186 The Genetics of " Rogues " among Culinary Peas ( Pisum sativum ) .
By W. Bateson , M.A. , F.R.S. , and Caroline Pellew , Student of the John Innes Horticultural Institution .
( Received January 5 , 1920 .
) In 1915 we described observations and experiments on rogue peas .
Our results up to the end of 1914 were as follows:\#151 ; ( 1 ) Rogues arise sporadically from self-fertilised seeds of various races of typical high-class peas .
( 2 ) These rogues are characterised by pointed leaflets and upward-curving pods .
Their foliar organs , stipules , leaflets , sepals , petals , and carpels are all narrower than those of the types .
It is especially characteristic of rogues that their leaflets end in points instead of the emarginate apices proper to the types .
Their seeds are , on an average , slightly smaller .
Such plants have a curiously wild appearance , which contrasts greatly with the ampler proportions of the types .
( 3 ) Rogues , self-fertilised , produce rogues exclusively .
( 4 ) Rogues crossed with types , whichever way the cross be made , give plants which as seedlings show evident indications of the type-characters , having parts much larger than those of rogues at the same age .
But these plants at an early stage , usually at some node below that at which the first flower is borne , change to rogues , producing stipules , leaves , and eventually pods , like those of rogues .
( 5 ) The offspring produced by the self-fertilisation of these Fi plants are exclusively rogues .
In some way unexplained the type-characters contributed , whether by the father or by the mother , are excluded at an early stage , so that they do not reappear in the germ-lineage .
In the variety which we have chiefly studied , Sutton 's Early Giant ( a strain of the well-known Gradus ) , plants intermediate between types and rogues are not uncommon .
We have scarcely ever seen such plants in other varieties .
In them a change in the foliar parts occurs like that described in the Ei plants ; but though most frequently completed at or near the first flowering node , their transformation may be deferred to a much higher level .
On such plants the shape of the pods also often changes progressively , lower pods being straighter and upper pods curved like those of rogues .
These intermediate plants form a fairly definite class , though they grade insensibly into both the types and the rogues .
In recording breeding results we have classified the plants into a graduated series of five groups , but for the present Genetics of " Rogues " among Culinary Peas ( Pisum sativum ) .
187 they may be reckoned as of two kinds : .
( 1 ) those in which none of the leaflets have points ; and ( 2 ) those in which pointed leaflets sooner or later appear .
This in practice forms a convenient distinction which can almost always be applied .
The two groups thus distinguished on somatic characters differ essentially in their genetic properties : " non-pointed " plants give on self-fertilisation what in previous publications we have called A families , consisting predominantly of types , or at least non-pointed plants , but the " pointed " give B families containing predominantly rogues .
Families in which the proportion of the two kinds approaches equality occur but rarely .
Intermediate plants of composite or mosaic nature are not very uncommon , and the mosaicism may be of varying extent .
In extreme cases , a whole branch may differ from the main stem , and a " pointed " branch giving a B family may arise from a " non-pointed " main stem giving an A family , or vice versa .
The adequate discussion of these irregularly composite plants must be deferred .
We have been chiefly occupied with a study of the genetic peculiarities connected with the gradual and regular assumption of the rogue characters by the intermediate plants which sooner or later acquire pointed leaflets .
From the fact that the change was progressive , we surmised that if the position of origin of the seeds on such plants were determined , it would be found that the few non-pointed plants which they produce would be found to come predominantly from seeds in the lower pods , and the pointed from the upper pods .
As the result of a preliminary trial we found that this rule was in fact followed , and this observation was published* at the end of 1915 .
Further experience has abundantly confirmed it .
But the converse conclusion which we also announced at the same time , to the effect that in the case of A families containing only a small number of rogues these come predominantly from the upper pods , has proved to be erroneous .
In families where the pointed plants are proportionally few , the seeds from which they come are not derived from any particular level on the parent plants , and the observation on which we relied was only an accident of the special case .
Since the publication of these first results the work has been continued along several lines .
