host prefference and attack pattern of dendroctonus rhizophagus

PLANTÐINSECT INTERACTIONS

Host Preference and Attack Pattern of Dendroctonus rhizophagus(Coleoptera: Curculionidae: Scolytinae): A Bark Beetle Specialist on

Pine Regeneration

GUILLERMO SANCHEZ-MARTINEZ1 AND MICHAEL R. WAGNER2

Environ. Entomol. 38(4): 1197Ð1204 (2009)

ABSTRACT Pine seedlings and saplings are seldom attacked by bark beetles of the genusDendroc-tonus. However, Dendroctonus rhizophagus (Coleoptera: Curculionidae: Scolytinae) Thomas andBright speciÞcally attacks pine seedlings and causes conspicuous mortality in naturally regeneratedstands in the Sierra Madre Occidental, northern Mexico. We evaluated the host preference and attackof D. rhizophagus under Þeld conditions. We tried to establish any relationship between tree growthor host size and the number of attacking beetles. Generally, only one pair of beetles attacked eachof the seedlings regardless of host size; however, a signiÞcant positive linear relationship between hostsize and adult brood size was observed. We found that this species preferred the best growing seedlingsin our study sites.

KEY WORDS bark beetles, Dendroctonus rhizophagus, host preference, pine seedlings, tree mor-tality

Because of their tree killing capacity, bark beetle spe-cies in the genera Dendroctonus and Ips (Coleoptera:Curculionidae: Scolytinae) are among the most inten-sively studied forest insects (Coulson 1979, Berryman1982, Wood 1982, Mattson and Haack 1987, Raffa andBerryman 1987). Byers (1995) estimated a total of3,800 published papers from 1970 to 1995 on bark andambrosia beetles. About one third of these publica-tions relate to the four most aggressive Dendroctonusspecies of North America:Dendroctonus frontalisZim-merman,Dendroctonus ponderosaeHopkins,Dendroc-tonus brevicomis LeConte, andDendroctonus pseudot-sugae Hopkins (Byers 1995). A bark beetle species isconsidered aggressive if it must kill its host for suc-cessful reproduction and feeding (Berryman et al.1989, Price 1997, Logan et al. 1998, Six and Paine 1999).An aggressive bark beetle species can successfullycolonize healthy trees, whereas nonaggressive speciesfeed on dying or dead trees (Raffa et al. 1993).

Studies on the behavior of aggressive bark beetlespecies have led to the identiÞcation of the mainfactors regulating the population dynamics of barkbeetles (Coulson 1979; Berryman 1982, 1997; Raffa andBerryman 1983). Theoretical models indicate that, atendemic population levels, bark beetles are conÞnedto stressed trees because tree resistance exerts a majorregulating force (Berryman 1982, 1991, 1997; Cates

and Alexander 1982, Sturgeon and Mitton 1982). Atepidemic levels, bark beetles overcome tree resistancethrough chemical communication and “mass attack”(Coulson 1979; Berryman 1982, 1997; Borden 1982;Raffa and Berryman 1983; Berryman et al. 1985, 1989).The concept of mass attack for bark beetles includesthe arrival of a large number of beetles on the tree hostina shortperiodof time thatoverwhelms treedefenses(Christiansen et al. 1987). For Dendroctonus and Ipsspecies, McHugh et al. (2003) used as criteria of massattack, if attacks occurred in �75% of the tree bolecircumference.

In addition to the theoretical arguments, the liter-ature on bark beetle population dynamics indicatesmost Dendroctonus and Ips species do not reach out-break populations on very young trees because theyoffer a limited resource for food and reproduction.Attacks on regeneration are considered rare, excep-tional, or unfeasible (Berryman 1982, Miller and Bor-den 1985, Berryman et al. 1989). However, in theSierra Madre Occidental, Northern Mexico,Dendroc-tonus rhizophagus Thomas and Bright exclusively killsseedlings and young saplings of Apache pine (Pinusengelmannii Carr.), Durango pine (Pinus durangensisMartinez), Pinus leiophylla Schlecht and Cham., andArizona pine (Pinus arizonica Engelm) (Thomas andBright1970,Estrada-Murrieta1983).Estrada-Murrieta(1983) tallied �2,000,000 seedlings killed by D. rhi-zophagus during 1977Ð1983 in Mesa del Huracan(29�38� N, 108�14� W), Chihuahua, comprising an in-fested surface of 12,000 ha. More recent statistics in-dicate that D. rhizophagus causes the most signiÞcantinfestations in the state of Chihuahua (�12,500 ha

