HORTSCIENCE 40(6):1598–1606. 2005. for ALS–PDC must be put to rest before the horticulture industry can expect increased use of this magnifi cent specimen plant. Cycad Toxins and Neurological One troublesome facet of past research has been the highly variable concentration Diseases in Guam: Defi ning of various toxins from batch to batch within a study, and from study to study within and Theoretical and Experimental among research teams (e.g., Palm and Cycad Societies of Australia, 1993, and citations therein). Indeed, this variability has been a Standards for Correlating Human reason for failures to embrace the cycad hy- pothesis. However, this variability need not be Disease with Environmental Toxins viewed through the eyes of a skeptic using an 1 enigmatic fi lter. We believe the documented Thomas E. Marler variability would have been explicable if College of Natural and Applied Sciences, University of Guam, UOG Station, background biotic and abiotic habitat factors Mangilao, Guam 96923 had been recorded, and appropriate sampling procedures had been used. Vivian Lee and Christopher A. Shaw We focus on several steryl glucoside sub- Department of Ophthalmology, University of British Columbia, Vancouver, types in our research (see Marler et al., 2005) BC, Canada and the present paper for two reasons. First, other known cycad toxins such as methyl azoxyl- Additional index words. ALS–PDC, BMAA, cycad, Cycas micronesica, Guam, neurotoxins, methanol (MAM) and β-methyl-amino-alanine secondary compounds, steryl glucosides (BMAA) have been studied for decades in spite of the fact that most do not generate behav- Abstract. Consumption of Cycas micronesica seed tissue has been associated with the amyo- ioral or pathological outcomes that resemble trophic lateral sclerosis–parkinsonism dementia complex (ALS–PDC) of the Western Pacifi c. ALS–PDC (reviewed by Wilson et al., 2002). However, failures to document vital plant and neighborhood descriptors and pronounced Our current work strongly implicates steryl variability in toxin concentrations noted within and among studies obfuscate decades of glucoside variants as contributing to ALS–PDC- research on this subject. We discuss the theoretical and experimental constraints of plant like outcomes in cycad fed mice (Khabazian et tissue sampling in relation to human disease research. Comparisons are made between this al., 2002; Shaw and Wilson, 2003; Wilson et approach and methods used throughout the history of ALS–PDC research, most notably al., 2002), mirroring in vivo data obtained by very recent reports concerning -methyl-amino-alanine. Methods for studying possible plant β earlier researchers (Ambike and Rao, 1967). neurotoxins need to be standardized and must follow rigorous criteria to be valid in principle. Second, the well-known cycad toxins such as Our discussions reveal why these criteria are essential and highlight the impact that natural MAM and BMAA are removed from cycad variations have on environmental toxin quantifi cation and interpretation. Past research tissue during the water-leaching used in tradi- on cycad toxins is defi cient on experimental and theoretical grounds, and interpretation of tional fl our preparation methods, but the steryl published data is dominated by ambiguities. This area of study as conventionally conceived glucosides are insoluble in water. Our ongoing and carried out needs transforming. We argue that future empirical studies should honor research includes all known neurotoxins in cycad appropriate plant science standards concomitantly with medical science standards. This dual tissue, since the role of these metabolites in focus will ensure appropriate sampling scheme, sample size, and reporting of background cycad plants has never been addressed in past plant and community factors known to infl uence phenotypic plasticity. research (Brenner et al., 2003). Theoretical and The island of Guam claims an endemic toxins have been published throughout recent experimental standards described in the pres- cycad species and an unwelcome high historical decades, and many of these were in pursuit of a ent study will be used to defi ne procedural and incidence of neurological disease among the reasonable link between exposure to cycad com- conceptual concerns as other candidate toxins Chamorro residents. The fi rst written record pounds and the prevalence of ALS–PDC (e.g., are given full scrutiny. suggesting that consumption of the cycad seed Kurland, 1993 and citations therein). However, Our aim in the present paper is to describe tissue was a probable cause of the neurologi- no article within this sizeable body of literature the multitude of omissions and shortcomings cal diseases was made by Spanish Lieutenant adequately described fi eld sampling methods, that have defi ned historical and ongoing research Colonel Felipe de la Corte in 1865 (as cited by habitat, conspecifi c and interspecifi c neighbors methods in this fi eld and illustrate the need Rogers, 1995). His opinions were ignored until of sampled plants, scaling and allometric rela- for appropriate theoretical and experimental the mid-20th century when the disease now tions, plant life history characteristics and events, directions in future research to remove the am- known as ALS–PDC and the cycad now known and other factors that are known to infl uence biguities that are defi ning of this large literature as Cycas micronesica K.D. Hill fi rst received secondary metabolism of plants. Unfortunately, base. Modifying the patterns of past research joint attention from the medical research com- this assertion holds true for the most recent addressing the link between the Guam cycad munity (reviewed in Kurland, 1988; Kurland et publications in this fi eld (e.g., Banack and Cox, and Guam’s neurological diseases to include al., 1994; Mabry, 2001). Consumption of cycad 2003; Yagi, 2004), despite the exposure of this appropriate plant science standards is critical megagametophyte tissue as processed fl our was manifest shortcoming of inadequate sampling for gaining a complete understanding of the commonly practiced, and the resulting inadver- methods by Zhang et al. (1996). chemical ecology in these stunning plants. tent ingestion of cycad toxins was considered the The endemic C. micronesica has not been We begin with two sections discussing the vehicle for human exposure (Whiting, 1963). popular in Guam’s horticulture industry. Al- various plant and habitat characteristics that Numerous articles and review chapters on cycad ternatively, C. revoluta Thunb. is one of the likely infl uence phenotypic plasticity in cycad most common landscape specimens in the biochemistry, factors that have been ignored Received for publication 29 Dec. 2004. Accepted commercial and residential landscape. Our in past research. for publication 27 Apr. 2005. We thank J. Chung, C. recent surveys revealed that the purported Melder and N. Dongol for assistance. Support pro- vided by USDA CSREES Special Grant in Tropical/ link between the native cycad and the dreaded Ignoring what Cannot be Ignored: Subtropical Agricultural Research to TEM and U.S. human disease is partly to blame for the lack Aerial Factors Army Medical and Materiel Command (DAMD17- of interest in using this handsome plant in 02-1-0678), NSERC Canada, and Scottish Rite the local horticulture industry. As a result, Control over known and probable biases Charitable Foundation of Canada to CAS. unanswered questions surrounding the hy- must be included in the sampling protocols in 1Corresponding author; e-mail tmarler @ uog.edu. pothesis that cycad toxin exposure is causal environmental research to ensure accuracy and

1598 HORTSCIENCE VOL. 40(6) OCTOBER 2005

OOctober.indbctober.indb 11598598 88/11/05/11/05 88:55:17:55:17 AAMM COVER STORIES

