proceed as follows: it may be argued that (1) the genetic effect of irradiation mediated through exposed mothers would be greater than that mediated through exposed fathers and hence the status of the father's exposure may be disregarded without undue loss of information. It may be further argued that (2) the absence of evidence for hetero- geneity between cities permits pooling the ob- servations from the two cities. Finally, (3) we have advanced reasons for believing that parents who were unexposed may not be in pari materia

OCR for page 88
90 Genetic EJec`.r of Atomic Bombs Chapter VII TABLE 7.1 THE FREQUENCY OF MALE BIRTHS BY PARENTAL EXPOSURE AND CITY (Unrelated parents) Hiroshima Fathers 1 2 3 4-5 Total Jn17,2941,50059639519,785 1 <:9,00577231320910,299 up.5207.5147.5252.5291.5205 (n5,3681,8503852447,847 2] ~2,8329411971224,092 I P.5276.5086.5117.5000.5215 n2,1854245211573,287 34 ~1,114220268801,682 -up.5098.5189.5144.5096.5117 (n1,1172021101171,546 4-5 g c35711095456790 up.5112.5396.4909.4786.5110 In 25,964 3,9761,612913 32,465 Total: ~13,522 2,042832467 16,863 up .5208 .5136.5161.5115 .5194 o Nagasaki Fathers ~_ . 1 2 3 4-5Total In14,6102,17024313917,16 14 ~7,6081,120129758,93~ (p.5207.5161.5309.5396.5204 rn9,3164,14427317813,911 24 ~4,8492,1121401037,209 up.5205.5097.5128.5787.517c In74727994351,155 34 ~3601345114559 up.4819.4803.5426.4000.4840 In5591163528738 4-5 J ~279561815368 Ip.4991.4828.5143.5357.4986 In 25,232 6,709645380 32,966 Total] ~13,096 3,422338207 17,063 up .5190 .5101.5240.5447 .5176

OCR for page 88
Analysis of the Sex Ratio Data with parents who were exposed, and therefore terminations occurring to mothers who were un- exposed should be rejected. If now we proceed on these three assumptions we may test whether the sex ratio is lower among the offspring of all mothers in exposure classes 3, 4, and 5 than it is among the offspring of all mothers in ex- posure class 2. The pooling of exposure class 3 with 4 and 5 serves to increase the number of observations which may be brought to bear on the question of a maternal exposure effect, and may be interpreted as a conservative step since TABLE 7.2 CHI-SQUARE ANALYSIS OF THE FREQUENCY OF MALE BIRTHS BY CITY AND PARENTAL EXPOSURE (Unrelated parents) a. All exposure cells (4 X 4) Total Interactions, first order CF CM ME Main effects city ~c) 1 Mother (M) 3 Father (F) 3 DF %2 31 20.118 3 1.498 3 2.582 9 3.606 b. Excluding parents with exposure 1 Total Interactions, first order - CF CM MF Main effects city tc' 1 Mother (M) 2 Father (F) 2 17 8.485 2 1.940 2 1.099 4 1.675 0.219 .50-.70 5.327 .10-.20 2.630 .30-.50 0.002 .95-.98 0.082 .95-.98 0.127 .90_.95 it lowers the mean dose in the "heavily" exposed group. The results of such a comparison are presented in Table 7.3 (Comparison 1) from which we note a X2=3.925 which on a one- tailed test of significance corresponds to a prob- ability level of 0.02. There may, then, be an effect of maternal exposure on the sex ratio. We could now ask whether there exists ancillary evidence which may shed light on the reality of this observed difference. Let us consider first what may be termed the "internal" evidence, that is to say, evidence which can be obtained from further analysis of the data given in Table 7.1. To this end four other comparisons are given in Table 7.3. These comparisons are as follows (the numbers corre 91 spend to the numbers given in Table 7.3~: a comparison of the sex ratio among terminations occurring to (2) mothers in class 2 with mothers in classes 3, 4, and 5 bat eliminating all those observations where the father was un exposed, (3) mothers in class 2 whose spouses were unexposed with mothers in classes 3, 4, and 5 whose spouses were also unexposed, (4) mothers in class 2 whose spouses were unex posed with mothers in class 2 whose spouses were exposed, and