CONTROL OF VAMPIRE BY TOPICAL APPLICATION OF AN , CHLOROPHACINONE’

Samuel B. Linhart, M.S.;2 Rad Flores Crespo, B.S.;3 and G. Clay Mitchell, Ph.D4

Common vampire bats were mist-netted, topically treated with an anticoagulant paste (chlorophacinone), and released. They returned to their roosts, where they contaminated other vam- pires. Control by this method in caves and on ranches reduced local vampire populations by about 97 per cent.

Introduction netting effective in large-scale control efforts. They normally roost in numerous scattered sites, It is estimated that one million head of Latin most of which are hard to locate and contain America’s cattle die annually from trans- relatively few bats (the common vampire’s cave- mitted by vampire bats ( rotundus, type sites usually contain between 10 and 200, Diaemus your@, and Diphylla ecaudata), whose while tree roosts generally have between 5 and range extends from northern Argentina to tropi- 50). In addition, vampires may frequently move cal Mexico. Effective and economical methods from site to site (1,2) and commonly share their are needed to reduce populations in roosts with beneficial bat species. Thus, selective those areas where bat-borne rabies is a problem. control within the roosts is difficult even when In 1968, a cooperative program to develop such these refuges are found. control methods was undertaken jointly by the However, certain behavioral characteristics of Mexican Agriculture and Livestock Ministry’s Desmodus rotundus suggesteda new approach to National Institute of Livestock Research and the control. These characteristics are as follows: ’ Bureau of Sports Fisheries and (a) The bats groom themselves intensively; Wildlife, under a sponsoring agreement with the captive spend an average of two to three U.S. Agency for International Development. The hours a day in this activity (3,4,5). In grooming, program has concentrated on ways of controlling vampires characteristically scratch the body the common vampire bat (Desmodus rotundus), rapidly with a hind foot, inserting the foot into the principal vector of the disease. the mouth every S-10 seconds; they also clean Vampire bats pose unusual control problems. their wing membranes and thumbs with the The animals are nocturnal and highly mobile, and tongue (4,5). their roosts are frequently located in rugged, (b) The bats use the same small niches within heavily vegetated terrain. Because they feed ex- their roosts (usually a one to three foot crevice or clusively on the fresh blood of live vertebrates, concavity in the ceiling of a cave, mine, etc.) for conventional control techniques cannot be used; extended periods of time. They roost very close and they rarely concentrate at feeding sites in together, and body contact is normal between sufficiently large numbers to make trapping or vampires sharing the, sameroosting site. (c) Although other bat species may share the ‘Also appearingin Spanishin Boletin de la Oficina Sanitaria Panamericana, Vol. LXXII, No. 2 (August, same roost, they use different niches. We have 1972). neither observed nor seen reported any instance 2Wildlife Biologist, Denver Wildlife Research of vampire bats sharing a niche with any other Center, Bureau of Sport Fisheries and Wildlife, U.S. Department of the Interior. bat species. 3Biologist, Instituto National de Investigaciones (d) Although it is difficult or impossible to Pecuarias, Secretaria de Agricultura y Ganaderi’a, capture the entire population of vampire bats Mbxico, D.F., Mexico. 4Wildlife Biologist, Bureau of Sport Fisheries and present in a given area, we have found it quite Wildlife, U.S. Department of the Interior. easy to capture small numbers at specific feeding 31 32 ENGLISH EDITION-BOLETIN DE LA OSP . Vol. VI, No. 2, I972 sites by placing mist nets around tethered or concavity for roosting. Metal window screen was corralled cattle. attached to the basin “roof’ so as to supply a These behavioral characteristics suggested surface from which the bats could hang. The bats that release of a few mist-netted vampires treated were permitted to feed at will on defibrinated with a slow-acting toxicant could have a signifi- bovine blood, provided daily in plastic feed- cant impact on the bat population. The topically ers attached to the side of the box. applied toxicant would not prevent them from A 25watt red light needed for television returning to their roosts, and would be trans- transmission was placed in the box; observations ferred by contact to other vampires using the by closed-circuit television confirmed that same roosting sites. Individual and mutual grooming behavior within the colony was similar grooming would then result in ingestion of the to that reported by Flores Crespo et al. (4) and toxicant. This technique would not require Greenhall (5). One night the bats’ activities were