Annual Review of Plant Physiology 11:  191-238  (June 1960)
VERNALIZATION AND ITS RELATIONS TO DORMANCY1
1The survey of literature pertaining to this review was concluded September 1959.
P. CHOUARD
Laboratory of Plant Physiology, the Sorbonne, University of Paris, Paris, France

HISTORY AND DEFINITION OF VERNALIZATION

In temperate countries, the seed of winter cereals must be planted before the end of winter in order to fruit within 12 months of sowing. By contrast, spring cereals, which generally can not survive the winter cold, flower soon after a spring sowing. The very notion of vernalization originated from this early agricultural experience.

Systematic research was undertaken on this subject as early as 1857, by Klippart (128). He showed that among the various climatic factors of winter, the determining factor is the cold temperature to which the young plants are subjected for a few weeks; this makes the winter cereals capable of flowering soon after the return of a warmer temperature. This notion was extended to other plants by von Seelhorst in 1898 (237) and much more extensively by Gassner in 1918 (89), who showed the great variety of "cold requirements" (measured by minimal chilling time) displayed by numerous species known as "biennials" or "winter annuals" as opposed to "spring plants" or "summer annuals" which have no significant chilling requirement for flowering. Gassner also showed that in winter cereals the early swollen germinating seed is already sensitive to the specific cold effect. Finally, in 1928, Lysenko (155) established that slight imbibition (e.g., 50 parts water to 100 parts dry matter) makes the cereal seed susceptible to this action of cold without inducing the excessive germination that could prevent use of a sowing machine. It was Lysenko who gave a name to the phenomenon. As spring cereals are called Jarovoe in Russian (from "Jar": formerly fire, or the god of spring), he called the process which makes a winter cereal behave like a spring cereal "Jarovization," and translated the word in his own English, French, and German translations into "vernalization" (Latin vernum meaning spring). Some French authors use the word "printanisation" but vernalization is most commonly used. In the next 20 years, a tremendous number of studies were undertaken on vernalization, mostly by agronomists, to quantitize the data on the process in cereals and in numerous other cultivated plants, in order to facilitate growing a crop following late sowing.

From their study of vernalization, Lysenko and his followers built their famous phasic theory of the stadial development of plants (156, 157) and [192] felt they had confirmed Mitchourin's theory of the transmissibility of certain characters acquired by adaptation. Two main groups have studied the physiology of vernalization on an accurate experimental basis: Gregory and Purvis (97 to 105, 206 to 215) in London using the winter rye Petkus; and Melchers, Lang et al. (139 to 146, 168 to 178) at Tübingen, with the biennial strain of henbane (Hyoscyamus niger). During all the development of that literature, "vernalization" was used with various meanings. Such disorder has to be clarified to prevent never-ending confusion. The original restricted meaning was: the substitution of chilling of a plant for the natural exposure to winter in order to make possible the initiation of flower primordia later. Vernalization induces or hastens the development of the capacity for flowering and although its action is not visible at first, it appears as an aftereffect. Chilling is the normal agent.

The derived meaning is: any physiological effect of chilling corresponding to the awakening of nature in spring. In this context any effect of chilling that has a forcing effect is called vernalization. A further derived meaning is: any physiological action stimulating the capacity for flowering, whatever the agent. In this sense vernalization was obtained by heat (e.g., 205) or by cold, by long days or short days, by light or dark, by nutrition, etc.

In this review, the original restricted meaning is used because it is the most precise and the nearest to the initial idea; this is implicitly admitted in the works of Gregory, Purvis, Melchers, and Lang, and is explicitly used by Wellensiek (274).

We reject the first derived meaning for it is expressed by the "breaking of dormancy." During dormancy, all organs or their primordia pre-exist, but either do not grow or grow slowly. The breaking of dormancy (often induced by chilling) removes growth inhibitions and allows an active growth as soon as favourable conditions return but does not directly cause the formation of new kinds of organs as does vernalization. The breaking of dormancy belongs to the physiology of growth or of vegetative development; vernalization belongs to the physiology of reproductive development. As an example, chilling potatoes to increase tuber yield, though commonly but wrongly named "vernalization," merely utilizes an aftereffect of cold on vegetative vigour and has nothing to do with the specific aim of vernalization which is to induce the capacity for flowering. Inducing this capacity in a nondormant rosette of biennial black henbane is true vernalization. Inducing dormant flower buds of lilac to grow and expand is breaking dormancy-not vernalization. Chilling is used in both of the examples, and relations between these two distinct but related phenomena will be discussed later.

I also reject the second derived meaning for it covers very different physiological phenomena such as photoperiodism, nutrition, or normal acceleration of growth and ageing by temperature rise. These phenomena are not related to the primary process of vernalization; i.e., a transformation of a winter plant into a spring plant. For instance, I shall not discuss here the so-called "heat-vernalization" which, in practice, has referred to such various [193] notions as normal growth acceleration by moderate heating, the phenomenon of thermoperiodicity, and sometimes experimental errors (241).

Another point to be emphasized concerning vernalization is that the process is revealed by an aftereffect. It is essentially a "preparatory process" to flowering, but is not the flower initiation process itself. In photoperiodism, the inductive photoperiod (or "euperiod") prepares for, as well as initiates, flower primordia. In bulbous Irises, subgenus Xiphion, studied by Blaauw (16), temperatures lower than 17° to 20°C. are necessary to initiate flower primordia, and this kind of chilling is not vernalization. During true vernalizing chilling, a sufficiently cold temperature is used, not to initiate flower primordia, but to create the capacity for subsequent flowering.

In summary, I will deal here with "vernalization" defined as follows: "The acquisition or acceleration of the ability to flower by a chilling treatment."

General reports or reviews on vernalization have been published during the last 20 years in Russian (157, 219), German (107, 139, 141, 172, 173, 177), French (36, 38, 41, 42, 63, 239, 274), Dutch (271, 277), and English (98, 140, 178, 183, 257, 279). In 1952 Lang published a general study on the physiology of flowering containing an excellent review of vernalization in the Annual Review of Plant Physiology (140). The story of vernalization of Petkus rye and henbane is so well described by Gregory (98) that no long description is necessary. The most commonly read book on this topic is probably Vernalization and Photoperiodism (183) published in 1948 by Murneek and Whyte; that fact explains why the present study deals mainly with investigations conducted between 1948 and 1959. Numerous important studies published before 1948 and here omitted appear in other references (140, 183, 279).

VARIOUS ASPECTS OF VERNALIZATION IN DIFFERENT SPECIES OF PLANTS

In order to work on a reliable basis I will first review the experimental data now available about vernalization in typical different species.

VERNALIZATION OF CEREAL GRAINS WITH SPECIAL REFERENCE TO THE RYE PETKUS

Cereals are monocarpic plants, growing initially as short rosettes but with elongated stems, the culms, when flower primordia have been formed. Nevertheless, they must be studied separately from other ordinary rosette plants like henbane because they have the peculiar ability to be vernalized in the seed stage, and the very large numbers of researches dealing with these plants have been conducted by trials using grain vernalization. The classical researches of Gregory and Purvis upon the rye Petkus are well known, so that only a short résumé is necessary.