lastly (5) mothers in class 2 with mothers in classes 3, 4, and 5 but elimi nating the single cell where mothers were in class 2 and had unexposed spouses, i.e. the M2F, cell. The arguments underlying these comparisons and the results obtained are as follows: P Comparison 2: The exclusion of mothers in .90_.95 class 1 is based on the assumption that this group is not comparable to groups 2 to 5 with .50_.70 regard to a number of concomitants, including .30-.50 age of mother, parity, and origin. From Tables 90-95 5.3 and 5.6 it should be noted that exclusion of mothers in class 1 alone leaves a series of cells, namely, mothers 2, 3, 4, and 5 and fathers 1, in which the mothers are clearly younger and have had fewer pregnancies, on (3 X 3) the average, than is true of the mothers in the .95_.9g' remaining exposure cells. While the effect of parity and maternal age on the sex ratio is cer .30_.50 tainly not striking, what effect there is suggests .50-.70 that younger mothers have a higher proportion 70- 80 of male births, and that first parities more fre quently terminate in male infants (Ciocco, 1938~. In this connection, it is interesting to note that in both cities one of the highest sex ratios among the numerically large cells occurs in that cell where the mothers are the youngest and have had the fewest pregnancies (the mothers 2-fathers 1 cell). It would be informa tive, then, to know whether a significant differ ence would obtain if the differences in maternal age and parity were further reduced by elimi nating those cells in which the fathers were unexposed. Comparison 2 in Table 7.3 affords information on this point. We find not only no significant difference when unexposed fathers are rejected but also that the absolute value of the difference is smaller than that obtained when unexposed fathers were included in the compari son. The direction of the difference is, however, still consistent with the genetic argument. It may, of course, be argued that the lack of a

OCR for page 88
~2 significant effect in this comparison is not unex- pected because of the depletion of the sample, and the fact that the absolute difference is smaller could merely reflect an effect of fathers. Comparison 3: The comparison which would maximize an effect due to mother's exposure if TABLE 7.3 SELECTED COMPARISONS REGARDING THE EFFECT OF IRRADIATION ON SEX-RATIO (See text for a.rs~mptior~s Underlying the selection of comparisons. ) (1) Comparison: Ma with Ms ~ (summing over cities and fathers) Sex of infant r Ma 11,296 10,462 (0.5192) or M(3 5) 3,399 3,327 (0.5054) Total 14,695 13,789 28,484 X~, = 3.925* Number of males per Total 100 females 21,758 107.99 6,726 102.18 (2) Comparison: ~ with Ms ~ (summing over cities and fathers but eliminating fathers 1) Sex of infant ~ ~e' Ma 3,615 3,459 (0.5110) M(3 5) 1,075 1,043 1 (0.5076) Total 4,690 4,502 X21 = 0.786 Number of males per Total 100 females 7,074 104.50 2,118 103.09 9,192 (3) Comparison: M~F1 with M: GF~ (summing over . . . ~ rem ce 0 x au cities J Sex of infant Number of males per Total 100 females 14,684 109.69 M~F1 7,681 7,003 (0.5231) Ma 5F1 2,324 2,284 (0 5043) Total 10,005 9,287 X21 = 4.9374 4,608 101.73 19,292 (4) Comparison: M:F1 with M:F~- (summing over cities ) Sex of infant M~F1 7,681 7,003 (0.5231) _ 5 3,615 3,459 (0.5110) Total 11,296 10,462 21,758 X21 = 2.782 Genetic Efects of Atomic Bombs Chapter N7II TABLE 7.3-Continued (5)Comparison: Ma with Ma 6 (summing over cities, fathers, but excluding fathers 1, mothers 2 cell) , ^ Sex of infant ~2 3,615 3,459 (0.5110) 3,399 3,327 6,726 102.18 (0.5054) [Total 7,014 6,786 X21 = 0.444 Number of males per Total 100 females 14,684 109.69 7,074 104.50 Number of males per Total 100 females 7~074 104.50 13,800 the effect of father's exposure was not negligible would be a comparison of the sex