Impact of an Invasive Predatory Lizard on the Endangered Hawaiian Tree Snail Achatinella Mustelina: a Threat Assessment

Home , Achatinella mustelina, Platydemus manokwari, Trioceros

Vol. 24: 115–123, 2014 ENDANGERED SPECIES RESEARCH Published online May 8 doi: 10.3354/esr00589 Endang Species Res

Impact of an invasive predatory lizard on the endangered Hawaiian tree snail Achatinella mustelina: a threat assessment

Luciano M. Chiaverano, Brenden S. Holland*

Pacific Biosciences Research Center, University of Hawaii at M noa, 1993 East-West Road, Honolulu, Hawaii 96822, USA

ABSTRACT: All 41 species of the Hawaiian tree snail genus Achatinella were listed as endangered in the early 1980s, primarily due to predation by invasive species. Today, only 9 species are estimated to remain. The recent discovery of A. mustelina in gut contents of Jackson’s chameleons Trioceros jacksonii xantholophus, a predatory arboreal African lizard introduced to Hawaii (USA) in 1972, was the first documentation of environmental impact by any reptile on native Hawaiian taxa. In this study, we assessed the predatory threat posed by chameleons to A. mustelina by (1) determining shell digestion and passage rates in daily fed and fasting individuals and (2) using these data to estimate time of ingestion of snails found in guts of field-collected chameleons. Our results indicate that daily food intake increases passage rate. Well fed chameleons passed largely intact A. mustelina shells in ~4 to 5 d, whereas starved individuals retained shells in their stomachs until complete digestion in ~8 d. Shell mass was digested at ~12.5% d−1. In total, 8 A. mustelina were found in 45 field-collected lizards, and estimated time of ingestion, based on shell condition, was ~3 to 4 d prior to capture in all cases. Given a predator density of 45 per 2000 m2 (0.5 acre) and assuming a feeding frequency of 8 snails every 3 to 4 d, we conservatively estimate that 45 chameleons could consume 730 to 974 snails yr−1, per 2000 m2. Given the low growth and reproductive rates of Hawaiian tree snails, this level of impact is probably unsustain- able even over the short term, and immediate control action is warranted.

KEY WORDS: Jackson’s chameleon · Trioceros jacksonii xantholopus · Endemic invertebrate conservation · Resource management · Invasive species control

Resale or republication not permitted without written consent of the publisher

INTRODUCTION 2009). Prior to the intentional release of the predatory snail Euglandina rosea from its native southeastern The terrestrial ecosystems of the Hawaiian Islands USA in 1955 by the Board of Agriculture (Mead 1979, (USA) rank among the most isolated and threatened Cowie 2001, Holland et al. 2008), the Oahu endemic on Earth (Dobson et al. 1997). Geographic isolation genus Achatinella represented a highly diverse combined with habitat diversity has resulted in extra - group, comprising 41 described species (USFWS ordinary levels of endemism: approximately 95% of 1993). Due primarily to predatory activity of invasive the native terrestrial Hawaiian plant and animal spe- taxa, including the black rat Rattus rattus and the cies are endemic (Carlquist 1970). In recent decades, flatworm Platydemus manokwari, the entire genus the Hawaiian flora and fauna have also become rec- Achatinella was placed on the US Fish and Wildlife ognized for high extinction levels across numerous Service Endangered Species list in the early 1980s groups, including land snails, with losses of up to (USFWS 1981). By 1993, 21 species had gone extinct, 90% of the described species (Cowie 2001, Holland and 18 of the 20 remaining species persisted only

*Corresponding author: [email protected] © Inter-Research 2014 · www.int-res.com 116 Endang Species Res 24: 115–123, 2014

