Palatability and Suitability of the Ulmaceae for Spongy Moth Larval Development (Lepidoptera: Lymantriidae)¹

Fifty-nine Ulmus and six Zelkova taxa, varieties, and cultivars were evaluated for suitability and preference for spongy moth (Lymantria dispar) larvae. Ulmus ‘Morton Stalwart’-Commendation^TM and U. ‘Morton Red Tip’-Danada Charm^{TM, ‘}Frontier’, ‘Homestead’, ‘Regal’, and U. elliptica were the most suitable for spongy moth larval development and pupation. Ulmus gausennii, and European elm U. carpinifolia were the least suitable. The remaining Ulmus and Zelkova taxa were either unsuitable (i.e. larvae failed to pupate) or were intermediate in suitability for larval development. Physical leaf characteristics (i.e. leaf thickness and toughness) differed by geographic origin with Asian and European elm leaves thicker than North American elms, and Asian elm leaves tougher than European and North American elm leaves. Overall, leaf thickness and toughness were related, but did not appear to affect larval survival, pupal fresh weight, or dry frass production. Additionally, Asian elm leaves appear to possess both glandular and straight trichomes, but their presence did not affect larval suitability. Hybridization appears to contribute to an ‘U. pumila and U. carpinifolia factor’ which significantly affected host suitability and preference. Results reported here, suggest there is a rich pool of Ulmus and Zelkova taxa for use in tree breeding programs in order to develop resistant trees to invasive lethal insect pests such as Asian long-horned beetle (ALB) [Anoplophora glabripennis (Motschulsky, 1853)], emerald ash borer (EAB) (Agrilus lanipennis Fairmaire, 1888), and diseases including Dutch elm disease (DED) (Ophiostoma spp.), and elm yellows (Candidatus Phytoplasma ulmi) for use in landscapes and urban reforestation efforts, and for their broad hardiness range, ability to tolerate variety of soil conditions, and minimal maintenance in harsh urban environments.

Since its introduction into the United States in the mid-19^th century, the spongy moth (Lymantria dispar, L.) (Lepidoptera: Lymantriidae) is still considered one of the most destructive and persistent pest of nursery crops, landscape plants, and rural and urban forest trees. Spongy moth larvae feed on over 500 woody plant species (Forbush and Fernald 1896, Liebhold et al. 1995), preferring oaks (Quercus spp.) (Barbosa and Krischik 1987), and certain other hardwood species (Elkinton and Liebhold 1990, Shields, et al. 2003). Chronic defoliation can lead to trees that may be predisposed to abiotic factors such as drought, and biotic factors including lethal secondary wood-boring insects and pathogens, for example, the two-lined chestnut borer (Agrilus bilineatus) Weber) and pathogens such as Armillaria root rot [Armillaria mellea, (Vahl) P. Kumm]. Common to most woody landscape and forest trees, host plant resistance for L. dispar has historically been compiled from anecdotal field studies and observations (Forbush and Fernald 1896, Mosher 1915) and while helpful, these studies require some level of interpretation as for host susceptibility, preference, and suitability. More recently, Liebhold et al. (1995) ranked over 600 North American angiosperm and gymnosperm tree species as to their susceptibility to the spongy moth based on previous field and laboratory tests. These susceptibility and suitability rankings were based on spongy moth defoliation, larval abundance, growth and survival, larval foliage preference, pupal weights, and ratio of pupal weight on the host to the pupal weight on white oak (Q. alba L.) (standard host) (Liebhold et al. 1995). Concurrently, extensive studies have focused on the relative susceptibility and suitability of oaks (Quercus spp.) because of the importance of the genus in the ecology and sustainability of eastern and mid-western North American and European forests (Barbosa and Capinera 1977, Barbosa and Greenblatt 1979, Barbosa et al. 1983, Lechowicz and Jobin 1983, Lechowicz and Mauffette 1986, Mauffette et al. 1983, Miller and Hanson 1989a,b, Montgomery 1991, Miller et al. 1991, Davidson et al. 1999, Foss and Rieske 2003, Rieske et al. 2003, Shields et al. 2003, Miller 2008, Milanovic et al. 2014). However, of the over 600 tree species listed in Liebhold et al. (1995), only U. americana and U. parvifolia are listed with suitability rankings of 2 and 1, respectively, and no rankings are provided for Zelkova taxa.

In recent years, a wide variety of Asiatic woody plant species, including Ulmus and Zelkova taxa, has been discovered and developed for possible use in tree breeding programs and for increasing plant diversity in urban forests and landscapes (Ware 1992, 1995). With the recent loss of millions of North American ash (Fraxinus spp.) trees to the emerald ash borer (EAB) (Agrilus planipennis, Fairmaire), and maples (Acer spp.) and other hardwoods to the Asian long-horned beetle (ALB) [Anoplophora glabripennis (Motschulsky, 1853)], there is an important and critical need for the development, and availability of new urban tree species for replanting in EAB and/or ALB affected areas, and to minimize the economic impact on communities and municipalities for the need for chemical protection from spongy moth defoliation (Poland and McCullough 2006, Raupp et al. 2006, Sydnor et al. 2007, Kovacs et al. 2010, Sadof et al. 2011, Vannatta et al. 2012, Hauer and Peterson 2017, Herms et al. 2019). With the exception of the aforementioned North American species of U. americana and the Asian species of U. parvifolia (Liebhold et al. 1995), to the best of our knowledge, no comprehensive studies have been conducted on the relative suitability and preference of recently acquired and developed Ulmus and Zelkova taxa of North American, Eurasian, and Asian parentage for spongy moth development and reproduction. Here, we report the results of a series of no-choice (NC) and multiple-choice (MC) laboratory feeding bioassays to determine the relative feeding suitability and preference of Ulmus and Zelkova taxa for spongy moth larval development and reproduction. Results from this study will contribute to the use of existing and additional Ulmus and Zelkova taxa in landscape and urban forest plantings, which will minimize the need for application of chemical insecticides, and contribute to more diverse landscapes and urban forest settings.

