Crop Production and Environment (1960) 231-244
R. O. Whyte

CHAPTER XV

BREEDING AND ANALYSES OF ENVIRONMENTAL REQUIREMENTS

Developmental Physiology in Relation to Formal Genetics

This is the section of the whole problem of developmental physiology that has caused most disagreement and controversy between the various schools of thought and interpretation. One chapter is completely inadequate to deal with the many fundamental issues that arise in the discussions and, in addition, the subject awaits detailed and unbiased consideration by a specialist in all respects of plant genetics and the theories of evolution. The forthcoming review by the Imperial Bureau of Plant Breeding and Genetics is awaited with great interest.

The basic criticism levelled by the supporters of phasic development at what they call the Mendelism-Morganism school of formal genetics would appear to be that the geneticists belonging to the latter have, in the very large number of experiments that have been carried out, concerned themselves with a study of many purely morphological and growth-type characters, with little or no regard to the physiological characters of a plant. The claim is put forward that it is the physiological inheritance of a plant that is important, both genetically and economically, and that many if not all of the morphological characters that have been the subject of economic plant breeding for many years are merely the expression of the type of inheritance of the physiological (developmental) characters under the environment of the experiment. The natural sequence in this argument is to say that many if not all the pure lines that have been developed are not true pure lines at all, as they are thoroughly heterozygous for the physiological characters and segregate out for these as soon as the environmental or other conditions permit.

In place of the conventional methods of plant breeding based upon factorial analysis, cytogenetics, the gene theory and similar methods that have been the foundation of genetical research for so long, it is argued that plant breeding should now be based upon a thorough knowledge of the developmental period and phases (or environmental requirement), obtained by phasic analysis, and that intelligent crossing of parents with the required developmental phases will provide a progeny with the requisite seasonal behaviour, growth type, morphological characters, yield, etc.

Another outcome of this research on the developmental period of plants is the emphasis on the method of vegetative hybridization in the Soviet Union. It is claimed that these graft-hybrids are true hybrids in that they show segregation, and this is used as an argument disproving the theory 'that the heredity of an organism is due to a special discrete object located in the chromosomes' (Lysenko). The connection between this result and the research on grafting in relation to hormonal exchange between plants noted in Chapters X and XI is obvious, and indicates how immediately applicable in crop production will be the results of this most specialized branch of developmental research.

The fact that it is possible to change a plant, even by the loosest form of contact-grafting, from a short-day to a long-day type or vice versa, from a 'late' to an 'early' type, and so on, seems to cast some doubt on the ability to regard these expressions of reaction to the environment as themselves based directly upon genetical factors and genes. It appears to be a question whether breeders who regard earliness and lateness, spring and winter habit, length of the thermo-phase or photo-phase as inheritable characters are not all talking about the same thing, namely the individual requirement of a plant or a variety as regards the individual factors of the environment, without a sufficiency of which its progress to maturity (development) would be impossible.

While fully realizing the scope and complexity of this whole question, some examples will now be discussed to indicate how the 'formal' geneticists and the 'developmental' geneticists might both regard and interpret the same phenomena. Reference will be made first to a study made at the Research Institute of Plant Physiology, London.

Inheritance of Spring and Winter (or Annual and Biennial) Habit

In the report by Purvis (1939) on this research, it is stated that vernalization would appear to convert a winter rye into a spring rye within one generation as far as physiological behaviour is concerned, that is it induces in winter rye a physiological condition leading to early flowering that is already inherent in spring rye.

Purvis states that a review of the literature would support the dominance of the spring habit, but varying results have been reported. Tschermak (1923) found winter habit dominant in wheat, but not in barley or rye; Spillman (see Washington, 1909) obtained an intermediate type in the F1 generation of a wheat cross and a 1:2:1 segregation in the F2, and similar results were obtained in various wheat and barley hybrids by Ollson, Schafer, McCall and Hull (see Washington, 1920).

