Seasonal Occurrence of Adult Carabid Beetles (Coleoptera: Carabidae) in West-Central Illinois

Carabid beetles (Coleoptera: Carabidae), including ground beetles and tiger beetles, comprise a highly diverse family with ∼40,000 described species (Lövei and Sunderland 1996). Most species are generalist predators (Larochelle and Larivière 2003), although phytophagy (Honek et al. 2003) and fungivory (Johnson and Cameron 1969) are also well represented in the family. Many carabids have close associations with specific environments and are important as bioindicators, and the family includes important biological control agents of agricultural pests (McCravy 2018, Willand et al. 2011). In addition, carabids have become model organisms for field studies examining functional traits (characteristics that contribute to short-term performance or long-term adaptation) such as trait reassembly during prairie restoration (Barber et al. 2017). Despite a large amount of research on carabids of the midwestern United States (McCravy and Lundgren 2011), relatively few studies have focused on temporal activity patterns.

Information on seasonal patterns of carabid species diversity is important for several reasons. Knowledge of temporal patterns can facilitate optimization of sampling periods for studies of carabid ecology and conservation (Werner and Raffa 2003), inform analyses of carabid co-occurrence with pest species for biological control purposes (Suenaga and Hamamura 2001), and provide information on potential impacts of exotic species on native carabids with which they may overlap temporally (Niemelä et al. 1997). Potential impacts of activities such as pesticide applications and prescribed burning may also be evaluated in relation to carabid activity. Carabids are an important food resource for many vertebrates (Larochelle 1975a, 1975b, 1980) which could be adversely affected by changes in carabid seasonal activity patterns. Carabid behaviors such as diel activity patterns can also vary with season (Willand and McCravy 2006). Furthermore, studies of carabid seasonal activity patterns can provide important information on potential effects of anthropogenic climate change. Evidence suggests that there have been major changes in the composition of the British carabid fauna associated with climate change (Coope 1978, 1987) and that carabids will likely respond to climate change via distributional shifts rather than physiological adaptation (Butterfield 1996). Knowledge of carabid temporal activity patterns could be useful in predicting interactions within new carabid assemblages.

Like much of the midwestern United States, the Illinois landscape has undergone dramatic changes due to cultivation. Over 80% of Illinois has been converted to cropland or pasture over the past 160 yr (Iverson 1988). The resulting landscape is dominated by an agricultural matrix with habitat fragments. Carabids are a focal group from both conservation and applied biological control standpoints in the Midwest (McCravy and Lundgren 2011). In this study, we examined seasonal activity of adult carabids in west-central Illinois in a habitat matrix of forest and reconstructed tallgrass prairie. Our objectives were to characterize the seasonal occurrence of carabids in west-central Illinois and to compare carabid species composition and diversity across months.

Carabids were collected in pitfall traps during May–October 2005 and 2006 at Western Illinois University's Alice L. Kibbe Life Science Research Station (N 40°21′52″, W 91°24′19″) in Hancock Co., IL. Each pitfall trap consisted of two 473-ml plastic cups (Dart Container Corp., Mason, MI), one nested inside the other, and placed in the ground with the rim flush with the ground surface. Traps were filled with approximately 150 ml of a 50/50 mixture of propylene glycol and water. A total of 162 pitfall traps were set in 18 3 × 3 arrays, encompassing an area of approximately 60 ha of reconstructed tallgrass prairie, oak–hickory forest, and early successional forest, with a variety of prescribed burning regimes. Traps within arrays were 5 m apart, and each array was 40–75 m from the nearest array. Traps were operated for seven consecutive days each month (08–15 May, 04–11 June, 09–16 July, 06–13 August, 10–17 September, and 08–15 October 2005; 12–19 May, 08–15 June, 13–20 July, 10–17 August, 16–23 September, and 09–16 October 2006) (see Willand et al. [2011] for further site and trapping details). Carabids were identified using Lindroth (1961–1969) and a synoptic reference collection.

