Vo]~uM~ XVI JANUAIlY, 1926 No. 2
I-IEIgEDITAI~u PP~EDISPOSITIONS TO DIZYGOTIC TWIN-BII%THS IN NOIgWEGIAN PEASANT FAMILIES.
B~~ KI~ISTINE BONNEVIE AN]) ASLAUQ S VE]~])]gUP, Univ. lust. A~'v.forJz~z., Oslo ~.
(With Tbirteell. r]~ e xt-figm:'es,, Nine Tables and Three ]?la,tes.)
TABLE O.[i' CONTENTS CI{AP, PA G J'~ I. Inbroduetion 125 (ct,) I-Iistorical Notes 125 (b))'laterb~l 132 (a) Indications of heredity . 135 II. Nabure of the hereditary predisposition' of twin.families 136 (a) None- or dizygotie twin-birghs 137 (b) Dizygotie twill-births in relation to fertility 150 Ill. Type of heredity of dizygotie twin-births 155 (c~) Sex-lhak,~ge 159 (b) Dolmnanee 162 (c) Reeessivi~y 164 Atmex: Do dizygotie twin-births a,lw~ys represent~ 6ne ~md the s~me character? 171 IV. Discussion 17& Summary 181 Literature 182 Explanation of Plates . 188
CHAPTER L
INTt~O DUCTION. (a) Hi~to~'ical 3/otes. Oux present knowledge with regard to the occurrence and nature of multiple births in man, has been reached during the last century through investigations following three main lines, statisticat, clinical and 9enea- ~ogicat respectively. (1) The statistical ilxvestigations o~ multiple births are chiefly based on the great statistics of Veit (1855) for Prussia and of Wapp~us I1859) for a whole series of European countries, each of these statistics covering many millions of births. The gener~l frequency of mtdtiple--or especially of twin-births was here at once estab]ished as amounting to about 1 in 80 1 The new name of :Kristlanis,. Jotu,n. of ~en. xvI 9 126 Hereditc~ry PredisjJoeit, io~8 ~o Dizyyo~ie Tzv'i~z-bi~"~hs births, representing a percentage of the total number of births varying between 0.93 and 1.~. The sexual ratio for twins was shown to be very nearly the same as for single births. A large mlmber of later statistics from various countries or districts, have served to suppot~ these resN~s of Veit and Wapp~us. An interesting investigation was made by Duncau (1865 a,), regarding twin-births in relation to the age of the mothers. As a result he was able (p. 772), "to s~aCe the law tha~ from the earliest childbearing period till the age of 40 is reached, that is, till a period when fecundity has become extraordinarily diminished, the fertility of mothers in twins graduNly increases." I:llc also raised the question about a causal connection between twin-births and the mmaber of pregnancies, a question which in a following paper (1865 b), was answered by him in the affirmative, "the increased frequency of twin-bearing as the number of the pregnancy increases" being here considered "as a law of the production of twins" (p. 929). The relation of twin-births, finally, go the general fertility of the twin-mothers was left by him as a question to which the material at his disposal could give no definite answer. Upon this point his work was continued by Puech (187~), who after a very thorough statistical review of the occm'rence of mtfltiple births in the various European cotmtries states the following restflts (p. 239): "...la fgconditg et l'aptitude aux grossesses mNtiples sont deux carat- tares contingents, .variant d'individu s individu, de contrge ~ contrge, et de peuple ~ peuple, prgsentant l'tme et l'au~re des oscillations en parfaite concordance, mais grant subordonnges l'une k l'autre. En dgfinitive, le degrg de fgconditg des femmes est la loi, d'aprgs laquelle se repartissent les grossesses mNtiples. Quant aux autres agents, ils n'ont d'effet que par leur mode d'action sur celle-d." The same question was investigated and. discussed also by Prinzing (1907), who supports the view of local differences with regard to twin- production without, however, considering the fertility of ~hc race as a sufficient explanation. With regard to t~'iplets statistical investigations corresponding to ~hose above mentioned have been made by lVlirabeau (189~) and by Saniter (1901), while the occm'rence of ~z~ad~'uples was made the subject of investigation by Bra~sfir6m (191@ (2) Parallel ~o and amplifying these statistical investigations, clinical evidence with regard to multiple births has been equally developed especially through the investigations of Spaeth (1860), Ahlfeld (1875), Strassmann (1889-1908), P~umpe (1891). K. BoNN~wm a~D A. Syrup,praise 127 Of special interest are here the results reached witch regard be the placenta and the secandines of these births, to the effect that the number of chorions should be considered decisive for the nature of the mtfltiple birth. Two originally separate placentas may often fuse to one, and with very few exceptions each one of the embryos, whether they develop from one or from ~;wo or nmre eggs, are covered by a separate grunion; but all investigators unanimously maintain that the number of eggs developing may be determined by the presence of a corresponding nmnber of chorions. A furl, her result, equMly beyond dispute is that twins (or triplets, etc.) contained within one and the same chorion, that is monozygo~ic twins, are found to be always of the same sex, while in the dizygotie twin-births, pairs of di:fferent se~ are also found. lgumpe (1891), iu his statistics upon clinically investigated material, has found that dizygotic twins are born (p. 346) "vorwiegend yon lVliittern im vorgeschritteuen Mter (tiber 25 aahre)," whiie mono- zygotic twins are fotmd to be "in jedem Geschlechtsalger gleieh oft beobachtet." The sexual proportions of roche- and of dizygotic twins separately, stated by Spaebh (1860), Strassmann (1889), Rumpe (1891), Br6m (1891), Quenzel (1898), Sehapiro (1912), Neuhauser (1913), l%~[eyer (!916) will, as well as ~;he papers of Weinbe N (1902-1912), be considered in a following chapter 1. With. regard to the etiology of mttlbiple births ~he authors seem go agree in looking upon the monozygotie group as due to some unknown anomaly in the early development of the egg (see especially Stoekard 1921). For the dizygotic twin-births the basis has been sought above all in the ovary of the twin-mother, the mtdtiple bir~h being considered as caused either by a simultaneous rupture of two Graafian follicles, or by the existence in one follicle of more ~han one egg. Investigations of ovaries have proved th~, as an anomaly, a structural basis may be found for a bursting of two or more follicles at a time (t]Tellin 1895, Patellani 1896, Olshausen u. Veil 1899), as well as for two or more eggs being contained in one and the same follicles (Schr~n 1863, v. ~'ranclu6 1898, v. Bmnm 1902, Arnold 1912). In both cases the ovaries have been characterised also by the maintainance up to the adult age of an abtmdanee of primordial follicles, thus giving a picture which is normally found only in young individuals. Such ovaries have been described as
1 The very sigaifican~ book of T. Wedetw~ng on ~qe.r Prolmrtlon and its Varlations i~ Relation to Ante-Natal, Mortalety, Os~o, 1924=, has appeared too ]M;e f~o be considered ia ~;his p~per. 9--2 128 Hereditary Predispositions to JOizygotic Twin-births being similar to those of mtfltiparous animals, and the relatively frequent occurrence in man of mtflbiple births has been taken as a sign, that the uniparity has not yet been fully established (I-Iellln, Strassmann a.o.). Strassmann (1908) draws our attention to the :fact that in eases of an urea'us bicornis which, as is well known, may be fotmd as an anomaly in man, twin-births (Dunning 1889) occur in a propm'tion of 1 in 18 births, instead of the usua,] h'equency of t : 80. While such ovarial predisposition to twin-births generally has been considered as an anomaly, the opinion has on the other hand been mainta,ined by Neugebaur (]913) a~(l quite recently also by Davenport (1920), that a simultaneous ovulation el! two or more eggs in reality is a more common phenomenon than is generally admitted, one of the two eggs being either not fertilised or ~,b some early stage checked ht its development. "Oft genug machte ich die Erfahrung," says Neugebaur (p. 1063): "hgtte ich die Placenta nicht untersucht, so ware die Zwillings- schwangerschaft ganz iibersehen worden." It is of interest in this connection to glance at the corresponding relations in mammals, known to us especially through the papers of Kgppeli (1908), l-lamnmnd (1914) and Kiipfer (1920, 1923). The domestic animals investigated all differ from each other wRh regard to the nmuerical relations between corpora lutes in the ovaries and embryos in the uterus. In the domestic swine the number of Grac~an follicles burst during a heat-period is very high, average 16', the number of embryos varying around half this number (Kiipfer 1920). In sheep, on the other hand, there is a more general correspondence between the two figttres, even if in about 30 per cent. of all cases the number of cotT. lug. may surpass the number of embryos by one, or in exceptional cases even by two or three (Kiipfer 1923). In cattle, finally, the correspondence between the nmnbers of co~'p, h& in the ovaries and embryos in the uterus is, normally, found to be absolute (Lillie 1917, Kiipfer 1920, 1923). The discrepancy between ovary and uterus, so striking in domestic swine, ,mus~ according to the interesting investigations of Kgppeli (1908) be looked upon as an effect of domestication i~self, not only the fertility but the whole sexual cycle of domestic swine being widely different from those of wild swine. In wild swine a heat-period occurs once a year, in the autumn, and the number of yotmg born is in average four, while in the domestic swine heat-periods return every three weeks and the number
1 Aeem~ling to Mill and Donoghue (1914) the number of eggs simuR~meously oxmlaterl may hi Marsupials (Dasyur'us) amount even to 35. K. ]~ONN]~]VIE AND A. ~VEI%DlgUP 129 of ripe follicles in .the ovary in each period has been raised to an average of 16. No wonder, then, that there may occur some discrepancy between ovary and uterus with regard to the speed of their adaptation to this new mode of life. As demonstrated by I-Iammond (1914) atrophic foetuses are found in the uterus of the sow which are (p. 270) "not all of the same size but appear te h.ave dropped out olle by one in a, manner such as wotdd occur in a struggle for nourishment with the increasing btflk of the foeCuses," while other cases seem t,o indicate the*; also "reduced sdtMity" may play a part in this atrophy of ~lm embryos. Both phenomena would :find a natural explanation as fe,~tures following the adaptation to domestic life. Similar, but less conspicuous, differences wi~,h regard to the heat periods have been shown (Kgppeli, Kiipfer) to exist also between the more natural and the highly cultivated races of sheep and cattle, while in goats as well as in dogs and cats the ovarial cycle has remMned that of wild animals. The experiments of Xgppeli have proved, however, that even in goats the ripening of ovarial follicles, and therefore also the occurrence of heat, may be accelerated through over-feeding. An interesting fact, also revealed through the investigations of K~;ppeli, is that the number of primary follicles in the ovary is con- siderably reduced during domestication, while with the acceleration of their ripening the size as well as the weight of the ovary is very con- sider~bly augmented. The transition from wild to the domesticated mode of life seems, therefore, go have a similar effect on the ovary as that demonstrated by yon Hansemann (1912) for the development of an individual to sexual ripeness, and eharacterised by him as "ein Kampf der Eier in den Ovarien." Each ripening @raafian follicle induces upon its surroundings an effect causing atretieal degeneration of a whole series of other follicles. Now, returning to the question raised by Neugebaur and Davenport (see above p. 128) wi~h regard t,o t,be actual number of double ovulations in man, we may, no doubt, compare his sexual cycle with that of domesticated animals. In the mNtiparous domesticated swine with its high degree of over- feeding, the number of young born was found to be considerably lower than that of follicles burst, mid at the same time the size of each litter was fotmd to have a wide range of variation, from 6 to 19 young (Kgppeli 1908). In sheep there was a much more stable relatioa between the number 130 Hereditary Precl'ist~ositr to D~:zygotlc Tzoin-births of ovulations and that of foetuses, as well as in regard to the number of young born, the sheep thus standing between the multiparous and uni- parous animals in so far as it displays a very high percentage (24. per cent. [marshall 190'.t], 36-60 per cent. [Bell 1904-, 19].2]) of twin- or triplet- births. The number of ovulations was found to surpass that o~ embryos in about 30 per cent. of all cases, while in the great mKjority of cases these numbers agreed. In cattle the ,~nipar~l:q as well as t~e agreemen~ between ovulations and embryos were sJ~own to be very stable, twin-birtJas occurring (Jones and ]%ouse 1920) in a percentage of 0"r only, and scarcely any deviation having been found betwee~ the numbers of owtlations and embryos. In horses ~he percentage of twin-birl~hs will probably prove to be similar to that of cattle, although the groups of material so far investigated (see 17$obertson 1913) are too small to give reliable results. With regard to un@arity man should, with its 1.2-14 per cent. twin-births, range near the cattle. If we may judge, therefore, from ~he stability here displayed in the correspondence between ovtflations and embryos, it seems probable that we shah find such correspondence also in man. Such conclusion agrees indeed very well with a long series of investigations of the ovaries of pregnant women. (3) Clinical investigators have in their commtmications given many examples of an individual, as well as of a hereditary disposition ~o multiple births, while other authors, especially Weinberg (1902-12), ha#e treated the question as to heredity by an ingenious use of statistical material. But for a proper treatment, the application of genealogical investigation is m~nifestly reqmred. Among the authors who have made genealogical study the main object of their work may be mentioned in the first place Goehlert (1897), v. Speyr (1894), Rosenfeld (i903), Oliver (1912), Jordan (1914) and Meyer (1916). Some of the pedigrees given by ~hese autJ~ors comprise a groat ~umber of individuals, belonging to a relatively large number of generations. This is the case especially with the pedigrees of princely families presented by Goehlert and by Neyer. But these pedigrees show, on the other hand, the very serious disadvantage that very often the female lines of the families have not been worked out. For hereditary questions, and especially for questions so complicated as that of the heredity o~ twin-bilChs this deficiency will make the pedigrees more or less useless. As a matter of fact ~o conclusion has been reached as.yet with regard to the type of heredity of twin-births. A review of the most recent titera- If. ]30NNEW~ A~O A. Svm~m~uP 131 ture shows great disagreement to exist in tile opinions, and even in the evidence, wi~h regard to dominance or recessiveness, as well as to the question of the responsibility of each of the twin-parents in bhe heredity. A consideration of the heredity of hvin- (or multiple-) births will, however, at once reveal a series of difficulties which explain why no definite solution has yet been reached. Among such di:ff[oulties should in die first place be mentioned: (1) The double nature of twin-births, the mono- and dizygotic, representing gene~ically absolutely dit/erent phenomena, while pheno- typically a distinction between both groups may in many cases prove cluite impossible. (2) The fact tha~, bereclitary disposition for twin-births will not, prove its e~stenee until in the next genez'adon. All women wiChout, or with only few, descendants shm~ld, therefore, be excluded from our con- siderations. (3) The fact, further, that so far as dizygotie twin-births are con- cerned the phenotypical appearance of an hereditary disposition must, according'to its nature, be se:c-limited. No man can, even if carrying the .~win-hereclity as a h0mozygote, directly prove the emstence of such disposition inthe F 1 generation. In the following generations also the inheritance may, if condttcted through males only, remain imperceptible to the investigator. The dittlcttlties hez'e mentioned will, to anyone familiar with genetics, prove very.serious--so serious indeed, that absolutely definite restdts as to the type of heredity may be tmattainable. The many'interesting questions connected with hereditary disposition to multiple births play, however, a considerable part in scientific discussion, and any contribution may be of value. The authors of this paper have, therefore, made use of the occasion offered by the natural conditions of our country--with its more or less isolated valleys--for making an investigation on the occurrence and heredity of twin-births. The populatlon of such a valley may ol%en genedca!Iy be considered as forming one great family; by ms/ring use of church~registers and archives if, has been possible, therefore, to collect a material genealogically accurate and at the same time large enough to form the basis of statistical treatment. If after all, our results will not be found to solve de~zitely the question, we believe this to be due to the nature of the problems before us more than t~ the material or to the methods used. ] 32/tereditc~ry 2~recllspositions to Dizygotic ~',win-births
(5) Mc~tericd. During earlier investigations on Norwegian peasant families con- ducted by the "Institut for Arvelighetsforskning" of the University of Oslo, we have been struck by the fact stated also in other countries (Prinzing 1907, Puech ]87~), that in the families of one parish twin- birbhs may occur considerably more frequently than in those of other valleys, although this d ifl!erence does not necessarily make ills appearance in the statisticM picture of the parish fallen as a whole. Thus, in one great family from Sogn, in which all births had been noted for 73 marriages, two marriages only (2.74, per cent.) were found with twin- births. Another family from Drangedal wii)h 310 marriages investigal~ed had no twin-birth at all. A third lane family, or rather a series of genetically intertangled families from Ringebu, with 380 fully in- vestigated marriages, proved in comparison with the other two to represent a twin-family, showing twins (or triplets) in no less than 71 (I9.5 per cent.) of all marriages. This family from Ringebu was taken as a starting-point of our investigations upon the heredity of a predisposition for twin-production, and a preliminary account (Bonnevie 1919) has been given of the main results so far reached. Later on the work has been continued through an investigation also of a series of families living wil)hin ~eldal and lgennebu, partly also in Orkedalen, these three parishes Forming together the valley of the river Orkla, which falls into the Trondhjemsfjord. The frequent occurrence within these families of twin- (or triplet-) births had been pointed out by'Dr StSren (1901), the clistrict-physician of Neldal, who in various ways has rendered good services to the knowledge of the poptflation of his district. As in l~ingebu, our investigations in ~{eldalen and l~ennebu have been made partly in the field, dmqng the summers of 1919 and 1920, much information having been obtained through public officers and from represent,'~f,ives of the families investigal~ed. Every information, thus obtained has, however, been controlled through the church-registers, and by :far the mos~ import~mt p~rt of our pedigrees are b~sed upon the church-registers only, or upon the "Records of division of inherited property," both going back to the second half of ~he seventeenth century. In our pedigrees, therefore, the descent of twin-pairs has been traced back to the time mentioned, while sideSranches of the same families were also worked out. l~eferring to the oldest ancestors, ~he families have been grouped under separate headings (in Neldal faro. 1-21, in (~.,//~,, 0
~ ,~'q,,~
+ . ? ' s~ I~ ...... ~ .....
d . ~ L ;" ..
