Light pollution and invasive alien species are both posing a threat to many ecosystems around the world. Understanding the effect both factors have on native ecosystems is important for preventing loss of native species. In the United Kingdom, street lighting is being changed from narrow to broad spectrum without much knowledge of how this will affect ecosystems. The horse chestnut leaf miner, Cameraria ohridella, is an invasive alien species posing a threat to the horse chestnut tree, Aesculus hippocastnum. This study looked at whether artificial light intensity affected the feeding activity of the horse chestnut leaf miner and whether artificial light intensity affected the growth of the horse chestnut tree. Leaflet samples and artificial light intensity data were gathered from a range of trees in different environments. ImageJ was used to measure aspects of the leaf, giving the propotion of leaf miner infection and length of the leaf. There was a moderate negative relationship between artificial light intensity and leaf miner feeding activity. There was no relationship between artificial light intensity and the size of the leaf. Cameraria ohridella appears to prefer lower artificial light intensities and the horse chestnut tree appears to not be affected by artificial light intensity.
1. Introduction
Pollution in all forms is presenting many problems for ecosystems all over the world. Light pollution is particularly problematic and a growing threat (Holzhauer et al., 2015). Organisms in all environments use light to time specific processes and the introduction of artificial light across most of the globe has begun to severely disrupt these (Gaston et al., 2013). Street lights have become more of an issue over the past century as they have started to be used all night, every night of the year (Gaston et al., 2013). Light pollution is predicted to increase at a rate of 6 per cent each year (Hölker et al., 2010). As their adaptations to the new level of artificial light are relatively new, little is known about how organisms are handling this sudden shift in their environment (Gaston et al., 2013).
The introduction of invasive alien species is another potential threat to ecosystems around the world. Although this human intervention has increased species richness in many places (Mooney & Cleland, 2001) the introductions have caused populations of native species to be hugely reduced or wiped out completely (Jeschke et al., 2014; Mooney & Cleland, 2001). Without complete knowledge of an invasive alien species, it makes it very difficult to prevent them from altering the native ecosystem (Jeschke et al., 2014). Some organisms such as viruses, pose more of a threat than others as they do not have to be established populations to affect a change on the receiving ecosystem (Jeschke et al., 2014). The increase in movement of humans and goods around the world has facilitated the rapidly increasing spread of invasive alien species (Mooney & Cleland, 2001).
The horse chestnut leaf miner, Cameraria ohridella, is a prime example of an invasive alien species. Its origin is unknown (Grabenweger et al., 2005) and was first discovered in Macedonia in 1985 (Deschka & Dimic, 1986). It was first recorded in the United Kingdom in 2002 (UK Forestry Commission, 2016). It infects the horse chestnut tree, Aesculus hippocastnum, causing leaf loss much earlier in its lifecycle than should happen (Grabenweger et al., 2005). A horse chestnut tree consists of leaves made up of five to seven leaflets. One adult moth will lay up to 40 eggs per leaflet and each leaflet could have up to 300 eggs on it from multiple adults (UK Forestry Commission, 2016). It takes about 4 weeks for their development of the leaf miner to be completed and during that time the larvae will eat away at the inside of the horse chestnut leaflet (UK Forestry Commission, 2016) creating mines. Although the leaf miner activity does not seem to be destructive to the horse chestnut tree, there is concern around how much of an effect the pest has on the tree’s reproduction and what this will mean for the tree in the long term (Percival et al., 2011). Despite itself not being native to the United Kingdom, the horse chestnut tree has become part of the rich biodiversity of the country. It is a staple across parks and recreational areas and is aesthetically important (Percival et al., 2011).
