| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Anatomy and Morphology |
Department of Plant Biology, 228 Plant Science, Cornell University, Ithaca, New York 14853 USA
Received for publication March 4, 2003. Accepted for publication May 30, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: abortive ovules • acorn • anatomy • Cerris • Fagaceae • flower • fruit • Quercus • seed
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, 1989
; Hufford, 1992
), molecular characters only (Chase et al., 1993
; Manos and Steele, 1997
; Chen et al., 1998
; Kallersjo et al., 1998
; Qiu et al., 1998
; Soltis et al., 2000
), as well as combined morphological and molecular characters (Nandi et al., 1998
). Families in this clade include Betulaceae (birches), Casuarinaceae (she-oaks), Fagaceae (oaks, beeches), Juglandaceae (walnuts), Myricaceae (wax myrtles), and Nothofagaceae (southern beeches), as well as the monotypic families Rhoipteleaceae and Ticodendraceae. Termed the "higher" Hamamelidae (henceforth referred to as the HH clade; Crane and Blackmore, 1989
) after the subclass Hamamelidae of Cronquist (1981) and Takhtajan (superorder; 1997), this clade was later described as "Fagales" (APG, 1998
); a name that has historically been applied to circumscriptions within the HH clade. To avoid potential taxonomic confusion, the unambiguous term HH clade will be used in this study. All flowers produced in this clade are unisexual, with pistillate flowers, interpreted as inferior and syncarpous, maturing into fruits that are pseudomonomerous with only one seed (Fig. 1). Some taxa produce only one ovule per flower, but more commonly, multiple ovules (one or two ovules per carpel) are formed in the flower and all but one abort before fruit maturation. These abortive ovules are in an apical position (distal) in the mature fruit in all the taxa found in the HH clade except for certain groups within the genus Quercus (the oaks) of the family Fagaceae, which can have apical abortive ovules, basal abortive ovules, or abortive ovules with a variable position that are neither strictly basal nor apical and are described as lateral (Table 1). Even within these different oak groups, ovule position can be variable: for example, the red oaks (Quercus section Lobatae) typically have apical abortive ovules, yet at least 10 species in Mexico and Central America have lateral to basal abortive ovules. (Trelease, 1924
; Muller, 1942a
; K. C. Nixon, Cornell University, unpublished data).
|
|
; Trelease, 1924
; Muller, 1942b
; Nixon, 1993
, 1997
) also making use of this feature to distinguish among groups in the oaks, including the most recent complete treatment of genus Quercus by Camus (1934–1954
) who divided the genus into two subgenera, Cyclobalanopsis and Quercus. In the Flora of China (Huang et al., 1994), Cyclobalanopsis is treated at the level of genus, and although this treatment does not conflict with accepted natural groups in the oaks, the current study will follow the classification and ranks of Camus according to her monograph (Table 1), with some modifications due to priority of names (Voss, 1981
). There is relatively little known about the anatomy and maturation of the flowers/fruits in subgenus Cyclobalanopsis, so the remainder of this paper will focus on subgenus Quercus.
Basal abortive ovule position (e.g., Fig. 2) figured prominently as a putative synapomorphy and derived character in the morphological analyses that proposed uniting the white oaks of subgenus Quercus section Quercus and the white oaks of the Cerris group (including Q. ilex) into a larger subgenus Quercus section Quercus sensu lato (s.l.) (Fig. 3; Nixon, 1984
, 1993
). Intersectional hybridization events between these two groups, which are very rare in subgenus Quercus, of both natural and artificial origin (Cottam et al., 1982
; Boavida et al., 2001
) have also focused on the similarities between these two sections. So far, molecular analyses based on ITS sequence data have not supported a circumscription of subgenus Quercus section Quercus s.l., with cladograms from these analyses separating the "white" oaks into two separate clades and basal abortive ovules independently derived in both groups (Figs. 4, 5; Manos et al., 1999
, 2001
). Based on these analyses, this article will treat the "white" oaks s.l. as two discrete entities and refer to them as section Cerris (entirely Old World, including the Ilex, Suber, and Cerris groups) and section Quercus sensu stricto (s.s.) (both Old World and New World). In order to understand the significance of basal abortive ovule position and how it relates to the sections Cerris and Quercus s.s., additional data on the processes that underlie fruit development is needed.
