The Phenology of Growth and Reproduction in Plants

Vol. 1/1, pp. 78–91 Perspectives © Gustav Fischer Verlag, 1998 in Plant Ecology, Evolution and Systematics

The phenology of growth and reproduction in plants

Michael Fenner

Biodiversity & Ecology Division, School of Biological Sciences, University of Southampton, Bassett Cres- cent East, Hampshire SO16 7PX, UK; email: [email protected]

Abstract

The study of phenological aspects of plants involves the observation, recording and interpretation of the timing of their life history events. This review considers the phenology of leafing, flowering and fruit production in a range of species and communities. The selective forces (both abiotic and biotic) that influence the timing of these events are discussed. Within the limits imposed by phylogenetic constraints, the phenological patterns (timing, frequency, duration, degree of synchrony, etc.) of each phase are probably the result of a compromise between a variety of selective pressures, such as seasonal climatic changes, resource availability, and the presence of pollinators, predators and seed dispersers. Many studies on flowering times stress the role of interactions between plant species which share pollinators or predators. The timing of fruiting plays a key role in controlling the abundance and variety of obligate frugivores in many tropical communities. The importance of long-term recording is stressed, particularly in species which fruit irregularly. An understanding of the phenology of plants is crucial to the understanding of community function and diversity.

Key words: Leafing, flowering, fruiting, pollination, seed predation, synchrony

Lo! Sweeten’d with the summer light, tory events occurring in response to water The full-juiced apple, waxing over-mellow, availability. Drops in a silent autumn night. Phenology (from the Greek phainein, to All its allotted length of days, show or appear) is the study of the timing of The flower ripens in its place, these life-history events. In plants, bud-burst, Ripens and fades, and falls, and hath no toil, leaf-expansion, abscission, flowering, fertili- Fast-rooted in the fruitful soil. sation, seedset, fruiting, seed dispersal and – Tennyson, The Lotus Eaters germination all take place in due season. For the most part, these events are too familiar to attract any special attention. Only when the Introduction expected pattern is broken, for example by out-of-season flowering or the loss of fruit One of the most familiar of all natural phe- due to a late frost, is our attention drawn to nomena is the cycle of events associated the importance of timing of growth and repro- with the passage of the seasons. This is duction in the life of plants. especially so in temperate and polar regions Phenological studies are as important to where annual changes in temperature are our understanding of species interactions and great and are accompanied by corresponding community function as the spatial aspects are. cycles in the growth and reproduction of the In Europe there is a long tradition of phenolog- flora. In tropical regions, seasons are often ical recording, which goes back to the last marked by differences in rainfall, with life-his- century, with an international network of phe- Phenology of growth and reproduction in plants 79 nological gardens. The information gathered in nature, it may not be easy to identify the rel- has mainly been used as a kind of bioindicator evant ones. The long-term cause of the timing of climatic patterns (Schnelle 1965–1986, of a life-history event is the selection pressure 1977). However, data about recurring life-his- which has resulted in the evolution of the tory events are still relatively scarce for most plant’s idiosyncratic phenology. These pres- communities. Some of the reasons for this are sures may range from abiotic constraints (e.g., pointed out by Lechowicz & Koike (1995). seasonally unfavourable temperatures, erratic Firstly, the data are hard won. They can only rainfall), to biotic pressures such as seasonal be accumulated by many years of repetitive presence of predators, pollinators and dis- recording at the same site. Phenological stud- persers. Many phenological studies are basi- ies need long-term planning and funding, and cally concerned with answering the question so do not lend themselves to the short-term ‘Why now, rather than earlier or later?’ The requirements of thesis projects, grant cycles, timing of a particular life-history event may be and the priorities of awarding bodies. Never- the result of selection for exploiting a favo- theless some excellent long-term studies have urable slot sandwiched between two more been published (e.g. Norton & Kelly 1988; hazardous periods. Allen & Platt 1990) and are a tribute to the per- Most life-history events occur over a sistence of their authors. period of time, and their timing can be de- The phenology of plant communities can scribed and quantified in terms of a number be studied by dealing with particular life-his- of parameters (Newstrom et al. 1994). These tory stages separately, such as leafing, flow- are (1) the frequency of occurrence, which is ering, fruiting, seed dispersal and germina- often annual; (2) the time of occurrence, tion. Each of these events occurs in its own which includes the starting date of the earliest calendar slot, but there is clearly some inter- individuals and the date of peak activity; (3) dependence between them. Fruiting must the duration of the event; (4) the magnitude wait upon flowering; seed dispersal cannot (both the mean and variability); and (5) the preceed fruiting. Each phenomenon can be degree of synchrony both within and between studied at different levels of organisation. For species. In general, where the climate is example, we can study the phenology of flow- highly seasonal, as in temperate regions and ering in a whole plant community, or in a guild in the dry tropics, the phenology of whole of species sharing a pollinator, or in a popula- communities tends to follow a broadly similar tion of a particular species, or in the flowers pattern from year to year, though modifica- on a single plant. Even an individual flower tions occur due to climatic differences be- undergoes a sequence of events (bud burst, tween years. Leaf flushing tends to be rela- stigma ripening, pollen shedding, fertilization, tively constant from year to year, but repro- fruit swelling, etc.) whose phenology differs duction can be erratic, especially in harsh cli- between species and is presumed to be the mates (Allen & Platt 1990). In the wet tropics result of natural selection. At each level, the where annual seasons are less marked, a constraints and selective forces which influ- great variety of phenological patterns are ence timing are different. This approach to seen: in different species, flowering may be examining phenological information at suc- continual, sub-annual, annual or supra-an- cessively finer levels of organisation is well nual, and may occur regularly or irregularly exemplified in the review of flowering pat- (Newstrom et al. 1994). The great variety of terns by Primack (1985). patterns encountered in leafing, flowering For individual species it is important to dis- and fruiting in any given habitat may be a re- tinguish between the physiological and eco- flection of diverse biotic and abiotic selective logical/evolutionary answers to the question of pressures operating in all communities. why leafing, flowering or fruiting occurs at a This review examines some of the inter- particular time. The immediate cause of a phe- esting questions which arise in relation to the nological event is the organism’s detection of phenology of plants, and attempts to summa- the environmental cues which trigger the ap- rize some of the answers which have been propriate response, e.g. a flowering response put forward. The main questions dealt with to short or long days (Bernier et al. 1981). here are: Physiological responses are in principle easy 1. How and why do plants differ from each to explore experimentally, though with so other in the timing of their life history many potential cues operating simultaneously events? 80 M. Fenner