"Petkus," a cultivated variety of Secale cereale consists of two races. The spring race is a typical annual rosette plant. When planted during the [194] spring under long days, the primary axis produces seven leaf primordia and then forms the young ear, recognizable by double ridges on the sides of the apex. In short days, it produces about 25 leaves before caring and may be described as a quantitative long-day plant. The winter race is a typical biennial rosette plant. When planted during spring, under long-day conditions, the primary axis produces 25 leaf primordia before caring, and 16 under short-day conditions; but if the seed has been previously chilled, it grows exactly like the spring variety (205, 206). The vernalizing effect of chilling has been studied over long experimental periods in this plant. Ear primordia have been formed under long-day conditions after the seventh leaf, even sometimes after the sixth, when chilling has been extended up to 90 days (94). Thus, in both strains the first six or seven lateral primordia formed by the apex are pre-determined as leaves, and "ripeness to flowering" is reached at the sixth or seventh leaf. From the seventh to the twenty-fifth leaf level, primordia are "labile" and can either give leaves and tillers exclusively, according to the photoperiodic treatment in the spring race, or in the winter race, according to the degree of vernalization and the subsequent photoperiod in accordance with the above interactions. The "labile" character of the seventh to twenty-fifth primordia is evidenced by the preinflorescent structure observed in the apex during this period (204). It is to be noticed that short days can here replace part of vernalization (94, 161, 206, 207, 266) as if the plant was a "facultative short-day" plant in relation to flower initiation. This is true at least as long as the winter strain has not been chilled; many cereal species show the same behaviour but it has not been shown that this last point is universally true. In all winter cereals, however, after vernalization the plant is a quantitative long-day plant.

The effectiveness of the vernalizing treatment may be measured either as the inverse of the cold exposure necessary to obtain the same result, e.g., the level of complete vernalization defined by caring after the seventh leaf under long days; or by the reduced number of leaves formed before ear initiation (this reduction ranging from 25 to 7 leaves) ; or else by the "score" (206, 207, 210), a statistical evaluation of the results observed on caring and emerging.

The effectiveness of chilling depends on several conditions.

  1. There must be a sufficient imbibition of water not necessarily the maximum, but at least the minimum imbibition necessary to allow the vernalizing process while low enough to hinder seedling growth, according to Lysenko's and Tolmacev's method (156, 258). In cereals, the appropriate water content is about 45 to 50 per cent of the dry weight, which means that 30 parts of water are normally required for 100 parts of air-dried mature seeds, with some variations according to the species or variety (80, 84, 97, 99, 106, 152). The two methods, maximum or minimum imbibition, do not always give exactly the same results in vernalization (94).
  2. The vernalizing process cannot begin before a period of activation [195] (the time necessary to induce the first apparent respiratory exchanges in the recently imbibed grain) of about 10 to 24 hr. at 15 to 18°C. after the beginning of soaking.
  3. The presence of air (oxygen) is absolutely required.
  4. An appropriate duration and temperature of chilling are needed. The effective low temperature is between +1 and 6 to 7°C. with a widespread optimum. It decreases from 0 to -4°C and disappears completely below -6°C. It also decreases above +7°C. and disappears completely between 12 and 14°C. (89, 107, 108, 210, 220, 222, 223). Above 15 to 17°C. the devernalizing effect of warm temperatures begins (214, 219).

In immature embryos, as in mature seeds or seedlings, the time required for complete vernalization is about the same: about 40 to 45 days in the winter strain of Petkus rye. A longer vernalizing chilling does not greatly modify the result; however, if it exceeds three or four months it seems to become less effective (210). When the treatment is applied to an older plant the minimal exposure to vernalizing chilling that leads effectively to the immediate acquiring of the flowering ability, decreases with age (81).

Respiration inhibitors and growth inhibitors inhibit the vernalizing action when they are applied at concentrations that completely stop respiration and growth. But low hydration or imbibition with a highly osmotic solution greatly slows down growth without reducing anything in the vernalizing process (155, 258). From this fact originated Lysenko's absolute distinction between growth and development (156, 279). By contrast, some other inhibitors applied at moderate rates can inhibit vernalization without stopping growth significantly (28, 48) and can be called "anti-vernalizing" agents.

The age of the plant in which chilling can effectively induce vernalization is extremely precocious. The immature embryo within the milky caryopse can be vernalized, either because of the cold nights occurring during late summer maturation in subarctic countries (131, 288), or local chilling of the ear on stalk (100).

Gentle drying of the wet seeds during chilling stops the vernalizing process but does not delete the acquired stage of vernalization. Interruptions of chilling by mild temperatures around 14 to 15°C. do not markedly alter the stage already reached in the vernalizing process. Alternating temperatures do not reduce the minimum number of required chilling days. Interruptions at 20°C. partly throw back the characters acquired but strengthen them when prolonged (86, 233). Under natural conditions the various cereal races have different chilling-time requirements that correspond to their adaptation to their native climate and to the fact that resistance to death by cold decreases during caring (219).

Devernalization instigated numerous discussions. The Russian authors observed that hot temperatures partly offset the vernalization effects (8, 77, 95, 255) ; but under the influence of the development of all plants through successive irreversible stages (7, 157 to 159), devernalization was [196] denied in various forms and nuances or interpreted as not being a real stage reversion (219, 248, 252). Purvis and Gregory have greatly enlightened the problem (101, 102, 214, 215). Two techniques ensure rapid devernalization: (a) exposure to about 20°C. in a damp atmosphere lacking oxygen (this exposure must not exceed about 1 to 5 days for it is most severe and would eventually kill the plants (102, 104) ; and (b) the exposure in ordinary damp atmosphere at about 35°C. for a longer period of about 8 to 12 days (101, 105, 213).

The effectiveness of both these treatments depends on the interval between the end of vernalizing chilling and the beginning of devernalization. When vernalization has been thoroughly achieved, a return to 15 to 20°C. fixes the vernalized state in a few days; at 4 to 5°C. the fixing requires several weeks. During the fixation process further devernalization loses its effectiveness (214). Revernalization may be achieved by another chilling.

The embryo itself directly perceives the chilling vernalizing effect (103, 207, 242), not the endosperm as was first thought (132, 165 to 167). Excised embryos and mere cauline fragments of the embryo, even when separated in the dry state before any diffusion of endosperm products, can be vernalized when imbibed (103, 113, 130, 190, 207). They must, however, be supplied with the sugars normally coming from endosperm. With their own metabolites, excised embryos can be but partly vernalized, even less when they have been starved in warm conditions without sugar before chilling. A sugar supply gives them back the complete ability of being vernalized (provided that starvation has not been excessive), but then a certain lag time will be observed which corresponds to the synthesis from sugar of some metabolites required for vernalization (99, 208 to 210). In winter cereals and some other perennials, minerals (3, 135, 185b) or vitamins (56) may enhance vernalizability, perhaps in relation to more active respiration. The meaning of the physico-chemical properties presented by vernalized meristems will be discussed later. The earlier emergence of the first leaves of cereals and most of all the smaller size of the first two leaves characterizes the vernalized state (108).

The vernalized state is not here transmitted by grafting, which is not generally successful in monocotyledons. But the vernalized state is communicated without reduction through cell divisions, i.e., in an autocatalytical way to all meristematic tissues originating at the apex of the embryo, e.g., to tillers even though they are formed after the vernalizing treatment (18, 108,207).

In cereals the organs already formed (young leaves, young culms) are most responsive to the elongating effect of gibberellin, but this substance cannot cause them to flower and cannot therefore replace vernalization by chilling (154). On the contrary, soaking nonchilled winter-strain seeds in natural extracts from vernalized embryos of rye (215) or of vernalized peas (116, 117) seems to be able to replace part of the cold vernalizing effect. [197]

But for slight quantitative differences, these fundamental rules of the vernalization of winter rye Petkus are the same for all cereals; especially for wheat (31, 82, 85, 147, 152, 161, 162, 229, 266, 282 to 284), barley (12, 70), oats (155 to 159, 202, 219, 239), rice (58, 244, 245) and different cereals (60, 89, 90, 155 to 159, 202, 219, 239, 252, 266).