No-choice (NC) larval laboratory feeding trials were conducted using newly hatched first instar larvae as previously described by Miller and Wiegrefe (2021). Fifty-nine different elm taxa were evaluated for relative resistance to larval feeding (refer to Table 1 for a listing of Ulmus and Zelkova taxa, hybrids, varieties, and cultivars tested). Candidate elms are growing at The Morton Arboretum, Lisle, IL, and ranged from a height of 3 to 10 m (9.8 – 32.8 ft) with a diameter at breast height (dbh) (1.37 m) of 5 to 20 cm (1.9 – 7.9 in).

Leaves for the laboratory bioassays were randomly collected from ground level from the canopy of the tree at all four cardinal directions. The leaf samples included the terminal 15 cm (5.9 in) of branches. Samples consisted of an equal portion of actively growing and senescent foliage for each tree. Only fully expanded leaves were used. Leaf samples were taken in this way to compensate for leaf quality within trees. Leaf samples were held in cold storage in plastic bags at 5 C (41 F) for a maximum of 2 days. Leaves collected from each test tree were combined for the laboratory bioassays. Three individual trees (replicates) of each taxon were evaluated. Malus sylvestris and Quercus palustris, both highly preferred hosts of the spongy moth (L. dispar), served as reference species.

Newly emerged first and second instar spongy moth larvae were used in the no-choice (NC) and multiple-choice (MC) feeding trials, and were reared from overwintering egg masses collected from infested trees on the grounds of The Morton Arboretum, Lisle, IL. Upon collection, the eggs masses were placed in brown kraft paper bags, and held in a refrigerator at approximately 40 F until needed. In early May, the egg masses were placed in an incubator under a photoperiod of 16:8 (L:D) hour at approximately 25 C (77 F). Upon hatching, each larva was randomly selected and one larva each was placed in each of 10 snap-cap plastic vials (2.5 cm × 7.5 cm) (1 in × 3 in) with foliage from the test taxon. The vials were examined daily for larval mortality, evidence of feeding, and pupation. Foliage was replaced every 2 days. Vials were placed in plastic trays, and were held in an incubator under a photoperiod of 16:8 (L:D) hour at approximately 25 C (77 F). Each of the three trees (replicates) of each taxon was assayed with 10 individual larvae for a total of 30 spongy moth larvae per taxon. The bioassay for a given larva was terminated at adult emergence. Larval longevity was defined as the difference in days from the date the larva was introduced to the foliage until death or pre-pupation. At the time of pupation (within 12 hr), each individual pupa was weighed (nearest 0.01 g) to obtain the pupal fresh weight. The proportion of larvae reaching pupation was calculated by recording each larva that pupated in each petri dish within a given taxon for all three single tree replicates. At the termination of the no-choice (NC) larval feeding trial, the remaining leaf tissue in each petri dish was removed leaving only the fecal pellets. Fecal pellets were dried in an oven at 50 C (122F) and then weighed (nearest 0.1 mg).

As previously described by Miller and Wiegrefe (2021), ten first instar spongy moth larvae were placed into each of 10 plastic petri dishes (0.6 cm × 15.0 cm) (0.24 in by 6.0 in). Each petri dish served as a replicate. A total of nine studies were conducted. Depending on the study, two to four leaf discs (2.54 cm diameter), with each disc representing one each of the different elm taxon choices, were placed into each dish and randomly arranged around the perimeter. Within each dish, the larvae had equal access to all foliage discs. The petri dishes were placed in clear plastic bags to prevent drying of the leaf discs and were held in an incubator under a photoperiod of 16:8 (L:D) hour ∼25 C (77 F). Condensation of water on the lid of the petri dish indicated a high relative humidity. The dishes were examined daily for 3 days. Each day, the foliage discs were removed from the dishes, replaced, and visually evaluated using a defoliation template for the proportion of leaf tissue removed by larval feeding. New foliage discs were arranged randomly around the perimeter of each dish to eliminate possible bias.

As previously described by Miller and Ware (2022), prior to using the leaves for the NC laboratory feeding trials, 10 leaves for each taxon were measured for leaf thickness, and inner and outer leaf toughness. Leaves were collected in the field as previously described. Leaf thickness was determined by using a Vernier caliper to measure the thickness of each leaf (nearest mm) approximately one-half the distance from the leaf margin to the mid-rib. Inner and outer leaf toughness was determined to the nearest gram using a Chatillon^TM digital force meter (pentrometer) (Greensboro, N.C.) applied to within 0.5 cm from the edge of the leaf for measuring outer toughness, and in the center of the leaf adjoining the mid-rib for inner toughness, respectively.

Suitability of each taxon for spongy moth larvae was defined by larval longevity, percent larval survival, percent of larvae pupating, pupal fresh weight, percent adult emergence, and dried fecal pellet weight in the no-choice (NC) larval feeding trials. Preference was measured using the percent of leaf tissue removed in the multiple-choice (MC) larval feeding trials. The male and female suitability ratio was calculated as the quotient of the MPW and FPW for each respective Ulmus taxon over the MPW and FPW of the reference taxon, Q. palustris, respectively. Male and female suitability rankings were defined using the following scale: suitability ranking 1 > 0.9; 2 > 0.5 to 0.9; 3 < 0.5 (Liebhold, et al. 1995, Miller and Hanson 1989a,b).

