Fasciation in Invading Common Mullein, Verbascum thapsus (Scrophulariaceae): Testing the Roles of Genetic and Environmental Factors1

Shahin Ansari2,3 and Curtis C. Daehler 2,4

Abstract: In Hawai‘i, Verbascum thapsus L. exhibits high rates of fasciation, which could have ecological and evolutionary consequences for spread of this noxious weed. Fasciated plants produce more seed capsules on average; however, the cause of fasciation in V. thapsus is not known. This study investigated w­hether fasciation in V. thapsus has a simple genetic basis, or whether it is caused by physical damage or pathogenic bacteria. Plants derived from self-pollinated fas- ciated and normal plants were grown in a field common garden and subjected to mechanical damage (simulated herbivory) and natural herbivory. Bacteria cul- tured from normal and fasciated plants were compared, and field plants were inoculated with a slurry of fasciated tissue. In the common garden, 31% of plants developed fasciation, but fasciation did not follow a simple monogenic pattern of inheritance. Artificial damage substantially reduced fasciation rates; damaged plants were between 1.3 and 32 times less likely to become fasciated, compared with undamaged plants. Bacterial isolates were similar between normal and fas- ciated plants and no inoculated plants developed fasciation, suggesting that bac- teria do not cause fasciation. Fasciated and normal plants often grow less than 1 m apart, indicating that climatic factors are not inducers of fasciation. Localized combinations of environmental conditions in Hawai‘i may promote frequent and persistent fasciation.

Analogous to cancer in animals, fascia- White 1943). Morphologically, fasciation tion is a term used to describe abnormal or may be manifested as a flat, banded, or rib- uncontrolled proliferation of cells and tissue bonlike expansion of the stem and /or the in- growth in plants. Fasciation can occur in florescence of a plant. According to Gorter roots, leaves, flowers, fruits, cotyledons, and (1965), “true fasciations” result from a lateral underground stems, and it is particularly no- expansion of the apical meristem. Linear fas- ticeable in the stems and branches of plants ciations are the most common type, wherein having indeterminate growth ( Jones 1935, the stem, in addition to being broadened in one plane, is often twisted and curled (La- motte et al. 1988). Besides disorganized tissue 1 This research was supported by a grant from the Ecology, Evolution, and Conservation Biology (EECB) growth, a most striking feature of fasciated Program at University of Hawai‘i. Manuscript accepted plants is an unusual increase in weight and /or 11 February 2011. volume of tissue ( White 1943). Fasciation was 2 Department of Botany, 3190 Maile Way, Honolulu, extensively studied in the early twentieth cen- Hawai‘i 96822. tury under the subject of plant teratology 3 Current address: SWCA Environmental Consul- tants, 201 Merchant Street, Honolulu, Hawai‘i 96813. ( Wordsell 1915, 1916). Fasciation is wide- 4 Corresponding author (e-mail: daehler@hawaii spread geographically and occurs in at least .edu). 107 plant families; it has been reported from herbs, shrubs, vines, and trees, although the most striking fasciations probably occur Pacific Science (2011), vol. 65, no. 4:451 – 463 doi: 10.2984/65.4.451 among biennials ( White 1943, 1948). © 2011 by University of Hawai‘i Press Fasciation in plants may have a genetic All rights reserved b­asis, or it could be caused by an environmen-

