1111 1111 11111111111 .,.._.,..,.

1293 01063 0451
lbkAKY

Michigan nStan:
Univemity

rHESlS

 

14.7
flmW’G-Bf \W‘PH‘ m- .

This is to certify that the

thesis entitled

MICROSPOROGENESIS IN HORDEUM COMPRESSUM:
MICROSPORE AND TAPETUM FROM
THE TETRAD STAGE T0 POLLEN MITOSIS

 

presented by

Joanne Holland Whallon

has been accepted towards fulfillment
of the requirements for

Doctor of Philosophy Botany & Plant Pathology

degree in

% Z222 <7Z‘

Major professor

 

 

 

Date N190”. Al/7 77

0-7639

 

\

 

 

OVERDUE FINES:
25¢ per day per item

RETURNING LIBRARY MATERIALS:

Place in book return to remove
charge from circulation records

 

 

C) Copyright by
JOANNE HOLLAND NHALLON

1979

 

 

MICROSPOROGENESIS IN HORDEUM COMPRESSUM:
MICROSPORE AND TAPETUM FROM
THE TETRAD STAGE T0 POLLEN MITOSIS

By

Joanne Holland Whallon

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Botany and Plant Pathology

l979

 

(,1

(-1 // a W

A

ABSTRACT

MICROSPOROGENESIS IN HORDEUM COMPRESSUW
MICROSPORE AND TAPETUM FROM
THE TETRAD STAGE T0 POLLEN MITOSIS

By

Joanne Holland Whallon

Microsporogenesis from the tetrad stage until the time of
the first pollen mitosis was investigated in Hordeum compressum
by light microscopy as well as by transmission and scanning electron
microscopy. fl. compressum is the first barley species, and the
first wild grass species, to be so investigated. It was found
that both microspore and tapetum bear a strong general resemblance
to the microspore and tapetum of other grasses for which comparable“
information is available. The development of the microspore wall,
however, appears to differ in several details from that reported
for other grasses, and may represent a different ontogenetic pathway.
In the tapetum, differences in the timing of senescence compared
to the cultivated species was apparent. A previously unreported
type of cytoplasmic structure was observed in some tapetal cells.
The ultrastructure of the developing tapetal orbicular wall supports
the hypothesis that the wall is an evolutionary specialization
rather than either a vestigial capacity of a tissue that was once
sporogenous, or a random, nonspecific configuration of excess

sporopollenin.

ACKNOWLEDGMENTS

The successful completion of this dissertation has been made
possible by many people, and to all of them I extend my sincere
thanks. I am particularly grateful to the members of my committee:
to Dr. William Tai, my major professor, for introducing me to
the study of ultrastructure, for permitting me then to wander
far afield from his major interests, and for his constant encouragement;

to Dr. Gary Hooper for patient instruction and enlightening discussions;

 

and to Drs. Clifford Pollard and William Hooker for long years
of interest and support.
Mrs. June Mack provided me with much valuable advice on the
fine points of transmission electron microscopy.
Plates 47-5l were taken by Royce Hazard, who also instructed
me in the use of the scanning electron microscope.
The pollen fertility assessment was done by Dr. Rye-Ho Huang.
Part of this work was supported by a grant from the Michigan
State University Electron Optics Committee, and the work was made
much easier by the purchase of equipment for my use by the Department

of Botany and Plant Pathology.

TABLE OF CONTENTS

List of Figures .................................................. iv
Introduction ..................................................... l
Materials and Methods ............................................ 4
Results .......................................................... 6
Pollen Fertility ............................................ 6
Light Microscope Observations of the Gametophyte ............ 6
Tetrad Stage ............................................. 6

Early Vacuolate Stage .................................... 9
Midvacuolate Stage ....................................... 9

Late Vacuolate Stage/Pollen Mitosis ...................... 10

Light Microscope Observations of the Tapetum ................ l2
Electron Microscope Observations of the Gametophyte ......... l3
Tetrad Stage ............................................. l3

Early Vacuolate Stage .................................... 24
Midvacuolate Stage ....................................... 30

Late Vacuolate Stage/Pollen Mitosis ...................... 34
Electron Microscope Observations of the Tapetum ............. 4O
Cytoplasm ................................................ 40

Cell Walls ............................................... 5l

Scanning Electron Microscope Observations of Whole Anthers.. 61

Discussion ....................................................... 68
Summary .......................................................... 94
Bibliography ..................................................... 96

LIST OF FIGURES

Figure Page

l-8. Stages of microspore development, phase contrast

 

microscopy .............................................. 8
l. Tetrad stage.
2. Tetrad stage.
3. Early vacuolate stage.
4. Early vacuolate stage.
5. Midvacuolate stage.
6. Midvacuolate stage.
7. Late vacuolate stage/pollen mitosis.
8. Late vacuolate stage/pollen mitosis.

9. Tetrad of microspores encased in callose ................... l5
10. Two microspores of a tetrad ................................ l8
ll. Microspore with new wall developing under special callose

wall ....................................................... l8
l2-l5. Early wall development in tetrad stage .................. 21

12. Vesicles lying outside the plasma membrane.

13. Microspore with paramural body.

l4. Fibrils extend from area of plasma membrane toward
callose.

15. Microspore with vesicles, fibrillar deposits and some
amorphous material.

l6-l8. More advanced wall development in tetrad stage .......... g 23
T9. Exine of early vacuolate microspore ........................ 23
20. Early vacuolate microspores ................................ 26
22l. Early midvacuolate microspore .............................. 29
22-23. Microspores at time of pollen mitosis ................... 32
24-26. Stages of exine development ............................. 36

24. Early midvacuolate stage.
25. Late midvacuolate stage.
26. Pollen mitosis.

27-29. Stages in pore development ............................... 39
27. Early vacuolate microspore.
28. Midvacuolate microspore.
29. Microspore at pollen mitosis.

30-38. Tapetal maturation.

30-3l. Tapetum at microspore tetrad stage ....................... 42
32. Early vacuolate microspore stage ............................ 45
33. Early midvacuolate microspore stage ......................... 45

 

34-36. Detail of tapetal cytoplasm at early midvacuolate stage.. 47

34. Overview of cytoplasm.
35. Cytoplasmic plate complex enlarged from Figure 34.
36. Adjacent cells connected by cytoplasmic band.

37-38. Stages in tapetal degeneration ........................... 50

37. Late midvacuolate stage.
38. Advanced degeneration at time of pollen mitosis.

39-42. Orbicule development ..................................... 53

39. Late tetrad stage.

40. Late tetrad stage.

4l. Early vacuolate microspore stage.
42. Early midvacuolate microspore stage.

43-46. 'Tapetal membrane' development ........................... 57

43. Late midvacuolate microspore stage.

44. Orbicule at time of pollen mitosis.

45. Tapetal cell at time of pollen mitosis.
46. Detail of crystal-like structure.

47-48. Cross-sections of individual locules ..................... 6O
47. Tapetum not yet degenerate.

48. Tapetum degenerate. 'Tapetal membrane' adhering to
surface of pollen grain.

49. Microspores dislodged from an anther ........................ 63
50. Longitudinal section of anther .............................. 63
5l-52. Orbicules ................................................ 65

5l. Orbicules on the locular face of the tapetum.
52. Tangential section of the locular face of a tapetal

cell.
53. Open anther ................................................. 67
54. Single microspore with attached reticulum ................... 67

55. Diagram of the microspore wall at the time of pollen
mitosis ..................................................... 69

 

vi

Introduction

From the early l860's until the early 1960's, the study of
pollen was largely the province of taxonomists. As a result of
their efforts, descriptions of the pollen morphology of thousands
of species entered the literature. Even with the relatively
limited resolution and depth of field of the light microscope,
it became apparent that the pollen wall is a highly complex structure.

Cytologists augmented the information supplied by the taxonomists,

 

adding some details on the developmental processes.

In recent years pollen studies have taken a new direction.
Not only are we able to see the sculpturing of the exine much
more clearly, but we are also able to observe the ultrastructural
details of its development, to examine the living cell hidden
inside, and to follow developments in associated anther tissues,
particularly the tapetum. In addition, it has become possible
to begin a correlation of biochemical studies with ultrastructural
details.

At least some stage of microsporogenesis has by now been
studied at the ultrastructural level in several species representing
several different families. A partial listing would include ggta

vulgaris (Hoefert, l969), Capsicum annuum (Horner and Rogers,

 

l974), Citrus limon (Horner and Lersten, l97l), Cosmos bipinnatus
(Dickinson, l976), Helleborus foetidus (Echlin and Godwin, l969),

Ipomoea purpurea (Godwin, Echlin and Chapman, l967), Lilium longiflorum

(Heslop-Harrison, 1968b; Dickinson, 1970, 1976), Parkinsonia aculeata

 

(Larson and Lewis, 1962), Pings banksiana (Dickinson, 1971),

Silage pendula (Heslop-Harrison, 1963), Sorghum bicolor (Christensen,
Horner and Lersten, 1972), Tradescantia bracteata (Mepham and

Lane, 1970), and Egg gays (Skvarla and Larson, 1966). Somewhat

less attention has been devoted to the tapetum, but nevertheless
Antirrhinum maius (Lombardo and Carraro, 1976), A!§flé.§2£ll§ (Steer,
l977a,b), Egg; vulgaris (Hoefert, 1971), Helleborus foetidus

(Echlin and Godwin, 1968), Paeonia tenuifolia (Marquardt, Barth

and von Rahden, l968), Pings banksiana (Dickinson and Bell, 1972,
1976), Sgillginon-scripta (Rodriguez—Garcia, 1978) and IEEQEET
cantia bracteata (Mepham and Lane, 1969) are among those in which
this tissue has been investigated.

Subsequent to these ultrastructural studies has come renewed
interest in the adaptive significance of the large numbers of
variations encountered, and a concomitant interest in dysfunctions
that might result in a reduced production of viable pollen. As
a result of these investigations we now have, for example, evidence
that the pollen wall acts as a repository for a number of proteins,
some derived from the gametophyte and others from the sporophyte,
which ensure the success or failure of the pollen-stigma interaction
(Heslop-Harrison, Heslop-Harrison, Knox and Howlett, 1973; Dickinson
and Lewis, 1973; Heslop-Harrison, Knox and Heslop-Harrison, 1974;

Heslop-Harrison, 1976) and we also know that the causes of sterility

 

are many and varied, from improper development of the intine
(DeVries and Ie, 1970) to failure of tapetal mitochondria (Narmke
and Lee, 1977; Lee, Gracen and Earle, 1979).

Barley (genus Hordeum), although it is an economic crop of
considerable significance, has up to this time not been the subject
of such investigations. The present work is an ultrastructural
study of part of the sequence of events of microsporogenesis in
H, compressum, a species that has not undergone heavy selection

pressure for cultivation.

Materials and Methods

 

Eleven cloned plants of Hordeum compressum were maintained

 

in field and greenhouse over a three-year period. Spikes which
had not yet begun to emerge from the boot were cut, stripped of
enclosing boot tissues and plunged intact into fixative. Occasionally
anthers were dissected out before fixation, but as this did not
noticeably improve fixation and did greatly complicate the process
of following the developmental sequence, it was not done as a
general rule.

Because satisfactory fixation of grass anthers is difficult
to achieve, a variety of different fixatives and buffers was
tried. The most successful results were obtained with a solution
of 5% glutaraldehyde either in 0.1M phosphate buffer at pH 7.4
with 6-10% sucrose, or in 0.01M phosphate buffer at pH 7.4 with
6 drops of 0.1M CaCl2 per 40 ml. of buffer. Fixation was carried
out in the cold for 10-18 hours. Following two 30 minute rinses
with buffer, spikes were postfixed in 1 or 2% 0504 for 4-12 hours,
rinsed in water and dehydrated either in a graded acetone series
(30 minute steps) or, for scanning electron microscopy (SEM),
with three five-minute rinses of acidified dimethoxypropane (Johnson,
Hooper, Holdaway and Rasmussen, 1976).

