Journal of Tropical Ecology (2012) 28:233–242. © Cambridge University Press 2012 doi:10.1017/S0266467412000119

Demography of micronesica on following introduction of the armoured scale

Thomas E. Marler∗,1 and John H. Lawrence†

∗ Western Pacific Tropical Research Center, CNAS, University of Guam, UOG Station, Mangilao, Guam 96923, USA † United States Department of Agriculture, Natural Resources Conservation Service, Mongmong, Guam, 96910, USA (Accepted 4 February 2012)

Abstract: Following the 2003 invasion of the armoured scale Aulacaspis yasumatsui to Guam, changes to population traits of the dominant Cycas micronesica were determined. Belt transects with a width of 4 m and an average length of 120 m were established in October 2004 to document mortality until January 2011. Stem height, basal diameter and leaf number were also measured for each plant and used to determine density, demography and allometric relationships. Allometric traits and a left-skewed demographic structure of the pre-invasion C. micronesica habitat documented a thriving population with high recruitment potential. Aulacaspis yasumatsui dispersed into the study site 4 mo after the initial census. All seedlings were killed within 9 mo and all juvenile were killed within 40 ± 10 mo. Mortality reached 92% by 6 y after chronic scale infestations. Allometry and demography of the 2011 survivors described a collapsing C. micronesica population of stressed and reproductively challenged trees with no recruitment. This classic example of the enemy release hypothesis has resulted in a homogeneous decline in plant density from 2007–2011. The trend predicts extirpation of C. micronesica from west Guam habitats by 2019.

Key Words: Cycadaceae, endangered species, Guam, invasion ecology, oceanic islands, population dynamics, survival

INTRODUCTION producer and exporter of Cycas revoluta Thunb., an economically significant landscape standard globally. Populations of many island endemic species are Exportation of C. revoluta resulted in the sequential threatened by human-induced environmental changes. dispersal of A. yasumatusi in numerous settings globally. The challenge confronting island ecologists is to clearly Aulacaspis yasumatsui was documented in Hawaii in 1998 identify causal mechanisms and unambiguous metrics to (Heu et al. 1999), in Guam in 2003, and its dispersal measure and document changes in population viability. from the initial outbreak site within urban landscapes The unintentional introduction of alien species to insular into native limestone forest habitat was confirmed in early areas is one means by which human activity disrupts 2005 (Terry & Marler 2005). ecologicalprocesses,andsmalloceanicislandsareexceed- The endemic range for Cycas micronesica (Hill 2004) ingly vulnerable to these disruptions. Specialist arthropod extends from some islands in the Republic of as the herbivore invaders are particularly threatening to plants southern boundary through Yap State in the Federated in insular areas because the native plants lack a shared States of , Guam, and Rota in the Marianas evolutionary history with the invader (Elton 1958), and as the northern boundary. Forest inventories in 2002 the invaders are not constrained by the natural pressures revealedC.micronesicawasthemostabundanttreespecies of their native range (Keane & Crawley 2002). in Guam’s various habitats (Donnegan et al. 2004). Cycas The armoured specialist scale Aulacaspis yasumatsui micronesica is the only native host of A. yasumatsui on was described from collections in Thailand (Takagi 1977) Guam, raising concerns about a possible decline of the and is known to attack several genera. It was cycad and its mutualists, as well as potential negative unintentionally introduced into southern Florida about cascading ecosystem responses (Moir et al. 2011). The 20 y ago (Howard et al. 1999). Florida was a major unique attributes and dominant status of this cycad valid- ate the prediction that its removal from native forests will exert manifold ecological changes. These changes require 1 Corresponding author. Email: [email protected] experimental and conceptual effort to be fully understood. 234 THOMAS E. MARLER AND JOHN H. LAWRENCE

Demographic studies are an important tool for Plot establishment and multi-trophic developments evaluating the condition of plant populations and monitoring changes in population status over time We established four belt transects with a width of 4 m (Schemske et al. 1994). Our objective in this study and an average length of 120 m (range 102–142 m) in was to assess recent changes in the demography of October 2004. The initial number of plants within the Guam’s C. micronesica population, monitoring what transects was 687. We re-visited the study site weekly started as a healthy, stable cycad population. We thereafter to ensure accurate assessment of the initial quantified size-based patterns of mortality and changes dispersal of A. yasumatsui within the study site. This to the demography and allometric traits of this occurred in January 2005. We purposely introduced the population after successive migrations of A. yasumatsui. armoured scale predator Blaisdell Long-term infestations of this armoured scale lead (Coleoptera, Coccinellidae) to Guam in November 2004, to sequential defoliation events with each successive and made a release at the study site in February flush of leaves generating fewer and smaller leaves, a 2005 (Moore et al. 2005). We also discovered Chilades process that ultimately results in plant death (Terry pandava Horsfield, a Cycas-specific butterfly that feeds & Marler 2005). The roots and pachycaul stems of on young, expanding tissues, in the study site in 2005 healthy are comprised mostly of live parenchyma (Moore et al. 2005). Its discovery validated yet another tissues containing abundant starch (Marler et al. 2010, problematic alien introduction to Guam. An ephemeral Norstog & Nicholls 1997). Therefore, our hypothesis irruption of Dihammus marianarum Aurivillius (Marler & was the initial pool of non-structural carbohydrate as Muniappan 2006) occurred within the study site in late dictated by size of the plant correlates with length 2007. This native longhorn beetle is a borer that feeds of time from initial infestation until plant death. Our on C. micronesica stems, primarily on stressed plants. results will inform ecological principles of rare plant A temporary decline in the R. lophanthae population populations in small islands and fragmented continental occurred in late 2008, which initiated a secondary populations. irruption of A. yasumatsui. This epidemic was also short- lived.