451 452 PACIFIC SCIENCE · October 2011 tal factor or bacterial infection. Rhodococcus bean (Vicia faba) have all been reported to in- fascians is a pathogenic bacterium that causes duce fasciation ( White 1948). fasciation in several monocots and dicots Verbascum thapsus L., commonly referred ranging across 40 families and 86 genera (Put- to as common mullein or great mullein, is a nam and Miller 2007). Rhodococcus fascians in- Eurasian biennial that is classified as a noxious duces a range of malformations in the host, weed in Hawai‘i. It is currently confined to including leaf deformation, leaf inhibition, the island of Hawai‘i; noxious weed regula- witch’s brooms, fasciations, and leafy galls tions may have helped prevent its establish- ( Vereecke et al. 2000). ment on other Hawaiian islands. High rates of When fasciation is not caused by bacteria, extreme fasciation in V. thapsus (up to 35%) it often has a genetic basis. Several true- were first reported in Hawai‘i by Juvik and breeding homozygous races of fasciated plants J­uvik (1992). Those authors observed fascia- were reported by White (1943, 1948), includ- tion in surveys made during the early 1980s, ing in garden peas (Pisum sativum umbella­ and fasciation remained common during the tum), cockscomb (Celosia cristata), tobacco course of this study (Figure 1), which docu- ( Nicotiana tabacum), common balsam (Impa­ ments its persistence in Hawai‘i for at least tiens balsamina), annual phlox (Phlox drum­ 30 yr. mondii ), and Japanese morning glory (Pharbi­ Fasciation in V. thapsus has not been re- tis nil ). Several studies have shown that ported in its natural range in Europe, nor is it fasciation has a monogenic pattern of Mende- frequent in its introduced range in North lian inheritance (Dwivedi and Singh 1990, America (Reinartz 1984, Juvik and Juvik 1992, Tyagi and Gupta 1991, Knights 1993, Gold- Parker et al. 2003). No fasciation has been man 1998), and a few have identified the spe- documented in introduced populations on the cific gene responsible ( Wyongyai et al. 1984, island of La Réunion, which is climatically Laufs et al. 1998, Tang and Knap 1998). In similar to Hawai‘i ( Juvik and Juvik 1992). some natural populations of Bauhinia sp. and The cause of fasciation in V. thapsus in Phorodendron sp., fasciation has become fixed Hawai‘i is not known. This study tested the ( White 1943), resulting in morphologically following three hypotheses: (1) fasciation has distinct populations, perhaps with ecological a simple genetic basis; (2) mechanical injury and evolutionary repercussions. predisposes the plants to fasciation; and (3) Compared with the genetic and patho­ fasciation is induced by pathogenic bacteria. logical basis of fasciation, environmental We also compared capsule production be- causes of fasciation are less well understood. tween fasciated and normal V. thapsus to test Damage, herbivory, and increases in available for a potential fitness difference in fasciated nutrients are a few of the environmental fac- plants. tors that have been suggested to induce fascia- tion ( White 1948). In Oenothera species, mi- croscopic injuries caused by a moth (Mompha materials and methods sp.) were regarded as the primary cause of fas- Seed Collection ciation, and an increase in nutrient availability also increased the number of fasciated plants Verbascum thapsus has a mixed mating system; (Knox 1908). In Raphanus raphanistrum, every the flowers are pollinated by insects, but de- fasciated shoot was associated with beetle lar- layed selfing occurs toward the end of the day vae burrowing through the center and then in unvisited flowers (Donnelly et al. 1998). upward to the growing point (Molliard 1900). Controlled self-pollinations were performed Nematodes were reported to induce stem fas- to obtain pure lines derived from normal and ciation in Henry’s lily (Lilium henryi ) (Stumm- fasciated plants and to facilitate genetic anal­ Tegethoff and Linskens 1985). Intentionally ysis. Thirty five each of normal and fasciated, injuring the growing tip of wereke (Ibervillea mature V. thapsus plants were identified from sonorae), crushing the young stem of pansies seven sites along an elevational gradient from (Viola tricolor), and cutting the root tip of fava 1,829 m to 2,765 m on Mauna Kea (island of Figure 1. Fasciated V. thapsus on the island of Hawai‘i: (A) pasture vegetation on Mauna Kea, ca. 2,400 m elevation, with normal inflorescence visible in the background; B( ) natural vegetation on younger lava substrate, Mauna Kea, ca. 2,100 m elevation, with endemic shrubs (Sophora chrysophylla) in background. 454 PACIFIC SCIENCE · October 2011

