Hereditas 39: 1-18 (1953)

Åke Gustafsson


THE conception of natural selection, and its consequence the struggle for life, has in the course of time aroused the minds immensely. Here is not the place of analysing or reviewing all speculations, assumptions or conclusions. There is in my opinion no need to deny that the rude method of stamping-out, by which the »unfit» are eliminated from life, still forms a central theme of the evolutionary theory. It was vividly depictured by DARWIN himself. Let me cite from the »Origin of species» (Chapter III): »We behold the face of nature bright with gladness, we often see superabundance of food; we do not see, or we forget, that the birds which are idly singing round us mostly live on insects or seeds, and are thus constantly destroying life; or we forget how largely these songsters, or their eggs, or their nestlings, are destroyed by birds and beast of prey; we do not always bear in mind that though food may be now superabundant, it is not so at all seasons of each recurring year».

In the following thesis not the method of stamping-out what is unfit will be considered, but the cooperation of fit and less fit to make the population, the entirety fitter than the partaking constituents themselves and alone. The mechanisms behind this cooperative increase in reproductive or vegetative fitness are barely touched upon in recent research. In many cases they are of a purely passive nature, making use, for instance, of all available space by means of differences in root and stem development, but now and then there may also occur some sort of a co-adaptation by biochemical means, an active attraction of biotypes or species to one another.

In this paper some examples will be given relating to an intraspecific type of cooperation. The theories and speculations involved in their interpretation will not be discussed.

An agricultural plant like barley forms an excellent material for the study of production capacity, of agricultural fitness, under varied environmental conditions, in sparse or dense stands, with regard to competition within and between strains, or even with other species like oats. The production capacity can be exactly measured in vegetative properties, for instance tillering, or in generative characteristics like number of ovules and seeds, or kernel weight.


The first part of this study deals with the important problem whether mixtures produce more than pure varieties. This is indeed a problem that has interested plant breeders and farmers for a long period of time. It was often considered that mixtures of strains and species, also within the self-fertilizing species, augment the yield per unit area (RÜMKER, 1892; MONTGOMERY, 1912; FRUWIRTH, 1923; ENGLEDOW and RAMIAH, 1930; ENGELKE, 1935; NUDING, 1936; JONES, 1948). This same doctrine has been essential to the Soviet-Russian ideas about competition within and between species [see, for instance, the results and compilations of CEDIK-TOMASEVIC, 1951: »Ausserdem geben die Mischungen von 2 Varietäten (einer Art) grössere Erträge, als gleichzeitig ausgelegte Komponenten im Reinanbau».] However, LYSSENKO, in his report of 1948, states as his opinion that neither competition, nor its contrast, here called cooperation, occurs within a species. (»Ablehnung des Kampfes und der gegenseitigen Hilfe der Individuen innerhalb einer Art.») There is no need to dwelt on this unexpected controversy. It would be fatal, however, to forget about the empirical back-ground of CEDIK-TOMASEVIC's study. A negative attitude towards an improvement by means of blending characterized the conclusions of FRANKEL (1939), who in careful experiments analysed the yield properties of mixtures and pure strains. »In all nine trials the yields of the blends are very closely similar to the expectation ... Only in two out of the nine varieties is there any indication of an increased yield due to blending» (p. 254). CARDON (1921) did not notice any increase in yield after the mixing of wheat, oats and barley. His yield data are however rather irregular.

In order to study the problem very carefully the Golden, Maja and Bonus barleys were tested in 1951 with regard to their behaviour in mixtures and in purity. All three varieties are fairly related. Maja is a product of Golden and the Moravian Binder barley, Bonus a product of Maja x Seger x Opal, all of these derived from Golden crosses. The yield of the varieties runs as: Golden—100, Maja—115,2, Bonus—124,9 (FRÖlER, 1951). The figures refer to Svalöf conditions.

The seeds were sown with the help of marking-boards, in dense or in sparse stands, with two kernels or one kernel per hole of the board, with 200 or 500 kilograms of saltpetre per hectare. After the time of harvest the number of plants, spikes, kernels and kernel weights were determined for every plot, except in the case of two series (1 and 2), where two kernels per hole were sown. Here the number of plants could not be considered. The same set of experiments has been repeated and largely extended in 1952.

The series were: (1) 2 kernels per hole, 5 X 15 cm, 200 kg salt-petre, (2) like 1 but 500 kg, (3) 1 kernel. 5 X 15 cm, 200 kg, (4) like 3 but 500 kg, (5) 1 kernel, 10 X 15 cm. 200 kg, (6) like 5 but 500 kg, (7) 1 kernel, 10 X 30 cm, 200 kg., (8) like 7 but 500 kg. In this way it has been possible to trace directly the influence of density, manuring and mixture on seed production. Every plot consisted of plants from 90 sown kernels, except for plots 1 and 2, which were made from 180 sown kernels. There were consequently eight different plots of every strain and mixture, making in all 48 different plots.

