Influence of Plant Host Quality on Fitness and Sex Ratio of the Wheat Stem Sawfly (Hymenoptera: Cephidae)

Transcription

1 PLANTÐINSECT INTERACTIONS Influence of Plant Host Quality on Fitness and Sex Ratio of the Wheat Stem Sawfly (Hymenoptera: Cephidae) H. A. CÁRCAMO, 1, 2 B. L. BERES, 1 F. CLARKE, 3 R. J. BYERS, 1 H.-H. MÜNDEL, 1 K. MAY, 1 AND R. DEPAUW 3 Environ. Entomol. 34(6): 1579Ð1592 (2005) ABSTRACT A resurgence of the wheat stem sawßy, historically the most important pest of wheat in the northern Great Plains, has been observed in western Canada over the past decade. Host plant resistance in the form of solid-stemmed cultivars remains the primary management strategy for this pest. The objectives of this study were to determine the effects of wheat cultivar on wheat stem sawßy Þtness and sex ratio. The cultivars studied varied with respect to stem diameter and pith expression and included representatives of the red seed, solid-stemmed spring and red seed, hollow-stemmed common and durum wheat classes. We present results from a primary study conducted in southern Alberta from 2001 to 2003 and a similar smaller study conducted in 1987 and All of the solid-stemmed cultivars (AC Eatonia, AC Abbey, Lancer, Leader) reduced female weights, size, fecundity, and in some cases, larval survivorship in the cut stubs and delayed date of adult emergence in the laboratory. Males were less responsive to this aspect of host quality. The hollow-stemmed durum cultivar AC Navigator had similar negative effects and deserves further study. A number of other hollow-stemmed wheats (McKenzie, AC Intrepid, Katepwa, and the durum AC Avonlea) had intermediate but inconsistent negative effects on sawßy Þtness. The varieties that maximized sawßy Þtness were AC Cadillac, CDC Teal, and Kyle (durum). Sawßy sex ratios were affected by stub diameter in our current and historical cultivar studies. Larger diameter stubs produced signiþcantly more females and smaller diameter stubs produced more males. Furthermore, stubs that failed to produce adults had signiþcantly smaller diameters than those that produced females but similar to those that produced males. The effect of the solid stem trait on sex ratio was inconsistent. Only Lancer had a male-biased sex ratio, whereas the other varieties had no consistent pattern. These results, however, are similar to other published reports that have noted inconsistent effects of the solid trait on sex ratio. Planting a solid-stemmed cultivar in a rotation that includes broad-leaved crops is recommended to reduce sawßy damage and future populations. KEY WORDS wheat stem sawßy, host quality, fecundity, sex ratio, solid-stemmed wheat THE WHEAT STEM SAWFLY, Cephus cinctus Norton, has been a pest of wheat on the northern Great Plains of North America (Fig. 1) since the early days of settlement (Fletcher 1896). A considerable body of literature has been amassed on its basic biology and management over the past 90 yr (see reviews by Holmes 1979, 1982, Weiss and Morrill 1992). Its taxonomy and origin, however, remains a subject of some debate (Ivie and Zinovjev 1996). A brief overview of the biology and management based on these reviews is presented here. Sawßies overwinter as diapausing mature larvae in cut stubs just above the crown but below ground level. In Canada, adults emerge, mate, and oviposit from the middle of June to early July. Males 1 Agriculture and Agri-Food Canada, Research Center, Lethbridge, Alberta, Canada. 2 Corresponding author: Agriculture and Agri-Food Canada, Lethbridge Research Centre, PO Box 3000, 5403Ð1 Ave S., Lethbridge, AB, Canada T1J 4B1 ( carcamoh@agr.gc.ca). 3 Agriculture and Agri-Food Canada, Research Center, Swift Current, Saskatchewan, Canada. appear a few days before females and are more common along the edges of the Þelds. The sawßy is a weak ßyer, which further enhances the edge effect. Larvae mine the top internodes Þrst. When the kernel moisture reaches around 50% (usually during the middle of August in southern Canada), the larvae migrate to the bottom internodes using light for orientation (Holmes 1975) and cut a characteristic ring a few centimeters above the crown. They plug the stem at the ring with frass and construct an overwintering chamber. Sawßy larvae in diapause are very cold hardy and can supercool to 22 C (Salt 1944). The grain yield loss caused by larval mining is twofold. Larval tunneling interferes with nutrient transfer capability of the conductive tissues and also weakens the stems. The results are smaller, lighter kernels with an estimated weight loss of 5Ð30% (Wallace and Mc- Neal 1966). Lighter kernels result in loss of weight per unit volume, which is a grading factor that determines the price received by the producer. The most severe form of loss is caused by the stems that are girdled and X/05/1579Ð1592$04.00/ Entomological Society of America

2 1580 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 Fig. 1. Historical area (shaded portion) of economic damage to wheat by the wheat stem sawßy in the northern Great Plains of North America. Study sites indicated by the numbers. break off and lodge to the ground (Holmes 1982). Furthermore, losses increase as a result of increased harvesting time, fuel, and equipment wear associated with the extra costs of gathering up stems that have lodged to the ground. These costs are directly proportional to the proportion of the stems lodged over. In Canada, the sawßy caused severe losses to wheat until the 1950s when a solid-stemmed cultivar with high resistance to the larva, appropriately named Rescue, was introduced (Platt et al. 1948). The genetic resistance of solid-stemmed cultivars, combined with climatic factors, reduced sawßy populations below economic levels for the following few decades in the Canadian prairies. As a result, farmers (and plant breeders) shifted their attention to higher yielding, susceptible cultivars. Small resurgences of this pest were observed in the 1980s (Byers and May 1991); however, during the 1990s, populations increased steadily and have now reached damaging levels throughout much of its historical range in Alberta and Saskatchewan. Several factors may have contributed to its increase, the most important being the widespread planting of susceptible thick, hollow-stemmed cultivars over the past 40 yr. Furthermore, reduced tillage practices currently leave the stubble relatively undisturbed after harvest, which may increase the overwintering success of