1989 1990

.;••

Natural Environment Research Council Report of the Institute of Terrestrial Ecology 1989/1990

Natural Environment Research Council Contents

Science Reports 33 - 70

Environmental pollution 33 The effect of hill cloud on pollutant deposition Variations in solution chemistry under four tree species at Gisburn Effects of acid waters on the survival of frogs Spray drift impacts on non‐target organisms Climate change ‐ the mild winter of 1989‐90

Population ecology 47 The effect of a plant compound on reproduction in rabbits Competition between red and greay squirrels Predator/prey relationships and species conservation: the otter Effects of roads on butterfly and burnet populations Land inheritance and population cycles of red grouse Effects of land use changes on wildlife

Community ecology 59 Beetles on Snowdon Scrub succession at Castor Hanglands NNR Nitrogen cycling in montane tropical rain forest

Scientific services 66 Automated dry/wet deposition separator of aerial precipitates (ADDSOAP) Pollution research facilities at ITE's Bangor Research Unit Analytical chemistry group Environmental pollution

This Research Programme aims to study The next report discusses another, concentrations of major anions, sulphate the input of pollutants to terrestrial and although very different, impact study (SO421 and nitrate (NO3.), with which it is freshwater ecosystems, their transfer which examines the influence of associated, shows largest values in the through these systems and through food agricultural pesticides on non-target east Midlands and East Anglia and the chains, and their impact on plant and insects in areas adjacent to sprayed smallest values in north-west Scotland In animal populations. ground Such information is important marked contrast, the wet deposited because many of our insect species acidity shows largest values in the high- A group at ITE Edinburgh has been survive in refuges such as hedgerows, rainfall areas of north and west Britain. investigating the deposition of and because many protected sites, eg The areas most sensitive to acidic inputs atmospheric pollutants for more than 15 Nature Reserves, where insects breed include many of the high-raiinfall, high- years. Early estimates of the amount of are surrounded by agricultural land. The altitude regions of Wales, Cumbria and acidic deposition falling on the different results of such studies help to defme the western Scotland (Acid Waters Review parts of the UK were based mainly on size of the buffer zone required to Group 1986). However, the rainfall collection and analysis of rainfall in open protect sites from the impacts of drifting chemistry monitoring networks from collectors. Working with others, the spray. which the acidic inputs are estimated are Edinburgh group has shown that this confined to low ground (<400 m). In approach is unsatisfactory as it ignores, The final paper illustrates an area of order to estimate wet deposition on hills, or grossly underestimates, the work which will become increasingly rainfall composition has been assumed to contribution of a number of important important in ITE in the coming years - remain constant with altitude. deposition processes. The first article in the assessment of the impacts of climatic this section discusses one of the change. In this example, the influence of A collaborative project between ITE and processes that leads to enhanced the mild winter of 1988-89 on plant, the University of Manchester Institute of deposition of acidic pollutants in upland insect and animal populations is Science and Technology (UMIST) is areas. This kind of work has led to new examined for indications of the changes testing this assumption, in field studies of estimates of the amounts of acid which might be expected if such mild the rain chemistry and hill cloud pollutants being deposited in the UK, and winters become a regular feature of composition with altitude in Cumbria and has shown that in some areas the total British weather. southern Scotland. The measurements deposition is almost double the earlier show that, in the presence of hill cloud estimates. enveloping the high ground, rainfall from higher levels intercepts the hill (cap) The amount of acidic pollution deposited The effect of hill cloud on cloud droplets and increases the amount at a given location is partly determined pollutant deposition of rain at the higher elevations by the by the nature of the vegetation canopy. (This work was partly funded by the well-known seeder-feeder process The composition of the incoming rainfall Department of the Environment) (Figure 29). This process is the principal is then modified as it passes through the cause of the orographic enhancement in vegetation canopy and the underlying The deposition of acidity and sulphate in rainfall over the uplands in western soils. The second article illustrates the rain is being monitored routinely Britain. Although the seeder-feeder impact that variations in tree species can throughout the UK (Review Group on mechanism is widely recognised, the have an atmospheric deposition and on Acid Rain 1990). The geographical ionic composition of the hill cloud and its the changes in chemistry which take pattern in mean acidity, along with the effect on wet deposition is new. In place as solutes are transferred through the soil/plant system. The choice of species can have a major impact on inputs and solute transfers to waters. Frontal cloud Rain (seeder) Acidification of surface waters in parts of Cap cloud 200 m 800 m the UK has led to a decline in S0i- populations of salmonids and in the diversity of invertebrate populations. NOi There has been little work, however, on H+ the impact of acid waters on amphibians. The third article describes experiments which examined the response of captive populations of the common frog to Valley -200 m variations in acidity; it illustrates the difficulties inherent in any attempt to Figure 29. Diagrammatic representation of the seeder-feeder process for assess impacts of pollutants on populations. enhancing concentrations of ions in rain on hills

33 addition to marine-derived sodium ion moist air is lifted over the hills. The cloud magnitude of this increase varies with (NAt), chloride (a) and sulphate (SO42 ), droplets are typically 3-15 Ilrn radius meteorological conditions but is typically the atmospheric aerosols contain SO42 and, unlike the aerosols on which they a factor of two between the valley (<200 NO3 and hydrogen ion (H+), resulting were nucleated, they are efficiently m) and the summit (847 m) sites. The hills from anthropogenic emissions of sulphur washed out by falling rain. At the sites of in Cumbria at 850 m above sea level are dioxide (SO2) and oxides of nitrogen measurements in Cumbria (Great Dun in cloud on some part of 220 days in each (N0x). These aerosols are present in the Fell and Corney Fell), the concentrations year, and the annual duration of cloud at size range 0.1-1.0 I,tum radius; they are of major ions in hill cloud are large such sites is in the order of 2000 h; these too small to deposit from the atmosphere relative to those in rain, as shown in sites are generally in cloud during frontal by gravitational sedimentation, and are Figure 29. rain. The enhancement of wet deposition not washed efficiently from the is, therefore, a common feature of this atmosphere by falling rain. They do, The larger concentrations in cap cloud area. The mechanism of seeder-feeder however, form efficient nuclei for the than rain upwind of the hill lead to an enhancement of wet deposition over formation of cloud droplets, and are increased concentration of major ions in Great Dun Fell has been simulated using readily activated into cloud droplets as rain with altitude (Fowler et al. 1988). The a mathematical model (Choularton et al.

Occult deposition: sulphur Wet deposition: sulphur

mg m-2 yr-1 mg m-2 yr-1

above 2000 above 160 1700 - 2000 80 - 160 1400 - 1700 LII 40 - 80 1100 - 1400 20 - 40 800 - 1100 10 - 20 bebw 800 below 10

N MENEM MI M-M -:M

Figure 30. Annual inputs of atmospheric sulphur in rain and snow Figure 31. Annual inputs of atmospheric sulphur by cloud water to the UK (excluding marine sources). The map includes the effects deposition to the UK (excluding marine sources), 1986-88 of seeder-feeder 'scavenging' of hill cloud. 1986-88

34 1988) which reproduces quantitatively the feeder cloud water was assumed to be a Cumbria, and the west central Highlands observed changes in wet deposition with factor of two larger than that in the of Scotland, the wet deposition of altitude. Both the experiments and the seeder rainfall (at Great Dun Fell the sulphate, nitrate, ammonium and acidity model have been successfully extended average concentration of feeder cloud is increased by up to 70% ove7 the to a series of hills across Cumbria, from was four times that of the seeder rain). earlier estimates. On the lower ground the coast at Corney Fell to the Pennines The enhanced rainfall at upland sites (<300 m), the effects are ,/ery small and at Great Dun Fell. over coastal rainfall rates was assumed to do not significantly alter the map. result entirely from seeder-feeder The satisfactory simulation of the scavenging. Using these simplifying These new estimates of wet deposition observed wet deposition from these assumptions, revised wet deposition are particularly important for studies of studies has been used to extrapolate to estimates incorporating seeder-feeder stream- and groundwater chemistry in the upland areas of the rest of the effects have been calculated for the catchments. The improved understanding country (Fowler et al. 1990). For this entire country, and Figure 30 shows the of inputs is being used to construct extrapolation the average concentration results for non-marine sulphate. At catchment budgets for sites being of individual ions in the 'scavenged' mountain-top sites in Snowdonia. monitored intensively in Wales. However, further work is necessary to show whether the extrapolation to remote areas of north and west Scodand is correct.

The hill cloud droplets are deposited efficiently on vegetation, and studies of droplet deposition in collaboration with the Department of Applied Physics at UMIST have shown the drop size dependence of the deposition rate. The measurements use micrometeorological techniques above the plant canopy to estimate rates of droplet deposition. The equipment shown in Plate il includes profile anemometers, three-dimensional sonic anemometers, cloud droplet sizing instruments (Knollenberg FSSP), and a rapid-response cloud hquid water content instrument (PVM). These studies have enabled a model to be developed of the deposition of cloud water on hills in the UK (Gallagher et a/. 1988; -Fowler et al 1990). The model uses the aerodynamic roughness of vegetation, deduced from vegetation height with wind speed. and cloud statistics, to estimate annual inputs to each 20 km grid square for the country. The results for cloud deposition of sulphur are shown in Figure 31.

These studies of wet deposition and cloud water deposition form a component of the Department of the EnvHonment's pollutant deposition and effects research programme, and have been incorporated in the recent review of acid deposition in the United Kingdom (ROAR 1990).

D Fowler

References

Acid Waters Review Grcinapi 1986. Acidity in United Kingdom fresh waters. Londbn: HMSO.

Choularton, T.W., Gay, M.N., Fowler, D., Cape, Ineil,Ca, 1988. The influence of altitude on wet deposition, comparison between field measurement at Great Dun Fell and the predictions of a seeder-feeder model. Atmos Environ., 22, 1363-1371.

Plate 11. Equipment to measure rates of cloud droplet deposition on a high-elevation forest in Fowler, D., Cape, Deith, the Scottish borders, using micrometeorological techniques Chou lexicon, T.W., Gay, :KJ. 6.,.:q(usces,

35 1988. The influence of altitude on rainfall Table 6. Contribution of stemflow as a percentage of total annual input (stemflow + throughfall) composition at Great Dun Fell. Atmos. to the forest floor, Gisburn Environ., 22, 1355-1362. Water volumes Acidity Fowler, D., Morse, A.P., Gallagher, M.W. & Scots Norway Scots Norway Choularton, T.W. 1990. Measurements of pine spruce Oak Alder pine spruce Oak Alder cloud water deposition on vegetation using a lysimeter and a flux-gradient technique. Tellus, 42. In press.

Gallagher, M.W., Choularton, T.W., Morse, A.P. & Fowler, D. 1988. Measurements of the deposited acidity. Relatively more water Water chemistry profiles - size dependence of cloud droplet deposition is contributed by stemflow in spruce The acidity of water measured at any at a hill site. Q. Jl R. met. Soc., 19, 1291-1304. (15-16%) than in the other stands stage in its passage through the profile, Review Group on Acid Rain. 1990. Review (10-12%), although this difference is not from incident rain to soil runoff, will be group of acid deposition in the United reflected in the acidity. The proportion of the net outcome of acidifying and Kingdom. London: HMSO. the total annual acidity reaching the neutralising processes. The former forest floor as stemflow is higher for both include acid pollutants captured by the deciduous broadleaves (c 25%) than in canopy (mainly SO42- and nitrate (NO3)), the two conifers - not through absolute the creation of organic acids associated Variations in solution increases in stemflow acidity, but with organic matter turnover and the chemistry under four tree because of reductions in throughfall release of NO3 (especially following N acidity during the summer in-leaf period. fixation in the case of alder), and the species at Gisburn The stemflow percentage for Scots pine imbalance in ion uptake by roots. In (c 14%) is the smallest in acidity terms, practice, inputs of gaseous ammonia It is well established that the acidity and with Norway spruce intermediate at c (NH3) - converted to the ammonium chemistry of rainwater change as it 20%. For all four stands, between-year (NI-14) ion and, therefore, having some passes through a forest canopy, and that variation was slight, suggesting that theoretical neutralising potential - also the proportion of such water flowing stemflow percentage (volume and have an acidifying effect. Its nitrification down the main branches and bole acidity) is a function of stand to NO3 by bacteria in the forest floor or (stemflow) may be modified to a different characteristics rather than weather. This soil, and its role in enhancing the canopy extent from the water falling only through finding contrasts with interception absorption of sulphur dioxide (SO2) the canopy (throughfall). These percentages, which are very dependent through co-deposition can both lead to modifications to water chemistry are not, on the pattern and intensity of rainfall. increases in acidification. The various however, the only ones occurring as the acidifying influences are mainly water passes further through the As an example of the importance of counteracted by exchange and ecosystem, prior to leaving the site as stemflow in the transfer of other ions from adsorption processes. The different drainage water: the forest floor and the canopy to the forest floor, the data for species of tree at Gisburn appear to vary different horizons of the soil also exert alder are shown in Table 7. Such transfer considerably in these respects, and an their influence. Few studies compare the depends not only on the relative attempt is made here to interpret the relative importance of these various quantities of water involved, but on ionic variations in acidity (with both species modifications to solution chemistry for concentrations. Because stemflow and profile position) by reference to the different species at the same site. Data contains higher concentrations than changes in the other measured ions. for acidity and solute chemistry are throughfall of sodium (Na), chloride (C1), presented here for stands of four tree sulphate (SO42-) and calcium (Ca), in Figure 32 summarises the acidity (shown species - Scots pine (Pihus sylvestris), addition to acidity, the importance of the as H ion concentration) in the water of Norway spruce (Picea abies), alder stemflow pathway for these ions is the various sampled strata (Figure 33) (Alnus glutinosa) and oak (Quercus greater than that indicated by the under the four tree species. In the petraea) - established at Gisburn in 1955. proportion of stemflow water alone. throughfall, there is a clear distinction Conversely, stemflow represents a between conifers, in which acidity is The comparisons between throughfall smaller than expected route for both identical and appreciably greater than in and stemflow were made in the two-year forms of nitrogen (N), phosphate- the incident rain, and broadleaves, which period, November 1983 to October 1985. phosphorus (PO4-P) (averaged for the also have almost identical acidities to The more complete water chemistry two years) and potassium (K). each other, although appreciably less profiles for bulk precipitation, throughfall than the conifers and marginally less than (only), forest floor leachate and soil However, whereas the above in the rain. In the forest floor waters, the leachates from the base of the Ah horizon percentages apply to the unit area of a conifer/broadleaf split disappears: acidity and from within the B/C horizon were stand, stemflow is cdncentrated round the under pine and alder rises sharply, that determined in 1987 as part of a CASE bases of the trees. Therefore, the under oak much less so, and spruce studentship by M Iles, jointly supervised localised effects of stemflow must be scarcely at all. Relative differences by ITE and Professor H G Miller considerable, especially in respect of between species alter yet again as some (University of Aberdeen). acidity inputs. neutralisation occurs in the upper mineral

Stemflow and throughfall Table 7. Contribution of stemflow as a percentage of tOtal annual flux (stemflow + throughfall) comparisons to the forest floor for other ions. Alder stands, Gisburn The contributions of stemflow, per unit area of stand, to the total input of Na K Ca Mg PO4 NO4 NH4 CI SO42- throughfall and sternflow are shown for 11.7 8.8 12.7 12.4 5.9 4.7 4.7 11.5 18.7 each tree species and year in Table 6, .in Nov 1983-Oct 1984 Nov 1984-Oct 1985 11.8 7.9 14.6 10.0 11.3 4.1 4.5 12.4 19.2 terms of both water volume and

36 . . . • inddrii ecipitition releases are similar): the two conifers thereby diverge markedly in acidity • - - . terms within this stmturn. The large rise in acidity associated with the forest floor waters of alder, which become almost as acid as those of pine, mainly relates to the fixation of N and its subsequent nitrification to NO3. Despite the partial neutralisation of this NO3- based acidity by a large concomitant increase in the release of all three bases, the net acidity remains high. Waters under oak are the least acid of the four species throughout the profile. As a deciduous broadleaf, pollutant SO42" (and NO3) inputs are low under oak, and the concentration of NO3 in both the forest floor and soil waters remains as low as . . . that in rain, reflecting low organic matter Stemflow Throughfall • .. . . TF turnover and associated N mineralisation in this species. Forest floor (L+F+H) FF Although the high levels of acidity in the forest floor are subsequently reduced in Upper mineral soil (loamy) Ah A the Ah mineral soil horizon, in all four species, this reduction is not generally associated with any overall change in the Lower mineral soil (gleyed clay) B/C 15 crn1 BC total concentration of base cations. Although Ca levels do rise slightly, this Figure 32. Profile showing strata in which waters increase is counterbalanced by the were sampled for chemical analysis sharply reduced release of K in this stratum. Rather, in this upper soil horizon, aluminium (Al) mobilisation plays the soil, acidity under alder falling especially concentrations of total organic carbon in major neutralising role, the different steeply. Only in the deeper, clayey B/C waters from this horizon, followed by levels of which - in this horizon only - horizon is the buffering marked enough sharp falls in soil waters. Spruce again reflect those for acidity in the four to cause convergence of all four species apparently counteracts this potential species. In contrast, in the B/C mineral to a level of acidity less than that of the acidification in the forest floor much soil horizon, Al concentrations decline incoming rainwater. How are these better than pine, by additional Ca again, and buffering is provided by differences reflected in the other ions, release (in this stratum, Mg and K considerable increases in Ca release, presented as concentrations in Table 8?

