THE JOURNAL OF EXPERIMENTAL ZOOLOGY 276:186-192 (1996) Olfaction and the Homing Ability of Pigeons in the Southeastern United States VERNER P. BINGMAN AKD SILVANO BENVENUTI Department of Psychology, Bowlitzg Green State University, Bowling Green, Ohio 43403 ABSTRACT The importance of atmospheric odors for homing pigeon navigation was tested using birds from a loft located in Savannah, GA, in the southeastern United States. When released from a familiar training site, control pigeons and pigeons given intranasal injections of zinc sulfate to produce mosmia both displayed good homeward orientation and homed quickly. When released from three unfamiliar release sites, in contrast, control birds tended to orient southeast, while zinc sulfate-treated birds were more likely to fly northwest. More importantly, while the majority of control pigeons returned to the home loft, few of the zinc sulfate-treated birds returned. The good performance of both groups from the familiar site indicates that zinc sulfate treatment does not impair the general motor ability or motivation of homing pigeons. Therefore, the observed impairment in homing success of the zinc sulfate-treated pigeons from the unfamiliar locations presuriably reflects an impaired ability to use atmospheric odors to navigate home. As such, the data support the hypothesis that successful homing pigeon navigation is based on the perception of atmospheric odors and that olfactory navigation is the primary mechanism used by pigeons over a broad range of geographic areas to approximate their relative position with respect to home from unfamiliar locations. o 1996 Wiley-Liss, Inc. The ability of homing pigeons to return home from distant, unfamiliar release sites is thought to be based on a two-step navigational process consisting of independent map and compass mechanisms (Kramer, '52; Wa.llraff, '90). Although the sensory basis of the compass mechanisms has been we11 described, both the sun and the earth's magnetic field are used (Kramer, '53;Wiltschko and Wiltschko, 'SO), and understanding the sensory basis of the navigational map has been more difficult (Gould, '82; Wallraff, '90; Able, '96). A large number of experiments have been carried out in Italy, Germany, and the United States, the results of which offer compelling evidence that airborne, olfactory cues play a critical role in the operation of the homing pigeon navigational map (for review see Papi, '90; Wallraff, '90). Despite this vast body of supporting evidence, the olfactory hypothesis as an explanation for the operation of the homing pigeon navigational map has been the source of ccmsiderable controversy (Schmidt-Koenig, '87; Able, '96). The most serious argument directed against the olfactory hypothesis is that atmospheric odors may not be similarly important as a navigational cue for all pigeons. For example, it, has been argued that olfaction is less important in some regions of Germany and the United States than perhaps in Italy 0 1996 WILEY-LISS, INC. (Wiltschko et al., '87a) or that the use of olfaction is dependent on the type of experience a pigeon has while young (Wiltschko et al., '87b). However, studies by other workers have contradicted these findings (Benvenuti and Brown, '89; Benvenuti et al., '90; Bingman and Mackie, '92). In summary, although most researchers agree that olfaction plays an important role in homing pigeon navigation (Able, '96), there remains some question with respect to how ubiquitous atmospheric odors are as a navigational cue. To address this issue, one obvious research strategy is to examine the importance of olfaction for pigeon navigational behavior in a number of different geographical regions. In the present study, we report the results of an experiment designed to examine the importance of atmospheric odors for the navigational performance of homing pigeons from Savannah, GA, which lies near the Atlantic coast of the southeastern United States. Savannah is interesting in two respects. First, it is in the United States where olfaction has been previously reported to be perhaps less important as a navigational cue (Wiltschko et al., '87a; but Received February 29, 1996;revision accepted May 24, 1996. Address reprint requests to Verner P. Bingman, Department of Psychology, Bowling Green State University, Bowling Green, OH 43403. PIGEON OLFACTORY NAVIGATION IN GEORGIA see Benvenuti and Brown, '89; Bingman and Mackie, '92). Second, Savannah lies in a coastal, hot, and humid environment that might be environmentally well suited for learning the spatial distribution of atmospheric odors and consequently an olfactory navigational map (Waldvogel, '87). This latter consideration would lead one to expect a map strongly influenced by atmospheric odors in this region. Indeed, the results are consistent with the hypotheses of an important role for olfaction in homing pigeon navigation in the southeastern United States. MATERIALS AND METHODS Experimental animals A pool of 53 