Insectes soc. 46 (1999) 58–65 0020-1812/99/010058-08 $ 1.50+0.20/0 © Birkhäuser Verlag, Basel, 1999 Insectes Sociaux Research article Colony-level and season-specific variation in cuticular hydrocarbon profiles of individual workers in the ant Formica truncorum J. Nielsen 1, J.J. Boomsma 1, *, N.J. Oldham 2, 3, H.C. Petersen 4 and E.D. Morgan 2 1 2 3 4 Department of Ecology and Genetics, University of Aarhus, Ny Munkegade, DK-8000 Aarhus C, Denmark, e-mail: Koos.Boomsma@biology.aau.dk Department of Chemistry, Keele University, ST5 5BG Keele, Staffordshire, England Present address: Max-Planck Institut für Chemische Ökologie, Tatzendpromenade 1 a, D-07745 Jena, Germany ADBOU, SDU Odense University, University of Odense, Campusvej 55, DK-5230 Odense M, Denmark Received 14 August 1997; revised 20 February 1998; accepted 16 April 1998. Summary. Cuticular hydrocarbon profiles of individual workers of the ant Formica truncorum were measured and found to contain relatively few hydrocarbons. Pentacosane, heptacosane, nonacosane and hentriacontane dominated the mixture, but small amounts of the corresponding alkenes were also present. Principal component analysis and nested analysis of variance showed that workers from different colonies varied significantly in quantitative aspects of their cuticular hydrocarbon profiles. Furthermore, differences between habitat-patches within populations and (to some extent) between populations were also detected. Finally, workers from the same colony, sampled only a few months apart, were found to be different in the quantitative composition of their cuticular hydrocarbon profiles, emphasising the importance of collecting samples from a colony at a single point in time. Key words: Cuticular hydrocarbons, ants, gas chromatography, mass spectrometry. Introduction The surface of the exoskeleton of insects is covered with a lipid layer which at least partly consists of hydrocarbons. The lipid layer provides protection against desiccation, but the prime function of the hydrocarbons has remained unclear (Lockey, 1988). It has been hypothesised that hydrocarbons keep the cuticular lipid layer slightly fluid, so that the insects can more easily cope with variable temperatures. However, * Author for correspondence. this does not explain why insects apparently regulate their cuticular hydrocarbon content and why they change their cuticular hydrocarbon mixture throughout the developmental stages (Blomquist et al. 1987). In social insects, part of the elaborate variation in cuticular hydrocarbons has been hypothesised to be connected to their secondary role as recognition substances (Howard and Blomquist, 1982; Blomquist et al., 1987; Howard, 1993), mediating communication and behavioural processes (Espelie et al., 1991; Jaisson, 1991; Butts et al., 1995). The postpharyngeal glands of ants have been shown to contain essentially the same hydrocarbon mixture as in the cuticular lipid layer (Bagnères and Morgan, 1991). A recent study on Cataglyphis niger has shown that cuticular hydrocarbons may also originate from cells associated with the epidermis and are transported from there to the cuticular lipid layer and to the postpharyngeal gland (Soroker et al., 1995). Ants are known to transfer cuticular hydrocarbons between nestmates by physical contact, grooming and trophallaxis (Vienne et al., 1995), so that nestmates, in spite of their genetic differences, acquire a similar hydrocarbon profile or “gestalt” (Crozier, 1987), which can be used for discriminating between nestmates and non-nestmates. Nestmate recognition is widespread in social insects and has been documented in bees (Breed, 1983; Breed et al., 1988; Breed and Julian, 1992), wasps (Ryan et al., 1985; Gamboa et al., 1986; Venkataraman et al., 1988; Singer and Espelie, 1992; Lorenzi et al., 1997; Ruther et al. 1998) and ants (Bonavita-Cougourdan et al., 1987; Stuart, 1987; Morel and Blum, 1988; Peeters, 1988; Crosland, 1990; Dahbi and Lenoir, 1998). A number of recent studies have also made explicit connections between social insect behaviour and hydrocarbon profiles (Clément et al., 1987; Hefetz et al., 1994; Espelie et al., 1994; Breed et al., 1995). Some studies 0.014 (2%) 0.042 (7%) 0.085 (24%) 0.043 (9%) 0.566 (98%) 0.534 (93%) 0.271 (76%) 0.450 (91%) 0.001 (0%) – – – 0.041 (2%) – – – 0.069 (3%) 0.028 (3%) – 0.034 (4%) 0.435 (20%) 0.155 (15%) 0.085 (24%) 0.293 (33%) 1.606 (75%) 0.879 (83%) 0.271 (76%) 0.557 (63%) 10 21 21 15 7 10 36 46 Mols 1.86 ± 0.18 (10) 1.79 ± 0.31 (21) 1.77 ± 0.29 (21) 1.47 ± 0.31 (14) 0.012 (8%) 0.007 (6%) 0.010 (8%) 0.007 (6%) 0.138 (92%) 0.105 (94%) 0.115 (92%) 0.116 (94%) 0.002 (0%) 0.002 (0%) 0.003 (0%) 0.006 (1%) 0.013 (2%) 0.007 (2%) 0.040 (2%) 0.013 (2%) 0.068 (10%) 0.055 (10%) 0.194 (10%) 0.040 (8%) 0.212 (31%) 0.226 (42%) 0.262 (14%) 0.134 (25%) 0.398 (57%) 0.244 (46%) 1.409 (74%) 0.342 (64%) 8 15 20 20 1.81 ± 0.15 (8) 1.81 ± 0.20 (15) 1.80 ± 0.17 (19) 1.68 ± 0.15 (19) 1 2 3 4 Tvärminne PC23 PC13 PC5A PC4A PC3A PC2A PC1A Sample size Samples were taken from two populations of Formica truncorum, one in Mols Bjerge near Aarhus, Denmark and another in Tvärminne near Hanko, Finland. Sample sizes are given in Table 1. Live ants from four colonies in the Finnish population were collected in September 1994 from two neighbouring islands (Mellanskär and Sundholmen; 1 – 1.5 km apart) in the Tvärminne archipelago and kindly made available by