358 ' Croonian Lecture :\#151 ; Genetic Segregation .
By W. Bateson , M.A. , F.R.S. ( Received September 20 , \#151 ; Lecture delivered June 17 , 1920 .
) The latqr developments of Mendelian analysis have been in the main an attempt to elucidate the scope and nature of Segregation .
Mendel proved the existence of characters determined by integral or unit factors .
Their integrity is maintained by segregation , the capacity , namely , to separate unimpaired after combination with their opposites .
Our first aim has been to discover specifically what characters behave in this way , whether there is any limit to the scope of segregation , or any characters or classes of characters which are determined by elements unable to segregate simply .
The second object has been to decide the time and place in the various life-cycles at which segregation occurs .
It is with the latter problem that I propose to deal more particularly in this Lecture , but a brief consideration of the range of characters , subject to segregation , is appropriate .
1 .
The Scope of Segregation .
Of the classes of features by which animals and plants are distinguished , most have now been shown to be dependent on segregable elements .
It is perhaps noticeable that we have no quite clear proof that the factors governing differences in number , meristic characters in the strict sense , commonly behave so simply as those determining other characteristics .
There are examples of parts repeated in series , such as the extra toe of the fowl ( a dominant ) and the leaf of the monophyllous Strawberry with a single leaflet ( a recessive ) which have a factorial inheritance , but the resulting terms , especially the heterozygotes , are indefinite .
In the polydactylous fowl , for example , the heterozygote has a great variety of shapes .
The hind toe is most often represented by two sub-equal digits , but the duplicity may be so slight as to appear externally only as a division of the nail .
It may also assume another very different form , in which there is only a single many-jointed digit representing the usual pair .
In the monophyllous Strawberry the homozygous recessive , whether before or after immediate extraction from the heterozygote , shows fluctuations to a 2- or 3-leaved conditiqn.* Such fluctuations are common among forms distinguished by meristic peculiarities .
There is not the uniformity and simplicity which is such a striking feature of * Richardson , 0 .
W. , 'Jour .
Gen./ vol. 3 , p. 171 ( 1914 ) .
Genetic Segregation .
variations in colour and many other substantive characteristics .
The evidence as to meristic differences is however still scanty , and it is too soon to decide what importance should be attached to this preliminary impression .
With more confidence we recognise that merely quantitative differences seldom , if ever , have a perfectly simple inheritance .
There are two obvious interpretations : ( 1 ) that the factors do not usually segregrate clean ; ( 2 ) that the number of factors involved is so large that their effects are masked .
Adequate discussion of these two possibilities could only be given at considerable length .
On the whole , I incline to the former alternative , but the material for a decision scarcely exists as yet .
Certain examples should be noticed in which , though the most obvious differentiating feature is quantitative , the underlying physiological distinction is more probably to be referred fo a qualitative factor .
Height in certain plants is a good illustration .
It is ostensibly a quantitative feature , and the tails segregate clean from the dwarfs .
But in various cases , Peas and Sweet Peas , the dwarfs are also a darker green .
The dwarf of Campanula especially ( about 8 inches high ) , is a plant so strikingly different from the type ( 2-3 feet high ) , that it is sold as a species " C. The leaves of the dwarf are an intense , dark green.* This variety is , of course , a recessive and segregates without intermediates , yet , if the qualitative distinctions were less evident it might easily be classed as a variety in quantity .
But the critical distinction is certainly qualitative and the great difference in size is consequential .
Though in such cases segregation is complete , it may almost be said to be characteristic of purely quantitative distinctions that one or other of the original parental types fails to reappear in its extreme form after a cross .
So usual is this feature in quantitative segregation that the phenomenon must have special significance .
Another fact is beginning to emerge which must contribute to the shaping of a conception of the physiological nature of segregation .
We have learnt that groups or complexes of factors may segregate whole .
To such a complex the distinction in sex is due , but in certain cases it may break up .
The occurrence of a large spur in fertile hens ( Leghorns , for instance ) must be regarded as due to the absence of that part of the sex-complex which in the normal hen inhibits the growth of the spur .
In ordinary fowls the whole inhibiting group remains on the female side , but the spur-inhibiting element can evidently separate from the rest .