The potentially acidifying Cl ion is more- or-less fully balanced by its associated Na ion, and hence both of these ions can be ignored in the present discussion. Similarly, in the rain, throughfall, and (except for alder) the forest floor water, NO3 is balanced by NH4, PO4 levels range from insignificant to small, and hence SO42- represents the main acidifying anion, except in the mineral soil strata where NO3 may appreciably cr exceed NH4. Despite the greater ci) 80 .• • concentration of 5042- in spruce 4 • " throughfall compared with that of pine (the denser spruce canopy possibly 260 -o trapping more pollutant sulphate-sulphur (SO42--S) through dry and occult 40 deposition and/or having a more leaky canopy for SO42-), the resulting net 0/ acidity is identical in the two conifers. 20 This result can be attributed to the greater leaching from spruce canopies, relative to those of pine, of all three 0 bases, Ca, Mg, and especially K, in TF FF A BC exchange for some of the SO42-- Level generated acidity. The forest floor is the Figure 33. Annual means of acidity (H ion concentration Reg main seat of organic acid production for 1-1), based on fortnightly sampling, December 1986-December all species, reflected in the steep rises in 198Z at Gisburn

37 Table 8 Mean annual concentrations (Åeq 1-I) of major ions in water profiles under four tree species at Gisburn, 1987 TF = throughfall; FF = forest floor leachate; A = Ah horizon (upper mineral soil) leachate; BC = B/C horizon (lower mineral soil) leachate

Species K Al NH4-N H pH 8 <1 27 37 4.43

Oak 34 <1 43 23 4.63 34 20 39 60 4.22 10 58 1 37 4.43 6 18 0 5 5.29

Alder 32 <1 25 30 4.52 66 40 44 136 3.87 20 119 <1 63 4.20 7 24 0 9 5.04

Spruce 95 7 99 72 4,14 74 40 61 82 4.09 15 105 1 70 4.16 8 34 0 7 5.15

Pine 57 14 99 71 4.14 72 38 136 150 3.82 33 167 3 107 3.97 13 80 0 20 4.70

Note Pine data based on only six replicates per 26 occasions other species 18 replicates Aluminium is assumed to be Al+++ which is especially high in the two waters under oak (production of NO3 and organic acids) conifer soils where it is supplemented by will almost certainly vary with site, as some increase in magnesium (Mg). K The likely predominance of drainage may the acidifying imbalance between levels are low in all B/C waters. from the B/C horizon in summer would cation and anion uptake. Nevertheless, permit the greater buffering of the acidity for polluted sites, evergreen conifers are The soil water data provide an in this deeper clayey soil in all four likely to have more adverse effects than indication of the chemistry of drainage species, but at the cost of increased Ca deciduous broadleaves on the acidity of waters from the different stands. In losses. In contrast, relatively large soil waters and the other associated general, high levels of acidity, Al and species differences remain in the B/C solute levels noted here. Similarly, NO3 are to be avoided from the point of waters for NO3, Mg, and Ca. These through its N fixing properties, alder is view of aquatic biota and water differences can be summarised as follows. likely to lead to more of the undesirable potability; losses of base cations also consequences than other native represent reductions in the site's fertility Acidity of Ah horizon waters: broadleaves. Its use as a buffer between status. (The soil waters contain zero or pine > spruce = oak conifer stands and watercourses could negligible levels of PO4 and ammonium- well, therefore, be counter-productive. nitrogen (NH4-N).) Al concentration in Ah horizon waters: A H F Brown pine > alder = oak As a result of the acidifying influences within these four ecosystems, and the NO3 concentration in Ah and B/C horizon release of other ions in response to them, waters: the chemistry of the drainage waters is alder > pine = oak likely to vary appreciably according to Effects of acid waters on the tree species. In addition, however, survival of frogs because the chemistry of the soil waters Base cation concentrations in B/C horizon (This work was funded by the Central usually differs markedly between the Ah waters: Electricity Generating Board and the Nature and the B/C horizons, the depth at which spruce = pine > alder > oak Conservancy Council) the drainage originates will also be important in terms of potential losses of Can these tree species differences for In certain upland regions in Britain, solutes. At present, this depth is not Gisburn be extrapolated to other sites? surface waters are either naturally acid known for Gisburn, but is likely to vary Such extrapolation would need great or susceptible to acidification: many such with season, water probably flowing care because the relative degree of waters have become more acid as a mainly from the less-buffered upper soil 'scavenging' of acidic atmospheric result of the deposition of airborne very wet conditions) in winter when pollutants by the different species of tree pollutants, and changes in soil water and these gley soils are normally saturated - notably the apparently large difference drainage associated with afforestation. lower down. Waters under pine are the between evergreen conifers and Numerous laboratory studies have most acidic and most Al mobilisation deciduous broadleaves - may vary with demonstrated sublethal and lethal effects occurs at either soil depth, together with pollution levels. The relative responses of of acid conditions on embryonic and high levels of total base cations the trees to these potentially acidifying larval amphibians, and evidence of acid- (comparable with those from spruce); pollutant anions by cation exchange related stress and mortality has been NO3 mobilisation is second only to that could also vary with tree age. Further, found in the field. However, a lasting under alder. Conversely, levels of acidity the endogeneous acidifying factors impact of acid conditions on amphibian and most other ions are lowest in the soil related to organic matter turnover populations has only been inferred. 38 ITE has a long-standing commitment to D 1 mg Ca carry out ecotoxicology studies where the primary objective is to benefit wildlife. 4 mg Ca rl As part of a study of the impact of acidification on the common frog (Rana temporaria), an experiment was carried out at ITE Monks Wood in 1989 which was designed to evaluate the single and combined effects of some conditions associated with acid waters on the quality and quantity of froglets arising from populations of tadpoles. 0 Earlier laboratory experiments had 90 180 360 90 180 360 90 180 360 shown that low pH and elevated levels of Non-acid pools Acid pools Acid pools + Al dissolved aluminium could reduce rates Figure 34. Tadpole survival during the period prior to the of growth and development of individual earliest metamorphosis. Each bar shows the mean for three pools tadpoles (Cummins 1989). The laboratory Values below bars show initial numbers of tadpoles per pool. results supported the hypothesis that Early survival was lower in acid water than in neutral water, and lowered tadpole density, resulting from lower still in acid water with aluminium added. In acid water, acid-related embryo mortality, could lead survival was lower at the lower of the two calcium concentrations to faster growth and development among replicated in three randomised blocks (A, 8 and C). surviving tadpoles, in spite of the direct suppressive effects of low pH. The latest experiment was performed out-of-doors, earliest metamorphosis, decreased with achieved metamorphosis (Figure 35). with larger populations of tadpoles, and increasing density. This effect of density included a greater number of variables was weakest in acid pools with additional Acid pools in block A were more acid than previous experiments. aluminium, where early mortality was than those in blocks B and C. In blocks B high. Thus, density-dependent mortality and C, early mortality in acid pools was appeared to be ameliorated by mortality. offset by a reduced density-dependent Common frog tadpoles were raised to caused by acid conditions. Mortality was mortality, to the extent that there was no metamorphosis in 54 pools, each of 200 strongly density-dependent later in the significant net effect of acid conditions on litres, arranged in three blocks of 18, in experiment; there was a significant the proportion of introduced tadpoles which initial tadpole density and water negative correlation between the number metamorphosing. In block A, early chemistry were manipulated. Water in of pre-metamorphic survivors in a pool mortality in acid pools was greater than the pools was made up to simulate and the proportion of those tadpoles that the resultant reduction in density- acidified and unacidified natural soft waters. In each block, one-third of the pools were initially circumneutral (pH 7), and two-thirds were acid (pH 4.1 or pH 90 • OM A • • 90 Tadpoles 4.3); aluminium (800 g Al 1-1) was added • 1 80 to six of the 12 acid pools. Two further A360 in pool treatments - initial calcium concentration (1 or 4 mg Ca 1-/) and initial tadpole density (90, 180 or 360 tadpoles pool-1)- a were applied factorially to the six pools • A with each combination of pH and • • A •• • aluminium. The experiment thus A

comprised 18 treatment combinations in III • an incomplete factorial design, which was • replicated in three randomised blocks (A, B and C).

• A Early in the experiment, tadpole survival

(Figure 34) was reduced by low pH A P<0.0001), and reduced further by the • addition of aluminium (P<0.006). In non- acid pools, early survival did not differ significantly between calcium treatments, but in the acid pools early survival was 0 lower at the lower level of calcium 0 60 120 180 240 300 360 (P=0.08) - the effect of calcium on Tadpole (nos) survival was greatest in the block where Figure 35. The relationship between number of tadpoles acid pools began at pH 4.1 (A) rather surviving in a pool prior to the earliest metamorphosis (nd and than pH 4.3 (B and C). Calcium has been the proportion of those tadpoles reaching metamorphosis shown to ameliorate sodium loss by acid- (p = In the Figure, 0 = arcsin p. Different symbols stressed tadpoles (Freda & Dunson 1984). distinguish initial tadpole densities: 90 tadpoles pool -1; 180 tadpoles pool-'; 360 tadpoles pool'. The greater the number of tadpoles in a pool at the time when metamorphosis began, the The proportion of tadpoles surviving the smaller the proportion of those tadpoles that survived to become early part of the experiment, prior to the froglets 39 dependent mortality later on, and survival increased as density was increased, even prove beneficial in some respects, to metamorphosis was significantly lower whereas, in non-acid pools, the but further acidification could produce a in acid pools than in non-acid pools. intermediate initial density produced the sudden and rapid decline in recruitment highest biomass of froglets (Figure 36). as early, acid-related mortality became In acid and non-acid pools, the mean size Thus, within the range of conditions in critical. of froglets decreased as initial tadpole this experiment, froglet biomass was density increased. Mean size at limited by density effects at C P Cummins metamorphosis was greater in acid pools circumneutral pH, and by effects of acid than in non-acid ones; furthermore, conditions at low pH. among acid pools, mean size at References metamorphosis was greater when High tadpole densities in non-acid pools Cummins, C.P. 1989. Interaction between the aluminium was added, and also at the also led to delayed metamorphosis. Analyses of the numbers of froglet that effects of pH and density on growth and lower of the two levels of calcium. These development in Rana temporaria tadpoles. effects of pH, aluminium and calcium can had left pools by the time an unscreened, Funct. Ecol., 3, 45-52. all be attributed to reduced tadpole control pool dried up indicated a greater density, resulting from mortality caused impact of desiccation on non-acid pools Freda, J. & Dunson, W.A. 1984. Sodium by the acid conditions. In acid pools, than on acid ones. balance of amphibian larvae exposed to low most of the mortality that occurred took environmental pH. Physiol. Zool., 57, 435-443. place early in the experiment, and the These results demonstrate that simple consequent reduction of density allowed comparisons of the outputs of froglets surviving tadpoles to grow more rapidly from pools in more or less acid areas than tadpoles in equivalent, non-acid may fail to detect acid stress on frog pools. In the non-acid pools, survival was reproduction. A range of acid conditions Spray drift impacts on non- high initially, leading to poorer growth may exist in which acid-related mortality, target organisms and greater mortality in the later stages in the embryonic or early larval stages, is (This work was partly funded by the Nature of development. Consequently, even offset by a consequent relaxation of Conservancy Council and the Department of though survival to metamorphosis in acid crowding effects. Although the relative the Environment) and non-acid pools was similar under importance of these factors is likely to some treatments, the greater mean size vary, according to resource availability Agricultural sprays are designed to kill of tadpoles in acid pools meant that and predation pressure, for example, the pests, weeds and diseases within the froglet biomass was greater there than in absolute effect of acid conditions on early crops where they occur. Some non-acid pools (Figure 36). survival ultimately limits the scope for compounds are effective against a wide successful reproduction. Acidification range of species, and are therefore likely Overall, increasing initial tadpole density within the range of conditions in which to affect non-target plants or led to an increase in the biomass of compensation can occur may have a invertebrates, including those of froglets produced. However, there was a much smaller impact on recruitment to conservation value which occur in field significant interaction between density frog populations than would be predicted margins and adjoining habitats. The and low pH. In acid pools, biomass from simple laboratory studies - it may hazard to such species depends on their intrinsic susceptibility to a compound, and to the extent of spray drift downwind of a sprayed area. The possible effects of spray drift were reviewed by Williams et al. (1987), but data are needed in order 30 to assess buffer zones for protecting El 1 mg Ca rl sensitive species. Research haS, • 4 mg Ca 1-1 therefore, been done on this topic with respect to higher plants, ferns, lichens, terrestrial and freshwater invertebrates, and initial woik on native plants is described by Marrs et al. (1989). This report sumrnarises the work done on insecticides and terrestrial insects.

There are three main aspects:

1. choosing species whose phenology and behaviour may render them vulnerable to insecticides;

2. measuring the relative toxicity of different compounds to these selected 90 180 360 90 180 360 90 180 360 species; Non-acid pools Acid pools Acid pools +Al Figure 36. Total live mass (biomass) of newly metamorphosed 3. measuring drift deposition and froglets. Each bar shows the mean for three pools. Values below mortality profiles downwind of sprayed bars show initial numbers of tadpoles pool-1. Biomass was limited areas in relation to site factors and by crowding effects in non-acid pools, whereas initial tadpole density limited biomass in acid conditions. Overall, in spite of the operating conditions. early mortality caused by acid conditions, acid pools were capable of producing a froglet biomass at least as great as that Butterfly larvae were chosen as the test from non-acid pools organisms because of their conservation for diagram 40 interest, and because there is good information available on their occurrence in field margins (Dover 1986) and on their general ecology (eg Thomas 1986 Pollard, Hall & Bibby 1986). For experimental purposes, the large white butterfly (Pieris brassicae) (Plate 12) was found to be particularly suitable as it can be bred continuously and its larvae are produced in large numbers. Three other, non-pest, species were used for comparison, namely green-veined white (Pieris napi), common blue (Polyommatus icarus) and hedge brown (Pyronia tithonus), all of which occur in field margins or adjacent habitats.

Laboratory experiments were carried out initially on two-day-old large white larvae (Plate 12 inset) with nine insecticides to determine LD50 from topical applications. Fewer insecticides were similarly tested on the other three species but, because of their smaller size at hatching, the tests were made when the larvae had reached Plate 13. Spraying carrots with tnazophos a comparable size.

insects were taken back to the This approach was also used to assess laboratory, and mortality was assessed the effects of drift from actual crop over three to four days. Experimental and spraying. Two bioassays were done for logistic models were fitted to the data to triazophos applied to carrots for carrot fly estimate 11950 distances' and 1D10 control (525 g ai ltia-') at the Arthur 4, distances', ie the distances at which 50% Rickwood Experimental Husbandry Farm and 10% mortality would occur under the at Mepal, Cambridgeshire (Plate 14). One prevailing conditions. LD10 distances are bioassay was done for cypermethrin (25 more appropriate for determining buffer g ai ha-1) plus demeton-S-methyl (325 g ai zones but can be estimated less ha i) applied to brussels sprouts at Upper accurately than LD50 distances. Caldecote, Bedfordshire.

Plate 12. Large white butterfly, with (inset) larvae killed by spray drift

Field bioassays were then done under experimental conditions at ITE Monks Wood. Commercial insecticide formulations were sprayed on an open area of short grass, using standard hydraulic equipment and maximum recommended application rates (boom length 6 m, height above ground 80 cm, pressure 2 bar, speed 6 km If ', discharging spray at 200 litres ha 1). Four passes were made covering an area of 70 m x 24 m to simulate the cumulative effect of drift from the outer zone of a sprayed field. Wind speed and direction were monitored on a portable anemometer. Target insects on cabbage (Brassica spp.) plants or on horse-radish (Armoracia rusticana) leaves were placed at intervals between 0 m and 32 m Plate 14, Experimental array of horse-radish leaves and caterpillars, with anemometer and downwind (Plate 13). After spraying, the collectors for measuring spray drift deposits 0-32 m downwind of a sprayed area

41 Table 9. Topical LD50 doses (ug insect-1) of nine insecticides for young larvae of four butterfly deposition, as determined from the water species sensitivity papers, was clearly related to (OP=organophosphorus, Ca=carbamate, OC=organochlorine, Py=pyrethroid, CH=chitin wind speed (Figure 37). Thus, inhibitor) approximately equal deposits were recorded at 2 m, 4 m, 8 m and 16 m distances for wind speeds of 2.5 m, 3.5 m, 4.0 m and 5.3 m s-1 respectively. At the lowest wind speed of 2.0 m s-1, the LD50 distance for large white was about 1 m and there was negligible kill beyond 4 m. At the highest wind speed of 5.3 m s-1, the drift killed over 95% of larvae up to 16 m, and no LD50 distance could be estimated. Intermediate wind speeds gave intermediate results, with high kills

Water-sensitive papers (Ciba Geigy 5.2 cm x 7.6 cm) were placed on both sides of the target plants at Monks Wood to register drift deposition. These papers gave a quick visual guide to drift, and image analysis provided relative estimates of deposition at different 1.0 distances downwind.