The number of harvested plants is slightly superior for the pure strains, viz. 87,3 per cent of the sown seeds as against 86,2 per cent for the mixtures. The difference lacks significance.

The lumped data on spike and kernel number, kernel weights and 1000-kernel weight are as follows:

  Pure strains Mixtures Difference
No. of spikes 6998 7185 + 2,7%
"   "   kernels 164467 169057 + 2,8
Total kernel weight 6688 gr 6936 gr + 3,7
1000-kernel weight 40,66 41,03 + 0,9

In all examined properties the mixtures are somewhat superior, although the differences are not very striking. The higher number of spikes and kernels work together with the larger seed size to increase production by some four cent.

Maja barley appears to be specially effective. The mixture of Maja and Golden lies 1,7 per cent higher than the best parent, Maja. Similarly, the mixture of Maja and Bonus surpasses the top strain, Bonus, by 2,4 per cent. The third mixture, Golden and Bonus, attains an intermediate position of the parents. In the experiments reported here the relative yields of Golden, Maja, and Bonus run as 100:110:116. The analysis implies that in this case two close-yielding varieties work better in mixture than do two distant-yielding types.

  Difference from
best parent
Difference from
mean of parents
Golden + Maja + 1,7% + 6,4%
Golden + Bonus - 7,8 + 0,1
Maja + Bonus + 2,4 + 5,5

If the materials are analysed in more detail, distinct regularities can be traced (Table 1).

The mixtures (G + M, G + B, M + B) react in a similar manner. With dense sowings the mixtures are superior to the mean of the parents by 10 per cent, if the manuring is low. With high manurings the yield of the mixtures falls considerably (with circa 7%). After sparse sowings and low manurings the mixtures are equal to the strains. With 500 kg manures they are again superior by 10 per cent.

The dense sowings comprise twelve plots of mixtures: six out of series 1 and 3, six out of series 2 and 4. The relative values of series 1 and 3 differ significantly from those of 2 and 4 (P < 0,01). Similarly the sparse sowings comprise twelve plots of mixtures. The six relative values of 5 and 7 differ from those of 6 and 8 (P ≥ 0,05). If we consider all 24 values, distributed into four groups as to density and manuring, there is a pronounced interaction of sowing density, manuring and strain mixture as to render a statistical value of 0,01 > P > 0,001. Consequently, we may state that strain mixtures react in specific manners to the varied types of environment. It also follows that there is a distinct interreaction of genotypes. Under certain conditions the strains oppose, under other conditions they improve upon each other.

TABLE 1. The relative effects of sowing distance and manuring on the production of pure strains in barley and their mixtures. Strains = 100.

  Dense sowings   Sparse sowings
Seed distance 5 X 15 cm 5 X 15 cm   10 X 15 cm 10 X 15 cm
Kernels/hole 2 2 1 1   1 1 1 1
Manuring,kg/ha 200 500 200 500   200 500 200 500
Series 1 2 3 4   5 6 7 8
All mixtures/strains 105,0 92,5 115,0 93,3   101,0 112,6 100,6 107,3


  Dense sowings   Sparse sowings
Manuring, kg/ha 200 500   200 500
Series 1+3 2+4   5+7 6+8
G + M/strains 108,2 87,7   106,1 113,8
G + B/strains 103,7 94,4   92,5 108,8
M + B/strains 118,3 96,6   103,0 106,0
Average 110,1 92,9   100,5 109,5

(G = Golden, M = Maja. B = Bonus barley.)

It is of fundamental value to know with certainty what lies behind the different production capacity of the mother strains. The average tillering of the strains amounts to 4,2, 4,1 and 4,3 spikes per plant, i. e. they lie on the same level. The increase in production rather depends on the higher number of kernels per spike and plant (84 → 89 → 96 kernels per plant) and the higher 1000-kernel weight (43,8 → 45,8 → 46,8 gr). These two properties work together in such a way as to raise the production of Maja barley and even more of Bonus barley to a very high level. The kernel weights per harvested plant is like 100:110 :123 for the three strains, i. e. the same proportion for Golden and Bonus as in the Svalöf yield trials for a sequence of years (FRÖIER, l. c.).

The change in manuring from 200 to 500 kilograms per hectare results in a pronounced increase of tillering. For the lumped material it amounts to about forty per cent. This in its turn causes an increase in the number of kernels per plant. Since the 1000-kernel weight rises with the higher manuring, the increase in kernel number combined with the larger seed weight makes the total production considerably higher (37 and 47 per cent increase of weight per plant and plot, respectively).

