THE SOUTHWESTERNNATURALISTTHENATURALIST SOUTHWESTEN46(3):282-2946(3) :282 294SEPTEMBER SEPTEMBER20012001

DIAGNOSTIC PHYTOLITHS FOR A PONDEROSA PINE-BUNCHGRASS COMMUNITY NEAR FLAGSTAFF, ARIZONA

BECKYK. KERNS*

School of Forestry,College of EcosystemScience and Management, Northern Arizona University, Flagstaff, AZ 86011-5018 Present address:Pacific NorthwestResearch Station, United States Department of Agriculture,Forest Service, Corvallis, OR 97331 * Correspondent:[email protected]. us

ABSTRACT-Phytolithanalysis could play an important role in understanding vegetation dynamics in southwestern ponderosa pine ()forests, which have been dramatically altered by fire suppression and other factors. My objectives were to develop a phytolith reference collection and classification system for a ponderosa pine-bunchgrass community found near Flagstaff, Ari- zona. I examined 27 species of grasses found in and around the study area and ponderosa pine for diagnostic phytoliths. Twenty other species common to the area were examined for redundant phytolith forms. Eight phytolith forms were identified, including a diagnostic phytolith for pon- derosa pine, the spiny body. The general subfamily system validated by numerous re- searchers is applicable to this community. Examination of phytolith shape frequencies show that for 7 species in the subfamily , and 1 species in the , very few (0 to 5%) nondiagnostic phytolith forms were present. Nondiagnostic phytoliths, particularly rondels, were more common (7 to 22%) for the 3 species from the subfamily. This result is consistent with the observation by other authors that all grasses produce rondel. forms and indi- cates that rondels will be over-represented in phytolith assemblages in comparison to actual veg- etation. The ponderosa pine spiny body appears to be a useful diagnostic for this area and vege- tation reconstructions using soil phytolith assemblages based on the system developed in this study could be used to understand grass-tree and grass vegetation dynamics.

RESUMEN-Los estudios de fitolitos pueden jugar un papel importante en la comprensi6n de la dinamica de la vegetaci6n en los bosques de pino ponderosa (Pinus ponderosa) del sudoeste de los Estados Unidos, los que han sufrido modificaciones drasticas debido a la eliminaci6n de in- cendios y a otros factores. Mis objetivos fueron desarrollar una colecci6n de referencia fitolita y un sistema de clasificaci6n para una comunidad de pino ponderosa y zacate cercana a Flagstaff, Arizona. Examine 27 especies de pastos encontradas en y alrededor del sitio de estudio y en pino ponderosa para fitolitos diagn6sticos. Se estudiaron otras 20 especies comunes en el area para ver si presentaban formas de fitolitos redundantes. Se identificaron 8 formas de fitolitos, incluyendo un fitolito diagn6stico de pino ponderosa: el cuerpo espinoso. El sistema general de subfamilias de Poaceas anteriormente presentado por otros investigadores es valido para esta comunidad. Examinaci6n de la frecuencia de formas de fitolitos indic6 que para 7 especies de la subfamilia Pooideae y una especie en la subfamilia Panicoideae presentaron muy pocos fitolitos no diagn6s- ticos (0 a 5%). Estos mismos fitolitos, especialmente los circulares (rondels), fueron mas comunes (7 a 22%) en las tres especies de la subfamilia Chloridoideae. Este resultado coincide con la observaci6n de otros autores de que todos los pastos producen rondels, e indica que es esperable que las formas circulares esten sobre-representadas en el banco de fitolitos en comparaci6n con la abundancia de los pastos en la comunidad vegetal. El cuerpo espinoso del pino ponderosa parece ser una forma diagn6stica 6til para esta zona. El sistema presentado en este trabajo puede usarse para reconstruir la vegetaci6n, y asi entender mejor la dinamica vegetal de pasto-arbol y pasto.

The use of diagnostic phytoliths, or phytolith role in environmental reconstruction and in- analysis, is playing an increasingly important terpretation of community dynamics September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrass community 283