Hawai‘i). The plants were selected in the e­arly rosette stage. Goats were observed to nib- flowering stage. On each plant, several un- ble on V. thapsus inflorescences, although V. opened flower buds were tagged. The plants thapsus is not a preferred browse plant for un- were visited early the next morning to per- gulates. Because the degree of potential de- form controlled self-pollinations on the struction due to ungulates was uncertain, part tagged flowers that had opened. Stigmas of (25 by 25 m2) of the field garden was enclosed freshly opened flowers were inspected to con- with 1.5 m tall galvanized steel wire fence firm the absence of pollen, which is easily (hog wire) designed to exclude ungulates, and v­isible, and intrafloral self-pollination was the remainder was unfenced and exposed to e­ffected using a fine brush. Following self- natural damage by ungulates. pollination, a 2 mm mesh was used to enclose In September 2000, near the start of the the part of the inflorescence containing the rainy season, the seedlings were transplanted self-pollinated flowers to prevent subsequent to the field. Three hundred and twelve seed- cross-pollination. At least three flowers were lings (26 each of normal and fasciated mater- self-pollinated on each plant. The mesh was nal lines; 52 lines × six replicates) were trans- removed from the plants when capsules had planted in the field common garden inside the formed. Capsules were collected when they fence, and 369 seedlings (22 normal and 19 were brown and dry, just before dehiscing. If fasciated maternal lines; 41 lines × nine repli- multiple seed capsules were collected from a cates) were transplanted in the adjoining field plant, they were pooled. common garden outside the fence in a com- plete randomized design. A distance of 1 m was maintained between the seedlings in a Common Garden row, with 1.5 m between rows. The garden A greenhouse facility at 2,100 m elevation on site was relatively open, with only herbaceous the island of Hawai‘i was used to establish vegetation. To avoid unnecessary disturbance,­ seedlings for the field common garden experi- the site was weeded only to remove plants that ment. Seeds from 31 normal and 29 fasciated were overtopping the transplanted rosettes. plants provided seedlings. Initially, five to 10 Inside the fence, two of the six replicates seeds from each of the 60 maternal plants within each line were randomly assigned to a were surface-sown on potting soil (Pro-Mix, mechanical damage treatment (to simulate Premier Horticulture Ltd., Canada) in each ungulate damage). The mechanical damage of 30 dibble tubes (38 mm diameter, 203 mm treatment was applied initially in October long) in a randomized block design. Seedlings 2000 to rosettes that were 10 cm or larger and in each tube were later randomly thinned to again 1 month later. If the rosettes that were three seedlings at the cotyledon stage and designated to receive the damage treatment then to one seedling in the four-leaf stage, re- were not 10 cm in October 2000, then the sulting in 30 replicate seedlings from each of treatment was applied when they reached 10 the 60 maternal lines. An automatic misting cm in diameter. All damaged plants were system watered the seedlings daily, for 10 min damaged once again when they were first re- every 2 hr during daylight hours. Plants were corded as bolting and flowering. Mechanical periodically rotated to ensure even watering. damage was applied using pliers to crush and The site of the field common garden was remove the top 2 to 3 cm of the tip of bolting determined based on the logistics of carrying and mature plants, or in rosettes the central water to the plants during the initial stages leaves were pinched off with pliers. Plants a­fter transplanting. The site also needed to be outside the fence were exposed to natural free from disturbance by military training ac- damage, especially from large ungulates. Nat- tivities and hunting. Ungulates including cat- ural damage on the plants both inside and tle, goats, and pigs were present in the area, outside the fence was recorded on all moni- and some V. thapsus plants were observed with toring dates throughout the 3-yr duration of broken inflorescences, torn outer leaves, or the experiment. The presence or absence of damaged central leaves among plants in the fasciation was noted on each plant throughout Fasciation in Verbascum thapsus · Ansari and Daehler 455 the experiment, and a plant was considered Field Survey and Inoculations fasciated irrespective of the degree of fascia- tion. Although fasciation can affect rosettes in A field survey was conducted across four sites ranging from 1,900 m to 2,440 m elevation on V. thapsus, fasciation is most often not appar- ent until the bolting and inflorescence stage; Mauna Kea to compare the number of cap- hence in this study, fasciation rates are re­ sules produced in normal versus fasciated V. ported only among plants that survived to the thapsus plants. Fifteen each of normal and fas- bolting stage or beyond. ciated plants were randomly selected at each site. To minimize counting errors, a blue per- manent marker was used to mark each capsule Bacterial Isolations on the plant during counting. Ten each of normal and fasciated plants were In February 2003, a field inoculation ex- sampled from one site on Mauna Kea at an periment was conducted at 2,485 m on Mauna elevation of 2,485 m. Fresh tissue samples Kea at a site that had a high frequency of fas- were haphazardly collected in pairs consisting ciation. Fasciated tissue samples from 10 fas- of normal and fasciated plants. First a mature ciated plants were collected and manually normal plant was selected followed by the mashed together with sterile