We describe a semi-quantitative approach of measuring characteristics of corticolous (bark-dwelling) arthropod communities. We placed commercially manufactured sticky traps on tree boles to estimate abundance, total length (a surrogate to biomass), richness, and Shannon diversity for comparison among tree species.
Terrestrial arthropods play an important role in our environment. Quantifying arthropods in a way that allows for a precise index or estimate of density requires a method with high detection probability and a consistent sampling area. We used manufactured sticky traps to compare abundance, total length (a surrogate for biomass), richness, and Shannon diversity of corticolous arthropods among the boles of 5 tree species. Efficacy of this method was adequate to detect variation in corticolous arthropods among tree species and provide a standard error of the mean that was <20% of the mean for all estimates with sample sizes from 7 to 15 individual trees of each species. Our results indicate, even with these moderate sample sizes, the level of precision of arthropod community metrics produced with this approach is adequate to address most ecological questions regarding temporal and spatial variation in corticolous arthropods. Results from this method differ from other quantitative approaches such as chemical knockdown, visual inspection, and funnel traps in that they provide an indication of corticolous arthropod activity over a relatively long-term, better including temporary bole residents, flying arthropods that temporarily land on the tree bole and crawling arthropods that use the tree bole as a travel route from the ground to higher forest foliage. Furthermore, we believe that commercially manufactured sticky traps provide more precise estimates and are logistically simpler than the previously described method of directly applying a sticky material to tree bark or applying a sticky material to tape or other type of backing and applying that to the tree bark.
Terrestrial arthropods play an important role in our environment. In addition to being of scientific interest in their own right, arthropods can be both detrimental and beneficial to other trophic levels (i.e., crops, horticultural plants, native vegetation, and food for insectivorous organisms1,2,3,4). Thus, understanding the factors that influence arthropod community development and abundance is critical to farmers5, pest control managers6, foresters4, plant biologists7, entomologists8, and wildlife and conservation ecologists that study community dynamics and manage insectivorous organisms9. Arthropod communities vary in species composition and abundance both temporally and spatially across a variety of ecological landscapes including plant communities, plant species, and across various regions of individual plants. For example, studies have demonstrated significant differences in arthropod community metrics between the roots, bole and stems, and foliage, within the same individual tree10,11. These findings are not surprising considering that different parts of the same plant, e.g., leaves versus barks of a tree, provide different resources for which arthropods have adapted to exploit. Thus, each part of the plant can support a different arthropod community. Because foliage dwelling arthropods can have such a large socioeconomic and environmental impact, substantial effort has been expended to measure community metrics using both qualitative and quantitative approaches12. Alternatively, much less effort has been expended to develop approaches of quantifying corticolous (bark-dwelling) arthropod communities.
Like foliage-dwelling arthropod communities, corticolous arthropod communities can be important from both a socioeconomic and environmental viewpoint. Some forest diseases that are caused or facilitated by corticolous arthropods can be detrimental to economically viable timber harvest4. Additionally, corticolous arthropods can be an important component of the food chain in forest communities13,14. For example, forest dwelling arthropods are the primary food source for many insectivorous bark gleaning song birds15,16. Thus, understanding the factors that influence communities of corticolous arthropods is of interest to foresters and both basic and applied ecologists.
Understanding factors that influence arthropod community composition and abundance often requires the capture of individuals. Capture techniques can generally be categorized into qualitative techniques that only detect presence of a species for estimates of species range, richness, and diversity17, or semi-quantitative and quantitative techniques that allow for an index or estimate of abundance and density of individuals within a taxonomic group18,19. Semi-quantitative and quantitative techniques allow researchers to estimate or at least consistently sample a specified sample area and estimate probability of detection or assume detection probability is non-directional and adequate as to not obscure the researcher's ability to detect spatial or temporal variation in abundance. Semi-quantitative and quantitative techniques for quantifying corticolous arthropods include suction or vacuum sampling of a specific area20,21,22, systematic counting of visible arthropods18,23, sticky traps24, various funnel or pot-type traps8,25, and entrance or emergent holes26,27.
A number of spatial and temporal factors are thought to lead to variation in corticolous arthropod communities11,14,28,29. For example, texture of tree bark is thought to influence the community structure of tree-dwelling arthropods14. Because of the more diverse surface area of the trunks of trees with more furrowed bark, trees with more furrowed bark are thought to support a greater diversity and abundance of arthropods14.