VERNALIZATION OF BIENNIAL DICOTYLEDONS IN ROSETTES

The classical case of henbane.—Our classical information concerning vernalization was based on henbane as much as on Petkus rye, because of Melchers' and Lang's remarkable work (92, 139 to 146, 168 to 178, 228, 287). They worked together in Tübingen with other collaborators for a long time. In brief, I will only emphasize the differences observed between the behaviour of henbane and the cereals.

Here again the species, Hyoscyamus niger consists of two physiological races, the biennial and the annual strains. The important differences relative to the cereals follow below.

  1. Neither the seed nor the immature embryo have yet been vernalized; chilling is only effective on plants at least ten days old and in the rosette stage. The capacity for vernalization increases until the thirtieth day and then stays constant for a long time. That is to say, in the terms used by Wellensiek, this plant presents a "juvenile phase" for the first ten days of life after germination since the plant cannot feel the vernalizing effect of cold before the rosette is ten days old; or in my own terms, the plant reaches its "ripeness for vernalization" at that age.
  2. The leafy plant is strictly a long-day type.
  3. The winter strain has an obligate requirement for vernalization.
  4. Partial replacement of vernalization by previous exposure to short days is not possible.
  5. This plant shows a remarkable ability for grafting and for the transmission of vernalization from donor to receptor through grafts (168 to 172).

The effectiveness of vernalization can be estimated by the inverse of the number of days between the end of chilling and the beginning of shooting at 20 to 25°C. under long days. The optimum temperature for vernalization is lower and more widespread for the longer chilling periods. As an example of temperature and exposure period, it ranges between +3 and +6°C. for 42 days of treatment yielding complete vernalization and a short lag time before shooting of about ten days. Over-vernalization does not seem detrimental. Daily alternate temperatures have no increased effect (186b). During chilling, oxygen is indispensable and light is useful.

Henbane is also known as having a particular photoperiodic behaviour: its critical photoperiod ranges from 11.5 hr. at 20.5°C. to 8.5 hr. at 11.5°C. When plants are quite vigorous, with reserve food in the root, defoliation [198] is followed by flowering even under short days. Suppressing respiration by anaerobiosis in the dark, or sugar infiltration into leaves, also induces flowering under short days. In the biennial variety, vernalization is acquired by chilling the vegetative point and not by chilling the leaves or roots only. Acquired vernalization is maintained for a long time even after more than 100 days under short-day conditions at 20°C. It can be acquired by split chilling periods interrupted by mild temperatures.

As in cereals, devernalization can be obtained either by anaerobiosis or more easily by raising the temperature to 35°C. for a few days, provided that the devernalization treatment be applied immediately after the end of chilling. An interval of 3 to 4 days at 20°C. is sufficient to fix the vernalized state definitively. After devernalization, revernalization is also possible.

Graft transmission of the vernalized state is one of the most extraordinary properties displayed by henbane (169 to 171, 176 to 178). A donor scion can be grafted generally by side-grafting onto the rosette; the scion either is a branch or a separate leaf from one of the numerous Solanaceous species among which all possible types of vernalization or photoperiodic requirements are found. By this means, Melchers succeeded in making a nonvernalized, biennial, black henbane receptor flower by grafting with a donor. The donor could be a fragment from an individual of the same strain previously vernalized and exposed to long days, or a fragment of the annual strain, or a fragment from the Hyoscyamus albus species which is indifferent to vernalization and photoperiodism, or from annual species, or from varieties that are photoperiodically insensitive (such as Petunia hybrida, Nicotiana silvestris), or even Nicotiana tabacum var. Maryland Mammoth, the famous short-day annual. Since a leaf of this latter plant exposed to long days cannot bring its own buds to flower, it is not a donor of the so-called flowering hormone or "florigen"; however, because it induces flowering in nonvernalized black henbane, it must contain another sort of hormone which is still only speculative, called "vernaline" by Melchers (172). Zeevaart (289) confirmed these facts and demonstrated the same results using a potato branch as donor. The potato donor is more active when defoliated. Tomato, on the other hand, does not transmit the flowering ability under these same conditions. Except for the tobacco, Maryland Mammoth, the donor must always be floriferous itself. Reciprocally, a nonvernalized, long-day, black henbane cannot be used as an effective donor to make a leafless Maryland Mammoth branch flower. The black henbane does act as a donor if previously vernalized. The theoretical concept concerning vernalization have been based mostly on these remarkably successful experiments on the graft transmission of the vernalized state. But, as has just been seen, their success is not general and in receptors other than henbane, such a transmission is only quite rarely successful. [199]

Neither the serological properties (106) nor the measurements of respiratory intensity (112) have shown any significant differences between nonvernalized and vernalized plants.

Lang first successfully substituted treatment with gibberellin for vernalizing chilling using the biennial black henbane (142, 143). Whereas the formation of flower primordia precedes shooting in vernalization by chilling, shooting begins first and flowering follows later when gibberellin is used (263).

Other biennial plants in rosettes.—Studies with other species show that the classical data based on henbane need modification.

The industrial importance of beets resulted in numerous studies some time ago (12, 24, 51, 182, 192, 250, 251). Sugar beets, as well as mangelwurzel and other analogues, generally have an absolute vernalization requirement and their flowering depends absolutely on long days. Industrial sugar production utilizes these properties: plants must be sown late enough to suffer less cold than would vernalize them yet early enough to enjoy the longest possible period for vegetable growth and sugar accumulation in the tap-root. A sowing date can be fixed according to the variety and climate. By contrast, intentional vernalization yields flowers and seeds within the very year of sowing. Stout (251) showed that grafting a vernalized stock on a nonvernalized stock transmitted the flowering ability to the latter; this was confirmed by Margara (164).

More recent studies (275) show that vernalization can be achieved in very young plants and even in seeds, but responsiveness increases with age, beginning with the second month. Chilling rosettes deprived of most of their tap-roots did not give a vernalizing effect even though chilling only the root has no effect at all; therefore, in order to respond to the vernalization treatment, the vegetative point must receive what the root supplies to it (135).

Margara (164) extended his investigations to the various wild and cultivated Beta species that display the greatest variety in their physiological requirements. While the flowering of sugar beets and analogues generally requires absolute long days, the induction of shooting after vernalization can also occur in short days.

Short days show a quick and strong devernalizing effect which can act long after the end of chilling and when vernalization is already expressed by shooting or even by early flowering. The return to short days brings back the formation of perched rosettes and none of these plants reflower or grow if they are placed under long days again; everything must be revernalized by chilling. However, this revernalization is particularly difficult in more or less devernalized perched axillary rosettes. Gibberellin induces shooting but not flowering without chilling even for those perched rosettes about one metre high. In spite of prolonged gibberellin treatment, shooting finally stops, perched rosettes form again and now seem resistant to prolonged vernalization treatment. [200]

In Beta maritima, which requires less vernalizing chilling, the terminal bud is somehow protected against initiation of flowers and the flowering that follows vernalization occurs much like that described later for Geum, by the shooting of axillary buds while the main rosette ensures a maintenance of the perennial state. In this species, gibberellin entirely replaces cold in inducing flowering as well as shooting.

