Abstract. Introduced populations of Western Green Lizard (Lacerta bilineata) and

Transcription

1 An Investigation into the Effects of the Western Green Lizard (Lacerta bilineata) and the Common Wall Lizard (Podarcis muralis) Introduced onto Boscombe Cliffs, Dorset, U.K. By Simon Mole A thesis submitted in partial fulfilment of the requirements for the Bachelor of Science in Wildlife Management, 2008, from Sparsholt College, Great Britain. 0

2 Abstract Introduced populations of Western Green Lizard (Lacerta bilineata) and Common Wall Lizard (Podarcis muralis) were discovered on Boscombe cliffs in Since then both species have been recorded as breeding successfully, but a more comprehensive investigation of their population was required. In this investigation, for the first time, a static survey path was followed over the cliff top adjacent to where they were first recorded in an attempt to plot their current territory and compare their numbers against the native Viviparous Lizard (Zootoca vivipara). Fifteen surveys were undertaken between April and September 2007 and at their completion had recorded 214 non-native and 44 native Lacertids. The range covered by these species showed that the two introduced species dominated the central area of the site with the native species only found in large numbers on the periphery. When these recordings were compared to previous records held by the Herpetological Conservation Trust the number of introduced species showed a marked increase (P. muralis increased between 2002 and 2007 by 40% and L. bilineata by 36%; compared with Z. vivipara which dropped by 75% over the same period). This increase was also shown when compared as a percentage against Z. vivipara sightings. A longer survey period is however required to prove any detrimental effects on the native lizard numbers. The vegetation at sites where L. bilineata was observed was surveyed and it was clear that vegetation structure rather than the species of vegetation was a larger factor on the lizards distribution. With this in mind it will be quite possible for both species to spread unimpeded along the cliffs. The spread of these introduced lizards may also affect a Sand Lizard (Lacerta agilis) colony farther west of this site, but that will require more on-going investigation to determine the competitive impact that may occur there. i

3 Acknowledgements I would like to thank everyone that helped make this investigation possible: my lecturers at Sparsholt College Andy Quayle, Sue Fitzpatrick and David Lock, Dr Chris Gleed-Owen at the Herpetological Conservation Trust and my wife Lucy Mole. ii

4 Table of Contents Introduction...page 1 Background page 1 Non-Indigenous Reptiles in Great Britain.page 2 Aims of Investigation.page 4 Method...page 7 Surveying for Lacertids......page 7 Selecting the Survey Area page 7 Surveying the Site for Lacertids page 8 Surveying the Vegetation and Structure of the Vegetation..page 11 Results..page 14 Lacertids Populations page 14 Lacertids Populations page 16 Environmental Factors.page 20 Microhabitat.page 22 Vegetation page 24 Discussion... page 28 Current Distribution of the Non-Indigenous Lacertids page 28 Effects on the Zootoca vivipara Population.....page 31 Limiting Factors on the Distribution of the Non-Indigenous Lacertids...page 35 Environmental Factors...page 35 Vegetation and Management of the Site....page 40 Possible Effects on Lacerta agilis page 44 Critique page 46 Conclusion...page 47 References... page 49 iii

5 Table of Illustrations Figure 1. Herpetological Conservation Trust historical records of Lacerta bilineata, Podarcis muralis and Zootoca Vivipara on Boscombe cliffs...page 7 Figure 2. Aerial photograph of survey area showing suitable scrub and survey patches. Suspected introduction site also shown...page 8 Figure 3. Aerial photograph of survey area showing survey route through the site. Also showing alternant starting positions.page 8 Figure 4. Aerial photograph of survey site showing structural mix of all vegetation present.page 12 Figure 5. An aerial photograph of the whole survey site showing all sightings of Lacerta bilineata, Podarcis muralis and Zootoca vivipara for page 14 Figure 6. Thematic maps of Lacerta bilineata, Podarcis muralis and Zootoca vivipara counts. Numbers in brackets indicate frequency of counts...page 15 Figure 7. The percentage of each species observed over all fifteen surveys in 2007.page 16 Table 1. The demographic breakdown of Lacerta bilineata, Podarcis muralis and Zootoca vivipara populations for page 16 Table 2. The total counts of all species seen between 2002 and 2006 within this investigation s survey area..page 17 Figure 8. A graph showing the average number of Z. vivipara, L. bilineata and P. muralis seen per visit from 2002 to 2007 page 17 Table 3. Number of observed and expected counts of each species over the five year period with Chi-squared calculations..page 18 Figure 9. An aerial map showing central locations used for comparing introduced Lacertids percentages against Z. vivipara numbers....page 19 Figure 10. A graph showing the percentage of L. bilineata and P. muralis against the observed numbers of Z. vivipara; showing trend lines...page 20 Figure 11. A graph showing the correlation between average air temperature and Lacertid observations for each survey undertaken in 2007.page 21 Figure 12. A graph showing the median basking surface temperatures for all three species of Lacertids. Also showing inter-quartile range and ranges...page 22 Table 4. ANOVA test showing results from one-way test when comparing log temperatures against species...page 23 iv

6 Table 5. Counts of all vegetative species and non-vegetative substrate recorded as a dominant feature in proximity to all Lacerta bilineata, Podarcis muralis and Zootoca vivipara sightings page 24 Figure 13. Graph showing the percentage of vegetative and non-vegetative substrate preferences for Lacerta bilineata, Podarcis muralis and Zootoca vivipara.. page 24 Figure 14. A graph showing the occurrence of plant species in quadrats where Lacerta bilineata were present or absent page 25 Figure 15. A graph showing the vegetation structure at all twenty sites that were shown to have L. bilineata present and to have L. bilineata absent page 26 Figure 16. Maps created in MapInfo showing the expansion of L. bilineata and P. muralis over the five year period ( ). Z. vivipara page 28 Table 6. Counts of both L. bilineata and P. muralis from 2002 to 2007 page 29 Table 7. Counts of Z. vivipara from 2002 to page 32 Figure 17. An aerial map showing the edge of the L. bilineata expansion, between 2004 and 2007, across the site pushing out Z. vivipara in the process... page 32 Figure 18 Structural map overlaid with species observation map showing preference for higher scrub cover..page 41 Figure 19. Location of non-native species and L. agilis colony.page 44 v

7 Introduction Background In these modern times we are more than aware of the effects of introducing species into areas of the planet where they do not naturally occur. Like all things that involve hindsight the event has already occurred and unfortunately the damage may already be irreversible. It is the varying degrees of damage that are of interest to those who wish to conserve species affected by introductions. The second leading cause, after loss of habitat, of the decline of biodiversity is the effect of species introduced beyond their native ranges (Meffe et al, 1997). There are numerous examples from around the planet of the destruction caused by introduced species. One of the most notorious herpetological examples is that of the Cane Toad (Chaunus marinus) which was introduced to Australia to combat the Cane Beetle (Dermolepida albohirtum) and has now spread from its introduction sites in Queensland nearly three thousand kilometres to the Kimberley in North Western Australia, devouring its way through many different native species en route. Another, less well known, example is that of the Brown Tree Snake (Boiga irregularis) which is believed to have been accidentally introduced to the Pacific island of Guam by stowing away on military ships (Savidge, 1988). This snake has now spread over the whole of Guam, which had no previous terrestrial snake species, causing massive declines in native birds (the Guam Rail [Gallirallus owstoni] and the Micronesian Kingfisher [Todiramphus cinnamominus] are now both extinct on Guam [Patrick, 2001]) and small mammals since its introduction. The only benefit of this snake s introduction is that it has eradicated some of Guam s other introduced reptile species. 1

8 Non-Indigenous Reptiles in Great Britain There have been many cases of non-indigenous reptiles finding themselves in the wild in Great Britain; sometimes by accident and other times intentionally. Beebee and Griffiths (2000) list the following reptilian species that are known escapees or releases in this country: Common Wall Lizard (Podarcis muralis), Aesculapian Snake (Zamenis longissima), Green Lizard (Lacerta viridis), North American Garter Snake (Thamnophis sirtalis), European Pond Terrapin (Emys orbicularis) and Red-Eared Terrapin (Trachemys scripta elegans). Not all of these species have had success establishing breeding colonies. There are two Lacertid species that have been introduced to the cliffs at Boscombe (50 o 43 N, 1 o 49 W) and they are the Common Wall Lizard (Podarcis muralis) and the Western Green Lizard (Lacerta bilineata), see appendix 1 and 2 for identification. Both of these species naturally occur to the northwest coast of France, where they are at the northern edge of their range (Arnold & Ovenden, 2004), but there is no fossil evidence to suggest they colonised this country after the last glaciation. The DNA from both species was examined by Deichsel, Gleed-Owen & Mayer (2007) to try to determine from where these animals had originated. Once segments of their DNA had been identified they were compared with existing homologous regions of sequences from samples across Europe to try to find a relation. A comparison was found and it was clear that both species had originated from northern Italy close to the Slovenian border. There is only one species of native Lacertid found within this investigation s survey area and that is the Viviparous Lizard (Zootoca vivipara), although it is worth mentioning that a colony of Sand Lizards (Lacerta agilis) is present west of this site (see appendix 3 for identification of both these species). There have been deliberate attempts to introduce both L. 2