Anthers prepared for transmission electron microscopy (TEM)
were embedded in Epon, sectioned with a diamond knife on a Reichert

ultramicrotome, double-stained with uranyl acetate and lead citrate

and viewed in either a Philips 300 or a Philips 201 electron microscope
operating at 60 KV. Monitor sections were stained with toluidine

blue and photographed with a Zeiss Photomicroscope II using phase
contrast optics. Anthers prepared for SEM were critical-point

dried in a Sorvall critical-point drying apparatus, mounted on

a stub using Scotch brand double sticky tape, coated with gold

alloy in a sputter coater and viewed in an I. S. I. Super II.

The anthers were mechanically broken open or cut with a razor

blade after mounting on the stub but before gold coating.

What appeared to be developmental abnormalities within the
anthers appeared fairly frequently, compounding the problems
resulting from troublesome fixation. These imperfect anthers
upon study yielded many valuable insights that are indirectly
a part of this dissertation, but their presence made determination
of the events of normal microsporogenesis difficult. The final
decision as to whether a given observation should or should not
be included as normal was thus influenced by the quality of fixation,
by how well the observation in question fitted into a continuum
with earlier and later stages whose identification was more secure,
and by the availability of duplicate observations, but these

criteria could not be applied equally in all instances.

Results

Pollen Fertility

A total of 2,402 grains was collected, stained and counted.

Of these, 1,988 or 78.6%, gave a positive response to IZ-KI.

Light Microscope Observations of the Gametophyte

 

Tetaad Stage

Sections through anthers which had completed meiosis (Figures
1, 2) revealed locules almost completely filled with tetrads in
a variety of non-spherical shapes surrounded by a considerable
amount of callose. The size of the tetrads relative to the size
of the locule ensured that each tetrad was contiguous with the
tapetum. Furthermore, the tetrads displayed some polarity of
orientation, so that many of the individual microspores were also
adjacent to the tapetum. This was particularly evident in cross-
section, where the common wall between two adjacent cells frequently
coincided with the locule radius. In longitudinal section, one
division plane was often more or less parallel to the long axis
of the anther, with the second one perpendicular to it.

In each cell of the tetrad, a centrally located nucleus,
with one or two nucleoli, and some small vacuoles could be seen.
No cytoplasmic connections were visible between any of these

haploid cells.

 

Figures 1-8. Stages of microspore development, phase contrast
microscopy. C, callose, M, microspore, T, tapetum; solid arrow
indicates pore. Bar = 10 pm.

Figures 1 and 2. Tetrad stage. Microspores still enclosed in
callose wall. Open arrow in Figure 2 indicates a tetrad with
one degenerating microspore.

Figures 3 and 4. Early vacuolate microspore stage. Callose has
disappeared, microspores have rounded, up wall is evident. Note
close association of microspores with tapetum (Figure 4 thin
arrows).

Figures 5 and 6. Midvacuolate microspore stage. Microspore in
Figure 5 shows nuclear displacement (open arrow). Orbicules are
visible on inner tapetal face. Two layers of microspore wall
are visible in Figure 6 (open arrow). Tapetal cells in Figure

6 are pulling apart.

Figure 7 and 8. Late vacuolate stage/pollen mitosis. Large
vacuole occupies central area of each microspore, nucleus is
located opposite pore region, tapetum is degenerate; 'tapetal
membrane' lines locule (open arrow).

l or”; or?"
315225.,

   

Eaaty Vacuotate Stage
As the callose began to break down, the individual cells
became progressively more rounded, and by the time the callose
was completely gone, it was difficult to discern any special
orientation of the microspores within the locule. Figures 3 and
4 show early vacuolate microspores. At this stage, no callose
was present. Each microspore was surrounded by a thin, dark-staining
wall. Hordeum pollen, like that of other grasses, has a single
pore, and those microspores in which the pore was visible showed
a 3 dark thickening of the wall at the pore periphery. The wall
seemed to be very malleable, since wherever one microspore came
in contact with another, or with the tapetum, it was often considerably
flattened. The tapetum in such cases showed no indentation or
other signs of pressure from the microspore. Several microspores
were in very close contact with the tapetum. The orientation
of the microspores within the locule, as judged by the direction
in which the pores faced, seemed to be random.
The contents of these young microspores was similar to that
of the microspores in the tetrads, except for an increase in the

size of the vacuoles.

Mtduaeuotate Stage

In midvacuolate microspores (Figures 5, 6), found in somewhat

older and larger anthers, the vacuoles had begun to coalesce,

 

and some nuclei showed signs of lateral displacement. The direction
of nuclear displacement seemed random, with respect both to the
cell itself and to the anther as a whole. The orientation of
the pores at this stage, however, was much less random than previously;
more often than not they faced outward rather than inward. The
wall was much thicker and more intensely stained, and the pores
appeared correspondingly more developed. The microspores seldom
made contact with each other or with the tapetum, presumably owing
to locule enlargement, and when they did they were much less
deformed than in the early vacuolate stage.

Development in the latter part of this stage (Figure 6) was
marked by increasing vacuolization, more nuclear displacement ‘
and additional thickening of the microspore wall. Two layers
of the wall, with an intervening space, could be resolved. A
common sight at this and later stages was the almost total collapse

of some of the microspores, presumably a fixation artifact.

Late Vacuotate Stage - Patten Mttoata

By the time of pollen mitosis (Figures 7, 8), a single large
vacuole filled the center of each microspore, leaving only a thin
layer of cytoplasm adjacent to the plasma membrane. Most of the
nuclei were in late prophase or prometaphase, some were in interphase,
and a few had completed mitosis. Each nucleus had migrated to
a position approximately opposite the pollen pore. The pores

themselves, as well as a considerable amount of adjacent wall

 

surface, were in close contact with the inner lining of the locule.
The pressure of the still-enlarging microspores had pushed aside
the remnants of the tapetal cells, and frequently impinged on

the middle layer, or, more rarely, on the endothecium.

The layers of the pollen grain wall were much thicker than
earlier, and, in contrast to the previous stages, stained dissimilarly,
with the inner layer being much darker than the outer one. The
dark inner layer passed uninterrupted beneath the pore, and in
the pore region, was greatly thickened and protruded into the

pore proper, where the operculum appeared to be resting directly

 

on it.

The maximum number of microspores seen in locule cross sections
was six. Where the microsporangia became narrower, the number
decreased to four, and eventually, at the ends of the anthers,
to one. Although the locule did not appear crowded, packing was
fairly close, such that in all stages the individual microspores
wre probably held in position by the presence of adjacent cells.

From the time of tetrad formation until pollen mitosis, both
microsporangia and microspores enlarged considerably. The anthers
increased in both length and width by at least 100%, while the
diameters of the microspores increased 150 - 200%. From the
latter figures it can be calculated that the volume of the individual
microspores increased 20 to 25 times. These values are only

approximate, since the measurements were made on thick sections

12

of Epon-embedded tissue. The fixation and embedding procedures
cause shrinkage of the anthers, and it is not known what factors
influence the shrinkage, nor which tissues are the most susceptible
to shrinkage.

In the tetrad stage, it was not uncommon to see tetrads in
which one or more of the member cells showed clear signs of degeneration
(Figure 2). After the callose had disappeared, however, all the
microspores of a given locule usually displayed the same characteristics
-- that is, either all appeared to be in good condition, or all
did not, and whatever abnormality was displayed by one was displayed

by all.

Light Microscope Observations of the Tapetum

 

From the time the microspores were in the tetrad stage,
through the midvacuolate stage, the tapetal cells formed a very
prominent layer owing to the fact that, unlike all the other cells
of the anther wall, they were not vacuolated. There was no evidence
of tapetal dimorphism - all the cells were approximately the same
size and shape, and all were binucleate.

The tapetum of Hordeum compressum is of the parietal or secretory

 

rather than the plasmodial or invasive type: that is, at no time
does it separate from the anther wall and become interspersed
with the microspores in the locule. Like all parietal tapeta,

it breaks down at some point after meiosis has been completed.

13

In the light microscope this breakdown first became evident prior

- to pollen mitosis, during the latter part of the midvacuolate

stage (Figure 6), and was evidenced by a somewhat disorganized
appearance of the cells, by their separation from each other,

and by indentations caused by the enlarging microspores. By the
time of pollen mitosis (Figures 7, 8), very little could be seen

of the tapetal cells except patches of dark-staining and apparently
degenerate cytoplasm.

A unique feature of the tapetum in this and some other plants
is the development on the locule face of structures called orbicules
or Ubisch bodies (Ubisch 1927, Kosmath 1927). The first detectable
sign of these in the light microscope was the appearance, at the
beginning of the midvacuolate stage (Figure 5), of small granules
lining the entire locule. These increased in size as the microspores
matured. By the time of pollen mitosis (Figures 7, 8), hollow
centers could be seen in some of them. At that time also, they
appeared to be attached to a thin, dark layer which in some places
seemed to rest on, and in other places had pulled away from, the
remnants of the tapetal cells. This layer, which has been given
the name 'tapetal membrane' (Banerjee, 1967), with its attached

orbicules lined the entire inner wall of the locule.
_§1ectron Microscope Observations of the Gametophyte

Temad Stage

The most prominent feature of the young tetrad cells was

 

l4

Figure 9. Tetrad of microspores encased in callose. C, callose;
L, locule; Mi, mitochondria; Os, osmiophilic bodies; P, plastids.
Bar = 2 um.

l5

   

 

the centrally located nucleus (Figures 9, 10) with two nucleoli

(Figure 10). The chromatin appeared to be in a partially dispersed

state. Nuclear pores were numerous, but no particular pattern
of their distribution was discerned. The cytoplasm was fairly
dense and contained small mitochondria, plastids with little
internal membrane development, free ribosomes, numerous small
vesicles, some rough endoplasmic reticulum (ER) and some smooth
ER. Dictyosomes were infrequent and quite small, consisting of
only three or four plates about one gym in diameter. There were
also in each cell several osmiophilic droplets or bodies around
which a membrane could not be distinguished. These ranged in
size up to onepm in diameter, and did not seem to occupy any
particular location within the cell. Also present were a number

of small organelles whose identity could not be established with

certainty, but which were thought to be promitochondria or proplastids.

The plasma membrane was smooth for the most part; occasionally
indentations suggestive of vacuoles either forming or discharging
were visible there.

Each individual tetrad member, as well as the whole tetrad,
was totally enveloped by callose. The thickness of the callose
sheath between any two adjacent tetrad members was about one um.
No cytoplasmic connections of any kind were seen traversing the
callose walls between these haploid cells. Outside the callose

sheath were more osmiophilic bodies, smaller in size (ca. 500

 

 

Figure 10. Two microspores of a tetrad. CP, remnants of cell
plate; N, nucleus; Nu, nucleolus. Bar = 1 pm.

Figure 11. Microspore with new wall developing under special
callose wall. M, microspore; N, nucleus of tapetal cell (nucleus
of microspore is out of plane of section); Os, osmiophilicbodies;
T, tapetum; W, developing microspore wall; arrow indicates pro-
bable pro-orbicule lying just under tapetal plasma membrane (cf.
text p. 51 and Figures 31, 40). Bar = l um.

l8

 

 

nm) but otherwise similar to those within the microspore cytoplasm.
These were usually seen between the callose and the remnants of
the primary wall which had once separated the sporogenous cells
from the tapetum, and occasionally in other places, such as between
adjacent tetrads (Figure 11).

The tetrads which were examined fell into one of two groups.
In the first and presumably younger group a close examination
of the area immediately exterior to the plasma membrane (Figure
10) revealed that the callose sheath was in direct contact with
the plasma membrane. Some sections contained membranes or what
appeared to be small membrane-bound vesicles or both, just outside
the plasma membrane. Around the tetrads in the second group
however (Figure 11), the callose was not in direct contact with
the plasma membrane, but was separated from it by a narrow electron-
translucent space within which a variety of structures could be
seen. Some cells had collections of membrane-bound vesicles in
this area; others had what looked like thin fibrils extending
from the plasma membrane toward the callose; still others had
both vesicles and collections of fibrils (Figures 12-15). Paramural
bodies (i.e. multivesicular bodies adjacent to the wall) were
also seen (Figure 13). In most instances, however, this space
between the plasma membrane and the callose wall was largely
filled except for some lateral discontinuities by a layer of amorphous,

osmiophilic material (Figures 16-18). The source of this material

 

 

20

Figures 12-15. Early wall development in tetrad stage. C, callose;
F, fibrils; PB, paramural body; Ve, vesicles. 0.2 Bar = 0.2 pm.