Having established the fact stated above , that in the B families the non-pointed plants come chiefly from the bottom pods , we endeavoured to ascertain more exactly the positions from which they do come .
Choosing , therefore , a number of intermediate plants which we thought likely to throw B families , we emasculated each flower on the stems in succession , as far as possible , pollinating them from type plants * ' Proc. Roy .
Soc. , ' B , vol. 89 , p. 174 .
188 Prof. W. Bateson and Miss C. Pellew .
Genetics of , belonging to families which had been selected as being the most rogue-free.* Since , thus fertilised , every egg-cell genetically rogue , will produce a rogue , and every egg-cell genetically type will produce a type , the resulting family , sown in order , will show the distribution of the rogue- and type-characters among the egg-cells of the intermediate parent .
The seeds from each pod were sown separately , and we expected to find that the non-pointed would come almost exclusively from the lowest pods .
This expectation , however , was not borne out by the result , for the proportion of non-pointed among the egg-cells did not diminish noticeably at least up to the tenth flower .
Since the same plants , if self-fertilised , very rarely produce a non-pointed plant from flowers higher than the fourth , a presumption was raised that the male and female sides must be differently constituted , and that the proportion of pointed gametes would be found to change with the level much more rapidly on the male side .
Tests were therefore instituted , using pollen of the successive flowers of the intermediate plants and trying it on the ovules of types .
The experiment has not been made on a scale sufficient to determine the true average ratios of non-pointed to pointed in successive flowers , but that the proportion of pointed increases rapidly from about the third flower is clearly proved .
Work of this kind is exceedingly laborious , and the wastage from various causes must inevitably be very large .
Of the plants chosen for operation some necessarily prove to be of the wrong kinds , throwing , for example , all rogues and all the work done on such plants turns out in the following year to be inapplicable .
We considered ourselves fairly fortunate in having raised 1016 cross-bred plants from which to estimate the output of the female side , and 468 plants derived from the male side of the pointed intermediates , together with 1572 plants from flowers of known position self-fertilised.f The results are set out in Table I , where the actual numbers from each flower are given with the ratios deduced from them .
As to the female side of the flowers the evidence may be regarded as significant up to the 14th flower .
The proportion of non-pointed female gametes remains about constant at a ratio of rather more than 1 in 2 up to the tenth flower , above which it diminishes .
On the male side the proportion of non-pointed gametes is about 1 in 5 for the first two flowers and then drops sharply and * The proportion of rogues ( or , more strictly , pointed ) thrown by ordinary types is , in our experience , about 2 per cent. In the purer type-families selected for crossing the percentage would be even lower than this , so that for practical purposes they may be regarded as rogue-free .
t An attempt was made on a considerable scale , by sowing the peas of each pod in their original order , to test whether there is any orderly arrangement of the seeds .
This work , which added greatly to our task , led to no positive result .
" Rogues " among Culinary Peas ( Pisum sativum ) .
189 progressively in succeeding flowers , falling again suddenly after the fifth flower .
We are satisfied that the record for the eighth flower is quite abnormal , and may be disregarded .
Further observations will no doubt modify the figures here given , but the general trend of the results is unmistakable .
Table I.\#151 ; Offspring of Pointed Intermediate Plants ( selfed or crossed with type ) grouped according to the ordinal position of the flowers from which the seeds were taken.* Ordinal No. of flower .
Cross-fertilised .
Self-fertilised .
Type-father .
Type-mother .
Non- Pointed .
Ratio .
Non- Pointed .
Ratio .
Non- Pointed .
