Foraging Populations and Distances of the Desert Subterranean Termite, Heterotermes Aureus (Isoptera: Rhinotermitidae), Associated with Structures in Southern Arizona

HOUSEHOLD AND STRUCTURAL INSECTS Foraging Populations and Distances of the Desert Subterranean Termite, Heterotermes aureus (Isoptera: Rhinotermitidae), Associated with Structures in Southern Arizona

1 2 PAUL B. BAKER AND MICHAEL I. HAVERTY

J. Econ. Entomol. 100(4): 1381Ð1390 (2007) ABSTRACT MarkÐreleaseÐrecapture studies were conducted on foraging populations of Hetero- termes aureus (Snyder) (Isoptera: Rhinotermitidae) associated with three structures in Tucson, AZ. Foraging population estimates ranged from 64,913 to 307,284 termites by using the Lincoln Index and from 75,501 to 313,251 termites using the weighted mean model. The maximum distance between monitors ranged from 26 to 65 m, with minimum total foraging distance ranging between 297 and 2,427 m. Characterizations of the cuticular hydrocarbons of foraging groups were qualitatively identical. Quantitative similarities within sites and differences among sites suggested that each site was occupied by a single colony during the sampling period. The colony at each site had a proportion of soldiers (0.135, 0.069, and 0.040) that was signiÞcantly different from the colonies at each of the other sites. From this study, we question the assumption of equal mixing of marked H. aureus foragers throughout the occupied collars around structures.

KEY WORDS colony density, colony size, cuticular hydrocarbons, markÐreleaseÐrecapture, soldier proportions

Subterranean termites have signiÞcant economic im- tion of foraging populations requires knowledge of pact worldwide. In the United States, subterranean foraging biology. However, our ability to understand termites cost consumers at least US$1.5 billion (Su and the population and foraging dynamics of H. aureus is Scheffrahn 1990). Species of Reticulitermes, Copto- restricted by its cryptic nature. Direct and indirect termes formosanus Shiraki, and Heterotermes aureus methods of sampling H. aureus in undeveloped, native (Snyder), are among the most economically impor- environments have produced asymmetrical results tant pests of structures in the mainland United States ranging from 23,000 to 300,000 individuals per colony (Su and Scheffrahn 1990, Baker and Bellamy 2006), yet on the same site near Tucson, AZ (Haverty et al. 1975, there are relatively few studies of colony demograph- Haverty and Nutting 1975, Jones 1990b). ics or foraging characteristics. With the development We report here estimates of the foraging popula- of long-lasting dyes for marking foragers (Su et al. tions and foraging distances, as well as soldier propor- 1991) or ßuorescent paint markers (Forschler 1994), tions, of colonies of H. aureus associated with struc- markÐrecapture and markÐreleaseÐrecapture meth- tures in three different locations in Tucson, AZ. We ods have been used extensively to study and monitor used a markÐreleaseÐrecapture protocol similar to that subterranean termites worldwide, but not without used to assist the development of baits for control of controversy (Thorne et al. 1996, Evans et al. 1998). subterranean termites (Su and Scheffrahn 1994, Getty However, despite drawbacks, including wide varia- et al. 2000) and characterizations of cuticular hydro- tions in the estimated number of termites, Baroni- carbons (Haverty et al. 1996) and soldier proportions Urbani et al. (1978) suggested this technique repre- to associate foraging groups within a colony. An un- sents at least a practical approach for estimating social derstanding of the magnitude of the foraging distances insect population sizes. and populations is key to effective deployment of baits In Arizona, H. aureus is the most economically im- for protection of structures in urban settings, as well portant termite pest of structures. Control by destruc- as understanding the role H. aureus plays in natural systems. 1 Corresponding author: Department of Entomology, University of Arizona, Tucson, AZ 85720 (e-mail: [email protected]). Materials and Methods 2 Chemical Ecology of Forest Insects, PaciÞc Southwest Research Station, U.S. Department of AgricultureÐForest Service, P.O. Box 245, Sites. Three structures located in Tucson, AZ, with Berkeley, CA 94701; and Division of Organisms and the Environment, active infestations of H. aureus were chosen for this Department of Environmental Science, Policy, and Management, College of Natural Resources, 137 Mulford Hall 3114, University of study. None of the structures had received a termiti- California, Berkeley, CA 94720. cide application in the previous 5 yr. Site 1 is the

0022-0493/07/1381Ð1390$04.00/0 ᭧ 2007 Entomological Society of America 1382 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

Fig. 1. Diagram of collar deployment at site 1, the University of ArizonaÕs Environmental Research Laboratory (Tucson, AZ).

