Planta (1984)160:129 142
P l a n t a 9 Springer-Verlag 1984
Ultrastructural and histochemical studies on guard cells Alison C. Wille and William J. Lucas Department of Botany, University of California, Davis, CA 95616, USA
Abstract. Serial thick sections of guard cells from Vicia faba L., Nicotiana tabacum L., Alliurn cepa L., Zea mays L. and Beta vulgaris L. were obtained systematically (600-800 nm) and viewed with the transmission electron microscope in an effort to demonstrate the presence or absence of a symplastic transport pathway within the stomatal complex. Eight to ten stomata from each species were examined, and no continuous plasmodesmata were found connecting guard cells to sister guard cells or to adjacent epidermal or subsidiary cells. Continuous plasmodesmata were observed in immature guard cells, but were sealed (truncated) during the development of the mature cell wall. Histochemical stains, phosphotungstic acid and silver methenamine, were used to demonstrate differentiation within the mature guard-cell wall. The structural differentiation of the stomatal apoplastic region is discussed in relation to functional specialization. Plasma-membrane elaborations or plasmalemmasomes were identified in the guard cells of Zea, and it is suggested that these structures may function in ion transport.
Key words: Cell wall structure - Guard cell Stoma - Plasmalemmasome - Plasmodesma.
Introduction
Studies on symplastic connection between guard cells and adjacent subsidiary or epidermal cells have produced conflicting reports, and the structure of the guard-cell wall and its possible influence on apoplastic ion transport have not been investiAbbreviations: PTA-HCl=phosphotungstic acid and hydrochloric acid; SM =silver methenamine; UA-LC=uranyl acetate and lead citrate
gated. Plasmodesmatal connections between guard cells and adjacent cells were first reported on the basis of light-microscopic studies (Waterkeyn and Bienfait 1966; Litz and Kimmins 1968; Inamdar et al. 1973). The presence of pit callose, fluorescent when stained with aniline blue, was taken as evidence for plasmodesmata in one study (Peterson and Hambleton 1978). The improved resolution available with transmission electron microscopy provides direct means for visualizing plasmodesmata, but most ultrastructural studies reported the absence of cytoplasmic connections in mature guard cells (Brown and Johnson 1962; Srivastava and Singh 1972; Singh and Srivastava 1972; A1laway and Setterfield 1972; Ziegler et al. 1974; Peterson and Hambleton 1978; Rutter and Willmer 1979; Galatis and Mitrakos 1980). Plasmodesmata have been identified in the guard cells of three species, Viciafaba and Nicotiana tabacum (Pallas and Mollenhauer 1972a, b) and Commelina communis (Fujino and Jinno 1972), but these reports are not entirely convincing. In the former study, there is reason to believe, based on the thickness of the cell walls and absence of a well-developed vacuolar system, that the guard cells in which continuous plasmodesmata were demonstrated were, in fact, immature. In the latter study, the quality of the micrographs was insufficient to interpret accurately the images of the alleged plasmodesmata. Continuous plasmodesmata have been observed to connect developing guard cells with adjacent cells, and plasmodesmata that appear truncated have been observed in the walls of mature guard cells (Willmer and Sexton 1979). However, no published micrographs contain both continuous plasmodesmata and a view of the stomatal pore that demonstrates that it had opened for the first time. The resolution of the light microscope is not sufficient for the accurate visualization of plasmo-
130