On the other hand , when in the cocks of certain breeds ( as occasionally in Wyandottes ) there is no spur , we * The ovary projects in a curious way above the sepals so much , that were the plant not a Campanula , botanists would describe this ovary as superior .
Mr. W. Bateson .
may perhaps conjecture that this element has been transferred to the male side.* The presence of the characteristically masculine comb and wattles in the male Sebright , which is otherwise henny , shows that another factor similarly detachable governs their development .
To the breaking up of large compound factors the production of intermediate types , such as occur among the colour-varieties of plants , is in all likelihood due .
The Sweet Pea and the Snapdragon have now an innumerable series of such colour-forms which may be represented as having arisen by the disintegration of the various anthocyanins .
That , at least , is the simplest way in which their origin can be conceived .
To the final result many qualifying elements contribute , and these may naturally be separate entities .
But change in the amount of the same colouring material , and diminution in the mere extent of its distribution , are common features of these graduated series .
As the cultural development of the species progresses , more and more of these quantitative intermediates appear , and are selected , until a practically continuous series is produced .
Although the interrelations of the whole series can be represented by a factorial scheme , the assumption that each of these factors had an aboriginal individuality appears gratuitous .
In Antirrhinum , for instance , the ordinary self-coloured flower segregates as a single unit from the white .
But there are " Delila " forms having the " face " coloured and the " throat " white .
Another variety has the " lip " coloured and the peripheral parts white , and to this again there is an almost exact counterpart in which the peripheral areas are coloured and the lip nearly white , and between these again there are further intergrades .
Apart from factors modifying its quality , the colour of the whole corolla , segregating as a single entity from the white , would without hesitation be represented as depending on a single factor .
Subsequent experience that this entity can break up into an indefinite number of fractions is not evidence that the original representation was wrong .
This reasoning applies to a great range of phenomena .
In view of the chromosome theory of linkage , it is therefore worth remarking that we do not find linkage especially frequent among these * After much experiment the genetics of spur-development are still very obscure .
In the course of a long series begun by crossing spurred Leghorn hens with a spurless Wyandotte cock ( giving F , normal in both sexes ) , neither form has reappeared in F2 with its original development .
Very rarely a hen with minute spurs has appeared , and occasionally a cock with the spurs sensibly reduced .
Nor from spurred hens xF1(j has anything approaching the original types been raised .
Conceivably the detached element is able to join again with the rest of the female complex .
\lt ; renetic Segregation .
fractional factors .
Have they , then , been distributed among different chromosomes ?
If in Antirrhinum , the colour of the face and of the throat were lately parts of a single factor for the total flower-colour , would not linkage between them be expected ?
Nevertheless , in cases of this sort , so far as I know , linkages have not been found with any special frequency .
The segregation of a group of differences\#151 ; and presumably factors\#151 ; in combination has lately been shown by Renner* to occur with extraordinary frequency in the ( Enotheras , and this peculiarity is without doubt at the bottom of the difficulties which have beset the genetic analysis of these plants .
The complexes are in several forms or species not borne equally by the two sexes of the same plant , and most of them are unable to exist in the homozygous state .
These discoveries greatly elucidate the ( Enothera problem .
For instance , not only Giln .
but biennis , muricata , and others also , are not homozygous types , but heterozygotes of a special kind .
Consequently , the production by them of " mutants " is not capable of the simple interpretation originally applied to them by de Vries .
Renner suggests that the mutants arise owing to some interchange between the complexes which at present we cannot investigate , but whatever be the exact manner of their origin we cannot regard them as genuine examples of the production of novel forms by a homozygous type .
Before leaving this part of the subject , we may notice that the supposition that segregation is concerned solely with characters of a superficial or trivial nature has been long ago disproved .
Baur 's Antirrhinums , the study of which was continued by Lotsy , were an excellent demonstration to the contrary , for they provided many illustrations of segregation in features , the " specific value " of which no systematist would question .
If further evidence were needed , it may be found in the fine series of experiments lately published by Heribert-Nilssonf in which , contrary to the belief long ago entertained by Wichura , show that , when Fa families are raised in adequate numbers , very striking segregation occurs in the species-crosses .