2. Why is the same stage apparently ‘stag- ranged from synchronous to continuous in gered’ between some sympatric species? different species and that there were large 3. Why do some stages occur on an annual differences in longevity. In seasonal climates, cycle while others do not? the production of new leaves tends to be con- 4. What are the constraints on the timing or centrated for most species into a short period frequency of a particular phase? of the year. This is the case both in evergreen 5. How does the phenology of the plants af- communities (such as the boreal conifer for- fect the primary consumers? est) and in deciduous forests (both temperate 6. How might global climatic change affect and tropical). Usually leafing is linked with the phenology of plants and hence the some climatic feature such as rainfall (Lieber- functioning of whole communities? man & Lieberman 1984; Bullock & Solís-Ma- gallanes 1990), irradiance (Wright & Van An excellent introduction to the broad topic Schaik 1994), temperature (Brooke et al. of phenological patterns in plants is provided 1996), or photoperiod (Loubry 1994). Where by Rathcke & Lacey (1985). For the flowering water is not a limiting factor, irradiance may phenology of tropical forests, Newstrom et al. play a predominant role. In a major study of (1994) have successfully reduced the great eight disparate tropical forests, Wright & Van variety of patterns found to a small number of Schaik (1994) show that leaf (and flower) pro- identifiable types. Also in tropical forests, Van duction coincides with seasonal peaks of irra- Schaik et al. (1993) have synthesized a diance. In three cases peak irradiance oc- range of data on phenological adaptation to curred in the drier season. Leaf production abiotic conditions, and discuss the commu- can be artificially induced in some cases by nity effects of phenological events. This re- irrigation during the dry season (Wright 1991; view deals with the three major phenophases Tissue & Wright 1995). However, in many in the life-cycle of plants (leafing, flowering tropical deciduous forests, flushing takes and fruiting) and attempts to identify the se- place during the dry season, well before the lective forces which may be responsible for arrival of the rains. For example in the sea- the patterns encountered. However, the pos- sonally dry forests of southern India leaf sible influence of selective forces in determin- flushing peaks in the hottest, driest month of ing phenological patterns must be interpreted the year (May), greening the canopy well in in the light of the phylogenetic inheritance of advance of the onset of the monsoon (Singh the species. As with all other genetically con- & Singh 1992; Murali & Sukumar 1993). Dry trolled characteristics, the timing of life history season leafing is also reported by Frankie et events may be largely influenced by evolu- al. (1974) for an evergreen forest in the low- tionary constraints, with selection merely lands of Costa Rica. The production of leaves modifying a predetermined underlying pat- before the arrival of the rain is achieved by tern. The issue of phylogenetic inertia is dis- drawing on internal reserves of water which cussed in the context of fruit characters in buffer the trees against the impact of sea- Lee et al. (1991). sonal drought (Borchert 1994). Al-Mufti et al. (1977) provide one of the few detailed studies of shoot phenology in Temporal patterns in leafing herbaceous communities. They show that in Community patterns productive, less disturbed, temperate communities, the biomass peaks in mid-summer. The timing and duration of bud burst, leaf ex- In low-fertility sites, plants tend to be ever- pansion, maturity, senescence and fall of in- green, with some species showing no de- dividual leaves are all crucial to the fitness of tectable seasonal peak in shoot expansion. plants. Rather few long-term studies have The phenology of growth in many temperate been made of the life histories of leaves. In bulbous plants (and grasses with large cells aseasonal climates, such as parts of the wet and high DNA content) appears to be con- tropics, leafing is continuous on a community trolled by their ability to grow at low tempera- scale, though a wide range of behaviour is tures in early spring by expanding cells which found between species. Lowman (1992) have been pre-formed in the preceding sum- monitored individual leaves of five species of mer. This contrasts with other herbaceous rain forest trees in Australia for up to twelve plants which have small cells and genomes, years. He found that flushing behaviour and delayed (summer) growth because they Phenology of growth and reproduction in plants 81 need higher temperatures for cell division the individuals which conform to the least (Grime & Mowforth 1982). The link between dangerous calendar slot. nuclear DNA and rate of shoot expansion in A second possible strategy for reducing spring has been shown in a limestone grass- herbivory is the synchronous production of land community by Grime et al. (1985). leaves throughout a population, thereby sati- ating the herbivores. This was investigated in Herbivory as a selective force in the Central American sub-canopy tree Gusta- leaf flushing via superba (Lecythidaceae) by Aide (1991). Most individuals produce their leaves A major selective force in determining the tim- synchronously with the rest of the population. ing of leafing may be the effect of herbivory. Some individuals, however, are found to be Young leaves are especially vulnerable to out of synchrony with their conspecifics, and attack by insects (Lieberman & Lieberman suffered significantly more herbivore damage. 1984; Lowman 1992), with the majority of life- Another phenological feature seen in time damage being suffered during the first many tropical evergreen trees is the delay in month (Aide 1993). Highest damage may greening of the leaves after expansion. A occur on early-expanding rather than on juve- study by Kursar & Coley (1992) using data nile leaves, at least in European shrubs (R.V. from 175 species on Barro Colorado Island, Jackson, J. Kollmann & P.J. Grubb, unpubl. Panama, indicates that delayed greening is data). A number of life history traits seem to most common in shade species and rare in have resulted from selection due to herbivory. light-demanding gap species. This was inter- One is the production of new leaves at a preted as a possible anti-herbivore defence period when the herbivores (mainly insects) mechanism. By delaying the input of chloro- are at their least abundant. This is the sup- phyll, nitrogen and other valuable resources posed advantage of dry season flushing. Ob- until after the leaves have acquired their servations by Murali & Sukumar (1993) show structural defences, the plants limit their that late flushing individual trees in the Indian losses to herbivory. The benefit presumably monsoon forest suffered significantly higher outweighs the disadvantage of the temporary damage from insects compared with those absence of photosynthetic activity, and the that flushed earlier because the late-flushing balance of advantage is reckoned to be espe- trees were damaged by the herbivorous in- cially favourable in shade leaves. sects which emerged with the rains. Therefore, the timing of leafing seems to The temporal escape from herbivory by be a compromise between exploiting the means of flushing in the dry season was in- most favourable environmental conditions vestigated experimentally by Aide (1992) and the avoidance of herbivory. In many using populations of the understorey shrub cases, herbivory may simply be avoided by Hybanthus prunifolius (Violaceae) in Pana- the satiation effect of synchronous flushing. ma. Individuals of this species respond to dry season rain by producing leaves. Some indi- Leaf fall viduals were experimentally prevented from flushing until the wet season. A comparison of For deciduous species, the seasonal loss of cohorts of dry-season and wet-season leaves leaves brings about a period of dormancy showed that wet season leaves suffered during which photosynthesis is generally sus- three times as much herbivore damage. pended. The cost of shedding the leaves is In temperate communities, individuals usually mitigated by the reabsorption of min- which flush either earlier or later than the erals preceding abscission (Delarco et al. mean of the population can find themselves 1991; Chidumayo 1994). The timing of leaf at risk of herbivory. For example, in Picea fall is often linked with some change in envi- glauca damage by the bud moth Zeiraphera ronmental conditions such as water availabil- canadensis has been shown to be greatest ity (Williams et al. 1997), temperature (Rusch amongst early leafing individuals (Quiring 1993) or photoperiod (Loubry 1994). For ex- 1994). In contrast, late