(Rovner, 1971; Carbone, 1977; Kurmann, 1985; place phytolith forms within groups character- Jiang and Piperno, 1998; Fredlund et al., 1998; istic of the grass subfamilies: Festucoid (Pooi- Kealhofer and Penny, 1998; McClaran and Um- deae), Chloridoid (Chloridoideae), and Pani- lauf, 2000). Phytoliths are plant silica remains coid (Panicoideae); however, later research in- that can be diagnostic at various taxonomic lev- dicated that a single plant may produce many els and occur throughout the plant kingdom types of phytolith forms (multiplicity) and a (Piperno, 1988). Upon death and decay of particular form may be produced by a number plant material, phytoliths remain in the depo- of different plant taxa (redundancy-Rovner, sitional environment and can be used to ex- 1983). Multiplicity and redundancy complicate amine vegetation history and dynamics. As a the concept of diagnostic phytolith forms. For method for determining prior reference con- example, species in the genus of the Pooi- ditions, phytoliths have several advantages over deae subfamily (C3) produce lobate forms other types of plant fossils such as . For characteristic of the C4 species in the Panicoi- example, they are persistent in oxidizing envi- deae subfamily. However, detailed morpholog- ronments such as soil (Wilding et al., 1977). ical examination and the use of liquid mounts Moreover, it is difficult to use pollen analysis for 3 dimensional phytolith examination has to distinguish ecologically diverse groups with- eliminated confusion regarding Stipa bilobates in the grass family (Ritchie, 1976; Kurmann, (Brown, 1984; Fredlund and Tieszen, 1994). 1985). Phytolith production in gymnosperms often Prior to Euro-American settlement, frequent has been described as poor, and as being a less (2 to 12 per year), low-intensity, surface fires promising area of study for phytolith mor- characterized semi-arid ponderosa pine (Pinus phology compared to flowering (Piper- ponderosa)-bunchgrass communities typical of no, 1988); however, some researchers have iso- the Intermountain West (Cooper, 1960; Dieter- lated diagnostic conifer forms, most notably ich, 1980; Covington and Moore, 1994a, 1994b; the bordered pit tracheary elements present in Swetnam and Baisan, 1996; Covington et al., Pinus and Picea, and asterosclereids produced 1997; Fule et al., 1997). Beginning in the early by Pseudotsuga menziesii (Brydon et al., 1963; part of this century, fire suppression, overgraz- Klein and Geis, 1978). Bozarth (1993) exam- ing, and a warm and wet climatic period led to ined 4 species of conifers and identified several an irruption of pine regeneration (Cooper, forms which occur only in conifers. These in- 1960; White, 1985; Savage et al., 1996; Mast et cluded thin plates with wavy margins common al., 1999). Concurrent changes in understory and possibly unique to Picea glauca, silicified plant communities are likely; however, quanti- tracheary elements with bordered pits found tative studies regarding prior conditions for in Pinus banksiana and P glauca, and the spiny understory plant communities in ponderosa irregular body diagnostic to P banksiana. Nor- pine forests of the Southwest are lacking. Un- gen (1972:50) observed a possible diagnostic derstanding vegetation history is a crucial step form for ponderosa pine he called the "spiny in determination of reference conditions, a elliptical blob." critical component of ecosystem management The first objective of this study was to ex- and restoration ecology (Kaufmann et al., amine local grass species and establish a work- 1994; Morgan et al., 1994; Christensen et al., able Poaceae short-cell classification system in 1996). the context of existing literature. I was also in- could an Phytolith analysis play important terested in developing a diagnostic phytolith role in in understanding vegetation dynamics for ponderosa pine. Although this paper fo- these communities; however, examination of cuses on diagnostic phytoliths from the Po- local flora is before required phytolith analysis aceae and ponderosa pine, 20 other species can be reliably applied. Several researchers common to the area were examined for redun- have Poaceae short-cell investigated diagnostic dant phytolith forms. phytoliths in other ecosystems (Met- Twiss et calfe, 1960; al., 1969; Brown, 1984; METHODS AND MATERIALS-The study area was lo- Mulholland, 1989; Mulholland and Rapp, cated within the Fort ValleyExperimental Forest, 10 1992; Fredlund and Tieszen, 1994). Early re- km northwest of Flagstaff, Arizona. This area was searchers (e.g., Twiss et al., 1969) attempted to chosen because it has never been harvested (hazard 284 The SouthwesternNaturalist vol. 46, no. 3

