[colloquial, first use ca. 1703 by John Ray]
Gymnosperms (from the Greek, γυμνόσπερμος, meaning "naked seed" because the seeds do not develop within an ovary).
The following table describes the relative hierarchy of gymnosperm taxonomy and provides links to descriptions of each order and family.
This database provides information on taxa belonging to the gymnosperms. The gymnosperms have long been recognized as a distinctive group of plants, but there has also been a long (and ongoing) debate about whether the principal groups within the gymnosperms share a common ancestor, and if so, whether that common ancestor is or is not shared with the flowering plants, also called angiosperms. If they do share a common ancestor, distinct from the flowering plants, then the gymnosperms are a unique lineage of plants. Otherwise, the term "gymnosperm" has no taxonomic significance. In any event, the term "gymnosperm" has practical significance because it refers to a relatively small and highly distinctive assemblage of plants that are for the most part very ancient in lineage and reduced to a largely relictual distribution, but which nonetheless have great cultural and ecological importance at a global scale—a significance far greater than their species diversity alone would imply.
The principal point of contention in the taxonomic debate is whether the gnetophytes Ephedra, Gnetum and Welwitschia share a common ancestor with the angiosperms or the conifers (which are usually, but not always, assumed to share a common ancestor with the cycads and Ginkgo). The former opinion is called the "anthophyte" hypothesis, where anthophytes are defined to be a clade including both angiosperms and gnetophytes. The latter opinion is the gnetifer (or the variant 'gnepine,' gnetophytes plus Pinaceae) hypothesis, where gnetifers are a clade including both gnetophytes and the conifers. A succinct online review of the subject is provided by Castillo (2001), and cladograms clarifying the anthophyte and gnetifer hypotheses are provided by Burleigh and Mathews (2004). Every year or so a new paper is published supporting one hypothesis or another, and the accumulation of molecular data seems to be swinging in favor of the gnetifer or gnepine hypothesis, but there are still recent morphology studies supporting the anthophyte hypothesis (e.g., Friis et al. 2007). Christenhusz et al. (2011), reviewing studies available to date, simply assign each of the four gymnosperm groups to a subclass under the seed plants, Equisetopsida. They also note that under the gnepine hypothesis it would be appropriate to put the gnetophytes and Pinaceae in one subclass, and all other confers in another subclass. However, few taxonomists would be so rash as to assert that an end is in sight. It appears likely that Darwin's "abominable mystery", the origin of the angiosperms, will remain a mystery for a few more years. Until it is resolved, we won't quite know what to make of the gnetophytes.
I largely adopt, for taxa at the rank of family and above, the classification proposed by Christenhusz et al. (2011). It is still a conservative view, for instance in not assuming validity of the gnepine hypothesis, but is well supported by both molecular and morphological evidence. I should note that much different classification schemes exist. For instance, Reveal (1998) subdivided the gymnosperms into 4 divisions, 6 classes, and 12 orders. I suppose he has reconsidered, since he is a co-author on the Christenhusz paper.
Although molecular studies have done much to improve our understanding of evolutionary relationships between extant gymnosperms, changing relationships at evolutionary timescales are best elucidated through the fossil record. The following points summarize the chronology of gymnosperm evolution:
- Late Devonian (385-359 million years ago [my]): Earliest seed plants arise (Hill 2005).
- Middle Pennsylvanian (310 my): Earliest conifers, cycads, and Ginkgo arose at about the same time, perhaps from a precursor in the Cordaitales, a plant that bore seeds in conelike structures. These earliest gymnosperms were not referrable to any families now existing. The Cordaitales persisted until the mass extinction at the end of the Triassic (201 my) (Hilton 2014).
- Permian (299-251 my): First appearance of Ginkgoales. The group is most diverse and abundant during the Jurassic (Thomas and Spicer 1986 in Hill 2005).
- Lower Permian (299-271 my): More recognizable cycads appear, e.g. Crossozamia (Gao and Thomas 1989); cycad diversity and abundance peaked during the Mesozoic (Hill 2005).
- Middle Permian (271-260 my): First evidence of Gnetophytes, limited to pollen evidence (Crane 1988); macrofossils attributable to Gnetophytes appear only much later (Cretaceous).
- Upper Triassic (235-202 my): First appearance of the Bennettiales. This was one of the most abundant and diverse gymnosperm groups in the Mesozoic, although it became extinct (along with the dinosaurs) at the end of the Cretaceous. Evolutionary relationships are unclear but the Bennettiales may have been most closely related to the Gnetales (Hill 2005).