Dominance of spring habit, with a 3:1 segregation or an approximation thereto in the F2 generation is reported by the following (references from Purvis, 1939): Nihlson Ehle (1917), Takahasi (1924), Fruwirth (1923), Tschermak (1906). Purvis also refers to examples of dominance of spring habit associated with less simple ratios. Cooper (1923) crossed two spring wheats with three winter varieties, and found the spring habit dominant in the F1 generation in all six cases. In the F2 generation, however, the segregation ratio was 13:3 with Marquis as the spring wheat parent, and 3:1 with Manitoba 169 as the spring parent. Cooper postulates that a dominant winter gene is present, but that its operation is suppressed by an inhibiting gene present in both the spring forms. It is suggested that the winter gene is present in one spring variety but absent in the other. Gaines (see Washington, 1917) offers a similar explanation of his results from crossing two spring barleys, when he obtained three winter-type plants in every sixteen plants of the F2 generation. Purvis continues: The assumption that the winter gene is dominant but can be inhibited provides an explanation of the different types of inheritance described above. If W represents the winter gene and I the inhibitor, then the gametes of winter forms would have the genetic constitution Wi while those of spring forms might be Wi, WI, or wI, and these, crossed with the winter forms, would give (a) 3:1 winter dominance, (b) 3:1 spring dominance, and (c) 13:3 spring dominance respectively. This does not, however, justify the statement that the winter habit is dominant. If the inhibiting gene has the effect of curtailing the long vegetative period associated with the "winter" gene, then it is not clear on what grounds one may discriminate between such a gene and one carrying the spring habit.'

Again Purvis notes that this relatively simple inheritance in which one gene delays flowering beyond the minimum period, while a second gene may render the first inoperative, 'thus restoring the status quo', may possibly apply in a number of cases. Much more complex ratios have, however, also been recorded, which Purvis considers indicate the possible operation of multiple factors. Reference is made to the work of Vavilov and Kuznecova (1921) and of Aamodt (1923).

The following summary of Aamodt's results of a cross between Marquis and Kanred wheats is taken from Purvis' paper, in order that it may fit in properly with her own discussion and her experimental results. The F1 generation seed from this cross was sown in autumn, so that its earliness could not be assessed. The plants were harvested individually and 5,250 plants were sown in the spring of the following year. Of these 4,808 eared before autumn and 442 remained vegetative. Purvis notes that the apparently anomalous ratio of 10.8:1 resembles that of 9.6:1 recorded by Vavilov and Kuznecova (1921). Heading of the 4,808 annual plants was spread over a period of 8 weeks, the earliest flowering at the same time as the spring parent. Selected families were harvested from the groups heading in each week, and F3 plants were grown separately from their seed. This revealed that the earliest group in the F2 generation was homozygous, all giving spring plants, and the proportion of heterozygotes increased as the heading dates of the F2 parent became later. In addition, some plants were homozygous for much later flowering dates than the spring parent, and the later groups included plants that were homozygous for the winter habit. Aamodt regarded his results as a demonstration of partial dominance of the spring habit, and considered that the operation of multiple factors is responsible for the varied inheritance of the spring and winter habit in wheat.

After quoting the conclusion of Kajanus (1927) that the genetic basis of the spring and winter types is by no means clear, Purvis states her own conclusions that there is clearly no generalization which can be applied to all cereals, although most of the evidence points to complete or partial dominance of the spring habit. It is suggested that possibly in the diploid genus, rye, a single pair of factors is concerned, while in the tetraploid genera, wheat and barley, two or many pairs may operate. 'Since, so far as is known, no qualitative differences appear in the response of the temperate cereals to external physiological factors controlling flower production, it would appear probable that the genetic factors controlling flower formation are throughout identical, and therefore a general agreement in dominance might reasonably be expected. Until more is known of the actual physiological processes concerned in flowering and their relationship to genes, any attempt at explanation must be largely speculative.'

The above paragraphs are from the introductory discussion on the inheritance of spring and winter habit in Purvis' paper (1939), after which is given a description of her own experiment, in which reciprocal crosses were made between Petkus winter rye and Petkus spring rye. In the F1 generation, the spring habit was completely dominant, while in the F2 generation, 'spring' and 'winter' plants (quotes inserted by Purvis) occur in approximately the ratio of 3:1. There is a dispersion of flowering dates similar to that observed by Aamodt (1923), but very much less marked. This is tentatively explained on the basis of independent segregation of factors for early and late ripening as distinct from spring and winter habit. In this, Purvis is agreeing with Tschermak (1923), who lists early and late ripening as a pair of factors distinct from the annual or biennial habit.

