The history and control of the pine beauty moth, Panolis flammea (D. & S.) (Lepidoptera: Noctuidae), in Scotland from 1976 to 2000


Barry J. Hicks, College of the North Atlantic, 4 Pike's Lane, Carbonear, Newfoundland, Canada A17 1A7. Tel.: +1 709 596 6139; fax: +1 709 596 2688; e-mail: bhicks@


1 The pine beauty moth, Panolis flammea, has been a serious pest of lodgepole pine plantations in Scotland since 1976. It historically feeds on native Scots pine throughout Europe but population levels of P. flammea on this host have never been high enough to cause tree mortality in the U.K.

2 This paper reviews recent advances in the biology of the pest and documents control programmes from 1976 to 1999.

3 There has been practically uninterrupted population monitoring of P. flammea from 1977 to the present day in Scottish lodgepole pine plantations. Intervention with chemical spraying has often been necessary.

4 The population data suggest that populations of P. flammea may have had a cyclic pattern over the monitoring period, with outbreaks occurring at regular intervals of between 6 and 7 years.

5 The amplitude of population cycles was large during the 1970s and 1980s, but has dampened in recent years. Natural enemies are believed to contribute to this trend. Fungal disease, specifically, appears to have had a greater effect on pest populations in recent years than in the past and is suggested to have contributed significantly to the population dynamics observed since 1990.


The pine beauty moth, Panolis flammea (Lepidoptera: Noctuidae), is the U.K.'s most serious pest of established forests (Day & Leather, 1997). It is indigenous in the U.K., where it feeds on Scots pine (Pinus sylvestris) (South, 1961). There have been numerous recorded P. flammea outbreaks on P. sylvestris in Continental Europe over a period of almost 200 years (Klimetzek, 1979), whereas it has never reached pest status on Scots pine in Britain.

In 1973, P. flammea was recorded for the first time on lodgepole pine (Pinus contorta) from areas of northern Scotland where it had never previously been recorded (Stoakley, 1977). This tree species is native to north-west North America and was introduced into Scotland in the mid 19th century. However, it was not until the 1950s and 1960s that it was planted extensively (Lines, 1976). Lodgepole pine proved to be a good choice for afforestation of wet, acid and infertile soils of the north region. In addition to its initial value as a cash crop, it had the advantage of improving the soil sufficiently to be able to plant Sitka spruce, Picea sitchensis, which is the main commercial tree species in Britain. Outbreaks of P. flammea have occurred many times in the extensive lodgepole pine plantations of northern Scotland (Fig. 1), caused damage to thousands of hectares and, in some cases, caused heavy tree mortality. The purpose of this paper is to give an up-to-date history of the outbreaks and control programmes targeted against the pine beauty moth. In addition, recent advances in the understanding of its biology are reviewed.

Figure 1.

Locations of Panolis flammea outbreaks in Scotland from 1976–2000. Some of the areas indicated have had multiple outbreaks.

Panolis flammea life history and biology

In Scotland, pine beauty moth adults fly from late March to May while further south, in England, adults have been known to fly as early as the end of February (Stoakley, 1977). The eggs are laid in short rows on needles produced in previous years. The larvae emerge from the eggs at the end of May or early June and begin to feed on new growth where they consume the internal tissues at the base of developing needle pairs with their heads within the needle and their abdomens exposed. Fourth- and fifth-instar larvae feed on the older foliage (Watt, 1987). By about mid-July, the larvae are fully grown, drop to the soil surface where they burrow into the litter and pupate inside a hibernaculum at the soil–litter interface (Watt & Leather, 1988).