Ratio .
pointed .
pointed .
pointed .
i 57 54 1 :1 *09 9 42 1:4-6 18 269 1:15-0 2 30 42 1-4 17 83 4-9 21 240 11-4 3 65 46 0-7 8 56 7-0 18 231 12 -8 4 56 64 1 -1 3 30 10 -0 15 228 16:2 5 32 46 1'4 7 86 12 -3 4 218 54 -5 6 26 37 1 -4 2 59 29-5 \#151 ; 144 7 48 43 0-9 1 23 23 -0 1 91 8 34 41 1 -2 6 28 4-6 \#151 ; 71 9 34 46 1 -3 \#151 ; 8 10 25 27 1-1 11 10 31 3*1 12 15 23 1-5 13 10 23 2*3 14 5 27 5 4 15 2 8 16 3 3 17 2 1 1 454 562 1:1-2 53 415 1 : 7-8 77 1492 1:19 -4 The series on the female side is much longer , because the fertilisations were largely done in 1916 , a year in which the plants grew very tall .
The height of a plant depends greatly on the weather , and in 1917-19 short plants were the rule .
Naturally it can only be on tall plants that the contrast between the composition of the offspring derived from the lower and * The material used in this Table is necessarily somewhat heterogeneous .
It includes-offspring of plants which , whether selfed or crossed , have given at least one plant as high as Class 3 .
To have extended it lower would have involved the inclusion of families from plants giving all rogues , and such plants could not exhibit the gradational effect which would thus have been to some extent masked .
Nevertheless , a few of these ought , no doubt , to have been included as genetically comparable .
The figures given for the results of self-fertilisation are not all derived from the plants actually used for the crosses , but include any results from flowers of known position borne by comparable plants .
190 Prof. W. Bateson and Miss C. Pellew .
The Genetics of the upper pods can be fully displayed .
In Table II a number of such families are set out .
To save space and labelling the pods were generally Table II.\#151 ; Families produced by Pointed Intermediate Plants ( self-fertilized ) which gave a small minority of non-pointed members .
Results tabulated according to the flowering nodes from which the seeds were taken .
Register No. of parent .
Ordinal No. of flowers .
Non-pointed .
Classes .
Pointed .
\#151 ; i. 2 .
3 .
4 .
5 .
3676 Main stem 1-7 _ 1 [ 28 8-11 \#151 ; \#151 ; \#151 ; \#151 ; 27 12-17 \#151 ; \#151 ; \#151 ; \#151 ; 28 Branch from 7th node 1-5 \#151 ; 6 1 2 22 6-8 \#151 ; 1 \#151 ; \#151 ; 20 9-15 \#151 ; \#151 ; \#151 ; \#151 ; 39 413/ 6 Main stem 1-4 5 !
5 9 5-8 \#151 ; 1 \#151 ; 13 9-12 \#151 ; \#151 ; 1 \#151 ; \#151 ; 16 13-15 \#151 ; \#151 ; \#151 ; \#151 ; 16 Branch from 9th node 1-3 \#151 ; \#151 ; 6 \#151 ; 8 4-6 \#151 ; \#151 ; \#151 ; 17 7-9 \#151 ; \#151 ; \#151 ; \#151 ; 22 1784/ 6 1-4 1 3 12 5-9 \#151 ; 1 3 \#151 ; 21 1 10-14 \#151 ; \#151 ; \#151 ; \#151 ; 25 1787/ 6 Main stem 1-4 4 16 5-8 \#151 ; \#151 ; \#151 ; \#151 ; 22 9-15 I \#151 ; \#151 ; \#151 ; 26 Branch from 8th node 1-7 | \#151 ; 4 2 \#151 ; 33 8-11 \#151 ; \#151 ; \#151 ; 15 192'/ 6 1-6 4 2 11 7-9 \#151 ; \#151 ; \#151 ; \#151 ; 22 10-14 \#151 ; \#151 ; \#151 ; \#151 ; 21 15-19 \#151 ; \#151 ; \#151 ; \#151 ; 24 2053/ 6 1-5 1 1 3 18 6-9 1 2 2 \#151 ; 23 10-15 \#151 ; \#151 ; 2 \#151 ; 35 258]/ ( 3 Main stem 1-8 1 3 31 9-13 \#151 ; \#151 ; \#151 ; \#151 ; 28 14-21 \#151 ; \#151 ; 1 35 Branch from 8th node 1-2 1 3 \#151 ; \#151 ; 8 3-7 \#151 ; \#151 ; 1 1 36 3492/ 6 1-7 3 4 7 \#151 ; 11 8-13 \#151 ; \#151 ; \#151 ; \#151 ; 33 14-15 \#151 ; \#151 ; \#151 ; \#151 ; 10 278\gt ; /6 1\#151 ; 5 _ 7 8 \#151 ; 11 6-9 \#151 ; \#151 ; 6 \#151 ; 25 1 10-16 \#151 ; \#151 ; 1 1 2 1 11 " Rogues " among Culinary Peas ( Pisum sativum ) .