' "?k'\ ~{, ?'q
_ J \ k :~ Yig. i. l~Io,p ~howing "~he long ~nd n~,rrow va.iley inelucl,ing Dhe 12a.lJ~he~ of lgen~eh~, iKeldal a.nd Orkd~l, with ll8~mes of f~l,ills whore impo~llD twin-f~milies h~ve been, o~' are s~fll, liviug. 134 tIereditm"y P~'edis#ositions to D[zygotic Twin-births 1%ennebu faro. 1-1~). But it is indeed very diffic~flt to tell how much belongs to one family and how much to another. One gets the impression that bhe native popMation of the whole valley might, if the church- registers had reached further back, have been acknowledged as descending from a relatively very small number of families, once immigrated into the valley. The pedigrees published in this paper all represent more or less restricted fragments of the full family records, which are preserved in the "Institut for Arvelightesforskning," Krisliania. It wend, of course, be of interest to compare our resMts Mso with those of other investigators and from other countries. And, indeed, in the very extens?ve pedigrees published by Lundberg (1918)on a Swedish family from Blekinge we have before us a material which, with regard to twin-births, may on sgveral points be used for a comparison, and which may st~,tistieally be added to the resNts from our Norwegian family-material. The family-material used in the following discussion consists,
Mo.~c~ ~,,~,~2~,3.
Po~,.,~t2.o 90~ 35~4 I (~77 2~733 107 52. I.~ 90~ 5877 12Z 0 .L~.
1880 i &.,.,..-,,~ .r..<,.LoA.. 1203~- 3S9 107 *t~,& 1259Y0 18~-8 I b'37
TABn~ i. Survey of m~teriM invesllg~%ted. therefore, of elaborate pedigrees from three Norwegian parishes: .IVfeldal, l~ennebu, l~ingebu, use having been made also of the Swedish Blekinge family, described by Lundberg. Besides these large pedigrees we have also a series of small families from the same parishes, which probably but not with absolute certainty are related to the others. As will be seen from Table I the whole family-material comprises thus 1%034 births with a sum botal of 466 twin-births, shill-born twins here included. K. BONN~r ASP A. ~w~J~Dgvp 135 Besides the families from 'the districts mentioned, an isolated case of multiple twin-production from Elverum, one woman having no less than eight twin-births, will also be considered. Several questions, as for instance the relations between mono- and dizygotic twin-births, need i~or their solution a comparison with the statistics of other populations. We have obtained such matmial from two sources, matdng use, first by the ldnd permission of the Director of the "Women's I-Iospital," Kris~iania, Professor Kr. Brandt, to in- vestigate the registers of this hospitM for a century with 537 twin- births. We obtained also leave from the "Central ONce of S~atisties" for a special research within the Norwegian population with regard to t~win-birbhs for each of the bwo years 1916 and 1917, including 1848 twin- births, living or dead. We are greatly indebted to the Directors (ff these two institutions for their assistance. (c) Ndications of heredity. Indications of heredity, as playing aparg in the production of twins, occur so often in scientific literature as well as in general experienc% theot no special argument on that score is needed. A few figiLres may sutfice. As will be seen from Table V, line 9, col. 1 (p. 59), mtdtiple births occurred within the Norwegian populaLion in ~he two years 1916-17 in a percentage df !'46, which may be considered a normal rate of twin-births in Norway. The next colts:an (2) of the same line proves the percentage of twim- birbhS:within Neldal and Rennebu for the period t750-1895 to be even lower than the normal percentage of Norway, via. 1-2~ per cent. But, a~er all, the rate of twin-births, within each of the great Norwegian f~milies investigated, from Neldal, l%ennebu and t~ingebu (Col. 3-5), is raised to a percen|~age of 2.85-3.91. The whole Norwegian family- material taken t, ogether (col. 6) shows, therefore, a percentage of twin- births of 3.25, more than twice that of the Norwegian poptxlation. In all the families invest, igated a thorough analysis has proved the grea) majority of twin-births to be accumMated upon certain genetical lines, while other great branches of the families may show no .single multiple birth. A few such twin-lines extracted from our ftfll pedigrees as well as one extracted from ehe pedigrees of Ltmdborg are shown in Pls. XV--XVIL Calculation from such lines (col. 7) shows a further rise of the twin-birth rage up to 8-23 per cent. Such a gradual rise of the percentage of twin-births, when passing from the whole popMation to special families, and within these to more 136 Hereditary Predispositions to Dizygotic Twin-births isolated lines, proves without doubt ~he hereditary nature of multiple blahs. That tendency to twin-births does not exist within all families is on the other hand demonstrated by the fact, already mentioned, that in two other great families of Norwegian peasants, from Sogn and from Drangedal with 313 and 800 births respectively, twin-births occurred in the first family in two marriages only, or in 0.6 per cent. of all births, while in the other no single twin-birth, has been noted. The figures here quoted support the view maintained by earlier attthors that the e~istenee in man of a hereditary 2redis2osition for t~vin-births 1 is obvious. Diffictflties are however, as already mentioned, raised from the very nature of the twin-births, these representing two groups according to whether the twins are monozygotie, or dizygotie. The nature of these two groups of twin-births is entirely different, while the two kinds of twins themselves are not always easily distingnishablC. The first, and perhaps the greatest, dii~culty to be overcome in an investigation upon the heredity of twin-births is there:fore, that of determining the true nature of the twin-material before us. The importance of such distinction is evident, a conclusion drawn with regard to the heredity of one kind of twin-births having no value wh.atever for our rmderstancgng of the occurrence of the other. One of them may a ~riori be supposed to be hereditary and the other perhaps not, or they may both prove to be caused by some hereditary tendency; but if so, their type of heredity may again be entirely different. Nany of the contradictions found in literature with regard to the heredity of twin-births will certainly prove to be due to this confusion.
OKAPTEI~ II.
I~ATUKE 0F TKE I-~EI~EDITAP~Y FI~ZDIS]~0SITION OF TWIIq-9~Ai~[ILIZS. l~efore entering upon a study of the type of heredity we must deride, if possible, the natm'e of the ~win.-bir~hs characteristic of and. heredil;ary in tl~e family-material before us. We require a distinction between mono- c~nd dizygotic twin-births; and it will further be of interest to investigate the theory of Puech (]87Q as to a 2ossib~e causat connection between gene~'al fertility and frecluency of twin-births. 1 Statistical investigations of a similar predisposition in sheep have been undertaken wi[h positive results by ~ietz ~nd ~oberts (1915) and Wriedt (1917)--and less conclusively for horses, by I~ober~son (1913) and by Stroever (1917). Taking his st~rting-poi~ in the phenotypic~l simil~fity Fisher (1919) has made win ~tempt to eliminate the genetieal difference between mono- and dizygotic twins. K. BON~EVIE A~D A. SvEgD~ur 137
(cb) Mono- o~' dizyyotic twin-births. No clinical observations are at our disposal with regard to the true nature of the twin-bilChs of the peasant families before us. The question whether they should be considered as roche- or dizygotic must, tl~erefore, be solved tlhrough a statistical study only, based on the fact tha~ in twin-births the sexual-proportion has repeatedly been stated to be somewhat the same as in single births. Upon this basis Weinberg (1902) has introduced his "Diffel:ence- method," according to which the number of mono- and dizygotic twin- births within a given population can be approximately decided. This "difference--method" is founded upon the fact that monozyyot@ twin-pairs are always of t;he same sex (cYd or ~o), while among cEzyyot,ie twins the sexual-proportion is supposed ~o be the same as anaong sist~ers and brothers in generM, that is appro~mately an equal number of males and females, the distribution of whom on the twin-pairs should expeotedly be that of 1 4c~ + 2 d9 + 1 99. If therefore in a given population of twin-pairs to the number of bisexual twin-pairs is added a similar number o:[ the unisexual ones, we should get the whole number of dizygotic twins, while the remainder of tmisexual twin-pairs would represent the monozygotic ones. The,same method has, independently been used also by Cobb (1915) while Dunoker (1915), without knowledge of Weinberg's method, has intr0dueed a more elaborate eMculation in which allowance is made also for the surplus of males expressed in the general sexualoproporgion of man (I06 c? : 100 ?). Such statistical methods can of course, as aelrnowledged also by the authors, give fully reliable restflts only when working with a large materfal; the smalIer the figures, the less exact the results. In order to test the value of both methods we have~ from the literal)ure, collected a number of cases in which information has been given in one and the same material of the distribution of sex as well as of the clinical results wil;h regard to the secundines (see Table II, p. 138). FN1 informations have been given by P~umpe (1891) for 166"twin-pairs as well as by Meyer (1916) for 250 pairs. To these are in Table II added also a number of 71 twin-pairs, for which full informations have been given by Women's I{ospitals in Kris~iania. Other authors, as Spaeth (1860), Br6m (1891) Quenzel (1898), Schapiro (1912), Neuhguser (1913) h.ave given similar information without, however, specifying the numbers of male and female twin-pairs, within the grouo of unise~u~,l oue~, Thi~ 138 ttereditcery Predispositions to Dizygotic Twin-births difference is of no significa~lee for.otu" testing of the "di:ffe:r:ence-method," wh~l.e for making use of' the ca[c~:,lation of Dtmeker l)he exact number of
II
o.- I! I'!
~o
N- ;=iN SN
m ~
'~o~
I o
male and female twin-pMrs woNd be necessary. Some of the attthors mentioned have for the sexual-proportion used a material of twin-pMrs somewhat larger than lhat for which the chorions have been investigated. K. Bo~1~v1~ x~D A. Sv~wul~ 139 The percentage will therefore, more than the absolute numbers of mono- and dizygotio twin-births, serve as a 1)asis of comparison between the ~esults of observation and of calculation. As will be seen from Table II, the percentage of dizygotic twin-births found tllrough the "difference-method" of Weinberg may in the various small pop~flations vary very considerably from that really observed (by S]?aeth 57-6 per cent. dizygotic twin-births instead of 75.2 per cent.; l~{eyer 984 per cent. iins~ead of 84.-4. ]?el: cent., etc.). But in the large matel~al obtained through adding all the :figures given, the resNts of the ,d~:fference-method" appear to be very satisfactory, the calculated ntlmbers of di- and of monozygotic twin-pairs differing little more than 1 per cent. from those observed (79-1. per cent. : 78.0 per cent. and 20"9 per. cent. : 22 per cent.). The material represented in the three first lines of Table II, including specia ! references to the mmabers: of male and female twin-pairs, has been calculated also according to the method of DunokerL It appears, ho~gever, that in this material of 4-87 twin-pairs the di~erenee-method of. Weinberg brings us nearer to the actually observed figures than does the method of Duncker. Even if, t;heoretically, a more exact result should be expected from a met]led of calculation, which takes into Consideration also the sexual-proportion of the popNation investigated, experienc6 has proved that because of the relatively gre~t mortality of twin-emgryos, the sexual-proportion of twin-populations may be fmmd to Vary within a range considerably wider than that of single births. The possible advantage of the method of Dunoker will, therefore, for its documentation need a very large material. According to the above considerations the "difference-method" of Weinberg has been used for all calculations with regard f,o mono- and dizygotic twin-births. The r'esults given in Table ]iI Mso show the obvious fact that the calculated values can oMy be approMmate and, of course, ~he less exact, the smaller the population considered. The objection made by I~Ieyer (1916, p. 328), on accmmt of.250 twin,pairs in which the sexas well as the number of chorions has been stated (of. Table H), against the validity of the difference-method will therefore scarcely ]?rove justified. As first proved bF Duncan (1865), and as demonstrated also in our preliminary report (Bonnevie 1919, Figs. 2-3) upon the l~ingebu family The cMculation of Duncker is based upe~ the formula ~' =l-L-~'~f,, in which ~V repre- sents ~he number of dizygotic ~wln-blrt~hs, f" the number of bisexuM twin-pMrs (c?-~-), while cZis the sexuM-proporgionof l~he~win~population inve.~glgM)ed (leo. ci~. pc. 507-511}. 14-0 tle~"ecEta~'y Precti~2ositio~~s to Dizygotic Twfn-births
~he:re seems to be some correlation between ~he age of the ~win-mother and the mlmber of ~win-bir~hs. Our 1%ingebu investigation has also suppozted the view maintained by l%umpe (1891) that upon ~his poin~ a difference exisbs between roche- and dizygotic twin-births. Here again, therefore, our whole materiM of ~win-births has for analysis been grouped according t,o the age of the t~win-mogher, as will be seen in Table II1, where ghe twin-births are arranged in groups corresponding each (;o five years o:l: age of (;he nmghe:l:, while in Table IIi a-b l;he full analysis (per year) of ~he same ma~eriM Will be found. All curves, of Figs. 2-6, have been constructed ttpo~ series found in Table lIi.