With this article we report a new semi-quantitative approach of enumerating corticolous arthropods that could be used to describe and test hypotheses regarding variation in corticolous arthropod communities across time and space with adequate precision to detect differences among tree species. Using sticky traps attached to the trunks of trees, we compared the abundance, total length (a surrogate for body mass), richness, and diversity of the arthropod community on the bole of white oak (Quercus alba), pignut hickory (Carya glabra), sugar maple (Acer saccharum), American beech (Fagus grandifolia), and tulip poplar (Liriodendron tulipifera) trees, trees that vary in bark texture.
This study was conducted in the Ozark and Shawnee Hills ecological sections of the Shawnee National Forest (SNF) in southwestern Illinois. During July 2015, we identified 18 (9 dominated by oak/hickory and 9 dominated by beech/maple) sites with the USFS stand cover map for the SNF (allveg2008.shp) in ArcGIS 10.1.1. In the xeric sites, the dominant species were pignut hickory and white oak and in mesic sites, the dominant species were American beech, sugar maple, and tulip poplar. To compare bole arthropod community among tree species, at each data collection site, we identified the three of the five (white oak, pignut hickory, sugar maple, American beech and tulip poplar) focal species trees >17 cm diameter at breast height (d.b.h.) closest to the center of a 10 m radial circle. If fewer than three appropriate trees were present, the circle was expanded and the closest tree fitting the criteria was selected. For each tree chosen, we installed four sticky traps at breast height, one facing in each cardinal direction: north, south, east and west.
We collected arthropod data from the boles of 54 individual trees (12 pignut hickories, 15 white oaks, 8 American beeches, 12 sugar maples, and 7 tulip poplars) among the 18 sites. We grouped arthropods according to a simplified guild classification by diagnostic morphological characteristics indicative of closely related orders from current phylogenetic records, similar to that of "operational taxonomic units"30,31 (Appendix A). Based on this classification, we captured representatives of 26 guilds in our traps that were each in place for 9 days (Appendix A). Because our study focused on trophic interactions between tree species, corticolous arthropods, and bark-gleaning birds, we removed all arthropods smaller than 3 mm from analysis because their importance as a food resource is minimal for bark-gleaning birds. We used a mixed model that included either arthropod length (surrogate to body mass), abundance, Shannon diversity and, richness as the dependent variable, tree species and effort (proportion of tree covered with traps) as fixed variables, and site as a random variable. Because all traps from a single tree were combined as one sample, individual trees were not included as a random variable.
1. Placement of a trap on the tree
2. Removing the trap from the tree
3. Laboratory analysis
Based on the mixed model results, the model that included tree species best explained variation in total arthropod length, abundance, and diversity, neither of independent variables explained substantial variation in richness, although the models that included tree species trapping effort were competitive with the null model (Table 1). In addition, proportion of the tree trapped appears to have no influence on abundance, total length, and Shannon diversity, with only minimal influence on richness (Table 1). The standard error of the mean (SEM) for total arthropod length varied from 4% of the mean in tulip poplar to 17% in sugar maple (Table 2). Abundance had similar levels of variation within species where the SEM was 7% of the mean in tulip poplar and 18% in sugar maple (Table 2). Conversely, variability in arthropod richness and diversity was much lower within species of tree in that SEM of richness ranged from 4% of the mean for pignut hickory to 9% of the mean in American beech, while diversity ranged from 4% of the mean in American beech to 7% of the mean in tulip poplar.
Dependent variable | Model | K | AIC | ΔAIC |
Richness | Null | 2 | 210.56 | 0 |
Tree species | 7 | 211.69 | 1.13 | |
Effort | 3 | 211.93 | 1.37 | |
Total body length | Tree species | 7 | 719.69 | 0 |
Null | 2 | 727.00 | 7.31 | |
Effort | 3 | 728.96 | 9.27 | |
Abundance | Tree species | 7 | 495.55 | 0 |
Null | 2 | 501.04 | 5.48 | |
Effort | 3 | 503.04 | 7.48 | |
Diversity | Tree species | 7 | 28.78 | 0 |
Null | 2 | 37.31 | 8.52 | |
Effort | 3 | 38.72 | 9.93 |
Table 1: Model results. Results of a mixed model analysis of covariance (ANCOVA) with corticolous arthropod richness, total body length, abundance, or Shannon diversity as the dependent variable, tree species and proportion of tree covered by traps (effort) as the independent fixed variables, and individual site as the independent random variable. K = number of model parameters, AIC = estimated Akaike's Information Criterion, and ΔAIC = the difference in AIC points form the model to the most parsimonious model.