Another relevant study discussed by Purvis should be quoted here. Melchers (1936-7) confirmed the work of Correns (1904) in finding in the species Hyoscyamus niger the existence of two races which are annual and biennial in their flowering behaviour. The biennial race itself includes two forms, one of which may be vernalized. After being sown in spring, the type that can be vernalized flowers during the summer, while the other forms a rosette of leaves and a tuber, and flowers only after a winter rest. Purvis quotes Correns as finding that hybrid plants formed rosettes and yet began to shoot late in autumn.

It is obvious from this abridged version of the discussion and results from Purvis' paper that the attempt to explain the particular phenomena at issue on a genetical basis, using simple multiple or inhibitory factors, has not been markedly successful. What would appear to be desirable now would be a series of experiments to prove whether or not the views of the phasic development school could contribute anything to a final and satisfactory explanation.

Criticism Based upon Developmental Physiology

This alternative Russian view, while not necessarily denying the existence of genetics as a subject, states that research on genetics and plant breeding has up till recently been based upon the theories of Mendel and Morgan, theories the soundness of which are regarded with great suspicion, to state the situation mildly. It is stated that what is inheritable is not a time of flowering or an annual or biennial habit, but a developmental period of a given type, individual and distinct types for every plant or at least for every variety. By 'developmental period' is understood that period characterized by a succession of requirements of environmental factors under the influence of which a plant can proceed to sexual reproduction and the formation of mature fruit. There can be little or no objection to this statement that every plant has a developmental period, and that to complete it every plant has its own environmental requirements, varying in degree, intensity, period of reaction and so forth. This would not necessarily mean the acceptance of a strict succession of discrete phases, one dependent upon low or high temperature, another on daylength or other variation in the supply of light.

The phasic development theory, however, upon which the criticism of formal genetics is based, does assume a definite and unchangeable succession of phases in the progress of a plant towards sexual maturity. It is also stated that these phases are inherited independently. Thus it is argued that, although plants may have been selected or bred for a definite total length of the vegetative period, such as the number of days from sowing to flowering or seed maturity, rarely if ever have pure lines been based upon the inheritance of the individual phases. For this reason, lines that may be 'pure' for total length of the vegetative period frequently show degeneration or resolution into a series of hitherto masked types when cultivated under conditions differing from those in which they were created.

Again, it is not necessary to accept the phasic development theory in toto, as the term 'requirement of low temperature' may be used in place of 'thermo-phase', and 'requirement of daylength' in place of photo-phase. The argument is that these 'requirements' are heritable characters in the constitution of a plant; as they have been ignored in most genetical studies, they have remained frequently in a heterozygous condition. Upon this basis would be explained the puzzling behaviour of crosses made to study earliness or lateness, or the annual and biennial habit, as well as the degeneration of pure lines noted above.

It would appear to be desirable to make what the Russian biologists call phasic analyses, or what might here be called, less controversially, analyses of environmental requirements for development. This would provide the data necessary to support or contradict the alternative interpretations of the results of experiment noted in the early paragraphs of this chapter. If, as the Russians state, the requirements of a plant for temperature (so many hours below a certain temperature for wheat, for example) or for light (so many long days or short days, or short days followed by long days) are inheritable characters, then all the material of wheat, barley, rye and other plants already studied for inheritance of spring or winter habit, or earliness and lateness, would require to be reinvestigated on this basis. The filial generations would also have to be grown in a series of controlled environments or at different latitudes, in order that the potentialities of all plants for full development would be fully expressed.

From the Russian point of view, the segregation ratios noted in experiments such as those of Aamodt (1923) and Purvis (1939) ought to differ according to the environment in which they are grown, and the extent to which that environment was attuned to the individual developmental requirements of the plants. It is possible that the inexplicable behaviour of the Australian and English wheats described on p. 228 may also be due to the heterozygous condition of the developmental characters of the plants concerned, and to the fact that some particular component of the environment was inhibiting the normal expression of the developmental period at one or other of the latitudes at which the experiments were made.

