LB 144 G1: Fluffy's Revised Half-Draft

Revised Half-Draft by: Chloe Gordon

Despite being highly advanced compared to numerous other kinds of animals, many of the traits common in humans can be found and better understood by looking at another species (Lattal, 2001). For this reason, it is of great importance to study and better understand the behaviors exhibited in other animals. The Sciuridae family, more broadly known to laymen as the squirrel family, is one species of animal which has some behavioral characteristics also found in humans. With squirrels being so prevalent and behaviorally affected by man’s urban environments, it is vital to study them in order to create the best possible cohabitable landscape for our two species (McCleery and Parker, 2011). When observing a community of squirrels, it is not uncommon to discover a social dominance hierarchy within the observed group. Dominance has typically been determined by the way in which squirrels react in the presence of other squirrels. For example, a squirrel was considered to be dominant if it showed some form of aggressive behavior (pursuing in a chase, lunging, etc.) towards another squirrel in any setting. If the victim squirrel’s response was to be chased away, then it was assumed that the squirrel which exhibited aggressive behavior was dominant (Pardo et al., 2014). Social dominance among groups of squirrels has been documented numerous times and in multiple subspecies, such as eastern gray squirrels (Sciurus carolinensis), tassel-eared squirrels (Sciurus aberti), and fox squirrels (Sciurus niger) (Allen and Aspey, 1986; Farentinos, 1972). However, it is a squirrel’s ability to use non-auditory gestures to display social dominance which makes them relevant to study and compare to humans. Similar to a human’s use of their hands to convey nonverbal messages, squirrels will often use their tails to communicate messages which signal readiness for copulation, aggression, alarm signals, etc. (McCloskey and Shaw, 1977). Prior research has suggested that these gestural messages are communicated through the tightness and the portion bent in a squirrel’s tail giving credit to the idea that squirrels can communicate through various positionings of their tails (Pardo et al., 2014). For our study, we looked at squirrels found within Michigan State University’s campus. While gray squirrels, fox squirrels, red squirrels (Tamiasciurus hudsonicus), and northern flying squirrels (Glaucomys sabrinus) can all be found in this area, fox squirrels are the most common. Due to this, fox squirrels were the source of data in this study.

We also compared homologous behaviors of humans and squirrels. Similar to squirrels, humans execute a large amount of nonverbal communications in their everyday lives. Often, these non-verbal gestures are in the form of co-speech gestures (hand or arm movements, head nodding, or change in posture) throughout a conversation (Cristilli et al., 2004). We more specifically observed the different types of co-speech gestures and watched how people communicate during conversation in public. There are three main types of gestures that can be observed; adapters, symbolic gestures, and conversational gestures (Krauss et al., 1996). The first type of hand movements (that are not considered gestures) are known as adapters. These are most commonly observed behaviors in speakers, including fidgeting, tapping or rubbing their hands, as well as frequently adjusting their clothing (Krauss et al., 1996). Secondly, there are symbolic gestures, also known as emblems (Willems et al., 2007). Emblems are gestures that have a specific meaning without speech, such as nodding "yes" or giving a "thumbs up"(Willems et al., 2007). Although symbolic gestures are most commonly used without speech, they are sometimes used along with speech (Cristilli et al., 2004). Lastly, there are conversational gestures which fall directly between adapters and symbolic gestures. Conversational gestures are more commonly seen during conversation, but unlike emblems they do not occur in the absence of speech (Krauss et al., 1996). We have observed conversational gestures that took place during conversation. Another important aspect of human communication we studied is eye contact. From all of these human movements, it is important to note that how one displays a certain movement can help to determine how dominant that person appears to be. For example, keeping strong eye contact while having a conversation, fast responses, or "confident" posture all contribute to how competent or more dominant a person may appear to be (Ridgeway, 1987).

