Causes of masting and its extent in perennial plants

Production of huge numbers of seeds by woody plants anywhere is always a notable event, seldom more so than in the tropics where complete areas of bamboo suddenly come into flower, or entire forests are swamped with seeds, as in the dipterocarp forests of Malaysia. Discussion so far has centred around intermittent synchronous production of large seed crops in particular species, but a number of authors have examined masting records, some collected over long periods of time, for a wide variety of trees. The intention has been to test the validity of the concept, ascertain the factors causing it, and establish the range of seeding patterns present in perennial plants in general and trees in particular. It is also important to realize that the resources released in mast years have effects throughout the forest ecosystem.

There is an enormous literature on masting, but a good modern summary is provided by Kelly and Sork (2002), who set out the major hypotheses very clearly, provide 570 masting data sets worldwide and give coefficients of variation (CVs; estimates of how variable masting is over periods of time) that throw light on the relevance of the various models put forward. Masting is best developed in species whose year on year fruit production shows high CVs.

As Fig. 4.10 indicates, weather and resources are both clearly involved in mast seeding, though the extent of their involvement is less clear. It might be that plants simply respond to variable weather by flowering more in years when it is favourable. Alternatively there may be selective advantages to masting (ultimate factors) that modify plant responses to weather and internal resource levels (proximate factors) that enhance interannual variation in seed production. The oldest model for mast seeding is the resource matching hypothesis (RM) which states that in the absence of selection for or against masting, seed crops vary in response to environmental variation. If this theory is accurate, years with good weather should involve good growth and high seed production. Jack pine Pinus banksiana growing in Quebec fits this pattern with strong correlations across years among reproduction, growth and good growing conditions. The reproductive pattern of this tree, however, is unusual in that it is strongly serotinous (see Sections 3.7.1 and 4.6.2) and the lifetime seed crop remains on the tree until fire occurs. In such a case the exact year of seed production is irrelevant and resource matching is thus an appropriate strategy. In most species, however, this does not occur and there is evidence of 'switching' of resources from growth to reproduction in mast years giving a negative correlation between growth and reproduction (good years for reproduction are bad for growth), thus refuting the RM hypothesis. Such results have been widely reported for many trees including firs Abies, maples Acer, birches Betula, Dacrydium (Rimu from New Zealand), beeches Fagus, spruces Picea, pines Pinus and Douglas fir Pseudotsuga. Koenig and Knops (1998) reviewed 298 data sets for northern hemisphere conifers and found negative correlations between growth and reproduction to be widespread. The conclusion is that switching, and hence masting, is much more common than pure resource matching. So the next question is, what triggers masting?

Masting requires synchrony amongst plants; this is usually in response to a weather cue. In tropical dipterocarps it is triggered by night temperatures dropping 2 °C over 3 nights (Ashton et al., 1988), while in other species flowering may be triggered by drought or fire; none of these factors indicates an increase in resources. Masting is a reproductive strategy requiring a type of resource allocation strategy that exaggerates variation between years. If the plant has a physiological mechanism that alters flowering effort in relation to an environmental signal, why does it do so in this way? The prediction made by Janzen (1971) that in masting species selection should lead flowering to be hypersensitive to both weather variables and levels of nutrient reserves in the plant has been fully justified. It is now clear that proximate mechanisms integrating weather cues and resource utilization can produce more variable patterns of reproduction than expected from simple resource matching (Kelly and Sork, 2002). Why is this an advantage?

If mast seeding is to be selectively advantageous an economy of scale (EOS) is required (Janzen, 1978). The three EOS hypotheses with the most experimental testing are wind pollination, predator satiation and animal dispersal. These will be defined and reviewed after a description of three hypotheses that apply only in specialized situations.

(1) The environmental prediction hypothesis is that trees will reproduce strongly in years that will be favourable for seedling establishment. If this occurs the next question is, what is the cue that triggers masting? Unlike trees, which are poly-carpic (= iteroparous) and accordingly fruit many times, many species of tropical Asian bamboos are monocarpic (= semelparous); these grasses flower and fruit only once and then die. The length of time between germination and flowering ranges from 3-120 years, according to species, and synchronized flowering occurs not only in local populations but also in widely transplanted individuals of the same genetic stock. What can have favoured the evolutionary development of such a genetic pattern, so different from that of masting trees where masting is largely environmentally induced?

(2) The bamboo fire circle hypothesis postulates that the synchronized death of bamboos after masting encourages fire, which prevents trees from out-competing the bamboo whose seedlings are able to establish rapidly on open sites.

(3) The predator cleansing hypothesis points to the way in which the synchronized death of bamboos reduces the densities of herbivores feeding on leaves of adult plants.