--'39 -- ~r -~ A 1~. .~ 1 B&.AAx, ~G97 ~3228 329~ 3017~
..9i
zV f'q o~o - ~'~ 6 75 IZ~ IZb 7b 55 12 q78 3I,~jY• 0,Dr3 g v~ N ~IZ 7Y 4~1 3I,Y Z9 DL - 7y~ ~ ta-cx~ 7. 112. ~78 31K, 3&Z 150 ~ ~%5-8 3t~ ~ O,l~h
8 ~-~ - .[6~k~ 7 117 IYr 112 9E 29 l 537 ~%38 ~ 0,P.G q e,oJ-r l&.',a.=Iya~ Z~,t, ~78,~, 5t~5,7 3~%1 307~1 93~8 ~,~ l'/3b ~_ Io c~ ~/o o) i o,v, I,ba l,Tz bt I,~ Q,~ 0,57 g . y,J ].~.5:.,~d. 7 ~5" 7~ t7 19 ~ X l~3 tb, G3 +- O,~l~ to I%3 ioo .3~tZ 7b 2~ O 3bI~ 30, b9 ~- 0~309 8o s~r O G7,.78% 108 ~7 5- 380 3t, D~_+ o, az~ Ib c~,X Jo~ ~ = 173~ Z~8~ tq%~ 333,5" Y~Ob,(~ iY ~ ~ o} I O, fi5 0,8~6 1,013 b3Y t,Z~ 2~21 1,3 7 +~ I ~r(+l) tLl~" _+O,~'c,G ~,~ 0 Z0 ll, O~7o
9K5 I00 80 88,~y~1
Lg. o~ bt,.,;..~ .~L~t,6~ 10 3~7 Gs"~i 673 G~I Zg~ Z7 t6~3
176,9 5%% I~l tOO0 ~4 Ut T~,~ (~= IO00) %6 107,7 Z33,9 Z71,9 25~,3 118,1 I1,~" #OO0 %~BL1~ III. Frequency of births, i;win-births, roche- and dizygotic Lwin-]~irtl~s, of the v~u:ious groups of ma%erial invcsUgated, arranged ~ecording to the age of the twin- mothers/The material upon which the series of Ulis fi~ble (except ser. 1 and 23 tM~en directly from the "Official 8tatisUcs") have been based is found in Table ill a-b, pl~n~ed ~t the end of tiffs paper. (pp. 187-188.) The figures ~dded in ( ) in ser. 1S and 20 repl'esen~ a smM1 number of twin-births whose sexes are unknown, but which are included in the figures of set. 15; they have been distributed between the groups of roche- and dizygobic twin-births accordfl]g to ~ahe relative size of the lat~er. K. BONNEV:[E AND A. SV~,I~gRuP ].41 As already mentioned above, our ma~eria,1 includes, besides the f~mily-material, the ~win-births occurring wif,hin a period of a htmdred years in the Women's Hospital of Kristiania, as well as of those of the Twin-mat.erial (of. Table III).
5"00.
80- 6O
400-
80.
64.
~00-
60: q.(]-
200 -
80- /x ~o0- / " // x\k #0- // \\\ ~0 : /,' IDO- / /" "''~ "~ ~0'~ !"// ," k " L," %~ ~ I l'"
Fig. 2. CJu~ves of 8win-birbhs, arranged accordfl~g to ~he age of ~he me#hers, wi~lfin each of ~he three groups of maberial investigated, whole Norwegian popttlation in the two years 1916-17. Each of these ghree groups of our material is represent;ed in one o~ 6he curves o~ Fig'. 2. These curves show very charac~erisbie di~fl!erenees. The mammtun of ~Win-bir~hs in the clinical maberial :falls between 24 ~z~cl 29 years of age ~[otu'n. of C4en, xvr IU 142 tTereditary Predispositions to Dizygotic Twin-births of the twin-mother; in that of the Norwegian poptdagion the greatest number of twin-births is found with mothers 29-34 years old, while in the family-materiaI the maximum nmnber of twins are borne by mothers 34--39 years old. The curve of the whole population being based upon no less than 1736 twin-pairs 1 and from a wholly representative material shmfld here be taken as the norm, in relation to which the characteristics of the two other curves shmfld be considered. The peculiarities of the clinical material are easily mlders~ood when remembering, what has been repeatedly mentioned also by previous authors, that such material will always be of a mm'e or less select nature. In the client61e of a Lying-in Hospital young women will as primiparae always be mm'e numerous proportionally than in the population as a whole--and the absolute number of twin-births will in clinical material be raised also for the reason that dangerous comp~cations occ]~ more often in m~fltiple than in single births. Of more general interest are the peculiarities of the family-material. But before entering into an analysis of this question, it will be necessary to s~udy more thoroughly the occ~u'rence and the nature of the twin- births in the whole Norwegian pop~ation, considering both the relations of twin-births ~o births in general, and the statistical relations between mono- and dizygotic twin-births. The relation between births and twin-biq'ths in Norway is plainly demonstrated in Fig. 3, the two c~rves of which are based upon the series 23 and 24 of Table III, or in other words upon the two series for births (set. 1) and for twin-births (set. 2) both reduced to the same size, viz. n = 1000. :Both series are seen to form approximately symmetrical curves failing steeply down towards the beginning as well as towards ~he end of the fertile age of the mothers. But ab the same Lime the curve of the ~win-birth is seen to be independent of ~he bir~h-c~rve in so far as its maximum is reached at a later age of the moghers, between 29 and 84 years instead of ~he 24-29 years ms,mum of ~he bir~h-rate. In Table Ill the average age of the mothers is shown to be for the birth~curve (set. 1) 31.03 :~ 0.004., while for the bwin-bir~h cTlrve (set. 2) it is 32.28 J: 0.145. Such independence between the two curves proves, in fldJ agreement wi~h the view maintained by Duncan (1865), ~he existence in ~he pop~flation of a p~edisposition fo~" twin-bi~'ths varying with the age of 1 Of ghe to~al 1878 twin-pah's (cf. Table I), 112 pMrs for which the mogher's gge was ~mknown~have been excluded, K. BONNEVIE AND A. ~VEEDIgUP 143 the mothers, while in its remflts dependent also upon bhe birth-ra~e characteristic of each age-group. In order ~o get a clear picture of these relations it will be necessary go consider not only the absolute numbers of maltiple birfl~s wJhich must necessarily be strongly influenced by the birth-rate, but above all the ~'elation between births and t,win-births within each age-group of the o~ 3o0 4s~t~ 80 ~
(~O-
frO-
20
~00 i I ~ \\ 80 i Ii "~ \ \\
40 \ \ \ 20 \
I00- \\
80" \ 60
~0 //
~0- / I
Fig. 3. Curves of births aad twfl:&irths in relMio: to gho ~go of the mothers, within t~he population of Norway 1916-17. For faciL[ga6iug the comlo~riaml the frequencies o]~ both series have been reduced to n= 1000. mothers, that is the number of twin-births expressed in percentage of the total number of birttts. A c~rve of tkis kind is given in Fig. 4 b, based on series 3 of Table III, while in Fig. 4 a the absolute numbers of twin-births are demonstrated. The prec]_isposition for twin-bii4hs is here seen to be very slight among yoking me%hers but to increase qu,ite regulc~rZy wgt7t 10--2 144 H~reditary Predi,~ositions to Dizyyotie Twin-bh'ths their grozving aye, ~he curve reaching i~s maximum in mo~hers, 34 and 39 ye~:s old; ~f~er this ~ime ~he disposition for ~bwin-bir~hs seems ~o be slowly diminishing. Norway 1916-17 (of. Table III a, set. 2, 4, 6).
~0o
{,o
hoe ~.
t~o- f / ~0. / / 300 '
80 /
4-0
/! :too~o~ l/ [i/i/] ! ! \ \ ] I x\ ~oo /" / \ \ \ 80 ' : r / 6o /,,/. / lto 1,I / ~oo I//) f~
Fig. 4 a. Analysis of the occurrence of f;win-births (including still-born) in ~;he Norwegian populal;ion: I;he frequencies of monozygo~io and of dizygotio ~wimbir~;hs as well as of their sum to~;al within the various age-periods of ~he mo~;hers. In l~he bwo lower curves of }~ig. 4 a-b are demonsbrat~ed ghe resull~s of an analysis of ~he bwin-bizths, according be bhe "difference-me,hod" of Weinberg, the material having been divided into bwo groups representing ~he roche- and d~zyyotic ~win-bil~hs respectively. K, BONNEVIE AN]) A. SV]~]t~DI~IP ]45
Norway 1916-17 (off Table III).
...... D,~.- ~j,~.~4~.~. (.~.~. ~ ) 20 "L~' "1;6
~,0 ~ /- jf..~ \ o,6-
/ r
Fig. l.b. The s~me freclueno~es as in Fig. 4 a expressed in per oon~. of ~he aumber of births within oaoh age-period,
Norway ]916-17: F~rst births (Table III, sot. 3). Twimbir~hs in % of bil%hs.
3,0- 8- ~1761766-
g- 2,0-
.~- 2= 1,0- 8-
--~- j - ~ , , , ".~ , ,
Fig. 4:c. Frequency o~ "first birf~hs" ~r_~angeclin t~he same way ~s in Fig. ~b. 146 tIeredita~"g I)redisl)ositio~s t,o D#~gyotie ~%vin-birtAs
The absolute numbers of each group shown in Fig. 4 c~ are found in series 4 and 6 of Table III. A characteristic difference is seen to exist between them, the curve of inonozygotic twin-births reaching its maxi- mmn already in mothers 2~-29 years old, with an average of 3145 -_t: 0.2~3, while that of dizygotic bi~r is constantly rising up to an age of 34-39 (av. 32.6 • 0.16@ The sum totals of mono- and dizygotic twin-births make 27.53 per cent. and 72.47 per cent. of the whole number of twin- pairs. A very interesting supplement to this picture is given in the two lower curves of Fig. 4-b proving the ~'ey,a,letr i~crease cmcl fcdl of the disposition for twin-births Co be due to the d'i:~ygotio twin-bi~"&s only, this curve (Table III, set. 7) following very closely the course of the main curve. The tendency for ~wnozygotio twin-births (Table III, ser. 5) seems, on the other hand, to be i~dependent of the age of the mother, the curve here taldng an approximately horizontal course. Irregularities at the extreme ends of these curves will be of no real importance, considering the rapid decrease in the absolute number not only of multiple births but of bilnbhs in general by mothers more than 40 years old. This result supports the conclusion drawn by Rumpe (1891) from his material of some 160 clinically investigated twin-bi~$hs (s. 346) "d ass zweieiige Zwillinge w~rwiegend yon IXitttern im vorgeschrittenen Alter (tiber 25 Jahren) geb oren werden," and hu'ther What "einefige Zwillinge...in j edem Geschlechtsal~er gleich og beobach~et (werden)." The dependence upon the age of the mother here demonstrated for the predisposition of twin-births might of course mean either a direct causality existing between age and twin-births, or an ind#ect one if not the age itself but the number of previous births by the same mother should prove to be decisive for the development of such predisposition. This question has, as already mentioned, been raised byDuncan (1865 a-b) who, after some hesitation in his first paper, finishes by declaring in the second (p. 929) thaW: '~ increased frequency of twin-bearing as tlm nmnber of the pregnancy increases may...be now regarded as a law of the production of twins." A definite solution of this question will, of course, meet with the ditliculty that a high frequency will never be met with except in older women; the parts played by each one of these causes will, therefore, not easily be decided. The only way of eliminating one of the two factors will be that of considering the primiparae as an isolated group, women of all ages (within the period of fertility) here being met with. Table IV gives our material for investigating the twin-births among K. BONNEVIE AND A. SVEI~DRUP 147 1st pregnancies in relation to 'the age of the mothers, illustrate([ for the whole population of Norway also in the curve of Fig. ~. c. In this curve (of. series 3 of Table IV) the percentage of twln-births among 1st pregnancies is seen to augment with the age of the mothers, following very near the same line as the curve of the general predis-
B,b~b~ I Iq~7 1078% 8(o76 3ff91 IZS(, ~.% 2.{~7t
PHG 17 " ~%,~J: I 3 O,~M 0,':/9 1,3q I,~}8 1,7q I,~.'~ M.tJ,l...~,.~,.~ Lt Z i B 17 II i~ 5 5q
T~]~ IV. Frequency of "flrs~ bir~h~" h~ Hm Norwegian population, ~nd in ~he f~mi]y. m~teri~], ~m'~nged ~eoording to the ~ge of the mothers. position for twin-births (Fig. ~ b). The slight difference between both ct~rves Mth regard to mothers 34-39 years old wi]l not surprise when taking into consideration the low number of primiparae among mothers of'this age. It see~us, therefore, absolutely certain that the age of the mothers plays an important and direct part in the development of a predis- position for t~cin-blrths. This fact does not, it is true, exclude the Possibility that a high number of pregnancies may also be of some significance. The similarity of the two curves of Fig. 4 b and c, proves, however, that not much room is left for any influence other than that of the age of the mother. The results here reached with regard to the predisposition for twin- births existing in ~he Norwegian population, and demonstrated in the curves of Fig. ~ a-c, may serve as a basis of comparison when turning, now, to a consideration of our family-material. Fig. 5 a, the curves of Which ~re based upon series 15, 18 and 20 of Table III, gives an illustration of the absolute numbers of twin-births in the family-material as wel] as the numbers of dizygotic and mono- zygotic twin-births, calculated according to the "diEerence-method." The most characteristic features of these curves wilI be seen first in the relatively very high age (34-39 years) of the mothers giving the maximal fr@quency of twin-births. Secondly the curve of dizygotic twins is I~S Hereditary Predi~2)ositions to Dizycyotic 1'wi~z-births extraordinary with regard to its height, as well as with regard to its shape. It covers in 'ghe family-material no less than 88.95 per cent. of the whole field, while 72.1-7 was the percentage of dizygo6ic twin-births in the whole population; and the shape of the dizygotic curve of the family-material is, especially for the older women, very near the same as bhat of all twin-births taken together, so near, indeed, that we are justified in saying that the dizygotic twin-births are here alone responsible for the deviation from the typical curve (Fig. 1- c~) of the whole pop~flation.
Fanfily.m~eriM (cfi TM)Ie III).
qo
80 70 .\ G0 ~'0 ~0 3o ~0 I0
Fig. 5 c~. Twi~-bh%hs(8~ill-bora inoluded) in bhe "~amfly-materiM"~rr~nged as ~ Fig. ~ c~.
These characteristic features of the family-material are even more clearly demonstrated in tim curves of Fig. 5 5, s~owing the number of twin-births within each q uinquemfiM age-period of l~he mothers (Table III, ser. 15) expressed in percentage of the "normal birth-rate" of the Norwegian population, as found in series 1 of Table III. In order to facilitate a comparison with our results from the whole population, the series of twin-births in the family-materiM has before finding the per- K. BON~EvIE ~ND A, SvEgDgm) 1.49 centages been calculated for ~ = 1736, that is for a total number of twin-births like that of the Norwegian population (Table III, ser. 16). The predisposition curves of the family-material, thus shown in Fig. 5 b, emphasize once more, and wiOh all desirable clearness, the heavy preponderance of dizygotic twin-births. We find here a predisposition for twin-births, wN.ch already in young mothers largely surpasses that of the whole popNation and which rises rapidly, lrom an a.ge of abou~ 2d years, to a very conspicuous maximum reached, as in the ctu've of Fig, ~ b, ~t the age of 3~-39 but lasting in the family-materiM Mso during
Family-mager~M (el. T~b[e III).
