Journal of Chemical Ecology, Vol. 7, No. 1, 1981
HOST SELECTION BEHAVIOR OF BARK BEETLES (COLEOPTERA: SCOLYTIDAE) ATTACKING Pinus ponderosa, WITH SPECIAL EMPHASIS ON THE WESTERN PINE BEETLE, Dendroctonus brevicomis 1
H E N R Y A. M O E C K , 2'5 D A V I D L. W O O D , 3 and KENNETH Q. LINDAHL, JR. 4 2Department of the Environment, Canadian Forestry Service, Pacific Forest Research Centre, 506 West Burnside Road, Victoria, British Columbia, V8Z lM5, Canada 3Department of Entomological Sc&nces, University of California, Berkeley, California 94720 4Group in Bostatistics, University of California, Berkeley, California 94720 (Received November 27, 1979; revised February 20, 1980)
Abstract--Detection of weakened hosts from a distance by bark beetles through olfaction was investigated in field experiments. No significant numbers of Scolytidae were attracted to anaerobically treated pine bolts, stem disks, or sugar and ponderosa pine bark including phloem. Treatment of living trees with cacodylic acid induced attacks by Dendroctonus brevicomis, D. ponderosae, Ips latidens, Gnathotrichus retusus, and l~'tyophthorus scalptor, beginning two weeks after treatment. There was no significant difference between landing rates of D. brevicomis and D. ponderosae on screened treated trees and screened controls. There was a significant increase in landing rates of G. retusus and I. latidens, because both species had penetrated the screen and produced pheromones. Tree frilling alone did not increase the landing rate of bark beetles. Freezing of the lower trunk with dry ice did not increase significantly the landing rate of D. brevicomis, D. ponderosae, G. retusus, or I. latidens on screened trees, whereas unscreened frozen trees were attacked by all four species. There was no significantly higher landing rate by D. brevicomis, D. ponderosae, L ~These studies were supported in part by the U.S.D.A. Forest Service, Cooperative State Research Services (2598-RRF,W-l 10) and the National Science Foundation and Environmental Protection Agency through a grant (NSF GB-34719/BM575-04223) to the University of California; and by the Canada Department of the Environment. The findings, opinions and recommendations are not necessarily those of the University of California or the funding agencies. 5From a thesis submitted by H.A. Moeck to the University of California, Berkeley, in partial fulfillment of the requirements for the degree Doctor of Philosophy in Entomology. 49 0098-0331/ 81/ 01004)049503.00/0 9 1981PlenumPublishingCorporation
50
MOECK ET AL.
paraconfusus, L latidens, G. retusus, or Hylurgops subcostulatus on screened trees evidencing symptoms of severe infection by the root pathogen Verticicladiella wagenerii, than on symptomless trees. These experiments show that D. brevicornis, D. ponderosae, L paraconfusus, L latidens, and G. retusus land, apparently indiscriminately, on healthy and stressed hosts. Thus, in these species host discrimination must occur after landing and prior to sustained feeding. Key Words--Primary attraction, tree predisposition, Dendroctonus, Ips, Gnathotrichus, Pityophthorus, Coleoptera, Scolytidae, Buprestidae, Ver-
ticicladiella, Pinus ponderosa.
INTRODUCTION
Bark and ambrosia beetles have been shown to be selective in their colonization of host plants, both with respect to the tree species and the physiological condition of the tree. With few exceptions (Rudinsky, 1962) bark beetles infest trees that have been weakened by biotic or abiotic agents such as drought (Blackman, 1924; Craighead, 1925; Ferrell, 1978; Hall, 1958; Kalkstein, 1976; King, 1972; Merker, 1952; St. George, 1929, 1930; Thomas, 1957), excess rainfall or flooding (Hetrick, 1949; Kalkstein, 1976; King, 1972; Lorio and Hodges, 1968), winter injury (Waiters, 1955), lightning (Anderson and Anderson, 1968; Hodges and Pickard, 1971; Johnson, 1966a,b; Lorio and Yandle, 1978; Schmitz and Taylor, 1969; St. George, 1930), smog (Stark et al., 1968), mechanical injury (Bennett, 1965; Gara and Holsten, 1975; Hines and Heikkenen, 1977; St. George, 1930; Thatcher, 1960), fire (Furniss, 1965; Hall and Eaton, 1961; Miller and Patterson, 1927; Swain, 1968), competition (Clements, 1953; Coulson et al., 1974; Sartwell and Stevens, 1975; Thomas, 1957), defoliation (Dewey et al., 1974; Drouin and Turnock, 1967; Schultz and Allen, 1977; Wickman, 1963; Wright, 1976), diseases (Bega et al., 1966; Cobb et al., 1974; Davidson, 1964; Felix et al., 1971; Ferrell, 1974; Ferrell and Smith, 1976; Goheen, 1976; Hertert et al., 1975; Hetrick, 1949; Jftrgensen and Petersen, 1951; Lorio, 1966; Merker, 1952; Nuorteva and Laine, 1968; Partridge and Miller, 1972; Stark and Cobb, 1969; Wagener and Cave, 1946; Wagener and Mielke, 1961; Wright et al., 1956), and herbicides (Chansler et al., 1970; Drew, 1977; Goldman et al., 1978; McGhehey and Nagel, 1967; Newton and Holt, 1971; Oliver, 1970; Svihra, 1974). Time of tree felling influences the subsequent incidence or severity of attacks by ambrosia beetles (Annila, 1975; Dyer, !967; Dyer and Chapman, 1965) and other insects (Johnson, 1964; Johnson and Zingg, 1969; Morley, 1939; Pfeffer, 1957). The trap tree method of bark beetle control is based upon the preference of various beetle species for trees treated in various ways, such as felling in sun or shade, girdling, and injection of the herbicide, cacodylic acid (Buffam, 1971; Buffam et al., 1973; Buffam and Yasinski, 1971; Frye and Wygant, 1971; Massey and
HOST SELECTION BEHAVIOR OF BARK BEETLES
51
Wygant, 1954; Nagel et al., 1957; Sedlaczek, 1921; Stelzer, 1970; Svihra, 1968; Whitten and Baker, 1939). Tree risk-rating systems were based on the observation that trees of a certain age, vigor, and appearance were preferentially killed by bark beetles (Johnson, 1972; Keen, 1936, 1943, 1946; Keen and Salman, 1942; Mogren, 1955; Person, 1928; Salman and Bongberg, 1942; Schenk and Benjamin, 1964; Struble, 1965). The mechanism of the above selectivity has long been postulated to be initiated by odors emanating from the host which the insects recognize and follow to the source. This has been termed "primary attraction," and efforts of many researchers have been directed toward the isolation and identification of these attractants (Adlung, 1958, 1960; Chararas, 1959; D~issler and Henker, 1959; Heikkenen and Hrutfiord, 1965; Kangas et al., 1967; Merker, 1955; Rudinsky, 1966; Yasunaga, 1962). In most cases, however, there is a lack of experimental evidence that primary attraction was occurring in the field, as opposed to the alternative hypothesis of initial random landing of beetles on various hosts and nonhosts, followed by sustained feeding on suitable hosts (Wood, 1972, and references cited therein). The primary attraction picture was further complicated by the discovery that many scolytid species produce powerful attractant pheromones (Borden et al., 1975). This makes the interpretation of earlier studies difficult, since it is not known whether strict precautions were taken to keep bark beetles from producing pheromones in trees or logs evaluated for primary attraction. Primary attraction has been demonstrated in the ambrosia beetles Trypodendron lineaturn (Olivier) (Borden et al., 1968; Chapman, 1962, 1966; Francia and Graham, 1967; Francke and Heeman, 1974; Graham, 1968; Moeck, 1970b), Xyloterus (Trypodendron) domesticus L. (Kerck, 1972), Xyleborus saxeseni (Ratzeburg) (Souto, 1974), Gnathotrichus sulcatus (LeConte) (Borden and Stokkink, 1973; Cade et al., 1970; Chapman, 1966), and G. retusus (LeConte) (Chapman, 1966), as well as the bark beetles Hylastes nigrinus (Mannerheim) (Chapman, 1966; Rudinsky and ZethnerMr 1967; Souto, 1974), Dryocoetes autographus (Ratzeburg) (Chapman, 1966), Pseudohylesinus nebulosus (LeConte) (Rudinsky, 1966; Souto, 1974; Stoszek, 1973), P. grandis Swaine (Rudinsky, 1966), P. granulatus (LeConte) (Souto, 1974), Leperisinus fraxini Pz..(Sch6nherr, 1970), Scolytus multistriatus (Marsham) (Peacock et al., 1971), S. quadrispinosus Say (Goeden and Norris, 1964), Ips typographus L. (Rudinsky et al., 1971), Dendroctonus pseudotsugae Hopkins (Chapman, 1964; Jantz and Rudinsky, 1966; Rudinsky, 1966), and D. rufipennis (Kirby) (Moeck, 1978). In all of the above studies, cut host material was used. Further, attractant compounds have been identified for T. lineatum (Bauer and Vit~, 1975; Moeck, 1970b, 1971; Nijholt and Schfnherr, 1976), X. domesticus (Kerck, 1972), and G. sulcatus (Cade et al., 1970). Conversely, other field experiments in which cut host material was