Motor control is an important part of being able to make the nonverbal communications observed during the study. There is a complex physiological background of being able to make and understand these nonverbal gestures in both humans and in animals. The forkhead box protein P2 (FOXP2), coded for by the FOXP2 gene, has been shown to be very important for language and neural development in humans and in many animals (Campbell et al., 2008). Homologs of FOXP2 are found in many vertebrates, but the human version of FOXP2 is thought to be what allows humans to articulate words in a way that other animals cannot. Some known aberrant forms of human FOXP2 cause significant speech disabilities caused by an inability to articulate properly (Enard et al., 2002). It is hypothesized that FOXP2 was initially used in ancient vertebrates for neuromotor development and to control coordinated movements and that vertebrates ended up co-opting its mechanisms for communication purposes in many species (Fisher & Marcus, 2006). FOXP2 has been shown to still be very important in neural development in rodents. Mice with a faulty FOXP2 gene were generated; the mice with two copies of the faulty gene were unable to create innate vocalizations, had impaired motor-learning skills, and had smaller cerebella than those with the standard mouse FOXP2 gene (Fujita et al., 2007; Groszer et al., 2008). While we have not been able to find research on FOXP2 in any squirrel models, the gene has been shown to be important in motor and lingual functions in animals and, more specifically, rodents (Fisher & Marcus, 2006). The gene found in rodents would be expected to be found in some similar form in squirrels because muroid rodents are closely related to sciuroid rodents, and FOXP2 is very well conserved throughout vertebrates in general (Fisher & Marcus, 2006). FOXP2 has been shown to be very important in human motor skills and language, as previously stated, and could potentially be very important in squirrel’s ability to communicate nonverbally.

Observational studies of dominance amongst squirrels took place over several weeks. In our study, we looked at nonverbal dominance communication in fox squirrels in order to compare it to nonverbal dominance communication in humans. We will then briefly explore the importance of the FOXP2 gene in relation to nonverbal communication. No manipulations occured throughout this study. East Campus, North Campus, and River Trail were the areas where we conducted the observations for our study. We hypothesize that both fox squirrels and humans transfer dominance information through nonverbal communication. From our hypotheses, we predict that dominant fox squirrels with lower tail portion bent will show higher levels of aggression towards subordinate fox squirrels with higher tail portion bent because dominant eastern gray squirrels showed higher levels of aggressive toward subordinate eastern gray squirrels with large tail portions bent (Pardo et al., 2014). We predict that dominant fox squirrels with higher tail tightness will show higher levels of aggression towards subordinate fox squirrels with lower tail tightness because dominant eastern gray squirrels with high tail tightness were observed to exhibit a higher level of aggression (Pardo et al., 2014). We predict that intrusive gestures, outward gestures, and amount of eye contact will correlate with dominance classification in conversational settings for human subjects because these were found to be correlated with dominance in task-oriented settings for human subjects (Ridgeway, 1987). The objectives of this study are to compare the tail positioning behaviors of squirrels to their dominance classification, to compare gesture frequency and amount of eye contact in humans with their dominance classifications, and to further explore if FOXP2 could be integral to these homologous nonverbal communication behaviors. Both types of nonverbal communication indicate a form of information transfer between organisms within a population in regards to dominance. The results of this study may indicate a similar means of nonverbally communicating dominance status between squirrels and humans.

