Tuatara (Sphenodon punctatus) feeding ecology in the presence of kiore (Rattus exulans)

Transcription

1 New Zealand Journal of Zoology ISSN: (Print) (Online) Journal homepage: Tuatara (Sphenodon punctatus) feeding ecology in the presence of kiore (Rattus exulans) Graham T. Ussher To cite this article: Graham T. Ussher (1999) Tuatara (Sphenodon punctatus) feeding ecology in the presence of kiore (Rattus exulans), New Zealand Journal of Zoology, 26:2, , DOI: 1.18/ To link to this article: Published online: 3 Mar 21. Submit your article to this journal Article views: 237 View related articles Citing articles: 12 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 9 January 218, At: 1:27

2 New Zealand Journal of Zoology, 1999, Vol. 26: / $7./ The Royal Society of New Zealand Tuatara (Sphenodon punctatus) feeding ecology in the presence of kiore (Rattus exulans) Downloaded by [ ] at 1:27 9 January 218 GRAHAM T. USSHER School of Environmental and Marine Sciences University of Auckland Private Bag 9219 Auckland, New Zealand g.ussher@auckland.ac.nz Abstract The diet of tuatara was investigated on kiore-inhabited Lady Alice Island in the Hen and Chickens Group, northern New Zealand, between 1993 and Both dietary (targeted) and inedible (incidental) items were eaten by tuatara. Dietary items recorded were exclusively invertebrate in origin. Estimates of environmental availability of invertebrates and indicated that the prey consumed were strongly selected by size and by taxa. Beetles, insect larvae, arachnids and weta comprised the greatest proportion of total diet, appeared in the greatest number of stomachs and were taken in excess of their abundance. Most prey were >1 mm in length, despite an abundance of smaller prey in the environment. The risk of predation by tuatara was greatest for terrestrial invertebrates and least for arboreal species. The composition of diets by habitat was largely similar between mid-successional kanuka (Kunzea ericoides) and late successional mixed broadleaf forest. The size distribution of prey eaten by tuatara in kanuka habitat during autumn, and the low number of tuatara yielding dietary samples, are discussed in terms of food competition with kiore. Overall, the foraging behaviour of tuatara was not obviously different on rodent-free compared to kiore-inhabited islands, either indicating that food competition is insufficient to influence diet, or that tuatara are unable to change their feeding behaviour under a higher degree of competitive pressure for prey items. Clear support for either hypothesis is lacking. Z9819 Received 12 May 1998, accepted 17 November 1998 Keywords dietary analysis; feeding ecology; tuatara; Sphenodon punctatus; kiore; Rattus exulans INTRODUCTION The two species of tuatara (Sphenodon) are restricted to approximately 3 islands around the coast of New Zealand (Cree & Butler 1993). As ground-based primary predators, tuatara consume a wide range of prey species. Foods include seabird adults, chicks and eggs (Walls 1981), skinks, geckos, frogs (Newman 1977), and even other tuatara (Daugherty & Cree 199; Fraser 1993), as well as an enormous range of invertebrates (Walls 1981). Studies of their foraging ecology and feeding behaviour have focused on tuatara from rodent-free Stephens Island, where large beetles (> 1 mm in length) dominate adult and juvenile diets (Walls 1981; Fraser 1993). Tuatara on Stephens I. are highly selective of prey group, but feed opportunistically within these groups. Nine islands which support tuatara also have, or had until recently, populations of the introduced Polynesian rat or kiore (Rattus exulans) (Cree & Butler 1993; R. Pierce pers. comm.). The kiore has been implicated in the local extinction or decline of a wide range of native vertebrates and invertebrates (Atkinson & Moller 199, and refs. therein), although much of this evidence remains circumstantial (Craig 1986). As with tuatara, the animal component of the kiore diet consists almost exclusively of invertebrates (Bettesworth 1972; Bunn 1979; Bunn & Craig 1989; Newman & McFadden 199; Roberts & Craig 199), of which beetles are among the most frequent and abundant prey (Ussher 1995). Three islands that still support kiore (or did until recently), also have substantial populations of northern tuatara (S. punctatus punctatus) (Cree & Butler 1993). General descriptions of the diets of tuatara in the presence of kiore have been published (Newman & McFadden 199) but little is known of the specific prey groups or of the feeding behaviour

