On the other hand, it is also true that Mr. von Osten knew nothing whatever of the finer, more minute signals, such as the final jerk, the head-movement upward, downward, etc., and it is difficult to conceive how he might have gained any knowledge of them. We might perhaps conceive of four possible sources. He might have come upon them by chance. But it is extremely improbable that in the million of possible forms of signaling he should have hit upon those that at the same time represent the natural expressive movements. Or he might have derived a knowledge of them through a study of the pertinent literature. I have searched diligently for such a source, in both the old and the modern literature, but in vain. From the sixteenth century on, there is a series of accounts of horses that were able to spell and to solve problems in arithmetic, and the reports on learned dogs go back even to the time of Justinian, in the middle of the sixth century.[107] All of these animals were kept for purpose of speculation and were exhibited for pecuniary reasons only. Nor does one read that any person could work with these animals off-hand, which was the characteristic feature of the Osten horse.[AN] In many cases we find mention made of the signs to which the animals reacted. Thus for the beginning or stopping of the animal's scraping or tapping, the signals were respectively raising and lowering of the eyes on the part of the trainer,[113] lowering and raising of the whip[114] or of the arm, stepping forward and backward,[115] and as a closing signal a slight bending forward.[116] The signals for beginning and ceasing to bark in the case of dogs, were the trainer's commands to "speak", and, at the same time, his looking at the dog, and then looking away for a closing sign;[117] or a mouth-movement on the part of the trainer and then a withdrawing of the left hand which had been resting on the hip.[118] Among the signals for nodding and shaking the head we find the following mentioned: raising and lowering the hand or arm[119] or the whip;[120] a movement of the hand toward the horse's nose, as a signal for nodding, and an arm-movement as a signal for shaking the head.[121] For this last, we find recommended also a slight breathing upon the animal,[122] and--in the case of dogs--a mouth-movement simulating blowing, or a turn of the fingers.[123] (We will not dwell upon the many signals for selecting objects, which are mentioned, since we have already discussed this point on page 230f). In all these instances it is plain that we have to do with purely voluntary and "artificial" signals. The only example of involuntary signs which Mr. von Osten could have found in literature, was that of Huggins's dog, which need not be considered here, since, as was said on page 177, the really effective signs in that case were not discovered. A third means by which Mr. von Osten might have gained a knowledge of the involuntary, natural expressive signs, would have been