DEVELOPMENT OF THE POLLEN IN THE ANTARCTIC FLOWERING PLANT COLOBANTHUS QUITENSIS ( KUNTH ) BARTL .

Colobanthus quitensis (Kunth) Bartl. produced two types very small bisexual fl owers. In the Antarctic natural conditions chasmogamic and cleistogamic fl owers most often form fi ve stamina with short fi laments. Two microsporangia with a three-layer wall form in the anther. Microspore mother cells, which develop into microspores after meiosis, form inside the microsporangium. Microsporocytes of Colobanthus quitensis are surrounded with a thick callose layer, the special wall. After meiosis, the callose wall is dissolved and microspores are released from the tetrad. The production of proorbicules, orbicules and peritapetal membrane, and the construction of a complex sporoderm with numerous apertural sites were observed. When microspore and pollen protoplasts underwent necrosis, probably as a result of temperature and osmotic stress, sporoderm layers formed around microspores, and the cell tapetum did not disintegrate. However, woody wall layers did not accumulate in endothecium cells.


INTRODUCTION
The summer season is considered to last for about 5 months, more or less from November to March, taking especially day and night length into account.Considering the length of the snow-cover period in the areas uncovered by ice, the growth season of seminiferous plants occurring here may be shortened by 1-2 months.

Light and electron microscopy
The material for light and electron microscopy was prepared by a standard method, i.e. fi xed in an osmic acid-glutaraldehyde solution.The material was dehydrated in an alcohol-acetone series, and embedded in Poly Bed 812 epoxy resin.Semi-thin and ultra-thin sections were prepared in a Leica (Ultracut R) ultramicrotome, using glass and diamond knives, stained with toluidine blue, and toluidine blue and azure B, following P e a r s e (1962), and then examined and photographed under a light microscope, models Olympus and Nikon Optiphot II.All preparations to be photographed under a light microscope were mounted in glycerin.Ultra--thin sections (60 -90 nm) were placed onto a copper wire gauze (300 meshes) and contrasted with a saturated aqueous solution of uranyl acetate and lead citrate, following R e y n o l d s (1973).Observations and electronograms were made in a JEOL JEM 100S transmission electron microscope.

RESULTS
Colobanthus quitensis fl owers usually produced 5 stamina with short fi laments.
In the anther there were two microsporangia, which had a regular circular shape in cross section.Four cell layers were visible in the microsporangium wall.The most external layer was epidermis under which, almost on its entire length, there was a layer of starch--fi lled cells which transformed into endothecium.Only at some places there was a middle layer of parenchyma.A nutritive tissue, tapetum (Fig. 1), developed inside.The central part of the microsporangium was occupied by archesporial tissue from which, as a result of meiosis, microspores were formed.Before meiosis began, microspore mother cells constructed special callose walls.Callose surrounded microsporocytes until the tetrad stage.After meiosis, when callose disintegrated, microspores were released from the tetrads, and a complex cell wall, sporoderm, formed on the surface of their protoplasts.
The developing mother cells of the male gametophyte, from the premeiotic stage, through meiosis, to the stage of double-cell pollen grains (Fig. 2), were surrounded and nourished by the cell tapetum.Cells of the secretory tapetum contained one or two nuclei.They retained their individuality until the gametophyte stage.In the thick cytoplasm of tapetal cells, before microspores were released from the tetrads, osmophillic droplets appeared and moved inside the pollen sac (Fig. 3).Apart from small lipid droplets, vast areas containing such material visible within the cytoplasm and on the borderline with the pollen sac, could be observed in the tapetal cells.This material, in the form of proorbicules, moved from various areas of the tapetum towards pollen grains, and in the form of bigger structures, orbicules, joined the developing sporoderm (Fig. 5 ).A dozen or so of apertural sites devoid of electron-dense sporopollenin material (Fig. 5, ap) differentiated in the forming sporoderm.In the semi-thin sections, stained with toluidine blue, proorbicules and orbicules were visible in the form of blue-green drops (Fig. 5 and Fig. 6).Orbicules settled on the surface of pollen grains until the fi nal stage of tapetal disintegration and atrophy.At the same time, endothecium with conspicuous ridge-like wall thickenings was already fully developed in the microsporangium wall (Fig. 5).
It is worth noting that, in the case of microspore or pollen protoplast necrosis, as a result of temperature or osmotic stress, tapetum did not disintegrate, and endothecium with woody ridges did not develop, although layers were almost normally deposited on the sporoderm (Fig. 3).
When pollen was mature and contained a bicellular and tricellular male gametophyte, Colobanthus quitensis microsporangia burst open at a specifi c site, stomium, on the inner (adaxial) side of the anther (Fig. 4 and Fig. 7).After anther dehiscence some pollen reached the surface of the receptive cells on the fi ve-part feathery stigma, while some remained inside the microsporangium and germinated (Fig. 4).

DISCUSSION
A small number of microspore mother cells differentiated in each of the two microsporangia of the Colobanthus quitensis anther.However, already in the early fi rst meiotic prophase these cells were surrounded with a thick callose wall and were isolated in this way until the tetrad stage.Callose began to hydrolise when the construction of sporoderm started, just as in other plants (H e s l o p -H a r r i s o n et al. 1986 ).
Microsporgenesis in Deschampsia antarctica was completely different.In the microsporocyte cell walls of the Antarctic hair grass, deposition of a thick callose layer was not observed.From the stage of sporogenic tissue cells, through all the meiotic stages, to tetrads, microsporocyte cell walls of the Antarctic hair grass remained thin.This feature distinguishes Deschampsia antarctica from other grasses, in which callose special walls of meiotic cells are common (M c Q u a d e and Yo u n g -R o t t l e r , 1984; R o d k i e w i c z et al.

