Adapted from the Dissertation for BSc Zoology, a novel research project, produced only with the observational aid and expertise of Dr. David Crosse, Dr. Nicole Goodey, Prof. Michael Cant, Daniel Blumgart, and Ignatius W. van Rooyen.
Abstract
Social group cohesion might occur when individual organisms can gain genetic fitness only by supporting group survival and productivity. While commonly observed, the function of allogrooming in social insects is not well understood. It has been suggested to benefit nestmate recognition, essential to group cohesion. Haplodiploid Hymenoptera have been favoured in cooperation and nestmate recognition studies but using diploid eusocial organisms can bring further insight into cooperation behaviour, as mechanism of social evolution in these species may differ from those in Hymenoptera. Here we aim to examine the role of allogrooming in maintaining group cohesion. By observing isolated individuals from colonies of the termite Pterotermes occidentis, upon their return to the colony, we found a significant yet temporary increase in allogrooming in response to isolation. We hypothesise that allogrooming might facilitate nestmate Cuticular Hydrocarbon exchange and thus may play an important role in social cohesion, a vital component of cooperation in social animals.
Introduction
The distribution of resources and competitors in an ecosystem determines animal behaviour in a community, and survival may depend heavily on the use of limited resources that need to be defended from rivals. Social insects often form colonies made up of related individuals that work together to use limited resources for the maintenance and growth of their colony, and when there is more than one colony in close proximity, inter-colony competition for the same limited resources can exist.
Being able to choose between cooperative or antagonistic behaviour when meeting with a conspecific is important for the group cohesion of social insects, and so a means of nestmate recognition is required. The recognition of, not only nest-mates but also instar and reproductive status within a colony, is also crucial for the structure, efficient function and security of an individual member as well as the whole colony: it prevents intra- and inter-specific parasitism, and the theft of valuable colony resources.
Social insects are thought to use chemical recognition cues to distinguish between nestmates, and non-nestmates competing for the same resources. Cuticular Hydrocarbons (CHCs), found on the exoskeleton of insects, have been suggested to play a particularly important role in chemical recognition cues. Thus, changes in relative CHC abundance between individuals may affect the behaviour towards each other.
Previous studies have shown wide variation in CHC profiles between insect individuals from different colonies and that social grooming (allogrooming) within the social colony of many different species is frequent. It is thought that on of the reasons allogrooming may be necessary is to maintain a common colony profile via the constant exchange of naturally individual and variable CHC profiles.
For this study we used a species of Arizonan drywood termite, Pterotermes occidentis, the workers of which constantly groom each other under natural conditions. We hypothesised that isolation of an individual would lead to an individual’s relative CHC profile differing from that of its colony upon reunion. We predicted that after an isolated termite is reunited with its colony, members would detect the difference in CHC profile, leading to the individual experiencing a higher than average frequency of grooming for a short time before grooming returns to average levels. Aggressive behaviour was not predicted in this species as aggression was very seldom observed during preliminary study.
Aims
To examine the role of allogrooming for the maintenance of group cohesion in a species of novel drywood termite, Pterotermes occidentis (Dictyoptera: Kalotermidiae) by testing, via experimental manipulation, whether isolation has an effect on the overall rate of allogrooming observed between isolated individuals and their colony, upon return.
Hypotheses
a. Allogrooming rate will rise from a baseline rate when an individual is reintroduced into its colony after a short isolation period, compared to a non-isolated individual which will maintain a constant allogrooming rate throughout the observation period.
b. There will be a larger increase in Experienced allogrooming than Expressed allogrooming for isolated termites.
c. Allogrooming rate will return to the baseline rate, after a time period of reintroduction of an individual to the colony.
We proposed that such changes in grooming frequency within colonies may be due to natural individual changes in cuticular hydrocarbon profiles of all colony members, leading to a need for continuous CHC transfer to maintain a homogenous intra-colonial chemical profile (Boulay et al. 2000; Nielsen et al., 1999).
