Journal of Chemical Ecology, Vol. 15, No. 5, 1989
RESPONSES BY CANIDS
TO S C E N T G L A N D
S E C R E T I O N S OF T H E W E S T E R N RATTLESNAKE
DIAMONDBACK
(Crotalus atrox)
PAUL J. WELDON l and DANIEL B. FAGRE2 Departments of 1Biology and 2Wildlife and Fisheries Sciences Texas A&M University College Station, TX 77843 (Received December 22, 1987; accepted July 28, 1988)
Abstract--Many snakes discharge malodorous fluids from paired scent glands in the base of the tail when they are disturbed. A number of authors suggest that these secretions repel predators. Scent gland secretions of the western diamondback rattlesnake (Crotalus atrox), or dichloromethane extracts of them, were presented to coyotes (Canis latrans) in three field tests, and to domestic dogs (Canis familiaris) in two kennel tests, to determine whether responses of possible benefit to snakes are elicited. Free-ranging coyotes visited and robbed and rolled at stations containing scent gland secretions in perforated plastic capsules more frequently than at those containing untreated or dichloromethane-treated capsules. Responses to dichloromethane extracts of scent gland secretions subjected to rotary evaporation were not significantly different from those to dichloromethane. Pure and mixed breeds of dogs presented with filter papers treated with dichloromethane or a dichloromethane extract of scent gland secretions mouthed (licked, bit, or ate) secretion-treated papers more frequently. Staffordshire terriers presented with filter papers treated with dichloromethane or dichloromethane extracts of snake scent gland and alligator (Alligator mississippiensis) paracloacal gland secretions exhibited urination postures to snake secretion-treated papers more frequently than to dichloromethane-treated papers, but responses to snake- and alligator-treated papers did not differ significantly. There was no indication that canids are repelled by scent gland secretions. Key Words--Canidae, coyote, Canis latrans, dog, Canisfamiliaris, rattlesnake, Crotalus atrox, scent glands.
1589 0098-0331/'89/0500- [ 589506.00/0 9 1989 Plenum Publishing Corporation
1590
W~LDONAND FAGRE INTRODUCTION
Reptiles produce a variety of secretions thought to repel predators. In only a few stUdies, however, have their natural products been tested for antipredator activity (Eisner et al., 1977; Kool, 1981; Weldon, 1988). Scent glands are paired exocrine organs in the tail of all snakes (Bellairs, 1950; Whiting, 1969). These glands typically release fluids, which are thought to discourage predators (Fitch, 1960, 1965; Burkett, 1966; Klauber, 1972), from two openings at the posteriolateral margin of the cloaca when snakes are disturbed. Several species of ants (Gehlbach et al., 1968) and several snakes (Watkins et al., 1969), including ophiophagous species, are repelled by the scent gland secretions of the Texas blindsnake (Leptotyphlops dulcis). A preliminary test indicated that some North American carnivores are reluctant to approach or eat food treated with desert kingsnake (Lampropeltis getulus splendida) secretions (Price and LaPointe, 1981), but there are no quantitative tests of the reactions by mammalian predators to these substances. Rattlesnakes (genus Crotalus) typically discharge, and sometimes spray, scent gland fluids when molested (Klauber, 1972; Duvall et al., 1985). Since many mammals are resistant to snake venom (Perez et al., 1978), defenses other than striking and envenomation--such as scent gland discharge--may be useful against these predators. Canids are one group of carnivores with which rattlesnakes interact. We report here tests of the reactions of coyotes (Canis latrans) and domestic dogs (Canis familiaris) to scent gland secretions of the western diamondback rattlesnake (Crotalus atrox). This study was conducted to determine whether responses of possible benefit to snakes are elicited in canids by scent gland secretions. Our field tests were designed primarily to test coyotes, although data on other vertebrates visiting our scent stations are presented. The series of field experiments involved different methods of chemical sample preparation or presentation, since the results of each test suggested questions that were addressed by modifications in subsequent designs. None of the field tests involved direct observation of the behavior of coyotes. Rather, inferences were made regarding coyotes' responses on the basis of impressions left in the soil or from other telltale deposits. Domestic dogs were tested to allow direct observation of canids' responses to scent gland secretions. METHODS AND MATERIALS
Coyotes Responses to rattlesnake scent gland secretions were conducted at the Chaparral Wildlife Management Area, a 6151-hectare ranch located in the Rio Grande Plains region of Texas in Dimmit and La Salle counties, and at the La
C A N I D RESPONSES T O S N A K E S E C R E T I O N S
1591
Copita Research Area, a 1093-hectare ranch located in Jim Wells County, approximately 15 km south of Alice, Texas. The average annual rainfall at the Chaparral and La Copita areas is 84 and 68 cm, respectively. The vegetation of both areas is dominated by honey mesquite (Prosopis glandulosa) and includes brazil (Condalia obovata) and prickly pear cactus (Opuntia rigidula). In addition to coyotes, eastern cottontail rabbits (Sylvilagusfloridanus), striped skunks (Mephitis mephitis), raccoons (Procyon lotor), badgers (Taxidea taxus), javelina (Tayassu tajacu), and rodents were observed at both sites at the time of testing. Five C. atrox were seen during the study in the Chaparral area. C. atrox also occurs at the La Copita area, but none were observed during our tests. First Field Study. The purpose of this study was to assess the repellency of scent gland secretions. Secretions from adult C. atrox, captured during March in Nolan and Brazos counties, Texas, were collected by manually pressing the base of the tail and allowing the scent gland exudates to flow into glass vials. The area of the cloaca from which scent gland secretions are released is shown in Figure 1. Two milliliters of dichloromethane (CH2C12) were added to each vial, and they were stored at - 2 0 ~ A metal spatula was used to apply 5-6 mg of the scent gland secretions or a commercially available coyote attractant, Carman's Canine Distant Call Lure (CDCL; Russ Carman, Milford, Pennsylvania), to cellulose sponges (1 • 1 • 0.5 cm) that had been immersed in water,
FIG. 1. Cloacal area of female Crotalus atrox showing papilla (arrow) from which scent gland secretions are discharged.
1592
WELDON AND FAGRE
boiled for 20 min, and dried; control sponges received 0.3 ml of CH2C12. The sponges were placed into perforated plastic capsules. One side of each capsule was pierced by a nail to secure it in the soil. All capsules were prepared within 24 hr of the first day of field presentations. They were kept in sealed glass jars until used. Coyote responses to scent gland secretions were evaluated at the Chaparral area during July and August by using a scent station survey method, similar to that described in tests of coyote attractants (Turkowski et al., 1979, 1983). One hundred twenty scent stations were established, each within 2 m of roads extending 80 km into the study site, approximately 0.5 km apart on alternate sides of the road. Each station was comprised of two circular clearings (diameter -- 1 m), 1-2 m apart, created by removing vegetation and sifting soil. A soft brush was used to smooth the soil surface so that tracks and other signs of animal activity could be discerned. A scent capsule was placed by hand (covered with a plastic glove) at the center of each circle. One clearing at each scent station received a capsule containing CDCL. Of the other clearings, half received capsules containing sponges treated with rattlesnake scent gland secretions; the other half received capsules containing sponges treated with CH2C12. Capsules containing scent gland secretions and plain CH2C12 were placed at alternate scent stations. Only one clearing per scent station typically is prepared during field surveys (e.g., Turkowski et al., 1979). We used two clearings during our first field study, one for the CDCL to attract coyotes. We expected coyotes to avoid scent gland secretions (cf., Price and LaPointe, 1981). We therefore deemed it necessary in this initial study to attract subjects to the proximity of scent gland samples so that an aversion could be distinguished from cases in which capsules were ignored. This experiment was not designed to evaluate the attractiveness of scent gland secretions, a property that was considered in subsequent tests. Scent stations were inspected between 0700 and 1200 hr for evidence of animal activities occurring over the previous 24 hr during each of five consecutive days. The following information was scored from marks or deposits left within a 1 m diameter of the center of each circle: (1) visiting vertebrates, identified by their footprints, (2) defecation, (3) urination, as evidenced by small circular depressions (sometimes still moist when examined) in the soil, (4) rubbing and/or rolling, as evidenced by body depressions in soil and hairs deposited on ground or attached to scent capsules (Figure 2), (5) scratching and digging, and (6) pulling up capsules. Contaminated, damaged, or missing capsules were replaced each day, and the soil surface was smoothed by a brush. Scent stations on which tracks could not be identified, including those disturbed by wind or rain, were classified as "inoperable" and were reset with no data recorded from either clearing at the station.
