Landmarks
When Martin Michener and I began to follow homing pigeons by airplane back in the early 1960s we thought that knowing the paths pigeons took from release to home would give us important insights into their navigation. What we found was that pigeons, released at the same location, generally followed somewhat different routes home. (Fig. 1) We believe that these routes were really different even for the same pigeon because, for light colored pigeons, we could watch them fly just at the level of the tree tops. A few pigeons on occasion flew much higher, but generally flight was close to the ground. That means that the birds couldn't be following a consistent set of familiar landmarks to find their way back to the loft even on their training flights, Michener and Walcott (1967)
. There was no evidence that pigeons consistently followed a prominent road from the release point, nor did it seem likely that distant landmarks were usually important. Recently, Lipp et al. (2004) have been using GPS systems to track pigeons. They find that their pigeons not only follow roads but also railroads. Only as they approach the loft do they desert the road. This behavior certainly differs from that we saw in our Lincoln, Massachusetts pigeons or those we followed around Ithaca, New York. Clearly, at least some pigeons make use of landmarks in their homing.
Yet the experiments of Schlichte and Schmidt-Koenig (1971) show that pigeons wearing frosted lenses would still orient toward home and many could even find the loft. Tracking these pigeons by airplane showed that they made straight tracks toward the loft but many landed within about 10 km of the loft (Schmidt-Koenig and Walcott, 1978). For these birds visual landmarks were not required either at the release point or to fly a straight path towards the loft; it was only in the final stages of finding the loft itself that they seemed crucial.
Individuality
Taking individual pigeons that had repeatedly flown over the same course to a more distant, unfamiliar release site revealed a number of different strategies. One bird flew for a few miles in the compass direction that would have led it home from the familiar release site. It then stopped, perched for about 20 minutes, and when it resumed flight, headed directly for home. Another pigeon that had been trained from Orange, Masachusetts and regularly passed a prominent hill (Mt. Wachusett) on its way home was released at Worcester airport. Instead of flying East toward home, it headed North flying toward Mt. Wachusett even though that was not the way home. Having flown around the mountain, it turned and headed for home (Fig. 2). Released in other locations, this pigeon always headed for the nearest mountain before flying home. Yet another pigeon always oriented very accurately toward the loft, but as it got within about 10 km it would stop, find a person out gardening and we would get a telephone call to come and collect it! Another pigeon released for the first time at Worcester airport started out headed Southwest and made a huge circular diversion to the South.
I offer these stories only as evidence that different individual pigeons raised and trained under identical conditions may well prefer to use or weight their orientation cues differently. When working with large groups of pigeons, these individual idiosyncrasies become submerged in the group data. It is only by studying individuals that they become apparent.
Compass systems
Sun compass
There is general agreement that homing pigeons use the sun as a compass reference. A six-hour clock shift results in roughly a 90° error in the pigeon's orientation. (Schmidt-Koenig, 1961, 1979 [summary]; Keeton, 1974). Yet if one looks at releases at different sites, the actual amount of shift varies widely. At some release points, the bearings are shifted much more than 90°, at others much less (Keeton and Walcott, unpublished data). Whether this is the effect of a different "release site bias" (Keeton, 1973) at the different sites or represents a difference in the effect of clock shifts on the sun compass is not clear.
Magnetic compass
Under overcast skies, Keeton (1974) showed that clock shifts have no effect on the pigeon's orientation. This means that the pigeons could not, in some mysterious way see the sun through the clouds, for, if they did, their bearings would be shifted. Further, it means that whatever the basis of their "map," it doesn't depend on an accurate sense of time. Yet the pigeons were still well oriented towards home. Keeton's also showed that pigeons with small bar magnets glued on their backs were often disoriented and would home more slowly than controls wearing brass bars. Keeton's (1971,1972) original results included only those releases in which the birds carrying brass bars were oriented. Bruce Moore (1988) reanalyzed all Keeton's data and included all the releases, even those in which the brass bar controls were disoriented. Under these conditions the difference both in orientation and homing time between brass and magnet birds disappears. This does not invalidate Keeton's essential point that when there is a difference between pigeons carrying magnets and those carrying brass bars, it is the magnet birds that show either poor orientation or slower homing.
Under sunny skies, magnets had very little effect except when pigeons were suddenly released at much longer than their accustomed distances. Under these conditions magnets caused poor orientation even under sun (Keeton, 1972 and Walcott, 1974, 1980a). Interestingly, young pigeons, released at an unfamiliar location for the first time under sunny skies, were disoriented by magnets (Keeton, 1972). This implies that young pigeons, like older birds under overcast, are using the earth's magnetic field. Wiltschko and Wiltschko (1981, 1990) have shown that young pigeons learn the sun-compass by reference to the direction of the earth's magnetic field. But why should they make this shift? Magnetic information should be available all the time, not just when the sun shines. There must be some important reason for pigeons to learn the complicated rules of sun movement rather than continuing to rely on the magnetic compass. Cochran et al. (2004) have shown that Catharus thrushes calibrate their magnetic compass on a daily basis using twilight cues, apparently just the reverse of what homing pigeons do.
