Ann. Rev. Plant Physiol. 3, 265-306 (1952)
PHYSIOLOGY OF FLOWERING¹
ANTON LANG^2,3
Kerckhoff Laboratories of Biology, California Institute of Technology,
Pasadena, California

VERNALIZATION AND ITS INTERRELATION WITH PHOTOPERIODISM

The mechanism of vernalization.—The changes which take place during the low-temperature treatment itself will be discussed first in the treatment of vernalization for reasons of clarity. The changes which follow the treatment will be taken up afterwards. The finding that vernalization is reversed by high temperature is of paramount importance for the understanding of the mechanism of the process [devernalization (61, 62, 124, 191, 212)]. Within the thermoinductive range and within a certain range of treatment times we find typical optimum curves for the temperature effect in vernalization (72, 120, 1.89, 212). Devernalization, in turn, becomes more effective as temperature is increased (124, 212). Over a particular temperature range, a temperature coefficient of more than three was found [Stout (212)]. This suggests that we are dealing with two antagonistic processes, one promotive and the other inhibitory with respect to ultimate floral initiation, and that both processes are of a chemical nature. Low temperature is apparently necessary because the promotive process is less accelerated than is the antagonist as temperature increases and may have a lower temperature optimum.

For more specific ideas the following three findings are significant: (a) as the low-temperature treatment progresses, the reversibility of vernalization in winter annuals (winter cereals) decreases and finally disappears entirely (5, 62, 76); (b) in biennial Hyoscyamus niger⁷ reversibility seems to disappear if the plants after the end of the cold treatment are kept for five or more days at room temperature (124); and (c) while with comparatively short times of treatment the effect of varying temperatures is different, results tend to equalize with prolonged treatment periods, so that in almost any temperature which is effective at all, the same level of vernalization seems to be finally reached (120, 189). These results are in agreement with an interpretation of vernalization suggested by Lang & Melchers (124), which can be formulated as follows:

Floral initiation depends in some manner on the attainment of a definite condition, probably the formation of a specific chemical substance (C). C is produced by way of an intermediary, B. B, however, can be destroyed, [291] inactivated, or sidetracked in some other way (to D), and thus be lost for the production of C. The three processes differ in their temperature dependence in such a way that in high temperatures all of B is diverted to D, whereas in lower temperatures at least part of B remains available for conversion to C. (A, B, and C or D need not be successive links in a straight chemical chain, but may be related in some indirect fashion; for example, B may be an enzyme catalyzing the synthesis of C from a precursor of its own.)

Chemical changes involved in vernalizalion.—Our insight into the biochemical changes which take place during the action of low temperature are based almost exclusively on the work of Purvis and Gregory on winter rye (Secale cereale). They find, first of all, that low temperature treatment of whole grains is effective right from the start of the treatment and occurs without a supply of any outside material (except oxygen, see below). Excised embryos, however, require the supply of an adequate kind and amount of sugar, and the first two low-temperature weeks (approximately) have no vernalizing effect (lag period) (74, 187, 188, 189; see also 110). Second, the sugar supply to excised embryos is particularly critical in the earlier period of the low-temperature treatment (188, 189). Third, the lag period disappears if the embryos are left attached to their endosperms for the first few days of the cold treatment, but it is not affected by any treatment given at room temperature (189). Fourth, cold treatment is effective only in the presence of sonic oxygen (63, 65, 75). Last, if the cold treatment is interrupted by periods at room temperature given under an atmosphere of nitrogen its effectiveness is lost (75). Similar interruptions given in air reduce the effect of the cold periods much less (provided the durations are not too long) (216); periods in nitrogen given at the vernalization temperature are similarly harmless (75).