Many geneticists are inclined to the view that segregable characters should be pictured as implanted on an irreducible base which is outside the scope of segregation , but no means have yet been devised for testing the reality of this conception .
2 .
The Moment of Segregation .
The next question is to determine when in the various life-cycles segregation can occur .
Obviously it is a phenomenon of cell-division .
If we * ' Zts .
f. ind .
Abstammungs- und Vererbungslebre , ' xviii , 1917 , p. 121 .
t Lund 's ' Universitets Arsskrift , ' N.F. , Avd .
2 , vol. 14 , November 28 , 1918 .
Mr. W. Bateson .
knew nothing of the genetics of plants we might confidently adopt the view which Morgan has so successfully developed , that normal segregation and redistribution happen exclusively in the process of reduction .
Though unconvinced , I cannot deny that linkage and crossing-over may well be represented provisionally , as effected during synapsis .
The scheme previously offered by Punnett and myself as a diagrammatic plan capable of representing these phenomena is certainly far less attractive .
There is evidence that in certain plants , e.g. , Matthiola , the composition of the families derived from single pods shows very great and perhaps irregular fluctuations , and the normal ratios for those families is only found by taking the average of many , but no sufficient statistical examination of such cases has yet been made .
In some suitable case estimations of the offspring derived from individual anthers might be of value in this connexion .
Renner , by examining the starches of the pollen-grains in CEnotheras , has lately made visible that di-morphism , of which we had previously genetic proof , and perhaps this novel and striking observation might be used for the purpose of mapping the distribution of such a character among the pollen-grains .
Meanwhile , it must be granted that no indication that gametic linkage results from somatic differentiation has yet been obtained .
When , however , we examine the view that linkage of factors is a consequence of their association in a chromosome , we must observe that there is no body of evidence that the number of linkage-systems agrees with that of the chromosomes , a primary postulate of Morgan 's theory .
Drosophila is the only example which has been adequately investigated .
The cytological appearances are not readily consistent with the other postulate of Morgan 's case , that crossing-over is effected by anastomosis of chromosomes and exchange of materials between them .
In our present ignorance of the physical nature of the factors we are not obliged to assume that an actual transference of material is an essential condition for an exchange of properties ; but since Morgan 's suggestion is made in that form we are bound to notice how difficult it is to interpret the visible phenomena of cytology in accordance with that hypothesis .
Without personal familiarity with cytology no one can have a confident opinion .
I observe , however , that in his most recent publication on these subjects , E. B. Wilson* gives a very emphatic " counsel of caution , " remarking that writers on genetics have taken too much for granted , and that for the present the genetic development of the chromosome theory has far outrun the cytological .
To a layman the visible appearance of chromosomes is scarcely suggestive of the prodigious material heterogeneity demanded , and the general course * ' Amer .
Not .
, ' p. 210 , May-June , 1920 .
Genetic Segregation .
of cytological evidence seems to indicate that the role of the chromosomes is passive rather than active .
Though showing features of regularity , they are capable of very wide variations without transgressing the limits of viability , which could scarcely be the case were every detail in their organisation critical .
The appearance of chromosomes is not to me suggestive of strings of beads of extreme heterogeneity , but rather of strands of some more or less homogeneous substance ; and in so far as numerical and geometrical order is exhibited by them , it would , in my opinion , be more proper to compare this regularity with that seen , for example , in drying mud or in the formation of prisms of basalt than to attribute to it a more fu ndamental meaning .
Leaving these speculative considerations , and limiting our inquiry to the concrete question , at what moment in the cycle does genetic segregation occur , we reach a perfectly definite answer : that whatever future research may decide as to the occurrence of segregation in animals\#151 ; which , for aught we know , may always be effected at the reduction division\#151 ; there is no such limitation in plants .
We are now thoroughly familiar with a large group of examples in which the genetic properties of the male and female cells of the same plant are quite different .
In these , at all events , the reduction-division cannot be the moment of the segregation by which these characters are distributed .
The first case detected was in Matthiwhere Miss Saunders ' results proved that in the double-throwing singles the pollen carries exclusively doubleness , the eggs being mixed , some single and some double .
A similar condition was shown to exist in regard to the cream and white plastids respectively , the pollen grains bearing exclusively cream .
