General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (Rev. 3/06) Page 1 of 40

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code PN0921

2. Project title

Assessing pesticide risks to amphibians and reptiles

3. Contractororganisation(s)

Central Science LaboratorySand HuttonYorkYO41 1LZ          

54. Total Defra project costs £ 18,631(agreed fixed price)

5. Project: start date................ 08 January 1997

end date................. 31 March 1998

SID 5 (Rev. 3/06) Page 2 of 40

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

1. This study reviews the behaviour of amphibians and reptiles in the agricultural environment, assesses their potential exposure to pesticides, and examines critically whether they are protected by risk assessments conducted for other types of animals.

2. The current status of reptiles and amphibians was reviewed and the results of the arable pond survey used to determine the distribution of amphibians in the arable habitat.

3. Of the holdings with ponds, 266 (63%) indicated that frogs were present. This figure is equivalent to over a third of all arable farms surveyed not only having ponds but having frogs present as well. A total of 1,083 ponds were recorded on the 266 holdings with frogs present.

4. Of the holdings with ponds, 189 (45%) indicated that toads were present. This figure is equivalent to almost a quarter of all arable farms surveyed not only having ponds but having toads present as well. A total of 892 ponds were recorded on the 189 holdings with toads present.

5. Of the holdings with ponds, 159 (38%) indicated that newts were present. This figure is equivalent to a fifth of all arable farms surveyed not only having ponds but having newts present as well. A total of 741 ponds were recorded on the 159 holdings with newts present.

6. Toxicity data was obtained for use in comparisons with fish and birds. The majority of data relevant to UK species was for aquatic phases with few data for standard 96h LC50 studies, most of which were for the introduced Xenopus laevis (African clawed frog) and only four values for native tadpoles, three for Bufo bufo (common toad) and one for Rana temporaria (common frog). LD50 data for adult UK species was found for only five compounds. No toxicity data for any UK reptiles was found.

7. A comparison of LC50 data for fish and larval amphibians indicated that the uncertainty factor of 100 placed around risk assessments for fish would make them protective of amphibians on most occasions. It is therefore recommended that fish assessments are used to protect larval

8. There appears to be little correlation between the amphibian and birds data presented here indicating that bird toxicity may be a poor predictor. These data suggest that it would be unwise to use bird data routinely to represent amphibians for which there are no data without using a large uncertainty factor. While this might be conservative for some compounds such as OP

SID 5 (Rev. 3/06) Page 3 of 40

compounds, it may be less so for other compounds such as pyrethroids to which amphibians seem far more sensitive.

9. For reptiles there is even less data in the literature although there seems to be more similarity with values for birds at least for those compounds for which data is available. While bird data may be a better surrogate for reptile data, the very small data set does not allow any test of the robustness of such an approach.

10. Methods for estimating food intake of amphibians and reptiles based on published allometric equations for field metabolic rate, energy and moisture content of different foods and assimilation rates were developed. Estimates of exposure of frogs via this route were calculated along with toxicity exposure ratio (TER) values and compared with birds. Given the relative differences in food intake between birds and amphibians/reptiles it is likely that the apparent risk would usually be lower unless these species are more sensitive to the particular compound.

11. For amphibians especially it is likely that dermal exposure would be far more important and this route is currently not routinely considered for birds. Factors affecting the degree of dermal exposure (e.g. distance travelled, mode of locomotion) were considered and methods of estimating worst case dermal exposure based on simple assumptions about absorption developed. These were used to estimate exposure to two pesticides for which we have LD50 data for a UK species.

12. A possible risk assessment process based on these results was suggested along with recommendations about further work. This included obtaining more data for adult amphibians and reptiles, collecting data to refine the worst-case estimate of dermal exposure and the development of specialised tests where necessary.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

SID 5 (Rev. 3/06) Page 4 of 40

INTRODUCTION

EU Directive 94/414 requires that plant protection products have no unacceptable effects on the environment particularly on non-target species. To date terrestrial risk evaluations have concentrated on birds, mammals, earthworms, non-target arthropods and honeybees. There is concern over the decline of amphibians and reptiles in the UK agricultural environment. There are a number of factors which may contribute to this including changes in husbandry and habitat removal. However, declines in amphibian populations in North America have been suspected to be linked to pesticide usage. There is therefore a need to consider whether specific attention should be paid to amphibians and reptiles in risk assessment.

It is sometimes assumed that amphibians and reptiles are protected by risk assessments conducted for fish and birds. However, there are a number of reasons to question this assumption. For example there has been a decline in amphibian numbers on agricultural land (Hall and Henry 1992, Watt et al. 1995). Also, assessment of risks to aquatic organisms may be relevant to amphibians in water but adult frogs, toads and newts often forage on land and may thus be directly exposed to pesticides applied as granules or as sprays on the surface of crops or soil, or through consumption of contaminated invertebrates. Amphibians have a large surface area in relation to their size and their damp, highly permeable skin would potentially aid pesticide uptake. In addition, pellets and granules may adhere directly to the skin. Overland migrations of large numbers of amphibians from hibernation sites to breeding ponds during the spring months make them particularly susceptible to early season pesticide applications.

The diet of reptiles differs in composition from the birds most frequently considered in risk assessment and the potential for dermal exposure is presumably significantly higher for reptiles due to greater contact with residues on soil and plant surfaces. In addition, specialist amphibian-eating species, such as the grass snake, may be in danger from the cumulative effects of the consumption of pesticide-contaminated frogs, toads and newts. Amphibians and reptiles are rarely submitted to the UK Wildlife Incident Investigation Scheme (WIIS), which therefore has limited opportunity to detect pesticide effects on them. However, there have been two cases in which dead frogs were found to contain residues of metaldehyde.

This study reviews the behaviour of amphibians and reptiles in the agricultural environment, assess their potential exposure to pesticides, and examine critically whether risk assessments for fish and birds provide protection for amphibians and reptiles. This involved defining reasonable worst case exposure scenarios, identifying practical methods of assessing dermal and oral exposure, assessing correlations between existing toxicity data, determining risk for a small number of pesticides and comparing the results with existing assessments for fish and birds and examine critically whether they are protected by risk assessments conducted for other types of animals. The aim of the study was to assist PSD in deciding whether to increase the attention given amphibians and reptiles in pesticide approvals and if so what further research might be necessary to develop new risk assessment methods.

CURRENT STATUS OF REPTILES AND AMPHIBIANS IN THE UK

Indigenous species

The number of indigenous species of reptiles and amphibians within the UK is limited to two species of frog, two toads, three newts, five lizards and three snakes (Table 1). Most of these species are relatively widespread, but three are more restricted including Rana dalmatina (agile frog) which is found only on the Channel Islands.

Non-indigenous species

Two species of frog, Rana ridibunda (marsh frog), and Rana esculenta (edible frog), are possible introductions from Europe but this is currently under discussion. Similarly Rana lessonae (pool frog) has been recorded in Norfolk but is not included in the list above as it may be either an introduced population or a relict population from a time when its distribution was larger. All three species are commonly found on the continent; interestingly Rana esculenta is a fertile triploid hybrid between Rana ridibunda and Rana lessonae.

Other introduced species include Xenopus laevis (African clawed frog), Hyla arborea (striped green tree frog), Alytes obstetricans (midwife toad), Bombina bombina (fire bellied toad,, Bombina variegata (yellow bellied toad), Triturus alpestris (alpine newt), Natrix tessalatus (dice snake), Trachemys scripta elegans (red eared terrapin), and Podarcis muralis (wall lizard). The latter species is also naturally occurring on the Channel Islands, as Lacerta viridis (green lizard).

Protection of amphibians and reptiles under UK legislation

The Triturus cristatus (great crested newt) is in decline over the whole of Europe and is one of the key species identified in the Government Response to the UK Steering Group Report on Biodiversity. Other species included in

5

this list are Bufo calamita (natterjack toad), and the Lacerta agilis (sand lizard). Surprisingly Coronella austriaca (smooth snake), which does have protection under the Wildlife and Countryside Act is not included.

Although only four species are currently identified in the Government Response to the UK Steering Group Report on Biodiversity all species are protected in some way by various wildlife protection acts, directives and conventions (Arnold 1995). Many species, including Triturus cristatus, Bufo calamita, Lacerta agilis, Coronella austriaca and Vipera berus are showing a decline in range and population numbers throughout Europe (Corbett 1989). Similar results were published in 1982 showing declines in Triturus helveticus, Vipera berus and Natrix natrix between 1959 and 1960-1973 (Cooke and Arnold 1982). This trend was based on presence or absence within 10km squares and do not give any indication of relative abundance. Declines in the many of the commoner species of reptiles and amphibians in Britain have also been reported by Hilton-Brown and Oldham (1991). In the light of these widespread declines English Nature (now part of Natural England) provided funding for a number of translocation, captive breeding and management projects for the four protected species and for other species occurring in the same areas (Biodiversity: The UK Steering Group Report Volume 2: Action Plans, 1995).

Table 1. Amphibian and Reptile species present in the UK. Key - N = Native; I = Introduced; ? = Origin unsure (probably introduced). Letters in italics indicate no records since 1970. All other records indicate occurrences since 1970. References Arnold (1995), Corbett (1989)

England Wales Scotland Ireland C. Isles DistributionUrodelaTriturus alpestris (alpine newt) ITriturus cristatus (great crested newt) N N N WidespreadTriturus helveticus (palmate newt) N N N N WidespreadTriturus vulgaris (smooth newt) N N N N N WidespreadAnuraAlytes obstetricans (midwife toad) IBombina bombina (fire Bellied toad) IBombina variegata (yellow Bellied toad)

I

Bufo bufo (common toad) N N N N WidespreadBufo calamita (natterjack toad) N N N N RestrictedHyla arborea (striped Tree frog) ?Rana esculenta (edible frog) ?Rana lessonae (pool frog) ?Rana dalmatina (agile frog) N RestrictedRana ridibunda (marsh frog) IRana temporaria (common frog) N N N N N WidespreadXenopus laevis (African clawed frog) ISauriaAnguis fragilis (slow worm) N N N N WidespreadLacerta agilis (sand lizard) N N I RestrictedLacerta viridis (green lizard) I NLacerta vivipara (common lizard) N N N N WidespreadPodarcis muralis (wall lizard) ? NSerpentesCoronella austriaca (smooth snake) N RestrictedNatrix natrix (grass snake) N N N N WidespreadNatrix tessellatus (dice snake) IVipera berus (adder) N N N WidespreadCheloniaTrachemys scripta elegans (red-eared terrapin)

I

GLOBAL POPULATION TRENDS

Worldwide populations of amphibians are declining, mainly due to effects of mans activities and habitat destruction. However there have been marked declines in species occurring in relatively undisturbed areas and there is a possibility that this may be a response to widespread environmental degradation (Blaustein and Wake 1995).

6

Amphibian species are often suggested as being “useful indicator species” (Baker 1997). Some of these declines may be real or they may simply be a function of short term monitoring. Long term monitoring of species such as Triturus cristatus show large population fluctuations and the wrong conclusions may be reached using short-term studies (Cooke 1995). For a fuller discussion of the factors important in amphibian decline see Linder et al. (2003).

Because of the global concern for amphibian species a Declining Amphibian Populations Task Force (DAPTF) was set up for comment and dissemination of information. The DAPTF merged into the new IUCN/SSC Amphibian Specialist Group (ASG) in June, 2006 (www.amphibians.org.)

THE IMPORTANCE OF PESTICIDE APPLICATIONS TO INDIGENOUS HERPETOFAUNA

Cooke (1997) divides the problems associated with pesticide applications into two main categories, Direct and Indirect effects.