The small crucifer, Arabidopsis thaliana, in rosette presents the advantages of quick growth and small dimensions, and displays all degrees of vernalization requirements from none at all to absolute (136, 138, 290). Napp-Zinn's investigations (184 to 186b) recently made the variety "Stockholm" of this carefully selected species a classical research material. This variety absolutely requires vernalization to be able to flower, but it shows successive possible periods for vernalization, devernalization, and anti-vernalization.

Normally, the treatment consists in soaking very small seeds for 24 hrs. at 20°C. followed by 9 to 15 weeks from -3.5° to +4 to 5°C. A long previous soaking of seeds delays flowering (anti-vernalizing effect), with the optimum vernalization obtained by 10 weeks at 2°C. Heat devernalization can be completely effected immediately after chilling no matter what chilling period is applied to the seeds. Heat treatment after germination can also delay flower formation, whether or not vernalization occurs later. The vernalized state is fixed by a stabilizing treatment at 20°C. for seven days. Revernalization is possible. The number of leaves on the main stem is related to the time of flowering: heat treatment before or after vernalization or after stabilization reduces the number of leaves by about 10 to 20 leaves compared to that of untreated plants flowering at the same age.

Gibberellic acid applied to the seeds of races that have low or facultative chilling requirements replaces the action of cold if the treated plants are grown under long days. Indoleacetic acid has a similar but weaker effect on plants grown under natural autumn days after the treatment, and a still weaker effect under long-day conditions. This action may sometimes be additive to that effected by gibberellin, whereas triiodobenzoic acid is antagonistic.

In contrast, true "winter" varieties that have an obligate chilling requirement do not respond to gibberellin. Young rosettes of nonvernalized plants of these varieties that are more resistant than older ones to vernalization by chilling, are more responsive to gibberellic acid, which induces them to shoot and flower. Then, as these rosettes grow and become susceptible once again to the vernalizing effect of cold, they grow less susceptible to gibberellic acid. Moreover, contrary to the situation in henbane, when cold and gibberellin cause shooting and flowering they simultaneously determine the initiation of the inflorescence and of shooting (228).

Salad lettuce (Latuca sativa), depending on the variety, responds in different degrees to vernalization, which is not, however, an absolute [201] requirement for flowering (243, 267). They normally have a long vegetative growth period as a rosette, the leaves of which constitute the head. Under long warm days the axis shoots up, the head-leaves separate, and flowers appear. Head formation, followed by shooting, depends on light intensity, period, and colour, as well as the relative temperatures of day and night. The response to these factors, however, is modified by vernalization. The extent of this effect depends on the temperature to which the seeds (after soaking), or the seedlings, are exposed. Exposure of the Great Lakes variety for a minimum of 13 days, or better for 20 days, at about 3°C., followed by growth at a night temperature above 13°C. results in a rapid acceleration of shooting and flowering without head formation. At lower night temperatures a head is formed but flowering occurs sooner than in nonvernalized controls. Short interruptions of the cold treatment for a few hours per day with temperatures of 20° to 35°C. do not decrease the effectiveness of the vernalizing treatment (cf. mainly Rappaport and Wittwer (216, 217). Vernalized seeds release half as much indoleacetic acid as nonvernalized seeds when extracted with ethyl ether (87).

Other salad plants, like endive (Cichorium endivia), which respond quantitatively to vernalization and photoperiodism, are induced to bolt and flower by use of gibberellin (109).

Spinacia oleracea is considered to be a strictly annual, long-day plant with no vernalization requirement. Chilling the wet seeds of certain varieties however (e.g., Matador), between 0° and 5°C. for five days, causes plants to flower earlier under subsequent long days and to produce more flowers and seeds (126). The flowering of other varieties is generally accelerated by vernalization of wet seeds for two weeks at 3°C. (127, 187). Chilling treatments of 10 to 40 days speed up shooting, compared to the control, if shooting takes place under long 17-hour days preceded by a short-day period of growth. Such acceleration is less marked, when compared to the control, if growth is entirely carried out under continuous days. Plants of the variety Nobel will flower under short days when the seeds have been vernalized (264).

Campanula medium was the object of a remarkable study by Wellensiek et al. (268, 270, 272). The seed is not vernalizable and rosettes maintained at a warm temperature under short or long days never flower. If such rosettes are exposed at the age of four months to cold temperature for at least four weeks and are then returned to a warm temperature combined with long days, they shoot and flower. If, upon the return to the warm temperature, they are subjected to short days, they remain in rosette. These results show this plant to be an absolute biennial that flowers only after vernalization followed by long days, similar to the biennial Hyoscyamus niger. C. medium differs from H. niger in that neither defoliation nor grafting can make it flower without previous chilling. On the contrary, if this plant is grown in rosette up to the age of three months in a warm temperature and [202] under short days, and then exposed to the warm temperature again but in combination with long days, it shoots and flowers. Here the early growth under short days replaces the vernalization by chilling. This property may be called "Wellensiek's phenomenon."

Gibberellin causes most C. medium plants to shoot under long days but after a certain number of repeated applications, growth stops when the plants are about 20 to 30 cm. and perched rosettes form with no flowering (273). I personally obtained contrary results with certain individuals after the same treatment (43). This is not in contradiction to Wellensiek for I probably had plants from unselected horticultural strains, whereas Wellensiek's plants were selected genetically from ancestors that gave a variety of reactions to chilling. Among my own plants, some not only show "Wellensiek's phenomenon," but are also unable to flower under long days without vernalization in spite of gibberellin applications. In such cases, some of my plants, after they have formed the perched rosette, resume shooting when given repeated gibberellin applications and then after 20 to 30 cm. of elongation, set in rosette again for a few weeks and so on-until they are several metres high. This is of considerable interest since it demonstrates the different responses given by separate individuals that differ by slight genetical characteristics not evident under ordinary cultivation.

Oenothera biennis and other Oenothera constitute one of the most complex genera in relation to vernalization. The species must be determined most carefully. Mrs. Picard and I have observed (35, 39, 42, 43, 201, 202) that Oenothera biennis (in the pure race obtained from the Munich botanical garden) has an intense and obligate requirement for chilling followed by long days if it is to flower. Up to now, natural and complex winter chilling proved much more effective than attempts at artificial chilling and alternating temperatures appeared more effective than continuous cold. Seed vernalization has been ineffective. After natural vernalization, the plants exposed to short days remain in rosette but soon show that they have lost their vernalized state, for they do not react any more to long-day exposures. After a long devernalizing period under short days, their old rosettes even seem particularly difficult if not impossible to revernalize. Gibberellic acid repeatedly applied to nonvernalized rosettes causes shooting to a height of 20 to 30 cm., under long days, followed by another perched rosette which then seems almost fully resistant to further treatments.

With Oenothera lamarckiana (from the forest of Fontainebleau), and O. parviflora (from Hooge Weluve near Arnhem), on the other hand, even though the chilling and long-day requirements are also qualitatively obligate, they are much easier to satisfy quantitatively. Artificial chilling in a cold room under low illumination for 12 hr. per day at 3 to 4°C. for I to 2 months generally succeeds. A more effective treatment appears to be with temperatures alternating every week or half week at 3° and 10 to [203] 12°C. successively. Late devernalization by short days is also evident and revernalization is possible. Repeated gibberellin applications to nonvernalized plants induce shooting and finally flowering quite rapidly under long days; and sometimes even under short days but then much more slowly. Neither Wellensiek's phenomenon, nor flowering after defoliation as in henbane, could be observed in any of these plants (42, 201).