Measures of suitability and preference were subjected to analysis of variance (ANOVA) by using taxon as the main effect. The percent of larvae pupating was arcsin transformed before analysis to correct for non-normality. Means of significant effect (5%) were compared with Dunn’s test. A coefficient of correlation was calculated for the rankings for larval development time with larval longevity, proportion of larvae pupating, pupal fresh weights, percent adult emergence, and fecal pellet weights. All data are presented as original means ± SEM. Data were analyzed using Jandel Scientific (1992).

A summary of the suitability of Ulmus and Zelkova taxa for spongy moth larval longevity and dry frass weight is presented in Table 2 for larvae that did not reach pupation, and in Table 3 for larval longevity (days), percent male and female larval survival to pupation, percent pupation and adult male and female emergence, pupal fresh weights, and dry frass weight.

Averaged across all 59 Ulmus species tested in this study, spongy moth larvae lived a mean of six days (range=3-12 days), and no larvae pupated when feeding on 66% of Ulmus taxa (Table 2); two exceptions being larvae feeding on ‘Accolade’ (3% larval survival), and the European species, U. microcarpa (5% of larval survival), but none of the larvae pupated (data not shown). Additionally, averaged across all six Zelkova taxa tested, spongy moth larvae feeding on these taxa lived a mean of four days (range=3-4 days), and none reach pupation (Table 2). Our results are consistent with a study by Miller and Hanson (1989a,b) who found that 1^st instar spongy moth larvae feeding on Z. serrata died, and 2^nd instars developed into pre-pupae, but then died.

Consistent with other suitability measures, larvae that lived the longest on highly suitable hosts (i.e. ‘Commendation’, ‘Danada Charm’) produced significantly greater dry frass weights (F=86.7, P<0.001) compared with all of the remaining Ulmus taxa. The reverse was also true. Dry frass weights for the highly preferred standards of M. sylvestris and Q. palustris were 2,188 g and 1,498 g, respectively. Dry frass weight was a moderate predictor of larval longevity (R=0.53, P<0.001) (Tables 2 and 3).

Larvae lived significantly longer (F=8.6; P<0.001) (mean=28 days; range=20-35 days) when feeding on the simple and complex hybrids of ‘Frontier’, ‘Pioneer’, ‘Regal’, ‘Urban’ elm; the North American cultivar, U. americana ‘Liberty’; and European taxa of U. elliptica, U. foliaceae, and U. sukaczevii compared with larvae feeding on the remaining Ulmus taxa and living less than 11 days. Larvae feeding on U. gausennii, ‘Homestead’, ‘Vanguard’, U. carpinifolia, and U. procera lived an intermediate number of days (mean= 16 days; range=14-18 days) (Table 3). Significantly more male and female larvae survived to pupation when feeding on U. gaussennii, ‘Triumph’, U. carpinifolia, U. procera, and U. sukaczevii compared with the remaining Ulmus taxa (male larvae, F=5.7, P<0.001, female larvae, F=14.2, P<0.001). Larvae feeding on the highly preferred standards of M. sylvestris and Q. palustris lived a mean of 16 and 10 days respectively with greater than 30% to 60% of male and female larvae pupating (Table 3). Collectively, for all Ulmus taxa evaluated, percent larval longevity was moderately correlated with larval survival (R=0.69, P<0.001), but was a weak predictor of percent pupation (R=0.30, P=0.02). Percent larval survival was highly related to the percentage of larvae pupating (R=0.81, P<0.001).

A mean of 10% (range=3%-17%) of male larvae, and no female larvae pupated when feeding on Asian taxa of U. gausennii, U. laciniata, U. lamellosa and U. szechuanica. A mean of 20% (range=3-40%) of male larvae and a mean of 14% (range=3%-23%) of female larvae pupated when feeding on the complex Eurasian hybrids of ‘Commendation’ ‘Danada Charm’, ‘Frontier’, ‘Homestead’, ‘Regal’, ‘Triumph’, and ‘Vanguard’ (Table 3). Thirty-seven percent of the remaining male pupae and 23% of female pupae; and 40% of male pupae and 13% of female pupae emerged as adults having developed on the complex Eurasian hybrids ‘Commendation’ and ‘Danada Charm’, respectively. A mean of 24% of male and 10% of female larvae pupated when feeding on the European taxa of U. elliptica, U. foliaceae, U. procera, and U. sukeczvii. Twenty-seven percent (27%) of male larvae feeding on the North American cultivar U. americana ‘Liberty’ pupated, but no female larvae pupated. For the highly preferred standards of M. sylvestris and Q. palustris, 60% and 37% of male larvae and 30% and 33% of female larvae pupated, respectively, and a like percentage of adult male and female moths emerged. The percentage of larvae pupating was highly correlated with percent adult emergence (R=0.94, P<0.001).