Literature Review
Cooperation and Altruism
In nature there is an expected tendency for selection to take place on self-interested behaviour and so when it is found that a species has evolved cooperative and altruistic behaviour despite the individuals costs involved, as is the case with many species across animal taxa, Evolutionary Biologists take a close interest.
A basic situation where supporting group survival and productivity is the only way for an individual to gain genetic fitness encourages the occurrence of social group cohesion (West, 2006). Such situations have been widely studied (Garcia and DeMonte, 2013; Abbot et al., 2011; Korb, 2009; van Veelen, 2009; Dugatkin, 1997).
They are thought to develop when the success of one animal is bound to that of another within a group (often due to matched altruistic behaviour between individuals that share genetic kinship) which increases inclusive fitness and group success while individual direct fitness is reduced. It is suggested that direct fitness interests can be overcome and cooperation achieved if matched behaviour leads to higher indirect and group fitness than can be achieved through selfish behaviour (Frank, 2009).
Altruism has been shown to exist in many different animal social groups (Biedermann and Taborsky, 2011; Duffy, 2002; Faulkes and Bennett, 2001) For example, the study of microorganism social evolution by Kreft (2004) showed that the economical use of common resources increased bacteria group fitness but could only do so at the cost of a decrease in individual growth rate and thus direct individual fitness.
In more complex organisms, recognition based interactions take place between group members and suppression of internal competition within a group influences the occurrence of social group cohesion (Frank, 2009; Tekleab et al., 2009). A study by Bennett et al. (1996) on Cryptomys camarensis, the eusocial Damaraland mole-rat, in which reproduction in a colony is restricted to a breeding pair, showed that not only is there a strong natural aversion to inbreeding, but also repression of ovulation in non-reproductive females by the presence of a reproductive female.
Repression of competition or conflict management behaviour by dominant individuals is essential for group cohesion (Flack et al., 2005). Conflict management often comes in the form of social grooming (allogrooming) of relatives. Madden and Clutton-Brock (2009) demonstrated that in meerkat (Suricata suricatta) societies in which the survival of pups depend on effective group hierarchy management, the frequency of a secondary conflict management strategy is increased when the primary strategy of stress mediation, allogrooming in response to non-resource-focused antagonism by dominant individuals, is reduced. Replacing one strategy with another has a balancing effect, preventing social disruption.
Allogrooming and Recognition Cues
Allogrooming is a grooming behaviour that takes place between two individuals from the same social group, having both a receiving and providing component exchanged between individuals.
The roles of social grooming are researched in primates at length (Grueter et al., 2013; Lehmann et al., 2006) with less extensive study existing for other mammals such as meerkats (Kutsukake & Clutton-Brock, 2010), badgers (Hewitt et al., 2009), impala (Connor, 1995), cows (Val-Laillet et al., 2009; Sato et al., 1993) and other species such as birds (Lewis et al., 2007; Radford and Du Plessis, 2006).
Even though it is one of the most commonly observed interactions in social insect colonies, the functions of insect allogrooming remain less well understood. Lenoir et al. (2001a) and others (Crozier et al., 2010;Blomquist and Howard, 2003; Boulay et al. 2000) suggest that allogrooming may benefit nestmate recognition.It is known that insect recognition is chemically based (van Wilgenbug et al., 2012; Smith and Bree, 1995) and variable over time (Lahav et al., 2001a) with cuticular hydrocarbons (CHCs) playing a particulary central role (Chaline et al., 2005; Akino et al., 2005; Ruther et al., 2002). Allogrooming is thus proposed by Lenoir et al. (2001a) as not only serving an important hygienic purpose in insect colonies (Fefferman et al., 2007), but also as a mechanism for maintaining a homogenous yet distinct colony odour by facilitating the exchange of CHCs between individuals in a colony.