1 Corresponding author: Instituto Nacional de InvestigacionesForestales, Agrõcolas y Pecuarias, Campo Experimental Pabellon, Km.32.5 Carr. Ags.-Zac., Pabellon de Arteaga, Ags. C.P. 20660, Mexico(e-mail: [email protected]).

2 School of Forestry, Northern Arizona University, Box 15018, Flag-staff, AZ 86011-5018.

0046-225X/09/1197Ð1204$04.00/0 � 2009 Entomological Society of America

infested from 1987 to 1994) in comparison with otherbark beetle species.

Although D. rhizophagus was described more thanthree decades ago (Thomas 1966, Thomas and Bright1970, Wood 1982), primary literature is limited to itstaxonomy, histology, and phylogeny (Diaz et al. 1998;Kelley and Farrell 1998; Zuniga et al. 1998, 2002; Sali-nas-Moreno et al. 2004). From a genetic point of view,D. rhizophagus is closely related to Dendroctonusvalens LeConte (Kelley and Farrell 1998, Zuniga et al.2002); however, these species show distinctive bio-logical differences. D. valens usually breeds in lowerboles and stumps of large trees of a broad range ofspecies but rarely kills it host (Thomas 1966, Pajaresand Lanier 1990, Kelley and Farrell 1998), whereasD.rhizophagus, a more specialized species (Kelley andFarrell 1998, Zuniga et al. 2002), attacks pine regen-eration and kills its host as the result of a successfulattack (Thomas and Bright 1970, Estrada-Murrieta1983).D. rhizophagus showsaclear adaptation towith-stand winter temperatures by synchronizing itsphloem consumption and development. During sum-mer and fall, larvae consume most of the phloem in thestem; during winter, they consume the phloem in theroots and then overwinter (Estrada-Murrieta 1983).Thus, to successfully reproduce,D. rhizophagus has toconsume the roots, resulting in the death of its host.

We consider that biological and ecological infor-mation on D. rhizophagus can be of great importanceto the development of theory on bark beetle popula-tion dynamics. For instance, descriptions by Thomasand Bright (1970) and Estrada-Murrieta (1983) indi-cate that this species does not mass attack. However,the capacity to build up populations to outbreak levelsis evident by the number of killed trees and extensiveinfested area (Estrada-Murrieta 1983. In addition, sev-eral years of Þeld observations by the Þrst authorindicate that D. rhizophagus seems to prefer the bestgrowing trees within a given young cohort. Wethought that if these premises hold true, the factorsregulating the population dynamics ofD. rhizophagusmust be different from those regulating the populationdynamics of the few other aggressive bark beetle spe-cies. In this paper, we provide empirical evidence tosupport the hypotheses that D. rhizophagus (1) pre-fers the best growing seedlings within a given youngstand and (2) though an aggressive species, lacks massattack behavior typical of other bark beetles.

Materials and Methods

Two Þeld studies were conducted during 1993Ð1995within natural forested areas in Madera, Chihuahua(29�12� N, 108�12� W), northern Mexico, where out-breaks by D. rhizophagus are common. Pine speciesgrowing in this region include P. engelmannii, P. ari-zonica, P. durangensis, P. leiophylla, andP. chihuahuanaamong others. Forests are under communal ownershipcalled “Ejidos,” managed by semiprivate forestry agen-cies (Forestry Units). The extent of theD. rhizophagusproblem is shown by the fact that 156 sanitation prac-tices were implemented from 1989 to 1995 in our

area of study (Pest Control Records of ForestryUnits 2 and 10).Study on Host Preference. A Þeld study was estab-

