8.3.1 A time to flower
Survival of many plant species depends on setting seed well in advance of seasonal environmental extremes including frost, heat or drought and particularly during pollen formation and pollination. Synchrony of flowering is also beneﬁcial especially for outbreeding species which must time their reproduction to coincide with flowering of other individuals or genotypes and often with the presence of insect and bird pollinators. The natural light and temperature environment provide much of the seasonal information essential for control of flowering time, but plant age or maturity can also be important.
(a) Plant maturity and flowering time
Many plants grow vegetatively for periods ranging from weeks to years and then flower autonomously, apparently without identiﬁable environmental control. Flowering of 25–30 year-old bamboo is one such example: no environmental cue is known for this species. Perhaps it has its own built-in developmental clock which determines flowering time as in some annuals which flower autonomously. In contrast, other species may flower late due instead to inappropriate cultural or environmental treatments. In this instance, flowering may not occur irrespective of whether the juvenile phase has ended.
In some species, flowering occurs after the apex has produced a particular number of leaves. This apparent leaf counting may reflect an interplay between older leaves and the roots. In tobacco, for instance, proximity of the roots to the main shoot apex is critical. Plants remain vegetative until the shoot apex is more than ﬁve to seven leaves above the roots or above a zone of experimentally induced root formation on the stem (McDaniel 1980).
Extremely fast flowering without any apparent juvenility is seen in some desert annual plants. They may germinate and reproduce rapidly after rainfall, forming as few as two or three leaves and then flowering. The terminal shoot apex and all axillary apices may become floral. More often, however, such rapid flowering is restricted to either lateral or terminal meristem(s), leaving a second population of meristems avail-able for further growth and reproduction if favourable conditions persist (Hayashi et al. 1994).
With some agricultural crops bred for earliness of flowering, such as soybean and rice, early maturity may have resulted from a shortening of the juvenile phase (Evans 1993) rather than from changes in sensitivity to environmental cues. Thus, for some crop plants, duration of juvenility can in-fluence chronological and developmental time from seed germination to flowering, regardless of other physiological controls of flowering.
As an adaptation for survival, juvenility is an advantage and a single gene controlling its duration is known in Pisum (Murfet 1985). Embryonic flowering (Emf) may peform a similar role in Arabidopsis. As discussed later, several other floral-speciﬁc genes also influence aspects of this floral transition. In contrast to the abbreviated juvenile phase of annuals, perennials such as apple or mango have a juvenile phase often lasting ﬁve to eight years. Various cultural and environmental manipulations including drought, nitrogen fertilisation, stem girdling, grafting and CO2 enrichment can reduce this period in conifers (see Pharis and King 1985). The juvenile period of some Eucalyptus species can also be shortened from two to three years to 9–12 months if grafted cuttings are exposed to cool inductive conditions and treated with an inhibitor of gibberellin biosynthesis. Endogenous gibberellin A1 (GA1) levels were lowered by this treatment (Moncur and Hasan 1994) so high gibberellin levels may be one component of prolonged juvenility in Eucalyptus. We will see later that in other species gibberellins may promote flowering, so we need to make clear distinctions between species, process (breaking juvenility or inducing flowering) and even the type of gibberellin (see Pharis and King 1985).
(b) Flowering time and environment: photothermal input
Environmental factors that limit plant growth may also pro-foundly influence flowering time. Suboptimal growth conditions may delay flowering and give an apparent ex-tended juvenile phase, and often light intensity, light duration and temperature are major limitations. Thus, a summation of both inputs (the photothermal sum) over all or part of the calendar year helps to characterise the growing season. Photo-thermal sums indicate whether there is adequate time from sowing to seed maturation for an annual crop or wild plant species. The yearly cycle of solar radiation highlights how this varies with latitude (Figure 8.19). There are losses due to cloud and to atmospheric interception. Of the remaining sun-light, the visible/photosynthetic component is about 45% and the rest is ‘heat’. The calculation of photothermal units integrates these heat and visible light inputs. For example, although daily photosynthetic flux at extreme latitudes may be high in summer, the growing season is extremely short.
