Establishment and range expansion of invasive Cactoblastis cactorum (Lepidoptera: Pyralidae: Phycitinae) in Texas

1 Introduction

Most herbivorous insects are ecological specialists (Futuyma and Moreno 1988) on a narrow range of the host plants available to them in a community (Forister et al. 2015). This global pattern characterizes many species of Lepidoptera (Dyer et al. 2007). Host plant ecological specialization can be driven by an insect’s behavioral adaptations for the host environment (Bernays 1998; Knolhoff and Heckel 2014), ability to overcome host chemical defenses (Kessler and Kalske 2018; van der Linden et al. 2021), and tolerance of environmental extremes in the host’s range (Forister et al. 2012). These drivers are affected by herbivore natural enemies (Abdala-Roberts et al. 2019; Mooney et al. 2012) in a tri-trophic context that can become disconnected when organisms enter a novel system (Stireman & Singer 2018).

Widely distributed Lepidoptera tend to colonize additional novel hosts (Jahner et al. 2011), and this pattern can strengthen at expanding geographic range margins (Singer and Parmesan 2021). Many globally invasive herbivorous insects consume a phylogenetically diverse range of host plants while others remain restricted to novel host plants with particular chemical or physiological characteristics (Garnas et al. 2016). Insects restricted to hosts within one genus, or one family, in their native range are unlikely to acquire phylogenetically or biochemically dissimilar hosts in novel environments (Dyer et al. 2007; Sunny et al. 2015). However, non-native specialists can cause extensive damage to novel hosts within their fundamental dietary niche because they have a predisposition to tolerate host chemical defenses (Sedio et al. 2020) and have been released from the constraints present in their native range (Callaway and Ridenour 2004; Keane and Crawley 2002). The cactus moth, Cactoblastis cactorum Berg (Lepidoptera: Pyralidae: Phycitinae), is a host specialist herbivore of several prickly pear cactus (Opuntia, Cactacaeae) in its native range that has expanded its diet to include many novel prickly pear hosts as its range has expanded (Hight et al. 2002; Morrison et al. 2020; Simmonds and Bennett 1966).

C. cactorum is native to Argentina, western Uruguay, southern Paraguay, and Brazil (Mann 1969). This species is multivoltine; gravid females regularly oviposit over 100 eggs that are adhered together into an ‘egg stick’ (Legaspi et al. 2009) that is laid on the cladode or tip of an Opuntia spine and is characteristic of cactus feeding pyralid moths (Figure 1). Larvae hatch 2 to 3 weeks after oviposition. Neonate larvae penetrate the host cuticle and begin feeding gregariously beneath the cuticle before burrowing into the cladode where they tunnel until all vascular tissue is consumed, leaving a fairly transparent cladode that desiccates and falls off the plant (Figure 1). Once a cladode has been consumed or decayed from secondary microbial infection (Starmer et al. 1988), the larvae burrow through the joint into an adjacent cladode or leave the damaged cladode and travel to a fresh cladode where they re-enter the tissue and continue feeding internally. Damage to host plants can cause desiccation of whole sections of cladodes (Figure 1). Larvae develop through five instars and exit the plant to form a cocoon and pupate in leaf litter near the base of the host plant. Larval development varies seasonally from 100 to 265 days depending on temperature and precipitation (Mann 1969). Larvae have a patchy distribution in the native range, presumably because this species coevolved with a suite of native natural enemies, competitors, and host plant defenses that control population growth (Dodd 1940; Folgarait et al. 2018; Mann 1969; Varone et al. 2019). Unlike sympatric Cactoblastis congeners, which are completely host-specific (i.e., monophagous), C. cactorum is oligophagous, consuming eight Opuntia species or varieties, as well as narrow-jointed prickly pear Salmonopuntia salmiana (J.Parm. ex Pfeiff.) P.V. Heath (Cataceae), in its native range (Mann 1969; Varone et al. 2014). Native North American cactus moths, such as Melitara and Olycella species (Pyralidae: Phycitinae), have life histories that are very similar to C. cactorum (Mann 1969).