O/o o} ~ ~ ~ 4,.~s (~.~,~.iv )
4,~
,~, ~.J "\
ar I , : ...... i ~"" " 't
~ig, 5 b. TWill-blr~hs in ~h6 ' F~mily-m~geriM : Firsg birfihs. Twin-birbhs in ~o of birbhs (T~ble IV, set. 5). Wok I 1,0 0,8 0,#- fJ O,Z- f Fig. 5 c. Twirt.bh'~hs ~mong "first.bh~hs" in ~he "f~mily.m~erial," expressed in per cent. of births. We feel, after the above demonstration, fully justi/ied in considering an inc~ea,&d predisposition for dizygotie twin-b#ths to be ~he character inheritable in our twi~-families. Such 1)redisl0ositio~ must, however, as generally maintained also by earlier investigators, be due to some characteristic feature in the ovary of the twin-mother causing the existence, simMtaneously, in her uterus of two (or more) fertilised eggs, while, so far as dizygotic twin-births are concerned, the father of the twins will be of no significance. The question about monozygotie twin-births, their nature and their possible heredity, will in what follows be left out of consideration, our whole attention being turned towards the twin-mothers as ~'epresentatives of an hereditary p'redispo~'ition for dizygotic births. (b) Dizygotie twin-birth~" in relation to fertility. As already mentioned above, Puech (187~) strongly emphasized ]]is view of a causal connection existing between multiple births and the general fertility of an individual woman, of a race or of a certain disbrict. A priori, such causal connection wotfld for dizygotie twin-births, seem very probable. A predisposition to double ovtflations which has i~s structflraI basis in the existence in the adult ovary of an unusual number K. Bo~r~zv,z a~) A. 8wm)ecuP 151 ~. VI.~.,L,,.,[. mdA,~,L R.~-,~.A,~ R.;~pq,.,.,. V.,~,~i.~. T~.,.;.,~- R ~-,.,w$ ,.,- I,%1 I -IZ 1-22 ~,v,.,,wv ' )116-1917 17,','0 -18~1} Y~o,~a./v &t.~.r,..be Lr [ I ~ 781t'~t 131YY 36~1b" 28~t0 qO00 I0q8~ 1618 S.e~. fu',-~.,.,.,-,-,. 2. }Z78~tt 658q 3S60 ~8~Z 3801 Iolqb I60Y ----- ~ O* .3 6~'7 qS 3311 1781 I't~% 19~b Yl69 76l I,o.~. ~ ~" y~ oj. ~ q 51,,~. 5o,z~ 50,03 YR~t~ ~l,,tt, So,TI 47,,~] Q 5 62099 .3~.78 1779 I~00 lSff~ YO~)t 8~tq g,~ ~ o~ L (9 ztS, ff7 49,7Y :~tq,q7 ~9,1~h z/8~ytt ., q g, lL~ ~2~.ff9 BGv~ 7 12.5978 I~q9@ 35~lq" 2733 3878 lOftY lhq~ $ .t.t..~.~ ,L~~[- 8 1848 "x) 161 I01 107 12Z ,330 1~5 I L,,a. ~'o ~l 7 q 1,46 l,i't 2,85" 5~91 3~lff 3~9..F 8,23 ' ~5.~... ,w, Jv.*.o,,,,, 10 112. 3 13 2. 6, gl 7 ..... ,~.~..-D-,~.e.,-,~- II 1736 158 88 105- 116 309 116 9 9 ~] ~o~ 5~ ~z ~z ~6 qo ~z . . I 5~17 OL- a~ 16 6%'1 56 33 V3 ~t6 13z ~'Z [,-&.,bh.~ ...... Plo-r ~, 18 578 (~~ ~6 (') ~%01 § I 9t~ (~,t) qS- O) I~,0) ~"~. 7. ~ Ii I't 1.7,s~ ~,%, ~,~o + 025 ~.0,~.~ I#,~G I0,~ C~, % oj 7 to 0,:~o O#G 0,7~ + 0,o~ o,Gr 0,~7 0,87 TA~SE V. S~a~m~mM da~a r%~rdmg h~hs (hne~ I-7) and ~wm&~%hs (lines 8--23) m ~he whole Norwegian population (col. I), in ~he whole vaUey of ~eklM and Rennebu (col 2), in each of ~ho ~wmJam hoe of ~'MdM, ~mmebu ~11d l~mg~bu (eels. 3-5), in ~he whole family-man,riM (ool. 6) ~nd in spedM ~win-lines (ool. 7). The following f~el,s should be noted: The numbers of births (line 7) have been found by subtr~mtlon of the sum total of multiple bLrths (llne 8) from that of individuMs born (line 1), remembm~ng that eaoh twiu-birth makes the number of indivtduMs sarpass by 1 (or in e~se of triplets by 2) the number of births. The numbers of mono- (llne 18) and dizygotio (lhle 21) twin-bhChs have been calcu- lated throug k the differenoe-method of Weinberg on the numbers Of mono- and bisexual ones (lines 14 and 16). The twin-births, whose sex was mdmowu (line 10) h~ve been added in ( ) ~o the groups of mono- and dizygotio twin-births (line 18 ~nd 21), dis~Sbuted between ~hese groups according ~o theh' relative size. Those figures have been inohtded in tim lmn~bers of mono. and dizygotio twin-births when eMmda~ing ~heir persentage of ~he total number o~ births (lines 20 and 23). 152 Hereditary P,redi~2)ositions to Dizygotic Twin,.bi~"ths r ~1: 4-l d A < < rn ,~i -- r ! !1 il N K. BONNEVIF~ aND A. Sv]~D~UP 153 of follicles (see above, p. 129), might also prove its effect in the occurrence of a great number of single births. This supposition is strongly supported, further, through the ~mani~ incus results reached in multiparous mammMs1 that "the frequency of prodltcgion of litters" as well as "the average size of the litter" both increase wi~h the age of the female, at least tti? to a certain point." (.Jones and t~ouse 1920, p. 265.) Of special interest too is ~he evidence from various races of sheep~, these animals forming a link between mtfltiparous and tmiparous oNanisms , in so far as twins and triplets are born in relatively high percentages besides their normally single births. The predisposition to mNtiple births is here known to vary characteristically from one race to an.other and go be hereclieary within special lines. Itere again we find an increase, with the age of the mother, of general fertility as well as of production of twins, and the question obviously arises whether these two phenomena (here referring to clizygotic twin-births only) shoNd not be considered as different expressions of one and the same hereditary character. This question was 2or ottr family-material raised in the preliminary report already (Bonnevie 1919), 380 fl~lly invesbigabed marriages from ~he l~ingebu family being arranged according to the number of births in each of them, and compared with 96 marriages belonging to twin- lines. The average number of children was fmmd to be somewhat higher in tl~e latter (6.72 ~: 0.25), than in the former (6-38 ~: 0.1~:); but the Whole materiM then at ore" disposal was too restricted, and the probable errors of the vaktes fotmd were, therefore, so high as to reduce very considerably the significance of this result. Once again, ~herefore, ~he same question has been investigated, th~s Lime upon 1178 apparently fN1 ~abernities chosen withou~ any selection from ~he wholefamily-matericd of geldal, t~emaebu and l~ingeba (Table VII set. 1), as compared with similar fraternities from special twin- families. These latter have been collected in ~hree differe~ ways, either (set. 2) from whole m~in-fc~migies comprising all their gene~ical lines with or without "twin-heredity," or (set. 3) from "twin-lines" only, like those of the pedigrees on Pie. XV--XVII, or finally (set. 4:) thrmlgh a selection of "twin-fraternities" chosen at random from the whole family-material. In each case some 125 fraternities have been used for investigating the fertility of twin-material. The four groups of fraternities above mentioned have been arranged 1 If~ng (2927, 2924:) ill ra~s; Ellinger (2921) ia swine. "~ tt.iegz ~nd ~obor~s (2925), Wge4g (]927. 2.qg~L J?.=b ~:~ 'lf~: ~ 154 Hereditc~ry Predispositions to Dizygotic Twin-births according to their number of bilChs, twins alway~ being counted as one birth only. As will be seen from Table VI the average fertility in the whole family-qnate~'ial proves to be 6.09 ! 0.065, while in the "twin- families" it is 6.78 • 0.222 and in the "twin-lines 6.76 • 0.176. In both these oases the difference in the fertility between twin-families and the general population is worth noting, even if it can scarcely be called convincing, the differences of the average nmnbers of births amounting ~o, or Co somewhat more ~han, three times their probable error. It should be remembered that the twin-lines are also included in the family- m.a~erial, which may perhaps have helped to raise its fertility. For a direct comparison of the fertility the distribution of fraternities has within each group of ma~eriaI been expressed also in percentage of Fertility @ *st. IS- ~\\ a \ ~. Z x Fig. 6. Fertilit~y, number of births per fr~brni~y, within the whole family-msr as compared wi~h ~ha~a of special ~win.fam/lies (eL T~ble gI, ser. 1" and 2"). their sum totals (Table VI, ser. 1"-4"); the two curves of Fig. 6 con- strutted upon such series give a graphical illustration of the fertility os the whole family-material as compared with that of the twin-families. It is not without interest that the rise of the average fertility found in twin-families as well as in twin-tines does not prove Co be valid for the twin-fraternities. This may of course be Clue to the scantiness of material; but at any rate it proves that even if there may 'be some causal con- uection between high fertility and a predisposition for dizygotic twins, such connection does not exist in form of any linkage between two hereditary factors. If such had been the case the mothers of twin- fraternities ought to represent also the highest fertility. In the following chapter, therefore, the predisposition for dizygotic twin-births may be investigated as an independently hereditable character, its type of heredity being discussed on the basis of the actual occurrence of twin-births in our family-material. K. BONNEVIE AND A. SvJaP~DI~I~P 155 0HABTEI{ III. TYPE OP I-~EIIEDITY OF DIZYGOTIC TWIIg-}311%THS. In the foregoing chapter we found the character hereditary in our :family-material to be that of a pred/isposition to dizggogc twi,n-b#'ths, or in other words to a double ovtflation, and consequently the adn-mother will be the person primarily interesting as carrying the hereditary/!actor. She, however, may have received this factor from any one of her parents or earlier ancestors whether males or females. Therefore, during ~he genealogical preparation of our WOrk we have paid special attention to tracing the ancestry of every single'twin-mother as far back as possible. According to church-registers as well as to "Igecords of division of inherited property" a great ntmaber of families have been reconstructed up to the second half of the 1.7th century. These various families prove to be interrelated so as to form together a comparatively small number of very ex~ensive pedigrees (from Neldal, I~ennebu and I~ingebu) all containing some 9-10 generations and amounting together to some 10,000 indi~4duals 1 (see Table I, p. 42). From such ~pproximately complete pedigrees it has been possible for a considerable number of twin-mo~hers of later generations to trace their whole ancestry 5-6 generations backwards. For twin-mo~hers belonging to older generations the ancestry available is, of course, less complete, while in many cases, also, some branch of the family which has immigrated from other parts of the cmmtry has not been followed htrther back. For each one of some 200 twin-pairs from Meldal and Rennebu tables of descent like that of Fig. 7 have been constructed, special stress having been laid upon tracing the ancestry of l~he twin-mothers. Besides this, the ancestry of about 80 twin-mothers from l~ingebu has been thoroughly investigated. In these tables the ancestry also of the twin-father will often be found when belonging to some of the families occurring on our pedigrees, but special investigations have as a rule not been made as to the ancestry of the twin-father. As will be seen from Fig. 7, each table of descent comprises in all six ancestor-generations, beginning with the parents of the twin-pair, which means that space has been arranged for 126 ancestors, all genera- tions taken together. To facilitate reference the numbering of these 1 Ig is, of corn.re, impossible to pubNsh such pedigrees i~ full. They will therefore be kep~ in ~he "Ias~i~u~ :for ArveHghe~sforsknhlg" of ~he Universiby or' K~'is~iani~. 156 Hereditary _Predispositions to Dizyyotie Tw'in-births aj ,-...... 8 b0LI '0'W o~ ~ ~o 00LI '~'0'M -~ -3,.3-,~,:[.----e~ 8 "-A'~'g ~-_~. ,p tr691 '~'I'V v~ ~,q c, - ~',I bg t-4 ol- . ~:~ s OL 9Y :d ~- i',.- '%| ~.i - .~. ~.o~ L9 s~ IN- c~ :.Cb~ ~o~sj "-\ "(l'I L891 Oh sv _ ~1 ~o~ - ~o A9 ~ b. ~ @ p g951 '~'I o.~- ~ II~l 'q'~'I ,,~ ~ - ~,'q'I'~ ~. 2- .o ~o~ "1~.'~ ~ ~" - me: dd "g "I ~ ~ o o ~o '~ 'O b , 2 ~0~i YI'O .~ ~ - ~ j~ '8 'l '~'~3"[ ,p ~'~1 "CJLEq o~9~ 'O'D 'B~II'TJJV 2 ~ = o ILgl '(3 "l ag "Z'O '~d a2~- C~ "-1'2'8. .~1 ~ " ~ ~1 c~~ - m _~- ~L~l "W '~ -i ' ,~ ~.~ "TI LL 'ffLV3 s ~l ~'r3 .2 tm o ,P "~'V -I b9 K. I~0NNEVIE AND A. ~u 1.57 spaces has been made conbinuously through all generations beginning with the parents of the twins. In these schemes have been noted for each ancestor nob only name and birth-year but also family-munber in ~he pedigrees and several o~her da~a, as year of marriage, place of living, etC., which have been omif~ed in Fig. 7. Of importance for the question of heredity is ~he introduction into ~hese schemes also of special symbols showing ~he relationship of an ancestor ~o other l~win-mof,hers. An exip]ana~ion o:~ these symbols which will be found also on the pedigrees, is given in Table VII2. 0 " ,Ko~ T?2o ~.~ ~ ~ ~ ~-~oU~ F ~ ;0 ~ ZN ,, TABL~ VII. ,Surveyof markings used on a,ncestry-schemes •ud on pedigrees to show the degree of rMationship ~ri~h a twiu-mother. Join'n, of Gem xvl 11 158 He,reditary Predi~spositions to Dizygotic Twin-births The symbols call attention to all individu~ls who, because of relation- ship with twin-mothers, might be suspected to be conductors of the hereditary factor. Without an exact knowledge of tile type of heredity it will, however, be impossible to tell to what degree of relationship such probability should be ex~cnded. In an investigation of the descent of a predisposition to dizygotic twin-births, considered as a character, three diffictfltiss must be borne in mind. (1) This character must ])e sex-limitediu its phenotypicM ~ppearance. I:[ carried by men, it may ]?ass unperceived through several generations and then, perhaps, suddeuly appear again in :~ woman t)elonging to at1 a.pparently free line. This difficulty may be partially obviated by the supposition of an occurrence of men-conductors in a number equal to that of the women carrying the hereditary :factor, the latter alone being counted and the final number of twin-mothers being doubled in order to give the supposed number of "twin-parents." (2) The phenotypieal proof of the existence of tMnning'-heredity does not, even in a woman, appear until the following generation. All women without children should, therefore be left out of consideration. (3) k double ovulation seems, according to our expm~ence, to be of a rather rare occurrence even in women carrying the hereditary predis- position for dizygotic twin-births, "repeated" twins occta'ring in only 7 per cent. of all twin-mothers of our material. Therefore, a woman who because of her heredity ought to be a "twin-mother" may perhaps not have manifested this character although she may hays had several single births. A.s a matter of fact, not only the childless women but also all those who have borne less ~han five children have been trea~ed separafMy; tNs limitation is based on ~he official statistics of Norway (1916-17) that more than 70 per cent. of all twin-births (1216 out of 1709) occur ~mong the first five pregnancies. The probability of a woman being a twin-mother without having shown it through her births, will ~hus be very considerably diminished by such exclusion of families with less than five children. This restriction of the material considered is, of course, very severe, reducing the number of individuals to a small fraction only of those really be:re. But at the same time th6 restriction is in no way selective, and the results reached will, in spite of the reduced figures, gain very much in reliability. K. Bo~57~vn~ aND A. Swamure 159 The three resl;ric~ions mentioned must, however they are