by Ian Woiwod, Phil Gould & Kelvin Conrad
Although many of the rarest British moths are now receiving special conservation attention we still know far too little about the current fortunes of the more common and widespread species. Because of the dedication of an army of volunteers throughout the UK and the foresight of entomologists at Rothamsted Experimental Station over the last 70 years, we are one of the few countries in the world to have any quantitative information on this important topic. In this article we describe the long-running national light-trap network, lay to rest some common misunderstandings about its origin and purpose, and present some results from recent work on moth declines that underline the vital importance of collecting such data into the future.
The Origins of The Rothamsted Insect Survey
Rothamsted Research (previously Rothamsted Experimental Station), near St Albans in Hertfordshire, is the oldest agricultural research station in the world. Amongst other things, it is internationally known for its long term ‘Classical’ field experiments on the effects of fertilisers that were started on the Rothamsted farm in 1843 and are still going strong and providing new biological insights 160 years later (e.g. Lewis 1992; http://www.rothamsted.bbsrc.ac.uk/resources/ClassicalExperiments.html ). Rothamsted also has a long history of entomological studies, and in the 1930s and 40s the head of the Entomology Department, C.B. Williams (of butterfly migration fame), ran a trap beside one of the fields on the Rothamsted farm from which all the moths were counted on a daily basis. Results from this, and other traps on the estate, were published in an influential series of papers covering many aspects of moth biology. These included: the effects of weather and moonlight on catches, the flight times both through the night and year, the influence of trap height on catch and the statistical measurement of moth diversity (long before the word biodiversity was even coined). These papers were highly original for their time and are still well worth seeking out by anyone interested in moth ecology and light-trapping (e.g. Williams 1935, 1936, 1939, 1940a, 1940b, 1944).
The story could well have ended there had not a young entomologist, Roy (L.R.) Taylor, come to Rothamsted in the 1950s. He was particularly interested in how and why populations fluctuated from year to year, but also realised that many insect groups, including moths, can be very mobile. He therefore knew that he needed to study populations of as many species as possible at different places to see how populations were synchronised and if the synchrony related to species’ mobility. These were very revolutionary ideas at the time but there were no suitable insect population data available with which to test them. One practical answer was to enlist the large number of enthusiasts interested in moths who could run traps and provide Rothamsted with the data. At the time, the robust and relatively cheap ‘Rothamsted’ light trap designed by C.B. Williams (Williams 1948; Fig 1) was the obvious choice and as a start, in 1960, one was placed beside the same Rothamsted field that C. B. Williams had sampled in the 1930s and 40s. As the trap was identical in design and light source, this immediately provided unique information on long-term changes on farmland moth populations.
The results were both surprising and alarming, as there had been an overall decline during the 1950s of about 70% in the total number of larger moths, and many once-common species at that site had either become very rare or had disappeared. No comparable data exists from elsewhere in Britain so we don’t know how widespread or synchronised any farmland moth decline was during this period, but we do know that the current moth population at that site is now very typical of arable and grassland fields in the UK (Woiwod 1991; Fig. 2). It therefore seems very likely that the decline of farmland moths in the 1950s was widespread, probably a result of the rapid agricultural intensification that took place at that time in response to real post-war food shortages. Included in this intensification were increased mechanisation leading to hedge removal and field enlargement, and the widespread adoption of herbicides and insecticides. All these factors are likely to have been highly detrimental to our moth fauna.
From these initial results it became apparent that in addition to providing invaluable research data, moth populations needed to be monitored much more widely if we were to understand population changes as they occurred. It was to fulfil both these purposes that the Rothamsted Insect Survey (RIS) national light-trap network was set up, building slowly in the early 1960s. The light-trap network was created by enlisting volunteers to run traps throughout Britain and by 1968 there was a good national coverage. In all over 450 sites have been sampled with 54 sites operating for 15 or more years. Currently, 87 traps are in operation (Fig. 3).