University of ArizonaÕs Environmental Research Lab- paver was placed on top. Approximately 5 mo after the oratory located near the Tucson International Airport. establishment of the collars, a tripleÐmarkÐrecapture This 70 m2, split-level structure, with a half basement, program (Su et al. 1993) was used to delimit colonies is made of slump block walls (Fig. 1). Site 2 is called and estimate the foraging populations and distances. the Shrimp House, and it is located on the West Ag- In addition, Sentricon monitoring stations (Dow Agro- ricultural Campus of the University of Arizona (Tuc- sciences LLC, Indianapolis, IN) were placed equidis- son, AZ). This concrete block structure covers 139.4 tance or 3.3 m between collars, but they were used m2, and it was built 3.2 m below the soil surface with only for evaluating foraging distances when occupied Ϸ0.7 m above ground (Fig. 2). Site 3, the Olson res- by marked termites. Sentricon Termite Colony Elim- idence, is a 337-m2 home with a standard ßoating slab, ination System is a Dow AgoSciences product devel- and it is constructed of slump blocks (Fig. 3). oped for the commercial termite market for use by Colonies, Population Estimates, and Foraging Dis- pest management professionals. tances. Size and dispersion of foraging populations Termites were captured, marked, released, and re- were estimated with markÐreleaseÐrecapture studies. captured over a period of Þve consecutive weeks in Monitoring stations (collars) were placed at 3.3-m the late summerÐfall 2000. The number of workers and intervals around each structure. Collars consisted of soldiers was determined by direct counts of individ- 15.8-cm-long by 15.8-cm-diameter polyvinyl chloride uals. One collar with at least 200 termites was selected pipe forced 2 mm into the ground. Within each collar and force-fed Þlter paper (Whatman no. 1, Whatman, we placed a 10- by 120-cm section of rolled cardboard Maidstone, United Kingdom) with 0.1% (wt:wt) Nile (B ßute, SF Roll Corp., Tucson Container Corp., Tuc- Blue A (Sigma-Aldrich, St. Louis, MO) (Su et al. 1993, son, AZ). A piece of 2.5-by 1.5- by 10-cm ash (Fraxinus Forschler and Townsend 1996) in an 85-mm petri dish sp.) was positioned in the center, and cardboard was in complete darkness at 25ЊC for 7 d. After 7 d, blue wrapped around the wood and held in place with a termites were counted, and they were returned to the rubber band. Once the rolled cardboard was placed monitoring collar from which they had been collected inside the collar, a 16- by 16- by 1.5-cm concrete brick (Su et al. 1993). August 2007 BAKER AND HAVERTY:FORAGING POPULATIONS OF THE DESERT SUBTERRANEAN TERMITE 1383

Fig. 2. Diagram of collar deployment at site 2, the Shrimp House, West Agricultural Campus, University of Arizona (Tucson, AZ).

Collections were made every 7 d after the release of tances between all possible pairs of collars occupied the marked termites, for 21 d. Termites were collected by the colony. These measurements assume foraging from all collars and brought back to the laboratory for galleries between collars for sites 1 and 2 were along sorting and counting. Termites found within a collar solid structure guidelines, such as block walls, whereas that contained marked termites at any time within the at site 3 it was calculated based on a straight line, and 21-d sampling period were considered nestmates. All thus is a conservative estimate. foraging nestmates within these collars were fed dyed Relationship of Foraging Groups by Quantiﬁcation Þlter paper for another 7 d and released back into their of Cuticular Hydrocarbon Mixtures. Termites were corresponding location. The weighted mean model gathered from individual monitoring collars to char- (WMM) was used to estimate termite foraging pop- acterize the cuticular hydrocarbons of each foraging ulations by counting the numbers of marked and un- group. Our hypothesis was that foraging groups from marked termites collected and the number of marked the same colony would have cuticular hydrocarbon termites released at each cycle (Begon 1979, Haverty mixtures that were quantitatively similar and that et al. 2000). As a comparison, population estimates were quantitatively different from foraging groups based on the Lincoln Index (LI) (Bailey 1951) were that belonged to a different colony. To validate this made using the Þrst markÐreleaseÐrecapture cycle. methodology, we compared the cuticular hydrocar- For the purposes of this article, we considered a bon mixtures of foraging groups from the three sites, colony to be foraging groups of H. aureus sharing because we were certain that the sites shared no col- interconnected galleries (Su and Scheffrahn 1996). onies in common. Our deÞnition assumes that these foragers also are Termite samples were collected 2 or 6 mo after the associated with other conspeciÞcs involved in coop- markÐreleaseÐrecapture cycle, and they were brought erative rearing of offspring (Wilson 1971). The max- to the laboratory at the University of Arizona, where imum foraging distance between collars used by a the termites were separated from cardboard and any colony was determined by measuring (or calculating) other debris. Samples of 50Ð200 foragers (pseuder- the linear distance between the two connecting col- gates or workers) were placed in separate vials, frozen, lars that are furthest apart. The minimum total forag- and then dried (Haverty et al. 1996). Once the ter- ing distance potentially traveled by members of a mites were completely dry, specimens were placed in colony was computed as the sum of the linear dis- separate, labeled, tightly capped 20-ml scintillation 1384 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