The essential point of the study goes to show that mixtures may be superior to the pure strains, that this phenomenon often has a genetical back-ground, insofar as different strains react in a varying manner with one another — here Maja barley is the best combiner and that the behaviour of the strain mixtures is directly influenced by the quality of the environmental factors (density, manuring).

In this series of experiments the factorial differences of the genotypes involved are no doubt very complex despite the close relationship of the strains. GUSTAFSSON et al. (1950) studied the behaviour of monofactorial mutants with regard to competitive ability. They showed that several mutations, homozygous or heterozygous, viable or lethal, interact with stand density, so as to be superior to the normal homozygote under certain conditions, inferior under others. In the analysis to follow a similar test material was studied with regard to the yielding capacity in mixture, e. g. segregating progenies of two monofactorial mutations.


In a recent paper GUSTAFSSON (1952) studied the viability reactions of two induced mutations, alone and in combination. One was the erectoides mutation, ert29 (HAGBERG, NYBOM and GUSTAFSSON, 1952), characterized by a change in the spike density. It gives a certain decrease in viability, if adequately analysed. The other mutation, viridis, is a chlorophyll lethal, the recessive dying in the seedling stage. The two mutations are loosely linked, with circa 30 per cent crossing over. They arose in one and the same X1 progeny.

In GUSTAFSSON'S paper it was shown that ert29, when heterozygous, decreased diploid viability. The chlorophyll lethal, vir, on the other hand, increased diploid viability when heterozygous. In the haploid state the chlorophyll lethal was eliminated to circa six per cent. ert29 showed no such gamete elimination. The result was that the two factors ert29 and vir behave entirely opposite to each other, both in the haploid and the diploid state.

In the year 1951 new experiments were begun and then fully analysed in 1952. The arrangements were as follows: In series 1-3 (v. below) seeds were sown of Maja barley in every tenth plot — 1, 11, 21, 31, etc. — each plot in series 1 and 2 comprising three rows and sixty seeds, in series 3 thirty seeds, in series 4 one hundred and twenty seeds, followed by randomized offspring of the normal homozygote (AABB), of the erectoides monohybrid (AaBB), of the lethal monohybrid (AABb), of the dihybrid (AaBb), of the erectoides homozygote (aaBB), and of the erectoides homozygote possessing the lethal (aaBb). In series 4 Maja barley was compared to the normal homozygote (AABB) and the erectoides mutant (aaBB).

Series 1: One kernel per hole of the marking-board, seed distance 5 X 15 cm, 300 kg saltpetre per hectare, 63 (64) offspring plots: Maja — 8, AABB — 6, AaBB — 7, AABb — 7, AaBb — 24, aaBB — 5, aaBb — 6. This arrangement corresponds to the so-called standard environment of GUSTAFSSON (1. c.), except for the fact that the manuring amounted to 300 instead of 200 kg.

Series 2: Like series 1 but 600 kg saltpetre, 51 (56) offspring plots: Maja — 7, AABB — 5, AaBB — 6, AABb — 3, AaBb — 21, aaBB — 5, aaBb — 4.

Series 3: Like 1 but seed distance 10 X 15 cm, 53 (55) offspring plots: Maja — 7, AABB — 2, AaBB — 6, AABb — 8, AaBb — 20, aaBB — 6, aaBb — 4.

Series 4: Two kernels per hole, 5 X 15 cm, 300 kg saltpetre. 12 (15) plots: Maja — 2, AABB — 5, aaBB — 5.

Comparison of Maja barley and the normal homozygote (AABB). — This analysis was made in order to settle whether the original irradiation had in any way influenced viability of the normal homozygote extracted from the segregating material of the X2 generation. Let me once and for all conclude that the normal segregate of the dihybrid (AaBb → AABB) is equal to the mother strain in production. It was tested not merely in the series and plots described above but also in a few yield trials of the Barley Department of the Swedish Seed Association. The following illustrates the case:

  1951, series 1-4 1950-1951, yield trials
No. of
No. of
Maja barley 229,3 97,7 24 6,56 104,1 3
Normal homozygote  235,4 100 18 6,30 100 3

Grand total: Maja barley = 100,9
Normal homozygote = 100

In the following discussions for this reason, when dealing with plot production, the values with regard to Maja barley have been added to those of the normal homozygote.


The erectoides mutation. — In the summer of 1951 all material was classified with regard to the mode of segregation, whether it contained recessives of the erectoides or viridis factor, or both of them. Offspring producing viridis mutants were then tested indoors. In the summer of 1952 a great many progenies were further analysed with regard to the occurrence of normal homozygotes and monoheterozygotes of ert29.