larvae and the emergence of adults the following spring throughout the entire Þeld. Consecutive dry years in some areas (1999Ð2000) may have also contributed to the resurgence through negative effects on natural enemies (Criddle 1922, Holmes 1982). Host plant tolerance in the form of solid-stemmed cultivars still represents the primary management strategy to reduce damage and suppress populations of the wheat stem sawßy. Results from plant breeding and agronomic studies have shown that solid-stemmed cultivars such as those currently recommended in Canada, AC Abbey and AC Eatonia, can reduce lodging levels by 50% or more relative to susceptible cultivars (DePauw et al. 1994). However, none of the currently recommended solid-stemmed and susceptible cultivars available in Canada have been studied in detail to determine their effects on the Þtness of adult sawßy and potential effects on population dynamics. Effects of host on sex ratios and Þtness of sawßy can be expected from the known responses of the sawßy to previously registered cultivars in Canada (Holmes 1982) and the United States (Weiss and Morrill 1992). Stem diameter has been shown to inßuence sex of sawßy with more females produced from thicker stems and more males from thinner stems in some studies (Wall 1952, Morrill et al. 2000) but not in others (McGinnis 1950, Jacobson and Farstad 1952, Holmes and Peterson 1963). Because previous solidstemmed cultivars had thinner stems than hollowstemmed cultivars, one would expect a male-biased sex ratio in these cultivars. However, previous Þndings have been inconclusive. For example, Holmes and Peterson (1963) reported cultivar effects, although they failed to detect stem diameter effects, and Farstad et al. (1949) also reported cultivar effects on sex ratios but made no reference to stem diameters. Wall (1952), however, found a clear pattern of more females from thick stems but did not consider cultivar effects in her work. Female fecundity is also reduced by solid-stemmed cultivars. According to early studies in Lethbridge, Canada, females produced an average of 27 and 37 eggs in the cultivars Rescue (solid) and Thatcher (hollow), respectively (Holmes 1979, 1982). Holmes also stated that, in general, females from thinner stems had fewer eggs. Morrill et al. (2000) showed that thin stems of the winter wheat cultivar Redwin produced signiþcantly smaller females with fewer eggs and reduced longevity relative to those that emerged

3 December 2005 CÁRCAMO ET AL.: HOST EFFECTS ON C. cinctus 1581 Table 1. Stem lumen, wheat class, and literature reference for wheat cultivars assessed for influence of host stem quality on fitness and sex ratio of the wheat stem sawfly Cultivar Stem type Class Reference Current cultivar study AC Avonlea Hollow CWAD Clarke et al AC Navigator Hollow CWAD Clarke et al Kyle Hollow CWAD Townley-Smith et al AC Barrie Hollow CWRS McCaig et al AC Cadillac Hollow CWRS DePauw et al AC Intrepid Hollow CWRS DePauw et al CDC Teal Hollow CWRS Hughes and Hucl 1993 Katepwa Hollow CWRS Campbell and Czarnecki 1987 McKenzie Hollow CWRS Graf et al AC Abbey Solid CWRS DePauw et al AC Eatonia Solid CWRS DePauw et al Historical cultivar study Lancer Solid CWRS DePauw et al 1986 Leader Solid CWRS DePauw et al Laura Hollow CWRS DePauw et al Katepwa Hollow CWRS Campbell and Czarnecki 1987 Biggar Hollow CPS-red DePauw et al Fielder Hollow CWSWS Sunderman and Bruinsma 1975 Red Bobs Hollow Red spring DePauw et al CWAD, Canada Western Amber Durum; CWRS, Canada Western Red Spring; CPS-red, Canada Prairie Spring Red; CWSWS, Canada Western Soft White Spring. from thicker stems. Solid-stemmed winter wheat cultivars have also been shown to negatively impact Þtness of the wheat stem sawßy (Morrill et al. 1994) in Montana where this insect has adjusted its lifecycle to become a pest of winter wheat (Morrill and Kushnak 1996). Our objectives were to compare the effects of a number of wheat cultivars on wheat stem sawßy Þtness and sex ratio. We selected cultivars that represented a cross-section of solid and hollow-stemmed Canada Western Red Spring class (CWRS) and durum wheats. Durum cultivars (e.g., AC Navigator) were of particular interest because this class has historically shown some inherent resistance to lodging relative to hollow-stemmed CWRS cultivars (Wallace and McNeal 1966, Bouhssini et al. 1987, Miller et al. 1993). The variety McKenzie contains some solidstem heritage from Amidon (Kappel 1997, Graf 2003), and was of particular interest given anecdotal observations of resistance from industry and previous agronomic trials (Beres, Cárcamo, and Byers, unpublished data). Materials and Methods The main portion of this work was conducted from 2001 to 2003 and will be referred to as the current cultivars study, which used recently released cultivars. We also report on two other smaller studies conducted in 1987 and 1991, which are referred to as the historical cultivars study. Current Cultivars Study Sites. Impacts of wheat cultivar on sawßy were studied from 2001 to 2003 at the permanent sawßy nursery of Agriculture and Agri-Food Canada, established in the 1950s (Peterson et al. 1968), located near, Alberta. is 10 km west of Lethbridge in the dark-brown soil zone and mixed grass prairie ecozone (Ecological StratiÞcation Working Group 1995). In 2001 and 2002, we sampled plots near the village of, 100 km southeast of Lethbridge, in southern Alberta. All sites occur within the area of historical sawßy damage (Fig. 1) and have experienced a severe resurgence in sawßy populations since Plant Material. Cultivars evaluated at most sites from 2001 to 2003 (Table 1) included currently grown CWRS (bread wheat) hollow-stemmed, susceptible to sawßy damage (AC Barrie, CDC Teal, AC Cadillac, AC Intrepid, Katepwa, McKenzie), the two most recently registered solid-stemmed CWRS cultivars (also bread wheat) that are more resistant to sawßy damage (AC Eatonia and AC Abbey) and three durum cultivars (pasta wheat) grown by many producers in the area (Kyle, AC Avonlea, and AC Navigator). Experimental Design, Agronomic Practices, and Sampling. The experimental design consisted of a randomized block with all cultivars appearing once in each of four to six replicate blocks. The experimental unit consisted of a single row,3 m in length and spaced 46 cm apart. Blocks were planted on 5-m centers and buffered by 2 m of winter wheat. The plots were planted in the third week of April of each year using a self-propelled plot seeder conþgured with doubledisc seed openers. Herbicides were chosen based on the weed spectrum present and applied in early June at label rates. No insecticides were applied at any of the sites. To assess the impact of cultivar on sawßy adult Þtness, we excavated the plants from a 1-m row in late May or early June from treatments grown the previous season. Plants were stored 2 wkat10 C until processed. Stems cut by the sawßy larvae, hereafter referred to as stubs, had their top outside diameter