Fig. 3. Diagram of collar deployment at site 3, the Olson residence (Tucson, AZ). vials (Wheaton ScientiÞc, Millville, NJ), and they Relationship of Foraging Groups by Examining Sol- were shipped overnight to the USDAÐForest Service dier Proportions. Numbers of soldier and worker ter- laboratory in Albany, CA. The hydrocarbons from mites in each collection were used to calculate the each termite sample were extracted, characterized, proportion of the total represented by soldiers. The and quantiÞed in the same manner as reported in statistical signiÞcance of the Þxed effects (sites 1, 2, Haverty et al. (1996). and 3), sampling period (Þrst, second, or third sam- In the text, we use shorthand nomenclature to iden- ple), and group (collars with dyed termites or collars tify individual hydrocarbons or mixtures of hydrocar- without dyed termites) were tested using the gener- bons. This shorthand uses a descriptor for the location alized linear mixed model at ␣ ϭ 0.05 (SAS Institute of methyl groups (X-me); the total number of carbons 2002). Regression analyses using foraging groups from (CXX) in the hydrocarbon component, excluding the collars with and without dyed termites were con- methyl branch(es); and the number of double bonds ducted to test the slopes (i.e., soldier proportions) after a colon (CXX:Y). Thus, heptacosane becomes among sites and among foraging groups from collars n-C27; 9-methylheptacosane becomes 9-meC27; with or without dyed termites at each site. Statistical 11,19-dimethylnonacosane becomes 11,19-dimeC29; signiÞcance between and/or among slopes was deter- and nonacosadiene becomes C29:2. mined by constructing 95% conÞdence limits. Gas chromatographic-mass spectrometric peak ar- eas were converted to percentages of the total hydrocarbon fraction. We tested the relationship among the Results and Discussion foraging groups and sites by using cluster analysis of the cuticular hydrocarbon mixtures. The presence of Mark–Release–Recapture Studies. Both H. aureus coeluting compounds precluded exact quantiÞcation and Gnathamitermes perplexus (Banks) (Termitidae) of many individual hydrocarbons; therefore, we used foragers were collected in collars at all three sites; the percentage of each peak as the response variable. however, the two species never occupied the same The Euclidean distance for the 12 foraging groups collar at the same time. In addition, G. perplexus for- from the three sites was calculated using all hydro- agers represented Ͻ1.4% of the foraging termites col- carbons (R Development Core Team 2004). lected at any site. They were easily distinguished from August 2007 BAKER AND HAVERTY:FORAGING POPULATIONS OF THE DESERT SUBTERRANEAN TERMITE 1385

Table 1. Summary of foragers released and recaptured for each of three cycles of mark–release–recapture at three sites in Tucson, AZ

No. No. collars No. collars with Marked foragers First Second Third Site collars with foragers marked foragers releaseda recaptureb recaptureb recaptureb 1 16 7 7 782 ϩ 4,170 4,498 (30) 1,325 (9) 49 (8) 2 39 23 8 371 ϩ 1,422 3,290 (9) 3,584 (2) 654 (5) 3 30 16 10 273 ϩ 3,000 13,557 (12) 7,959 (4) 2,419 (27)

a Number of marked foragers initially released ϩ no. of foragers subsequently released. b Total number of termites recovered with the number of marked termites in parentheses.