At about the same time the light-trap network was being established, a complementary network of 12.2m (40ft) high suction traps was also being set up to monitor aphids. These two networks became known as the Rothamsted Insect Survey and have provided a wealth of scientific information on the way insect populations change over time and space, the responses of insects to climate and environmental change and the measurement and analysis of biodiversity (Taylor 1986; Woiwod & Harrington 1994). More recently a joint grant with Butterfly Conservation from the Esmée Fairbairn Trust has enabled us to look at trends in individual species of common moths in much more detail. This work is still in progress and some interesting and rather worrying results are beginning to emerge (Conrad et al. 2004, see later).
The Rothamsted Trap Design and The Need For Dead Samples
Although the Rothamsted-style light trap is not now widely used by amateur moth recorders, it has proved ideal for long-term, quantitative, standardised, national monitoring. One important feature of the trap is its opaque roof, which not only protects the sample from bad weather but has also been shown by careful experimentation to improve the consistency of catches from night to night compared to the more widely used Robinson MV traps (and probably many of the ‘open-topped’ traps commonly in use today). This rather surprising effect is thought to be related to changes in the height of flight of some of the more powerful moths, such as noctuids, which can be very influenced by local weather conditions from night to night, making the sample size much more erratic and therefore less representative of actual population size. As the Rothamsted trap only catches a relatively small sample from low level in the immediate vicinity of the trap it is little influenced by this effect (Tayor & French 1974; Intachat & Woiwod 1999).
The Rothamsted traps use 200w clear tungsten-filament bulbs, partly for continuity with the historic data and because the relatively small samples obtained are practical to deal with, without any likelihood of harming moth populations. This latter point is particularly important when running a scheme where the samples have to be killed for later identification, often away from where the traps are run. In an increasingly conservation-conscious world the necessity for taking dead samples is often questioned, but we have found no practical way of obtaining an extensive set of daily data from live sampling. The time and commitment required from expert volunteer trap operators would be just too great.
In the spirit of the earliest days of the network, experienced trap operators often identify much of the catch themselves, sending only difficult or unusual specimens to Rothamsted to confirm identification. Some operators post the contents of the traps to volunteer identifiers, and about half of the sites send their catches to the Insect Survey where they are identified by Phil Gould.
There can be conservation issues associated with regular live trapping using MV-type traps, as many moths tend to settle on vegetation around such traps and it doesn’t take birds and bats long to discover this convenient food supply. The Rothamsted trap is particularly benign in this respect because it is placed on a stand 1.2m above the ground (Fig. 1) which, combined with the opaque roof and base, mean that moths approaching the trap either go in or enter a dark zone above or below the trap and continue on their way without. For scientific, long-term quantitative monitoring of moth populations it has become increasingly apparent over the years that the Rothamsted choice of trap design, with its small but consistent sample size and simple operation, is as near optimum as possible for such work.
The concern about taking dead samples is not a new one and the whole subject of the ethics of insect collecting was addressed by C.B. Williams in the 1950s, who actually used samples from a Rothamsted trap to illustrate the lack of a long-term effect of such sampling on moth populations (Williams 1952). All of his points are still valid. In his and our view, insects should never be killed casually or thoughtlessly nor should sampling have any detrimental effects at the population level. We are therefore careful to ensure no vulnerable or very localised populations are sampled and are happy to enter a dialogue with anyone concerned about the issue. In our view this careful approach is fully justified by the results of conservation importance emerging from the unique RIS datasets.
The Garden Tiger story
The RIS moth data has been widely used in research over the years and at the last count well over 600 publications have made use of it in one way or another (Woiwod & Harrington 1994). However, until recently, relatively little analysis had gone into one of the most pressing topics of interest to all conservationists—just what has been happening to our moth fauna? The first detailed study of a single species using RIS data was of the Garden Tiger Arctia caja (Fig. 4), a species once regarded as common and widespread but now known to have declined and largely disappeared from many areas in England. Analysis confirmed that not only are there 30% fewer RIS sites recording the species now than 30 years ago but even at inhabited sites there has been a 30% decline in abundance. The decline has been particularly marked in the south-east, whereas other areas, such as Scotland, are less affected. It should perhaps be pointed out that the species has never been caught in large numbers in RIS samples and the change represented a sudden drop from an average of only 4.2 individuals per occupied site per year up to 1983, to an average of 3.0 individuals after 1984 (Conrad et al. 2002). Further analysis has shown that wet winters and warm springs are particularly detrimental to this species, and that changes in distribution and abundance may be an unfortunate consequence of recent climate change related to large-scale weather patterns over the North Atlantic (Conrad et al. 2003). The reason for winter and spring weather being so critical for this species are not fully understood but are likely to be related to overwintering survival of the young hibernating larvae.