Thermal sums (based on a heat sum above a 10°C base) have been used in the USA to predict the likely penalty in flowering time, and hence in yield, from growing long-season (late flowering) corn varieties at a higher latitude (Figure 8.20). To maintain yield, breeders have had to obtain lines with shorter growing seasons, in this case selecting varieties with more rapid early seedling growth and therefore re-quiring smaller thermal sums. Similar approaches with other crops such as soybean have used data from analysis of ﬁeld environments and controlled-environment studies (see Evans 1993).
Photothermal responses for perennial crops are more complex, partly because flowering may relate to current and previous years’ environmental conditions. Controlled-environment experiments help us unravel some of the interactions. In vines such as grape and kiwifruit, the extent of bud dormancy can be determined on cuttings taken from ‘winter’ canes and transferred to controlled-environment cabinets. This enables prediction of timing of ﬁeld budburst for each cultivar (see Section 8.1.3).
Another approach with perennial plants involves collection of ﬁeld flowering and temperature data over a number of years at different latitudes. For two ericaceous shrubs a heat sum model predicted flowering times at eight ﬁeld sites in Canada (Reader 1983) and similar heat sum relationships have been shown for another 15 species at 200 latitudinal sites in Alberta. The earliest spring flowering species had the smallest heat sum for flower opening.
Information on climate and plant responses to the environment provides one way to estimate global re-productive potential. In equatorial zones, temperature and irradiance change less over the year (Figure 8.19) and time of flowering may instead reflect seasonal rainfall patterns. In warmer temperate zones, early spring flowering and adaptation to intermediate heat sums can ensure reproduction prior to high summer temperatures and drought stress, but a second favourable climatic window is autumn. At high latitudes or at altitude, growth and flowering occur during midsummer.
Although these ideas can explain seasonality of flowering, photothermal relationships match best to the period of development up to flower opening (Reader 1983). They apply less well to floral induction, which is often a response to speciﬁc episodes of high or low temperature and/or to seasonal change in daylength. Assessment of such responses is best studied in controlled-environment chambers where each component can be varied independently. In this way we can reveal effects on flowering of seasonal changes in amount and duration of daylight, the ‘photo’ component of photothermal responses. As shown in Figure 8.21, flowering response of the grass Lolium temulentum varies with irradiance at the time of exposure to a single inductive long day. Increase in photo-synthetic input is beneﬁcial but is not the major limiting factor for flowering. Rather, daylength (photoperiod duration) is the major determinant of flowering in this and many other species.
(c) Daylength and flowering time
As long ago as 1914, scientists recognised that daylength regulated flowering time of hops (Humulus japonicus) and by 1920 two Americans, Garner and Allard, had demonstrated daylength control of flowering of many species. They termed the species either short- or long-day plants (SDPs or LDPs). SDPs flower in response to a decrease in daylength, that is, an increasing length of the daily dark period and a shortening photoperiod; LDPs flower in response to increasing photo-period. As well as causing flowering, daylength can also influence winter dormancy of buds, tuberisation, leaf growth, germination, anthocyanin pigmentation and sex expression.
Change in daylength is identical from year to year (Figure 8.22) and so provides precise information on season. Thus a photoperiodic plant can time reproduction to avoid mid-summer drought, autumn cold or late spring frosts. Summer flowering at higher latitudes typically will involve a response to long days. In the tropics, daylength changes little, so selection pressure could be for daylength insensitivity or short-day response, provided plants could measure such small changes in daylength. Withrow (1959) calculated that to measure seasonal time to within one week required a 1–3% precision in measurement of daylength. Only a 4–12% precision was required for accuracy to within a month. In the tropics, a 1–3% accuracy would mean distinguishing photo-periods differing by 7–21min around a 12h daylength. Remarkably, several species including some tropical plants do show such accuracy. In studies with rice, a tropical SDP, flowering occurred 30 to 50 d later when the photoperiod was increased by only 10min, from 11h 50min to 12h (Dore 1959).