The laboratory tests with the large white Wind speed (m s-1) butterfly showed a 700-fold range in toxicity, from dimethoate (least toxic) to 4, diflubenzuron (most toxic) (Sinha, Lakhani • 4.0 5.3 & Davis 1990). Three-day-old green- F iNo veined white were found to be two to three times less sensitive, and a field trial 8 0.1 with large white and green-veined white cr) together confirmed this finding. The other two butterfly species were likewise less sensitive weight for weight (Table 9). •

In an initial series of field trials with four compounds (Sinha et a). 1990), dfflubenzuron was again the most toxic. At a rather high wind speed of 4.25 m s-1, .01 it gave an LD50 distance of 6.7 m for 2.5 2.0 I large white. In all other cases, the LD50 • distance was <1 m (Table 10). Likewise, • the estimate for LDIO distance for diflubenzuron was almost 16 m, while for the other compounds it was <2 m. 12 4 6 8 12 16 Downwind distance (m) A second series of field trials compared Figure 37 Spray drift deposition on water-sensitive papers at different drift effects from diflubenzuron at five wind speeds. Five trials with diflubenzuron. Means and ranges from paired different wind speeds. The drift receptors

Table 10. Estimated downwind distances (dso and d10 from exponential models) at which 50% for the first few metres and then a and 10% of large white butterfly larvae are killed by spray drift from four insecticides under decline (Table 11, Figure 38). similar experimental conditions. (Mortality assessed after one day for fenvalerate and after three days in other cases) This effect was also seen in the field bioassay for cypermethrin plus demeton- S-methyl. The estimated LD50 distance here was about 5 m for a wind speed of 2.5 m s-/, while estimates for LD10 distance were between 14 m and 20 m, depending on the model used. The two field bioassays for triazophos gave

42 Table 11. Estimated downwind distances distance for these three sprays, in the Dr S N Sinha, from the Indian Agricultural (from logistic and exponential models) at absence of hedges, etc, were 27 ni, 11 m Institute, made a valuable contribution to which 50% and 10% of large white larvae are and 20 m respectively. Buffer zones killed by spray drift from diflubenzuron at the work in 1988. He was supported by a different wind speeds under similar which are adequate for large white are British Council fellowship. experimental conditions likely to be adequate for green-veined white, common blue and hedge brown B N K Davis, T J Yates and K H Lakhani dso larvae under similar conditions of Wind speed Logistic/ Logistic/ exposure. (m s-1) exponential exponential References The most extreme case of drift hazard 4.0 5.5/5.4 5.5/15.7 was recorded when the mean wind Dover, J. 1986. The benefits of unsprayed 5.3 >16 No estimate speed was 5.3 m sl (at the upper limit of 2.0 1.1/1.0 2.613.2 headlands: butterflies on farmland. Annu. Rev. 3.5 7.617.7 26.9/25.3 force 3 on the Beaufort scale). Game Conserv. 1985, 65-68. 2.5 8.0/9.3 17.3/28.0 Diflubenzuron spray drift caused 97% kill at 16 m under these conditions, which Marrs, R. H., Williams, C. T., Frost, A. J. & should clearly be avoided. Plant, H. A. 1989. Assessment of the effects of herbicide spray drift on a range of plant estimates for LD50 distance of 2.4 m and The study shows that some of the most species of conservation interest. Environ. 4.1 m at similar low wind speeds. widely used insecticides, such as Pollut, 59, 71-86. pirimicarb, phosalone and dimethoate, On the basis of contact toxicity and pose very little risk to butterflies. Pollard, E., Hall, M. L. & Bibby, T. J. 1986. maximum approved rate of application, Diflubenzuron is extremely toxic but is Monitoring the abundance of butterflies pirimiphos-methyl and triazophos would little used in Britain. The greatest field 1976-1986. (Research and Survey in Nature appear to be the most hazardous to large Conservation no.2.) Peterborough: Nature hazards to butterflies may, therefore, be Conservancy Council, white larvae in the field, followed by from triazophos, cypermethrin, fenitrothion and then diflubenzuron. fenitrothion and pirimiphos-methyl, which Sinha, S.N., Lalchani, K.H. & Davis, B.N.K. However, the last three of these are all used quite widely. Buffer zones of 1990. Studies on the toxicity of insecticidal drift compounds also act as stomach poisons, the order of 20 m may be needed for to the first instar larvae of the large white while pirimiphos-methyl has strong triazophos and cypermethrin under butterfly (Pieris brassicae) (Lepidoptera: vapour action, so their overall relative conditions of moderate wind. Pieridae). Ann. appl. Biol., 116, 27-41. toxicity in the field may be different. Of the four compounds tested More field-scale tests are required on Thomas, J. A. 1986. RSNC guide to butterflies experimentally - phosalone, fenitrothion, these compounds. Further work is also of the British Isles. Rushden: Hamlyn. fenvalerate and diflubenzuron - only needed on the effects of hedges in diflubenzuron produced significant deflecting or trapping spray drift Williams, C. T., Davis, B. N. K., Marrs, R. H. mortality of larvae more than 2 m from deposits, and on the extent to which & Osborn, D. 1987. Impact of pesticide drift. the sprayer. spray drift may penetrate woodland (NERC contract report to the Nature margins. Most important, however, is the Conservancy Council.) Abbots Ripton: Institute of Terrestrial Ecology. The results for diflubenzuron, at wind need to measure the effects of spray drift speeds of 2.0-4.0 m s, were comparable from aerial spraying, as desk study with those for triazophos (at half the full estimates suggest that buffer zones recommended application rate) and for should be considerably greater for some cypermethrin at wind speeds of 1.5-2.7 m insecticides to afford protection to Climate change - the mild s. The highest logistic estimates of LIDio sensitive species. winter of 1988-89 (This work was funded by the Department of the Environment)

100 -•-•—• 5.3 • • The exceptionally mild winter of 1988-89 provided an opportunity to observe the responses of the UK flora and fauna to 80 the type of warm conditions that could occur in the 21st century as a result of an enhanced greenhouse effect. This co- ordinated project collected observations • • on temperatures (University of East to. Anglia), hydrology (Institute of Hydrology), ecological impacts (ITE and the Unit of Comparative Plant Ecology Wind speed (m s-1) (UCPE), University of Sheffield), • freshwater impacts (Institute of 20 2.5 Freshwater Ecology), and impacts on agriculture and horticulture (University of \. • Nottingham). Over 100 people were 2.0 • 0 • consulted. 012 4 8 16 24 Downwind distance (m) The mean temperature in central England in the period December 1988 to Figure 38. Mortality of large white butterfly larvae exposed to March 1989 was 2.6°C above the 1961-80 diflubenzuron spray drift at three different wind speeds (see Table I I) average, and was the second mildest

43 -4 4 -12 -8 -4 4 -12 -8

61 -

— I 49 49 49 49 -8 0 4 -12 -4 0 4 -12 Longitude Longitude Figure 39 Difference between the mean temperatures in December 1988-February 1989 in the UK and the long-term mean (1961-80) for December-February DC) The greatest temperature anomaly is shown over the Firth of Forth (dara by courtesy of P D Jones, Climatic Research Unit. Umversity of East Anglia) winter recorded since 1659. In December-February there were 29 10000 fewer days below 5°C than the 1961-80 mean, including 23 consecutive days above 5°C during the Christmas-New 9000 Year period. The centre of the greatest Above 200 anomaly (warmest compared with 175 - 200 1961-80) was over north-eastern England 150 - 175 and eastern Scotland (Figure 39). The 8000 125 - 150 winter was characterised by westerly air 100 125 flows (18 more westerly days than 75 - 100 normal) 7000 50 - 75 Below 50 Mild winters in Britain tend to be relatively wet. 1988-89 was especially 6000 unusual in being dry. Over England and Wales, the period from November to January was the driest since 1879, and 5000 followed a relatively dry autumn. By contrast, the north-west of Scotland recorded more ram than usual (Figure 4000 40). with incessant driving rain causing heavy red deer (Cervus elaphus) mortalities. By mid-February, there was concern over the water resources in 3000 England, but groundwater levels increased following heavy rainfall in March and April 2000

Observations of up to 17 native plant species at the six 1TE Stations, and of 1000 over 200 plants in East Lothian (by Professor F T Last), showed that most species that normally flower before April 0 flowered several days or weeks earlier 200 1200 2200 3200 4200 5200 6200 than normal in 1989 (Figure 41). Easting Figure 40. Precipitation in December 1988-February 1989 as a percentage However, some plant species flowered or of the 1961-80 average (data by courtesy of N W Arnell, Institute of came into leaf at the same time as Hydrology)

44 120 vernal species which have large DNA amounts

Conifer seedlings in forest nurseries did not 'harden off' as expected About 2-3 million transplants died in cold storage, 100 and a similar number died after outplanting.

Many invertebrate species were active during the winter. Black and green 80 aphids built up large populations, and aPhicides were quickly sold out to farmers. There were also major 1989 outbreaks of the green spruce aphid (Elatobium abietinum) on spruce (Picea 60 spp.). Moths were unusually active in March-April, and many hibernating butterfly species were sighted during the winter. Species such as the red admiral (Vanessa atalanta) and painted lady 40 (Vanessa cardw) successfully overwintered in southern England Slug populations were higher than for about Mean for ten years, partly as a result of the wet 1978-88 summer and autumn of 1988 and the continued growth of vegetation and crops during the mild winter..

20 11 The Royal Society for Nature Month Conservation recorded common frogs .JF M A MJJ AS (Rana temporaria) spawning several 0 weeks or months earlier than in previous 0 5 10 15 20 25 30 35 40. years (Figure 42). Reptiles were also Week active during the winter, but some Figure 41. Number of species flowering during 1989 in a garden in East suffered from fungal disease or from cold Lothian compared with the mean number in 1978-88. The vertical bars denote weather in March-April 1989. two standard deviations (data by courtesy of Professor F T Last) Bird movements and behaviour were unusual. Oystercatchers (Haematopus ostralegus) returned to their breeding normal, presumably because their spring flowered at the same time as in previous grounds early; and many bird species phenology is controlled more by years. exhibited full breeding behaviour in mid- daylength and winter chilling than by winter. spring temperatures. For example, Observations on plant growth during blackthorn (Prunus spinosa) flowering March in Derbyshire supported the Freshwaters were about 2°C warmer than and horse chestnut (Aesculus theoretical expectation of UCPE that average, and no ice cover was recorded hiPpocastanum) leafing were very early, species with small amounts of DNA per on most lakes. However, the mild winter whereas honesty (Lunaria annua) and cell grew mUch earlier than normal, and had no dramatic effect on freshwater most apple (Malus spp.) varieties hence competed more vigorously with biology. Model simulations suggested that

, ODecember MJanuary

February

N March

II April

1984 1985 1989 Figure 42. Phenological records of first date of spawning of the common frog in 1984, 1985 and 1989 (data by courtesy of the Royal Society for Nature Conservation)

45 the warm temperatures in 1988-89 would Department of the Environment has have had little effect on the seasonal commissioned a further set of dynamics of Daphnia, or on the long-term observations on the impact of the hot survival of brown trout (Salmo trutta). summer of 1989 and the mild, wet winter of 1989-90. The mild winter had large effects on agriculture, partly by affecting crop M G R Cannell growth and development, and partly by affecting the spread and survival of diseases, pests and weeds.

Pigeon (Callumba spp.) numbers were high, but pigeon damage was generally less than normal on crops like oilseed rape because alternative food supplies were plentiful. In contrast, rooks (Corvus frugilegus) damaged many autumn-sown crops as they foraged for leather-jackets (Tipulidae) that were usually active.. Rat (Rattus spp.) and mouse (Apodemus spp., Micromys minutus) populations were higher than normal, and Rentokil reported record sales; however, the mild winter may have been only one of several contributory factors. Many weeds survived the winter, as did volunteer cereals and potatoes which could carry pathogen inoculum.

Winter wheat crops were 2-3 weeks more advanced in the early part of the season, but they flowered only 7-10 days early and yields were near normal. However, some cereal crops were affected early and severely with mildew, yellow rust and barley yellow dwarf virus. Unusually large quantities of fungicides (as well as aphicides and molluscides - for slugs) were used.

Grass continued to grow throughout the winter, and sheep, which cause less • trampling damage, were kept out late and turned out early. Some stock keepers were deceived by the lushness of the grass, and failure to feed supplements led to pregnancy toxaemia, milk fever, and hypomagnesaemia.

In general, early vegetables and flowers were oversupplied and prices were low. Disease and early regrowth caused problems in stored onions.

Overall, the mild winter revealed many aspects of the natural and managed environment that respond to unseasonally warm winter conditions, and those aspects that fail to respond. It also revealed a complex pattern of 'winners' and 'losers': some organisms benefitted from growth, development or activity in the winter, others did not. The complex set of observations posed many questions that need to be addressed in a more scientific manner and to be confirmed by further observations. To that end, a programme of research on the impacts of climatic warming on the UK flora and fauna has begun in ITE, and the

46 Population ecology

Most of ITE's research on population much wider programme aimed at not ovulate spontaneously. The surge of ecology is concerned with understanding understanding the various pressures on luteinizing hormone (LH), which induces the factors that influence the abundance plant and animal life in present-day Britain. ovulation, only occurs in response to, and and distribution of wild plants and immediately following, copulation. Thus, animals. Such an understanding is the reproductive condition of female essential for the effective conservation of The effect of a plant rabbits cannot be determined simply by wildlife, and is helpful in the design of measuring plasma levels of reproduCtion control programmes for pest species. It compound on reproduction in hormones, as in other female mammals. also enables us to predict more rabbits However, an assessment of reproductive accurately the effects of land use condition can be obtained by giving a changes on particular species. Research begun in the 1960s in the USA single injection of gonadotrophin- on the montane vole (Microtus montanus) releasing hormone (GnRH), which causes The six projects described below are all first suggested that a plant compound a surge of LH in exactly the same way as concerned with species of importance to may affect mammalian reproduction. This copulation. LH levels reach a peak 15 man, either as pests or as conservation vole inhabits upland grasslands in minutes after injection, and this surge priorities. Thus, the rabbit is a western North America. Breeding begins causes the ovaries to produce the steroid widespread pest of farmland-whose each year within days of the first flush of hormone progesterone, which itself reproductive rate is legendary. The work grass growth, yet the time that this flush reaches a peak about 60 minutes after described forms part of a larger project occurs can vary by more than one month injection. designed to understand the various between years. Consequently, the voles factors which influence rabbit breeding must use some cue other than, or in Most of the work on photoperiodic rates. The work on squirrels is addition to, the change in daylength to control of seasonal breeding in mammals, concerned with factors affecting time their breeding. It has been and the involvement of melatonin, has numbers, and examines why the suggested that some factor in the new involved species which show introduced grey squirrel has so plant growth itself may trigger breeding, spontaneous oestrus cycles. The successfully replaced the native red and, eventually, a compound was isolated objectives of this study were (i) to squirrel in most of Britain. Another which seemed to act in this way (Sanders investigate the, seasonal cycle of popular mammal which has declined in et al. 1981). Field studies using food response to GnRH in a reflex ovulator, numbers is the otter. However, a good treated with the compound appeared to the rabbit; (ii) to determine if melatonin population still remains in Shetland and in confirm the hypothesis (Berger et al. modulates this cycle; and (iii) to see if other parts of Scotland, where the effects 1981), although the experimental design the plant compound 6-MBOA affects the of habitat, food supply and pollutants are was open to criticism. cycle by mimicking the action of being studied to provide information of melatonin. Compounds of similar use in conservation. The compound isolated was 6-methoxy-2- molecular structure to natural hormones benzoxazolanine (6-MBOA). This is an often act either in the same way as the The article on butterflies examines the indoleamine, and, as such, has a similar natural hormone (agonists) or by impact of road construction on structure to melatonin, which is a inhibiting the natural hormone populations. It shows that several species hormone in vertebrates. Melatonin is (antagonists). Whether a compound acts can thrive on the flowers present on important in the photoperiodic responses as an agonist or antagonist depends on verges, and that roads are less of a of mammals, and is produced by the the exact molecular structure and its barrier to dispersal than was once pineal gland which lies just beneath the effect on metabolism and binding to feared. It also suggests ways in which top of the skull. Melatonin synthesis is natural receptors. road verges could be adapted at the inhibited by daylight, so large amounts construction stage to support a wider are only produced during the night. Experiment 1 range of butterfly species. Consequently, there is a daily cycle of melatonin concentrations in blood, with Female European rabbits (Oryctolagus The last two papers discuss some effects high levels at night and low levels during cuniculus) were kept individually in large of land use changes on certain bird the day. In addition, there is a seasonal cages, with free access to food and species, showing how the lapwing has cycle, with high levels for a longer water. The cages were kept outdoors, declined under modern agriculture, while period of each day during winter, and for protected from rain and wind, but the oystercatcher has increased. The a shorter period during summer. It has exposed to natural daylight. Each rabbit effects of modern forestry on the been demonstrated in a number of was given a monthly injection of GnRH, capercaillie are also examined, as are mammals that this annual cycle modulates and blood samples were collected in the the effects of changed grazing pressure the timing of seasonal breeding. hour following the injection. on certain moorland species. Concentrations of LH and progesterone Unlike most mammals, rabbits do not were measured in the plasma from these These projects form a small part of a have oestrus cycles, and consequently do samples by radioimmunoasSay.