distilled water closest fasciated individual. The top 15 cm of using a pestle and mortar in the field. The re- each inflorescence was cut and collected in a sultant thick slurry was used to inoculate 10 sterile plastic bag with minimal handling. All normal plants in each of three life stages: the bags were stored in a cooler and brought r­osettes, bolting, and flowering. Inoculated to the laboratory the following day. Pieces of plants were paired with controls of similar ap- pearance and size. For rosettes, the leaves in V. thapsus tissue from these collected samples were immersed in 0.6% hypochlorite solution the center were pinched off using pliers and for 10 min for surface sterilization and then the slurry from fasciated plants was applied rinsed with sterile distilled water. Two sets of using a brush. For bolting and flowering indi- bacterial isolations were made using these viduals, a superficial cut was made on the side i­nflorescence samples. For the first set, four of the stalk using the tip of the pliers, and the small (5 mm) pieces from each inflorescence top 3 to 4 cm was also mildly crushed with the were placed in petri dishes containing nutri- pliers. The top 5 to 10 cm of the stalk includ- ent agar medium with 0.5% yeast extract. In ing the wound was coated with the slurry. All the second set, small pieces of superficial tis- control plants were wounded in a similar sue from these plants were dissected and com- manner but brushed with sterile distilled minuted in sterile water to prepare a slurry, w­ater. After 5 months the plants were scored which was vigorously shaken and then used to for morphological abnormalities and fascia- streak nutrient agar plates. All the inoculated tion. plates were incubated at 25°C. After 7 days the plates were scored for the presence and Statistical Analyses physical appearance of bacterial colonies. Rhodococcus fascians, the recognized causative Common garden results were analyzed by lo- agent of pathogen-induced fasciation, appears gistic regression (Systat 11, Systat Software as orange, punctiform colonies on nutrient Inc., Chicago, Illinois), in which morphology agar plates incubated at 25°C for 7 days of the garden plants (fasciated or normal) was (O­dura 1975). In this study, the bacteria were the dependent variable and parental plant not identified to species, but the bacterial morphology, artificial damage, and natural ­colonies were screened for the color and damage were independent categorical vari- ­morphology of R. fascians. A third set of plate ables. Rates of bacterial isolation from fasci- inoculations was also made using slurry pre- ated and normal plants were compared using pared from fasciated and normal plants col- chi-square exact tests (StatExact 4, Cytel Soft- lected at an elevation of 1,800 m on Mauna ware Inc., Cambridge, Massachusetts). Out- Kea. comes of the field inoculation experiment 456 PACIFIC SCIENCE · October 2011 were also analyzed by chi-square exact tests. 95% confidence interval of the odds ratio Capsule numbers were compared between from logistic regression indicates that dam- normal and fasciated plants across sites using aged plants were between 1.3 and 32 times analysis of variance (ANOVA) followed by less likely to become fasciated. Artificial dam- post hoc contrasts. age did not interact with parental morphology (Table 1), indicating that the inhibitory ef- fects of artificial damage on fasciation did not results depend on morphology of the parental plant. In total, 13% of plants were naturally Common Garden d­amaged. Damage ranged from apparently A total of 95 plants survived and reached at chewed leaves to a broken stalk. Natural dam- least the bolting stage by the end of the ex- age was not statistically associated with fascia- periment. This sample size was inadequate for tion in the logistic regression (P = .23, Table testing between-line variation; therefore, the 1); among the seven plants that had natural influence of parental morphology (fasciated damage without artificial damage, none de- or normal) on morphology of progeny was veloped fasciation (Figure 2), whereas the tested by pooling all lines of fasciated or overall fasciation rate for undamaged plants ­normal plants. Survival rate did not differ was 48%. A chi-square exact test reveals that by parental plant morphology (χ2 = 0.20, these rates are statistically different (χ2 = 5.84, P = .33). df = 1, P = .035 without midpoint correction, Overall, 31% of plants developed fascia- P = .027 with midpoint correction) and con- tion. There was no statistical difference in the sistent with the inhibitory effects observed in fasciation rate inside and outside the fence artificially damaged plants. Logistic regres- (t-ratio = -1.56, P = .12); therefore, all plants sion combining artificial and natural damage were combined for logistic regression (Table identifies a significantly lower probability of 1). Morphology of the parental plant was not fasciation associated with damage (P = .009) statistically linked to morphology of progeny but no effect of parental morphology (P = .18) in the common garden (P = .26, Table 1, Fig- and no interaction between parental mor- ure 2). Although the regression coefficient for phology and damage (P = .86). parental morphology was positive, it had a large standard error (Table 1). Bacteria Isolations Thirty-seven percent of the analyzed plants were in the artificial damage treatment. None of the bacterial colonies isolated from Contrary to our initial prediction, artificial fasciated or normal V. thapsus resembled the damage was associated with lower rates of colony phenotype of Rhodococcus fascians. Iso- f­asciation (P = .023, Table 1, Figure 2). The lations made directly from dissected plant