By adopting this outlook, it may also be possible to explain the behaviour of improved 'pasture' strains of herbage plants under conditions different from those obtaining at the centre at which they were produced. It may also be possible to plan the breeding of pasture grasses on the basis of analysis of phases (or environmental requirements). A 'pasture' type of grass is necessarily a 'late' type, continuing to provide vegetative tillers throughout the grazing season. It is difficult to obtain high seed yields from them, the strains being known as 'shy seeders'. In multiplying seed of these strains, everything is done to maintain this late quality, by deferring the date of cutting as far as possible, for example. In selecting pasture types, the herbage breeder is actually selecting types upon the basis of the developmental physiology, although probably in no case has an analysis of the environmental requirements and reactions of the original material been made. Further, the selection of pasture types on the basis of their developmental physiology has been made only under one set of environmental conditions; it is not known whether that particular environment may be preventing the full expression of the plant's potentialities, and its true nature may not be revealed until the so-called 'pure' strain or combination of strains is grown in a different environment, whether natural or artificial.

Inheritance Studies on Duration of Developmental Stages

One example will be taken of what may be regarded as an intermediate stage between those geneticists who concern themselves with total developmental period and those who base their genetical work on a theory of the existence of individual developmental phases of a qualitative nature. Powers and Lyon (1941) studied the inheritance of the duration of three stages of development in certain crosses involving varieties of tomato (Lycopersicon esculentum and L. pimpinellifolium). The three stages recognized by these workers were: (1) number of days from sowing to first bloom; (2) number of days from first bloom to first fruit set; and (3) number of days from first fruit set to first complete change of colour of any fruit. The sum of these three stages is regarded as a measure of the earliness-of-maturity character.

The purposes of the investigation were to determine whether these natural biological periods in the development of the tomato plant are definite sub-characters and to ascertain the efficiency of the fit between obtained and theoretical means, based upon certain formulae, as a method of determining whether the effects of the genes differentiating the quantitative characters are arithmetically or geometrically cumulative. Throughout the study and interpretation of the data, an attempt was made to keep in mind the probability that the genes bring about the differentiation of a character by initiating, either directly or indirectly, developmental processes which no doubt in many cases interact among themselves.

From the experimental data it was found that the earliness-of-maturity character, as measured by number of days from sowing to first complete change of colour of any fruit, is composed of at least the 'three stages of development' noted at the beginning of this section. This fact is regarded as having considerable importance in breeding early varieties of tomatoes. The cross Danmark x Johannisfeuer is used as an example. Danmark has a short period of development from planting of the seed to first bloom but a long period of development from first fruit set to first complete change of colour of any fruit, whereas Johannisfeuer has a long period of development from planting of the seed to first bloom, but a short one from first fruit set to first complete change of colour of any fruit.

As the F1 generations of the crosses between the tomato varieties Danmark x Red Currant and Johannisfeuer x Red Currant have short periods for both these stages of development, it would appear that a variety could be produced re-combining the genes for short duration of the stage from sowing to first bloom, characteristic of Danmark, with the genes for short duration of the stage from first fruit set to first complete change of colour of fruit, characteristic of Johannisfeuer. 'Adverse linkage relationships would complicate the problem. Also, the nature of the interactions of the re-combined genes would determine whether a strain could be developed that possessed a short period for both of these stages of development. At the present time neither the linkage relationships nor the nature of the interactions of the genes are known.'

Solidago sempervirens is a short-day species. Goodwin (1944) working at the University of Rochester, New York, has isolated three strains from natural populations; these can be distinguished from one another by their dates of floral initiation, which are controlled primarily by the length of the photoperiod, and by their rates of bud development. From an analysis of F1 and F2 hybrid populations, it is estimated that the minimal number of gene substitutions which must determine the above physiological difference between strains approximates the haploid number of chromosomes, which is nine; it is considered probable that these genes are located in many if not all the linkage groups.

Before proceeding to a description of a Russian experiment conducted on the basis of the phasic development theory, passing reference may be made to the work of Baranov (1939), which produces evidence to prove the formative role of the environment, which is the basis of Lysenko's method of 'training' of plants to change their genotype.