47 At the beginning of the experiment in 60 spring, the GnRH injections caused a a) 30 0 0.6 large peak of LH at 15 min and of 3 -o progesterone at 60 min (Figure 43). Both R. responses declined rapidly immediately 40733 0. cc: 0 after the summer solstice. The LH 20 Ow 0.4 response gradually increased during CD autumn, but that of progesterone did not a increase again until the following spring. CD The decrease in response after the 10 3 0.2 (g: solstice is due to the decrease in ca daylength (food availability was constant), and may be mediated by melatonin. It is perhaps paradoxical that a decrease in 0 0 0 1AM987 JJ A SONDJ F M A M daylength inhibits the response, and yet 1988 the response resumes under the shorter days of autumn. The short days appear to Figure 43. An injection of gonadotrophin-releasing hormone (GnRH) causes switch on the reproductive system in release of luteinizing hormone (L1-1)from the pituitary gland. Maximum concentration of LH in blood occurs about 15 min after injection. This LH surge rabbits and other species, in terms of causes release of progesterone from the ovaries, which reaches maximum endogenous GnRH synthesis in the concentration in the blood about 60 min after injection. This Figure shows hypothalmus, but subsequent exposure to seasonal changes in the response of female rabbits to an injection of GnRH long days seems to render them (n=8). The open bars show plasma concentrations of LH 15 min after injection, responsive to a decrease in daylength and the solid bars progesterone concentrations 60 min after injection. The line causing a transient switch off. The represents change in ovarian weight in free-living rabbits (from Boyd & Myhill 1987) delayed return of the progesterone response is because the ovaries do not mature until spring, and so are unable to respond to the LH surge. 30 LM 60 Experiment 2 20 Female rabbits were kept individually in 40 cages indoors and exposed to artificial 0 long days. Each was given an injection of 9 GnRH, and blood samples were taken. 20 -o 5 The animals were then assigned to one DC of four groups. The first group was kept '9) DC on long days; the second group was kept o -a on long days and treated with melatonin; ca the third was kept on long days and L6 cnco treated with 6-MBOA; and the last group 60 was transferred to short days. Two 8 futther injections of GnRH were given, the first two weeks after the beginning of 401 treatment and the second four weeks later. Again, blood saniples were taken 10 and plasma levels of LH and . 20 progesterone measured. 0 The rabbits kept on long days showed no 1 2 3 1 2 3 0 significant change in response to GnRH, GnRH injection nos neither did those treated with 6-MBOA Figure 44. Responses of female rabbits to injections of GnRH. All rabbits were (Figure 44). Those treated with melatonin ininelly kept on long days and given an injection of GnRH Two weeks later, the and those transferred to short days both different treatments began. Group L (n =5) was kept on long days, group L6 (n=6) showed decreases in the response of LH was kept on long days and treated with 6-MBOA, group LM (n=4) was kept on long and of progesterone. So, the effect of days and treated with melatonin, and group S (n=5) was transferred to short days. short days is mediated by melatonin, but Another injection of GnRH was given two weeks after treatment began, and a further 6-MBOA does not mimic this effect. injection was given four weeks later The open bars show plasma LH concentrations 15 min after injection, and the solid bars progesterone concentrations after 60 min These two experiments are examples of a series done on female rabbits. Several experiments also involved male rabbits. The conclusions were that rabbits are and so was not a melatonin antagonist. References highly photoperiodic and that melatonin The original work suggesting possible mediates the effects of daylength. effects of 6-MBOA is open to criticism, • Berger, P.J., Negns, N.C., Sanders, E.H. & and, since our study began, other results Gardner, P.D. 1981. Chemical triggering of However, 6-MBOA had no effect reproduction in Microtus montanus. Science, whatever: it did not mimic short day have been published which were either NY, 14, 69-70. effects in animals on long days, and so inconclUsive or contradictory. was not a melatonin agonist, neither did it Boyd, LL. & Myhill, D.G. 1987. Seasonal mimic long days in animals on short days, A S Dawson and C J Bray changes in condition, reproduction and 48 fecundity in the wild European rabbits seeds in deciduous woodland, their (Oryctolagus cuniculus). J ZoolPond., 212, 4- native habitat in North America. Red 223-234: squirrels, in contrast, did only 33% of Sanders, E.H., Gardner, their foraging on the ground. Arboreal P.D., Berger, PJ. 3- Negus, N.C. 1981. 6-Methoxybenzoxazolanine: foraging is appropriate for much of the a plant derivative that stimulates reproduction red squirrel Eurasian range, where seeds in Microtus montanus. Science, NY , 2114, 67-69. are available in conifer canopy throughout the year, but did this feeding behaviour prevent them from competing with grey squirrels in deciduous woods?

From 1983, another CASE studentship (This work was partly supported by an NERC enabled J L Holm to work on red CASE studentship with Royal Holloway and 0 squirrels in three oak/hazel woods on the Bedford New College, London) 1 2 34 56 Isle of Wight, for comparison with grey Pre-breeding density (nos ha--1) squirrels in three similar woods on the North American grey squirrels (Sciurus Figure 45. The range of densities, mainland. Squirrel numbers and carolinensis) may have been released in and of numbers of young produced breeding success were monitored by Britain as early as the 1830s, but only per adult female, for red squirrels in trapping in each wood (Plate 15), and became permanently established Scots pine on Furzey Island ( ), foraging was assessed by recording 30- following an introduction to Woburn in compared with grey squirrels in fix ranges about four times a year. The 1890. After 30 years, the species was oak/hazel woods (----). Mean values abundance of the main tree seeds was established in 20 different areas, but its are where the lines cross estimated each winter by counting acorn major range expansion occurred during cups and hazel shells in 50 cm x 50 cm the 1940s and 1950s (Lloyd 1983). From quadrats under oak and hazel canopy. the late 1940s onwards, there was a Nevertheless, there are marked corresponding rapid contraction in the differences between red and grey The number of hazel bushes was counted British range of the native red squirrel squirrels in deciduous woodland, both in too, in four grids of 25 m x 25 m, spaced (Sciurus vulgaris). By 1989, red squirrels foraging behaviour and demography. The equally through each 100 m grid square. were abundant only south of Yorkshire, in first evidence came from comparing Oaks were either counted in the centre some Welsh conifer forests, and on the radio-tagged grey squirrels in Monks of each square, in a 50 m x 50 m Isle of Wight, with small populations in Wood with red squirrels studied at quadrat, or mapped throughout the whole East Anglia and on three islands in Poole Merlewood by M Tonkin, a CASE student wood. Harbour. with the University of Bradford. Monks Wood is predominantly oak (Quercus Grey squirrels ate many hazel nuts Why have the red squirrels been spp.), ash (Fraxinus excelsior), field before they were ripe, and had generally replaced? It is important to answer this maple (Acer campestre), and hazel consumed this crop by October, after question, because, if the mechanism can (Corylus avellana), fairly similar to the which they fed mainly on acorns (Plate be understood, then steps might be oak, ash, yew (Taxus baccata) and hazel 16). Red squirrels ate some acorns, but possible to halt or even to reverse the woodland near Merlewood. Grey squirrel often left them part-eaten. They fed process. An important clue is that red foraging was 86% on the ground, which is mainly on hazel nuts through the winter squirrels can persist in conifers despite appropriate for recovering fallen tree or, when these were scarce, foraged in the presence of greys, and even recolonise such areas, but seem never to recolonise deciduous or mixed woodland from the edge of their range. This fact suggests that their disappearance is not simply a matter of disease, or of aggression by grey squirrels (Gurnell 1987). On the other hand, a mathematical model incorporating a coefficient of competition fitted quite well to the observed pattern of red squirrel replacement (Okuba et al. 1989).

Squirrels are dependent for food on tree seed crops during much of the year. MacKinnon (1978) noted that red squirrel densities tended to be lower than grey squirrels, and suggested that red squirrels were conservative breeders in order to reduce the scramble for food when seeds were few. However, when British Petroleum commissioned ITE to survey red squirrels on Furzey Island, where Scots pine (Pinus sylvestris) trees were producing heavy crops beside the oil rigs, red squirrels could breed as well as grey squirrels at similar densities in deciduous woodland (Figure 45). Plate 15. Trapped red squirrel in oak/hazel woodland

49 Table 12. Squirrel pre-breeding density, breeding success (with standard errors for both) in oak/hazel woods, plus mean tree seed availability. Correlations of squirrel parameters to tree seed availabilities are also given

Squirrels ha T Young per adult Tree seeds female ha No. of Correlation Correlation site years Mean Acorn Hazel Mean Acorn Hazel Acorns Hazel nuts

Grey squirrel 15 2.31 P<0.02 NS 0.69 P<0.002 NS 37000 9000 (0.40) (0.28) (17000) (2000)

Red squirrel 9 0.86 NS P<0.002 0.37 NS P<0 001 76000 17000 (0.09) (0.17) (23000) (11000) Plate 16. An excellent .i.t:orn crop: not such good food for red squirrels oaks for the fungus Vuilleminia, which and productivity in conifers on Furzey develop 'seed-less' oaks to annul the grows under dead bark. The foraging Island (Figure 45). grey squirrel advantage, which could preferences were quantified by a also reduce bark-stripping? Alternatively, computer program, which used tree Acorns contain polyphenols (tannins), are there polyphenol-tolerant red density in each 100 m square to estimate which inhibit digestion and can kill squirrels elsewhere in Eurasia which the mean tree density at each radio- horses. This ingredient is probably a could be introduced to Britain? Could location, for comparison with mean tree defence against seed-eating animals. such squirrels be produced by gene density within each polygon range outline transfer from grey squirrels? There is a (Figure 46). Where there were significant However, pigs thrive on acorns, and it pressing need to compare red and grey seemed possible that grey squirrels had squirrels in conifers. If red squirrels have evolved a similar adaptation to their an advantage there, reintroductions might native deciduous habitat. Six young be practical in new southern conifer squirrels of each species were, therefore, forests, at least where there has been no tested for a month in captivity on an public pressure to plant oak fringes. acorn diet, three at a time, while the others remained on a mixed diet of hazel R E Kenward, T Parish and J L Holm nuts, peanuts, sunflower seeds, apple and carrot. Unlike grey squirrels, red References 0 squirrels on the mixed diet did not eat oa acorns, but both species gained weight, Gurnell, J. 1987. The natural history of • as did grey squirrels on the acorn diet. squirrels. London: Christopher Helm. In comparison, red squirrels eating Lloyd, A.G. 1983. Past and present acorns alone showed significant weight 100 m distributions of red and grey squirrels. losses and reduced survival. Analysis of Mammal Rev, 13, 69-80. faeces showed that the longest-surviving red squirrel was digesting acorns only MacKinnon, K. 1978. Competition between 78% as efficiently as the grey squirrels. between red squirrels Sciurus vulgaris and a grey Figure 46. Structure analysis of The most dramatic difference, however, grey squirrels Sciurus carolmensis Mammal squirrel range in Monks Wood, showing was that red squirrel faeces contained Rev., 8, 185-190. the radio-locations (+) and convex polygon range outline of a grey squirrel 87% of the polyphenols in the original Okubo, A., Maini, P.R., Wi11iaisoai Wit.E. overlain on the 100 m x 100 m grid acorns, whereas faeces from the grey Murray, J.D. 1989. A diffusion model for squares used to estimate tree and seed squirrels had only 37%. British squirrels. Proc. R. Soc B, 238,113- 125. densities. 0 marks the main drey The findings confirm the hypothesis that red squirrels are the victim of a plant preferences, these were always for red defence mechanism, which grey squirrels squirrels to forage in hazel-rich areas, but can overcome. Red squirrels in deciduous for grey squirrels to forage where oak woods appear to depend on hazel, which, density was relatively high. like conifers, was an early coloniser of northern Eurasia as the ice sheets (This work was partly funded by the World The foraging preferences were reflected retreated, and which was being lost Wide Fund for Nature) in squirrel demography. The pre- through coppicmg just when red squirrels breeding density and spring breeding were disappearing most rapidly from The numbers of otters (Lutra lutra) have success were significantly related to the British woodland. In winter, after the hazel declined dramatically over the last few density of hazel nuts but not acorns for crop has been eaten, red squirrels would decades; the animal has now red squirrels, and to acorns but not hazel have less suitable food than the invading disappeared from several countries in nuts for grey squirrels (Table 12). grey squirrels, because most British Europe, and it is rare in many others Although there were, on average, more deciduous woodland is still rich in oaks. where it was previously common. Several than twice as many acorns for red reasons are put forward for the decline, squirrels as for grey squirrels, red This study suggests a bleak future for red but evidence is generally lacking. squirrels did not exploit this food, and squirrels in British deciduous and mixed Environmental pollution is considered to had only 37% of the density and 54% of woodland, but many questions remain be important, as well as habitat change, the productivity of the grey squirrels unanswered. Can oak varieties be found disturbance, and mortality caused (Table 12) - far less than their densities with less toxic acorns, or could one directly by man. As otters are at the top

50 of a food chain, they can be important bottom-living species of fish, especially in mid-summer. hitters were small (mean indicators of environmental change. eelpout (Zoarces vivipmus) and five- 1.8), and the numbers of cubs reared in Furthermore, the otter is considered to bearded rockling (Ciliate mustela). These the study area over the winters be an endangered species, and it attracts fishes were active at night, and easiest correlated closely with fish densities the much public interest and concern. To for otters to catch during the inactive previous summer (Figure 47), ie at the understand the possible mechanisms for phase of the fishes in daytime, which time of year when the mothers suckle the change in its fortunes, ITE is studying probably explains the diurnal activity of their cubs. This correlation suggested factors limiting otter populations, their otters in Shetland. Fish biomass was that fish populations, and year-to-year interactions with the fish on which they assessed with small funnel traps, over a changes therein, had an important effect prey, and the role of habitat. These three-year period; it was highest in mid- on the recruitment of young otters. factors were first studied in a relatively summer, and lowest in early spring. The healthy population in Shetland, and main food species of otters were found in research is currently being carried out in different habitats; for instance, eelpout Mortality an agricultural area in mainland Scotland. was found mostly along sheltered coasts Adult mortality was studied using corpses and in the summer, rockling mostly in collected from all over Shetland. Most Numbers in Shetland exposed areas and more often in winter. otters died naturally, although traffic also As expected, this pattern of food took a substantial toll. Most natural deaths Previously, there have been no availability affected the extent and occurred in spring (Figure 48), during successful attempts at estimating numbers success of foraging by otters in different of otters because of the problems of parts of their ranges. 40 observing the animals. In Shetland, otters live along the coast, are normally active o Natural in daytime, and individuals can be Social organisation and • Violent recognised. One study area contained reproduction several otter ranges, and a good The spatial organisation of the otter was correlation was found between the complicated, a demonstration of the number of otter dens or 'holts', and the difficulty in distinguishing between social number of resident adults. On average, and solitary species of carnivore. There there were 3.0 holts per resident female, were small female group territories (4 km which made it possible to use the of coast, or larger), in which group numbers of holts along the 1400 km of members were probably related Shetland coast to estimate the total (mothers/daughters, or sisters) (Plate 17), numbers of otters on the islands. The where each had a different core area, coast was classified into six habitats and all had access to the various habitats (strata), which were sampled for holts with different prey at varying times of with different intensities; the best habitat year. Male ranges were larger, and was sampled most intensively, and the overlapped with several female group oof holts were counted in about one-third of ranges. Males also used a different the coast. It was estimated that there habitat from females: they were much were 1185 holts in Shetland (n=13, 95% more likely to be found on the exposed 0 donfidence limits); allowing for observed JF MA MJ JA S ND coasts, whilst females used the sheltered Month sex ratios and numbers of non-resident bays. Within a female group range, otters, this estimate translated into several or all females reproduced, mostly Figure 48. Percentage of otter 700-900 animals. Although the Shetland carcases found in different months, population is probably one of the most separately, for violent deaths (n=53, mostly road kills) and natural deaths important in Europe, it is almost totally (n=49) isolated, fairly small, and therefore 18 vulnerable (for instance to disasters such as an oil spill or epidemic). 16 1987 and immediately after the time when The large majority of these otters lived 14 food was in shortest supply, and when on undrained coasts with peat, where 1986 the animals were underweight. The age they had easy access to freshwater, 12 • of dead otters was determined from which the animals need for washing. A incremental rings in the teeth. Although series of experiments with captive otters the potential lifespan of otters is 15 years showed that regular washing in or more, the mean life expectancy of an freshwater is essential for otters living in adult otter in Shetland was only three seawater, in order to maintain the years. Annual mortality increased linearly thermo-insulation of their fur. Numbers of with age, unlike mortality in most otters were very low on Shetland coasts mammals. This increase was possibly a which had been drained for agriculture, result of the accumulating effects of toxic and the absence of freshwater is a compounds in the body. However, the probable cause. The low numbers of available evidence pointed tb food otters on coasts elsewhere, such as in shortage as the most important ultimate .0 Orkney and the east coast of Scotland, 0 1 cause of death. can be similarly explained. Eelpout, rockling (nos rn-‘) Figure 47. Numbers of otter cubs The livers of dead otters contained traces Prey populations present M winter in a study area in of DDE, HEOD (dieldrin), BHC (lindane), Shetland, in relation to different prey cadmium and lead, but none in The Shetland otters fed mainly on small, densities the preceding summer concentrations which would have affected

51 males appears to be different from that of females in freshwater also, adult males being more likely to inhabit the large rivers, females the small burns as well as lochs. Both sexes are much less likely to use bolts' than the coastal otters; in fact, some litters are born in reed couches, and some otters rarely, if ever, sleep underground.

