Sulfur Knapweed Moth

Spotted knapweed is an alien weed in North America that first arrived in the late 1800s, probably in ship ballast and as a contaminant of alfalfa seed (Sheley et al. 1999). It was long called Centaurea maculosa in the biological control literature; however the plants in North America are perennial, polycarpic and tetraploid (2n = 36), whereas C. maculosa is biennial, strictly monocarpic and diploid (2n = 18) (Ochsmann 2001; Mráz et al. 2011). The North American plants correspond to Centaurea stoebe subsp. micranthos, which is native to eastern and central Europe. The names "C. biebersteinii" and "C. micranthos" are obsolete synonyms, and the diploid "C. maculosa" is now recognized as Centaurea stoebe subsp. stoebe. Here I will use the name spotted knapweed to refer to the invasive tetraploid species. Spotted knapweed has invaded over 2 million ha of rangeland in the western USA and Canada, where it displaces desirable vegetation, reducing forage and plant diversity (Duncan et al. 2004). The annual cost due to infestation of over 2 million acres of rangeland and wildland in Montana was estimated to be over $14 million (Hirsch and Leitch 1996). Management of knapweeds using herbicides or controlled grazing with sheep or goats is generally too expensive or unfeasible, so there was strong support to develop a classical biological control program which started in 1961 (Story 1995). Surveys in Europe identified 20 prospective agents, of which 12 were eventually released in North America (Story 2002).

Agapeta zoegana  (Lepidoptera: Tortricidae), is a root boring moth that is native to Europe, occurring predominantly in semi-natural, undisturbed sites (e.g. dry grassland, steppe, and south-facing slopes) and ruderal habitats (e.g. gravel quarries) in areas with a moderately humid temperate climate or with an arid sub-continental climate (Muller et al. 1988). Historically, it has been recorded to feed on the roots of Scabiosa columbaria, Centaurea jacea, C. maculosa, C. nigra, C. paniculata, Jurinea linearifolia, and probably C. scabiosa. It was tested on 56 plant species in five families, including closely related native species and species of economic importance such as safflower and artichoke, to determine if it would be safe to use as a biological control agent (Muller et al. 1988). It only developed on a few closely related species of Centaurea including spotted knapweed and to a lesser extent on diffuse knapweed, C. vallesiaca, and C. nigrescens. It did not develop on S. columbaria or C. jacea, raising questions about these historical host records. The moth was first released in the United States in 1984 in Montana. It has been released throughout the western USA, and has become established in California, Colorado, Idaho, Minnesota, Montana, Nebraska, North Dakota, Oregon, South Dakota, Utah, Washington, and Wyoming. However, it is not known to be established after releases in Arizona, Indiana, Nevada, North Dakota and Wisconsin (Story 2002, 2004).

This species appears to have two generations per year in Austria and Hungary, but only one in Montana and British Columbia, probably due to a shorter summer (Muller et al. 1988, Powell et al. 2000; Story 2002; Corn et al. 2009). Adult moths emerge from knapweed roots in early July through early September, peaking in early August in Montana (Story et al. 1991). The sex ratio is 1:1. Mating takes place within 24 hours after emergence, and mated females begin laying eggs the next day. Adults live for about 9 to 14 days and females lay about 75 eggs (up to 290), mostly within a 4-day period. The eggs are laid singly or in small clusters of up to 3 on the surfaces of the stem and leaves, predominantly in crevices, of knapweed plants. They prefer to oviposit on rosettes (81%) compared to flowering bolted plants. The eggs hatch in 7 to 10 days at 23-24°C, and the larvae move immediately to the root crown and mine into the root. There are 6 larval instars. Larvae are capable of killing small rosettes and then moving at least 10 cm to another knapweed plant to feed. The larvae overwinter in the root, complete development the following spring, and pupate by midsummer. The moths are strong fliers and will disperse to nearby knapweed patches. Methods to rear moths in field cages are described by Story et al. (1994), who estimated the cost to be $1.32 per insect.

In surveys in Austria and Hungary, A. zoegana was the most common root feeder, after 4 pooled Apion spp., with an average density of 23.6 larvae per 100 knapweed roots (Muller et al. 1988).  Predation of larvae was observed but has not been quantified, and parasitism of large larvae in pupae ranged from 11 to 60% in Europe (Muller et al. 1988). A Bracon species caused 20 and 29% parasitism of larvae at 2 of 9 sites surveyed in Montana (Herron-Sweet et al. 2015). 

A survey of 86 releases in Montana, Idaho and Washington found no effect of release number (49 to 1945 adults), spotted knapweed patch size or plant density on establishment success (Clark et al. 2001a). Highest establishment rates were associated with clay loam or sandy clay loam soils (100%) and even-age forests (85%), and roadside strips of knapweed had lower success than continuous nonlinear patches. Agapeta zoegana has also established on diffuse knapweed, infesting 17 and 20% of plants at two sites in central Montana 3 to 9 years after the first releases (Smith 2004), although infestation rates were much lower in British Columbia (0 to 10%; Myers et al. 2009).

The presence of the insect can be determined by looking for larvae and larval feeding damage in the roots of knapweed plants, by searching visually for adults resting on plants, or by sweep net (Clark et al. 2001b). Sweep net was the most efficient (regarding observations/min.), but it had the lowest detection rate of the three methods. Detection by visual searching had a higher recovery rate, and root dissection took the most time. An analysis of natural infestation of spotted knapweeds in Montana indicated that larvae are preferentially found in plants with root diameter greater than 3.5 mm (Smith and Story 2003), which suggests that it is most productive to sample large plants if just trying to detect the presence of infestation.  