ratio among the children of mothers in class 2 whose hus- bands were unexposed with the ratio among the children of mothers in classes 3 to 5 whose husbands were also unexposed. Moreover, from Tables 5.3 and 5.6 we note that these cells are not too dissimilar with regard to mean maternal age and parity. This comparison results in a X2 = 4~937, which is clearly significant. The occurrence of a more striking difference here than in comparison 1 could be interpreted as indicating that the effect of father's exposure is not negligible. Comparison 4: If the effect of father's ex- posure is not negligible, as the preceding com- parison might suggest, we could appraise this effect by holding mother's exposure constant (class 2) and then contrasting the offspring of unexposed fathers with those of exposed fathers. This should give rise, on genetic theory, to a higher sex ratio among the children of exposed fathers than among unexposed fathers. From comparison 4, however, we note that the direc- tion of change is opposite to hypothesis, but this difference is not significant. The compari- son does, however, raise questions as to the validity of using as a "control" the mothers 2-fathers 1 data. The latter is further borne out by a comparison of M2F, with M2F2 wherein we would expect no differences, whereas we find a striking though not significantly increased sex ratio in the M2F, cell (X~=3.221~. Comparison; 5: Comparison 4 leads one to conclude that it would be informative to con- trast the sex ratio when mothers are in class 2 with mothers in classes 3 to 5 tent eliminating the single cell in which the mothers were 2 and the fathers were unexposed. When this is done, a non-significant difference results. This further suggests the critical role played by the rejected

OCR for page 88
Analysis of the Sex Ratio Data cell in the tests of significance. At this point it seems appropriate to introduce evidence on the sex ratio available from the Japanese vital sta- tistics which further imply the occurrence of an inexplicable inflation of the sex ratio in the mothers 2-fathers 1 cell. In Table 7.4 are given the national average sex ratio and the average sex ratio for Hiroshima and Nagasaki for the period 1935-1952. In terms of numbers of males per 100 females we note that at no time during the period 1935-1952 do the combined city averages or the national average equal much less exceed, the 109.7 found in the M2F, cell.2 The sex ratio obtaining in Kure during the years 1948-1950 was 0.5133 or 105.4 males per 100 females, a figure also lower than that observed in the M2F, cell. It should be noted, however, that the data from "heavily" exposed mothers deviate from the national average as well as the city averages in the manner antici- pated by genetic theory. - Let us turn now to the "external" evidence bearing on the difference we have been dis cussing, namely, the evidence afforded by com parisons other than those arising from Table 7.1. Several sources of evidence are pertinent here, namely, the stillbirth data, the early termi nation data, and the sex ratio among births occurring prior to the period covered by the Genetics Program. If the depression of the sex ratio observed among terminations to "heavily" exposed mothers is due to induced sex-linked lethal genes, then we might logically expect to find evidence of increased male mortality among stillborn infants or among pregnancies termi- nating before 21 weeks of gestation. In this connection, it is interesting to note that the sex- mother interaction in Table 9.5 (the stillbirth data) reveals no evidence of heterogeneity among the sexes with regard to mother's ex- posure. We have previously stated (Sec. 2.8) that an attempt was made to obtain data on pregnancies terminating before 21 weeks of gestation. We have further stated that because