Figure 12. Vesicles lying outside plasma membrane. Note thin
horizontal line between open arrows (see discussion), here and
in Figure 14

Figure 13. Microspore with paramural body.

Figure 14. Fibrils extend from area of plasma membrane toward
callose.

Figure 15. Microspore with vesicles, fibrillar deposits (thin
arrows) and some amorphous material (open arrow).

 

21

 

 

22

Figures 16-18. More advanced wall development in tetrad stage.
C, callose; M, microspore; possible development of two (open
arrow) and three (closed arrow) layers. Bar = 0.2 pm.

Figure 19. Exine of early vacuolate microspore. Ca, cavus; FL,
foot-layer; L, locule; S, spinule; Te, tectum; thin arrows in-
dicate channels in foot-layer, heavy arrow indicates lamella.
Bar = 0.2 um.

23

. ///,

w/n

fl

 

24

was not apparent, but the cells which possessed it were often
more highly vacuolated than those which did not. The stellate
configuration of the vacuoles, however, indicated the possibility
of a fixation artifact (Matile, 1975).

Some cells had much thicker accumulations of this osmiophilic
material than other cells. Occasionally profiles were seen which
were suggestive of the development of a second osmiophilic layer

separated from the first by a non-osmiophilic area (Figure 18).

Eaaty Vacuotate Stage

By the time the microspores had reached the early vacuolate
stage (Figure 20), no remnants of callose were seen. The nucleus
of each cell was still centrally located. The chromatin was
perhaps slightly more dispersed than at the previous stage. ER
profiles were somewhat more frequent, and sometimes appeared in
pairs. There were many more vacuoles of all sizes.

Surrounding each microspore at this stage was a very definite
rudimentary exine (Figures 19, 20) consisting of two layers, an
outer tectum and an inner foot-layer, separated by an intervening
space, the cavus. Spinules protruded above the tectum, and columellae
were seen to continue inward from some of the spinules and to
end in the foot-layer. The width of this wall, from the base
of the foot-layer to the top of the tectum, but not including
the spinules, measured between 80 and 100 nm; the spinules protruded

above the tectum an additional 80 nm. Both tectum and foot-layer,

25

Figure 20. Early vacuolate microspores. ER, (rough) endoplasmic
reticulum; L, locule; 0, orbicule; Po, pore; T, tapetum; V, vacuole.
Bar = 2 um. ~

 

 

26

/

2%

Maw/2..., .,

/ /./z a .,
”Wee/7.

fl . /
,7 w, y/fl/Hfl/K

 

27

but particularly the former, had an irregular and discontinuous
appearance. The inner edge of the foot-layer (i.e., the side
toward the plasma membrane)was very smooth and regular, and seemed
in some places to be demarcated by a faint layer about 25 A wide.
Also at the base of the foot-layer smooth lamellar structures

were frequently observed which sometimes seemed to be in the
process of becoming embedded in the foot-layer. The width of
these lamellae approximated that of the plasma membrane. Although
it constituted the innermost part of the wall at this stage, the
foot-layer was only rarely in direct contact with the plasma
membrane; for the most part the two were separated by an uneven
space of moderate electron density. The plasma membrane presented
a smooth profile.

At the pore area, the foot-layer branched into a multi—layered
structure (Figures 20, 27). Each layer consisted of a lamella
thickened on both sides with an electron-dense material; the more
externally positioned lamellae had greater amounts of this encrusting
material. The lamellae did not extend across the center of the
pore area, but met and curved downward, leaving a clear, undecorated
area, the 'pore membrane' about two pm in diameter. In the center
of the 'pore membrane' a small decoration, the incipient operculum,
could be seen. The lamellae were located in, or probably caused,

a raised ring or annulus around the area of the 'pore membrane'.

Beneath the 'pore membrane' and the annulus the plasma membrane

28

Figure 21. Early midvacuolate microspore. D, dictyosome; ER,
(rough) endoplasmic reticulum; Mi, mitochondria; N, nucleus; 0,
orbicule; T, tapetum; V, vacuole; evagination of nuclear membrane
at open arrow. Bar = 2 pm.

 

29

. , 7 /// x,
I ..//fl/;

4/ t, ,

./ / ////

#4 //yy/////
/

.2 7/

.21....

//fl,

.7. ,
$4.1.

/

. . , ,2
awe/.25 22/ . .

«7.6.
2
.46..
.1.

 

 

 

 

 

 

 

_30

was considerably withdrawn, and the space thus created was filled
with material continuous with and apparently identical to the
previously mentioned material of moderate electron density which
underlay the non-apertured wall, and additionally in some cases

with many small vesicles.

Mtdvaeuotate Stage

In those microspores which had progressed to the midvacuolate
stage (Figure 21), there were still a large number of very small
vacuoles, but there were also some much larger vacuoles, and some
of these appeared to be autophagic. The nuclei, frequently somewhat
laterally displaced, contained very fine chromatin fibrils.

Nuclear evaginations were quite common. Sometimes it was difficult
to discern whether these evaginations made connections with any
other organelle or not; at other times they were extensive enough
to have attached ribosomes, and occasionally they contained within
themselves structures consistent with the idea that budding of
mitochondria or plastids might be occurring (Figure 21, open
arrow). The cytoplasm of these cells also contained strongly
osmiophilic bodies, but unlike the ones seen in the tetrads, these
were not smooth in outline.

At this stage (Figure 24) the thickness of the wall was almost
triple what it had been in the early vacuolate stage (Figure 19),
an increase which was observed in foot-layer, cavus and tectum,

but not in spinules. From the base of the foot-layer to the top

31

Figures 22 and 23. Microspores at time of pollen mitosis. Ch,
chromatin; Mi, mitochondria; N, nucleus; Nu, nucleolus; Os, osmio-
philic body; Po, pore; TM, tapetal membrane; V, vacuole. Bar

= 2 11m.

 

 

33

of the tectum measured about 250 nm. The spinules were no larger
than previously and again projected about 80 nm above the tectum.
The tectum no longer seemed to be discontinuous, but it was interrupted
at frequent intervals by channels about 50 nm in diameter connecting
the external environment of the cell with the cavus. These channels
were almost always unbranched and occupied about one-third of
the interspinule area. The spinules themselves were not channelled,
nor, usually, were the areas of the tectum immediately beneath
them. The cavus itself was electron-translucent except for a
thin continuous layer in its approximate center, and except for
the columellae connecting tectum and foot-layer. The foot-layer,
about three fourths as thick as the tectum, contained only a few
channels. The thin line subtending the base of the foot-layer
was much darker than in the early vacuolate stage, and was continuous
under the whole foot-layer, even under the channels.

The positions of the foot-layer and the plasma membrane
relative to each other were about the same as in the previous
stage. The plasma membrane was still smooth, but many small vacuoles
could be seen lying immediately beneath it (Figures 21, 28).

In the early vacuolate stage (Figure 27) the encrusting
deposits on the pore lamellae were about the same thickness as
the tectum and foot-layer. By the midvacuolate stage, however,
the outer three or four lamellae had become much thicker than

the tectum (Figure 28). The entire pore area including the annulus

34

had increased in width and depth, but the number of individual
lamellae remained the same. Although the tectum of the annulus
bore the same pattern of channels as did the rest of the tectum,
the encrusted lamellae differed from the foot-layer in that they
contained no channels at all.

As the midvacuolate stage progressed, the exine nearly doubled
in width, to about 400 nm (Figure 25). Again, this increase
reflected changes in size of tectum, cavus and foot-layer, but
not spinules. The channels in the tectum and foot-layer had not
narrowed. The dark line at the base of the foot-layer was still
apparent, even under the channels. Both the inner side of the
tectum and the outer side of the foot-layer were much more irregular
than previously. The cavus seemed to contain some fibrillar

material.

Late Uaeuotate Stage - Patten Mttoata
The very large, centrally located vacuole characteristic
of this stage pressed all the other cell organelles against the
pollen grain wall (Figures 22, 23). The cytoplasm thus compressed
was heavily populated with organelles, including large numbers
of dictyosomes, and some osmiophilic deposits. The nucleus, surrounded
by a thin layer of cytoplasm, protruded into the vacuole. No
unusual chromosomal features were apparent.
The exine at this stage measured about 700 nm in thickness

(Figure 26). Except for a thin layer lining the cavus, the channels

 

35

Figures 24-26. Stages of exine development. Ca, cavus; Col,
columella; FL, foot-layer; G, granular material between plasma
membrane and foot-layer; L, locule; PM, plasma membrane; S, spin-
ule; Te, tectum. Bar = 0.2 um. Note thin dark line subtending
foot-layer in Figures 24 and 25 (arrows).

Figure 24. Early midvacuolate stage.
Figure 25. Late midvacuolate stage.

Figure 26. Pollen mitosis.

/

.2 WWW/7 1'
15%4/fi/2’1/j
MW

. I
/ //

«4,

”a. a
re /

#1,
9/

36

 

saw" u

37

and the outer side of the tectum, it was noticeably less osmiophilic
than in earlier stages. All the surfaces were smoother; the spinules,
although still pointed at their tips, protruded much less angularly.
Some of the channels appeared to be occluded, but it may be that
they merely passed out of the plane of the section. Although
many of the channels seemed to be narrower than those of the
earlier stages, that appearance may also have been the result
of their passing out of the plane of section. The contents of
the cavus were coarser than earlier, and the columellae were much
thicker. The space between the base of the foot-layer and the
plasma membrane had narrowed, and was filled with a granular
substance of moderate electron density. This same granular substance
continued into the area under the pore, filling the entire area
between the plasma membrane and the 'pore membrane' (Figure 29).
It must correspond, therefore, to the darker component of the
pollen wall that was observed in the light microscope (Figures
7,8). (Thus, the inner of the two layers in Figures 7 and 8 is
apparently not the same as the inner layer in Figure 6. The dark
inner layer in Figures 7 and 8 probably represents the beginning
of the endexine, which by the intensity of its stain is obscuring
the foot-layer, i.e., the inner of the two layers visible in Figure
6.)

In the pore area (Figure 29) the lamellae were so thickened

as to have almost obliterated the spaces between them, although

 

38

Figures 27-29. Stages in pore development. G, granular material;
L, locule; Lm, lamellae; O, orbicule; Op, operculum; PM, plasma
membrane; Te, tectum. Bar = 1 11m.

Figure 27. Early vacuolate microspore.

Figure 28. Midvacuolate microspore.

Figure 29. Microspore at pollen mitosis.

 

39

 

40

The space between the outermost lamella and the tectum remained.
The pillow of granular material (endexine?) on which the operculum
rested had protruded into the pore so that the top of the operculum

was almost level with the top of the annulus.

Electron Microscope Observations of the Tapetum

 

CytoptaAm

As microsporogenesis was proceeding from the tetrad stage
to pollen mitosis, the most striking feature of the tapetal cytoplasm
was the abundant rough ER (Figures 30-35). There was a very heavy
population of free ribosomes as well, some of which were in a
helical rather than a polysome configuration. Dictyosomes were
common, particularly at the early midvacuolate stage. Mitochondria
were abundant and always in the 'orthodox' configuration. Plastids,
however, were rather poorly developed. Although large in size
and many in number, each individual plastid had very few internal
membranes and was largely filled with an amorphous substance of
medium electron density. Occasional inclusions, some electron-
dense and some electron-transparent, were contained within some
of the plastids. Although there were none of the large vacuoles
characteristic of most plant cells, these cells did possess a
considerable number of small vacuoles or vesicles. The two tapetal

nuclei were usually more or less centrally located on the cell‘s

 

41

Figures 30-38. Tapetal maturation.

