In: Van Driesche, R., et al., 2002, Biological Control of Invasive Plants in the Eastern United States, USDA Forest Service Publication FHTET-2002-04, 413 p.
Nature of Damage
Economic damage. Lost pasturage in many states, especially states with hilly terrain and pastures on steep slopes, has resulted in significant reduction in potential beef production. This thorned bramble now infests more than 45 million acres throughout the eastern United States (Underwood et al., 1996). Chalamira and Lawrence (1984) reported that multiflora rose was the highest priority agricultural problem in West Virginia. Experimental multiflora control programs in West Virginia during 1980 and 1981 indicated that more than 36,500 hectares were heavily infested and that a ten-year eradication program using herbicides would cost more than $40 million (Williams and Hacker, 1982). Similar burdens and costs were reported from neighboring states; to date, multiflora has been declared a noxious weed in Illinois, Iowa, Kansas, Maryland, Missouri, Ohio, Pennsylvania, Virginia, Wisconsin, and West Virginia (Amrine and Stasny, 1993).
Ecological damage. Multiflora rose has invaded a large number of habitats, from hillside pastures, fence rows, right-of-ways, and roadsides to forest edges and the margins of swamps and marshes (Scott, 1965). A single, vigorous, mature plant can produce up to half a million achenes (seeds) annually. Where plants have become well established, a huge seed bank develops that can continue to produce seedlings for at least twenty years after removal of mature plants. Severe multiflora rose infestations have lowered land values for agriculture, forestry, and recreation (Underwood et al., 1996). Since the 1960s, multiflora rose has become one of the most noxious weeds in the eastern United States. It is especially troublesome in regions with steep slopes, which prevent access by tractors or mowers for cutting this weed. Multiflora rose forms dense, impenetrable thickets in many regions of the eastern United States. At least ten states have passed noxious weed laws against it, and it is illegal to plant it in many areas (Amrine and Stasny, 1993; Fawcett, 1980; Klimstra, 1956; Kriebel, 1987; Williams and Hacker, 1982; Underwood et al., 1996). Many state publications and web sites list cultural and chemical methods for controlling multiflora rose, but biological control has been a neglected management option (Lingenfelter and Curran, 1995; Underwood and Stroube, 1986; Underwood et al, 1996).
Extent of losses. In West Virginia, projected costs to farmers for controlling multiflora rose from 1981 to 1982 exceeded $40 million (Williams and Hacker, 1982); at today’s rates, this cost would exceed $48 million. Similar costs accrue to most eastern states and control costs continue to rise as this noxious weed continues to spread.
In eastern North America, multiflora rose is abundant from the Great Plains (where the species has been planted as wind breaks) to the east coast. It occurs from northern Texas, Arkansas, Mississippi, Alabama, and Georgia in the south, north to the New England coast, central New York, southern Michigan, Wisconsin, and Minnesota. It occurs only as plantings south of central Georgia, probably because of the lack of cold temperatures needed to stimulate seed germination. The plant’s northern distribution is limited by its sensitivity to severe cold temperatures.
Background Information On The Pest Plant
Multiflora rose is in the Subfamily Rosoïdeae, Tribe Roseae. Rosa is the only known genus in the tribe. The most closely related plants are members of the tribes Potentilleae (Sections Rubinae, Potentillinae, Dryadinae), Cercocarpaceae, Ulmariëae, and Sanguisorbeae. The most common genera that would be most closely related to Rosa are Rubus, Potentilla, Fragaria, Geum, Dryas, Adenostema, Purshia, Cercocarpus, Alchemilla, Agrimonia, and Poterium.
Multiflora rose was first described from Japan. It is a stout, thorny, diffusely branched, perennial shrub with numerous arching stems (canes) arising from the crown; plants may reach 3 m height and 6.5 m diameter. Twigs are reddish to green, 1.5 cm in diameter and armed with numerous, recurved thorns; thornless clones occur sparsely throughout the eastern United States. Leaves are odd-pinnately compound, 8 to 11 cm long, divided into five to 11 sharply toothed, ovate to oblong leaflets. Basal petioles are 1.0 to 1.3 cm long and have finely dissected, usually glandular stipules. Large clusters of showy, fragrant, white to pink 2.5 cm flowers occur in dense to sparse panicles that appear in late May or June. Panicles contain six to 100 (average of 63) hypanthia or hips that are glabrous to pubescent, develop during the summer, and become bright red by mid-September; hips contain an average of seven (one to 21) achenes. Hypanthia become soft after frost and eventually become leathery, remaining on the plant through the winter. Achenes are yellowish to tan, somewhat irregular in shape, about 2 to 4 mm long by 2 mm wide, and enclosed in sharp spicules. Winter-feeding birds often consume fruits by January. Seeds are attacked by the rose seed chalcid, Megastigmus aculeatus var. nigroflavus Hoffmeyer (Hymenoptera: Torymidae) in many areas (see below).