Fie. 2. A plastic capsule containing Crotalus atrox scent gland secretions. The depression in the soil and hairs on the capsule indicate robbing and rolling by a coyote. ts~
z
9
9
z
9
z
f3
1594
WELDON ANDFAGRE
Second Field Study. The purpose of this study was to assess the attractiveness of scent gland secretions. Scent gland secretions were obtained, stored at -20~ and applied to sponges, as previously described; control sponges were untreated. Capsules containing untreated sponges or ones treated with scent gland secretions or CDCL were placed at stations at the La Copita area during March. Each scent station consisted of one circular clearing (diameter = 1 m) of smooth, sifted soil, located within 2 m and on alternate sides of the road. A total of 51 scent stations, 17 for each condition, were established. Stations were monitored for three consecutive days, where measures described for the first field study were scored. Third Field Study. The purpose of this study was to determine whether behaviorally active chemicals from scent gland secretions can be extracted in an organic solvent and can survive rotary evaporation. Scent gland secretions were obtained and stored as described above. Approximately 80 ml of CH2C12 were added to the secretions, they were filtered, and the filtrate was placed on a rotary evaporator to remove the solvent. A 1.0 mg/ml solution of the residue of the scent gland secretions was prepared in CH2C12, and 0.3 ml of this solution was applied to sponges; control sponges received 0.3 ml of CH2C12. Scent stations, comprised of one circular clearing each, were prepared at the Chaparral area during November. A capsule containing a sponge treated with scent gland secretion extract, CH2C12, or CDCL was placed at each station. Sixty-six scent stations, 22 for each condition, were prepared and monitored for three consecutive days using measures described for the previous studies. Domestic Dogs Scent gland secretions, collected in vials as previously described, were transported on dry ice and stored at - 7 0 ~ A total of 30 ml of CH2C12 was added to the vials in 5-ml aliquots. After the addition of each aliquot, the vials were shaken for 1 min, a pipette was inserted beneath a layer of insoluble materials to draw off the CH2C12, and the extract was filtered. The concentration of CHzC12-soluble substances in the filtrate was determined by transferring 5 ml of the solution to a vial, immersing the vial in a water bath (45~ under N2, and weighing the residue after drying for 24 hr in vacuo. A 5 mg/ml solution of secretions was prepared from the original (unevaporated) solution in an attempt to conserve volatile compounds. Filter papers (diameter = 4.3 cm) each were treated with 0.5 ml of the scent gland solution; control papers received 0.5 ml of CHzC12. The papers were air-dried for 5 rain under a hood, placed into glass Petri dishes, covered, and kept at - 7 0 ~ for several hours before use.
CANID RESPONSES TO SNAKE SECRETIONS
1595
A total of 102 domestic dogs (Canis familiaris), pure and mixed breeds, was tested. They were housed individually in 1.2 • 2.4 x 2.7-m kennels, provided with water ad libitum, and fed between 1030 and 1200 hr. No information on the history of the subjects was available. All dogs had occupied their kennels at least 24 hr before testing. Each dog was presented with snake-scented and control filter papers taped to the center of a 7 x 13-cm metal plate. A plate with filter paper was inserted under the front of the kennel into a 15 x 15-cm area designated by a chalk line drawn on the floor. The dogs' tendencies to enter the chalked area and paw at, sniff (by placing the snout within an estimated 5 cm from the paper), mouth (lick, bite, or eat the paper), and rub and roll on the paper were scored for 2 rain to each condition. Two minutes elapsed between presentations, which were counterbalanced. All testing was done between 1200 and 1600 hr. A separate test was done with 19 adult ( > 10 months old) Staffordshire terriers, maintained in their home kennels (1.5 x 4.9 x 1.8-m to 4.9 x 7.6 x 2.4-m) for at least three months. The sex of subjects was recorded. Dogs were presented with filter papers treated with plain CH2CI 2 and CH2C12 extracts of secretions of rattlesnake scent glands or American alligator (Alligator mississippiensis) paracloacal glands. Paracloacal glands are paired exocrine organs (not homologous with scent glands) in the cloacal walls of crocodilians, and are suspected of containing pheromones. An analysis of these secretions indicates a variety of esters, 3-methylbutanoates, free fatty acids, and other compounds (Weldon et al., 1988). The paracloacal gland secretions, collected from adult male alligators during September in Port Arthur, Texas, were included here as an additional control condition. Both rattlesnake and alligator secretions were stored in vials with several ml of CH2C12. They were filtered after adding 30 ml in 5-ml aliquots of CH2C12 to the samples, as described above. Aliquots of snake scent gland and paracloacal gland extracts were weighed, and the solutions from which they were derived were dissolved in CH2C12 to give 6.3 mg/ml solutions, 0.5 ml of which was applied to filter papers. Control papers received 0.5 ml of CH2C12. The papers were air-dried for at least 5 min before being presented to the dogs. The filter papers were taped to metal plates and placed in the dogs' home kennels by their owners. Each dog was presented with control, snake-scented, and alligator-scented papers for 2 rain each in a Latin-square design. Sniffing, mouthing, pawing, rubbing and rolling, and urination postures (leg-lifting and squatting) were scored; since it was not always clear whether micturition occurred, we conservatively scored the occurrence of only the associated squatting or leg-lifting postures. One to two minutes elapsed between presentations. Tests were run between 1000 and 1700 hr, before the dogs were fed.
1596
WELDON AND FAGRE RESULTS
Coyotes First Field Study. The chi-square test was used in all field data analyses. Omitting those scent stations classified as inoperable, 756 observations remained for analysis. The most common vertebrates visiting scent stations over the fiveday test period are listed in Table 1. Visitation rates to scent stations by all species combined did not differ significantly (P > 0.40) from day to day. Visitation rates of coyotes for days 1-5 were 21%, 36%, 36%, 27%, and 32%, respectively (P > 0.50). Coyotes visited 33 % of the scent stations in the Chaparral area. There was
TABLE 1. FREQUENCY OF RESPONSE (%) BY COYOTES AND OTHER MAMMALS IN SOUTH TEXAS TO SCENT STATIONS WITH RATTLESNAKE SCENT GLAND SECRETIONS, DICHLOROMETHANE, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%)
Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Rodents Visits Lagomorphs Visits
Rattlesnake secretion (N = 193)
Dichloromethane (N = 185)
CDCL (N = 378) ~
30.6 2.6 3.1 1.6 22.3 9.3 b
33.5 4.9 0.0 0.0 20.5 2.2 ~
33.1 3.2 1.3 1.1 25.7 13.6 b
4.1
2.7
4.5
1.6
2.2
2.4
3.1
3.7
3.5
80.3
73,5
78,3
8.8
9.2
6.4
CDCL was present at every station so that there were twice as many observations as for the other two treatments. bMeans with different letters are significantly (P < 0.05) different.
C A N ID RESPONSES T O SNAKE S E C R E T I O N S
1597
no evidence on the basis of visitations alone that coyotes discriminated between treatments (P > 0.50), indicating their failure to avoid rattlesnake scent gland chemicals (Table 1). A significantly greater frequency of rubbing and rolling occurred at snake secretion-scented stations than occurred at control stations (P < 0.01). No significant differences were detected in the frequency with which snake secretion-scented or control capsules were disturbed (P > 0.60), or in urinations that occurred at snake secretion-scented or control stations (P > 0.20). Defecation and ground-scratching occurred at frequencies too low to permit analysis. A comparison of coyotes' responses to capsules treated with snake scent gland secretions and coyote attractant, CDCL, on adjoining circles indicated no significant differences between these conditions with respect to rates of visitation (P > 0.50), capsule-pulling (P > 0.30), urination (P > 0.10), defecation (P > 0.10), or ground-scratching (P > 0.60). Coyotes rubbed and rolled at 9 % of snake secretion-scented stations, but this response did not differ significantly (P > 0.10) from the rate at which this occurred (14%) at CDCL stations. Coyotes rubbed and rolled during 30% of their visits to snake secretion-scented stations; they did so at 41% of the CDCL stations visited. Other taxa, particularly rodents, were frequent visitors at scent stations (Table 1). There were no differences in visitation rates at snake secretion-scented and control stations for rodents (P = 0.14) or lagomorphs (P = 0.89), indicating a failure to avoid rattlesnake scent gland chemicals. We also failed to detect significant differences in visitations at snake secretion-scented and control stations for raccoons (P > 0.60), striped skunks (P > 0.20), and javelina (P > 0.50), but visitation rates for these taxa were low. No behavioral responses other than capsule-pulling were recorded for animals other than coyotes, and in many of these instances the capsules appeared to have been kicked or trampled inadvertently. Almost all of the capsules remained within the 1 m circle where species other than coyotes were involved. Coyotes typically carried the capsule at least several meters. There were 18 instances in which the capsules with any of the three stimuli were pulled by all other species combined, for a frequency of 2 %, in contrast to 24 % for coyotes. Second Field Study. After eliminating inoperable stations at La Copita, a total of 146 observations remained for analysis (Table 2). Visitation rates of coyotes for days 1-3 were 22%, 10%, and 14%, respectively. No significant differences were detected in the frequency with which coyotes visited snake secretion- and CDCL-treated stations (P > 0.05), but visits to both of these were higher than to the control stations (P = 0.05). Nearly all capsule-pulling was directed at scent gland secretions and CDCL; the difference between these was slight. Rubbing and rolling was not observed at control stations, but it occurred at 10% and 11% of the snake secretion- and CDCL-treated
1598
WELDON AND FAGRE
TABLE 2. FREQUENCYOF RESPONSE (%) BY COYOTES AND OTHER MAMMALSAT LA COPITA RESEARCH AREA TO SCENT STATIONS WITH RATTLESNAKESCENT GLAND SECRETIONS, UNTREATED CONTROL, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%) Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Lagomorphs Visits
Rattlesnake secretion (N = 50)
Untreated control (N = 51)
CDCL (N = 45)
22.0 a 2.0 0.0 0.0 14.0 10.0
5.8b 0.0 0.0 0.0 1.9 0.0
17.8~ 0.0 0.0 0.0 13.3 11.1
20.0
5.8
17.8
0.0
0.0
2.2
10.0
3.9
6.6
44.0
35.3
53.3
a Means with different letters are significantly (P < 0.05) different.