In Oenothera suaveolens, O. longiflora, O. stricta, and others the chilling requirement is not obligate. After a long delay, flowering may occur under long days without previous chilling. Other species are annual without any vernalization requirement, but they do require long days; still others that are annual are almost day-neutral. Up to now, grafting has not succeeded in inducing flowering by nonchilled plants that have an obligate vernalization requirement.

Celery (Apium) requires vernalization for flowering; as a result, active leaf growth occurs during the first year of culture. Exposure of young plants to chilling at 4° to 10°C. for a few weeks, followed by a return to warm temperatures, causes flowering within the first year of culture. Long ago it was shown that when the collar alone is chilled, with the main root remaining at the warm temperature, vernalization was effected, whereas chilling the root alone had no effect at all. It was one of the first examples (1930) to demonstrate that vernalization is sensed by the vegetative growing point, not by leaves or roots (54).

Carrots require different degrees of vernalization depending upon the variety of carrot in question (132). Recent Russian studies (6; 133 to 135; 219, 221) indicate that:

  1. Varieties from southern U.S.S.R. (Armenia and Daghestan), after the wet seeds are chilled at 2°C. for 50 to 80 days, have about one-third of the individuals flowering the same year.
  2. In varieties from more northern countries, seed chilling first seemed totally ineffective; but the following winter, after storing the isolates in a cool cave and then setting out in a hot greenhouse to accelerate growth, the plants from chilled seeds started initiating flowers two weeks later, but the others did not, suggesting that seed chilling had achieved part of the vernalizing process which would have been completed during further storage at a cool temperature (219).
  3. Chilling carrot isolates from a first crop causes complete vernalization of whole plants as well as of those from which the bud has been separated from the tap root after chilling and then planted alone. By comparison, when the bud has been chilled after removal of four-fifths of the root, the plant is not vernalized. The sugar-supplying root seems to play the same part here as does endosperm in the vernalization of cereal caryopses. The same fact had been reported for beets (134).
  4. Nonvernalized carrots grafted to flowering carrots or to an annual umbellifer ready to flower under long days (the dill) flowered under these [204] conditions (133, 135). Gibberellin may induce bolting and flowering without chilling (143).

Digitalis purpurea has long been recognized as requiring vernalization followed by long days to achieve flowering (5). The vernalization requirement still seems obligate; but the long-day requirement is less mandatory, at least in certain strains where I observed flowering after vernalization under short days. Wittwer, Bukovak et al. (280) showed that gibberellin causes these plants to shoot and flower without vernalization. I also observed this and showed that flowering is rapid in long days with frequent flower anomalies and the occurrence of perched rosettes on some axillary buds; but gibberellin also causes these plants to flower very late under short days (43). The other Digitalis species have various chilling requirements.

Scrofularia vernalis is a remarkable plant in many respects. It has a mandatory cold requirement. It is not seed vernalizable but natural chilling or artificial chilling, although less easily done, achieves vernalization of young rosettes as well as older ones, even those several years old. But after vernalization this plant is absolutely day neutral, proving that a long-day requirement is not necessarily linked with a mandatory vernalization requirement, as was often thought. Moreover, gibberellin causes considerable elongation of stems, even those of very young rosettes; this phenomenon is often associated with tumors issuing from cell proliferations in the medulla and cortex. Axillary shoots elongate too, but no evidence of flower initiation could ever be observed from such treatments. There appears here to be a sort of incompatibility between elongation caused by gibberellin and flower initiation whereas shooting after chilling is always followed by flowering (39, 41, 42, 43, 123).

Numerous other plants have been studied (38, 39, 43, 47). Many campanulaceous plants have a mandatory vernalization requirement, for example Symphysandra, Campanula persicaefolia [as was shown by Wellensick (273)], C. alliariaefolia, C. primulaefolia, etc. In the last two, gibberellin has no effect either on shooting or on flowering. Reseda luteola requires extensive chilling, but gibberellin causes it to shoot and finally flower provided that it is injected into the stem at abnormally high concentrations (43). Isatis tinctoria absolutely requires vernalization and gibberellin injections can only make it shoot with difficulty. Lunaria biennis [studied by Wellensiek (271, 273) ] has an absolute chilling requirement that diminishes with age, it is not seed vernalizable, and does not respond to gibberellin. This is also true for Eryngium variifolium (47) which, furthermore, doesn't grow at all when treated with gibberellin.

Several plants are remarkable because of the important modification exhibited by leaves after the chilling treatment. Cichorium intybus (274), the bitter Chicory, is seed-vernalizable (two weeks at 5°C.) [studied by Hartmann (111)]. After vernalization, the leaves are hairy and more serrated, they stand erect rather than spread out, and the plant flowers under long days with the new leaves sensing the photoperiod. [205]

Crepis biennis and Senecio jacobea, which absolutely require vernalization, at least as rosettes, show similar modifications (39, 43, 47).

These examples could go on without end. The preceding are sufficient to show that the vernalization process is much more complex and varied than suggested by the comparison between classical winter rye and black henbane.

VERNALIZATION OF CAULESCENT BIENNIAL OR MONOCARPIC PLANTS

Contrary to rosette plants, caulescent plants form internodes of various lengths from the very beginning of growth. Their flowering is not associated, therefore, as in rosette plants, with the deep change which represents the shift of a brachyblast into a long stem. Several remarkable plants of this category require vernalization.

Euphorbia lathyris.—This species is remarkable for its regular elongation. It forms internodes about one-half cm. long under short days, and 1 to 2 cm. long under long days. As long as the plant has not been chilled, it keeps growing regularly producing stems that can reach two or three metres in length or more. Some individuals with different genetic factors finally flower under these conditions when quite old; but most of them do not flower except after chilling. Natural or artificial chilling, whether applied to young or old plants, causes vernalization. After returning to a warm temperature, there soon occurs the formation of one apex flower and, by the subterminal branching of the shoot, the inflorescence. After vernalization, flowering is strictly day neutral. Gibberellin applied to nonvernalized plants causes an initial rapid elongation of internodes, which then returns to normal a few weeks after the end of the treatment. This is repeated with every application of gibberellin, but no flowering is induced (43).

Various species of Iberis, particularly intermedia ssp. Durandii.—After the initial growth the nonvernalized plant branches into numerous very leafy shoots that never flower under long days until the plant is chilled. After chilling and returning to a warm temperature, flowering occurs equally soon under short days as under long days, but there are many more flowers formed under long-day conditions. If continuously grown under short days, without chilling, a few flowers appear after a long time. This is another example of those intriguing relations between short days and the vernalizing effect of chilling. Gibberellin has no effect on flowering even though it is very effective in increasing elongation (43).

Cabbages (Brassica).—Cabbages are botanically and physiologically highly variable. Their various chilling requirements have been extensively studied from an agricultural and horticultural point of view (52, 71, 78, 129, 180, 200, etc.). Certain species or varieties are rosette plants with tap-roots; for example, turnips, in which shooting and flowering are accelerated by seed vernalization, and "choux-navets" and rapes "rutabagas" which obligately require vernalization of leafy plants before initiating [206] flowers. By contrast, other stay caulescent; such as the fodder strains of Brassica oleracea, or the oil-seed cabbages (B. campestris) "Coiza" and "Navette" which are annual strains with no vernalization requirement, and strictly biennial strains which obligately require chilling of leafy plants. Others are caulescent first and then form perched rosettes with numerous leaves. Among these are the ordinary garden cabbages (B. oleracea capitata) that show various chilling requirements depending on the strain. Gibberellin is an effective substitute for chilling in certain strains, sometimes causing an extensive elongation and occasionally flowering, but it is ineffective in other cases. A natural gibberellin-like substance has been found in the young floriferous shoots of certain cabbages (153). There are biennial cabbages that flower without chilling after grafting on flowering cabbages (135) or on annual mustards or on annual B. crenata (55).