Each cane on a large plant may contain 40 to 50 pannicles. Each pannicle can contain as many as 100 hypanthia or hips (average of about 50) and each hip, an average of seven seeds (range of one to 22). Thus each large cane can potentially produce up to 17,500 seeds. Seeds remain viable for a number of years (Evans, 1983; Underwood et al, 1996). We have found as many as 90% of the seed to be viable, in the absence of drought, stress, and seed chalcids. The abundant floral production of this plant may be the result of the plant’s evolution in the presence of its seed predator, the multiflora rose seed chalcid, Megastigmus aculeatus var. nigroflavus Hoffmeyer (Hymenoptera: Torymidae). In Asia, the chalcid may infest 95% of the achenes or seeds (Weiss, 1917). The chalcid reproduces by parthenogenesis (female:male ratio is 200:1), possibly a mechanism to match the huge resource (Shaffer, 1987). Multiflora rose is moderately winter-hardy, tolerant to many North American insects and diseases, and grows rapidly into dense thorny thickets favorable for many species of wildlife. Its abundant fruits are food to deer and birds. The flowers produce large amounts of golden, sweet-tasting pollen that can be harvested by fitting bee hives with pollen traps (Amrine unpublished). The plant has a vigorous root system capable of checking erosion, and if carefully planted and mechanically trimmed, multiflora rose can make living fences capable of restraining some species of livestock (Dugan, 1960). It is still planted as a living fence in southern Delaware to separate herds of horses. Because of these traits, multiflora rose was widely planted throughout the eastern United States from the 1930s until the 1960s as living fences, for erosion control, and to protect and feed native wildlife. In West Virginia, more than 14 million plants were planted in the 1940s to 1960s (Dugan, 1960), and in North Carolina, more than 20 million were planted (Nalepa, 1989). Only a few states (e.g., Kentucky) refused to promote this plant. Consequently, many areas of Kentucky are relatively free of the weed. Since the plant was distributed as rooted cuttings and not from seed, no seed chalcids were distributed.
Some early experiments were conducted to show that spread of multiflora seed by birds was minimal. However, the birds chosen were chickens, doves, pigeons, turkeys, and their relatives—all of which have gizzards containing stones that grind seeds. Songbirds were not tested as potential seed dispersers. Robins, mockingbirds, starlings, red-winged blackbirds, and other species feed heavily on multiflora rose hips in fall and winter, and, because of the numerous spicules in each hip, seeds pass rapidly through their digestive tracts and remain intact. Passage of seeds through digestive tracts of songbirds increases the germination rate, while bird feces provides fertilizer to seedlings (Lincoln, 1978; Scott, 1965).
Analysis of Related Plants in the Eastern United States
According to the Synonymized Checklist of the Vascular Flora of the United States, Puerto Rico, and the Virgin Islands and the Texas A&M University Bioinformatics Working Group on the Rosaceae (part of BONAP, theBiota of North America Program), there are 82 species or subspecies of roses that are either native to the eastern United States, have escaped from cultivation, or are grown in gardens. In addition, there are some 8,000 registered cultivars of roses, worldwide, with many new ones registered annually. The following is a list of roses occurring in this region:
None of the above roses are known to be rare or endangered; many have ranges restricted to mountains, to the northern regions, to marshes, to deserts or to the west. Several introduced roses have become noxious weeds. The Macartney rose (Rosa bracteata Wendland) was imported into Texas from eastern Asia and has become a noxious weed along the Gulf Coast, infesting more than 500,000 acres of productive grasslands in 40 southeastern Texas counties (Scott, 1965). The Cherokee rose (Rosa laevigata Michaux), another introduced plant from China (however, the State Flower of Georgia), became a severe weed in the Black Belt region (several counties characterized by rich, dark soil) in central Alabama. Land covered by the weed in nine counties could have produced 1.5 million pounds of beef annually, if in productive pasture (Scott, 1965). Rosa canina L., a native of Europe and west Asia, has been introduced into most of the eastern United States; it is widely dispersed and occasionally found to be abundant, but has shown no weedy propensity in the east. The large hips of R. canina are valued by natural food enthusiasts. Rosa eglanteria L., another native of Europe has become widely dispersed in the United States; it is very weedy in New Zealand. Rosa rugosa Thunb., another non-indigenous rose from China, has been introduced throughout the eastern US; this species is commonly cultivated as an ornamental species rose; it has escaped and become abundant along the northeast coast, especially in Long Island, New York (Amrine, pers. observ., 2001) and Maine (Peck, 2001).
Common native roses in eastern North America include the prickly rose, Rosa acicularis Lindl. (in mountains and northern regions), the smooth rose (Rosa blanda Aiton), the prairie rose (R. setigera), the swamp rose (Rosa palustris Marsh), the Virginia rose (R. virginiana) and the pasture rose (Rosa carolina L.). None of these native roses have become weeds except in rare instances. Abundant natural controls and seed predators probably prevent them from becoming weeds. The introduced roses, Rosa eglanteria L., R. canina, and R. rugosa, all ornamental species, have escaped and are commonly found in many areas, but have not been observed to be significant weeds.
Only the genus Rosa occurs in the tribe Roseae. The most closely related plants are members of the tribes Potentilleae (Sections Rubinae, Potentillinae, Dryadinae), Cercocarpaceae, Ulmariëae, and Sanguisorbeae. Thus, genera most closely related to Rosa are Rubus (blackberries, raspberries, brambles; probably more than 75 species occur in eastern North America), Potentilla (cinquefoil, 15 species), Fragaria (strawberries, five species), Geum (avens, 10 species), Dryas (mountain avens, two species in western North America), Adenostema (chamise, ribbonwood; two species in California), Purshia (antelope bush, two species in western North America), Cercocarpus (mountain mahogany, five species in western North America), Alchemilla (lady’s mantle, parsley-piert; three to four naturalized species in eastern North America), Agrimonia (beggar-ticks, about 10 species in eastern North America), Poterium (burnet, one species naturalized in eastern North America) and Filipendula (meadowsweet, two or three native or naturalized species in eastern North America). Some of the Rubus are occasionally attacked by the rose stem girdler, Agrilus aurichalceus aurichalceus Redtenbacher; none of the other arthropods or diseases affecting multiflora rose, discussed herein, occur on any of these related plants.
History of Biological Control Efforts in the Eastern United States
Area of Origin of Weed
As mentioned above, R. multiflora originated in eastern Asia. It is native to Japan, Korea and northeast China and a wide variety of other deciduous-forest podzol areas of eastern Asia that are similar to those of the eastern United States (Good, 1964). It also occurs in similar areas of Europe.