stations, respectively. Coyotes rubbed and rolled during 71% and 83 % of their visits to stations with scent gland secretions and C D C L , respectively. Evidence o f visits by other vertebrates was indicated at the scent stations (Table 2). No significant differences in visits were noted where the sample sizes for these taxa were adequate for chi-square comparisons. Third Field Study. After eliminating inoperable stations at the Chaparral area, a total of 196 observations remained for analysis (Table 3). Visitation rates o f coyotes for days 1-3 were 2 6 % , 21%, and 38%, respectively. The C D C L stations were visited by coyotes more frequently than were either snake secretion-scented or CH2C12-treated stations (P < 0.05). Capsulepulling and rubbing and rolling were the only behavioral measures recorded in numbers adequate for analysis. Evidence o f both responses was observed more frequently at the C D C L stations than at those treated with either scent gland secretions or CH2C12 (P < 0.05). The frequency with which snake secretionscented capsules were pulled did not differ significantly (P < 0.05) from that
1599
CANID RESPONSESTO SNAKESECRETIONS
TABLE 3. FREQUENCYOF RESPONSE (%) BY COYOTES AND OTHER MAMMALS AT CHAPARRAL WILDLIFE MANAGEMENTAREA TO SCENT STATIONS WITH EXTRACT OF RATTLESNAKE SCENT GLANDS SECRETIONS (AFTER ROTARY EVAPORATION), DICHLOROMETHANE, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%) Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Rodents Visits Lagomorphs Visits
Rattlesnake secretion (N = 66)
Dichloromethane (N = 66)
CDCL (N = 64)
19.7 ~ 3.0 0.0 1,5 13.6a 9.0"
21.2 a 7.5 0.0 0.0 16.6" 0,0 n
45.3 b 10.9 3.1 9.3 34.4 ~' 39.0 b
I0.6
6.0
17.1
1.5
0.0
0.0
6.0
3.0
4.7
60.6
63.6
59.4
40.9 a
40.0 ~
21.9 b
Means with different letters are significantly (P < 0.05) different.
o b s e r v e d w i t h CH2C12-treated capsules; h o w e v e r , no rubbing and rolling was o b s e r v e d in r e s p o n s e to CH2C12-treated capsules. R a c c o o n s visited 11% o f the scent stations, m o s t frequently those containing C D C L (P > 0.10) (Table 3). L a g o m o r p h s and rodents w e r e frequent visitors, but o n l y the f o r m e r e x h i b i t e d differences in visitation rates for the three conditions, visiting C D C L stations less frequently than either snake secretiono r CH2C12-scented stations (P < 0.05). O t h e r species visited stations too infrequently for c h i - s q u a r e analyses.
Domestic Dogs O f the 102 dogs tested in the k e n n e l survey, 77 e x h i b i t e d one or m o r e o f the scored b e h a v i o r s during either control or e x p e r i m e n t a l presentations. O n l y data f r o m these r e s p o n d i n g individuals are p r e s e n t e d and subjected to statistical
1600
WELDON AND FAGRE
TABLE 4. RESPONSES OF 77 DOMESTIC DOGS, PURE AND MIXED BREEDS, DURING 2MIN PRESENTATIONS OF FILTER PAPERS TREATED WITH DICHLOROMETHANE AND DICHLOROMETHANE EXTRACT OF RATTLESNAKE SCENT GLAND SECRETIONS
Treatment
.~ times entered square (No. of dogs)
No. dogs pawing
~7 sniffing bouts (No. of dogs)
No. dogs mouthing
No. dogs rubbing and rolling
Dichloromethane Scent gland
1.5 (59) 1.9 (66)
14 15
1.3 (66) 1.8 (63)
15 26
4 8
analyses (Table 4). The chi-square test indicates that dogs mouthed filter papers treated with the CH2C12 extract of scent gland secretions more than controls (P < 0.05). No other significant differences between conditions were detected. Of the 19 Staffordshire terriers tested, 12 exhibited one or more of the scored behaviors during either control or experimental presentations. Only data from these responding individuals are presented here and subjected to statistical analyses. Rubbing and rolling was observed during 0, 1, and 3 sessions, and urination postures during 1, 2, and 6 sessions with CH2CI> paracloacal gland secretions, and scent gland secretions, respectively. All urination postures, except one in which a female urinated directly on an alligator-scented paper, were exhibited by males. No significant differences were detected between conditions for rubbing and rolling. The Fisher exact-probability test, which was used due to the small sample size, indicates that urination postures occurred during significantly more sessions with scent gland secretions than with CHzC12 (P < 0.05), but responses to snake- and alligator-treated papers did not differ significantly (0.10 > P > 0.05). DISCUSSION
The reactions of predatory mammals to snake scent gland secretions were observed by Price and LaPointe (1981), who presented desert kingsnake (Lampropeltis getulus splendida) exudates on food to badgers, coyotes, and other North American carnivores. The coyotes' responses were not specifically described, but Price and LaPointe (1981) characterized the overall reactions to scent gland secretions by the carnivores they examined as "aversionary." The results of our first two field experiments indicate that coyotes rubbed and rolled on rattlesnake scent gland secretions, as they did on the coyote attractant, CDCL. In our third field test, however, rubbing and rolling did not occur to scent gland extracts at a rate significantly different from that to CH2CI>
CANID RESPONSES TO SNAKE SECRETIONS
t 601
although this behavior was indicated at a significantly higher rate to CDCL. This suggests that the chemicals eliciting rubbing and rolling are not retained (in sufficient quantities) following our methods of extraction, rotary evaporation, and field presentation. Domestic dogs in both our kennel tests also failed to rub and roll on CHzC12 extracts of scent gland secretions significantly more often than on control papers. Scent gland compounds insoluble in CH2C12, i.e., glyco- or mucoproteins and acid mucopolysaccarides (Whiting, 1969; Blum et al., 1971), likely were denatured or removed by our extraction methods. These substances may provide a matrix for the retention of volatile, behaviorally active chemicals. Bullard (1982) states that most, if not all, canids rub and roll on putrefying materials. Fox and Cohen (1977) report rubbing and rolling by a variety of canids to natural and synthetic substances, including skatole and propionic acid. Although sterols and fatty acids have been indicated in snake scent glands (Blum et al., 1971; Simpson et al., 1988), Oldak (1976) failed to detect these or other nonpolar lipids in his thin-layer chromatograms of crotaline scent gland secretions, including those of a rattlesnake. Razakov and Sadykov (1986) identified fatty acids from the scent glands of the mamushi (Agkistrodon halys), indicating that these compounds are present in at least some crotalines. Chemicals in the secretions of C. atrox need to be identified and tested as synthetic analogs to determine those that elicit canid responses. In our study, Staffordshire terriers exhibited urination postures more frequently to rattlesnake secretion extracts than to CH2C12, hut snake- and alligator-treated papers did not differ significantly in their effects. Urination (or urination posturing) was not scored during our first kennel survey, and it occurred at too low a frequency for statistical analysis during our field studies. The tendency of canids to urinate on conspecific urine deposits often is interpreted as a form of scent marking. Urination also may be elicited by " n o v e l " odors, such as those arising from other species' feces (Fox and Cohen, 1977). Kool (1981) noted that the dingo (C. familiaris) urinated on a turtle's (Chelodina longicollis) Rathke's gland secretions, which are hypothesized to deter predators. In Kool's study, as with ours, the adaptive significance of eliciting urination in canids is not obvious. Neither are the aversive properties of scent gland secretions evident from our results. In the first field study, designed to measure the potential repellency of snake scent gland secretions, coyotes visited capsules containing these exudates at a rate that was not significantly different from that to capsules containing CHzC12. In the second field study, designed to measure the potential attractiveness of rattlesnake scent gland secretions, coyotes visited stations containing secretion extracts significantly more often than those containing untreated capsules and at a rate not significantly different from that to CDCL. This is worth noting since CDCL has been developed specifically as a lure for these
1602
WELDON AND FAGRE
carnivores and is one of the most effective commercial attractants available (Turkowski et al., 1983). We did not test the palatability of untreated scent gland secretions in our experiments, thus, we cannot relate our results directly to those of Price and LaPointe (1981). However, the tendency of dogs to mouth snake secretionscented filter papers and that of coyotes in the field to pull up (presumably by mouth) snake secretion-scented capsules does not support the hypothesis that these materials are distasteful. Palatability studies using whole, untreated scent gland substances are needed. Scent gland secretions could interfere with predation in ways other than by having aversive effects. Rubbing and rolling, for example, could distract canids from predatory attack. Even the elicitation of mouthing could be advantageous for snakes if scent gland chemicals direct attacks toward the tail, and away from their head and trunk, in a way analogous to that in which visually directed predators are diverted by snake tail displays (Greene, 1987). Insight into the mechanisms by which scent gland secretions benefit snakes could be obtained by observing interactions between snakes and predators and, perhaps, by comparing the survivorship of intact snakes with that of those from which these glands have been surgically removed. Scent glands are homologous among extant snakes. Since the common ancestor of Recent snakes lived during the Upper Cretaceous (Rage, 1987) this can be considered a minimum age for these organs. Carnivores did not appear until during the Tertiary (Ewer, 1973). Given this chronology, it is clear that snake scent glands did not evolve as a response to these mammalian predators. The antipredator function(s) of the exudates produced in these organs needs to be tested on other taxa, e.g., varanid lizards and crocodilians. Functions and effects other than the production of predator repellents have been attributed to scent glands, including that of pheromone release. Although experiments suggest that blindsnake (Leptotyphlops dulcis) scent gland chemicals attract conspecifics (Watkins et al., 1969), attempts to elicit trail-following with these substances in other snakes have failed (reviewed in Ford, 1986). Graves and Duvall (1988) report that airborne presentations of scent gland chemicals from the prairie rattlesnake (C. viridus) elicit increased heart rates in conspecifics, and they suggest, following Brisbin's (1968) observations on a common kingsnake (Lampropeltis getulus), that these organs release alarm pheromones. Even so, whether scent gland chemicals function primarily to alert conspecifics is uncertain since alarm properties often are derived secondarily from compounds released during predatory attack (Weldon, 1983).
