Received: 4 May 2019 | Revised: 7 January 2020 | Accepted: 29 January 2020 DOI: 10.1002/ece3.6112 ORIGINAL RESEARCH A novel method to measure hairiness in bees and other insect pollinators Laura Roquer-Beni1 | Anselm Rodrigo1 | Xavier Arnan1 | Alexandra-Maria Klein2 | Felix Fornoff2 | Virginie Boreux2 | Jordi Bosch1 1 CREAF, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Catalunya, Abstract Spain Hairiness is a salient trait of insect pollinators that has been linked to thermoregu- 2 Nature Conservation and Landscape lation, pollen uptake and transportation, and pollination success. Despite its po- Ecology, University of Freiburg, Freiburg, Germany tential importance in pollination ecology, hairiness is rarely included in pollinator trait analyses. This is likely due to the lack of standardized and efficient methods to Correspondence Laura Roquer-Beni, CREAF, Edifici C, measure hairiness. We describe a novel methodology that uses a stereomicroscope Universitat Autònoma de Barcelona, equipped with a live measurement module software to quantitatively measure two Cerdanyola del Vallès, ES-08193 Catalunya, Spain. components of hairiness: hair density and hair length. We took measures of the two Email: l.roquer@creaf.uab.cat hairiness components in 109 insect pollinator species (including 52 bee species). Funding information We analyzed the relationship between hair density and length and between these BiodivERsA/FACCE-JPI, Grant/Award two components and body size. We combined hair density and length measures to Number: 2014-74; Ministerio de Economía y Competitividad, Grant/Award Number: calculate a hairiness index and tested whether hairiness differed between major PCIN-2014-145-C02 and RYC-2015-18448; pollinator groups and bee genera. Body size was strongly and positively correlated Federal Ministry of Education and Research, Grant/Award Number: 01LC1403; AGAUR, to hair length and weakly and negatively correlated to hair density. The correlation Grant/Award Number: 2018FI_B2_00152 between the two hairiness components was weak and negative. According to our hairiness index, butterflies and moths were the hairiest pollinator group, followed by bees, hoverflies, beetles, and other flies. Among bees, bumblebees (Bombus) and mason bees (Osmia) were the hairiest taxa, followed by digger bees (Anthophorinae), sand bees (Andrena), and sweat bees (Halictini). Our methodology provides an ef- fective and standardized measure of the two components of hairiness (hair density and length), thus allowing for a meaningful interpretation of hairiness. We provide a detailed protocol of our methodology, which we hope will contribute to improve our understanding of pollination effectiveness, thermal biology, and responses to climate change in insects. KEYWORDS functional diversity, functional trait, pilosity, pollinating efficiency, protocol, thermoregulation This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. Ecology and Evolution. 2020;10:2979–2990.  www.ecolevol.org | 2979 2980 | ROQUER-BENI et al. 1 | I NTRO D U C TI O N Notwithstanding the importance of hairiness in pollinator ecol- ogy, information on how to measure this trait is scarce and inconsis- Functional traits are morphological, physiological, or phenological tent across studies (Moretti et al., 2017). As a result, hairiness data characteristics measurable at the individual level, which are believed are mostly lacking in pollinator data bases and, when available, are to influence the fitness of an organism, to be a response of the organ- not comparable across studies. Some studies use thorax hair length ism to environmental changes or to reflect the effect of the organism as a measure of hairiness (Peat et al., 2005; Peters et al., 2016). on ecosystem function (Violle et al., 2007). A growing number of Others use the percentage of body surface covered by hair (Kühsel, studies are exploring the relationship between functional trait diver- 2015; Phillips et al., 2018). However, these measures do not account sity, environmental change and species composition, and emphasiz- for the two components of hairiness (hair length and hair density; ing the importance of functional diversity in ecosystem processes Moretti et al., 2017). Other studies do consider both components, (e.g., Arnan, Cerdá, Rodrigo, & Retana, 2013; Elmqvist et al., 2003; but use a semi-quantitative scale (Woodcock et al., 2019). Finally, Hooper et al., 2005; Petchey & Gaston, 2006; Suding et al., 2008). Stavert and collaborators (2016) proposed an innovative method Trait-based studies were mostly pioneered by plant ecologists, and that uses a measure of entropy obtained from images of the insect's extensive trait data bases are available for many plant taxa and com- body surface as a proxy for hairiness. However, we could not apply munities (e.g., Díaz et al., 2007; Lavorel & Garnier, 2002; McIntyre, this method to pollinators with shiny cuticles, which yielded high lev- Lavorel, Landsberg, & Forbes, 1999). Compared with plants, we els of entropy due to light reflection. know much less about animal functional diversity, especially of ter- The aim of our study was to develop a method to quantitatively restrial invertebrate communities (Moretti et al., 2017; Parr et al., measure hairiness in insect pollinators. We describe procedures to 2017). This is partly caused by the lack of standardized protocols measure hair density and hair length and propose a simple hairiness to measure functional traits in terrestrial invertebrates (Didham, index integrating both components. These procedures are then syn- Leather, & Basset, 2016; Moretti et al., 2017). Consequently, some thesized in a standardized protocol. We apply this protocol to three important traits are often not measured or measured in ways that different body parts of 109 insect pollinator species and show that are not comparable across studies, rendering databases insufficient, our methodology discriminates pollinator groups and bee genera in non-uniform, or taxa-limited. terms of hairiness. Finally, because body size is another functional Pollinators play a key role in the functioning of terrestrial eco- trait that has been related to both thermoregulation (Heinrich, systems and provide an essential ecosystem service in terms of crop 1993; Stone & Willmer, 1989) and pollination effectiveness (Jauker, pollination (Klein et al., 2007). However, several studies in Europe Speckmann, & Wolters, 2016; Kandori, 2002; Willmer & Finlayson, and North America have shown that pollinator diversity is declin- 2014), we explore the relationship between hair density, hair length, ing (Bartomeus & Winfree, 2013; Biesmeijer, 2006; Colla, Gadallah, and body size. Richardson, Wagner, & Gall, 2012; Powney et al., 2019). Within this context, functional traits are increasingly being incorporated in pol- linator studies. Various studies have established links between en- 2 | M ATE R I A L S A N D M E TH O DS vironmental changes and species susceptibility (Murray, Kuhlmann, & Potts, 2009; Roulston & Goodell, 2011) and between biodiversity 2.1 | Insect specimens and ecosystem functioning (Fontaine, Dajoz, Meriguet, & Loreau, 2006; Gagic et al., 2015). Commonly used functional traits in pol- We used a collection of pollinator insects from Sweden, Germany, linator ecology studies include body size, mouthpart length, social- and Spain composed of 109 species including Anthophila (bees; 52 ity, trophic specialization (lecty), voltinism, flight period, and nesting species), Syrphidae (hoverflies, 27), Bombyliidae (bee-flies, 2), other habits (e.g., Aguirre-Gutiérrez et al., 2016; Coutinho, Garibaldi, & flies (9), Coleoptera (beetles, 9), Lepidoptera (butterflies and moths, Viana, 2018; De Palma et al., 2015; Woodcock et al., 2019). 