De Vries observed a comparable arrangement among the CEnotheras , and Renner has lately shown that the phenomenon is widely spread in that group , thereby providing a consistent interpretation of much that was formerly obscure in the genetic behaviour of these plants .
In Campanula carpatica Miss Pellew proved that the pollen grains of the hermaphrodite form called pelviformis carry exclusively femaleness , and preponderantly white flower-colour ( the plant being heterozygous for blue ) .
The ease of Petunia investigated by Miss Saunders* is somewhat pecidiar in the fact that in the heterozygous singles the male side carries the dominant singleness only , since in those instances to which the conception of dominance can be applied , it is the male which commonly carries the recessive .
Segregation of these characters cannot in plants so organised be supposed to take place later than the constitution of the male and female organs , and therefore the reduction * 'Jour .
Gen. , ' vol. ] , p. 57 ( 1910 ) .
Mr. W. Bateson .
division cannot be the one critical moment .
The suggestion has been made that germ-cells of the missing kinds may be in reality formed and perish before reaching a functional stage .
As regards the ( Enotheras , where shrivelled pollen grains abound , this conjecture is very plausible and probably correct ; but when as in the other cases here quoted , the pollen grains are uniformly sound , the hypothesis is inapplicable and without evidential support .
Moreover , even if it were true that certain classes of germ-cells perish in one or other of the sexes , that would hardly alleviate the difficulty , for this differential viability would remain to be be accounted for , being itself a phenomenon of segregation .
Begonia Davisii* is another curious illustration in which segregation must occur even earlier .
This plant is a wild , true-breeding species , with ordinary single flowers .
All the pollen grains however carry doubleness , and used on the female flowers of doubles give offspring all double ( single being the dominant ) .
The pollen of this plant is as uniform and perfect as those of any species I have ever seen .
We must therefore conclude that the segregation by which singleness separates from doubleness is effected not later than the formation of the rudiments of the male and female flowers .
Cytologieal investigation may no doubt show that the distinctions between the genetic properties of the male and female are associated with visible nuclear differences , but I see no reason for anticipating that such differences must exist .
Cells which differ in their genetic potentialities must of course differ in physical constitution , but there is no reason to suppose that this difference need be in any way dependent on chromosome structure .
As regards Campanula carpatica " pelviformis " and Begonia Davisii experiment has shown that the peculiar kind of segregation which they exhibit may recur in their offspring .
In the Begonia , if the female of Davisii be fertilised with pollen of an ordinary double tuberous Begonia , the doubleness so introduced stays on the male side just as the doubleness of its own male does , and a plant so bred has its pollen all double .
But if the male of Davisii be used on the female of an ordinary single , there is no restriction of doubleness to either sex of the offspring .
The peculiarity of Davisii must therefore be attributed to the special .
properties of its female side .
The Campanula case is complex and has not yet been fully explored , but at least from the female side of pelviformis plants have been raised which retain the properties of the mother as regards the distribution of the white and blue colours .
We have at the John limes Institution been lately investigating a similar case in flax , which though comparable has some special features .
A dwarf * 'Jour .
Gen. , ' vol. 8 , p. 199 ( 1919 ) .
Genetic Segregation .
flax ( Linum usitatissimum ) of unknown origin , presumably a stray seedling of one of the varieties grown for oil , was fertilised with pollen from our tall fibre strain .
Both parents breed true to the fully hermaphrodite condition with anthers perfectly formed , and the Fi plants were normal in this respect .
F2 consisted of hermaphrodites , and a recessive form with aborted anthers , generally contabescent and not dehiscing at all , but having occasionally a few grains of good pollen .
The ratio was a normal 3 : 1 .
The recessive , having-occasional grains of pollen , self-fertilised , gave similar plants with anthers wholly or almost wholly aborted .
The normal F2 hermaphrodites gave in F3 families which showed that some of the F2 plants were pure normals , others heterozygous in the ordinary way .
But when the recessives were fertilised with pollen from three several varieties of tall fibre flax , only recessives were produced .
These tall flaxes therefore are normally heterozygous for the recessive or " sub-female " condition , and this in segregation is permanently relegated to the male side of the plant , while the female side takes the hermaphrodite factor .