in restricted forest patches, in very low numbers (US- from the guts of field-collected chameleons were par- FWS 1993). Recent surveys continue to reveal a dete- tially digested, although no method was available to riorating scenario (Holland & Hadfield 2007, D. Sischo estimate time of ingestion. Information regarding pers. comm.), and currently we estimate that only 9 digestion rates and relationships of passage time to Achatinella species remain (B. Holland pers. obs.). feeding regime of tree snail shells is critical in terms Primary threats to Hawaiian tree snails come from of gaining an understanding and assessing the sever- invasive predators, including rats (Rattus spp.), the ity of the threat posed by Jackson’s chameleons. For wolf snail Euglandina rosea, the flatworm P. manok- example, if shells and fragments of crushed shells wari, and, based on a recent revelation, Jackson’s require relatively long periods of time to digest or chameleon Trioceros jacksonii xantholophus (Hol- pass, i.e. multiple weeks or months, then due to their land et al. 2010). Jackson’s chameleons are native to shear bulk, the shells might impede further feeding, montane habitats of Kenya and Tanzania (Spawls & thus limiting or reducing potential impact on endan- Rotich 1997, Necas 1999). In 1972, several dozen gered tree snail fauna by chameleons. However, if individuals of this subspecies, the largest of 3, were the rate of digestion/passage is relatively rapid, i.e. imported and released on the Hawaiian island of only a few days, this suggests a potentially elevated Oahu via the pet trade (McKeown 1991, Waring impact of chameleon predation on rare snails, as a 1997). In addition to this well-documented release, it given population density of lizards would tend to has been reported that hundreds of chameleons born consume more snails in a given amount of time. in pet stores loca ted on the west coast of the USA Data on shell digestion rates may provide im portant were also imported and released at numerous locali- information in determining the time of consumption, ties in Hawaii by a number of individuals in the early enhance our ability to assess predatory threat, and 1970s, with the intention of establishing a stock of allow inference regarding movements and dispersal animals to be exported for the pet trade (Kraus et al. of chameleons (Chiaverano et al. 2014). In the present 2012). As a result of subsequent decades of human- study, we carried out laboratory experiments to eval- mediated transport and distribution of specimens uate digestion and passage of A. mus telina shells by among islands, as well as ongoing releases into the Jackson’s chameleons under different feeding sce- wild, T. j. xantholophus spread rapidly throughout narios. We then applied these data to A. mustelina re- most of the state, becoming established in multiple covered from the guts of Jackson’s chameleons col- areas on all major Hawaiian Islands (with the possi- lected in the field to estimate time of ingestion and to ble exception of Kauai) by the mid 1990s (McKeown more accurately assess pre datory impact and under- 1991, Kraus et al. 2012). Although chameleons have stand threats posed by Jackson’s chameleons to en- not been legally expor ted from Kenya to Hawaii dangered Hawaiian tree snails. since 1981 (McKeown 1996), and regulations took effect in the late 1990s aimed at thwarting transport among islands and prohibiting conveyance out of the MATERIALS AND METHODS state, Jackson’s chameleons are still legally sold in pet stores (authors’ pers. obs.). Specimen collection for laboratory shell Despite the fact that Jackson’s chameleons are digestion trials predators, and are known to feed on a variety of invertebrate and small vertebrate species, they were Jackson’s chameleons were collected by hand at not considered a threat to the native Hawaiian fauna night (collection permit: ODF-092611R) from the (McKeown 1996) until recently, when it was demon- Tantalus area (Fig. 1) for feeding trials. Individuals strated for the first time that when present in pristine were brought to the Hawaiian Tree Snail Conserva- native Hawaiian forest, chameleons prey on numer- tion Lab (University of Hawaii, Manoa) and main- ous species of endemic arthropods and gastropods, tained under uniform conditions in individual plastic including the Critically Endangered (IUCN 2013) cages (210 × 120 × 120 mm) with tree branches for Oahu tree snail Achatinella mustelina (Holland et al. climbing and perching. Cages had perforated lids 2010). Since then, Jackson’s chameleons have been and were modified by removal of a rectangular por- considered a species of conservation concern in the tion of the floor and replacing it with 1 mm screen, Hawaiian Islands, and studies of their impacts on glued into the bottom cage floor to facilitate air flow native species have recently begun (Kraus & Preston and water drainage. Cages were hand-misted with 2012, Kraus et al. 2012). In a recent study by Holland tap water at regular intervals and kept at constant et al. (2010), tree snail shells that were recovered temperature (24°C) in the lab. Chiaverano & Holland: Impact of chameleons on endangered snails 117