3 118 New Zealand Journal of Zoology, 1999, Vol. 26 Downloaded by [ ] at 1:27 9 January 218 involved. This study aims to describe the feeding behaviour and foraging ecology of tuatara sympatric with, or allopatric with, kiore. It also aims to provide baseline information for a larger study assessing the impacts of kiore on native biota. STUDY AREA Tuatara diet and the availability of invertebrates were studied on Lady Alice I. (138 ha) in the Hen and Chickens Group (35 5'S, 'E), situated 1 km off the coast of Whangarei, northern New Zealand. Because tuatara are inactive over the colder winter months (Cree & Butler 1993) visits to the island were restricted to the spring (mid-september), summer (late February) and autumn (mid-may) months of 1993 and Lady Alice I. is the most modified of the Chickens Group. In the early 19s, forest was cleared and stock farming established at one end of the island (Hayward & McCallum 1984). Mid-successional kanuka (Kunzea ericoides) forest and late successional mixed broadleaf forest are now the two dominant and structurally distinct vegetation types. To account for possible differences in foraging ecology between habitats, diets of tuatara were described in both. Movement of tuatara between study sites or adjacent forest types during foraging was considered unlikely, because the sites were separated by approximately 1 km; adult tuatara seldom move far (see Newman & McFadden in press), and there were non-sampled, 2 m wide 'buffer' strips between habitats. The least disturbed (late successional) mixed broadleaf forest was dominated by kohekohe (Dysoxylum spectabile), karaka (Cornynocarpus laevigata), puriri (Vitex lucens), tawapou (Planchonella novo-zelandicd) and pohutukawa (Metrosideros excelsa). The understorey consisted of immature forms of the above canopy trees, plus kawakawa (Macropiper excelsum) and karamu (Coprosma spp.). The mid-successional regenerating forest was dominated by mature, aging kanuka below which broadleaf species, kawakawa, hangehange (Geniostoma rupestre) and karamu grew. Study sites were sampled over two week periods at each season and at similar times within seasons during both years. MATERIALS AND METHODS Tuatara diets Tuatara typically shelter in underground burrows (usually seabird burrows) by day, and emerge at night to feed. Individuals active on the forest floor were caught by hand after methodically searching the same areas of habitat on each visit. The stomachs of all individuals caught were sampled. Stomach contents were flushed (see Legler & Sullivan 1979) out onto a fine sieve (.5 mm mesh) using a modified garden pressure sprayer (brand Cambrian) and the regurgitated food preserved in 7% alcohol until analysis. Tests made before sampling indicated that a single stomach-flushing was sufficient to obtain the entire stomach contents. Samples were placed on petri dishes and divided into quadrats, and then separated into specific dietary constituents. Their relative volumes were estimated to the nearest 5% under a stereo-microscope. Items were initially separated into animal and non-animal groups, after which the animal component was divided into vertebrate and invertebrate constituents and identified to Order or Family level. Unknown species were added to a reference collection of invertebrates on the study island. Prey were categorised according to body length of whole specimens (collected from the study areas) into three classes, <3 mm, 3-1 mm or >1 mm in length. Snails are eaten whole by tuatara, which allows easy estimation of size. Invertebrate availability Environmental abundance and diversity of invertebrates was estimated simultaneously with tuatara diets, using pan-traps to sample litter-dwelling invertebrates. Pan-traps consisted of a clear plastic dish (approximately 24 x 16 X 4.5 cm) placed flat on top of the leaf litter and filled Vi full with a mixture of water and detergent (to ensure removal of water surface tension). The upper lip of each trap was made level with the leaf litter by packing the external sides of each trap with soil and leaf litter. Twenty pan-traps were placed at 1 m intervals in broadleaf habitat. Invertebrates were not sampled from kanuka habitat, because sampling effort was pre-determined by a wider study of kiore interactions with tuatara. Traps were set for periods of 4-5 days on each trip and emptied every 24 hours. Analysis Individual tuatara ate relatively few (numbers and diversity) food items compared with the extensive