repeatability. Following is a short list of factors research to clarify the infl uence of seed age on level is controlling of many plastic responses, that defi ne the aerial environment of cycad plants neurotoxin concentration is warranted. Seed age including synthesis and allocation of second- and that elicit phenotypic plasticity of plant has never been reported in past publications in ary compounds (e.g., Callaway et al., 2003; secondary chemistry. These are the major fac- this area of research. Herms and Mattson, 1992). Moreover, the tors that we have documented and standardized The timing and magnitude of resource alloca- level of incident light at microsites within a for our own research (Marler et al., 2005). We tions within a plant are determined by modular plant canopy is known to infl uence reproduc- suggest documenting each of these factors as organization, with each growth module behav- tive growth characteristics of those microsites a minimum for reliable interpretation of future ing semi-autonomously (Herms and Mattson, (Childers, 1995). For example, seeds on the research on cycad secondary metabolism. 1992). The synchrony of these growth modules north and south orientations of a cycad plant The chemical constituency of a plant organ among C. micronesica individuals within a land- may experience less incident light than on the observed at one point in time is the result of scape is increased by typhoon damage (Hirsh east and west orientations. Moreover, infl uence many processes in plant life history. For concen- and Marler, 2002). With each successive growth of shade from neighboring plants on cycad tration, pool size, and allocation of secondary module thereafter, the population becomes less secondary metabolism is unknown. We suggest compounds among organs of cycad plants, his- synchronized. The initial cost of constructing that recording incident light or percent shade is tory of herbivory may elicit immense infl uence a leaf is generally greater for species with required to establish legitimacy of a data set. on instantaneous metabolite concentration. For evergreen leaves, large leaves, and leaves that Historical ALS–PDC research has ignored this example, herbivory may increase the products are sclerophyllous (Chabot and Hicks, 1982). requirement. of secondary metabolism of some species so Cycas micronesica leaves are distinguished Comparing a compound in tissue from one strongly that fi tness is compromised (Heil and by all three of these characteristics. Although plant with that in tissue from another plant Baldwin, 2002). Moreover, herbivory of leaf no resource budgets have been determined to without ensuring the shade level of each plant tissue may elicit a plastic response of secondary date to quantify these dynamics for cycads, the is homogeneous (e.g., Banack and Cox, 2003) compounds found in reproductive structures substantial costs and payback time no doubt cannot be justifi ed. In fact, one of the underly- (Baldwin and Karb, 1995). No testing has infl uence the partitioning of assimilates. We ing justifi cations Bryant et al. (1983) used to been conducted to determine the direction and suggest a third factor of plant life history that propose the carbon/nutrient balance hypothesis magnitude of plastic responses of cycad tissue must be recorded to obtain unambiguous data was the known infl uence of shade on phenotypic to herbivory. Until this testing is completed, a is the growth history of these episodic events variation in secondary metabolism. detailed description of herbivory history as a before sampling dates. Timing of past growth Addressing the above-ground biases. minimum measure of precaution is essential to fl ushes has never been reported in past publica- Throughout the decades of research on C. mi- account for consumer damage biases. This can- tions in this area of research. cronesica secondary metabolites, 100% of the not occur without a lengthy history of observing Scaling relations among vegetative organs articles omitted essential information about plant the plants that are sampled. at the individual plant level follow consistent life history, size, fecundity, and sun exposure. A second factor in plant life history that patterns that apply across species and ecological Considering the fact that phenotypic plasticity must be known to control background bias conditions (Niklas, 2004). Likewise, reproduc- of plants in response to these and other factors when quantifying secondary compounds in tive growth size may scale with plant size for a may cause characteristics to vary by orders of cycad plants is age of reproductive fl ushes broad taxonomic range that includes angiosperm magnitude (see Callaway et al., 2003), exclu- that are present on sampling dates. In Guam’s and nonangiosperm species (Niklas, 1993; Pe- sion of this information in past research is not environment, the exterior of a C. micronesica ters et al., 1988). Zotz (2000) reported that lack acceptable. We propose the failure to report seed changes from light green initially through of attention to plant size in physiology studies has these infl uential factors may fully explain the a dark green phase to bronze by ca. 8 months the potential to introduce inappropriate ambigui- variability in neurotoxin concentration from of age (T. Marler, unpublished data). Sun ex- ties in interpretation of results. The relationship study to study. Certainly, future attempts to posure infl uences the rate of this color change. between reproductive growth characteristics and determine distribution of secondary compounds Many seeds do not increase in volume beyond plant size is unknown for cycad plants. Until among various tissues of cycad plants must 8 months, and exterior phenotype changes little more is learned about distributional biases of include this information for conclusions to be until the sarcotesta becomes dark brown before secondary compounds among competing organs meaningful. More importantly, the sample size seed abscission. The timing of this change from and the infl uence of plant size on theses biases, needed to ensure accuracy will be minimized bronze to brown is highly variable among plants, we suggest that height and any measure of stem by these efforts to control background biases, but occurs from 17 to 21 months under typical girth are minimum requirements to account for and this will directly improve conservation conditions (T. Marler, unpublished). Thus, the these mass- or volume-related sample biases. No measures as this research continues. exterior phenotype of seeds may appear similar measure of plant size has been included in past Our approach was to use these descriptors to from 8 months to greater than 21 months. Our publications in this area of research. standardize the plants included in each experi- ability to estimate age of seeds observed on Crop load has been studied extensively in ment (see Marler et al., 2005), thereby ensuring previously unvisited plants is nil within this fruit science, and many horticultural practices these factors were relatively homogeneous age range, despite our extensive experiences in in orchard management are based on this throughout the population. An appropriate observing behavior of the species since 1997. knowledge (e.g., Childers, 1995). The crop load alternative strategy would be to treat the nu- Our ability to estimate age of seeds after the affects varied characteristics such as distribution merical data from these factors as co-factors color transition from bronze to dark brown is of assimilates, fruit size, fruit chemistry, and in covariance analysis. Both approaches would also nil, despite the fact that this transition occurs total plant yield. Position of fruits within an remove the bias from each mean, but would not as early as 17 months yet seeds may remain on infl orescence or within a canopy affects matu- inappropriately suggest causal relationships. plants >30 months. ration and sampling procedures across a wide Our ongoing research is designed to determine Our research to date (Marler et al., 2005) does range of architecture types (De Silva et al., 2000; the direct infl uence of these varied plant factors not allow a defi nitive assessment of seed age and Miles et al., 1996; Viti et al., 2003). Despite this on neurotoxin concentration in C. micronesica secondary metabolism, since initial objectives broad knowledge base indicating fecundity and tissue. We believe this information will lead were designed to defi ne methods for sampling location of a reproductive organ within the plant toward an understanding of the role of these schemes. However, a comparison of the means infl uence many characteristics, past ALS–PDC metabolites in the plant. from the 12-month-old tissue with the means researchers have ignored the need to document from 17- or 19-month-old megagametophyte this information from sampled cycad plants. We Ignoring what Cannot be Ignored: tissue indicate the younger seeds contain greater suggest that recording the number of seeds on a Below-ground Factors concentration of steryl glucosides (see Marler sampled plant should be treated as a minimum et al., 2005). These observations are evidence measure of precaution to account for probable ALS–PDC incidence in Guam peaked in that seed age must be known for repeatability seed load biases. the mid-20th century, and incidence has been and removal of ambiguity, but also that further The level of sun exposure at the whole plant declining since this time (see Kurland, 1993).