homing pigeons, reared at a loft in Savannah, GA, was used. It is worth noting that the location of the loft was in the backyard of a house located in a heavily wooded residential neighborhood. The local environment around the loft was very reminiscent of the garden loft location of the Wiltschko laboratory in Frankfurt, Germany, where pigeons are reported to acquire a navigational map relatively uninfluenced by atmospheric odors (Wiltschko et al., '87b). The birds were all born in the spring of 1995. They were housed in a loft that had an opening facing the north (trap and standing area) but was otherwise closed. Beginning at about 2 months of age, the birds underwent a modest training regimen which consisted of continual free flights around the loft, followed by four group releases from different directions not more than 10 km from home. Following this preliminary training, all pigeons were then given five training releases from a location (familiar site, see below) 34 km west of the home loft. The first three training releases from this location were group releases. For the last two training releases, birds were released in groups of three. Following training, the experimental releases began. Experimental releases On September 29, 1995, 24 pigeons were arbitrarily chosen and subjected t o a n intranasal injection of zinc sulfate heptahydrate dissolved in distilled water. Zinc sulfate alters or destroys the olfactory mucosa, effectively rendering pigeons anosmic until regeneration takes place (Benvenuti e t al., '92; Schlund, '92). Zinc sulfate administration is generally considered an anosmic procedure with few if any side effects (Benvenuti and Gagliardo, in preparation). Two 187 milliliters of a 4% solution was injected into each choana and the solution was allowed to flow out of the nostril (Henvenuti et al., '92). Twenty-four control pigeons were subjected to the same treatment except they were injected with avian Ringer solution. On September 30, 1995, 12 zinc sulfate and 12 control pigeons were released from the familiar training site west of home (release site A). Later on the same day, the remaining pigeons plus those that had returned from the familiar training site experimental release were taken to a location approximately 10 km east of the loft and group released. This was done in an attempt to eliminate a possible acquired tendency on the part of the pigeons t o fly east as a result of the five training releases from the familiar site. On October 1, 1995, 10 zinc sulfate birds and 10 control birds, which were not released from the familiar site, were released from an unfamiliar release site 51 km southwest of home (release site B). On October 2, 1995, a third release took place from an unfamiliar site 54 km northwest of the home loft (release site C). Of the 13 control birds released, 11 participated in the familiar training site release (A) and 2 were involved in an experimental release for the first time. Of the 12 zinc sulfatetreated animals, 10 participated in the familiar training site release (A) and 2 were involved in an experimental release for the first time. After the first three releases, the remaining zinc sulfate-treated birds (N = 4) were each given an additional 1 ml of zinc sulfate solution in each choana. Further, 3 previous control birds and 5 experimentally native birds were given 2 ml of the zinc sulfate solution in each choana. Therefore, for the last experimental release on October 6, 1995, from a location 34 km north of home (release site D, actually in South Carolina), 12 zinc sulfate-treated birds were released. Twelve control birds that had returned from releases B and C also participated. Except for release C, which took place under sunny conditions, all releases took place under partly cloudy, humid conditions with winds less than 20 km/h. However, the disk of the sun was always visible when a bird was released. For all experimental releases, pigeons were released singly, alternating between treatment groups, and followed with binoculars (10 x 40) until disappearing from view. Vanishing bearings and vanishing times were recorded with compass and stopwatch, respectively. Arrival times were recorded by an observer stationed at the loft. 188 V.P. BINGMAN AND S . BENVENUTI Statistics Statistical analysis of the vanishing bearings ence in homing time approached significance (U = 41.5, P c 0.10 two-tailed). No difference in homing success was recorded (Table 2). was performed with a Ftayleigh test (Batschelet, From the unfamiliar release sites the pattern ,811. Between group calmparisons of vanishing was again very different. Homing times were poor bearings were performed with a Watson U2-test (Fig. 21, and because few zinc sulfate birds re(Batschelet, '811, homing times [only from the faturned, no between group statistical comparisons miliar training site release (A) because zinc sulwere made. However, there was a striking befate-treated birds generally did not return from