CHIMPANZEES. 31 



the simiine intelligence by a series of special experiments. It was found, however, 

 that such experiments were seriously hampered by the effects on the creature of the 

 visits of the numbers of people who were constantly passing in and out of the room 

 in which she was kept ; and there is consequently but little doubt that, under more 

 favourable circumstances, the results obtained would have been more remarkable 

 than they are. Dr. Romanes, having secured the assistance of the keepers, caused 

 them to ask " Sally " repeatedly for one, two, or three straws, which she was to pick 

 up and hold out from among the litter strewing her cage. Tlie number of straws 

 asked for was constantly varied, and never followed any regular order ; and when 

 the correct number was presented the animal was rewarded by a piece of fruit, 

 while if the number was incorrect her offer was refused. " In this way," observes 

 Dr. Romanes,! « i^-^^ ^-^^^ y^.^^ taught to associate these three numbers witli their 

 names. Lastly, if two or three straws were demanded, she was taught to hold one 

 or two in her mouth until she had picked up tlie remaining straw, and then to 

 hand the two or three straws together. This prevented any error arising from her 

 interpretation of vocal tones. As soon as the animal undei-stood what was required, 

 and had learnt to associate these three numbers with tlieir names, she never failed 

 to give the number of straws asked for. Her education was then extended in a 

 similar manner from three, to four, and from four to five straws. Here I allowed 

 her education to terminate. But more recently one of the keepers has endeavoured 

 to advance her instruction as far as ten. The result, however, is what might have 

 been anticipated. Altliough she very rarely makes any mistake in handing out 

 one, two, three, four, or five straws, according to the number asked for, and although 

 she is usually accurate in handing out as many as six or seven, when the numbers 

 eight, nine, or ten are named, the result becomes more and more uncertain, so as to 

 be suggestive of guess-work. It is evident, however, that she understands the 

 words seven, eight, nine, and ten to betoken numbers higher than those below them ; 

 and if she is asked for any of these numbers, she gives some number that is above 

 six and not more than ten ; but there is no such constant accuracy displayed in 

 handing out the exact number named, as is the case below six. On the whole, then, 

 while there is no doubt that this animal can accuratelj' compute any number of 

 straws up to five, beyond five the accuracy of her computation becomes progressively 

 diminished. It is to be noticed that the ape exhibits some idea of multiplication , 

 for she very frequently doubles over a long straw so as to make it apjaear as two 

 straws. Any of the rare errors which she makes in dealing with numbers below 

 six are almost invariably due to her thus endea-vouring to duplicate her straws. In 

 this connection it is to be remembei-ed that, owing to the method above described, 

 when any high number is demanded, a considerable tax is imposed upon her 

 patience ; and as her movements are deliberate, while her store of patience is small, 

 it is evident that the doubling of the straws is intended to save trouble by getting 

 the sum completed with greater rapidity. Of course we do not recognise these 

 doubled straws as equivalent to two straws, and therefore the persistency with 

 which she endeavours to palm them off as such is the more noteworthy. Moreover, 

 I am disposed to think that the uncertainty which attends her dealing with the 

 numbers six and seven is more due to her losing patience than to her losing count ; 



1 We have somewhat abbreviated the extract. 