Detection of daylength involves a photoreceptor called phytochrome. This pigment detects very low energies of visible light, especially red and far-red wavelengths. The con-sequence is that major daily and seasonal fluctuations in photosynthetic light intensity do not influence measurement of daylength. So sensitive is phytochrome that at latitudes up to 40° plants respond to twilight radiation for about 20min after sunset and before sunrise (Salisbury and Ross 1983). At high latitudes, the midsummer sun may never set as far as phytochrome sensing is concerned. We return to discussions of phytochrome in Section 8.4.
The duration of daily light/darkness which is effective for flowering may be very precise or very broad. Such contrasting patterns are illustrated in Figure 8.23 along with typical long-day, short-day, intermediate, ambiphotoperiodic or day-neutral (indifferent) responses. Daylength-indifferent types represent less than 15% of the 150 or so grass species reviewed by Evans (1964), although this proportion may be an underestimate as ‘observed’ day-neutral responses might not always be reported.
Within a species there can be large differences in photo-period response, as in the LDP Phleum pratense (Figure 8.23). The full range of daylength response types may even be found within a single species. For example, in a controlled environ-ment study of 30 ecological races of the Australian grass Themeda australis, Evans and Knox (1969) found that low-latitude strains, from 6° to 15°S, behaved as SDPs (Figure 8.24). Races from more southerly origins to 43° were LDPs with some responsive to vernalisation (see later). This ecotypic variability exempliﬁes heritability and adaptability of environ-mentally responsive flowering and appears to have aided reproductive success of Themeda. If the species migrated to Australia via Asia and New Guinea, it would probaly have adapted from a short-day response to day neutrality or sensitivity to long day and to vernalisation.
Some plants will flower after just one cycle of the appropriate daylength. Cocklebur (Xanthium strumarium) and Japanese morning glory (Pharbitis nil) are classic examples of SDPs responding to one short day, or more correctly, one long night. Similar single-cycle responses are found for LDPs such as Lolium temulentum (Figure 8.21). Other species require several days (e.g. soybean, strawberry) or weeks of exposure to the appropriate daylength (e.g. Geraldton wax, chrysanthemum). In some plants a sequence of short days must precede long days (SLDP) as for some clovers (e.g. Trifolium repens) and grasses (e.g. Poa pratensis). Conversely, some species respond as long–short-day plants (LSDP) in-cluding Aloe, Bryophyllum and some mosses and liverworts (see summaries in Lang 1965; Vince-Prue 1975). Some dual photoperiodic responses may be satisﬁed simultaneously so that flowering is best at intermediate daylengths (e.g. some sugar cane genotypes). The converse is also known, ambi-photoperiodic response, with best flowering at either short or long days but not at intermediate daylengths (Figure 8.23). Separation in time occurs in some grasses which respond to short days for primary induction leading to a microscopically visible inflorescence but later to long days for subsequent development to anthesis (Heide 1994).
(d) Low temperature and flowering time
Although growth is limited by low temperature, scientists in the mid-nineteenth century recognised that floral initiation of many species requires exposure to cold. For a temperate cereal such as wheat, low-temperature exposure of imbibed grain caused winter lines to flower like their spring wheat counter-parts. We term this response vernalisation, meaning ‘to become spring-like’.
Vernalisation-responsive species include winter annuals, biennials and perennials. Many are also LDPs including some grasses and species with a rosette growth habit. Effective temperatures for vernalisation range between -6°C and 14°C, with most temperate species responding best between 0°C and 7°C. In all cases, these temperatures are below those optimal for growth. Floral primordia are sometimes initiated during the cold period, as in brussels sprout, turnip, stock and bulbous iris. Alternatively, cold treatment is a preparatory phase enabling later initiation of flowers.
Generally, prolonged exposures of one to three months are required for vernalisation but this varies with temperature and species. However, as with photoperiodic species, some respond to a single cold day, for example chervil. In Geum, the vernalisation period depends on meristem location, ranging from two to three months in axillary meristems to one year for the terminal apex. Heterogeneity of floral response of meristems has clear adaptive beneﬁts, whether for perennation as with Geum or for opportunistic responses to rainfall as for desert ephemerals (see above).