Segregation in regard to the same recessive may take place in one of two ways .
It may be unilateral , as it is when in heterozygous association with the female of the tall flaxes , or it may be ambilateral and unrestricted to either sex when it is in association with the female of the oil flax .
We must infer that the female halves of the two types differ in some critical respect which decides the manner of the segregation .
Unilaterality may also show itself as a difference in the closeness of linkage on the two sex-sides of the same plant , and no doubt this fact may have a bearing on the interpretation of the foregoing cases .
The late R. P. Gregory discovered the first case of this , in Primula sinensis .
He found that the linkage between magenta colour and short style was closer in the eggs than in the pollen .
Recent work on a larger scale has given 10'9 :1 as the female linkage and 6'4 : 1 for the pollen .
A similar difference has been also found for the linkage between green stigma and " reddish " stem ( as opposed to dark red ) , the value being 29-8 :1 for the eggs and 41-7:1 for the pollen .
In both examples , individual families show wide fluctuations , and these values should for the present be regarded as approximate only .
Whatever be their meaning , they show that some segregation has occurred in the formation of the two sets of sexual organs , such that the process of gametic differentiation is not the same in both .
Besides these examples of differentiation between the male and female sides , there are others proving that segregation may occur at other stages in somatic development .
The most obvious examples are the variegated plants .
I have discussed this subject elsewhere in connexion with reversible Mr. W. Bateson .
3 ( \gt ; 6 periclinal " chimeras " of white over green which produce shoots having the white enclosed in the green.* To these must now be added the cases in which the plants arising from adventitious buds differ from the plants which produce them .
I have described one of these examples in Bouvardia .
The pinkish white " Bridesmaid " gives the red flowered " Hogarth " from its root-cuttings .
Three similar occurrences have been found in fancy Pelargoniums .
The root-cuttings of a white flowered variety , " Pearl , " give a red-flowered form , very like " Mine .
Thibaut .
" " Mrs. Gordon , " which is a full rose-pink , with whitish edges , gives from its root-cuttings flowers in which the two posterior petals are marked with dark red , not unlike the variety " Cardiff .
" A more striking case is that of " Escot , " which gives from its roots plants with bright pinkish red marks , those of the original being purplish red .
The most curious feature of this case lies in the increased size both of the plant and the flowers coming from the roots , and it is scarcely possible to see the petals of " Escot , " which are characteristically rolled back , side by side with those of the root-form , which are not only larger but also flat , without surmising that this rolling back is an expression of the greater size of the larger petal contained within the smaller , causing a want of correlation between the growth of the inner and outer tissues .
Buckling or crumpling of leaves through want of correlation was a conspicuous feature of some of Winkler 's " graft-hybrids , " made from nigrum and S. lycopersicum , when the larger tomato was enclosed within the smaller species .
We have had a precisely similar example in a salinon-fringed Pelargonium bred by Mr. Jarman , of Chard .
The leaves are obviously buckled , the petals are laciniated , and the female parts aborted , though the anthers are perfect .
This male and deformed flower is proper to the outer tissues only ; for on two occasions the plants have produced shoots with large flat leaves and normal hermaphrodite flowers with their petals entire .
Obviously , this normal plant was enclosed within a skin of the fringed type .
In all these examples , a somatic segregation has occurred which determines the genetic potentialities .
The interpretation that they are periclinal chimseras is probably correct for the most part .
The fringed Pelargonium is obviously of this nature .
Nevertheless , the fact that a root-cutting consistently produces a certain type of plant which is not the original does not prove that the distribution is periclinal .
Another possibility is well illustrated by the case of a variegated Spircea having the * ' Jour .
Gen. , ' vol. 7 , p. 93 ( 1919 ) .
Genetic Segregation .
stem , petioles , and ( basal ) centres of the leaves without chlorophyll.* The growing point has the power of laying down green tissue in the lateral areas only , the internodal regions being albinotic .
Eoot-cuttings from this form give albino plants which die after the development of two or three small leaves .
Now in this case we can see the distribution of the green and white respectively , and we recognise that the roots give albino plants because they belong wholly to the albinotic area .
On similar lines it is possible to interpret the Bouvardiaand other cases .
The distribution of the two types in the same plant may be such that one is limited to the root and internodes , while the other is in the nodal structures .