desmata, while the high resolution obtained with the transmission electron microscope requires sections so thin that several hundred must be observed in order to inspect an entire cell. Serial thick sections have been shown to provide sufficient contrast and resolution to visualize plasmodesmata. Through this technique, it has been possible to carry out a thorough search for plasmodesmata in 20-30 stomata in five plant species. The fact that not a single continuous plasmodesma was encountered provides strong evidence that plasmodesmatal communication does not play an important role in stomatal function. The absence of cytoplasmic communication between guard cells and adjacent cells has important implications in relation to the mechanism of stomatal ion transport. With plasmodesmata being absent, transport of ions and carbohydrates to and from guard cells must be entirely apoplastic. It follows that the guard-cell wall may be involved in specialized interactions with ions as they are transported between guard cells and adjacent cells. Histochemical reagents have been used to demonstrate the differentiation of sieve-element walls into regions with distinct structure and composition (Lucas and Franceschi 1982). We have applied similar techniques to the walls of guard cells to demonstrate structural specialization of various apoplastic regions. We propose that the specific structure of the guard-cell wall has functional implications with respect to intercellular communication between guard cells and epidermal cells, and merits further study. Materials and methods Plant materials and culture techniques. AlIium cepa L. (grown from bulb sets; Coast Brand, Aurora, Ore., USA), Beta vulgaris L. cv. USH10 (Spreckels Sugar Co., Woodland, Cal., U S A ; Lot 3102) and Vicia faba L. cv. Long Pod (W. Atlee Burpee Co., Warmister, Penn., U S A ; Lot 517602) were maintained in a growth chamber under the following conditions: 27~
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells day-17~ night temperatures, 75%-100% relative humidity and a 16-h photoperiod at 180 gmol m -2 s -1. The soil used was a 2:1 : 1 (by vol.) mixture of medium vermiculite, peatmoss and sand, and the plants were watered daily and fertilized with full-strength Hoagland's solution (Hoagland and Arnon 1950) once per week. Fully expanded leaves were harvested when the plants were four to eight weeks old. In addition, young leaves of B. vulgaris were harvested when less than 2 cm long for observation of developing guard cells. Allium cepa L. cv. White Lisbon (Northrup King Seeds, Fresno, Cal., USA) was grown from seed in vermiculite, and the seedlings were maintained under constant light at 52 gmol m - 2 s-1. Cotyledons were harvested after 5 d, and the region between the hook and the terminally attached seed coat was prepared for observation of developing guard cells (Palevitz and Hepler 1976). Phaseolus vulgaris L. cv. Light Red Kidney (University of California, Davis, California Crop Improvement), Nicotiana rustica L. (U.C. Davis, Botany Department Greenhouse Facility), Commelina coelistis L. (Kew Gardens, Richmond, Surrey, UK), and Zea mays L. (A632 x [C3640 x Ohio 43] ; Crow's Hybrid Corn Co., Milford, Ill., USA), and Allium porrum L. (grown from bulbs) were grown and maintained in our greenhouse facility where they were watered three times per week with 0.5-strength Hoagland's solution.
Tissue preparation. Tissue from fully expanded leaves was cut with a tissue slicer (Mickle Laboratory Engineering Co. ; Gomshall, Surrey, UK) into 1.0 mm x 0.3 mm sections. After cutting, the tissue was immersed immediately, at room temperature, in a fixative containing glutaraldehyde (2%), depolymerized paraformaldehyde (3%, for Allium spp.) and sodium-cacodylate buffer (50 mM, pH 7.2). Vacuum infiltration was occasionally employed to facilitate penetration of the fixative. Following fixation (3-10 h, 4 ~ C), tissue was post-fixed with 1% osmium tetroxide in sodium-cacocylate buffer (25 mM, pH 7.2) for 1-3 h, dehydrated through a graded acetone series, and embedded in Spurr's resin (Spurr 1969). Serial thick sectioning. To determine with certainty the presence or absence of plasmodesmata, guard cells were serial-sectioned, and each section was carefully examined for plasmodesmata. To decrease the number of sections required to serial-section a pair of guard cells completely, sections were cut at a thickness which produced green to red interference colors (approx. 