trees have been felled), is currently closed to grazing All grasses listed in Table 1 were examined for (area was grazed between 1876 and 1910: Covington diagnostic short-cell identification. For many species et al., 1997), and is proximal to an ongoing ecosys- several plants were available for examination; how- tem restoration project (Covington et al., 1997). The ever, some species were limited to 1 sample (Table area consists of gently rolling topography (0 to 5% 1). Because future application of this system was for slopes) with an average elevation of 2,250 m. Mean examination of soil phytolith assemblages, only Po- annual precipitation is 56.7 cm with approximately aceae short-cells were considered, because they are half of that falling as snow in the winter, and the fairly equal in size and silicification and would be other half as monsoonal rains in July through Sep- equivalently resistant to post-depositional degrada- tember (Schubert, 1974). Mean annual air temper- tion. Phytolith shape frequencies of the leaf epider- ature is 7.5?C with an average of 94 frost-free grow- mis were determined for a subsample of grasses. ing days. Soils are mapped as a complex of fine, Species were chosen to represent the most common smectitic Typic Argiborolls and Mollic Eutroboralfs for the study area and also to span a range of sub- that developed on Tertiary basalt flows and cinders families and tribes. If more than 1 plant was collect- (Miller et al., 1995). Present-day plant community ed then the sample was randomly chosen. To pre- structure in the study area is characterized by small pare the slide, ash material was well mixed, and for patches (0.02 to 0.29 ha) of larger old-growth trees each sample, the same amount of material was in clumps of 3 or more interspersed with dense placed on the slide. This amount was determined stands of younger smaller trees (germination after through trial and error using a small scoop with a settlement, ca. 1870), and relict patches (<0.01 ha) mark at the appropriate depth. Slides were then sys- of open areas with bunchgrasses and other herba- tematically scanned until 100 disarticulated short- ceous plants (White, 1985; Kerns, 1999). cells were counted. If the 100th short-cell was count- Grass species were collected from the study area ed in the middle of a field with additional short-cells, and nearby sites, resulting in a collection of 27 spe- the entire field was counted to reduce bias. cies for diagnostic Poaceae short-cell phytolith in- Photographs were made using a LEO 435VP Scan- vestigation (Table 1). Development of this collection ning Electron Microscope at the Northern Arizona was based on personal field experience, consultation University Electron Microscope Facility (Figs. 1-8). with other ecologists and botanists, examination of The entire reference collection of dry ash, micro- herbarium notes, and prior botanical studies. Al- scope slides, and card catalog, is housed by the Eco- though this species list is not exhaustive for grasses logical Restoration Institute, School of Forestry, found in ponderosa pine forests of northern Arizona Northern Arizona University. it does include common species from a wide variety Phytolith Classification-Eight categories of phyto- of habitats. Several introduced species were also col- lith forms from leaf and needle material were iden- lected. To ensure that the diagnostic phytolith form tified, and these identifications were consistent with developed for ponderosa pine was not redundant in existing literature (Metcalfe, 1960; Twiss et al., 1969; other species in the study area, an additional 20 spe- Brown, 1984; Mulholland, 1989; Mulholland and cies of trees, shrubs, forbs, and grass floral material Rapp, 1992; Fredlund and Tieszen, 1994). Forms are were examined (Table 2). A botanist (J. Springer or pictured in Figs. 1-8 and summarized in Table 3 J. M. Rominger, Northern Arizona University) con- along with their representative taxa. Because nu- firmed all plant identifications. merous researchers have published grass short-cell Plant samples were separated into leaf, needle, or classification systems, I tried to use previously pub- seedhead (if present), soaked and washed with de- lished nomenclature to avoid further confusion. In ionized water, and dried at 70?C for 24 h. Samples the following section I provide more detailed de- were then weighed into crucibles and dry oxidized scriptions of the forms and point out usage of terms overnight at 400?C. Initially, ash was washed in 1M for comparative purposes. Table 4 shows phytolith HCL to remove calcium oxalate (M. Umlauf, North- shape frequencies for the species subsampled. ern Arizona University, pers. comm.). However, sam- Description of Diagnostic Forms-Morphological ter- ples did not react to this treatment, and no differ- minology is standard. The adaxial/abaxial outline is ence between washed and unwashed samples was de- the appearance of the phytolith when looking di- tected under the microscope. Therefore, this step rectly down at the top/bottom. The bottom is de- was dropped. Ash material was then mounted in fined as the largest, flattest face, or the face not dec- Canada Balsam. After mounting, slides were placed orated or pointed. The top is opposite to the bot- into a low temperature oven (35 to 40?C) for 48 to tom. Sides are the longest pair of opposite faces, and 72 h and cooled. Slides were examined at 400X us- the ends are the other pair of faces. When I refer to ing a Standard 20 Zeiss biological microscope. I ex- a phytolith shape in cross-section, this is viewing the amined phytolith forms in a semi-liquid mount that phytolith on one of its sides. Conversely, the end allowed the phytolith to be rotated and viewed in 3 view is looking down on one of its ends. dimensions. 1) Rondel. This form is more or less circular to 285 September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrass community

TABLE --Grass species collected and examined for this study. Systematics and common names conform to Allred (1993).

Taxa Species Common name Poaceae Pooideae Aveneae Koeleriamacrantha prairie Junegrass Poeae tectorumab cheatgrass B. ciliatus fringed brome arizonicac Arizona fescue compressat'b Canada bluegrass P fendlerianc( muttongrass P pratensisa,b Kentucky bluegrass Triticeae Agropyrondesertoruma b,d crested wheatgrass A. smithizb western wheatgrass elymoidesce bottlebrush squirreltail Hordeumjubatumbd foxtail barley Stipeaef Stipa comatcf needle-and-thread pringlebTe,g Pringle's needlegrass Aristida purpureab purple three-awn Aristida sp. three-awn Panicoideae Andropogoneae Schizachyriumscopariumb little bluestem Paniceae Panicum bulbosum bulb panicum Chloridoideae Cynodonteae (Chlori- blue grama deae) Eragrostideae Blepharoneurontricholepisc pine dropseed Lycurus phleoidesbd wolftail Muhlenbergia minutissimac least muhly M. montanaF mountain muhly M. rigens deergrass M. virescensbg screw leaf muhly M. wrightizb spike muhly Sporoboluscryptandrusb,d sand dropseed S. interruptusb,g black dropseed a Introduced (United States Department of Agriculture and Natural Resource Conservation Service, 2000). b Only one plant examined. c Very common perennial species in study area. d Not found in study area. e Synonymous with Sitanion hystrix, Hesperostipacomata, and Stipa pringlei, respectively. f Some authors (i.e., Watson and Dallwitz, 1992) place the within the Arundinoideae. g Species not listed in Allred (1993), name conforms with McDougall and Haskell (1973). oval in adaxial/abaxial outline and asymmetric in All the C3 (Pooideae) grasses in this study commonly cross-section (Fig. 1). The top can be flat, horned produce rondels, but Festuca arizonica, Poa fendleri- (2 points), decorated (more than 2 points or com- ana, and Stipa comatahave high (>90%) percentages plex), or crested. The base can be flat, but is more of rondels (Table 4). typically slightly to highly indented. Mulholland 2) Pyramid. The pyramidal form is roughly square (1989) also used the term rondel. This form in- to rectangular in adaxial/abaxial view, edges are cludes the la, Id, le categories ofTwiss et al. (1969), smooth and it is trapezoidal in cross-section. This conical and keel variants of Fredlund and Tieszen form is not a true pyramid because the top is trun- (1994), and Brown's (1984) non-sinuous trapezoids. cated (Fig. 2). Fredlund and Tieszen (1994) use the 286 TIheSouthwestern Naturalist vol. 46, no. 3