Tree species | Richness | Total length | Shannon diversity | Abundance | ||||||||
X | SE | % of mean |
X | SE | % of mean |
X | SE | % of mean |
X | SE | % of mean |
|
Sugar maple (N = 12) | 8.33 | 0.59 | 7% | 365.20 | 63.69 | 17% | 1.59 | 0.09 | 6% | 45.45 | 8.15 | 18% |
Pignut Hickory (N = 12) | 7.83 | 0.30 | 4% | 573.90 | 81.58 | 14% | 1.24 | 0.07 | 6% | 70.09 | 10.10 | 14% |
Tulip Poplar (N = 7) | 8.75 | 0.49 | 6% | 195.35 | 7.09 | 4% | 1.73 | 0.12 | 7% | 25.67 | 1.87 | 7% |
American Beach (N = 8) | 8.29 | 0.81 | 9% | 349.91 | 38.45 | 11% | 1.53 | 0.06 | 4% | 47.00 | 5.32 | 11% |
White Oak (N = 15) | 9.07 | 0.42 | 4% | 407.38 | 40.16 | 10% | 1.64 | 0.09 | 5% | 50.57 | 5.26 | 10% |
Table 2: Parameter estimates from the most parsimonious model in Table 1. The mean (X), SEM, and percentage of SEM for each community metric of corticolous arthropods captured on 5 species of trees using commercially manufactured sticky traps in the Shawnee National Forest in southern Illinois.
Although alternative techniques such as suction or sweep nets have been used, most previously published attempts at quantifying arthropods on tree boles used some version of either quantifying arthropods by visually inspecting tree boles in the field, using chemical pesticides to kill arthropods in a specified area then quantifying the recovered arthropods, or placing funnel traps or a sticky substance directly onto the tree19,23,25,35,36. Each of these approaches have benefits and shortcomings.
With chemical knockdown, a pesticide is sprayed over a predefined area and arthropods are allowed to drop onto a drop cloth as they die, where they are then collected and quantified19. Alternatively, with visual location, live arthropods are located in the predefined area and collected by hand for later quantification23. Both of these methods are instantaneous relative to our method, thus provide a more quantifiable estimate of area sampled for use in estimating density. A further attribute to chemical knockdown as well as visual inspection is, because it is somewhat instantaneous, the estimate is limited to the time at which the survey was conducted. Because it only samples arthropods present at the time of sampling, this method provides an accurate estimate of the size of area sampled, facilitating an estimate of density. These approaches, however, often disregard variation in the non-resident arthropod population, arthropods that temporarily inhabit the tree boles such as flying arthropods or arthropods that use the surface of tree boles as travel routes from the ground to higher forest foliage. Because many of the arthropods that influence other trophic levels use bark for short periods as part-time residence, nearly instantaneous samples from the visual observation and chemical knockdown method likely will not adequately depict the entire suit of arthropods that use tree bark as a substrate8,35,36.
To better depict the corticolous arthropod community that occurs over longer periods, longer-term methods such as funnel and sticky traps have been developed25,26,27,28,29,30,31,35,36. Funnel traps are attached to tree boles and are designed to funnel arthropods into bottles of preservative, thus are beneficial in that they can be used for long periods of time (weeks to potentially months) while still preserving the arthropods. The limitation of these traps is their limited ability to trap flying arthropods that land on the tree boles. Alternatively, sticky traps are effective at capturing both crawling and flying arthropods.
With the original sticky traps, a sticky material was placed directly onto the tree to trap both crawling and flying arthropods over a predetermined time37. While this approach was effective at trapping both crawling and flying arthropods, it is difficult to spread the exact same amount of material for each trap, thus maintain a consistent sampling area and trapped arthropods have to be identified and quantified in the field under often less than ideal weather conditions, potentially leading to additional variation in the estimates due to misidentification or miscounting. An improvement was offered by Collins et al.36 when they spread the sticky material on tape, then, after trapping for a predetermined amount of time, covered the tape with cellophane and removed the tape so arthropod identification and quantification could be conducted later in the laboratory, where conditions were much more appropriate for the activity. While this method is an improvement over the previously described methods, it is still messy, and still difficult to consistently spread the same amount of sticky material at each trap. As an improvement to this method, we propose using commercially manufactured sticky traps to address both of these deficiencies.