4 1582 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 measured as the average of three angles (0, 90, and 135 ) with a digital caliper (thickness is used hereafter to indicate stem diameter and not the stem wall). Stubs were placed in individual plastic centrifuge vials (Diamed Lab Supplies) at 20 C, under a photoperiod of a 16:8-h light-dark regimen. After emergence, the adult sawßies were allowed2dtomature before killing them by freezing at 20 C. Response variables included Julian date of emergence, sex, fresh weight, and right and left forewing lengths measured from the base of the thorax to the tip using an ocular micrometer. Females were dissected to determine egg numbers. Historical Cultivars In a preliminary study of the relationship between host stem diameter and the sex and size of the resultant adult sawßies, stubs containing overwintered larvae were collected in April 1987 from plots of Red Bobs at the site. Red Bobs is a highly susceptible hollow-stemmed cultivar (DePauw et al. 1995) of historical interest that was used to maintain a high population of sawßy. The top outside diameter of each stub was measured with a micrometer caliper and placed in a vial at room temperature. Adults were killed by freezing at 20 C on the day of emergence, sexed, and weighed, and their wing lengths measured with an ocular micrometer. On 22 May 1991, the solid-stemmed cultivars, Leader and Lancer (CWRS), and the hollow-stemmed cultivars Laura and Katepwa (CWRS), Biggar (Canada Prairie Spring Red), and Fielder (Soft White Spring), were sampled from the nursery to assess impacts on sawßy Þtness (Table 1). At least 30 adults of each sex were reared individually from each cultivar (except for Biggar) following the methodology described above. Sixty uncut stems were randomly selected from each variety, and the outside stem diameters were measured at the middle of the third internode. Stem solidness of these uncut stems was quantiþed from cross-sections, cut by hand with a sharp single-edge razor blade, midway between the second and third basal nodes, and visually rated from 1 to 5: (1) hollow, (2) lumen less than one-thirds Þlled, (3) lumen onethird to two-thirds Þlled, (4) lumen more than twothirds Þlled, and (5) solid. All adults were sexed, and a subsample of 10 frozen females was weighed from each variety. On 13 April 1992, these same cultivars were sampled from the above nursery to compare effects on sawßy sex ratios. At least 200 stems from each variety were pooled from six plots, and the cut stubs were separated from the clumps and placed in a cage at 22 C and a 16:8-h light-dark photoperiod. Data Analysis SAS (version 8.2) PROC MIXED (SAS Institute 1996) was performed on each of the sawßy response variables (y ijk ) measured in each location in each year. A split plot analysis with subsamples and the main plots (cultivars) in a randomized complete block design was performed on the sex-differentiated variables (stub diameter, emergence date, wing size, and body weight). Sawßy sex was the subplot and the number of samples of each sex was the subsample. The adult survival recorded for each stub, 0 (dead) and 1 (alive), were analyzed with PROC GLIMMIX with the ERROR BINOMIAL and the LINK PROBIT (Littell et al. 1996). Analysis of covariance (ANCOVA) was performed with stem diameter as the covariate to determine the effect of cultivars if they had the same stem diameter. The form of the covariate was determined by Þrst testing whether each slope was equal to zero. If not equal to zero, we tested whether the slopes were equal to each other, and if so, we used the common slope model described by Littell et al. (1996). Depending on the model, cultivar, sawßy sex, and cultivar by sawßy sex were considered Þxed, and replications, replication by cultivar, replication by sawßy sex, and replication by cultivars by sawßy sex were considered random. Satterthwaites degrees of freedom (Steel and Torrie 1980) were estimated with the option DDEM SATTERTH. Comparisons were performed with the t-test in the PDIFF option, which used the appropriate error term for each difference between two treatment means (SE) (Steel and Torrie 1980). Differences were considered signiþcant at 0.05 and meaningful when (Sokal and Rohlf 1981). Only a priori tests for cultivars of interest, as deþned in the objectives, were considered when the F-test was not signiþcant (Chew 1976). The number of subsamples available in each plot varied; therefore, different errors terms were required for each combination of difference. The appropriate errors were used for the differences discussed but to save space the range of the SEs of the differences between treatments means is presented to provide an indication of the variation. Pearsons correlations between adult weight and wing length and between egg load and female fecundity in each year and location were calculated for each trait to determine the consistency of response in different environments. For the historical study conducted in 1987, contingency tables and 2 tests of association were performed (SAS version 5; SAS Institute 1985) on sex of sawßy (male and female) against date of emergence (before 8 May, 8Ð9 May, 10Ð11 May, and after 12 May) and against stem diameter ( 2.2, 2.2Ð2.5, 2.5Ð2.7, 2.7 mm). Furthermore, the diameters of stems that produced males, females, or no adults and the weight of males and females were compared using the nonparametric Krukal-Wallis test (Steel and Torrie 1980). The relationships among stem diameter, weight, body, and wing lengths with eggs were studied with correlation analyses. For the data from adults that emerged in 1991, we also relied on the Kruskall-Wallis test (Steel and Torrie 1980) to compare the cultivars because replicate plots were not sampled separately. These latter analyses were performed using Systat version 10.2 (Systat Software 2002). In 1992, sex ratios were only compared numerically because replicate plots were pooled for each cultivar inside a cage.