The radio-location studies have revealed the importance to otters of reed beds, even as far as several kilometres from open water, for resting and breeding as well as for feeding (Plate 18). The animals' use of such habitat has largely gone unnoticed in studies relying on the distribution of faeces ('spraints') as an indicator for habitat selection. Faeces were found to be an unreliable indicator of otter distribution and activity. Reed beds, incidentally, have disappeared in many parts of the country, which could mean the destruction of an important part Plate 17 Female otter with four cubs in Shetland, in winter. The numbers of cubs in an area of the otters' habitat. Small islands in are correlated with fish densities earlier in the year rivers were also shown to be important.

the animals. Some contained enough PCBs to have had a possible negative influence on reproduction, but, of the compounds studied, only mercury was likely to have affected the overall level of mortality. Mercury occurs naturally in the sea, especially in the Shetland area, and otters do not have the detoxification mechanism for methyl mercury, involving selenium, which is found in other marine mammals. Mercury accumulated in otters with age (Figure 49), and in several otters aged five years or more sufficient mercury was found to have had a possibly toxic effect, This observation suggests that, although otter populations in the sea reach densities higher than anywhere else, the species is not optimally adapted to this habitat.

The correlations found in the Shetland studies suggested that, in areas where otters were common, fish populations were likely to have profound effects on recruitment and mortality, whatever the Plate 18. Otter habitat in NE Scotland. Reed beds are an important component of the habitat, exact mechanisms. Toxic compounds may especially for providing shelter play a role secondary to that of food shortage, but the data suggest that a relatively small increase in the level of some compounds, as pollutants, could Scotland show important similarities and cause accumulations affecting overall differences with the otters in the sea in mortality. An important conclusion for the Shetland. As the animals in freshwater The present study demonstrated aspects conservation management of otter habitat are largely nocturnal, they are being of the food limitation mechanism in a was the effect of land drainage: otters studied with the aid of radio-location, population of mammals, with the need freshwater, especially underground, using implanted radio-transmitters (under possibility of prey availability affecting eg in peat, in order to be able to exploit licence from the Nature Conservancy both mortality and recruitment. However, the fish populations in the sea. Council and Home Office). The ranges with the dispersion of otter populations, it are very much larger in freshwater than also indicated that there were overriding in Shetland: 25-40 km for females, and up effects of habitat. Moreover, effects of to 70 km for males. Again, several compounds such as methyl mercury in The populations of otters living in lakes, females may share a range, especially in the environment suggested that there rivers and small streams in north-east lakes. Interestingly, the habitat of the could be a complicated interaction

52 both rasra, serious consideration was given to building a butterfly bridge' across the route.

50 An initial investigation of this problem was made in the summer of 1989, on major trunk and holiday routes in Dorset and Hampshire (Munguira & Thomas 1990). The verges of 12 roads were analysed for 14 weeks, and the density and diversity of butterfly and burnet moth populations were measured along fixed 3 0 transects The results were compared 20 o 0 with the volume of traffic on these roads, 0 o° 8 0 the nature of the adjoining land, and five 0 0 00 0 0 o characteristics of the verges themselves: cbo 10 their topography, width, the abundance co° 0° o 00 of nectar sources, the number of o breeding habitats of different species that cPce ocbo 0 each contained, and whether the verges 0 1 2 3 4 5 6 7 8 9 10 were sheltered by hedgerows. Studies Age (years) were also made to determine whether the species found along roadsides were Figure 49. Concentration of mercury (ppm dry weight) in livers of otters in breeding on and supported by the relation to age verges, or were simply temporary visitors to flowers. Finally, two sites were intensively studied to discover whether between food availability and other wheré numbers of otters are declining. In adults crossed the roads and, if so, how factors affecting otter populations. terms of conservation, the studies show many were killed by vehicles, using which types of habitat need to be standard mark-release-recapture (MRR) These mechanisms, mutatis mutandis, are protected and why; they indicate which techniques on the adults. likely also to affect populations in fresh- fish species may limit numbers, and how water. The food of otters in the various at least some pollutants can affect otter It was immediately clear that the verges study areas of north-east Scotland numbers in their interaction with the prey of some very busy roads contained large consists very largely of salmonid fishes populations. populations of butterflies and burnets. For (Salrno spp.) at all times of year, with example, more than 2770 adults of the eels (Anguilla anguilla) predominant in H Kruuk meadow brown (Maniola jurtina) were the lakes, and frogs (Rana temporaria) for recorded per hectare on the peak day of a short period in spring. The effects of emergence on one dual carriageway - a otters on prey populations, and vice higher density than has been measured versa, are currently being analysed, to Effects of roads on butterfly in any of the studies of this species in determine the role of food as a factor and burnet populations 'natural' habitats. High densities of limiting numbers in this more agricultural marbled white (Melanargia galathea) habitat, and possible interaction with There has been much speculation about (Plate 19), common blue (Polyommatus othe.r predators of salmonids. the effects of roads on butterflies. On the icarus) and Essex skipper (Thymelicus one hand, it is often argued that new lineola) were also recorded on some The role of toxic compounds in the constructions will harm species that roadsides. In addition, some verges environment (especially pollutants) is also breed on neighbouring land, even if the supported a wide range of species. The being studied in freshwater, although route itself does not destroy existing richest example was a cutting for a present data suggest that direct effects breeding sites. The main concerns are bypass excavated six years earlier on otters are not of overriding that exhaust fumes might poison nearby through intensive farmland that had importance, at least in north-east populations, and that sites on either side previously been almost bereft of Scotland. However, it is possible that of a road will become isolated, with butterflies. This site now contained at pollutants affect otter populations in a adults reluctant to cross this supposed least 23 species of butterfly (40% of the much more subtle way than has barrier. On the other hand, developers UK species) and two species of burnet. previously been assumed. Several metals, argue that butterfly habitats are created Most were breeding in self-contained including mercury, may affect thermo- by new road schemes, because the populations that were wholly supported regulation in mammals in the laboratory, verges and cuttings that are generated by the verge, with eggs and larvae found even when present at relatively low are usually won from intensively farmed within 1 m of the busy road. Other levels, and preliminary studies on otters land that supports few, if any, species. verges had an extremely poor fauna. suggest that the demands of thermo- regulation may play a vital role in the Proponents and opponents of road When the reasons for this variation were adjustment of otter populations to energy schemes have been able to use analysed, it became clear that neither the resources. butterflies as a cause célèbre mainly volume of traffic nor the nature of the because few data existed on which to adjacent land had any bearing on the These complicated relationships will be base their cases. For example, there was density or diversity of butterflies and the subject of future studies, and the sufficient concern for butterflies to be a burnets on the verges in between. results of this research in Scotland are contributary factor in decisions to reroute Indeed, both measurements of expected to be relevant also in many the M40 near Oxford, and to delay the Lepidopteran richness were negatively other parts of Europe, especially in areas extension of the M3 at Winchester. In (but not significantly) correlated with the

53 14

0 — 12

E 10

8 0

0 0 4

0 WO 0 0

Plate 19. Marbled white, found breeding in 0 F large numbers on the verges of several major 0 10 20 30 40 45 0 5 10 15 20 25 roads in Dorset Traffic (vehicles min-i) Habitat Figure 50. The number of different species of butterfly and burner moth fbund on 12 southern road verges compared with (i) the volume of traffic on each road. and (i1) the variety of butterfly and burnet breeding habitats that each richness of the adjoining land for verge contained butterflies and with the volume of traffic (Figure 50i). movement of butterflies, and how few of small reduction in movement was usually In contrast, the condition of the verge the adults were killed by vehicles. In recorded. Nevertheless, between 10% had a strong bearing on the butterflies some cases, there was no discernible and 30% of all the marked adults that and burnets that were present the five barrier effect (Figure 51), although a were later recaptured had crossed roads variables measured together explained in the study area. These high figures 82% of the variation in the number of suggest that the existence of a busy route species found on different verges, and between two populations does not add 90% of the variation in the density of significantly to their isolation, although it adults. The variety of species present should be noted that most of the was particularly closely correlated with butterflies studied were common species. the range of different breeding habitats Some rarities are notoriously sedentary, that existed on the 12 verges (Figure and additional research is needed on 50ii), although variation in the density of whether new roads form a significant Lepidoptera was more closely correlated barrier for such species. with verge width. Thus, the wider verges held significantly higher densities of It was possible to record the mortalities adults per unit area of land, but only inflicted by car strike on only one busy supported a large number of species if road. Cars killed between 0.6% and 1.9% the specialised (and differing) breeding of the adults of those species that live in Plate 20. Small blue: at least five colonies of requirements of many species were closed populations, and about 7% of this rarity are supported by the verges of present. species that are migratory or nomadic. major southern roads or motorways. It is a species whose requirements could easily be These figures are insignificant compared It was surprising to find how slight an incorporated into the design of many more to the other natural mortalities impediment these roads were to the road cuttings experienced by adult butterflies.

The overall conclusion of the study is that many of the fears that have been raised over proposals for new roads are groundless, although there are certain qualifications. Most of this research involved fairly common species of

*N

N 10 m Plate 21. Adonis blue: at least two colonies of Figure 51. The movements of marked meadow brown butterflies between this rarity breed on road verges It is a 42 subareas of an access roundabout on either side of a dual carriageway species whose requirements could be The road formed no barrier to movements even though a vehicle passed incorporated into the design of several approximately every 30 seconds proposed road schemes

54 butterfly and burnet. and inadequate data Weals, T., ell, S. & Frost, A. 1981. Creating can reasonably suppose that the size of a exist on the mobility of rarities. It is clear, attractive grasslands using native plant cock's territory affects his reproductive however, that entire populations of local species. Attingham Park: Nature Conservancy success, ie the number of his sons Council. or rare species can be supported by recruited to the breeding population in road verges, so long as their specialised the following year. If the reproductive habitats exist there. We recorded success of the whole population, in any grayling (Hipparchia semele), silver- one year, is averaged, the proportion of studded blue (Plebejus argus) and brown young cocks recruited to the spring argus (Aricia agestis) on the 14 study population in the following year is related sites in Dorset and Hampshire, and to the average territory size of cocks in elsewhere found several large the current spring population. populations of small blue (Cupido In some parts of their range, the numbers minimus) (Plate 20) and two of Adonis of red grouse (Lagopus I. scoticus) This argument is based on a preliminary blue (Lysandra bellargus) (Plate 21) that fluctuate in cycles of fairly regular analysis of data on individual grouse were wholly supported by the verges of periodicity of about six years and of territories and performance on a 1.8 km2 busy roads. varying amplitude. What are the causes study area at Kerloch Moor, of the cyclicity? Moss and Watson (1985) Kincardineshire. The data include the At present, however, it is very rare for considered, in turn, the effects of food, mapped territorial boundaries of roadsides to support scarce species, cover, predation, parasites and disease. individually tabbed cocks and, wherever because most new verges and cuttings They concluded that each of these factors possible, the relative positions of fathers are uniformly graded, given a thick layer can be ruled out as proximate causes of and sons (Figure 52). of (expensive) topsoil, and then often the cyclic pattern in numbers. It is now seeded with alien grasses or cultivars - proposed that the pattern could be As a rule, males remained on the same practices which provide the explained by the joint action of two site. Those old cocks that survived and requirements only of common species of factors: territorial competition and were still territorial the following sprind butterfly and burnet. It is now clear that a differential behaviour towards kin and (40-60% of the autumn number) usually much wider range of butterfly habitats non-kin. The evidence leads to the retained a part of, or augmented, the could be created, including those of assumption that the cycle in red grouse is territory which they had held the national rarities, if the traditional smooth driven by the male sex. Each autumn, old previous year. Out of 213 records of the topography of roadsides were altered to cocks and young cocks hatched the central co-ordinates of the territories of provide sheltered pockets containing previous spring join in the contest for cocks surviving to the following year, the patches of thin and deep soils, and if territories. The cock has the dual role of distance moved in 195 records was these pockets were seeded with wild defending his territory from the equivalent to less than one territory flower mixes. Cuttings in the chalk and incursions of other cocks (Plate 22), and away. limestone of southern England offer the of attracting and guarding a hen; the greatest potential, but probably almost hens are attracted to cocks with larger Paternal and fraternal proximity was also every new scheme could be improved territories. normal. In 14 out of 38 records of upon. However, to fulfill the potential of a territoriality in surviving fathers and their new roadside, butterfly and other wildlife The function of territorial behaviour may sons, the sons gained territories adjacent habitats must be incorporated into the be to provide an adequate supply of to their father's territory and a further 14 original design, before the earth-moving some resource, such as food, cover, and settled at a distance equivalent to less plant has departed. ITE is uniquely space for sexual display, or for some than two territories away. In the case of qualified to advise on this potential other purpose. Whatever its function, one orphaned cocks, eleven out of 28 enrichment of the British countryside, because of its pioneering research on the establishment of wild plant swards (Wells, Bell & Frost 1981) and its detailed analyses of the requirements of different butterflies (Thomas 1984, 1989; Munguira & Thomas 1990),

J A Thomas and M L Mungira* (*Departamento de Biologia, Universidad Autonoma de Madrid)

References

Munguira, ML. & Thomas, IA.. 1990. The exploitation of road verges by butterfly and burnet populations, and the effect of roads on adult dispersal and mortality. J appt Ecol. In press.

Thomas, J.A. 1984. The conservation of butterflies in temperate countries: past efforts and lessons for the future. Symp. R. ent. Soc. bond., 11, 333-53.

Thomas, J.A. 1989. The Hamlyn guide to butterflies of the British Isles. 2nd ed. London: Hamlyn. Plate 22. A cock red grouse

55 These speculations on differential 500 behaviour between kin and non-kin of red grouse are consistent with the 0 400 100 m .N observations of Green (1983) on the red- cn ''•...... • legged partridge (Alectoris rufa). g 300 200 In any one year, the chance that a 0 territorial father has a son in the 100 Son territorial population in the subsequent year is assumed to be directly related, first, to the size of the father's territory 0 20 40. 60 80 100 and, second, to the family relatedness of Year '• Father Son the neighbours, to the extent that they Figure 53. Segment of simulated recognise each other as father, son or population fluctuations for a survival rate brother. The interplay of these two distributed randomly as a normal variate, factors, territory size and kinship, is with mean=0.5 and SD=0.05 sufficient to explain the cyclical rise and • . : ..... fall in population size...... •• There is no need for a change in the It is suggested that, at the low point of rules governing the behaviour or in the Figure 52. Territories of a red grouse the cycle, the territory holders are few genotype of the cocks, as suggested by father, his three sons and other and well spaced out, territory sizes are Chitty (1967), between low and high neighbouring cocks at Kerloch Moor in large, young cocks can more easily densities; the slowing of recruitment and spring 1974 establish territories close to those of their the decline in population size can be fathers, and the level of recruitment is explained simply by variations in the high. As the population increases, the level of aggressiveness between territories established on the study area average territory size decreases, and neighbours. According to the model, this were more or less coincident with their smaller territories will lead to a reduced aggressiveness depends upon the natal territory, and a further eleven recruitment rate of young cocks. This degree of family relatedness and upon settled at a distance equivalent to less reduction, in turn, together with the territory size. This conclusion may apply than two territories away. Of the 29 average 57% loss through the mortality of more generally. Any species exhibiting instances of marked full brothers older cocks, will reduce the degree of fidelity to natal territory and differential obtaining territories, eight were adjacent kinship between neighbours. As his own behaviour between kin and non-kin has and only five settled at a distance territory becomes smaller, an individual the potential to show cyclical, fluctuations equivalent to more than four territories cock, though still behaving differentially in population size. away. towards kin and non-kin, would be more likely to attack or threaten other cocks: 'Cyclical fluctuations in vole (Microtus A model has been built based on these Combined with the shrinkage of territory spp.) numbers may be similarly observed factors, but most statements are size, a reduced degree of kinship would explained. Charnov and Finerty (1980) speculations which have yet to be tested. thus lead to a further reduction in suggested that aggression between recruitment rate. As the population unrelated individuals may be a factor The data are consistent with the idea that further increases, this falling recruitment causing population cycles. Subsequently, competition amongst father and sons is continues, until the recruitment rate Boonstra et al. (1987). demonstrated that less intense than that between unrelated equals the replacement rate. At this fidelity to natal territory, usually among cocks. Father, sons and brothers may be point, the male territorial population is at females, occurs in most microtine relatively tolerant in the territorial its peak. Then, the effect of the greater species. The situation is thus similar to contest, directing their aggressiveness to aggressiveness between neighbours, that found in red grouse, except that, in unrelated neighbours. associated with the continuing decrease voles, it may be the female which drives in kinship, is likely to tip the population the cycles. With increasing population densities and into decline. The population should smaller territory sizes, however, the continue to decline until the effect of M D Mountford and A Watson tolerance between fathers, sons and increasing territory size cancels that of brothers is likely to decrease, but decreasing kinship. Cock numbers thus competition between the members of a reach a low point, after which the References family group will probably still be less recruitment rate rises again, and the next cycle begins. Boonstra, R., Krebs, G.J., Baines, M.S., than that between unrelated cocks. In Johnson, M.L. & Caine, I.T.M. 1987. Natal order of priority, a cock is likely first to philopatry and breeding systems in voles attempt to meet his own territorial The ITE mathematical model shows (Microtus spp.). j Anim. Ecol., 56, 655-673. requirements, then those of his cycles with a period of about six years father/son/brother, and last of all to (Figure 53). Charnov, E.L. & Finerty, J.P. 1980. Vole accept unrelated cocks as neighbours. population cycles: a case for kin-selection? The hen's strategy seems quite different. Oecologia, 45, 1-2. The outcome of the territorial contest She apparently invests little effort in the presumably determines the cock's establishment of territorY, but chooses Chitty, D. 1967. The natural selection of self- subsequent success at mating with a hen between the male territories on offer, and regulatory behaviour in animal populations. Proc. ecol. Soc. ABU:, 2, 51-78. and producing offspring, and the mutual then directs most of her energies to the tolerance of father and son should give production of young. If the territory is too Green, RE. 1983. Spring dispersal and them an advantage in inclusive fitness small relative to others nearby, she . agonistic behaviour of the red-legged over a pair of unrelated, and moves on in search of a more attractive partridge (Alectoris rufa). j Zool., 201, consequently more hostile, neighbours. territory (Moss, Watson & Parr 1988). 541-555.

