nestmate relatedness in a communal bee, perdita texana ......described elsewhere (neff and danforth...

Nestmate Relatedness in a Communal Bee, Perdita texana (Hymenoptera: Andrenidae), Based onDNA FingerprintingAuthor(s): Bryan N. Danforth, John L. Neff and Percival Barretto-KoSource: Evolution, Vol. 50, No. 1 (Feb., 1996), pp. 276-284Published by: Society for the Study of EvolutionStable URL: http://www.jstor.org/stable/2410799 .

Accessed: 12/09/2013 11:00

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Society for the Study of Evolution is collaborating with JSTOR to digitize, preserve and extend access toEvolution.

http://www.jstor.org

This content downloaded from 128.253.38.173 on Thu, 12 Sep 2013 11:00:55 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ssevol

http://www.jstor.org/stable/2410799?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Evolution, 50(1), 1996, pp. 276-284

NESTMATE RELATEDNESS IN A COMMUNAL BEE, PERDITA TEXANA (HYMENOPTIERA: ANDRENIDAE), BASED ON DNA FINGERPRINTING

BRYAN N. DANFORTH1"4, JOHN L. NEFF2, AND PERCIVAL BARRETTO-KO3 IDepartment of Entomology, Comstock Hall, Cornell University, Ithaca, New York 14853

E-mail: [email protected] 2Central Texas Melittological Institute, 7307 Running Rope, Austin, Texas 78731

3Genome Variation Analysis Facility, 233 Biotechnology Building, Cornell University, Ithaca, New York 14853

Abstract.-Perdita texana is a facultatively communal bee species with up to 28 females per nest. We used multilocus DNA fingerprinting to test the hypothesis that nestmates are more closely related to each other than are nonnestmates. The mean band sharing proportion among pairwise nestmate comparisons did not differ significantly from the mean among nonnestmate comparisons [P = 0.787 (df = 484)]. Although mean band sharing proportions did not differ among nestmates and nonnestmates, some nestmates show very high band sharing proportions (in excess of the upper 95% confidence limit for the nonnestmate mean). These individuals almost certainly are related, probably as half-sib sisters, however, they comprise a very small percentage of the nestmate populations. Our results indicate that kin selection is unlikely to play an important role in the evolution and maintenance of communal nesting. Communal societies most likely arise because of the mutualistic benefits of cooperative nesting, including accelerated nest founding and improved nest defense.

Key words.-DNA fingerprinting, kin-selection, nestmates, Perdita texana, sociality.

Received March 29, 1994. Accepted September 19, 1994.

In social organisms, the genetic relatedness among inter- acting individuals may have important fitness implications. Hamilton (1964a,b) introduced the concept of inclusive fitness in order to incorporate the combined effects of individual reproduction and the reproduction of related individuals. An individual's overall fitness includes both a direct effect (Awj), the change in fitness resulting from its own actions, and an indirect effect

[E (r, X AWob the accumulated fitness effects of its behavior on the fitness of other individuals weighted by the coefficient of relatedness, r,y (Ross and Carpenter 1991). Behavioral interactions between two individuals (x and y) should be favored by selection when x suffers no decrease in individual fitness that is not more than offset by an increase in indirect fitness gained through y:

A wx + E (rxy X A wy) > 0.

Stated in this way, the concept of inclusive fitness may be applied to all forms of behavioral interactions, from minimal cooperation to the evolution of sterile workers.

Communality, or the sharing of a nest by several to many reproductive females with each female constructing and pro- visioning her own cells and laying eggs, is a form of social organization that has received little attention by students of social evolution. Communal nesting appears to be a stable form of social behavior that is phenetically but not evolu- tionarily intermediate between solitary and eusocial behavior (Kukuk and Eickwort 1987). In light of Hamilton's inequality, the existence of apparently stable communal associations of female Hymenoptera raises a number of evolutionary ques- tions, most importantly, what roles do nestmate relatedness and kin selection play in determining the behavioral interactions among communal nestmates?

4 Corresponding author.

Communal associations among nestmates are known to occur in diverse lineages of bees, including the Hylaeinae, Nomiinae (Wcislo 1993), Halictinae [especially the Agapos- temon complex and Australian species (Knerer and Schwarz 1976)], Andrenidae, Oxaeidae, Anthophoridae (especially Exomalopsini, Eucerini and Xylocopinae) and Megachilidae (Michener 1974; Eickwort 1981), as well as some sphecid (Crabroninae and Philanthinae; Evans et al. 1976; Evans and Hook 1986), masarine (Zucchi et al. 1976), and pompilid (Wcislo et al. 1988) wasps. Although most studies infer the existence of communal behavior from the presence of developed ovaries in all females sharing a nest, only two studies have confirmed the reproductive status of nestmates using artificial observation nests (Kukuk and Schwarz 1987, 1988; Danforth 1991 a,b).

Although communal insects appear to present some precursors of more highly social forms of behavior, such as nest sharing, nestmate relatedness is usually thought to be too low to lead to significant indirect fitness effects. Abrams and Eickwort (1981), for example, inferred low levels of nestmate relatedness in Agapostemon virescens from the high rate of nest switching by females (> 70% of adults switched nests in some populations). The mutual advantages of shared nest construction (Eickwort 1981; Michener 1974) or shared nest defense (Abrams and Eickwort 1981) are thought to be the primary factors favoring communal associations and are not predicated on the assumption that nestmates are close kin (Lin and Michener 1972).