191 Table II\#151 ; continued .
Classes .
# Ordinal Register No. of parent .
No. of Non-pointed .
Pointed .
flowers .
1 .
2 .
3 .
4 .
5 .
256/ 7 1-2 2 2 .
6 3-5 \#151 ; \#151 ; \#151 ; \#151 ; 15 6-7 \#151 ; \#151 ; \#151 ; \#151 ; 17 112717 1-2 3 3 1 1 1 3-4 \#151 ; 3 \#151 ; 4 1 5-6 \#151 ; \#151 ; \#151 ; \#151 ; 9 7-8 \#151 ; \#151 ; \#151 ; \#151 ; 3 18377 1-2 1 \#151 ; \#151 ; 14 3-4 \#151 ; \#151 ; \#151 ; 1 12 5-7 \#151 ; \#151 ; \#151 ; \#151 ; 16 1855/ 7 1-2 2 4 2 1 3-6 1 3 10 5 9 7-8 \#151 ; 1 3 3 3 9 \#151 ; \#151 ; \#151 ; 1 3 26577 1-3 5 2 7 4-5 \#151 ; \#151 ; \#151 ; 1 7 6-7 \#151 ; \#151 ; \#151 ; \#151 ; 8 45377 1-2 2 3 1 3 3-4 \#151 ; \#151 ; 1 1 7 5-7 \#151 ; \#151 ; \#151 ; \#151 ; 9 457V7 1-2 _ 1 9 3-4 \#151 ; 1 \#151 ; \#151 ; \#151 ; 11 5-6 \#151 ; \#151 ; \#151 ; \#151 ; 5 7-8 \#151 ; \#151 ; \#151 ; \#151 ; 12 9-11 \#151 ; \#151 ; \#151 ; \#151 ; 5 25478 1 __ ___ 1 6 2\#151 ; 3 \#151 ; 1 1 1 8 4 \#151 ; \#151 ; \#151 ; 1 8 5-6 \#151 ; \#151 ; \#151 ; 1 15 7-10 \#151 ; \#151 ; \#151 ; 1 22 3734/ 8 1 _ 1 1 1 1 3-4 \#151 ; \#151 ; \#151 ; 9 5-6 \#151 ; \#151 ; \#151 ; \#151 ; 14 7-8 \#151 ; \#151 ; \#151 ; \#151 ; 11 55678 1-3 1 4 1 2 12 4-5 1 1 4 7 6\#151 ; 8 \#151 ; \#151 ; 2 2 10 60378 1 _ 7 2-4 \#151 ; 1 \#151 ; \#151 ; 13 5-8 \#151 ; \#151 ; \#151 ; \#151 ; 19 6219/ 8 1-2 1 13 3-4 - \#151 ; \#151 ; 13 5-9 \#151 ; 1 22 192 Prof. W. Bateson and Miss C. Pellew .
The Genetics of grouped into three or four successive lots .
This was done before we set about determining the actual ratios for each flower , and though the numbers are not available for that purpose , they provide a larger total of observations than could have been reached had the material been more subdivided .
It will be noticed that a progressive decline ( towards the rogues ) can be often seen in the offspring of branches as well as on the main stems .
In Table II the plants are classified in the five classes which we have arbitrarily used .
Class 1 is the true type ; Class 2 differs from it slightly and evasively in foliage , but chiefly in having some curvature in the pods ; Class 3 are the pointed intermediates of the kind which show the gradational changes ; Class 4 contains plants below Class 3 and approaching Class 5 , which last are the genuine rogues .