flushing individuals of ample, in an Australian tropical savanna, Quercus geminata were found to be most Williams et al. (1997) recorded a wide degree susceptible to a leaf miner moth, Stilbosis of deciduousness, but in each species leaf quadricustatella (Mopper & Simberloff 1995). fall was coincident with the attainment of sea- Clearly, selection will favour the survival of sonal minima in leaf water potential. In con- 82 M. Fenner trast, Loubry (1994) found that individual tion of a temperate rain forest in Chile that trees in a tropical rain forest had their own flowering (and fruiting) was largely con- periodicity with respect to leaf fall, and that it strained by seasonal variables. was not linked with a dry season. Leaf fall in this case was more closely correlated with Species patterns variations in daylength, even though these variations are small at 5° latitude. Although The pattern of flowering of individual species leaf fall is broadly linked to external condi- varies widely. In some species it is staggered tions, a number of studies indicate that there over a long period; in others there is a much is also an endogenous component to its tim- more pronounced peak with relatively few in- ing (Porras 1991; Loubry 1994; Williams et al. dividuals occurring in the early and late parts 1997). of the flowering period. The various types are well exemplified in a study of eight sympatric species in the Myrtaceae family in Central Flowering phenology Brazil by Proença & Gibbs (1994). These Community patterns ranged from mass-flowering (synchronous over a short period), pulsed flowering (re- The timing of flowering is one of the most peated flushes), to steady-state flowering (a widely investigated aspects of the phenology few per day over a long period). Each pattern of plant life-cycles, and has been studied on may be a response to a different set of selec- every scale, from the level of the community tive pressures. A strong peak of flowering (Murali & Sukumar 1994) to that of the indi- suggests a disadvantage to out-of-season in- vidual flower (Herrera 1995a). In most plant dividuals, possibly due to pollination reduc- communities, although at least some species tion in an out-crossing species. In the mass- will be in flower throughout the growing sea- flowering Andean shrub Befaria resinosa son, there is a tendency for peaks of flower- Melampy (1987) showed that the proportion ing to occur. In wet tropical forests flower pro- of flowers that set fruit was positively corre- duction may coincide with peaks of irradiance lated with flower abundance. Augspurger (Wright & Van Schaik 1994). In the season- (1981) induced individuals of the shrub Hy- ally dry tropical forests flowering is often con- banthus prunifolius to flower out of synchrony centrated in the transition from the late dry to with the general population, and found that the early wet season (Murali & Sukumar these plants had a lower seed set and greater 1994). A particularly marked example of the seed loss to predation than the general popu- concentration of community-wide flowering is lation. In a situation in which out-of-season found in the ground flora of temperate decid- individuals are at a disadvantage, there uous woodlands, where a pronounced peak would be a strong stabilizing selective pres- of flowering occcurs in spring before the tree sure in favour of synchrony. canopy closes (Heinrich 1976). Flower pro- Where flowering periods are more stag- duction and maintenance requires the expen- gered, other selective pressures may be op- diture of energy to form non-photosynthetic erating. Variation in weather from year to year tissue and nectar, the cost of which is consid- or changes in pollinator or seed predator erable (Ashman & Schoen 1997). In many abundance may result in inconsistent selec- cases its timing may be largely determined by tive pressures on flowering times, favouring seasonal changes in resource availability. In early, average or late individuals in different some studies, climatic factors (such as some years. In the tropical shrub Erythroxylum ha- measure of cumulative heat sum) are the vanense selection pressure on flowering time best predictors of flowering (Diekmann 1996; varies in direction from year to year (Domin- White 1995). Abiotic control of flowering time guez & Dirzo 1995). Brody (1997) also points may be strongest in harsh environments with out that the net effect of pollinators and seed a short growing season. For example, in predators varies both in time and space. alpine plant communities there may be a high Even in any one year the advantage of an in- level of selection in favour of early flowering dividual having a particular flowering time in in most species because of the restricted relation to the general population may be time available subsequently for seed matura- finely balanced. Zimmerman (1988) for ex- tion (Totland 1993). Smith-Ramírez & Ar- ample notes that late-flowering plants of mesto (1994) concluded from an investiga- Polemonium foliosissimum were more likely Phenology of growth and reproduction in plants 83 to outcross than the general population. The vidual species, though rather few records potential benefit of this may be counterbal- have been made for wild plants. Even in the anced by the possible disadvantages of late tropics, a response to changes in daylength blooming (less time for seed ripening, etc.). are not unknown. For example, Njoku (1958) Rathcke & Lacey (1985) point out that at least reports that daylength variations of 15 min- some degree of asynchronous flowering in a utes (as at Ibadan, 7°26′N), is sufficient to ini- population would have the benefit of promot- tiate flowering in many Nigerian herbs and ing outcrossing by forcing pollinators to move shrubs. A bout of cool temperature in tropical between individuals. The highest levels of rainforest can induce flowering in Diptero- outcrossing of a mass-flowering species pro- carpaceae in Malaysia (Ashton et al. 1988) bably occurs at the beginning and end of the and in a range of trees in Gabon (Tutin & Fer- blooming season (Carpenter 1976). Asyn- nandez 1993). In temperate climates, warm chrony within individuals would have the temperatures often act as flowering triggers, same effect (Rathcke & Lacey 1985). especially in mast seeding trees (Allen & Platt If selection has long-term evolutionary ef- 1990). Rain triggers flowering in the shrub fects, flowering times would have to be under Erythroxylum havanense in Mexico (Domin- genetic control. The fact that differences in guez & Dirzo 1995), and in the shrub Hyban- flowering time between individuals in natural thus prunifolius in Panama (Augspurger populations are often maintained from year to 1981). Experiments with desert ephemerals year (e.g. in Polemonium foliosissimum, Zim- in Namaqualand show that drought can bring merman & Gross 1984) suggests genetic forward the flowering date in late sown control. Various studies show that these dif- species (Steyn et al. 1996). ferences are maintained in cultivation, for example, in Rumex crispus (Akeroyd & Briggs Competition for pollinators and 1983), Solidago canadensis (Pors & Werner interference between species 1989) and Centaurea species (Lack 1982). Widén (1991) went one stage further by The extent to which flowering times of sym- showing that rank order of flowering in indi- patric species are influenced by competition viduals of the perennial herb Senecio integri- for pollinators has been much debated over folius was significantly correlated with that of the years. It is hypothesized that competition their progeny. Waser (1983) cites a number could be reduced by avoiding overlap in flow- of examples of the genetic control of flower- ering times. This might be expected to result ing time, and even suggests it to be a rela- in character displacement between coexist- tively flexible character. ing species sharing the same pollinators. A number of studies indicate that inter-specific Environmental cues for flowering exploitative competition for pollinators does occur (see examples cited in Rathcke 1983). The immediate cause of flowering is the de- Another form of competition between plants velopment of flowering initials in the appropri- sharing a pollinator is the interference which ate meristematic tissues. In some cases flow- occurs when pollen from one species is de- ering may simply result from the plant having posited on the stigma of another. Reduced reached a critical size or developmental seed set in the field as