lished in Madera, Chihuahua, northern Mexico, inAugust 1994, after the period of attack ofD. rhizopha-gus. Four naturally regenerated Apache pine (P. en-gelmannii) stands were selected. Three stands withinEjido Madera (29�12� N, 108�12� W) were �1 km apart.The fourth stand within Ejido El Largo (29�21� N,108�28� W) was �50 km distant from the other three.In each stand, we randomly selected a point, whichbecame the center of a 100-m2 plot. For each treewithin the plots, we recorded the attack condition(attacked or nonattacked), height, stem diameter atthe ground level (DGL), length of terminal shoot, andage (by counting number or whorls). Our plots con-tained mostly a cohort of pine regeneration. Becausethis was a Þeld study, bark beetles naturally selectedtheir hosts within the plots.

The study was considered as a two-factor completerandomized design. The attack condition of the treewas considered as factor A with two levels (1 � attack0 � no attack). Site location was considered as factorB with four levels (site 1Ð4). Individual trees repre-sented the experimental units. DGL, tree height,length of the terminal shoot, and age were the re-sponse variables. Mean radial growth was estimated bytaking the quotient between stem radius and tree age(mean radial growth). Data were analyzed throughmultivariate analysis of variance (MANOVA) at the0.05 experimental-wise � level, followed by univariateF-tests for each factor, using SYSTAT 7.0 (SYSTAT1997) and validation of the analyses with JMP 3.0 (SASInstitute 1996).Study on Attack Pattern. This study was conducted

during 1993Ð1994 within natural forested areas of theEjido Madera (29�12� N, 108�12� W). A random sampleof 42 currently infested trees from threeP. engelmanniistands was collected in September 1993. The ßightperiod of this bark beetle occurs from mid-June tomid-August. Infested trees at this time of year have agreen and healthy appearance but are identiÞed by adistinctive single entry hole at the base of the stem thatis usually covered with white frass. Sample trees werecarefully extracted from the ground, labeled, andtransported to a local Research Station of the National(Mexican) Institute on Forestry, Agriculture and An-imal Husbandry Research (INIFAP). Trees were mea-sured for length of terminal shoot, height, DGL, andage (by number of whorls). All the attacking barkbeetles, larvae, and eggs were extracted from thephloem and counted. Tree measurements were usedto derive two growth indices. The Þrst index wasconstructed by taking the quotient between stem ra-dius and tree age (mean radial growth). The secondindex was deÞned as ratio between current annualheight growth (CAG) and mean annual height growth(MAG). As pointed out by Avery and Burkhart (1994)and Smith et al. (1997), CAG and MAG can be ex-pressed in any unit of measurement. In this study, thelength of the terminal shoot represented CAG,whereas MAG was the tree height divided by age.

1198 ENVIRONMENTAL ENTOMOLOGY Vol. 38, no. 4

To capture a broader representation of host sizes,another random sample of 95 currently infested treesfrom four stands was collected in August 1994. Thesame procedure was followed for tree measurementsand for the extraction of attacking beetles, eggs, andlarvae.

Standard forestry measurements had to be modiÞedbecause trees attacked byD. rhizophagusare relativelysmall. For instance, we measured the diameter at theground level instead of diameter at breast high be-cause most trees were �1.5 m tall. Similarly, the pe-riodic growth ratio vigor index, deÞned by the meanannual growth in the past 5 yr divided by mean annualgrowth of the previous 5 yr (Mahoney 1978), wasimpractical because most trees were �10 yr old. Treesfrom the 1993 sample ranged from 0.54 to 1.63 m inheight (mean � 0.99, SD � 0.26, n� 42) and from 3.8

to 7.9 cm in DGL (mean � 5.95, SD � 0.43, n � 42).The 1994 sample contained a wider range of tree sizes.Height ranged from 0.23 to 2.46 m (mean � 1.10, SD �0.52, n � 95) and DGL from 2.6 to 11.2 cm (mean �5.66, SD � 1.49, n � 95).