To evaluate the potential role of kin selection in the maintenance of communal bee societies we measured relatedness in a communal andrenid, Perdita texana, using multilocus DNA fingerprinting. DNA fingerprinting is an effective method for estimating relatedness (Lynch 1988; Reeve et al. 1992) and has been applied to studies of humans (Chakraborty and Jin 1993), naked mole rats (Reeve et al. 1990), wolves (Leh- man et al. 1992), condors (Geyer et al. 1993), wasps (Pfennig

276

C) 1996 The Society for the Study of Evolution. All rights reserved.



NESTMATE RELATEDNESS IN P. TEXANA 277

and Reeve 1993), and bees (Blanchetot 1991, 1992; Mueller et al. 1994).

The nesting and mating behavior of P. texana have been described elsewhere (Neff and Danforth 1992; Danforth and Neff 1992) but can be summarized as follows. Perdita texana is a ground nesting, univoltine, facultatively communal bee species found in western Texas and southern Oklahoma. As many as 28 females may share a single nest, but most nests have no more than five females, and approximately 25% of the nests are occupied by a single female. There is no apparent reproductive division of labor among nestmates, as all resi- dents of communal nests forage and have developed ovaries. Nests consist of a single main tunnel and several laterals leading to brood cells. Females provision a single cell per day with pollen and nectar collected exclusively from Opun- tia (Platyopuntia) spp. Mating takes place primarly on flow- ers; however, some males enter the nests of females at the end of the foraging period and remain in the nest until the following morning, suggesting that mating may take place within nests.

MATERIALS AND METHODS

Specimens for this study were collected at Pedernales Falls State Park, Blanco County, Texas, on May 26 and 27, 1993. All nests were located along a dirt road and nests were separated by no more than 5 m. Individual females were captured in plastic cups as they departed nests and were transferred to labeled vials and placed on ice. The data presented below are based on a total of 76 females from 13 nests. The distribution of nest sizes from this collection are shown in Figure lB along with the presumed "natural" distribution of nest sizes (Fig. IA) obtained from nest excavations in 1990.

Prior to DNA extraction, the gut was removed from all females by dissection to eliminate possible contamination from cactus DNA. Following dissection, specimens were briefly frozen in liquid nitrogen and ground in individual 1.5 ml Eppendorf tubes in the presence of 2x CTAB extraction buffer (Saghai-Maroof et al. 1984) and 100 [Lg of proteinase K. Specimens were incubated for 2 h at 55?C and then ex- tracted with chloroform-isoamylalchohol, phenol-chloroform-isoamylalchohol, and chloroform-isoamylalchohol, in that order. The DNA was precipitated with 2.5 volumes of ice cold ethanol and 1/10 volume 3M sodium acetate, washed once in 80% ethanol and resuspended in 50 [L Tris-EDTA (pH 7.6) buffer. DNA extractions yielded from 1.43 to 6.58 pLg DNA [x = 3.01 + 0.193 pLg (n = 30)] for males and from 1.43 to 9.2 pLg DNA [x = 4.72 + 0.256 pLg (n = 32)] for females.

To generate multilocus fingerprints, total DNA from individual females was digested with 20 units of restriction enzyme (see below) in the presence of 2.5 mM spermidine and 1 pLg RNase in a total reaction volume of 150 [L for roughly 6 h and then concentrated using protocols in Aquadro et al. (1992). One third of the sample (roughly 1.5 pLg DNA) was loaded on a 0.8% agarose gel and run at 40 V for 32 h in 1 x TAE gel buffer (Sambrook et al. 1989). DNA was transferred to Schleicher and Schuell Nytran+ (0.45 [Lm) nylon membranes as described in Aquadro et al. (1992) and

1 I I I I I I I I I I I I I I I

A. 8

c 6

E 4-

z

2

10 d 3 6 8 11111

8

4- 0

z

2

Number of females per nest

FIG. 1. A. Frequency distribution of nest sizes (number of females per nest) based on excavations in Coleman, Texas in 1990. B. Fre- quency distribution of nest sizes (number of females per nest) used in the present study.

baked at 80?C for 30-45 min, as recommended by the manufacturer.

For chemiluminescent detection we used the Boehringer Mannheim Genius System and followed protocols in the Gen- ius System Users Guide to Filter Hybridization (version 2.0), except where indicated below. Probe labeling was performed using polymerase chain reaction (PCR) amplification as described in Emanuel (1991) and Lanzillo (1990). PCR-labeling gives improved probe sensitivity as compared to random primed labeling and gave sharper bands with lower background in our study. For the M13 probe, we used two primers (5'-CCCTTATTAGCGTTTGCCAT-3' and 5'-GGTACATG- GGTTCCTATT-3') to amplify a 643-bp fragment including the two tandem repeat regions of the M13 gene III (Vassart et al. 1987). For the Jeffreys 33.6 and 33.15 probes (Jeffreys et al. 1985a,b), we used plasmid clones (pSPT 19.6 and 19.15) containing the Jeffreys core sequences (Carter et al. 1989) kindly provided by Dr. C. F Aquadro for amplification and labeling. We used primers homologous to the SP6 (5'-CA-