|
). Although there have been numerous studies of flower and fruit maturation in the oaks, the studies that sampled species in section Cerris only documented morphological features (Kaul, 1985
), limited the descriptions to diagnostic features, and did not describe maturation patterns (Camus, 1934–1954
; Soepadmo, 1968
) or only described fruit development up to early embryogenesis (Corti, 1954
, 1955
, 1959
; Scaramuzzi, 1960
; Bianco, 1961
; Boavida et al., 1999
). To obtain a complete developmental series of fruit maturation in section Cerris and to allow direct comparison of features described in the Mogensen study, we sampled a species of section Cerris, a cultivated Quercus acutissima Carruth. individual, over a growing season. This species is native to temperate East Asia and is commonly planted as an ornamental in North America. Commonly called the Sawtooth Oak, this species exhibits consistent biennial fruit maturation (Fig. 6), as opposed to some other species in section Cerris that have been reported to exhibit both annual and biennial maturation patterns (Corti, 1955
; Elena-Rossello et al., 1993
; Boavida et al., 1999
).
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Fixation
Collection materials were fixed for a minimum of 7 d in a dehydrating FAA formulation (70% ethanol, 5% acetic acid, 5% formalin) to promote rapid penetration into the fruits (Johansen, 1940
; O'Brien and McCully, 1981
; Ruzin, 1999
) and were then transferred to 70% ethanol for long-term storage.
Pre-infiltration, infiltration, and embedding
Specimens were dehydrated, pre-infiltrated, infiltrated, and embedded using existing protocols (Gerrits and Smid, 1983
; Gerrits and Horobin, 1996
) for the Technovit 7100 kit (Kulzer, Germany), which uses glycol methacrylate as the embedding medium.
Sectioning
Embedded specimens were sectioned on a manual microtome in increments of 4 µm using Ralph-Bennett glass knives (Bennett et al., 1976
). Sections were placed in sequence on a pool of water on a microscope slide and allowed to air dry on a slide warmer set to 60°C. No pretreatment of the slides was necessary except for a swipe with 70% alcohol to remove dust. The next day, the slides were prepared for staining. At least one slide from each specimen was set aside as a control or for later comparative staining techniques. A more detailed account of the methods used in this study is in preparation and will be published separately.
Staining
A staining solution of toluidine blue O (TBO) 0.05% was prepared from a standard recipe (Feder and O'Brien, 1968
) with an acetate buffer (pH 4.4) substituted for the benzoate buffer. Handled this way, TBO stains lignin and sclerified structures a light aquamarine or turquoise color, cellulosic structures stain a dark blue, and proteins stain a red-purple to violet. No mounting media or coverslips were used, and immersion oil was placed directly on the sections for microphotography (O'Brien and McCully, 1981
). A combination of aqueous alcian blue and weak aqueous safranin O (Ruzin, 1999
) was used to verify that the structures stained by TBO could be compared to previous studies that used fast green and safranin O stains. Acid fuchsin stain was used to verify the staining of lignin by TBO (O'Brien and McCully, 1981
).
Photography
Photographs of hand-sections were obtained with a Wild dissecting microscope (Leica, Wetzlar, Germany) using a Nikon 995 digital camera (Tokyo, Japan) and a microscope uniadapter (ECO-1832D, Zarf Enterprises, Spokane, Washington, USA). Photomicrographs were obtained with an Olympus BX60 compound microscope (Tokyo, Japan) using video-capture under the control of a microcomputer, with scale calibrated using a stage micrometer.
