WITH THE EEADER. 467 



trouble. But the experiment upon the colt was so success- 

 ful, he driving around in the building without breeching, 

 entirely gentle and fearless, the people requested that he 

 be driven outside. My assistant, not being able to explain 

 the principles and conditions of success so that they could 

 understand (a dangerous omission which I feared he would 

 make, but which I supposed I had guarded against), finally 

 consented, being persuaded he would have no trouble in 

 doing so. When the doors were opened, the horse drove 

 all right until he got just outside, when he made a spring 

 forward, pulled away, and made directly for the canal, which 

 was close by. He jumped from the bank to the bottom, a 

 distance of twenty-two feet, crossed, and soon disappeared 

 down the tow-j)ath. Being apprised of the trouble, I told 

 the owner I would pay any damages sustained by the horse, 

 and gave orders to have him caught and brought back. 

 Fortunately the horse was unharmed, and I again subjected 

 him to treatment, when he was driven without difficulty. 



I made the incident an opportunity for explaining that 

 a very common cause of failure is in supposing that because 

 a horse has submitted to treatment in one place, and ap- 

 pears perfectly docile, he will prove equally so in all places ; 

 that there would be no more difficulty in driving a horse 

 out of doors than in a building, provided he could be worked 

 there as in the building. But the difficulty was, we could 

 not do this on account of the crowd that always hung 

 around, making it impossible to have sufficient privacy to 

 subject him to treatment. 



The public, somehow, could not understand this condi- 

 tion, and it was continually the cause of more or less trouble 

 to me. A very bad case would frequently be brought for- 

 ward to be experimented upon, when by subjecting to treat- 

 ment he could be made to submit to being hitched up and 

 driven around in the building perfectly manageable ; but 