a result of hetero- stage, though this usually depends on a specific pollen transfer by shared pollinators cumulative heat-sum, such as degree-days is well established in a number of cases above a threshold (Diekmann 1996). In other (Galen & Gregory 1989; Waser 1978a,b). Its cases, flowering may be a physiological re- existence would provide a selection pressure sponse to a very specific environmental cue. for the separation of the flowering times of Examples of such cues are short or long two sympatric species. There is some evi- days, warm or cold temperatures, or an dence for this. Carpenter (1976) describes episode of rain or drought (Bernier et al. the case of two trees in Hawaii (Metrosideros 1981). A shared response in a population to a collina and Sophora chrysophylla) which specific cue has the advantage of resulting in share a pollinator (honeycreepers). The two synchronous flowering. In temperate regions species grow together at some elevations where daylength varies markedly with the and separately at others. Where they are seasons, a photoperiodic response may be a sympatric their flowering times are sequen- major determinant of flowering time for indi- tial; where they grow separately their flower- 84 M. Fenner ing periods overlap. This is at least circum- species also appear to have a staggered se- stantial evidence of local character displace- quence (Ashton et al. 1988), as do meadow ment. Another intriguing case in which a plant communities in the Rocky Mountains coexisting species pair shares a pollinator is (Pleasants 1980). However, there are difficult reported by Waser (1983). Ipomopsis aggre- statistical problems involved in testing the gata and Penstemon barbatus grow together significance of these sequences. A common on isolated mountains in south-west USA. approach is to compare the data with a ran- They are both pollinated by hummingbirds dom sequence. When this is done, few if any and always flower sequentially. The hypothe- examples can be distinguihed from a random sis that their separate flowering times may be sequence statistically. This was the case, for due to character displacement is supported example, with the analysis of Stiles’s data by by Waser’s discovery that the order of flower- Poole & Rathcke (1979), though this is ro- ing ‘flip-flops’ from place to place. This varia- bustly challenged by Stiles (1979). The tests tion in the sequence of flowering may be due applied in these cases may not always be ap- to the genetic characteristics of the founding propriate. Cole (1981) and Gleeson (1981) populations, but the consistency of the diver- modified the test and obtained different re- gence in flowering time strongly suggests sults. Useful accounts of the technicalities of local character displacement. applying tests to detect orderly sequences in Another interaction between plants which flowering are given by Gleeson (1981), Es- may provide a selection pressure affecting tabrook et al. (1982) and Pleasants (1990). flowering times is the advantage conferred on Rathcke (1984) also did a random model test one species by the presence of another on the flowering phenologies of plants from a species in flower at the same time. The joint swamp in Rhode Island, but again found that floral display may result in increased visita- an apparently ‘promising’ sequence could not tion rates, leading to enhanced seed set for be distinguished from random. one or both species. There is some evidence The present position with regard to charac- that this may occur (Thompson 1978, 1981; ter displacement due to competition for polli- 1982; Schemske 1981), but it is difficult to nators is that it is unproven, at least for long standardize all other factors that affect seed sequences, but that some cases certainly set. Facilitation, if effective, would provide a suggest a tendency towards it. It is probably selective force to converge flowering times rare in any case for a large group of sympatric between species. An interesting form of this plants to share exactly the same pollinators, ‘facilitation’ is the benefit that early flowering and since so many other forces influence plants confer on those flowering later in the flowering times, it would probably be difficult season, by supporting the initial pollinators to identify any tendency which did exist. (which survive to pollinate the later species). Cases involving pairs of species (Carpenter This ‘sequential mutualism’ was nicely 1976; Waser 1983) seem quite well estab- demonstrated by Waser & Real (1979) in Del- lished. The phenomenon would be most con- phinium nelsonii and Ipomopsis aggregata. vincing if shown in a guild of related plants. The complexity of these interactions is indi- However, it should be borne in mind that a cated by the fact that these two species were group of sequentially pollinated species may also shown to compete for pollination. be the consequence of what Janzen calls Many studies in which the flowering times ‘ecological fitting’ (Janzen 1985), resulting of species within communities are recorded from the arrival of widespread species whose show that there is often a temporal sequence local survival is due to their ‘fitting in’ to the ex- of flowering, with each species peaking in isting community. They were fortuitously pre- turn. When presented graphically, with the adapted, therefore, character displacement species placed in order of mean flowering due to competition for pollinators is unneces- date, the sequence often appears to be non- sary to account for their flowering time. It is random (that is, more evenly spaced and with possible that many, if not most, pollination less overlap than would be expected in a ran- guilds arise in this way. dom arrangement). Stiles (1977) gives a much-quoted example of a guild of humming- Predispersal seed predation bird-pollinated plants in a Costa Rican rain forest which show an orderly sequence of Another important potential biotic influence flowering peaks. Sympatric Dipterocarp on flowering time is pre-dispersal seed pre- Phenology of growth and reproduction in plants 85 dation. It can provide a selective pressure if it occurred alone than when it was accompa- is concentrated at a particular time within the nied by Polemonium foliosissimum. The com- flowering period. A concentration of flower- plexities of the interactions between plants head or seed predation in early flowering indi- sharing common pollinators and common viduals has been reported for Polemonium predators, combined with the effects of an- foliosissimum (Zimmerman 1980; Brody nual variations in climate on each of these or- 1997), Ipomopsis aggregata (Brody 1997), Si- ganisms separately, make it particularly diffi- lene vulgaris (Pettersson 1991), and Actaea cult to assess long-term net selective pres- spicata (Eriksson 1995). In some cases early sures on flowering time. and late season individuals suffer most damage. Fenner (1985) found that in a population Inherent constraints on flowering of Centaurea nigra, early and late capitula were much more prone to seed predation Each life-history stage of a plant is to some than those which flowered in mid-season. extent dependent upon or constrained by the This may have been due simply to a satura- preceeding and succeeding stages. For most tion effect at the peak of abundance, but the species, the time of flowering may be con- effect would be to stabilize flowering times strained by the timing of other phenophases around the mid-season date. In Baptisia aus- such as growth, seed dispersal or seed ger- tralis Evans et al. (1989) recorded greater mination (Johnson 1993). An important con- flower damage early and late in the season, straint occurs at the end of the season when giving a narrow window in time for flowering late flowering individuals may not have to occur. Significantly, however, the timing of enough time for fruit maturation (Helenurm & this window shifts from year to year, probably Barrett 1987; Rathcke 1988; Kudo 1993). selecting for the maintenance of a range of Dieringer (1991) showed that late flowers of flowering times within the population. The op- Agalinis strictifolia, a hemiparasitic grassland posite pattern has been recorded in Bartsia annual herb, produced fewer fruits than early alpina (Molau et al. 1989) and Erigeron glau- or middle-flowering individuals. Seed size cus (English-Loeb & Karban 1992), those was found to decline in late-flowering plants early and late flowers suffer less damage of Ranunculus adoneus (Galen & Stanton from