Overall, female spongy moth pupae were significantly heavier (mean=0.816g; range=0.610-1.290g) by 2.5 times compared with male spongy moth pupae (mean=0.323g; range=0.255-0.564g) (t = 22287.000, P = <0.001) (Table 3). Male pupae developing on U. szechuanica, ‘Commendation’, ‘Danada Charm’, ‘Frontier’, ‘Homestead’, ‘Pioneer’, ‘Regal’, U. elliptica, and U. foliaceae were significantly heavier compared with male pupae developing on U. gausennii, U. laciniata, U. lamellosa, ‘Triumph’, U. americana cultivars ‘Jefferson’ and ‘Liberty’, U. carpinifolia, and U. sukaczevii. Male pupae developing on ‘Vangard’, and U. procera had intermediate pupal fresh weights (PFW) (F=4.3; P<0.001). Female pupae developing on ‘Danada Charm’ and ‘Frontier’ and the highly preferred standard of Q. palustris were significantly heavier compared with pupae developing on ‘Homestead’, ‘Vanguard’, U. elliptica, and U. foliaceae, Female pupae developing on ‘Commendation’ and the highly preferred standard, M. sylvestris had intermediate PFWs (F=4.5; P<0.001). Male and female PFWs for the highly preferred species of M. sylvestris and Q. palustris were 0.527g and 0.536g, and 0.865g and 1.097g, respectively and were comparable with other studies (Miller and Hanson 1987, Miller and Wiegrefe 2021) (Table 3). PFWs were moderately correlated with larval longevity (R=0.53, P<0.001) and percent adult emergence (R=0.64, P<0.001).

Significantly more male and female moths emerged when completing development on ‘Commendation’ and ‘Danada Charm (13% to 40%) compared with all of the remaining Ulmus taxa (male moths, F=13.8, P<0.001, female moths, F=6.8, P<0.001) (Table 3). Additionally, compared with all of the remaining Ulmus taxa, a significantly greater number of male moths completed development on U. gaussennii, but no female moths emerged when developing on this taxon. Percent adult emergence for both male and female moths for the reference species of M. sylvestris and Q. palustris was slightly higher (30% to 60%), respectively. Adult female spongy moths laid significantly more eggs per ovipositing female when developing on the highly preferred standard of Q. palustris (516 eggs) compared with ’Commendation’ (255 eggs) and the highly preferred standard of Malus spp. (227 eggs). Ovipositing females laid an intermediate number of eggs when developing on ‘Danada Charm’ (323 eggs) (F=20.9; P<0.001) (Table 4). Ninety percent to 100% of emerging females laid eggs on ‘Commendation’, ‘Danada Charm’, and Q. palustris (standard) compared with only 50% of emerging adult females laying eggs on M. sylvestris (standard), but there was no significant difference in the percent females ovipositing (F=1.0; P=0.15) (Table 4). The number of eggs laid per ovipositing female was highly correlated with the percentage of females ovipositing (R=0.96, P=0.04).

Overall, female larvae had a significantly higher (t=2.9; P=0.008) survival rate (47%) compared with male larvae (41%), and the mean percent larval survival (i.e. larvae living to pupation) differed significantly (F=8.6; P<0.001) by geographic origin (Table 5). Spongy moth larvae feeding on simple and complex Eurasian hybrids and European taxa both had a mean survival rate of 31% (Eurasian range=0%-48%; European range=0%-56%) compared with larvae feeding on Asian taxa with <2% of larvae surviving to pupation (range=0%-64%). Note that spongy moth larvae survived to pupation on only four out of 26 Asian taxa tested, namely U. gausennii (male=47%; female=63%), U. laciniata (male=7%; female=10%), U. lamellosa (male=37%; female=0%), and U. szechuanica (male=3%; female=7%) (Tables 3 and 5). Larvae feeding on North American taxa had an intermediate mean percent survival rate of 14% (range=0%-39%), but like the Asian taxa, the only larvae that survived to pupation were those feeding on the U. americana cultivar ‘Liberty’ (Tables 3 and 5).

Suitability ratios (SRs), for male and female spongy moth larvae that pupated, when feeding on Ulmus taxa, are summarized in Table 6. The overall mean male pupal SR was 0.77 (moderate) and 0.75 (moderate) for female pupae. The mean male SR for Asian taxa was 0.61 (moderate) (range=0.48 to 0.86). No female larvae pupated when feeding on Asian taxa. The mean male pupal SR was 0.84 (range=0.71-1.05) (moderate), and 0.81 (range=0.75-0.84 (moderate) for female larvae developing on simple and complex Eurasian hybrids. Male larvae feeding on the European taxa of U. elliptica, U. foliaceae, and U. sukaczevii had a mean male pupal SR of 0.84 (range=0.68-0.91) (moderate), and a mean female pupal SR of 0.67 (range=0.59-0.78) (moderate) (Table 6). These results suggest that the majority of the Ulmus taxa tested in this study have low (SR< 0.5) to moderate (SR=0.5 to 0.9) host suitability for male and female spongy moth larvae feeding, development, and pupation. Host suitability significantly affected male and female pupal weights. (t = 22287.000, P = <0.001) with female pupae being 2.5 times heavier than male pupae (male = 0.323 g versus female = 0.816 g). Overall, male pupae had a low suitability ranking (value=3, pupa < 700 mg) and female pupae had a moderate suitability ranking (value=2, 700 mg < pupa < 1,000 mg) (Miller and Hanson 1989a,b). Suitability rankings for male pupae for both of the standards of Malus sylvestris and Q. palustris were 3 (0.527g and 0.536g), and 2 (0.857g) and 1 (1.097g) for female pupae, respectively. Male and female pupal weights in our study were comparable with other studies for another highly preferred host, Q. macrocarpa (male=0.339g, female=0.886g) (Miller and Hanson, 1989a,b, Miller and Wiegrefe 2021b). Percent male and female larval development survival was not correlated with male and female suitability ratings (SRs) (male, R=0.21, P=0.45; female, R=0.55, P=0.20). However, both male and female pupal fresh weights (PFW) were very strong predictors of suitability (male: R=0.85, P<0.001, female: R=0.99, P<0.001). Liebhold et al. (1995) does not provide any SR values for Ulmus taxa, therefore prohibiting any direct comparisons among Ulmus taxa evaluated in this study, but Liebhold et al. (1995) does indicate that 11 Ulmus taxa have moderate suitability rankings (value=2), with the exception of U. parvifolia which is rated low (value=1).