However, nestmate recognition mechanisms and their functions in group cooperation require further investigation as conflicting theories exist. Kirchner and Minkley (2002) argue that nestmate recognition is mediated through chemicals released by gut symbionts, while alternate explanations emphasising the effects of complex interactions between behavioural and environmental stimuli with genes (Bos et al., 2011; Adams, 1991; Blaustein, 1983).
Being highly eusocial insects, Hymenoptera have been particularly favoured in the study of kin selection driven cooperation, altruism and nestmate recognition. Due to their haplodiploidy, female hymenopteran workers share 75% relatedness with full sisters compared to the 50% relatedness shared by diploid siblings, and thus many Hymenoptera social groups present unusually high genetic kinship.
Using cooperative Hymenopteran society models, it has been shown that, unlike groups of the most simplistic unicellular organisms (Kreft, 2004; Velicer and Yu, 2003) recognition-based interaction exists between individuals of complex societies, allowing potentially altruistic acts to be directed at specific recipients (Wlodarczyk 2012; van Zwede and d’Ettorre, 2010; Fletcher, 2008). As a result, individually costly behaviour patterns, such as those that transfer recognition cues between individuals are expected to increase overall group productivity and enhance inclusive fitness levels.
The Lower Termite, Pterotermes occidentis
Hymenoptera have been favoured in the study of kin selection, cooperation, altruism and nestmate recognition. They are often highly eusocial insects and due to their haplodiploidy, female hymenopteran workers are haploid and share 75% relatedness with full sisters compared to the 50% relatedness shared by diploid siblings. Thus, Hymenoptera social groups tend to present unusually high genetic kinship.
However, the use of homoploid social organisms in attempts to demonstrate a general underlying mechanism of cooperation is important, offering more insight into the evolution of cooperation. If the evolution of cooperative behaviour and social structure depends on group genetic relatedness (Gadagkar et al., 1991), principles that are true for haplodiploid hymenoptera may not necessarily apply to homoploid eusocial species such as the African and Damaraland mole-rats (Burland et al., 2002; Lovegrove, 1991) or termites (Korb, 2008; Thorne, 1997; Bartz, 1979) that all produce only diploid offspring. Like in Hymenoptera species, eusociality is exhibited in termites despite their diploidy, with high levels of cooperation, brood care, a worker caste and repressed reproduction (Thorne, 1997).
Pollock (1996) argues that cooperation under kin selection theory in Hymenoptera can be reduced in the presence of high social competition associated conflict leading to kin avoidance and preferred interaction with non-relatives, when a stable matched behaviour strategy exists between non-relatives. Lower termites are then especially suitable as alternative model subject species of broad concepts like recognition aids for cooperative group organisation since they lack substantial intra-colonial aggression (Olugbemi, 2012; Korb, 2005) which may otherwise be a cause of potentially confounding effects in experimental studies of nest-mate behaviour.
This project focused on the wingless, immature individuals of the colony: pseudergates (some times referred to as the ‘worker caste’).
Selection and Marking Methods
Only the immature worker caste (pseudergates) was used. From each colony 16 individuals were randomly selected, weighed and uniquely marked using different colour combinations of enamel point dots on the head, thorax and abdomen.
Mark retention was limited and variable due peeling quality of the paint and higher grooming rates experienced by all marked individuals. An optimum period of 2 days existed between marking and first observation: allowing paint to dry and termite recovery after a large amount of handling, yet not losing the mark before the final observation. Marking each individual with multiple dots also reduced the need to re-mark.
At marking, termite restraint was essential in preventing the smearing of paint across important delicate body parts, potentially adversely affecting behaviour and mortality. However, P. occidentis are soft bodied and difficult to handle with forceps for long periods of time, and each focal individual was exposed to a controlled amount of CO2 (10-20%) gas, previously shown as suitable insect anesthetic (Nilson et al., 2006) for 30 seconds, allowing for the application of small paint dots on the head, thorax and abdomen using the round head of a needle dipped in the enamel paint.
