False Ring Formation in Eastern Hemlock (Tsuga canadensis) After Hemlock Woolly Adelgid (Adelges tsugae) and Elongate Hemlock Scale (Fiorinia externa) Feeding

Hemlock forests in the eastern United States are threatened by two sessile invasive herbivores: the elongate hemlock scale, Fiorinia externa Ferris (Hemiptera: Diaspididae; ‘EHS’) and the hemlock woolly adelgid Adelges tsugae Annand (Hemiptera: Adelgidae; ‘HWA’). EHS and HWA occupy similar feeding guilds but have enormously different effects on tree health. EHS reduces hemlock growth and causes needle discoloration and loss, but only causes tree mortality under high EHS densities (McClure 1980b). In contrast, HWA has devastated stands of hemlocks on the east coast of the United States. Although EHS reduces fitness of the tree and can kill already stressed trees (McClure 1980), HWA is known to kill hemlocks in as few as four years (McClure 1991). The mechanism by which HWA and EHS kill trees is not yet elucidated and little is known as to the physiological effects each invasive has on hemlock. For the first part of my master’s research, I focused on differences in abnormal wood production among uninfested trees, EHS-infested trees and HWAinfested trees at the branch level. Specifically, I measured false ring density, ring growth and earlywood:latewood ratios in the two most recently deposited growth rings. Branches from HWA-infested trees had 30% more false ring than branches from EHS-infested trees and 50% more than branches from uninfested trees. In contrast, growth and earlywood:latewood ratios did not differ among treatments. This result suggests that two invasive insects from similar feeding guilds have differing effects on false ring formation in eastern hemlock. These false rings may be the product of a systemic plant hypersensitive response to feeding by HWA on hemlock braches. If false rings are responsible for or symptomatic of hemlock water stress, this may provide a potential explanation for the relatively large effect of HWA infestations on tree health. For the second part of my master’s thesis I looked at the impact of HWA on eastern hemlock anatomy and physiology. Specifically, I looked at growth and production of new buds on terminal and side branches in hemlock infested with and without HWA. We found that trees infested with HWA have significantly less new growth and fewer new buds. Additionally, I measured water potential, photosynthesis and stomatal conductance in trees infested with and without HWA during diapause and immediately after HWA resumes feeding. HWA undergoes summer diapause while still attached to eastern hemlock and it is unknown if this ‘inactive’ period affects tree health. We found that actively feeding HWA exacerbate reductions in photosynthesis and stomatal conductance, but not water potential. The presence of HWA, irrespective of feeding activity, decreases eastern hemlock water potential, photosynthesis, and stomatal conductance. Additionally, water potential and stomatal conductance were negatively correlated with HWA density. These data indicate that HWA negatively impacts tree health even when not actively feeding and depleting carbon reserves. These results also suggest that HWA-infested trees are water stressed, shedding light on possible mechanisms behind HWA-induced death.


LIST OF TABLES
. Adelges tsugae has a greater effect on tree health than F. externa, but the mechanism underlying their differential effect is unknown. We explored their effects by assessing growth ring formation in branches of trees that had been experimentally infested for four years with A. tsugae, F. externa, or neither insect. We measured false ring density, ring growth, and earlywood:latewood ratios in the two most recently deposited growth rings. Branches from A. tsugae-infested trees had 30% more false rings than branches from F. externa-infested trees and 50% more than branches from uninfested trees. Branches from F. externa-infested trees and control trees did not differ in false ring formation. Radial growth and earlywood: latewood ratios did not differ among treatments. Our results show that two invasive herbivores with piercingsucking mouth parts have differing effects on false ring formation in eastern hemlock.
These false rings may be the product of a systemic plant hypersensitive response to feeding by A. tsugae on hemlock stems. If false rings are responsible for or symptomatic of hemlock water stress, this may provide a potential explanation for the relatively large effect of A. tsugae infestations on tree health.