The 1951 plots segregating for the erectoides mutation and fully analysed in 1952 gave the following results with regard to genotype viability, in this case kernel weight/plant (normal homozygote = 100):

  AABB AaBB aaBB Average
plant yield
No. of
Series 1 100 90,6 ± 6,3 81,3 ± 6,7 4,42 gr 180
»   2 100 102,2 ± 14,0 80,7 ± 9,1 4,07 » 56
»   3 100 98,8 ± 9,1 96,9 ± 9,9 7,18 » 168
Average 100 97,2 86,3 404

The results of series 1 (the standard test) entirely agree with the data given by GUSTAFSSON (l. c., p. 272). Here the proportion runs as 100:91:81, there as 100:94:79. Consequently, in standard sowings with 200-300 kg saltpetre the erectoides mutant is inferior to the mother strain by 20 per cent. The monoheterozygote is intermediate in production. With a higher manuring (series 2) the heterozygote is apparently more productive, but no statistical significance of the fact is found. With a less degree of competition (series 3) there is an increase in relative production of the mutant. The heterozygote is still intermediary. The mutant is definitely better than in the dense sowings.

In the offspring of dihybrid plants no three but nine different genotypes are competing, three of which (AAbb, Aabb, aabb) die at an early stage. Six consequently survive. Without any detailed analysis it may be stated that the production of AABB, AaBB and aaBB is in the proportion of 100:98:78, i. e. figures similar to those given above.

Other facts than those reported add up to the following conclusion: Erectoides 29 is in the case of a high stand density definitely inferior to the competing genotypes. In the case of a more open growth with less competition (sparse stands) it may approach the normal homozygote. This holds true of pure cultures as well, either in small plots or in normal yield trials. 21 small plots gave an average relative production of 92,1 %, 5 plots from yield trials the relative production of 93,6 %.

The viridis mutation. — In the following survey all the 1951 data on the viridis heterozygote have been gathered. They concern monohybrid offspring exclusively — AABb → AABB, AABb and Aabb , where the AAbb genotype is lethal.

  AABB AABb Average
No. of
Series 1 100 102,7 ± 6,8 5,20 gr 314
» 2 100 110,8 ± 9,5 6,07 » 122
» 3 100 112,3 ± 6,6 8,11 » 182
Average 100 108,6 618

Evidently the lethal heterozygote is by 8 or 9 per cent superior to its mother strain. In the material studied by GUSTAFSSON (1952) it was to circa 20 per cent superior. Nevertheless the new material confirms the chief results of the previous study.

It is interesting to note that the viridis heterozygote is more productive in the lack of competition (series 3) than in the corresponding crowded conditions of series 1. This contrasts to the spontaneous lethals albina 7 and xantha 3 arisen out of Golden barley (GUSTAFSSON, NYBOM and VON WETTSTEIN, 1950), which are superior under crowded conditions but inferior when competition relaxes.

Erectoides and the viridis mutation. — In the 1949 experiments (GUSTAFSSON, l. c., p. 271) the data taken at their face value indicated that the genotype, homozygous for ert29 but heterozygous for vir (aaBb), was inferior to its sister genotype aaBB, homozygous for ert29 and normal chlorophyll formation. However, the difference was not significant (P = 0,1). In the new experiments this point was specially studied, with the following result.

  aaBB aaBb Average
No. of plants
Series 1 100 107,7 ± 6,5 4,73 gr 267
» 2 100 97,3 ± 7,2 5,34 150
» 3 100 109,5 ± 12,1 7,12 81
Average 100 104,8   498

The analysis gives no clear-cut answer to the problem. The standard errors are high and the differences fairly slight. The material certainly indicates a weak superiority for the viridis lethal also with the erectoides genotype. Like the case with the lethal in the normal genotype its greatest effect is when the crowding is relaxed (series 3). On the other hand, if we average all the data from 1940 and 1951, we obtain the following relative values: 88,5, 107,7, 97,3, 109,5, with a mean of l00,8, i. e. almost the exact production of the homozygous erectoides genotype (aaBB). The erectoides genotype shows a wide variation in production. This is even more so for the lethal heterozygote, in changed environments and different years.

The offspring of dihybrid plants contain three different genotypes, possessing heterozygous lethals, viz. AABb, AaBb and aaBb. The values of relative production are as 98:104:82 (AABB = 100). The erectoides homozygote aaBB is to 1,3 per cent inferior to its lethal heterozygote. Consequently here, too, the lethal heterozygote appears to exert a slightly beneficial effect.