8 pairs were formed from the 16 individuals selected per colony by matching termites as closely as possible by weight, and each individual was allocated either to the Test or Control treatment group. Individuals from the Test treatment would be isolated and reintegrated with the colony, whereas those from the Control treatment would undergo no isolation but would otherwise experience exactly the same procedure as Test individuals in every other respect in order to account for effects such as handling stress, and marking on termite behaviour as well as creating a point of reference to which Test termite behaviour could be compared.
This having been done for each colony, a total of 32 pairs of focal individuals resulted. The unique marking, weight, colony number, pair number and treatment group allocation of each individual was recorded for future reference, and all focal individuals returned to their colonies after selection and marking.
Demonstration of the colours used for marking under natural light (used to identify termites outside of observation conditions) and red light under which termite behaviour was observed (used to identify focal individuals under observation).
Two marked focal individuals among unmarked nest-mates.
Data Collection Methods
Allogrooming Measure
Due to its frequency and distinction from other behaviours, allogrooming in this species was easily observable and accurately timed. Observations were done in a demountable observation room kept at 26±5°C, under red light with minimum sound and movement vibrations. These conditions were controlled to resemble the natural climate in which P. occidentis is normally found.
Observation intervals lasted 45 minutes, for each individual the trial period was 7 days long and observations took place on Day 0, Day 4 and Day 7, which involved the collection of pre-isolation, post-isolation and post-reintegration grooming rates respectively for Test individuals. Observations for Control individuals took place at the same times.
Test and Control individuals from the same pair were always observed on the same day, though observation times of day were not matched.
Control individuals were returned to their colony after every observation whereas Test individuals were isolated after observation on Day 0, and only returned to the colony on Day 4 and Day 7.Isolation was achieved by placing the individual inside a clear container (5mm x 3mm x 3mm) with a previously unused piece of wood and covered with gauze to allow ventilation but prevent escape. The isolation containers were maintained in the same incubator as the main colonies.
Due to the relatively large area of wood within main colonies and large total number of termites in all but one of the four colonies used, accurate observation of particular individuals within a main colony was impossible under time constraint and observer limitations. As a result observation arenas were made from transparent Perspex (110mm x 110mm x 50mm) and random colony samples were used instead. Each arena held a single 8mm thick slice of wood, particular to the observed colony, elevated on a transparent stand at approximately a 45° angle to a mirrored surface. This reduced the chances for focal individuals to be out of sight, yet allowed termites to move freely.
Before an observation session, a random colony sample of 50 individuals from the three larger colonies would be placed in their observation arena and positioned in the preheated observation demountable to acclimatise for 45 minutes before observations began. The smallest colony contained less than 50 individuals in total and was used in its entirety.
While allowing the colony sample to acclimatise, the focal individuals were identified and placed in a gridded tray, labelled with the identification numbers and mark-colour sequences. These were allowed a 30 minute recovery time in the observation room after handling.
When the first focal individual was placed in an observation arena with the colony sample, a stop-watch was started and during the next 45 minutes, the start and end times of any allogrooming event was recorded, along with whether it was expressed or experienced with respect to the focal individual. Allogrooming consists of two components that interact with each other, the performer and the recipient, at the same time. Since grooming can also be expressed and experienced at the same time by a single individual, it was important to make a distinction between these two types of allogrooming in recorded data.
At times when multiple grooming events were taking place at once, the start and end time of each was recorded, even if they overlapped. In the event that a focal individual was out of sight, the duration was recorded and later deducted from the total observation time during data analysis.
Once the observation of a given individual was complete, the individual was not removed from the arena as this would cause disturbance to the entire colony sample. Instead a sample increase was noted and the next observation was continued. When the observations for a particular colony sample was complete, Test and Control focal individuals were either returned to the colony (on Days 0 and 7) or Test termites were placed in separate prepared isolation pots while Control individuals were returned to the colony (on Day 4).
Colonies are kept in incubators. Termitaria (custom designed by Wilmie van Rooyen and produced by Ignatius van Rooyen) are used to house colonies, preventing termite injury and death from being trapped under wooden slices when termitaria are handled. These termitaria also allow quick access to termites and visibility from all sides, while preventing termite escape.