TABLE2-Other conifer species, understory shrubs, and forbs examined to check for redundant phytolith forms. Except for Pinus ponderosa, one plant was examined for each species listed. Nomenclature conforms to McDougall and Haskell (1973) and common names are from the PLANTS National Database (United States Department of Agriculture and Natural Resource Conservation Service, 2000).

Family Species Common name Asteraceae Antennaria parvifolia small-leaf pussytoes Bahia dissecta ragleaf bahia Cirsium wheeleri Wheeler's thistle Hieracium fendleri yellow hawkweed Senecio multilobatuse lobeleaf groundsel Solidago sparsiflorca threenerve goldenrod Townsendiasp. Townsend daisey Viguiera multifloraa showy goldeneye Cupressaceae Juniperus monosperma oneseed juniper Cyperaceae Carex occidentalis western sedge Fagaceae Quercusgambelii Gambel oak Pinaceae Abies lasiocarpab'c corkbark fir Pinus edulisc twoneedle pinyon P ponderosa ponderosa pine Pseudotsuga menziesiic Douglas-fir Rhamnaceae Ceanothusfendleri Fendler's ceanothus Rosaceae Potentilla subviscosa Navajo cinquefoil Salicaceae Populus tremuloides quaking aspen Salicaceae Salix scoulerianac Scouler's willow Saxifragaceae Ribes sp.c currant a new name Packera multilobata, Solidago velutina, and Heliomerismultiflora var. multiflora, respectively (United States Department of Agriculture and Natural Resource Conservation Service, 2000). b Abies lasiocarpa var. arizonica (United States Department of Agriculture and Natural Resource Conserva- tion Service, 2000). c Not found in the study area.

pyramid name for this form; it is equivalent to the 3) Stipeae Pyramid. The Stipeae pyramid has a rectangles of Mulholland (1989), lb and Ig types of bilobate adaxial/abaxial outline and has often been Twiss et al. (1969), and Brown's (1984) non-sinuous confused with bilobate forms commonly produced trapezoids. The pyramid was a moderately uncom- by the Panicoideae (Mulholland, 1989; Fredlund mon form for grass taxa examined for this study, but and Tieszen, 1994). The more recent practice of 3 is produced by nearly all Pooideae grasses in small dimensional viewing makes distinguishing between quantities and at a greater frequency by Elymus ely- Panicoid bilobates and Stipeae pyramids less prob- moides and Festuca arizonica (Table 4). lematic because the cross-section of the Stipeae pyr-

TABLE3-Diagnostic phytolith forms, their dominant locally representative subfamilies and genera, and photosynthetic pathway. Genera are listed from most common to least common for that phytolith form.

Photosynthetic Phytolith form Subfamily or tribe Representative genera pathway Rondel Pooideae Festuca, Poa, all Pooideae C3 Pyramid Pooideae All C3 Stipeae pyramid Stipeae tribe Piptochaetium, Stipa C3 Crenate Pooideae Bromus, Koeleria,Elymus, Piptochaetium C3 Panicoid lobate Panicoideae Schizachyrium,Panicum C4 Simple bilobate varies Aristida, all Chlorideae, Stipeae Tribe C4/C3 Saddle Chlorideae All C4 Spiny body Pinaceae Pinus ponderosa C3 September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrasscommunity 287

TABLE 4-Phytolith shape frequencies from leaf epidermis material for selected species. Data based on a 100-short cell count from 1 plant for each species.

% Nondi- Stipeae Panicoid Simple agnostic Species Rondel Pyramid pyramid Crenate lobate bilobate Saddle forms Pooideae Koeleriamacrantha 37 5 - 53 - -5 5 Bromus ciliatus 7 - - 93 - - 0 Festuca arizonica 94 6 - -- 0 Poa fendleriana 96 3 -1 - - 0 Elymus elymoides 71 3 - 27 - - - 0 Stipa comata 99 -1 0 Piptochaetiumpringlei 14 2 53 31 - - 0 Panicoideae Panicum bulbosum 2 - -1 85 12 3 Chloridoideae Bouteloua gracilis 11 2 87 11 Blepharoneurontricholepis 7 - - - 5 88 7 21 - -1 4 74 22

amid is trapezoidal, with the top being smaller than which distinguished this form from other bilobates. the base, resembling a pyramid with a bilobate base I used the tribe designation because the Stipeae pyr- (Fig. 3). Panicoid bilobates are very flat in cross-sec- amid was commonly produced in Piptochaetiumprin- tion and symmetric. Although several researchers glei, which is in the Stipeae tribe (Table 1). I added have observed morphological differences between the pyramid modifier to reflect the actual 3 dimen- the Stipeae-type bilobates and true bilobates (Twiss sional shape of the phytolith, and because it is con- et al., 1969; Brown, 1984; Mulholland, 1989), Fred- sistent with Pooideae/C3 subfamily diagnostic lund and Tieszen (1994) used the term Stipa-type, shapes. However, within the flora examined in this