effected, all work in the same direction, that namely of reducing the actual occurrence of twin-births below the frequency expected according ~o the laws of heredity. (a) Sex-linked inhe,ritance. The possibility of a causal relation between fsrtiliby and dizygot, ie ~win-birl)hs raises ~hs question of a l?ossible l?arallcl with Pearl's rsstd~s. As shown by him (].915) l~he ferf, ility in the common fowl is inherited ]?argly as a se~-li,nl~;ed character. (err. 2, el, P]. XVI) (z,zq I I c,,,? "~ "" , I,ttt laZZ 18gg 1863 1887 Fig. 8. Sm~ll fr~gmenb of a pedigree (see explanation ~o Pls. XV, Xu Even in our material small pedigrees may be found which, isolated, might seem to indicate that our character is a sex-~i,nlcec~ domincm~. Thus in f,he pedigree of Fig. 8 (a small part of the larger pedigree of P1. XVI) we find in the first generation a wdman, who has in her first marriage five, and in a second marriage four children: in aII, four sons and five daughters, none of them being twins. This woman as well as 160 H.ereditary Predispositions to Dizygotic Twin-births both her husbands belong to twin-families all. three having twin-mothers among their nieces. On the supposition of a sex-linked inheritance, the predisposition for dizygot,ic t,win-birt,hs should pass from a man with his sex-chromosome, to all his daughters, while none of his sons should become conductors. In the pedigree of Fig. 8 this might seem to be the case: among the five daughters of the first generation one died young and all the others proved to be t,win-mothers; the %nr sons have, on t,he ot,her hand, no twin-mothers among their descendants. On the same assumption of a sex-linked in]~erit,ance the daught,ers should be expect,ed to pass thei{' predisposition t,o half of their daughters as well as to half of their sons. In fraternity 1, 4 (Fig. 8) one of the gwo daughters (I, 4T) is a twin-mob,her and her son (1, ~m) again a conduct,or, the other daughter has borne no twins. The twin-mot,her (1, ,tin) appearing among tJae children of the latter night have received a new inheritance from her father who is seen to be himself the brother of a twin-mother of another family. The only case, which in the pedigree of Fig. 8 would not fNly agree with the supposit,ion of sex-linked dominance is that of the twin-mother, no. I,,mm. Her father should be supposed not to be a conductor, and if the inheritance had been carried by the mother, she ought herself to be a twin-mother, a possibility which is, however, nob excluded even if she has so far borne no twins. But pedigrees like t,hat, of Fig. 8 are mere except,ions and a review of the whale material makes it certMn that the predisposition for dizygotic twin-birghs cannot be considered a dominant sex-linked character. The inadequacy of such hypothesis is, as an example, proved by the small pedigree of Fig. 9, taken from P1. XVII. In the first generation we have a fraternity of eight children, among whom t,here are no twins ; t,he mother therefore is in all probability not a f,win-mother. On the supposition of sex-linked dominance the heredity should t,herefore have been introduced by the father, and presumably carried on to all his daughters and to none of his sons. Only one of the daughters (2, 1) appea,rs, however, t,o be ~ t,win-mother although her two sist,ers have nine and five children respectively, while on the ot,her hand, two sons appear as conductors, one of them (2, s) having three, the or,her (2, 3) two twin-mothers among their children. In case of a recessive sex-Iin]ced inheritance a woman should never become a twin-mother except when receiving the factor in question from both her parents, while a man should receive his inheritance from the mother only. A tr.ansference of the inheritance through K, BoNN~v~ AND A, Svm~)~ 161 ,.3 J 62 Hered'[tary Predis1) ositions to D~zygotic ~win-births males in two consecutive generations would, thus, not ~gree with this supposition. As a matter of fact it would, however, be impossible to establish m~i~cient proof :for such relations. The conditions of the women wm~d here be just the same as in case of pure recessivity and no sex-linkage can be proved by their inheritance only; and as for the men, the limitation of the phenotypical appearance of the character in question to the female sex makes it impossible to decide in each special[ c~se, whether they are carrying the hereditary factor or not. (b) Do~ninctnce. In c~se of pm:e dominance the hereditary ch,~r~cter ought to appear in a series of successive generations. A.cknowledging the difficulties, mentioned above, caused by the sex-limitation of the character in El j za 17(~ IA l'Z~l 18%3 I~ I, II.~ 183% L t '.lizl Fig. 10. Fragment of pedigree (see expla.nation to Pls. XV, XVI). cluestion, we shotfld nevertheless expect to meet, frequently, with cases in which such transference is visible. Looking over our large pedigrees we find, indecd, lour or five cases, K. Bom,~Ew~ ~k~D A. SvEmDmup in which twin-mothers appear in at least three successive generations. Examples of such inheritance will be found on P1. XVI, from l~ennebu; one case from l~ingebu is reproduced also in Fig. 10. We find here one twin-mother (no. 1, ll) who among her nine children has two daughters (1, ill and 1, llf) proving to be twin-mothers; one of these has, again, a twin-mother (no. 1, ].]9.1) among iher children. The zti.rgt mentioned twin-mother (1, ll) is :l!m:l;iher, daughter of a man (no. 1, l) wlho is the brother of a twin-mother (no. 1, s) and /~herefore, very pro- bably, himself a conductor of the hereditary factor. Here we have, thus, no less than fora' successive generations displaying the character in question giving, if considered al(me, a good illusgration of pure dominance. M l S 38 9(~ 89 I ~38 TA]BLE VIII. Statistics oil the p~rengs of 238 gwin-mol,hers (see text). Attention should, however, at once be drawn to the fact that at the same time, even in this pedigree, the necessary conditions are found also :for a recessive inheritance, the parents of each twin-mother both belonging to ~wi~t-lines. On the wt~.ole, and taking into consideration the large number of twin-mothers (more than ~100) in the Norwegian family-material, the cases indic&ring pure dominance are remarkably few. But acknowledging on the other hand the many different reasons for a phenotypical sup- pression of this character (p. 158) as well as the possible e~stence o~ modifying factors suppressing the dominance phenotypically it will be 164 tte~"editaxy Predi,?positions to Dizy,qotie Tzoin-bi~'ths impossible, from the pedigrees alone, to reach any definite decision with regard to the question of dominance. In order to advance a further step an investigation has therefore been undertaken with regard to the parents of twin-mothers (Table VIII), starting from the supposition that, in case of pure dominance, a twin- mother ought to have received her inheritance from each of her parents in the proportion i : 1. As shown in Table VIII, 35 twin-mothers have, in the Norwegian pedigrees, been found to be themselves daughters of ~win-mothers, thus having' evidently received their inheritance from the mother. Among the other fraternities including twin-mothers, all those (89) have been left out of consideration in which the number of children is less than five (see above, p. 158). No less than ll,l fraternities consisting of five or more individuals still remain in which a twin-mother is not seen to have received her inheritance from the mother, and where, therefore, the father should be supposed to be the conductor. But this proportion 35 : 114 is very far from the expected 1 : 1, so far indeed that this difference makes the assumption of a dominant heredity very improbable, allowance being made for all the above reservations. (c) Recessivity. On the assumption of recessivity each twin-mother should be con- sidered a recessive homozygote, while both her parents must be con- ductors of the hereditary factor, that is either hetero- or recessive homozygotes. This condition has been investigated for all the twin- mothers occurring in ore" pedigrees, the ancestry of each being thoroughly studied. In order to estimate the hereditary qualities of the parents of a twin-mother we have fmmd it necessary to know her ancestors for at least three generations with all obtainable information as to their genealogical relations, fraternities, descendants of sisters and brothers, e~c. For many twin-mothers we could not, of coarse, obtain fult knowledge. The results of our investigation with regard ~o 176 twin-mothers from i~eldal and l~ennebu are given in Table IX. A full knowledge of the ancestors, within at least three generations, has been reached for 46 twin-motllers all showing in the ancestry of both parents the presence of twin-heredity. 50 other twin-mothers belong to families well known on one side, either of their father (24 cases) or of their mother (26 cases) and here again, twinlheredity is without ex- ception proved to exist on the well ascertained side of. their ancestry, K. BONNEVIIfl AND A, ~VEI~DNUP 165 and very often also on the other side. For 80 twin-mothers, finally, it has proved impossible to clear up three generations of their ances~ry on either side; but even here twin-heredity has been fmmd to exist in 19 cases upon both sides, and in 15 and 31 cases on the side of their father or their mother respectively, while in ten cases twin-births are found to occm' within, or among the descendants of the fraternity of the twin- mother in question. Summing up, we find among 176 t;win-mothers from lc~lXa~.tz ,tor Bod~,l~t~ % ~q [9 7 lq I"1 o-'et,r lq I~ 3) ~0 8O 86 2Y 50 IO Ta~L~ IX. Survey of ancestry of p~renbs of 176 twin-mo~hers from $ieldM and l~ennebu. ~oldal and Rennebu no less than 86, both parents of whom belong to twimlines, while for 75 others the same may be said with regard to the side of their ancestry which has been flflly investigated. For five twin- mot~hers only, with very insufllciently known ancestry, nothing can be said with regard to the probability of their parents being heterozygous. Conspicuous illush'ations of such twin-mother ancestry are given in Fig. 11 a, 5, one from ~eldaI, the other from P~ennebu. To these results may be added also those from I~ingebu (Bonnevie 1919) where, besides 21 twin-mothers with insufficiently known ancestry, "67 twin-mothers, the ascendants of whom are known through several generations on one (30 cases) or on both (37 cases) sides, are without exception shown to descend from twin-families through both parents or through the one of them which is known." On the whole therefore our investigations have, so far as it has been possibl e to reach a smCficient knowledge of the ancestry, exhibited without exception the conditions necessary for a recessive heredity to exist in our family-material. This result does no% of course, in itself prove the recessive heredity 166 ttereditary Predispositions to Dizygotic Twin-births of dizygot, ic twin-births. Considering the high degree of inbreeding in the poptdation of Norwegian v~leys i Ois to be expected that a doable- sided heterozygosis might be found in any family, even outside the twin-lines. Since in the very large series of cases investigated, the whole evidence is consistent with the hypothesis that the capacity for dizygotic twin-births descends as a recessive, nothing incompatible with qoj, 8 C. o AE I'/O~ " d 176~ 't----0 d' og-- - ~ ~,~ M m,~"'" 'd. d P0 ' ~ BP @ -r N ? daJ dJa I'~G7 ~RJ IG8X 9 1708aa )l-{J d dTO I'I01 / X o030~ HR .g ,,a~ # 17.t,~ )all d r 9 "~ o~RH d ?mBE P 1706MB ; dBP 1683 *no~ daa oCA ~689 1703 Fig, 1.1. c,~. that representation having been seen (see p. 162), the conclusion mus~ be regarded as well founded, Our investigagion with regard to recessivity shoNd not, however, be finished without taldng a survey also on the ge~ecdogica~ side of our material. We very much regret the necessity of restricting our demonstration of tlJs material to the examples fotmd in the twin-lines of Pls. XV, XVI, K. ]-J0NNEV~ aND A, SvE~P~up 167 the twin-births here collected representing only ~ small fragmeng o17 those found on our ftLLI,pedigrees. For saving space we have a]so been obliged to omi~ in these plates everything which was not of direct significance wibh regard ~o the geuotypical relations of Dwin-mothers; all lines withouD any twin-births, having earlier been utilised for ~he statistical results, have therefore here been excluded. For comparison with oni' own O ~ ~ i~ 0 5~s # RM? o9~0~'~~ ~%~ ~o--~, -~ / ,;<,' ~E A "o oo~ va @MO I'iO& Fig. l l b. Fig. 11 c4 b. Examples of ancestry-schemes,a from i~eld~l, h from t~,ermebu,showin~ gwlrt- mothers (~) a/way~to desoerKlfrom p~l'en~a belonging t,o ~win-lhies. n:aberial we have i~ P1. xgII shown also the twin-lines of the large Blek{nge family described by Lundberg (1913). Unlike the two other pedigrees P1. XVII contains no[ only a fragment bu[ the whole number of twin-mothers born within ~he large Blekinge family. The three pedigrees of Pls. XV--XVII, representing each some spadG oally characteristic features, will in the following be considered separately. 