L. Sundström (see Sundström, 1993 for details on the Finnish population). The Danish population at Mols had much smaller colonies than the one in Finland. In this population, samples from four colonies were collected in July 1994. For one Danish colony (# 7) samples from the same age cohort of workers were also taken in May 1994. The colonies in the Danish populations were taken from two habitat patches, characterised by different vegetation, which were ca 1 km apart. Only noncallow adult workers were used for gas-chromatographic analysis, i.e. workers whose exoskeletons had completely chitinised. This choice was made because the cuticular hydrocarbon pattern of callow workers is different from that of mature workers in other ant species (Provost et al., 1993). Most populations of Formica truncorum are dominated by monogynous (single-queen) colonies, although highly polygonous populations are occasionally found (Sundström, 1995). The colonies analysed in this study come from such monogynous populations, which Worker headwidth ± SD (n) Field sites and sampling Colony Methods Population have addressed the variation in hydrocarbon profiles of individual workers among colonies of bees and wasps in the same population (Getz and Page, 1991; Espelie et al., 1990, 1994; Singer et al., 1992; Layton et al., 1994; Butts et al., 1995), but only one such study has been done on ants (Bonavita-Cougourdan et al., 1991). Also studies addressing temporal variation in hydrocarbon profiles are relatively rare and currently restricted to ants (Vander Meer et al., 1989; Provost et al., 1993). This study presents an analysis of cuticular hydrocarbon variation among nestmate individuals of the ant Formica truncorum. We have chosen to concentrate exclusively on the cuticular hydrocarbons and to avoid any possibility of including chemicals that are glandular secretions and that were not on the cuticle itself. Hence our use of the solid probe technique, because extraction with hexane runs the risk of including glandular hydrocarbons, particularly from the postpharyngeal and Dufour glands. A detailed justification of this method and a review of the advantages of the solid probe technique are given in Bagnères and Morgan (1991). In addition to advantages of accuracy and avoidance of contamination, another reason for preferring the solid-sampling technique was that much of the functional significance of cuticular hydrocarbons is assumed to be connected to nestmate recognition via mutual antennation among workers. In that context it would also be undesirable to have traces of glandular secretions, which are likely to serve other than recognition purposes, included in the analysis. The species specific hydrocarbon profile of Formica truncorum workers is shown to be remarkably simple. Variation of hydrocarbon profiles within and among colonies is quantified in a Danish and a Finnish population, each with two habitat patches. In addition, we provide evidence for substantial differences in hydrocarbon profiles between spring samples and summer samples of workers of the same age cohort in the same colony. Research article Table 1. Basic statistics on worker headwidth (sample size in brackets), followed by the number of individuals analysed for cuticular hydrocarbons and the eigenvalues from within-colony principal component analysis using standardized values of all hydrocarbons available in each sample (PC1-6 with subscript A). Numbers in brackets are percentages variance explained. As the data are compositional, the last eigenvalue is always equal to zero. The two final columns give the eigenvalues and percentages explained variance of the first and second principal component axis when the analysis was restricted to the 3 major peaks that could be quantified in all colonies Insectes soc. Vol. 46, 1999 59 60 J. Nielsen et al. was previously documented by allozyme analysis (Sundström, 1994; Nielsen, 1995). From the same allozyme studies it was also known that Formica truncorum queens mate with 1–4 males. In order to make the estimates of colony-level differences in cuticular hydrocarbon profiles conservative, we have only analysed samples from colonies which were known to have multiple mated queens, i.e. a subset of colonies with maximal within-colony genetic diversity. Cuticular hydrocarbons in Formica truncorum being disturbed by small peaks from pollution in the sample. Before running any samples collected in the field, a test sample made from a mixture of linear hydrocarbons C20 to C30 was chromatographed, in order to establish retention times of the cuticular hydrocarbons and thereby their exact identity. Statistics: Compositional data and their standardisation Handling and storage of samples After collecting, all individual ants were divided into head, mesosoma, and gaster. Worker mesosomas were stored in Eppendorff plastic tubes and kept at – 70°C until submitted to gel (allozyme) electrophoresis to determine their patriline (Nielsen, 1995). The detailed results of this study will be reported elsewhere. The heads were used for sizemeasurements (headwidth excluding the compound eyes), whereas the antennae and gasters were used for assessing individual cuticular hydrocarbon profiles. To this end, individual gasters and antennae were stored in glass tubes and kept at – 70°C until analysis by gas chromatography. Throughout this procedure, all material to be used for gas chromatography