Figures 30 and 31. Tapetum at microspore tetrad stage. C, callose;
ER, (rough) endoplasmic reticulum; M, microspore; ML, cell of
middle layer; N, nucleus; P, plastid; incipient orbicules at heavy
arrows. Bar = l um.

42

 

W,
2...:

f

 

43

long axis. The proportion of euchromatin to heterochromatin
(i.e., dispersed vs. condensed) in these nuclei was about the

same as that of the other cells in the anther wall, and much lower
than that in the microspores.

Although none of the various types of organelles was confined
exclusively to one area of the cell, their distribution was not
totally random. In the early tetrad stage (Figure 30), the stacks
of rough ER tended to enclose pockets of cytoplasm containing
mitochondria and plastids. By the early vacuolate stage (Figure
32) this arrangement was no longer so apparent, and by the midvacuo-
late stage (Figures 33, 34) nearly every plastid seemed to be
closely associated with, and sometimes almost surrounded by, an
ER cisterna. These cisternae were often smooth, but sometimes
rough and occasionally bifacial.

Some of the features seen in the early midvacuolate stage
are illustrated in more detail in Figures 34-36. The smooth
cisternae ensheathing the plastids, and cytoplasmic connections
between cells, are prominent. Unusual subcellular structures,
which appeared to be a combination of cytoplasmic plates and
flattened vesicles, were observed in several tapetal cells of
this anther. These structures have tentatively been named 'cytoplasmic
plate complexes.’ They consisted of flat or curved cytoplasmic
plates surrounded by a flattened sac whose boundaries paralleled

those of the cytoplasmic plates (Figures 34-36). At their ends

44

Figures 32 and 33. Tapetum at early vacuolate and early midvac-

uolate microspore stages. D, dictyosome; ER, (rough) endoplasmic
reticulum; L, locule; Mi, mitochondrion; N, nucleus; 0, orbicule;
P, plastid. Bar = 1 pm.

Figure 32. Early vacuolate microspore stage. Cytoplasm of ad-
jacent cells confluent at heavy arrow.

Figure 33. Early midvacuolate microspore stage. Remnants of
primary wall at double arrows.

45

 

 

 

46

Figures 34-36. Detail of tapetal cytoplasm at early midvacuolate
stage. CPX, cytoplasmic plate complex; 0, dictyosome; eC, ensheath-
ing cisternae; ER, (rough) endoplasmic reticulum; hP, helical
polysome; Mi, mitochondrion; NP, nuclear pore; P, plastid; Pl,

plate area of cytoplasmic plate complex; Sa, sac or cisternal

area of cytoplasmic plate complex; V, small vacuole. Bar = 0.5

pm.

Figure 34. Overview of cytoplasm.
Figure 35. Cytoplasmic plate complex enlarged from Figure 34.

Figure 36. Adjacent cells connected by cytoplasmic band (arrow).
Cytoplasmic plate complex in upper cell.

47

 

48

the sacs were dilated and frequently continuous with larger vesicles
which usually appeared to contain some fibrillar material. Occas-
ionally some of the sacs seemed to connect with the ER (Figures

34, 35). The membranes of both the plates and the surrounding

sacs were trilaminar, as were those of the larger vesicles; the
non-cytoplasmic face of these membranes (and of the plasma membrane)
stained more intensely than the cytoplasmic face did (Figure35).
The width of the plates (about 30 nm) and of the cisternal areas
(about 15 nm) was very regular, much more so than that of the

ER and dictyosomes in these cells. There was a suggestion of

some substructure, perhaps running at right angles to the long

axis, in both the cytoplasmic and non-cytoplasmic areas. There
were no ribosomes within or adjoining any part of the complexes.

The frequency of these complexes was low, since in any one
thin section there were always cells which seemed to have none,
and two was the maximum number seen in a single section of a
single cell. No attempt was made to determine by serial sections
the total number present in an entire cell, or whether there were
any cells that had none at all.

Tapetal degeneration began to be evident in the late mid-
vacuolate stage (Figure 37) with the appearance of 'myelin' figures
between the radial walls. At that time there was a great increase
in the number of plastids and some decrease in the number of mitochon-

dria. By the time of pollen mitosis (Figure 38) the tapetal

 

49

Figures 37 and 38. Stages in tapetal degeneration. D, dictyo-
some; E, microspore exine; En, endothecial cell; ER, (rough)
endoplasmic reticulum; Mi, mitochondria; My, myelin figure; N,
nucleus; 0, orbicule; P, plastid; TM, tapetal membrane; V, vacuole.
Bar = 1 pm.

Figure 37. Late midvacuolate stage.

Figure 38. Advanced degeneration at time of pollen mitosis.

W

' 22

/

 

51

cytoplasm had been reduced to about one-tenth of its former amount.
The plasma membrane did not rupture, and some organelles within
the cells were still identifiable.
Premature degeneration and other abnormalities were not as
common in the tapetum as in the microspores. No examples were
seen of degeneration which was only premature: all prematurely
degenerating tapetal cells displayed a type of degeneration different 1
than that which normally began in the late midvacuolate stage. i
Abnormal tapetal degeneration was not always accompanied by abnormal
microsporogenesis; sometimes the developing microspores appeared

normal (Figure 9).

Cett Waite

At the end of meiosis II, thin primary walls were present
at the radial and outer tangential surfaces of the tapetal cells
(Figure 31). The radial walls subsequently disintegrated (Figures
32, 33) although remnants of the middle lamella frequently persisted
at their distal ends, and broad cytoplasmic connections were estab-
lished between adjacent cells, making the tapetum a syncytium
(Figures 32, 33, 36, 37). The outer tangential plasma membrane
showed a tendency to withdraw at intervals from its wall, leaving
pockets which appeared empty except for the presence of an occasional
vesicle (Figure 33).

The course of development of the inner tangetial wall was

much more complex. At the end of meiosis II this wall appeared

52

Figures 39-42. Orbicule development. C, callose; L, locule;

M, microspore; N, nucleus; NP, nuclear pore; 0, orbicules; PW,
remnant of primary wall; T, tapetum; TR, tapetal reticulum. Bar
= 0.2 pm.

Figure 39. Late tetrad stage.

Figure 40. Late tetrad stage. Note electron-dense material at
orbicule surface (arrow). Microspore has some wall development.

Figure 41. Early vacuolate microspore stage. Orbicule has prom-
inent spinules.

Figure 42. Early midvacuolate microspore stage. Orbicule with
spinules and channels. Incipient tapetal reticulum on locular
face of tapetal plasma membrane.

 

53

Vfl/y/(yh

aw.

i/

.747

27642 2... .

//,// a,

W/
94/ /

;%

 

 

54

much thicker than any of the other tapetal walls, but appeared

to be undergoing a process of dissolution and was fibrillar and
granular in appearance (Figures 39, 40). The middle lamella
remained intact, and as mentioned previously, osmiophilic bodies
were trapped between it and the callose of the microspores. The
middle lamella subsequently disappeared, but the fibrillar remnants
of the rest of the wall were still apparent even at the time of
pollen mitosis (Figures 33, 37, 38).

Shortly after the tapetal wall began disintegrating, the
tapetal plasma membrane acquired a very active configuration,
with a large number of pockets or depressions. Small homogeneous
droplets or spherical bodies of a medium electron density about
150 nm in diameter were present in many of these depressions
(Figure 39). These droplets appeared to have a boundary of some
sort, but not a trilaminar one. It was not possible to identify
any comparable structures within the cytoplasm.

At the late tetrad stage, when wall development was apparent
in the microspores, the depressions in the tapetal plasma membrane
contained spherical bodies about 200 nm in diameter which were
electron-translucent (Figure 40). Many gave the appearance of
having been recently expelled from the cytoplasm, and occasionally
one could be seen lying just beneath the plasma membrane (Figure
11). Small patches of a moderately electron-dense material were

frequently seen close to the surface of these droplets (Figure

55

40).

By the early vacuolate stage, the spherical bodies measured
about 300 nm, and were still electron-translucent (Figure 41).
Accretions of electron-dense material, presumably sporopollenin
or protosporopollenin, decorated the periphery of all the droplets
identifying them with certainty as developing orbicules. In size
and shape these accretions resembled the spinules on the developing
exine in the same anthers. The subsequent development of these
orbicules closely paralleled that of the tectum of the developing
microspore exine. In the early midvacuolate stage, the orbicular
covering, 60 to 100 nm thick, contained both channels and spinule
projections (Figure 42); the contents of the orbicule were no
longer electron-translucent. The channels, which had a diameter
of about 45 nm, and thus were a little narrower than those in
the tectum, appeared to be open to the exterior but not to the
interior. As the midvacuolate period progressed the covering
thickened (Figure 43). The orbicular contents lost their uniformity
and became more electron-translucent at their peripheries. By
the time of pollen mitosis (Figure 44) the orbicular covering
was indistinguishable from the pollen tectum in width (about 300
nm), size of channels (25-45 nm) and spinule-type projections.

The center of the orbicules at that stage contained fibrillar
material comparable to that in the locule itself, and the channels

appeared patent. Although there was frequent direct contact

56

Figures 43-46. 'Tapetal membrane' development. L, locule; ML,
cell of middle layer; 0, orbicule; T, tapetum; TR, tapetal reti-
culum. Bar = 0.2 15m.

Figure 43. Late midvacuolate microspore stage.

Figure 44. Orbicule at time of pollen mitosis. Tapetal reticulum
well developed.

Figure 45. Tapetal cell at time of pollen mitosis. Note layer
between reticulum and plasma membrane containing crystal-like
structures (arrows).

Figure 46. Detail of crystal-like structure.

 

 

57

 

 

58

between the orbicules and the pollen exine at this time, there
were never any signs of actual merging of the two.

In the midvacuolate microspore stage, on the locular face
of the tapetal cells between the orbicules, an incipient electron-
dense reticulum was visible (Figure 42). This layer continued
to develop and eventually became quite conspicuous (Figures 38,
52). At the time of pollen mitosis (Figure 38) this reticulum
with the attached orbicules seemed to form an integral unit which
lined the entire locule. It did not rest directly on the tapetal
cell remnants, but was separated from them by a region approximately
150 nm in width (Figure 38). This region was not empty; it contained
a material of light electron density within which there were some
very electron-dense areas and also some structures which looked
as though they might have held crystals in the process of formation
(Figures 45, 46). Both longitudinal and cross sections of this
region looked the same. This is perhaps the initiation of the
outer of the two layers of the 'tapetal membrane' as described
by Banerjee (1967).

At all stages of development, many of the orbicules were
not quite wholly sheathed by sporopollenin. In the area not
sheathed, contact was maintained between the enclosed droplet
and the tapetal plasma membrane (Figure 42). These openings were

still present to some extent at pollen mitosis.

 

59

Figures 47 and 48. Cross-sections of individual locules. M,
microspore; Po, pore; T, tapetum; TM, tapetal membrane. Bar =
1 um.

Figure 47. Tapetum not yet degenerate.

Figure 48. Tapetum degenerate. 'Tapetal membrane' adhering to
surface of pollen grain.

6O

 

 

61

Scanning Electron Microscope Observations of Whole Anthers

 

Several of the observations made in the light and transmission
electron microscopes were corroborated with the SEM. The close
packing of the microspores within the anther was much more striking
than in comparable TEM sections (Figures 47-50). Orientation
polarity was clearly evident (Figures 49, 50); microspores which
did not have their pores in contact with the tapetum were uncommon.
The closeness of the association between the microspores and the
'tapetal membrane' was especially apparent. Where microspores
had been dislodged, indentations left in the tapetum by the annuli
were clearly visible (Figure 53), and in some places patches of
the reticulum could be seen clinging to the microspores (Figure
48). The orbicules varied in size, some being about twice as
large as others; their distribution appeared to be random (Figure

51).

62

Figure 49. Microspores dislodged from an anther. Note orienta-
tion of pores. Bar = l 'um.

Figure 50. Longitudinal section of anther. Note orientation
of pores (arrows). Bar = 1 pm.

g ~v‘~\\\’\~ \,

 

,2
\ V§\&R&\n

\ \.

so

'.