5), Vespidae (wasps, 3), and Symphyta (saw-flies, 2). Bees comprised One particularly important trait in pollinator insects is hairiness the following genera: Andrena (20 species), Bombus (10), Lasioglossum (pilosity). Hairiness creates an insulation layer that mitigates convec- (8), Halictus (3), Osmia (5), Apis (1), Anthophora (1), Eucera (2), Xylocopa tive loss of heat generated by the vibration of thoracic muscles, thus (1), and Nomada (1). For the analyses, Lasioglossum and Halictus spe- playing an essential role in thermoregulation (Heinrich, 1993; May, cies (Halictini) were grouped together, as well as Anthophora, Eucera, 1979). Some studies have found differences in hair length between and Xylocopa species (Anthophorinae) (Table S1). Because bees show bees from different climates (Peat, Darvill, Ellis, & Goulson, 2005) marked sexual dimorphism, we only worked with females. and along elevation gradients (Peters, Peisker, Steffan-Dewenter, & Hoiss, 2016), suggesting that hairiness could act as a response trait to climatic changes. Hairiness can also be considered an effect 2.2 | Body size trait involved in pollen collection and transfer (Amador et al., 2017; Müller, 1995; Thorp, 2000), potentially affecting pollination effec- We used body length as an estimator of body size (mean ± SE sam- tiveness (Phillips, Williams, Osborne, & Shaw, 2018; Stavert et al., ple size = 5.44 ± 0.39 specimens per pollinator species). For bees, in 2016; Woodcock et al., 2019). addition to body length, we measured intertegular span (hereafter ROQUER-BENI et al. | 2981 ITS) using a stereomicroscope (mean ± SE sample size = 6.62 ± 0.68 The time spent measuring hairiness (hair length + hair density in specimens per species). ITS is the most commonly used estimator of three 0.1 mm2 areas of the three body parts) was about 15 min per body mass in bee studies (Cane, 1987; Cariveau et al., 2016; Kendall specimen. A detailed protocol describing our method can be found et al., 2019; Osorio-Canadas et al., 2016). For this reason, we used in Appendix 1. ITS in the analyses involving only bees and body length in the analy- ses involving all pollinators. ITS and body length were highly corre- lated in bees (Spearman ρ = 0.87; p < .001). 2.4 | Data analysis All analyses were conducted in R v.3.3.2 (R Core Team, 2016), first 2.3 | Hairiness with all pollinator species, and then with bees only. The two components of hairiness (hair density and hair length) were measured in three body parts: the dorsal surface of the mesothorax, the 2.4.1 | Relationships between hair density, hair ventral surface of the thorax, and the face (Figure 1a-d). We selected length, and body size these body parts because the flight muscles involved in endogenous heat production are located in the thorax (Heinrich, 1993) and because For each body part separately, we tested whether hair density and the thorax and the face act as surfaces of pollen exchange in the polli- hair length were correlated. We also tested the correlation of each of nation of many flower species (Willmer, 2011). We measured a mean of these two hairiness components with body size. Because hair density three specimens per species (mean ± SE = 2.96 ± 0.11). Measurements and length were only weakly correlated (see results), we calculated, for were taken with the stereomicroscope LEICA M165C equipped with each body part, a hairiness index (hair density × hair length). Finally, we a LAS live measurement module software (Leica Microsystems). This examined whether the hairiness components and the hairiness index module allows taking length and surface measurements on live images of the three different body parts were correlated. We used either the in real units (Figure 1e,f). Pearson or Spearman correlation depending on data distribution. 2.3.1 | Hair density (number of hairs/mm2) 2.4.2 | Hairiness comparisons among pollinator groups and bee taxa In each of the three above-mentioned body parts, we selected 3 rep- resentative areas of approximately 0.1 mm2 and counted the number We calculated the coefficients of variation (SD/mean × 100) of the of hairs in each area. In some cases, notably in species with high hair two hairiness components within and between species, separately density and in specimens in which hairs formed clumps due to ma- for each body part. Because our method accounts for hair loss and nipulation during capture and/or preservation, it was easier to count since we only measured female specimens, we expected greater hairs at their insertion points, usually signaled by a micropore on the variability between than within species. cuticle (Figure 1e, see Appendix 1 for details). Counting micropores We explored whether hair density, hair length, and hairiness has the added advantage that can be applied to specimens that have index differed between pollinator groups and bee taxa using one-way lost hairs (e.g., due to aging; Bosch & Vicens, 2006; Southwick, 1985; ANOVA and Kruskal-Wallis tests (depending on data distribution), fol- or to poor manipulation) to obtain a measure of original hair cover. lowed by post hoc tests for multiple comparisons (Tukey's and Dunn's The results of these three measurements were used to calculate tests, respectively). We analyzed each hairiness component of each a mean hair density for each body part. In some species, hairiness body part separately. Log (X + 1) and square-root transformations patterns were clearly not uniform within a body part (notably in the were applied to improve normality and homoscedasticity of model re- face). In these cases (11.3% of the 327 species/body parts we mea- siduals if needed. Pollinator groups and bee taxa with three or fewer 2 sured), we sampled approximately 0.1 mm of the area occupied by species (bee-flies, saw-flies, wasps, Apis, Nomada) were excluded from each hairiness pattern separately, and the overall hair density mean the analyses but their values are provided in the figures. was weighted by the area occupied by each hairiness pattern. 