Of these classes , 1 and 2 are non-pointed ; 3 , 4 , and 5 are pointed .
There is a complete series of intergrades , both in somatic appearance , and in genetic constitution as indicated by the families produced .
For Table II we have chosen those families which contained at least one non-pointed plant .
The progressive decline towards rogues affects the lower classes , and not merely the frequency of the non-pointed plants .
The same feature is also shown by the families containing only pointed plants , which we have not included in this epitome .
It appears also in the distribution of the several pointed classes resulting from crossing the intermediates as female with type , the pointed plants thus produced by the lower pods being chiefly Classes 3 and 4 , the corresponding plants from the upper pods being chiefly rogues of Class 5 .
This feature , which is not shown in Table I , was most conspicuous when the tall intermediates were used .
Inasmuch as the relative proportions of the several kinds of offspring produced by Class 3 plants ( and doubtless by Class 4 also ) depend on the absolute height of the parents , it is impossible to give any significant figure for these proportions , but the averages produced by the self-fertilisation of the several classes are roughly as follows :\#151 ; Class 1 Non-pointed 47 to Pointed .
1 " 2* 12 1 " 3 1 10-20 " 4 1 160 " 5 0 All rogues .
* Class 2 in our first paper ( 'Jour .
Genetics , ' vol. 5 ) was not the exact equivalent of that subsequently adopted .
We had not then appreciated that the points are the best criterion , and we previously included in Class 3 plants which we now know to belong genetically rather with Class 2 .
The inclusion of such plants in Class 2 , of course , lowei 's the average offspring of that class .
" Rogues " among Culinary Peas ( Pisum sativum ) .
193 Since plants can be found giving every proportion of non-pointed to pointed , the actual ratios observed in any group of plants mean little unless the group can be accurately defined .
Such definition on somatic characters is not strictly possible .
In choosing plants for crossing , we took those which , judged on their early stages , might be expected to throw a small minority of non-pointed plants , for our object was to examine the gradational effect , which can only be manifested by plants of that composition .
In Table II , however , we have plants chosen when ripening , on their adult characters , selected further on account of their being well grown , and the result is that these , as tested by their offspring , proved to be of a somewhat higher constitution than those chosen for crossing .
For any plants which did eventually show the rogue " points " are not too high to show the gradational effect , though judged as young plants , before they had assumed the points , they would not have been known to be available for the purposes for which the crosses were made .
The higher the level at which the points appear , on the whole , the higher is the proportion of non-pointed plants found among the offspring , but this correspondence is rough , and only becomes apparent when long series of families are examined .
Recapitulation .
The three chief phenomena may be recapitulated:\#151 ; 1 .
Eeciprocal crosses between type and rogue give plants which , as they develop , turn into rogues .
2 .
Though the characters of the type are certainly introduced , manifesting their presence by affecting the form of the young Fi plant , they very rarely* take part in the germ-lineage , being apparently left behind in the lower nodes .
3 .
Plants really intermediate between type and rogue nevertheless exist , but never breed even approximately true .
Their germ-cells may be either type , intermediate ( of at least two kinds ) , or rogue .
The proportion of gametes carrying type-characters is different on the male and female sides .
In both sexes the ratio shows gradational change .
Of the egg-cells of 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 character\#151 ; above which level the proportion declines .
Of the pollen in the lowest two flowers , only about 20 per cent , is type-bearing , and the proportion diminishes rapidly in each successive flower above this level .
* The exceptions mentioned in our first paper ( 'Jour .
Genetics , ' vol. 5 , pp. 29-30 , 1915 ) , should probably be thus interpreted .
A few others have since been met with which hereafter will need detailed description .
194 Prof. W. Bateson and Miss C. Pellew .
Genetics of Discussion .
Features somewhat comparable with the first two of these peculiarities have been recorded , but the third is , so far as we are aware , as yet without parallel .
Biffen* has observed the total " suppression " of a character , grey .chaff , introduced by Rivet wheat , in a cross with Polish ; and , in variegated Capsicum , crossed with green , Ikenof never recovered the green in later generations .