Evans et al. (1991) reviewed research on the pine beauty moth biology in the 1970s and 1980s. The major emphasis on research in that period was placed on the population dynamics of this pest on both lodgepole and Scots pine. A significant early finding was that outbreaks apparently tended to occur on lodgepole pine growing in deep poorly drained peat (Stoakley, 1979). In investigating the reasons for this link, the possible role of plant stress was assessed. Researchers hypothesized that trees growing in poorly drained soil were stressed and indications from the literature at the time suggested that the foliage of stressed trees may be better for some insect herbivores by accumulating more available nitrogen (White, 1974, 1984) and by reducing the plant's secondary compounds, which act as anti-herbivore defensive compounds (Rhoades, 1983, 1985). Although trees growing in deep peat were stressed, amelioration of these conditions did not affect the population development of the moth (Leather, 1993). Watt (1989a, b) showed that Scots pine is a better host than lodgepole pine for P. flammea and that the stressful conditions of growing in deep peat did not increase either the nitrogen content or decrease the tannin content of lodgepole pine, thus discounting the role of host quality in determining host and stand associations. One possible explanation for the difference in populations between the two tree species is that the survival of pupae is better in deep peat than in the litter layers beneath Scots pine (Leather, 1984; Evans et al., 1991). The optimum microclimate for pupal survival (i.e. temperature and moisture) is best provided on the peat sites that have deeper litter layers with good insulating properties (Evans et al., 1991).

Since the late 1980s, the attention of researchers has focused on natural enemies as potential determinants of the differential survival of the moth in forests of the two host plants. Predator exclusion experiments indicated that predation was higher in Scots pine sites than in lodgepole pine sites (Watt, 1989b). Aegerter (1997) showed that Scots pine plantations have both a greater number and diversity of parasitoids than lodgepole pine plantations. The difference between the forest types increased over the summer months until almost an order of magnitude difference in parasitoid numbers was observed between them. However, the presence of key parasitoid species between the forests was not different and thus was not considered to contribute to the difference observed in the population dynamics of the pine beauty moth between forests that had outbreaks and those that had not (Aegerter, 1997). Furthermore, the overall rates of larval parasitism was not shown to be significantly different between the two forest types, with larval parasitism rates around 30% in each (Aegerter, 1997).

Life table work on P. flammea populations showed that predation strongly affected survival of eggs and larvae (Barbour, 1987; Watt & Leather, 1988). Mortality of larvae rose to 99% by the beginning of pupation (Watt & Leather, 1988). Predation of larvae by arachnids was investigated by Docherty (1993). Scots pine sites had significantly more species of spiders and harvestmen than lodgepole pine sites. Docherty & Leather (1997) concluded that the difference observed in arachnid predators was due to differences in ground vegetation cover at each site. The number of harvestmen entering the canopy did not differ between Scots and lodgepole sites, but Docherty & Leather (1997) believed that the greater abundance of alternative prey items available to arachnids in Scots pine sites helped to maintain a larger population and may have increased the number of pine beauty moth larvae consumed.

The diversity and abundance of carabid beetles is significantly greater in Scots pine sites than in lodgepole pine sites (Walsh, 1990; Walsh et al., 1993). Feeding experiments showed that three Carabus species were potential pupal predators and that they were abundant when P. flammea pupae were present. The differences observed in pupal mortality between Scots and lodgepole pines are therefore consistent with the prevalence of the Carabus species, which may contribute towards the regulation of P. flammea in Scots pine (Walsh et al., 1993).

The record of a fungal epizootic in 1998 from field populations in Scotland indicated that the impact and diversity of the natural enemies affecting P. flammea may have changed (Hicks & Watt, 2000). The fungal diseases affecting the larvae are relatively recent occurrences in P. flammea in Scotland.

Population history of P. flammea: monitoring methods

Low numbers of the pine beauty moth were regularly observed by Forestry Commission personnel in yearly pupal samples taken to monitor the population of the pine looper moth, Bupalus piniaria L., in Scots pine forests from the 1950s onwards (Bevan & Brown, 1978).

When several outbreaks of pine beauty moth occurred during 1976–77 in the north of Scotland, it showed that there was an immediate threat to lodgepole pine in these areas, and the Forestry Commission undertook a systematic pupal survey in northern Scotland.

