Introduction

Seed plants show a diversity of adaptations for seed dispersal. These adaptations may involve physical dispersal agents (transport by wind, water, gravity, etc.) or biotic dispersal agents (self-dispersal, or more commonly transport by animals), often combined in various ways, simultaneously, sequentially or as alternatives. The adaptations that are possible or appropriate for a particular species are constrained by its structure and ecology, by its physical and biotic environment, by its history, and by genotypic limits to the extent to which it can vary in response to selective pressures. Adaptations for dispersal interact with and may be conditional upon adaptations for pollination (e.g. the position, number, structure and phenology of flowers) and for seedling establishment (seed size and number, seed protection, seed shape, seed dormancy and seed placement) (Fenner, 1985; Primack, 1987). Seed size has a particularly close interaction with seed dispersal mechanisms (Harper, Lovell & Moore, 1970).

Adaptations for seed dispersal must thus be seen in the context of the complex web of competing and interacting adaptations, constraints and possibilities that affect and ultimately determine the biology of a species. Within a species, the factors affecting seed dispersal may not be uniform. Different populations or individuals of a species may, for example, grow in widely different physical and biotic environments. Even within a single population, the biotic and physical environment may change greatly through time. A population of a pioneering tree species like Rhamnus catharticus or Salix atrocinerea, for example, may pass from a colonizing stage in an open habitat to a late-successional stage in closed woodland.

Introduction

This paper focuses on events post-pollination, and will only describe the forms of pollen, stigma, style, ovary, and embryo sac where structure is relevant to the process being discussed. Data have been included on a selective basis; the paper, therefore, must not be regarded as an exhaustive analysis. For further information, a number of reviews have recently appeared relating to pollination and fertilization. These include reviews of the male gametophyte (Mascarenhas, 1989), pollen, pistil and reproductive function (Knox & Williams, 1986), pollen germination and tube growth (Heslop-Harrison, 1987; Steer & Steer, 1989), micro- and megasporogenesis and fertilization (Fougere-Rifot, 1987), the cytological basis of the plastid inheritance in angiosperms (Hagemann & Schroder, 1989), and molecular aspects of fertilization (Cornish, Anderson & Clarke, 1988). Self-incompatibility has also been deliberately given a low priority in this paper because excellent reviews have been published (Gibbs, 1986; Heslop-Harrison, 1983; Clarke et al., 1989) and in order to emphasize other aspects of the process of fertilization and its control. Where relevant in any discussion, self-incompatibility is briefly considered.

The pollen–stigma interaction

Stigma receptivity

The development of receptivity of the stigmatic surface may vary widely between species from several days prior to anthesis to several days after the flower opens (Palser, Rouse & Williams, 1989). In ‘wet’ stigmas receptivity often coincides with accumulation of stigmatic secretion and in most stigmas with a positive response for non-specific esterase (Heslop-Harrison and Shivanna, 1977).

Introduction

The pattern of resource allocation in plants and its regulation is of great practical importance to agriculturalists and horticulturalists, and is of conceptual interest to ecologists and evolutionary biologists. This chapter discusses ecological and physiological aspects of reproductive allocation and includes an approach to allocation based on plant developmental morphology; the improvement of agricultural yield in terms of increased allocation to reproductive sinks is also evaluated.

It is generally considered that nutritional and energetic resources are allocated in a strategic manner that maximizes fitness (Harper, 1977; Fitter, 1986; Bazzaz et al., 1987). Resources are allocated to three activities: growth, reproduction and defence. The inherent pattern of allocation is considered to reflect the effect of past selection pressures in optimizing both fecundity and survival. Thus genetic differences in the pattern of resource allocation tend to be broadly associated with different ecological strategies. For example, Harper (1977) has argued that in a crowded community where there will be intense competition for resources a successful individual needs vigorous shoot and root growth to maximize resource acquisition; in this case fecundity will be associated with, and depend upon, a major allocation of resources to vegetative growth. In contrast in more open environments with relatively little competition from neighbouring individuals allocation to vegetative growth becomes less important and so a large proportion of the total resource may be invested in reproductive growth and development, thereby maximizing the chance of leaving descendants.