these data are deficient in a number of respects 2 It should be noted that the data in Table 7.4 rep- resent only liveborn infants. Since the data presented in Table 7.1 are for all births, and since there is a relative excess of males among stillbirths, the sex ratios given in Table 7.4 are not strictly comparable with those in Table 7.1, being biased in a downward direction. It seems unlikely that allowance for this bias would bring the M2F, cell findings in line. 93 they will not be presented in this report. How- ever, for what it is worth, a preliminary analy- sis of these data failed to reveal evidence that "heavily" exposed mothers more frequently abort or miscarry male infants than do unex- posed or "lightly" exposed mothers. If, then, one accepts the observed "exposure effect" as real, it is necessary to postulate that the induced sex-linked lethals are either gametic lethals or act so early in embryogenesis that the affected feti are resorbed or ejected at a time when the sex of the abortus cannot be readily determined. Finally, when in a preliminary analysis of the data in 1953 it seemed likely that a signifi- cant relationship between maternal exposure and the sex ratio might exist, an effort was made to collect and scrutinize other data which would bear on this relationship. To this end, data were gleaned from the records of the Genetics Pro- gram as well as other ABCC records pertinent to the sex ratio among births occurring in Hiro- shima and Nagasaki following the atomic bomb- ings but prior to the inception of the Genetics Program. The data to be presented in this section were collected from the histories of reproductive per- formance prior to approximately June, 1948 obtained on: 1. Those women who had been delivered of a registered or unregistered infant in the years 1948-1952 and known to ABCC through the Genetics Program. This information was available on only those parents who had pro- duced a child on whom a "Genetics Long Form" was available (see Sec. 2.2 ~ . 2. All adults seen by the Commission under Project ME-55 (a random sample of adults exposed at distances of less than 2,000 meters and their controls). 3. Those women who had reported a spon- taneous abortion and were seen under the study of early terminations mentioned in Section 2.8. These three sources of data were cross-refer- enced to prevent duplication. In all, informa- tion was obtained on 8,824 pregnancies termi- nating between April, 1946 and approximately dime, 1948. The former date was selected to avoid inclusion of pregnancies wherein the fetus might have been exposed in liters, and the latter date was so chosen as to prevent over- lapping of these data with the Genetics Pro- gram. The sex of these infants is given in Table

OCR for page 88
94 Genetic EfJects of Atomic Bombs Chapter VII -. -% o _ 00 (-~ cr~ ~ ~\ \C~ ~ ~ 0 or - - to ~ vow 0 0 (3,) ~ _1 cat 0 co ~ ~ ~ ~ Go ON GN red 0 0 ~ ~ 0 ~ via ~ ~ O ~ ~ 00 us vet ~o ~ oo 00 oo v' ~s ~ ~ 0 ~ 00 ~ ~ ~ ~ ~ . - ~ t{N 00 ~ O ~ O ~ ~ 1 1 Q 1-~ oo 1- ~ ~ ~ ~ ~ _ ~ ~ ~ c~ ~ ~ ~ ~ ~ ~ r~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ _4 o O cd G~ O ~ c~ CN ~ O oo X O O O O O O O O O U) cn G~ O G~ O ~ ~ ~ ~ O <} cx ~ oo ~ 0 ~ U-\ ~ 1 1 1 oo 1- ~ oO ~ u~ ~1 ~ G~ 1 1 1 ~ ~ ~ ~ ~ ~ ~ ~ - . ~ ~ ~ ~ ~ ~ - ~= - - tJ~ ._ ~ cd V~ Cd I ~ ~ V~ ax ~ ~ ~ ~ ~ ~ u~ G ~- O ~ ~ 1_ ~ ~ ~ [_ _ ~ ~m~d4 O ~ trY 00 Co - 1,^ ~ ~ ~ =^ ~ 00 r~ O - c~ ~ 1~ I GN ~ 0 00 > ~ ,,\ ~ =) \O ~ ~ ~ O O O O O O O O O [S4 X _1 _4 ~ - ~ ~ ~ ~ ~ O :,~ C~ .- E~ ~ cn 0 o E~ ~ 1 ~1 r~ ~ co V~ ~ ~ oo oo ~ O 00 cd ~ ~ ~ ~ ~ 00 ~ u~ ~ ^ ~ ^ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ G~ O ~ - ~ ~ ~ ~ ~ ~o ~ ~ ~ 1 1 00 0 0 ~1 ~ 0 1- ~1 1 - I_ ~ I_ ~ ~ ~- Cx ~ 0N o 1 1 ~ ~ ~ O 1 1 10 .11 1 1 00 I1 1 1 1 o - - ,~ oo oo oo ~ r~ ~o o V-N o I G~ r~ ~ 0 oo ~^ ~ C ~r o CN I oo ox ~r ~ , oo ~ ~V~ I oo oo ~ o oo oo 1- 1 _, . . . . Cx ~ G~ o o o o ax I V~ o ~r I ~r ~ 0 r~ 0= G~ r~ r~ o - ~ GN ~ ~ '-~ 00 ~ ~ ~O ~ 00 00 0 GN . . . . . . . . . . U~ ~ ~ ~ ~ U^` ~ ~ V~ ~ U~ V~ Xooo ooooo U) I G~ o a~ U~ C~ U~ U~ ~^ o C~ U~ o ~4 C: . . : : . - 4_, o ed ~gS o