Changes in other common and widespread species
Using the Garden Tiger work as a springboard, we have now been able to extract population trends using the full RIS moth dataset. Analysing records over the 35 year period from 1968 to 2002, and using very strict criteria for site and species inclusion to ensure the validity of the results, it has been possible to estimate population change for 337 species of macro-moth (Conrad et al. 2004). The results are very worrying, not only for the readers of this journal but also for anyone interested more widely in the conservation of British wildlife biodiversity because moth adults and larvae are such an important food source for mammals, birds and even other insects (Fox et al. in prep).
The percentage of species with >10% decreases in five years (54%) was more than double those with >10% increases in five years (22%) and summed across all species macro-moth abundance has declined by almost a third over the 35 year period (Fig. 5). However, it is clear that declines are not uniform across Britain with the strongest decline in the south, particularly the south-east, and the fewest declines in the north. In fact, overall total moth abundance in the north has remained fairly stable with species in decline being balanced out by those increasing.
At the species level there is particular cause for concern. Just over 20% (71) of the 337 species are declining at rates greater than 3.5% per year, a rate which is generally regarded as cause for serious conservation concern (IUCN World Conservation Union 2001). It shouldn’t be forgotten that these are all species which are generally regarded as common and widespread. Until recently, none has given cause for concern or been thought to warrant any conservation priority. Hopefully this will change with the wider publication of these alarming results (Fox et al. in prep.).
It is not possible to list all 337 species with their trend details here, but as a taster Table 1 lists the top 10 of both decreasing and increasing species. More species and further examples will be given in The State of Britain’s Larger Moths report which is due to be published soon (Fox et al. in prep).
What is Causing These Changes to Our Common Moth Fauna?
The patterns of decline and increase in some of our common moths are intriguing at the very least and undoubtedly result from a variety of factors. However, let’s start at the beginning with some explanations for our results that have been suggested but we are sure are incorrect. Following recent media coverage of our preliminary results it didn’t take long before some wit phoned Radio 2’s Wake up to Wogan to offer the solution that it was all our fault, i.e. we were catching so many moths that it was hardly surprising that they were getting rarer. Well, for reasons already given above, we are sure that’s not so. In addition to our small sample size and the early analysis by C.B. Williams (1951), we have looked very carefully at longer runs of data without finding any consistent pattern that could be ascribed to a trapping-out effect. Apart from anything else the patterns of decline don’t support such a scenario, with different rates of change in different regions even for the same species. For example, Scalloped Hazel Odontopera bidentata has strongly declined in the south-east, remained stable in the south-west and increased significantly in the north (Fig. 6). Such examples (together with the basic understanding that insect population dynamics are driven by very high natural birth and death rates in each generation) exclude any possibility that our small light-trap samples are having the slightest influence on moth abundance over the long term.
However, we do need to consider very carefully other possible biases that could be influencing our results and producing unwelcome artefacts. There are several candidates. One suggestion is that our pattern of sites has changed over the years and that the results may just be a reflection of that. As with all volunteer-based recording systems, it is certainly true that there is a turnover of sites. We have therefore repeated the analyses using only the longer running sites (i.e. those with at least 10-, 15- or 20-year runs of data) and our conclusions remain the same, confirming that trap turnover is not a significant factor (Conrad et al. 2004, and unpublished).
Another suggestion is that there might have been a change in the effectiveness of the light source in our traps. This is unlikely as the technology used in making the tungsten bulbs we use is long established and the bulbs are replaced several times throughout the year so that bulb age is not likely to be a problem. Again, the varied response of the same species in different regions also mitigates against this idea being valid.
Finally, there is the question of competition from light pollution. As light-trap sampling relies on the attraction of moths to light, it has been suggested that such traps may have become less effective as the background illumination has increased and our results might just reflect that. This is a serious concern and we have used available satellite data on changes in surface light emission between 1992 and 2000 to compare moth trends at light-trap sites with no change or actual decreases in background illumination to those with increasing levels of light emission. We found no difference in moth trends between the two light pollution categories so can be pretty sure that our results are robust in this respect (Conrad et al. 2004, and unpublished data).
