1000 Cassius Stare.jpg

As part of Dr. Beach's long-term study of the behavioral endocrinology of sexually differentiated behaviors in dogs (conducted at the University of California, Berkeley), we investigated sex differences in social rank and aggressiveness. The findings proved to be so rewarding, that observational experiments on the development of social hierarchies became the focus of the research program for nearly a decade.

Social rank was assigned primarily from results of dyadic bone tests — observations of two dogs with one bone, and from general observations of the pack at large, e.g., which dogs had prime access to other valued and limited commodities such as, favorite resting places, food pans, water supply, toys etc. Standard dyadic bone tests comprised two parts: firstly, the Equal Opportunity test, in which the bone (a large meaty ox tail) was thrown to land equidistant between the two dogs and then, the ensuing interactions were observed for 90 seconds (e.g., possession, approach, vocalizations, threats etc.); and secondly, the Affirmative Action test, wherein the bone was given to the underdog for 30 seconds and then the higher ranking individual was released from a holding cage and the interactions were observed for a further 90 seconds.


Male-Male Dyads
An examination of inter-male relationships demonstrated an absolutely linear hierarchy, with extremely infrequent variation during the nearly ten years of testing. Regardless of the specific situation, ownership of a bone, or any valued commodity, was almost always pre-decided by the rank of the adult dogs concerned.

Female-Female Dyads

Adult females had a similar linear hierarchy but compared with males, they exhibited greater variation from day to day. Female success was often dependent on the specific situation. For example, if both females had equal opportunity to take possession of the bone, its ownership was decided primarily by rank. However, if one bitch had already established possession, it was not uncommon for her to successfully defend the bone, even against much higher-ranking females.

Male-Female Dyads
In view of their greater size and strength, it was not surprising to find that males were generally higher ranking than females and won nearly 80% of the tests. Even so, some females were very high ranking and in some groups, the top dog was female. Also, if a bitch had already established possession, it was not uncommon for her to successfully defend the bone, even against much higher-ranking males. It appeared, that “possession” was nine-tenths of female law. Indeed, it would seem that female dogs often invoke the First Bitch Amendment to Male Hierarchical Law, namely, “I have it and you don’t!”

Group Structure

Males and females had separate distinct hierarchies within the group hierarchy — an almost fixed linear hierarchy for males and a more flexible linear hierarchy for females. Within the group as a whole, the rank order of males was always the same and the order of females was usually the same, but the relationship between males and females often varied from day to day, such that male and female hierarchies slid up and down and/or expanded or shrunk relative to each other.  For instance, the top two dogs were usually Ken and Eddie, with Cassius and Joan both vying for third place. However, on occasions, high-ranking females held all the top three positions. It was as if on some days the females just told the males not to push it ...and they didn't.

By and large the social machinations of the group ran fairly smoothly and the outcome of most doggy disagreements was pre-decided by relative rank. Noisy and protracted squabbles were infrequent and largely restricted to the middle-order males. Pilo-erection, growling, snarling, snapping and fighting were indicative of the insecurity and lack of confidence normally associated with middle-ranking dogs, especially males. Obviously, high-ranking dogs — the true top dogs — have little occasion to threaten, or fight, and low-ranking dogs would be silly to do so.

Middle-ranking male dogs wasted a lot of time “posturing” over the bone, showing off and loudly advertising their prize. High-ranking dogs and females, however, just got down to the important business at hand — chewing the bone. Moreover, confident in their rank and possession, most high-ranking male dogs were quite willing to share a bone with others, although by and large, only females were willing to try and share a high-ranking male's bone. Typically, a bitch would voluntarily relinquish possession as soon as an high-ranking male sauntered up to chew the meaty end, whereupon the female would recommence chewing, but this time, at the thin boney end of the ox-tail bone.  Although some bitches were eager to share a higher-ranking male's bone, bitches were not inclined to share their bone with any other dog. In fact, many bitches could be extremely successful at keeping a bone from dogs that were much higher-ranking. For example, when approached by a stiff-legged, growly-wowly, middle-ranking male, it was not uncommon for a bitch to hunker down, continue tucking in, and to simply ignore the male's macho protestations until he exhausted his growling and eventually wandered away.