Within the Direct effects category he includes two incidents recorded under the Wildlife Incident Investigation Scheme, (WIIS), on both occasions metaldehyde was involved in a lethal poisoning of frogs (Greig-Smith et al. 1990, Fletcher et al. 1994). These cases may have been secondary poisoning where the target species, slugs, were ingested. Other potential problem areas are the use of fumigants, aluminium phosphide and sodium cyanide, for rodent and rabbit control in burrows being used as hibernacula or day time refuges for herpetofauna, aquatic or ditch side herbicides and their effect on tadpoles or breeding populations and the use of granular fertilisers, although not pesticides, are also important.

Indirect effects include the use of herbicides, the problem of spray drift and associated problems with loss of available poolside or terrestrial habitat.

RECENT SURVEYS OF UK POPULATIONS

There are a number of organisations, other than the Open University, who have previously studied, or are currently studying native amphibian and reptile species.

Leicester Polytechnic, now De Montfort University have been working for a number of years on amphibians. In 1986 they commenced a survey of Triturus cristatus populations in ponds and in 1993 they combined this with a survey monitoring all amphibian and reptile populations in and around ponds. Most of the data for these surveys were collected on a voluntary basis by amateurs from a number of herpetological societies (Swan and Oldham 1993a, 1993b).

John Moores University in Liverpool have recently initiated a Pond Life survey and a survey of Cheshire ponds. This survey has not been directly targeted at reptiles and amphibians but both groups are recorded where found.

Both the Joint Nature Conservation Committee (JNCC) and Institute of Terrestrial Ecology (ITE, now Centre for Ecology and Hydrology, CEH) have been involved in the pond surveys conducted by De Montfort University and have recently produced the Provisional Atlas of Amphibians and Reptiles in Great Britain (Arnold 1995). The bulk of the data has been collected by volunteers with the collation and interpretation of results conducted by the ITE.

The ITE was also involved in conducting surveys on comparative changes in pond numbers in lowland Britain. The base line for these surveys is a Countryside Commission survey conducted in 1984, recording the number and location of ponds within 10km grids in lowland Britain. Comparisons are also made between ponds in arable and grassland environments. In order to establish changes in pond numbers the DOE has funded surveys of ponds only in 1990 and hedgerows and ponds in 1992/1993. A further survey on the quantitative assessment of ponds in the same area is planned.

The pesticide usage survey has conducted an assessment of the number of ponds and pond management during the arable surveys of 1994 and 1996. In the 1996 survey amphibian presence or absence data was collected following discussions with the farmer, the results obtained are presented later in this report. A comparative survey is currently being undertaken in conjunction with a survey of 550 grassland and fodder farms.

ASSOCIATION WITH AGRICULTURAL HABITATS

Introduction

This is probably one of the most difficult areas to assess with only limited work being conducted in specific cropped areas, most information has been derived from recent studies in the Netherlands, (Strijbosch 1980) and Great

7

Britain, (Arnold 1995). The latter paper derives much of its data from a report of amphibian and reptile surveys presented to English Nature (Swan and Oldham 1993a, 1993b) and includes reports from observers around the country. For many species the association with agricultural habitats is not available and data have been obtained from a large number of sources, in particular Appleby (1971), Cooke (1986), Frazer (1983), Stafford (1989) and Wisniewski (1989).

The reporting of amphibian species is further complicated by the fact that all GB species need to return to the water to breed. Many of the records included in the book published by Arnold (1995), were of breeding populations and as such the proportion of those recorded in terrestrial habitats was reduced by these observations. This latter comment is not made to detract from the content of the book which has proved invaluable in mapping the distribution of native species.

During their terrestrial phase many amphibians are secretive and rarely seen, observations are normally made at the start of the breeding season during the post hibernation migration and during the main part of the breeding season which can last for several weeks in some species. During the breeding season some species can be easily observed particularly those that advertise their presence with mating calls. In contrast reptile species are almost wholly terrestrial and a good source of information on this group is the book published by Arnold (1995).

Although there is a loose association with arable and grassland for a number of species, in most cases unimproved grassland is the more important crop. In many cases field boundaries are the most important habitats and arable fields may simply be incidental migration routes. However species such as Triturus cristatus have been recorded in wheat fields, possibly foraging (Cooke 1986). Using pit fall traps Cooke showed that adult newts will move up to 20m into a wheat crop, pre-harvest, but are much less likely to do so after harvest, possibly because of the mechanical operations made at this time. In addition Oldham and Swan (1992) report that Bufo bufo (the common toad), will forage at night in arable fields and will retreat to scrub refugia during the day.

The importance of a variety of habitats in both the terrestrial and aquatic environments has been stressed on more than one occasion, (Oldham and Nicholson 1986), and the actual importance to amphibians and reptiles of a crop monoculture is very limited. Studies by Swan and Oldham (1993), show that the percentage occupancy of water bodies is influenced by the surrounding habitat and that in all cases unimproved grassland is a more favourable habitat for the commoner amphibian species than either improved grassland or arable land (Table 2).

Table 2. Percentage occupancy of water bodies associated with terrestrial habitat features (Swan and Oldham 1993a)

Species Improved grassland

Unimproved grassland

Arable Woodland

Rana temporaria (common frog) 52 58 39 61Bufo bufo (common toad) 13 29 27 32Triturus vulgaris (smooth newt) 19 29 22 16Triturus helveticus (palmate newt) 8 10 4 19Triturus cristatus (great crested newt) 10 14 8 11

A study of amphibians in agricultural lowlands (Beebee 1981) showed quite clearly that ponds wholly surrounded by arable fields, pasture, arable and pasture mixtures or even dense woodland rarely produced amphibians. Beebee (1981) suggested the best situations for newts were those in which arable crops were absent but included various combinations of woodland, pasture and scrub around the pools. Studies of Triturus alpestris and Triturus vulgaris in the Netherlands have also confirmed the studies of Beebee, showing that newts will only rarely migrate through intensively used arable or pasture land but will utilise hedges as a corridor through unsuitable areas (Gelder and Grooten 1992).

A survey 69 agricultural workers in Cambridgeshire showed that 76% thought agricultural chemicals were responsible for the decline in frogs and toads in their area, with a further 37% implicating improved drainage as the major problem (Cooke and Ferguson 1976).

Information on amphibians in agricultural habitats obtained during a recent arable crops pesticide usage survey is reported later in this document.

Habitat preferences of amphibian and reptile species in the UK are shown in Table 3. Although there is a loose association with arable and grassland for a number of species, in most cases grassland is the more important crop. In many cases the field boundaries are the most important habitats and arable fields may simply be incidental migration routes. However species such as Triturus cristatus have been recorded in wheat fields, possibly foraging (Wisniewski 1989).

8

Table 3. Habitat preferences of amphibian and reptile species in the UK. Primary habitats are listed in approximate order of importance. Primary and breeding habitats associated with agriculture are printed in bold. However the terrestrial habitat data is much more relevant to reptile species where data is readily available (Arnold 1995), rather than for amphibians which are more readily observed during their breeding period in water.

OrderSpecies*

Primary habitat Breeding habitat Association with crops

References

UrodelaT. alpestris - - - -T. cristatus Scrubland, hedge bases, ditches,

woodland, gardens, arable landGently flowing waters, reservoirs, ditches, larger ponds, 30-55cm deep.

Arable/grass farmland Wisniewski (1989), Arnold (1995), Cooke (1986)

T .helveticus Deciduous forest, heath, orchards, gardens, farmland, wet grassland and marshes

Permanent, shaded, unpolluted, clearwater bodies. Over 80cm deep

Wet grassland Wisniewski (1989), Arnold (1995)

T .vulgaris Upland, lowland, woodland, plantation, heathland, marsh, farmland and urban

Shallow temporary ponds/brooks Arable/grass farmland Wisniewski (1989), Arnold (1995), Marnell (1998)

AnuraA. obstetricans - - - -B. bombina - - - -B. variegata - - - -B. bufo Woodlands, wetlands, uplands,

grassland, arable land, moorland, gardens

Ponds, lakes reservoirs, occasionally ditches. Normally deeper water (33cm).

Arable/grassland Frazer (1983), Slater (1992), Strjbosch (1980), Oldham and Swan (1992), Cooke (1975)

B. calamita Heathland, sand dune and upper saltmarsh

Shallow water (8-15cm), dune slacks (occasionally brackish).

Limited - golf courses Anon. (1995)

H. arborea - - - -R. esculenta Mainly aquatic - any type of water bodies Ponds, lakes, ditches, dykes, slow flowing

rivers, fenlandField boundaries arable/grass

Frazer (1983)

R. lessonae Mainly aquatic - small pools, marshland Small pools, (woodland and open), marshland, fenland

Field boundaries arable/grass

Frazer (1983)

R. dalmatina - - - -R. ridibunda Mainly aquatic - ditches, dykes

(occasionally brackish)Ditches, dykes - mainly grazing marsh, arable land not as suitable.

Field boundaries arable/grass

Frazer (1983), Menzies (1962)

R. temporaria Woodland, marsh, grassland, gardens Ponds, ditches, slow flowing rivers, temp. pools, shallow water (18cm)

Arable/grassland Frazer (1983), Strijbosch (1980), Marnell (1998), Cooke (1975)

X. laevis - - - -SauriaA. fragilis Woodland, heath/moorland, scrub, dry

stone walls, hedgerows, railways, gardensOvo-viviparous. Juveniles occur in damper situations.

Arable/grassland Frazer (1983), Stafford (1989), Arnold (1995)

L. agilis Heathland and derivatives, sand dunes South facing sandy slopes used for egg deposition (oviparous).

Limited-golf courses House and Spellerberg (1983)

9

OrderSpecies*

Primary habitat Breeding habitat Association with crops

References

L. viridis - - - -L. vivipara Heath/moorland, woodland, sand

dunes/shore, banks, damp meadowsSouth facing banks. Viviparous. Arable/grassland Frazer (1983), Simms (1970),

Arnold (1995)P. muralis - - - -SerpentesC. austriaca Sandy heaths Basks on South facing banks. Ovo-

viviparous.Limited association Arnold (1995), Appleby (1971)

N. natrix Woodland, heath/moor, water body, marsh, garden, scrub

Eggs are normally laid in manure/compost heaps/decaying vegetation.

Arable/grassland Arnold (1995), Appleby (1971)

N. tessellatusV. berus Heath/moor, woodland, scrub Basks on South facing banks. Ovo-

viviparous.Arable/grassland Arnold (1995), Appleby (1971)

CheloniaT s..elegans - - - -

* see Table 1 for full species and common names

10

Arable Pond Survey, 1996

Methods

The following data were collected during a pesticide usage survey on arable crops. Pesticide usage data are collected as part of the FEPA requirements for the post registration monitoring of pesticides. A full account of the methodologies involved in sample selection, surveying procedures and raising factors has been reported elsewhere (Thomas et al. 1997).

Farmers were asked a number of questions regarding the number, distribution and management of ponds on their holding. In addition they were also asked for information on the presence or absence of frogs, toads or newts. This latter question was a difficult one to ask as there was no verification of observations by any members of the survey team and many of the farmers had no idea whether or not they had any amphibian species present. It was further complicated by the fact that to the untrained eye there are only slight morphological differences between frogs and toads, however the presence of frog spawn was very good in identifying frog populations.

Of the three groups, newts are the most likely to be under recorded due to their secretive behaviour, even during the breeding season. In addition the migration of newt species is nowhere near as intense as the mass migrations of common frogs and toads that also advertise their presence with mating calls.

Information was not collected in this survey that identified the locations of the ponds at which each amphibian species was observed.