In addition to our samples of currently attackedtrees, a random sample of 18 previously attacked trees(1993 attack, already killed trees) with mature broodwas collected in mid-June of 1994, a few days beforethe beginning of the ßight period. New beetles at thistime are aggregated inside the root collar and alongthe roots. Trees were measured in the Þeld (height,DGL, age, and number of entry holes) and carefullyextracted from the ground to reduce the possibility oflosing any specimen. All the adult beetles were ex-tracted from the tree, collected in vials, and latercounted.

Fig. 1. Characteristics of trees preferred by D. rhizophagus in four different P. engelmanni stands (sites) in Madera Chih.,northern Mexico. For each site, attacked and nonattacked trees were growing together within a 100-m2 sampling plot. Multivariateadditive difference by the attack (WilksÕ �* � 0.915, F� 12.602, P� 0.0001) and site factor (WilksÕ �* � 0.307, Approximate F�66.984, P � 0.0001) indicate that D. rhizophagus preferred the best growing trees regardless of site.

Table 1. Multivariate ANOVA for the detection of differencesbetween attacked and nonattacked trees, among four sites wherethe trees were growing, and for the interactive effect of the twofactors on multiple response variables of P. engelmannii (hosts ofD. rhizophagus)

Source ofvariation

WilksÕ � Fdf (factor

error)P

Attack 0.915 Exact 12.60 5 681 �0.0001Site 0.307 Approximate 66.98 15 1,880 �0.0001Attack � site 0.943 Approximate 2.71 15 1,880 0.0004

Table 2. Univariate F tests for the interactive effect of attackand site factors on all the response variables of P. engelmannii (hostof D. rhizophagus)

Response variable Fdf

PInteraction Error

Mean radial growth 3.836 3 685 0.010Length of terminal shoot 1.767 3 685 1.520Tree height 0.526 3 685 0.665Diameter at ground level 1.615 3 685 0.185Age 2.156 3 685 0.092

August 2009 SANCHEZ-MARTINEZ AND WAGNER: HOST PREFERENCE AND ATTACK PATTERN OF D. rhizophagus 1199

Data sets (1993 and 1994) were analyzed separatelythrough linear regression analysis using the generallineal model: y � X� �, where ynxm � responsevariable; Xn(r1) � matrix of predictor variables;�(r1)n � matrix of intercept and slopes for the dif-ferent changes on X; and � � error.

The predictor variables for currently infested treeswere the mean radial growth and CAG:MAG ratio.The response variables were the absolute number ofcolonizing beetles/tree and total immature brood(number of hatching larvae eggs)/tree. In the caseof old attacked trees, the predictor variables wereDGL and height, and the response variable was thenumber of new adult beetles.

Results

Host Preference. Trees in this study had an overallaverage SD of 1.05 0.4696 m high, 4.80 1.359 cmDGL, 20.85 9.18 cm length of terminal shoot,0.3278 0.079 radial growth ratio, and 7.93 0.101 yrof age (n � 693).

The two-way MANOVA indicated a signiÞcant in-teraction effect by attack and site in at least one of theresponse variables (Table 1). Based on Johnson andWichern (1998), we conducted Þve univariate two-way ANOVAs to see if the interaction was present inall the response variables and give the appropriateinterpretation. We found that the only response withsigniÞcant interaction was mean radial growth (Table2); therefore, we used the results from MANOVA tointerpret the simultaneous additive effects of site andattack on height, DGL, age, and terminal shoot. Under

this scenario, there was a signiÞcant multivariate dif-ference between the attacked and nonattacked treesand at least one site had trees with different growingattributes (Table 1).

Univariate F tests also indicated that every responsevariable contributed to the multivariate difference bythe attack and location factors, conÞrming that thosetrees that were preferred by D. rhizophagus had thebest current annual growth (terminal shoot), the larg-est height and diameter, and were slightly older withinthe young cohort (Fig. 1).

