Target-Specific Fire Ant Suppression for Preservation of Native Species
and Non-Target Organisms such as the Horned Lizard
Jerry L. Cook, PhD
Department of Biological Sciences, Sam Houston State University, Huntsville, Texas 77341-2116

The red imported fire ant, Solenopsis invicta Buren, has spread through the southeastern United States during the past 80 years, since its accidental introduction into Mobile, Alabama. Because Solenopsis invicta was introduced without its natural enemies, and is an aggressive predatory species, it has wrecked havoc with indigenous populations ranging from small invertebrates to man (Porter et al. 1997). This invasion has decimated the indigenous ant fauna primarily by competitive replacement (Porter and Savignano 1990). This then allowed S. invicta populations to continue to expand to the point where it is commonly found at levels of higher than 100 mounds per acre, and has been recorded at over 1,000 mounds per acre (Jerry Cook unpublished data). This high density of fire ants is capable of damaging the biodiversity of many ecosystems (Porter and Savignano 1990).

One apparent casualty of the S. invicta invasion appears to be the Texas Horned Lizard, Phrynosoma cornutum, which is now listed as a threatened species in Texas and Oklahoma. Phrynosoma cornutum relies almost exclusively on a diet of harvester ants from the genus Pogonomyrmex (Blackshear and Richerson 1999). The harvester ants are being directly impacted by S. invicta, along with many other ant species (Porter and Savignano 1990, Mann 1994, Pennisi 2000). There may also be some direct effect on the Texas horned lizard by S. invicta, since ingestion of fire ants is toxic to many species due to its alkaloid poisons (Contreras and Labay 1999). However, it is uncertain if P. cornutum will feed upon S. invicta. The effect of S. invicta on Pogonomyrmex harvester ants and the Texas horned lizard is just one example of a much greater problem that the red imported fire ant is causing for ecosystems throughout the southeastern United States.

In October 1998 a program sponsored by the Texas Army National Guard was started to determine how to best eliminate fire ants from training facilities (specifically Camp Swift, Camp Bowie, Camp Maxey, and Ft. Wolters) and not further harm the natural ecosystem. The challenge of this project was to eliminate S. invicta while not harming native species, including closely related native ants. The common practice of using broadcast baits for fire ant treatment is unsuitable for this type of project because native ants will also feed on these toxic baits and, by eliminating S. invicta, native ants could also be removed. A broadcast bait program may not only eliminates native ants that are important to the ecosystem (Folgarait 1998), but this in turn can remove the only real competitors of S. invicta. After the elimination of ant competitors, new S. invicta colonies usually move back into the empty niche and are then able to exist at much higher population levels than before the chemical treatment. The other drawback to this chemical treatment regime is that it can be expensive and put toxic chemicals into the environment. Therefore, the Texas Army National Guard program had as a goal to find a way to manage S. invicta without eliminating other native species. To evaluate the results of the program, Pogonomyrmex species populations were monitored.

The concept of this program was to find a technique to use commercially available ant baits and to apply them in a way that would not impact other species. In turn, if this program were a success, it would conserve populations of Texas horned lizards, harvester ants, and other native species. In the first year of this program, a prescription treatment method was identified that appeared to effectively manage fire ants while not impacting native ants.

This program continued in the Army National Guard fiscal year 2000 with the goal of evaluating the program developed in the previous year and to attempt to conserve harvester ant and Texas horned lizard populations. The four sites being utilized for this program all had harvester ant populations at the start of the study. The harvester ant species found at all of these sites was Pogonomyrmex barbatus (F. Smith) and Camp Swift had an additional species, Pogonomyrmex comanche Wheeler.

Pogonomyrmex comanche is becoming a rare species and may be more in danger of extinction than even P. cornutum (Jerry Cook unpublished data). This ant species was historically found in several sandy regions of central Texas and Oklahoma (Cole 1968, Taber 1998). It has now been eliminated from most of these historical regions and has only been positively identified at two current locations, Lost Pines State Park (near Bastrop, Texas) and Camp Swift. Presumably, the reasons for its decline are habitat loss and, direct competition and predation by S. invicta. The status of P. comanche was unknown before this project. Pogonomyrmex barbatus is also being impacted by the same pressures as P. comanche, but it also occurs outside the invasion area of S. invicta and has a much larger range. Thus, it is not presently threatened, but its populations are declining, as are most ants in the area invaded by S. invicta.