3 | R E S U LT S 2.3.2 | Hair length (mm) 3.1 | Hairiness components: hair density and hair The length of 8–9 hairs of each body part was measured using the length length measuring tool of the software (Figure 1f). Again, in body parts with clearly distinct hairiness patterns, hair length of 8–9 hairs was Hair density of all pollinators ranged from 0 to 5,797.6 hairs/mm2 measured separately for each part and the overall mean hair length (mean ± SE = 428.0 ± 32.3) and that of bees from 63.5 to 1,052.2 was weighted by the surface occupied by each hairiness pattern. (mean ± SE = 333.4 ± 11.6). Hair length of all pollinators ranged from 2982 | ROQUER-BENI et al. F I G U R E 1 Pinned Andrena haemorrhoa female (a), body parts in which hairiness was measured (b: dorsal surface of the mesothorax, c: ventral surface of the mesothorax, d: face), and close-up images of the dorsal mesothorax showing measurements of hair density (e) and hair length (f) (a) (b) (c) (d) (e) (f) 0.01 to 1.91 mm (mean ± SE = 0.48 ± 0.02) and that of bees from pollinators: ρ = 0.47–0.67; bees: r = .61–.74). The hairiness index was 0.09 to 1.58 mm (mean ± SE = 0.62 ± 0.03). also strongly correlated across body parts (all pollinators: r = .78– Dorsal thorax hair density and length were weakly and negatively .88; bees: r = .77–.85). correlated (all pollinators: r = −.25, p < .01; bees: r = −.48; p < .001; Figure 2a,b). Ventral thorax and face hair density and length were also weakly and negatively correlated for bees (r = −.38 and −.39, 3.4 | Differences in hairiness across pollinator respectively; p < .01), but not for all pollinators (p > .08). groups and bee taxa Since both hairiness components and the hairiness index were cor- 3.2 | Relationship between hairiness related among body parts, hereafter, we only show results of the components and body size dorsal region of the thorax (the analysis of the ventral region of the thorax and the face yielded similar results; see Tables S4 and S5; Dorsal thorax hair density and body size were negatively correlated, Figures S3 and S4). weakly for all pollinators (r = −.20, p < .05, Figure 2c) and moderately The coefficient of variation of dorsal thorax hairiness compo- for bees (r = −.55, p < .001, Figure 2d). The analysis of ventral thorax nents was much higher between species (all pollinators: 73.0%– and face hairiness yielded similar results (Table S2). Conversely, hair 76.7%; bees: 53.4%–60.7%) than within species (all pollinators: length and body size were positively and strongly correlated in all 17.1%–18.7%; bees: 17.0%–18.3%; Table S4). three body parts (all pollinators, r = .67–.70, p < .001, Figure 2e; bees, r = .89–.93, p < .001, Figure 2f). 3.4.1 | All pollinators 3.3 | Hairiness comparisons among body parts We found clear differences among pollinator groups in the two hairiness components and the hairiness index (Figures 3 and 4a). Hairiness was positively correlated across body parts (all p < .001, Butterflies and moths were the group with the highest hair density, Table S3). Correlation coefficients were higher for hair length followed by bees, hoverflies, and beetles; other flies were the group (all pollinators: r = .91–.95; bees: r = .96) than for hair density (all with the lowest hair density (ANOVA, F4,97 = 6.9, p < .001, Figure 3a). ROQUER-BENI et al. | 2983 F I G U R E 2 Scatter plots showing the All pollinators (109 species) Bees (52 species) relationship between hair density and (a) (b) r = –.25 r = –.48 hair length (a and b), between hair density log (DT hair density + 1) log (DT hair density + 1) p < .01 p < .001 and body size (c and d) and between hair length and body size (e and f); of the dorsal surface of the mesothorax for all pollinators (a, c, and e; 109 species) and for bees only (b, d, and f; 52 species). Each point corresponds to one species. See Table S2 for results of the ventral surface of the mesothorax and the face. log (DT hair length + 1) log (DT hair length + 1) DT, dorsal surface of the mesothorax; ITS, intertegular span r = –.20 (c) r = –.55 (d) log (DT hair density + 1) log (DT hair density + 1) p < .05 p < .001 log (body length + 1) log (ITS + 1) (e) (f) r = .89 r = .67 p < .001 log (DT hair length + 1) log (DT hair length + 1) p < .001 log (body length + 1) log (ITS + 1) Antophorinae All pollinators Bees Hover Flies Andrena Beetles Bee Flies Bombus Apis Bees Butterflies and Moths Saw Flies Halictini Nomada Other Flies Wasps Osmia Butterflies and moths and bees had longer hair than any of the other 4 | D I S CU S S I O N pollinator groups (F4,97 = 11.3, p < .001, Figure 3c). According to the hairiness index, butterflies and moths were the hairiest group The aim of our study was to establish a standard practical procedure followed by bees, hoverflies, beetles, and other flies (F4,97 = 27.2, to quantitatively measure the two components of hairiness and to p < .001, Figure 3e). promote the use of this important trait in pollinator studies. 3.4.2 | Bees 4.1 | Advantages of the method Our measures of hairiness also yielded clear differences among The methodology we describe has several advantages. First, it pro- bee taxa (Figures 3 and 4b). Halictini had the highest hair density, vides a quantitative measure of hair density and length. Second, followed by Bombus, Osmia, and Andrena, and Anthophorinae had it is a non-invasive methodology; specimens remain undamaged. the lowest density (F4,45 = 16.25, p < .001, Figure 3b). This pattern Therefore, our methodology can be applied not only to dead speci- changed completely for hair length. Bombus and Anthophorinae had mens, but also to live (anesthesized) insects. Third, it can be applied the longest hair followed by Osmia and Andrena, and Halictini had to specimens in which the hair cover has been altered due to ma- the shortest hair (F4,45 = 67.9, p < .001, Figure 3d). According to the nipulation. Tufts of clumped hair are common in specimens that have hairiness index, the hairiest taxa were Bombus and Osmia, followed been exposed to high concentrations of ethyl acetate in killing jars by Anthophorinae, Andrena and Halictini (F4,45 = 13.5, p < .001, and in specimens that have been kept in water or alcohol (e.g. speci- Figure 3f). mens obtained from pan of malaise traps). Fourth, it accounts for 2984 | ROQUER-BENI et al. All pollinators (109 species) Bees (52 species) F I G U R E 3 Mean ± SE hair density (a and b), hair length (c and d), and hairiness a DT hair density (hairs/mm2 ) DT hair density (hairs/mm2 ) (a) (b) index (e and f) of the dorsal surface of a 900 the mesothorax (DT) of various pollinator 600 groups and bee taxa. Different letters bc b indicate significant differences among 600 b 400 b ab groups (post hoc Tukey's tests, p < .05). b c Groups with fewer than three species (in c 300 gray) were not included in the analyses 200 BE BT BM OF HF BF SF WA AN BO HA OS ANT AM NS a 1.5 (c) a (d) 1.5 DT hair length (mm) DT hair length (mm) b ab 1.0 a 1.0 c b b b 0.5 0.5 d 0.0 BE BT BM OF HF BF SF WA AN BO HA OS ANT AM NS a (e) (f) 500 a ab DT hairiness index DT hairiness index 1,000 400 300 bc 500 c 200 c b cd d c 100 0 BE BT BM OF HF BF SF WA AN BO HA OS ANT AM NS All pollinators Bees Bees (BE) Andrena (AN) Beetles (BT) Bombus (BO) Butterflies and Moths (BM) Halictini (HA) Other Flies (OF) Osmia (OS) Hover Flies (HF) Antophorinae (ANT) Bee Flies (BF) Apis (AM) Saw Flies (SF) Nomada (NS) Wasps (WA) hair loss due to aging or poor preservation. Certain pollinator groups, This amount of time can be reduced if, depending on the objectives notably bee-flies, tend to lose hairs during capture and manipula- of the study, fewer body parts are considered. tion. Thus, if needed, it is possible to obtain a measure of original hair density (as opposed to actual hair density). Fifth, measurements are taken directly from the specimen rather than from photographs. 