As with photoperiodism, dependence of flowering on vernalisation changes with latitude. For example, a vernalis-ation response appears only in high-latitude ecotypes of Themeda australis (Figure 8.24) and is likewise more important for species and ecotypes from higher altitudes. European thistle (Cirsium vulgare) collected from the Mediterranean to Scandinavia exhibit vernalisation requirements predominantly in lines from colder, more northerly sites (Weselingh et al. 1994). In addition to latitude effects in the grass Phalaris aquatica, there is a superimposed altitudinal cline.
Leaves sense photoperiod, but perception of low tempera-tures resulting in vernalisation responses can be by the shoot apex instead. Chilling of leaves is usually ineffective (Bernier et al. 1981). However, cold-treated leaf cuttings of species such as Lunaria and Thlaspi arvense, and even chicory root explants, regenerate plants which flower without further vernalisation (Metzger 1988). One hypothesis is that vernalisation responses may be initiated only at sites with potential for cell division, that is, meristems or regenerating tissues. On the other hand, in pea and sweet pea, there is clear evidence of transmission of vernalisation signals across graft unions (Table 8.8). In these experiments, perception of cold must have occurred in cells other than those in the responding shoot apex. These species also exhibit normal shoot apex vernalisation responses, so there can be two different mechanisms of low-temperature sensing.
The presence of water and metabolic activity are essential requirements for vernalisation. We deduce this from vernalis-able species which can respond during seed germination. Radish seed, for example, cannot be vernalised when dry or in a nitrogen atmosphere.
The vernalised state is quite stable in seeds of some species: they can be dried after cold treatment, even stored for long periods, and then sown without loss of response. However, particularly with marginal vernalisation, temperatures immediately following often need to remain below 25°C to prevent devernalisation. High temperature up to 40°C for a few days sometimes annuls a preceding cold exposure (Bernier et al. 1981). Indeed, devernalisation every summer may reset the flowering of perennial plants so that they require renewed vernalisation each winter.
Photoperiod requirements post-vernalisation are diverse. Many winter annuals or biennials require long days following vernalisation. For example, vernalised Hyoscyamus will not flower under short days but under long days promptly forms flowers, even with 300 short days between vernalisation and induction. In contrast, sensitivity of spinach to inductive long days is altered following cold treatments with a shortening of the critical day length from 14h to 8h. A few cold-responsive plants, such as chrysanthemum, require short days after vernalisation.
The genetics of vernalisation range from simple to very complex depending on the species. For example, a single locus distinguishes the biennial, cold-requiring strain of Hyoscyamus from its annual counterpart. By comparison, vernalisation of hexaploid wheat involves at least three loci (Vrn 1, 3 and 4), probably reflecting its genetic complexity.
Pea and Arabidopsis normally respond both to photoperiod and to vernalisation. Of the many late-flowering mutants known, some are vernalisation responsive, including gigas (gi) in pea and luminidependens (ld) in Arabidopsis. There are also vernalisation-unresponsive and early-flowering mutants. One simple explanation is that the wild-type products of some of these genes are inhibitors of floral induction or initiation or, conversely, stabilise vegetative growth.
Vernalisation may involve decreased DNA methylation allowing activation of suites of genes including some involved in synthesis of gibberellins. For example, extending the earlier work of Hirono and Redei (1966), Burn et al. (1993) found that vernalisation-responsive late-flowering mutants of Arabidopsis treated with the demethylating agent 5-azacytidine flower earlier than unvernalised controls. From this result, they concluded that demethylation occurs during vernalisation and leads to selective derepression of genes required for flowering.
Cool temperature response
In addition to classic vernalisation responses, there are many reports of species, especially from warm climates where near-freezing temperatures are infrequent, which flower if exposed to temperatures from 10°C to 20°C. For some tropical fruit crops (e.g. mango, avocado, lychee, longan), especially those grown in the subtropics (latitude 23°–30°) where substantial seasonal temperature changes occur, floral induction results from exposure to night temperatures of 10–15°C. Because tropical species are relatively under-researched compared with their temperate counterparts, physiologists have yet to decide whether these cool responses have similar mechanisms to temperate vernalisation but are adapted to a different temperature range. Another possibility is that flower initiation and development are blocked/reversed by higher tempera-tures, so low temperature could merely be a passive condition permitting expression of an innate capacity to flower. This may be the case for Acacia and rice flower (see King et al. 1992) but for Pimelea ferruginea, which flowers if exposed to temperatures below a daily average of 16–18°C for ﬁve to seven weeks, the response is inductive and higher temperature does not cause loss of developing flowers (King et al. 1992).