Spearman correlation analyses were conducted to test for associations of monthly carabid abundance between years and species richness between years, using SigmaPlot, Version 13. Pitfall trap collections are a function of carabid activity levels as well as density, and so are best viewed in terms of “activity-density” (Maelfait and Baert 1975, Thiele 1977). Herein, “abundance” is used in this sense.

Shannon indices (H) (Shannon 1948) were calculated for all months combined and for each month, and compared between months using the t test developed by Hutcheson (1970). Holm's step-down procedure (Holm 1979) was used to adjust for paired comparisons by comparing the nth smallest P value with 0.05/number of comparisons + 1 – n. The Shannon index is a diversity measure that quantifies the uncertainty or entropy of a community based on the number of species (species richness) and the relative abundance of the species (evenness). Greater Shannon values indicate greater diversity. Evenness values were obtained by dividing the Shannon value by the natural logarithm of species richness. Evenness can range from 0 to 1, with greater values representing greater evenness. Effective number of species (ENS) for each month and for all months combined was derived using exp(H). H and other entropy measures are nonlinear, making interpretation problematic. ENS is the number of equally abundant species necessary to produce the observed Shannon value, eliminating nonlinearity and allowing more meaningful comparisons of effect size (Jost 2007).

One-factor permutational multivariate analysis of variance (PerMANOVA; Anderson 2001) was used to compare species composition (species present and their relative abundances) among months. PerMANOVA is a multivariate analysis of variance that provides a test for differences in mean within-group distances among groups, using a permutation-based F test. For PerMANOVA, species with fewer than three nonzero cells were excluded from analysis to reduce sparsity, and values for each species were relativized by the maximum value to reduce heterogeneity and effects of dominant species (Peck 2016). PerMANOVA was conducted using PC-Ord, Version 7. Spearman correlations, using SigmaPlot Version 13, were conducted to compare species composition between years, both within months and overall.

A total of 2,712 carabids were collected overall, representing 27 genera and 54 species (Table 1). In 2005, 2,073 carabids were collected, representing 25 genera and 47 species; in 2006, 639 carabids were collected, representing 22 genera and 32 species. Abundance and species richness of carabids were consistently greater in 2005 than in 2006, and were greater in May–July than in August–October. Monthly abundance was significantly correlated between years (ρ = 1.000, P = 0.00278, n = 6), but richness was not (ρ = 0.771, P = 0.103, n = 6).

Shannon index values and ENS were similar between years overall, but varied substantially between years among months (Table 2). In August and October Shannon values were significantly higher in 2005, but the 2006 value was highest in June. When yearly data were pooled, Shannon index values were lower in July and August than in the remaining months, as were ENS and evenness (Table 2). There was no significant correlation of monthly values between years for Shannon values or for ENS (ρ = 0.143, P = 0.803, n = 6, each comparison). There was also no significant correlation of monthly ENS and species richness (ρ =0.029, P =0.919, n = 6).

PerMANOVA revealed significant differences in carabid species composition among months (F = 2.1867; df = 5, 6; P = 0.0044). Variation in species composition among months accounted for 37.238% of total variation. All pairwise comparisons between months were nonsignificant (P ≥ 0.3268, each comparison). Species composition was significantly correlated between years overall (ρ = 0.573, P < 0.0001, n = 54), and for May and June (May: ρ = 0.492, P = 0.0110, n = 26; June: ρ = 0.414, P = 0.00912, n = 39), but not for the remaining months (July: ρ = 0.305, P = 0.145, n = 24; August: ρ = 0.116, P = 0.638, n = 18; September: ρ = 0.315, P = 0.198, n = 18; October: ρ = 0.0446, P = 0.861, n = 17).

Of the 17 species for which at least 40 individuals were collected (Figs. 1–4), all but two (Cicindela sexguttata F. and Notiophilus novemstriatus LeConte) were most abundant in 2005 (Table 1). Cyclotrachelus seximpressus (LeConte) and Synuchus impunctatus (Say) were collected exclusively in 2005. The two most abundant species, Cyclotrachelus sodalis (LeConte) and Chlaenius platyderus Chaudoir, comprised 54.3% of total captures, with C. sodalis numbers peaking in July and C. platyderus in June (Fig. 1).