Brussel sprouts (B. oleracea gemmifera) were studied by Stokes and Verkerk (249, 260). They showed that if this plant cannot flower without chilling it is not because of a real vernalization requirement, but because flowering cannot occur in ordinary warm or mild conditions. Ripeness to flowering appears after a juvenile stage and requires 30 nodes. From then on, exposure to cold temperatures of 5 to 10°C. induces flowering. Exposures taking place at such temperatures therefore initiate flower primordia and are not a "preparatory" process as is vernalization. Chinese cabbage (or Pak-choi: B. sinensis) seems to be similar (286). White mustard (Sinapis alba) is an annual plant in which wet-seed chilling slightly speeds up flowering (59). Indian mustard (B. juncea) shows a higher vernalization requirement which may be satisfied by seed chilling; immediate heat devernalization is not effective (238) as is true of cereals.

Many other cases of vernalization requirements of various degrees have been reported for other crucifers and in particular for radishes (61), Lunaria biennis (271), stocks (Matthiola incana), Erysimum perofskianum, Cheiranthus Allionii, the Siberian wallflower, etc. (79, 203, 273, 274).

Tomatoes.—Tomatoes are considered typical annual plants with a remarkable responsiveness to daily photo-thermo-periods. However, they display some noticeable phenomena related to vernalization (93, 124) the stimulating effect of wet-seed chilling has been reported by Russian authors particularly e.g., 10 days at 0°C. or 14 days at alternating temperatures of 18°C. and 0°C. in relation to crop precocity and yield (253, 254, 265). In other cases, particularly in Ailsa Craig and Potentate, other authors (18, 281) showed that wet-seed chilling had no effect. After the cotyledons open, however, if the seedling is exposed to 8°C. for nine days instead of being maintained under warm conditions, the first inflorescence will be formed one or two nodes earlier. When chilling is extended for a few more days, this first inflorescence will contain more flowers. Then, by successive periods from week to week or more, similar chillings stimulate the earliness of further inflorescences. Nightly cold temperature is mostly effective dur­ing [207] this kind of vernalization (13, 261), and for tomato growth generally.

Annual legumes.—Numerous cases of vernalization have been reported in caulescent annual legumes for example: sweet peas (125, 239), vetches (160, 239), lentils (130), chick peas (122, 194 to 196, 239), broad beans (240), etc. In many of these plants and particularly in vetches (239) the treatment increases the rapid growth of lower axillary buds, gives a more branched habit to plants, and makes them flower earlier.

Finally, the garden pea (Pisum sativum) reveals most interesting phenomena; the first flower is formed on a precise, genetically defined node for each variety (114). The next lower nodes bear aborted shoots that have a tendency to flower. Lower nodes later form the axillary leafy shoots. Often, dwarfed early varieties will generally present their first flower at a low node, the seventh for example, and are strictly indifferent to chilling (193). By comparison, late varieties that flower on higher nodes (the tenth or seventeenth for example) as Zelka, Unica, Greenfast. Alaska, Téléphone, etc. are markedly influenced by the chilling of recently soaked seed (62, 117, 121, 197). Exposing plants to a few days of 3° to 5°C. temperatures prior to growing them under normal conditions induces the appearance of the first flower 1, 2, or 3 nodes lower than in controls. Leopold & Guernsey (150) have shown that this effect is increased by soaking the seeds destined for chilling in an extremely diluted auxin and thiamine solution and that the effect increases with concentrations ranging from 10-7M to 10-4M; but triiodobenzoic acid increases the vernalizing effect at a concentration of 10-7M, whereas at higher concentrations it reduces or inhibits the chilling vernalizating effect. After vernalization by chilling, even when enhanced by these chemicals, exposure to 39°C. for 18 to 24 hrs. totally removes the vernalizing effect. The same thing happens after a total lack of carbon dioxide for at least 1 hr. immediately after the vernalizing treatment. The carbon dioxide appears to be necessary to fix the vernalizing effect.

Highkin (117) showed that if seeds were exposed to 20°C. for some days after soaking and then subjected to chilling instead of being directly exposed to chilling immediately after imbibition, the vernalizing effect of chilling was completely prevented by a 5-day exposure to a warm temperature. He further showed that if the water used to soak pea seeds for 25 days at 4°C, was then used to soak other pea seeds which were immediately planted without chilling, it caused a significant reduction in the age of the appearance of the first flower (116). It was reported that this could also be obtained with the soaking water of chilled winter rye seeds and that these various soaking waters could speed up flowering not only in the pea, but also in other vernalization-requiring plants such as winter cereals. The active substance could be guanosine or something similar (118).

Gibberellin also effects the nodal placement of the first flower in peas. It greatly increases the internode elongation of dwarf varieties (this effect [208] was sometimes called "vegetative vernalization," a possibly erroneous usage). More & Bonde (17, 181) showed that on late dwarf Téléphone peas, gibberellin applications delayed flowering slightly by bringing the first flower node to a somewhat higher rank. When seeds from this variety are chill-vernalized, the rank of the first flower node is lowered; but if gibberellin is applied at the same time as chilling in the soaking water, it prevents the chilling vernalizing effect. When it is applied immediately after chilling, it destroys the effect of chilling and corresponds to a "reproductive devernalization."

Paton & Barber (198) and Haupt (114) showed that grafting a late variety on an early one, or even on a flowering late variety, causes an accelerated scion flowering by lowering the rank of the first flowering node. This effect occurs even though the scion is set with its own cotyledons and leaves. However, the effect is stronger when the scion is defoliated. The stock used as donor is much more effective if it keeps its own leaves but it is still effective if set in interstock, the stock being from a late variety. These observations show that the fixing of the first flowering node in the pea depends mainly on a stimulating substance, but that an inhibiting substance may also occur, and that the final result may come from a balance between both effects.

VERNALIZATION OF PERENNIAL OR POLYCARPIC PLANTS

The vernalization requirement in perennials has often been cited incidentally (23, 188, 199). However, the association of perennial growth (= polycarpism) and a vernalization requirement appears contradictory. In a plant that requires vernalization, the effect of winter cold is transmitted to the buds arising from the treated bud; therefore, when spring arrives, all of the buds should ultimately produce flowers and the plant should be monocarpic. Since the plant is perennial, one of the following alterations of the normal vernalizing process is likely to have occurred (42): (a) some of the buds are devernalized during summer; (b) some of the buds are not vernalized during winter chilling; (c) vernalization is not indefinitely autocatalytic among the second generation of buds derived from adventitious buds arising in the tissues that were not meristematic at the time of vernalization. Let us see what is the result of studies, especially those of the English authors with Chrysanthemum and of the French authors with various genera such as Geum, Scabiosa, Teucrium, etc.