The following conclusions can be drawn from the first three points above. First, the processes which take place during the cold treatment and which ultimately result in floral initiation depend on the presence of some substance which, in entire grains, is supplied from the endosperm, but which, in excised embryos, must be synthesized by the embryo itself. This synthesis depends on the presence of sugar, very probably as an energy substrate. Second, this substance persists in the embryo only at low temperature while in higher temperatures it is subject to some sort of loss. We thus find a situation basically identical with the one indicated by the temperature studies and we are justified in assuming, tentatively, that the processes revealed in the two approaches are identical. The last two points indicate, in addition, that process II and possibly process I involve oxidative reactions, whereas process III seems to be fully effective in the absence of oxygen. A further [292] separation and characterization of the individual processes may be accomplished by the use of metabolic blocks. Certain enzyme poisons (arsenate and possibly fluoride) seem to inhibit vernalization in a specific way (46).

The cold-induced floral stimulus and its relationship to the daylength-controlled stimulus.—The temperature and metabolic studies carried out in vernalization show that a certain induced state must be reached in the plants for floral initiation to proceed and that the attainment of this state depends on the balance of several individual processes, some promotive and some antagonistic. This situation is the same as in photoperiodism, except of course that the environmental factor controlling the balance is not daylength but temperature.

In photoperiodism, the condition reached during induction was a transmissible floral stimulus—florigen. This fact was basically quite evident, since in the case of daylength the sites of perception and of response are clearly separated. In the action of low temperature, there is no comparable separation. The low-temperature effect in winter annuals can be perceived by the tissues of the embryo (see above and 71), the shoot tissue alone apparently being fully sufficient (186, 200), and in biennials by the stem tip alone (47, 54, 148, 149). However, if cold-treated individuals of biennial Hyoscyamus are grafted to noncold-treated ones, the latter are induced to form flowers [Melchers (148, 149)]. Thus, in the outcome of vernalization, too, a transmissible floral stimulus appears in the plant and we have to ask ourselves what the relationship of this stimulus is to florigen.

Some plants possess a low-temperature and a daylength (usually a long day) requirement simultaneously. If either of the requirements is of a strictly qualitative kind, as in biennial Hyoscyamus, floral initiation will take place only if thermal induction precedes photoinduction.⁸ If biennial Hyoscyamus is grafted to Nicotiana tabacum var. Maryland Mammoth, a short-day plant which requires no low-temperature treatment, that is, a summer annual, and if the latter is prevented from flower formation by the non-inductive daylength (long-day) conditions, the receptor forms flowers only if the donor has been vernalized and is exposed to the inductive daylength (157; see 156, p. 161; see also 211). Thus, the vernalization changes must be completed before florigen can be formed.

The simplest possibility would then be that the stimulus transmitted from vernalized to nonvernalized Hyoscyamus plants is florigen, for it is conceivable that the immediate cold-induced state, while promoting the formation of florigen, remains localized in the cold-treated plants. However, grafting experiments reciprocal to those described in the last paragraph, with [293] nonvernalized biennials as receptors and daylength-dependent summer annuals as donors, prove that this explanation is not correct. Such grafts were first made by Melchers (150) using again biennial Hyoscyamus (as receptor) and Maryland Mammoth tobacco (as donor). Flowers were formed by the receptors regardless of whether or not the donor was photoinduced and capable of floral initiation. We have to asssume the existence of two different transmissible stimuli: florigen, the production of which in long- and short-day plants is controlled by daylength, and another stimulus, the production of which in biennials is controlled by low temperature. To stress their separate natures, Melchers (150) has called the temperature-controlled stimulus "vernalin."

If this situation is general, a summer annual long-day plant maintained on short days should induce floral initiation in a noncold-treated biennial just as readily as the long-day-treated short-day plant did. In long-day plants, however, florigen formation is actively inhibited under short-day conditions, and this inhibition somehow acts at a distance. Thus, the situation is not so simple as it appears, and grafts between annual and biennial Hyoscyamus gave at first negative results (153). Some recent experiments, however, suggest that by appropriate timing the two effects can to some extent be separated. If the short-day-treated donor was removed 12 days after grafting, flowers were formed by 50 per cent of the receptors (151). The presence of vernalin can be demonstrated in this way in long-day summer annuals, too.