Figure 1:

Neonate Cactoblastis cactorum larvae eclosing from their egg stick (A), late instar larva inside a cladode (B), fresh frass exuded from opening constructed on a Opuntia engelmannii cladode (C), cladode with tissue hollowed out by gregarious larvae located within the red insert (D), desiccated cladodes with heavy larval feeding damage and frass pile to the right (E), female moth on top and male on bottom (F). Photo credits: Invasive Species Research Lab at The University of Texas Austin (A–E), and Jeff Lotz at the Florida Department of Agriculture and Consumer Services (F).

C. cactorum was deployed as a biological control agent of invasive Opuntia in Australia beginning in the 1920s where it contributed to bringing populations of introduced North American species Opuntia stricta (Haw.) Haw. and Opuntia ficus-indica (L.) Mill. to manageable densities (Mann 1969; Zimmermann et al. 2000, 2004]). Biological control of invasive Opuntia with C. cactorum was not as effective in some regions of southern Africa, possibly due to host plant mismatches and higher predation levels (Hoffmann et al. 2020; Petty 1948; Zimmermann et al. 2004). Control of invasive Opuntia in Australia and southern Africa was achieved by augmenting mass C. cactorum releases with repeated introductions of cochineal bugs (Dactylopius spp., Hemiptera: Dactylopiidae), another Opuntia specialist insect group native to the Americas (Hoffmann et al. 2020). Following the successes in Australia and southern Africa, C. cactorum was introduced to control invasive Opuntia populations in Kenya where it did not establish, and the islands of Hawai’i, Mauritius, New Caledonia, St. Helena, and Ascension with varying degrees of success (Hoffmann et al. 2020).

In the late 1950s, C. cactorum was used to control native Opuntia species on Caribbean islands including Antigua, Montserrat, the Cayman Islands, Nevis and St. Kitts. The moth is assumed to have dispersed naturally to Cuba, Hispaniola, Puerto Rico, the U.S. Virgin Islands, and the Bahamas by the 1980s, although human movement cannot be ruled out (Hoffmann et al. 2020; Zimmermann et al. 2004). The moth had established a population in the Florida Keys by 1989 (Habeck and Bennett 1990). It may have dispersed naturally from the Caribbean via windborne dispersal as it is thought to have done between Hawai’ian Islands (Garcia-Tuduri et al. 1971; Pemberton 1995) or via hurricane winds (Andraca-Gómez et al. 2015, 2020]). Unintentional, human-assisted movement via the commercial Opuntia trade, or airline passenger luggage are also possibilities (Pemberton 1995).

Since 1989, C. cactorum has spread throughout Florida, north to coastal South Carolina and west to the Louisiana Gulf coast (Simonsen et al. 2008). The southeastern USA species O. stricta and Opuntia humifusa (Raf.) Raf. have experienced increased mortality (Baker and Stiling 2009; Jezorek et al. 2012) and downward shifts in size structure and reproductive output (Sauby et al. 2017) following C. cactorum establishment. During this period, the sterile male technique was developed (Carpenter et al. 2001), an approach designed to prevent the spread of viable C. cactorum adults and eradicate isolated populations beyond the leading edge of the southeastern USA invasion front and on Isla Mujeres and Isla Contoy in Mexico (Bello-Rivera et al. 2021). In addition, the United States Department of Agriculture, Animal and Plant Inspection Service (USDA-APHIS) was mandated to eradicate Opuntia populations infested with C. cactorum larvae (Plant Protection Act 2000). These management efforts provided time for research, host-specificity testing, and development of a biological control agent for introduced C. cactorum populations, Apanteles opuntiarum Martínez & Berta (Hymenoptera: Braconidae), a host-specific, larval koinobiont parasitoid wasp of C. cactorum in its native range (Awad et al. 2019; Folgarait et al. 2018; Varone et al. 2015). Release of the agent in North America was pending regulatory approval as of December 2024. Since its arrival in the southeastern USA, anxiety has existed about the eventuality of C. cactorum invading arid dryland biomes of the western USA and Mexico where Opuntia is represented by over 80 native species that support a diversity of trophic webs (Hight et al. 2002; Morrison et al. 2020; Simonson et al. 2005). Moreover, damage to commercial Opuntia operations could substantially impact local economies and human livelihoods, particularly in Mexico where it is a significant agricultural commodity that was valued at over $200 million USD in 2022 (Valdez 2022).