HORTSCIENCE VOL. 40(6) OCTOBER 2005 1599

OOctober.indbctober.indb 11599599 88/11/05/11/05 88:55:20:55:20 AAMM Ongoing research designed to determine the of the closely related strains that formed each The complexities of how conspecifi c and in- relationship of a historical peak in cycad tissue geographic group. terspecifi c neighbors infl uence a plant’s growth ingestion by Guam residents several decades A Nostoc strain isolated from symbiotic and physiology via below-ground communica- ago to a subsequent peak in neurological tissue and maintained in axenic cultures can- tion are beginning to be understood (reviewed disease incidence among those residents must not be considered a germane strain without in Callaway, 2002; Callaway et al., 2003). employ methods that are relevant to the seed subsequent validation. Isolation of an incon- Recent evidence reveals sophisticated plant collection environment during that era in his- sequential symbiont or nonsymbiotic Nostoc root communication dexterity that includes tory. Following is a short list of environmental from the symbiotic plant tissue is likely. For an ability to distinguish conspecifi c roots from factors that impact rhizosphere dynamics and Azolla–Nostoc symbioses (e.g., Gebhardt and self roots (e.g., Falik et al., 2003; Gersani et are known to infl uence secondary chemistry of Nierzwicki-Bauer, 1991), isolated Nostoc al., 2001). These root interactions do not elicit plants. We believe these factors render the use strains were regarded and studied as relevant straightforward plant responses resulting from of source plants outside of native habitats of Azolla symbionts until proper examination a neighbor’s competitive or facilitation attri- little to no value in this line of research. ultimately proved they were distinct from butes. Rather, a suite of benefi cial and negative Physical and chemical properties of litterfall major symbionts. Isolated Nostoc strains from effects of a plant’s habitat interplay spatially infl uence myriad habitat processes that affect C. micronesica plants grown in ex situ gardens and ephemerally to determine the net-posi- phenotypic plasticity of plant biochemistry (e.g., and maintained as axenic cultures (e.g., Murch tive or net-negative effect of conspecifi c and Loranger et al., 2002; Madritch and Hunter, et al., 2004) are not only alien genotypes, but interspecifi c neighbors (Holzapfel and Mahall, 2002; Santa Regina and Tarazona, 2001). For should justifi ably be treated as immaterial to 1999). Positive attributes of close neighbors are example, Allison and Vitousek (2004) report a the symbiotic system until they have been being tapped to modify traditional reforestation 50-fold variation in leaf litter decay rates among proven to be major symbionts. efforts through use of facilitating nurse plants 11 species, with native species exhibiting much In addition to cyanobionts, cycad roots are (e.g., Castro et al., 2002). slower decay rates than alien invader species. also profi cient hosts of mycobionts (Fisher and Secondary metabolism is included in the Our initial studies (Marler et al., 2005) are the Vovides, 2004; Muthukumar and Udaiyan, assemblage of responses under the infl uence fi rst to ensure that interspecifi c neighbors of 2002). Thus, the cycad symbiotic system is of neighbors. For example, allocation costs to sampled plants were endemic or indigenous, and actually a tripartite system, where each organ- secondary compounds of Nicotiana attenuata to document these providers of detritus inputs ism interacts with the other two. Research on were not observed in the absence of neighbors, during the years before sampling dates. The physiology of this system must address the but were extensive when plants were grown relevance of past reports including compounds interactions among all three organisms (e.g., with neighbors in environmentally realistic from tissue sampled from C. micronesica plants Tsimilli-Michael et al., 2000). The many issues conditions (van Dam and Baldwin, 2001). in ex situ botanic gardens is suspect, but even discussed above concerning relevance of ‘alien’ Ehlers and Thompson (2004) report that co-oc- past studies that failed to document the habitat detritus providers, soil microorganisms, and curring plant species adapt to one another. No biodiversity of sampled cycad plants in Guam cyanobionts also applies to mycorrhizae that doubt the multifaceted levels of root communi- have ignored this relevant factor. colonize C. micronesica roots in an ex situ gar- cation combine with the chemical relations of Detritovores and other edaphic microorgan- den. In light of the tripartite system, however, detritus inputs by neighbors to select for this isms involved in biodegradation of detritus and the lack of relevance of ex situ research may evolutionary phenomenon. These root interac- biosynthesis of glomalin and other chemicals be even more magnifi ed. Indeed, both of the tions among co-occurring plants indicate the that enter the soil matrix are critical in defi ning microbionts in ex situ sites are distinct from biochemistry of C. micronesica plants grow- soil physicochemical properties and cycling those that comprise the indigenous tripartite ing in remote locations devoid of indigenous of abiotic and biotic factors in the plant’s system in Guam’s environment. Knowledge neighbors cannot be considered relevant to rooting environment (e.g., Caesar-TonThat of the cycad–mycorrhizae tripartite system biochemistry of these plants in habitat where and Cochran, 2001; Rillig et al., 2001). The with Nostoc is lagging behind that of the Guam’s residents harvested seeds for human ecological effi ciency of Guam’s indigenous legume–mycorrhizae tripartite system with consumption several decades ago. edaphic microorganisms in relation to those that nodule-inducing bacteria, but this does not jus- Edaphic factors within the natural range of a happen to surround the rooting environment of tify ignoring the likelihood that the three-way cycad species provided many of the evolution- a cycad plant via anthropogenic actions in ex interactions are as controlling of cycad growth ary selective pressures that defi ne phenotype situ botanic gardens is virtually unknown. and development as are those in legumes. Fisher and genotype. Moreover, minimal differences The cycad–cyanobiont relationship has and Vovides (2004) point out the probability in habitat abiotic factors may profoundly unique features that include a high degree of that cycads may require mycorrhizal coloniza- modify the complex positive and negative specifi city within a coralloid root structure tion for realistic Nostoc function. interactions among plant neighbors (Choler et (Costa et al., 1999) and a possible direct infl u- We use these truths to suggest biochemistry al., 2001; Schenk et al., 2003). Thus, transfer ence of phenolic compounds on the formation of C. micronesica in the absence of indigenous of the entire suite of indigenous organisms of symbiosis and metabolism (Lobakova et al., microbionts must be viewed as artifacts until inclusive of those listed above to a remote 2004). The access to fi xed nitrogen afforded by genetic and phenotypic geographic distribu- garden would still fall short of providing the this symbiosis may be one of the reasons that tions of the microbiont populations are deter- basics of an experimental setting from which cycads were selected for survival throughout mined. To illustrate, indigenous cyanobionts relevant physiology data could be obtained. tens of millions of years (Norstog and Nich- and mycobionts could be isolated from a Guam Transfer of the edaphic substrates that serve ols, 1997). Based on the results of Ow et al. habitat and transported to a remote garden as the matrix for the fi ne-tuned biological in- (1999), numerous Nostoc strains found in ex setting. These microbionts could then be used teractions known to infl uence plant physiology situ gardens would likely invade the conspicu- to inoculate specimens of C. revoluta or any would also be required. ous invasion zones of C. micronesica coralloid other Cycas species to develop a functioning The Guam habitats that C. micronesica roots. However, presumption that behavior of tripartite system. Reasonable discussion of has colonized are varied. The major island these ‘alien’ cyanobionts is representative of biochemistry data from this tripartite system perimeter forested habitats are from Ritidian behavior within the indigenous cycad–Nostoc would not infer that the results relate to bio- soil series, the interior habitats in northern association in Guam is not justifi ed. Genetic chemistry of an indigenous C. micronesica Guam are from Guam soil series, the upland diversity among Nostoc strains isolated from tripartite system in habitat, even though two habitats in southern Guam are from Akina cycad species within their natural range is not of the three organisms would be indigenous to soil series, and the ravine forests in southern known to date, and caution should prevail until Guam. Similarly, reasonable discussion of bio- Guam are dominated by Agfayan soil series this diversity is known. Indeed, Nilsson et al. chemistry data from an ex situ C. micronesica (Young, 1988). These substrates and their rock (2000) reported that genotype of cyanobiont specimen in a tripartite system with two alien outcrops are highly contrasting in chemistry isolates from Gunnera species grouped by geo- microbionts would not infer that the results and physical characteristics. The Ritidian and graphic region, indicating a limited distribution are relevant to the real world. Guam soil series are both alkaline and contigu-