FIG. 1-End view of a rondel from Festuca arizonica. 288 TheSouthwestern Naturalist vol. 46, no. 3

FIG. 2-Cross-sectional view of a pyramid from Piptochaetiumpringlei. study, only 2 species from the Stipeae tribe were col- produced by S. comata (Table 4). These data are sup- lected and examined, S. comata and Piptochaetium ported by Mulholland's (1989) observation that S. pringlei (may also be known as Stipa pringlei); thus it comata had less than 5% of this form. remains to be seen if this shape is truly diagnostic 4) Crenate. Crenates have sinuous adaxial/abaxial for the Stipeae tribe. Both of these species produce edges and an asymmetric cross-section (Fig. 4). They the Stipeae pyramid, although it is not commonly are frequently longer than wide (although square

FIG. 3-Looking down on the top of a Stipeae pyramid from Piptochaetiumpringlei. September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrasscommunity 289

FIG. 4-Crenate in adaxial/abaxial view from Bromus ciliatus. and bullet shapes are common) and range from thin lund and Tieszen, 1994). Ninety-three percent of to thick in cross-section. The term crenate was used short-cells found in leaf epidermis tissue of Bromus by Fredlund and Tieszen (1994) and is equivalent to ciliatus were of this form (Table 4). the term sinuate used by Mulholland (1989), sinu- 5) Panicoid Lobate. This designation includes sev- ous trapezoid by Brown (1984), and the lh type of eral forms diagnostic to the Panicoideae subfamily, Twiss et al. (1969). Many Pooideae grasses produce including the bilobate, cross, and polylobate. Pani- crenates (Twiss et al., 1969; Mulholland, 1989; Fred- coid bilobates (Fig. 5) exhibit a double lobed adax-

-^/'1 ,,,1 ,11W"'N''' ??11???1~1I?~ 1 ' 1?~C~P~~~TZB;s! !D 1- 1wN -`V <962~?~Ei~`Pa~riJIw ^-- % G.Aiptm, Aai

FIG. 5--Adaxial/abaxial views of a panicoid lobate in situ from Schizachyriumscopanium. 290 TheSouthwestern Naturalist vol. 46, no. 3

??i ' ~ ?:?~-~:'.. ...: ..

?,- . .i -.': 1.

-: > .I

FIG. 6-Adaxial/abaxial view of a simple bilobate from Aristida sp.

ial/abaxial outline and indented or straight ends (1994) and emphasizes 3 planes of symmetry, which and a well developed lateral plane of symmetry are the 2 planes of the adaxial and abaxial surfaces, (Fredlund and Tieszen, 1994). This form has also and cross-sectional symmetry. The adaxial/abaxial been referred to as the dumbbell and is diagnostic outline of a saddle (using light microscopy) has for grasses in the Panicoideae subfamily (Twiss et al., been described as a battle-axe with double edges 1969; Mulholland 1989). Crosses have a minimum (Twiss et al., 1969). Grasses in the Chloridoideae of 3 indentations and are no more than 9.16 mi- subfamily commonly produce saddles (Table 4). crons longer than wide (see Pearsall, 1978; Pearsall, 8) Ponderosa pine spiny body type. For the flora 1986; Piperno, 1988; Mulholland, 1993, for extensive examined for this study, the spiny body type is di- discussions of crosses). Crosses were not common in agnostic for ponderosa pine (Fig. 8), and based on the Panicoideae grasses examined in this study. Mul- an examination of the literature, a newly identified tilobate forms were encountered most often in the phytolith form. Characterization and description of grass Panicum bulbosum, and can be differentiated this form was based on 5 different samples of needle from the crenate form. Multilobates are lobed, material from 5 trees spanning various levels of ma- whereas crenates are sculpted and have more nu- turity. Nomenclature is based on Bozarth's (1993) merous indentations. Crenates are also trapezoidal description and photograph of spiny irregular bod- in cross-section whereas multilobates are flatter and ies from P banksiana, which appear to be very similar more symmetric. to the ponderosa pine spiny body form, and spiny 6) Simple Bilobate. The simple bilobate can be bodies may be diagnostic of the genus Pinus. Spiny distinguished from the Panicoid bilobate by round- bodies are irregularly shaped, round to irregularly ed, rather than indented or straight ends, and asym- elliptical (typically shaped like a dolphin) and com- metric adaxial/abaxial outline (Fredlund and Tiesz- pletely silicified. Spines can be somewhat rounded en, 1994). The shank on the simple bilobate is gen- to sharp, forked or simple, and occur abundantly erally much longer (Fig. 6). Simple bilobates are typ- over the body of the phytolith. This form is similar ically produced by members of the Chloridoideae in size to most short-cells (Fig. 8). A scan (314-cell) subfamily and some Pooideae grasses (Fredlund and of an entire slide of ponderosa pine leaf material Tieszen, 1994). Species in the genus Aristida pro- from 1 randomly chosen sample indicated that 19% duce many types of simple bilobate forms. of the silicified material produced was of the spiny 7) Saddle. The saddle has been consistently de- body variety, and only 1.9% was the bordered pit fined by many researchers (Metcalfe, 1960; Twiss et tracheary elements diagnostic for many conifers. al., 1969; Brown, 1984; Mulholland, 1989; Fredlund Other species, including 5 conifers, were exam- and Tieszen, 1994; Fig. 7). My definition of the sad- ined and did not produce any form that could be dle is the same as used by Fredlund and Tieszen confused with the spiny body diagnostic for ponde- September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrass community 291