(e) Water stress and nutrition
In some species including Lolium, Pharbitis and Xanthium, floral induction and development are blocked by water stress (see Bernier et al. 1981). For Lolium, an 8h stress inhibited flowering only if given at the time of the long day, not one day before or after. Shoot apex abscisic acid (ABA; see Chapter 9) content increased transiently up to 10-fold in association with the brief water stress (King and Evans 1977). Furthermore, ABA inhibited flowering if applied at the time of the long day. Later in flower development, water stress or ABA application can result in sterility in wheat. The problem is morphologically aberrant pollen, but seeds are still set if plants are hand pollinated (Morgan 1980b).
By contrast, positive responses of flowering to water stress are also known. For the geophyte Geophila renaris, growth under water-limited conditions for two months causes flowering (see Bernier et al. 1981). Similarly, water stress coupled with enhanced photosynthetic conditions, high tem-perature and gibberellin application can cause precocious flowering in some conifers (Pharis and King 1985). In mango trees grown in the tropics with little temperature variation, seasonal flowering appears to be promoted by water stress during the dry season. This may relate to trees having an extended period of suspended growth during which ability to flower gradually develops, for example as a result of accumulation of stored carbohydrate.
Nutritional status of plants has little direct influence on floral initiation, although in many species there are effects on flower number and on fruit and seed development. For example, pollen fertility in wheat is reduced by excesses and deﬁciencies of trace elements including copper and boron (reviewed by Graham and Nambiar 1981). In strawberry, plant size and fruit and flower number increase as nitrogen supply is increased (Guttridge 1969), but the supply of nitrogen during early stages of flower initiation may enhance vegetative growth not flowering. Such complex responses make it dif-ﬁcult to argue that transition to flowering requires low-nitrogen status coupled with enhanced carbon supply. Numerous studies have failed to demonstrate an inverse relationship between nitrogen supply and flowering and, as noted above, there are often positive effects on floral development (see Bernier et al. 1981). Perhaps a unique response to nitrogen is the dramatic increase in flowering of apple supplied with nitrogen but only if supplied as ammonia (Grasmanis and Leeper 1967). Overall, mineral nutrients, while essential for growth, may not speciﬁcally regulate flowering.
(f) Environmental and seasonal synchronisation of flowering
Species in their natural environments
Control of seasonal flowering time may be as simple as the acquisition of a long-day or short-day photoperiodic response, or of both as in LSDP where exposure ﬁrst to long summer days is essential to guarantee flowering in the short days of autumn. Alternatively, floral development may occur in spring when both temperature and irradiance increase rapidly to permissive levels (Figure 8.19). A vernalisation requirement allows for spring flowering, or for summer flowering when combined with a long-day response.
Often, a combination of short day then long day, as well as temperature, is important in synchronisation of flowering of perennial grasses (Heide 1994). Comparison of environmental tolerances of Bromus inermis, a species adapted to lower latitudes, and Poa pratensis, an arctic–alpine species, highlights how these inputs determine survival. For flowering, both species require short-day or low-temperature exposure followed by long days. The short-day response is strict in Bromus and, because of intolerance to low temperatures, it will never flower at the high latitude of Tromsö (69°39¢N), as shown by its climate phototherm (Figure 8.25). The response of Poa, by contrast, overlaps an arctic phototherm (Tromsö) but this species is intolerant of the higher summer tem-peratures at lower latitudes. Dual induction responses also enable high-latitude-adapted species to initiate inflorescence primordia in autumn short days. The outcome is to maximise the number of summer days available for seed development because anthesis proceeds rapidly in the following summer long days, even in the short, cool arctic growing season.