INTRODUCTION
Herbivores can alter plant physiology directly through tissue and nutrient removal and indirectly through the induction of increased chemical (Bezemer et al. 2003, Kaplan et al. 2008) and/or morphological defenses (Levin 1973). Although such responses vary between herbivores, alterations in plant physiology are especially likely with invasive or other species that reach high densities on their host plants (Sakai et al. 2001). Although herbivore-induced changes to plant structure are most commonly thought to involve architectural shifts resulting from bud/branch mortality or altered height/radial increments (Traw andDawson 2002, Sopow et al. 2003), herbivory may also induce changes in woody plant tissues in the stems of conifers and other woody plants ).
In conifers, false rings are thick-walled xylem cells that appear as dark bands of latewood flanked on both sides with earlywood (Copenheaver et al. 2006). False rings occur within an annual ring but, although they resemble the end of an annual ring, do not occur on a yearly or seasonal basis. Normal rings are composed of large, thin-walled cells formed early in the growing season and small, thick-walled cells formed later in the year. These true rings are characterized by an abrupt increase in cell size at the start of the new growing season, while false rings are identified by a slow increase in cell diameter and decrease in cell wall thickness adjacent to the false ring (Copenheaver et al. 2006). Like compression wood, false rings have thick-walled xylem cells that increase resistance to water flow (Bolton and Petty 1978). False rings are associated with water stress and insect infestation, and have been observed in conifers such as Pinus sylvestris and P. banksiana (Hollingsworth and Hain 1992, Cherubini et al. 2003, Copenheaver et al. 2006. Drought may induce false rings by reducing photosynthesis and stopping cambial activity during the summer (Cherubini et al. 2003). During periods of water stress, small, thick-walled cells are formed in the wood; if conditions become more favorable, subsequent cells will be larger with thinner walls (Wimmer et al. 2000). This alternation in cell size may appear as a false ring. In support of this, Wimmer et al. (2000) found that false rings were associated with periods of alternating wet and dry months.
Although not all herbivores induce changes in ring formation (Priya andBhat 1997, Heijari et al. 2010), certain insects have been linked to their occurrence.
Increased densities of the balsam woolly adelgid (Adelges piceae Ratz.) are correlated with the formation of rotholz rings, a type of abnormal wood similar to compression wood, in Fraser fir (Abies fraseri) (Hollingsworth and Hain 1992), and rotholz rings are only found near areas of adelgid feeding. Since these rings contain cells that do not conduct water and balsam woolly adelgid feeding is also associated with an increase in non-conducting heartwood (Arthur and Hain 1986), the resulting water stress may eventually kill the tree (Hollingsworth and Hain 1991). The formation of rotholz rings may defend against low-density adelgid infestations by forming necrotic tissue around the feeding site that isolates and starves the insects (Arthur and Hain 1985). With many points of adelgid feeding, however, so much of the stem may become nonconductive that the increased water stress actually kills the tree (Arthur and Hain 1985, McClure 1988. The hemlock woolly adelgid (Adelges tsugae Annand (Hemiptera: Adelgidae)) is an invasive hemipteran herbivore that is causing high mortality of eastern hemlock (Tsuga canadensis) across the eastern United States (Orwig et al. 2002). Adelges tsugae was first reported in Virginia in the 1950's (Souto et al. 1996) and has since spread rapidly along the east coast, now ranging from northern Georgia to Maine Cheah 1999, USFS 2008). Adelges tsugae completes two generations per year in its invaded range (McClure 1989), and feeds on eastern hemlock at the base of the needle petiole by inserting its stylet bundle into xylem ray parenchyma tissue (Young et al. 1995 Although little is known about why these species differ in their impact, there is evidence that A. tsugae induces an especially pronounced hypersensitive response in the tree (Radville et al. 2011). The hypersensitive response is a plant defense response that increases the levels of reactive oxygen species such as superoxide anions, hydroxyl radicals, and hydrogen peroxide (H 2 O 2 ), thereby