Acknowledgments--B. Beaver helped with preliminary observations on dogs' reactions to scent gland chemicals. L. MacDonald and M. Phi Dao lent valuable help in testing domestic dogs. R.S. Gray, H.B. Harper, and C. Toth kindly allowed us to test their Staffordshire terriers and
CANID RESPONSES TO SNAKE SECRETIONS
1603
assisted in the process. G. Drew, D. Martin, and C. Roberts offered expert field assistance in tests with coyotes. P. Freed, C. Roberts, J.T. Simpson, G. Wills, and T. Wilson helped collect scent gland secretions. D. Synatske and the staff at the Chaparral Wildlife Management Area, Texas Parks and Wildlife Department, allowed us to conduct our studies at their facilities. R.W. Bullard and H.W. Greene commented on earlier drafts of the manuscript. This study was supported by grants from the Whitehall Foundation (to P.J.W.), the Texas Agricultural Experiment Station (through D.B.F.; contribution #23293), and a Texas A&M minigrant (to D.B.F.). To all we are grateful.
REFERENCES BELLAIRS, A. D'A. 1950. The limbs of snakes with special reference to the hind limb rudiments of Trachyboa boulengeri. Br. J. Herpetol. 4:73-83. BLUM, M.S., BYRD, J.B., TRAVIS,J.R., WATK1NS,J.F., II, and GEHLBACH, F.R. 1971. Chemistry of the cloacal sac secretion of the blind snake Leptotyphlops duleis. Comp. Biochem. Physiol. 38:103-107. BRISBIN, I.L., JR. 1968. Evidence for the use of postanal musk as an alarm device in the king snake, Lampropeltis getulus. Herpetologica 24:169-170. BULLARD, R.W. 1982. Wild canid associations with fermentation products. Ind. Eng. Chem. Prod. Res. Div. 21:646-655. BULLARD, R.W., TURKOWSKI,F.J., and KtLBURN, S.R. 1983. Responses of free-ranging coyotes to lures and their modifications. J. Chem. Ecol. 9:877-888. BURKETT, R.D. 1966. Natural history of the cottonmouth moccasin, Agkistrodon piscivorus (Reptilia). Univ. Kans. Publ. Mus. Nat. Hist. 17:435-491. DUW,LL, D., KING, M.B., and GUTZWILLER, K.J. 1985. Behavioral ecology and ethology of the prairie rattlesnake. Nat. Geogr. Res. 1: 80-111. EISNER, T., CONNER, W.E., HICKS, K., DODGE, K.R., ROSENBERG, H.I., JONES, T.H., COHEN, M., and MEINWALD, J. 1977. Stink of the stinkpot turtle identified: r acids. Science 196:1347-1349. EWER, R.F. 1973. The Carnivores. Cornell University Press, Ithaca, New York. FITCH, H.S. 1960. Autecology of the copperhead. Univ. Kans. Publ. Mus. Nat. Hist. 13:85-288. Fn'CH, H.S. 1965. An ecological study of the garter snake, Thamnophis sirtalis. Univ. Kans. Publ. Mus. Nat. Hist. 15:493-564. FORD, N.B. 1986. The role of pheromone trails in the sociobiology of snakes, pp. 309-312, in D. Duvall, D. Muller-Schwarze, and R.M. Silverstein (eds.). Chemical Signals in Vertebrates 4: Ecology, Evolution, and Comparative Biology. Plenum Press, New York. Fox, M.W., and COHEN, J.A. 1977. Canid communication, pp. 728-748, in T.A. Sebeok (ed.). How Animals Communicate. Indiana University Press, Bloomington. GEHLBACr~, F.R., WATKINS, J.F., II, and RENO, H.W. 1968. Blind snake defensive behavior elicited by ant attacks. BioScience 18:784-785. GRAVES, B.M., and DUVALL, D. 1988. Evidence of an alarm pheromone from the cloacal sacs of prairie rattlesnakes. Southwest Nat. 33:339-346. GREENE, H.W. 1987. Antipredator mechanisms in reptiles, pp. 1-151, in C. Gans and R.B. Huey (eds.). Biology of the Reptilia, Vol. 16. Alan R. Liss, New York. KLAUBER, L.M. 1972. Rattlesnakes, Their habits, Life Histories, and Influence on Mankind, Vol. 2. University of California Press, Berkeley. KOOL, K. 1981. Is the musk of the long-necked turtle, Chelodina longicollis, a deterrent to predators? Aust. J. Herpetol. 1:45-53.
1604
WELDON AND FAGRE
OLDAK, P.D. 1976. Comparison of the scent gland secretion lipids of twenty-five snakes: Implications for biochemical systematics. Copeia 1976:320-326. PEREZ, J.C., HAWS, H.C., GARCIA, V.E., and JENNINCS, B.M., III. 1978. Resistance of warmblooded animals to snake venoms. Toxicon 16:375-383. PRICE, A.H., and LAPOINTE, J.L. 1981. Structure-functional aspects of the scent gland in Lampropeltis getulus splendida. Copeia 1981:138-146. RAGE, J.C. 1987. Fossil history, pp. 51-76, in R.A. Seigel, J.T. Collins, and S.S. Novak, (eds.). Snakes: Ecology and Evolutionary Biology. Macmillan, New York. RAZAKOV, R.R., and SADYKOV,A.S. 1986. Study of complex mixtures of natural substances by the refocusing and DADI methods. VI. Components of the secretion of the pre-anal [sic] gland of some poisonous snakes. Khim. Prir. Soedin 4:421-423. SIMPSON,J.T., WELDON, P.J., and SHARP,T.R. 1988. Identification of major lipids from the scent gland secretions of Dumeril's ground boa (Acrantophis dumerili Jan) by gas chromatographymass spectrometry. Z. Naturforsch. 43c:914-917. TURKOWSKI,F.J., POPELKA,M.L., GREEN, B.B., and BULLARD,R.W. 1979. Testing the response of coyotes and other predators to odor attractants, pp. 255-269, in J.R. Beck (ed.). Test Methods for Vertebrate Pest Control and Management Materials. ASTM STP 680. American Society for Testing and Materials, Philadelphia, Pennsylvania. TURKOWSKI,F.J., POPELKA,M.L., and BULLARD, R.W. 1983. Efficacy of odor lures and baits for coyotes. Wildl. Soc. Bull. 11:136-145. WATKINS,J.F., II, GEHLBACH,F.R., and KROLL,J.C. 1969. Attractant-repellent secretions of blind snakes (Leptotyphlops dulcis) and their army ant prey (Neivamyrmex nigrescens). Ecology 50:1098-1102. WELDON, P.J. 1983. The evolution of alarm pheromones, pp. 309-312, in D. Multer-Schwarze and R.M. Silverstein, (eds.). Chemical Signals in Vertebrates 3. Plenum Press, New York. WELDON, P.J. 1988. Feeding responses by snappers (genus Lutjanus) to the yellow-bellied sea snake (Pelamis platurus). Zool. Sci. 5:443-448. WELDON, P.J., SHAFAGATI,A., and WHEELER, J.W. 1988. Lipids from the paracloacal glands of the American alligator (Alligator mississippiensis). Lipids 23:727-729. WHmNG, A.M., 1969. Squamate cloacal glands: Morphology, histology and histochemistry. PhD dissertation. Pennsylvania State University.