tween group difference in homing success (Table the unfamiliar sites] with a Mann-Whitney U-test, 2, the number of birds returning, excluding those and homing success with a chi-square test. that returned with bird(s) from the other group, RESULTS compared to the number of birds that did not return). Basically, the majority of control birds reVanishing bearings turned from all three unfamiliar locations, while Except for a narrow styetch of salt marshes that few zinc sulfate-treated birds returned. A differborder the Atlantic Ocean, coastal Georgia and ence in homing success was found for all three South Carolina are characterized by extensive releases (at least P c 0.05). Indeed, for all three pine/oak forests. As a result, we were unsuccess- releases, 22 of a possible 33 control birds returned, ful in finding any location where birds could be while only 5 of a possible 30 zinc sulfate-treated followed until vanishing in all directions. The four birds returned. release sites used generally allowed us t o follow Taken together, the homing success data stand the pigeons for a reasonable distance (at least 1 in strong support of the hypothesis that atmokm), but it was difficult to keep birds until van- spheric odors are a critical source of navigational ishing because they would regularly disappear be- information that permits homing pigeons to navihind distant trees. Most of the vanishing bearings gate home from unfamiliar locations in the southrecorded were based on birds lost behind trees. eastern United States. However, because we were able to follow all birds for at least 1min, the recorded bearings probably DISCUSSION provide a reasonable, albeit not totally satisfacThe general failure of the zinc sulfate-treated tory, indication of the pigeons' directional prefer- birds to return home from the unfamiliar release ences (Wallraff, '67). sites provides one more piece of evidence to the From the familiar site (A) both groups oriented growing body of research findings supporting the homeward and no differences in vanishing bear- critical role of atmospheric odors in homing piings were recorded (Fig. 1, Table 1).From the geon navigation (Papi, '90; Wallrd, '90;Able, '96). three unfamiliar release sites (B-D) the results However, one must be cautious with any kind of were quite different. The control birds consistently anosmic procedure that may potentially cause side oriented to the southeast regardless of release site effects unrelated to navigation that would impair (Fig. 1, Table 1).In contrast, the zinc sulfate- homing performance. The performance of the zinc treated birds oriented northwest from two release sulfate-treated birds from the familiar release site, sites (B, C), where their orientation differed from however, clearly indicates that the experimental controls at the P c 0.10 level. From release site treatment did not lead to any impairment that D, however, they oriented southeast in a manner would preclude flying home. This finding is conindistinguishable from controls. sistent with other studies (Benvenuti et al., '92; In summary, the control birds oriented in a man- Benvenuti and Gagliardo, in preparation) demonner suggesting a strong tendency to fly southeast strating no effect of zinc sulfate treatment on hom[preferred compass direction (PCD); Wallraff, '781 ing ability from familiar sites where presumably immediately upon release. Zinc sulfate produced non-olfactory information acquired during previanosmia, causing a change in that preference t o ous training flights can direct the pigeons home. the northwest, but not under all conditions. Nonetheless, the zinc sulfate-treated birds from the familiar site were slightly slower than the conHoming time and homing success trols. The small absolute difference in homing From familiar release site A, the median hom- time (Fig. 2) could be used t o argue that the zinc ing time of the control pigeons was 42 min and sulfate treatment may have had a modest, nonthe zinc sulfate birds 55 min (Fig. 2). This differ- specific effect on homing. However, a similar dif- 189 PIGEON OLFACTORY NAVIGATION IN GEORGIA N N 47O N 54 km 34 km Fig. 1. Vanishing bearings of control (open dots) and zinc sulfate-treated (closed dots) pigeons from the familiar training location (A) and three unfamiliar release sites (€3-D). Each dot represents the vanishing bearing of one bird. Arrows in each circle represent the mean vectors for each group (open heads, controls; filled heads, zinc sulfate-treated). The length of an arrow corresponds to the mean vector length, which can be read with the scale in A (solid lines, P < 0.05; dashed lines, P > 0.05; see Table 1). Arrows outside each diagram and associated numbers identify the distance and direction home. North (N) is at the top of each diagram. ference was not observed in another experiment where pigeons with similar experience were given intranasal administration of zinc sulfate (Benvenuti et al., '92). Therefore, because the