Study organisms

Professor Andrew Radford: Full research details.

Current Research

1. Group decision-making

Many of the advantages of group living, including lowered predation rates and enhanced foraging success, can only be fully realised when group members remain in close contact. Synchronisation of activity is therefore important for group cohesion if, for example, group members are to follow a common travel direction or leave a particular site together. Although groups of animals often need to coordinate their actions, very little is known about the underlying mechanisms by which the relevant decisions are made. A recent theoretical model hypothesised that a group could reach a decision either despotically or democratically. Although several observational studies have indicated that behavioural and vocal voting might be involved in the group decision-making of non-human vertebrates, none have tested this experimentally.

I am therefore investigating the decision-making process when groups of pied babblers move to a new foraging site. Pied babblers are excellent subjects for this research because they live in permanent groups of 3-15 individuals, they spend much of their time foraging on the ground as a close-knit unit, and they can be habituated to the close presence of observers. It is therefore possible to make behavioural observations and sound recordings from only a few feet away, to quantify individual foraging success directly (by recording the capture rate of identifiable prey items) and to conduct a variety of experimental manipulations. The open habitat also makes video filming easy, allowing a precision of analysis unobtainable through observation alone. Furthermore, babblers can be trained to jump onto a scale, allowing weights to be gathered repeatedly during the course of the day (providing accurate and regular measures of foraging success and condition).

My observational data suggest that babblers may use a democratic vocal voting system when deciding to move to a new foraging site, with an individual's calling dependent on its current foraging success. Subordinates 'vote' more frequently, but it is dominants which tend to lead the group to a new area. I am currently planning supplementary feeding and call-playback experiments to distinguish clearly between democratic and despotic possibilities.

2. Alarm-call use and development

A variety of mammalian studies have shown that alarm calls not only warn others of the presence of predators, but that they can be highly specific with respect to the type of predator, the degree of threat, or both. There has, however, been a paucity of such studies on birds. My observations and sound recordings suggest that pied babbler alarm calls convey information about the nature of the threat; different calls are produced in different situations. I am using playback experiments to compare the importance of threat urgency and predator type in pied babblers, and model presentations to confirm the observational finding that 'mobbing' calls are given to some predators (e.g. cape cobras (Naja nivea) and yellow mongooses (Cynictis penicillata)), while 'flight' calls are given to others (e.g. martial eagles (Polemaetus bellicosus)).

Future work in this area will target contextual vocal learning - how young birds learn to give alarm calls in the appropriate context and to respond appropriately to the alarm calls of others. I will be considering the developing responses of young babblers to the alarm calls of both their own species and those given by crimson-breasted shrikes (Laniarius atrococcineus) and fork-tailed drongos (Dicrurus adsimilis), two species which commonly associate with babbler groups. Preliminary evidence shows that babblers do respond to the alarm calls of these species, but they must learn that shrike calls tend to be reliable, while those of drongos are less so (drongos use false calls to kleptoparasitise the babblers when they are foraging).

3. Function of foraging vocalisations

Individuals of many group-living species produce and exchange frequent vocalisations when foraging, but their exact function is often difficult to determine. I am investigating the function of two such calls given by pied babblers. The first is a 'recruitment' call, used to announce the presence of particularly rich and divisible food sources (e.g. a termite nest). Assessment of individual foraging success in different patches and the relationship with calling suggest that a threshold of food availability is required before the call is given. This threshold varies depending on the individual and the circumstances (e.g. dry versus wet seasons). Similarly, individuals differ in their likelihood of responding depending on individual success and condition, as well as current food availability.

Using food-supplementation and playback experiments, as well as observational data, I have shown that the call is most prevalent when there are independent fledglings within the group (i.e. young individuals that are no longer being directly fed, but which are still poor foragers). The recruitment call may therefore act as a form of extended parental care.

The second 'close' call is given throughout foraging bouts by all individuals. Theoretically it might be used to maintain spacing between group members or to prevent scattering of individuals. Observations and call playbacks suggest that this particular call is best described as a 'contact' call, being used to maintain cohesiveness of the group as it forages; the pattern of calling allows each individual to orientate itself in relation to the centre of the group. Individuals are more likely to give the call when they can hear the calling of others, particularly when their sight is obscured by vegetation. Calling rates decrease when an individual becomes somewhat separated from the group.

4. Allopreening

Allopreening is a widespread phenomenon among birds but, unlike allogrooming in mammals, virtually no detailed studies have investigated its functions. In green woodhoopoes, a group-living African bird species, allopreening appears to serve a dual purpose: allopreening of the head and neck region (inaccessible to the recipient itself) seems to be primarily for hygienic purposes; in contrast, allopreening of the rest of the body may serve a primarily social function. Allopreening of these accessible body parts peaks at certain times of year, is more apparent in larger groups and dominants are the recipients far more than subordinates, which themselves conduct more of the active preening. This suggests that the allopreening acts to enhance social cohesion within the group, rather than to maintain dominance relationships.