60(~800 nm). The stomata of Vicia faba, Nicotiana tabacum, Allium cepa, Zea rnays, and Beta vulgaris were serial-sectioned with a diamond knife on a Porter-Blum MT-2 ultramicrotome (Ivan Sorvall, Newton, Conn., USA). After sectioning, the specimens were mounted on 0.5 m m x 2 . 0 m m slotted Cu-Rh,
)
Figs. 1-9. Plasmodesmata between leaf cells of various species. Sections were cut at a thickness that produced red-green interference colors (600-800 nm) and were viewed unstained at 100 kV. Fig. 1. Plasmodesmata in primary pit field between mesophyll cell (above) and epidermal cell (below) in Allium cepa. x 15,000; b a r = 1 gm. Fig. 2. Plasmodesma between epidermal cells of Vieia faba. x 40,000; bar=0.25 gm. Fig. 3. Plasmodesmata in a primary pit field between subsidiary cell (above) and epidermal cell (below) of Zea mays. x 40,000; bar = 0.25 gm. Fig. 4. Truncated plasmodesmata between guard cell (below) and epidermal cell (above) of V. faba. x 40,000; bar = 0.25 gm. Fig. 5. Truncated plasmodesmata between guard cell (below) an epidermal cell (above) of Nicotiana tabacum, x 25,000; bar=0.5 lain. Fig. 6. Truncated plasmodesmata between guard cell (below) and epidermal cell (above) of Beta vulgaris, x25,000; bar=0.5 gm. Figs. 7-9. Serial thick sections of the guard cell wall of V. faba demonstrating that plasmodesmata are indeed truncated, x 15,000; bars = 1 gm
Figs. 11)-12. Truncated plasmodesmata between epidermal cells (top) and guard cells (below) in various species Sections post-stained with uranyl acetate and lead citrate; bars = 0.1 gm. Fig. 10. Truncated plasmodesma in guard-cell wall of Beta vulgaris, x 100,000. Fig. 11. Guard-cell plasmodesmata in V.faba. x 100,000. Fig. 12. Guard-cell plasmodesmata in N. tabacum, x 80,000
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells
131
132
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells Formvar-coated grids (Ernest F. Fullam, Schenectady, N.Y., USA). Unstained sections were scanned for plasmodesmata at 100 kV on a JEOL (Tokyo, Japan), model 100S transmission electron microscope. Carbohydrate histochemistry. For general ultrastructural observation, silver sections were cut with a diamond knife on a 2128 Ultratome (LKB Instruments, Washington, D.C., USA) and mounted on uncoated grids. Sections were stained with lead citrate (Reynolds 1963) and uranyl acetate (2% in 50%, v/v, methanol in water). Gold sections mounted on nickel grids were oxidized by floating on 1% periodic acid in water for 20-30 min at room temperature and stained by floating, for 1 h, in the dark, at 60 ~ C, on drops of staining solution containing 0.1% silver nitrate and 3% hexamethylenetetramine titrated to pH 9 with 5% sodium borate (Rambourg 1967). The grids were then rinsed with distilled water and floated on 0.5% aqueous sodium-thiosulfate solution for 5 rain. Controls were run to demonstrate the specificity of the silver-methanamine reaction. Sections were oxidized with 10% hydrogen peroxide (15-30 min) rather than periodic acid, before incubation in silver-methanamine medium; alternatively, sections were stained without oxidation. Gold sections mounted on uncoated nickle grids were bleached (to remove osmium) by floating the grids on drops of 10% hydrogen peroxide for 15-30 min at room temperature. After washing in distilled water, the grids were floated on drops of staining solution (1% phosphotungstic acid in 10% hydrochloric acid; Hayat 1975) for 10 min in the dark.