Mapping characters
The species from the current study (Quercus acutissima) and the Mogensen (1965)
study (Q. alba, Q. velutina) were substituted for their respective sections on the appropriate terminals of a tree derived from the most recent molecular analysis that includes the major groups in Quercus (Manos et al., 2001
). Characters of fruiting habit, flowers, fruits, and seeds were scored and optimized on the tree using standard optimization procedures (Fitch, 1971
) in WinClada (Nixon, 1999–2002
). All 20 characters were nonadditive, with 16 binary and four multistate characters. The matrix and character coding are available from the first author.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
A more detailed format that is designed to facilitate easy recovery of the data as a parallel series of features within each developmental stage is presented in the Appendix.
Flowers of current year
Stage 0, Winter dormancy: 21 March–25 April
Pistillate flowers of the current year were either not yet differentiated on the inflorescence axis or were differentiating on the inflorescence axis within closed terminal buds.
Stage 1, Bud break: 2–9 May (Figs. 7–9)
At bud break, there were two subopposite flowers on each inflorescence axis in most cases, although some collections had up to five flowers arranged as decussate, subopposite pairs. Each pistillate flower was in the axil of one large bract that was usually flanked by two ephemeral secondary bracts on the inflorescence axis. The cupule could be seen subtending the flower with one to two scale cycles (Fig. 7). The most prominent features of the flowers were the (usually) three styles that were slightly recurved, tapering to slightly flared apices with scattered druses visible in section. Stigmatic surfaces on the styles stained an intense violet and were present from the style apices down adaxial central furrows to the point where the styles fused together (Figs. 7–9). Ungerminated pollen grains, 24–32 µm in diameter, were present on the stigmatic surfaces other as well as other surfaces of the flowers (Figs. 7–9). Unicellular-lignified and multicellular-glandular trichomes and a blue-staining surface distinguished the perianth lobes from the styles in section. At the base of the pistil, the ovary was not yet differentiated or distinguishable from the style bases (Fig. 7).
|
Stages 3–7, Quiescence: 23 May–15 August
The styles and perianth senesced completely in stage 3, with cells variously losing cohesion, becoming compressed, suberized, or filled with tannins. For the rest of the current growing season, the flowers remained relatively unchanged except for some enlargement of the ovary at the base of the styles.
Stage 8, Quiescent: 29 August–26 September (Fig. 14)
Before the onset of winter dormancy, each pistillate flower appeared enveloped within a cupule with only the styles and perianth exserted. Cupules at this stage had at least six cycles of awl-shaped, imbricate scales. The outermost scales were highly sclerified with the scale tips appressed to the pistil below the perianth (Fig. 14). The styles and perianth were often broken so that only the bases of the styles remained. In cases where the styles or perianth were still present at the apex of the flower, the outermost cells were either filled with tannins or suberized (Fig. 14; at arrows). Within the styles, cells internal to the tannin-filled or suberized cell rows were highly vacuolated and some had druses. At the base of the styles, some cells remained parenchymatous and stained intensely (Fig. 14). The ovary was still not well-differentiated at the base of the styles, but locular spaces delimited by intruded septa were more obvious. Unicellular trichomes were now present on the walls of the locules (Fig. 14). No ovule primordia were visible.
Flowers of previous year
Stage 0, Winter dormancy: 21 March–25 April (e.g., see Fig. 14)
At the conclusion of winter dormancy, the pistillate flowers formed in the previous year were very similar to those of stage 8 of the current year (see earlier).
Stages 1–2, Quiescent: 2–16 May
No change from stage 0.