- Triassic, earliest Norian (228-225 my): The first appearance of Podocarpaceae, the extinct species Rissikia media (Townrow 1967 in Rothwell et al. 2012).
- Lower Jurassic (202-176 my): First appearance of Ginkgo, the oldest extant conifer genus (Tralau 1968)
- Jurassic, Sinemurian (197-190 my): First record of Cupressaceae, the extinct species Austrohamia minuta (Escapa et al. 2008 in Rothwell et al. 2012).
- Jurassic, Pliensbachian (190-183 my): First record of Taxaceae, the extinct species Paleotaxus rediviva (Nathorst 1908 in Rothwell et al. 2012).
- Jurassic, Pliensbachian (190-183 my): First record of Araucariaceae, the extant genus Araucaria (Arrondo and Petriella 1980 in Rothwell et al. 2012).
- Jurassic, Kimmeridgian (156-151 my): First record of Pinaceae, the extinct species Eathiestrobus mackenziei (Rothwell et al. 2012).
- Lower Cretaceous (130 my): First appearance of Pinus, the extinct species Pinus yorkshirensis from a fossil cone preserved in the Wealden Formation in Yorkshire, UK (Ryberg et al. 2012).
- Cretaceous, Aptian (125-112 my): First Gnetales macrofossils, in the form of Drewria potomacensis, an extinct species with leaves resembling those of Gnetum (Crane and Upchurch 1987 in Hill 2005).
- Cretaceous, Conacian (89-86 my): First record of Sciadopityaceae, the extant genus Sciadopitys (Ohsawa et al. 1991 in Rothwell et al. 2012).
Although morphologically, ecologically and taxonomically diverse, and individually very numerous, the gymnosperms are not represented by a great many species. Here is a table summarizing diversity at the family, genus and species level. In this treatment there are 947 species (as of 2002.07.07), or about as many as may be found in the largest genera of flowering plants (such as Acacia). Of this total, one species represents the Ginkgos, 68 represent the Gnetophytes, 289 are Cycads, and the remaining 589 species are Conifers. At this time, those taxon figures are debatable within ±5-10%. There are not likely to be many more species discovered. There have lately been notable discoveries (Cupressus vietnamensis, discovered in 2001 in Vietnam), and the affinity of certain cycads, some podocarps and many gnetophytes for tropical rainforest environments means that new discoveries may be made in those taxa as well.
As a practical matter, this database generally treats species as valid if there is decent ground for debate as to whether a species is distinct, or should be treated as a subspecies or variety. Such debates are common because, for many taxa, up-to-date taxonomic treatments are simply not available. Where I can find such a treatment, then I tend to give it considerable credence; but there are some taxa that have not received monographic treatment for many decades.
I am also sometimes questioned about the distinction between subspecies and varieties. Generally I prefer to discriminate subspecies if they represent spatially and ecologically disjunct groups that show minimal evidence of recent introgression or other genetic exchange. I prefer to use the term variety to represent morphologically distinctgroups that may be local in extent, share a common range or ecological setting, or display evidence of active gene flow between the taxa. Of course there are many intermediate conditions and there is still an element of subjectivity. Nonetheless I feel it is better to have scientific criteria in designation of varieties and subspecies; in the past, the two seem to have been discriminated mainly on the basis of personal preferences.
Plants are classified into two large groups, vascular plants and nonvascular plants. Nonvascular plants lack tissues that have been specialized for the transport of water and nutrients between distant parts of the plant. They include all algae, mosses, and their allies. Vascular plants have such transport systems. They include the vast majority of familiar terrestrial plants and can be divided into three major groups according to major features of their reproductive systems. The simplest and oldest of these is the Pterophytina or ferns and fern allies. The next group is the Pinophyta or Gymnospermae, which are the focus of this presentation. The last and by far the largest group (about 250,000 described species), is the Angiospermae.
Gymnosperms are woody plants, either shrubs, trees, or, rarely, vines (some gnetophytes). They differ from flowering plants in that the seeds are not enclosed in an ovary but are exposed within any of a variety of structures, the most familiar being cones. The gymnosperms display four architectural models of tree growth, which determine the basic form and appearance of the plant; see Tomlinson (1983) for more detail.
A more anatomically detailed discussion of the comparative morphology of gymnosperms and related vascular plants can be found in the On-Line Biology Book or by clicking on Google Search (Gymnosperm Life Cycle). It seems odd, but apart from that naked seed, the gymnosperms really do not share many attributes. Wood and leaf, habit and habitat, physiology and anatomy - all are quite variable between the four major groups. We might have chosen to put "the gymnosperms" together in one group because they are all relicts, survivors of an ancient flora. They all seem strange to children of the Cenozoic savannas.