FIG. 7-Cross-sectional view of saddles from Blepharoneurontricholepis. rosa pine (Table 2). Because some authors have not- heads were present: Poa fendleriana, Festuca arizonica, ed that floral bracts of grasses have dendriform and Elymus elymoides, Muhlenbergia montana, Blepharoneu- asteriform phytoliths, which may be confused with ron tricholepis,Schizachyrium scoparium, and Bouteloua the spiny body type produced by ponderosa pine (D. gracilis. No phytolith forms were noted that could be Pearsall, University of Missouri, Columbia, pers. confused with the spiny body produced by ponde- comm.; see also Piperno, 1988), material from seed rosa pine, which can be distinguished from grass flo- heads was examined for the species where seed ral silicified material by having an irregular round

FIG. 8-The spiny body from Pinus ponderosa. 292 The SouthwesternNaturalist vol. 46, no. 3 to elliptical shape, irregularly and forked spines, hav- splitting the species in Bromus may be helpful ing no flat surfaces, and being completely silicified. in determining the approximate date of intro- duction of nonnative aggressive species such as RESULTS AND DISCUSSION-Examination of (cheatgrass). In addition, fur- grass flora from this study indicate that the ther examination of grass phytoliths from flo- general subfamily system originally defined by ral structures may reveal additional diagnostic Twiss et al. (1969), and validated by other re- forms. searchers (Brown, 1984; Mulholland, 1989; Pi- The newly identified ponderosa pine spiny perno, 1988; Fredlund and Tieszen, 1994) can body appears to be a useful diagnostic for this be applied to the ponderosa pine-bunchgrass area. Although Norgen (1972) noted this form community examined. Examinations of phyto- may be diagnostic, no other references to this lith shape frequencies indicate that for 7 spe- species were found in the literature. It is pos- cies in the subfamily Pooideae, and 1 species sible that this form is common to other species in the Panicoideae, very few (0 to 5%) nondi- in the genus Pinus, although spiny bodies were agnostic phytolith forms were present. Nondi- not observed in Pinus edulis, the other pine agnostic phytoliths, particularly rondels, were species examined. Results from this study sup- more common (7 to 22%) for the 3 species port other research suggesting that diagnostic examined from the Chloridoideae subfamily. phytoliths for conifers can be isolated and co- This result is consistent with the observation nifer species should not be ignored or gener- that all grasses produce rondel forms (Mulhol- ally categorized as nondiagnostic (Brydon et land, 1989; Fredlund and Tieszen, 1994) and al., 1963; Klein and Geis, 1978; Bozarth, 1993). indicates that rondels might be over-represent- Examination of soil phytolith assemblages ed in phytolith assemblages from the study based on the system developed in this study area, and saddles under-represented in com- could be used to understand grass-tree and parison to actual vegetation. grass vegetation dynamics. Although this study The potentially confusing Stipeae pyramid focused on grass and ponderosa pine phyto- was not common (<1%) for S. comata and was liths, additional examination of forb and shrub much more common in P pringlei. Interesting- species, including floral parts, may reveal other ly, taxonomists consider the Stipeae tribe tax- important diagnostic forms. In addition, ex- onomically confusing in general (Kellogg and amination of multiple individuals of the same Campbell, 1987; Watson and Dallwitz, 1992). species from a variety of soil types, and under- Barker et al. (1995) point out that the Stipeae standing differences in phytolith form and size tribe has been placed both within the Arun- due to environmental factors, would strength- dinoideae and the Pooideae by various re- en the use of phytolith analysis as a quantitative searchers. Using gene sequencing, Barker et al. tool for historical vegetation reconstructions. (1995) reported that the Arundinoideae (in- This researchwas supported by state funds provid- Stipa, Aristida, and was cluding Danthonia) poly- ed by the School of Forestry,Northern Arizona Uni- that within the sub- phyletic, meaning lineages versity,the United States Department of the Interior family show affinities with all 3 of the other Bureau of Land Management,Arizona State Office, subfamilies. These authors also suggest that and several Northern Arizona GraduateCollege Or- placement of the Stipeae tribe in the Pooideae ganized Research Grants. I thank my major profes- needs further investigation. The Stipeae pyra- sor, M. M. Moore, for support and review of this pa- For field and assistance I would like mid phytolith form clearly reflects the poly- per. laboratory to L. Labate. I thank D. Pearsall for phyletic nature of species the Stipeae tribe. acknowledge sharing her phytolith expertise with me at the Uni- With further study of phytolith forms from versity of Missouri, Columbia. For botanical exper- local flora, or of grass application morphomet- tise, I would like to acknowledgeJ. Springer and J. ric a more Po- techniques, specific phytolith Rominger. S. Hart and M. Wagner provided lab use aceae classification system could be developed. and space for me at NAU. For example, there might be potential to break out the larger and indented rondel form noted LITERATURECITED for P cre- fendleriana and establish diagnostic ALLRED,K. W. 1993. A field guide to the grasses of nate forms to separate Bromus and Koeleriama- New Mexico. Agricultural Experiment Station, crantha. Separating these genera and further New Mexico State University,Las Cruces. September 2001 Kerns-Phytoliths for a ponderosa pine-bunchgrasscommunity 293