56 Moss, R. & Watson, A. 1985. Adaptive value around the British coast, mainly on bare reintroduced in 1837 in Perthshire, and of spacing behaviour in population cycles of or stony ground. In northern Britain, spread rapidly through the new conifer red grouse and other animals In: Behavioural particularly NE Scotland, it spread inland ecology- ecological consequences of adaptive plantations which were, at that time, behaviour, edited by RIVL Sibly & RH. Smith, along river systems in the last century, mainly of Scots pine (Pinus sylvestris). 275-294. Oxford Blackwell Scientific. nesting on banks and shingles A Large-scale planfing from 1919 by the progression from riverine habitats to Forestry Commimion and, more recently, Moss, R., Watson, A. & Parr, R. 1988. Mate nesting on farmland far inland has been by private concerns used a high choice by hen red grouse Lagopus lagopus noted especially since the 1940s. On percentage of fast-growing North with an excess of cocks - role of territory size these farmlands, the oystercatcher nests American species, notably Sitka spruce and food quality. Ibis 130, 545-552. alongside the lapwing (Vanellus (Picea sitchensis). This practice has vanellus), which, in contast, is resulted in large tracts of closely spaced, decreasing in number, as in many other uniform, even-aged trees which shade parts of Britain. Both species were out the ground vegetation when the studied near Banchory, Kincardineshire, canopy closes (although a .mosaic of • Effects of land use changes on land farmed on a 5-7-year rotation of mixed-age trees may develop in the leys, root crops and spring-sown cereals, second rotation). on wildlife and on the adjoining moorland edge. The (P-art of this work was done in collaboration two species differed in their use of Droppings consisting entirely of spruce with the Macaulay Land Use Research Institute habitat during the breeding season in needles have been found in forest with (MLUR1), and was funded by the Department two important respects. 64% Sitka spruce and 7% Scots pine, and of Agriculture and Fisheries for Scotland) captive birds seem to subsist as well First, although they nested on all upon spruce as on pine. The structure of The effects of changes in land use upon available substrates, there was a clear the forest and the growth form of the wildlife can be immediate, ie when a and consistent preference by trees may be as important for grass field is ploughed for cereal oystercatchers for recently cultivated capercaillie as the food value of tree production, or longer term, through, for land, and by lapwings for the wetter, species which comprise the forest. Our example, afforestation or the removal of broken ground, particularly that of the objective is, therefore, to determine sheep from hill grazings. Some animals moorland edge. Many early lapwing which features in the structure encourage are favoured, while others cannot adapt nests on farmland were destroyed by and support populations of males and disappear. Some species may be cultivation, but, as oystercatchers started (females appear to be more adaptable in 'pre-adapted' to exploit novel, man- to nest 22-45 days later than lapwings, their choice of habitat). To this end, modified environments, especially when they avoided most field operations. The cocks were counted on their display the changed environments closely flat, bare, open fields closely resemble grounds (leks) in spring in a wide range resemble their ancestral fiabitats. Others, the ancestral coastal sites of of forest types. Habitat maps are being such as ungulates, are able to modify oystercatchers, at least when the nests made covering a circle of 1 km radius their own environment, especially when were initiated; indeed, so do the flat roofs from each lek, which should include the their populations are uncontrolled. of some buildings in Aberdeen, where home range of most of the displaying up to 120 pairs of oystercatchers now birds. These maps are used as a basis Whatever the cause, a change in land nest each year. for sampling the forest structure in use affects the species composition of the relation to capercaillie use. The height, vegetation, and, equally important for Second, lapwing chicks are largely spacing, shape and clumping of the trees wildlife, the structure of the habitat dependent upon wet and/or grazed fields are measured, as well as features of the provided by the plants usually changes with a short sward to which they are ground vegetation. Principal component also. Structure' may be considered as taken by their parents. As more land is scores derived from thesp measurements having vertical and horizontal drained or used for the cultivation of will be used to describe forest structure components. Vertical structure is autumn-sown cereals, and less grazing quantitatively. concerned with the depth and layering of land is required for cattle, suitable fields the vegetation, from the soil surface up to to rear young have declined. There is an early indication that the the canopy of ground layer, shrub or Oystercatchers, however, are unusual openness of a forest is important for tree. Horizontal structure involves the among waders in that they may carry capercaillie. Not only are adults able to shape (slender, branching, bushy) and food back to their young. They are, fly between the trees, but sufficient light patchiness of individual plants or therefore, better adapted to exploit reaches the forest floor for a ground associations of plants within the overall habitats normally considered unsuitable layer of vegetation to grow. This ground community. Techniques are being for wader chicks. In summary, the layer is typical of old forests and of great developed for quantifying these concepts lapwing, once so typical of the British importance for chicks, which feed on the so that the significance for wildlife of farming scene, is not as well adapted as vegetation (notably blaeberry (Vaccinium change in land use can be inferred from the oystercatcher to overcome changes myrtillus)) and associated insects. vegetation composition and structure. in farming practice, with the result that its numbers are decreasing in many areas, In the following contrasting examples whilst those of the oystercatcher are Moorland drawn from recent studies on Scottish increasing. The grazing of rough grasslands and farmland, conifer forest and moorland, moorland by sheep has played a major aspects of adaptability, land use change, Forest role in determining the vegetation and habitat structure are illustrated in composition and structure of the Scottish three major environments. Throughout Europe and the USSR, most uplands. Decades of continuous grazing populations of capercaillie (Tetrao and burning to provide fresh growth for Farmland urogallus) are closely associated with old sheep and red grouse (Lagopus 1. coniferous forests. The bird probably scoticus) have prevented the The oystercatcher (Fiaematopus became extinct in Britain in the establishment of shrubs and trees. Large ostralegus) is a wading bird which nests eighteenth century. It was successfully areas of heather (Calluna vulgaris)-

57 dominant moorland have been converted Site A to grassland, sometimes through 0 Grass • Dwarf shrub intentional 'improvement', but often as an >50 unplanned consequence of overgrazing 0 Moss • Trees / shiubs and overburning. A gradual change is 41-50 I 0 Other Oa' currently taking place, in which sheep 31-40 stocks are being reduced in the uplands z 21-30 as production is increased in the CD lowlands. The effects of this change on 11-20 the vegetation and the consequences for 6-10 r) wildlife in the uplands need to be 3 understood. 0-5

Although the main pathways for 100 50 0 50 100 secondary vegetation succession after the Mean cover per plot (%) removal of sheep have been studied for the main upland vegetation types (Miles, Site B Welch & Chapman 1978), much of the component work on which our understanding has been based was done >50 on a small scale and involved exclosure 41-50 plots. Extrapolating from these plots to ary entire hillsides is difficult, as exclosures 31-40 a. take no account of the ranging behaviour 21-30 Fl of the herbivores which can change the 11-20 ?) habitat. Furthermore, workers have paid c7; scant attention to the vertical structure of 6-10 -6 the vegetation, which is likely to have 3 far-reaching effects on ground-dwelling 0-5 — fauna. 150 100 50 0 50 100 In the collaborative study with MLURI, Mean cover per plot (%) the changes in vegetation composition Figure 54. Vegetation profiles of dry heathland at two locations where sheep and structure on hillsides are being grazing was removed over 20 years ago (reduced treatment), compared with quantified, following the short-, medium- adjacent control areas. Red deer graze both reduced and control treatments and long-term removal of sheep (Figure at site A; these sites have very similar profiles. Red deer have been 54). The data are being compared with excluded from the reduced area at site B, and the profiles differ markedly changes in indices of abundance of key wildlife species. The sampling procedure involves measuring the height, cover and Factors which complicate this basic Reference layering of all the plant species in such pattern include burning and the vegetation communities as wet or dry • presence or absence of red deer Miles, J., Welch, D. & Chapman, S.B. 1978. heath, grassland, flushed grassland and (Cervus elaphus) and lagomorphs. At Vegetation and management in the uplands. bracken (Pteridium aquilinum). The several sites, virtually no changes in In: Upland land use in England and Wales, 77-95. (Publication CCP III.) Cheltenham: numbers of pellet groups in sheep, deer,- vegetation were found, even after 20 Countryside Commission. lagomorph and grouse droppings are years without sheep (Figure 54, Site A), counted, and a search is made for runs of but at most of these sites there was field vole (Microtus agrestis) and other significantly more red deer dung on the signs of their activity. Sites in eleven areas where sheep had been removed. areas in eastern, western, and northern One effect of removing sheep appears to Scotland have been chosen, covering be an increase in the distribution and regions with different soils and climate. numbers of red deer, with the result that Each site has an 'experimental' area little structural change in the vegetation where sheep were removed from one to occurs on such sites. 20 years ago, and a control area which is still grazed by livestock. Conclusion On heaths and dry grasslands, the Such studies as these described are general effects of sheep reduction, in the improving our ability to predict the effect absence of other factors, are that the of land use change on certain species vegetation grows taller, moss and litter which might be at risk, or which may strata increase in depth, heather invades benefit from such changes. The defmition grassy sites, and woody shrubs and trees of habitat structure in relation to animals' eventually enter the community. The requirements also provides targets for increase in bulk of the grass, litter and management to enhance wildlife moss layers provided habitat for voles in populations, eg in Environmentally 1989, a year of low vole -numbers; a Sensitive Areas. significant increase in vole activity was found only one year after grazing had N Picozzi, D C Catt and D Hope* been reduced on a recently fenced site. (*MLURI)

58 Comnumity ecology

Community ecology is a complex area of Beetles on Snowdon In I989, weekly collections of beetles on science concerned with the interactions Snowdon were made by pitfall trapping between plant and animal communities. The work of the montane ecology group in closely grazed montane fescue The increasing anxiety over regional and at ITE Bangor includes research on the (Festuca) grassland at altitudes of 660 rn, global issues has led to a growing effects of gradients of altitude (and thus 860 m, 980 m and 1055 m above sea demand for predictive work at the of climate) on plants and animals, and level. community level. studies of exclusively montane habitats and species. Current research on beetles Some species were found only at the The following three articles illustrate the on Snowdon can thus be divided into an higher altitudes, others mainly at lower broad nature of ITE's approach to the examination of the relationship between altitudes, and yet others at a wide range subject. All have conservation value. beetle species and altitude, and a study of altitudes. The ground beetle Patrobus of the rare Snowdon rainbow beetle assimilis and the rove beetle The study of beetles in Snowdonia relates (Chrysolina cerealis). Geodromicus longipes are both the distribution of beetle species to exclusively high-altitude species (Figure altitude, especially the factors limiting 55). The ground beetle Pterostichus such distribution. Knowledge of the Beetle species and altitude madidus, however, is a very common climatic limits on their range and about low-altitude species (Figure 55), which is their interactions with one another are The objective is to determine why high- found at the lowest of the four sites. The important in order to predict the altitude species occur where they do, species Carabus problematicus, similar to consequences of any possible climatic and why lowland species are unable to the common violet ground beetle change. extend farther up the mountains. The (Carabus violaceus), is known to be a answers to these questions should wide-ranging species, found above The second study of scrub succession provide some indication of the beetle's altitudes of about 300 m. The fewer shows the value of long-term studies in likely response to changing climate. number of beetles found in the lower ecological research and manipulative field trials. The study followed the succession of vegetation for 16 years after clearance of old dense scrub, with i. Patrobus assimilis Geodromicus longtpes particular emphasis on rotational management for conservation parposes. Seral communities such as scrub are 35 7 dynamic and must be treated so, and not 30 60 as static systems, as in a plagioclimax. 25 5 Too often, scrub communities arise by — 20 40 chance, following changes in land use or -1? 15 tu 30 grazing pressures. Active rotational cp 20 management provides a means of 5 10 creating habitats and landscapes to maximise species and habitat richness. 0

By contrast, the third article describes nitrogen cycling in a montane tropical Pterostichus madidus iv. Carabus problematicus rain forest in Costa Rica. Field and 9 laboratory trials were conducted to 20 determine if soil nitrogen processes were 8 reduced with increasing altitude, and to 70 identify the factors affecting those 0 60 processes. The raising of temperature 50 alone had little effect on mineralisation or fi) 10 40 nitrification, the important determinants 30 being soil aeration and moisture. This 5 2 study is relevant to the conservation of 1 tropical forests, as drying of the soil VMS 0 following clearances could result in 1055 980 860 660 1055 980 860 660 increased mineralisation and leaching Altitude (m) losses, making regeneration in montane Figure 55. Numbers of four beetle species collected in 1989 at four areas difficult. sites of decreasing altitude on Snowdon 59 sites on Snowdon seems to be related to i. Nebna gyllenhali the lack of shelter for this large species 100 (about 20 mm in length) at these 90 particular sites (Figure 55). 80 Differences in life cycles are also apparent between species. For example, 70 -z-n- some species, such as the ground beetle o 60 Nebria gyllenhali, appear to be autumn — breeders, whereas others, such as the 50 rove beetle Othius angustus, appear to 40 be spring breeders (Figure 56). co Differences in the timing of the life cycle 30 can also occur within a species, in populations at different altitudes, and 20 these differences are now being 10 examined. 0 Jun Jul Aug Sep Oct To what extent does climate limit the distribution of the species? The Othius angustus responses of the beetles to gradients of 16 temperature and humidity are currently being studied. At the same time, the microclimatic conditions of the various 14 sites, and, within the sites, beneath stones and in crevices where the beetles shelter 12 from severe conditions, are also being studied. —10(I)

coa) 8 "cr) The Snowdon rainbow beetle (Plate 23) a) co 6 is a leaf beetle found in Snowdonia but nowhere else in Britain. It is also found in central Europe. 4

Our present interest in the species is 2 because it is one of the protected species (Schedule 5 of the Wildlife and 0 Countryside Act 1981) for which the Jun Jul Aug Sep Oct Nature Conservancy Council in Wales is Month (1989) preparing a monitoring strategy. In Figure 56. Numbers of two beetle species collected on particular, we need to know the optimal Snowdon at different dates in 1989: (i) an autumn breeder, and time of year and time of day at which it (ii) a spring breeder can be found, in order to estimate changes in its population from year to year.

There are several interesting problems relating to the Snowdon beetle. It is restricted to one main site, and yet its host plant, wild thyme (Thymus drucei), is common in the uplands and near the coast of Britain.

The beetle's life history is to be studied in captivity. In particular, its behaviour at different times of day and night will be monitored to see when it is most likely to be found in the field. Its behavioural response to microclimatic conditions, especially temperature and humidity, will be examined. The microclimatic conditions of the site will also be compared with apparently similar sites on Snowdon, to determine whether this particular site has any peculiarities.