Questions on the location of ponds were made in order to obtain the maximum amount of data regarding habitat and wildlife corridors. Ponds in cropped fields were normally those in an arable crop but they could have been in any other combinable/harvestable crop and were not located alongside a field boundary. Ponds in the boundary locations were in fields running alongside hedges or woodland. Grass fields could either have been temporary leys or permanent pasture. Ponds in woodland and farmyards are relatively self-explanatory. Other locations included scrub land, wasteland and unspecified areas.

Results

From a survey of 772 arable farms in England and Wales, 422 (55%), indicated that they had at least one pond on the holding (Table 4). It can be seen from that most ponds were recorded in the Eastern region, but this is more a reflection of the distribution of arable crops rather than ponds.

The data show that on average three ponds can be found on each arable farm, however in Midlands and Western, Eastern and South Eastern the average increases to four ponds and in Northern region and Wales decreases to two ponds per farm. Over 70 ponds were recorded on one holding, but 35% of holdings with ponds recorded only one pond, 24% two ponds, 13% three ponds, 9% four ponds with the remainder having five or more ponds.

Most ponds, 28%, were recorded as being in grass fields, however a further 27% were found in cropped fields, 19% in woodland, 12% on field boundaries with the remainder being in farmyards and other locations.

11

Table 4. Distribution and location of ponds on arable farms visited in 1996

Region Number of farms

Cropped fields

Grass fields

Woodland Boundaries Farmyard Other locations

Total ponds

Average number of

pondsNorthern 79 33 54 29 27 7 4 154 2

Midlands and western

82 87 132 36 34 7 24 320 4

Eastern 150 196 143 123 37 22 113 634 4

South Eastern

56 59 45 56 38 6 9 213 4

South Western

49 16 31 28 39 10 1 125 3

Wales 6 1 10 0 1 1 0 13 2

Total 422 392 415 272 176 53 150 1,458 3

Percentage of all ponds

- 27 28 19 12 4 10 - -

The distribution of ponds on arable farms with frogs, toads and newts present is shown in Table 5.

Table 5. Presence of frogs, toads and newts on arable farms

Region Number of farms

% with frogs % with toads % with newts

Northern 79 65 47 47Midlands and western 82 59 37 39Eastern 150 66 52 34South eastern 56 68 39 41South western 49 49 41 31Wales 6 83 33 17Total 422 63 45 38

Frogs in ponds on arable holdings

The distribution of ponds on holdings with frogs was very much in line with the distribution of ponds on all arable farms. There were only slight differences in the percentages recorded in any one category. The distribution of frogs showed an even distribution throughout the country (in line with arable farms and their pond distribution) and within different locations on the farm. This is in line with the geographical distribution of frogs, and in particular the common frog, Rana temporaria, which has been recorded in all 10km grids within England and Wales.

Of the holdings with ponds, 266 (63%) indicated that frogs were present. This figure is equivalent to over a third of all arable farms surveyed not only having ponds but having frogs present as well. A total of 1,083 ponds were recorded on the 266 holdings with frogs present.

No information on the location of frogs within each of the locations was available; the data available mainly show presence or absence. However, 44 (17%), of arable farmers indicated that frogs were present on their holding and that all ponds on the same holding were to be found in cropped fields. These 44 arable farms accounted for a total of 91 ponds or 8% of the total number of ponds.

In the vast majority of cases the frog species present is likely to be Rana temporaria, with the possible exceptions being some farms in the South Eastern region, particularly Kent, and a few locations in Eastern region. It is a very visible species spawning in shallow water and producing distinctive egg masses that are present for several days before hatching. In many cases these egg masses will float to the surface, particularly when the water temperature rises.

12

Toads in ponds on arable holdings

Again the distribution of ponds on holdings with toads was very much in line with the distribution of ponds on all arable farms and there were only slight differences in the percentages recorded in any one category. The distribution of toads followed the distribution of arable farms and their pond distribution, including the location of ponds within the farm. This is in line with the geographical distribution of Bufo bufo, the species most likely to be encountered. The only other native species of toad being Bufo calamita, which has a much more restricted range and is normally only encountered in heathland, sand dune and upper salt marsh.

Of the holdings with ponds, 189 (45%) indicated that toads were present. This figure is equivalent to almost a quarter of all arable farms surveyed not only having ponds but having toads present as well. A total of 892 ponds were recorded on the 189 holdings with toads present.

No information on the location of toads within each of the locations was available; the data available mainly show presence or absence. However, 33 (17%), of arable farmers indicated that toads were present on their holding and that all ponds on the same holding were to be found in cropped fields. These 33 arable farms accounted for a total of 71 ponds or 8% of the total number of ponds.

Bufo bufo is not as readily seen as Rana temporaria, and its behaviour may account for the smaller number of recorded arable farms with toads. It is highly visible during breeding time when it will migrate over long distances to preferred spawning sites. However encounters are normally only made on roads and not at the pond. Spawn is often deposited in deeper water where it is attached to submerged vegetation, it rarely floats to the surface and can only normally be found with careful searching. Males will often call and give away the presence of a colony to observers nearby.

Newts in ponds on arable holdings

As for frogs and toads the distribution of ponds on holdings with newts was very much in line with the distribution of ponds on all arable farms and there were only slight differences in the percentages recorded in any one category. The distribution of newts again followed the distribution of arable farms and their pond distribution, including the location of ponds within the farm.

Of the holdings with ponds, 159 (38%) indicated that newts were present. This figure is equivalent to a fifth of all arable farms surveyed not only having ponds but having newts present as well. A total of 741 ponds were recorded on the 159 holdings with newts present.

No information on the location of newts within each of the locations was available; the data available mainly show presence or absence. However, 27 (17%), of arable farmers indicated that newts were present on their holding and that all ponds on the same holding were to be found in cropped fields. These 27 arable farms accounted for a total of 61 ponds or 8% of the total number of ponds.

It is difficult to establish which species of newt would have been present in the arable ponds. All three native species have a wide distribution throughout England and Wales. Two species, Triturus vulgaris and Triturus helveticus are very similar and only readily distinguished during the breeding season. The third and largest species, Triturus cristatus, is very distinctive but regularly uses deeper ponds in which to breed and as such is often difficult to observe. All three species are secretive and only regularly observed by spending periods of time at the breeding ponds. It is likely that because of the difficulty involved in observing newts that the figures show an under recording of newt populations present.

DETAILED INFORMATION FOR SELECTED SPECIES

In order to derive the species to be selected as representative of the UK species at risk much of the information presented earlier in this document was used. In particular the distribution of species has been of key importance and no detailed information has been included for restricted or introduced species. Those species that showed some affinities with agricultural habitats and those occurring in the 1996 arable pond survey have also been included.

Detailed species information for Rana temporaria, Bufo bufo, Triturus cristatus, Triturus vulgaris, Lacerta vivipara and Natrix natrix was provided in the full report to PSD. The main sources of information were Arnold (1995) for distribution information and Frazer (1983) for information relating to the ecology and biology of individual species.

13

RISK ASSESSMENT METHODS

The aim of this section is to consider whether existing risk assessments methods for fish and birds provide protection for amphibians and reptiles. The usefulness of aquatic risk assessments for fish will be considered by comparing toxicity values with the most sensitive aquatic phase of amphibians. Risk assessment methods currently used for birds will be compared with those for reptiles and the terrestrial phases of amphibians considering assessments of dietary exposure and potential dermal exposure where possible. Before this can be done it is necessary to consider the availability of toxicity data such as 96h LC50 data for aquatic phases and LD50 data for reptiles and terrestrial phases of amphibians. Further to this it is necessary to determine the availability of data for UK species and variability of data due to species sensitivity, differences in sensitivity between life stages and differences between active ingredients and formulations.

Toxicity data

Availability of toxicity data

Main sources of toxicity data available include the Reptile and Amphibian Toxicity Literature (RATL) database (Pauli et al. 2000), EPA AQUIRE aquatic toxicity database (http://www.epa.gov/ecotox/), the review by Hefenist et al. (1989) and the SETAC publication ‘Ecotoxicology of Amphibians and Reptiles’ (Sparling et al. 2000). Useful information contained in more recent data from papers is summarised in Appendices in the full report to PSD. The vast majority of the data is LC50 data for aquatic phases of amphibians (mainly larvae/tadpoles) with relatively little LD50 data for terrestrial phases.

Availability of toxicity data for UK species.

The toxicity data obtained from the above sources is shown in Tables 6 and 7. Again the majority of the values found were for aquatic phases with few data for standard 96h LC50 studies, most of which were for the introduced Xenopus laevis and only four values for native tadpoles (three for Bufo bufo and one for Rana temporaria). LD50 data for adult UK species was found for only five compounds (Table 7). No toxicity data for any UK reptiles was found.

14

Table 6. LC50 data for native (Rana temporaria, Bufo bufo, Triturus cristatus) and introduced (Xenopus laevis, Rana ridibunda) amphibian species found in the UK.

Chemical Species Common name Stage Duration Toxicity (mg/l) Referencecarbaryl Bufo bufo Common toad 26-37 day tadpoles 24h 20.5-21.8 Marchal-Ségault (1976)DDT Bufo bufo Common toad 26-37 d tadpole 24h 0.7-1.3 Marchal-Ségault (1976)fenthion Bufo bufo Common toad 26-37 d tadpole 24h 2.1-2.6 Marchal-Ségault (1976)paclobutrazol Bufo bufo Common toad tadpole 24h 15.6 Liu et al. (1996)tetramethyl pyrophosphate Rana ridibunda Marsh frog adult (18-45g) 24h 34 Edery and Schatzberg-Porath (1960)

deltamethrin Rana temporaria Common frog tadpole 24h 13.35 (8.45-21.03) Thybaud (1990)HCH, gamma Rana temporaria Common frog tadpole 24h 8.63 (7.58-9.84) Thybaud (1990)carbaryl Xenopus laevis African clawed frog embryo 24h 4.7 (18°C) Elliot-Feeley and Armstrong (1982)

fenitrothion Xenopus laevis African clawed frog embryo 24h>10 (18-25˚C); 0.33 (30˚C)

Elliot-Feeley and Armstrong (1982)

thiram (F) Xenopus laevis African clawed frog stage 47 24h 0.017 Seuge et al. (1983)thiram (F) Xenopus laevis African clawed frog stage 53 24h 0.025 Seuge et al. (1983)carbaryl Bufo bufo Common toad 26-37 day tadpoles 48h 18.2-20.8 Marchal-Ségault (1976)DDT Bufo bufo Common toad 26-37 d tadpole 48h 0.5-0.8 Marchal-Ségault (1976)fenthion Bufo bufo Common toad 26-37 d tadpole 48h 2.0-2.2 Marchal-Ségault (1976)deltamethrin Rana temporaria Common frog tadpole 48h 19.61 (13.35-29.3) Thybaud (1990)HCH, gamma Rana temporaria Common frog tadpole 48h 5.88 (5.39-6.43) Thybaud (1990)paclobutrazol Bufo bufo Common toad tadpole 48h 14.4 Liu et al. (1996)mercuric chloride Xenopus laevis African clawed frog adult 48h 0.1 Sloof et al. (1983)phenol, pentachloro Xenopus laevis African clawed frog 3-4wk post hatch 48h 0.26 Slooff and Baerselman (1980) thiram (F) Xenopus laevis African clawed frog stage 47 48h 0.014 Seuge et al. (1983)thiram (F) Xenopus laevis African clawed frog stage 53 48h 0.022 Seuge et al. (1983)DDT Bufo bufo Common toad 26-37 d tadpole 72h 0.3-0.5 Marchal-Ségault (1976)paclobutrazol Bufo bufo Common toad tadpole 72h 11 Liu et al. (1996)thiram (F) Xenopus laevis African clawed frog not specified 72h 0.013 Seuge et al. (1983)thiram (F) Xenopus laevis African clawed frog not specified 72h 0.021 Seuge et al. (1983)carbaryl Bufo bufo Common toad 26-37 day tadpoles 96h 16.8-20.6 Marchal-Ségault (1976)fenthion Bufo bufo Common toad 26-37 d tadpole 96h 1.8-2.2 Marchal-Ségault (1976)paclobutrazol Bufo bufo Common toad tadpole 96h 9.1 Liu et al. (1996)Cyanatryn Rana temporaria Common frog tadpole 96h 30mg/l Haddow et al. (1974) DDT Rana temporaria Common frog adult 96h 24mg/l Harri et al. (1979)Naphthalene Xenopus laevis African clawed frog larvae (3 wk) 96h 2.1 Edmisten and Bantle (1982)