52
MOECK ET AL,
presented have failed to demonstrate primary attraction for Dendroctonus brevicomis LeConte (Vitd and Gara, 1962), D. frontalis Zimm. (Vit~ and Pitman, 1968), D. rufipennis (Gara and Holsten, 1975), Ips paraconfusus Lanier (Wood, 1963; Wood and Vit6; 1961), L calligraphus Germ. (Wilkinson, 1964), L avulsus Eichh. (Vit6 et al., 1964), L grandicollis Eichh. (Vitd et al., 1964; Wilkinson, 1964), L pini (Say) (Anderson, 1948), L acuminatus Gyll. (Bakke, 1967), L borealis Swaine, and Polygraphus rufipennis (Kirby) (Gara and Holsten, 1975). Furthermore, the only studies to date which utilized standing trees naturally predisposed to bark beetle attack have also failed to demonstrate primary attraction for Dendroctonus ponderosae Hopkins and D. brevicomis (Wood, 1972, 1976). From the above, it would appear that some scolytid species are capable of orienting to weakened hosts through primary attraction, whereas other species either do not have or use this capability, at least as far as can be determined with field methods used to date. However, Person (1931) and others (Miller and Keen, 1960) reported that D. brevicomis was attracted to fermenting pine phloem in laboratory experiments. Also, preliminary tests using olfactometers designed by Moeck (1970a) and Wood and Bushing (1963) indicated that some individuals of D. brevicomis, D. ponderosae, and L pini responded positively to odors of Pinus ponderosa Laws. phloem treated anaerobically for approximately 7 hr by the method of Graham (1968) (Moeck, unpublished). It could therefore be argued that since these bark beetle species appear to be capable of responding to host odors in the laboratory, the negative field results were due to the lack of attractants in the cut host material or trees at the time they were presented, or beetles were not flying at the time of the test. In bolts, the age since cutting may have been too little or too great. Living trees, although in a class shown to be highly susceptible to bark beetle infestation (Wood, 1972, 1976), may not have been attractive at the time (both within and between seasons) the studies were conducted. The following field experiments on primary attraction were designed to test these possibilities by anaerobic treatment of host material (Graham, 1968), artificial predisposition to bark beetle attack of living, apparently healthy trees, and selection of severely diseased trees that had a high probability of being attacked by bark beetles during the current season (Goheen, 1976). M E T H O D S AND M A T E R I A L S
Experiment 1 Treatment I, Aug. 17-20, 1971. A living P. ponderosa, 25 cm diameter at breast height (dbh), was felled and cut into 75-cm-long bolts. Five bolts were treated anaerobically in the following manner: the bolts were placed into a
HOST SELECTION BEHAVIOR OF BARK BEETLES
53
clean 55-gal steel drum, which was evacuated with two small compressors in series on the vacuum mode to a pressure of approximately 70 mm Hg and left outside for 12 hr overnight. Five bolts were placed in a coldroom at approximately 4~ for the same period to serve as untreated controls. In order to monitor for flight activity of D. brevicomis, two bolts were prepared by manually infesting them with 25 females each (Bedard et al., 1969). All bolts were covered with 1.6 • 1.4-mm mesh aluminum screen to prevent volunteer attacks. The bolts were placed upright on 10-gal drums used as support platforms and surrounded by I m e of Stikem-Special| 6.4-ram mesh hardware cloth (Bedard and Browne, 1969). The bolts were deployed approximately 90 m apart in two lines on ridges about 0.4 km apart near Bass Lake, Madera County, California. All insects longer than 1 mm were picked from the traps after 18 and 42 hr. Treatment 2, Aug. 25-Sept. 8, 1971. A living P. ponderosa, 23 cm dbh, was felled and cut into seven 70-cm-long bolts. Five bolts were treated anaerobically as in treatment 1 for 24 hr. Two D. brevicomis monitor bolts were prepared as above. The next day another living P. ponderosa of similar diameter was cut into five bolts to serve as untreated controls. All bolts were screened to keep out volunteers. The bolts were placed upright on the ground or on old stumps at the forest edges, Blodgett Forest Research Station, E1 Dorado County, California, and surrounded with 0.74 m 2 of Stikem-coated hardware cloth. All insects longer than 1 mm were picked from the traps after l, 2, 5, 7, and 13 days. Treatment 3, Sept. 1-15, 1971. A living sugar pine ( Pinus lambertiana Dougl.), 30 cm dbh, was cut and all the bark, including phloem, from the stem below the live crown was removed. Half of the approximately 7 m 2 of bark was treated anaerobically as in treatment 1 for 21 hr, and the remaining half was left in a screened box as the control. The treated and untreated bark was stacked on the ground in a forest opening at Blodgett Forest in the vicinity of logging slash containing the brood of Ips spp. and other bark beetle species. The bark was covered with 0.71 X 0.71 X 0.91-m screen cages, and one empty cage was used as another control. The sides and top of each cage were covered with 2.5 m 2 of Stikemcoated hardware cloth. The cages were placed at the vertices of an equilateral triangle with 46-m sides. After 13 days, the insects were washed from the traps with hot kerosene (Browne, 1978). Treatrnent 4, Sept. 14-24, 1971. A living P. ponderosa, 30 cm dbh, was felled, and the bark, including phloem, was removed. Half of the approximately 6 m 2 of bark was treated anaerobically as in treatment 1 for 20 hr, and the remaining half was left in a screened box as a control. The treated and untreated bark was stacked inside 0.36 X 0.74 • 0.74-m cardboard boxes, the tops of which were covered with 1.6 • 1.4-mm mesh aluminum screen; one
54
MOECK ET AL.
empty box served as additional control. One square meter of Stikem-coated hardware cloth was placed in triangular fashion on top of each box. The boxes were placed at the vertices of an equilateral triangle with 30-m sides immediately inside the forest edge at Blodgett Forest. After 9 days, the insects were washed from the traps with hot kerosene. Samples of fresh bark and that treated anaerobically in the drum and separately in a desiccator in the laboratory for the same time period were frozen for subsequent gaschromatographic analysis of ethanol content (Moeck, 1970). Treatment 5, Oct. 8-14, 1971. A living P. ponderosa, 36 cm dbh, was felled; part of the lower stem was cut into 5-cm-thick disks and part of the upper stem was cut into four bolts 0.6 m long. Half of the disks were treated anaerobically as in treatment 1 for 20 hr and the remaining half was left as a control. Two D. brevicomis and two I. paraconfusus monitor bolts were prepared as in treatment 1, with 30 females and 20 males each, respectively. The bolts were screened to keep out volunteers. Untreated and treated disks were placed in four screened boxes as in treatment 4 and, with two empty control boxes, were arranged in two triangular groups in the Pilgrim Creek area of the McCloud Flats near McCloud, Siskiyou County, California. The monitor bolts were placed upright on the ground. Bolts and boxes were each provided with 0.5 m 2 of Stikem-coated hardware cloth. To monitor for ambrosia beetles, six traps with 950 ml of 10% ethanol in water and six water control traps, each with 0.25 m 2 of Stikem-coated hardware cloth, were deployed in the same area, 30-60 m from the boxes. At each site, a test and a control trap were placed approximately 5 m apart on 30-cm squares of corrugated cardboard on the ground (Moeck, 1971). After 5 days, the insects were washed from the traps with hot kerosene. One flesh and one anaerobically treated disk were frozen for subsequent gas-chromatographic analysis of ethanol content. Gas Chromatography. Wood and bark samples frozen in experiment I, treatments 4 and 5, were freeze-dried (Moeck, 1970), and the condensates obtained were analyzed on a Varian Aerograph model 2700, using a 6.35-mm-OD • 1.52-m copper column packed with 100/120 mesh Porapak Q| Nitrogen carrier gas was applied at a pressure of 17 psi; air and hydrogen flow rates to the flame-ionization detector were 300 and 30 ml/min, respectively. Injector temperature was 200~ column temperature was 120~C, and detector temperature was 202 ~C; sample size was 5 #1. Ethanol at 1%, 0.1%, 0.01% and 0.001% in water (v/v) were used as standards.