That considerations of this kind are not fantastical is proved by the genetical phenomena seen in the case of " rogues " in culinary peas , which Miss Pellew and I have been investigating for a number of years , f The rogue is a peculiar , wild-looking plant , differing in various ways from the type , chiefly in having pointed leaflets .
Crosses between it and the type give plants which in their lower parts are intermediate , though turning into rogues as they develop .
The self-fertilised offspring of rogues and also of these Fi plants are always rogues , and evidently the type-characters introduced from the type parent are left behind in the lower parts .
Such a case may perhaps be compared with the condition seen in the variegated and we may fairly conjecture that if it were possible to raise root-cuttings from the Fi peas , they would produce types .
A more gradual exclusion of the type-elements in the lower parts is seen in certain intermediates .
These may scarcely differ from types in the lower parts , though changing to rogues , sometimes abruptly , sometimes gradually , as the series of flowering nodes is developed .
Reciprocal crosses between the successive flowers of such plants and the flowers of types has shown that , together with the gradational change in the somatic structures , there is also a gradational change in the proportion of gametes bearing the rogue and type characters respectively .
This proportion and the rapidity of the change differ on the male and female sides .
Of the egg cells in the lower flowers , up to about the 10th flowering node , rather more than 50 per cent , carry the type-characters\#151 ; or at least the non-pointed leaflets\#151 ; but above this level the proportion of types declines .
Of the pollen in the lowest two flowers only * This is somewhat like the Pelargonium named by Messrs. Cannell " Freak of Nature , " in which the chlorophyll has a closely similar distribution , and it , like the Spine a , is sterile on both male and female sides .
In this I have never seen pollen , but very rarely a fruit is formed , which , no doubt , is due to an occasional development of a bud in the green area , an occurrence frequent in variegated plants .
Whether these fruits contain viable seeds is not yet known .
t ' Roy .
Soc. Proc. , ' B , vol. 91 , p. 186 ( 1920 ) .
VOL. XCI.\#151 ; B. 2 G Genetic Segregation .
about 20 per cent , is type-bearing and the proportion diminishes rapidly in each successive flower above that level .
In all the examples given hitherto the segregation is in diploid tissues , but a comparable phenomenon has been proved by Collins to occur in the haploid axis of a moss ( Funaria ) .
In a dioecious moss , as the Marchals have shovyn , sex-segregation occurs at spore-formation , the division in which reduction is effected .
This , of course , agrees with cytological expectation , though so far as I know , the details have .not been observed .
But from the leaves of mosses placed in nutrient fluids new plants may be raised without great difficulty , and Collins found that the ( perigonial ) leaves surrounding the male organ thus propagated , produce exclusively male axes.* He has since raised similar cultures from the ( perichsetial ) leaves surrounding the female organ , and , as related in the paper following this , from them monoecious plants resulted .
The proof is thus complete that in a haploid tissue a segregation of sex can occur .
The inference may be drawn that the factors for other characters may similarly be liable to segregate in the haploid state .
In this connexion I may mention a case which though as yet obscure , perhaps fulfils this expectation .
In botanic gardens a variegated maiden-hair fern capillus- Veneris ) is grown which has wedges of white tissue irregularly distributed in the segments .
This plant produces spores freely , f and these give rise to prothallia which in several cultures raised here have always been entirely green .
But when ferns arise from these green prothallia by the sexual process , they are of three kinds , green , white or variegated like the parent plants .
The fact that the prothallia should be all green is entirely unexpected and creates a distinct problem , but it is evident that segregation must occur either in some of the cell-divisions by which the prothallia proliferate , or in those by which the gametes are formed , in either case in haploid tissue .
This segregation is essentially different from that by which the differentiation of organs , such as the archegonia and antheridia is accomplished , inasmuch as it relates to elements determining the characters of the next generation .
From the evidence given it is clear that in a wide view of living things segregation cannot be exclusively a property of the reduction-division , and for the present , it should be regarded as a possibility which may occur at any division in the life-cycle .
* ' Jour .
Gen. , ' vol. 8 , p. 145 ( 1918-9 ) .
t I have not satisfied myself that spores are produced in sori on the white areas .