9 bilineata and P. muralis into Great Britain over the years with varying degrees of success. One attempt at releasing L. bilineata dates back to the nineteenth century (Lever, 1979). All introductions of L. bilineata, apart from the population at Boscombe, have not persisted. P. muralis introductions have, on the whole, fared a lot better with many breeding colonies established around the south of England. There are a number of colonies on the Isle of Purbeck, plus two colonies on Portland in Dorset and another at Shoreham in Sussex. There are also colonies farther east in Kent and London, to name but two. There are also a number of colonies in Surrey located around the Reigate and Nutfield area (Wycherley & Anstis, 2001). The largest and, most likely, longest established is the colony in and around Ventnor on the Isle of Wight; this colony may have been present for over 100 years (Quayle & Noble 2000). It has been suggested that this colony, as it has persisted for so long as that no one knows its true origin, might be a native population. There is, however, no fossil or even sub-fossil evidence to support this theory (Gleed-Owen, 2007). P. muralis has proved to be a good coloniser in other countries too, as shown by Allan et al (2006), Burke & Deichsel (2008), Münch (2001) and Walker & Deichsel (2005). L. bilineata has proven less successful in colonising high numbers of locations, with only one record from Kansas, U.S.A. (Burke & Deichsel, 2008 and Gubanyi, 2000). Knowing that they are capable of successfully breeding in this country s climate there was concern when a population of P. muralis was discovered on the cliffs at Boscombe in July of 2002 (Gleed-Owen, 2004) only two kilometres from the population of L. agilis. L. bilineata was also observed at the same time, but it was not known then whether they would be able to successfully breed or not. That was five years ago and the population is still there. Under the Wildlife and Countryside Act of 3

10 1981 this was an illegal release of both L. bilineata and P. muralis as the act states that any person who releases or allows to escape into the wild any animal which is a) of a kind not ordinarily in and is not a regular visitor to Great Britain in a wild state, or b) included in Part 1 of Schedule 9 (animals which are established in the wild), shall be guilty of an offence (as stated at This investigation is not however concerned with how these lizards got there, because as mentioned at the start we are dealing with hindsight and they are already well established. Time and effort is far better spent on a far more important question: what effects are they having on the local herpetofauna? The site at Boscombe is part of a larger stretch of coastline that extends from Hengistbury Head westwards to Poole Harbour. This disjointed cliff top habitat consists of the expected vegetation species such as Marram Grass (Ammophila arenaria) plus numerous introductions such as Holm Oak (Quercus ilex). The cliff face itself is south facing at a 45 o angle, this combined with the dense scrub cover provides perfect reptile habitat. Chris Gleed-Owen s account of where he first observed these lizards (the zigzag path leading from the cliff top down to Boscombe beach) was used as a central point for this investigation. This focus will be used as a starting point for all anticipated spread of both L. bilineata and P. muralis. As reptiles are ectothermic animals they are forced into hibernation by this country s cold winters, this means the best time to observe reptiles in Great Britain is between April and September and this is when this investigation was undertaken. Aims of Investigation There are a number of questions that need to be answered to find out whether the introduced Lacertid species were affecting the native Z. vivipara and had the potential to spread farther than their current range along the cliffs. Where the nonnative species of Lacertids were found on the cliff top was the most important issue 4

11 that needed answering as this is the starting point from where this investigation can be explored further. What are the habitat preferences of the introduced Lacertids and does this affect their distribution? Within its natural range P. muralis is at home in a wide variety of habitats where it is sufficiently dry and sunny (Street, 1979). As its name suggests it is frequently found basking or foraging on stone walls and is quite at home in urban environments. It is a very active and alert predator feeding mostly on invertebrate prey such as spiders, butterflies and beetles, but is also opportunistic feeding on fruit or discarded human waste such as ice cream. L. bilineata is typically found, within its natural range, in and around dense bushy vegetation with good exposure to sun (Arnold & Ovenden, 2004). Being bigger then P. muralis means that L. bilineata can predate larger food items as well as its main prey items, invertebrates, such as nestling birds; it is also omnivorous and will feed on fruit when it is available. With the issue of habitat preferences at Boscombe answered, more questions can be asked about their potential to increase their territory. The first would be, are they spreading and if so where are they spreading to and how quickly are they getting there? Questions could then be asked about limiting factors and other environmental effects on the population. On the European mainland both of these species are widespread and have successfully colonised much of the central and southern areas of the continent due to their generalist habitat requirements; their only limiting factors being temperature, competition from other Lacertid species and urbanisation in the case of L. bilineata. Both L. bilineata and P. muralis have a number of species that prey on them. Mammalian predators include Mustelids, Hedgehogs (Erinaceus europaeus) and Shrews (Sorex spp) and avian species such as Gulls (Larus spp), Corvids and raptors such as Kestrels (Falco tinnunculus). They will also fall prey to Ophidian predators. As this survey site has a large visitor presence Bournemouth 5

12 Borough Council has to manage the site accordingly and this tends to involve removal of the vegetation when it get too big (mostly Gorse [Ulex europaeus]). This management has a direct influence on the habitat at this site and could play a major role in the spread or restriction of both the introduced species. Assessing the effects on the native Z. vivipara during this territory expansion process would be the hardest question to answer given the time span of this investigation. A longer time period is ideally needed to monitor population growth effectively with the same consistent surveying techniques used in this investigation always being applied. For this investigation records have been acquired from the Herpetological Conservation Trust s database tracking the sightings of the introduced species from These records are, however, random sightings with no structured survey technique being applied to them, but they are reliable. On completion of this investigation educated assumptions should be able to be made from these findings about the effects on the Z. vivipara population and on the population of L. agilis, if they increase their range that far. L. agilis has a far more specialised habitat requirement in Great Britain than it does in Europe (in Europe all of these species naturally overlap into the same geographic range quite happily due to L. agilis losing its dependence on sandy soils for oviposition) and this lack of environmental adaptability may lead it to struggle in light of new competition from these introduced species. With all of these questions answered it will be known if the null hypothesis can be rejected or not; the null hypothesis being:- The spread of the non-native Lacertids on Boscombe cliffs is not influenced by current habitat management and their spread will not have a detrimental effect on the native Herpetofauna already present. 6

13 Method Surveying for Lacertids Selecting the Survey Area The survey location was chosen taking into account the proximity of the suspected introduction site for both P. muralis and L. bilineata: the zigzag path leading down to the beach adjacent to the Boscombe Cliff Bowling Club. There were two possible areas that could have been surveyed: the cliffs themselves or the cliff tops. Due to the health and safety issues of surveying the cliffs the cliff top vegetation was selected instead; which is unfortunate as the cliffs themselves are likely to be more optimal habitat for lizards. There was, however, opportunity to survey a portion of the cliff, from its top edge, with little risk. Using a previous sightings map, provided by the Herpetological Conservation Trust, the survey area was selected just beyond the existing sightings of both species. This was done to show if either species was spreading beyond their known locations (Figure 1) or whether the population was static. The survey area was then split into six segments to make discussion of the whole area easier and to illustrate changes in population densities from the introduction site outwards. Figure 2. The Herpetological Conservation Trust historical records of Lacerta bilineata (green spots), Podarcis muralis (yellow spots) and Zootoca Vivipara (brown spots) on Boscombe cliffs. 7

14 Surveying the Site for Lacertids A preliminary visit to the site determined where the suitable habitat was located and this information was then marked onto an aerial photograph of the survey area. Suitable habitat comprised of mature Gorse (Ulex europaeus) thickets, Marram Grass (Ammophila arenaria) tussocks and an herbaceous species layer covering the cliff edge. The remaining habitat between the vegetation patches comprised of open short and long sward grassland. Figure 2 shows the survey area split into the six patches and the suitable vegetation described from the first site visit. From this overview map of the site a transect (survey path) was plotted onto it, using the information shown to survey the site covering as much of the suitable habitat as possible: see figure 3. Figure 3. Aerial photograph of survey area showing survey route through the site. Also showing alternant starting positions. 8

15 Once the transect was plotted the surveying could begin. As this survey was dealing with ectothermic species the weather played a pivotal role in determining when surveys should be conducted. As the purpose of this investigation was not to establish presence or absence of lizard species, as this was already known, surveys were always undertaken on favourable days for observing reptiles i.e. towards the end of the morning, early and late in the year, and earlier in the morning during the hotter summer months. This also meant that surveying started in April as this can be the best time of year for spotting reptiles, as when they are fresh out of hibernation they are not so alert and easier to observe. The transect (survey path) had alternating starting points, at the west end for one survey and then the east end for the next, as they were all conducted in the morning and as a general rule it will become hotter towards midday. Alternating the start points eliminated a bias to an optimal temperature being reached at the same point in each of the surveys. The process of surveying was simple and based on an observe and record principle. A global positioning system (G.P.S.) was used to plot the start of each transect (survey path) walked and the track function used to plot the length of each survey. The time, air temperature (a baking thermometer was used for this measurement), wind strength (the Beaufort scale was adopted for this), wind direction and cloud cover (a scale of zero to eight was used to represent the cloud cover e.g. 0/8 no clouds to 8/8 full cloud cover) were also recorded. The weather conditions for the previous three days were also recorded (source as rain in the past twenty four hours can have an impact on reptile observations; this is however not such a factor early in the season and has more effect later into the summer. The transect (survey path) was then walked and all species of Lacertids observed being recorded on the G.P.S.. The observations were all made within one metre of the survey path, but it must be allowed that some 9