Results
Visualization of plasmodesmata in thick sections Since it is not standard technique to observe plasmodesmata with the transmission electron microscope in thick sections (600-800 nm), it was necessary to verify that they can be unambiguously identified under these conditions. Epidermal cells (Figs. 1-3) were used to demonstrate that plasmodesmata are clearly visible in thick sections viewed at 100 kV. While resolution was generally poor, contrast was quite high, even without post-staining. Four sets of serial sections (two paradermal and two transverse) were cut from epidermal tissue of each of five species; Viciafaba, Nicotiana tabacum, Allium cepa, Zea mays and Beta vulgaris. Blocks sectioned in paradermal orientation were
133
approx. 0.5 mm 2, contained three to six stomata, and required 3 0 4 0 sections for a complete epidermal series. Transverse sections of comparable size contained one to three guard-cell pairs, and a complete series was represented in approx. 50 sections. Occasionally some sections were lost as a consequence of separation from the ribbon or breakage of the Formvar film, but never more than five sections, and in each species (except V. faba) at least one complete series was preserved. A total of eight to ten stomata from at least two different leaves from each species were thoroughly examined, and not a single continuous plasmodesma was observed to connect a guard cell to its sister guard cell or to an adjacent epidermal or subsidiary cell. The starch-lacking species, A. cepa, lacked continuous plasmodesmata, as did the four starch-containing species. While continuous plasmodesmata were never observed in mature guard cells, truncated plasmodesmata were common and appeared distinct (Figs. 4-6). Sections viewed in series (Figs. 7-9) demonstrate that the truncated appearance is not an artifact of oblique planes of sectioning. In addition, guard-cell plasmodesmata were always truncated by the interior region of the guard cell wall; further evidence that they were not merely sectioned at oblique angles. Finally, thick sections were approximately ten times the thickness of the average plasmodesma, making it highly unlikely that the plasmodesmata were not contained within a single section. The detailed structure of truncated plasmodesmata was viewed in silver sections (Figs. 10-12). Some appeared singly (Fig. 10) while others occurred in clusters and were connected by a characteristic sinus (Figs. 11, 12). The sinus was occasionally associated with a relatively electron-dense region of the wall (Fig. 11) which may indicate the location of the middle lamella. The guard-cell wall was typically thicker toward the guard-cell side of the sinus giving the overall indication that plasmodesmata had been truncated by deposition of the guard-cell wall (Figs. 11, 12). Truncated plasmo-
P l a n t a 9 Springer-Verlag 1984
Ultrastructural and histochemical studies on guard cells Alison C. Wille and William J. Lucas Department of Botany, University of California, Davis, CA 95616, USA
Abstract. Serial thick sections of guard cells from Vicia faba L., Nicotiana tabacum L., Alliurn cepa L., Zea mays L. and Beta vulgaris L. were obtained systematically (600-800 nm) and viewed with the transmission electron microscope in an effort to demonstrate the presence or absence of a symplastic transport pathway within the stomatal complex. Eight to ten stomata from each species were examined, and no continuous plasmodesmata were found connecting guard cells to sister guard cells or to adjacent epidermal or subsidiary cells. Continuous plasmodesmata were observed in immature guard cells, but were sealed (truncated) during the development of the mature cell wall. Histochemical stains, phosphotungstic acid and silver methenamine, were used to demonstrate differentiation within the mature guard-cell wall. The structural differentiation of the stomatal apoplastic region is discussed in relation to functional specialization. Plasma-membrane elaborations or plasmalemmasomes were identified in the guard cells of Zea, and it is suggested that these structures may function in ion transport.
Key words: Cell wall structure - Guard cell Stoma - Plasmalemmasome - Plasmodesma.