Stage 3, Resume development: 23 May (Figs. 15, 16)
In the cupule, new scale primordia were seen as well as vascular tissue and sclereid clusters in the cupule base (Fig. 15). Styles and perianth lobes were most often broken, with portions that remained sclerified or suberized. The ovary, with its locules and developing ovules, could now be distinguished from the style bases (Fig. 15). Some endocarp differentiation in the ovary could be seen as a uniseriate epidermis. Unicellular, lignified and unlignified trichomes filled the locules (Figs. 15, 16). Two collateral ovule primordia were present on each septum, oriented perpendicular to the axis of the ovary (Fig. 15; at arrow). At later stages, these became pendulous with the apices of the primordia oriented toward the base of the ovary (Fig. 16). Six ovule primordia were normally present, although one specimen only had four ovules in two locules. Each funiculus appeared sessile.
|
Stage 5, Megasporogenesis: 27 June (Figs. 21, 22)
Cupule and scales remained unchanged relative to Stage 4. The ovary remained unchanged relative to Stage 4 except for the differentiation of isolated sclereid clusters in the abscission zone at the base of the ovary (Fig. 21). The ovules were expanded to fill the locules by the end of this stage, and most of the ovules examined had one megaspore mother cell that was visible as a lighter-staining cell 2–3 cells below the apex of the nucellus (Fig. 22). Ovules were crassinucellar with the inner and outer integuments 5–6 cells wide at the level of the megaspore mother cell.
Stage 6, Megagametogenesis: 4–11 July (Fig. 23)
Cupule and scales remained unchanged relative to Stage 4. Differentiation of the exocarp was visible with an external uniseriate epidermis and 3–4 rows of small and densely staining cells oriented perpendicularly to the epidermis, forming a palisade layer. The numerous cells (50–100) in the mesocarp were not as densely staining as the exocarp, with some cells highly vacuolated, some with druses, and some vascular tissue present. The 5–10 cells that formed the endocarp were parenchymatous and contained tannins. Lignified unicellular trichomes filled in the spaces in the locules around the ovules. Vascular tissue could be seen in the base of each septum. Embryo sacs of various developmental stages were visible in many of the ovules as lighter-staining tissue with free nuclei (Fig. 23). As the embryo sacs developed, each nucellus deteriorated in a basipetal direction, beginning with cells that were immediately basal to the embryo sac, to form a caecum (Fig. 23). By the end of this stage, the ovules had expanded to fill the locules.
Stage 7, Mature embryo sacs: 18 July–15 August (Figs. 24–30)
By the end of this stage, the cupule base was more sclerenchymatous, with more numerous and larger sclereid clusters and more cells showing tannin accumulation. The outermost scale cycles became reflexed away from the ovary due to adaxial surface expansion (25 July; e.g., see Fig. 6). The exocarp showed basipetal sclerification of the palisade layer beginning from the base of the styles (25 July) to about 50% of the ovary. The mesocarp was mostly parenchymatous, with some highly vacuolated cells, some scattered druses, and some vascular tissue present. Below the style bases, an apical portion of the mesocarp was sclerified into a cone-shaped structure (1 August; Fig. 24, at arrows). The sclerification of the abscission zone was mostly complete, forming a disklike structure at the base of the ovary composed of clusters of sclereids that were interdigitated with parenchymatous cells (Fig. 24). Lignified unicellular trichomes from the endocarp filled spaces in the locules around the ovules (Figs. 24, 25, 27, 30), and the parenchymatous septa that were present from the apex to the base of the ovary had significant vascular tissue differentiation in their bases (Fig. 30). In each ovule with an embryo sac, the caecum was completely filled by the embryo sac (Figs. 24–29) except for a remnant of the nucellus, the basal lateral densely staining hypostase (Fig. 29). The egg apparatus with two synergids and an egg cell was present at the apex of the ovule (Figs. 26, 28), with the central cell usually midway down the caecum (Figs. 26, 29). The antipodals were not located in any of the samples. Most of the flowers had at least one ovule with a viable embryo sac with two syngergids, an egg, and a central cell. Ovules that did not have a viable embryo sac were either underdeveloped or were aborted and were smaller than the ovules with viable embryo sacs. Toward the end of this stage, endosperm was usually visible in at least one ovule, even if zygotes or embryos could not be located in the embryo sac. When endosperm was present in an ovule, its inner integument was either highly vacuolated or broken down around the embryo sac so that only the outer integument remained, except for a small plug of the inner integument in the micropyle (Figs. 24, 25). By the end of this stage, one ovule with endosperm was usually much larger than the others. The rest of the ovules showed signs of abortion, as inner integuments and internal structures coagulated to form amorphous, dark-staining structures contained within outer integuments (Fig. 27, at arrows). A zygote was tentatively identified in one sample (1 August, not illustrated).