Experiment 2 July 19-Oct. 27, 1972. This study was conducted in flat terrain (McCloud Flats) near McCloud, California, in young densely stocked P. ponderosa stands containing high-level populations of D. brevicomis in some areas. Five
HOST SELECTION BEHAVIOR OF BARK BEETLES
55
plots were selected, each with the following treatments: (1) an untreated tree; (2) an untreated tree with the bole screened to the base of live crown with 1.6 • 1.4-mm mesh aluminum screen; (3) a tree with an axe frill at the base; (4) a tree treated with 150-200 ml of undiluted Silvisar| 510 (cacodylic acid; several times the recommended dose was used to ensure rapid tree kill; Oliver, 1970) in an axe frill at the base; (5) a tree treated with cacodylic acid as in 4 and the bole screened; (6) a tree with the lower stem packed with dry ice to interrupt the upward flow of water and thus stimulate drought stress; (7) a tree with the lower stem packed with dry ice as in 6 and the bole screened. Trees were screened as follows (Figure 1): the trees were climbed with the aid of Swedish cone-picking ladders (with the pointed spacers padded to prevent damage to the bole or screen) and pruned of dead branches to the base of the live crown. The outer bark of the stem just below the first live branch was then shaved smooth with a drawknife and wrapped with two layers of 5cm-wide heating duct tape. The aluminum screen (purchased in 1.22 X 30.5-m rolls) was cut to proper length, with one edge tapered to fit the tree, and then stapled at 30-cm intervals along one inside edge to lath strips. The screen was hoisted to the top with braided nylon line looped over a branch and tacked to the tree at the proper height (Wood, 1976). The ladders were then removed, the screen wrapped around the tree and the ladders repositioned. The screen was then stapled to the tree at the prepared top strip and to the lath strips at 2to 3-cm intervals down the length of the tree. Alternatively, after the ladders were removed, the stapling was done by a man controlling his own descent in a harness attached to a rope looped around the stem and over a live branch. At the base, the screen was stapled and taped to the smoothed bark. Lower stem freezing was accomplished as follows: most of the outer bark was removed with a drawknife to a height of about 0.6 m. A corrugated cardboard sleeve, 15 cm greater in diameter than the stem and 30 cm long, was then stapled, tied, and taped to the tree just above the root collar and covered with 5-cm-thick foil-faced fibreglass building insulation. Approximately 30 kg crushed dry ice were placed in the sleeve which was then tied with a drawstring. The dry ice was replenished after 1- to 3-day intervals. Tree moisture stress was monitored with a pressure b o m b (PMS Instrument Co., Corvallis, Oregon); readings were taken on previous years' needle fascicles (Johnson and Nielsen, 1969) from branch tips cut with pole pruners or shot down with a .222-caliber rifle equipped with a telescopic sight. Moisture stress was determined at dawn, and on Aug. 21, at 1000, 1200, 1400, and 1600 hr (PDT). Comparison readings were taken on nearby healthy untreated trees. All trees were provided with 6 Stikem-coated hardware cloth traps, each 30.5 X 61 cm, two each at 2 m, midbole and base of live crown; middle and upper traps were raised and lowered with braided nylon line (after the method developed by Bedard, unpublished). Total trapping surface was 1.12 m 2/ tree. Traps were placed on the trees on July 28 (plot 5) and July 31 (plots 1-4).
56
M O E C K ET AL.
FIG. 1. Screened diseased ponderosa pine, experiment 3, Blodgett Forest, California, 1973.
HOST SELECTION BEHAVIOR OF BARK BEETLES
57
Cacodylic acid treatment and frilling treatment were applied on Aug. 8 (plot 5) and Aug. 9 (plots 1-4). Freezing treatment was applied discontinuously between Aug. 15 and Sept. 29. Scolytidae, Cleridae, Trogositidae (Coleoptera), and Siricidae (Hymenoptera) were picked from the traps after 9, 17, 22, 39, and 53 days and all insects were washed from the traps with hot kerosene after 88 days. In August 1973, all cacodylic acid-treated and frozen trees were cut, and the status of insect infestations beneath the bark was determined. All other trees were checked for phloem and crown condition. Trap catch data were analyzed as follows: pairwise comparisons under the different treatment regimes were made with the Wilcoxon signed rank test for matched pairs (Lehmann, 1975, Chap. 3). Matching was based on trapping period and plot, i.e., comparison pairs consisted of trap catches during the same period on trees in the same plot. A large number of ties and zeros occurred in the data, necessitating the use of midranks. Since the published tables for the signed rank test are generally incorrect when midranks are used, a computer program was developed to calculate the exact significance probabilities of test results. 6
Experiment 3 April 26-Sept. 14, 1973. This study was conducted at the University of California Blodgett Forest Research Station in the central Sierra Nevada mountains in the Gaddis Creek basin (elev. 1300 m). Forest cover is a mixed conifer type, with P. ponderosa 50-80 years old predominating. A high population of D. brevicomis has persisted in this area since at least 1955 and appears to be intimately associated with a spreading infection of the root pathogen, Vertieicladiella wagenerii Kendrick (Cobb, et al., 1974; Stark and Dahlsten, 1970). A 4-year study by Goheen (1976) has indicated that a tree with 50% or more of its circumference at the root collar stained by V. wagenerii, as determined by arch punch sampling early in the season, had a 0.45 probability of being attacked by bark beetles in the same year. This method was therefore chosen to select trees naturally predisposed to bark beetle attack and to study landing rates throughout the season. In late April and early May, two groups of trees were chosen: 20 healthy P. ponderosa (mean dbh 34.7 + 1.9 cm) showing no V. wageneriistain in a 5% linear sample (6.4-mm arch punch sample every 12.7 cm) at the root collar, and 35 severely diseased P. ponderosa with 50% or more of the samples showing V. wagenerii stain in the sapwood. Trees surrounding the study trees were also sampled to ensure the absence of severely diseased trees which, if attacked by beetles, would influence landing rates on the study trees. Twenty 6For a discussion of the difficulty in dealing with ties and zeros, see Lehmann (1975) pp. 18-23, 60, 129-132. Information about the computer program may be obtained by contacting K. Q. Lindahl.
58
MOECK ET AL.
of the diseased trees (D) (mean dbh 34.6 +__ 1.5 cm, mean percentage V. wagenerii stain 74.9 + 2.9) were screened (DS) to the base of the live crown by the ladder method as described in experiment 2 (Figure 1). The healthy (H) and remaining diseased trees (D) (mean dbh 44.2 + 2.6 cm, mean percentage V. wagenerii stain 62.6 ___3.2) were left unscreened (HU and DU), and all were pruned of dead branches to 11 m with pole pruners. The latter group served as a check on the proportion of diseased trees coming under insect attack. All trees were provided with six Stikem-coated hardware cloth traps, each 30.5 • 61 cm, two each at 1 m, midbole, and base of live crown; middle and upper traps were raised and lowered as described in experiment 2. Total trapping surface was 1.12 m2/tree. All insects longer than 1 mm, except Psocoptera, were picked from the traps after each of 12 trapping periods which varied from 1 to 2 weeks in duration. Collections 2-5 were from half of the traps on each tree, whereas collection 6 was a cumulative count for periods 2-5 inclusive from the other half of the traps. Collections 1 and 7-12 were from all traps, for a total of 13 collections. In September, screened diseased trees with fading foliage or insect attack in the crown were cut, and the status of the insect infestation was determined. The landing rate data for each insect species under investigation were statistically analyzed as follows: the observations were first sorted into 3 groups: namely, groups HU, DS, and DU, containing data from healthy unscreened, diseased screened, and diseased unscreened trees, respectively, except those subject to direct insect attack or pheromone interference from nearby attacked trees. Evidence of this was provided by boring dust on the traps or under the screen, presence of fresh pitch tubes on the tree or galleries in the bark as revealed by tree dissection at the end of the experiment; data from these trees for the affected trapping periods were placed into a fourth group A (attacked). All observations were then standardized to a trap-treeday basis by dividing the numbers of insects caught by the number of days between observations. The resulting landing rates for groups HU, DS, and DU were then compared by the Kruskal-Wallis test (Lehmann, 1975). When this test was significant, indicating differences among the groups, multiple comparisons were made to determine which pairs of groups differed from one another (Miller, 1966). Groups HU and A were compared by the Wilcoxon rank-sum test (Lehmann, 1975). Where data were available the sexes were analyzed separately. All tests were judged significant if the significance probability was less than a = 0.05. RESULTS
Experiment 1 Very few Scolytidae were caught on any of the test materials except on traps containing manually infested bolts (Tables 1-3). These beetles were
HOST SELECTION BEHAVIOR OF BARK BEETLES
59
TABLE 1. CATCH OF SCOLYTIDAE ON STICKY TRAPS CONTAINING PONDEROSA PINE BOLTS THAT WERE UNTREATED, ANAEROBICALLY TREATED, OR INFESTED WITH D, b r e v i c o m i s FEMALES, EXPERIMENT 1, TREATMENTS 1 AND 2, BASS LAKE AND BLODGETT FOREST~ CALIFORNIA, AUGUST-SEPTEMBER 1971
No. and species of Scolytidae trapped at
Treatment no.