Materials and Methods
Sites for experimental and control plots were selected at Camp Swift, Fort Wolters, and Camp Bowie. Plots were not selected at Camp Maxey because no harvester ants were found. Each plot contained fire ant mounds and native ants including mounds of harvester ants, Pogonomyrmex sp. (in all cases the harvester ant species was P. barbatus, but at Camp Swift a site was selected that also included P. comanche). Treatment and control plots were comparable in habitat type and relative ant density. The plots were separated by at least 100 feet to provide a buffer zone between treatment and control. The plot size was at least ½ acre and was marked with stakes that remained throughout the test. At each National Guard site, there was a treatment plot and a non-treatment (control) plot.

Prior to any treatments a survey was conducted to establish ant populations. For this survey, pitfall traps and bait traps were used. Nine evenly spaced pitfall traps were used in each plot. These traps consisted of tubes submerged in the soil, so that their opening is at ground level. The tubes were partially filled with a propylene glycol solution that is safe to vertebrates that might get into the traps. The propylene glycol acted as a killing agent and preservative for invertebrates, and will not quickly evaporate. Control and test sites had pit-fall traps utilized for the same period of time. Traps were collected, taken to the laboratory and all ants were identified, counted and recorded. Bait traps consisted of putting a test tube baited with meat (Vienna sausage) onto the ground. Nine of these traps were evenly spaced and put in each of the test plots for about 45 minutes. The vials were capped, taken to the laboratory, frozen, and the ants were then identified, counted and recorded.

Prior to treatments all fire ant mounds within the plot were located and flagged. Mounds were counted and recorded. The S. invicta mounds in the treatment plots were then individually treated by sprinkling 1 to 5 tablespoons of bait product around the mound, depending on mound size. Treatments were made using Amdro® (hydramethylnon), a slow acting stomach poison. Re-treatments were made at a later date to any existing fire ant mounds within the treatment plot. Control plots had mounds identified, counted and recorded.

Surveys were made approximately 3 months later in the same manner as the pre-survey. This time frame is such that control of fire ant should be achieved, if control occurs. The results were compared with the pre-survey populations. A more meaningful comparison will be made one year after treatment, in fall 2000. This survey will be at the same time, seasonally, as the initial survey and will be a more accurate comparison to pre-treatment conditions. However, this will occur in the next year cycle and the earlier analysis was used to complete the final report for this funding year. To get an accurate assessment of this program, it will have to run for more than one year to allow ant populations to adjust to changed conditions, such as relatively fire ant free areas of the treatment plots. Plots will continue to be treated and monitored after the initial survey to allow for this continuance. Any decrease in a native ant species will be evaluated and determined if it is a probable result of the treatment method.

Results
Camp Maxey. At the beginning of this project in 1998 there were only a small number of harvester ants, Pogonomyrmex barbatus, present at Camp Maxey (1999 final project report). All of these harvester ants were in regions associated with firing ranges where a program has now been going for the past two years to control fire ants using broadcast bait applications. In the process of these treatments all P. barbatus have been eliminated. This year, I extensively surveyed Camp Maxey for additional harvester ants, but none were located.