METHODS Measurements and data analysis

Study site We visited the site approximately annually to conduct a typeIright-censoringapproach(Pyke&Thompson1986) Following the 2003 unintentional introduction of A. in which we documented plant mortality to January yasumatsui to Guam, we scouted various native forest 2011. Plants in the initial census were tagged to ensure habitats to locate a representative study site where we accuracy during subsequent data collection. In addition could determine population-level responses to anticipated to the survival data, we measured stem height, basal migrations of A. yasumatsui. We selected a high-density, diameter and leaf number for each individual plant. As an healthy population in western Guam where Cycas estimation of reproductive effort, sex was assigned to each micronesica stem density was in excess of 3500 plants individual using any signs of reproductive structures. ha−1. This site resides within the federally managed This approach included documenting any persistent GuamNationalWildlifeRefuge,centredat13.645278◦N, reproductive structures from past reproductive events, 144.858333◦E. The complex of soils is dominated by not just the current season’s structures. a clayey-skeletal, gibbsitic, non-acid, isohyperthermic We calculated the density of plants from each census Lithic Ustorthents (Young 1988). This soil supports within height categories. The data were organized as most of the highest-density cycad habitats on Guam. a two-factor (height category × month of census) The C. micronesica population is the most genetically ANOVA using SAS Proc GLIMMIX with AR(1) covariance isolated population on Guam, and deserves paramount structure, with month treated as a repeated-measures conservation efforts (Cibrian-Jaramillo´ et al. 2010). variable. In addition, we organized juvenile plants less Severalecologicalstudieshavebeenconductedatthissite, than 100 cm in height into four size categories then providing a more robust understanding of the site than recorded the number of months required for each many other candidate sites. Its vegetation is classified as height category to reach 100% mortality. A single- secondary limestone forest, with Aglaia, Ficus, Meiogyne, factor ANOVA was performed using SAS Proc GLM. We Macaranga, Neisosperma, Ochrosia, Pandanus, Premna and compared 2004 and 2011 frequency data among various Pisonia being the common native genera. In addition to size categories for stem height, stem diameter and leaf invasive insects, this habitat is severely threatened by two number with the Wilcoxon rank-sum test (Wilcoxon feral ungulate species and invasive plants. 1945) using SAS Proc NPAR1WAY. For describing size invasion alters cycad demography 235

Figure 1. Demographic characteristics of Guam’s Cycas micronesica population in October 2004 (left column, n = 687) 4 mo prior to the establishment of the invasive Aulacaspis yasumatsui in the study area and in January 2011 (right column, n = 57) after 72 mo of scale infestation. Percentage of individuals within height categories in 2004 (a) and 2011 (b). Percentage of individuals within stem diameter categories in 2004 (c) and 2011 (d). Percentage of individuals within total leaf number categories in 2004 (e) and 2011 (f). distributions, we calculated the skewness and kurtosis diameter and total leaf number (Figure 1, Table 1). The (Groeneveld & Meeden 1984, Joanes & Gill 1998) then degree of skewness and kurtosis was greatest for leaf employed the Lilliefors goodness-of-fit test for normality number. Variation in plant density over time revealed the (Lilliefors 1967). For allometric relations, we used log10- main effects of height category (F = 31.0, df = 5,126, P ≤ transformed data to determine log-log linear relationships 0.0001) and census date (F = 128, df = 6,126, P ≤ between stem height and diameter, stem height and leaf 0.0001) were significant, as was their interaction (F = number, and stem diameter and leaf number for the 2004 40.1, df = 30,126, P ≤ 0.0001). The initial decline and 2011 data. First we determined the slope of each in density of seedling and juvenile plants 1–100 cm regression for each transect. Thereafter a dataset was in height was extreme, whereas the decline in density created with the slope as the response variable and a of plants above 100 cm in height was more gradual single-factorANOVAwasperformedusingSASProcGLM. (Figure 2). All seedlings within the transects were killed within 9 mo after the initial A. yasumatsui infestation.

RESULTS Timing of juvenile mortality Population demography The number of months required for juvenile plants to The 2004 C. micronesica population presented a left- reach 100% mortality differed among four 25-cm height skewed frequency distribution for stem height, stem basal increments (F = 6.4, df = 4,15, P ≤ 0.0033). Mortality of 236 THOMAS E. MARLER AND JOHN H. LAWRENCE

Table 1. Characteristics of frequency distributions among size categories Table 2. Time required for juvenile Cycas micronesica for Cycas micronesica stem height, stem diameter and leaf number in plants to suffer 100% mortality following Aulacaspis October 2004, 4 mo prior to the establishment of Aulacaspis yasumatsui yasumatsui establishment, as documented by 25-cm- in the study area (Before) and in January 2011, 72 mo following the increment height categories, P ≤ 0.0033, n = 4 invasion event (After). transects. Least square means with the same letter Characteristic Before After are not significantly different. Stem height Size category (cm) Time (mo) Lilliefors test statistic 0.444 0.328 1–25 25 A Skewness 1.99 0.601 26–50 31 AB Kurtosis 6.56 3.47 51–75 34 AB Stem diameter 76–100 40 B Lilliefors test statistic 0.425 0.247 Skewness 1.66 −0.635 Kurtosis 4.61 2.67 that for the 2004 population (Figure 3). Additionally, the Leaf number range of the data for all three regressions in 2011 was Lilliefors test statistic 0.517 0.450 severely constricted by the plant mortality after 6 y of A. Skewness 3.78 158 Kurtosis 17.8 27 000 yasumatsui infestations.