difference was so small, a more likely explanation for the difference is that even from familiar sites, in the southeastern United States atmospheric odors may make a navigational contribution, and the marginally slower homing times of the zinc sulfatetreated birds were a consequence of not being able to use that information. Although the purpose of this experiment was to examine the importance of atmospheric odors for navigation in the southeastern United States, from a discussion perspective, the normative homing data from the controls are also of interest. It has often been noted that the homing performance of pigeons in Italy is unusually good compared t o other locations, and it has been suggested that the coastal environment characteristic of Italy in some way makes for easier "olfactory" navigation (Waldvogel, '87). Therefore, it was expected that the birds in coastal Georgia would show similarly impressive navigation. They did not. Indeed, examining Figure 2 and Table 2 one is struck by the poor homing times and homing success of the V.P. BINGMAN AND S.BENVENUTI 190 TABLE f. Analysis of vanishing bearings for the four experimental releases' N Familiar site A (H = go", D = 34 km) Control ZnSO4 Unfamiliar Site B (H= 47", 13 = 5 1 km) Control ZnS04 Unfamiliar Site C (H = 138", D = 54 km) Control ZnSOd Unfamiliar Site D (H = 180°, D = 34 km) Control ZnSOd a 11(12) 11(12) Group difference r ggo***.. 1050x*x* 0.85 0.81 P>O.lO 128"" 0.56 322""** 0.79 11(13) 11(12) 11goh-s 326""* 0.40 0.60 P<O.lO 12 (12) 10 (12) 1130*V"* 1270**** 0.83 0.78 P>O.10 7 (10) 9 (10) P<O.lO 'N t number of vanishing bearings recorded (number in parentheses = number of birds released); a = means vector angle (north = 360");NS = non-significant; *P<O.10; **P.:O.O5; **"P<0.01; ****PcO.OOl as measured by the Rayleigh test); r = mean vectbr length. Between group comparisons were performed with a Watson-Williams U2-test. control birds. In this context, it is worth noting t h a t the environment in Georgia differs from Tuscany, Italy, in two respects. As mentioned earlier, the loft of the pigeons was located in a wooded residential area with little exposure to unobstructed winds and was similar to the typical rearing conditions of t h e Wiltschko laboratory in Frankfurt, Germany. In Pisa (Tuscany), the lofts are located in a n open field with full exposure to passing winds. Wind exposure plays a critical role in olfactory navigational ma p learning (e.g., Baldaccini e t al., '74) and limited exposure t o winds may impair development of a n olfactory FAMILIAR e, A 0r L A 1 map (Wiltschko et al., '87b), contributing perhaps to the relatively poor performance of the control birds. Additionally, coastal Georgia is characterized by what appeared to be continuous pine/oak forests for at least 100 km in all directions away from the coast. This homogeneous environment, in contrast to the area around Pisa with its many mountains and varying habitats, may provide insufficient "spatial atmospheric heterogeneity" to enable birds to learn a precise olfactory map that would permit efficient navigation home. Another notable feature of the behavior of the control birds was the consistent southeasterly ori- 2 3 4I SITE 5I HR LATER -@QmL LOST UNFAMILIAR SITES 0 .wy Fig. 2. Homing times of control (open dots) and zinc sulfate-treated (filled dots) pigeons from the familiar training location (A) and three unfamiliar release sites (B-I>). Each dot represents the homing time for one bird. Note: Pigeons that returned with a bird(s) from the other treatment group were excluded. Pigeons that returned with a bird(s) from the same treatment group were treated as independent data points. PIGEON OLFACTORY NAVIGATION IN GEORGIA TABLE 2. Summary of homing success Familiar site A Control ZnS04 Unfamiliar site B Control ZnS04 Unfamiliar site C Control ZnS04 Unfamiliar site D Control ZnSOl Returned' Lost Group difference 11 10 1 2 P>O.10 7 1 1 5 P<O.O2 7 1 6 11 P<O.O2 8 3 4 9 P<0.05 'Birds that returned together with a bird(s) from the other treatment group were excluded from the analysis. Birds that returned together with a bird from the same treatment group were treated as independent data points. Between group comparions were performed with a chi-square test. 191 CONCLUSIONS The data presented in this paper indicate that for pigeons in coastal Georgia atmospheric odors represent a critical source of environmental information that enables pigeons t o determine their relative location with respect to home. Additionally, the data suggest that the pigeons in this region show a strong tendency to vanish in a southeasterly direction and that this southeasterly PCD is useful in getting birds t o the coast and likely influenced by the perception of atmospheric odors. Taken together, the data support a large body of evidence identifying atmospheric odors as a critical element in homing pigeon navigation. ACKNOWLEDGMENTS We sincerely thank Burt Oostlander, Wendy Oostlander, and Dick