Great Basin Naturalist

Volume 59 Number 2 Article 11

4-30-1999

Assortative mating in soldier (Cantharidae, Chauliognathus): test of the mate-choice hypothesis

Ruth Bernstein University of Colorado, Boulder

Stephen Bernstein University of Colorado, Boulder

Follow this and additional works at: https://scholarsarchive.byu.edu/gbn

Recommended Citation Bernstein, Ruth and Bernstein, Stephen (1999) "Assortative mating in soldier beetles (Cantharidae, Chauliognathus): test of the mate-choice hypothesis," Great Basin Naturalist: Vol. 59 : No. 2 , Article 11. Available at: https://scholarsarchive.byu.edu/gbn/vol59/iss2/11

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Great Basin Naturalist 59(2), ©1999, pp. 188-192

ASSORTATIVE MATING IN SOLDIER BEETLES (CANTHARIDAE: CHAULIOGNATHUS): TEST OF THE MATE-CHOICE HYPOTHESIS

Ruth Bernsteinl and Stephen Bernsteinl

ABSTH.v:r.-SoJdier beetles of 2 species, Ch(luliognathus basalis and C. deceptus, were examined to test the Crespi hypothesis that positive assortative mating by size is caused by mate choice. Specifically, we tested the prediction that if mate choice involves choosing the largest mate available, then mating individuals v!'ill be larger than nonmating individ­ uals. Four samples were taken, at different times during the mating season, from each of2 sites. Each sample consisted ofmating pairs, nonmating males, ancl nonmating females. Some ofthe samples contained beetles of both species; others contained beetles of a single specie.~_ For each gender elytron lengths of mating individuals were compared with elytron lengths of nonmating individuals. \Ve found no effect ofmating status (mating ys_ nomnating) on elytron lengths in sam­ ples that exhihited assortative mating (which occurs where 2 species coexist). Surprisingly, we found a consistent effect of mating status on elytron lengths in samples that did not exhibit a.ssortative mating (which occurs where only 1 species exists). Our results do not support the mate-choice hypothesis. Instead, mate choice and assortative mating appear to be alternative mating patterns in which mate choice occurs \'ihere a single species exists and assortative mating occurs where 2 species coexist.

Key words: mate choice, (lSS01-tative mating, soldier beetles, Chauliognathus deceptus, Chauliognathus basalis.

Positive assortative mating by size occurs a preference for larger males (Crespi 1989), when the body sizes of mating pairs are more \Vhen male choice occurs in soldier beetles, it similar than if they mated at random. This is most likely because larger females carry mating pattern has been observed in soldier more eggs (Ridley 1983), Wnen female choice beetles (Chauliognathus), as reported by Mason occurs, the apparent preference may be due to (1972), McCauley and Wade (1978), McLain the superior ability of larger males in over­ (1982, 1984, 1985), and Bernstein and Bern­ coming the higher "resistance to mating" of stein (1998), The ultimate cause of positive larger females (McCauley and Wade 1978, assortative mating by size may be sexual selec­ McCauley 1981), A prediction of the mate­ tion, in which differences in reproductive suc­ choice hypothesis is that the mean size ofmat­ ce!'iS, cau!'ied by competition over mates, are ing individuals will be larger than the mean related to body size (Andersson 1994), Alter­ size of nonmating individuals (Arnqvist et al. natively, positive assortative mating by size 1996), Most data on do, in fact, may, be an artif~lct of environmental factors, show this mating pattern, either for females or such as temporal or spatial covariance ofbody for both sexes (Crespi 1989), size among mating males and females. In the study described herein, we exam­ Many hypotheses have been offered to ex­ ined the mate-choice hypothesis in 2 species plain assortative mating in arthropods (Crespi of soldier beetles (Chauliognathus deceptus Le 1989), One ofthese is the mate-choice hypotlle­ Conte and C. basalis Fender), These beetles sis, hased on sexual selection, in which indi­ mate conspicuously, occur in large populations, viduals choose larger mates because they ben­ and remain coupled for many hours, (In a pre­ ellt reproductively and m'e differently capable liminary study vdth marked pairs, "ve found ofexercising choice (Darwin 1871, Ridley 1983), that 68% of mated pairs remained together for Male choice involves a large-male mating ad­ more than 5 hand 34% for more than 17 h), In vantage in male-male competition for females a previous study (Bernstein and Bernstein 1998) comhined with a preference for larger females; we found positive assortative mating by body female choice involves hlctors that increase size in some populations but not in others. large-female pairing probabihty combined with Arnqvist et aL (1996) lists, for the mate-choice

