Summary Reforestation requires a constant supply of high quality seed. Different methods have been used to measure seed and fruit production as a proportion of that which potentially could be produced. These estimates coupled with developmen-tal studies permit determination of when and to what extent different biological constraints reduce seed and fruit produc-tion. These vary among species, years and sites. Overcoming biological constraints may be possible in seed orchards and seed production areas. Biological constraints include: (1) peri-odic or inadequate floral initiation, (2) asynchronous develop-ment and flowering, (3) floral abortion, (4) ovule abortion, (5) embryo abortion, and (6) failure of seeds and fruits to mature and our inability to determine maturity. Solutions to these constraints are varied but all must begin with an under-standing of reproductive biology of each species.
Keywords: embryo, floral initiation, flowering, fruit produc-tion, maturaproduc-tion, ovule, reforestation.
Introduction
Reforestation requires a constant supply of high quality tree seed or vegetative propagules. However, seed is often in short supply, as a result of much seed being of low quality with variable maturity and a limited storage life. Seed production is variable among species and regions, but the causes of these variations are often uncertain. Even in seed orchards many constraints exist and mating is usually not random. The seed produced may represent only a portion of the genetic diversity intended by the phenotypic selection and orchard design (El-Kassaby 1992).
Seed production research must first identify species where there are biological constraints to seed or fruit production and then determine the biological constraints for each species in a particular location. Most seed production research has been done on temperate conifers (Owens and Blake 1985, Owens 1991). We know little about the reproductive biology and seed production of tropical species, particularly the tropical hard-woods about which there is now increased interest.
Identifying biological constraints
The reproductive phenology, from floral initiation through pollination and seed or fruit maturity, of temperate conifers is well known, but less is known about temperate hardwoods.
Even the basic phenologies of the reproductive cycles are poorly understood for tropical conifers and hardwoods. Most temperate and many tropical forest trees exhibit indirect flow-ering in which there is a period of dormancy between floral initiation and pollination. However, there are at least four other groups of tropical flowering periodicities that are recognized (Janzen 1978): (1) everflowering species, such as Ficus, that initiate flowers throughout the year and, therefore flowers, seeds and fruits at various stages of development are always present; (2) nonseasonal-flowering species that exhibit flow-ering periodicity among plants and from branch to branch; (3) gregarious-flowering species that initiate floral buds continu-ously, but these remain closed for weeks or months until environmental conditions are favorable for anthesis, some-times resulting in anthesis at the same time over a wide geo-graphical area, as in Coffea; and (4) seasonal-flowering species that flower in response to rainy or dry seasons or subtle seasonal variations in day length or temperature, as has been proposed for certain dipterocarps (Ashton et al. 1988).
Frequency of flowering (Table 1) (Owens et al. 1991a), which is usually measured by frequency of seed or fruit pro-duction at the end of the reproductive cycle, is variable. Seed production is measured as seed per hectare or per tree with little concern for the amount of seed produced relative to the potential seed that could be produced. More accurate, though more laborious methods are available. For example, in coni-fers, seed efficiency (SEF), or seed to ovule ratio (S/O), meas-ures the filled seed produced per cone as a percentage of the seed potential (number of fertile ovules per cone) (Owens et al. 1990, 1991a). In flowering plants, reproductive success (RS), which is a measure of the fruit to flower ratio (Fr/Fl) multiplied by the S/O ratio, has been used (Weins et al. 1987). Both of these approaches make it possible to rank species, clones or individual trees with respect to RS and identify quickly major constraints to seed or fruit production (Table 2).
However, use of SEF or RS does not give a complete picture of the constraints or when they occur in the reproductive cycle. Careful dissections or detailed anatomical techniques (Owens et al. 1991b) are needed to determine the times and possible causes of losses (Figure 1).