Two basic types of relationship between the two stimuli can then be visualized: (a) vernalin is a precursor of florigen, entering into its formation; and (b) vernalin acts as a catalyst of florigen formation. Attempts to extract vernalin (153) or to obtain a transmission without tissue union (155) have met with no more success than in the case of florigen. We have, therefore, only indirect means to attempt a decision between the two alternatives, but the evidence available is not sufficient to justify a definitive conclusion.

Unlike that of daylength, the effect of low temperature upon floral initiation is of an unmistakably quantitative nature: the longer the treatment, the faster the response, until the optimum level is reached (55, 120, 189, 190). We may safely state (cf. p. 289) that the primary product of vernalization is related to floral initiation in an indirect manner. One may, for example, think of this product as an enzyme catalyzing the synthesis of some compound necessary for floral initiation: the longer the low temperature acts, the more enzyme is formed, and the sooner will the threshold amount of the compound be attained.

At present, however, we do not have any conclusive evidence showing that vernalin is, in fact, the primary product of vernalization. It is just as probable that it is a secondary product catalyzed by some original, coldinduced state localized in the cells. Since vernalin is transmissible, this possibility may even appear as the more plausible one.

On the other hand, even if we were certain that vernalin is a secondary [294] product of vernalization, we should not be entitled to state definitively that it is a direct florigen precursor. In nonvernalized and in incompletely vernalized winter cereals, flower formation is frequently accelerated by limited periods of short-days ["short-day vernalization" (7, 147, 185, 190, 222)]. If fully vernalized, these plants usually behave like straightforward long-day plants, although in one case flower formation is reported to have been promoted in fully vernalized winter wheat (Triticum aestivum) by a period of continuous darkness (235). It appears that the conditions which are optimal for florigen formation are not optimal for the functioning of vernalin. It is possible that vernalin is light-sensitive, but only to such a degree that the sensitivity is not apparent unless the rate of vernalin formation or the amount of vernalin present is low. The existence of short-day vernalization does not rule out the possibility that vernalin is a precursor of florigen, but it does indicate that further processes may intervene. Before this possibility is settled definitive conclusions are premature. It also remains to be seen whether short-day vernalization proves a further clue for the understanding of short-day plants or whether we are dealing with an entirely separate phenomenon.

Photo periodism and vernalization in retrospect.—The analysis of photoperiodism and vernalization enables us to understand fairly well the general mechanism of the action of daylength and low temperature in relation to floral initiation. What significance do these results have for our understanding of floral initiation in general? The crucial point, as the grafting experiments show, is that flower formation can be induced in a long-day plant and in a short-day plant not only by another long- or short-day plant, but also by a day-neutral plant in which floral initiation is independent of daylength (cf. Table I, p. 270). Likewise, the floral stimulus which in biennial plants appears after low-temperature treatment is also present in summer annuals. This has already been referred to and is substantiated by the summary of grafts given in Table III. In both cases the stimuli pass with equal ease between plants of the same species, between different species, and between different genera. Transmission from host to parasite also seems to be possible (56, 100). The limiting factor thus is not taxonomic relationship but apparently the compatibility of the tissues of donor and receptor.⁹ We thus find that the stimuli demonstrated in plants dependent on daylength and in plants dependent on low temperature are not specific for these plants, but are operative and are identical in all plants. Daylength and temperature [295] dependence are only special cases in which the formation of the stimuli can be disrupted at specific points by the effect of specific environmental conditions. The significance of the study of the two phenomena lies in the fact that they are toeholds for penetrating into the general physiology of floral initiation.