Fig. 2. Force-feeding empty Achatinella mustelina tree snail shells to Jackson’s chameleons Trioceros jacksonii xantholophus for laboratory digestion and passage rate trials. Sexual dimorphism of this invasive lizard is evident in the figure; the chameleon on the left is a female (lacking horns), the one on the right is a male (1 rostral, 2 ocular horns)

swallow snail shells whole (B. Holland et al. unpub- lished data); therefore, we force-fed chameleons intact shells. Water was used to initiate the swallow- Fig. 1. Main Hawaiian Islands, USA (top), with detailed view ing reflex. Once the shell was ingested, time was of Oahu (bottom), showing collection sites for Jackson’s chameleons Trioceros jacksonii xantholophus used for the recorded as time 0. laboratory feeding trials (TA: Tantalus), and for the gut In order to test different feeding scenarios, chame - content analysis (PH: Puu Hapapa) leons were then randomly divided into 2 groups, each corresponding to a different feeding treatment. Group I was fed live cockroaches Nauphoetacinerea Experimental approach: feeding trials sp., Pycnoscelusindicus sp., and Diploptera punctata (length: 1−2 cm), twice per day, until satiated, while Prior to the experiment, chameleons were starved Group II was not fed during the trial. Water was pro- for 36 h, each individual was weighed, measured for vided twice per day in both groups. To assess timing snout-to-vent length (SVL) using a caliper, and of the shell digestion and passage process, 3 to 4 sexed. Feeding trials consisted of force-feeding each replicate chameleons from each group were eutha- chameleon 1 fresh Achatinella mustelina shell (Fig. 2), nized at the following intervals: 3, 4, 5, 7, and 8 d. which had been collected as empty ground shells Chameleons that did not pass the shell were dis- from the northern Waianae Mountains (Puu Hapapa; sected and the position of the shell or shell fragments Fig. 1) (USFWS Endangered Species permit no. (PSH) within the digestive tract (1 = stomach, 2 = TE043638-8). This site is within the natural range proximal half of intestine, 3 = distal half of intestine, and current distribution of A. mustelina as well as a 4 = passed) was recorded. Shells and or shell frag- number of other rare endemic snails, arthropods, ments were then washed with 99% ethanol and air- birds and plants. Since chameleons and tree snail dried, and the final weight was recorded (FWE). The shells varied in size, shells were selected and stan- percentage of digested shell mass (DSM) was deter- dardized with initial weight (IWE) of 0.5% of total mined per chameleon using the following formula: chameleon weight. Force-feeding was conducted by DSM: (IWE − FWE) × 100 / IWE (1) manually opening the mouth of the restrained lizard and introducing the shell, orientated with the apex Shells and shell fragments were digitally photo - downward, into the mouth, and then gently pushing graphed, and images were used to separate shells the shell into the esophagus with a narrow (0.5 cm into 5 discrete categories based on condition (CON): diameter) probe (Fig. 2). Although Jackson’s chame - (1) intact shell with pitted and patchy, somewhat leons masticate most arthropod prey, they tend to degraded periostracum (~22% digested mass), (2) 118 Endang Species Res 24: 115–123, 2014