The UK Forestry Commission (2016) describes the basic life cycle of Cameraria ohridella. However, little more is known about this species. Its response to increased light pollution has certainly not been documented. Studies on other leaf miners and their response to light have been conducted but they report conflicting findings. Connor (2006) studied Cameraria hamadrydella, a leaf mining moth of white oak, Quercus alba, and discovered C. hamadrydella had no preference between light and shaded leaves. However, there was a greater number of mines on shaded leaves compared to leaves in the light, implying that another factor affects larval survival (Connor, 2006). Potter (1992) studied Phytomyza ilicicola, a leaf miner of American holly and found no difference between leaf miner abundance on shaded and unshaded leaves. Collinge and Louda (1988) investigated the activity of Scaptomyza nigrita, a leaf mining fly, on bitter cress and found that, despite the adult flies preferring plants in the sun, the abundance of mines was higher on leaves in artificial shade. Bultman and Faeth (1988) found that leaf miner activity in response to light varies between leaf miner species. C. ohridella is a diurnal species (Fischer et al., 2012) and so it would be expected to be drawn to trees in environments with higher light intensities.
Interestingly, light availability has been shown to affect the size of a leaf. Connor (2006) showed that white oak leaves in shade were around 30% larger than their counterparts in direct sunlight. This study also investigates whether there is a difference in leaf size between horse chestnut trees in continually light areas and trees in dark areas. If continual light does influence the growth of the horse chestnut tree in the same way as Connor (2006) discovered with white oak, the introduction of excessive street lighting may harm the population of horse chestnut trees that the United Kingdom has. A reduction in leaf size would reduce the overall productivity of the tree.
The present study looks at whether artificial light intensity affects the feeding activity of the horse chestnut leaf miner as well as the impact artificial light intensity can have on the general growth of a tree. By collecting leaflet samples from a selection of trees under a variety of light intensities, we were able to investigate our two hypotheses:
(1) feeding activity of the horse chestnut leaf miner, Cameraria ohridella, will increase with higher light intensities and
(2) leaflets will be of greater size under lower light intensities.
2. Methods
2.1 Sample Site
Before any sampling or data collection, a pilot search to find a range of horse chestnut trees under different artificial light intensities was implemented. No consideration was given to any other factors such as size or age. 30 trees were decided upon. All of the trees were in Bristol, United Kingdom and ranged in location from Durdham Downs down into Clifton Village. The distance between the furthest trees was 1.7 miles. Each tree was given a number and its location was recorded using Google MapsTM. The trees selected were across a range of night light, from completely unlit by street lighting to being surrounded by street lighting.
2.2 Leaflet Collection
In order to accurately document which trees the leaves were from, they were collected before collecting the light data. Over a period of seven days, visits to the 30 predetermined trees were made. A horse chestnut tree leaf consists of five to seven leaflets. To maintain consistency, the central leaflet was taken from each leaf and cut at the base of the leaflet using scissors. It was this leaflet that was kept for later analysis. At each tree, ten leaflets were collected. To decide which leaflets to take, it was determined which leaflet was lowest on the exterior of the tree. The nine remaining leaflets were then collected from around the initial leaflet within a 30cm diameter. Once the leaflets from one tree had been collected, they were put into a bag which was labelled with the tree number. A photograph was then taken standing under the tree at the point the leaves were collected from, holding the numbered bag.
The leaves were taken into the laboratory on the same day they were collected and photographed using a Canon EOS M3 camera (Canon, United Kingdom). This was to record the leaf in the way it was found in case any miners that were still alive continued to eat the leaf once it had been removed from the tree. They were photographed by mounting the camera on a tripod and keeping the camera settings consistent so all of the photographs would be of the same quality when it came to the analysis. The leaflets were placed on a white background with the tree number sticker next to it so they could be categorised correctly once all of the photographs were taken. There was also a 30cm ruler in shot to allow a scale bar to be used when it came to analysing the leaflets.
2.3 Light Collection
Once all of the leaflets were photographed, light collection began. Between October 23rd and November 14th, a four week period, light data was collected from the trees around Clifton Village on a Sunday and from the trees around Durdham Downs on a Monday. It was ensured that it was completely dark before any data collection began. Because of this, data from all 30 trees were unable to be collected in one night due to safety concerns. To prevent this having too much of an effect on the data, it was ensured data was collected on consecutive days. Data was collected using an ISO-TECH ILM-01 Light Meter (RS Components Ltd, Northants, United Kingdom) which measures in Lux. By collecting over four weeks, the moon cycle was able to be taken into account.