278 BRYAN DANFORTH ET AL.

TACGATTTAGGTGACACTATAG-3') and T7 (5 '-TAA- TACGACTCACTATAGGGAGA-3') promoter regions flank- ing the polycloning site. The pSPT 19.6 yields an approximately 800 bp fragment and the pSPT 19.15 yields an approximately 650 bp fragment. We followed the PCR labeling protocol of Emanuel (1991 ) for a 100 ,ul reaction, except that for the Jeffreys probe labeling reactions we lowered the final MgCl2 concentration to 1.0 mM. Template DNA was linear- ized prior to use (with EcoRI for M13mpl8 RF DNA and with Scal for the pSPT plasmid DNAs) and 1 ng was used per reaction. Cycle conditions were as follows: 20-25 cycles at 94?C for 45 s, 55?C for 45 s, and 72?C for 1 min. The yield of labeled probe was determined by the Boehringer Mannheim probe labeling assay. We typically generate 500 ng of labeled probe per 1-ng template DNA. For a typical hybridization volume of 15 ml we add 10 ,ul of the probe labeling reaction (roughly 5 ng/ml final probe concentration).

Prehybridization and hybridization were performed in a Hybaid oven at 55-57?C for 24 h each following protocols in Aquadro et al. (1992). Membranes were washed twice in 2x wash (0.3M NaCl, 3mM sodium citrate, 0.1% SDS) at room temperature, once in 2X wash at 55?C and finally in 0.5)X wash at room temperature. Long blocking (2-3 h) dur- ing the detection stage was found to significantly reduce background problems. Following exposure of the membrane to antidigoxigenin alkaline phosphatase (diluted 1:10,000 in Genius buffer 2) for 30 min, we applied Lumi-phos 530 diluted 1:10 in Genius buffer 3 (contrary to manufacturer's recommendations). Ten to 20 ml of diluted Lumi-phos 530 was sufficient to completely cover a 20 X 26 cm membrane, and several membranes could be placed sequentially into a plexiglass tray containing the diluted Lumi-phos 530.

Following the application of Lumi-phos 530, membranes were placed into heavy-gauge acetate sheet protectors (Joshua Meiers cat. no. 06198) for approximately 12 h prior to exposure to X-ray film in order to allow the photogenic reaction to reach its maximum intensity. Exposures were typically 15- 45 min.

An initial screening of four-cutter restriction enzymes (RsaI, Alul, HinfI, HaeIII, and Sau3AI) and three probes (M13, Jeffreys 33.6, and 33.15) indicated that HaeIII digests in combination with the 33.15 probe gave the maximum number of sharp, scorable bands, and the results presented below are based on this enzyme/probe combination (Fig. 2).

A number of precautions were taken to reduce the subjec- tivity inherent in scoring multilocus DNA fingerprints. First, individuals were randomly arranged on gels such that the person scoring the gel was unaware of which lanes represented nestmate comparisons and which nonnestmate comparisons. Second, at least one duplicate individual was in- cluded on each gel (in nonadjacent lanes) to evaluate the error rate in scoring. Third, approximately 50 pg of Boeh- ringer Mannheim Molecular Weight Marker X (cat. no. 1498 037) was added to each sample prior to running the gel. This molecular weight marker contained a range of fragment sizes from 100 bp to over 12 kb with intermediate fragment sizes differing by roughly 1 kb. Following the initial probing with the Jeffries 19.15 probe, membranes were washed twice in 0.1% SDS heated to 95?C (Bruford et al. 1992) to remove bound probe and reprobed with random-prime labeled mo-

1 2 3 4 5 6 7 8 9101112

<- 12216 <-11198i g <-- 1 018

4 <--~~~9162

b~~~~~~ ~<- 61 08 ...l..._ <712

< 2036

FIG. 2. Multilocus DNA fingerprint of Perdita texana. Two nests are represented on the gel: bees from nest 7 are in lanes 1, 6, 7, 9, and 11; bees from nest 1 are in lanes 2, 4, 5 8, and 12. Lanes 3 and 10 are replicate lanes of a bee collected from another nest. Roughly 1.5 pRg DNA was loaded per lane, 30 min exposure. Bands were scored in the size range from 3 to 12 kb. The locations of the Boehringer Mannheim molecular weight marker bands are shown to the right of the gel (numbers indicate fragment size in base pairs).

lecular weight marker X. The resulting image could be placed beneath the fingerprint for scoring. Using the molecular weight marker improved our ability to assess whether bands of similar mobility in widely separated lanes were indeed the same bands. A similar method was described by Jones et al. (1991) but with a lambda Hindlll molecular weight marker.

RESULTS

We scored from 66 to 89 (x = 79.17 ?3.89, N = 6) variable bands per gel ranging in size from 2.5 to 12 kb (Fig. 2). On average, individual bees showed 25.84 ? 0.53 (N = 79)




TABLE 1. Mean band sharing for nonnestmate and nestmate comparisons. Sample size, N, refers to the total number of pairwise comparisons made.