15

Chemical Species Common name Stage Duration Toxicity (mg/l) Referenceacrolein Xenopus laevis African clawed frog tadpole 96h 0.007 Holcombe et al. (1987)dieldrin Xenopus laevis African clawed frog juvenile 96h 0.047 (0.039-0.056) Schuytema et al. (1991)dieldrin Xenopus laevis African clawed frog tadpole 96h 0.040 (0.035-0.047) Schuytema et al. (1991)dieldrin Xenopus laevis African clawed frog tadpole 96h 0.049 (0.033-0.073) Schuytema et al. (1991)malaoxon Xenopus laevis African clawed frog embryo 96h 0.9 (0.7-1.0) Snawder and Chambers (1989) malathion Xenopus laevis African clawed frog embryo 96h 10.9 (10.6-11.3) Snawder and Chambers (1989) paraoxon Xenopus laevis African clawed frog embryo 96h 1.2 (1.1-1.3) Snawder and Chambers (1989) parathion Xenopus laevis African clawed frog embryo 96h 14.7 (13.5-16.0) Snawder and Chambers (1989) phenol, 2,4,6 trichloro Xenopus laevis African clawed frog tadpole 96h 1.2 Holcombe et al. (1987)thiram (F) Xenopus laevis African clawed frog stage 47 96h 0.013 Seuge et al. (1983)thiram (F) Xenopus laevis African clawed frog stage 53 96h 0.021 Seuge et al. (1983)maneb Triturus cristatus Great crested newt adult male 8.8h 125 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult female 16h 125 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult male 19h 75 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult female 19.5h 100 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult female 25.5h 75 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult male 28h 100 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult male 76h 50 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult female 7d 50 Zaffaroni et al. (1978)maneb Triturus cristatus Great crested newt adult male 10d 25 Zaffaroni et al. (1978)azinphos-methyl Xenopus laevis African clawed frog 2wk tadpole 4d 2.94 Schuytema et al. (1995)dieldrin Xenopus laevis African clawed frog embryo 7d 0.168 (0.156-0.182) Schuytema et al. (1991)

dieldrin Xenopus laevis African clawed frog 12d tadpole 10d0.0029 (0.0025-0.0034)

Schuytema et al. (1991)

dieldrin Xenopus laevis African clawed frog embryo 14d 0.028 (0.238-0.034) Schuytema et al. (1991)dieldrin Xenopus laevis African clawed frog embryo 21d 0.015 (0.012-0.018) Schuytema et al. (1991)

dieldrin Xenopus laevis African clawed frog tadpole 24d 0.0055 (0.0045-0.0066) Schuytema et al. (1991)

dieldrin Xenopus laevis African clawed frog tadpole 28d0.0109 (0.0073-0.0161)

Schuytema et al. (1991)

ethylene thiourea Xenopus laevis African clawed frog tadpole (5-12 d) 10d 100 Birch and Mitchell (1986)

cypermethrin Rana temporaria Common frog tadpole-metamorphosis Not stated 0.0065 Paulov (1990)

16

Table 7. LD50 data for adults of native (Rana temporaria) and introduced (Xenopus laevis, Rana ridibunda) species

Chemical Species Common name Route Duration Toxicity (mg/kg) ReferenceDDT Rana temporaria Common frog oral 20d 7.6 Harri et al. (1979)

fenitrothion Rana temporaria Common frog injection not given2400 (males)

Gromysz-Kalkowska and Szubartowska (1993)

fenitrothion Rana temporaria Common frog injection ales not given2220 (females)

Gromysz-Kalkowska and Szubartowska (1993)

paraoxon Rana ridibunda Marsh frog injection 7d 91 Edery and Schatzberg-Porath (1960)sodium monofluoroacetate Xenopus laevis African clawed frog injection not given > 500.0 Chenoweth (1949)

tetrachlorvinphos Rana temporaria Common frog injection not given151-192 (males > females)

Gromysz-Kalkowska and Szubartowska (1993)

17

Assessment of aquatic risk

Selection of data

Species sensitivity

Examination of the toxicity data for one endpoint (96h LC50 for larvae/tadpoles) where compounds had been tested on several species (e.g. Table 6) indicated an apparent wide range in sensitivity between species, sometimes of several orders of magnitude. This may be due to differences between age of the animals, differences in study method or differences in test material. In order to be conservative the lowest value or a calculated HC5 value should be used for comparisons. The latter is preferable to avoid bias due to varying sample size for different pesticides.

Sensitivity of amphibian life stages

Several studies have identified differences between the sensitivity of amphibian life stages (e.g. Greulich and Pflugmacher 2003, Harris et al. 2000). Toxicity data for studies where different life stages were tested allowing direct comparisons to be made are shown in Table 8. To be most protective for amphibians it will be necessary to use the most sensitive stage for comparison with fish data.

Table 8. Comparison of toxicity of pesticides to different life stages of amphibians

LC50 (mg/l)

Species Chemical Egg EmbryoLarva/

tadpoleSub-adult Adult

Testduration Reference

Bufo americanus Glyphosate isopropylamine 6.4 1.7 96h Edgington et al.

(2004)

Microhyla ornata carbofuran 44.23 13.47 96h Pawar and Katdare (1984)

Microhyla ornata fenitrothion 3.21 1.14 96h Pawar and Katdare (1984)

Microhyla ornata HCH, gamma 23.37 7.27 96h Pawar and Katdare (1984)

Rana catesbeiana

3-trifluoromethyl-4-nitro phenol 0.95 12.99 24h Kane et al. (1993)

Rana clamitans Glyphosate isopropylamine 4.1 1.4 96h Edgington et al.

(2004)

Rana pipiens Glyphosate isopropylamine 7.5 1.1 96h Edgington et al.

(2004)Rana sphenocephala

campheclor, toxaphene 0.651 0.065 0.378 96h Pierce and Wooten

(1992) Rana sphenocephala endrin 0.025 0.006 0.005 96h Pierce and Wooten

(1992) Rana sphenocephala endrin 0.006 0.005 96h Leftwich and Lilly

(1992) Xenopus laevis Carbaryl 15.25 1.73 96h Zaga et al. (1998)

Xenopus laevis Glyphosate isopropylamine 7.9 0.88 96h Edgington et al.

(2004)

Larvae/tadpoles are consistently more sensitive than embryos, and more limited data suggest they are also more sensitive than eggs, sub-adults and adults. It is suggested by Edgington et al. (2004) that the presence of functional gills in the larvae makes them more susceptible to the formulation due to uptake. These differences suggest that larvae/tadpole data should be used in comparisons with fish.

Technical vs. formulation toxicity

Formulations often appear to be far more toxic than technical material and additional components of waterborne pesticides can increase toxicity by more than 2 orders of magnitude (Mayer and Ellersieck 1986). Howe et al. (2004) demonstrated higher toxicity of glyphosate formulations when compared to technical material. This causes problems when selecting data from databases (e.g. RATL) for analysis as it is often not clear whether the test material was a formulation or technical grade material. Even where fields for test substance type, purity etc. are included in the database (e.g. EPA AQUIRE) the data is often incomplete. Where possible the studies that can be

18

identified as technical material should be selected but this would reduce the number of values available for comparison. Use of HC5 rather than lowest values where sufficient data is available would also reduce the impact of very low values due to the effects of formulations.

Previous comparisons of amphibian and fish toxicity data

To test whether aquatic risk assessments for fish protect amphibians it would be necessary to compare the toxicity of pesticides to both groups. Birge et al. (2000) compared the embryo-larval toxicity for 25 amphibian species with that of rainbow trout for 34 inorganic elements and 27 organic compounds (447 tests) to classify species as very sensitive to tolerant compared to fish. Comparison was made by calculating chemical hazard index values (CHI50) calculated as follows:

CHI50 = amphibian LC50/fish LC50.

Amphibian LC50 by this measure was lower than fish LC50 in 36% to 63% of cases depending on the fish species compared suggesting greater sensitivity in a large proportion of comparisons. In this comparison the lowest species specific CHI50 in comparisons with rainbow trout was 0.19 for metals and 1.40 for organics which provides some indication that the safety factor of 100 put around RA for fish may encompass at least most amphibian risks. The only UK species in this study was Rana temporaria for organics which had the lowest CHI50 value in comparison with rainbow trout (1.40) and was classified as ‘sensitive’.

Comparison of most sensitive amphibian with rainbow trout

To investigate the relationship between amphibian and fish toxicity the ratio of amphibian and fish toxicity was calculated to determine how frequently the toxicity to amphibians would fall outside the uncertainty factor of 100 placed around aquatic TERs. Given that larvae/tadpoles appear to be consistently more sensitive than other aquatic life stages, these were used for comparison with fish to ensure that the outcome was conservative. The distribution of the log ratio between amphibian and trout toxicity (log10 [amphibian LC50/trout LC50]) was examined in each case to determine how frequently amphibians were more than 100 times more sensitive than fish (e.g. log10 [amphibian LC50/trout LC50] < -2).

Toxicity data (96h LC50) for larvae/tadpoles were compared with those for rainbow trout obtained from the Pesticide Manual (13th edition). In order to be conservative the lowest value was used where more than one species of amphibian had been tested. Only 96h LC50 data for trout where an actual value was given were used in the analysis (e.g. excluding > value). Collating amphibian data from the above sources, and excluding those that were clearly identifiable as formulation, provided 33 values for analysis. At an amphibian/trout ratio of 0.01, p = 0.045 indicating that over 95% of amphibian toxicity values fell within a factor of 100 of trout toxicity. To confirm this, the analysis was repeated using the quality-checked dataset produced for the DEMETRA project. This provided 28 values for analysis providing a value for p at 0.01 of 0.018. Taking this further, only data which was identified as technical material was compared with the DEMETRA data set in the same way. This left 13 values and in this case the p value at a ratio of 0.01 was 0.047 indicating that using this measure over 95% of values fell within a factor of 100 despite the small size of the sample.

Assessment of the protection of untested amphibians

The above analysis compared data for tested amphibians with those of trout. However, it is likely that assessments will need to be made of the risk to species that have not been tested. To do this the first dataset above was examined to determine the number of compounds for which there were at least two species tested and data for trout in the Pesticide Manual which left 19 compounds. Secondly the HC5 was calculated on log toxicity data using the excel function NORMINV with a p value of 0.05 to determine the 5% value. In some cases only two values were available for a compound which was not ideal but in these cases the estimated HC5 will be below the lowest value making the comparison conservative. These HC5 values were then back transformed for calculation of amphibian/trout ratios as before and the distribution of the log ratios examined to estimate the frequency with which values might fall outside the safety factor of 100. In this case the p value at the threshold was 0.108 indicating that fish risk assessments would be protective on over 89% of occasions. When this was repeated using the quality checked fish data, 13 values were available and these provided a p value of 0.084. As above this estimate is likely to be conservative as only ‘known’ formulation values were removed and assuming that formulations are likely to be more toxic any remaining ones in the dataset would tend to skew the distribution further to the left and make the fish data appear less protective.