1600 HORTSCIENCE VOL. 40(6) OCTOBER 2005

OOctober.indbctober.indb 11600600 88/11/05/11/05 88:55:23:55:23 AAMM ous in some areas. The slightly acid Agfayan (Young, 1988). These seasonal differences in of representatives from 10 cycad genera under series is contiguous with the highly acid Akina water availability exert considerable control homogeneous edaphic and climatic conditions. series in many locations. Thus, gene fl ow may over plant growth and development, and soil Indeed, addressing an objective this restrictive be substantial between the habitats that cover a characteristics largely defi ne the severity of requires an ex situ collection, since no cycad gradient overlapping the contiguous geologic seasonal drought stress. Habitats in Ritidian and habitat worldwide includes these genera in substrates. Indeed, transects in two of the fi ve Guam series soils dry rapidly following rainfall, close proximity. But this research must carry habitats we have been studying since 2002 but habitats in Akina and Agfayan series soils the caveat that no relevance to plant behavior include these gradients. However, the habitats retain water reserves. Until more is known about in habitat is inferred. on Guam soil series are disjunct from habitats cycad secondary metabolism and environmental An overview of the cycad environment on Agfayan or Akina series, and habitats on cues such as seasonal water availability and in the ex situ settings sampled by Banack Ritidian series are disjunct from habitats on diel or seasonal temperature patterns, research and Cox (2003) provide a working example. Agfayan series. Not only are the physico- should be confi ned to Guam’s environment and The two Florida botanic gardens from which chemical characteristics highly contrasting, dates of sampling must be recorded. With this samples were obtained use many preplant the geographical isolation of the cycad habitats information documented, historic climatologi- procedures that create a rooting environment based on the contrasting substrates adds to the cal data can be retrieved to draw conclusions more heavily infl uenced by anthropogenic ac- possibility that diversity of fragmented cycad about the infl uence of weather patterns before tions than geologic actions. Individual plants populations is clustered genetically. Numerous sampling dates from previous reports. are grouped with conspecifi c neighbors in plants have evolved specifi c genotypes that Addressing the neighborhood biases. A monoculture planting beds, or are surrounded are adapted to their microenvironment more review of the collective literature on cycad by alien species including judiciously managed so than conspecifi c individuals from other toxins reveals that 100% of these articles fail turfgrass. The planting sites are less shaded microenvironments (Marler, 2002). Moreover, to account for these environmental factors than Guam’s cycad habitats. All plants receive there is no reason to assume the microbionts that largely infl uence plant physiology. These routine fertilizer applications. The use of tissue from C. micronesica originating from Guam’s omissions have the potential to render these from plants that have received routine doses alkaline soils are symbiotically active in former studies irrelevant to the questions being of fertilizers to study secondary metabolism Guam’s acid soils, or vice versa. Similarly, studied or incomplete at best. Plant secondary cannot be reasonably viewed as appropriate, there is no reason to assume the infl uences of chemistry is determined not only by intrinsic unless the research questions are restricted interspecifi c neighbors found in alkaline soil genetic and phenotypic components, but also to the infl uences of differential fertilization. habitats are similar to those infl uences in acid by many community-level components. The One of the underlying justifi cations Bryant et soil habitats, and vice versa. environmental factors described above and al. (1983) used to propose the carbon/nutrient These realities shape appropriate C. micro- many others clearly reveal a plant’s chemical balance hypothesis was the well-known infl u- nesica research in two manners. First, research ecology is played out in a world much larger ence of fertilizers on phenotypic variation in on Guam must document habitat descriptors than the plant itself. For example, Hevia et secondary metabolism. Lastly, these Florida for methods to be repeatable and interpreta- al. (1999) argue that nutrient use effi ciency sites are ground zero for the American inva- tion of biochemistry data to be meaningful. can be validly studied only when comparing sion of Aulacaspis yasumatsui Takagi, the For example, Brenes-Arguedas and Coley species that co-occur in the same habitats. Asian cycad scale that was fi rst discovered (2005) report that many environmental fac- In this context, Norstog and Nicholls (1997) there before 1996 (Howard et al., 1999). All tors explained spatial variation of secondary warned that studying growth characteristics of the Cycas plants in one of the ex situ sites chemistry of Quararibea asterolepis Pitt. at the of cycad specimens in ex situ cultivation is of receive comprehensive sprays of a growth landscape level. Second, the use of individuals little value for making inferences about the regulator as a portion of the methods used in ex situ collections is not advisable based on behavior of a species in its natural habitat. for control of this deadly scale (J. Haynes, genetic fundamentals, unless microsite and soil They extend this warning to include a cycad personal communication). Virtually nothing is type of the original seed collection in Guam is specimen growing in a site within its natural known about the impact of synthetic pesticide described in detail for each accession. Many range that may have had anthropogenic infl u- spray materials on secondary metabolism of of the provinces, especially in southern Guam, ences in past or present. C. micronesica. contain more than one of these soil series, These warnings ring true for Guam because The preceding paragraph provides graphic and a record of the province of seed origin is C. revoluta plants have been introduced to the evidence that biochemistry of tissue collected insuffi cient. island for many decades (Stone, 1970). This from plants in botanic gardens cannot be con- Plants growing in sites with Akina soils cycad species is currently a ubiquitous mem- sidered relevant to a plant’s native environment. must contend with a rhizosphere that is char- ber of the urban landscape. The vast majority Ignoring this reality is an atrocity when tissues acterized by overabundance of various metals of these specimens were brought to Guam as from some plants that have received extensive and aluminum toxicity tendencies. Indeed, landscape-ready container specimens from off- horticultural inputs are compared with tissues these chemical characteristics of Akina soils island nurseries. In all likelihood, alien Nostoc from other plants that have not (e.g., Banack may be one of the major reasons that some and mycorrhizae strains accompanied these alien and Cox, 2003). This approach is not consistent indigenous species do not occur in habitats on cycads during the high volume imports geared with scientifi c standards, and variation among this substrate (Marler and Lawrence, 2004). toward satisfying the landscape industry. Thus, sample means is likely more infl uenced by The possibility that aluminum, iron, or other any C. micronesica in the urban landscape may the heterogeneous anthropogenic treatments soil-derived metal ions interact with primary be colonized by these alien strains of microbi- among the sample sites than inherent variation toxins in diet-induced pathology (e.g., Garruto onts, and in all likelihood would be supplying among organ and tissue categories. et al., 1989; Zhang et al., 1996) cannot be experimental artifacts if they were targeted as One other item that requires attention is sex- discounted. Thus, the overabundance of these the source of seeds for studying the relationship ual compatibility. Guam’s cycad is a member of metals in the Akina soil series in relation to between cycad toxins and ALS–PDC. the C. rumphii species subsection, a complex the other soils series in Guam may add another Certainly ex situ collections are invaluable uniquely defi ned by the presence of a spongy variable that complicates interpretation of past for protecting and preserving genetic diversity, fl otation layer within the seed [Hill, 1994; publications. This is yet another example of for display purposes, and for research that is not Marler et al., 2005 (Fig. 1B)]. Little is known why microsite descriptions are needed to strongly infl uenced by genetic × environment about compatibility among the various species authenticate each study. interactions. But plant physiology is under that comprise this subsection, but cycads have Guam’s seasonal changes are under the utmost control of environment and genetic weak fertility barriers (Hill, 1993) and compat- direct infl uence of heterogeneous rainfall pat- × environment factors. Thus, physiology re- ibility among some or all of these species is terns. August through October generally receive search in off-site collections must be restricted highly likely (K. Hill, personal communica- >30 cm of precipitation per month, and Febru- to appropriate objectives. For example, Marler tion). For this reason we are discouraging the ary through April receive <10 cm per month and Willis (1997) compared leaf physiology introduction of any of these species to the

HORTSCIENCE VOL. 40(6) OCTOBER 2005 1601

OOctober.indbctober.indb 11601601 88/11/05/11/05 88:55:26:55:26 AAMM Guam landscape industry (Marler, 2002). The certainties at this stage. Epidemiology of these Brenner et al. (2003) state that other physi- ‘civilian’ gene pool in Guam is unadulterated, human diseases strongly suggests that exposure ological roles for BMAA should be investi- but introducing a sexually compatible species to an environmental factor is partly causal, C. gated, such as possible endogenous signaling carries the risk of escaped alien genes taking micronesica tissues that have been ingested in compounds. We suggest this assertion holds up residence in the C. micronesica population the traditional diet contain neurotoxins that true for all of the identifi ed compounds, and in habitat. Specimens of C. micronesica plants maintain integrity through the traditional fl our that the pathological responses of animals to in ex situ collections are surrounded by other preparation and cooking procedures, no defi n- the known C. micronesica toxins are incidental Cycas species that may be sexually compatible. ing plant function has been revealed for any to evolutionary pressures that selected for these Wind pollination has not been disproven for of the studied compounds, and no compelling interesting metabolites. The continued study Cycas species, and the likelihood of genetic evidence can be used to discount the hypothesis of secondary metabolism in cycads would be contamination cannot be ignored. The chance that direct exposure to cycad compounds is the best served by answering questions within the of harvesting interspecifi c Cycas seed material causal environmental factor (e.g., Brenner et context of the various resource models, such as in ex situ settings does not allow their use in al., 2003; Mabry, 2001). Moreover, our animal the growth-differentiation balance hypothesis, studies of secondary compounds in a manner model induces neurodegeneration following the carbon-nutrient balance hypothesis, or the that infers relevance to in situ locales. exposure to cycad tissue in a ready and repeat- optimal partitioning hypothesis (Bloom et Clearly, studying cycad specimens in able manner (Shaw and Wilson, 2003). If the al., 1985; Herms and Mattson, 1992). These native locales is the sole valid approach cycad hypothesis is even marginally correct, rely in diverse ways on the premise that a for determining secondary metabolism in applying results from the Guam model toward fundamental tradeoff in resource or ecological an unambiguous manner. The use of in situ eventual prevention of many neurological costs exists between growth and differentiation plants for this research also mandates two diseases appears to be possible. However, processes including secondary metabolism, or further considerations. First, appropriate in ongoing research must be conducted by using that response to limiting factors elicits a shift situ methods must exclude cycad individuals appropriate plant science standards that include to make all resources equally limiting among positioned within the sphere of infl uence of any adequate control over sampling methods (e.g., tissues or processes. alien arthropod or plant species. These alien Marler et al., 2005) and description of habitat The known functions of plant secondary organisms were not likely infl uencing cycad and plant factors that have the potential to compounds include structural support, attrac- plant chemistry several decades ago when affect secondary metabolism. tion of pollinators or dispersers, temporary the human ingestion of seed fl our reached a Nomenclature. Recent publications (e.g., assimilate storage and transport, hormone peak. This is becoming increasingly diffi cult as Chen et al., 2002; Cox and Sacks, 2002; regulation, and myriad other functions (e.g., more alien species become established in the Zaninovic, 2004) continue to sustain the past Herms and Mattson, 1992). Cantatore et al. Guam environment, and points to the urgency confusion about nomenclature of this eminent (2000) propose that steryl glucosides are car- of continuing this research in a timely man- cycad. The Guam cycad is Cycas micronesica riers of glucose across membranes and serve ner. Second, since the geologic substrates and (Hill, 1994), and is a member of the C. rumphii as glucose donors. The loading of immature interspecifi c indigenous neighbors that defi ne species complex. Every member of this com- seed tissues with these compounds (Marler et Guam’s cycad habitats are highly contrasting, plex has been erroneously called C. circinalis al., 2005) lends itself to the possibility that they these habitat descriptors must accompany each L. or C. rumphii Miq. at some point. However, do indeed serve as carbohydrate donors for sampling to ensure reliability. Throughout the the distribution of C. rumphii is restricted long-term development of tissue throughout the long history of ALS–PDC research until the to a band of island nations centered on the lengthy embryo and megagametophyte matura- present study, 100% of the articles provided Moluccan island group, and the distribution tion period. The decline in steryl glucosides no information on this subject. of C. circinalis is restricted to India (Hill and with seed age substantiates the possibility that Stevenson, 2005; K. Hill, personal communica- early storage pools are metabolized to act as Past Oversights and Future Directions tion, A. Lindstrom, personal communication). donors of resources. Alternatively, Kunimoto Cycas micronesica is endemic to Guam and et al. (2002) provide evidence that steryl glu- The text heretofore amply describes some surrounding islands. If considered within the cosides are mediators in signal transduction to pertinent plant and environmental factors that broad context of various disciplines, through- induce benefi cial stress responses. In general, have been ignored in past research on the rela- out years, and among countries, the number sterols are involved in regulating membrane tionship between cycad toxins and ALS–PDC. of times the Guam cycad has been mentioned properties (e.g., Hartmann, 2003). All members Clearly, this line of research as convention- in the literature perhaps eclipses that of any of the C. rumphii species subsection produce ally conceived is in need of a major shift in other cycad species. Certainly, the continued seeds with a fl otation layer and are designed methods employed. Broadly, the established use of incorrect names to refer to this species to withstand extended periods of fl oating in pattern of zealous focus on animal toxicology unduly perpetuates confusion. ocean currents. The ability of these seeds to and epidemiology methods while disregarding The decades of extensive research on the maintain viability and integrity of tissue func- relevant plant and environmental methods link between cycad toxins and Guam’s neu- tion throughout extended exposures to extreme needs to be reassessed. We next discuss these rological diseases have attracted a plethora osmotic potentials may involve the proposed issues in general. of research scientists in many fi elds. Until signal transduction or membrane regulating Cycas micronesica research. Cycads are presently, this research has progressed with properties of these interesting compounds. the oldest living seed plants and are an opulent the exclusion of a local plant scientist. This Considering the many functions of second- source of information for interpreting plant example of continued misapplication of out- ary compounds, we agree with Brenner et al. biology, yet have been largely ignored in recent dated nomenclature validates the assertions of (2003) that searching for plant functions other developments of plant research (see Brenner Bawa et al. (2004) that research teams address- than anti-herbivory functions is justifi ed for et al., 2003). Similarly, Stone (1970) reported ing topics in tropical biology need to include each identifi ed cycad metabolite. In contrast, that the physiology of Guam’s plants was virtu- members of the local community. Banack and Cox (2003) opine that the neuro- ally unknown, yet more than 3 decades later Secondary compounds in cycads. Questions toxins found in C. micronesica tissues have, the number of physiology studies on Guam’s concerning the function of secondary com- indeed, evolved as anti-herbivory compounds. endemic and indigenous plants remains in the pounds in cycads have not been addressed to An abundance of BMAA in the sarcotesta is single digits. Continued research on this valu- date (Brenner et al., 2003). Pathogenic reactions evidence they use for this opinion, since this able endemic species promises to provide much by many herbivores have been demonstrated is the tissue consumed by fruit bats. We have useful information about horticulture, human for many of these cycad compounds. However, already argued that their data are ambiguous neurological diseases, tropical ecosystems, that animal response is not defi nitive evidence and statistical validity is impossible since their and chemical ecology. that the compounds embody a defensive role methods included the use of a sample size of The historical research on Guam’s neuro- or that herbivore pressures in evolutionary one (Marler et al., 2005). However, our data logical diseases collectively supports several history elicited their selection. also reveals that sarcotesta tissue is, in general,