Distribution and Ecology
Living gymnosperms are distributed worldwide (excepting Antarctica), with a majority, particularly the conifers, in temperate and subarctic regions.
Gymnosperms include many of the largest trees on earth. Ginkgo can grow very large. Among the conifers, Sequoiadendron giganteum is the largest of all trees, but many other giants exist as well; see the sidebar for A Tale of Big Tree Hunting In California.
The age of a gymnosperm is usually difficult to determine. Precise numbers are only available if someone recorded the date of planting. Many trees have (approximately) annual rings, permitting approximate determination of the age of an individual, but in harsh environments many years may pass before the first ring is made. For example, Ken Lertzman (pers. comm. 1990) found mountain hemlock seedlings up to 18 years old (based on bud scar counts) that had yet to form their first ring. The greatest age known from ring counts is about 5000 years, for Pinus longaeva. Several other species, such as Thuja plicata and Taxus baccata, might achieve comparable ages but we cannot tell because they live in moist climates where the tree's woody heart rots away while the rest of the tree is in the prime of life, so that no record of the rings is left behind. For further detail on the oldest known gymnosperms, see the article How Old Is That Tree?.
Many other trees, including most tropical species, and all non-tree gymnosperms (Cycadales and Gnetales), do not have regular annual rings. Age estimates for such plants are usually based on extrapolation of observed growth rates. Such extrapolation may lead to estimation errors of several hundred percent (usually overestimation, because trees grow more slowly as they age).
Many species of plants are clonal, so that numerous different individuals are genetically identical. Among the oldest, a clone age of 43,600 years has been estimated for Lomatia tasmanica, a Tasmanian tree (DEWHA 2010). A clone age estimate of 10,500 years has been proposed for Huon pine, Lagarostrobos franklinii, and a comparable age may be reached in the redwood, Sequoia sempervirens, which is only one of many gymnosperms that can sprout from fallen stems or cut stumps. Some cycads produce clones by producing buds at the base, which may become separated from the main plant and then root. The oldest gymnosperm clones are likely to be found among species prone to vegetative reproduction in an area of extraordinarily stable climate, such as cycads of the genus Encephalartos growing in South Africa.
Dendrochronology, the study of tree rings, is necessarily restricted to trees that form annual rings. This includes hundreds of species. Nearly all gymnosperms living in temperate or boreal climates have been investigated for use in dendrochronology, and most have proven suitable, with the exceptions found mainly in areas where extreme stress from cold, drought or other factors frequently prevents the formation of an annual ring. Tropical gymnosperms have been less widely studied, but generally, those growing in climates with pronounced wet and dry seasons form annual rings and those with year-round rainfall are not usable.
There are few gymnosperms that were not used by aboriginal peoples; the main exceptions seem to have been in areas where there were few or no people, or in tropical areas where angiosperm diversity far exceeded gymnosperm diversity. Two outstanding examples include:
- Cycads (Cycadales), Bunya pines (Araucaria bidwillii) and stone pines (Pinus subsection Cembroides), all of which produced edible nuts that were a staple food for local cultures.
- Western redcedar (Thuja plicata), which filled virtually all needs of Pacific Northwest Coast peoples except providing food.
Today, of course, gymnosperms provide the world with softwood lumber and with most of the wood pulp supply. Most of this timber comes from a handful of species, notably Pinus radiata, the most widely planted tree in the world, although it has one of the smallest natural distributions.
Gymnosperm diversity is best observed at botanical gardens and arboreta. Conifers, at least, are also very well represented in ornamental collections throughout the temperate zones of the world.
All trees, indeed all vascular plants, possess a fundamental architecture prescribed by their genetic code. This architecture remains constant throughout the plant's life, providing a uniform functional structure even as the plant experiences immense changes in size, from seed to tree, and even in spite of calamities that may befall the plant during its life. Every tree can be defined by one of 23 different architectural models. This is a complex topic, not yet fully explored, and I have not yet found a source that classifies the gymnosperms according to their various architectural models; for further information see this excellent blog and the additional sources cited therein.
Arrondo, O. G., and B. Petriella. 1980. Alicurá, nueva localidad plantífera Liásica, de la provincia de Neuquén, Argentina. Ameghiniana 17:200-215.