H. aureus and thus they were not included in any of the foraging populations because of the very small num- results. bers collected (less than Þve). Nonetheless, marked During this study, we collected and marked 11,230 foragers were found in Sentricon stations 1 and 8, H. aureus foragers in total, and we released 89.2% of and they were included in the calculation of forag- them (10,018) back into their original collars. The ing distances (Fig. 1; Table 2). The maximum for- return of the marked foragers was very low (Table 1). aging distance of 26 m and minimum of total forag- Because of the small return of marked foragers (Յ11 ing distance of 297 m (Table 2) were based on all with a mean of 2.12 per occupied collar) in the collars, collars and the two Sentricon stations that con- we decided to estimate the population two different tained marked termites during at least one sampling ways: 1) include only the monitoring collars with period (Table 2; Appendix 1). marked termites assuming the other collars were oc- Site 2. Marked foragers were initially released into cupied by a different colony or colonies, and 2) in- collar 19; additional marked foragers were released 2 clude all collars at the site as one colony assuming the wk later into collars 20 and 37 (Fig. 2; Appendix 2). No collars without marked foragers did not receive them termites, marked or unmarked, were ever recovered simply because of the small return of marked individ- from the initial release point. One week after the uals (Table 1; Appendices 1Ð3). release of marked termites only Þve of the 14 occupied Site 1. Marked termites were initially released into collars had marked termites in them. The next week, collar 9; additional marked foragers were released 2 none of the four collars originally occupied by marked wk later into collars 8Ð12 and 16 (Fig. 1; Appendix 1). foragers contained marked termites; only one collar One week after the release of marked termites, six of (27) contained marked foragers. During the third the seven collars that were occupied during the 3-wk week, only three collars (12, 17, and 23) contained sampling period had marked termites in them. The marked termites. None of the collars was consistently following week, six of the seven collars were occupied, occupied over the 3-wk sampling period. but only three contained marked termites. During the Of the 23 occupied collars, only eight ever con- third week, only two collars were occupied, but both tained marked termites (Table 1). Sixteen of the total collars had marked termites in them. Only one collar 39 collars were never occupied at this site during the (16) furthest from the release point was consistently 3-wk sampling period. Because not all of the occupied occupied and contained marked termites all 3 wk. All collars contained marked foragers during this sam- occupied collars at this site had marked termites at pling period, we would infer that site 2, Shrimp House, least once during the 3-wk sampling period. Nine was “infested” by at least two colonies, if we were to collars were never occupied. Overall recapture rate of assume random foraging among the collars and uni- marked foragers was 47 of 4,952 released for a return form mixing of the marked termites (Su et al. 1984). of 0.94% (Table 1). Foraging populations were esti- However, Jones (1990a) inferred that mixing of mated to be 118,030 by the LI and 100,410 by the marked foragers of H. aureus might not be uniform in WMM (Table 2). a natural setting; thus, the assumption that marked H. Termites were collected from Sentricon stations; aureus foragers would mix uniformly around struc- however, we did not include them in the estimates of tures is also doubtful. Because the return of marked foragers was so low, we hesitate to assume uniform Table 2. Population estimates based on Lincoln Index and mixing of marked foragers among the collars. There- weighted mean model and foraging distances for H. aureus colonies fore, it is not unreasonable to assume that all of the at three sites in Tucson, AZ occupied collars were used by only one colony. Weighted Max Min total Only 16 marked foragers were recovered of 1,793 Lincoln Site mean distance foraging Index released, for a return of 0.89% (Table 1). If we were model (m) distance (m) to consider only the collars with marked foragers, Only collars with marked foragers estimated foraging population was 64,913 by the LI Site 1 118,030 110,410 26 297 and 75,501 by the WMM, with a maximum foraging Site 2 64,913 75,501 62 1,193 Site 3 226,272 252,205 35 880 distance of 62 m and a minimum total foraging distance All collars with foragers of 1,193 m (Table 2). However, if we were to include Site 1 118,030 110,410 26 297 all of the collars with termites in them, the foraging Site 2 98,797 175,596 66 2,427 population estimate would be 98,797 by the LI and Site 3 307,284 313,251 35 1,130 175,596 by the WMM, with a maximum distance be- 1386 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4 tween monitoring collars of 66 m and a minimum total foraging distance of 2,427 m (Table 2). Site 3. Marked foragers were initially released into collar 19; additional marked termites were released 2 wk later into collars 15, 18, 20, and 21 (Fig. 3; Appendix 3). No termites, marked or unmarked, were ever recovered from the initial release point. One week after the release of marked termites only four of the 14 occupied collars contained marked termites in them. The next week, none of the four collars originally occupied by marked foragers contained marked individuals. Moreover, two of the four stations were unoccupied. However, 10 previously occupied collars, in addition to one previously unoccupied collar, contained foragers; only three contained marked foragers. During the third week after release of marked termites, only nine collars were occupied; six contained marked termites. Only four of the collars were consistently occupied over the 3-wk sampling period. All Fig. 4. Dendrogram from cluster analysis based on Eu- but one of the occupied collars were occupied at least clidian distance of hydrocarbons extracted from 12 foraging twice. Of the 16 collars occupied by foragers, only 10 groups collected from sites 1, 2, and 3. ever contained marked termites (Table 1; Appendix 3). Because not all of the occupied collars were visited similar to one another than they were to those from by marked foragers during the 3-wk sampling period, site 1 or site 3. The same is true for sites 1 and 3. We we would infer that the site 3 was “infested” by at least collected cuticular hydrocarbon samples from collars two colonies, if we were to assume random foraging that were not previously occupied by marked foragers and uniform mixing of the marked termites (Su et al. during the markÐreleaseÐrecapture sampling, primar- 1984). However, as with the site 2, the return of ily because termites were present when we sampled 2 marked termites was so low, we hesitate to make an or 6 mo later. Because of the similarity of the cuticular emphatic inference about the number of colonies. hydrocarbon mixtures of these to the others from the Only 43 marked foragers were recovered of 3,273 same site that were occupied during the 3-wk sampling released for a return of 1.31% (Table 1). If we were to period, we are conÞdent that they are from the same consider only the collars with marked foragers the colonies. However, the cuticular hydrocarbons of the estimated foraging population was 226,272 by the LI, foragers collected in collar 7 from site16moafter and 252,205 by the WMM, with a maximum foraging population sampling was completed are so different distance of 35 m and a minimum total foraging distance from the others at the same site that we concluded that of 880 m (Table 2). Based on the number of marked two colonies, rather than one, “infested” site 1. foragers recovered, it is not unreasonable to assume This is the Þrst attempt to use the quantities of that we should include all of the collars. Thus, the cuticular hydrocarabons to connect or associate for- foraging population estimate would be 307,284 by the aging groups of the same species from the same gen- LI and 313,251 by the WMM, with a maximum distance eral area. Our results are very encouraging. Our in- between monitoring collars of 35 m and a minimum ference would be stronger if corroborated with total foraging distance of 1,130 m (Table 2). agonistic bioassays and microsatellite analyses. These Colonies per Structure Inferred from Character- latter two tests were not part of the original protocol. ization of Cuticular Hydrocarbons. We sampled 12 These results do, however, have positive implications foraging groups of H. aureus, three groups from site 1, for estimating colony populations and foraging dis- Þve groups from site 2, and four groups from the site tance. 