was kept well away from plastic, as chemicals from plastic contaminate hydrocarbon profiles. Preparation of samples for gas chromatography and mass spectrometry The gaster from each individual ant worker was dissected and cleaned mechanically by hand in distilled water to remove all inner tissue, leaving only the cuticula. This cuticula plus the antennae from each individual were packed in a microcapillary glass tube, that was sealed at both ends (Morgan, 1990) and placed in a larger glass tube with the individual’s unique code on it. Usually, a leg or an antenna from each animal is enough for a complete analysis of the cuticular hydrocarbon pattern (Bagnères and Morgan, 1990). However, Formica truncorum workers appeared to contain only relatively small amounts of cuticular hydrocarbons, so that it was necessary to use the cuticle of an entire ant-gaster to exceed the threshold value required by the mass spectrometer. Analysis of the cuticular hydrocarbon profiles The substances present in the lipid layer of Formica truncorum workers were identified using gas chromatography-mass spectrometry (GCMS), using the solid sample technique described by Morgan (1990). Gas chromatography-mass spectrometry was performed on a Hewlett Packard 5890 Gas Chromatograph and a 5870 Mass Selective Detector. The system was controlled by a Hewlett Packard (series 300) computer. The sample was heated in the injector to 220°C for 4 minutes before crushing. The oven was programmed to raise the temperature from 120°C to 270°C at 7 °C per minute and to hold at 270°C for 10 minutes. Helium was used as carrier gas at a flow rate of 1 ml per minute (see Oldham et al., 1994 for further details). The equipment used for quantitative analysis of cuticular hydrocarbons was a Carlo Erba Instrumentazione Fractovap series 4160 gas chromatograph with a flame ionisation detector and a Milton Roy CI 4000 computing integrator. Each sample was loaded into the solid sampler where it was heated for 3 minutes at 250°C before crushing. The split vent was closed before crushing the sample and reopened 2 minutes later. Helium was used as carrier gas with a flow rate of 30 ml per minute through the split vent and a flow rate of 2 ml per minute through the column. The initial temperature of the oven was 180°C. This temperature was held constant for the first 2 minutes followed by an increase of 10 °C per minute until the final temperature was reached at 320 °C. The final temperature was held for 3 minutes and total running time was 19 minutes per sample. The temperature of the detection port was 300 °C. The integrator had a delay time of 9 minutes and a minimum area of 2000 units in order to prevent the analyses from The amounts of single hydrocarbons in each ant sample were presented by the integrator as areas under chromatogram peaks. These areas are given in absolute figures, where the magnitude of an area under a peak depends on the amount of tissue analysed. Standardisations adjusting for such total concentration effects normally involve dividing each hydrocarbon area with the sum of all hydrocarbon areas for the focal ant, resulting in proportional data which add up to one. However, this poses problems for multivariate statistics (Aitchison, 1986), as the compositional nature of such data may induce spurious correlations. A more complicated standardisation was therefore required, and in the present study we followed the one proposed by Aitchison (1986, p. 78 – 79): Z ij = ln (Yij /g(Yj )) where Zij is the standardised peak area i for individual ant j, Yij is the observed peak area i for individual ant j, and g(Yj) is the geometric mean of all peak areas for ant j included in the analysis. Overall analyses of quantitative variation in peak volumes was done by principal component analysis. The analyses involving all 8 colonies had to be restricted to the volumes of the three major peaks, because peak C 31:0 in the Danish colony 36 was missed due to an unfortunate failure of the gas chromatograph. The minor peaks C 27:1 , C 29 :1 and C 31:1 were insufficient to be quantified in some of the colonies (more often in the Danish than in the Finnish colonies) and had to be excluded from the overall analyses. However, the analysis of seasonal variation among two samples of the same colony (# 7) could be based on all peaks except C27:1 . For the further analysis of three standardised peak volumes, we used nested ANOVAs of the “pure Model II” type with unequal sample sizes. ANOVAs were carried out according to Sokal and Rohlf (1981 Box 10.5) using GLIM (Crawley, 1993) and were done on (1) the standardised data for each peak, and (2) scores on the first and second principal component axis describing the total variation across all peaks. In order to check for body-size related variation in hydrocarbons between populations and colonies, ANCOVA’s were performed with “colony” as factor and “headwidth” as covariate, assuming that headwidth is a reasonable predictor of body size. In addition, F-tests and t-tests were applied using methods described by Zar (1984), unless otherwise stated. Results Cuticular hydrocarbon profiles in Formica truncorum Figure 1 shows a typical cuticular hydrocarbon profile obtained from a single worker of Formica truncorum. The profile is dominated by four saturated alkane peaks (Cn : 0), with minor contributions from three unsaturated alkene (Cn : 1) peaks. The major peaks