Biological constraints and solutions
Constraints and their relative importance vary among species, sites and years. Constraints result from developmental
irregu-Constraints to seed production: temperate and tropical forest trees
J. N. OWENS
Graduate Center for Forest Biology, University of Victoria, Victoria, British Columbia V8W 2Y2, Canada
Received May 31, 1994
larities causing abortion, or abnormal or delayed development of a reproductive structure. An anomaly may not become manifest until long after its cause has occurred. The longer the reproductive cycle, the more opportunity there is for climatic and other factors to affect seed and fruit production adversely (Owens 1991).
Failure to initiate floral structures
A common cause of low seed production is a failure to initiate cones or flowers. Seasonal flowering trees do not flower fre-quently. They have a particular developmental stage and sea-son at which floral initiation occurs that is mediated through the interaction of developmental, physiological and environ-mental events, many of which are poorly understood. In trees with distinct vegetative buds (Table 1), floral initiation usually only occurs in some of these buds long before floral structures are clearly visible. In many north-temperate conifers (Fig-ure 2) and a few hardwoods, the time and method of floral initiation have been determined by anatomical study (Owens and Blake 1985). In most cases, floral initiation and early development of floral structures occur before winter dor-mancy. In the few tropical forest trees that have been studied, Pinus caribaea Morelet. (Harrison and Slee 1991), Eucalyptus
(Moncur and Boland 1989) and some dipterocarps (Ng 1981, Ngamkhajorniway and Wasuwanich 1986, Ashton et al. 1988), as in temperate conifers and hardwoods, floral initiation ap-pears to be correlated with lateral shoot elongation and stages of vegetative bud development. Such correlations are useful because they allow us to predict the time of floral initiation based on external morphology.
Floral enhancement is usually achieved by manipulating the environmental factors known to promote flowering or by ap-plication of fertilizers or plant growth regulators (PGR). These methods have been quite successful in north-temperate coni-fers (Owens and Blake 1985, Ross and Pharis 1985) but have had little success in tropical hardwoods. Notable exceptions are some fruit trees and Eucalyptus (Griffin et al. 1993, Mon-cur 1994), where flowering has been promoted by paclobutra-zol. These flower enhancing methods may only be applicable to potted trees, seed production areas and seed orchards where trees are small and environmental conditions can be controlled. Environmental factors known to affect flowering include light intensity, temperature, water stress and nutrients. Photo-period may influence flowering indirectly through its control of shoot growth and bud development (Owens and Blake 1985), but it is unlikely to affect flowering in tropical forest trees. Increasing light is a common practice in north-temperate
Table 1. Bud types, age and frequency of flowering in selected hardwood and conifer trees.1
Family/ Common name Buds simple (S) Buds terminal (T) First good Intervals between Botanical name or mixed (M) or axillary (A) crops (years) seed crops (years)
Aceraceae
Acer platanoides Norway maple M T, A 25--30 1--3
Platanus occidentalis American sycamore M A 25--30 1--3
Betulaceae
Alnus glutinosa Common alder S A 15--20 2--3
Betula pendula Silver birch S T, A 15 1--3
Dipterocarpaceae
Dipterocarpus spp. M A 20--30 1--6
Fagaceae
Fagus sylvatica Common beech M A 50--60 5--15
Quercus petraea Sessile oak M A 40--50 2--5
Salicaceae
Populus tremuloides Quaking aspen S A 50--702 4--5
Salix nigra Black willow S A 25--752 1--3
Oleaceae
Fraxinus excelsior Common ash M A 25--30 3--5
Verbenaceae
Tectona grandis Teak M T 20--25 1--3
Cupressaceae
Chamaecyparis lawsoniana Lawson cypress S T 20--25 4--5
Thuja plicata Western red cedar S T 20--25 2--3
Pinaceae
Abies grandis Grand fir S A 40--45 3--5
Larix decidua European larch S T 25--30 3--5
Picea abies Norway spruce S T, A 30--35 3--5
Pinus contorta Lodgepole pine M A 15--20 2--3
Pinus caribaea Caribean pine M A 10--15 1--3
Pseudotsuga menziesii Douglas-fir S A 30--35 5--7
Tsuga heterophylla Western hemlock S A, T 25--30 3--4
seed production areas in which irradiation to the crown is increased by thinning, proper spacing and crown management. High solar irradiance also enhances flowering in branches of teak (Tectona grandis L.F.) (Nanda 1962). High temperature at the time of floral initiation may enhance flowering in temper-ate trees but timing is important (Ross 1989). Temperature-en-hanced flowering has been accomplished by tenting (see Owens and Blake 1985), proper location of seed orchards and moving potted trees into greenhouses (Philipson 1983, Ross 1988). However, this technique may be of limited value in tropical hardwoods. In contrast, low winter temperatures have enhanced flowering in Eucalyptus nitens (Deane & Maiden) Maiden grafts in conjunction with paclobutrazol application (Moncur and Hasan 1994). Water stress and root factors have also been used to enhance flowering in temperate hardwoods (Holmsgaard and Olson 1966) and conifers (Ross 1991). Water stress may be crudely controlled by proper drainage, site selec-tion and root pruning in soil-grown trees and by withholding water in potted trees. Results with tropical trees have not been reported. Nutrient availability may be manipulated through fertilizer application in seed production areas and seed or-chards (see Owens and Blake 1985).