Gel number sl, (nonnestmates) ? SEM (N) sl (nestmates) ? SEM (N) P (df)

1 0.517 0.016 (75) 0.552 ? 0.025 (30) 0.232 (103) 2 0.485 ? 0.016 (63) 0.494 ? 0.030 (28) 0.770 (89) 3 0.492 ? 0.017 (60) 0.478 ? 0.019 (31) 0.615 (89) 4 0.544 ? 0.022 (43) 0.535 ? 0.033 (23) 0.817 (64) 5 0.452 ? 0.021 (47) 0.461 ? 0.025 (31) 0.794 (76) 6 0.464 ? 0.016 (35) 0.467 + 0.019 (20) 0.902 (53) Overall 0.494 + 0.008 (323) 0.498 + 0.011 (163) 0.787 (484)

bands. The similarity index, or band sharing probability, sxy, was calculated for all possible pairwise comparisons on each of the six gels. Sxy is twice the number of bands shared by two individuals divided by the sum of the number of bands scored in each individual (Burke and Bruford 1987; Burke 1989; Gibbs et al. 1990). Table 1 shows the mean band sharing probabilities for nonnestmate and nestmate comparisons broken down by gel and the overall means. Figure 3 shows the overall distributions. Nestmates and nonnestmates do not differ in mean band-sharing probability. In fact, the band sharing probabilities are almost indistinguishable. When the data are broken down by gel the same pattern emerges; there is no significant difference in mean band-sharing between nestmates and nonnestmates on any of the six gels (Table 1).

The band-sharing data shown in Figure 3 violate at least two assumptions of parametric hypothesis testing. First, nei- ther the nestmate (g1 = 0.873; P < 0.001) nor the nonnestmate (gj = 0.734; P < 0.001) distributions are normally distributed. Second, because the band sharing data include all possible pairwise comparisons on each gel, our data violate the assumption of independence among variates (Lynch 1990, 1991). Interdependence arises because the value of Sxy derived from a comparison of individuals X and Y is not independent of the value of syz derived from a comparison of individuals Y and Z, because both data points share a common individual, Y.

To correct for the combined problems of nonnormality and nonindependence, we applied a permutation test to the data (Lehman 1986, p. 230). In the permutation test, we randomly sampled one set of 163 Sxy values (call this the "pseudo- nestmate" sample) and another set of 323 sxy values (call this the "pseudononnestmate" sample) from the data set without regard to the true nature of the sxy values. We calculated the mean of each of these samples and subtracted the "pseu- donestmate" mean from the "pseudononnestmate" mean, giving a value D, the deviation between the two means. This procedure is repeated 1000 times to generate a distribution of D values with a mean of 0 and some standard deviation. We then used this distribution to test the hypothesis that our observed D value (0.498 - 0.494 = 0.004) deviates significantly from the mean, 0. The permutation test solves the problem of nonindependence because it is based on deviations between means. Means, unlike variances, should not be affected by interdependence among all possible pairwise comparisons.

Figure 4 shows the results of the permutation test. The observed difference between the nestmate mean and the nonnestmate mean does not differ significantly from 0 (P =

Non-nestmates 0.20 - 0.494? O.08 (323)

o 015 02 04 0. . .

U 0.10

0.05

0.0 0.2 0.4 0.6 0.8 1.0 Band sharing probability

0.25 n estmates 0498 0.011 (1l63)

0.20 -

0.15- C)

LL 0.10

0.05

0.0 0.2 0.4 0.6 0.8 1.0 Band shanng probability

FIG. 3. Frequency distributions of band-sharing probability for nonnestmates and nestmates. Overall data are shown. Dashed vertical lines indicate 95% confidence limits based on the standard deviation among nonnestmate band-sharing probabilities (lower limit =0.2294; upper limit =0.7586).




Ob$erved Mean = 0.000 / ldeviation = 0.004 std dev.=0.013 p076n

0.20 - n=1000

One-tailei threshold o . - g 0.15 of signific ance (D=0.0215)

Two-tail d threshold of signi cance (D=0.0255)

LJL 0.10

0.05

-0.045 -0.025 -0.005 0.015 0.035 0.055 D value

FIG. 4. Distribution of 1000 D-values from permutation test. See text for discussion.

0.7566) and falls far short of the one-tailed threshold of significance.

Table 2 presents the mean band sharing probabilities for each nest. In two cases, individuals from the same nest were run on two different gels (nests 3 and 9) because these were the largest nests (14 individuals per nest) and dividing them between gels gave a more even balance of nonnestmate and nestmate comparisons. The mean band sharing among individuals in the same nest run on different gels did not differ significantly from each other (nest 3: t = 1.459, df = 40, P

TABLE 2. Mean band sharing among nestmates by nest. The last column indicates the number of observations that fall in the upper 5% of the nonnestmate distribution (i.e., that fall above the upper 95% confidence interval based on a one-tailed test). The probabilities of obtaining the observed number of values in excess of the upper 95% confidence interval are shown in parentheses.

Number observed > Nest number s, + SEM (N) upper 95% CI

1 0.440 0.023 (10) 0 2 0.474 + 0.038 (21) 4 (P = 0.087) 3 0.473 0.021 (42) 5 (P = 0.311) 3 (gel 3) 0.504 + 0.025 (21) 3 (gel 5) 0.443 + 0.034 (21) 5 0.615 (1) 0 6 0.425 0.023 (10) 0 7 0.495 + 0.028 (10) 0 8 0.542 0.019 (6) 0 9 0.547 + 0.021 (42) 6 (P = 0.156) 9 (gel 1) 0.562 + 0.026 (21) 9 (gel 4) 0.532 ? 0.034 (21) 10 0.724 (1) 1 (P = 0.084) 15 0.500 ? 0.031 (10) 0 16 0.400 (1) 0 18 0.468 ? 0.058 (6) 0 24 0.657 ? 0.130 (3) 1 (P = 0.25 1)

Total 17

= 0.152; nest 9: t = 0.693, df = 40, P = 0.492), as one would expect if each gel contained a representative sample of individuals.