capturing large numbers of vampires or finding recorded at half-hour intervals, from 7:30 p.m. their roosts, and would probably be safe for to 12:30 a.m. At each interval during this period, other bat species using the same refuges, To from one-third to one-half of the 20 bats were evaluate the potential of this technique, the labo- observed grooming. Despite considerable unoc- ratory and field tests described in the pages that cupied space in the inverted metal basin, follow were carried out. grooming commonly took place with the bats very close together or touching each other, and Laboratory Tests body contact was frequent. Several instances were also noted of bats licking each other, and of The toxicant chosen for evaluation was one bat folding its wings about another. These chlorophacinone (2-[(p-chlorophenyl)phen- observations indicated that any material adher- ylacetyll-1,3-indandione),-a slow-acting anti- ing to the body of one colony member would coagulant originally developed in France for con- promptly spread to the others, who would soon trolling rats and mice. The chlorophacinone was ingest it. suspended in 1 per cent gum tragacanth in saline solution and administered by gavage.The median Selection of a Carrier lethal dose (LDse) was determined by the Thompson-Weil method (6, 7). The acute oral Tests were then conducted to select a carrier LDs ,-,for Desmodus was found to be 3.06 mg/kg for the toxicant that would adhere well to the (95 per cent confidence limits: 0.97-9.64 treated bat, transfer easily on contact, and be mg/kg). The highest dose administered (140 ingested during grooming. Ten candidate carriers mg/kg) required 1.5 days to kill all bats; the were mixed with an inert fluorescent powder lowest doses (3.7 and 2.5 mg/kg) required 9-16 (either an acrylic polyester paint pigment or a days and several bats survived. commercial metallic tracer) so that persistence and transfer of the material could be easily deter- Observation of Grooming Behavior mined by scanning with ultraviolet light. Each carrier was applied to the back of an individually Laboratory tests confirmed that material caged bat; each bat and its cage were examined applied to one vampire would be transferred to under ultraviolet light several times a day for two other vampires with which it roosted. Twenty or three days. The three most promising carrier Desmodus were placed in a simulated roost in the mixtures tested (petroleum jelly and the fluores- form of a plywood box measuring 25 in. x 26 in. cent paint pigment; a saturated solution of the x 36 in.; the box, lined with plastic sheeting to pigment in acetone; and a sticky mixture of the facilitate cleaning, was provided with a clear plas- pigment in mineral oil, rosin, and polyethylene) tic door for observation. A circular hole in the were then introduced into the 20-bat colony. In top was covered with an inverted metal basin 13 each test, one bat was removed and about 1.5 ml inches in diameter and 6 inches deep, to provide a of the carrier was smeared on its back; it was Desmodus rotundus, the common vampire bat. returned to the cage, and the colony was exam- and the uropatagium. The clear plastic door of ined under ultraviolet light several times a day for the cage was covered with opaque paper to re- two days. duce external stimuli, and the treated bat was The flourescent marker in both the petroleum reintroduced onto the floor of the box. The jelly and the acetone-pigment mixture was trans- colony was then checked daily for 31 days to ferred to all or nearly all of the 20 bats in the record mortality. colony within 24 hours. Fluorescence was most The results of this test are summarized in noticeable on the forearms, legs, and feet. The Table 1. The treated bat was found dead the acetone-pigment mixture gradually powdered or TABLE l-Deaths in captive colony of 20 flaked from the back of the treated bat; this vampires bats, following treatment of one bat promoted a very uniform distribution of mate- with 50 mg of chlorophacinone dissolved in 1.5 rial, but also resulted in some losses;considerable ml of petroleum jelly. amounts of pigment were found on the floor of Days Cumulative after Number percentage the box. The petroleum jelly, which was not so treatment of deaths of dead bats uniformly distributed but which adhered very well to the fur, was therefore selected for use 1 1 (treated bat) 5 2-4 0 5 with the toxicant. 5 4 25 6 1 30 Test with Chlorophacinone 7 5 55 8 1 60 Fifty milligrams of chlorophacinone were 9 4 80 dissolved in 1.5 ml of melted petroleum jelly, and 10 0 80 11 1 85 the solution was permitted to cool until it solid- 12-13 0 85 ified. A single bat was removed from the 20-bat 14 2 95 colony, and the chlorophacinone paste was Total 19 95 applied to its back between the scapular region 34 ENGLISH EDITION-BOLETIN DE LA OSP . Vol. VI, No. 2, 2972