Sampling procedures were established in selected locations each covering approximately 30 ha. Within these locations, 11 positions were randomly selected within an area of about 0.3 ha and a portable frame, 30 cm × 30 cm × 15 cm deep, was used to sample the forest litter and peat. During the winter of 1977, 1200 ha of lodgepole pine plantations were surveyed using this sampling regime. On the basis of observed defoliation and tree mortality patterns, the Forestry Commission decided that a control threshold of 15 pupae/m2 would be used to determine if spraying was required to prevent tree mortality. This method of pupal sampling was carried out in high risk areas of Scotland from 1977 to 1993.

From 1993 until the present, pheromone traps augmented by selective pupal surveys have been deployed to monitor pine beauty moth in Scotland. Funnel traps, purchased from AgriSense-BCS (Mid Glamorgan, U.K.), were baited with a lure containing 25 µg Z-9-tetradecenyl acetate + 2.5 µg Z-11-tetradecenyl acetate. The traps were placed in the forests during the third or fourth week of March and left until early May, when they were emptied and all moths counted. Funnel traps have the advantage of not becoming saturated with moths so that an accurate total count can be obtained. Over several years of pheromone trapping and pupal sampling, estimates of the relationship between pupal density and the number of moths captured in traps have been made, resulting in the estimate that 100 moths captured corresponds to 1 pupa/m2 (D. Barbour, unpublished data). This relationship has been used in relating pheromone trap catches to the equivalent numbers of pupae in this paper.

To determine if a significant trend occurred in the population of P. flammea in Scotland, all of the data from the 13 long-term monitoring sites were pooled. Data from forests that were sprayed with insecticide in the year that they were sprayed plus the year after spraying were removed from the analysis of the pooled data. An average number of pupae from the 13 sites for each year was determined. An autocorrelation function (AFC) (Box & Jenkins, 1976) was obtained from the yearly means and a test for statistical significance of no serial correlation over the sampling period was determined using the Ljung-Box Q statistic (Minitab, version 12.1). Autocorrelation, defined as correlation of n with n + 1, n + 2,…, n + 20 (n = the average number of pupae in year 1977) allowed us to determine if a significant pattern of correlation occurs in the data set. The Ljung-Box Q statistic allows the statistical test of the null hypothesis that no serial correlation occurs up to the lag specified. While this statistic does not allow inferences on the shape of serial correlation in the data (i.e. whether there is a cyclic pattern or not) it does show whether the data are random or not.

Population history of P. flammea: results

Thirteen sites in northern Scotland (Fig. 2) have had almost entirely uninterrupted monitoring since the outbreaks of P. flammea began in 1976. The population levels at those sites, based on pupal sampling and pheromone trapping, are presented in Fig. 3. In addition, the data for all monitoring sites throughout Scotland from 1977 until 1999 have been pooled in Fig. 4 to show both the maximum number of pupae and the adjusted mean number (sprayed forests removed) of pupae per m2. Often the mean numbers within a forest were below the spray threshold, but localized abundance warranted a spray programme.

Figure 2.

Location of long-term monitoring sites for Panolis flammea in northern Scotland.

Figure 3.

Figure 3.

Populations of Panolis flammea from selected lodgepole pine plantations in Scotland. Dark solid bars represent the average number of pupae/m2 and the dark striped bars the average number of moths in pheromone traps (×10−2); the light bars represent the maximum number of pupae/m2. The threshold number for spraying (pupae/m2) is indicated by the broken line.

Figure 3.

Figure 3.

Populations of Panolis flammea from selected lodgepole pine plantations in Scotland. Dark solid bars represent the average number of pupae/m2 and the dark striped bars the average number of moths in pheromone traps (×10−2); the light bars represent the maximum number of pupae/m2. The threshold number for spraying (pupae/m2) is indicated by the broken line.

Figure 4.

The population of Panolis flammea from 13 recording sites in Pinus contorta plantations throughout Scotland from 1977–1998. Forests that were sprayed with insecticide were removed from the dataset. See legend to Fig. 3.