16 lizards may have been missed within this parameter due to the observer not seeing animals within the denser patches of vegetation. As well as the location of each individual sighted basic information was also recorded, such as species (Podarcis muralis, Lacerta bilineata and Zootoca vivipara were the only known Lacertid species present), sex and age. Four cohorts were identified: 1 st year young (juveniles born in 2007), juvenile (less than a year old), immature (over one year but not showing full adult markings) and adult. Along with this basic data other information was recorded:-! Surface temperature of where the individual was basking, if within reach (a baking thermometer was used for these measurements).! Was the individual sheltered from the wind?! The aspect of the basking site.! The dominant vegetation surrounding the basking site.! Was the animal paired obviously with another (mate guarding)? When identification of species was not possible those lizards were not recorded, as this data would have no use in the analysing process. Only true species identification could be used to show where the lizards where spreading to. Once the transect (survey path) had been walked and all of the sightings recorded the time was once again noted, along with the air temperature. The finish point was also recorded on the G.P.S. and the length of the track noted to make sure that there was the same effort (in this case area covered) put into each of the surveys. When the surveys were completed the information gathered from them was plotted onto the software MapInfo ; this was done to show the species seen and their location on the survey site. It could then be seen whether the sightings that were occurring were starting to repeat in their locations. Once repetition started to occur it 10

17 was known that enough surveys had been done; fifteen surveys in total were completed. These surveys were spaced out over the course of an ordinary reptile season i.e. April to September. The surveys were conducted, as much as possible, evenly throughout this time. Due to an un-seasonally poor summer most of the surveys were completed towards the beginning and the end of the season. This spread also aimed to pick up a better representation of the demographics of the populations e.g. last year s juveniles (seen at the beginning of the season), immature animals and this year s juveniles (seen at the end of the season). Surveying the Vegetation and Structure of the Vegetation Once it had been established where all the lizards had been seen, from all fifteen surveys, a more detailed investigation into the vegetation where they were being found could commence. Only L. bilineata was used in the surveying of vegetation because they had, on the whole, been seen in more accessible locations. Ten known locations where L. bilineata had been recorded were selected. The selection was aimed at representing a range of lizards from all the different patches (i.e. 1-6) within the survey area (if they had been observed within that patch) and the two main differing habitats where they had been found: Gorse thickets and Marram Grass tussocks. To determine the size of quadrat that was to be used for surveying two sites were visited and a progressive doubling technique was applied. This involved laying the quadrat, which was divided into twenty-five parts (each part ten centimetres by ten centimetres), where the lizard was observed and then counting, using just the one square first, how many plant species fell within that square. This included any leaves that grew into the square too. The same species count was then done for two squares and then for three and so on. Using this technique the number of 11

18 species observed within each individual size rises quickly, then it levels out to form an optimal number; at this point the size of quadrat that is best suited for surveying that particular vegetation is shown. The optimal size for this survey was twenty centimetres by twenty centimetres. Once the size of quadrat was established the surveying of the vegetation could commence. At each of the ten sites the quadrat was laid where the lizard had been observed and all the different species of vegetation within it were recorded. To make a comparison to the vegetation where L. bilineata was observed ten locations were also selected within the survey area where it was not seen. These sites consisted of ten similar locations to where they had been seen, i.e. Gorse thickets and Marram Grass tussocks. Again the same method of surveying was completed recording all different plant species within a twenty centimetre by twenty centimetre quadrat. Figure 4. Aerial photograph of survey site showing structural mix of all vegetation present. The management of Boscombe Cliffs is undertaken by Bournemouth Borough Council, with the Gorse patches being cut every few years to minimise fire risk on the cliff top. To create a more detailed map of the existing vegetation structure the whole site was surveyed with all vegetation being observed and its approximate height was 12

19 plotted onto an aerial photograph. For this to be achieved the denotations on figure 4 were gauged by a visual estimation of the vegetation height by the observer, using his body to gauge it i.e. vegetation below the knee was recorded at under 50 centimetres, between the knee and the head was 50 to 180 centimetres and above the head was recorded as over 180 centimetres. This, along with the previous vegetation surveys, created an overview of all the aspects of the surveyed area to allow comparisons to be made in relation to the locations of all Lacertids observed. 13

20 Results Lacertids Populations 2007 After fifteen surveys a total of two hundred and fifty eight Lacertid observations were made; these comprised solely of the three target species L. bilineata, P. muralis and Z. vivipara. The coordinates for all observations were then loaded onto the software MapInfo to produce a population overview of the whole Figure 5. An aerial photograph of the whole survey site showing all sightings of Lacerta bilineata (green spots), Podarcis muralis (yellow spots) and Zootoca vivipara (brown spots) for survey site, figure 5. This map clearly shows that L. bilineata is distributed all along the top of the cliff edge and has increased its range northwards on the cliff top fanning out from the introduction site. P. muralis is only occupying a small area of the cliff top around the introduction site, with a couple of lone sightings farther out. Z. vivipara was located in small numbers around the central introduction site but was only found in large numbers on the periphery of the survey area. The spot map shown above has the advantage of representing the range of the lizards seen, but the disadvantage of numbers not being so easily read due to overlapping spots. For a clearer representation of numbers thematic maps were generated in MapInfo. The survey area was arbitrarily split into polygons with smaller 14

21 polygons near the centre where the high concentrations of sightings were located and larger polygons drawn on the edges of the survey area where low concentrations of sightings were located (figure 6). These maps dramatically show the partitioning of Z. a) b) c) Figure 6. Thematic maps of (a) Lacerta bilineata, (b) Podarcis muralis and (c) Zootoca vivipara counts. Numbers in brackets indicate frequency of counts. 15

22 vivipara on the periphery of the survey area whilst higher counts of both L. bilineata and P. muralis are concentrated around the central introduction area. The total count of L. bilineata was 104, with 110 P. muralis and only 44 Z. vivipara, see figure 7 for percentages. This shows that the introduced Lacertids are present in much higher numbers 17% than the native Lacertid within 40% the chosen survey area. If the populations are looked at demographically (table 1) L. 43% bilineata and P. muralis have a balanced sex ratio (binomial test, Lacerta bilineata Podarcis muralis Zootoca vivipara Figure 7. The percentage of each species observed over all fifteen surveys in Lb Pm Zv females males adult u/s immature juvenile last year juvenile this year Total Table 1. The demographic breakdown of Lacerta bilineata, Podarcis muralis and Zootoca vivipara populations for p = 0.2 for both species) whereas Z. vivipara shows a more skewed sex ratio towards the females (binomial test, p = 0.002). Of the immature lizards observed there was a relatively even spread over all three species. With regard to juvenile observations, last year s L. bilineata and P. muralis seem to have reproduced very successfully compared to Z. vivipara but in comparison this year s Z. vivipara has had the most breeding success with no observations of juvenile L. bilineata being made. Lacertids Populations Since their discovery on Boscombe Cliffs in 2002 (Gleed-Owen, 2004) the populations of L. bilineata and P. muralis have been surveyed each year with 16

23 differing degrees of surveying pressure. Both the introduced species were first seen on the zigzag path on the cliff face just south of this investigation s survey site, therefore 2002 (7 visits) 2003 (11 visits) 2004 (42 visits) 2005 (32 visits) 2006 (10 visits) Lb Pm Zv Total Table 2. The total counts of all species seen between 2002 and 2006 within this investigation s survey area. L. bilineata, although being present, was not recorded on table 2 above in 2002, as it took that first year for them to spread into this investigation s survey area. This table shows a marked increase in both the populations of introduced Lacertids. When including the records from 2007 the average number of each species shows a dramatic change. Z. vivipara in 2002 made up 93% of all lizard sightings but by 2007 made up only 17%, whereas L. bilineata rose from 0% in 2002 to 40% in 2007 and P. muralis also expanded from 8% in 2002 to 43% in Figure 8 looks at these figures from a Average Number seen per Visit (7 visits) 2003 (11 visits) 2004 (42 visits) 2005 (32 visits) 2006 (10 visits) 2007 (15 visits) Year Lacerta bilineata Podaric muralis Zootoca vivipara Figure 8. A graph showing the average number of Z. vivipara, L. bilineata and P. muralis seen per visit from 2002 to