A Buse Plate 23. The Snowdon rainbow beetle

60 (The Nature Conservancy Council collaborated in the setting up and general management of the plots at Castor, and provided some financial contributions to the research in 1973-76 and 1985)

In ecological studies, it is rarely possible to follow changes in the field over a long period of time. Although the long-term approach provides key information in understanding factors which maintain biodiversity, there are few examples of such data in the scientific literature because of the resources required in terms of manpower, time and finance for continual monitoring (Likens 1987). Results may not be produced quickly; also, such work is often empirical because of natural events which cannot be foreseen, and which prove to have a vital bearing on later events. Any disadvantages, however, are outweighed by new insights gained, as long as the Plate 24 Aerial photograph of plots 1-10 at Castor Hanglands NNR in 1988. (Plot I. top righ( recording has clear objectives, and was only a trial plot and is an irregular clearing. plot 2, top left, was cleared in 1974 and has quantitative, reproducible methods of been in succession since 1979: positions of transects are visible from trampling dunng recording, data collection are adopted. and some deer paths can be seen)

Since 1974, successional vegetation work at Castor Hanglands, but some mean numbers were 122 (240 species following clearance of old dense scrub in results and discussion follow. overall). There were many interesting sequential plots has been studied at species of disturbed ground (Plate 25), Castor Hang lands National Nature some of which had rarely or never Reserve, near Peterborough. The previously been recorded from the objectives are to record colonisation and The old scrub before clearance was Reserve. Annuals or biennials were change in a rotational management dense, casting heavy shade, and only 57 frequent, eg henbane (Hyoscyamus system for the conservation of seral scrub flowering plant species were recorded. niger), fluellen (Kicicda spp.), stone communities. One plot of 50 m x 50 m of Even in the first year after clearance, the parsley (Sison amomum), mullein the old scrub (c 30 years in 1974) was mean number of species per plot was (Verbascum thapsus) and mignonette cleared every year from 1974 to 1988 108 (a total of 244 species for all 12 of the (Reseda spp.). Others persisted for (Plate 24). All plots have four random first year plots). In the second year, the longer, eg milk-vetch (Astragalus permanent belt transects, of 0.5 m x 25 m, one in each quadrant. The transects are marked by buried metal markers (an important feature of long-term work is reliable permanent marking systems in the field). All woody plants in the transects were recorded by searching through each metre of vegetation; seedlings and regrowth were distinguished and heights measured. The dominant flowering plants were noted, together with total numbers in transects and plots.

Plots received differing initial treatments after clearance, because it was decided that immediate regrowth of the shrubs should be reduced. The first four plots were mowed in July and September for four years after clearance. Such frequent mowing seriously reduced the seeding of early successional plants. Therefore, a second group of four plots was mowed once in late September for four years after clearance. The remaining four plots were left unmown.

It is not possible here to describe all the Plate 25. Early successional plants _M the second year after clearance of plot 13 in 1987 61 glycyphyllos), deadly nightshade (Atropa 4000 4 • belladonna) and pepper saxifrage End of mowing Mowed in July and September (Si laum silaus). Many of these species \. apparently germinated from the soil seed bank, responding to the opportunities presented by the bared ground following clearante. 200 3A,

Twenty-two species of woody plants have SA \ been recorded. Most of the common Ja• species occurred predominantly as 2 o • seedlings in the first year. Blackthorn \ • 'A. C/3 _ (Prunus spinosa) regrew from cut stumps o - 0 and suckers. Figure 57 shows the total (O 6 0 numbers of woody plants in the plots and -00 End of mowing Mowed in September the subsequent losses. There were 0 9 4 striking differences between plots in the .! 2000 initial numbers recorded. These differences were partly related to meteorological factors, because seedling D- numbers were mrrelated in the different a) plots in particy .ar years: 1979 and 1985 7 8 - - 0 were good years for several species .775 0 (Table 13). Preliminary analyses show cc Ex that seedling numbers relate more to the 12 No treatment volume of seed produced by the parent bushes than to the weather for germination. Another aspect of initial seedling occurrence was the relationship to topography. Where plots had ancient 2000 ridge and furrow, there were generally 134; 11 C.,. more woody plants on the ridges (Figure 10° •. N 58); few species were associated with the • \ 0 hollows, notably willow (Salix spp.). tn, • • 0 Succession in later years 1 3 5 7 9 11 13 15 Species numbers were highest in the first Years after clearance or second years and then fell slowly, reaching a mean of 76.8 per plot after ten Figure 57 Total numbers of woody plants, as rooting frequency, in the transects of plots 2-12 and the deer exclosure (Ex) for all years of recording. years. From the results so far, the (Plot I was a trial plot and recording was discontinued: deér exclosure had untreated plots are losing species faster only two transects and figures doubled for direct comparison) than the mowed plots. The sequence of dominants was from dog's mercury (Mercurialis perennis) and rough-stalked meadow-grass (Poa trivialis) (with disturbed ground species, eg stinking Plate 26). In the next few years, The influence of deer on the woody qroundsel (Senecio viscosus), St John's vegetative reproduction by suckering is plants has proved far greater than wort (Hypericum spp.), creeping expected to become more important; expected, and so a deer exclosure was buttercup (Ranunculus repens), ground indeed, there was a slight rise in set up in 1985. The contrast in growth has ivy (Glechoma hederacea), creeping numbers of woody plants in the oldest been remarkable. For example, after five thistle (Cirsium arvense) and musk thistle plot (Figure 57). years, the mean height of unbrowsed (Carduus nutans)), through to Yorkshire fog (Holcus lanatus), tufted hair-grass (Deschampsia cespitosa), and towards bush grass (Calamagrostis epigejos) and Table 13. Total numbers of seedlings of six species in the first three years after clearance to woody plants. The speed of the change show years of abundance depended greatly on the initial Species colonisers, and transition matrices are Year Plots Comel Hawthorn Ash Rose Bramble Willow being used to analyse the process in more detail. Tall dense grasses suppress 1977 2-4 51 64 1 336 740 3 the establishment of most woody plant 1978 3-5 48 34 0 284 76 0 seedlings, and, if the initial woody plants 1979 4-6 97 93 35 1160 693 50 fail to grow on, succession can be 1980 5-7 35 27 6 106 41 1 deflected to a longer phase of grassland 1981 6-8 13 21 18 27 40 10 (Ward 1990). This deflection has 1982 7-9 155 14 16 928 328 7 occurred at Castor because of the 1983 8-10 51 9 7 87 54 0 0 mowing and deer (Cervus nippon) 1984 9-11 236 32 378 318 333 1985 10-12 495 91 21 1296 1089 27 browsing (eg plot 2 is still grassland 1986 11-13 145 67 63 505 180 0 eleven years after mowing ceased -

62 blackthorn was 75.9 cm (n=229, SD 38.8), seral change. The initiation phase and while the browsed plants were 22.5 cm the subsequent early competition (n=463, SD 10.6). In 1989, all woody between herbs and grasses and woody plants in the transects were examined for plants are critical to the course of the damage, revealing that 40% (n=1939) had succession. Indeed, scrub-resistant tall been browsed, Many undamaged plants dense grassland might have a function in were short and hidden in the denser some species-poor amenity grasslands for grass, or were protected by clumps of reducing scrub invasion. thorny bramble (Rubus) and gorse (Ulex) which acting as 'nurse shrubs', shielding the more palatable woody species. Ephemeral early successional Spacing patterns generated in this way communities of bare/disturbed ground probably persist through to older scrub. are less often included in management plans for Nature Reserves, and in some cases could be combined with rotational scrub management (Ward 1990). The old Management of seral communities like scrub should continue until there is a scrub must take account of dynamic high proportion of bared ground caused change: it is not practical to maintain a by shading, to enable greater activation, steady state like a plagioclimax. The with less competition, for species of the management of succession requires soil seed bank following clearance. Early information about the timing of successional species need to set seed colonisation and the persistence of well to replenish the seed bank for a species of plants and animals during later rotational clearance.

Plate 26. Transect A in plot 2 in 1989, eleven 10 Willow years after mowing ceased, with tall dense Spindle-tree grasses and herbs suppressing colomsation by 10 Hawthorn woody plant seedlings. Marker canes stand in buried metal tubes at each end of the transect

Gorse 10 Browsing and grazing animals may Elder greatly influence the course of 10 succession. In management terms, it is 20 Cornel advisable to consider whether the objectives of scrub rotational management should include the early 20 Cherry control of deer and/or rabbit (Oryctolagus cuniculus) numbers.

-(7-80 Bramble The Castor trial has stimulated the development of ideas about the 60 relationships between management and Ce the landscaping of dynamic systems. Too 40 many scrub communities have resulted by chance in areas which have remained 20 ungrazed. The rotational management of structurally diverse vegetation provides an opportunity for reshaping and Rose landscaping on Nature Reserves and 40 amenity sites in ways which will also maximise the species richness of seral 20 communities.

L K Ward

40 References 20 =5 0 2 4 6 8 10 12 14 16 18 20 22 24 Likens, G.E., ed. 1987. Long-term studies in ecology - approaches and alternatives. Distance along transect (m) London: Springer-Verlag. Stinreng groundsel Wood fragments cPU Bare ground Ward, L.K. 1990. Management of grassland - Dog's mercury scrub mosaics. In: Calcareous grasslands, Figure 58. Numbers of woody plants in each metre of transect D in plot 12 in ecology and conservation, edited by S.H. the first year after clearance (1985) (with soil profile of ridge and furrow, and Hillier, D. Wells & D.W.H. Walton. Bluntishant key to dominants) Bluntisham Books. In press.

63 Nitrogen cycling in montane Both ammonium- and nitrate-nitrogen can 2. incubated under field conditions at the be taken up directly by plants. collection site for between 12 and 15 tropical rain forest Mineralisation and nitrification are, thus, days and then extracted with 1 M KC1, (This project was part of a multidisciplinary key processes likely to control the as above; study on the ecology of tropical mountains, productivity of trees. Both processes are and was done in conjunction with Operation mediated by soil micro-organisms and 3. incubated under laboratory conditions Raleigh and colleagues from the Organization are reversible. Generally, mineralisation at the base of the mountain and of Tropical Studies (OTS) at La Se lva Field extracted as in treatment 2 above. Station in Costa Rica. The study was partly occurs when the soil organic matter and funded by the award of a Travelling soil micro-organisms act as a source of Field and laboratory mineralisation and Fellowship to R H Mans from the British inorganic nitrogen, ie when the reaction Ecological Society) is positive, and immobilisation takes nitrification rates were then calculated by place when soil organic matter acts as a difference (incubated - initial It is well known that the species sink, ie the reaction is negative. concentrations) and expressed on a 14- composition and structure of moist day basis for comparative purposes. tropical forests change with altitude This study tested the hypothesis that soil (Grubb 1977), and one of the hypotheses nitrogen processes were reduced at 4. air-dried, sieved to pass through a 2 put forward to explain these changes is altitude in moist tropical rain forest on a mm mesh and brought back to Britain that nutrient cycling is reduced at high single mountain, Volcan Barva, in Costa for chemical analysis. The rate of altitude. In particular, the transformations Rica, Central America, and assessed the mineralisation was estimated in of nitrogen (N) in the soil were thought to environmental factors which might affect incubations made under 'improved' change with altitude (Marrs et al. 1988). such processes. At six altitudes (100 m, conditions, ie where soil aeration and In most ecosystems, soil nitrogen occurs 500 m, 1000 m, 1500 m, 2000 m and 2600 moisture conditions were ameliorated mainly in organic matter, and it is made m) on the mountain, permanent study by adding acid-washed sand and a available for plant uptake through the plots of 1 ha were established (Table 14). known volume (6 ml) of deionised processes of mineralisation and In January 1986, ten replicate surface soil water. nitrification. Mineralisation is the process samples were collected from the plots, where soil organic nitrogen is broken and treated as follows: The three ways of measuring down by soil micro-organisms to produce mineralisation, therefore, represent a inorganic ammonium-nitrogen, and 1. extracted with 1 M KC1 for an initial series of estimates where two nitrification describes the subsequent measure of inorganic ammonium- and environmental parameters which were reaction which forms nitrate-nitrogen. nitrate-nitrogen concentrations; thought most likely to limit the

Table 14. Some characteristics of the six 1 ha plots sampled on Volcan Banta, Costa Rica

Basal area of The five leading families Plot woody species Height of Height of No. of individuals No. of species ranked by number of altitude Prevailing Prevailing (?-.10 cm dbh) main canopy tallest tree of woody species of woody plants woody species (m) slope (°) aspect (m2) (m) (m) (<10 cm dbh) (>10 cm dbh) cm dbh)

100 5 23 35-40 45 494 112 Mimosaceae Lauraceae Annonaceae Euphorbiaceae Meliaceae 500 0 24 38-43 50 391 121 Mimosaceae Meliaceae Rubiaceae Lauraceae Annonaceae 1000 20 NE 31 35-40 45 546 106 Lauraceae Mimosaceae Rubiaceae Moraceae Meliaceae 1500 15 N 27 30-35 38 553 63 Lauraceae Euophorbiaceae Myrtaceae Rubiaceae Cyatheaceae 2000 15 NW 27 25-30 35 448 59 Melastomataceae Lauraceae Rubiaceae Euphorbiaceae Piperaceae 2600 15 N 25 20-25 32 503 32 Lauraceae Melastomataceae Araliaceae Cunoniaceae Symplocaceae

64 mineralisation process in montane soils i. Field Mans, R.H., Proctor, J., Hawley, A. & were alleviated temperature in the _30 Maantford, M.D. 1988. Changes in soil laboratory incubations, and both '(I) • nitrogen-mineralization and nitrification along temperature and moisture/aeration in the a's.20 an attitudinal transect in tropical rain forest in 'improved' incubations. Costa Rica j EcoL, 76, 466-48a 7110

Nitrogen mineralisation under the cp 0 'irnproved' conditions increased markedly ii. Laboratory with increasing altitude (Figure 59i), from .5 20 150 tug to 760 pg N gi 14 daysl between To' 100 m and 2600 m. Under field k= 10 • conditions, the opposite result was found, the rate declining with increasing altitude 1000 2000 from 38 pg to -2 pg N og-1 14 daysl Altitude (m) (Figure 59ii). Figure 6(1 Changes in nitrification rates under (i) field, and (it) Under field conditions, there was a laboratory incubation along the change from mineralisation to altitudinal transect on Volcan Barva, immobilisation at the highest altitude, and Costa Rica; the outer lines show the similar results were found in the approximate 95% confidence limits, and the closed circles represent the laboratory study (Figure 59ii). mean values

i. 'Improved' completely reversed, and the rates were 1000 between one and three orders of magnitude greater. The implication of this 800 finding is that soil structure and aeration • 600 • may be the major factors limiting the breakdown of soil organic nitrogen in -— 400 • these montane soils. The soils from high altitude in Costa Rica were very wet, -o 200 with moisture contents (80%) double those found in the lowland forest sites. 7 Moreover, the montane soils had much higher concentrations of total nitrogen ii. Field and laboratory and carbon, and extractable cations 8, 50 (Mans et a). 1988). 2 40 2 30 These conclusions have some relevance 20 10 for the wise resource management of 0 montane tropical forest vegetation. The -10 combination of high rainfall and the high 0 1000 2000 store of both nutrients suggests that such Altitude (m) ecosystems will be particularly Figure 59. Changes in vulnerable to erosion and leaching losses nitrogen/mineralisation ih soils along the after forest destruction. This vulnerability altitudihal transect on Volcan Larva, was to be expected. However, the Costa Rica, under (I) 'improved' dramatic increase in breakdown of the incubation conditions, and (h) field incubation (fitted curve and soil organic nitrogen under the approximate 95% confidence limits) and 'improved' conditions implies that any laboratory incubation (solid vertical management change which dries out the lines = means + 95% confidence soil, either directly by draining or limits). The closed circles represent the indirectly if the soil surface dries when mean values of the raw data for the soil temperatures are increased following respective curves forest clearance, could stimulate mineralisation and intensify leaching losses. Thus, the regeneration of montane Under field and laboratory conditions, the tropical forests may be particularly nitrification rate also decreased with difficult, and may take much longer than increasing altitude (Figure 60), from 22 forests at low altitude. pg to 3 pg N g - 1 14 days-i. R H Marrs and M D Mountford A major conclusion from this study was that the rates of nitrogen mineralisation, or indeed of nitrification, did not improve References when temperature alone was ameliorated. When the soils were dried Grabb, P.J. 1977. Control of forestgrowth and and aeration improved, the relationship distribution on wet tropical mountains. Annu. between altitude and mineralisation was Rev. Ecol. Syst., 8, 83-107.