Most species showed greatest abundance in spring or early summer (Figs. 1 – 4). Platynus decentis (Say), Amara cupreolata Putzeys, and Trichotichnus fulgens (Csiki) showed greatest overall numbers in May, declining thereafter (Figs. 1, 3, 4). The tiger beetle Cicindela sexguttata was most common in spring, especially June, and absent from fall collections, as was Galerita janus (F.) (Figs. 1, 2). Poecilus lucublandus (Say), Cyclotrachelus seximpressus, Chlaenius emarginatus Say, and Dicaelus elongatus Bonelli were most abundant in June (Figs. 3, 4), and Pterostichus stygicus (Say) was most abundant in June and September samples (Fig. 2).

Among later season species, almost all of the tiger beetle Megacephala virginica L. were collected in July (Fig. 2). Most Pterostichus permundus (Say) and N. novemstriatus were collected in September and October, with relatively small numbers collected in the spring (Figs. 2 and 4). Most S. impunctatus were collected in September (Fig. 4).

Numbers of carabids collected varied greatly between years. This is not unusual; pitfall trap collections of carabids can vary by more than an order of magnitude in different years (Jones 1979). Even though carabid numbers were substantially lower in 2006 than in 2005, the second year of sampling yielded seven previously undetected species overall, and from one to eight previously undetected species each month. This suggests that 2 yr of sampling provided substantial benefits in documenting the carabid fauna of Kibbe Station. Likewise, monthly sampling proved to be important since species composition varied significantly among months, and many carabid species showed strongly seasonal activity patterns at this site. Relative abundance and species richness among months were consistent between years overall, with both of these measures highest in May–July and lower in August–October. June sampling yielded the greatest abundance and species richness; about two-thirds and three-fourths of total species richness were collected during this month in 2005 and 2006, respectively. Other studies of carabids in the U.S. Upper Midwest (Epstein and Kulman 1990, Werner and Raffa 2003), Alberta, Canada (Niemelä et al. 1992), and northwestern Arkansas Ozark Highlands (Hamilton et al. 2018) have also obtained greatest trap catches early in the season. These seasonal patterns suggest that trapping throughout the growing season is necessary to sufficiently sample carabid diversity at this location, but for limited monitoring early season sampling is best. The relatively low numbers collected in August suggest low activity at that time. Temperate carabid activity is often low in summer, when some species may enter aestivation (Lövei and Sunderland 1996).

Shannon index values and corresponding evenness and ENS values also showed a seasonal pattern with lower values in July and August. This pattern appears to have been driven by the dominance of C. sodalis, which accounted for about two-thirds of all carabid captures during these months. Shannon and ENS values were relatively high in October, when species richness was low, but evenness was high. This may in part be a result of relatively low sample sizes and consequent lower limits on variation in relative species abundances. ENS values did not correlate well with species richness, again probably due to the dominance/evenness factors described above. Monthly Shannon (and associated ENS) values also did not correlate significantly between years, probably due in large part to the much lower 2006 values for August and October. These low values corresponded to very low species richness (six and five species, respectively) for these 2 mo.

Species composition varied significantly among months, further underscoring the importance of sampling throughout the activity season. Species composition was consistent between years overall, but only for May and June among particular months. Relatively low sample sizes in the later months could be a reason. It is also possible that carabid communities increase in entropy as the season progresses, and that this occurs at variable rates in different years.

Among individual species, seasonal activity patterns generally agreed with those reported in the literature. It should be noted that, because our sampling was restricted to a 1-week period near the middle of each month, our data represent discrete samples rather than continuous samples as has been done in most other studies. Therefore, comparisons with other studies should be done with caution. In southeastern Minnesota, Epstein and Kulman (1990) found C. sodalis to be most common from late June through July, and again from mid-August to mid-September, when their sampling ended. In our study, C. sodalis numbers peaked in July samples, with few individuals collected in September and October. Epstein and Kulman (1990) suggest that this species, as well as S. impunctatus, overwinter as larvae. They collected S. impunctatus adults in greatest numbers in July and August, with declining numbers in September. Werner and Raffa (2003) found a similar pattern in northeastern Wisconsin and the Upper Peninsula of Michigan, although they did not sample in September. In our study, S. impunctatus was only collected in 2005, with 23 of 32 specimens collected in September, perhaps due to the longer growing season in west-central Illinois.