What is more interesting than the overall mean band sharing among nestmates is the distribution of band sharing probabilities for each nest (Fig. 5). In roughly one-third of all nests, the distribution of sxy appears bimodal (nests 2, 3, 9, and 18), and in some nests (nests 2, 3, 9, and 24), there are pairs of individuals whose band-sharing probability exceeds the upper 95% confidence limit on nonnestmate comparisons (Table 2).

To test the hypothesis that the distributions shown in Figure 5 have significantly higher variance in sxy than one would expect by chance, we applied a bootstrap approach in which we randomly sampled n observations from the data set of nestmate sxy values, where n corresponds to the actual number of pairwise comparisons within nests (cf. Table 2). The standard deviation in sxy was calculated for each subsample, and the procedure was repeated 1000 times with replacement. These samplings generated normal distributions of 1000 standard deviations in sxy for various sample sizes, ranging from n = 3 (nest 24) to n = 42 (nests 3 and 9, Table 2). Using these normal distributions, we calculated an upper 95% confidence limit with which to compare our observed standard deviations. Only nest 2 showed a standard deviation significantly higher than one would expect by chance in a one- tailed test of significance (P < 0.05).

In a second bootstrap analysis, we attempted to test the hypothesis that the distributions shown in Figure 5 were skewed to the right such that significantly more values fell above the upper 95% confidence interval for nonnestmate sxy values than one would expect by chance. In this analysis, we randomly sampled n observations from the data set of nonnestmate sxy values, where n corresponds to the actual number of pairwise comparisons within nests (Table 2). The number