Methods

Revised Half-Draft: Brandon Llewellyn

For observing tail positioning and dominance interactions between fox squirrels (Sciurus Niger), 20 dominance interactions were observed around the Michigan State University Campus in East Lansing, Michigan; 15 interactions in the River Trail Neighborhood, 4 interactions were observed in East Neighborhood, and 1 interaction was observed in North Neighborhood. Since each interaction involves two fox squirrels, a maximum of 40 fox squirrels were observed. The fox squirrels were not tagged, so it is possible that some were recorded multiple times. The sexes of the squirrels was not determined nor recorded. When making observations, researchers did not approach the fox squirrels closer than approximately 10 meters; however, if any fox squirrels approached the researchers closer than 10 meters, the researchers attempted to stay still and not disturb the squirrel. While observing, researchers recorded video of squirrel activity including at least 5 seconds before the dominance interaction occurred and until the dominance interaction was completed. The video was analyzed at a later date where the tail tightness and tail portion bent of the dominant and subordinate squirrel were quantified along with the level of aggression. The dominant squirrel was determined to be the squirrel that aggressed towards the other squirrel, and the subordinate squirrel was determined to be the squirrel that was aggressed towards and fled. Both the tail tightness and tail portion bent characteristics were quantified according to previous research on tail position in eastern grey squirrels (Sciurus carolinensis) (Pardo et al., 2014). The aggression levels were quantified into three categories: low, medium, and high. Low aggression was categorized as an interaction where the dominant squirrel did not run directly towards the subordinate squirrel but rather in an arcing path or when the dominant squirrel stopped chasing before the subordinate squirrel fled. Medium aggression was categorized as an interaction in which the dominant squirrel ran directly towards the subordinate squirrel and stopped chasing as soon as the subordinate squirrel fled. High aggression was categorized as an interaction where the dominant squirrel ran directly towards the subordinate squirrel and continued to chase even after the subordinate squirrel fled. All researchers were present when the data was quantified in order to reduce individual bias. When recording observation, the researchers also took note of the weather conditions, time, and location. All interactions were observed between the hours of 9am and 3pm. Interaction were recorded on October 17th, 20th, 28th, 29th, and 30th of 2016 and on November 1st of 2016.

For observing nonverbal dominance communication in humans (Homo sapiens) seven conversations between two human subjects were recorded for five minutes around the MSU campus; four conversations were recorded in Holmes Hall, two conversations were recorded in the MSU Union, and one interaction was recorded in Shaw Hall. Researchers candidly video recorded conversations between two human subjects from a distance. Two researchers were required to record the conversations from two angles in order to capture the faces of both conversing subjects. Conversations captured included a mix of sexes; conversations between two females, two males, and one of each sex were captured. The video recordings were quantified when all researchers were present in order to reduce individual bias. The dominance of each subject was quantified based on posture. When positioned with broader shoulder and enlarged chest, a subject was classified as dominant compared to another subject who was positioned with more slumped shoulders and constricted posture. Eye contact was measured by recording the time in seconds in which the subject was looking at the face of the other subject. The frequency of outward and intrusive gestures were also recorded. An outward gesture was classified as a gesture where the subject’s hands raise above their shoulders with their elbows out or where their hands are farther away from their torso than the approximate length of their forearm. An intrusive gesture was classified as a gesture where the subject gestures such that their hands reach within the approximate length of their forearm of the other subject. Conversations were recorded between the hours of 12pm and 9pm. Conversations were recorded on October 12th of 2016 and on November 12th, 14th, 15th, 16th, and 17th of 2016. For both squirrel and human recordings were captured using a Canon SX610 HS, an Apple iPhone 6, or a Samsung Galaxy S 6+.

For the analysis of the squirrel data, a chi-squared test of independence was used. For each tail position characteristic, tail tightness and tail portion bent, three chi-squared tests of independence were performed: one comparing situations where the dominant squirrel had a higher quantification (D>S) than the subordinate to ones where the dominant squirrel had an equal quantification (D=S), one comparing D>S situations to situations where the dominant squirrel had a lower quantification than the subordinate squirrel (D<S), and one comparing D=S situations to D<S situations. The number of interactions at each aggression level, low, medium, and high, were used to compare the situations. For the analysis of the human data, a t-test was used to determine significance. For eye contact, the mean seconds of eye contact held for dominant subjects and for subordinate subjects was calculated from the seven conversations recorded. The mean frequency of outward gestures per five minutes and intrusive gestures per five minutes was calculated for dominant subjects and for subordinate subjects from the seven conversations recorded. Three separate t-tests were performed on the human data. A t-test was performed to compare the dominant mean to the subordinate mean for eye contact, for outward gestures, and for intrusive gestures. Statistical significance was determined when the p-value was found to be lower than 0.05.