The genotype viability in dihybrid offspring. — Owing to the low number of normal homozygotes appearing, viz. one normal segregate out of twelve surviving individuals in the case of free combination and less in the case of factor repulsion, there is a great labour involved when trying to obtain sufficient material. The interpretation of the data is not easy. They are given with all reserve.

  AABB AaBB AABb AaBb aaBB aaBb Yield/
No. of plants
Series 1 100 102,3 104,7 106,0 78,4 81,4 4,97 219
» 2 100 94,2 100,0 103,1 81,7 92,0 5,25 175
» 3 100 96,3 89,1 102,8 82,3 72,9 8,10 165
Average 100 97,6 97,9 104,0 80,8 82,1  
No. of plants 55 86 72 186 72 88 559

Except for the production of the erectoides genotype aaBB, the figures scarcely agree with those given by GUSTAFSSON (1952, p. 272).

In the old analysis, however, the kernel yields were not examined, merely the number of spikes and kernels. In addition, there were only six normal plants present, giving a high statistical uncertainty. Especially interesting in the new analysis is the high production of the dihybrid itself, which is 4 per cent better, on the average, than the normals. In all three series it lies higher than any other genotype. Still the proportions are statistically uncertain.

Summarizing the data given, we conclude with regard to the production capacity of individual genotypes, either in pure cultures, in monohybrid or in dihybrid segregation: (1) the erectoides homozygote is by 20 per cent inferior in segregating progenies of normal density, but no more than 5-10 per cent in pure cultures, (2) the erectoides heterozygote AaBB lies closer to the normal homozygote, but is still by at least 3 or 4 per cent inferior, (3) the erectoides type heterozygous for the lethal, aaBb, is similar to the erectoides homozygote aaBB, possibly somewhat superior, (4) the normal type heterozygous for the lethal, AABb, is distinctly superior to its homozygote AABB by 10 per cent or more, and (5) the double heterozygote AaBb appears to be slightly superior to all other genotypes in the case of a dihybrid segregation.


The erectoides monohybrid. — In Table 2 the data do not refer to the production of individual genotypes but of genotype mixtures, the mother plants of the plots having the constitution AABB, AaBB, aaBB.

TABLE 2. Production of the monohybrid offspring in relation to its corresponding homozygotes.

per plot (gr)
(rel. fig.)
(rel. fig.)
Per cent of harvested plants
Series 1 233,0 107,2 98,7 86,3 94,3 88,3
» 2 262,8 99,4 83,6 90,4 93,1 83,0
» 3 199,3 105,5 94,1 93,0 98,9 92,8
Average 100 104,0 92,1 89,9 95,4 88,0

Most interesting figures relate to the offspring of the erectoides monohybrid AaBB, with its three segregated genotypes of AABB, AaBB and aaBB. The yield data of this offspring lie in all three series above the calculated mean of the individual genotype productions, but also, on an average, above the normal homozygote with no less than four per cent. If we apply the yield figures found for the three genotypes within a segregating progeny, we should expect the AaBB offspring of series 1 to produce 90,8 per cent of what AABB is producing, of series 2 to produce 97,0 per cent and of series 3 to produce 98,5 per cent, i. e. an averaged yield of 95,4 per cent of AABB. Since we know the production of the individual offspring plants of the normal homozygote in pure culture, we can compare these values with those actually found within the segregating progenies.

Individual plant production in pure cultures and genotype mixtures.
(Relative figures.)

  Pure cultures Genotype mixtures
All plots
Series 1 100 (4,50 gr) 96,4 100,0 91,3
» 2 107,6 98,0 107,3 91,3
» 3 158,9 150,0 163,1 150,7

The higher production of the monohybrid offspring depends on two factors: (1) a higher number of harvested plants per plot, 95 % as against 90 % of the seeds sown, (2) on a higher production capacity of the individual plants in genotype mixtures. If we scrutinize the figures above, we find that the genotypes AABB + AaBB lie on the same level as or higher than the corresponding normal plants from the pure cultures. Taking into consideration that AaBB is inferior to AABB by some three per cent, we may conclude that also AABB itself produces more in genotype mixtures than in pure cultures.

It is, however, not quite appropriate to lump all yield values and to compare the lumped figures for genotype after genotype, since the percentage of harvested plants is so different. If we for that sake divide the materials into groups of different density, according to the number of harvested plants per plot, we obtain the values of Table 3.

TABLE 3. Production of the erectoides monohybrid in relation to the two homozygotes and the occurring plant density.
(Calculated as harvested plants/sown seeds.)