Camp Swift. Camp Swift has the unique distinction among Texas Army National Guard training facilities of having two harvester ants, Pogonomyrmex comanche and P. barbatus. These ants are living sympatricly at one site, which was used as the test site for this program. This site is located at N 30° 16.786', W 97° 19.133'. Another population of P. comanche is located at N 30° 17.087', W 97° 18.571'. These two sites are 1.06 Km apart, but there are no other harvester ants in between. Because reproductive flights are usually less than this distance, these two populations are most likely isolated. There are several other areas that have P. barbatus in small colonies within Camp Swift, including sites near the headquarters building. A treatment for fire ants was made on April 20. There were only 8 S. invicta mounds found within the treatment plot. This is a decrease from 35 mounds in spring, 1999 and 22 mounds in fall, 1999. Previous treatments were made after the fall 1999 mound counts. The control site went from 33 S. invicta mounds in spring, 1999 to 38 mounds in spring 2000. Only 18 S. invicta mounds were observed in fall 2000, but the area was still in a severe drought, and many mounds may have been missed. All harvester ants are now gone from the control site (originally there were 6 P. comanche and 4 P. barbatus mounds). Harvester ants present in the treatment plot started in spring, 1999 at 12 colonies of P. comanche and 4 colonies of P. barbatus. By fall, 2000 these numbers had changed to 11 colonies of P. comanche and 2 colonies of P. barbatus. Mating flights of P. comanche were witnessed on June 15 at Camp Swift, but it is too early to know if any new colonies have been successfully started. It is usually the next spring before the incipient colony is large enough to be easily noticed. The first few months of a colony consist of the newly mated queen tending the first batch of brood using reserves from degenerating flight muscles as a nutrition source. There is no foraging until this first brood becomes adults around three months later. These results of mound counts in these plots are given as part of Table 5 and Table 6.

Pitfall traps collected more species than bait traps. When fire ants were present, they tended to exclude other ants from the bait traps, so bait traps can be utilized best to evaluate relative numbers of fire ants present. Pitfall traps give a better picture of the diversity, because they are a more random collecting method, collecting anything that falls into them. Results at Camp Swift are given in Table 1. The spring collection was a time when ants were very active and probably represents a more realistic picture of the ant diversity. The fall samples were collected after a severe drought and record temperature period had occurred, probably yielding unrealistically low levels of ant activity.

Table 1. Ant totals collected in traps at the treatment and control sites at Camp Swift. The numbers provided are totals from 9 pitfall traps and 9 bait traps. Note that conditions were very dry in the fall 2000 collection and comparisons are not realistic between spring and fall.

Camp Bowie. Camp Bowie has many Pogonomyrmex barbatus sites throughout the base and P. cornutum sightings are rare, but not unusual. There was no effort made to evaluate numbers of P. cornutum, but personnel at the base reported sightings. Camp Bowie is unique among the four facilities used for these tests because fire ants are not prevalent throughout the base. In fact, fire ants were found in only three regions, around the headquarters, buildings and firing ranges; around a pond and wooded area centered at N 31°38.494', W 98° 56.299'; and one S. invicta mound was sited at on open field, N31° 38.596', W 98° 56.438'. The rest of the base had no S. invicta activity observed. The second site (around the pond) had no native ants observed.

Pogonomyrmex barbatus colonies were surveyed at several sites in May 2000. The site at N31° 38.596', W 98° 56.438' had about 3 colonies per hectare of P. barbatus and many mounds of Linepithema sp. and Forelius sp. The site on the firing range had numerous harvester ants and will be reported upon below, as it was the test site for this program. A site at N 31° 36.895', W 98° 55.472' had about 10 colonies of P. barbatus per hectare and several other native ant mounds, but no S. invicta mounds. This concentration of P. barbatus is quite high for a harvester ant with large colonies. Harvester ants were also observed at several other areas throughout the base.

Harvester ant populations at the treatment site have remained stable throughout the test. In fall 1998, there were 3 P. barbatus colonies within the treatment site and those colonies have remained throughout the test. The control site had one P. barbatus colony, which is also still present in 2000. Solenopsis invicta colonies fluctuated as reported in Table 6. The survey of ant species at Camp Bowie is given in Table 2. The site used for this project has had relatively high densities of fire ants for many years (Jim Hillegus personal communication). This site did not have a large biodiversity of ant species collected in the traps.

Table 2. Ant totals collected in traps at the treatment and control sites at Camp Bowie. The numbers provided are totals from 9 pitfall traps and 9 bait traps. Note that conditions were very dry in the fall 2000 collection and comparisons are not realistic between spring and fall.