inducing cell death in herbivore-colonized areas in order to isolate and starve feeding organisms (Heath 2000, Liu 2010. The cue for this response is often the presence of a foreign substance indicative of herbivore feeding (reviewed in (Cornelissen et al. 2002), and the ensuing localized tissue death is a particularly effective response to sessile herbivores (Karban and Baldwin 1997). This response has been shown to reduce plant damage caused by balsam woolly adelgids, bark beetles, and a host of other herbivore species (Fernandes 1990, Ollerstam andLarsson 2003). In the case of A. tsugae, Radville et al. (2011) found that infestation caused a larger localized hypersensitive response (measured as hypothesize that A. tsugae infestation will induce a greater degree of false ring formation than will infestation with F. externa or neither insect (control). Through this research, we hope to provide insight into why feeding by A. tsugae is much more damaging than feeding by F. externa. Because every section was asymmetrical, we measured branch radius in three different axes and averaged them to calculate a mean branch radius. We followed the same procedure to calculate the mean thickness of the 2009 and 2010 growth rings and the mean thickness of latewood in the 2009 and 2010 tree rings of each branch. We calculated the width of each ring's earlywood by subtracting the latewood thickness from the mean ring width, and the earlywood:latewood ratio by dividing the thickness of each ring's earlywood by its latewood.

Statistical analysis:
The unit of replication for our analyses was the mean response per tree per treatment (21 replicates). Data were square-root transformed when necessary to improve normality; variances were homogenous between treatments. When analyzing data on branch diameter and total annual growth rings, we used ANOVA to test for the main effects of treatment (A. tsugae-only, F. externaonly, and control) and location within the experimental grid (included as a blocking variable), and for their two-way interaction. All other data were analyzed using repeated measures ANOVA to test for the main effects of treatment, location, time (either the 2009 or 2010 growth ring), and their interactions. We performed means separation tests, where appropriate, using Tukey's HSD. Statistical analyses were performed using JMP 9.0.0 (SAS 2010). When initial p-values are significant, we report both the initial p-value as well as the p-value corrected for multiple comparisons at α=0.05 using step-up FDR, a sequential Bonferroni-type procedure (Benjamini and Hochberg 1995).

RESULTS
Branch size and age: There were no treatment-level differences in either branch radius or age, measured as the number of annual growth rings (Tables 1, 2A-B). Branches averaged 0.61+0.036 [SE] cm in diameter and had similar numbers of annual growth rings. There was no effect of tree location within the experimental grid (Table 2A-B), and no significant treatment*location interactions.
Adelges tsugae-infested trees averaged 0.96 false rings/growth ring, significantly more than in either control or F. externa-infested trees (0.48 and 0.66, respectively; Tukey's HSD, P < 0.05). In contrast, F. externa-infested trees did not differ from the uninfested controls (Fig. 3). There was a marginally significant effect of tree location within the experimental grid (Table 3A), but no significant change in false ring density across time (Table 3A). All two-and three way interactions were nonsignificant (Table 3A).
Earlywood, latewood, and ring width: There were no treatment-level differences in the width of earlywood (Tables 1, 3B), latewood (Tables 1, 3C), or the annual rings (Tables 1, 3D). There was also no effect of treatment on the earlywood:latewood ratio (Tables 1, 3E). These four variables did not change over time and were unaffected by tree location within the experimental grid (Tables 1, 3B-E). There were no significant two-or three-way interactions (Tables 1, 3B-E).

DISCUSSION
While both insects have piercing-sucking mouth parts, infestation by F. externa and A. tsugae had markedly different effects on wood formation. Branches from A. tsugae-infested trees had a greater number of false rings than branches from uninfested trees (Fig. 3). Branches infested with A. tsugae had 50% more false rings than control branches and 30% more false rings than F. externa-infested trees. In contrast, infestation by F. externa did not significantly increase false ring formation.