 

 

64

Figures 51 and 52. Orbicules. O, orbicule; T, tapetum; TR,
tapetal reticulum. Bar = l uflL

Figure 51. Orbicules on the locular face of the tapetum.

Figure 52. Tangential section of the locular face of a tapetal
cell.

65

 

66

Figure 53. Open anther. Note close packing of microspores, and
indentations in tapetum where microspores have been dislodged
(arrows). M, microspore; T, tapetum. Bar = l um.

Figure 54. Single microspore with attached reticulum. O, orbi-
cule; Op, operculum; S, spinules; TR, tapetal reticulum. Bar
=1pm.

67

 

68

Discussion

 

When viewed from the outside, the mature pollen grain of

H, compressum is one of the least elaborate in the plant kingdom.

 

It is spherical, or nearly so, and its most conspicuous feature

is its single pore. The pore is surrounded by a raised rim, or
annulus, and has a small spherical decoration, the operculum,

at its center. The surface of the grain, including the annulus
and the operculum, is smooth except for a dense population of
small spinules, more or less evenly distributed, which rise about
50 nm above the surface and occupy about one-fourth of the surface
area.

At maturity the pollen wall has an inner and an outer component,
the intine and the exine. It is the exine which, because it is
composed mainly of the extraordinarily resistant family of compounds
collectively called sporopollenin, is capable of persisting over
extremely long periods of time with little change.

A cross-section of the gametophyte of H, compressum at the

 

first mitosis shows the wall to be composed solely of exine, the
intine having not yet developed. The several subdivisions of

the exine can be distinguished at this stage and are diagrammed

in Figure 55. The outermost layer is the tectum and from it

extend the spinules that are apparent in surface view. Immediately
beneath the tectum is a space, the cavus, and beneath that the

foot-layer. The cavus is traversed at irregular intervals by

69

tectum

 

spinule

 

channel

 

cavus
columella
foot-layer

 

 

 

 

plasma membrane

 

 

Figure 55. Diagram of the microspore wall at the time of
pollen mitosis.

'rather thick columns, called columellae. Other authors (Skvarla
and Larson, 1966) have concluded that originally the spinules

are the tips of the columellae, and that during the expansion
which the exine undergoes as the microspore enlarges, the spinules
and columellae sometimes become separated. The entire unit,
columella plus spinule, is a baculum. All the aforementioned
areas of the exine, with the exception of the cavus, exhibit the
same staining reaction, and together they comprise the ektexine.

A number of narrow channels traverse the tectum, connecting the
cavus with the external environment. The foot-layer is also
channelled, but much less extensively so. Beneath the foot-layer,

occupying most of the space between it and the plasma membrane,

70

is a moderately osmiophilic granular material.

The pollen walls of other species have the same basic subdivisions,
although the final product may be more or less elaborate than
that of H. compressum. The development of this highly complex
wall has been traced in several species, and in all cases has
been found to begin while the microspores are still joined in
a tetrad formation. For the purposes of discussion, the develop—
mental process can be divided into three stages: (1) the early
development of the layers lying parallel to the plasma membrane,
(2) the development of the radial components, and (3) the growth
of the various components.

The initial sign of wall development, a layer laid down
outside the tetrad microspore plasma membrane underneath the
special callose wall, is called the primexine; the word was coined
by Heslop-Harrison (1963) to indicate the importance of this
material in the development of the mature exine. The term 'primexine'
is somewhat ambiguous, since in most cases it does not refer to
a structure with clearly defined characteristics, but to a structure
or structures whose characteristics vary considerably from species
to species. It is usually finely fibrillar in nature, although
in some species with certain fixatives, especially KMnO4, it may
appear translucent (Heslop—Harrison, l968a,b). In some species
it is of uniform consistency and distribution (Christensen, Horner

and Lersten, 1972) and in others not (Horner and Pearson, 1978).

 

71

It has been shown in some species to be cellulosic in nature
(Heslop-Harrison, 1968a,b), although variation in staining reac-
tions from species to species suggests the possibility of other
compositions. Its upper and lower surfaces are said to act as
a template for the formation of the tectum and foot-layer of the
mature exine (Heslop-Harrison, 1968a), while the bacula (or colum-
ellae) arise as radially oriented, dense, rod-like structures
within it (Heslop-Harrison, l968a; Christensen, Horner and Lersten,
1972; Horner and Pearson, 1978).

A surprising variety of configurations was observed in the

initial stages of wall development in H, compressum. Some of

 

these are particularly interesting because of their similarities

to wall development in other types of cells. For example, membrane-
bound vesicles lying outside the plasma membrane can be found

in both normally growing tissue, such as Aygfla coleoptiles (Heyn,

1971, 1972), and experimental tissue, such as callus. In Andrographis

 

paniculata callus (Bowes and Butcher, 1967) vesicles are embedded

 

in the walls. In carrot callus (Halperin and Jensen, 1967) one
of the first signs of wall breakdown is the g; ggyg appearance
of small membrane-bound vesicles in the region of the middle
lamella. Furthermore, around both the dividing and non-dividing
cells of proembryos derived from carrot callus, the cell wall

is separated from the plasma membrane by an electron-translucent

space, as was the case for some of the H, compressum microspores.

 

72

Within this space, in addition to vesicles, a thin, faint, horizontal
dense layer is visible, similar to the one sometimes seen around
the microspores (Figures 12, 14). This may be the structure which
Skvarla and Larson (1966) labelled the tectum template in ggg
Hgys microspores. Its presence in carrot callus, whose cells
do not develop a tectum, would presumably nullify that interpretation.

The microspores showing early wall development were also
the ones in which paramural bodies were seen. Paramural bodies
have been found in Z, mgyg tetrads (Skvarla and Larson, 1966)
at about the same stage of development, and indeed are a common
feature in many types of plant cells (Roland, 1973). Although
it is far from certain, the paucity of dictyosomes in H. compressum
microspore tetrads would tend to indicate that these particular
paramural bodies were plasmalemmasomes (derived from the plasma
membrane) rather than lomasomes (derived from cytoplasmic membranes).
According to Marchant and Robards (1968) plasmalemmasomes are
more likely than lomasomes to be involved in enzymic processes,
and may function in secondary modifications of the cell wall.
On the other hand, whether these vesicles were moving out of the
cell, or perhaps into it as has been suggested by Thomson (1967),
and what their relationship is to the vesicles lying outside the
plasma membrane, is really a matter of conjecture.

Some of the H, compressum microspores exhibited a type or

stage of early wall development that seems to correspond to the

73

primexine described in Silene pendula and some other species

 

(Heslop-Harrison, l968a), where a thin electron-translucent area
lyfing parallel to the surface of the plasma membrane is interrupted
at intervals by radially oriented deposits of a fibrillar nature

(Figure 15). But in H, compressum wall development in apparently

 

more advanced microspores still encased in callose consisted of
a discontinuous layer of amorphous osmiophilic material (Figures
16-18). This material was not 99 the surfaces of the electron-
translucent areas, but within them, and radially directed rods,
rather than being prominent, were not visible. It would seem,“

therefore, that in H, compressum the electron-translucent areas

 

may represent sites where a substance secreted by the microspore

has caused a localized degradation of callose, and that within

these electron-translucent areas chemical reactions occurring

between the breakdown products and materials secreted by the cell

result in the deposition of the amorphous osmiophilic substance.

Also, materials from the locule may diffuse through the callose

and enter into this reaction; there are conflicting reports about

the impermeability of callose (Mepham, 1970; Southworth, 1971).

The fibrillar areas in Figure 15 could be residual areas of callose.
Alternatively, preformed osmiophilic material may be discharged

from vesicles into the translucent areas. Arguing against this

possibility is the fact that the profile of the plasma membrane

was not as convoluted as would be expected in a cell actively

74

engaged in such secretion.

The dense osmiophilic material varied in width from cell
to cell, but even when it attained its maximum width, lateral
discontinuities persisted in it. This is contrary to the situation

reported for Sorghum bicolor (Christensen, Horner and Lersten,

 

1972) and some other species where the primexine becomes a uniform

layer. If this osmiophilic layer in H, compressum is a primexine,

 

as seems likely, then we can assume that sporopollenin will be
laid down on its upper and lower surfaces, that bacula will develop,
most probably in the discontinuities, and that the primexine will
ultimately disappear leaving the profile of the developing exine
seen in Figure 19.

There is the possibility of another ontogenetic pathway.
The osmiophilic layer in Figure 18 is separated from the plasma
membrane by a less darkly-staining region, and in at least one
area there seem to be two osmiophilic layers separated by a spacer
layer. These configurations could be interpreted to mean that
the initial layer of dense material is the rudimentary tectum,
and that following the secretion of its components two other
layers are secreted: a 'spacer' layer that initially occupies
the region that will become the cavus, followed by another osmio-
philic layer which will become the foot-layer. Should this inter-
pretation be correct, it would raise the possibility that_H. compressum

is totally lacking a primexine, and would tend to corroborate

 

the observations of Rowley and Dahl (1977) who found no primexine

in Artemisia vulgaris. In that event, the role that has been

 

postulated for the primexine, namely that it is a necessary inter-
mediate and probably a template in the development of the mature
exine(Heslop-Harrison, 1968 a,b), would have to be reassessed.

In various plants the eccentric location of certain organelles,
usually but not always ER, has been associated with the laying
down of primexine and with the imprinting upon it of a template
which will control the pattern of the mature exine (Heslop-Harrison,
1963; Skvarla and Larson, 1966; Vazart, 1970). No such organelle
association was seen in this study. The possibility of microtubule
involvement has not been ruled out, however, since fixation was
carried out in the cold, which would have caused the depolymerization
of any microtubules that might have been present.

Radially oriented structures usually begin to appear while
the microspores are still encased in the special callose wall,
although not necessarily at the same time as that the concentric
components of the primexine appear. In many species, including
z, Hgys (Skvarla and Larson, 1966) and S. bicolor (Christensen,
Horner and Lersten, 1972), bacula, or columellae as the case may
be, evidently arise as thickenings within the primexine. In

species that possess spines (e.g. Helianthus annuus) unambiguous

 

spine (but not spinule) templates may be present as outward extensions

of the primexine (Horner and Pearson, 1978). The ontogeny of

76

the bacula in H, compressum is unresolved. As mentioned earlier,

 

it is generally held that the bacula are merely columellae whose
tips or spinules project above the tectum. Yet the fact is that
the profiles of the spear-shaped spinules as they appeared in

the early vacuolate microspore stage of H. compressum often gave

 

the impression that they were separate entities which had come

to rest in the discontinuities of the tectum: their density was
slightly different than that of the tectum and foot-layer, they
had a smooth outline which sometimes did not connect to the tectum,
and, unlike the tectum and foot-layer, they displayed an osmiophlic
boundary layer. Although they frequently did seem to extend to

the foot-layer, or to rest atop columellae, such was not always

the case. Sometimes a spinule was displaced slightly to one side
of a columella; at other times no columella was visible beneath

a spinule. The very strong resemblance of the spinules in size,
shape and staining characteristics to similar structures on the
developing tapetal orbicules is a curious fact, and one that makes
a primexine origin doubtful. A more likely explanation would

seem to be that both the rudimentary exine and the surface of

the orbicules have specific areas that for one reason or another
act as nucleating centers for the crystallization of precursors
that are present in the locular fluid. This would be consistent
with the points mentioned, and also with the facts that (l) the

spear-like structures appeared comparably well developed on all

77

parts of the exine as well as on the tapetum, and (2) no spear-
like structures were seen in the exine of those tetrads still
encased in callose.

Following the dissolution of the special callose wall, all
components of the ektexine underwent considerable growth, and
the tectum and foot-layer became channelled. Both phenomena,
but particularly the latter, are poorly understood. It has been
shown that sporopollenin precursors are formed in the tapetum
of at least some plants (Brooks and Shaw, 1978; Rodriguez-Garcia,
1978), and it is presumed that they are secreted by the tapetum
into the locule after which they polymerize on any suitable surface.