digested apex, extensive pitting, some perforations Statistical analyses through the shell (~44% digested mass), (3) only the main body whorl remaining (~60% digested mass), In the laboratory shell digestion trials, differences (4) only shell fragments remain (~82% digested in CON and DSM between fed and fasting cha - mass), and (5) shell completely digested. meleons within a given time interval were directly compared using Mann-Whitney and t-tests, respectively. Both CON and DSM were then compared over Specimen collections for gut contents 24 h time intervals using a Friedman 1-way ANOVA (non-para metric ordinal data) and a regular 1-way Jackson’s chameleons were collected in the field ANOVA, respectively. Tukey tests were used for by hand from a 2000 m2 area that lacks public access, pairwise comparisons. Pearson correlations were located in the central Waianae Mountain Range (Puu used to assess the significance of the relationship Hapapa, Fig. 1) at about 800 m elevation. This site between both CON and DSM over time. In addition, represents a native rain forest characterized by a second ANOVA was applied to compare DSM dense, high canopy (up to 90% canopy cover), domi- among categories of shell condition, and post hoc nated by endemic trees and shrubs such as Metro - pairwise comparisons were made using Tukey tests. sideros polymorpha, Urera glabra, Pisonia spp., Ilex spp., Cyanea sp., and Perrotetia sandwi censis, and the tallest canopy exceeds 15 m in height. A long- RESULTS term concerted botanical restoration effort has been underway here since 2009, including out-planting Laboratory component: shell digestion trials many rare plant species, weed removal, and invasive rat, slug and flatworm control (Oahu Army Natural We used 37 adult Jackson’s chameleons in labora- Resource Program pers. comm.). This collection local- tory feeding experiments: 15 females and 22 males. ity is the site of a recently completed invasive preda- There was no significant difference in SVL between tor-proof fence, engineered and built to protect en - fed (Group I) and fasting (Group II) chameleons dangered tree snails from rats, predatory gastropods, (t1,35 = −1.3, p = 0.12) used in this trial. Twelve flatworms, and chameleons. Chameleon surveys and chameleons were used in Group I, while 25 individu- collection occurred weekly from July 2012, when the als were included in Group II. For Group I, treatment barrier was closed, until February 2013, when the was carried out for 5 d (see details below), and for last chameleon was removed from inside the barrier. Group II, 8 d. Surveys and conservation activities within the preda- Our feeding trials indicated that food intake signif- tor-proof structure have continued on a weekly basis, icantly shortened the passage time of the shell and no chameleons have been observed since then. through the digestive tract of Jackson’s chameleons. Once the barrier was closed, chameleons were By Day 3, the shell had moved to the proximal por- unable to pass in either direction, thus the number of tion of the intestine in 50% of fed (Group I) individu- chameleons within the perimeter during the collec- als (i.e. Position 2) and in the other 50%, the shell was tion period did not increase, with the only potential found in the distal half of the intestine (i.e. Position 3; replacement of animals collected from within being Fig. 3). On Day 4, shells in 33% of individuals of due to births. Group I were located in the distal portion of the intes- All chameleons collected were brought to the lab tine and 67% of the shells had passed (i.e. Position 4), where they were euthanized, dissected, and their gut while 100% of the shells had passed by Day 5 contents sorted and examined under dissecting sco - (Fig. 3). By contrast, in the fasting treatment (Group pes. All snail shells found were separated, identified, II), 100% of the shells remained in the stomach of cleaned, dried, and digitally photographed. Photo- chameleons (i.e. Position 1) for the duration of the graphs were used to assess shell condition (CON) trial (Fig. 3). and compared to condition of shells used in labora- The process of shell digestion was not affected by tory digestion trials to assign degradation category feeding treatment (i.e. fed daily versus fasting). On and estimate consumption time. All aspects of cap- Day 3, CON and DSM did not vary significantly ture, handling, and euthanasia of Jackson’s chame - between Groups I and II (CON: Mann-Whitney, U = leons were conducted in accordance with protocols 6, Z = 0.98; DSM: t-test: t = 1.1; p > 0.05 for all tests). of the University of Hawaii Institutional Animal Care Due to the fact that shells had passed the stomach by and Use Committee (IACUC). Day 3 in fed individuals, CON and the percentage of Chiaverano & Holland: Impact of chameleons on endangered snails 119

Fig. 3. Position of endangered tree snail shells in guts of daily fed (fed; n = 15) and fasting (non-fed; n = 12) Jackson’s cha me leons Trioceros jacksonii xantholophus during laboratory digestion trials, 3, 4, and 5 d after ingestion. Shells remained in the stomach of all 12 non-fed individuals for the duration of the experiment (8 d, not shown), whereas after Day 5, all shells in fed chameleons were passed

DSM remained constant until the end of the trial (CON: r = 0.01; χ2 = 2.1. DSM: r = 0.02; ANOVA:

F1,15 = 1.43; p > 0.05 for all tests; Fig. 4). Contrarily, since shells remained in the stomach of starved chameleons until complete digestion (see above), both CON and DSM increased significantly over time (CON: r = 0.82; DSM: r = 0.84, p < 0.05) in these indi- Fig. 4. (A) Shell condition and (B) percentage of digested viduals, with significant differences detected among mass of Achatinella mustelina shells over time in daily fed d s χ2 ( ) and fasting ( ) Jackson’s chame leons Trioceros jacksonii days (CON: = 34.4, p < 0.01; DSM: F1,22 = 22.43, p < xantholophus. Points show the mean values, whiskers 0.01; Fig. 4). CON varied significantly among all indicate SE. Different lower case letters indicate significant days, with the exception of the period between Days differences in shell condition and percentage of digested 5 and 7 (Fig. 4A). A similar pattern was observed for shell mass among days in starved individuals (ANOVA, DSM, with no significant differences detected be - Tukey test, p < 0.01). Asterisks indicate no significant differences in both parameters over time for fed specimens tween Days 7 and 8 (Fig. 4B). The percentage of DSM (ANOVA, p > 0.05) varied significantly among CON Categories 1 to 4 −1 (F3,25 = 50.7, p < 0.001; Fig. 4), increasing by ~20% d after Day 3. In 8 d, 66% of shells fed to individuals of individuals (nos. 59, 73, and 74) had a single shell, Group II, independent of size, were completely while individual no. 71 had 5 shells, giving a total of digested (Category 5) and could not be detected, 8 shells. The position of the shell within the gut, shell whereas 34% were found in the form of small frag- size, and shell condition per individual is summa- ments (Category 4; Fig. 5). Thus, the average rate at rized in Table 1. Only 1 shell was found in the proxi- which Achatinella muste lina shell mass was digested mal portion of the intestine (Position 2), which dis- by Jackson’s chame leons in the laboratory was played a partially digested apex, extensive pitting, 12.5% d−1. and holes passing through the shell, typical of shell Category 2. The rest of the shells had only the main body whorl remaining (Category 3), and were found Gut content analysis of field-collected specimens in the stomach (Position 1) of the chameleons (Fig. 6, Table 1). Interestingly, of the 5 A. mustelina shells Of 45 Jackson’s chameleons collected from Puu found in the stomach of chameleon no. 71, 1 was Cat- Hapapa (Fig. 1), 4 individuals (8.9%) had at least 1 egory 1, 1 was Category 3, and 3 were Category 2 Achatinella mustelina shell present in the gut. Three (Fig. 6, Table 1). 120 Endang Species Res 24: 115–123, 2014

Table 1. Achatinella mustelina. Summary of shells found in the stomachs of 45 Jackson’s chameleons Trioceros jacksonii xantholophus collected from the Waianae Mountain Range (Puu Hapapa, see Fig. 1), on the Hawaiian Island of Oahu. Chameleon no. 71 had 5 shells in its stomach (labeled A to E here). For a description of the shell condition categories, see ‘Materials and methods’

ID No. of Shell Position Shell shells dimensions within condition in gut (l × w, mm) gut category

59 1 6.58 × 6.01 Stomach 3 71 5 A: 10.09 × 8.98 Stomach 1, 2, 3 B: 6.72 × 6.33 C: 6.39 × 4.01 D: 5.47 × 4.29 E: 7.82 × 5.94 73 1 7.28 × 4.65 Proximal 2 intestine 74 1 4.17 × 3.86 Stomach 3

Fig. 5. Endangered Oahu tree snail Achatinella mustelina shells at different stages of digestion, expressed as categories of shell condition (CON: 1−4, see ‘Experimental approach: feeding trials’ in the ‘Materials and methods’ for a detailed description of each category; see Fig. 3 for shell condition versus time). Category 1: ~22% digested mass. Category 2: ~44% digested mass. Category 3: ~60% digested mass. Category 4: ~82% digested mass. Category 5 (not shown): shell completely digested

DISCUSSION

In light of the recent discovery that Jackson’s chameleons prey on native island species (Holland et al. 2010), this investigation was designed as a follow- up study aimed at gaining insights to further assess their potential predatory threat, specifically on the endangered Hawaiian tree snail Achatinella mustelina. Thus, this study represents the first attempt to quantitatively evaluate impact and assess the threat by predation on an endangered tree snail by an invasive lizard. Our feeding trials allowed us to determine the passage time and digestion rates of A. mus telina shells under different feeding scenarios, and to de- velop a sequence of objective, discrete categories of shell condition that can be assigned to estimate resi- dence time in the gut of this species. These re sults Fig. 6. Achatinella mustelina. Shells from gut contents of were useful not only in estimating the time of con- field-collected Jackson’s chameleons Trioceros jacksonii xantholophus, collected at Puu Hapapa in the Waianae sumption of snails consumed by chameleons collected Mountains, Oahu (see Fig. 1). Each image shows 2 views for from the wild, but also for scaling up and quantifying each of 8 shells found in the stomachs of 45 chameleons ex- the potential impact of Jackson’s chame leons on en- amined during this study (see Table 1 for dimensions and as- dangered Hawaiian tree snails in a given area. signed shell condition category). Shell 1: Category 3 found in stomach of individual no. 59. Shell 3: Category 3 found in Passage time of A. mustelina shells through the gut stomach of individual no. 74. The remaining shells were of Jackson’s chameleons was dramatically shortened found in the stomach of individual no. 71. Shell 2: Category by the amount of food supply (i.e. fed versus fasting). 3, shell 6: Category 1, shells 4, 7, and 8: Category 2 Chiaverano & Holland: Impact of chameleons on endangered snails 121