TABLE 1 Logistic Regression Testing Influence of Parental Morphology (Fasciated or Normal), Natural Damage (Present or Absent), and Artificial Damage Treatment ( Yes or No) on Morphology of Plants (Fasciated or Normal) in a Field Common Garden

Factor Estimate S.E. t-Ratio P-Value

Parental morphology 0.615 0.550 1.12 .26 Natural damage -1.024 0.856 -1.20 .23 Artificial damage -1.857 0.820 -2.27 .023 Artificial damage × parental morphology 0.211 1.206 0.18 .86 Constant -0.452 0.356 -1.27 .20

Note: Log Likelihood of the model is -51.21 versus -58.45 for the constant only model (χ2 = 14.48, df = 3, P = .002). Fasciation in Verbascum thapsus · Ansari and Daehler 457

Figure 2. Fasciation rates in undamaged plants versus plants exposed to one type of damage. ns, no statistical differ- ence between progeny of fasciated and normal mothers; *, significant difference between damaged and undamaged plants (P < .05).

t­issue from 2,485 m elevation on Mauna Kea Field Survey and Inoculations showed the presence of only white and yellow bacterial colonies, and there were no signifi- Across all four sites, fasciated plants averaged cant differences in the frequency of occur- more than twice the number of capsules and rence of these colonies between normal and seeds as normal plants, and within each site, fasciated plants (Figure 3). Isolations derived fasciated plants always averaged more cap- from the slurry revealed five different colony sules (Figure 4). Among the field-inoculated morphologies, including an orange morph, plants, none developed fasciation irrespective but rates of isolation did not differ between of the treatment (slurry or water control) and fasciated and normal plants (Figure 3). The growth stage (Table 2). The damaged stem same five colony morphs were also isolated and leaf tissue immediately adjacent to the from plants collected at the 1,800 m site, but site of wounding inoculation sometimes ap- again there were no differences in isolation peared crinkled, but there were no differences rates between normal and fasciated plants between treatment and controls in the rate or (data not shown). For plants collected at 1,800 extent of localized leaf deformation (Table 2). m, orange colonies were cultured from slur- Irrespective of treatment, there was 100% ries derived from 20% of the fasciated plants survival among rosettes and there were no and 25% of the normal plants (no statistical significant differences between the slurry difference), and the orange colonies were not treatment and controls in the survival of punctiform, as would be expected for R. fas­ plants at the bolting and flowering stage cians. (T­able 2). 458 PACIFIC SCIENCE · October 2011

Figure 3. Frequency of occurrence of different-colored colonies from normal and fasciated plants collected at 2,485 m on Mauna Kea (T, cultured directly from tissue; L, cultured from the tissue slurry). Rhodococcus fascians, a pathogenic bacteria that can cause fasciation, is expected to form orange colonies. There were no significant differ- ences in isolation rates between normal and fasciated plants.

discussion ents would have produced progeny that were 100% fasciated (if recessive) or either 75% or Does Propensity toward Fasciation Have a 100% fasciated (if dominant), and fasciation Genetic Basis? should have occurred at a much lower rate (0% if fasciation was dominant, and 25% if Although significantly higher capsule and fasciation was recessive) in progeny of normal seed production in fasciated V. thapsus sug- parents. In this study, only about half of un- gests that fasciation could be ecologically im- damaged progeny of fasciated parents were portant and help promote the success of this fasciated, and this rate was similar among invasive weed, fasciation does not appear to progeny derived from selfed normal plants. be heritable. If fasciation in V. thapsus were These results indicate that, unlike in some controlled by a single recessive or dominant a­gricultural plants such as soybean (Glycine gene, then the self-pollinated fasciated par- max), mung bean (Vigna radiata), lentil (Lens Fasciation in Verbascum thapsus · Ansari and Daehler 459

Figure 4. Comparison of capsule production in normal versus fasciated V. thapsus across four sites ranging from 1,900 m to 2,440 m on the island of Hawai‘i. Fifteen fasciated and 15 normal plants were measured at each site. All differences are statistically significant P( < .05, ANOVA followed by post hoc contrasts). Error bars indicate 1 SE.

TABLE 2 Survival, Development of Deformed Leaves, and Fasciation among Field Plants Inoculated with a Slurry Extracted from Fasciated Plant Tissues

% of Plants with Deformed Leaves at Point % of Plants That Survived of Injury % Fasciated

Stage Control Inoculated Control Inoculated Control Inoculated

Rosette 100 100 60 60 0 0 Bolting 50 70 ns 10 10 0 0 Flowering 20 50 ns 0 0 0 0

Note: ns, no statistical difference, P > .05.

culinaris), chickpea (Cicer arietinum), and red a simple biallelic pattern of inheritance. Fur- beet (Beta vulgaris) (Albertsen et al. 1983, thermore, the lack of difference in fasciation Dwivedi and Singh 1990, Tyagi and Gupta rate between progeny of fasciated and normal 1991, Knights 1993, Goldman 1998), the fas- plants provides no evidence at all for herita- ciated phenotype in V. thapsus does not follow bility. 460 PACIFIC SCIENCE · October 2011