Varieties of barley were tested in the Pamir, at an elevation of 3,860 m.; the forty-five varieties that had been reproduced in the Leningrad region (Pushkin) failed to attain full ripeness, and only twelve of these reached wax ripeness. Plants from seed of the same varieties that had been reproduced near Tashkent ripened fully. In subsequent years it was noted that in all cases the plants from seeds obtained under dry southern conditions ripened earlier and more completely; the shortening of the total vegetative period was due to shortening of the period from earing to ripening (the generative phase), as it was found that the period from seed germination to earing was in most cases longer than that in plants from seeds reproduced in northern latitudes. Subsequent reproduction of their progeny in the Pamir still further accentuated the initial differences. The conclusion drawn is that the environmental conditions of the place of reproduction may, in 2 or 3 years, induce profound changes not only in the phenotype, but also in the genotype of plants.

Breeding Based upon Analysis of Environmental Requirements

Two papers by Reimers on analysis of the respective environmental requirements of different species of lettuce (Lactuca saliva var. capitata) and the use of this information in breeding work will be used to exemplify the outlook of the phasic development school. In the first (1938) Reimers refers to earlier research by other authors in which it was established that both shooting and flowering in lettuce can be accelerated by keeping the germinating seeds at low temperature and cultivating the plant under conditions of prolonged day; shortening the duration of day retards development. Reference is also made to the attempts of Bremer to apply the variation in daylength to the study of genetic differences between winter, spring and summer forms of lettuce.

Reimer himself approached the same question of the cause of differences in time of shooting from the viewpoint of the theory of phasic development, and paid particular attention to the thermo- and photo-phases. Several varieties of lettuce were selected which differ sharply in earliness and are typical representatives of the various biological groups within the sub-species. The vernalization experiments were made on the varieties Ideal, Eier Gelber, Steinkopf and Trotzkopf. The photo-phase was studied in the same varieties with the exception that Steinkopf was replaced by Gr. Blonde Paresseuse. Ideal is regarded as a typical spring variety, Gr. Blonde Paresseuse as a typical summer variety, and the others as intermediate in type.

The purpose of the experiment, carried out at the Laboratory of Physiology of the U.S.S.R. Research Institute of Vegetable Production, Moscow, is stated by Reimer as follows:

1. To determine the duration of the thermo- and photo-phases in the representatives of the different groups.
2. To discover whether it would be possible to split the lettuce varieties into biological strains by the method of phasic analysis.
3. To ascertain the response to vernalization of varieties differing in length of vegetative period, when germinated seeds and green plants were used [this variant was introduced because some vegetables (cabbage, turnips) give a better response to vernalization when the treatment is applied to green plants rather than germinated seed].

Analysis of Temperature Requirement

Seeds were germinated until about 5 per cent showed the first appearance of rootlets; the seeds were then transferred to a room with a temperature kept within the range of 2.5 to 5 C., for periods of 10, 20 and 30 days respectively. Vernalized and control seeds were sown on May 5th. Growth became visible on May 11th to 12th. Different variants did not indicate varietal differences in appearance of 1st, 3rd and 5th leaf. The total number of leaves formed until the formation of the flower shoot was equal in all variants. Shooting began on June 10th.

Vernalization for 10 days caused an acceleration of shooting in the variety Ideal of from 3 to 5 days, but none of the other varieties were able to complete their thermo-phase in this time. It is concluded that the thermo-phase in the variety Ideal is passed in 10 days, but that in the other varieties it is longer, taking about 20 days.

Experiments on green plants indicated that, although the optimal temperature for completion of the thermo-phase by seeds is from +2 to +5 C., the same temperature has a deleterious effect upon green plants, inhibiting their growth and development. A comparison is made with other vegetables. As already noted, cabbage and turnips cannot be vernalized in the form of germinated seeds; radish can be vernalized both as green plants and as germinated seeds, while lettuce can be vernalized only in the form of germinated seeds.

Analysis of Light Requirement

Experiments to discover the conditions under which the photo-phase proceeds and also its duration were made on the lettuce varieties Ideal, Eier Gelber, Trotzkopf and Gr. Blonde Paresseuse. The plants were sown in plots in the open, and were exposed to daylight from the appearance of young growth to the end of vegetation: (a) for 10 hours; (6) for 14 hours; (c) from sunrise to sunset; and (d) during the whole day plus continuous illumination during the night. Shortening of the day greatly retarded shooting in the early (winter) varieties, but had little effect on the late (summer) varieties. The biological difference between the winter and summer varieties of lettuce is that the former have a short thermo-phase and a long photo-phase, while the latter have a long thermo-phase and a short photo-phase.