Thus, the daily ingestion of prey items (i.e. twice per shells through the gut of individuals from this area. day) besides the shell apparently facilitated the pas- Further support is given by the fact that 1 chameleon sage of stomach contents into the intestine, thus (no. 73) had an A. mustelina shell located within the reducing the time shells remained in the stomach, proximal portion of the intestine exhibiting a state where digestion takes place. In all fed Jackson’s of digestion characteristic of Category 2 (Fig. 5, chameleons, shells had already moved out of the Table 1), which was achieved in ~3 d during our lab- stomach by Day 3 and were passed in 5 d, displaying oratory trials (Fig. 4A). Therefore, the actual impact a Category 2 digestion condition. In contrast, in fast- of Jackson’s chameleons on endangered tree snails ing individuals, shells remained in the stomach for may be far more substantial than this low number the duration of the experiment, achieving relatively would suggest, due not only to the relatively short advanced states of digestion (Categories 3 and 4), passage time these lizards exhibit when well fed, but including complete digestion (Category 5), in ~8 d also to the relatively short time required to com- (Fig. 2). These results are in concordance with those pletely dissolve an A. mustelina shell (~8 d). obtained by Waldschmidt et al. (1986) on the lizard We estimated the density of Jackson’s chameleons Uta stansburiana, with daily fed individuals able to in this area (Puu Hapapa) by extrapolating from the pass food within ~5 d, but taking twice as long when 45 chameleons collected per 2000 m2 without re - food ration was interrupted. It is well documented placement (see ‘Materials and methods’), giving a that food assimilation and passage time in lizards is value of 1 chameleon per 44 m2. We might assume affected by body temperature and body size (SVL; that at any given time, about 9% of these chameleons Van Damme et al. 1991, Xiang et al. 1996); however, prey on snails, or an instructive way to view the data in the present study, all specimens were housed at is that every 3 to 4 d, 45 chameleons can potentially the same controlled temperature, and SVL did not consume 8 snails. Extrapolating these figures to a vary significantly between fed and fasting individu- period of 1 yr, for every 2000 m2, chameleons at this als. In other words, the inference here is that the density could consume 730 to 974 tree snails. How- ingestion of prey items can have an indirect effect on ever, it should be noted that this estimate may be shell digestion, reducing the time of exposure to the conservative, because the density of chameleons low pH environment of the stomach, and suggests could be substantially higher in the future, or in other that in areas where food is readily available, such as localities. For example, Kraus et al. (2012) reported a the Puu Hapapa collection site in this study (Fig. 1), density of 173 chameleons per 2000 m2 on the island shells will likely stay in the gut of Jackson’s cha - of Maui. At this population density, we estimate a meleons for significantly shorter periods than in loss rate of 2806 to 3744 snails yr−1, per 0.2 ha. The areas/ patches where prey is scarce. estimated feeding frequency of 72 h may also be conservative, since our gut content analyses revealed that Jackson’s chameleons have the potential to feed Threat assessment on A. mustelina at an even higher frequency, snail density permitting. For instance, cha me leon no. 71 Despite a recent effort to model future range expan- had 5 shells in its stomach in various states of de- sion for Jackson’s chameleons (Rödder et al. 2011), gradation, corresponding to Categories 1, 2, and 3. detailed geographic distribution and population den- According to our laboratory experiments, shells in sities of this species have yet to be documented. The the stomach went from Category 1 to 3 in ~48 h, sug- population at Puu Hapapa is evidently established, gesting that this individual likely ingested A. mus - healthy, well fed, and clearly reproductive (authors’ telina on 3 consecutive days. pers. obs.); it may therefore represent a useful model Extrapolation of this estimated rate of loss of for other island locations that share basic habitat endangered snails across the landscape is alarming, characteristics. The observed incidence of endan- and this projection is devastating when considered in gered A. mustelina tree snail predation in field- light of snail species on the brink of extinction, with collected lizards might at first appear relatively low, remarkably slow natural growth and reproductive since only 4 of the 45 (8.9%) field- collected cha me - rates, and extremely protracted replacement poten- leons had A. mustelina shells in their guts. However, tial. A. mustelina suffers 98% natural mortality dur- these chameleons were uniformly well fed and their ing the first 4 yr of life, reaching maturity in 5 to 7 yr, stomach contents revealed diverse and abundant and must reproduce for 12 yr to achieve replacement prey items (authors’ unpubl. data), suggesting condi- (Hadfield & Miller 1989). Another factor that renders tions leading to a relatively short passage time of A. mustelina particularly vulnerable to chameleons is 122 Endang Species Res 24: 115–123, 2014