Total trapping time
5 bolts untreated
1 2
42 hr 13 days
1 9 D. valens 1 9 Lparaeonfusus
2 bolts infested with 25 9 D. brevicomis each ~
5 bolts treated anaerobically b
13c~ 119 D.brevicornis 14c~ 9 9 D. brevicomis I cJ G. retusus
0 0
~18 of 25 and 19 of 25 females in treatment 1, and 21 of 25 and 18 of 25 females in treatment 2, producing frass. bAnaerobic treatment 12 and 24 hr in treatments 1 and 2, respectively.
responding to pheromones produced by female D. brevicomis (Table 1) as has been demonstrated by Bedard et al. (1969), and by male I. paraconfusus (Table 3), as shown by Wood and Vitd (1961). Failure to catch D. brevicomis on female-infested bolts in treatment 5 (Table 3) remains unexplained, but Bedard and Wood (unpublished) found that such bolts were attractive in spring and early summer, but not in the fall, the season during which this test was conducted. Thus in treatment 5 no conclusion with regard to attraction of D. brevicomis can be drawn, since the beetles' presence was not verified. The few scolytids caught on untreated or treated host materials probably arrived by chance, since as many were trapped on the control traps that contained no host materials. The gas chromatography results indicate that the large-scale anaerobic treatment of bolts, bark, and wood disks was ineffective in producing the high (0.4%) ethanol concentration obtained in earlier studies (Moeck, 1970b). The reasons for this are not known. The extract from P. ponderosa bark frozen 4 hr after cutting contained 0.011% ethanol and after 20 hr of anaerobic treatment in the drum it contained 0.026%, whereas the sample from the desiccator in the laboratory contained 0.40%. Similarly, treatment of the P. ponderosa disks increased the ethanol concentration only slightly from 0.003% to 0.006%. However, the absence of a response to the 10% ethanol-inwater mixture indicates that ethanol produced by anaerobiosis would not be attractive by itself even at 0.4%, the maximum concentration expected under natural conditions (Moeck, 1970b). However, other products produced under these conditions may have been attractive. Although no proof was obtained that ambrosia beetles and bark beetles other than D. brevicornis and L paraconfusus were flying during the
13
9
3
4
1 9 D. brevicomis 1 Pseudopityophthorus sp.
1 c~ L paraconfusus 1 Xyle6orus scopulorum 1 Hylastes graeilis 1 9 D. brevicomis 0
Untreated bark ~
Empty cage
Anaerobically treated bark b
0
2 Scolytus sp. 1 Pseudopityophthorus sp.
aP. tarnbertiana bark in treatment 3, P. ponderosa bark in treatment 4. bAnaerobic treatment of P. lambertiana bark for 21 hr in treatment 3, and of P. ponderosa bark for 20 hr in treatment 4.
Total trapping time (days)
Treatment no.
No. and species of Scolytidae trapped at
TABLE 2. CATCH OF SCOLYTIDAE ON STICKY TRAPS ON EMPTY CAGES AND ON CAGES CONTAINING UNTREATED OR ANAEROBICALLY TREATED PONDEROSA PINE OR SUGAR PINE BARK, EXPERIMENT 1, TREATMENTS 3 AND 4, BLODGETT FOREST, CALIFORNIA, SEPTEMBER 1971
(3
O~
61
HOST SELECTION BEHAVIOROF BARK BEETLES TABLE 3. CATCH OF SCOLYTIDAE ON STICKY TRAPS CONTAINING WATER, 10~ ETHANOL 1N WATER, D. brevicomis AND ]. paraconfusus INFESTED PONDEROSA PINE BOLTS AND UNTREATED AND ANAEROBICALLY TREATED PONDEROSA PINE STEM DISKS, EXPERIMENT l, TREATMENT 5, MCCLOUD, CALIFORNIA, OCTOBER 1971 ~
Treatment 2 empty cages 6 water traps 6 ethanol (10%) traps 2 P. ponderosa bolts infested with 30 9 D. brevicomis eachb 2 P. ponderosa bolts infested with 20 c~ L paraconfusus eachc 2 boxes with untreated P.ponderosa disks 2 boxes with P. ponderosa disks anaerobically treated for 20 hr
No. and speciesof Scolytidaetrapped 0 1 9 G. retusus, 1 9 G. sulcatus 0 1 Hylurgops subcostulatus 16 ~ 18 9 L paraconfusus
1 9 L paraconfusus 1 9 G. retusus 1Hylurgops reticulatus
~Total trapping time 5 days. b28 of 30 and 28 of 30 females producing frass. c16 of 20 and 18 of 20 males producing frass.
experimental period, their absence is believed to be highly unlikely at the elevation and time of year of these experiments. Further, these experiments demonstrate that very little, if any, attraction is produced in host material subjected to the most extreme stress, i.e., killing of the tree. The duration of experiment 1, treatment 2 (13 days) was believed to be adequate for natural aging processes to take place and render the bolts suitable for attack, yet no bark beetles were trapped. Host suitability was indicated by the boring activity of D . b r e v i c o m i s females in the monitor bolts and the presence of an active pair o f D. p o n d e r o s a e in an 8-cm gallery on Aug. 28 in the top of the tree cut on Aug. 25. Treatments 3-5 were an attempt to increase the surface area for evaporation of host volatiles, including terpenes, yet no significant catches were recorded for periods up to 13 days. Thus the question of whether anaerobic processes that produce higher alcohol levels in wood a n d / o r phloem results in attraction of bark beetles remains unanswered by these experiments. If primary attraction occurs, it must be at a very low level and must be produced under conditions other than those imposed in experiment 1, i.e., longer periods of aging or anaerobic treatment. Experiment
2
Tree screening did not appear to materially influence the landing rate of Scolytidae (Table 4, columns 1 and 2), confirming previous findings by Wood
4 6
2 t 10 2 0 1
Dendroctonus ponderosae d 9 Gnathotrichus retusus Hylastes gracilis Hylastes tenuis Hylurgops reticulatus
Unscreened control ~
Dendroctonus brevicomis d 9
Species
4 8 2 1 0 1
2 3
Screened control"
2 2 2 1 0 0
3 4
Unscreened frill only
16 19 379 1 1 3
8 13
Unscreened frill with cacodylic acid
9 6 38 b 5 0 4
2 2
Screened frill with cacodylic acid
Tree treatment (5 trees each) and no. of beetles trapped
13 17 28 3 0 9
13 25
Unscreened frozen
4 2 1 1 0 0
2 3
Screened frozen
TABLE 4. CATCH OF SCOLYTIDAE AND PLATYPODIDAE ON UNSCREENED AND SCREENED PONDEROSA PINES THAT WERE HEALTHY OR TREATED WITH CACODYLIC ACID, FRILLING, OR LOWER STEM FREEZING WITH DRY ICE, EXPERIMENT 2, McCLOUD, CALIFORNIA, JULY-OCTOBER 1972
> t~
b~
ON
63 629 0. 10
5 1
1 5
0 0 3
1 1
1
1 18
aIncludes control trees and test trees pr io r to treatment. bBeetles penetrated screen (see text).
Total all species No. of trap-tree-days Mean No. per trap-tree-day
Hy[urgops subcostulatus Ips latidens Phloeosinus spp. Pityophthorus scalptor P~'tyophthorus serratus Pseudohylesinus sp. Pseudopityophthorus sp. Scolytus spp. Xyleborus saxeseni Xyleborus scopulorum Platypus wilsoni d 3 3 69 580 0.12
0 0 6 0 0
0 26 0 6 4
2 3 39 394 0.10
0 8 1 5 3 0 1 0 0 2
1 21 2 4 8 1 0 1 0 0 0 1 147 355 0.4I
0 63 b 0 10 1 1 0 0 0 0 2 1 144 394 0.37
2 171 0 21 0 0 1 1 0 2 5 3 646 394 1.64
1 1 46 355 0.13
0 6 7 1 0 1 0 0
0 16 b
t~
).