1602 HORTSCIENCE VOL. 40(6) OCTOBER 2005

OOctober.indbctober.indb 11602602 88/11/05/11/05 88:55:29:55:29 AAMM more heavily endowed with steryl glucosides since the years of highest incidence (reviewed so-called excitotoxic cell death. Barring some than is megagametophyte tissue (Marler et al., in Kurland, 1988; Kurland et al., 1994; Mabry, still unknown mechanism by which a fruit bat 2005). We suggest that this spatial distribution 2001). This trend fi ts the spike in direct cycad could either sequester or detoxify BMAA, it does not support a chemical defensive role tissue consumption during World War II years is distinctly unlikely these mammals could and offer several lines of reasoning for this (Palm and Cycad Societies of Australia, 1993 accumulate high concentrations of the toxin suggestion. and citations therein). Alternatively, copious over time without infl icting the same neural While the cycad sarcotesta is not homolo- consumption of Guam bats continued through damage on themselves as has been speculated gous with angiosperm pericarp, it is considered the 1960s and consumption of imported bats to occur in human ALS–PDC victims. Dif- analogous in many respects (Brenner et al., continued through the early 1990s (Wiles, ferential sequestration of BMAA in bats vs. 2003; Norstog and Nicholls, 1997). One of 1992). During the period 1975–89, 65% of other mammalian species is unknown and those analogies is attraction of seed dispersers. these imported bats originated in Rota, Palau, remains at best speculative. Moreover, puta- Indeed, recognizing this as an attribute of the and Yap and were comprised of Pteropus tive detoxifi cation mechanisms in bats would sarcotesta is fundamental to validating the Cox mariannus, the very species of bat from render BMAA nontoxic regardless of where and Sacks (2002) hypothesis. Preferentially Guam (Wiles, 1992). The endemic range of the molecule next appeared in the food chain. loading a defensive chemical into the very C. micronesica also includes all three of these Overall, it is diffi cult to conceive of a less likely tissue that attracts dispersers does not con- islands (A. Lindstrom, personal communica- candidate for neurotoxin ‘biomagnifi cation’ form to a logical interpretation of a defensive tion). Thus, exposure to a C. micronesica toxin than BMAA. strategy. On the contrary, tissues consumed by via bat consumption would not have declined A detailed look at partitioning of chemicals a seed disperser are equipped with a suite of throughout these decades of importing bats within plant tissues is useful for shedding light biochemicals that attract and reward dispers- to satisfy the dietary demands of Guam resi- on putative chemical defense. For example, the ers. For example, C. micronesica sarcotesta dents. Neurotoxic exposure to BMAA via bat steryl glucosides appear in abundance in the tissue emits pungent fruity volatiles after the consumption clearly does not fi t the patterns sclerotesta tissue (Marler et al., 2005). This seed exterior morphs from bronze to dark of disease incidence. tissue is analogous to the stony endocarp of a brown. Seeds are odorless before this phase. The original manuscript by Cox and Sacks drupe, and is not at risk of herbivore consump- A sound interpretation of this characteristic is (2002) which fi rst proposed the hypothesis of tion because of its physical characteristics. that sarcotesta from Guam’s cycad is outfi tted an involvement of bats as cycad toxin vectors Thus, evolutionary herbivore pressures were for attracting seed dispersers that are drawn by contained estimated historical numbers rather not likely to select for preferential loading of aroma rather than sight, and those volatiles are than real data. These authors purport to show these compounds in the sclerotesta, and their a signal to dispersers of seed maturity. Indeed, that the decline in the bat population parallels abundance in this tissue points to the probability ethology of fruit bats indicates bat fruits pro- the decline in the incidence of ALS–PDC, the of an alternative plant function. duce large seeds with drab color and strong latter data compiled by neuroepidemiologists Correlating human diseases with plant tox- odors (van der Pijl, 2002). These descriptors working on Guam. While the human data are ins. An emerging view in neurological disease are characteristic of C. micronesica seeds, well established, the subjective estimations research, particularly concerning age-depen- and we suggest the notion of loading this tis- about bat population are tenuous. Scrutiny of dent disorders, is that environmental factors sue that attracts a disperser and is consumed their plots (Cox and Sacks, 2002) reveals an likely play a key role in disease genesis (e.g., by a disperser with a toxin for the purpose of accounting of bat populations that span the Shaw and Wilson, 2003). In this view, various deterring consumption by the disperser does majority of the 20th century, yet factual data natural and synthetic toxins, acting against a not appear plausible. The presence of second- points do not exist before about 1960 or sub- backdrop of variations in genetic susceptibility, ary compounds truly involved in defense are sequent to ca. 1986. The authors have drawn initiate biochemical cascades which culminate linked to increased plant fi tness (e.g., Ohnmeiss their graph with a sharp extrapolated upward in neural cell death. As discussed throughout and Baldwin, 2000), not linked to eliciting defl ection to a fabricated point of 60,000 bats this article, abundant evidence suggests that this pathology in dispersers that improve seedling in the early 1920s. They do not attempt to is true for ALS–PDC at the very least and may recruitment potential. justify this nonexistent data point until a year be more broadly true for neurological diseases Cycad neurotoxins are unlikely candidates later (Monson et al., 2003), when they opine outside of Guam. A similar view concerning the for defensive compounds because of failures that Guam could have supported this many bats impact of gene × environment interactions has to demonstrate direct effects on herbivore based on their estimated percentage of forest been posited for various forms of cancer and performance. The majority of research on coverage in the early 1900s. This curve forms cardiovascular disease (Willet, 2002). defensive secondary chemistry has examined the entire initial basis for the bat hypothesis, As this nascent aspect of medical sci- this direct infl uence on herbivore performance yet rests solely on a fabricated curve using ence progresses, establishing theoretical and (Karban and Baldwin, 1997). Thus, defensive an estimated bat population that could have experimental standards for correlating any compounds are directly or indirectly linked been supported by an estimated forest size. environmental toxin to any human disease to decreased herbivory. A foundation of the Clearly without real data, there is no justifi able is warranted. We use the discussions in the Cox and Sacks (2002) hypothesis includes claim to any correlation of bat population with present paper as an example toward this end. long-term ingestion of sarcotesta tissue by fruit ALS–PDC incidence. A wholesale focus on animal systems and bats in order for ‘biomagnifi cation’ (we believe Scrutiny of BMAA itself reveals an ex- disease epidemiology is not advisable if rel- bioaccumulation is a more appropriate term) tremely weak agonist at the N-methyl-D-as- evant botanical and habitat factors are ignored. to occur. A truly defensive chemical would partate (NMDA) receptor (Weiss et al., 1989), This is precisely what has happened over the deter herbivory long before bioaccumulation is one of the three primary ionotropic receptors decades with the Guam cycad research. On given the time to transpire. Surely a compound to the neurotransmitter glutamate. This clas- the contrary, research addressing aspects of inducing a pathological response in humans sification scheme for glutamate receptor the environmental source of a proposed toxin many years after ingestion is equally unlikely subtypes is largely independent of mammalian should progress concomitantly with controlled to demonstrate a direct effect on herbivore species. Although some differences occur in research on animal responses to the toxin. For deterrence. subunit composition during development, the this to proceed effi ciently, a plant scientist Several aspects of the BMAA biomagnifi ca- functions, agonists, binding properties, and that is established in the geographic region tion hypothesis (Cox and Sacks, 2002) and the induced excitotoxicity are quite similar across of interest should be included in the research defensive notions of BMAA are problematic. species (Watanabe, 1996). BMAA applied to team (Bawa et al., 2004). First, the spike in incidence of ALS–PDC in mammalian neurons in vitro evokes activation We mention one practical outcome as an the populations in Guam and Rota in the mid- of NMDA receptors leading to an infl ux of example of benefi ts from employing this dual 20th century followed by a continuing decline Ca2+ followed by a membrane depolarization. approach in the early stages of studying on a in incidence of the disease indicates a decline Over-stimulation of NMDA receptors by glu- putative toxin. Various compounds isolated in exposure to the causal environmental toxin tamate, BMAA, or other agonists can lead to from plants may have toxic and/or therapeutic