Introduction
Studies on symplastic connection between guard cells and adjacent subsidiary or epidermal cells have produced conflicting reports, and the structure of the guard-cell wall and its possible influence on apoplastic ion transport have not been investiAbbreviations: PTA-HCl=phosphotungstic acid and hydrochloric acid; SM =silver methenamine; UA-LC=uranyl acetate and lead citrate
gated. Plasmodesmatal connections between guard cells and adjacent cells were first reported on the basis of light-microscopic studies (Waterkeyn and Bienfait 1966; Litz and Kimmins 1968; Inamdar et al. 1973). The presence of pit callose, fluorescent when stained with aniline blue, was taken as evidence for plasmodesmata in one study (Peterson and Hambleton 1978). The improved resolution available with transmission electron microscopy provides direct means for visualizing plasmodesmata, but most ultrastructural studies reported the absence of cytoplasmic connections in mature guard cells (Brown and Johnson 1962; Srivastava and Singh 1972; Singh and Srivastava 1972; A1laway and Setterfield 1972; Ziegler et al. 1974; Peterson and Hambleton 1978; Rutter and Willmer 1979; Galatis and Mitrakos 1980). Plasmodesmata have been identified in the guard cells of three species, Viciafaba and Nicotiana tabacum (Pallas and Mollenhauer 1972a, b) and Commelina communis (Fujino and Jinno 1972), but these reports are not entirely convincing. In the former study, there is reason to believe, based on the thickness of the cell walls and absence of a well-developed vacuolar system, that the guard cells in which continuous plasmodesmata were demonstrated were, in fact, immature. In the latter study, the quality of the micrographs was insufficient to interpret accurately the images of the alleged plasmodesmata. Continuous plasmodesmata have been observed to connect developing guard cells with adjacent cells, and plasmodesmata that appear truncated have been observed in the walls of mature guard cells (Willmer and Sexton 1979). However, no published micrographs contain both continuous plasmodesmata and a view of the stomatal pore that demonstrates that it had opened for the first time. The resolution of the light microscope is not sufficient for the accurate visualization of plasmo-
130
desmata, while the high resolution obtained with the transmission electron microscope requires sections so thin that several hundred must be observed in order to inspect an entire cell. Serial thick sections have been shown to provide sufficient contrast and resolution to visualize plasmodesmata. Through this technique, it has been possible to carry out a thorough search for plasmodesmata in 20-30 stomata in five plant species. The fact that not a single continuous plasmodesma was encountered provides strong evidence that plasmodesmatal communication does not play an important role in stomatal function. The absence of cytoplasmic communication between guard cells and adjacent cells has important implications in relation to the mechanism of stomatal ion transport. With plasmodesmata being absent, transport of ions and carbohydrates to and from guard cells must be entirely apoplastic. It follows that the guard-cell wall may be involved in specialized interactions with ions as they are transported between guard cells and adjacent cells. Histochemical reagents have been used to demonstrate the differentiation of sieve-element walls into regions with distinct structure and composition (Lucas and Franceschi 1982). We have applied similar techniques to the walls of guard cells to demonstrate structural specialization of various apoplastic regions. We propose that the specific structure of the guard-cell wall has functional implications with respect to intercellular communication between guard cells and epidermal cells, and merits further study. Materials and methods Plant materials and culture techniques. AlIium cepa L. (grown from bulb sets; Coast Brand, Aurora, Ore., USA), Beta vulgaris L. cv. USH10 (Spreckels Sugar Co., Woodland, Cal., U S A ; Lot 3102) and Vicia faba L. cv. Long Pod (W. Atlee Burpee Co., Warmister, Penn., U S A ; Lot 517602) were maintained in a growth chamber under the following conditions: 27~
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells day-17~ night temperatures, 75%-100% relative humidity and a 16-h photoperiod at 180 gmol m -2 s -1. The soil used was a 2:1 : 1 (by vol.) mixture of medium vermiculite, peatmoss and sand, and the plants were watered daily and fertilized with full-strength Hoagland's solution (Hoagland and Arnon 1950) once per week. Fully expanded leaves were harvested when the plants were four to eight weeks old. In addition, young leaves of B. vulgaris were harvested when less than 2 cm long for observation of developing guard cells. Allium cepa L. cv. White Lisbon (Northrup King Seeds, Fresno, Cal., USA) was grown from seed in vermiculite, and the seedlings were maintained under constant light at 52 gmol m - 2 s-1. Cotyledons were harvested after 5 d, and the region between the hook and the terminally attached seed coat was prepared for observation of developing guard cells (Palevitz and Hepler 1976). Phaseolus vulgaris L. cv. Light Red Kidney (University of California, Davis, California Crop Improvement), Nicotiana rustica L. (U.C. Davis, Botany Department Greenhouse Facility), Commelina coelistis L. (Kew Gardens, Richmond, Surrey, UK), and Zea mays L. (A632 x [C3640 x Ohio 43] ; Crow's Hybrid Corn Co., Milford, Ill., USA), and Allium porrum L. (grown from bulbs) were grown and maintained in our greenhouse facility where they were watered three times per week with 0.5-strength Hoagland's solution.