Introduction

The effects of the environment on the number and quality of the offspring produced by organisms is the fundamental concern of evolutionary ecology. For plants, the number and quality of their offspring may be limited by external environmental factors, such as pollination and seed predation, or internally by the availability of resources which, in turn, may be influenced by weather and soil conditions, herbivory, disease and competition. Each of these factors may vary over time and location. In response to these varying environmental conditions, plants may adjust their resource commitment to seed production by altering (i) the number (or gender) of flowers that differentiate, (ii) the number of ovules per flower, (iii) the number of fruits and seeds that abort, and (iv) the mass of the mature seeds (see, for example, Adams, 1967; Primack, 1978; Lloyd, 1980; Stephenson, 1981, 1984; Thomson, 1989). In the parlance of our colleagues who study animals, plants can potentially alter (i) clutch size (a clutch is defined as offspring in more or less the same developmental phase that are simultaneously drawing resources from the maternal plant), (ii) the interval between successive clutches, and/or (iii) the mass of the offspring at the termination of maternal care (Stearns, 1976; Burley, 1980).

The goals of this Chapter are to review the individual mechanisms by which plants regulate maternal investment under various environmental conditions, to review the directions in which these mechanisms are regulated with respect to one another, and to review the consequences of the different mechanisms of regulation on offspring performance (quality).

Interactions with endogenous receptors

The existence of functional endogenous receptors for lectins in plants has often been suggested. Such receptors, by definition, are good inhibitors of lectins from the same plant and therefore their presence ought to be relatively easy to demonstrate. Indeed, the occurrence of endogenous inhibitors in the seeds of Vicia cracca had been suggested as early as 1960 (Renkonnen, 1960) and similar observations have been made for the seeds of Dioclea sclerocarpa (Pusztai & Moreira, unpublished observations).

Most legume seeds contain appreciable amounts of α-galactosidic oligosaccharides, such as raffinose, stachyose and verbascose, which are good inhibitors of galactose-specific lectins. In seeds such as soyabean, which contain both such lectins and also α-galactosides, the oligosaccharides may be regarded as natural lectin-binders. The lectin, which is a major example of the N-acetyl-D-galactosamine/galactose-specific class of lectins co-exists with raffinose-type oligosaccharides in the seeds. These naturally occurring sugars are about 14% as effective in inhibiting the haemagglutination activity of the lectin as N-acetyl-D-galactosamine (Pereira et al., 1974). It is thus not surprising that less than 20% of the lectin can be absorbed on to guar gum-based affinity absorbents from non-dialysed extracts of soyabean. However, both by dialysis and/or precipitation of the proteins with ammonium sulphate followed by dialysis, most of the natural lectin-binders are removed from the soyabean extracts. The lectin in the extract is then selectively and quantitatively absorbed by the affinity column. Similar results can be obtained if the seed meal is first washed by 60% aqueous ethanol to remove the bulk of low molecular weight glycosides before the extraction of the lectin (Pusztai, Stewart & Watt, in press).

Although plant lectins have been known for the last 100 years, we are only beginning to understand their functions in the life-cycle of plants. The fact that lectin genes have been conserved well during evolution, especially in Leguminoseae, and that lectin proteins coded for by the genes are homologous, argues very strongly that lectins may have important, albeit mainly unknown, roles in the life of the plant. In the past, many such roles have been proposed. Thus lectins may act as defensive agents, carbohydrate transporters, recognition agents, storage proteins or growth regulators, etc (for references see Pusztai et al., 1983a; Etzler, 1986). However, experimental evidence in support of such roles is still scanty at best.

Whatever role(s) lectin may play, it is important to find out their localization and distribution in various parts of the plant. Temporal changes in their distribution may also help to formulate ideas about the functions of lectins. In addition, any proposed role will also have to be compatible with their ultrastructural localization and any possible changes in it during the various stages of plant development. Owing to their great importance, studies on the localization of lectins have been reviewed extensively (Pusztai et al., 1983a; Etzler, 1985, 1986).

Localization in the plant and cellular location

Seeds

Lectins are generally most abundant in the seeds of plants, especially in the seeds of Leguminoseae and Graminaceae. However,lectins are by no means confined to seeds. Variable, although usually small amounts of lectins or lectin-like, immunochemically cross-reacting proteins (CRM) are also present in vegetative parts of plants.