HORTSCIENCE VOL. 40(6) OCTOBER 2005 1603

OOctober.indbctober.indb 11603603 88/11/05/11/05 88:55:32:55:32 AAMM potentials. For example, the various steryl glu- Finally, in future research designed for conspecifi c and interspecifi c neighbors, recent cosides described in the present article appear to correlating human diseases with plant toxins, weather patterns, horticultural inputs, and be toxic to neural cells, but may have quite dif- the issue of sampling error must include many other factors may entirely explain their ferent outcomes in other organ systems (Shaw respect for sample size. In our initial reports reported variation in secondary compounds. In and Wilson, 2003). The mandatory protocols (Marler et al., 2005) we obtained samples for contrast to their interpretations that tissue type using animal systems to explore the medical each experiment from plants that conformed (Banack and Cox, 2002) or species (Yagi, 2004) and therapeutic applications of plant molecules to a consistent seed age, plant size, seed load, account for variation in sample group means, require the selective and economical harvesting level of canopy cover, and plant life history. these undocumented plant and habitat factors of targeted molecules. This research will be Despite this focus on removing probably bi- that infl uence phenotypic plasticity may fully ideally served by determining the most effi cient ases, we still observed considerable variation explain the variation in sample group means. and cost-effective protocols for harvesting the attributed to experimental error. We suggest We argue that our approach serves as a molecules. In our studies, the data relate to the the inconsistencies in past research may be model for defi ning theoretical and experimental established and ongoing research on metabo- entirely explained by ignoring the need to approaches to accurately correlate any hu- lites in mature seed megagametophyte tissue, account for these sources of bias, ignoring man disease with environmental toxins. The the source of fl our that historically entered the the need for a deliberate sampling scheme, methods described will enable researchers human diet in Guam. However, we demonstrate and ignoring the need to defi ne an adequate to accurately correlate toxin distributions in that immature 1-year-old seeds may yield a sample size. Our results underscore the reality the human environment with incidence of greater quantity of steryl glucosides than the that secondary compounds of C. micronesica disease. For continued ALS–PDC research, mature seeds that have been used throughout plants must be studied with inclusion of this other candidate neurotoxins should now be the decades for harvesting putative toxins (see information for conclusions to be accurate and given full scrutiny using these procedural Marler et al., 2005). We also demonstrate that consequential. and conceptual guidelines. The failures of sarcotesta tissue may yield a greater quantity past ALS–PDC research teams to honor fun- of steryl glucosides than megagametophyte tis- Conclusions damental plant science standards, to adhere sue. Thus, the costs of isolating and harvesting to acceptable sampling protocols, and to these neurotoxins for animal toxicology and Methods for studying possible plant toxins document mandatory background plant and epidemiology studies may be considerably need to be standardized and must follow cer- neighborhood information cannot continue if reduced if protocols are modifi ed from using tain rigorous criteria to be valid in principle. we are to successfully use this area of study mature megagametophyte tissue to including These criteria include testing to determine to improve human health issues. immature sarcotesta tissue. adequate sample size, honoring the need for We suggest these implications from studies a sampling scheme that avoids known biases, Literature Cited extend beyond ALS–PDC research. Indeed, recognizing the protected and sensitive nature Allison, S.D. and P.M. Vitousek. 2004. Rapid nutri- attempts at correlating exposure to any plant of cycad populations, and factoring in known ent cycling in leaf litter from invasive plants in molecule with any human disease is best served plant and habitat features that infl uence plant Hawaii. Oecologia 141:612–619. by determining the most effi cient and economi- secondary metabolism. Ambike, S.H. and M.R.R. Rao. 1967. Studies on a cal protocol for harvesting the molecules. This We have applied this empirical approach phytosterolin from the bark of Ficus religiosa. cannot be accomplished by following past to our study with steryl glucoside variants Indian J. Pharmacol. 29:91–94. examples of ignoring the importance of the (Marler et al., 2005) due to evidence that Baldwin, I.T. and M.J. Karb. 1995. Plasticity in plant and the plant’s habitat in defi ning those such molecules in cycad are in part causal allocation of nicotine to reproductive parts in Ni- protocols. to ALS–PDC. We suggest that this approach cotiana attenuata. J. Chem. Ecol. 21:897–909. Banack, S.A. and P.A. Cox. 2003. Distribution of Zhang et al. (1996) recognized the defi cien- be viewed as mandatory for clarifying the the neurotoxic nonprotein amino acid BMAA cies in past research on Guam’s neurological need to establish satisfactory theoretical and in Cycas micronesica. Bot. J. Linn. Soc. diseases and described the need to begin docu- experimental protocols. Our results reveal that 143:165–168. menting various aspects of cycad sample his- these criteria are essential and highlight the Bawa, K.S., W.J. Kress, N.M. Nadkarni, and N.M. tory, including specifi cs such as location of the impact that natural variations can have on toxin Nadkarni. 2004. Beyond paradise—Meeting the sampled plant. Thus, the continued disregard quantifi cation in plants, and more importantly, challenges in tropical biology in the 21st century. for documenting and reporting these plant and on the interpretation of results. Biotropica 36:437–446. habitat characteristics (e.g., Banack and Cox, Past research on the relationship between Bloom, A.J., F.S. Chapin, III, and H.A. Mooney. 2003) is unfortunate. Our hope is our initial cycad compounds and ALS–PDC fails to meet 1985. Resource limitation in plants—An economic analogy. Annu. Rev. Ecol. Syst. studies describing the need for appropriate these criteria. Throughout decades of articles on 16:363–392. sampling protocol (Marler et al., 2005) and this subject, documentation of plant and habitat Brenes-Arguedas, T. and P.D. Coley. 2005. Pheno- the description of various factors that infl uence factors known to infl uence plant physiology typic variation and spatial structure of secondary phenotypic plasticity of plants contained in the has not occurred before our studies (Marler chemistry in a natural population of a tropical present paper validate the reality of the Zhang et al., 2005). The obvious conclusion is that tree species. Oikos 108:410–420. et al. (1996) assertions and will bring forth that researchers and reviewers have assumed these Brenner, E.D., D.W. Stevenson, and R.W. Twigg. needed change in the defi cient methods that plant and environment characteristics that elicit 2003. Cycads: evolutionary innovations and the have plagued the ALS–PDC research since phenotypic plasticity in plant physiology are role of plant-derived neurotoxins. Trends Plant its inception. irrelevant. Unfortunately, ongoing research Sci. 8:446–452. Bryant, J.P., F.S. Chapin, III, and D.R. Klein. These issues also must address the neces- continues to be plagued with these omissions. 1983. Carbon/nutrient balance of boreal plants sity of including conservation measures in Banack and Cox (2003) present intertissue in relation to vertebrate herbivory. Oikos all decisions. Many plant species that are of variation of several compounds by sampling 40:357–368 interest in horticulture, ethnobotany or medi- individuals in different geographic locations Caesar-TonThat, T.C. and V.L. Cochran. 2001. cal research are threatened or endangered, and for each tissue category, yet report no plant or Role of saprophytic basidiomycete soil fungus indiscriminate harvesting of tissue for isolating landscape descriptors to allow repeatability in aggregate stabilization, p. 575–579. In: D.E. a metabolite cannot be justifi ed regardless of their methods and ensure homogeneity of Stott, R.H. Mohtar, and G.C. Steinhardt (eds.). of the relative importance to human health. these infl uential factors among the individu- Sustaining the global farm. Certainly, research on any newly identifi ed als that were sampled. Yagi (2004) presents Callaway, R.M. 2002. The detection of neighbors by plants. Trends Ecol. Evol. 17:104–105. metabolite should address plant and habitat interspecies variation of several compounds, Callaway, R.M., S.C. Pennings, and C.L. Richards. factors at the early stages of the research in yet reports no plant or landscape descriptors to 2003. Phenotypic plasticity and interactions order to fully honor mandatory conservation ensure homogeneity of these infl uential factors among plants. Ecology 84:1115–1128. measures when developing strategic and ef- among the species. Heterogeneity of plant his- Cantatore, J.L., S.M. Murphy, and D.V. Lynch. fi cient research protocols. tory, fecundity, allometry, shade, ontogenesis, 2000. Compartmentation and topology of glu-