  Series I Series 2 Series 3
  80-90% 90-100% 80-90% 90-100% 80-90 % 90-100%
AABB-offspring 239,1 gr 221,9 gr 235,2 gr 208,3 gr 220,5 gr 196,5 gr
No. of plots 9 5 2 10 1 8
AaBB-offspring 247,8 250,8 255,0 264,1 - 213,0
No. of plots 2 5 2 4 - 6
aaBB-offspring [190,8] 256,0 203,3 285,5 148,0 195,5
No. of plots 2 3 4 1 1 5


    80-90 % 90-100 %
  AABB 100 100
Series 1-3. Relative values: AaBB 106,0 106,6
  aaBB 77,8 107,1

The results are immediately evident: (1) The erectoides progeny (aaBB) is much better when a full number of plants is present. In fact, it is by some 7 per cent better than the normal type in the case of full density but is inferior in sparse stands by some 20 per cent. The difference appears to be significant (0,05 > P > 0,01). This indicates that the erectoides genotype is a good yielder in the case of a dense stand. When for some reason, however, there occur gaps in the stand or the germination is depressed, it shows a poor compensation ability by means of an increased tillering. This is in agreement with other experiments and explains why erectoides cannot compete in mixed stands but is better in pure cultures, relatively seen.

(2) The monohybrid mixture (the AaBB offspring) is better yielding than the normal type both in dense and sparse stands. It is, on an average, superior by 6 or 7 per cent in both density groups. Also here the significance appears to be reliable (0,05 > P > 0,01). Consequently, the monohybrid mixture owes its superiority not only to the higher plant percentage, but to a better yielding per plant, since it is superior both when the plant number is maximal and when it is decreased.

The lethal monohybrid. — The data are summarized in the following survey:

  Yield of AABB-offspring
per plot
Yield of AABb-Offspring
(relative figures)
Per cent harvested
Per cent harvested
plants, AABb
Series 1 233,0 gr 102,0 86,1 % 74,4 %
» 2 202,8 » 94,3 90,4 » 68,3 »
» 3 199,3 » 92,5 92,9 » 74,8 »
Average 100 96,3 89,3 % 72,4 %

In spite of the much decreased number of plants per plot the lethal heterozygote is entirely on the level with the normal type in series 1, the standard arrangement. However, it is distinctly inferior in the case of a sparse stand (series 3). In series 2, where the number of harvested plants is extremely low, it also falls down. Nevertheless its yield capacity is astonishingly good. Apparently it has a high compensating ability, fills out the gaps arisen from the death of the chlorophyll recessives, when the gaps are not too pronounced as in the case of 2 and 3.

This ability of compensation is more pronounced in the genotype AABb, the lethal heterozygote, than in AABB.

Average plant production. (Relative figures.)

Homozygote offspring Heterozygote offspring
        AABB: 115,7
Series 1 100 (4,50 gr) 117,7  /
        AABb: 118,8
AABB: 127,2
» 2 107,6 134,3  /
        AABb: 140,1
» 3 158,9 180,2  /
        AABb: 188,3

It is plain that a high manuring (series 2) causes a rise in average plant production, although smaller than with a wide seed sowing. In all three kinds of environment the heterozygote compensates better than the normal type to an increase in available plant space.

What has here been concluded, is confirmed by a detail analysis of the different density groups. There is in this instance, too, especially in series 1 but also noticeable in series 2 and 3, an evident occurrence of a genotypical cooperation.

Erectoides and the lethal heterozygote. — In this case there is but a weak compensation for the decrease in plant number, even in the standard arrangement of series 1 and with the high manuring of series 2. The decrease in yield is very pronounced, when the wide sowing of series 3 is considered.

  Yield of
per plot
Yield of
(rel. fig.)
% of harvested
plants, aaBB
% of harvested
plants, aaBb
Series 1 229,9 gr 92,1. 88,3 75,3
» 2 219,8 » 91,2 83,0 62,5
» 3 187,6 » 76,8 92,8 67,5
Average 100 86,7 87,3 68,4

The two erectoides genotypes do not possess the same compensating ability as the corresponding normal genotypes. In addition, the aaBb plants are less superior to the erectoides homozygote aaBB. There is however a small degree of genotype cooperation apparent when the different percentage groups of harvested plants per plot are compared for the two types of offspring.

Finally, it is impossible at present to discuss the results of the dihybrid offspring plots. So much can be said, however, that the individual genotypes have, on an average, a lower compensation ability than the normal lethal heterozygote but a higher than the erectoides lethal heterozygote, for the corresponding percentage groups of harvested plants.