Population trait changes

The distribution of size categories in the 2004 population differed significantly from that in the 2011 population for stem height (Wilcoxon statistic = 37, P ≤ 0.0301), stem diameter (Wilcoxon statistic = 40, P ≤ 0.0061), and leaf number (Wilcoxon statistic = 36, P ≤ 0.0453). Following 6 y of A. yasumatsui damage this population of C. micronesica exhibited a unimodal pattern for stem height (Figure 1). The degree of skewness declined to 30% of the 2004 values, and the degree of kurtosis declined to about half of the 2004 values (Table 1). The left-skewed 2004 distribution of stem diameter was reversed by Figure 2. The influence of plant height size categories (cm) on survival of 6 y of damage to become a right-skewed pattern Cycas micronesica following the establishment of Aulacaspis yasumatsui (Table 1, Figure 1). The left-skewed 2004 distribution in western Guam. The x-axis refers to January of each calendar year. of leaf numbers was retained in the 2011 population Arrow on x-axis marks the initial infestation of A. yasumatsui in the (Figure 1). In fact, this skewed pattern was magnified study habitat. by the response to years of A. yasumatsui damage as evidenced by a 42-fold increase in skewness the1–25-cmcategoryrequiredabout2y,whilethatofthe (Table 1). The lowest leaf number category increased 76–100-cm category required a mean of more than 3 y from c. 75% of the 2004 population to more than (Table 2). The census confirming ultimate loss of the final 90% of the 2011 population. Furthermore, the three juvenile plant was the January 2010 census. At this stage largest leaf number categories from the 2004 census when all of the juvenile plants had been killed, the mature declined to zero by the 2011 census. The lowest plants above 100 cm in height were at 73% mortality. leaf number categories in 2004 were due to the prevalent seedling and juvenile populations. In contrast, the 2011 data did not include any plants less than Allometric patterns 100 cm in height, so the trend of declining leaf number per plant occurred within a remaining population exclusively Differences in the slope of the 2004 versus 2011 comprisedofmatureplants.Lillieforsteststatisticrevealed populations of C. micronesica were apparent for stem strong evidence against normality for every population height versus diameter (F = 55.9, df = 1,6, P ≤ 0.0003), test (Table 1). The 2011 results portray the absence stem height versus leaf number (F = 130, df = 1,6, P ≤ of seedling and juvenile plants and predict imminent 0.0001) and stem diameter versus leaf number (F = population decline. 143.3, df = 1,6, P ≤ 0.0001). The 2011 population In order to increase our understanding of how 6 y of A. exhibited a slope for height versus diameter that was 40%, yasumatsui attack changed the C. micronesica populations, a slope for height versus leaf number that was 10%, and we calculated mean ± SE for several before–after variables a slope for diameter versus leaf number that was 13% of (Table 3). Total plant density declined to 8% of initial Scale insect invasion alters cycad demography 237

Table 3. Characteristics of Guam’s Cycas micronesica population in October 2004, 4 mo prior to the establishment of Aulacaspis yasumatsui in the study area (Before) and in January 2011, 72 mo following the invasion event (After). n = 4 transects, mean ± SE. Characteristic Before After Plant density (stems ha−1) 3580 ± 153 290 ± 63 Mean mature tree leaf number 5 ± 820± 3 Mean mature tree height (cm) 244 ± 10 258 ± 5 Total stem basal area (m2 ha−1) 93.7 ± 12.5 28.2 ± 6.3 Quotient sexed trees 0.49 ± 0.04 0.33 ± 0.09

The male/female quotient among these sexed individuals was 0.73 in 2004 and 0.63 in 2011, signifying an increase in the ability to identify signs of female plants relative to that of male plants.