In general, gibberellins (GA) promote flowering in conifers. Cones have been induced in several species of Cupressaceae and Taxodiaceae with GA3, often without adjunct treatments,
and in the Pinaceae, with the less polar GA4, GA7 and GA9, but
usually with a mixture of GA4/7 and adjunct treatments (see
Owens and Blake 1985). However, GAs often inhibit
flower-ing in hardwoods (Jackson and Sweet 1972, Ross and Pharis 1985, Pharis and Ross 1986, Singh and Kumar 1986). In E. nitens, application of paclobutrazol reduced the concentra-tion of endogenous GAs in apical tissue and enhanced flower-ing in trees maintained outside over winter (Moncur and Hasan 1994).
The timing of floral induction treatments is important and must precede reproductive bud determination (Figure 2). Be-cause biochemical changes begin before anatomical differ-ences are visible, accurate phenological studies that correlate floral initiation with vegetative shoot growth are essential so that floral induction or enhancement treatments can be applied precisely and cost effectively.
Pollination
Most temperate forest trees are wind pollinated, and the period of anthesis is short and cued by specific environmental events such as temperature, whereas for many tropical forest trees, distinct environmental cues are lacking. In tropical Agathis and Pinus, pollen shed and seed-cone receptivity are not well synchronized resulting in long periods of pollination with only a small proportion of cones shedding pollen and being recep-tive at one time. A pollen cloud, which is necessary for good seed set in many wind-pollinated species, is absent. This major constraint (Figure 1) may be overcome in seed orchards by supplemental mass pollination (SMP) (Webber 1991); how-ever, this technique cannot be applied in most seed production areas or in natural stands where trees are large.
Table 2. Reproductive success (RS) within Acacia auriculiformis A. Cunn. ex Benth. ×A. mangium hybrid.1
Spike Flowers per spike % Flower Fruits per Fruit per Ovules per Seed per Seed per RS2 survival spike flower ovary pod ovule
Prepollination Postpollination
1 100 1 1 1 0.01 16 12 0.75 0.0075
2 100 6 6 2 0.02 16 14 0.87 0.0174
3 95 9 9 3 0.03 16 5 0.31 0.0093
4 94 15 16 1 0.01 16 7 0.43 0.0043
5 78 14 18 3 0.03 16 8 0.50 0.0150
6 97 5 5 4 0.04 16 8 0.50 0.0200
7 98 21 21 2 0.02 16 7 0.44 0.0088
8 98 18 18 2 0.02 16 12 0.75 0.0150
9 100 5 5 5 0.05 16 11 0.69 0.0340
10 110 6 5 1 0.01 16 8 0.50 0.0045
11 96 6 6 8 0.08 16 8 0.50 0.0410
12 101 14 14 6 0.05 16 6 0.35 0.0175
13 107 19 18 2 0.02 16 14 0.35 0.0055
14 73 14 19 2 0.03 16 16 0.31 0.0083
15 105 12 11 1 0.01 16 13 0.81 0.0730
16 89 3 3 2 0.02 16 7 0.44 0.0080
17 113 17 15 6 0.05 16 4 0.25 0.0120
18 81 9 11 3 0.04 16 5 0.31 0.0110
19 92 4 4 10 0.11 16 10 0.62 0.0678
20 96 11 11 4 0.04 16 9 0.56 0.0224
Mean 96.7 10.2 10.8 3.3 0.034 16 9.2 0.58 0.0196
In tropical hardwoods that are mostly pollinated by insects, it is essential to know the phenology of the trees and pollinators (Bawa et al. 1985). Pollinators and host trees may be very specific and synchronized, or there may be a variety of polli-nators visiting a variety of plants over long periods of time. Because pollinators may rely on other plant species as nesting sites and for food at other times of the year, the weed-free understory of north-temperate seed orchards may be undesir-able in tropical hardwood seed orchards. Supplementing polli-nators, especially if they are colonial bees, is a viable alternative (Moncur 1993). In many tropical hardwood seed orchards, it may be advantageous to have widely spaced trees with open, managed crowns to enhance pollination and an understory of selected herbaceous or shrubby species to en-hance the pollinator population.