Nothing observed in H, compressum could be considered inconsistent

 

with this mode of action. Initially in this species only the
rudimentary exine and tapetal orbicules present suitable surfaces
for sporopollenin polymerization. At a later stage the plasma
membranes on the inner tangential face of the tapetal cells also
become covered with sporopollenin.

There is evidence that microspores themselves are also capable
of secreting sporopollenin precursors and probably contribute
at least part of the exine (Echlin, 1971; Gabara, 1974). This
is especially likely to be true in plants where the exine is
highly developed before the microspores are released from the

callose (e.g. Helianthus annuus (Horner and Pearson, 1978), Trades-

 

cantia bracteata (Mepham and Lane, 1970)). It seems likely that

 

78

in H. compressum the microspore contributes to the sporopollenin
of the foot-layer at least, since it appears to be built up from
the inside, and to the endexine when it ultimately develops.

Much attention has been focused on the structure and function
of the 'tapes' or 'lamellae' which have been observed frequently
in the developing foot-layer, particularly in the pore area, and
also, but much less frequently, in the tectum and orbicules.
Some investigators claim that sporopollenin is deposited only
on lamellae (Heslop-Harrison, 1975). Lamellae were clearly visible
in H. compressum, especially during the early vacuolate period,
in the foot layer of the non-apertured wall and of the pore, and
as reported for other species, they seemed to become embedded
and subsequently obscured in the foot-layer of the non-apertured
wall, but remained visible in the pore area. It is not clear
whether the dark line subtending the foot-layer in the midvacuolate
stage is such a lamella or is a separate phenomenon. It is thought
that the lamellae are produced by the plasma membrane and travel
from the plasma membrane to the foot-layer (Dickinson, 1976).
Whether this is also true in H. compressum has not been determined
but it may be true since some lamellae were seen very close to
the plasma membrane.

The evidence from a great number of investigators suggests
that for sporopollenin polymerization to occur one of two things

must be present, either a special surface or pre-existing sporopol-

79

lenin. Sporopollenin is not deposited on every available surface.
It is frequently not deposited on the tapetal cells themselves,

and it is never deposited on the microspore plasma membrane.

On the other hand, once it has been deposited in the tectum or

on the tapetal orbicules, it usually continues to thicken without
the obvious intervention of a new polymerizing surface. (For

an exception see Echlin and Godwin, 1968). These facts have

caused Mepham (1970) to advance the theory that the polymerization
process may require two enzymes, an initiator and a chain elongator.
Apparently the tapetal orbicules, at least some primexines and

the lamellae, contain an initiator or an initiating site. It
remains enigmatic that the foot-layer should require successive
layers of lamellae, and it would seem that in those species where
repeated layers of lamellae are produced some other purpose must
also be served by them. It is worth noting that the production

of layers of lamellae is not unique to pollen grain walls: a similar
configuration has been observed during the development of new

walls around tomato protoplasts (Willison and Cooking, 1972).

The debate in the literature at present centers on whether
lamellae are essential for polymerization of sporopollenin. The
lamellae have been reported often enough that the reality of their
existence is no longer questioned; it is assumed that they are
of unknown composition. The possibility that the lamellae themselves

could be the fixed ultrastructural representation of sporopollenin

 

80

precursors in the process of polymerization does not seem to have
been given much consideration. In view of the difficulty of
locating a source for the lamellae, particularly for those which
occur in the orbicules, it would seem reasonable to consider such
a possibility.

The channels that penetrated the tectum and the foot-layer
were on the whole fairly straight (although a few were branched),
non-tapering, and of similar size. Many of them persisted as
the tectum and foot-layer thickened. (In the late midvacuolate
and late vacuolate stages some appeared to be occluded, but may
have simply passed out of the plane of the section). It is not
known how such channels arise. Rowley (1976) has reviewed three
processes which may result in tectal perforation, which presumably
could be applied to the foot-layer as well. The first of these
is incomplete coalescence of the tectum as it spreads over the
primexine. Perforations thus formed can be distinguished by their
size variations as well as by their shapes: funnel-form in longitudi-
nal section and angular in cross-section. The H. compressum channels
do not fit this description. A second possibility is that a once
complete tectal sheet may become reticulate by stretching, as
has been reported for Sparganium (Banerjee, Rowley and Allessio,
1965). Stretching cannot be invoked in the case of H, compressum
since it would not explain the channels, identical to those of

the exine, that were found on the tapetal orbicules, which did

81

not enlarge. Nor did the channels continue to increase in size

or number as would be expected from continuing microspore enlargement.
The third mechanism, and the one strongly favored by Rowley,

is the exclusion of sporopollenin from certain areas by the presence

of rod-shaped entities which he considers to be part of a plasma

membrane glycocalyx. Work done by Roland and Vian (1971) using

periodic acid Schiff-thiocarbohydrazide-silver protein (PATSP)

and phosphotungstic acid (PTA) at low pH indicates the presence

of a coat on the plasma membrane of plant cells similar to that

described in animal cells (Rambourg, 1971; Winzler, 1970), as

well as the presence of strands connecting the plasma membrane

to the cell wall. Rowley and coworkers (Rowley, 1973a, 1976;

Rowley and Dahl, 1977; Rowley and Skvarla, 1975), applying similar

techniques to microspores of several species, have found, among

other things, that microchannels in Epilobium and Artemisia stain

with lanthanum nitrate and other cations, indicating the presence

of acid polysaccharides. It is not clear whether this hypothesis

can be applied to grasses. The idea that a rodshaped element

may be responsible for sporopollenin exclusion is appealing because

it would explain why the channels persist rather than become

occluded as one might expect if they were awash in sporopollenin

precursors. In H. compressum, very narrow, somewhat opaque channels

 

were seen in the early vacuolate stage (Figure 19). Whether these

expanded to become the channels visible in later stages is very

82

uncertain.

The function of the channels in grasses is unknown. Studies
done on a variety of species have shown that molecules with molecular
weights as high as 500,000 can cross both aperturate and non-aperturate
walls (Rowley, 1973b, 1975). Thus there is at present no reason

to assume that the channels in H. compressum are not capable of

 

transporting materials. Since all areas of the cell inside the

wall are metabolically active, it would seem advantageous for

all areas to have easy access to metabolites. On the other hand,

it would be premature to dismiss the possibility that the channels
primarily serve some other function, or that they might be in

whole or in part artefactual. Inducible channels have been reported

in the nexine of Epilobium angustifolium (Rowley, 1973a, 1976).

 

If we assume that the channels are real structures engaged

in transport, then, in H. compressum at least, more transport

 

apparently occurs between the locule and the cavus than between

the cavus and the microspore. Among the things we might reasonably
expect to find transported into the cavus are sporopollenin precursors,
because the columellae increase considerably in both length and

width. If the columellae do not grow by accretion of new material,
then they must do so as a result of material flow from the tectum

or the foot layer or both. No attention has been given to either

the growth of the columellae or the increase in width of the

cavus. Does the cavus enlarge because the columellae grow, or

83

is the situation the reverse? Is it possible, for example, that
the cavus contains a hygroscopic carbohydrate which swells as

a consequence of absorbing water from the locule, thus forcing
the tectum and foot-layer farther and farther apart? Although
the cavus of acetolyzed pollen is empty (hence its name), and

although at the time of pollen mitosis the contents of the cavus

of H. compressum appeared similar to the contents of the locule,

 

in earlier stages its characteristics were different from those
of the locule and we have to assume that the contents of the two
areas were not the same and were not freely interchangeable.
We need not expect to find the contents of the cavus to be the
same for all species, identical in each and every species, since
it apparently serves a function - storage of material derived
from the tapetum - in some species (Heslop-Harrison, Heslop-Harrison,
Knox and Howlett, 1973; Howlett, Knox and Heslop-Harrison, 1973)
which it does not serve in grasses, where its function is presently
unknown.

Although exine development has been investigated at the TEM
level in a number of species, fairly complete information is

available for only two grass species, Sorghum bicolor (Christensen,

 

Horner and Lersten, 1972; Christensen and Horner, 1974) and Egg
mays (Skvarla and Larson, 1966). Some information on later stages

is available for 393 annua (Rowley, 1962, 1964) and Triticum aestivum

 

(DeVries and Ie, 1970), and there are a few isolated pictures

 

84

available of Avena sativa (Gunning and Steer, 1975; Steer, 1977),

 

Agropyron intermedians (Young et al., 1979), and Phleum pratense

 

 

(Rowley and Skvarla, 1974). Since Sorghum, Zea and Hordeum are

 

strikingly similar although not identical at the time of pollen
mitosis - there are some subtle differences in channelling and
surface texturing (Andersen and Bertelsen, 1972; Grant, 1972;
Rajendra et al., 1978) - it is disconcerting to find that some
stages of early exine development in H. compressum resemble those
of S. bicolor, while others are more similar to those of ;, HHQL,
especially so since the latter was fixed in KMnO4. It is highly
likely, however, that identical preparative techniques would go
a long way toward bringing the three grasses into closer agree-
ment while making it possible to determine exactly what the differences
are. The problem of interpretation, not only for the grasses
but also for all the species that have been studied, is compounded
by an almost total lack of cytochemical data. What seems to be
needed is a series of extensive cytochemical studies such as those
done recently by Rowley and Dahl (1977) on Artemisia. Such studies
which, in addition to locating sporopollenin, make possible the
identification of proteins, lipids and various classes of carbo-
hydrates, ought to resolve many of the ambiguities and contradictions
that presently exist.

The ultrastructure of the microspore cytoplasm and nucleus

was unremarkable and, as far as can be determined from the scant

85

literature available, similar to that of other grasses. The usual
organelles were present, with the exception of dictyosomes, which
were scarce until the time of pollen mitosis. The lack of dictyosomes
is surprising in a cell whose major activity seems to be the
making of a complex wall, particularly if a cellulosic primexine
is in fact being produced. In~Artemisia (Rowley and Dahl, 1977)
the initial exine structures contain large amounts of protein,
a fact that might be consistent with a dearth of dictyosomes.
It may be that dictyOSome appearance at the time of pollen mitosis
is a preparation for the secretion of the intine, which in all
species is reported to be cellulosic and which in H. compressum
is laid down after the first pollen mitosis. The nature of the
osmiophilic bodies seen in several microspores and the occasional
appearance of pairs of parallel ER cisternae are without explanation
at present. Growth in cell size was concomitant with increasing
vacuolization and was accompanied by the migration of the nucleus
to the side of the cell away from the germinal pore.

It was abundantly clear that at the time of pollen mitosis
the majority of the microspores were so oriented that the pore
was in contact with the tapetum; most of the microspore tetrads
also seemed to have one surface near the tapetum. Although Banerjee
and Barghoorn (1971) have advanced the view that the microspores
become oriented after their release from the callose, it seems

reasonable to suppose that the pore ordinarily develops adjacent

86

to the tapetum and remains there throughout the period of pollen
development. This has been proposed by Christensen and Horner
(1974) as one of several polar events occurring within the developing
microspore. The random orientation of the microspores in the
early and midvacuolate stages could easily have resulted from
disturbances during fixation. But whether or not the position
of the pore adjacent to the tapetum has any functional significance
is a moot point. The factor determining pore location has been
looked for in several species, and in some grasses has been reported
to be, or to be related to, the orientation of the meiotic spindle
(Rajendra et al., 1978). Thus the pore-tapetum relationship could
be the consequence of a polarization several generations of cells
earlier and have no particular significance to the plant at this
stage.