Areas Surveyed for Natural Enemies and Natural Enemies Found
Hindal and Wong (1988) surveyed West Virginia for arthropods and diseases occuring on multiflora rose. They found several insects and diseases, of which the following were noted: the rose seed chalcid, Megastigmus aculeatus var. nigroflavus Hoffmeyer (Hymenoptera: Torymidae), introduced from Japan; a native raspberry cane borer, Oberea bimaculata Olivier (Coleoptera: Cerambycidae); a native tortricid hip borer, Grapolita packerdi Zeller (Lepidoptera: Tortricidae); a native powdery mildew (Sphaerotheca sp.); several native fungi that cause cankers (species of Epicoccum, Leptosphaeria, Phoma, and Phomopsis); and several introduced European stem gall forming species, from which bacteria were cultured that were similar to Agrobacterium tumefasciens (E. F. Sm. et Towns.) Conn. Of these, only the seed chalcid appeared to present any possibility of significant biological control. Mays and Kok (1988) found the seed chalcids in roses in Virginia, and Shaffer (1987) reported finding the seed chalcid in all counties of West Virginia that were surveyed as well as in Indiana, Kentucky, Maryland, Ohio, and Pennsylvania. To our knowledge, no surveys have been conducted for natural enemies of multiflora rose in eastern Asia. Consequently, surveys of natural enemies associated with this rose in its native range and compilation from the literature of its known natural enemies, both typical early steps of most plant biological control projects, have not been done.
Host Range Tests and Results
Results of host range tests for the eriophyid mite Phyllocoptes fructiphilus Keifer (vector of rose rosette disease [RRD]) and the rose rosette disease virus are given in Tables 1 through 3. Most native roses in the midatlantic region have been tested and can not be infected with RRD; all are excellent hosts for the mite. Most ornamental roses are capable of sustaining the mite and of being infected by RRD. Many cultivars are very susceptible to RRD and these are indicated in the tables in bold type. Only members of Rosa can be infected with RRD or serve as hosts for the mite. A large number of other rosaceous plants have been tested for RRD susceptibility and mite acceptance. All tests, including backgrafts to multiflora rose, have been negative. None of the other rosaceous plants support the mite. A number of grafted rosaceous plants have been grown at the West Virginia University Horticulture Farm since 1989; to date, none have shown any symptoms of RRD and backgrafts have been negative. The rose seed chalcid has only been found in seed from multiflora rose; apparently differences in the hips and/or times of flowering prevent the chalcid from successfully developing in seeds of other roses.
Citations: 1)Allington et al., 1968, 2)Crowe, 1983, 3)Doudrick and Millikan, 1983 , 4)Gergerich and Kim, 1983, 5)Keifer, 1966, 6)Thomas and Scott,1953, 7)Viehmeyer, 1961, 8)Wagnon, 1966, 9)Wagnon, 1970, 10)Amirine et al., 1995, 11)Finkes, 1991, 12)Worden, 1988, 13)Epstein and Hill, 1998, 14)Sauer, 2001, 15)Peck, 2001, 16)Epstein and Hill, 1994, 17)Epstein and Hill, 1999, 18)Peck, 2002, 19)Higgins, 2001.
Abbreviations: CL = Large-flowered climber, CM = Climbing miniature, Engl. R. = English rose, FL = Floribunda, G = Grandiflora, HT = Hybrid tea, M = Miniature, P = Pollyanna, PT = Portland rose, RB = Rambler, SH = Shrub, R = Resistant, S = Susceptible, T = Tolerant.
Results: NONE of the above plants were successfully infected with RRD; Amrine and Stasny (unpublished) showed that back grafts were negative.
0 - mites lived less than 3 days (unsuitable).
1 - mites lived for a week without laying eggs (unsuitable).
2 - mites laid eggs (suitable).
To our knowledge, no intentional releases were made of any of the insects, mites, or pathogens discussed in the following section; all are either native North American species or, as in the case of the rose stem girdler and the multiflora rose seed chalcid, were accidentally introduced. Rose rosette disease has been transmitted to target multiflora roses by grafting and by mite releases in Iowa and West Virginia (Amrine and Stasny, 1993; Epstein and Hill, 1994b, 1995b; Amrine et al., 1995; Epstein, 1995; Epstein et al., 1997). Because of the susceptibility of many ornamental roses to RRD and P. fructiphilus (Tables 1 and 3), this work has been opposed by the American Rose Society and by rosarians in general (Obrycki, 1995; Philley, 1995; Peck, 2001; Pagliai, pers. comm.). However, augmentation research has provided valuable information on the potential spread of RRD. Experimental increase of the rose seed chalcid was successful in West Virginia; infestation increased in one season from 3.2 to 77.5% (see section on multifloral rose seed chalcid under Biology and Ecology of Key Natural Enemies).
Biology and Ecology of Key Natural Enemies
Four agents have been found in the United States that show potential for biological control of multiflora rose. These are a “virus” that causes rose rosette disease, an eriophyid mite (P. fructiphilus) that transmits this virus, a seed chalcid (M. aculeatus var. nigroflavus) that lays its eggs in rose hips and whose larvae feed on immature seeds, and a stem girdler (Agrilus aurichalceus aurichalceus Redtenbacher [Coleoptera: Buprestidae]) that kills multiflora rose canes.
published a U.S. map showing RRD’s known distribution as far east as Ohio, Pennsylvania, Tennessee, and West Virginia in 1994. This native pathogen has caused a fatal epidemic in Rosa multiflora from the Great Plains as far east as Berks County, Pennsylvania and Queen Annes County, Maryland, in the Delmarva peninsula (Fig. 2) (Amrine and Stasny, 1993; Epstein and Hill, 1995a, 1999).