Incompatibility responses
It has generally been thought that incompatibility does not occur in conifers (Sedgley and Griffin 1989). Rather, postzy-gotic self-inviability has been the only proposed mechanism. Multiple fertilization takes place and simple polyembryony occurs, but most embryos resulting from selfing abort during early development, whereas embryos resulting from cross fer-tilization usually survive. In most conifers, if a high degree of selfing has occurred, there will be a significant reduction in seed production. However, we have found that, in Douglas-fir, incompatibility may occur through reduced pollen germina-tion and pollen tube growth or by prevengermina-tion of fertilizagermina-tion (Takaso and Owens 1994). The incompatibility mechanism is less precise than in angiosperms and results from secretions
Figure 1. The seed efficiencies and causes for seed loss in three conifers (adapted from Colangeli and Owens 1990, and Owens et al. 1990, 1991a).
from the integument, nucellus, megagametophyte or egg. These secretions may only decrease the possibility of fertiliza-tion from self and certain out-cross pollen rather than com-pletely blocking it. Therefore, incompatibility in conifers may only reduce seed production.
In hardwoods, incompatibility is genetically controlled by the gametophyte or sporophyte and, consequently, may occur at different times. Sporophytic incompatibility reactions result because the coating materials in the pollen exine are derived from the male sporophyte parent during pollen development. The interaction of the coating materials and the female sporo-phyte tissues may prevent pollen from adhering to the stigma, prevent pollen hydration or germination, or inhibit pollen tube penetration of the stigma (Figure 3). Gametophytic incompati-bility reactions result because the pollen intine, formed en-tirely by the gametophyte, reacts with the female sporophyte stigma, style or ovule tissues inhibiting pollen-tube growth (Sedgley and Griffin 1989). Any of these reactions may pre-vent fertilization by incompatible pollen thus reducing seed production, but they may also increase seed quality. These are prezygotic (prefertilization) events.
Selfing, which is an important cause of reduced seed pro-duction in most conifer seed orchards, can be reduced by SMP (Webber 1991) and proper seed orchard design and manage-ment. In tropical hardwood seed orchards, selfing resulting in flower, young fruit or ovule abortion may be reduced and seed production increased by proper spacing of trees and increasing the number of pollinators.
Floral abortion
Young pine seed cones that are inadequately pollinated are usually shed (cone drop) during the first year (Sweet 1973). In other conifers, unpollinated seed cones usually continue to mature, but they may be small and produce no filled seed (Owens et al. 1991a). In hardwoods, there are many more flowers than fruits produced. Floral abortion may occur at many stages but is most frequent at or shortly after pollination. In small-flowered hardwoods, abundant flowers serve as at-tractants for pollinators. Unpollinated flowers usually abscise immediately, greatly reducing the Fr/Fl ratio (Table 2). Incom-patibility may also result in flower abortion in a manner and time similar to unpollinated flowers. Adverse climatic condi-tions, commonly low temperatures in temperate regions and heavy rainfall and winds in tropical regions, and predators also cause flower loss.