Adults can be collected by night lighting: hanging a sheet and illuminating it with a bright light or black light (Fitzpatrick 1989; Story et al. 2001). Fluorescent and ultraviolet light were equally attractive to the moth. Catches were highest 1 to 2 hours after sunset. However, most of the adults attracted are males, so this is not a good strategy for collecting insects to move to a new location.

Females produce a pheromone that attracts males. The main component is Z-11-tetradecenyl acetate (Z-11:AC), which is a common component of pheromones of tortricid moths (Tóth et al. 1985). Pheromone traps charged with Z-11:AC can trap male A. zoegana, although other species of tortricids are also likely to be caught. This method was found to be more consistent than sampling roots, visual transects and sweep netting, but required two site visits (to set up and take down traps) (Sturdevant and Dewey 2002).

Larvae feeding on potted spotted knapweed plants stimulated compensatory growth of roots (Muller 1989). When newly hatched larvae were transferred to first-year spotted knapweed plants in a field garden experiment near Basel Switzerland, they caused no measurable reduction of shoot mass, root mass or shoot/root ratio (Steinger and Müller-Schärer 1992).  In another field garden experiment in Switzerland using first-year spotted knapweed plants exposed to varying numbers of moths in cages (0, 2, 4 or 8 pairs), intermediate levels of exposure slightly decreased plant height and survival during the first season, although this was trivial compared to the impact of grass competition (Müller-Schärer 1991). Furthermore, at plots with high density of grass and knapweed, exposure to 2 pairs of moths increased knapweed plant survival, number of shoots per plant and number of shoots per area as compared to controls without moths (Müller-Schärer 1991), suggesting that the plants compensate for insect damage. A field garden experiment in Montana showed that growth of a perennial grass, Festuca idahoensis, was reduced more by spotted knapweed plants in plots where moths had been released than in plots without moths, suggesting that the herbivory stimulated greater interspecific competitive effects, the reverse of what is desired (Callaway et al. 1999). Another experiment showed that knapweed plants attacked by A. zoegana produced more flowerheads and slightly decreased reproduction of nearby F. idahoensis plants (Ridenour and Callaway 2003). When infested plants were transplanted to the field for a year, infestation failed to reduce growth of spotted knapweed, but it reduced the growth of nearby grasses by 28.5% vs. 7.2% by uninfested knapweed plants (Thelen et al. 2005). These effects were associated with higher concentrations of catechin; however, the chemical extraction methods used in these studies appear to be erroneous (Blair et al. 2005), which led to posting of an erratum in 2010 (Bais et al. 2003). Proper methods showed much lower concentrations of catechin associated with spotted knapweed, and application of 10x the concentration of catechin that was previously reported to cause 100% mortality only slightly reduced growth of F. idahoensis (Blair et al. 2005).

Comparison of spotted knapweed plants in a field in Montana with an established population of A. zoegana (at least 4 years after the first release) to those in a field without the moth found plants to have 43% less biomass, 29% fewer stems and 43% fewer flowerheads in the presence of the moth (Story et al. 2000). However, direct comparison of infested vs. uninfested plants within a site showed that infested plants were larger (112% above-ground biomass) and 22% older, suggesting that the moths preferentially attack larger plants. However, field experiments in Switzerland indicated that moths lay more eggs on rosettes and plants with one stem compared to multi-stemmed plants, although this did not result in differences in larval densities (Collins and Müller-Schärer 2012). In the Montana study, there was no impact of the moth on knapweed rosette density, whereas high grass density plots had 35% fewer bolted knapweed plants and 38% fewer rosettes and 50% fewer seedlings as compared to low grass density (Story et al. 2000). Effects of the moth and grass competition were not additive. In a survey of 12 sites in Montana where the moth was released, spotted knapweed stem density declined at all sites between the pre-release (1991-1993) and post-release (1997-1998) evaluations, but this decline was not correlated to establishment of the moth (Clark et al 2001c).

Although diffuse knapweed has been well controlled in parts of British Columbia (Gayton and Miller 2012; Bourchier and van Hezewijk 2013), infestation of roots by A. zoegana was low (0 to 10%) (Myers et al. 2009), so it is unlikely that it had any significant impact in that region.

A test of the effect of fall versus spring application of two herbicides (2,4-D and clopyralid) indicated that spring applications may be compatible if delayed until late spring but that fall applications kill too many larvae (Story and Stougaard 2006). Larvae appeared to tolerate moribund plants, but usually died in plants killed by herbicide. On the other hand, if there are areas that cannot be sprayed, they could serve as a refuge for the moth, which would allow it to persist. Targeted grazing of spotted knapweed by sheep in July or August in northwestern Montana did not affect infestation rates of A. zoegana, which was an average of 1.2 larvae per m2 (Mosley et al. 2016). Sheep grazing in July and August removed 96 and 98% of knapweed flowerheads, respectively. Larvae in infested knapweed plants survived a wildfire in Montana, presumably because they were protected underground and the knapweed plants were not killed (Sturdevant and Dewey 2002).

Three years after releasing A. zoegana in areas with high grass density (50% cover) vs. low density (10%), the percentage of knapweed plants infested with A. zoegana in the high grass-cover plots was nearly twice that of the low grass-cover plots (Story et al. 2000). Plots with high grass density contained fewer bolted knapweed plants (35%), fewer rosettes (38%), and fewer seedlings (50%) than plots with low grass density. Presence of the moth had no effect on knapweed rosette density, but appeared to reduce the number of bolted knapweed plants in plots with low grass density by 39% and increase the density of knapweed seedlings in the spring by 65%.