Paired coils
Walcott and Green (1974) and Visalberghi and Alleva (1979) used paired coils to generate a uniform magnetic field around the pigeon's head. Initially, the orientation of pigeons wearing coils generating a small 0.1 gauss field, about 1/6th of the earth's magnetic field, was compared to the orientation of birds with coils but with no current flowing through them and hence, no magnetic field. Under sunny skies birds wearing the coils with current had more scattered vanishing bearings than those without. (Walcott, 1977). A second series of experiments using coils generating earth strength magnetic fields utilized two different field polarities. By reversing the polarity of the battery either of two magnetic fields could be generated: either a NUP (in which a magnetic compass placed between the two coils had its North seeking needle pointing towards the coil on the top of the pigeon's head) or a SUP with the opposite polarity. This difference in polarity had only a small effect under sunny skies. Pigeons released with an earth strength magnetic field around their heads vanish in slightly different directions depending on the polarity of the magnetic field. This is a consistent and repeatable effect and takes place under sunny skies (Walcott, 1977). While the difference in orientation between pigeons carrying NUPs and SUPs was only 18 to 34 degrees at vanishing, the time course of their directional choices reveals much greater differences shortly after release (Fig. 3). Pigeons released North and South of the loft with NUP coils vanish some 18 to 34 degrees to the left of Sups and birds with no current. When released from the West, NUPs and birds with no current vanished 19 degrees to the right of the SUPs.
Repeating the same experiment under overcast showed that pigeons wearing NUP coils headed away from home while birds with SUP coils homed normally (Walcott and Green, 1974; Visalberghi and Alleva, 1979). This finding strengthened the idea that pigeons unable to see the sun rely on the earth's magnetic field as a compass. The dominance of the sun compass becomes particularly dramatic when tracking a pigeon equipped with NUP head coils and it is vanishing in the direction away from home. If the sun suddenly becomes visible for a moment, the pigeon turns and flies directly over the release point heading for home.
Yet the dichotomy between sun and magnetic field is not a simple alternative. Wiltschko and Wiltschko (2001) report that when 6 hr clock-shifted pigeons released under sun showed a deflection of less than the customary 90°, equipped them with bar magnets restored their full 90° shift. They interpret this result to mean that the shifted pigeons were using both their sun and magnetic compasses simultaneously. As Walcott (1977) suggested, there seems be an interaction between the sun and the magnetic compass systems causing the deflection shown by the NUP and SUP pigeons under sun.
Map
Release site bias
Pigeons released at various sites around Ithaca, New York rarely head exactly towards home (Windsor [1975]; see Walcott [1996] for a complete map). Pigeons exhibit what Keeton (1973) called a "release site bias." At Castor Hill, New York, 160 km NNE of Ithaca, for example, a substantial clockwise shift in vanishing bearings takes place under both sun and overcast. This observation seems to eliminate any chance that the bias was a sun compass effect. Furthermore, the clockwise deflection was similar for other birds from other lofts. Remarkably, the amount of bias has varied over the years. As Figure 4 shows, the bias at Castor Hill slowly varies from year to year. For any particular year, it was generally greatest for inexperienced birds released there for the first time and diminished with older. more experienced pigeons. Following these birds by airplane showed that they headed West 25–32 km before they changed course quite abruptly and headed South toward the loft. There were a few birds in the Cornell loft that exhibited no bias whatever at Castor Hill and simply flew straight home. The offspring of these no-bias birds also showed no deflection at Castor Hill. What these birds were doing that differed from what our usual pigeons did is totally mysterious. But the fact that it could be inherited suggests that it has a genetic basis.
Jersey Hill: a Bermuda Triangle
Probably the most dramatic example of a release point problem for Cornell pigeons is the release site at the Jersey Hill Fire tower near Hornell, New York 135 km West of the loft. At this otherwise unremarkable release site, Cornell pigeons vanish in all directions (Fig. 5) and the vast majority never return home. Following some of these birds by airplane, Bruce Moore showed that they wandered over the countryside mostly never getting closer to home. A few finally made it home by flying in a hugely indirect route. Pigeons from several other lofts, even Cornell pigeons raised in other lofts fly straight home from Jersey Hill (Fig. 6). Even pigeons from Trumansburg, 18 km West of Ithaca are well oriented and home normally. Jersey Hill seems to pose problems only for pigeons raised in the Cornell pigeon lofts on Turkey Hill in Ithaca, New York (Walcott and Brown, 1989)! Perhaps pigeons compare some feature remembered from the home loft with that at the release site. For most pigeons this works well, but for Cornell pigeons at Jersey Hill, it fails to provide useful information. Why Cornell pigeons flying incorrect paths from Jersey Hill fail to correct them is puzzling since they fly over areas from which Cornell pigeons home normally.