Having established that the observed changes in moth populations are almost certainly genuine we can now pose the really crucial question. Just what has caused such a large imbalance between declines and increases in our common moths? The honest answer is that we don’t really know. Further work is now required to fully understand what is happening, particularly if we are to have any possibility of reversing the trend. Some relationships between trends and the life-history characteristics of groups of species have been found (Conrad et al. 2004). For example, many lichen-feeding species have in general been increasing, presumably as a result of lichen populations responding to reduced sulphur dioxide levels in the air. Conifer feeders have also been doing well, which might also be expected. In contrast, species that overwinter as eggs have been doing particularly badly and yet species with adults that overwinter have been increasing. Perhaps here we are beginning to see effects that might be related to recent changes in our British climate.
The detailed analysis already done on the Garden Tiger certainly suggested that climate change was an important component of its decline in abundance and change in distribution. However, there are almost certainly other factors at work in other species, which might include: land-use change, agricultural intensification (e.g. herbicide and insecticide use), urbanisation, and eutrophication from air pollution by nitrates. Our old enemy light pollution may also be implicated, not in this context as a competitor with the traps but as a pollutant in its own right, preventing normal breeding behaviour of many species. Teasing out the relative importance of these various factors is vitally important but will not be easy, particularly as many are closely correlated.
How can you help?
Long-term population monitoring of any species rich insect group is never going to be easy. However, we hope from the above account that the importance of the RIS light-trap network is apparent. We are always in need of volunteer trap operators or identifiers to fill in gaps in our coverage and replace sites that have been lost from our network for various reasons. As little as 5 minutes per day is required for daily trap operation and all equipment is provided and running costs met. The commitment comes in keeping the traps running continuously (although samples can be accumulated when necessary) and wanting to make a really useful contribution to our future knowledge of moth populations. Please contact us at Rothamsted or e-mail for further details.
The continuing, long-term research of the Rothamsted Insect Survey is supported by the UK Biotechnology and Biological Sciences Research Council (BBSRC), from which Rothamsted Research receives grant-aided support. We are also grateful to the Lawes Agricultural Trust. Recent research has been funded under the MAFS Initiative of the BBSRC, and through a collaborative study with Butterfly Conservation, funded by the Esmée Fairbairn Foundation. We would like to thank the many dedicated volunteers and staff at Rothamsted who have been involved in the RIS light-trap network over the years. Without their loyal support we would know a great deal less.
Conrad, K. F., Woiwod, I. P. & Perry, J. N., 2002. Long-term decline in abundance and distribution of the garden tiger moth (Arctia caja) in Great Britain. Biological Conservation, 106: 329–337.
Conrad, K. F., Woiwod, I. P. & Perry, J. N., 2003. East Atlantic teleconnection pattern and the decline of a common arctiid moth. Global Change Biology 9: 125–130.
Conrad, K. F., Woiwod, I. P., Parsons, M., Fox, R. & Warren, M. S., 2004. Long-term population trends in widespread British moths. Journal of Insect Conservation 8: 119–136.
Fox, R., Conrad, K. F., Parsons, M., Warren, M. S. & Woiwod, I. P., in prep. The state of Britain’s larger moths. Butterfly Conservation and Rothamsted Research, Wareham, Dorset.
Intachat, J. & Woiwod, I. P., 1999. Trap design for monitoring moth biodiversity in tropical rainforests. Bulletin of Entomological Research 89, 153–163.
IUCN World Conservation Union. 2001. IUCN Red List categories and criteria: Version 3.1. IUCN Species Survival Commission, Gland, Switzerland and Cambridge, UK.
Lewis, T., 1992. 150 years of research at Rothamsted: practice with science exemplified. Journal of the Royal Agricultural Society of England 153: 107–118.
Taylor, L. R. & French, R. A., 1974. Effect of light trap design and illumination on samples of moths in an English woodland. Bulletin of Entomological Research 63: 583–594.