OCR for page 88
Analysis of the Sex Ratio Data 95 7.5 where the terminations are distributed by of city, but no effect of maternal or paternal city and parental exposure. exposure. The exposure categories given in Table 7.5 are different from those presented elsewhere in this report, being the categories in use during the preliminary analysis. No attempt has been been made to take into account shielding due to slight differences in the way in which the three different segments of the data were collected. The radiation categories are here defined as follows: (A) Not exposed, i.e., not present in the city at the time of bombing. (B) Exposed, no symptoms, over 2,545 meters from the hypocenter. The actual worth of these data is conjectural. All of the information is anamnestic, and the misreporting of the sex of an infant, particularly those who were stillborn or who died at an early age, is a very real possibility. The preceding paragraphs serve to point up two items of general interest. Firstly, they indi- cate the interpretive difficulties which arise when one begins to select specific cells or groups of cells on which to base comparisons. Secondly, they seine to indicate that while an elegant ge- netic argument can be advanced for expecting changes in the sex ratio consequent to parental TABLE 7.5 THE FREQUENCY OF MALE BIRTHS AMONG INFANTS BORN AFTER APRIL, 1946 BUT PRIOR TO JUNE, 1 948, BY PARENTAL EXPOSURE Fathers Hiroshima Nagasaki Mothers ~A- ~, A AB. CD, E Total AB. CD, ETotal ~ n2,38021481 2,675 2,002292532,347 AN m1,23911241 1,392 1,037142291,208 tp.5206.5234.5062 .5204 .5180.4863.5472.5147 En73840369 1,210 1,072682871,841 B. C] m37722130 628 53136342936 (p.5108.5484.4348 .5190 .4953.5323.4828.5084 En34573127 545 1253348206 D, Ej m2013758 296 752222119 (p.5826.5068.4567 .5431 .6000.6667.4583.5777 ~ n3,463690277 4,430 3,1991,0071884,394 Total] m1,817370129 2,316 1,643527932,263 (p.5247.5362.4657 .5228 .5136.5233.4947.5150 (C) Exposed, no symptoms, 1,845-2,544 me- ters from the hypocenter. (D) Exposed, no symptoms, 1,844 or less me- ters from the hypocenter. (E) Exposed, symptomatic, i.e., reporting epi- lation and/or petechiae and/or gingivitis. It should be noted that the average exposure in categories A, B. and E will be essentially un- changed from those in categories 1, 2, and 5. But the average exposure in categories C and D will be less than those in 3 and 4. Inspection of these data, to the extent that trends are apparent, reveals exposure differences which are diametrically opposed to those seen in Table 7.1. The changes observed with re- spect to both maternal and paternal exposure are contrary to those predicted by the genetic hy- pothesis. The analysis of these data (see Table 7.6) reveals no significant interactions, an effect TABLE 7.6 CHI-SQUARE ANALYSIS OF THE FRE- QUENCY OF MALE BIRTHS AMONG INFANTS BORN AFTER APRIL, 1946 BUT PRIOR TO JUNE, 1948 DF Total 17 Interactions, first order CM ............ CF ............ ME ............ Main effects Cities .......... Mothers Hiroshima .... Nagasaki ..... 2 % 228.999 2 2 4 1 2 2 1.034 0.700 8.848 5.347 1.035 3.559 Sum 4 Fathers Hiroshima 2 4.168 Nagasaki 2 0.616 Sum 4 p <.001 .50-.70 .70-.80 .05-.10 .02-.05 .50-.70 .10-.20 4.594 .30-.50 .10-.20 .70-.80 4.784 .30-.50

OCR for page 88