5 December 2005 CÁRCAMO ET AL.: HOST EFFECTS ON C. cinctus 1583 Table 2. Cultivars used in study of sawfly fitness, no. dead larvae (d), and sample sizes of adults Site and year Variety Adults d Adults d Adults d a Adults d a Hard Red Spring Solid stems AC Abbey NP NP b 58 37ab AC Eatonia NP NP a 24 24ab Hollow stems AC Barrie d 57 23abc AC Cadillac d 70 24bcd AC Intrepid NP NP cd 65 33ab Katepwa bc 88 30bcd McKenzie bc 99 16cd CDC Teal bc d Durums (hollow) AC Avonlea NP NP NP NP 6 13cd 14 18bcd AC Navigator NP NP NP NP 6 12bc 9 25a Kyle NP NP NP NP 12 18cd 7 29ab a Statistical tests using PROC GLIMMIX (SAS Institute 1996), revealed highly signiþcant effects (P 0.05) of cultivar on mortality at, df 9,27, F 3.64 in 2002 and df 10,48, F 3.32 in Cultivars not sharing letters indicate signiþcant differences between the average mortality (proportions means not shown). NP, cultivar not planted. Results Current Cultivars The number of stubs and adults that emerged from each meter of row sampled varied. Stubs from the cultivars AC Eatonia, and to a lesser extent, AC Abbey, AC Navigator, and AC Intrepid, produced few adults (Table 2). The diameter of the durum stubs was thicker than the bread wheats at in 2002 and 2003, but the differences were signiþcant only in 2002 (P 0.05; Table 3). Stem diameter affected the sex of the sawßy produced in each environment with males emerging from smaller diameter stems fairly consistently to signiþcantly so, depending on the site (P 0.05; Table 3). Stem diameter of the cultivars did not signiþcantly interact with sex. However, stubs that failed to produce adults had signiþcantly smaller diameters than those that produced females but did not differ from those that produced males (P 0.05). Fewer adults emerged from the solid stem varieties (Table 2). In fact, at two of the three site-years where the solid-stemmed cultivars were included in the study, no adults were recovered (AC Eatonia at 2002, AC Abbey at 2002). Relative to the hollow-stemmed cultivars, there was a trend toward high numbers of dead larvae in stubs of AC Navigator (durum), which formed a cluster along with the solid-stemmed cultivars during both years of the study at (Table 2). Emergence date was signiþcantly earlier for males than females at three of the four site-years (Table 4). There was a trend, although not signiþcant, of slower development in AC Eatonia compared with AC Cadillac in 2002 at (Table 4). At in 2003, adult sawßies from the hollow-stemmed cultivars, AC Avonlea and AC Cadillac, emerged signiþcantly earlier than those that developed in the solid-stemmed AC Eatonia (Table 4; P 0.05). This was the only site and year where sex had no effect on emergence date. Adult weights and wing lengths were signiþcantly correlated to each other at all sites and years (R values ranged from 0.74 to 0.99; P 0.05); therefore, only weights are presented because this variable is more accurately measured (Table 5). The weight of the adult sawßies differed signiþcantly (P 0.05) among cultivars at in 2001 and in 2002 after adjusting for stem diameters. Males were consistently smaller than females, regardless of cultivar, although in some instances the differences were not signiþcant. Solid-stemmed cultivars produced signiþcantly to numerically smaller adults than hollow-stemmed cultivars at in 2002 and 2003 with or without stem diameter as a covariate. McKenzie ( 2002), AC Avonlea ( 2002), and Katepwa ( 2002) produced slightly larger adults than AC Abbey. Males generally responded little to cultivars (data not shown). Egg load was signiþcantly correlated with female weights (R values ranged from 0.76 to 0.99; all P 0.05; Fig. 2). Hence, the responses of egg load (Table 6) to cultivars were very similar to the responses of the weight variable (Table 5). Cultivar had a signiþcant effect on egg load at two of the four site-years (Table 6); however, the solid-stemmed cultivars were not planted at in 2001 and were underrepresented at both sites in 2002 (Table 2). Stem diameter had a signiþcant covariate effect (P 0.01) at three site-years, with in 2003 being the exception; for this year location, the slope of the covariate equaled zero so unadjusted data are presented (Table 6). At in 2001, cultivar and diameter affected egg loads (P 0.05), and when adjusted to a common stem diameter, McKenzie females with an estimated average of 19 eggs differed signiþcantly

6 1584 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 Table 3. stem sawfly Least square means of diameters (mm) of stubs for cultivars at the two study sites used to assess cultivar effects on wheat Cultivar Hollow durums AC Avonlea NP NP 2.69 bc 2.63 a AC Navigator NP NP 2.79 ab 2.45 a Kyle NP NP 2.90 a 2.71 a Hollow bread wheats AC Barrie 2.05 a 2.10 a 2.49 cd 2.54 a AC Cadillac 2.22 a 2.01 a 2.46 d 2.58 a AC Intrepid NP 1.60 a 2.42 d 2.49 a CDC Teal 2.40 a 2.16 a 2.52 cd 2.66 a Katepwa 2.31 a 1.94 a 2.48 cd 2.62 a McKenzie 2.13 a 2.05 a 2.49 cd 2.62 a Solid bread wheats AC Abbey NP NE a 2.44 d 2.52 a AC Eatonia NP 1.95 a NE 2.52 a Means of sex dead NA 1.79 A 2.55 A 2.55 A Across female 2.39 B 2.03 B 2.64 B 2.68 B Cultivars male 2.05 A 1.87 AB 2.52 A 2.50 A SED Ð Ð Ð Ð0.10 SED Ð Ð Ð0.042 Statistical results: Num df Den df F Test Num df Den df F Test Num df Den df F Test Num df Den df F Test Cultivar (C) ** Sex(S) ** ** ** C*S Sex and cultivar had no interactions; therefore, sex, including stems with dead larvae, were averaged. *, ** F-test signiþcant at P 0.05 and 0.01, respectively. Means not sharing letters within the same case type are signiþcantly different (P 0.05). 2 Range of SE of the difference between mean of sex of each cultivar when they varied. 