The first known observation of C. cactorum in Texas was made in the town of Alvin in July 2017 (29.38832 °N, 95.27123 °W; www.iNaturalist.org). The USDA-AHPIS responded in the summer of 2018 by eradicating Opuntia patches found in Brazoria, Chambers, Colorado, and eastern Matagorda counties (USDA-APHIS, personal communications). In November 2018, we discovered that C. cactorum had dispersed to the south-west as far as the Texas Parks and Wildlife Mad Island Wildlife Management Area (28.67263 °N, 96.04723 °W) and Collegeport (28.72526 °N, 96.16983 °W) in western Matagorda County, Texas. The C. cactorum dispersal pathway into Texas is unknown, but there are several possibilities. Prevailing winds that blow from the south-east to the north-west across the Gulf of Mexico may have facilitated atmospheric dispersal of C. cactorum across unsuitable habitat in the Mississippi River region, or across the Gulf of Mexico, to suitable host plant patches along the south-east Texas Gulf Coast (e.g., Andraca-Gómez et al. 2015, 2020]; Lander et al. 2014; Siljamo et al. 2020). Again, alternatives include tropical storm systems (Andraca-Gómez et al. 2015, 2020]) or human-facilitated dispersal (Pemberton 1995).

Our main question was, “how did the distribution of invasive C. cactorum change from 2017 to 2023 in Texas?” We investigated this question with a 4-year survey to monitor the expansion of the C. cactorum population, and we used local weather data to correlate the observed dispersal pattern with seasonal temperature variation. Our objectives were to 1) document the novel range occupied by C. cactorum, 2) measure the yearly expansion rate to the south-west, and 3) leverage that data to set up testable predictions about biological factors that may affect C. cactorum population expansion.

2 Materials and methods

3 Results

3.1 Abundance and distribution

We set 3,785 pheromone traps over 214 trapping days that caught 871 C. cactorum moths between April 2020 and December 2023. Less than 1 % of the traps deployed caught moths and the number of individuals per trap varied from 1 to 24 moths. Total moth abundances summed across the study period were highest from June to July (n = 211 total) and late August to September (n = 251 total) with an average of five individuals per trap ±3 (SEM) during these maximum abundance peaks (Figure 2). Moths were consistently collected from April to December annually, suggesting continuous and overlapping larval development and moth activity (Figure 2). There was no significant difference in the number of pheromone traps deployed as distances from the Texas Gulf Coast increased which showed that sampling effort was consistent across the study region (Pearson correlation, p = 0.86, n = 515; Supplementary Figure 1). In addition, there was no correlation between the number of traps deployed weekly and the number of moths captured, which showed that observed abundances were not due to spatial sampling bias (Pearson correlation, p = 0.11, n = 515; Supplementary Figure 1). The most inland moth samples were trapped 151 and 153 km from the coast along Interstate-10 in Columbus (29.840154 °N, 96.544558 °W) and Schulenberg, Texas (29.69442 °N, 96.968397 °W), respectively. Moth abundance declined significantly with increasing distance from the coast (Spearman correlation, p < 0.001, r = −0.25, n = 515; Supplementary Figure 1A). Moths were significantly more abundant in the area 0–80 km from the Texas Gulf Coast than the area 80–160 km from the coast (Welch t-test, p < 0.001, t = 4.4, df = 272.8; Figure 4). For context on spatial variation in moth abundance, 78 % of the individuals were trapped 0–80 km from the coast (n = 678) while 22 % were trapped greater than 80 km from the coast (n = 193). In addition, DD were significantly higher 40 km from the coast (mean = 61 DD ± 0.6 SEM) than 120 km inland where there were 43 DD below 0 °C from 2020 to 2023 (mean = 51 DD ± 0.7 SEM; Welch t-test, p < 0.001, t = −11.4, df = 2,871; Supplementary Figure 2).