Nesti I Nest 2 03 3 0 -

Nest 15 : Nest 18 :~~~~015

020.2

LL L 01

0.05

0.3-

FIG. 5. Freqency distriutions ofbnd-shain 3 rbblt foniiulnests 6ahdvria ie niaeupr9%cniec

~~~~~~0.2~~~~~~~~~~~~.

LL~~~~~~~~~~~L 0.1~~~~~~~~~~~0

Nest 7I 0.3 02-Ns9

c~~~~~~~~~~~~~~> 025 0)2

LL L 0.1

0.4-

limit~~~~~~~~~~~Ns base ontesadrNeito mn onetaebn-hrn rbbltest (uprlmt=076)1Masaeidctdwt

0.20 -

03

0.16 I ci~c

1-'-oio a) LL

005

0.0 0.2 0.4A 0.'6 0.8 1.0 00 02 04 06 08 10o

Band sharing probability Band sharing probability

FIG. 5. Frequency distributions of band-sharing probability for individual nests. Dashed vertical lines indicate upper 95% confidence limit based on the standard deviation among nonnestmate band-sharing probabilities (upper limit =0.7161). Means are indicated with arrow heads along x-axis.

of observations that fell above the upper 95% confidence interval based on the overall nonnestmate data set (0.7161) was tallied, and the procedure repeated 1000 times. These bootstrap samplings generated a normal distribution for the number of sxy values one could expect by chance to fall above the 95% confidence interval. The results of this test are shown in the last column of Table 2. None of the nests deviated significantly from expectation (Table 2).

These results indicate that, although there are some individuals within nests that appear to be close relatives, most

of the high sxy values among nestmates can be accounted for by chance alone. Nest 2 (Fig. 5) is the exception and indicates that at least some nests contain relatives. The presence of related females within nests most likely arises when sisters reuse their natal nest. The number of nonrelated "joiners"; however, swamps out any potential for elevated mean relatedness among nestmates.

There is no significant relationship between mean band- sharing probability among nestmates and nest size (r = 0.229; n = 13, P > 0.05).




One can use the mean band sharing probabilities among nestmates and nonnestmates to calculate an estimate of the mean relatedness among nestmates (Reeve et al. 1992). If mean band-sharing among nonnestmates accurately reflects mean band sharing among nonrelatives, we estimate mean relatedness among nestmates to be 0.0079 [r = (Snestmates -

Snonnestmates)( 1 - Snonnestmates) = (0.498 - 0.494)/(1 -

0.494)]. We can use the results of the permutation test along with

the above equation to calculate the average relatedness among nestmates that we would have detected as statistically significant given our data set. The permution test indicates that a deviation of 0.0215 from the mean of nonnestmate comparisons would have been significant in a one-tailed test (Fig. 4). This corresponds to an observed mean band sharing among nestmates of 0.5155 or a mean relatedness among nestmates (using Reeve et al. 1992) of 0.0425. This value falls midway between the relatedness among first cousins (0.0625) and second cousins (0.0156; relatedness values based on the assumption that females mate multiply).

DISCUSSION

To the extent that these results hold for other communal bees, it would seem that kin selection or any indirect fitness benefits associated with nest sharing play a minimal role in the evolution or maintenance of communal societies. Similar results have been obtained in a study of Australian communal halictids (Kukuk and Sage 1994). Kukuk and Sage used allo- zyme data to calculate nestmate relatedness in Lasioglossum (Chilalictus) hemichalceum, a communal species with up to 20 adult females per nest. Their results, based on two poly- morphic loci, indicated that nestmate relatedness is indistinguishable from zero in reproductively active colonies and was only slightly greater than zero (0.13) in a larger data set of colonies in various stages of maturity.

Our results are consistent with quantitative models for the maintenance of egalitarian societies (Vehrencamp 1983a,b; Keller and Reeve 1994). These models relate the degree of reproductive skew (the distribution of direct reproduction among colony members) to a number of factors including ecological constraints to solitary nesting, colony productivity, and relatedness among colony members. One counter- intuitive prediction that these models make is that reproductive skew will be positively correlated with relatedness among colony members. This is because when relatedness between dominant and subordinate individuals is low, subordinate individuals have little to gain, in terms of indirect fitness, in societies in which they cannot share directly in reproduction. Under conditions of low relatedness and high skew, subordinates may simply choose to leave the colony. Ecological constraints may limit their ability to found new nests but, at least in P. texana, joining other, previously es- tablished nests is almost certainly an option.

How would low levels of nestmate relatedness arise? Based on previous behavioral studies (Neff and Danforth 1992), it was unclear whether nests were founded at the beginning of each season primarily through the reuse of natal nests, which would tend to give non-0 relatedness values, or primarily through dispersal of females from their natal nest and sub-

sequent cofounding of new nests, which would give low to 0 values of relatedness. In a species such as P. texana, in which females mate multiply throughout their lives and even throughout a single foraging trip (Neff and Danforth 1992), sisters are expected to be half sibs and therefore have an average relatedness of 0.25. We assumed at the outset that the maximum level of nestmate relatedness we could obtain would be 0.25 (a single-female nest giving rise to half-sib sisters who remain in the natal nest the following season, with no unrelated females admitted into the nests). Given that mean band sharing among nonnestmates in our study was 0.494, one can use Reeve et al.'s (1992) equation relating band-sharing values to relatedness to estimate the expected band-sharing probability among half-sib sisters to be 0.6205. Although mean band sharing among nestmates (0.498) is con- siderably below this value, Figure 5 reveals that in several nests there are groups of individuals with band-sharing values equal to or in excess of 0.6205 (e.g., nests 2, 3, 7, 9, and 18) suggesting that natal nest reuse, although not common, does occur. We suspect that the observed low level of nestmate relatedness arises primarily through dispersal of females from their natal nests at the start of each season.