24 different angle, the average number seen per visit from 2002 to There is a marked decline in Z. vivipara numbers (there were on average 5.29 seen per survey in 2002 compared with only 2.93 in 2007) whilst L. bilineata and P. muralis numbers have increased in the last five years (L. bilineata from 1.45 in 2002 to 6.8 in 2007, and P. muralis from 0.43 in 2002 to 7.47 in 2007). a) b) c) Numbers Observed Total Lb Pm Zv Total Numbers Expected Total Lb Pm Zv Total Chi-squared (χ 2 ) Total χ 2 Lb Pm Zv Total Table 3. Number of observed (a) and expected (b) counts of each species over the five year period with Chi-squared calculations (c). (df = 10, Lb p= 0.002, Pm p= <0.000 & Zv p= <0.000.) A Chi-Squared test was performed on this data to see if there was a significant difference between the average number of each species seen per survey between 2002 and 2007; the null hypothesis being that the average number seen per survey will increase as expected. Table 3 (c) shows that the population of L. bilineata has not fluctuated far from its expected amount over the five year period. Populations of P. muralis on the other hand have had some large rises (2007 χ 2 = 31.29) and falls (2006 χ 2 = 0.01). Overall, though, the population shows a marked unexpected change and taking account of the readings from figure 8 it is that of an increase in numbers. Z. vivipara shows a steady population rate (the reading for 2002 is biased to the fact that 18

25 both the introduced species were not yet well established on the site) until 2007 when there was a rise (χ 2 = 18.03); overall this species shows the greatest total difference (χ 2 = 69.91). The null hypothesis can therefore be rejected for P. muralis and Z. vivipara but not for L. bilineata. To try and eliminate the bias of survey pressure differences a comparison was made by calculating L. bilineata and P. muralis as a percentage against Z. vivipara. Even though a different number of surveys were completed for each year between 2002 and 2007 the percentage difference between the introduced Lacertids and Z. vivipara are still comparable due to there being similar pressures to obtain both these sets of data. A Figure 9. An aerial map showing central locations used for comparing introduced Lacertids percentages against Z. vivipara numbers. central location (figure 9) within the survey area was chosen, as one of these polygons has always shown a recording of one of the species between 2002 and For L. bilineata there was a percentage range of 77 over the five year period when comparing their numbers as a percentage of Z. vivipara numbers. The lowest percentage occurred in 2002 (0%, not yet spread onto survey site) and the highest in 2004 (77%), this gave a median percentage of 56% (lower quartile 45.5% and upper quartile 68%), figure 10 (over page). P. muralis had a smaller percentage range (52) with the minimum percentage occurring in 2002 (19%) and the maximum in 2007 (71%). Overall this also gave a median of 56% with lower quartile of 39.8% and the upper quartile being 67%. 19

26 Percentage Year Lacerta bilineata Podarcis muralis Linear (Podarcis muralis) Linear (Lacerta bilineata) Figure 10. A graph showing the percentage of L. bilineata and P. muralis against the observed numbers of Z. vivipara; showing trend lines. Environmental Factors Whilst conducting the surveys a number of other environmental factors were recorded along with the basic species, age and sex information. These readings were designed to give an overview of each species preferences for activity and a basic look at the specific microhabitat they occupied. All the species in this survey are ectothermic so one of the biggest influences on activity is air temperature. The air temperature was taken at the beginning and the end of each survey. The average temperatures recorded ranged from 18 o C to 22.5 o C with the overall average temperature being recorded at 20.1 o C for all surveys. These average survey temperatures were then compared to the number of Lacertids seen on each survey with L. bilineata showing the highest correlation (Pearson correlation, r = 0.558, df = 14, p = 0.031), Z. vivipara in the middle (Pearson correlation, r = 0.211, df = 14, p = 0.211) and P. muralis showing the weakest correlation (Pearson correlation, r = 0.128, df = 14, p = 0.649). This shows that the temperature was only statistically significant 20

27 to L. bilineata. There was a significant correlation (Pearson correlation, r = 0.52, p = 0.047) shown for all species involved when compared with average survey temperature (figure 11), showing it to be a statistically significant factor in activity. Three other environmental factors were recorded whilst conducting the surveys; they were wind strength, wind direction and cloud cover. There was a range from 2 (light breeze, Beaufort scale) to 5 (fresh breeze, Beaufort scale) recorded over all surveys and only a very weak correlation showing no significance (Spearman correlation r = 0.165, p = 0.558) when comparing wind strength to Lacertid count per survey. There was a mix of differing wind directions measured from all the surveys which had no effect on the numbers of observations made. The cloud cover had a range of 0/8 to 6/8 which led to a modest correlation which just failed to reach significance (Spearman correlation r = 0.493, p = 0.062). 35 r = 0.52 p= Count of all Species (Lb, Pm & Zv) Air Temperature (Degrees Celsius) Survey 0 Figure 11. A graph showing the correlation between average air temperature and Lacertid observations for each survey undertaken in

28 Microhabitat The overall median basking temperature was 23 o C for all species involved (Anderson-Darling test for normality, p=<0.005, skewed ). For L. bilineata (n=92*) the minimum surface temperature recorded was 20 o C and the maximum was 27 o C (figure 12). This gave a median temperature of 22 o C. For P. muralis (n=75*) the minimum surface temperature recorded was 21 o C and the maximum being 28 o C. This gave a median temperature of 23 o C. Finally Z. vivipara (n=37*) had a minimum basking surface temperature of 20 o C and a maximum of 25 o C. The median of these measurements is 22 o C Median Temperature Lacerta bilineata Podarcis muralis Zootoca vivipara Figure 12. A graph showing the median basking surface temperatures for all three species of Lacertids. Also showing inter-quartile range and ranges. *Counts differ from previous total species figures given due to fact that temperature readings could not always be taken due to lizard positions i.e. basking on a branch. 22

29 These readings were subjected to a One-Way ANOVA test to look for any significance. Table 4 shows that when comparing the log temperature against the species f= 11.9 and p= <0.000, there was a statistically significant correlation between basking temperatures and the particular species in question. The mean temperatures and standard deviations show that L. bilineata and Z. vivipara are similar in their basking temperature requirements but P. muralis has a preference for hotter surfaces as their standard deviation falls outside that of the other two species. There were also two other microhabitat factors recorded, the first being whether the sites was wind sheltered. There were 125 sites which were wind sheltered and 133 that were not. Of the four different basking directions recorded the highest proportion for lizards were observed on the southerly side of the vegetation (n=150). The next highest number was on the south eastern side (n=93). There were also records on the easterly and south westerly sides of the vegetation, n=5 and 10 respectively. One-way ANOVA: Log Temp_1 versus Spp Source DF SS MS F P Spp Error Total S = R-Sq = 10.59% R-Sq(adj) = 9.70% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev Lb (-----*----) Pm (-----*-----) Zv ( * ) Pooled StDev = Table 4. ANOVA test showing results from one-way test when comparing log temperatures against species. 23

30 Vegetation From the visual survey of surrounding habitat L. bilineata showed five differing vegetation types, P. muralis had three non-vegetative substrates along with five vegetative types and Z. vivipara had a vegetative mix preference much the same as the two introduced species, see table 5. Lb Pm Zv Bare Ground Bramble (Rubus fruticosus) Fence Post Gorse (Ulex europaeus) Rye & Bent spp (Lolium & Agrostis) Holm Oak (Quercus ilex) Herb Spp Lyme Grass (Elymus arenarius) Marram Grass (Ammophila arenaria) Wall Table 5. Counts of all vegetative species and non-vegetative substrate recorded as a dominant feature in proximity to all Lacerta bilineata, Podarcis muralis and Zootoca vivipara sightings. Figure 13 shows the percentage of these vegetative species and non-vegetative substrates utilised by all three species. For L. bilineata one vegetative type stands out Percentage Bare Ground Rubus fruticosus Fence Post Ulex europaeus Lolium & Agrostis spp Quercus ilex Mixed Herb Spp Elymus arenarius Ammophila arenaria Wall Lacerta bilineata Podarcis muralis Zootoca vivipara Figure 13. Graph showing the percentage of vegetative and non-vegetative substrate preferences for Lacerta bilineata, Podarcis muralis and Zootoca vivipara. 24

31 as a favourite; the species being U. europaeus (63.5 percent of all sightings). The second closest preference being the grass mix (26.9 percent). Whilst P. muralis and Z. vivipara show a preference for two vegetation types (U. europaeus 39.1 percent and the grass mix 38.2 percent for P. muralis and 36.4 percent and 29.5 percent respectively for Z. vivipara) Z. vivipara also showed a preference for R. fruticosus (20.5 percent). These results from the lizard surveys were then followed up with further surveying concentrating primarily on the vegetation within the microhabitat of L. bilineata. From the ten sites visited with quadrats where L. bilineata had been observed twelve different plant species were recorded. The most prevalent species were U. europaeus and Agrostis spp. with counts of both species totalling seven. There was also a number of single, possibly random recordings of species such as Sweet Alyssum (Lobularia maritima) and White Stonecrop (Sedum album) within the locations where L. bilineata had been observed. In the quadrats where L. bilineata had not been sighted nine different species were recorded. Again here U. europaeus and Count Agrostis spp Rubus fruticosus Hypochoeris radicata Ulex europaeus Hedera helix Ammophila arenaria Species Rumex acetosella Plantago lanceolata Lolium spp Raphanus raphanistrum Lobularia maritima Sedum album Lb Present Sites 25 Lb Absent Sites Figure 14. A graph showing the occurrence of plant species in quadrats where Lacerta bilineata were present or absent.