Applying chlorophacinone paste to the back of a Desmodus r.

following morning, and 18 of the remaining 19 tion, in a mixed agricultural and forested area bats died in 5-14 days. Blood consumption by about one mile west of the village of Tepoztlan, the colony was reduced the first 2 days, but municipality of Tepoztlan. There are two en- thereafter stayed about normal for the number trances but one is inaccessible. The single of live bats remaining. Dead bats showed signs of roosting site within the cave was estimated to subcutaneous hemorrhaging, characteristic of contain 50-60 common vampire bats; no other anticoagulant poisoning, on the extremities and bat specieswere present. on the body where the patagium5 joins the A six-meter, four-shelf mist net was placed in trunk. On the basis of these encouraging results, front of the cave entrance at 8:35 p.m., about 1 field investigations were undertaken in Mexico to hour after dark, on 18 March 1971. We initially test the technique on wild vampires netted at planned to net and treat vampires as they re- cave entrances and around cattle corrals. turned to the roost after feeding, thus simulating the type of situation anticipated on cattle Field Work ranches. In this instance, however, the net was erected too early in the night, and all bats cap- Cavetests tured were leaving the cave to forage. Since we felt a minimum of disturbance was desirable, we Two cavesin the State of Morelos were initial- removed the net at 9:50 p.m., by which time 16 ly chosen as study sites. (Annual precipitation in vampires had been captured. Six were released, this area is 20-40 inches.) Cueva de.los Muiiecos, and the remaining ten were left in a cage inside one of those selected, is a small cave measuring the entrance until 7:15 the following morning. roughly 85 ft. x 40-50 ft. x 6-10 ft. It is located At that time, six of the ten (the former com- on the side of a cliff at about 5,400 feet eleva- prising about 10 per cent of the number esti- mated to be roosting in the cave) were each ‘Wing membrane. treated with a mixture of 1.5 ml petroleum jelly Linhart et al. . CONTROL OF VAMPIRE BATS 3.5 and 50 mg chlorophacinone, as previously de- nine of which were returning to the cave en- scribed; all ten were then released back into the gorged with blood. Six of the engorged bats, or cave. about 10 per cent of the vampires estimated Eight days after treatment we found only using the cave, were each treated with the chloro- 12-15 live vampires in the cave and noted a phacinone paste. All 13 bats were then banded marked difference in their evasive behavior. and released into the cave. Vampires normally attempt to hide by flying, Eight days after treatment, we found a num- hopping, or crawling to other parts of the cave ber of dead vampires on the floor or hanging when a light shines on them. However,all but one from the roof and walls of the cave. Some were in of the 12-15 vampires we saw were apparently or directly beneath the roost sites, and others unable to fly and made no attempt to flee from were up to 56 feet away. We found three dying their original roosting site when a headlamp vampires unable to fly; two were hemorrhaging beam was shone on them for S-10 minutes. Their externally. An aborted fetus was found directly movements were slow, and they had difficulty beneath one of the roost sites. Fourteen days climbing. A number ofvampire wings were found after treatment we were able to find only one live on the floor of the cave, but we were unable to and apparently normal vampire, although the identify the predator or scavenger that appar- size of the cave and the presence of numerous ently ate dead or dying bats as they fell from the bats of other species made visual estimates roost. Fourteen days after treatment, only one difficult. live vampire was found in the cave. The total A total of 94 dead vampires was found in this wings and intact carcasses recovered from this cave. This is a minimum estimate of mortality, cave represented at least 27 dead vampires; but since a few dead bats could have been missed in since the upper portions of the cave were inacces- the rock rubble covering the cave floor, and some sible and rock rubble on the floor made recovery may have died while foraging outside the cave. of wings difficult, it was impossible to determine Forty-two of the dead vampires were later exam- the total number of bats killed. ined in the laboratory, and all showed evidence Cueva de1 Salitre, the other cave selected, is of hemorrhaging on the body, forearms, or wing located at about 5,300 feet elevation in the membranes. A thorough search of the cave failed municipality of Emiliano Zapata, about eight to reveal dead bats of any other species, other miles south of Tezoyuca on the road to Zacate- than old skeletons or severely desiccated pet, State of Morelos. It has a single entrance and remains. is a long, very large, Y-shaped cave located be- In this cave, application of the anticoagulant neath the valley floor, and measuring roughly resulted in the death of at least 15 vampires for 870 ft. x 2540 ft. x 13-20 ft. Brush,small trees, each vampire treated. This result, in a large cave and cultivated sugarcane are the dominant vege- with several vampire roosting sites, clearly tation in this locality. We originally estimated demonstrated the soundness of our original con- that about 60 common vampire bats were cept. No bats of other species appeared to have roosting in the cave, but subsequent mortality been affected, suggesting that the technique is counts showed this estimate was too low. Sight- specific for vampires. ings of bats and presence of fresh excrement Several months after the above work was beneath roosts indicated the vampires were using completed, one of us (Flores Crespo) treated a five distinct roosting sites within the cave. Up to third cave (Cueva de1 Puente Negro) located a thousand or more bats of other species, in- about four miles from the town of Coquimatlln cluding Natalus stramineous, Pteronotus in the State of Colima. This cave, measuring parnellii, and jamaicensis, were also approximately500 ft.x IO-100 ft. X 1 S-20 ft., roosting throughout the cave. contained an exceptionally large Desmodus col- A 6-meter mist net was placed at the cave ony estimated at about 2,000 individuals located entrance at 11:OO p.m. on 19 March 1971. in two roosting sites. In addition, several hundred Thirteen vampires were captured by midnight, thousand other bats, including Pteronotus davyi, 36 ENGLISH EDITION-BOLETIN DE LA OSP Vol. VI, No. 2, 1972 ’