1604 HORTSCIENCE VOL. 40(6) OCTOBER 2005

OOctober.indbctober.indb 11604604 88/11/05/11/05 88:55:34:55:34 AAMM cosylceramide synthesis. Biochem. Soc. Trans. daceae) in New Guinea and the Western Pacifi c. procedures. HortScience 40(6):1607–1611. 28:748–750. Austral. Syst. Bot. 7:543–567. Marler, T.E. and L.E. Willis. 1997. Leaf gas-exchange Castro, J., R. Zamora, J.A. Hodar, and J.M. Gomez. Hill, K.D and D.W. Stevenson. 2005. World list of characteristics of sixteen cycad species. J. Amer. 2002. Use of shrubs as nurse plants: A new cycads. 4 May 2005. http://plantnet.rbgsyd.gov. Soc. Hort. Sci. 122:38–42. technique for reforestation in Mediterranean au/PlantNet/cycad/wlist.html. Miles, D.B., G.S. Smith, and S.A. Miller. 1996. mountains. Restoration Ecol. 10:297–305. Hirsh, H. and T. Marler. 2002. Damage and recov- Within plant sampling procedures—Fruit varia- Chabot, B.F. and D.J. Hicks. 1982.The ecology of leaf ery of Cycas micronesica after Typhoon Paka. tion in kiwifruit vines. Ann. Bot. 78:289–294. life spans. Annu. Rev. Ecol. Syst. 13:229–259. Biotropica 34:598–602. Monson, C.S., S. A. Banack, and P.A. Cox. 2003. Chen, K.-M., U.K. Craig, C.-T. Lee, and R. Had- Holzapfel, C. and B.E. Mahall. 1999. Bidirectional Conservation implications of Chamorro con- dock. 2002. Cycad neurotoxins, consumption of facilitation and interference between shrubs sumption of fl ying foxes as a possible cause fl ying foxes, and ALS–PDC disease in Guam. and annuals in the Mojave Desert. Ecology of amyotrophic lateral sclerosis–Parkinsonism Neurology 59:1664–1665. 80:1747–1761. dementia complex in Guam. Conserv. Biology Childers, N.F. 1995. Modern fruit science: Orchard Howard, F.W., A. Hamon, M. McLaughlin, T. 17:678–686. and small fruit culture. Horticultural Publ., Weissling, and S.-L. Yang. 1999. Aulacaspis Murch, S.J., P.A. Cox, and S.A. Banack. 2004. A Gainesville, Fla. yasumatsui (Hemiptera: Sternorrhyncha: mechanism for slow release of biomagnifi ed Choler, P., R. Michalet, and R.M. Callaway. 2001. Fa- Diaspididae), a scale insect pest of cycads cyanobacterial neurotoxins and neurodegenera- cilitation and competition on gradients in Alpine recently introduced into Florida. Fla. Entomol. tive disease in Guam. Proc. Natl. Acad. Sci. USA plant communities. Ecology 82:3295–3308. 82:14–27. 101:12228–12231. Costa, J.-L., P. Paulsrud, and P. Lindblad. 1999. Karban, R. and I.T. Baldwin. 1997. Induced re- Muthukumar, T. and K. Udaiyan. 2002. Arbuscular Cyanobiont diversity within coralloid roots of sponses to herbivory. Chicago Univ. Press, mycorrhizas in cycads of southern India. Mycor- selected cycad species. FEMS Microbiol. Ecol. Chicago, Ill. rhiza 12:213–217. 28:85–91. Khabazian, I., J.S. Bains, D.E. Williams, J. Cheung, Niklas, K.J. 1993. The allometry of plant reproduc- Cox, P.A. and O.W. Sacks. 2002. Cycad neurotoxins, J.M.B. Wilson, B.A. Pasqualotto, S.L. Pelech, tive biomass and stem diameter. Amer. J. Bot. consumption of fl ying foxes, and ALS–PDC R.J. Andersen, Y.-T. Wang, L. Liu, A. Nagai, S.U. 80:461–467. disease in Guam. Neurology 58:956–959. Kim, U.-K. Craig, and C.A. Shaw. 2002. Isolation Niklas, K.J. 2004. Plant allometry: is there a grand De Silva, H.N., A.J. Hall, W.M. Cashmore, and of various forms of steryl β-D-glucoside from unifying theory? Biol. Rev. 79:871–889. D.W. Tustin. 2000. Variation of fruit size and the seed of Cycas circinalis: Neurotoxicity and Nilsson, M., B. Bergman, and U. Rasmussen. 2000. growth within an apple tree and its infl uence on implications for ALS–parkinsonism dementia Cyanobacterial diversity in geographically re- sampling methods for estimating the parameters complex. J. Neurochem. 82:516–528. lated and distant host plants of the genus Gunnera. of mid-season size distributions. Ann. Bot. Kunimoto, S., W. Murofushi, H. Kai, Y. Ishida, Arch. Microbiol. 173:97–102. 86:493–501. A. Uchiyama, T. Kobayashi, S. Kobayashi, H. Norstog, K.J. and T.J. Nicholls. 1997. The biology of Ehlers, B.K. and J. Thompson. 2004. Do co-occur- Murofushi, and K. Murakami-Murofushi. 2002. the cycads. Cornell Univ. Press, Ithaca, N.Y. ring species adapt to one another? The response Steryl glucoside is a lipid mediator in stress- Ohnmeiss, T.E. and I.T. Baldwin. 2000. Optimal de- of Bromus erectus to the presence of differ- responsive signal transduction. Cell Structure fense theory predicts the ontogeny of an induced ent Thymus bulgaris chemotypes. Oecologia Funct. 27:157–162. nicotine defense. Ecology 81:1765–1783. 141:511–518. Kurland, L.T. 1988. Amyotrophic lateral sclerosis Ow, M.C., M. Gantar, and J. Elhai. 1999. Recon- Falik, O., P. Reides, M. Gersani, and A. Novoplansky. and Parkinson’s disease complex on Guam linked stitution of a cycad–cyanobacterial association. 2003. Self/non-self discrimination in roots. J. to an environmental toxin. Trends Neurosci. Symbiosis 27:125–134. Ecol. 91:525–531. 11:51–53. Palm and Cycad Societies of Australia. 1993. Fisher, J.B. and A.P. Vovides. 2004. Mycorrhizae are Kurland, L.T. 1993. Cycas circinalis as an etiologic The biology, structure, and systematics of the present in cycad roots. Bot. Rev. 70:16–23. risk factor in amyotrophic lateral sclerosis and Cycadales. Proc. 2nd Intl. Conf. Cycad Biol- Garruto, R.M., S.K. Shankar, R. Yanagihara, A.M. other neurodegenerative diseases on Guam, p. ogy, Townsville, Queensland, Australia, 22–28 Salazar, H.L. Amyx, and D.C. Gajdusek. 1989. 29–36. In: D.W. Stevenson and K.J. Norstog July 1990. Low-calcium, high-aluminum diet-induced (eds.). Proceedings of Cycad 90, the Second Peters, R.H., S. Cloutier, D. Dube, A. Evens, P. pathology in cynomologus monkeys. Acta International Conference on Cycad Biology. Hastings, H. Kaiser, D. Kohn, and B. Sarwer- Neuropathol. 78:210–219. Palm and Cycad Soc. Austral., Ltd. Milton, Foner. 1988. The allometry of the weight of Gebhardt, J.S. and S.A. Nierzwicki-Bauer. 1991. Queensland. fruit on trees and shrubs in Barbados. Oecologia Identifi cation of a common cyanobacterial Kurland, L.T, K. Radhakrishnan, D.B. Williams, and 74:612–616. symbiont associated with Azolla spp. through S.C. Waring. 1994. Amyotrophic lateral sclero- Rillig, M.C., S.F. Wright, M.S. Torn, and K.A. molecular and morphological characterization of sis–parkinsonism–dementia complex on Guam: Nichols. 2001. Unusually large contribution free-living and symbiotic cyanobacteria. Appl. Epidemiological perspectives, p. 109–130. In: A. of arbuscular mycorrhizal fungi to organic Environ. Microbiol. 57:2141–2146. Williams (ed.). Motor neuron disease. Chapman matter pools in tropical forest soils. Plant Soil Gersani, M., J.S. Brown, E.E. O’Brien, G.M. Maina, and Hall, London. 233:167–177. and Z. Abramsky. 2001. Tragedy of the com- Lobakova, E.S., G.A. Dubravina, and N.V. Za- Rogers, R.F. 1995. Destiny’s Landfall—A history of mons as a result of root competition. J. Ecol. goskina. 2004. Formation of phenolic compounds Guam. Univ. Hawaii Press, Honolulu. 89:660–669. in apogeotropic roots of cycad plants. Russ. J. Santa Regina, I. and T. Tarazona. 2001. Nutrient Hartmann, M.-A. 2003. Sterol metabolism and Plant Physiol. 51:486–493. cycling in a natural beech forest and adja- functions in higher plants, p. 183–211. In: G. Loranger, G., J.-F. Ponge, D. Imbert, and P. Lavelle. cent planted pine in northern Spain. Forestry Daum (ed.). Topics in current genetics 6: Lipid 2002. Leaf decomposition in two semi-evergreen 74:11–28. metabolism and membrane biogenesis. Springer- tropical forests: Infl uence of litter quality. Biol. Schenk, H.J., C. Holzapfel, J.G. Hamilton, and B.E. Verlag, Heidelberg. Fertility Soils 35:247–252. Mahall. 2003. Spatial ecology of a small desert Heil, M. and I.T. Baldwin. 2002. Fitness costs of Mabry, T.J. 2001. Selected topics from forty years shrub on adjacent geological substrates. J. Ecol. induced resistance: Emerging experimental of natural products research: Betalains to fl avo- 91:383–395. support for a slippery concept. Trends Plant noids, antiviral proteins, and neurotoxic nonpro- Shaw, C.A. and J.M.B. Wilson. 2003. Analysis of Sci. 7:61–67. tein amino acids. J. Nat. Prod. 64:1596–1604. neurological disease in four dimensions: Insight Herms, D.A and W.J. Mattson. 1992. The dilemma Madritch, M.D. and M.D. Hunter. 2002. Pheno- from ALS–PDC epidemiology and animal mod- of plants: To grow or defend. Qrtly. Rev. Biol. typic diversity infl uences ecosystem function- els. Neurosci. Biobehavioral Rev. 27:493–505. 67:283–335. ing in an oak sandhills community. Ecology Stone, B.C. 1970. The fl ora of Guam. Micronesica Hevia, F., M.L. Minoletti O., K.L.M. Decker, and 83:2084–2090. 6:1–659. R.E.J. Boerner. 1999. Foliar nitrogen and phos- Marler, T.E. 2002. Conservation of local genotypes Tsimilli-Michael, M., P. Eggenberg, B. Biro, K. phorus dynamics of three Chilean Nothofagus when planting native plants on Guam. Microne- Koves-Pechy, I. Voros, and R.J. Strasser. 2000. (Fagaceae) species in relation to leaf lifespan. sica Suppl. 6:55–60. Synergistic and antagonistic effects of arbus- Amer. J. Bot. 86:447–455. Marler, T.E. and J.H. Lawrence. 2004. Chemical cular mycorrhizal fungi and Azospirillum and Hill, K.D. 1993. History of the taxonomy of Austra- limitations of yoga root growth in an acid soil. Rhizobium nitrogen-fi xers on the photosynthetic lian cycads, p. 250–259. Proc. 2nd Intl. Conf. on Micronesica 37:157–161. activity of alfalfa, probed by the polyphasic Cycad Biology. Palm and Cycad Soc. Austral. Marler, T.E., V. Lee, and C.A. Shaw. 2005. Spatial chlorophyll a fl uorescence transient O-J-I-P. Ltd. Milton, Queensland. variation of steryl glucosides in Cycas microne- Appl. Soil Ecol. 15:169–182. Hill, K.D. 1994. The Cycas rumphii complex (Cyca- sica plants—Within- and among-plant sampling van Dam, N.M. and I.T. Baldwin. 2001. Competition