The various dyadic relationships represent the building blocks for group structure. However, observations of the group as a whole revealed additional variations and complexities, generally favoring lower-ranking individuals. For example, when the bone was thrown to the group of twenty dogs, opportunistic lower-ranking animals would often get a few bites of meat before a higher-ranking dog expropriated the bone. Ken, Joan and Doris had an interesting triadic relationship. Doris would always be the first to get to the bone and would manage a few gobbles prior to Joan's charge. Ken would then saunter up to Joan and casually take possession of his bone, whereupon Joan would stiffly retreat, and Doris would approach with rump wagging and tail between her legs to share the bone with her good buddy Kenny.

Although the social hierarchy provided general guidelines for group interaction, there were many unique rank-reversals, especially between dogs that were close friends. Joe and Cassius were an interesting pair. Born just one week apart, they had always roomed together: they ate together, played together and always slept together and... whenever there was a disturbance in the pack, it usually involved Cassius growling and snarling at Joe. Joe rarely responded to these displays and usually left Cassius to stamp his little paws and have his tiny temper tantrums. Although Cassius was generally regarded as being higher-ranking to Joe, it suspiciously seemed California Joe was just too cool to care that much about possessions and consumerism and for the most part, if Cassius (with yon lean and hungry look) was going to get his knickers in a twist over a measly bone, then Joe was prepared to let him have it. However, twice I saw Joe stop Cassius mid-growl with nothing more than a mere glance.     

At one interesting point in the study, immediately after the Top Dog Ken had died of old age, the rank order of the top five male dogs showed a perfect inverse correlation with bodyweight, i.e., the top dog, “Fast” Eddie was the smallest, the #2 dog Cassius was the second smallest, middle-ranking Joe was the middle-sized dog, down to Whip — the underdog — the largest in the pack! Investigating established hierarchies sometimes produces surprising results. The above hierarchy was simply a fluke. The more important variable was age and indeed, the male hierarchy had a 100% direct correlation with age.  And of course, age dictates relative weight when growing up. For example, as adults, Whip was more than twice the weight of Eddie but when Whip was learning the ropes at eight weeks of age, Eddie was already an experienced six-year old and four or five times Whip’s size.

Our interesting male hierarchy prompted us to retest with several other groups of adult male dogs and then to start a developmental study.


When a hungry neonate is confronted with an occupied teat, he has a couple of options: to supplant the resident nurser, or, to search for an unoccupied teat. An analysis of “teat expropriations” and “teat defenses” revealed a rudimentary linear hierarchy as early as two weeks of age. High rank was strongly correlated with body weight. Larger pups were able to supplant others and hence had primary access to the bitch's milk thus, further increasing their weight advantage. Rank was also correlated with sex; male pups (often heavier) were usually higher ranking than females.

Intra-Litter Pup-Pup Relationships
Dyadic and group bone tests revealed, that within each litter, both the top dog and underdog were irrefutably established by eight weeks of age. Top- and bottom-ranking dogs have unique social positions, since both may generalize about their social relationships. The top dog assumes he is higher ranking than the rest of the pups, and the bottom dog learns that he is lower ranking to all. It was surprisingly common for both the top dog and underdog to be male. However, middle-order pups experience a more complicated social scenario, since they are higher ranking to some individuals but lower ranking to others. The middle-order relationships were not firmly established until the pups were three months of age, where after each litter had a stable linear hierarchy, with rank strongly correlated with sex and weight — male pups and/or heavier pups tending to be higher ranking.