II )qJaI'tI1Wl It or Environll"lental, I'rlllu!l1t(ml. and Or!(anhHlk l\iolo!!)'. (j niV(:'~ity ofColorado. Boulcl<;l; CO 80,109,

188 1999] MATE CHOICE IN SOWIER BEETLES 189 hypothesis, the following predictions: (1) the ba.salis, an unexpected result since the 2 form of assorlative mating will be true, (2) species cannot be distinguished in the Held. mating males will be larger than nonmating Males were identiHed to species by the shape males, and (3) mating females will be larger of the copulatory organ, using the key pro­ than noomating fema.les. The first prediction vided in Fender (1964); females were identi­ ha, been tested (Bernstein and Bel11stein 1998) fled by species-specific correlations between and found true: the correlation between body length of the posterior elytron spot and length sizes ofmates is true (linear) rather than appal"· of the elytron (Ilernstein and Bernstein 1998). ent (in which the variance in male size changes 11,e right elytron of each was severed with increasing female size). Here, we test the from the body and its maximum length was remaining 2 predictions, specifically that mat­ measured to the nearest 0.001 mm, lIsing a ing individuals are larger than nonmating indi­ binocular microscope with an eyepiece viduals in the populations that exhibit assorta­ micrometer. tive mating but not in the populations that do The eflects of sampling time aod mating not exhibit assortative mating. status (mating versus nonmaling) on body size were analyzed by 2-factor analyses ofvariance. METHODS Sampling time represents both progression of the mating ~ea.son and variations in assortative We eolJccted mating and nonmating beetles mating, as only the first 2 samples at the from 2 sites, one on the plains and the other in canyon site exhibited this matin~ pattern a canyon, within 30 km of Boulder, Colorado. (Bernstein and Bernstein 1998). All samples The plains site is a meadow (elevation 1760 m) were redutrt'CI in size (hy random elimination) near Eldorado Springs, where beetles were to the smallest sample in order to meet the leeding and mating on sunllower (Helillnthus recommended equal sample sizes for the 2­ annum L.) blossoms. The canyon site is a lactor AN OVA. We hegan with 24 groups: a roadside in South St. Vrain Canyon (elevation male group and a female group fi'om each of 1830 m) near Lyons, where beetles were feed­ the 4 samples of C. deceptw! at the canyon ing and mating on blossoms of rabbitbrush site, Ii-om each of the 4 samples of C. ha.salis at [Chrysotharn",", TUtuseosus (Pall.) Britt.]. Eacll the canyon site, and from each of the 4 sam­ sampling site encompassed an area of less ples of C. ba.salis at the plains site. Five of the than 0.5 ha. At each site 4 samples were taken 24 groups were eliminated from analysis, 4 at 1-wk intervals during the approximately because they were too small (the male group month-long mating season. All beetles in each and female group from samples 3 and 4 of C. sample were collected on a single morning, ba.salis at the canyon site) and 1 because it was between 0900 hand 1030 h. At this time of so mnch smaller than ti,e other samples (sam­ day beetles are too cool for rapid locomotion ple 4 of C. ba.salis females at the plains site) and so are likely to have been coupled since at that we preferred to eliminate it rather than least the previous evening. The sluggish condi­ greatly reduce the other samples. Where sig­ tion also prevents sampling bias, as any beetle nificant eOects of sampling time were found, (regardless of size or mating statns) is easily the means of mating and of nonmating indi­ captured by sliding it li'om a blossom into a viduals were compared llsing the Newman­ collecting vial. Whenever possible, each sample Kenis multiple-comparison test (Zar 1996) with consisted of 40 mating pairs, 40 nonmating the level ofsignincancc set at 0.05. males. and 40 nOll11l3ting females. Samples of 10 presenting a summary of our results, we mating pairs are the same ones reported in an compare the pattern ofassortative mating with earlier publication in which patterns of assnr­ the pattern ofsize diflerences between mating tative mating were described (Bernstein and and nonmating individuals among the samples. Bernstein 1998). Part of this comparison involved testing for Beetles were frozen within a few hours after differences among the Pearson product corre­ capture and tben preserved in 70% alcohol. lation coefficients (r), our measure of assorta­ Plains samples consisted entirely of Chauliog­ tive mating. For these tests we lollowed the I'IlI1.htts ba.salUJ. Canyon samples consisted of 2 statistical prncedures descIibed by Zar (1996) species, ChauliogTUtthu., deceptus and C. with the level of significance set at 0.05. 190 GREAT BASIN NATURALIST [Volume 59