The tapetum itself is one of the most puzzling of all plant
tissues. Its various unusual physical features combined with
its location make clear that it plays a highly specialized and
significant role which must somehow be concerned with the develop-
ment of the microspore, although very little progress has been
made toward determining what that role is. The tapetum is the
site of synthesis of sporopollenin precursors in some species
(Brooks and Shaw, 1978; Rodriguez-Garcia, 1978) and in some species
is the site of synthesis of recognition proteins which are later

transferred to the pollen grain and function in pollen-stigma

87

interactions (Heslop-Harrison, Heslop-Harrison, Knox and Howlett,

1973; Dickinson and Lewis, 1973; Heslop-Harrison, Knox and Heslop-

Harrison, 1974). Maheshwari (1950) suggested that the tapetum

is important in the nutrition of the developing microspore, and

it is frequently labelled a "nurse" tissue; that designation seems

inappropriate since the tapetum has neither chloroplasts nor

starch reserves. The tapeta of different species follow a sequence

of events which is in a general sense fairly predictable, but

correlation of these events with the development of the microspore

is complicated by variation in their timing from species to species.
Among the uncommon features which characterize the tapetum

of H. compressum and many other species are: (1) a binucleate

 

condition; (2) the lack of an extensive vacuole system; (3) a

high proportion of ER, predominantly rough; (4) dissolution of

the primary walls followed by a period when the inner tangential
and radial sides of the cells have no walls whatsoever; (5) subse-
quent development on the inner tangential face of a highly unusual
wall; and (6) broad cytoplasmic connections between many of the
tapetal cells.

The first three of these six, although noteworthy, are not
extraordinary. The binucleate or multinucleate condition is a
type of polyploidy found in many different kinds of plant and
animal cells, and there are no satisfactory explanations to account

for it. In addition, some of the individual nuclei in the tapetum

88

have more than two nucleoli, and during the karyokinesis which
results in the binucleate condition, the chromosomes appear unusually
large (Whallon, unpublished). Both observations suggest that
each individual nucleus may also be polyploid.

The absence of a large vacuole may be correlated with the
lack of wall structure and with the proximity of the locular
fluid. Since the vacuole and the wall together function in osmoregu-
lation, any plant cell without a wall and with a large vacuole
could be expected to burst. Thus, the absence of a vacuolar
system may be necessary if the cell is to survive without a wall.
A large vacuole also usually serves as a repository of waste
products. At or shortly after the time that a vacuolar system
might ordinarily be developing, the plasma membrane comes into
direct contact with the locular fluid as a result of wall dissolution.
Presumably such waste products as the cell needs to get rid of
can then be excreted directly into the locule. Once the tapetal
reticulum has formed, contact between the plasma membrane and
the locule must be severely restricted, and it is possible that
the consequent build-up of poisons contributes to the cell's
death. It is not likely to be the primary cause of death since
all tapetal cells, even those in species where a reticulum does
not develop, senesce prematurely.

Many of the speculations about tapetal function center around

the large amount of ER, particularly rough ER, that develops in

89

these cells. It is generally accepted that the tapetum is a

secretory tissue, and although sporopollenin is suspected of being

the primary product, definitive proof of this is lacking. Unusually
close association between ER and plastids, such as that shown

by the ensheathing cisternae in Figure 34, has been reported in

A, égtiyg (Steer, 1977b), where the ER also ensheaths the mitochondria.

In Lycopersicon peruvianum the association is even more intimate

 

(Pacini and Cresti, 1976).

It is possible that the 'cytoplasmic plate complexes‘ represent
a localized area of modified ER. Previous published studies have
not shown any similar structures, with the possible exception
of the 'zytoplasmavesikel' described by Marquardt gt 31. (1968).
The complexes bear some resemblance to dictyosomes, but their
association with the larger vesicles, and their enclosed cyto-
plasmic plates, indicate that they are something other than specialized
dictyosomes. Morphologically they are somewhat similar to the
so-called 'dense compartments' found in megakaryocytes of stressed
rats (Bentfeld-Barker and Bainton, 1977); the function of dense
compartments is also unknown.

The tapetum of H. compressum bears a strong overall resemblance

 

to those of H. sativa and I. aestivum. But apparently in H. sativa,
and perhaps in I. aestivum, the tapetum reaches its fullest development
after the pollen wall has completed most of its growth (Steer,

1977b; OeVries and Ie, 1970). This is not true in H, compressum,

 

 

90

nor in many other species (Echlin, 1971). Intine development
also precedes tapetal degeneration in I. aestivum, but follows

it in H, compressum.

 

One of the most unusual features of the maturing anther is
the layer of orbicules which completely covers the inner tangential
wall of the tapetum. These structures, whose function remains
an enigma, are present in many, but not all, species. The ontogeny
of the orbicules is approximately the same in all species, although
there is some dispute about whether the droplets or cores (pro-
orbicules) upon which the sporopollenin aggregates are performed
in the tapetum (Echlin and Godwin, 1968; Steer, 1977a). In H.

compressum, the first-formed droplets could not be seen within

 

the tapetal cytoplasm (Figure 39), although the later-formed,
electron-translucent droplets could (Figures 40, 11); it is not
possible to conclude that the earlier droplets were in fact pro-
orbicules. There have been attempts to dismiss the orbicules

as an accidental product of the tapetal cell's ability to secrete
sporopollenin precursors from all its surfaces in combination
with the peculiar environment presented by the locule (Dickinson
and Bell, 1972). However, it is apparent in grasses, more so
than in some other species, that the patterning of the orbicule
is identical to that of the tectum on the developing pollen grains.
Thus, there is clearly something special about the orbicule as

a polymerizing surface, since sporopollenin which accumulates

91

on other tapetal surfaces either has a different pattern (e.g.,

the tapetal reticulum in H, compressum) or lacks a pattern altogether

 

(e.g. that produced at the outer tangetial surface in H, 533112).
An examination of Figure 51 shows that the orbicules vary in size,
some being at least twice as large as others.

The 'tapetal membrane' which developed beneath the orbicules
as the tapetal cells senesced consisted, at the time of pollen
mitosis, of two layers, an inner (i.e., toward the locule), fenestrated
reticulum, and an outer, non-fenestrated, backing layer. The
'membrane' seems to serve the function of enclosing the pollen
grains in a sac, for reasons unknown. A similar structure, called
the 'peritapetal membrane', develops in non-grass species at the
outer tangential side of the tapetum (Heslop-Harrison, 1969).

In most species remnants of the tapetal cells are ultimately
deposited on the pollen grains as either 'pollenkitt' - a sticky
homogenous substance composed largely of carotenes, characteristic
of entomophilous species, - or as tryphine - a nonhomogenous
substance composed of remnants of cytoplasmic elements. The
presence of the 'tapetal membrane' on the inner side of the tapetum
would seem to prevent any tapetal remnants from reaching the
microspores unless the backing layer becomes fenestrated at some
later time.

In grasses, there is a very close association between the

orbicules and the reticulum of the tapetal membrane, but this

92

does not necessarily imply a functional relationship. Some species

(e.g. Pinus banksiana, Beta vulgaris) with a peritapetal membrane

 

 

also have orbicules on the locular face of the tapetum.

In the plasmodial or invasive type of tapetum, the cytoplasm
of adjacent cells becomes confluent as a result of breakdown of
the plasma membranes (Mepham and Lane, 1969). Such an occurrence
has not usually been mentioned as characteristic of parietal

tapeta, although it has been reported to occur in Pinus banksiana

 

by Dickinson and Bell (1972), and has been mentioned in passing

by Young et a1. (1979) with regard to Agropyron intermedians.

 

Broad cytoplasmic bands connecting adjacent cells were clearly

present in H. compressum from the early vacuolate microspore stage

 

onwards.

Good progress has been made toward an understanding of events
in the microspore, and there is every reason to believe that
progress will continue to be good. An approach that I believe
might be fruitful, especially with regard to the developing wall,
would be to look for similarities between the microspore and other
cells of the plant. But no concepts of microsporogenesis can
be satisfactory until we also understand the tapetum, and at present
we know very little about the tapetum. Considerable difficulty
arises because there are some very great differences among the
tapeta of those species which have been studied to date, and also

because the tapeta of some species contain highly anomalous structures,

 

93

such as the nucleolar tubules of Beta vulgaris (Hoefert, 1971).
Ultrastructural studies need to be done on many more species.

A technique for isolating tapetal cells so that biochemical assays
on a time course could be performed, especially if combined with

ultrastructural studies, ought to prove extremely valuable.

 

94

Me

From combined light and electron microscopic observations

on developing anthers of H, compressum, it appears to be true

 

that:

1. The microscope resembles that of other grasses in number and
kind of organelles, pattern of vacuolization, migration of the
nucleus, and orientation relative to the tapetum.

2. In the development of the microspore, a primexine may be
lacking; bacula probably arise after callose has dissolved from
precursors present in the locule. Lamellae are present in both
the apertured and non-apertured parts of the wall during the
period of sporopollenin deposition. The exine channels are not
the result of incomplete tectal and foot-layer spreading, nor

of stretching consequent to an increase in size of the microspore.
3. In the tapetal cells, which are parietal (secretory) and
binucleate, extensive ER develops, and there is no large vacuolar
system. Broad cytoplasmic bands connect the tapetal cells to

one another. A cytoplasmic structure found in some tapetal cells
during the midvacuolate stage may represent a previously undescribed
organelle, or an undescribed configuration of a known organelle.
4. Tapetal degeneration occurs earlier than in H, ggtiyg or I.
aestivum. The development of the 'tapetal membrane' prior to
rupturing of the tapetal plasma membrane may prevent any remnants

of the tapetal cells from reaching the pollen as tryphine.

95

5. Tapetal pro-orbicules are preformed in the tapetal cytoplasm;

the subsequent development of the orbicules is practically identi-
cal to that of the microspore tectum. This fact combined with

the pattern of sporopollenin in the tapetal reticulum, supports

the hypothesis that the orbicules represent an evolutionary speciali-
zation rather than a vestigial capacity of a tissue that was once
sporogenous, or a random, non-specific configuration of excess

sporopollenin.

 

BIBLIOGRAPHY

96

Andersen, S. T. and F. Bertelsen. 1972. Scanning electron microscope
studies of pollen cereals and other grasses. Grana 12:79-86.

Banerjee, U. C. 1967. Ultrastructure of the tapetal membranes in
' grasses. Grana Palynol 7:365-377.

 

Banerjee, U. C. and E. S. Barghoorn. 1971. The tapetal membranes in
grasses and Ubisch body control of mature exines. In Pollen,
development and physiology, J. Heslop-Harrison, ed. London: Butter-
worths, pp. 126-127.

 

 

Banerjee, U. C., J. R. Rowley and M. L. Allessio. 1965. Exine plasticity
during pollen grain maturation. _g. Palynol. 1:70-89.

Bentfeld-Barker, M. E. and D. F. Bainton. 1977. Ultrastructure of
rat megakaryocytes after prolonged thrombocytopenia. .g. Ultra-
struct. Res. 61:201-214.

 

Bowes, B. G. and D. N. Butcher. 1967. Electron microscopic observations
on cell wall inclusions in Andrographis paniculata callus. .2.
Pflanzegphysiol. 58:86-89.

 

 

Brooks, J. and G. Shaw. 1978. Sporopollenin: a review of its chemistry,
paleochemistry and geochemistry.' Grana 17:91-97.

Christensen, J. E. and H. T. Horner, Jr. 1974. Pollen pore development
and its spatial orientation during microsporogenesis in the grass
Sorghum bicolor. Amer. H, H95. 61:604-623

 

Christensen, J. E., H. T.Horner, Jr. and N. R. Lersten. 1972. Pollen
wall and tapetal orbicular wall development in Sorghum bicolor
(Gramineae). Amer. H, H93. 59:43-58.

 

OeVries, A. Ph. and T. S. Ie. 1970. Electron microscopy on anther
tissue and pollen of male-sterile and fertile wheat (Triticum
aestivum L.). Euphytica 19:103-120.

Dickinson, H. G. 1970. Ultrastructural aspects of primexine formation
in the microspore tetrad of Lilium longiflorum. Cytobiologie
1:437-449.

 

 

Dickinson, H. G. 1971. The role played by sporopollenin in the dev-
elopment of pollen in Pinus banksiana. In Sporopollenin, J. Brooks
.e§_gl. eds. London: Academic Press, pp. 31-68.

 

 

Dickinson, H. G. 1976. Common factors in exine deposition. In The
evolutionary significance gf_the exine (Linnean Soc. Symp. SE? ‘
Hg, 1), I. Ferguson and J. Muller, eds. London: Academic Press,

 

 

97

pp. 67-89.