Ft. Wolters. Harvester ants at Ft. Wolters are all Pogonomyrmex barbatus and are found in densities less than occur at Camp Bowie or Camp Swift. Throughout the base the highest P. barbatus densities were about 5 per hectare (which is still considered a high density for this species). Pogonomyrmex barbatus was found in five distinct regions of Ft. Wolters. One of the locations does not have a recorded GPS location, but was found near the firing ranges. Other locations of P. barbatus were N 32° 50.964', W 98° 02.525' ( 2 colonies or P. barbatus and about 20 colonies of S. invicta per hectare; N 32° 51.019', W 98° 02.652' (1 colony of P. barbatus and about 40 colonies of S. invicta per hectare); N 32° 52.122'm W 98° 02.479' (5 colonies of P. barbatus and about 10 colonies of S. invicta per hectare; and N 32° 51.405', W 98° 01.874' (1 colony of P. barbatus and about 40 colonies of S. invicta per hectare). These numbers will support P. cornutum populations and there have been several sightings at Ft. Wolters (Sgt. York personal communication).

The bait test at Ft. Wolters showed a decrease in fire ant colonies from 54 colonies to 12 colonies in the first year of the project. In spring 2000, S. invicta colonies had increased to 26 colonies and again reduced to 6 in late summer 2000 (however this last figure may be an underestimate due to extreme weather conditions). The control site ranged around 40 colonies, except for the final summer survey of 14 mounds. The number of P. barbatus colonies in the treatment plot has remained at two throughout the test. In the control plot, P. barbatus started with one colony, which was lost in the summer of 2000. Results of the prescription bait treatment program are given in Tables 5 and 6. Ant surveys collected at baits and in pitfall traps are shown in Table 3.

Overall summary
The prescription bait treatment program has the cumulative results listed in Table 4. These results are provided to survey the effectiveness of the protocol. Solenopsis invicta was reduced by prescription treatments by an average of 92 percent from spring, 1999 to fall, 2000; although this result may partially be a result of weather. A more accurate approximation of the reduction would be somewhere between a 58 and 92 percent reduction. In these same treatment plots about 6 percent of the harvester ant mounds were lost. In untreated control plots fire ant levels changed in concentration somewhere between declining by 58% to increasing by 9 percent. In control plots, about 67 percent (an average of the sites) of harvester ant colonies were lost when insecticide treatment was not provided. Since many more harvester ants were found at one site (Camp Swift) in control plots the actual percent lost was about 88 percent of all Pogonomyrmex species in the non-treated plots.

Table 3. Ant totals collected in traps at the treatment and control sites at Ft. Wolters.

Table 4. Percent change in mound densities between treatment and untreated control plots. Solenopsis invicta and combined Pogonomyrmex species were evaluated from spring 1999 to fall, 2000. Results for S. invicta may be skewed downward because of dry conditions making it difficult to identify mounds. Therefore, columns with "*" are determined from spring 1999 to spring 2000, just prior to the spring treatment. Results for the Pogonomyrmex sp. should be accurate because even in the extreme weather conditions, these ants were active and colonies could be accurately identified.

Table 5. Number of harvester ants in plots at three Texas National Guard facilities. Camp Swift results include both Pogonomyrmex barbatus and P. comanche. Camp Maxey is not included because there are no remaining harvester ants.

Table 6. Number of Solenopsis invicta mounds in plots at three Texas National Guard facilities. Camp Maxey is not included because there are no remaining harvester ants for continuation of the project.

* Fall 2000 results were collected in September, when it was still extremely dry, making it possible that S. invicta mounds may have been missed.

Back to Results

Discussion
Some of the results for 2000 are rather misleading because of the severe drought and high temperatures of late summer and early fall. It is unfortunate that the timing of this report was such that it is hard to make sense of the fire ant numbers. However, it is clear that there has been a decline in S. invicta numbers from spring 1999 to fall 1999, and even though there was naturally some rebound in colony number before the spring 2000 treatment, these numbers were lower than spring 1999. Actual numbers of S. invicta colonies will likely be even much different between treatment and untreated control sites, when the weather conditions make it easier to count colonies. The observed fall 2000 numbers are almost certainly more than are reported, but mounds were simply not apparent on the surface. This is also demonstrated in the numbers of S. invicta colonies appearing to reduce by over 50 percent in the untreated control colonies. In reality, these numbers are probably very close to the spring mound counts.