Figure 2:

Temporal variation in Cactoblastis cactorum abundances in Texas from 2020 to 2023. Total moths captured at pheromone traps are summed by the week that they were collected and measured on left y-axis. Average degree days (DD) are plotted with a continuous blue line and measured on the right y-axis. Vertical dashed lines indicate the period from approximately April to December each year when pheromone traps were deployed; data that fall outside of this period were from roadside or sentinel site larval surveys.

The Texas C. cactorum population has expanded rapidly from the point of initial discovery in 2017 at Alvin, Texas (www.iNaturalist.org). By 2020, the infested area had expanded to 20,831 km². From 2020 to 2023, the area increased by 31 % to 27,218 km², to the west and south-west. The expansion followed the Texas Gulf Coast south-west to the north-eastern edge of Corpus Christi Bay and slowed considerably along a north to south invasion front that approximately parallels US Route 183. The increase in area was only 0.27 % from 2022 to 2023 (Figure 3; Supplementary Table 2).

Figure 3:

Cactoblastis cactorum range size expansion in southeast Texas from 2017 to December 2023.

The largest concentration of moths trapped since 2022 were around the towns of Victoria, Goliad, and Refugio, Texas at the south-western edge of the invasion front (average 1 moth/trap ±0.2 SEM). No moths have been detected north of Interstate-10 between Columbus and Seguin, Texas. The area east and north of Alvin, Texas has not been sampled because the project objective was to monitor moth population expansion to the west and south-west. However, no reports of moth presence in Texas have been made to the USDA-APHIS east or north of the study area since C. cactorum established in Texas, to our knowledge (Figure 3).

3.2 Range expansion

The C. cactorum population dispersed to the south-west at an average of 46.2 km per year (range = 45.9–47.1 km year) during the initial establishment in Texas between 2017 and 2020 (Figure 3, Supplementary Table 2). The population dispersed an average of 7.5 km per year (range = 2.2–13.9 km year) to the south-west between 2020 and 2023, with only an average spread of 3 km from 2022 to 2023 (Figure 3, Supplementary Table 2). Seasonal moth abundances correlated with DD between April and December (Pearson correlation, t = 1.9, df = 84, p = 0.05, r = 0.20; Figure 2). While moth presence in the study area was continuous from spring to fall, and fairly consistent across years, we observed three to four flight peaks with observable increases in moth abundance each year (Figure 4). There were four seasonal flight peaks 0–80 km from the coast and three flight peaks 80–160 km inland from the coast. The number of moths trapped in these flight peaks varied among years (Figure 2). Flight peaks 1 and 2 coincided close to the coast and further inland, while the third inland flight peak occurred over a longer period and stopped sooner than on the coast (Figure 4).

Figure 4:

Variation in the total number of Cactoblastis cactorum caught in pheromone traps from 2020 to 2023 at sites 0–80 km from the Texas Gulf Coast (A) and 80–160 km inland from the coast (B). Moth abundances were totaled by the week of the year that they were trapped. Seasonal flight peaks are highlighted in red. Vertical dashed lines indicate the period from approximately April to December each year when pheromone traps were sampled; data that fall outside of this period were from roadside or sentinel site larval surveys.

Discrete flight peaks are expected to correspond with complete Pyralidae generations (Göttig and Herz 2017; Santos et al. 2023). So, we interpreted three discrete flight peaks as three principal C. cactorum generations in a year. Using three main generations per year as the mode, the C. cactorum population dispersed to the southwest at an average of 15.4 km/generation (range = 15.3–15.7 km/generation) during the initial establishment from 2017 to 2020. The population only dispersed an average of 2.51 km/generation (range = 0.7–4.6 km/generation) to the south-west between 2020 and 2023 (Figure 3, Supplementary Table 2).