9
.
HOST SELECTION BEHAVIOR OF BARK BEETLES (COLEOPTERA: SCOLYTIDAE) ATTACKING Pinus ponderosa, WITH SPECIAL EMPHASIS ON THE WESTERN PINE BEETLE, Dendroctonus brevicomis 1
H E N R Y A. M O E C K , 2'5 D A V I D L. W O O D , 3 and KENNETH Q. LINDAHL, JR. 4 2Department of the Environment, Canadian Forestry Service, Pacific Forest Research Centre, 506 West Burnside Road, Victoria, British Columbia, V8Z lM5, Canada 3Department of Entomological Sc&nces, University of California, Berkeley, California 94720 4Group in Bostatistics, University of California, Berkeley, California 94720 (Received November 27, 1979; revised February 20, 1980)
Abstract--Detection of weakened hosts from a distance by bark beetles through olfaction was investigated in field experiments. No significant numbers of Scolytidae were attracted to anaerobically treated pine bolts, stem disks, or sugar and ponderosa pine bark including phloem. Treatment of living trees with cacodylic acid induced attacks by Dendroctonus brevicomis, D. ponderosae, Ips latidens, Gnathotrichus retusus, and l~'tyophthorus scalptor, beginning two weeks after treatment. There was no significant difference between landing rates of D. brevicomis and D. ponderosae on screened treated trees and screened controls. There was a significant increase in landing rates of G. retusus and I. latidens, because both species had penetrated the screen and produced pheromones. Tree frilling alone did not increase the landing rate of bark beetles. Freezing of the lower trunk with dry ice did not increase significantly the landing rate of D. brevicomis, D. ponderosae, G. retusus, or I. latidens on screened trees, whereas unscreened frozen trees were attacked by all four species. There was no significantly higher landing rate by D. brevicomis, D. ponderosae, L ~These studies were supported in part by the U.S.D.A. Forest Service, Cooperative State Research Services (2598-RRF,W-l 10) and the National Science Foundation and Environmental Protection Agency through a grant (NSF GB-34719/BM575-04223) to the University of California; and by the Canada Department of the Environment. The findings, opinions and recommendations are not necessarily those of the University of California or the funding agencies. 5From a thesis submitted by H.A. Moeck to the University of California, Berkeley, in partial fulfillment of the requirements for the degree Doctor of Philosophy in Entomology. 49 0098-0331/ 81/ 01004)049503.00/0 9 1981PlenumPublishingCorporation
50
MOECK ET AL.
paraconfusus, L latidens, G. retusus, or Hylurgops subcostulatus on screened trees evidencing symptoms of severe infection by the root pathogen Verticicladiella wagenerii, than on symptomless trees. These experiments show that D. brevicornis, D. ponderosae, L paraconfusus, L latidens, and G. retusus land, apparently indiscriminately, on healthy and stressed hosts. Thus, in these species host discrimination must occur after landing and prior to sustained feeding. Key Words--Primary attraction, tree predisposition, Dendroctonus, Ips, Gnathotrichus, Pityophthorus, Coleoptera, Scolytidae, Buprestidae, Ver-
ticicladiella, Pinus ponderosa.
INTRODUCTION
Bark and ambrosia beetles have been shown to be selective in their colonization of host plants, both with respect to the tree species and the physiological condition of the tree. With few exceptions (Rudinsky, 1962) bark beetles infest trees that have been weakened by biotic or abiotic agents such as drought (Blackman, 1924; Craighead, 1925; Ferrell, 1978; Hall, 1958; Kalkstein, 1976; King, 1972; Merker, 1952; St. George, 1929, 1930; Thomas, 1957), excess rainfall or flooding (Hetrick, 1949; Kalkstein, 1976; King, 1972; Lorio and Hodges, 1968), winter injury (Waiters, 1955), lightning (Anderson and Anderson, 1968; Hodges and Pickard, 1971; Johnson, 1966a,b; Lorio and Yandle, 1978; Schmitz and Taylor, 1969; St. George, 1930), smog (Stark et al., 1968), mechanical injury (Bennett, 1965; Gara and Holsten, 1975; Hines and Heikkenen, 1977; St. George, 1930; Thatcher, 1960), fire (Furniss, 1965; Hall and Eaton, 1961; Miller and Patterson, 1927; Swain, 1968), competition (Clements, 1953; Coulson et al., 1974; Sartwell and Stevens, 1975; Thomas, 1957), defoliation (Dewey et al., 1974; Drouin and Turnock, 1967; Schultz and Allen, 1977; Wickman, 1963; Wright, 1976), diseases (Bega et al., 1966; Cobb et al., 1974; Davidson, 1964; Felix et al., 1971; Ferrell, 1974; Ferrell and Smith, 1976; Goheen, 1976; Hertert et al., 1975; Hetrick, 1949; Jftrgensen and Petersen, 1951; Lorio, 1966; Merker, 1952; Nuorteva and Laine, 1968; Partridge and Miller, 1972; Stark and Cobb, 1969; Wagener and Cave, 1946; Wagener and Mielke, 1961; Wright et al., 1956), and herbicides (Chansler et al., 1970; Drew, 1977; Goldman et al., 1978; McGhehey and Nagel, 1967; Newton and Holt, 1971; Oliver, 1970; Svihra, 1974). Time of tree felling influences the subsequent incidence or severity of attacks by ambrosia beetles (Annila, 1975; Dyer, !967; Dyer and Chapman, 1965) and other insects (Johnson, 1964; Johnson and Zingg, 1969; Morley, 1939; Pfeffer, 1957). The trap tree method of bark beetle control is based upon the preference of various beetle species for trees treated in various ways, such as felling in sun or shade, girdling, and injection of the herbicide, cacodylic acid (Buffam, 1971; Buffam et al., 1973; Buffam and Yasinski, 1971; Frye and Wygant, 1971; Massey and
HOST SELECTION BEHAVIOR OF BARK BEETLES
51
Wygant, 1954; Nagel et al., 1957; Sedlaczek, 1921; Stelzer, 1970; Svihra, 1968; Whitten and Baker, 1939). Tree risk-rating systems were based on the observation that trees of a certain age, vigor, and appearance were preferentially killed by bark beetles (Johnson, 1972; Keen, 1936, 1943, 1946; Keen and Salman, 1942; Mogren, 1955; Person, 1928; Salman and Bongberg, 1942; Schenk and Benjamin, 1964; Struble, 1965). The mechanism of the above selectivity has long been postulated to be initiated by odors emanating from the host which the insects recognize and follow to the source. This has been termed "primary attraction," and efforts of many researchers have been directed toward the isolation and identification of these attractants (Adlung, 1958, 1960; Chararas, 1959; D~issler and Henker, 1959; Heikkenen and Hrutfiord, 1965; Kangas et al., 1967; Merker, 1955; Rudinsky, 1966; Yasunaga, 1962). In most cases, however, there is a lack of experimental evidence that primary attraction was occurring in the field, as opposed to the alternative hypothesis of initial random landing of beetles on various hosts and nonhosts, followed by sustained feeding on suitable hosts (Wood, 1972, and references cited therein). The primary attraction picture was further complicated by the discovery that many scolytid species produce powerful attractant pheromones (Borden et al., 1975). This makes the interpretation of earlier studies difficult, since it is not known whether strict precautions were taken to keep bark beetles from producing pheromones in trees or logs evaluated for primary attraction. Primary attraction has been demonstrated in the ambrosia beetles Trypodendron lineaturn (Olivier) (Borden et al., 1968; Chapman, 1962, 1966; Francia and Graham, 1967; Francke and Heeman, 1974; Graham, 1968; Moeck, 1970b), Xyloterus (Trypodendron) domesticus L. (Kerck, 1972), Xyleborus saxeseni (Ratzeburg) (Souto, 1974), Gnathotrichus sulcatus (LeConte) (Borden and Stokkink, 1973; Cade et al., 1970; Chapman, 1966), and G. retusus (LeConte) (Chapman, 1966), as well as the bark beetles Hylastes nigrinus (Mannerheim) (Chapman, 1966; Rudinsky and ZethnerMr 1967; Souto, 1974), Dryocoetes autographus (Ratzeburg) (Chapman, 1966), Pseudohylesinus nebulosus (LeConte) (Rudinsky, 1966; Souto, 1974; Stoszek, 1973), P. grandis Swaine (Rudinsky, 1966), P. granulatus (LeConte) (Souto, 1974), Leperisinus fraxini Pz..(Sch6nherr, 1970), Scolytus multistriatus (Marsham) (Peacock et al., 1971), S. quadrispinosus Say (Goeden and Norris, 1964), Ips typographus L. (Rudinsky et al., 1971), Dendroctonus pseudotsugae Hopkins (Chapman, 1964; Jantz and Rudinsky, 1966; Rudinsky, 1966), and D. rufipennis (Kirby) (Moeck, 1978). In all of the above studies, cut host material was used. Further, attractant compounds have been identified for T. lineatum (Bauer and Vit~, 1975; Moeck, 1970b, 1971; Nijholt and Schfnherr, 1976), X. domesticus (Kerck, 1972), and G. sulcatus (Cade et al., 1970). Conversely, other field experiments in which cut host material was