TABLE III
LIST OF FLOWER-INDUCTION GRAFTS BETWEEN NONVERNALIZED COLD-REQUIRING
PLANTS AND NONCOLD-REQUIRING OR VERNALIZED COLD-REQUIRING PLANTS

An attempt has been made in this review to present the available evidence in a maximally integrated form, but without engaging in speculation unless it helps to clarify the issue. Several authors have advanced hypotheses which are intended to explain in a more specific manner either the entire sequence of the processes underlying floral initiation or certain phases of this sequence (7, 20, 25, 73, 79, 80, 123, 190, 208). Some of these hypotheses are, essentially, generalized transliterations of the experimental findings, while others involve rather detailed assumptions about the chemical or physiological changes leading to floral initiation and about their interrelations. The value of such hypotheses depends largely on the extent of our factual information. If we try to summarize this information, we will have the balance sheet which follows.

We have recognized several individual processes participating in floral [296] initiation—those involved in its daylength and low-temperature dependence. We cannot be sure, however, that we know all the processes involved in these phenomena and still less that they are all the processes which take part in floral initiation. We know that in either case the processes control the formation of transmissible stimuli. The stimulus appearing as the outcome of the daylength-dependent complex of processes in long-day plants is identical with that in short-day plants. It is the direct outcome of the activity of these processes and controls floral initiation in a direct manner. However, we cannot yet integrate the processes which control the production of this stimulus into one comprehensive story. The stimulus appearing as the outcome of the low-temperature dependent complex of processes is different from and has to be present before the former stimulus can be produced; but we do not know if this stimulus is the direct outcome of the low-temperature dependent processes or the exact relation of the two stimuli. The stimuli are most likely specific chemical compounds, but their exact nature is quite unknown. We have recognized the relation of certain chemical changes to photo- and thermoinduction, thus providing us with the first chemical approaches to flower formation; but this line of work is at its very beginning. We must finally bear in mind that, whereas the general features of photoperiodism and vernalization are well established and have been found to be identical in numerous plants, the physiological work on the two phenomena has been done in many cases with but one or two representatives of each type; its basis, therefore, is very narrow.

This balance shows clearly that there are many broad gaps in our knowledge. Consequently, a hypothesis designed to visualize the relations more clearly and to help in finding new approaches, is perfectly legitimate. If a specific picture is presented, however, it must be borne in mind that it is based not so much on the presence of positive proof as on the lack of sufficient information; it will therefore usually be one of several possibilities among which no decision can yet be made. Such hypotheses, therefore, easily run the risk of offering pseudosolutions of a problem which can be solved only by further experimentation.

SELF-PERPETUATING EFFECTS IN FLORAL INITIATION

In concluding the treatment of photoperiodism and vernalization, a feature must be discussed which may well prove to be one of the most interesting in the entire field. It is presently believed that the appearance of specific, autocatalytic materials is a crucial factor in development, accounting for the stability of the characteristics of differentiated cells and tissues. In floral initiation, also, the existence of such materials appears to be possible. If this possibility is borne out, the situation would be particularly promising, for the formation of the self-perpetuating material would be under the control of definite environmental conditions and thus be accessible to experimental variation. [297]

Indications of self-perpetuating effects are present both in daylength and low-temperature action. In Xanthium, the floral stimulus is not only transmitted to noninduced parts of the plant, but these parts continue to produce flowers indefinitely, even if the induced plant part is removed (81). This ability has been passed without any noticeable decrease through several graft "generations" (9; see 11). It is thus difficult to conceive that this ability is caused by the carry-over of a limited, although perhaps ample, amount of the original stimulus. The indirectly induced parts apparently continue to produce the stimulus themselves.

The evidence for self-perpetuating effects in vernalization is more indirect, but even more suggestive. The cold treatment of winter cereals can be applied at amazingly early stages in the life of the plant. It is fully effective not only in seeds in which germination has just begun and then further growth prevented by reduction of the water supply (144), but likewise in seeds developing on the mother plant or on cut-off ears (52, 53, 74, 111, 112). The changes which are caused by thermal induction thus persist through a considerable length of time in which the plant increases enormously in mass. If the treatment is given during embryonic growth, it is more effective in the younger than in the later stages (1). Thus, no loss whatsoever occurs in the subsequent development. Furthermore, if a plant is vernalized in those early stages, the induced state is not limited to the main shoot, but is present to the same degree in the side shoots (tillers), even though they are differentiated long after the end of the treatment. The low-temperature-induced changes seem to be perpetuated as the plant proceeds with its growth and to be maintained in all its cells.