The carbohydrate specificity of lectins is established most conveniently by the Landsteiner hapten-inhibition technique (Landsteiner, 1962; Goldstein & Hayes, 1978; Goldstein & Poretz, 1986). In this method the inhibitory effectiveness of various mono- and oligosaccharides of known composition is compared in a recognized and convenient reaction of lectins, such as haemagglutination. Initially, on the basis of their reactivity with monosaccharides that differed in configuration at C-3 or C-4 of the pyranose ring, Mäkelä (1957) suggested that lectins could be divided into four classes. Thus, for example, concanavalin A, which reacts with D-mannose and/or D-glucose, belongs to group III. Soyabean lectin, whose specificity is for N-acetyl-D-galactosamine and/or D-galactose, is classified as group II. According to Mäkelä's scheme the L-fucose-binding lectins are members of group I (Fig. 1.1). However, no group IV-specific lectins have been found in Nature so far. In more recent studies, although still within the four basic classes of Mäkelä, the definition of the specificity of lectins has been further refined and extended. Accordingly, from their reaction with simple sugars, lectins are first classified into broad groups of lectin classes as either mannose/ glucose-specific, or N-acetylglucosamine-specific, or N-acetylgalactosamine/galactose-specific or fucose-specific lectins. This is then followed by a more precise classification based on extensive investigations with a great number of oligosaccharides of known composition and structure. From such studies of the most complementary carbohydrate structure, the recognition and binding site of the lectins within the four main classes, is obtained.

Lectins are known to react chiefly with the non-reducing end of oligoand polysaccharides.

Most of our knowledge of the primary structure of plant lectins has been gathered in the last decade from extensive chemical studies with lectins isolated from leguminous plants. The comparative, evolutionary and chemotaxonomical aspects of such studies are of particularly special value for our understanding of the potential functional significance of lectins and gene alterations in plants during evolution.

Primary sequences

A comparison of partial or complete primary sequences of lectins of diverse origin, including one- or two-chain lectins, has shown up extensive homologies between different lectins (Strosberg et al., 1986a,b). Thus, for example, the amino terminal sequences of the β-chains of two-chain lectins from Vicieae and of one-chain lectins from other tribes of Papilionaceae or Caesalpinoideae, compared with the appropriate sequence of concanavalin A (residues of 123–147), are highly homologous (Strosberg et al., 1983; 1986a,b). It has also become clear from sequence comparisons that lectins from the same tribe are more homologous to each other than to members of other tribes. Rather interestingly, phenylalanine residues in positions 6 and 11 are conserved in all 26 lectins used for comparison studies. Residue 5 is also found to be highly conserved and occurs in 23 of the 26 lectins studied.

Thus, the comparison of even partial sequences has turned out to be very rewarding. However, more is to be expected from comparisons of the structure of fully sequenced lectins, and the value of such structural comparisons is likely to grow in future. Up until now, Strosberg and his colleagues have already compared the full primary amino acid sequences of seven leguminous lectins.

It is just over a hundred years ago that the first plant lectin was described by Stillmark (Stillmark, 1888, 1889). Working with extracts from castor bean (Ricinus communis) he obtained a preparation which agglutinated red blood cells. As more and more of such substances were later discovered in other plants, and as a common name for them, the term, haemagglutinin, was proposed by Elfstrand (1898). The striking toxicity of some of the haemagglutinins, such as ricin and abrin (Abrus precatorius) has made the effects of these substances relatively easy to test on animals. Ehrlich, who is usually considered to be the father of immunology, has shownthat rabbits fed with small amounts of seeds containing the toxin became partially immune to the toxicity, thus demonstrating that the haemagglutinins were also antigenic. Landsteiner and Raubitschek (1908) showed later that not all haemagglutinins need necessarily be as toxic as ricin or abrin. For example, the agglutinins obtained from common beans (Phaseolus vulgaris), peas (Pisum sativum), lentils (Lens culinaris), etc, were relatively non-toxic, water soluble proteins. It is now known that such haemagglutinating proteins are found in all taxonomic groups of the Plant Kingdom and that they are not all overtly toxic.

The next momentous step in the history of haemagglutinins was the realization that some of the haemagglutinins agglutinated blood cells only from some groups of individuals within the ABO blood group system, without affecting cells from other groups (Renkonnen, 1948; Boyd & Reguera, 1949).