predators. This would favour a pro- 1991). Primack (1985) neatly demonstrated longed period of blooming. In addition to se- the apparent constraint on flowering time im- lection for flowering time, predation may imposed by the need to allow a sufficient period pose other selective pressures. In Bartsia for fruit to ripen. He found a significant nega- alpina (Molau et al. 1989) and Tripleurosper- tive correlation between flowering date and mum inodorum (Fenner 1985) the larger in- volume of fruit amongst 115 insect-pollinated florescences have been shown to be prefer- species in the British flora. That is, species entially attacked. with larger fruit flowered earlier. Since larger Ideally, predation and pollination data fruit are assumed to require a longer period to should be combined in the same study (as in mature, the implication of the data is that fruit Pettersson 1991; English-Loeb & Karban size imposes at least some influence on flow- 1992; Brody 1997), so that the net effect of ering time. these two forces can be assessed. Zimmer- An important constraint on the flowering man (1980) gives an interesting example of a phenology of a species may be its taxonomic trade-off between pollination and predation in affinities. Related plants share similar inher- Polemonium foliosissimum in which the pre- ent design constraints which would limit their dators (fly larvae) preferentially attack those potential evolutionary response to selection. plants which receive highest numbers of polli- Analyses of various floras confirm that this nation visits. The balance of advantage to the phylogenetic effect on flowering time is plant shifts during the season. Another com- strong. Its influence has been demonstrated plicating factor which needs to be considered in the floras of South West Australia (Bell & in all these studies is the effect of the pres- Stephens 1984), North America and Japan ence of other plant species which share the (Kochmer & Handel 1986) and the Cape same seed predator. In Brody’s (1997) study (Johnson 1993). Coexisting congenerics of two species which share a common sometimes show rather similar phenologies, dipteran seed predator, Ipomopsis aggregata as for example, Commiphora species in East suffered much more damage in sites where it African dry forest vegetation (Fenner 1982). 86 M. Fenner

Clearly, the results of all studies in which the Since seed dissemination is the major influence of various putative selective pres- function of the fruit, we might expect its phe- sures is tested, should be interpreted in the nology to be influenced by selective pres- light of these phylogenetic limitations. sures which would favour successful dispersal. A number of studies have sought to test the hypothesis that wind-dispersed species Fruiting phenology release their disseminules (whether fruits or Selective pressures on timing individual seeds) at the most appropriate season. There is a general tendency, at least As with flowering, the phenology of fruiting is in the case of seasonally dry tropical forests, governed by its own set of constraints. Jor- for wind dispersed species to shed their dano (1992) has compiled a summary of fruit- seeds during the dry season (Ibarra-Man- ing phenology in a wide range of ecosystems, ríquez et al. 1991; De Lampe et al. 1992; which indicated that fruiting peaks generally Morellato & Leitao 1996). Dry season disper- occur during periods of low photosynthetic sal may be more effective as it coincides with activity or after periods of high rates of re- leaflessness. The timing of shedding of these serve accumulation. There is a broad divide wind-dispersed fruits contrasts with that of between temperate and tropical fruiting pat- fleshy or dry heavy fruits which are produced terns. In temperate regions fruiting has a uni- throughout the year (De Lampe et al. 1992). modal peak in late summer or autumn, with a For fleshy fruits which attract dispersal mean duration of less than 1.5 months. In agents, the timing of ripening would be ex- tropical communities there are usually some pected to coincide with the presence of the species in fruit at any given time of the year, dispersers. In the tropics there are numerous and individual species tend to have longer specialised frugivores which eat fruit through- fruiting periods, with a mean of more than 4.0 out the year. The continuous availability of months (Jordano 1992). These differences the fruit is a necessary part of the mutualism probably reflect the nature of the contraints between plant and disseminator. In contrast, on fruit timing in the two regions. The highly in temperate dispersers (mainly birds), fru- seasonal occurence of favourable tempera- givory is highly seasonal. In many temperate tures and irradiance in the middle latitudes regions the abundance of frugivorous birds confines the necessary resource build-up for varies seasonally, reaching a peak in autumn fruit production and ripening to a restricted during migration to winter quarters. A number period. In most cases it has to be squeezed in of studies record the concentration of fruiting between flowering time and the onset of win- in bird-dispersed species at this time: in ter. In a temperate Australian community, forests of eastern USA (Stiles 1980), in Illi- French (1992) suggests that fruit production nois (Thompson & Willson 1979) and in is largely controlled by the accumulation of northern and southern Europe (Snow 1971; enough photosynthate, which can only occur Herrera 1984). Snow (1971) points out that towards the end of the growing season. In in- fruits of bird-dispersed species ripen later at dividual cases, flowering time may further re- increasingly southerly latitudes in Europe, in strict the fruiting calendar slot. Late-flowering conformity with the southerly migration of fru- Rhododendron aureum in northern Japan givorous birds in autumn. The phenology of fails to set fruits because of the onset of fleshy fruit production in the Mediterranean autumn frost and snow, preventing fruit ripen- closely matches that of the seasonal abun- ing (Kudo 1993). Physical limitations on fruit- dance of frugivorous avian dispersers (Her- ing time are probably less acute in the tropics, rera 1995b). though seasonal rainfall is probably the single most important abiotic factor controlling re- Mast fruiting source availability there (Borchert 1994). Tim- ing of fruiting may also be under some selec- The production of fruits by an individual plant tive pressure to disperse seeds at the start of is often irregular. For many species, fruit the rainy season to facilitate germination crops appear only once every few years in an (Garwood 1983). However, in at least some unpredictable pattern, with several barren tropical rain forests fruiting has virtually no seasons in between. There is a tendency for seasonal pattern when viewed from the com- all the plants in a population to produce fruit munity level (Putz 1979; Hilty 1980). in the same year. This phenomenon, called Phenology of growth and reproduction in plants 87

‘masting’, is seen in a range of long-lived giving rise to severe shortages which few fru- species, and is characteristic of many tem- givores can survive. Those species which perate broadleaved trees (Hilton & Packham fruit in the season of general scarcity there- 1986; Allen & Platt 1990) and conifers (Nor- fore have a central role in the maintenance of ton & Kelly 1988). Irregular supra-annual community function and diversity. They are fruiting also occurs in the tropical Diptero- true ‘keystone species’ in the sense that their carpaceae, montane grasses and perennial removal would cause major community im- herbs (Primack 1985). Mast fruiting is linked poverishment. Howe (1984) and Terborgh to the formation of flowering initials in re- (1986) give examples of providers of ‘key- sponse to environmental cues which occur stone plant resources’. In many tropical only at irregular intervals (Fenner 1991). Sev- forests this role falls to members of the fig eral hypotheses which attribute an adaptive genus (Ficus) whose continuous fruiting may significance to mast behaviour have been put be partly a consequence of its obligate mutu- forward. It is often seen as a mechanism of alism with wasp pollinators (Lambert & Mar- seed predator deterence, in which the preda- shall 1991; Milton 1991). A knowledge of fruit- tors are alternately starved and satiated (Sil- ing phenology is of central importance for vertown 1980). The fact that species which conservation management