1996), as in other angiosperms (P a c i n i et al. 1985).
In the microspore of Colobanthus quitensis, after the fi rst mitosis, the formed generative nucleus does not break off the sporoderm, but remains fi rmly attached to the inner surface of the pollen wall.After mitosis, the generative cell stays close to the sporoderm, but it is separated from the vegetative cell by a very delicate wall with a little amount of fi brous material.The typical wall      1996).The further development of Colobanthus quitensis pollen takes place in a similar way to other species with trinucleate pollen.The generative cell remains fi rmly attached to the cell wall and, in this position, divides mitotically producing two sperm cells.This is different in the microspores of many described angiosperms with binucleate pollen (M a s c ar e n h a s , 1989; R o d k i e w i c z et al., 1996).In these species, soon after mitosis, sometimes after an hour for example, the generative cell breaks off the sporoderm and becomes completely immersed in the cytoplasm of the vegetative cell.There it often changes its shape from lenticular to fusiform and pointed at the ends.This shape can be changed and is maintained by cytoskeleton microtubules running parallel to its long axis.In its further development, in pollen or pollen-tube, the generative cell undergoes one more mitosis giving rise to two sperm cells.The elongated shape and transfer of sperm cells and vegetative nucleus, that is the male germ unit, in the pollen-tube of fl owering plants is of cytoskeletal origin (Va n L a m m e r e n et al. 1985; 1989; B e dn a r a , 2003).When two sperm cells are produced after the second mitosis, apart from the common wall or at least plasmatic membranes inherited from the generative cell, they are related by a group of phragmoplast microtubules and, additionally, cortical microtubules which were polymerised earlier and do not disappear despite the beginning of the next mitosis.
The male germ unit of Antarctic pearlwort was observed in a transmission electron microscope.Unfortunately, we failed to photograph the only object.Nevertheless, the morphologic variation of sperm cells, with features as in Plumbago zeylanica (R u s s e l , 1984) or Brassica (D u m a s et al. 1985; C h a r z y ń s k a et al. 1989), were clearly visible.One sperm cell of Antarctic pearlwort contained mainly plastids, and in the cytoplasm of the other cell, its elongated protrusion stuck in the vegetative nucleus, particularly mitochondria could be observed (G i e ł w a n o w s k a , 2005).Some of the mature trinucleate pollen of Antarctic pearlwort spills out of the anther onto the surface of the feathery stigma and germinates there.The pollen which remains in the microsporangium germinates inside the anther, as is the case in other cleistogamic fl owers (Giełwanowska et al. 2006).

CONCLUSIONS
1. Cleistogamy in Colobanthus quitensis was induced by low temperatures, high air humidity and strong winds.
2. A small number of microspores differentiated in the Colobanthus quitensis microsporangium is typical of cleistogamic species.
Colobanthus quitensis (Kunth) Bartl. is the only fl owering dicot of the Antarctica.It has grown together with Deschampsia antarctica Desv.(Poaceae) for a few thousand years.It effectively colonises surfaces of the Antarctic geobotanical zone uncovered by melting ice (E d w a r d s , 1974; A l b e r d i et al. 2002).In the extremely hostile Antarctic conditions, Colobanthus quitensis (Caryophyllaceae) fl owers profusely almost every year.After pollination and fertilization it produces seeds.Antarctic pearlwort seeds are often dead due to unfavourable microhabitat conditions and the short cold summer (Convey, 1996; G i e ł w a n o w s k a et al. 2006).
Irrespective of other factors affecting the plants, this time is critical for the formation of generative cells and seeds (E d w a r d s , 1974; G i e ł w a n o w s k a et al. 2006).Macroscopic observations of fl ower and seed development in Antarctic pearlwort and Antarctic hair grass were conducted in various areas of the Maritime Antarctic and in different growth seasons.Greene and Holtom (1971) described the fl ower formation in Colobanthus quitensis and Deschampsia antarctica on the South Orkney Islands.B r o w n (1912) and H o l d g at e (1964) observed specimens of Antarctic pearlwort and hair grass which fl owered and dispersed seeds in the communities of South Georgia, while Holtom and G r e e n e (1967) on some of the South Shetland Islands.As only few descriptions of generative cell development in Antarctic fl owering plants have been published (G i e ł w a n o w s k a , 2005; G i e ł w a n o w s k a et al. 2005; G i e ł w a n o w s k a et al. 2006), in this work we present some stages of microsporogenesis and pollen development in Colobanthus quitensis.

Fig. 2 .
Fig. 2. Binucleate pollen grains are visible in the microsporangium.Tapetum is disintegrated (ta).2000 x.Fig. 3. Microsporangium with disturbed pollen development.There is no developed endotecium in the microsporangium wall.Tapetum with dense, osmophilic drops (arrows) is visible.Pollen grains have well-developed sporoderm and degenerated contents (arrowheads).2000 x.Fig. 4. Longitudinal section of a closed fl ower bud with pollen grains in open microsporangia.Some of the pollen grains are germinating.1200 x.

Fig. 7 .
Fig. 7. Ultrastructure of bicellular pollen grains.Cell wall separating a generative cell and a vegetative cell (arrow) is visible.Dense cytoplasm and a nucleus with a nucleole.18 000 x.