To determine the order in which light data would be collected, a random number generator was used, allowing the elimination of investigator bias. By using a random number generator, trees were visited at different times of night, preventing results being affected by reoccurrances such as traffic lights. Using the photo taken when the leaflets were collected, the light meter was positioned underneath the point of the tree the leaflets were collected from. The light meter was then held at arms length to the point at which the leaves were taken from. If the collection points were close to the ground, the light meter was still held below the branches so it could still be held out at arms length. Holding the light meter out at arms length prevented shadows being cast across the reader and affecting the reading. When the trees were by a road, data was collected when the road was clear from cars to prevent any extra light being cast on the reader and affecting the data. Once all of the light data was recorded, averages for each tree were calculated. The fourth set of data was discounted due to no leaves being left on the tree, increasing the amount of light that would reach the collection points. A super moon event also occurred on the last two days of collection, adding extra light that was not usually present.
2.4 Image Analysis
The imaging software ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA) was used to analyse the leaflet photographs that had been taken. A global scale bar was set for all leaflets using the ruler from the photograph. The length of the leaflet was measured from the base to the tip. The area of the leaflet was also measured by splitting the colour channels and using either red or blue, depending on what was best for that specific leaflet. Contrast levels were changed until the leaflet was all the same colour and then the particles were analysed, giving an accurate area of the leaflet. If there was damage to the leaflets, such as rips or holes, the area of the remaining leaflet was measured. Despite almost all leaflets being infected with the fungus Guignardia aesculi, the resulting leaf blotch from this infection was ignored. To measure the infected areas, the freehand tool was used to draw around the mined areas of the leaflet. The proportion of leaf miner infection per leaflet was then calculated. Once every leaflet had been analysed, the average proportion of leaf miner infection per tree was calculated. By using the tree as the unit of replication, pseudoreplication was not a factor.
3. Results
In total, a sample size of 300 leaves from 30 trees was collected and four nights of artificial light intensity collection around both Clifton Village and Durdham Downs were completed, making eight nights in total. The final set of data were discounted as all of the leaves had fallen from the trees and there was a supermoon event which could have interfered with the data. All 300 leaves were photographed on the day they were collected to guard against any extra leaf miner infection being included in the image analysis. The distribution of average artificial light intensity (Shapiro-Wilk; 0.452, d.f. = 30, p<0.001) and average proportion of leaf miner infection (Shapiro-Wilk; 0.711, d.p. = 30, p<0.001) were non-parametric. In order to make the data parametric, the average artificial light intensity was transformed using log10 (Shapiro-Wilk; 0.964, d.f. = 30, p = 0.384) and the average proportion of leaf miner infection was transformed using a fourth root transformation (Shapiro-Wilk; 0.973, d.f. = 30, p = 0.613). Figure 1 shows that the higher the artificial light intensity, the lower the infection of the tree (Pearson Correlation; -0.459, n = 30, p = 0.011).
The relationship between leaflet area and the artificial light intensity at a site was also examined. The average leaflet area was non-parametric (Shapiro-Wilk; 0.854, d.f. = 30, p = 0.001) and so was transformed using log10 (Shapiro-Wilk; 0.937, d.f. = 30, p = 0.077). Figure 2 shows there is no relationship between the artificial light intensity at a site and the area of the leaflet (Pearson Correlation; 0.083, n = 30, p = 0.664).
4. Discussion
4.1 Artificial Light Intensity and Feeding Activity
This study has shown that there is a relationship between artificial light intensity and the feeding activity of the horse chestnut leaf miner, Cameraria ohridella (Fig. 1). By gathering leaflets of horse chestnut trees from a variety of locations at varying artificial light intensities, it was demonstrated that there is a moderate negative correlation between the artificial light intensity and leaf miner activity (Fig. 1). This was a surprising result as it was expected that leaf miner feeding activity would increase with increased artificial light intensity. It is worth considering therefore, the possible reasons for this negative correlation.