generally, but is of share a common seed predator tend to mast special value in tropical forests which support in the same year provides some support for a diversity of fruit-dependent species. this hypothesis, since those which do not conform to the mass fruiting would attract the predators (Ashton et al. 1988). An alternative Conclusions (or possibly additional) advantage would be more effective pollination by massed flower- From the current literature on phenological ing (Norton & Kelly 1988). A further consider- patterns in plant communities it is possible to ation is that supra-annual fruiting may just be draw a few tentative conclusions and to indi- a consequence of the need to build up re- cate certain useful directions for future work. serves between crops. The cost of fruiting is One point which stands out is the need for high and known to have a depressive effect long-term research in this area. For example, on growth in subsequent years (Haase a single-year study showing that early and 1986). late flowers suffer more predation than mid- season flowers would suggest that selection Keystone fruit providers would result in syncronous flowering. How- ever, if several years are recorded and it is The phenology of fruiting has very important found that the timing of the ‘window’ of re- consequences for community diversity be- duced predation shifts from year to year (as cause of the dependence of frugivores on a in Baptisia australis, Evans et al. 1989) our constant supply of food. Many tropical forests conclusion would be that selection favours support a range of specialist fruit-eating asynchrony. With masting species, few con- species, such as birds, primates, bats and in- clusions at all can be drawn for at least a cou- sects. The availability of fruit in most commu- ple of decades. One very good reason for col- nities varies considerably through the year lecting long-term masting data is to monitor (Friedel et al. 1993), and many frugivores can the biological effect (if any) of global climate tide themselves over a period of scarcity by change. Masting species are likely to be par- various behavioural devices, such as dietary ticularly sensitive to changes in the frequency switching (to leaves, nectar, etc.), a change in of the climatic cues which trigger flowering. range, migration and seasonal breeding (Van An increase in mean temperature would in- Schaik et al. 1993). It is probable that the car- crease the frequency of fruit production in rying capacity for frugivores is set by the nar- some species (possibly having deleterious rowness of these ‘bottle necks’ or ‘pinch effects of exhaustion), and would decrease points’ in fruit availability. For example, the fruiting in those species which require a cold- Urucu terra firme forest in Brazil has a re- temperature cue, such as Dacrydium cu- markably low density of frugivores compared pressinum (Norton & Kelly 1988). As major with comparable forests elsewhere. Peres economic species (e.g. most softwoods) (1994) suggests that this may be linked to the would be affected, as well as many species of strongly seasonal production of fruit there, ecological importance, this aspect of repro- 88 M. Fenner ductive phenology would seem worthy of keep in mind the contraints imposed by their more extensive investigation (Fenner 1991; phylogenetic inheritance. The flowers that Lechowicz 1995). The reproductive phenol- bloom in spring may do so largely because ogy of whole communities may be markedly their ancestors have bequeathed to them a affected by climate change, e.g. montane genetic constitution for doing so. It is highly grasslands (McKone et al. in press) and probable that these phylogenetic constraints tropical forests (Reich 1995). severely limit the ‘room for manoeuvre’ avail- A number of studies illustrate the complex- able for selection and hence adaptation ity of the interacting forces which may collec- (Bradshaw 1984). The selective forces which tively influence flowering times. The effect of we investigate (seasonal availability of re- competition between plants for pollinators sources, competition for pollinators, avoid- must be seen in the context of possible facili- ance of loss to seed predators) may have tation occurring at the same time (Waser & only marginal effects on highly conservative Real 1979). The effect of seed predators genomes. should take into account other plant species sharing those predators (Brody 1997). Super- Lo! In the middle of the wood, imposed on these biotic influences are the The folded leaf is woo’d from out the bud abiotic constraints, chiefly the timing of re- With winds upon the branch, and there source availability. Since we cannot expect Grows green and broad, and takes no care, every investigation to be comprehensive, fu- Sun-steep’d at noon, and in the moon ture work might profitably be concentrated on Nightly dew-fed; and turning yellow systems in which we already have a good Falls, and floats adown the air. grounding. Investigations which are too nar- rowly focussed are in danger of missing some unconsidered influence which may be References of over-riding importance. The results of most phenological investi- Aide, T.M. (1991) Synchronous leaf production and gations have been expressed in some quanti- herbivory in juveniles of Gustavia superba. Oecolo- tative way, such as percent reduction in seed gia, 88, 511–514. set due to competition for pollinators, or per- Aide, T.M. (1992) Dry season leaf production: An escape from herbivory. Biotropica, 24, 532–537. cent seed loss by predation. Rather few have Aide, T.M. (1993) Patterns of leaf development and looked at the qualitative consequences of the herbivory in a tropical understory community. Ecol- differential survival of certain genotypes. In ogy, 74, 455–466. some cases the disadvantages of a lower Akeroyd, J.R. & Briggs, D. (1983) Genecological studies rate of survival in early or late individuals may of Rumex crispus L. Garden experiments using be more than compensated by the long-term transplanted material. New Phytologist, 94, 309–323. Allen, R.B. & Platt, K.H. (1990) Annual seedfall varia- genetic advantages of higher rates of out- tion in Nothofagus solandri (Fagaceae), Canter- crossing. For example, Zimmerman (1988) bury, New Zealand. Oikos, 57, 199–206. demonstrated increased outcrossing in late Al-Mufti, M.M., Sydes, C.L., Furness, S.B., Grime, J.P. flowering Polemonium foliosissimum individ- & Band, S.R. (1977) A quantitative analysis of shoot uals. In Bartsia alpina seed predation is con- phenology and dominance in herbaceous vegeta- centrated on the self-pollinated flowers, tion. Journal of Ecology, 65, 759–791. Ashman, T.-L. & Schoen, D.J. (1997) The cost of floral thereby selecting in favour of the outcrossed longevity in Clarkia tembloriensis: an experimental individuals which then form a larger propor- investigation. Evolutionary Ecology, 11, 289–300. tion of the seed pool (Molau et al. 1989). An Ashton, P.S., Givnish, T.J. & Appanah, S. (1988) Stag- understanding of how these genetic differ- gered flowering in the Dipterocarpaceae: new in- ences in the surviving progeny affect their rel- sights into floral induction and the evolution of mast fruiting in the aseasonal tropics. American Natural- ative fitness is central to assessing the ef- ist, 132, 44–66. fects of the various mortality agents on flow- Augspurger, C.K. (1981) Reproductive synchrony of a ering times. The importance of considering tropical shrub: Experimental studies on effects of the long-term evolutionary consequences of pollinators and seed predators on Hybanthus pruni- ecological processes is stressed by Brad- folius (Violaceae). Ecology, 62, 775–788. shaw (1984). Bell, D.T. & Stephens, L.J. (1984) Seasonality and phenology of kwongan species. Kwongan: Plant Finally, one must emphasize again that in Life of the Sandplain (eds. J.S. Pate & J.S. Beard), our search for selective forces, which shape pp. 205–226. University of Australia Press, Ned- the phenology of plants, we must always lands. Phenology of growth and reproduction in plants 89