The three major EOS hypotheses are:

(a) The pollination efficiency hypothesis that masting should be strongly selected in species that are able to reach greater pollination efficiency through synchronized above-average flowering effort. If this hypothesis is true the percentage fruit set should be higher when flower density is higher. Pollination is also likely to be more sensitive to flowering density in obligate outcrossers (must cross-pollinate) such as dioecious (male and female on separate plants) and self-incompatible species. Masting is more likely to occur or evolve in trees growing in unproductive habitats where reserves are expended in occasional large efforts, while trees in much more productive sites might well be able to make a large flowering effort year on year. Kelly and Sork (2002) conclude that wind pollination often provides an EOS but that animal pollination does not.

(b) The predator satiation hypothesis (Janzen, 1971) concludes that seed predators cause selection for masting when larger seed crops synchronized among individuals experience lower percentage seed predation, as indeed occurs in beech (Section 4.4.1). As long ago as 1942 Salisbury noted that in beech and oaks the only seeds escaping predation were those produced in mast years, and that if a species had a constant seed crop its natural enemies could increase their numbers to the point where all seeds were destroyed every year. The mobility of the seed predator species is an important factor in the equation. Birds and mammals can move very considerable distances whereas some invertebrates, including Cydia fagiglandana (Fig. 4.11) normally do not. The high CVs of predator-dispersed plants were consistent with the trees benefiting from predator satiation rather than dispersal.

(c) The animal dispersal hypothesis is that masting should be selected against in plants dispersed by frugivores (animals eating only a proportion of the seeds they cache) that are saturated by large fruit production, creating diseconomies of scale (Janzen, 1971; Silvertown, 1980). An oak, for example, dependent upon birds and rodents to carry the seeds away and cache them underground (see Box 4.1), could have many seeds go to waste if they overwhelm the dispersing animals. Indeed scatter-hoarding birds collected 89% of seeds from the canopy of singleleaf pinyon pine Pinus monophylla in Arizona in a low-seed year but only 43% in a high-seed year (Vander Wall, 1997). Herrera et. al. (1998) analysed 296 data sets describing annual variation in seed output by 144 species of woody plants. They concluded that seed production was slightly more variable amongst windpollinated trees than those pollinated by animals. Plants dispersed by frugivores (animals that consumed some seed directly but concealed many more, some of which gave rise to seedlings later) showed less variation in annual seed production than those that were dispersed by natural processes such as the wind, or by animals that were predominantly seed predators. Moreover, oaks may benefit by drawing animals such as squirrels from considerable distances as decribed in (b) above. In the case of the pinyon pine, the birds responsible, the Clark's nutcracker Nucifraga columbiana, cache such large numbers of seeds (see Box 4.1) that the needs of the tree to reproduce are undoubtedly met even in mast years when more than half the seeds go to waste. The extra seeds not cached may act as a signal to attract the birds to feed.

Which of these hypotheses are true for plants? Sork (1993) finds support for the pollination efficiency hypothesis for several wind-pollinated temperate trees (e.g. beech) but not for oaks. She suggests that in their case, predator satiation more adequately explains the selection pressures driving masting behaviour. It is important to recognize that the three main hypotheses are not mutually exclusive; elements of one or more may apply to any particular masting plant.

Behaviour comparable to that of the masting seen in trees is extremely rare in animals, though there are a few insect species whose reproduction is both synchronized and semelparous (i.e. they breed just once and then die). The six species of cicadas are an excellent example of this (Heliovarra et al., 1994). Synchronized iteroparity (repeated breeding) in which a long-lived animal will synchronously breed in some years but not in others, even given an adequate food supply, is very unusual indeed, though known in the cases of the kaka and kakapo Strigops habroptilus parrots. Both breed only in response to a masting food crop, regardless of supplementary artificial feeding, so perhaps they respond to the same environmental cue as the masting plants they are associated with.

Finally, Kelly and Sork (2002) note that the global pattern of masting shows the highest seed crop variability at mid latitudes and in the southern hemisphere, with a decrease towards the equator. Seed production is markedly affected by rainfall, whose pattern of annual variability is similar. Koenig and Knops (2000) examined 443 data sets showing patterns of annual seed production by northern hemisphere trees, finding that global patterns of annual rainfall and summer temperatures are generally similar to those exhibited by annual tree seed production but are more normally distributed. This, together with the inverse relationship between growth and reproductive effort shown by so many trees, supports the hypothesis that variability in annual seed production (masting) is indeed an evolved strategy.

Worm Farming

Worm Farming

Do You Want To Learn More About Green Living That Can Save You Money? Discover How To Create A Worm Farm From Scratch! Recycling has caught on with a more people as the years go by. Well, now theres another way to recycle that may seem unconventional at first, but it can save you money down the road.

Get My Free Ebook


Post a comment