TABLE 1. 1\vo-factor ANOVA results for males: effects of sampling time and mating status (mating versus lioumating) 00 elytron lenhrth (NS :::0 P > 0.05; NA ;=: not applicable).

Species Effect of Effect of Interaction Newman-Keuis test (site) sampling time mating status effect (P < 0.05)

~Jating; C. deceptus F3.2)(; = 4.270 F1,216 =; 16.050 F3,216 = 4.981 Xl = X2 = x3 = X4 (canyon) IF < 0.01) (P < 0.0005) IP < 0.(025) Nonmating: Xl = x2 > X3 = X4

C. basalis FJ,.3fj = 0.380 F 1,36 = 1.264 F 1,36 = 0.157 NA (canyon) INS) (NS) (NS)

C. basalis F = 0.239 1,240 F ,240 = 0.221 NA 3240• F = 34.476 3 (plains) INS) IP < 0.0005) (KS)

TABLE 2. Two-(tdor ANOYA results for females: effects of sampling time and mating status (mating versus lioumating) on elytron length (NS = P > 0.05; NA = not applicable).

Species Efli:~<.:t of Effect of Interaction Newman-KeuIs test (site) sampling time mating status effect (P < 0.05)

C. deceptus F,1,176 = F ,176 = 30.827 F = 8.474 Mating: Xl := X = 1: = X4 4.752 1 3176• z 3 (canyon) (P < 0.005) (P < 0.(005) IF < 0.0(05) Nonmating; Xl = X2 > x3:= 1:4

C. hasalis 1<'1,36 = 0.124 F 36 = 0.270 F = 0.108 NA J• I36• (canyon) INS) (NS) INS)

C. basalis F ,.2.'34 F 15.410 := NA 2 =0.014 l234• = F2234• 0.534 (plains) (NS) (P < 0.(005) INS)

RESULTS there are no effects of sampling time, mating status, or interaction on the elytron length of Results of the 2-factor analyses of variance either males or females. For C. basalis at the are presented for males in Table 1 and females plains site, there is a significant effect of mat~ in Table 2. Each table shows the effect on ing status on elytron length but no effect of mean elytron length of sampling time, mating sampling time nor of an interaction (i.e., mat­ status, and interaction between sampling time ing and nonmating individuals are different in and mating status. Results of the Newman~ size, and this difference remains the same KeuIs test for differences among sampling across samples). times are provided for each ANOVA with a Results of our studies are summarized in significant difference among samples. Table 3 where, for each sample, we give the Males and females exhibit the same relative abundance of a congener, correlation ANOYA pattern. In C. deceptus there are sig­ coefficient (r) of elytron lengths of mating nificant effects of sampling time and mating pairs, size ratio of mating to nonmating males, status, as well as the interaction between sam­ and size ratio of mating to nonmating females. pling time and mating status, on elytron lengths. For both species, samples 1 and 2 from the An interaction effect indicates that size differ­ canyon site have higher correlation coeffi­ ences between mating and nonmating individ­ cients, higher percentages of the less abun­ uals occur in some samples but not in others. dant species, and virtually no size differences According to the Newman-Keuls multiple­ between mating and nonmating individuals. comparisons tests, elytron lengths of mating For C. deceptus in samples 3 and 4 from the individuals remain the same across samples, canyon site and for all c. basalis samples from whereas in nonmating individuals there is a the plains site, correlation coefficients are low, significant diflerence between samples 2 and percentages of the less abundant species are 3 (i.e., nonmating individuals are smaller in low, and mating and nonmating individuals samples 3 and 4 than in samples 1 and 2). For are more different in size than in samples 1 C. basalis at the canyon site (samples 1 and 2), and 2 from the canyon. There are significant 1999] MATE CHOICE IN SOLDIER BEETLES 191

TABLE 3. Summary ofsample values: percent ofthe less abundant species (n == total number ofindividuals in the sam- ple), strength of assortative mating (coefficient r of the correlation between elytron lengths of mates; n = number of pairs), and size ratios ofmated and unmated individuals (n = number ofratios = n ofgroups in the ANOVA).