A summary of the nine MC studies is presented in Table 7. In MC study #1, spongy moth larvae preferred ‘Commendation’ over ‘Danada Charm’ (t=8.2; P<0.001). When larvae were given a choice in MC study #2 between ‘Commendation’ and ‘Accolade’, significantly more ‘Commendation’ and U. pumila leaf tissue was consumed compared to larvae feeding on ‘Accolade’ (F=19.9, P<0.001). In MC study #3, when larvae were presented a choice between ‘Accolade’ and ‘Danada Charm’, neither were preferred with <10% of leaf tissue consumed. In MC study #4, spongy moth larvae were allowed to choose between ‘Vanguard’ and its U. japonica and U. pumila parentage. Significantly more leaf tissue was consumed by larvae feeding on U. pumila as compared to larvae feeding on ‘Vanguard’ and U. japonica (F=18.4, P<0.001). Spongy larvae preferred U. davidiana foliage over the other Davidiana complex species of U. japonica and U. wilsoniana in MC study #5 (F=17.5, P<0.001). When examining elm taxa of Eurasian parentage, ‘Frontier’ was more highly preferred over ‘Pioneer’ and U. carpinifolia in MC study #6 (F=4.4, P=0.04). In MC study #7, the complex hybrid ‘Homestead’ was highly preferred followed by U. pumila. The complex hybrid, ‘Regal’, was least preferred and ‘Urban’ elm was moderately preferred (F=7.8, P=0.002). In MC study #8, there was no significant difference in preference when larvae fed on ‘Patriot’, U. wilsoniana, and ‘Urban’ elm. In the final MC study, #9, as before, ‘Frontier’ was highly preferred as compared to U. carpinifolia, which was moderately preferred and U. parvifolia which was least preferred (F=8.9, P=0.004). For the Ulmus taxa evaluated in the multiple-choice (MC) preference studies, larval longevity was highly correlated with percent larval survival (R=0.83, P<0.001) for the same taxa evaluated in the NC feeding studies.

A summary of elm leaf thickness and inner and outer leaf toughness by geographic origin and taxon is presented in Tables 8 and 9, respectively. Asian and European elms had significantly thicker leaves (1.5X) compared with North American elms (F=70.0, P<0.001), and Asian elm leaves were significantly tougher than European and North American elms by a factor of 1.3X and 1.4X for inner and outer leaf toughness, respectively (inner toughness: F=31.1, P<0.001 and outer toughness: F=39.1, P<0.001). There was no significant difference between inner and outer leaf toughness for all elms tested (Table 8). Leaf thickness was moderately correlated with inner and outer leaf toughness (inner R=0.36; P=0.02; outer R=0.42, P<0.006), and inner leaf toughness was strongly correlated with outer leaf toughness (R=0.94, P<0.001). Of the 21 Asian elm taxa evaluated, U. castaneifolia had significantly thicker leaves and U. gausenni, U. glaucescens, U. japonica, U. pumila, and U. szechuanica; and the complex hybrids of ‘Accolade’, ‘Commendation’, and ‘Vanguard’ had the thinnest leaves. Leaves of all of the remaining 12 Asian elm taxa were intermediate in thickness (F=353.3, P<0.001). Leaves of U. bergmaniana, U. gausenni, U. japonica, and U. pumila were the most tender (least tough) compared with the toughest leaves of U. glaucescens var. lasiophylla, U. macrocarpa, and ‘Triumph’. Leaves of all of the remaining Asian elm taxa were intermediate in inner and outer leaf toughness (inner toughness: F=254.4, P<0.001; outer toughness: F=303.0, P<0.001) (Table 9).

For the simple and complex Eurasian hybrids, Frontier, Homestead and Regal had significantly thinner and more tender leaves than Patriot, Pioneer, and Prospector (thickness, F = 47.4, P<0.001; inner toughness, F=17.2, P<0.001; outer toughness, F=17.8, P =<0.001). Among European elms, U. foliaceae had the thinnest leaves and U. procera and U. sukaczevii had significantly thicker leaves (F=44.2, P<0.001). Ulmus elliptica and U. procera had significantly tougher leaves compared with U. sukaczevii (inner toughness, F=30.6, P<0.001; outer toughness, F=4.9, P<0.001).The North American elm U. thomasii had significantly thinner leaves compared with U. alata and U. crassifolia (F=40.4, P<0.001). There was no significant difference in inner leaf toughness for all of the North American elms tested in this study, but outer leaf toughness for U. alata was significantly greater compared with U. americana and U. thomasii (F=4.4, P<0.005).

For all of the elm taxa tested, larval longevity was not correlated with leaf thickness (R=0.16, P=0.32), inner leaf toughness (R=0.30, P=0.06), or outer leaf toughness (R=0.29, P=0.07). Likewise, percent larval survival was not correlated with leaf thickness (R=0.14, P=0.40) or inner (R=0.17, P=0.30) and outer leaf toughness (R=0.22, P=0.17). Additionally, leaf thickness and leaf toughness were not good predictors of the amount of dry frass weight produced (thickness: R=0.14, P=0.0.41; inner toughness: R=0.01, P=0.97, outer toughness: R=0.01, P=0.87) or pupal fresh weight (thickness R=0.15, P=0.37; inner toughness: R=0.11, P=0.51, outer toughness: R=0.10, P=0.47).

In summary, of the 59 Ulmus and six Zelkova taxa, hybrids, varieties, and cultivars tested in this study, the vast majority of Ulmus and all of the Zelkova taxa were found not to be suitable for spongy larval development, pupation, and adult reproduction. Results from this study, provides new insight into the potential of Ulmaceae for use in future tree breeding programs.