When considering the results of this analysis it must be remembered that:

a) This procedure assumes that the species sensitivity distribution (SSD) is log normal, which is usually a good fit for other aquatic organisms.

19

b) These HC5’s will be very approximate because some are calculated from only two species. However, errors resulting from this should cancel out over a number of pesticides unless the tested species are a biased sample.

Conclusion on aquatic risk

The above analysis indicates that the uncertainty factor of 100 placed around risk assessments for fish would make them protective of amphibians on most occasions. It is therefore recommended that fish assessments are used to protect larval amphibians against acute risk unless specific concerns arise.

Assessment of terrestrial risk

Comparison with bird toxicity

Toxicity of pesticides to adult amphibians along with mallard and quail data where found are shown in Table 9. Clearly the availability of acute toxicity data for adults is poor compared to larvae.

20

Table 9. Acute toxicity to adult amphibians compared to birds.

Toxicity (mg/kg)Chemical Species Common name Route Amphibian Mallard QuailBaygon (propoxur) Rana catesbeiana Bullfrog Oral 595 9.4 a 28.3 (JQ)

d

Carbaryl Rana catesbeiana Bullfrog Oral >4000 >2564 a 2230 (JQ) b

Chlorpyrifos Rana catesbeiana Bullfrog Oral >400 76 a 32 (BWQ) c

DDT Rana catesbeiana Bullfrog Oral >2000 >2240 a -Rana temporaria European common frog Oral 7.6Deltamethrin, cis Rana pipiens Northern leopard frog Injection 0.35 >4640 b -Demeton Rana catesbeiana Bullfrog Oral 562 7.2 a 8.5 (JQ)

d

Diazinon Rana catesbeiana Bullfrog Oral >2000 3.5 a 5.2 (BWQ) c

Dicrotophos Rana catesbeiana Bullfrog Oral 2000 4.2 a 4.3 (JQ) d

Fenitrothion Rana temporaria European common frog Injection 2220 1190 b 23.6 b

Mexacarbate Rana catesbeiana Bullfrog Oral 566 3 a 3.2 (JQ) d

Nabam Rana catesbeiana Bullfrog Oral 420 >2560 a -Paraquat Rana esculenta edible frog Injection 260 199 b 175 (BWQ)

b

Phorate Rana catesbeiana Bullfrog Oral 85 0.6 a 7 (BWQ) c

Phosphamidon Bufo arenarum Common toad Injection 1195 3.8 b 3.6-7.5 (JQ) b

Pyrethrin Rana pipiens Northern leopard frog Injection 5.8 >1000 b 7070 (JQ)

s-Fenpropathrin Rana pipiens Northern leopard frog Injection 0.27 1089 b -

Sodium monofluoroacetate

Limnodynastes tasmaniensis Marbled frog Injection 60

5.9 a -Rana catesbeiana Bullfrog Oral 54Rana pipiens Northern leopard frog Injection 150Xenopus laevis African clawed frog Injection > 500.0

ss-fenvalerate Rana pipiens Northern leopard frog Injection 0.13 9932 b -Strychnine Rana catesbeiana Bullfrog Oral 2.2 2 a -Temephos Rana catesbeiana Bullfrog Oral >2000 79 a 27.4 (BWQ)

c

TEPP Rana catesbeiana Bullfrog Oral 112 3.6 a -Tetrachlorvinphos Rana temporaria European common frog Injection 151 >2000 b -

a Hall and Henry (1992)b Pesticide Manual 13th editionc DEMETRAd Smith (1987)BWQ = bobwhite quailJQ = Japanese quail

21

There appears to be little correlation between the amphibian and birds data presented here indicating that bird toxicity may be a poor predictor. However, as well as the small size of the data set there are many other potential causes of variability such as differences in species sensitivity among the amphibians tested and differences in test conditions (e.g. test substance type, temperature). Also, much of the LD50 data for adult amphibians is from an injected rather than oral dose which makes comparisons problematic. However, comparisons of oral data alone do not seem to improve the correlation although there is some indication that the amphibians are far less sensitive to organophosphorus compounds in many cases. An analysis of the distribution of log 10 [amphibian LD50/mallardLD50] as conducted for the comparison of larvae with fish indicated that at a log ratio of –1 (assuming an uncertainty factor of 10), p = 0.32 suggesting that only 68% of amphibian values fell within a factor of 10 of mallard toxicity. If the uncertainty factor was set at 100 this increases to 82%.

The difference between the ratios for OPs and pyrethroids was examined by conducting a t-test on the log ratios. This was significant (t = 7.64, p < 0.001) with values for OPs far greater than those for pyrethroids. This suggests that while an uncertainty factor of 10 may be suitable when considering OPs, a larger value may be more suitable for pyrethroids. The lack of data prevents conclusions for other classes of compounds.

For reptiles there is even less data available (Table 10) although there seems to be more similarity with values for birds at least for these compounds. Given the small size of the data set it is not possible to make meaningful comparisons.

Table 10. Acute oral toxicity of pesticides to reptiles and mallard ducks.

Chemical Species Toxicity(mg/kg) Source Mallard

(mg/kg) Source

Azinphos-methyl Anolis carolinensis 98 Hall and Henry

(1992) 136 Hall and Henry (1992)

Malathion Anolis carolinensis 2324 Hall and Henry

(1992) 1485 Hall and Henry (1992)

Malathion Lacerta parva 169.8 Ozelmas and Akay (1995) 1485 Smith (1987)

methyl-parathion Anolis carolinensis 82.7 Hall and Henry

(1992) 10 Hall and Henry (1992)

Parathion Anolis carolinensis 8.9 Hall and Henry

(1992) 2.1 Hall and Henry (1992)

Parathion Gallotia galloti >7.5 Sanchez et al. (1997 1.4-2.4 Smith (1987)

Phosphamidon Calotes versicolor 2.3

Meenakshi and Karpagaganapathi (1996)

3.8 Pesticide Manual 13th edition

Trichlorfon Gallotia galloti >100 Fossi et al. (1995) 36.8 Smith (1987)

These data suggest that it would be unwise to use bird data routinely to represent amphibians for which there are no data without using a large uncertainty factor. While this might be conservative for some compounds such as OP compounds, it may be less so for other compounds such as pyrethroids to which amphibians seem far more sensitive. Bird data may be a better surrogate for reptile data but the very small data set does not allow any test of the robustness of such an approach.

The dermal route of exposure appears potentially important for amphibians and reptiles (see below), but there appear to be little or no data on dermal toxicity of pesticides to amphibians and reptiles (with the exception of a few studies with organochlorine compounds). Mineau (2002) derived a dermal toxicity index (DTI) for birds, by regressing the ratio of 1000 x acute oral to dermal toxicity (for species where both have been tested) on physico-chemical properties (Kow, molecular weight and molecular volume) and the rat dermal toxicity index for the same pesticides. A similar approach could be considered for amphibians and reptiles, but would first require generation of data on amphibian or reptile dermal toxicity for a representative sample of pesticides. In the meantime, any assessment of dermal exposure will have to be very approximate, based on extrapolation from oral toxicity data (see below).

Food intake

One of the main routes of exposure for reptiles and post metamorphic amphibians is likely to be through consumption of contaminated food. At present this is the only route considered for birds and it should be possible to make predictions in a similar way. However, the estimation of daily intake may be more problematic for amphibians and reptiles as there may be far higher daily variation in food intake and energy requirements due to

22

the effects of temperature on activity and metabolism (Robinson et al. 1983). Also, there may be large periods of the year when no food is consumed due to hibernation. The latter may be taken account of by only considering risk during the active period but the effects of temperature may be harder to include in an assessment.

Estimating energy requirements

Before estimating the potential dietary exposure of post-metamorphic amphibians and reptiles it is necessary to obtain estimates of daily food intake. One source of information is chapter 2 of the EPA Wildlife Exposure Factors Handbook (WEFH, http://cfpub.epa.gov/ncea/cfm/wefh.cfm?ActType=default) which summarizes metabolic rate and food intake data for a range of US amphibians and reptiles for use in risk assessment. While some of these are based on allometric estimates of basal metabolic rate (BMR) for poikilotherms (e.g. Robinson et al. 1983) there appears to be a lack of the type of field metabolic rate (FMR) data for amphibians that is available for other vertebrate groups. Chapter 3 of the handbook presents equations for calculating food intake rates for iguanid lizards based on field metabolic rates (Nagy 1987) as the only data of this type available for reptiles and amphibians. This has since been superceded by the allometric equations presented in Nagy et al. (1999) which allow calculation of field metabolic rates and 95% confidence limits for ‘all reptiles’ and subgroups such as ‘all lizards’, Lacertidae etc..

All reptiles kJ/day = 0.196(g body mass)0.889

All lizards kJ/day = 0.190(g body mass)0.916

Lacertidae kJ/day = 0.1586(g body mass)1.009

Using the equation for lizards we can calculate daily energy requirements for a range of animal sizes covering the list of main species from the first section. Estimated FMR and 95% confidence limits are shown in Table 11.

Table 11. Daily energy expenditure based on the allometric equation for all lizards in Nagy et al. (1999).

Daily energy expenditure (kJ/day)Weight (g) Estimated lower 95% upper 95%

2 0.36 0.07 1.734 0.68 0.17 2.636 0.98 0.28 3.418 1.28 0.40 4.1210 1.57 0.51 4.7915 2.27 0.80 6.4020 2.95 1.10 7.9625 3.62 1.38 9.5030 4.28 1.66 11.0535 4.93 1.93 12.6040 5.57 2.19 14.1745 6.21 2.45 15.7550 6.84 2.70 17.35

100 12.90 4.84 34.40150 18.71 6.59 53.12

While these are the best data available for estimating the field metabolic rate of reptiles, it is not certain that they are suitable for use for amphibians. The WEFH provides an allometric equation for estimating the BMR for amphibians and reptiles at 20C based on Robinson et al. (1983).

BMR (Watts) = 0.19(kg body mass)0.76

Taigen (1983) used regression analysis on data for 17 anuran amphibians to generate allometric equations for resting metabolic rate (RMR) and maximum metabolic rate (MMR).

RMR (VO2 ml/h) = 0.092(g body mass)0.73

MMR (VO2 ml/h) = 0.75(g body mass)0.91

Using the conversion factor 1ml(O2) = 20.1 J and converting to kJ/d the various estimates were compared (Table 12).

23

Table 12. Comparisons of estimated metabolic rate for lizards (Nagy et al. 1999), amphibians and reptiles (Robinson et al. 1983) and anuran amphibians (Taigen 1983).

Lizards Amphibians and reptiles

Anuran amphibians

Weight (g) FMR(kJ/d)

BMR (kJ/d)

RMR (kJ/d)

MMR (kJ/d)

Average(kJ/d)

2 0.36 0.15 0.07 0.68 0.384 0.68 0.25 0.12 1.28 0.706 0.98 0.34 0.16 1.85 1.018 1.28 0.42 0.20 2.40 1.3010 1.57 0.50 0.24 2.94 1.5915 2.27 0.67 0.32 4.25 2.2920 2.95 0.84 0.40 5.53 2.9625 3.62 0.99 0.47 6.77 3.6230 4.28 1.14 0.53 7.99 4.2635 4.93 1.28 0.59 9.20 4.9040 5.57 1.42 0.66 10.38 5.5245 6.21 1.55 0.71 11.56 6.1450 6.84 1.68 0.77 12.72 6.75

100 12.90 2.85 1.28 23.90 12.59150 18.71 3.88 1.72 34.57 18.15

While the estimates of RMR for amphibians (Taigen 1983) were around half of those of the Robinson (1983) estimate of BMR, the estimates of MMR were considerably higher. The average of the RMR and MMR for amphibians, which could be used as an estimate of FMR, was remarkably similar to those of Nagy et al. (1999) for lizards. This suggests that the Nagy (1999) values may be suitable for estimating the field metabolic rate of amphibians for our purposes and if we wish to be conservative the MMR values from Taigen (1983) could be used rather than the more extreme upper 95% values from Nagy.