Although our results indicate that fascia- Because the frequency of fasciation was tion in V. thapsus in Hawai‘i probably does not initially reported to increase with elevation in have a simple genetic basis, a multiallelic ge- Hawai‘i, Juvik and Juvik (1992) proposed that netic basis for fasciation has been reported in diurnal frost damage at high elevation might several plant species (Gottschalk and Hussein cause fasciation. However, a high frequency 1977, Eenink and Garretsen 1980, Albertsen of fasciation (50%) in populations located at et al. 1983), and fasciation may develop as a persistently warm sites near sea level (50 m) result of gene by environment interaction has subsequently been observed (S.A., unpubl. (Robinson 1987). It was beyond the scope of data), so factors other than frost must be this study to investigate the possibility of such ­involved. complex scenarios in V. thapsus. Is Fasciation in V. thapsus Induced by Bacteria? Does Damage Predispose Plants to Fasciation? In this study, none of the isolated bacterial If physical damage caused fasciation, then colonies had cultural characteristics similar to damaged plants should have had higher rates those of Rhodococcus fascians. However, because­ of fasciation, but this study found the oppo- it is sometimes difficult to isolate bacteria site. Feral goats were observed around the ex- from infected tissue, the isolation results perimental garden several times during the alone provide rather weak evidence against course of the study (S.A., pers. obs.). Never- the bacteria hypothesis. In contrast, the fail- theless, natural damage rates were low (around ure of all field-inoculated plants to develop 10%), and this damage did not promote fas- fasciation provides stronger evidence that fas- ciation. ciation in V. thapsus is not caused by a patho- In a study in North America ( Naber and gen. The bacteria can be spread easily from Aarssen 1998), no fasciations were reported in infected plants, seeds, and soil on hands or V. thapsus following experimental removal of tools (Putnam and Miller 2007). Two forms the apical meristems. In Lilium henryi, wound- of R. fascians bacteria are known. One of them ing of healthy plants in various stages of de- is epiphytic, only colonizing the plant surface, velopment by either damaging the growing affecting shoot meristem formation; the other point, decapitation, or defoliation also did not form is endophytic, and this form is required induce fasciation (Stumm-Tegethoff and Lin- for symptom maintenance (Goethals et al. skens 1985). The evidence that mechanical 2001, Vereecke et al. 2002). Plants infected damage causes fasciation is mainly anecdotal with R. fascians are often substantially debili- ( White 1943, 1948). tated by the pathogen. In Chrysanthemum Insects have been reported to cause fascia- maximum and Pisum sativum, plants infected tion in some plants (Knox 1908, Binggelli with R. fascians show hypertrophied shoots, 1990). We conducted field surveys of V. and they are stunted and appear weak com- t­hapsus on the island of Hawai‘i from sea level pared with noninfected plants (Odura 1975). to 3,000 m, but little or no evidence of insect Infection at the seedling stage may result in damage was observed. A few plants in a popu- complete inhibition of seedling growth, in- lation at Pu‘uwa‘awa‘a Natural Area Reserve cluding inhibition of root growth and thick- (545 m) showed evidence of seed predation; ening of the hypocotyls ( Vereecke et al. however, no insects were observed on those 2000). The secondary symptoms accompa­ plants. The seed predation occurred after nying hypertrophied shoot formation are capsules had matured and therefore cannot stunted growth, reduction in flower size, and be linked to fasciation. Dissection of fasci­ reduced yield (Odura 1975). Contrary to ated plants also did not reveal any activity of these debilitating symptoms, in V. thapsus the sucking or internally feeding insects or nema- fasciated plants in the field appear to be more todes. Hence it is unlikely that fasciation in V. robust, taller, and larger than normal plants, t­hapsus is caused or mediated by insects or and they produce more capsules. Greater n­ematodes. yield in fasciated plants has not been reported Fasciation in Verbascum thapsus · Ansari and Daehler 461 in bacteria-associated fasciation, but it is ated V. thapsus develop a degree of stem sometimes seen when fasciation is caused by woodiness that is not seen in normal plants other factors (e.g., Tyagi and Gupta 1991). ( Juvik and Juvik 1992). Woodiness may be a Potential fitness trade-offs associated with by-product of longer life span, which may higher yield in nonbacterial fasciations have benefit island colonizers by allowing more not been investigated. opportunities for cross-pollination and escape from detrimental inbreeding (Bo¨hle et al. 1996), but a reproductive advantage a­ssociated Potential Environmental Factors with greater investment in reproduction (al- Detailed breeding experiments by DeVries lowed for by woody stalks) would also greatly (1908) and White (1948) showed that in some benefit a colonist during island invasion. This plants fasciation was expressed only under is especially true when risks of juvenile mor- certain environmental conditions. Muncie tality are relatively low, as was true in high- and Patel (1930) suggested that in sweet peas, elevation populations in Hawai‘i where V. fasciation was induced by overcrowding of thapsus is unusually long-lived (Ansari and their roots. In a separate experiment they fur- Daehler 2010). Although we were unable to ther showed that wounding by puncturing the demonstrate a simple genetic basis for fascia- base of stems with a sterile needle, combined tion in invading V. thapsus, our statistical with crowding, led to fasciation rates of 28% p­ower to detect small degrees of heritability in comparison to 17% in undamaged controls was low, and we did not test for gene-by-­ (χ2 = 5.90, df = 1, P = .015). Muncie and Patel environment interactions. If fasciation does (1930) concluded that abnormal growing con- have an element of heritability, then we ex- ditions induce fasciation in sweet peas. pect to see a long-term pattern of increasing Contrary to Muncie and Patel’s (1930) gigantism and adaptive radiation in V. thapsus findings on peas, increased spacing and favor- as it invades Hawai‘i, as has been observed in able growing season increased fasciation rates numerous other island colonizations by ro- in flax Linum( usitatissimum) (Shibuya 1939). sette plants over evolutionary time (Carlquist Many early researchers speculated that high 1974). levels of nutrition induce fasciation, but none provided any experimental evidence for this (Hus 1908, Knox 1908, White 1943, 1948). acknowledgments Several other explanations such as changing We are grateful to the Environmental Divi- the photoperiod and increased nutrition with sion of Pöhakuloa Training Area for provid- water shortage have been speculated to in- ing greenhouse space and access to the field duce fasciation (Binggelli 1990). Zinc defi- sites. We also thank Monica Mejia Chang, ciency has also been reported to induce fascia- Dave Faucet, Teresa Restom-Gaskill, and tion in plants (Rance et al. 1982) while also Shelly Lammers for helping set up the field causing a characteristic suite of debilitating experiment. The manuscript was improved by symptoms that are not observed in V. thapsus. comments from Rob Cowie and four anony- Coarse-scale environmental factors in mous reviewers. Hawai‘i, such as a year-round growing season and high ultraviolet radiation, may promote fasciation, with susceptibility modified by Literature Cited l­ocal factors such as disturbance, exposure, or perhaps genetics. A localized contributing Albertsen, M. C., T. M. Curry, R. G. Palmer, factor is supported by the fact that normal and and C. E. Lamotte. 1983. Genetics and fasciated plants commonly occur in close comparative growth morphology of fascia- proximity (<1 m) in the field. The ecological tion in soybeans (Glycine max [L] Merr). or evolutionary significance of fasciation in Bot. Gaz. 144:263 – 275. natural populations has rarely been examined. Ansari, S., and C. C. Daehler. 2010. Life his- In addition to producing more capsules, fasci- tory variation in a temperate plant invader, 462 PACIFIC SCIENCE · October 2011