|
|
study onto a tree derived from molecular data are presented in Fig. 42. Tree length for the morphological characters alone was 31 steps, with a consistency index (CI) of 0.74, and a retention index (RI) of 1.0. Total ambiguities (characters coded as either missing values or as full polymorphisms) in the matrix was 21%, and 12 characters were uninformative (either pleisiomorphic or autapomorphic) as mapped on this tree. Of the remaining characters that were informative, two characters appear as parallelisms between Q. acutissima (section Cerris) and Q. alba (section Quercus s.s.): character 13 (locules are filled with ovules for much of development; CI = 0.50) and character 16 (basal abortive ovules; CI = 0.66).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
work that provided a framework for the current study.
Shared features among sections Cerris, Quercus s.s., and Lobatae
For all three species, Q. acutissima, Q. alba, and Q. velutina, the meristems of the pistillate inflorescence axes axes form at the end of the growing season and remain quiescent in terminal buds until the following spring. These reproductive meristems are usually indistinguishable from vegetative meristems until a few weeks before the buds break. The pistillate inflorescences, ranging in length from almost sessile up to several centimeters long, are formed on the main branch in the axils of leaves distally to the staminate inflorescences. The pistillate inflorescence axis bears one to several flowers (as observed in a wider sample of species) that are mostly sessile and in decussate subopposite pairs (Trelease, 1924
; Langdon, 1939
; Muller, 1942b
; Turkel, 1950
; Corti, 1954
, 1955
, 1959
; Bianco, 1961
; Sharp and Sprague, 1967
; Soepadmo, 1968
; Fey, 1981
; Kaul, 1985
; Boavida et al., 1999
).
At anthesis of the pistillate flowers, the ovary is an insignificant part of the pistil, and each flower is essentially three styles with a perianth (Fig. 7), with the cupule present as a toruslike structure at the base of the pistil before the first few cupule scales are formed (Turkel et al., 1955
; Scaramuzzi, 1960
; Bianco, 1961
; Fey, 1981
; Cecich, 1997
; Boavida et al., 1999
, 2001
). Stigmatic surfaces are restricted in area to the style apices and along adaxial grooves on the styles and consist of the nonpapillate, "dry" type with a pellicle (Heslop-Harrison, 1971
). The styles senesce within a month after pollination. Although pollen tube penetration into the styles (Turkel et al., 1955
; Stairs, 1964
; Cecich, 1997
) was not directly observed in the current study, the adherence of Quercus-type pollen to the stigmatic surfaces, attached empty pollen grains that have presumably discharged their microgametophytes, and one specimen with pollen tube growth within the style (Fig. 11) inferred that pollination was successful.