4.2 | Relationship between hair density, hair For this reason, our method is not affected by shininess, a common length and body size feature of the cuticle of many pollinator insects. Sixth, our method discriminates the two components of hairiness, thus allowing for a We found a negative (albeit weak) correlation between hair length meaningful interpretation of the functional and evolutionary conse- and density for all three measured body parts in bees and for the quences of hairiness. dorsal region of the thorax in all pollinators. Accounting for the two Although our method may appear to be time-consuming, a components of hairiness would be redundant if these two variables trained person can process a specimen (8–9 measures of hair were highly correlated. Some groups such as beetles and Halictini length + 3 measures of hair density in 3 body parts) in just 15 min. had very short hair but very high hair density (Figure 3). Other ROQUER-BENI et al. | 2985 F I G U R E 4 Pollinator groups (a) and bee taxa (b) ordered by increasing Hairiness index hairiness from left to right. (Photograph credits Nicolas J. Vereecken [all bees], (a) Adrià Miralles [hoverflies] and Laura Roquer-Beni [other flies, beetles and All pollinators lepidopterans]. All images used with Beetles permission.) Other flies Bees Hoverflies Lepidopterans (b) Bees Andrena Antophorinae Osmia Bombus Halictini groups such as butterflies and moths and bees of the genera Bombus transport, and deliver pollen is likely to be influenced by the two and Osmia, had long hair and high hair density. hairiness components. Longer hairs provide a greater surface for Accounting for the two components of hairiness is also import- pollen grain adherence, and hair spacing (the inverse of hair den- ant because hair length and hair density may be differently related to sity) may be important in relation to pollen grain size (Haider, Dorn, body size. A positive relationship between hair length and body size Sedivy, & Müller, 2014; Roberts & Vallespir, 1978), which shows is expected due to allometric and mechanical constraints (movement great variability among plant taxa (Willmer, 2011). would be impaired in a small animal with long hair). Previous studies A link between hairiness and pollination effectiveness has been have found a positive relationship between hair length and body size found in some studies (Phillips et al., 2018; Stavert et al., 2016). at the intraspecific level in bumblebees (Goulson et al., 2002; Peat Given the positive correlation between body size and hair length, et al., 2005). Our results show that this relationship holds at the inter- studies exploring the relationship between hairiness and pollina- specific level and when pollinators from different orders are consid- tion effectiveness should account for body size, which, along with ered. Positive relationships between body size and length of various flower-handling behavior and visit duration, has also been shown to appendages are common in insects (proboscis: Cariveau et al., 2016; affect pollination effectiveness (Jauker et al., 2016; Kandori, 2002; Kunte, 2007; legs: Kaspari & Weiser, 1999; Teuscher, Brändle, Traxel, & Phillips et al., 2018; Willmer & Finlayson, 2014). Accordingly, in pol- Brandl, 2009 and wings: Bosch & Vicens, 2002; Bullock, 1999). lination studies, hairiness measures should target the body parts The relationship between hair density and body size, on the other involved in pollen transfer, which depend on flower morphology, pol- hand, is less straightforward. We cannot think of any a priori reason why linator body size and intra-floral foraging behavior (Araujo, Medina, hair density should differ between large and small animals. We found that & Gimenes, 2018; Beattie, Breedlove, & Ehrlich, 1973; Bosch, 1992; the relationship between hair density and body size was weak and neg- Solís-Montero & Vallejo-Marín, 2017). ative, especially in bees. Interestingly, studies on various groups of mam- mals have also found that small species tend to have denser (and shorter) fur (Sandel, 2013; Schwartz & Rosenblum, 1981; Steudel, Porter, & Sher, 4.4 | Hairiness as a response trait 1994). We suggest that the negative relationship between hair length and hair density, rather than indicative of a direct trade-off, can be explained Our methodology can also be important for studies on thermal biol- through the relationship between these two variables and body size. ogy and studies exploring the geographical distribution of pollinator Given that small animals cannot have long hair due to the above-men- communities and populations and their response to climate change. tioned mechanical constraints, the evolutionary pathway to achieve high Some pollinators generate heat endogenously by contracting their levels of hairiness in small animals is through increased hair density. flight muscles (Heinrich, 1993), and hairiness provides an insulation layer around the body surface that slows convective heat loss (May, 1979). As with pollination effectiveness, both components of hairiness 4.3 | Hairiness as an effect trait (length and density) are likely to contribute to the creation and mainte- nance of this insulation layer and therefore to influence thermoregula- Our methodology and our results have important implications for tion (Steudel et al., 1994; Wasserman & Nash, 1979). Consequently, we studies on pollination effectiveness. The ability to incorporate, would expect pollinator species and populations to be hairier in colder 2986 | ROQUER-BENI et al. climates. Again, given the correlation between hair length and body Stockholm) for providing pollinator specimens. This research is part size, studies addressing the distribution of pollinators in relation to cli- of the EcoFruit project funded through the 2013-2014 BiodivERsA/ mate should account for body size. Body size is strongly related to the FACCE-JPI joint call (agreement# BiodivERsA-FACCE 2014-74), with ability to generate heat and fly at low temperatures both at the intra- the Spanish Ministerio de Economía y Competitividad (MINECO, and interspecific levels (Bishop & Armbruster, 1999; Heinrich, 1993; project# PCIN-2014-145-C02) and German Federal Ministry of Osorio-Canadas et al., 2016; Stone, 1993; Stone & Willmer, 1989). Education and Research (PT-DLR/BMBF, grant# 01LC1403). LRB Both body size and hair length of bumblebees have been shown to be was funded by an AGAUR grant (FI grant# 2018FI_B2_00152) of greater in species from colder areas along latitudinal (Peat et al., 2005) Generalitat de Catalunya and by a SUPER-B COST project (Action# and elevational gradients (Peters et al., 2016). Since most endogenous FA1307). XA was funded by a Ramón y Cajal grant from the Spanish heat is produced by the flight muscles (Heinrich, 1993), measures of MINECO (grant# RYC-2015-18448). hairiness in thermal biology studies should mainly target the thorax, although other body parts (head, abdomen) have also been shown to C O N FL I C T O F I N T E R E S T be involved in heat loss (Cooper, 1985; Heinrich & Buchmann, 1986; None declared. Roberts & Harrison, 1998). AU