Rose rosette disease is a mite-transmitted, graftable “virus” that produces fragments of double-stranded RNA in rose tissue (Frist, pers. comm.; Di et al., 1990; Hill et al., 1995). Various structures found in electron microscope micrographs have been tentatively identified as the agent (Gergerich and Kim, 1983), but none have been conclusively proven to be the agent. It has not yet been taxonomically characterized (Epstein and Hill, 1999). Symptoms of RRD in multiflora rose include red, purplish or dark green veinal pigmentation (Fig. 3); production of bright red lateral shoots (Fig. 4); enlarged stems and stipules; dense, yellowish, dwarfed foliage; and premature development of lateral buds producing many
Rose rosette disease is transmitted by the eriophyid mite, P. fructiphilus (Figs. 8, 9, and 10), which develops in high numbers on shoots of RRD-infected multiflora roses and other rose species (Amrine et al., 1988). Phyllocoptes fructiphilus was first described from Rosa californica Cham. et Schlechtend. in California by Keifer (1940). Since that date, it often has been found associated with RRD in roses throughout the United States (Amrine and Stasny, 1993; Epstein and Hill, 1994b; Epstein and Hill, 1995a, 1999; Amrine et al., 1995; Amrine,1996). The mite often occurs in the absence of the virus, producing no visible symptoms on rose plants. It only develops on tender, rapidly growing tissue and is aerially disseminated (Zhao 2000). Doudrick (1984) and Doudrick et al. (1983) claimed that Phyllocoptes fructiphilus could not transmit RRD to multiflora roses. They conducted transmission tests by
Phyllocoptes adalius Keifer is a mite very similar to P. fructiphilus and also occurs on many roses in the eastern United States; P. adalius occurs as a vagrant, usually on the underside of mature leaf blades of many species and varieties of roses. It has been thoroughly tested as a vector, but can not transmit RRD (Kharboutli, 1987; Kassar and Amrine, 1990; Amrine et al., 1995). Rose rosette disease was first found in West Virginia in 1989, and spread throughout the state by 2000 (Brown and Amrine, unpub.). Several predators, a parasitic fungus, and drought appear to have affected field populations of P. fructiphilus in West Virginia and may have slowed the spread of RRD.
Rose rosette disease can also be transmitted by grafting, and experiments in Iowa have shown that this approach can be used to augment the virus in dense stands of multiflora rose (Epstein and Hill, 1994b; Epstein and Hill, 1995b, 1995d, 1998, 1999; Epstein et al., 1997). Obrycki et al. (2001) are conducting new trials of RRD releases and augmentation in southern Iowa to reduce multiflora rose in pastureland. They indicate that releases will not be made in areas with ornamental roses. Much of this work has been opposed by the American Rose Society and by rosarians in general (Harwood, 1995; Obrycki, 1995; Philley, 1995; Peck, 2001; Sauer, 2001; Pagliai, pers. comm.). However, augmentation research by Epstein et al. has provided valuable information on the potential spread of RRD from multiflora to ornamental roses.
Mites overwinter as adult females on living, green rose tissue (Amrine and Hindal, 1988; Amrine et al., 1995). In early spring, the mites move from wintering sites (clumps of overwintering foliage, loose bark on live stems, old or loose bud scales, etc.) onto developing shoots to lay eggs. A favorite oviposition site is between the stem and basal petiole of young leaves appressed to stems. Females live about 30 days and lay about one egg per day. Eggs hatch in three to four days and the development of each immature stage (protonymph and deutonymph) requires about two days (Kassar and Amrine, 1990; Kassar, 1992). Thus, in warm weather, one generation may be produced per week. Development is continuous throughout the season until weather turns cold in the fall and mites seek protective wintering sites on the plants. Overwintering mites will die if host canes die, as they require green stem or leaf tissue.
In May, 1987, Amrine et al. (1990) began a long-term study at Clifty Falls State Park in Madison, Indiana. The site was heavily infested with both healthy and RRD-symptomatic multiflora roses. A total of 180 multiflora rose plants were marked and visited monthly during the growing season for the next five years. The initial average density was 1,200 plants per acre and, at the beginning of the study, 30% of plants were symptomatic and 1% had been killed by RRD. The infection increased each year and leveled off to 94% by September 1991 with a mortality of 88%. The average longevity of infected plants was 22.4 months (range three to 48 months). Mite populations were 14 times larger on symptomatic plants compared to healthy plants in 1987 and 1988. Mite populations were low and sporadic in April and gradually increased to peak abundance by September in most years. At peak abundance, nearly all RRD-symptomatic plants (98%+) were infested with mites. The average number of mites per symptomatic shoot in September of each year (1987 to 90) was 112, 30, 112, and 6.6 respectively (mite density on healthy plants was usually below 10 per shoot). The low average number in 1988 (30) resulted from a severe drought that killed mites on desiccated foliage. The low fall density in 1990 (6.6) resulted from unusually cold weather in December 1989 (-31°C), which killed nearly all above ground RRD-symptomatic canes and thus killed most of the overwintering mites. By the end of the study (1994), 97% of the marked plants were dead or symptomatic and the density of live multiflora roses had dropped to about 800 per acre, many of which were new, small plants.
As of 2001, RRD was present in multiflora roses in all counties in West Virginia and was found as far east as Berks County, Pennyslvania, Queen Anne and Talbot Counties, Maryland and Manassas Battlefield, Virginia (Fig. 2). The disease is probably present in Delaware, New Jersey, New York, and other eastern states. It is likely that RRD will be present throughout the eastern United States within ten years. RRD will have a very significant effect on multiflora rose populations, potentially reducing numbers by 90% or more throughout the region. In each local area, the RRD epidemic is likely to continue until multiflora rose stands are killed. Young seedlings will then sprout and reach moderate size before RRD again reinfects the stand. In Madison, Indiana, for example, a survey in 1994 found that while more than 97% of the original large plants had died of RRD, the infection rate of the abundant, newly sprouted plants was only 20 to 25%. The low percent infection rate reflects the slow build up of the infection in new plants. A visit to Clifty Falls State Park on 26 May 2002 revealed an estimated density of 200-400 multiflora roses per acre with an infestation of 60% RRD. Much of the original grassland has become early stage forest, which will shade future germinating seed and the resulting plants. As another example, RRD was first discovered in Monongalia County, West Virginia in 1989; as of July 2001, the average infection rate throughout the county was 10 to 20%. We expect to see infection rates equivalent to Madison, Indiana (30%) within five more years.