P. psilotis, P. rubiginosa, soricina, ing, two or three observers recorded the location and waterhousii, were present. On 24 and number of fresh vampire bites on each ani- + July 1971, 90 Desmodus were captured in one mal, and all animals were returned to pasture. . 6-meter mist net, treated with chlorophacinone The same animals were corralled again in late ’ paste, and released back into the cave. The cave afternoon, and the corral was surrounded by was revisited eight days later on 1 August, and an about 100 linear meters of mist netting two estimated 2,000 dead vampires were recovered. meters high. All vampires caught during the night The ratio of treated bats to total dead vampires were caged, banded, treated with the chloro- .. recovered was therefore approximately 1:22. No phacinone p&e, and released at least 45 minutes dead bats other than Desmodus were found. before dawn. Release well before dawn is impor- tant, as we have observed that vampires released Ranch tests at sunrise or during daylight hours, in addition to being harassed by , seek the nearest dark Bats attacking livestock at Ranch0 Don refuge and probably do not return to their +. Tomas were treated on 15 March 1971. Ranch0 normal roosting site until dark. Don Tomas is a 2,500-acre, intensively managed Ranch0 Don Tomis was visited 15 days after ’ ranch in the eastern foothills of the Sierra Madre treatment and Ranch0 San Ricardo 13 days after Oriental. Situated at about 700 feet elevation,it treatment. The cattle were again corralled, bite is near the village of Micas in the municipality of counts were made, mist nets were erected, and all Valles, State of San Luis Potosi. Part of the ranch vampires caught were banded and released. 1’ is situated within a long, narrow, partially cleared There is evidence that vampires prefer to for- valley about 9-13 x 0.6 miles in size, on either age during the darkest hours of the night when ’ side of which are heavily forested hills 650-l ,300 there is no moonlight (I, 8). This factor was feet high. A fenced pasture in the valley con- considered. At Ranch0 Don Tom&, there were 4 tained about 140 Zebu cows and calves that were l/4 hours of complete darkness the night of subjected nightly to moderate vampire attack. treatment and 8 l/4 hours the night of post- This herd was selected for study. Mature cattle treatment evaluation; at Ranch0 San Ricardo, it located a few kilometers down the valley were was completely dark all night long on both A attacked very lightly, and a check of other stock nights. It would therefore seem unlikely that in adjacent valleys revealed no fresh vampire variations in vampire foraging behavior due to bites. We were unable to find vampire roosts in moonlight influenced our results. the vicinity of the study area, although radio- The results of these tests are summarized in telemetry showed that they were located on an Table 2. At both ranches, the number of bites adjacent hillside. recorded and the number of vampire bats netted rt The second ranch, Ranch0 San Ricardo, was had decreased by 95 per cent or more two weeks ! treated on 24 March 1971. This is a 250-acre after treatment. At Ranch0 Don Tomas, a cow- dairy ranch in the northern part of the State of boy reported finding one dead vampire and a Oaxaca; it has less than 660 feet elevation, and is second unable to fly at a corral about two-thirds located near the village of Vicente in the munici- of a mile from the treatment site. He said that pality of Acatlan de Perez Figueroa. About 60 blood was present on both bats, which suggested i adult crossbred cattle, subjected nightly to hemorrhaging from exposure to the anticoagu- , moderate vampire attack, were used for the lant. A letter from the owner of Ranch0 San study (calves were placed in illuminated stalls at Ricardo, about a month after the test there night and were not attacked). We made no ended, stated that vampire attack had been re- attempt to find local vampire roosts, since an duced on several neighboring ranches as well as earlier search by other investigators had been his own, indicating that such control may extend ’ unsuccessful. over a fairly wide area. The same procedures were followed at both In more recent field tests, another anticoagu- ranches. Cattle were corralled early in the morn- lant, diphenadione (2-diphenylacetyl-1, Linhart et al. . CONTROL OF VAMPIRE BATS 37