4 Ussher Feeding Ecology of Tuatara 119 Fig. 1 Proportion of captured tuatara yielding stomach samples on Lady Alice Island. Numbers of animals sampled appear above columns. 3 4 i I T 2 Kanuka habitat Mixed broadleaf habitat Downloaded by [ ] at 1:27 9 January 218 species diversity of the tuatara diet generally. This, and the small number of samples for each sampling period, prevented any analysis of specific prey groups within individual seasons and between habitats. Because preliminary observations indicated that most invertebrates recorded within the trapped and diet groups were insects, most non-insect items were pooled by Class and most insect species by Order. Class Malacostraca (non-insect) was divided into Order Amphipoda and Order Isopoda, because their mode of locomotion differs (jumping vs. crawling respectively) and hence so does their vulnerability to predation by tuatara. Similarly, weta (insect Order Orthoptera) were divided into ground-dwelling and tree dwelling species. The inability of pan-trapping to sample some invertebrate foods eaten by tuatara meant that some prey groups were excluded from the foraging strategy analyses, such as insect larvae, worms, gastropods, and tree weta (SubFamily Stenopelmatidae). The efficiency of sampling (the relative proportions of prey groups or number of prey groups recorded) was calculated from Fagen and Goldman's estimate of sampling coverage (Lehner 1979), where 1. = total coverage and <1. = reduced coverage (see also Clode & MacDonald 1995). Non-parametric techniques were used for most analyses because of small sample sizes. Foraging success and size-related prey selection was analysed using the PC based Statistical Analysis System (SAS) Categorical Data Analysis (CATMOD) procedure. This uses a maximum likelihood Analysis of Variance (ANOVA) function to test for changes in proportions of specified variables (e.g. presence of food in stomachs, specific prey) between treatments (between habitats and similar seasons pooled for both years). Interactive effects of habitat and season were non-significant for all models, allowing removal following preliminary analysis. Zero values in data-sets were approximated to.1 during analysis, meaning that the differences in final p-values between approximations and actual zero values were at least one order of magnitude less than the minimum alpha level of significance set for all analyses (.25 two-tailed tests). The numbers of prey items eaten by tuatara were normalised by logtransformation before analysis using Generalised Linear Models ANOVA. RESULTS Altogether, 128 tuatara from two habitats were stomach-flushed over six sampling periods, yielding 79 samples. The smallest tuatara sampled was 183 mm snout-vent length (SVL), and the largest 272 mm SVL. A wide range of organic and inorganic material was found in stomachs, ranging from animal remains, seeds and fruits to live and dead leaves, twigs, dirt, stones and sand. The proportion of animal material present varied greatly between samples, although in most stomachs (63%) non-animal material comprised 15% or less of the volume consumed. Of the 79 samples, one consisted entirely of non-animal material, 16 (2%) consisted entirely of animal material, and the remaining 62 were a mixture. Most tuatara had not fed recently and had empty stomachs (Fig. 1). Foraging success was defined as the proportion of tuatara containing animal food and was largest during spring and smallest during summer. There was no significant difference in foraging success between tuatara from the two habitats at