Tissue preparation. Tissue from fully expanded leaves was cut with a tissue slicer (Mickle Laboratory Engineering Co. ; Gomshall, Surrey, UK) into 1.0 mm x 0.3 mm sections. After cutting, the tissue was immersed immediately, at room temperature, in a fixative containing glutaraldehyde (2%), depolymerized paraformaldehyde (3%, for Allium spp.) and sodium-cacodylate buffer (50 mM, pH 7.2). Vacuum infiltration was occasionally employed to facilitate penetration of the fixative. Following fixation (3-10 h, 4 ~ C), tissue was post-fixed with 1% osmium tetroxide in sodium-cacocylate buffer (25 mM, pH 7.2) for 1-3 h, dehydrated through a graded acetone series, and embedded in Spurr's resin (Spurr 1969). Serial thick sectioning. To determine with certainty the presence or absence of plasmodesmata, guard cells were serial-sectioned, and each section was carefully examined for plasmodesmata. To decrease the number of sections required to serial-section a pair of guard cells completely, sections were cut at a thickness which produced green to red interference colors (approx. 60(~800 nm). The stomata of Vicia faba, Nicotiana tabacum, Allium cepa, Zea rnays, and Beta vulgaris were serial-sectioned with a diamond knife on a Porter-Blum MT-2 ultramicrotome (Ivan Sorvall, Newton, Conn., USA). After sectioning, the specimens were mounted on 0.5 m m x 2 . 0 m m slotted Cu-Rh,
)
Figs. 1-9. Plasmodesmata between leaf cells of various species. Sections were cut at a thickness that produced red-green interference colors (600-800 nm) and were viewed unstained at 100 kV. Fig. 1. Plasmodesmata in primary pit field between mesophyll cell (above) and epidermal cell (below) in Allium cepa. x 15,000; b a r = 1 gm. Fig. 2. Plasmodesma between epidermal cells of Vieia faba. x 40,000; bar=0.25 gm. Fig. 3. Plasmodesmata in a primary pit field between subsidiary cell (above) and epidermal cell (below) of Zea mays. x 40,000; bar = 0.25 gm. Fig. 4. Truncated plasmodesmata between guard cell (below) and epidermal cell (above) of V. faba. x 40,000; bar = 0.25 gm. Fig. 5. Truncated plasmodesmata between guard cell (below) an epidermal cell (above) of Nicotiana tabacum, x 25,000; bar=0.5 lain. Fig. 6. Truncated plasmodesmata between guard cell (below) and epidermal cell (above) of Beta vulgaris, x25,000; bar=0.5 gm. Figs. 7-9. Serial thick sections of the guard cell wall of V. faba demonstrating that plasmodesmata are indeed truncated, x 15,000; bars = 1 gm
Figs. 11)-12. Truncated plasmodesmata between epidermal cells (top) and guard cells (below) in various species Sections post-stained with uranyl acetate and lead citrate; bars = 0.1 gm. Fig. 10. Truncated plasmodesma in guard-cell wall of Beta vulgaris, x 100,000. Fig. 11. Guard-cell plasmodesmata in V.faba. x 100,000. Fig. 12. Guard-cell plasmodesmata in N. tabacum, x 80,000
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells
131
132
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells
A.C. Wille and W.J. Lucas: Ultrastructural studies on guard cells Formvar-coated grids (Ernest F. Fullam, Schenectady, N.Y., USA). Unstained sections were scanned for plasmodesmata at 100 kV on a JEOL (Tokyo, Japan), model 100S transmission electron microscope. Carbohydrate histochemistry. For general ultrastructural observation, silver sections were cut with a diamond knife on a 2128 Ultratome (LKB Instruments, Washington, D.C., USA) and mounted on uncoated grids. Sections were stained with lead citrate (Reynolds 1963) and uranyl acetate (2% in 50%, v/v, methanol in water). Gold sections mounted on nickel grids were oxidized by floating on 1% periodic acid in water for 20-30 min at room temperature and stained by floating, for 1 h, in the dark, at 60 ~ C, on