Ovule abortion
The S/O ratio is extremely variable among species and years. In conifers, the number of ovules per cone may vary from a few in the Cupressaceae to several hundred in some Pinaceae and Araucariaceae. In temperate regions, early ovule abortion may result from low temperatures during spring pollination (Owens et al. 1990). Seed insects that lay eggs in conelets at pollination (Hedlin 1974) and diseases (Sutherland et al. 1987) destroy many ovules and young seeds. These losses can be minimized through good seed orchard management.
Most conifers have variable numbers of poorly developed ovules at the base and tip of cones. This is a developmental phenomenon with no apparent solution. Lack of pollination is the major cause of ovule abortion in most conifers (Figure 1). Unpollinated ovules often develop normally until the time of fertilization in many Pinaceae, Cupressaceae and Arau-cariaceae. Seed coats may partially develop and seeds, which may externally appear normal and filled, are actually empty or contain only the degenerated megagametophyte (Owens et al. 1990, 1991a). In Pinus, unpollinated ovules abort before fer-tilization leaving small, rudimentary ‘‘seeds’’ (Owens et al. 1982). It has been estimated that if more than 20% of ovules in a pine cone abort, the cone will abort (Sarvas 1962). In coni-fers, the major solution to ovule abortion is to ensure adequate cross-pollination through seed orchard design and manage-ment.
In tropical hardwoods, the number of ovules per flower varies (Tectona (Hedegart 1976) and Pterocarpus (Troup 1921) have four, Acacia has 16 (Table 2), and dipterocarps generally have six (Ashton 1988)), whereas seed per fruit is generally much less (Tectona and Pterocarpus have one to two, Acacia has two to 16 (Table 2), and dipterocarps generally have one (Ashton 1985)). In some temperate hardwoods such as species of Quercus and Betula (Mogensen 1975) in which the number of ovules is four and six, respectively, the number of surviving ovules appears to be fixed at one; however, in most species it is variable. Weins et al. (1987) concluded that the number of surviving ovules in herbaceous species is fixed in many inbreeders but random in many outcrossing species. This has not been tested in hardwoods. Ovule abortion decreases seed production, but it is also an important means of selection to ensure the survival of remaining ovules (Shaanker et al. 1988). Abortion has been interpreted as a trade-off between seed number and seed quality (Stephenson and Winsor 1986) that permits the parent plant to match seed production with available resources.
Ovule abortion in hardwoods is commonly caused by a lack of pollen, but other more subtle factors are also involved (Shaanker et al. 1988). For example, ovules closest to the source of nutrients have the best chance of survival. Biochemi-cally, the more fertilized ovules, the greater the plant growth regulator production, which may result in a stronger sink for nutrients resulting in higher ovule survival in that fruit (Sweet 1973). Developmentally, ovule abortion often results from lack of fertilization due to incompatibility reactions (Dumas and Knox 1983). Generally, the mechanisms by which ovule, em-bryo and seed ontogeny are controlled are poorly understood in forest trees. Only recently has gene regulation of these events begun to be studied in a few herbaceous angiosperms in which the development and genetics are well understood (Reiser and Fischer 1993, Thomas 1993). Such studies are not yet possible in forest trees.