3 Range of SE of the difference between means of sexes presented when they varied. Abbreviations: NP, not planted; NE, no emergence of adults; NA, data not collected. from AC Barrie, CDC Teal, and AC Cadillac. This latter cultivar had an average of 43 eggs per female. Katepwa females averaged 26 eggs, which was not signiþcantly different from McKenzie but differed signiþcantly from AC Cadillac (P 0.05). No solidstemmed cultivars were planted at in In 2002 at, cultivar had no signiþcant effect on egg load (Table 6); however, the cultivar AC Abbey was effective at killing all the larvae in the few stubs found (Table 2). AC Eatonia was almost as effective in killing the sawßy larvae because only two larvae survived to become females, which had relatively low fecundity at an average of 17 eggs (Table 6). At in 2002, cultivar and stem diameter had a highly signiþcant effect on egg load (Table 6). No adults emerged from the samples collected from the solid-stemmed cultivar AC Eatonia. In the other solidstemmed cultivar, AC Abbey, females produced an average of 25 eggs, which differed signiþcantly from those that developed in the following cultivars: Kyle, AC Barrie, AC Cadillac, and CDC Teal. AC Avonlea, AC Navigator, and AC McKenzie were signiþcantly lower than Kyle, AC Barrie, AC Cadillac, and CDC Teal. Katepwa females, averaging 33 eggs, were signiþcantly lower than AC Cadillac and CDC Teal (Table 6; P 0.05). In 2003 at, cultivar had a meaningful effect, only at P 0.06, but stem diameter did not impact egg loads (P 0.10). Overall egg loads were higher than other location-years and ranged from an average of 31Ð32 in the two solid-stemmed cultivars to an average of 41Ð44 in CDC Teal, Katepwa, and AC Cadillac. Females from the cultivars AC Barrie and the two solid-stemmed cultivars had signiþcantly lower fecundity than females that emerged from CDC Teal, Katepwa, and AC Cadillac (P 0.05; Table 6). Historical Cultivars At in 1987, the diameters of stems from Red Bobs that produced males were signiþcantly thinner than those that produced females (P 0.05), but neither of these differed from the intermediate-sized stems that failed to produce adults (P 0.05; Fig. 3a). Males had signiþcantly lighter weights than females (P 0.05; Fig. 3b) and emerged before females so that by 8 May 50% of all males had emerged, whereas 50% of the females emerged after 10 May (data not shown). Wing and body lengths were highly correlated with weight (Fig. 4; R 0.84), and all of these body size measurements were signiþcantly correlated with egg load (R 0.76). Stem diameter was weakly but signiþcantly correlated with sawßy egg load (Fig. 4; R 0.38; P 0.05). The sex ratio in Red Bobs favored females at 54%.

7 December 2005 CÁRCAMO ET AL.: HOST EFFECTS ON C. cinctus 1585 Table 4. Least square means of emergence date (Julian day) for wheat stem sawfly reared under laboratory conditions at room temp (22 C) collected from two environments: (2002 and 2003) and (2001 and 2002) Cultivar Hollow durums AC Avonlea NP NP 194 a 170 a AC Navigator NP NP 196 a 176 bc Kyle NP NP 194 a 184 d Hollow bread wheats AC Barrie 169 a 185 a 193 a 179 bc AC Cadillac 168 a 182 a 193 a 175 b AC Intrepid NP 185 a 194 a 178 bc CDC Teal 169 a 185 a 193 a 177 bc Katepwa 170 a 185 a 194 a 179 bc McKenzie 174 a 186 a 193 a 177 bc Solid bread wheats ACAbbey NP NE 194 a 179 bc AC Eatonia NP 186 a NE 182 dc Males 168 A 184 A 193 A 178 A Females 172 B 186 B 195 B 178 A SED 1 1.4Ð Ð Ð Ð3.2 SED Statistical results: Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Cultivar (C) ** Sex (S) * * ** C*S ** *, ** F-test signiþcant at P 0.05 and 0.01, respectively. 1 Range of SE of the difference between mean of each cultivar when they varied. 2 SE of the difference between means of sexes presented. Abbreviations: NP, not planted; NE, no emergence of adults. In the other historical study, adults that emerged in the spring of 1991 were clearly affected by the solidstemmed trait of the cultivars (Table 7). The solidstemmed cultivars Leader and Lancer (solid ratings 4.0 and 4.6, respectively) produced signiþcantly smaller adults (Kruskall-Wallis and Duncan tests, P 0.05) in terms of wing length and weight and females with fewer eggs compared with the four hollowstemmed cultivars (solid ratings 2.3). Fielder produced the largest adults and most fecund females with egg loads around 52 per female and signiþcantly higher than both the solid and the other three hollow-stemmed cultivars (P 0.05, Kruskall-Wallis test; Table 7). This study was repeated in 1991Ð1992, but only sex ratio data were found in the archives. Sex ratios appeared to differ among cultivars (Table 7). The solid-stemmed Leader had a clear male-biased sex ratio with 38% females, whereas the hollow-stemmed cultivars, Biggar and Fielder, favored females at 69 and 63% of the population, respectively. Unexpectedly, the other solid-stemmed cultivar Lancer had an unbiased ratio at 52% females, which was similar to the other hollowstemmed cultivars such as Laura and Katepwa that had 42 and 53% females, respectively. Discussion Fitness The main objective for both the current and historical studies was to assess cultivar effects, particularly those with the solid-stemmed trait, on a number of sawßy population parameters, including date of emergence, proportion of adults that emerged from the stubs, adult size (wing length and weight), and female fecundity. These traits are expected to reßect sawßy Þtness, and their responses to cultivars will improve our understanding of resistance mechanisms, sawßy population dynamics, and future pest status of this insect in western Canada. In our analyses of the more recent data set, we used stem diameter as a covariate because this variable is known to inßuence Þtness of this insect (Morrill et al. 2000). In general, although stem diameter often had a signiþcant effect on the response variables, the effects of cultivar remained signiþcant. Therefore, we are conþdent that our statements on cultivar differences were not masked or confounded by variation in stem diameters. The proportion of adults that emerged (survivorship) from the stubs differed among cultivars but was difþcult to compare statistically because of heterogeneous sample sizes. Agronomic studies to be published elsewhere (Beres, Cárcamo, and Byers, unpublished data) have determined that the solid-stemmed cultivars included in this study can substantially reduce damage (and yield losses) in terms of the number of stems cut by the sawßy. Hence, there were few stubs per sample for the solid-stemmed cultivars to assess emergence and other Þtness parameters. Therefore, our results for survivorship and other parameters must be considered conservative estimates of the impacts of

