Nature 196: 1302-1304. (1962)
Department of Agricultural Botany, University of Wales, Aberystwyth,
and Earhart Plant Research Laboratories. Biology Division,
California Institute of Technology, Pasadena

EXPERIMENTS with many thousands of plants carried out over a period of nine years have shown that heritable changes are induced in flax plants of the variety Stormont Cirrus when they are supplied with different combinations of nitrogen, phosphorus and potassium, and when they are grown for at least the first five weeks in a heated greenhouse1. The original variety is a plastic genotroph which is changed into a large stable genotroph, or a small stable genetroph, depending on the nutrients supplied. The large stable genotroph is three or more times the weight of the small stable genotroph and both have remained unchanged in inheritance for seven generations irrespective of the fertilizers that have since been applied. Both behave as two distinct genetic types, giving equilinear inheritance in their reciprocal crosses and no transmission through their reciprocal grafts. The plastic genotroph is intermediate in weight, but taller than the other two, and all plants are highly uniform within each genotroph.

Studies have been made on the variation in the F1 and F2 of the reciprocal crosses between the large (L) and small (S) stable genotrophs. Table 1 gives the results of separate crossing experiments with different groups of parent plants carried out in different years at Aberystwyth. The F2 was examined first and the coefficients of variation, which compensate for differences in the means, show that the variation in both reciprocal crosses is higher than within the individual parent crosses, suggesting genic segregation. On the other hand, there is a marked difference between the two reciprocal crosses, the S x L variation being higher than any of the other three types of cross. In addition, the same pattern of variation is obtained in the F1 of these crosses, and in the F1 of other crosses, in plant weight and plant height. The correlation between the F1 and F2 coefficients of variation of the four types of crosses is extremely high (r = 0.87 for F1 weight/F2 weight; r = 0.97 for F1 height/F2 weight). Genic segregation therefore is probably not the sole expIanation for the increase in variation in the F2 of the reciprocal crosses.

Coefficient of variation (C of V, standard deviation divided by mean) is averaged over each set of experiments

    L x L L x S S x L S x S
  Mean Variance Mean Variance Mean Variance Mean Variance
F2 plant weight Exp. 1 33 167 27 172 28 353 7 11
Exp. 2 35 135 20 81 22 103 10 7
Exp. 4 36 84 22 53 28 209 11 18
  Total 104 386 69 306 78 705 28 36
  C of V 0.33 0.42 0.55 0.37
F1 plant weight Exp. 4 64 293 50 203 53 240 17 11
Exp. 5 65 233 49 153 47 228 21 31
Exp. 6 103 1,130 81 1,166 75 2,503 26 382
  Total 232 1,656 180 1,522 175 2,961 64 424
  C of V 0.28 0.32 0.43 0.20
F1 plant height Exp. 1 85 12 81 19 77 30 71 8
Exp. 2 61 15 61 19 60 41 52 18
  Total 146 30 142 38 137 71 123 26
  C of V 0.05 0.06 0.09 0.06

To check whether the F2 variation was at least in part due to the variation in the F1, five F1 plants were selected from each of the four types of crosses, and F2 families of five plants were grown from each in winter in a heated greenhouse. Table 2 shows that the positive correlation between F1 plant weights and F2 family means is virtually complete in the S x L cross, and high in the L x S cross, while the small amount of variation among the F2 family means of each of the parent crosses is negatively correlated with the F1 plant weights, a maternal relationship commonly occurring under winter conditions. There is, therefore, an immediate breakdown, or instability, in the F1 of crosses between the two stable genotrophs producing to all appearances random, heritable variation, which presumably arises from the opposing nature of the factors or processes determining the two genotrophs, or the destruction of their delicate balance by chance metabolic variation. Since factors in addition to nutrition appear to be involved, the various forms may be referred to as 'epitrophs'. There is evidence from exploratory experiments at the Earhart Laboratories that the F1 instability may be arrested if the F1 plants are grown under certain conditions but it is not clear at present which are the relevant environmental factors. The large genotroph used for those crosses was induced at Aberystwyth by applying nitrogen, phosphorus and potassium to the plastic genotroph, and the small by applying nitrogen and potassium, but taking into account soil fertility and soil interactions, it was suspected that the balance of nitrogen and phosphorus was particularly important, nitrogen inducing the large genotroph and phosphorus the small. At the Earhart Laboratories nitrogen and phosphorus were added separately to different groups of plants of the plastic genotroph grown in a gravel and vermiculite mixture with dilute Hoagland's solution in different day/night temperature (°C) combinations. The progeny of the nitrogen and phosphorus treated plants were approximately equal in weight to the large and small genotrophs respectively when the treated parent plants were grown in 27/15 and 23/23 temperatures, progressively smaller induced changes occurring in 27/11, 27/7 and 31/7, and none in 19/11. This confirms the importance of the balance of nitrogen and phosphorus, in the absence of soil interactions, for the induction of changes transmitted to the first generation and also shows that high day and night temperatures are necessary. On the other hand, these changes were not inherited by the second generation grown at Aberystwyth so that there remain other unknown environmental factors. Those impermanent changes are nevertheless of additional interest and are of some significance in view of the temperature requirements for partial reversion mentioned here.