32 Agrostis spp made up a large part of the composition of the vegetation (counts of 8 and 6 respectively). There were also a couple of anomalies recorded here too; they were Quercus ilex and Elymus arenarius. Overall for the species that were recorded in both the present and the absent quadrat surveys seven of them were found in both. The difference between counts of these species, when comparing them, ranged from 0 to 3. Figure 14 (previous page) shows that there was very little difference (counts of one to three) in the species located within the quadrats at both the L. bilineata present and the L. bilineata absent sites. Comparison in Vegetation Structure Between the Sites where Lacerta bilineata was Observed and from where it was Absent Count Scrub < 50 cm Scrub cm Scrub 180 cm + Trees Vegetation Description Lb Present Sites Lb Absent Sites Figure 15. A graph showing the vegetation structure at all twenty sites that were shown to have L. bilineata present and to have L. bilineata absent. As well as looking at the vegetation species a more detailed look at the structure of the vegetation over the whole survey area was also conducted. Figure 15 shows that of the ten sites where L. bilineata was observed four of them were over 180 centimetres, three of them between 50 and 180 centimetres, two under 50 centimetres and only one under trees. From the sites where L. bilineata is absent the vegetation 26

33 was all under 180 centimetres and half of that was less than 50 centimetres, showing a moderate avoidance of low scrub cover by L. bilineata. 27

34 Discussion Current Distribution of Non-Indigenous Lacertids The results of this investigation have shown that both populations of L. bilineata and P. muralis are well established at Boscombe and have been breeding successfully and increasing their range since their introduction in 2002, as shown in figure 16. As explained in the Method the optimal habitat for all the Lacertids Figure 16. Maps created in MapInfo showing the expansion of L. bilineata (green spots) and P. muralis (yellow spots) over the five year period ( ). Z. vivipara (brown spots). 28

35 surveyed is more likely to have been the cliffs themselves, as they are south facing, at a 45 o angle, comprising of patchy vegetation with patches of bare sand in between. It appears that when the cliff population reaches carrying capacity they spill out onto the cliff top looking for new territory. This can be seen to the west of the site where L. bilineata has colonised the Marram Grass verges to the path adjacent to the cliff edge. Elsewhere on the site L. bilineata seems to have been slower to colonise the patches of gorse away from the cliff edge and central introduction site; but is now slowly starting to do so. Even though there are only short distances between vegetation patches the intense anthropogenic pressure may limit its rapid spread; but not stop it. P. muralis is still only found on the cliff top around the original introduction site and has not spread too far from there. The reasons for this may be due to the habitat preferences; this will be dealt with in more detail later on. The short distances to the gorse seem to have been no problem for them to achieve, it seems that when longer distances over open ground have to be made they are more reluctant to make the journey. Perhaps if the population were to grow higher putting extra pressure on territories this gap would be bridged; this may be the reason for the occasional lone sighting of P. muralis in Again it should be emphasised that P. muralis has successfully spread farther west and east from the introduction site along the cliffs where there is nothing to impede its spread; proving that it does have the capability to increase its territory size where it is available. Table 6 shows that the counts of both L. bilineata and P. muralis have also increased over the five year period; but only at Species 2002 (7 visits) 2003 (11 visits) 2004 (42 visits) 2005 (32 visits) 2006 (10 visits) 2007 (15 visits) Lb Pm Table 6. Counts of both L. bilineata and P. muralis from 2002 to

36 face value. The visiting pressures have, however, differed greatly over this time. Before the discovery of the two introduced species in 2002 there was little or no interest in this stretch of coastline, herpetologically speaking. It was known that Z. vivipara occurred there, but being of no conservation issue regular surveys to the area were not done. This all changed with the discovery of the non-native Lacertids. There is therefore likely to be a bias in counts due to the differing numbers of surveys completed in a year. There was also no record made of individuals, meaning that double counting will almost certainly have occurred. This means that population estimates would be inaccurate if they were worked out from these records but trends in the population could be seen if compared to the number of visits. These count records are reflected in the size of the range of each species over the same time span i.e count of P. muralis equalled 3; the smallest area covered, whereas 2007 equalled 112; the largest area covered. Since their introduction in 2002 both L. bilineata and P. muralis have increased their range over the cliff top. L. bilineata has spread the farthest due to its ability to colonise the Marram Grass on the site, whilst P. muralis is still restricted to the central vegetation patches. The populations of both species have increased over this time as well, as dispersing juveniles have colonised new areas on the site. The Chi-Squared test (page 18) performed in the results shows that the population of L. bilineata has increased as expected (except in 2004; probably due to the exceptional number of site visits that year) and has continued to reproduce steadily over the five years. These results are to be expected from an animal that has been moved to what would likely be the far edge of its range (if the English Channel did not exist) i.e. some years it will do well and others badly, but overall, without some major stochastic event, its population should tick over evenly given the longevity of 30

37 individual lizards. The same could be said for P. muralis but the Chi-Squared test showed a big fluctuation in its results for this species. Again there was the 2004 anomaly (explained by visit numbers) but there was also a large χ 2 value given for This is likely to do with the breeding success of the species. As P. muralis is oviparous it relies on the external environment, in this case the number of sunny days, to incubate its eggs and as the summer of 2006 was particularly good it seems likely that P. muralis had a very successful breeding season, which is reflected in the large χ 2 value calculated. Why L. bilineata did not have this increase in recruitment too would require further investigation on the oviposition sites of this species at Boscombe, but is likely to be due to differing incubation parameters of this species. Effects on the Zootoca vivipara Population As it has been previously mentioned, making effective estimations of the populations of Lacertids from the current records would be risky due to the number of biases present. Again though, assumptions can be made on what effect the increasing numbers of non-indigenous Lacertids might be having on the native Z. vivipara found on the site. Both figures 8 and 10 (pages 17 and 20 respectively) in the results show an apparent rise in non-native Lacertids against Z. vivipara; but does this really show a drop in their population? There is no doubt that L. bilineata and P. muralis have increased their numbers but there is only circumstantial evidence that this is having a serious effect on Z. vivipara numbers. The dip in the population may only be a natural fluctuation and the results shown as a percentage against one another are always going to suggest an apparent decline for populations of a certain species that do not occur in large densities against those that do. It could also be argued that this investigation was biased towards seeing L. bilineata and P. muralis due to the times when the surveys were completed, i.e. warm sunny days that were ideal for both 31

38 introduced species. All the surveys were conducted with this fact in mind and that fact applies to past records also. This is why these Z. vivipara records are comparable as Species 2002 (7 visits) 2003 (11 visits) 2004 (42 visits) 2005 (32 visits) 2006 (10 visits) 2007 (15 visits) Zv Table 7. Counts of Z. vivipara from 2002 to 2007 past observations and those of this investigation were conducted under the same conditions. Table 7 shows the counts of Z. vivipara over the five years from 2002 to At face value there is not much change between the 2002 and the 2007 counts, but it is this greater number of site visits which accounts for the increased observations of Z. vivipara. One particular a) comparison is worth making between 2004 (which had many visits) and 2007, is that of the area with Marram Grass lining the footpath to the west of the site. In 2004 many b) Z. vivipara were recorded here, but by 2007 hardly any records were made in the same Figure 17. An aerial map showing the edge of the L. bilineata expansion, between 2004 (a) and 2007 (b), across the site pushing out Z. vivipara in the process. area (figure 17). This point seems to represent the current edge of the expansion of L. bilineata and it is only a short distance north to the next patch of Gorse where Z. vivipara is now found in large numbers. The Chi-Squared test on page 18 gives the greatest χ 2 values to Z. vivipara, which suggests they have had the greatest change in population over this period. This is likely to be due to the fact that the test would have expected more sightings in 2002 of the non-natives but they had not reached this area yet and would therefore not have been able to affect the test. So it appears that the introduced Lacertids are not 32