Species with annual fruit maturation (e.g., Q. alba) have a delay of approximately 1 mo between pollination and syngamy, while species with biennial fruit maturation (e.g., Q. acutissima and Q. velutina) are quiescent 
13 mo between pollination and syngamy. Despite the difference in quiescence time, ovule initiation and maturation in the flowers are similar in all three species. Typically, there are three locules in the pistillate flowers, with each locule containing two collateral ovules that at maturity are oriented with the micropyle toward the apex of the flower (epitropous) with the raphe on the ventral side of the ovule (Figs. 17, 18, 21). This final ovule orientation is best described as semianatropous (Corti, 1954
). Soon after the nucellus is distinguishable within the developing ovule, one or more megaspore mother cells, or megasporocytes, can be seen within the apex of the nucellus (Benson, 1893
; Langdon, 1939
; Turkel, 1950
; Rebuck, 1952
; Hjelmqvist, 1953
; Corti, 1954
, 1955
, 1959
; Turkel et al., 1955
; Scaramuzzi, 1960
; Bianco, 1961
; Stairs, 1964
; Mogensen, 1965
; Fey, 1981
; Boavida et al., 1999
). Only one megaspore mother cell was found in each ovule in Q. acutissima (Figs. 18, 21, 22). Although not directly observed in the current study, megagametogenesis is of the monosporic, Polygonum-type with a seven-celled, eight-nucleate embryo sac (Hjelmqvist, 1953
; Corti, 1959
; Bianco, 1961
; Stairs, 1964
), although several authors have mistakenly described the process as tetrasporic, Adoxa-type (Conrad, 1900
; Bagda, 1952
; Turkel et al., 1955
). The ephemeral antipodals were not detected in the current study, although some authors illustrated the position of these three nuclei as apical to the senescing nucellus (Benson, 1893
; Langdon, 1939
; Turkel, 1950
; Rebuck, 1952
; Hjelmqvist, 1953
; Corti, 1954
, 1955
, 1959
; Turkel et al., 1955
; Scaramuzzi, 1960
; Bianco, 1961
; Stairs, 1964
; Mogensen, 1965
; Fey, 1981
; Boavida et al., 1999
). A few studies that examined Q. alba (section Quercus s.s.), Q. rubra from section Lobatae, and other species from section Cerris reported tracheids or procambium-like tissue in the hypostase or surrounding tissue (Benson, 1893
; Langdon, 1939
; Corti, 1954
, 1955
; Turkel et al., 1955
), but this was not observed in the current study.
During the maturation of the ovules and embryo sacs, the pollen tubes (Figs. 10, 11), are quiescent in a basal area of the styles that remains parenchymatous. After megasporogenesis and during megagametogenesis, the pollen tubes resume growth and enter the ovules through the micropyle (porogamy; Benson, 1893
; Kerner Von Marilaun, 1895
; Conrad, 1900
; Klebelsberg, 1910
; Hjelmqvist, 1953
; Corti, 1954
; Mogensen, 1972
; Cecich, 1997
; Boavida et al., 1999
). Pollen tube growth into the ovules as well as the resulting syngamy was not directly observed in the current study, but there was no evidence of ovular disruption that would indicate that the pollen tubes entered the embryo sac other than through the micropyle. All of the ovules in the ovary usually have embryo sacs of various developmental stages at the time of syngamy, but only one ovule matures into the seed. The abortive ovules can be distinguished in the early stages of abortion from the surviving ovule by the loss of cohesion of the embryo sac and the coagulation of the inner integument and contents into a densely staining structure as seen in Q. acutissima (Figs. 27, 30, 36, 37). The abortive ovules remain attached to their respective septal remnants (Conrad, 1900
; Langdon, 1939
; Corti, 1954
, 1955
, 1959
; Turkel et al., 1955
; Scaramuzzi, 1960
; Bianco, 1961
; Stairs, 1964
; Mogensen, 1965
, 1972
, 1973
, 1975a
, b
; Brown, 1971
; Brown and Mogensen, 1972
; Fey, 1981
; Boavida et al., 1999
).