TABLE 2-Results of treating vampire bats netted at cattle corrals with chlorophacinone/petroleum jelly (50 mg/1.5 ml per bat).

Ranch0 Don Tomis Rancho San Ricardo

15 days 13 days Before after Before after treatment treatment treatment treatment

Number of cattle examined: Adults 71 61 56 59 Calves 46 62 0 0 Number of fresh vampire bites: Adults 16 1* 37 1 Calves 53 0 Average number of bites/: Adults 0.2 0.02 0.7 0.02 Calves 1.2 0 Number of vampires netted: 51 3 34 1 Number of vampires treated and released: 54 34 Per cent reduction in vampire bites: Adults 93.6 97.3 Calves 100 Per cent reduction in vampires netted: 94.7 97.1

*One other animal had a wound, but it could not be definitely attributed to vampire attack.

3-indandione), has also been very effective in be effective. appear ideal for this reducing vampire populations on ranches in purpose, but considerations of economy, effec- Mexico and Brazil. This compound, used in the tiveness, and safety should dictate the final same manner and at the same concentration as choice. Whatever toxicant is selected, it should the chlorophacinone, has resulted in 95 to 100 be thoroughly tested in the laboratory and under per cent reduction in vampire bat bite counts. a variety of field conditions; sites tested should also include areaswhere vampires normally roost Discussion and Recommendations in trees. Ideally, such studies should be done on The concept of controlling vampire bats by ranches where vampire attack is severe and mist-netting a few individuals, treating them with roosting sites are known. Treatment and release anticoagulant, and releasing them to contami- of captured bats near cattle could then be corre- nate other vampires within their roost appears lated with subsequent mortality at roosts. The very promising, in view of these preliminary duration and extent of control should be investi- results. Our tests on two ranches resulted in gated at regular intervals after treatment by bite reductions in vampire activity exceeding 95 per counts and mist-netting, both at the treatment cent, following one night’s work by a crew of site and in surrounding areas. three men. The technique thus appears both Further studies on the technique’s safety more effective and far less time-consuming than should also be conducted. Because bats as a conventional control by roost treatment or mist- group are highly beneficial in insect control and netting alone. (Schmidt etal. (9), working at two in other ways, there is a particular need for data ranches in Puebla and Oaxaca, Mexico, report confirming that the procedure does not harm that control at roosts and intermittent netting non-vampire species sharing the same roosts. In around cattle corrals over a six-month period this regard, persons using the technique should reduced vampire bites by 70-80 per cent.) also be instructed in proper use of mist nets, Our tests indicate that chlorophacinone and safe bat-handling procedures, and identification diphenadione will control vampire bats; theoreti- of vampires-so that beneficial species caught in cally at least, any relatively potent, slow-acting mist nets are not accidentally treated with toxicant combined with a suitable carrier would toxicant and released. 38 ENGLISH EDITION-BOLETIN DE LA OSP . Vol. VI, No. 2, I972