39 affecting Z. vivipara numbers until you look at the thematic distribution maps of all three species (page 15). These tell a different story about the interaction between L. bilineata, P. muralis and Z. vivipara where Z. vivipara is being pushed to the periphery of the site where the introduced species do not occur. If this continues, with Z. vivipara being pushed to the edge of the expanding wave of non-natives, eventually due to urbanisation isolating this coastal strip, making it finite, the non-natives will also spread to its end and Z. vivipara is unlikely to be able to compete. All three of these species share a pretty similar ecology, the main difference being that Z. vivipara, as its name suggests, is viviparous. This adaptation gives it an advantage in colder climates as it can reproduce more successfully even if there is a poor summer. This is backed up by Uller and Olsson s (2003) study of Z. vivipara in Sweden where they showed that this lizard was well adapted for endurance at relatively low temperatures. This study was, however, undertaken on the extreme of their natural range and the environmental pressure exerted there would be greater than those they would experience in Great Britain, but it shows they can compete under harsh conditions. If this is the only advantage Z. vivipara has over the two other Lacertids, in light of climate change leading to a rise in average temperatures meaning better incubation success for them, Z. vivipara is unlikely to be able to compete successfully in the long run or will only be found in small numbers along the cliffs at Boscombe. This inability to compete at a low density was recorded by Münch (2001) who describes how an introduced population of P. muralis displaced native populations of Z. vivipara and L. agilis in the Dortmund area of western Germany; just outside the native range of P. muralis. Both these native species were however present only in small numbers prior to the release of P. muralis. 33

40 A study of introduced Lacertids by Burke and Deichsel (2008) recorded seven successful introductions of P. muralis and one introduction of L. bilineata in North America. Most of these introductions have been deliberate and have been met with little concern to the possible effects on the local herpetofauna; introduced reptiles in the north urban areas offer excellent opportunities for study by amateur herpetologists, because such species are often common, conspicuous and subject to few legal restrictions. In most of these cases the introduced species are limited by the surrounding habitat being unsuitable and some populations are so small they are likely to become extinct anyway. There are other locations, however, where there is possibility for territory expansion and this could affect native reptiles. In this instance P. muralis and L. bilineata have been moved across biogeographic realms, but this is not the case at Boscombe. This does however demonstrate another example of the ability of these species (especially P. muralis) to colonise new areas and possibly affect native species in the process. If the current trends continue at Boscombe Z. vivipara will soon be at smaller densities too and it is possible they also could go the way of the Dortmund population and they only had to compete against the one introduced species. With all the evidence of the success of colonising new areas the future does not look good for Z. vivipara, nor does it bode well for the L. agilis population farther west of this site, as they too are only present in small numbers. 34

41 Limiting Factors on the Distribution of the Non-Indigenous Lacertids The habitat at Boscombe is extremely good for lizards, but there are however still annual or daily fluctuations in climate and the environment that could act as limiting factors to the distribution of the Lacertids. Environmental Factors When the average air temperatures for each survey were compared to the numbers of different species being observed, some of the species showed a strongly significant correlation between these two factors. L. bilineata was shown to be the only species of the three to have a significant dependence on air temperature. Although found on the Channel Islands and the north coast of France L. bilineata does not naturally occur this far north and being a larger animal would need a longer period to thermoregulate. This could, however, lead to a bias in sightings of different species on the same day, i.e. good surveying conditions for L. bilineata might not have picked up as many P. muralis and Z. vivipara, as they had already reached an active temperature and would no longer be basking and therefore harder to see. This dependence on warmer temperatures is the most likely reason why other introductions of this species have failed over the years but, as already mentioned, the cliffs at Boscombe are almost unique in Great Britain due to their geographical location and topography and have proven more than adequate for supporting populations of L. bilineata. The other two species of Lacertids (P. muralis and Z. vivipara) did not have the same high significance in correlation between sightings and air temperature. Both are smaller than L. bilineata and would therefore not need the same length of time to reach an active body temperature and they are also both far more tolerant to colder 35

42 conditions than L. bilineata. P. muralis has even been observed during the winter months basking in the sun next to patches of frost at Boscombe (Gleed-Owen, 2007). As Z. vivipara is native to the British Isles it is more tolerant of low temperatures. This means that it is unlikely that higher temperatures (those most commonly recorded during the surveys) influenced the activity to the same extent in P. muralis and Z. vivipara as it did in L. bilineata. Whether this investigation was therefore biased to seeing L. bilineata may have been likely as the average air temperature for all fifteen surveys was recorded at between 18 o C and 22.5 o C, but as all the surveys were undertaken in this small range of 4.5 o C variation the sightings are still comparable to the other species, as the conditions were near enough consistent. These were only average temperatures, too, which were taken over roughly an hour (the time it took to complete a survey) so there would have been enough variation over this timescale, and differing survey dates, to eliminate bias towards a particular species. Overall, as would be expected for ectothermic species, temperature will play a large role in activity and ultimately breeding success. A study of an accidental introduction of P. muralis on Vancouver Island, British Columbia, Canada, (Allan et al, 2006) showed that here temperature was a key influence on their breeding success. The native species of concern here was the Northern Alligator Lizard (Elgaria coerulea). The Wall Lizards here escaped on, in terms of latitude, the far northern limit of their equivalent climatic range in Canada. They have been able to breed successfully on Vancouver Island but experienced the same limitations that the population in Boscombe will encounter in terms of temperature, although to a greater extent. This limiting factor (temperature) seems to be the main barrier to their spread on Vancouver Island and as Northern Alligator Lizard is viviparous (the same as Z. vivipara) they, too, have the ecological edge over their non-native competitors. 36

43 Unfortunately for the native species at Boscombe the effect of temperature, as it currently occurs, is not felt to the same extreme as it is in Canada and although it has an effect on activity it puts little pressure on their breeding success. Rainfall (which is predicted to increase over the summer months as a result of climate change) may act more as a limiting factor on these oviparous species as their eggs will not incubate successfully. This year (2007) three of the surveys were conducted in September, when the young of that year would have hatched. No L. bilineata juveniles were recorded and very few P. muralis; in contrast to juvenile Z. vivipara which were seen in greater numbers. This contrasted with the sightings of first year juveniles, for both L. bilineata and P. muralis, recorded at the beginning of the season; they were found in high numbers after the especially dry summer of Van Damme et al (1992) showed that differing incubating temperature can have dramatic effects on recruitment for P. muralis. From their study they demonstrated that for embryogenesis to occur the eggs required temperatures of between 24 o C and 35 o C (with high mortality at both extremes) at varying durations (days with sunlight). The higher temperatures required only 12.5 days before hatching occurred whilst the lower temperatures needed much longer, with up to 73.6 days at 24 o C. As the summer of 2007 was a particularly poor one, in regard to the number of sunny days, this is bound to have affected recruitment for P. muralis at Boscombe. Another influence temperature has, which relates to the overall fitness of P. muralis, was demonstrated by Braña & Ji (2007). They demonstrated that differing temperatures during incubation resulted in different phenotypes (head, tail, limb proportions) all of which could disadvantage P. muralis in the wild. The unpredictable weather of current years could therefore be affecting P. muralis in this country and lessening its competitive advantages over the other species, but there is no sign of any competitive edge loss at Boscombe. Z. vivipara, 37

44 being viviparous, does not have these reproductive environmental constraints due to its being able to self-regulate incubation temperature by its own position in the environment (although still reliant on weather conditions). This is, however, aided by a varied topography, which the habitat at Boscombe provides. The other two environmental factors recorded during this investigation were wind strength and cloud cover. The wind strength was not significant in influencing the number of sightings, but this was due to the fact that surveys were only conducted on days when conditions were deemed suitable. Days with greater wind strength would most likely have produced poorer sighting numbers, but those days were not used for surveying. The cloud cover just failed to reach significance, but again the days that were deemed suitable for observing Lacertids were chosen, not giving a true range of factors against which to compare sightings. There was the slight significance, however, that due to the thermoregulatory needs of the species concerned, cloud free days are better for quick achievement of optimal body temperature. The climate at Boscombe also directly affects the microhabitats occupied by all three Lacertids and this in turn directly affects their ecology and breeding success. The air temperature findings demonstrated an effect on activity levels from species to species; basking temperature (substrate) too can have a similar effect. L. bilineata and Z. vivipara showed lowest substrate temperature requirements whilst P. muralis showed the highest. These temperature requirements again are a reflection of each species thermoregulatory needs and adaptations to the local environment. The ANOVA test confirmed this correlation between species and basking temperature, emphasising the fact that P. muralis requires warmer basking sites than the other two species. As this site and the surrounding habitat is composed of free draining sandy soils, which are prone to warming quickly in direct sunlight, there are numerous 38

45 suitable basking locations for all the species concerned. This, combined with the patchy vegetation on the cliffs and adjacent cliff top, means there are no physical landscape barriers to impede their spreading in both directions along the cliffs for many kilometres. Two other features of the microhabitats where the Lacertids were located were also recorded; whether the site was sheltered from the wind and the direction the site was facing. There was a rough split, half and half, for all the sites with regard to whether it was wind sheltered or not. It is likely that even if a site was exposed to the wind, from the recorder s point of view, the lizard, only a few centimetres off the ground, would not have been affected unduly by the wind when it was in amongst the vegetation. Whether the sightings of Lacertids varied due to the basking site being wind sheltered would also be affected by the direction of the site, as there is no point for an ectothermic species to choose a site that is sheltered from the wind if it is also shaded from the sun. The benefit of choosing a site in the sun is greater than that of choosing a site sheltered from the wind. This leads on to the next microhabitat feature; basking direction. Again here the position of the sun would dictate the location of the basking site, depending on the time of day and the time of year. This means that the surveys would have to have been completed at differing times of the day to comprise a true representation between basking direction and sighting success. The survey route also did not encompass all sides of the vegetation and therefore these results will always be biased if any statistical test were to be run on them. Overall temperature plays the biggest part in the environmental factors that could limit the spread of the non-native Lacertids. One aspect of this is the local area and surrounding habitat which must have locations for thermoregulation and suitable oviposition sites; these requirements are all met here meaning that this is no barrier to 39