Estimating food intake

For bird and reptile insectivores Nagy et al. (1999) recommend the use of a conversion factor of 18.0 kJ/g of dry matter (DM) to convert these energy values to dry food intake e.g.

Daily food intake (DM) =

To convert this to fresh matter (FM) they assume that animal foods such as insects usually contain 70% water so that FM = 3.33 x DM. Further information on energy content, moisture content for a wide variety of foods along with assimilation efficiencies for birds are available in Crocker et al. (2002) and are shown in Table 13.

Table 13. Energy and moisture content of animal food types (Crocker et al. 2002)

Food type kJ/g dry weight MoistureSmall mammals 21.7 0.686Bird/mammal carrion 22.6 0.688Arthropods 21.9 0.705Caterpillars 21.7 0.794Soil invertebrates 19.3 0.846Fish 20.7 0.711Aquatic invertebrates 19.6 0.773

Using these values would allow for more refined estimates of consumption based on the diet of the species under consideration. For passerine birds the assimilation efficiency is given as 0.76 which for arthropods provides a slightly more conservative estimate of kJ/gDM of 16.6. Using these values daily food intake can be calculated as follows:

24

Daily food intake (wet g) = Daily energy expenditure (kJ)Energy in food (kJ/g) x (1 - moisture) x Assimilation efficiency

Where moisture and assimilation efficiency are proportions between 0 and 1.

Using these values and the daily energy expenditure calculations above we can estimate the daily food intake and food intake rate (FIR) for animals of different sizes and different food types. The most important food items for amphibians and small reptiles are likely to be arthropods and soil invertebrates. Tables 14 and 15 provide estimated food intake and FIR for these food types. Note that the confidence intervals derive from those for daily energy requirement and do not include uncertainty in other parts of the calculation.

Table 14. Estimated daily food intake (FM) and FIR for amphibians/reptiles feeding on arthropods based on allometric equations for lizards in Nagy et al. (1999) and energy content, moisture content, and assimilation efficiency for passerine birds from Crocker et al. (2002).

g FM/day FIRWeight (g) Estimate lower 95% upper 95% Estimate lower 95% upper 95%

2 0.073 0.015 0.353 0.037 0.008 0.1764 0.138 0.035 0.536 0.034 0.009 0.1346 0.200 0.058 0.694 0.033 0.010 0.1168 0.260 0.081 0.838 0.032 0.010 0.10510 0.319 0.104 0.976 0.032 0.010 0.09815 0.462 0.164 1.304 0.031 0.011 0.08720 0.602 0.223 1.621 0.030 0.011 0.08125 0.738 0.282 1.936 0.030 0.011 0.07730 0.872 0.338 2.250 0.029 0.011 0.07535 1.005 0.393 2.566 0.029 0.011 0.07340 1.135 0.447 2.885 0.028 0.011 0.07245 1.265 0.499 3.208 0.028 0.011 0.07150 1.393 0.549 3.534 0.028 0.011 0.071

100 2.628 0.986 7.006 0.026 0.010 0.070150 3.810 1.342 10.819 0.025 0.009 0.072

Table 15. Estimated daily food intake (FM) and FIR for amphibians/reptiles feeding on soil invertebrates based on allometric equations for lizards in Nagy et al. (1999) and energy content, moisture content, and assimilation efficiency for passerine birds from Crocker et al. (2002).

g FM/day FIRWeight (g) Estimate lower 95% upper 95% Estimate lower 95% upper 95%

2 0.159 0.033 0.766 0.079 0.016 0.3834 0.299 0.077 1.166 0.075 0.019 0.2916 0.434 0.125 1.508 0.072 0.021 0.2518 0.565 0.175 1.823 0.071 0.022 0.22810 0.693 0.226 2.122 0.069 0.023 0.21215 1.005 0.356 2.834 0.067 0.024 0.18920 1.308 0.485 3.524 0.065 0.024 0.17625 1.605 0.612 4.207 0.064 0.024 0.16830 1.896 0.735 4.891 0.063 0.025 0.16335 2.184 0.855 5.578 0.062 0.024 0.15940 2.468 0.971 6.272 0.062 0.024 0.15745 2.749 1.084 6.973 0.061 0.024 0.15550 3.028 1.193 7.682 0.061 0.024 0.154

100 5.713 2.143 15.229 0.057 0.021 0.152150 8.283 2.917 23.517 0.055 0.019 0.157

The slightly lower energy content combined with the higher moisture content of soil invertebrates means that the FIR for this food type is over twice that for arthropods.

Comparisons with birds

25

Clearly the FIRs calculated here are far lower than those predicted for birds. For instance the standard tier 1 insectivorous bird considered in SANCO/4145/2000 (Anon. 2002) has a FIR of 1.04 which is over 30 times higher than that predicted for a 10g lizard feeding on arthropods and 10 times the upper 95% confidence limit for this value. It would seem that unless the amphibian or reptile being considered is far more sensitive than birds, then the calculated TER based on these levels of intake would most likely always be lower suggesting that they would be protected by the bird assessment. However, while dermal exposure is considered to be a potential risk for birds even though it is not currently assessed, it is likely to be an even more important route of exposure for amphibians especially, and is likely to significantly increase exposure above that predicted by the dietary route alone.

Birge et al. (2000) suggest that variability in feeding patterns make even dietary exposure estimates difficult. For example, amphibian/reptile food consumption may be punctuated (e.g. snakes etc.) due to relatively low DEE so that ‘average’ assessments of energy required from food may underestimate worst case ingestion. Patterns of feeding in amphibians are likely to be less predictable than birds due to the effects of temperature on activity and energy expenditure (Robinson et al. 1983). Presumably the toxicity as well could vary due to metabolic changes (e.g. rate of metabolism of contaminants). However, these potential variations in feeding levels seem unlikely to exceed the 30 times differences seen in intake rates when compared to birds.

Dermal exposure

Dermal exposure of wildlife has been shown to be potentially important in studies with birds. Driver et al. (1991) demonstrated that exposure via the dermal route could not only increase the estimate of risk but also extend the period over which effects occurred when compared to oral exposure alone. Mineau (2002) showed that predictions of mortality for birds after pesticide treatment were improved when the dermal risk was taken into account. Given the lack of hairs, feathers or scales which may form a partial barrier to exposure, and the presence of specialised structures to enhance absorption (e.g. dermal patches) it can be assumed that should amphibians come into contact with a contaminated substrate that at least some exposure will occur. Cowman and Clark (1995, 1996) conducted studies where adult toads of Bufo valliceps and Bufo velatus were exposed ventrally to sublethal concentrations of carbofuran (Furadan 4F) and trifluralin (Treflan). Carbofuran at levels of 5.5% of application rate had severe behavioural effects.. It might be expected that the keratinised scales of reptiles may form some barrier such that the rate of uptake would be reduced compared to amphibian. Given the relatively low metabolic rates of amphibians and reptiles compared to birds and mammals and the consequently reduced oral exposure it is likely that dermal exposure would form a much larger proportion of the total exposure. Also, there may be a far greater chance of direct overspray of amphibians and reptiles when in the crop than would be expected for birds (except perhaps chicks of ground nesting species) due to a reduced ability to escape quickly enough.

However, estimation of dermal uptake is complex compared to dietary exposure as it depends on the surface area in contact with the contaminated substrate/medium, the water potential of the animal and substrate/medium, the physical properties of the chemical (e.g. molecular size) and the permeability of the skin. A formula for calculating the rate of uptake of a dissolved contaminant based on the assumption that it will move into the animal at the same rate as water is absorbed and is at the same concentration as the pore water (or soil concentration if this is all that is available) are presented in Birge et al. (2000).

where:dm/dt = rate of uptake of water by amphibians (kg x s-1)A = area in contact with substrate (m2)R = resistance to water uptake (s x m2 x Pa x kg-1)s = water potential of soil (Pa)a = water potential of amphibian (Pa)

A similar formula is also given for absorption in water.

where:dm/dt = rate of uptake of water by amphibians (kg x s-1)Aw = area in contact with water (m2)R = resistance to water uptake (s x m2 x Pa x kg-1)w = water potential of water (Pa)

26

a = water potential of amphibian (Pa)

(From: Feder and Burggren 1992)

Allometric equations also exist for calculation of the surface area of amphibians as follows.

SAskin (cm2) = 13.52 Wt0.579 (g) all frogs (based on Hutchinson et al. 1968)*

SAskin (cm2) = 8.42 Wt0.694 (g) salamanders (Whitford and Hutchinson 1967)

* note: equations as listed in the source paper and presented in the Wildlife Exposure Factors Handbook appear to be incorrect. For example the formula for all frogs is presented as: SA = 1.131 Wt0.579 where it should be in the form log(SA) = 1.131 + 0.579 log (Wt) as for salamander equation in Whitford and Hutchinson (1967).

Due to the similarity in shape, the equation for salamanders would seem a good starting point for estimating the area of newts and lizards in the absence of other data, while the area of a snake could be estimated as that of a cylinder if length and average diameter are known.

Estimation of dermal exposure

Even if we make the simple assumption of absorption with water as above, and all the parameters are known, the wide range of possible scenarios of soil type, soil moisture content, hydration state of the animal and time in contact with the substrate make it impossible to produce a robust estimate of uptake that could be used in risk assessment.

A simple and conservative assumption would be that the animal instantaneously absorbs any contaminant it contacts. This would require information on soil concentration (e.g. based on application rate), an estimate of the area of animal in contact with the soil, and the distance travelled. This would allow calculation of the total area of contaminated soil contacted as the animal moves around its environment and therefore the maximum amount of contaminant that could be absorbed.

Several studies indicate that amphibians have the potential to move considerable distances in a single day. Oldham and Swan (1992) found that toads (Bufo bufo) with ingested radio transmitters would move between 0 and 180m between daily recordings during the active season in spring and summer (but were not tracked during migratory phase). Some were found to forage in arable fields at night up to 20m from the scrub they took refuge in during daylight. This suggests that total distance moved across the fields could be 40m or more. In a German study of the movements of Rana temporaria using implantable transmitters the daily movement during migration ranged from 0 to 155m (Seitz et al. 1992). Mean weight of frogs was 37g (range 26.7g to 46.3g). Total migration distance from the origin over the course of the study ranged from 10 to 460m.

It seems likely that migratory movements in spring would represent a worst case e.g. migrating long distances across fields during spring applications. For example a study of migrating Bufo bufo found that the maximum distance covered in one night was 440m (Van-Gelder et al. 1986). In a study of the movements of Bufo calamita in Spain the maximum distances between daily recordings after the breeding season ranged from 125 to 353m (Miaud et al. 2000). Of course, it is unlikely that all of these large scale movements were within treated fields but from these studies it would seem reasonable to consider the effects of movements of up to 100m across treated areas.

Using this approach for a hypothetical amphibian weighing 40g and 5cm wide moving through an area treated at 0.5kg/ha active ingredient we can illustrate the exposure for a given range of % absorption and travel distance (Figure 1).