A serious limitation to the use of RRD as a biological control agent is its ability to infect ornamental roses. Many species and varieties of roses are susceptable to the vector and to RRD (Tables 1 and 3). However, other plants in the Rosaceae have been found to be immune to the RRD agent (Table 2). Rose breeders and gardeners throughout the eastern United States will need to reduce local stands of multiflora rose for a one or two mile radius in order to lower the risk of infestation by airborne mites, which transmit RRD. Thomas and Scott (1953), Allington et al. (1968), Amrine et al. (1995), and Epstein and Hill (1998, 1999) listed varieties and cultivars of ornamental roses that are particularly susceptible to RRD and its vector (also, see listings in bold type in Table 1). Avoiding planting of these varieties can help reduce injury to adjacent ornamental roses. Peck (2001) listed Cygon 2E (citing work by Dr. George Philley, Plant Pathologist, Texas A&M, Overton, Texas) as a treatment for protecting ornamental roses; other chemicals such as Avid (abamectin) may prove effective in controlling the mites. Thomas and Scott (1953), Allington et al. (1968), and Amrine et al. (1995) discussed varieties of roses resistant to RRD. This information can be used to incorporate resistance into new rose varieties.
Multiflora Rose Seed Chalcid, Megastigmus aculeatus var. nigroflavus Hoffmeyer (Hymenoptera: Torymidae)
infestation rate of 74.1% (range of 64 to 79%). The seed chalcids likely had reached the new plantings in droppings of birds that fed on the hips produced in 1990. In fall 1991, numbers of seed chalcids in the plot were augmented by placing about 1,500 rose hips (average of seven seeds per hip), which had an infestation rate of 79%. In December 1992, the seed infestation rate in the plot reached 77.5% (20 samples; 20 hips each; range of 57 to 93%).
Suggestions that this seed chalcid will infest the seed of other roses seem unfounded. Torymid infestation of seeds of R. setigera, R. palustris, R. carolina, or Rosa canina L., in our study sites over the past 15 years have not been observed. Balduf (1959) reported rearing a dark form of Megastigmus aculeatus from Rosa eglanteria and R. virginiana; these were not reported to be Megastigmus aculeatus var. nigroflavus Hoffmeyer. Only R. multiflora seems to be susceptible, either because of timing of bloom (late May to early June for multifloras in West Virginia versus July for the others), or because the fruits of other roses are too large or thick for the chalcids’ ovipositors to penetrate. Of 31 states in the eastern United States sampled by the author, the chalcid was found in all except Florida, Louisiana, Mississippi, Texas, and northern New England. It will continue to spread by feeding birds until all stands of multiflora roses are infested. Weiss’s report (1917) about seed from Japan having 95% infestation indicates the probable potential for this seed chalcid to infest the seed of multiflora rose.
It is virtually certain that RRD will greatly reduce the density of multiflora rose. No multifloras have been found that are resistant to the disease (Amrine et al., 1990, Amrine and Stasny,1993; Epstein and Hill, 1998). The reduced populations of multiflora rose remaining after the RRD epidemic are likely to be infested by the seed chalcid at the same rate (90 to 95%) as plants in Korea and Japan. Multiflora rose will then be another occasional plant in the environment, and not the noxious weed that it is today. We estimate that this scenario will transpire within the next three to five decades. Farmers and others wanting eradication of multiflora rose desire human intervention to increase the rate of spread of the disease, the mite and the torymid into uninfested areas. However, rosarians desire that all augmentation work with RRD and the mite cease.
Rose Stem Girdler, Agrilus aurichalceus Redtenbacher (Coleoptera: Buprestidae)
Synonyms for this species include Agrilus viridis L., Agrilus viridis var. fagi Ratz., Agrilus communis var. rubicola Abeille, Agrilus rubicola Abeille, and Agrilus politus Say. Many reports of this insect in Rubus (brambles) were made under the name Agrilus ruficolis (Fabricius), the red-necked cane borer, whose symptoms are nearly identical. The two beetles are distinctive and easily separated. This small brownish-golden, metallic buprestid beetle is about 5 to 9 mm long (Fig. 16). It is a non-indigenous species from Europe that has been established throughout eastern North America and is abundant at several sites in Delaware, Indiana, Maryland, Ohio, Pennsylvania, Virginia, and West Virginia. It caused a small degree of control of multiflora rose in Ohio and West Virginia (Amrine and Stasny, 1993). All plant tissue beyond the point where the stem is girdled was killed, including developing rose hips and seeds. Borers overwinter in the previous year’s canes, pupate in April, and emerge as adults in May. Douglas and Cowles (2001) state that development may require two seasons, which is contrary to all other reports. Adults can be found on multiflora rose foliage in sunny mornings. Females oviposit on the bark of new canes in May and June. Larvae hatch and burrow under the bark, moving upward from the oviposition site (Fig. 17). The initial burrowing does not kill the cane but by late July the infested stems begin to wilt, and by August-September, canes beyond the girdle die and appear as brown “flags” on rose bushes (Figs. 18-19.)
The largest infestation we observed was a site with 20% of canes infested (Fayette County, Ohio, 1988). Large numbers of larvae were found to be parasitized; 22 parasitic wasps emerged from 45 canes held for the emergence of 23 beetles. These parasites were Ptinobius magnificus (Ashmead) (Pteromalidae)(determined by E. E. Grissell), Eurytoma magdaldis Ashmead (Eurytomidae) (new host record, determined by E. E. Grissell), Leluthia astigma (Ashmead) (Braconidae) (determined by P. M. Marsh), Metapelma schwarzi (Ashmead) (Eulophidae) (new host record, determined by M. E. Schauff), and Tetrastichus agrili Crawford (Eulophidae) (determined by M. E. Schauff). The last was most abundant. Because of relatively low incidence and high parasitization, we believe that this insect will have only minor importance as a biological control agent of multiflora rose.
