Home

Flowering periodicity

Tree Autecology and Biology

Temperate Forest Ecology

Tropical Forest Ecology

Forestry

Sand Dune Ecology

East Usambaras

Pitcairn Islands

Over the past century the periodicity in flowering and fruiting of temperate forest trees has been the focus of much attention (e.g. Sarvas 1955, Davis 1957, Rohmeder 1957, Matthews 1963) and particularly in relation to mast years (e.g. Silvertown 1980). For economic reasons the alternate bearing observed in fruit trees, particularly apples, has drawn much interest (e.g. Monselise & Goldschmit 1982) while European foresters have been interested in the main timber trees. Variation in flowering and fruiting has been investigated in several broadleaves including Betula spp. (Sarvas 1952, Kinnaird 1974, Pelham et al. 1984), Fagus sylvatica (Matthews 1955, Holmsgaard 1962, Wachter 1964, Nilsson & Wüstlung 1987), Fraxinus excelsior (Tapper 1992) and Quercus spp. (e.g. Maurer 1964) and several conifer species (e.g. Koski & Tallqvist 1978, Leikola et al. 1982). In Denmark, Andersen (1974) investigated several deciduous tree species between 1967 and 1972 and found that most species flowered abundantly every second year in forest conditions. In Wales Hyde (1951), using the amount of areal pollen over six years, found that large pollen catches occurred by-early peaks in only three out of seven investigated deciduous tree species.

The factors underlying the existence of mast years have been discussed in relation to seed predation by Janzen (1971) and Silvertown (1980) and wind pollination by Smith et al. (1990). Fenner (1991) provides an overview of the factors potentially regulating the periodicity of flowering and fruiting in forest trees. The main factors involved in masting include seed predator escape (i.e. more seeds survive), flowering synchronization (i.e. the probability of successful pollination is increased), environmental conditions (e.g. spring frost kills flowers), and resource allocation (i.e. costs of reproduction are high).

Tree species with little economic importance have been neglected including species such as sycamore and ash. Fenner (1991) wonders whether "wind dispersed species like ash, hornbeam, lime and maple `mast'? If so, what cues induce bumper crops?" In sycamore Evans (1984) stated that the interval between large seed crops is one to three years and this figure is widely accepted (e.g. Philipson 1990, Savill 1991). In the literature there is the occasional reference to sycamore flowering and fruiting intensity. In 1976 abundant flowering was recorded in Great Britain and the seed crop was described as "massive" not only in sycamore but in most tree species (Lines 1977). Montgomery et al.'s (1991) data indicate that 1984 was an exceptional fruiting year in Northern Ireland. High sycamore seedling numbers (up to 154/m2) recorded in a Belgium forest (sycamore constituted 3-8% of total basal area) also indicate a very high seed production in 1976 (van Miegroet et al. 1981). In Germany sycamore was recorded as producing a good seed crop in southern parts of the country in 1983 and it flowered well in 1984 but much variation was observed with a decrease in flowering intensity at higher altitudes (Eicke 1984).

The main difficulty in the study of the variation in flowering and fruiting and the search for factors regulating their periodicity and intensity is the gathering of quantitative data over a long period of time. Foresters who, until recently, were the only workers to gather data on this topic, have long records but only used subjective measurements (usually a scale of 1 to 4). Among some of the more recent studies, that by Norton & Kelly (1988), investigating the seed production in Rimu over 33 years, did not include any assessment of intra-population variation in fruiting.

In this Section the annual variation in sycamore flowering is investigated. As opposed to most other investigations the data are 1/ quantitative, expressed in terms of the number of buds flowering each year, 2/ includes more than one population, 3/ investigates intra-population variation, but 4/ only includes a run of population data over seven years although data for up to 15 years is available for some trees.

METHODS

Three populations (Ballyversal, Murlough Bay and Portstewart) from the north coast of Ireland were used in this study. At each site three branches per tree were collected and brought back to the laboratory in 1986. The estimates of flowering intensity were made using the method described in Section 2.6.2. Flowering intensity in any one year is expressed as the average percentage of buds flowering. Population flowering intensity is the average percentage flowering of all the trees.

In order to compare flowering in sycamore in Ireland (non native range) and its native range (Alps) a site was selected in the Swiss Prealps. Because all trees investigated in Ireland have well developed lower branches on at least 40% of the crown, it was thought that forest trees would not be adequate for a comparison of flowering intensity. Instead a population of free standing sycamores on a north facing scree slope were chosen at Le Mollard (about 10km north of Montreux) at an altitude of about 1300m. Due to time and physical constraints only one branch from each of five trees was sampled in summer 1988. Because the trees were free standing it is believed that inter-branch variation in flowering intensity is limited, though the small number of trees investigated may not be truly representative of the population.