C 25 :0 (pentacosane), C 27 :0 (heptacosane) and C 29 :0 (nonacosane) could be accurately quantified for a representative sample of workers in all eight colonies, but the smaller peaks (C31 :0 , hentriacontane, and the alkene peaks) could only be quantified for some of the colonies. Basic statistics on sample size and worker headwidth are given in Table 1, together with the eigenvalues and percentages of explained variance of the principal component Insectes soc. Vol. 46, 1999 Research article Figure 1. A typical cuticular hydrocarbon profile of a worker of the ant Formica truncorum. Peaks 1 –4 are the saturated hydrocarbons (alkanes) C 25 :0 , C 27 :0 , C29:0 and C 31 :0 which always occurred in appreciable quantities. Peaks 5 –7 are the unsaturated hydrocarbons (alkenes) C 27 : 1 , C 29 : 1 and C31:1 , which occurred in concentrations that were often too low to be measured accurately axes in the colony-by-colony analyses. Note that the number of principal components calculated reflect the number of peaks that could be scored for each colony (maximum six), and that the last principle component is equal to zero per definition, because of the compositional nature of the data. The peak-wise analysis of variation at the three largest peaks is given in Table 2. Significant variance components at the population level were detected for C 27:0 and the second principle component, but these were no longer significant when the habitat patches were included as a separate level in the ANOVAs. In the latter analysis, significant variance components at the patch level were found for C 29: 0 and both principal component axes (Table 2). Colony-level differences were significant for most variables tested (Table 2). Principal component analysis of the quantitative variation in standardised peak volumes for the eight colonies produced the 61 patterns given in Figure 2. In contrast to the data in Table 1, the results given here are from a single PCA on the entire data set, presented in two figures (Fig. 2a, b) to improve resolution. Figure 2a shows the variation within and among the four colonies in the Finnish population. Figure 2b shows the results for the four colonies from the Danish population. Here, the total variation of all data points is considerably larger than that in the Finnish population (which is indicated by the ellipse in Fig. 2b). This visual impression from Figure 2 was confirmed by analysis of the standardised generalised variances (Zhivotovsky, 1988), showing an approximately six fold difference in variation between the two populations (F130, 122 = 5.76; p < 0.00005). It appears that the Danish material forms two separate clusters of two colonies each, representing the two habitat patches of about 1 km apart, with a distinctly different vegetation. Patch-level differences among the two islands in Finland were less pronounced, with the two colonies in the upper left part of Fig. 2a being from one island (Mellanskär) and the two colonies in the bottom right part of Figure 2a from the other island (Sundholmen). Overall, three clusters of nests seem to be clearly separated, where colonies in one Danish habitat patch differ from the Finnish population mostly at the first principal component axis and colonies in the second Danish habitat patch differ from the Finnish population mostly at the second principal component axis (Fig. 2b). Worker headwidths (Table 1) in the Danish population were also more variable than in the Finnish population, both with regard to differences within colonies and with respect to differences among colonies, but the overall means at the population-level were not significantly different (t = 0.90; P = 0.369). ANCOVAs on the standardised scores for C 25 :0 , C 27 :0 , C 29 :0 , and first (PC1) and second (PC2) principal component axes on worker headwidth showed no significant effects in any of the cases (F1, 118 = 2.52, 0.01, 2.77, 2.76, 0.07, respectively; all P’s ≥ 0.1), although some heterogeneity Table 2. Results of nested ANOVAs of quantitative variation in cuticular hydrocarbons of individual workers within and among colonies of the ant Formica truncorum. In both populations, there were two patches with two colonies each. The first row refers to analyses with patches included as a separate level, whereas the second row refers to analysis ignoring patch structure. F-statistics, degrees of freedom (given as integers) and significances (* = P < 0.05; ** = P < 0.025; *** = P < 0.01) were determined according to Sokal and Rohlf ’s (1981) recommendations for unequal sample sizes Among populations Among patches within populations Among colonies within patches within populations Standardised values for C25 F1, 2 = 1.084 (ns) F1, 6 = 2.625 (ns) F2, 4 = 9.403 (ns) – F4, 122 = 3.797 *** F6, 122 = 15.136 *** Standardised values for C27 F1, 2 = 14.247 (ns) F1, 5 = 22.014 *** F2, 4 = 2.139 (ns) – F4, 122 = 1.324 (ns) F6, 122 = 1.826 (ns) Standardised values for C29 F1, 2 = 0.311 (ns) F1, 6 = 0.819 (ns) F2, 4 = 18.495 ** – F4, 122 = 2.570 * F6, 122 = 18.333 *** First principal component F1, 2 = 0.691 (ns) F1, 6 = 1.740 (ns) F2, 4 = 12.273 * – F4, 122 = 3.432 ** F6, 122 = 17.128 *** Second principal component F1, 2 = 3.220 (ns) F1, 6 = 8.289 * F2, 4 = 12.883 ** – F4, 122 = 0.635 (ns) F6, 122 = 3.151 *** 62 J. Nielsen et al. Cuticular hydrocarbons in Formica truncorum a b Figure 2. Individual and colony-level variation in hydrocarbon