The pupal survey and pheromone trap data for the pooled sites suggested that the population of P. flammea in Scotland subjectively resembles a cycle (Fig. 4). The autocorrelation function of the data supported the possibility of the population being cyclical with a period of about 7 years (Fig. 5). However, due to the short data set, we cannot be certain that the pattern observed was cyclical. Significance testing of the null hypothesis that no autocoorelation had occurred over the sampling period (lag years = 20) using the Ljung-Box Q statistic showed that the autocorrelations were indeed significantly different from zero (LBQ = 32.56, d.f. = 20, P = 0.038). Thus, the data were not random. To test the assumption of a 7- year pattern, ACF analysis was repeated on the data that was adjusted to eliminate the cycle using a difference filter of 7 years. In this case, the statistical testing showed these autocorrelations not to be different from zero (lag years = 14) (LBQ = 11.48, d.f. = 14, P = 0.648), thus indicating a possible cyclic trend of the data.

Figure 5.

Autocorrelation function for Panolis flammea populations from lodgepole pine plantations in Scotland. The dashed lines represent 95% confidence limits for individual correlations.

The cyclic trend observed in the pooled population data can be observed from the individual forest sites. Although much local variation in population size occurred, many of the long-term monitoring sites appear to have had fluctuating populations that peaked every 5–8 years (Fig. 3C–H). Some of the forests have had the peaks occurring at longer intervals (Fig. 4A,B). Of the long-term monitoring sites, North Dalchork (Fig. 4H) has had the largest population size over the sampling years.

The history of P. flammea management in Scotland

The first outbreak of pine beauty moth in Scotland occurred at Rimsdale (Fig. 2) in 1976, when a significant area of lodgepole pine was severely defoliated (Stoakley, 1977). Attempts to control P flammea began in 1977. The first attempts were with the biological control agent Bacillus thuringiensis Berliner (Dipel®) but failed due to the feeding behaviour of the early instar larvae. These larvae feed on internal tissues of needles (Stoakley, 1979) and therefore the spray programme failed because larvae did not ingest enough bacterial toxin to cause significant mortality. The insecticide that was eventually used was the organophosphate fenitrothion. It proved to be effective in reducing the outbreak populations of P flammea up until the late 1980s (Table 1). However, environmental concern over the use of this insecticide resulted in other more environmentally acceptable control measures being tested and used. The chitin synthesis inhibitor diflubenzuron (Dimilin®) is presently the insecticide of choice of the Forestry Commission against P. flammea.

Table 1.  A summary of the forest area sprayed with insecticides against the pine beauty moth, Panolis flammea, in Scotland from 1977 to 2000
Year of sprayControlling agentArea sprayed (ha)
1977Dipel® (Bacillus thuringiensis Berliner)530
Dimilin® (diflubenzuron)Experimental
Bacilus thuringiensisExperimental
Total 25828

The frequency and amount of insecticide used has declined in the 1990s. While pine beauty moth numbers appeared to be rising in 1998, several naturally occurring fungal species were shown to cause a collapse in these populations (Hicks & Watt, 2000). Four fungal species were identified for the first time from P. flammea larvae in Britain. The fungi Entomophaga aulicae, Batkoa major, Nomuraea rileyi and Beauveria bassiana together reduced the larval numbers by 88% in some areas. Pupal samples taken in the autumn indicated a drastic reduction in P. flammea in areas where numbers were high during the previous autumn (Hicks & Watt, 2000). Further larval collecting in 1999 verified that a crash had occurred in the field population of P. flammea. This was the first study to show that fungal pathogens can affect P. flammea populations and reduce the numbers dramatically. Bioassays using B. bassiana against several life stages of P. flammea have indicated that this fungal species has good potential to be explored further as a biological control agent for P. flammea (B. Hicks, unpublished data).