3, albeit 2Ð6 mo after population sampling. The cu- Colonies per Structure Inferred from Proportion of ticular hydrocarbon mixture of H. aureus is composed Soldiers. Only the site Þxed effect resulted in a sta- ϭ Ͻ of n-alkanes, dienes, 2/4-, 3-, 5-, and internally tistically signiÞcant difference (F2,71 5.82; P 0.005) branched monomethyl alkanes, and various internally in proportion of soldiers. Thus, each site had a pro- branched dimethylalkanes, and 5,19-dimeC27. The portion of soldiers that was statistically signiÞcant most abundant components were 9-meC27, 9-meC29 from each of the others; sampling period and presence and 11,19-dimeC29. All 12 samples were qualitatively or absence of marked termites did not affect soldier identical, i.e., all samples had the same hydrocarbons. proportions. The mean proportion of soldiers at the There were quantitative differences in the cuticular site 1 was 0.135 (SE ϭ 0.038), whereas samples at sites hydrocarbon mixtures among the various collars that 2 and 3 had mean proportions of 0.069 (SE ϭ 0.009) we sampled. and 0.040 (SE ϭ 0.009), respectively. Cuticular hydrocarbon mixtures were more similar At site 1, the correlation between number of sol- among collars from the same site than among those diers and the total number of foragers was highly from different sites, with one exception: site 1, collar signiÞcant (r2 ϭ 0.88, P Ͻ 0.005; n ϭ 14). Because all 7 (Fig. 4). Thus, all of the site 2 samples were more collars sampled contained marked termites during at August 2007 BAKER AND HAVERTY:FORAGING POPULATIONS OF THE DESERT SUBTERRANEAN TERMITE 1387 least one sampling period, we did not compare the pied by only one colony each. Regardless, we reported regression of collars with and without marked foragers population estimates and foraging distances for sites 2 at this site. and 3 by using both assumptions (Table 2). At site 2, the correlation between the number of Our population estimates of H. aureus for the three soldiers and the total number of foragers was highly sites, based on the LI and the WMM, seem to fall signiÞcant (r2 ϭ 0.86, P Ͻ 0.005; n ϭ 34). At this site, within the ranges reported by Haverty and Nutting the regression of foraging groups with dyed termites (1975) and Jones (1990b), despite the presence of the was not statistically different from that of foraging structures. All of the previous estimates of H. aureus groups without dyed termites (Y ϭ 4.19 ϩ 0.052X, r2 ϭ Þeld populations and foraging territories (Haverty and 0.93, P Ͻ 0.005 [n ϭ 13] versus Y ϭ 1.51 ϩ 0.050X, r2 ϭ Nutting 1975; Jones and Nutting, 1989; Jones 1990a,b) 0.78, P Ͻ 0.005 [n ϭ 21]). Therefore, we feel it is were made under native desert conditions. In com- reasonable to assume that the foragers in all of the parison with other estimates of subterranean termite collars at this site were from the same colony, because foraging populations in North America, our data, and the soldier proportions are so similar between the two those of others (Jones 1990b), support an equivalent groups and the return of dyed termites was low, such range in numbers of termites/colony from 5,000 to that dyed termites might not have found their way to 500,000 foragers of Reticuliterms spp. (Forschler and all collars sampled. Townsend 1996, Haverty et al. 2000). However, col- At site 3, the correlation between the number of onies of the Formosan subterannean termite, C. for- soldiers and the total number of foragers was highly mosanus, can range from 1 to 6 million (Su and Schef- signiÞcant (r2 ϭ 0.64, P Ͻ 0.005; n ϭ 34). At this site, frahn 1988, Grace et al. 1996). the regression of foraging groups with dyed termites Of the 10,018 termites that were dyed blue, only 106 was not statistically different from that of foraging were recaptured from all three sites for a 1.06% return. groups without dyed termites (Y ϭ 7.16 ϩ 0.022X, r2 ϭ This is considerably lower that the 2.49% reported by 0.61, P Ͻ 0.005 [n ϭ 22[ versus Y ϭ 3.51 ϩ 0.026X, r2 ϭ Jones (1990b) under native desert conditions using 0.80, P Ͻ 0.005 [n ϭ 12]). As with site 2, we feel it is Sudan Red, whereas Forschler and Townsend (1996), reasonable to assume that the foragers in all monitor- Thorne et al. (1996), and Haverty et al. (2000) had ing collars were from the same colony for the same returns of 6.15, 6.67, and 5.39%, respectively, for Re- reasons. ticuliterms spp. Sornnuwat et al. (1996) reported 1.83% This is a signiÞcant deviation from the hypothesis return of marked termites from a foraging populations presented by Haverty (1977) that each termite species of Coptotermes gestroi Wasmann in an urban area, a has a Þxed proportion of soldiers. Conditions at each proportion more reßective of our results. Haverty et site, whether biological or physical, resulted in differ- al. (1974) reported H. aureus foraging moderately high ent proportions of soldiers, as well as soldier propor- in late summer and early fall responding to the in- tions higher than those previously reported for H. crease in rain and favorable soil temperatures during aureus (Nutting et al. 1973; Haverty and Nutting 1975). this period. In general, we collected the greatest num- However, signiÞcant differences in soldier propor- ber of foragers during the Þrst collection, with the tions among colonies of the same species should be third collection substantially lower. However the third expected. Haverty (1979) and Haverty and Howard collection gave the highest percentage of marked ter- (1981) demonstrated that soldier development in or- mites. Continual disturbance probably caused this de- phaned groups of workers of the rhinotermitid species crease in the number of foragers captured. C. formosanus and Reticulitermes ﬂavipes (Kollar) was Maximum foraging distance for the three sites signiÞcantly different among colonies within each of ranged from 26 to 62 m, and it is likely a consequence these species. Oster and Wilson (1978) theorized that of the greatest distance of the collars. Minimum total caste ratios should be ßexible in mature colonies to foraging distances between sites varied widely from accommodate environmental variations. However, at 297 to 2,427 m. Thus, the size of the occupied structure the time of their publication, they had not seen a and the number of occupied collars greatly affects this systematic study relating soldier/worker ratios to nest metric. The greater distances traveled by H. aureus structure or features of the environment. Thus, we foragers at the site 2 were based on a 3.2-m-deep conclude that it is likely that all collars sampled at each basement and our assumption that the termites trav- site contained foragers from the same colony. The eled