5 12 New Zealand Journal of Zoology, 1999, Vol. 26 Downloaded by [ ] at 1:27 9 January 218 the same season (CATMOD: season df=l x 2= 5.56 P >.5; habitat df=l % 2 =.62 P >.1). Coverage of the range of dietary constituents by sampling was high in both habitats (kanuka:.95, broadleaf:.87), suggesting that increasing the number of samples would not have greatly increased the proportions of prey groups recorded or the numbers of prey groups recorded. Diversity of foods eaten Only invertebrate food items were found in the 78 samples containing animal prey. Most prey belonged to the Phylum Arthropoda (85.6%), and the remainder were either earthworms (Phylum Annelida: 7.%) or slugs and snails (Phylum Mollusca: 7.4%). Insects dominated the Arthropod prey (86.%) (Appendix 1), while spiders and harvestmen (Class Arachnida: 9.1%), slaters (Order Isopoda: 1.4%), ground hoppers (Order Amphipoda: 1.2%), millipedes (Class Diplopoda: 1.7%) and centipedes (Class Chilopoda:.6%), made up the remainder. The proportion of prey items which could not be identified was very small (.85 % = 3 items). Of these, one was a beetle, and was included within the Order Coleoptera in discussions of general consumption patterns, and the other two were Arthropod in origin. Abundance of prey eaten Insect larvae (Orders Coleoptera, Lepidoptera and Diptera) and beetles (Order Coleoptera) were the two prey groups consumed in the greatest quantities (irrespective of individual diets), and together they comprised over 5% of the total number of items listed in Appendix 1. The next most numerous prey items were arachnids, worms, ants and wasps (Order Hymenoptera) and weta (Order Orthoptera). There was no significant difference between the relative proportions of prey groups eaten by tuatara in kanuka and broadleaf habitats (Wilcoxon signed rank sum test T=75 n=21 P >.1). Many of the same prey species which numerically dominated the total items eaten (above) were also eaten most often by individual tuatara (% frequency of occurrence of samples containing a given item). Beetles were consistently eaten by the most tuatara in most samples (4 out of 6) (Table 1). Insect larvae, followed by arachnids and weta, were the next most frequently represented prey groups. As implied Table 1 Frequency of occurrence (% of samples with given item) of invertebrate groups in the stomachs of tuatara in kanuka and broadleaf habitats on Lady Alice Island. n= number of tuatara sampled. Prey Group Season Beetles Insect Larvae Arachnids Weta Snails/ slugs Worms Ants/ Wasps Moths Other Insects Other Arthropods summer n=l Kanuka 1 autumn n= spring n= summer n= Broadleaf autumn n= spring n= Table 2 Seasonal differences in the numbers of food items eaten by tuatara from two habitats on Lady Alice Island. Animals whose stomachs did not contain food were excluded from the analysis. Season Summer Autumn Spring Kanuka mean + S.E. 1.8 ± ±2. Mean number of food Forest n items in tuatara stomachs Broadleaf Forest mean ± S.E. n 3.2 ± ± ±

6 Ussher Feeding Ecology of Tuatara 121 Fig. 2 Selection of invertebrate groups by tuatara in mixed broadleaf forest, Lady Alice Island. Only those prey groups for which availability estimates were possible (from pan-trapping) are shown. Altogether, 29 prey items were eaten by tuatara and 1957 comparative specimens were caught in pan-traps. Cock- Hemiptera Amphip(X)S Bnsllelails Flies Rove beetles Invertebrate Group Downloaded by [ ] at 1:27 9 January 218 by this, there was a significant relationship between the rate prey groups were eaten (proportional composition) and how widespread they were among diets (frequency of occurrence) (Spearman's Correlation Coefficient=.94, df=19, P <.1). Overall, insect larvae, beetles, arachnids and weta (in that order) were the prey groups most likely to be eaten by tuatara in either habitat. Tuatara in kanuka habitat ate significantly greater numbers of prey during spring than other seasons (Table 2) (ANOVA F= df=43 P <.1) but those in broadleaf habitat did not (ANOVA F=.9 df=29 P >.5). Tuatara in the two habitats ate similar numbers of items in the same season (ANOVA summer F=2.22 df=23 P>.1, autumn F=3.2 df=23 P >.5, spring F=4.51 df=26 P >.25). Size of prey eaten Overall, 77% of prey individuals were >1 mm in length. The remaining 23% of items were 3-1 mm, and none was below 3 mm in length. Large prey comprised a similar proportion of diets throughout all seasons in both habitats (CATMOD kanuka df=2 X 2 =3.42 P >.1, broadleaf df=2 % 2 =4.94 P >.5). The size distributions of prey were similar between habitats in spring (CATMOD df=l % 2 =.41 P >.5) and summer (CATMOD df= 1 % 2 = 1.14 P >.1) but during autumn tuatara in kanuka forest ate proportionally fewer large prey than did tuatara from broadleaf forest (kanuka: large prey 63%, broadleaf: large prey 88%; (CATMOD df=l % 2 =5.49 P =.2)). Feeding preferences Prey groups were not eaten in the same proportions that they were estimated to be available. Particularly abundant invertebrates such as amphipods, flies and rove beetles (Family Staphylinidae) appeared in disproportionately lower numbers in tuatara diets while other beetles, litter dwelling weta (SubFamily Rhaphidophorinae and SubFamily Stenopelmatidae) and arachnids were eaten in numbers exceeding their relative availability (Fig. 2). Additionally, most prey (76.1%) in the diets of tuatara from broadleaf forest were large, even though few large invertebrates were available (8.