27

Figure 1. Estimated dermal exposure assuming a range of % absorption from the surface of a 5cm wide track across an area treated at 0.5kg/ha. Each line represents a different track length ranging from 1m to 100m. This graph can be used to obtain a conservative estimate of dermal exposure for different combinations of % absorption and distance travelled.

0

1000

2000

3000

4000

5000

6000

7000

0 10 20 30 40 50 60 70 80 90 100

Absorption (%)

mg/k

g

1m5m

10m

20m

30m

40m

50m

60m

70m

80m

90m

100m

Clearly there are a vast range of possible scenarios which could affect the estimate of % absorption and the 100% assumption is likely to overestimate the level of exposure due to the factors below.

Factors reducing simple estimate of contact exposure:

Overestimation of the area of ground contacted when stationary. Not all of lower part of animal in contact with ground when moving (e.g. walking). Movement in leaps/hops so not all soil along animal’s track contacted. Reduced soil concentration due to interception by crop, soil absorption, residue decay etc..

Surface area in contact with the soil when stationary

Even if the simple estimate of a 5cm circle as the plan area of an amphibian in this example was an accurate value it is unlikely that the entire area would contact the ground as the main points of contact (for a frog or toad) would be the feet and the rear of the animal while resting. Better data on the actual surface area in contact would help to refine the estimate. Other species such as newts and snakes would seem likely to have more contact with the soil relative to total surface area.

Surface area in contact with the soil while moving

As above the area in contact with soil is likely to be lower than predicted and most likely lower than for a stationary animal as only the feet are in contact with the ground. Again, better data on the area of contact by feet would help refine this estimate of uptake.

Movements in leaps or hops

Amphibians moving by leaping would greatly reduce the area of contaminated ground contacted for a given distance. Studies of the effects of temperature on jumping in Rana pipiens and Bufo americanus indicated that over the range 5C to 25C the mean jumping distances ranged from approximately 20cm to 60cm for Rana pipiens. and 10cm to 20cm for Bufo americanus (Renaud and Stevens 1983). At 15C mean jump distances were approximately 41cm and 18cm respectively. Wilson et al. (2000) developed allometric scaling relationships for jumping distance of Limnodynastes peronii (the striped marsh frog). These predict that at 24C a 40g frog would

28

have a maximum leaping distance of approximately 90cm. The possible effects of this type of locomotion in our example are relatively easy to estimate (Table 16).

Table 16. Proportion of total possible track contacted assuming an area of contact described by a 5cm diameter circle at each jump.

Leap distance(cm)

% of track contacted

Proportion of track contacted

10 39.3 0.39320 19.6 0.19630 13.1 0.13140 9.8 0.09850 7.9 0.07960 6.5 0.06570 5.6 0.05680 4.9 0.04990 4.4 0.044

100 3.9 0.039

Even a relatively small leap of 20cm would reduce the estimate of exposure in the Figure 1 by over 80%. Given the range of possibilities from walking to leaping/hopping or a combination of the two it is difficult to make accurate predictions of contact especially given the potential effects of sublethal exposure on behaviour.

Reduced soil concentration

Soil concentration is likely to be lower than that predicted from the application rate due to losses in spraying, absorption into the soil, adsorption to soil particles and residue decay. Also, interception by the crop with factors ranging from 0 (no crop) to 0.75 (full canopy, depending on crop) would potentially reduce the estimated exposure by up to 75%.

Comparison of dermal exposure with toxicity

As stated earlier, there is almost no existing data on dermal toxicity of pesticides to amphibians and reptiles. Dermal LD50s are generally higher than oral toxicities, due to lower absorption dermally than from the gut. However, the method proposed above for estimating exposure includes a factor to allow for partial absorption (x axis in Figure 1). Therefore, the resulting estimate of exposure needs to be compared to an estimate of the internal LD50, i.e. toxicity based only on the absorbed dose. This should be reasonably approximated by LD50s for injected doses, which are available for some pesticides. Where only an acute oral LD50 is available, or where the assessment is based on extrapolation from an avian acute oral LD50, this may over-estimate the internal LD50 due to <100% absorption from the gut, and therefore under-estimate risk.

Conclusion on terrestrial risk

The above calculations allow a worst-case estimate of terrestrial risk despite the absence of data necessary to define the dermal exposure. The list of factors that suggest that such an exercise will overestimate exposure may imply that it is of limited use in risk assessment. However, if using the above set of worst case conditions based on contact with the soils and travel distance, the lethal dose is not reached then this would seem to be a robust determination of low risk.

COMPARATIVE RISK ASSESSMENT

Amphibians and fish

The analysis in the earlier section indicates that in most cases where a risk assessment has been conducted for fish, it should also be protective for amphibians. Where the fish risk assessment triggers concern that cannot be removed by refining the exposure assessment (which would increase the TER equally for fish and amphibians), or where there is reason to suspect that amphibians may be more sensitive then consideration could be given to conducting toxicity tests with amphibian larvae.

29

Amphibians and birds

The selection of species for comparison was limited by the small amount of data available for UK species. The above assessments for dietary and dermal exposure were used to predict the exposure of Rana temporaria to two compounds (tretrachlorvinphos and fenitrothion) for which there is LD50 data.

Dietary exposure

Using the estimates of consumption proposed above, the estimated TER for the standard insectivorous bird (blue tit) form SANCO (2000) was compared with estimates for the common frog Tables 17 and 18.

30

Table 17. Comparison of TERs for the standard insectivorous bird and common frogs (Rana temporaria) of different weights feeding on arthropods following application of tetrachlorvinphos. RUD values for insects are as in SANCO/4145/2000. Consumption, FIR and TER are shown for both mean and upper 95% estimates of FMR using the lizard equation in Nagy et al. (1999).

Species Weight Consumption (g) FIR Rate Diet RUD Toxicity TER  (g) Mean upper 95% Mean upper 95% (kg/ha)     (mg/kg) Mean upper 95%

Blue tit 10 10.4 - 1.04 - 0.5 Small insects 52 1500* 55.5 -Common frog 10 0.32 0.98 0.032 0.098 0.5 Small insects 52 151 182.1 59.5Common frog 20 0.60 1.62 0.030 0.081 0.5 Small insects 52 151 193.0 71.6Common frog 30 0.87 2.25 0.029 0.075 0.5 Small insects 52 151 199.7 77.4Common frog 40 1.14 2.89 0.028 0.072 0.5 Small insects 52 151 204.6 80.5Common frog 10 0.32 0.98 0.032 0.098 0.5 Large insects 14 151 676.4 221.0Common frog 20 0.60 1.62 0.030 0.081 0.5 Large insects 14 151 717.0 266.1Common frog 30 0.87 2.25 0.029 0.075 0.5 Large insects 14 151 741.8 287.6Common frog 40 1.14 2.89 0.028 0.072 0.5 Large insects 14 151 759.9 299.0* lowest bird value from Pesticide Manual

31

Table 18. Comparison of TERs for the standard insectivorous bird and common frogs (Rana temporaria) of different weights feeding on arthropods Following application of fenitrothion. RUD values for insects are is in SANCO/4145/2000. Consumption, FIR and TER are shown for both mean and upper 95% estimates of FMR using the lizard equation in Nagy et al. (1999).

Species Weight Consumption (g) FIR Rate Diet RUD Toxicity TER  (g) Mean upper 95% Mean upper 95% (kg/ha)     (mg/kg) Mean upper 95%

Blue tit 10 10.4 - 1.04 - 0.5 Small insects 52 23.6* 0.9 -

Blue tit 10 10.4 - 1.04 - 0.5 Small insects 52 1190** 44.0 -Common frog 10 0.32 0.98 0.032 0.098 0.5 Small insects 52 2220 2677.4 874.7Common frog 20 0.60 1.62 0.030 0.081 0.5 Small insects 52 2220 2837.9 1053.3Common frog 30 0.87 2.25 0.029 0.075 0.5 Small insects 52 2220 2936.2 1138.5Common frog 40 1.14 2.89 0.028 0.072 0.5 Small insects 52 2220 3008.0 1183.7Common frog 10 0.32 0.98 0.032 0.098 0.5 Large insects 14 2220 9944.5 3248.9Common frog 20 0.60 1.62 0.030 0.081 0.5 Large insects 14 2220 10540.7 3912.1Common frog 30 0.87 2.25 0.029 0.075 0.5 Large insects 14 2220 10905.8 4228.7Common frog 40 1.14 2.89 0.028 0.072 0.5 Large insects 14 2220 11172.6 4396.4

* toxicity to bobwhite quail** toxicity to mallard ducks

32

The relatively low FIR for these species compared to birds means that unless amphibians are significantly more sensitive, TERs and apparent risk will inevitably be far lower. For tetrachlorvinphos where it would appear that frogs are more sensitive there was still a far higher TER based on mean values.

Dermal exposure

Using this approach for two compounds (tetrachlorvinphos and fenitrothion) for which we have toxicity data for UK species (Rana temporaria) we can investigate the influence of the various factors on the predicted exposure. For tetrachlorvinphos, assuming an area of a 5cm circle (19.6cm2) and a bodyweight of 40g, it is clear that if all the pesticide applied was absorbed the animal would be close to a lethal dose without moving (Table 19).

Table 19. LD50 data for two pesticides and Rana temporaria (the common frog). The number of LD50’s that could be absorbed from a 5cm wide circle of soil (assuming an application rate of 0.5kg a.i./ha and a bodyweight of 40g) are also shown.

Compound LD50 LD50’s per 5cm circletetrachlorvinphos 151mg/kg 0.650

fenitrothion 2220mg/kg 0.044

Taking the LD50 by injection (Table 19) as an estimate of internal LD50, the estimated TER for dermal uptake by the same 5cm wide animal weighing 40g moving across an area treated at 0.5kg/ha with tetrachlorvinphos or fenitrothion is shown in Figures 2 and 3.

Figure 2. Effects of track length and % absorption on estimated dermal exposure to tetrachlorvinphos expressed as a TER assuming absorption from the surface of a 5cm wide track across an area treated at 0.5kg/ha. Each line represents a different track length ranging from 1m to 100m.

0.01

0.1

1

10

100

1000

0 10 20 30 40 50 60 70 80 90 100% absorption

TE

R

10m

20m

30m50m

100m

5m

1m

Figure 3. Estimated dermal exposure to fenitrothion measured as the number of LD50s assuming a range of % absorption from the surface of a 5cm wide track across an area treated with fenitrothion at 0.5kg/ha. Each line represents a different track length ranging from 1m to 100m.

0.1

1

10

100

1000

10000

0 10 20 30 40 50 60 70 80 90 100% absorption

TE

R

10m

20m30m50m

100m

5m

1m

33

Clearly for tetrachlorvinphos a lethal dose would be accumulated after a very small distance assuming that it is absorbed readily. For fenitrothion even assuming 100% absorption animals would have to travel over 30m before a lethal dose was obtained. The effects of including interception factors can easily be included and in this case an interception of 0.7 as might occur at full canopy would suggest that the animal would not receive a lethal dose even after 100m. However, in both cases the trigger value for the TER of 10 are readily breached unless the animal move only a short distance or % absorption can be reduced to very low values on the basis of other information.

The effects of a jumping style of locomotion of different distances compared to a continuous track are shown in Figures 4 and 5.

Figure 4. Estimated dermal exposure measured as the number of LD50s per m travelled across an area treated with tetrachlorvinphos at 0.5kg/ha assuming a range of % absorption from the surface. Each line represents a different ‘hop’ distance with 0cm representing contact with a complete 5cm wide track and 25/50cm representing two hop lengths assuming contact area of a 5cm circle at each hop.