168 tlereditary Predispositions to Dizygotic Twln.births The Dombu family, Fam. 2, the twin-lines of which are represented on PI. XV, is an old, and therefore also a widely spread, family within the parish of Meldal. It contains upon our full pedigrees 326 fraternities with in all 1196 births, including 34 twin-births, the percentage of twin- births thus making 2.8 per cent. or somewhat below the average of our whole twin-family material (3.25 per cent.). The "twin-lines" of this family, represented on PI. XV, contain's 96 fraternities with 425 births, among which 29 or 7.5 per cent. twin-births, a percentage which is, again, somewhat lower than that (8.23 per cent.) of the whole group of twin-lines considered on Table V (col. 7, line 9). Upon the whole, therefore, we have before us on PI. X V a real twin-family, but at the same time a rather Iow-graded one. On the pedigree of P1. XV the number of twin-mothers (29) amounts to 6.3 per cent. of all individuals (457). A survey of the pedigree of PI. xg proves first of all the result, already mentioned (p. 165), that whenever a twin-mother occurs, we find either both her parents or at all events the one of them whose ancestry is known, to have other twin-mothers among their relatives. The first fraternity of this pedigree (2) does not include any twin- mother but it is, on the assumption of a recessive heredity, certain that at least one of the parents must have been a heterozygote, one of the daughters (2, s) being the mother of a twin-mother (2, 32). Also in other lines (2, 1 and 2, 4) twin-mothers occur in early generations, making it probable that the same recessive heredity is present even here. In the line 2, 5, the twin-mothers are found only in later generations, and it would not seem impossible that in each of the three branches of this line (2, 51, 2, 5~ and 2, 5s) the twin-birth heredity had been introdticed from outside. The fact however that twin-mothers are born in each of these three branches all descending from the same fraternity as 2, 3 and that, on the other hand, an introductio~ from outside to both parents of the twin-mother would need in each case at least two crossings with conductors from other branches, all this indicates a recessive heredity to be present also in the first ancestress (2, 5) of this family b~'anch. Upon the whole, the Dombu family gives an example of a typical twin-family with a relatively low number of homozygotes, as woNd be the case if, for example, in the first ancestor-pair, one heterozygote has been crossed with a dominant homozygote, and if the inbreeding with other twin-lines has not been very strong. A different type of twin-family is represented in the Grindal family, Faro. 2 of Rennebu (PI. XVI). The percentage of twin-births is here raised above the average to more than 10 per cent., more than 9 per cent. K. BONNEVIE AND ~. ~VEI~DI~UP 169 of the 472 individuMs of the pedigree being twin-me,hers. In P1. XVI b is shown the ancestry also of some individuals coming from o Sher twin- families (within the "twin-lines" of this family) of the .same, high-value type. In these families we have before us not only several cases of twin- mothers occurring in a series of successive generations, but we find also 'a number of cases with two, o1: even 3-4, twin-mothers in the same fraternity. Pedigrees like t,his might at first sigh~ make one think of a dominant type of heredity (see p. 162) but each separate case of twin- births fits at the same time equally well inf,o a scheme of reecssive heredity as working under a strong inbreeding beSween twinqines. The two types of twin f~unilies demonstrated in ~he pedigrees of P]s. XV, XVI are, in fact, good representatives of the whole pop~dMion of the valley investigated: there are ~o be found large twin-families of the Dombu type (Pl. XV) with scattered strains of latent heredity for clizygoCie twin-births carried through their various branches, bu~ with no great number of twin-mother,s, and there are branches in which the number of t,win-mothers may be raised considerably above the average of all twin-families t~ken together, and which through inter-crossing may bring about, results like tha~ of the Grinclal family (P1.XVIa-b). The di'fferenee between these t,wo typesis, on the assumption of a recessive heredity, easily explained through variations in the genotypieal starting-poinfi of the family-branch in question, between the two extremes DR x DD and RR x R_R, and through variations also in the degree of inbreeding between twin-lines. With regard to the Bleldnge family (P1. XVII) described by Lundberg (1913) we fNly agree with this author when he says (p. 427): "Es ist augenscheinlich, dass eine etwas stgrkere An/age fitr Zwi]lingsgebmn%n seit glteren Zeiten innerhalb dieses Geseldechtes nicht, existiert, denn da sol/ten wit eine weitaus grSssere Anzahl Zwi]linge innerhalb desselben erwarten dtirfen, zumal die Verwandtenehen sehr gew/ilmlieh sin& Es sieht deshalb aus, als ob die Anlage ftir gwillingsgeburf, en erst verh~lt- nissmgssig spat dem Ges&Llechte zugefiihrt worden sei." In Ltmdborg's admirably ftfll and extensive pedigrees we find "das grosse Geschlecht" arranged in "Linie I--gI," each line descending from one of six sisters and brothers, borne 1745-64 bythe first ancestors. Line V, which had already, 1799, left the parish, has not been worked out. To this large family, including 1909 bir~hs, of which 1.466 per cent. are twin-births, are added a few small pedigrees Of Lundberg's, one or more individuals of which are married into the "laNe" family. 170 Hereditcery 2:'redis2ositions to Dizygotic Twin-births With regar d to Lhe distribution of 28 twin- (and 1 triple-) births, Lundberg finds "dass Zwillingsgeburten in den verschiedenen LiNen, wenn man deren Grtsse beNcksiohtigt ziemlich gleichmgssig verteilt vorkommen." (Table XXIV, p. ,128.) In order to compare and control our own rest,Its with regard to the herecliby of twin-births we have ventured to rearrange the pedigrees of Lundberg so as to brace the ancestry of each twin-mother as far back as possible. Suchrecons f~ruction is found on Pl. XVII, in which to facilitate reference control each fraternity carries, instead of the usual decimal-number, the number of the plate and falnily used by Lundberg in his ~elas. It shotfld be menLioned also that only such twin-mothers have been considered by us whose birth is to be found in the pedigrees. From the pedigree on PI. XVII it is evident that no less than half of the 24 twin-mothers considered may genotypically be included in the two small pedigrees of "Eskil Svensson" (Lundberg I, a 14) and of "Faro. B" which are seen to have, in directly descending lines, eight and four twin-mothers respectively. Among the six lines of the "large family" lines V and VI have given no proof of twin-birth heredity, while at the same time in line I, III and IV some hereditary factor for twin-births must have e~sted from rather early generations. It might, for example, under the assumption of a recessive heredity, be supposed that, in line i the twin-birth heredity has been introduced through the male parent of faro. 2, 1, in line III through the female of faro. 15, l, and in line IV through the male of faro. 28, 1. Even in line II hereditary twin-factors must exist in order to explain the occurrence of the three twin-mothers in faro. 20, i (line IV) ; the heredity might here perhaps have been introduced through the male parent of faro. 9, 1. None of the persons mentioned are wi~h regard to their genealogical relations known well enough to justify any conclusion as go whether or nob they belong to twin-families. The same resNt might, however, have been reached also through the supposition of a recessive heredity existing already from ~he beginning, in one of the first parents of the "large family," and spreading in the first generation to at least four of the six children (lines I--IV); with regard ~o line V we know, as already mentioned, very little. If, after all, we think Lundberg right in his conclusion that the "large family" cannot be considered a twin-family, it is because of the large number of family-branches (not represented on Pl. XVII) in which no twin-heredity has shown its existence. Even in a very low-grade "twin-family" starting from a cross DR • DD, we should expect a K. Bo~EvI~ A~I) A. Sv}~oI~lJl, 171 spreading in the various family-branches far more eonspi cuous than that here fmmd. The mere statistical result of no more than 1"i66 per cent. twin$irths, will in comparison urith the 3.25 per cent. in ~he Norwegian twin-families, in itself be suNcient to prove that the occurrence of hereditary twin-births in the Blekinge family must be supposed to be restricted to certain branches only. This family thus supplements the two hypes of bwin-families represented oll P1. I alld P1. ]I, the three )edigrees finding, in all their diversity, a, full explanation iD the assump- tion of a recessive heredity for dizygo5ic twin-birl, hs. d~.NNEX" DO DIZYs TWIN-BII~(I?HS ALWAYS I~[!IPP~ESENT ONE aND Tr~ S~km,: cm~kc~r~ld As mentioned in chapter ] of this paper (p. 135) we have, besides ~his family material, considered also an isolated case of such predis- position, a woman from Elverum who with her eight twin-births occupies an exceptional position among all our twin-mothers. Special stress has been laid, therefore, upon investigating the question whether or not we have before us, in this case, a hereditary character of the same nature as that of the other twin-mothers in our material. Fig. 12 gives our present, knowledge about the b~rths and the family- relations of this woraan. As will be seen from the pedigree the tirst f0ta' pairs of twins were born within ~he lapse of less ~han 3 years (24/4 1905--27/2 1908); then follow, within the next 't years (until 18/9 1912) fear single births, and then again at a somewhat slower rate during 6 years (until 29/3 1919) four new twinJoirths, and, finally, after 1 year a single birth. This youngest child is, when this is written, nearly ~t years old; the age of the mother is, however, no more than 41 years, and the number of births may perhaps still be augmented. At any rate she has, so far, borne 21 children withhl the lapse of 15 years. Ten of these children died during their first year, all the twins having, as told. by the midwife, been born about one month too early, add the mother being, from the first birth ml, absolutely incapable of suckling her childrml. Four of the twin-pairs are bisexual, the remaining four pairs uni- sexual; but according to bhe midwife two ehorions have been present in all twin-bil~hs. We have therefore, no doubt, before us a case of some anomaly in the ovaries making a double ovtflation the rNe instead of an exception, probably due either to a simttltaneous bursting of two follicles Or to the existence of follicles containing more than one egg. Ovaries showing both these kinds of anomNy have, as mentioned in the intro- 172 Hereditary Predispositions to Dizygotic Twin.births i 8 o gh'd eo 2~ca B~ bb N K. BONNEWE AND A. SVEI~OI~UP 173 duction (p. 128), been demonstrated already by previous authors. The lack of milk-secretion may perhaps Mso be seen in a causal connection with the anomalous ovary. The question whether this character can be considered transmissible has, of course, been of special interest to the authors of this paper. In- vestigations have been made in the field as well as in the church-books, but so :far wi~h an absolutely negative res~dt; no twins have been found in the family, and none of the twins occurring within the last century, in two parishes, ih~ve been found to descend h:om any known ancestor of the ~win-mol~her in quesl;ion. The value of this negative resNt is uot absolute, in. so far as no more than three ancestor-general~ions have been investigated and, especially on the side of the mother of the twin-mother, with n.o satisfactory result. Possibly in some o~her parish ~win-mothers might be found descending from the great grand-parents of this twin-mother. But even so, the pedigree before us (Fig. 12) makes it very little probable that the predisposition for twin-births shodd in this "case have been inherited. A dominant heredity ought o.ertainly to have become apparent within severn of the fraternities (some large) of the pedigree. In case of a recessive heredity, on the other hand, both parents of the twin-mother should be considered heterozygotes, an assumption which may, as already mentioned, perhaps not be excluded for her mother, though for her father that seems very improbable. ]?Ie descends from ancestors living through generations within the parish of Elverum and even if the genealogical investigation has not been carried further back than the second half of the 18th century, the fact that none of the later twin-biiShs of this parish can be traced back to any of the known fraternif,ies makes it practically certain f,hat~ ~he father of our twin- mother does not belong t~o any twin-family and that he is himself no heterozygote. Upon the whole, therefore, it seems that we have before us in this woman afirst apu)earance of an anomaly in the ovary causing, as a rule, double ovulations. The question whether ~his anomaly should be considered as hereditary cannot be decided tmtil one or more new genera- tions may be added to our pedigree. Similar eases from cattle (Pearl 1912) as well as from. man (Stocks (1861), Strassmann (1889), I%osenfeld (1903), Brattstrhm (1917), Daven- port (t919)), have, however, occasionally been described in literature, or even in newspapers ; and in so far as ~hese desm~ptions may be considered reliable, t,he anomaly seems to be inherital)le. Journ. of Gem xvI 12 t74 Hereditary PredislJositions to Dizygotic Twin-births Special mention should be made of the one case only, whichhas been described by Davenport (1919). Here we find in a family from Cleveland one woman, three times married, with in all 15 mldtiple births, out of which 7 were twin-births, 5 ti{ple and 3 quadruple births; no birth was single. "Thus the proposita has averaged nearly three children at a biigh and has had thirty pregnancies inside of twelve years in the last mating, in addition to the eleven by the two earlier matings." This very extraordinary case is evidently of the same di-(poly-) zygotic natnre as the one described in this paper. In the two foregoing generations "ordy twins, triplets and quadruplets" were Said to have been born, aTld even in the latest generation one ~win-birtt~ has occurred. The hea'editary nature of the anomaly seems, therefore, in this case evident, and the whole pedigree gives a strong i,)dicat[on of a dominant type of heredity. The whole appearance of such am)malous predisposition to multiple births makes us incline to consider it a character different from and independent of the more common hereditary tendency to dizygotic twin-births. CKAPTEI~ IV. DISOUSSION. We have now to consider how far our conclusions are in agreement with the i'esults of previous authors. Look~g over the genealogical investigations on twin-heredity we shall, as already mentioned in the introduction, find many of these pedigrees more or less incomplete, in so far as very often the ancestry was imperfectly investigated. From such incomplete pedigrees no reliable conclusions can, of course, be drawn. The most complete pedigrees on twin-heredity fmmd in literature are those of Meyer (1916). With regard to a comparison between pedigrees from clif~erent populations, it ought, however, to be remembered that we do not, a priori, know whether the twin-blithe considered are always comparable. When in our material, we felt justified in considering the hereditary predisposition to be to dizygotic twin$irths, tJfis view has been ha.sod upon the thorough statistical investigations of chapter II. Similar investigations w~/1, indeed, be necessary also for each new twin-population before deciding the nature of the twin-births. Thus in the large pedigree of Meyer (1916, Taft IV) ot' princely families the number of unisexual twin-births (29) make more than three times that of bisexual ones (9), that is, we shmdd according to the I(. BONNEVIIg AND A. ~V}glt,Dll,UP 175 difference-method have here at least as many monozygotic twin-birt,hs as dizygotie ones. This result, may, of course, be due to mere chance, the number of twin-births here considered being too small to give :reliable results. It shows, however, that care should be taken not to draw general conclusions from inadequate ~amily-material. An apparent discrepancy in our own preliminary results (Bosnevie, 1919) to which attention was drawn by Dahlberg 0923) arose tJ~rough the scarcity of tl~e t,Mn-mat,erial then invest,fag ted. Dahlberg has pointed out,, t,hat in tlle Norwegian J~,ingebu family there seen'ted to be an augment,at,ion not only of dizygotic, but also of monozygot,ic twin-births, wllen compared with the frecluency of twins within the wh(/le population. Statistical investigations upon our whole family-material (Table V, p. 59 of this paper) have; however, with regard to this question given the following results: The frequencies normal to the Norwegian population (O~cicd Statistics of No~',w~y, 1916-1.7) are found in column 1 of Table V. The percentage of twin- (mtfltiple-) births as compared with the total number of births, is 1.