that fact that both species are exclusively arboreal source management in all regions where chameleons (USFWS 1993, Losos et al. 1993) and thus they either have become established or have the potential occupy shared habitat. Jackson’s chame leons are to be released in the future. cruise foragers capable of actively searching for food (Hagey et al. 2010), which translates into a potential Acknowledgements. We thank the staff of the Oahu Army Natural Resources Program, especially V. Costello, S. Joe, increased possibility of an encounter with the se - J. Tanino, J. Rohrer, and K. Kawelo, for their commitment to dentary tree snails. Thus far, no specific or effective Hawaiian tree snail conservation and their diligence and management strategy has been developed for Jack- dedication to invasive predator control and habitat stabiliza- son’s chameleons in Hawaii. The only control meas- tion. We thank members of the Hawaiian Tree Snail Conser- ure available is manual search and capture, and vation Lab, including M. J. Wright and T. A. Hooper Smith, B. Dietrich, C. Artiles, and C. Howard for lab assistance. This effective protection of native resources includes project was supported by US Army Cooperative Agree- installation of predator exclosure structures specifi- ment — W9126G-11-2-0066. All handling and care of cha - cally designed to prevent entry by these efficient meleons was conducted in compliance with IACUC protocol climbers. no. 1105-2. Although there are currently no restrictions on pur- LITERATURE CITED chase, capture, or keeping Jackson’s chameleons as pets in Hawaii, release is illegal. The public, how- Carlquist SJ (1970) Hawaii: a natural history: geology, ever, is generally unaware of this (authors’ pers. climate, native flora and fauna above the shoreline. obs.), and there have been no prosecutions because Natural History Press, Garden City, NY of the lack of enforcement. Jackson’s chameleons Carpenter AI, Rowcliffe JM, Watkinson AR (2004) The dy na - mics of the global trade in chameleons. Biol Conserv 120: make challenging pets, due primarily to the diversity 291−301 of prey species required as food, and therefore many Chiaverano LM, Wright MJ, Holland BS (2014) Home on the pet chame leons are released (Kraus et al. 2012). This range: Movement behavior is habitat-dependent in inva- has likely led to ongoing range expansions in the sive Jackson’s chameleons in Hawaii. J Herpetol 48 (in press) Hawaiian Islands, where chameleons are available Cowie RH (2001) Invertebrate invasions on Pacific Islands via the pet trade. Due to the combination of a large and the replacement of unique native faunas: a synthesis and possibly expanding range (Rödder et al. 2011), of the land and freshwater snails* Contribution no. 2001- predation on diverse taxa (Holland et al. 2010, Kraus 001 of Bishop Museum’s Pacific Biological survey. Biol Invasions 3:119–136 & Preston, 2012, Kraus et al. 2012), and high popula- Dobson AP, Rodriguez JP, Roberts WM, Wilcove DS (1996) tion densities, this species poses a serious threat to Geographic distribution of endangered species in the Hawaiian fauna. As such, it is high time that the state United States. Science 275: 550−553 legislature acknowledges this problem and alters Hadfield MG, Miller SE (1989) Demographic studies on regulations accordingly, such that chameleons are no Hawaii’s endangered tree snails: Partulina proxima. Pac Sci 43:1−16 longer available via the pet trade. Hagey TJ, Losos JB, Harmon LJ (2010) Cruise foraging of invasive chameleon (Chamaeleo jacksonii xantholophus) in Hawaii. Breviora 519: 1−7 Regional implications outside of Hawaii Holland BS (2009) Island flora and fauna: snails. In: Gillespie R, Clague DA (eds) The encyclopedia of islands. Science Publishing Group, University of California Press, Los Jackson’s chameleons represent one of the most Angeles, CA, p 537−542 popular species in the pet trade on an international Holland BS, Hadfield MG (2007) Molecular systematics of scale. Between 1977 and 1982, more than 80 000 the endangered O‘ahu tree snail Achatinella mustelina Jackson’s chameleons were exported from Kenya, (Mighels 1845): synonymization of subspecies and esti- mation of gene flow between chiral morphs. Pac Sci 61: with most of the specimens destined for the US mar- 53−66 ket, as well as, for example, Japan, several European Holland BS, Christensen CC, Hayes KA, Cowie RH (2008) countries, and South Africa (Carpenter et al. 2004). In Biocontrol in Hawaii: a response to Messing (2007). Proc addition, release of chameleons into the wild has Hawaiian Entomol Soc 40: 81−83 Holland BS, Montgomery SL, Costello V (2010) A reptilian been reported at a number of other localities, includ- smoking gun: first record of invasive Jackson’s chame - ing the Marshall Islands, Morro Bay (California), leon (Chamaeleo jacksonii) predation on native Hawai- South Africa, and the Florida Keys. Thus, this preda- ian species. Biodivers Conserv 19: 1437−1441 tor could pose a serious threat to native fauna glob- IUCN (2013) IUCN Red List of Threatened Species. Version 2013.2. www.iucnredlist.org ally, and particularly to threatened or endangered Kraus F, Preston D (2012) Diet of the invasive lizard Cha - taxa in tropical and subtropical ecosystems. Impact maeleo jacksonii (Squamata: Chamaeleonidae) at a wet- assessment data therefore have relevance for re - forest site in Hawaii. Pac Sci 66: 397−404 Chiaverano & Holland: Impact of chameleons on endangered snails 123