BARKER,N. P., H. P. LINDER,AND E. H. HARLEY.1995. FULE, P. Z., W. W. COVINGTON,AND M. M. MOORE. Polyphyly of Arundinoideae (Poaceae): evidence 1997. Determining reference conditions for eco- from rbcL sequence data. Systematic Botany 20: system management of southwestern ponderosa 423-435. pine forests. Ecological Applications 7:895-908. BOZARTH,S. 1993. Biosilicate assemblages of boreal JIANG,Q., ANDD. PIPERNO.1998. Phytolith analysis of forests and aspen parklands. In: Pearsall, D. M., an archaeological site (Longshan Period) in Zhu- and D. R. Piperno, editors. Current research in madian City, Henan Provence, China: paleoenvi- phytolith analysis. The University Museum of Ar- ronmental and cultural implications. Geoar- chaeology and Anthropology, University of Penn- chaeology 9:409-417. sylvania, Philadelphia. Pp. 95-105. KAUFMANN,M. R., R. T. GRAHAM,D. A. BOYCE,JR., W. BROWN,D. A. 1984. Prospects and limits of a phyto- H. MOIR, L. PERRY,R. T. REYNOLDS,R. L. BASSETT, lith key for grasses in the central United States. P. MEHLHOP,C. B. EDMINSTER,W. M. BLOCK,AND Journal of Archaeological Science 11:345-368. P. S. CORN. 1994. An ecological basis for ecosys- BRYDON,J. E., W. G. DORE, ANDJ. S. CLARK.1963. tem management. United States Department of Silicified plant asterosclereids preserved in soil. Agriculture Forest Service Rocky Mountain For- Soil Science Society of America Journal 27:476- est Range and Experiment Station General Tech- 477. nical Report GTR RM-246, Fort Collins, Colora- CARBONE,V. A. 1977. Phytoliths as paleoecological do. indicators. Annals, New York Academy of Science KEALHOFER,L., ANDD. PENNY.1998. A combined pol- 288:195-205. len and phytolith record for fourteen thousand CHRISTENSEN,N. L., A. M. BARTUSKA,J. H. BROWN,S. years of vegetation change in northeastern Thai- CARPENTER,C. D'ANTIONIA,R. FRANCIS,J. E. land. Review of Paleobotany and Palynology 103: FRANKLIN,J. A. MACMAHON,R. Noss, D. J. PAR- 83-93. SONS,C. H. PETERSON,M. G. TURNER,AND R. G. KELLOGG,E. A., ANDC. S. CAMPBELL.1987. Phyloge- WOODMANSEE.1996. The report of the Ecological netic analysis of the Gramineae. In: Soderstrom, Society of America committee on the scientific T. R., K. W. Hilu, C. S. Campbell, and M. E. Bark- basis for ecosystem management. Ecological Ap- worth, editors. Grass systematics and evolution. plications 6:665-691. Smithsonian Institution Press, Washington, D.C. COOPER,C. F. 1960. Changes in vegetation, structure, Pp. 310-322. and growth of southwestern pine forests since KERNS,B. K. 1999. Phytolith assemblages and soil white settlement. Ecological Monographs 30: characteristics from a southwestern ponderosa 129-164. pine/bunchgrass community. Unpublished Ph.D. COVINGTON,W. W., ANDM. M. MOORE.1994a. Post- dissertation, Northern Arizona University, Flag- settlement changes in natural fire regimes and staff. forest structure: ecological restoration of old- KLEIN,R. L., ANDJ. W. GEIS. 1978. Biogenic silica in growth ponderosa pine forests. Journal of Sus- the Pinaceae. Soil Science 126:45-156. tainable Forestry 2:153-181. KURMANN,M. 1985. An opal phytolith and palyno- COVINGTON,W. W., ANDM. M. MOORE.1994b. South- morph study of extant and fossil soils in Kansas western ponderosa pine forest structure: changes (U.S.A.). Paleogeography, Paleoclimatology, Pa- since Euro-American settlement. Journal of For- leoecology 49:217-235. estry 92:39-47. MAST,J. N., P. Z. FULE,M. M. MOORE,W. W. COVING- COVINGTON, W. W., P. Z. FULE,M. M. MOORE,S. C. TON, AND A. E. M. WALTZ.1999. Restoration of HART,T. E. KOLB,J. N. MAST, S. S. SACKETT,AND presettlement age structure of an Arizona pine M. R. WAGNER.1997. Restoration of ecosystem forest. Ecological Applications 9:228-239. health in southwestern ponderosa pine forests. MCCLARAN,M. P., AND M. UMLAUF.2000. Desert Journal of Forestry 95:25-29. grassland dynamics estimated from carbon iso- DIETERICH,J. H. 1980. Chimney Spring forest fire topes in grass phytoliths and soil organic matter. history. United States Department of Agriculture Journal of Vegetation Science 11:71-76. Forest Service Rocky Mountain Forest Range and MCDOUGALL,M. B., ANDH. S. HASKELL.1973. Seed Experiment Station Research Paper RM-222, Fort plants of northern Arizona. Museum of Northern Collins, Colorado. Arizona, Flagstaff. FREDLUND,G. G., AND L. T. TIESZEN.1994. Modern METCALFE,C. R. 1960. Anatomy of the Monocotyle- phytolith assemblages from the North American dons. I. Gramineae. Claredon Press, Oxford, Great Plains. Journal of Biogeography 21:321- United Kingdom. 335. MILLER,G., N. AMBOS,P. BONESS,D. REYHER,G. ROB- FREDLUND,G. G., C. B. BOUSMAN,AND D. K. BOYD. ERTSON, K. SCALZONE,R. STEINKE,AND T. SUBIRGE. 1998. The Holocene phytolith record from Mor- 1995. Terrestrial ecosystem survey of the Cocon- gan Playa. Plains Anthropologist 43:187-200. ino National Forest. United States Department of 294 TheSouthwestern Naturalist vol. 46, no. 3