RESPONSES BY CANIDS
TO S C E N T G L A N D
S E C R E T I O N S OF T H E W E S T E R N RATTLESNAKE
DIAMONDBACK
(Crotalus atrox)
PAUL J. WELDON l and DANIEL B. FAGRE2 Departments of 1Biology and 2Wildlife and Fisheries Sciences Texas A&M University College Station, TX 77843 (Received December 22, 1987; accepted July 28, 1988)
Abstract--Many snakes discharge malodorous fluids from paired scent glands in the base of the tail when they are disturbed. A number of authors suggest that these secretions repel predators. Scent gland secretions of the western diamondback rattlesnake (Crotalus atrox), or dichloromethane extracts of them, were presented to coyotes (Canis latrans) in three field tests, and to domestic dogs (Canis familiaris) in two kennel tests, to determine whether responses of possible benefit to snakes are elicited. Free-ranging coyotes visited and robbed and rolled at stations containing scent gland secretions in perforated plastic capsules more frequently than at those containing untreated or dichloromethane-treated capsules. Responses to dichloromethane extracts of scent gland secretions subjected to rotary evaporation were not significantly different from those to dichloromethane. Pure and mixed breeds of dogs presented with filter papers treated with dichloromethane or a dichloromethane extract of scent gland secretions mouthed (licked, bit, or ate) secretion-treated papers more frequently. Staffordshire terriers presented with filter papers treated with dichloromethane or dichloromethane extracts of snake scent gland and alligator (Alligator mississippiensis) paracloacal gland secretions exhibited urination postures to snake secretion-treated papers more frequently than to dichloromethane-treated papers, but responses to snake- and alligator-treated papers did not differ significantly. There was no indication that canids are repelled by scent gland secretions. Key Words--Canidae, coyote, Canis latrans, dog, Canisfamiliaris, rattlesnake, Crotalus atrox, scent glands.
1589 0098-0331/'89/0500- [ 589506.00/0 9 1989 Plenum Publishing Corporation
1590
W~LDONAND FAGRE INTRODUCTION
Reptiles produce a variety of secretions thought to repel predators. In only a few stUdies, however, have their natural products been tested for antipredator activity (Eisner et al., 1977; Kool, 1981; Weldon, 1988). Scent glands are paired exocrine organs in the tail of all snakes (Bellairs, 1950; Whiting, 1969). These glands typically release fluids, which are thought to discourage predators (Fitch, 1960, 1965; Burkett, 1966; Klauber, 1972), from two openings at the posteriolateral margin of the cloaca when snakes are disturbed. Several species of ants (Gehlbach et al., 1968) and several snakes (Watkins et al., 1969), including ophiophagous species, are repelled by the scent gland secretions of the Texas blindsnake (Leptotyphlops dulcis). A preliminary test indicated that some North American carnivores are reluctant to approach or eat food treated with desert kingsnake (Lampropeltis getulus splendida) secretions (Price and LaPointe, 1981), but there are no quantitative tests of the reactions by mammalian predators to these substances. Rattlesnakes (genus Crotalus) typically discharge, and sometimes spray, scent gland fluids when molested (Klauber, 1972; Duvall et al., 1985). Since many mammals are resistant to snake venom (Perez et al., 1978), defenses other than striking and envenomation--such as scent gland discharge--may be useful against these predators. Canids are one group of carnivores with which rattlesnakes interact. We report here tests of the reactions of coyotes (Canis latrans) and domestic dogs (Canis familiaris) to scent gland secretions of the western diamondback rattlesnake (Crotalus atrox). This study was conducted to determine whether responses of possible benefit to snakes are elicited in canids by scent gland secretions. Our field tests were designed primarily to test coyotes, although data on other vertebrates visiting our scent stations are presented. The series of field experiments involved different methods of chemical sample preparation or presentation, since the results of each test suggested questions that were addressed by modifications in subsequent designs. None of the field tests involved direct observation of the behavior of coyotes. Rather, inferences were made regarding coyotes' responses on the basis of impressions left in the soil or from other telltale deposits. Domestic dogs were tested to allow direct observation of canids' responses to scent gland secretions. METHODS AND MATERIALS
Coyotes Responses to rattlesnake scent gland secretions were conducted at the Chaparral Wildlife Management Area, a 6151-hectare ranch located in the Rio Grande Plains region of Texas in Dimmit and La Salle counties, and at the La
C A N I D RESPONSES T O S N A K E S E C R E T I O N S
1591
Copita Research Area, a 1093-hectare ranch located in Jim Wells County, approximately 15 km south of Alice, Texas. The average annual rainfall at the Chaparral and La Copita areas is 84 and 68 cm, respectively. The vegetation of both areas is dominated by honey mesquite (Prosopis glandulosa) and includes brazil (Condalia obovata) and prickly pear cactus (Opuntia rigidula). In addition to coyotes, eastern cottontail rabbits (Sylvilagusfloridanus), striped skunks (Mephitis mephitis), raccoons (Procyon lotor), badgers (Taxidea taxus), javelina (Tayassu tajacu), and rodents were observed at both sites at the time of testing. Five C. atrox were seen during the study in the Chaparral area. C. atrox also occurs at the La Copita area, but none were observed during our tests. First Field Study. The purpose of this study was to assess the repellency of scent gland secretions. Secretions from adult C. atrox, captured during March in Nolan and Brazos counties, Texas, were collected by manually pressing the base of the tail and allowing the scent gland exudates to flow into glass vials. The area of the cloaca from which scent gland secretions are released is shown in Figure 1. Two milliliters of dichloromethane (CH2C12) were added to each vial, and they were stored at - 2 0 ~ A metal spatula was used to apply 5-6 mg of the scent gland secretions or a commercially available coyote attractant, Carman's Canine Distant Call Lure (CDCL; Russ Carman, Milford, Pennsylvania), to cellulose sponges (1 • 1 • 0.5 cm) that had been immersed in water,
FIG. 1. Cloacal area of female Crotalus atrox showing papilla (arrow) from which scent gland secretions are discharged.
1592
WELDON AND FAGRE
boiled for 20 min, and dried; control sponges received 0.3 ml of CH2C12. The sponges were placed into perforated plastic capsules. One side of each capsule was pierced by a nail to secure it in the soil. All capsules were prepared within 24 hr of the first day of field presentations. They were kept in sealed glass jars until used. Coyote responses to scent gland secretions were evaluated at the Chaparral area during July and August by using a scent station survey method, similar to that described in tests of coyote attractants (Turkowski et al., 1979, 1983). One hundred twenty scent stations were established, each within 2 m of roads extending 80 km into the study site, approximately 0.5 km apart on alternate sides of the road. Each station was comprised of two circular clearings (diameter -- 1 m), 1-2 m apart, created by removing vegetation and sifting soil. A soft brush was used to smooth the soil surface so that tracks and other signs of animal activity could be discerned. A scent capsule was placed by hand (covered with a plastic glove) at the center of each circle. One clearing at each scent station received a capsule containing CDCL. Of the other clearings, half received capsules containing sponges treated with rattlesnake scent gland secretions; the other half received capsules containing sponges treated with CH2C12. Capsules containing scent gland secretions and plain CH2C12 were placed at alternate scent stations. Only one clearing per scent station typically is prepared during field surveys (e.g., Turkowski et al., 1979). We used two clearings during our first field study, one for the CDCL to attract coyotes. We expected coyotes to avoid scent gland secretions (cf., Price and LaPointe, 1981). We therefore deemed it necessary in this initial study to attract subjects to the proximity of scent gland samples so that an aversion could be distinguished from cases in which capsules were ignored. This experiment was not designed to evaluate the attractiveness of scent gland secretions, a property that was considered in subsequent tests. Scent stations were inspected between 0700 and 1200 hr for evidence of animal activities occurring over the previous 24 hr during each of five consecutive days. The following information was scored from marks or deposits left within a 1 m diameter of the center of each circle: (1) visiting vertebrates, identified by their footprints, (2) defecation, (3) urination, as evidenced by small circular depressions (sometimes still moist when examined) in the soil, (4) rubbing and/or rolling, as evidenced by body depressions in soil and hairs deposited on ground or attached to scent capsules (Figure 2), (5) scratching and digging, and (6) pulling up capsules. Contaminated, damaged, or missing capsules were replaced each day, and the soil surface was smoothed by a brush. Scent stations on which tracks could not be identified, including those disturbed by wind or rain, were classified as "inoperable" and were reset with no data recorded from either clearing at the station.