A variant that could not be controlled when collecting artificial light intensity data was the wavelength of the light present. As technology has advanced, street lighting has moved from narrow spectrum lighting into broad spectrum lighting (Davies et al., 2013). This recent changeover has brought about changes in insect ecosystems that have not been investigated thoroughly (Davies et al., 2013). Throughout the sample site, there are four different types of street lamp used: ConstantColor™ CMH White Light StreetWise™ Lamps (GE Lighting, Europe), SON High Pressure Sodium Lamps (Philips, Europe), Philips GRN LED light source (Phillips, United Kingdom) and Urbis Axia Lanterns with LED neutral white light source (Urbis Schréder, United Kingdom). Broad spectrum lighting, such as High Pressure Sodium lamps, allow insects to see within more of their spectral range (Davies et al., 2013). If the wavelength of the lights present is different across the sample site, falling in or out of the adult moth’s spectral range, the adult female’s activity could be influenced. This in turn could partially explain some of the distribution seen in Fig. 1.
There is also variation between the broad spectrum lights that are used. With advances of technology, LED lighting is becoming more readily available (Pawson & Bader, 2014). Studies have shown that more insects are attracted to LED lights than the High Pressure Sodium lights that are being replaced (Pawson & Bader, 2014). Not all of the street lighting around the sample site had been replaced with LED lighting. Further research is required to determine whether wavelength and light type are factors when looking at the activity of horse chestnut leaf miners, Cameraria ohridella.
Another species in the same genus as the horse chestnut leaf miner, Cameraria hamadryadellam is known to lay its eggs just after dark (Connor, 2006). This species shows no discrimination between light and dark environments when laying its eggs, in line with its nocturnal lifestyle (Connor, 2006). Other species of leaf miner have also been shown not to have no preference of shaded and unshaded areas (Potter, 1992). However, C. ohridella does not follow this same pattern. There are suggestions based on the eyes of C. ohridella that this species follows a diurnal lifestyle (Fischer et al., 2012). Their diurnal lifestyle may mean that instead of searching for somewhere to lay their eggs based on the tree species, they are using other cues, such as photochemical (Johne et al., 2008), to determine the suitability of a leaflet. It is suggested that foliage quality may decrease with prolonged exposure to light, releasing altered chemical signals (Connor, 2006). The ability of C. ohridella to distinguish chemical cues may be part of the reason for the negative relationship between artificial light intensity and feeding activity (Fig.1).
Larval survival in Cameraria hamadryadellam is also seen to be affected by artificial light intensity with more larvae surviving in shade than in light (Connor, 2006). As previously stated, C. hamadryadellam show no preference between light and shaded trees when it comes to egg laying so another factor must be affecting larval survival such as chemical alterations of leaves in higher light intensities, reducing the quality of the leaf (Connor, 2006). Again, if Cameraria ohridella follows the same patterns, this could be an explanation for why there is a negative correlation between artificial light intensity and feeding activity.
Horse chestnut trees are also prone to infection by the fungus Guignardia aesculi, another invasive alien species to the UK introduced within the last century (Plenk, 1996). There has been little research on interactions between Cameraria ohridella and Guignardia aesculi. However, C. ohridella are known to be able to detect changes in the chemical make up of the horse chestnut leaflet post infection with the fungus which leads to a reduction in egg laying on that infected leaflet by the leaf miner (Johne et al., 2008). Infection of a leaflet with G. aesculi also causes the amount of area available to the leaf miner to reduce (Johne et al., 2008). Further studies investigating the influence of G. aesculi infection on C. ohridella behaviour need to be carried out.
There are also studies that have shown that the proportion of leaf miner larvae on a leaf can be affected by the presence of natural predators (Grabenweger et al., 2005). Only 3 species of bird and one species of cricket are known to be main predators of the leaf miner larvae (Grabenweger et al., 2005). All four species have a negative impact on the leaf miner population with birds predating up to 4% of the population (Grabenweger et al., 2005). Predation was not taken into account in the present study. This would need to be considered in the next study as different populations of horse chestnut tree may have different levels of predation, affecting the amount of feeding activity from the leaf miner.
4.2 Artificial Light Intensity and Leaflet Area
This study also suggests that there is no relationship between artificial light intensity and leaf area (Fig. 2). Other studies have found that leaves of other trees such as the white oak, Quercus alba, and bittercress plants grow larger and longer in shade (Connor, 2006; Collinge & Louda, 1988) but this study does not reflect that. The horse chestnut tree, Aesculus hippocastanum, may not follow the same growth patterns as these other species of tree but more focussed research into this needs to be carried out before any conclusions can be made.