Embryo, seed and fruit development and maturation
Despite numerous studies of embryogenesis in conifers (Singh 1978), there have been few studies designed specifically to determine the times and possible causes of embryo abortion
(Dogra 1967, 1984, Colangeli and Owens 1990, Owens et al. 1990, 1991a). Our knowledge of flowering plant embryogeny is much more complete in herbaceous species than in hard-woods (Wardlaw 1955, Davis 1966, West and Harada 1993). But as in conifers, there are still relatively few studies to determine times and causes of embryo abortion. We know almost nothing about tropical hardwood embryogeny. The interaction of the embryo and megagametophyte in conifers is complex and poorly understood. Embryos may abort followed by degeneration of the megagametophyte and vice versa (Co-langeli and Owens 1990). Polyembryony occurs in most coni-fers and increases the number of embryos per ovule and hence the possibility of one embryo surviving. Therefore, although early embryo abortion is a common occurrence in most ovules, seed production only decreases when all embryos abort. In contrast, true polyembryony is rare in hardwoods (Gifford and Foster 1989), therefore, any embryo abnormality is likely to cause seed loss. In conifers and hardwoods, the frequency of embryo abortion tends to decrease as embryos develop.
Our understanding of the molecular biology underlying the control of embryo development and subsequently abnormali-ties and abortion in herbaceous species is rapidly expanding (Goldberg et al. 1989, Thomas 1993). Seed proteins are the area of focus because they provide distinct markers of the DNA regulated embryogenic events. Results from these studies may provide insight into embryo development in forest trees.
Once embryos, seeds, cones and fruits are fully enlarged, most must undergo a period of maturation. Maturation may require certain developmental, physiological and biochemical changes before seeds or fruits can be harvested. For example, Tectona requires about 120 days from pollination until fruit maturity. Fruit growth and development occur for the first 50 days and fruit maturation occurs during the final 70 days (Hedegart 1976). However, it is difficult to distinguish imma-ture fruits from maimma-ture fruits. Easily distinguishable external morphological features, such as size, shape, texture and color of seeds, cones or fruits, must be correlated with internal embryo, endosperm, megagametophyte, seed coat or fruit wall ontogeny and water content. Increases in abscisic acid (ABA) occur as fruits and seeds mature. Although ABA content is a good indicator of maturity, it is not easily identified and meas-ured. The accumulation of storage products, most commonly proteins but in some seeds, carbohydrates or lipids, is corre-lated with increases in ABA. These can easily be identified by simple histochemical tests (Owens et al. 1991b) on tissue slices. Such studies can assist in developing field criteria for identifying mature seeds and fruits for harvest.
Conclusions
it could be more informative to determine reproductive success in relation to reproductive potential using fruit to flower and seed to ovule ratios. This information, coupled with more detailed studies of ontogeny, would allow constraints to be identified and solutions sought.
Acknowledgments
The research that formed the basis of most of this paper was supported by a research grant from the Natural Sciences and Engineering Re-search Council of Canada. Some unpublished information on tropical forest trees was provided by Mr. Prasert Sornsathapornkul and Mr. Suwan Tangmitcharoen, graduate students under Canadian Interna-tional Development Agency scholarships.
References
Ashton, P.S. 1988. Dipterocarp biology as a window to the under-standing of tropical forest structure. Annu. Rev. Ecol. Syst. 19:347--370.
Ashton, P.S., T.J. Givnish and S. Appanah. 1988. Staggered flowering in the Dipterocarpaceae: new insights into floral induction and the evolution of mast fruiting in the aseasonal tropics. Am. Nat. 132:44--66.
Bawa, K.S., D.R. Perry and J.H. Beach. 1985. Reproductive biology of tropical lowland rain forest trees. I. Sexual systems and incom-patibility mechanisms. Am. J. Bot. 72:331--345.
Colangeli, A.M. and J.N. Owens. 1990. Cone and seed development in a wind-pollinated, western hemlock (Tsuga heterophylla) clone bank. Can. J. For. Res. 20:1432--1437.
Davis, G.L. 1966. Systematic embryology of the angiosperms. Wiley, New York, 528 p.
Dogra, P.D. 1967. Seed sterility and disturbances in embryogeny in conifers with particular reference to seed testing and tree breeding in Pinaceae. Stud. For. Suec. 45:1--97.
Dogra, P.D. 1984. The embryology, breeding systems and seed steril-ity in Cupressaceae----a monograph. In Glimpses in Plant Research. Aspects of Reproductive Biology. Ed. P.K.K. Nair. Vikas Publishing House PVT Ltd., New Delhi, India, pp 1--127.