Cultivation of lettuce plants at different daylengths had a marked effect upon the morphological characters and behaviour of different plants within a variety. Under a natural daylength as well as in a 24-hour day, plants of all varieties remained uniform in all respects; when grown under a 14-hour day, and even more so under a 10-hour day, their behaviour and appearance were very diverse. Plants belonging to one and the same variety now differed markedly in general shape and other morphological characters, as well as in time of shooting. Thus a variety that has been carefully selected under a normal daylength in one region is nevertheless a population or mixture composed of biological races differing with regard to their physiological properties. In other words, they are heterozygous for the phases of development but this heterozygosity is not manifested so long as the strain is cultivated under the same conditions as those in which it was produced.


PLATE 21
The use of cereals in Canada to produce herbage of high nutritive value for conversion
into dried grass pellets for poultry feed (see p. 254). Photos: J. M. Appleton.


PLATE 22
Types of Poa pratensis grown at Weibullsholm, S. Sweden. That on left (few panicles) brought from
N. Sweden, and therefore out of optimal environment for flowering. Photo: E. Åkerberg 1928.

Selection Following Phasic Analysis

The experiments described above were continued (Reimer, 1939) to establish whether the length of the photo-phase (or the daylength requirement) observed in a given lettuce plant when it was grown in a reduced daylength would prove to be a hereditary character under the same conditions of development. If this was so, it was regarded as possible to breed for a new variety differing in its behaviour from the initial one, with a character such as absence of bolting. It was concluded that the simplest method to avoid early formation of floriferous shoots would be to select plants after the length of their photo-phase had been determined by phasic analysis. Two different sets of the varieties Ideal and Eier Gelber were selected on the following basis.

Ideal
No. 1. Various plants grown in natural daylength ('initial variety').
No. 2. The plant which had been the first to form a floriferous shoot in a 14-hour day (short-phase plant).
No. 3. The plant which had been the last to start stem formation in a 14-hour day (long-phase plant).

Eier Gelber
No. 4. Initial variety (as No. 1).
No. 5. Short-phase plant (as No. 2).

The seeds were sown in two prepared beds, and the plants grown in one of the beds received the natural daylength, the others 14-hour day. The plants grown in normal day were distinctive in form and size, no morphological differences between the plants of the initial variety and the other strains being observed. Again there were striking differences within the initial variety when grown in a 14-hour day, but within the progeny of the strains selected for short photo-phase and long photo-phase respectively there were no differences, irrespective of the daylength they received.

These results are only preliminary, but are regarded as proving the possibility and even the necessity of applying the method of phasic analysis to selection of lettuce plants, as has been done by Lysenko in the case of cereals (ref. No. 1 in Reimer, 1939, and Imperial Agricultural Bureaux, 1935).

Breeding Based upon Photoperiodic Reaction and Temperature Relations

The inheritance of the photoperiodic reaction has been studied by Lang and Melchers for Hyoscyamus niger (1943), Driver for potatoes (see Cambridge, 1944) and Sylven for forest trees (see p. 324) to mention only a few cases. Driver states that a large number of genes are involved, but the short-day determining genes seem to be dominant over the long-day ones. There is a great variation in the gene make-up of varieties. Whereas earlier varieties on selfing give a large percentage of long-day types, later varieties give a greater percentage of short-day types, while some South American varieties appear to be almost pure short-day forms. It is possible that some of the short-day types from South America may be of value in breeding potatoes for tropical countries, while breeding for the higher latitudes may be more successful if based upon the long-day or day-neutral types. Breeding for the warmer tropical climates may have to be based upon heat-resisting species, as the majority of the most promising Andean species are accustomed to temperatures no higher than those of the best northern potato-growing areas, and will probably be adversely affected by high temperatures.

Bell (1945) is using the reactions of sugar beet to light and temperature in his study of bolting in that crop. This is in two parts as follows:

(a) The isolation of strains showing extreme resistance to bolting.