46 their spread. The other aspect is local weather conditions, which are under great debate as to what the future will hold for the climate here and all around the world. From the latter of the two aspects there could be two outcomes that would have differing effects on the non-native Lacertids. If there are hotter drier summers their recruitment will be successful and they will easily expand their range, but if summers are wetter they are unlikely to fare as well. Where there is prey, predators too will be found and they are an essential part of all ecosystems. The main source of predation on Lacertids at Boscombe is likely to come from avian predators such as Larids, Corvids and Kestrels (Falco tinnunculus). There may also be cases of domestic cats coming across from the urban dwelling close to the site, but due to the high pressure of dog walkers this is not very likely. The dogs themselves seemed to pay little attention to the basking lizards and the lizards seemed not to mind the dogs either. Over the course of this investigation only one L. bilineata was observed with an autotomised tail suggesting that predation attempts of the lizards may not be that commonplace. This is most likely due to the spiky Gorse thickets preferred by all species when basking. Vegetation and Management on Site As in most habitats vegetation is a vital component in determining the species composition to be found there. Whether it was the species of vegetation or the structure of the vegetation present at Boscombe that could provide clues to further limiting factors to the spread of the non-native Lacertids was focused on in this investigation. When comparisons were conducted on the vegetative species composition between sites where L. bilineata was observed, and where they were not, it was clear that these lizards were not biased towards a particular species mix. The habitat at 40

47 Boscombe is not as botanically rich in comparison to other habitats, for example, chalk downland, due to the marine environment leading to only hardy species growing there. The only large trees are introduced species such as Holm Oak (Quercus ilex), Sycamore (Acer pseudoplatanus) and Pine (Pinus spp) but there are however large amounts of Gorse scrub and many differing herbaceous species forming a solid understorey. In between the scrub there are open patches of grassland of a long sward. This vegetation species range supports a large variety of invertebrates which in turn means there is plenty of prey for the lizards. As the whole cliff habitat from Poole Harbour eastwards to Hengistbury Head is very similar, there is always going to be plenty of prey available to the non-native Lacertids. Due to there also being little change in the soil composition over the area the species mix does not a) b) Figure 18. Structural map (a) overlaid with species observation map (b) showing preference for higher scrub cover. 41

48 change much either, meaning that the species of vegetation over the whole site and beyond are suitable for sustaining Lacertids. It was however still clear that there were some areas where the non-native species had not yet colonised, so instead of looking at the plant species the structure of the vegetation was investigated instead. When a structural map of the site was overlaid with the spot map generated in MapInfo, showing the differing species, it was clear that the higher scrub patches supported the higher numbers of Lacertids (figure 18, previous page). The native habitat requirements for L. bilineata and P. muralis, described by Street (1979) and Arnold and Ovenden (2004), state that they both can be found in a wide variety of habitats. P. muralis is the more urbanised of the two species but still is found in greater numbers in more natural surroundings. The dominant common factor which both species require is that of hot dry areas within the habitat where they can thermoregulate and oviposit. These areas are going to be determined by the shade of the surrounding vegetation. As Gorse grows older it becomes leggy, which in turn opens up the ground to the sun making such sites described above available whilst still providing the protection of cover. This structural nature of the gorse has dictated the spread of the non-native Lacertids from the central introduction site on the cliff top. They have spread farther along the cliff face where access does not allow the same clearance of scrub that occurs on the cliff top leading to optimal habitat being found all along the length of the cliff face. Furthermore the coastal conditions act as management here, firstly wind and salt not allowing large trees to block out the light and erosion keeping areas of open ground available. There is one exception to this spread, dictated by gorse structure on the site, and that is a number of L. bilineata found within the Marram Grass (Ammophila arenaria) tussocks to the west of the site. Here the Marram has facilitated their spread along the cliff top, making use of another aspect 42

49 of the habitat which has yet been largely ignored by P. muralis. Here L. bilineata seems to have successfully colonised the shorter vegetation; but the Marram acts just like a smaller version of the gorse providing thick cover from predators, plenty of prey and sunny spots for basking. This again provides extra evidence that the structure of the vegetation is a limiting factor to the spread of the non-natives as the Marram Grass provides a good halfway point for dispersal to the vegetation beyond, but they have not made that small distance successfully yet. When the Gorse on the periphery grows to a suitable size (not the thick closed canopy condition in which it grows when young) it will only be a matter of time until it is colonised by the non-native Lacertids already so close to it. The structure of the vegetation at Boscombe is determined by the management activities of Bournemouth Borough Council. Due to the high human presence at this site a certain amount of habitat management is conducted in order to maintain its appearance (Clark, 2007). On the whole the site is left to succeed on its own with the only intensive management consisting of mowing grass close to the path and maintenance of public facilities, such as park benches. The Gorse is cut back when it becomes too big, but this is not done to an annual management plan. As they have a limited budget there are not always the resources available to do the work and the Gorse is allowed to overgrow. This overgrown state appears, however, to be the preferred successional stage of the non-native Lacertids and the management is unknowingly limiting their distribution when it does occur; as is shown by the Gorse in the north west corner of the site (see figure 17) which was cut back a few years ago and has not been colonised by either L. bilineata or P. muralis. 43

50 Possible Effects on Lacerta agilis The introduction of L. bilineata and P. muralis onto the cliffs at Boscombe has already started to show signs that they are affecting the native Z. vivipara there, but an issue that is of greater conservation concern is the effect they may have on the population of L. agilis to the west of this site (figure 19). Populations of L. agilis have Figure 19. Location of non-native species and L. agilis colony. declined dramatically in Great Britain over the last century, mostly due to habitat loss. As a result they are protected from capture, killing, disturbance and trading as a listed schedule 5 species on the Wildlife and Countryside Act 1981 (House & Spellerberg, 1983). There have already been sightings recorded of P. muralis close to the site but as yet L. bilineata has taken longer to spread that far. The only real barrier to their spread is the car park at the base of Boscombe pier; but taking into account both these species tolerance of humans and ability to expand their territories this would present little problem to them. So if, or when, they do reach the colony of L. agilis, what will be the likely effect? The best way to try to second guess these effects is to look at and compare the different species ecologies and see where any competition might occur. Firstly, there are the habitat requirements of the three species in question and it has already been established that the non-native species are quite capable of breeding 44

51 successfully in the available habitat all along the cliffs at this site. L. agilis is normally associated with lowland heath in the south of England but it is also at home in the sand dunes of Merseyside. The common denominator at both of these sites is sandy soils; the latter of these two national locations represents the habitat at Boscombe more closely. The same limiting factor the non-natives Lacertids have encountered in this country, i.e. that of a warm enough substrate in which their eggs can be incubated, also applies to L. agilis. This means there will be competition between the differing species for suitable habitat which would encompass the following ecological needs: male territories, oviposition sites and a source of food. As all three species take similar food items, invertebrates, there would also be extra competition here too. Even though this cliff habitat is extremely rich in invertebrate prey it would be unlikely to support four different Lacertid species in the one place at the current predicted expansion rate. Münch (2001) showed that if these introduced species are competing with our native Lacertids, when they occur at low numbers, they will cause their local extinction. The only advantage L. agilis (and Z. vivipara) will have on the two nonnative species is the fact that they are both found farther north over their whole range and are therefore better adapted to colder climates. The unpredictable weather of the future may be enough to let them hold out at this site during those wetter years when the non-natives can only achieve poor recruitment. Overall L. agilis is likely to suffer a dramatic loss in population when, or if, (most likely when), the non-natives reach them. Whether they will be able to compete in the long term against them is possible but unlikely, but only time will tell. 45

52 Critique As with all investigations there are always some aspects that could be changed with the benefit of hindsight; this investigation being no exception. It would have been beneficial to have completed some surveys outside of the normal April to September season to try to establish each species extreme tolerances to environmental conditions. This, however, would have increased the number of unsuccessful visits to the site (i.e. no Lacertids being recorded) and would not have been practical due to the distance that had to be travelled to reach the site. The surveys could also have been completed at differing times of the day to try to eliminate bias towards a certain species. The question of whether L. bilineata suffered poor recruitment in 2007 could have been enforced in 2008 by surveying at the start of the season for any juveniles of this species. An absence of juveniles would back up the belief that poor summers (in terms of hours of sunlight, not temperature) exert a major limiting factor on L. bilineata, as was suspected during this investigation. More sampling of vegetation when investigating at the species level could have allowed more statistical tests to have been run on the data collected. With more samples better correlations and significances could have been established between vegetative species and lizard species, if there were any. This could have given greater insight into further limiting factors affecting the spread of both introduced species. There are also other ecological factors that could have been looked at to try to establish possible limiting factors affecting the Lacertids, such as diet, hibernacula sites, male territory size and population census using marked recapture techniques. All of these would have required a longer period of study than this investigation would have allowed but could be used to improve future research on the subject. This longer study time (for example, three to five years) would be the key to establishing 46