Results

Revised Half-Draft by: Eric Zelichowski

Fox Squirrel Tail Portion Bent

The data represented in Figure 1A is from 20 different dominance interactions between fox squirrels. The data quantifies the portion of the fox squirrels’ tails that were bent according to the quantifications described herein. Levels of aggression were split into three different categories: low, medium, and high. The quantifications for aggression were also described herein. Using a chi-squared test of independence, we looked at levels of aggression and related them to tail portion bent in dominant fox squirrels compared to subordinate fox squirrels as described herein. The first chi-squared test of independence for tail portion bent compared D>S situations to D=S situations based on the quantity of interactions at each aggression level; a chi-squared value of 1.33 was calculated with a degree of freedom of 2. Additionally, we calculated a p-value of 0.513. N=8. For the second chi-squared test of independence, we compared D>S situations to D<S situations based on the quantity of interactions at each aggression level. From this, we calculated a chi-squared value of 6.74 with a degree of freedom of 2. Additionally, we calculated a p-value of 0.0344. N=14. For the final chi-squared test, we compared D<S situations to D=S situations based on the quantity of interactions at each interaction level. From this information, we calculated a chi-squared value of 3.89 with a degree of freedom of 2. Additionally we found the p-value to be 0.143. N=18. Statistical significance was only met when comparing D>S situations to D<S situations where the p-value was calculated to be 0.0344. This finding is in line with our prediction that dominant fox squirrels with lower portion bent rankings will aggress more on subordinate fox squirrels with higher portion bent rankings (Pardo et al, 2014). Due to the fact that our sample size was small, predicted results were also generated in order to better illustrate our originally expected data based on our prediction. These predicted results are displayed in Figure 1B. Staying in line with the prediction, the expected results show higher levels of aggression for scenarios where dominant fox squirrels had lower tail portion bent quantifications than subordinate fox squirrels (Pardo et al., 2014).

Fox Squirrel Tail Tightness

Figure 2A also represents collected data from 20 dominance interactions between fox squirrels. Figure 2A compares tail tightness to the quantity of interactions for each level of aggression. The tail tightness and levels of aggression quantifications have been described herein. Chi-squared tests of independence were used to determine statistical significance. The first chi-squared test compared D>S situations to D=S situations based on the quantity of interactions at each aggression level. From this analysis, we ended up calculating a chi-squared value of 0.958 with a degree of freedom of 2. Additionally, we calculated a p-value of 0.619. N=17. For the second chi-squared test we ran, we compared D>S situations to D<S situations based on the number of interactions observed at each aggression level. We calculated a chi-squared value of 2.22 with a degree of freedom of 2. Additionally, we calculated a p-value of 0.330. N=16. For our third and final chi-squared test, we looked at D<S situations compared to D=S situations based on the amount of observed interactions at each aggression level. From this analysis, we calculated a chi-squared value of 1.56 with a degree of freedom of 2. Additionally, we also calculated a p-value of 0.459. N=7. Since all of the calculated p-values were above 0.05, there were no statistically significant differences between any of the situations.Predictive data which relates aggression level to tail tightness quantifications in dominant and subordinate fox squirrels is represented in Figure 2B. We predicted that dominant fox squirrels with a higher quantification of tail tightness would agress more towards subordinate fox squirrels with lower quantification of tail tightness because dominant eastern gray squirrels with high tail tightness were observed to be more aggressive (Pardo et al., 2014). As such, our predictive data which related level of aggression to various comparisons between tail tightness in dominant and subordinate fox squirrels, levels of aggression increased as tail tightness in dominant fox squirrels increased and tail tightness in subordinate squirrels decreased.