3.3 Apparent competition with native cactus moths

Our compilation of native cactus moth records revealed abundant Melitara species observations from the north-west to the south-west of the current invasive C. cactorum population range in south-east Texas (Figure 5). Melitara density increases substantially at the western edge of the invasion front along the Interstate-37 corridor as well a smaller increase to the east towards Galveston Bay and the outlying barrier islands. The Melitara density records are very low in the center of the Texas C. cactorum population invaded range as well as to the north and east of the range where Opuntia host plant density is low in deciduous woodlands. Opuntia density is also low in the current invasion zone as a result of heavy agricultural land use, industrialization, and mechanical removal by landowners (authors personal observations).

Figure 5:

Heat map estimating the density of native cactus moths (Melitara) in Texas. No color indicates no Melitara records, cooler colors indicate lower density, and warmer colors indicate higher density. The Cactoblastis cactorum range as of December 2023 is indicated with a dashed line.

4 Discussion

C. cactorum experienced an enormous southwesterly population expansion in Texas at a rate of 46.2 km per year, or 15.4 km per generation, at its peak dispersal rate period from 2017 to 2020. The moth was expected to arrive at the Rio Grande Valley of Texas, on the USA-Mexico border in approximately 6.9 years at this rate (Supplementary Table 2). It was estimated that C. cactorum was capable of expanding its range in Florida by 50–75 km year (Stiling 2002), a rate comparable with that reported here from 2017 to 2020. However, the decreased expansion rate of 7.5 km per year that we observed from 2020 to 2023 is more similar to dispersal rates observed in Australia (6.4–9.6 km year; Dodd 1940) and South Africa (1.2–2.4 km year; Petty 1948). Lower dispersal rates from 2020 to 2023 were likely affected by high mortality during extreme weather events in Texas which included winter storms in February 2021 and 2023 as well as long-term regional drought and excessive heat from summer to fall of 2023. C. cactorum is unlikely to tolerate low temperatures in North American deserts and high plains where winters are colder, longer, with more extreme fluctuations than the region where C. cactorum is native (Hight and Carpenter 2009; Pérez-De la O et al. 2020; Simonson et al. 2005). Our finding of fewer moths and flight peaks further inland from the coast where temperatures were lower supported this pattern.

Moth abundance varied considerably across 4 years of monitoring (Figure 2). Abundances were lower in 2021 and 2022 than in 2020 or 2023. We speculate that this pattern was largely driven by Winter Storm Uri in 2021 which brought 1 week where the temperature rarely exceeded 0 °C across most of Texas. These extreme temperatures negatively affected C. cactorum directly by causing many individuals to freeze and indirectly by causing massive dieback of Opuntia in the region (authors personal observations). Many Opuntia that were not sheltered from snow and ice lost much of their above ground tissue and did not re-grow enough material to support cactus moth larvae until 2023. This boom-and-bust cycle of population growth is typical of invasive insects that have recently established in a novel environment (Strayer et al. 2017). In this case, boom-and-bust population growth demonstrated that while C. cactorum is vulnerable to environmental extremes, the population was capable of quickly recovering and reaching population sizes that are comparable to those before the disturbance event (Figure 2).

C. cactorum has two generations per year in its native range (Mann 1969) as well as in the introduced ranges in Australia (Dodd 1940) and South Africa (Petty 1948). In Florida, there are three generations per year with earlier annual initiation of flight activity and overlapping generations at southern sites with higher average annual temperature (Hight and Carpenter 2009). Three climate variables summarizing temporal variation in temperature extremes explained 94 % of the variation in a recent North America C. cactorum niche distribution model (Pérez-De la O et al. 2020). These empirical and theoretical findings are consistent with our observations that moth abundances and flight activity increased closer to the Texas Gulf Coast where average annual temperatures were significantly higher (Figure 4, Supplementary Figure 2).