along the physical guides of the slump block wall, reason(s) for the signiÞcant differences in soldier pro- not in a straight line from collar to collar. Whereas at portions among sites is unknown. site 3, we assumed a straight line between collars At site 1, all termites collected in the collars during because of a ßoating slab only 10 cm in depth. the 3-wk sampling period were, without question, Estimates of foraging populations of subterranean from a single colony, because each collar that con- termites are commonly made worldwide by using tained termites contained dyed termites at one time or markÐreleaseÐrecapture techniques. There is, how- another. At both sites 2 and 3, it is possible that more ever, considerable debate and controversy over the than one colony was present, because not all of the accuracy, validity, and usefulness of these estimates collars that were visited by foragers contained dyed (Baroni-Urbani et al. 1978, Thorne et al. 1996, Evans termites (Table 1). However, the evidence from char- et al. 1998). We made our estimates of foraging acterization of the cuticular hydrocarbons and soldier populations of H. aureus with knowledge and ap- proportions strongly suggest that each site was occu- preciation of both sides of the argument and in- 1388 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4 cluded our raw data (Appendices 1Ð3) so that others Acknowledgments can draw their own conclusions. QuantiÞcation of We thank Ruben Marchosky for technical assistance with foraging distances, range, area or territory also can the data; L. J. Nelson for technical assistance characterizing also debated. the cuticular hydrocarbons of H. aureus; J. A. Baldwin for The concept of foraging range seems to be a more assistance with the cluster and regression analyses; and M. K. appropriate descriptor of where termites are work- Rust, B. T. Forschler, C. Schal, and three anonymous review- ing than an arbitrary foraging territory (Thorne ers for helpful reviews of an earlier draft of this manuscript. 1998). Constructing a polygon around the collars or This research was made possible, in part, by donations from monitoring stations occupied or used by a colony Dow AgroSciences. includes a great deal of real estate that the termites are not necessarily using or defending. In our study, References Cited we have chosen to use the metric minimum total foraging distance proposed by Haverty et al. (2000) Baker, P. B., and Bellamy, D. 2006. Field and laboratory as a variance of foraging range. This is the distance evaluation of persistence and bioavailability of soil ter- potentially traveled by members of a colony, and it miticides to desert subterranean termite Heterotermes was computed as the sum of the linear distances aureus (Isoptera: Rhinotermitidae). J. Econ. Entomol. 99: 1345Ð1353. between all possible pairs of collars occupied and Bailey, N.T.J. 1951. On estimating the size of mobile popu- defended by a colony. lations from capture-recapture data. Biometrika 38: 293Ð We know that the tunneling galleries do not follow 306. straight lines, at the same depth in the soil, between Baroni-Urbani, C., G. Jones, and G. J. Peakin. 1978. Empir- collars or other food sources. Rather these galleries are ical data and demographic parameters, pp. 5Ð44. In M. V. a vast network of interconnected tunnels and cham- Brian [ed.], Production ecology of ants and termites. bers within the soil, at various depths, and they avoid Cambridge University Press, London, United Kingdom. Begon, M. 1979. Investigating animal abundance: capture-re- various obstacles, interspersed among food sources. capture for biologists. University Park Press, Baltimore, MD. We recognize that the density of collars could affect Evans, T. A., M. Lenz, and P. V. Gleeson. 1998. Testing this parameter. Furthermore, it would be naieve to assumptions of mark-recapture protocols for estimating state that all pairs of collars are connected with their population size using Australian mound-building, subter- own separate tunnel; some collars could be connected ranean termites. Ecol. Entomol. 23: 101Ð121. in series that would make the total foraging distance Forschler, B. T. 1994. Fluorescent spray paint as a topical marker on subterranean termites (Isoptera: Rhinoter- smaller. Nonetheless, minimum total foraging distance mitidae). Sociobiology 24: 27Ð38. provides an additional metric for a more complete Forschler, B. T., and M. L. Townsend. 1996. Mark-release- picture of overall colony size for comparing colonies recapture estimates of Reticulitermes spp. (Isoptera: Rhi- of the same, or different, species of termites, and it notermitidae) colony foraging populations from Georgia, does not include an area or volume of soil that termites U.S.A. Environ. Entomol. 25: 952Ð962. do not use. Getty, G. M., M. I. Haverty, K. A. Copren, and V. R. Lewis. The results of this study continue to support the 2000. Response of Reticulitermes spp. (Isoptera: Rhino- termitidae) in northern California to baiting with concept that estimates of foraging populations based hexaßumuron with Sentricon Termite Colony Elimina- on markÐreleaseÐrecapture are somewhat controver- tion System. J. Econ. Entomol. 93: 1498Ð1507. sial (Thorne et al. 1996, Evans et al. 1998); we did not Grace, J. K., C.H.M. Tome, T. G. Shelton, R. J. Oshiro, and Þnd evidence that marked foragers mix equally among J. R. Yates. 1996. Baiting studies and considerations with foraging sites (collars) occupied by the same colony. Coptotermes formosanus (Isoptera: Rhinotermitidae) in However, we feel that markÐreleaseÐrecapture does Hawaii. Sociobiology 28: 511Ð520. retain a practical use for estimating foraging popula- Haverty, M. I. 1977. The proportion of soldiers in termite colonies: a list and bibliography. Sociobiology 2: 199Ð216. tions of subterranean termites (Baroni-Urbani et al. Haverty, M. I. 1979. Soldier production and maintenance of 1978, Su and Scheffrahn 1996, Tsunoda et al. 1998). soldier proportions in laboratory experimental groups of Still foraging population estimates of H. aureus seem to Coptotermes formosanus Shiraki. Insectes Soc. 26: 69Ð84. be similar whether around structures or in native, Haverty, M. I., and R. W. Howard. 1981. Production of undisturbed deserts. Maximum foraging distance is soldiers and maintenance of soldier proportions by probably a function of the size of the foraging “terri- laboratory experimental groups of Reticulitermes tory or area, whereas the minimum total foraging ﬂavipes (Kollar) and Reticulitermes virginicus (Banks) ” “ ” (Isoptera: Rhinotermitidae). Insectes Soc. 28: 32Ð39. distances are strongly affected by the number of for- Haverty, M. I., and W. L. Nutting. 1975. Density, dispersion, aging sites (in our case collars); the more available and composition of desert termite foraging populations sites or the greater their density, the minimum total and their relationship to superÞcial dead wood. Environ. foraging distance will be larger. Although, it is still Entomol. 4: 480Ð486. unclear whether the structures affect the dispersion of Haverty, M. I., G. M. Getty, K. A. Copren, and V. R. Lewis. foraging sites and can make the total foraging dis- 2000. Size and dispersion of colonies of Reticulitermes spp. (Isoptera: Rhinotermitidae) in a wildland and a res- tances larger because the structure itself can exclude idential location in northern California. Environ. Ento- a signiÞcant portion of the territory (in all three di- mol. 29: 241Ð249. mensions), but it does provide guides (and gaps) for Haverty, M. I., J. K. Grace, L. J. Nelson, and R. Y. Yamamoto. foraging galleries. 1996. Intercaste, intercolony, and temporal variation August 2007 BAKER AND HAVERTY:FORAGING POPULATIONS OF THE DESERT SUBTERRANEAN TERMITE 1389