Leaf tissue chemical analysis, and its influence on host suitability and preference was beyond the scope of this study and was not examined. However, we will discuss the effect of physical plant traits (i.e. leaf thickness, toughness, and pubescence), the relationship between elm host plant phenology and larval feeding, and possible effects of hybridization on host suitability and preference of Ulmus taxa.

Physical plant traits, such as leaf toughness and thickness, can greatly influence invertebrate feeding and usually correlate with leaf fiber and lignin content (Graca and Zimmer 2005). Agrawal and Fishbein (2006) found leaf toughness could be used to predict herbivory of many plants including milkweeds (Asclepias spp.). Raupp (1985) found that the toughness of willow (Salix spp.) leaves reduced leaf beetle feeding due to mandibular wear. In field defoliation trials, Bosu et al. (2007) observed that the thicker and tougher leaves of Asian elm taxa experienced much less defoliation by the elm leaf beetle (Pyrrhalta luteola Muller). Potter and Kimmerer (1988) found that the thick glabrous cuticle and tough leaf margins of the American holly (Ilex opaca Aiton) leaves were more important than leaf margin spines in deterring edge feeding caterpillars. In a previous study by Miller and Ware (2022), they determined that Asian and European elm leaves were significantly thicker than North American elm leaves, and Asian elm leaves were tougher than European and North American elm leaves. Additionally, elm leaf thickness was weakly correlated with inner and outer leaf toughness, but inner and outer leaf toughness was strongly related. Recently, Miller and Wiegrefe (2021) found that spongy moth development was not correlated with the thickness of Carpinus leaves but was related to inner and outer leaf toughness. Additionally, Miller and Ware (2014) determined that the thicker and tougher leaves of most Asian elm taxa and hybrids were less susceptible to feeding by the elm leafminer (Fenusa ulmi Sundevall). In the present study, elm leaf thickness and inner and outer leaf toughness were not good predictors of larval longevity, pupation, pupal fresh weights, or frass production. Schneider (1980) reported that it takes approximately three days for newly hatched fall cankerworm (Alsophila pometaria Harris) larvae to starve to death which is consistent with our results, in this study, where spongy moth larvae lived an average of only 3-5 days when feeding on a majority of Ulmus and Zelkova taxa.

Our results are similar to larval feeding studies by Miller (2000) and Miller et al. (2001) where they found that the majority of these same Asian elm taxa were unsuitable for spring cankerworm [Paleacrita vernata (Peck, 1795)] and fall cankerworm [Alsophila pometaria (Harris, 1841)] larval development and pupation. Similarly, less than 30% of spongy moth larvae feeding on thin, thick, tough or tender leaves of ‘Regal’, ‘Frontier’, U. foliaceae, and U. sukaczevii survived, and none pupated or emerged as adult moths. In contrast, spongy moth larvae feeding on either the thin, thick, or tougher leaves of ‘Commendation’ and ‘Danada Charm’ had over 50% larval survival, and an equal proportion of larvae developing into adults. Studies by Cunningham and Dix (1983) and Dix et al. (1996) found that Siberian elm (U. pumila) clones with thicker leaves had significantly higher rates of percent defoliation and thinner Siberian elm leaves experienced less feeding by the spring cankerworm (Paleacrita vernata) (Peck). They suggest that spongy moth larvae may prefer to feed on trees with thicker leaves because of the increased feeding efficiency that would be offered by leaves with a greater weight per unit area. Their findings are consistent with our study where U. pumila leaves were significantly thinner and less tough compared with other Ulmus taxa, and spongy moth larvae only lived a mean of five days, and none reached pupation. Spongy moth larvae tend to feed inward from the leaf margins, toward the mid-vein, eventually consuming the entire leaf. It would seem reasonable that more tender elm leaves would be more palatable for younger larvae to feed on, but that does not appear to be the case in this study. It is possible that older larvae are able to feed on tougher and thicker leaves, but again, that was not consistently observed in this study. Thinner leaves on trees may also mature more rapidly, thus accumulating tannins and cellulose faster than trees with smaller and thicker leaves (Feeny 1970). Feeny (1970) also documented the deterioration in quality of oak leaves as food for Lepidoptera larvae during leaf maturation. The development of Lepidoptera larvae on oaks was adversely affected by tannin and cellulose accumulation in the leaf tissues and cell walls, respectively.

The amount of leaf tissue removed in the multiple-choice (MC) studies by larvae feeding on Asian taxa was consistent with larval longevity and percent larval survival in the no-choice (NC) feeding trials, suggesting that other factors (i.e. leaf chemistry, pubescence) besides leaf thickness and toughness may be responsible for host preference and may play a more important role in the suitability and preference of elms for spongy moth larval development. Paluch et al. (2009) found significant differences in the total phenolic content of 13 of some of the same Asian elm taxa and cultivars evaluated here, but no significant differences were identified between spongy moth larval longevity, pupal fresh weight, or percent adult emergence with regard to total leaf phenolic content. In another study, Paluch et al. (2008) found that Asian elms belonging to the David complex (i.e. U. propinqua, U. japonica, U. wilsoniana) possessed a much larger diversity of leaf terpenes, phenolics, alkaloids, and lipids. More specifically, the diversity of lipids showed a significant effect relating to David complex parentage suggesting that elms belonging to this complex are better suited for spongy moth larval development than other elms. However, in this study, none of the larvae feeding on members of the David complex survived to pupation. It appears that host suitability relies on a complexity of leaf chemical composition, and that generalist herbivorous insects, such as the spongy moth, might selectively use host plants containing a larger diversity of compounds (Schultz 1988, Paluch et al. 2008).