Cicindela sexguttata was collected predominantly in June in our study. This species also has a seasonal peak in June in South Dakota and Nebraska (Carter 1989, Spomer et al. 2008). Cicindela sexguttata is unusual among North American Cicindelinae in that adults are observed almost exclusively in spring/early summer, and rarely in fall (Knisley and Schultz 1997). This species has a 2-yr life cycle, but adults remain in the pupal burrows, emerging the following spring (Knisley and Schultz 1997, Shelford 1908). Platynus decentis was collected primarily in May and June in our study, with similar patterns found in southeastern Minnesota (Epstein and Kulman 1990), northeastern Wisconsin and the Upper Peninsula of Michigan (Werner and Raffa 2003), and Alberta, Canada (Niemelä et al. 1992). This is a forest species that overwinters as adults in forests and forest edges, under embedded stones, in leaf litter, and in tree stumps (Larochelle 1972, Larochelle and Larivière 2003). Gravid females of this species have been found in Illinois in mid-May (Gilbert 1957). Amara cupreolata and Poecilus lucublandus also overwinter as adults (Epstein and Kulman 1990, Larochelle and Larivière 2003). We collected A. cupreolata exclusively in May–July, with greatest numbers in May. Epstein and Kulman (1990) collected this species primarily throughout May and to a lesser extent in June. Poecilus lucublandus had distinct spring and late-season activity periods in southeastern Minnesota, with none captured from early July to mid-August (Epstein and Kulman 1990). We collected relatively large numbers of this species in our mid-July 2005 sampling, with numbers declining in later samples. It is possible that a period of activity could have occurred between our July and August samplings. We collected C. seximpressus in 2005 only. Allen and Thompson (1977) found a bimodal activity pattern for this species in northwestern Arkansas woodlands, with relatively large numbers in June and July, and then again in September and October. In our study, 38 of 48 C. seximpressus were collected in June. The remainder were distributed throughout the later samples, with no apparent peak.

The relatively large numbers of Pterostichus stygicus collected in June and September samples in our study suggests a bimodal seasonal pattern, particularly during 2005 when sample size was greatest. Gilbert (1957) suggested that, in northeastern Illinois (Cook Co.), females of this species either oviposited in the current spring, or in the previous year and then overwintered even though they had already reached sexual maturity. Collections of P. permundus likewise were suggestive of a bimodal pattern in our study, although only 6% of specimens were collected in June. Allen and Thompson (1977) found a bimodal activity pattern for this species in woodlands of northwestern Arkansas, but with greatest numbers in June. Megacephala virginica is a summer tiger beetle, with peak abundance in August according to Knisley and Schultz (1997). In Nebraska, Carter (1989) recorded greatest numbers in August and especially September. In our study, almost all M. virginica were collected in July.

Our results show that west-central Illinois carabid communities vary temporally, with abundance and species richness greatest early in the season. Species composition was fairly consistent across years. Abundance and species richness differed greatly between years, but relative numbers among months were similar between years. Carabid species showed distinct seasonal patterns of occurrence, but for the most part were similar to patterns found for these species in other midwestern U.S. studies.

The authors thank Seán Jenkins (Western Illinois University), Ben Wodika (Illinois State University), and Jamie Palmer (Purdue University) for assistance with field work, and Robert Davidson (Carnegie Museum of Natural History) for identification of unknown carabids. We thank Dean Corgait (Illinois Department of Natural Resources) for assistance with the funding application process. Generous financial support of this research was provided by the Illinois Wildlife Preservation Fund (Illinois Department of Natural Resources, Contract RC06L21W) and the Western Illinois University Graduate Research Council.