Sample 1 Sample 2 Sample 3 Sample 4 C. deceptus (canyon site) % less abundant 29.4 32.5 10.6 13.1 species in sample (n ~ 160) (n = 160) (n = 160) (n = 160) r values 0.421 0.651 0.157 -0.019 (differences among (n ~ 30) (n ~ 28) (n ~ 38) (n ~ 36) the values: P < 0.05) male size ratio: 1.000 1.002 1.057 1.064 mated + unmated (n ~ 28) (n = 28) (n = 28) (n = 28) female size ratio: 1.006 1.001 1.087 1.067 mated + unmated (n ~ 23) (n = 23) (n = 23) (n = 23)

C. basalis (canyon site) % less abundant 29.4 32.5 10.6 13.1 species in sample (n ~ 160) (n ~ 160) (n ~ 160) (n ~ 160) r values 0.533 0.529 too few too few (differences among (n ~ 10) (n ~ 12) the values: NS) male size ratio: 1.022 1.013 too few too few mated + unmated (n ~ 10) (n ~ 10) female size ratio: 1.003 1.015 too few too few mated + unmated (n ~ 10) (n ~ 10)

C. basalis (plains site) % less abundant 0 0 0 0 species in sample (n ~ 160) (n ~ 160) (n = 160) (n ~ 124) r values 0.042 0.095 0.168 "'{).088 (differences among (n ~ 40) (n = 40) (n ~ 40) (n ~ 40) the values: NS) male size ratio: 1.048 1.042 1.047 1.033 mated + unmated (n ~ 31) (n = 31) (n ~ 31) (n ~ 31) female size ratio: 1.039 1.037 1.020 too few mated + unmated (n ~ 40) (n = 40) (n ~ 40)