%). Weta: In total, 23 (6.5%) prey items identified were the remains of weta. Most of these (78%) were litter dwelling weta, and the remainder were tree weta. Only litter weta of >1 mm body length were eaten in all seasons and habitats, although smaller litter weta (3-1 mm) were present in all seasons and habitats sampled (range: 1-5/1 trap nights). Tree weta were eaten in both habitats but not in all seasons. Three were eaten during summer and one during spring. Remains from one summer sample also yielded 35 fully formed tree weta eggs. Insect larvae: Most larvae eaten were of moths (89.7%) followed by fly (6.%) and beetle larvae (4.3%). Moth larvae were eaten predominantly during spring in both habitats (kanuka 96% spring; broadleaf 71%), but there was no seasonal pattern of consumption for beetle and fly larvae. Larvae >1 mm in length comprised 97% of those eaten, with only 4 out of 116 larvae falling in the smaller (3-1 mm) size range. All four of these smaller larvae were >7 mm in length. Beetles: Many different beetle groups were eaten by tuatara in both habitats. Over both habitats most (71%) beetles eaten were large (>1 mm) and belonged to only three families (Scarabaeidae 67%, Cerambycidae 18% and Elateridae 15%). Within broadleaf habitat only, some beetle groups were eaten in numbers disproportionate to their estimated availability (e.g. Scarabaeidae, Cerambycidae and Bostrichoidea; Table 3). Other groups, such as Staphylinidae and Scaphidiidae, were abundant in

7 122 New Zealand Journal of Zoology, 1999, Vol. 26 Downloaded by [ ] at 1:27 9 January 218 the environment but were not eaten by tuatara. Remaining groups appeared to be consumed in numbers relative to their availability. Representation of large beetles in diets reflected the relative composition found in the environment for Scarabaeidae, Elateridae and Carabidae (Table 3), but not for Cerambycidae. DISCUSSION Feeding behaviour Tuatara on Lady Alice I. are opportunistic foragers, but their choices are heavily influenced by prey size and prey density. The invertebrate groups most often selected were insect larvae, beetles, arachnids and weta, indicating that invertebrates that live in or are attracted to leaf litter are most likely to be eaten by tuatara. The method used by tuatara to detect and catch prey may contribute significantly to the range of prey types consumed (or omitted) from their diets. Visual cues are of prime importance for foraging tuatara (Walls 1981; Meyer-Rochow & Teh 1991), although olfactory stimuli are used to detect some prey, such as eggs and carrion (Walls 1981). Prey are caught by a sudden lunge, although a large number of initial attempts miss the intended food (Walls 1981). It is to be expected that only those prey with the optimal characteristics for identification and successful predation, such as large size and relatively slow movement, will appear in diets with the greatest frequency, whereas those that are small and capable of sudden evasive manoeuvres (such as amphipod hoppers and rove beetles) should be eaten in the least quantities. This is indeed the case for the tuatara on Lady Alice I., where prey are predominantly large (>1 mm) and with limited means of escape. This pattern is demonstrated particularly well by the beetles. Groups with large numbers of large individuals (e.g. Scarabaeidae) or those which were terrestrial and slow moving (e.g. Elateridae and Curculionidae) were most often represented. Many more large Cerambycidae were eaten than were counted by the pan-traps. Adult Cerambycidae forage in the canopy at night but take refuge under bark by day. Their habitats when emerging at dusk may make them susceptible to predation by tuatara, but in the canopy they were out of reach of the