  L x L L x S S x L S x S
Variance Variance of F plants 4,196 7,626 3,587 865
Variance of F2 family means 1 203 83 11
fF1/F2 -0.19 0.62 0.94 -0.67

At Aberystwyth the large and small genotrophs were grown under conditions giving a greater day/night temperature difference, maximum natural day length and maximum light intensity during the first five weeks of their growth, and the two following generations were grown under approximately similar conditions. By the third generation the difference in weight between the two genotrophs had practically vanished although other characteristic differences such a branching habit, leaf size and colour remained distinct. At the Earhart Laboratories the original large ad small stable genotrophs were grown in day/night temperatures of 27/15 and 27/7 in 50 per cent Hoagland's solution, and the two following generations in all combinations of these temperatures as shown in Table 3, where the differences between the large and small genotrophs in the second generation are entered. The differences between the two sets of differences, one set for the second generation plants grown in 27/15, and the other for those grown in 27/7 in addition to assessing the responses in this generation remove a large amount of extraneous maternal effects of the previous generation's treatments, and show that the difference between the large and small genotrophs grown consistently in 27/7 is greatly reduced compared with those grown consistently in 27/15. Those grown previously for one generation in each of the two temperature combinations show intermediate differences. This partial reversion is probably not permanent and appears from these and other observations to occur when parents and their progeny are grown consistently in the same environment of strongly contrasting day/night temperatures (27/7). In view of these results, the induction and transmission to the first generation of only small changes at 31/7 and 27/7 temperature combinations mentioned here are not unexpected.


Parental generation temperature 27/15 27/15 27/7 27/7
First generation temperature 27/15 27/7 27/15 27/7
Second generation grown in 27/7 (a) 18 18 10 10
Second generation grown in 27/15 (b) 10 16 9 15
Difference (a-b) 8 7 1 -5

Altogether there are here three classes of heritable change. Class I: changes which are maintained largely in those environments inducing them, that is, partial temporary reversion of the large and small genotrophs. Class II: the induction of the large and small stable genotrophs from the plastic genotroph. Class III: the heritable variation arising inmiodiately in crosses between the large and small genotrophs. Classes II and III are almost certainly nuclear with the cytoplasm having a subsidiary role. It is not clear whether intermediate genotrophic forms induced by certain fertilizer and temperature treatments, or some of the epitrophs, are stable or transitional to the three genotrophs mentioned earlier, but it is notable that all the variation originates from the influence of the environment on a hitherto uniform variety of flax.

These results are more profitably discussed when other crossing experiments are completed. Numerous comparisons can be made with other forms of heritable change, cellular differentiation, paramutation and similar phenomena, but it is particularly interesting to note that the variation in Class III is strikingly similar to the variation reported by Sonnoborn2, Kimball3, Nanney4, and other workers, among caryonides of the unicellular organisms Paramecium aurelia and Tetrahymena piriformis, where exconjugants give rise to daughter cells of different mating types during successive mitotic divisions. Variation of this type may therefore be common and perhaps responsible for party of the variation normally ascribed to classical segregation and recombination.

Varieties of other crop plant, unlike flax varieties, have been bred for characters such as plant weight and seed yield1, and they may be equivalent to the stable, rather than to the plastic flax genotrophs, so that the probability of Class I changes occurring would be higher than those of Class II. One example is known; Highkin5 has reported that peas of two varieties deteriorate in size and vigour when grown for several generations in constant day/night temperatures, but are restored after two or three generations in a fluctuating day/night temperature. Speculatively, assuming similar metabolic changes to those in flax, the pea varieties could have arisen initially from plastic pea genotrophs grown at a reasonably high temperature with excess phosphorus — not improbable conditions. On the other hand, it may be simply a question of breeding.

It is probable that individuals in natural populations vary in their possession of heritable factors or processes which can be altered or stabilized by the environment. Mutation anywhere in the heredity material, or crossing between unlike individuals, may stabilize some factors to give stable genotrophs, and break down others to give plastic genotrophs, the environment thereby directly altering the characteristics of a population, as distinct from its indirect action in natural selection. Although it is possible that heterozygous individuals in natural populations may be buffered to a greater extent against all three classes of heritable change, such changes may be widespread, and potent factors in biological evolution.

  1. Durrant, A., Heredity, 17, 27 (1962).
  2. Sonneborn, T. M., Proc. Amer. Phil. Soc., 79, 411 (1938).
  3. Kimball, R. F., J Exp. Zool., 81, 165 (1939).
  4. Nanney, D. L., Genetics, 40(3): 388 (1955).
  5. Highkin, H. R. Amer. J. Bot., 45, 626 (1958).