The situation is not so clear in other cases. Photoperiodic induction can persist for extended periods of time, but in most plants it is definitely not permanent, provided the inductive treatment was not so extended as to determine all growing points of the plant to floral initiation (see 44, 45, 121, 143). According to Chouard (45), some LDP (Helianthemum guttatum, Nigella damascena, etc.), once photoinduced, do not revert to vegetative growth; but other cases comparable to that of Xanthium in which there is a maintenance of the induction through several graft generations have so far not been reported. In biennial plants, cold treatment of germinating seed is either not effective (cf. 151), or the effect is comparatively slight (175, 227). Treatment during seed development has not been reported. In Hyoscyamus the effect of optimal thermal induction does not noticeably decrease in the course of at least 100 days (124), but it has been observed that the effect of a suboptimal induction disappears in a shorter period (121). However, even in these cases the presence of a self-perpetuating material is not ruled out, for its rate of reproduction may lag behind the growth rate of the plant so that a continuous refilling is required.

We are thus led to consider the possibility that the effects of both photoand thermoinduction are self-perpetuating. Purvis (189) believes that the [298] course of vernalization itself—that is, the increase of the effect during the actual cold treatment—is autocatalytic. The multiplication of the self-perpetuating materials would be closely co-ordinated with the growth of the plant, for both in Xanthium and in winter cereals, differences in the degree of induction persist throughout the life of the plant as differences in the intensity and regularity of flowering or in the earliness of the response. We would seem to be dealing with materials which multiply regularly along with cell division. The multiplication would seem mainly to occur in the meristematic tissues, for in Xanthium only those noninduced parts of the plant in a menstematic condition at the time of the initial photoinduction are capable of the indirect induction (134). In the case of Xanthium we could assume that the self-perpetuating effect is identical with the floral stimulus which appears as the result of photoinduction. This would indicate that florigen is autocatalytic and would be another reason why it cannot be detached from living cells. In vernalization, the autocatalytic material may not be identical with the transmissible stimulus, but may be strictly intracellular, for we cannot decide whether vernalin is a direct or a secondary product of cold action (see above, p. 293).

THE LATER STAGES OF FLOWERING

The later stages of flower development.—Our insight into the physiology of the later stages of flowering is very deficient and consists largely of isolated pieces of evidence. The separation of some of the stages may seem questionable, particularly that between floral initiation and organization and that between floral maturation and anthesis. In most cases, a floral primordium, once initiated, continues to develop at least into a complete, microscopical bud, and once a flower has matured, it enters anthesis quasi automatically. In a few instances, however, this close relationship is broken, indicating that the stages do differ in some specific way. We have already seen that in Xanthium, floral initiation involves a reduction of the auxin level in the dark periods of photoinduction, whereas the further development of the inflorescence primordia is dependent on a sufficient auxin supply (p. 279). Similarly, in certain plants the unfolding of the flowers depends on a definite environmental stimulus and withholding this stimulus results in a delay of anthesis [for example, in Cereus grandiflorus (199)].

The stimuli of floral initiation and the later stages of flowering.—Since floral initiation is determined by specific stimuli, an obvious question is whether or not these stimuli also have some part in the further development of the flower. Most evidence indicates that they have. The time of floral initiation does not depend on the amount of photoinduction once the threshold value has been reached (p. 289); but both the number of flowers initiated and the rate and degree of their development usually increase as the inductive treatment is extended (for example, 79, 95, 121). In vernalization, not only the time to initiation is shortened as the duration of thermal [299] induction is increased, but also the quantity of flower formation [for example, the number of spikelets formed (7)] seems to increase. Some observations, however, indicate that the floral-initiating stimulus is not the only factor which controls the subsequent development of the flowers. The development of the differentiated floral primordia towards anthesis depends in some plants on environmental conditions different from those which determine initiation (60). Hyoscyamus plants which have received a suboptimal thermal- or photoinduction produce great numbers of floral primordia, which, however, develop only to small defective buds (121). In Fragaria, a substance has been extracted from photoinduced plants which promotes maturation of the initiated buds (205).