52
MOECK ET AL,
presented have failed to demonstrate primary attraction for Dendroctonus brevicomis LeConte (Vitd and Gara, 1962), D. frontalis Zimm. (Vit~ and Pitman, 1968), D. rufipennis (Gara and Holsten, 1975), Ips paraconfusus Lanier (Wood, 1963; Wood and Vit6; 1961), L calligraphus Germ. (Wilkinson, 1964), L avulsus Eichh. (Vit6 et al., 1964), L grandicollis Eichh. (Vitd et al., 1964; Wilkinson, 1964), L pini (Say) (Anderson, 1948), L acuminatus Gyll. (Bakke, 1967), L borealis Swaine, and Polygraphus rufipennis (Kirby) (Gara and Holsten, 1975). Furthermore, the only studies to date which utilized standing trees naturally predisposed to bark beetle attack have also failed to demonstrate primary attraction for Dendroctonus ponderosae Hopkins and D. brevicomis (Wood, 1972, 1976). From the above, it would appear that some scolytid species are capable of orienting to weakened hosts through primary attraction, whereas other species either do not have or use this capability, at least as far as can be determined with field methods used to date. However, Person (1931) and others (Miller and Keen, 1960) reported that D. brevicomis was attracted to fermenting pine phloem in laboratory experiments. Also, preliminary tests using olfactometers designed by Moeck (1970a) and Wood and Bushing (1963) indicated that some individuals of D. brevicomis, D. ponderosae, and L pini responded positively to odors of Pinus ponderosa Laws. phloem treated anaerobically for approximately 7 hr by the method of Graham (1968) (Moeck, unpublished). It could therefore be argued that since these bark beetle species appear to be capable of responding to host odors in the laboratory, the negative field results were due to the lack of attractants in the cut host material or trees at the time they were presented, or beetles were not flying at the time of the test. In bolts, the age since cutting may have been too little or too great. Living trees, although in a class shown to be highly susceptible to bark beetle infestation (Wood, 1972, 1976), may not have been attractive at the time (both within and between seasons) the studies were conducted. The following field experiments on primary attraction were designed to test these possibilities by anaerobic treatment of host material (Graham, 1968), artificial predisposition to bark beetle attack of living, apparently healthy trees, and selection of severely diseased trees that had a high probability of being attacked by bark beetles during the current season (Goheen, 1976). M E T H O D S AND M A T E R I A L S
Experiment 1 Treatment I, Aug. 17-20, 1971. A living P. ponderosa, 25 cm diameter at breast height (dbh), was felled and cut into 75-cm-long bolts. Five bolts were treated anaerobically in the following manner: the bolts were placed into a
HOST SELECTION BEHAVIOR OF BARK BEETLES
53
clean 55-gal steel drum, which was evacuated with two small compressors in series on the vacuum mode to a pressure of approximately 70 mm Hg and left outside for 12 hr overnight. Five bolts were placed in a coldroom at approximately 4~ for the same period to serve as untreated controls. In order to monitor for flight activity of D. brevicomis, two bolts were prepared by manually infesting them with 25 females each (Bedard et al., 1969). All bolts were covered with 1.6 • 1.4-mm mesh aluminum screen to prevent volunteer attacks. The bolts were placed upright on 10-gal drums used as support platforms and surrounded by I m e of Stikem-Special| 6.4-ram mesh hardware cloth (Bedard and Browne, 1969). The bolts were deployed approximately 90 m apart in two lines on ridges about 0.4 km apart near Bass Lake, Madera County, California. All insects longer than 1 mm were picked from the traps after 18 and 42 hr. Treatment 2, Aug. 25-Sept. 8, 1971. A living P. ponderosa, 23 cm dbh, was felled and cut into seven 70-cm-long bolts. Five bolts were treated anaerobically as in treatment 1 for 24 hr. Two D. brevicomis monitor bolts were prepared as above. The next day another living P. ponderosa of similar diameter was cut into five bolts to serve as untreated controls. All bolts were screened to keep out volunteers. The bolts were placed upright on the ground or on old stumps at the forest edges, Blodgett Forest Research Station, E1 Dorado County, California, and surrounded with 0.74 m 2 of Stikem-coated hardware cloth. All insects longer than 1 mm were picked from the traps after l, 2, 5, 7, and 13 days. Treatment 3, Sept. 1-15, 1971. A living sugar pine ( Pinus lambertiana Dougl.), 30 cm dbh, was cut and all the bark, including phloem, from the stem below the live crown was removed. Half of the approximately 7 m 2 of bark was treated anaerobically as in treatment 1 for 21 hr, and the remaining half was left in a screened box as the control. The treated and untreated bark was stacked on the ground in a forest opening at Blodgett Forest in the vicinity of logging slash containing the brood of Ips spp. and other bark beetle species. The bark was covered with 0.71 X 0.71 X 0.91-m screen cages, and one empty cage was used as another control. The sides and top of each cage were covered with 2.5 m 2 of Stikemcoated hardware cloth. The cages were placed at the vertices of an equilateral triangle with 46-m sides. After 13 days, the insects were washed from the traps with hot kerosene (Browne, 1978). Treatrnent 4, Sept. 14-24, 1971. A living P. ponderosa, 30 cm dbh, was felled, and the bark, including phloem, was removed. Half of the approximately 6 m 2 of bark was treated anaerobically as in treatment 1 for 20 hr, and the remaining half was left in a screened box as a control. The treated and untreated bark was stacked inside 0.36 X 0.74 • 0.74-m cardboard boxes, the tops of which were covered with 1.6 • 1.4-mm mesh aluminum screen; one
54
MOECK ET AL.
empty box served as additional control. One square meter of Stikem-coated hardware cloth was placed in triangular fashion on top of each box. The boxes were placed at the vertices of an equilateral triangle with 30-m sides immediately inside the forest edge at Blodgett Forest. After 9 days, the insects were washed from the traps with hot kerosene. Samples of fresh bark and that treated anaerobically in the drum and separately in a desiccator in the laboratory for the same time period were frozen for subsequent gaschromatographic analysis of ethanol content (Moeck, 1970). Treatment 5, Oct. 8-14, 1971. A living P. ponderosa, 36 cm dbh, was felled; part of the lower stem was cut into 5-cm-thick disks and part of the upper stem was cut into four bolts 0.6 m long. Half of the disks were treated anaerobically as in treatment 1 for 20 hr and the remaining half was left as a control. Two D. brevicomis and two I. paraconfusus monitor bolts were prepared as in treatment 1, with 30 females and 20 males each, respectively. The bolts were screened to keep out volunteers. Untreated and treated disks were placed in four screened boxes as in treatment 4 and, with two empty control boxes, were arranged in two triangular groups in the Pilgrim Creek area of the McCloud Flats near McCloud, Siskiyou County, California. The monitor bolts were placed upright on the ground. Bolts and boxes were each provided with 0.5 m 2 of Stikem-coated hardware cloth. To monitor for ambrosia beetles, six traps with 950 ml of 10% ethanol in water and six water control traps, each with 0.25 m 2 of Stikem-coated hardware cloth, were deployed in the same area, 30-60 m from the boxes. At each site, a test and a control trap were placed approximately 5 m apart on 30-cm squares of corrugated cardboard on the ground (Moeck, 1971). After 5 days, the insects were washed from the traps with hot kerosene. One flesh and one anaerobically treated disk were frozen for subsequent gas-chromatographic analysis of ethanol content. Gas Chromatography. Wood and bark samples frozen in experiment I, treatments 4 and 5, were freeze-dried (Moeck, 1970), and the condensates obtained were analyzed on a Varian Aerograph model 2700, using a 6.35-mm-OD • 1.52-m copper column packed with 100/120 mesh Porapak Q| Nitrogen carrier gas was applied at a pressure of 17 psi; air and hydrogen flow rates to the flame-ionization detector were 300 and 30 ml/min, respectively. Injector temperature was 200~ column temperature was 120~C, and detector temperature was 202 ~C; sample size was 5 #1. Ethanol at 1%, 0.1%, 0.01% and 0.001% in water (v/v) were used as standards.