Commercially produced sticky traps have been used to trap flying arthropods over water38, at various elevations of vascular vegetation39, and in the foliage of trees40, but to our knowledge have not been used to sample arthropods on tree bark. Commercially produced sticky traps provide an improvement over previously used approaches in that the sticky material is adhered to the cardboard backing in the factory and because they are commercially manufactured, the surface area of the material is very consistent. Additionally, the traps can be placed on the trees with the trap intact, preventing flying arthropods from landing directly on the trap, as was done in our study, or the cardboard cover could be removed so the trap is catching both crawling arthropods and flying arthropods landing directly on the trap. Additionally, the traps are easily removed from the tree, covered with cellophane and transported to the laboratory where they can be stored in a freezer and quantified at a later date. The trap's stiff cardboard construction also facilitates viewing the traps in the laboratory under a dissecting microscope allowing for more precise identification, quantification and measurements of arthropods, reducing some of the detection error that would likely occur when conducting this activity in the field. Finally, sticky material on trees can become saturated, reducing the ability of the trap to capture arthropods41. The method we describe allows researchers to easily replace the sticky traps to maintain effectiveness, allowing long-term monitoring of individual trees.
As demonstrated by our results, this approach appears to provide adequate precision to address most ecological or environmental questions regarding variation in corticolous arthropod communities. Detection of arthropods from sticky traps used to quantify corticolous arthropods with this method was adequately precise to provide an SEM that was <20% of the mean for all community metrics used in this study. This level of precision was achieved with a reasonable sample sizes of only 7 to 15 individual trees. With this level of precision and moderate sample sizes, we detected differences in total length (a surrogate for biomass), total abundance, total richness, and Shannon diversity among species of trees. We did not partition the variance between measurement error (variance associated with variation in proportion of area trapped among traps or variation in detection probability) and variance among individual trees within a tree species, however, these results clearly indicate that this method has adequate detection probability to prevent measurement error from obscuring results to important ecological or environmental questions.
We describe this method as being semi-quantitative because although we believe our detection probability to be high and provide adequate precision to address most ecological questions, we have no way of estimating detection probability. Thus, we have no way of estimating potential negative bias associated with our point estimates. Additionally, a fully quantitative method that could be used to estimate overall abundance or density, requires an accurate estimate of the sampling area42. Unlike the visual inspection or chemical knockdown methods, the sampling area with funnel traps and with this method is uncertain because it is not instantaneous, the traps are placed on the tree for a predetermined amount of time and arthropods going about their normal activities are trapped when they cross the surface of the sticky traps. Thus, the size of area being trapped is dependent on the activity level of the arthropods. Arthropod activity level varies with time of day, by season, by species, or by individual8. Because arthropod activity level varies, the sampling area will vary based on activity level. It will be important for researchers to consider how activity level influences inference from results when using both this and the funnel trap method. We argue, however, neither methods that provide a more accurate estimate of the sampling area because they are more instantaneous nor methods that provides a less accurate estimate of sampling area but a better depiction of the arthropod community over time is better. Instead, the two types of methods address different questions. The chemical knockdown and visual inspection methods describe the community during a very specific point in time, while the funnel and sticky trap methods describe the community over a period of hours or days, depending on how long the traps are left in place. We believe, however, when researchers are interested in identifying and describing spatial and temporal variation of corticolous arthropod communities utilizing the bark surface over a substantial time (days to weeks), the method described here is the most convenient and accurate approach.
Finally, the primary objective of our original study was to better understand how mesophication of southeastern deciduous forests is likely to impact forest dwelling insectivorous birds and mammals, thus, we combined arthropods into guilds43. We see no reason, however, why these capture techniques could not be used to quantify arthropods at the species or any other taxonomic level.
The authors have nothing to disclose.
The authors would like to thank the U.S. Department of Agriculture Forest Service for funding this project through USFS Agreement 13-CS-11090800-022. Support for ECZ was provided by NSF-DBI-1263050. ECZ assisted in the development of the research concept, collected all field data, conducted laboratory analysis, and produced the original manuscript. MWE assisted in the development of the research concept and study design, assisted in directing field data collection and laboratory analysis, and heavily edited the manuscript. KPS assisted with study design, directed the field and laboratory work, assisted with data analysis, and reviewed the manuscript.
Straight Draw Bark Shaver, 8" | Timber Tuff | TMB-08DS | |
PRO SERIES Bulk Mouse & Insect Glue Boards | Catchmaster | #60m | |
Staple gun | Stanley | TR45D |