Left: Observation area (custom designed by Wilmie van Rooyen and produced by Ignatius van Rooyen) with wood elevated on serrated edges so that termites could move freely without being trapped between wood and perspex. The arena was placed over a mirror so that all sides of the wood could be observed at once. Right: Observation took place under red light in a temperature an noise controlled room.
Solid Phase Microextraction
In addition to behavioural observation, Solid-Phase Microextraction (SPME) was used to collect cuticular hydrocarbon samples from focal termites to be used in a separate but parallel future investigation of changes in individual cuticular hydrocarbon profiles in response to isolation (Singer, 1998). For this, all focal termites were swabbed with a thin polymer-coated fused fiber which collected cuticular hydrocarbon samples from the individuals.being alive nor dead could be ruled out).
The results of this study are not presented or discussed here, however attaining the cuticular hydrocarbon sample required that focal individuals were mechanically restrained for a short amount of time before each observation, the handling stress of which possibly affected the general behaviour that was then observed and the possible effects on the results of this study will be discussed.
Mortality Survey
Data Analysis Methods
Allogrooming Analysis
Out of the 32 termites pairs which were selected and marked for observations, 25 pairs of complete allogrooming data sets were obtained due to the death of one or both individuals in 7 of the pairs during the course of the experiment. All statistical tests were performed using the statistical analysis software package, R.
A series of 11 different statistical tests on the same two sets of data were performed at a 95% confidence level. This meant that there was a 43% chance of Type 1 error (1 – 0.9511) occurring.
Using the Start and End times of each allogrooming bout, the duration of allogrooming bouts were calculated (End Time – Start Time=Grooming Duration). Grooming Durations for each individual on each observation day were then used to calculate the proportion of time the focal termite experienced grooming (Sum of Experienced Grooming Durations/Total Observation
Duration=Proportion) and the proportion of time the focal termite expressed grooming (Sum of Expressed Grooming Duration/Total Observation Duration=Proportion). The proportion of time spend allogrooming was representative of the allogrooming rate. First the collected data was illustrated in line-graph form, comparing the mean proportion of time spent experiencing and expressing grooming on each observation day (0, 4, and 7) for Test and Control treatment groups. It was determined that Expressed Grooming data was not normally distributed while Experienced Grooming data was normal. This was showed in the form of histograms and doing an analysis of variance on all the data using the Shapiro-Wilk normality test.
Independent paired difference tests were then used to determine whether there is a significant difference in means between Test and Control treatments, on Day 0, 4 and 7, for Experienced Grooming time proportions (Paired t-tests) and Expressed time proportions (Wilcoxon Signed rank tests).
By plotting means against variances for each observation day, it was found that the variance increased with mean. Consequently, Kruskal-Wallis rank sum tests were used to demonstrate whether there is significant difference in means of observation days within Test and Control treatments, for Experienced Grooming proportions and Expressed Grooming proportions. Some differences were shown to exist and consequently, paired difference tests (Paired t-tests for Experienced Grooming data and Wilcoxon Signed rank tests for Expressed Grooming data), were used to determine where these differences in means lay and what their magnitudes are: for each treatment group we did the appropriate paired difference tests with Day 0 versus Day 4, Day 4 versus Day 7, and Day 0 versus Day 7.
Mortality Survey Analysis
Post-experimental mortality data was analysed using the Chi-squared test to determine whether treatment group had an effect on mortality.
Results
Allogrooming
The data for the proportion of observation time that focal individuals were groomed by colony members (Experienced Grooming) and groomed colony members (Expressed Grooming) show that Test and Control individuals spent a greater proportion of time being groomed than grooming overall (Fig.1)
Referring to Table 1, the difference for mean Experienced Grooming proportions between Test and Control treatments was found not to be significant on Day 0 (P=0.32), nor on Day 7 (P=0.86), but was significant on Day 4 (P=0.04). For the mean Experessed Grooming proportions, once again no significant difference was determined between Test and Control treatments on Day 0 (P=0.67) or Day 7 (P=0.48) but on Day 4 the difference was again significant (P=0.031) as shown in Table 1.