drops of staining solution containing 0.1% silver nitrate and 3% hexamethylenetetramine titrated to pH 9 with 5% sodium borate (Rambourg 1967). The grids were then rinsed with distilled water and floated on 0.5% aqueous sodium-thiosulfate solution for 5 rain. Controls were run to demonstrate the specificity of the silver-methanamine reaction. Sections were oxidized with 10% hydrogen peroxide (15-30 min) rather than periodic acid, before incubation in silver-methanamine medium; alternatively, sections were stained without oxidation. Gold sections mounted on uncoated nickle grids were bleached (to remove osmium) by floating the grids on drops of 10% hydrogen peroxide for 15-30 min at room temperature. After washing in distilled water, the grids were floated on drops of staining solution (1% phosphotungstic acid in 10% hydrochloric acid; Hayat 1975) for 10 min in the dark.
Results
Visualization of plasmodesmata in thick sections Since it is not standard technique to observe plasmodesmata with the transmission electron microscope in thick sections (600-800 nm), it was necessary to verify that they can be unambiguously identified under these conditions. Epidermal cells (Figs. 1-3) were used to demonstrate that plasmodesmata are clearly visible in thick sections viewed at 100 kV. While resolution was generally poor, contrast was quite high, even without post-staining. Four sets of serial sections (two paradermal and two transverse) were cut from epidermal tissue of each of five species; Viciafaba, Nicotiana tabacum, Allium cepa, Zea mays and Beta vulgaris. Blocks sectioned in paradermal orientation were
133
approx. 0.5 mm 2, contained three to six stomata, and required 3 0 4 0 sections for a complete epidermal series. Transverse sections of comparable size contained one to three guard-cell pairs, and a complete series was represented in approx. 50 sections. Occasionally some sections were lost as a consequence of separation from the ribbon or breakage of the Formvar film, but never more than five sections, and in each species (except V. faba) at least one complete series was preserved. A total of eight to ten stomata from at least two different leaves from each species were thoroughly examined, and not a single continuous plasmodesma was observed to connect a guard cell to its sister guard cell or to an adjacent epidermal or subsidiary cell. The starch-lacking species, A. cepa, lacked continuous plasmodesmata, as did the four starch-containing species. While continuous plasmodesmata were never observed in mature guard cells, truncated plasmodesmata were common and appeared distinct (Figs. 4-6). Sections viewed in series (Figs. 7-9) demonstrate that the truncated appearance is not an artifact of oblique planes of sectioning. In addition, guard-cell plasmodesmata were always truncated by the interior region of the guard cell wall; further evidence that they were not merely sectioned at oblique angles. Finally, thick sections were approximately ten times the thickness of the average plasmodesma, making it highly unlikely that the plasmodesmata were not contained within a single section. The detailed structure of truncated plasmodesmata was viewed in silver sections (Figs. 10-12). Some appeared singly (Fig. 10) while others occurred in clusters and were connected by a characteristic sinus (Figs. 11, 12). The sinus was occasionally associated with a relatively electron-dense region of the wall (Fig. 11) which may indicate the location of the middle lamella. The guard-cell wall was typically thicker toward the guard-cell side of the sinus giving the overall indication that plasmodesmata had been truncated by deposition of the guard-cell wall (Figs. 11, 12). Truncated plasmo-