Spongy moth, and spring and fall cankerworm egg hatch occurs between 100 and 200 DD₅₀ and which corresponds closely with adult elm leafminer emergence (Miller and Ware 2014, Orton, 2007). In an elm leafminer study by Miller and Ware (2014), they observed considerable variability in spring elm leafing out phenology for many of the same less suitable Asian elm taxa evaluated in this study. For example, less suitable elms, such as U. davidiana, U. japonica, U. lamellosa, U. pumila, and others tend to begin leafing out at approximately the same time as initial spring (P. vernata) (Peck) and fall cankerworm (Alsophila pometaria Harris) egg hatch, and adult leafminer emergence, but suffer no or negligible feeding damage from either insect (Miller and Ware 2014, Miller et al. 2001a). Other less preferred Asian, European, and North American elms, such as U. parvifolia,U. wilsoniana, U. carpinifolia, and U. thomasii, are typically breaking bud or just flowering, and generally do not leaf out until after cankerworm and spongy moth larval feeding is well underway, and peak adult elm leafminer emergence has passed. This phenological pattern suggests that these taxa possibly avoid feeding damage, and appear less susceptible (Miller and Ware 2014, Miller et al. 2001). Conversely, field observations have revealed that European species such as U. elliptica, U. glabra, U. laciniata, and U. sukaczevii, consistently leaf out at approximately the same time of initial cankerworm, spongy moth, and elm leafminer insect feeding and oviposition, respectively, and these same species were found to be suitable for spongy moth larval development, and moderately to highly susceptible to elm leafminer feeding damage. Witter and Waisamer (1978) also observed that population levels of tortricid caterpillars on early flushing aspen clones were higher than those on late flushing aspen clones. Because the emerging 2nd-stage tortricid larvae were synchronized with their food source, they had higher survival and less dispersal than larvae on unsynchronized leaves. The strong correlation between leafing out phenology of the aforementioned Ulmus taxa and their relative susceptibility, suitability, and preference suggests a spectrum of apparent resistance for spongy moth larval feeding and development. Many of these same Asian taxa evaluated in this study have also been shown to be less suitable for other leaf-feeding insects such as the elm leaf beetle (Pyrrhalta luteola Muller), and Japanese beetle (Popillia japonica Newman, 1841) (Miller and Ware 1997, 1999, 2014; Miller et al. 1999, 2001a,b, Bosu et al. 2007, Bosu and Wagner 2007, Condra et al. 2010, Griffin et al. 2017). These phenological traits should be more fully explored in future elm tree breeding programs.

Hybridization may have both a positive and/or negative influence on host suitability and preference for spongy moth larvae (Paige and Capman 1993, Orians 2000, Cheng et al. 2010, Lopez-Caamal 2014). In this study, U. carpinifolia, U. pumila, and members of the David complex (i.e. U. japonica, U. wilsonsiana, U. propinqua) appear to affect feeding preference by spongy moth larvae compared with other Ulmus taxa. For example, in the multiple-choice (MC) laboratory feeding bioassays, when a much less suitable or preferred taxon such as U. parvifolia, is crossed with a preferred taxon, such as U. carpinifolia, to form ‘Frontier’ suitability and preference increases significantly. Conversely, when the more suitable and preferred U. carpinifolia is hybridized with the much less suitable U. glabra to form ‘Pioneer’, there is no appreciable change in suitability or preference. Further, when the preferred U. pumila is included in complex hybrids, such as ‘Homestead’, ‘Regal’, ‘Commendation’, ‘Urban elm’, and ‘Vanguard’, larval suitability and preference also increases, but the opposite is also true for the U. pumila × U. japonica hybrids of ‘Cathedral’ and ‘New Horizon’ where preference decreased. When the very unsuitable and much less preferred U. wilsoniana or U. japonica are both included in the complex hybrids of ‘Patriot’, ‘Accolade’, and ‘Danada Charm’, suitability and preference is significantly reduced. Results reported here suggest that U. carpinifolia and U. pumila tend to increase preference in simple and complex hybrids, but members of the David elm complex such as U. wilsoniana and U. japonica tend to decrease suitability and preference. Hybrids containing U. glabra and U. parvifolia parentage do not appear to have a significant effect on the suitability or preference for spongy moth larval feeding and development. The reason for this hybrid effect is not clear. A similar phenomenon, the “U. pumila and U. carpinifolia factors” has been observed in other studies with elms (Ulmus spp.). For example, when Siberian elm (U. pumila) is incorporated into hybrids with other taxa it appears to affect the susceptibility, suitability, and preference of elms to an extensive herbivorous insect guild including Japanese beetle, elm leaf beetle spring and fall cankerworm, elm leafminer (Miller and Ware 1997, 1999, 2014, Miller et al. 1999 2001a,b, Bosu and Wagner, 2007, Bosu et al. 2007, Condra et al. 2010; Griffin et al. 2017). In all likelihood, elm leaf chemistry may play a role in host preference and suitability (Paluch et al., 2008, 2009).

While leaf trichomes may provide a general defense, the effect of pubescence may be positive, negative, or non-existent depending on the herbivore species, and can be influenced by genotype, the environment, leaf development stage, and previous feeding (Levin 1973, Pillemer and Tingey 1976, Roberts et al. 1979, Southwood 1986, Peter et al. 1995, Dix et al. 1996, Werker 2000, Dalin and Bjorkman 2003, Bosu and Wagner 2007, Dalin et al. 2008, Kariyat, et al. 2018). Conversely, trichomes may also play a role in attracting herbivores to the host plant (Southwood 1986) and may also help protect plants from insect herbivores specifically as it relates to feeding, growth, survival, and oviposition. Feeding may be negatively correlated with trichome density, but is generally considered a “soft weapon” in plant defense compared to other plant traits (Levin 1973, 2006, Potter et al. 1998, Miller and Ware 1999, Miller 2000, Miller et al. 2001a,b, Dalin et al. 2008).