A significant difference existed between observation days within both Test and Control treatment groups (P=0.0022 and P=0.046 respectively) for Experienced Grooming proportions (Table 2).
For the Test group, mean Experienced Grooming proportions were not significantly different between Day 0 and Day 7, but significant differences lay between Day 4 and Day 0, and between Day 4 and Day 7. The same pattern was found for the Control treatment group but at a lower significance (Table 2).
For Expressed Grooming proportions, a significant difference was found between days within the Test group (P=0.0025) but not within the Control group (P=0.093), as shown in Table 2. For the Test group, it was indicated that that there was a significant difference between Day 0 and Day 4 and between Day 4 and Day 7, but not between Day 0 and Day 7 (Table 2.).
Although, no significant difference in grooming proportions between observation days was found for the Control group using the Kruskal-Wallis rank sum test, Fig. 1 does illustrate some elevation in mean Expressed Grooming on Day 4. This was investigated and no significance existed in the differences between Day 0 and Day 4, nor between Day 7 and Day 0, but between Day 4 and Day 7, the difference was of some significance (P=0.024) as shown in Table 2.
To sum up all the data, the above described relationships are illustrated in Fig. 1. Larger peaks in mean Experienced Grooming and mean Expressed Grooming time proportions on Day 4, as well as higher levels of significance in the differences between Day 0 and Day 4 (Table 2), are shown for the Test treatment group than for the Control treatment group; and on Day 7 declines in mean grooming proportions resembled those found on Day 0.
Mortality
Finally, Fig. 2 illustrating post-experimental termite mortality data from all 64 initially marked focal individuals indicates that there may be a relationship between the treatment group which individuals belonged to and their mortality.
It might be concluded from the bar graph that more Test individuals were found dead, and fewer were found alive compared to Control individuals. Also, that more Control individuals remained unfound and so potentially alive, which wasn’t the case for Test individuals. Equal amounts of Test and Control individuals were found indistinguishable from other marked individuals within the colony.
Discussion
In this experiment we tested whether there would be a change in allogrooming rate experienced and expressed by individuals of P. occidentis individuals upon reintroduction to the colony after a period of isolation. Our aim was to assess the role of allogrooming in the maintenance of group cohesion for a lower termite species. We found a significant increase in mean allogrooming rate in response to isolation from the colony for both components of allogrooming.
As we predicted, our results show no significant difference on Day 0 and a significant allogrooming increase after 4 days of isolation. We also observed no aggressive reactions from the colony towards returning individuals. This is feasibly indicative of a change in the isolated individual, detected by its colony upon return, triggering a behavioural response.
We suggest that during isolation, no opportunity existed for the homogenisation of CHC profiles between the colony and the isolated individual. A slight change in the individual’s profile distinct from that of the colony may have occurred as have been shown possible in various ant species (Nielsen et al., 1999; Liu et al. 2001; Vander Meer et al., 1989), and could have been detected by the colony. Longer, more frequent grooming bouts between the focal individual and colony members could be an attempt to re-homogenise the colony’s chemical profile.
These results support the findings of Lenoir et al. (2001a) and Boulary et al. (2004), showing that colony members of two ant species, Aphaenogaster senilis and Camponotus fellah, react sociably towards returning individuals after a period of short isolation (approximately 3-10 days) and that increased social behaviour was likely to promote the re-homogenisation of colony odour.