Despite the difference in false ring formation, there were no between-treatment differences in annual ring width or earlywood and latewood production. Perhaps the most likely explanation for our findings is that the increased number of false rings in A. tsugae-infested branches is a consequence of plant hypersensitivity, a defense mechanism against sessile herbivores and pathogens ). The hypersensitive response induces cell death by increasing the reactive oxygen species (Heath 2000), which isolates the herbivore or pathogen and prevents it from establishing a suitable nutritional site (Wong and Berryman 1977, , Bonello et al. 2006 A. tsugae densities were consistently lower than F. externa densities for the duration of the study (Figure 1), a fact that suggests even low A. tsugae densities induce a greater degree of false ring formation than higher F. externa densities. Although unproven, it has also been suggested that a component of A. tsugae saliva is 'toxic' (Young et al. 1995) and that it injects chemicals during feeding that adversely affect  (Bolton and Petty 1978). Compression wood conducts water less efficiently than does normal wood (Spicer and Gartner 1998), and there is also evidence that insect-induced false rings impede water transport. Mitchell (1967) found that trees infested by the balsam woolly adelgid absorbed and transported less dye (a proxy for water) than uninfested subalpine and grand fir trees. Rotholz rings appeared to inhibit dye transport and infested trees had half as many conducting tree rings (Mitchell 1967). Puritch (1971) showed that balsam woolly adelgid interfered with the water conduction in grand fir, evident in the reduced permeability of sapwood in balsam woolly adelgid-infested trees. Since balsam woolly adelgid and its hosts are closely related to A. tsugae and eastern hemlock, it seems reasonable to assume that false rings formed in A. tsugaeinfested eastern hemlock will correlate with changes in water transport efficiency (an idea first suggested by Walker-Lane 2009). If the false rings produced in A. tsugaeinfested trees are indicative of water stress, this may explain why A. tsugae has such a severe impact on tree mortality.     Cross-section of T. canadensis branch with multiple false rings (indicated by arrows).

TABLES
All photographs were taken using a PixeLink PL-A662 camera attached to an Olympus SZX12 microscope at 400x magnification. Am. 88: 827-835.

INTRODUCTION
Many sap-feeding insects have long lasting physiological impacts on their host plant. These physiological changes are driven by both changes in plant nutrients (McClure 1980a, Masters andBrown 1992) and the production of secondary chemicals (Karban andMyers 1989, Haukioja et al. 1990). By removing nutrients from the plants' xylem or phloem, sap-feeding insect herbivores reduce plant growth, decrease photosynthesis rates and decrease plant reproduction (Candolfi et al. 1993, Meyer 1993. Sap-feeding insects are detrimental to trees (Vranjic andGullan 1990, Smith andSchowalter 2001), yet there is minimal literature that quantifies the impact of sap-feeding herbivores on woody species (reviewed in Zvereva et al. 2010).
Conifers may be especially susceptible to sap-feeders because unlike deciduous trees that store resources in their roots, stems, and other tissues inaccessible to sap feeders, evergreens allocate more storage to foliage (Chapin et al. 1990, Krause and Raffa 1996, Hester et al. 2004). The lack of such stored resources makes conifers vulnerable to herbivore attacks and in some cases, intense sap-feeding events can even result in tree death , Furuta and Aloo 1994, Paine 2000.
The invasive non-native hemlock woolly adelgid (Adelges tsugae Annand Hemiptera: Adelgidae) is a specialist sap-feeding insect currently decimating eastern hemlock (Tsuga canadensis (L.) Carrière), a foundation species in eastern North American forests. Feeding occurs through the insertion of the stylet bundle at the base of a needle into the ray parenchyma tissue (Young et al. 1995). Once HWA selects a feeding place, it remains sessile throughout its entire life cycle. HWA can kill mature hemlocks within four years of infestation (McClure 1991). As eastern hemlock stands disappear, they are replaced by deciduous hardwood species such as birch Foster 1998, Catovsky andBazzaz 2000). This disappearance has major impacts on ecosystem processes that can result in the regional homogenization of forest structure (Ellison et al. 2005) and declines of bird (Tingley et al. 2002) and invertebrate (Snyder et al. 2002) biodiversity (but see Ingwell et al. 2012).