8 1586 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 Table 5. Least square mean weights of adult wheat stem sawfly from two environments: (2002 and 2003) and (2001 and 2002) Cultivar Sex Raw Raw Adjusted Adjusted Hollow durums AC Avonlea 5.11 ab a AC Navigator 5.53 abcd 8.56 a Kyle 6.30 bcd 7.33 a Hollow-stemmed bread wheats: AC Barrie 4.32 ab 4.16 a 6.62 cd 8.18 a AC Cadillac 6.99 c 3.95 a 6.82 d 9.56 a AC Intrepid NP 3.00 a 5.88 abcd 8.85 a CDC Teal 5.62 ab 4.77 a 6.58 cd 8.59 a Katepwa 4.19 c 3.19 a 5.60 abc 8.57 a McKenzie 3.87 c 3.10 a 5.16 ab 7.26 a Solid bread wheats AC Abbey NP NE 4.49 a 6.88 a AC Eatonia NP 2.87 a 7.32 a Females 6.59 A 4.79 A 7.71 A A Males 3.40 B 2.37 B 3.86 B 5.61 B SED Ð Ð Ð Ð1.78 SED (raw values) (on stem adjusted) Statistical results: Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Cultivar (C) ** * Sex (S) ** * ** ** C*S Stem Diameter ** * Means not sharing letters within same case type are signiþcantly different at P *, ** F-test signiþcant at P 0.05 and 0.01, respectively. 1 Data and statistics presented for reßect values adjusted for stem diam, whereas raw values are presented for given the lack of signiþcant effects of stem diam on weights at this site. 2 Slopes for the covariate at in 2001 and 2002 were not equal. 3 Range of SE of the difference between mean of cultivar when they varied. 4 SE of the difference between means of sexes presented. Abbreviations: NP, not planted; NE, no emergence of adults. solid-stemmed cultivars on sawßy. Despite these constraints, we observed consistent trends of lower survivorship to the adult stage in the solid-stemmed cultivars similar to those reported for earlier registered cultivars such as Rescue (Holmes and Peterson 1960, 1961, 1963). We also noted that some hollow-stemmed cultivars, such as AC Navigator, McKenzie, and AC Intrepid, were associated with lower survivorship of the sawßy depending on the site and the year; some of these cultivars deserve more detailed investigation. The average diameter of the stems that failed to produce live adults were thinner than those that produced females but often thicker than those stems that produced males. An experimental study is needed to determine if female sawßy larvae have reduced survivorship in stems with intermediate diameter. Male sawßy emerged 2Ð4 d earlier than females, which was consistent with our Þeld observations. There were no consistent differences in emergence dates among cultivars, although we observed some nonsigniþcant trends toward earlier emergence in some hollowstemmed cultivars relative to the solid cultivars. Sawßy fecundity was clearly affected by wheat cultivar in both studies. Although stem diameter was important, cultivar remained a signiþcant factor after covariate analyses. The solid-stemmed cultivars had the most pronounced effect because in some years no females emerged from these cultivars. The few females that emerged had consistently lower egg loads compared with hollow-stemmed cultivars. These results were expected, because the solid-stemmed cultivars included here, like all their genetic predecessors, were derived from the same germplasm line (S615), which has been shown to reduce sawßy Þtness (Platt and Farstad 1946, Holmes 1982, DePauw et al. 1994). In general, a cultivar such as AC Eatonia will suffer less than one-half of the lodging damage relative to susceptible cultivars (DePauw et al. 1994). Holmes (1979) reported that the solid-stemmed cultivar Rescue produced females with an average of 27 eggs, and those from the hollow-stemmed cultivar Thatcher produced 37 eggs. Our results from both studies were within this range; for example, at in 2002, adults from larvae that developed in 2001 had an average of 25 eggs from the solid-stemmed AC Abbey and an average of 41 from the hollow-stemmed AC Cadillac. Also, as reported by Holmes (1979, 1982), during drought years such as those experienced by the

9 December 2005 CÁRCAMO ET AL.: HOST EFFECTS ON C. cinctus 1587 Fig. 2. Relationship between female weights and egg loads at 2001, 2003, and with sites pooled in larvae in 2001 at, Þtness can decrease considerably as shown by the low egg production observed in all the cultivars this year at this site. In contrast, larvae at in 2002 experienced a low stress year because of ample moisture, and the females that emerged in 2003 had more eggs. The impact of drought seems to be buffered in the susceptible cultivars because the number of eggs did not ßuctuate as widely as in the solid-stemmed cultivars; they ranged from around 30Ð40 for the most susceptible cultivars such as CDC Teal and AC Cadillac. Past studies have shown that hollow-stemmed cultivars exhibit differences in susceptibility to the wheat stem sawßy (Platt and Farstad 1946, McGinnis 1950). Some of the cultivars in our study sometimes had a detrimental effect on sawßy Þtness. Females from the hollow-stemmed cultivars AC Navigator, AC Avonlea, McKenzie, Katepwa, AC Intrepid, and AC Barrie had similar egg loads, but sometimes females from these cultivars were signiþcantly less fecund than those from other hollow-stemmed cultivars in at least 1 yr at one site. McKenzie was derived from a cross with Amidon (Graf et al. 2003), a cultivar registered in the United States that is reported to have partial resistance to the sawßy (Kappel 1997). Among the durum cultivars, AC Navigator and AC Avonlea have been observed to have greater resistance in commercial Þelds than the durum Kyle and some of the hollow-stemmed spring wheats (personal observations). Furthermore, our data showed that AC Navigator often had a negative impact on sawßy survivorship and other population parameters and deserves further study to clarify the nature of the resistance. The genotype by environment interaction in the expression of the solid pith trait and the subsequent resistance to the sawßy has been reported for earlier solid-stemmed cultivars as well (Platt 1941). Similarly, the recently registered cultivar AC Abbey and to a lesser extent AC Eatonia will sometimes have reduced resistance to the sawßy; reduced exposure to sunlight during the critical stem elongation stage (Platt 1941) or to other severe environmental conditions is thought to cause this variation. Nevertheless, despite this variability, a solid-stemmed cultivar like AC Eatonia will suffer about one-half the damage compared with the hollow-stemmed cultivars (DePauw et al. 1994). Future sawßy populations can be reduced by planting these solid-stemmed cultivars and by diversifying the crop rotations to include nonhost broad leaf crops. Sex Ratio We explored two aspects of host quality that can inßuence wheat stem sawßy sex ratio. First, stem diameters varied within cultivars because the main stem within a clump is usually thicker than the subsequent tillers. Second, cultivars varied with respect to stub