The realization made almost simultaneously by Boyd (Boyd & Reguera, 1949) and Renkonnen (1948) that some plant lectins are blood group-specific has done more than any other to turn the studies of lectins from an interesting, though somewhat esoteric field of study into a major and multidisciplinary science. Thus, Nowell discovered in 1960 that preparations of a lectin from the seeds of Phaseolus vulgaris (PHA) have strong mitogenic activity on peripheral lymphocytes (Nowell, 1960). Not long after this, Aub and his colleagues recognized by a chance observation that some transformed cells have greater affinity for lectins than the cell lines from which they originate (Aub, Tieslau & Lankester, 1963). Obviously, both these important effects are the result of specific reactions between cell surface carbohydrate structures and the appropriate lectins. This initial specific recognition of and binding to cell surface receptors by lectins occurs both in vivo and in vitro and has wide-ranging and dramatic effects on cells.

These two major discoveries led to an upsurge of research activity by scientists from all fields of biological sciences, particularly, cell biologists interested in immunology and cancer research. As the composition and the physical state of the cell surface is known to change throughout the life-cycle of the cell and to alter radically during cancerous or precancerous transformation, it is not surprising to find that lectins, which can specifically recognize such surface receptors, are exploited as welladaptable and excellent tools for monitoring and studying these changes.

Agglutination

Agglutination is one of the most characteristic and important reactions of those plant lectins which have at least two carbohydrate-binding sites.

Insulin-mimicking effects on adipocytes and other cells

The finding that lectins, such as concanavalin A, wheatgerm agglutinin, PHA and others, can mimic the effects of insulin on adipocytes is one of the most dramatic examples of interactions between lectins and cell membranes. Thus, these lectins stimulate the synthesis of triglycerides and the transport and oxidation of glucose in fat cells in vitro (Cuatrecasas & Tell, 1973; Pusztai & Watt, 1974; Livingstone & Purvis, 1980). Similar effects have also been observed in vivo (Katzen et al., 1981).

It is clear that, depending on their carbohydrate specificities, various lectins display significant differences in their insulin-mimicking activities. These differences become clearer if the insulin effect is not tested by the commonly used glucose oxidation assay, but rather by assessing their effects on lipogenesis or from their inhibition of lipolytic effects of hormones, such as epinephrine or other lipolytic hormones, on rat or hamster adipocytes (Ng, Li & Yeung, 1989). Rat cells lack functional αadrenergic receptors, while hamster adipocytes possess both α- and βdrenergic receptors (Carpene, Berlan & Lafontan, 1983) and these variations help to differentiate between the insulin effects of the different lectins. Thus, the D-mannose and N-acetyl-D-glucosamine-specific lectins generally have both potent antilipolytic and lipogenic activities while the fucose-binding lectins possess antilipolytic activity only. With some exceptions, most of the D-galactose-specific lectins show slight insulinlike effects. Furthermore, the N-acetyl-D-galactosamine-specific Dolichos biflorus and Maclura pomifera lectins exert antilipolytic activity in hamster but not in rat adipocytes.

While vines and woody lianas are important growth forms in tropical forest ecosystems, and a noticeable component in many temperate forest ecosystems, they are largely ignored in discussions of the floras of arid and semi-arid ecosystems of the world. Nevertheless, vines represent a small but fascinating component of these floras, and provide insights into modes of adaptive strategies in the tolerance of drought and heat stress. In this chapter we first review the variation in growth forms of arid zone vines, then describe the floristic diversity of vines in mediterranean-type and desert ecosystems, and finally discuss the adaptations of arid zone vines to environmental stress.

Growth forms

Vine species in arid and semi-arid areas of the world occur in a remarkably broad range of growth forms although none of these is unique to such areas. At one extreme are woody lianas, which may occur as both evergreen and deciduous species. Species with this growth form may grow up over shrubs to form their own canopies in full sun, as do tropical lianas, or may grow high into woodland trees while remaining within the shaded canopy of their host species. Ephedra pedunculat, unusual as a liana in the Gymnospermae (Cutler, 1939), has been observed to use telephone poles as a substrate in the Chihuahuan Desert region of south Texas (S. Carlquist, personal communication).

At the other extreme are annual vine species that commonly trail along the ground or grow up into the lower branches of shrubs or suffrutescent species.