We conducted TukeyÕs honestly signiÞcant differ-ence (HSD) tests to Þnd out how general tree at-tributes differed among the locations. Trees on sites 2and 3 had similar characteristics. Trees of site 1 wereof equal age and DGL as those in sites 2 and 3 but had

Fig. 2. Mean radial growth of attacked and nonattackedtrees byD. rhizophagus in four differentP. engelmannii stands(sites) in Madera Chih., northern Mexico. For each site,attacked and nonattacked trees were growing together withina 100-m2 sampling plot. A signiÞcant interaction between theattack and site factor was observed (F � 3.836, P � 0.01),because the difference between the mean radial growth at-tacked and nonattacked trees in site 1 was not as marked as inthe other sites. However, because in every case,D. rhizophagusselected those trees with larger mean radial growth, this statis-tical interaction isnotsigniÞcant fromthebiological standpoint.Mean radial growth of attacked and nonattacked trees wassigniÞcantly different (F � 30.60, P � 0.000).

Fig. 3. Relationship between growth of young P. en-gelmanni and the absolute number of attacking bark beetles(D. rhizophagus) in naturally infested trees in Madera, Chih.,northern Mexico, in 1993. Two growth indices are shown. Themean radial growth (A) is the quotient between stem radius atthe ground level and tree age. Current annual growth:meanannual growth ratio (B) refers to the ratio between currentheight growth (length of terminal shoot) and total tree height.Regression analyses indicate that the number of attacking bee-tles does not respond to either of these growth indices. Dotsrepresent a sample size of 42 observations.


smaller height and current growth. Finally, site 4 hadthe smallest but the oldest trees in comparison to theother sites, which indicates that this was the poorestsite in our study (Fig. 1).

The additive signiÞcant effects by the attack andlocation factors and the lack of interaction betweenthese two factors showed conclusively that, regardlessof the host differences among sites, D. rhizophaguspreferred the best growing trees within each site.

We pointed out earlier that the univariate F-test forradial growth ratio indicated a signiÞcant interactionby the factors. However, as shown in Fig. 2, the in-teraction appears because the mean radial growth ofnonattacked trees in site 2 was smaller than that of site1, whereas the opposite occurred for attacked trees.Because the levels of the factor location are not gra-dients that can be regulated in the statistical sense, thisstatistical interaction simply indicates that the differ-ence between attacked and nonattacked trees in site1 was not as strong as in the other sites, but in everycase, D. rhizophagus selected the best-growing seed-lings (Fig. 2).

Attack Pattern. For the 1993 data set, regressionanalyses showed no linear relationship betweenany of the tree growth indices and the absolutenumber of attacking beetles in currently infested trees(Fig. 3). Likewise, there was no linear relationship be-tween tree growth and immature bark beetle brood size(Fig. 4).

For the 1994 data set, the regression between meanradial growth and number of attacking beetles in cur-rently infested trees was signiÞcant, but the variationon the response variable accounted for by mean radialgrowth was minimal (Fig. 5A). The regression be-tween CAG:MAG ratio and the number of attackingbeetles in currently infested trees was not signiÞcant(Fig. 5B). Similarly, there was no signiÞcant linearrelationship between mean radial growth and imma-

Fig. 4. Relationship between growth of P. engelmanni andthe immature brood of D. rhizophagus in naturally infestedregeneration in Madera, Chih., northern Mexico, in 1993. Twogrowth indices are shown. The mean radial growth (A) is thequotient between stem radius at the ground level and tree age.Current annual growth:mean annual growth ratio (B) refers tothe ratio between current height growth (length of terminalshoot) and total tree height. Regression analyses were not sig-niÞcant. Dots represent a sample size of 42 observations.

Fig. 5. Relationship between growth of young P. en-gelmanni and the absolute number of attacking bark beetles(D. rhizophagus) in naturally infested trees in Madera, Chih.,northern Mexico, in 1994. Two growth indices are shown.The mean radial growth (A) is the quotient between stemradius at the ground level and tree age. Current annualgrowth:mean annual growth ratio (B) refers to the ratiobetween current height growth (length of terminal shoot)and total tree height. Regression analyses indicate only aweak relationship between mean radial growth and the num-ber of attacking beetles (A). Dots represent a sample size of95 observations.


ture bark beetle brood size nor between CAG:MAGratio and immature brood size (Fig. 6).

In contrast to the results in currently infestedtrees, we found a signiÞcant positive linear relation-ship between host size and adult brood size in oldinfested trees (Fig. 7), despite the fact that all 18infested trees had just one attack (entry hole).These results suggest that, although parent beetlesdo not regulate the immature brood according tohost size, the adult brood size of D. rhizophagus isadjusted according to tree size probably throughintraspeciÞc competition.