65 Scientific services

Large areas of work of ITE's are of today's pressing environmental current form of ADDSOAP in field trials underpinned by ideas and equipment problems. over the past five years. developed and supported by scientific services. Many of the newly emerging During initial development, solutions problems of the environment require focused on the detection of air moisture, more precise ways of measuring, before Automated dry/wet surface wetness, and impact. The latter, solutions can be suggested. although novel, appeared promising. deposition separator of aerial Experiments with thin stretched The three articles in this section highlight precipitates (ADDSOAP) diaphragms and a microphone exploited one aspect of the work of scientific the self-resonance effects of metal sheet services. They illustrate how instruments Much research is dependent on structures and vibration transducers in are being deployed by themselves, in instrumentation. However, apparently contact with them. Selective filtering of purpose-built structures and in laboratory simple requests dan be technically transducer signals and summing vibration conditions, to solve some of the current demanding and require an iterative patterns produced a programmable problems. sequence of development and trial. One precipitation sensor, where the size and illustration is an application of research rate of impacts, or their combined effect, The atmospheric deposition of pollutants on atmospheric deposition, where it could be set as a threshold for cover takes place both in the wet and dry state. became increasingly important to opening. It is this precipitation sensing It is important to distinguish between distinguish wet deposition in rain from module which has been named these two states as they have been found the dry deposited gases and particulates, ADDSOAP; it can be linked to a variety to affect vegetation in different ways. The and occult deposition, ie mist, fog and of cover mechanisms, can drive first article describes an instrument cloud water. It is often preferable to motorised valves, and can even drive a which is able to separate wet and dry separate the 'true' wet deposition by motor-powered wheel or Pulley for deposition by the use of an automatically covering the funnel between rainfall moving a structure when a pre-selected activated cover. It is able to sense when events, or providing some means to precipitation threshold occurs. raindrops start falling, and then moves its wash down the funnel at the start of a cover over the collecting funnel. rain event in order to remove Development of ADDSOAP has accumulated dry deposits, keeping progressed over the years, from the In modern pollution research, emphasis is these washings separate from the original metallic funnel cover which now concentrating on the more subtle following rainfall. Manual control of a acted as a vibration sensing board when effects of long-term low-level exposures fimnel cover is neither practical nor cost- rain landed on it. The vibrations were of phytotoxic gases on vegetation. The effective. Collectors have, therefore, transmitted down the cover support to an second article describes the facilities at been developed with automatically electronic sensor. The processed signal ITE Bangor for carrying out this type of activated covers, the movement of the then controlled the motor drive to research. A range of equipment has cover being controlled via some form of position the cover. Although this device been specially designed to fit into rainfall sensor. was compact and battery-powered, the hemispherical glasshouses. Gas control cover did not fit tightly on the water and delivery systems are housed in a The commonest form of sensor detects collection funnel, and insects were able separate building, with gas analysers and the moisture from raindrops, generally as to seek shelter and add to the nutrients a sampling system for monitoring dome a conductive path between of the collected sample. Also, although gas levels, and the supply and sample electrodes/conductors. However, a large the device responded quickly to impacts, lines are ducted underground to the sensing area is required if the initial and it did not detect fog or fine drizzle. glasshouses. most contaminated raindrops of a shower Modifications were made which largely are to be captured. There are also other overcame these problems. The modified The final article discusses the services practical problems, eg detecting instrument had a separate rain impact provided by the Institute's analytical raindrop size, contamination and power sensing box, a ventilated wet and dry chemiStry group, and illustrates its requirements. As a result, this form of thermistor psychrometer to open the response to the high demand for the sensor was rejected at an early stage in cover under high humidity conditions, analysis of aqueous solutions, as a result the development of ADDSOAP. and a counterbalanced cover plate which of increased interest in water quality. revolved through about 2200 to form a Autosamplers and computerised control Ideally, a sensor should detect, within good fit on the funnel. However, this with data processing systems have been seconds, the start and finish of a rain design was not ideal in windy conditions; used to increase throughput. event and the type of rainfall. It should the cover could be buffeted as it swung require no maintenance, be .reliable, and over between the open and closed state, These three articles emphasise how a have low power consumption if battery or could be lifted from the seating on the wide range of instruments and technical operation is considered. These funnel. In the open state, deposits on the knowledge are being used to solve some requirements have been fulfilled in the exposed underside of the cover could be

66 transferred to the funnel. The wind direction sensor and time recorder of rainfall occur, and vegetation may be psychrometer also posed maintenance would allow the catchment of plumes of cloaked for prolonged periods in highly problems, with a lack of water to the wet industrial pollutants. Warren Springs acidic mists. ITE's new pollution research wick and thermistors changing their Laboratory is currently evaluating a facilities at Bangor have been designed characteristics. range of precipitation sensors, including to allow studies of the effects on plants of the ADDSOAP, with a view to uprating both low, controlled concentrations of In 1986, an improved device was built the performance of their nationwide gaseous pollutants and acidic mists. (ITE Annual Report 1985/86, p139) for distribution of wet-only precipitation research into acid waters in Wales, using sensors. Studies of the effects of gaseous experience gained from the two previous pollutants on vegetation are conducted in models. It incorporates a separate The ADDSOAP concept has developed four hemispherical glasshouses, sounding board, which houses the control over the years to its present reliable manufactured by Rosedale Engineering' electronics and a synthetic fibre form. One can only guess at future Ltd, of Filey, Yorkshire, and marketed humidistat for moisture detection. The requirements, and adapt the device under the tradename 'Solardome l(Plate funnel cover is designed on a four-rod accordingly. 27). The system is based on one support, offset parallelogram principle, so originally designed in 1976 at the that the top of the cover is always D G Benham University of Lancaster (Ashenden et al. exposed to the elements; the cover lifts 1982) and later modified (Lucas, Cottam off the funnel but remains horizontal until & Mansfield 1987). The domes of the almost clear of the funnel, when it tilts Bangor glasshouses are smaller, which down the side out of the way below the provides a number of operational funnel rim. An optional system has two advantages. Each dome has a diameter of funnels, one for wet and the other for dry 3.05 m, a floor surface area of 7.31 m2, collection, with the parallelogram arms (The experimental work was partly funded by and is coTistructed of glass mounted in an moving the cover from one funnel to the the Welsh Office) anodised aluminium framework (Figure other. Years of field trials have shown 61). The domes are ventilated by two that wind or insect problems are of little Assessments of the potential toxicity of 'Sonoxcarb' fan filter units (Machine significance, and the ventilated humidity pollutants to vegetation have changed Control Ltd, Horsham, Sussex), each of sensor has proved reliable and substantially in recent years. Emphasis which supplies two domes. Ambient air is maintenance-free. The stainless steel has moved away from studies of acute drawn through charcoal filters, to remove impact sensor has no moving parts or injury to plants, caused by high any background levels of pollutants, and surfaces to contaminate, and thus concentrations of the phytotoxic gases — then pushed, via a flow splitter, along the requires no maintenance. The electronics sulphur dioxide (SO2), nitrous oxide (NO2) main air supply trunking to each dome. control the mechanical positioning of the and ozone (03) —to investigations of the The air then enters a mixing chamber cover, so there are no switches to go more subtle effects of long-term low-level where controlled quantities of pollutant wrong or out of alignment. The exposures. In addition, the importance of gases may be added before it is pushed electronics can be programmed for the wet deposition has been recognised as a into the dome and released around the desired size and frequency of impacts to major route for the influx of potentially circumference via a ring of lay-flat operate the cover; the cover is replaced damaging levels of sulphur, nitrogen and perforated polythene tube. Finally, the air over the funnel within a minute of these hydrogen ions into ecosystems. The leaves the system via a central chimney conditions ceasing. Use of the humidistat importance of this route is particularly cowl at the top of the dome. The rate of with impact sensing will capture almost evident in the uplands, where high levels air movement through each dome is all wet deposition; using the humidistat and impact sensor in opposition, the cover will open only when fog, dew or mist (occult deposition) occur, closing as soon as impacts are detected. The cover can operate equally well in the reverse mode, keeping the funnel dry or discarding occult deposition.

An eight-bottle carousel and timer have also been produced to sample portions of a shower, or for daily sampling. The whole apparatus can run for several weeks from a charged car battery, which can be routinely changed or, for continuous use, recharged from a solar panel or domestic mains supply.

A number of ADDSOAPs are now in operation around Britain, and, although such a device has a specialised and limited market, current trends towards pollution control and environmental consciousness are likely to extend the market. Applications of ADDSOAPs have, so far, included building weathering, acid rain, nutrient input, airborne metals and trace element studies. Integration with a Plate 27 Solardome' fumigation system

67 systems control building and pumped underground to the Meatme.nr bays, 2/7-22, where they are ultrasonically atomised In A a compressed air sffeam (Plate 20 .

The facilities have been used for a range of experiments aimed at hieterminiug effects of pollutants on Me growth and //\ ,------..".-- physiology of plants. One exolillPle IS an 6 i D 1,27,, C \ B investigation of the effecD of low concentrations of SO2, NO2 and 03, ---- together with the additional stresses _ imposed by wet acid deposition, on Me 12,97.0 ens) Figure 61. Diagram of 'Solardorne' fumigation chamber. 11,fan filter umh B. growth of white clover (lIftacalicT!") flow sphtter; C, air ducting; D, mixing chamber; E, lay-Ilat tubing F, cv 'Grasslands Huial. The e. enment was ventilation cowl; G duct to second Solardome chamber conducted on 6-week-old . Dings grown in 7.5 cm diametr • _ contabarig John Innes no. 2 potting COMQ0St Atmospheres within the glass'nouses wefe set to provide treatments of adjustable up to two complete air differential regulation-type controllers changes per minute. with needle valves (Rosemount Ltd, i. charcoal filtered an (cormol) Stockport, Cheshire). All domes are fitted ii SO2+NO2 Gas control and delivery systems, with supply lines for the three pollutants iii. 03 together with gas analysers and a and PTFE sampling lines. The levels of iv. SO2+NO2+03. sampling system for momt ci ing dome gas pollutants in air sampled from the domes levels, are housed nearby in Me systerfs are measured using Monitor Laboratories Concentrations of SC- J1i0? were control building, and all supply a-nd analysers for 03 and NO2, and a Meloy maintained at 40 pplo, ror two ai 80 bob on sample lines are ducted underground to flame photometric analyser for 502. additional peaks, one of the domes. Ozone is generated by a Thursdays, and one of 3 __ 777L80 bob, bona laboratory ozonator (Wallace & Tiernan, Studies on the effects of acid mists on immediately followed by J. lc at 110 Tonbridge, Kent); SO2 and NO2 are plants are conducted in a closed 7.3 m x 03, on Tuesdays. Monitorino Me dome. supplied from cylinders of compressed, 3.3 m polythene tunnel, divided internally atmospheres ensured tiler concentraMons pre-diluted gas, as opposed to Me pure by polythene sheeting into four treatment remained +2 ppb of 3fre ciesired _levels. gases used in the earlier, Drger bays, Each bay is fitted with a Sonicore There were three evperrmeMal 01oct‹3 of 'Solardome' systems (see Lucas, Cottam & atomising station (Lucas Dawe 20 plants in each dome, -Me roositions of Mansfield 1987). The advantage of this Ultrasonics, London), which can provide which were rotated weekly Five olanis system is an easier control of pollutant low volumes (0.4-8.0 1 11-I) of treatment from each block were allocated to concentrations, without the need for solutions at particle sizes in the range receive one of four acid Toast CreatmeriLs. expensive mass flow contmllers. Instead, 5-30 mn (equivalent to fogs/clouds). For all misting treatments, svants were flows are maintained using constant 2ilreatment solutions are made up in the

10

8

2

0 Control SN 03 Figure 62 Mean dry weight yields of white clover after exposed for 15 weeks to different aerial pollutants and a •If Plate 28. Misting un 7/2 oceration mist treatments

68 taken from the domes and transferred to and change. Reviews have been autosampler and computerised the polythene tunnel misting system. The neces.Rary to respond to the heavy and control/data processing system have mist fteatments were three 4 h exposures changing demands for analytical services enabled the grouping of all water (Mondays, Wednesdays and Fridays from ecological scientists and to update methods, and provided the option to each week) of: the book Chemical analysis of ecological operate the analysers unattended materials (Allen 1989), which has become overnight. i pH 2.5 a standard text since it was first ii pH 3.5 published in 1974. Improving laboratory operations to iii. pH 4.5 increase sample throughput inevitably iv. pH 5.6 This text was based on the routine results in greater demands in terms of methods pioneered and developed from data handling and report preparation. In which provided a total wet deposition of agricultural and water quality methods, anticipation of these demands, our 6 mm per week Mist solutions were applied and modified in ITE's Merlewood strategy has been to upgrade each made in deionised water by adding laboratory for conservation experimental analytical instrument to incorporate a equimolar concentrations of arnmonium work The book contained separate personal computer (PC)-based data nitrate and sulphuric acid. Supplementary sections on soil, vegetation and water handling package, thus enabling the use watering of plants was done by watering analysis, and covered all aspects from of purpose-designed software from the can at a rate of 24 mm per week, using a sample collection to calculation of results. different manufacturers, and then to link simulated acid rain solution of pH 4.5 Full practical details were provided for each data system to the main Station (made up by additions of sulphuric acid well-tried and tested methods, with computing network Thin-wire ethernet and nitric acids in the ratio of 7:3, by manual and semi-automated options to has been installed, and each PC volume). allow users to select a procedure to communicates with the Station's VAX match their facilities. The second edition mainframe computer via PCSA, which Plants were removed from the treatments now includes theory, methodology and provides a network communication for after 15 weeks of exposure, their roots practical hints, based on 30 years of automatic and transparent file storage. washed free of soil, and then dry weights experience in supplying chemical data, a Data collation/manipulation and report obtained. The results are shown in Figure vital component to ecological research. generation can then be completed using 62. An analysis of variance of the data The text has been completely revised any of the extensive range of facilities revealed significant differences (P<0.001) and updated to reflect the rapid growth and packages available on the mainframe between both gaseous and acid mist in technology and the change in machine. treatments, and a significant (P<0.001) emphasis from conservation to applied interactive effect of gases x mists on ecology. The section on instrumental It is still too early to assess the full total plant dry weights. Averaged over all procedures has been extended to benefits of upgrading the automated misting treatments, there were reductions incorporate liquid chromatographic analytical equipment, but, even during in total plant dry weights of 16.6% for methods and a vast experience in the difficult commissioning period, there S02+ NO2, 43.6% for 03 and 59.4% for automated methodology. Advances in has been a gradual increase in SO2+NO2+03, in comparison with control computing, statistics and data evaluation throughput. Our records show that 9000 plants. For misting treatments, a lower procedures promoted a radical revision samples (7000 natural waters) have been yield was averaged over all gas of the final chapter, whilst the emphasis analysed for 70 000 determinations, an treatments at pH 2.5, compared to pH 4.5 in water analysis was shifted towards the increase of 40% over the last two years, and pH 5.6. The significant interaction of handling of solutions from nutrient cycling gas x mist treatments revealed that yield at similar staffing levels. and pollution studies. reductions which were attributed to the During the last year, the group made a more acidic mist treatments (pH 2.5 and Several of the analytical facilities major contribution to cycling/leaching 3.5) were not significant in the 03 and available to the group have been studies carried out under contract in S02+ NO2+03 treatments; the severely upgraded over the last three years. Devi Ila Forest, by developing a harmful effects of these gaseous pollution During 1989, the major commissioning methodology to measure sulphur and 5- treatments appear to override any effects work on an inductively coupled 35. Labelled S was applied to soil under of the mists. plasma/optical emission spectrometer trees, and its progress was followed T W Ashenden, C R Rafarel and S A Bell (ICP/OES) was completed. This through the ecosystem. Sample instrument uses a sequential scan to preparation techniques were developed perform multi-element analysis for the for leachates (throughfall and stemflow References wide range of applications required. solutions), pine needles, and soils to Methods have been developed and overcome interferences in the counting Ashenden, T.W., Tabner, P.W., Williams, P., tested for routine analysis of major procedure. Pre-concentration of leachates Whitmore, M.E. & Mansfield, T.A. 1982. A cations on aqueous solutions (sodium, was required to detect precisely both the large-scale system for fumigating plants with calcium, magnesium, iron and aluminium). S-35 by alpha counting (using a high SO2 and NO2. Environ. Pollut B, 3, 21-26. ICP/OES has been particularly useful for ionic tolerant fluor) and the sulphate ion Lucas, P.W., Cottam, D.A. & Mansfield, T.A. trace metal analysis in soil and chromatographically. Matrix interference 1987. A large-scale fumigation system for vegetation. In some instances (eg boron), proved to be a major problem in investigating interactions between air pollution the sensitivity and performance of the vegetation acid digests, but was and cold stresson plants. Environ. Pollut, 43, instrument have proved far superior to eventually avoided by using an oxygen 15-28. other methods available bomb combustion procedure. Soil leaching was simulated by sequential High demand for the analysis of aqueous aqueous extraction, pre-concentration of solutions has also required reorganisation combined extracts, and then Analytical chemistry group of the anion determinations, previously measurement of S components. carried out on dispersed continuous flow Over the past few years, the group has equipment. Upgrading the equipment As a laboratory involved in long-term been going through a process of review and the addition of a large capacity studies, it is important that the chemical

69 data produced are reliable and of consistent quality. In addition to the rigorous internal quality control procedures developed, the group paiticipates in national and international standardisation schemes to monitor performance and contribute to the scientific community as a whole. This year we have advised on the methodology for EEC certified reference material, and set up a scheme for a large 'international programme studying tropical soil fertility. Our work has also extended to the Sudan, where we have advised on the development of ecological analysis and provided technical support to commission equipment.

In addition to supporting contract work led by other ITE scientists, the group itself continues to generate income for the Institute by providing a service for sister organisations, such as the British Antarctic Survey and the Nature Conservancy Council, as well as universities and private companies.

A P Rowland

Reference

Allen, S.E., ed. 1989. Chemical analysis of ecological materials. Oxford: Blackwell Scientific.

70 ISBN: 1 85531 028 7 For further information please contact:

Institute of Terrestrial Ecology (North) Bush Estate Penicuik Midlothian EH26 OQB Telephone: (031)4454343/6

Institute of Terrestrial Ecology (South) Monks Wood Experimental Station Abbots Ripton Huntingdon PE 17 2LS Telephone: (04873) 381/8

Natural Environment Research Council £6.00net