Dumas, C. and R.B. Knox. 1983. Callose and determination of pistil viability and incompatibility. Theor. Appl. Genet. 67:1--10. El-Kassaby, Y.A. 1992. Domestication and genetic diversity----should
we be concerned? For. Chron. 68:687--700.
Gifford, E.M. and A.S. Foster. 1989. Morphology and evolution of vascular plants, 3rd Edn. W.H. Freeman and Co., New York, 626 p. Goldberg, D.J., S.J. Barker and L. Parez-Grav. 1989. Regulation of
gene expression during plant embryogenesis. Cell 56:149--160. Griffin, A.R., P. Whiteman, T. Rudge, I.P. Burgess and M. Moncur.
1993. Effect of paclobutrazol on flower-bud production and vegeta-tive growth in two species of Eucalyptus. Can. J. For. Res. 23:640--647.
Harrison, D.L.S. and M.U. Slee. 1991. Gibberellin A4/7 enhanced
flowering in Pinus caribaea var. hondurensis. Can. J. For. Res. 21:788--793.
Hedegart, T. 1976. Breeding systems, variation and genetic improve-ment of teak (Tectona grandis L.f). In Tropical Trees. Eds. J. Burley and B.T. Styles. Academic Press, London, pp 109--123.
Hedlin, A.F. 1974. Cone and seed insects of British Columbia. Envi-ron. Can., Can. For. Serv. Inf. Rep. BC-X- 90, 61 p.
Holmsgaard, E. and H.C. Olsen. 1966. Experimental induction of flowering in beech. Forst. Forsogsraes. Dan. 30:3--17.
Jackson, D.I. and G.B. Sweet. 1972. Flower initiation in temperate woody plants. A review based largely on the literature of conifers and deciduous fruit trees. N.Z. For. Serv. Reprint No. 508, pp 9--24. Janzen, D.H. 1978. Seeding patterns of tropical trees. In Tropical Trees as Living Systems. Eds. P.B. Tomlinson and M.H. Zimmermann. Cambridge University Press, Cambridge, U.K., pp 83--128. Moncur, M.W. 1994. Flower induction and enhancement in tropical
species. In Proceedings: International Symposium on Genetic Con-servation and Production of Tropical Forest Tree Seed. Eds. M. Drysdale, S.E.T. John and A.C. Yap. Asean-Canada Forest Tree Seed Center, Chaingmai, Thailand.
Moncur, M.W. and D.J. Boland. 1989. Floral morphology of Eucalyp-tus melliodoro A. Cunn. ex. Schau. and comparison with other eucalypt species. Aust. J. Bot. 37:125--135.
Moncur, M.W. and O. Hasan. 1994. Floral induction in Eucalyptus nitens. Tree Physiol. 14:1303--1312.
Nanda, K.K. 1962. Some observations on growth, branching behav-iour and flowering of teak (Tectona grandis L.F.) in relation to light. Indian For. 88:207--218.
Ng, F.S.P. 1981. Vegetation and reproductive phenology of diptero-carps. Malays. For. 44:197--221.
Ngamkhajomawat, S. and P. Wasuwanich. 1986. Preliminary studies on flower development in Dipterocarpus alatus roxb. Embryon 2:56--61.
Owens, J.N. 1991. Flowering and seed set. In Physiology of Trees. Ed. A.S. Raghavendra. John Wiley and Sons, pp 247--271.
Owens, J.N. and M.D. Blake. 1985. Forest tree seed production. A review of literature and recommendations for future research. Envi-ron. Can., Can. For. Serv. Inf. Rep. PI-X-53, 161 p.
Owens, J.N., A.M. Colangeli and S.J. Morris. 1990. The effect of self, cross and no pollination on ovule, embryo, seed and cone develop-ment in western red cedar (Thuja plicata Donn). Can. J. For. Res. 20:66--75.
Owens, J.N., A.M. Colangeli, and S.J. Morris. 1991a. Factors affect-ing seed set in Douglas-fir (Pseudotsuga menziesii (Mirb) Franco). Can. J. Bot. 69:229--238.