in cuticular hydrocarbons of Coptotermes formosanus (Isoptera: Rhinotermitidae) in an urban environment. Shiraki (Isoptera: Rhinotermitidae). J. Chem. Ecol. 22: Sociobiology 14: 353Ð359. 1813Ð1843. Su, N.-Y., and R. H. Scheffrahn. 1990. Economically impor- Haverty, M. I., J. P. La Fage, and W. L. Nutting. 1974. Sea- tant termites in the United States and their control. sonal activity and environmental control of foraging of Sociobiology 17: 77Ð94. the subterranean termite, Heterotermes aureus (Snyder), Su, N.-Y., and R. H. Scheffrahn. 1994. Field evaluation of in a desert grassland. Life Sci. 15: 1091Ð1101. hexaßumuron bait for population suppression of sub- Haverty, M. I., W. L. Nutting, and J. P. La Fage. 1975. Den- terranean termites (Isoptera: Rhinotermitidae). J. sity of colonies and spatial distribution of foraging terri- Econ. Entomol. 87: 389Ð397. tories of the desert subterranean termite, Heterotermes Su, N.-Y., and R. H. Scheffrahn. 1996. A review of the aureus (Snyder). Environ. Entomol. 4: 105Ð109. evaluation criteria for bait-toxicant efÞcacy against Jones, S. C. 1990a. Delineation of Heterotermes aureus Þeld colonies of subterranean termites (Isoptera). So- (Isoptera: Rhinotermitidae) foraging territories in a So- ciobiology 28: 521Ð530. noran Desert grassland. Environ. Entomol. 19: 1047Ð1054. Su, N.-Y., P. M. Ban, and R. H. Scheffrahn. 1991. Evaluation Jones, S. C. 1990b. Colony size of the desert subterranean of twelve dye markers for population studies of the east- termite Heterotermes aureus (Isoptera: Rhinotermitidae). ern and Formosan subterranean termites (Isoptera: Rhi- Southwest. Nat. 35: 285Ð291. notermitidae). Sociobiology 19: 349Ð362. Jones, S. C., and W. L. Nutting. 1989. Foraging ecology of Su, N.-Y., P. M. Ban, and R. H. Scheffrahn. 1993. Foraging subterranean termites in the Sonoran Desert, pp. 79Ð106. populations and territories of the eastern subterranean In J. O. Schimdt [ed.], Interactions among plants and termite (Isoptera: Rhinotermitidae) in southeastern Flor- animals in the western deserts. University of New Mexico ida. Environ. Entomol. 22: 1113Ð1117. Press, Albuquerque, NM. Su, N.-Y., M. Tamashiro, J. Yates, and M. I. Haverty. 1984. Nutting, W. L., M. I. Haverty, and J. P. La Fage. 1973. For- Foraging behavior of the Formosan subterranean termite aging behavior of two species of subterranean termites in (Isoptera: Rhinotermitidae). Environ. Entomol. 13: 1466Ð the Sonoran Desert of Arizona, pp. 298Ð301. In Proceed- 1470. ings, 7th International Congress of the International Thorne, B. L. 1998. Biology of subterranean termites in the Union for the Study of Social Insects, 10Ð15 September genus Reticulitermes. Part I, pp. 1Ð30. In NPCA research 1973, London, United Kingdom. Birkha¨user, Basel, Swit- report on subterranean termites. National Pest Control zerland. Association, Dunn Loring, VA. Oster, G. F., and E. O. Wilson. 1978. Caste and ecology in the Thorne, B. L., E. Russek-Cohen, B. T. Forschler, N. L. social insects. Princeton University Press, Princeton, NJ. Breisch, and J.F.A. Traniello. 1996. Evaluation of markÐ R Development Core Team. 2004. A language and environ- releaseÐrecapture methods for estimating forager popu- ment for statistical computing. R Foundation for Statistical lation size of subterranean termite (Isoptera: Rhinoter- Computing, Vienna, Austria. (http://www.R-project.org). mitidae) colonies. Environ. Entomol. 25: 938Ð951. SAS Institute. 2002. SAS/STAT 2002Ð2003. Software, ver- Tsunoda, K., H. Matsuoka, and T. Yoshimura. 1998. Colony sion 9.1.3 of the SAS System for Windows. SAS Institute, elimination of Reticulitermes speratus (Isoptera: Rhinot- Cary, NC. ermitidae) by bait application and the effect of foraging Sornnuwat, Y., K. Tsunoda, T. Yoshimura, M. Takahashi, and territory. J. Econ. Entomol. 91: 1383Ð1386. C. Vongkaluang. 1996. Foraging populations of Copto- Wilson, E. O. 1971. The insect societies. The Belknap Press termes gestroi (Isoptera: Rhinotermitidae) in an urban of Harvard University Press, Cambridge, MA. area. J. Econ. Entomol. 89: 1485Ð1490. Su, N.-Y., and R. H. Scheffrahn. 1988. Foraging population and territory of the Formosan subterranean termite Received 22 September 2006; accepted 27 April 2007.