The results that are encouraging are those of harvester ant conservation. There has been a loss of three Pogonomyrmex barbatus colonies at Camp Swift during the treatment time, but 94 percent of all Pogonomyrmex colonies have been conserved. This loss may be a natural fluctuation, or it could be a result of the remaining fire ants within the plot. However, mating flights were observed this year and if any of these were successful, they should be apparent by later this fall, or certainly by spring 2001. If some of these are able to establish it may be because predation and competition by fire ants are becoming less in the treatment plots.

One outcome of this project that has already been demonstrated is that broadcast bait treatments are detrimental to harvester ants, and probably to other native ants also. At Camp Maxey, the only harvester ants were associated with disturbed areas on the firing ranges. These areas were part of another program to treat fire ants along these ranges using less labor intensive broadcast baiting. The harvester ants at these areas were eliminated, along with many fire ants. However, the results are that fire ants return (although at lower numbers if the bait application is regularly applied), but harvester ants do not. Other native ant species are probably not found in these treatment areas after the two years of broadcast baitings. The result is that these treatments must be continued or fire ants will soon be found in numbers even larger than before the start of the program because there is no longer competition from other ant species.

Surveys of other ants in both the prescription bait treatment plots and their associated control plots do not really mean much at this time. They do show that the treatment program does not appear to be affecting large numbers of native ants, but even this is hard to say at this time. These results were gathered to provide a good starting point for the evaluation of this program in regards to affects of the entire ant population. Because the fall results were skewed by weather conditions, comparisons are really meaningless. The biodiversity in these small plots is ample for good evaluations of this treatment program and its effects on native ants and will provide numbers needed for a good statistical analysis in next years program.

National Guard management policies do not have to be changed to maintain the work on this project and preservation of harvester ants and horned lizards. Harvester ants and horned lizards prefer disturbed habitats. Training exercises should not greatly affect the maintenance of these populations. However, other native species may be more adversely affected. Thus far, there is little information on the effects on other native ant species, but this information may be forthcoming in next years project. An important concern is the use of pesticides in these areas. Many native ants, such as harvester ants, are much more susceptible to insecticides than the red imported fire ant. Treatment strategies other than prescription bait treatments will almost certainly impact harvester ants and in turn, impact populations of the Texas horned lizard. As stated earlier, elimination of all ant species in an area makes it easier for re-establishment by red imported fire ants. To maintain native ants, the only insecticide use should be such that it is limited to the target population.

References Cited
Blackshear, S. D. and J. V. Richerson. 1999. Ant diet of the Texas horned lizard (Phrynosoma cornutum) from the Chihuahuan desert. Texas Journal of Science, 51: 147-152.

Cole, A. C. 1968. Pogonomyrmex Harvester Ants: A Study of the Genus in North America. The University of Tennessee Press, Knoxville, TN. 222 pp.

Contreras, C. and A. Labay. 1999. Rainbow trout kills induced by fire ant ingestion. Texas Journal of Science, 51: 60-61

Folgarait, P. J. 1998. Ant biodiversity and its relationship to ecosystem functioning: a review. Biodiversity and Conservation, 7:1211-1244.

Mann, C. C. 1994. Fire ants parlay their queens into a threat to biodiversity. Science, 263: 60-61.

Pennisi, E. 2000. When fire ants move in, others leave. Science, 289: 231.

Porter, S. D. and D. A. Savignano. 1990. Invasion of polygyne fire ants decimates native ants and disrupts arthropod community. Ecology, 71: 2095-2106.

Porter, S. D., D. F. Williams, R. S. Patterson, and H. G. Fowler. 1997. Intercontinental differences in the abundance of Solenopsis fire ants (Hymenoptera: Formicidae): escape from natural enemies? Environmental Entomology, 26: 373-384.

Taber, S. W. 1998. The World of the Harvester Ants. Texas A&M University Press, College Station, TX. 213 pp.

Back to Materials and Methods

Back to Results

Back to Discussion

Back to Contents

Click  to view the entire report in ".pdf" format so the document can be easily and accurately printed.. Adobe Acrobat Reader will launch automatically. If you do not have Adobe Acrobat Reader, click