Many toxicants, including anticoagulants, are them to contaminate other vampires within potentially hazardous to predators and scaven- their roost was investigated in kdbOmtOry and gers that eat poisoned animals. Our observations field tests with common vampire bats (Des- modus rotundus). In these tests, a mixture of in Cueva de 10s Mufiecos showed that one or about 1.5 ml of petroleum jelly and 50 mg of more species will eat dead or dying vampires. chlorophacinone (2-[(p-chlorophenyl) phenyl- Therefore, attempts should be made to deter- acetyl]-1,3-indandionex an anticoagulant, was mine the hazards, if any, of secondary poisoning. applied to each bat. In the laboratory, applying Finally, we recommend that this technique this mixture to 1 of 20 vampires in a captive colony killed 19 of the 20 bats in 14 days. not be used for bat control in the few local areas Field tests in Mexico with bats captured, (such as parts of northern Argentina) where vam- treated, and released at three caves resulted in pires roost in wells. Because contamination of the death of nearly all vampires in each colony. well water could result in a hazard to human In two of the caves, at least 15-22 vampires health, other control techniques, such as fumiga- were killed for each one treated; no mortality was observed among other bat species seen tion of the wells, should be investigated. roosting in the caves. In tests at two ranches, mist nets were placed around corralled cattle Summary for one night, and every vampire captured was treated and released. At both locations, both The concept of controlling vampire bats by the number of vampires netted and the number mist-netting a few individuals, applying a slow- of vampire bites on cattle had declined by more acting toxicant to their backs, and releasing than 95 per cent two weeks after treatment.0

ACKNOWLEDGEMENTS

The Chempar Chemical Company, Inc., 260 studies on their ranches. Gilbert Holguin, Luis Madison Avenue, New York, New York, 100 16, Huerta, and Antonio Vargas provided labora- kindly provided the chlorophdcinone. Diphe- tory and field assistance, and Ann Jones edited nadione was provided by the Velsicol Chemical the manuscript. Corporation, 341 East Ohio Street, Chicago, This research was supported by funds pro- Illinois, 60611. We thank Mr. Malcom Niven vided to the U.S. Bureau of Sport Fisheries and and Sr. Miguel Angel Perez, owners of Ranch0 Wildlife by the U.S. Agency for International Don Tom& and Ranch0 San Ricardo, respec- Development under PASA RA(ID) l-67. tively, who were extremely helpful during the

REFERENCES

(1) Villa, R., B. Los murcXagos de MPxico. Institute (6) Thompson, W. R. “Use of Moving Averages and de Biologia, Universidad National Autonoma Interpolation to Estimate Median-Effective de Mexico (Mexico D. F., Mexico), 1966. Dose. I. Fundamental Formulas, Estimation of (2) Wimsatt, W. A. “Transient Behavior, Nocturnal Error, and Relation to Other Methods.” Bact Activity Patterns, and Feeding Efficiency of Rev II: 115145,1947. Vampire Bats (Desmodus rotundus) Under (7) Weil, C. S. “Tables for Convenient Calculations of Natural Conditions.” JMammal50(2) :233-244, Median-Effective Dose (LD5o or EDso) and 1969. Instructions in Their Use.” Biometrics (3) Flores Crespo, R., R. J. Burns, and S. B. Linhart. 8:249-252, 1952. “Comportamiento de1 Vampiro (Desmodus (8) Crespo, J. A., J. M. Vanella, B. D. Blood, and J. rotundus) durante su alimentacibn en ganado M. de Carlo. “Observaciones ecologicas de1 vam- bovino en cautiverio.” Tee Pecuar Mex piro (Desmodus rotundus rotundus) (Geoffroy) 18:40-44, 1971. en el norte de Cordoba.” Revista Museo Argen- (4) Flores Crespo, R., S. B. Linhart, and R. J. Burns. tino Ciencias Naturales “Bernardino Rivadavia” “Comportamiento de1 vampiro (Desmodus e Institute National de Investigation de las rotundus) en cautiverio.” The Southwest Natur, Ciencias Naturales. Ciencias Zoologicas 19 72. (in press.) 6(4):131-160,1961. (5) Greenhall, A. M. “Notes on Behavior of Captive (9) Schmidt, U., A. M. Greenhall, and W. Lopez- Vampire Bats.” Mammalia 29(4) :44 I-451, Forment. “Vampire Bat Control in Mexico.” 1965. Bijdr. Dierk. 40(I) :74-76.1970.