~6 (line 9), while, according to the difference-method of Weinbe N, 27.53 per cent. of all twin-births should be considered as monozygotie (line 19) and 72.4-7 per cent. as dizygotic (line 22). In other words 0-t0 per cent. of all births are supposed t,o be monozygotie twin-births (line 20), while dizygotic t,win-birt,hs seem to occur wit,h a frequency of 1.06 per cent,, of aH birt,hs (line 28). In column 2, gixdng the figures regarding the whole population (1750- 1895) of the valley in which our new investigations have been made (Heldal-P, ennebu (Fig. 1)), we fred the general frequency of twin-births t,o be somewhat lower, 1.2~: per cent. (line 9) than the norm given by t,he whole population, but the relation between the frequencies of roche- and dizygof,ic births is somewhat, the same, vim 0-36 per cent,, and 0.88 per cent. of a,ll births (lines 20, 23). Corresponding figures :for tim family-material are found, for each ~amily alone in eels. 3-5, and for all families together in col 6, while in col. 7 the figures are given for special twin-lines of the families in- vesi,ig@ed. The widely varying results reached for each separate family (eels. 3-5) through the difference-method will not surprise us if we remember that such calculation can be reliable only when based upon large figures (cf. also Table II), which is not the case for each family or for the twin- lines, Iotmd in line 8 of Table V. By summing up the restflts of all the :families (col. 6), the influence of chance is, however, largely reduced, 1~9--2 176 lgereditco W lb'edis2ositions to .Dizygotic Twin-bi'rths and here we find monozygotic twin-births in 0.~7 per cent. of all births, a resulf which is seen to be very near that oi the whole Norwegian population. Even here the number of twin-births is too low to be accepted as a definite restdt; it does not, however, suggest an increase in the number of monozygotic twin-births in our family-material. The whole increase (line 9) from the normal percentage of twin-births (1.4:6 per cont.) to that of the twin-families (3-25 per cenL) is, on the contrary, seen to be clue to I~he corresponding increase of the percentage of di- zygotic births (lin e 23) from 1.06 per cent. to 2.78 per cent. When in our investigations upon the heredity of dizygotic twin-births the twin-fathers have been left out of consideration we are, as already mentioned, in ftdl agreement with previous authors both clinical and gynaecological. In a few recent papers upon heredity of twin-births the view has, however, been maintained that the twin-~a~her might play a more or less important pal4 not only in the production of mono-, but even of dizygotic twin-births. Thus IXeyer (1916) after having (p. 299) declared with regard to dizygotic twin-births that "diese Stgrung in der Funktion des Ovariums natiirlich nut dem weiblichen Organismus selbst innewohnen (kann)" finds himself "auf Grtmd des Stammbaum-materials" obliged "auch diesen Vorbehalt ale ansicher and fraglich bezeichnen zu mtissen."-- "@ewiss ist es uns," he continues, "bei den zur Verfiigtmg stehenden Theorien der Entstehtmg Zweiefiger nicht Ieieht, uns sine reehte Vor- stelhmg fiir direkte v~tterliche Beteiligung zu machen, aber bei ~{usterung der Stammbgtune drs sich einem geradezu sine l%eihe yon Fallen auf, bei denen die Nichtmitwirhmg der selbst deutlich belasteten Vgter yon Zwillingsp~rchen, also sicher zweieiigen, einem ale sine gezwtmgcne Auslegung erscheinen m6chte." The author ~hen points out no less than ten different cases in his pedigrees which, according to his opinion, necessitate the conchlsion tt~at the twin-fathers acted as conductors. After careful consideration of all these cases, we do not, however, agree with his conclusion. In the first case mentioned (S~ammbaum VII, 62) both parents belong to twin-families, the mother being a sister of another twin-mother, twins existing also among the sisters and brothers of her mother, and her grandmother being herself a twin. In this case the fact that also the father belongs go a twin-family should not be sufficient to "necessitate" any conclusion with regard to his influence on the two twin-births among his own children. K. BoNN~vm am) A. SVEaDaUr 177 k sin~lar case, both parents belonging to twin-families, is represented also in Stb.VIII, 160. In six other cases, mentioned by ~eyer, the family of ~he mother is either not at all, or at any rate only imperfectly known (Stb.VII, 69, 70; Stb.VIII, 151, 198, 199, 22@ There remain still two crossings (Stb.VIII, lll+ 112) in both of which ~he same man, belonging to ~ twin-family, is the ~win-father. /A3___~ qT"3,511 Fig. lB. Fragmen~of pedigree fromI~fllgeb,u(of. T~bleVII, and ex ~l~nabio~to Pls. XV, Xu Here again, however, the ancestry of 5he twin-mothers is in one case absohltely unknown, in the other known only on the side of her father, while no information is given with. regard to her mother. None of the ~en cases, therefore, quot~ed in supporl, of Meyer's view, appear to us as being in any way conclusive in favour of a hereditary influence of the twin--father upon dizygotic twin-births. We find in our own -family-material a whole series of c~ses in which no~ only the mother but also the father of twins belong to twin-families. If, for instance, in the pedigree of Fig. 13 the ancestry of the twin-mothers had not been known, we should have had before us no less than ~hree families (3, ~2; 178 flereditary Predis~)ositions to Dizygotic Twin-births 3, ~2; 3, ml) in wlfich gwin-bilShs might seem to have been caused by the lathers. With full knowledge also of the ancestry of the ~win-mo~hers, we find, however, in MI these c~ses in the family of ~he mothers alone a fully su:mcient basis for the hereditary appearance of twins, each of the five ewin-mothers of this pedigree being born of parents both belonging to gwin-falniHes. Also Davenport, (1920) draws, from his statistics upon a material of "repeater-families" (with more than one pair of twins), the conclusion that (p. 125) "the father has about as much influence in the production of l;wins as 6he mo~,her." This ma~eriM of Davenport has, as it seems, been collected by field- workers, and ~he "repeM;ers" belong to a great many different families. No distinction has been made between roche- and dizygotic twln-bid;hs; no information is given either wil)h regard to a possible relationship between both parents, or with regard go the number of cases in which not only the fathers bug the mothers also of "repeated" twins belong to ewin-falnilies. It is clear, therefore, that the frequency-figures given in the paper of Davenport cannot p.rov~ anything with regard to the conductors of ~he tenclency to mono- or dizygogic twin-births. Even if it mllst be acknow- ledged ~hat the statistics of Davenport represent a Surprisingly high number of cases in which the fathers of "repeated" twins belong themselves to twin-families, we fully agree with Weinberg (1924) who maintains that the material of Davenport is small in comparison with his own, which proves beyond[ doubt the overwhelming hereditary infuence of the twin-mothers. Davenport (1920 a) sets foiCh also an hypothesis to explain how the father may have an influence upon the production of twins (dizygotic). "It may well be," he says (p. 125) "that two eggs are simMtaneously ovulated much more frequently f,han at present recognized and that the comparative rarity of twin-births is due either to a failure of fertilization of one egg, or ~,o a failure of development of one egg." That such "failures" of fertilization or of development really exist he concludes from an investigation of "birthdnterval.s" within a certain generation of "the Cotton family." Assuming the normal interval between birf, hs to be about two years, he looks upon "the frequent intervals of 3, 4, 5 and even more years" as examples of "failures," although he admits the possibility also of other explanations. In analogy with the results reached upon certain animals, he thinks ~he failures due to "lethd factors," representing (p. 128) "widespread phenomena even K. BONNEVIE AND A. SVNgDgUP 179 in human germ cells," and accouni~ing "for I~ ecrU,sin proport~ion of long in~ervals between births, of early miscarriages, and of s~erile unions." According ~o ~his view he looks upon (p. 128) %wins ~ha~ are born" as "~he residuum of a grea~er number of ~wins ~ha~ s~ar~ in ~heir de- velopmenL and of a s~i11 grea~er number of pairs of eggs simu1~aneously owfia~ed," reaching a~ las~ ~he conclusion (p. 129) "~ha~ among humans ~he cases of ~win-repea~ing fraternities are those in which ~here are no or few le~hal factors in ~he germ cells, so ~ha~ ~here is a maximum fertilization and developinen~ of ~he eggs laid"--and further "~ha~ fainilies which easily produce ~wins do so no~ only because in ~he mother ~he eggs were laid in pairs, bu~ also because in ~he father ~he Sl?erin is ac~ive, abundan~ and wi~hou~ lethal factors, so ~ha~ ~he number of eggs fertilized and brough~ ~o flfil ~erm approaches a, maximum." The question here ~ouched by Davenpor~ abou~ ~he numerical relations between actual and obse~'~ed production of ~wins is one of grea~ in~eresL bu~ a~ ~he same tfime one which will no~ easily be definitely solved. A~ ~his place we shall refer ~o i~ only in so far as his hypothesis lnigh~ raise an objection ~o ~he view here ~aken. The reasonings of Davenport, seem t~o be based upon ~wo di~eren~ assumpl~ions, viz. : (1) t~,haI~a double owdal~ion is a common occurrence in man, and (2) t~hal~ I~he relal~ively rare occurrence of I~wi~-birl~hs, as well as a nabural ocma'rence of long birl~h-inl~ervals, is due t~o lebhal facl~ors in one or bo~h germ-cells causing (a) a failure of fertilization, or (b) a failure of developmen~ of ~he embryo. Wi~h regard ~o ~he firs~ asst~mp~ion abou~ ~he ~requency of double ovulations a ~horough investigation of ~he literature, anatomical as well as gynaecological, proves ~he ovulation of single eggs ~o be ~ypical for ~he human ovary. Descriptions of double and mul~iple ovtfla~ions may be found (see Arnold 1912); bu~ such examples are always considered as exceptions from ~he general rule. BuL on ~he o~her hand, i~ is very probable ~ha~ ~he actual production of ~wins may, even in man, be considerably larger ~han ~ha~ obse~'~ed-- one or bo~h ~win-embryos being checked in ~heir developmen~ a~ an early s~age (Neugebaur 1913). Hammond (1914-) in his discussion of ~he cause of.a~rophic foetuses in swine comes ~o ~he conclusion, ~ha~ (p. 273) "possibly ~he a~rophy of ~he embryo may be caused by reduced vitality due ~o something inheren~ in ~lie foetus, and ~kis may often be augmented by inbreeding." This leads us ~o ~he second assumption of Davenpor~ as ~o lethal 180 Tfereditary Predisposit,~ons to Dizyyotic Twin-births factors causing (a) a failure of fertilization, or (b) a failure of development --the latter par~ of which has found a suppm~b already in the results of Neugebaur and I-Iammond. A failure of fertilization, as caused by lethal factors'in the sperms may of coume also occur, leading to more or less complete sterility of certain crossings. But in his hypothesis Davenport seems to consider the failure of fertilization of an egg a common occurrence in free breeding, and upon I~his point our experience from Norwegian pegsanl>families does nob support his view. In order to tes~ the birth-intervals in such families, the number of years during which the five first births have occurred has been no~ed :for one whole generation of our l~ingebu family, comprising 91 fraternities with 5-10 or even more children. Here no less {,hun 58 mothers, or 63.7 per cent., have borne five children within 11 years, an average birth- intervat of two years, while further 16 mothers or in all 74 (81.3 per cent.), have borne ~heir fifth child before the end of 6he t3th year--and even this last lapse of time gives no great space for "failures of fm• In 86 other fraternities, chosen at random from the church-books, the corre- sponding figttres are not much lower, viz. 55.8 per cent. with five births within i1 years, and 77.9 per cenl~, with five births within 13 years. Taking into consideration also the many ditterent causes, physical and psychical, which may serve to augment the bitCh-intervals of a human family, a failure of fm~ilization does not seem to be any common feature in free breeding. But if a double ovtdation from the htmaan ovary represents an ex- ception from the general rtfle, and if the abtmdanee and activity of the sperm should normally be considered suflleient for fertilization of the eggs present in the tube, then the explanation given by Davenport for the part p!ayed by the father in twin-production has lost its principal base, and does, therefore, not raise any objection I~o our considering the twi~-mother as being alone responsible for the production of dizygotie twins. A thorough investigation of the whole literature concerning twin- births has, indeed, revealed to us no facts opposed to the views main- tained in this paper, nor to the assumption reached I,hat in the Norwegian peasant families investigated a predisposition for dizygotie twin-births is inherited as a, probably monohybt{d, recessive character, while at the same time exceptional cases of an excessive multiparity may occur which do not follow this genera/law of heredity. If. ]30NN~V~I~ AND A. Sv~RDNtrP 181 (1) The results of this investigation have been based upon a geuea- logical, as well as on a statistical, treatment of a aeries of Norwegian peasant families inhabFdng two neighbouring parishes, iVIddM and ~,ennebu; near Trondhjemsfjord, and the parish of tZingebu in ~-ud- brandsdalen. The pedigrees of t/rose families include some 10,000 births among which are some 430 ~win-birthsL Use has been made also of ~he twin-branches of the Swedish Blekinge family described by Lundberg (1913). For a comparison, and :[or ext~enslon of our s~,atlstical restflts, we have used t;he material of the O~eictl Sta[,~;stics of Norwc~y :for 1916-17, as well as of that of the "Women's ]:Iospital" of Kristiania for ] 81(;-~ 916, including 18'-1-8 and 537 t~win-bit%hd respectively (see Table I, p. 42). (2) Within the populations of the parishes investigated twin-births have been found not equally sea~tered but aeeumNated on certain families and within these on certain twin-lines. While the percentage of twin-births in the Norwegian population proved to be 1.g6, the twin- families gave 3-25 per sent. and in special twin-lines the percentage was raised even to 8.23 (see Table V, p. 59). Such aecnmtflation of twin- births in special genetical lines is in itself a strong indication of heredity. (3) Before investigating the type of heredity, the question had go be solved about/abe nature of the twin-births before us. The numbers of m0no- and dizygotic twin-bil%hs calctflated according to Weinberg's difference-method, and arranged according to the age of the twin-mother, have proved beyond doubt that the character inheritable in these families is that of a 2redis2ositio~z for dizygotie twin-births, or in other words, for a double ovulation from the ovary of the twin-mother (see Tables III, IV and Figs. 2-6). (4:) Invest,igations have been made also wi~,h regard t,o a possible causal connection bet,ween twin-births and general felqdJity. Although fertility seems t,o be somewhat, higher wit,hin twin-familles than in the whole poptflation, no evidence has been found indldating the twin- mothers to be at t,he same time the women with the highest fertility (Table VI, Fig. 7). (5) The l;ype of heredity has been invest;igated through a thorough study of the genealogical relations of all ~win-mothers, the twin-relations also of each ancestor being pointed out through special markings on the pedigrees and on ~he ancestry-schemes (Figs. 8-11, Table VII). 