T H O R S ' C O N T R I B U T I O N S LRB, JB, AR, and XA conceived the study; LRB conducted the 4.5 | Hairiness in trait-based studies measurements, analyzed the data, and led the preparation of the manuscript with the help from JB, AR, and XA. All authors provided Functional diversity studies typically characterize species based on insect specimens, contributed to the development of the methodol- suites of traits. Ideally, these traits should be biologically meaningful, ogy, and/or provided input during the preparation of the manuscript. easy to measure and comparable across taxa. In principle, and until All authors provided final approval for publication. we have a better understanding of the mechanistic effects of hair length and hair density of different body regions on various ecologi- DATA AVA I L A B I L I T Y S TAT E M E N T cal functions, we suggest keeping these two measures as separate The data generated in this study are available at Dryad Digital traits in a multitrait space. Otherwise, if a single measure is desirable, Repository: https​://doi.org/10.5061/dryad.3ffbg​79f0 they can be combined into a single trait (hairiness index). ORCID Laura Roquer-Beni https://orcid.org/0000-0001-8454-6745 4.6 | Concluding remarks Anselm Rodrigo https://orcid.org/0000-0001-6341-0363 Xavier Arnan https://orcid.org/0000-0002-9904-274X We have developed a standardized procedure to measure hairiness Alexandra-Maria Klein https://orcid.org/0000-0003-2139-8575 and explored the relationships between hairiness components and Felix Fornoff https://orcid.org/0000-0003-0446-7153 between hairiness and body size. Overall, these relationships were Virginie Boreux https://orcid.org/0000-0002-0450-6063 similar when analyzing only bees and when analyzing all pollina- Jordi Bosch https://orcid.org/0000-0002-8088-9457 tors. Importantly, in addition to insect pollinators, our methodology can be applied to other groups of terrestrial arthropods and can REFERENCES be used to explore the relationships between hairiness and other Aguirre-Gutiérrez, J., Kissling, W. D., Carvalheiro, L. G., WallisDeVries, ecological functions besides those discussed above. Hairiness M. F., Franzén, M., & Biesmeijer, J. C.. (2016). Functional traits help to explain half-century long shifts in pollinator distributions. Scientific has been shown to act as a physical and sensory barrier against Reports, 6(1), 1–13. https​://doi.org/10.1038/srep2​4 451 predators and parasites in caterpillars (Castellanos, Barbosa, Zuria, Amador, G. J., Matherne, M., Waller, D., Mathews, M., Gorb, S. N., & Hu, Tammaru, & Christman, 2011; Lindstedt, Lindström, & Mappes, D. L. (2017). Honey bee hairs and pollenkitt are essential for pollen 2008; Sugiura & Yamazaki, 2014) and moths (Shen, Neil, Robert, capture and removal. Bioinspiration and Biomimetics, 12(2), 026015. https​://doi.org/10.1088/1748-3190/aa5c6e Drinkwater, & Holderied, 2018). We hope our methodology will Araujo, L. S., Medina, A. M., & Gimenes, M. (2018). Pollination efficiency foster the inclusion of this important trait in insect data bases and on Ipomoea bahiensis (Convolvulaceae): Morphological and be- will contribute to our understanding of the importance of hairiness havioural aspects of floral visitors. Iheringia. Série Zoologia, 108, 1–5. in insect ecology. https​://doi.org/10.1590/1678-4766e​2018012 Arnan, X., Cerdá, X., Rodrigo, A., & Retana, J. (2013). Response of ant functional composition to fire. Ecography, 36(11), 1182–1192. https​ AC K N OW L E D G E M E N T S ://doi.org/10.1111/j.1600-0587.2013.00155.x We would like to thank Siyu Huang for her help with the hairiness Bartomeus, I., & Winfree, R. (2013). Pollinator declines: Reconciling measurements and D. García (Unidad Mixta de Investigación en scales and implications for ecosystem services. F1000Research, 2, Biodiversidad [CSIC-UO-PA], University of Oviedo), M. Miñarro 1–5. https​://doi.org/10.12688/​f1000​resea​rch.2-146.v1 Beattie, A. J., Breedlove, D. E., & Ehrlich, P. R. (1973). The ecology of the (Servicio Regional de Investigación y Desarrollo Agroalimentario, pollinators and predators of Frasera speciosa. Ecology, 54(1), 81–91. Asturias), P. A. Hambäck and U. Samnegård (University of https​://doi.org/10.2307/1934376 ROQUER-BENI et al. | 2987 Beck, J., & Kitching, I. J. (2007). Correlates of range size and dispersal of plant communities. PLoS Biology, 4(1), 0129–0135. https​://doi. ability: A comparative analysis of sphingid moths from the Indo- org/10.1371/journ​al.pbio.0040001 Australian tropics. Global Ecology and Biogeography, 16(3), 341–349. Gagic, V., Bartomeus, I., Jonsson, T., Taylor, A., Winqvist, C., Fischer, C., https​://doi.org/10.1111/j.1466-8238.2007.00289.x … Bommarco, R. (2015). Functional identity and diversity of animals Biesmeijer, J. C. (2006). Parallel declines in pollinators and insect-pol- predict ecosystem functioning better than species-based indices. linated plants in Britain and the Netherlands. Science, 313(5785), Proceedings of the Royal Society B: Biological Sciences, 282(1801), 351–354. https​://doi.org/10.1126/scien​ce.1127863 20142620. https​://doi.org/10.1098/rspb.2014.2620 Bishop, J. A., & Armbruster, W. (1999). Thermoregulatory abilities of Alaskan García-Barros, E. (2000). Body size, egg size, and their interspe- bees: Effects of size, phylogeny and ecology. Functional Ecology, 13, cific relationships with ecological and life history traits in but- 711–724. https​://doi.org/10.1046/j.1365-2435.1999.00351.x terflies (Lepidoptera: Papilionoidea, Hesperioidea). Biological Bosch, J. (1992). Floral biology and pollinators of three co-occurring Journal of the Linnean Society, 70(2), 251–284. https​://doi. Cistus species (Cistaceae). Botanical Journal of the Linnean Society, org/10.1111/j.1095-8312.2000.tb002​10.x 109(1), 39–55. https​://doi.org/10.1111/j.1095-8339.1992.tb002​57.x García-Barros, E. (2015). Multivariate indices as estimates of dry body Bosch, J., & Vicens, N. (2002). Body size as an estimator of production weight for comparative study of body size in Lepidoptera. Nota costs in a solitary bee. Ecological Entomology, 27(2), 129–137. https​:// Lepidopterologica, 38(1), 59–74. https​://doi.org/10.3897/nl.38.8957 doi.org/10.1046/j.1365-2311.2002.00406.x Goulson, D., Peat, J., Stout, J., Tucker, J., Darvill, B., Derwent, L. C., & Bosch, J., & Vicens, N. (2006). Relationship between body size, provision- Hughes, W. O. H. (2002). Can alloethism in workers of the bum- ing rate, longevity and reproductive success in females of the solitary blebee, Bombus terrestris, be explained in terms of foraging effi- bee Osmia cornuta. Behavioral Ecology and Sociobiology, 60(1), 26–33. ciency? Animal Behaviour, 64(1), 123–130. https​://doi.org/10.1006/ https​://doi.org/10.1007/s00265-005-0134-4 anbe.2002.3041 Bullock, S. (1999). Relationships among body size, wing size and mass Haider, M., Dorn, S., Sedivy, C., & Müller, A. (2014). Phylogeny and flo- in bees from a tropical dry forest in Mexico. Journal of the Kansas ral hosts of a predominantly pollen generalist group of mason bees Entomological Society, 72, 426–439. (Megachilidae: Osmiini). Biological Journal of the Linnean Society, Cane, J. H. (1987). Estimation