However, both studies noted aggressive behaviour of colony members towards returning nestmates after a longer period of isolation (approximately 10-40 days). Aggression has been shown as part of kin recognition systems (Ichinose and Lenoir, 2009;Suarez et al., 2002) and a lack thereof in some social insect species, including P. occidentis, can be puzzling since strong intraspecific resource competition often exists in eusocial insects due to valuable food sources (Shellman-Reeve, 1997). Effective kin recognition could allow distinction between nestmates and non-nestmates (Sobotnik et al., 2008; Lenoir et al 2001b) and assessment of the most appropriate interaction to display such as social rejection (Kaib et al., 2002) or acceptance (Holway et al., 1998)
We used a short isolation period and valuable future study lies in lengthening and varying this isolation period, to explore the reactions towards nestmates in P. occidentis after a longer time away from the colony. These results can then be analysed in the light of genetic relatedness. For example, Kidokoro-Kobayashi et al. (2012), suggests that unicoloniality might explain the lack of aggression seen in some social insect species, while Fuller et al (2004) showed that, since both genetic and environmental factors change with geographic location, aggression towards non-nestmates increases with geographic distance from a nest. It is explained that colonies closer to each other are more likely to share genetic similarities due to colony separations earlier in time. This makes the location of a colony’s origin also an important aspect of future study, as this was not included as a factor in our work.
Allogrooming has hygienic purposes (Fefferman et al., 2007) the effect of which was also not considered during our experiments and holds potential for future investigation discussed further on. In dampwood termite colonies which have high pathogen loads, allogrooming has an important role in the removal of pathogens and parasites (Rosengaus et al., 1998), and has been reported infrequent in species of drywood colonies which have naturally very low pathogen loads (Rosengaus et al., 2003). However, allogrooming is still frequently observed in P. occidentis colonies possibly supporting the hypothesis that it is needed to facilitates the exchange of CHCs between individuals of eusocial insect colonies (Crozier et al., 2010; Lenoir et al., 2001a; Boulay et al. 2000) and thus, its role in group cohesion, possibly via kin recognition, can be essential for cooperation in social animals (Garcia and De Monte, 2013).
In addition, we show fluctuation in allogrooming rate for the non-isolated control individuals, which was unexpected. On Day 4 the increase in grooming rate of non-isolated individuals was significant for both components of allogrooming. These were of lower significance than the increase seen for isolated individuals, and the difference between the increased rates of isolated and non-isolated groups on Day 4 was significant.
A possible explanation may be that during a given observation, a non-isolated individual may have been inadvertently affected by a possible continuous response of the colony sample to the presence of an isolated individual during a prior observation. Lenoir et al. (1999) discussed a chemical integration period which exists in newer members of ant colonies and we postulate a similar occurrence in termites.
A way in which such an effect could be minimised is by renewing the colony sample before each observation, allowing time between observations for the colony to recover from the response to the isolated individual.
For both isolated and non-isolated groups, the increase in mean experienced allogrooming rate was more significant than the mean increase in expressed allogrooming rate. However, the variance of allogrooming rate across Day 0, 4 and 7, within the isolated group was more significant than within the control group.
This was predicted as we expected the colony members to focus on isolated individuals when returned. Since we recorded all observations in respect to individuals, and each termite can only groom one other individual while being groomed by many, a higher increase for experienced allogrooming can be recorded. As stated above, fluctuation in grooming rate for non-isolated individuals we found was not predicted but this concept may apply to the similar pattern in allogrooming found for the non-isolated group.
After a period of isolated termite reintegration with their colonies, the allogrooming rates of all individuals had lowered significantly by Day 7, and closely resembled the basal rate recorded on Day 0.The difference between rates recorded before isolation on Day 0 and after the reintegration period on Day 7 was not significant for either component of allogrooming, in either treatment group.
As already mentioned, we explain these results by proposing a mechanism involving CHCs. For example, if the presence of an anomalous CHC profile might trigger increased allogrooming, homogeneity may be achieved in the colony-specific CHC profile again, after some time. This could then mean that the cue for increased allogrooming also dissipated and allogrooming returns to a normal rate as a result (Singer, 1998; Vander Meer et al., 1989).
We discovered that both isolated and non-isolated individuals experienced overall more allogrooming than they expressed, even at the basal rate collected on Day 0. This finding was surprising because if our sample of focal individuals accurately represented the colony populations they were sampled from, it would not be rational for every termite to received more grooming on average, than it performs.