Perennial plants in rosettes.—In the vegetative stage, Geum urbanum and one other Geum form short rosettes consisting of leaves without stipules; flowers occur on long axillary stems bearing leaves with stipules. The plants remain perennial by a rosette main axis which seems absolutely protected against all possibilities of flower formation. We showed [Chouard (43, 47), Chouard & Weber (50), Lé (149), Weber (269)] that before the autumn [209] cold, no axillary bud had yet been transformed into an inflorescence primordium. The normal chilling time for vernalization is at least five weeks, at +10 to +4°C. This chilling confers a flowering capacity to the buds located in the axil of the younger visible leaves and some of the still younger leaves. Everything occurs as if the youngest, barely differentiated buds were indifferent to chilling vernalization. They react to vernalization as they become more differentiated; finally, as they grow older, their state is definitively fixed, either in the vegetative stage if they have not been vernalized in time, or as inflorescences if they have been vernalized in time.

Older axillary buds on the outer part of the rosette generally stop growing because of the correlative inhibition exerted by the active terminal bud. If the latter is taken off, outer axillary buds start growing—but in rosette only. Then, if they have undergone chilling, those among their own axillary buds which were responsive at that time and were vernalized, flower.

When the vernalizing temperature is below 1°C. no initiation appears during the minimum chilling period, but then, after warming up to 4°C. to as high as 30 to 35°C, the evolution of vernalized buds into inflorescence primordia followed by the shooting of these inflorescences and their flowering occurs slowly. The process occurs more quickly at cool temperatures and at the normal rate at warm temperatures. Hot temperatures cause accelerated secondary branching. Vernalization occurs around 4°C. and is immediately followed by a slow tissue differentiation which goes on actively even at such a low temperature.

The most easily vernalizable buds are those which, at the time of the chilling treatment, consist of their own apex meristem with a few roughly differentiated leaf primordia. If the chilling time is extended to 10 to 15 weeks, more buds are vernalized, particularly those that were too young at the beginning of the treatment. If the treatment is prolonged for 30 to 50 weeks, a remarkable phenomenon occurs : the terminal apex itself, as well as the older outer buds of the rosette, is vernalized. It follows that all the existing buds become vernalized. They form inflorescences after a return to warm temperatures and also form a terminal inflorescence (which never occurs in any Geum in any natural climatic condition of the world). These polycarpic plants become monocarpic and die after complete flowering of all their buds.

By taking off a leaf with its axillary bud and planting it, the bud can be freed from correlative inhibition and reveal whether it is vernalized or not. Inflorescences then form several leaves with stipules before ending in a flower at their apex. Without exception all axillary buds of an inflorescence are vernalized and able to flower. In a few exceptional cases, we could obtain adventitious buds on the petiole of leaves with stipules on flowering shoots. These newly formed buds are not vernalized and only give rise to rosettes.

Flower initiation, growth, and flowering following vernalization are [210] absolutely day-neutral. We could never obtain devernalization either by short-day treatment or by a high temperature (35°C.) even when applied immediately after chilling. The continuance of the vernalized state in an autocatalytic manner in all successive generations of buds of inflorescences is indefinite. Cutting young flower buds or treating them at a high temperature (30 to 35°C.) induces considerable axillary branching in the inflorescence yet all successive generations of buds still can flower. Shifting nonvernalized plants from short days to long days does not cause any flowering (that is to say this plant does not show Wellensiek's phenomenon).

Up to now, gibberellin has not caused Geum to flower. In the absence of chilling gibberellin induces the shooting of some axillary buds or even of the terminal bud, which then forms an erect leafy stem. Leaves take on some of the form of flowering shoot leaves but do not have stipules. In spite of repeated gibberellin treatments, these elongated shoots finally form perched rosettes.

Certain individuals 1 to 10 per 1000 at the most) in Geum urbanum show a reduced vernalization requirement. These individuals grow one or two inflorescences without chilling, after 1 to 3 years of growth in warm temperatures. They may sometimes be induced to flower by gibberellin. They seem to belong to physiological races which differ genetically from the common type. In G. bulgaricum and G. intermedium this situation is true of most individuals, whereas in such species as C. canadense, G. album, G. macrophyllum, etc. we have only observed up to now a mandatory vernalization requirement, similar to that generally found in G. urbanum.

In Geum, mature achenes are dormant due to tegument inhibitions. Excised mature embryos are not dormant. Neither achenes nor excised embryos have yet been chill-vernalized. But young seedlings having only two small leaves in addition to the cotyledons can respond to vernalization and can flower on a young axillary shoot.

Geum, and especially Geum urbanum, must be considered as a new type of vernalization-requiring plant groups, like Petkus rye and henbane, which were the first two classical types.

Scabiosa succisa (=Succisa pratensis = S. praemorsa) behaves much like G. urbanum particularly in relation to the permanent inability of the terminal bud rosette to flower. Normally, vernalizing chilling induces the ability for further growth into an inflorescence in the two main axillary buds at their very young responsive age. Here however, under long days, after a sufficient delay for differentiation, inflorescences grow rapidly with long internodes and then flower. Under short days, the phenomena are slower, internodes are shorter, and flowering occurs much later. After vernalization, this plant therefore is a "quantitative long-day plant."

We first used plants from the "forest of Breteuil," in Perche. All plants showed the following remarkable reactions. When grown under short days at a warm temperature up to the age of a vigorous rosette the plants, when [211] placed under long days at warm temperature, soon flowered as if vernalized, thus reproducing "Wellensiek's phenomenon" as found in Campanula me­dium. Moreover, repeated gibberellin treatments induced the shooting of axillary buds as if they were vernalized. Under these conditions, they flowered either slowly with short days or faster with long days and finally attained about the same length. Experiments carried out on individuals from other countries, such as Burgundy and the forests of "Brie," showed much greater vernalization requirements. In these races, gibberellin induced temporary shooting as in Geum, but no flowering, and Wellensiek's phenomenon did not occur. We have separated different clones from different individuals of the same population: each one behaves consistently but may differ physiologically one from another. Again this example shows how different the behavior of different strains of the same Linnean species can be (35, 42, 45, 47, 115).

Perennial plants in rosettes with terminal inflorescence.—Contrary to both preceding types, the following plants normally flower after transformation of the main axis into an inflorescence. These perennials, which require vernalization of various degrees, are probably more numerous than formerly thought. Many perennial meadow Graminaceae, studied by Cepikova (23) and more recently by some others (53, 168, 199) but mainly by Wycherley (285) initiate inflorescences on main shoots after chilling. Cynosurus cristatus requires 15 to 30 days at 5°C. or less and active flower initiation may follow, even at such low temperatures. It is markedly speeded up by long days; returning to short days can force some of the inflorescences back into rosettes. Lolium perenne is easily vernalized by only 15 days at 5°C. Dactylis glomerata requires a much longer chilling period. Poa pratensis, Dactylis glomerata, as well as a strain of Cynosurus and probably Agropyrum repens seem to be vernalizable but on shoots brought into a state of short and stout shoots by a short-day treatment (53, 88, 188, 195, 199, 285). 1 also observed mandatory or substantial vernalization requirements in Poa supina, various Festuca, Bromus, etc. (47). Their behavior seems to display significant differences from clone to clone and will soon need careful investigation, especially the requirement for short days before or during chilling.