10 1588 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 Table 6. Least square means of egg loads of wheat stem sawfly from two environments: (2002 and 2003) and (2001 and 2002) Hollow durums Adjusted Raw Adjusted Adjusted 1 AC Avonlea NP NP 30 ab 35 ab AC Navigator NP NP 30 ab 40 abcd Kyle NP NP 37 cde 38 abcd AC Barrie 28 b 21 a 39 cde 33 a AC Cadillac 43 c 22 a 41 de 44 d A Intrepid NP 23 a 35 abcd 38 abcd CDC Teal 32 b 29 a 43 e 41 bcd Katepwa 26 ab 21 a 33 abc 42 cd McKenzie 19 a 17 a 28 ab 36 ab AC Abbey NP NE 25 a 32 a AC Eatonia NP 17 a NE 31 a SED 2 2.6Ð Ð Ð Ð5.9 Statistical results: Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Num DF Den DF F Test Cultivar (C) ** *** Stem diam *** *** *** Adjusted values (standardized for a commonstem diam) are presented when stem diam had a signiþcant covariate effect on the response variable Aegg load@. *, ** F-test signiþcant at P 0.01 and 0.001, respectively. 1 Slopes for the covariate at in 2003 were equal to zero. 2 Range of SE of the difference between mean of each cultivar. Abbreviations: NP, not planted; NE, no emergence of adults. diameters (durums thicker than bread wheats) and pith expression (hollow- and solid-stemmed). Variability of stub diameters within cultivars affected sex ratio. In general, females emerged from thicker stems compared with males within a given cultivar in our study of Red Bobs in 1987 and other cultivars in 2001Ð2003. Similar effects of stem diameter on sex ratios were reported by Wall (1952) for each of three hollow-stemmed cultivars and three solidstemmed lines and by Morrill et al. (2000) for the hollow-stemmed, winter wheat cultivar Redwin. However, McGinnis (1950) found no differences in sex ratio between main stems and tillers in the two hollow-stemmed varieties, Thatcher and Red Bobs. He suggested that lower survivorship of males in Red Bobs, which had a lower lodging level, explained the higher incidence of females in this variety compared with Thatcher that had a male-biased ratio. Holmes and Peterson (1963) found no covariate effect of stem diameter on sex ratio despite differences in stem diameters within the cultivars they selected and with their seeding date treatments. It is not clear if the male stem diameters were compared directly against the female stem diameters in the latter study. The differences in sex ratios between thick and thin stems within cultivars are mediated by a combination of insectðhost interactions. First, like other Hymenoptera, wheat stem sawßies are haplo-diploid with only fertilized eggs turning into females with the full complement of 18 chromosomes (Mackay 1955). Uniparental, parthenogenetically reproduced, totally female populations are the exception and have only been reported once (Farstad 1938). To maximize host use and subsequent Þtness, females are thought to lay fertilized eggs in stems with larger diameter (McGinnis 1950, Wall 1952), which produce bigger, more fecund females as found in our study and elsewhere (Morrill et al. 2000) and with other hymenopterans that lay female eggs in bigger hosts (Charnov et al. 1981, King 1988). Second, because females are larger than males they require more resources and can be expected to be more sensitive to host quality (Charnov 1982), and in the case of the sawßy, they may die more often than males in thin stems. Our study and that of Morrill and Weaver (2000) support this hypothesis in the sense that only females had reduced Þtness (weights) in hosts of lower quality. Also, stems that failed to produce adults were signiþcantly thinner than those stems that produced females but not from those that produced males. One may speculate that stems of intermediate size receive female eggs but lack sufþcient resources to produce an adult. A direct test of these hypotheses is possible but cumbersome given the difþculty in sexing larvae (Holmes 1970), especially dead ones. A third mechanism that drives the sex ratio in favor of males in secondary and tertiary thin tillers is the fact that thin stems will develop last and will often be available to females that may have spent their spermatozoa or to females that emerged too late. These females may be virgins because males emerge Þrst and disappear before the end of the female ßight period (Jacobson and Farstad 1952, Holmes and Peterson 1963). Unmated females thus will lay only male eggs (McGinnis 1950) that will survive in the vacant thin stems or will be cannibalized by the female larvae laid

11 December 2005 CÁRCAMO ET AL.: HOST EFFECTS ON C. cinctus 1589 Fig. 4. Correlations of adult population parameters for sawßy reared from Red Bobs in 1987 from plant material collected from the nursery. LOGSTEM, logarithm of stem diameter; WTMG, average weights in milligrams; BODYL, average body lengths; WINGL, average wing length; EGGS, average number of eggs per female. Fig. 3. (A) Diameters of stems with dead larvae, male, or female sawßy in the Red Bobs wheat cultivar in (B) Weight of males and females. by the earlier cohort of mated females in the more advanced thicker stems (Holmes and Peterson 1963). The combination of all of these factors, some intrinsic to the insect and others resulting from the interaction with host development, explain the strong sex bias toward females in thick stems and males in thinner stems within a cultivar observed in most studies. In the current cultivars study we found no cultivar effects on sex ratio of the sawßy despite differences in diameters between hollow-stemmed durums and hollow-stemmed bread wheats. This result may be explained by the distribution of relative stem diameters that should be similar among cultivars, given the tillering patterns of wheat grown under similar conditions. This assumes that a female lays an egg of a given sex following a relative sex allocation decision rule for each clump of stems: the primary stem gets a female, the thinnest a male, and the intermediate(s) usually a female. Because females tend to be more responsive to host quality as shown by our Þtness data and elsewhere within the Hymenoptera (Charnov 1982), the latter will experience lower survivorship as a result of the decreased host stem diameter. This scenario would result in an unbiased or slightly female biased ratio regardless of variety and would explain the results for our