Fie. 2. A plastic capsule containing Crotalus atrox scent gland secretions. The depression in the soil and hairs on the capsule indicate robbing and rolling by a coyote. ts~
z
9
9
z
9
z
f3
1594
WELDON ANDFAGRE
Second Field Study. The purpose of this study was to assess the attractiveness of scent gland secretions. Scent gland secretions were obtained, stored at -20~ and applied to sponges, as previously described; control sponges were untreated. Capsules containing untreated sponges or ones treated with scent gland secretions or CDCL were placed at stations at the La Copita area during March. Each scent station consisted of one circular clearing (diameter = 1 m) of smooth, sifted soil, located within 2 m and on alternate sides of the road. A total of 51 scent stations, 17 for each condition, were established. Stations were monitored for three consecutive days, where measures described for the first field study were scored. Third Field Study. The purpose of this study was to determine whether behaviorally active chemicals from scent gland secretions can be extracted in an organic solvent and can survive rotary evaporation. Scent gland secretions were obtained and stored as described above. Approximately 80 ml of CH2C12 were added to the secretions, they were filtered, and the filtrate was placed on a rotary evaporator to remove the solvent. A 1.0 mg/ml solution of the residue of the scent gland secretions was prepared in CH2C12, and 0.3 ml of this solution was applied to sponges; control sponges received 0.3 ml of CH2C12. Scent stations, comprised of one circular clearing each, were prepared at the Chaparral area during November. A capsule containing a sponge treated with scent gland secretion extract, CH2C12, or CDCL was placed at each station. Sixty-six scent stations, 22 for each condition, were prepared and monitored for three consecutive days using measures described for the previous studies. Domestic Dogs Scent gland secretions, collected in vials as previously described, were transported on dry ice and stored at - 7 0 ~ A total of 30 ml of CH2C12 was added to the vials in 5-ml aliquots. After the addition of each aliquot, the vials were shaken for 1 min, a pipette was inserted beneath a layer of insoluble materials to draw off the CH2C12, and the extract was filtered. The concentration of CHzC12-soluble substances in the filtrate was determined by transferring 5 ml of the solution to a vial, immersing the vial in a water bath (45~ under N2, and weighing the residue after drying for 24 hr in vacuo. A 5 mg/ml solution of secretions was prepared from the original (unevaporated) solution in an attempt to conserve volatile compounds. Filter papers (diameter = 4.3 cm) each were treated with 0.5 ml of the scent gland solution; control papers received 0.5 ml of CHzC12. The papers were air-dried for 5 rain under a hood, placed into glass Petri dishes, covered, and kept at - 7 0 ~ for several hours before use.
CANID RESPONSES TO SNAKE SECRETIONS
1595
A total of 102 domestic dogs (Canis familiaris), pure and mixed breeds, was tested. They were housed individually in 1.2 • 2.4 x 2.7-m kennels, provided with water ad libitum, and fed between 1030 and 1200 hr. No information on the history of the subjects was available. All dogs had occupied their kennels at least 24 hr before testing. Each dog was presented with snake-scented and control filter papers taped to the center of a 7 x 13-cm metal plate. A plate with filter paper was inserted under the front of the kennel into a 15 x 15-cm area designated by a chalk line drawn on the floor. The dogs' tendencies to enter the chalked area and paw at, sniff (by placing the snout within an estimated 5 cm from the paper), mouth (lick, bite, or eat the paper), and rub and roll on the paper were scored for 2 rain to each condition. Two minutes elapsed between presentations, which were counterbalanced. All testing was done between 1200 and 1600 hr. A separate test was done with 19 adult ( > 10 months old) Staffordshire terriers, maintained in their home kennels (1.5 x 4.9 x 1.8-m to 4.9 x 7.6 x 2.4-m) for at least three months. The sex of subjects was recorded. Dogs were presented with filter papers treated with plain CH2CI 2 and CH2C12 extracts of secretions of rattlesnake scent glands or American alligator (Alligator mississippiensis) paracloacal glands. Paracloacal glands are paired exocrine organs (not homologous with scent glands) in the cloacal walls of crocodilians, and are suspected of containing pheromones. An analysis of these secretions indicates a variety of esters, 3-methylbutanoates, free fatty acids, and other compounds (Weldon et al., 1988). The paracloacal gland secretions, collected from adult male alligators during September in Port Arthur, Texas, were included here as an additional control condition. Both rattlesnake and alligator secretions were stored in vials with several ml of CH2C12. They were filtered after adding 30 ml in 5-ml aliquots of CH2C12 to the samples, as described above. Aliquots of snake scent gland and paracloacal gland extracts were weighed, and the solutions from which they were derived were dissolved in CH2C12 to give 6.3 mg/ml solutions, 0.5 ml of which was applied to filter papers. Control papers received 0.5 ml of CH2C12. The papers were air-dried for at least 5 min before being presented to the dogs. The filter papers were taped to metal plates and placed in the dogs' home kennels by their owners. Each dog was presented with control, snake-scented, and alligator-scented papers for 2 rain each in a Latin-square design. Sniffing, mouthing, pawing, rubbing and rolling, and urination postures (leg-lifting and squatting) were scored; since it was not always clear whether micturition occurred, we conservatively scored the occurrence of only the associated squatting or leg-lifting postures. One to two minutes elapsed between presentations. Tests were run between 1000 and 1700 hr, before the dogs were fed.
1596
WELDON AND FAGRE RESULTS
Coyotes First Field Study. The chi-square test was used in all field data analyses. Omitting those scent stations classified as inoperable, 756 observations remained for analysis. The most common vertebrates visiting scent stations over the fiveday test period are listed in Table 1. Visitation rates to scent stations by all species combined did not differ significantly (P > 0.40) from day to day. Visitation rates of coyotes for days 1-5 were 21%, 36%, 36%, 27%, and 32%, respectively (P > 0.50). Coyotes visited 33 % of the scent stations in the Chaparral area. There was
TABLE 1. FREQUENCY OF RESPONSE (%) BY COYOTES AND OTHER MAMMALS IN SOUTH TEXAS TO SCENT STATIONS WITH RATTLESNAKE SCENT GLAND SECRETIONS, DICHLOROMETHANE, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%)
Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Rodents Visits Lagomorphs Visits
Rattlesnake secretion (N = 193)
Dichloromethane (N = 185)
CDCL (N = 378) ~
30.6 2.6 3.1 1.6 22.3 9.3 b
33.5 4.9 0.0 0.0 20.5 2.2 ~
33.1 3.2 1.3 1.1 25.7 13.6 b
4.1
2.7
4.5
1.6
2.2
2.4
3.1
3.7
3.5
80.3
73,5
78,3
8.8
9.2
6.4
CDCL was present at every station so that there were twice as many observations as for the other two treatments. bMeans with different letters are significantly (P < 0.05) different.
C A N ID RESPONSES T O SNAKE S E C R E T I O N S
1597
no evidence on the basis of visitations alone that coyotes discriminated between treatments (P > 0.50), indicating their failure to avoid rattlesnake scent gland chemicals (Table 1). A significantly greater frequency of rubbing and rolling occurred at snake secretion-scented stations than occurred at control stations (P < 0.01). No significant differences were detected in the frequency with which snake secretion-scented or control capsules were disturbed (P > 0.60), or in urinations that occurred at snake secretion-scented or control stations (P > 0.20). Defecation and ground-scratching occurred at frequencies too low to permit analysis. A comparison of coyotes' responses to capsules treated with snake scent gland secretions and coyote attractant, CDCL, on adjoining circles indicated no significant differences between these conditions with respect to rates of visitation (P > 0.50), capsule-pulling (P > 0.30), urination (P > 0.10), defecation (P > 0.10), or ground-scratching (P > 0.60). Coyotes rubbed and rolled at 9 % of snake secretion-scented stations, but this response did not differ significantly (P > 0.10) from the rate at which this occurred (14%) at CDCL stations. Coyotes rubbed and rolled during 30% of their visits to snake secretion-scented stations; they did so at 41% of the CDCL stations visited. Other taxa, particularly rodents, were frequent visitors at scent stations (Table 1). There were no differences in visitation rates at snake secretion-scented and control stations for rodents (P = 0.14) or lagomorphs (P = 0.89), indicating a failure to avoid rattlesnake scent gland chemicals. We also failed to detect significant differences in visitations at snake secretion-scented and control stations for raccoons (P > 0.60), striped skunks (P > 0.20), and javelina (P > 0.50), but visitation rates for these taxa were low. No behavioral responses other than capsule-pulling were recorded for animals other than coyotes, and in many of these instances the capsules appeared to have been kicked or trampled inadvertently. Almost all of the capsules remained within the 1 m circle where species other than coyotes were involved. Coyotes typically carried the capsule at least several meters. There were 18 instances in which the capsules with any of the three stimuli were pulled by all other species combined, for a frequency of 2 %, in contrast to 24 % for coyotes. Second Field Study. After eliminating inoperable stations at La Copita, a total of 146 observations remained for analysis (Table 2). Visitation rates of coyotes for days 1-3 were 22%, 10%, and 14%, respectively. No significant differences were detected in the frequency with which coyotes visited snake secretion- and CDCL-treated stations (P > 0.05), but visits to both of these were higher than to the control stations (P = 0.05). Nearly all capsule-pulling was directed at scent gland secretions and CDCL; the difference between these was slight. Rubbing and rolling was not observed at control stations, but it occurred at 10% and 11% of the snake secretion- and CDCL-treated
1598
WELDON AND FAGRE
TABLE 2. FREQUENCYOF RESPONSE (%) BY COYOTES AND OTHER MAMMALSAT LA COPITA RESEARCH AREA TO SCENT STATIONS WITH RATTLESNAKESCENT GLAND SECRETIONS, UNTREATED CONTROL, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%) Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Lagomorphs Visits
Rattlesnake secretion (N = 50)
Untreated control (N = 51)
CDCL (N = 45)
22.0 a 2.0 0.0 0.0 14.0 10.0
5.8b 0.0 0.0 0.0 1.9 0.0
17.8~ 0.0 0.0 0.0 13.3 11.1
20.0
5.8
17.8
0.0
0.0
2.2
10.0
3.9
6.6
44.0
35.3
53.3
a Means with different letters are significantly (P < 0.05) different.