Discussion

In this study, we showed that D. rhizophagus, al-though an aggressive species, follows an attack strat-egy that is completely different from congeneric mass

attacking beetles. It specializes in seedlings and veryyoung saplings, prefers the best growing trees withina given young tree cohort, but lacks the trait of massattack. Because no attacks were observed in largesaplings, pole sized, or adult trees, we believe theproximate and ultimate regulating force of the popu-lation dynamics ofD. rhizophagus is the availability ofyoung tree cohorts (of preferred species) withinstands and forests and the femaleÕs reproductive ca-pacity. We also believe that an age (�10 yr old) andsize (�3 m tall) threshold must limit host suscepti-bility. The empirical evidence in this study and theanecdotal reports of other bark beetle species attack-ing young conifers (Schenk et al. 1976, Furniss andCarolin 1977) gives us the opportunity to enrich thegeneral theory (under development) on bark beetlepopulation dynamics.

We consider that the exceptions and even the gen-eralized patterns on bark beetle population dynamicsshould be frequently revisited to provide strength inthe development of a general theory of bark beetlepopulation dynamics. We conclude that D. rhizopha-gus is a good example to show that bark beetles notonly have the ability to reproduce on seedlings andsaplings but can become specialists. It is a good speciesfor the study of bark beetle population dynamics be-cause it gives the opportunity to solve the sample sizeproblem and the cohort problem discussed by Price etal. (1990). Because of the lack of mass attack and thesmall host size, the entire number of attacking beetlesand entire immature and mature brood per tree can becounted as we did in this study. Finally because thisspecies attacks young cohorts of trees, studies on thepositive and negative impacts on stand structure arenecessary.

Fig. 6. Relationship between growth of P. engelmanniand the immature brood of D. rhizophagus in naturally in-fested regeneration in Madera, Chih., northern Mexico, in1994. Two growth indices are shown in this Þgure. The meanradial growth (A) is the quotient between stem radius at theground level and tree age. Current annual growth:mean an-nual growth ratio (B) refers to the ratio between currentheight growth (length of terminal shoot) and total tree height.Regression analyses showed no signiÞcant linear relationshipbetweenanyof the twogrowth indicesandthe immaturebroodsizeproducedbytheabsolutenumberofattackingbeetles.Dotsrepresent a sample size of 95 observations.

Fig. 7. Relationship between the size of trees (P. en-gelmannii) infested by D. rhizophagus and the emergingmature brood size after the host was killed. All trees in thissample (n � 18) were naturally infested. Notice that theabsolute number of new adult beetles per tree was recorded.The number of attacks per tree was one for all trees.


Acknowledgments

We thank J. D. Bailey (Northern Arizona University,School of Forestry, Flagstaff), K. M. Clancy (USDAÐForestService, Flagstaff), and P. W. Price (Northern Arizona Uni-versity, Biology Department, Flagstaff) for helpful review ofa former version of this manuscript. This research was fundedby the Instituto Nacional de Investigaciones Forestales,Agrõcolas y Pecuarias (Mexico), Campo ExperimentalMadera, and the Northern Arizona University, McIntire-Stennis Program. A. D. Pereda (Forestry Unit Number 2,Ejido El Largo), S. E. Paz (Forestry Unit Number 10, EjidoMadera), and H. R. Bolanos (Forestry Unit Number 5, SanJuanito) provided valuable reports on the control of out-breaks of D. rhizophagus, and we are very thankful to them.P. P. Gastelum helped in the collection and dissection ofsample trees.

References Cited

Avery, T. E., and H. E. Burkhart. 1994. Forest measure-ments. McGraw-Hill, New York.

Berryman, A. A. 1982. Population dynamics of bark beetles,pp. 264Ð314. In J. B. Mitton and K. B Sturgeon (eds.), Barkbeetles in North American conifers: a system for the studyof evolutionary biology. University of Texas Press, Austin,TX.