Taylor, L. R., 1986. Synoptic dynamics, migration and the Rothamsted Insect Survey. Journal of Animal Ecology 55: 1–38.
Williams, C. B., 1935. The times of activity of certain nocturnal insects, chiefly Lepidoptera, as indicated by a light trap. Transactions of the Royal Entomological Society of London 83: 523–555.
Williams, C. B., 1936. The influence of moonlight on the activity of certain nocturnal insects, particularly of the family noctuiidae, as indicated by a light trap. Philosophical Transactions of the Royal Society of London (B) 226: 357–389.
Williams, C. B., 1939. An analysis of four years captures of insects in a light trap, Part 1: General survey; sex proportions; phenology; and time of flight. Transactions of the Royal Entomological Society of London 89: 79–132.
Williams, C. B., 1940a. The number of insects caught in a light trap at Rothamsted during four years 1933–1937. Proceedings of the Royal Entomological Society of London (A) 15: 78–80.
Williams, C. B., 1940b. An analysis of four years captures of insects in a light trap, Part II: The effect of weather conditions on insect activity; and the estimation and forecasting of changes in the insect population. Transactions of the Royal Entomological Society of London 90: 227–306.
Williams, C. B., 1944. Some applications of the logarithmic series and the index of diversity to ecological problems. Journal of Ecology 32: 1–44.
Williams, C. B., 1948. The Rothamsted light trap. Proceedings of the Royal Entomological Society of London (A) 23: 80–85.
Williams, C. B., 1952. Some notes on killing insects for collections and for scientific research. The Entomologist 85: 271–279.
Woiwod, I. P., 1991. The ecological importance of long-term synoptic monitoring. In The Ecology of Temperate Cereal Fields (ed. L. G. Firbank, N. Carter, J. F. Darbyshire & G. R. Potts), pp. 275–304. Oxford: Blackwell.
Woiwod, I. P. & Harrington, R., 1994. Flying in the face of change: The Rothamsted Insect Survey. In Long-term Experiments in Agricultural and Ecological Sciences (ed. R. A. Leigh & A. E. Johnston), pp. 321–342. Wallingford: CAB International.
Table 1. The top ten moths that have decreased and increased 1968–2002
|Satin Beauty||Deileptenia||ribeata||SE, SW|
|Treble Brown Spot||Idaea||trigeminata||SE, SW|
|Scarce Footman||Eilema||complana||SE, SW|
|Peacock Moth||Semiothisa||notata||Mainly SE, SW|
|Juniper Carpet||Thera||juniperata||Scarce everywhere|
|Grey Shoulder-knot||Lithophane||ornitopus||SE, SW|
|Broad-bordered Yellow Underwing||Noctua||fimbriata||Mainly SE, SW|
|Devon Carpet||Lampropteryx||otregiata||SE, SW|
|Dusky Thorn||Ennomos||fuscantaria||England, Wales|
|The Spinach||Eulithis||mellinata||Mainly England, Wales|
|Figure of Eight||Diloba||caeruleocephala||Mainly England, Wales|
|The Anomalous||Stilbia||anomala||Mainly England, Wales|
Ian Woiwod, Phil Gould and Kelvin Conrad, The Rothamsted Insect Survey, Rothamsted Research, Harpenden, Hertfordshire, AL5 2JQ
Figure 1. RIS light traps are operated at a wide variety of sites throughout Britain. This one at Glencoe in Scotland has been in operation for the last 10 years.
Figure 2. Mean annual catches of the longest running RIS trap (Barnfield), in 3 year categories, compared with average values for RIS arable and grassland sites.
Figure 3. Distribution of RIS light-trap sites in 2003.
Figure 4. The Garden Tiger Arctia caja
Figure 5. The total number of moths across all sites shows an overall decline of 32% between 1968-2002. The index of annual abundance measures relative changes in population size between years and is proportional to total moth abundance. Therefore, a 10% change in the annual index indicates a 10% change in total moth abundance.
Figure 6. Trends of the Scalloped Hazel in 3 regions showing different patterns of change in each region. The index of annual abundance, as in Figure 5, shows relative changes in annual abundance and therefore, the solid lines reflect rates of change in total population size for the species.