Loft location
That the location of the pigeon loft makes an important difference in the pigeon's orientation is shown by two experiments. The Wiltschkos had a loft of pigeons in the garden at the Frankfurt zoological institute. Making these pigeons anosmic had essentially no effect on either the pigeons' orientation or homing. Their siblings raised in a loft on the roof of the Institute, four stories above, when made anosmic neither oriented nor homed. This difference could be explained by the difference in the access to wind and the odors it carries to the two lofts. The loft in the garden was sheltered from the wind. Pigeons growing up in this environment may have learned to rely on non-olfactory cues. Pigeons raised in the loft on the roof where there was free access to the wind and the characteristic odors that it brought may have grown up convinced of the utility of olfactory cues (Wiltschko and Wiltschko, 1989).
Magnetic anomalies
Another example of the same kind of phenomenon is the results obtained by releasing pigeons at magnetic anomalies (Wagner [summarized 1983]; Kiepenheuer [1982, 1986] and Walcott [1978]). These anomalies are local distortions in the earth's magnetic field caused by changes in the magnetic permeability of the underlying bed rock. Homing pigeons released at such locations, even under sunny skies are generally disoriented until they leave the distorted area. They then head home normally. That the disorientation at anomalies is really a consequence of the distorted magnetic field is supported by the finding that the degree of disorientation is correlated with the strength of the anomaly (Fig. 7). These results were obtained with pigeons raised in our old lofts at Fox Ridge Farm, in Lincoln, Massachusetts. Astonishingly, Cornell pigeons raised in Ithaca proved to be better oriented at magnetic anomalies than at control, magnetically normal sites!
To determine whether this result was caused by some difference between Ithaca, New York and Boston, Massachusetts pigeons, we established a new loft of pigeons at Codman Farms in Lincoln, Massachusetts our original site being no longer available. This loft was stocked with both Boston and Ithaca birds and when they were tested at magnetic anomalies both groups of pigeons flew straight home. Magnetic anomalies had no effect. We then put a second loft at Fox Ridge Farm a hundred meters or so from our old lofts. It and the Codman Farm loft were stocked with sibling pigeons and then the birds were trained together. When both groups were tested at the anomaly at Iron Mine Hill, Codman Farm birds were well oriented but Fox Ridge Farm pigeons vanished randomly (Fig. 8). This was a particularly dramatic result because the two groups of birds were siblings, they had been trained together and the only difference between the two groups was the separation of their lofts by about 2.5 km. Why should there be any difference at all? Looking at the magnetic map of Lincoln, Massachusetts (Fig. 9) showed that the Fox Ridge Farm loft was located on a steep magnetic gradient. There was a variation of 450 nT over a distance of 1 km. The Codman Farm loft, on the other hand, was located in a relative magnetic flatland with a gradient of 80 nT/km. Could it be that pigeons growing up on the steep magnetic hillside at Fox Ridge farm found that the magnetic field provided useful information whereas pigeons at Codman Farm relied on other cues (Walcott, 1992)? Releasing pigeons at a series of magnetic anomalies did not improve their orientation at the Iron Mine anomaly. Yet after returning from Iron Mine, pigeons released there a second time were well oriented (Lednor and Walcott, 1988).
If Fox Ridge Farm pigeons are using magnetic information as part of their map, it is curious that they were not bothered by head coils which generated continuous magnetic noise (Lednor and Walcott, 1983). Nor did magnets have any effect on the disorientation of these birds released at the magnetic anomaly at Iron Mine (Walcott, 1980b).
That pigeons are disoriented at magnetic anomalies even under sun suggests that the earth's magnetic field might play a role in the "map." While Wallraff (1983, 1996, 1999) has argued cogently against this idea, the anomaly experiments certainly suggest some use of magnetic fields by at least some pigeons.
The picture that emerges from all these observations suggests that pigeons might be opportunists. They may have the capacity to use a number of different orientation cues (See Wiltschko and Wiltschko, 1994, 1995). Which ones they use might well be a consequence of the characteristics and environment of their home loft as well as individual characteristics of the particular pigeon. If we add to that, the variation in cues available from one release site to another we end up with a complex and variable set of possibilities. This makes a great deal of sense if we consider the evolutionary origin of homing pigeons. Their ancestors nested on cliffs and foraged for seeds in the fields. They had to return home with food for the young. Modern racing pigeons have been bred for this ability to return home since 4000 bc. Eight thousand years of strong selection both for pigeons that return (those that fail, don't breed!) and for doing it speedily might be expected to result in a robust system of homing involving many different cues and the flexibility to choose those that are most useful in any given environment. Given the importance of returning to familiar territory to breed and winter, it isn't surprising that migratory birds also appear to rely on multiple cues.
Answers to all your Biology Questions
