USDA Herding Dog Study

@Linda Rohem, reprinted with permission, originally published in AHBA newsletter

When reading the book The New Knowledge of Dog Behavior by Clarence Pfaffenberger, I was intrigued to find a reference to a project involving herding dogs that had taken place just before World War II under the auspices of the U.S. Department  of Agriculture.  The study was conducted by Dr. Walker M. Dawson at the USDA facility in Beltsville, Maryland.

I tried to track down information regarding this project, but a letter to the USDA in Beltsville, Maryland brought no reply.  However, an English Shepherd breeder with whom I corresponded had been in con­tact with Dr. Dawson some years before, and sent me a copy of information sent to her. Their correspondence had been in 1975, and the material he enclosed with his letter to her was dated 1965 with an identification number of ARS 44-163.   It was about 3-1/2 pages long and titled Studies of the Inheritance of Intelligence and Temperament in Dogs.

The introductory paragraph states, “In the fall of 1935, the Bureau of Animal Industry of the U.S. Dept. of Agriculture started work on a project to study the inheritance of intelligence  and temperament in farm animals and their relation to other characters of economic importance.   Dogs were used as the experi­mental animals.  Sheep herding  was studied as one of the manifestations of intelligence since it was of economic  importance  in agriculture.” One of the breeds used was the Puli (a Puli breeder told me that this came about because of the interest in the breed by a member of the Roosevelt family);  Pulik were also crossbred to the German Shepherd and the Border Collie (as examples of other herding dogs), and with the Chow Chow (to be used as a control). Later, a pair of Turkish sheep dogs and a litter from the pair were included. According to the Pfaffenberger book, English Shepherds were involved.  “Dr. Dawson had assembled,  after great care, a colony of the dogs which we used to have on the farm and called shepherds, English shepherds, or farm collies.”  The material Dr. Dawson sent to my acquaintance mentions the Border Collie rather than the English Shepherd, however,  so it is unclear whether there was some confusion between English Shepherds and Border Collies, or whether both were used.

Dr. Dawson’s material continues, “The tests given to the dogs were:  Learning to walk on a leash at 4 mos. old; reactions toward sheep, without being trained, at 6 mos.; learning to lie down to one whistle signal and get up and come to the trainer at another whistle, at 8 mos.; training to herd sheep with the trainer at 10 mos.; reactions to a stranger at 11 mos.; reactions in a multiple-choice apparatus at 1 year.

“These tests gave results on many characteristics,  but because of the war it was necessary to discontinue the project… it has not been possible to analyze most of the data or to draw conclusions as to the value of the tests.  A wide variation was found in the reactions of the dogs in all the tests…

“Results from the ‘puppy sheep test,’ in which all or part of a litter of puppies was placed in a small paddock with a few sheep at 6 mos. of age, indicated quite strongly that dogs which were aggressive toward the sheep at this age remained so unless they were controlled through later training or were made nonaggressive by being hurt by the sheep. Some of the puppies which were afraid of the sheep at 6 mos., however, later became aggressive toward them. Results indicate that dogs or puppies which were afraid and timid toward the trainer and people were not necessarily  so toward the sheep, and vice versa.”

“Preliminary analysis on the sheep herding test showed that the dogs differed so much in their reactions toward the sheep that they had to be divided  into four groups for training and testing:   (1) dogs which were so aggressive toward sheep that they had to be brought under control by training before they could be allowed off leash with the sheep; (2) dogs which were interested in the sheep but either were not very aggressive or were very easily controlled;  (3) dogs which had to have their interest in the sheep stimulated  or their courage bolstered by having another dog with them; and (4) dogs which had so little interest in the sheep or were so frightened by them that their interest could not be stimulated  by use of another dog.  An approximate separation of all the dogs (131) into the four groups gave 18, 41, 16 and 25 percent, respectively. The distribution of the dogs by breed groups, however, differs quite materially from the total. Thus, only 11 percent  of the Pulis were in the first group and 36 percent of the Pulis were in the last group. 37 percent  of the German Shepherd X Puli crossbreeds were too aggressive and 12 percent showed no interest in the sheep. With our conditions and each dog being given a relatively short training period, approximately 50 percent of the dogs in group 1, 37 percent of the dogs in group 2, and 5 percent of the dogs in group 3 showed definite promise of becoming at least fair sheep dogs if given adequate training.  This is approximately one-fourth of all the dogs tested. From our experience in giving additional training in herding to some of the promising dogs, we doubt if we would have been able to make good sheep dogs out of more than half the dogs we thought showed promise.

“Quite marked differences were noticeable in the tendency of dogs of the different  breeds to harm the sheep under the conditions of the sheepherding test. It is believed this is due to inherited differences in the way the dogs bite. The Collies, for example, seemed to have a tendency to nip the sheep much of the time, rather than really biting them, and thus did little damage. At least some of the German Shepherd X Puli crossbreeds seemed to slash and tear when they bit, and this was apt to result in serious wounds.  The Turkish dogs, while they sometimes grabbed the sheep, tended to hold on without tearing and seldom injured the sheep.”

Results also were given with regard to the other tests mentioned, but with no comments regarding breeds other than with reference to the multiple-choice apparatus. This test involved the dogs finding which one of four doors was unlocked, with the escape door never being the same twice in succession, to assess which dogs were able to find their way out with less repetition. It was stated that the Puli X GSD’s scored significantly higher than the purebred Pulis or the Puli X Chows.

In his accompanying letter, Dr. Dawson described the setup of the sheep test:
“I usually tested several puppies from a litter together. I placed the puppies in a pen 50 to 100 ft. square with 3 sheep and made notes of their reactions. I got out of sight and watched them through a window. Look particularly for interest in the sheep, courage or aggressiveness and what the shepherds call ‘eye’ (the pup keeps its eyes fixed on the sheep which it should approach at least a little way toward them while they face it).  Dogs may change as they grow up so the second test should be when the dog is about a year old … in addition look for ability or willingness to obey commands even when at some distance from the trainer. I trained the dogs to go with the trainer on a leash, to heel, to lie down to one note on a whistle and to come to the trainer to two notes on a whistle before we tried the dog with sheep . . . We believe that dogs that are aggressive but can be controlled even when excited and at a distance from the trainer and with a strong eye are the best prospects for further work.  To train a top notch dog takes a tremendous  amount of work and time so one can train only a few such dogs. How such a complex combination  of the right amount of aggressiveness with the right sensitivity to control is inherited I can not guess.  Even in the sheep herding breeds I understand the shepherds find at best only a few which can become good herders.”

Ultimately, the project was closed down due to America’s entry into the war, and the USDA never published results of the study.