96 Genetic EJects of Atomic Bombs Chapter VII exposure, the sex ratio, as a variable, leaves much to be desired. A voluminous literature purports to show any number of factors which can alter the sex ratio; adequate explanation for the peculiar variations which occur in the sex ratio due to those factors have not been advanced. In view of the unsatisfactory state in which we find ourselves with regard to the effect of parental irradiation on the sex ratio, data con- tinue to be collected in Japan which are perti- nent to this problem. The present findings fail to confirm unequivocally the apparently signifi- cant eRect of exposure on sex ratio reported in a preliminary note (Neel et al., 1953), although the direction of the observed effect remains the same. Among possible reasons for the disparity the following should be considered: (1) the different (improved) classification of parental exposure employed in the present analysis, (2) the superior statistical techniques which have become available since that preliminary report, and (3) the accumulation of additional data. It is worth noting that the direction of devia- tion among the cells is generally consistent with that to be expected from genetic theory, namely, a reduction in the sex ratio with increasing ma- ternal exposure, and an augmentation of the sex ratio with increasing paternal exposure. 7.5 Summary. No significant association between sex ratio and parental exposure is demonstrable in either the comparison involving non-exposed parents or in the comparison among differing classes of parental exposure. The direction of deviation is, in general, con- sistent with the genetic hypothesis. Note added in proof. Since the previous por- tion of this chapter was written it has been pos- sible to bring certain additional evidence to bear on the question of an altered sex ratio. Briefly, the source of this evidence is as follows: When in 1953 it was decided that the Genetics Program should be terminated for the reasons advanced in Section 2.11, there existed some question as to whether or not the sex ratio had been signifi- canlly altered among the pregnancies occurring to exposed parents. It seemed desirable that addi- tional observations be made with regard to the sex of infants being born to exposed and non- exposed parents in Hiroshima and Nagasaki. Through the cooperation of the municipal au- thorities, during 1954 and 1955 access was ob- tained to the birth records on both live and stillborn infants. From these records the requi- site information to identify the infant as well as his or her parents was copied along with the weight of the child, the sex, and place in sib- ship. Parental exposure was secured from (1) the master file on exposed individuals main- tained by the ABCC, or (2) a home visit if one or both parents were not listed in the master file. In the years 1954 and 1955 there were 22,710 births concerning which, during 1954, 1955, and the early part of 1956, it was possi- ble to obtain the information listed above. These data, which are presented in Table 7.7, were analyzed during the summer of 1956. The series is continuous with that previously reported; no infant is recorded in both sets of data. In Table 7.8 are presented the results of an analysis of the combined data, that is, of the data available for 1948-1955. From the latter table we note that when all exposure cells are considered there is no evidence of heterogeneity between the two sets of data nor evidence for a parental exposure effect. When those infants one or both of whose parents were unexposed are excluded, a somewhat different situation ob- tains. We note a significant difference in the sex ratio between the two sets of data with male births being somewhat more common in 1954-1955. Since none of the interactions in- volving time is significant, there is no evidence that this increase is at the expense of a par- ticular city or parental exposure. The "sum" test is, therefore, a valid measure of the effect of parental exposure on the sex ratio. Neither father's nor mother's exposure can be shown to exert a