Schultz whose hospitality entation. This southeasterly PCD (Wallraff, '78) and homing pigeon knowledge made the experiis interesting because it is perpendicular to and ment both a success and pleasure to carry out. therefore the shortest route to the coast, which is This work was supported by grants from NIH oriented northeast-southwest. In other words, (1R03 MH52315) t o V.P.B., NATO (CRG 950084) given the location of the loft just a few kilometers to V.P.B. and F. Papi, and the Italian National from the coast, a navigational strategy of first orito F. Papi. Research Council enting in the direction of the coast and then moving parallel to the coast either northeast or LITERATURE CITED southwest is a reasonable, albeit not the most efficient, navigational strategy. Although the ten- Able, K.P. (1996) The debate over olfactory navigation by homing pigeons. J. Exp. Biol., 199:121-124. dency to move in a specific compass direction is not as strong in Pisa, a tendency t o move in an Baldaccini, N.E., S. Benvenuti, V. Fiaschi, P. Ioale, and F. Papi (1974) Pigeon homing: Effects of manipulation of approximate southwesterly direction there has sensory experience at home site. J. Comp. Physiol., been well described (Ioal6, '95). An approximate 94:85-96. southwesterly PCD in Pisa is also perpendicular Batschelet, E. (1981) Circular Statistics in Biology. Academic Press, New York. to the approximate northwest-southeast oriented coastline there, suggesting that pigeons living Benvenuti, S., and LA. Brown (1989) The influence of olfactory deprivation on homing of experienced and inexperinear coasts may generally display initial orientaenced American pigeons. Behaviour, 111~113-128. tion biased in the direction of the coast. Benvenuti, S., V. Fiaschi, A. Gagliardo, and P. Luschi (1990) Of further interest is that the southeasterly oriPigeon homing: A comparison between groups raised under different conditions. Behav. Ecol. Sociobiol., 27:93-98. entation of the Georgia birds appeared dependent on the perception of atmospheric odors. The anos- Benvenuti, S., P. IoalB, A. Gagliardo, and F. Bonadonna (1992) Effects of zinc sulfate-induced anosmia on homing behaviour mic, zinc sulfate-treated birds orientated northwestof pigeons. Comp. Biochem. Physiol., 103A:519-526. erly from two unfamiliar release sites. However, Bingman, V.P., and A. Mackie (1992) Importance of olfaction they did orient southeasterly from release site D. for homing pigeon navigation in Ohio, USA. Ethol. Ecol. Evol., 4:395-399. Although highly speculative, we wish t o suggest that, in coastal Georgia, southeasterly orientation Gould, J.L. (1982) The map sense of pigeons. Nature, 296~205-211. is triggered by some atmospheric quality associ- Ioale, P. (1995) Preferred compass directions of homing piated with the Georgia coast. In the absence of such geons in Italy. Boll. Zool., 62:13-21. information, the initial orientation of the birds is Kramer, G. (1952) Experiments on bird orientation. Ibis, 94.265-285. biased to the northwest. But why did the zinc sulfate-treated birds orient southeast from release Kramer, G. (1953) Die Sonnenorientierung der Vogel. Verh. Dtsch. Zool. Ges., 195277-84. site D? Although we have no explanation for this, Papi, F. (1990) Olfactory navigation in birds. Experientia, it is worth noting that spontaneous 180" reversals 46:352-363. in PCD directions have been reported (WallrafY, '91). Schlund, W. (1992) Intra-nasal zinc sulfate irrigation in pi- 192 V.P. BINGMAN AND S.BENVENUTI geons: Effects on olfactory capabilities and homing. J. Exp. Biol., 164:171-187. Schmidt-Koenig, K. (1987) Bird navigation: Has olfactory orientation solved the problem? Q. Rev. Biol., 62:31-47. Waldvogel, J.A. (1987) 0lfa.ctory navigation in homing pigeons: Are the current models atmospherically realistic? Auk, 104:369-379. Wallraff, H.G. (1967) The present status of our knowledge about pigeon homing. In: Proceedings of the XIV Ornithological Congress. D.W. Snow, ed. Blackwell Scientific Publications, Oxford, pp. 331-358. Wallraff, H.G. (1978) Preferred compass directions in initial orientation of homing pigeons. In: Animal Migration, Navigation, and Homing. K. Schmidt-Koenig and W.T. Keeton, eds. Springer-Verlag, Berlin, pp. 171-183. Wallraff, H.G. (1990) Navigation by homing pigeons. Ethol. E d . E d . , 2~81-116. Wallraff, H.G. (1991) Conceptual approaches to avian navigation systems. In: Orientation in Birds. P. Berthold, ed. Birkhauser Verlag, Basel, pp. 128-165. Wiltschko, W., and R. Wiltschko (1988)Magnetic orientation in birds. In: Current Ornithology, Vol. 5. R.F. Johnston, ed. Plenum Press, New York, pp. 67-121. Wiltschko, W., R. Wiltschko, and C. Walcott (1987a) Pigeon homing: Different effects of olfactory deprivation in different countries. Behav. Ecol. Sociobiol.,21:333-342. Wiltschko, W., R. Wiltschko, M. Gruter, and U. Kowalski (1987b) Pigeon homing: Early experience determines what factors are used for navigation. Naturwissenschaften, 74:196-198.