Human Dominance

In our study, data was collected from seven conversations between human subjects and analyzed for gestural behavior and eye contact quantifications as described herein. To analyze all of the gesture and eye contact data, we utilized T-tests. We required a p-value of 0.05 or lower to declare statistical significance. Eye contact data is displayed in Figure 3A. For our eye contact analysis, we found that eye contact between individuals varied from 104.1 seconds to 261.0 seconds. After analysis, we discovered that on average dominant subjects kept 198.0 seconds of eye contact per interaction while subordinate subjects kept 144.0. We also calculated a standard deviation of 61.5 for dominant individuals and 61.3 for subordinate individuals and a standard error of 23.3 for dominant subjects and 23.2 for subordinate subjects. Additionally, we retrieved a t-statistic of 1.66 with a degree of freedom of 12.0 which leads to a p-value of 0.122 showing that there was no statistical significance between dominant and subordinate eye contact. Outward and intrusive gesture data are displayed in Figure 3B. From the data we recorded, on average, dominant individuals displayed 11 outward gestures per interaction while subordinate individuals displayed 4.43. We also calculated a standard deviation of 3.51 for dominant individuals and 2.44 for subordinate individuals and a standard error of 1.33 for dominant subjects and 0.922 for subordinate subjects. Additionally, we retrieved a t-statistic of 4.07 and a degree of freedom of 10.7 that were used to calculate a p-value of 0.00197 which shows a statistically significant difference between dominant and subordinate frequencies of outward gestures. From our data on intrusive gestures, we calculated an average of 0.429 gestures per interaction for dominant individuals and 0.0 for subordinate individuals. We also calculated a standard deviation of 0.787 for dominant subjects and 0.0 for subordinate subjects and a standard error of 0.922 for dominant individuals and 0.0 for subordinate individuals. Additionally, we calculated a t-statistic of 1.44, a degree of freedom of 6, and a p-value of 0.199 showing that the difference between intrusive gesture frequencies in our data was not statistically significant. Although not all were statistically significant, the trends in the data go along with our prediction that intrusive gestures, outward gestures, and amount of eye contact will correlate with dominance classification in conversational settings for human subjects because these were found to be correlated with dominance in task-oriented settings for human subjects (Ridgeway, 1987).

References

Allen, D. S., W. P. Aspey, 1986. Determinants of social dominance in eastern gray squirrels

(Sciurus carolinensis): a quantitative assessment. Animal Behavior 34: 81-89.

Campbell, P., R. L. Reep, M. L. Stoll, A. G. Ophir, S. M. Phelps. 2008. Conservation and diversity

of Foxp2 expression in muroid rodents: Functional implications. The Journal of Comparative

Neurology 512: 84-100.

Cristilli, C. and S. Carlomagno. 2004. Conceptual and Lexical Factors in the Production of

Speech and Conversational Gestures: Neuropsychological Evidence. International Gesture

Workshop 70-76.

Enard, W., M. Przeworski, S. E. Fisher, C. S. L. Lai, V. Wiebe, T. Kitano, A. P. Monaco, and

S. Paabo. 2002. Molecular evolution of FOXP2, a gene involved in speech and language. Nature

418: 869-872.

Farentinos, R. C. 1972. Social dominance and mating activity in the tassel-eared squirrel (Sciurus

aberti ferreus). Animal Behavior 20: 316-326.

Fisher, S. E. and G. F. Marcus. 2006. The eloquent ape: Genes, brains and the evolution of

language. Nature Reviews Genetics 7: 9-20.

Fujita, E., Y. Tanabe, A. Shiota, M. Ueda, K. Suwa, M. Y. Momoi, T. Momoi. 2007. Ultrasonic

vocalization impairment of Foxp2 (R552H) knockin mice related to speech-language disorder and

abnormality of Purkinje cells. PNAS 105: 3117-3122.

Groszer, M., D.A. Keays, R. M. J. Deacon, J. P. de Bono, S. Prasad-Mulcare, S. Gaub, M.G.