Recruitment

This site contained C. micronesica seedlings at the density of 602 ha−1 in 2004. These were all killed within 9 mo of scale infestation, and the census in 2005 revealed a complete lack of seedlings. As the R. lophanthae predator began to exert some biological control of the damage by A. yasumatsui, but before the mature plant health declined substantially, a few new seedling recruits were apparent in this habitat. This was revealed as 118 seedlings ha−1 in January 2007 and 28 seedlings ha−1 in January 2008. All of these seedlings were killed by direct A. yasumatsui damage before the three-leaf stage. As mature plant Figure 3. Allometric relations of Guam’s Cycas micronesica population in October 2004, 4 mo prior to the establishment of the invasive Aulacaspis health continued to decline with each passing year, seed yasumatsui in the study area and in January 2011, after 72 mo of scale production and quality declined. As a result, no new infestation. Height versus diameter slope was 1.62 for 2004 and 0.65 seedlings have been observed persisting in this habitat for 2011 (a). Height versus leaf number was 1.28 for 2004 and 0.11 for since 2008. 2011 (b). Diameter versus leaf number was 0.71 for 2004 and 0.09 for An unusual increase in recruitment into the 1–100- 2011 (c). cm height category occurred in 2010. This was not the normal growth increment increase that promoted plant density, and mean leaf number for mature plants a cohort of individuals from the seedling category into declined to 24% of initial leaf number. Range in leaf the juvenile category. Instead, it occurred because many number for mature plants was 21–229 in 2004 and 1–72 mature plants lost most of their stem height due to in 2011. These characteristics of changes in leaf number the D. marianarum tunnelling that peaked in 2007– explain the drastic changes in the regressions of height 2008. This is the typical historical response to this versus leaf number and diameter versus leaf number pest where the entire upper portion of the stem and (Figure 3). In 2004 all of the individuals with few leaves leaves dies back to about 1 m height. The phenomenon were seedlings and small juveniles, so the y-intercept was decreased calculated plant density within some of the near the origin as both factors scaled together. In contrast, taller categories, but not due to mortality. It also increased the y-intercepts for both height and diameter were quite plant density in the 1–100-cm category, but not due to large for 2011 plants because all of the individuals with normal growth increment recruitment. This also partly few leaves were large mature specimens. explains the atypical height versus diameter slope for the Mean height of mature trees exhibited a non-significant 2011populationbyconstrictingtheheightrangewithout increase following 6 y of A. yasumatsui damage, while a parallel constriction of the diameter range (Figure 3). total stem basal area declined to 30% of the initial basal area during the same period (Table 3). About half of the mature tree population could be sexed in 2004, but only DISCUSSION about a third of the 2011 population could be sexed (Table 3), indicating a decline in reproductive effort We present a north-western Pacific island case study among the few remaining C. micronesica plants on Guam. describing the consequences of a predicted invasion that 238 THOMAS E. MARLER AND JOHN H. LAWRENCE empirically supports the contention that island-endemic rather than showing a left-skewed trend whereby the plant populations are acutely susceptible to invasive smallest mature trees were preferentially killed. Third, herbivores (Meyer & Butaud 2009). This is also a classic although we cannot determine the timing of 100% example of the enemy release hypothesis (Elton 1958, mortality for any of the mature tree size categories to Liu & Stiling 2006, Norghauer et al. 2011). First, A. date, we can determine the time to reach 50% mortality. yasumatsui populations within the scale’s native range co- The transition to at least 50% mortality occurred in occur with healthy native Cycas populations that exhibit the January 2007 census for 101–200-cm, 201–300- scale infestations but are not severely threatened (Tang cm and 401+ cm categories. This level of mortality for etal.1997).Second,manymatureC.micronesicatreesdied the 301–400-cm category was evident in the January during less than 1 y of A. yasumatsui infestation before R. 2009 census. Therefore, the benefits of being large are lophanthae became well-established, but after its effective very clear for juvenile plants, but less clear for mature establishment the R. lophanthae predator has allowed plants. Other factors may explain which mature trees are many trees to survive 6 y of A. yasumatsui infestation. most susceptible to A. yasumatsui damage, such as the Third, C. micronesica plants growing in ex situ plantings available pool of carbon reserves at the time of initial in Thailand where A. yasumatsui is controlled by various infestation (Galiano et al. 2011). natural enemies do not exhibit greater susceptibility to A. Determining the influences of climatic, edaphic, biotic yasumatsui infestations than other congeneric plants (A. J. and demographic factors on stand-level die-back may Lindstrom,¨ unpubl. data). Our results empirically validate aid in understanding the underlying causes of tree death the designation of endangered status for C. micronesica by (Mueller-Dombois 1986, 1987). Mortality of the mature the International Union for Conservation of Nature and plants in our study appears to conform to Manion’s con- Natural Resources. ceptual model of tree disease (Manion 1981). This model Initial establishment of A. yasumatsui in the study site positsthatpre-disposingfactorsrenderanindividualmore preceded that of R. lophanthae. This enabled a brief period or less vulnerable to subsequent stresses, which then of A. yasumatsui infestation without biological control. decrease vigour and increase vulnerability prior to impos- Unfortunately the size differential between R. lophanthae ition of the next stress. Ultimately a newly imposed stress and A. yasumatsui limits ongoing efficacy of biological can cause a final vigour decline and subsequent death. control, as the scale is able to find locations on the The chronic attack by A. yasumatsui imposed the major Cycas plant body that the predator cannot access (Marler stress factor that contributed to this ongoing phase of & Moore 2010). Regardless, the long-term protection matureC.micronesicaplantmortality.However,otherbio- afforded by this predator is undoubtedly responsible for logical factors contribute to the compilation of stressors. the mature trees that are still alive today. Contributing factors adding to the A. yasumatsui pressure include the chronic pressure of Chilades pandava (Moore et al. 2005) and Erechthias sp. (Marler & Size-related impacts Muniappan 2006) infestations. The combination of these three alien arthropod pests, all of which are very recent The first year of C. micronesica mortality following unintentional introductions on Guam, does not allow C. A. yasumatsui infestation was comprised primarily of micronesica plants the time to adequately recover carbon seedlings. As time passed the loss of seedlings was from the