of bee size using intertegular span 111(1), 78–91. https​://doi.org/10.1111/bij.12186​ (Apoidea). Journal of the Kansas Entomological Society, 60, 145–147. Heinrich, B. (1993). The hot-blooded insects: Mechanisms and evolution of Cariveau, D. P., Nayak, G. K., Bartomeus, I., Zientek, J., Ascher, J. S., Gibbs, thermoregulation. Cambridge, MA: Harvard University Press. J., & Winfree, R. (2016). The allometry of bee proboscis length and Heinrich, B., & Buchmann, S. (1986). Thermoregulatory physiology of its uses in ecology. PLoS ONE, 11(3), 1–13. https​://doi.org/10.1371/ the carpenter bee, Xylocopa varipuncta. Journal of Comparative journ​al.pone.0151482 Physiology B, 156(4), 557–562. https​://doi.org/10.1007/BF006​91042​ Castellanos, I., Barbosa, P., Zuria, I., Tammaru, T., & Christman, M. C. Hooper, D. U., Chapin, F. S., Ewel, J. J., Hector, A., Inchausti, P., Lavorel, (2011). Contact with caterpillar hairs triggers predator-specific de- S., … Wardle, D. A. (2005). Effects of biodiversity on ecosystem func- fensive responses. Behavioral Ecology, 22(5), 1020–1025. https​://doi. tioning: A consensus of current knowledge. Ecological Monographs, org/10.1093/behec​o/arr085 75(1), 3–35. https​://doi.org/10.1890/04-0922 Colla, S. R., Gadallah, F., Richardson, L., Wagner, D., & Gall, L. (2012). Jauker, F., Speckmann, M., & Wolters, V. (2016). Intra-specific body size Assessing declines of North American bumble bees (Bombus spp.) determines pollination effectiveness. Basic and Applied Ecology, 17(8), using museum specimens. Biodiversity and Conservation, 21(14), 714–719. https​://doi.org/10.1016/j.baae.2016.07.004 3585–3595. https​://doi.org/10.1007/s10531-012-0383-2 Kandori, I. (2002). Diverse visitors with various pollinator importance Cooper, P. D. (1985). Temperature regulation of honey bees (Apis mellif- and temporal change in the important pollinators of Geranium thun- era) foraging in the Sonoran desert. Journal of Experimental Biology, bergii (Geraniaceae). Ecological Research, 17, 283–294. https​://doi. 114, 1–15. Retrieved from http://jeb.biolo​gists.org/conte​nt/114/1/1. org/10.1046/j.1440-1703.2002.00488.x short​ Kaspari, M., & Weiser, M. D. (1999). The size-grain hypothesis and inter- Coutinho, J. G. D. E., Garibaldi, L. A., & Viana, B. F. (2018). The influence specific scaling in ants. Functional Ecology, 13(4), 530–538. https​:// of local and landscape scale on single response traits in bees: A me- doi.org/10.1046/j.1365-2435.1999.00343.x ta-analysis. Agriculture, Ecosystems & Environment, 256, 61–73. https​ Kendall, L. K., Rader, R., Gagic, V., Cariveau, D. P., Albrecht, M., Baldock, ://doi.org/10.1016/J.AGEE.2017.12.025 K. C. R., … Bartomeus, I. (2019). Pollinator size and its consequences: De Palma, A., Kuhlmann, M., Roberts, S. P. M., Potts, S. G., Börger, L., Robust estimates of body size in pollinating insects. Ecology and Hudson, L. N., … Purvis, A. (2015). Ecological traits affect the sen- Evolution, 9(4), 1702–1714. https​://doi.org/10.1002/ece3.4835 sitivity of bees to land-use pressures in European agricultural land- Klein, A.-M., Vaissière, B. E., Cane, J. H., Steffan-Dewenter, I., scapes. Journal of Applied Ecology, 52(6), 1567–1577. https​://doi. Cunningham, S. A., Kremen, C., & Tscharntke, T. (2007). Importance org/10.1111/1365-2664.12524​ of pollinators in changing landscapes for world crops. Proceedings of Díaz, S., Lavorel, S., de Bello, F., Quétier, F., Grigulis, K., & Robson, T. M. the Royal Society B: Biological Sciences, 274(1608), 303–313. https​:// (2007). Incorporating plant functional diversity effects in ecosystem doi.org/10.1098/rspb.2006.3721 service assessments. Proceedings of the National Academy of Sciences Kühsel, S. (2015). Pollinator trait diversity: functional implications at dif- of the United States of America, 104(52), 20684–20689. https​://doi. ferent land-use intensities and environmental conditions. PhD Thesis. org/10.1073/pnas.07047​16104​ Technische Universität. Didham, R. K., Leather, S. R., & Basset, Y. (2016). Circle the bandwag- Kunte, K. (2007). Allometry and functional constraints on proboscis ons—challenges mount against the theoretical foundations of applied lengths in butterflies. Functional Ecology, 21(5), 982–987. https​://doi. functional trait and ecosystem service research. Insect Conservation org/10.1111/j.1365-2435.2007.01299.x and Diversity, 9(1), 1–3. https​://doi.org/10.1111/icad.12150​ Lavorel, S., & Garnier, E. (2002). Predicting changes in community compo- Elmqvist, T., Folke, C., Nystrom, M., Peterson, G., Bengtsson, J., Walker, sition and ecosystem functioning from plant traits. Functional Ecology, B., & Norberg, J. (2003). Response diversity, ecosystem change, and 16(5), 545–556. https​://doi.org/10.1046/j.1365-2435.2002.00664.x resilience. Frontiers in Ecology and the Environment, 1(9), 488. https​:// Lindstedt, C., Lindström, L., & Mappes, J. (2008). Hairiness and warn- doi.org/10.2307/3868116 ing colours as components of antipredator defence: Additive or in- Fontaine, C., Dajoz, I., Meriguet, J., & Loreau, M. (2006). Functional di- teractive benefits? Animal Behaviour, 75(5), 1703–1713. https​://doi. versity of plant-pollinator interaction webs enhances the persistence org/10.1016/J.ANBEH​AV.2007.10.024 2988 | ROQUER-BENI et al. May, M. L. (1979). Insect thermoregulation. Annual Review of of Physical Anthropology, 152(1), 145–150. https​://doi.org/10.1002/ Entomology, 24(1), 313–349. https​://doi.org/10.1146/annur​ ajpa.22333​ ev.en.24.010179.001525 Schwartz, G. G., & Rosenblum, L. A. (1981). Allometry of primate hair McIntyre, S., Lavorel, S., Landsberg, J., Forbes, T. D. A. (1999). Disturbance density and the evolution of human hairlessness. American Journal response in vegetation—towards a global perspective on functional of Physical Anthropology, 55(1), 9–12. https​://doi.org/10.1002/ traits. Journal of Vegetation Science, 10(5), 621–630. https​://doi. ajpa.13305​50103​ org/10.2307/3237077 Shen, Z., Neil, T. R., Robert, D., Drinkwater, B. W., & Holderied, M. Miller, W. E. (1991). Body size in North American Lepidoptera as related W. (2018). Biomechanics of a moth scale at ultrasonic frequen- to geography. Journal of the Lepidopterists' Society, 45, 158–168. cies. Proceedings of the National Academy of Sciences of the United Moretti, M., Dias, A. T. C., de Bello, F., Altermatt, F., Chown, S. L., Azcárate, States of America, 115(48), 12200–12205. https​://doi.org/10.1073/ F. M., … Berg, M. P. (2017). Handbook of protocols for standardized pnas.18100​25115​ measurement of terrestrial invertebrate functional traits. Functional Solís-Montero, L., & Vallejo-Marín, M. (2017). Does the morphological Ecology, 31(3), 558–567. https​://doi.org/10.1111/1365-2435.12776​ fit between flowers and pollinators affect pollen deposition? An ex- Müller, A. (1995). Morphological Specializations in Central European Bees perimental test in a buzz-pollinated species with anther dimorphism. for the Uptake of Pollen from Flowers with Anthers Hidden in Narrow Ecology and Evolution, 7(8), 2706–2715. https​://doi.org/10.1002/ Corolla Tubes (Hymenoptera: Apoidea). Entomologia Generalis, 20(1– ece3.2897 2), 43–57. https​://doi.org/10.1127/entom.gen/20/1995/43 Southwick, E. E. (1985). Bee hair structure and the effect of hair on me- Murray, T. E., Kuhlmann, M., & Potts, S. G. (2009). Conservation ecol- tabolism at low temperature. Journal of Apicultural Research, 24(3), ogy of bees: Populations, species and communities. Apidologie, 40(3), 144–149. https​://doi.org/10.1080/00218​839.1985.11100664 211–236. https​://doi.org/10.1051/apido/​2009015 Stavert, J. R., Liñán-Cembrano, G., Beggs, J. R., Howlett, B. G., Pattemore, Nylin, S., Wiklund, C., Wickman, P.