differences (X2 = 9.571) among correlation which individnals prefer larger mates. A pre­ coefficients (r) of the 4 samples of C. deceptus diction of this hypothesis (in addition to aSSOf­ at the canyon site. There is no significant dif­ tative mating) is that mating individuals are ference (z = 0.011) between correlation coeffi­ larger than nonmating individuals because cients of the 2 samples of c. basalis in the larger individuals mate more often and/or canyon, nor are there significant differences remain coupled for longer periods of time (X2 = 0.309) among the 4 samples of C. basalis than the less desirable, smaller individuals. in the plains site. We tested the specific prediction that mating individuals are larger than nonmating individ­ DISCUSSION uals in populations that exhibit positive assor­ tative mating and found the prediction to be One hypothesis to explain positive assorta­ false for 2 species of soldier beetles (c. basalis tive mating is the mate-choice hypothesis in and c. deceptus). In populations ofthese species 192 GHEAT BASIN NATUHALlST [Volume 59 that exhibit assortative mating (as measured by advantage of being paired, regardless of mate corrrelations of the elytron lengths of mates), size, under these conditions. In our study the there are no differences between elytron lengths advantage of having a large mate may be ofmating and lioumating individuals. In popu­ countered by the disadvantage of mating with lations that do not exhibit assortative mating, an individual of the wrong species. 'rhe rela­ however, mating males and females are larger tionship between inhibition ofmate choice and than lioumating males and females. stimulation of assortative mating, however, Is mate choice the usual mating pattern in remains unclear. soldier beetles? Where mate choice (as defined by larger mating than noumating individuals) LITERATURE CITED was examined in previous studies, all involv­ ing C. pennsylvanicus, the results were mixed. ANDERSSON, M. 1994. Sexual selection. Princeton Univer­ Mason (1980) found evidence offemale choice sity Press, Princeton, NJ. 599 pp. ARNQVIST, G., L. ROWE, J.J. KnUPA, AND A. SnL 1996. (larger mating than nonmating males) in 2 of3 Assortative mating by size; a meta-analysis ofmating populations sampled in northern New York, patterns in water striders. Evolutionary Ecology 10: but no evidence ofmale choice. McLain (1982) 265-284. found both male and female choice in all 6 BEHNSTEIN, n., AND S. BERNSTEIN. 1998. Assortative mat­ ing by size in two species of Chauliognathus Henz populations sampled in northern Georgia. A (Coleoptera; Cantharidae). Southwestern Naturalist later study (McLain 1985) found no evidence 43,62-69. of female choice (only males were measured) CRESPI, BJ. 1989. Causes of assortative mating in arthro­ in a cline of15 populations in northern Georgia. pods, Behavior 38:980-1000. DARWIN, C. 1871. The descent of man, and selection in A tentative interpretation of our results is relation to sex. John Murray, London. that mate choice is the normal mating pattern FENDEH, K.M. 1964. The Chauliognathini ofAmerica north in soldier beetles, but that presence of a con­ of Mexico (Coleoptera-Cantharidae). Northwest Sci­ gener on the same host plant inhibits this nor­ ence 38:52-64, 95-106. mal behavior and triggers assortative mating. MASON, L.G, 1972. Natural populations and assorta­ tive mating. Amel;can Midlwld Naturalist 88:150-157. What is the evidence that assortative mating . 1980. Sexual :>election and the evolution of pair­ ,md mate choice arc alternative mating pat­ -~b'ondingin soldier beetles. Evolution 34;174-180. terns in soldier beetles? Three previous stud­ MCCAULEY, D.E. 1881. Application of the Kence-Bryant ies, all involving C. pennsylvanicus" have eval­ model of mating behavior to a natural population of uated both mating patterns in the same popu­ soldier beetles. American Naturalist 117:400-402. MCCAULEY, D.E., AND M.J. WADE. 1978. Female choice lations. McLaiu (1982) found both male and and the mating structure of a natural population of female choice, but no asssortative mating, in 6 the , Clutuliognathus pennsylooni,cus. populations he sampled in Georgia (a result Evolution 32:771-775. that lends support to our interpretation). How­ McLAIN, D.K. 1981. Interspecific interference competi­ tion and mate choice in the soldier beetle, Chauli­ ever, in a later study McLain (1985) found nei­ ognathl1R pennsylvanicus. Behavioral Ecology and ther mate choice nor assortative mating in a Sociobiology 9:65-66. cline of 15 populations in Georgia. McCauley _---,,,. 1982. Density dependent sexual selection and and Wade (1978) found both mate choice (by positive phenotypic assortative mating in natural populations of the soldier beetle, Clw,uliogna.thus males and by females) and assortative mating pennsylvanicus. Evolution 36:1227-1235. in populations they studied in Illinois. Thus, ___. 1984. Host plant rnorpholo~,ry, speciation, and the the results on C. pennsylvanicus do not form a economics of mate choice in the soldier beetle, consistent pattern, Chauliognathus pennsylvanicus. Evolutionary The­ Our results suggest that the soldier beetles ory 7:63-67. ----,C"~. 1985. Clina! variation in morphology and assorta­ we studied exhibit mate-choice behavior except tive mating in the soldier beetle, Chauliognathus in the presence of a congener. McLain (1981), pentL~yl1X/.nicu8 (Coleoptera: Canthm·idae). Biological in studies of C. pennsylvanicus, also found Journal ofthe Linnaean Society 2Ei:10,'J-117. inhibition ofmate-choice behavior (in females) HlDLEY, M. 1983. The explanation of organic diversity: the by the presence of other species, in this case compm'ative method and adaptations for mating. Clarendon Press, Oxford. wasps. Wasps were less aggressive toward pairs ZAR, lH. 1996. Biostatistical analysis. 3n1 edition. Prentice­ ofbeetles than toward individuals, so that paired Hall, Inc., Upper Saddle Rivel; NJ. 662 pp. females were able to feed more efficiently than unpaired females. Thus, the advantage of Received 31 March 1998 Accepted 29 June 1998 mating with a larger male is countered by the