What factors favor cooperative nest initiation by unrelated females? In other words, what are the potential ecological constraints to solitary nest founding? There are at least three reasons why cooperative nest founding, even by unrelated individuals, may be advantageous. First, female P. texana, like many other communal bees, have a very short season of adult activity. In P. texana, oligolectic foraging on Opuntia places severe time constraints on new nest founding. Opuntia plants bloom in central Texas for 2 to 3 wk and female reproductive success may be partly dependent on founding a nest rapidly. The soil at all sites studied is typically very hard and rapid construction of the main burrow system may be facilitated by cooperation among several females.

Second, multifemale nests may benefit from improved nest defense against parasites and predators. Neff and Danforth (1992) observed cleptoparasitic Sphecodes manni (Halicti- dae) entering nests of P. texana at Pedernales Falls State Park, although found no evidence of successful parasitism in nest excavations. However, levels of parasitism may vary consid- erably from site to site, and we cannot exclude parasitism as an important factor in favoring communal behavior.

Finally, communal nesting may be favored by what Gad- agkar (1990, p. 17) termed "the advantage of assured fitness returns." In multifemale nests a female who dies before the end of the active flight season may benefit from the presence of unrelated nestmates if they prevent the nest from being invaded by predators following her death. Although Gad- agkar's assured fitness hypothesis probably applies best to vespid wasps, such as Polistes, which build aboveground, exposed nests, his reasoning could be applied to subterranean nests where even closed cells may be vulnerable to digging predators, such as ants, or parasites, such as mutillid wasps.

The above hypotheses all invoke a mutualistic benefit of cooperative nesting. The importance of cooperative nest defense has been emphasized by other authors (cf. Lin and Michener 1972), but in the context of the origins of eusociality. Increasingly it is becoming evident that communal lineages of bees and wasps do not give rise to eusocial species




(e.g., Kukuk and Eickwort 1987; Alexander et al. 199 1; Pack- er 1993). Eusociality appears to arise far more commonly from solitary or semisocial precursors favoring the "famil- ial" route to eusociality over the parasocial route (Lin and Michener 1972). That communal nesting does not appear to give rise to eusociality suggests that communal behavior is not an intermediate form of social organization, but instead represents a stable alternative form of social behavior primarily favored by mutualistic interactions among nonrelated individuals.

ACKNOWLEDGMENTS

This research was conducted in the Genome Variation Analysis Facility, B. May, Director. We are grateful to C. F Aquadro, S. Bogdanowicz, and C. R. Freeman for advice on all aspects of this research, B. Sawyer for technical help, and G. Churchill, Biometrics Unit, Cornell University, for advice on statistical analysis of DNA fingerprinting data. G. C. Eick- wort, U. G. Mueller, C. R. Freeman, W. T. Wcislo, and B. May provided helpful comments on earlier versions of the manuscript. This research was entirely supported by a Na- tional Science Foundation Postdoctoral Research Fellowship in Environmental Biology to B.N.D. (DEB-9201921).

LITERATURE CITED

Abrams, J., and G. C. Eickwort. 1981. Nest switching and guarding by the communal sweat bee Agapostemon virescens (Hymenop- tera, Halictidae). Insectes Sociaux 28:105-116.

Alexander, R. D., K. M. Noonan, and B. J. Crespi. 1991. The evolution of eusociality. Pp. 3-44 in P. W. Sherman, J. U. M. Jarvis, and R. D. Alexander, eds. The biology of the naked mole rat. Princeton University Press, Princeton, NJ.

Aquadro, C. F, W. A. Noon, and D. J. Begun. 1992. RFLP analysis using heterologous probes. Pp. 115-157 in A. R. Hoelzel, ed. Molecular genetic analysis of populations. IRL Press, Oxford University Press, Oxford.

Blanchetot, A. 1991. Genetic relatedness in honey bees as estab- lished by DNA fingerprinting. Journal of Heredity 82:391-396.

1992. DNA fingerprinting analysis in the solitary bee Megachile rotundata: Variability and nest mate genetic relationships. Genome 35:681-688.

Bruford, M. W., 0. Hanotte, J. F. Y. Brookfield, and T. Burke. 1992. Single-locus and multilocus DNA fingerprinting. Pp. 225-269 in A. R. Hoelzel, ed. Molecular genetic analysis of populations. IRL Press, Oxford University Press, Oxford.

Burke, T 1989. DNA fingerprinting and other methods for the study of mating success. Trends in Ecology and Evolution 4:139-144.

Burke, T., and M. W. Bruford. 1987. DNA fingerprinting in birds. Nature 327:149-152.

Carter, R. E., J. H. Wetton, and D. T. Parkin. 1989. Improved genetic fingerprinting using RNA probes. Nucleic Acids Re- search 17:5867.

Chakraborty, R., and L. Jin. 1993. Determination of relatedness between individuals by DNA fingerprinting. Human Biology 65: 875-895.

Danforth, B. N. 199 la. Female foraging and intra-nest behavior of a communal bee, Perdita portalis Timberlake (Hymenoptera: An- drenidae). Annals of the Entomological Society of America 84: 537-548.

. 1991b. The morphology and behavior of dimorphic males in Perdita portalis (Hymenoptera: Andrenidae). Behavioral Ecology and Sociobiology 29:235-247.

Danforth, B. N., and J. L. Neff. 1992. Male polymorphism and polyethism in Perdita texana (Hymenoptera: Andrenidae). An- nals of the Entomological Society of America 85:616-626.

Eickwort, G. C. 1981. Presocial insects. Pp. 199-280 in H. R. Hermann, eds. Social insects, Vol. 2. Academic Press, New York.

Emanuel, J. R. 1991. Simple and efficient system for synthesis of non-radioactive nucleic acid hybridization probes using PCR. Nucleic Acids Research 19:2790.

Evans, H. E., and A. W. Hook. 1986. Nesting behavior of Australian Cerceris digger wasps, with special emphasis on nest reutili- zation and nest sharing. Sociobiology 11:275-302.

Evans, H. E., R. W. Mathews, J. Alcock, and M. A. Fritz. 1976. Notes on the nests and prey of two subspecies of Cerceris ru- fimana Taschberg (Hymenoptera: Sphecidae: Cercerini). Journal of the Kansas Entomological Society 49:126-132.