Figure 5. Estimated dermal exposure measured as the number of LD50s per m travelled across an area treated with fenitrothion at 0.5kg/ha assuming a range of % absorption from the surface. Each line represents a different ‘hop’ distance with 0cm representing contact with a complete 5cm wide track and 25/50cm representing two hop lengths assuming contact area of a 5cm circle at each hop.

Clearly the extent of dermal uptake could be estimated to be very high or very low relative to the lethal dose depending on the values chosen to estimate it.

Another approach to identify the risk is to calculate the maximum distance that could be travelled before a trigger value (e.g. TER = 10) is breached and consider whether this is plausible e.g.

TL = Bwt x LD50 TER x TW x (APPRATE x 100) x ABS

where:

TL track length for TER of interestTER TER of interest (e.g. 10)TW track width (m)Bwt bodyweight (kg)LD50 toxicity (mg/kg)APPRATE application rate (kg a.i./ha)*ABS absorption as a proportion (e.g. % absorption/100)

* this is multiplied by 100 in the equation to convert to mg/m2

This provides the track length where the entire track is contacted. To adjust for reduced contact (e.g. due to hopping locomotion) this track length should be divided by the estimate of proportion of track contacted (e.g. as in Table 16 above)

The results of carrying out these calculations for the two examples under consideration where TER = 10 are shown in Table 20.

Table 20. Effects of hop distance from 0cm (continuous 5cm wide track) to 50cm on the maximum distance that can be travelled before the estimated TER drops below 10.

Maximum track length for TER ≥ 10 (m)% Tetrachlorvinphos Fenitrothion

absorption 0cm 25cm 50cm 0cm 25cm 50cm1 24.16 153.75 307.49 355.2 2260.4 4520.75 4.83 30.75 61.50 71.0 452.1 904.1

10 2.42 15.37 30.75 35.5 226.0 452.120 1.21 7.69 15.37 17.8 113.0 226.030 0.81 5.12 10.25 11.8 75.3 150.740 0.60 3.84 7.69 8.9 56.5 113.050 0.48 3.07 6.15 7.1 45.2 90.460 0.40 2.56 5.12 5.9 37.7 75.370 0.35 2.20 4.39 5.1 32.3 64.680 0.30 1.92 3.84 4.4 28.3 56.590 0.27 1.71 3.42 3.9 25.1 50.2

100 0.24 1.54 3.07 3.6 22.6 45.2

These results indicate clearly that for tetrachlorvinphos in this example, very short movements would lead to a significant dose.

Calculating total exposure and TER

Using the above models total exposure via dietary and oral routes can be defined as:

ETE = DIETARY + (TW x TL x (APPRATE x 100) x INT x CONTACT x ABS)Bwt

TER = LD50 / ETE

LD50 toxicity (mg/kg)Bwt bodyweight (kg)DIETARY predicted dietary exposure (FIR x APPRATE x RUD)TW track width (m) based on estimate of width of animal (0.05m in our examples)TL track length being considered (m)APPRATE application rate (kg/ha)*INT interception factor for crop CONTACT contact factor based on actual contact with track where there is less than 100% due to actual

area in contact or hopping locomotion (may be calculated as a combination of both e.g. proportion of contact x hop factor)

ABS absorption as a proportion (e.g. % absorption/100) based on any evidence that would suggest that <100% of pesticide contacted would be absorbed.

* this is multiplied by 100 in the equation to convert to mg/m2

For an extreme worst case estimate of dermal exposure, as for our continuous track, exposure would be defined by TW and APPRATE assuming a TL of say 100m and setting INT, CONTACT and ABS to 1. It would also be possible to include an estimate of overspray exposure based on estimated plan area and application rate (e.g. as in Table 19).

CONCLUSIONS AND RECOMMENDATIONS

Toxicity data

While the comparisons of amphibian larvae with fish, and reptiles with birds may indicate some similarities in toxicity, amphibians at least are known to be more resistant to OP compounds and more sensitive to some other compounds than could be predicted from toxicity to other groups (Hall and Henry 1992). Most of the available

data is for larval anurans where such differences in susceptibility may be detected, but where there is scant data, such as toxicity data for reptiles and post-metamorphic amphibians, these predictions made using toxicity values for other species should be made with far more caution.

Risk assessment process

Given the results of our analysis we recommend:

1. Use of fish assessments to protect larvae/tadpoles.2. Use of estimated intake data to predict dietary exposure of adult amphibians and reptiles.3. Use of simple dermal exposure calculations to test whether exposure via this route for credible worst

case conditions is likely to put animals at risk.

However, the paucity of toxicity data for adult amphibians and reptiles makes it difficult to conduct terrestrial risk assessments. The comparison of bird and amphibian LD50 data indicates that bird data may be a poor predictor except for compounds such as OPs where amphibians show low sensitivity. One possibility might be to use bird data with a high uncertainty factor (e.g. 100) as a screening assessment. It is likely that the dermal risk will be the most significant factor and this is the area in need of the most refinement. The simplest option for addressing this would be the collection of data on the actual area in contact with the ground while stationary and during locomotion to refine the estimate of contact with the treated soil. This could be obtained with captive animals in arenas perhaps using dyes to record tracks. Further work would be necessary to determine actual absorption under semi-natural conditions.

Given the results of this study we suggest the following risk assessment process for amphibians and reptiles.

Tier 1

Worst case

Assume animal moving 100m across a treated area, assume 100% absorption, use bird toxicity (assuming no amphibian value available), use application rate as soil concentration and calculate TER

If TER < 10 (for anticholinesterase pesticides) or 100 (for other classes of pesticides) consider refinements.

Unconservative case

Assume stationary frog absorbs pesticide from the soil it is standing on (if consider plan area, e.g. as 5cm circle in example, this would be equivalent to overspray.

If TER < 10 (for anticholinesterases) or 100 (for others) then there is a high concern.

Larvae

Use fish assessment. If there is concern go to Tier 2.

Tier 1a

Make calculations/plot graphs (e.g. Figure 5) taking account of factors affecting estimate of % absorption such as method of locomotion, interception factor etc..

Consider plausibility/remoteness of risk in context of crop, available ecological data etc..

Tier 2

Larvae

Refine exposure assessment in same way as for fish. If concern remains, consider conducting toxicity tests with larval stages.

Adults

1) Get better generic data on ecology/behaviour (possible research may be needed).

This would include data on behaviour and contact with the ground to refine worst case estimates of adsorption and data on movements and use of arable fields. This may require fieldwork such as a radio-tracking study of the type that has been conducted for birds and mammals to determine an estimate of PT (Crocker et al. 2002).

2) Measurement of:

a) toxicityb) absorption

It would be possible to make these measurements in the laboratory using semi-field arena type studies subject to consideration of ethical issues. This should be conducted at a range of rates to check the margin of safety for other untested species. If possible the possibility of recovery and re-exposure should also be considered.

Further work

Based on the assessments presented above, the greatest sources of uncertainty concerning risks to amphibians and reptiles comprise:

Oral and dermal toxicity to adult stages

This could be addressed either by requesting studies from notifiers on a case-by-case basis or by testing a range of pesticides of different classes, to establish generic extrapolation factors between birds and amphibians/reptiles (including appropriate allowance for uncertainty, and for variation between different species of amphibians and reptiles). To minimise use of animals, it may be preferable to use LD50s based on dosing by injection to assess the sum of dermal and dietary exposure, rather than considering oral and dermal toxicity separately.

Dermal exposure

Work could be conducted to refine estimates of dermal exposure. The simplest refinement would be to obtain accurate measurements of contact with the ground to improve the estimate of maximum dose that could be absorbed making the assumptions about absorption with water. This could be backed up by arena or field studies to address track lengths and modes of locomotion (e.g. patterns of walking, hopping etc.).

Design of special studies for regulatory assessment

If it were intended to request special studies to assess exposure and effects on amphibians/reptiles then it would be desirable to conduct preliminary research to establish a practical protocol, as a basis for advising notifiers on appropriate design and evaluating submitted studies.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

Anon. (1995) Biodiversity: The UK Steering Group Report. Volume 2: Action Plans. HMSO, London.

Anon (2002) Working document Guidance document on Risk Assessment for Birds and Mammals Council Directive 91/414/EEC SANCO/4145/2000. (See http://europa.eu.int/comm/food/fs/ph_ps/pro/wrkdoc/wrkdoc19_en.pdf)

Appleby L G (1971) British Snakes. John Baker (Publishers) Ltd. London.

Arnold H R (1995) Atlas of amphibians and reptiles in Britain. HMSO, London.

Avery R A (1966). Food and feeding habits of the common lizard (Lacerta vivipara) in the West of England. J. Zool. Lond. 149:115-121.

Avery R A (1968) Food and feeding relationships of three species of Triturus (Amphibia: Urodela) during the aquatic phase. Oikos 19:408-412.

Avery R A (1971). Estimates of food consumption by the lizard Lacerta vivipara Jacquin. J. Anim. Ecol. 40:351-365.

Baker J (1997) The Declining Amphibian Populations Task Force. British Herpetological Society Bulletin 59:35-38.

Ballasina D (1984) Amphibians of Europe. David and Charles (Publishers) Ltd.

Beebee T J C (1971) Changes in status of the great crested newt Triturus cristatus in the British Isles. Brit. J. Herpetol. 5:515-521.

Beebee T J C (1981) Habitats of the British amphibians (4): Agricultural lowlands and a general discussion of requirements. Biological Conservation 21:127-139.

Berrill M, Coulson D, McGillivray L and Pauli B (1998) Toxicity of endosulfan to aquatic life stages of anuran amphibians. Environ. Toxicol. Chem. 17(9):1738-1744

Birch W X. and Mitchell R A. (1986) Effects of nabam on developing Xenopus laevis embryos: minimum concentration, biological stability, and degradative products. Arch. Environ. Contam. Toxicol. 15:637-645.

Birge W J, Westerman A G and Spromberg J A (2000) Comparative toxicology and risk assessment of amphibians. In: Ecotoxicology of amphibians and reptiles. Sparling DW, Linder G and Bishop CA eds. SETAC Press, Pensacola. pp. 727-791.

Blaustein A R and Wake D B (1995) The puzzle of declining amphibian populations. Scientific American 272:56-61

Chenoweth M B. (1949) Monofluoroacetic acid and related compounds. J.Pharmacol.Exp.Ther. 97:383-424

Cooke A S and Arnold H R (1982) National changes in status of the commoner British amphibians and reptiles before 1974. British Journal of Herpetology. 6(6):206-207

Cooke A S and Ferguson P F (1976) Changes in status of the frog (Rana temporaria) and the toad (Bufo bufo) on part of the East Anglian fenland in Britain. Biol. Conserv. 9:191-198.

Cooke A S (1975) Spawn site selection and colony size of the frog (Rana temporaria) and the toad (Bufo bufo). J. Zool. 175:29-38.

Cooke A S (1985) The warty newt Triturus cristatus at Shillow Hill: Numbers and Density. Report of the Huntingdon Fauna and Flora Soc. 37:22-25

Cooke A S (1986) The warty newt Triturus cristatus at Shillow Hill: ranging on arable land. Report of the Huntingdon Fauna and Flora Soc. 38:40-44

Cooke A S (1995) A comparison of survey methods for crested newts (Triturus cristatus) and night counts at a secure site, 1983-1993. The Herpetological Journal 5(2):221-228.

Cooke A S (1997) Safeguarding herpetofauna from agrochemical use in the designed landscape. In English Nature Science Series No. 30. Opportunities for amphibians and reptiles in the designed landscape. Eds. Bray, R and Gent, T.

Corbett K (1989) Conservation of European Reptiles and Amphibians. Helm, London.

Cowman D E and Clark D R (1995) Comparative toxicities of 2 common agricultural chemicals to toads.