Baum, C. A. French, H. Nicod, J. A. Coventry, W. Enard, M. Fray, S. D. M. Brown,. M. Nolan, S.

Paabo, K. M. Channon, R. M. Costa, J. Eilers, G. Ehret, J. N. P. Rawlins,and S. E. Fisher. 2008.

Impaired synaptic plasticity and motor learning in mice with a point mutation implicated in human

speech deficits. Current Biology 18: 354-362.

Krauss, Robert M., Yihsiu Chen, and Purnima Chawla. 1996. Nonverbal Behavior and Nonverbal

Communication: What Do Conversational Hand Gestures Tell Us?, 392-396.

Lattal, K. A. 2001. The Human Side of Animal Behavior. The Behavior Analysis 24:147-161.

McCleery, R. A. and I. D. Parker. 2011. Influence of the urban environment on fox squirrel range

overlap. Journal of Zoology 285: 239-246.

McCloskey, R. J. and K. C. Shaw. 1977. Copulatory behavior of the fox squirrel. Journal of

Mammalogy 58: 663-665.

Pardo, M. A., S. A. Pardo, W. M. Shields. 2014. Eastern gray squirrels (Sciurus carolinensis)

communicate with the positions of their tails in an agonistic context. The American Midland

Naturalist 172.2: 359.

Willems, R. M., and P. Hagoort. 2007. Neural Evidence for the Interplay between Language,

Gesture, and Action: A Review. Brain and Language 101:278-289

Figures

Revised Half-Draft: Eric Zelichowski

Figure 1. Correlation Between Squirrel Aggression Level and Portion of Squirrel Tail Bent.
_{(A)A bar graph represents data from 20 different squirrel interactions. The portion bent quantification between dominant and subordinate fox squirrels is related to the level of aggression recorded from various interactions. In our graphs, error bars are not present because we are not looking at the calculated means from our data. Three chi-squared tests of independence were used to determine the p-value for situations relating (D>S) and (D=S), (D>S) and (D<S), and (D=S) and (D<S). For situations relating (D>S) and (D=S), a p-value of 0.513 was calculated. For situations relating (D>S) and (D<S), a p-value of 0.0344 was calculated. For situations relating (D=S) and (D<S), a p-value of 0.143 was calculated.
(B)A bar graph represents predicted results if more data had been collected.}

Figure 2. Correlation Between Squirrel Aggression Level and Tail Tightness
_{(A)This graph is based off of data collected from 20 different squirrel-squirrel interactions.The tail tightness quantification between dominant and subordinate fox squirrels is related to the level of aggression recorded from various interactions. There are no error bars in this graph, due to the fact that we are not looking at approximated population means. Three chi-squared tests of independence were used to determine the p-value for situations relating (D>S) and (D=S), (D>S) and (D<S), and (D=S) and (D<S). For situations related (D>S) and (D=S), a p-value of 0.619 was calculated. For situations relating (D>S) and (D<S), a p-value of 0.330 was calculated. For situations relating (D=S) and (D<S), a p-value of 0.143 was calculated.
(B)This graph depicts the predicted results of our study if more data had been collected.}

Figure 3. Nonverbal Dominance Communication Frequency in Humans
_{(A) In this graph, seven trials of eye contact observations were averaged. Eye contact was determined between two subjects. Dominance classifications were made based on the postures of the subjects. This data was also collected over a five minute period. Error bars are represented in the graph by standard error. A t-test was performed on the data to determine significance. The p-value was found to be 0.122. (B) In this graph, seven trials pertaining to frequency of outward and intrusive gestures were averaged. Dominance classifications were made based on the postural positions of the subjects. This data was collected over a five minute period. Error bars are represented in this graph by standard error. A t-test was performed on the data to determine significance. The p-value was calculated to be 0.0019 for outward gestures and 0.199 for intrusive gestures.}

Figure 4. Information Transfer and Homologous Studies in Biology: (Download Here)

Appendix

Original Half-Draft