4.3 Limitations
The different street light types present across the sample site is a cause for concern when evaluating this study. With such a variation in reaction to the different light types present around Durdham Downs and Clifton Village (Pawson & Bader, 2014), the possibility that this was a factor when the adult female was laying her eggs cannot be ignored. Further research looking at street light types and Cameraria ohridella feeding activity should be carried out to determine whether the change in street lighting across the country from High Pressure Sodium lamps to LED lights will affect the infection of horse chestnut trees. Although Cameraria ohridella causes no mortality of the horse chestnut tree, it is still potentially detrimental to the health of such a precious tree (Gilligan & Passoa, 2014).
The time of year that this study was carried out also restricted what data could be collected. If this study was repeated in the spring and summer months, data on where adult moths physically lay their eggs could be collected, allowing similar information to the study on Cameraria hamadryadellam by Connor (2006) to be collected. This would provide further information on whether the adult females have a preference between high and low artificial light intensities or whether higher artificial light intensities cause mortality among larvae.
Another advantage of collecting data earlier in the year and over a more prolonged period would be to allow analysis of multiple generations of C. ohridella. The leaf miner is known to have up to 5 overlapping generations in a year with the last generation able to remain as pupae for up to 7 months (UK Forestry Commission, 2016; Ferracini & Alba, 2008) and survive in temperatures as cold as -23oC (UK Forestry Commission, 2016). Monitoring a number of generations will give more insight into the life cycle of C. ohridella as well as strengthening the evidence of the relationship between artificial light intensity and feeding activity.
The amount of data collected also limited what statistical tests could be run. Although there are different street light types present in Clifton Village and on Durdham Downs, there was insufficient data to run comparison tests. There were only seven sets of data from Clifton Village compared to the 23 sets from Durdham Downs. If this study was to be repeated, even sets of data should be collected across a variety of street light types to allow for appropriate comparisons to be carried out. This study was limited in the amount of data which could be collected due to several time restraints. Frequent average light intensity data should be collected for as long as necessary until a consistent average is reached.
Data should also be collected across a variety of different environments. As Grabenweger et al. (2005) showed, the presence of natural predators can affect the leaf miner population by 4%. By collecting artificial light intensity and proportion data as well as recording the presence of predators across a variety of environments, it can be determined whether the presence of C. ohridella’s natural predators have an effect on their feeding activity.
Another restriction that was faced was being unable to collect all artificial light intensity data on one night. Because data collection was done on foot, a lot of time was wasted walking between sites. It was unsafe to be out collecting data past midnight and because of the distances between the trees, it was unfeasible to collect all data in one night. To counteract this, light data was collected on consecutive nights. However, it would have been ideal to collect all of the light data in one night and therefore allowing for more than one night of light data collection in a week. If this study is repeated, light data should be collected all on one night with at least two nights of light data collection per week. This would allow for a more precise average to be calculated.
4.4 Conclusions
The presence of Cameraria ohridella in and around the UK is a problem. Although the horse chestnut tree is currently only listed as Near Threatened on the IUCN Red List of Threatened Species (Khela, 2013), the impact of the C. ohridella infection could have a massive effect on the population (Khela, 2013). Their parasitism rate is surprisingly low as the adults emerge once the majority of parasitoids have died meaning they have no natural predators (Girardoz et al., 2006). With such little information on the life cycle of the leaf miner and its behaviour, it is hard to know how to control them and prevent them from becoming more detrimental. From this study we have shown that C. ohridella appear to have a preference for horse chestnut trees in dark environments than light environments. More work needs to be done in order to determine what causes them to hold this preference. Some reasons could be due to the change in chemical signalling from trees subjected to prolonged light, a different level of attraction to certain broad spectrum lighting or the presence/absence of natural predators in a specific area. By determining what causes them to feed more prolifically, their infection of the horse chestnut tree can be better understood. This would then facilite tackling the infection and preventing the demise of the horse chestnut tree.
2017-1-24-1485268499