Field to nursery transplantations have often demonstrated environmental influences on flowering, as noted above for vernalisation of Cirsium arvense. Alternatively, controlled en-vironment studies of the type used by Evans and Knox have revealed ecotypic adaptation of flowering in Themeda (Figure 8.24). Rarely have the two approaches been combined. Either photothermal models have been used to assess ﬁeld flowering data or laboratory environmental response proﬁles have been incorporated into empirical models predicting ﬁeld response. However, with Pimelea ferruginea grown simultaneously in controlled environments and in the ﬁeld over winter (King et al. 1996), there was a close match between effective tempera-tures for flowering in the ﬁeld and laboratory. In addition, evidence for adaptation to small (4°C) temperature dif-ferences came from a high-latitude ecotype from 31°S which was unable to flower when transplanted to the warmer extreme of the species distribution (28°S).
Predicting flowering time of ﬁeld crops
Phototherms only broadly deﬁne the tolerance of a species to its environment. A more deﬁnitive approach uses rates of response of flowering to photoperiod and temperature based on constants derived from controlled environments. Threshold limits are also imposed to constrain models to response envelopes of the sort illustrated in Figure 8.25. In a broader study (Lawn et al. 1995), six crop species (soybean, cowpea, mungbean, chickpea, barley and lentil) sown at different latitudes and times flowered in the ﬁeld at times which correlate well with those predicted from a simple linear additive model (Lawn et al. 1995). However, such models make no allowance for effects of light intensity and extreme conditions outside the threshold limits which can be important for flowering, for example vernalisation or warm temperatures.
Commercial nursery floriculture
Prior information on environmental response has been crucial to nursery production of potted flowering plants including the SDPs chrysanthemum and poinsettia. However, there may be inevitable compromises in some of the complex protocols required for commercial production of an Australian SDP, Geraldton wax. Its critical photoperiod is about 13h, so the maximum tolerable daylength would be about 12h from sunrise to sunset plus 20min each pre-dawn and twilight (Dawson and King 1993). Thus, in summer, glasshouse black-out curtains are used to maintain the inductive short day, but this is obviously not an option for ﬁeld-grown plants. Glass-house summer temperatures exceeding 35–40°C, well above the optimum for the species, are another problem. As a comparison, optimal mean daily temperature for chrysanthemum is about 21°C (Pearson et al. 1993). Con-sequently, greenhouses are often shaded to avoid costly cooling, but then lower photosynthetic input may result in poorer flowering.
Flowering of woody horticultural species
Prolonged juvenility of woody species is a problem for growers and breeders of tree and vine crops. However, there are so many uncontrolled variables in the ﬁeld that it can be difﬁcult to identify the inductive factors. Yields can be severely depressed by inappropriate timing of practices such as pruning, irrigation and fertilisation. Furthermore, inductive conditions may be required for several months. One solution for mango, lychee, olive and citrus has involved the use of controlled environments and ‘mini’ plants grown from cuttings. These showed that cool temperatures were required for induction, a response similar to Pimelea and many other ornamental and woody species.
For some species, microscopic examination of shoot meristems has augmented our ability to make decisions on practical management of flowering. For example, in kiwifruit (Actinidia) and stone fruits (Prunus spp.) floral induction occurs in the previous growing season, whereas in many subtropical species no initiation takes place until winter. In the case of kiwifruit, it was discovered that late summer pruning was removing many of the floral apices (Snowball 1995).
Clearly, knowledge of environmental effects on flowering has been essential for development of nursery, orchard and agricultural crops. Particularly for ﬁeld crops, breeders have selected for day-neutral responses. For glasshouse crops, genotype and environment have often been altered. The future offers many opportunities for applying our knowledge of daylength and photothermal responses.
Plants depend on natural daylength changes (e.g. short day, long day, short day→long day, long day→short day and/or low temperatures to regulate timing of reproduction. Progressively shorter days in autumn, for example, are likely to cause flowering in LSDPs. A requirement for low temperature (vernalisation) can ensure bienniality in spring-germinating species. Many warm-adapted species appear to depend on cool rather than cold temperature for spring flowering.