HWA has a spring and a summer generation in its invaded range and each generation passes through four larval instars before becoming adults. The spring generation completes its life cycle between April -June and lays eggs that become the summer generation. The summer generation hatches in July and remains on hemlock until the following April when the cycle starts again (McClure 1989). While the spring generation feeds continuously throughout their shorter life cycle, the summer generation enters diapause in July immediately after hatching. Summer diapause in HWA is primarily induced by temperature (Salom 2001). The summer generation stays dormant until October, when HWA break diapause and resume feeding throughout the winter until April.
Diapausing insects undergo a period of arrested development characterized by metabolic depression (Triplehorn and Johnson 2005). Although diapause is often associated with winter, summer diapause also occurs in a wide range of insect taxa (reviewed in Masaki 1980). Insects in the Adelgidae experience diapause (Havill and Foottit 2007), but only two species (Adelges tsugae and Adelges piceae) are known to go through summer diapause (Amman 1962, McClure 1989. Summer diapause is induced by a range of environmental factors that include photoperiod and temperature. In addition to abiotic influences, biotic factors such as host plant quality also affect the induction and length of insect diapause (Dalin and Nylin 2012). For example, leaf toughness influences the likelihood of diapause in the swallowtail butterfly Byasa alcinous (Takagi and Miyashita 2008). Hunter and McNeal (1997) found that host plant species and the nutritional content of the diet influences the induction of, and mortality during, diapause in the lepidopteran herbivore Choristoneura rosaceana.
While many studies focus on the effect of the plant host on the insect in diapause, the question remains whether the presence of a diapausing insect impacts host-plant physiology and performance.
The hemlock-HWA interaction provides an ideal system to explore the impact of sap-feeding herbivores on trees during and after diapause on their host plant. This is because HWA experiences summer diapause with its stylet imbedded at the base of hemlock needles. The aim of this study was to determine how HWA impacts hemlock performance during diapause and active feeding. Specifically, we assessed hemlock growth during the tree growing season, when HWA from the spring are feeding. Once the summer generation emerged and entered diapause (which coincides with the end of the host plant's growing season) we measured physiological plant responses such as water potential, photosynthesis, and stomatal conductance. These responses were measured again immediately after HWA resumed feeding. Water Potential: On September 8 and October 27, 2012 we measured predawn shoot water potential on 12 randomly chosen trees per treatment in the HWAinfested and control treatments. We chose to take physiological measurement (water potential and gas exchange) in the autumn for the following reasons: (1) eastern hemlock photosynthesize year round (Hadley 2000); (2)  Gas Exchange: We measured gas exchange on a terminal branch (2012 growth) on each tree used to quantify water potential. Measurements were conducted between 9:00-11:00 am on September 9 and October 26, 2012. In branches from HWA-infested trees, we counted the number of HWA present/ per cm on the sampled foliage. After each measurement, foliage inside the leaf chamber was excised, placed on a white piece of paper, and photographed; we quantified total leaf area using imageJ 1.44 software (Abràmoff et al. 2004). To determine gas exchange rates we used a CIRAS-2 portable photosynthesis system (PP systems, Haverhill, MA, USA) with a 2.5 cm 2 leaf chamber and a CIRAS-2 LED light source of 1500 µmolm -2 s -1 , a CO 2 concentration of 390 ppm, air flow rate at 350 cm 3 s -1 and leaf temperature of 25° C.

