ARITHMETICAL POWEES OF ANIMALS. 281 



two, and three dark bands. This we continued for 

 another three months, or, say, allowing for intermissions, 

 ten weeks ; but, to my surprise, entirely without success, 

 for we altogether failed to make Van understand what 

 we wanted. I was rather disappointed at this, as, if 

 it had succeeded, the plan would have opened out many 

 interesting lines of inquiry. Still, in such a case one 

 ought not to wish for one result more than another, 

 as, of course, the object of all such experiments is 

 merely to elicit the truth, and our result in the present 

 case, though negative, is very interesting. I do not, 

 however, regard it as by any means conclusive, and 

 should be glad to see it repeated. If the result proved 

 to be the same, it would certainly imply very little 

 power of combining even extremely simple ideas. 



Can Animals count ? 



I then endeavoured to get some insight into the 

 arithmetical condition of the dog's mind. On this 

 subject I have been able to find but liltle in any of 

 the standard works on the intelligence of animals. 

 Considering, however, the very limited powers of 

 savage men in this respect — that no Australian 

 language, for instance, contains numerals even up to 

 four, no Australian being able to count his own fingers 

 even on one hand — we cannot be surprised if other 

 animals have made but little progress. Still, it is 

 curious that so little attention should have been 

 directed to this subject. Leroy, who, though he ex- 

 presses the opinion that " the nature of the soul of 

 animals is unimportant," was an excellent observer, 

 mentions a case in \Ahich a man was anxious to shoot 