Saxifraga rotundifolia obligately requires vernalizing chilling; the subsequent flower initiation is photoperiodically day-neutral, but the length of the flowering shoots and the number of flowers are much greater with long than short days. Many perennial Crucifers behave similarly, except for certain rosette types, e.g., Draba aizoides and D. hispanica, which normally grow as small rosettes with numerous pointed leaves. These must be vernalized and then they initiate flowers even at relatively low temperatures and finally flower irrespective of the photoperiod. When maintained constantly in a warm temperature, the rosettes grow a short, sometimes flexible stem with numerous leaves, that never bears flowers. Here gibberellin increases [212] elongation without inducing flowering. It does not appear that the axillary buds can be vernalized by chilling. In the genus Erysimum, numerous perennial species also require vernalization, either obligately, or almost so. Without chilling, certain species remain as mere rosettes while others, especially at warm temperatures, grow extensive leafy stems of various lengths. Gibberellin generally, increases elongation without inducing the flowering that occurs normally after vernalization irrespective of the photoperiod. Dianthus, which we will refer to again, can be classified here. Various species of Primula, many of Potentilla, Agrimonia, Lychnis coronaria, viscaria, and flos-cuculi, Cardamine amara, various species of Silene, Scrofularia alata, etc. under observation for numerous years, proved to have mandatory vernalization requirements. In Eryngium variifolium, the vernalization requirement seems important and gibberellin has been totally ineffective in causing even the least elongation or modification of the plant habit (43, 46, 47).

In Scrofularia alata, a perennial absolutely requiring vernalization, gibberellin induces considerable elongation under either long or short days, but no flowering. However, after an insufficient period of chilling that has no apparent effect, gibberellin induces both elongation and flowering. The small seeds may be vernalized, contrary to those of henbane which cannot be (148). When fully studied, Scrofularia alata could become the other exemplifying type for perennial rosette plants requiring vernalization.

Scabiosa canescens (= suaveolens), after germination, produces simple obovate leaves that spread flat on the ground. After chilling, newly formed leaves are deeply pinnatisect into fine segments. The plant then grows and flowers quickly with long days, slowly under short days. Without vernalization, plants grown from seeds from an apparently homogeneous natural population display great genetic variability. Some plants of this category remain in rosettes for years with simple young leaves; others grow longer after a certain time lag but retain the simple leaves on a long stem for prolonged periods; others spontaneously form divided leaves quite early while growing either slightly or extensively and finally flower late. This plant, perennial according to the nonvernalized or devernalized shoots with simple leaves formed at the base, illustrates very well the leaf polymorphism that follows vernalization as was reported for bitter Cichorium. It also shows that the elongation and the different leaf forms that necessarily precede flowering, can either be linked or separate. Therefore, they are possibly independent phenomena (43, 47).

Caulescent perennial plants.—Before vernalization, certain of these have a short stem, differing little from an elongated rosette. Dianthus (carnation) (43, 47), displays all possible types of vernalization requirements: none in D. superbus which keeps flowering again and again under long days from bottom sprouts; low in D. gallicus, seguieri, campestris, etc. which elongate into creeping or stoloniferous stems and flower quite abundantly, but earlier [213] and still more abundantly after vernalization; and important in D. arenarius, attenuatus, geminiflorus, and caryophyllus, which develop short stems of which the strongest shoots produce 1 or 2 flowers after 2 to 4 years of growth under warm conditions. In addition, the requirement is mandatory for numerous species with rare individual exceptions: D. deltoides forms long, very leafy shoots that never flower without vernalization or else flower very late and poorly; D. coesius and D. carthusianorum, remain as very short stems with heavy masses of leaves; D. neglectus and graniticus (with slightly longer stems) and D. squarrosus, with long stems, never flower without vernalization. D. barbatus (268, 271) forms stalks with very dense leaves and must be vernalized except for one strain that keeps flowering constantly and which, through a graft made by Waterschoot (268), induced a vernalization-requiring plant to flower. Gibberellin appears not to induce flowering in carnation in place of vernalization. Cultivated carnations (from D. caryophyllus) have more moderate requirements for flowering, e.g., everflowering carnation (14) must have formed 9 to 11 pairs of leaves before it can be vernalized. When chilling is then supplied, they can flower under long days at the fourteenth node. Flowering is markedly later under short days or without chilling.

Many Leucanthemum, mainly L. cebennense (= monspeliense) (43, 47), and probably Pyrethrum cinerariaefolium (91) form short bushy stems that only bear flowers on the terminal axis after vernalization under essentially day-neutral conditions. However, the pedicels are very short under short days and very long under long days. In these plants, as in carnation, very young or insufficiently robust shoots at the time of chilling are insensitive to vernalization and thus ensure the perennial condition.

We also find typically caulescent perennial species (42, 43, 47), Teucrium scorodonia is one of the best examples, forming suckers and leafy shoots. It never flowers unless previously vernalized. After vernalization it requires long days, although not always. The terminal shoots terminate in inflorescences. The secondary shoots, leafy lower sprouts, and subterranean stems or suckers maintain the perennial condition. Repeated applications of gibberellin can induce an elongation of several metres, but no flowering. Teucrium scordium continues in the form of creeping shoots without flowers when nonvernalized (47).

Many woody plants (bushes or even trees) have similar requirements (47), e.g., Medicago arborea, in which a chilled twig flowers without this ability being transmitted to the other nonchilled twigs of the same plant. Wellensiek also showed vernalization requirements for the woody crucifer Cheiranthus cheiri, the wallflower (273, 274). I found Alyssoides (=Vesicaria) utriculatus to also require vernalization. Obligate vernalization requirements are also found in various species of Cistus, in Salvia (officinalis, triflora, lavandulaefolia), in Lavatera (47), etc. It is also very likely that large trees such as the olive tree (Olea europea) have varying vernalization requirements depending on the variety (4, 110). [214]

In other species, such as Geum and Succisa, the apical bud is protected against vernalization, but they are typically caulescent. An example is Anagallis tenella, which covers the ground with stoloniferous stems, mandatorily requires a long vernalization and then the axillary buds flower only under long days. Vernalization does not seem to extend to axillary buds other than those responsive at the time of treatment. Radicant Lysimachia nemorum and non-radicant Saxifraga hypnoides behave similarly (47).

Chrysanthemum hortorum or morifolium, the garden chrysanthemums (33, 49, 203) flower in short days more or less obligately (late varieties) or only prefer short days (early varieties). Some of them (22) flower irrespective of temperature; others are stimulated by warm temperatures (> 12°C.) and others by chilling (< 12°C.). Although short days induce flowering, chrysanthemums also require various degrees of preparation. I have already shown (33) that certain varieties respond to short days much earlier and more effectively when they have previously received a long-day treatment. Schwabe (230 to 232) showed that certain varieties require a preparatory chilling period of various extents which can be considered as vernalization. The treatment is more effective when given at night and for about three weeks around 4 to 6°C. These phenomena have also been carefully studied by Vince (262, 263) who showed in particular that only certain varieties may undergo the preparatory phase partly with cold temperature treatment and partly with long-day treatment. Many cold-requiring varieties grow in warm temperatures as creeping shoots, forming short sprouts with layered leaves, as in rosettes.

Removing the rosette terminal bud in certain of these releases axillary shoots that are then directly influenced by short days (233, 234, 263). Most often, axillary buds are vernalized simultaneously with the terminal bud, but not always down to the base of the stem. Moreover, exposure to high or low light intensities, or growth in the ground (suckers) causes devernalization, which ensures the perennial state. This is most effective at temperatures less than 18°C., even if the plants receive additional soluble sugars. However, heat does not devernalize in high light intensities. The vernalized state is not graft transmitted. Gibberellin often suppresses the branching stage, but the long shoots still develop flowers in a few varieties, even under short days. Picking the flower buds formed under short days sometimes induces flowers directly on young suckers, but this is not true for most varieties. The varietal responses are so varied that the interesting investigations recently completed are now giving rise to further studies [cf. recent works, Cathey (22), Chan (29), Doorenbos (72), Schwabe (230 to 236), Vince (262, 263)].