Owens, J.N., S.J. Simpson and M. Molder. 1982. Sexual reproduction of Pinus contorta. II. Postdormancy ovule, embryo, and seed devel-opment. Can. J. Bot. 60:2071--2083.
Owens, J.N., P. Sornsathapornkul and S. Tangmitchareon. 1991b. Manual: studying flowering and seed ontogeny in tropical forest trees. ASEAN-Canada Forest Tree Seed Center Project, Muak-Lek, Saraburi, Thailand, 134 p.
Pharis, R.P. and S.D. Ross. 1986. Pinaceae. Hormonal promotion of flowering. In Handbook of Flowering, Vol. 5. Ed. A. Halevy. C.R.C. Press, Boca Raton, FL, pp 2--41.
Philipson, J.J. 1983. The role of gibberellin A4/7, heat and drought in
the induction of flowering in Sitka spruce. J. Exp. Bot. 34:291--302. Reiser, L. and R.L. Fischer. 1993. The ovule and the embryo sac. Plant
Cell 5:1291--1301.
Ross, S.D. 1988. Effects of temperature, drought, gibberellin A4/7, and
their timing on flowering in potted Picea engelmannii and P. glauca
grafts. Can. J. For. Res. 18:163--171.
Ross, S.D. 1989. Temperature influences on reproductive processes in conifers. In Climate Applications in Forest Renewal and Forest Production. Eds. D.C. MacIver, R.B. Street and A.N. Auclair. Can. Gov. Print. Center, Ottawa, pp 40--43.
Ross, S.D. and R.P. Pharis. 1985. Flower induction in crop trees: different mechanisms and techniques with special reference to conifers. In Attributes of Trees as Crop Plants. Eds. M.G.R. Can-nell, J.E. Jackson and J.C. Gordon. Institute of Terrestrial Ecology, Monkswood Experimental Station, Abbots Ripton, Huntingdon, U.K., pp 383--397.
Sarvas, R. 1962. Investigations on the flowering and seed crop of
Pinus sylvestris. Commun. Inst. For. Fenn. 53:1--198.
Sedgley, M. and A.R. Griffin. 1989. Sexual reproduction of tree crops. Academic Press Ltd., London, 378 p.
Shaanker, R.V., K.N. Ganeshaiah and K.S. Bawa. 1988. Parent--off-spring conflict, sibling rivalry, and brood size patterns in plants. Annu. Rev. Ecol. Syst. 19:177--205.
Singh, K. and S. Kumar. 1986. Manipulation of flowering in wood plants by the use of gibberellin. Indian J. For. 9:331--336. Stephenson, A.G. and J.A. Winsor. 1986. Lotus corniculatus regulates
offspring quality through selective fruit abortion. Evolution 40:453--458.
Sutherland, J.R., T. Miller and R.S. Quinard. 1987. Cone and seed diseases of North American conifers. North American Forestry Commission Publication No. 1, Victoria, B.C., Canada, 77 p.
Sweet, G.B. 1973. Shedding of reproductive structures in forest trees.
In Shedding of Plant Parts. Ed. T.T. Kozlowski. Academic Press, New York, pp 341--382.
Takaso, T. and J.N. Owens. 1994. Effects of ovular secretions on pollen in Pseudotsuga menziesii (Pinaceae). Am. J. Bot. 81:504--513.
Thomas, T.L. 1993. Gene expression during plant embryogenesis and germination: an overview. Plant Cell 5:1401--1410.
Troup, R.S. 1921. The silviculture of Indian trees, Vol. 1. Clarendon Press, Oxford, U.K., pp 287--292.
Wardlaw, C.W. 1955. Embryogenesis in plants. Methuen, London, 381 p.
Webber, J.E. 1991. Interior spruce pollen management manual. Land Management Report. No. 70., B.C. Ministry of Forests, 25 p. West, M.A.L. and J.J. Harada. 1993. Embryogenesis in higher plants:
an overview. Plant Cell 5:1361--1369.