Amrine and Stasny (1993) found girdled rose stems on Rosa multiflora only. Douglas and Cowles (2001) report that it occurs on R. rugosa and R. hugonis in Connecticut. Agrilus aurichalceus aurichalceus often was found attacking canes of Rubus (blackberries, raspberries, brambles) (Hutson, 1932; Mundinger, 1941; Davis, 1963). Brussino and Scaramozzino (1982) reported it attacking Rubus fruticosus L., Rubus caesius L., and Rosa idaeus L. in Piedmont, Italy, where it also attacked Rosa alpina L., R. canina, Rosa damascena Mill., and R. rugosa . It has also been listed as attacking Ribes, Grossularia, Crataegus, and Prunus in North America and Europe (Garlick, 1940; Rejzek, 2001); however, these records are in error, and probably represent different species of Agrilus (Brussino and Scaramozzino, 1982).
Evaluation of Project Outcomes
Establishment and Spread of Agents
All four agents have been well established in the eastern United States and should eventually be found in all dense stands of the weeds. Rose rosette disease and Phyllocoptes fructiphilus have been found as far east as Berks County, Pennsylvania, Queen Anne and Talbot Counties, Maryland, and Manassas Battlefield, Virginia. Careful surveys would probably find both agents as far east as New Jersey and southern New York. The rose seed chalcid has been found in 30 eastern states, and it probably is found in all regions where multiflora rose has become established. The rose stem girdler also is found in most areas of eastern North America and in Utah.
Suppression of Target Weed
Amrine et al. (1990) showed that RRD and P. fructiphilus have excellent potential to reduce multiflora rose. Rate of infection of 180 marked plants in Clifty Falls State Park increased from 30% in 1986 to 94% in 1990. Mortality of marked roses increased from 2% to 94% in the same period. However, germination by the vast seed burden replaces most roses killed by RRD. When seed chalcids significantly infest multiflora seed, then reduction will become apparent. This RRD epidemic has now reached equivalent levels in many parts of West Virginia, and it is expected that in the next decade, similar reduction of dense stands of multiflora rose will occur. The rate of infestation of the rose seed chalcid is increasing in all areas surveyed. In some areas of West Virginia, rates of seed infestation now exceed 80% (Amrine, unpub.).
Recovery of Native Plant Communities
In Clifty Falls State Park, multiflora rose was not replaced by native plant species, but by another invasive exotic plant, Japanese honeysuckle (Lonicera japonica Thunb.). This weed has covered nearly all of the old dead roses and has invaded nearly every part of the park (as of 1994). In many areas of West Virginia, multiflora rose has been replaced by the noxious weeds, Tatarian honeysuckle (Lonicera tatarica L.), autumn olive (Elaeagnus umbellata Thunb.), Japanese honeysuckle, and Japanese knotweed (Polygonum cuspidatum Siebold et Zucc.). If the alien invasive weeds can be controlled or eliminated, then native vegetation should recover.
Millions of dollars now spent annually by farmers in many eastern states to control multiflora rose will be saved when the plant is eventually controlled. In West Virginia, during 1980 and 1981, more than 36,500 hectares were heavily infested with multiflora and a ten-year eradication program using herbicides was estimated to cost more than $40 million (Williams and Hacker, 1982). The same or increased acreage is now infested, and allowing for inflation this cost has probably doubled. However, monetary savings will be slow to develop because of the slow natural spread of both the epidemic and biological control agents.
Recommendations for Future Work
Much work remains to be done to survey for the distribution and intensity of infection/infestation of RRD and P. fructiphilus in multiflora roses. The RRD epidemic in multiflora rose stands is expected to increase greatly over the next few decades. Studies and data are not available showing potential recovery of pastureland/farmland and savings involved; this work should be done in areas where significant mortality due to RRD has occurred (Missouri, Illinois, Indiana). Dense stands of multiflora rose will need to be controlled to prevent infection of ornamental roses with RRD. To quote R. Hartzler, “reduction of multiflora rose densities should be a common goal for rose growers and landowners” (Obrycki et al., 2001). Horticulturalists need to breed RRD-tolerant or RRD-resistant roses (Zary, 1995). The rose seed chalcid, now found throughout the eastern United States, should be intentionally released in areas wherever infestation rates are below 50 to 60%. Risk to other rose species from this seed chalcid appears to be minimal, but host range studies should be conducted to confirm the chalcid¢s suspected high specificity. This insect’s high potential to reduce seeding of multiflora rose justifies its increased distribution. Even if not deliberately spread, its range will increase by birds. Eventually, multiflora rose will be reduced to low levels, occurrence of RRD will become minimal, as in California, Wyoming and Utah, where it originated, and problems for farmers and rosarians alike should be greatly reduced.
I would like to dedicate this chapter to the memory of Professor Dale F. Hindal, West Virginia University Plant Pathologist, recently deceased. Our discussion of this possible research at a football brunch in 1985 began our work with multiflora rose, RRD, Phyllocoptes fructiphilus, and Megastigmus nigroflavus. His insights, humor, deep knowledge of plant pathology and entomology, and excellent teaching ability launched our adventures into this work. I also express my gratitude to the many acarologists, rosarians, plant pathologists, agronomists, and others who have shared important information regarding the mites, multiflora rose, the seed chalcid, the rose stem girdler and other arthropods discussed in this chapter. This work could not have been done without the expert assistance of my assistant, Terry Stasny. Much of our research was supported financially by the West Virginia Department of Agriculture; all of our research was conducted as part of Hatch Project 317, the Davis College of Agriculture, Forestry and Consumer Sciences, at West Virginia University.