Dickinson, H. G. and P. R. Bell. 1972. The role of the tapetum in
the formation of sporopollenin-containing structures during micro-
sporo-genesis in.Pinus banksiana. Planta (Ber1.) 107:205-215.

 

Dickinson, H. G. and P. R. Bell. 1976. Changes in the tapetum of Pinus
banksiana accompanying formation and maturation of the pollen.
.HHH.IEQE. (London) 40:1101-1109.

 

Dickinson, H. G. and D. Lewis. 1973. The formation of the tryphine
coating the pollen grains of Raphanus and its properties relating
to the self-incompatibility system. Proc. ng. Soc. London, Ser.
H, Biol. S21. 184:149-165.

 

Echlin, P. 1971. Production of sporopollenin by the tapetum. In
Sporopollenin, J. Brooks §§_§l,, eds. London: Academic Press,
pp. 220-248.

 

Echlin, P. and H. Godwin. 1968. The ultrastructure and ontogeny of
pollen in Helleborus foetidus L. I. The development of the tape-
tum and Ubisch bodies. H, Cell §gi, 3:161-174.

 

Echlin, P. and H. Godwin. 1969. The ultrastructure and ontogeny of
pollen in Helleborus foetidus L. III. The formation of the pollen
grain wall. 9, Cell Sgi, 5:459-477.

 

Gabara, B. 1974. A possible role for the endoplasmic reticulum in
exine formation. Grana 14:16-22.

Godwin, H., P. Echlin and B. Chapman. 1967. The development of the
pollen grain wall in Ipomoea purpurea (L.)Roth. {Hgy. Palaeobot.
Palynol. 3:181-195.

 

Grant, C. A. 1972. A scanning electron microscopy survey of some
Maydeae pollen. Grana 12:177-184.

Gunning, B. E. S. and M. Steer. 1975. Ultrastructure and the biology
9f_plant cells. LondonzArnold.

 

Halperin, W. and W. A. Jensen. 1967. Ultrastructural changes during
growth and embryogenesis in carrot cell cultures. .g. Ultrastruct
Hgé, 18:428-443.

 

Heslop-Harrison, J. 1963. An ultrastructural study of pollen wall
ontogeny in Silene pendula. Grana palynol. 427-24.

 

 

98

Heslop-Harrison, J. l968a. Pollen wall development. Science 161:230-
237. ““““

Heslop-Harrison, J. 1968b. Wall development within the microspore
tetrad of Lilium longiflorum. Canad. H. pr. 46:1185-1192.

 

Heslop-Harrison, J. 1969. An acetolysis-resistant membrane investing
tapetum and sporogenous tissue in the anthers of certain Compo-
sitae. Canad. H. pr, 47:541-542.

Heslop-Harrison, J. 1971. The pollen wall: structure and development.
In Pollen, development and physiology, J. Heslop-Harrison, ed.
London: Butterworths, pp. 75-98.

Heslop-Harrison, J. 1975. The physiology of the pollen grain surface.
Proc. pr. Soc. London, Ser. H, Biol. H91. 190:275-299.

 

 

Heslop-Harrison, J. 1976. The adaptive significance of the exine.
In The evolutionary significance_pf the exine. (Linnean Soc. Symp.
Series Hp, 13, I Ferguson and J. Muller, eds. London: Academic
Press, pp. 27-37.

 

Heslop-Harrison, J., Y. Heslop-Harrison, R. B. Knox and B. Howlett.
1973. Pollen-wall proteins: 'gametophytic' and 'sporophytic'
fractions in the pollen walls of the Malvaceae. Ann. Bot. (London)
37:403-412.

 

Heslop-Harrison, J., R. B. Knox and Y. Heslop-Harrison. 1974. Pollen-
wall proteins:exine-held fractions associated with the incompati-
lity response in Cruciferae. Theor. Appl. Genetics. 44:133-137.

 

 

Heyn, A. N. J. 1971. Observations on the exocytosis of secretory
vesicles and their products in coleoptiles of Avena. H, Ultra-
struct. Res. 37:69-81.

 

Heyn, A. N. J. 1972. Intra- and extracytoplasmic microtubules in
coleoptiles of Avena. ‘H. Ultrastruct. Res. 40:433-457.

 

Hoefert, L. L. 1969. Ultrastructure of Beta pollen. I. Cytoplasmic
constituents. Amer. H, pr, 56:363-368.

Hoefert, L. L. 1971. Ultrastructure of tapetal cell ontogeny in Beta.
Protoplasma 73:397-406.

 

Horner, H. T., Jr. and N. R. Lersten. 1971. Microsporogenesis in
Citrus limon (Rutaceae). Amer. H, pr. 58:72-79.

 

Horner, H. T., Jr. and C. 8. Pearson. 1978. Pollen wall and aperture

99

development in Helianthus annuus (Compositae: Heliantheae) Amer.
_H..pr. 65:293-309.

 

Horner, H. T., Jr. and M. A. Rogers. 1974. A comparative light and
electron microscopic study of microsporogenesis in male-fertile
and cytoplasmic male-sterile pepper (Capsicum annuum). Canad.
'H..Hp£, 52:435-441.

 

Howlett, B. J., R. B. Knox and J. Heslop-Harrison. 1973. Pollen-wall
proteins: release of the allergen anitgen E from intine and exine
sites in pollen grains of ragweed and Cosmos. .H. Cell H21. 13:603-
619.

Johnson, W. S., G. R. Hooper, B. F. Holdaway and H. P. Rasmussen. 1976.
2,2,Dimethoxypropane, a rapid dehydrating agent for scanning elect-
ron microscopy. Micron 7:305-306.

Kosmath, L. 1927. Studien uber das antherentapetum. Oesterr. Bot.
.g. 76:235-241.

 

Larson, 0. A. and C. W. Lewis. 1962. Pollen wall development in Park-
insonia aculeata. Grana Palynol. 3(3):21-27.

 

 

Lee, S. L. J., V. E. Gracen and E. D. Earle. 1979. The cytology of
pollen abortion in C-cytoplasmic male-sterile corn anthers. Amer.
H, pr, 66:656-667.

Lombardo, G. and L. Carraro. 1976. Tapetal ultrastructural changes
during pollen development. I. Studies on Antirrhinum maius.
Caryologia 29:113-125.

 

 

Maheshwari, P. 1950. ‘Hp Introduction Hp the Embryolggy pf_Angiosperms.
New YorkchGraw-Hill.

 

 

 

Marchant, R. and A. W. Robards. 1968. Membrane systems associated
with the plasmalemma of plant cells. .flflfl°.§9£- (London) 32:457-
471.

Marquardt, H., 0. M. Barth and U. von Rahden. 1968. Zytophotometrische
und elektronen mikroskopische beobachtungen uber die tapetumzellen
in den antheren von Paeonia tenuifolia. Protoplasma 65:407-421.

 

 

Matile, Ph. 1975. The lytic compartment pf plant cells. Wien and
New York: Springer-Verlag.

 

 

Mepham, R. H. 1970. Development of the pollen grain wall: further
work with Tradescantia bracteata. Protoplasma 71:39-54.

 

 

100

(Mepham, R. H. and Lane, G. R. 1969. Formation and development of the
tapetal periplasmodium in Tradescantia bracteata. Protoplasma
68:175-192.

 

Mepham, R. H. and G. R. Lane. 1970. Observations on the fine structure
of developing microspores of Tradescantia bracteata. Protoplasma
70:1-20.

Pacini, E. and M. Cresti. 1976. Close association between plastid
and endoplasmic reticulum cisterns during pollen grain development
in Lycopersicon peruvianum wild tomato species. .H. Ultrastruct.
.Hgg. 57:260-265. ‘

 

 

Rajendra, B. R., A. S. Tomb, K. A. Mujeeb and L. S. Bates. 1978.
Pollen morphology of selected Tritiaceae and two intergeneric
hybrids. Pollen H_Spores 20:145-156.

Rambourg, A. 1971. Morphological and histochemical aspects of glycoproteins
at the surface of animal cells. Int. Rev. Cytol. 31257-113.

 

Rodriguez-Garcia, M. I. 1978. Elektronenmikroskopische untersuchungen
von tapetum und meiocyten wahrend der mikrosporogenesis bei Scilla
nonsepipta. Pollen H_Spores 20:467-484.

 

Roland, J. C. 1973. The relationship between the plasmalemma and the
plant cell wall. Inter. Rev. Cytol. 36:45-92.

 

Roland, J. C. and B. Vian. 1971. Reactivite du plasmalemme vegetal.
Etude cytochimique. Protoplasma 73:121-137.

 

Rowley, J. R. 1962. Nonhomogenous sporopollenin in microspores of
Hpg annua. Grana Palynol. 3(3):3-20.

 

Rowley, J. R. 1964. Formation of the pore in pollen of Hpg_annua.
In Pollen Physiology and Fertilization, H. Linskens, ed. Amsterdam:
North Holland Publishing Co., pp. 59-73.

 

Rowley, J. R. 1973a. Formation of pollen exine bacules and microchannels
on a glycocalyx. Grana 31:129-138.

Rowley, J. R. 1973b. Translocation through the pollen wall. .H. Ultra-
struct. .ng. 44:449-450.

Rowley, J. R. 1975. The permeability of the pollen wall to exogenous
protein tracers. .H. Ultrastruct. Res. 50:394.

 

Rowley, J. R. 1976. Dynamic changes in pollen wall morphology. In
The evolutionary significance pf_the exine (Linnean Soc. Symp.

 

 

101

Series Hp._l), I. Ferguson and J. Muller, eds. London: Academic
Press, pp. 39-65.

Rowley, J. R. and A. 0. Dahl. 1977. Pollen development in Artemisia
vulgaris with special reference to glycocalyx material. Pollen
H Spores 19:169-284.

Rowley, J. R. and J. J. Skvarla. 1974. Plasma membrane-glycocalyx
origin of Ubisch body wall. Pollen H_Spores 16:441-445.

 

Rowley, J. R. and J. J. Skvarla. 1975. The glycocalyx and initiation
of exine spinules on microspores of Canna. Amer. H._pr. 62:479-
485.

Skvarla, J. J. and D. A. Larson. 1966. Fine structural studies of
Zea mays pollen. I: Cell membranes and exine ontogeny. Amer.
EFT'HHH, 53:1112-1125.

Southworth, D. 1971. Incorporation of radioactive precursors into
developing pollen walls. In Pollen, development and h siolo ,
J. Heslop-Harrison, ed. London: Butterworths, pp. 115-120.

Steer, M. 1977a. Differentiation of the tapetum in Avena. I. The
cell surface. .H. Cell H31. 25:125-138.

Steer, M. 1977b. Differentiation of the tapetum in Avena. II. The
endoplasmic reticulum and Golgi apparatus. 1H. Cell H31. 28:71-
86.

 

 

Thomson, W. W. 1967. Electron microscope studies on some modifications
of the plasmalemma in oranges. .H. Ultrastruct. Res. 17:475-486.

 

Ubisch, G. V. 1927. Zur entwicklungsgeschichte der antheren. Planta
3:490-495.

Vazart, B. 1970. Morphogenese du sporoderm et participation des mito-
chondries a la mise en place de la primexine dans la pollen de
Linum usitatissimum L. Compt. Rend. Hebd. Seances-Acad. Sci.
270:3210-3212.

 

Warmke, H. E. and S. L. J. Lee. 1977. Mitochondrial degeneration in
Texas cytoplasmic male-sterile corn anthers. .H. Heredity 68:213-
222.

Willison, J. H. M. and E. C. Cocking. 1972. The production of micro-
fibrils at the surface of isolated tomato-fruit protoplasts.
Protoplasma 75:397-403.

 

102

Winzler, R. J. 1970. Carbohydrates in cell surfaces. Inter. Rev.
Cyt l. 29:77-125.

 

Young, 8. A., J. Schulz-Schaeffer and T. W. Carroll. 1979. Anther
and pollen development in male-sterile intermediate wheatgrass

plants derived from wheat x wheatgrass hybrids. Canad. H. pr.
57:602-618.

 

 

MICHIGAN STATE UNIV. LIBRARIES
11111111111111111HHHWIW11111111111111Hill
31293010630451