pan-traps. Influence of kiore There was little difference between the diets of tuatara on Lady Alice I. and on rodent-free Stephens I. (Walls 1981; Fraser 1993). Both used opportunistic, size-related and density-dependent foraging strategies. This suggests that either the foraging strategy of the tuatara is relatively inflexible under increased competition for food resources from kiore, or that the intensity of competition from kiore is insufficient to promote change. Some of the seasonal and habitat differences in the size and number of prey eaten by tuatara could be linked to kiore population dynamics. Kiore on northern offshore islands (such as the Chickens Group) go through an annual cyclic pattern of abundance (Craig & Bunn 1989). Numbers of kiore are lowest during winter months, increase in early spring, reach a maximum density during autumn, and Table 3 Distribution of beetle groups eaten by tuatara in mixed broadleaf forest, Lady Alice Island, by family and by body size. 'Large specimens' are >1 mm body length. Beetle Family Stomachs n=29 % Composition Pan-traps n=286 Stomachs % Large Specimens n Pan-traps n Scarabaeidae Cerambycidae Curculionidae Bostrichoidea Elateridae Carabidae Ostomidae Crysomelidae Tenebrionidae Scaphidiidae Staphylinidae

8 Ussher Feeding Ecology of Tuatara 123 Downloaded by [ ] at 1:27 9 January 218 fall to low densities when food becomes exhausted during late autumn (Bunn 1979; Moller & Craig 1987; Newman & McFadden 199). Population increase depends on the availability of high protein seeds and fruits (Craig & Bunn 1989), which are abundant but seasonally restricted in early successional habitats. When they are available, kiore can reach high densities, but then switch increasingly to animal food sources as plant sources become exhausted (Roberts & Craig 199; Ussher 1995). In mature habitats, such as the broadleaf forest in this study, population fluctuations are dampened (Craig 1986) because the supply of seeds and fruits is limited and seasonally dispersed (Speed 1986). On Lady Alice I. greater numbers of food items were found in tuatara diets during spring in kanuka habitat, when kiore are at comparatively low densities and therefore offering reduced competition. Similarly, there was a significant reduction in the proportion of large prey eaten by tuatara from kanuka forest during autumn compared with broadleaf forest, when kiore are at high densities in early successional habitat. The generally small proportion of tuatara yielding dietary samples in both habitats may also be due competition with kiore for food items. The removal of kiore from Lady Alice in 1995 has offered the opportunity to experimentally test this hypothesis. The results of that study will be presented separately. This study sampled only tuatara >18 mm SVL. Analysis of tuatara diet on Stephens I. indicates that the diversity and size of prey consumed is similar in both adult (>18mm SVL) and juvenile ( mm SVL) age classes. However, post-hatchling tuatara (< 12 mm SVL) take smaller prey (Fraser 1993). The present study shows that small prey (<1 mm) were still abundant in the environment despite the presence of kiore (which also eat both large and small prey: Ussher 1995). The significance of possible dietary interactions between tuatara and kiore is the focus of an ongoing study on the Chicken Islands. Given the lack of evidence for direct kiore predation on tuatara young and eggs on kiore-inhabited Islands, many authors have theorised about the influence of dietary competition on food availability and quality for female tuatara and the effects this may have on breeding success (Newman & McFadden 199; Newman et al. 1994; Cree et al. 1995). Monitoring of invertebrate populations and the diets of tuatara following the removal of kiore from islands in the Chickens Group will help assess the nature and degree of influence that kiore may have on this ancient reptile. ACKNOWLEDGMENTS Many thanks to the Whangarei Department of Conservation staff and to the many volunteers for assisting with field work. Also thanks to Don Cowley and Chris Green who helped to identify invertebrate specimens and remains and to John Craig, Don Newman, Ray Pierce and David Towns and an anonymous referee for commenting on drafts of the manuscript. Financial support for this project was provided by the New Zealand Lotteries