Voltinism is a strong predictor of insect dispersal rate, where species with multiple generations per year spread faster than those with only one (Fahrner and Aukema 2018). Another important component of dispersal rate is the distance that insects can travel (Evans 2016). A laboratory flight mill study of C. cactorum corroborated field observations from Florida by demonstrating that total distances flown by female moths was 2.2 km on average, with no significant difference in distance traveled by mated and unmated females (Sarvary et al. 2008). At this dispersal rate, the Florida population could expand its area 6.6 km year in one direction with the mode of three generations per year (Hight and Carpenter 2009). This annual, individual flight-based dispersal rate does not account for the observed dispersal rate of in Florida (50–75 km year; Stiling 2002) or during the period of rapid population expansion in Texas from 2017 to 2020 (mean = 46.2 km year; Supplementary Table 2). The per generation dispersal rate estimated in Florida would not add up to the annual expansion rate that we observed even if there was complete overlapping and continuous generations in all 9 months of the year where we observed moth activity (2.2 km generation × 9 months = 19.8 km year). Therefore, another factor must have facilitated long distance dispersal.

Prevailing winds are expected to move an insect a much greater distance than an individual moth can travel via its own flight (LeBrun et al. 2008; Sarvary et al. 2008) and are an alternative dispersal model to flight-driven population expansion (Hight and Carpenter 2009). Wind governs the dispersal dynamics of introduced flying insects such as wasps (Lander et al. 2014), beetles (Chase et al. 2017), and flies (LeBrun et al. 2008). Prevailing winds may actually explain the records of C. cactorum females traveling 24 km to oviposit in Australia (Dodd 1940). In south-east Texas, non-native insect population edges that expanded with the prevailing winds dispersed a greater distance than populations that dispersed into the wind (LeBrun et al. 2008). Prevailing winds consistently blow from the south-east to north-west for most of the year across the Gulf of Mexico from eastern Florida to the Gulf Coast of Texas and Mexico (ncei.noaa.gov). The fact that observed dispersal rates exceed average distances that C. cactorum could fly on their own may indicate that prevailing winds facilitated long-distance C. cactorum dispersal to patches with suitable resources within Texas, as well as throughout Florida.

Understanding the relationship between dispersal and environmental conditions is important to C. cactorum management because generations may proceed faster, and flight periods may be extended, or even become continuous, in south and south-west Texas, and in Mexico where winters are warmer and Opuntia species occur at higher densities than their current range (Simonson et al. 2005). Dryland regions of North America with abundant Opuntia host plants may create high resource patch connectivity on the landscape that extends west to the Mojave and Sonoran Deserts, and south to lowland tropical forests in Mexico (Crowl et al. 2008). Mountains such as the Rocky Mountains in New Mexico and the Sierra Madre cordilleras in Mexico may present barriers to western dispersal if the moths are primarily dispersing on their own because high-elevation conditions might be too cold for flight (Hight and Carpenter 2009). However, the possibility of long-distance dispersal via wind reported here, and speculated about in other regions invaded by C. cactorum (Andraca-Gómez et al. 2015, 2020]; Garcia-Tuduri et al. 1971; Pemberton 1995), indicate that prevailing winds could move moths past these physical barriers if there are no other factors limiting dispersal. Understanding how dispersal is affected by the interaction between wind and other environmental factors that vary seasonally (e.g. LeBrun et al. 2008) will be critical for efficient deployment of C. cactorum biological control agents. We predict that other factors contributing to annual and seasonal variation in C. cactorum abundances include relative humidity, precipitation, and biogeochemical factors that indirectly affect C. cactorum by altering host plant quality. Additional niche distribution modeling that expands on the conclusions reached by Pérez-De la O et al. (2020) would allow us to parse out how environmental and biological factors limit the C. cactorum distribution in North America.