-O., & Garcia-Barros, E. (1993). D. E., & Bartomeus, I. (2016). Hairiness: The missing link between Absence of trade-offs between sexual size dimorphism and early pollinators and pollination. PeerJ, 4, e2779. https​://doi.org/10.7717/ male emergence in a butterfly. Ecology, 74(5), 1414–1427. https​://doi. peerj.2779 org/10.2307/1940071 Steudel, K., Porter, W. P., & Sher, D. (1994). The biophysics of Bergmann's Osorio-Canadas, S., Arnan, X., Rodrigo, A., Torné-Noguera, A., Molowny, rule: A comparison of the effects of pelage and body size variation R., & Bosch, J. (2016). Body size phenology in a regional bee fauna: on metabolic rate. Canadian Journal of Zoology, 72(1), 70–77. https​:// A temporal extension of Bergmann's rule. Ecology Letters, 19(12), doi.org/10.1139/z94-010 1395–1402. https​://doi.org/10.1111/ele.12687​ Stone, G. N. H. (1993). Endothermy in the solitary bee Anthophora Parr, C. L., Dunn, R. R., Sanders, N. J., Weiser, M. D., Photakis, M., Bishop, plumipes: Independent measures of thermoregulatory ability, costs T. R., … Gibb, H. (2017). GlobalAnts: A new database on the geogra- of warm-up and the role of body size. Journal of Experimental Biology, phy of ant traits (Hymenoptera: Formicidae). Insect Conservation and 174, 299–320. Diversity, 10(1), 5–20. https​://doi.org/10.1111/icad.12211​ Stone, G. N., & Willmer, P. G. (1989). Warm-up rates and body tempera- Peat, J., Darvill, B., Ellis, J., & Goulson, D. (2005). Effects of cli- tures in bees: The importance of body size, thermal regime and phy- mate on intra-and interspecific size variation in bum- logeny. Journal of Experimental Biology, 147(1), 303–328. Retrieved ble-bees. Functional Ecology, 19(19), 145–151. https​://doi. from http://jeb.biolo​gists.org/cgi/conte​nt/abstr​act/147/1/303 org/10.1111/j.0269-8463.2005.00946.x Suding, K. N., Lavorel, S., Chapin, F. S., Cornelissen, J. H. C., Díaz, S., Petchey, O. L., & Gaston, K. J. (2006). Functional diversity: Back to ba- Garnier, E., … Navas, M.-L. (2008). Scaling environmental change sics and looking forward. Ecology Letters, 9(6), 741–758. https​://doi. through the community-level: A trait-based response-and-effect org/10.1111/j.1461-0248.2006.00924.x framework for plants. Global Change Biology, 14(5), 1125–1140. https​ Peters, M. K., Peisker, J., Steffan-Dewenter, I., & Hoiss, B. (2016). ://doi.org/10.1111/j.1365-2486.2008.01557.x Morphological traits are linked to the cold performance and distri- Sugiura, S., & Yamazaki, K. (2014). Caterpillar hair as a physical barrier bution of bees along elevational gradients. Journal of Biogeography, against invertebrate predators. Behavioral Ecology, 25(4), 975–983. 43(10), 2040–2049. https​://doi.org/10.1111/jbi.12768​ https​://doi.org/10.1093/behec​o/aru080 Phillips, B. B., Williams, A., Osborne, J. L., & Shaw, R. F. (2018). Shared Teuscher, M., Brändle, M., Traxel, V., & Brandl, R. (2009). Allometry be- traits make flies and bees effective pollinators of oilseed rape tween leg and body length of insects: Lack of support for the size- (Brassica napus L.). Basic and Applied Ecology, 32, 66–76. https​://doi. grain hypothesis. Ecological Entomology, 34(6), 718–724. https​://doi. org/10.1016/j.baae.2018.06.004 org/10.1111/j.1365-2311.2009.01124.x Powney, G. D., Carvell, C., Edwards, M., Morris, RKA., Roy, H. E., Thorp, R. W. (2000). The collection of pollen by bees. Plant Systematics Woodcock, B. A., & Isaac, NJB. (2019). Widespread losses of polli- and Evolution, 222(1–4), 211–223. https​://doi.org/10.1007/BF009​ nating insects in Britain. Nature Communications, 10(1), https​://doi. 84103​ org/10.1038/s41467-019-08974-9 Violle, C., Navas, M. L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., R Core Team. (2016). R: A language and environment for statistical comput- & Garnier, E. (2007). Let the concept of trait be functional!. Oikos, ing. Retrieved from https​://www.r-proje​c t.org/ 116(5), 882–892. https​://doi.org/10.1111/j.2007.0030-1299.15559.x Roberts, R. B., & Vallespir, S. R. (1978). Specialization of hairs bearing Wasserman, D., & Nash, D. J. (1979). Variation in body size, hair pollen and oil on the legs of bees (Apoidea: Hymenoptera)1. Annals length, and hair density in the deer mouse Peromyscus manicula- of the Entomological Society of America, 71(4), 619–627. https​://doi. tus along an altitudinal gradient on JSTOR. Holarctic Ecology, 2(2), org/10.1093/aesa/71.4.619 115–118. Retrieved from https​://www.jstor.org/stabl​e/36826​ Roberts, S. P., & Harrison, J. F. (1998). Mechanisms of thermoregula- 65?read-now=1&seq=2#page_scan_tab_contents tion in flying bees. American Zoologist, 38(3), 492–502. https​://doi. Willmer, P. (2011). Pollination and floral ecology. Princeton, NJ: Princeton org/10.1093/icb/38.3.492 University Press. Roulston, T. H., & Goodell, K. (2011). The role of resources and risks in Willmer, P. G., & Finlayson, K. (2014). Big bees do a better job: Intraspecific regulating wild bee populations. Annual Review of Entomology, 56(1), size variation influences pollination effectiveness. Journal of Pollination 293–312. https​://doi.org/10.1146/annur​ev-ento-120709-144802 Ecology, 14, https​://doi.org/10.26786/​1920-7603(2014)22 Sandel, A. A. (2013). Brief communication: Hair density and body mass in Woodcock, B. A., Garratt, M. P. D., Powney, G. D., Shaw, R. F., Osborne, mammals and the evolution of human hairlessness. American Journal J. L., Soroka, J., … Pywell, R. F. (2019). Meta-analysis reveals that ROQUER-BENI et al. | 2989 pollinator functional diversity and abundance enhance crop polli- nation and yield. Nature Communications, 10(1), 1481. https​://doi. How to cite this article: Roquer-Beni L, Rodrigo A, Arnan X, org/10.1038/s41467-019-09393-6 et al. A novel method to measure hairiness in bees and other insect pollinators. Ecol Evol. 2020;10:2979–2990. S U P P O R T I N G I N FO R M AT I O N https​://doi.org/10.1002/ece3.6112 Additional supporting information may be found online in the Supporting Information section. APPENDIX 1 Pre-treatment of the specimens The method can only be applied to dry specimens. Apart from this, P R OTO C O L TO M E A S U R E H A I R I N E S S I N B E E S A N D no special pre-treatment is necessary. The method works well even OT H E R I N S EC T P O L L I N ATO R S with specimens that have lost hair due to aging or poor manipulation, This protocol follows the structure of the “Handbook of protocols as well as with specimens with clumps of hair (e.g., specimens initially for standardized measurement of terrestrial invertebrate functional kept in alcohol that have not been properly dried). The method can traits” (Moretti et al., 2017). We focus on