Board. REFERENCES Bettesworth, D. J. 1972: Rattus exulans on Red Mercury Island. Tane 18: Bunn, T. J. 1979: The effects of food supply on the breeding behaviour and population ecology of kiore on Tiritiri Matangi Island. Unpublished MSc. Thesis, University of Auckland, Auckland, New Zealand. Bunn, T. J.; Craig, J. L. 1989: Population cycles of Rattus exulans, population changes, diet and food availability. New Zealand Journal of Zoology 16: Clode, D.; MacDonald, D. W. 1995: Evidence for food competition between mink (Mustela vison) and otter (Lutra lutra) on Scottish Islands. Journal of the Zoological Society London 237: Craig, J. L. 1986: The effects of kiore on other fauna. In: Wright A. E.; Beever R. E. ed. The offshore Islands of Northern New Zealand. New Zealand Department of Lands and Survey information series 16: Craig, J. L.; Bunn, T. J. 1989: Theeffects of experimental manipulation of food supplies on a population of Rattus exulans. New Zealand Journal of Zoology 16: Cree, A.; Butler, D. 1993: Tuatara recovery plan (Sphenodon spp.). Threatened species recovery plan series 9. Wellington, New Zealand, Department of Conservation. Cree, A.; Daugherty, C. H.; Hay, J. M. 1995: Reproduction of a rare New Zealand reptile, the tuatara Sphenodon punctatus, on rat-free and rat-inhabited islands. Conservation biology 9: Daugherty, C.; Cree, A. 199: Tuatara: A survivor from the dinosaur age. New Zealand Geographic: April- June 6: Fraser, J. 1993: Diets of wild tuatara (Sphenodon punctatus) on Stephens Island. Unpublished MSc thesis. University of Otago, New Zealand. Hayward, B. W.; McCallum, J. 1984: Offshore island research group trip to the Chickens (Marotere) Islands, north-east New Zealand, new year Tane 3:

9 124 New Zealand Journal of Zoology, 1999, Vol. 26 Downloaded by [ ] at 1:27 9 January 218 Lehner, P. N. 1979: Handbook of ethological methods. New York, Garland STPM Press. Legler, J. M.; Sullivan L. J. 1979: The application of stomach-flushing to lizards and anurans. Herpetologica 35: Meyer-Rochow, V. B.; Teh, K. L. 1991: Visual predation by tuatara (Sphenodon punctatus) on the beach beetle (Chaerodes trachyscelides) as a selective force in the production of distinct colour morphs. Tuatara 31: 1-8. Newman, D. G. 1977: Some evidence of the predation of Hamilton's frog (Leiopelma hamiltoni (McCulloch)) by tuatara (Sphenodon punctatus (Grey)) on Stephens Island Proceedings of the New Zealand Ecological Society 24: Newman, D. G.; McFadden, I. 199: Seasonal fluctuations of numbers, breeding and food of kiore (Rattus exulans) on Lady Alice Island (Hen and Chickens Group), with a consideration of kiore: tuatara (Sphenodon punctatus) relationships in New Zealand. New Zealand Journal of Zoology 17: Newman, D. G.; Watson, P. R.; McFadden, I. 1994: Egg production by tuatara on Lady Alice and Stephens Island, New Zealand. New Zealand Journal of Zoology 21: Roberts, M.; Craig, J. L 199: The demography of kiore (Rattus exulans) in three habitats. New Zealand Journal of Zoology 17: Speed, H. 1986: Demography and diet of kiore (Rattus exulans) on Little Barrier Island. Unpublished MSc. thesis, University of Auckland, New Zealand. Ussher, G. T. 1995: Feeding ecology and dietary interactions of tuatara and kiore on the Chicken Islands. Unpublished MSc. thesis, University of Auckland, New Zealand. Walls, G. Y. 1981: Feeding ecology of the tuatara (Sphenodon punctatus) on Stephens Island, Cook Strait. New Zealand Journal of Ecology 4:

10 Ussher Feeding Ecology of Tuatara 125 Appendix 1 Percentage composition of prey items retrieved from 78 tuatara (total items = 353). Downloaded by [ ] at 1:27 9 January 218 Phylum Annelida Mollusca Arthropoda Taxonomic level to which Class Oligochaeta Gastropoda Chilopoda Diplopoda Malacostraca Arachnida Hexapoda prey were identified Order Amphipoda Isopoda Araneae Opiliones Microcoryphia Orthoptera Phasmida Mantodea Blattaria Hemiptera Homoptera Coleoptera Diptera Lepidoptera Hymenoptera Family Snails Slugs Total Other Inverts Talitridae Litter weta Tree weta Lygaeidae Pentatomidae Cicadidae Bostrichidae Elateridae Carabidae Cerambycidae Crysomelidae Curculionidae Ostomidae Pselaphidae Scarabaeidae Tenebrionidae unidentified beetles Larvae Adults Larvae Adults Larvae Formicidae Vespidae Other Hymenoptera Unidentified Arthropods Invertebrates < 1 mm Total Arthropods % Total Items