Gadagkar, R. 1990. Evolution of eusociality: the advantage of assured fitness returns. Philosophical Transactions of the Royal Society of London, Biological Sciences 329:17-25.

Geyer, C. J., 0. A. Ryder, L. G. Chemnick, E. A Thompson. 1993. Analysis of relatedness in California condors from DNA fingerprints. Molecular Biology and Evolution 10:571-589.

Gibbs, H. L., P. J. Weatherhead, P T. Boag, B. N. White, L. M. Tabak, D. J. Hoysak. 1990. Realized reproductive success of polygynous red-winged blackbirds revealed by DNA markers. Science 250:1394-1397.

Hamilton, W. D. 1964a. The genetical evolution of social behavior, I. J. Theoretical and Experimental Biology 7:1-16.

. 1964b. The genetical evolution of social behavior, II. The- oretical and Experimental Biology 7:17-52.

Jeffreys, A. J., V. Wilson, and S. L. Thein. 1985a. Hypervariable "minisatellite" regions in human DNA. Nature 314:67-73.

1985b. Individual-specific "fingerprints" of human DNA. Nature 316:76-79.

Jones, C. S., C. M. Lessells, and J. R. Krebs. 1991. Helpers-at- the-nest in European bee-eaters (Merops apiaster): a genetic analysis. Pp. 169-192 in T. Burke, G. Dolf, A. J. Jeffreys, and R. Wolff, eds. DNA fingerprinting: Approaches and applications, Birkhauser Verlag, Basel, Switzerland.

Keller, L., and H. K. Reeve. 1994. Partitioning of reproduction in animal societies. Trends in Ecology and Evolution 9:98-102.

Knerer, G., and M. Schwarz. 1976. Halictine social evolution: the Australian enigma. Science 194:445-448.

Kukuk, P. E, and G. C. Eickwort. 1987. Alternative social structures in halictine bees. Pp. 555-556 in J. Eder and H. Rembold, eds. Chemistry and biology of social insects. Peperny, Mtinchen, Germany.

Kukuk, P. E, and G. K. Sage. 1994. Reproductivity and relatedness in a communal halicine bee, Lasioglossum (Chilalictus) hemichalceum. Insectes Sociaux 41:443-455.

Kukuk, P. E, and M. Schwarz. 1987. Intranest behavior of the communal sweat bee Lasioglossum (Chilalictus) erythrurum (Hy- menoptera: Halictidae). Journal of the Kansas Entomological Society 60:58-64.

. 1988. Macrocephalic male bees as functional reproduc- tives and probable guards. Pan-Pacific Entomologist 64:131- 137.

Lanzillo, J. J. 1990. Preparation of digoxigenin-labelled probes by the polymerase chain reaction. BioTechniques 8:621-622.

Lehman, E. L. 1986. Testing statistical hypotheses. 2d ed. Wiley, New York.

Lehman, N., P. Clarkson, L. D. Mech, T. J. Meier, and R. K. Wayne. 1992. A study of the genetic relationships within and among wolf packs using DNA fingerprinting and mitochondrial DNA. Behavioral Ecology and Sociobiology 30:83-94.

Lin, N., and C. D. Michener. 1972. Evolution of sociality in insects. Quarterly Review of Biology 47:131-159.

Lynch, M. 1988. Estmation of relatedness by DNA fingerprinting. Molecular Biology and Evolution 5:584-599.

. 1990. The similarity index in DNA fingerprinting. Mo- lecular Biology and Evolution 7:478-484.

. 1991. Analysis of population genetic structure by DNA fingerprinting. Pp. 113-126 in T. Burke, G. Dolf, A. J. Jeffreys, and R. Wolff, eds. DNA Fingerprinting: Approaches and applications. Birkhauser Verlag, Basel, Switzerland.

Michener, C. D. 1974. The social behavior of the bees. Harvard University Press, Cambridge, MA.



284 BRYAN DANFORTH ET A'

Mueller, U. G., G. C. Eickwort, and C. F. Aquadro. 1994. DNA fingerprinting analysis of parent-offspring conflict in a bee. Pro- ceedings of the National Academy of Science, USA 91:5143- 5147.

Neff, J. L., and B. N. Danforth. 1992 (1991). The nesting and foraging behavior of Perdita texana Cresson (Hymenoptera: An- drenidae). Journal of the Kansas Entomological Society 64:394- 405

Packer, L. 1993. Multiple-foundress associations in sweat bees. Pp. 215-233 in L. Keller, ed. Queen number and sociality in insects. Oxford University Press, Oxford.

Pfennig, D. W., and H. K. Reeve. 1993. Nepotism in a solitary wasp as revealed by DNA fingerprinting. Evolution 47:700-704.

Reeve, H. K., D. E Westneat, W. A. Noon, P. W. Sherman, and C. F Aquadro. 1990. DNA "fingerprinting" reveals high levels of inbreeding in colonies of the eusocial naked mole rat. Proceed- ings of the National Academy of Science, USA 87:2496-2500.

Reeve, H. K., D. E Westneat, and D. Queller. 1992. Estimating average within-group relatedness from DNA fingerprints. Mo- lecular Ecology 1:223-232.

Ross, K. G., and J. M. Carpenter. 1991. Population genetic structure, relatedness, and breeding systems. Pp. 451-479 in K. G. Ross, and R. W. Mathews, eds. The social behavior of wasps. Cornell University Press, Ithaca, NY.

Saghai-Maroof, M. A., K. M. Soliman, R. A. Jorgensen, and R. W. Allard. 1984. Ribosomal DNA spacer-length polymorphism in barley: Mendelian inheritance, chromosomal locations, and pop-

ulation dynamics. Proceedings of the National Academy of Sci- ence, USA 81:8014-8018.

Sambrook, J., E. E Fritsch, and T. Maniatis. 1989. Molecular clon- ing, 2d ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Vassart, G., M. Georges, R. Monsieur, H. Brocas, A. S. Lequarre, and D. Christophe. 1987. A sequence in M13 phage detects hypervariable minisatellites in human and animal DNA. Science 235:683-684.

Vehrencamp, S. L. 1983a. A model for the evolution of despotic versus egalitarian societies. Animal Behavior 31:667-682.

. 1983b. Optimal degree of skew in cooperative societies. American Zoologist 23:327-335.

Wcislo, W. T 1993. Communal nesting in a North American pearly- banded bee, Nomia tetrazonata, with notes on nesting behavior of Dieunomia heteropoda (Hymenoptera: Halictidae: Nomiinae). Annals of the Entomological Society of America 86:813-821.

Wcislo, W. T., M. J. West-Eberhard, and W. G. Eberhard. 1988. Natural history and behavior of a primitively social wasp, Au- plopus semialatus, and its parasite, Irenangelus eberhardi (Hy- menoptera: Pompilidae). Journal of Insect Behavior 1:247-260.

Zucchi, R., S. Yamane, and S. F. Sakagami. 1976. Preliminary notes on the habits of Trimeria howardi, a neo ropical communal ma- sarid wasp, with description of the mature larva (Hymenoptera: Vespoidea). Insecta Matsumurana rN.S.] 8:47-57.

Corresponding Editor: J. Werren



nestmate relatedness in a communal bee, perdita texana ......described elsewhere (neff and danforth...

Documents