Verbascum thapsus, along a tropical eleva- C. W. Smith, and J. T. Tunison, eds. Alien tional gradient in Hawai‘i . Biol. Invasions plant invasions in native ecosystems of 12:4033 – 4047. Hawai‘i: Management and research. Uni- Binggelli, P. 1990. Occurrence and causes of versity of Hawai‘i Press, Honolulu. fasciation. Cecidology 5:57 – 62. Knights, E. J. 1993. Fasciation in chickpea: Bo¨hle, U. R., H. H. Hilger, and W. F. Mar- Genetics and evaluation. Euphytica tin. 1996. Island colonization and evolu- 69:163 – 166. tion of the insular woody habit in Echium Knox, A. D. 1908. The induction, develop- L. (Boraginaceae). Proc. Natl. Acad. Sci. ment and heritability of fasciation. Carne- U.S.A. 93:11740 – 11745. gie Inst. Washington Publ. 98. Carlquist, S. 1974. Island biology. Columbia Lamotte, C. E., T. M. Curry, R. G. Palmer, University Press, New York. and M. C. Albertsen. 1988. Developmental DeVries, H. 1908. The mutation theory: Ex- anatomy and morphology of fasciation in periments and observations on the origin the soybean (Glycine max). Bot. Gaz. of species in the vegetable kingdom. Open 149:389 – 407. Court Publishing Company, Chicago. Laufs, P., J. Dockx, J. Korneberger, and J. Donnelly, S. E., C. J. Lortie, and L. W. Aars- Traas. 1998. Mgoun1 and mgoun2: Two sen. 1998. Pollination in Verbascum thapsus genes required for primordium initiation (Scrophulariaceae): The advantage of be- at the shoot apical and floral meristems in ing tall. Am. J. Bot. 85:1618 – 1625. Arabidopsis thaliana. Development (Camb.) Dwivedi, S., and D. P. Singh. 1990. Inheri- 125:1253 – 1260. tance of fasciation in mungbean ( Vigna ra­ Molliard, M. 1900. Cas de virescence et de diata (L). Wilczek). Indian J. Genet. 50:81 – fasciation d’origine parasitaire. Rev. Gen. 82. Bot. 12:323 – 328. Eenink, A. H., and F. Garretsen. 1980. Re- Muncie, J. H., and M. K. Patel. 1930. Fas­ search on the inheritance of fasciation in ciation of sweet peas. Am. J. Bot. 17:218 – lettuce (Lactuca sativa L.). Euphytica 230. 29:653 – 660. Naber, A. C., and L. W. Aarssen. 1998. Ef- Goethals, K., D. Vereecke, M. Jaziri, M. V. fects of shoot apex removal and fruit Montagu, and M. Holsters. 2001. Leafy h­erbivory on branching, biomass and re- gall formation by Rhodococcus fascians. Annu. production in Verbascum thapsus (Scrophu- Rev. Phytopathol. 39:27 – 52. lariaceae). Am. Midl. Nat. 140:42 – 54. Goldman, I. L. 1998. Inheritance of ffs, a gene Odura, K. A. 1975. Factors affecting epidemi- conditioning fasciated flower stem in red ology of bacterial fasciation of Chrysanthe­ beet. J. Am. Soc. Hortic. Sci. 123:632 – 634. mum maximum. Phytopathology 65:719 – Gorter, C. J. 1965. Origin of fasciation. En- 721. cycl. Plant Physiol. 15:330 – 351. Parker, I. M., J. Rodriguez, and M. E. Loik. Gottschalk, W., and H. A. S. Hussein. 1977. 2003. An evolutionary approach to under- Fasciated peas: Unusual mutants for breed- standing the biology of invasions: Local ing and research. J. Nucl. Agric. Biol. adaptation and general purpose genotypes 6:27 – 33. in the weed Verbascum thapsus. Conserv. Hus, H. 1908. Fasciations of known causa- Biol. 17:59 – 72. tion. Am. Nat. 42:81 – 97. Putnam, M. L., and M. L. Miller. 2007. Rho­ Jones, D. F. 1935. The similarity between fas- dococcus fascians in herbaceous perennials. ciations in plants and tumors in animals Plant Dis. 91:1064 – 1076. and their genetic basis. Science ( Washing- Rance, S. J., D. M. Cameron, and E. R. Wil- ton, D.C.) 81:75 – 76. liams. 1982. Correction of crown disorders Juvik, J. O., and S. P. Juvik. 1992. Mullein of Pinus caribaea var. hondurensis by applica- (Verbascum thapsus): The spread and adap- tion of zinc. Plant Soil 65:293 – 296. tation of a temperate weed in the montane Reinartz, J. A. 1984. Life history variation of tropics. Pages 254 – 270 in C. P. Stone, common mullein ( Verbascum thapsus). I. Fasciation in Verbascum thapsus · Ansari and Daehler 463