Burleigh, J. G. and S. Mathews. 2004. Phylogenetic signal in nucleotide data from seed plants: implications for resolving the seed plant tree of life. American Journal of Botany 91: 1599-1613. Available online, accessed 2007.11.26.
Castillo, G. R. H. 2001. Seed plant phylogeny and the anthophyte hypothesis. http://www.biology.ualberta.ca/courses.hp/biol606/OldLecs/Lecture2001.03.HC.html, accessed 2007.11.26.
Christenhusz, M. J. M., J. L. Reveal, A. Farjon, M. F. Gardner, R. R. Mill, and M. W. Chase. 2011. A new classification and linear sequence of extant gymnosperms. Phytotaxa 19:55-70.
Crane, P. R. 1988. Major clades and relationships in "higher" gymnosperms. Pp. 218-272 in C. B. Beck (ed.), Origin and Evolution of Gymnosperms. New York: Columbia University Press.
Crane, P. R. and G. R. Upchurch Jr. 1987. Drewria potomacensis gen. et sp. nov., an Early Cretaceous member of Gnetales from the Potomac Group of Virginia. American Journal of Botany 74:1723-1738.
[DEWHA] Department of the Environment, Water, Heritage and the Arts (2010). Lomatia tasmanica in Species Profile and Threats Database, Department of the Environment, Water, Heritage and the Arts, Canberra. www.environment.gov.au/sprat, accessed 2010.03.29.
Escapa, I. , R. Cúneo, and B. Axsmith. 2008. A new genus of the Cupressaceae (sensu lato) from the Jurassic of Patagonia: Implications for conifer megasporangiate cone homologies. Review of Palaeobotany and Palynology 151:110-122.
Friis, E.M., P.R. Crane, K.R. Pedersen, S. Bengtson, P.C.J. Donoghue, G.W. Grimm, and M. Stampanoni. 2007. Phase-contrast X-ray microtomography links Cretaceous seeds with Gnetales and Bennettitales. Nature 450: 549-553.
Gao, Z. and B. A. Thomas. 1989. A review of fossil cycad megasporophylls, with new evidence of Crossozamia pomel and its associated leaves from the lower Permian of Taiyuan, China. Review of Palaeobotany and Palynology 60(3):205-223.
Hill, Ken. 2005. Diversity and evolution of gymnosperms. Pp. 25-44 in R. J. Henry, Plant Diversity and Evolution. Cambridge, MA: CABI Publishing.
Hilton, J. 2014. The origin of modern conifer families. www.birmingham.ac.uk/research/activity/geosystems/projects/conifer-families/index.aspx, accessed 2014.12.06.
Nathorst, A. G. 1908. Paläobotanische Mitteilungen, 7: Über Palissya, Stachyotaxus und Palaeotaxus. Kungliga Svenska Vetenskapsakademiens Handlingar 43:3-37.
Ohsawa, T., M. Nishida, and H. Nishida. 1991. Structure and affinities of the petrified plants from the Cretaceous of northern Japan and Saghalien IX. A petrified cone of Sciadopitys from the Upper Cretaceous of Hokkaido. Journal of Phytogeography and Taxonomy 39:97-105.
Rothwell, G. W., G. Mapes, R. A. Stockey, and J. Hilton. 2012. The seed cone Eathiestrobus gen. nov.: fossil evidence for a Jurassic origin of Pinaceae. American Journal of Botany 99(4):708-720.
Ryberg, P. E., G. W. Rothwell, R. A. Stockey, J. Hilton, G. Mapes, and J. B. Riding. 2012. Reconsidering relationships among stem and crown group Pinaceae: oldest record of the genus Pinus from the Early Cretaceous of Yorkshire, United Kingdom. International Journal of Plant Sciences 173(8):917-932.
Thomas, B. A., and Spicer, R. A. 1986. The Evolution and Palaeobiology of Land Plants. London: Croom Helm. ix, 309p.
Tomlinson, P. B. 1983. Tree Architecture: New approaches help to define the elusive biological property of tree form. American Scientist 71(2):141-149. Available http://harvardforest.fas.harvard.edu/sites/harvardforest.fas.harvard.edu/files/publications/pdfs/Tomlinson_AmScientist_1983.pdf, accessed 2017.08.12.
Townrow, J. A. 1967. On Rissikia and Mataia - Podocarpaceous conifers from the lower Mesozoic of southern lands. Papers and Proceedings of the Royal Society of Tasmania 101:103-136.
Tralau, H. 1968. Evolutionary trends in the genus Ginkgo. Lethaia 1:63–101.