Experiment 2 July 19-Oct. 27, 1972. This study was conducted in flat terrain (McCloud Flats) near McCloud, California, in young densely stocked P. ponderosa stands containing high-level populations of D. brevicomis in some areas. Five
HOST SELECTION BEHAVIOR OF BARK BEETLES
55
plots were selected, each with the following treatments: (1) an untreated tree; (2) an untreated tree with the bole screened to the base of live crown with 1.6 • 1.4-mm mesh aluminum screen; (3) a tree with an axe frill at the base; (4) a tree treated with 150-200 ml of undiluted Silvisar| 510 (cacodylic acid; several times the recommended dose was used to ensure rapid tree kill; Oliver, 1970) in an axe frill at the base; (5) a tree treated with cacodylic acid as in 4 and the bole screened; (6) a tree with the lower stem packed with dry ice to interrupt the upward flow of water and thus stimulate drought stress; (7) a tree with the lower stem packed with dry ice as in 6 and the bole screened. Trees were screened as follows (Figure 1): the trees were climbed with the aid of Swedish cone-picking ladders (with the pointed spacers padded to prevent damage to the bole or screen) and pruned of dead branches to the base of the live crown. The outer bark of the stem just below the first live branch was then shaved smooth with a drawknife and wrapped with two layers of 5cm-wide heating duct tape. The aluminum screen (purchased in 1.22 X 30.5-m rolls) was cut to proper length, with one edge tapered to fit the tree, and then stapled at 30-cm intervals along one inside edge to lath strips. The screen was hoisted to the top with braided nylon line looped over a branch and tacked to the tree at the proper height (Wood, 1976). The ladders were then removed, the screen wrapped around the tree and the ladders repositioned. The screen was then stapled to the tree at the prepared top strip and to the lath strips at 2to 3-cm intervals down the length of the tree. Alternatively, after the ladders were removed, the stapling was done by a man controlling his own descent in a harness attached to a rope looped around the stem and over a live branch. At the base, the screen was stapled and taped to the smoothed bark. Lower stem freezing was accomplished as follows: most of the outer bark was removed with a drawknife to a height of about 0.6 m. A corrugated cardboard sleeve, 15 cm greater in diameter than the stem and 30 cm long, was then stapled, tied, and taped to the tree just above the root collar and covered with 5-cm-thick foil-faced fibreglass building insulation. Approximately 30 kg crushed dry ice were placed in the sleeve which was then tied with a drawstring. The dry ice was replenished after 1- to 3-day intervals. Tree moisture stress was monitored with a pressure b o m b (PMS Instrument Co., Corvallis, Oregon); readings were taken on previous years' needle fascicles (Johnson and Nielsen, 1969) from branch tips cut with pole pruners or shot down with a .222-caliber rifle equipped with a telescopic sight. Moisture stress was determined at dawn, and on Aug. 21, at 1000, 1200, 1400, and 1600 hr (PDT). Comparison readings were taken on nearby healthy untreated trees. All trees were provided with 6 Stikem-coated hardware cloth traps, each 30.5 X 61 cm, two each at 2 m, midbole and base of live crown; middle and upper traps were raised and lowered with braided nylon line (after the method developed by Bedard, unpublished). Total trapping surface was 1.12 m 2/ tree. Traps were placed on the trees on July 28 (plot 5) and July 31 (plots 1-4).
56
M O E C K ET AL.
FIG. 1. Screened diseased ponderosa pine, experiment 3, Blodgett Forest, California, 1973.
HOST SELECTION BEHAVIOR OF BARK BEETLES
57
Cacodylic acid treatment and frilling treatment were applied on Aug. 8 (plot 5) and Aug. 9 (plots 1-4). Freezing treatment was applied discontinuously between Aug. 15 and Sept. 29. Scolytidae, Cleridae, Trogositidae (Coleoptera), and Siricidae (Hymenoptera) were picked from the traps after 9, 17, 22, 39, and 53 days and all insects were washed from the traps with hot kerosene after 88 days. In August 1973, all cacodylic acid-treated and frozen trees were cut, and the status of insect infestations beneath the bark was determined. All other trees were checked for phloem and crown condition. Trap catch data were analyzed as follows: pairwise comparisons under the different treatment regimes were made with the Wilcoxon signed rank test for matched pairs (Lehmann, 1975, Chap. 3). Matching was based on trapping period and plot, i.e., comparison pairs consisted of trap catches during the same period on trees in the same plot. A large number of ties and zeros occurred in the data, necessitating the use of midranks. Since the published tables for the signed rank test are generally incorrect when midranks are used, a computer program was developed to calculate the exact significance probabilities of test results. 6
Experiment 3 April 26-Sept. 14, 1973. This study was conducted at the University of California Blodgett Forest Research Station in the central Sierra Nevada mountains in the Gaddis Creek basin (elev. 1300 m). Forest cover is a mixed conifer type, with P. ponderosa 50-80 years old predominating. A high population of D. brevicomis has persisted in this area since at least 1955 and appears to be intimately associated with a spreading infection of the root pathogen, Vertieicladiella wagenerii Kendrick (Cobb, et al., 1974; Stark and Dahlsten, 1970). A 4-year study by Goheen (1976) has indicated that a tree with 50% or more of its circumference at the root collar stained by V. wagenerii, as determined by arch punch sampling early in the season, had a 0.45 probability of being attacked by bark beetles in the same year. This method was therefore chosen to select trees naturally predisposed to bark beetle attack and to study landing rates throughout the season. In late April and early May, two groups of trees were chosen: 20 healthy P. ponderosa (mean dbh 34.7 + 1.9 cm) showing no V. wageneriistain in a 5% linear sample (6.4-mm arch punch sample every 12.7 cm) at the root collar, and 35 severely diseased P. ponderosa with 50% or more of the samples showing V. wagenerii stain in the sapwood. Trees surrounding the study trees were also sampled to ensure the absence of severely diseased trees which, if attacked by beetles, would influence landing rates on the study trees. Twenty 6For a discussion of the difficulty in dealing with ties and zeros, see Lehmann (1975) pp. 18-23, 60, 129-132. Information about the computer program may be obtained by contacting K. Q. Lindahl.
58
MOECK ET AL.
of the diseased trees (D) (mean dbh 34.6 +__ 1.5 cm, mean percentage V. wagenerii stain 74.9 + 2.9) were screened (DS) to the base of the live crown by the ladder method as described in experiment 2 (Figure 1). The healthy (H) and remaining diseased trees (D) (mean dbh 44.2 + 2.6 cm, mean percentage V. wagenerii stain 62.6 ___3.2) were left unscreened (HU and DU), and all were pruned of dead branches to 11 m with pole pruners. The latter group served as a check on the proportion of diseased trees coming under insect attack. All trees were provided with six Stikem-coated hardware cloth traps, each 30.5 • 61 cm, two each at 1 m, midbole, and base of live crown; middle and upper traps were raised and lowered as described in experiment 2. Total trapping surface was 1.12 m2/tree. All insects longer than 1 mm, except Psocoptera, were picked from the traps after each of 12 trapping periods which varied from 1 to 2 weeks in duration. Collections 2-5 were from half of the traps on each tree, whereas collection 6 was a cumulative count for periods 2-5 inclusive from the other half of the traps. Collections 1 and 7-12 were from all traps, for a total of 13 collections. In September, screened diseased trees with fading foliage or insect attack in the crown were cut, and the status of the insect infestation was determined. The landing rate data for each insect species under investigation were statistically analyzed as follows: the observations were first sorted into 3 groups: namely, groups HU, DS, and DU, containing data from healthy unscreened, diseased screened, and diseased unscreened trees, respectively, except those subject to direct insect attack or pheromone interference from nearby attacked trees. Evidence of this was provided by boring dust on the traps or under the screen, presence of fresh pitch tubes on the tree or galleries in the bark as revealed by tree dissection at the end of the experiment; data from these trees for the affected trapping periods were placed into a fourth group A (attacked). All observations were then standardized to a trap-treeday basis by dividing the numbers of insects caught by the number of days between observations. The resulting landing rates for groups HU, DS, and DU were then compared by the Kruskal-Wallis test (Lehmann, 1975). When this test was significant, indicating differences among the groups, multiple comparisons were made to determine which pairs of groups differed from one another (Miller, 1966). Groups HU and A were compared by the Wilcoxon rank-sum test (Lehmann, 1975). Where data were available the sexes were analyzed separately. All tests were judged significant if the significance probability was less than a = 0.05. RESULTS
Experiment 1 Very few Scolytidae were caught on any of the test materials except on traps containing manually infested bolts (Tables 1-3). These beetles were
HOST SELECTION BEHAVIOR OF BARK BEETLES
59
TABLE 1. CATCH OF SCOLYTIDAE ON STICKY TRAPS CONTAINING PONDEROSA PINE BOLTS THAT WERE UNTREATED, ANAEROBICALLY TREATED, OR INFESTED WITH D, b r e v i c o m i s FEMALES, EXPERIMENT 1, TREATMENTS 1 AND 2, BASS LAKE AND BLODGETT FOREST~ CALIFORNIA, AUGUST-SEPTEMBER 1971
No. and species of Scolytidae trapped at
Treatment no.