Synonymy

Schoenetum Newbould 1960; Roydon Common Zone 5a Daniels & Pearson 1974p.p.; Schoenetum nigricantis Ivimey-Cook et al. 1975; Schoeno-Juncetum subnodulosi ericetosum (Allorge 1922) Wheeler 1980Z? p.p.

Constant species

Anagallis tenella, Erica tetralix, Molinia caerulea, Narthecium ossifragum, Schoenus nigricans, Aneura pinguis, Campylium stellatum, Scorpidium scorpioides, Sphagnum auriculatum, S. subnitens.

Physiognomy

The Schoenus nigricans-Narthecium ossifragum mire includes mildly calcicole Schoenus vegetation of southwest lowland Britain which cannot readily be integrated in the Schoenetum. As in that community, Schoenus is usually a strong dominant here, and its grey-green, semievergreen foliage often enables stands to be picked out at a distance from their usual context of brownor straw-coloured Sp/zagnwm-dominated vegetation or heaths. And the tussock habit of the plant, with its robust rootstocks crowned by densely-caespitose clumps of shoots (Sparling 1962a, 1968), again gives the vegetation its distinctive structural character, with stools packed at varying densities and separated by systems of runnels. The tussock tops provide their characteristic niche for calcifuges but, in this community, the calcicolous element is confined to areas of close contact with the irrigating waters, other distinctive plants making an appearance around the bases of the tussocks.

Species occurring frequently in both communities are few in number and the most obvious is Molinia caerulea, which is generally abundant here, reinforcing the tussocky character of the vegetation. Typically, mixtures of Schoenus and Molinia cover the bulk of the ground without any Juncus subnodulosus, a species that is quite often co-dominant in the Schoenetum. Indeed, no other rushes or sedges of medium stature play an important role in this community: there is occasionally some J. acutiflorus, but it is not usually abundant. Other elements conspicuous by their absence are the various small sedges and other herbs that commonly line the runnels of the Schoenetum and provide much of its calcicolous stamp, and the taller mesophytic plants that give that community much of its colour: of these only Carex panicea and C. demissa are occasionally found, and sometimes Dactylorhiza incarnata ssp. pulchella.

Among other vascular plants represented here, only Anagallis tenella provides a floristic link between the two communities.

With something as complex and variable as vegetation, no key can pretend to offer an infallible short cut to diagnosis. The following should thus be seen as simply a crude guide to identifying the types of mire vegetation in the scheme and must always be used in conjunction with the data tables and community descriptions. It relies on floristic (and, to a lesser extent, physiognomic) features of the vegetation and demands a knowledge of the British vascular flora and, in many cases here, of bryophytes. It does not make primary use of any habitat features, though these may provide a valuable confirmation of a diagnosis.

Because the major distinctions between the vegetation types in the classification are based on inter-stand frequency, the key works best when sufficient samples of similar composition are available to construct a constancy table. It is the frequency values in this (and, in some cases, the ranges of abundance) which are then subject to interrogation with the key. Most of the questions are dichotomous and notes are provided at particularly awkward choices or where confusing zonations are likely to be found.

Samples should always be taken from homogeneous stands and be 2 × 2m or 4 × 4m according to the scale of the vegetation or, where complex mosaics occur, of identical size but irregular shape. Very small springs or flushes can be sampled in their entirety.

1 At least some Sphagna constant and often forming an important structural element in the vegetation 2

Sphagna at most occasional and patchily prominent and often altogether absent 34

2 Sphagnum auriculatum constant and locally abun dant in spring-head, soakway or pool vegetation with some of Hypericum elodes, Potamogeton polygonifolius, Juncus bulbosus, Ranunculus flammula, Deschampsia cespitosa, Saxifraga stellaris, Ranunculus omiophyllus, Montia fontana, Anthelia julacea, Philonotis fontana, Scapania undulata, Marsupella emarginata, Dicranella palustris 14

Sphagnum auriculatum can be frequent but not with the above 3

3 Sphagna occurring as an extensive carpet or in patches with graminoid monocotyledons (some of Scirpus cespitosus, Eriophorum vaginatum, E. angustifolium, Molinia caerulea) and sub-shrubs (some of Erica tetralix, Calluna vulgaris, Myrica gale, Vaccinium spp., Empetrum sspp.) 17