more recent study. Alternatively, if females followed an absolute sex allocation decision rule, where stems below a critical diameter all get male eggs, those varieties that have a larger proportion of stems below a threshold value would produce a very male-biased population. Furthermore, the reduced host quality (solid pith) of even the thickest stems may increase female mortality and further skew the population toward males. Our results from both studies and those of Wall (1952) do not support this idea. However, a more rigorous test would be to determine sex ratios in large monoculture Þelds where females do not have the choice to select adjacent plots or rows with the varieties that have thicker stems. An experimental approach under controlled conditions could help elucidate this question as well. Differences in sex ratios between solid- and hollowstemmed varieties have been reported in the literature (Farstad et al. 1949, Holmes and Peterson 1963). In our extensive recent study that included six hollowstemmed bread wheats, three hollow-stemmed durum, and two solid-stemmed bread wheat cultivars, we found no such differences. Solid-stemmed cultivars such as AC Eatonia and AC Abbey, however, are effective at reducing the number of cut stems (De- Pauw et al. 1994), and the number of adults available

12 1590 ENVIRONMENTAL ENTOMOLOGY Vol. 34, no. 6 Table 7. Population parameters of male and female sawflies from the historical cultivar stubs collected from on 22 May 1991 (30 or 31 specimens measured for each sex, except for Biggar [15 males]) and sex ratio of adults from stubs collected on 13 April 1992 Cultivar Egg loads Mean weights of adults (mg) 1991 Collection 1992 Collection Mean length of left forewing ( 10Ð1 reading of 6 by 15 optic) Mean SE Males SE Females SE Males SE Females SE Sex ratio No. adults Fielder 51.87a a a a a Laura 44.55ab a ab a ab Katepwa 39.87b a b a ab Biggar 39.23b a b a bc Leader a 30.71bc b c b cd Lancer a 25.00c b c b d Entries are means of pooled individuals with 1 SE. Means within columns not sharing letters are signiþcantly different (P 0.05, Kruskall-Wallis nonparametric test). a Solid-stemmed bread wheat. Percent females for these two varieties may have been too small to determine sex ratios. In the historical 1987 study, the hollow-stemmed cultivars Fielder and Biggar had a female biased sex ratio compared with the solidstemmed cultivar Leader. This difference was not consistent among the other varieties in these two classes. Wall (1953) reported a similar study in which stem diameter affected sex ratio allocation within cultivars, but cultivar differences were inconsistent. In the study of Wall (1953), one of three hollow-stemmed cultivars, Red Bobs, had a female-biased sex ratio and H4191, one of three solid-stemmed cultivars had a male-biased sex ratio, and there seemed to be no general pattern with respect to the sex ratio of the cultivar and its mean stem diameter (see Tables 3 and 4 in Wall 1952). Holmes and Peterson (1963) reported a higher proportion of females in hollow-stemmed than in solid-stemmed bread wheat cultivars and a decreasing female incidence with delayed seeding date. The latter pattern was also observed by Jacobson and Farstad (1952) and was explained by late emerging females that were unmated or had spent their sperm reserves and were forced to lay male eggs. According to Holmes and Peterson (1963), the reduction of females in solid-stemmed cultivars resulted from high mortality of the Þrst cohort of eggs (with a female bias) laid in the lower internodes (Holmes and Peterson 1961) during the beginning of the ßight period and higher survivorship of later egg cohorts (male bias) laid in higher internodes that tend to have less pith expression (DePauw et al. 1994, 1999). In hollowstemmed cultivars, the early female dominated egg cohort hatches successfully, and the female larva cannibalizes those that hatch later from male eggs laid by unmated females (Holmes and Peterson 1963). Our results on sawßy Þtness support these arguments, given that females are more likely to respond to cultivar quality than males. Therefore, because female eggs tend to be laid in the main stem, which is the one with best solid pith expression, more females are expected to die and more males to survive in the thinner secondary tillers, thereby skewing the sex ratio toward males. Our recent study, however, did not support this hypothesis, although it was weakly supported by our study from 1987 and the study of Wall (1953). Some of the inconsistencies may be related to the variability of the solid pith expression in most cultivars, which is affected by climatic factors (Platt 1941, Holmes 1984) and small sample sizes from the solid cultivars. The cultivar Golden Ball illustrates the difþculty of predicting the impact of a solid-stemmed cultivar on sex ratio. This cultivar is a thick, solid-stemmed, slow growing durum with very stable pith expression and low levels of lodging at harvest (Platt and Farstad 1946), yet it consistently produces a female-dominated sex ratio among the few adults that survive (Farstad et al. 1949, Holmes and Peterson 1963, Cárcamo, Beres, Clarke, DePauw, Byers, unpublished data). Holmes and Peterson (1960, 1963) explained the prevalence of females in Golden Ball as a result of the occurrence of hollow internodes that promoted the survivorship of earlier laid female eggs. However, other researchers including two of our co-authors (F.C. and B.B.) have consistently found all internodes to be solid in recent plantings of Golden Ball, which would not explain the female-biased sex ratio. The tall stature and its thick stems may attract the earlier mated females to lay female eggs on this cultivar under garden type experiments with multiple cultivars and may explain the incidence of more females. In conclusion, our plot studies showed that the reduction in host quality found in solid-stemmed wheat cultivars will cause surviving larvae to develop into smaller females with lower fecundity relative to those that develop in hollow-stemmed cultivars. Selecting a solid-stemmed cultivar is recommended to reduce lodging damage at the end of the growing season and to reduce future populations below economically damaging levels. Additional pest management tactics such as biological and cultural control need to be developed and integrated with host plant resistance.