profile of individual workers of Formica truncorum, projected according to the variation at the first and second principal component axis. Lines connect scores of individual samples with the mean scores for their colony. (a) Variation within and among the four colonies from the Finnish population (from left to right on the horizontal axis colony 4, 3, 1 and 2, respectively). (b) Variation within and among the four colonies from the Danish population (from left to right on the horizontal axis colony 7, 10, 46 and 36, respectively). The ellipse defines the area where the Finnish colonies would have been if they had been plotted in this figure as well among regression slopes was detected for C 25 :0 (F7, 111 = 2.99; P < 0.01) and for PC1 (F7, 111 = 2.79; P < 0.05). Variation between spring and summer samples from the same colony Figure 3. Season-specific variation in cuticular hydrocarbon profile of workers of Formica truncorum as observed in colony 7 from the Danish population. The black dots are the ants collected in May, whereas the open circles are the ants collected in July. Note that the pattern of variation of the July samples is somewhat different than in Figure 2 b, because the principle component axes are mainly determined by the (different) other samples included in the respective analyses For one colony from the Danish population (# 7), quantitative variation for cuticular hydrocarbon patterns was obtained from two samples collected in May and July 1994, respectively. In a principal component analysis these two subsamples separated virtually without overlap (Fig. 3). A discriminant analysis based on six standardised peak values showed a significant difference between the May and July sample (P < 0.001). This implies that it is apparently very important to have samples from the same time-interval when comparing differences among colonies and among populations. This was the case for our comparisons among colonies within each of the two populations, but not for the comparisons between the Danish and Finnish populations. For comparison, colony 7 (the 10 individuals samples in July) is the one most towards the left in Figure 2b. The apparent change in the overall direction of the scatter for colony 7 in Figure 2b and Figure 3 is due to the fact that principal component analysis reveals differences relative to the other samples in the analysis. Insectes soc. Vol. 46, 1999 Discussion Low amounts and diversity of hydrocarbons on the cuticle of Formica truncorum workers In previous work it has been shown that the cuticular hydrocarbons of insects are easily sampled from a few abdominal sclerites, one or two antennae, or a wing (Bagnères and Morgan, 1990), using the solid sampling method applied in our study. This method gives highly reproducible results, and avoids accidental contamination from the postpharyngeal glands (of ants, which contains a similar quantity of the same or similar mixture of hydrocarbons) (Bagnères and Morgan, 1991) or from the Dufour gland (of bumblebees, which also contains a similar mixture) (Oldham et al., 1994), or from contamination on the legs from nest substrates. The total amount and number of components varies with species between e.g. as few as 10 in Manica rubida and 42 in Myrmica rubra (Bagnères and Morgan, 1991). The low diversity of the Formica truncorum cuticular hydrocarbon profiles is thus not exceptional when compared to at least one other species of ant. It can not be explained by the fact that the absolute quantity of hydrocarbons was low, as Oldham et al. (1994) quantified 26 hydrocarbons on a single antenna of Bombus terrestris by using the same solid-sampling technique. Spatial variation in cuticular hydrocarbon profiles The results of Table 2 and Figure 2 indicate that environmental factors are important determinants of variation in cuticular hydrocarbon patterns, but that variation at the population and colony level is not just a simple function of geographical distance. Nests in the Danish population that are in different habitat patches but only 1 km apart may easily be more different from each other than another pair of nests of which one is situated in Denmark and the other in Finland. This suggests that local soil and food conditions may have quite some influence on the quantitative blend of cuticular hydrocarbons produced in each colony. Interestingly, however, not all peaks are equally affected by environmental differences at the various scales. The two-level ANOVA, with colonies nested within populations, and leaving patches out, gave a significant variance component for heptacosane (C 27:0) (P < 0.01) among the two geographically separated populations (reflected also in the second principal component axis, P < 0.05). On the other hand, there is no significant quantitative difference among populations for pentacosane (C 25:0) and nonacosane (C 29:0) (reflected in the first principal component axis), but here the genetic and environmental differences become more clearly expressed among patches and colonies. We note, however, that the population level differences may have been somewhat underestimated, because the minor peaks had to be excluded from the quantitative analyses (Table 1). On the other hand, any differences at this level need to be considered with caution because the samples in Denmark and Finland were taken at different times of the year (Fig. 3, see also below). Research