In a previous study, Miller and Ware (2022), found some interesting contrasts in elm leaf trichome density and morphology on the abaxial (lower) and adaxial (upper) leaf surfaces of the pubescent elm taxa of U. chenmoui, U. glauscens, U. lamellosa, U. macrocarpa, and U. propinqua (Fig. 1 and 2 a,b,c,d). Namely, trichomes on the lower leaf surface of U. chenmoui are of medium density, uneven length, and not all are upright. On the upper leaf surface, trichomes are scattered with bulbous bases, some curving and of uneven lengths. Trichomes on the lower leaf surface of U. glaucescens (Fig. 1d, 2d) and upper leaf surface of U. propinqua (Fig. 1c) are very dense and thick, fine, straight, and of even length, with some curvature on the upper leaf surface. In contrast, the leaf trichomes on the lower and upper surfaces (Fig. 2 a,c) of U. macrocarpa are of medium density, but have a bulbous base which may suggest they are glandular, and possess feeding deterrent chemicals making the leaf less palatable. The trichomes on U. lamellosa leaves are scattered and many tend to be broken off during the growing season, giving the leaves a very rough or sandpapery feel (author’s personal observations) (Fig. 1a, 2b). However, in this study, leaf trichome density (i.e. pubescence) does not appear to be a good indicator of suitability for spongy moth larvae. With the exception of U. gausenii, and U. lamellosa, neither pubescent, or non-pubescent Asian elm taxa were suitable for spongy moth larval development and pupation. Our results are consistent with studies by Miller (2000) and Miller et al. (2001) where they found that these same Asian elm taxa were not suitable for spring and fall cankerworm development. They also found, in multiple choice studies, that there was no significant difference in leaf tissue consumed for spring and fall cankerworm larvae feeding on leaf discs of U. glaucescens var. lasiophylla, U. lamellosa, and U. macrocarpa with trichomes present versus leaf discs with trichomes removed. Conversely, Miller and Wiegrefe (2021) found that pubescent leaves of certain Carpinus taxa were less suitable and less preferred by spongy moth larvae, but there was no relationship between the amount of leaf tissue removed and trichome density.

Liebhold et al. (1995) does not provide any specific SR values for Ulmus taxa, therefore prohibiting any direct comparisons for specific Ulmus taxa evaluated in this study. However, using SR’s from this study, as previously described by Montgomery (1991) and Liebhold et al. (1995), we found a strong correlation between both male and female pupal weights and host suitability. Suitability appears to have a significant effect on both male and female pupal weights with females weighing nearly 2.5 times more than males. Of all of the elm taxa tested in this study, the complex Eurasian hybrids of ‘Commendation’, ‘Danada Charm’, ‘Frontier’, and ‘Vanguard’ appear to be moderately suitable (value=2 or 0.5 < ratio <0.9) (Montgomery, 1991) for larval development, and of these same taxa, only ‘Commendation’ and ‘Danada Charm’ appear to be suitable for spongy moth reproduction. With the above exceptions, all of the remaining Ulmus and Zelkova tested here do not appear to be suitable for spongy larval development and reproduction.

In conclusion, many of the aforementioned taxa are presently in the horticulture trade and are performing well as parkway and landscape trees due to their resistance to Dutch elm disease (DED), and elm yellows, and most sap-feeding and leaf-feeding insect pests (Smalley and Guries 1993, Sinclair et al. 2000, Jacobs et al. 2003, Costello et al. 2004, Bosu et al. 2007, Herath, et al. 2010, Griffin, et al. 2017). However, there may be some use limitations for some of the above taxa in settings with chronic or high spongy moth, Japanese beetle, elm leaf beetle, elm leafminer, and elm flea weevil [Orchestes ulmi (Linnaeus, 1758)] populations (Miller 2000, Miller et al. 2001a,b, Condra et al. 2010, Potter and Redmond 2013, Griffins et al. 2017). Elm and Zelkova spp. selections should be based on hardiness zones and growth characteristics when planning for future urban tree and landscape plantings, and overall performance should be tempered with tolerance and/or resistance to local insect and disease pressure.

Some of the Asian elms species evaluated in this study including U. szechuanica, and U. parvifolia and cultivars, members of the David elm complex (i.e. U. davidiana, U. japonica, U. propinqua, and U. wilsoniana), North American taxa of U. alata, U. crassifolia, U. serotina, U. thomasii; European taxa of U. glabra, U. microcarpa, U. rubraU. sukaczevii; and Zelkova taxa and cultivars should be considered in future tree breeding efforts because of their desirable horticultural attributes, their ability to tolerate harsh urban growing conditions and their resistance to DED and herbivorous insect pests particularly where spongy moth populations are sporadic or absent (Ware 1992, 1995, Miller et al. 1999, 2001, Miller 2000, Warren 2000, Hartman, et al. 2007, Brady et al. 2008, McPherson et al. 2008, Dirr 2009, Condra et al. 2010, Potter and Redmond 2013, Griffin et al. 2017, Giblin and Gilman 2021). It is apparent that physical leaf factors do not play a major role in the susceptibility, suitability, and preference of Ulmaceae for spongy moth larval development and reproduction; therefore, more research is needed to better identify possible chemical factors involved with Ulmaceae host plant resistance, and how these factors might be incorporated into a more comprehensive tree breeding program.