Insights into the lower dispersal rates and novel range occupation from 2021 to 2024 may be deduced from the observation that native cactus moth population densities, including C. cactorum in its native range, are patchy in the Americas (Mann 1969; authors personal observations). Patchy distributions indicate that top–down regulation by natural enemies and/or bottom–up regulation via host plant quality and toxicity are occurring for these organisms. The low number of Melitara observations in the area occupied by C. cactorum probably reflects reality, as opposed to a sampling bias, for two reasons (Figure 4). One, that region has substantial agricultural and industrial footprints, which are expected to lower insect abundance and species richness (Wagner et al. 2021). Two, amateur and professional observations of various organisms are made in this area regularly, which indicates that a lack of native cactus moth observations actually reflects a low density in the region. Other patterns in Figure 4 are most certainly artefacts of human settlement and property ownership. One, Melitara has a low density in south and far west Texas, with the exception of areas around population centers in Corpus Christi, Laredo, and the Rio Grande valley, as well as state and national parks. This pattern is most certainly a false negative because most of that region is comprised of large, private ranches where field observations are not generally made public. Two, dense population centers in central Texas, from Austin through San Antonio to Corpus Christi, increase the likelihood that people will observe cactus moths or their larvae. This pattern supports our hypothesis that Melitara, and by extension their parasitoids, become more common as the density of Opuntia host plants increases in the Hill Country and Tamaulipan thornscrub of central and south-west Texas. This preliminary evidence suggests that native cactus moth apparent competition with C. cactorum may be occurring via shared parasitoid assemblages.

Parasitism by coevolved Hymenoptera and Diptera accounted for 20 % of C. cactorum mortality in Argentina, its native range (Varone et al. 2019) and native generalist egg-parasitoids have been reared from C. cactorum in Florida (Paraiso et al. 2011). North American cactus moths, including Melitara, are parasitized by wasps and flies in at least five families (Fernandez-Triana et al. 2013; Mann 1969; Morales-Gálvez et al. 2022; Villegas-Luján et al. 2024). Given the shared biology and evolutionary proximity of American cactus moths, we hypothesize that some native North American cactus moth parasitoids can utilize C. cactorum as a host, and thus provide a level of natural biological control. In North America, cases of native parasitoids having a negative impact on invasive insect pests have been reported for the gypsy moth (Timms et al. 2012), the cotton leafworm (Chabaane et al. 2015), the brown marmorated stink bug (Rondoni et al. 2017), and the bagrada bug (Felipe-Victoriano et al. 2019). Testing this hypothesis also will indicate if C. cactorum is providing a food-web subsidy to native parasitoids (Pearson and Callaway 2003), which increases their population densities and thus causes declines in native cactus moth abundances via apparent competition (Carvalheiro et al. 2008; Holt 1977). Another top–down control mechanism for introduced insects is novel entomopathogens (Dara et al. 2019). Entomopathogens are predicted to be rare in low density insect populations and thus facilitate relatively patchy distributions (Anderson and May 1981). We have observed low levels of Beauveria (Hypocreales: Cordycipitaceae) fungal infection of larval Melitara in the isolated microenvironment of their closed cactus pad galleries (authors personal observations). Indeed, Mann (1969) also recognized that increases in precipitation enabled fungal growth which increased Melitara mortality via fungal infection. Ongoing and future studies documenting natural enemies of native cactus moths and C. cactorum in Texas will reveal if there is direct evidence that apparent competition or other top–down mechanisms are slowing population spread at the south-western range border in Texas.

Currently, C. cactorum host plants are O. stricta on the Texas coastline and O. engelmannii in the interior region Texas. This area has a low abundance of the small statured O. macrorhiza and introduced O. ficus-indica and C. cactorum completes development on all of these hosts (Morrison et al. 2020). Expansion at the north-eastern edge of the Texas C. cactorum range has slowed and this rate change correlates with a gradual transition from O. engelmannii as the most abundant host plant to O. humifusa, the most abundant prickly pear species in the interior of the southeastern United States. C. cactorum completes its life cycle naturally on O. humifusa in Florida (Jezorek et al. 2012), but larvae may perform poorly on this species relative to performance on O. engelmannii, which is abundant to the south and west (Jezorek et al. 2010). Novel host plants with high concentrations of secondary metabolites can reduce feeding efficiency or increase mortality of invasive caterpillars and thus act as biochemical barrier to dispersal onto novel host plants (Dean et al. 2022; Wang et al. 2020). There is significant variation in nutritional quality and defensive trait values among populations of O. engelmannii, the most widespread North American prickly pear cactus species (Morrison et al. 2020; Morrison unpublished data). Variation in biochemical defenses and other host qualities may also explain why C. cactorum does not use all of the potential host plant species in its native range, including the abundant low-growing species Opuntia sulphurea G. Don ex Salm-Dyck (Mann 1969). Another aspect that may contribute to bottom–up control of cactus moths is that Opuntia species, including O. engelmannii, hybridize (Grant & Grant 1971; Griffith 2001). Hybridization can introgress plant defensive traits and produce novel chemical cocktails that are toxic to, or deter feeding by, specialist herbivores that have coevolved to tolerate certain metabolites that are specific to the parental species (Orians 2000).