53 the long term effects on the native populations of Lacertids. As long as the same static path and other survey parameters (number of surveys, weather conditions, etc) were observed a comparison of the population for all species involved could be made giving good indicators of increases or decreases in numbers. This in turn would give stronger evidence to the effects felt by Z. vivipara by the introduction of L. bilineata and P. muralis. Conclusion This investigation has answered the questions raised at the beginning of the study, but has also raised new ones along the way. It has been established that since their introduction in 2002 both L. bilineata and P. muralis have increased their range well beyond their initial introduction site and have bred successfully over this time too. Their increasing numbers do appear to have affected the native Z. vivipara found on site, pushing them outwards from the central area now predominantly occupied by the non-native species. These periphery areas are now the only places where Z. vivipara can be found in high numbers and once the non-natives have increased their range to the edge of the available habitat Z. vivipara will no longer be common on Boscombe cliffs. It is clear that the local topography and climate are satisfactory for all three Lacertids species at Boscombe, providing them with all their ecological needs. This habitat also extends beyond this investigation s area meaning there is suitable habitat for further expansion of their territory. The only limiting factors affecting their spread seem to be climate and management of the site. The climate, which affects recruitment rates, is not predictable and this will be even more influential in the future. The management is not directly aimed at limiting the spread of these non-native Lacertids but when the Gorse is cut back it does limit the suitable habitat available to them. There is, 47

54 however, always going to be suitable habitat available on the cliff face due to the difficulty in accessing it to perform management. From here the non-native species have, and could, spread farther, all along the cliffs to the east and west, with very few obstacles in their way to stop them. During this spread they will reach the L. agilis colony that is located just to the west of Boscombe pier and it is likely that these lizards will suffer the same fate as Z. vivipara. It is unlikely that the introduced species could now be eradicated from the cliffs, not without some kind of total clearance of all the vegetation; this, however, would impact on many other species. Even if all the scrub were to be removed from the cliff top, the cliff face (which would be extremely difficult and dangerous to clear) would act as a source population for re-colonising the cliff top. The only thing that can be done is to monitor the remaining native species and hope they can compete. Z. vivipara is however not of priority for conservation in Great Britain (although it is now listed as a B.A.P. [Biodiversity Action Plan] species) and neither is L. agilis, on a global scale. This case does, however, provide another example of how illegal anthropogenic interference in an ecosystem could have adverse long term effects, not just on the congeneric species but also on other taxa found there. Some introductions are neutral, most die, but it is often too late for native species by the time invasive species are established. Is it worth the risk? L. bilineata and P. muralis are now well established on Boscombe cliffs and have been shown to be detrimentally affecting the population of Z. vivipara; meaning that the null hypothesis can be rejected. This investigation will also help to map their spread for any further research and could be used to predict trends in the populations of the non-natives and native species. On a global scale this is another good example of why the moving of species outside their natural range is never a good idea. 48

55 References Allan, G. M., Prelypchan, C. J. & Gregory, P. T. (2006) Population profile of an introduced species, the common wall lizard (Podarcis muralis), on Vancouver Island, Canada. Canadian Journal of Zoology Arnold, N. & Ovenden, D. (2004) A Field Guide to the Reptiles and Amphibians of Britain and Europe. 2 nd edition. Great Britain. Harper Collins Publishers Ltd. Beebee, T. & Griffiths, R. (2000) Amphibians and Reptiles. Great Britain. Harper Collins Publishers Ltd. Braña, F. & Ji, X. (2007) The selective basis for increased egg retention: early incubation temperature determines hatchling phenotype in Wall Lizards (Podarcis muralis) Biological Journal of the Linnean Society Burke, R. L. & Deichsel, G. (2008) Lacertid Lizards Introduced into North America: History and Future. Unpublished. Clark, S. (2007) [discussion on management practices on Boscombe cliffs] Personal communication, 20 th November Deichsel, G., Gleed-Owen, C. P. & Mayer, W. (2007) Lacerta bilineata (Western Green Lizard) and Podarcis muralis (Common Wall Lizard) United Kingdom, Dorset. Herpetological Review. 38 (1) Gleed-Owen, C. P. (2004) Green Lizards and Wall Lizards on Bournemouth cliffs. Herpetological Bulletin Gleed-Owen, C. P. (2007) [discussion on non-native Lacertids activities] Personal communication, October Gubanyi, J. (2000) A breeding Colony of Western Green Lacertas (Lacerta bilineata) Confirmed in Southwestern Topeka (Kansas). Transactions of the Kansas Academy of Science. 103 (3-4), House, S. M. & Spellerberg, I. F. (1983) Ecology and conservation of the Sand Lizard (Lacerta agilis) habitat in southern England. Journal of Applied Ecology Joint Nature Conservation Council. Wildlife and Countryside Act 1981, Chapter 69 [online] Available from: [Accessed 20 November 2007] Lever, C. (1979) The Naturalized Animals of the British Isles. Great Britain. Paladin. Meffe, G. K., Ronald Carroll, C. et al. (1997) Principles of Conservation Biology. 2 nd edition. United States of America. Sinauar Associates Inc. 49

56 Münch, D. (2001) Do allochthone Common Wall Lizards jeopardize autochtone Sand and Viviparous Lizards? Dortmunder Beiträge zur Landesskunde Patrick, L. (2001) Introduced Species Summary Project, Brown Tree Snake (Boiga irregularis) [online]. U.S.A.: Columbia University. Available from: irregularis.html [Accessed 19 November 2007] Quayle, A. & Noble, M. (2000) The Wall Lizard in England. British Wildlife. 12 (2) Savidge, J. A. (1988) Food habits of Boiga irregularis, an introduced predator on Guam. Journal of Herpetology. 22(3) Street, D. (1979) Reptiles of Northern and Central Europe. Great Britain. Batsford Ltd. Uller, T. & Olsson, M. (2003) Life in the land of the midnight sun: are northern lizards adapted to longer days? Oikos. 101 (2) Van Damme, R., Bauwens, D., Braña, F. & Verheyen, R. F. (1992) Incubation temperature differentially affects hatching time, egg survival and hatchling performance in the lizard Podarcis muralis. Herpetologica. 48 (2) Walker, Z & Deichsel, G. (2005) Podarcis muralis (Common Wall Lizard). USA: Indiana. Herpetological Review. 36 (2) Wycherley, J & Anstis, R. (2001) Amphibians and reptiles of Surrey. Great Britain. Surrey Wildlife Trust. 50

57 Female Green Lizard (Lacerta bilineata) in Gorse (Ulex europaeus) thicket. Snout to vent length normally 13 centimetres, tail twice as long as body. Note twin dorsolateral lines which give this species its scientific name. Male Green Lizard in Marram (Ammophila arenaria) tussock. Same length as female but with larger head and swelling at base of tail. Note blue throat patch possessed by males in breeding condition. Immature Green Lizard showing intermediate markings between juvenile and adult form. Juvenile Green Lizard basking on Bramble (Rubus fruticosus) leaves. Chocolate brown in colour with green throat. Note the twin dorsolateral lines appearing down the back. Total length of juveniles 10 centimetres. Appendix 51 1

58 Female Common Wall Lizard (Podarcis muralis) climbing wall. Mostly brown in colour with mottled sides, also showing slight dorsal stripe. Note the tail length which is over double the body length. Male Common Wall Lizard basking on wall. Up to 7.5 centimetres snout to vent length in adult animals. Males have a green mottled back pattern; there is also a brown form in parts of this species range. It has a wide distribution in Europe from Northern Spain east to Bulgaria. Simon Mol Immature Common Wall Lizard emerging from vegetation. Starting to show adult patterning but lacking colouration as yet. Juvenile Common Wall Lizard basking on rock. Good climbers at all ages these lizards are quite at home on artificial substrates. Total length of juveniles 7 ti t Appendix 52 2

59 Female Viviparous Lizard (Zootoca vivipara) basking. Both sexes are brown in colour with females normally possessing dorsal and lateral striping. Snout to vent length 6.5 centimetres, with a smaller tail proportionately compared with Common Wall Lizard. Male Viviparous Lizard showing spotted pattern. Males are the same length as females but possess larger heads and swelling at the base of the tail. Juvenile Viviparous Lizards are very small, 5 centimetres at birth. They are a metallic brown / black in colour. Although not seen within this investigation s survey area the Sand Lizard (Lacerta agilis) has a colony to the west of the site in Boscombe and is of most concern with regard to the spread of the two introduced Lacertid species. The male (left) and the female (right) are pictured above. Appendix 53 3