In a series of papers the present author and his collaborators have tried to exemplify the complex modes of interaction between genotypes, as well as genotypes and environment, also when monofactorial differences are studied. A specially significant phenomenon deals with the so-called Montgomery effect. This implies that a special genotype, although inferior in maximal production capacity, nevertheless might be superior in competition. This feature was studied by GUSTAFSSON et al., 1950, in some monofactorial mutations in barley, viable as well as lethal. Another significant result of recent research has indicated that a mutant may react in a most unexpected way to a drastic change in environment. Here the ancestor line becomes scarcely viable and is often surpassed in viability by a normally inferior or even semilethal mutant. In addition, mutations lethal or semilethal, when homozygous, may stimulate in heterozygous condition and make the heterozygote superior in production to its normal homozygote.

In this paper some further contributions have been rendered. Take, for instance, the lethal viridis mutation, induced in Maja barley by X-rays. It causes one fourth of the monohybrid offspring to die at an early seedling stage. In spite of the necessarily thinned stand the heterozygous lethal conditions such a high compensation ability of the monoheterozygote and, by genotype cooperation, of the normal homozygote, too, as to make the population reach or surpass the production level of what the normal homozygote does in pure cultures. Here it is worth while to refer to the publication of ENGLEDOW and RAMIAH (1930), who found that 20 to 30 per cent of the possible maximal yield of a cereal field is regularly lost by gaps and thinnings in the population. Some ten years later TEDIN and ANDERSSON (1943) noted that neighbours of the existing gaps in a barley field were able to compensate for approximately one third of the total loss in production, caused by plant accidents. The remaining two thirds of the loss could not be regained by an increase in tillering. The proportion of 1/3 regain, 2/3 loss, of course, depends on the choice of genotype, the type of climate and the actual environmental conditions. The viridis heterozygote of Maja barley, discussed here, and by compensation and genotype cooperation, its normal homozygote, too, together recapture more than 100 per cent of the loss. This is when the sowing arrangements are of the standard type, i. e. at a normal density. The seed distance being exceptionally large or the plant elimination heavily accentuated, the offspring of the lethal heterozygote are incapable of compensating the entire loss.

Even more pronounced is the cooperation when the offspring of the erectoides monohybrid are considered. Here the mixture of 1/4 AABB, 2/4 AaBB and 1/4 aaBB produces four per cent more than the normal homozygote itself, despite the fact that AaBB is by three per cent inferior and aaBB by fourteen per cent inferior to the AABB genotype of the same offspring. Under the standard arrangement the surplus lies higher, viz. between seven and eight per cent. In his paper of 1912, MONTGOMERY stated that two varieties in competition gave »a greater number of plants at harvest». This is valid here, too. (The strain mixtures, however, gave a lower plant percentage than the pure strains. Cf. p. 3.) In addition, there is a distinct gain in plant productivity, compared to the conditions in pure cultures.

Further points of interest concern the behaviour of the lethal and the erectoides heterozygotes. It has again been found that the genotype heterozygous for the lethal, AABb, increases production above that of its homozygote, but that the erectoides heterozygote, AaBB, on the other hand, decreases production. This illustrates once more »that there are no definite means of calculating from the viability of the homozygous mutation, how it will react in the heterozygous state» (GUSTAFSSON, 1952, p. 272). Interesting, too, is the fact that the induced lethal is equally effective in sparse as well as in dense plant stands. This is in contrast to the spontaneous lethals previously studied in this respect (GUSTAFSSON, NYBOM and VON WETTSTEIN, 1950), which are superior under competition but inferior when competition relaxes. The reason, why the viridis lethal increases yield, is so far not settled. There is, however, a weak indication that the superiority is connected with an increase in chlorophyll pigments (NYBOM, unpubl.), similar to what was found in the chlorophyll heterozygote of Antirrhinum studied by STUBBE and PIRSCHLE (1941).

The behaviour of the ert29 homozygote further illustrates the complexity of the viability reaction. Under heavy competition (dense stands) the mutant is by 20 per cent inferior to its ancestor. When competition is relaxed, but still in segregating progenies, it gains in relative production up to a 4 per cent point of inferiority. Similar is the case in pure stands, either in the small plots of series 1-4 or in regular yield trials, where its inferiority stops at 5-10 per cent. This contrasts to the bright-green mutant 4 out of Bonus barley (GUSTAFSSON and NYBOM, 1950) that is superior to its ancestor in dense stands but distinctly inferior in yield trials or in such mixed stands where competition relaxes.

These data further argue the need of extreme caution when evolutionary trends are concluded out of the behaviour of some single mutation studied. In barley every mutation, so far studied, whether morphological or physiological, drastic or minor, viable or lethal, appears to possess a reaction mode of its own, changing with the type of environment and the genotypical milieu.