Berryman, A. A. 1991. Population theory: an essential in-gredient in pest prediction, management, and policy-making. Am. Entomol. 37: 138Ð142.

Berryman, A. A. 1997. On the principles of population dy-namics and theoretical models. Am. Entomol.: 147Ð151.

Berryman, A. A., B. Dennis, K. F. Raffa, and N. C. Stenseth.1985. Evolution of optimal group attack, with particularreference to bark beetles (Coleoptera: Scolytidae). Ecol-ogy 66: 898Ð903.

Berryman, A. A., K. F. Raffa, J. A. Millstein, and N. C. Stens-eth. 1989. Interaction dynamics of bark beetle aggrega-tion and conifer defense rates. Oikos 56: 256Ð263.

Borden, J.H. 1982. Aggregation pheromones, pp. 74Ð139. InJ. B. Mitton and K. B. Sturgeon (eds.), Bark beetles ofNorth American conifers: a system for the study of evo-lutionary biology. University of Texas Press, Austin, TX.

Byers, J. 1995. Host-tree chemistry affecting colonization inbark beetles, pp. 154Ð213. In R. T. Carde and W. J. Bell.(eds), Chemical ecology of insects 2. Chapman & Hall,New York.

Cates, R. G., and H. Alexander. 1982. Host resistance andsusceptibility, pp. 212Ð263. In J. B. Mitton and K. B. Stur-geon (eds.), Bark beetles of North American conifers: asystem for the study of evolutionary biology. Universityof Texas Press, Austin, TX.

Christiansen, E., R. H. Waring, and A. A. Berryman. 1987.Resistance of conifers to bark beetle attack: searching forgeneral relationship. For. Ecol. Manag. 22: 89Ð106.

Coulson, R. N. 1979. Population dynamics of bark beetles.Annu. Rev. Entomol. 24: 417Ð447.

Diaz, E., R. Cisneros, G. Zuniga, and E. Uria-Galicia. 1998.Comparative anatomical and histological study of thealimentary canal of Dendroctonus parallelocollis, D. rhi-zophagus, and D. valens (Coleoptera: Scolytidae). Ann.Entomol. Soc. Am. 91: 479Ð487.

Estrada-Murrieta, O. 1983. Biologõa del descortezador delrenuevo de pino Dendroctonus rhizophagus T. y B. (Col.:Scolytidae) en la region Mesa del Huracan, Chih. Tesis deLicenciatura. Universidad Autonoma Chapingo, Mexico.

Furniss, R. L., and V. M. Carolin. 1977. Western forest in-sects. Misc. Publ. 1339. U.S. Forest Service, Washington,DC.

Johnson, R. A., and D. W. Wichern. 1998. Applied multi-variate statistical analysis. Prentice Hall, Upper SaddleRiver, NJ.

Kelley, S. T., andB.D. Farrell. 1998. Is specialization a deadend? The phylogeny of host use in Dendroctonus barkbeetles (Scolytidae). Evolution 52: 1731Ð1743.

Logan, J. A., P. White, B. J. Bentz, and J. A. Powell. 1998.Model analysis of spatial patterns in mountain pine beetleoutbreaks. Theor. Popul. Biol. 53: 236Ð255.

Mahoney,R.L. 1978. Lodgepole pine/mountain pine beetlerisk classiÞcation methods and their application, pp. 106Ð110. In A. A. Berryman, G. D. Amman, and R. W. Stark(eds.), Theory and practice of mountain pine beetle man-agement in Lodgepole pine forests. University of Idaho,Moscow, ID.

Mattson, W. J., and R. A. Haack. 1987. The role of drought inoutbreaks of plant-eating insects. Bioscience 37: 110Ð118.

McHugh, C. W., T. E. Kolb, and J. L. Wilson. 2003. Barkbeetle attacks on ponderosa pine following Þre in north-ern Arizon. Eviron. Entomol. 32: 510Ð522.

Miller, D. R., and J. H. Borden. 1985. Life history and bi-ology of Ips latidens (Leconte) (Coleoptera: Scolytidae).Can. Entomol. 117: 859Ð871.