Allington, W. B., R. Staples, and G. Viehmeyer. 1968. Transmission of rose rosette virus by the
Amrine, Jr., J. W., 1996. Phyllocoptes fructiphilus and biological control of multiflora rose, pp. 741-749.
Amrine, Jr., J. W. and D.F. Hindal. 1988. Rose rosette: a fatal disease of multiflora rose. Circular 147.
Amrine, Jr., J. W. and T. A. Stasny. 1993. Biological control of multiflora rose, pp. 9-21. In McKnight, B.
Amrine, Jr., J. W., D. F. Hindal, T. A. Stasny, R. L. Williams, and C. C. Coffman. 1988. Transmission of
Amrine, Jr., J. W., D. F. Hindal, R. Williams, J. Appel, T. Stasny, and A. Kassar. 1990. Rose rosette as
Amrine, Jr., J. W., A. Kassar, and T. Stasny. 1995. Phyllocoptes fructiphilus (Acari: Eriophyidae) the
Balduf, W. V. 1959. Obligatory and facultative insects in rose hips. Their recognition and bionomics.
Brown, T. 1994. United States distribution of rose rosette disease, pp. 53-56. In Epstein, A. and J. H.
Brussino, G., and P. L. Scaramozzino. 1982. La presenza in Piemonte di Agrilus aurichalceus Redt.
Chalamira, L. R. and L. D. Lawrence. 1984. Agricultural research needs and priorities as perceived by
Crowe, F. J. 1983. Witches’ broom of rose: a new outbreak in several central states. Plant Disease 67:
Davis, D. W. 1963. The control of the rose stem girdler on raspberries in Utah. Journal of Economic
Di, R., J. H. Hill, and A. H. Epstein. 1990. Disease-specific double-stranded RNA associated with the
Doudrick, R. L. 1984. Etiological studies of rose rosette. M. S. thesis, University of Missouri, Columbus,
Doudrick, R. L. and D. F. Millikan. 1983. Some etiology and symptomological aspects of rose rosette.
Douglas , S. M. and R. S. Cowles (eds.). 2001. Plant Pest Handbook. Connecticut Agricultural
Dugan, R. F. 1960. Multiflora rose in West Virginia. West Virginia University Agriculture Experiment
Epstein, A. H. 1995. Biological control of multiflora rose with rose rosette disease. Pm-1615.
Epstein, A. H. and J. H. Hill. 1994a. Characteristics and control of rose rosette disease. The American
Epstein, A. H. and J. H. Hill. 1994b. Characterization of the rose rosette disease causal agent: potential
Epstein, A. H. and J. H. Hill. 1995a. Rose rosette disease - historical aspects and current status, pp.
Epstein, A. H. and J. H. Hill. 1995b. Field augmentation of rose rosette disease: Results and risk
Epstein, A. H., and J. H. Hill. 1995c. Some physiological and structural changes observed in rose
Epstein, A. H. and J. H. Hill. 1995d. The biology of rose rosette disease: A mite-associated disease of
Epstein, A. H. and J. H. Hill. 1997. Regional Research Project, S-268, Evaluation and Development of
Epstein, A. H. and J. H. Hill. 1998. Rose rosette disease. The American Rose Magazine February
Epstein, A. H. and J. H. Hill. 1999. Status of rose rosette disease as a biological control for multiflora
Epstein, A. H., J. H. Hill, and J. J. Obrycki. 1993. Rose rosette disease. Iowa State University,
Epstein, A. H., J. H. Hill, and F. W. Nutter, Jr. 1997. Augmentation of rose rosette disease for the
Evans, J. E. 1983. A literature review of management practices for multiflora rose (Rosa multiflora).
Fawcett, R. S. 1980. Today’s weed—multiflora rose. Weeds Today 11: 22-23.
Garlick, W. G. 1940. Notes on the rose stem girdler, Agrilus communis rubicola AB. The Canadian
Gergerich, R. C. And K. S. Kim. 1983. A description of the causal agent of rose rosette disease.
Good, R. 1964. The Geography of the Flowering Plants, 3rd ed. John Wiley and Sons, New York.
Harwood, C. 1995. Rose industry concerns regarding rose rosette disease. Pp. 79-80 In: Epstein, A.,
Hill, J. H., A. H. Epstein, J. R. Seibel, and W. A. Miller. 1995. Causal agent of rose rosette disease, pp.
Hindal, D. F., J. W. Amrine, R. L. Williams, and T. A. Stasny. 1988. Rose rosette disease on multiflora
Hindal, D. F. and S. M. Wong. 1988. Potential biocontrol of multiflora rose, Rosa multiflora. Weed
Hutson, R. 1932. A new pest on raspberry. Quarterly Bulletin of the Michigan Agricultural Experiment
Kassar, A. 1992. A study of the biology of Phyllocoptes fructiphilus Keifer, and its ability to transmit
Kassar, A. and J. W. Amrine, Jr. 1990. Rearing and development of Phyllocoptes fructiphilus.
Keifer, H. H. 1939. Eriophyid studies VII. Bulletin of the California Department of Agriculture 28:484-505.
Keifer, H. H. 1940. Eriophyid studies VIII. Bulletin of the California Department of Agriculture 29(1): 21-
Keifer H. H. 1966. Eriophyid studies B-21. Bureau of Entomology, California Department of Agriculture,
Kharboutli, M. S. 1987. Eriophyid mites (Acari: Eriophyidae) on Rosa multiflora Thunb. in West Virginia
Klimstra, W. D. 1956. Problems in the use of multiflora rose. Transactions of the Illinois Academy of
Kriebel, R. C. 1987. Plain ol’ Charlie Deam: Pioneer Hoosier Botanist. Purdue University Press, West
Liberty Hyde Bailey Hortorium. 1976. Hortus Third, a Concise Dictionary of Plants Cultivated in the
Lincoln, Jr., W. C. 1978. The effect of the digestive tract on the germination of multiflora rose seed.