naturally long-lived leaves following initial sequentially followed by 100% mortality of plants within leaf construction. This forces the plant into a chronic the height range 1–25 cm, 26–50 cm, 51–75 cm, and state of carbon deficit. In addition to these leaf feeders, finally 76–100 cm. These results for seedling and juvenile the stem borer D. marianarum (Marler & Muniappan plants up to 100 cm in height were consistent with 2006) is a native longhorn beetle that preferentially our hypothesis. Moreover, they lend evidence for the attacks stressed plants, a behaviour common among hypothesis that carbon limitations and starvation are other Cerambycidae stem borers. Feral ungulates are also universal mechanisms of tree decline following biotic or threatening the health of remaining plants, as the pig (Sus abiotic stresses (McDowell et al. 2008). scrofa L.) consumes stems and the deer (Cervus mariannus Unlike juvenile plant mortality, the mature plants Desmarest) consumes leaves and reproductive organs, appeared to exhibit no overt relationship of size category adding to the multiple compounding insults from alien to survivorship. First, the mean height of mature trees peststothisislandendemic. did not change during the time frame of this study (Table 3), despite mortality and removal from the population of 75% of the mature trees. Second, 19% Stand structure of the mature trees in excess of 100 cm in height died within 9 mo of A. yasumatsui infestation, and these plants Understanding the structure of a plant population enables were distributed randomly among mature size categories the assessment of regeneration processes and reflects on Scale insect invasion alters cycad demography 239 the population health. The left-skewed size structure of Other cycads Guam’s C. micronesica population prior to A. yasumatsui damage is often called reverse-J shape, conforming to Cycads are the most threatened group of plant a Type III curve according to Deevey (1947) or Type I species on Earth (Hoffmann et al. 2010). Despite this curve according to Bongers et al. (1988). This structure unfortunate conservation condition and the importance suggests copious fecundity and favourable site conditions of demographic studies to assess population status, only for establishment and survival of seedlings. Indeed, only a few of the c. 300 described cycad species (Donaldson one-third of the 2004 population was comprised of 2003) have been the subject of demographic studies. mature trees greater than 100 cm in height. The leaf Merely one of these focused on a congeneric of our number size class distribution exhibited more asymmetry study species, when Keppel (2001) determined height and a broader peak in the distribution curve than did class distribution of one Cycas seemannii population. Other the stem height or stem diameter size class distribution, relevant studies included Ceratozamia matudae (Perez-´ as evidenced by the extent of skewness and kurtosis Farrera & Vovides 2004a, 2004b), Ceratozamia mirandae (Groeneveld & Meeden 1984, Joanes & Gill 1998). (Perez-Farrera´ & Vovides 2004b, Perez-Farrera´ et al. The ability of larger plants to delay mortality following 2006), Dioon edule (Vovides 1990), Dioon merolae (Lazaro-´ A. yasumatsui infestation generated extreme changes Zermeno˜ et al. 2011), Zamia debilis (Negron-Ortiz & in population structure. Therefore, only 6 y after A. Breckon 1989) and Zamia soconuscensis (Perez-Farrera´ & yasumatsui migrated into the study site a left-skewed Vovides 2004b). C. micronesica population exhibiting healthy trees with Without exception, these studies revealed populations bountiful active recruitment, this same population was that generally conformed to the classic left-skewed size transformed into a unimodal stem height structure and class structure. Our report, therefore, stands out as the a right-skewed stem diameter structure comprised of first to describe a unimodal population status for a cycad unhealthy trees exhibiting no recruitment. habitat as evidence of a severely threatened population. Although the general trend of the 2004 population was Furthermore, ours is the first study to our knowledge clearly left-skewed, this cycad population exhibited more that has quantified the size-structure response of a cycad than twice as many trees within the 201–300-cm height population to a biotic or abiotic disturbance by comparing category than within the 101–200-cm height category pre- and post-disturbance status within the same habitat. (Figure 2). This anomaly in what would have otherwise been a smooth curve may signify an ephemeral pulse in recruitment at some historical period that enabled the Prognosis greater number of 201–300-cm trees, or alternatively a more recent decline in recruitment that resulted in the Our results provide empirical evidence that supports smaller number of 101–200-cm trees. Perhaps habitat the endangered status of C. micronesica and highlight destruction during establishment of copra plantations in concerns about ecosystem processes that will be impacted the 20th century and as a result of military operations by the loss of this native cycad. Gestalt comparison and construction during and following World War II, of a Guam forest before and after the consequences of or the cumulative impacts of these events combined this invasion (Figure 4) reveals immeasurable changes with natural climatic events may explain these results. to biotic and abiotic characteristics. An immediate Alternatively, the extent of ecosystem destruction during research focus on changes to habitat- and ecosystem- the direct hit of several major typhoons striking Guam level processes that result from the epidemic mortality of in the past century may partly explain the anomaly. this species is warranted, as the changes in population Unfortunately, no reliable historical records are available density are of such great magnitude and speed. The that permit us to definitively establish any single cause window of opportunity to quantify these rapid changes or any series of events that influenced C. micronesica has already been compromised for this island population, demography or the subsequent recovery of this highlighting the importance of securing data prior to habitat. predicted disturbances to serve as benchmarks. Death of a portion of juvenile C. micronesica plants Compounding negative impacts visited on this and shortly after the A. yasumatsui invasion caused significant other native plant communities are anticipated with changes on allometric relations for the juvenile survivors future waves of biological invasions associated with the (Niklas & Marler 2008). Here we have extended the nascent military build-up on Guam (Marler & Moore results to include the changes in allometry for the entire 2011). Guam’s dominant native trees are ecologically population after 100% removal of the juveniles and 75% well adapted to local soils and climate, and have removal of the mature plants. The slope and range in fundamental cultural and economic significance. The data were severely decreased by the selective removal of continued loss of ecologically significant plants and C. micronesica plants by the invasive pest. habitat will continue to be disastrous for the island, 240 THOMAS E. MARLER AND JOHN H. LAWRENCE