For the remainder of the discussion, any ovule with a multicellular embryo will be referred to as a seed, with its respective ovary now a fruit. Multicellular embryos were common in this study and assumed to be the result of syngamy and not apomixis or parthenogenesis (Johri, 1984
). In the embryo sac, the free-nuclear or coenocytic endosperm proliferates before the first division of the zygote and can have vacuoles of various sizes (Hjelmqvist, 1957), although only one specimen showed possible evidence of a vacuole in Q. acutissima (see Fig. 33, outline). In one collection (29 August) of Q. acutissima, fruits with almost mature seeds (e.g., Fig. 38) were collected at the same time as a fruit that had a heart-shaped embryo (Figs. 33, 34), supporting the observation made by other authors that the level of maturity in fruits from an individual tree can vary widely at any given time (Conrad, 1900
; Langdon, 1939
; Corti, 1954
, 1955
, 1959
; Turkel et al., 1955
; Scaramuzzi, 1960
; Bianco, 1961
; Stairs, 1964
; Mogensen, 1965
; Fey, 1981
).
In the current study, a concentration of cellular endosperm was observed at the chalazal end of the embryo sac in the more mature fruits and may be haustorial (Figs. 35, 40). Adjacent to this chalazal concentration of cellular endosperm is the hypostase. Several authors have discussed the possibility that the hypostase functions as a haustorium for the growing embryo (Benson, 1893
; Langdon, 1939
; Hjelmqvist, 1953
; Corti, 1954
; Mogensen, 1973
). Whether the chalazal cellular endosperm or the hypostase function separately or together as a haustorium is not known at the present time.
Differences between section Cerris and sections Quercus s.s. and Lobatae
A comparison of the historically important features of section Cerris to both the sections Quercus s.s. and Lobatae is presented here with features unique to section Cerris described first, followed by those features that group section Cerris with either section Quercus s.s. or section Lobatae (Table 2).
|
; Camus, 1934–1954
; Muller, 1942b
; Soepadmo, 1968
, 1972
; Maleev, 1970
; Kaul, 1985
; Nixon, 1997
).
It is possible to compare the time delay between pollination and syngamy in both annual- and biennial-fruiting species in subgenus Quercus if the time of quiescence of the biennial species (1 yr) is subtracted from the time between pollination and syngamy. The approximate 1 mo time delay that remains is comparable across both annual- and biennial-fruiting taxa (Conrad, 1900
; Corti, 1954
; Turkel et al., 1955
; Stairs, 1964
; Mogensen, 1965
; Sharp and Sprague, 1967
; Cecich, 1997
). Syngamy usually occurs 1 mo after pollination in both the annual-fruiting (e.g., Q. alba) and biennial-fruiting species (e.g., Q. velutina), but in the current study, no zygotes or embryos could be located in the fruits of Q. acutissima for at least 2 mo after pollination occurred. A 3–5 mo delay in syngamy for section Cerris has been reported previously for an annual-fruiting form of Q. aegilops (Scaramuzzi, 1960
). It is possible that the delay in syngamy observed in the current study may be environmental and is not a consistent phylogenetic characteristic. Samples of other Quercus species that were collected for anatomical studies at the same time and from the same study area as Q. acutissima may provide some answers to this particular question.
One feature that was noticeable early in the development of the pistillate flower in Q. acutissima was the lignification of the trichomes in the locules. A large number of trichomes were lignified before megasporogenesis and fractured during sectioning to form artifact spaces in the ovary (e.g., Figs. 17, 18, 21). Because lignification patterns of endocarp trichomes have not been followed in previous studies, this feature needs to be examined more closely in the future. In a preliminary examination of the original slides provided by Mogensen (1965)
, endocarp trichomes do not appear to lignify until after syngamy, with trichomes in Q. alba lignifying prior to Q. velutina.
In the current study, one of the most mature fruits collected from Q. acutissima contained a seed that had an embryo shoot apex with leaf or scale primordia buttresses (Fig. 39). In section Quercus s.s., the shoot apex forms several leaf or scale primordia before the acorn falls from the tree, and in studied species of section Lobatae, the shoot apex remains naked (Mogensen, 1965
; Sutton, 1969
; Sutton and Mogensen, 1970
). A larger sampling of mature acorns and the examination of seedling germination morphology in Q. acutissima and other species from section Cerris may reveal whether these buttresses are a consistent state that is intermediate between section Lobatae and section Quercus s.s. or if the embryo shoot apex in section Cerris later forms leaf or scale primordia before falling from the tree.