HORTSCIENCE VOL. 40(6) OCTOBER 2005 1605

OOctober.indbctober.indb 11605605 88/11/05/11/05 88:55:37:55:37 AAMM mediates cost of jasmonate-induced defenses, N (BOAA) on cultured cortical neurons. Brain adult mice fed washed cycad fl our. Neuromol. acquisition and transgenerational plasticity in Res. 497:64–71. Med. 1:207–221. Nicotiana attenuata. Funct. Ecol. 15:406–415. Whiting, M.G. 1963. Toxicity of cycads. Econ. Bot. Yagi, F. 2004. Azoxyglycoside content and β-gly- van der Pijl, L. 2002. Principles of dispersal in higher 17:271–302. cosidase activities in leaves of various cycads. plants, p. 309–321. In: R.L. Chazdon and T.C. Wiles, G.J. 1992. Recent trends in the fruit bat Phytochemistry 65:3243–3247. Whitmore (eds.). Foundations of tropical forest trade on Guam, p. 53–60. In: D.E. Wilson and Young, F.J. 1988. Soil survey of Territory of Guam. biology. Univ. Chicago Press, Chicago. G.L. Graham (eds.). Pacifi c island fl ying foxes: U.S. Dept. Agr. Soil Cons. Serv. Viti, R., S. Bartolini, and R. Guerriero. 2003. The Proceedings of an international conservation Zaninovic, V. 2004. Possible etiologies for tropical infl uence of sampling from different canopy posi- conference. U.S. Fish Wildlife Serv. Biol. Rpt. spastic paraparesis and human T lymphotropic tions on the evaluation of fl ower bud anomalies 90(23). virus I-associated myelopathy. Braz. J. Med. and dormancy. Fruits 58:117–126. Willet, W.C. 2002. Balancing life-style and genom- Biol. Res. 37:1–12. Watanabe, M. 1996. Developmental regulation of ics research for disease prevention. Science Zhang, Z.X., D.W. Anderson, N. Mantel, and G.C. ionotropic glutamate receptor gene expression 296:695–698. Roman. 1996. Motor neuron disease on Guam: and functional correlations, p. 73–90. In: C.A. Wilson, J.M., I. Khabazian, M.C. Wong, A. Seyeda- geographical and familial occurrence, 1956–85. Shaw (ed.). Receptor dynamics in neural devel- likhani, J.S. Bains, B.A. Pasqualotto, D.E. Wil- Acta Neurol. Scand. 94:51–59. opment. CRC Press, Boca Raton, Fla. liams, R.J. Andersen, R.J. Simpson, R. Smith, Zotz, G. 2000. Size-related intraspecifi c variability Weiss, J.H., J.Y. Koh, and D.W. Choi. 1989. Neu- U.K. Craig, L.T. Kurland, and C.A. Shaw. in physiological traits of vascular epiphytes and rotoxicity of beta-N-methylamino-L-alanine 2002. Behavioral and neurological correlates its importance for plant physiological ecology. (BMAA) and beta-N-oxalylamino-L-alanine of ALS–parkinsonism dementia complex in Perspectives Plant Ecol. Evol. Syst. 3:19–28.

1606 HORTSCIENCE VOL. 40(6) OCTOBER 2005

OOctober.indbctober.indb 11606606 88/11/05/11/05 88:55:40:55:40 AAMM