BaurJ interpreted certain wholly white and wholly green offspring obtained in crosses between white-skinned and green plants by the very probable suggestion that the green or white characters might have been omitted by somatic segregation , though the inference was scarcely capable of direct proof .
wings also saw something possibly analogous in Humulus , traces of variegation appearing in lower leaves of plants , which subsequently became green .
None of these examples , however , are strictly comparable , but Winge 's case perhaps comes nearest .
A gradational change in the numerical proportions of the gametes at the successive nodes has not , we believe , been elsewhere observed .
The distribution of the type gametes borne by the pointed intermediates corresponds , no doubt , in a loose way with the distribution of vegetative vigour .
But weak growth on type plants is not specially likely to bear rogues , nor do rogues , however luxuriantly growing , produce types .
Aetiological interpretations of this kind are inconsistent with all that we have learnt of genetic principle .
Nevertheless , the coincidence may not be without significance .
Where the output of the several sorts of gametes is so uncertain no gametic system of a Mendelian kind can be propounded .
The most that can be expected of such a system in this case is that it should qualitatively represent the distinction between the genetic nature of the classes here called 2 and 3 .
Since type ( T ) and rogue ( R ) gives Fi rogue , neither of these intermediates can receive a rogue gamete in fertilization .
Since also these two classes differ from the type , the gametes composing either of them cannot both be typebearing .
From this reasoning it appears to be practically certain that two sorts of intermediate gametes must exist , the one more type-like , T ' , the other more rogue-like , R ' .
TT ' will then represent Class 2 and TR ' Class 3 ; but where , as in this case , there is no clear discontinuity no analysis can be pressed .
All that can be positively stated is that there are two sorts of intermediate gametes , that both are unstable , being incapable of constituting a stable zygote , * * * S * 'Jour .
Genetics , ' vol. 5 , p. 227 ( 1916 ) .
+ Ibid , .
, vol. 6 , p. 201 ( 1917 ) .
+ Especially ' Zeits .
f. ind .
Abstammungslehre , ' vol. 4 , Heft 2 ( 1910 ) .
S 'C .
R. Lab .
Carlsberg , ' vol. 14 , No. 3 , p. 11 ( 1919 ) .
" Rogues " among Culinary Peas ( Pisum sativum ) .
195 .
and that when united in fertilization with a type gamete , the resulting zygote is nearer the type than any combination of intermediate gametes would be .
Class 2 plants throw perceptibly more pointed plants than do the real types -(see p. 192 ) but they do not show the gradational phenomenon as Class 3 plants do .
Evidently the gradual extrusion of the type at the successive nodes must be a process similar to that by which the type is extruded in the base of Fi , differing only in that it is prolonged over a longer series of nodes .
The actual data suggest that the gradational phenomenon occurs with greatest intensity in the more rogue-like of the intermediate plants and is less well marked in those which , judged by their offspring , are higher ( more type-like ) in genetic composition , being absent altogether in Class 2 .
Discussion of these points cannot be given in brief .
As it seemed possible that the types might be tetraploid , counts of .chromosomes were made by Miss Nesta Thomas , but the number found in both rogue and type was the same ( seven for the haploid number).* The persistent recurrence of rogues among the offspring of types must indicate some liability to an error in cell-division .
Once the abnormality has .occurred , of which pointed leaflets are the ostensible indication , there is a progressive change in successive generations such that , assuming equal fecundity in all classes , the progeny would in a few generations consist of rogues in overwhelming proportions .
It was a commonplace of practical \#166 ; breeders and of conventional evolutionists that when selection is suspended , a breed " degenerates .
" This doctrine , promulgated , as it commonly was , without any reservation as to crossing or reference to critical purity of line is fallacious as an expression of physiological truths however much the objective .consequences may seem to fulfil the prophecy .
In the present example the popular conception of degeneration is precisely realised .
So far as we know it unique .
* Since confirmed by Prof. K. Matsui .