A further set of factors that may interact with Opuntia host quality could be the presence of extensive fungal and arthropod damage on cladodes. We have documented damage (unpublished) from multiple arthropods, Hemiptera: Chelinidea vittiger Uhler and Narnia femorata Stål (Coreidae), Dactylopius opuntiae Cockerell (Dactylopiidae), Diaspis cf. echinocacti Bouché (Diaspididae); Lepidoptera: five Melitara species, Dyotopasta yumaella Kearfott (Tineidae), Loxomorpha flavidissimalis Grote (Crambidae), and Marmara opuntiella Busck (Gracillariidae); Coleoptera: Moneilema armatum LeConte (Cerambycidae), and several common fungi (Colletotrichum sp. (Glomerellales: Glomerallaceae), Phyllosticta sp. (Botryospaeriales: Botryospaeriaceae), Hendersonia sp. (Pleosporales: Phaeospaeriaceae) and Stephensea sp. (Pezizales: Pyronemataceae)). In common with our observation that Melitara are more prevalent in the south-west of the study area, our observations of fungal and arthropod damage generally also follow this geographic pattern. We speculate that female C. cactorum may select healthy cladodes and avoid plants with extensive damage. Alternately, some forms of fungal or arthropod damage may induce plant defenses that suppress larval development. As such, these other plant associates may provide an alternate source that drives variation in Opuntia quality. Variation in North American Opuntia host quality and native range host usage highlights the need to further characterize Texas Opuntia chemistry (Kumar and Sharma 2020; Morrison et al. 2020) and leverage that information to test whether Opuntia species and populations at the boundary of the C. cactorum invasion front are barriers to invasion.

Documenting the distribution and spread of the Texas C. cactorum population has provided the information necessary to select biological control agent field release sites that will maximize effectiveness of the program. We will continue monitoring C. cactorum in Texas in anticipation of biological control agent field releases. Furthermore, we plan to monitor C. cactorum and introduced biocontrol parasitoid population sizes and distributions in Texas following field release to evaluate the effectiveness of the program and provide information that allows us to select different deployment sites as the program proceeds.

The C. cactorum biological control agent research program included over 20 years of studies on comparative native and introduced range biology (Bloem et al. 2005; Hight and Carpenter 2009; Srivastava et al. 2019), A. opuntiarum native range biology (Folgarait et al. 2018), candidate agent field host range testing (Varone et al. 2015) and nontarget host choice testing (Florida Department of Agriculture and Consumer Services, unpublished), and the development of scalable and repeatable mass wasp-rearing protocols (Awad et al. 2019; Goñalons et al. 2014; Varone et al. 2020). Trapping sites where C. cactorum was detected within the Texas range overlap with regions predicted to be environmentally suitable for A. opuntiarum in a recent ecological niche model (Pérez-De la O et al. 2020). This ecological niche model also predicted that the fundamental niche of C. cactorum extends west to the Pacific Coast at Baja California and south through nopal growing regions of North-Central Mexico, while the A. opuntiarum distribution is considerably more limited. Regardless of why the C. cactorum population dispersal has stalled at the time of publication, this environmental mismatch between biological control agent and target highlights the urgent need to release A. opuntiarum in the field.