article 63 To our knowledge, the only other study directly comparing within and among colony variation in hydrocarbon profiles of individual workers is the one by Butts et al. (1995) on the hornet, Vespa crabro. In this study, the resolution per individual was higher, because more peaks were available for analysis, but the number of colonies (6) and the number of workers per colony (6) analysed was lower than the sample sizes in our study. Superficially, the pattern of variation found in Vespa crabro is similar to the one in Formica truncorum. Some colonies separate completely in 2-dimensional principal component plots, whereas others partially overlap. In some respects, the social systems of Vespa crabro and Formica truncorum may be similar, as both are characterised by single queen (monogyne) colonies. However, it remains unknown whether hornet queens were also mated with several males as were the queens of Formica truncorum whose offspring were analysed in this study. In more detail, it seems that the colony-level variation among individual workers is more discrete in hornets than in Formica truncorum. The reason for this difference may be the already mentioned higher number of variables (peak volumes) in the principal component analysis, the genetically more discrete nature of hornet colonies (in case Vespa crabro would prove to have mostly or exclusively single queen mating), or a more variable environment in hornets than in wood ants. Whatever is the case, it seems that variation in cuticular hydrocarbon profiles of individual workers could provide hornet workers with more accurate tools for nestmate recognition than workers of Formica truncorum. Whilst Table 2 demonstrates that pentacosane (C 25 :0) and nonacosane (C 29 :0) occur in colony-specific proportions, the overlap of individual scores of Formica workers from different colonies within the same habitat patch (Fig. 2) suggests that cues other than cuticular hydrocarbons are used in order to achieve accurate nestmate recognition. That nestmate recognition is accurate in this species can be inferred from the fact that strange and unexpected allozyme genotypes were not found in subsequent samples from these colonies spanning a total of seven years (L. Sundström and J.J. Boomsma, in prep.). On the other hand, increasing genetic variation within colonies as found in polygynous populations of Formica truncorum is known to results in reduced discrimination of workers between nestmate and non-nestmate queens (Sundström, 1997). Temporal variation in cuticular hydrocarbon profiles Belonging to different castes or having different agerelated tasks may expose workers from the same colony to different environments within and around their nest, which may lead to differences in cuticular hydrocarbon patterns (Bonavita-Cougourdan et al., 1994). Differences in tasks in relation to age are well known in honey bees (Winston, 1987) and ants (Wilson, 1971). Task differences have also been reported between matrilines in polygynous ants (Snyder, 1992), and between patrilines in honey bees (Robinson and Page, 1988, Frumhoff and Baker, 1988). Also this factor may have contributed to the relatively high within-colony varia- 64 J. Nielsen et al. tion found in Formica truncorum. Ant workers are more long-lived than workers of honey bees or hornets, so that our Formica samples were relatively heterogeneous with respect to age. From the single Danish colony in our study where two samples were available (Fig. 3), it seems clear that temporal variation in individual hydrocarbon profiles may be considerable and that this may induce considerable errors when comparing colony level patterns from samples that have not been collected within a short period of time. Temporal differences of this kind may be a general phenomenon, as similar cases have been reported for fire ants (Vander Meer et al., 1989). In addition, differences in diet have also been shown to induce changes in hydrocarbon profiles in larvae of the non-social insect Manduca sexta (Espelie and Bernays, 1989). As already mentioned, some of the differences between our Danish population (sampled in July) and the Finnish population (sampled in September), may therefore have been due to temporal differences in diet and general nest environment. However, our main conclusions are unlikely to have been much affected by this complication, as populationlevel differences in hydrocarbon profiles were mostly nonsignificant (see previous section). Genetic versus environmental factors determining variation in hydrocarbon profiles Studies on honey bees by Page et al. (1991) have demonstrated that variation in cuticular hydrocarbon profiles has a significant genetic component. These authors conclude that in spite of active exchange of hydrocarbons between workers of the same nest, each individual still maintains enough of its own hydrocarbon profile to represent a genetically distinct entity. This implies that part of the colony-level variation observed in our study (Fig. 2) and in Butts et al. (1995) may be genetic, as the colony queen and her mate(s) represent genetically discrete entities. However, colonies also have different environments, so that the present data on Formica truncorum do not allow us to partition the among-colony variance of hydrocarbon patterns into a genetic and an environmental component. In order to do this, it would be necessary to partition