Inter-Litter Pup-Pup Relationships

Each litter had grown up with mum in individual indoor/outdoor runs. When three of the litters were 10-, 12- and 16-weeks-old respectively, they were transferred into a large outdoor living area along with 12 adult dogs, (including the sires and dams). Relationships between littermates remained stable. Initially, litter rank had some carry-over effect, determining the puppies overall ranking in the large group. For example, the top dogs from each litter became the three highest ranking pups of the new group and similarly, the underdogs from each litter became the three lowest ranking dogs of the litters combined. Once the puppies got to know each other however, after a week of living together, some low-ranking pups from older litters realized they were higher ranking than some of the high-ranking pups from younger litters. Being a big fish in a little pond, or a little fish in a little pond, appeared to have a temporary influence on the pups' interactions with other puppies. But after socializing with numerous puppies of various ages, the most important determinants of relative rank between pups from different litters were age and sex; male and/or especially, older puppies were usually higher ranking. At this age, even a small difference in age represented a considerable weight advantage.

Pup-Adult Relationships

One day after testing, we noticed pups competing for a bone in the group enclosure. The top puppy James won the bone, whereupon all the other dogs circled James and watched him chewing. Obviously, none of the pups were going to take the bone away and it seemed like the adult dogs were somehow inhibited from doing so. After a while, Doris, a low ranking but opportunistic female, inched forward and gently tugged at the bone. James growled, snarled, flustered and blustered, but eventually gave up the gambit. The instant James relinquished possession, the bone rapidly changed paws a number of times — from Doris to Joan to Cassius and eventually, ending up with Eddie, the top dog. The above incident prompted a two-year series of bone tests between adults and pups (from these and other litters).

In Equal Opportunity tests, adult dogs always captured the bone and never even let the pups come close. However, in the early Affirmative Action tests with two-month-old puppies starting in possession of the bone, adult females never attempted to take the bone away and adult males only did so in 40% of the tests. By the time the “pups” were six months old, however, adult females expropriated the bone in 60% of tests and adult males always took it away. It was apparent that adult dogs, bitches especially, showed leniency towards young pups in social situations. The termination of this “puppy license” is cued by rising testosterone levels in male pups at four- to five-months of age, which reach a peak around 10 months (4-5 ng/ml) before declining to adult levels (1-2 ng/ml).
When puppies approached adolescence, they were continually harassed by adult dogs. Adult males especially targeted male adolescents. This stressful phase of social development is mercifully short, because the pups quickly learn to display active and exaggerated appeasement in order to allay harassment by adults, i.e., the pups learn their station in life before they become serious competition on the social scene. Even so, several maturing adolescents, especially the high-ranking males, started to challenge older, low-ranking females. In our studies, all challenges against adult males were unsuccessful, even though, when full grown, most of the new generation turned out to be larger than the old guard.

Developmental rank-reversals are more likely to occur on the domestic scene with the enormous disparity in size between different breeds and with human intervention. Human intervention tends to exacerbate most social problems. Nonetheless, it is not uncommon for an extremely small adult dog to maintain higher rank over a much larger, but younger dog of a different breed.

Relative size and strength is the most important determinant of rank at each stage in development. However, once pups (of the same breed) have grown up and assumed their relative positions within an established adult hierarchy, there may be no correlation between rank and adult weight. Social hierarchies must always be viewed in a developmental context. Indeed, the above-mentioned adult male hierarchy, which was negatively correlated with adult weight, showed a perfect positive correlation with age. Thus, although in adulthood, Cassius was larger than Eddie, for the first six months of Cassius' life, he was a mere slip of a pup compared with three-year-old Eddie.

Developmental Nolo Contendre! is the name of the game. Puppies grow up within an established social culture where they simply cannot compete with older animals. Thus, puppies become a part of the culture.

Certainly, puppies may squabble and scrap amongst themselves when establishing hierarchies. But hierarchies exist to prevent disagreements and fights between adult dogs. Moreover, adult hierarchies are absolutely not maintained by aggression. On the contrary, hierarchies exist to prevent aggression. To cite canine social structure as rationale for employing human aggression and dominance in dog training is completely unfounded.

This article is based on Dr. Dunbar's Behavior column in the July 1989 issue of the American Kennel Gazette. Reprinted with permission of the author and the American Kennel Club.