Suitable adjustment of the length of the time for which seedlings are exposed to continuous light treatment while growing slowly under low temperature conditions during the winter results in a physiological resolution of the material. Under such conditions, a certain proportion of the plants will show no tendency to bolt after the young plants are transplanted into the field in the spring. Such non-bolting plants have been selected, grown for seed in the normal way the following year under strict isolation, but allowing the plants to interpollinate. The seed from each plant is harvested separately, for analysis of cluster characters, and the best lines retained for observation of bolting properties and root characters in the following year.

Very early sowing in the field of material obtained in this way has shown great reductions in the number of bolters. In one experiment in which sowing took place on February 8th, the average number of bolters in 42 lines obtained from non-bolting plants was just over 8 per cent, while the original material from which the non-bolters had been selected showed over 70 per cent of bolters. Re-selection on root characters is then practised within the lines most resistant to bolting.

(b) The study of the developmental physiology of selected lines in breeding material.

The physiology of the biennial habit in sugar beet is very significant in breeding work. The rate of development of the root with the consequent seasonal increment in root weight and sugar percentage is obviously of basic importance in attempting to increase the sugar production by selection. Plants pass into the reproductive stage, i.e. bolt, when they have reached a certain physiological stage when growing under field conditions, and this stage is determined by the developmental balance in relation to the biennial habit. In Great Britain early maturation of the roots is most desirable, but it is necessary to ensure that early root maturation does not accompany a tendency to bolt.

All selected lines are subjected to continuous light treatment in the seedling stage and are then transplanted into the field. Detailed observations are made on each line with regard to earliness, time spread, and total number of bolters. Periodic counts are made for dates of anthesis of individual plants within each line, and by this means a reasonably accurate assessment is obtained of the developmental physiology of the material. This study is conducted at the same time that the lines are being tested in normal field sowings for assessment of the root and general vegetative characters, and the scheme enables a valuable comparison to be made, in the same year and under the same conditions, of the two phases of growth and development. It has been found that individual lines show very characteristic behaviour in relation to bolting after the continuous light treatment.

It is well known that plant breeders use the photoperiodic behaviour of plants in breeding types with different flowering dates. One or other of the parents used in a cross is grown in an environment differing to some extent from normal in order that the pollen of the male parent is produced at the same time as the ovule is ready for fertilization, or vice versa. A somewhat unusual variant of this method is used by Fennell (1945) in his grape breeding programme at the Inter-American Institute of Agricultural Sciences, Costa Rica. Since grape pollen retains its potency for a very short time, one of the greatest problems of the plant breeder is to ensure a supply of fresh, viable pollen on the date when the seed parent is in flower. This is particularly the case in the humid tropics, an environment alien to most cultivated varieties; frequently it is practically impossible to grow to flowering size certain varieties of North American and European grapes most desirable for breeding. The prolonged dormancy resulting from insufficient low temperature in the warmer climates often retards the flowering of most varieties of the labrusca and vinifera species until a date too late for breeding purposes. This difficulty is overcome by a combination of the grafting technique with a knowledge of the varietal differences in the growth and flowering habits of both proposed parent vines, and a careful selection of stock vines especially as to whether they are early, medium or late in budding out.

The Russian workers, Vinogradova and Novoderizkina (1941), base their breeding work on herbage grasses on the environmental requirements of the species used. The whole breeding procedure can be much hastened by means of the glasshouse which enables a second generation to be raised, additional to that grown out in the field. Perennial grasses before being put in the glasshouse must have passed through the vernalization stage. Plants to be transferred to the glasshouse are usually put in pots or boxes at the end of August or beginning of September in order that they may be hardened and begin vegetative growth. Under the conditions described, the plants were left in the open until the end of November or beginning of December after which they were transferred to a glasshouse having a temperature of 3 to 6 C. where they were left long enough for the soil to thaw. Then the temperature was raised to between 12 and 15 C. From the end of December to the beginning of January the period of daylight was supplemented by electric light, ensuring a period of illumination amounting to 16 or 17 hours, for which purpose 230-watt lamps were used. The plants flowered early in March. It is advisable to raise the temperature to between 20 and 25 C. during the emergence of the spikes. The seeds ripen by the end of April or beginning of May.

This discussion of the relationship between developmental physiology and plant breeding merely touches the fringes of a large and complex subject, which may lead to problems such as the existence of different kinds of genes (Darlington, 1944) and related subjects.