between genetic entities sharing the same environment, for example by comparing workers with the same mother but with a different father within the same colony. An example of such a set-up has recently been provided for honey bees by Arnold et al. (1996). Acknowledgements We thank Lotta Sundström for providing the samples from the Tvärminne population in Finland, Diethe Ortius and Jes S. Pedersen for comment on the manuscript, Camilla Håkansson for measuring headwidths of ants and the Mols Laboratory, Femmøller, Denmark for providing facilities during the fieldwork. The work reported here was supported by a grant from the Danish Natural Science Research Council to J.J. B. and through the research network “Social Evolution” of the Universities of Aarhus, Firenze, Keele, Sheffield, Uppsala, Würzburg and the ETH Zürich, financed by the European Commission via the Training and Mobility of Researchers (TMR) programme. Cuticular hydrocarbons in Formica truncorum References Aitchison, J., 1986. The Statistical Analysis of Compositional Data. Chapman and Hall, London. pp. 416. Arnold, G., B. Quenet, J.-M. Cornuet, C. Masson, B. De Schepper, A. Estoup and P. Gasqui, 1996. Kin recognition in honeybees. Nature 379: 498. Bagnères, A.-G. and E.D. Morgan, 1990. A simple method for the analysis of insect cuticular hydrocarbons, J. Chem. Ecol. 16: 3263– 3276. Bagnères, A.-G. and E.D. Morgan, 1991. The postpharyngeal glands and the cuticle of Formicidae contain the same characteristic hydrocarbons. Experientia 47: 106–111. Blomquist, G.J., D.R. Nelson and M. de Renobales, 1987. Chemistry, biochemistry, and physiology of insect cuticular lipids. Arch. Insect Biochem. Physiol. 6: 227–265. Bonavita-Cougourdan, A., J.-L. Clément and C. Lange, 1987. Nestmate recognition: The role of cuticular hydrocarbons in the ant Camponotus vagus Scop. J. Entomol. Sci. 22: 1–10. Bonavita-Cougourdan, A., G. Theraulaz, A.-G. Bagnères, M. Roux, M. Pratte, E. Provost and J.-L. Clément, 1991. Cuticular hydrocarbons, social organization and ovarian development in a polistine wasp: Polistes dominulus Christ. Comp. Biochem. Physiol. 100 B: 667– 680. Bonavita-Cougourdan, A., E. Provost, G. Rivière, A.-G. Bagnères and J.-L. Clément, 1994. Nestmate recognition in ants: Plasticity of signals and genetic factors. In: Les Insectes Sociaux (A. Lenoir, G. Arnold and M. Lepage, Eds.), Sorbonne, Paris. p. 142. Breed, M.D., 1983. Nestmate recognition in honey bees. Anim. Behav. 31: 86–91. Breed, M.D. and G.E. Julian, 1992. Do simple rules apply in honey bee nestmate discrimination? Nature 357: 685–686. Breed, M.D., K.R. Williams and J.H. Fewell, 1988. Comb wax mediates the acquisition of nest-mate recognition in honey bees. Proc. Natl. Acad. Sci. USA 85: 8766–8769. Breed, M.D., M.F. Garry, A.N. Pearce, B.E. Hibbard, L.B. Bjostad and R.E. Page, Jr.,1995. The role of wax comb in honey bee nestmate recognition. Anim. Behav. 50: 489–496. Butts, D.P., M.A. Camann and K.E. Espelie, 1995. Workers and queens of the European hornet Vespa crabro L. have colony-specific cuticular hydrocarbon profiles (Hymenoptera: Vespidae). Insectes. soc. 42: 45–55. Clément, J.-L., A. Bonavita-Cougourdan and C. Lange, 1987. Nestmate recognition and cuticular hydrocarbons in Camponotus vagus Scop. In: Chemistry and Biology of Social Insects. (J. Eder and H. Rembold, Eds.), Verlag J. Peperny, Munich, Germany. pp. 473–474. Crawley, M.J., 1993. GLIM for Ecologists. Blackwell Scientific Publications, Oxford, UK. pp. 379. Crosland, M.W.J., 1990. Variation in ant aggression and kin discrimination ability within and between colonies. J. Insect Behav. 3: 359–379. Crozier, R.H., 1987. Genetic aspects of kin recognition: Concepts, Models, and Synthesis. In: Kin recognition in animals (D.J.C. Fletcher and C.D. Michener, Eds.), Wiley & Sons. pp. 55–73. Dahbi, A. and A. Lenoir, 1998. Queen and colony odour in the multiple nest ant species, Cataglyphis iberica (Hymenoptera, Formicidae). Insectes soc. 45: 301–313. Espelie, K.E. and E.A. Bernays, 1989. Diet-related differences in the cuticular lipids of Manduca sexta larvae. J. Chem. Ecol. 15: 2003– 2017. Espelie, K.E., J.W. Wenzel and G. Chang, 1990. Surface lipids of the social wasp Polistes metricus Say and its nest and nest pedicel and their relation to nestmate recognition. J. Chem. Ecol. 16: 2229–2241. Espelie, K.E., E.A. Bernays and J.J. Brown, 1991. Plant and insect cuticular lipids serve as behavioral cues for insects. Arch. Insect Biochem. Physiol. 17: 223–233. Insectes soc. Vol. 46, 1999 Espelie, K.E., G.J. Gamboa, T.A. Grudzien and E.A. Bura, 1994. Cuticular hydrocarbons of the paper wasp, Polistes fuscatus: a search for recognition pheromones. J. Chem. Ecol. 20: 1677–1687. Frumhoff, P.C. and J. Baker, 1988. A genetic component to division of labour within honey bee colonies. Nature 333: 358–361. Gamboa, G.J., H.K. Reeve and D.W. Pfennig, 1986. The evolution and ontogeny of nestmate recognition in social wasps. Ann. Rev. Entomol. 31: 431–454. Getz, W.M. and R.E. Page, Jr., 1991. Chemosensory kin-communication systems and kin recognition in honey bees. Ethology 87: 298–315. Hefetz, A., V. Soroker, C. Vienne, E. Nowbahari and C. Errard, 1994. The glandular source and the mode of transfer of nestmate recognition cues in ants. In: Les Insectes Sociaux (A. Lenoir, G. Arnold and M. Lepage, Eds.), Sorbonne, Paris, p. 216. Howard, R.W., 1993. Cuticular hydrocarbons and chemical communication. In: Insect lipids. Chemistry, Biochemistry and Biology (D.W. Stanley-Samuelson and D.R. Nelson, Eds.), University of Nebraska Press, Lincoln, Nebr. pp. 179–226. Howard, R.W. and G.J. Blomquist, 1982. Chemical ecology and biochemistry of insect hydrocarbons. Ann. Rev. Entomol. 