RESULTS

In Ireland all trees flowered every year between 1980 and 1986, but large variations were observed between years and between trees. Table 1 shows some of the variation observed ranging from constant high flowering (e.g. P19) to trees exhibiting clear bi-yearly peaks in flowering (e.g. P21), whereas others do not exhibit obvious regular cycles (e.g. P26, P27). In all observed trees flowering intensity in 1984 was greater than 75% whereas in other years a very wide variation in flowering intensity between trees could be observed.

Table 1. Mean and standard of deviation in flowering intensity (% of buds flowering) in a selected number of sycamore trees at Portstewart.

Table 2. Mean and standard of deviation in flowering intensity in three Irish populations (average percentage of flowering buds in the population).

At Le Mollard four of the five trees were found not to flower during at least one year and the variation in flowering intensity between trees appeared to be greater than in Ireland.

The flowering intensity at the Irish and Swiss sites show the same pattern of bi-yearly flowering. Flowering in Ireland in 1984 was exceptional with an intensity of over 90%, but in other years there was a large variation between populations. In the Alps the trees studied flowered less than in Ireland but 1984 was also an exceptionally good flowering year. The years 1978, 1987 and 1988 had flowering intensities of less than 5%.

DISCUSSION

Flowering in Ireland occurs every year but flowering intensity shows bi-yearly peaks with a very pronounced peak in 1984. Individual trees all exhibit a peak in 1984 but in other years most trees show different patterns of flowering, but with some individuals showing typical bi-yearly cycles. In sycamore the inflorescence initiation takes place the previous summer, starting in mid-July according to Anderson & Guard (1964), and the summer of 1983 was particularly hot. The mean average July-August temperature was more than 1oC warmer in 1983 than in other years. The prolific flowering of 1976 reported by Lines (1977) followed another hot summer.

Observations on urban trees growing in Montreux further support the view that warm summers the year prior to flowering are necessary for `mast' flowering to occur. Montreux is situated on the shore of the Léman on south-west facing slopes and is backed by the Alps. The combination of these physical factors means that the Montreux climate is much warmer, particularly in spring, than elsewhere in the region. Sycamore observed along one of the town's streets produce every year a very abundant crop of fruits and include the largest infructescences I have ever observed.

The intensity of flowering in Ireland and Switzerland follow the same cycles. Trees in both regions exhibit an exceptional flowering year in 1984. Eicke's (1984) observations that flowering in 1984 in Germany, following a hot 1983 summer, was also good support the views that 1/ 1984 was a excellent flowering year in sycamore in the whole of western Europe and 2/ that a hot summer is necessary during the previous growing season. However the Irish data indicate that an oceanic climate is more favourable for the regular induction of flowering.

When one examines the yearly variation in flowering intensity of individual trees it appears that a warm summer induces all individuals in a population to flower synchronously. During the following years, as the environmental conditions (i.e. average summer temperatures) are less favourable for inflorescence initiation, some individuals start to flower out of phase. Variable flowering intensities are then observable until a hot summer occurs. Obviously a longer run of flowering data is necessary to confirm this. During years following `average' summers developmental constraints and resource allocation are likely to be factors strongly affecting flowering intensity.

Exceptional flowering years in sycamore appears to fit with Fenner's (1991) conclusion that in nearly all cases of good flowering years in trees, the event has been linked with the temperature at bud formation and either to unusually low or high temperature depending on the species. The control of flowering by climatic variables is also supported by the fact that intervals between mast years increase towards the edge of the range or with altitude (Fenner 1991), which has yet to be ascertained for sycamore. Here it is worth noting that sycamore flowers more profusely and regularly in the invaded range than in its native habitats. Le Mollard is situated in the heart of the region where sycamore dominated forests are the most common (Clot pers. comm.).

In beech, buds are initiated in the July of the summer prior to flowering and if temperature is high a proportion of the buds will produce reproductive structures (Fenner 1991). Holmsgaard & Olsen (1960) related seed crop size to this high temperature of the previous summer but also to the April rainfall of the seed year, as well as the size of the previous year's crop. Gaumann (1935 in Ellenberg 1988) estimated that a tree requires 2/3rd of the total carbohydrate assimilated in one year in order to produce a full crop of seeds. In sycamore the size of the seed production is likely to affect the next year's flowering intensity. This is suggested by trees which never or only occasionally set seed (they are male flowering or protandrous Mode C, see sex expression section for details) and flower very profusely every year. For instance tree P21 over a period of 11 years had a flowering intensity which varied only between 91% and 100%. Trees which set seed following every flowering event never flower with such an intensity.