stations, respectively. Coyotes rubbed and rolled during 71% and 83 % of their visits to stations with scent gland secretions and C D C L , respectively. Evidence o f visits by other vertebrates was indicated at the scent stations (Table 2). No significant differences in visits were noted where the sample sizes for these taxa were adequate for chi-square comparisons. Third Field Study. After eliminating inoperable stations at the Chaparral area, a total of 196 observations remained for analysis (Table 3). Visitation rates o f coyotes for days 1-3 were 2 6 % , 21%, and 38%, respectively. The C D C L stations were visited by coyotes more frequently than were either snake secretion-scented or CH2C12-treated stations (P < 0.05). Capsulepulling and rubbing and rolling were the only behavioral measures recorded in numbers adequate for analysis. Evidence o f both responses was observed more frequently at the C D C L stations than at those treated with either scent gland secretions or CH2C12 (P < 0.05). The frequency with which snake secretionscented capsules were pulled did not differ significantly (P < 0.05) from that
1599
CANID RESPONSESTO SNAKESECRETIONS
TABLE 3. FREQUENCYOF RESPONSE (%) BY COYOTES AND OTHER MAMMALS AT CHAPARRAL WILDLIFE MANAGEMENTAREA TO SCENT STATIONS WITH EXTRACT OF RATTLESNAKE SCENT GLANDS SECRETIONS (AFTER ROTARY EVAPORATION), DICHLOROMETHANE, AND COYOTE ATTRACTANT (CARMAN'S CANINE DISTANT CALL LURE)
Response (%) Animals and behavioral response Coyote Visits Urination Defecation Groundscratching Capsule pulling Rubbing/rolling Raccoon Visits Striped skunk Visits Javelina Visits Rodents Visits Lagomorphs Visits
Rattlesnake secretion (N = 66)
Dichloromethane (N = 66)
CDCL (N = 64)
19.7 ~ 3.0 0.0 1,5 13.6a 9.0"
21.2 a 7.5 0.0 0.0 16.6" 0,0 n
45.3 b 10.9 3.1 9.3 34.4 ~' 39.0 b
I0.6
6.0
17.1
1.5
0.0
0.0
6.0
3.0
4.7
60.6
63.6
59.4
40.9 a
40.0 ~
21.9 b
Means with different letters are significantly (P < 0.05) different.
o b s e r v e d w i t h CH2C12-treated capsules; h o w e v e r , no rubbing and rolling was o b s e r v e d in r e s p o n s e to CH2C12-treated capsules. R a c c o o n s visited 11% o f the scent stations, m o s t frequently those containing C D C L (P > 0.10) (Table 3). L a g o m o r p h s and rodents w e r e frequent visitors, but o n l y the f o r m e r e x h i b i t e d differences in visitation rates for the three conditions, visiting C D C L stations less frequently than either snake secretiono r CH2C12-scented stations (P < 0.05). O t h e r species visited stations too infrequently for c h i - s q u a r e analyses.
Domestic Dogs O f the 102 dogs tested in the k e n n e l survey, 77 e x h i b i t e d one or m o r e o f the scored b e h a v i o r s during either control or e x p e r i m e n t a l presentations. O n l y data f r o m these r e s p o n d i n g individuals are p r e s e n t e d and subjected to statistical
1600
WELDON AND FAGRE
TABLE 4. RESPONSES OF 77 DOMESTIC DOGS, PURE AND MIXED BREEDS, DURING 2MIN PRESENTATIONS OF FILTER PAPERS TREATED WITH DICHLOROMETHANE AND DICHLOROMETHANE EXTRACT OF RATTLESNAKE SCENT GLAND SECRETIONS
Treatment
.~ times entered square (No. of dogs)
No. dogs pawing
~7 sniffing bouts (No. of dogs)
No. dogs mouthing
No. dogs rubbing and rolling
Dichloromethane Scent gland
1.5 (59) 1.9 (66)
14 15
1.3 (66) 1.8 (63)
15 26
4 8
analyses (Table 4). The chi-square test indicates that dogs mouthed filter papers treated with the CH2C12 extract of scent gland secretions more than controls (P < 0.05). No other significant differences between conditions were detected. Of the 19 Staffordshire terriers tested, 12 exhibited one or more of the scored behaviors during either control or experimental presentations. Only data from these responding individuals are presented here and subjected to statistical analyses. Rubbing and rolling was observed during 0, 1, and 3 sessions, and urination postures during 1, 2, and 6 sessions with CH2CI> paracloacal gland secretions, and scent gland secretions, respectively. All urination postures, except one in which a female urinated directly on an alligator-scented paper, were exhibited by males. No significant differences were detected between conditions for rubbing and rolling. The Fisher exact-probability test, which was used due to the small sample size, indicates that urination postures occurred during significantly more sessions with scent gland secretions than with CHzC12 (P < 0.05), but responses to snake- and alligator-treated papers did not differ significantly (0.10 > P > 0.05). DISCUSSION
The reactions of predatory mammals to snake scent gland secretions were observed by Price and LaPointe (1981), who presented desert kingsnake (Lampropeltis getulus splendida) exudates on food to badgers, coyotes, and other North American carnivores. The coyotes' responses were not specifically described, but Price and LaPointe (1981) characterized the overall reactions to scent gland secretions by the carnivores they examined as "aversionary." The results of our first two field experiments indicate that coyotes rubbed and rolled on rattlesnake scent gland secretions, as they did on the coyote attractant, CDCL. In our third field test, however, rubbing and rolling did not occur to scent gland extracts at a rate significantly different from that to CH2CI>
CANID RESPONSES TO SNAKE SECRETIONS
t 601
although this behavior was indicated at a significantly higher rate to CDCL. This suggests that the chemicals eliciting rubbing and rolling are not retained (in sufficient quantities) following our methods of extraction, rotary evaporation, and field presentation. Domestic dogs in both our kennel tests also failed to rub and roll on CHzC12 extracts of scent gland secretions significantly more often than on control papers. Scent gland compounds insoluble in CH2C12, i.e., glyco- or mucoproteins and acid mucopolysaccarides (Whiting, 1969; Blum et al., 1971), likely were denatured or removed by our extraction methods. These substances may provide a matrix for the retention of volatile, behaviorally active chemicals. Bullard (1982) states that most, if not all, canids rub and roll on putrefying materials. Fox and Cohen (1977) report rubbing and rolling by a variety of canids to natural and synthetic substances, including skatole and propionic acid. Although sterols and fatty acids have been indicated in snake scent glands (Blum et al., 1971; Simpson et al., 1988), Oldak (1976) failed to detect these or other nonpolar lipids in his thin-layer chromatograms of crotaline scent gland secretions, including those of a rattlesnake. Razakov and Sadykov (1986) identified fatty acids from the scent glands of the mamushi (Agkistrodon halys), indicating that these compounds are present in at least some crotalines. Chemicals in the secretions of C. atrox need to be identified and tested as synthetic analogs to determine those that elicit canid responses. In our study, Staffordshire terriers exhibited urination postures more frequently to rattlesnake secretion extracts than to CH2C12, hut snake- and alligator-treated papers did not differ significantly in their effects. Urination (or urination posturing) was not scored during our first kennel survey, and it occurred at too low a frequency for statistical analysis during our field studies. The tendency of canids to urinate on conspecific urine deposits often is interpreted as a form of scent marking. Urination also may be elicited by " n o v e l " odors, such as those arising from other species' feces (Fox and Cohen, 1977). Kool (1981) noted that the dingo (C. familiaris) urinated on a turtle's (Chelodina longicollis) Rathke's gland secretions, which are hypothesized to deter predators. In Kool's study, as with ours, the adaptive significance of eliciting urination in canids is not obvious. Neither are the aversive properties of scent gland secretions evident from our results. In the first field study, designed to measure the potential repellency of snake scent gland secretions, coyotes visited capsules containing these exudates at a rate that was not significantly different from that to capsules containing CHzC12. In the second field study, designed to measure the potential attractiveness of rattlesnake scent gland secretions, coyotes visited stations containing secretion extracts significantly more often than those containing untreated capsules and at a rate not significantly different from that to CDCL. This is worth noting since CDCL has been developed specifically as a lure for these
1602
WELDON AND FAGRE
carnivores and is one of the most effective commercial attractants available (Turkowski et al., 1983). We did not test the palatability of untreated scent gland secretions in our experiments, thus, we cannot relate our results directly to those of Price and LaPointe (1981). However, the tendency of dogs to mouth snake secretionscented filter papers and that of coyotes in the field to pull up (presumably by mouth) snake secretion-scented capsules does not support the hypothesis that these materials are distasteful. Palatability studies using whole, untreated scent gland substances are needed. Scent gland secretions could interfere with predation in ways other than by having aversive effects. Rubbing and rolling, for example, could distract canids from predatory attack. Even the elicitation of mouthing could be advantageous for snakes if scent gland chemicals direct attacks toward the tail, and away from their head and trunk, in a way analogous to that in which visually directed predators are diverted by snake tail displays (Greene, 1987). Insight into the mechanisms by which scent gland secretions benefit snakes could be obtained by observing interactions between snakes and predators and, perhaps, by comparing the survivorship of intact snakes with that of those from which these glands have been surgically removed. Scent glands are homologous among extant snakes. Since the common ancestor of Recent snakes lived during the Upper Cretaceous (Rage, 1987) this can be considered a minimum age for these organs. Carnivores did not appear until during the Tertiary (Ewer, 1973). Given this chronology, it is clear that snake scent glands did not evolve as a response to these mammalian predators. The antipredator function(s) of the exudates produced in these organs needs to be tested on other taxa, e.g., varanid lizards and crocodilians. Functions and effects other than the production of predator repellents have been attributed to scent glands, including that of pheromone release. Although experiments suggest that blindsnake (Leptotyphlops dulcis) scent gland chemicals attract conspecifics (Watkins et al., 1969), attempts to elicit trail-following with these substances in other snakes have failed (reviewed in Ford, 1986). Graves and Duvall (1988) report that airborne presentations of scent gland chemicals from the prairie rattlesnake (C. viridus) elicit increased heart rates in conspecifics, and they suggest, following Brisbin's (1968) observations on a common kingsnake (Lampropeltis getulus), that these organs release alarm pheromones. Even so, whether scent gland chemicals function primarily to alert conspecifics is uncertain since alarm properties often are derived secondarily from compounds released during predatory attack (Weldon, 1983).