Agriculture Forest Service, Southwestern Region, ROVNER,I. 1983. Plant opal phytolith analysis: major Flagstaff, Arizona. advances in archaeobotanical research. In: Schif- MORGAN, P., G. H. APLET, J. B. HAUFLER, H. C. HUM- fer, M. editor. Advances in archaeological meth- PHRIES,M. M. MOORE,AND W. D. WILSON.1994. od and theory. Academic Press, Inc., San Diego, Historical range of variability: a useful tool for California. Pp. 225-266. evaluating ecosystem change. Journal of Sustain- SAVAGE,M., P. M. BROWN,ANDJ. FEDDEMA.1996. The able Forestry 2:87-111. role of climate in a pine forest regeneration MULHOLLAND, S. C. 1989. Phytolith shape frequen- pulse in the southwestern United States. Ecosci- cies in North Dakota grasses: a comparison to ence 3:310-318. Silviculture of southwestern general patterns. Journal of Archaeological Sci- SCHUBERT, G. H. 1974. ence 16:489-511. ponderosa pine: the status-of-our-knowledge. Forest MULHOLLAND,S. C. 1993. A test of phytolith analysis United States Department of Agriculture Service Research RM-123. at Big Hidatsa, North Dakota. In: Pearsall, D. M., Paper T. ANDC. H. BAISAN.1996. Historical and D. R. Piperno, editors. Current research in SWETNAM, W., fire in the southwestern United phytolith analysis. The University Museum of Ar- regime patterns States since AD 1700. In: C. D., editor. Fire chaeology and Anthropology, University of Penn- Allen, effects in southwestern forests. United States De- sylvania, Philadelphia. Pp. 131-146. of Forest Service General MULHOLLAND,S. C., AND G. R. RAPP. 1992. A mor- partment Agriculture Technical GTR RM-246, Fort Collins, Col- phological classification of grass silica-bodies. In: Report orado. 11-32. Mulholland, S. C., and G. R. Rapp, editors. Phy- Pp. TwIss, P. C., E. SUESS,AND R. M. SMITH.1969. Mor- tolith systematics, emerging issues. Plenum Press, classification of Soil New York, New York. Pp. 65-89. phological grass phytoliths. Science Society of AmericaJournal 33:109-114. NORGEN, A. 1972. Opal phytoliths as indicators of UNITED STATESDEPARTMENT OF AGRICULTUREAND soil and history. Unpublished age vegetative NATURALRESOURCE CONSERVATION SERVICE. 2000. Ph.D. dissertation, State University, Cor- Oregon National PLANTS database. National Plant Data vallis. Center, Baton Rouge, Louisiana, http:// PEARSALL,D. M. 1978. Phytolith analysis of archaeo- soils: evidence for maize cultivation in for- plants.usda.gov/plants. logical WATSON,L., AND M. DALLWITZ.1992. The mative Ecuador. Science 199:177-178. J. grass genera of the world. C.A.B. International, Wal- PEARSALL,D. M. 1986. Paleoethnobotany, a hand- lingford, Oxon, United Kingdom. book of procedures. Academic Press, Inc., San WHITE, A. S. 1985. Presettlement regeneration pat- California. Diego, terns in a southwestern ponderosa pine stand. PIPERNO, D. R. 1988. Phytolith analysis, an archaeo- Ecology 66:589-594. and Academic logical geological perspective. WILDING, L. P., N. E. SMECK, AND L. R. DREES. 1977. San California. Press, Inc., Diego, Silica in soils. In: Dixon, J. B., and S. B. Weed, RITCHIE,J. C. 1976. The late Quaternary vegetational editors. Minerals in soil environments. Soil Sci- of the western interior of Canada. Cana- history ence Society of America, Madison, Wisconsin. Pp. dian Journal of Botany 54:1793-1818. 471-552. ROVNER, I. 1971. Potential of opal phytoliths for use in paleoecological reconstructions. Quaternary Submitted28 April 2000. Accepted12 October2000. Research 1:343-359. AssociateEditor was David B. Wester