Reduction of the intensity of flowering is frequently accompanied by leaf-like development (phyllody) of the leaf organs (bracts) in the flower region. In Kalanchoë, however, the two phenomena have different causative mechanisms: the number of flowers depends on the amount of the floral stimulus, but the phyllody depends on the shortness of the dark periods which the plants receive either during or after photoinduction (89).

Flower development and auxin.—The growth of the flowers proceeds at first by cell division; later, cell elongation becomes a major factor, except in the ovary which may continue to grow to maturity almost exclusively by cell division (cf. 101, 202, 203, 207, 219, 220, 221). A priori, it is to be expected that auxin plays an essential part, at least in floral maturation. It seems, furthermore, that auxin is already essential in the division period of growth, for the ovaries of Cucumis anguria (gherkin) contain auxin from the very earliest stages of growth, and the decrease in growth rate prior to full maturity (before fertilization) coincides with a drop in the auxin content (172). It appears likely that the separation of the division and the elongation periods of growth in organs with a limited growth capacity is in general not warranted. The transition of the growth of Cucurbit ovaries from division to elongation—which occurs, moreover, in different tissues at different timesis not at all reflected in the growth rate of the whole ovary (204). It seems that cell division in such organs is an essentially passive process, determined by the size increase of the cells and the supply of materials to the fruit. The behavior of Xanthium shows that in the youngest stages the flowers or inflorescences are dependent on an auxin supply from other parts of the plant. In later stages, they produce auxin themselves, and this auxin may also control the growth of the flower or inflorescence stalks (6, 7, 97, 209). The time of onset of auxin production apparently varies in different species. In Secale it seems to be as late as the time of ear emergence (97). Most of the auxin is formed in the anthers (97, 234), where it undergoes a definite developmental cycle (97). If the anthers are removed, the growth of the entire flower bud (236) or of certain of its parts, like the pistil and the hypanthium of Oenothera flowers (226), may cease. The auxin of the anthers seems to serve as an auxin source for the other parts of the flower, with the exception of the ovary [300] which produces its own auxin (see above). Removal of the anthers is effective only prior to a certain point of development; later, the flower parts have apparently received enough auxin to finish development.

Differential effects in the development of male and female flower parts and flowers.—In general, the development of the different parts of a flower is highly co-ordinated and if conditions are unfavorable all parts cease growing simultaneously. In certain cases, however, a differential response of anthers and ovaries has been observed. In Lycopersicum esculentum (tomato) (102) and in Bryophyllum (194) low light intensity causes failure of anther development, while in Capsella bursa-pastoris the same effect is caused by unfavorable temperature and water supply conditions (115). Particularly interesting are those cases in which the sex differentiation in monoecious plants is affected. In Xanthium (171) and in Ambrosia (146) the ratio of female and male inflorescences or flowers is increased with extended short-day treatment; the total number may remain the same. In Cucurbita pepo (Acorn squash), low temperatures and short photoperiods favor the production of female over male flowers (173). Such cases may prove a tool for studying floral organization, a stage which has so far defied any experimental approach. It is possible that the effect of the environmental conditions is exerted through changes in the auxin level. Some of the effective conditions (short days, low temperatures) are believed to reduce the auxin level (although the evidence is hardly conclusive). In Cucumis sativus (cucumber) and in Cucurbita, auxin application causes formation of female flowers in sites which are normally occupied by male ones (116, 173).

Back to part 1

Literature Cited