Total trapping time
5 bolts untreated
1 2
42 hr 13 days
1 9 D. valens 1 9 Lparaeonfusus
2 bolts infested with 25 9 D. brevicomis each ~
5 bolts treated anaerobically b
13c~ 119 D.brevicornis 14c~ 9 9 D. brevicomis I cJ G. retusus
0 0
~18 of 25 and 19 of 25 females in treatment 1, and 21 of 25 and 18 of 25 females in treatment 2, producing frass. bAnaerobic treatment 12 and 24 hr in treatments 1 and 2, respectively.
responding to pheromones produced by female D. brevicomis (Table 1) as has been demonstrated by Bedard et al. (1969), and by male I. paraconfusus (Table 3), as shown by Wood and Vitd (1961). Failure to catch D. brevicomis on female-infested bolts in treatment 5 (Table 3) remains unexplained, but Bedard and Wood (unpublished) found that such bolts were attractive in spring and early summer, but not in the fall, the season during which this test was conducted. Thus in treatment 5 no conclusion with regard to attraction of D. brevicomis can be drawn, since the beetles' presence was not verified. The few scolytids caught on untreated or treated host materials probably arrived by chance, since as many were trapped on the control traps that contained no host materials. The gas chromatography results indicate that the large-scale anaerobic treatment of bolts, bark, and wood disks was ineffective in producing the high (0.4%) ethanol concentration obtained in earlier studies (Moeck, 1970b). The reasons for this are not known. The extract from P. ponderosa bark frozen 4 hr after cutting contained 0.011% ethanol and after 20 hr of anaerobic treatment in the drum it contained 0.026%, whereas the sample from the desiccator in the laboratory contained 0.40%. Similarly, treatment of the P. ponderosa disks increased the ethanol concentration only slightly from 0.003% to 0.006%. However, the absence of a response to the 10% ethanol-inwater mixture indicates that ethanol produced by anaerobiosis would not be attractive by itself even at 0.4%, the maximum concentration expected under natural conditions (Moeck, 1970b). However, other products produced under these conditions may have been attractive. Although no proof was obtained that ambrosia beetles and bark beetles other than D. brevicornis and L paraconfusus were flying during the
13
9
3
4
1 9 D. brevicomis 1 Pseudopityophthorus sp.
1 c~ L paraconfusus 1 Xyle6orus scopulorum 1 Hylastes graeilis 1 9 D. brevicomis 0
Untreated bark ~
Empty cage
Anaerobically treated bark b
0
2 Scolytus sp. 1 Pseudopityophthorus sp.
aP. tarnbertiana bark in treatment 3, P. ponderosa bark in treatment 4. bAnaerobic treatment of P. lambertiana bark for 21 hr in treatment 3, and of P. ponderosa bark for 20 hr in treatment 4.
Total trapping time (days)
Treatment no.
No. and species of Scolytidae trapped at
TABLE 2. CATCH OF SCOLYTIDAE ON STICKY TRAPS ON EMPTY CAGES AND ON CAGES CONTAINING UNTREATED OR ANAEROBICALLY TREATED PONDEROSA PINE OR SUGAR PINE BARK, EXPERIMENT 1, TREATMENTS 3 AND 4, BLODGETT FOREST, CALIFORNIA, SEPTEMBER 1971
(3
O~
61
HOST SELECTION BEHAVIOROF BARK BEETLES TABLE 3. CATCH OF SCOLYTIDAE ON STICKY TRAPS CONTAINING WATER, 10~ ETHANOL 1N WATER, D. brevicomis AND ]. paraconfusus INFESTED PONDEROSA PINE BOLTS AND UNTREATED AND ANAEROBICALLY TREATED PONDEROSA PINE STEM DISKS, EXPERIMENT l, TREATMENT 5, MCCLOUD, CALIFORNIA, OCTOBER 1971 ~
Treatment 2 empty cages 6 water traps 6 ethanol (10%) traps 2 P. ponderosa bolts infested with 30 9 D. brevicomis eachb 2 P. ponderosa bolts infested with 20 c~ L paraconfusus eachc 2 boxes with untreated P.ponderosa disks 2 boxes with P. ponderosa disks anaerobically treated for 20 hr
No. and speciesof Scolytidaetrapped 0 1 9 G. retusus, 1 9 G. sulcatus 0 1 Hylurgops subcostulatus 16 ~ 18 9 L paraconfusus
1 9 L paraconfusus 1 9 G. retusus 1Hylurgops reticulatus
~Total trapping time 5 days. b28 of 30 and 28 of 30 females producing frass. c16 of 20 and 18 of 20 males producing frass.
experimental period, their absence is believed to be highly unlikely at the elevation and time of year of these experiments. Further, these experiments demonstrate that very little, if any, attraction is produced in host material subjected to the most extreme stress, i.e., killing of the tree. The duration of experiment 1, treatment 2 (13 days) was believed to be adequate for natural aging processes to take place and render the bolts suitable for attack, yet no bark beetles were trapped. Host suitability was indicated by the boring activity of D . b r e v i c o m i s females in the monitor bolts and the presence of an active pair o f D. p o n d e r o s a e in an 8-cm gallery on Aug. 28 in the top of the tree cut on Aug. 25. Treatments 3-5 were an attempt to increase the surface area for evaporation of host volatiles, including terpenes, yet no significant catches were recorded for periods up to 13 days. Thus the question of whether anaerobic processes that produce higher alcohol levels in wood a n d / o r phloem results in attraction of bark beetles remains unanswered by these experiments. If primary attraction occurs, it must be at a very low level and must be produced under conditions other than those imposed in experiment 1, i.e., longer periods of aging or anaerobic treatment. Experiment
2
Tree screening did not appear to materially influence the landing rate of Scolytidae (Table 4, columns 1 and 2), confirming previous findings by Wood
4 6
2 t 10 2 0 1
Dendroctonus ponderosae d 9 Gnathotrichus retusus Hylastes gracilis Hylastes tenuis Hylurgops reticulatus
Unscreened control ~
Dendroctonus brevicomis d 9
Species
4 8 2 1 0 1
2 3
Screened control"
2 2 2 1 0 0
3 4
Unscreened frill only
16 19 379 1 1 3
8 13
Unscreened frill with cacodylic acid
9 6 38 b 5 0 4
2 2
Screened frill with cacodylic acid
Tree treatment (5 trees each) and no. of beetles trapped
13 17 28 3 0 9
13 25
Unscreened frozen
4 2 1 1 0 0
2 3
Screened frozen
TABLE 4. CATCH OF SCOLYTIDAE AND PLATYPODIDAE ON UNSCREENED AND SCREENED PONDEROSA PINES THAT WERE HEALTHY OR TREATED WITH CACODYLIC ACID, FRILLING, OR LOWER STEM FREEZING WITH DRY ICE, EXPERIMENT 2, McCLOUD, CALIFORNIA, JULY-OCTOBER 1972
> t~
b~
ON
63 629 0. 10
5 1
1 5
0 0 3
1 1
1
1 18
aIncludes control trees and test trees pr io r to treatment. bBeetles penetrated screen (see text).
Total all species No. of trap-tree-days Mean No. per trap-tree-day
Hy[urgops subcostulatus Ips latidens Phloeosinus spp. Pityophthorus scalptor P~'tyophthorus serratus Pseudohylesinus sp. Pseudopityophthorus sp. Scolytus spp. Xyleborus saxeseni Xyleborus scopulorum Platypus wilsoni d 3 3 69 580 0.12
0 0 6 0 0
0 26 0 6 4
2 3 39 394 0.10
0 8 1 5 3 0 1 0 0 2
1 21 2 4 8 1 0 1 0 0 0 1 147 355 0.4I
0 63 b 0 10 1 1 0 0 0 0 2 1 144 394 0.37
2 171 0 21 0 0 1 1 0 2 5 3 646 394 1.64
1 1 46 355 0.13
0 6 7 1 0 1 0 0
0 16 b
t~
).
9
.