Study
Statistical Analyses: All statistical analyses were performed using JMP 10.0 with each data point being the mean response variable per tree per sampling date. We used repeated-measures ANOVA with treatment and branch type (terminal or side) for branch growth and number of secondary buds. We used repeated-measures ANOVA to analyze the main effects of treatment and time, and the treatment*time interaction, on the following variables: water potential, net photosynthesis, stomatal conductance, and evaporation. We used linear regression to assess the correlation between HWA density and water potential, photosynthesis, and stomatal conductance for both time points. We checked all data for normality, sphericity and homogeneity of variance and log transformed water potential data in order to meet ANOVA assumptions. For analyses that did not meet the assumptions of sphericity, we report univariate Greenhouse-Geisser corrected p-value is reported. The critical P value used in this study was P < 0.05.

RESULTS
Growth: By the end of the growing season, terminal branches on control trees were 41% longer than HWA-infested terminal branches ( Figure 1A). In contrast, HWA infestation significantly affected side branch growth: side branches on control trees were 57% longer than on HWA-infested trees ( Figure 1B)

DISCUSSION
Diapause, a period of arrested growth and metabolic depression (Hahn and Denlinger 2011), allows insects to survive in otherwise-unsuitable environments (Andrewartha 1952). Insect diapause can occur at any stage in insect development and in locations that include soil, leaf litter, and on a host plant. Although insects that diapause on their host plant should have little or no effect on plant performance when in dormancy, we are unaware of any literature exploring the impact of 'inactive' insect presence on plant health. Our study investigates effects of an insects' active period (feeding) and inactive period (diapause) on plant performance. Our results showed that active feeding by HWA, an invasive sap-feeding herbivore, had a predictably detrimental impact on hemlock growth and physiology. The fact that HWA decreased water potential, photosynthesis and stomatal conductance further suggests that it induces symptoms of water stress in eastern hemlock. These symptoms are magnified when HWA is actively feeding versus in diapause (inactive period).
HWA had significant impacts on hemlock growth. The spring progrediens generation settles and begins to feed during the start of the hemlocks' growing season.
By the end of the growing season, terminal branches on control trees were 41% longer and had 56% more new buds than HWA-infested trees. The effect of HWA was even more pronounced on side branches. Side branches on uninfested trees grew 56% more and had 120% more new buds than HWA-infested trees. These results suggest that HWA-infested trees have significantly less lateral branching than uninfested trees.
This finding matters because lateral branching can be an effective herbivore deterrent: the increasingly complex structure and lateral spread of branches can make it difficult for herbivores to navigate (Vesey-FitzGerald 1973, Archibald andBond 2003).
Increased tree architecture can also promote tolerance to herbivory by increasing sectored subunits within a plant and augmenting resource capture (reviewed in Stowe et al. 2000). In addition to HWA, white-tail deer also feed on eastern hemlock (Eschtruth and Battles 2008 HWA-induced changes in hemlock anatomy and physiology likely accentuate the impact of HWA-induced water stress. Changes in water status alter gas exchange rates. For instance low water potential in plants is coupled with decreases in photosynthetic activity and stomatal conductance (Farquhar and Sharkey 1982, Epron and Dreyer 1993, Dang et al. 1997) and water stress in conifers such as Abies spp.
greatly reduces net photosynthesis (Puritch 1973). Congruently, increased photosynthetic rates in galled leaves is suggested to be explained by improved water relations (Fay et al. 1993). We showed a reduction in water potential and concomitant reductions in photosynthesis and stomatal conductance in HWA-infested trees. Sapfeeding insects generally tend to decrease photosynthesis (reviewed in Zvereva et al. 2010), and we found a similar result, HWA-infested trees had lower photosynthetic rates; a difference that was magnified when HWA were actively feeding.