Kraus F, Medeiros A, Preston D, Jarnevich CS, Rodda GH USFWS (United States Fish and Wildlife Service) (1981) (2012) Diet and conservation implications of an invasive Endangered and threatened wildlife and plants; listing chameleon, Chamaeleo jacksonii (Squamata: Chamae - the Hawaiian (Oahu) tree snails of the genus Achatinella leonidae) in Hawaii. Biol Invasions 14: 579−593 as endangered species. Fed Reg 46:3178–3182 Losos JB, Walton BM, Bennett AF (1993) Trade-offs be tween USFWS (United States Fish and Wildlife Service) (1993) sprinting and clinging ability in Kenyan cha me leons. Recovery Plan: Oahu tree snails of the genus Acha ti - Funct Ecol 1993:281–286 nella. US Department of the Interior, US Fish and McKeown S (1991) Jackson’s chameleons in Hawaii are the Wildlife Service, Portland, OR recently described Mt. Kenya subspecies, Chamaeleo Van Damme R, Bauwens R, Verheyen RF (1991) The thermal jacksonii xantholophus. Bull Chicago Herpetol Soc 26:49 dependence of feeding behavior, food consumption and McKeown S (1996) A field guide to reptiles and amphibians gut-passage time in the lizard Lacerta vivipara Jacquin. in the Hawaiian Islands. Diamond Head Publishing, Los Funct Ecol 5: 507−517 Angeles, CA Waldschmidt SR, Jones SM, Porter WA (1986) The effect of Mead AR (1979) Economic malacology: with particular body temperature and feeding regime on activity, pas- reference to Achatina fulica. In: Fretter V, Peake J (eds) sage time, and digestive coefficient in the lizard Uta Pulmonates, Vol 2B. London Academic Press, London stansburiana. Physiol Zool 59: 376−383 Necas P (1999) Chameleons: Nature’s hidden jewels. Chi- Waring GH (1997) Preliminary study of the behavior and maira, Frankfurt am Main ecology of Jackson’s chameleons of Maui, Hawaii. Re - Rödder D, Bohme W, Lotters S (2011) ‘Chamaliens’ on the port for USGS/BRD/PIERC Haleakala Field Station. Ha - Hawaiian Islands: spatial risk assessment for the invasive waii Ecosystems at Risk (HEAR) project. www. hear.org Jackson’s chameleon (Chamaeleonidae). Salamandra 47: Xiang J, Weiguo D, Pingyue S (1996) Body temperature, 36−42 thermal tolerance and influence of temperature on sprint Spawls S, Rotich D (1997) An annotated checklist of the speed and food assimilation in adult grass lizards, Taky- lizards of Kenya. J East Afr Nat Hist 86:61−83 dromus septentrionalis. J Therm Biol 3:55−61

Editorial responsibility: Luca Luiselli, Submitted: December 13, 2013; Accepted: January 28, 2014 Rome, Italy Proofs received from author(s): April 29, 2014