Appendix 1. Termites collected for each of three dates at site 1, the Environmental Research Laboratory

Collara 18 Aug. 2000b 25 Aug. 2000b 1 Sept. 2000b 8 243 (4) 219 (0) 9 18 (1) 51 (5) 10 467 (3) 265 (0) 11 675 (10) 12 300 (1) 47 (0) 16 (4) 15 157 (3) 16 2,825 (11) 586 (1) 33 (4) Totals 4,498 (30) 1,325 (9) 49 (8)

On 11 August 2000, 782 marked termites were released into collar 9. On 25 August 2000, 4170 marked termites were released into collars 8Ð12 and 16. a Collars 1Ð7, 13, and 14 were never occupied by foragers. b Total number of termites recovered with the number of marked termites in parentheses. 1390 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 4

Appendix 2. Termites collected for each of three dates at Site Appendix 3. Termites collected for each of three dates at site 2, the Shrimp House 3, the Olson residence

Collara 7 Sept. 2000b 14 Sept. 2000b 21 Sept. 2000b Collara 25 Sept. 2000b 2 Oct. 2000b 9 Oct. 2000b 7 200 (0) 1 988 (0) 645 (0) 8 240 (0) 18 (0) 5 1,389 (0) 226 (0) 9 21 (1) 7 3,827 (0) 405 (1) 318 (0) 11 26 (0) 10 1,703 (0) 651 (2) 216 (2) 12 94 (3) 11 455 (0) 1,010 (0) 265 (4) 13 56 (0) 15 1,411 (1) 1,116 (0) 14 159 (0) 105 (0) 16 1,027 (0) 285 (0) 16 0 (0) 17 77 (0) 281 (0) 17 27 (3) 34 (0) 18 (1) 18 266 (4) 15 (5) 18 64 (2) 20 1,595 (6) 477 (0) 19 40 (0) 21 578 (1) 27 (7) 20 544 (2) 789 (0) 25 534 (0) 289 (7) 22 81 (0) 27 107 (0) 23 230 (0) 87 (1) 28 828 (0) 453 (0) 43 (0) 24 306 (0) 1 (0) 29 111 (0) 1,267 (1) 25 25 (0) 30 1,020 (2) 26 124 (0) Totals 13,507 (12) 7,979 (4) 2,419 (27) 27 86 (0) 233 (2) 29 125 (0) 876 (0) 127 (0) On 18 September 2000, 273 marked termites were released into 30 1,066 (0) 83 (0) collar 19. On 2 October 2000, 3,000 marked termites were released into 33 62 (0) collars 15, 18, 20, and 21. 35 172 (0) a Collars 2Ð4, 6, 8Ð9, 12Ð14, 22Ð24, and 26 were never occupied by 37 1,398 (1) foragers. Total 3,290 (9) 3,584 (2) 654 (5) b Total number of termites recovered with the number of marked termites in parentheses. On 31 August 2000, 371 marked termites were released into collar 19. On 14 September 2000, 1422 marked termites were released into collars 20 and 37. a Collars 1Ð6, 10, 15, 16, 21, 28, 31Ð32, 34, and 38Ð39 were never occupied by foragers. b Total number of termites recovered with the number of marked termites in parentheses.