Latitudinal differences in population dy- Vereecke, D., K. Cornelis, W. Temmerman, namics and timing of reproduction. J. Ecol. M. Jaziri, M. V. Montagu, M. Holsters, 72:897 – 912. and K. Goethals. 2002. Chromosomal Robinson, R. W. 1987. Inheritance of oppo- l­ocus that affects pathogenicity of Rhodo­ site leaf arrangement in Cucumis sativus L. coccus fascians. J. Bacteriol. 184:1112 – 1120. Department of Horticultural Sciences, Vereecke, D., B. Sylvia, C. Simon-Mateo, D. New York State Agricultural Experiment Inze, M. V. Montagu, K. Goethals, and M. Station, Cucurbit Genetics. Cooperative Jaziri. 2000. The Rhodococcus fascians – plant Report 10. interaction: Morphological traits and bio- Shibuya, T. 1939. The occurrence of fascia- technological applications. Planta (Berl.) tion in flax stem in relation to the environ- 210:241 – 251. ment. J. Soc. Trop. Agric., Taiwan 11:227 – White, O. E. 1943. The biology of fasciation 236. and its relation to abnormal growth. J. Stumm-Tegethoff, B. F. A., and H. F. Lin- Hered. 36:11 – 22. skens. 1985. Stem fasciation in Lilium hen­ ——— . 1948. Fasciation. Bot. Rev. 14:319 – ryi caused by nematodes. Acta Bot. Neerl. 358. 34:83 – 93. Wordsell, W. C. 1915. The principles of Tang, Y., and H. T. Knap. 1998. Fasciation plant teratology. Vol. 1. Ray Society, Lon- mutation enhances meristematic activity don. and alters pattern formation in soybean. ——— . 1916. The principles of plant teratol- Int. J. Plant Sci. 159:249 – 260. ogy. Vol. 2. Ray Society, London. Tyagi, B. S., and P. K. Gupta. 1991. Induced Wyongyai, W., F. Tadahiko, and S. Matsu- mutations for fasciation in lentil (Lens culi­ moto. 1984. Morphological characteristics naris Med.). Indian J. Genet. 51:326 – and growth habit of fasciated soybean. Jpn. 331. J. Crop Sci. 53:371 – 378.