Chaw, S.M., C.L. Parkinson, Y. Cheng, T.M. Vincent, and J.D. Palmer. 2000. Seed plant phylogeny inferred frrom all three plant genomes: monophyly of extant gymnosperms and origin of Gnetales from conifers. Proc Natl Acad Sci USA 97:4086-4091. Available online HERE (2007.11.26).
The University of Sydney School of Biological Sciences provides a good overview of the gymnosperm life cycle (2006.03.20).
Wang, Xiao-Quan, and Jin-Hua Ran. 2014. Evolution and biogeography of gymnosperms. Molecular Phylogenetics and Evolution 75:24-40.
White 1994 gives a beautifully illustrated account of the fossil origins of the major gymnosperm and angiosperm groups.
Last Modified 2018-01-13
Double fertilization evolved in the Gnetales and in the angiosperms, but it differs in the two groups of plants. This phenomenon occurs in the extant Gnetales Ephedra(Friedman, 1990, 1992, 1994, 1998Friedman, 1990Friedman, 1992Friedman, 1994Friedman, 1998) and Gnetum (Carmichael and Friedman, 1996). The product of the second fertilization in Ephedra and Gnetum is a diploid (supernumerary) embryo; thus, two diploid genetically identical embryos are formed in the female gametophyte after the two sperm nuclei are released (Friedman, 1998). In Ephedra, one sperm nucleus unites with the egg and the other with the ventral canal nucleus (Friedman, 1990, 1992Friedman, 1990Friedman, 1992). An egg is not formed in Gnetum; thus, the sperm nuclei unite with two free haploid nuclei at the micropylar end of the female gametophyte (Carmichael and Friedman, 1996). Only one of the embryos survives to seed maturity in Ephedra and Gnetum (Friedman and Carmichael, 1996). The female gametophyte supplies nourishment to the developing embryo in both Ephedra and Gnetum, but development of the female gametophyte is not completed in Gnetum until after fertilization has occurred (Carmichael and Friedman, 1996).
The product of the second fertilization in angiosperms is a non-embryo polyploid tissue called endosperm, which supplies food to the developing embryo. Development of endosperm and reduction of the female gametophyte to form the embryo sac occurred in angiosperms after their divergence from gymnosperms (Friedman, 1992). Thus, in development of the angiosperm seed there has been “a tendency to shift the dependency of the early embryo directly to the parent sporophyte and away from an intermediate gametophyte generation or some modified form of that generation” (Steeves, 1983).
Numerous attempts have been made to find fossils of angiosperms that date from the Jurassic and even the Triassic. Although many intriguing plant fossils have been found in rocks of these periods, none has been judged to be unquestionably the remains of an angiosperm (Hughes, 1976; Tiffney, 1984). However, some Triassic and Jurassic fossils have characteristics of both gymnosperms (mostly) and angiosperms (Stewart and Rothwell, 1993). Thus, true angiosperms are not found in the fossil record until the Early Cretaceous (Hickey and Doyle 1977; Doyle, 1978; Tiffney, 1984; Doyle and Donoghue, 1987; Sun et al., 2002, 2011Sun et al., 2002Sun et al., 2011), and the oldest unambiguous angiosperm fossils (mostly pollen) are 140–130 Mya (Soltis et al., 2005).
Fossil fruits and seeds from the Lower Cretaceous said to be angiosperms include those of Onoana, Nyssidium and Kenella (Hughes, 1976). However, in a summary of Cretaceous seed and fruit fossils with presumed angiosperm affinities, Tiffney (1984) concluded that there was insufficient evidence for members of these three genera to be called angiosperms. Further, there also seemed to be questions about whether or not Aptian Stage fossils of Carpolithus, Onoana and Prototrapa and Albian fossils of Araliaecarpum, Caricopsis and Carpolithus were angiosperms (Tiffney, 1984). Fossil “narrow pseudo-syncarpous carpels” of Leefructus with affinities to the Ranunculaceae are from the latest Barremian and earliest Aptian (Sun et al., 2011). Fossils of Caspiocarpus paniculiger (dehiscent follicle and seed) and Ranunculaecarpus quinquiecarpellatus (dehiscent follicle) are from Albian deposits in Kazakhstan and near the Kolyma River in far eastern Siberia (Russia), respectively (Tiffney, 1984). A number of genera, including Carpites, Laurus, Platanus and “Salix,” have been collected in Cenomanian deposits (Tiffney, 1984).