significant effect. It is particularly interesting to examine the combined data under those assumptions given in Section 7.4, Comparisons 1 and 3, which gave rise to a significant effect of mother's exposure on the sex ratio. If in the combined data father's exposure is ignored, mother's exposure classes 3, 4, and 5 are pooled, and those infants whose mothers are in exposure class 1 are rejected, no significant effect of mother's exposure can be demonstrated (X2 = 1.634, DF = 1, and the probability on a one-tailed test is greater than 0.10~. This is, of course, at variance with the findings for the 1948-1953 data given in Sec- tion 7.4. Stated in terms of regression analysis, if sex ratio is regressed on "mean" maternal exposure using as the estimated doses 8, 60,

OCR for page 88
Analysis of the Sex Ratio Data TABLE 7.7 THE FREQUENCY OF MALE BIRTHS BY PARENTAL EXPOSURE AND CITY 1954 - 1955: UNRELATED PARENTS Hiroshima Fathers ~_ 1 2 3 4-5Total ren6,2007442751507,369 14 ~3,254386156783,874 ~ P.5248.5188.5673.5200.5257 rn1,790435120942,439 24 ~93324261541,290 CP.5212.5563.5083.5745.5289 6571217750905 34 ~358654029492 LO LP.5449.5372.5195.5800.5436 rn374603423491 4-5; ~188351214249 P.5027.5833.3529.6087.5071 n9,0211,36050631711,204 Total: ~4,7337282691755,905 VP.5247.5353.5316.5521.5270 Nagasaki v, Q) o A ., 1234-5Total in5,440754112696,375 1- ~2,81338855353,291 LP.5171.5146.4911.5072.5162 n3,1361,039100804,355 2 131,57956468472,258 LP.5035.5428.6800.5875.5185 n369913016506 3 1320446136269 CP.5528.5055.4333.3750.5316 n218291310270 4-5 IS'1231384148 ~ P.5642.4483.6154.4000.5481 n 9,163 1,913255175 11,506 Totali: 4,719 1,01114492 5,966 LP 5150 .5285.5647.5257 .5185 97

OCR for page 88
98 Genetic Ejects of Atomic Bombs Chapter VII and 200 reps for exposure categories 2, 3, and 4-5 respectively, one obtains a regression co- efficient equal to -0.55 ~ /100 reps for the combined data and-0.81~o/100 reps for the 1948-1953 data alone. The former is not sic- nificantly different from zero even on a one- tailed test whereas the latter is significant at the 5 per cent level if a one-tailed test is used. The net effect of the 195~1955 data is, then, to render more questionable the occurrence of a significant effect of parental exposure on the sex ratio in the Hiroshima-Nagasaki data. TABLE 7.8 CHI-SQUARE ANALYSIS OF THE FRE QUENCY OF MALE BIRTHS DURING THE YEARS 1948 - 1955 BY TIME OF BIRTH, CITY, AND PA RENTAL EXPOSURE: UNRELATED PARENTS. TWO INTERVALS OF TIME ARE RECOGNIZED, NAMELY, 1948 - 1953 AND 1954 - 1955. All exposure cells Source DF X2 P Total 63 60.900 Interactions, First Order TC ................. TM ................ TF .................. CM ................. CF .................. ME ............ Main Effects Time (T) ..... City (C) .. Mother (M) Father (F) 1 3 3 3 3 9 0.760 4.758 4.845 1.453 1.589 3.310 ....... 1 1.203 1 1.109 3 2.291 3 1.431 TABLE 7.8 Confirmed Excluding Parents with Exposure 1 Source DF x2 Total 35 39.421 Interactions, First Order TC . . . TM TF . . . CM . . CF . . . ME . . ... ... ... . . . .... .. .. Main Effects Time (T) City (C) 1948-1953 1954-1955 Sum ....... Mother (M) 1948-1953 1954-1955 >.50 Sum . . . Father (F) 1948-1953 1954-1955 2 ... 2 4 p >.25 0.009 1.851 0.273 2.330 3.546 4.285 1 1 0.002 1 0.008 . 2 0.010 0.082 2.810 2.892 0.126 0.449 0.575 .90-.95 .30-.50 .80-.90 .30-.50 .10-.20 .30-.50 001-.01 .95-.98 .90-.95 >.99 .95-.98 .20-.30 .50-.70 .90-.95 .70-.80 .30-.50 .10-.20 .10-.20 We are indebted to Dr. Robert Holmes, cur 50-70 rent director of the ABCC, and Dr. Lowell 95_9~3 Woodbury, chief of the Statistics Unit of the ABCC, for their cooperation in connection with 20 30 these later aspects of the sex-ratio program. We 20- 30 also acknowledge with pleasure Mrs. Jean Oku- 50_70 moto's efficient supervision of the collection of .50_.70 these data. .95-.98