1 8~ill-born included. 182 Heredit~ry Predisj~ositions to Dizygotic Twin-births (6) Alt]:mugh in exceptional cases, on small parts of our pedigrees, the occurrence of twin-mothers might be explained as due ~o a dominan~ sex-linked inheritance (Fig. 9), and although some restricted parts of our pedigrees might seem to indicate a pure dominan~ type of heredity (Fig. 11 a, b) such assumptions find no support in the numerical restflts (Table VIII). (7) As for recessivity--twin-mothers being themselves recessive homozygotes--the necessary conditions: that both her parents should be conductors of the hereditary factor, have so far as our results reach, been found to be present in 0m' family-material (Table IX). (8) In the pedigrees of Pls. XV, XVI, in which only the "twin-lines" have been included, examples are given of low- and of high-value twin- families, which, may restfll~ from various combinations r~mging between DR • DD and RR • RR. PI. xgII gives, after a rearrangement of the pedigrees of Lundborg (1913), some twin-branches introduced into ~he Blekinge family which is probably, in itself, no twin-family. (9) As an addition, has been investigated also a separate case of excessive predisposition for ~win-births, a woman from Elveram having borne no less than 8 pairs of twins. No twin-births were found among her relatives and it seems very improbable that her parents shotfld, both, have been heterozygotes. The question whether this anomaly shouId be considered as ]mredlt:~ry cannot be decided until new generations are known. LITEI%ATUI%E. A~mFELD,F. (1875-77). "Beitr. zm' Lelu'e yon den ZMllingen." Arch. f. C~jn. Bd. vii, IX, xI. --(1902). "Wie s~ell~ sieh das Zahtenverhgl~n~ss d. einefigen Zwfllinge zu den zweieiigen?" Zeitsclw Geb. h. ~. Gyn. Bd. XSVrL A~_RNOLD,LLOYD (1912). "Adult Human Ovaries with Follicles containing severM Oocy~es." Anat. 7~ec. Vol. VL BELL, A. G~A~ (1904). "The Multi-nippled Sheep of Beinn Blu'eagh." Science, Vo1. xix. p. 767; and ibid., 1912, ~xxv[. BosrN~vI~, K. (1919). "Ore tvillingfOdslersArvelighet." 2V. Mat. 3~ Laegev. B~A~s~I~5~, E. (191~). "Ein Fall yon viereiigen Vierlinge nebst einigen Beobach- ~ungen betreffs der Vierlingsgebm:~en bn Mlgemehlen." Monatseh~'. f. Geb. ~. C-yn. Bd. xr~. B~EI~ (1891). "Uber Zwfllings-gebur6en(her. aus K~zmal'sky's Klinik)." Zen~rbL f. @~. No. 8. CoB]3, ~I. W. (1915). "Evklenee beal'ing upon the origin of human twins." flcienee, N.S. No. 41. K. BONNEVlE AND A. SVEI~DI~UP 183 CJORu I%. (1895). "The influence of Inheritance on the Tendency to have Twins." Lancet, p. 1105. ])A]~L]3ERCr ~. (1923). "Twins and Hereclity." tIeredilas, Bd. IV. DAVDNPOR% C~. B. (1919). "A Strain producing Nfultiple Births." Journ. Here& Vol. x. -- (1920 a). "Influence of the t3'[ale on the production of Twins." Amer. Naturalist, Vol. i~]:v. --(]920 b). "Herediby in ~win births." Proc. Soc. Exp. Biol. Med. Vol. xv]J:. DREJE)~, ]?. (]895). "Ore Tvilllnger." N. Mctg. Laeg. Vids/c. Tiilegshef~e 1895. DuNo~, 5{. (1865 a). "On aome Laws of ~he Producl~io~ of Twins." l~din, l~fed. Journ. No. 117. -- (1865 b). "On ~he eomparagive Frequency of Twin-Bearing in Different Preg- nancies." Edin. Med. Journ. No. 118. DUNO~],m, G. (1915). "Die Frequenzverteilung d. Ceschlschtskombina~ionen bei l~'l:ehrlingsgebm'~en des Menschen und des Schwei~ls." l~iof Uentrcdbl. Bd. xxxv. DUN;N]:N(~(1389). "Doppel~er Uberus u. Vagina." Zentrbl. f. (/yn. No. 44. E:LLINC~Etr T. (1921). "The Ildiuence of Age on Fertility in Swine." P'roc. Nat. Acazl. So. Vol. vii. F:ESmIER, I%. A. (1919). "The Genesis of Twins." Genetics, Vol. IV. :FRANqU]~, O'~wo v. (1898). "Besehreibmlg einiger seltener Eierstoekpr~tpara~e." Zeitschr. f. CVeb. u. Gyn. Bd. xxxrx. G~ELWOFW (19207). Om tvillingsvangers/xtbet og tvillingf&lselen (2 dok~oor-fore- ]esninger). Kj6bhn. GOEr~DRW, V. (1897). "Die Zwillinge. Ein Beitrag zm" Physiologic des l~lensehen." Virchozo's ArcMv. Bd. ~xxw. t-L,~rOND, J. (1914). "On some fae~,ors cont,'oiling ferSili~y in Domesbic Az~imals." Journ. Agr. Sc. Vol. w. t{ANSnmtNN, D. v. (1912). "Uber den Kampf der Eier in den Ovarien.'~ Arch. Entw. Mech. Bd. xxxv. tt~I~E, W. (1899). "No~oe on ~ohe Fertility of differen~o Breeds of Sheep, wi~oh I%emarks on ~ohePrewlenee of Abor~oion and Barrmmess ~oherein." Proc. Roy. Soc. London, Vol. LV. t{DLLIiff (1895). Ursache der Mult,ipa, ritat der un@aaren Tie~'e iiberhaupt und der Zzoillingsschwangerschaft beim Menschen insbesondere. 3/[iinehen. tilL, J. P. and DONO~UD, Cm~s. JcI. (1914~). "The I%eproduc~ive Cycle in ~he Marsupial Dasyurus viverrinua" Quart. Journ. Mitt. Sc. Vol. lax. JONES, V. It. and I%OUSE, E. (1920). "The I%elabion of Age of Dam ~o obsmwed Feetmdity in Domesticated Animals." Journ. 1)alr. So. Vol. izr. JORDAN, H. E. (1914~). "tteredi~ary Lef~handedness wi~h a no~Ge on Twimfing." Journ,. Genet. Vol. Iv. I~2PDL~, J'. (] 908). Beitr. zur zt na,t. "a,. Physiol. d. Ovarien yon wihllebenden u. gezghmten Wiederk~ucrn ~. Schweinen. Ina~g. Disser~o. Bern. K~NO, H. D. (1917). "The I%ela~oion of Age bo Ferifili~y in the I%a~." Anat. Re~. Vol. xx. I~c;, It. D. (1924). "Littler Production ~nd t;he sex ra~io in various s~rains of ra~s." Anat. Rec. Vol, xxv~. KLEImV3I(JtI:TER,L. (1871). Die Lehre yon den Zwillingen. Paris. 18~ Hereditary Predispositions to Dizygotic Tzoin-births IEih,~ER, M;. (1920). "]3eit~r. zur morphol, dcr weiblichen Gcschl. organe bei den S~ugetieren. (Uber das Atd~rzten gelber Korper am Ovarium des domestizier6en P~indes u. Schwoines.)" Viert. Jahresb. d. Natl. Ges. Zi~rich, J~in'g. 65. ---(1923). "]3ci~r. zm' Korphologie d. weibl. Geschlechtsorgane bei don Sgnge- tieren." Ibid. Jahrg. 68. LAm~ER, ]3. (1920). "lV[nl6iple births ~mong t~he Chinese." ~u China Rssisw, Vol. ~, L~i[~TZl,]~, W. (/891). Om svangerslcab ogjb'dsd ved tvillinqer. I(jSbhn. LmT~I~, F~K (1917). "The Free-IV[ar~in, ~ S~udy of ~he Action of Sex J:[ormones in ~he Foe~M Life of G~61e." Journ. Ex29. Zool. Vol. xxrcr. Lv~m~o:l~e~, J:L (/912). "Uber die Erblichkei~sverhgl~nisse der kon,%it~utionellen (heredi~tren) T~ubstummheit~ trod einige WorSe tiber die ]3edettttmg der Erblich- keitsforschtmg fth' she t~'alflcheitsletu'e." ArJ~. if. Rass. Gesellsch..Biol. ]3d. Ix. -- (1913). l]Iedizinisch-biologisd~e ]Za,milienCbrschungen innerhalb eines 2232.k@figen Jtaue~'ngeschlechtes in ,S'chweden (lg~'ovinz Blet;i'nge). gen;~. MAgS~, F. It. A. (1901). "Prelim. Communicg~ion on the Oes~rous Cycle and ~he Form~tion of ~he Corpus In~teum in t~he Sheep." P'roc. Roy. Noc. London, Vol. L XV1Ts --(1904). "The Oes~rous Cycle and ~he Forraa~i0n of the 6'orl)us Luteum in ~he Sheep." Phil. Trans. Roy. Nee. London, Ser. ]3, Vol. oxeye. ~m~, H. (1916). "Zur ]3iologie der Zwill~ge." Zeitschr f. C~eb. u. ~2n. ]3d. ]~xxxx, N]~vo~am~, Fmv. (1913). "t~amfistischer ]3eitrag zur Frage der tmgewshnlichen Fruchgbarkeit des Weibes." Zeitsd~r. 3~ ~jn. 1913. N~mzfivs~, P. (1913). ~ber Zwillingsschwa,ngersehaften. Disser~. ~[fmchen. N]~w?zx~ (I917). Biology of Twins. Univ. Chicago Press. 0LIVIng, & (1912). "The tIerer tendency to ~wflufing." The Eugen. t~eview, Vol. Iv. OLSIL~US]~N, 12,. trod VIii'z, J. (1899). Lehrbueh de~' Geb~rtshi~lfe (s. 219): Die mehrfaehe Gebu~'t. ]3onn. P~.~J]~.r~, S. (1896). "Die mel~achen Sehw~ngersch~f~en, etc." Zeitseb'./. (gob. u. ~jn. ]3d. xxxv. P]~5 (1912). "Triplet Calves." Ann. ~ep. Yffann. Agr. Exp. ~qtat. 1912 ~nd 1915. (1915). "~[endelian Inheritance of Fectmd_ity in the Domestic Fowl, and Average Flock Prodncgion." Amer. Naturalist. -- (1917). "Ferti~ty and Age in ghe Domestic Fowl." P~'oo. Nat. Aea& No. Vol. n:r. P~z~, F. (1907). "Die 5rtliche~ gerschiedenheitm~ der Zwillingshgufigkeig und deren Ursachen." Zeitsehr. ,f. Gob. u. (@n. ]3d. LX. Pv~e~r, A. (1874). "Des Accouchemen~s Kultiples en France et dans los Princip~les Coa~r4es de l'Em'ope." Ann. d'tIygiane Publ. Set. ~. T. 4I. -- (1377). "De la l%epe6~ion des Accouchemea~s ~ultiples." Ann. de ~jn. T. 7. Qu]~Nz~ (1898). Die Zwilli~ujsgeburten der Hallenser Univ.-Klinilc u. Poliklinilc yon 1887-97. Dissert. Halle. 1-g]~vss, P. (1872). "Zm" Lein'e yon den Zwillingcn." Arch. f. (r ]3d. Iv. 1-gn~Tz, It. L. and I'r E, (1915). "Degree of l%esemblance of Parents and Offspring with l~especg to ]3ir~.h as Twins for l%egist,ered Sin'opshiz'e Sheep." Journ. Agr. 2esea~.ch, Vol. ~v. .tgo~s, E. (1921). "Fertility in Shropshire Sheep." Ibid. Vol. xxn. J[.(. ]~0NNEVIE AND A. SVEI~DI{UP 185 Ro~R~soN, J. ]3. (1913). "Twimfing in Thoroughbred Brood ~{ares." The Bloodstodc Breeder's l~eview, Vol. II. ROS~N:~'ELD, S. (1903). "Zttr Frage d. vererblichen Anlage zu KelMfllgsgebm'~en." Zeitschr. Gob. u. Gyn. Bd. 50. ~u~r2~ (1891). "Ueber eftfige Unterschiede zwischen oftleiigen ~md zweieiigen Zwillingen." Ibid. Bd. xx~. I~m~rlN, A. (1901). "Die Z~Bings- m Drillings-Geburten in Preussen im letzten JahrzehnC." 1)outs'oh. Mcd. Wodtsc/tr. Jalu'g. 27. S~tSlT~I~, !%. (1901). "Drillingsgebm.ten. Eineiige Drillinge." [bid. Bd. xnw. Soms~mo, ]3. (1912). Zwitti~gs- und i~lehrti'ngs-geb~o'ten in der Kgl. Cha,~td-I~'rauen- l~lini/~ veto 1 Jan. ]905 bls 1 Jan. ]910. Dissert. Berlin. Som~6N, O. (1863). "Beitr. zm" Kennbniss d. Ana~omle u. Physiologie des Eiers~oeks d. S~mge~hiere." Z~dtschr. wis,s. Zool. Bd. xm S~i('2g, J. (1860).' "Studien tiber Zv~ftlge." Zeitsdtr. d. K. K. Ges. d. Aertzte, Wien. Jahrg. 16. S~m~B, Tm v. (1894:). Die meh~fadten Uebzo'teu in ihre~ erblid~eu BezieAungen. Disser~. Basel. STOe~]~]), C~. 1~. (1921). "Developmental l~a~e and Structtu'al Expression: An Experimental S~udy of Twins, Double l'~[onsters and Single Deformgbies, and t0ho Interaction s,mong Embryonic Organs during theft' Origin s,nd Development." Amer. Journ. Anent. Vol. xxv~. $xooms, A. W. (1861). "Sterility in Twin Sisters." Lancet, p. 78. S~'O~O~mL, W, (1899). '"0bet Theflungsvorggnge in Primordi~l-Eiern bei einer E~vaehsenen." A~'d~. Miler. Anat. ]3(]. m~n S~RAss~m% P. (1889). Zur Lehre vo~ der mehrfa~hen Sd~wcm.gersJmft. ])isser~. Berlin. -- (1904). "Die mehrfaehe Sehwangersoh~ft." Winckels Handb. d. Geburtsh. --(1908). "Die Anthropologisehe Bedeuttmg tier ~Ietn']inge." Zeitsdm Ethnol. Vol. x~. S:r6R:~N, E. (1901). "Studier over s,rve]]ghet. Om m'velighetsforholdene red tier- f5dsler." Tidslcr. f. norslce Icege,foren. S~(mv~, A. W. (1917). Des A,~ftreten u. die Vererbung yon Meh~iings-Geburten beim Vollblutpferde. Berlin. V~I% G. (1855). "Bei~r. ztt~ Gebtt~tshfflfl/chen Sta~istik." Monatsdm J'. Geburts. l:~unde, Bd. v-w. WAK~:~Y,~2~. (1895). "The ft~fluence of inheritance on the Tendency ~o have Twins." Lancet, p. ]289. W,~2Ps J. E. (1859). Allg. Bev611cerungsstatistg~. Leipzig. Wm~/~G', W. (1902). "Problemc der N:ehrlings-geb~rtens~atis~ik." Zeitsdm Gob. u. Cqp~. Bd. x~v::. --(1902). "Beitr. zm" Physiologie u. Pathologic d. l~iehr]ingsgebm'~en boim ]~enschen." Pfliggers Arch. Bd. LXXXV]:rr. -- (1903). "Neue BeiCrage zttr Lehre yon den Z~llingen." Zsitschr. f. Geb. u. @n. Bd. x~vm. (1909). "Die Anlage z, ~ekrlingsgeburten beim ~{enschen." Arch. 1Pass. Biol. Bd, w. (1912). "Wei~ere ]3eitrage zm' Theorie des Verorbung." Ibid, Bd. ~x. 186 Hereditary Predispositions to Dizyyotie Twin-bi,rths WEZNBJmC~, W. (1924). "Vererbung u. Ausselffgkgoren bel mensehliehen Zwillingen." Zeitschr. in&d#. Abst. s. 78. Bd. xxxm W~m:, Om~. (1917). fleretning: Under~d'/~'elserovertvillingj?~dslerho~sau. Kris~iania. --(1923). "Trilling- og gvilllngfSdsler hes ssu." Tidskr. Nots#. Landbr. I{. 7. Zw~Er~, E. (1914). "Erfahrungen ~n den legzgen 10,000 (4eburge~ mig besonderer I~er[icksichgignng des Algersbildes." Arch. f. Gyn. Bd. o:[. EXPLANATION OF PLATES XV--XVII. P;r~s. XV, XVI. Pedigrees of ~win-fami]ies, Fam. 2 frem l~feldal (Pl. XV) and Faro. 2 frem Gmmebu (i~1. XVI). Only ~he "twia-lhies" (genegical lines leading te twin-bit~bl~s) ~re demonstrated en these pedigrees, while for individuals not entering into these lines, abe po,s~er~ty has been emitted. The full pedigrees are kept hi "Insgigut for Arvelighetsferslming" ef the Uni- versity, Xristia~da. The deoimM-system used fer fraternities and ind[viduals will scarcely need any explanation. All gwlmmethers are en the pedigrees marked with black, while for their parents and earlier ~neesgors markings of relationship with other twin-mothers have been used as explaflled on Table VII, p. 157. A nete ef interrogation means ghsg the person in question belongs to an un]mewn family, or that his ancestry has not been cleared up for so much as thi'ee generagiens. ~Years of birth h~ve, as a rule, been noted for twins '.rod for gwin-meghers only; the figures added ge all t~win-pab's, ~bove the llne cennecting the twins, means the number ef each twin-birth in the series of births of the fraternity. :Figures in ( ) added, obliquely, to some persons married into the family me,ms the number of the fraternity te which the person belongs. Such figures have ~og been adcled except when this fraternity belongs go ene of the twin-families of ]?.Is. XV and XVI. When on P1. XV1 (2, 111 and 2, 1121) reference is made to M. 2, 3213 and ~I. 2, 5313, this means the ~Ieldal family, P1. XV. Some names of the prhmip~d farms on which the gwin-famih'es have been, or are, living will be feund added te special branches of the f~milies. To the genneba family of PI. XVI a, an annex (P1. XVI b) has been added, as containing abe ancestry of several persons married inge the fanfily mentiened. Ps. XVII. Pedigree shewing the gwin-1)ranehes of the Bleldnge family, based upon t~ rearrangement ef the very extensive eriginM pedigrees of LundbeN (I913). As in Pls. XV, XVI aN descent from persons ne~ ef gene~ical flt~eres~ fer twin-births has been lefg out. The usual deeimal-desigalation of fraternities has on this pedigree been enfigged itx order te facilitate centreI of eur rearrangemenf~--fer the designations used by Lundberg himself; XLV 3 means, for instance, faro. 3 on Pl. XLV ef Ltmdberg's Atlas. Even here ghe marldngs ef relationship used on Pls. XV, XVI and explMned on %~ble VII, p. 65, have iu a few eases been intredaeed en the pedigree. K. }3011Nl#VI~ ~tN9 A, ~Vl~I~I)I, UP ].87 - ~:2 ' I N i I N .i:- ...... I ~ I I ~ I I i I ,-,o 4-io- ,- - 4- - ~l,,:, - - !-I ,~.': ~ i~" - i ~-,~ ~- - ~o ~1' ~ o - -- I -- i ' ' _ ;2 i.l' - I 14" I .,~: ~ ~ .~1 ~._ ~ N -"I) ~= _~. ,; ~.~- .., ~ o ~ ~1~ ~*o- I~,~ ~_ ~, ,~_ ~ 2- I N ~Y,N -~ ~, ~ F,~I~ 4- U.-, ,.,q '.% ~#1# S ~.~'_~ ~.~ ~ 2- ~ -4-4- -~ ~ ,.~ ,". _--2"- # a- .,-.. = ,.4 ~4-1~ I' o~1r 'oo~o~ . . bo, o~ I. I_, 'f3 c,,-'b ~ ~..,~} 23 ,. o ,. § u.oo g- ,.~ ~,.,~,q.-.o3~ d.od,. "q'"%" Cq, 21 188 .f:[eredita~'y _Predispositions to Dizygotic Twi~'~-births 'n,3 'l. o,:b~ 0 0 (D ~c~o olo o~o oo ~ i I~ 0 e i~ 0 Q ,:~o o ooo o(F:~ e io oo I~lr o o -- .~:~ o -o. r,*( v :-: ~,", u ~ --~ 4 q+ I. ~. '~" i .). 0j ~ I ~-~ ~.~ 0## c-3 4-~ 0) "r~ob0 b-t m'-d~ d i ~- 4-~-~ ,~-~ ,~_ ,I. -~i- ~q ~ ~-'~ --I" ~"~ -i- c~,~ ~---- ~ -I-l~ ~ ~c~ ~r ~ ~__ ~ ~'~_i ~ ~l ~'~ ~ "'r "" ~ - ~ -~'i ~ ~ ~"' '-I {. .,~ 4" ~N + , ! ~ o Oo , ro o~ o~" . ZO O~ Os ,-~ c.J I i @ o 0.) >< k-- ~ ~ -~.~-- ~ ~I ,I,I 13., 04 c~ Z X _3 0 c.:.3 I-- Z !/l.J... o o ~?- ~-I~ ~o-. ~-L~f ~ ~ "1 ~ I~ ~" ~ ~/ ~ I + ~-I ~ ~o-I ~, ...... ~-J .~ ~ ~ 0 Do z t- O -11 0"3 prl m CO CO 0 ir- X 0 bo ~D r- -4 m X ~o b~ LLI _J c~ 0 C Z i-" t o F Z m --t o ! ~d t" Z 0 = I F" -.-I m ) !