predicts a typhoon event will occur prior to the predicted 2019 extirpation. The poor health status of remaining trees indicates that the natural resilience of this taxon to typhoon damage has been decreased by the chronic A. yasumatsui damage, a contention that has yet to be tested. In contrast to these factors that may increase mortality rate, a factor that has the potential to ward off extirpation is the purposeful introduction of a second and hopefully moreeffectivebiologicalcontrolagent.Aparasitoidwould be a better candidate for a second biological control agent, as a second predator species may be impaired by the same limitations that prevent R. lophanthae from being completely effective. DuringthetimeframeofthisGuamstudy,A.yasumatsui invaded Rota in 2007 and Palau in 2008. Our striking results serve as a benchmark for comparative purposes as the influence of this pest on the C. micronesica populations in Rota and Palau should be determined in the future.

ACKNOWLEDGEMENTS

This project was made possible in part by National Science Foundation SGER No. 0646896, USDA CSREES Project No. 2003-05495, US Forest Service Project No. 06-DG- 11052021-206, No. 09-DG-11052021-173 and No. 10- DG-11059702-095 to TEM. This material was made possible, in part, by a Cooperative Agreement from the Figure 4. General appearance of a high density Cycas micronesica habitat prior to the invasion of Aulacaspis yasumatsui (a). General appearance of United States Department of Agriculture’s Animal and the same habitat 29 mo after the initial infestation of A. yasumatsui (b). Plant Health Inspection Service (APHIS). It may not necessarily express APHIS’ views. underscoring the importance of empirical studies to inform conservation decisions. LITERATURE CITED The surviving C. micronesica population on Guam is rooted in an expiring timer to a bomb of unknown BONGERS, F., POPMA, J., MEAVE, J. & CARABIAS, J. 1988. Structure ecological impacts, with a short but uncertain amount of and composition of the lowland rain forest of “Los Tuxtlas”, Mexico. time left on the timer. If the ongoing negative population Vegetatio 74:55–88. density trajectory established over the past 4 y is CIBRIAN-JARAMILLO,´ A., DALY, A. C., BRENNER, E., DESALLE, sustained, extirpation will occur in 2019. However, an in- R. & MARLER, T. E. 2010. When North and South don’t mix: crease in biotic disturbances would likely reduce this time genetic connectivity of a recently endangered oceanic cycad, Cycas frame, an outcome that may materialize as the invasion micronesica, in Guam using EST-microsatellites. Molecular Ecology of yet another Cycas pest or as another irruption of one of 19:2364–2379. the existing pests. Abiotic disturbances that may similarly DEEVEY, E. S. 1947. Life tables for natural populations of animals. shorten the time frame include damage by a tropical Quarterly Review of Biology 22:283–314. cyclone or a severe dry season. Concerns about the threat DONALDSON, J. S. 2003. Status survey and conservation action plan of tropical cyclone damage are especially warranted for IUCN/SSC. Cycad Specialist Group, IUCN, Gland. 86 pp. Guam habitats, as the average return time for typhoons DONNEGAN, J. A., BUTLER, S. L., GRABOWIECKI, W., HISEROTE, over a 54-y monitoring period was 4.7 y (Guard et al. B. A. & LIMTIACO, D. 2004. Guam’s forest resources, 2002.Resource 1999). Endemic species such as C. micronesica cannot es- Bulletin PNW-RB-243. U.S. Department of Agriculture, Forest cape damage during the more powerful typhoons (Marler Service, Pacific Northwest Research Station, Portland. 32 pp. & Hirsh 1998). However, in the absence of subsequent ELTON, C. S. 1958. The ecology of invasions by animals and plants. disturbances a healthy C. micronesica plant is resilient Chapman Hall, London. 181 pp. to typhoon damage (Hirsh & Marler 2002). Guam has GALIANO, L., MART´INEZ-VILALTA, J. & LLORET, F. 2011. Carbon not experienced a major typhoon since the unintentional reserves and canopy defoliation determine the recovery of Scots pine introduction of A. yasumatsui in 2003, so probability 4 yr after a drought episode. New Phytologist 190:750–759. Scale insect invasion alters cycad demography 241