Pajares, J. A., and G. N. Lanier. 1990. Biosystematics of theturpentine beetles Dendroctonus terebrans and D. valens(Coleoptera: Scolytidae). Ann. Entomol. Soc. Am. 83:171Ð188.

Price, P. W. 1997. Insect ecology. John Wiley, New York.Price, P. W., N. Cobb, T. P. Craig, G. W. Fernandez, J. K.Itami, S. Mopper, and R. W. Preszler. 1990. Insect her-bivore population dynamics on trees and shrubs: newapproaches relevant to latent and eruptive species andlife table development, pp. 1Ð138. In E. A. Bernays (ed.),Insect-plant interactions, vol. II. CRC, Boca Raton, FL.

Raffa, K. F., and A. A. Berryman. 1983. The role of host plantresistance in the colonization behavior and ecology of barkbeetles (Coleoptera: Scolytidae). Ecol. Monogr. 53: 27Ð49.

Raffa, K. F., and A. A. Berryman. 1987. Interacting selectivepressures in conifer-bark beetle systems: a basis for re-ciprocal adaptations. Am. Nat. 129: 234Ð262.

Raffa, K. F., T.W. Phillips, and S.M. Salom. 1993. Strategiesand mechanisms of host colonization by bark beetles, pp.103Ð128. In T. D. Schowalter and J. M. Filip (eds.), Bee-tleÐpathogen interactions in conifer forests. Academic,London, United Kingdom.

Salinas-Moreno, Y.,M.G.Mendoza,M.A.Barrios,R.Cisneros,J.Macıas-Samano, andG. Zuniga. 2004. Areography of thegenus Dendroctonus (Coleoptera: Curculionidae: Scolyti-nae) in Mexico. J. Biogeogr. 31: 1163Ð1177.

SAS Institute. 1996. JMP starts statistics software. SAS In-stitute, Duxbury, NY.

Schenk, J. A., A. A. Berryman, and J. W. Dale. 1976. Anevaluation of a Scolytus tsugae (Coleoptera: Scolytidae)infestation in a small stand of Douglas-Þr saplings inNorthern Idaho. Can. Entomol. 108: 1079Ð1083.

Six, D. L., and T. D. Paine. 1999. Phylogenetic comparisonof ascomycete mycangial fungi and Dendroctonus barkbeetles (Coleotera: Scolytidae). Ann. Entomol. Soc. Am.92: 159Ð166.

Smith, D. M., B. C. Larson, M. J. Kelty, and P.M.S. Ashton.1997. The practice of silviculture: applied forest ecology,John Wiley, New York.

Sturgeon, K. B., and J. B. Mitton. 1982. Evolution of barkbeetle communities, pp. 350.384. In J. B. Mitton and K. B.Surgeon (eds.), Bark beetles of North American conifers:


a system for the study of evolutionary biology. Universityof Texas Press, Austin, TX.

SYSTAT. 1997. SYSTAT 7.0 software.Thomas, J. B. 1966. Some Scolytidae from the Sierra Madre

Occidental in Mexico. Can. Entomol. 98: 871Ð875.Thomas, J. B., and D. E. Bright. 1970. A new species ofDendroctonus (Coleoptera: Scolytidae) from Mexico.Can. Entomol. 102: 479Ð483.

Wood, S. L. 1982. The bark and ambrosia beetles of Northand Central America (Coleoptera: Scolytidae), a taxo-nomic monograph. Great Basin Nat. 6.

Zuniga, G., R. Cisneros, and Y. Salinas-Moreno. 1998. Chro-mosome numbers of Dendroctonus parallelocollis and D.rhizophagus (Coleoptera: Scolytidae) and new reports ofDendroctonus populations from Mexico and Guatemala.Ann. Entomol. Soc. Am. 91: 392Ð394.

Zuniga, G., R. Cisneros, J. L. Hayes, and J. Macias-Samano.2002. Karology, geographic distribution, and origin of thegenus Dendroctonus Ericsson (Coleoptera: Scolytidae).Ann. Entomol. Soc. Am. 95: 267Ð275.

Received 24 March 2009; accepted 8 June 2009.


host prefference and attack pattern of dendroctonus rhizophagus

Documents