I am currently examining how intragroup allopreening is influenced by intergroup conflict. Although numerous studies have investigated the increase in affiliative behaviour following conflict between group members, virtually none have looked at the importance of conflict between groups, even though these are common in many social species. It appears that intragroup allopreening increases following intergroup conflicts, particularly those that are long in duration or lost, and those involving strange groups as opposed to neighbours. Moreover, not all group members increase their allopreening to the same extent: the postconflict increases are the result of more allopreening of subordinate helpers by the dominant pair. This may be because the dominant pair are trying to encourage the helpers to participate in future intergroup conflicts; they may be trading stress-reducing allopreening for assistance in intergroup conflict.


1.Mediation of foraging conflict within groups

All habitats have a finite level of food resources. As a consequence, foraging conflict may arise within a species for these limited resources. Green woodhoopoe groups consist of a breeding pair and up to 10 non-breeding helpers of both sexes. Group members forage as a close-knit unit for much of the time, creating opportunities for both intersexual and intrasexual conflict over foraging resources. I showed that foraging conflict is minimised both through niche partitioning and by the use of sexually dimorphic vocalisations, providing one of the first demonstrations of the function of a food-related call in a social species.

Individuals preferentially use different foraging techniques, dependent on their bill length, with juvenile males changing their preferred foraging tactics as their bills grow. A specific ‘kek’ call acts as a vocal signal of foraging niche, regulating spacing between foraging competitors (i.e. individuals that forage using the same preferred techniques). Individual foragers increase their calling rate only in response to foraging competitors, which subsequently move away. Maintenance of such spacing is important because individuals suffer a lowered success rate when foraging in close proximity to a competitor.

2. Mediation of territorial conflict between groups

Most studies of asymmetric contests have focused on individuals, so I used the territorial vocal chorusing displays of green woodhoopoes to investigate experimentally the key factors determining the duration and outcome of inter-group interactions. Contests, when groups of 2-12 individuals give cackling choruses alternately, are generally decided within five minutes (‘short’) or take longer than 15 minutes to reach an outcome (‘extended’). Resident groups tend to win short contests, even though there is no obvious difference in the chorusing of the two protagonists, perhaps because intruders are only checking territory occupancy or the possibility of breeding vacancies. During extended contests, resident groups match the length of each chorus given by an intruding group, losing the contest and retreating only when they fail to maintain this matching. On such occasions, intruders remain in the resident’s territory to forage and examine roost holes. It is easier for larger groups to maintain a certain length of chorusing, because each individual makes a smaller personal contribution and thus is likely to suffer a lower metabolic cost. In extended contests, the relative size of the two groups is therefore critical, with the larger group typically winning. During contests, both sexes respond more strongly to same-sex intruders, while subordinate individuals cackle for longer than dominants, presumably because different individuals suffer varying foraging and/or breeding costs from the inclusion of additional group members.

I am continuing to investigate inter-group interactions in the pied babbler, where contests are complicated by the production of four different vocalisations within a chorus and the fact that not all group members participate in all contests.

3. Facultative sex ratio manipulation

Mothers would often benefit from producing more of one sex than the other. Although some species show an astonishing ability to skew their sex ratio adaptively, the trends found in many vertebrate studies have proved inconsistent. I demonstrated the importance of large sample sizes and data from multiple years by showing great variation in both the significance and the directional trend when considering potential factors influencing the sex ratio of great tit (Parus major) broods. Although some factors (e.g. male age and female tarsus length) were significant predictors in some years, there were no consistent trends and no factors were significant when combining data from all five years.

Even though several studies have provided convincing evidence of facultative sex-ratio manipulation, evidence for a mechanism by which such a bias is achieved is equivocal at best. Using 30 years of data, colleagues and I showed that springbok (Antidorcas marsupialis) mothers preferentially produce daughters when they are in good condition (the opposite trend to that found in a number of other polygynous, size-dimorphic ungulates). Furthermore, we were able to examine culled individuals and show that selective re-absorption of embryos is unlikely to be the mechanism by which deviations from an equal sex ratio are achieved. Hence, either differential implantation occurs or females are able to influence the sex of the sperm fertilising an egg.


1. Kuruman River Reserve, Northern Cape Province, South Africa

This is the site of the pied babbler research and is a 3500 hectare area situated 17 km south of the South-Africa Botswana border and 30 km west of the small settlement of Van Zylsrus. There are currently 14 colour-ringed, habituated groups of babblers on the reserve with several more groups targeted for imminent work. The study site consists of sparsely vegetated dune country on either side of the (dry) Kuruman river, which makes observations particularly easy. For further information about work on the babblers see The Pied Babbler Research Project pages.

In addition to the babblers there are a number of habituated meerkat groups on the reserve, which have formed the basis of study for the last 10 years. See the University of Cambridge meerkat pages.

2. Morgan’s Bay, Eastern Cape Province, South Africa

This is where the green woodhoopoe research is conducted. Within a 33 km2 area, there are 28 colour-ringed groups which inhabit the thick riverine forests that form belts along the river valleys. Territories are linearly arranged along these river courses. Canopy height varies between 8 and 15 m, with an average height of 9.6 m. Forest sections are separated by areas of open grassland, much of which is grazed by cattle. The mean annual rainfall is 982 mm, with well-defined peaks in October and March, and the average temperature is 23°C. My work on the woodhoopoes is in collaboration with Prof. Morné du Plessis (formerly of the Percy FitzPatrick Institute of African Ornithology, University of Cape Town, now at the World Wildlife Fund, South Africa.