GROENEVELD, R. A. & MEEDEN, G. 1984. Measuring skewness and MITTERMEIER, R. A., REID, G. M., RODRIGUEZ, J. P., ROSENBERG, kurtosis. The Statistician 33:391–399. A. A., SAMWAYS, M. J., SMART, J., STEIN, B. A. & STUART, S. N. GUARD, C., HAMNETT, M. P., NEUMANN, C. J., LANDER, M. A. & 2010. The impact of conservation on the status of the world’s SIEGRIST, H. G. 1999. Typhoon vulnerability study for Guam.WERI vertebrates. Science 330:1503–1509. Technical Report 85. University of Guam. HOWARD, F. W., HAMON, A., MCLAUGHLIN, M. & WEISSLING, HEU, R. A., CHUN, M. & NAGAMINE, W. T. 1999. Sago palm scale.New T. 1999. Aulacaspis yasumatsui (Homoptera: Sternorrhyncha: Pest Advisory No. 99-01. Hawaii Dept. of Agriculture, Honolulu, ), a scale insect pest of cycads recently introduced into Hawai’i. Florida. Florida Entomologist 82:14–27. HILL, K. D. 1994. The complex (Cycadaceae) in New JOANES, D. N. & GILL, C. A. 1998. Comparing measures of sample Guinea and the Western Pacific. Australian Systematic Botany 7:543– skewness and kurtosis. The Statistician 47:183–189. 567. KEANE, R. M. & CRAWLEY, M. J. 2002. Exotic plant invasions and the HIRSH, H. & MARLER, T. 2002. Damage and recovery of Cycas enemy release hypothesis. Trends in Plant Science 17:164–170. micronesica after Typhoon Paka. Biotropica 34:598–602. KEPPEL, G. 2001. Notes on the natural history of Cycas seemannii HOFFMANN, M., HILTON-TAYLOR, C., ANGULO, A., BOHM,¨ M., (Cycadaceae). South Pacific Journal of Natural History 19:35–41. BROOKS,T.M.,BUTCHART,S.H.M.,CARPENTER,K.E.,CHANSON, LAZARO-ZERME´ NO,˜ J. M., GONZALEZ-ESPINOSA,´ M., MENDOZA, A., J., COLLEN, B., COX, N. A., DARWALL, W. R. T., DULVY, N. K., MART´INEZ-RAMOS, M. & QUINTANA-ASCENCIO, P. F. 2011. HARRISON, L. R., KATARIYA, V., POLLOCK, C. M., QUADER, S., Individual growth, reproduction and population dynamics of Dioon RICHMAN, N. I., RODRIGUES, A. S. L., TOGNELLI, M. F., VIE,´ merolae (Zamiaceae) under different leaf harvest histories in Central J.-C., AGUIAR, J. M., ALLEN, D. J., ALLEN, G. R., AMORI, G., Chiapas, Mexico. Forest Ecology and Management 261:427–439. ANANJEVA, N. B., ANDREONE, F., ANDREW, P., ORTIZ, A. L. LILLIEFORS, H. 1967. On the Kolmogorov–Smirnov test for normality A., BAILLIE, J. E. M., BALDI, R., BELL, B. D., BIJU, S. D., BIRD, with mean and variance unknown. Journal of the American Statistical J. P., BLACK-DECIMA, P., BLANC, J. J., BOLANOS,˜ F., BOLIVAR-G., Association 62:399–402. W., BURFIELD, I. J., BURTON, J. A., CAPPER, D. R., CASTRO, F., LIU, H. & STILING, P. 2006. Testing the enemy release hypothesis: a CATULLO, G., CAVANAGH, R. D., CHANNING, A., CHAO, N. L., review and meta-analysis. Biological Invasions 8:1535–1545. CHENERY, A. M., CHIOZZA, F., CLAUSNITZER, V., COLLAR, N. J., MANION, P. 1981. Tree disease concepts. Prentice Hall, Englewood Cliffs. COLLETT, L. C., COLLETTE, B. B., FERNANDEZ, C. F. C., CRAIG, 399 pp. M. T., CROSBY, M. J., CUMBERLIDGE, N., CUTTELOD, A., MARLER, T. E. & HIRSH, H. 1998. Guam’s Cycas micronesica population DEROCHER, A. E., DIESMOS, A. C., DONALDSON, J. W., ravaged by Supertyphoon Paka. HortScience 33:1116–1118. DUCKWORTH, J. W., DUTSON, G., DUTTA, S. K., EMSLIE, R. MARLER, T. E. & MOORE, A. 2010. Cryptic scale infestations on Cycas H., FARJON, A., FOWLER, S., FREYHOF, J., GARSHELIS, D. L., revoluta facilitate scale invasions. HortScience 45:837–839. GERLACH, J., GOWER, D. J., GRANT, T. D., HAMMERSON, G. MARLER, T. E. & MOORE, A. 2011. Military threats to terrestrial A., HARRIS, R. B., HEANEY, L. R., HEDGES, S. B., HERO, J.- resources not restricted to wartime: a case study from Guam. Journal M., HUGHES, B., HUSSAIN, S. A., ICOCHEA M., J., INGER, R. of Environmental Science & Engineering 5:1198–1214. F., ISHII, N., ISKANDAR, D. T., JENKINS, R. K. B., KANEKO, MARLER, T. E. & MUNIAPPAN, R. 2006. Pests of Cycas micronesica leaf, Y., KOTTELAT, M., KOVACS, K. M., KUZMIN, S. L., LA MARCA, stem, and male reproductive tissues with notes on current threat E., LAMOREUX, J. F., LAU, M. W. N., LAVILLA, E. O., LEUS, status. Micronesica 39:1–9. K., LEWISON, R. L., LICHTENSTEIN, G., LIVINGSTONE, S. R., MARLER, T. E., LINDSTROM,¨ A. & FISHER, J. B. 2010. Stem tissue LUKOSCHEK, V., MALLON, D. P., MCGOWAN, P. J. K., MCIVOR, dimensions correlate with ease of horticultural management for six A., MOEHLMAN, P. D., MOLUR, S., ALONSO, A. M., MUSICK, Cycas species. HortScience 45:1293–1296. J. A., NOWELL, K., NUSSBAUM, R. A., OLECH, W., ORLOV, N. L., McDOWELL, N., POCKMAN, W. T., ALLEN, C. D., BRESHEARS, D. D., PAPENFUSS, T. J., PARRA-OLEA, G., PERRIN, W. F., POLIDORO, COBB, N., KOLB, T., PLAUT, J., SPERRY, J., WEST, A., WILLIAMS, B. A., POURKAZEMI, M., RACEY, P. A., RAGLE, J. S., RAM, M., D. G. & YEPEZ, E. A. 2008. Mechanisms of plant survival and RATHBUN, G., REYNOLDS, R. P., RHODIN, A. G. J., RICHARDS, S. mortality during drought: why do some plants survive while others J., RODR´IGUEZ, L. O., RON, S. R., RONDININI, C., RYLANDS, A. succumb to drought? New Phytologist 178:719–739. B., DE MITCHESON, Y. S., SANCIANGCO, J. C., SANDERS, K. L., MEYER, J.-Y. & BUTAUD, J.-F. 2009. The impacts of rats on the SANTOS-BARRERA, G., SCHIPPER, J., SELF-SULLIVAN, C., SHI, Y., endangered native flora of French Polynesia (Pacific Islands): drivers SHOEMAKER, A., SHORT, F. T., SILLERO-ZUBIRI, C., SILVANO, D. of plant extinction or coup de graceˆ species? Biological Invasions L., SMITH, K. G., SMITH, A. T., SNOEKS, J., STATTERSFIELD, A. 11:1569–1585. J., SYMES, A. J., TABER, A. B., TALUKDAR, B. K., TEMPLE, H. MOIR, M. L., VESK, P. A., BRENNAN, K. E. C., KEITH, D. A., MCCARTHY, J., TIMMINS, R., TOBIAS, J. A., TSYTSULINA, K., TWEDDLE, D., M. A. & HUGHES, L. 2011. Identifying and managing threatened UBEDA, C., VALENTI, S. V., VAN DIJK, P. P., VEIGA, L. M., VELOSO, invertebrates through assessment of coextinction risk. Conservation A., WEGE, D. C., WILKINSON, M., WILLIAMSON, W. A., XIE, F., Biology 25:787–796. YOUNG, G. E., AKC¸AKAYA, H. R., BENNUN, L., BLACKBURN, T. M., MOORE, A., MARLER, T., MILLER, R. H. & MUNIAPPAN, R. 2005. BOITANI, L., DUBLIN, H. T., DA FONSECA, G. A. B., GASCON, C., Biological control of cycad aulacaspis scale on Guam. The Cycad LACHER, T. E., MACE, G. M., MAINKA, S. A., MCNEELY, J. A., Newsletter 28:6–8. 242 THOMAS E. MARLER AND JOHN H. LAWRENCE