A possible reason for a lack in expressed allogrooming could be due to handling stress experienced during Solid-Phase Microextraction hydrocarbon sampling since this was a physically disruptive process for this soft-bodied species.
Though quantifiable data was not collected, there seemed to be variation in the stress experienced through SPME sampling as it is not a precise process, whereas this variation was not obviously reflected in the difference in overall expressed and experienced allogrooming.
On the other hand, all focal individuals were marked and during observation, un-marked (non-focal) colony members were noted as particularly attentive to the paint markings when grooming. This causes us to put forth a more plausible explanation, similar to one mentioned earlier: all focal individuals received more frequent grooming than they were able to perform at a given time, even before isolations took place, and we speculate that this may be a result of the hygienic function served by allogrooming.
In light of the fact that some drywood termite species exhibit very low levels of allogrooming, and their colonies are relatively pathogen free, potential fascinating questions are raised such as the reasons some drywood termites retain allogrooming while others do not and the practical degree to which group cohesion is affected by allogrooming. These hold the potential for manipulation experiments using markings on individuals within a colony and testing a response.
However, since there is also little research on the effects of enamel paint (or other marking mediums) and CO2 as an anaesthetic on termites, it may well be that these also had an effect on their behaviour. In the past, CO2 been shown effective and safe for invertebrate use (Cooper, 2011), possibly even blocking the effects of handling stress (Pankiw and Page, 2003) however some studies point out possible disadvantages Dombrowski et al. 2013; Colinet and Renault, 2012).
A study by Miramontes and DeSouza (1996) has demonstrated increased mortality rate due to social isolation the termite, Nasutitermes cf. aquiline. The results of our mortality survey do not support this for short isolations of this species as it showed no significant relationship between treatment and the mortality which occurred up to 40 days after the final observation. However, a number of focal individuals had lost their marks by the time the mortality survey was taken and so were indistinguishable from other termites and surveying at shorter time intervals would present more useful results.
Colony polymorphism, caste of focal individual and colony size may all affect social behaviour (Nandi et al., 2013; Aviles et al., 2004; Crosland et al., 1997; Oster and Wilson, 1978) but were not controlled for in this study. Since focal individuals and colony samples were randomly selected from each colony, and the colony sizes ranged from 30-170 individuals, the representation of each colony’s behaviour patter n was reasonable. Future more advanced statistical analyses may be done using methods which account for mixed effects such as caste and colony size.
The importance of group cohesion in survival and productivity of cooperative insect groups has been thoroughly described (Hamilton et al., 2009) and a recent study by Lihoreau and Rivault (2009) showed the value of kin recognition to the cohesion of a social life in the cockroach, Blattella germanica. The kin recognition properties of CHCs have been shown to exist even in sub-social species such as Stegodyphus lineatus, a sub-social spider (Grinsted et al., 2011).
Other than pathogen and parasite control, allogrooming may serve an adaptive role as a mechanism of group-specific CHC exchange between individuals within a colony, which would enable nestmate discrimination and protection of resources. There is also evidence that the CHCs are influenced not only by the environment (Wagner et al., 2001; Silverman and Liang, 2001) but also genetics (van Zweden et al., 2010; Dronnet et al., 2006; Page et al., 1991). If kin selection plays a role in whether or not one individual displays allogrooming behaviour, and as a result influences the cohesion of a group according to a genetic relationship, there may be inclusive fitness benefits for allogrooming.
Finally, having obtained good indication that allogrooming can fluctuate in accordance with whether or not an individual has had recent contact with its nestmates, and given that allogrooming can be widely recognised in many social species, we allogrooming as having a potentially recognition based function, essential for the effective cohesion of a social group. Through the continued interest in group behaviour and the study of its common underlying principles, we can come to gain better understanding around the evolution of cooperation under the laws of natural selection.
In Theory 