Acknowledgments--B. Beaver helped with preliminary observations on dogs' reactions to scent gland chemicals. L. MacDonald and M. Phi Dao lent valuable help in testing domestic dogs. R.S. Gray, H.B. Harper, and C. Toth kindly allowed us to test their Staffordshire terriers and
CANID RESPONSES TO SNAKE SECRETIONS
1603
assisted in the process. G. Drew, D. Martin, and C. Roberts offered expert field assistance in tests with coyotes. P. Freed, C. Roberts, J.T. Simpson, G. Wills, and T. Wilson helped collect scent gland secretions. D. Synatske and the staff at the Chaparral Wildlife Management Area, Texas Parks and Wildlife Department, allowed us to conduct our studies at their facilities. R.W. Bullard and H.W. Greene commented on earlier drafts of the manuscript. This study was supported by grants from the Whitehall Foundation (to P.J.W.), the Texas Agricultural Experiment Station (through D.B.F.; contribution #23293), and a Texas A&M minigrant (to D.B.F.). To all we are grateful.
REFERENCES BELLAIRS, A. D'A. 1950. The limbs of snakes with special reference to the hind limb rudiments of Trachyboa boulengeri. Br. J. Herpetol. 4:73-83. BLUM, M.S., BYRD, J.B., TRAVIS,J.R., WATK1NS,J.F., II, and GEHLBACH, F.R. 1971. Chemistry of the cloacal sac secretion of the blind snake Leptotyphlops duleis. Comp. Biochem. Physiol. 38:103-107. BRISBIN, I.L., JR. 1968. Evidence for the use of postanal musk as an alarm device in the king snake, Lampropeltis getulus. Herpetologica 24:169-170. BULLARD, R.W. 1982. Wild canid associations with fermentation products. Ind. Eng. Chem. Prod. Res. Div. 21:646-655. BULLARD, R.W., TURKOWSKI,F.J., and KtLBURN, S.R. 1983. Responses of free-ranging coyotes to lures and their modifications. J. Chem. Ecol. 9:877-888. BURKETT, R.D. 1966. Natural history of the cottonmouth moccasin, Agkistrodon piscivorus (Reptilia). Univ. Kans. Publ. Mus. Nat. Hist. 17:435-491. DUW,LL, D., KING, M.B., and GUTZWILLER, K.J. 1985. Behavioral ecology and ethology of the prairie rattlesnake. Nat. Geogr. Res. 1: 80-111. EISNER, T., CONNER, W.E., HICKS, K., DODGE, K.R., ROSENBERG, H.I., JONES, T.H., COHEN, M., and MEINWALD, J. 1977. Stink of the stinkpot turtle identified: r acids. Science 196:1347-1349. EWER, R.F. 1973. The Carnivores. Cornell University Press, Ithaca, New York. FITCH, H.S. 1960. Autecology of the copperhead. Univ. Kans. Publ. Mus. Nat. Hist. 13:85-288. Fn'CH, H.S. 1965. An ecological study of the garter snake, Thamnophis sirtalis. Univ. Kans. Publ. Mus. Nat. Hist. 15:493-564. FORD, N.B. 1986. The role of pheromone trails in the sociobiology of snakes, pp. 309-312, in D. Duvall, D. Muller-Schwarze, and R.M. Silverstein (eds.). Chemical Signals in Vertebrates 4: Ecology, Evolution, and Comparative Biology. Plenum Press, New York. Fox, M.W., and COHEN, J.A. 1977. Canid communication, pp. 728-748, in T.A. Sebeok (ed.). How Animals Communicate. Indiana University Press, Bloomington. GEHLBACr~, F.R., WATKINS, J.F., II, and RENO, H.W. 1968. Blind snake defensive behavior elicited by ant attacks. BioScience 18:784-785. GRAVES, B.M., and DUVALL, D. 1988. Evidence of an alarm pheromone from the cloacal sacs of prairie rattlesnakes. Southwest Nat. 33:339-346. GREENE, H.W. 1987. Antipredator mechanisms in reptiles, pp. 1-151, in C. Gans and R.B. Huey (eds.). Biology of the Reptilia, Vol. 16. Alan R. Liss, New York. KLAUBER, L.M. 1972. Rattlesnakes, Their habits, Life Histories, and Influence on Mankind, Vol. 2. University of California Press, Berkeley. KOOL, K. 1981. Is the musk of the long-necked turtle, Chelodina longicollis, a deterrent to predators? Aust. J. Herpetol. 1:45-53.
1604
WELDON AND FAGRE
OLDAK, P.D. 1976. Comparison of the scent gland secretion lipids of twenty-five snakes: Implications for biochemical systematics. Copeia 1976:320-326. PEREZ, J.C., HAWS, H.C., GARCIA, V.E., and JENNINCS, B.M., III. 1978. Resistance of warmblooded animals to snake venoms. Toxicon 16:375-383. PRICE, A.H., and LAPOINTE, J.L. 1981. Structure-functional aspects of the scent gland in Lampropeltis getulus splendida. Copeia 1981:138-146. RAGE, J.C. 1987. Fossil history, pp. 51-76, in R.A. Seigel, J.T. Collins, and S.S. Novak, (eds.). Snakes: Ecology and Evolutionary Biology. Macmillan, New York. RAZAKOV, R.R., and SADYKOV,A.S. 1986. Study of complex mixtures of natural substances by the refocusing and DADI methods. VI. Components of the secretion of the pre-anal [sic] gland of some poisonous snakes. Khim. Prir. Soedin 4:421-423. SIMPSON,J.T., WELDON, P.J., and SHARP,T.R. 1988. Identification of major lipids from the scent gland secretions of Dumeril's ground boa (Acrantophis dumerili Jan) by gas chromatographymass spectrometry. Z. Naturforsch. 43c:914-917. TURKOWSKI,F.J., POPELKA,M.L., GREEN, B.B., and BULLARD,R.W. 1979. Testing the response of coyotes and other predators to odor attractants, pp. 255-269, in J.R. Beck (ed.). Test Methods for Vertebrate Pest Control and Management Materials. ASTM STP 680. American Society for Testing and Materials, Philadelphia, Pennsylvania. TURKOWSKI,F.J., POPELKA,M.L., and BULLARD, R.W. 1983. Efficacy of odor lures and baits for coyotes. Wildl. Soc. Bull. 11:136-145. WATKINS,J.F., II, GEHLBACH,F.R., and KROLL,J.C. 1969. Attractant-repellent secretions of blind snakes (Leptotyphlops dulcis) and their army ant prey (Neivamyrmex nigrescens). Ecology 50:1098-1102. WELDON, P.J. 1983. The evolution of alarm pheromones, pp. 309-312, in D. Multer-Schwarze and R.M. Silverstein, (eds.). Chemical Signals in Vertebrates 3. Plenum Press, New York. WELDON, P.J. 1988. Feeding responses by snappers (genus Lutjanus) to the yellow-bellied sea snake (Pelamis platurus). Zool. Sci. 5:443-448. WELDON, P.J., SHAFAGATI,A., and WHEELER, J.W. 1988. Lipids from the paracloacal glands of the American alligator (Alligator mississippiensis). Lipids 23:727-729. WHmNG, A.M., 1969. Squamate cloacal glands: Morphology, histology and histochemistry. PhD dissertation. Pennsylvania State University.
