ii PREFACE TO THE SECOND EDITION. 



simple and straightforward explanation of 

 the evidence afforded by the letters, we must 

 choose one or other of the following solutions 

 of the problem. These stories may of course 

 all be (i) hoaxes ; (2) instances of faulty 

 observation ; (3) coincidences, i.e., the chance 

 linking of the utterance of certain words 

 with certain acts by dogs, though in reality 

 the words and acts had nothing to do with 

 each other ; (4) hallucination on the part of 

 the reporters they were so convinced that 

 the dogs could understand, that they ima- 

 gined the dogs to do things which in reality 

 they never did. Now, it must be observed 

 that each of these explanations would be, 

 primd facie, perfectly tenable if there were 

 only one story of a dog understanding human 

 speech. But it is difficult to see how any of 

 them will hold good for all the instances re- 

 corded, both in the new and in the old portion 

 of the present book. It cannot seriously be 

 maintained that all the stories are hoaxes, 

 or that all the reporters were incapable 

 of relating what they saw, or, again, that 

 all the facts stated were due to mere co- 