Bernier, G., Kinet, J.-M. & Sachs, R.M. (1981) The Fenner, M. (1991) Irregular seed crops in forest trees. Physiology of Flowering. CRC Press Inc., Florida. Quarterly Journal of Forestry, 85, 166–172. Borchert, R. (1994) Soil and stem water storage deter- Frankie, G.W., Baker, H.G. & Opler, P.A. (1974) Com- mine phenology and distribution of tropical dry for- parative phenological studies of trees in tropical est trees. Ecology, 75, 1437–1449. wet and dry forests in the lowlands of Costa Rica. Bradshaw, A.D. (1984) The importance of evolutionary Journal of Ecology, 62, 881–913. ideas in ecology – and vice versa. Evolutionary French, K. (1992) Phenology of fleshy fruits in a wet Ecology (ed. B. Shorrocks), pp. 1–25. Symposium sclerophyll forest in southeastern Australia: Are BES 23, Blackwells Scientific Publications, Oxford. birds an important influence? Oecologia, 90, Brody, A.K. (1997) Effects of pollinators, herbivores, 366–373 and seed predators on flowering phenology. Ecol- Friedel, M.H., Nelson, D.J., Sparrow, A.D., Kinloch, J.E., ogy, 78, 1624–1631. & Maconochie, J.R. (1993) What induces central Brooke, M. de L., Jones, P.J., Vickery, J.A. & Waldren, Australian arid zone trees and shrubs to flower and S. (1996) Seasonal patterns of leaf growth and fruit? Australian Journal of Botany, 41, 307–319. loss, flowering and fruiting on a subtropical central Galen, C. & Gregory, T. (1989) Interspecific pollen Pacific island. Biotropica, 28, 164–179. transfer as a mechanism for competition: conse- Bullock, S.H. & Solís-Magallanes, J.A. (1990) Phe- quences of foreign pollen contamination for seed nology of canopy trees of a tropical deciduous set in an alpine wildflower, Polemonium viscosum. forest in Mexico. Biotropica, 22, 22–35. Oecologia, 81, 120–123. Carpenter, F.L. (1976) Plant-pollinator interactions in Galen, C. & Stanton, M.L. (1991) Consequences of Hawaii: pollination energetics of Metrosideros col- emergence phenology for reproductive success in lina (Myrtaceae). Ecology, 57, 1125–1144. Ranunculus adoneus (Ranunculaceae). American Chidumayo, E.N. (1994) Phenology and nutrition of Journal of Botany, 78, 978–988. Miombo woodland trees in Zambia. Trees – Struc- Garwood, N.C. (1983) Seed germination in a seasonal ture and Function, 9, 67–72. tropical forest in Panama: A community study. Eco- Cole, B.J. (1981) Overlap, regularity, and flowering logical Monographs, 53, 159–181. phenologies. American Naturalist, 117, 993–997. Gleeson, S.K. (1981) Character displacement in flow- De Lampe, M.G., Bergeron, Y., McNeil, R. & Leduc, A. ering phenologies. Oecologia, 51, 294–295. (1992) Seasonal flowering and fruiting patterns in Grime, J.P. & Mowforth, M.A. (1982) Variation in tropical semiarid vegetation of north-eastern genome size – an ecological interpretation. Nature, Venezuela. Biotropica, 24, 64–76. 299, 151–153. Delarco, J.M., Escudero, A. & Garrido, M.V. (1991) Ef- Grime, J.P., Shacklock, J.M.L. & Band, S.R. (1985). fects of site characteristics on nitrogen retransloca- Nuclear DNA contents, shoot phenology and tion from senescing leaves. Ecology, 72, 701–708. species co-existence in a limestone grassland Diekmann, M. (1996) Relationship between flowering community. New Phytologist, 100, 435–445. phenology of perennial herbs and meteorological Haase, P. (1986) Flowering records of some subalpine data in deciduous forests of Sweden. Canadian trees and shrubs at Arthur’s Pass, New Zealand. Journal of Botany, 74, 528–537. New Zealand Journal of Ecology, 9 19–23. Dieringer, G. (1991) Variation in individual flowering Heinrich, B. (1976) Flowering phenologies: bog, wood- time and reproductive success of Agalinis stricti- land, and disturbed habitats. Ecology, 57, 890–899. folia (Scrophulariaceae). American Journal of Bo- Helenurm, K. & Barrett, S.C.H. (1987) The reproduc- tany, 78, 497–503. tive biology of boreal forest herbs. 2. Phenology of Dominquez, C.A. & Dirzo, R. (1995) Rainfall and flow- flowering and fruiting. Canadian Journal of Botany, ering synchrony in a tropical shrub: variable selec- 65, 2047–2056. tion on the flowering time of Erythroxylum hava- Herrera, C.M. (1984) A study of avian frugivores, bird- nense. Evolutionary Ecology, 9, 204–216. dispersed plants, and their interactions in Mediter- English-Loeb, G.M. & Karban, R. (1992) Consequences ranean scrublands. Ecological Monographs, 54, of variation in flowering phenology for seed head her- 1–23. bivory and reproductive success in Erigeron glaucus Herrera, C.M. (1995a) Floral biology, microclimate, (Compositae). Oecologia, 89, 588–595. and pollination by ectothermic bees in an early- Eriksson, O. (1995) Asynchronous flowering reduces blooming herb. Ecology, 76, 218–228. seed predation in the perennial forest herb Actaea Herrera, C.M. (1995b) Plant-vertebrate seed dispersal spicata. Acta Oecologica, 16, 195–203. systems in the Mediterranean: ecological, evolu- Estabrook, G.F., Winsor, J.A., Stephenson, A.G. & tionary, and historical determinants. Annual Review Howe, H.F. (1982) When are two phenological pat- of Ecology and Systematics, 26, 705–727. terns different? Botanical Gazette, 143, 374–378. Hilton, G.M. & Packham, J.R. (1986) Annual and re- Evans, E.W., Smith, C.C. & Gendron, R.P. (1989) Tim- gional variation in English beech mast (Fagus syl- ing of reproduction in a prairie legume: seasonal vatica L.). Arboricultural Journal, 10, 3–14. impacts of insects consuming flowers and seeds. Hilty, S.L. (1980) Flowering and fruiting periodicity in a Oecologia, 78, 220–230. premontane rain forest in Pacific Colombia. Fenner, M. (1982) Aspects of the ecology of Acacia- Biotropica, 12, 292–306. Commiphora woodland near Kibwesi, Kenya. Jour- Howe, H.F. (1984) Implications of seed dispersal by nal of the East African Natural History Society & animals for tropical reserve management. Biologi- National Museum of Kenya, 175, 1–11. cal Conservation, 30, 261–281. Fenner, M. (1985) Seed Ecology. Chapman & Hall, Ibarra–Manríques, G., Sánchez-Garfias, B. & Gonzá- London. lez-Garcia, L. (1991) Fenologia de lianas y arboles 90 M. Fenner

anemocoros en una selva calido-humeda de Méx- bivory in an oak population: the role of plant phenol- ico. Biotropica, 23, 242-254. ogy and insect performance. Ecology, 76, Janzen, D.H. (1985) On ecological fitting. Oikos, 45, 1233–1241. 308–310. Morellato, P.C. & Leitao, H.F. (1996) Reproductive Johnson, S.D. (1993) Climatic and phylogenetic deter- phenology of climbers in a south-eastern Brazilian minants of flowering seasonality in the Cape flora. forest. Biotropica, 28, 180–191. Journal of Ecology, 81, 567–572. Murali, K.S. & Sukumar, R. (1993) Leaf flushing phe- Jordano, P. (1992) Fruits and frugivory. Seeds: The nology and herbivory in a tropical dry deciduous Ecology of Regeneration in Plant Communities (ed. forest, southern India. Oecologia, 94, 114–119. M. Fenner), pp. 105–156. CAB International, Ox- Murali, K.S. & Sukumar, R. (1994) Reproductive phe- ford. nology of a tropical dry forest in Mudumalai, south- Kochmer, J.P. & Handel, S.N. (1986) Constraints and ern India. Journal of Ecology, 82, 759–767. competition in the evolution of flowering phenology. Newstrom, L.E., Frankie, G.W. & Baker, H.G. (1994) A Ecological Monographs, 56, 303–325. new classification for plant phenology based on Kudo, G. (1993) Relationship between flowering time flowering patterns in lowland tropical rain forest and fruit set of the entomophilous alpine shrub, trees at La Selva, Costa Rica. Biotropica, 26, Rhododendron aureum (Ericaceae), inhabiting 141–159. snow patches. American Journal of Botany, 80, Njoku, E. (1958) The photoperiodic response of some 1300–1304. Nigerian plants. Journal of the West African Sci- Kursar, T.A. & Coley, P.D. (1992) Delayed greening in ence Association, 4, 99–111. tropical leaves: an antiherbivore defense. Biotro- Norton, D.A. & Kelly, D. (1988) Mast seeding over 33 pica, 24, 256–262. years by Dacrydium cupressinum Lamb. (rimu) Lack, A.J. (1982) Competition for pollinators in the (Podocarpaceae) in New Zealand: the importance ecology of Centaurea scabiosa L. and Centaurea of economies of scale. Functional Ecology, 2, nigra L. I. Variation in flowering time. New Phytolo- 399–408. gist, 91, 297–308. Peres, C.A. (1994) Primate responses to phenological Lambert, F.R. & Marshall, A.G. (1991) Keystone char- changes in an Amazonian terre firme forest. acteristics of bird–dispersed Ficus in a Malaysian Biotropica, 26, 98–112. lowland rain forest. Journal of Ecology, 79, Pettersson, M.W. (1991) Flower herbivory and seed 793–809. predation in Silene vulgaris (Caryophyllaceae): Ef- Lechowicz, M.J. (1995) Seasonality of flowering and fects of pollination and phenology. Holarctic Ecol- fruiting in temperate forest trees. Canadian Journal ogy, 14, 45–50. of Botany, 73, 175–182. Pleasants, J.M. (1980) Competition for bumblebee Lechowicz, M.J. & Koike, T. (1995) Phenology and pollinators in Rocky mountain plant communities. seasonality of woody plants: an unappreciated