MUELLER-DOMBOIS,D.1986.Perspectivesforanetiologyofstand-level HERNANDEZ´ -JONAPA,´ R. & VILLALOBOS-MENDEZ,´ S. M. 2006. dieback. Annual Review of Ecology and Systematics 17:221–243. Demography of the cycad Ceratozamia mirandae (Zamiaceae) under MUELLER-DOMBOIS, D. 1987. Natural dieback in forests. Bioscience disturbed and undisturbed conditions in a biosphere reserve of 37:575–583. Mexico. Plant Ecology 187:97–108. NEGRON-ORTIZ, V. & BRECKON, G. J. 1989. Population structure in PYKE, D. A. & THOMPSON, J. N. 1986. Statistical analysis of survival Zamia debilis (Zamiaceae) I. Size classes, leaf phenology, and leaf and removal rate experiments. Ecology 67:240–245. turnover. American Journal of Botany 76:891–900. SCHEMSKE, D., HUSBAND, B., RUCKELSHAUS, M., GOODWILLIE, NIKLAS, K. J. & MARLER, T. E. 2008. Sex and population differences C., PARKER, I. & BISHOP, J. 1994. Evaluating approaches to in the allometry of an endangered cycad species, Cycas micronesica the conservation of rare and endangered plants. Ecology 75:584– (Cycadales). International Journal of Plant Sciences 169:659–665. 606. NORGHAUER, J. M., MARTIN, A. R., MYCROFT, E. E., JAMES, A. TAKAGI, S. 1977. A new species of Aulacaspis associated with a cycad & THOMAS, S. C. 2011. Island invasion by a threatened tree in Thailand (Homoptera: Cocoidea). Insecta Matsumurana New Series species: evidence for natural enemy release of mahogany (Swietenia 11:63–72. macrophylla) on Dominica, Lesser Antilles. PLoS ONE 6(4): e18790. TANG, W., YANG, S.-L. & VATCHARAKORN, P. 1997. Cycads of NORSTOG, K. J. & NICHOLLS, T. J. 1997. The biology of the cycads. Cornell Thailand. Nong Nooch Tropical Garden and the Cycad Conservation University Press, New York. 363 pp. Company. Bangkok, Thailand. 34 pp. PEREZ-FARRERA,´ M. A. & VOVIDES, A. P. 2004a. Spatial distribution, TERRY, I. & MARLER, T. 2005. Paradise lost? Tipping the scales against population structure, and fecundity of Ceratozamia matudai Lundell Guam’s Cycas micronesica. The Cycad Newsletter 28:21–23. (Zamiaceae) in El Triunfo Biosphere Reserve, Chiapas, Mexico. The VOVIDES, A. P. 1990. Spatial distribution, survival, and fecundity of Botanical Review 70:299–311. Dioon edule (Zamiaceae) in a tropical deciduous forest in Veracruz, PEREZ-FARRERA,´ M. A. & VOVIDES, A. P. 2004b. Ecology of cycads in Mexico, with notes on its habitat. American Journal of Botany Southern Mexico. Pp. 112–120 in Lindstrom,¨ A. J. (ed.). Proceedings 77:1532–1543. of the Sixth International Conference on Cycad Biology. Nong Nooch WILCOXON, F. 1945. Individual comparisons by ranking methods. Tropical Garden, Thailand. 216 pp. Biometrics Bulletin 1:80–83. PEREZ-FARRERA,´ M. A., VOVIDES, A. P., OCTAVIO-AGUILAR, YOUNG, F. J. 1988. Soil survey of Territory of Guam. United States P., GONZALEZ-ASTORGA,´ J., DE LA CRUZ-RODR´IGUEZ, J., Department of Agriculture Soil Conservation Service. 166 pp.