Potential synapomorphies between sections Cerris/Lobatae and sections Cerris/Quercus s.s
Style morphology of the pistillate flower in subgenus Quercus has historically been an important character in distinguishing among the major groups (Oersted, 1871
; Trelease, 1924
; Camus, 1934–1954
; Reece, 1938
; Muller, 1942b
; Tillson and Muller, 1942
; Soepadmo, 1968
; Maleev, 1970
; Tucker, 1980
; Nixon, 1984
, 1993
; Kaul, 1985
). The styles in Q. acutissima, with slender and thin apices that remain straight or may become slightly curved, have closest affinities to the styles seen in section Lobatae (e.g., Q. velutina). The styles do not lengthen appreciably in Q. alba with the style apices spreading out laterally to become rounded and blunt.
The pistillate flowers in both Q. acutissima and Q. velutina have perianth lobes that are longer than they are wide and can extend up the style column and be mistaken for styles in longitudinal section (e.g., Figs. 10, 12). The perianth in Q. alba is considerably shorter, with tepals as long as they are wide.
Although the perianth tepals of Q. acutissima and Q. alba are not morphologically similar in height/width proportions, they are morphologically similar in that the perianth lobes do not invaginate at the base to form a skirt or flange. This flange can be seen in biennial-fruiting taxa in the section Lobatae, such as Q. velutina, which seals the parenchymatous flower during quiescence by interlocking the perianth with the outermost cupule scales (Oersted, 1871
; Trelease, 1924
; Camus, 1934–1954
; Muller, 1942b
; Soepadmo, 1968
; Tucker, 1980
; Nixon, 1984
, 1993
; Kaul, 1985
).
Distinguishing between the different layers in the fruit wall can be arbitrary, particularly in the early stages of ovary development, but for discussion purposes, the mature fruit layers described by Harz (1885) and Soepadmo (1968)
for the genus Quercus were used. The exocarp and mesocarp of Q. acutissima can be compared to the exocarp and mesocarp of Q. alba and Q. velutina as separate layers, but when the two layers combine at fruit maturity to form the "shell" of the nut, the structure seen in Q. acutissima is best interpreted as a different character state when compared to the other two species. The exocarp palisade layer in Q. acutissima and Q. alba is thin with short sclereids and is a minor component of the fruit wall when compared to the robust and dense palisade layer of Q. velutina. The mesocarp in Q. acutissima and Q. velutina has some scattered groups of sclereids at maturity, but is relatively parenchymatous at maturity compared to the mesocarp in Q. alba. The total composition of the "shell" of the nut at maturity results in three different fruit wall compositions across the three species. At maturity, the exocarp in Q. velutina, the mesocarp in Q. alba (Mogensen, 1965
), or the combination of both layers in Q. acutissima makes up the mature "shell." Because this is one of the few studies to compare the patterns of fruit wall differentiation across several sections, this observation needs further verification with more sampling.
The absence or presence of a pubescent or tomentose endocarp, historically cited as a useful character in differentiation among the sections of subgenus Quercus, can be problematic unless applied consistently. For example, when the fruit of section Quercus s.s. is opened, the seed coat often detaches from the seed, adhering to the fruit wall, and the inside of the seed coat is then mistaken for the endocarp (Nixon, 1984
, 1997
). The true endocarp in section Quercus s.s. is mostly glabrous at maturity, although sometimes patches of short and sparse pubescence can be found. The trichome type, size, and form seen in Q. acutissima are similar to trichomes in Q. alba, but in Q. acutissima the endocarp remains tomentose at maturity. In contrast, the endocarp in Q. velutina is very tomentose and is densely covered by trichomes