27: 149–172. Jaisson, P., 1991. Kinship and fellowship in ants and social wasps. In: Kin Recognition (P.G. Hepper, Ed.), Cambridge University Press, Cambridge, UK. pp. 60–93. Layton, J.M., M.A. Camann and K.E. Espelie, 1994. Cuticular lipid profiles of queens, workers and males of the social wasp Polistes metricus Say are colony specific. J. Chem. Ecol. 20: 2307–2321. Lockey, K.H., 1988. Lipids of the insect cuticle: Origin, composition and function. Comp. Biochem. Physiol. 89 B: 595–645. Lorenzi, M.C., A.-G. Bagnères, J.-L. Clément and S. Turillazzi, 1997. Polistes biglumis bimaculatus epicuticular hydrocarbons and nestmate recognition (Hymenoptera, Vespidae). Insectes soc. 44: 123–128. Morgan, E.D., 1990. Preparation of small-scale samples from insects for chromatography. Analytica Chimica Acta 236: 227–235. Morel, L. and M.S. Blum, 1988. Nestmate recognition in Camponotus floridanus callow worker ants: Are sisters or nestmates recognised? Anim. Behav. 36: 718–725. Nielsen, J., 1995. Cuticular hydrocarbon patterns in relation to assessment of relatedness asymmetry in the ant Formica truncorum. MSc thesis, University of Aarhus, Denmark. Oldham, N.J., J. Billen and E.D. Morgan, 1994. On the similarity of the Dufour gland secretion and the cuticular hydrocarbons of some bumblebees. Physiol. Entomol. 19: 115–123. Page, R.E., Jr., R.A. Metcalf, R.L. Metcalf, E.H. Erickson, Jr. and R.L. Lampman, 1991. Extractable hydrocarbons and kin recognition in honeybee (Apis mellifera L.). J. Chem. Ecol. 17: 745–756. Peeters, C.P., 1988. Nestmate discrimination in a ponerine ant (Rhytidoponera sp. 12) without a queen caste and with low intra-nest relatedness. Insectes soc. 35: 34–46. Provost, E., G. Rivière, M. Roux, E.D. Morgan and A.-G. Bagnères, 1993. Change in the chemical signature of the ant Leptothorax lichtensteini Bondroit with time. Insect Biochem. Molec. Biol. 23: 945–957. Robinson, G.E. and R.E. Page, Jr. 1988. Genetic determination of guarding and undertaking in honey-bee colonies. Nature 333: 356–358. Research article 65 Ruther, J., S. Sieben and B. Schricker, 1998. Role of cuticular lipids in nestmate recognition of the European hornet Vespa crabro L. (Hymenoptera, Vespidae). Insectes soc. 45: 169–179. Ryan, R.E., T.J. Cornell and G.J. Gamboa, 1982. Nestmate recognition in the bald-faced hornet, Dolichovespula maculata (Hymenoptera: Vespidae). Z. Tierpsychol. 69: 19–26. Singer, T.L. and K.E. Espelie, 1992. Social wasps use nest paper hydrocarbons for nestmate recognition. Anim. Behav. 44: 63–68. Singer, T.L., M.A. Camann and K.E. Espelie, 1992. Discriminant analysis of cuticular hydrocarbons of social wasp Polistes exclamans Viereck and surface hydrocarbons of its nest paper and pedicel. J. Chem. Ecol. 18: 785–797. Snyder, L.E., 1992. The genetics of social behavior in a polygynous ant. Naturwissenschaften 79: 525–527. Sokal, R.R. and F.J. Rohlf, 1981. Biometry. Second edition. W.H. Freeman and Company, New York. pp. 776. Soroker, V., C. Vienne and A. Hefetz, 1995. Hydrocarbon dynamics within and between nestmates in Cataglyphis niger (Hymenoptera: Formicidae). J. Chem. Ecol. 21: 365–378. Stuart, R.J., 1987. Individual workers produce colony-specific nestmate recognition cues in the ant Leptothorax curvispinosus. Anim. Behav. 35: 1062–1069. Sundström, L., 1993. Genetic population structure and sociogenetic organisation in Formica truncorum (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 33: 345–354. Sundström, L., 1994. Sex ratio bias, relatedness asymmetry and queen mating frequency in ants. Nature 367: 266–268. Sundström, L., 1995. Sex allocation and colony maintenance in monogyne and polygyne colonies of Formica truncorum (Hymenoptera: Formicidae): the impact of kinship and mating structure. Amer. Nat. 146: 182–201. Sundström, L., 1997. Queen acceptance and nestmate recognition in monogyne and polygyne colonies of the ant Formica truncorum. Anim. Behav. 53: 499–510. Vander Meer, R.K., D. Saliwanchik and B. Lavine, 1989. Temporal changes in colony cuticular hydrocarbon patterns of Solenopsis invicta. Implications for nestmate recognition. J. Chem. Ecol. 15: 2115–2125. Venkataraman, A.B., V.B. Swarnalatha, P. Nair and R. Gadagkar, 1988. The mechanism of nestmate discrimination in the tropical social wasp Ropalidia marginata and its implications for the evolution of sociality. Behav. Ecol. Sociobiol. 23: 271–279. Vienne, C., V. Soroker and A. Hefetz, 1995. Congruency of hydrocarbon patterns in heterospecific groups of ants: transfer and/or biosynthesis? Insectes soc. 42: 267–277. Wilson, E.O., 1971. The Insect Societies. Belknap Press of Harvard University Press, Cambridge, Mass. pp. 548. Winston, M.L., 1987. The Biology of the Honey Bee. Harvard University Press, Cambridge, Mass. pp. 281. Zar, J.H., 1984. Biostatistical Analysis. Second edition. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. pp. 718. Zhivotovsky, L.A., 1988. Some methods of analysis of correlated characters. In: Proceedings of the Second International Conference on Quantitative Genetics (B.B. Weir, E.J. Eisen, M.M. Goodman and G. Namkoong, Eds.), Sinauer, Sunderland, Mass., pp. 423–432.