This paper documents the population history from the time P. flammea outbreaks began in 1976 until the present day. Pupal counts and pheromone trap data suggest that the population is cyclic on P. contorta in Scotland. The pattern observed in the population dynamics of P. flammea from some individual forests appears, to some extent, to be an artefact of the spray programmes as outbreak populations were nearly always treated with insecticide. However, when the data are adjusted to remove the effect of insecticide spraying, there still remains a cyclic trend in the population. An autocorrelation of the adjusted means of the pooled data also shows a trend that resembles a cyclic pattern with a period length of approximately 7 years. This was supported by the observation of no significant autocorrelation in the data when they were adjusted to compensate for the 7 years cycle. Moreover, in cases where high density populations were left untreated in the past, the population normally declined during the following year due to starvation of late instar larvae and dispersal from heavily defoliated areas (Watt et al., 1989). These observations, together with data compiled during life table work, led researchers during the 1980s to conclude that natural enemies played a minor role in the population dynamics of P. flammea in lodgepole pine plantations (Watt & Leather, 1988). However, Watt & Hicks (2000) suggested that this line of thinking may have been incorrect and that natural enemies had been playing a greater role in the population dynamics of P. flammea than had previously been thought. By examining the population levels of P. flammea for several years after spraying, Watt & Hicks (2000) showed that the populations continued to decline or remained low for 2–3 years post-spray without additional spraying. They concluded that a delayed density-dependent factor (presumably natural enemies) was the most likely cause of this trend.

Although, based on limited data, we must be cautious about applying the term ‘cycle’ to the trend observed in the abundance of Scottish P. flammea populations, it is interesting to compare these results with those observed elsewhere for the same species. Panolis flammea has shown cyclic population dynamics on P. sylvestris in Germany since the 1800s (Klimetzek, 1979). The periodicity observed in Scotland (6–7 years) is comparable to the periodicity of P. flammea in Germany (8.1 ± 0.4 years). Biotic factors, namely insect parasitoids and fungal diseases, were shown to cause the collapse of many German populations in the past and to have contributed to the cyclic pattern observed (Barbour, 1987).

The frequency and impact of P. flammea outbreaks have declined in the 1990s. In addition, the oscillations that can be observed also appear to be dampening during the same time. The dampening of the fluctuations may be the result of the action of the entomopathogenic fungi that have been shown to affect populations of P. flammea in Scotland. During the summer of 1990, P. flammea larvae were collected from several locations in Scotland and no fungal disease was observed (Aegerter, 1997). Other researchers working on P. flammea populations in Scotland during the 1980s and 1990s did not record any fungal disease in the larvae until 1993 (Watt & Leather, 1988; Heritage et al., 1994, and unpublished data; D. Barbour, unpublished data). Whether the fungi have always been present in Scottish lodgepole pine forests at very low frequency or whether they have more recently dispersed to these forests is unknown. However, with the observation of new or different strains of fungi affecting P. flammea, we may be seeing a potential natural regulatory factor for this pest species in the future. It has been suggested that pathogens of high pathogenicity may be able to suppress insect populations below their economic thresholds and that even if suppression of the insect population by the pathogen is unstable, leading to cycles, these cycles may not cause damage at their peaks (Bowers et al., 1993). Intensive monitoring of the pest population, particularly over the duration of a full cycle, will determine whether these statements hold true.


Intensive monitoring of P. flammea populations in Scotland using pupal sampling and pheromone trapping suggested that the population behaviour resembles a cycle with periods of between 6 and 7 years. Although the amplitude of some of the cycles has been reduced with the application of chemical insecticides, some of the reductions were probably due to the action of natural processes. Several factors may be responsible for the observed population dynamics of P. flammea. These factors have been discussed by other authors (Watt & Leather, 1988; Evans et al., 1991). However, it is clear from the data presented here and elsewhere (Hicks & Watt, 2000) that our understanding of the population dynamics of this pest has changed in recent times. The impact of fungi in potentially regulating populations has recently been more noticeable than in the past. Future work should involve determining the impact of these fungal species on other natural enemies of the pine beauty moth. In conclusion, the use of selected fungal species should be investigated further as possible biological control agents against this destructive forest pest.


The authors would like to thank the Forestry Commission personnel who collected data on the abundance of eggs and pupae. This research was partially supported by a Rothermere Foundation postgraduate fellowship to B.H.