Lingenfelter, D. D. And W. S. Curran. 1995. Multiflora rose management in grass pastures (an
Mays, W. T. and L.-K. Kok. 1988. Seed wasp on multiflora rose, Rosa multiflora, in Virginia. Weed
Milliron, M. E. 1949. Taxonomic and biological investigations in the genus Megastigmus. American
Mundinger, F. G. 1941. Two buprestid cane-borers of brambles with experiments on control. Journal of
Nalepa, C. A. 1989. Distribution of the rose seed chalcid Megastimus aculeatus var. nigroflavus
Obrycki, J. J. 1995. Conflicts of interest in weed biological control: lessons for Rose Rosette, pp. 75-76.
Obrycki, J. J., R. Hartzler, D. Frey, and S. Thompson. 2001. New project addresses multiflora rose
Peck, A. 2001. Rose rosette: the early years. Heritage Roses 26(2): 5-7.
Peck, A., A. Boudoin, M. A. Hansen, J. W. Amrine. 1001. Recent spread of rose rosette disease in the
Philley, G. L. 1995. Concerns of Extension. pp. 77-78. In Epstein, A. H. and J. H. Hill (eds.).
Rehder, A. 1936. On the history of the introduction of woody plants into N. America. National
Rejzek, Roman. 2001. Entomology. Prague, Czechoslovakia., Agrilus aurichalceus aurichalceus
Sauer, R. 2001. Rose Rosette Disease Revisited. Heritage Roses 26(2):2-4 + map.
Scott, R. F. 1965. Problems of multiflora rose spread and control. Transactions, 30th North American
Shaffer, D. F. 1987. A study of the biocontrol of Rosa multiflora Thunb. utilizing the rose-seed chalcid
Szafone, R. 1991. Vegetation management guidelines: multiflora rose (Rosa multiflora Thunb.). Natural
Thomas, E. A. and C. E. Scott. 1953. Rosette of rose. Phytopathology 43: 218-219.
Tipping, P. W., A. B. Sindermann. 2000. Disease Notes. Natural and augmented spread of rose rosette
Underwood, J. F., M. M. Loux, J. W. Amrine, Jr., and W. B. Bryan. 1996. Multiflora rose control. Ohio
Underwood, J.F. and E.W. Stroube. 1986. Multiflora rose control. Ohio State University, Cooperative
Viehmeyer, G. 1961. A “new” disease of roses. American Rose Annual 46: 98-101.
Wagnon, H. K. 1966. Report of a new or unusual plant pathogen: Witches’ Broom (Rosette of Rose).
Wagnon, H. K. 1970. Report of new or unusual plant pathogen: Witches’ Broom (Rosette of Rose).
Weiss, H. B. 1917. Megastigmus aculeatus Swed., introduced into New Jersey from Japan. Journal of
Williams, R. L. and J. D. Hacker. 1982. Control of multiflora rose in West Virginia. Proceedings of the
Willis, J. C. 1985. A Dictionary of the Flowering Plants and Ferns, 8th ed. Cambridge University Press,
Zary, K. W. 1995. Breeding for resistance to rose rosette. Pp. 81-82 In: Epstein, A., J. H. Hill (eds)
Zhao, S. 2000. Study of dispersal and diversity of eriophyid mites (Acari: Eriophyoidea). Ph.D.
Selected Websites for Multiflora Rose
BONAP, Biota of North America Program, Texas A&M University Bioinformatics Working Group.
Calflora Database, Digital Library Project, University California Berkeley.
Douglas , S. M. and R. S. Cowles. (eds.). 2001. Plant Pest Handbook, Connecticut Agricultural
Hortiplex. 2002. http://hortiplex.gardenweb.com/plants/
Illinois Department of Conservation, http://www.inhs.uiuc.edu/edu/VMG/mrose.html.
Invasive Plant Council of New York State, New York State Museum and the New York State Biodiversity
Lingenfelter, D. D. and W. S. Curran. 1995. Multiflora Rose Management in Grass Pastures (An
Missouri Department of Conservation.
Nebraska. File G980: PLANT DISEASES, A-4, Ornamentals.
Peck, Anne. 2001. Rose Rosette, a web book. http://web.ntown.net/~apeck.
Roebuck, Field. 2001. Watch out for Rose Rosette. http://froebuck.home.texas.net/newpage2.htm.
Rejzek, Roman. 2001. Entomology. Prague, Czechoslovakia., Agrilus aurichalceus aurichalceus
Synonymized Checklist of the Vascular Flora of the United States, Puerto Rico, and the Virgin Islands,
Underwood, J. F., M. M. Loux, J. W. Amrine, Jr., and W. B.Bryan. 1996. Multiflora rose control. Ohio
USDA, NRCS. 2001. The PLANTS Database, Version 3.1 (http://plants.usda.gov). National Plant Data
Vascular Plant Image Gallery, Bioinformatics Working Group, Texas A&M University, College Station
Virginia Native Plant Society. http://www.vnps.org/invasive/invmulti.htm.
Weeds Gone Wild, Alien Plant Invaders of Natural Areas, PCA Alien Plant Working Group, National
Whitelaw, Mark. 2001. Landscape Consulting Services, Fort Worth, Texas. Rose Rosette Disease
Wildflower Nirvana. http://www.wfnirvana.com/master.html;
Wisconsin Department of Natural Resources:
West Virginia Extension Service, West Virginia University, Morgantown, West Virginia, Control of
|Bargeron, C.T., D.J. Moorhead, G.K. Douce, R.C. Reardon & A.E. Miller |
(Tech. Coordinators). 2003. Invasive Plants of the Eastern U.S.:
Identification and Control. USDA Forest Service - Forest Health
Technology Enterprise Team. Morgantown, WV USA. FHTET-2003-08.