There is consequently a wide complexity of genotype interreactions. The stamping-out method of natural selection, so expressively depicted by DARWIN, has a counter-part in the mutual cooperation of genotypes, within or between species. If the results of this paper are valid not only for the agricultural field, but can be transferred also to the happenings in free nature, one important conclusion turns up. Although a certain mutant, variety or species, owing to an inferiority in production, should according to advance calculations be doomed to extinction, it will be tolerated or even favoured as a member of a genotype community, since it enhances fitness of its superiors and of the population as a whole. This is specially illustrated by the monohybrid offspring of ert29, AaBB, as well as the Golden-Maja and Maja-Bonus mixtures. A population, lacking in inferior genotypes and in heterogeneity, should according to this view often be less fit in total production than a corresponding population hiding harmful genes or producing inferior genotypes.

To finish this sketchy analysis: There is successively spreading in agricultural praxis a belief that a superior cereal stock should not be built upon a single genotype, but rather on a couple of or on numerous genotypes, selected so as to cooperate, also in self-fertilizing organisms. Such a view obtains some support from this study. The plant breeder making hybrid corn or sugar beet selects the inbred lines or strains which are specially apt in intercrossings. In a similar way the wheat, oats and barley breeders have in future to select the lines or strains that interreact in an improving way upon the production of the population and by that lead forward to a gain in total yield.


  1. CARDON, P. V. 1921. Grain mixtures and root crops under irrigation. — Agr. Exp. Sta. Montana Bull. 143: 1-14.
  2. CEDIK-TOMASEVIC, Z. F. 1951. Die Resultate der Versuche mit Artmischungen von Getreide. — Agrobiologia 1: 100-121. (Cited after »Züchter», 21, 1951: 348-349.)
  3. DARWIN, CH. 1859. On the origin of species. — Reprint of the first edition; London, 1950. )With a foreword by C. D. DARLINGTON.)
  4. ENGELKE, H. 1935. Cited after FRANKEL, 1939.
  5. ENGLEDOW, F. L. and RAMIAH, K. 1930. Investigations on yield in cereals. VII. A study of development and yield of wheat based upon varietal comparison. — Journ. Agr. Sci. 20:265-344.
  6. FRANKEL, O. H. 1939. Analytical yield investigations on New Zealand wheat. IV. Blending varieties of wheat. — Journ. Agr. Sci. 29: 249-261.
  7. FRUWIRTH, C. 1923. In »Deutsche Landwirtschaftliche Presse». Cited after LINDBLAD, E. 1923. Blandsäd eller renhestånd. — Lantmannen 6:701-702.
  8. FRÖlER, K. 1951. Korn. — In »Svensk Växtförädling», Stockholm: 169-208.
  9. GUSTAFSSON, A. 1952. Mutations, environment and evolution. – Cold Spring Harbor Symp. Quant. Biol. 16, 1951:263-281.
  10. GUSTAFSSON, A. and NYBOM, N. 1950. The viability reaction of some induced and spontaneous mutations in barley. — Hereditas XXXVI: 113-133.
  11. GUSTAFSSON, A., NYBOM, N. and VON WETTSTEIN, U. 1950. Chlorophyll factors and heterosis in barley. — Hereditas XXXVI: 383-302.
  12. HAGBERG, A., NYBOM, N. and GUSTAFSSON, A. 1952. Allelism of erectoides mutations in barley. — Hereditas XXXVIII: 510-512.
  13. JONES, E. T. 1948. Cereal production. — Welsh Plant Breed. Sta., Aberystwyth: I-46.
  14. LYSSENKO, T. D. 1918. Die Situation in der biologischen Wissenschaft. — Deutsche Übersetzung. 2. AufI. Berlin, 1951: 1-456.
  15. MONTGOMERY, E. G. 1912. Competition in cereals. — Bull. Agr. Exp. Sta. Nebraska 127: 3-22.
  16. NUDING, J. 1936. Cited after FRANKEL, 1939.
  17. RÜMKER, K. VON. 1892. Cited after FRANKEL, 1939.
  18. STUBBE, H. und PIRSCHLE, K. 1941. Über einen monogen bedingten Fall von Heterosis bei Antirrhinum majus. — Ber. deutsch. bot. Ges. 58:546-5,18.
  19. TEDIN, O. and ANDERSSON, E. 1943. Urvalsstudier hos korn. — Sv. Uts. För. Tidskr. (Svalöf) 53: 98-110.


Introduction 1
The cooperation of strains 2
Yield interaction in mono- and dihybrids 6
    Comparison of Maja barley and the normal homozygote (AABB) 7
  Genotype viability 7
    The erectoides mutation 7
    The viridis mutation 8
    Erectoides and the viridis mutation 9
    The genotype viability in dihybrid offspring 9
  The reaction of genotype mixtures 10
    The erectoides monohybrid 10
    The lethal monohybrid 12
    Erectoides and the lethal heterozygote 13
Conclusions 14
Literature cited 17