Fossil angiosperm seeds from the Albian and Cenomanian stages are small, e.g., Ranunculaecarpus quinquiecarpellatus, 1.5 mm long×0.6 mm wide and Carpites liriophylli, 1.4 mm long×0.6 mm wide (Tiffney, 1984). Further, these small seeds were produced in follicles or capsules (Tiffney, 1984, 1986Tiffney, 1984Tiffney, 1986). In their review of angiosperm radiation in the Cretaceous, Wing and Boucher (1998) concluded that diversification of angiosperm families was much faster in the second than in the first part of the Cretaceous. In fact, fossils of many extant families have now been found in Cretaceous-aged rocks. Fruits and/or seeds (or parts of them) are among the many kinds of fossil plant parts found in the Cretaceous of various extant families, including Amaranthaceae, Aquifoliaceae (Collinson et al., 1993), Ceratophyllaceae (Dilcher, 1989), Fagaceae? (Herendeen et al., 1995), Juglandaceae (Hermanova et al., 2011), Hamamelidaceae, Lythraceae (Estrada-Ruiz et al., 2009), Menispermaceae (Collinson et al., 1993), Musaceae (Rodriguez-de la Rosa and Cevallos-Ferriz, 1994), Phytolaccaceae (Cevallos-Ferriz et al., 2008), Sabinaceae, Sapindaceae (Collinson et al., 1993) and Trimeniaceae (Yamada et al., 2008).
In younger rocks/sediments, seeds/fruits of many other extant plant families have been found. Some examples are: Late Paleocene: 10 families from Sentinel Butte Formation in North Dakota (USA) (Crane et al., 1990), Icacinaceae (Pigg et al., 2008), Nymphaeaceae (Taylor et al., 2006), Ranunculaceae (Pigg and DeVore, 2005); Paleocene/Eocene boundary: Lythraceae, Nyssaceae, Vitaceae (Fairon-Demaret and Smith, 2002); Early Eocene: Annonaceae, Boraginaceae, Caprifoliaceae, Euphorbiaceae, Flacourtiaceae, Icacinaceae, Magnoliaceae, Menispermaceae, Nymphaeaceae, Rosaceae, Theaceae and Vitaceae (Chandler, 1964); Middle Eocene: Araceae (Smith and Stockey, 2003), Cornaceae (Stockey et al., 1998), Fagaceae (Mindell et al., 2009), Lauraceae (Little et al., 2009), Nymphaeaceae (Cevallos-Ferriz and Stockey, 1989) and Salicaceae (Manchester et al., 2006); 34 different families have been found in the Middle Eocene Clarno Nut Beds in Oregon (USA) (Manchester, 1994); Oligocene-Miocene: Alismatales (Estrada-Ruiz and Cevallos-Ferriz, 2007); Early Miocene: Sargentodoxaceae (Tiffney, 1993; Traverse, 1994); and Middle Miocene: Alismataceae (Haggard and Tiffney, 1997), Anacardiaceae, Annonaceae, Chrysobalanaceae (Tiffney et al., 1994), Fagaceae (Borgardt and Pigg, 1999) and 42 families from Denmark (Friis, 1985).
In the major radiation of angiosperms that occurred in the Late Cretaceous and Early Tertiary, many of the modern families and genera first appeared in the fossil record (Tiffney, 1981, 1986Tiffney, 1981Tiffney, 1986; see Wing and Boucher, 1998). Many, but not all, of these new genera had large (up to 50,000–100,000 mm3) seeds (Tiffney, 1986). Haig and Westoby (1991) note that for both fossil and extant species, the smallest gymnosperm seeds are larger than the smallest angiosperm seeds. These authors suggest that time of fertilization in angiosperms is more efficient with regard to allocation of resources than it is in gymnosperms. In gymnosperms, a female gametophyte that contains food reserves for the embryo is produced prior to fertilization, while in angiosperms the food-supplying tissue (endosperm) is not produced until after fertilization. Thus, if fertilization does not occur fewer resources are lost via ovule abortion in angiosperms than in gymnosperms. Seeds of angiosperms can be smaller than those of gymnosperms because the costs of pollination are reduced substantially in angiosperms (Haig and Westoby, 1991). However, there is a limit to how small seeds can be, and this may be determined partly by “accessory costs” (i.e., the costs of pollen capture and ovules that abort). With a decrease in seed size, accessory costs increase, and consequently allocation of food reserves to the developing embryo decreases. Minimum seed size is the point at which any further decrease in resources allocated to the embryo would reduce chances of seedling survival (Haig and Westoby, 1991).