insect pollinators, but our also be applied to anesthesized live specimens. methodology can also be applied to other groups of arthropods. Body parts Definition and relevance Hairiness is best measured in body parts with flat surfaces, but the Insect hairiness (pilosity) is the collective presence of hairs, scales, method can be applied to any body part. Hairiness may strongly dif- seta or bristles growing from the cuticle (Moretti et al., 2017). fer among body parts. The target body part(s) should be decided Hairiness creates an insulation layer that mitigates the convective based on the objectives of the study. A study on pollination effec- loss of heat generated by the vibration of thoracic muscles, thus tiveness should target body parts directly involved with the uptake playing an essential role in thermoregulation (Heinrich, 1993; May, and transfer of pollen. A study on thermoregulation should empha- 1979). Some studies have found a negative relationship between bee size body parts involved in heat generation (thorax) and dissipation hair length and ambient temperature (Peat, Darvill, Ellis, & Goulson, (abdomen, head, appendages). 2005; Peters, Peisker, Steffan-Dewenter, & Hoiss, 2016), suggest- We found hairiness to be correlated across the face, the dorsal ing that hairiness acts as a response trait to climate. Hairiness can surface of the thorax and the ventral surface of the thorax. also be considered an effect trait involved in pollen collection and transfer (Amador et al., 2017; Müller, 1995; Thorp, 2000), potentially Measurement of hairiness components affecting pollination effectiveness (Phillips, Williams, Osborne, & Hairiness can be decomposed in hair density and hair length (Moretti Shaw, 2018; Stavert et al., 2016; Woodcock et al., 2019). Hairiness et al., 2017). may also be involved in antipredator strategies. In caterpillars and moths, hairs have been shown to provide mechanical and sensory Hair density defence against predators and parasites (Castellanos, Barbosa, Zuria, Counting all hairs in an entire body part is unpractical. We found that Tammaru, & Christman, 2011; Lindstedt, Lindström, & Mappes, averaging hair counts of three areas of ca. 0.1 mm2 each provides a 2008; Shen, Neil, Robert, Drinkwater, & Holderied, 2018; Sugiura good measure of hair density for a given body part. The area sam- & Yamazaki, 2014). pled (and if needed the area of the entire body part) can be measured with the appropriate tool in the software. Technical support Counting hairs can be complicated when hair density is high and Measurements should be taken with a stereomicroscope with a when hairs are long and form clumps due to manipulation during magnification range of 20–80× and a camera connected to a com- capture and/or preservation. In these cases, it is easier to count puter equipped with a software module that allows taking length the micropores of the cuticle in which hairs are inserted (Figure and surface measurements directly on live images in real units. We A1). Because insect cuticles may display a large variety of micro- used a LEICA M165C and the live measurement module from Leica sculpture patterns, including several types of punctuations, it is Microsystems. Alternatively, other stereomicroscopes and software very important to spend some time to identify the correct type of packages could be used (e.g., free software ImageJ). Measurements micropores before starting the counts. Counting micropores has can also be taken from microscope pictures (instead of live images) the added advantage that it provides a measure of original hairi- but this is a more time-consuming alternative. If real units are not ness even in specimens that have lost hair due to aging or poor provided by the software, then the level of magnification must be manipulation (e.g., bee-flies typically lose a lot of hair during ma- accounted for. nipulation). In some cases, it may be practical to rub the cuticle of 2990 | ROQUER-BENI et al. F I G U R E A 1 Hair micropores in the dorsal thorax region of Scaeva albomaculata (left) and of Andrena haemorrhoa (right) Hairiness index We found hair length and hair density to be only weakly (and nega- tively) correlated. For this reason, we recommend reporting meas- ures of the two components of hairiness (hair density and hair length) separately. However, in studies in which the two compo- nents are suspected to have a similar effect on function, we propose a hairiness index combining the two hairiness components (hair den- sity × hair length). Body size We found that body size was positively correlated to hair length and, to a lesser extent, negatively correlated to hair density. Body size is strongly related to heat generation and dissipation and the ability to fly at low temperatures (e.g., Bishop & Armbruster, 1999; Heinrich, 1993; Osorio-Canadas et al., 2016; Peat et al., 2005; Peters F I G U R E A 2 Measurement of hair length on the dorsal thorax et al., 2016; Stone & Willmer, 1989), and may influence pollination region of Xylocopa violacea effectiveness (Jauker, Speckmann, & Wolters, 2016; Kandori, 2002; Phillips et al., 2018; Willmer & Finlayson, 2014). For these reasons, the insect with an insect pin to detach hairs, thus facilitating mi- we recommend accounting for body size in studies measuring hairi- cropore counts. If instead of original hairiness a measure of actual ness. Appropriate measures of body size for bees include intertegu- hairiness is desirable, then only micropores with standing hairs lar span (interspecific level, Cane, 1987), head width and forewing should be counted. length (intraspecific level, Bosch & Vicens, 2002). Forewing length Sometimes hairiness patterns are distinctly non-uniform across a and wingspan are appropriate measures for butterflies (Beck & body part, notably in the face. In these cases, the area occupied by Kitching, 2007; García-Barros, 2000, 2015; Miller, 1991; Nylin, each hairiness pattern can be sampled separately. The percentage of Wiklund, Wickman, & Garcia-Barros, 1993). When comparing spe- surface occupied by each different density should be reported and the cies from different insect orders, body length is probably the most overall mean of hair density for that body part should be weighted by suitable measure. the area occupied by each hairiness pattern. Other considerations Hair length We found that measuring three specimens per species allowed us The length of a hair can be measured using the length measuring to discriminate between pollinator groups and bee genera. Sample tool of the software (Figure A2). Hair length is best measured from sizes should be increased in studies addressing intra-specific a side view. We recommend measuring 5–10 hairs in each focal variability. body part. Because many pollinator species (notably bees) show marked sex- ual dimorphism, males and females should be measured separately. Duration of the measurements Measures of hairiness on the thorax dorsal region and body size Following the above-mentioned recommendations, a trained person of 109 pollinator species from Spain, Germany and Sweden are pro- can measure hair density and hair length of a body part in 5 min. vided in Table S1.