ele- Ecology, 61, 1446–1459. ment in global change research. Canadian Journal Pleasants, J.M. (1990) Null models tests for competi- of Botany, 73, 147–148. tive displacement: the fallacy of not focusing on the Lee, W.G., Grubb, P.J. & Wilson, J.B. (1991). Patterns whole community. Ecology, 71, 1078–1084. of resource allocation in fleshy fruits of nine Euro- Poole, R.W. & Rathcke, B.J. (1979) Regularity, ran- pean tall shrub species. Oikos, 61, 307–315. domness, and aggregation in flowering phenolo- Lieberman, D. & Lieberman, M. (1984) The causes gies. Science, 203, 470–471. and consequences of synchronous flushing in a dry Porras, R.C. (1991) Phenology of Quercus seemannii tropical forest. Biotropica, 16 193–201. Lieb (Fagaceae) in Cartago, Costa Rica. Revista Loubry, D. (1994) La phénologie des arbres caduci- de Biologia Tropical, 39, 243–248. foliés en forêt guyanaise (5o de latitude nord): Illus- Pors, B. & Werner, P.A. (1989) Individual flowering tration d’un déterminisme à composantes endo- time in a goldenrod (Solidago canadensis): Field gène et exogène. Canadian Journal of Botany, 72, experiment shows genotype more important than 1843–1857. environment. American Journal of Botany, 76, Lowman, M.D. (1992) Leaf growth dynamics and her- 1681–1688. bivory in five species of Australian rain-forest Primack, R.B. (1985) Patterns of flowering phenology canopy trees. Journal of Ecology, 80, 433–447. in communities, populations, individuals and single McKone, M.J., Kelly, D. & Lee, W.G. (in press) Cli- flowers. The Population Structure of Vegetation mate change could decrease seed production of (ed. J. White), pp. 571–593. Junk, Dordrecht. masting species via population responses by spe- Proença, C.E.B. & Gibbs, P.E. (1994) Reproductive cialist seed-predator insects. Global Change Bio- biology of eight sympatric Myrtaceae from central logy. Brazil. New Phytologist, 126, 343–354. Melampy, M.N. (1987) Flowering phenology, pollen Putz, F.E. (1979) Aseasonality in Malaysian tree phe- flow and fruit production in the Andean shrub Be- nology. Malaysian Forester, 42, 1–24. faria resinosa. Oecologia, 73, 293–300. Quiring, D.T. (1994) Influence of inter-tree variation in Milton, K. (1991) Leaf change and fruit production in time of budburst of white spruce on herbivory and six neotropical Moraceae species. Journal of Ecol- the behaviour and survivorship of Zeiraphera cana- ogy, 79, 1–26. densis. Ecological Entomology, 19, 17–25. Molau, U., Eriksen, B & Knudsen, J.T. (1989) Pre-dis- Rathcke, B.J. (1983) Competition and facilitation persal seed predation in Bartsia alpina. Oecologia, among plants for pollination. Pollination Biology 81, 181–185. (ed. L.A. Real), pp. 305–338. Academic Press, New Mopper, S. & Simberloff, D. (1995) Differential her- York. Phenology of growth and reproduction in plants 91

Rathcke, B.J. (1984) Patterns of flowering phenolo- temperate fruit/bird interactions: phenological gies: testability and causal inference using a ran- strategies. Evolution, 33, 973–982. dom model. Ecological Communities, Conceptual Tissue, D.T. & Wright, S.J. (1995) Effect of seasonal Issues and the Evidence (eds. D.R. Strong, D. water availability on phenology and the annual Simberloff, L.G. Abele, & A.B. Thistle), pp. shoot carbohydrate cycle of tropical forest shrubs. 383–393. Princeton University Press, Princeton, Functional Ecology, 9, 518–527. New Jersey. Totland, O. (1993) Pollination in alpine Norway: flower- Rathcke, B.J. (1988) Flowering phenologies in a shrub ing phenology, insect visitors, and visitation rates in community: competition and constraints. Journal of two plant communities. Canadian Journal of Ecology, 76, 975–994. Botany, 71, 1072–1079. Rathcke, B. J. & Lacey, E.P. (1985) Phenological pat- Tutin, C.E.G. & Fernandez, M. (1993) Relationships terns of terrestrial plants. Annual Review of Ecol- between minimum temperature and fruit production ogy and Systematics, 16, 179–214. in some tropical forest trees in Gabon. Journal of Reich, P.B. (1995) Phenology of tropical forests: pat- Tropical Ecology, 9, 241–248. terns, causes, and consequences. Canadian Jour- Van Schaik, C.P., Terborgh, J.W. & Wright, S.J. (1993) nal of Botany, 73, 164–174. The phenology of tropical forests: adaptive signifi- Rusch, V.E. (1993) Altitude variation in the phenology cance and consequences for primary consumers. of Nothofagus pumilio in Argentina. Revista Chilena Annual Review of Ecology and Systematics, 24, de Historia Natural, 66, 131–141. 353–377. Schemske, D.W. (1981) Floral convergence and polli- Waser, N.M. (1978a) Competition for pollination and nator sharing in two bee-pollinated tropical herbs. sequential flowering in two Colorado wildflowers. Ecology, 62, 946–954. Ecology, 59, 934–944. Schnelle, F. (1965–1986) Beiträge zur Phänologie Eu- Waser, N.M. (1978b) Interspecific pollen transfer and ropas. Berichte des Deutschen Wetterdienstes, Of- competition between co-occurring plant species. fenbach (Main). Oecologia, 36, 223–236. Schnelle, F. (1977) Ergebnisse aus den internationa- Waser, N.M. (1983) Competition for pollination and flo- len phänologischen Gärten. Berichte des Deut- ral character differences among sympatric plant schen Wetterdienstes Nr. 144, Offenbach (Main). species: a review of evidence. Handbook of Experi- Silvertown, J.W. (1980) The evolutionary ecology of mental Pollination Biology (eds. C.E. Jones & R.J. mast seeding in trees. Biological Journal of the Lin- Little), pp. 277–293. Van Nostrand-Reinhold, New nean Society, 14, 235–250. York. Singh, J.S. & Singh, V.K. (1992) Phenology of season- Waser, N.M. & Real, L.A. (1979) Effective mutualism ally dry tropical forest. Current Science, 63, 684–689. between sequentially flowering plant species. Na- Smith-Ramírez, C. & Armesto, J.J. (1994) Flowering ture, 281, 670–672. and fruiting patterns in the temperate rain-forest of White, L.M. (1995) Predicting flowering of 130 plants Chiloe, Chile: ecologies and climatic constraints. at eight locations with temperature and daylength. Journal of Ecology, 82, 353–365. Journal of Range Management, 48, 108–114. Snow, D.W. (1971) Evolutionary aspects of fruit-eating Widén, B. (1991) Environmental and genetic influ- birds. Ibis, 113, 194–202. ences on phenology and plant size in a perennial Steyn, H.M., Van Rooyen, N., Van Rooyen, M.W. & herb, Senecio integrifolius. Canadian Journal of Theron, G.K. (1996) The phenology of Namaqua- Botany, 69, 209–217. land ephemeral species: the effect of water-stress. Williams, R.J., Myers, B.A., Muller, W.J., Duff, G.A. & Journal of Arid Environments, 33, 49–62. Eamus, D. (1997) Leaf phenology of woody Stiles, E.W. (1980) Patterns of fruit presentation and species in a North Australian tropical savanna. seed dispersal in bird-disseminated woody plants Ecology, 78, 2542–2558. in the eastern deciduous forest. American Natural- Wright, S.J. (1991) Seasonal drought and the phenol- ist, 116, 670–686. ogy of understory shrubs in a tropical moist forest. Stiles, F.G. (1977) Coadapted competitors: the flower- Ecology, 72, 1643–1657. ing seasons of hummingbird-pollinated plants in a Wright, S.J. & Van Schaik, C.P. (1994) Light and the tropical forest. Science, 198, 1177–1178. phenology of tropical trees. American Naturalist, Stiles, F.G. (1979) Regularity, randomness, and 143 192–199. aggregation in flowering phenologies: a reply. Zimmerman, M. (1980) Reproduction in Polemonium: Science, 203, 471. pre-dispersal seed predation. Ecology, 61, Terborgh, J. (1986) Community aspects of frugivory in 502–506. tropical forests. Frugivores and Seed Dispersal Zimmerman, M. (1988) Nectar production, flowering (eds. A. Estrada & T. H. Fleming), pp. 371–384. phenology, and strategies for pollination. Plant Re- Junk, Dordrecht. productive Ecology (eds. J. Lovett-Doust & L. Thompson, J.D. (1978) Effect of stand composition on Lovett-Doust), pp. 157–178. Oxford University insect visitation in two-species mixtures of Hiera- Press, Oxford. cium. American Midland Naturalist, 100, 431–440. Zimmerman, M. & Gross, R.S. (1984) The relationship Thompson, J.D. (1981) Spatial and temporal compo- between flowering phenology and seed set in an nents of resource assessment of flower-feeding in- herbaceous perennial plant, Polemonium foliosissi- sects. Journal of Animal Ecology, 50, 49–59. mum Gray. American Naturalist, 111, 185–191. Thompson, J.D. (1982) Pattern of visitation by animal pollinators. Oikos, 39, 241–250. Received 19 January 1998 Thompson, J.N. & Willson, M.F. (1979) Evolution of Revised version accepted 6 April 1998