The HWA-induced changes in hemlock physiology we observed may result from the formation of a higher number of false rings on HWA-infested branches (Gonda-King et al. 2012). False rings are bands of abnormal wood within an annual ring that consist of thick-walled xylem cells and which may hinder water transport efficiency (Mitchell 1967). While our data showed that actively-feeding HWA had a greater impact on hemlock physiology and growth, HWA's impact on physiology was visible even during diapause, as shown by a 10% and 41% reduction in photosynthetic rates and stomatal conductance, respectively, in HWA-infested trees. Once HWA emerged from diapause and began feeding, photosynthesis and stomatal conductance were 56% and 70% lower in infested trees. HWA's impact on photosynthesis and stomatal conductance during diapause might be explained by lasting prior changes in nutrient allocation from when HWA was actively feeding. Sap-feeding insects often act as resource sinks (Inbar et al. 1995, Kaplan et al. 2011) that can compete with natural plant sinks (i.e. actively growing tissues). HWA feeding alters local and systemic foliar nitrogen content (Stadler et al. 2005, Miller-Pierce 2010, Gómez et al. 2012b). Nitrogen sink competition between feeding herbivores and new foliage alters leaf nutrient status, resulting in decreased photosynthesis (Larson 1998). Since HWA feed on photosynthate from xylem ray parenchyma cells (Young et al. 1995) that transfer and store nutrients it is likely HWA induce greater sinks when feeding as opposed to in diapause. This is consistent with the idea that competition between plant and herbivore sinks reduces net photosynthesis. The act of HWA feeding does not magnify reductions in water potential. If changes in water potential are driven by false rings formation, this would have a permanent impact on water relations as changes to wood anatomy are not transient.
Our results illustrate that HWA presence, whether in diapause or actively feeding, negatively impacts hemlock health. The during-diapause impact of HWA may be caused by physical injury or chemical cues associated with the initial stylet insertion by HWA. Stylet insertion by other sucking insects can cause long-lasting anatomical and physiological changes (Ladd and Rawlins 1965, Ecale and Backus 1995, Shackel et al. 2005, and this may be the case for HWA as well. HWA secretes a salivary sheath when feeding at the base of hemlock needles that Young et al. (1995) suggests may be 'toxic' and responsible for the disproportional negative impact on hemlock growth caused by HWA feeding. In support of the 'toxic' saliva hypothesis, Radville et al. (2011) found that HWA elicits a systemic hypersensitive response in hemlock.
Despite finding that HWA continues to impact hemlock performance during diapause, we cannot determine if this is due to the presence of inactive HWA on the plant or if these effects are residual long-term impacts from prior HWA infestations.
Unfortunately, there is no true 'control' to compare impact of HWA diapause versus HWA feeding because the HWA life cycle always has a feeding period before diapause. A short-term feeding event always occurs prior to HWA entering diapause and short-term feeding events by herbivores can have lasting impacts on host plant health. For example, Cinara pseudotsugae, a sap sucking aphid, reduced root and shoot growth on its host tree and physiological impacts were still evident one year after herbivory (Smith and Schowalter 2001). It is possible that prior HWA feeding events have had long-lasting impacts on hemlock physiology and this is reflected during the diapause period. Regardless of the mechanism driving decreased hemlock performance when HWA are in diapause, these data show that HWA has a lasting impact on hemlock performance.
While the detrimental impact of HWA on hemlock has long been recognized, the effect of HWA on hemlock physiology has received less attention. Even less recognition has been given to the impact of HWA during diapause. Our results suggest that HWA-infested trees are water stressed due to decreased hemlock growth, water potential, photosynthesis and stomatal conductance. Reductions in photosynthesis and stomatal conductance are expectedly magnified during periods of HWA feeding.
These physiological changes in HWA-infested trees may shed light on possible mechanisms behind HWA-induced death. We suggest taking long-term measurements on HWA-infested trees, from infestation to mortality, to better clarify the mechanism of HWA-induced death. Our study is the first to address the impact of HWA in diapause versus actively feeding and our results suggest that HWA has a lasting physiological impact on hemlock regardless of feeding activity. : Correlation between HWA and water potential. HWA density is negatively correlated with water potential in September (R 2 =0.123), when in diapause, and in October (R 2 =0.313) when actively feeding.